From dbda92d16f8655044e082930e4e9d244b87fde77 Mon Sep 17 00:00:00 2001 From: "Bu, Yitian" Date: Mon, 18 Feb 2013 12:53:37 +0000 Subject: printk: Fix rq->lock vs logbuf_lock unlock lock inversion commit 07354eb1a74d1 ("locking printk: Annotate logbuf_lock as raw") reintroduced a lock inversion problem which was fixed in commit 0b5e1c5255 ("printk: Release console_sem after logbuf_lock"). This happened probably when fixing up patch rejects. Restore the ordering and unlock logbuf_lock before releasing console_sem. Signed-off-by: ybu Cc: Peter Zijlstra Cc: stable@vger.kernel.org Link: http://lkml.kernel.org/r/E807E903FE6CBE4D95E420FBFCC273B827413C@nasanexd01h.na.qualcomm.com Signed-off-by: Thomas Gleixner --- kernel/printk.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/printk.c b/kernel/printk.c index 267ce780abe8..e698e80d8428 100644 --- a/kernel/printk.c +++ b/kernel/printk.c @@ -1358,9 +1358,9 @@ static int console_trylock_for_printk(unsigned int cpu) } } logbuf_cpu = UINT_MAX; + raw_spin_unlock(&logbuf_lock); if (wake) up(&console_sem); - raw_spin_unlock(&logbuf_lock); return retval; } -- cgit v1.2.3 From 45ceebf77653975815d82fcf7cec0a164215ae11 Mon Sep 17 00:00:00 2001 From: Paul Gortmaker Date: Fri, 19 Apr 2013 15:10:49 -0400 Subject: sched: Factor out load calculation code from sched/core.c --> sched/proc.c This large chunk of load calculation code can be easily divorced from the main core.c scheduler file, with only a couple prototypes and externs added to a kernel/sched header. Some recent commits expanded the code and the documentation of it, making it large enough to warrant separation. For example, see: 556061b, "sched/nohz: Fix rq->cpu_load[] calculations" 5aaa0b7, "sched/nohz: Fix rq->cpu_load calculations some more" 5167e8d, "sched/nohz: Rewrite and fix load-avg computation -- again" More importantly, it helps reduce the size of the main sched/core.c by yet another significant amount (~600 lines). Signed-off-by: Paul Gortmaker Acked-by: Peter Zijlstra Cc: Frederic Weisbecker Link: http://lkml.kernel.org/r/1366398650-31599-2-git-send-email-paul.gortmaker@windriver.com Signed-off-by: Ingo Molnar --- kernel/sched/Makefile | 2 +- kernel/sched/core.c | 569 ------------------------------------------------- kernel/sched/proc.c | 578 ++++++++++++++++++++++++++++++++++++++++++++++++++ kernel/sched/sched.h | 8 + 4 files changed, 587 insertions(+), 570 deletions(-) create mode 100644 kernel/sched/proc.c (limited to 'kernel') diff --git a/kernel/sched/Makefile b/kernel/sched/Makefile index deaf90e4a1de..54adcf35f495 100644 --- a/kernel/sched/Makefile +++ b/kernel/sched/Makefile @@ -11,7 +11,7 @@ ifneq ($(CONFIG_SCHED_OMIT_FRAME_POINTER),y) CFLAGS_core.o := $(PROFILING) -fno-omit-frame-pointer endif -obj-y += core.o clock.o cputime.o idle_task.o fair.o rt.o stop_task.o +obj-y += core.o proc.o clock.o cputime.o idle_task.o fair.o rt.o stop_task.o obj-$(CONFIG_SMP) += cpupri.o obj-$(CONFIG_SCHED_AUTOGROUP) += auto_group.o obj-$(CONFIG_SCHEDSTATS) += stats.o diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 58453b8272fd..bfa7e77e0b50 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -2056,575 +2056,6 @@ unsigned long nr_iowait_cpu(int cpu) return atomic_read(&this->nr_iowait); } -unsigned long this_cpu_load(void) -{ - struct rq *this = this_rq(); - return this->cpu_load[0]; -} - - -/* - * Global load-average calculations - * - * We take a distributed and async approach to calculating the global load-avg - * in order to minimize overhead. - * - * The global load average is an exponentially decaying average of nr_running + - * nr_uninterruptible. - * - * Once every LOAD_FREQ: - * - * nr_active = 0; - * for_each_possible_cpu(cpu) - * nr_active += cpu_of(cpu)->nr_running + cpu_of(cpu)->nr_uninterruptible; - * - * avenrun[n] = avenrun[0] * exp_n + nr_active * (1 - exp_n) - * - * Due to a number of reasons the above turns in the mess below: - * - * - for_each_possible_cpu() is prohibitively expensive on machines with - * serious number of cpus, therefore we need to take a distributed approach - * to calculating nr_active. - * - * \Sum_i x_i(t) = \Sum_i x_i(t) - x_i(t_0) | x_i(t_0) := 0 - * = \Sum_i { \Sum_j=1 x_i(t_j) - x_i(t_j-1) } - * - * So assuming nr_active := 0 when we start out -- true per definition, we - * can simply take per-cpu deltas and fold those into a global accumulate - * to obtain the same result. See calc_load_fold_active(). - * - * Furthermore, in order to avoid synchronizing all per-cpu delta folding - * across the machine, we assume 10 ticks is sufficient time for every - * cpu to have completed this task. - * - * This places an upper-bound on the IRQ-off latency of the machine. Then - * again, being late doesn't loose the delta, just wrecks the sample. - * - * - cpu_rq()->nr_uninterruptible isn't accurately tracked per-cpu because - * this would add another cross-cpu cacheline miss and atomic operation - * to the wakeup path. Instead we increment on whatever cpu the task ran - * when it went into uninterruptible state and decrement on whatever cpu - * did the wakeup. This means that only the sum of nr_uninterruptible over - * all cpus yields the correct result. - * - * This covers the NO_HZ=n code, for extra head-aches, see the comment below. - */ - -/* Variables and functions for calc_load */ -static atomic_long_t calc_load_tasks; -static unsigned long calc_load_update; -unsigned long avenrun[3]; -EXPORT_SYMBOL(avenrun); /* should be removed */ - -/** - * get_avenrun - get the load average array - * @loads: pointer to dest load array - * @offset: offset to add - * @shift: shift count to shift the result left - * - * These values are estimates at best, so no need for locking. - */ -void get_avenrun(unsigned long *loads, unsigned long offset, int shift) -{ - loads[0] = (avenrun[0] + offset) << shift; - loads[1] = (avenrun[1] + offset) << shift; - loads[2] = (avenrun[2] + offset) << shift; -} - -static long calc_load_fold_active(struct rq *this_rq) -{ - long nr_active, delta = 0; - - nr_active = this_rq->nr_running; - nr_active += (long) this_rq->nr_uninterruptible; - - if (nr_active != this_rq->calc_load_active) { - delta = nr_active - this_rq->calc_load_active; - this_rq->calc_load_active = nr_active; - } - - return delta; -} - -/* - * a1 = a0 * e + a * (1 - e) - */ -static unsigned long -calc_load(unsigned long load, unsigned long exp, unsigned long active) -{ - load *= exp; - load += active * (FIXED_1 - exp); - load += 1UL << (FSHIFT - 1); - return load >> FSHIFT; -} - -#ifdef CONFIG_NO_HZ_COMMON -/* - * Handle NO_HZ for the global load-average. - * - * Since the above described distributed algorithm to compute the global - * load-average relies on per-cpu sampling from the tick, it is affected by - * NO_HZ. - * - * The basic idea is to fold the nr_active delta into a global idle-delta upon - * entering NO_HZ state such that we can include this as an 'extra' cpu delta - * when we read the global state. - * - * Obviously reality has to ruin such a delightfully simple scheme: - * - * - When we go NO_HZ idle during the window, we can negate our sample - * contribution, causing under-accounting. - * - * We avoid this by keeping two idle-delta counters and flipping them - * when the window starts, thus separating old and new NO_HZ load. - * - * The only trick is the slight shift in index flip for read vs write. - * - * 0s 5s 10s 15s - * +10 +10 +10 +10 - * |-|-----------|-|-----------|-|-----------|-| - * r:0 0 1 1 0 0 1 1 0 - * w:0 1 1 0 0 1 1 0 0 - * - * This ensures we'll fold the old idle contribution in this window while - * accumlating the new one. - * - * - When we wake up from NO_HZ idle during the window, we push up our - * contribution, since we effectively move our sample point to a known - * busy state. - * - * This is solved by pushing the window forward, and thus skipping the - * sample, for this cpu (effectively using the idle-delta for this cpu which - * was in effect at the time the window opened). This also solves the issue - * of having to deal with a cpu having been in NOHZ idle for multiple - * LOAD_FREQ intervals. - * - * When making the ILB scale, we should try to pull this in as well. - */ -static atomic_long_t calc_load_idle[2]; -static int calc_load_idx; - -static inline int calc_load_write_idx(void) -{ - int idx = calc_load_idx; - - /* - * See calc_global_nohz(), if we observe the new index, we also - * need to observe the new update time. - */ - smp_rmb(); - - /* - * If the folding window started, make sure we start writing in the - * next idle-delta. - */ - if (!time_before(jiffies, calc_load_update)) - idx++; - - return idx & 1; -} - -static inline int calc_load_read_idx(void) -{ - return calc_load_idx & 1; -} - -void calc_load_enter_idle(void) -{ - struct rq *this_rq = this_rq(); - long delta; - - /* - * We're going into NOHZ mode, if there's any pending delta, fold it - * into the pending idle delta. - */ - delta = calc_load_fold_active(this_rq); - if (delta) { - int idx = calc_load_write_idx(); - atomic_long_add(delta, &calc_load_idle[idx]); - } -} - -void calc_load_exit_idle(void) -{ - struct rq *this_rq = this_rq(); - - /* - * If we're still before the sample window, we're done. - */ - if (time_before(jiffies, this_rq->calc_load_update)) - return; - - /* - * We woke inside or after the sample window, this means we're already - * accounted through the nohz accounting, so skip the entire deal and - * sync up for the next window. - */ - this_rq->calc_load_update = calc_load_update; - if (time_before(jiffies, this_rq->calc_load_update + 10)) - this_rq->calc_load_update += LOAD_FREQ; -} - -static long calc_load_fold_idle(void) -{ - int idx = calc_load_read_idx(); - long delta = 0; - - if (atomic_long_read(&calc_load_idle[idx])) - delta = atomic_long_xchg(&calc_load_idle[idx], 0); - - return delta; -} - -/** - * fixed_power_int - compute: x^n, in O(log n) time - * - * @x: base of the power - * @frac_bits: fractional bits of @x - * @n: power to raise @x to. - * - * By exploiting the relation between the definition of the natural power - * function: x^n := x*x*...*x (x multiplied by itself for n times), and - * the binary encoding of numbers used by computers: n := \Sum n_i * 2^i, - * (where: n_i \elem {0, 1}, the binary vector representing n), - * we find: x^n := x^(\Sum n_i * 2^i) := \Prod x^(n_i * 2^i), which is - * of course trivially computable in O(log_2 n), the length of our binary - * vector. - */ -static unsigned long -fixed_power_int(unsigned long x, unsigned int frac_bits, unsigned int n) -{ - unsigned long result = 1UL << frac_bits; - - if (n) for (;;) { - if (n & 1) { - result *= x; - result += 1UL << (frac_bits - 1); - result >>= frac_bits; - } - n >>= 1; - if (!n) - break; - x *= x; - x += 1UL << (frac_bits - 1); - x >>= frac_bits; - } - - return result; -} - -/* - * a1 = a0 * e + a * (1 - e) - * - * a2 = a1 * e + a * (1 - e) - * = (a0 * e + a * (1 - e)) * e + a * (1 - e) - * = a0 * e^2 + a * (1 - e) * (1 + e) - * - * a3 = a2 * e + a * (1 - e) - * = (a0 * e^2 + a * (1 - e) * (1 + e)) * e + a * (1 - e) - * = a0 * e^3 + a * (1 - e) * (1 + e + e^2) - * - * ... - * - * an = a0 * e^n + a * (1 - e) * (1 + e + ... + e^n-1) [1] - * = a0 * e^n + a * (1 - e) * (1 - e^n)/(1 - e) - * = a0 * e^n + a * (1 - e^n) - * - * [1] application of the geometric series: - * - * n 1 - x^(n+1) - * S_n := \Sum x^i = ------------- - * i=0 1 - x - */ -static unsigned long -calc_load_n(unsigned long load, unsigned long exp, - unsigned long active, unsigned int n) -{ - - return calc_load(load, fixed_power_int(exp, FSHIFT, n), active); -} - -/* - * NO_HZ can leave us missing all per-cpu ticks calling - * calc_load_account_active(), but since an idle CPU folds its delta into - * calc_load_tasks_idle per calc_load_account_idle(), all we need to do is fold - * in the pending idle delta if our idle period crossed a load cycle boundary. - * - * Once we've updated the global active value, we need to apply the exponential - * weights adjusted to the number of cycles missed. - */ -static void calc_global_nohz(void) -{ - long delta, active, n; - - if (!time_before(jiffies, calc_load_update + 10)) { - /* - * Catch-up, fold however many we are behind still - */ - delta = jiffies - calc_load_update - 10; - n = 1 + (delta / LOAD_FREQ); - - active = atomic_long_read(&calc_load_tasks); - active = active > 0 ? active * FIXED_1 : 0; - - avenrun[0] = calc_load_n(avenrun[0], EXP_1, active, n); - avenrun[1] = calc_load_n(avenrun[1], EXP_5, active, n); - avenrun[2] = calc_load_n(avenrun[2], EXP_15, active, n); - - calc_load_update += n * LOAD_FREQ; - } - - /* - * Flip the idle index... - * - * Make sure we first write the new time then flip the index, so that - * calc_load_write_idx() will see the new time when it reads the new - * index, this avoids a double flip messing things up. - */ - smp_wmb(); - calc_load_idx++; -} -#else /* !CONFIG_NO_HZ_COMMON */ - -static inline long calc_load_fold_idle(void) { return 0; } -static inline void calc_global_nohz(void) { } - -#endif /* CONFIG_NO_HZ_COMMON */ - -/* - * calc_load - update the avenrun load estimates 10 ticks after the - * CPUs have updated calc_load_tasks. - */ -void calc_global_load(unsigned long ticks) -{ - long active, delta; - - if (time_before(jiffies, calc_load_update + 10)) - return; - - /* - * Fold the 'old' idle-delta to include all NO_HZ cpus. - */ - delta = calc_load_fold_idle(); - if (delta) - atomic_long_add(delta, &calc_load_tasks); - - active = atomic_long_read(&calc_load_tasks); - active = active > 0 ? active * FIXED_1 : 0; - - avenrun[0] = calc_load(avenrun[0], EXP_1, active); - avenrun[1] = calc_load(avenrun[1], EXP_5, active); - avenrun[2] = calc_load(avenrun[2], EXP_15, active); - - calc_load_update += LOAD_FREQ; - - /* - * In case we idled for multiple LOAD_FREQ intervals, catch up in bulk. - */ - calc_global_nohz(); -} - -/* - * Called from update_cpu_load() to periodically update this CPU's - * active count. - */ -static void calc_load_account_active(struct rq *this_rq) -{ - long delta; - - if (time_before(jiffies, this_rq->calc_load_update)) - return; - - delta = calc_load_fold_active(this_rq); - if (delta) - atomic_long_add(delta, &calc_load_tasks); - - this_rq->calc_load_update += LOAD_FREQ; -} - -/* - * End of global load-average stuff - */ - -/* - * The exact cpuload at various idx values, calculated at every tick would be - * load = (2^idx - 1) / 2^idx * load + 1 / 2^idx * cur_load - * - * If a cpu misses updates for n-1 ticks (as it was idle) and update gets called - * on nth tick when cpu may be busy, then we have: - * load = ((2^idx - 1) / 2^idx)^(n-1) * load - * load = (2^idx - 1) / 2^idx) * load + 1 / 2^idx * cur_load - * - * decay_load_missed() below does efficient calculation of - * load = ((2^idx - 1) / 2^idx)^(n-1) * load - * avoiding 0..n-1 loop doing load = ((2^idx - 1) / 2^idx) * load - * - * The calculation is approximated on a 128 point scale. - * degrade_zero_ticks is the number of ticks after which load at any - * particular idx is approximated to be zero. - * degrade_factor is a precomputed table, a row for each load idx. - * Each column corresponds to degradation factor for a power of two ticks, - * based on 128 point scale. - * Example: - * row 2, col 3 (=12) says that the degradation at load idx 2 after - * 8 ticks is 12/128 (which is an approximation of exact factor 3^8/4^8). - * - * With this power of 2 load factors, we can degrade the load n times - * by looking at 1 bits in n and doing as many mult/shift instead of - * n mult/shifts needed by the exact degradation. - */ -#define DEGRADE_SHIFT 7 -static const unsigned char - degrade_zero_ticks[CPU_LOAD_IDX_MAX] = {0, 8, 32, 64, 128}; -static const unsigned char - degrade_factor[CPU_LOAD_IDX_MAX][DEGRADE_SHIFT + 1] = { - {0, 0, 0, 0, 0, 0, 0, 0}, - {64, 32, 8, 0, 0, 0, 0, 0}, - {96, 72, 40, 12, 1, 0, 0}, - {112, 98, 75, 43, 15, 1, 0}, - {120, 112, 98, 76, 45, 16, 2} }; - -/* - * Update cpu_load for any missed ticks, due to tickless idle. The backlog - * would be when CPU is idle and so we just decay the old load without - * adding any new load. - */ -static unsigned long -decay_load_missed(unsigned long load, unsigned long missed_updates, int idx) -{ - int j = 0; - - if (!missed_updates) - return load; - - if (missed_updates >= degrade_zero_ticks[idx]) - return 0; - - if (idx == 1) - return load >> missed_updates; - - while (missed_updates) { - if (missed_updates % 2) - load = (load * degrade_factor[idx][j]) >> DEGRADE_SHIFT; - - missed_updates >>= 1; - j++; - } - return load; -} - -/* - * Update rq->cpu_load[] statistics. This function is usually called every - * scheduler tick (TICK_NSEC). With tickless idle this will not be called - * every tick. We fix it up based on jiffies. - */ -static void __update_cpu_load(struct rq *this_rq, unsigned long this_load, - unsigned long pending_updates) -{ - int i, scale; - - this_rq->nr_load_updates++; - - /* Update our load: */ - this_rq->cpu_load[0] = this_load; /* Fasttrack for idx 0 */ - for (i = 1, scale = 2; i < CPU_LOAD_IDX_MAX; i++, scale += scale) { - unsigned long old_load, new_load; - - /* scale is effectively 1 << i now, and >> i divides by scale */ - - old_load = this_rq->cpu_load[i]; - old_load = decay_load_missed(old_load, pending_updates - 1, i); - new_load = this_load; - /* - * Round up the averaging division if load is increasing. This - * prevents us from getting stuck on 9 if the load is 10, for - * example. - */ - if (new_load > old_load) - new_load += scale - 1; - - this_rq->cpu_load[i] = (old_load * (scale - 1) + new_load) >> i; - } - - sched_avg_update(this_rq); -} - -#ifdef CONFIG_NO_HZ_COMMON -/* - * There is no sane way to deal with nohz on smp when using jiffies because the - * cpu doing the jiffies update might drift wrt the cpu doing the jiffy reading - * causing off-by-one errors in observed deltas; {0,2} instead of {1,1}. - * - * Therefore we cannot use the delta approach from the regular tick since that - * would seriously skew the load calculation. However we'll make do for those - * updates happening while idle (nohz_idle_balance) or coming out of idle - * (tick_nohz_idle_exit). - * - * This means we might still be one tick off for nohz periods. - */ - -/* - * Called from nohz_idle_balance() to update the load ratings before doing the - * idle balance. - */ -void update_idle_cpu_load(struct rq *this_rq) -{ - unsigned long curr_jiffies = ACCESS_ONCE(jiffies); - unsigned long load = this_rq->load.weight; - unsigned long pending_updates; - - /* - * bail if there's load or we're actually up-to-date. - */ - if (load || curr_jiffies == this_rq->last_load_update_tick) - return; - - pending_updates = curr_jiffies - this_rq->last_load_update_tick; - this_rq->last_load_update_tick = curr_jiffies; - - __update_cpu_load(this_rq, load, pending_updates); -} - -/* - * Called from tick_nohz_idle_exit() -- try and fix up the ticks we missed. - */ -void update_cpu_load_nohz(void) -{ - struct rq *this_rq = this_rq(); - unsigned long curr_jiffies = ACCESS_ONCE(jiffies); - unsigned long pending_updates; - - if (curr_jiffies == this_rq->last_load_update_tick) - return; - - raw_spin_lock(&this_rq->lock); - pending_updates = curr_jiffies - this_rq->last_load_update_tick; - if (pending_updates) { - this_rq->last_load_update_tick = curr_jiffies; - /* - * We were idle, this means load 0, the current load might be - * !0 due to remote wakeups and the sort. - */ - __update_cpu_load(this_rq, 0, pending_updates); - } - raw_spin_unlock(&this_rq->lock); -} -#endif /* CONFIG_NO_HZ_COMMON */ - -/* - * Called from scheduler_tick() - */ -static void update_cpu_load_active(struct rq *this_rq) -{ - /* - * See the mess around update_idle_cpu_load() / update_cpu_load_nohz(). - */ - this_rq->last_load_update_tick = jiffies; - __update_cpu_load(this_rq, this_rq->load.weight, 1); - - calc_load_account_active(this_rq); -} - #ifdef CONFIG_SMP /* diff --git a/kernel/sched/proc.c b/kernel/sched/proc.c new file mode 100644 index 000000000000..bb3a6a0b8623 --- /dev/null +++ b/kernel/sched/proc.c @@ -0,0 +1,578 @@ +/* + * kernel/sched/proc.c + * + * Kernel load calculations, forked from sched/core.c + */ + +#include + +#include "sched.h" + +unsigned long this_cpu_load(void) +{ + struct rq *this = this_rq(); + return this->cpu_load[0]; +} + + +/* + * Global load-average calculations + * + * We take a distributed and async approach to calculating the global load-avg + * in order to minimize overhead. + * + * The global load average is an exponentially decaying average of nr_running + + * nr_uninterruptible. + * + * Once every LOAD_FREQ: + * + * nr_active = 0; + * for_each_possible_cpu(cpu) + * nr_active += cpu_of(cpu)->nr_running + cpu_of(cpu)->nr_uninterruptible; + * + * avenrun[n] = avenrun[0] * exp_n + nr_active * (1 - exp_n) + * + * Due to a number of reasons the above turns in the mess below: + * + * - for_each_possible_cpu() is prohibitively expensive on machines with + * serious number of cpus, therefore we need to take a distributed approach + * to calculating nr_active. + * + * \Sum_i x_i(t) = \Sum_i x_i(t) - x_i(t_0) | x_i(t_0) := 0 + * = \Sum_i { \Sum_j=1 x_i(t_j) - x_i(t_j-1) } + * + * So assuming nr_active := 0 when we start out -- true per definition, we + * can simply take per-cpu deltas and fold those into a global accumulate + * to obtain the same result. See calc_load_fold_active(). + * + * Furthermore, in order to avoid synchronizing all per-cpu delta folding + * across the machine, we assume 10 ticks is sufficient time for every + * cpu to have completed this task. + * + * This places an upper-bound on the IRQ-off latency of the machine. Then + * again, being late doesn't loose the delta, just wrecks the sample. + * + * - cpu_rq()->nr_uninterruptible isn't accurately tracked per-cpu because + * this would add another cross-cpu cacheline miss and atomic operation + * to the wakeup path. Instead we increment on whatever cpu the task ran + * when it went into uninterruptible state and decrement on whatever cpu + * did the wakeup. This means that only the sum of nr_uninterruptible over + * all cpus yields the correct result. + * + * This covers the NO_HZ=n code, for extra head-aches, see the comment below. + */ + +/* Variables and functions for calc_load */ +atomic_long_t calc_load_tasks; +unsigned long calc_load_update; +unsigned long avenrun[3]; +EXPORT_SYMBOL(avenrun); /* should be removed */ + +/** + * get_avenrun - get the load average array + * @loads: pointer to dest load array + * @offset: offset to add + * @shift: shift count to shift the result left + * + * These values are estimates at best, so no need for locking. + */ +void get_avenrun(unsigned long *loads, unsigned long offset, int shift) +{ + loads[0] = (avenrun[0] + offset) << shift; + loads[1] = (avenrun[1] + offset) << shift; + loads[2] = (avenrun[2] + offset) << shift; +} + +long calc_load_fold_active(struct rq *this_rq) +{ + long nr_active, delta = 0; + + nr_active = this_rq->nr_running; + nr_active += (long) this_rq->nr_uninterruptible; + + if (nr_active != this_rq->calc_load_active) { + delta = nr_active - this_rq->calc_load_active; + this_rq->calc_load_active = nr_active; + } + + return delta; +} + +/* + * a1 = a0 * e + a * (1 - e) + */ +static unsigned long +calc_load(unsigned long load, unsigned long exp, unsigned long active) +{ + load *= exp; + load += active * (FIXED_1 - exp); + load += 1UL << (FSHIFT - 1); + return load >> FSHIFT; +} + +#ifdef CONFIG_NO_HZ_COMMON +/* + * Handle NO_HZ for the global load-average. + * + * Since the above described distributed algorithm to compute the global + * load-average relies on per-cpu sampling from the tick, it is affected by + * NO_HZ. + * + * The basic idea is to fold the nr_active delta into a global idle-delta upon + * entering NO_HZ state such that we can include this as an 'extra' cpu delta + * when we read the global state. + * + * Obviously reality has to ruin such a delightfully simple scheme: + * + * - When we go NO_HZ idle during the window, we can negate our sample + * contribution, causing under-accounting. + * + * We avoid this by keeping two idle-delta counters and flipping them + * when the window starts, thus separating old and new NO_HZ load. + * + * The only trick is the slight shift in index flip for read vs write. + * + * 0s 5s 10s 15s + * +10 +10 +10 +10 + * |-|-----------|-|-----------|-|-----------|-| + * r:0 0 1 1 0 0 1 1 0 + * w:0 1 1 0 0 1 1 0 0 + * + * This ensures we'll fold the old idle contribution in this window while + * accumlating the new one. + * + * - When we wake up from NO_HZ idle during the window, we push up our + * contribution, since we effectively move our sample point to a known + * busy state. + * + * This is solved by pushing the window forward, and thus skipping the + * sample, for this cpu (effectively using the idle-delta for this cpu which + * was in effect at the time the window opened). This also solves the issue + * of having to deal with a cpu having been in NOHZ idle for multiple + * LOAD_FREQ intervals. + * + * When making the ILB scale, we should try to pull this in as well. + */ +static atomic_long_t calc_load_idle[2]; +static int calc_load_idx; + +static inline int calc_load_write_idx(void) +{ + int idx = calc_load_idx; + + /* + * See calc_global_nohz(), if we observe the new index, we also + * need to observe the new update time. + */ + smp_rmb(); + + /* + * If the folding window started, make sure we start writing in the + * next idle-delta. + */ + if (!time_before(jiffies, calc_load_update)) + idx++; + + return idx & 1; +} + +static inline int calc_load_read_idx(void) +{ + return calc_load_idx & 1; +} + +void calc_load_enter_idle(void) +{ + struct rq *this_rq = this_rq(); + long delta; + + /* + * We're going into NOHZ mode, if there's any pending delta, fold it + * into the pending idle delta. + */ + delta = calc_load_fold_active(this_rq); + if (delta) { + int idx = calc_load_write_idx(); + atomic_long_add(delta, &calc_load_idle[idx]); + } +} + +void calc_load_exit_idle(void) +{ + struct rq *this_rq = this_rq(); + + /* + * If we're still before the sample window, we're done. + */ + if (time_before(jiffies, this_rq->calc_load_update)) + return; + + /* + * We woke inside or after the sample window, this means we're already + * accounted through the nohz accounting, so skip the entire deal and + * sync up for the next window. + */ + this_rq->calc_load_update = calc_load_update; + if (time_before(jiffies, this_rq->calc_load_update + 10)) + this_rq->calc_load_update += LOAD_FREQ; +} + +static long calc_load_fold_idle(void) +{ + int idx = calc_load_read_idx(); + long delta = 0; + + if (atomic_long_read(&calc_load_idle[idx])) + delta = atomic_long_xchg(&calc_load_idle[idx], 0); + + return delta; +} + +/** + * fixed_power_int - compute: x^n, in O(log n) time + * + * @x: base of the power + * @frac_bits: fractional bits of @x + * @n: power to raise @x to. + * + * By exploiting the relation between the definition of the natural power + * function: x^n := x*x*...*x (x multiplied by itself for n times), and + * the binary encoding of numbers used by computers: n := \Sum n_i * 2^i, + * (where: n_i \elem {0, 1}, the binary vector representing n), + * we find: x^n := x^(\Sum n_i * 2^i) := \Prod x^(n_i * 2^i), which is + * of course trivially computable in O(log_2 n), the length of our binary + * vector. + */ +static unsigned long +fixed_power_int(unsigned long x, unsigned int frac_bits, unsigned int n) +{ + unsigned long result = 1UL << frac_bits; + + if (n) for (;;) { + if (n & 1) { + result *= x; + result += 1UL << (frac_bits - 1); + result >>= frac_bits; + } + n >>= 1; + if (!n) + break; + x *= x; + x += 1UL << (frac_bits - 1); + x >>= frac_bits; + } + + return result; +} + +/* + * a1 = a0 * e + a * (1 - e) + * + * a2 = a1 * e + a * (1 - e) + * = (a0 * e + a * (1 - e)) * e + a * (1 - e) + * = a0 * e^2 + a * (1 - e) * (1 + e) + * + * a3 = a2 * e + a * (1 - e) + * = (a0 * e^2 + a * (1 - e) * (1 + e)) * e + a * (1 - e) + * = a0 * e^3 + a * (1 - e) * (1 + e + e^2) + * + * ... + * + * an = a0 * e^n + a * (1 - e) * (1 + e + ... + e^n-1) [1] + * = a0 * e^n + a * (1 - e) * (1 - e^n)/(1 - e) + * = a0 * e^n + a * (1 - e^n) + * + * [1] application of the geometric series: + * + * n 1 - x^(n+1) + * S_n := \Sum x^i = ------------- + * i=0 1 - x + */ +static unsigned long +calc_load_n(unsigned long load, unsigned long exp, + unsigned long active, unsigned int n) +{ + + return calc_load(load, fixed_power_int(exp, FSHIFT, n), active); +} + +/* + * NO_HZ can leave us missing all per-cpu ticks calling + * calc_load_account_active(), but since an idle CPU folds its delta into + * calc_load_tasks_idle per calc_load_account_idle(), all we need to do is fold + * in the pending idle delta if our idle period crossed a load cycle boundary. + * + * Once we've updated the global active value, we need to apply the exponential + * weights adjusted to the number of cycles missed. + */ +static void calc_global_nohz(void) +{ + long delta, active, n; + + if (!time_before(jiffies, calc_load_update + 10)) { + /* + * Catch-up, fold however many we are behind still + */ + delta = jiffies - calc_load_update - 10; + n = 1 + (delta / LOAD_FREQ); + + active = atomic_long_read(&calc_load_tasks); + active = active > 0 ? active * FIXED_1 : 0; + + avenrun[0] = calc_load_n(avenrun[0], EXP_1, active, n); + avenrun[1] = calc_load_n(avenrun[1], EXP_5, active, n); + avenrun[2] = calc_load_n(avenrun[2], EXP_15, active, n); + + calc_load_update += n * LOAD_FREQ; + } + + /* + * Flip the idle index... + * + * Make sure we first write the new time then flip the index, so that + * calc_load_write_idx() will see the new time when it reads the new + * index, this avoids a double flip messing things up. + */ + smp_wmb(); + calc_load_idx++; +} +#else /* !CONFIG_NO_HZ_COMMON */ + +static inline long calc_load_fold_idle(void) { return 0; } +static inline void calc_global_nohz(void) { } + +#endif /* CONFIG_NO_HZ_COMMON */ + +/* + * calc_load - update the avenrun load estimates 10 ticks after the + * CPUs have updated calc_load_tasks. + */ +void calc_global_load(unsigned long ticks) +{ + long active, delta; + + if (time_before(jiffies, calc_load_update + 10)) + return; + + /* + * Fold the 'old' idle-delta to include all NO_HZ cpus. + */ + delta = calc_load_fold_idle(); + if (delta) + atomic_long_add(delta, &calc_load_tasks); + + active = atomic_long_read(&calc_load_tasks); + active = active > 0 ? active * FIXED_1 : 0; + + avenrun[0] = calc_load(avenrun[0], EXP_1, active); + avenrun[1] = calc_load(avenrun[1], EXP_5, active); + avenrun[2] = calc_load(avenrun[2], EXP_15, active); + + calc_load_update += LOAD_FREQ; + + /* + * In case we idled for multiple LOAD_FREQ intervals, catch up in bulk. + */ + calc_global_nohz(); +} + +/* + * Called from update_cpu_load() to periodically update this CPU's + * active count. + */ +static void calc_load_account_active(struct rq *this_rq) +{ + long delta; + + if (time_before(jiffies, this_rq->calc_load_update)) + return; + + delta = calc_load_fold_active(this_rq); + if (delta) + atomic_long_add(delta, &calc_load_tasks); + + this_rq->calc_load_update += LOAD_FREQ; +} + +/* + * End of global load-average stuff + */ + +/* + * The exact cpuload at various idx values, calculated at every tick would be + * load = (2^idx - 1) / 2^idx * load + 1 / 2^idx * cur_load + * + * If a cpu misses updates for n-1 ticks (as it was idle) and update gets called + * on nth tick when cpu may be busy, then we have: + * load = ((2^idx - 1) / 2^idx)^(n-1) * load + * load = (2^idx - 1) / 2^idx) * load + 1 / 2^idx * cur_load + * + * decay_load_missed() below does efficient calculation of + * load = ((2^idx - 1) / 2^idx)^(n-1) * load + * avoiding 0..n-1 loop doing load = ((2^idx - 1) / 2^idx) * load + * + * The calculation is approximated on a 128 point scale. + * degrade_zero_ticks is the number of ticks after which load at any + * particular idx is approximated to be zero. + * degrade_factor is a precomputed table, a row for each load idx. + * Each column corresponds to degradation factor for a power of two ticks, + * based on 128 point scale. + * Example: + * row 2, col 3 (=12) says that the degradation at load idx 2 after + * 8 ticks is 12/128 (which is an approximation of exact factor 3^8/4^8). + * + * With this power of 2 load factors, we can degrade the load n times + * by looking at 1 bits in n and doing as many mult/shift instead of + * n mult/shifts needed by the exact degradation. + */ +#define DEGRADE_SHIFT 7 +static const unsigned char + degrade_zero_ticks[CPU_LOAD_IDX_MAX] = {0, 8, 32, 64, 128}; +static const unsigned char + degrade_factor[CPU_LOAD_IDX_MAX][DEGRADE_SHIFT + 1] = { + {0, 0, 0, 0, 0, 0, 0, 0}, + {64, 32, 8, 0, 0, 0, 0, 0}, + {96, 72, 40, 12, 1, 0, 0}, + {112, 98, 75, 43, 15, 1, 0}, + {120, 112, 98, 76, 45, 16, 2} }; + +/* + * Update cpu_load for any missed ticks, due to tickless idle. The backlog + * would be when CPU is idle and so we just decay the old load without + * adding any new load. + */ +static unsigned long +decay_load_missed(unsigned long load, unsigned long missed_updates, int idx) +{ + int j = 0; + + if (!missed_updates) + return load; + + if (missed_updates >= degrade_zero_ticks[idx]) + return 0; + + if (idx == 1) + return load >> missed_updates; + + while (missed_updates) { + if (missed_updates % 2) + load = (load * degrade_factor[idx][j]) >> DEGRADE_SHIFT; + + missed_updates >>= 1; + j++; + } + return load; +} + +/* + * Update rq->cpu_load[] statistics. This function is usually called every + * scheduler tick (TICK_NSEC). With tickless idle this will not be called + * every tick. We fix it up based on jiffies. + */ +static void __update_cpu_load(struct rq *this_rq, unsigned long this_load, + unsigned long pending_updates) +{ + int i, scale; + + this_rq->nr_load_updates++; + + /* Update our load: */ + this_rq->cpu_load[0] = this_load; /* Fasttrack for idx 0 */ + for (i = 1, scale = 2; i < CPU_LOAD_IDX_MAX; i++, scale += scale) { + unsigned long old_load, new_load; + + /* scale is effectively 1 << i now, and >> i divides by scale */ + + old_load = this_rq->cpu_load[i]; + old_load = decay_load_missed(old_load, pending_updates - 1, i); + new_load = this_load; + /* + * Round up the averaging division if load is increasing. This + * prevents us from getting stuck on 9 if the load is 10, for + * example. + */ + if (new_load > old_load) + new_load += scale - 1; + + this_rq->cpu_load[i] = (old_load * (scale - 1) + new_load) >> i; + } + + sched_avg_update(this_rq); +} + +#ifdef CONFIG_NO_HZ_COMMON +/* + * There is no sane way to deal with nohz on smp when using jiffies because the + * cpu doing the jiffies update might drift wrt the cpu doing the jiffy reading + * causing off-by-one errors in observed deltas; {0,2} instead of {1,1}. + * + * Therefore we cannot use the delta approach from the regular tick since that + * would seriously skew the load calculation. However we'll make do for those + * updates happening while idle (nohz_idle_balance) or coming out of idle + * (tick_nohz_idle_exit). + * + * This means we might still be one tick off for nohz periods. + */ + +/* + * Called from nohz_idle_balance() to update the load ratings before doing the + * idle balance. + */ +void update_idle_cpu_load(struct rq *this_rq) +{ + unsigned long curr_jiffies = ACCESS_ONCE(jiffies); + unsigned long load = this_rq->load.weight; + unsigned long pending_updates; + + /* + * bail if there's load or we're actually up-to-date. + */ + if (load || curr_jiffies == this_rq->last_load_update_tick) + return; + + pending_updates = curr_jiffies - this_rq->last_load_update_tick; + this_rq->last_load_update_tick = curr_jiffies; + + __update_cpu_load(this_rq, load, pending_updates); +} + +/* + * Called from tick_nohz_idle_exit() -- try and fix up the ticks we missed. + */ +void update_cpu_load_nohz(void) +{ + struct rq *this_rq = this_rq(); + unsigned long curr_jiffies = ACCESS_ONCE(jiffies); + unsigned long pending_updates; + + if (curr_jiffies == this_rq->last_load_update_tick) + return; + + raw_spin_lock(&this_rq->lock); + pending_updates = curr_jiffies - this_rq->last_load_update_tick; + if (pending_updates) { + this_rq->last_load_update_tick = curr_jiffies; + /* + * We were idle, this means load 0, the current load might be + * !0 due to remote wakeups and the sort. + */ + __update_cpu_load(this_rq, 0, pending_updates); + } + raw_spin_unlock(&this_rq->lock); +} +#endif /* CONFIG_NO_HZ */ + +/* + * Called from scheduler_tick() + */ +void update_cpu_load_active(struct rq *this_rq) +{ + /* + * See the mess around update_idle_cpu_load() / update_cpu_load_nohz(). + */ + this_rq->last_load_update_tick = jiffies; + __update_cpu_load(this_rq, this_rq->load.weight, 1); + + calc_load_account_active(this_rq); +} diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index ce39224d6155..a38ee0a0650e 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -10,8 +10,16 @@ #include "cpupri.h" #include "cpuacct.h" +struct rq; + extern __read_mostly int scheduler_running; +extern unsigned long calc_load_update; +extern atomic_long_t calc_load_tasks; + +extern long calc_load_fold_active(struct rq *this_rq); +extern void update_cpu_load_active(struct rq *this_rq); + /* * Convert user-nice values [ -20 ... 0 ... 19 ] * to static priority [ MAX_RT_PRIO..MAX_PRIO-1 ], -- cgit v1.2.3 From 8527632dc95472adb571701e852479531c0567a2 Mon Sep 17 00:00:00 2001 From: Paul Gortmaker Date: Fri, 19 Apr 2013 15:10:50 -0400 Subject: sched: Move update_load_*() methods from sched.h to fair.c These inlines are only used by kernel/sched/fair.c so they do not need to be present in the main kernel/sched/sched.h file. Signed-off-by: Paul Gortmaker Acked-by: Peter Zijlstra Cc: Frederic Weisbecker Link: http://lkml.kernel.org/r/1366398650-31599-3-git-send-email-paul.gortmaker@windriver.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 18 ++++++++++++++++++ kernel/sched/sched.h | 18 ------------------ 2 files changed, 18 insertions(+), 18 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index c61a614465c8..08a554dd3e90 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -113,6 +113,24 @@ unsigned int __read_mostly sysctl_sched_shares_window = 10000000UL; unsigned int sysctl_sched_cfs_bandwidth_slice = 5000UL; #endif +static inline void update_load_add(struct load_weight *lw, unsigned long inc) +{ + lw->weight += inc; + lw->inv_weight = 0; +} + +static inline void update_load_sub(struct load_weight *lw, unsigned long dec) +{ + lw->weight -= dec; + lw->inv_weight = 0; +} + +static inline void update_load_set(struct load_weight *lw, unsigned long w) +{ + lw->weight = w; + lw->inv_weight = 0; +} + /* * Increase the granularity value when there are more CPUs, * because with more CPUs the 'effective latency' as visible diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index a38ee0a0650e..f1f6256c1224 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -892,24 +892,6 @@ static inline void finish_lock_switch(struct rq *rq, struct task_struct *prev) #define WF_FORK 0x02 /* child wakeup after fork */ #define WF_MIGRATED 0x4 /* internal use, task got migrated */ -static inline void update_load_add(struct load_weight *lw, unsigned long inc) -{ - lw->weight += inc; - lw->inv_weight = 0; -} - -static inline void update_load_sub(struct load_weight *lw, unsigned long dec) -{ - lw->weight -= dec; - lw->inv_weight = 0; -} - -static inline void update_load_set(struct load_weight *lw, unsigned long w) -{ - lw->weight = w; - lw->inv_weight = 0; -} - /* * To aid in avoiding the subversion of "niceness" due to uneven distribution * of tasks with abnormal "nice" values across CPUs the contribution that -- cgit v1.2.3 From 424c93fe4cbe719e7fd7169248d2b648c493b68d Mon Sep 17 00:00:00 2001 From: Nathan Zimmer Date: Thu, 9 May 2013 11:24:03 -0500 Subject: sched: Use this_rq() helper It is a few instructions more efficent to and slightly more readable to use this_rq()-> instead of cpu_rq(smp_processor_id())-> . Size comparison of kernel/sched/fair.o: text data bss dec hex filename 27972 122 26 28120 6dd8 fair.o.before 27956 122 26 28104 6dc8 fair.o.after Signed-off-by: Nathan Zimmer Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1368116643-87971-1-git-send-email-nzimmer@sgi.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 6 ++---- kernel/sched/rt.c | 2 +- 2 files changed, 3 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index c61a614465c8..f2c9c0c3406c 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -5418,10 +5418,9 @@ static inline void nohz_balance_exit_idle(int cpu) static inline void set_cpu_sd_state_busy(void) { struct sched_domain *sd; - int cpu = smp_processor_id(); rcu_read_lock(); - sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd); + sd = rcu_dereference_check_sched_domain(this_rq()->sd); if (!sd || !sd->nohz_idle) goto unlock; @@ -5436,10 +5435,9 @@ unlock: void set_cpu_sd_state_idle(void) { struct sched_domain *sd; - int cpu = smp_processor_id(); rcu_read_lock(); - sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd); + sd = rcu_dereference_check_sched_domain(this_rq()->sd); if (!sd || sd->nohz_idle) goto unlock; diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c index 127a2c4cf4ab..7aced2e3b085 100644 --- a/kernel/sched/rt.c +++ b/kernel/sched/rt.c @@ -472,7 +472,7 @@ static int rt_se_boosted(struct sched_rt_entity *rt_se) #ifdef CONFIG_SMP static inline const struct cpumask *sched_rt_period_mask(void) { - return cpu_rq(smp_processor_id())->rd->span; + return this_rq()->rd->span; } #else static inline const struct cpumask *sched_rt_period_mask(void) -- cgit v1.2.3 From 1b1d2fb4444231f25ddabc598aa2b5a9c0833fba Mon Sep 17 00:00:00 2001 From: Colin Cross Date: Mon, 6 May 2013 23:50:08 +0000 Subject: lockdep: remove task argument from debug_check_no_locks_held The only existing caller to debug_check_no_locks_held calls it with 'current' as the task, and the freezer needs to call debug_check_no_locks_held but doesn't already have a current task pointer, so remove the argument. It is already assuming that the current task is relevant by dumping the current stack trace as part of the warning. This was originally part of 6aa9707099c (lockdep: check that no locks held at freeze time) which was reverted in dbf520a9d7d4. Original-author: Mandeep Singh Baines Acked-by: Pavel Machek Acked-by: Tejun Heo Signed-off-by: Colin Cross Signed-off-by: Rafael J. Wysocki --- kernel/exit.c | 2 +- kernel/lockdep.c | 17 ++++++++--------- 2 files changed, 9 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index af2eb3cbd499..e59756275000 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -835,7 +835,7 @@ void do_exit(long code) /* * Make sure we are holding no locks: */ - debug_check_no_locks_held(tsk); + debug_check_no_locks_held(); /* * We can do this unlocked here. The futex code uses this flag * just to verify whether the pi state cleanup has been done diff --git a/kernel/lockdep.c b/kernel/lockdep.c index 1f3186b37fd5..e16c45b9ee77 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -4090,7 +4090,7 @@ void debug_check_no_locks_freed(const void *mem_from, unsigned long mem_len) } EXPORT_SYMBOL_GPL(debug_check_no_locks_freed); -static void print_held_locks_bug(struct task_struct *curr) +static void print_held_locks_bug(void) { if (!debug_locks_off()) return; @@ -4099,22 +4099,21 @@ static void print_held_locks_bug(struct task_struct *curr) printk("\n"); printk("=====================================\n"); - printk("[ BUG: lock held at task exit time! ]\n"); + printk("[ BUG: %s/%d still has locks held! ]\n", + current->comm, task_pid_nr(current)); print_kernel_ident(); printk("-------------------------------------\n"); - printk("%s/%d is exiting with locks still held!\n", - curr->comm, task_pid_nr(curr)); - lockdep_print_held_locks(curr); - + lockdep_print_held_locks(current); printk("\nstack backtrace:\n"); dump_stack(); } -void debug_check_no_locks_held(struct task_struct *task) +void debug_check_no_locks_held(void) { - if (unlikely(task->lockdep_depth > 0)) - print_held_locks_bug(task); + if (unlikely(current->lockdep_depth > 0)) + print_held_locks_bug(); } +EXPORT_SYMBOL_GPL(debug_check_no_locks_held); void debug_show_all_locks(void) { -- cgit v1.2.3 From 18ad0c6297df1d671ecea83b608cd9e432642a05 Mon Sep 17 00:00:00 2001 From: Colin Cross Date: Mon, 6 May 2013 23:50:10 +0000 Subject: freezer: shorten freezer sleep time using exponential backoff All tasks can easily be frozen in under 10 ms, switch to using an initial 1 ms sleep followed by exponential backoff until 8 ms. Also convert the printed time to ms instead of centiseconds. Acked-by: Pavel Machek Acked-by: Tejun Heo Signed-off-by: Colin Cross Signed-off-by: Rafael J. Wysocki --- kernel/power/process.c | 26 +++++++++++++++----------- 1 file changed, 15 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/power/process.c b/kernel/power/process.c index 98088e0e71e8..fc0df8486449 100644 --- a/kernel/power/process.c +++ b/kernel/power/process.c @@ -30,9 +30,10 @@ static int try_to_freeze_tasks(bool user_only) unsigned int todo; bool wq_busy = false; struct timeval start, end; - u64 elapsed_csecs64; - unsigned int elapsed_csecs; + u64 elapsed_msecs64; + unsigned int elapsed_msecs; bool wakeup = false; + int sleep_usecs = USEC_PER_MSEC; do_gettimeofday(&start); @@ -68,22 +69,25 @@ static int try_to_freeze_tasks(bool user_only) /* * We need to retry, but first give the freezing tasks some - * time to enter the refrigerator. + * time to enter the refrigerator. Start with an initial + * 1 ms sleep followed by exponential backoff until 8 ms. */ - msleep(10); + usleep_range(sleep_usecs / 2, sleep_usecs); + if (sleep_usecs < 8 * USEC_PER_MSEC) + sleep_usecs *= 2; } do_gettimeofday(&end); - elapsed_csecs64 = timeval_to_ns(&end) - timeval_to_ns(&start); - do_div(elapsed_csecs64, NSEC_PER_SEC / 100); - elapsed_csecs = elapsed_csecs64; + elapsed_msecs64 = timeval_to_ns(&end) - timeval_to_ns(&start); + do_div(elapsed_msecs64, NSEC_PER_MSEC); + elapsed_msecs = elapsed_msecs64; if (todo) { printk("\n"); - printk(KERN_ERR "Freezing of tasks %s after %d.%02d seconds " + printk(KERN_ERR "Freezing of tasks %s after %d.%03d seconds " "(%d tasks refusing to freeze, wq_busy=%d):\n", wakeup ? "aborted" : "failed", - elapsed_csecs / 100, elapsed_csecs % 100, + elapsed_msecs / 1000, elapsed_msecs % 1000, todo - wq_busy, wq_busy); if (!wakeup) { @@ -96,8 +100,8 @@ static int try_to_freeze_tasks(bool user_only) read_unlock(&tasklist_lock); } } else { - printk("(elapsed %d.%02d seconds) ", elapsed_csecs / 100, - elapsed_csecs % 100); + printk("(elapsed %d.%03d seconds) ", elapsed_msecs / 1000, + elapsed_msecs % 1000); } return todo ? -EBUSY : 0; -- cgit v1.2.3 From 613f5d13b569859171f0896fbc73ee0bfa811fda Mon Sep 17 00:00:00 2001 From: Colin Cross Date: Mon, 6 May 2013 23:50:11 +0000 Subject: freezer: skip waking up tasks with PF_FREEZER_SKIP set Android goes through suspend/resume very often (every few seconds when on a busy wifi network with the screen off), and a significant portion of the energy used to go in and out of suspend is spent in the freezer. If a task has called freezer_do_not_count(), don't bother waking it up. If it happens to wake up later it will call freezer_count() and immediately enter the refrigerator. Combined with patches to convert freezable helpers to use freezer_do_not_count() and convert common sites where idle userspace tasks are blocked to use the freezable helpers, this reduces the time and energy required to suspend and resume. Acked-by: Tejun Heo Acked-by: Pavel Machek Signed-off-by: Colin Cross Signed-off-by: Rafael J. Wysocki --- kernel/freezer.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'kernel') diff --git a/kernel/freezer.c b/kernel/freezer.c index c38893b0efba..8b2afc1c9df0 100644 --- a/kernel/freezer.c +++ b/kernel/freezer.c @@ -110,6 +110,18 @@ bool freeze_task(struct task_struct *p) { unsigned long flags; + /* + * This check can race with freezer_do_not_count, but worst case that + * will result in an extra wakeup being sent to the task. It does not + * race with freezer_count(), the barriers in freezer_count() and + * freezer_should_skip() ensure that either freezer_count() sees + * freezing == true in try_to_freeze() and freezes, or + * freezer_should_skip() sees !PF_FREEZE_SKIP and freezes the task + * normally. + */ + if (freezer_should_skip(p)) + return false; + spin_lock_irqsave(&freezer_lock, flags); if (!freezing(p) || frozen(p)) { spin_unlock_irqrestore(&freezer_lock, flags); -- cgit v1.2.3 From 56467c7697f5aef6974501fbe2c3e63674583549 Mon Sep 17 00:00:00 2001 From: Colin Cross Date: Mon, 6 May 2013 23:50:18 +0000 Subject: futex: use freezable blocking call Avoid waking up every thread sleeping in a futex_wait call during suspend and resume by calling a freezable blocking call. Previous patches modified the freezer to avoid sending wakeups to threads that are blocked in freezable blocking calls. This call was selected to be converted to a freezable call because it doesn't hold any locks or release any resources when interrupted that might be needed by another freezing task or a kernel driver during suspend, and is a common site where idle userspace tasks are blocked. Acked-by: Tejun Heo Acked-by: Thomas Gleixner Acked-by: Darren Hart Signed-off-by: Colin Cross Signed-off-by: Rafael J. Wysocki --- kernel/futex.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/futex.c b/kernel/futex.c index b26dcfc02c94..d710fae8abbe 100644 --- a/kernel/futex.c +++ b/kernel/futex.c @@ -61,6 +61,7 @@ #include #include #include +#include #include @@ -1807,7 +1808,7 @@ static void futex_wait_queue_me(struct futex_hash_bucket *hb, struct futex_q *q, * is no timeout, or if it has yet to expire. */ if (!timeout || timeout->task) - schedule(); + freezable_schedule(); } __set_current_state(TASK_RUNNING); } -- cgit v1.2.3 From b0f8c44f30e58c3aaaaaf864d5c3d3cc2e8a4c2d Mon Sep 17 00:00:00 2001 From: Colin Cross Date: Mon, 6 May 2013 23:50:19 +0000 Subject: nanosleep: use freezable blocking call Avoid waking up every thread sleeping in a nanosleep call during suspend and resume by calling a freezable blocking call. Previous patches modified the freezer to avoid sending wakeups to threads that are blocked in freezable blocking calls. This call was selected to be converted to a freezable call because it doesn't hold any locks or release any resources when interrupted that might be needed by another freezing task or a kernel driver during suspend, and is a common site where idle userspace tasks are blocked. Acked-by: Tejun Heo Acked-by: Thomas Gleixner Signed-off-by: Colin Cross Signed-off-by: Rafael J. Wysocki --- kernel/hrtimer.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/hrtimer.c b/kernel/hrtimer.c index fd4b13b131f8..3ee4d06c6fc2 100644 --- a/kernel/hrtimer.c +++ b/kernel/hrtimer.c @@ -47,6 +47,7 @@ #include #include #include +#include #include @@ -1545,7 +1546,7 @@ static int __sched do_nanosleep(struct hrtimer_sleeper *t, enum hrtimer_mode mod t->task = NULL; if (likely(t->task)) - schedule(); + freezable_schedule(); hrtimer_cancel(&t->timer); mode = HRTIMER_MODE_ABS; -- cgit v1.2.3 From a2d5f1f5d941593e61071dc78e9de228eda5475f Mon Sep 17 00:00:00 2001 From: Colin Cross Date: Mon, 6 May 2013 23:50:20 +0000 Subject: sigtimedwait: use freezable blocking call Avoid waking up every thread sleeping in a sigtimedwait call during suspend and resume by calling a freezable blocking call. Previous patches modified the freezer to avoid sending wakeups to threads that are blocked in freezable blocking calls. This call was selected to be converted to a freezable call because it doesn't hold any locks or release any resources when interrupted that might be needed by another freezing task or a kernel driver during suspend, and is a common site where idle userspace tasks are blocked. Acked-by: Tejun Heo Signed-off-by: Colin Cross Signed-off-by: Rafael J. Wysocki --- kernel/signal.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index 113411bfe8b1..50e41075ac77 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -2848,7 +2848,7 @@ int do_sigtimedwait(const sigset_t *which, siginfo_t *info, recalc_sigpending(); spin_unlock_irq(&tsk->sighand->siglock); - timeout = schedule_timeout_interruptible(timeout); + timeout = freezable_schedule_timeout_interruptible(timeout); spin_lock_irq(&tsk->sighand->siglock); __set_task_blocked(tsk, &tsk->real_blocked); -- cgit v1.2.3 From cee22a15052faa817e3ec8985a28154d3fabc7aa Mon Sep 17 00:00:00 2001 From: Viresh Kumar Date: Mon, 8 Apr 2013 16:45:40 +0530 Subject: workqueues: Introduce new flag WQ_POWER_EFFICIENT for power oriented workqueues Workqueues can be performance or power-oriented. Currently, most workqueues are bound to the CPU they were created on. This gives good performance (due to cache effects) at the cost of potentially waking up otherwise idle cores (Idle from scheduler's perspective. Which may or may not be physically idle) just to process some work. To save power, we can allow the work to be rescheduled on a core that is already awake. Workqueues created with the WQ_UNBOUND flag will allow some power savings. However, we don't change the default behaviour of the system. To enable power-saving behaviour, a new config option CONFIG_WQ_POWER_EFFICIENT needs to be turned on. This option can also be overridden by the workqueue.power_efficient boot parameter. tj: Updated config description and comments. Renamed CONFIG_WQ_POWER_EFFICIENT to CONFIG_WQ_POWER_EFFICIENT_DEFAULT. Signed-off-by: Viresh Kumar Reviewed-by: Amit Kucheria Signed-off-by: Tejun Heo --- kernel/power/Kconfig | 20 ++++++++++++++++++++ kernel/workqueue.c | 13 +++++++++++++ 2 files changed, 33 insertions(+) (limited to 'kernel') diff --git a/kernel/power/Kconfig b/kernel/power/Kconfig index 5dfdc9ea180b..46455961a88f 100644 --- a/kernel/power/Kconfig +++ b/kernel/power/Kconfig @@ -263,6 +263,26 @@ config PM_GENERIC_DOMAINS bool depends on PM +config WQ_POWER_EFFICIENT_DEFAULT + bool "Enable workqueue power-efficient mode by default" + depends on PM + default n + help + Per-cpu workqueues are generally preferred because they show + better performance thanks to cache locality; unfortunately, + per-cpu workqueues tend to be more power hungry than unbound + workqueues. + + Enabling workqueue.power_efficient kernel parameter makes the + per-cpu workqueues which were observed to contribute + significantly to power consumption unbound, leading to measurably + lower power usage at the cost of small performance overhead. + + This config option determines whether workqueue.power_efficient + is enabled by default. + + If in doubt, say N. + config PM_GENERIC_DOMAINS_SLEEP def_bool y depends on PM_SLEEP && PM_GENERIC_DOMAINS diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 4aa9f5bc6b2d..8068d97ce141 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -272,6 +272,15 @@ static cpumask_var_t *wq_numa_possible_cpumask; static bool wq_disable_numa; module_param_named(disable_numa, wq_disable_numa, bool, 0444); +/* see the comment above the definition of WQ_POWER_EFFICIENT */ +#ifdef CONFIG_WQ_POWER_EFFICIENT_DEFAULT +static bool wq_power_efficient = true; +#else +static bool wq_power_efficient; +#endif + +module_param_named(power_efficient, wq_power_efficient, bool, 0444); + static bool wq_numa_enabled; /* unbound NUMA affinity enabled */ /* buf for wq_update_unbound_numa_attrs(), protected by CPU hotplug exclusion */ @@ -4085,6 +4094,10 @@ struct workqueue_struct *__alloc_workqueue_key(const char *fmt, struct workqueue_struct *wq; struct pool_workqueue *pwq; + /* see the comment above the definition of WQ_POWER_EFFICIENT */ + if ((flags & WQ_POWER_EFFICIENT) && wq_power_efficient) + flags |= WQ_UNBOUND; + /* allocate wq and format name */ if (flags & WQ_UNBOUND) tbl_size = wq_numa_tbl_len * sizeof(wq->numa_pwq_tbl[0]); -- cgit v1.2.3 From 0668106ca3865ba945e155097fb042bf66d364d3 Mon Sep 17 00:00:00 2001 From: Viresh Kumar Date: Wed, 24 Apr 2013 17:12:54 +0530 Subject: workqueue: Add system wide power_efficient workqueues This patch adds system wide workqueues aligned towards power saving. This is done by allocating them with WQ_UNBOUND flag if 'wq_power_efficient' is set to 'true'. tj: updated comments a bit. Signed-off-by: Viresh Kumar Signed-off-by: Tejun Heo --- kernel/workqueue.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 8068d97ce141..16ca2d3dd29f 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -314,6 +314,10 @@ struct workqueue_struct *system_unbound_wq __read_mostly; EXPORT_SYMBOL_GPL(system_unbound_wq); struct workqueue_struct *system_freezable_wq __read_mostly; EXPORT_SYMBOL_GPL(system_freezable_wq); +struct workqueue_struct *system_power_efficient_wq __read_mostly; +EXPORT_SYMBOL_GPL(system_power_efficient_wq); +struct workqueue_struct *system_freezable_power_efficient_wq __read_mostly; +EXPORT_SYMBOL_GPL(system_freezable_power_efficient_wq); static int worker_thread(void *__worker); static void copy_workqueue_attrs(struct workqueue_attrs *to, @@ -4987,8 +4991,15 @@ static int __init init_workqueues(void) WQ_UNBOUND_MAX_ACTIVE); system_freezable_wq = alloc_workqueue("events_freezable", WQ_FREEZABLE, 0); + system_power_efficient_wq = alloc_workqueue("events_power_efficient", + WQ_POWER_EFFICIENT, 0); + system_freezable_power_efficient_wq = alloc_workqueue("events_freezable_power_efficient", + WQ_FREEZABLE | WQ_POWER_EFFICIENT, + 0); BUG_ON(!system_wq || !system_highpri_wq || !system_long_wq || - !system_unbound_wq || !system_freezable_wq); + !system_unbound_wq || !system_freezable_wq || + !system_power_efficient_wq || + !system_freezable_power_efficient_wq); return 0; } early_initcall(init_workqueues); -- cgit v1.2.3 From fa3ca07e96185aa1496b405472399a2a2a336a17 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Sun, 14 Apr 2013 11:36:56 -0700 Subject: cgroup: refactor hierarchy_id handling We're planning to converting hierarchy_ida to an idr and use it to look up hierarchy from its id. As we want the mapping to happen atomically with cgroupfs_root registration, this patch refactors hierarchy_id init / exit so that ida operations happen inside cgroup_[root_]mutex. * s/init_root_id()/cgroup_init_root_id()/ and make it return 0 or -errno like a normal function. * Move hierarchy_id initialization from cgroup_root_from_opts() into cgroup_mount() block where the root is confirmed to be used and being registered while holding both mutexes. * Split cgroup_drop_id() into cgroup_exit_root_id() and cgroup_free_root(), so that ID release can happen before dropping the mutexes in cgroup_kill_sb(). The latter expects hierarchy_id to be exited before being invoked. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 56 +++++++++++++++++++++++++++++++++++--------------------- 1 file changed, 35 insertions(+), 21 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 2a9926275f80..dbc84f7d23b8 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1426,13 +1426,13 @@ static void init_cgroup_root(struct cgroupfs_root *root) list_add_tail(&cgrp->allcg_node, &root->allcg_list); } -static bool init_root_id(struct cgroupfs_root *root) +static int cgroup_init_root_id(struct cgroupfs_root *root) { - int ret = 0; + int ret; do { if (!ida_pre_get(&hierarchy_ida, GFP_KERNEL)) - return false; + return -ENOMEM; spin_lock(&hierarchy_id_lock); /* Try to allocate the next unused ID */ ret = ida_get_new_above(&hierarchy_ida, next_hierarchy_id, @@ -1448,7 +1448,18 @@ static bool init_root_id(struct cgroupfs_root *root) } spin_unlock(&hierarchy_id_lock); } while (ret); - return true; + return 0; +} + +static void cgroup_exit_root_id(struct cgroupfs_root *root) +{ + if (root->hierarchy_id) { + spin_lock(&hierarchy_id_lock); + ida_remove(&hierarchy_ida, root->hierarchy_id); + spin_unlock(&hierarchy_id_lock); + + root->hierarchy_id = 0; + } } static int cgroup_test_super(struct super_block *sb, void *data) @@ -1482,10 +1493,6 @@ static struct cgroupfs_root *cgroup_root_from_opts(struct cgroup_sb_opts *opts) if (!root) return ERR_PTR(-ENOMEM); - if (!init_root_id(root)) { - kfree(root); - return ERR_PTR(-ENOMEM); - } init_cgroup_root(root); root->subsys_mask = opts->subsys_mask; @@ -1500,17 +1507,15 @@ static struct cgroupfs_root *cgroup_root_from_opts(struct cgroup_sb_opts *opts) return root; } -static void cgroup_drop_root(struct cgroupfs_root *root) +static void cgroup_free_root(struct cgroupfs_root *root) { - if (!root) - return; + if (root) { + /* hierarhcy ID shoulid already have been released */ + WARN_ON_ONCE(root->hierarchy_id); - BUG_ON(!root->hierarchy_id); - spin_lock(&hierarchy_id_lock); - ida_remove(&hierarchy_ida, root->hierarchy_id); - spin_unlock(&hierarchy_id_lock); - ida_destroy(&root->cgroup_ida); - kfree(root); + ida_destroy(&root->cgroup_ida); + kfree(root); + } } static int cgroup_set_super(struct super_block *sb, void *data) @@ -1597,7 +1602,7 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, sb = sget(fs_type, cgroup_test_super, cgroup_set_super, 0, &opts); if (IS_ERR(sb)) { ret = PTR_ERR(sb); - cgroup_drop_root(opts.new_root); + cgroup_free_root(opts.new_root); goto drop_modules; } @@ -1641,6 +1646,10 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, if (ret) goto unlock_drop; + ret = cgroup_init_root_id(root); + if (ret) + goto unlock_drop; + ret = rebind_subsystems(root, root->subsys_mask); if (ret == -EBUSY) { free_cg_links(&tmp_cg_links); @@ -1684,7 +1693,7 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, * We re-used an existing hierarchy - the new root (if * any) is not needed */ - cgroup_drop_root(opts.new_root); + cgroup_free_root(opts.new_root); if (((root->flags | opts.flags) & CGRP_ROOT_SANE_BEHAVIOR) && root->flags != opts.flags) { @@ -1702,6 +1711,7 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, return dget(sb->s_root); unlock_drop: + cgroup_exit_root_id(root); mutex_unlock(&cgroup_root_mutex); mutex_unlock(&cgroup_mutex); mutex_unlock(&inode->i_mutex); @@ -1754,13 +1764,15 @@ static void cgroup_kill_sb(struct super_block *sb) { root_count--; } + cgroup_exit_root_id(root); + mutex_unlock(&cgroup_root_mutex); mutex_unlock(&cgroup_mutex); simple_xattrs_free(&cgrp->xattrs); kill_litter_super(sb); - cgroup_drop_root(root); + cgroup_free_root(root); } static struct file_system_type cgroup_fs_type = { @@ -4642,7 +4654,9 @@ int __init cgroup_init(void) /* Add init_css_set to the hash table */ key = css_set_hash(init_css_set.subsys); hash_add(css_set_table, &init_css_set.hlist, key); - BUG_ON(!init_root_id(&rootnode)); + + /* allocate id for the dummy hierarchy */ + BUG_ON(cgroup_init_root_id(&rootnode)); cgroup_kobj = kobject_create_and_add("cgroup", fs_kobj); if (!cgroup_kobj) { -- cgit v1.2.3 From 54e7b4eb15fc4354d5ada5469e3db4a220ddb3ed Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Sun, 14 Apr 2013 11:36:57 -0700 Subject: cgroup: drop hierarchy_id_lock Now that hierarchy_id alloc / free are protected by the cgroup mutexes, there's no need for this separate lock. Drop it. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 23 +++++++++++++++++------ 1 file changed, 17 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index dbc84f7d23b8..3ef677d314bc 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -189,9 +189,13 @@ struct cgroup_event { static LIST_HEAD(roots); static int root_count; +/* + * Hierarchy ID allocation and mapping. It follows the same exclusion + * rules as other root ops - both cgroup_mutex and cgroup_root_mutex for + * writes, either for reads. + */ static DEFINE_IDA(hierarchy_ida); static int next_hierarchy_id; -static DEFINE_SPINLOCK(hierarchy_id_lock); /* dummytop is a shorthand for the dummy hierarchy's top cgroup */ #define dummytop (&rootnode.top_cgroup) @@ -1430,10 +1434,12 @@ static int cgroup_init_root_id(struct cgroupfs_root *root) { int ret; + lockdep_assert_held(&cgroup_mutex); + lockdep_assert_held(&cgroup_root_mutex); + do { if (!ida_pre_get(&hierarchy_ida, GFP_KERNEL)) return -ENOMEM; - spin_lock(&hierarchy_id_lock); /* Try to allocate the next unused ID */ ret = ida_get_new_above(&hierarchy_ida, next_hierarchy_id, &root->hierarchy_id); @@ -1446,18 +1452,17 @@ static int cgroup_init_root_id(struct cgroupfs_root *root) /* Can only get here if the 31-bit IDR is full ... */ BUG_ON(ret); } - spin_unlock(&hierarchy_id_lock); } while (ret); return 0; } static void cgroup_exit_root_id(struct cgroupfs_root *root) { + lockdep_assert_held(&cgroup_mutex); + lockdep_assert_held(&cgroup_root_mutex); + if (root->hierarchy_id) { - spin_lock(&hierarchy_id_lock); ida_remove(&hierarchy_ida, root->hierarchy_id); - spin_unlock(&hierarchy_id_lock); - root->hierarchy_id = 0; } } @@ -4656,8 +4661,14 @@ int __init cgroup_init(void) hash_add(css_set_table, &init_css_set.hlist, key); /* allocate id for the dummy hierarchy */ + mutex_lock(&cgroup_mutex); + mutex_lock(&cgroup_root_mutex); + BUG_ON(cgroup_init_root_id(&rootnode)); + mutex_unlock(&cgroup_root_mutex); + mutex_unlock(&cgroup_mutex); + cgroup_kobj = kobject_create_and_add("cgroup", fs_kobj); if (!cgroup_kobj) { err = -ENOMEM; -- cgit v1.2.3 From 1a574231669f8c3065c83974e9557fcbbd94b8a6 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Sun, 14 Apr 2013 11:36:58 -0700 Subject: cgroup: make hierarchy_id use cyclic idr We want to be able to lookup a hierarchy from its id and cyclic allocation is a whole lot simpler with idr. Convert to idr and use idr_alloc_cyclc(). Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 28 ++++++++-------------------- 1 file changed, 8 insertions(+), 20 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 3ef677d314bc..dcb417c6c242 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -194,8 +194,7 @@ static int root_count; * rules as other root ops - both cgroup_mutex and cgroup_root_mutex for * writes, either for reads. */ -static DEFINE_IDA(hierarchy_ida); -static int next_hierarchy_id; +static DEFINE_IDR(cgroup_hierarchy_idr); /* dummytop is a shorthand for the dummy hierarchy's top cgroup */ #define dummytop (&rootnode.top_cgroup) @@ -1432,27 +1431,16 @@ static void init_cgroup_root(struct cgroupfs_root *root) static int cgroup_init_root_id(struct cgroupfs_root *root) { - int ret; + int id; lockdep_assert_held(&cgroup_mutex); lockdep_assert_held(&cgroup_root_mutex); - do { - if (!ida_pre_get(&hierarchy_ida, GFP_KERNEL)) - return -ENOMEM; - /* Try to allocate the next unused ID */ - ret = ida_get_new_above(&hierarchy_ida, next_hierarchy_id, - &root->hierarchy_id); - if (ret == -ENOSPC) - /* Try again starting from 0 */ - ret = ida_get_new(&hierarchy_ida, &root->hierarchy_id); - if (!ret) { - next_hierarchy_id = root->hierarchy_id + 1; - } else if (ret != -EAGAIN) { - /* Can only get here if the 31-bit IDR is full ... */ - BUG_ON(ret); - } - } while (ret); + id = idr_alloc_cyclic(&cgroup_hierarchy_idr, root, 2, 0, GFP_KERNEL); + if (id < 0) + return id; + + root->hierarchy_id = id; return 0; } @@ -1462,7 +1450,7 @@ static void cgroup_exit_root_id(struct cgroupfs_root *root) lockdep_assert_held(&cgroup_root_mutex); if (root->hierarchy_id) { - ida_remove(&hierarchy_ida, root->hierarchy_id); + idr_remove(&cgroup_hierarchy_idr, root->hierarchy_id); root->hierarchy_id = 0; } } -- cgit v1.2.3 From 857a2beb09ab83e9a8185821ae16db7dfbe8b837 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Sun, 14 Apr 2013 20:50:08 -0700 Subject: cgroup: implement task_cgroup_path_from_hierarchy() kdbus folks want a sane way to determine the cgroup path that a given task belongs to on a given hierarchy, which is a reasonble thing to expect from cgroup core. Implement task_cgroup_path_from_hierarchy(). v2: Dropped unnecessary NULL check on the return value of task_cgroup_from_root() as suggested by Li Zefan. Signed-off-by: Tejun Heo Acked-by: Greg Kroah-Hartman Acked-by: Li Zefan Cc: Kay Sievers Cc: Lennart Poettering Cc: Daniel Mack --- kernel/cgroup.c | 32 ++++++++++++++++++++++++++++++++ 1 file changed, 32 insertions(+) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index dcb417c6c242..6b2b1d945df2 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1827,6 +1827,38 @@ out: } EXPORT_SYMBOL_GPL(cgroup_path); +/** + * task_cgroup_path_from_hierarchy - cgroup path of a task on a hierarchy + * @task: target task + * @hierarchy_id: the hierarchy to look up @task's cgroup from + * @buf: the buffer to write the path into + * @buflen: the length of the buffer + * + * Determine @task's cgroup on the hierarchy specified by @hierarchy_id and + * copy its path into @buf. This function grabs cgroup_mutex and shouldn't + * be used inside locks used by cgroup controller callbacks. + */ +int task_cgroup_path_from_hierarchy(struct task_struct *task, int hierarchy_id, + char *buf, size_t buflen) +{ + struct cgroupfs_root *root; + struct cgroup *cgrp = NULL; + int ret = -ENOENT; + + mutex_lock(&cgroup_mutex); + + root = idr_find(&cgroup_hierarchy_idr, hierarchy_id); + if (root) { + cgrp = task_cgroup_from_root(task, root); + ret = cgroup_path(cgrp, buf, buflen); + } + + mutex_unlock(&cgroup_mutex); + + return ret; +} +EXPORT_SYMBOL_GPL(task_cgroup_path_from_hierarchy); + /* * Control Group taskset */ -- cgit v1.2.3 From cb65537ee1134d3cc55c1fa83952bc8eb1212833 Mon Sep 17 00:00:00 2001 From: David Howells Date: Fri, 10 May 2013 19:50:26 +0100 Subject: Add wait_on_atomic_t() and wake_up_atomic_t() Add wait_on_atomic_t() and wake_up_atomic_t() to indicate became-zero events on atomic_t types. This uses the bit-wake waitqueue table. The key is set to a value outside of the number of bits in a long so that wait_on_bit() won't be woken up accidentally. What I'm using this for is: in a following patch I add a counter to struct fscache_cookie to count the number of outstanding operations that need access to netfs data. The way this works is: (1) When a cookie is allocated, the counter is initialised to 1. (2) When an operation wants to access netfs data, it calls atomic_inc_unless() to increment the counter before it does so. If it was 0, then the counter isn't incremented, the operation isn't permitted to access the netfs data (which might by this point no longer exist) and the operation aborts in some appropriate manner. (3) When an operation finishes with the netfs data, it decrements the counter and if it reaches 0, calls wake_up_atomic_t() on it - the assumption being that it was the last blocker. (4) When a cookie is released, the counter is decremented and the releaser uses wait_on_atomic_t() to wait for the counter to become 0 - which should indicate no one is using the netfs data any longer. The netfs data can then be destroyed. There are some alternatives that I have thought of and that have been suggested by Tejun Heo: (A) Using wait_on_bit() to wait on a bit in the counter. This doesn't work because if that bit happens to be 0 then the wait won't happen - even if the counter is non-zero. (B) Using wait_on_bit() to wait on a flag elsewhere which is cleared when the counter reaches 0. Such a flag would be redundant and would add complexity. (C) Adding a waitqueue to fscache_cookie - this would expand that struct by several words for an event that happens just once in each cookie's lifetime. Further, cookies are generally per-file so there are likely to be a lot of them. (D) Similar to (C), but add a pointer to a waitqueue in the cookie instead of a waitqueue. This would add single word per cookie and so would be less of an expansion - but still an expansion. (E) Adding a static waitqueue to the fscache module. Generally this would be fine, but under certain circumstances many cookies will all get added at the same time (eg. NFS umount, cache withdrawal) thereby presenting scaling issues. Note that the wait may be significant as disk I/O may be in progress. So, I think reusing the wait_on_bit() waitqueue set is reasonable. I don't make much use of the waitqueue I need on a per-cookie basis, but sometimes I have a huge flood of the cookies to deal with. I also don't want to add a whole new set of global waitqueue tables specifically for the dec-to-0 event if I can reuse the bit tables. Signed-off-by: David Howells Tested-By: Milosz Tanski Acked-by: Jeff Layton --- kernel/wait.c | 88 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 88 insertions(+) (limited to 'kernel') diff --git a/kernel/wait.c b/kernel/wait.c index 6698e0c04ead..ce0daa320a26 100644 --- a/kernel/wait.c +++ b/kernel/wait.c @@ -287,3 +287,91 @@ wait_queue_head_t *bit_waitqueue(void *word, int bit) return &zone->wait_table[hash_long(val, zone->wait_table_bits)]; } EXPORT_SYMBOL(bit_waitqueue); + +/* + * Manipulate the atomic_t address to produce a better bit waitqueue table hash + * index (we're keying off bit -1, but that would produce a horrible hash + * value). + */ +static inline wait_queue_head_t *atomic_t_waitqueue(atomic_t *p) +{ + if (BITS_PER_LONG == 64) { + unsigned long q = (unsigned long)p; + return bit_waitqueue((void *)(q & ~1), q & 1); + } + return bit_waitqueue(p, 0); +} + +static int wake_atomic_t_function(wait_queue_t *wait, unsigned mode, int sync, + void *arg) +{ + struct wait_bit_key *key = arg; + struct wait_bit_queue *wait_bit + = container_of(wait, struct wait_bit_queue, wait); + atomic_t *val = key->flags; + + if (wait_bit->key.flags != key->flags || + wait_bit->key.bit_nr != key->bit_nr || + atomic_read(val) != 0) + return 0; + return autoremove_wake_function(wait, mode, sync, key); +} + +/* + * To allow interruptible waiting and asynchronous (i.e. nonblocking) waiting, + * the actions of __wait_on_atomic_t() are permitted return codes. Nonzero + * return codes halt waiting and return. + */ +static __sched +int __wait_on_atomic_t(wait_queue_head_t *wq, struct wait_bit_queue *q, + int (*action)(atomic_t *), unsigned mode) +{ + atomic_t *val; + int ret = 0; + + do { + prepare_to_wait(wq, &q->wait, mode); + val = q->key.flags; + if (atomic_read(val) == 0) + ret = (*action)(val); + } while (!ret && atomic_read(val) != 0); + finish_wait(wq, &q->wait); + return ret; +} + +#define DEFINE_WAIT_ATOMIC_T(name, p) \ + struct wait_bit_queue name = { \ + .key = __WAIT_ATOMIC_T_KEY_INITIALIZER(p), \ + .wait = { \ + .private = current, \ + .func = wake_atomic_t_function, \ + .task_list = \ + LIST_HEAD_INIT((name).wait.task_list), \ + }, \ + } + +__sched int out_of_line_wait_on_atomic_t(atomic_t *p, int (*action)(atomic_t *), + unsigned mode) +{ + wait_queue_head_t *wq = atomic_t_waitqueue(p); + DEFINE_WAIT_ATOMIC_T(wait, p); + + return __wait_on_atomic_t(wq, &wait, action, mode); +} +EXPORT_SYMBOL(out_of_line_wait_on_atomic_t); + +/** + * wake_up_atomic_t - Wake up a waiter on a atomic_t + * @word: The word being waited on, a kernel virtual address + * @bit: The bit of the word being waited on + * + * Wake up anyone waiting for the atomic_t to go to zero. + * + * Abuse the bit-waker function and its waitqueue hash table set (the atomic_t + * check is done by the waiter's wake function, not the by the waker itself). + */ +void wake_up_atomic_t(atomic_t *p) +{ + __wake_up_bit(atomic_t_waitqueue(p), p, WAIT_ATOMIC_T_BIT_NR); +} +EXPORT_SYMBOL(wake_up_atomic_t); -- cgit v1.2.3 From 5d33b883aed81c6fbcd09c6f7c3619eee850a7e2 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 25 Apr 2013 20:31:43 +0000 Subject: clocksource: Always verify highres capability If a clocksource has a (wrong) high rating, but can't be used as a timebase for oneshot tick mode, it is unconditionally selected even when the system is already in oneshot tick mode. This causes full system failure. Verify the clocksource selection against the oneshot mode. Signed-off-by: Thomas Gleixner Acked-by: John Stultz Cc: Magnus Damm Link: http://lkml.kernel.org/r/20130425143435.635040849@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/time/clocksource.c | 31 ++++++++++++++++++++++++++----- 1 file changed, 26 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c index c9583382141a..dda5c7130d93 100644 --- a/kernel/time/clocksource.c +++ b/kernel/time/clocksource.c @@ -553,6 +553,26 @@ static u64 clocksource_max_deferment(struct clocksource *cs) #ifndef CONFIG_ARCH_USES_GETTIMEOFFSET +static struct clocksource *clocksource_find_best(bool oneshot) +{ + struct clocksource *cs; + + if (!finished_booting || list_empty(&clocksource_list)) + return NULL; + + /* + * We pick the clocksource with the highest rating. If oneshot + * mode is active, we pick the highres valid clocksource with + * the best rating. + */ + list_for_each_entry(cs, &clocksource_list, list) { + if (oneshot && !(cs->flags & CLOCK_SOURCE_VALID_FOR_HRES)) + continue; + return cs; + } + return NULL; +} + /** * clocksource_select - Select the best clocksource available * @@ -563,12 +583,14 @@ static u64 clocksource_max_deferment(struct clocksource *cs) */ static void clocksource_select(void) { + bool oneshot = tick_oneshot_mode_active(); struct clocksource *best, *cs; - if (!finished_booting || list_empty(&clocksource_list)) + /* Find the best suitable clocksource */ + best = clocksource_find_best(oneshot); + if (!best) return; - /* First clocksource on the list has the best rating. */ - best = list_first_entry(&clocksource_list, struct clocksource, list); + /* Check for the override clocksource. */ list_for_each_entry(cs, &clocksource_list, list) { if (strcmp(cs->name, override_name) != 0) @@ -578,8 +600,7 @@ static void clocksource_select(void) * capable clocksource if the tick code is in oneshot * mode (highres or nohz) */ - if (!(cs->flags & CLOCK_SOURCE_VALID_FOR_HRES) && - tick_oneshot_mode_active()) { + if (!(cs->flags & CLOCK_SOURCE_VALID_FOR_HRES) && oneshot) { /* Override clocksource cannot be used. */ printk(KERN_WARNING "Override clocksource %s is not " "HRT compatible. Cannot switch while in " -- cgit v1.2.3 From ba919d1caa2e624eb8c6cae1f2ce0a253e697d45 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 25 Apr 2013 20:31:44 +0000 Subject: clocksource: Let timekeeping_notify return success/error timekeeping_notify() can fail due cs->enable() failure. Though the caller does not notice and happily keeps the wrong clocksource as the current one. Let the caller know about failure, so the current clocksource will be shown correctly in sysfs. Signed-off-by: Thomas Gleixner Acked-by: John Stultz Cc: Magnus Damm Link: http://lkml.kernel.org/r/20130425143435.696321912@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/time/clocksource.c | 6 +++--- kernel/time/timekeeping.c | 5 +++-- 2 files changed, 6 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c index dda5c7130d93..1923a340bd91 100644 --- a/kernel/time/clocksource.c +++ b/kernel/time/clocksource.c @@ -611,10 +611,10 @@ static void clocksource_select(void) best = cs; break; } - if (curr_clocksource != best) { - printk(KERN_INFO "Switching to clocksource %s\n", best->name); + + if (curr_clocksource != best && !timekeeping_notify(best)) { + pr_info("Switched to clocksource %s\n", best->name); curr_clocksource = best; - timekeeping_notify(curr_clocksource); } } diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index 98cd470bbe49..da6e10c7a378 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -648,14 +648,15 @@ static int change_clocksource(void *data) * This function is called from clocksource.c after a new, better clock * source has been registered. The caller holds the clocksource_mutex. */ -void timekeeping_notify(struct clocksource *clock) +int timekeeping_notify(struct clocksource *clock) { struct timekeeper *tk = &timekeeper; if (tk->clock == clock) - return; + return 0; stop_machine(change_clocksource, clock, NULL); tick_clock_notify(); + return tk->clock == clock ? 0 : -1; } /** -- cgit v1.2.3 From 09ac369c825d9d593404306d59062d854b321e9b Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 25 Apr 2013 20:31:44 +0000 Subject: clocksource: Add module refcount Add a module refcount, so the current clocksource cannot be removed unconditionally. Signed-off-by: Thomas Gleixner Cc: John Stultz Cc: Magnus Damm Link: http://lkml.kernel.org/r/20130425143435.762417789@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/time/timekeeping.c | 19 ++++++++++++++----- 1 file changed, 14 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index da6e10c7a378..933efa4071c3 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -627,11 +627,20 @@ static int change_clocksource(void *data) write_seqcount_begin(&timekeeper_seq); timekeeping_forward_now(tk); - if (!new->enable || new->enable(new) == 0) { - old = tk->clock; - tk_setup_internals(tk, new); - if (old->disable) - old->disable(old); + /* + * If the cs is in module, get a module reference. Succeeds + * for built-in code (owner == NULL) as well. + */ + if (try_module_get(new->owner)) { + if (!new->enable || new->enable(new) == 0) { + old = tk->clock; + tk_setup_internals(tk, new); + if (old->disable) + old->disable(old); + module_put(old->owner); + } else { + module_put(new->owner); + } } timekeeping_update(tk, true, true); -- cgit v1.2.3 From f5a2e34375a5e2b711aea488ac3ae50eeba6d57c Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 25 Apr 2013 20:31:45 +0000 Subject: clocksource: Allow clocksource select to skip current clocksource Preparatory patch for clocksource unbind support. Split out code from clocksource_select and modify it, so it skips the current clocksource on request and tries to find a fallback clocksource. Convert all existing users. No functional change. Signed-off-by: Thomas Gleixner Cc: John Stultz Cc: Magnus Damm Link: http://lkml.kernel.org/r/20130425143435.834965397@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/time/clocksource.c | 31 ++++++++++++++++++++----------- 1 file changed, 20 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c index 1923a340bd91..9782997cb6cf 100644 --- a/kernel/time/clocksource.c +++ b/kernel/time/clocksource.c @@ -553,7 +553,7 @@ static u64 clocksource_max_deferment(struct clocksource *cs) #ifndef CONFIG_ARCH_USES_GETTIMEOFFSET -static struct clocksource *clocksource_find_best(bool oneshot) +static struct clocksource *clocksource_find_best(bool oneshot, bool skipcur) { struct clocksource *cs; @@ -566,6 +566,8 @@ static struct clocksource *clocksource_find_best(bool oneshot) * the best rating. */ list_for_each_entry(cs, &clocksource_list, list) { + if (skipcur && cs == curr_clocksource) + continue; if (oneshot && !(cs->flags & CLOCK_SOURCE_VALID_FOR_HRES)) continue; return cs; @@ -573,26 +575,20 @@ static struct clocksource *clocksource_find_best(bool oneshot) return NULL; } -/** - * clocksource_select - Select the best clocksource available - * - * Private function. Must hold clocksource_mutex when called. - * - * Select the clocksource with the best rating, or the clocksource, - * which is selected by userspace override. - */ -static void clocksource_select(void) +static void __clocksource_select(bool skipcur) { bool oneshot = tick_oneshot_mode_active(); struct clocksource *best, *cs; /* Find the best suitable clocksource */ - best = clocksource_find_best(oneshot); + best = clocksource_find_best(oneshot, skipcur); if (!best) return; /* Check for the override clocksource. */ list_for_each_entry(cs, &clocksource_list, list) { + if (skipcur && cs == curr_clocksource) + continue; if (strcmp(cs->name, override_name) != 0) continue; /* @@ -618,6 +614,19 @@ static void clocksource_select(void) } } +/** + * clocksource_select - Select the best clocksource available + * + * Private function. Must hold clocksource_mutex when called. + * + * Select the clocksource with the best rating, or the clocksource, + * which is selected by userspace override. + */ +static void clocksource_select(void) +{ + return __clocksource_select(false); +} + #else /* !CONFIG_ARCH_USES_GETTIMEOFFSET */ static inline void clocksource_select(void) { } -- cgit v1.2.3 From 29b5407819f59731c9423238fae03b756822708c Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 25 Apr 2013 20:31:45 +0000 Subject: clocksource: Split out user string input Split out the user string input for clocksource override. Preparatory patch for unbind. [ jstultz: Fix an off by one error ] Signed-off-by: Thomas Gleixner Cc: John Stultz Cc: Magnus Damm Link: http://lkml.kernel.org/r/20130425143435.895851338@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/time/clocksource.c | 37 +++++++++++++++++++++++-------------- 1 file changed, 23 insertions(+), 14 deletions(-) (limited to 'kernel') diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c index 9782997cb6cf..d7f1a45c2fa5 100644 --- a/kernel/time/clocksource.c +++ b/kernel/time/clocksource.c @@ -174,7 +174,8 @@ clocks_calc_mult_shift(u32 *mult, u32 *shift, u32 from, u32 to, u32 maxsec) static struct clocksource *curr_clocksource; static LIST_HEAD(clocksource_list); static DEFINE_MUTEX(clocksource_mutex); -static char override_name[32]; +#define CS_NAME_LEN 32 +static char override_name[CS_NAME_LEN]; static int finished_booting; #ifdef CONFIG_CLOCKSOURCE_WATCHDOG @@ -838,6 +839,23 @@ sysfs_show_current_clocksources(struct device *dev, return count; } +static size_t clocksource_get_uname(const char *buf, char *dst, size_t cnt) +{ + size_t ret = cnt; + + /* strings from sysfs write are not 0 terminated! */ + if (!cnt || cnt >= CS_NAME_LEN) + return -EINVAL; + + /* strip of \n: */ + if (buf[cnt-1] == '\n') + cnt--; + if (cnt > 0) + memcpy(dst, buf, cnt); + dst[cnt] = 0; + return ret; +} + /** * sysfs_override_clocksource - interface for manually overriding clocksource * @dev: unused @@ -852,22 +870,13 @@ static ssize_t sysfs_override_clocksource(struct device *dev, struct device_attribute *attr, const char *buf, size_t count) { - size_t ret = count; - - /* strings from sysfs write are not 0 terminated! */ - if (count >= sizeof(override_name)) - return -EINVAL; - - /* strip of \n: */ - if (buf[count-1] == '\n') - count--; + size_t ret; mutex_lock(&clocksource_mutex); - if (count > 0) - memcpy(override_name, buf, count); - override_name[count] = 0; - clocksource_select(); + ret = clocksource_get_uname(buf, override_name, count); + if (ret >= 0) + clocksource_select(); mutex_unlock(&clocksource_mutex); -- cgit v1.2.3 From 7eaeb34305dee26634f7c98ae62646da5cebe91d Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 25 Apr 2013 20:31:46 +0000 Subject: clocksource: Provide unbind interface in sysfs With the module refcount held for the current clocksource there is no way to unload the module. Provide a sysfs interface which allows to unbind the clocksource. One could argue that the clocksource override could be (ab)used to do so, but the clocksource override cannot be used from the kernel itself, while an unbind function can be used to programmatically check whether a clocksource can be shutdown or not. The unbind functionality uses the new skip current feature of clocksource_select and verifies that a fallback clocksource has been installed. If the clocksource which should be unbound is the current clocksource and no fallback can be found, unbind returns -EBUSY. This does not support the unbinding of a clocksource which is used as the watchdog clocksource. No point in fostering crappy hardware. Signed-off-by: Thomas Gleixner Cc: John Stultz Cc: Magnus Damm Link: http://lkml.kernel.org/r/20130425143435.964218245@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/time/clocksource.c | 73 +++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 73 insertions(+) (limited to 'kernel') diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c index d7f1a45c2fa5..791d1aeb17ac 100644 --- a/kernel/time/clocksource.c +++ b/kernel/time/clocksource.c @@ -440,6 +440,11 @@ static int clocksource_watchdog_kthread(void *data) return 0; } +static bool clocksource_is_watchdog(struct clocksource *cs) +{ + return cs == watchdog; +} + #else /* CONFIG_CLOCKSOURCE_WATCHDOG */ static void clocksource_enqueue_watchdog(struct clocksource *cs) @@ -451,6 +456,7 @@ static void clocksource_enqueue_watchdog(struct clocksource *cs) static inline void clocksource_dequeue_watchdog(struct clocksource *cs) { } static inline void clocksource_resume_watchdog(void) { } static inline int clocksource_watchdog_kthread(void *data) { return 0; } +static bool clocksource_is_watchdog(struct clocksource *cs) { return false; } #endif /* CONFIG_CLOCKSOURCE_WATCHDOG */ @@ -628,6 +634,11 @@ static void clocksource_select(void) return __clocksource_select(false); } +static void clocksource_select_fallback(void) +{ + return __clocksource_select(true); +} + #else /* !CONFIG_ARCH_USES_GETTIMEOFFSET */ static inline void clocksource_select(void) { } @@ -803,6 +814,29 @@ void clocksource_change_rating(struct clocksource *cs, int rating) } EXPORT_SYMBOL(clocksource_change_rating); +/* + * Unbind clocksource @cs. Called with clocksource_mutex held + */ +static int clocksource_unbind(struct clocksource *cs) +{ + /* + * I really can't convince myself to support this on hardware + * designed by lobotomized monkeys. + */ + if (clocksource_is_watchdog(cs)) + return -EBUSY; + + if (cs == curr_clocksource) { + /* Select and try to install a replacement clock source */ + clocksource_select_fallback(); + if (curr_clocksource == cs) + return -EBUSY; + } + clocksource_dequeue_watchdog(cs); + list_del_init(&cs->list); + return 0; +} + /** * clocksource_unregister - remove a registered clocksource * @cs: clocksource to be unregistered @@ -883,6 +917,40 @@ static ssize_t sysfs_override_clocksource(struct device *dev, return ret; } +/** + * sysfs_unbind_current_clocksource - interface for manually unbinding clocksource + * @dev: unused + * @attr: unused + * @buf: unused + * @count: length of buffer + * + * Takes input from sysfs interface for manually unbinding a clocksource. + */ +static ssize_t sysfs_unbind_clocksource(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count) +{ + struct clocksource *cs; + char name[CS_NAME_LEN]; + size_t ret; + + ret = clocksource_get_uname(buf, name, count); + if (ret < 0) + return ret; + + ret = -ENODEV; + mutex_lock(&clocksource_mutex); + list_for_each_entry(cs, &clocksource_list, list) { + if (strcmp(cs->name, name)) + continue; + ret = clocksource_unbind(cs); + break; + } + mutex_unlock(&clocksource_mutex); + + return ret ? ret : count; +} + /** * sysfs_show_available_clocksources - sysfs interface for listing clocksource * @dev: unused @@ -925,6 +993,8 @@ sysfs_show_available_clocksources(struct device *dev, static DEVICE_ATTR(current_clocksource, 0644, sysfs_show_current_clocksources, sysfs_override_clocksource); +static DEVICE_ATTR(unbind_clocksource, 0200, NULL, sysfs_unbind_clocksource); + static DEVICE_ATTR(available_clocksource, 0444, sysfs_show_available_clocksources, NULL); @@ -948,6 +1018,9 @@ static int __init init_clocksource_sysfs(void) error = device_create_file( &device_clocksource, &dev_attr_current_clocksource); + if (!error) + error = device_create_file(&device_clocksource, + &dev_attr_unbind_clocksource); if (!error) error = device_create_file( &device_clocksource, -- cgit v1.2.3 From a89c7edbe7d7aa80f507915f3dd801211b116b79 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 25 Apr 2013 20:31:46 +0000 Subject: clocksource: Let clocksource_unregister() return success/error The unregister call can fail, if the clocksource is the current one and there is no replacement clocksource available. It can also fail, if the clocksource is the watchdog clocksource and I'm not going to provide support for this. Signed-off-by: Thomas Gleixner Cc: John Stultz Cc: Magnus Damm Link: http://lkml.kernel.org/r/20130425143436.029915527@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/time/clocksource.c | 33 ++++++++++++--------------------- 1 file changed, 12 insertions(+), 21 deletions(-) (limited to 'kernel') diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c index 791d1aeb17ac..31b90332f47b 100644 --- a/kernel/time/clocksource.c +++ b/kernel/time/clocksource.c @@ -389,28 +389,17 @@ static void clocksource_enqueue_watchdog(struct clocksource *cs) static void clocksource_dequeue_watchdog(struct clocksource *cs) { - struct clocksource *tmp; unsigned long flags; spin_lock_irqsave(&watchdog_lock, flags); - if (cs->flags & CLOCK_SOURCE_MUST_VERIFY) { - /* cs is a watched clocksource. */ - list_del_init(&cs->wd_list); - } else if (cs == watchdog) { - /* Reset watchdog cycles */ - clocksource_reset_watchdog(); - /* Current watchdog is removed. Find an alternative. */ - watchdog = NULL; - list_for_each_entry(tmp, &clocksource_list, list) { - if (tmp == cs || tmp->flags & CLOCK_SOURCE_MUST_VERIFY) - continue; - if (!watchdog || tmp->rating > watchdog->rating) - watchdog = tmp; + if (cs != watchdog) { + if (cs->flags & CLOCK_SOURCE_MUST_VERIFY) { + /* cs is a watched clocksource. */ + list_del_init(&cs->wd_list); + /* Check if the watchdog timer needs to be stopped. */ + clocksource_stop_watchdog(); } } - cs->flags &= ~CLOCK_SOURCE_WATCHDOG; - /* Check if the watchdog timer needs to be stopped. */ - clocksource_stop_watchdog(); spin_unlock_irqrestore(&watchdog_lock, flags); } @@ -841,13 +830,15 @@ static int clocksource_unbind(struct clocksource *cs) * clocksource_unregister - remove a registered clocksource * @cs: clocksource to be unregistered */ -void clocksource_unregister(struct clocksource *cs) +int clocksource_unregister(struct clocksource *cs) { + int ret = 0; + mutex_lock(&clocksource_mutex); - clocksource_dequeue_watchdog(cs); - list_del(&cs->list); - clocksource_select(); + if (!list_empty(&cs->list)) + ret = clocksource_unbind(cs); mutex_unlock(&clocksource_mutex); + return ret; } EXPORT_SYMBOL(clocksource_unregister); -- cgit v1.2.3 From 7172a286ced0c1f4f239a0fa09db54ed37d3ead2 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 25 Apr 2013 20:31:47 +0000 Subject: clockevents: Get rid of the notifier chain 7+ years and still a single user. Kill it. Signed-off-by: Thomas Gleixner Cc: John Stultz Cc: Magnus Damm Link: http://lkml.kernel.org/r/20130425143436.098520211@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/time/clockevents.c | 35 +++-------------------------------- kernel/time/tick-broadcast.c | 5 ++--- kernel/time/tick-common.c | 30 +++++------------------------- kernel/time/tick-internal.h | 7 ++++--- 4 files changed, 14 insertions(+), 63 deletions(-) (limited to 'kernel') diff --git a/kernel/time/clockevents.c b/kernel/time/clockevents.c index c6d6400ee137..dd70b4842c62 100644 --- a/kernel/time/clockevents.c +++ b/kernel/time/clockevents.c @@ -15,7 +15,6 @@ #include #include #include -#include #include #include "tick-internal.h" @@ -23,10 +22,6 @@ /* The registered clock event devices */ static LIST_HEAD(clockevent_devices); static LIST_HEAD(clockevents_released); - -/* Notification for clock events */ -static RAW_NOTIFIER_HEAD(clockevents_chain); - /* Protection for the above */ static DEFINE_RAW_SPINLOCK(clockevents_lock); @@ -232,30 +227,6 @@ int clockevents_program_event(struct clock_event_device *dev, ktime_t expires, return (rc && force) ? clockevents_program_min_delta(dev) : rc; } -/** - * clockevents_register_notifier - register a clock events change listener - */ -int clockevents_register_notifier(struct notifier_block *nb) -{ - unsigned long flags; - int ret; - - raw_spin_lock_irqsave(&clockevents_lock, flags); - ret = raw_notifier_chain_register(&clockevents_chain, nb); - raw_spin_unlock_irqrestore(&clockevents_lock, flags); - - return ret; -} - -/* - * Notify about a clock event change. Called with clockevents_lock - * held. - */ -static void clockevents_do_notify(unsigned long reason, void *dev) -{ - raw_notifier_call_chain(&clockevents_chain, reason, dev); -} - /* * Called after a notify add to make devices available which were * released from the notifier call. @@ -269,7 +240,7 @@ static void clockevents_notify_released(void) struct clock_event_device, list); list_del(&dev->list); list_add(&dev->list, &clockevent_devices); - clockevents_do_notify(CLOCK_EVT_NOTIFY_ADD, dev); + tick_check_new_device(dev); } } @@ -290,7 +261,7 @@ void clockevents_register_device(struct clock_event_device *dev) raw_spin_lock_irqsave(&clockevents_lock, flags); list_add(&dev->list, &clockevent_devices); - clockevents_do_notify(CLOCK_EVT_NOTIFY_ADD, dev); + tick_check_new_device(dev); clockevents_notify_released(); raw_spin_unlock_irqrestore(&clockevents_lock, flags); @@ -433,7 +404,7 @@ void clockevents_notify(unsigned long reason, void *arg) int cpu; raw_spin_lock_irqsave(&clockevents_lock, flags); - clockevents_do_notify(reason, arg); + tick_notify(reason, arg); switch (reason) { case CLOCK_EVT_NOTIFY_CPU_DEAD: diff --git a/kernel/time/tick-broadcast.c b/kernel/time/tick-broadcast.c index 24938d577669..3500caaa0bfd 100644 --- a/kernel/time/tick-broadcast.c +++ b/kernel/time/tick-broadcast.c @@ -64,7 +64,7 @@ static void tick_broadcast_start_periodic(struct clock_event_device *bc) /* * Check, if the device can be utilized as broadcast device: */ -int tick_check_broadcast_device(struct clock_event_device *dev) +void tick_install_broadcast_device(struct clock_event_device *dev) { struct clock_event_device *cur = tick_broadcast_device.evtdev; @@ -72,7 +72,7 @@ int tick_check_broadcast_device(struct clock_event_device *dev) (tick_broadcast_device.evtdev && tick_broadcast_device.evtdev->rating >= dev->rating) || (dev->features & CLOCK_EVT_FEAT_C3STOP)) - return 0; + return; clockevents_exchange_device(tick_broadcast_device.evtdev, dev); if (cur) @@ -90,7 +90,6 @@ int tick_check_broadcast_device(struct clock_event_device *dev) */ if (dev->features & CLOCK_EVT_FEAT_ONESHOT) tick_clock_notify(); - return 1; } /* diff --git a/kernel/time/tick-common.c b/kernel/time/tick-common.c index 5d3fb100bc06..dbf4e18d5101 100644 --- a/kernel/time/tick-common.c +++ b/kernel/time/tick-common.c @@ -208,11 +208,11 @@ static void tick_setup_device(struct tick_device *td, /* * Check, if the new registered device should be used. */ -static int tick_check_new_device(struct clock_event_device *newdev) +void tick_check_new_device(struct clock_event_device *newdev) { struct clock_event_device *curdev; struct tick_device *td; - int cpu, ret = NOTIFY_OK; + int cpu; unsigned long flags; raw_spin_lock_irqsave(&tick_device_lock, flags); @@ -275,18 +275,14 @@ static int tick_check_new_device(struct clock_event_device *newdev) tick_oneshot_notify(); raw_spin_unlock_irqrestore(&tick_device_lock, flags); - return NOTIFY_STOP; + return; out_bc: /* * Can the new device be used as a broadcast device ? */ - if (tick_check_broadcast_device(newdev)) - ret = NOTIFY_STOP; - + tick_install_broadcast_device(newdev); raw_spin_unlock_irqrestore(&tick_device_lock, flags); - - return ret; } /* @@ -360,17 +356,10 @@ static void tick_resume(void) raw_spin_unlock_irqrestore(&tick_device_lock, flags); } -/* - * Notification about clock event devices - */ -static int tick_notify(struct notifier_block *nb, unsigned long reason, - void *dev) +void tick_notify(unsigned long reason, void *dev) { switch (reason) { - case CLOCK_EVT_NOTIFY_ADD: - return tick_check_new_device(dev); - case CLOCK_EVT_NOTIFY_BROADCAST_ON: case CLOCK_EVT_NOTIFY_BROADCAST_OFF: case CLOCK_EVT_NOTIFY_BROADCAST_FORCE: @@ -404,21 +393,12 @@ static int tick_notify(struct notifier_block *nb, unsigned long reason, default: break; } - - return NOTIFY_OK; } -static struct notifier_block tick_notifier = { - .notifier_call = tick_notify, -}; - /** * tick_init - initialize the tick control - * - * Register the notifier with the clockevents framework */ void __init tick_init(void) { - clockevents_register_notifier(&tick_notifier); tick_broadcast_init(); } diff --git a/kernel/time/tick-internal.h b/kernel/time/tick-internal.h index f0299eae4602..60742fe6f63d 100644 --- a/kernel/time/tick-internal.h +++ b/kernel/time/tick-internal.h @@ -18,6 +18,8 @@ extern int tick_do_timer_cpu __read_mostly; extern void tick_setup_periodic(struct clock_event_device *dev, int broadcast); extern void tick_handle_periodic(struct clock_event_device *dev); +extern void tick_notify(unsigned long reason, void *dev); +extern void tick_check_new_device(struct clock_event_device *dev); extern void clockevents_shutdown(struct clock_event_device *dev); @@ -90,7 +92,7 @@ static inline bool tick_broadcast_oneshot_available(void) { return false; } */ #ifdef CONFIG_GENERIC_CLOCKEVENTS_BROADCAST extern int tick_device_uses_broadcast(struct clock_event_device *dev, int cpu); -extern int tick_check_broadcast_device(struct clock_event_device *dev); +extern void tick_install_broadcast_device(struct clock_event_device *dev); extern int tick_is_broadcast_device(struct clock_event_device *dev); extern void tick_broadcast_on_off(unsigned long reason, int *oncpu); extern void tick_shutdown_broadcast(unsigned int *cpup); @@ -102,9 +104,8 @@ tick_set_periodic_handler(struct clock_event_device *dev, int broadcast); #else /* !BROADCAST */ -static inline int tick_check_broadcast_device(struct clock_event_device *dev) +static inline void tick_install_broadcast_device(struct clock_event_device *dev) { - return 0; } static inline int tick_is_broadcast_device(struct clock_event_device *dev) -- cgit v1.2.3 From 7126cac426137633e470167524e7bcb590fd49b3 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 25 Apr 2013 20:31:48 +0000 Subject: clockevents: Simplify locking Now that the notifier chain is gone there are no other users and it's pointless to nest tick_device_lock inside of clockevents_lock because there is no other use case. Signed-off-by: Thomas Gleixner Cc: John Stultz Cc: Magnus Damm Link: http://lkml.kernel.org/r/20130425143436.162888472@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/time/tick-common.c | 22 +++++----------------- 1 file changed, 5 insertions(+), 17 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-common.c b/kernel/time/tick-common.c index dbf4e18d5101..170a4bdfa99e 100644 --- a/kernel/time/tick-common.c +++ b/kernel/time/tick-common.c @@ -33,7 +33,6 @@ DEFINE_PER_CPU(struct tick_device, tick_cpu_device); ktime_t tick_next_period; ktime_t tick_period; int tick_do_timer_cpu __read_mostly = TICK_DO_TIMER_BOOT; -static DEFINE_RAW_SPINLOCK(tick_device_lock); /* * Debugging: see timer_list.c @@ -206,16 +205,14 @@ static void tick_setup_device(struct tick_device *td, } /* - * Check, if the new registered device should be used. + * Check, if the new registered device should be used. Called with + * clockevents_lock held and interrupts disabled. */ void tick_check_new_device(struct clock_event_device *newdev) { struct clock_event_device *curdev; struct tick_device *td; int cpu; - unsigned long flags; - - raw_spin_lock_irqsave(&tick_device_lock, flags); cpu = smp_processor_id(); if (!cpumask_test_cpu(cpu, newdev->cpumask)) @@ -273,8 +270,6 @@ void tick_check_new_device(struct clock_event_device *newdev) tick_setup_device(td, newdev, cpu, cpumask_of(cpu)); if (newdev->features & CLOCK_EVT_FEAT_ONESHOT) tick_oneshot_notify(); - - raw_spin_unlock_irqrestore(&tick_device_lock, flags); return; out_bc: @@ -282,7 +277,6 @@ out_bc: * Can the new device be used as a broadcast device ? */ tick_install_broadcast_device(newdev); - raw_spin_unlock_irqrestore(&tick_device_lock, flags); } /* @@ -311,9 +305,7 @@ static void tick_shutdown(unsigned int *cpup) { struct tick_device *td = &per_cpu(tick_cpu_device, *cpup); struct clock_event_device *dev = td->evtdev; - unsigned long flags; - raw_spin_lock_irqsave(&tick_device_lock, flags); td->mode = TICKDEV_MODE_PERIODIC; if (dev) { /* @@ -325,26 +317,20 @@ static void tick_shutdown(unsigned int *cpup) dev->event_handler = clockevents_handle_noop; td->evtdev = NULL; } - raw_spin_unlock_irqrestore(&tick_device_lock, flags); } static void tick_suspend(void) { struct tick_device *td = &__get_cpu_var(tick_cpu_device); - unsigned long flags; - raw_spin_lock_irqsave(&tick_device_lock, flags); clockevents_shutdown(td->evtdev); - raw_spin_unlock_irqrestore(&tick_device_lock, flags); } static void tick_resume(void) { struct tick_device *td = &__get_cpu_var(tick_cpu_device); - unsigned long flags; int broadcast = tick_resume_broadcast(); - raw_spin_lock_irqsave(&tick_device_lock, flags); clockevents_set_mode(td->evtdev, CLOCK_EVT_MODE_RESUME); if (!broadcast) { @@ -353,9 +339,11 @@ static void tick_resume(void) else tick_resume_oneshot(); } - raw_spin_unlock_irqrestore(&tick_device_lock, flags); } +/* + * Called with clockevents_lock held and interrupts disabled + */ void tick_notify(unsigned long reason, void *dev) { switch (reason) { -- cgit v1.2.3 From 8c53daf63f56791ed47fc585206ef3049489612f Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 25 Apr 2013 20:31:48 +0000 Subject: clockevents: Move the tick_notify() switch case to clockevents_notify() No need to call another function and have duplicated cases. Signed-off-by: Thomas Gleixner Cc: John Stultz Cc: Magnus Damm Link: http://lkml.kernel.org/r/20130425143436.235746557@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/time/clockevents.c | 28 ++++++++++++++++++++++++- kernel/time/tick-common.c | 50 ++++----------------------------------------- kernel/time/tick-internal.h | 5 ++++- 3 files changed, 35 insertions(+), 48 deletions(-) (limited to 'kernel') diff --git a/kernel/time/clockevents.c b/kernel/time/clockevents.c index dd70b4842c62..0e3a8448e115 100644 --- a/kernel/time/clockevents.c +++ b/kernel/time/clockevents.c @@ -404,10 +404,36 @@ void clockevents_notify(unsigned long reason, void *arg) int cpu; raw_spin_lock_irqsave(&clockevents_lock, flags); - tick_notify(reason, arg); switch (reason) { + case CLOCK_EVT_NOTIFY_BROADCAST_ON: + case CLOCK_EVT_NOTIFY_BROADCAST_OFF: + case CLOCK_EVT_NOTIFY_BROADCAST_FORCE: + tick_broadcast_on_off(reason, arg); + break; + + case CLOCK_EVT_NOTIFY_BROADCAST_ENTER: + case CLOCK_EVT_NOTIFY_BROADCAST_EXIT: + tick_broadcast_oneshot_control(reason); + break; + + case CLOCK_EVT_NOTIFY_CPU_DYING: + tick_handover_do_timer(arg); + break; + + case CLOCK_EVT_NOTIFY_SUSPEND: + tick_suspend(); + tick_suspend_broadcast(); + break; + + case CLOCK_EVT_NOTIFY_RESUME: + tick_resume(); + break; + case CLOCK_EVT_NOTIFY_CPU_DEAD: + tick_shutdown_broadcast_oneshot(arg); + tick_shutdown_broadcast(arg); + tick_shutdown(arg); /* * Unregister the clock event devices which were * released from the users in the notify chain. diff --git a/kernel/time/tick-common.c b/kernel/time/tick-common.c index 170a4bdfa99e..84c7cfca4d7d 100644 --- a/kernel/time/tick-common.c +++ b/kernel/time/tick-common.c @@ -284,7 +284,7 @@ out_bc: * * Called with interrupts disabled. */ -static void tick_handover_do_timer(int *cpup) +void tick_handover_do_timer(int *cpup) { if (*cpup == tick_do_timer_cpu) { int cpu = cpumask_first(cpu_online_mask); @@ -301,7 +301,7 @@ static void tick_handover_do_timer(int *cpup) * access the hardware device itself. * We just set the mode and remove it from the lists. */ -static void tick_shutdown(unsigned int *cpup) +void tick_shutdown(unsigned int *cpup) { struct tick_device *td = &per_cpu(tick_cpu_device, *cpup); struct clock_event_device *dev = td->evtdev; @@ -319,14 +319,14 @@ static void tick_shutdown(unsigned int *cpup) } } -static void tick_suspend(void) +void tick_suspend(void) { struct tick_device *td = &__get_cpu_var(tick_cpu_device); clockevents_shutdown(td->evtdev); } -static void tick_resume(void) +void tick_resume(void) { struct tick_device *td = &__get_cpu_var(tick_cpu_device); int broadcast = tick_resume_broadcast(); @@ -341,48 +341,6 @@ static void tick_resume(void) } } -/* - * Called with clockevents_lock held and interrupts disabled - */ -void tick_notify(unsigned long reason, void *dev) -{ - switch (reason) { - - case CLOCK_EVT_NOTIFY_BROADCAST_ON: - case CLOCK_EVT_NOTIFY_BROADCAST_OFF: - case CLOCK_EVT_NOTIFY_BROADCAST_FORCE: - tick_broadcast_on_off(reason, dev); - break; - - case CLOCK_EVT_NOTIFY_BROADCAST_ENTER: - case CLOCK_EVT_NOTIFY_BROADCAST_EXIT: - tick_broadcast_oneshot_control(reason); - break; - - case CLOCK_EVT_NOTIFY_CPU_DYING: - tick_handover_do_timer(dev); - break; - - case CLOCK_EVT_NOTIFY_CPU_DEAD: - tick_shutdown_broadcast_oneshot(dev); - tick_shutdown_broadcast(dev); - tick_shutdown(dev); - break; - - case CLOCK_EVT_NOTIFY_SUSPEND: - tick_suspend(); - tick_suspend_broadcast(); - break; - - case CLOCK_EVT_NOTIFY_RESUME: - tick_resume(); - break; - - default: - break; - } -} - /** * tick_init - initialize the tick control */ diff --git a/kernel/time/tick-internal.h b/kernel/time/tick-internal.h index 60742fe6f63d..06bfc8802dfb 100644 --- a/kernel/time/tick-internal.h +++ b/kernel/time/tick-internal.h @@ -18,8 +18,11 @@ extern int tick_do_timer_cpu __read_mostly; extern void tick_setup_periodic(struct clock_event_device *dev, int broadcast); extern void tick_handle_periodic(struct clock_event_device *dev); -extern void tick_notify(unsigned long reason, void *dev); extern void tick_check_new_device(struct clock_event_device *dev); +extern void tick_handover_do_timer(int *cpup); +extern void tick_shutdown(unsigned int *cpup); +extern void tick_suspend(void); +extern void tick_resume(void); extern void clockevents_shutdown(struct clock_event_device *dev); -- cgit v1.2.3 From ccf33d6880f39a35158fff66db13000ae4943fac Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 25 Apr 2013 20:31:49 +0000 Subject: clockevents: Add module refcount We want to be able to remove clockevent modules as well. Add a refcount so we don't remove a module with an active clock event device. Signed-off-by: Thomas Gleixner Cc: John Stultz Cc: Magnus Damm Link: http://lkml.kernel.org/r/20130425143436.307435149@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/time/clockevents.c | 1 + kernel/time/tick-broadcast.c | 3 +++ kernel/time/tick-common.c | 4 ++++ 3 files changed, 8 insertions(+) (limited to 'kernel') diff --git a/kernel/time/clockevents.c b/kernel/time/clockevents.c index 0e3a8448e115..89e394caa769 100644 --- a/kernel/time/clockevents.c +++ b/kernel/time/clockevents.c @@ -357,6 +357,7 @@ void clockevents_exchange_device(struct clock_event_device *old, * released list and do a notify add later. */ if (old) { + module_put(old->owner); clockevents_set_mode(old, CLOCK_EVT_MODE_UNUSED); list_del(&old->list); list_add(&old->list, &clockevents_released); diff --git a/kernel/time/tick-broadcast.c b/kernel/time/tick-broadcast.c index 3500caaa0bfd..0e374cd2e0ef 100644 --- a/kernel/time/tick-broadcast.c +++ b/kernel/time/tick-broadcast.c @@ -19,6 +19,7 @@ #include #include #include +#include #include "tick-internal.h" @@ -73,6 +74,8 @@ void tick_install_broadcast_device(struct clock_event_device *dev) tick_broadcast_device.evtdev->rating >= dev->rating) || (dev->features & CLOCK_EVT_FEAT_C3STOP)) return; + if (!try_module_get(dev->owner)) + return; clockevents_exchange_device(tick_broadcast_device.evtdev, dev); if (cur) diff --git a/kernel/time/tick-common.c b/kernel/time/tick-common.c index 84c7cfca4d7d..433a1e11d13b 100644 --- a/kernel/time/tick-common.c +++ b/kernel/time/tick-common.c @@ -18,6 +18,7 @@ #include #include #include +#include #include @@ -257,6 +258,9 @@ void tick_check_new_device(struct clock_event_device *newdev) goto out_bc; } + if (!try_module_get(newdev->owner)) + return; + /* * Replace the eventually existing device by the new * device. If the current device is the broadcast device, do -- cgit v1.2.3 From 501f867064e95f9a6f540e60705be0937280e7ec Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 25 Apr 2013 20:31:49 +0000 Subject: clockevents: Provide sysfs interface Provide a simple sysfs interface for the clockevent devices. Show the current active clockevent device. Signed-off-by: Thomas Gleixner Cc: John Stultz Cc: Magnus Damm Link: http://lkml.kernel.org/r/20130425143436.371634778@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/time/clockevents.c | 86 +++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 86 insertions(+) (limited to 'kernel') diff --git a/kernel/time/clockevents.c b/kernel/time/clockevents.c index 89e394caa769..0a23f4f29934 100644 --- a/kernel/time/clockevents.c +++ b/kernel/time/clockevents.c @@ -16,6 +16,7 @@ #include #include #include +#include #include "tick-internal.h" @@ -460,4 +461,89 @@ void clockevents_notify(unsigned long reason, void *arg) raw_spin_unlock_irqrestore(&clockevents_lock, flags); } EXPORT_SYMBOL_GPL(clockevents_notify); + +#ifdef CONFIG_SYSFS +struct bus_type clockevents_subsys = { + .name = "clockevents", + .dev_name = "clockevent", +}; + +static DEFINE_PER_CPU(struct device, tick_percpu_dev); +static struct tick_device *tick_get_tick_dev(struct device *dev); + +static ssize_t sysfs_show_current_tick_dev(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct tick_device *td; + ssize_t count = 0; + + raw_spin_lock_irq(&clockevents_lock); + td = tick_get_tick_dev(dev); + if (td && td->evtdev) + count = snprintf(buf, PAGE_SIZE, "%s\n", td->evtdev->name); + raw_spin_unlock_irq(&clockevents_lock); + return count; +} +static DEVICE_ATTR(current_device, 0444, sysfs_show_current_tick_dev, NULL); + +#ifdef CONFIG_GENERIC_CLOCKEVENTS_BROADCAST +static struct device tick_bc_dev = { + .init_name = "broadcast", + .id = 0, + .bus = &clockevents_subsys, +}; + +static struct tick_device *tick_get_tick_dev(struct device *dev) +{ + return dev == &tick_bc_dev ? tick_get_broadcast_device() : + &per_cpu(tick_cpu_device, dev->id); +} + +static __init int tick_broadcast_init_sysfs(void) +{ + int err = device_register(&tick_bc_dev); + + if (!err) + err = device_create_file(&tick_bc_dev, &dev_attr_current_device); + return err; +} +#else +static struct tick_device *tick_get_tick_dev(struct device *dev) +{ + return &per_cpu(tick_cpu_device, dev->id); +} +static inline int tick_broadcast_init_sysfs(void) { return 0; } #endif + +static int __init tick_init_sysfs(void) +{ + int cpu; + + for_each_possible_cpu(cpu) { + struct device *dev = &per_cpu(tick_percpu_dev, cpu); + int err; + + dev->id = cpu; + dev->bus = &clockevents_subsys; + err = device_register(dev); + if (!err) + err = device_create_file(dev, &dev_attr_current_device); + if (err) + return err; + } + return tick_broadcast_init_sysfs(); +} + +static int __init clockevents_init_sysfs(void) +{ + int err = subsys_system_register(&clockevents_subsys, NULL); + + if (!err) + err = tick_init_sysfs(); + return err; +} +device_initcall(clockevents_init_sysfs); +#endif /* SYSFS */ + +#endif /* GENERIC_CLOCK_EVENTS */ -- cgit v1.2.3 From 45cb8e01b2ecef1c2afb18333e95793fa1a90281 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 25 Apr 2013 20:31:50 +0000 Subject: clockevents: Split out selection logic Split out the clockevent device selection logic. Preparatory patch to allow unbinding active clockevent devices. Signed-off-by: Thomas Gleixner Cc: John Stultz Cc: Magnus Damm Link: http://lkml.kernel.org/r/20130425143436.431796247@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/time/tick-broadcast.c | 25 ++++++++++++---- kernel/time/tick-common.c | 69 +++++++++++++++++++++++--------------------- 2 files changed, 56 insertions(+), 38 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-broadcast.c b/kernel/time/tick-broadcast.c index 0e374cd2e0ef..d067c01586f5 100644 --- a/kernel/time/tick-broadcast.c +++ b/kernel/time/tick-broadcast.c @@ -65,19 +65,34 @@ static void tick_broadcast_start_periodic(struct clock_event_device *bc) /* * Check, if the device can be utilized as broadcast device: */ +static bool tick_check_broadcast_device(struct clock_event_device *curdev, + struct clock_event_device *newdev) +{ + if ((newdev->features & CLOCK_EVT_FEAT_DUMMY) || + (newdev->features & CLOCK_EVT_FEAT_C3STOP)) + return false; + + if (tick_broadcast_device.mode == TICKDEV_MODE_ONESHOT && + !(newdev->features & CLOCK_EVT_FEAT_ONESHOT)) + return false; + + return !curdev || newdev->rating > curdev->rating; +} + +/* + * Conditionally install/replace broadcast device + */ void tick_install_broadcast_device(struct clock_event_device *dev) { struct clock_event_device *cur = tick_broadcast_device.evtdev; - if ((dev->features & CLOCK_EVT_FEAT_DUMMY) || - (tick_broadcast_device.evtdev && - tick_broadcast_device.evtdev->rating >= dev->rating) || - (dev->features & CLOCK_EVT_FEAT_C3STOP)) + if (!tick_check_broadcast_device(cur, dev)) return; + if (!try_module_get(dev->owner)) return; - clockevents_exchange_device(tick_broadcast_device.evtdev, dev); + clockevents_exchange_device(cur, dev); if (cur) cur->event_handler = clockevents_handle_noop; tick_broadcast_device.evtdev = dev; diff --git a/kernel/time/tick-common.c b/kernel/time/tick-common.c index 433a1e11d13b..c34021650348 100644 --- a/kernel/time/tick-common.c +++ b/kernel/time/tick-common.c @@ -205,6 +205,37 @@ static void tick_setup_device(struct tick_device *td, tick_setup_oneshot(newdev, handler, next_event); } +static bool tick_check_percpu(struct clock_event_device *curdev, + struct clock_event_device *newdev, int cpu) +{ + if (!cpumask_test_cpu(cpu, newdev->cpumask)) + return false; + if (cpumask_equal(newdev->cpumask, cpumask_of(cpu))) + return true; + /* Check if irq affinity can be set */ + if (newdev->irq >= 0 && !irq_can_set_affinity(newdev->irq)) + return false; + /* Prefer an existing cpu local device */ + if (curdev && cpumask_equal(curdev->cpumask, cpumask_of(cpu))) + return false; + return true; +} + +static bool tick_check_preferred(struct clock_event_device *curdev, + struct clock_event_device *newdev) +{ + /* Prefer oneshot capable device */ + if (!(newdev->features & CLOCK_EVT_FEAT_ONESHOT)) { + if (curdev && (curdev->features & CLOCK_EVT_FEAT_ONESHOT)) + return false; + if (tick_oneshot_mode_active()) + return false; + } + + /* Use the higher rated one */ + return !curdev || newdev->rating > curdev->rating; +} + /* * Check, if the new registered device should be used. Called with * clockevents_lock held and interrupts disabled. @@ -223,40 +254,12 @@ void tick_check_new_device(struct clock_event_device *newdev) curdev = td->evtdev; /* cpu local device ? */ - if (!cpumask_equal(newdev->cpumask, cpumask_of(cpu))) { - - /* - * If the cpu affinity of the device interrupt can not - * be set, ignore it. - */ - if (!irq_can_set_affinity(newdev->irq)) - goto out_bc; - - /* - * If we have a cpu local device already, do not replace it - * by a non cpu local device - */ - if (curdev && cpumask_equal(curdev->cpumask, cpumask_of(cpu))) - goto out_bc; - } + if (!tick_check_percpu(curdev, newdev, cpu)) + goto out_bc; - /* - * If we have an active device, then check the rating and the oneshot - * feature. - */ - if (curdev) { - /* - * Prefer one shot capable devices ! - */ - if ((curdev->features & CLOCK_EVT_FEAT_ONESHOT) && - !(newdev->features & CLOCK_EVT_FEAT_ONESHOT)) - goto out_bc; - /* - * Check the rating - */ - if (curdev->rating >= newdev->rating) - goto out_bc; - } + /* Preference decision */ + if (!tick_check_preferred(curdev, newdev)) + goto out_bc; if (!try_module_get(newdev->owner)) return; -- cgit v1.2.3 From 03e13cf5ee60584fe0c831682c67212effb7fca4 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 25 Apr 2013 20:31:50 +0000 Subject: clockevents: Implement unbind functionality Provide a sysfs interface to allow unbinding of clockevent devices. The device is unbound if it is unused or if there is a replacement device available. Unbinding of broadcast devices is not supported as we don't want to foster that nonsense. If no replacement device is available the unbind returns -EBUSY. Unbind is available from the kernel and through sysfs, which is necessary to drop the module refcount. Signed-off-by: Thomas Gleixner Cc: John Stultz Cc: Magnus Damm Link: http://lkml.kernel.org/r/20130425143436.499216659@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/time/clockevents.c | 125 ++++++++++++++++++++++++++++++++++++++++++++ kernel/time/clocksource.c | 9 ++-- kernel/time/tick-common.c | 24 +++++++++ kernel/time/tick-internal.h | 7 +++ 4 files changed, 161 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/time/clockevents.c b/kernel/time/clockevents.c index 0a23f4f29934..38959c866789 100644 --- a/kernel/time/clockevents.c +++ b/kernel/time/clockevents.c @@ -25,6 +25,13 @@ static LIST_HEAD(clockevent_devices); static LIST_HEAD(clockevents_released); /* Protection for the above */ static DEFINE_RAW_SPINLOCK(clockevents_lock); +/* Protection for unbind operations */ +static DEFINE_MUTEX(clockevents_mutex); + +struct ce_unbind { + struct clock_event_device *ce; + int res; +}; /** * clockevents_delta2ns - Convert a latch value (device ticks) to nanoseconds @@ -245,6 +252,90 @@ static void clockevents_notify_released(void) } } +/* + * Try to install a replacement clock event device + */ +static int clockevents_replace(struct clock_event_device *ced) +{ + struct clock_event_device *dev, *newdev = NULL; + + list_for_each_entry(dev, &clockevent_devices, list) { + if (dev == ced || dev->mode != CLOCK_EVT_MODE_UNUSED) + continue; + + if (!tick_check_replacement(newdev, dev)) + continue; + + if (!try_module_get(dev->owner)) + continue; + + if (newdev) + module_put(newdev->owner); + newdev = dev; + } + if (newdev) { + tick_install_replacement(newdev); + list_del_init(&ced->list); + } + return newdev ? 0 : -EBUSY; +} + +/* + * Called with clockevents_mutex and clockevents_lock held + */ +static int __clockevents_try_unbind(struct clock_event_device *ced, int cpu) +{ + /* Fast track. Device is unused */ + if (ced->mode == CLOCK_EVT_MODE_UNUSED) { + list_del_init(&ced->list); + return 0; + } + + return ced == per_cpu(tick_cpu_device, cpu).evtdev ? -EAGAIN : -EBUSY; +} + +/* + * SMP function call to unbind a device + */ +static void __clockevents_unbind(void *arg) +{ + struct ce_unbind *cu = arg; + int res; + + raw_spin_lock(&clockevents_lock); + res = __clockevents_try_unbind(cu->ce, smp_processor_id()); + if (res == -EAGAIN) + res = clockevents_replace(cu->ce); + cu->res = res; + raw_spin_unlock(&clockevents_lock); +} + +/* + * Issues smp function call to unbind a per cpu device. Called with + * clockevents_mutex held. + */ +static int clockevents_unbind(struct clock_event_device *ced, int cpu) +{ + struct ce_unbind cu = { .ce = ced, .res = -ENODEV }; + + smp_call_function_single(cpu, __clockevents_unbind, &cu, 1); + return cu.res; +} + +/* + * Unbind a clockevents device. + */ +int clockevents_unbind_device(struct clock_event_device *ced, int cpu) +{ + int ret; + + mutex_lock(&clockevents_mutex); + ret = clockevents_unbind(ced, cpu); + mutex_unlock(&clockevents_mutex); + return ret; +} +EXPORT_SYMBOL_GPL(clockevents_unbind); + /** * clockevents_register_device - register a clock event device * @dev: device to register @@ -487,6 +578,38 @@ static ssize_t sysfs_show_current_tick_dev(struct device *dev, } static DEVICE_ATTR(current_device, 0444, sysfs_show_current_tick_dev, NULL); +/* We don't support the abomination of removable broadcast devices */ +static ssize_t sysfs_unbind_tick_dev(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count) +{ + char name[CS_NAME_LEN]; + size_t ret = sysfs_get_uname(buf, name, count); + struct clock_event_device *ce; + + if (ret < 0) + return ret; + + ret = -ENODEV; + mutex_lock(&clockevents_mutex); + raw_spin_lock_irq(&clockevents_lock); + list_for_each_entry(ce, &clockevent_devices, list) { + if (!strcmp(ce->name, name)) { + ret = __clockevents_try_unbind(ce, dev->id); + break; + } + } + raw_spin_unlock_irq(&clockevents_lock); + /* + * We hold clockevents_mutex, so ce can't go away + */ + if (ret == -EAGAIN) + ret = clockevents_unbind(ce, dev->id); + mutex_unlock(&clockevents_mutex); + return ret ? ret : count; +} +static DEVICE_ATTR(unbind_device, 0200, NULL, sysfs_unbind_tick_dev); + #ifdef CONFIG_GENERIC_CLOCKEVENTS_BROADCAST static struct device tick_bc_dev = { .init_name = "broadcast", @@ -529,6 +652,8 @@ static int __init tick_init_sysfs(void) err = device_register(dev); if (!err) err = device_create_file(dev, &dev_attr_current_device); + if (!err) + err = device_create_file(dev, &dev_attr_unbind_device); if (err) return err; } diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c index 31b90332f47b..6d05b00410cc 100644 --- a/kernel/time/clocksource.c +++ b/kernel/time/clocksource.c @@ -31,6 +31,8 @@ #include #include +#include "tick-internal.h" + void timecounter_init(struct timecounter *tc, const struct cyclecounter *cc, u64 start_tstamp) @@ -174,7 +176,6 @@ clocks_calc_mult_shift(u32 *mult, u32 *shift, u32 from, u32 to, u32 maxsec) static struct clocksource *curr_clocksource; static LIST_HEAD(clocksource_list); static DEFINE_MUTEX(clocksource_mutex); -#define CS_NAME_LEN 32 static char override_name[CS_NAME_LEN]; static int finished_booting; @@ -864,7 +865,7 @@ sysfs_show_current_clocksources(struct device *dev, return count; } -static size_t clocksource_get_uname(const char *buf, char *dst, size_t cnt) +size_t sysfs_get_uname(const char *buf, char *dst, size_t cnt) { size_t ret = cnt; @@ -899,7 +900,7 @@ static ssize_t sysfs_override_clocksource(struct device *dev, mutex_lock(&clocksource_mutex); - ret = clocksource_get_uname(buf, override_name, count); + ret = sysfs_get_uname(buf, override_name, count); if (ret >= 0) clocksource_select(); @@ -925,7 +926,7 @@ static ssize_t sysfs_unbind_clocksource(struct device *dev, char name[CS_NAME_LEN]; size_t ret; - ret = clocksource_get_uname(buf, name, count); + ret = sysfs_get_uname(buf, name, count); if (ret < 0) return ret; diff --git a/kernel/time/tick-common.c b/kernel/time/tick-common.c index c34021650348..5edfb4806032 100644 --- a/kernel/time/tick-common.c +++ b/kernel/time/tick-common.c @@ -205,6 +205,17 @@ static void tick_setup_device(struct tick_device *td, tick_setup_oneshot(newdev, handler, next_event); } +void tick_install_replacement(struct clock_event_device *newdev) +{ + struct tick_device *td = &__get_cpu_var(tick_cpu_device); + int cpu = smp_processor_id(); + + clockevents_exchange_device(td->evtdev, newdev); + tick_setup_device(td, newdev, cpu, cpumask_of(cpu)); + if (newdev->features & CLOCK_EVT_FEAT_ONESHOT) + tick_oneshot_notify(); +} + static bool tick_check_percpu(struct clock_event_device *curdev, struct clock_event_device *newdev, int cpu) { @@ -236,6 +247,19 @@ static bool tick_check_preferred(struct clock_event_device *curdev, return !curdev || newdev->rating > curdev->rating; } +/* + * Check whether the new device is a better fit than curdev. curdev + * can be NULL ! + */ +bool tick_check_replacement(struct clock_event_device *curdev, + struct clock_event_device *newdev) +{ + if (tick_check_percpu(curdev, newdev, smp_processor_id())) + return false; + + return tick_check_preferred(curdev, newdev); +} + /* * Check, if the new registered device should be used. Called with * clockevents_lock held and interrupts disabled. diff --git a/kernel/time/tick-internal.h b/kernel/time/tick-internal.h index 06bfc8802dfb..be1690eaecff 100644 --- a/kernel/time/tick-internal.h +++ b/kernel/time/tick-internal.h @@ -11,6 +11,8 @@ extern seqlock_t jiffies_lock; #define TICK_DO_TIMER_NONE -1 #define TICK_DO_TIMER_BOOT -2 +#define CS_NAME_LEN 32 + DECLARE_PER_CPU(struct tick_device, tick_cpu_device); extern ktime_t tick_next_period; extern ktime_t tick_period; @@ -23,9 +25,14 @@ extern void tick_handover_do_timer(int *cpup); extern void tick_shutdown(unsigned int *cpup); extern void tick_suspend(void); extern void tick_resume(void); +extern bool tick_check_replacement(struct clock_event_device *curdev, + struct clock_event_device *newdev); +extern void tick_install_replacement(struct clock_event_device *dev); extern void clockevents_shutdown(struct clock_event_device *dev); +extern size_t sysfs_get_uname(const char *buf, char *dst, size_t cnt); + /* * NO_HZ / high resolution timer shared code */ -- cgit v1.2.3 From bdc7119f1bdd0632d42f435941dc290216a436e7 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 24 May 2013 10:55:38 +0900 Subject: cgroup: make cgroup_is_removed() static cgroup_is_removed() no longer has external users and it shouldn't grow any - controllers should deal with cgroup_subsys_state on/offline state instead of cgroup removal state. Make it static. While at it, make it return bool. Signed-off-by: Tejun Heo --- kernel/cgroup.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index a19419f4af1a..501974823b33 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -226,7 +226,7 @@ static int css_refcnt(struct cgroup_subsys_state *css) } /* convenient tests for these bits */ -inline int cgroup_is_removed(const struct cgroup *cgrp) +static inline bool cgroup_is_removed(const struct cgroup *cgrp) { return test_bit(CGRP_REMOVED, &cgrp->flags); } -- cgit v1.2.3 From 53fa5261747a90746531e8a1c81eeb78fedc2f71 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 24 May 2013 10:55:38 +0900 Subject: cgroup: add cgroup->serial_nr and implement cgroup_next_sibling() Currently, there's no easy way to find out the next sibling cgroup unless it's known that the current cgroup is accessed from the parent's children list in a single RCU critical section. This in turn forces all iterators to require whole iteration to be enclosed in a single RCU critical section, which sometimes is too restrictive. This patch implements cgroup_next_sibling() which can reliably determine the next sibling regardless of the state of the current cgroup as long as it's accessible. It currently is impossible to determine the next sibling after dropping RCU read lock because the cgroup being iterated could be removed anytime and if RCU read lock is dropped, nothing guarantess its ->sibling.next pointer is accessible. A removed cgroup would continue to point to its next sibling for RCU accesses but stop receiving updates from the sibling. IOW, the next sibling could be removed and then complete its grace period while RCU read lock is dropped, making it unsafe to dereference ->sibling.next after dropping and re-acquiring RCU read lock. This can be solved by adding a way to traverse to the next sibling without dereferencing ->sibling.next. This patch adds a monotonically increasing cgroup serial number, cgroup->serial_nr, which guarantees that all cgroup->children lists are kept in increasing serial_nr order. A new function, cgroup_next_sibling(), is implemented, which, if CGRP_REMOVED is not set on the current cgroup, follows ->sibling.next; otherwise, traverses the parent's ->children list until it sees a sibling with higher ->serial_nr. This allows the function to always return the next sibling regardless of the state of the current cgroup without adding overhead in the fast path. Further patches will update the iterators to use cgroup_next_sibling() so that they allow dropping RCU read lock and blocking while iteration is in progress which in turn will be used to simplify controllers. v2: Typo fix as per Serge. Signed-off-by: Tejun Heo Acked-by: Serge E. Hallyn --- kernel/cgroup.c | 62 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 62 insertions(+) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 501974823b33..b87c7a5a5497 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -2975,6 +2975,55 @@ static void cgroup_enable_task_cg_lists(void) write_unlock(&css_set_lock); } +/** + * cgroup_next_sibling - find the next sibling of a given cgroup + * @pos: the current cgroup + * + * This function returns the next sibling of @pos and should be called + * under RCU read lock. The only requirement is that @pos is accessible. + * The next sibling is guaranteed to be returned regardless of @pos's + * state. + */ +struct cgroup *cgroup_next_sibling(struct cgroup *pos) +{ + struct cgroup *next; + + WARN_ON_ONCE(!rcu_read_lock_held()); + + /* + * @pos could already have been removed. Once a cgroup is removed, + * its ->sibling.next is no longer updated when its next sibling + * changes. As CGRP_REMOVED is set on removal which is fully + * serialized, if we see it unasserted, it's guaranteed that the + * next sibling hasn't finished its grace period even if it's + * already removed, and thus safe to dereference from this RCU + * critical section. If ->sibling.next is inaccessible, + * cgroup_is_removed() is guaranteed to be visible as %true here. + */ + if (likely(!cgroup_is_removed(pos))) { + next = list_entry_rcu(pos->sibling.next, struct cgroup, sibling); + if (&next->sibling != &pos->parent->children) + return next; + return NULL; + } + + /* + * Can't dereference the next pointer. Each cgroup is given a + * monotonically increasing unique serial number and always + * appended to the sibling list, so the next one can be found by + * walking the parent's children until we see a cgroup with higher + * serial number than @pos's. + * + * While this path can be slow, it's taken only when either the + * current cgroup is removed or iteration and removal race. + */ + list_for_each_entry_rcu(next, &pos->parent->children, sibling) + if (next->serial_nr > pos->serial_nr) + return next; + return NULL; +} +EXPORT_SYMBOL_GPL(cgroup_next_sibling); + /** * cgroup_next_descendant_pre - find the next descendant for pre-order walk * @pos: the current position (%NULL to initiate traversal) @@ -4137,6 +4186,7 @@ static void offline_css(struct cgroup_subsys *ss, struct cgroup *cgrp) static long cgroup_create(struct cgroup *parent, struct dentry *dentry, umode_t mode) { + static atomic64_t serial_nr_cursor = ATOMIC64_INIT(0); struct cgroup *cgrp; struct cgroup_name *name; struct cgroupfs_root *root = parent->root; @@ -4217,6 +4267,14 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, goto err_free_all; lockdep_assert_held(&dentry->d_inode->i_mutex); + /* + * Assign a monotonically increasing serial number. With the list + * appending below, it guarantees that sibling cgroups are always + * sorted in the ascending serial number order on the parent's + * ->children. + */ + cgrp->serial_nr = atomic64_inc_return(&serial_nr_cursor); + /* allocation complete, commit to creation */ list_add_tail(&cgrp->allcg_node, &root->allcg_list); list_add_tail_rcu(&cgrp->sibling, &cgrp->parent->children); @@ -4304,6 +4362,10 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) * removed. This makes future css_tryget() and child creation * attempts fail thus maintaining the removal conditions verified * above. + * + * Note that CGRP_REMVOED clearing is depended upon by + * cgroup_next_sibling() to resume iteration after dropping RCU + * read lock. See cgroup_next_sibling() for details. */ for_each_subsys(cgrp->root, ss) { struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; -- cgit v1.2.3 From 75501a6d59e989e5c286716e5b3b66ace4660e83 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 24 May 2013 10:55:38 +0900 Subject: cgroup: update iterators to use cgroup_next_sibling() This patch converts cgroup_for_each_child(), cgroup_next_descendant_pre/post() and thus cgroup_for_each_descendant_pre/post() to use cgroup_next_sibling() instead of manually dereferencing ->sibling.next. The only reason the iterators couldn't allow dropping RCU read lock while iteration is in progress was because they couldn't determine the next sibling safely once RCU read lock is dropped. Using cgroup_next_sibling() removes that problem and enables all iterators to allow dropping RCU read lock in the middle. Comments are updated accordingly. This makes the iterators easier to use and will simplify controllers. Note that @cgroup argument is renamed to @cgrp in cgroup_for_each_child() because it conflicts with "struct cgroup" used in the new macro body. Signed-off-by: Tejun Heo Acked-by: Serge E. Hallyn Reviewed-by: Michal Hocko --- kernel/cgroup.c | 25 +++++++++++++++++++------ 1 file changed, 19 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index b87c7a5a5497..fefc41c1a147 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -3031,6 +3031,11 @@ EXPORT_SYMBOL_GPL(cgroup_next_sibling); * * To be used by cgroup_for_each_descendant_pre(). Find the next * descendant to visit for pre-order traversal of @cgroup's descendants. + * + * While this function requires RCU read locking, it doesn't require the + * whole traversal to be contained in a single RCU critical section. This + * function will return the correct next descendant as long as both @pos + * and @cgroup are accessible and @pos is a descendant of @cgroup. */ struct cgroup *cgroup_next_descendant_pre(struct cgroup *pos, struct cgroup *cgroup) @@ -3050,11 +3055,9 @@ struct cgroup *cgroup_next_descendant_pre(struct cgroup *pos, /* no child, visit my or the closest ancestor's next sibling */ while (pos != cgroup) { - next = list_entry_rcu(pos->sibling.next, struct cgroup, - sibling); - if (&next->sibling != &pos->parent->children) + next = cgroup_next_sibling(pos); + if (next) return next; - pos = pos->parent; } @@ -3069,6 +3072,11 @@ EXPORT_SYMBOL_GPL(cgroup_next_descendant_pre); * Return the rightmost descendant of @pos. If there's no descendant, * @pos is returned. This can be used during pre-order traversal to skip * subtree of @pos. + * + * While this function requires RCU read locking, it doesn't require the + * whole traversal to be contained in a single RCU critical section. This + * function will return the correct rightmost descendant as long as @pos is + * accessible. */ struct cgroup *cgroup_rightmost_descendant(struct cgroup *pos) { @@ -3108,6 +3116,11 @@ static struct cgroup *cgroup_leftmost_descendant(struct cgroup *pos) * * To be used by cgroup_for_each_descendant_post(). Find the next * descendant to visit for post-order traversal of @cgroup's descendants. + * + * While this function requires RCU read locking, it doesn't require the + * whole traversal to be contained in a single RCU critical section. This + * function will return the correct next descendant as long as both @pos + * and @cgroup are accessible and @pos is a descendant of @cgroup. */ struct cgroup *cgroup_next_descendant_post(struct cgroup *pos, struct cgroup *cgroup) @@ -3123,8 +3136,8 @@ struct cgroup *cgroup_next_descendant_post(struct cgroup *pos, } /* if there's an unvisited sibling, visit its leftmost descendant */ - next = list_entry_rcu(pos->sibling.next, struct cgroup, sibling); - if (&next->sibling != &pos->parent->children) + next = cgroup_next_sibling(pos); + if (next) return cgroup_leftmost_descendant(next); /* no sibling left, visit parent */ -- cgit v1.2.3 From 4eedb77a9cd8f2e68b31c8b9a20524a50727c16f Mon Sep 17 00:00:00 2001 From: Paul Bolle Date: Fri, 17 May 2013 10:51:33 +0200 Subject: locking: Fix copy/paste errors of "ARCH_INLINE_*_UNLOCK_BH" The Kconfig symbols ARCH_INLINE_READ_UNLOCK_IRQ, ARCH_INLINE_SPIN_UNLOCK_IRQ, and ARCH_INLINE_WRITE_UNLOCK_IRQ were added in v2.6.33, but have never actually been used. Ingo Molnar spotted that this is caused by three identical copy/paste erros. Eg, the Kconfig entry for INLINE_READ_UNLOCK_IRQ has an (optional) dependency on: ARCH_INLINE_READ_UNLOCK_BH were it apparently should depend on: ARCH_INLINE_READ_UNLOCK_IRQ instead. Likewise for the Kconfig entries for INLINE_SPIN_UNLOCK_IRQ and INLINE_WRITE_UNLOCK_IRQ. Fix these three errors. This never really caused any real problems as these symbols are set (or unset) in a group - but it's worth fixing it nevertheless. Reported-by: Ingo Molnar Signed-off-by: Paul Bolle Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1368780693.1350.228.camel@x61.thuisdomein Signed-off-by: Ingo Molnar --- kernel/Kconfig.locks | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/Kconfig.locks b/kernel/Kconfig.locks index 44511d100eaa..d2b32ac27a39 100644 --- a/kernel/Kconfig.locks +++ b/kernel/Kconfig.locks @@ -138,7 +138,7 @@ config INLINE_SPIN_UNLOCK_BH config INLINE_SPIN_UNLOCK_IRQ def_bool y - depends on !PREEMPT || ARCH_INLINE_SPIN_UNLOCK_BH + depends on !PREEMPT || ARCH_INLINE_SPIN_UNLOCK_IRQ config INLINE_SPIN_UNLOCK_IRQRESTORE def_bool y @@ -175,7 +175,7 @@ config INLINE_READ_UNLOCK_BH config INLINE_READ_UNLOCK_IRQ def_bool y - depends on !PREEMPT || ARCH_INLINE_READ_UNLOCK_BH + depends on !PREEMPT || ARCH_INLINE_READ_UNLOCK_IRQ config INLINE_READ_UNLOCK_IRQRESTORE def_bool y @@ -212,7 +212,7 @@ config INLINE_WRITE_UNLOCK_BH config INLINE_WRITE_UNLOCK_IRQ def_bool y - depends on !PREEMPT || ARCH_INLINE_WRITE_UNLOCK_BH + depends on !PREEMPT || ARCH_INLINE_WRITE_UNLOCK_IRQ config INLINE_WRITE_UNLOCK_IRQRESTORE def_bool y -- cgit v1.2.3 From ab573844e3058eef2788803d373019f8bebead57 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Wed, 1 May 2013 17:25:44 +0200 Subject: perf: Fix hw breakpoints overflow period sampling The hw breakpoint pmu 'add' function is missing the period_left update needed for SW events. The perf HW breakpoint events use the SW events framework to process the overflow, so it needs to be properly initialized in the PMU 'add' method. Signed-off-by: Jiri Olsa Reviewed-by: Peter Zijlstra Cc: H. Peter Anvin Cc: Oleg Nesterov Cc: Arnaldo Carvalho de Melo Cc: Ingo Molnar Cc: Paul Mackerras Cc: Corey Ashford Cc: Frederic Weisbecker Cc: Vince Weaver Cc: Stephane Eranian Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1367421944-19082-5-git-send-email-jolsa@redhat.com Signed-off-by: Ingo Molnar --- kernel/events/core.c | 2 +- kernel/events/hw_breakpoint.c | 5 +++++ 2 files changed, 6 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 9dc297faf7c0..e0dcced282e4 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -4961,7 +4961,7 @@ static DEFINE_PER_CPU(struct swevent_htable, swevent_htable); * sign as trigger. */ -static u64 perf_swevent_set_period(struct perf_event *event) +u64 perf_swevent_set_period(struct perf_event *event) { struct hw_perf_event *hwc = &event->hw; u64 period = hwc->last_period; diff --git a/kernel/events/hw_breakpoint.c b/kernel/events/hw_breakpoint.c index a64f8aeb5c1f..966a241e8616 100644 --- a/kernel/events/hw_breakpoint.c +++ b/kernel/events/hw_breakpoint.c @@ -612,6 +612,11 @@ static int hw_breakpoint_add(struct perf_event *bp, int flags) if (!(flags & PERF_EF_START)) bp->hw.state = PERF_HES_STOPPED; + if (is_sampling_event(bp)) { + bp->hw.last_period = bp->hw.sample_period; + perf_swevent_set_period(bp); + } + return arch_install_hw_breakpoint(bp); } -- cgit v1.2.3 From 9e6302056f8029f438e853432a856b9f13de26a6 Mon Sep 17 00:00:00 2001 From: Stephane Eranian Date: Wed, 3 Apr 2013 14:21:33 +0200 Subject: perf: Use hrtimers for event multiplexing The current scheme of using the timer tick was fine for per-thread events. However, it was causing bias issues in system-wide mode (including for uncore PMUs). Event groups would not get their fair share of runtime on the PMU. With tickless kernels, if a core is idle there is no timer tick, and thus no event rotation (multiplexing). However, there are events (especially uncore events) which do count even though cores are asleep. This patch changes the timer source for multiplexing. It introduces a per-PMU per-cpu hrtimer. The advantage is that even when a core goes idle, it will come back to service the hrtimer, thus multiplexing on system-wide events works much better. The per-PMU implementation (suggested by PeterZ) enables adjusting the multiplexing interval per PMU. The preferred interval is stashed into the struct pmu. If not set, it will be forced to the default interval value. In order to minimize the impact of the hrtimer, it is turned on and off on demand. When the PMU on a CPU is overcommited, the hrtimer is activated. It is stopped when the PMU is not overcommitted. In order for this to work properly, we had to change the order of initialization in start_kernel() such that hrtimer_init() is run before perf_event_init(). The default interval in milliseconds is set to a timer tick just like with the old code. We will provide a sysctl to tune this in another patch. Signed-off-by: Stephane Eranian Signed-off-by: Peter Zijlstra Cc: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo Link: http://lkml.kernel.org/r/1364991694-5876-2-git-send-email-eranian@google.com Signed-off-by: Ingo Molnar --- kernel/events/core.c | 114 +++++++++++++++++++++++++++++++++++++++++++++++---- 1 file changed, 106 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index e0dcced282e4..97bfac7e6f45 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -170,6 +170,8 @@ int sysctl_perf_event_sample_rate __read_mostly = DEFAULT_MAX_SAMPLE_RATE; static int max_samples_per_tick __read_mostly = DIV_ROUND_UP(DEFAULT_MAX_SAMPLE_RATE, HZ); +static int perf_rotate_context(struct perf_cpu_context *cpuctx); + int perf_proc_update_handler(struct ctl_table *table, int write, void __user *buffer, size_t *lenp, loff_t *ppos) @@ -658,6 +660,98 @@ perf_cgroup_mark_enabled(struct perf_event *event, } #endif +/* + * set default to be dependent on timer tick just + * like original code + */ +#define PERF_CPU_HRTIMER (1000 / HZ) +/* + * function must be called with interrupts disbled + */ +static enum hrtimer_restart perf_cpu_hrtimer_handler(struct hrtimer *hr) +{ + struct perf_cpu_context *cpuctx; + enum hrtimer_restart ret = HRTIMER_NORESTART; + int rotations = 0; + + WARN_ON(!irqs_disabled()); + + cpuctx = container_of(hr, struct perf_cpu_context, hrtimer); + + rotations = perf_rotate_context(cpuctx); + + /* + * arm timer if needed + */ + if (rotations) { + hrtimer_forward_now(hr, cpuctx->hrtimer_interval); + ret = HRTIMER_RESTART; + } + + return ret; +} + +/* CPU is going down */ +void perf_cpu_hrtimer_cancel(int cpu) +{ + struct perf_cpu_context *cpuctx; + struct pmu *pmu; + unsigned long flags; + + if (WARN_ON(cpu != smp_processor_id())) + return; + + local_irq_save(flags); + + rcu_read_lock(); + + list_for_each_entry_rcu(pmu, &pmus, entry) { + cpuctx = this_cpu_ptr(pmu->pmu_cpu_context); + + if (pmu->task_ctx_nr == perf_sw_context) + continue; + + hrtimer_cancel(&cpuctx->hrtimer); + } + + rcu_read_unlock(); + + local_irq_restore(flags); +} + +static void __perf_cpu_hrtimer_init(struct perf_cpu_context *cpuctx, int cpu) +{ + struct hrtimer *hr = &cpuctx->hrtimer; + struct pmu *pmu = cpuctx->ctx.pmu; + + /* no multiplexing needed for SW PMU */ + if (pmu->task_ctx_nr == perf_sw_context) + return; + + cpuctx->hrtimer_interval = + ns_to_ktime(NSEC_PER_MSEC * PERF_CPU_HRTIMER); + + hrtimer_init(hr, CLOCK_MONOTONIC, HRTIMER_MODE_REL_PINNED); + hr->function = perf_cpu_hrtimer_handler; +} + +static void perf_cpu_hrtimer_restart(struct perf_cpu_context *cpuctx) +{ + struct hrtimer *hr = &cpuctx->hrtimer; + struct pmu *pmu = cpuctx->ctx.pmu; + + /* not for SW PMU */ + if (pmu->task_ctx_nr == perf_sw_context) + return; + + if (hrtimer_active(hr)) + return; + + if (!hrtimer_callback_running(hr)) + __hrtimer_start_range_ns(hr, cpuctx->hrtimer_interval, + 0, HRTIMER_MODE_REL_PINNED, 0); +} + void perf_pmu_disable(struct pmu *pmu) { int *count = this_cpu_ptr(pmu->pmu_disable_count); @@ -1506,6 +1600,7 @@ group_sched_in(struct perf_event *group_event, if (event_sched_in(group_event, cpuctx, ctx)) { pmu->cancel_txn(pmu); + perf_cpu_hrtimer_restart(cpuctx); return -EAGAIN; } @@ -1552,6 +1647,8 @@ group_error: pmu->cancel_txn(pmu); + perf_cpu_hrtimer_restart(cpuctx); + return -EAGAIN; } @@ -1807,8 +1904,10 @@ static int __perf_event_enable(void *info) * If this event can't go on and it's part of a * group, then the whole group has to come off. */ - if (leader != event) + if (leader != event) { group_sched_out(leader, cpuctx, ctx); + perf_cpu_hrtimer_restart(cpuctx); + } if (leader->attr.pinned) { update_group_times(leader); leader->state = PERF_EVENT_STATE_ERROR; @@ -2555,7 +2654,7 @@ static void rotate_ctx(struct perf_event_context *ctx) * because they're strictly cpu affine and rotate_start is called with IRQs * disabled, while rotate_context is called from IRQ context. */ -static void perf_rotate_context(struct perf_cpu_context *cpuctx) +static int perf_rotate_context(struct perf_cpu_context *cpuctx) { struct perf_event_context *ctx = NULL; int rotate = 0, remove = 1; @@ -2594,6 +2693,8 @@ static void perf_rotate_context(struct perf_cpu_context *cpuctx) done: if (remove) list_del_init(&cpuctx->rotation_list); + + return rotate; } #ifdef CONFIG_NO_HZ_FULL @@ -2625,10 +2726,6 @@ void perf_event_task_tick(void) ctx = cpuctx->task_ctx; if (ctx) perf_adjust_freq_unthr_context(ctx, throttled); - - if (cpuctx->jiffies_interval == 1 || - !(jiffies % cpuctx->jiffies_interval)) - perf_rotate_context(cpuctx); } } @@ -6001,7 +6098,9 @@ skip_type: lockdep_set_class(&cpuctx->ctx.lock, &cpuctx_lock); cpuctx->ctx.type = cpu_context; cpuctx->ctx.pmu = pmu; - cpuctx->jiffies_interval = 1; + + __perf_cpu_hrtimer_init(cpuctx, cpu); + INIT_LIST_HEAD(&cpuctx->rotation_list); cpuctx->unique_pmu = pmu; } @@ -7387,7 +7486,6 @@ perf_cpu_notify(struct notifier_block *self, unsigned long action, void *hcpu) case CPU_DOWN_PREPARE: perf_event_exit_cpu(cpu); break; - default: break; } -- cgit v1.2.3 From 62b8563979273424d6ebe9201e34d1acc133ad4f Mon Sep 17 00:00:00 2001 From: Stephane Eranian Date: Wed, 3 Apr 2013 14:21:34 +0200 Subject: perf: Add sysfs entry to adjust multiplexing interval per PMU This patch adds /sys/device/xxx/perf_event_mux_interval_ms to ajust the multiplexing interval per PMU. The unit is milliseconds. Value has to be >= 1. In the 4th version, we renamed the sysfs file to be more consistent with the other /proc/sys/kernel entries for perf_events. In the 5th version, we handle the reprogramming of the hrtimer using hrtimer_forward_now(). That way, we sync up to new timer value quickly (suggested by Jiri Olsa). Signed-off-by: Stephane Eranian Signed-off-by: Peter Zijlstra Cc: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo Link: http://lkml.kernel.org/r/1364991694-5876-3-git-send-email-eranian@google.com Signed-off-by: Ingo Molnar --- kernel/events/core.c | 63 ++++++++++++++++++++++++++++++++++++++++++++++++---- 1 file changed, 59 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 97bfac7e6f45..53d1b300116a 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -723,13 +723,21 @@ static void __perf_cpu_hrtimer_init(struct perf_cpu_context *cpuctx, int cpu) { struct hrtimer *hr = &cpuctx->hrtimer; struct pmu *pmu = cpuctx->ctx.pmu; + int timer; /* no multiplexing needed for SW PMU */ if (pmu->task_ctx_nr == perf_sw_context) return; - cpuctx->hrtimer_interval = - ns_to_ktime(NSEC_PER_MSEC * PERF_CPU_HRTIMER); + /* + * check default is sane, if not set then force to + * default interval (1/tick) + */ + timer = pmu->hrtimer_interval_ms; + if (timer < 1) + timer = pmu->hrtimer_interval_ms = PERF_CPU_HRTIMER; + + cpuctx->hrtimer_interval = ns_to_ktime(NSEC_PER_MSEC * timer); hrtimer_init(hr, CLOCK_MONOTONIC, HRTIMER_MODE_REL_PINNED); hr->function = perf_cpu_hrtimer_handler; @@ -6001,9 +6009,56 @@ type_show(struct device *dev, struct device_attribute *attr, char *page) return snprintf(page, PAGE_SIZE-1, "%d\n", pmu->type); } +static ssize_t +perf_event_mux_interval_ms_show(struct device *dev, + struct device_attribute *attr, + char *page) +{ + struct pmu *pmu = dev_get_drvdata(dev); + + return snprintf(page, PAGE_SIZE-1, "%d\n", pmu->hrtimer_interval_ms); +} + +static ssize_t +perf_event_mux_interval_ms_store(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count) +{ + struct pmu *pmu = dev_get_drvdata(dev); + int timer, cpu, ret; + + ret = kstrtoint(buf, 0, &timer); + if (ret) + return ret; + + if (timer < 1) + return -EINVAL; + + /* same value, noting to do */ + if (timer == pmu->hrtimer_interval_ms) + return count; + + pmu->hrtimer_interval_ms = timer; + + /* update all cpuctx for this PMU */ + for_each_possible_cpu(cpu) { + struct perf_cpu_context *cpuctx; + cpuctx = per_cpu_ptr(pmu->pmu_cpu_context, cpu); + cpuctx->hrtimer_interval = ns_to_ktime(NSEC_PER_MSEC * timer); + + if (hrtimer_active(&cpuctx->hrtimer)) + hrtimer_forward_now(&cpuctx->hrtimer, cpuctx->hrtimer_interval); + } + + return count; +} + +#define __ATTR_RW(attr) __ATTR(attr, 0644, attr##_show, attr##_store) + static struct device_attribute pmu_dev_attrs[] = { - __ATTR_RO(type), - __ATTR_NULL, + __ATTR_RO(type), + __ATTR_RW(perf_event_mux_interval_ms), + __ATTR_NULL, }; static int pmu_bus_running; -- cgit v1.2.3 From 2b923c8f5de6722393e614b096d5040b6d4eaf98 Mon Sep 17 00:00:00 2001 From: Stephane Eranian Date: Tue, 21 May 2013 12:53:37 +0200 Subject: perf/x86: Check branch sampling priv level in generic code This patch moves commit 7cc23cd to the generic code: perf/x86/intel/lbr: Demand proper privileges for PERF_SAMPLE_BRANCH_KERNEL The check is now implemented in generic code instead of x86 specific code. That way we do not have to repeat the test in each arch supporting branch sampling. Signed-off-by: Stephane Eranian Signed-off-by: Peter Zijlstra Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Arnaldo Carvalho de Melo Link: http://lkml.kernel.org/r/20130521105337.GA2879@quad Signed-off-by: Ingo Molnar --- kernel/events/core.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 53d1b300116a..a0780b3a3d50 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -6481,11 +6481,6 @@ static int perf_copy_attr(struct perf_event_attr __user *uattr, if (!(mask & ~PERF_SAMPLE_BRANCH_PLM_ALL)) return -EINVAL; - /* kernel level capture: check permissions */ - if ((mask & PERF_SAMPLE_BRANCH_PERM_PLM) - && perf_paranoid_kernel() && !capable(CAP_SYS_ADMIN)) - return -EACCES; - /* propagate priv level, when not set for branch */ if (!(mask & PERF_SAMPLE_BRANCH_PLM_ALL)) { @@ -6503,6 +6498,10 @@ static int perf_copy_attr(struct perf_event_attr __user *uattr, */ attr->branch_sample_type = mask; } + /* kernel level capture: check permissions */ + if ((mask & PERF_SAMPLE_BRANCH_KERNEL) + && perf_paranoid_kernel() && !capable(CAP_SYS_ADMIN)) + return -EACCES; } if (attr->sample_type & PERF_SAMPLE_REGS_USER) { -- cgit v1.2.3 From 0c1061733aa0303e6536c0bc7f86d68f5eb55446 Mon Sep 17 00:00:00 2001 From: Juri Lelli Date: Wed, 15 May 2013 11:04:10 +0200 Subject: rtmutex: Document rt_mutex_adjust_prio_chain() Parameters and usage of rt_mutex_adjust_prio_chain() are already documented in Documentation/rt-mutex-design.txt. However, since this function is called from several paths with different semantics (related to the arguments), it is handy to have a quick reference directly in the code. Signed-off-by: Juri Lelli Cc: Clark Williams Cc: Steven Rostedt Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1368608650-7935-1-git-send-email-juri.lelli@gmail.com Signed-off-by: Ingo Molnar --- kernel/rtmutex.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'kernel') diff --git a/kernel/rtmutex.c b/kernel/rtmutex.c index 1e09308bf2a1..0dd6aec1cb6a 100644 --- a/kernel/rtmutex.c +++ b/kernel/rtmutex.c @@ -145,6 +145,19 @@ int max_lock_depth = 1024; /* * Adjust the priority chain. Also used for deadlock detection. * Decreases task's usage by one - may thus free the task. + * + * @task: the task owning the mutex (owner) for which a chain walk is probably + * needed + * @deadlock_detect: do we have to carry out deadlock detection? + * @orig_lock: the mutex (can be NULL if we are walking the chain to recheck + * things for a task that has just got its priority adjusted, and + * is waiting on a mutex) + * @orig_waiter: rt_mutex_waiter struct for the task that has just donated + * its priority to the mutex owner (can be NULL in the case + * depicted above or if the top waiter is gone away and we are + * actually deboosting the owner) + * @top_task: the current top waiter + * * Returns 0 or -EDEADLK. */ static int rt_mutex_adjust_prio_chain(struct task_struct *task, -- cgit v1.2.3 From c7e99fc75de8882bc4104455ace366d9d3599a96 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 May 2013 09:28:02 +0200 Subject: clockevents: Define CS_NAME_LEN unconditionally Unbreak architectures which do not use clockevents, but require to build some of the core timekeeping infrastructure Reported-by: Ingo Molnar Signed-off-by: Thomas Gleixner --- kernel/time/tick-internal.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-internal.h b/kernel/time/tick-internal.h index be1690eaecff..bc906cad709b 100644 --- a/kernel/time/tick-internal.h +++ b/kernel/time/tick-internal.h @@ -6,13 +6,13 @@ extern seqlock_t jiffies_lock; +#define CS_NAME_LEN 32 + #ifdef CONFIG_GENERIC_CLOCKEVENTS_BUILD #define TICK_DO_TIMER_NONE -1 #define TICK_DO_TIMER_BOOT -2 -#define CS_NAME_LEN 32 - DECLARE_PER_CPU(struct tick_device, tick_cpu_device); extern ktime_t tick_next_period; extern ktime_t tick_period; -- cgit v1.2.3 From 41261b6a832ea0e788627f6a8707854423f9ff49 Mon Sep 17 00:00:00 2001 From: Gerald Schaefer Date: Fri, 24 May 2013 18:07:49 +0200 Subject: sched/autogroup: Fix race with task_groups list In autogroup_create(), a tg is allocated and added to the task_groups list. If CONFIG_RT_GROUP_SCHED is set, this tg is then modified while on the list, without locking. This can race with someone walking the list, like __enable_runtime() during CPU unplug, and result in a use-after-free bug. To fix this, move sched_online_group(), which adds the tg to the list, to the end of the autogroup_create() function after the modification. Signed-off-by: Gerald Schaefer Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1369411669-46971-2-git-send-email-gerald.schaefer@de.ibm.com Signed-off-by: Ingo Molnar --- kernel/sched/auto_group.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/auto_group.c b/kernel/sched/auto_group.c index 64de5f8b0c9e..4a073539c58e 100644 --- a/kernel/sched/auto_group.c +++ b/kernel/sched/auto_group.c @@ -77,8 +77,6 @@ static inline struct autogroup *autogroup_create(void) if (IS_ERR(tg)) goto out_free; - sched_online_group(tg, &root_task_group); - kref_init(&ag->kref); init_rwsem(&ag->lock); ag->id = atomic_inc_return(&autogroup_seq_nr); @@ -98,6 +96,7 @@ static inline struct autogroup *autogroup_create(void) #endif tg->autogroup = ag; + sched_online_group(tg, &root_task_group); return ag; out_free: -- cgit v1.2.3 From c5405a495e88d93cf9b4f4cc91507c7f4afcb901 Mon Sep 17 00:00:00 2001 From: Neil Zhang Date: Thu, 11 Apr 2013 21:04:59 +0800 Subject: sched: Remove redundant update_runtime notifier migration_call() will do all the things that update_runtime() does. So let's remove it. Furthermore, there is potential risk that the current code will catch BUG_ON at line 689 of rt.c when do cpu hotplug while there are realtime threads running because of enabling runtime twice while the rt_runtime may already changed. Signed-off-by: Neil Zhang Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1365685499-26515-1-git-send-email-zhangwm@marvell.com Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 3 --- kernel/sched/rt.c | 40 ---------------------------------------- kernel/sched/sched.h | 1 - 3 files changed, 44 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index bfa7e77e0b50..79e48e6a9385 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -6285,9 +6285,6 @@ void __init sched_init_smp(void) hotcpu_notifier(cpuset_cpu_active, CPU_PRI_CPUSET_ACTIVE); hotcpu_notifier(cpuset_cpu_inactive, CPU_PRI_CPUSET_INACTIVE); - /* RT runtime code needs to handle some hotplug events */ - hotcpu_notifier(update_runtime, 0); - init_hrtick(); /* Move init over to a non-isolated CPU */ diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c index 7aced2e3b085..8853ab17b750 100644 --- a/kernel/sched/rt.c +++ b/kernel/sched/rt.c @@ -699,15 +699,6 @@ balanced: } } -static void disable_runtime(struct rq *rq) -{ - unsigned long flags; - - raw_spin_lock_irqsave(&rq->lock, flags); - __disable_runtime(rq); - raw_spin_unlock_irqrestore(&rq->lock, flags); -} - static void __enable_runtime(struct rq *rq) { rt_rq_iter_t iter; @@ -732,37 +723,6 @@ static void __enable_runtime(struct rq *rq) } } -static void enable_runtime(struct rq *rq) -{ - unsigned long flags; - - raw_spin_lock_irqsave(&rq->lock, flags); - __enable_runtime(rq); - raw_spin_unlock_irqrestore(&rq->lock, flags); -} - -int update_runtime(struct notifier_block *nfb, unsigned long action, void *hcpu) -{ - int cpu = (int)(long)hcpu; - - switch (action) { - case CPU_DOWN_PREPARE: - case CPU_DOWN_PREPARE_FROZEN: - disable_runtime(cpu_rq(cpu)); - return NOTIFY_OK; - - case CPU_DOWN_FAILED: - case CPU_DOWN_FAILED_FROZEN: - case CPU_ONLINE: - case CPU_ONLINE_FROZEN: - enable_runtime(cpu_rq(cpu)); - return NOTIFY_OK; - - default: - return NOTIFY_DONE; - } -} - static int balance_runtime(struct rt_rq *rt_rq) { int more = 0; diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index f1f6256c1224..c806c61a1261 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -1041,7 +1041,6 @@ static inline void idle_balance(int cpu, struct rq *rq) extern void sysrq_sched_debug_show(void); extern void sched_init_granularity(void); extern void update_max_interval(void); -extern int update_runtime(struct notifier_block *nfb, unsigned long action, void *hcpu); extern void init_sched_rt_class(void); extern void init_sched_fair_class(void); -- cgit v1.2.3 From 77bd39702f0b3840cea17681409270b16a3b93c0 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 12 Apr 2013 01:50:58 +0200 Subject: sched: Update rq clock before migrating tasks out of dying CPU Because the sched_class::put_prev_task() callback of rt and fair classes are referring to the rq clock to update their runtime statistics. There is a missing rq clock update from the CPU hotplug notifier's entry point of the scheduler. Signed-off-by: Frederic Weisbecker Cc: Li Zhong Cc: Steven Rostedt Cc: Paul Turner Cc: Mike Galbraith Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1365724262-20142-2-git-send-email-fweisbec@gmail.com Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 79e48e6a9385..7bf0418dc60f 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -4378,6 +4378,13 @@ static void migrate_tasks(unsigned int dead_cpu) */ rq->stop = NULL; + /* + * put_prev_task() and pick_next_task() sched + * class method both need to have an up-to-date + * value of rq->clock[_task] + */ + update_rq_clock(rq); + for ( ; ; ) { /* * There's this thread running, bail when that's the only -- cgit v1.2.3 From 71b1da46ff70309a2ec12ce943aa0d192d2c8f0c Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 12 Apr 2013 01:50:59 +0200 Subject: sched: Update rq clock before setting fair group shares Because we may update the execution time in sched_group_set_shares()->update_cfs_shares()->reweight_entity()->update_curr() before reweighting the entity while setting the group shares and this requires an uptodate version of the runqueue clock. Signed-off-by: Frederic Weisbecker Cc: Li Zhong Cc: Steven Rostedt Cc: Paul Turner Cc: Mike Galbraith Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1365724262-20142-3-git-send-email-fweisbec@gmail.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index f62b16dfba63..f76ca21711bb 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -6107,6 +6107,9 @@ int sched_group_set_shares(struct task_group *tg, unsigned long shares) se = tg->se[i]; /* Propagate contribution to hierarchy */ raw_spin_lock_irqsave(&rq->lock, flags); + + /* Possible calls to update_curr() need rq clock */ + update_rq_clock(rq); for_each_sched_entity(se) update_cfs_shares(group_cfs_rq(se)); raw_spin_unlock_irqrestore(&rq->lock, flags); -- cgit v1.2.3 From 1ad4ec0dc740c4183acd6d6e367ca52b28e4fa94 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 12 Apr 2013 01:51:00 +0200 Subject: sched: Update rq clock before calling check_preempt_curr() check_preempt_curr() of fair class needs an uptodate sched clock value to update runtime stats of the current task of the target's rq. When a task is woken up, activate_task() is usually called right before ttwu_do_wakeup() unless the task is still in the runqueue. In the latter case we need to update the rq clock explicitly because activate_task() isn't here to do the job for us. Signed-off-by: Frederic Weisbecker Cc: Li Zhong Cc: Steven Rostedt Cc: Paul Turner Cc: Mike Galbraith Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1365724262-20142-4-git-send-email-fweisbec@gmail.com Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 7bf0418dc60f..46d00172ae4a 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -1365,6 +1365,8 @@ static int ttwu_remote(struct task_struct *p, int wake_flags) rq = __task_rq_lock(p); if (p->on_rq) { + /* check_preempt_curr() may use rq clock */ + update_rq_clock(rq); ttwu_do_wakeup(rq, p, wake_flags); ret = 1; } -- cgit v1.2.3 From 1a55af2e45cce0ff13bc33c8ee99da84e188b615 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 12 Apr 2013 01:51:01 +0200 Subject: sched: Update rq clock earlier in unthrottle_cfs_rq In this function we are making use of rq->clock right before the update of the rq clock, let's just call update_rq_clock() just before that to avoid using a stale rq clock value. Signed-off-by: Frederic Weisbecker Cc: Li Zhong Cc: Steven Rostedt Cc: Paul Turner Cc: Mike Galbraith Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1365724262-20142-5-git-send-email-fweisbec@gmail.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index f76ca21711bb..1c8762a5370c 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -2319,12 +2319,14 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq) se = cfs_rq->tg->se[cpu_of(rq_of(cfs_rq))]; cfs_rq->throttled = 0; + + update_rq_clock(rq); + raw_spin_lock(&cfs_b->lock); cfs_b->throttled_time += rq->clock - cfs_rq->throttled_clock; list_del_rcu(&cfs_rq->throttled_list); raw_spin_unlock(&cfs_b->lock); - update_rq_clock(rq); /* update hierarchical throttle state */ walk_tg_tree_from(cfs_rq->tg, tg_nop, tg_unthrottle_up, (void *)rq); -- cgit v1.2.3 From 78becc27097585c6aec7043834cadde950ae79f2 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 12 Apr 2013 01:51:02 +0200 Subject: sched: Use an accessor to read the rq clock Read the runqueue clock through an accessor. This prepares for adding a debugging infrastructure to detect missing or redundant calls to update_rq_clock() between a scheduler's entry and exit point. Signed-off-by: Frederic Weisbecker Cc: Li Zhong Cc: Steven Rostedt Cc: Paul Turner Cc: Mike Galbraith Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1365724262-20142-6-git-send-email-fweisbec@gmail.com Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 6 +++--- kernel/sched/fair.c | 44 ++++++++++++++++++++++---------------------- kernel/sched/rt.c | 8 ++++---- kernel/sched/sched.h | 10 ++++++++++ kernel/sched/stats.h | 8 ++++---- kernel/sched/stop_task.c | 8 ++++---- 6 files changed, 47 insertions(+), 37 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 46d00172ae4a..36f85be2932b 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -667,7 +667,7 @@ void sched_avg_update(struct rq *rq) { s64 period = sched_avg_period(); - while ((s64)(rq->clock - rq->age_stamp) > period) { + while ((s64)(rq_clock(rq) - rq->age_stamp) > period) { /* * Inline assembly required to prevent the compiler * optimising this loop into a divmod call. @@ -1328,7 +1328,7 @@ ttwu_do_wakeup(struct rq *rq, struct task_struct *p, int wake_flags) p->sched_class->task_woken(rq, p); if (rq->idle_stamp) { - u64 delta = rq->clock - rq->idle_stamp; + u64 delta = rq_clock(rq) - rq->idle_stamp; u64 max = 2*sysctl_sched_migration_cost; if (delta > max) @@ -2106,7 +2106,7 @@ static u64 do_task_delta_exec(struct task_struct *p, struct rq *rq) if (task_current(rq, p)) { update_rq_clock(rq); - ns = rq->clock_task - p->se.exec_start; + ns = rq_clock_task(rq) - p->se.exec_start; if ((s64)ns < 0) ns = 0; } diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 1c8762a5370c..3ee1c2e4ae60 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -704,7 +704,7 @@ __update_curr(struct cfs_rq *cfs_rq, struct sched_entity *curr, static void update_curr(struct cfs_rq *cfs_rq) { struct sched_entity *curr = cfs_rq->curr; - u64 now = rq_of(cfs_rq)->clock_task; + u64 now = rq_clock_task(rq_of(cfs_rq)); unsigned long delta_exec; if (unlikely(!curr)) @@ -736,7 +736,7 @@ static void update_curr(struct cfs_rq *cfs_rq) static inline void update_stats_wait_start(struct cfs_rq *cfs_rq, struct sched_entity *se) { - schedstat_set(se->statistics.wait_start, rq_of(cfs_rq)->clock); + schedstat_set(se->statistics.wait_start, rq_clock(rq_of(cfs_rq))); } /* @@ -756,14 +756,14 @@ static void update_stats_wait_end(struct cfs_rq *cfs_rq, struct sched_entity *se) { schedstat_set(se->statistics.wait_max, max(se->statistics.wait_max, - rq_of(cfs_rq)->clock - se->statistics.wait_start)); + rq_clock(rq_of(cfs_rq)) - se->statistics.wait_start)); schedstat_set(se->statistics.wait_count, se->statistics.wait_count + 1); schedstat_set(se->statistics.wait_sum, se->statistics.wait_sum + - rq_of(cfs_rq)->clock - se->statistics.wait_start); + rq_clock(rq_of(cfs_rq)) - se->statistics.wait_start); #ifdef CONFIG_SCHEDSTATS if (entity_is_task(se)) { trace_sched_stat_wait(task_of(se), - rq_of(cfs_rq)->clock - se->statistics.wait_start); + rq_clock(rq_of(cfs_rq)) - se->statistics.wait_start); } #endif schedstat_set(se->statistics.wait_start, 0); @@ -789,7 +789,7 @@ update_stats_curr_start(struct cfs_rq *cfs_rq, struct sched_entity *se) /* * We are starting a new run period: */ - se->exec_start = rq_of(cfs_rq)->clock_task; + se->exec_start = rq_clock_task(rq_of(cfs_rq)); } /************************************************** @@ -1515,7 +1515,7 @@ static void update_cfs_rq_blocked_load(struct cfs_rq *cfs_rq, int force_update) static inline void update_rq_runnable_avg(struct rq *rq, int runnable) { - __update_entity_runnable_avg(rq->clock_task, &rq->avg, runnable); + __update_entity_runnable_avg(rq_clock_task(rq), &rq->avg, runnable); __update_tg_runnable_avg(&rq->avg, &rq->cfs); } @@ -1530,7 +1530,7 @@ static inline void enqueue_entity_load_avg(struct cfs_rq *cfs_rq, * accumulated while sleeping. */ if (unlikely(se->avg.decay_count <= 0)) { - se->avg.last_runnable_update = rq_of(cfs_rq)->clock_task; + se->avg.last_runnable_update = rq_clock_task(rq_of(cfs_rq)); if (se->avg.decay_count) { /* * In a wake-up migration we have to approximate the @@ -1625,7 +1625,7 @@ static void enqueue_sleeper(struct cfs_rq *cfs_rq, struct sched_entity *se) tsk = task_of(se); if (se->statistics.sleep_start) { - u64 delta = rq_of(cfs_rq)->clock - se->statistics.sleep_start; + u64 delta = rq_clock(rq_of(cfs_rq)) - se->statistics.sleep_start; if ((s64)delta < 0) delta = 0; @@ -1642,7 +1642,7 @@ static void enqueue_sleeper(struct cfs_rq *cfs_rq, struct sched_entity *se) } } if (se->statistics.block_start) { - u64 delta = rq_of(cfs_rq)->clock - se->statistics.block_start; + u64 delta = rq_clock(rq_of(cfs_rq)) - se->statistics.block_start; if ((s64)delta < 0) delta = 0; @@ -1823,9 +1823,9 @@ dequeue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags) struct task_struct *tsk = task_of(se); if (tsk->state & TASK_INTERRUPTIBLE) - se->statistics.sleep_start = rq_of(cfs_rq)->clock; + se->statistics.sleep_start = rq_clock(rq_of(cfs_rq)); if (tsk->state & TASK_UNINTERRUPTIBLE) - se->statistics.block_start = rq_of(cfs_rq)->clock; + se->statistics.block_start = rq_clock(rq_of(cfs_rq)); } #endif } @@ -2100,7 +2100,7 @@ static inline u64 cfs_rq_clock_task(struct cfs_rq *cfs_rq) if (unlikely(cfs_rq->throttle_count)) return cfs_rq->throttled_clock_task; - return rq_of(cfs_rq)->clock_task - cfs_rq->throttled_clock_task_time; + return rq_clock_task(rq_of(cfs_rq)) - cfs_rq->throttled_clock_task_time; } /* returns 0 on failure to allocate runtime */ @@ -2159,7 +2159,7 @@ static void expire_cfs_rq_runtime(struct cfs_rq *cfs_rq) struct rq *rq = rq_of(cfs_rq); /* if the deadline is ahead of our clock, nothing to do */ - if (likely((s64)(rq->clock - cfs_rq->runtime_expires) < 0)) + if (likely((s64)(rq_clock(rq_of(cfs_rq)) - cfs_rq->runtime_expires) < 0)) return; if (cfs_rq->runtime_remaining < 0) @@ -2248,7 +2248,7 @@ static int tg_unthrottle_up(struct task_group *tg, void *data) #ifdef CONFIG_SMP if (!cfs_rq->throttle_count) { /* adjust cfs_rq_clock_task() */ - cfs_rq->throttled_clock_task_time += rq->clock_task - + cfs_rq->throttled_clock_task_time += rq_clock_task(rq) - cfs_rq->throttled_clock_task; } #endif @@ -2263,7 +2263,7 @@ static int tg_throttle_down(struct task_group *tg, void *data) /* group is entering throttled state, stop time */ if (!cfs_rq->throttle_count) - cfs_rq->throttled_clock_task = rq->clock_task; + cfs_rq->throttled_clock_task = rq_clock_task(rq); cfs_rq->throttle_count++; return 0; @@ -2302,7 +2302,7 @@ static void throttle_cfs_rq(struct cfs_rq *cfs_rq) rq->nr_running -= task_delta; cfs_rq->throttled = 1; - cfs_rq->throttled_clock = rq->clock; + cfs_rq->throttled_clock = rq_clock(rq); raw_spin_lock(&cfs_b->lock); list_add_tail_rcu(&cfs_rq->throttled_list, &cfs_b->throttled_cfs_rq); raw_spin_unlock(&cfs_b->lock); @@ -2323,7 +2323,7 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq) update_rq_clock(rq); raw_spin_lock(&cfs_b->lock); - cfs_b->throttled_time += rq->clock - cfs_rq->throttled_clock; + cfs_b->throttled_time += rq_clock(rq) - cfs_rq->throttled_clock; list_del_rcu(&cfs_rq->throttled_list); raw_spin_unlock(&cfs_b->lock); @@ -2726,7 +2726,7 @@ static void __maybe_unused unthrottle_offline_cfs_rqs(struct rq *rq) #else /* CONFIG_CFS_BANDWIDTH */ static inline u64 cfs_rq_clock_task(struct cfs_rq *cfs_rq) { - return rq_of(cfs_rq)->clock_task; + return rq_clock_task(rq_of(cfs_rq)); } static void account_cfs_rq_runtime(struct cfs_rq *cfs_rq, @@ -3966,7 +3966,7 @@ int can_migrate_task(struct task_struct *p, struct lb_env *env) * 2) too many balance attempts have failed. */ - tsk_cache_hot = task_hot(p, env->src_rq->clock_task, env->sd); + tsk_cache_hot = task_hot(p, rq_clock_task(env->src_rq), env->sd); if (!tsk_cache_hot || env->sd->nr_balance_failed > env->sd->cache_nice_tries) { @@ -4322,7 +4322,7 @@ static unsigned long scale_rt_power(int cpu) age_stamp = ACCESS_ONCE(rq->age_stamp); avg = ACCESS_ONCE(rq->rt_avg); - total = sched_avg_period() + (rq->clock - age_stamp); + total = sched_avg_period() + (rq_clock(rq) - age_stamp); if (unlikely(total < avg)) { /* Ensures that power won't end up being negative */ @@ -5261,7 +5261,7 @@ void idle_balance(int this_cpu, struct rq *this_rq) int pulled_task = 0; unsigned long next_balance = jiffies + HZ; - this_rq->idle_stamp = this_rq->clock; + this_rq->idle_stamp = rq_clock(this_rq); if (this_rq->avg_idle < sysctl_sched_migration_cost) return; diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c index 8853ab17b750..8d85f9ac4262 100644 --- a/kernel/sched/rt.c +++ b/kernel/sched/rt.c @@ -886,7 +886,7 @@ static void update_curr_rt(struct rq *rq) if (curr->sched_class != &rt_sched_class) return; - delta_exec = rq->clock_task - curr->se.exec_start; + delta_exec = rq_clock_task(rq) - curr->se.exec_start; if (unlikely((s64)delta_exec <= 0)) return; @@ -896,7 +896,7 @@ static void update_curr_rt(struct rq *rq) curr->se.sum_exec_runtime += delta_exec; account_group_exec_runtime(curr, delta_exec); - curr->se.exec_start = rq->clock_task; + curr->se.exec_start = rq_clock_task(rq); cpuacct_charge(curr, delta_exec); sched_rt_avg_update(rq, delta_exec); @@ -1345,7 +1345,7 @@ static struct task_struct *_pick_next_task_rt(struct rq *rq) } while (rt_rq); p = rt_task_of(rt_se); - p->se.exec_start = rq->clock_task; + p->se.exec_start = rq_clock_task(rq); return p; } @@ -1997,7 +1997,7 @@ static void set_curr_task_rt(struct rq *rq) { struct task_struct *p = rq->curr; - p->se.exec_start = rq->clock_task; + p->se.exec_start = rq_clock_task(rq); /* The running task is never eligible for pushing */ dequeue_pushable_task(rq, p); diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index c806c61a1261..74ff659e964f 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -548,6 +548,16 @@ DECLARE_PER_CPU(struct rq, runqueues); #define cpu_curr(cpu) (cpu_rq(cpu)->curr) #define raw_rq() (&__raw_get_cpu_var(runqueues)) +static inline u64 rq_clock(struct rq *rq) +{ + return rq->clock; +} + +static inline u64 rq_clock_task(struct rq *rq) +{ + return rq->clock_task; +} + #ifdef CONFIG_SMP #define rcu_dereference_check_sched_domain(p) \ diff --git a/kernel/sched/stats.h b/kernel/sched/stats.h index 2ef90a51ec5e..17d7065c3872 100644 --- a/kernel/sched/stats.h +++ b/kernel/sched/stats.h @@ -61,7 +61,7 @@ static inline void sched_info_reset_dequeued(struct task_struct *t) */ static inline void sched_info_dequeued(struct task_struct *t) { - unsigned long long now = task_rq(t)->clock, delta = 0; + unsigned long long now = rq_clock(task_rq(t)), delta = 0; if (unlikely(sched_info_on())) if (t->sched_info.last_queued) @@ -79,7 +79,7 @@ static inline void sched_info_dequeued(struct task_struct *t) */ static void sched_info_arrive(struct task_struct *t) { - unsigned long long now = task_rq(t)->clock, delta = 0; + unsigned long long now = rq_clock(task_rq(t)), delta = 0; if (t->sched_info.last_queued) delta = now - t->sched_info.last_queued; @@ -100,7 +100,7 @@ static inline void sched_info_queued(struct task_struct *t) { if (unlikely(sched_info_on())) if (!t->sched_info.last_queued) - t->sched_info.last_queued = task_rq(t)->clock; + t->sched_info.last_queued = rq_clock(task_rq(t)); } /* @@ -112,7 +112,7 @@ static inline void sched_info_queued(struct task_struct *t) */ static inline void sched_info_depart(struct task_struct *t) { - unsigned long long delta = task_rq(t)->clock - + unsigned long long delta = rq_clock(task_rq(t)) - t->sched_info.last_arrival; rq_sched_info_depart(task_rq(t), delta); diff --git a/kernel/sched/stop_task.c b/kernel/sched/stop_task.c index da5eb5bed84a..e08fbeeb54b9 100644 --- a/kernel/sched/stop_task.c +++ b/kernel/sched/stop_task.c @@ -28,7 +28,7 @@ static struct task_struct *pick_next_task_stop(struct rq *rq) struct task_struct *stop = rq->stop; if (stop && stop->on_rq) { - stop->se.exec_start = rq->clock_task; + stop->se.exec_start = rq_clock_task(rq); return stop; } @@ -57,7 +57,7 @@ static void put_prev_task_stop(struct rq *rq, struct task_struct *prev) struct task_struct *curr = rq->curr; u64 delta_exec; - delta_exec = rq->clock_task - curr->se.exec_start; + delta_exec = rq_clock_task(rq) - curr->se.exec_start; if (unlikely((s64)delta_exec < 0)) delta_exec = 0; @@ -67,7 +67,7 @@ static void put_prev_task_stop(struct rq *rq, struct task_struct *prev) curr->se.sum_exec_runtime += delta_exec; account_group_exec_runtime(curr, delta_exec); - curr->se.exec_start = rq->clock_task; + curr->se.exec_start = rq_clock_task(rq); cpuacct_charge(curr, delta_exec); } @@ -79,7 +79,7 @@ static void set_curr_task_stop(struct rq *rq) { struct task_struct *stop = rq->stop; - stop->se.exec_start = rq->clock_task; + stop->se.exec_start = rq_clock_task(rq); } static void switched_to_stop(struct rq *rq, struct task_struct *p) -- cgit v1.2.3 From 1eaff67266b6b6c97bbd33cf2c20577822836413 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 May 2013 09:48:46 +0200 Subject: clocksource: Implement clocksource_select_fallback() for CONFIG_ARCH_USES_GETTIMEOFFSET=y commit 7eaeb34305 (clocksource: Provide unbind interface in sysfs) implemented clocksource_select_fallback() which is not defined for CONFIG_ARCH_USES_GETTIMEOFFSET=y. Add an empty inline function for that. Reported-by: Ingo Molnar Reported-by: fengguang.wu@intel.com Signed-off-by: Thomas Gleixner --- kernel/time/clocksource.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c index 6d05b00410cc..e713ef7d19a7 100644 --- a/kernel/time/clocksource.c +++ b/kernel/time/clocksource.c @@ -632,6 +632,7 @@ static void clocksource_select_fallback(void) #else /* !CONFIG_ARCH_USES_GETTIMEOFFSET */ static inline void clocksource_select(void) { } +static inline void clocksource_select_fallback(void) { } #endif -- cgit v1.2.3 From 7b959fc582741227a1c4cba710d6aff8fb183128 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Wed, 22 May 2013 18:34:09 +0900 Subject: kprobes: Fix to free gone and unused optprobes Fix to free gone and unused optprobes. This bug will cause a kernel panic if the user reuses the killed and unused probe. Reported at: http://sourceware.org/ml/systemtap/2013-q2/msg00142.html In the normal path, an optprobe on an init function is unregistered when a module goes live. unregister_kprobe(kp) -> __unregister_kprobe_top ->__disable_kprobe ->disarm_kprobe(ap == op) ->__disarm_kprobe ->unoptimize_kprobe : the op is queued on unoptimizing_list and do nothing in __unregister_kprobe_bottom After a while (usually wait 5 jiffies), kprobe_optimizer runs to unoptimize and free optprobe. kprobe_optimizer ->do_unoptimize_kprobes ->arch_unoptimize_kprobes : moved to free_list ->do_free_cleaned_kprobes ->hlist_del: the op is removed ->free_aggr_kprobe ->arch_remove_optimized_kprobe ->arch_remove_kprobe ->kfree: the op is freed Here, if kprobes_module_callback is called and the delayed unoptimizing probe is picked BEFORE kprobe_optimizer runs, kprobes_module_callback ->kill_kprobe ->kill_optimized_kprobe : dequeued from unoptimizing_list <=!!! ->arch_remove_optimized_kprobe ->arch_remove_kprobe (but op is not freed, and on the kprobe hash table) This doesn't happen if the probe unregistration is done AFTER kprobes_module_callback is called (because at that time the op is gone), and kprobe-tracer does it. To fix this bug, this patch changes kprobes_module_callback to enqueue the op to freeing_list at kill_optimized_kprobe only if the op is unused. The unused probes on freeing_list will be freed in do_free_cleaned_kprobes. Note that this calls arch_remove_*kprobe twice on the same probe. Thus those functions have to check the double free. Fortunately, most of arch codes already checked that except for mips. This will be fixed in the next patch. Signed-off-by: Masami Hiramatsu Cc: Timo Juhani Lindfors Cc: Ananth N Mavinakayanahalli Cc: Anil S Keshavamurthy Cc: Frank Ch. Eigler Cc: systemtap@sourceware.org Cc: yrl.pp-manager.tt@hitachi.com Cc: David S. Miller Cc: "David S. Miller" Link: http://lkml.kernel.org/r/20130522093409.9084.63554.stgit@mhiramat-M0-7522 [ Minor edits. ] Signed-off-by: Ingo Molnar --- kernel/kprobes.c | 30 ++++++++++++++++++++---------- 1 file changed, 20 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/kprobes.c b/kernel/kprobes.c index 3fed7f0cbcdf..bddf3b201a48 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -467,6 +467,7 @@ static struct kprobe *__kprobes get_optimized_kprobe(unsigned long addr) /* Optimization staging list, protected by kprobe_mutex */ static LIST_HEAD(optimizing_list); static LIST_HEAD(unoptimizing_list); +static LIST_HEAD(freeing_list); static void kprobe_optimizer(struct work_struct *work); static DECLARE_DELAYED_WORK(optimizing_work, kprobe_optimizer); @@ -504,7 +505,7 @@ static __kprobes void do_optimize_kprobes(void) * Unoptimize (replace a jump with a breakpoint and remove the breakpoint * if need) kprobes listed on unoptimizing_list. */ -static __kprobes void do_unoptimize_kprobes(struct list_head *free_list) +static __kprobes void do_unoptimize_kprobes(void) { struct optimized_kprobe *op, *tmp; @@ -515,9 +516,9 @@ static __kprobes void do_unoptimize_kprobes(struct list_head *free_list) /* Ditto to do_optimize_kprobes */ get_online_cpus(); mutex_lock(&text_mutex); - arch_unoptimize_kprobes(&unoptimizing_list, free_list); + arch_unoptimize_kprobes(&unoptimizing_list, &freeing_list); /* Loop free_list for disarming */ - list_for_each_entry_safe(op, tmp, free_list, list) { + list_for_each_entry_safe(op, tmp, &freeing_list, list) { /* Disarm probes if marked disabled */ if (kprobe_disabled(&op->kp)) arch_disarm_kprobe(&op->kp); @@ -536,11 +537,11 @@ static __kprobes void do_unoptimize_kprobes(struct list_head *free_list) } /* Reclaim all kprobes on the free_list */ -static __kprobes void do_free_cleaned_kprobes(struct list_head *free_list) +static __kprobes void do_free_cleaned_kprobes(void) { struct optimized_kprobe *op, *tmp; - list_for_each_entry_safe(op, tmp, free_list, list) { + list_for_each_entry_safe(op, tmp, &freeing_list, list) { BUG_ON(!kprobe_unused(&op->kp)); list_del_init(&op->list); free_aggr_kprobe(&op->kp); @@ -556,8 +557,6 @@ static __kprobes void kick_kprobe_optimizer(void) /* Kprobe jump optimizer */ static __kprobes void kprobe_optimizer(struct work_struct *work) { - LIST_HEAD(free_list); - mutex_lock(&kprobe_mutex); /* Lock modules while optimizing kprobes */ mutex_lock(&module_mutex); @@ -566,7 +565,7 @@ static __kprobes void kprobe_optimizer(struct work_struct *work) * Step 1: Unoptimize kprobes and collect cleaned (unused and disarmed) * kprobes before waiting for quiesence period. */ - do_unoptimize_kprobes(&free_list); + do_unoptimize_kprobes(); /* * Step 2: Wait for quiesence period to ensure all running interrupts @@ -581,7 +580,7 @@ static __kprobes void kprobe_optimizer(struct work_struct *work) do_optimize_kprobes(); /* Step 4: Free cleaned kprobes after quiesence period */ - do_free_cleaned_kprobes(&free_list); + do_free_cleaned_kprobes(); mutex_unlock(&module_mutex); mutex_unlock(&kprobe_mutex); @@ -723,8 +722,19 @@ static void __kprobes kill_optimized_kprobe(struct kprobe *p) if (!list_empty(&op->list)) /* Dequeue from the (un)optimization queue */ list_del_init(&op->list); - op->kp.flags &= ~KPROBE_FLAG_OPTIMIZED; + + if (kprobe_unused(p)) { + /* Enqueue if it is unused */ + list_add(&op->list, &freeing_list); + /* + * Remove unused probes from the hash list. After waiting + * for synchronization, this probe is reclaimed. + * (reclaiming is done by do_free_cleaned_kprobes().) + */ + hlist_del_rcu(&op->kp.hlist); + } + /* Don't touch the code, because it is already freed. */ arch_remove_optimized_kprobe(op); } -- cgit v1.2.3 From 26cb63ad11e04047a64309362674bcbbd6a6f246 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 28 May 2013 10:55:48 +0200 Subject: perf: Fix perf mmap bugs Vince reported a problem found by his perf specific trinity fuzzer. Al noticed 2 problems with perf's mmap(): - it has issues against fork() since we use vma->vm_mm for accounting. - it has an rb refcount leak on double mmap(). We fix the issues against fork() by using VM_DONTCOPY; I don't think there's code out there that uses this; we didn't hear about weird accounting problems/crashes. If we do need this to work, the previously proposed VM_PINNED could make this work. Aside from the rb reference leak spotted by Al, Vince's example prog was indeed doing a double mmap() through the use of perf_event_set_output(). This exposes another problem, since we now have 2 events with one buffer, the accounting gets screwy because we account per event. Fix this by making the buffer responsible for its own accounting. Reported-by: Vince Weaver Signed-off-by: Peter Zijlstra Cc: Al Viro Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Link: http://lkml.kernel.org/r/20130528085548.GA12193@twins.programming.kicks-ass.net Signed-off-by: Ingo Molnar --- kernel/events/core.c | 37 ++++++++++++++++++++----------------- kernel/events/internal.h | 3 +++ 2 files changed, 23 insertions(+), 17 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 9dc297faf7c0..ae752cd4a086 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -2917,7 +2917,7 @@ static void free_event_rcu(struct rcu_head *head) kfree(event); } -static void ring_buffer_put(struct ring_buffer *rb); +static bool ring_buffer_put(struct ring_buffer *rb); static void free_event(struct perf_event *event) { @@ -3582,13 +3582,13 @@ static struct ring_buffer *ring_buffer_get(struct perf_event *event) return rb; } -static void ring_buffer_put(struct ring_buffer *rb) +static bool ring_buffer_put(struct ring_buffer *rb) { struct perf_event *event, *n; unsigned long flags; if (!atomic_dec_and_test(&rb->refcount)) - return; + return false; spin_lock_irqsave(&rb->event_lock, flags); list_for_each_entry_safe(event, n, &rb->event_list, rb_entry) { @@ -3598,6 +3598,7 @@ static void ring_buffer_put(struct ring_buffer *rb) spin_unlock_irqrestore(&rb->event_lock, flags); call_rcu(&rb->rcu_head, rb_free_rcu); + return true; } static void perf_mmap_open(struct vm_area_struct *vma) @@ -3612,18 +3613,20 @@ static void perf_mmap_close(struct vm_area_struct *vma) struct perf_event *event = vma->vm_file->private_data; if (atomic_dec_and_mutex_lock(&event->mmap_count, &event->mmap_mutex)) { - unsigned long size = perf_data_size(event->rb); - struct user_struct *user = event->mmap_user; struct ring_buffer *rb = event->rb; + struct user_struct *mmap_user = rb->mmap_user; + int mmap_locked = rb->mmap_locked; + unsigned long size = perf_data_size(rb); - atomic_long_sub((size >> PAGE_SHIFT) + 1, &user->locked_vm); - vma->vm_mm->pinned_vm -= event->mmap_locked; rcu_assign_pointer(event->rb, NULL); ring_buffer_detach(event, rb); mutex_unlock(&event->mmap_mutex); - ring_buffer_put(rb); - free_uid(user); + if (ring_buffer_put(rb)) { + atomic_long_sub((size >> PAGE_SHIFT) + 1, &mmap_user->locked_vm); + vma->vm_mm->pinned_vm -= mmap_locked; + free_uid(mmap_user); + } } } @@ -3676,9 +3679,7 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma) WARN_ON_ONCE(event->ctx->parent_ctx); mutex_lock(&event->mmap_mutex); if (event->rb) { - if (event->rb->nr_pages == nr_pages) - atomic_inc(&event->rb->refcount); - else + if (event->rb->nr_pages != nr_pages) ret = -EINVAL; goto unlock; } @@ -3720,12 +3721,14 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma) ret = -ENOMEM; goto unlock; } - rcu_assign_pointer(event->rb, rb); + + rb->mmap_locked = extra; + rb->mmap_user = get_current_user(); atomic_long_add(user_extra, &user->locked_vm); - event->mmap_locked = extra; - event->mmap_user = get_current_user(); - vma->vm_mm->pinned_vm += event->mmap_locked; + vma->vm_mm->pinned_vm += extra; + + rcu_assign_pointer(event->rb, rb); perf_event_update_userpage(event); @@ -3734,7 +3737,7 @@ unlock: atomic_inc(&event->mmap_count); mutex_unlock(&event->mmap_mutex); - vma->vm_flags |= VM_DONTEXPAND | VM_DONTDUMP; + vma->vm_flags |= VM_DONTCOPY | VM_DONTEXPAND | VM_DONTDUMP; vma->vm_ops = &perf_mmap_vmops; return ret; diff --git a/kernel/events/internal.h b/kernel/events/internal.h index eb675c4d59df..5bc6c8e9b851 100644 --- a/kernel/events/internal.h +++ b/kernel/events/internal.h @@ -31,6 +31,9 @@ struct ring_buffer { spinlock_t event_lock; struct list_head event_list; + int mmap_locked; + struct user_struct *mmap_user; + struct perf_event_mmap_page *user_page; void *data_pages[0]; }; -- cgit v1.2.3 From a6572f84c5b135d9b6df279ed3c8de028bd1edd9 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 17 May 2013 10:31:04 +0800 Subject: watchdog: Disallow setting watchdog_thresh to -1 In old kernels, it's allowed to set softlockup_thresh to -1 or 0 to disable softlockup detection. However watchdog_thresh only uses 0 to disable detection, and setting it to -1 just froze my box and nothing I can do but reboot. Signed-off-by: Li Zefan Acked-by: Don Zickus Link: http://lkml.kernel.org/r/51959668.9040106@huawei.com Signed-off-by: Ingo Molnar --- kernel/sysctl.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 9edcf456e0fc..b0a1f99907f3 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -120,7 +120,6 @@ extern int blk_iopoll_enabled; /* Constants used for minimum and maximum */ #ifdef CONFIG_LOCKUP_DETECTOR static int sixty = 60; -static int neg_one = -1; #endif static int zero; @@ -814,7 +813,7 @@ static struct ctl_table kern_table[] = { .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_dowatchdog, - .extra1 = &neg_one, + .extra1 = &zero, .extra2 = &sixty, }, { -- cgit v1.2.3 From 84f9f3a15611536537d59060818a2354d5039ff3 Mon Sep 17 00:00:00 2001 From: Stanislaw Gruszka Date: Thu, 2 May 2013 15:34:33 +0200 Subject: sched: Use swap() macro in scale_stime() Simple cleanup. Reported-by: Peter Zijlstra Signed-off-by: Stanislaw Gruszka Cc: Frederic Weisbecker Link: http://lkml.kernel.org/r/1367501673-6563-1-git-send-email-sgruszka@redhat.com Signed-off-by: Ingo Molnar --- kernel/sched/cputime.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c index cc2dc3eea8a3..94691bcd7364 100644 --- a/kernel/sched/cputime.c +++ b/kernel/sched/cputime.c @@ -515,9 +515,8 @@ static cputime_t scale_stime(u64 stime, u64 rtime, u64 total) for (;;) { /* Make sure "rtime" is the bigger of stime/rtime */ - if (stime > rtime) { - u64 tmp = rtime; rtime = stime; stime = tmp; - } + if (stime > rtime) + swap(rtime, stime); /* Make sure 'total' fits in 32 bits */ if (total >> 32) -- cgit v1.2.3 From cfeaa93f8a13ae9117ae20933a38a406de80849e Mon Sep 17 00:00:00 2001 From: Gerlando Falauto Date: Mon, 6 May 2013 14:30:17 +0000 Subject: genirq: Generic chip: Remove the local cur_regs() function Since we already have an irq_data_get_chip_type() function which returns a pointer to irq_chip_type, use that instead of cur_regs(). Signed-off-by: Gerlando Falauto Cc: Andrew Lunn Cc: Joey Oravec Cc: Lennert Buytenhek Cc: Russell King - ARM Linux Cc: Jason Gunthorpe Cc: Holger Brunck Cc: Ezequiel Garcia Acked-by: Grant Likely Cc: Sebastian Hesselbarth Cc: Jason Cooper Cc: Arnd Bergmann Cc: devicetree-discuss@lists.ozlabs.org Cc: Rob Herring Cc: Ben Dooks Cc: Gregory Clement Cc: Simon Guinot Cc: linux-arm-kernel@lists.infradead.org Cc: Thomas Petazzoni Cc: Jean-Francois Moine Cc: Nicolas Pitre Cc: Rob Landley Cc: Maxime Ripard Link: http://lkml.kernel.org/r/20130506142539.010164766@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/irq/generic-chip.c | 31 +++++++++++++++++-------------- 1 file changed, 17 insertions(+), 14 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/generic-chip.c b/kernel/irq/generic-chip.c index c89295a8f668..0e6ba789056c 100644 --- a/kernel/irq/generic-chip.c +++ b/kernel/irq/generic-chip.c @@ -16,11 +16,6 @@ static LIST_HEAD(gc_list); static DEFINE_RAW_SPINLOCK(gc_lock); -static inline struct irq_chip_regs *cur_regs(struct irq_data *d) -{ - return &container_of(d->chip, struct irq_chip_type, chip)->regs; -} - /** * irq_gc_noop - NOOP function * @d: irq_data @@ -39,10 +34,11 @@ void irq_gc_noop(struct irq_data *d) void irq_gc_mask_disable_reg(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); + struct irq_chip_type *ct = irq_data_get_chip_type(d); u32 mask = 1 << (d->irq - gc->irq_base); irq_gc_lock(gc); - irq_reg_writel(mask, gc->reg_base + cur_regs(d)->disable); + irq_reg_writel(mask, gc->reg_base + ct->regs.disable); gc->mask_cache &= ~mask; irq_gc_unlock(gc); } @@ -57,11 +53,12 @@ void irq_gc_mask_disable_reg(struct irq_data *d) void irq_gc_mask_set_bit(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); + struct irq_chip_type *ct = irq_data_get_chip_type(d); u32 mask = 1 << (d->irq - gc->irq_base); irq_gc_lock(gc); gc->mask_cache |= mask; - irq_reg_writel(gc->mask_cache, gc->reg_base + cur_regs(d)->mask); + irq_reg_writel(gc->mask_cache, gc->reg_base + ct->regs.mask); irq_gc_unlock(gc); } @@ -75,11 +72,12 @@ void irq_gc_mask_set_bit(struct irq_data *d) void irq_gc_mask_clr_bit(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); + struct irq_chip_type *ct = irq_data_get_chip_type(d); u32 mask = 1 << (d->irq - gc->irq_base); irq_gc_lock(gc); gc->mask_cache &= ~mask; - irq_reg_writel(gc->mask_cache, gc->reg_base + cur_regs(d)->mask); + irq_reg_writel(gc->mask_cache, gc->reg_base + ct->regs.mask); irq_gc_unlock(gc); } @@ -93,10 +91,11 @@ void irq_gc_mask_clr_bit(struct irq_data *d) void irq_gc_unmask_enable_reg(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); + struct irq_chip_type *ct = irq_data_get_chip_type(d); u32 mask = 1 << (d->irq - gc->irq_base); irq_gc_lock(gc); - irq_reg_writel(mask, gc->reg_base + cur_regs(d)->enable); + irq_reg_writel(mask, gc->reg_base + ct->regs.enable); gc->mask_cache |= mask; irq_gc_unlock(gc); } @@ -108,10 +107,11 @@ void irq_gc_unmask_enable_reg(struct irq_data *d) void irq_gc_ack_set_bit(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); + struct irq_chip_type *ct = irq_data_get_chip_type(d); u32 mask = 1 << (d->irq - gc->irq_base); irq_gc_lock(gc); - irq_reg_writel(mask, gc->reg_base + cur_regs(d)->ack); + irq_reg_writel(mask, gc->reg_base + ct->regs.ack); irq_gc_unlock(gc); } @@ -122,10 +122,11 @@ void irq_gc_ack_set_bit(struct irq_data *d) void irq_gc_ack_clr_bit(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); + struct irq_chip_type *ct = irq_data_get_chip_type(d); u32 mask = ~(1 << (d->irq - gc->irq_base)); irq_gc_lock(gc); - irq_reg_writel(mask, gc->reg_base + cur_regs(d)->ack); + irq_reg_writel(mask, gc->reg_base + ct->regs.ack); irq_gc_unlock(gc); } @@ -136,11 +137,12 @@ void irq_gc_ack_clr_bit(struct irq_data *d) void irq_gc_mask_disable_reg_and_ack(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); + struct irq_chip_type *ct = irq_data_get_chip_type(d); u32 mask = 1 << (d->irq - gc->irq_base); irq_gc_lock(gc); - irq_reg_writel(mask, gc->reg_base + cur_regs(d)->mask); - irq_reg_writel(mask, gc->reg_base + cur_regs(d)->ack); + irq_reg_writel(mask, gc->reg_base + ct->regs.mask); + irq_reg_writel(mask, gc->reg_base + ct->regs.ack); irq_gc_unlock(gc); } @@ -151,10 +153,11 @@ void irq_gc_mask_disable_reg_and_ack(struct irq_data *d) void irq_gc_eoi(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); + struct irq_chip_type *ct = irq_data_get_chip_type(d); u32 mask = 1 << (d->irq - gc->irq_base); irq_gc_lock(gc); - irq_reg_writel(mask, gc->reg_base + cur_regs(d)->eoi); + irq_reg_writel(mask, gc->reg_base + ct->regs.eoi); irq_gc_unlock(gc); } -- cgit v1.2.3 From 899f0e66fff36ebb6dd6a83af9aa631f6cb7e0dc Mon Sep 17 00:00:00 2001 From: Gerlando Falauto Date: Mon, 6 May 2013 14:30:19 +0000 Subject: genirq: Generic chip: Add support for per chip type mask cache Today the same interrupt mask cache (stored within struct irq_chip_generic) is shared between all the irq_chip_type instances. As there are instances where each irq_chip_type uses a distinct mask register (as it is the case for Orion SoCs), sharing a single mask cache may be incorrect. So add a distinct pointer for each irq_chip_type, which for now points to the original mask register within irq_chip_generic. So no functional changes here. [ tglx: Minor cosmetic tweaks ] Reported-by: Joey Oravec Signed-off-by: Simon Guinot Signed-off-by: Holger Brunck Signed-off-by: Gerlando Falauto Cc: Andrew Lunn Cc: Lennert Buytenhek Cc: Russell King - ARM Linux Cc: Jason Gunthorpe Cc: Holger Brunck Cc: Ezequiel Garcia Acked-by: Grant Likely Cc: Sebastian Hesselbarth Cc: Jason Cooper Cc: Arnd Bergmann Cc: devicetree-discuss@lists.ozlabs.org Cc: Rob Herring Cc: Ben Dooks Cc: Gregory Clement Cc: Simon Guinot Cc: linux-arm-kernel@lists.infradead.org Cc: Thomas Petazzoni Cc: Jean-Francois Moine Cc: Nicolas Pitre Cc: Rob Landley Cc: Maxime Ripard Link: http://lkml.kernel.org/r/20130506142539.082226607@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/irq/generic-chip.c | 16 ++++++++++------ 1 file changed, 10 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/generic-chip.c b/kernel/irq/generic-chip.c index 0e6ba789056c..113d9ebfe0aa 100644 --- a/kernel/irq/generic-chip.c +++ b/kernel/irq/generic-chip.c @@ -39,7 +39,7 @@ void irq_gc_mask_disable_reg(struct irq_data *d) irq_gc_lock(gc); irq_reg_writel(mask, gc->reg_base + ct->regs.disable); - gc->mask_cache &= ~mask; + *ct->mask_cache &= ~mask; irq_gc_unlock(gc); } @@ -57,8 +57,8 @@ void irq_gc_mask_set_bit(struct irq_data *d) u32 mask = 1 << (d->irq - gc->irq_base); irq_gc_lock(gc); - gc->mask_cache |= mask; - irq_reg_writel(gc->mask_cache, gc->reg_base + ct->regs.mask); + *ct->mask_cache |= mask; + irq_reg_writel(*ct->mask_cache, gc->reg_base + ct->regs.mask); irq_gc_unlock(gc); } @@ -76,8 +76,8 @@ void irq_gc_mask_clr_bit(struct irq_data *d) u32 mask = 1 << (d->irq - gc->irq_base); irq_gc_lock(gc); - gc->mask_cache &= ~mask; - irq_reg_writel(gc->mask_cache, gc->reg_base + ct->regs.mask); + *ct->mask_cache &= ~mask; + irq_reg_writel(*ct->mask_cache, gc->reg_base + ct->regs.mask); irq_gc_unlock(gc); } @@ -96,7 +96,7 @@ void irq_gc_unmask_enable_reg(struct irq_data *d) irq_gc_lock(gc); irq_reg_writel(mask, gc->reg_base + ct->regs.enable); - gc->mask_cache |= mask; + *ct->mask_cache |= mask; irq_gc_unlock(gc); } @@ -250,6 +250,10 @@ void irq_setup_generic_chip(struct irq_chip_generic *gc, u32 msk, if (flags & IRQ_GC_INIT_MASK_CACHE) gc->mask_cache = irq_reg_readl(gc->reg_base + ct->regs.mask); + /* Initialize mask cache pointer */ + for (i = 0; i < gc->num_ct; i++) + ct[i].mask_cache = &gc->mask_cache; + for (i = gc->irq_base; msk; msk >>= 1, i++) { if (!(msk & 0x01)) continue; -- cgit v1.2.3 From af80b0fed67261dcba2ce2406db1d553d07cbe75 Mon Sep 17 00:00:00 2001 From: Gerlando Falauto Date: Mon, 6 May 2013 14:30:21 +0000 Subject: genirq: Generic chip: Handle separate mask registers There are cases where all irq_chip_type instances have separate mask registers, making a shared mask register cache unsuitable for the purpose. Introduce a new flag IRQ_GC_MASK_CACHE_PER_TYPE. If set, point the per chip mask pointer to the per chip private mask cache instead. [ tglx: Simplified code, renamed flag and massaged changelog ] Signed-off-by: Gerlando Falauto Cc: Andrew Lunn Cc: Joey Oravec Cc: Lennert Buytenhek Cc: Russell King - ARM Linux Cc: Jason Gunthorpe Cc: Holger Brunck Cc: Ezequiel Garcia Acked-by: Grant Likely Cc: Sebastian Hesselbarth Cc: Jason Cooper Cc: Arnd Bergmann Cc: devicetree-discuss@lists.ozlabs.org Cc: Rob Herring Cc: Ben Dooks Cc: Gregory Clement Cc: Simon Guinot Cc: linux-arm-kernel@lists.infradead.org Cc: Thomas Petazzoni Cc: Jean-Francois Moine Cc: Nicolas Pitre Cc: Rob Landley Cc: Maxime Ripard Link: http://lkml.kernel.org/r/20130506142539.152569748@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/irq/generic-chip.c | 17 ++++++++++------- 1 file changed, 10 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/generic-chip.c b/kernel/irq/generic-chip.c index 113d9ebfe0aa..da2a94191fc5 100644 --- a/kernel/irq/generic-chip.c +++ b/kernel/irq/generic-chip.c @@ -241,18 +241,21 @@ void irq_setup_generic_chip(struct irq_chip_generic *gc, u32 msk, { struct irq_chip_type *ct = gc->chip_types; unsigned int i; + u32 *mskptr = &gc->mask_cache, mskreg = ct->regs.mask; raw_spin_lock(&gc_lock); list_add_tail(&gc->list, &gc_list); raw_spin_unlock(&gc_lock); - /* Init mask cache ? */ - if (flags & IRQ_GC_INIT_MASK_CACHE) - gc->mask_cache = irq_reg_readl(gc->reg_base + ct->regs.mask); - - /* Initialize mask cache pointer */ - for (i = 0; i < gc->num_ct; i++) - ct[i].mask_cache = &gc->mask_cache; + for (i = 0; i < gc->num_ct; i++) { + if (flags & IRQ_GC_MASK_CACHE_PER_TYPE) { + mskptr = &ct[i].mask_cache_priv; + mskreg = ct[i].regs.mask; + } + ct[i].mask_cache = mskptr; + if (flags & IRQ_GC_INIT_MASK_CACHE) + *mskptr = irq_reg_readl(gc->reg_base + mskreg); + } for (i = gc->irq_base; msk; msk >>= 1, i++) { if (!(msk & 0x01)) -- cgit v1.2.3 From 966dc736b819999cd2d3a6408d47d33b579f7d56 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 6 May 2013 14:30:22 +0000 Subject: genirq: Generic chip: Cache per irq bit mask Cache the per irq bit mask instead of recalculating it over and over. Signed-off-by: Thomas Gleixner Cc: Thomas Petazzoni Cc: Andrew Lunn Cc: Russell King - ARM Linux Cc: Jason Cooper Cc: Arnd Bergmann Cc: Jean-Francois Moine Cc: devicetree-discuss@lists.ozlabs.org Cc: Rob Herring Cc: Jason Gunthorpe Cc: Gregory Clement Cc: Gerlando Falauto Cc: Rob Landley Acked-by: Grant Likely Cc: Maxime Ripard Cc: Ezequiel Garcia Cc: linux-arm-kernel@lists.infradead.org Cc: Sebastian Hesselbarth Link: http://lkml.kernel.org/r/20130506142539.227119865@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/irq/generic-chip.c | 23 ++++++++++++++--------- 1 file changed, 14 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/generic-chip.c b/kernel/irq/generic-chip.c index da2a94191fc5..957155cebbac 100644 --- a/kernel/irq/generic-chip.c +++ b/kernel/irq/generic-chip.c @@ -35,7 +35,7 @@ void irq_gc_mask_disable_reg(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); struct irq_chip_type *ct = irq_data_get_chip_type(d); - u32 mask = 1 << (d->irq - gc->irq_base); + u32 mask = d->mask; irq_gc_lock(gc); irq_reg_writel(mask, gc->reg_base + ct->regs.disable); @@ -54,7 +54,7 @@ void irq_gc_mask_set_bit(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); struct irq_chip_type *ct = irq_data_get_chip_type(d); - u32 mask = 1 << (d->irq - gc->irq_base); + u32 mask = d->mask; irq_gc_lock(gc); *ct->mask_cache |= mask; @@ -73,7 +73,7 @@ void irq_gc_mask_clr_bit(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); struct irq_chip_type *ct = irq_data_get_chip_type(d); - u32 mask = 1 << (d->irq - gc->irq_base); + u32 mask = d->mask; irq_gc_lock(gc); *ct->mask_cache &= ~mask; @@ -92,7 +92,7 @@ void irq_gc_unmask_enable_reg(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); struct irq_chip_type *ct = irq_data_get_chip_type(d); - u32 mask = 1 << (d->irq - gc->irq_base); + u32 mask = d->mask; irq_gc_lock(gc); irq_reg_writel(mask, gc->reg_base + ct->regs.enable); @@ -108,7 +108,7 @@ void irq_gc_ack_set_bit(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); struct irq_chip_type *ct = irq_data_get_chip_type(d); - u32 mask = 1 << (d->irq - gc->irq_base); + u32 mask = d->mask; irq_gc_lock(gc); irq_reg_writel(mask, gc->reg_base + ct->regs.ack); @@ -123,7 +123,7 @@ void irq_gc_ack_clr_bit(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); struct irq_chip_type *ct = irq_data_get_chip_type(d); - u32 mask = ~(1 << (d->irq - gc->irq_base)); + u32 mask = ~d->mask; irq_gc_lock(gc); irq_reg_writel(mask, gc->reg_base + ct->regs.ack); @@ -138,7 +138,7 @@ void irq_gc_mask_disable_reg_and_ack(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); struct irq_chip_type *ct = irq_data_get_chip_type(d); - u32 mask = 1 << (d->irq - gc->irq_base); + u32 mask = d->mask; irq_gc_lock(gc); irq_reg_writel(mask, gc->reg_base + ct->regs.mask); @@ -154,7 +154,7 @@ void irq_gc_eoi(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); struct irq_chip_type *ct = irq_data_get_chip_type(d); - u32 mask = 1 << (d->irq - gc->irq_base); + u32 mask = d->mask; irq_gc_lock(gc); irq_reg_writel(mask, gc->reg_base + ct->regs.eoi); @@ -172,7 +172,7 @@ void irq_gc_eoi(struct irq_data *d) int irq_gc_set_wake(struct irq_data *d, unsigned int on) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); - u32 mask = 1 << (d->irq - gc->irq_base); + u32 mask = d->mask; if (!(mask & gc->wake_enabled)) return -EINVAL; @@ -264,6 +264,11 @@ void irq_setup_generic_chip(struct irq_chip_generic *gc, u32 msk, if (flags & IRQ_GC_INIT_NESTED_LOCK) irq_set_lockdep_class(i, &irq_nested_lock_class); + if (!(flags & IRQ_GC_NO_MASK)) { + struct irq_data *d = irq_get_irq_data(i); + + d->mask = 1 << (i - gc->irq_base); + } irq_set_chip_and_handler(i, &ct->chip, ct->handler); irq_set_chip_data(i, gc); irq_modify_status(i, clr, set); -- cgit v1.2.3 From d0051816e619f8f082582bec07ffa51bdb4f2104 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 6 May 2013 14:30:24 +0000 Subject: genirq: irqchip: Add a mask calculation function Some chips have weird bit mask access patterns instead of the linear you expect. Allow them to calculate the cached mask themself. Signed-off-by: Thomas Gleixner Cc: Thomas Petazzoni Cc: Andrew Lunn Cc: Russell King - ARM Linux Cc: Jason Cooper Cc: Arnd Bergmann Cc: Jean-Francois Moine Cc: devicetree-discuss@lists.ozlabs.org Cc: Rob Herring Cc: Jason Gunthorpe Cc: Gregory Clement Cc: Gerlando Falauto Cc: Rob Landley Acked-by: Grant Likely Cc: Maxime Ripard Cc: Ezequiel Garcia Cc: linux-arm-kernel@lists.infradead.org Cc: Sebastian Hesselbarth Link: http://lkml.kernel.org/r/20130506142539.302898834@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/irq/generic-chip.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/generic-chip.c b/kernel/irq/generic-chip.c index 957155cebbac..5068fe3ae1af 100644 --- a/kernel/irq/generic-chip.c +++ b/kernel/irq/generic-chip.c @@ -240,6 +240,7 @@ void irq_setup_generic_chip(struct irq_chip_generic *gc, u32 msk, unsigned int set) { struct irq_chip_type *ct = gc->chip_types; + struct irq_chip *chip = &ct->chip; unsigned int i; u32 *mskptr = &gc->mask_cache, mskreg = ct->regs.mask; @@ -267,9 +268,12 @@ void irq_setup_generic_chip(struct irq_chip_generic *gc, u32 msk, if (!(flags & IRQ_GC_NO_MASK)) { struct irq_data *d = irq_get_irq_data(i); - d->mask = 1 << (i - gc->irq_base); + if (chip->irq_calc_mask) + chip->irq_calc_mask(d); + else + d->mask = 1 << (i - gc->irq_base); } - irq_set_chip_and_handler(i, &ct->chip, ct->handler); + irq_set_chip_and_handler(i, chip, ct->handler); irq_set_chip_data(i, gc); irq_modify_status(i, clr, set); } -- cgit v1.2.3 From 3528d82b684680b72fa31881c8c572c5a98b51de Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 6 May 2013 14:30:25 +0000 Subject: genirq: Generic chip: Split out code into separate functions Preparatory patch for linear interrupt domains. Signed-off-by: Thomas Gleixner Cc: Thomas Petazzoni Cc: Andrew Lunn Cc: Russell King - ARM Linux Cc: Jason Cooper Cc: Arnd Bergmann Cc: Jean-Francois Moine Cc: devicetree-discuss@lists.ozlabs.org Cc: Rob Herring Cc: Jason Gunthorpe Cc: Gregory Clement Cc: Gerlando Falauto Cc: Rob Landley Acked-by: Grant Likely Cc: Maxime Ripard Cc: Ezequiel Garcia Cc: linux-arm-kernel@lists.infradead.org Cc: Sebastian Hesselbarth Link: http://lkml.kernel.org/r/20130506142539.377017672@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/irq/generic-chip.c | 50 ++++++++++++++++++++++++++++++++--------------- 1 file changed, 34 insertions(+), 16 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/generic-chip.c b/kernel/irq/generic-chip.c index 5068fe3ae1af..3deb3333d53e 100644 --- a/kernel/irq/generic-chip.c +++ b/kernel/irq/generic-chip.c @@ -186,6 +186,19 @@ int irq_gc_set_wake(struct irq_data *d, unsigned int on) return 0; } +static void +irq_init_generic_chip(struct irq_chip_generic *gc, const char *name, + int num_ct, unsigned int irq_base, + void __iomem *reg_base, irq_flow_handler_t handler) +{ + raw_spin_lock_init(&gc->lock); + gc->num_ct = num_ct; + gc->irq_base = irq_base; + gc->reg_base = reg_base; + gc->chip_types->chip.name = name; + gc->chip_types->handler = handler; +} + /** * irq_alloc_generic_chip - Allocate a generic chip and initialize it * @name: Name of the irq chip @@ -206,17 +219,31 @@ irq_alloc_generic_chip(const char *name, int num_ct, unsigned int irq_base, gc = kzalloc(sz, GFP_KERNEL); if (gc) { - raw_spin_lock_init(&gc->lock); - gc->num_ct = num_ct; - gc->irq_base = irq_base; - gc->reg_base = reg_base; - gc->chip_types->chip.name = name; - gc->chip_types->handler = handler; + irq_init_generic_chip(gc, name, num_ct, irq_base, reg_base, + handler); } return gc; } EXPORT_SYMBOL_GPL(irq_alloc_generic_chip); +static void +irq_gc_init_mask_cache(struct irq_chip_generic *gc, enum irq_gc_flags flags) +{ + struct irq_chip_type *ct = gc->chip_types; + u32 *mskptr = &gc->mask_cache, mskreg = ct->regs.mask; + int i; + + for (i = 0; i < gc->num_ct; i++) { + if (flags & IRQ_GC_MASK_CACHE_PER_TYPE) { + mskptr = &ct[i].mask_cache_priv; + mskreg = ct[i].regs.mask; + } + ct[i].mask_cache = mskptr; + if (flags & IRQ_GC_INIT_MASK_CACHE) + *mskptr = irq_reg_readl(gc->reg_base + mskreg); + } +} + /* * Separate lockdep class for interrupt chip which can nest irq_desc * lock. @@ -242,21 +269,12 @@ void irq_setup_generic_chip(struct irq_chip_generic *gc, u32 msk, struct irq_chip_type *ct = gc->chip_types; struct irq_chip *chip = &ct->chip; unsigned int i; - u32 *mskptr = &gc->mask_cache, mskreg = ct->regs.mask; raw_spin_lock(&gc_lock); list_add_tail(&gc->list, &gc_list); raw_spin_unlock(&gc_lock); - for (i = 0; i < gc->num_ct; i++) { - if (flags & IRQ_GC_MASK_CACHE_PER_TYPE) { - mskptr = &ct[i].mask_cache_priv; - mskreg = ct[i].regs.mask; - } - ct[i].mask_cache = mskptr; - if (flags & IRQ_GC_INIT_MASK_CACHE) - *mskptr = irq_reg_readl(gc->reg_base + mskreg); - } + irq_gc_init_mask_cache(gc, flags); for (i = gc->irq_base; msk; msk >>= 1, i++) { if (!(msk & 0x01)) -- cgit v1.2.3 From 088f40b7b027dad6519712ff224a5798dd62a204 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 6 May 2013 14:30:27 +0000 Subject: genirq: Generic chip: Add linear irq domain support Provide infrastructure for irq chip implementations which work on linear irq domains. - Interface to allocate multiple generic chips which are associated to the irq domain. - Interface to get the generic chip pointer for a particular hardware interrupt in the domain. - irq domain mapping function to install the chip for a particular interrupt. Note: This lacks a removal function for now. [ Sebastian Hesselbarth: Mask cache and pointer math fixups ] Signed-off-by: Thomas Gleixner Cc: Thomas Petazzoni Cc: Andrew Lunn Cc: Russell King - ARM Linux Cc: Jason Cooper Cc: Arnd Bergmann Cc: Jean-Francois Moine Cc: devicetree-discuss@lists.ozlabs.org Cc: Rob Herring Cc: Jason Gunthorpe Cc: Gregory Clement Cc: Gerlando Falauto Cc: Rob Landley Acked-by: Grant Likely Cc: Maxime Ripard Cc: Ezequiel Garcia Cc: linux-arm-kernel@lists.infradead.org Cc: Sebastian Hesselbarth Link: http://lkml.kernel.org/r/20130506142539.450634298@linutronix.de Signed-off-by: Thomas Gleixner --- kernel/irq/generic-chip.c | 187 ++++++++++++++++++++++++++++++++++++++++++++-- kernel/irq/irqdomain.c | 6 -- 2 files changed, 181 insertions(+), 12 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/generic-chip.c b/kernel/irq/generic-chip.c index 3deb3333d53e..8743d62fded7 100644 --- a/kernel/irq/generic-chip.c +++ b/kernel/irq/generic-chip.c @@ -7,6 +7,7 @@ #include #include #include +#include #include #include #include @@ -244,12 +245,156 @@ irq_gc_init_mask_cache(struct irq_chip_generic *gc, enum irq_gc_flags flags) } } +/** + * irq_alloc_domain_generic_chip - Allocate generic chips for an irq domain + * @d: irq domain for which to allocate chips + * @irqs_per_chip: Number of interrupts each chip handles + * @num_ct: Number of irq_chip_type instances associated with this + * @name: Name of the irq chip + * @handler: Default flow handler associated with these chips + * @clr: IRQ_* bits to clear in the mapping function + * @set: IRQ_* bits to set in the mapping function + */ +int irq_alloc_domain_generic_chips(struct irq_domain *d, int irqs_per_chip, + int num_ct, const char *name, + irq_flow_handler_t handler, + unsigned int clr, unsigned int set, + enum irq_gc_flags gcflags) +{ + struct irq_domain_chip_generic *dgc; + struct irq_chip_generic *gc; + int numchips, sz, i; + unsigned long flags; + void *tmp; + + if (d->gc) + return -EBUSY; + + if (d->revmap_type != IRQ_DOMAIN_MAP_LINEAR) + return -EINVAL; + + numchips = d->revmap_data.linear.size / irqs_per_chip; + if (!numchips) + return -EINVAL; + + /* Allocate a pointer, generic chip and chiptypes for each chip */ + sz = sizeof(*dgc) + numchips * sizeof(gc); + sz += numchips * (sizeof(*gc) + num_ct * sizeof(struct irq_chip_type)); + + tmp = dgc = kzalloc(sz, GFP_KERNEL); + if (!dgc) + return -ENOMEM; + dgc->irqs_per_chip = irqs_per_chip; + dgc->num_chips = numchips; + dgc->irq_flags_to_set = set; + dgc->irq_flags_to_clear = clr; + dgc->gc_flags = gcflags; + d->gc = dgc; + + /* Calc pointer to the first generic chip */ + tmp += sizeof(*dgc) + numchips * sizeof(gc); + for (i = 0; i < numchips; i++) { + /* Store the pointer to the generic chip */ + dgc->gc[i] = gc = tmp; + irq_init_generic_chip(gc, name, num_ct, i * irqs_per_chip, + NULL, handler); + gc->domain = d; + raw_spin_lock_irqsave(&gc_lock, flags); + list_add_tail(&gc->list, &gc_list); + raw_spin_unlock_irqrestore(&gc_lock, flags); + /* Calc pointer to the next generic chip */ + tmp += sizeof(*gc) + num_ct * sizeof(struct irq_chip_type); + } + return 0; +} +EXPORT_SYMBOL_GPL(irq_alloc_domain_generic_chips); + +/** + * irq_get_domain_generic_chip - Get a pointer to the generic chip of a hw_irq + * @d: irq domain pointer + * @hw_irq: Hardware interrupt number + */ +struct irq_chip_generic * +irq_get_domain_generic_chip(struct irq_domain *d, unsigned int hw_irq) +{ + struct irq_domain_chip_generic *dgc = d->gc; + int idx; + + if (!dgc) + return NULL; + idx = hw_irq / dgc->irqs_per_chip; + if (idx >= dgc->num_chips) + return NULL; + return dgc->gc[idx]; +} +EXPORT_SYMBOL_GPL(irq_get_domain_generic_chip); + /* * Separate lockdep class for interrupt chip which can nest irq_desc * lock. */ static struct lock_class_key irq_nested_lock_class; +/** + * irq_map_generic_chip - Map a generic chip for an irq domain + */ +static int irq_map_generic_chip(struct irq_domain *d, unsigned int virq, + irq_hw_number_t hw_irq) +{ + struct irq_data *data = irq_get_irq_data(virq); + struct irq_domain_chip_generic *dgc = d->gc; + struct irq_chip_generic *gc; + struct irq_chip_type *ct; + struct irq_chip *chip; + unsigned long flags; + int idx; + + if (!d->gc) + return -ENODEV; + + idx = hw_irq / dgc->irqs_per_chip; + if (idx >= dgc->num_chips) + return -EINVAL; + gc = dgc->gc[idx]; + + idx = hw_irq % dgc->irqs_per_chip; + + if (test_bit(idx, &gc->installed)) + return -EBUSY; + + ct = gc->chip_types; + chip = &ct->chip; + + /* We only init the cache for the first mapping of a generic chip */ + if (!gc->installed) { + raw_spin_lock_irqsave(&gc->lock, flags); + irq_gc_init_mask_cache(gc, dgc->gc_flags); + raw_spin_unlock_irqrestore(&gc->lock, flags); + } + + /* Mark the interrupt as installed */ + set_bit(idx, &gc->installed); + + if (dgc->gc_flags & IRQ_GC_INIT_NESTED_LOCK) + irq_set_lockdep_class(virq, &irq_nested_lock_class); + + if (chip->irq_calc_mask) + chip->irq_calc_mask(data); + else + data->mask = 1 << idx; + + irq_set_chip_and_handler(virq, chip, ct->handler); + irq_set_chip_data(virq, gc); + irq_modify_status(virq, dgc->irq_flags_to_clear, dgc->irq_flags_to_set); + return 0; +} + +struct irq_domain_ops irq_generic_chip_ops = { + .map = irq_map_generic_chip, + .xlate = irq_domain_xlate_onetwocell, +}; +EXPORT_SYMBOL_GPL(irq_generic_chip_ops); + /** * irq_setup_generic_chip - Setup a range of interrupts with a generic chip * @gc: Generic irq chip holding all data @@ -354,6 +499,24 @@ void irq_remove_generic_chip(struct irq_chip_generic *gc, u32 msk, } EXPORT_SYMBOL_GPL(irq_remove_generic_chip); +static struct irq_data *irq_gc_get_irq_data(struct irq_chip_generic *gc) +{ + unsigned int virq; + + if (!gc->domain) + return irq_get_irq_data(gc->irq_base); + + /* + * We don't know which of the irqs has been actually + * installed. Use the first one. + */ + if (!gc->installed) + return NULL; + + virq = irq_find_mapping(gc->domain, gc->irq_base + __ffs(gc->installed)); + return virq ? irq_get_irq_data(virq) : NULL; +} + #ifdef CONFIG_PM static int irq_gc_suspend(void) { @@ -362,8 +525,12 @@ static int irq_gc_suspend(void) list_for_each_entry(gc, &gc_list, list) { struct irq_chip_type *ct = gc->chip_types; - if (ct->chip.irq_suspend) - ct->chip.irq_suspend(irq_get_irq_data(gc->irq_base)); + if (ct->chip.irq_suspend) { + struct irq_data *data = irq_gc_get_irq_data(gc); + + if (data) + ct->chip.irq_suspend(data); + } } return 0; } @@ -375,8 +542,12 @@ static void irq_gc_resume(void) list_for_each_entry(gc, &gc_list, list) { struct irq_chip_type *ct = gc->chip_types; - if (ct->chip.irq_resume) - ct->chip.irq_resume(irq_get_irq_data(gc->irq_base)); + if (ct->chip.irq_resume) { + struct irq_data *data = irq_gc_get_irq_data(gc); + + if (data) + ct->chip.irq_resume(data); + } } } #else @@ -391,8 +562,12 @@ static void irq_gc_shutdown(void) list_for_each_entry(gc, &gc_list, list) { struct irq_chip_type *ct = gc->chip_types; - if (ct->chip.irq_pm_shutdown) - ct->chip.irq_pm_shutdown(irq_get_irq_data(gc->irq_base)); + if (ct->chip.irq_pm_shutdown) { + struct irq_data *data = irq_gc_get_irq_data(gc); + + if (data) + ct->chip.irq_pm_shutdown(data); + } } } diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c index 5a83dde8ca0c..1db9e70f5488 100644 --- a/kernel/irq/irqdomain.c +++ b/kernel/irq/irqdomain.c @@ -16,12 +16,6 @@ #include #include -#define IRQ_DOMAIN_MAP_LEGACY 0 /* driver allocated fixed range of irqs. - * ie. legacy 8259, gets irqs 1..15 */ -#define IRQ_DOMAIN_MAP_NOMAP 1 /* no fast reverse mapping */ -#define IRQ_DOMAIN_MAP_LINEAR 2 /* linear map of interrupts */ -#define IRQ_DOMAIN_MAP_TREE 3 /* radix tree */ - static LIST_HEAD(irq_domain_list); static DEFINE_MUTEX(irq_domain_mutex); -- cgit v1.2.3 From e8bd834f73714378ef110a64287db1b77033c8da Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Wed, 29 May 2013 03:10:52 +0100 Subject: genirq: irqchip: Add mask to block out invalid irqs Some controllers have irqs that aren't wired up and must never be used. For the generic chip attached to an irq_domain this provides a mask that can be used to block out particular irqs so that they never get mapped. Signed-off-by: Grant Likely Link: http://lkml.kernel.org/r/1369793454-19197-2-git-send-email-grant.likely@linaro.org Signed-off-by: Thomas Gleixner --- kernel/irq/generic-chip.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'kernel') diff --git a/kernel/irq/generic-chip.c b/kernel/irq/generic-chip.c index 8743d62fded7..95575d8d5392 100644 --- a/kernel/irq/generic-chip.c +++ b/kernel/irq/generic-chip.c @@ -359,6 +359,9 @@ static int irq_map_generic_chip(struct irq_domain *d, unsigned int virq, idx = hw_irq % dgc->irqs_per_chip; + if (test_bit(idx, &gc->unused)) + return -ENOTSUPP; + if (test_bit(idx, &gc->installed)) return -EBUSY; -- cgit v1.2.3 From d671a605580d2caafc77f1a25bcf8435795df6fe Mon Sep 17 00:00:00 2001 From: Andreas Fenkart Date: Fri, 10 May 2013 12:21:30 +0200 Subject: genirq: Add kerneldoc for irq_disable. Document the lazy disable functionality. comment based on changelog of d209a699a0b975ad Signed-off-by: Andreas Fenkart Cc: balbi@ti.com Link: http://lkml.kernel.org/r/1368181290-1583-1-git-send-email-andreas.fenkart@streamunlimited.com Signed-off-by: Thomas Gleixner --- kernel/irq/chip.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'kernel') diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index cbd97ce0b000..a3bb14fbe5c6 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -213,6 +213,19 @@ void irq_enable(struct irq_desc *desc) irq_state_clr_masked(desc); } +/** + * irq_disable - Mark interupt disabled + * @desc: irq descriptor which should be disabled + * + * If the chip does not implement the irq_disable callback, we + * use a lazy disable approach. That means we mark the interrupt + * disabled, but leave the hardware unmasked. That's an + * optimization because we avoid the hardware access for the + * common case where no interrupt happens after we marked it + * disabled. If an interrupt happens, then the interrupt flow + * handler masks the line at the hardware level and marks it + * pending. + */ void irq_disable(struct irq_desc *desc) { irq_state_set_disabled(desc); -- cgit v1.2.3 From 6cffe00f7d4e24679eae6b7aae4caaf915288256 Mon Sep 17 00:00:00 2001 From: Todd Poynor Date: Wed, 15 May 2013 14:38:11 -0700 Subject: alarmtimer: Add functions for timerfd support Add functions needed for hooking up alarmtimer to timerfd: * alarm_restart: Similar to hrtimer_restart, restart an alarmtimer after the expires time has already been updated (as with alarm_forward). * alarm_forward_now: Similar to hrtimer_forward_now, move the expires time forward to an interval from the current time of the associated clock. * alarm_start_relative: Start an alarmtimer with an expires time relative to the current time of the associated clock. * alarm_expires_remaining: Similar to hrtimer_expires_remaining, return the amount of time remaining until alarm expiry. Signed-off-by: Todd Poynor Signed-off-by: John Stultz --- kernel/time/alarmtimer.c | 39 ++++++++++++++++++++++++++++++++++++++- 1 file changed, 38 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/time/alarmtimer.c b/kernel/time/alarmtimer.c index f11d83b12949..3e5cba274475 100644 --- a/kernel/time/alarmtimer.c +++ b/kernel/time/alarmtimer.c @@ -199,6 +199,12 @@ static enum hrtimer_restart alarmtimer_fired(struct hrtimer *timer) } +ktime_t alarm_expires_remaining(const struct alarm *alarm) +{ + struct alarm_base *base = &alarm_bases[alarm->type]; + return ktime_sub(alarm->node.expires, base->gettime()); +} + #ifdef CONFIG_RTC_CLASS /** * alarmtimer_suspend - Suspend time callback @@ -305,7 +311,7 @@ void alarm_init(struct alarm *alarm, enum alarmtimer_type type, } /** - * alarm_start - Sets an alarm to fire + * alarm_start - Sets an absolute alarm to fire * @alarm: ptr to alarm to set * @start: time to run the alarm */ @@ -324,6 +330,31 @@ int alarm_start(struct alarm *alarm, ktime_t start) return ret; } +/** + * alarm_start_relative - Sets a relative alarm to fire + * @alarm: ptr to alarm to set + * @start: time relative to now to run the alarm + */ +int alarm_start_relative(struct alarm *alarm, ktime_t start) +{ + struct alarm_base *base = &alarm_bases[alarm->type]; + + start = ktime_add(start, base->gettime()); + return alarm_start(alarm, start); +} + +void alarm_restart(struct alarm *alarm) +{ + struct alarm_base *base = &alarm_bases[alarm->type]; + unsigned long flags; + + spin_lock_irqsave(&base->lock, flags); + hrtimer_set_expires(&alarm->timer, alarm->node.expires); + hrtimer_restart(&alarm->timer); + alarmtimer_enqueue(base, alarm); + spin_unlock_irqrestore(&base->lock, flags); +} + /** * alarm_try_to_cancel - Tries to cancel an alarm timer * @alarm: ptr to alarm to be canceled @@ -394,6 +425,12 @@ u64 alarm_forward(struct alarm *alarm, ktime_t now, ktime_t interval) return overrun; } +u64 alarm_forward_now(struct alarm *alarm, ktime_t interval) +{ + struct alarm_base *base = &alarm_bases[alarm->type]; + + return alarm_forward(alarm, base->gettime(), interval); +} -- cgit v1.2.3 From 5c83545f24ab3dd67e0ae0e2b795fea750f08c34 Mon Sep 17 00:00:00 2001 From: Colin Cross Date: Tue, 21 May 2013 22:32:14 -0700 Subject: power: Add option to log time spent in suspend Below is a patch from android kernel that maintains a histogram of suspend times. Please review and provide feedback. Statistices on the time spent in suspend are kept in /sys/kernel/debug/sleep_time. Cc: Android Kernel Team Cc: Colin Cross Cc: Todd Poynor Cc: San Mehat Cc: Benoit Goby Cc: John Stultz Cc: Thomas Gleixner Signed-off-by: Colin Cross Signed-off-by: Todd Poynor [zoran.markovic@linaro.org: Re-formatted suspend time table to better fit expected values. Moved accounting of suspend time into timekeeping core. Removed CONFIG_SUSPEND_TIME flag and made the feature conditional on CONFIG_DEBUG_FS. Changed the file name to sleep_time to better fit terminology in timekeeping core. Changed seq_printf to seq_puts. Tweaked commit message] Signed-off-by: Zoran Markovic Signed-off-by: John Stultz --- kernel/time/Makefile | 1 + kernel/time/timekeeping.c | 2 ++ kernel/time/timekeeping_debug.c | 72 ++++++++++++++++++++++++++++++++++++++ kernel/time/timekeeping_internal.h | 14 ++++++++ 4 files changed, 89 insertions(+) create mode 100644 kernel/time/timekeeping_debug.c create mode 100644 kernel/time/timekeeping_internal.h (limited to 'kernel') diff --git a/kernel/time/Makefile b/kernel/time/Makefile index ff7d9d2ab504..d52ac8bf0006 100644 --- a/kernel/time/Makefile +++ b/kernel/time/Makefile @@ -7,3 +7,4 @@ obj-$(CONFIG_GENERIC_CLOCKEVENTS_BROADCAST) += tick-broadcast.o obj-$(CONFIG_TICK_ONESHOT) += tick-oneshot.o obj-$(CONFIG_TICK_ONESHOT) += tick-sched.o obj-$(CONFIG_TIMER_STATS) += timer_stats.o +obj-$(CONFIG_DEBUG_FS) += timekeeping_debug.o diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index 933efa4071c3..838fc0777b68 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -25,6 +25,7 @@ #include "tick-internal.h" #include "ntp_internal.h" +#include "timekeeping_internal.h" static struct timekeeper timekeeper; static DEFINE_RAW_SPINLOCK(timekeeper_lock); @@ -851,6 +852,7 @@ static void __timekeeping_inject_sleeptime(struct timekeeper *tk, tk_xtime_add(tk, delta); tk_set_wall_to_mono(tk, timespec_sub(tk->wall_to_monotonic, *delta)); tk_set_sleep_time(tk, timespec_add(tk->total_sleep_time, *delta)); + tk_debug_account_sleep_time(delta); } /** diff --git a/kernel/time/timekeeping_debug.c b/kernel/time/timekeeping_debug.c new file mode 100644 index 000000000000..802433a4f5eb --- /dev/null +++ b/kernel/time/timekeeping_debug.c @@ -0,0 +1,72 @@ +/* + * debugfs file to track time spent in suspend + * + * Copyright (c) 2011, Google, Inc. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + */ + +#include +#include +#include +#include +#include +#include + +static unsigned int sleep_time_bin[32] = {0}; + +static int tk_debug_show_sleep_time(struct seq_file *s, void *data) +{ + unsigned int bin; + seq_puts(s, " time (secs) count\n"); + seq_puts(s, "------------------------------\n"); + for (bin = 0; bin < 32; bin++) { + if (sleep_time_bin[bin] == 0) + continue; + seq_printf(s, "%10u - %-10u %4u\n", + bin ? 1 << (bin - 1) : 0, 1 << bin, + sleep_time_bin[bin]); + } + return 0; +} + +static int tk_debug_sleep_time_open(struct inode *inode, struct file *file) +{ + return single_open(file, tk_debug_show_sleep_time, NULL); +} + +static const struct file_operations tk_debug_sleep_time_fops = { + .open = tk_debug_sleep_time_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, +}; + +static int __init tk_debug_sleep_time_init(void) +{ + struct dentry *d; + + d = debugfs_create_file("sleep_time", 0444, NULL, NULL, + &tk_debug_sleep_time_fops); + if (!d) { + pr_err("Failed to create sleep_time debug file\n"); + return -ENOMEM; + } + + return 0; +} +late_initcall(tk_debug_sleep_time_init); + +void tk_debug_account_sleep_time(struct timespec *t) +{ + sleep_time_bin[fls(t->tv_sec)]++; +} + diff --git a/kernel/time/timekeeping_internal.h b/kernel/time/timekeeping_internal.h new file mode 100644 index 000000000000..13323ea08ffa --- /dev/null +++ b/kernel/time/timekeeping_internal.h @@ -0,0 +1,14 @@ +#ifndef _TIMEKEEPING_INTERNAL_H +#define _TIMEKEEPING_INTERNAL_H +/* + * timekeeping debug functions + */ +#include + +#ifdef CONFIG_DEBUG_FS +extern void tk_debug_account_sleep_time(struct timespec *t); +#else +#define tk_debug_account_sleep_time(x) +#endif + +#endif /* _TIMEKEEPING_INTERNAL_H */ -- cgit v1.2.3 From 45eacc692771bd2b1ea3d384e6345cab3da10861 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 15 May 2013 22:16:32 +0200 Subject: vtime: Use consistent clocks among nohz accounting While computing the cputime delta of dynticks CPUs, we are mixing up clocks of differents natures: * local_clock() which takes care of unstable clock sources and fix these if needed. * sched_clock() which is the weaker version of local_clock(). It doesn't compute any fixup in case of unstable source. If the clock source is stable, those two clocks are the same and we can safely compute the difference against two random points. Otherwise it results in random deltas as sched_clock() can randomly drift away, back or forward, from local_clock(). As a consequence, some strange behaviour with unstable tsc has been observed such as non progressing constant zero cputime. (The 'top' command showing no load). Fix this by only using local_clock(), or its irq safe/remote equivalent, in vtime code. Reported-by: Mike Galbraith Suggested-by: Mike Galbraith Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 2 +- kernel/sched/cputime.c | 6 +++--- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 58453b8272fd..e1a27f918723 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -4745,7 +4745,7 @@ void __cpuinit init_idle(struct task_struct *idle, int cpu) */ idle->sched_class = &idle_sched_class; ftrace_graph_init_idle_task(idle, cpu); - vtime_init_idle(idle); + vtime_init_idle(idle, cpu); #if defined(CONFIG_SMP) sprintf(idle->comm, "%s/%d", INIT_TASK_COMM, cpu); #endif diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c index cc2dc3eea8a3..b5ccba22603b 100644 --- a/kernel/sched/cputime.c +++ b/kernel/sched/cputime.c @@ -747,17 +747,17 @@ void arch_vtime_task_switch(struct task_struct *prev) write_seqlock(¤t->vtime_seqlock); current->vtime_snap_whence = VTIME_SYS; - current->vtime_snap = sched_clock(); + current->vtime_snap = sched_clock_cpu(smp_processor_id()); write_sequnlock(¤t->vtime_seqlock); } -void vtime_init_idle(struct task_struct *t) +void vtime_init_idle(struct task_struct *t, int cpu) { unsigned long flags; write_seqlock_irqsave(&t->vtime_seqlock, flags); t->vtime_snap_whence = VTIME_SYS; - t->vtime_snap = sched_clock(); + t->vtime_snap = sched_clock_cpu(cpu); write_sequnlock_irqrestore(&t->vtime_seqlock, flags); } -- cgit v1.2.3 From 521921bad1192fb1b8f9b6a5aa673635848b8b5f Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 16 May 2013 01:21:38 +0200 Subject: kvm: Move guest entry/exit APIs to context_tracking The kvm_host.h header file doesn't handle well inclusion when archs don't support KVM. This results in build crashes for such archs when they want to implement context tracking because this subsystem includes kvm_host.h in order to implement the guest_enter/exit APIs but it doesn't handle KVM off case. To fix this, move the guest_enter()/guest_exit() declarations and generic implementation to the context tracking headers. These generic APIs actually belong to this subsystem, besides other domains boundary tracking like user_enter() et al. KVM now properly becomes a user of this library, not the other buggy way around. Reported-by: Kevin Hilman Reviewed-by: Kevin Hilman Tested-by: Kevin Hilman Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Kevin Hilman Cc: Marcelo Tosatti Cc: Gleb Natapov Signed-off-by: Ingo Molnar --- kernel/context_tracking.c | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/context_tracking.c b/kernel/context_tracking.c index 65349f07b878..85bdde1137eb 100644 --- a/kernel/context_tracking.c +++ b/kernel/context_tracking.c @@ -15,7 +15,6 @@ */ #include -#include #include #include #include -- cgit v1.2.3 From 1a7f829f094dd7951e7d46c571a18080e455a436 Mon Sep 17 00:00:00 2001 From: Li Zhong Date: Fri, 17 May 2013 16:44:04 +0800 Subject: nohz: Fix notifier return val that enforce timekeeping In tick_nohz_cpu_down_callback() if the cpu is the one handling timekeeping, we must return something that stops the CPU_DOWN_PREPARE notifiers and then start notify CPU_DOWN_FAILED on the already called notifier call backs. However traditional errno values are not handled by the notifier unless these are encapsulated using errno_to_notifier(). Hence the current -EINVAL is misinterpreted and converted to junk after notifier_to_errno(), leaving the notifier subsystem to random behaviour such as eventually allowing the cpu to go down. Fix this by using the standard NOTIFY_BAD instead. Signed-off-by: Li Zhong Reviewed-by: Srivatsa S. Bhat Acked-by: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/time/tick-sched.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index f4208138fbf4..0cf1c1453181 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -306,7 +306,7 @@ static int __cpuinit tick_nohz_cpu_down_callback(struct notifier_block *nfb, * we can't safely shutdown that CPU. */ if (have_nohz_full_mask && tick_do_timer_cpu == cpu) - return -EINVAL; + return NOTIFY_BAD; break; } return NOTIFY_OK; -- cgit v1.2.3 From 0de358f1c2642710d41190b73fbc295e675c4ab8 Mon Sep 17 00:00:00 2001 From: Kamalesh Babulal Date: Thu, 30 May 2013 14:34:20 +0530 Subject: sched/fair: Remove unused variable from expire_cfs_rq_runtime() Commit 78becc2709 ("sched: Use an accessor to read the rq clock") introduces rq_clock(), which obsoletes the use of the "rq" variable in expire_cfs_rq_runtime() and triggers this build warning: kernel/sched/fair.c: In function 'expire_cfs_rq_runtime': kernel/sched/fair.c:2159:13: warning: unused variable 'rq' [-Wunused-variable] Signed-off-by: Kamalesh Babulal Acked-by: Frederic Weisbecker Acked-by: Paul Turner Cc: peterz@infradead.org Link: http://lkml.kernel.org/r/1369904660-14169-1-git-send-email-kamalesh@linux.vnet.ibm.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 3ee1c2e4ae60..143dcdbc47af 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -2156,7 +2156,6 @@ static int assign_cfs_rq_runtime(struct cfs_rq *cfs_rq) static void expire_cfs_rq_runtime(struct cfs_rq *cfs_rq) { struct cfs_bandwidth *cfs_b = tg_cfs_bandwidth(cfs_rq->tg); - struct rq *rq = rq_of(cfs_rq); /* if the deadline is ahead of our clock, nothing to do */ if (likely((s64)(rq_clock(rq_of(cfs_rq)) - cfs_rq->runtime_expires) < 0)) -- cgit v1.2.3 From f5d00c1f9adb350c24c5301600f7bf2da99b66de Mon Sep 17 00:00:00 2001 From: Jiri Bohac Date: Tue, 28 May 2013 15:29:03 +0200 Subject: tick: Remove useless timekeeping duty attribution to broadcast source Since 7300711e ("clockevents: broadcast fixup possible waiters"), the timekeeping duty is assigned to the CPU that handles the tick broadcast clock device by the time it is set in one shot mode. This is an issue in full dynticks mode where the timekeeping duty must stay handled by the boot CPU for now. Otherwise it prevents secondary CPUs from offlining and this breaks suspend/shutdown/reboot/... As it appears there is no reason for this timekeeping duty to be moved to the broadcast CPU, besides nothing prevent it from being later re-assigned to another target, let's simply remove it. Signed-off-by: Jiri Bohac Reported-by: Steven Rostedt Acked-by: Thomas Gleixner Cc: Steven Rostedt Cc: Thomas Gleixner Cc: Paul E. McKenney Cc: Peter Zijlstra Cc: Borislav Petkov Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/time/tick-broadcast.c | 4 ---- 1 file changed, 4 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-broadcast.c b/kernel/time/tick-broadcast.c index 0c739423b0f9..b4c245580b79 100644 --- a/kernel/time/tick-broadcast.c +++ b/kernel/time/tick-broadcast.c @@ -698,10 +698,6 @@ void tick_broadcast_setup_oneshot(struct clock_event_device *bc) bc->event_handler = tick_handle_oneshot_broadcast; - /* Take the do_timer update */ - if (!tick_nohz_full_cpu(cpu)) - tick_do_timer_cpu = cpu; - /* * We must be careful here. There might be other CPUs * waiting for periodic broadcast. We need to set the -- cgit v1.2.3 From cd38ca854de15b26eb91009137cbe157d8a8e773 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Mon, 3 Jun 2013 18:20:29 +0000 Subject: PM / Hibernate: print physical addresses consistently with other parts of kernel Print physical address info in a style consistent with the %pR style used elsewhere in the kernel. Commit 69f1d475cc did this for a similar printk in this file, but I must have missed this one. Signed-off-by: Bjorn Helgaas Signed-off-by: Rafael J. Wysocki --- kernel/power/snapshot.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c index 0de28576807d..7872a35eafe7 100644 --- a/kernel/power/snapshot.c +++ b/kernel/power/snapshot.c @@ -642,8 +642,9 @@ __register_nosave_region(unsigned long start_pfn, unsigned long end_pfn, region->end_pfn = end_pfn; list_add_tail(®ion->list, &nosave_regions); Report: - printk(KERN_INFO "PM: Registered nosave memory: %016lx - %016lx\n", - start_pfn << PAGE_SHIFT, end_pfn << PAGE_SHIFT); + printk(KERN_INFO "PM: Registered nosave memory: [mem %#010llx-%#010llx]\n", + (unsigned long long) start_pfn << PAGE_SHIFT, + ((unsigned long long) end_pfn << PAGE_SHIFT) - 1); } /* -- cgit v1.2.3 From 40b313608ad4ea655addd2ec6cdd106477ae8e15 Mon Sep 17 00:00:00 2001 From: Stephen Rothwell Date: Tue, 21 May 2013 13:49:35 +1000 Subject: Finally eradicate CONFIG_HOTPLUG Ever since commit 45f035ab9b8f ("CONFIG_HOTPLUG should be always on"), it has been basically impossible to build a kernel with CONFIG_HOTPLUG turned off. Remove all the remaining references to it. Cc: Russell King Cc: Doug Thompson Cc: Bjorn Helgaas Cc: Steven Whitehouse Cc: Arnd Bergmann Cc: Pavel Machek Cc: "Rafael J. Wysocki" Cc: Andrew Morton Signed-off-by: Stephen Rothwell Acked-by: Mauro Carvalho Chehab Acked-by: Hans Verkuil Signed-off-by: Greg Kroah-Hartman --- kernel/power/Kconfig | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/power/Kconfig b/kernel/power/Kconfig index 5dfdc9ea180b..9c39de095ba9 100644 --- a/kernel/power/Kconfig +++ b/kernel/power/Kconfig @@ -100,7 +100,6 @@ config PM_SLEEP_SMP depends on SMP depends on ARCH_SUSPEND_POSSIBLE || ARCH_HIBERNATION_POSSIBLE depends on PM_SLEEP - select HOTPLUG select HOTPLUG_CPU config PM_AUTOSLEEP -- cgit v1.2.3 From f12dc020149fad7087e119e54cffea668272bf7d Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 3 Jun 2013 19:13:02 -0700 Subject: cgroup: mark "tasks" cgroup file as insane Some resources controlled by cgroup aren't per-task and cgroup core allowing threads of a single thread_group to be in different cgroups forced memcg do explicitly find the group leader and use it. This is gonna be nasty when transitioning to unified hierarchy and in general we don't want and won't support granularity finer than processes. Mark "tasks" with CFTYPE_INSANE. Signed-off-by: Tejun Heo Acked-by: Li Zefan Cc: Johannes Weiner Cc: Michal Hocko Cc: Balbir Singh Cc: KAMEZAWA Hiroyuki Cc: cgroups@vger.kernel.org Cc: Vivek Goyal --- kernel/cgroup.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index fefc41c1a147..1e0f445b5b88 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4037,6 +4037,7 @@ static int cgroup_clone_children_write(struct cgroup *cgrp, static struct cftype files[] = { { .name = "tasks", + .flags = CFTYPE_INSANE, /* use "procs" instead */ .open = cgroup_tasks_open, .write_u64 = cgroup_tasks_write, .release = cgroup_pidlist_release, -- cgit v1.2.3 From cc5943a7816ba6c00639837a62131386619548dc Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 3 Jun 2013 19:13:55 -0700 Subject: cgroup: mark "notify_on_release" and "release_agent" cgroup files insane The empty cgroup notification mechanism currently implemented in cgroup is tragically outdated. Forking and execing userland process stopped being a viable notification mechanism more than a decade ago. We're gonna have a saner mechanism. Let's make it clear that this abomination is going away. Mark "notify_on_release" and "release_agent" with CFTYPE_INSANE. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 1e0f445b5b88..b3bb8a393642 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4052,6 +4052,7 @@ static struct cftype files[] = { }, { .name = "notify_on_release", + .flags = CFTYPE_INSANE, .read_u64 = cgroup_read_notify_on_release, .write_u64 = cgroup_write_notify_on_release, }, @@ -4073,7 +4074,7 @@ static struct cftype files[] = { }, { .name = "release_agent", - .flags = CFTYPE_ONLY_ON_ROOT, + .flags = CFTYPE_INSANE | CFTYPE_ONLY_ON_ROOT, .read_seq_string = cgroup_release_agent_show, .write_string = cgroup_release_agent_write, .max_write_len = PATH_MAX, -- cgit v1.2.3 From d5c56ced775f6bdc32b689b01c9c4f9b66e18610 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 3 Jun 2013 19:14:34 -0700 Subject: cgroup: clean up the cftype array for the base cgroup files * Rename it from files[] (really?) to cgroup_base_files[]. * Drop CGROUP_FILE_GENERIC_PREFIX which was defined as "cgroup." and used inconsistently. Just use "cgroup." directly. * Collect insane files at the end. Note that only the insane ones are missing "cgroup." prefix. This patch doesn't introduce any functional changes. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 47 ++++++++++++++++++++++++----------------------- 1 file changed, 24 insertions(+), 23 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index b3bb8a393642..bc53d5014b28 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4029,35 +4029,16 @@ static int cgroup_clone_children_write(struct cgroup *cgrp, return 0; } -/* - * for the common functions, 'private' gives the type of file - */ -/* for hysterical raisins, we can't put this on the older files */ -#define CGROUP_FILE_GENERIC_PREFIX "cgroup." -static struct cftype files[] = { +static struct cftype cgroup_base_files[] = { { - .name = "tasks", - .flags = CFTYPE_INSANE, /* use "procs" instead */ - .open = cgroup_tasks_open, - .write_u64 = cgroup_tasks_write, - .release = cgroup_pidlist_release, - .mode = S_IRUGO | S_IWUSR, - }, - { - .name = CGROUP_FILE_GENERIC_PREFIX "procs", + .name = "cgroup.procs", .open = cgroup_procs_open, .write_u64 = cgroup_procs_write, .release = cgroup_pidlist_release, .mode = S_IRUGO | S_IWUSR, }, { - .name = "notify_on_release", - .flags = CFTYPE_INSANE, - .read_u64 = cgroup_read_notify_on_release, - .write_u64 = cgroup_write_notify_on_release, - }, - { - .name = CGROUP_FILE_GENERIC_PREFIX "event_control", + .name = "cgroup.event_control", .write_string = cgroup_write_event_control, .mode = S_IWUGO, }, @@ -4072,6 +4053,26 @@ static struct cftype files[] = { .flags = CFTYPE_ONLY_ON_ROOT, .read_seq_string = cgroup_sane_behavior_show, }, + + /* + * Historical crazy stuff. These don't have "cgroup." prefix and + * don't exist if sane_behavior. If you're depending on these, be + * prepared to be burned. + */ + { + .name = "tasks", + .flags = CFTYPE_INSANE, /* use "procs" instead */ + .open = cgroup_tasks_open, + .write_u64 = cgroup_tasks_write, + .release = cgroup_pidlist_release, + .mode = S_IRUGO | S_IWUSR, + }, + { + .name = "notify_on_release", + .flags = CFTYPE_INSANE, + .read_u64 = cgroup_read_notify_on_release, + .write_u64 = cgroup_write_notify_on_release, + }, { .name = "release_agent", .flags = CFTYPE_INSANE | CFTYPE_ONLY_ON_ROOT, @@ -4095,7 +4096,7 @@ static int cgroup_populate_dir(struct cgroup *cgrp, bool base_files, struct cgroup_subsys *ss; if (base_files) { - err = cgroup_addrm_files(cgrp, NULL, files, true); + err = cgroup_addrm_files(cgrp, NULL, cgroup_base_files, true); if (err < 0) return err; } -- cgit v1.2.3 From 06d6b3cbdf94bc37732df83e7c25774370411a56 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 5 Jun 2013 17:15:11 +0800 Subject: cpuset: remove redundant check in cpuset_cpus_allowed_fallback() task_cs() will never return NULL. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cpuset.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 64b3f791bbe5..f0c884a0e574 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -2253,8 +2253,7 @@ void cpuset_cpus_allowed_fallback(struct task_struct *tsk) rcu_read_lock(); cs = task_cs(tsk); - if (cs) - do_set_cpus_allowed(tsk, cs->cpus_allowed); + do_set_cpus_allowed(tsk, cs->cpus_allowed); rcu_read_unlock(); /* -- cgit v1.2.3 From 40df2deb50570b288b7067b111af0aa9ca640e6f Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 5 Jun 2013 17:15:23 +0800 Subject: cpuset: cleanup guarantee_online_{cpus|mems}() - We never pass a NULL @cs to these functions. - The top cpuset always has some online cpus/mems. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cpuset.c | 29 +++++++---------------------- 1 file changed, 7 insertions(+), 22 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index f0c884a0e574..d753837cca33 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -304,53 +304,38 @@ static struct file_system_type cpuset_fs_type = { /* * Return in pmask the portion of a cpusets's cpus_allowed that * are online. If none are online, walk up the cpuset hierarchy - * until we find one that does have some online cpus. If we get - * all the way to the top and still haven't found any online cpus, - * return cpu_online_mask. Or if passed a NULL cs from an exit'ing - * task, return cpu_online_mask. + * until we find one that does have some online cpus. The top + * cpuset always has some cpus online. * * One way or another, we guarantee to return some non-empty subset * of cpu_online_mask. * * Call with callback_mutex held. */ - static void guarantee_online_cpus(const struct cpuset *cs, struct cpumask *pmask) { - while (cs && !cpumask_intersects(cs->cpus_allowed, cpu_online_mask)) + while (!cpumask_intersects(cs->cpus_allowed, cpu_online_mask)) cs = parent_cs(cs); - if (cs) - cpumask_and(pmask, cs->cpus_allowed, cpu_online_mask); - else - cpumask_copy(pmask, cpu_online_mask); - BUG_ON(!cpumask_intersects(pmask, cpu_online_mask)); + cpumask_and(pmask, cs->cpus_allowed, cpu_online_mask); } /* * Return in *pmask the portion of a cpusets's mems_allowed that * are online, with memory. If none are online with memory, walk * up the cpuset hierarchy until we find one that does have some - * online mems. If we get all the way to the top and still haven't - * found any online mems, return node_states[N_MEMORY]. + * online mems. The top cpuset always has some mems online. * * One way or another, we guarantee to return some non-empty subset * of node_states[N_MEMORY]. * * Call with callback_mutex held. */ - static void guarantee_online_mems(const struct cpuset *cs, nodemask_t *pmask) { - while (cs && !nodes_intersects(cs->mems_allowed, - node_states[N_MEMORY])) + while (!nodes_intersects(cs->mems_allowed, node_states[N_MEMORY])) cs = parent_cs(cs); - if (cs) - nodes_and(*pmask, cs->mems_allowed, - node_states[N_MEMORY]); - else - *pmask = node_states[N_MEMORY]; - BUG_ON(!nodes_intersects(*pmask, node_states[N_MEMORY])); + nodes_and(*pmask, cs->mems_allowed, node_states[N_MEMORY]); } /* -- cgit v1.2.3 From 67bd2c59850de20d0ecdc8084cbbfe34e53b6804 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 5 Jun 2013 17:15:35 +0800 Subject: cpuset: remove unnecessary variable in cpuset_attach() We can just use oldcs->mems_allowed. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cpuset.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index d753837cca33..dbef832e5e2d 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -1407,8 +1407,7 @@ static cpumask_var_t cpus_attach; static void cpuset_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) { - /* static bufs protected by cpuset_mutex */ - static nodemask_t cpuset_attach_nodemask_from; + /* static buf protected by cpuset_mutex */ static nodemask_t cpuset_attach_nodemask_to; struct mm_struct *mm; struct task_struct *task; @@ -1442,13 +1441,12 @@ static void cpuset_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) * Change mm, possibly for multiple threads in a threadgroup. This is * expensive and may sleep. */ - cpuset_attach_nodemask_from = oldcs->mems_allowed; cpuset_attach_nodemask_to = cs->mems_allowed; mm = get_task_mm(leader); if (mm) { mpol_rebind_mm(mm, &cpuset_attach_nodemask_to); if (is_memory_migrate(cs)) - cpuset_migrate_mm(mm, &cpuset_attach_nodemask_from, + cpuset_migrate_mm(mm, &oldcs->mems_allowed, &cpuset_attach_nodemask_to); mmput(mm); } -- cgit v1.2.3 From 249cc86db7492dc8de1d2eddebc6bcc4ab2a8e9e Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 5 Jun 2013 17:15:48 +0800 Subject: cpuset: remove cpuset_test_cpumask() The test is done in set_cpus_allowed_ptr(), so it's redundant. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cpuset.c | 19 +------------------ 1 file changed, 1 insertion(+), 18 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index dbef832e5e2d..51f8e1d5a2a9 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -783,23 +783,6 @@ void rebuild_sched_domains(void) mutex_unlock(&cpuset_mutex); } -/** - * cpuset_test_cpumask - test a task's cpus_allowed versus its cpuset's - * @tsk: task to test - * @scan: struct cgroup_scanner contained in its struct cpuset_hotplug_scanner - * - * Call with cpuset_mutex held. May take callback_mutex during call. - * Called for each task in a cgroup by cgroup_scan_tasks(). - * Return nonzero if this tasks's cpus_allowed mask should be changed (in other - * words, if its mask is not equal to its cpuset's mask). - */ -static int cpuset_test_cpumask(struct task_struct *tsk, - struct cgroup_scanner *scan) -{ - return !cpumask_equal(&tsk->cpus_allowed, - (cgroup_cs(scan->cg))->cpus_allowed); -} - /** * cpuset_change_cpumask - make a task's cpus_allowed the same as its cpuset's * @tsk: task to test @@ -835,7 +818,7 @@ static void update_tasks_cpumask(struct cpuset *cs, struct ptr_heap *heap) struct cgroup_scanner scan; scan.cg = cs->css.cgroup; - scan.test_task = cpuset_test_cpumask; + scan.test_task = NULL; scan.process_task = cpuset_change_cpumask; scan.heap = heap; cgroup_scan_tasks(&scan); -- cgit v1.2.3 From a73456f37b9dbc917398387d0cba926b4455b70f Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 5 Jun 2013 17:15:59 +0800 Subject: cpuset: re-structure update_cpumask() a bit Check if cpus_allowed is to be changed before calling validate_change(). This won't change any behavior, but later it will allow us to do this: # mkdir /cpuset/child # echo $$ > /cpuset/child/tasks /* empty cpuset */ # echo > /cpuset/child/cpuset.cpus /* do nothing, won't fail */ Without this patch, the last operation will fail. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cpuset.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 51f8e1d5a2a9..535dce685eec 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -856,14 +856,15 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs, if (!cpumask_subset(trialcs->cpus_allowed, cpu_active_mask)) return -EINVAL; } - retval = validate_change(cs, trialcs); - if (retval < 0) - return retval; /* Nothing to do if the cpus didn't change */ if (cpumask_equal(cs->cpus_allowed, trialcs->cpus_allowed)) return 0; + retval = validate_change(cs, trialcs); + if (retval < 0) + return retval; + retval = heap_init(&heap, PAGE_SIZE, GFP_KERNEL, NULL); if (retval) return retval; -- cgit v1.2.3 From c5a130325f13b219438cb100e2da71a3e31199f3 Mon Sep 17 00:00:00 2001 From: Chen Gong Date: Thu, 6 Jun 2013 15:20:51 -0700 Subject: ACPI/APEI: Add parameter check before error injection When param1 is enabled in EINJ but not assigned with a valid value, sometimes it will cause the error like below: APEI: Can not request [mem 0x7aaa7000-0x7aaa7007] for APEI EINJ Trigger registers It is because some firmware will access target address specified in param1 to trigger the error when injecting memory error. This will cause resource conflict with regular memory. So It must be removed from trigger table resources, but incorrect param1/param2 combination will stop this action. Add extra check to avoid this kind of error. Signed-off-by: Chen Gong Signed-off-by: Tony Luck --- kernel/resource.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/resource.c b/kernel/resource.c index d7386986e10e..77bf11a86c7d 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -409,6 +409,7 @@ int __weak page_is_ram(unsigned long pfn) { return walk_system_ram_range(pfn, 1, NULL, __is_ram) == 1; } +EXPORT_SYMBOL_GPL(page_is_ram); void __weak arch_remove_reservations(struct resource *avail) { -- cgit v1.2.3 From e44193d39e8d4d1de5d996fcd37ed75e5c704f10 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Sun, 9 Jun 2013 17:14:22 +0800 Subject: cpuset: let hotplug propagation work wait for task attaching Instead of triggering propagation work in cpuset_attach(), we make hotplug propagation work wait until there's no task attaching in progress. IMO this is more robust. We won't see empty masks in cpuset_attach(). Also it's a preparation for removing propagation work. Without asynchronous propagation we can't call move_tasks_in_empty_cpuset() in cpuset_attach(), because otherwise we'll deadlock on cgroup_mutex. tj: typo fixes. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cpuset.c | 29 +++++++++++++++++------------ 1 file changed, 17 insertions(+), 12 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 535dce685eec..e902473f76bf 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -59,6 +59,7 @@ #include #include #include +#include /* * Tracks how many cpusets are currently defined in system. @@ -275,6 +276,8 @@ static void schedule_cpuset_propagate_hotplug(struct cpuset *cs); static DECLARE_WORK(cpuset_hotplug_work, cpuset_hotplug_workfn); +static DECLARE_WAIT_QUEUE_HEAD(cpuset_attach_wq); + /* * This is ugly, but preserves the userspace API for existing cpuset * users. If someone tries to mount the "cpuset" filesystem, we @@ -1436,14 +1439,8 @@ static void cpuset_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) } cs->attach_in_progress--; - - /* - * We may have raced with CPU/memory hotunplug. Trigger hotplug - * propagation if @cs doesn't have any CPU or memory. It will move - * the newly added tasks to the nearest parent which can execute. - */ - if (cpumask_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed)) - schedule_cpuset_propagate_hotplug(cs); + if (!cs->attach_in_progress) + wake_up(&cpuset_attach_wq); mutex_unlock(&cpuset_mutex); } @@ -1555,10 +1552,6 @@ static int cpuset_write_resmask(struct cgroup *cgrp, struct cftype *cft, * resources, wait for the previously scheduled operations before * proceeding, so that we don't end up keep removing tasks added * after execution capability is restored. - * - * Flushing cpuset_hotplug_work is enough to synchronize against - * hotplug hanlding; however, cpuset_attach() may schedule - * propagation work directly. Flush the workqueue too. */ flush_work(&cpuset_hotplug_work); flush_workqueue(cpuset_propagate_hotplug_wq); @@ -2005,8 +1998,20 @@ static void cpuset_propagate_hotplug_workfn(struct work_struct *work) struct cpuset *cs = container_of(work, struct cpuset, hotplug_work); bool is_empty; +retry: + wait_event(cpuset_attach_wq, cs->attach_in_progress == 0); + mutex_lock(&cpuset_mutex); + /* + * We have raced with task attaching. We wait until attaching + * is finished, so we won't attach a task to an empty cpuset. + */ + if (cs->attach_in_progress) { + mutex_unlock(&cpuset_mutex); + goto retry; + } + cpumask_andnot(&off_cpus, cs->cpus_allowed, top_cpuset.cpus_allowed); nodes_andnot(off_mems, cs->mems_allowed, top_cpuset.mems_allowed); -- cgit v1.2.3 From 388afd8549dc8be0920e00ae9404341593b6bd7c Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Sun, 9 Jun 2013 17:14:47 +0800 Subject: cpuset: remove async hotplug propagation work As we can drop rcu read lock while iterating cgroup hierarchy, we don't have to do propagation asynchronously via workqueue. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cpuset.c | 69 +++++++++++++-------------------------------------------- 1 file changed, 16 insertions(+), 53 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index e902473f76bf..608fe1308b22 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -101,8 +101,6 @@ struct cpuset { /* for custom sched domain */ int relax_domain_level; - - struct work_struct hotplug_work; }; /* Retrieve the cpuset for a cgroup */ @@ -268,12 +266,7 @@ static DEFINE_MUTEX(callback_mutex); /* * CPU / memory hotplug is handled asynchronously. */ -static struct workqueue_struct *cpuset_propagate_hotplug_wq; - static void cpuset_hotplug_workfn(struct work_struct *work); -static void cpuset_propagate_hotplug_workfn(struct work_struct *work); -static void schedule_cpuset_propagate_hotplug(struct cpuset *cs); - static DECLARE_WORK(cpuset_hotplug_work, cpuset_hotplug_workfn); static DECLARE_WAIT_QUEUE_HEAD(cpuset_attach_wq); @@ -1554,7 +1547,6 @@ static int cpuset_write_resmask(struct cgroup *cgrp, struct cftype *cft, * after execution capability is restored. */ flush_work(&cpuset_hotplug_work); - flush_workqueue(cpuset_propagate_hotplug_wq); mutex_lock(&cpuset_mutex); if (!is_cpuset_online(cs)) @@ -1821,7 +1813,6 @@ static struct cgroup_subsys_state *cpuset_css_alloc(struct cgroup *cont) cpumask_clear(cs->cpus_allowed); nodes_clear(cs->mems_allowed); fmeter_init(&cs->fmeter); - INIT_WORK(&cs->hotplug_work, cpuset_propagate_hotplug_workfn); cs->relax_domain_level = -1; return &cs->css; @@ -1984,18 +1975,17 @@ static void remove_tasks_in_empty_cpuset(struct cpuset *cs) } /** - * cpuset_propagate_hotplug_workfn - propagate CPU/memory hotplug to a cpuset + * cpuset_hotplug_update_tasks - update tasks in a cpuset for hotunplug * @cs: cpuset in interest * * Compare @cs's cpu and mem masks against top_cpuset and if some have gone * offline, update @cs accordingly. If @cs ends up with no CPU or memory, * all its tasks are moved to the nearest ancestor with both resources. */ -static void cpuset_propagate_hotplug_workfn(struct work_struct *work) +static void cpuset_hotplug_update_tasks(struct cpuset *cs) { static cpumask_t off_cpus; static nodemask_t off_mems, tmp_mems; - struct cpuset *cs = container_of(work, struct cpuset, hotplug_work); bool is_empty; retry: @@ -2044,34 +2034,6 @@ retry: */ if (is_empty) remove_tasks_in_empty_cpuset(cs); - - /* the following may free @cs, should be the last operation */ - css_put(&cs->css); -} - -/** - * schedule_cpuset_propagate_hotplug - schedule hotplug propagation to a cpuset - * @cs: cpuset of interest - * - * Schedule cpuset_propagate_hotplug_workfn() which will update CPU and - * memory masks according to top_cpuset. - */ -static void schedule_cpuset_propagate_hotplug(struct cpuset *cs) -{ - /* - * Pin @cs. The refcnt will be released when the work item - * finishes executing. - */ - if (!css_tryget(&cs->css)) - return; - - /* - * Queue @cs->hotplug_work. If already pending, lose the css ref. - * cpuset_propagate_hotplug_wq is ordered and propagation will - * happen in the order this function is called. - */ - if (!queue_work(cpuset_propagate_hotplug_wq, &cs->hotplug_work)) - css_put(&cs->css); } /** @@ -2084,8 +2046,8 @@ static void schedule_cpuset_propagate_hotplug(struct cpuset *cs) * actively using CPU hotplug but making no active use of cpusets. * * Non-root cpusets are only affected by offlining. If any CPUs or memory - * nodes have been taken down, cpuset_propagate_hotplug() is invoked on all - * descendants. + * nodes have been taken down, cpuset_hotplug_update_tasks() is invoked on + * all descendants. * * Note that CPU offlining during suspend is ignored. We don't modify * cpusets across suspend/resume cycles at all. @@ -2128,21 +2090,26 @@ static void cpuset_hotplug_workfn(struct work_struct *work) update_tasks_nodemask(&top_cpuset, &tmp_mems, NULL); } + mutex_unlock(&cpuset_mutex); + /* if cpus or mems went down, we need to propagate to descendants */ if (cpus_offlined || mems_offlined) { struct cpuset *cs; struct cgroup *pos_cgrp; rcu_read_lock(); - cpuset_for_each_descendant_pre(cs, pos_cgrp, &top_cpuset) - schedule_cpuset_propagate_hotplug(cs); - rcu_read_unlock(); - } + cpuset_for_each_descendant_pre(cs, pos_cgrp, &top_cpuset) { + if (!css_tryget(&cs->css)) + continue; + rcu_read_unlock(); - mutex_unlock(&cpuset_mutex); + cpuset_hotplug_update_tasks(cs); - /* wait for propagations to finish */ - flush_workqueue(cpuset_propagate_hotplug_wq); + rcu_read_lock(); + css_put(&cs->css); + } + rcu_read_unlock(); + } /* rebuild sched domains if cpus_allowed has changed */ if (cpus_updated) @@ -2193,10 +2160,6 @@ void __init cpuset_init_smp(void) top_cpuset.mems_allowed = node_states[N_MEMORY]; register_hotmemory_notifier(&cpuset_track_online_nodes_nb); - - cpuset_propagate_hotplug_wq = - alloc_ordered_workqueue("cpuset_hotplug", 0); - BUG_ON(!cpuset_propagate_hotplug_wq); } /** -- cgit v1.2.3 From 5e1cda5b8ae93f5f02e8c5a30390ac9b4d2c20e6 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Wed, 29 May 2013 03:10:53 +0100 Subject: irqdomain: Relax failure path on setting up mappings Commit 98aa468e, "irqdomain: Support for static IRQ mapping and association" introduced an API for directly associating blocks of hwirqs to linux irqs. However, if any irq in that block failed to map (say if the mapping functions returns an error because the irq is already mapped) then the whole thing will fail and roll back. This is probably too aggressive since there are valid reasons why a mapping may fail. ie. Firmware may have a particular IRQ marked as unusable. This patch drops the error path out of irq_domain_associate(). If a mapping fails, then it is simply skipped. There is no reason to fail the entire allocation. v2: Still output an information message on failed mappings and make sure attempted mapping gets cleared out of the irq_data structure. Signed-off-by: Grant Likely Cc: Paul Mundt Cc: Benjamin Herrenschmidt Cc: Thomas Gleixner --- kernel/irq/irqdomain.c | 16 ++++------------ 1 file changed, 4 insertions(+), 12 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c index 20b677dd0b27..61d6d3c80fee 100644 --- a/kernel/irq/irqdomain.c +++ b/kernel/irq/irqdomain.c @@ -464,23 +464,15 @@ int irq_domain_associate_many(struct irq_domain *domain, unsigned int irq_base, /* * If map() returns -EPERM, this interrupt is protected * by the firmware or some other service and shall not - * be mapped. - * - * Since on some platforms we blindly try to map everything - * we end up with a log full of backtraces. - * - * So instead, we silently fail on -EPERM, it is the - * responsibility of the PIC driver to display a relevant - * message if needed. + * be mapped. Don't bother telling the user about it. */ if (ret != -EPERM) { - pr_err("irq-%i==>hwirq-0x%lx mapping failed: %d\n", - virq, hwirq, ret); - WARN_ON(1); + pr_info("%s didn't like hwirq-0x%lx to VIRQ%i mapping (rc=%d)\n", + of_node_full_name(domain->of_node), hwirq, virq, ret); } irq_data->domain = NULL; irq_data->hwirq = 0; - goto err_unmap; + continue; } } -- cgit v1.2.3 From 9bbf877d3b6b8c5991000296f40a3f0fe66fa89b Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Thu, 6 Jun 2013 12:10:24 +0100 Subject: irqdomain: Replace LEGACY mapping with LINEAR The LEGACY mapping unnecessarily complicates the irqdomain code and can easily be implemented with a linear mapping. By ripping it out and replacing it with the LINEAR mapping the object size of irqdomain.c shrinks by about 330 bytes (ARMv7) which offsets the additional allocation required by the linear map. It also makes it possible for current LEGACY map users to pre-allocate irq_descs for a subset of the hwirqs and dynamically allocate the rest as needed. Signed-off-by: Grant Likely Cc: Paul Mundt Cc: Benjamin Herrenschmidt Cc: Thomas Gleixner Cc: Rob Herring --- kernel/irq/irqdomain.c | 84 ++++---------------------------------------------- 1 file changed, 6 insertions(+), 78 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c index 61d6d3c80fee..1ac8cf41b9a5 100644 --- a/kernel/irq/irqdomain.c +++ b/kernel/irq/irqdomain.c @@ -82,13 +82,6 @@ void irq_domain_remove(struct irq_domain *domain) mutex_lock(&irq_domain_mutex); switch (domain->revmap_type) { - case IRQ_DOMAIN_MAP_LEGACY: - /* - * Legacy domains don't manage their own irq_desc - * allocations, we expect the caller to handle irq_desc - * freeing on their own. - */ - break; case IRQ_DOMAIN_MAP_TREE: /* * radix_tree_delete() takes care of destroying the root @@ -122,17 +115,6 @@ void irq_domain_remove(struct irq_domain *domain) } EXPORT_SYMBOL_GPL(irq_domain_remove); -static unsigned int irq_domain_legacy_revmap(struct irq_domain *domain, - irq_hw_number_t hwirq) -{ - irq_hw_number_t first_hwirq = domain->revmap_data.legacy.first_hwirq; - int size = domain->revmap_data.legacy.size; - - if (WARN_ON(hwirq < first_hwirq || hwirq >= first_hwirq + size)) - return 0; - return hwirq - first_hwirq + domain->revmap_data.legacy.first_irq; -} - /** * irq_domain_add_simple() - Allocate and register a simple irq_domain. * @of_node: pointer to interrupt controller's device tree node. @@ -213,57 +195,17 @@ struct irq_domain *irq_domain_add_legacy(struct device_node *of_node, void *host_data) { struct irq_domain *domain; - unsigned int i; - domain = irq_domain_alloc(of_node, IRQ_DOMAIN_MAP_LEGACY, ops, host_data); + pr_debug("Setting up legacy domain virq[%i:%i] ==> hwirq[%i:%i]\n", + first_irq, first_irq + size - 1, + (int)first_hwirq, (int)first_hwirq + size -1); + + domain = irq_domain_add_linear(of_node, first_hwirq + size, ops, host_data); if (!domain) return NULL; - domain->revmap_data.legacy.first_irq = first_irq; - domain->revmap_data.legacy.first_hwirq = first_hwirq; - domain->revmap_data.legacy.size = size; - - mutex_lock(&irq_domain_mutex); - /* Verify that all the irqs are available */ - for (i = 0; i < size; i++) { - int irq = first_irq + i; - struct irq_data *irq_data = irq_get_irq_data(irq); - - if (WARN_ON(!irq_data || irq_data->domain)) { - mutex_unlock(&irq_domain_mutex); - irq_domain_free(domain); - return NULL; - } - } + WARN_ON(irq_domain_associate_many(domain, first_irq, first_hwirq, size)); - /* Claim all of the irqs before registering a legacy domain */ - for (i = 0; i < size; i++) { - struct irq_data *irq_data = irq_get_irq_data(first_irq + i); - irq_data->hwirq = first_hwirq + i; - irq_data->domain = domain; - } - mutex_unlock(&irq_domain_mutex); - - for (i = 0; i < size; i++) { - int irq = first_irq + i; - int hwirq = first_hwirq + i; - - /* IRQ0 gets ignored */ - if (!irq) - continue; - - /* Legacy flags are left to default at this point, - * one can then use irq_create_mapping() to - * explicitly change them - */ - if (ops->map) - ops->map(domain, irq, hwirq); - - /* Clear norequest flags */ - irq_clear_status_flags(irq, IRQ_NOREQUEST); - } - - irq_domain_add(domain); return domain; } EXPORT_SYMBOL_GPL(irq_domain_add_legacy); @@ -492,10 +434,6 @@ int irq_domain_associate_many(struct irq_domain *domain, unsigned int irq_base, } return 0; - - err_unmap: - irq_domain_disassociate_many(domain, irq_base, i); - return -EINVAL; } EXPORT_SYMBOL_GPL(irq_domain_associate_many); @@ -575,10 +513,6 @@ unsigned int irq_create_mapping(struct irq_domain *domain, return virq; } - /* Get a virtual interrupt number */ - if (domain->revmap_type == IRQ_DOMAIN_MAP_LEGACY) - return irq_domain_legacy_revmap(domain, hwirq); - /* Allocate a virtual interrupt number */ hint = hwirq % nr_irqs; if (hint == 0) @@ -706,10 +640,6 @@ void irq_dispose_mapping(unsigned int virq) if (WARN_ON(domain == NULL)) return; - /* Never unmap legacy interrupts */ - if (domain->revmap_type == IRQ_DOMAIN_MAP_LEGACY) - return; - irq_domain_disassociate_many(domain, virq, 1); irq_free_desc(virq); } @@ -732,8 +662,6 @@ unsigned int irq_find_mapping(struct irq_domain *domain, return 0; switch (domain->revmap_type) { - case IRQ_DOMAIN_MAP_LEGACY: - return irq_domain_legacy_revmap(domain, hwirq); case IRQ_DOMAIN_MAP_LINEAR: return irq_linear_revmap(domain, hwirq); case IRQ_DOMAIN_MAP_TREE: -- cgit v1.2.3 From 0bb4afb45dd1add73ca643a865daa38716aeff0c Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Thu, 6 Jun 2013 14:23:30 +0100 Subject: irqdomain: Add a name field This patch adds a name field to the irq_domain structure to help mere mortals understand the mappings between irq domains and virqs. It also converts a number of places that have open-coded some kind of fudging an irqdomain name to use the new field. This means a more consistent display of names in irq domain log messages and debugfs output. Signed-off-by: Grant Likely --- kernel/irq/generic-chip.c | 1 + kernel/irq/irqdomain.c | 19 ++++++------------- 2 files changed, 7 insertions(+), 13 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/generic-chip.c b/kernel/irq/generic-chip.c index 95575d8d5392..ca98cc5d6308 100644 --- a/kernel/irq/generic-chip.c +++ b/kernel/irq/generic-chip.c @@ -305,6 +305,7 @@ int irq_alloc_domain_generic_chips(struct irq_domain *d, int irqs_per_chip, /* Calc pointer to the next generic chip */ tmp += sizeof(*gc) + num_ct * sizeof(struct irq_chip_type); } + d->name = name; return 0; } EXPORT_SYMBOL_GPL(irq_alloc_domain_generic_chips); diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c index 1ac8cf41b9a5..b1b5e6793fd2 100644 --- a/kernel/irq/irqdomain.c +++ b/kernel/irq/irqdomain.c @@ -410,12 +410,15 @@ int irq_domain_associate_many(struct irq_domain *domain, unsigned int irq_base, */ if (ret != -EPERM) { pr_info("%s didn't like hwirq-0x%lx to VIRQ%i mapping (rc=%d)\n", - of_node_full_name(domain->of_node), hwirq, virq, ret); + domain->name, hwirq, virq, ret); } irq_data->domain = NULL; irq_data->hwirq = 0; continue; } + /* If not already assigned, give the domain the chip's name */ + if (!domain->name && irq_data->chip) + domain->name = irq_data->chip->name; } switch (domain->revmap_type) { @@ -708,8 +711,6 @@ static int virq_debug_show(struct seq_file *m, void *private) { unsigned long flags; struct irq_desc *desc; - const char *p; - static const char none[] = "none"; void *data; int i; @@ -731,20 +732,12 @@ static int virq_debug_show(struct seq_file *m, void *private) seq_printf(m, "0x%05lx ", desc->irq_data.hwirq); chip = irq_desc_get_chip(desc); - if (chip && chip->name) - p = chip->name; - else - p = none; - seq_printf(m, "%-15s ", p); + seq_printf(m, "%-15s ", (chip && chip->name) ? chip->name : "none"); data = irq_desc_get_chip_data(desc); seq_printf(m, data ? "0x%p " : " %p ", data); - if (desc->irq_data.domain) - p = of_node_full_name(desc->irq_data.domain->of_node); - else - p = none; - seq_printf(m, "%s\n", p); + seq_printf(m, "%s\n", desc->irq_data.domain->name); } raw_spin_unlock_irqrestore(&desc->lock, flags); -- cgit v1.2.3 From cef5075c8c238ffd04c86a77a5a9bdbd18031137 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Wed, 11 Jul 2012 17:24:31 +0100 Subject: irqdomain: merge linear and tree reverse mappings. Keeping them separate makes irq_domain more complex and adds a lot of code (as proven by the diffstat). Merging them simplifies the whole scheme. This change makes it so both the tree and linear methods can be used by the same irq_domain instance. If the hwirq is less than the ->linear_size, then the linear map is used to reverse map the hwirq. Otherwise the radix tree is used. The test for which map to use is no more expensive that the existing code, so the performance of fast path is preserved. It also means that complex interrupt controllers can use both the linear map and a tree in the same domain. This may be useful for an interrupt controller with a base set of core irqs and a large number of GPIOs which might be used as irqs. The linear map could cover the core irqs, and the tree used for thas irqs. The linear map could cover the core irqs, and the tree used for the gpios. v2: Drop reorganization of revmap data Signed-off-by: Grant Likely Cc: Paul Mundt Cc: Benjamin Herrenschmidt Cc: Thomas Gleixner Cc: Rob Herring --- kernel/irq/irqdomain.c | 107 ++++++++++++++----------------------------------- 1 file changed, 29 insertions(+), 78 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c index b1b5e6793fd2..5a1d8ec8509e 100644 --- a/kernel/irq/irqdomain.c +++ b/kernel/irq/irqdomain.c @@ -34,22 +34,24 @@ static struct irq_domain *irq_default_domain; * to IRQ domain, or NULL on failure. */ static struct irq_domain *irq_domain_alloc(struct device_node *of_node, - unsigned int revmap_type, + unsigned int revmap_type, int size, const struct irq_domain_ops *ops, void *host_data) { struct irq_domain *domain; - domain = kzalloc_node(sizeof(*domain), GFP_KERNEL, - of_node_to_nid(of_node)); + domain = kzalloc_node(sizeof(*domain) + (sizeof(unsigned int) * size), + GFP_KERNEL, of_node_to_nid(of_node)); if (WARN_ON(!domain)) return NULL; /* Fill structure */ + INIT_RADIX_TREE(&domain->revmap_data.tree, GFP_KERNEL); domain->revmap_type = revmap_type; domain->ops = ops; domain->host_data = host_data; domain->of_node = of_node_get(of_node); + domain->revmap_data.linear.size = size; return domain; } @@ -81,22 +83,12 @@ void irq_domain_remove(struct irq_domain *domain) { mutex_lock(&irq_domain_mutex); - switch (domain->revmap_type) { - case IRQ_DOMAIN_MAP_TREE: - /* - * radix_tree_delete() takes care of destroying the root - * node when all entries are removed. Shout if there are - * any mappings left. - */ - WARN_ON(domain->revmap_data.tree.height); - break; - case IRQ_DOMAIN_MAP_LINEAR: - kfree(domain->revmap_data.linear.revmap); - domain->revmap_data.linear.size = 0; - break; - case IRQ_DOMAIN_MAP_NOMAP: - break; - } + /* + * radix_tree_delete() takes care of destroying the root + * node when all entries are removed. Shout if there are + * any mappings left. + */ + WARN_ON(domain->revmap_data.tree.height); list_del(&domain->link); @@ -223,20 +215,11 @@ struct irq_domain *irq_domain_add_linear(struct device_node *of_node, void *host_data) { struct irq_domain *domain; - unsigned int *revmap; - revmap = kzalloc_node(sizeof(*revmap) * size, GFP_KERNEL, - of_node_to_nid(of_node)); - if (WARN_ON(!revmap)) + domain = irq_domain_alloc(of_node, IRQ_DOMAIN_MAP_LINEAR, size, ops, host_data); + if (!domain) return NULL; - domain = irq_domain_alloc(of_node, IRQ_DOMAIN_MAP_LINEAR, ops, host_data); - if (!domain) { - kfree(revmap); - return NULL; - } - domain->revmap_data.linear.size = size; - domain->revmap_data.linear.revmap = revmap; irq_domain_add(domain); return domain; } @@ -248,7 +231,7 @@ struct irq_domain *irq_domain_add_nomap(struct device_node *of_node, void *host_data) { struct irq_domain *domain = irq_domain_alloc(of_node, - IRQ_DOMAIN_MAP_NOMAP, ops, host_data); + IRQ_DOMAIN_MAP_NOMAP, 0, ops, host_data); if (domain) { domain->revmap_data.nomap.max_irq = max_irq ? max_irq : ~0; irq_domain_add(domain); @@ -257,28 +240,6 @@ struct irq_domain *irq_domain_add_nomap(struct device_node *of_node, } EXPORT_SYMBOL_GPL(irq_domain_add_nomap); -/** - * irq_domain_add_tree() - * @of_node: pointer to interrupt controller's device tree node. - * @ops: map/unmap domain callbacks - * - * Note: The radix tree will be allocated later during boot automatically - * (the reverse mapping will use the slow path until that happens). - */ -struct irq_domain *irq_domain_add_tree(struct device_node *of_node, - const struct irq_domain_ops *ops, - void *host_data) -{ - struct irq_domain *domain = irq_domain_alloc(of_node, - IRQ_DOMAIN_MAP_TREE, ops, host_data); - if (domain) { - INIT_RADIX_TREE(&domain->revmap_data.tree, GFP_KERNEL); - irq_domain_add(domain); - } - return domain; -} -EXPORT_SYMBOL_GPL(irq_domain_add_tree); - /** * irq_find_host() - Locates a domain for a given device node * @node: device-tree node of the interrupt controller @@ -359,17 +320,13 @@ static void irq_domain_disassociate_many(struct irq_domain *domain, irq_data->domain = NULL; irq_data->hwirq = 0; - /* Clear reverse map */ - switch(domain->revmap_type) { - case IRQ_DOMAIN_MAP_LINEAR: - if (hwirq < domain->revmap_data.linear.size) - domain->revmap_data.linear.revmap[hwirq] = 0; - break; - case IRQ_DOMAIN_MAP_TREE: + /* Clear reverse map for this hwirq */ + if (hwirq < domain->revmap_data.linear.size) { + domain->linear_revmap[hwirq] = 0; + } else { mutex_lock(&revmap_trees_mutex); radix_tree_delete(&domain->revmap_data.tree, hwirq); mutex_unlock(&revmap_trees_mutex); - break; } } } @@ -421,16 +378,12 @@ int irq_domain_associate_many(struct irq_domain *domain, unsigned int irq_base, domain->name = irq_data->chip->name; } - switch (domain->revmap_type) { - case IRQ_DOMAIN_MAP_LINEAR: - if (hwirq < domain->revmap_data.linear.size) - domain->revmap_data.linear.revmap[hwirq] = virq; - break; - case IRQ_DOMAIN_MAP_TREE: + if (hwirq < domain->revmap_data.linear.size) { + domain->linear_revmap[hwirq] = virq; + } else { mutex_lock(&revmap_trees_mutex); radix_tree_insert(&domain->revmap_data.tree, hwirq, irq_data); mutex_unlock(&revmap_trees_mutex); - break; } irq_clear_status_flags(virq, IRQ_NOREQUEST); @@ -667,13 +620,6 @@ unsigned int irq_find_mapping(struct irq_domain *domain, switch (domain->revmap_type) { case IRQ_DOMAIN_MAP_LINEAR: return irq_linear_revmap(domain, hwirq); - case IRQ_DOMAIN_MAP_TREE: - rcu_read_lock(); - data = radix_tree_lookup(&domain->revmap_data.tree, hwirq); - rcu_read_unlock(); - if (data) - return data->irq; - break; case IRQ_DOMAIN_MAP_NOMAP: data = irq_get_irq_data(hwirq); if (data && (data->domain == domain) && (data->hwirq == hwirq)) @@ -696,13 +642,18 @@ EXPORT_SYMBOL_GPL(irq_find_mapping); unsigned int irq_linear_revmap(struct irq_domain *domain, irq_hw_number_t hwirq) { + struct irq_data *data; BUG_ON(domain->revmap_type != IRQ_DOMAIN_MAP_LINEAR); /* Check revmap bounds; complain if exceeded */ - if (WARN_ON(hwirq >= domain->revmap_data.linear.size)) - return 0; + if (hwirq >= domain->revmap_data.linear.size) { + rcu_read_lock(); + data = radix_tree_lookup(&domain->revmap_data.tree, hwirq); + rcu_read_unlock(); + return data ? data->irq : 0; + } - return domain->revmap_data.linear.revmap[hwirq]; + return domain->linear_revmap[hwirq]; } EXPORT_SYMBOL_GPL(irq_linear_revmap); -- cgit v1.2.3 From 1aa0dd94ca07df818cf14588c9031ab1d7fd84d3 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Sat, 8 Jun 2013 12:03:59 +0100 Subject: irqdomain: Eliminate revmap type The NOMAP irq_domain type is only used by a handful of interrupt controllers and it unnecessarily complicates the code by adding special cases on how to look up mappings and different revmap functions are used for each type which need to validate the correct type is passed to it before performing the reverse map. Eliminating the revmap_type and making a single reverse mapping function simplifies the code. It also shouldn't be any slower than having separate revmap functions because the type of the revmap needed to be checked anyway. The linear and tree revmap types were already merged in a previous patch. This patch rolls the NOMAP or direct mapping behaviour into the same domain code making is possible for an irq domain to do any mapping type; linear, tree or direct; and that the mapping will be transparent to the interrupt controller driver. With this change, direct mappings will get stored in the linear or tree mapping for consistency. Reverse mapping from the hwirq to virq will go through the normal lookup process. However, any controller using a direct mapping can take advantage of knowing that hwirq==virq for any mapped interrupts skip doing a revmap lookup when handling IRQs. Signed-off-by: Grant Likely --- kernel/irq/generic-chip.c | 5 +---- kernel/irq/irqdomain.c | 55 +++++++++++++++++++---------------------------- 2 files changed, 23 insertions(+), 37 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/generic-chip.c b/kernel/irq/generic-chip.c index ca98cc5d6308..4b011064e146 100644 --- a/kernel/irq/generic-chip.c +++ b/kernel/irq/generic-chip.c @@ -270,10 +270,7 @@ int irq_alloc_domain_generic_chips(struct irq_domain *d, int irqs_per_chip, if (d->gc) return -EBUSY; - if (d->revmap_type != IRQ_DOMAIN_MAP_LINEAR) - return -EINVAL; - - numchips = d->revmap_data.linear.size / irqs_per_chip; + numchips = d->revmap_size / irqs_per_chip; if (!numchips) return -EINVAL; diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c index 5a1d8ec8509e..c38be78fceb4 100644 --- a/kernel/irq/irqdomain.c +++ b/kernel/irq/irqdomain.c @@ -25,7 +25,6 @@ static struct irq_domain *irq_default_domain; /** * irq_domain_alloc() - Allocate a new irq_domain data structure * @of_node: optional device-tree node of the interrupt controller - * @revmap_type: type of reverse mapping to use * @ops: map/unmap domain callbacks * @host_data: Controller private data pointer * @@ -34,7 +33,7 @@ static struct irq_domain *irq_default_domain; * to IRQ domain, or NULL on failure. */ static struct irq_domain *irq_domain_alloc(struct device_node *of_node, - unsigned int revmap_type, int size, + int size, const struct irq_domain_ops *ops, void *host_data) { @@ -46,12 +45,11 @@ static struct irq_domain *irq_domain_alloc(struct device_node *of_node, return NULL; /* Fill structure */ - INIT_RADIX_TREE(&domain->revmap_data.tree, GFP_KERNEL); - domain->revmap_type = revmap_type; + INIT_RADIX_TREE(&domain->revmap_tree, GFP_KERNEL); domain->ops = ops; domain->host_data = host_data; domain->of_node = of_node_get(of_node); - domain->revmap_data.linear.size = size; + domain->revmap_size = size; return domain; } @@ -67,8 +65,7 @@ static void irq_domain_add(struct irq_domain *domain) mutex_lock(&irq_domain_mutex); list_add(&domain->link, &irq_domain_list); mutex_unlock(&irq_domain_mutex); - pr_debug("Allocated domain of type %d @0x%p\n", - domain->revmap_type, domain); + pr_debug("Added domain %s\n", domain->name); } /** @@ -88,7 +85,7 @@ void irq_domain_remove(struct irq_domain *domain) * node when all entries are removed. Shout if there are * any mappings left. */ - WARN_ON(domain->revmap_data.tree.height); + WARN_ON(domain->revmap_tree.height); list_del(&domain->link); @@ -100,8 +97,7 @@ void irq_domain_remove(struct irq_domain *domain) mutex_unlock(&irq_domain_mutex); - pr_debug("Removed domain of type %d @0x%p\n", - domain->revmap_type, domain); + pr_debug("Removed domain %s\n", domain->name); irq_domain_free(domain); } @@ -216,7 +212,7 @@ struct irq_domain *irq_domain_add_linear(struct device_node *of_node, { struct irq_domain *domain; - domain = irq_domain_alloc(of_node, IRQ_DOMAIN_MAP_LINEAR, size, ops, host_data); + domain = irq_domain_alloc(of_node, size, ops, host_data); if (!domain) return NULL; @@ -230,10 +226,9 @@ struct irq_domain *irq_domain_add_nomap(struct device_node *of_node, const struct irq_domain_ops *ops, void *host_data) { - struct irq_domain *domain = irq_domain_alloc(of_node, - IRQ_DOMAIN_MAP_NOMAP, 0, ops, host_data); + struct irq_domain *domain = irq_domain_alloc(of_node, 0, ops, host_data); if (domain) { - domain->revmap_data.nomap.max_irq = max_irq ? max_irq : ~0; + domain->revmap_direct_max_irq = max_irq ? max_irq : ~0; irq_domain_add(domain); } return domain; @@ -321,11 +316,11 @@ static void irq_domain_disassociate_many(struct irq_domain *domain, irq_data->hwirq = 0; /* Clear reverse map for this hwirq */ - if (hwirq < domain->revmap_data.linear.size) { + if (hwirq < domain->revmap_size) { domain->linear_revmap[hwirq] = 0; } else { mutex_lock(&revmap_trees_mutex); - radix_tree_delete(&domain->revmap_data.tree, hwirq); + radix_tree_delete(&domain->revmap_tree, hwirq); mutex_unlock(&revmap_trees_mutex); } } @@ -378,11 +373,11 @@ int irq_domain_associate_many(struct irq_domain *domain, unsigned int irq_base, domain->name = irq_data->chip->name; } - if (hwirq < domain->revmap_data.linear.size) { + if (hwirq < domain->revmap_size) { domain->linear_revmap[hwirq] = virq; } else { mutex_lock(&revmap_trees_mutex); - radix_tree_insert(&domain->revmap_data.tree, hwirq, irq_data); + radix_tree_insert(&domain->revmap_tree, hwirq, irq_data); mutex_unlock(&revmap_trees_mutex); } @@ -399,7 +394,9 @@ EXPORT_SYMBOL_GPL(irq_domain_associate_many); * * This routine is used for irq controllers which can choose the hardware * interrupt numbers they generate. In such a case it's simplest to use - * the linux irq as the hardware interrupt number. + * the linux irq as the hardware interrupt number. It still uses the linear + * or radix tree to store the mapping, but the irq controller can optimize + * the revmap path by using the hwirq directly. */ unsigned int irq_create_direct_mapping(struct irq_domain *domain) { @@ -408,17 +405,14 @@ unsigned int irq_create_direct_mapping(struct irq_domain *domain) if (domain == NULL) domain = irq_default_domain; - if (WARN_ON(!domain || domain->revmap_type != IRQ_DOMAIN_MAP_NOMAP)) - return 0; - virq = irq_alloc_desc_from(1, of_node_to_nid(domain->of_node)); if (!virq) { pr_debug("create_direct virq allocation failed\n"); return 0; } - if (virq >= domain->revmap_data.nomap.max_irq) { + if (virq >= domain->revmap_direct_max_irq) { pr_err("ERROR: no free irqs available below %i maximum\n", - domain->revmap_data.nomap.max_irq); + domain->revmap_direct_max_irq); irq_free_desc(virq); return 0; } @@ -617,17 +611,13 @@ unsigned int irq_find_mapping(struct irq_domain *domain, if (domain == NULL) return 0; - switch (domain->revmap_type) { - case IRQ_DOMAIN_MAP_LINEAR: - return irq_linear_revmap(domain, hwirq); - case IRQ_DOMAIN_MAP_NOMAP: + if (hwirq < domain->revmap_direct_max_irq) { data = irq_get_irq_data(hwirq); if (data && (data->domain == domain) && (data->hwirq == hwirq)) return hwirq; - break; } - return 0; + return irq_linear_revmap(domain, hwirq); } EXPORT_SYMBOL_GPL(irq_find_mapping); @@ -643,12 +633,11 @@ unsigned int irq_linear_revmap(struct irq_domain *domain, irq_hw_number_t hwirq) { struct irq_data *data; - BUG_ON(domain->revmap_type != IRQ_DOMAIN_MAP_LINEAR); /* Check revmap bounds; complain if exceeded */ - if (hwirq >= domain->revmap_data.linear.size) { + if (hwirq >= domain->revmap_size) { rcu_read_lock(); - data = radix_tree_lookup(&domain->revmap_data.tree, hwirq); + data = radix_tree_lookup(&domain->revmap_tree, hwirq); rcu_read_unlock(); return data ? data->irq : 0; } -- cgit v1.2.3 From fa40f377577752b83252b9d2b3165d4acee0eb7c Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Sat, 8 Jun 2013 12:57:40 +0100 Subject: irqdomain: Clean up aftermath of irq_domain refactoring After refactoring the irqdomain code, there are a number of API functions that are merely empty wrappers around core code. Drop those wrappers out of the C file and replace them with static inlines in the header. Signed-off-by: Grant Likely --- kernel/irq/irqdomain.c | 127 ++++++++++++++----------------------------------- 1 file changed, 36 insertions(+), 91 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c index c38be78fceb4..e0db59e2eef6 100644 --- a/kernel/irq/irqdomain.c +++ b/kernel/irq/irqdomain.c @@ -23,8 +23,11 @@ static DEFINE_MUTEX(revmap_trees_mutex); static struct irq_domain *irq_default_domain; /** - * irq_domain_alloc() - Allocate a new irq_domain data structure + * __irq_domain_add() - Allocate a new irq_domain data structure * @of_node: optional device-tree node of the interrupt controller + * @size: Size of linear map; 0 for radix mapping only + * @direct_max: Maximum value of direct maps; Use ~0 for no limit; 0 for no + * direct mapping * @ops: map/unmap domain callbacks * @host_data: Controller private data pointer * @@ -32,10 +35,10 @@ static struct irq_domain *irq_default_domain; * register allocated irq_domain with irq_domain_register(). Returns pointer * to IRQ domain, or NULL on failure. */ -static struct irq_domain *irq_domain_alloc(struct device_node *of_node, - int size, - const struct irq_domain_ops *ops, - void *host_data) +struct irq_domain *__irq_domain_add(struct device_node *of_node, + int size, int direct_max, + const struct irq_domain_ops *ops, + void *host_data) { struct irq_domain *domain; @@ -50,23 +53,16 @@ static struct irq_domain *irq_domain_alloc(struct device_node *of_node, domain->host_data = host_data; domain->of_node = of_node_get(of_node); domain->revmap_size = size; + domain->revmap_direct_max_irq = direct_max; - return domain; -} - -static void irq_domain_free(struct irq_domain *domain) -{ - of_node_put(domain->of_node); - kfree(domain); -} - -static void irq_domain_add(struct irq_domain *domain) -{ mutex_lock(&irq_domain_mutex); list_add(&domain->link, &irq_domain_list); mutex_unlock(&irq_domain_mutex); + pr_debug("Added domain %s\n", domain->name); + return domain; } +EXPORT_SYMBOL_GPL(__irq_domain_add); /** * irq_domain_remove() - Remove an irq domain. @@ -99,30 +95,28 @@ void irq_domain_remove(struct irq_domain *domain) pr_debug("Removed domain %s\n", domain->name); - irq_domain_free(domain); + of_node_put(domain->of_node); + kfree(domain); } EXPORT_SYMBOL_GPL(irq_domain_remove); /** - * irq_domain_add_simple() - Allocate and register a simple irq_domain. + * irq_domain_add_simple() - Register an irq_domain and optionally map a range of irqs * @of_node: pointer to interrupt controller's device tree node. * @size: total number of irqs in mapping * @first_irq: first number of irq block assigned to the domain, - * pass zero to assign irqs on-the-fly. This will result in a - * linear IRQ domain so it is important to use irq_create_mapping() - * for each used IRQ, especially when SPARSE_IRQ is enabled. + * pass zero to assign irqs on-the-fly. If first_irq is non-zero, then + * pre-map all of the irqs in the domain to virqs starting at first_irq. * @ops: map/unmap domain callbacks * @host_data: Controller private data pointer * - * Allocates a legacy irq_domain if irq_base is positive or a linear - * domain otherwise. For the legacy domain, IRQ descriptors will also - * be allocated. + * Allocates an irq_domain, and optionally if first_irq is positive then also + * allocate irq_descs and map all of the hwirqs to virqs starting at first_irq. * * This is intended to implement the expected behaviour for most - * interrupt controllers which is that a linear mapping should - * normally be used unless the system requires a legacy mapping in - * order to support supplying interrupt numbers during non-DT - * registration of devices. + * interrupt controllers. If device tree is used, then first_irq will be 0 and + * irqs get mapped dynamically on the fly. However, if the controller requires + * static virq assignments (non-DT boot) then it will set that up correctly. */ struct irq_domain *irq_domain_add_simple(struct device_node *of_node, unsigned int size, @@ -130,33 +124,25 @@ struct irq_domain *irq_domain_add_simple(struct device_node *of_node, const struct irq_domain_ops *ops, void *host_data) { - if (first_irq > 0) { - int irq_base; + struct irq_domain *domain; + + domain = __irq_domain_add(of_node, size, 0, ops, host_data); + if (!domain) + return NULL; + if (first_irq > 0) { if (IS_ENABLED(CONFIG_SPARSE_IRQ)) { - /* - * Set the descriptor allocator to search for a - * 1-to-1 mapping, such as irq_alloc_desc_at(). - * Use of_node_to_nid() which is defined to - * numa_node_id() on platforms that have no custom - * implementation. - */ - irq_base = irq_alloc_descs(first_irq, first_irq, size, - of_node_to_nid(of_node)); - if (irq_base < 0) { + /* attempt to allocated irq_descs */ + int rc = irq_alloc_descs(first_irq, first_irq, size, + of_node_to_nid(of_node)); + if (rc < 0) pr_info("Cannot allocate irq_descs @ IRQ%d, assuming pre-allocated\n", first_irq); - irq_base = first_irq; - } - } else - irq_base = first_irq; - - return irq_domain_add_legacy(of_node, size, irq_base, 0, - ops, host_data); + } + WARN_ON(irq_domain_associate_many(domain, first_irq, 0, size)); } - /* A linear domain is the default */ - return irq_domain_add_linear(of_node, size, ops, host_data); + return domain; } EXPORT_SYMBOL_GPL(irq_domain_add_simple); @@ -184,11 +170,7 @@ struct irq_domain *irq_domain_add_legacy(struct device_node *of_node, { struct irq_domain *domain; - pr_debug("Setting up legacy domain virq[%i:%i] ==> hwirq[%i:%i]\n", - first_irq, first_irq + size - 1, - (int)first_hwirq, (int)first_hwirq + size -1); - - domain = irq_domain_add_linear(of_node, first_hwirq + size, ops, host_data); + domain = __irq_domain_add(of_node, first_hwirq + size, 0, ops, host_data); if (!domain) return NULL; @@ -198,43 +180,6 @@ struct irq_domain *irq_domain_add_legacy(struct device_node *of_node, } EXPORT_SYMBOL_GPL(irq_domain_add_legacy); -/** - * irq_domain_add_linear() - Allocate and register a linear revmap irq_domain. - * @of_node: pointer to interrupt controller's device tree node. - * @size: Number of interrupts in the domain. - * @ops: map/unmap domain callbacks - * @host_data: Controller private data pointer - */ -struct irq_domain *irq_domain_add_linear(struct device_node *of_node, - unsigned int size, - const struct irq_domain_ops *ops, - void *host_data) -{ - struct irq_domain *domain; - - domain = irq_domain_alloc(of_node, size, ops, host_data); - if (!domain) - return NULL; - - irq_domain_add(domain); - return domain; -} -EXPORT_SYMBOL_GPL(irq_domain_add_linear); - -struct irq_domain *irq_domain_add_nomap(struct device_node *of_node, - unsigned int max_irq, - const struct irq_domain_ops *ops, - void *host_data) -{ - struct irq_domain *domain = irq_domain_alloc(of_node, 0, ops, host_data); - if (domain) { - domain->revmap_direct_max_irq = max_irq ? max_irq : ~0; - irq_domain_add(domain); - } - return domain; -} -EXPORT_SYMBOL_GPL(irq_domain_add_nomap); - /** * irq_find_host() - Locates a domain for a given device node * @node: device-tree node of the interrupt controller -- cgit v1.2.3 From 1400ea86025a22862f97e7fe544433751b43ecec Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Thu, 6 Jun 2013 22:20:44 +0100 Subject: irqdomain: Beef up debugfs output This patch increases the amount of output produced by the irq_domain_mapping debugfs file by first listing all of the registered irq domains at the beginning of the output, and then by including all mapped IRQs in the output, not just the active ones. It is very useful when debugging irqdomain issues to be able to see the entire list of mapped irqs, not just the ones that happen to be connected to devices. Signed-off-by: Grant Likely --- kernel/irq/irqdomain.c | 35 ++++++++++++++++++++++++++++++----- 1 file changed, 30 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c index e0db59e2eef6..280b8047d8db 100644 --- a/kernel/irq/irqdomain.c +++ b/kernel/irq/irqdomain.c @@ -596,12 +596,29 @@ static int virq_debug_show(struct seq_file *m, void *private) { unsigned long flags; struct irq_desc *desc; - void *data; + struct irq_domain *domain; + struct radix_tree_iter iter; + void *data, **slot; int i; - seq_printf(m, "%-5s %-7s %-15s %-*s %s\n", "irq", "hwirq", + seq_printf(m, " %-16s %-6s %-10s %-10s %s\n", + "name", "mapped", "linear-max", "direct-max", "devtree-node"); + mutex_lock(&irq_domain_mutex); + list_for_each_entry(domain, &irq_domain_list, link) { + int count = 0; + radix_tree_for_each_slot(slot, &domain->revmap_tree, &iter, 0) + count++; + seq_printf(m, "%c%-16s %6u %10u %10u %s\n", + domain == irq_default_domain ? '*' : ' ', domain->name, + domain->revmap_size + count, domain->revmap_size, + domain->revmap_direct_max_irq, + domain->of_node ? of_node_full_name(domain->of_node) : ""); + } + mutex_unlock(&irq_domain_mutex); + + seq_printf(m, "%-5s %-7s %-15s %-*s %6s %-14s %s\n", "irq", "hwirq", "chip name", (int)(2 * sizeof(void *) + 2), "chip data", - "domain name"); + "active", "type", "domain"); for (i = 1; i < nr_irqs; i++) { desc = irq_to_desc(i); @@ -609,12 +626,15 @@ static int virq_debug_show(struct seq_file *m, void *private) continue; raw_spin_lock_irqsave(&desc->lock, flags); + domain = desc->irq_data.domain; - if (desc->action && desc->action->handler) { + if (domain) { struct irq_chip *chip; + int hwirq = desc->irq_data.hwirq; + bool direct; seq_printf(m, "%5d ", i); - seq_printf(m, "0x%05lx ", desc->irq_data.hwirq); + seq_printf(m, "0x%05x ", hwirq); chip = irq_desc_get_chip(desc); seq_printf(m, "%-15s ", (chip && chip->name) ? chip->name : "none"); @@ -622,6 +642,11 @@ static int virq_debug_show(struct seq_file *m, void *private) data = irq_desc_get_chip_data(desc); seq_printf(m, data ? "0x%p " : " %p ", data); + seq_printf(m, " %c ", (desc->action && desc->action->handler) ? '*' : ' '); + direct = (i == hwirq) && (i < domain->revmap_direct_max_irq); + seq_printf(m, "%6s%-8s ", + (hwirq < domain->revmap_size) ? "LINEAR" : "RADIX", + direct ? "(DIRECT)" : ""); seq_printf(m, "%s\n", desc->irq_data.domain->name); } -- cgit v1.2.3 From 016a8d5be6ddcc72ef0432d82d9f6fa34f61b907 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 28 May 2013 17:32:53 -0400 Subject: rcu: Don't call wakeup() with rcu_node structure ->lock held This commit fixes a lockdep-detected deadlock by moving a wake_up() call out from a rnp->lock critical section. Please see below for the long version of this story. On Tue, 2013-05-28 at 16:13 -0400, Dave Jones wrote: > [12572.705832] ====================================================== > [12572.750317] [ INFO: possible circular locking dependency detected ] > [12572.796978] 3.10.0-rc3+ #39 Not tainted > [12572.833381] ------------------------------------------------------- > [12572.862233] trinity-child17/31341 is trying to acquire lock: > [12572.870390] (rcu_node_0){..-.-.}, at: [] rcu_read_unlock_special+0x9f/0x4c0 > [12572.878859] > but task is already holding lock: > [12572.894894] (&ctx->lock){-.-...}, at: [] perf_lock_task_context+0x7d/0x2d0 > [12572.903381] > which lock already depends on the new lock. > > [12572.927541] > the existing dependency chain (in reverse order) is: > [12572.943736] > -> #4 (&ctx->lock){-.-...}: > [12572.960032] [] lock_acquire+0x91/0x1f0 > [12572.968337] [] _raw_spin_lock+0x40/0x80 > [12572.976633] [] __perf_event_task_sched_out+0x2e7/0x5e0 > [12572.984969] [] perf_event_task_sched_out+0x93/0xa0 > [12572.993326] [] __schedule+0x2cf/0x9c0 > [12573.001652] [] schedule_user+0x2e/0x70 > [12573.009998] [] retint_careful+0x12/0x2e > [12573.018321] > -> #3 (&rq->lock){-.-.-.}: > [12573.034628] [] lock_acquire+0x91/0x1f0 > [12573.042930] [] _raw_spin_lock+0x40/0x80 > [12573.051248] [] wake_up_new_task+0xb7/0x260 > [12573.059579] [] do_fork+0x105/0x470 > [12573.067880] [] kernel_thread+0x26/0x30 > [12573.076202] [] rest_init+0x23/0x140 > [12573.084508] [] start_kernel+0x3f1/0x3fe > [12573.092852] [] x86_64_start_reservations+0x2a/0x2c > [12573.101233] [] x86_64_start_kernel+0xcc/0xcf > [12573.109528] > -> #2 (&p->pi_lock){-.-.-.}: > [12573.125675] [] lock_acquire+0x91/0x1f0 > [12573.133829] [] _raw_spin_lock_irqsave+0x4b/0x90 > [12573.141964] [] try_to_wake_up+0x31/0x320 > [12573.150065] [] default_wake_function+0x12/0x20 > [12573.158151] [] autoremove_wake_function+0x18/0x40 > [12573.166195] [] __wake_up_common+0x58/0x90 > [12573.174215] [] __wake_up+0x39/0x50 > [12573.182146] [] rcu_start_gp_advanced.isra.11+0x4a/0x50 > [12573.190119] [] rcu_start_future_gp+0x1c9/0x1f0 > [12573.198023] [] rcu_nocb_kthread+0x114/0x930 > [12573.205860] [] kthread+0xed/0x100 > [12573.213656] [] ret_from_fork+0x7c/0xb0 > [12573.221379] > -> #1 (&rsp->gp_wq){..-.-.}: > [12573.236329] [] lock_acquire+0x91/0x1f0 > [12573.243783] [] _raw_spin_lock_irqsave+0x4b/0x90 > [12573.251178] [] __wake_up+0x23/0x50 > [12573.258505] [] rcu_start_gp_advanced.isra.11+0x4a/0x50 > [12573.265891] [] rcu_start_future_gp+0x1c9/0x1f0 > [12573.273248] [] rcu_nocb_kthread+0x114/0x930 > [12573.280564] [] kthread+0xed/0x100 > [12573.287807] [] ret_from_fork+0x7c/0xb0 Notice the above call chain. rcu_start_future_gp() is called with the rnp->lock held. Then it calls rcu_start_gp_advance, which does a wakeup. You can't do wakeups while holding the rnp->lock, as that would mean that you could not do a rcu_read_unlock() while holding the rq lock, or any lock that was taken while holding the rq lock. This is because... (See below). > [12573.295067] > -> #0 (rcu_node_0){..-.-.}: > [12573.309293] [] __lock_acquire+0x1786/0x1af0 > [12573.316568] [] lock_acquire+0x91/0x1f0 > [12573.323825] [] _raw_spin_lock+0x40/0x80 > [12573.331081] [] rcu_read_unlock_special+0x9f/0x4c0 > [12573.338377] [] __rcu_read_unlock+0x96/0xa0 > [12573.345648] [] perf_lock_task_context+0x143/0x2d0 > [12573.352942] [] find_get_context+0x4e/0x1f0 > [12573.360211] [] SYSC_perf_event_open+0x514/0xbd0 > [12573.367514] [] SyS_perf_event_open+0x9/0x10 > [12573.374816] [] tracesys+0xdd/0xe2 Notice the above trace. perf took its own ctx->lock, which can be taken while holding the rq lock. While holding this lock, it did a rcu_read_unlock(). The perf_lock_task_context() basically looks like: rcu_read_lock(); raw_spin_lock(ctx->lock); rcu_read_unlock(); Now, what looks to have happened, is that we scheduled after taking that first rcu_read_lock() but before taking the spin lock. When we scheduled back in and took the ctx->lock, the following rcu_read_unlock() triggered the "special" code. The rcu_read_unlock_special() takes the rnp->lock, which gives us a possible deadlock scenario. CPU0 CPU1 CPU2 ---- ---- ---- rcu_nocb_kthread() lock(rq->lock); lock(ctx->lock); lock(rnp->lock); wake_up(); lock(rq->lock); rcu_read_unlock(); rcu_read_unlock_special(); lock(rnp->lock); lock(ctx->lock); **** DEADLOCK **** > [12573.382068] > other info that might help us debug this: > > [12573.403229] Chain exists of: > rcu_node_0 --> &rq->lock --> &ctx->lock > > [12573.424471] Possible unsafe locking scenario: > > [12573.438499] CPU0 CPU1 > [12573.445599] ---- ---- > [12573.452691] lock(&ctx->lock); > [12573.459799] lock(&rq->lock); > [12573.467010] lock(&ctx->lock); > [12573.474192] lock(rcu_node_0); > [12573.481262] > *** DEADLOCK *** > > [12573.501931] 1 lock held by trinity-child17/31341: > [12573.508990] #0: (&ctx->lock){-.-...}, at: [] perf_lock_task_context+0x7d/0x2d0 > [12573.516475] > stack backtrace: > [12573.530395] CPU: 1 PID: 31341 Comm: trinity-child17 Not tainted 3.10.0-rc3+ #39 > [12573.545357] ffffffff825b4f90 ffff880219f1dbc0 ffffffff816e375b ffff880219f1dc00 > [12573.552868] ffffffff816dfa5d ffff880219f1dc50 ffff88023ce4d1f8 ffff88023ce4ca40 > [12573.560353] 0000000000000001 0000000000000001 ffff88023ce4d1f8 ffff880219f1dcc0 > [12573.567856] Call Trace: > [12573.575011] [] dump_stack+0x19/0x1b > [12573.582284] [] print_circular_bug+0x200/0x20f > [12573.589637] [] __lock_acquire+0x1786/0x1af0 > [12573.596982] [] ? sched_clock_cpu+0xb5/0x100 > [12573.604344] [] lock_acquire+0x91/0x1f0 > [12573.611652] [] ? rcu_read_unlock_special+0x9f/0x4c0 > [12573.619030] [] _raw_spin_lock+0x40/0x80 > [12573.626331] [] ? rcu_read_unlock_special+0x9f/0x4c0 > [12573.633671] [] rcu_read_unlock_special+0x9f/0x4c0 > [12573.640992] [] ? perf_lock_task_context+0x7d/0x2d0 > [12573.648330] [] ? put_lock_stats.isra.29+0xe/0x40 > [12573.655662] [] ? delay_tsc+0x90/0xe0 > [12573.662964] [] __rcu_read_unlock+0x96/0xa0 > [12573.670276] [] perf_lock_task_context+0x143/0x2d0 > [12573.677622] [] ? __perf_event_enable+0x370/0x370 > [12573.684981] [] find_get_context+0x4e/0x1f0 > [12573.692358] [] SYSC_perf_event_open+0x514/0xbd0 > [12573.699753] [] ? get_parent_ip+0xd/0x50 > [12573.707135] [] ? trace_hardirqs_on_caller+0xfd/0x1c0 > [12573.714599] [] SyS_perf_event_open+0x9/0x10 > [12573.721996] [] tracesys+0xdd/0xe2 This commit delays the wakeup via irq_work(), which is what perf and ftrace use to perform wakeups in critical sections. Reported-by: Dave Jones Signed-off-by: Steven Rostedt Signed-off-by: Paul E. McKenney --- kernel/rcutree.c | 17 +++++++++++++++-- kernel/rcutree.h | 2 ++ 2 files changed, 17 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 16ea67925015..b61d20c5ee7b 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1613,6 +1613,14 @@ static int __noreturn rcu_gp_kthread(void *arg) } } +static void rsp_wakeup(struct irq_work *work) +{ + struct rcu_state *rsp = container_of(work, struct rcu_state, wakeup_work); + + /* Wake up rcu_gp_kthread() to start the grace period. */ + wake_up(&rsp->gp_wq); +} + /* * Start a new RCU grace period if warranted, re-initializing the hierarchy * in preparation for detecting the next grace period. The caller must hold @@ -1637,8 +1645,12 @@ rcu_start_gp_advanced(struct rcu_state *rsp, struct rcu_node *rnp, } rsp->gp_flags = RCU_GP_FLAG_INIT; - /* Wake up rcu_gp_kthread() to start the grace period. */ - wake_up(&rsp->gp_wq); + /* + * We can't do wakeups while holding the rnp->lock, as that + * could cause possible deadlocks with the rq->lock. Deter + * the wakeup to interrupt context. + */ + irq_work_queue(&rsp->wakeup_work); } /* @@ -3235,6 +3247,7 @@ static void __init rcu_init_one(struct rcu_state *rsp, rsp->rda = rda; init_waitqueue_head(&rsp->gp_wq); + init_irq_work(&rsp->wakeup_work, rsp_wakeup); rnp = rsp->level[rcu_num_lvls - 1]; for_each_possible_cpu(i) { while (i > rnp->grphi) diff --git a/kernel/rcutree.h b/kernel/rcutree.h index da77a8f57ff9..4df503470e42 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -27,6 +27,7 @@ #include #include #include +#include /* * Define shape of hierarchy based on NR_CPUS, CONFIG_RCU_FANOUT, and @@ -442,6 +443,7 @@ struct rcu_state { char *name; /* Name of structure. */ char abbr; /* Abbreviated name. */ struct list_head flavors; /* List of RCU flavors. */ + struct irq_work wakeup_work; /* Postponed wakeups */ }; /* Values for rcu_state structure's gp_flags field. */ -- cgit v1.2.3 From 971394f389992f8462c4e5ae0e3b49a10a9534a3 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Sun, 2 Jun 2013 07:13:57 -0700 Subject: rcu: Fix deadlock with CPU hotplug, RCU GP init, and timer migration In Steven Rostedt's words: > I've been debugging the last couple of days why my tests have been > locking up. One of my tracing tests, runs all available tracers. The > lockup always happened with the mmiotrace, which is used to trace > interactions between priority drivers and the kernel. But to do this > easily, when the tracer gets registered, it disables all but the boot > CPUs. The lockup always happened after it got done disabling the CPUs. > > Then I decided to try this: > > while :; do > for i in 1 2 3; do > echo 0 > /sys/devices/system/cpu/cpu$i/online > done > for i in 1 2 3; do > echo 1 > /sys/devices/system/cpu/cpu$i/online > done > done > > Well, sure enough, that locked up too, with the same users. Doing a > sysrq-w (showing all blocked tasks): > > [ 2991.344562] task PC stack pid father > [ 2991.344562] rcu_preempt D ffff88007986fdf8 0 10 2 0x00000000 > [ 2991.344562] ffff88007986fc98 0000000000000002 ffff88007986fc48 0000000000000908 > [ 2991.344562] ffff88007986c280 ffff88007986ffd8 ffff88007986ffd8 00000000001d3c80 > [ 2991.344562] ffff880079248a40 ffff88007986c280 0000000000000000 00000000fffd4295 > [ 2991.344562] Call Trace: > [ 2991.344562] [] schedule+0x64/0x66 > [ 2991.344562] [] schedule_timeout+0xbc/0xf9 > [ 2991.344562] [] ? ftrace_call+0x5/0x2f > [ 2991.344562] [] ? cascade+0xa8/0xa8 > [ 2991.344562] [] schedule_timeout_uninterruptible+0x1e/0x20 > [ 2991.344562] [] rcu_gp_kthread+0x502/0x94b > [ 2991.344562] [] ? __init_waitqueue_head+0x50/0x50 > [ 2991.344562] [] ? rcu_gp_fqs+0x64/0x64 > [ 2991.344562] [] kthread+0xb1/0xb9 > [ 2991.344562] [] ? lock_release_holdtime.part.23+0x4e/0x55 > [ 2991.344562] [] ? __init_kthread_worker+0x58/0x58 > [ 2991.344562] [] ret_from_fork+0x7c/0xb0 > [ 2991.344562] [] ? __init_kthread_worker+0x58/0x58 > [ 2991.344562] kworker/0:1 D ffffffff81a30680 0 47 2 0x00000000 > [ 2991.344562] Workqueue: events cpuset_hotplug_workfn > [ 2991.344562] ffff880078dbbb58 0000000000000002 0000000000000006 00000000000000d8 > [ 2991.344562] ffff880078db8100 ffff880078dbbfd8 ffff880078dbbfd8 00000000001d3c80 > [ 2991.344562] ffff8800779ca5c0 ffff880078db8100 ffffffff81541fcf 0000000000000000 > [ 2991.344562] Call Trace: > [ 2991.344562] [] ? __mutex_lock_common+0x3d4/0x609 > [ 2991.344562] [] schedule+0x64/0x66 > [ 2991.344562] [] schedule_preempt_disabled+0x18/0x24 > [ 2991.344562] [] __mutex_lock_common+0x3d4/0x609 > [ 2991.344562] [] ? get_online_cpus+0x3c/0x50 > [ 2991.344562] [] ? get_online_cpus+0x3c/0x50 > [ 2991.344562] [] mutex_lock_nested+0x3b/0x40 > [ 2991.344562] [] get_online_cpus+0x3c/0x50 > [ 2991.344562] [] rebuild_sched_domains_locked+0x6e/0x3a8 > [ 2991.344562] [] rebuild_sched_domains+0x1c/0x2a > [ 2991.344562] [] cpuset_hotplug_workfn+0x1c7/0x1d3 > [ 2991.344562] [] ? cpuset_hotplug_workfn+0x5/0x1d3 > [ 2991.344562] [] process_one_work+0x2d4/0x4d1 > [ 2991.344562] [] ? process_one_work+0x207/0x4d1 > [ 2991.344562] [] worker_thread+0x2e7/0x3b5 > [ 2991.344562] [] ? rescuer_thread+0x332/0x332 > [ 2991.344562] [] kthread+0xb1/0xb9 > [ 2991.344562] [] ? __init_kthread_worker+0x58/0x58 > [ 2991.344562] [] ret_from_fork+0x7c/0xb0 > [ 2991.344562] [] ? __init_kthread_worker+0x58/0x58 > [ 2991.344562] bash D ffffffff81a4aa80 0 2618 2612 0x10000000 > [ 2991.344562] ffff8800379abb58 0000000000000002 0000000000000006 0000000000000c2c > [ 2991.344562] ffff880077fea140 ffff8800379abfd8 ffff8800379abfd8 00000000001d3c80 > [ 2991.344562] ffff8800779ca5c0 ffff880077fea140 ffffffff81541fcf 0000000000000000 > [ 2991.344562] Call Trace: > [ 2991.344562] [] ? __mutex_lock_common+0x3d4/0x609 > [ 2991.344562] [] schedule+0x64/0x66 > [ 2991.344562] [] schedule_preempt_disabled+0x18/0x24 > [ 2991.344562] [] __mutex_lock_common+0x3d4/0x609 > [ 2991.344562] [] ? rcu_cpu_notify+0x2f5/0x86e > [ 2991.344562] [] ? rcu_cpu_notify+0x2f5/0x86e > [ 2991.344562] [] mutex_lock_nested+0x3b/0x40 > [ 2991.344562] [] rcu_cpu_notify+0x2f5/0x86e > [ 2991.344562] [] ? __lock_is_held+0x32/0x53 > [ 2991.344562] [] notifier_call_chain+0x6b/0x98 > [ 2991.344562] [] __raw_notifier_call_chain+0xe/0x10 > [ 2991.344562] [] __cpu_notify+0x20/0x32 > [ 2991.344562] [] cpu_notify_nofail+0x17/0x36 > [ 2991.344562] [] _cpu_down+0x154/0x259 > [ 2991.344562] [] cpu_down+0x2d/0x3a > [ 2991.344562] [] store_online+0x4e/0xe7 > [ 2991.344562] [] dev_attr_store+0x20/0x22 > [ 2991.344562] [] sysfs_write_file+0x108/0x144 > [ 2991.344562] [] vfs_write+0xfd/0x158 > [ 2991.344562] [] SyS_write+0x5c/0x83 > [ 2991.344562] [] tracesys+0xdd/0xe2 > > As well as held locks: > > [ 3034.728033] Showing all locks held in the system: > [ 3034.728033] 1 lock held by rcu_preempt/10: > [ 3034.728033] #0: (rcu_preempt_state.onoff_mutex){+.+...}, at: [] rcu_gp_kthread+0x167/0x94b > [ 3034.728033] 4 locks held by kworker/0:1/47: > [ 3034.728033] #0: (events){.+.+.+}, at: [] process_one_work+0x207/0x4d1 > [ 3034.728033] #1: (cpuset_hotplug_work){+.+.+.}, at: [] process_one_work+0x207/0x4d1 > [ 3034.728033] #2: (cpuset_mutex){+.+.+.}, at: [] rebuild_sched_domains+0x17/0x2a > [ 3034.728033] #3: (cpu_hotplug.lock){+.+.+.}, at: [] get_online_cpus+0x3c/0x50 > [ 3034.728033] 1 lock held by mingetty/2563: > [ 3034.728033] #0: (&ldata->atomic_read_lock){+.+...}, at: [] n_tty_read+0x252/0x7e8 > [ 3034.728033] 1 lock held by mingetty/2565: > [ 3034.728033] #0: (&ldata->atomic_read_lock){+.+...}, at: [] n_tty_read+0x252/0x7e8 > [ 3034.728033] 1 lock held by mingetty/2569: > [ 3034.728033] #0: (&ldata->atomic_read_lock){+.+...}, at: [] n_tty_read+0x252/0x7e8 > [ 3034.728033] 1 lock held by mingetty/2572: > [ 3034.728033] #0: (&ldata->atomic_read_lock){+.+...}, at: [] n_tty_read+0x252/0x7e8 > [ 3034.728033] 1 lock held by mingetty/2575: > [ 3034.728033] #0: (&ldata->atomic_read_lock){+.+...}, at: [] n_tty_read+0x252/0x7e8 > [ 3034.728033] 7 locks held by bash/2618: > [ 3034.728033] #0: (sb_writers#5){.+.+.+}, at: [] file_start_write+0x2a/0x2c > [ 3034.728033] #1: (&buffer->mutex#2){+.+.+.}, at: [] sysfs_write_file+0x3c/0x144 > [ 3034.728033] #2: (s_active#54){.+.+.+}, at: [] sysfs_write_file+0xe7/0x144 > [ 3034.728033] #3: (x86_cpu_hotplug_driver_mutex){+.+.+.}, at: [] cpu_hotplug_driver_lock+0x17/0x19 > [ 3034.728033] #4: (cpu_add_remove_lock){+.+.+.}, at: [] cpu_maps_update_begin+0x17/0x19 > [ 3034.728033] #5: (cpu_hotplug.lock){+.+.+.}, at: [] cpu_hotplug_begin+0x2c/0x6d > [ 3034.728033] #6: (rcu_preempt_state.onoff_mutex){+.+...}, at: [] rcu_cpu_notify+0x2f5/0x86e > [ 3034.728033] 1 lock held by bash/2980: > [ 3034.728033] #0: (&ldata->atomic_read_lock){+.+...}, at: [] n_tty_read+0x252/0x7e8 > > Things looked a little weird. Also, this is a deadlock that lockdep did > not catch. But what we have here does not look like a circular lock > issue: > > Bash is blocked in rcu_cpu_notify(): > > 1961 /* Exclude any attempts to start a new grace period. */ > 1962 mutex_lock(&rsp->onoff_mutex); > > > kworker is blocked in get_online_cpus(), which makes sense as we are > currently taking down a CPU. > > But rcu_preempt is not blocked on anything. It is simply sleeping in > rcu_gp_kthread (really rcu_gp_init) here: > > 1453 #ifdef CONFIG_PROVE_RCU_DELAY > 1454 if ((prandom_u32() % (rcu_num_nodes * 8)) == 0 && > 1455 system_state == SYSTEM_RUNNING) > 1456 schedule_timeout_uninterruptible(2); > 1457 #endif /* #ifdef CONFIG_PROVE_RCU_DELAY */ > > And it does this while holding the onoff_mutex that bash is waiting for. > > Doing a function trace, it showed me where it happened: > > [ 125.940066] rcu_pree-10 3.... 28384115273: schedule_timeout_uninterruptible <-rcu_gp_kthread > [...] > [ 125.940066] rcu_pree-10 3d..3 28384202439: sched_switch: prev_comm=rcu_preempt prev_pid=10 prev_prio=120 prev_state=D ==> next_comm=watchdog/3 next_pid=38 next_prio=120 > > The watchdog ran, and then: > > [ 125.940066] watchdog-38 3d..3 28384692863: sched_switch: prev_comm=watchdog/3 prev_pid=38 prev_prio=120 prev_state=P ==> next_comm=modprobe next_pid=2848 next_prio=118 > > Not sure what modprobe was doing, but shortly after that: > > [ 125.940066] modprobe-2848 3d..3 28385041749: sched_switch: prev_comm=modprobe prev_pid=2848 prev_prio=118 prev_state=R+ ==> next_comm=migration/3 next_pid=40 next_prio=0 > > Where the migration thread took down the CPU: > > [ 125.940066] migratio-40 3d..3 28389148276: sched_switch: prev_comm=migration/3 prev_pid=40 prev_prio=0 prev_state=P ==> next_comm=swapper/3 next_pid=0 next_prio=120 > > which finally did: > > [ 125.940066] -0 3...1 28389282142: arch_cpu_idle_dead <-cpu_startup_entry > [ 125.940066] -0 3...1 28389282548: native_play_dead <-arch_cpu_idle_dead > [ 125.940066] -0 3...1 28389282924: play_dead_common <-native_play_dead > [ 125.940066] -0 3...1 28389283468: idle_task_exit <-play_dead_common > [ 125.940066] -0 3...1 28389284644: amd_e400_remove_cpu <-play_dead_common > > > CPU 3 is now offline, the rcu_preempt thread that ran on CPU 3 is still > doing a schedule_timeout_uninterruptible() and it registered it's > timeout to the timer base for CPU 3. You would think that it would get > migrated right? The issue here is that the timer migration happens at > the CPU notifier for CPU_DEAD. The problem is that the rcu notifier for > CPU_DOWN is blocked waiting for the onoff_mutex to be released, which is > held by the thread that just put itself into a uninterruptible sleep, > that wont wake up until the CPU_DEAD notifier of the timer > infrastructure is called, which wont happen until the rcu notifier > finishes. Here's our deadlock! This commit breaks this deadlock cycle by substituting a shorter udelay() for the previous schedule_timeout_uninterruptible(), while at the same time increasing the probability of the delay. This maintains the intensity of the testing. Reported-by: Steven Rostedt Signed-off-by: Paul E. McKenney Tested-by: Steven Rostedt --- kernel/rcutree.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index b61d20c5ee7b..35380019f0fc 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1451,9 +1451,9 @@ static int rcu_gp_init(struct rcu_state *rsp) rnp->grphi, rnp->qsmask); raw_spin_unlock_irq(&rnp->lock); #ifdef CONFIG_PROVE_RCU_DELAY - if ((prandom_u32() % (rcu_num_nodes * 8)) == 0 && + if ((prandom_u32() % (rcu_num_nodes + 1)) == 0 && system_state == SYSTEM_RUNNING) - schedule_timeout_uninterruptible(2); + udelay(200); #endif /* #ifdef CONFIG_PROVE_RCU_DELAY */ cond_resched(); } -- cgit v1.2.3 From d7f3e207397d7b4868e33d3f88396a06f4d5a8c7 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 18 Mar 2013 16:19:52 -0700 Subject: rcu: Convert rcutree.c printk calls This commit converts printk() calls to the corresponding pr_*() calls. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutree.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 35380019f0fc..1009c0ccd4b1 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -866,7 +866,7 @@ static void print_other_cpu_stall(struct rcu_state *rsp) * See Documentation/RCU/stallwarn.txt for info on how to debug * RCU CPU stall warnings. */ - printk(KERN_ERR "INFO: %s detected stalls on CPUs/tasks:", + pr_err("INFO: %s detected stalls on CPUs/tasks:", rsp->name); print_cpu_stall_info_begin(); rcu_for_each_leaf_node(rsp, rnp) { @@ -899,7 +899,7 @@ static void print_other_cpu_stall(struct rcu_state *rsp) smp_processor_id(), (long)(jiffies - rsp->gp_start), rsp->gpnum, rsp->completed, totqlen); if (ndetected == 0) - printk(KERN_ERR "INFO: Stall ended before state dump start\n"); + pr_err("INFO: Stall ended before state dump start\n"); else if (!trigger_all_cpu_backtrace()) rcu_dump_cpu_stacks(rsp); @@ -922,7 +922,7 @@ static void print_cpu_stall(struct rcu_state *rsp) * See Documentation/RCU/stallwarn.txt for info on how to debug * RCU CPU stall warnings. */ - printk(KERN_ERR "INFO: %s self-detected stall on CPU", rsp->name); + pr_err("INFO: %s self-detected stall on CPU", rsp->name); print_cpu_stall_info_begin(); print_cpu_stall_info(rsp, smp_processor_id()); print_cpu_stall_info_end(); -- cgit v1.2.3 From 6eaef633d77f50f031dd355ff5f91aaa1aaf9885 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Tue, 19 Mar 2013 10:08:37 -0700 Subject: rcu: Move code to apply callback-numbering simplifications The addition of callback numbering allows combining the detection of the ends of old grace periods and the beginnings of new grace periods. This commit moves code to set the stage for this combining. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutree.c | 118 +++++++++++++++++++++++++++---------------------------- 1 file changed, 59 insertions(+), 59 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 1009c0ccd4b1..c36e52dc091d 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -984,65 +984,6 @@ void rcu_cpu_stall_reset(void) rsp->jiffies_stall = jiffies + ULONG_MAX / 2; } -/* - * Update CPU-local rcu_data state to record the newly noticed grace period. - * This is used both when we started the grace period and when we notice - * that someone else started the grace period. The caller must hold the - * ->lock of the leaf rcu_node structure corresponding to the current CPU, - * and must have irqs disabled. - */ -static void __note_new_gpnum(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp) -{ - if (rdp->gpnum != rnp->gpnum) { - /* - * If the current grace period is waiting for this CPU, - * set up to detect a quiescent state, otherwise don't - * go looking for one. - */ - rdp->gpnum = rnp->gpnum; - trace_rcu_grace_period(rsp->name, rdp->gpnum, "cpustart"); - rdp->passed_quiesce = 0; - rdp->qs_pending = !!(rnp->qsmask & rdp->grpmask); - zero_cpu_stall_ticks(rdp); - } -} - -static void note_new_gpnum(struct rcu_state *rsp, struct rcu_data *rdp) -{ - unsigned long flags; - struct rcu_node *rnp; - - local_irq_save(flags); - rnp = rdp->mynode; - if (rdp->gpnum == ACCESS_ONCE(rnp->gpnum) || /* outside lock. */ - !raw_spin_trylock(&rnp->lock)) { /* irqs already off, so later. */ - local_irq_restore(flags); - return; - } - __note_new_gpnum(rsp, rnp, rdp); - raw_spin_unlock_irqrestore(&rnp->lock, flags); -} - -/* - * Did someone else start a new RCU grace period start since we last - * checked? Update local state appropriately if so. Must be called - * on the CPU corresponding to rdp. - */ -static int -check_for_new_grace_period(struct rcu_state *rsp, struct rcu_data *rdp) -{ - unsigned long flags; - int ret = 0; - - local_irq_save(flags); - if (rdp->gpnum != rsp->gpnum) { - note_new_gpnum(rsp, rdp); - ret = 1; - } - local_irq_restore(flags); - return ret; -} - /* * Initialize the specified rcu_data structure's callback list to empty. */ @@ -1359,6 +1300,45 @@ __rcu_process_gp_end(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_dat } } +/* + * Update CPU-local rcu_data state to record the newly noticed grace period. + * This is used both when we started the grace period and when we notice + * that someone else started the grace period. The caller must hold the + * ->lock of the leaf rcu_node structure corresponding to the current CPU, + * and must have irqs disabled. + */ +static void __note_new_gpnum(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp) +{ + if (rdp->gpnum != rnp->gpnum) { + /* + * If the current grace period is waiting for this CPU, + * set up to detect a quiescent state, otherwise don't + * go looking for one. + */ + rdp->gpnum = rnp->gpnum; + trace_rcu_grace_period(rsp->name, rdp->gpnum, "cpustart"); + rdp->passed_quiesce = 0; + rdp->qs_pending = !!(rnp->qsmask & rdp->grpmask); + zero_cpu_stall_ticks(rdp); + } +} + +static void note_new_gpnum(struct rcu_state *rsp, struct rcu_data *rdp) +{ + unsigned long flags; + struct rcu_node *rnp; + + local_irq_save(flags); + rnp = rdp->mynode; + if (rdp->gpnum == ACCESS_ONCE(rnp->gpnum) || /* outside lock. */ + !raw_spin_trylock(&rnp->lock)) { /* irqs already off, so later. */ + local_irq_restore(flags); + return; + } + __note_new_gpnum(rsp, rnp, rdp); + raw_spin_unlock_irqrestore(&rnp->lock, flags); +} + /* * Advance this CPU's callbacks, but only if the current grace period * has ended. This may be called only from the CPU to whom the rdp @@ -1381,6 +1361,26 @@ rcu_process_gp_end(struct rcu_state *rsp, struct rcu_data *rdp) raw_spin_unlock_irqrestore(&rnp->lock, flags); } +/* + * Did someone else start a new RCU grace period start since we last + * checked? Update local state appropriately if so. Must be called + * on the CPU corresponding to rdp. + */ +static int +check_for_new_grace_period(struct rcu_state *rsp, struct rcu_data *rdp) +{ + unsigned long flags; + int ret = 0; + + local_irq_save(flags); + if (rdp->gpnum != rsp->gpnum) { + note_new_gpnum(rsp, rdp); + ret = 1; + } + local_irq_restore(flags); + return ret; +} + /* * Do per-CPU grace-period initialization for running CPU. The caller * must hold the lock of the leaf rcu_node structure corresponding to -- cgit v1.2.3 From 398ebe6000c16135d12ce2ff64318f306ffb20b0 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Tue, 19 Mar 2013 10:53:14 -0700 Subject: rcu: Make __note_new_gpnum() check for ends of prior grace periods The current implementation can detect the beginning of a new grace period before noting the end of a previous grace period. Although the current implementation correctly handles this sort of nonsense, it would be good to reduce RCU's state space by making such nonsense unnecessary, which is now possible thanks to the fact that RCU's callback groups are now numbered. This commit therefore makes __note_new_gpnum() invoke __rcu_process_gp_end() in order to note the ends of prior grace periods before noting the beginnings of new grace periods. Of course, this now means that note_new_gpnum() notes both the beginnings and ends of grace periods, and could therefore be used in place of rcu_process_gp_end(). But that is a job for later commits. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutree.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index c36e52dc091d..54aba759b609 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1309,6 +1309,9 @@ __rcu_process_gp_end(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_dat */ static void __note_new_gpnum(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp) { + /* Handle the ends of any preceding grace periods first. */ + __rcu_process_gp_end(rsp, rnp, rdp); + if (rdp->gpnum != rnp->gpnum) { /* * If the current grace period is waiting for this CPU, -- cgit v1.2.3 From d34ea3221a0f34ed42eadabf054604bbcc7ecd27 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Tue, 19 Mar 2013 11:10:43 -0700 Subject: rcu: Rename note_new_gpnum() to note_gp_changes() Because note_new_gpnum() now also checks for the ends of old grace periods, this commit changes its name to note_gp_changes(). Later commits will merge rcu_process_gp_end() into note_gp_changes(). Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutree.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 54aba759b609..7eb2bc95300a 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1307,7 +1307,7 @@ __rcu_process_gp_end(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_dat * ->lock of the leaf rcu_node structure corresponding to the current CPU, * and must have irqs disabled. */ -static void __note_new_gpnum(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp) +static void __note_gp_changes(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp) { /* Handle the ends of any preceding grace periods first. */ __rcu_process_gp_end(rsp, rnp, rdp); @@ -1326,19 +1326,20 @@ static void __note_new_gpnum(struct rcu_state *rsp, struct rcu_node *rnp, struct } } -static void note_new_gpnum(struct rcu_state *rsp, struct rcu_data *rdp) +static void note_gp_changes(struct rcu_state *rsp, struct rcu_data *rdp) { unsigned long flags; struct rcu_node *rnp; local_irq_save(flags); rnp = rdp->mynode; - if (rdp->gpnum == ACCESS_ONCE(rnp->gpnum) || /* outside lock. */ + if ((rdp->gpnum == ACCESS_ONCE(rnp->gpnum) && + rdp->completed == ACCESS_ONCE(rnp->completed)) || /* w/out lock. */ !raw_spin_trylock(&rnp->lock)) { /* irqs already off, so later. */ local_irq_restore(flags); return; } - __note_new_gpnum(rsp, rnp, rdp); + __note_gp_changes(rsp, rnp, rdp); raw_spin_unlock_irqrestore(&rnp->lock, flags); } @@ -1377,7 +1378,7 @@ check_for_new_grace_period(struct rcu_state *rsp, struct rcu_data *rdp) local_irq_save(flags); if (rdp->gpnum != rsp->gpnum) { - note_new_gpnum(rsp, rdp); + note_gp_changes(rsp, rdp); ret = 1; } local_irq_restore(flags); @@ -1396,7 +1397,7 @@ rcu_start_gp_per_cpu(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_dat __rcu_process_gp_end(rsp, rnp, rdp); /* Set state so that this CPU will detect the next quiescent state. */ - __note_new_gpnum(rsp, rnp, rdp); + __note_gp_changes(rsp, rnp, rdp); } /* -- cgit v1.2.3 From efc151c33b971148894789dc7c5589dec46d4348 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 18 Mar 2013 16:24:11 -0700 Subject: rcu: Convert rcutree_plugin.h printk calls This commit converts printk() calls to the corresponding pr_*() calls. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutree_plugin.h | 45 ++++++++++++++++++++++----------------------- 1 file changed, 22 insertions(+), 23 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 3db5a375d8dd..207844ea0226 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -53,38 +53,37 @@ static char __initdata nocb_buf[NR_CPUS * 5]; static void __init rcu_bootup_announce_oddness(void) { #ifdef CONFIG_RCU_TRACE - printk(KERN_INFO "\tRCU debugfs-based tracing is enabled.\n"); + pr_info("\tRCU debugfs-based tracing is enabled.\n"); #endif #if (defined(CONFIG_64BIT) && CONFIG_RCU_FANOUT != 64) || (!defined(CONFIG_64BIT) && CONFIG_RCU_FANOUT != 32) - printk(KERN_INFO "\tCONFIG_RCU_FANOUT set to non-default value of %d\n", + pr_info("\tCONFIG_RCU_FANOUT set to non-default value of %d\n", CONFIG_RCU_FANOUT); #endif #ifdef CONFIG_RCU_FANOUT_EXACT - printk(KERN_INFO "\tHierarchical RCU autobalancing is disabled.\n"); + pr_info("\tHierarchical RCU autobalancing is disabled.\n"); #endif #ifdef CONFIG_RCU_FAST_NO_HZ - printk(KERN_INFO - "\tRCU dyntick-idle grace-period acceleration is enabled.\n"); + pr_info("\tRCU dyntick-idle grace-period acceleration is enabled.\n"); #endif #ifdef CONFIG_PROVE_RCU - printk(KERN_INFO "\tRCU lockdep checking is enabled.\n"); + pr_info("\tRCU lockdep checking is enabled.\n"); #endif #ifdef CONFIG_RCU_TORTURE_TEST_RUNNABLE - printk(KERN_INFO "\tRCU torture testing starts during boot.\n"); + pr_info("\tRCU torture testing starts during boot.\n"); #endif #if defined(CONFIG_TREE_PREEMPT_RCU) && !defined(CONFIG_RCU_CPU_STALL_VERBOSE) - printk(KERN_INFO "\tDump stacks of tasks blocking RCU-preempt GP.\n"); + pr_info("\tDump stacks of tasks blocking RCU-preempt GP.\n"); #endif #if defined(CONFIG_RCU_CPU_STALL_INFO) - printk(KERN_INFO "\tAdditional per-CPU info printed with stalls.\n"); + pr_info("\tAdditional per-CPU info printed with stalls.\n"); #endif #if NUM_RCU_LVL_4 != 0 - printk(KERN_INFO "\tFour-level hierarchy is enabled.\n"); + pr_info("\tFour-level hierarchy is enabled.\n"); #endif if (rcu_fanout_leaf != CONFIG_RCU_FANOUT_LEAF) - printk(KERN_INFO "\tExperimental boot-time adjustment of leaf fanout to %d.\n", rcu_fanout_leaf); + pr_info("\tExperimental boot-time adjustment of leaf fanout to %d.\n", rcu_fanout_leaf); if (nr_cpu_ids != NR_CPUS) - printk(KERN_INFO "\tRCU restricting CPUs from NR_CPUS=%d to nr_cpu_ids=%d.\n", NR_CPUS, nr_cpu_ids); + pr_info("\tRCU restricting CPUs from NR_CPUS=%d to nr_cpu_ids=%d.\n", NR_CPUS, nr_cpu_ids); #ifdef CONFIG_RCU_NOCB_CPU #ifndef CONFIG_RCU_NOCB_CPU_NONE if (!have_rcu_nocb_mask) { @@ -123,7 +122,7 @@ static int rcu_preempted_readers_exp(struct rcu_node *rnp); */ static void __init rcu_bootup_announce(void) { - printk(KERN_INFO "Preemptible hierarchical RCU implementation.\n"); + pr_info("Preemptible hierarchical RCU implementation.\n"); rcu_bootup_announce_oddness(); } @@ -490,13 +489,13 @@ static void rcu_print_detail_task_stall(struct rcu_state *rsp) static void rcu_print_task_stall_begin(struct rcu_node *rnp) { - printk(KERN_ERR "\tTasks blocked on level-%d rcu_node (CPUs %d-%d):", + pr_err("\tTasks blocked on level-%d rcu_node (CPUs %d-%d):", rnp->level, rnp->grplo, rnp->grphi); } static void rcu_print_task_stall_end(void) { - printk(KERN_CONT "\n"); + pr_cont("\n"); } #else /* #ifdef CONFIG_RCU_CPU_STALL_INFO */ @@ -526,7 +525,7 @@ static int rcu_print_task_stall(struct rcu_node *rnp) t = list_entry(rnp->gp_tasks, struct task_struct, rcu_node_entry); list_for_each_entry_continue(t, &rnp->blkd_tasks, rcu_node_entry) { - printk(KERN_CONT " P%d", t->pid); + pr_cont(" P%d", t->pid); ndetected++; } rcu_print_task_stall_end(); @@ -942,7 +941,7 @@ static struct rcu_state *rcu_state = &rcu_sched_state; */ static void __init rcu_bootup_announce(void) { - printk(KERN_INFO "Hierarchical RCU implementation.\n"); + pr_info("Hierarchical RCU implementation.\n"); rcu_bootup_announce_oddness(); } @@ -1883,7 +1882,7 @@ static void print_cpu_stall_fast_no_hz(char *cp, int cpu) /* Initiate the stall-info list. */ static void print_cpu_stall_info_begin(void) { - printk(KERN_CONT "\n"); + pr_cont("\n"); } /* @@ -1914,7 +1913,7 @@ static void print_cpu_stall_info(struct rcu_state *rsp, int cpu) ticks_value = rsp->gpnum - rdp->gpnum; } print_cpu_stall_fast_no_hz(fast_no_hz, cpu); - printk(KERN_ERR "\t%d: (%lu %s) idle=%03x/%llx/%d softirq=%u/%u %s\n", + pr_err("\t%d: (%lu %s) idle=%03x/%llx/%d softirq=%u/%u %s\n", cpu, ticks_value, ticks_title, atomic_read(&rdtp->dynticks) & 0xfff, rdtp->dynticks_nesting, rdtp->dynticks_nmi_nesting, @@ -1925,7 +1924,7 @@ static void print_cpu_stall_info(struct rcu_state *rsp, int cpu) /* Terminate the stall-info list. */ static void print_cpu_stall_info_end(void) { - printk(KERN_ERR "\t"); + pr_err("\t"); } /* Zero ->ticks_this_gp for all flavors of RCU. */ @@ -1948,17 +1947,17 @@ static void increment_cpu_stall_ticks(void) static void print_cpu_stall_info_begin(void) { - printk(KERN_CONT " {"); + pr_cont(" {"); } static void print_cpu_stall_info(struct rcu_state *rsp, int cpu) { - printk(KERN_CONT " %d", cpu); + pr_cont(" %d", cpu); } static void print_cpu_stall_info_end(void) { - printk(KERN_CONT "} "); + pr_cont("} "); } static void zero_cpu_stall_ticks(struct rcu_data *rdp) -- cgit v1.2.3 From 470716fc043aba2fea832334e58d5cd5d82288a3 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Tue, 19 Mar 2013 11:32:11 -0700 Subject: rcu: Switch callers from rcu_process_gp_end() to note_gp_changes() Because note_gp_changes() now incorporates rcu_process_gp_end() function, this commit switches to the former and eliminates the latter. In addition, this commit changes external calls from __rcu_process_gp_end() to __note_gp_changes(). Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutree.c | 31 +++---------------------------- kernel/rcutree_plugin.h | 2 +- 2 files changed, 4 insertions(+), 29 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 7eb2bc95300a..b04f134ab8bc 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1343,28 +1343,6 @@ static void note_gp_changes(struct rcu_state *rsp, struct rcu_data *rdp) raw_spin_unlock_irqrestore(&rnp->lock, flags); } -/* - * Advance this CPU's callbacks, but only if the current grace period - * has ended. This may be called only from the CPU to whom the rdp - * belongs. - */ -static void -rcu_process_gp_end(struct rcu_state *rsp, struct rcu_data *rdp) -{ - unsigned long flags; - struct rcu_node *rnp; - - local_irq_save(flags); - rnp = rdp->mynode; - if (rdp->completed == ACCESS_ONCE(rnp->completed) || /* outside lock. */ - !raw_spin_trylock(&rnp->lock)) { /* irqs already off, so later. */ - local_irq_restore(flags); - return; - } - __rcu_process_gp_end(rsp, rnp, rdp); - raw_spin_unlock_irqrestore(&rnp->lock, flags); -} - /* * Did someone else start a new RCU grace period start since we last * checked? Update local state appropriately if so. Must be called @@ -1393,9 +1371,6 @@ check_for_new_grace_period(struct rcu_state *rsp, struct rcu_data *rdp) static void rcu_start_gp_per_cpu(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp) { - /* Prior grace period ended, so advance callbacks for current CPU. */ - __rcu_process_gp_end(rsp, rnp, rdp); - /* Set state so that this CPU will detect the next quiescent state. */ __note_gp_changes(rsp, rnp, rdp); } @@ -1531,7 +1506,7 @@ static void rcu_gp_cleanup(struct rcu_state *rsp) ACCESS_ONCE(rnp->completed) = rsp->gpnum; rdp = this_cpu_ptr(rsp->rda); if (rnp == rdp->mynode) - __rcu_process_gp_end(rsp, rnp, rdp); + __note_gp_changes(rsp, rnp, rdp); nocb += rcu_future_gp_cleanup(rsp, rnp); raw_spin_unlock_irq(&rnp->lock); cond_resched(); @@ -2276,7 +2251,7 @@ __rcu_process_callbacks(struct rcu_state *rsp) WARN_ON_ONCE(rdp->beenonline == 0); /* Handle the end of a grace period that some other CPU ended. */ - rcu_process_gp_end(rsp, rdp); + note_gp_changes(rsp, rdp); /* Update RCU state based on any recent quiescent states. */ rcu_check_quiescent_state(rsp, rdp); @@ -2362,7 +2337,7 @@ static void __call_rcu_core(struct rcu_state *rsp, struct rcu_data *rdp, if (unlikely(rdp->qlen > rdp->qlen_last_fqs_check + qhimark)) { /* Are we ignoring a completed grace period? */ - rcu_process_gp_end(rsp, rdp); + note_gp_changes(rsp, rdp); check_for_new_grace_period(rsp, rdp); /* Start a new grace period if one not already started. */ diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 207844ea0226..f279148a0168 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -1628,7 +1628,7 @@ static bool rcu_try_advance_all_cbs(void) */ if (rdp->completed != rnp->completed && rdp->nxttail[RCU_DONE_TAIL] != rdp->nxttail[RCU_NEXT_TAIL]) - rcu_process_gp_end(rsp, rdp); + note_gp_changes(rsp, rdp); if (cpu_has_callbacks_ready_to_invoke(rdp)) cbs_ready = true; -- cgit v1.2.3 From ba9fbe955f026780e6b27c279dba7c86dfdcb7d5 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Tue, 19 Mar 2013 11:53:31 -0700 Subject: rcu: Merge __rcu_process_gp_end() into __note_gp_changes() This commit eliminates some duplicated code by merging __rcu_process_gp_end() into __note_gp_changes(). Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutree.c | 48 ++++++------------------------------------------ 1 file changed, 6 insertions(+), 42 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index b04f134ab8bc..ac8f03c41476 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1254,18 +1254,16 @@ static void rcu_advance_cbs(struct rcu_state *rsp, struct rcu_node *rnp, } /* - * Advance this CPU's callbacks, but only if the current grace period - * has ended. This may be called only from the CPU to whom the rdp - * belongs. In addition, the corresponding leaf rcu_node structure's - * ->lock must be held by the caller, with irqs disabled. + * Update CPU-local rcu_data state to record the beginnings and ends of + * grace periods. The caller must hold the ->lock of the leaf rcu_node + * structure corresponding to the current CPU, and must have irqs disabled. */ -static void -__rcu_process_gp_end(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp) +static void __note_gp_changes(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp) { - /* Did another grace period end? */ + /* Handle the ends of any preceding grace periods first. */ if (rdp->completed == rnp->completed) { - /* No, so just accelerate recent callbacks. */ + /* No grace period end, so just accelerate recent callbacks. */ rcu_accelerate_cbs(rsp, rnp, rdp); } else { @@ -1276,41 +1274,7 @@ __rcu_process_gp_end(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_dat /* Remember that we saw this grace-period completion. */ rdp->completed = rnp->completed; trace_rcu_grace_period(rsp->name, rdp->gpnum, "cpuend"); - - /* - * If we were in an extended quiescent state, we may have - * missed some grace periods that others CPUs handled on - * our behalf. Catch up with this state to avoid noting - * spurious new grace periods. If another grace period - * has started, then rnp->gpnum will have advanced, so - * we will detect this later on. Of course, any quiescent - * states we found for the old GP are now invalid. - */ - if (ULONG_CMP_LT(rdp->gpnum, rdp->completed)) { - rdp->gpnum = rdp->completed; - rdp->passed_quiesce = 0; - } - - /* - * If RCU does not need a quiescent state from this CPU, - * then make sure that this CPU doesn't go looking for one. - */ - if ((rnp->qsmask & rdp->grpmask) == 0) - rdp->qs_pending = 0; } -} - -/* - * Update CPU-local rcu_data state to record the newly noticed grace period. - * This is used both when we started the grace period and when we notice - * that someone else started the grace period. The caller must hold the - * ->lock of the leaf rcu_node structure corresponding to the current CPU, - * and must have irqs disabled. - */ -static void __note_gp_changes(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp) -{ - /* Handle the ends of any preceding grace periods first. */ - __rcu_process_gp_end(rsp, rnp, rdp); if (rdp->gpnum != rnp->gpnum) { /* -- cgit v1.2.3 From 63274cfb94aac109fc2490a70a96b26751608e57 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Tue, 19 Mar 2013 12:21:29 -0700 Subject: rcu: Eliminate check_for_new_grace_period() wrapper function One of the calls to check_for_new_grace_period() is now redundant due to an immediately preceding call to note_gp_changes(). Eliminating this redundant call leaves a single caller, which is simpler if inlined. This commit therefore eliminates the redundant call and inlines the body of check_for_new_grace_period() into the single remaining call site. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutree.c | 25 +++---------------------- 1 file changed, 3 insertions(+), 22 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index ac8f03c41476..b73014998b40 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1307,26 +1307,6 @@ static void note_gp_changes(struct rcu_state *rsp, struct rcu_data *rdp) raw_spin_unlock_irqrestore(&rnp->lock, flags); } -/* - * Did someone else start a new RCU grace period start since we last - * checked? Update local state appropriately if so. Must be called - * on the CPU corresponding to rdp. - */ -static int -check_for_new_grace_period(struct rcu_state *rsp, struct rcu_data *rdp) -{ - unsigned long flags; - int ret = 0; - - local_irq_save(flags); - if (rdp->gpnum != rsp->gpnum) { - note_gp_changes(rsp, rdp); - ret = 1; - } - local_irq_restore(flags); - return ret; -} - /* * Do per-CPU grace-period initialization for running CPU. The caller * must hold the lock of the leaf rcu_node structure corresponding to @@ -1749,8 +1729,10 @@ static void rcu_check_quiescent_state(struct rcu_state *rsp, struct rcu_data *rdp) { /* If there is now a new grace period, record and return. */ - if (check_for_new_grace_period(rsp, rdp)) + if (rdp->gpnum != rsp->gpnum) { + note_gp_changes(rsp, rdp); return; + } /* * Does this CPU still need to do its part for current grace period? @@ -2302,7 +2284,6 @@ static void __call_rcu_core(struct rcu_state *rsp, struct rcu_data *rdp, /* Are we ignoring a completed grace period? */ note_gp_changes(rsp, rdp); - check_for_new_grace_period(rsp, rdp); /* Start a new grace period if one not already started. */ if (!rcu_gp_in_progress(rsp)) { -- cgit v1.2.3 From ce3d9c03d1fa079678cc8df1517011e215517cda Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Tue, 19 Mar 2013 12:27:50 -0700 Subject: rcu: Inline trivial wrapper function rcu_start_gp_per_cpu() Given the changes that introduce note_gp_change(), rcu_start_gp_per_cpu() is now a trivial wrapper function with only one caller. This commit therefore inlines it into its sole call site. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutree.c | 14 +------------- 1 file changed, 1 insertion(+), 13 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index b73014998b40..391bd724cd77 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1307,18 +1307,6 @@ static void note_gp_changes(struct rcu_state *rsp, struct rcu_data *rdp) raw_spin_unlock_irqrestore(&rnp->lock, flags); } -/* - * Do per-CPU grace-period initialization for running CPU. The caller - * must hold the lock of the leaf rcu_node structure corresponding to - * this CPU. - */ -static void -rcu_start_gp_per_cpu(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp) -{ - /* Set state so that this CPU will detect the next quiescent state. */ - __note_gp_changes(rsp, rnp, rdp); -} - /* * Initialize a new grace period. */ @@ -1367,7 +1355,7 @@ static int rcu_gp_init(struct rcu_state *rsp) WARN_ON_ONCE(rnp->completed != rsp->completed); ACCESS_ONCE(rnp->completed) = rsp->completed; if (rnp == rdp->mynode) - rcu_start_gp_per_cpu(rsp, rnp, rdp); + __note_gp_changes(rsp, rnp, rdp); rcu_preempt_boost_start_gp(rnp); trace_rcu_grace_period_init(rsp->name, rnp->gpnum, rnp->level, rnp->grplo, -- cgit v1.2.3 From 05eb552bf5ed9e7277bdc9c273ed2f4e9b7dc3e5 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Tue, 19 Mar 2013 12:38:24 -0700 Subject: rcu: Move redundant call to note_gp_changes() into called function The __rcu_process_callbacks() invokes note_gp_changes() immediately before invoking rcu_check_quiescent_state(), which conditionally invokes that same function. This commit therefore eliminates the call to note_gp_changes() in __rcu_process_callbacks() in favor of making unconditional to call from rcu_check_quiescent_state() to note_gp_changes(). Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutree.c | 10 ++-------- 1 file changed, 2 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 391bd724cd77..7a5194ef90da 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1716,11 +1716,8 @@ rcu_report_qs_rdp(int cpu, struct rcu_state *rsp, struct rcu_data *rdp) static void rcu_check_quiescent_state(struct rcu_state *rsp, struct rcu_data *rdp) { - /* If there is now a new grace period, record and return. */ - if (rdp->gpnum != rsp->gpnum) { - note_gp_changes(rsp, rdp); - return; - } + /* Check for grace-period ends and beginnings. */ + note_gp_changes(rsp, rdp); /* * Does this CPU still need to do its part for current grace period? @@ -2184,9 +2181,6 @@ __rcu_process_callbacks(struct rcu_state *rsp) WARN_ON_ONCE(rdp->beenonline == 0); - /* Handle the end of a grace period that some other CPU ended. */ - note_gp_changes(rsp, rdp); - /* Update RCU state based on any recent quiescent states. */ rcu_check_quiescent_state(rsp, rdp); -- cgit v1.2.3 From 9a5739d73f9369ba1cdba3889ee4e2f87be25a46 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 28 Mar 2013 20:48:36 -0700 Subject: rcu: Remove "Experimental" flags After a release or two, features are no longer experimental. Therefore, this commit removes the "Experimental" tag from them. Reported-by: Paul Gortmaker Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutree_plugin.h | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 207844ea0226..6b3ccaae93ab 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -81,7 +81,7 @@ static void __init rcu_bootup_announce_oddness(void) pr_info("\tFour-level hierarchy is enabled.\n"); #endif if (rcu_fanout_leaf != CONFIG_RCU_FANOUT_LEAF) - pr_info("\tExperimental boot-time adjustment of leaf fanout to %d.\n", rcu_fanout_leaf); + pr_info("\tBoot-time adjustment of leaf fanout to %d.\n", rcu_fanout_leaf); if (nr_cpu_ids != NR_CPUS) pr_info("\tRCU restricting CPUs from NR_CPUS=%d to nr_cpu_ids=%d.\n", NR_CPUS, nr_cpu_ids); #ifdef CONFIG_RCU_NOCB_CPU @@ -91,19 +91,19 @@ static void __init rcu_bootup_announce_oddness(void) have_rcu_nocb_mask = true; } #ifdef CONFIG_RCU_NOCB_CPU_ZERO - pr_info("\tExperimental no-CBs CPU 0\n"); + pr_info("\tOffload RCU callbacks from CPU 0\n"); cpumask_set_cpu(0, rcu_nocb_mask); #endif /* #ifdef CONFIG_RCU_NOCB_CPU_ZERO */ #ifdef CONFIG_RCU_NOCB_CPU_ALL - pr_info("\tExperimental no-CBs for all CPUs\n"); + pr_info("\tOffload RCU callbacks from all CPUs\n"); cpumask_setall(rcu_nocb_mask); #endif /* #ifdef CONFIG_RCU_NOCB_CPU_ALL */ #endif /* #ifndef CONFIG_RCU_NOCB_CPU_NONE */ if (have_rcu_nocb_mask) { cpulist_scnprintf(nocb_buf, sizeof(nocb_buf), rcu_nocb_mask); - pr_info("\tExperimental no-CBs CPUs: %s.\n", nocb_buf); + pr_info("\tOffload RCU callbacks from CPUs: %s.\n", nocb_buf); if (rcu_nocb_poll) - pr_info("\tExperimental polled no-CBs CPUs.\n"); + pr_info("\tPoll for callbacks from no-CBs CPUs.\n"); } #endif /* #ifdef CONFIG_RCU_NOCB_CPU */ } -- cgit v1.2.3 From 026ad2835ce6202069e7aa0b11f5f1be4de34550 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 3 Apr 2013 22:14:11 -0700 Subject: rcu: Drive quiescent-state-forcing delay from HZ Systems with HZ=100 can have slow bootup times due to the default three-jiffy delays between quiescent-state forcing attempts. This commit therefore auto-tunes the RCU_JIFFIES_TILL_FORCE_QS value based on the value of HZ. However, this would break very large systems that require more time between quiescent-state forcing attempts. This commit therefore also ups the default delay by one jiffy for each 256 CPUs that might be on the system (based off of nr_cpu_ids at runtime, -not- NR_CPUS at build time). Updated to collapse #ifdefs for RCU_JIFFIES_TILL_FORCE_QS into a step-function definition as suggested by Josh Triplett. Reported-by: Paul Mackerras Signed-off-by: Paul E. McKenney --- kernel/rcutree.c | 18 ++++++++++++++++-- kernel/rcutree.h | 15 ++++++++++----- 2 files changed, 26 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 1009c0ccd4b1..f344d3c824a4 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -218,8 +218,8 @@ module_param(blimit, long, 0444); module_param(qhimark, long, 0444); module_param(qlowmark, long, 0444); -static ulong jiffies_till_first_fqs = RCU_JIFFIES_TILL_FORCE_QS; -static ulong jiffies_till_next_fqs = RCU_JIFFIES_TILL_FORCE_QS; +static ulong jiffies_till_first_fqs = ULONG_MAX; +static ulong jiffies_till_next_fqs = ULONG_MAX; module_param(jiffies_till_first_fqs, ulong, 0644); module_param(jiffies_till_next_fqs, ulong, 0644); @@ -3265,11 +3265,25 @@ static void __init rcu_init_one(struct rcu_state *rsp, */ static void __init rcu_init_geometry(void) { + ulong d; int i; int j; int n = nr_cpu_ids; int rcu_capacity[MAX_RCU_LVLS + 1]; + /* + * Initialize any unspecified boot parameters. + * The default values of jiffies_till_first_fqs and + * jiffies_till_next_fqs are set to the RCU_JIFFIES_TILL_FORCE_QS + * value, which is a function of HZ, then adding one for each + * RCU_JIFFIES_FQS_DIV CPUs that might be on the system. + */ + d = RCU_JIFFIES_TILL_FORCE_QS + nr_cpu_ids / RCU_JIFFIES_FQS_DIV; + if (jiffies_till_first_fqs == ULONG_MAX) + jiffies_till_first_fqs = d; + if (jiffies_till_next_fqs == ULONG_MAX) + jiffies_till_next_fqs = d; + /* If the compile-time values are accurate, just leave. */ if (rcu_fanout_leaf == CONFIG_RCU_FANOUT_LEAF && nr_cpu_ids == NR_CPUS) diff --git a/kernel/rcutree.h b/kernel/rcutree.h index 4df503470e42..4a39d364493c 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -343,12 +343,17 @@ struct rcu_data { #define RCU_FORCE_QS 3 /* Need to force quiescent state. */ #define RCU_SIGNAL_INIT RCU_SAVE_DYNTICK -#define RCU_JIFFIES_TILL_FORCE_QS 3 /* for rsp->jiffies_force_qs */ +#define RCU_JIFFIES_TILL_FORCE_QS (1 + (HZ > 250) + (HZ > 500)) + /* For jiffies_till_first_fqs and */ + /* and jiffies_till_next_fqs. */ -#define RCU_STALL_RAT_DELAY 2 /* Allow other CPUs time */ - /* to take at least one */ - /* scheduling clock irq */ - /* before ratting on them. */ +#define RCU_JIFFIES_FQS_DIV 256 /* Very large systems need more */ + /* delay between bouts of */ + /* quiescent-state forcing. */ + +#define RCU_STALL_RAT_DELAY 2 /* Allow other CPUs time to take */ + /* at least one scheduling clock */ + /* irq before ratting on them. */ #define rcu_wait(cond) \ do { \ -- cgit v1.2.3 From 4982969d965ec87b1887c86d2e0b3d81065e1d38 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 28 Mar 2013 16:56:53 -0700 Subject: rcu: Merge adjacent identical ifdefs Two ifdefs in kernel/rcupdate.c now have identical conditions with nothing between them, so the commit merges them into a single ifdef. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcupdate.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'kernel') diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c index 48ab70384a4c..faeea984dbaa 100644 --- a/kernel/rcupdate.c +++ b/kernel/rcupdate.c @@ -145,9 +145,6 @@ static struct lock_class_key rcu_sched_lock_key; struct lockdep_map rcu_sched_lock_map = STATIC_LOCKDEP_MAP_INIT("rcu_read_lock_sched", &rcu_sched_lock_key); EXPORT_SYMBOL_GPL(rcu_sched_lock_map); -#endif - -#ifdef CONFIG_DEBUG_LOCK_ALLOC int debug_lockdep_rcu_enabled(void) { -- cgit v1.2.3 From 99f88919f8fa8a8b01b5306c59c9977b94604df8 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Tue, 12 Mar 2013 16:54:14 -0700 Subject: rcu: Remove srcu_read_lock_raw() and srcu_read_unlock_raw(). These interfaces never did get used, so this commit removes them, their rcutorture tests, and documentation referencing them. Signed-off-by: Paul E. McKenney Reviewed-by: Lai Jiangshan Reviewed-by: Josh Triplett --- kernel/rcutorture.c | 39 --------------------------------------- 1 file changed, 39 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c index e1f3a8c96724..b1fa5510388d 100644 --- a/kernel/rcutorture.c +++ b/kernel/rcutorture.c @@ -695,44 +695,6 @@ static struct rcu_torture_ops srcu_sync_ops = { .name = "srcu_sync" }; -static int srcu_torture_read_lock_raw(void) __acquires(&srcu_ctl) -{ - return srcu_read_lock_raw(&srcu_ctl); -} - -static void srcu_torture_read_unlock_raw(int idx) __releases(&srcu_ctl) -{ - srcu_read_unlock_raw(&srcu_ctl, idx); -} - -static struct rcu_torture_ops srcu_raw_ops = { - .init = rcu_sync_torture_init, - .readlock = srcu_torture_read_lock_raw, - .read_delay = srcu_read_delay, - .readunlock = srcu_torture_read_unlock_raw, - .completed = srcu_torture_completed, - .deferred_free = srcu_torture_deferred_free, - .sync = srcu_torture_synchronize, - .call = NULL, - .cb_barrier = NULL, - .stats = srcu_torture_stats, - .name = "srcu_raw" -}; - -static struct rcu_torture_ops srcu_raw_sync_ops = { - .init = rcu_sync_torture_init, - .readlock = srcu_torture_read_lock_raw, - .read_delay = srcu_read_delay, - .readunlock = srcu_torture_read_unlock_raw, - .completed = srcu_torture_completed, - .deferred_free = rcu_sync_torture_deferred_free, - .sync = srcu_torture_synchronize, - .call = NULL, - .cb_barrier = NULL, - .stats = srcu_torture_stats, - .name = "srcu_raw_sync" -}; - static void srcu_torture_synchronize_expedited(void) { synchronize_srcu_expedited(&srcu_ctl); @@ -1983,7 +1945,6 @@ rcu_torture_init(void) { &rcu_ops, &rcu_sync_ops, &rcu_expedited_ops, &rcu_bh_ops, &rcu_bh_sync_ops, &rcu_bh_expedited_ops, &srcu_ops, &srcu_sync_ops, &srcu_expedited_ops, - &srcu_raw_ops, &srcu_raw_sync_ops, &sched_ops, &sched_sync_ops, &sched_expedited_ops, }; mutex_lock(&fullstop_mutex); -- cgit v1.2.3 From 127781d1ba1ee5bbe1780afa35dd0e71583b143d Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 27 Mar 2013 08:44:00 -0700 Subject: rcu: Remove TINY_PREEMPT_RCU TINY_PREEMPT_RCU adds significant code and complexity, but does not offer commensurate benefits. People currently using TINY_PREEMPT_RCU can get much better memory footprint with TINY_RCU, or, if they really need preemptible RCU, they can use TREE_PREEMPT_RCU with a relatively minor degradation in memory footprint. Please note that this move has been widely publicized on LKML (https://lkml.org/lkml/2012/11/12/545) and on LWN (http://lwn.net/Articles/541037/). This commit therefore removes TINY_PREEMPT_RCU. Signed-off-by: Paul E. McKenney [ paulmck: Updated to eliminate #else in rcutiny.h as suggested by Josh ] Reviewed-by: Josh Triplett --- kernel/rcutiny_plugin.h | 854 ------------------------------------------------ 1 file changed, 854 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutiny_plugin.h b/kernel/rcutiny_plugin.h index 8a233002faeb..29a4dd78c8bf 100644 --- a/kernel/rcutiny_plugin.h +++ b/kernel/rcutiny_plugin.h @@ -102,763 +102,6 @@ static void check_cpu_stalls(void) RCU_TRACE(check_cpu_stall_preempt()); } -#ifdef CONFIG_TINY_PREEMPT_RCU - -#include - -/* Global control variables for preemptible RCU. */ -struct rcu_preempt_ctrlblk { - struct rcu_ctrlblk rcb; /* curtail: ->next ptr of last CB for GP. */ - struct rcu_head **nexttail; - /* Tasks blocked in a preemptible RCU */ - /* read-side critical section while an */ - /* preemptible-RCU grace period is in */ - /* progress must wait for a later grace */ - /* period. This pointer points to the */ - /* ->next pointer of the last task that */ - /* must wait for a later grace period, or */ - /* to &->rcb.rcucblist if there is no */ - /* such task. */ - struct list_head blkd_tasks; - /* Tasks blocked in RCU read-side critical */ - /* section. Tasks are placed at the head */ - /* of this list and age towards the tail. */ - struct list_head *gp_tasks; - /* Pointer to the first task blocking the */ - /* current grace period, or NULL if there */ - /* is no such task. */ - struct list_head *exp_tasks; - /* Pointer to first task blocking the */ - /* current expedited grace period, or NULL */ - /* if there is no such task. If there */ - /* is no current expedited grace period, */ - /* then there cannot be any such task. */ -#ifdef CONFIG_RCU_BOOST - struct list_head *boost_tasks; - /* Pointer to first task that needs to be */ - /* priority-boosted, or NULL if no priority */ - /* boosting is needed. If there is no */ - /* current or expedited grace period, there */ - /* can be no such task. */ -#endif /* #ifdef CONFIG_RCU_BOOST */ - u8 gpnum; /* Current grace period. */ - u8 gpcpu; /* Last grace period blocked by the CPU. */ - u8 completed; /* Last grace period completed. */ - /* If all three are equal, RCU is idle. */ -#ifdef CONFIG_RCU_BOOST - unsigned long boost_time; /* When to start boosting (jiffies) */ -#endif /* #ifdef CONFIG_RCU_BOOST */ -#ifdef CONFIG_RCU_TRACE - unsigned long n_grace_periods; -#ifdef CONFIG_RCU_BOOST - unsigned long n_tasks_boosted; - /* Total number of tasks boosted. */ - unsigned long n_exp_boosts; - /* Number of tasks boosted for expedited GP. */ - unsigned long n_normal_boosts; - /* Number of tasks boosted for normal GP. */ - unsigned long n_balk_blkd_tasks; - /* Refused to boost: no blocked tasks. */ - unsigned long n_balk_exp_gp_tasks; - /* Refused to boost: nothing blocking GP. */ - unsigned long n_balk_boost_tasks; - /* Refused to boost: already boosting. */ - unsigned long n_balk_notyet; - /* Refused to boost: not yet time. */ - unsigned long n_balk_nos; - /* Refused to boost: not sure why, though. */ - /* This can happen due to race conditions. */ -#endif /* #ifdef CONFIG_RCU_BOOST */ -#endif /* #ifdef CONFIG_RCU_TRACE */ -}; - -static struct rcu_preempt_ctrlblk rcu_preempt_ctrlblk = { - .rcb.donetail = &rcu_preempt_ctrlblk.rcb.rcucblist, - .rcb.curtail = &rcu_preempt_ctrlblk.rcb.rcucblist, - .nexttail = &rcu_preempt_ctrlblk.rcb.rcucblist, - .blkd_tasks = LIST_HEAD_INIT(rcu_preempt_ctrlblk.blkd_tasks), - RCU_TRACE(.rcb.name = "rcu_preempt") -}; - -static int rcu_preempted_readers_exp(void); -static void rcu_report_exp_done(void); - -/* - * Return true if the CPU has not yet responded to the current grace period. - */ -static int rcu_cpu_blocking_cur_gp(void) -{ - return rcu_preempt_ctrlblk.gpcpu != rcu_preempt_ctrlblk.gpnum; -} - -/* - * Check for a running RCU reader. Because there is only one CPU, - * there can be but one running RCU reader at a time. ;-) - * - * Returns zero if there are no running readers. Returns a positive - * number if there is at least one reader within its RCU read-side - * critical section. Returns a negative number if an outermost reader - * is in the midst of exiting from its RCU read-side critical section - * - * Returns zero if there are no running readers. Returns a positive - * number if there is at least one reader within its RCU read-side - * critical section. Returns a negative number if an outermost reader - * is in the midst of exiting from its RCU read-side critical section. - */ -static int rcu_preempt_running_reader(void) -{ - return current->rcu_read_lock_nesting; -} - -/* - * Check for preempted RCU readers blocking any grace period. - * If the caller needs a reliable answer, it must disable hard irqs. - */ -static int rcu_preempt_blocked_readers_any(void) -{ - return !list_empty(&rcu_preempt_ctrlblk.blkd_tasks); -} - -/* - * Check for preempted RCU readers blocking the current grace period. - * If the caller needs a reliable answer, it must disable hard irqs. - */ -static int rcu_preempt_blocked_readers_cgp(void) -{ - return rcu_preempt_ctrlblk.gp_tasks != NULL; -} - -/* - * Return true if another preemptible-RCU grace period is needed. - */ -static int rcu_preempt_needs_another_gp(void) -{ - return *rcu_preempt_ctrlblk.rcb.curtail != NULL; -} - -/* - * Return true if a preemptible-RCU grace period is in progress. - * The caller must disable hardirqs. - */ -static int rcu_preempt_gp_in_progress(void) -{ - return rcu_preempt_ctrlblk.completed != rcu_preempt_ctrlblk.gpnum; -} - -/* - * Advance a ->blkd_tasks-list pointer to the next entry, instead - * returning NULL if at the end of the list. - */ -static struct list_head *rcu_next_node_entry(struct task_struct *t) -{ - struct list_head *np; - - np = t->rcu_node_entry.next; - if (np == &rcu_preempt_ctrlblk.blkd_tasks) - np = NULL; - return np; -} - -#ifdef CONFIG_RCU_TRACE - -#ifdef CONFIG_RCU_BOOST -static void rcu_initiate_boost_trace(void); -#endif /* #ifdef CONFIG_RCU_BOOST */ - -/* - * Dump additional statistice for TINY_PREEMPT_RCU. - */ -static void show_tiny_preempt_stats(struct seq_file *m) -{ - seq_printf(m, "rcu_preempt: qlen=%ld gp=%lu g%u/p%u/c%u tasks=%c%c%c\n", - rcu_preempt_ctrlblk.rcb.qlen, - rcu_preempt_ctrlblk.n_grace_periods, - rcu_preempt_ctrlblk.gpnum, - rcu_preempt_ctrlblk.gpcpu, - rcu_preempt_ctrlblk.completed, - "T."[list_empty(&rcu_preempt_ctrlblk.blkd_tasks)], - "N."[!rcu_preempt_ctrlblk.gp_tasks], - "E."[!rcu_preempt_ctrlblk.exp_tasks]); -#ifdef CONFIG_RCU_BOOST - seq_printf(m, "%sttb=%c ntb=%lu neb=%lu nnb=%lu j=%04x bt=%04x\n", - " ", - "B."[!rcu_preempt_ctrlblk.boost_tasks], - rcu_preempt_ctrlblk.n_tasks_boosted, - rcu_preempt_ctrlblk.n_exp_boosts, - rcu_preempt_ctrlblk.n_normal_boosts, - (int)(jiffies & 0xffff), - (int)(rcu_preempt_ctrlblk.boost_time & 0xffff)); - seq_printf(m, "%s: nt=%lu egt=%lu bt=%lu ny=%lu nos=%lu\n", - " balk", - rcu_preempt_ctrlblk.n_balk_blkd_tasks, - rcu_preempt_ctrlblk.n_balk_exp_gp_tasks, - rcu_preempt_ctrlblk.n_balk_boost_tasks, - rcu_preempt_ctrlblk.n_balk_notyet, - rcu_preempt_ctrlblk.n_balk_nos); -#endif /* #ifdef CONFIG_RCU_BOOST */ -} - -#endif /* #ifdef CONFIG_RCU_TRACE */ - -#ifdef CONFIG_RCU_BOOST - -#include "rtmutex_common.h" - -#define RCU_BOOST_PRIO CONFIG_RCU_BOOST_PRIO - -/* Controls for rcu_kthread() kthread. */ -static struct task_struct *rcu_kthread_task; -static DECLARE_WAIT_QUEUE_HEAD(rcu_kthread_wq); -static unsigned long have_rcu_kthread_work; - -/* - * Carry out RCU priority boosting on the task indicated by ->boost_tasks, - * and advance ->boost_tasks to the next task in the ->blkd_tasks list. - */ -static int rcu_boost(void) -{ - unsigned long flags; - struct rt_mutex mtx; - struct task_struct *t; - struct list_head *tb; - - if (rcu_preempt_ctrlblk.boost_tasks == NULL && - rcu_preempt_ctrlblk.exp_tasks == NULL) - return 0; /* Nothing to boost. */ - - local_irq_save(flags); - - /* - * Recheck with irqs disabled: all tasks in need of boosting - * might exit their RCU read-side critical sections on their own - * if we are preempted just before disabling irqs. - */ - if (rcu_preempt_ctrlblk.boost_tasks == NULL && - rcu_preempt_ctrlblk.exp_tasks == NULL) { - local_irq_restore(flags); - return 0; - } - - /* - * Preferentially boost tasks blocking expedited grace periods. - * This cannot starve the normal grace periods because a second - * expedited grace period must boost all blocked tasks, including - * those blocking the pre-existing normal grace period. - */ - if (rcu_preempt_ctrlblk.exp_tasks != NULL) { - tb = rcu_preempt_ctrlblk.exp_tasks; - RCU_TRACE(rcu_preempt_ctrlblk.n_exp_boosts++); - } else { - tb = rcu_preempt_ctrlblk.boost_tasks; - RCU_TRACE(rcu_preempt_ctrlblk.n_normal_boosts++); - } - RCU_TRACE(rcu_preempt_ctrlblk.n_tasks_boosted++); - - /* - * We boost task t by manufacturing an rt_mutex that appears to - * be held by task t. We leave a pointer to that rt_mutex where - * task t can find it, and task t will release the mutex when it - * exits its outermost RCU read-side critical section. Then - * simply acquiring this artificial rt_mutex will boost task - * t's priority. (Thanks to tglx for suggesting this approach!) - */ - t = container_of(tb, struct task_struct, rcu_node_entry); - rt_mutex_init_proxy_locked(&mtx, t); - t->rcu_boost_mutex = &mtx; - local_irq_restore(flags); - rt_mutex_lock(&mtx); - rt_mutex_unlock(&mtx); /* Keep lockdep happy. */ - - return ACCESS_ONCE(rcu_preempt_ctrlblk.boost_tasks) != NULL || - ACCESS_ONCE(rcu_preempt_ctrlblk.exp_tasks) != NULL; -} - -/* - * Check to see if it is now time to start boosting RCU readers blocking - * the current grace period, and, if so, tell the rcu_kthread_task to - * start boosting them. If there is an expedited boost in progress, - * we wait for it to complete. - * - * If there are no blocked readers blocking the current grace period, - * return 0 to let the caller know, otherwise return 1. Note that this - * return value is independent of whether or not boosting was done. - */ -static int rcu_initiate_boost(void) -{ - if (!rcu_preempt_blocked_readers_cgp() && - rcu_preempt_ctrlblk.exp_tasks == NULL) { - RCU_TRACE(rcu_preempt_ctrlblk.n_balk_exp_gp_tasks++); - return 0; - } - if (rcu_preempt_ctrlblk.exp_tasks != NULL || - (rcu_preempt_ctrlblk.gp_tasks != NULL && - rcu_preempt_ctrlblk.boost_tasks == NULL && - ULONG_CMP_GE(jiffies, rcu_preempt_ctrlblk.boost_time))) { - if (rcu_preempt_ctrlblk.exp_tasks == NULL) - rcu_preempt_ctrlblk.boost_tasks = - rcu_preempt_ctrlblk.gp_tasks; - invoke_rcu_callbacks(); - } else { - RCU_TRACE(rcu_initiate_boost_trace()); - } - return 1; -} - -#define RCU_BOOST_DELAY_JIFFIES DIV_ROUND_UP(CONFIG_RCU_BOOST_DELAY * HZ, 1000) - -/* - * Do priority-boost accounting for the start of a new grace period. - */ -static void rcu_preempt_boost_start_gp(void) -{ - rcu_preempt_ctrlblk.boost_time = jiffies + RCU_BOOST_DELAY_JIFFIES; -} - -#else /* #ifdef CONFIG_RCU_BOOST */ - -/* - * If there is no RCU priority boosting, we don't initiate boosting, - * but we do indicate whether there are blocked readers blocking the - * current grace period. - */ -static int rcu_initiate_boost(void) -{ - return rcu_preempt_blocked_readers_cgp(); -} - -/* - * If there is no RCU priority boosting, nothing to do at grace-period start. - */ -static void rcu_preempt_boost_start_gp(void) -{ -} - -#endif /* else #ifdef CONFIG_RCU_BOOST */ - -/* - * Record a preemptible-RCU quiescent state for the specified CPU. Note - * that this just means that the task currently running on the CPU is - * in a quiescent state. There might be any number of tasks blocked - * while in an RCU read-side critical section. - * - * Unlike the other rcu_*_qs() functions, callers to this function - * must disable irqs in order to protect the assignment to - * ->rcu_read_unlock_special. - * - * Because this is a single-CPU implementation, the only way a grace - * period can end is if the CPU is in a quiescent state. The reason is - * that a blocked preemptible-RCU reader can exit its critical section - * only if the CPU is running it at the time. Therefore, when the - * last task blocking the current grace period exits its RCU read-side - * critical section, neither the CPU nor blocked tasks will be stopping - * the current grace period. (In contrast, SMP implementations - * might have CPUs running in RCU read-side critical sections that - * block later grace periods -- but this is not possible given only - * one CPU.) - */ -static void rcu_preempt_cpu_qs(void) -{ - /* Record both CPU and task as having responded to current GP. */ - rcu_preempt_ctrlblk.gpcpu = rcu_preempt_ctrlblk.gpnum; - current->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_NEED_QS; - - /* If there is no GP then there is nothing more to do. */ - if (!rcu_preempt_gp_in_progress()) - return; - /* - * Check up on boosting. If there are readers blocking the - * current grace period, leave. - */ - if (rcu_initiate_boost()) - return; - - /* Advance callbacks. */ - rcu_preempt_ctrlblk.completed = rcu_preempt_ctrlblk.gpnum; - rcu_preempt_ctrlblk.rcb.donetail = rcu_preempt_ctrlblk.rcb.curtail; - rcu_preempt_ctrlblk.rcb.curtail = rcu_preempt_ctrlblk.nexttail; - - /* If there are no blocked readers, next GP is done instantly. */ - if (!rcu_preempt_blocked_readers_any()) - rcu_preempt_ctrlblk.rcb.donetail = rcu_preempt_ctrlblk.nexttail; - - /* If there are done callbacks, cause them to be invoked. */ - if (*rcu_preempt_ctrlblk.rcb.donetail != NULL) - invoke_rcu_callbacks(); -} - -/* - * Start a new RCU grace period if warranted. Hard irqs must be disabled. - */ -static void rcu_preempt_start_gp(void) -{ - if (!rcu_preempt_gp_in_progress() && rcu_preempt_needs_another_gp()) { - - /* Official start of GP. */ - rcu_preempt_ctrlblk.gpnum++; - RCU_TRACE(rcu_preempt_ctrlblk.n_grace_periods++); - reset_cpu_stall_ticks(&rcu_preempt_ctrlblk.rcb); - - /* Any blocked RCU readers block new GP. */ - if (rcu_preempt_blocked_readers_any()) - rcu_preempt_ctrlblk.gp_tasks = - rcu_preempt_ctrlblk.blkd_tasks.next; - - /* Set up for RCU priority boosting. */ - rcu_preempt_boost_start_gp(); - - /* If there is no running reader, CPU is done with GP. */ - if (!rcu_preempt_running_reader()) - rcu_preempt_cpu_qs(); - } -} - -/* - * We have entered the scheduler, and the current task might soon be - * context-switched away from. If this task is in an RCU read-side - * critical section, we will no longer be able to rely on the CPU to - * record that fact, so we enqueue the task on the blkd_tasks list. - * If the task started after the current grace period began, as recorded - * by ->gpcpu, we enqueue at the beginning of the list. Otherwise - * before the element referenced by ->gp_tasks (or at the tail if - * ->gp_tasks is NULL) and point ->gp_tasks at the newly added element. - * The task will dequeue itself when it exits the outermost enclosing - * RCU read-side critical section. Therefore, the current grace period - * cannot be permitted to complete until the ->gp_tasks pointer becomes - * NULL. - * - * Caller must disable preemption. - */ -void rcu_preempt_note_context_switch(void) -{ - struct task_struct *t = current; - unsigned long flags; - - local_irq_save(flags); /* must exclude scheduler_tick(). */ - if (rcu_preempt_running_reader() > 0 && - (t->rcu_read_unlock_special & RCU_READ_UNLOCK_BLOCKED) == 0) { - - /* Possibly blocking in an RCU read-side critical section. */ - t->rcu_read_unlock_special |= RCU_READ_UNLOCK_BLOCKED; - - /* - * If this CPU has already checked in, then this task - * will hold up the next grace period rather than the - * current grace period. Queue the task accordingly. - * If the task is queued for the current grace period - * (i.e., this CPU has not yet passed through a quiescent - * state for the current grace period), then as long - * as that task remains queued, the current grace period - * cannot end. - */ - list_add(&t->rcu_node_entry, &rcu_preempt_ctrlblk.blkd_tasks); - if (rcu_cpu_blocking_cur_gp()) - rcu_preempt_ctrlblk.gp_tasks = &t->rcu_node_entry; - } else if (rcu_preempt_running_reader() < 0 && - t->rcu_read_unlock_special) { - /* - * Complete exit from RCU read-side critical section on - * behalf of preempted instance of __rcu_read_unlock(). - */ - rcu_read_unlock_special(t); - } - - /* - * Either we were not in an RCU read-side critical section to - * begin with, or we have now recorded that critical section - * globally. Either way, we can now note a quiescent state - * for this CPU. Again, if we were in an RCU read-side critical - * section, and if that critical section was blocking the current - * grace period, then the fact that the task has been enqueued - * means that current grace period continues to be blocked. - */ - rcu_preempt_cpu_qs(); - local_irq_restore(flags); -} - -/* - * Handle special cases during rcu_read_unlock(), such as needing to - * notify RCU core processing or task having blocked during the RCU - * read-side critical section. - */ -void rcu_read_unlock_special(struct task_struct *t) -{ - int empty; - int empty_exp; - unsigned long flags; - struct list_head *np; -#ifdef CONFIG_RCU_BOOST - struct rt_mutex *rbmp = NULL; -#endif /* #ifdef CONFIG_RCU_BOOST */ - int special; - - /* - * NMI handlers cannot block and cannot safely manipulate state. - * They therefore cannot possibly be special, so just leave. - */ - if (in_nmi()) - return; - - local_irq_save(flags); - - /* - * If RCU core is waiting for this CPU to exit critical section, - * let it know that we have done so. - */ - special = t->rcu_read_unlock_special; - if (special & RCU_READ_UNLOCK_NEED_QS) - rcu_preempt_cpu_qs(); - - /* Hardware IRQ handlers cannot block. */ - if (in_irq() || in_serving_softirq()) { - local_irq_restore(flags); - return; - } - - /* Clean up if blocked during RCU read-side critical section. */ - if (special & RCU_READ_UNLOCK_BLOCKED) { - t->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_BLOCKED; - - /* - * Remove this task from the ->blkd_tasks list and adjust - * any pointers that might have been referencing it. - */ - empty = !rcu_preempt_blocked_readers_cgp(); - empty_exp = rcu_preempt_ctrlblk.exp_tasks == NULL; - np = rcu_next_node_entry(t); - list_del_init(&t->rcu_node_entry); - if (&t->rcu_node_entry == rcu_preempt_ctrlblk.gp_tasks) - rcu_preempt_ctrlblk.gp_tasks = np; - if (&t->rcu_node_entry == rcu_preempt_ctrlblk.exp_tasks) - rcu_preempt_ctrlblk.exp_tasks = np; -#ifdef CONFIG_RCU_BOOST - if (&t->rcu_node_entry == rcu_preempt_ctrlblk.boost_tasks) - rcu_preempt_ctrlblk.boost_tasks = np; -#endif /* #ifdef CONFIG_RCU_BOOST */ - - /* - * If this was the last task on the current list, and if - * we aren't waiting on the CPU, report the quiescent state - * and start a new grace period if needed. - */ - if (!empty && !rcu_preempt_blocked_readers_cgp()) { - rcu_preempt_cpu_qs(); - rcu_preempt_start_gp(); - } - - /* - * If this was the last task on the expedited lists, - * then we need wake up the waiting task. - */ - if (!empty_exp && rcu_preempt_ctrlblk.exp_tasks == NULL) - rcu_report_exp_done(); - } -#ifdef CONFIG_RCU_BOOST - /* Unboost self if was boosted. */ - if (t->rcu_boost_mutex != NULL) { - rbmp = t->rcu_boost_mutex; - t->rcu_boost_mutex = NULL; - rt_mutex_unlock(rbmp); - } -#endif /* #ifdef CONFIG_RCU_BOOST */ - local_irq_restore(flags); -} - -/* - * Check for a quiescent state from the current CPU. When a task blocks, - * the task is recorded in the rcu_preempt_ctrlblk structure, which is - * checked elsewhere. This is called from the scheduling-clock interrupt. - * - * Caller must disable hard irqs. - */ -static void rcu_preempt_check_callbacks(void) -{ - struct task_struct *t = current; - - if (rcu_preempt_gp_in_progress() && - (!rcu_preempt_running_reader() || - !rcu_cpu_blocking_cur_gp())) - rcu_preempt_cpu_qs(); - if (&rcu_preempt_ctrlblk.rcb.rcucblist != - rcu_preempt_ctrlblk.rcb.donetail) - invoke_rcu_callbacks(); - if (rcu_preempt_gp_in_progress() && - rcu_cpu_blocking_cur_gp() && - rcu_preempt_running_reader() > 0) - t->rcu_read_unlock_special |= RCU_READ_UNLOCK_NEED_QS; -} - -/* - * TINY_PREEMPT_RCU has an extra callback-list tail pointer to - * update, so this is invoked from rcu_process_callbacks() to - * handle that case. Of course, it is invoked for all flavors of - * RCU, but RCU callbacks can appear only on one of the lists, and - * neither ->nexttail nor ->donetail can possibly be NULL, so there - * is no need for an explicit check. - */ -static void rcu_preempt_remove_callbacks(struct rcu_ctrlblk *rcp) -{ - if (rcu_preempt_ctrlblk.nexttail == rcp->donetail) - rcu_preempt_ctrlblk.nexttail = &rcp->rcucblist; -} - -/* - * Process callbacks for preemptible RCU. - */ -static void rcu_preempt_process_callbacks(void) -{ - __rcu_process_callbacks(&rcu_preempt_ctrlblk.rcb); -} - -/* - * Queue a preemptible -RCU callback for invocation after a grace period. - */ -void call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) -{ - unsigned long flags; - - debug_rcu_head_queue(head); - head->func = func; - head->next = NULL; - - local_irq_save(flags); - *rcu_preempt_ctrlblk.nexttail = head; - rcu_preempt_ctrlblk.nexttail = &head->next; - RCU_TRACE(rcu_preempt_ctrlblk.rcb.qlen++); - rcu_preempt_start_gp(); /* checks to see if GP needed. */ - local_irq_restore(flags); -} -EXPORT_SYMBOL_GPL(call_rcu); - -/* - * synchronize_rcu - wait until a grace period has elapsed. - * - * Control will return to the caller some time after a full grace - * period has elapsed, in other words after all currently executing RCU - * read-side critical sections have completed. RCU read-side critical - * sections are delimited by rcu_read_lock() and rcu_read_unlock(), - * and may be nested. - */ -void synchronize_rcu(void) -{ - rcu_lockdep_assert(!lock_is_held(&rcu_bh_lock_map) && - !lock_is_held(&rcu_lock_map) && - !lock_is_held(&rcu_sched_lock_map), - "Illegal synchronize_rcu() in RCU read-side critical section"); - -#ifdef CONFIG_DEBUG_LOCK_ALLOC - if (!rcu_scheduler_active) - return; -#endif /* #ifdef CONFIG_DEBUG_LOCK_ALLOC */ - - WARN_ON_ONCE(rcu_preempt_running_reader()); - if (!rcu_preempt_blocked_readers_any()) - return; - - /* Once we get past the fastpath checks, same code as rcu_barrier(). */ - if (rcu_expedited) - synchronize_rcu_expedited(); - else - rcu_barrier(); -} -EXPORT_SYMBOL_GPL(synchronize_rcu); - -static DECLARE_WAIT_QUEUE_HEAD(sync_rcu_preempt_exp_wq); -static unsigned long sync_rcu_preempt_exp_count; -static DEFINE_MUTEX(sync_rcu_preempt_exp_mutex); - -/* - * Return non-zero if there are any tasks in RCU read-side critical - * sections blocking the current preemptible-RCU expedited grace period. - * If there is no preemptible-RCU expedited grace period currently in - * progress, returns zero unconditionally. - */ -static int rcu_preempted_readers_exp(void) -{ - return rcu_preempt_ctrlblk.exp_tasks != NULL; -} - -/* - * Report the exit from RCU read-side critical section for the last task - * that queued itself during or before the current expedited preemptible-RCU - * grace period. - */ -static void rcu_report_exp_done(void) -{ - wake_up(&sync_rcu_preempt_exp_wq); -} - -/* - * Wait for an rcu-preempt grace period, but expedite it. The basic idea - * is to rely in the fact that there is but one CPU, and that it is - * illegal for a task to invoke synchronize_rcu_expedited() while in a - * preemptible-RCU read-side critical section. Therefore, any such - * critical sections must correspond to blocked tasks, which must therefore - * be on the ->blkd_tasks list. So just record the current head of the - * list in the ->exp_tasks pointer, and wait for all tasks including and - * after the task pointed to by ->exp_tasks to drain. - */ -void synchronize_rcu_expedited(void) -{ - unsigned long flags; - struct rcu_preempt_ctrlblk *rpcp = &rcu_preempt_ctrlblk; - unsigned long snap; - - barrier(); /* ensure prior action seen before grace period. */ - - WARN_ON_ONCE(rcu_preempt_running_reader()); - - /* - * Acquire lock so that there is only one preemptible RCU grace - * period in flight. Of course, if someone does the expedited - * grace period for us while we are acquiring the lock, just leave. - */ - snap = sync_rcu_preempt_exp_count + 1; - mutex_lock(&sync_rcu_preempt_exp_mutex); - if (ULONG_CMP_LT(snap, sync_rcu_preempt_exp_count)) - goto unlock_mb_ret; /* Others did our work for us. */ - - local_irq_save(flags); - - /* - * All RCU readers have to already be on blkd_tasks because - * we cannot legally be executing in an RCU read-side critical - * section. - */ - - /* Snapshot current head of ->blkd_tasks list. */ - rpcp->exp_tasks = rpcp->blkd_tasks.next; - if (rpcp->exp_tasks == &rpcp->blkd_tasks) - rpcp->exp_tasks = NULL; - - /* Wait for tail of ->blkd_tasks list to drain. */ - if (!rcu_preempted_readers_exp()) { - local_irq_restore(flags); - } else { - rcu_initiate_boost(); - local_irq_restore(flags); - wait_event(sync_rcu_preempt_exp_wq, - !rcu_preempted_readers_exp()); - } - - /* Clean up and exit. */ - barrier(); /* ensure expedited GP seen before counter increment. */ - sync_rcu_preempt_exp_count++; -unlock_mb_ret: - mutex_unlock(&sync_rcu_preempt_exp_mutex); - barrier(); /* ensure subsequent action seen after grace period. */ -} -EXPORT_SYMBOL_GPL(synchronize_rcu_expedited); - -/* - * Does preemptible RCU need the CPU to stay out of dynticks mode? - */ -int rcu_preempt_needs_cpu(void) -{ - return rcu_preempt_ctrlblk.rcb.rcucblist != NULL; -} - -#else /* #ifdef CONFIG_TINY_PREEMPT_RCU */ - #ifdef CONFIG_RCU_TRACE /* @@ -895,79 +138,6 @@ static void rcu_preempt_process_callbacks(void) { } -#endif /* #else #ifdef CONFIG_TINY_PREEMPT_RCU */ - -#ifdef CONFIG_RCU_BOOST - -/* - * Wake up rcu_kthread() to process callbacks now eligible for invocation - * or to boost readers. - */ -static void invoke_rcu_callbacks(void) -{ - have_rcu_kthread_work = 1; - if (rcu_kthread_task != NULL) - wake_up(&rcu_kthread_wq); -} - -#ifdef CONFIG_RCU_TRACE - -/* - * Is the current CPU running the RCU-callbacks kthread? - * Caller must have preemption disabled. - */ -static bool rcu_is_callbacks_kthread(void) -{ - return rcu_kthread_task == current; -} - -#endif /* #ifdef CONFIG_RCU_TRACE */ - -/* - * This kthread invokes RCU callbacks whose grace periods have - * elapsed. It is awakened as needed, and takes the place of the - * RCU_SOFTIRQ that is used for this purpose when boosting is disabled. - * This is a kthread, but it is never stopped, at least not until - * the system goes down. - */ -static int rcu_kthread(void *arg) -{ - unsigned long work; - unsigned long morework; - unsigned long flags; - - for (;;) { - wait_event_interruptible(rcu_kthread_wq, - have_rcu_kthread_work != 0); - morework = rcu_boost(); - local_irq_save(flags); - work = have_rcu_kthread_work; - have_rcu_kthread_work = morework; - local_irq_restore(flags); - if (work) - rcu_process_callbacks(NULL); - schedule_timeout_interruptible(1); /* Leave CPU for others. */ - } - - return 0; /* Not reached, but needed to shut gcc up. */ -} - -/* - * Spawn the kthread that invokes RCU callbacks. - */ -static int __init rcu_spawn_kthreads(void) -{ - struct sched_param sp; - - rcu_kthread_task = kthread_run(rcu_kthread, NULL, "rcu_kthread"); - sp.sched_priority = RCU_BOOST_PRIO; - sched_setscheduler_nocheck(rcu_kthread_task, SCHED_FIFO, &sp); - return 0; -} -early_initcall(rcu_spawn_kthreads); - -#else /* #ifdef CONFIG_RCU_BOOST */ - /* Hold off callback invocation until early_initcall() time. */ static int rcu_scheduler_fully_active __read_mostly; @@ -1001,8 +171,6 @@ static int __init rcu_scheduler_really_started(void) } early_initcall(rcu_scheduler_really_started); -#endif /* #else #ifdef CONFIG_RCU_BOOST */ - #ifdef CONFIG_DEBUG_LOCK_ALLOC #include @@ -1020,25 +188,6 @@ void __init rcu_scheduler_starting(void) #ifdef CONFIG_RCU_TRACE -#ifdef CONFIG_RCU_BOOST - -static void rcu_initiate_boost_trace(void) -{ - if (list_empty(&rcu_preempt_ctrlblk.blkd_tasks)) - rcu_preempt_ctrlblk.n_balk_blkd_tasks++; - else if (rcu_preempt_ctrlblk.gp_tasks == NULL && - rcu_preempt_ctrlblk.exp_tasks == NULL) - rcu_preempt_ctrlblk.n_balk_exp_gp_tasks++; - else if (rcu_preempt_ctrlblk.boost_tasks != NULL) - rcu_preempt_ctrlblk.n_balk_boost_tasks++; - else if (!ULONG_CMP_GE(jiffies, rcu_preempt_ctrlblk.boost_time)) - rcu_preempt_ctrlblk.n_balk_notyet++; - else - rcu_preempt_ctrlblk.n_balk_nos++; -} - -#endif /* #ifdef CONFIG_RCU_BOOST */ - static void rcu_trace_sub_qlen(struct rcu_ctrlblk *rcp, int n) { unsigned long flags; @@ -1105,9 +254,6 @@ MODULE_LICENSE("GPL"); static void check_cpu_stall_preempt(void) { -#ifdef CONFIG_TINY_PREEMPT_RCU - check_cpu_stall(&rcu_preempt_ctrlblk.rcb); -#endif /* #ifdef CONFIG_TINY_PREEMPT_RCU */ } #endif /* #ifdef CONFIG_RCU_TRACE */ -- cgit v1.2.3 From 221304e95e1466fb49b630f67a719cc735ec5353 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 27 Mar 2013 08:59:52 -0700 Subject: rcu: Remove show_tiny_preempt_stats() With the removal of CONFIG_TINY_PREEMPT_RCU, show_tiny_preempt_stats() is now an empty function. This commit therefore eliminates it by inlining it. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutiny_plugin.h | 13 ------------- 1 file changed, 13 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutiny_plugin.h b/kernel/rcutiny_plugin.h index 29a4dd78c8bf..cf0bc22434c0 100644 --- a/kernel/rcutiny_plugin.h +++ b/kernel/rcutiny_plugin.h @@ -102,18 +102,6 @@ static void check_cpu_stalls(void) RCU_TRACE(check_cpu_stall_preempt()); } -#ifdef CONFIG_RCU_TRACE - -/* - * Because preemptible RCU does not exist, it is not necessary to - * dump out its statistics. - */ -static void show_tiny_preempt_stats(struct seq_file *m) -{ -} - -#endif /* #ifdef CONFIG_RCU_TRACE */ - /* * Because preemptible RCU does not exist, it never has any callbacks * to check. @@ -202,7 +190,6 @@ static void rcu_trace_sub_qlen(struct rcu_ctrlblk *rcp, int n) */ static int show_tiny_stats(struct seq_file *m, void *unused) { - show_tiny_preempt_stats(m); seq_printf(m, "rcu_sched: qlen: %ld\n", rcu_sched_ctrlblk.qlen); seq_printf(m, "rcu_bh: qlen: %ld\n", rcu_bh_ctrlblk.qlen); return 0; -- cgit v1.2.3 From 9acaac8ced57be8312cbf9f2a1e4f5e23b363493 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 27 Mar 2013 09:02:40 -0700 Subject: rcu: Remove rcu_preempt_check_callbacks() With the removal of CONFIG_TINY_PREEMPT_RCU, rcu_preempt_check_callbacks() is now an empty function. This commit therefore eliminates it by inlining it. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutiny.c | 1 - kernel/rcutiny_plugin.h | 8 -------- 2 files changed, 9 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutiny.c b/kernel/rcutiny.c index a0714a51b6d7..91782827775b 100644 --- a/kernel/rcutiny.c +++ b/kernel/rcutiny.c @@ -257,7 +257,6 @@ void rcu_check_callbacks(int cpu, int user) rcu_sched_qs(cpu); else if (!in_softirq()) rcu_bh_qs(cpu); - rcu_preempt_check_callbacks(); } /* diff --git a/kernel/rcutiny_plugin.h b/kernel/rcutiny_plugin.h index cf0bc22434c0..404b3a31e517 100644 --- a/kernel/rcutiny_plugin.h +++ b/kernel/rcutiny_plugin.h @@ -102,14 +102,6 @@ static void check_cpu_stalls(void) RCU_TRACE(check_cpu_stall_preempt()); } -/* - * Because preemptible RCU does not exist, it never has any callbacks - * to check. - */ -static void rcu_preempt_check_callbacks(void) -{ -} - /* * Because preemptible RCU does not exist, it never has any callbacks * to remove. -- cgit v1.2.3 From 47d65935a7f26f24417585e872e254c7ecc6596f Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 27 Mar 2013 09:05:34 -0700 Subject: rcu: Remove rcu_preempt_remove_callbacks() With the removal of CONFIG_TINY_PREEMPT_RCU, rcu_preempt_remove_callbacks() is now an empty function. This commit therefore eliminates it by inlining it. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutiny.c | 1 - kernel/rcutiny_plugin.h | 8 -------- 2 files changed, 9 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutiny.c b/kernel/rcutiny.c index 91782827775b..6f5a2a6cc63f 100644 --- a/kernel/rcutiny.c +++ b/kernel/rcutiny.c @@ -289,7 +289,6 @@ static void __rcu_process_callbacks(struct rcu_ctrlblk *rcp) *rcp->donetail = NULL; if (rcp->curtail == rcp->donetail) rcp->curtail = &rcp->rcucblist; - rcu_preempt_remove_callbacks(rcp); rcp->donetail = &rcp->rcucblist; local_irq_restore(flags); diff --git a/kernel/rcutiny_plugin.h b/kernel/rcutiny_plugin.h index 404b3a31e517..8b835b98114c 100644 --- a/kernel/rcutiny_plugin.h +++ b/kernel/rcutiny_plugin.h @@ -102,14 +102,6 @@ static void check_cpu_stalls(void) RCU_TRACE(check_cpu_stall_preempt()); } -/* - * Because preemptible RCU does not exist, it never has any callbacks - * to remove. - */ -static void rcu_preempt_remove_callbacks(struct rcu_ctrlblk *rcp) -{ -} - /* * Because preemptible RCU does not exist, it never has any callbacks * to process. -- cgit v1.2.3 From 58c4e69d43df91fd6a55bc070474aad6b7cfb18d Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 27 Mar 2013 09:11:12 -0700 Subject: rcu: Remove rcu_preempt_process_callbacks() With the removal of CONFIG_TINY_PREEMPT_RCU, rcu_preempt_process_callbacks() is now an empty function. This commit therefore eliminates it by inlining it. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutiny.c | 1 - kernel/rcutiny_plugin.h | 8 -------- 2 files changed, 9 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutiny.c b/kernel/rcutiny.c index 6f5a2a6cc63f..7fc2339b0859 100644 --- a/kernel/rcutiny.c +++ b/kernel/rcutiny.c @@ -314,7 +314,6 @@ static void rcu_process_callbacks(struct softirq_action *unused) { __rcu_process_callbacks(&rcu_sched_ctrlblk); __rcu_process_callbacks(&rcu_bh_ctrlblk); - rcu_preempt_process_callbacks(); } /* diff --git a/kernel/rcutiny_plugin.h b/kernel/rcutiny_plugin.h index 8b835b98114c..bfe992407803 100644 --- a/kernel/rcutiny_plugin.h +++ b/kernel/rcutiny_plugin.h @@ -102,14 +102,6 @@ static void check_cpu_stalls(void) RCU_TRACE(check_cpu_stall_preempt()); } -/* - * Because preemptible RCU does not exist, it never has any callbacks - * to process. - */ -static void rcu_preempt_process_callbacks(void) -{ -} - /* Hold off callback invocation until early_initcall() time. */ static int rcu_scheduler_fully_active __read_mostly; -- cgit v1.2.3 From 9dc5ad32488a75504349372330cc228d4dd678db Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 27 Mar 2013 10:11:15 -0700 Subject: rcu: Simplify RCU_TINY RCU callback invocation TINY_PREEMPT_RCU could use a kthread to handle RCU callback invocation, which required an API to abstract kthread vs. softirq invocation. Now that TINY_PREEMPT_RCU is no longer with us, this commit retires this API in favor of direct use of the relevant softirq primitives. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutiny.c | 14 +++++++++----- kernel/rcutiny_plugin.h | 33 --------------------------------- 2 files changed, 9 insertions(+), 38 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutiny.c b/kernel/rcutiny.c index 7fc2339b0859..4adc9e26da34 100644 --- a/kernel/rcutiny.c +++ b/kernel/rcutiny.c @@ -44,7 +44,6 @@ /* Forward declarations for rcutiny_plugin.h. */ struct rcu_ctrlblk; -static void invoke_rcu_callbacks(void); static void __rcu_process_callbacks(struct rcu_ctrlblk *rcp); static void rcu_process_callbacks(struct softirq_action *unused); static void __call_rcu(struct rcu_head *head, @@ -227,7 +226,7 @@ void rcu_sched_qs(int cpu) local_irq_save(flags); if (rcu_qsctr_help(&rcu_sched_ctrlblk) + rcu_qsctr_help(&rcu_bh_ctrlblk)) - invoke_rcu_callbacks(); + raise_softirq(RCU_SOFTIRQ); local_irq_restore(flags); } @@ -240,7 +239,7 @@ void rcu_bh_qs(int cpu) local_irq_save(flags); if (rcu_qsctr_help(&rcu_bh_ctrlblk)) - invoke_rcu_callbacks(); + raise_softirq(RCU_SOFTIRQ); local_irq_restore(flags); } @@ -277,7 +276,7 @@ static void __rcu_process_callbacks(struct rcu_ctrlblk *rcp) ACCESS_ONCE(rcp->rcucblist), need_resched(), is_idle_task(current), - rcu_is_callbacks_kthread())); + false)); return; } @@ -307,7 +306,7 @@ static void __rcu_process_callbacks(struct rcu_ctrlblk *rcp) RCU_TRACE(rcu_trace_sub_qlen(rcp, cb_count)); RCU_TRACE(trace_rcu_batch_end(rcp->name, cb_count, 0, need_resched(), is_idle_task(current), - rcu_is_callbacks_kthread())); + false)); } static void rcu_process_callbacks(struct softirq_action *unused) @@ -379,3 +378,8 @@ void call_rcu_bh(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) __call_rcu(head, func, &rcu_bh_ctrlblk); } EXPORT_SYMBOL_GPL(call_rcu_bh); + +void rcu_init(void) +{ + open_softirq(RCU_SOFTIRQ, rcu_process_callbacks); +} diff --git a/kernel/rcutiny_plugin.h b/kernel/rcutiny_plugin.h index bfe992407803..36fd83c544c8 100644 --- a/kernel/rcutiny_plugin.h +++ b/kernel/rcutiny_plugin.h @@ -102,39 +102,6 @@ static void check_cpu_stalls(void) RCU_TRACE(check_cpu_stall_preempt()); } -/* Hold off callback invocation until early_initcall() time. */ -static int rcu_scheduler_fully_active __read_mostly; - -/* - * Start up softirq processing of callbacks. - */ -void invoke_rcu_callbacks(void) -{ - if (rcu_scheduler_fully_active) - raise_softirq(RCU_SOFTIRQ); -} - -#ifdef CONFIG_RCU_TRACE - -/* - * There is no callback kthread, so this thread is never it. - */ -static bool rcu_is_callbacks_kthread(void) -{ - return false; -} - -#endif /* #ifdef CONFIG_RCU_TRACE */ - -static int __init rcu_scheduler_really_started(void) -{ - rcu_scheduler_fully_active = 1; - open_softirq(RCU_SOFTIRQ, rcu_process_callbacks); - raise_softirq(RCU_SOFTIRQ); /* Invoke any callbacks from early boot. */ - return 0; -} -early_initcall(rcu_scheduler_really_started); - #ifdef CONFIG_DEBUG_LOCK_ALLOC #include -- cgit v1.2.3 From 4879c84daa7bd6757b99ef76b30d4fcebccfcc6f Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 27 Mar 2013 10:18:04 -0700 Subject: rcu: Remove check_cpu_stall_preempt() With the removal of CONFIG_TINY_PREEMPT_RCU, check_cpu_stall_preempt() is now an empty function. This commit therefore eliminates it by inlining it. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutiny_plugin.h | 7 ------- 1 file changed, 7 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutiny_plugin.h b/kernel/rcutiny_plugin.h index 36fd83c544c8..bac3a6ecb991 100644 --- a/kernel/rcutiny_plugin.h +++ b/kernel/rcutiny_plugin.h @@ -82,8 +82,6 @@ static void check_cpu_stall(struct rcu_ctrlblk *rcp) rcp->jiffies_stall = jiffies + rcu_jiffies_till_stall_check(); } -static void check_cpu_stall_preempt(void); - #endif /* #ifdef CONFIG_RCU_TRACE */ static void reset_cpu_stall_ticks(struct rcu_ctrlblk *rcp) @@ -99,7 +97,6 @@ static void check_cpu_stalls(void) { RCU_TRACE(check_cpu_stall(&rcu_bh_ctrlblk)); RCU_TRACE(check_cpu_stall(&rcu_sched_ctrlblk)); - RCU_TRACE(check_cpu_stall_preempt()); } #ifdef CONFIG_DEBUG_LOCK_ALLOC @@ -182,8 +179,4 @@ MODULE_AUTHOR("Paul E. McKenney"); MODULE_DESCRIPTION("Read-Copy Update tracing for tiny implementation"); MODULE_LICENSE("GPL"); -static void check_cpu_stall_preempt(void) -{ -} - #endif /* #ifdef CONFIG_RCU_TRACE */ -- cgit v1.2.3 From 318bdcd95938ec3a530fc789da662ce159d50d46 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 27 Mar 2013 10:43:02 -0700 Subject: rcu: Consolidate rcutiny_plugin.h ifdefs This commit rearranges code in order to allow ifdefs to be consolidated in kernel/rcutiny_plugin.h, simplifying the code. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutiny_plugin.h | 86 +++++++++++++++++++++++-------------------------- 1 file changed, 40 insertions(+), 46 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutiny_plugin.h b/kernel/rcutiny_plugin.h index bac3a6ecb991..65ef1800f4fd 100644 --- a/kernel/rcutiny_plugin.h +++ b/kernel/rcutiny_plugin.h @@ -53,54 +53,10 @@ static struct rcu_ctrlblk rcu_bh_ctrlblk = { }; #ifdef CONFIG_DEBUG_LOCK_ALLOC +#include + int rcu_scheduler_active __read_mostly; EXPORT_SYMBOL_GPL(rcu_scheduler_active); -#endif /* #ifdef CONFIG_DEBUG_LOCK_ALLOC */ - -#ifdef CONFIG_RCU_TRACE - -static void check_cpu_stall(struct rcu_ctrlblk *rcp) -{ - unsigned long j; - unsigned long js; - - if (rcu_cpu_stall_suppress) - return; - rcp->ticks_this_gp++; - j = jiffies; - js = rcp->jiffies_stall; - if (*rcp->curtail && ULONG_CMP_GE(j, js)) { - pr_err("INFO: %s stall on CPU (%lu ticks this GP) idle=%llx (t=%lu jiffies q=%ld)\n", - rcp->name, rcp->ticks_this_gp, rcu_dynticks_nesting, - jiffies - rcp->gp_start, rcp->qlen); - dump_stack(); - } - if (*rcp->curtail && ULONG_CMP_GE(j, js)) - rcp->jiffies_stall = jiffies + - 3 * rcu_jiffies_till_stall_check() + 3; - else if (ULONG_CMP_GE(j, js)) - rcp->jiffies_stall = jiffies + rcu_jiffies_till_stall_check(); -} - -#endif /* #ifdef CONFIG_RCU_TRACE */ - -static void reset_cpu_stall_ticks(struct rcu_ctrlblk *rcp) -{ -#ifdef CONFIG_RCU_TRACE - rcp->ticks_this_gp = 0; - rcp->gp_start = jiffies; - rcp->jiffies_stall = jiffies + rcu_jiffies_till_stall_check(); -#endif /* #ifdef CONFIG_RCU_TRACE */ -} - -static void check_cpu_stalls(void) -{ - RCU_TRACE(check_cpu_stall(&rcu_bh_ctrlblk)); - RCU_TRACE(check_cpu_stall(&rcu_sched_ctrlblk)); -} - -#ifdef CONFIG_DEBUG_LOCK_ALLOC -#include /* * During boot, we forgive RCU lockdep issues. After this function is @@ -179,4 +135,42 @@ MODULE_AUTHOR("Paul E. McKenney"); MODULE_DESCRIPTION("Read-Copy Update tracing for tiny implementation"); MODULE_LICENSE("GPL"); +static void check_cpu_stall(struct rcu_ctrlblk *rcp) +{ + unsigned long j; + unsigned long js; + + if (rcu_cpu_stall_suppress) + return; + rcp->ticks_this_gp++; + j = jiffies; + js = rcp->jiffies_stall; + if (*rcp->curtail && ULONG_CMP_GE(j, js)) { + pr_err("INFO: %s stall on CPU (%lu ticks this GP) idle=%llx (t=%lu jiffies q=%ld)\n", + rcp->name, rcp->ticks_this_gp, rcu_dynticks_nesting, + jiffies - rcp->gp_start, rcp->qlen); + dump_stack(); + } + if (*rcp->curtail && ULONG_CMP_GE(j, js)) + rcp->jiffies_stall = jiffies + + 3 * rcu_jiffies_till_stall_check() + 3; + else if (ULONG_CMP_GE(j, js)) + rcp->jiffies_stall = jiffies + rcu_jiffies_till_stall_check(); +} + #endif /* #ifdef CONFIG_RCU_TRACE */ + +static void reset_cpu_stall_ticks(struct rcu_ctrlblk *rcp) +{ +#ifdef CONFIG_RCU_TRACE + rcp->ticks_this_gp = 0; + rcp->gp_start = jiffies; + rcp->jiffies_stall = jiffies + rcu_jiffies_till_stall_check(); +#endif /* #ifdef CONFIG_RCU_TRACE */ +} + +static void check_cpu_stalls(void) +{ + RCU_TRACE(check_cpu_stall(&rcu_bh_ctrlblk)); + RCU_TRACE(check_cpu_stall(&rcu_sched_ctrlblk)); +} -- cgit v1.2.3 From 2439b696cb5303f1eeb6aeebcee19e0056c3dd6e Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 11 Apr 2013 10:15:52 -0700 Subject: rcu: Shrink TINY_RCU by moving exit_rcu() Now that TINY_PREEMPT_RCU is no more, exit_rcu() is always an empty function. But if TINY_RCU is going to have an empty function, it should be in include/linux/rcutiny.h, where it does not bloat the kernel. This commit therefore moves exit_rcu() out of kernel/rcupdate.c to kernel/rcutree_plugin.h, and places a static inline empty function in include/linux/rcutiny.h in order to shrink TINY_RCU a bit. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcupdate.c | 26 +------------------------- kernel/rcutree_plugin.h | 26 ++++++++++++++++++++++++++ 2 files changed, 27 insertions(+), 25 deletions(-) (limited to 'kernel') diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c index 48ab70384a4c..0be1fa2ea521 100644 --- a/kernel/rcupdate.c +++ b/kernel/rcupdate.c @@ -104,31 +104,7 @@ void __rcu_read_unlock(void) } EXPORT_SYMBOL_GPL(__rcu_read_unlock); -/* - * Check for a task exiting while in a preemptible-RCU read-side - * critical section, clean up if so. No need to issue warnings, - * as debug_check_no_locks_held() already does this if lockdep - * is enabled. - */ -void exit_rcu(void) -{ - struct task_struct *t = current; - - if (likely(list_empty(¤t->rcu_node_entry))) - return; - t->rcu_read_lock_nesting = 1; - barrier(); - t->rcu_read_unlock_special = RCU_READ_UNLOCK_BLOCKED; - __rcu_read_unlock(); -} - -#else /* #ifdef CONFIG_PREEMPT_RCU */ - -void exit_rcu(void) -{ -} - -#endif /* #else #ifdef CONFIG_PREEMPT_RCU */ +#endif /* #ifdef CONFIG_PREEMPT_RCU */ #ifdef CONFIG_DEBUG_LOCK_ALLOC static struct lock_class_key rcu_lock_key; diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 207844ea0226..de701bbdb624 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -932,6 +932,24 @@ static void __init __rcu_init_preempt(void) rcu_init_one(&rcu_preempt_state, &rcu_preempt_data); } +/* + * Check for a task exiting while in a preemptible-RCU read-side + * critical section, clean up if so. No need to issue warnings, + * as debug_check_no_locks_held() already does this if lockdep + * is enabled. + */ +void exit_rcu(void) +{ + struct task_struct *t = current; + + if (likely(list_empty(¤t->rcu_node_entry))) + return; + t->rcu_read_lock_nesting = 1; + barrier(); + t->rcu_read_unlock_special = RCU_READ_UNLOCK_BLOCKED; + __rcu_read_unlock(); +} + #else /* #ifdef CONFIG_TREE_PREEMPT_RCU */ static struct rcu_state *rcu_state = &rcu_sched_state; @@ -1100,6 +1118,14 @@ static void __init __rcu_init_preempt(void) { } +/* + * Because preemptible RCU does not exist, tasks cannot possibly exit + * while in preemptible RCU read-side critical sections. + */ +void exit_rcu(void) +{ +} + #endif /* #else #ifdef CONFIG_TREE_PREEMPT_RCU */ #ifdef CONFIG_RCU_BOOST -- cgit v1.2.3 From 14961444696effb2e660fe876e5c1880f8bc3932 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Tue, 16 Apr 2013 07:49:22 -0700 Subject: rcu: Shrink TINY_RCU by reworking CPU-stall ifdefs TINY_RCU's reset_cpu_stall_ticks() and check_cpu_stalls() functions are defined unconditionally, and are empty functions if CONFIG_RCU_TRACE is disabled (which in turns disables detection of RCU CPU stalls). This commit saves a few lines of source code by defining these functions only if CONFIG_RCU_TRACE=y. Signed-off-by: Paul E. McKenney --- kernel/rcutiny.c | 4 ++-- kernel/rcutiny_plugin.h | 6 ++---- 2 files changed, 4 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutiny.c b/kernel/rcutiny.c index 4adc9e26da34..aa344111de3e 100644 --- a/kernel/rcutiny.c +++ b/kernel/rcutiny.c @@ -204,7 +204,7 @@ static int rcu_is_cpu_rrupt_from_idle(void) */ static int rcu_qsctr_help(struct rcu_ctrlblk *rcp) { - reset_cpu_stall_ticks(rcp); + RCU_TRACE(reset_cpu_stall_ticks(rcp)); if (rcp->rcucblist != NULL && rcp->donetail != rcp->curtail) { rcp->donetail = rcp->curtail; @@ -251,7 +251,7 @@ void rcu_bh_qs(int cpu) */ void rcu_check_callbacks(int cpu, int user) { - check_cpu_stalls(); + RCU_TRACE(check_cpu_stalls()); if (user || rcu_is_cpu_rrupt_from_idle()) rcu_sched_qs(cpu); else if (!in_softirq()) diff --git a/kernel/rcutiny_plugin.h b/kernel/rcutiny_plugin.h index 65ef1800f4fd..0cd385acccfa 100644 --- a/kernel/rcutiny_plugin.h +++ b/kernel/rcutiny_plugin.h @@ -158,15 +158,11 @@ static void check_cpu_stall(struct rcu_ctrlblk *rcp) rcp->jiffies_stall = jiffies + rcu_jiffies_till_stall_check(); } -#endif /* #ifdef CONFIG_RCU_TRACE */ - static void reset_cpu_stall_ticks(struct rcu_ctrlblk *rcp) { -#ifdef CONFIG_RCU_TRACE rcp->ticks_this_gp = 0; rcp->gp_start = jiffies; rcp->jiffies_stall = jiffies + rcu_jiffies_till_stall_check(); -#endif /* #ifdef CONFIG_RCU_TRACE */ } static void check_cpu_stalls(void) @@ -174,3 +170,5 @@ static void check_cpu_stalls(void) RCU_TRACE(check_cpu_stall(&rcu_bh_ctrlblk)); RCU_TRACE(check_cpu_stall(&rcu_sched_ctrlblk)); } + +#endif /* #ifdef CONFIG_RCU_TRACE */ -- cgit v1.2.3 From 34376a50fb1fa095b9d0636fa41ed2e73125f214 Mon Sep 17 00:00:00 2001 From: Ben Greear Date: Thu, 6 Jun 2013 14:29:49 -0700 Subject: Fix lockup related to stop_machine being stuck in __do_softirq. The stop machine logic can lock up if all but one of the migration threads make it through the disable-irq step and the one remaining thread gets stuck in __do_softirq. The reason __do_softirq can hang is that it has a bail-out based on jiffies timeout, but in the lockup case, jiffies itself is not incremented. To work around this, re-add the max_restart counter in __do_irq and stop processing irqs after 10 restarts. Thanks to Tejun Heo and Rusty Russell and others for helping me track this down. This was introduced in 3.9 by commit c10d73671ad3 ("softirq: reduce latencies"). It may be worth looking into ath9k to see if it has issues with its irq handler at a later date. The hang stack traces look something like this: ------------[ cut here ]------------ WARNING: at kernel/watchdog.c:245 watchdog_overflow_callback+0x9c/0xa7() Watchdog detected hard LOCKUP on cpu 2 Modules linked in: ath9k ath9k_common ath9k_hw ath mac80211 cfg80211 nfsv4 auth_rpcgss nfs fscache nf_nat_ipv4 nf_nat veth 8021q garp stp mrp llc pktgen lockd sunrpc] Pid: 23, comm: migration/2 Tainted: G C 3.9.4+ #11 Call Trace: warn_slowpath_common+0x85/0x9f warn_slowpath_fmt+0x46/0x48 watchdog_overflow_callback+0x9c/0xa7 __perf_event_overflow+0x137/0x1cb perf_event_overflow+0x14/0x16 intel_pmu_handle_irq+0x2dc/0x359 perf_event_nmi_handler+0x19/0x1b nmi_handle+0x7f/0xc2 do_nmi+0xbc/0x304 end_repeat_nmi+0x1e/0x2e <> cpu_stopper_thread+0xae/0x162 smpboot_thread_fn+0x258/0x260 kthread+0xc7/0xcf ret_from_fork+0x7c/0xb0 ---[ end trace 4947dfa9b0a4cec3 ]--- BUG: soft lockup - CPU#1 stuck for 22s! [migration/1:17] Modules linked in: ath9k ath9k_common ath9k_hw ath mac80211 cfg80211 nfsv4 auth_rpcgss nfs fscache nf_nat_ipv4 nf_nat veth 8021q garp stp mrp llc pktgen lockd sunrpc] irq event stamp: 835637905 hardirqs last enabled at (835637904): __do_softirq+0x9f/0x257 hardirqs last disabled at (835637905): apic_timer_interrupt+0x6d/0x80 softirqs last enabled at (5654720): __do_softirq+0x1ff/0x257 softirqs last disabled at (5654725): irq_exit+0x5f/0xbb CPU 1 Pid: 17, comm: migration/1 Tainted: G WC 3.9.4+ #11 To be filled by O.E.M. To be filled by O.E.M./To be filled by O.E.M. RIP: tasklet_hi_action+0xf0/0xf0 Process migration/1 Call Trace: __do_softirq+0x117/0x257 irq_exit+0x5f/0xbb smp_apic_timer_interrupt+0x8a/0x98 apic_timer_interrupt+0x72/0x80 printk+0x4d/0x4f stop_machine_cpu_stop+0x22c/0x274 cpu_stopper_thread+0xae/0x162 smpboot_thread_fn+0x258/0x260 kthread+0xc7/0xcf ret_from_fork+0x7c/0xb0 Signed-off-by: Ben Greear Acked-by: Tejun Heo Acked-by: Pekka Riikonen Cc: Eric Dumazet Cc: stable@kernel.org Signed-off-by: Linus Torvalds --- kernel/softirq.c | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/softirq.c b/kernel/softirq.c index b5197dcb0dad..3d6833f125d3 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -195,8 +195,12 @@ void local_bh_enable_ip(unsigned long ip) EXPORT_SYMBOL(local_bh_enable_ip); /* - * We restart softirq processing for at most 2 ms, - * and if need_resched() is not set. + * We restart softirq processing for at most MAX_SOFTIRQ_RESTART times, + * but break the loop if need_resched() is set or after 2 ms. + * The MAX_SOFTIRQ_TIME provides a nice upper bound in most cases, but in + * certain cases, such as stop_machine(), jiffies may cease to + * increment and so we need the MAX_SOFTIRQ_RESTART limit as + * well to make sure we eventually return from this method. * * These limits have been established via experimentation. * The two things to balance is latency against fairness - @@ -204,6 +208,7 @@ EXPORT_SYMBOL(local_bh_enable_ip); * should not be able to lock up the box. */ #define MAX_SOFTIRQ_TIME msecs_to_jiffies(2) +#define MAX_SOFTIRQ_RESTART 10 asmlinkage void __do_softirq(void) { @@ -212,6 +217,7 @@ asmlinkage void __do_softirq(void) unsigned long end = jiffies + MAX_SOFTIRQ_TIME; int cpu; unsigned long old_flags = current->flags; + int max_restart = MAX_SOFTIRQ_RESTART; /* * Mask out PF_MEMALLOC s current task context is borrowed for the @@ -265,7 +271,8 @@ restart: pending = local_softirq_pending(); if (pending) { - if (time_before(jiffies, end) && !need_resched()) + if (time_before(jiffies, end) && !need_resched() && + --max_restart) goto restart; wakeup_softirqd(); -- cgit v1.2.3 From ee23871389d51e07380d23887333622fbe7d3dd9 Mon Sep 17 00:00:00 2001 From: Ivo Sieben Date: Mon, 3 Jun 2013 12:12:02 +0200 Subject: genirq: Set irq thread to RT priority on creation When a threaded irq handler is installed the irq thread is initially created on normal scheduling priority. Only after the irq thread is woken up it sets its priority to RT_FIFO MAX_USER_RT_PRIO/2 itself. This means that interrupts that occur directly after the irq handler is installed will be handled on a normal scheduling priority instead of the realtime priority that one would expect. Fix this by setting the RT priority on creation of the irq_thread. Signed-off-by: Ivo Sieben Cc: Sebastian Andrzej Siewior Cc: Steven Rostedt Link: http://lkml.kernel.org/r/1370254322-17240-1-git-send-email-meltedpianoman@gmail.com Signed-off-by: Thomas Gleixner --- kernel/irq/manage.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index fa17855ca65a..e16caa81f887 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -840,9 +840,6 @@ static void irq_thread_dtor(struct callback_head *unused) static int irq_thread(void *data) { struct callback_head on_exit_work; - static const struct sched_param param = { - .sched_priority = MAX_USER_RT_PRIO/2, - }; struct irqaction *action = data; struct irq_desc *desc = irq_to_desc(action->irq); irqreturn_t (*handler_fn)(struct irq_desc *desc, @@ -854,8 +851,6 @@ static int irq_thread(void *data) else handler_fn = irq_thread_fn; - sched_setscheduler(current, SCHED_FIFO, ¶m); - init_task_work(&on_exit_work, irq_thread_dtor); task_work_add(current, &on_exit_work, false); @@ -950,6 +945,9 @@ __setup_irq(unsigned int irq, struct irq_desc *desc, struct irqaction *new) */ if (new->thread_fn && !nested) { struct task_struct *t; + static const struct sched_param param = { + .sched_priority = MAX_USER_RT_PRIO/2, + }; t = kthread_create(irq_thread, new, "irq/%d-%s", irq, new->name); @@ -957,6 +955,9 @@ __setup_irq(unsigned int irq, struct irq_desc *desc, struct irqaction *new) ret = PTR_ERR(t); goto out_mput; } + + sched_setscheduler(t, SCHED_FIFO, ¶m); + /* * We keep the reference to the task struct even if * the thread dies to avoid that the interrupt code -- cgit v1.2.3 From 58e8eedf18577c7eac722d5d1f190507ea263d1b Mon Sep 17 00:00:00 2001 From: Yoshihiro YUNOMAE Date: Tue, 23 Apr 2013 10:32:39 +0900 Subject: tracing: Fix outputting formats of x86-tsc and counter when use trace_clock Outputting formats of x86-tsc and counter should be a raw format, but after applying the patch(2b6080f28c7cc3efc8625ab71495aae89aeb63a0), the format was changed to nanosec. This is because the global variable trace_clock_id was used. When we use multiple buffers, clock_id of each sub-buffer should be used. Then, this patch uses tr->clock_id instead of the global variable trace_clock_id. [ Basically, this fixes a regression where the multibuffer code changed the trace_clock file to update tr->clock_id but the traces still use the old global trace_clock_id variable, negating the file's effect. The global trace_clock_id variable is obsolete and removed. - SR ] Link: http://lkml.kernel.org/r/20130423013239.22334.7394.stgit@yunodevel Signed-off-by: Yoshihiro YUNOMAE Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 8 +++----- kernel/trace/trace.h | 2 -- 2 files changed, 3 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 1a41023a1f88..e71a8be4a6ee 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -652,8 +652,6 @@ static struct { ARCH_TRACE_CLOCKS }; -int trace_clock_id; - /* * trace_parser_get_init - gets the buffer for trace parser */ @@ -2826,7 +2824,7 @@ __tracing_open(struct inode *inode, struct file *file, bool snapshot) iter->iter_flags |= TRACE_FILE_ANNOTATE; /* Output in nanoseconds only if we are using a clock in nanoseconds. */ - if (trace_clocks[trace_clock_id].in_ns) + if (trace_clocks[tr->clock_id].in_ns) iter->iter_flags |= TRACE_FILE_TIME_IN_NS; /* stop the trace while dumping if we are not opening "snapshot" */ @@ -3825,7 +3823,7 @@ static int tracing_open_pipe(struct inode *inode, struct file *filp) iter->iter_flags |= TRACE_FILE_LAT_FMT; /* Output in nanoseconds only if we are using a clock in nanoseconds. */ - if (trace_clocks[trace_clock_id].in_ns) + if (trace_clocks[tr->clock_id].in_ns) iter->iter_flags |= TRACE_FILE_TIME_IN_NS; iter->cpu_file = tc->cpu; @@ -5095,7 +5093,7 @@ tracing_stats_read(struct file *filp, char __user *ubuf, cnt = ring_buffer_bytes_cpu(trace_buf->buffer, cpu); trace_seq_printf(s, "bytes: %ld\n", cnt); - if (trace_clocks[trace_clock_id].in_ns) { + if (trace_clocks[tr->clock_id].in_ns) { /* local or global for trace_clock */ t = ns2usecs(ring_buffer_oldest_event_ts(trace_buf->buffer, cpu)); usec_rem = do_div(t, USEC_PER_SEC); diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 711ca7d3e7f1..20572ed88c5c 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -700,8 +700,6 @@ enum print_line_t print_trace_line(struct trace_iterator *iter); extern unsigned long trace_flags; -extern int trace_clock_id; - /* Standard output formatting function used for function return traces */ #ifdef CONFIG_FUNCTION_GRAPH_TRACER -- cgit v1.2.3 From d7880812b3594d3c6dcbe3cfd71dabb17347d082 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 10 Jun 2013 16:52:03 +0200 Subject: idle: Add the stack canary init to cpu_startup_entry() Moving x86 to the generic idle implementation (commit 7d1a9417 "x86: Use generic idle loop") wreckaged the stack protector. I stupidly missed that boot_init_stack_canary() must be inlined from a function which never returns, but I put that call into arch_cpu_idle_prepare() which of course returns. I pondered to play tricks with arch_cpu_idle_prepare() first, but then I noticed, that the other archs which have implemented the stackprotector (ARM and SH) do not initialize the canary for the non-boot cpus. So I decided to move the boot_init_stack_canary() call into cpu_startup_entry() ifdeffed with an CONFIG_X86 for now. This #ifdef is just a temporary measure as I don't want to inflict the boot_init_stack_canary() call on ARM and SH that late in the cycle. I'll queue a patch for 3.11 which removes the #ifdef if the ARM/SH maintainers have no objection. Reported-by: Wouter van Kesteren Cc: x86@kernel.org Cc: Russell King Cc: Paul Mundt Signed-off-by: Thomas Gleixner --- kernel/cpu/idle.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) (limited to 'kernel') diff --git a/kernel/cpu/idle.c b/kernel/cpu/idle.c index d5585f5e038e..bf2ee1aafa0e 100644 --- a/kernel/cpu/idle.c +++ b/kernel/cpu/idle.c @@ -5,6 +5,7 @@ #include #include #include +#include #include @@ -112,6 +113,21 @@ static void cpu_idle_loop(void) void cpu_startup_entry(enum cpuhp_state state) { + /* + * This #ifdef needs to die, but it's too late in the cycle to + * make this generic (arm and sh have never invoked the canary + * init for the non boot cpus!). Will be fixed in 3.11 + */ +#ifdef CONFIG_X86 + /* + * If we're the non-boot CPU, nothing set the stack canary up + * for us. The boot CPU already has it initialized but no harm + * in doing it again. This is a good place for updating it, as + * we wont ever return from this function (so the invalid + * canaries already on the stack wont ever trigger). + */ + boot_init_stack_canary(); +#endif current_set_polling(); arch_cpu_idle_prepare(); cpu_idle_loop(); -- cgit v1.2.3 From 9350de06be45a5a8b927ac6577c9d35de61c90ca Mon Sep 17 00:00:00 2001 From: Bernie Thompson Date: Sat, 1 Jun 2013 00:47:43 +0000 Subject: PM / wakeup: Adjust messaging for wake events during suspend This adds in a new message to the wakeup code which adds an indication to the log that suspend was cancelled due to a wake event occouring during the suspend sequence. It also adjusts the message printed in suspend.c to reflect the potential that a suspend was aborted, as opposed to a device failing to suspend. Without these message adjustments one can end up with a kernel log that says that a device failed to suspend with no actual device suspend failures, which can be confusing to the log examiner. Signed-off-by: Bernie Thompson Signed-off-by: Rafael J. Wysocki --- kernel/power/suspend.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/power/suspend.c b/kernel/power/suspend.c index bef86d121eb2..ece04223bb1e 100644 --- a/kernel/power/suspend.c +++ b/kernel/power/suspend.c @@ -269,7 +269,7 @@ int suspend_devices_and_enter(suspend_state_t state) suspend_test_start(); error = dpm_suspend_start(PMSG_SUSPEND); if (error) { - printk(KERN_ERR "PM: Some devices failed to suspend\n"); + pr_err("PM: Some devices failed to suspend, or early wake event detected\n"); goto Recover_platform; } suspend_test_finish("suspend devices"); -- cgit v1.2.3 From ad71d889b88055e61e3970a6744a271a51a94f42 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Tue, 30 Apr 2013 15:46:14 -0400 Subject: tracing: Add function probe to trigger a ftrace dump to console Add the "dump" command to have the ftrace buffer dumped to console if a function is hit. This is useful when debugging a tripple fault, where you have an idea of a function that is called just before the tripple fault occurs, and can tell ftrace to dump its content out to the console before it continues. Format is: :dump echo 'bad_address:dump' > /debug/tracing/set_ftrace_filter To remove this: echo '!bad_address:dump' > /debug/tracing/set_ftrace_filter Requested-by: Luis Claudio R. Goncalves Signed-off-by: Steven Rostedt --- kernel/trace/trace_functions.c | 59 ++++++++++++++++++++++++++++++++++++++---- 1 file changed, 54 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c index c4d6d7191988..d7c8719734b8 100644 --- a/kernel/trace/trace_functions.c +++ b/kernel/trace/trace_functions.c @@ -290,6 +290,13 @@ ftrace_stacktrace_count(unsigned long ip, unsigned long parent_ip, void **data) trace_dump_stack(STACK_SKIP); } +static void +ftrace_dump_probe(unsigned long ip, unsigned long parent_ip, void **data) +{ + if (update_count(data)) + ftrace_dump(DUMP_ALL); +} + static int ftrace_probe_print(const char *name, struct seq_file *m, unsigned long ip, void *data) @@ -327,6 +334,13 @@ ftrace_stacktrace_print(struct seq_file *m, unsigned long ip, return ftrace_probe_print("stacktrace", m, ip, data); } +static int +ftrace_dump_print(struct seq_file *m, unsigned long ip, + struct ftrace_probe_ops *ops, void *data) +{ + return ftrace_probe_print("dump", m, ip, data); +} + static struct ftrace_probe_ops traceon_count_probe_ops = { .func = ftrace_traceon_count, .print = ftrace_traceon_print, @@ -342,6 +356,11 @@ static struct ftrace_probe_ops stacktrace_count_probe_ops = { .print = ftrace_stacktrace_print, }; +static struct ftrace_probe_ops dump_probe_ops = { + .func = ftrace_dump_probe, + .print = ftrace_dump_print, +}; + static struct ftrace_probe_ops traceon_probe_ops = { .func = ftrace_traceon, .print = ftrace_traceon_print, @@ -425,6 +444,19 @@ ftrace_stacktrace_callback(struct ftrace_hash *hash, param, enable); } +static int +ftrace_dump_callback(struct ftrace_hash *hash, + char *glob, char *cmd, char *param, int enable) +{ + struct ftrace_probe_ops *ops; + + ops = &dump_probe_ops; + + /* Only dump once. */ + return ftrace_trace_probe_callback(ops, hash, glob, cmd, + "1", enable); +} + static struct ftrace_func_command ftrace_traceon_cmd = { .name = "traceon", .func = ftrace_trace_onoff_callback, @@ -440,6 +472,11 @@ static struct ftrace_func_command ftrace_stacktrace_cmd = { .func = ftrace_stacktrace_callback, }; +static struct ftrace_func_command ftrace_dump_cmd = { + .name = "dump", + .func = ftrace_dump_callback, +}; + static int __init init_func_cmd_traceon(void) { int ret; @@ -450,13 +487,25 @@ static int __init init_func_cmd_traceon(void) ret = register_ftrace_command(&ftrace_traceon_cmd); if (ret) - unregister_ftrace_command(&ftrace_traceoff_cmd); + goto out_free_traceoff; ret = register_ftrace_command(&ftrace_stacktrace_cmd); - if (ret) { - unregister_ftrace_command(&ftrace_traceoff_cmd); - unregister_ftrace_command(&ftrace_traceon_cmd); - } + if (ret) + goto out_free_traceon; + + ret = register_ftrace_command(&ftrace_dump_cmd); + if (ret) + goto out_free_stacktrace; + + return 0; + + out_free_stacktrace: + unregister_ftrace_command(&ftrace_stacktrace_cmd); + out_free_traceon: + unregister_ftrace_command(&ftrace_traceon_cmd); + out_free_traceoff: + unregister_ftrace_command(&ftrace_traceoff_cmd); + return ret; } #else -- cgit v1.2.3 From 90e3c03c3a09a7b176b3fe59d78f5d9755ac8e37 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Tue, 30 Apr 2013 19:00:46 -0400 Subject: tracing: Add function probe to trigger a ftrace dump of current CPU trace Add the "cpudump" command to have the current CPU ftrace buffer dumped to console if a function is hit. This is useful when debugging a tripple fault, where you have an idea of a function that is called just before the tripple fault occurs, and can tell ftrace to dump its content out to the console before it continues. This differs from the "dump" command as it only dumps the content of the ring buffer for the currently executing CPU, and does not show the contents of the other CPUs. Format is: :cpudump echo 'bad_address:cpudump' > /debug/tracing/set_ftrace_filter To remove this: echo '!bad_address:cpudump' > /debug/tracing/set_ftrace_filter Signed-off-by: Steven Rostedt --- kernel/trace/trace_functions.c | 44 ++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 44 insertions(+) (limited to 'kernel') diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c index d7c8719734b8..b863f93b30f3 100644 --- a/kernel/trace/trace_functions.c +++ b/kernel/trace/trace_functions.c @@ -297,6 +297,14 @@ ftrace_dump_probe(unsigned long ip, unsigned long parent_ip, void **data) ftrace_dump(DUMP_ALL); } +/* Only dump the current CPU buffer. */ +static void +ftrace_cpudump_probe(unsigned long ip, unsigned long parent_ip, void **data) +{ + if (update_count(data)) + ftrace_dump(DUMP_ORIG); +} + static int ftrace_probe_print(const char *name, struct seq_file *m, unsigned long ip, void *data) @@ -341,6 +349,13 @@ ftrace_dump_print(struct seq_file *m, unsigned long ip, return ftrace_probe_print("dump", m, ip, data); } +static int +ftrace_cpudump_print(struct seq_file *m, unsigned long ip, + struct ftrace_probe_ops *ops, void *data) +{ + return ftrace_probe_print("cpudump", m, ip, data); +} + static struct ftrace_probe_ops traceon_count_probe_ops = { .func = ftrace_traceon_count, .print = ftrace_traceon_print, @@ -361,6 +376,11 @@ static struct ftrace_probe_ops dump_probe_ops = { .print = ftrace_dump_print, }; +static struct ftrace_probe_ops cpudump_probe_ops = { + .func = ftrace_cpudump_probe, + .print = ftrace_cpudump_print, +}; + static struct ftrace_probe_ops traceon_probe_ops = { .func = ftrace_traceon, .print = ftrace_traceon_print, @@ -457,6 +477,19 @@ ftrace_dump_callback(struct ftrace_hash *hash, "1", enable); } +static int +ftrace_cpudump_callback(struct ftrace_hash *hash, + char *glob, char *cmd, char *param, int enable) +{ + struct ftrace_probe_ops *ops; + + ops = &cpudump_probe_ops; + + /* Only dump once. */ + return ftrace_trace_probe_callback(ops, hash, glob, cmd, + "1", enable); +} + static struct ftrace_func_command ftrace_traceon_cmd = { .name = "traceon", .func = ftrace_trace_onoff_callback, @@ -477,6 +510,11 @@ static struct ftrace_func_command ftrace_dump_cmd = { .func = ftrace_dump_callback, }; +static struct ftrace_func_command ftrace_cpudump_cmd = { + .name = "cpudump", + .func = ftrace_cpudump_callback, +}; + static int __init init_func_cmd_traceon(void) { int ret; @@ -497,8 +535,14 @@ static int __init init_func_cmd_traceon(void) if (ret) goto out_free_stacktrace; + ret = register_ftrace_command(&ftrace_cpudump_cmd); + if (ret) + goto out_free_dump; + return 0; + out_free_dump: + unregister_ftrace_command(&ftrace_dump_cmd); out_free_stacktrace: unregister_ftrace_command(&ftrace_stacktrace_cmd); out_free_traceon: -- cgit v1.2.3 From 8092e808a31839c502a52d391b15f31c1d8764f5 Mon Sep 17 00:00:00 2001 From: Harsh Prateek Bora Date: Fri, 24 May 2013 12:52:17 +0530 Subject: tracing/trivial: Consolidate error return condition Consolidate the checks for !enabled and !param to return -EINVAL in event_enable_func(). Link: http://lkml.kernel.org/r/1369380137-12452-1-git-send-email-harsh@linux.vnet.ibm.com Signed-off-by: Harsh Prateek Bora Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 27963e2bf4bf..db086f172cf5 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -2011,10 +2011,7 @@ event_enable_func(struct ftrace_hash *hash, int ret; /* hash funcs only work with set_ftrace_filter */ - if (!enabled) - return -EINVAL; - - if (!param) + if (!enabled || !param) return -EINVAL; system = strsep(¶m, ":"); -- cgit v1.2.3 From 238ae93d699d59876b470bf6455de22bcfaa9a1b Mon Sep 17 00:00:00 2001 From: Wang YanQing Date: Sun, 26 May 2013 16:52:01 +0800 Subject: tracing: Fix file mode of free_buffer Commit 4f271a2a60c748599b30bb4dafff30d770439b96 (tracing: Add a proc file to stop tracing and free buffer) implement a method to free up ring buffer in kernel memory in the release code path of free_buffer's fd. Then we don't need read/write support for free_buffer, indeed we just have a dummy write fop, and don't implement read fop. So the 0200 is more reasonable file mode for free_buffer than the current file mode 0644. Link: http://lkml.kernel.org/r/20130526085201.GA3183@udknight Acked-by: Vaibhav Nagarnaik Acked-by: David Sharp Signed-off-by: Wang YanQing Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 1a41023a1f88..5f4a09c12e0b 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -5935,7 +5935,7 @@ init_tracer_debugfs(struct trace_array *tr, struct dentry *d_tracer) trace_create_file("buffer_total_size_kb", 0444, d_tracer, tr, &tracing_total_entries_fops); - trace_create_file("free_buffer", 0644, d_tracer, + trace_create_file("free_buffer", 0200, d_tracer, tr, &tracing_free_buffer_fops); trace_create_file("trace_marker", 0220, d_tracer, -- cgit v1.2.3 From 7614c3dc74733dff4b0e774f7a894b9ea6ec508c Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 28 May 2013 20:01:16 -0400 Subject: ftrace: Use schedule_on_each_cpu() as a heavy synchronize_sched() The function tracer uses preempt_disable/enable_notrace() for synchronization between reading registered ftrace_ops and unregistering them. Most of the ftrace_ops are global permanent structures that do not require this synchronization. That is, ops may be added and removed from the hlist but are never freed, and wont hurt if a synchronization is missed. But this is not true for dynamically created ftrace_ops or control_ops, which are used by the perf function tracing. The problem here is that the function tracer can be used to trace kernel/user context switches as well as going to and from idle. Basically, it can be used to trace blind spots of the RCU subsystem. This means that even though preempt_disable() is done, a synchronize_sched() will ignore CPUs that haven't made it out of user space or idle. These can include functions that are being traced just before entering or exiting the kernel sections. To implement the RCU synchronization, instead of using synchronize_sched() the use of schedule_on_each_cpu() is performed. This means that when a dynamically allocated ftrace_ops, or a control ops is being unregistered, all CPUs must be touched and execute a ftrace_sync() stub function via the work queues. This will rip CPUs out from idle or in dynamic tick mode. This only happens when a user disables perf function tracing or other dynamically allocated function tracers, but it allows us to continue to debug RCU and context tracking with function tracing. Link: http://lkml.kernel.org/r/1369785676.15552.55.camel@gandalf.local.home Cc: "Paul E. McKenney" Cc: Tejun Heo Cc: Ingo Molnar Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Peter Zijlstra Acked-by: Paul E. McKenney Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 23 +++++++++++++++++++++-- 1 file changed, 21 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 6c508ff33c62..800a8a2fbddb 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -413,6 +413,17 @@ static int __register_ftrace_function(struct ftrace_ops *ops) return 0; } +static void ftrace_sync(struct work_struct *work) +{ + /* + * This function is just a stub to implement a hard force + * of synchronize_sched(). This requires synchronizing + * tasks even in userspace and idle. + * + * Yes, function tracing is rude. + */ +} + static int __unregister_ftrace_function(struct ftrace_ops *ops) { int ret; @@ -440,8 +451,12 @@ static int __unregister_ftrace_function(struct ftrace_ops *ops) * so there'll be no new users. We must ensure * all current users are done before we free * the control data. + * Note synchronize_sched() is not enough, as we + * use preempt_disable() to do RCU, but the function + * tracer can be called where RCU is not active + * (before user_exit()). */ - synchronize_sched(); + schedule_on_each_cpu(ftrace_sync); control_ops_free(ops); } } else @@ -456,9 +471,13 @@ static int __unregister_ftrace_function(struct ftrace_ops *ops) /* * Dynamic ops may be freed, we must make sure that all * callers are done before leaving this function. + * + * Again, normal synchronize_sched() is not good enough. + * We need to do a hard force of sched synchronization. */ if (ops->flags & FTRACE_OPS_FL_DYNAMIC) - synchronize_sched(); + schedule_on_each_cpu(ftrace_sync); + return 0; } -- cgit v1.2.3 From aaf6ac0f0871cb7fc0f28f3a00edf329bc7adc29 Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Fri, 7 Jun 2013 15:07:48 +0900 Subject: tracing: Do not call kmem_cache_free() on allocation failure There's no point calling it when _alloc() failed. Link: http://lkml.kernel.org/r/1370585268-29169-1-git-send-email-namhyung@kernel.org Signed-off-by: Namhyung Kim Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 7 +------ 1 file changed, 1 insertion(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index db086f172cf5..f57b01574a30 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -97,7 +97,7 @@ static int __trace_define_field(struct list_head *head, const char *type, field = kmem_cache_alloc(field_cachep, GFP_TRACE); if (!field) - goto err; + return -ENOMEM; field->name = name; field->type = type; @@ -114,11 +114,6 @@ static int __trace_define_field(struct list_head *head, const char *type, list_add(&field->link, head); return 0; - -err: - kmem_cache_free(field_cachep, field); - - return -ENOMEM; } int trace_define_field(struct ftrace_event_call *call, const char *type, -- cgit v1.2.3 From 11682a41618f8094cb7a9330b4b6a12ffaef5774 Mon Sep 17 00:00:00 2001 From: Marcus Gelderie Date: Tue, 4 Jun 2013 09:32:09 +0200 Subject: alarmtimer: Export symbols of functions declared in linux/alarmtimer.h Export symbols so they can be used by drivers/staging/android/alarm-dev.c if it is built as a module. So far alarm-dev is built-in but module support is planned (see drivers/staging/android/TODO). Signed-off-by: Marcus Gelderie [jstultz: tweaked commit message, also export newly added functions] Signed-off-by: John Stultz --- kernel/time/alarmtimer.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/time/alarmtimer.c b/kernel/time/alarmtimer.c index 3e5cba274475..eec50fcef9e4 100644 --- a/kernel/time/alarmtimer.c +++ b/kernel/time/alarmtimer.c @@ -204,6 +204,7 @@ ktime_t alarm_expires_remaining(const struct alarm *alarm) struct alarm_base *base = &alarm_bases[alarm->type]; return ktime_sub(alarm->node.expires, base->gettime()); } +EXPORT_SYMBOL_GPL(alarm_expires_remaining); #ifdef CONFIG_RTC_CLASS /** @@ -309,6 +310,7 @@ void alarm_init(struct alarm *alarm, enum alarmtimer_type type, alarm->type = type; alarm->state = ALARMTIMER_STATE_INACTIVE; } +EXPORT_SYMBOL_GPL(alarm_init); /** * alarm_start - Sets an absolute alarm to fire @@ -329,6 +331,7 @@ int alarm_start(struct alarm *alarm, ktime_t start) spin_unlock_irqrestore(&base->lock, flags); return ret; } +EXPORT_SYMBOL_GPL(alarm_start); /** * alarm_start_relative - Sets a relative alarm to fire @@ -342,6 +345,7 @@ int alarm_start_relative(struct alarm *alarm, ktime_t start) start = ktime_add(start, base->gettime()); return alarm_start(alarm, start); } +EXPORT_SYMBOL_GPL(alarm_start_relative); void alarm_restart(struct alarm *alarm) { @@ -354,6 +358,7 @@ void alarm_restart(struct alarm *alarm) alarmtimer_enqueue(base, alarm); spin_unlock_irqrestore(&base->lock, flags); } +EXPORT_SYMBOL_GPL(alarm_restart); /** * alarm_try_to_cancel - Tries to cancel an alarm timer @@ -375,6 +380,7 @@ int alarm_try_to_cancel(struct alarm *alarm) spin_unlock_irqrestore(&base->lock, flags); return ret; } +EXPORT_SYMBOL_GPL(alarm_try_to_cancel); /** @@ -392,6 +398,7 @@ int alarm_cancel(struct alarm *alarm) cpu_relax(); } } +EXPORT_SYMBOL_GPL(alarm_cancel); u64 alarm_forward(struct alarm *alarm, ktime_t now, ktime_t interval) @@ -424,6 +431,7 @@ u64 alarm_forward(struct alarm *alarm, ktime_t now, ktime_t interval) alarm->node.expires = ktime_add(alarm->node.expires, interval); return overrun; } +EXPORT_SYMBOL_GPL(alarm_forward); u64 alarm_forward_now(struct alarm *alarm, ktime_t interval) { @@ -431,7 +439,7 @@ u64 alarm_forward_now(struct alarm *alarm, ktime_t interval) return alarm_forward(alarm, base->gettime(), interval); } - +EXPORT_SYMBOL_GPL(alarm_forward_now); /** -- cgit v1.2.3 From 38ff87f77af0b5a93fc8581cff1d6e5692ab8970 Mon Sep 17 00:00:00 2001 From: Stephen Boyd Date: Sat, 1 Jun 2013 23:39:40 -0700 Subject: sched_clock: Make ARM's sched_clock generic for all architectures Nothing about the sched_clock implementation in the ARM port is specific to the architecture. Generalize the code so that other architectures can use it by selecting GENERIC_SCHED_CLOCK. Signed-off-by: Stephen Boyd [jstultz: Merge minor collisions with other patches in my tree] Signed-off-by: John Stultz --- kernel/time/Makefile | 1 + kernel/time/sched_clock.c | 215 ++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 216 insertions(+) create mode 100644 kernel/time/sched_clock.c (limited to 'kernel') diff --git a/kernel/time/Makefile b/kernel/time/Makefile index d52ac8bf0006..9250130646f5 100644 --- a/kernel/time/Makefile +++ b/kernel/time/Makefile @@ -4,6 +4,7 @@ obj-y += timeconv.o posix-clock.o alarmtimer.o obj-$(CONFIG_GENERIC_CLOCKEVENTS_BUILD) += clockevents.o obj-$(CONFIG_GENERIC_CLOCKEVENTS) += tick-common.o obj-$(CONFIG_GENERIC_CLOCKEVENTS_BROADCAST) += tick-broadcast.o +obj-$(CONFIG_GENERIC_SCHED_CLOCK) += sched_clock.o obj-$(CONFIG_TICK_ONESHOT) += tick-oneshot.o obj-$(CONFIG_TICK_ONESHOT) += tick-sched.o obj-$(CONFIG_TIMER_STATS) += timer_stats.o diff --git a/kernel/time/sched_clock.c b/kernel/time/sched_clock.c new file mode 100644 index 000000000000..aad1ae6077ef --- /dev/null +++ b/kernel/time/sched_clock.c @@ -0,0 +1,215 @@ +/* + * sched_clock.c: support for extending counters to full 64-bit ns counter + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ +#include +#include +#include +#include +#include +#include +#include +#include +#include + +struct clock_data { + u64 epoch_ns; + u32 epoch_cyc; + u32 epoch_cyc_copy; + unsigned long rate; + u32 mult; + u32 shift; + bool suspended; +}; + +static void sched_clock_poll(unsigned long wrap_ticks); +static DEFINE_TIMER(sched_clock_timer, sched_clock_poll, 0, 0); +static int irqtime = -1; + +core_param(irqtime, irqtime, int, 0400); + +static struct clock_data cd = { + .mult = NSEC_PER_SEC / HZ, +}; + +static u32 __read_mostly sched_clock_mask = 0xffffffff; + +static u32 notrace jiffy_sched_clock_read(void) +{ + return (u32)(jiffies - INITIAL_JIFFIES); +} + +static u32 __read_mostly (*read_sched_clock)(void) = jiffy_sched_clock_read; + +static inline u64 notrace cyc_to_ns(u64 cyc, u32 mult, u32 shift) +{ + return (cyc * mult) >> shift; +} + +static unsigned long long notrace cyc_to_sched_clock(u32 cyc, u32 mask) +{ + u64 epoch_ns; + u32 epoch_cyc; + + /* + * Load the epoch_cyc and epoch_ns atomically. We do this by + * ensuring that we always write epoch_cyc, epoch_ns and + * epoch_cyc_copy in strict order, and read them in strict order. + * If epoch_cyc and epoch_cyc_copy are not equal, then we're in + * the middle of an update, and we should repeat the load. + */ + do { + epoch_cyc = cd.epoch_cyc; + smp_rmb(); + epoch_ns = cd.epoch_ns; + smp_rmb(); + } while (epoch_cyc != cd.epoch_cyc_copy); + + return epoch_ns + cyc_to_ns((cyc - epoch_cyc) & mask, cd.mult, cd.shift); +} + +/* + * Atomically update the sched_clock epoch. + */ +static void notrace update_sched_clock(void) +{ + unsigned long flags; + u32 cyc; + u64 ns; + + cyc = read_sched_clock(); + ns = cd.epoch_ns + + cyc_to_ns((cyc - cd.epoch_cyc) & sched_clock_mask, + cd.mult, cd.shift); + /* + * Write epoch_cyc and epoch_ns in a way that the update is + * detectable in cyc_to_fixed_sched_clock(). + */ + raw_local_irq_save(flags); + cd.epoch_cyc_copy = cyc; + smp_wmb(); + cd.epoch_ns = ns; + smp_wmb(); + cd.epoch_cyc = cyc; + raw_local_irq_restore(flags); +} + +static void sched_clock_poll(unsigned long wrap_ticks) +{ + mod_timer(&sched_clock_timer, round_jiffies(jiffies + wrap_ticks)); + update_sched_clock(); +} + +void __init setup_sched_clock(u32 (*read)(void), int bits, unsigned long rate) +{ + unsigned long r, w; + u64 res, wrap; + char r_unit; + + if (cd.rate > rate) + return; + + BUG_ON(bits > 32); + WARN_ON(!irqs_disabled()); + read_sched_clock = read; + sched_clock_mask = (1 << bits) - 1; + cd.rate = rate; + + /* calculate the mult/shift to convert counter ticks to ns. */ + clocks_calc_mult_shift(&cd.mult, &cd.shift, rate, NSEC_PER_SEC, 0); + + r = rate; + if (r >= 4000000) { + r /= 1000000; + r_unit = 'M'; + } else if (r >= 1000) { + r /= 1000; + r_unit = 'k'; + } else + r_unit = ' '; + + /* calculate how many ns until we wrap */ + wrap = cyc_to_ns((1ULL << bits) - 1, cd.mult, cd.shift); + do_div(wrap, NSEC_PER_MSEC); + w = wrap; + + /* calculate the ns resolution of this counter */ + res = cyc_to_ns(1ULL, cd.mult, cd.shift); + pr_info("sched_clock: %u bits at %lu%cHz, resolution %lluns, wraps every %lums\n", + bits, r, r_unit, res, w); + + /* + * Start the timer to keep sched_clock() properly updated and + * sets the initial epoch. + */ + sched_clock_timer.data = msecs_to_jiffies(w - (w / 10)); + update_sched_clock(); + + /* + * Ensure that sched_clock() starts off at 0ns + */ + cd.epoch_ns = 0; + + /* Enable IRQ time accounting if we have a fast enough sched_clock */ + if (irqtime > 0 || (irqtime == -1 && rate >= 1000000)) + enable_sched_clock_irqtime(); + + pr_debug("Registered %pF as sched_clock source\n", read); +} + +static unsigned long long notrace sched_clock_32(void) +{ + u32 cyc = read_sched_clock(); + return cyc_to_sched_clock(cyc, sched_clock_mask); +} + +unsigned long long __read_mostly (*sched_clock_func)(void) = sched_clock_32; + +unsigned long long notrace sched_clock(void) +{ + if (cd.suspended) + return cd.epoch_ns; + + return sched_clock_func(); +} + +void __init sched_clock_postinit(void) +{ + /* + * If no sched_clock function has been provided at that point, + * make it the final one one. + */ + if (read_sched_clock == jiffy_sched_clock_read) + setup_sched_clock(jiffy_sched_clock_read, 32, HZ); + + sched_clock_poll(sched_clock_timer.data); +} + +static int sched_clock_suspend(void) +{ + sched_clock_poll(sched_clock_timer.data); + cd.suspended = true; + return 0; +} + +static void sched_clock_resume(void) +{ + cd.epoch_cyc = read_sched_clock(); + cd.epoch_cyc_copy = cd.epoch_cyc; + cd.suspended = false; +} + +static struct syscore_ops sched_clock_ops = { + .suspend = sched_clock_suspend, + .resume = sched_clock_resume, +}; + +static int __init sched_clock_syscore_init(void) +{ + register_syscore_ops(&sched_clock_ops); + return 0; +} +device_initcall(sched_clock_syscore_init); -- cgit v1.2.3 From 16e53dbf10a2d7e228709a7286310e629ede5e45 Mon Sep 17 00:00:00 2001 From: "Srivatsa S. Bhat" Date: Wed, 12 Jun 2013 14:04:36 -0700 Subject: CPU hotplug: provide a generic helper to disable/enable CPU hotplug There are instances in the kernel where we would like to disable CPU hotplug (from sysfs) during some important operation. Today the freezer code depends on this and the code to do it was kinda tailor-made for that. Restructure the code and make it generic enough to be useful for other usecases too. Signed-off-by: Srivatsa S. Bhat Signed-off-by: Robin Holt Cc: H. Peter Anvin Cc: Thomas Gleixner Cc: Ingo Molnar Cc: Russ Anderson Cc: Robin Holt Cc: Russell King Cc: Guan Xuetao Cc: Shawn Guo Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/cpu.c | 55 +++++++++++++++++++++++-------------------------------- 1 file changed, 23 insertions(+), 32 deletions(-) (limited to 'kernel') diff --git a/kernel/cpu.c b/kernel/cpu.c index b5e4ab2d427e..198a38883e64 100644 --- a/kernel/cpu.c +++ b/kernel/cpu.c @@ -133,6 +133,27 @@ static void cpu_hotplug_done(void) mutex_unlock(&cpu_hotplug.lock); } +/* + * Wait for currently running CPU hotplug operations to complete (if any) and + * disable future CPU hotplug (from sysfs). The 'cpu_add_remove_lock' protects + * the 'cpu_hotplug_disabled' flag. The same lock is also acquired by the + * hotplug path before performing hotplug operations. So acquiring that lock + * guarantees mutual exclusion from any currently running hotplug operations. + */ +void cpu_hotplug_disable(void) +{ + cpu_maps_update_begin(); + cpu_hotplug_disabled = 1; + cpu_maps_update_done(); +} + +void cpu_hotplug_enable(void) +{ + cpu_maps_update_begin(); + cpu_hotplug_disabled = 0; + cpu_maps_update_done(); +} + #else /* #if CONFIG_HOTPLUG_CPU */ static void cpu_hotplug_begin(void) {} static void cpu_hotplug_done(void) {} @@ -540,36 +561,6 @@ static int __init alloc_frozen_cpus(void) } core_initcall(alloc_frozen_cpus); -/* - * Prevent regular CPU hotplug from racing with the freezer, by disabling CPU - * hotplug when tasks are about to be frozen. Also, don't allow the freezer - * to continue until any currently running CPU hotplug operation gets - * completed. - * To modify the 'cpu_hotplug_disabled' flag, we need to acquire the - * 'cpu_add_remove_lock'. And this same lock is also taken by the regular - * CPU hotplug path and released only after it is complete. Thus, we - * (and hence the freezer) will block here until any currently running CPU - * hotplug operation gets completed. - */ -void cpu_hotplug_disable_before_freeze(void) -{ - cpu_maps_update_begin(); - cpu_hotplug_disabled = 1; - cpu_maps_update_done(); -} - - -/* - * When tasks have been thawed, re-enable regular CPU hotplug (which had been - * disabled while beginning to freeze tasks). - */ -void cpu_hotplug_enable_after_thaw(void) -{ - cpu_maps_update_begin(); - cpu_hotplug_disabled = 0; - cpu_maps_update_done(); -} - /* * When callbacks for CPU hotplug notifications are being executed, we must * ensure that the state of the system with respect to the tasks being frozen @@ -589,12 +580,12 @@ cpu_hotplug_pm_callback(struct notifier_block *nb, case PM_SUSPEND_PREPARE: case PM_HIBERNATION_PREPARE: - cpu_hotplug_disable_before_freeze(); + cpu_hotplug_disable(); break; case PM_POST_SUSPEND: case PM_POST_HIBERNATION: - cpu_hotplug_enable_after_thaw(); + cpu_hotplug_enable(); break; default: -- cgit v1.2.3 From cf7df378aa4ff7da3a44769b7ff6e9eef1a9f3db Mon Sep 17 00:00:00 2001 From: Robin Holt Date: Wed, 12 Jun 2013 14:04:37 -0700 Subject: reboot: rigrate shutdown/reboot to boot cpu We recently noticed that reboot of a 1024 cpu machine takes approx 16 minutes of just stopping the cpus. The slowdown was tracked to commit f96972f2dc63 ("kernel/sys.c: call disable_nonboot_cpus() in kernel_restart()"). The current implementation does all the work of hot removing the cpus before halting the system. We are switching to just migrating to the boot cpu and then continuing with shutdown/reboot. This also has the effect of not breaking x86's command line parameter for specifying the reboot cpu. Note, this code was shamelessly copied from arch/x86/kernel/reboot.c with bits removed pertaining to the reboot_cpu command line parameter. Signed-off-by: Robin Holt Tested-by: Shawn Guo Cc: "Srivatsa S. Bhat" Cc: H. Peter Anvin Cc: Thomas Gleixner Cc: Ingo Molnar Cc: Russ Anderson Cc: Robin Holt Cc: Russell King Cc: Guan Xuetao Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sys.c | 29 ++++++++++++++++++++++++++--- 1 file changed, 26 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/sys.c b/kernel/sys.c index b95d3c72ba21..2bbd9a73b54c 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -362,6 +362,29 @@ int unregister_reboot_notifier(struct notifier_block *nb) } EXPORT_SYMBOL(unregister_reboot_notifier); +/* Add backwards compatibility for stable trees. */ +#ifndef PF_NO_SETAFFINITY +#define PF_NO_SETAFFINITY PF_THREAD_BOUND +#endif + +static void migrate_to_reboot_cpu(void) +{ + /* The boot cpu is always logical cpu 0 */ + int cpu = 0; + + cpu_hotplug_disable(); + + /* Make certain the cpu I'm about to reboot on is online */ + if (!cpu_online(cpu)) + cpu = cpumask_first(cpu_online_mask); + + /* Prevent races with other tasks migrating this task */ + current->flags |= PF_NO_SETAFFINITY; + + /* Make certain I only run on the appropriate processor */ + set_cpus_allowed_ptr(current, cpumask_of(cpu)); +} + /** * kernel_restart - reboot the system * @cmd: pointer to buffer containing command to execute for restart @@ -373,7 +396,7 @@ EXPORT_SYMBOL(unregister_reboot_notifier); void kernel_restart(char *cmd) { kernel_restart_prepare(cmd); - disable_nonboot_cpus(); + migrate_to_reboot_cpu(); syscore_shutdown(); if (!cmd) printk(KERN_EMERG "Restarting system.\n"); @@ -400,7 +423,7 @@ static void kernel_shutdown_prepare(enum system_states state) void kernel_halt(void) { kernel_shutdown_prepare(SYSTEM_HALT); - disable_nonboot_cpus(); + migrate_to_reboot_cpu(); syscore_shutdown(); printk(KERN_EMERG "System halted.\n"); kmsg_dump(KMSG_DUMP_HALT); @@ -419,7 +442,7 @@ void kernel_power_off(void) kernel_shutdown_prepare(SYSTEM_POWER_OFF); if (pm_power_off_prepare) pm_power_off_prepare(); - disable_nonboot_cpus(); + migrate_to_reboot_cpu(); syscore_shutdown(); printk(KERN_EMERG "Power down.\n"); kmsg_dump(KMSG_DUMP_POWEROFF); -- cgit v1.2.3 From 637241a900cbd982f744d44646b48a273d609b34 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Wed, 12 Jun 2013 14:04:39 -0700 Subject: kmsg: honor dmesg_restrict sysctl on /dev/kmsg The dmesg_restrict sysctl currently covers the syslog method for access dmesg, however /dev/kmsg isn't covered by the same protections. Most people haven't noticed because util-linux dmesg(1) defaults to using the syslog method for access in older versions. With util-linux dmesg(1) defaults to reading directly from /dev/kmsg. To fix /dev/kmsg, let's compare the existing interfaces and what they allow: - /proc/kmsg allows: - open (SYSLOG_ACTION_OPEN) if CAP_SYSLOG since it uses a destructive single-reader interface (SYSLOG_ACTION_READ). - everything, after an open. - syslog syscall allows: - anything, if CAP_SYSLOG. - SYSLOG_ACTION_READ_ALL and SYSLOG_ACTION_SIZE_BUFFER, if dmesg_restrict==0. - nothing else (EPERM). The use-cases were: - dmesg(1) needs to do non-destructive SYSLOG_ACTION_READ_ALLs. - sysklog(1) needs to open /proc/kmsg, drop privs, and still issue the destructive SYSLOG_ACTION_READs. AIUI, dmesg(1) is moving to /dev/kmsg, and systemd-journald doesn't clear the ring buffer. Based on the comments in devkmsg_llseek, it sounds like actions besides reading aren't going to be supported by /dev/kmsg (i.e. SYSLOG_ACTION_CLEAR), so we have a strict subset of the non-destructive syslog syscall actions. To this end, move the check as Josh had done, but also rename the constants to reflect their new uses (SYSLOG_FROM_CALL becomes SYSLOG_FROM_READER, and SYSLOG_FROM_FILE becomes SYSLOG_FROM_PROC). SYSLOG_FROM_READER allows non-destructive actions, and SYSLOG_FROM_PROC allows destructive actions after a capabilities-constrained SYSLOG_ACTION_OPEN check. - /dev/kmsg allows: - open if CAP_SYSLOG or dmesg_restrict==0 - reading/polling, after open Addresses https://bugzilla.redhat.com/show_bug.cgi?id=903192 [akpm@linux-foundation.org: use pr_warn_once()] Signed-off-by: Kees Cook Reported-by: Christian Kujau Tested-by: Josh Boyer Cc: Kay Sievers Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/printk.c | 91 +++++++++++++++++++++++++++++++-------------------------- 1 file changed, 50 insertions(+), 41 deletions(-) (limited to 'kernel') diff --git a/kernel/printk.c b/kernel/printk.c index fa36e1494420..8212c1aef125 100644 --- a/kernel/printk.c +++ b/kernel/printk.c @@ -363,6 +363,53 @@ static void log_store(int facility, int level, log_next_seq++; } +#ifdef CONFIG_SECURITY_DMESG_RESTRICT +int dmesg_restrict = 1; +#else +int dmesg_restrict; +#endif + +static int syslog_action_restricted(int type) +{ + if (dmesg_restrict) + return 1; + /* + * Unless restricted, we allow "read all" and "get buffer size" + * for everybody. + */ + return type != SYSLOG_ACTION_READ_ALL && + type != SYSLOG_ACTION_SIZE_BUFFER; +} + +static int check_syslog_permissions(int type, bool from_file) +{ + /* + * If this is from /proc/kmsg and we've already opened it, then we've + * already done the capabilities checks at open time. + */ + if (from_file && type != SYSLOG_ACTION_OPEN) + return 0; + + if (syslog_action_restricted(type)) { + if (capable(CAP_SYSLOG)) + return 0; + /* + * For historical reasons, accept CAP_SYS_ADMIN too, with + * a warning. + */ + if (capable(CAP_SYS_ADMIN)) { + pr_warn_once("%s (%d): Attempt to access syslog with " + "CAP_SYS_ADMIN but no CAP_SYSLOG " + "(deprecated).\n", + current->comm, task_pid_nr(current)); + return 0; + } + return -EPERM; + } + return security_syslog(type); +} + + /* /dev/kmsg - userspace message inject/listen interface */ struct devkmsg_user { u64 seq; @@ -620,7 +667,8 @@ static int devkmsg_open(struct inode *inode, struct file *file) if ((file->f_flags & O_ACCMODE) == O_WRONLY) return 0; - err = security_syslog(SYSLOG_ACTION_READ_ALL); + err = check_syslog_permissions(SYSLOG_ACTION_READ_ALL, + SYSLOG_FROM_READER); if (err) return err; @@ -813,45 +861,6 @@ static inline void boot_delay_msec(int level) } #endif -#ifdef CONFIG_SECURITY_DMESG_RESTRICT -int dmesg_restrict = 1; -#else -int dmesg_restrict; -#endif - -static int syslog_action_restricted(int type) -{ - if (dmesg_restrict) - return 1; - /* Unless restricted, we allow "read all" and "get buffer size" for everybody */ - return type != SYSLOG_ACTION_READ_ALL && type != SYSLOG_ACTION_SIZE_BUFFER; -} - -static int check_syslog_permissions(int type, bool from_file) -{ - /* - * If this is from /proc/kmsg and we've already opened it, then we've - * already done the capabilities checks at open time. - */ - if (from_file && type != SYSLOG_ACTION_OPEN) - return 0; - - if (syslog_action_restricted(type)) { - if (capable(CAP_SYSLOG)) - return 0; - /* For historical reasons, accept CAP_SYS_ADMIN too, with a warning */ - if (capable(CAP_SYS_ADMIN)) { - printk_once(KERN_WARNING "%s (%d): " - "Attempt to access syslog with CAP_SYS_ADMIN " - "but no CAP_SYSLOG (deprecated).\n", - current->comm, task_pid_nr(current)); - return 0; - } - return -EPERM; - } - return 0; -} - #if defined(CONFIG_PRINTK_TIME) static bool printk_time = 1; #else @@ -1249,7 +1258,7 @@ out: SYSCALL_DEFINE3(syslog, int, type, char __user *, buf, int, len) { - return do_syslog(type, buf, len, SYSLOG_FROM_CALL); + return do_syslog(type, buf, len, SYSLOG_FROM_READER); } /* -- cgit v1.2.3 From f000cfdde5de4fc15dead5ccf524359c07eadf2b Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Wed, 12 Jun 2013 14:04:46 -0700 Subject: audit: wait_for_auditd() should use TASK_UNINTERRUPTIBLE audit_log_start() does wait_for_auditd() in a loop until audit_backlog_wait_time passes or audit_skb_queue has a room. If signal_pending() is true this becomes a busy-wait loop, schedule() in TASK_INTERRUPTIBLE won't block. Thanks to Guy for fully investigating and explaining the problem. (akpm: that'll cause the system to lock up on a non-preemptible uniprocessor kernel) (Guy: "Our customer was in fact running a uniprocessor machine, and they reported a system hang.") Signed-off-by: Oleg Nesterov Reported-by: Guy Streeter Cc: Eric Paris Cc: Al Viro Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/audit.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/audit.c b/kernel/audit.c index 21c7fa615bd3..91e53d04b6a9 100644 --- a/kernel/audit.c +++ b/kernel/audit.c @@ -1056,7 +1056,7 @@ static inline void audit_get_stamp(struct audit_context *ctx, static void wait_for_auditd(unsigned long sleep_time) { DECLARE_WAITQUEUE(wait, current); - set_current_state(TASK_INTERRUPTIBLE); + set_current_state(TASK_UNINTERRUPTIBLE); add_wait_queue(&audit_backlog_wait, &wait); if (audit_backlog_limit && -- cgit v1.2.3 From 736f3203a06eafd0944103775a98584082744c6b Mon Sep 17 00:00:00 2001 From: Chen Gang Date: Wed, 12 Jun 2013 14:05:07 -0700 Subject: kernel/audit_tree.c:audit_add_tree_rule(): protect `rule' from kill_rules() audit_add_tree_rule() must set 'rule->tree = NULL;' firstly, to protect the rule itself freed in kill_rules(). The reason is when it is killed, the 'rule' itself may have already released, we should not access it. one example: we add a rule to an inode, just at the same time the other task is deleting this inode. The work flow for adding a rule: audit_receive() -> (need audit_cmd_mutex lock) audit_receive_skb() -> audit_receive_msg() -> audit_receive_filter() -> audit_add_rule() -> audit_add_tree_rule() -> (need audit_filter_mutex lock) ... unlock audit_filter_mutex get_tree() ... iterate_mounts() -> (iterate all related inodes) tag_mount() -> tag_trunk() -> create_trunk() -> (assume it is 1st rule) fsnotify_add_mark() -> fsnotify_add_inode_mark() -> (add mark to inode->i_fsnotify_marks) ... get_tree(); (each inode will get one) ... lock audit_filter_mutex The work flow for deleting an inode: __destroy_inode() -> fsnotify_inode_delete() -> __fsnotify_inode_delete() -> fsnotify_clear_marks_by_inode() -> (get mark from inode->i_fsnotify_marks) fsnotify_destroy_mark() -> fsnotify_destroy_mark_locked() -> audit_tree_freeing_mark() -> evict_chunk() -> ... tree->goner = 1 ... kill_rules() -> (assume current->audit_context == NULL) call_rcu() -> (rule->tree != NULL) audit_free_rule_rcu() -> audit_free_rule() ... audit_schedule_prune() -> (assume current->audit_context == NULL) kthread_run() -> (need audit_cmd_mutex and audit_filter_mutex lock) prune_one() -> (delete it from prue_list) put_tree(); (match the original get_tree above) Signed-off-by: Chen Gang Cc: Eric Paris Cc: Al Viro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/audit_tree.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/audit_tree.c b/kernel/audit_tree.c index a291aa23fb3f..43c307dc9453 100644 --- a/kernel/audit_tree.c +++ b/kernel/audit_tree.c @@ -658,6 +658,7 @@ int audit_add_tree_rule(struct audit_krule *rule) struct vfsmount *mnt; int err; + rule->tree = NULL; list_for_each_entry(tree, &tree_list, list) { if (!strcmp(seed->pathname, tree->pathname)) { put_tree(seed); -- cgit v1.2.3 From 33ad801dfb5c8b1127c72fdb745ce8c630150f3f Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Sun, 9 Jun 2013 17:15:08 +0800 Subject: cpuset: record old_mems_allowed in struct cpuset When we update a cpuset's mems_allowed and thus update tasks' mems_allowed, it's required to pass the old mems_allowed and new mems_allowed to cpuset_migrate_mm(). Currently we save old mems_allowed in a temp local variable before changing cpuset->mems_allowed. This patch changes it by saving old mems_allowed in cpuset->old_mems_allowed. This currently won't change any behavior, but it will later allow us to keep tasks in empty cpusets. v3: restored "cpuset_attach_nodemask_to = cs->mems_allowed" Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cpuset.c | 61 ++++++++++++++++++++++++++++++++++----------------------- 1 file changed, 36 insertions(+), 25 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 608fe1308b22..2b4554588a04 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -88,6 +88,18 @@ struct cpuset { cpumask_var_t cpus_allowed; /* CPUs allowed to tasks in cpuset */ nodemask_t mems_allowed; /* Memory Nodes allowed to tasks */ + /* + * This is old Memory Nodes tasks took on. + * + * - top_cpuset.old_mems_allowed is initialized to mems_allowed. + * - A new cpuset's old_mems_allowed is initialized when some + * task is moved into it. + * - old_mems_allowed is used in cpuset_migrate_mm() when we change + * cpuset.mems_allowed and have tasks' nodemask updated, and + * then old_mems_allowed is updated to mems_allowed. + */ + nodemask_t old_mems_allowed; + struct fmeter fmeter; /* memory_pressure filter */ /* @@ -972,16 +984,12 @@ static void cpuset_change_task_nodemask(struct task_struct *tsk, static void cpuset_change_nodemask(struct task_struct *p, struct cgroup_scanner *scan) { + struct cpuset *cs = cgroup_cs(scan->cg); struct mm_struct *mm; - struct cpuset *cs; int migrate; - const nodemask_t *oldmem = scan->data; - static nodemask_t newmems; /* protected by cpuset_mutex */ - - cs = cgroup_cs(scan->cg); - guarantee_online_mems(cs, &newmems); + nodemask_t *newmems = scan->data; - cpuset_change_task_nodemask(p, &newmems); + cpuset_change_task_nodemask(p, newmems); mm = get_task_mm(p); if (!mm) @@ -991,7 +999,7 @@ static void cpuset_change_nodemask(struct task_struct *p, mpol_rebind_mm(mm, &cs->mems_allowed); if (migrate) - cpuset_migrate_mm(mm, oldmem, &cs->mems_allowed); + cpuset_migrate_mm(mm, &cs->old_mems_allowed, newmems); mmput(mm); } @@ -1000,25 +1008,26 @@ static void *cpuset_being_rebound; /** * update_tasks_nodemask - Update the nodemasks of tasks in the cpuset. * @cs: the cpuset in which each task's mems_allowed mask needs to be changed - * @oldmem: old mems_allowed of cpuset cs * @heap: if NULL, defer allocating heap memory to cgroup_scan_tasks() * * Called with cpuset_mutex held * No return value. It's guaranteed that cgroup_scan_tasks() always returns 0 * if @heap != NULL. */ -static void update_tasks_nodemask(struct cpuset *cs, const nodemask_t *oldmem, - struct ptr_heap *heap) +static void update_tasks_nodemask(struct cpuset *cs, struct ptr_heap *heap) { + static nodemask_t newmems; /* protected by cpuset_mutex */ struct cgroup_scanner scan; cpuset_being_rebound = cs; /* causes mpol_dup() rebind */ + guarantee_online_mems(cs, &newmems); + scan.cg = cs->css.cgroup; scan.test_task = NULL; scan.process_task = cpuset_change_nodemask; scan.heap = heap; - scan.data = (nodemask_t *)oldmem; + scan.data = &newmems; /* * The mpol_rebind_mm() call takes mmap_sem, which we couldn't @@ -1032,6 +1041,12 @@ static void update_tasks_nodemask(struct cpuset *cs, const nodemask_t *oldmem, */ cgroup_scan_tasks(&scan); + /* + * All the tasks' nodemasks have been updated, update + * cs->old_mems_allowed. + */ + cs->old_mems_allowed = newmems; + /* We're done rebinding vmas to this cpuset's new mems_allowed. */ cpuset_being_rebound = NULL; } @@ -1052,13 +1067,9 @@ static void update_tasks_nodemask(struct cpuset *cs, const nodemask_t *oldmem, static int update_nodemask(struct cpuset *cs, struct cpuset *trialcs, const char *buf) { - NODEMASK_ALLOC(nodemask_t, oldmem, GFP_KERNEL); int retval; struct ptr_heap heap; - if (!oldmem) - return -ENOMEM; - /* * top_cpuset.mems_allowed tracks node_stats[N_MEMORY]; * it's read-only @@ -1087,8 +1098,8 @@ static int update_nodemask(struct cpuset *cs, struct cpuset *trialcs, goto done; } } - *oldmem = cs->mems_allowed; - if (nodes_equal(*oldmem, trialcs->mems_allowed)) { + + if (nodes_equal(cs->mems_allowed, trialcs->mems_allowed)) { retval = 0; /* Too easy - nothing to do */ goto done; } @@ -1104,11 +1115,10 @@ static int update_nodemask(struct cpuset *cs, struct cpuset *trialcs, cs->mems_allowed = trialcs->mems_allowed; mutex_unlock(&callback_mutex); - update_tasks_nodemask(cs, oldmem, &heap); + update_tasks_nodemask(cs, &heap); heap_free(&heap); done: - NODEMASK_FREE(oldmem); return retval; } @@ -1431,6 +1441,8 @@ static void cpuset_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) mmput(mm); } + cs->old_mems_allowed = cpuset_attach_nodemask_to; + cs->attach_in_progress--; if (!cs->attach_in_progress) wake_up(&cpuset_attach_wq); @@ -1985,7 +1997,7 @@ static void remove_tasks_in_empty_cpuset(struct cpuset *cs) static void cpuset_hotplug_update_tasks(struct cpuset *cs) { static cpumask_t off_cpus; - static nodemask_t off_mems, tmp_mems; + static nodemask_t off_mems; bool is_empty; retry: @@ -2015,11 +2027,10 @@ retry: /* remove offline mems from @cs */ if (!nodes_empty(off_mems)) { - tmp_mems = cs->mems_allowed; mutex_lock(&callback_mutex); nodes_andnot(cs->mems_allowed, cs->mems_allowed, off_mems); mutex_unlock(&callback_mutex); - update_tasks_nodemask(cs, &tmp_mems, NULL); + update_tasks_nodemask(cs, NULL); } is_empty = cpumask_empty(cs->cpus_allowed) || @@ -2083,11 +2094,10 @@ static void cpuset_hotplug_workfn(struct work_struct *work) /* synchronize mems_allowed to N_MEMORY */ if (mems_updated) { - tmp_mems = top_cpuset.mems_allowed; mutex_lock(&callback_mutex); top_cpuset.mems_allowed = new_mems; mutex_unlock(&callback_mutex); - update_tasks_nodemask(&top_cpuset, &tmp_mems, NULL); + update_tasks_nodemask(&top_cpuset, NULL); } mutex_unlock(&cpuset_mutex); @@ -2158,6 +2168,7 @@ void __init cpuset_init_smp(void) { cpumask_copy(top_cpuset.cpus_allowed, cpu_active_mask); top_cpuset.mems_allowed = node_states[N_MEMORY]; + top_cpuset.old_mems_allowed = top_cpuset.mems_allowed; register_hotmemory_notifier(&cpuset_track_online_nodes_nb); } -- cgit v1.2.3 From 070b57fcacc9dfc23a180290079078373fb697e1 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Sun, 9 Jun 2013 17:15:22 +0800 Subject: cpuset: introduce effective_{cpumask|nodemask}_cpuset() effective_cpumask_cpuset() returns an ancestor cpuset which has non-empty cpumask. If a cpuset is empty and the tasks in it need to update their cpus_allowed, they take on the ancestor cpuset's cpumask. This currently won't change any behavior, but it will later allow us to keep tasks in empty cpusets. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cpuset.c | 76 ++++++++++++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 65 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 2b4554588a04..82ac1f862cbc 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -791,6 +791,45 @@ void rebuild_sched_domains(void) mutex_unlock(&cpuset_mutex); } +/* + * effective_cpumask_cpuset - return nearest ancestor with non-empty cpus + * @cs: the cpuset in interest + * + * A cpuset's effective cpumask is the cpumask of the nearest ancestor + * with non-empty cpus. We use effective cpumask whenever: + * - we update tasks' cpus_allowed. (they take on the ancestor's cpumask + * if the cpuset they reside in has no cpus) + * - we want to retrieve task_cs(tsk)'s cpus_allowed. + * + * Called with cpuset_mutex held. cpuset_cpus_allowed_fallback() is an + * exception. See comments there. + */ +static struct cpuset *effective_cpumask_cpuset(struct cpuset *cs) +{ + while (cpumask_empty(cs->cpus_allowed)) + cs = parent_cs(cs); + return cs; +} + +/* + * effective_nodemask_cpuset - return nearest ancestor with non-empty mems + * @cs: the cpuset in interest + * + * A cpuset's effective nodemask is the nodemask of the nearest ancestor + * with non-empty memss. We use effective nodemask whenever: + * - we update tasks' mems_allowed. (they take on the ancestor's nodemask + * if the cpuset they reside in has no mems) + * - we want to retrieve task_cs(tsk)'s mems_allowed. + * + * Called with cpuset_mutex held. + */ +static struct cpuset *effective_nodemask_cpuset(struct cpuset *cs) +{ + while (nodes_empty(cs->mems_allowed)) + cs = parent_cs(cs); + return cs; +} + /** * cpuset_change_cpumask - make a task's cpus_allowed the same as its cpuset's * @tsk: task to test @@ -805,7 +844,10 @@ void rebuild_sched_domains(void) static void cpuset_change_cpumask(struct task_struct *tsk, struct cgroup_scanner *scan) { - set_cpus_allowed_ptr(tsk, ((cgroup_cs(scan->cg))->cpus_allowed)); + struct cpuset *cpus_cs; + + cpus_cs = effective_cpumask_cpuset(cgroup_cs(scan->cg)); + set_cpus_allowed_ptr(tsk, cpus_cs->cpus_allowed); } /** @@ -920,12 +962,14 @@ static void cpuset_migrate_mm(struct mm_struct *mm, const nodemask_t *from, const nodemask_t *to) { struct task_struct *tsk = current; + struct cpuset *mems_cs; tsk->mems_allowed = *to; do_migrate_pages(mm, from, to, MPOL_MF_MOVE_ALL); - guarantee_online_mems(task_cs(tsk),&tsk->mems_allowed); + mems_cs = effective_nodemask_cpuset(task_cs(tsk)); + guarantee_online_mems(mems_cs, &tsk->mems_allowed); } /* @@ -1018,10 +1062,11 @@ static void update_tasks_nodemask(struct cpuset *cs, struct ptr_heap *heap) { static nodemask_t newmems; /* protected by cpuset_mutex */ struct cgroup_scanner scan; + struct cpuset *mems_cs = effective_nodemask_cpuset(cs); cpuset_being_rebound = cs; /* causes mpol_dup() rebind */ - guarantee_online_mems(cs, &newmems); + guarantee_online_mems(mems_cs, &newmems); scan.cg = cs->css.cgroup; scan.test_task = NULL; @@ -1405,6 +1450,8 @@ static void cpuset_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) struct cgroup *oldcgrp = cgroup_taskset_cur_cgroup(tset); struct cpuset *cs = cgroup_cs(cgrp); struct cpuset *oldcs = cgroup_cs(oldcgrp); + struct cpuset *cpus_cs = effective_cpumask_cpuset(cs); + struct cpuset *mems_cs = effective_nodemask_cpuset(cs); mutex_lock(&cpuset_mutex); @@ -1412,9 +1459,9 @@ static void cpuset_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) if (cs == &top_cpuset) cpumask_copy(cpus_attach, cpu_possible_mask); else - guarantee_online_cpus(cs, cpus_attach); + guarantee_online_cpus(cpus_cs, cpus_attach); - guarantee_online_mems(cs, &cpuset_attach_nodemask_to); + guarantee_online_mems(mems_cs, &cpuset_attach_nodemask_to); cgroup_taskset_for_each(task, cgrp, tset) { /* @@ -1434,9 +1481,11 @@ static void cpuset_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) cpuset_attach_nodemask_to = cs->mems_allowed; mm = get_task_mm(leader); if (mm) { + struct cpuset *mems_oldcs = effective_nodemask_cpuset(oldcs); + mpol_rebind_mm(mm, &cpuset_attach_nodemask_to); if (is_memory_migrate(cs)) - cpuset_migrate_mm(mm, &oldcs->mems_allowed, + cpuset_migrate_mm(mm, &mems_oldcs->mems_allowed, &cpuset_attach_nodemask_to); mmput(mm); } @@ -2186,20 +2235,23 @@ void __init cpuset_init_smp(void) void cpuset_cpus_allowed(struct task_struct *tsk, struct cpumask *pmask) { + struct cpuset *cpus_cs; + mutex_lock(&callback_mutex); task_lock(tsk); - guarantee_online_cpus(task_cs(tsk), pmask); + cpus_cs = effective_cpumask_cpuset(task_cs(tsk)); + guarantee_online_cpus(cpus_cs, pmask); task_unlock(tsk); mutex_unlock(&callback_mutex); } void cpuset_cpus_allowed_fallback(struct task_struct *tsk) { - const struct cpuset *cs; + const struct cpuset *cpus_cs; rcu_read_lock(); - cs = task_cs(tsk); - do_set_cpus_allowed(tsk, cs->cpus_allowed); + cpus_cs = effective_cpumask_cpuset(task_cs(tsk)); + do_set_cpus_allowed(tsk, cpus_cs->cpus_allowed); rcu_read_unlock(); /* @@ -2238,11 +2290,13 @@ void cpuset_init_current_mems_allowed(void) nodemask_t cpuset_mems_allowed(struct task_struct *tsk) { + struct cpuset *mems_cs; nodemask_t mask; mutex_lock(&callback_mutex); task_lock(tsk); - guarantee_online_mems(task_cs(tsk), &mask); + mems_cs = effective_nodemask_cpuset(task_cs(tsk)); + guarantee_online_mems(mems_cs, &mask); task_unlock(tsk); mutex_unlock(&callback_mutex); -- cgit v1.2.3 From 5c5cc62321d9df7a9a608346fc649c4528380c8f Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Sun, 9 Jun 2013 17:16:29 +0800 Subject: cpuset: allow to keep tasks in empty cpusets To achieve this: - We call update_tasks_cpumask/nodemask() for empty cpusets when hotplug happens, instead of moving tasks out of them. - When a cpuset's masks are changed by writing cpuset.cpus/mems, we also update tasks in child cpusets which are empty. v3: - do propagation work in one place for both hotplug and unplug v2: - drop rcu_read_lock before calling update_task_nodemask() and update_task_cpumask(), instead of using workqueue. - add documentation in include/linux/cgroup.h Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cpuset.c | 141 +++++++++++++++++++++++++++++++++++++++++++------------- 1 file changed, 110 insertions(+), 31 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 82ac1f862cbc..3473dd2580d1 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -874,6 +874,45 @@ static void update_tasks_cpumask(struct cpuset *cs, struct ptr_heap *heap) cgroup_scan_tasks(&scan); } +/* + * update_tasks_cpumask_hier - Update the cpumasks of tasks in the hierarchy. + * @root_cs: the root cpuset of the hierarchy + * @update_root: update root cpuset or not? + * @heap: the heap used by cgroup_scan_tasks() + * + * This will update cpumasks of tasks in @root_cs and all other empty cpusets + * which take on cpumask of @root_cs. + * + * Called with cpuset_mutex held + */ +static void update_tasks_cpumask_hier(struct cpuset *root_cs, + bool update_root, struct ptr_heap *heap) +{ + struct cpuset *cp; + struct cgroup *pos_cgrp; + + if (update_root) + update_tasks_cpumask(root_cs, heap); + + rcu_read_lock(); + cpuset_for_each_descendant_pre(cp, pos_cgrp, root_cs) { + /* skip the whole subtree if @cp have some CPU */ + if (!cpumask_empty(cp->cpus_allowed)) { + pos_cgrp = cgroup_rightmost_descendant(pos_cgrp); + continue; + } + if (!css_tryget(&cp->css)) + continue; + rcu_read_unlock(); + + update_tasks_cpumask(cp, heap); + + rcu_read_lock(); + css_put(&cp->css); + } + rcu_read_unlock(); +} + /** * update_cpumask - update the cpus_allowed mask of a cpuset and all tasks in it * @cs: the cpuset to consider @@ -925,11 +964,7 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs, cpumask_copy(cs->cpus_allowed, trialcs->cpus_allowed); mutex_unlock(&callback_mutex); - /* - * Scan tasks in the cpuset, and update the cpumasks of any - * that need an update. - */ - update_tasks_cpumask(cs, &heap); + update_tasks_cpumask_hier(cs, true, &heap); heap_free(&heap); @@ -1096,6 +1131,45 @@ static void update_tasks_nodemask(struct cpuset *cs, struct ptr_heap *heap) cpuset_being_rebound = NULL; } +/* + * update_tasks_nodemask_hier - Update the nodemasks of tasks in the hierarchy. + * @cs: the root cpuset of the hierarchy + * @update_root: update the root cpuset or not? + * @heap: the heap used by cgroup_scan_tasks() + * + * This will update nodemasks of tasks in @root_cs and all other empty cpusets + * which take on nodemask of @root_cs. + * + * Called with cpuset_mutex held + */ +static void update_tasks_nodemask_hier(struct cpuset *root_cs, + bool update_root, struct ptr_heap *heap) +{ + struct cpuset *cp; + struct cgroup *pos_cgrp; + + if (update_root) + update_tasks_nodemask(root_cs, heap); + + rcu_read_lock(); + cpuset_for_each_descendant_pre(cp, pos_cgrp, root_cs) { + /* skip the whole subtree if @cp have some CPU */ + if (!nodes_empty(cp->mems_allowed)) { + pos_cgrp = cgroup_rightmost_descendant(pos_cgrp); + continue; + } + if (!css_tryget(&cp->css)) + continue; + rcu_read_unlock(); + + update_tasks_nodemask(cp, heap); + + rcu_read_lock(); + css_put(&cp->css); + } + rcu_read_unlock(); +} + /* * Handle user request to change the 'mems' memory placement * of a cpuset. Needs to validate the request, update the @@ -1160,7 +1234,7 @@ static int update_nodemask(struct cpuset *cs, struct cpuset *trialcs, cs->mems_allowed = trialcs->mems_allowed; mutex_unlock(&callback_mutex); - update_tasks_nodemask(cs, &heap); + update_tasks_nodemask_hier(cs, true, &heap); heap_free(&heap); done: @@ -2048,6 +2122,7 @@ static void cpuset_hotplug_update_tasks(struct cpuset *cs) static cpumask_t off_cpus; static nodemask_t off_mems; bool is_empty; + bool sane = cgroup_sane_behavior(cs->css.cgroup); retry: wait_event(cpuset_attach_wq, cs->attach_in_progress == 0); @@ -2066,21 +2141,29 @@ retry: cpumask_andnot(&off_cpus, cs->cpus_allowed, top_cpuset.cpus_allowed); nodes_andnot(off_mems, cs->mems_allowed, top_cpuset.mems_allowed); - /* remove offline cpus from @cs */ - if (!cpumask_empty(&off_cpus)) { - mutex_lock(&callback_mutex); - cpumask_andnot(cs->cpus_allowed, cs->cpus_allowed, &off_cpus); - mutex_unlock(&callback_mutex); + mutex_lock(&callback_mutex); + cpumask_andnot(cs->cpus_allowed, cs->cpus_allowed, &off_cpus); + mutex_unlock(&callback_mutex); + + /* + * If sane_behavior flag is set, we need to update tasks' cpumask + * for empty cpuset to take on ancestor's cpumask. + */ + if ((sane && cpumask_empty(cs->cpus_allowed)) || + !cpumask_empty(&off_cpus)) update_tasks_cpumask(cs, NULL); - } - /* remove offline mems from @cs */ - if (!nodes_empty(off_mems)) { - mutex_lock(&callback_mutex); - nodes_andnot(cs->mems_allowed, cs->mems_allowed, off_mems); - mutex_unlock(&callback_mutex); + mutex_lock(&callback_mutex); + nodes_andnot(cs->mems_allowed, cs->mems_allowed, off_mems); + mutex_unlock(&callback_mutex); + + /* + * If sane_behavior flag is set, we need to update tasks' nodemask + * for empty cpuset to take on ancestor's nodemask. + */ + if ((sane && nodes_empty(cs->mems_allowed)) || + !nodes_empty(off_mems)) update_tasks_nodemask(cs, NULL); - } is_empty = cpumask_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed); @@ -2088,11 +2171,13 @@ retry: mutex_unlock(&cpuset_mutex); /* - * If @cs became empty, move tasks to the nearest ancestor with - * execution resources. This is full cgroup operation which will + * If sane_behavior flag is set, we'll keep tasks in empty cpusets. + * + * Otherwise move tasks to the nearest ancestor with execution + * resources. This is full cgroup operation which will * also call back into cpuset. Should be done outside any lock. */ - if (is_empty) + if (!sane && is_empty) remove_tasks_in_empty_cpuset(cs); } @@ -2114,10 +2199,9 @@ retry: */ static void cpuset_hotplug_workfn(struct work_struct *work) { - static cpumask_t new_cpus, tmp_cpus; - static nodemask_t new_mems, tmp_mems; + static cpumask_t new_cpus; + static nodemask_t new_mems; bool cpus_updated, mems_updated; - bool cpus_offlined, mems_offlined; mutex_lock(&cpuset_mutex); @@ -2126,12 +2210,7 @@ static void cpuset_hotplug_workfn(struct work_struct *work) new_mems = node_states[N_MEMORY]; cpus_updated = !cpumask_equal(top_cpuset.cpus_allowed, &new_cpus); - cpus_offlined = cpumask_andnot(&tmp_cpus, top_cpuset.cpus_allowed, - &new_cpus); - mems_updated = !nodes_equal(top_cpuset.mems_allowed, new_mems); - nodes_andnot(tmp_mems, top_cpuset.mems_allowed, new_mems); - mems_offlined = !nodes_empty(tmp_mems); /* synchronize cpus_allowed to cpu_active_mask */ if (cpus_updated) { @@ -2151,8 +2230,8 @@ static void cpuset_hotplug_workfn(struct work_struct *work) mutex_unlock(&cpuset_mutex); - /* if cpus or mems went down, we need to propagate to descendants */ - if (cpus_offlined || mems_offlined) { + /* if cpus or mems changed, we need to propagate to descendants */ + if (cpus_updated || mems_updated) { struct cpuset *cs; struct cgroup *pos_cgrp; -- cgit v1.2.3 From 88fa523bff295f1d60244a54833480b02f775152 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Sun, 9 Jun 2013 17:16:46 +0800 Subject: cpuset: allow to move tasks to empty cpusets Currently some cpuset behaviors are not friendly when cpuset is co-mounted with other cgroup controllers. Now with this patchset if cpuset is mounted with sane_behavior option, it behaves differently: - Tasks will be kept in empty cpusets when hotplug happens and take masks of ancestors with non-empty cpus/mems, instead of being moved to an ancestor. - A task can be moved into an empty cpuset, and again it takes masks of ancestors, so the user can drop a task into a newly created cgroup without having to do anything for it. As tasks can reside in empy cpusets, here're some rules: - They can be moved to another cpuset, regardless it's empty or not. - Though it takes masks from ancestors, it takes other configs from the empty cpuset. - If the ancestors' masks are changed, those tasks will also be updated to take new masks. v2: add documentation in include/linux/cgroup.h Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cpuset.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 3473dd2580d1..3b3fdfdd4d78 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -479,7 +479,7 @@ static int validate_change(const struct cpuset *cur, const struct cpuset *trial) */ ret = -ENOSPC; if ((cgroup_task_count(cur->css.cgroup) || cur->attach_in_progress) && - (cpumask_empty(trial->cpus_allowed) || + (cpumask_empty(trial->cpus_allowed) && nodes_empty(trial->mems_allowed))) goto out; @@ -1466,8 +1466,13 @@ static int cpuset_can_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) mutex_lock(&cpuset_mutex); + /* + * We allow to move tasks into an empty cpuset if sane_behavior + * flag is set. + */ ret = -ENOSPC; - if (cpumask_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed)) + if (!cgroup_sane_behavior(cgrp) && + (cpumask_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed))) goto out_unlock; cgroup_taskset_for_each(task, cgrp, tset) { -- cgit v1.2.3 From f047cecf2cfc9595b1f39c9aab383bb0682f5a53 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Thu, 13 Jun 2013 15:11:44 +0800 Subject: cpuset: fix to migrate mm correctly in a corner case Before moving tasks out of empty cpusets, update_tasks_nodemask() is called, which calls do_migrate_pages(xx, from, to). Then those tasks are moved to an ancestor, and do_migrate_pages() is called again. The first time: from = node_to_be_offlined, to = empty. The second time: from = empty, to = ancestor's nodemask. so looks like no pages will be migrated. Fix this by: - Don't call update_tasks_nodemask() on empty cpusets. - Pass cs->old_mems_allowed to do_migrate_pages(). v4: added comment in cpuset_hotplug_update_tasks() and rephased comment in cpuset_attach(). Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cpuset.c | 25 +++++++++++++++++++------ 1 file changed, 19 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 3b3fdfdd4d78..4c17d96bd3a5 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -1563,9 +1563,18 @@ static void cpuset_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) struct cpuset *mems_oldcs = effective_nodemask_cpuset(oldcs); mpol_rebind_mm(mm, &cpuset_attach_nodemask_to); - if (is_memory_migrate(cs)) - cpuset_migrate_mm(mm, &mems_oldcs->mems_allowed, + + /* + * old_mems_allowed is the same with mems_allowed here, except + * if this task is being moved automatically due to hotplug. + * In that case @mems_allowed has been updated and is empty, + * so @old_mems_allowed is the right nodesets that we migrate + * mm from. + */ + if (is_memory_migrate(cs)) { + cpuset_migrate_mm(mm, &mems_oldcs->old_mems_allowed, &cpuset_attach_nodemask_to); + } mmput(mm); } @@ -2152,10 +2161,12 @@ retry: /* * If sane_behavior flag is set, we need to update tasks' cpumask - * for empty cpuset to take on ancestor's cpumask. + * for empty cpuset to take on ancestor's cpumask. Otherwise, don't + * call update_tasks_cpumask() if the cpuset becomes empty, as + * the tasks in it will be migrated to an ancestor. */ if ((sane && cpumask_empty(cs->cpus_allowed)) || - !cpumask_empty(&off_cpus)) + (!cpumask_empty(&off_cpus) && !cpumask_empty(cs->cpus_allowed))) update_tasks_cpumask(cs, NULL); mutex_lock(&callback_mutex); @@ -2164,10 +2175,12 @@ retry: /* * If sane_behavior flag is set, we need to update tasks' nodemask - * for empty cpuset to take on ancestor's nodemask. + * for empty cpuset to take on ancestor's nodemask. Otherwise, don't + * call update_tasks_nodemask() if the cpuset becomes empty, as + * the tasks in it will be migratd to an ancestor. */ if ((sane && nodes_empty(cs->mems_allowed)) || - !nodes_empty(off_mems)) + (!nodes_empty(off_mems) && !nodes_empty(cs->mems_allowed))) update_tasks_nodemask(cs, NULL); is_empty = cpumask_empty(cs->cpus_allowed) || -- cgit v1.2.3 From 3fc3db9a3ae0ce108badf31a4a00e41b4236f5fc Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 12 Jun 2013 21:04:48 -0700 Subject: cgroup: remove now unused css_depth() Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 12 ------------ 1 file changed, 12 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index bc53d5014b28..d4a329f5874c 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -5227,18 +5227,6 @@ unsigned short css_id(struct cgroup_subsys_state *css) } EXPORT_SYMBOL_GPL(css_id); -unsigned short css_depth(struct cgroup_subsys_state *css) -{ - struct css_id *cssid; - - cssid = rcu_dereference_check(css->id, css_refcnt(css)); - - if (cssid) - return cssid->depth; - return 0; -} -EXPORT_SYMBOL_GPL(css_depth); - /** * css_is_ancestor - test "root" css is an ancestor of "child" * @child: the css to be tested. -- cgit v1.2.3 From 5abb8855734fd7b3fa7f91c13916d0e35d99763c Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 12 Jun 2013 21:04:49 -0700 Subject: cgroup: consistently use @cset for struct css_set variables cgroup.c uses @cg for most struct css_set variables, which in itself could be a bit confusing, but made much worse by the fact that there are places which use @cg for struct cgroup variables. compare_css_sets() epitomizes this confusion - @[old_]cg are struct css_set while @cg[12] are struct cgroup. It's not like the whole deal with cgroup, css_set and cg_cgroup_link isn't already confusing enough. Let's give it some sanity by uniformly using @cset for all struct css_set variables. * s/cg/cset/ for all css_set variables. * s/oldcg/old_cset/ s/oldcgrp/old_cgrp/. The same for the ones prefixed with "new". * s/cg/cgrp/ for cgroup variables in compare_css_sets(). * s/css/cset/ for the cgroup variable in task_cgroup_from_root(). * Whiteline adjustments. This patch is purely cosmetic. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 216 ++++++++++++++++++++++++++++---------------------------- 1 file changed, 109 insertions(+), 107 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index d4a329f5874c..1f5a4e101ed1 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -376,30 +376,32 @@ static unsigned long css_set_hash(struct cgroup_subsys_state *css[]) * compiled into their kernel but not actually in use */ static int use_task_css_set_links __read_mostly; -static void __put_css_set(struct css_set *cg, int taskexit) +static void __put_css_set(struct css_set *cset, int taskexit) { struct cg_cgroup_link *link; struct cg_cgroup_link *saved_link; + /* * Ensure that the refcount doesn't hit zero while any readers * can see it. Similar to atomic_dec_and_lock(), but for an * rwlock */ - if (atomic_add_unless(&cg->refcount, -1, 1)) + if (atomic_add_unless(&cset->refcount, -1, 1)) return; write_lock(&css_set_lock); - if (!atomic_dec_and_test(&cg->refcount)) { + if (!atomic_dec_and_test(&cset->refcount)) { write_unlock(&css_set_lock); return; } /* This css_set is dead. unlink it and release cgroup refcounts */ - hash_del(&cg->hlist); + hash_del(&cset->hlist); css_set_count--; - list_for_each_entry_safe(link, saved_link, &cg->cg_links, + list_for_each_entry_safe(link, saved_link, &cset->cg_links, cg_link_list) { struct cgroup *cgrp = link->cgrp; + list_del(&link->cg_link_list); list_del(&link->cgrp_link_list); @@ -421,45 +423,45 @@ static void __put_css_set(struct css_set *cg, int taskexit) } write_unlock(&css_set_lock); - kfree_rcu(cg, rcu_head); + kfree_rcu(cset, rcu_head); } /* * refcounted get/put for css_set objects */ -static inline void get_css_set(struct css_set *cg) +static inline void get_css_set(struct css_set *cset) { - atomic_inc(&cg->refcount); + atomic_inc(&cset->refcount); } -static inline void put_css_set(struct css_set *cg) +static inline void put_css_set(struct css_set *cset) { - __put_css_set(cg, 0); + __put_css_set(cset, 0); } -static inline void put_css_set_taskexit(struct css_set *cg) +static inline void put_css_set_taskexit(struct css_set *cset) { - __put_css_set(cg, 1); + __put_css_set(cset, 1); } /* * compare_css_sets - helper function for find_existing_css_set(). - * @cg: candidate css_set being tested - * @old_cg: existing css_set for a task + * @cset: candidate css_set being tested + * @old_cset: existing css_set for a task * @new_cgrp: cgroup that's being entered by the task * @template: desired set of css pointers in css_set (pre-calculated) * * Returns true if "cg" matches "old_cg" except for the hierarchy * which "new_cgrp" belongs to, for which it should match "new_cgrp". */ -static bool compare_css_sets(struct css_set *cg, - struct css_set *old_cg, +static bool compare_css_sets(struct css_set *cset, + struct css_set *old_cset, struct cgroup *new_cgrp, struct cgroup_subsys_state *template[]) { struct list_head *l1, *l2; - if (memcmp(template, cg->subsys, sizeof(cg->subsys))) { + if (memcmp(template, cset->subsys, sizeof(cset->subsys))) { /* Not all subsystems matched */ return false; } @@ -473,28 +475,28 @@ static bool compare_css_sets(struct css_set *cg, * candidates. */ - l1 = &cg->cg_links; - l2 = &old_cg->cg_links; + l1 = &cset->cg_links; + l2 = &old_cset->cg_links; while (1) { struct cg_cgroup_link *cgl1, *cgl2; - struct cgroup *cg1, *cg2; + struct cgroup *cgrp1, *cgrp2; l1 = l1->next; l2 = l2->next; /* See if we reached the end - both lists are equal length. */ - if (l1 == &cg->cg_links) { - BUG_ON(l2 != &old_cg->cg_links); + if (l1 == &cset->cg_links) { + BUG_ON(l2 != &old_cset->cg_links); break; } else { - BUG_ON(l2 == &old_cg->cg_links); + BUG_ON(l2 == &old_cset->cg_links); } /* Locate the cgroups associated with these links. */ cgl1 = list_entry(l1, struct cg_cgroup_link, cg_link_list); cgl2 = list_entry(l2, struct cg_cgroup_link, cg_link_list); - cg1 = cgl1->cgrp; - cg2 = cgl2->cgrp; + cgrp1 = cgl1->cgrp; + cgrp2 = cgl2->cgrp; /* Hierarchies should be linked in the same order. */ - BUG_ON(cg1->root != cg2->root); + BUG_ON(cgrp1->root != cgrp2->root); /* * If this hierarchy is the hierarchy of the cgroup @@ -503,11 +505,11 @@ static bool compare_css_sets(struct css_set *cg, * hierarchy, then this css_set should point to the * same cgroup as the old css_set. */ - if (cg1->root == new_cgrp->root) { - if (cg1 != new_cgrp) + if (cgrp1->root == new_cgrp->root) { + if (cgrp1 != new_cgrp) return false; } else { - if (cg1 != cg2) + if (cgrp1 != cgrp2) return false; } } @@ -527,14 +529,13 @@ static bool compare_css_sets(struct css_set *cg, * template: location in which to build the desired set of subsystem * state objects for the new cgroup group */ -static struct css_set *find_existing_css_set( - struct css_set *oldcg, - struct cgroup *cgrp, - struct cgroup_subsys_state *template[]) +static struct css_set *find_existing_css_set(struct css_set *old_cset, + struct cgroup *cgrp, + struct cgroup_subsys_state *template[]) { int i; struct cgroupfs_root *root = cgrp->root; - struct css_set *cg; + struct css_set *cset; unsigned long key; /* @@ -551,17 +552,17 @@ static struct css_set *find_existing_css_set( } else { /* Subsystem is not in this hierarchy, so we * don't want to change the subsystem state */ - template[i] = oldcg->subsys[i]; + template[i] = old_cset->subsys[i]; } } key = css_set_hash(template); - hash_for_each_possible(css_set_table, cg, hlist, key) { - if (!compare_css_sets(cg, oldcg, cgrp, template)) + hash_for_each_possible(css_set_table, cset, hlist, key) { + if (!compare_css_sets(cset, old_cset, cgrp, template)) continue; /* This css_set matches what we need */ - return cg; + return cset; } /* No existing cgroup group matched */ @@ -603,18 +604,18 @@ static int allocate_cg_links(int count, struct list_head *tmp) /** * link_css_set - a helper function to link a css_set to a cgroup * @tmp_cg_links: cg_cgroup_link objects allocated by allocate_cg_links() - * @cg: the css_set to be linked + * @cset: the css_set to be linked * @cgrp: the destination cgroup */ static void link_css_set(struct list_head *tmp_cg_links, - struct css_set *cg, struct cgroup *cgrp) + struct css_set *cset, struct cgroup *cgrp) { struct cg_cgroup_link *link; BUG_ON(list_empty(tmp_cg_links)); link = list_first_entry(tmp_cg_links, struct cg_cgroup_link, cgrp_link_list); - link->cg = cg; + link->cg = cset; link->cgrp = cgrp; atomic_inc(&cgrp->count); list_move(&link->cgrp_link_list, &cgrp->css_sets); @@ -622,7 +623,7 @@ static void link_css_set(struct list_head *tmp_cg_links, * Always add links to the tail of the list so that the list * is sorted by order of hierarchy creation */ - list_add_tail(&link->cg_link_list, &cg->cg_links); + list_add_tail(&link->cg_link_list, &cset->cg_links); } /* @@ -632,10 +633,10 @@ static void link_css_set(struct list_head *tmp_cg_links, * substituted into the appropriate hierarchy. Must be called with * cgroup_mutex held */ -static struct css_set *find_css_set( - struct css_set *oldcg, struct cgroup *cgrp) +static struct css_set *find_css_set(struct css_set *old_cset, + struct cgroup *cgrp) { - struct css_set *res; + struct css_set *cset; struct cgroup_subsys_state *template[CGROUP_SUBSYS_COUNT]; struct list_head tmp_cg_links; @@ -646,40 +647,40 @@ static struct css_set *find_css_set( /* First see if we already have a cgroup group that matches * the desired set */ read_lock(&css_set_lock); - res = find_existing_css_set(oldcg, cgrp, template); - if (res) - get_css_set(res); + cset = find_existing_css_set(old_cset, cgrp, template); + if (cset) + get_css_set(cset); read_unlock(&css_set_lock); - if (res) - return res; + if (cset) + return cset; - res = kmalloc(sizeof(*res), GFP_KERNEL); - if (!res) + cset = kmalloc(sizeof(*cset), GFP_KERNEL); + if (!cset) return NULL; /* Allocate all the cg_cgroup_link objects that we'll need */ if (allocate_cg_links(root_count, &tmp_cg_links) < 0) { - kfree(res); + kfree(cset); return NULL; } - atomic_set(&res->refcount, 1); - INIT_LIST_HEAD(&res->cg_links); - INIT_LIST_HEAD(&res->tasks); - INIT_HLIST_NODE(&res->hlist); + atomic_set(&cset->refcount, 1); + INIT_LIST_HEAD(&cset->cg_links); + INIT_LIST_HEAD(&cset->tasks); + INIT_HLIST_NODE(&cset->hlist); /* Copy the set of subsystem state objects generated in * find_existing_css_set() */ - memcpy(res->subsys, template, sizeof(res->subsys)); + memcpy(cset->subsys, template, sizeof(cset->subsys)); write_lock(&css_set_lock); /* Add reference counts and links from the new css_set. */ - list_for_each_entry(link, &oldcg->cg_links, cg_link_list) { + list_for_each_entry(link, &old_cset->cg_links, cg_link_list) { struct cgroup *c = link->cgrp; if (c->root == cgrp->root) c = cgrp; - link_css_set(&tmp_cg_links, res, c); + link_css_set(&tmp_cg_links, cset, c); } BUG_ON(!list_empty(&tmp_cg_links)); @@ -687,12 +688,12 @@ static struct css_set *find_css_set( css_set_count++; /* Add this cgroup group to the hash table */ - key = css_set_hash(res->subsys); - hash_add(css_set_table, &res->hlist, key); + key = css_set_hash(cset->subsys); + hash_add(css_set_table, &cset->hlist, key); write_unlock(&css_set_lock); - return res; + return cset; } /* @@ -702,7 +703,7 @@ static struct css_set *find_css_set( static struct cgroup *task_cgroup_from_root(struct task_struct *task, struct cgroupfs_root *root) { - struct css_set *css; + struct css_set *cset; struct cgroup *res = NULL; BUG_ON(!mutex_is_locked(&cgroup_mutex)); @@ -712,12 +713,12 @@ static struct cgroup *task_cgroup_from_root(struct task_struct *task, * task can't change groups, so the only thing that can happen * is that it exits and its css is set back to init_css_set. */ - css = task->cgroups; - if (css == &init_css_set) { + cset = task->cgroups; + if (cset == &init_css_set) { res = &root->top_cgroup; } else { struct cg_cgroup_link *link; - list_for_each_entry(link, &css->cg_links, cg_link_list) { + list_for_each_entry(link, &cset->cg_links, cg_link_list) { struct cgroup *c = link->cgrp; if (c->root == root) { res = c; @@ -1608,7 +1609,7 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, struct cgroupfs_root *existing_root; const struct cred *cred; int i; - struct css_set *cg; + struct css_set *cset; BUG_ON(sb->s_root != NULL); @@ -1666,8 +1667,8 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, /* Link the top cgroup in this hierarchy into all * the css_set objects */ write_lock(&css_set_lock); - hash_for_each(css_set_table, i, cg, hlist) - link_css_set(&tmp_cg_links, cg, root_cgrp); + hash_for_each(css_set_table, i, cset, hlist) + link_css_set(&tmp_cg_links, cset, root_cgrp); write_unlock(&css_set_lock); free_cg_links(&tmp_cg_links); @@ -1944,10 +1945,11 @@ EXPORT_SYMBOL_GPL(cgroup_taskset_size); * * Must be called with cgroup_mutex and threadgroup locked. */ -static void cgroup_task_migrate(struct cgroup *oldcgrp, - struct task_struct *tsk, struct css_set *newcg) +static void cgroup_task_migrate(struct cgroup *old_cgrp, + struct task_struct *tsk, + struct css_set *new_cset) { - struct css_set *oldcg; + struct css_set *old_cset; /* * We are synchronized through threadgroup_lock() against PF_EXITING @@ -1955,25 +1957,25 @@ static void cgroup_task_migrate(struct cgroup *oldcgrp, * css_set to init_css_set and dropping the old one. */ WARN_ON_ONCE(tsk->flags & PF_EXITING); - oldcg = tsk->cgroups; + old_cset = tsk->cgroups; task_lock(tsk); - rcu_assign_pointer(tsk->cgroups, newcg); + rcu_assign_pointer(tsk->cgroups, new_cset); task_unlock(tsk); /* Update the css_set linked lists if we're using them */ write_lock(&css_set_lock); if (!list_empty(&tsk->cg_list)) - list_move(&tsk->cg_list, &newcg->tasks); + list_move(&tsk->cg_list, &new_cset->tasks); write_unlock(&css_set_lock); /* - * We just gained a reference on oldcg by taking it from the task. As - * trading it for newcg is protected by cgroup_mutex, we're safe to drop - * it here; it will be freed under RCU. + * We just gained a reference on old_cset by taking it from the + * task. As trading it for new_cset is protected by cgroup_mutex, + * we're safe to drop it here; it will be freed under RCU. */ - set_bit(CGRP_RELEASABLE, &oldcgrp->flags); - put_css_set(oldcg); + set_bit(CGRP_RELEASABLE, &old_cgrp->flags); + put_css_set(old_cset); } /** @@ -2925,7 +2927,7 @@ static void cgroup_advance_iter(struct cgroup *cgrp, { struct list_head *l = it->cg_link; struct cg_cgroup_link *link; - struct css_set *cg; + struct css_set *cset; /* Advance to the next non-empty css_set */ do { @@ -2935,10 +2937,10 @@ static void cgroup_advance_iter(struct cgroup *cgrp, return; } link = list_entry(l, struct cg_cgroup_link, cgrp_link_list); - cg = link->cg; - } while (list_empty(&cg->tasks)); + cset = link->cg; + } while (list_empty(&cset->tasks)); it->cg_link = l; - it->task = cg->tasks.next; + it->task = cset->tasks.next; } /* @@ -4516,7 +4518,7 @@ int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss) struct cgroup_subsys_state *css; int i, ret; struct hlist_node *tmp; - struct css_set *cg; + struct css_set *cset; unsigned long key; /* check name and function validity */ @@ -4583,17 +4585,17 @@ int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss) * this is all done under the css_set_lock. */ write_lock(&css_set_lock); - hash_for_each_safe(css_set_table, i, tmp, cg, hlist) { + hash_for_each_safe(css_set_table, i, tmp, cset, hlist) { /* skip entries that we already rehashed */ - if (cg->subsys[ss->subsys_id]) + if (cset->subsys[ss->subsys_id]) continue; /* remove existing entry */ - hash_del(&cg->hlist); + hash_del(&cset->hlist); /* set new value */ - cg->subsys[ss->subsys_id] = css; + cset->subsys[ss->subsys_id] = css; /* recompute hash and restore entry */ - key = css_set_hash(cg->subsys); - hash_add(css_set_table, &cg->hlist, key); + key = css_set_hash(cset->subsys); + hash_add(css_set_table, &cset->hlist, key); } write_unlock(&css_set_lock); @@ -4653,13 +4655,13 @@ void cgroup_unload_subsys(struct cgroup_subsys *ss) */ write_lock(&css_set_lock); list_for_each_entry(link, &dummytop->css_sets, cgrp_link_list) { - struct css_set *cg = link->cg; + struct css_set *cset = link->cg; unsigned long key; - hash_del(&cg->hlist); - cg->subsys[ss->subsys_id] = NULL; - key = css_set_hash(cg->subsys); - hash_add(css_set_table, &cg->hlist, key); + hash_del(&cset->hlist); + cset->subsys[ss->subsys_id] = NULL; + key = css_set_hash(cset->subsys); + hash_add(css_set_table, &cset->hlist, key); } write_unlock(&css_set_lock); @@ -5006,7 +5008,7 @@ void cgroup_post_fork(struct task_struct *child) */ void cgroup_exit(struct task_struct *tsk, int run_callbacks) { - struct css_set *cg; + struct css_set *cset; int i; /* @@ -5023,7 +5025,7 @@ void cgroup_exit(struct task_struct *tsk, int run_callbacks) /* Reassign the task to the init_css_set. */ task_lock(tsk); - cg = tsk->cgroups; + cset = tsk->cgroups; tsk->cgroups = &init_css_set; if (run_callbacks && need_forkexit_callback) { @@ -5036,7 +5038,7 @@ void cgroup_exit(struct task_struct *tsk, int run_callbacks) if (ss->exit) { struct cgroup *old_cgrp = - rcu_dereference_raw(cg->subsys[i])->cgroup; + rcu_dereference_raw(cset->subsys[i])->cgroup; struct cgroup *cgrp = task_cgroup(tsk, i); ss->exit(cgrp, old_cgrp, tsk); } @@ -5044,7 +5046,7 @@ void cgroup_exit(struct task_struct *tsk, int run_callbacks) } task_unlock(tsk); - put_css_set_taskexit(cg); + put_css_set_taskexit(cset); } static void check_for_release(struct cgroup *cgrp) @@ -5453,12 +5455,12 @@ static int current_css_set_cg_links_read(struct cgroup *cont, struct seq_file *seq) { struct cg_cgroup_link *link; - struct css_set *cg; + struct css_set *cset; read_lock(&css_set_lock); rcu_read_lock(); - cg = rcu_dereference(current->cgroups); - list_for_each_entry(link, &cg->cg_links, cg_link_list) { + cset = rcu_dereference(current->cgroups); + list_for_each_entry(link, &cset->cg_links, cg_link_list) { struct cgroup *c = link->cgrp; const char *name; @@ -5483,11 +5485,11 @@ static int cgroup_css_links_read(struct cgroup *cont, read_lock(&css_set_lock); list_for_each_entry(link, &cont->css_sets, cgrp_link_list) { - struct css_set *cg = link->cg; + struct css_set *cset = link->cg; struct task_struct *task; int count = 0; - seq_printf(seq, "css_set %p\n", cg); - list_for_each_entry(task, &cg->tasks, cg_list) { + seq_printf(seq, "css_set %p\n", cset); + list_for_each_entry(task, &cset->tasks, cg_list) { if (count++ > MAX_TASKS_SHOWN_PER_CSS) { seq_puts(seq, " ...\n"); break; -- cgit v1.2.3 From 69d0206c793a17431eacee2694ee7a4b25df76b7 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 12 Jun 2013 21:04:50 -0700 Subject: cgroup: bring some sanity to naming around cg_cgroup_link cgroups and css_sets are mapped M:N and this M:N mapping is represented by struct cg_cgroup_link which forms linked lists on both sides. The naming around this mapping is already confusing and struct cg_cgroup_link exacerbates the situation quite a bit. >From cgroup side, it starts off ->css_sets and runs through ->cgrp_link_list. From css_set side, it starts off ->cg_links and runs through ->cg_link_list. This is rather reversed as cgrp_link_list is used to iterate css_sets and cg_link_list cgroups. Also, this is the only place which is still using the confusing "cg" for css_sets. This patch cleans it up a bit. * s/cgroup->css_sets/cgroup->cset_links/ s/css_set->cg_links/css_set->cgrp_links/ s/cgroup_iter->cg_link/cgroup_iter->cset_link/ * s/cg_cgroup_link/cgrp_cset_link/ * s/cgrp_cset_link->cg/cgrp_cset_link->cset/ s/cgrp_cset_link->cgrp_link_list/cgrp_cset_link->cset_link/ s/cgrp_cset_link->cg_link_list/cgrp_cset_link->cgrp_link/ * s/init_css_set_link/init_cgrp_cset_link/ s/free_cg_links/free_cgrp_cset_links/ s/allocate_cg_links/allocate_cgrp_cset_links/ * s/cgl[12]/link[12]/ in compare_css_sets() * s/saved_link/tmp_link/ s/tmp/tmp_links/ and a couple similar adustments. * Comment and whiteline adjustments. After the changes, we have list_for_each_entry(link, &cont->cset_links, cset_link) { struct css_set *cset = link->cset; instead of list_for_each_entry(link, &cont->css_sets, cgrp_link_list) { struct css_set *cset = link->cg; This patch is purely cosmetic. v2: Fix broken sentences in the patch description. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 226 ++++++++++++++++++++++++++++---------------------------- 1 file changed, 113 insertions(+), 113 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 1f5a4e101ed1..ef97bd0cd546 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -315,20 +315,24 @@ static void cgroup_release_agent(struct work_struct *work); static DECLARE_WORK(release_agent_work, cgroup_release_agent); static void check_for_release(struct cgroup *cgrp); -/* Link structure for associating css_set objects with cgroups */ -struct cg_cgroup_link { - /* - * List running through cg_cgroup_links associated with a - * cgroup, anchored on cgroup->css_sets - */ - struct list_head cgrp_link_list; - struct cgroup *cgrp; - /* - * List running through cg_cgroup_links pointing at a - * single css_set object, anchored on css_set->cg_links - */ - struct list_head cg_link_list; - struct css_set *cg; +/* + * A cgroup can be associated with multiple css_sets as different tasks may + * belong to different cgroups on different hierarchies. In the other + * direction, a css_set is naturally associated with multiple cgroups. + * This M:N relationship is represented by the following link structure + * which exists for each association and allows traversing the associations + * from both sides. + */ +struct cgrp_cset_link { + /* the cgroup and css_set this link associates */ + struct cgroup *cgrp; + struct css_set *cset; + + /* list of cgrp_cset_links anchored at cgrp->cset_links */ + struct list_head cset_link; + + /* list of cgrp_cset_links anchored at css_set->cgrp_links */ + struct list_head cgrp_link; }; /* The default css_set - used by init and its children prior to any @@ -339,7 +343,7 @@ struct cg_cgroup_link { */ static struct css_set init_css_set; -static struct cg_cgroup_link init_css_set_link; +static struct cgrp_cset_link init_cgrp_cset_link; static int cgroup_init_idr(struct cgroup_subsys *ss, struct cgroup_subsys_state *css); @@ -378,8 +382,7 @@ static int use_task_css_set_links __read_mostly; static void __put_css_set(struct css_set *cset, int taskexit) { - struct cg_cgroup_link *link; - struct cg_cgroup_link *saved_link; + struct cgrp_cset_link *link, *tmp_link; /* * Ensure that the refcount doesn't hit zero while any readers @@ -398,12 +401,11 @@ static void __put_css_set(struct css_set *cset, int taskexit) hash_del(&cset->hlist); css_set_count--; - list_for_each_entry_safe(link, saved_link, &cset->cg_links, - cg_link_list) { + list_for_each_entry_safe(link, tmp_link, &cset->cgrp_links, cgrp_link) { struct cgroup *cgrp = link->cgrp; - list_del(&link->cg_link_list); - list_del(&link->cgrp_link_list); + list_del(&link->cset_link); + list_del(&link->cgrp_link); /* * We may not be holding cgroup_mutex, and if cgrp->count is @@ -475,26 +477,26 @@ static bool compare_css_sets(struct css_set *cset, * candidates. */ - l1 = &cset->cg_links; - l2 = &old_cset->cg_links; + l1 = &cset->cgrp_links; + l2 = &old_cset->cgrp_links; while (1) { - struct cg_cgroup_link *cgl1, *cgl2; + struct cgrp_cset_link *link1, *link2; struct cgroup *cgrp1, *cgrp2; l1 = l1->next; l2 = l2->next; /* See if we reached the end - both lists are equal length. */ - if (l1 == &cset->cg_links) { - BUG_ON(l2 != &old_cset->cg_links); + if (l1 == &cset->cgrp_links) { + BUG_ON(l2 != &old_cset->cgrp_links); break; } else { - BUG_ON(l2 == &old_cset->cg_links); + BUG_ON(l2 == &old_cset->cgrp_links); } /* Locate the cgroups associated with these links. */ - cgl1 = list_entry(l1, struct cg_cgroup_link, cg_link_list); - cgl2 = list_entry(l2, struct cg_cgroup_link, cg_link_list); - cgrp1 = cgl1->cgrp; - cgrp2 = cgl2->cgrp; + link1 = list_entry(l1, struct cgrp_cset_link, cgrp_link); + link2 = list_entry(l2, struct cgrp_cset_link, cgrp_link); + cgrp1 = link1->cgrp; + cgrp2 = link2->cgrp; /* Hierarchies should be linked in the same order. */ BUG_ON(cgrp1->root != cgrp2->root); @@ -569,61 +571,64 @@ static struct css_set *find_existing_css_set(struct css_set *old_cset, return NULL; } -static void free_cg_links(struct list_head *tmp) +static void free_cgrp_cset_links(struct list_head *links_to_free) { - struct cg_cgroup_link *link; - struct cg_cgroup_link *saved_link; + struct cgrp_cset_link *link, *tmp_link; - list_for_each_entry_safe(link, saved_link, tmp, cgrp_link_list) { - list_del(&link->cgrp_link_list); + list_for_each_entry_safe(link, tmp_link, links_to_free, cset_link) { + list_del(&link->cset_link); kfree(link); } } -/* - * allocate_cg_links() allocates "count" cg_cgroup_link structures - * and chains them on tmp through their cgrp_link_list fields. Returns 0 on - * success or a negative error +/** + * allocate_cgrp_cset_links - allocate cgrp_cset_links + * @count: the number of links to allocate + * @tmp_links: list_head the allocated links are put on + * + * Allocate @count cgrp_cset_link structures and chain them on @tmp_links + * through ->cset_link. Returns 0 on success or -errno. */ -static int allocate_cg_links(int count, struct list_head *tmp) +static int allocate_cgrp_cset_links(int count, struct list_head *tmp_links) { - struct cg_cgroup_link *link; + struct cgrp_cset_link *link; int i; - INIT_LIST_HEAD(tmp); + + INIT_LIST_HEAD(tmp_links); + for (i = 0; i < count; i++) { link = kmalloc(sizeof(*link), GFP_KERNEL); if (!link) { - free_cg_links(tmp); + free_cgrp_cset_links(tmp_links); return -ENOMEM; } - list_add(&link->cgrp_link_list, tmp); + list_add(&link->cset_link, tmp_links); } return 0; } /** * link_css_set - a helper function to link a css_set to a cgroup - * @tmp_cg_links: cg_cgroup_link objects allocated by allocate_cg_links() + * @tmp_links: cgrp_cset_link objects allocated by allocate_cgrp_cset_links() * @cset: the css_set to be linked * @cgrp: the destination cgroup */ -static void link_css_set(struct list_head *tmp_cg_links, - struct css_set *cset, struct cgroup *cgrp) +static void link_css_set(struct list_head *tmp_links, struct css_set *cset, + struct cgroup *cgrp) { - struct cg_cgroup_link *link; + struct cgrp_cset_link *link; - BUG_ON(list_empty(tmp_cg_links)); - link = list_first_entry(tmp_cg_links, struct cg_cgroup_link, - cgrp_link_list); - link->cg = cset; + BUG_ON(list_empty(tmp_links)); + link = list_first_entry(tmp_links, struct cgrp_cset_link, cset_link); + link->cset = cset; link->cgrp = cgrp; atomic_inc(&cgrp->count); - list_move(&link->cgrp_link_list, &cgrp->css_sets); + list_move(&link->cset_link, &cgrp->cset_links); /* * Always add links to the tail of the list so that the list * is sorted by order of hierarchy creation */ - list_add_tail(&link->cg_link_list, &cset->cg_links); + list_add_tail(&link->cgrp_link, &cset->cgrp_links); } /* @@ -638,10 +643,8 @@ static struct css_set *find_css_set(struct css_set *old_cset, { struct css_set *cset; struct cgroup_subsys_state *template[CGROUP_SUBSYS_COUNT]; - - struct list_head tmp_cg_links; - - struct cg_cgroup_link *link; + struct list_head tmp_links; + struct cgrp_cset_link *link; unsigned long key; /* First see if we already have a cgroup group that matches @@ -659,14 +662,14 @@ static struct css_set *find_css_set(struct css_set *old_cset, if (!cset) return NULL; - /* Allocate all the cg_cgroup_link objects that we'll need */ - if (allocate_cg_links(root_count, &tmp_cg_links) < 0) { + /* Allocate all the cgrp_cset_link objects that we'll need */ + if (allocate_cgrp_cset_links(root_count, &tmp_links) < 0) { kfree(cset); return NULL; } atomic_set(&cset->refcount, 1); - INIT_LIST_HEAD(&cset->cg_links); + INIT_LIST_HEAD(&cset->cgrp_links); INIT_LIST_HEAD(&cset->tasks); INIT_HLIST_NODE(&cset->hlist); @@ -676,14 +679,15 @@ static struct css_set *find_css_set(struct css_set *old_cset, write_lock(&css_set_lock); /* Add reference counts and links from the new css_set. */ - list_for_each_entry(link, &old_cset->cg_links, cg_link_list) { + list_for_each_entry(link, &old_cset->cgrp_links, cgrp_link) { struct cgroup *c = link->cgrp; + if (c->root == cgrp->root) c = cgrp; - link_css_set(&tmp_cg_links, cset, c); + link_css_set(&tmp_links, cset, c); } - BUG_ON(!list_empty(&tmp_cg_links)); + BUG_ON(!list_empty(&tmp_links)); css_set_count++; @@ -717,9 +721,11 @@ static struct cgroup *task_cgroup_from_root(struct task_struct *task, if (cset == &init_css_set) { res = &root->top_cgroup; } else { - struct cg_cgroup_link *link; - list_for_each_entry(link, &cset->cg_links, cg_link_list) { + struct cgrp_cset_link *link; + + list_for_each_entry(link, &cset->cgrp_links, cgrp_link) { struct cgroup *c = link->cgrp; + if (c->root == root) { res = c; break; @@ -1405,7 +1411,7 @@ static void init_cgroup_housekeeping(struct cgroup *cgrp) INIT_LIST_HEAD(&cgrp->sibling); INIT_LIST_HEAD(&cgrp->children); INIT_LIST_HEAD(&cgrp->files); - INIT_LIST_HEAD(&cgrp->css_sets); + INIT_LIST_HEAD(&cgrp->cset_links); INIT_LIST_HEAD(&cgrp->allcg_node); INIT_LIST_HEAD(&cgrp->release_list); INIT_LIST_HEAD(&cgrp->pidlists); @@ -1604,7 +1610,7 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, BUG_ON(!root); if (root == opts.new_root) { /* We used the new root structure, so this is a new hierarchy */ - struct list_head tmp_cg_links; + struct list_head tmp_links; struct cgroup *root_cgrp = &root->top_cgroup; struct cgroupfs_root *existing_root; const struct cred *cred; @@ -1636,7 +1642,7 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, * that's us. The worst that can happen is that we * have some link structures left over */ - ret = allocate_cg_links(css_set_count, &tmp_cg_links); + ret = allocate_cgrp_cset_links(css_set_count, &tmp_links); if (ret) goto unlock_drop; @@ -1646,7 +1652,7 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, ret = rebind_subsystems(root, root->subsys_mask); if (ret == -EBUSY) { - free_cg_links(&tmp_cg_links); + free_cgrp_cset_links(&tmp_links); goto unlock_drop; } /* @@ -1668,10 +1674,10 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, * the css_set objects */ write_lock(&css_set_lock); hash_for_each(css_set_table, i, cset, hlist) - link_css_set(&tmp_cg_links, cset, root_cgrp); + link_css_set(&tmp_links, cset, root_cgrp); write_unlock(&css_set_lock); - free_cg_links(&tmp_cg_links); + free_cgrp_cset_links(&tmp_links); BUG_ON(!list_empty(&root_cgrp->children)); BUG_ON(root->number_of_cgroups != 1); @@ -1722,9 +1728,8 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, static void cgroup_kill_sb(struct super_block *sb) { struct cgroupfs_root *root = sb->s_fs_info; struct cgroup *cgrp = &root->top_cgroup; + struct cgrp_cset_link *link, *tmp_link; int ret; - struct cg_cgroup_link *link; - struct cg_cgroup_link *saved_link; BUG_ON(!root); @@ -1740,15 +1745,14 @@ static void cgroup_kill_sb(struct super_block *sb) { BUG_ON(ret); /* - * Release all the links from css_sets to this hierarchy's + * Release all the links from cset_links to this hierarchy's * root cgroup */ write_lock(&css_set_lock); - list_for_each_entry_safe(link, saved_link, &cgrp->css_sets, - cgrp_link_list) { - list_del(&link->cg_link_list); - list_del(&link->cgrp_link_list); + list_for_each_entry_safe(link, tmp_link, &cgrp->cset_links, cset_link) { + list_del(&link->cset_link); + list_del(&link->cgrp_link); kfree(link); } write_unlock(&css_set_lock); @@ -2908,12 +2912,11 @@ int cgroup_rm_cftypes(struct cgroup_subsys *ss, struct cftype *cfts) int cgroup_task_count(const struct cgroup *cgrp) { int count = 0; - struct cg_cgroup_link *link; + struct cgrp_cset_link *link; read_lock(&css_set_lock); - list_for_each_entry(link, &cgrp->css_sets, cgrp_link_list) { - count += atomic_read(&link->cg->refcount); - } + list_for_each_entry(link, &cgrp->cset_links, cset_link) + count += atomic_read(&link->cset->refcount); read_unlock(&css_set_lock); return count; } @@ -2922,24 +2925,23 @@ int cgroup_task_count(const struct cgroup *cgrp) * Advance a list_head iterator. The iterator should be positioned at * the start of a css_set */ -static void cgroup_advance_iter(struct cgroup *cgrp, - struct cgroup_iter *it) +static void cgroup_advance_iter(struct cgroup *cgrp, struct cgroup_iter *it) { - struct list_head *l = it->cg_link; - struct cg_cgroup_link *link; + struct list_head *l = it->cset_link; + struct cgrp_cset_link *link; struct css_set *cset; /* Advance to the next non-empty css_set */ do { l = l->next; - if (l == &cgrp->css_sets) { - it->cg_link = NULL; + if (l == &cgrp->cset_links) { + it->cset_link = NULL; return; } - link = list_entry(l, struct cg_cgroup_link, cgrp_link_list); - cset = link->cg; + link = list_entry(l, struct cgrp_cset_link, cset_link); + cset = link->cset; } while (list_empty(&cset->tasks)); - it->cg_link = l; + it->cset_link = l; it->task = cset->tasks.next; } @@ -3160,7 +3162,7 @@ void cgroup_iter_start(struct cgroup *cgrp, struct cgroup_iter *it) cgroup_enable_task_cg_lists(); read_lock(&css_set_lock); - it->cg_link = &cgrp->css_sets; + it->cset_link = &cgrp->cset_links; cgroup_advance_iter(cgrp, it); } @@ -3169,16 +3171,16 @@ struct task_struct *cgroup_iter_next(struct cgroup *cgrp, { struct task_struct *res; struct list_head *l = it->task; - struct cg_cgroup_link *link; + struct cgrp_cset_link *link; /* If the iterator cg is NULL, we have no tasks */ - if (!it->cg_link) + if (!it->cset_link) return NULL; res = list_entry(l, struct task_struct, cg_list); /* Advance iterator to find next entry */ l = l->next; - link = list_entry(it->cg_link, struct cg_cgroup_link, cgrp_link_list); - if (l == &link->cg->tasks) { + link = list_entry(it->cset_link, struct cgrp_cset_link, cset_link); + if (l == &link->cset->tasks) { /* We reached the end of this task list - move on to * the next cg_cgroup_link */ cgroup_advance_iter(cgrp, it); @@ -4625,7 +4627,7 @@ EXPORT_SYMBOL_GPL(cgroup_load_subsys); */ void cgroup_unload_subsys(struct cgroup_subsys *ss) { - struct cg_cgroup_link *link; + struct cgrp_cset_link *link; BUG_ON(ss->module == NULL); @@ -4654,8 +4656,8 @@ void cgroup_unload_subsys(struct cgroup_subsys *ss) * in loading, we need to pay our respects to the hashtable gods. */ write_lock(&css_set_lock); - list_for_each_entry(link, &dummytop->css_sets, cgrp_link_list) { - struct css_set *cset = link->cg; + list_for_each_entry(link, &dummytop->cset_links, cset_link) { + struct css_set *cset = link->cset; unsigned long key; hash_del(&cset->hlist); @@ -4688,7 +4690,7 @@ int __init cgroup_init_early(void) { int i; atomic_set(&init_css_set.refcount, 1); - INIT_LIST_HEAD(&init_css_set.cg_links); + INIT_LIST_HEAD(&init_css_set.cgrp_links); INIT_LIST_HEAD(&init_css_set.tasks); INIT_HLIST_NODE(&init_css_set.hlist); css_set_count = 1; @@ -4696,12 +4698,10 @@ int __init cgroup_init_early(void) root_count = 1; init_task.cgroups = &init_css_set; - init_css_set_link.cg = &init_css_set; - init_css_set_link.cgrp = dummytop; - list_add(&init_css_set_link.cgrp_link_list, - &rootnode.top_cgroup.css_sets); - list_add(&init_css_set_link.cg_link_list, - &init_css_set.cg_links); + init_cgrp_cset_link.cset = &init_css_set; + init_cgrp_cset_link.cgrp = dummytop; + list_add(&init_cgrp_cset_link.cset_link, &rootnode.top_cgroup.cset_links); + list_add(&init_cgrp_cset_link.cgrp_link, &init_css_set.cgrp_links); for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { struct cgroup_subsys *ss = subsys[i]; @@ -5454,13 +5454,13 @@ static int current_css_set_cg_links_read(struct cgroup *cont, struct cftype *cft, struct seq_file *seq) { - struct cg_cgroup_link *link; + struct cgrp_cset_link *link; struct css_set *cset; read_lock(&css_set_lock); rcu_read_lock(); cset = rcu_dereference(current->cgroups); - list_for_each_entry(link, &cset->cg_links, cg_link_list) { + list_for_each_entry(link, &cset->cgrp_links, cgrp_link) { struct cgroup *c = link->cgrp; const char *name; @@ -5481,11 +5481,11 @@ static int cgroup_css_links_read(struct cgroup *cont, struct cftype *cft, struct seq_file *seq) { - struct cg_cgroup_link *link; + struct cgrp_cset_link *link; read_lock(&css_set_lock); - list_for_each_entry(link, &cont->css_sets, cgrp_link_list) { - struct css_set *cset = link->cg; + list_for_each_entry(link, &cont->cset_links, cset_link) { + struct css_set *cset = link->cset; struct task_struct *task; int count = 0; seq_printf(seq, "css_set %p\n", cset); -- cgit v1.2.3 From f4f4be2bd2889c69a8698edef8dbfd4f6759aa87 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 12 Jun 2013 21:04:51 -0700 Subject: cgroup: use kzalloc() instead of kmalloc() There's no point in using kmalloc() instead of the clearing variant for trivial stuff. We can live dangerously elsewhere. Use kzalloc() instead and drop 0 inits. While at it, do trivial code reorganization in cgroup_file_open(). This patch doesn't introduce any functional changes. v2: I was caught in the very distant past where list_del() didn't poison and the initial version converted list_del()s to list_del_init()s too. Li and Kent took me out of the stasis chamber. Signed-off-by: Tejun Heo Cc: Kent Overstreet Acked-by: Li Zefan --- kernel/cgroup.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index ef97bd0cd546..d86a8477d56a 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -597,7 +597,7 @@ static int allocate_cgrp_cset_links(int count, struct list_head *tmp_links) INIT_LIST_HEAD(tmp_links); for (i = 0; i < count; i++) { - link = kmalloc(sizeof(*link), GFP_KERNEL); + link = kzalloc(sizeof(*link), GFP_KERNEL); if (!link) { free_cgrp_cset_links(tmp_links); return -ENOMEM; @@ -658,7 +658,7 @@ static struct css_set *find_css_set(struct css_set *old_cset, if (cset) return cset; - cset = kmalloc(sizeof(*cset), GFP_KERNEL); + cset = kzalloc(sizeof(*cset), GFP_KERNEL); if (!cset) return NULL; @@ -2475,10 +2475,12 @@ static int cgroup_file_open(struct inode *inode, struct file *file) cft = __d_cft(file->f_dentry); if (cft->read_map || cft->read_seq_string) { - struct cgroup_seqfile_state *state = - kzalloc(sizeof(*state), GFP_USER); + struct cgroup_seqfile_state *state; + + state = kzalloc(sizeof(*state), GFP_USER); if (!state) return -ENOMEM; + state->cft = cft; state->cgroup = __d_cgrp(file->f_dentry->d_parent); file->f_op = &cgroup_seqfile_operations; @@ -3511,7 +3513,7 @@ static struct cgroup_pidlist *cgroup_pidlist_find(struct cgroup *cgrp, } } /* entry not found; create a new one */ - l = kmalloc(sizeof(struct cgroup_pidlist), GFP_KERNEL); + l = kzalloc(sizeof(struct cgroup_pidlist), GFP_KERNEL); if (!l) { mutex_unlock(&cgrp->pidlist_mutex); return l; @@ -3520,8 +3522,6 @@ static struct cgroup_pidlist *cgroup_pidlist_find(struct cgroup *cgrp, down_write(&l->mutex); l->key.type = type; l->key.ns = get_pid_ns(ns); - l->use_count = 0; /* don't increment here */ - l->list = NULL; l->owner = cgrp; list_add(&l->links, &cgrp->pidlists); mutex_unlock(&cgrp->pidlist_mutex); -- cgit v1.2.3 From 54766d4a1d3d6f84ff8fa475cd8f165c0a0000eb Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 12 Jun 2013 21:04:53 -0700 Subject: cgroup: rename CGRP_REMOVED to CGRP_DEAD We will add another flag indicating that the cgroup is in the process of being killed. REMOVING / REMOVED is more difficult to distinguish and cgroup_is_removing()/cgroup_is_removed() are a bit awkward. Also, later percpu_ref usage will involve "kill"ing the refcnt. s/CGRP_REMOVED/CGRP_DEAD/ s/cgroup_is_removed()/cgroup_is_dead() This patch is purely cosmetic. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 30 ++++++++++++++---------------- 1 file changed, 14 insertions(+), 16 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index d86a8477d56a..84efb344fdf6 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -226,9 +226,9 @@ static int css_refcnt(struct cgroup_subsys_state *css) } /* convenient tests for these bits */ -static inline bool cgroup_is_removed(const struct cgroup *cgrp) +static inline bool cgroup_is_dead(const struct cgroup *cgrp) { - return test_bit(CGRP_REMOVED, &cgrp->flags); + return test_bit(CGRP_DEAD, &cgrp->flags); } /** @@ -300,7 +300,7 @@ static inline struct cftype *__d_cft(struct dentry *dentry) static bool cgroup_lock_live_group(struct cgroup *cgrp) { mutex_lock(&cgroup_mutex); - if (cgroup_is_removed(cgrp)) { + if (cgroup_is_dead(cgrp)) { mutex_unlock(&cgroup_mutex); return false; } @@ -892,7 +892,7 @@ static void cgroup_diput(struct dentry *dentry, struct inode *inode) if (S_ISDIR(inode->i_mode)) { struct cgroup *cgrp = dentry->d_fsdata; - BUG_ON(!(cgroup_is_removed(cgrp))); + BUG_ON(!(cgroup_is_dead(cgrp))); call_rcu(&cgrp->rcu_head, cgroup_free_rcu); } else { struct cfent *cfe = __d_cfe(dentry); @@ -2363,7 +2363,7 @@ static ssize_t cgroup_file_write(struct file *file, const char __user *buf, struct cftype *cft = __d_cft(file->f_dentry); struct cgroup *cgrp = __d_cgrp(file->f_dentry->d_parent); - if (cgroup_is_removed(cgrp)) + if (cgroup_is_dead(cgrp)) return -ENODEV; if (cft->write) return cft->write(cgrp, cft, file, buf, nbytes, ppos); @@ -2408,7 +2408,7 @@ static ssize_t cgroup_file_read(struct file *file, char __user *buf, struct cftype *cft = __d_cft(file->f_dentry); struct cgroup *cgrp = __d_cgrp(file->f_dentry->d_parent); - if (cgroup_is_removed(cgrp)) + if (cgroup_is_dead(cgrp)) return -ENODEV; if (cft->read) @@ -2831,7 +2831,7 @@ static void cgroup_cfts_commit(struct cgroup_subsys *ss, mutex_lock(&inode->i_mutex); mutex_lock(&cgroup_mutex); - if (!cgroup_is_removed(cgrp)) + if (!cgroup_is_dead(cgrp)) cgroup_addrm_files(cgrp, ss, cfts, is_add); mutex_unlock(&cgroup_mutex); mutex_unlock(&inode->i_mutex); @@ -2999,14 +2999,14 @@ struct cgroup *cgroup_next_sibling(struct cgroup *pos) /* * @pos could already have been removed. Once a cgroup is removed, * its ->sibling.next is no longer updated when its next sibling - * changes. As CGRP_REMOVED is set on removal which is fully + * changes. As CGRP_DEAD is set on removal which is fully * serialized, if we see it unasserted, it's guaranteed that the * next sibling hasn't finished its grace period even if it's * already removed, and thus safe to dereference from this RCU * critical section. If ->sibling.next is inaccessible, - * cgroup_is_removed() is guaranteed to be visible as %true here. + * cgroup_is_dead() is guaranteed to be visible as %true here. */ - if (likely(!cgroup_is_removed(pos))) { + if (likely(!cgroup_is_dead(pos))) { next = list_entry_rcu(pos->sibling.next, struct cgroup, sibling); if (&next->sibling != &pos->parent->children) return next; @@ -4383,7 +4383,7 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) * attempts fail thus maintaining the removal conditions verified * above. * - * Note that CGRP_REMVOED clearing is depended upon by + * Note that CGRP_DEAD assertion is depended upon by * cgroup_next_sibling() to resume iteration after dropping RCU * read lock. See cgroup_next_sibling() for details. */ @@ -4393,7 +4393,7 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) WARN_ON(atomic_read(&css->refcnt) < 0); atomic_add(CSS_DEACT_BIAS, &css->refcnt); } - set_bit(CGRP_REMOVED, &cgrp->flags); + set_bit(CGRP_DEAD, &cgrp->flags); /* tell subsystems to initate destruction */ for_each_subsys(cgrp->root, ss) @@ -5063,7 +5063,7 @@ static void check_for_release(struct cgroup *cgrp) int need_schedule_work = 0; raw_spin_lock(&release_list_lock); - if (!cgroup_is_removed(cgrp) && + if (!cgroup_is_dead(cgrp) && list_empty(&cgrp->release_list)) { list_add(&cgrp->release_list, &release_list); need_schedule_work = 1; @@ -5209,9 +5209,7 @@ __setup("cgroup_disable=", cgroup_disable); * Functons for CSS ID. */ -/* - *To get ID other than 0, this should be called when !cgroup_is_removed(). - */ +/* to get ID other than 0, this should be called when !cgroup_is_dead() */ unsigned short css_id(struct cgroup_subsys_state *css) { struct css_id *cssid; -- cgit v1.2.3 From ddd69148bdc45e5e3e55bfde3571daecd5a96d75 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 12 Jun 2013 21:04:54 -0700 Subject: cgroup: drop unnecessary RCU dancing from __put_css_set() __put_css_set() does RCU read access on @cgrp across dropping @cgrp->count so that it can continue accessing @cgrp even if the count reached zero and destruction of the cgroup commenced. Given that both sides - __css_put() and cgroup_destroy_locked() - are cold paths, this is unnecessary. Just making cgroup_destroy_locked() grab css_set_lock while checking @cgrp->count is enough. Remove the RCU read locking from __put_css_set() and make cgroup_destroy_locked() read-lock css_set_lock when checking @cgrp->count. This will also allow removing @cgrp->count. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 84efb344fdf6..1a68241ca835 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -407,19 +407,13 @@ static void __put_css_set(struct css_set *cset, int taskexit) list_del(&link->cset_link); list_del(&link->cgrp_link); - /* - * We may not be holding cgroup_mutex, and if cgrp->count is - * dropped to 0 the cgroup can be destroyed at any time, hence - * rcu_read_lock is used to keep it alive. - */ - rcu_read_lock(); + /* @cgrp can't go away while we're holding css_set_lock */ if (atomic_dec_and_test(&cgrp->count) && notify_on_release(cgrp)) { if (taskexit) set_bit(CGRP_RELEASABLE, &cgrp->flags); check_for_release(cgrp); } - rcu_read_unlock(); kfree(link); } @@ -4370,11 +4364,19 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) struct cgroup *parent = cgrp->parent; struct cgroup_event *event, *tmp; struct cgroup_subsys *ss; + bool empty; lockdep_assert_held(&d->d_inode->i_mutex); lockdep_assert_held(&cgroup_mutex); - if (atomic_read(&cgrp->count) || !list_empty(&cgrp->children)) + /* + * css_set_lock prevents @cgrp from being removed while + * __put_css_set() is in progress. + */ + read_lock(&css_set_lock); + empty = !atomic_read(&cgrp->count) && list_empty(&cgrp->children); + read_unlock(&css_set_lock); + if (!empty) return -EBUSY; /* @@ -5051,8 +5053,6 @@ void cgroup_exit(struct task_struct *tsk, int run_callbacks) static void check_for_release(struct cgroup *cgrp) { - /* All of these checks rely on RCU to keep the cgroup - * structure alive */ if (cgroup_is_releasable(cgrp) && !atomic_read(&cgrp->count) && list_empty(&cgrp->children)) { /* -- cgit v1.2.3 From 6f3d828f0fb7fdaffc6f32cb8a1cb7fcf8824598 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 12 Jun 2013 21:04:55 -0700 Subject: cgroup: remove cgroup->count and use cgroup->count tracks the number of css_sets associated with the cgroup and used only to verify that no css_set is associated when the cgroup is being destroyed. It's superflous as the destruction path can simply check whether cgroup->cset_links is empty instead. Drop cgroup->count and check ->cset_links directly from cgroup_destroy_locked(). Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 21 +++++---------------- 1 file changed, 5 insertions(+), 16 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 1a68241ca835..49bfd7b0bbda 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -408,8 +408,7 @@ static void __put_css_set(struct css_set *cset, int taskexit) list_del(&link->cgrp_link); /* @cgrp can't go away while we're holding css_set_lock */ - if (atomic_dec_and_test(&cgrp->count) && - notify_on_release(cgrp)) { + if (list_empty(&cgrp->cset_links) && notify_on_release(cgrp)) { if (taskexit) set_bit(CGRP_RELEASABLE, &cgrp->flags); check_for_release(cgrp); @@ -616,7 +615,6 @@ static void link_css_set(struct list_head *tmp_links, struct css_set *cset, link = list_first_entry(tmp_links, struct cgrp_cset_link, cset_link); link->cset = cset; link->cgrp = cgrp; - atomic_inc(&cgrp->count); list_move(&link->cset_link, &cgrp->cset_links); /* * Always add links to the tail of the list so that the list @@ -4370,11 +4368,11 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) lockdep_assert_held(&cgroup_mutex); /* - * css_set_lock prevents @cgrp from being removed while - * __put_css_set() is in progress. + * css_set_lock synchronizes access to ->cset_links and prevents + * @cgrp from being removed while __put_css_set() is in progress. */ read_lock(&css_set_lock); - empty = !atomic_read(&cgrp->count) && list_empty(&cgrp->children); + empty = list_empty(&cgrp->cset_links) && list_empty(&cgrp->children); read_unlock(&css_set_lock); if (!empty) return -EBUSY; @@ -5054,7 +5052,7 @@ void cgroup_exit(struct task_struct *tsk, int run_callbacks) static void check_for_release(struct cgroup *cgrp) { if (cgroup_is_releasable(cgrp) && - !atomic_read(&cgrp->count) && list_empty(&cgrp->children)) { + list_empty(&cgrp->cset_links) && list_empty(&cgrp->children)) { /* * Control Group is currently removeable. If it's not * already queued for a userspace notification, queue @@ -5422,11 +5420,6 @@ static void debug_css_free(struct cgroup *cont) kfree(cont->subsys[debug_subsys_id]); } -static u64 cgroup_refcount_read(struct cgroup *cont, struct cftype *cft) -{ - return atomic_read(&cont->count); -} - static u64 debug_taskcount_read(struct cgroup *cont, struct cftype *cft) { return cgroup_task_count(cont); @@ -5507,10 +5500,6 @@ static u64 releasable_read(struct cgroup *cgrp, struct cftype *cft) } static struct cftype debug_files[] = { - { - .name = "cgroup_refcount", - .read_u64 = cgroup_refcount_read, - }, { .name = "taskcount", .read_u64 = debug_taskcount_read, -- cgit v1.2.3 From 455050d23e1bfc47ca98e943ad5b2f3a9bbe45fb Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 13 Jun 2013 19:27:41 -0700 Subject: cgroup: reorder the operations in cgroup_destroy_locked() This patch reorders the operations in cgroup_destroy_locked() such that the userland visible parts happen before css offlining and removal from the ->sibling list. This will be used to make css use percpu refcnt. While at it, split out CGRP_DEAD related comment from the refcnt deactivation one and correct / clarify how different guarantees are met. While this patch changes the specific order of operations, it shouldn't cause any noticeable behavior difference. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 61 +++++++++++++++++++++++++++++++++------------------------ 1 file changed, 35 insertions(+), 26 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 49bfd7b0bbda..5a1ddecc3cfa 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4379,13 +4379,8 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) /* * Block new css_tryget() by deactivating refcnt and mark @cgrp - * removed. This makes future css_tryget() and child creation - * attempts fail thus maintaining the removal conditions verified - * above. - * - * Note that CGRP_DEAD assertion is depended upon by - * cgroup_next_sibling() to resume iteration after dropping RCU - * read lock. See cgroup_next_sibling() for details. + * removed. This makes future css_tryget() attempts fail which we + * guarantee to ->css_offline() callbacks. */ for_each_subsys(cgrp->root, ss) { struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; @@ -4393,8 +4388,41 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) WARN_ON(atomic_read(&css->refcnt) < 0); atomic_add(CSS_DEACT_BIAS, &css->refcnt); } + + /* + * Mark @cgrp dead. This prevents further task migration and child + * creation by disabling cgroup_lock_live_group(). Note that + * CGRP_DEAD assertion is depended upon by cgroup_next_sibling() to + * resume iteration after dropping RCU read lock. See + * cgroup_next_sibling() for details. + */ set_bit(CGRP_DEAD, &cgrp->flags); + /* CGRP_DEAD is set, remove from ->release_list for the last time */ + raw_spin_lock(&release_list_lock); + if (!list_empty(&cgrp->release_list)) + list_del_init(&cgrp->release_list); + raw_spin_unlock(&release_list_lock); + + /* + * Remove @cgrp directory. The removal puts the base ref but we + * aren't quite done with @cgrp yet, so hold onto it. + */ + dget(d); + cgroup_d_remove_dir(d); + + /* + * Unregister events and notify userspace. + * Notify userspace about cgroup removing only after rmdir of cgroup + * directory to avoid race between userspace and kernelspace. + */ + spin_lock(&cgrp->event_list_lock); + list_for_each_entry_safe(event, tmp, &cgrp->event_list, list) { + list_del_init(&event->list); + schedule_work(&event->remove); + } + spin_unlock(&cgrp->event_list_lock); + /* tell subsystems to initate destruction */ for_each_subsys(cgrp->root, ss) offline_css(ss, cgrp); @@ -4409,34 +4437,15 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) for_each_subsys(cgrp->root, ss) css_put(cgrp->subsys[ss->subsys_id]); - raw_spin_lock(&release_list_lock); - if (!list_empty(&cgrp->release_list)) - list_del_init(&cgrp->release_list); - raw_spin_unlock(&release_list_lock); - /* delete this cgroup from parent->children */ list_del_rcu(&cgrp->sibling); list_del_init(&cgrp->allcg_node); - dget(d); - cgroup_d_remove_dir(d); dput(d); set_bit(CGRP_RELEASABLE, &parent->flags); check_for_release(parent); - /* - * Unregister events and notify userspace. - * Notify userspace about cgroup removing only after rmdir of cgroup - * directory to avoid race between userspace and kernelspace. - */ - spin_lock(&cgrp->event_list_lock); - list_for_each_entry_safe(event, tmp, &cgrp->event_list, list) { - list_del_init(&event->list); - schedule_work(&event->remove); - } - spin_unlock(&cgrp->event_list_lock); - return 0; } -- cgit v1.2.3 From ea15f8ccdb430af1e8bc9b4e19a230eb4c356777 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 13 Jun 2013 19:27:42 -0700 Subject: cgroup: split cgroup destruction into two steps Split cgroup_destroy_locked() into two steps and put the latter half into cgroup_offline_fn() which is executed from a work item. The latter half is responsible for offlining the css's, removing the cgroup from internal lists, and propagating release notification to the parent. The separation is to allow using percpu refcnt for css. Note that this allows for other cgroup operations to happen between the first and second halves of destruction, including creating a new cgroup with the same name. As the target cgroup is marked DEAD in the first half and cgroup internals don't care about the names of cgroups, this should be fine. A comment explaining this will be added by the next patch which implements the actual percpu refcnting. As RCU freeing is guaranteed to happen after the second step of destruction, we can use the same work item for both. This patch renames cgroup->free_work to ->destroy_work and uses it for both purposes. INIT_WORK() is now performed right before queueing the work item. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 38 +++++++++++++++++++++++++++----------- 1 file changed, 27 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 5a1ddecc3cfa..df6814706cca 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -208,6 +208,7 @@ static struct cgroup_name root_cgroup_name = { .name = "/" }; */ static int need_forkexit_callback __read_mostly; +static void cgroup_offline_fn(struct work_struct *work); static int cgroup_destroy_locked(struct cgroup *cgrp); static int cgroup_addrm_files(struct cgroup *cgrp, struct cgroup_subsys *subsys, struct cftype cfts[], bool is_add); @@ -830,7 +831,7 @@ static struct cgroup_name *cgroup_alloc_name(struct dentry *dentry) static void cgroup_free_fn(struct work_struct *work) { - struct cgroup *cgrp = container_of(work, struct cgroup, free_work); + struct cgroup *cgrp = container_of(work, struct cgroup, destroy_work); struct cgroup_subsys *ss; mutex_lock(&cgroup_mutex); @@ -875,7 +876,8 @@ static void cgroup_free_rcu(struct rcu_head *head) { struct cgroup *cgrp = container_of(head, struct cgroup, rcu_head); - schedule_work(&cgrp->free_work); + INIT_WORK(&cgrp->destroy_work, cgroup_free_fn); + schedule_work(&cgrp->destroy_work); } static void cgroup_diput(struct dentry *dentry, struct inode *inode) @@ -1407,7 +1409,6 @@ static void init_cgroup_housekeeping(struct cgroup *cgrp) INIT_LIST_HEAD(&cgrp->allcg_node); INIT_LIST_HEAD(&cgrp->release_list); INIT_LIST_HEAD(&cgrp->pidlists); - INIT_WORK(&cgrp->free_work, cgroup_free_fn); mutex_init(&cgrp->pidlist_mutex); INIT_LIST_HEAD(&cgrp->event_list); spin_lock_init(&cgrp->event_list_lock); @@ -2991,12 +2992,13 @@ struct cgroup *cgroup_next_sibling(struct cgroup *pos) /* * @pos could already have been removed. Once a cgroup is removed, * its ->sibling.next is no longer updated when its next sibling - * changes. As CGRP_DEAD is set on removal which is fully - * serialized, if we see it unasserted, it's guaranteed that the - * next sibling hasn't finished its grace period even if it's - * already removed, and thus safe to dereference from this RCU - * critical section. If ->sibling.next is inaccessible, - * cgroup_is_dead() is guaranteed to be visible as %true here. + * changes. As CGRP_DEAD assertion is serialized and happens + * before the cgroup is taken off the ->sibling list, if we see it + * unasserted, it's guaranteed that the next sibling hasn't + * finished its grace period even if it's already removed, and thus + * safe to dereference from this RCU critical section. If + * ->sibling.next is inaccessible, cgroup_is_dead() is guaranteed + * to be visible as %true here. */ if (likely(!cgroup_is_dead(pos))) { next = list_entry_rcu(pos->sibling.next, struct cgroup, sibling); @@ -4359,7 +4361,6 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) __releases(&cgroup_mutex) __acquires(&cgroup_mutex) { struct dentry *d = cgrp->dentry; - struct cgroup *parent = cgrp->parent; struct cgroup_event *event, *tmp; struct cgroup_subsys *ss; bool empty; @@ -4423,6 +4424,21 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) } spin_unlock(&cgrp->event_list_lock); + INIT_WORK(&cgrp->destroy_work, cgroup_offline_fn); + schedule_work(&cgrp->destroy_work); + + return 0; +}; + +static void cgroup_offline_fn(struct work_struct *work) +{ + struct cgroup *cgrp = container_of(work, struct cgroup, destroy_work); + struct cgroup *parent = cgrp->parent; + struct dentry *d = cgrp->dentry; + struct cgroup_subsys *ss; + + mutex_lock(&cgroup_mutex); + /* tell subsystems to initate destruction */ for_each_subsys(cgrp->root, ss) offline_css(ss, cgrp); @@ -4446,7 +4462,7 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) set_bit(CGRP_RELEASABLE, &parent->flags); check_for_release(parent); - return 0; + mutex_unlock(&cgroup_mutex); } static int cgroup_rmdir(struct inode *unused_dir, struct dentry *dentry) -- cgit v1.2.3 From d3daf28da16a30af95bfb303189a634a87606725 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 13 Jun 2013 19:39:16 -0700 Subject: cgroup: use percpu refcnt for cgroup_subsys_states A css (cgroup_subsys_state) is how each cgroup is represented to a controller. As such, it can be used in hot paths across the various subsystems different controllers are associated with. One of the common operations is reference counting, which up until now has been implemented using a global atomic counter and can have significant adverse impact on scalability. For example, css refcnt can be gotten and put multiple times by blkcg for each IO request. For highops configurations which try to do as much per-cpu as possible, the global frequent refcnting can be very expensive. In general, given the various and hugely diverse paths css's end up being used from, we need to make it cheap and highly scalable. In its usage, css refcnting isn't very different from module refcnting. This patch converts css refcnting to use the recently added percpu_ref. css_get/tryget/put() directly maps to the matching percpu_ref operations and the deactivation logic is no longer necessary as percpu_ref already has refcnt killing. The only complication is that as the refcnt is per-cpu, percpu_ref_kill() in itself doesn't ensure that further tryget operations will fail, which we need to guarantee before invoking ->css_offline()'s. This is resolved collecting kill confirmation using percpu_ref_kill_and_confirm() and initiating the offline phase of destruction after all css refcnt's are confirmed to be seen as killed on all CPUs. The previous patches already splitted destruction into two phases, so percpu_ref_kill_and_confirm() can be hooked up easily. This patch removes css_refcnt() which is used for rcu dereference sanity check in css_id(). While we can add a percpu refcnt API to ask the same question, css_id() itself is scheduled to be removed fairly soon, so let's not bother with it. Just drop the sanity check and use rcu_dereference_raw() instead. v2: - init_cgroup_css() was calling percpu_ref_init() without checking the return value. This causes two problems - the obvious lack of error handling and percpu_ref_init() being called from cgroup_init_subsys() before the allocators are up, which triggers warnings but doesn't cause actual problems as the refcnt isn't used for roots anyway. Fix both by moving percpu_ref_init() to cgroup_create(). - The base references were put too early by percpu_ref_kill_and_confirm() and cgroup_offline_fn() put the refs one extra time. This wasn't noticeable because css's go through another RCU grace period before being freed. Update cgroup_destroy_locked() to grab an extra reference before killing the refcnts. This problem was noticed by Kent. Signed-off-by: Tejun Heo Reviewed-by: Kent Overstreet Acked-by: Li Zefan Cc: Michal Hocko Cc: Mike Snitzer Cc: Vivek Goyal Cc: "Alasdair G. Kergon" Cc: Jens Axboe Cc: Mikulas Patocka Cc: Glauber Costa --- kernel/cgroup.c | 165 +++++++++++++++++++++++++++++++++++--------------------- 1 file changed, 104 insertions(+), 61 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index ebbfc043153f..2e9da7bf25cb 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -63,9 +63,6 @@ #include -/* css deactivation bias, makes css->refcnt negative to deny new trygets */ -#define CSS_DEACT_BIAS INT_MIN - /* * cgroup_mutex is the master lock. Any modification to cgroup or its * hierarchy must be performed while holding it. @@ -213,19 +210,6 @@ static int cgroup_destroy_locked(struct cgroup *cgrp); static int cgroup_addrm_files(struct cgroup *cgrp, struct cgroup_subsys *subsys, struct cftype cfts[], bool is_add); -static int css_unbias_refcnt(int refcnt) -{ - return refcnt >= 0 ? refcnt : refcnt - CSS_DEACT_BIAS; -} - -/* the current nr of refs, always >= 0 whether @css is deactivated or not */ -static int css_refcnt(struct cgroup_subsys_state *css) -{ - int v = atomic_read(&css->refcnt); - - return css_unbias_refcnt(v); -} - /* convenient tests for these bits */ static inline bool cgroup_is_dead(const struct cgroup *cgrp) { @@ -4139,12 +4123,19 @@ static void css_dput_fn(struct work_struct *work) deactivate_super(sb); } +static void css_release(struct percpu_ref *ref) +{ + struct cgroup_subsys_state *css = + container_of(ref, struct cgroup_subsys_state, refcnt); + + schedule_work(&css->dput_work); +} + static void init_cgroup_css(struct cgroup_subsys_state *css, struct cgroup_subsys *ss, struct cgroup *cgrp) { css->cgroup = cgrp; - atomic_set(&css->refcnt, 1); css->flags = 0; css->id = NULL; if (cgrp == dummytop) @@ -4266,7 +4257,13 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, err = PTR_ERR(css); goto err_free_all; } + + err = percpu_ref_init(&css->refcnt, css_release); + if (err) + goto err_free_all; + init_cgroup_css(css, ss, cgrp); + if (ss->use_id) { err = alloc_css_id(ss, parent, cgrp); if (err) @@ -4331,8 +4328,12 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, err_free_all: for_each_subsys(root, ss) { - if (cgrp->subsys[ss->subsys_id]) + struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + + if (css) { + percpu_ref_cancel_init(&css->refcnt); ss->css_free(cgrp); + } } mutex_unlock(&cgroup_mutex); /* Release the reference count that we took on the superblock */ @@ -4360,6 +4361,48 @@ static int cgroup_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode) return cgroup_create(c_parent, dentry, mode | S_IFDIR); } +static void cgroup_css_killed(struct cgroup *cgrp) +{ + if (!atomic_dec_and_test(&cgrp->css_kill_cnt)) + return; + + /* percpu ref's of all css's are killed, kick off the next step */ + INIT_WORK(&cgrp->destroy_work, cgroup_offline_fn); + schedule_work(&cgrp->destroy_work); +} + +static void css_ref_killed_fn(struct percpu_ref *ref) +{ + struct cgroup_subsys_state *css = + container_of(ref, struct cgroup_subsys_state, refcnt); + + cgroup_css_killed(css->cgroup); +} + +/** + * cgroup_destroy_locked - the first stage of cgroup destruction + * @cgrp: cgroup to be destroyed + * + * css's make use of percpu refcnts whose killing latency shouldn't be + * exposed to userland and are RCU protected. Also, cgroup core needs to + * guarantee that css_tryget() won't succeed by the time ->css_offline() is + * invoked. To satisfy all the requirements, destruction is implemented in + * the following two steps. + * + * s1. Verify @cgrp can be destroyed and mark it dying. Remove all + * userland visible parts and start killing the percpu refcnts of + * css's. Set up so that the next stage will be kicked off once all + * the percpu refcnts are confirmed to be killed. + * + * s2. Invoke ->css_offline(), mark the cgroup dead and proceed with the + * rest of destruction. Once all cgroup references are gone, the + * cgroup is RCU-freed. + * + * This function implements s1. After this step, @cgrp is gone as far as + * the userland is concerned and a new cgroup with the same name may be + * created. As cgroup doesn't care about the names internally, this + * doesn't cause any problem. + */ static int cgroup_destroy_locked(struct cgroup *cgrp) __releases(&cgroup_mutex) __acquires(&cgroup_mutex) { @@ -4382,16 +4425,34 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) return -EBUSY; /* - * Block new css_tryget() by deactivating refcnt and mark @cgrp - * removed. This makes future css_tryget() attempts fail which we - * guarantee to ->css_offline() callbacks. + * Block new css_tryget() by killing css refcnts. cgroup core + * guarantees that, by the time ->css_offline() is invoked, no new + * css reference will be given out via css_tryget(). We can't + * simply call percpu_ref_kill() and proceed to offlining css's + * because percpu_ref_kill() doesn't guarantee that the ref is seen + * as killed on all CPUs on return. + * + * Use percpu_ref_kill_and_confirm() to get notifications as each + * css is confirmed to be seen as killed on all CPUs. The + * notification callback keeps track of the number of css's to be + * killed and schedules cgroup_offline_fn() to perform the rest of + * destruction once the percpu refs of all css's are confirmed to + * be killed. */ + atomic_set(&cgrp->css_kill_cnt, 1); for_each_subsys(cgrp->root, ss) { struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; - WARN_ON(atomic_read(&css->refcnt) < 0); - atomic_add(CSS_DEACT_BIAS, &css->refcnt); + /* + * Killing would put the base ref, but we need to keep it + * alive until after ->css_offline. + */ + percpu_ref_get(&css->refcnt); + + atomic_inc(&cgrp->css_kill_cnt); + percpu_ref_kill_and_confirm(&css->refcnt, css_ref_killed_fn); } + cgroup_css_killed(cgrp); /* * Mark @cgrp dead. This prevents further task migration and child @@ -4427,12 +4488,19 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) } spin_unlock(&cgrp->event_list_lock); - INIT_WORK(&cgrp->destroy_work, cgroup_offline_fn); - schedule_work(&cgrp->destroy_work); - return 0; }; +/** + * cgroup_offline_fn - the second step of cgroup destruction + * @work: cgroup->destroy_free_work + * + * This function is invoked from a work item for a cgroup which is being + * destroyed after the percpu refcnts of all css's are guaranteed to be + * seen as killed on all CPUs, and performs the rest of destruction. This + * is the second step of destruction described in the comment above + * cgroup_destroy_locked(). + */ static void cgroup_offline_fn(struct work_struct *work) { struct cgroup *cgrp = container_of(work, struct cgroup, destroy_work); @@ -4442,16 +4510,19 @@ static void cgroup_offline_fn(struct work_struct *work) mutex_lock(&cgroup_mutex); - /* tell subsystems to initate destruction */ + /* + * css_tryget() is guaranteed to fail now. Tell subsystems to + * initate destruction. + */ for_each_subsys(cgrp->root, ss) offline_css(ss, cgrp); /* - * Put all the base refs. Each css holds an extra reference to the - * cgroup's dentry and cgroup removal proceeds regardless of css - * refs. On the last put of each css, whenever that may be, the - * extra dentry ref is put so that dentry destruction happens only - * after all css's are released. + * Put the css refs from cgroup_destroy_locked(). Each css holds + * an extra reference to the cgroup's dentry and cgroup removal + * proceeds regardless of css refs. On the last put of each css, + * whenever that may be, the extra dentry ref is put so that dentry + * destruction happens only after all css's are released. */ for_each_subsys(cgrp->root, ss) css_put(cgrp->subsys[ss->subsys_id]); @@ -5100,34 +5171,6 @@ static void check_for_release(struct cgroup *cgrp) } } -/* Caller must verify that the css is not for root cgroup */ -bool __css_tryget(struct cgroup_subsys_state *css) -{ - while (true) { - int t, v; - - v = css_refcnt(css); - t = atomic_cmpxchg(&css->refcnt, v, v + 1); - if (likely(t == v)) - return true; - else if (t < 0) - return false; - cpu_relax(); - } -} -EXPORT_SYMBOL_GPL(__css_tryget); - -/* Caller must verify that the css is not for root cgroup */ -void __css_put(struct cgroup_subsys_state *css) -{ - int v; - - v = css_unbias_refcnt(atomic_dec_return(&css->refcnt)); - if (v == 0) - schedule_work(&css->dput_work); -} -EXPORT_SYMBOL_GPL(__css_put); - /* * Notify userspace when a cgroup is released, by running the * configured release agent with the name of the cgroup (path @@ -5245,7 +5288,7 @@ unsigned short css_id(struct cgroup_subsys_state *css) * on this or this is under rcu_read_lock(). Once css->id is allocated, * it's unchanged until freed. */ - cssid = rcu_dereference_check(css->id, css_refcnt(css)); + cssid = rcu_dereference_raw(css->id); if (cssid) return cssid->id; -- cgit v1.2.3 From c9e5fe66f5947c9e56dfc7655e5b4b127ca2120f Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 14 Jun 2013 11:18:27 +0800 Subject: cpuset: rename @cont to @cgrp Cont is short for container. control group was named process container at first, but then people found container already has a meaning in linux kernel. Clean up the leftover variable name @cont. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cpuset.c | 32 ++++++++++++++++---------------- 1 file changed, 16 insertions(+), 16 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 4c17d96bd3a5..654c95979028 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -116,9 +116,9 @@ struct cpuset { }; /* Retrieve the cpuset for a cgroup */ -static inline struct cpuset *cgroup_cs(struct cgroup *cont) +static inline struct cpuset *cgroup_cs(struct cgroup *cgrp) { - return container_of(cgroup_subsys_state(cont, cpuset_subsys_id), + return container_of(cgroup_subsys_state(cgrp, cpuset_subsys_id), struct cpuset, css); } @@ -433,7 +433,7 @@ static void free_trial_cpuset(struct cpuset *trial) static int validate_change(const struct cpuset *cur, const struct cpuset *trial) { - struct cgroup *cont; + struct cgroup *cgrp; struct cpuset *c, *par; int ret; @@ -441,7 +441,7 @@ static int validate_change(const struct cpuset *cur, const struct cpuset *trial) /* Each of our child cpusets must be a subset of us */ ret = -EBUSY; - cpuset_for_each_child(c, cont, cur) + cpuset_for_each_child(c, cgrp, cur) if (!is_cpuset_subset(c, trial)) goto out; @@ -462,7 +462,7 @@ static int validate_change(const struct cpuset *cur, const struct cpuset *trial) * overlap */ ret = -EINVAL; - cpuset_for_each_child(c, cont, par) { + cpuset_for_each_child(c, cgrp, par) { if ((is_cpu_exclusive(trial) || is_cpu_exclusive(c)) && c != cur && cpumask_intersects(trial->cpus_allowed, c->cpus_allowed)) @@ -1759,13 +1759,13 @@ static size_t cpuset_sprintf_memlist(char *page, struct cpuset *cs) return count; } -static ssize_t cpuset_common_file_read(struct cgroup *cont, +static ssize_t cpuset_common_file_read(struct cgroup *cgrp, struct cftype *cft, struct file *file, char __user *buf, size_t nbytes, loff_t *ppos) { - struct cpuset *cs = cgroup_cs(cont); + struct cpuset *cs = cgroup_cs(cgrp); cpuset_filetype_t type = cft->private; char *page; ssize_t retval = 0; @@ -1795,9 +1795,9 @@ out: return retval; } -static u64 cpuset_read_u64(struct cgroup *cont, struct cftype *cft) +static u64 cpuset_read_u64(struct cgroup *cgrp, struct cftype *cft) { - struct cpuset *cs = cgroup_cs(cont); + struct cpuset *cs = cgroup_cs(cgrp); cpuset_filetype_t type = cft->private; switch (type) { case FILE_CPU_EXCLUSIVE: @@ -1826,9 +1826,9 @@ static u64 cpuset_read_u64(struct cgroup *cont, struct cftype *cft) return 0; } -static s64 cpuset_read_s64(struct cgroup *cont, struct cftype *cft) +static s64 cpuset_read_s64(struct cgroup *cgrp, struct cftype *cft) { - struct cpuset *cs = cgroup_cs(cont); + struct cpuset *cs = cgroup_cs(cgrp); cpuset_filetype_t type = cft->private; switch (type) { case FILE_SCHED_RELAX_DOMAIN_LEVEL: @@ -1940,14 +1940,14 @@ static struct cftype files[] = { /* * cpuset_css_alloc - allocate a cpuset css - * cont: control group that the new cpuset will be part of + * cgrp: control group that the new cpuset will be part of */ -static struct cgroup_subsys_state *cpuset_css_alloc(struct cgroup *cont) +static struct cgroup_subsys_state *cpuset_css_alloc(struct cgroup *cgrp) { struct cpuset *cs; - if (!cont->parent) + if (!cgrp->parent) return &top_cpuset.css; cs = kzalloc(sizeof(*cs), GFP_KERNEL); @@ -2042,9 +2042,9 @@ static void cpuset_css_offline(struct cgroup *cgrp) * will call rebuild_sched_domains_locked(). */ -static void cpuset_css_free(struct cgroup *cont) +static void cpuset_css_free(struct cgroup *cgrp) { - struct cpuset *cs = cgroup_cs(cont); + struct cpuset *cs = cgroup_cs(cgrp); free_cpumask_var(cs->cpus_allowed); kfree(cs); -- cgit v1.2.3 From 29ce3785b22da47c49f4ef6e14b9014fa5dee261 Mon Sep 17 00:00:00 2001 From: James Bottomley Date: Wed, 8 May 2013 14:05:34 -0700 Subject: idle: Enable interrupts in the weak arch_cpu_idle() implementation PARISC bootup triggers the warning at kernel/cpu/idle.c:96. That's caused by the weak arch_cpu_idle() implementation, which is provided to avoid that architectures implement idle_poll over and over. The switchover to polling mode happens in the first call of the weak arch_cpu_idle() implementation, but that code fails to reenable interrupts and therefor triggers the warning. Fix this by enabling interrupts in the weak arch_cpu_idle() code. [ tglx: Made the changelog match the patch ] Signed-off-by: James Bottomley Reviewed-by: Srivatsa S. Bhat Link: http://lkml.kernel.org/r/1371236142.2726.43.camel@dabdike Signed-off-by: Thomas Gleixner --- kernel/cpu/idle.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/cpu/idle.c b/kernel/cpu/idle.c index bf2ee1aafa0e..e695c0a0bcb5 100644 --- a/kernel/cpu/idle.c +++ b/kernel/cpu/idle.c @@ -59,6 +59,7 @@ void __weak arch_cpu_idle_dead(void) { } void __weak arch_cpu_idle(void) { cpu_idle_force_poll = 1; + local_irq_enable(); } /* -- cgit v1.2.3 From 8aac62706adaaf0fab02c4327761561c8bda9448 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 14 Jun 2013 21:09:49 +0200 Subject: move exit_task_namespaces() outside of exit_notify() exit_notify() does exit_task_namespaces() after forget_original_parent(). This was needed to ensure that ->nsproxy can't be cleared prematurely, an exiting child we are going to reparent can do do_notify_parent() and use the parent's (ours) pid_ns. However, after 32084504 "pidns: use task_active_pid_ns in do_notify_parent" ->nsproxy != NULL is no longer needed, we rely on task_active_pid_ns(). Move exit_task_namespaces() from exit_notify() to do_exit(), after exit_fs() and before exit_task_work(). This solves the problem reported by Andrey, free_ipc_ns()->shm_destroy() does fput() which needs task_work_add(). Note: this particular problem can be fixed if we change fput(), and that change makes sense anyway. But there is another reason to move the callsite. The original reason for exit_task_namespaces() from the middle of exit_notify() was subtle and it has already gone away, now this looks confusing. And this allows us do simplify exit_notify(), we can avoid unlock/lock(tasklist) and we can use ->exit_state instead of PF_EXITING in forget_original_parent(). Reported-by: Andrey Vagin Signed-off-by: Oleg Nesterov Acked-by: "Eric W. Biederman" Acked-by: Andrey Vagin Signed-off-by: Al Viro --- kernel/exit.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index af2eb3cbd499..7bb73f9d09db 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -649,7 +649,6 @@ static void exit_notify(struct task_struct *tsk, int group_dead) * jobs, send them a SIGHUP and then a SIGCONT. (POSIX 3.2.2.2) */ forget_original_parent(tsk); - exit_task_namespaces(tsk); write_lock_irq(&tasklist_lock); if (group_dead) @@ -795,6 +794,7 @@ void do_exit(long code) exit_shm(tsk); exit_files(tsk); exit_fs(tsk); + exit_task_namespaces(tsk); exit_task_work(tsk); check_stack_usage(); exit_thread(); -- cgit v1.2.3 From c5cdc67a58a22c49f558b450c6f748251ceb2e7b Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Thu, 13 Jun 2013 22:19:43 +0000 Subject: irqdomain: Remove temporary MIPS workaround code The MIPS interrupt controllers are all registering their own irq_domains now. Drop the MIPS specific code because it is no longer needed. Signed-off-by: Grant Likely Cc: linux-mips@linux-mips.org Cc: linux-kernel@vger.kernel.org Patchwork: https://patchwork.linux-mips.org/patch/5458/ Signed-off-by: Ralf Baechle --- kernel/irq/irqdomain.c | 12 ------------ 1 file changed, 12 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c index 54a4d5223238..a341b3d433ad 100644 --- a/kernel/irq/irqdomain.c +++ b/kernel/irq/irqdomain.c @@ -665,18 +665,6 @@ unsigned int irq_create_of_mapping(struct device_node *controller, domain = controller ? irq_find_host(controller) : irq_default_domain; if (!domain) { -#ifdef CONFIG_MIPS - /* - * Workaround to avoid breaking interrupt controller drivers - * that don't yet register an irq_domain. This is temporary - * code. ~~~gcl, Feb 24, 2012 - * - * Scheduled for removal in Linux v3.6. That should be enough - * time. - */ - if (intsize > 0) - return intspec[0]; -#endif pr_warning("no irq domain found for %s !\n", of_node_full_name(controller)); return 0; -- cgit v1.2.3 From 336ae1180df5f69b9e0fb6561bec01c5f64361cf Mon Sep 17 00:00:00 2001 From: Stephen Boyd Date: Mon, 17 Jun 2013 15:40:58 -0700 Subject: ARM: sched_clock: Load cycle count after epoch stabilizes There is a small race between when the cycle count is read from the hardware and when the epoch stabilizes. Consider this scenario: CPU0 CPU1 ---- ---- cyc = read_sched_clock() cyc_to_sched_clock() update_sched_clock() ... cd.epoch_cyc = cyc; epoch_cyc = cd.epoch_cyc; ... epoch_ns + cyc_to_ns((cyc - epoch_cyc) The cyc on cpu0 was read before the epoch changed. But we calculate the nanoseconds based on the new epoch by subtracting the new epoch from the old cycle count. Since epoch is most likely larger than the old cycle count we calculate a large number that will be converted to nanoseconds and added to epoch_ns, causing time to jump forward too much. Fix this problem by reading the hardware after the epoch has stabilized. Cc: Russell King Signed-off-by: Stephen Boyd Signed-off-by: John Stultz --- kernel/time/sched_clock.c | 19 ++++++++----------- 1 file changed, 8 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/time/sched_clock.c b/kernel/time/sched_clock.c index aad1ae6077ef..a326f27d7f09 100644 --- a/kernel/time/sched_clock.c +++ b/kernel/time/sched_clock.c @@ -49,10 +49,14 @@ static inline u64 notrace cyc_to_ns(u64 cyc, u32 mult, u32 shift) return (cyc * mult) >> shift; } -static unsigned long long notrace cyc_to_sched_clock(u32 cyc, u32 mask) +static unsigned long long notrace sched_clock_32(void) { u64 epoch_ns; u32 epoch_cyc; + u32 cyc; + + if (cd.suspended) + return cd.epoch_ns; /* * Load the epoch_cyc and epoch_ns atomically. We do this by @@ -68,7 +72,9 @@ static unsigned long long notrace cyc_to_sched_clock(u32 cyc, u32 mask) smp_rmb(); } while (epoch_cyc != cd.epoch_cyc_copy); - return epoch_ns + cyc_to_ns((cyc - epoch_cyc) & mask, cd.mult, cd.shift); + cyc = read_sched_clock(); + cyc = (cyc - epoch_cyc) & sched_clock_mask; + return epoch_ns + cyc_to_ns(cyc, cd.mult, cd.shift); } /* @@ -160,19 +166,10 @@ void __init setup_sched_clock(u32 (*read)(void), int bits, unsigned long rate) pr_debug("Registered %pF as sched_clock source\n", read); } -static unsigned long long notrace sched_clock_32(void) -{ - u32 cyc = read_sched_clock(); - return cyc_to_sched_clock(cyc, sched_clock_mask); -} - unsigned long long __read_mostly (*sched_clock_func)(void) = sched_clock_32; unsigned long long notrace sched_clock(void) { - if (cd.suspended) - return cd.epoch_ns; - return sched_clock_func(); } -- cgit v1.2.3 From 37074c5a1b9979d05b9effc7634385fc0fa7ccc4 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Uwe=20Kleine-K=C3=B6nig?= Date: Wed, 12 Jun 2013 14:24:12 +0200 Subject: irq/generic-chip: fix a few kernel-doc entries MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Uwe Kleine-König Signed-off-by: Jiri Kosina --- kernel/irq/generic-chip.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/generic-chip.c b/kernel/irq/generic-chip.c index c89295a8f668..b34e7267b817 100644 --- a/kernel/irq/generic-chip.c +++ b/kernel/irq/generic-chip.c @@ -48,7 +48,7 @@ void irq_gc_mask_disable_reg(struct irq_data *d) } /** - * irq_gc_mask_set_mask_bit - Mask chip via setting bit in mask register + * irq_gc_mask_set_bit - Mask irq via setting bit in mask register * @d: irq_data * * Chip has a single mask register. Values of this register are cached @@ -66,7 +66,7 @@ void irq_gc_mask_set_bit(struct irq_data *d) } /** - * irq_gc_mask_set_mask_bit - Mask chip via clearing bit in mask register + * irq_gc_mask_clr_bit - Mask chip via clearing bit in mask register * @d: irq_data * * Chip has a single mask register. Values of this register are cached @@ -130,7 +130,7 @@ void irq_gc_ack_clr_bit(struct irq_data *d) } /** - * irq_gc_mask_disable_reg_and_ack- Mask and ack pending interrupt + * irq_gc_mask_disable_reg_and_ack - Mask and ack pending interrupt * @d: irq_data */ void irq_gc_mask_disable_reg_and_ack(struct irq_data *d) -- cgit v1.2.3 From 6db8e85c5c1f89cd0183b76dab027c81009f129f Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 14 Jun 2013 11:18:22 -0700 Subject: cgroup: disallow rename(2) if sane_behavior cgroup's rename(2) isn't a proper migration implementation - it can't move the cgroup to a different parent in the hierarchy. All it can do is swapping the name string for that cgroup. This isn't useful and can mislead users to think that cgroup supports proper cgroup-level migration. Disallow rename(2) if sane_behavior. v2: Fail with -EPERM instead of -EINVAL so that it matches the vfs return value when ->rename is not implemented as suggested by Li. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 2e9da7bf25cb..c2c64005bbc2 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -2508,6 +2508,13 @@ static int cgroup_rename(struct inode *old_dir, struct dentry *old_dentry, cgrp = __d_cgrp(old_dentry); + /* + * This isn't a proper migration and its usefulness is very + * limited. Disallow if sane_behavior. + */ + if (cgroup_sane_behavior(cgrp)) + return -EPERM; + name = cgroup_alloc_name(new_dentry); if (!name) return -ENOMEM; -- cgit v1.2.3 From 084457f284abf6789d90509ee11dae383842b23b Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Tue, 18 Jun 2013 18:40:19 +0800 Subject: cgroup: fix umount vs cgroup_cfts_commit() race cgroup_cfts_commit() uses dget() to keep cgroup alive after cgroup_mutex is dropped, but dget() won't prevent cgroupfs from being umounted. When the race happens, vfs will see some dentries with non-zero refcnt while umount is in process. Keep running this: mount -t cgroup -o blkio xxx /cgroup umount /cgroup And this: modprobe cfq-iosched rmmod cfs-iosched After a while, the BUG() in shrink_dcache_for_umount_subtree() may be triggered: BUG: Dentry xxx{i=0,n=blkio.yyy} still in use (1) [umount of cgroup cgroup] Signed-off-by: Li Zefan Signed-off-by: Tejun Heo Cc: stable@vger.kernel.org --- kernel/cgroup.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index c2c64005bbc2..0224f6b3103e 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -2798,13 +2798,17 @@ static void cgroup_cfts_commit(struct cgroup_subsys *ss, { LIST_HEAD(pending); struct cgroup *cgrp, *n; + struct super_block *sb = ss->root->sb; /* %NULL @cfts indicates abort and don't bother if @ss isn't attached */ - if (cfts && ss->root != &rootnode) { + if (cfts && ss->root != &rootnode && + atomic_inc_not_zero(sb->s_active)) { list_for_each_entry(cgrp, &ss->root->allcg_list, allcg_node) { dget(cgrp->dentry); list_add_tail(&cgrp->cft_q_node, &pending); } + } else { + sb = NULL; } mutex_unlock(&cgroup_mutex); @@ -2827,6 +2831,9 @@ static void cgroup_cfts_commit(struct cgroup_subsys *ss, dput(cgrp->dentry); } + if (sb) + deactivate_super(sb); + mutex_unlock(&cgroup_cft_mutex); } -- cgit v1.2.3 From 1c8158eeae0f37d0eee9f1fbe68080df6a408df2 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Tue, 18 Jun 2013 18:41:10 +0800 Subject: cgroup: fix umount vs cgroup_event_remove() race commit 5db9a4d99b0157a513944e9a44d29c9cec2e91dc Author: Tejun Heo Date: Sat Jul 7 16:08:18 2012 -0700 cgroup: fix cgroup hierarchy umount race This commit fixed a race caused by the dput() in css_dput_fn(), but the dput() in cgroup_event_remove() can also lead to the same BUG(). Signed-off-by: Li Zefan Signed-off-by: Tejun Heo Cc: stable@vger.kernel.org --- kernel/cgroup.c | 25 +++++++++++++++++++------ 1 file changed, 19 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 0224f6b3103e..7db2940bfc77 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -3821,6 +3821,23 @@ static int cgroup_write_notify_on_release(struct cgroup *cgrp, return 0; } +/* + * When dput() is called asynchronously, if umount has been done and + * then deactivate_super() in cgroup_free_fn() kills the superblock, + * there's a small window that vfs will see the root dentry with non-zero + * refcnt and trigger BUG(). + * + * That's why we hold a reference before dput() and drop it right after. + */ +static void cgroup_dput(struct cgroup *cgrp) +{ + struct super_block *sb = cgrp->root->sb; + + atomic_inc(&sb->s_active); + dput(cgrp->dentry); + deactivate_super(sb); +} + /* * Unregister event and free resources. * @@ -3841,7 +3858,7 @@ static void cgroup_event_remove(struct work_struct *work) eventfd_ctx_put(event->eventfd); kfree(event); - dput(cgrp->dentry); + cgroup_dput(cgrp); } /* @@ -4129,12 +4146,8 @@ static void css_dput_fn(struct work_struct *work) { struct cgroup_subsys_state *css = container_of(work, struct cgroup_subsys_state, dput_work); - struct dentry *dentry = css->cgroup->dentry; - struct super_block *sb = dentry->d_sb; - atomic_inc(&sb->s_active); - dput(dentry); - deactivate_super(sb); + cgroup_dput(css->cgroup); } static void css_release(struct percpu_ref *ref) -- cgit v1.2.3 From f57947d27711451a7739a25bba6cddc8a385e438 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Tue, 18 Jun 2013 18:41:53 +0800 Subject: cgroup: fix memory leak in cgroup_rm_cftypes() The memory allocated in cgroup_add_cftypes() should be freed. The effect of this bug is we leak a bit memory everytime we unload cfq-iosched module if blkio cgroup is enabled. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cgroup.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 7db2940bfc77..1d4f471de8d5 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -2889,7 +2889,8 @@ int cgroup_rm_cftypes(struct cgroup_subsys *ss, struct cftype *cfts) list_for_each_entry(set, &ss->cftsets, node) { if (set->cfts == cfts) { - list_del_init(&set->node); + list_del(&set->node); + kfree(set); cgroup_cfts_commit(ss, cfts, false); return 0; } -- cgit v1.2.3 From 0541881502a1276149889fe468662ff6a8fc8f6d Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Thu, 13 Jun 2013 13:17:02 -0700 Subject: range: Do not add new blank slot with add_range_with_merge Joshua reported: Commit cd7b304dfaf1 (x86, range: fix missing merge during add range) broke mtrr cleanup on his setup in 3.9.5. corresponding commit in upstream is fbe06b7bae7c. The reason is add_range_with_merge could generate blank spot. We could avoid that by searching new expanded start/end, that new range should include all connected ranges in range array. At last add the new expanded start/end to the range array. Also move up left array so do not add new blank slot in the range array. -v2: move left array to avoid enhance add_range() -v3: include fix from Joshua about memmove declaring when DYN_DEBUG is used. Reported-by: Joshua Covington Tested-by: Joshua Covington Signed-off-by: Yinghai Lu Link: http://lkml.kernel.org/r/1371154622-8929-3-git-send-email-yinghai@kernel.org Cc: v3.9 Signed-off-by: H. Peter Anvin --- kernel/range.c | 21 +++++++++++---------- 1 file changed, 11 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/range.c b/kernel/range.c index eb911dbce267..322ea8e93e4b 100644 --- a/kernel/range.c +++ b/kernel/range.c @@ -4,7 +4,7 @@ #include #include #include - +#include #include int add_range(struct range *range, int az, int nr_range, u64 start, u64 end) @@ -32,9 +32,8 @@ int add_range_with_merge(struct range *range, int az, int nr_range, if (start >= end) return nr_range; - /* Try to merge it with old one: */ + /* get new start/end: */ for (i = 0; i < nr_range; i++) { - u64 final_start, final_end; u64 common_start, common_end; if (!range[i].end) @@ -45,14 +44,16 @@ int add_range_with_merge(struct range *range, int az, int nr_range, if (common_start > common_end) continue; - final_start = min(range[i].start, start); - final_end = max(range[i].end, end); + /* new start/end, will add it back at last */ + start = min(range[i].start, start); + end = max(range[i].end, end); - /* clear it and add it back for further merge */ - range[i].start = 0; - range[i].end = 0; - return add_range_with_merge(range, az, nr_range, - final_start, final_end); + memmove(&range[i], &range[i + 1], + (nr_range - (i + 1)) * sizeof(range[i])); + range[nr_range - 1].start = 0; + range[nr_range - 1].end = 0; + nr_range--; + i--; } /* Need to add it: */ -- cgit v1.2.3 From 794611a1dfcb055d7d41ce133378dd8197d73e38 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Tue, 18 Jun 2013 18:53:53 +0800 Subject: cgroup: make serial_nr_cursor available throughout cgroup.c The next patch will use it to determine if a cgroup is newly created while we're iterating the cgroup hierarchy. tj: Rephrased the comment on top of cgroup_serial_nr_cursor. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cgroup.c | 18 ++++++++++-------- 1 file changed, 10 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 1d4f471de8d5..e6571ca822a0 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -198,6 +198,15 @@ static DEFINE_IDR(cgroup_hierarchy_idr); static struct cgroup_name root_cgroup_name = { .name = "/" }; +/* + * Assign a monotonically increasing serial number to cgroups. It + * guarantees cgroups with bigger numbers are newer than those with smaller + * numbers. Also, as cgroups are always appended to the parent's + * ->children list, it guarantees that sibling cgroups are always sorted in + * the ascending serial number order on the list. + */ +static atomic64_t cgroup_serial_nr_cursor = ATOMIC64_INIT(0); + /* This flag indicates whether tasks in the fork and exit paths should * check for fork/exit handlers to call. This avoids us having to do * extra work in the fork/exit path if none of the subsystems need to @@ -4222,7 +4231,6 @@ static void offline_css(struct cgroup_subsys *ss, struct cgroup *cgrp) static long cgroup_create(struct cgroup *parent, struct dentry *dentry, umode_t mode) { - static atomic64_t serial_nr_cursor = ATOMIC64_INIT(0); struct cgroup *cgrp; struct cgroup_name *name; struct cgroupfs_root *root = parent->root; @@ -4309,13 +4317,7 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, goto err_free_all; lockdep_assert_held(&dentry->d_inode->i_mutex); - /* - * Assign a monotonically increasing serial number. With the list - * appending below, it guarantees that sibling cgroups are always - * sorted in the ascending serial number order on the parent's - * ->children. - */ - cgrp->serial_nr = atomic64_inc_return(&serial_nr_cursor); + cgrp->serial_nr = atomic64_inc_return(&cgroup_serial_nr_cursor); /* allocation complete, commit to creation */ list_add_tail(&cgrp->allcg_node, &root->allcg_list); -- cgit v1.2.3 From e8c82d20a9f729cf4b9f73043f7fd4e0872bebfd Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Tue, 18 Jun 2013 18:48:37 +0800 Subject: cgroup: convert cgroup_cft_commit() to use cgroup_for_each_descendant_pre() We used root->allcg_list to iterate cgroup hierarchy because at that time cgroup_for_each_descendant_pre() hasn't been invented. tj: In cgroup_cfts_commit(), s/@serial_nr/@update_upto/, move the assignment right above releasing cgroup_mutex and explain what's going on there. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cgroup.c | 80 +++++++++++++++++++++++++++++++-------------------------- 1 file changed, 44 insertions(+), 36 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index e6571ca822a0..0ed7d8db6508 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1399,7 +1399,6 @@ static void init_cgroup_housekeeping(struct cgroup *cgrp) INIT_LIST_HEAD(&cgrp->children); INIT_LIST_HEAD(&cgrp->files); INIT_LIST_HEAD(&cgrp->cset_links); - INIT_LIST_HEAD(&cgrp->allcg_node); INIT_LIST_HEAD(&cgrp->release_list); INIT_LIST_HEAD(&cgrp->pidlists); mutex_init(&cgrp->pidlist_mutex); @@ -1414,12 +1413,10 @@ static void init_cgroup_root(struct cgroupfs_root *root) INIT_LIST_HEAD(&root->subsys_list); INIT_LIST_HEAD(&root->root_list); - INIT_LIST_HEAD(&root->allcg_list); root->number_of_cgroups = 1; cgrp->root = root; cgrp->name = &root_cgroup_name; init_cgroup_housekeeping(cgrp); - list_add_tail(&cgrp->allcg_node, &root->allcg_list); } static int cgroup_init_root_id(struct cgroupfs_root *root) @@ -2785,65 +2782,78 @@ static int cgroup_addrm_files(struct cgroup *cgrp, struct cgroup_subsys *subsys, return ret; } -static DEFINE_MUTEX(cgroup_cft_mutex); - static void cgroup_cfts_prepare(void) - __acquires(&cgroup_cft_mutex) __acquires(&cgroup_mutex) + __acquires(&cgroup_mutex) { /* * Thanks to the entanglement with vfs inode locking, we can't walk * the existing cgroups under cgroup_mutex and create files. - * Instead, we increment reference on all cgroups and build list of - * them using @cgrp->cft_q_node. Grab cgroup_cft_mutex to ensure - * exclusive access to the field. + * Instead, we use cgroup_for_each_descendant_pre() and drop RCU + * read lock before calling cgroup_addrm_files(). */ - mutex_lock(&cgroup_cft_mutex); mutex_lock(&cgroup_mutex); } static void cgroup_cfts_commit(struct cgroup_subsys *ss, struct cftype *cfts, bool is_add) - __releases(&cgroup_mutex) __releases(&cgroup_cft_mutex) + __releases(&cgroup_mutex) { LIST_HEAD(pending); - struct cgroup *cgrp, *n; + struct cgroup *cgrp, *root = &ss->root->top_cgroup; struct super_block *sb = ss->root->sb; + struct dentry *prev = NULL; + struct inode *inode; + u64 update_upto; /* %NULL @cfts indicates abort and don't bother if @ss isn't attached */ - if (cfts && ss->root != &rootnode && - atomic_inc_not_zero(sb->s_active)) { - list_for_each_entry(cgrp, &ss->root->allcg_list, allcg_node) { - dget(cgrp->dentry); - list_add_tail(&cgrp->cft_q_node, &pending); - } - } else { - sb = NULL; + if (!cfts || ss->root == &rootnode || + !atomic_inc_not_zero(&sb->s_active)) { + mutex_unlock(&cgroup_mutex); + return; } - mutex_unlock(&cgroup_mutex); - /* - * All new cgroups will see @cfts update on @ss->cftsets. Add/rm - * files for all cgroups which were created before. + * All cgroups which are created after we drop cgroup_mutex will + * have the updated set of files, so we only need to update the + * cgroups created before the current @cgroup_serial_nr_cursor. */ - list_for_each_entry_safe(cgrp, n, &pending, cft_q_node) { - struct inode *inode = cgrp->dentry->d_inode; + update_upto = atomic64_read(&cgroup_serial_nr_cursor); + + mutex_unlock(&cgroup_mutex); + + /* @root always needs to be updated */ + inode = root->dentry->d_inode; + mutex_lock(&inode->i_mutex); + mutex_lock(&cgroup_mutex); + cgroup_addrm_files(root, ss, cfts, is_add); + mutex_unlock(&cgroup_mutex); + mutex_unlock(&inode->i_mutex); + + /* add/rm files for all cgroups created before */ + rcu_read_lock(); + cgroup_for_each_descendant_pre(cgrp, root) { + if (cgroup_is_dead(cgrp)) + continue; + + inode = cgrp->dentry->d_inode; + dget(cgrp->dentry); + rcu_read_unlock(); + + dput(prev); + prev = cgrp->dentry; mutex_lock(&inode->i_mutex); mutex_lock(&cgroup_mutex); - if (!cgroup_is_dead(cgrp)) + if (cgrp->serial_nr <= update_upto && !cgroup_is_dead(cgrp)) cgroup_addrm_files(cgrp, ss, cfts, is_add); mutex_unlock(&cgroup_mutex); mutex_unlock(&inode->i_mutex); - list_del_init(&cgrp->cft_q_node); - dput(cgrp->dentry); + rcu_read_lock(); } - - if (sb) - deactivate_super(sb); - - mutex_unlock(&cgroup_cft_mutex); + rcu_read_unlock(); + dput(prev); + deactivate_super(sb); } /** @@ -4320,7 +4330,6 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, cgrp->serial_nr = atomic64_inc_return(&cgroup_serial_nr_cursor); /* allocation complete, commit to creation */ - list_add_tail(&cgrp->allcg_node, &root->allcg_list); list_add_tail_rcu(&cgrp->sibling, &cgrp->parent->children); root->number_of_cgroups++; @@ -4559,7 +4568,6 @@ static void cgroup_offline_fn(struct work_struct *work) /* delete this cgroup from parent->children */ list_del_rcu(&cgrp->sibling); - list_del_init(&cgrp->allcg_node); dput(d); -- cgit v1.2.3 From 00356bd5f0f5e04183fb15805eb29e97c2fc20ac Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 18 Jun 2013 11:14:22 -0700 Subject: cgroup: clean up cgroup_serial_nr_cursor cgroup_serial_nr_cursor was created atomic64_t because I thought it was never gonna used for anything other than assigning unique numbers to cgroups and didn't want to worry about synchronization; however, now we're using it as an event-stamp to distinguish cgroups created before and after certain point which assumes that it's protected by cgroup_mutex. Let's make it clear by making it a u64. Also, rename it to cgroup_serial_nr_next and make it point to the next nr to allocate so that where it's pointing to is clear and more conventional. Signed-off-by: Tejun Heo Cc: Li Zefan --- kernel/cgroup.c | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 0ed7d8db6508..65f333ebb572 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -203,9 +203,10 @@ static struct cgroup_name root_cgroup_name = { .name = "/" }; * guarantees cgroups with bigger numbers are newer than those with smaller * numbers. Also, as cgroups are always appended to the parent's * ->children list, it guarantees that sibling cgroups are always sorted in - * the ascending serial number order on the list. + * the ascending serial number order on the list. Protected by + * cgroup_mutex. */ -static atomic64_t cgroup_serial_nr_cursor = ATOMIC64_INIT(0); +static u64 cgroup_serial_nr_next = 1; /* This flag indicates whether tasks in the fork and exit paths should * check for fork/exit handlers to call. This avoids us having to do @@ -2803,7 +2804,7 @@ static void cgroup_cfts_commit(struct cgroup_subsys *ss, struct super_block *sb = ss->root->sb; struct dentry *prev = NULL; struct inode *inode; - u64 update_upto; + u64 update_before; /* %NULL @cfts indicates abort and don't bother if @ss isn't attached */ if (!cfts || ss->root == &rootnode || @@ -2815,9 +2816,9 @@ static void cgroup_cfts_commit(struct cgroup_subsys *ss, /* * All cgroups which are created after we drop cgroup_mutex will * have the updated set of files, so we only need to update the - * cgroups created before the current @cgroup_serial_nr_cursor. + * cgroups created before the current @cgroup_serial_nr_next. */ - update_upto = atomic64_read(&cgroup_serial_nr_cursor); + update_before = cgroup_serial_nr_next; mutex_unlock(&cgroup_mutex); @@ -2844,7 +2845,7 @@ static void cgroup_cfts_commit(struct cgroup_subsys *ss, mutex_lock(&inode->i_mutex); mutex_lock(&cgroup_mutex); - if (cgrp->serial_nr <= update_upto && !cgroup_is_dead(cgrp)) + if (cgrp->serial_nr < update_before && !cgroup_is_dead(cgrp)) cgroup_addrm_files(cgrp, ss, cfts, is_add); mutex_unlock(&cgroup_mutex); mutex_unlock(&inode->i_mutex); @@ -4327,7 +4328,7 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, goto err_free_all; lockdep_assert_held(&dentry->d_inode->i_mutex); - cgrp->serial_nr = atomic64_inc_return(&cgroup_serial_nr_cursor); + cgrp->serial_nr = cgroup_serial_nr_next++; /* allocation complete, commit to creation */ list_add_tail_rcu(&cgrp->sibling, &cgrp->parent->children); -- cgit v1.2.3 From 03c78cbebb323fc97295ff97dc5e009d56371d57 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 14 Jun 2013 11:17:19 +0800 Subject: cgroup: rename cont to cgrp Cont is short for container. control group was named process container at first, but then people found container already has a meaning in linux kernel. Clean up the leftover variable name @cont. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cgroup.c | 22 +++++++++++----------- 1 file changed, 11 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 65f333ebb572..1051c1f69674 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -5515,7 +5515,7 @@ struct cgroup_subsys_state *cgroup_css_from_dir(struct file *f, int id) } #ifdef CONFIG_CGROUP_DEBUG -static struct cgroup_subsys_state *debug_css_alloc(struct cgroup *cont) +static struct cgroup_subsys_state *debug_css_alloc(struct cgroup *cgrp) { struct cgroup_subsys_state *css = kzalloc(sizeof(*css), GFP_KERNEL); @@ -5525,23 +5525,23 @@ static struct cgroup_subsys_state *debug_css_alloc(struct cgroup *cont) return css; } -static void debug_css_free(struct cgroup *cont) +static void debug_css_free(struct cgroup *cgrp) { - kfree(cont->subsys[debug_subsys_id]); + kfree(cgrp->subsys[debug_subsys_id]); } -static u64 debug_taskcount_read(struct cgroup *cont, struct cftype *cft) +static u64 debug_taskcount_read(struct cgroup *cgrp, struct cftype *cft) { - return cgroup_task_count(cont); + return cgroup_task_count(cgrp); } -static u64 current_css_set_read(struct cgroup *cont, struct cftype *cft) +static u64 current_css_set_read(struct cgroup *cgrp, struct cftype *cft) { return (u64)(unsigned long)current->cgroups; } -static u64 current_css_set_refcount_read(struct cgroup *cont, - struct cftype *cft) +static u64 current_css_set_refcount_read(struct cgroup *cgrp, + struct cftype *cft) { u64 count; @@ -5551,7 +5551,7 @@ static u64 current_css_set_refcount_read(struct cgroup *cont, return count; } -static int current_css_set_cg_links_read(struct cgroup *cont, +static int current_css_set_cg_links_read(struct cgroup *cgrp, struct cftype *cft, struct seq_file *seq) { @@ -5578,14 +5578,14 @@ static int current_css_set_cg_links_read(struct cgroup *cont, } #define MAX_TASKS_SHOWN_PER_CSS 25 -static int cgroup_css_links_read(struct cgroup *cont, +static int cgroup_css_links_read(struct cgroup *cgrp, struct cftype *cft, struct seq_file *seq) { struct cgrp_cset_link *link; read_lock(&css_set_lock); - list_for_each_entry(link, &cont->cset_links, cset_link) { + list_for_each_entry(link, &cgrp->cset_links, cset_link) { struct css_set *cset = link->cset; struct task_struct *task; int count = 0; -- cgit v1.2.3 From 9bb5d40cd93c9dd4be74834b1dcb1ba03629716b Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 4 Jun 2013 10:44:21 +0200 Subject: perf: Fix mmap() accounting hole Vince's fuzzer once again found holes. This time it spotted a leak in the locked page accounting. When an event had redirected output and its close() was the last reference to the buffer we didn't have a vm context to undo accounting. Change the code to destroy the buffer on the last munmap() and detach all redirected events at that time. This provides us the right context to undo the vm accounting. Reported-and-tested-by: Vince Weaver Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/20130604084421.GI8923@twins.programming.kicks-ass.net Cc: Signed-off-by: Ingo Molnar --- kernel/events/core.c | 228 ++++++++++++++++++++++++++++++++--------------- kernel/events/internal.h | 3 +- 2 files changed, 159 insertions(+), 72 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index ae752cd4a086..b391907d5352 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -196,9 +196,6 @@ static void cpu_ctx_sched_in(struct perf_cpu_context *cpuctx, static void update_context_time(struct perf_event_context *ctx); static u64 perf_event_time(struct perf_event *event); -static void ring_buffer_attach(struct perf_event *event, - struct ring_buffer *rb); - void __weak perf_event_print_debug(void) { } extern __weak const char *perf_pmu_name(void) @@ -2917,7 +2914,8 @@ static void free_event_rcu(struct rcu_head *head) kfree(event); } -static bool ring_buffer_put(struct ring_buffer *rb); +static void ring_buffer_put(struct ring_buffer *rb); +static void ring_buffer_detach(struct perf_event *event, struct ring_buffer *rb); static void free_event(struct perf_event *event) { @@ -2942,15 +2940,30 @@ static void free_event(struct perf_event *event) if (has_branch_stack(event)) { static_key_slow_dec_deferred(&perf_sched_events); /* is system-wide event */ - if (!(event->attach_state & PERF_ATTACH_TASK)) + if (!(event->attach_state & PERF_ATTACH_TASK)) { atomic_dec(&per_cpu(perf_branch_stack_events, event->cpu)); + } } } if (event->rb) { - ring_buffer_put(event->rb); - event->rb = NULL; + struct ring_buffer *rb; + + /* + * Can happen when we close an event with re-directed output. + * + * Since we have a 0 refcount, perf_mmap_close() will skip + * over us; possibly making our ring_buffer_put() the last. + */ + mutex_lock(&event->mmap_mutex); + rb = event->rb; + if (rb) { + rcu_assign_pointer(event->rb, NULL); + ring_buffer_detach(event, rb); + ring_buffer_put(rb); /* could be last */ + } + mutex_unlock(&event->mmap_mutex); } if (is_cgroup_event(event)) @@ -3188,30 +3201,13 @@ static unsigned int perf_poll(struct file *file, poll_table *wait) unsigned int events = POLL_HUP; /* - * Race between perf_event_set_output() and perf_poll(): perf_poll() - * grabs the rb reference but perf_event_set_output() overrides it. - * Here is the timeline for two threads T1, T2: - * t0: T1, rb = rcu_dereference(event->rb) - * t1: T2, old_rb = event->rb - * t2: T2, event->rb = new rb - * t3: T2, ring_buffer_detach(old_rb) - * t4: T1, ring_buffer_attach(rb1) - * t5: T1, poll_wait(event->waitq) - * - * To avoid this problem, we grab mmap_mutex in perf_poll() - * thereby ensuring that the assignment of the new ring buffer - * and the detachment of the old buffer appear atomic to perf_poll() + * Pin the event->rb by taking event->mmap_mutex; otherwise + * perf_event_set_output() can swizzle our rb and make us miss wakeups. */ mutex_lock(&event->mmap_mutex); - - rcu_read_lock(); - rb = rcu_dereference(event->rb); - if (rb) { - ring_buffer_attach(event, rb); + rb = event->rb; + if (rb) events = atomic_xchg(&rb->poll, 0); - } - rcu_read_unlock(); - mutex_unlock(&event->mmap_mutex); poll_wait(file, &event->waitq, wait); @@ -3521,16 +3517,12 @@ static void ring_buffer_attach(struct perf_event *event, return; spin_lock_irqsave(&rb->event_lock, flags); - if (!list_empty(&event->rb_entry)) - goto unlock; - - list_add(&event->rb_entry, &rb->event_list); -unlock: + if (list_empty(&event->rb_entry)) + list_add(&event->rb_entry, &rb->event_list); spin_unlock_irqrestore(&rb->event_lock, flags); } -static void ring_buffer_detach(struct perf_event *event, - struct ring_buffer *rb) +static void ring_buffer_detach(struct perf_event *event, struct ring_buffer *rb) { unsigned long flags; @@ -3549,13 +3541,10 @@ static void ring_buffer_wakeup(struct perf_event *event) rcu_read_lock(); rb = rcu_dereference(event->rb); - if (!rb) - goto unlock; - - list_for_each_entry_rcu(event, &rb->event_list, rb_entry) - wake_up_all(&event->waitq); - -unlock: + if (rb) { + list_for_each_entry_rcu(event, &rb->event_list, rb_entry) + wake_up_all(&event->waitq); + } rcu_read_unlock(); } @@ -3582,23 +3571,14 @@ static struct ring_buffer *ring_buffer_get(struct perf_event *event) return rb; } -static bool ring_buffer_put(struct ring_buffer *rb) +static void ring_buffer_put(struct ring_buffer *rb) { - struct perf_event *event, *n; - unsigned long flags; - if (!atomic_dec_and_test(&rb->refcount)) - return false; + return; - spin_lock_irqsave(&rb->event_lock, flags); - list_for_each_entry_safe(event, n, &rb->event_list, rb_entry) { - list_del_init(&event->rb_entry); - wake_up_all(&event->waitq); - } - spin_unlock_irqrestore(&rb->event_lock, flags); + WARN_ON_ONCE(!list_empty(&rb->event_list)); call_rcu(&rb->rcu_head, rb_free_rcu); - return true; } static void perf_mmap_open(struct vm_area_struct *vma) @@ -3606,28 +3586,100 @@ static void perf_mmap_open(struct vm_area_struct *vma) struct perf_event *event = vma->vm_file->private_data; atomic_inc(&event->mmap_count); + atomic_inc(&event->rb->mmap_count); } +/* + * A buffer can be mmap()ed multiple times; either directly through the same + * event, or through other events by use of perf_event_set_output(). + * + * In order to undo the VM accounting done by perf_mmap() we need to destroy + * the buffer here, where we still have a VM context. This means we need + * to detach all events redirecting to us. + */ static void perf_mmap_close(struct vm_area_struct *vma) { struct perf_event *event = vma->vm_file->private_data; - if (atomic_dec_and_mutex_lock(&event->mmap_count, &event->mmap_mutex)) { - struct ring_buffer *rb = event->rb; - struct user_struct *mmap_user = rb->mmap_user; - int mmap_locked = rb->mmap_locked; - unsigned long size = perf_data_size(rb); + struct ring_buffer *rb = event->rb; + struct user_struct *mmap_user = rb->mmap_user; + int mmap_locked = rb->mmap_locked; + unsigned long size = perf_data_size(rb); - rcu_assign_pointer(event->rb, NULL); - ring_buffer_detach(event, rb); - mutex_unlock(&event->mmap_mutex); + atomic_dec(&rb->mmap_count); + + if (!atomic_dec_and_mutex_lock(&event->mmap_count, &event->mmap_mutex)) + return; + + /* Detach current event from the buffer. */ + rcu_assign_pointer(event->rb, NULL); + ring_buffer_detach(event, rb); + mutex_unlock(&event->mmap_mutex); + + /* If there's still other mmap()s of this buffer, we're done. */ + if (atomic_read(&rb->mmap_count)) { + ring_buffer_put(rb); /* can't be last */ + return; + } - if (ring_buffer_put(rb)) { - atomic_long_sub((size >> PAGE_SHIFT) + 1, &mmap_user->locked_vm); - vma->vm_mm->pinned_vm -= mmap_locked; - free_uid(mmap_user); + /* + * No other mmap()s, detach from all other events that might redirect + * into the now unreachable buffer. Somewhat complicated by the + * fact that rb::event_lock otherwise nests inside mmap_mutex. + */ +again: + rcu_read_lock(); + list_for_each_entry_rcu(event, &rb->event_list, rb_entry) { + if (!atomic_long_inc_not_zero(&event->refcount)) { + /* + * This event is en-route to free_event() which will + * detach it and remove it from the list. + */ + continue; } + rcu_read_unlock(); + + mutex_lock(&event->mmap_mutex); + /* + * Check we didn't race with perf_event_set_output() which can + * swizzle the rb from under us while we were waiting to + * acquire mmap_mutex. + * + * If we find a different rb; ignore this event, a next + * iteration will no longer find it on the list. We have to + * still restart the iteration to make sure we're not now + * iterating the wrong list. + */ + if (event->rb == rb) { + rcu_assign_pointer(event->rb, NULL); + ring_buffer_detach(event, rb); + ring_buffer_put(rb); /* can't be last, we still have one */ + } + mutex_unlock(&event->mmap_mutex); + put_event(event); + + /* + * Restart the iteration; either we're on the wrong list or + * destroyed its integrity by doing a deletion. + */ + goto again; } + rcu_read_unlock(); + + /* + * It could be there's still a few 0-ref events on the list; they'll + * get cleaned up by free_event() -- they'll also still have their + * ref on the rb and will free it whenever they are done with it. + * + * Aside from that, this buffer is 'fully' detached and unmapped, + * undo the VM accounting. + */ + + atomic_long_sub((size >> PAGE_SHIFT) + 1, &mmap_user->locked_vm); + vma->vm_mm->pinned_vm -= mmap_locked; + free_uid(mmap_user); + + ring_buffer_put(rb); /* could be last */ } static const struct vm_operations_struct perf_mmap_vmops = { @@ -3677,10 +3729,24 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma) return -EINVAL; WARN_ON_ONCE(event->ctx->parent_ctx); +again: mutex_lock(&event->mmap_mutex); if (event->rb) { - if (event->rb->nr_pages != nr_pages) + if (event->rb->nr_pages != nr_pages) { ret = -EINVAL; + goto unlock; + } + + if (!atomic_inc_not_zero(&event->rb->mmap_count)) { + /* + * Raced against perf_mmap_close() through + * perf_event_set_output(). Try again, hope for better + * luck. + */ + mutex_unlock(&event->mmap_mutex); + goto again; + } + goto unlock; } @@ -3722,12 +3788,14 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma) goto unlock; } + atomic_set(&rb->mmap_count, 1); rb->mmap_locked = extra; rb->mmap_user = get_current_user(); atomic_long_add(user_extra, &user->locked_vm); vma->vm_mm->pinned_vm += extra; + ring_buffer_attach(event, rb); rcu_assign_pointer(event->rb, rb); perf_event_update_userpage(event); @@ -3737,6 +3805,10 @@ unlock: atomic_inc(&event->mmap_count); mutex_unlock(&event->mmap_mutex); + /* + * Since pinned accounting is per vm we cannot allow fork() to copy our + * vma. + */ vma->vm_flags |= VM_DONTCOPY | VM_DONTEXPAND | VM_DONTDUMP; vma->vm_ops = &perf_mmap_vmops; @@ -6415,6 +6487,8 @@ set: if (atomic_read(&event->mmap_count)) goto unlock; + old_rb = event->rb; + if (output_event) { /* get the rb we want to redirect to */ rb = ring_buffer_get(output_event); @@ -6422,16 +6496,28 @@ set: goto unlock; } - old_rb = event->rb; - rcu_assign_pointer(event->rb, rb); if (old_rb) ring_buffer_detach(event, old_rb); + + if (rb) + ring_buffer_attach(event, rb); + + rcu_assign_pointer(event->rb, rb); + + if (old_rb) { + ring_buffer_put(old_rb); + /* + * Since we detached before setting the new rb, so that we + * could attach the new rb, we could have missed a wakeup. + * Provide it now. + */ + wake_up_all(&event->waitq); + } + ret = 0; unlock: mutex_unlock(&event->mmap_mutex); - if (old_rb) - ring_buffer_put(old_rb); out: return ret; } diff --git a/kernel/events/internal.h b/kernel/events/internal.h index 5bc6c8e9b851..ca6599723be5 100644 --- a/kernel/events/internal.h +++ b/kernel/events/internal.h @@ -31,7 +31,8 @@ struct ring_buffer { spinlock_t event_lock; struct list_head event_list; - int mmap_locked; + atomic_t mmap_count; + unsigned long mmap_locked; struct user_struct *mmap_user; struct perf_event_mmap_page *user_page; -- cgit v1.2.3 From e712209a9e0b70e78b13847738eb66fe37412515 Mon Sep 17 00:00:00 2001 From: Stephane Eranian Date: Thu, 6 Jun 2013 11:02:04 +0200 Subject: perf: Fix hypervisor branch sampling permission check Commit 2b923c8 perf/x86: Check branch sampling priv level in generic code was missing the check for the hypervisor (HV) priv level, so add it back. With this patch, we get the following correct behavior: # echo 2 >/proc/sys/kernel/perf_event_paranoid $ perf record -j any,k noploop 1 Error: You may not have permission to collect stats. Consider tweaking /proc/sys/kernel/perf_event_paranoid: -1 - Not paranoid at all 0 - Disallow raw tracepoint access for unpriv 1 - Disallow cpu events for unpriv 2 - Disallow kernel profiling for unpriv $ perf record -j any,hv noploop 1 Error: You may not have permission to collect stats. Consider tweaking /proc/sys/kernel/perf_event_paranoid: -1 - Not paranoid at all 0 - Disallow raw tracepoint access for unpriv 1 - Disallow cpu events for unpriv 2 - Disallow kernel profiling for unpriv Signed-off-by: Stephane Eranian Acked-by: Petr Matousek Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/20130606090204.GA3725@quad Signed-off-by: Ingo Molnar --- kernel/events/core.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index d0e0d0d2025f..aca95bce34c8 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -6573,8 +6573,8 @@ static int perf_copy_attr(struct perf_event_attr __user *uattr, */ attr->branch_sample_type = mask; } - /* kernel level capture: check permissions */ - if ((mask & PERF_SAMPLE_BRANCH_KERNEL) + /* privileged levels capture (kernel, hv): check permissions */ + if ((mask & PERF_SAMPLE_BRANCH_PERM_PLM) && perf_paranoid_kernel() && !capable(CAP_SYS_ADMIN)) return -EACCES; } -- cgit v1.2.3 From 03d8e80beb7db78a13c192431205b9c83f7e0cd1 Mon Sep 17 00:00:00 2001 From: Mischa Jonker Date: Tue, 4 Jun 2013 11:45:48 +0200 Subject: perf: Add const qualifier to perf_pmu_register's 'name' arg This allows us to use pdev->name for registering a PMU device. IMO the name is not supposed to be changed anyway. Signed-off-by: Mischa Jonker Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1370339148-5566-1-git-send-email-mjonker@synopsys.com Signed-off-by: Ingo Molnar --- kernel/events/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index aca95bce34c8..9c8920783317 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -6179,7 +6179,7 @@ free_dev: static struct lock_class_key cpuctx_mutex; static struct lock_class_key cpuctx_lock; -int perf_pmu_register(struct pmu *pmu, char *name, int type) +int perf_pmu_register(struct pmu *pmu, const char *name, int type) { int cpu, ret; -- cgit v1.2.3 From 873b4c65b519fd769940eb281f77848227d4e5c1 Mon Sep 17 00:00:00 2001 From: Vincent Guittot Date: Wed, 5 Jun 2013 10:13:11 +0200 Subject: sched: Fix clear NOHZ_BALANCE_KICK I have faced a sequence where the Idle Load Balance was sometime not triggered for a while on my platform, in the following scenario: CPU 0 and CPU 1 are running tasks and CPU 2 is idle CPU 1 kicks the Idle Load Balance CPU 1 selects CPU 2 as the new Idle Load Balancer CPU 2 sets NOHZ_BALANCE_KICK for CPU 2 CPU 2 sends a reschedule IPI to CPU 2 While CPU 3 wakes up, CPU 0 or CPU 1 migrates a waking up task A on CPU 2 CPU 2 finally wakes up, runs task A and discards the Idle Load Balance task A quickly goes back to sleep (before a tick occurs on CPU 2) CPU 2 goes back to idle with NOHZ_BALANCE_KICK set Whenever CPU 2 will be selected as the ILB, no reschedule IPI will be sent because NOHZ_BALANCE_KICK is already set and no Idle Load Balance will be performed. We must wait for the sched softirq to be raised on CPU 2 thanks to another part the kernel to come back to clear NOHZ_BALANCE_KICK. The proposed solution clears NOHZ_BALANCE_KICK in schedule_ipi if we can't raise the sched_softirq for the Idle Load Balance. Change since V1: - move the clear of NOHZ_BALANCE_KICK in got_nohz_idle_kick if the ILB can't run on this CPU (as suggested by Peter) Signed-off-by: Vincent Guittot Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1370419991-13870-1-git-send-email-vincent.guittot@linaro.org Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 21 +++++++++++++++++---- 1 file changed, 17 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 58453b8272fd..919bee68032b 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -633,7 +633,19 @@ void wake_up_nohz_cpu(int cpu) static inline bool got_nohz_idle_kick(void) { int cpu = smp_processor_id(); - return idle_cpu(cpu) && test_bit(NOHZ_BALANCE_KICK, nohz_flags(cpu)); + + if (!test_bit(NOHZ_BALANCE_KICK, nohz_flags(cpu))) + return false; + + if (idle_cpu(cpu) && !need_resched()) + return true; + + /* + * We can't run Idle Load Balance on this CPU for this time so we + * cancel it and clear NOHZ_BALANCE_KICK + */ + clear_bit(NOHZ_BALANCE_KICK, nohz_flags(cpu)); + return false; } #else /* CONFIG_NO_HZ_COMMON */ @@ -1393,8 +1405,9 @@ static void sched_ttwu_pending(void) void scheduler_ipi(void) { - if (llist_empty(&this_rq()->wake_list) && !got_nohz_idle_kick() - && !tick_nohz_full_cpu(smp_processor_id())) + if (llist_empty(&this_rq()->wake_list) + && !tick_nohz_full_cpu(smp_processor_id()) + && !got_nohz_idle_kick()) return; /* @@ -1417,7 +1430,7 @@ void scheduler_ipi(void) /* * Check if someone kicked us for doing the nohz idle load balance. */ - if (unlikely(got_nohz_idle_kick() && !need_resched())) { + if (unlikely(got_nohz_idle_kick())) { this_rq()->idle_balance = 1; raise_softirq_irqoff(SCHED_SOFTIRQ); } -- cgit v1.2.3 From 29bb9e5a75684106a37593ad75ec75ff8312731b Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 24 May 2013 15:23:40 -0400 Subject: tracing/context-tracking: Add preempt_schedule_context() for tracing Dave Jones hit the following bug report: =============================== [ INFO: suspicious RCU usage. ] 3.10.0-rc2+ #1 Not tainted ------------------------------- include/linux/rcupdate.h:771 rcu_read_lock() used illegally while idle! other info that might help us debug this: RCU used illegally from idle CPU! rcu_scheduler_active = 1, debug_locks = 0 RCU used illegally from extended quiescent state! 2 locks held by cc1/63645: #0: (&rq->lock){-.-.-.}, at: [] __schedule+0xed/0x9b0 #1: (rcu_read_lock){.+.+..}, at: [] cpuacct_charge+0x5/0x1f0 CPU: 1 PID: 63645 Comm: cc1 Not tainted 3.10.0-rc2+ #1 [loadavg: 40.57 27.55 13.39 25/277 64369] Hardware name: Gigabyte Technology Co., Ltd. GA-MA78GM-S2H/GA-MA78GM-S2H, BIOS F12a 04/23/2010 0000000000000000 ffff88010f78fcf8 ffffffff816ae383 ffff88010f78fd28 ffffffff810b698d ffff88011c092548 000000000023d073 ffff88011c092500 0000000000000001 ffff88010f78fd60 ffffffff8109d7c5 ffffffff8109d645 Call Trace: [] dump_stack+0x19/0x1b [] lockdep_rcu_suspicious+0xfd/0x130 [] cpuacct_charge+0x185/0x1f0 [] ? cpuacct_charge+0x5/0x1f0 [] update_curr+0xec/0x240 [] put_prev_task_fair+0x228/0x480 [] __schedule+0x161/0x9b0 [] preempt_schedule+0x51/0x80 [] ? __cond_resched_softirq+0x60/0x60 [] ? retint_careful+0x12/0x2e [] ftrace_ops_control_func+0x1dc/0x210 [] ftrace_call+0x5/0x2f [] ? retint_careful+0xb/0x2e [] ? schedule_user+0x5/0x70 [] ? schedule_user+0x5/0x70 [] ? retint_careful+0x12/0x2e ------------[ cut here ]------------ What happened was that the function tracer traced the schedule_user() code that tells RCU that the system is coming back from userspace, and to add the CPU back to the RCU monitoring. Because the function tracer does a preempt_disable/enable_notrace() calls the preempt_enable_notrace() checks the NEED_RESCHED flag. If it is set, then preempt_schedule() is called. But this is called before the user_exit() function can inform the kernel that the CPU is no longer in user mode and needs to be accounted for by RCU. The fix is to create a new preempt_schedule_context() that checks if the kernel is still in user mode and if so to switch it to kernel mode before calling schedule. It also switches back to user mode coming back from schedule in need be. The only user of this currently is the preempt_enable_notrace(), which is only used by the tracing subsystem. Signed-off-by: Steven Rostedt Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1369423420.6828.226.camel@gandalf.local.home Signed-off-by: Ingo Molnar --- kernel/context_tracking.c | 40 ++++++++++++++++++++++++++++++++++++++++ 1 file changed, 40 insertions(+) (limited to 'kernel') diff --git a/kernel/context_tracking.c b/kernel/context_tracking.c index 65349f07b878..66677003e223 100644 --- a/kernel/context_tracking.c +++ b/kernel/context_tracking.c @@ -71,6 +71,46 @@ void user_enter(void) local_irq_restore(flags); } +#ifdef CONFIG_PREEMPT +/** + * preempt_schedule_context - preempt_schedule called by tracing + * + * The tracing infrastructure uses preempt_enable_notrace to prevent + * recursion and tracing preempt enabling caused by the tracing + * infrastructure itself. But as tracing can happen in areas coming + * from userspace or just about to enter userspace, a preempt enable + * can occur before user_exit() is called. This will cause the scheduler + * to be called when the system is still in usermode. + * + * To prevent this, the preempt_enable_notrace will use this function + * instead of preempt_schedule() to exit user context if needed before + * calling the scheduler. + */ +void __sched notrace preempt_schedule_context(void) +{ + struct thread_info *ti = current_thread_info(); + enum ctx_state prev_ctx; + + if (likely(ti->preempt_count || irqs_disabled())) + return; + + /* + * Need to disable preemption in case user_exit() is traced + * and the tracer calls preempt_enable_notrace() causing + * an infinite recursion. + */ + preempt_disable_notrace(); + prev_ctx = exception_enter(); + preempt_enable_no_resched_notrace(); + + preempt_schedule(); + + preempt_disable_notrace(); + exception_exit(prev_ctx); + preempt_enable_notrace(); +} +EXPORT_SYMBOL_GPL(preempt_schedule_context); +#endif /* CONFIG_PREEMPT */ /** * user_exit - Inform the context tracking that the CPU is -- cgit v1.2.3 From e23ee74777f389369431d77390c4b09332ce026a Mon Sep 17 00:00:00 2001 From: Kirill Tkhai Date: Fri, 7 Jun 2013 15:37:43 -0400 Subject: sched/rt: Simplify pull_rt_task() logic and remove .leaf_rt_rq_list [ Peter, this is based off of some of my work, I ran it though a few tests and it passed. I also reviewed it, and added my SOB as I am somewhat a co-author to it. ] Based on the patch by Steven Rostedt from previous year: https://lkml.org/lkml/2012/4/18/517 1)Simplify pull_rt_task() logic: search in pushable tasks of dest runqueue. The only pullable tasks are the tasks which are pushable in their local rq, and no others. 2)Remove .leaf_rt_rq_list member of struct rt_rq and functions connected with it: nobody uses it since now. Signed-off-by: Kirill Tkhai Signed-off-by: Steven Rostedt Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/287571370557898@web7d.yandex.ru Signed-off-by: Ingo Molnar --- kernel/sched/rt.c | 82 ++++++++++------------------------------------------ kernel/sched/sched.h | 1 - 2 files changed, 16 insertions(+), 67 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c index 8d85f9ac4262..01970c8e64df 100644 --- a/kernel/sched/rt.c +++ b/kernel/sched/rt.c @@ -399,20 +399,6 @@ static inline struct task_group *next_task_group(struct task_group *tg) (iter = next_task_group(iter)) && \ (rt_rq = iter->rt_rq[cpu_of(rq)]);) -static inline void list_add_leaf_rt_rq(struct rt_rq *rt_rq) -{ - list_add_rcu(&rt_rq->leaf_rt_rq_list, - &rq_of_rt_rq(rt_rq)->leaf_rt_rq_list); -} - -static inline void list_del_leaf_rt_rq(struct rt_rq *rt_rq) -{ - list_del_rcu(&rt_rq->leaf_rt_rq_list); -} - -#define for_each_leaf_rt_rq(rt_rq, rq) \ - list_for_each_entry_rcu(rt_rq, &rq->leaf_rt_rq_list, leaf_rt_rq_list) - #define for_each_sched_rt_entity(rt_se) \ for (; rt_se; rt_se = rt_se->parent) @@ -509,17 +495,6 @@ typedef struct rt_rq *rt_rq_iter_t; #define for_each_rt_rq(rt_rq, iter, rq) \ for ((void) iter, rt_rq = &rq->rt; rt_rq; rt_rq = NULL) -static inline void list_add_leaf_rt_rq(struct rt_rq *rt_rq) -{ -} - -static inline void list_del_leaf_rt_rq(struct rt_rq *rt_rq) -{ -} - -#define for_each_leaf_rt_rq(rt_rq, rq) \ - for (rt_rq = &rq->rt; rt_rq; rt_rq = NULL) - #define for_each_sched_rt_entity(rt_se) \ for (; rt_se; rt_se = NULL) @@ -1066,9 +1041,6 @@ static void __enqueue_rt_entity(struct sched_rt_entity *rt_se, bool head) if (group_rq && (rt_rq_throttled(group_rq) || !group_rq->rt_nr_running)) return; - if (!rt_rq->rt_nr_running) - list_add_leaf_rt_rq(rt_rq); - if (head) list_add(&rt_se->run_list, queue); else @@ -1088,8 +1060,6 @@ static void __dequeue_rt_entity(struct sched_rt_entity *rt_se) __clear_bit(rt_se_prio(rt_se), array->bitmap); dec_rt_tasks(rt_se, rt_rq); - if (!rt_rq->rt_nr_running) - list_del_leaf_rt_rq(rt_rq); } /* @@ -1394,42 +1364,24 @@ static int pick_rt_task(struct rq *rq, struct task_struct *p, int cpu) return 0; } -/* Return the second highest RT task, NULL otherwise */ -static struct task_struct *pick_next_highest_task_rt(struct rq *rq, int cpu) +/* + * Return the highest pushable rq's task, which is suitable to be executed + * on the cpu, NULL otherwise + */ +static struct task_struct *pick_highest_pushable_task(struct rq *rq, int cpu) { - struct task_struct *next = NULL; - struct sched_rt_entity *rt_se; - struct rt_prio_array *array; - struct rt_rq *rt_rq; - int idx; - - for_each_leaf_rt_rq(rt_rq, rq) { - array = &rt_rq->active; - idx = sched_find_first_bit(array->bitmap); -next_idx: - if (idx >= MAX_RT_PRIO) - continue; - if (next && next->prio <= idx) - continue; - list_for_each_entry(rt_se, array->queue + idx, run_list) { - struct task_struct *p; + struct plist_head *head = &rq->rt.pushable_tasks; + struct task_struct *p; - if (!rt_entity_is_task(rt_se)) - continue; + if (!has_pushable_tasks(rq)) + return NULL; - p = rt_task_of(rt_se); - if (pick_rt_task(rq, p, cpu)) { - next = p; - break; - } - } - if (!next) { - idx = find_next_bit(array->bitmap, MAX_RT_PRIO, idx+1); - goto next_idx; - } + plist_for_each_entry(p, head, pushable_tasks) { + if (pick_rt_task(rq, p, cpu)) + return p; } - return next; + return NULL; } static DEFINE_PER_CPU(cpumask_var_t, local_cpu_mask); @@ -1703,12 +1655,10 @@ static int pull_rt_task(struct rq *this_rq) double_lock_balance(this_rq, src_rq); /* - * Are there still pullable RT tasks? + * We can pull only a task, which is pushable + * on its rq, and no others. */ - if (src_rq->rt.rt_nr_running <= 1) - goto skip; - - p = pick_next_highest_task_rt(src_rq, this_cpu); + p = pick_highest_pushable_task(src_rq, this_cpu); /* * Do we have an RT task that preempts diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index 74ff659e964f..029601a61587 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -361,7 +361,6 @@ struct rt_rq { unsigned long rt_nr_boosted; struct rq *rq; - struct list_head leaf_rt_rq_list; struct task_group *tg; #endif }; -- cgit v1.2.3 From 22b958d8cc5127d22d2ad2141277d312d93fad6c Mon Sep 17 00:00:00 2001 From: Michael Wang Date: Tue, 4 Jun 2013 14:23:39 +0800 Subject: sched: Refine the code in unthrottle_cfs_rq() Directly use rq to save some code. Signed-off-by: Michael Wang Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/51AD87EB.1070605@linux.vnet.ibm.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 143dcdbc47af..47a30be1fe83 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -2315,7 +2315,7 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq) int enqueue = 1; long task_delta; - se = cfs_rq->tg->se[cpu_of(rq_of(cfs_rq))]; + se = cfs_rq->tg->se[cpu_of(rq)]; cfs_rq->throttled = 0; -- cgit v1.2.3 From 8404c90d050733b3404dc36c500f63ccb0c972ce Mon Sep 17 00:00:00 2001 From: Michael Wang Date: Tue, 4 Jun 2013 14:24:08 +0800 Subject: sched: Femove the useless declaration in kernel/sched/fair.c default_cfs_period(), do_sched_cfs_period_timer(), do_sched_cfs_slack_timer() already defined previously, no need to declare again. Signed-off-by: Michael Wang Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/51AD8808.7020608@linux.vnet.ibm.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 4 ---- 1 file changed, 4 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 47a30be1fe83..c0ac2c3b56e1 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -2618,10 +2618,6 @@ static void check_cfs_rq_runtime(struct cfs_rq *cfs_rq) throttle_cfs_rq(cfs_rq); } -static inline u64 default_cfs_period(void); -static int do_sched_cfs_period_timer(struct cfs_bandwidth *cfs_b, int overrun); -static void do_sched_cfs_slack_timer(struct cfs_bandwidth *cfs_b); - static enum hrtimer_restart sched_cfs_slack_timer(struct hrtimer *timer) { struct cfs_bandwidth *cfs_b = -- cgit v1.2.3 From 0a0fca9d832b704f116a25badd1ca8c16771dcac Mon Sep 17 00:00:00 2001 From: Viresh Kumar Date: Tue, 4 Jun 2013 13:10:24 +0530 Subject: sched: Rename sched.c as sched/core.c in comments and Documentation Most of the stuff from kernel/sched.c was moved to kernel/sched/core.c long time back and the comments/Documentation never got updated. I figured it out when I was going through sched-domains.txt and so thought of fixing it globally. I haven't crossed check if the stuff that is referenced in sched/core.c by all these files is still present and hasn't changed as that wasn't the motive behind this patch. Signed-off-by: Viresh Kumar Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/cdff76a265326ab8d71922a1db5be599f20aad45.1370329560.git.viresh.kumar@linaro.org Signed-off-by: Ingo Molnar --- kernel/cpuset.c | 4 ++-- kernel/time.c | 2 +- kernel/workqueue_internal.h | 2 +- 3 files changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 64b3f791bbe5..902d13fc2b13 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -540,7 +540,7 @@ static void update_domain_attr_tree(struct sched_domain_attr *dattr, * This function builds a partial partition of the systems CPUs * A 'partial partition' is a set of non-overlapping subsets whose * union is a subset of that set. - * The output of this function needs to be passed to kernel/sched.c + * The output of this function needs to be passed to kernel/sched/core.c * partition_sched_domains() routine, which will rebuild the scheduler's * load balancing domains (sched domains) as specified by that partial * partition. @@ -569,7 +569,7 @@ static void update_domain_attr_tree(struct sched_domain_attr *dattr, * is a subset of one of these domains, while there are as * many such domains as possible, each as small as possible. * doms - Conversion of 'csa' to an array of cpumasks, for passing to - * the kernel/sched.c routine partition_sched_domains() in a + * the kernel/sched/core.c routine partition_sched_domains() in a * convenient format, that can be easily compared to the prior * value to determine what partition elements (sched domains) * were changed (added or removed.) diff --git a/kernel/time.c b/kernel/time.c index d3617dbd3dca..7c7964c33ae7 100644 --- a/kernel/time.c +++ b/kernel/time.c @@ -11,7 +11,7 @@ * Modification history kernel/time.c * * 1993-09-02 Philip Gladstone - * Created file with time related functions from sched.c and adjtimex() + * Created file with time related functions from sched/core.c and adjtimex() * 1993-10-08 Torsten Duwe * adjtime interface update and CMOS clock write code * 1995-08-13 Torsten Duwe diff --git a/kernel/workqueue_internal.h b/kernel/workqueue_internal.h index ad83c96b2ece..7e2204db0b1a 100644 --- a/kernel/workqueue_internal.h +++ b/kernel/workqueue_internal.h @@ -64,7 +64,7 @@ static inline struct worker *current_wq_worker(void) /* * Scheduler hooks for concurrency managed workqueue. Only to be used from - * sched.c and workqueue.c. + * sched/core.c and workqueue.c. */ void wq_worker_waking_up(struct task_struct *task, int cpu); struct task_struct *wq_worker_sleeping(struct task_struct *task, int cpu); -- cgit v1.2.3 From 4a850cbefa9592ddde3670a41c10c9576a657c43 Mon Sep 17 00:00:00 2001 From: Viresh Kumar Date: Tue, 4 Jun 2013 16:12:43 +0530 Subject: sched: Remove unused params of build_sched_domain() build_sched_domain() never uses parameter struct s_data *d and so passing it is useless. Remove it. Signed-off-by: Viresh Kumar Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/545e0b4536166a15b4475abcafe5ed0db4ad4a2c.1370436120.git.viresh.kumar@linaro.org Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index d8f071cc9f51..342e74419f8f 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5943,9 +5943,8 @@ static void __sdt_free(const struct cpumask *cpu_map) } struct sched_domain *build_sched_domain(struct sched_domain_topology_level *tl, - struct s_data *d, const struct cpumask *cpu_map, - struct sched_domain_attr *attr, struct sched_domain *child, - int cpu) + const struct cpumask *cpu_map, struct sched_domain_attr *attr, + struct sched_domain *child, int cpu) { struct sched_domain *sd = tl->init(tl, cpu); if (!sd) @@ -5985,7 +5984,7 @@ static int build_sched_domains(const struct cpumask *cpu_map, sd = NULL; for (tl = sched_domain_topology; tl->init; tl++) { - sd = build_sched_domain(tl, &d, cpu_map, attr, sd, i); + sd = build_sched_domain(tl, cpu_map, attr, sd, i); if (tl->flags & SDTL_OVERLAP || sched_feat(FORCE_SD_OVERLAP)) sd->flags |= SD_OVERLAP; if (cpumask_equal(cpu_map, sched_domain_span(sd))) -- cgit v1.2.3 From 22da956953f371c1ee7a578c31ed8c5702cb52b1 Mon Sep 17 00:00:00 2001 From: Viresh Kumar Date: Tue, 4 Jun 2013 15:41:15 +0530 Subject: sched: Optimize build_sched_domains() for saving first SD node for a cpu We are saving first scheduling domain for a cpu in build_sched_domains() by iterating over the nested sd->child list. We don't actually need to do it this way. tl will be equal to sched_domain_topology for the first iteration and so we can set *per_cpu_ptr(d.sd, i) based on that. So, save pointer to first SD while running the iteration loop over tl's. Signed-off-by: Viresh Kumar Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/fc473527cbc4dfa0b8eeef2a59db74684eb59a83.1370436120.git.viresh.kumar@linaro.org Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 342e74419f8f..137dcc03f66d 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5985,16 +5985,13 @@ static int build_sched_domains(const struct cpumask *cpu_map, sd = NULL; for (tl = sched_domain_topology; tl->init; tl++) { sd = build_sched_domain(tl, cpu_map, attr, sd, i); + if (tl == sched_domain_topology) + *per_cpu_ptr(d.sd, i) = sd; if (tl->flags & SDTL_OVERLAP || sched_feat(FORCE_SD_OVERLAP)) sd->flags |= SD_OVERLAP; if (cpumask_equal(cpu_map, sched_domain_span(sd))) break; } - - while (sd->child) - sd = sd->child; - - *per_cpu_ptr(d.sd, i) = sd; } /* Build the groups for the domains */ -- cgit v1.2.3 From 1c6321694074163b5863c13d71c19ca953a3fb08 Mon Sep 17 00:00:00 2001 From: Viresh Kumar Date: Mon, 10 Jun 2013 16:27:18 +0530 Subject: sched: Don't initialize alloc_state in build_sched_domains() alloc_state will be overwritten by __visit_domain_allocation_hell() and so we don't actually need to initialize alloc_state. Signed-off-by: Viresh Kumar Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/df57734a075cc5ad130e1ae498702e24f2529ab8.1370861520.git.viresh.kumar@linaro.org Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 137dcc03f66d..3de62649869b 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5969,7 +5969,7 @@ struct sched_domain *build_sched_domain(struct sched_domain_topology_level *tl, static int build_sched_domains(const struct cpumask *cpu_map, struct sched_domain_attr *attr) { - enum s_alloc alloc_state = sa_none; + enum s_alloc alloc_state; struct sched_domain *sd; struct s_data d; int i, ret = -ENOMEM; -- cgit v1.2.3 From c75e01288ce9c9a6b7beb6b23c07d2e4d1db8c84 Mon Sep 17 00:00:00 2001 From: Viresh Kumar Date: Mon, 10 Jun 2013 16:27:19 +0530 Subject: sched: Don't set sd->child to NULL when it is already NULL Memory for sd is allocated with kzalloc_node() which will initialize its fields with zero. In build_sched_domain() we are setting sd->child to child even if child is NULL, which isn't required. Lets do it only if child isn't NULL. Signed-off-by: Viresh Kumar Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/f4753a1730051341003ad2ad29a3229c7356678e.1370861520.git.viresh.kumar@linaro.org Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 3de62649869b..88c2c0ee5a52 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5955,8 +5955,8 @@ struct sched_domain *build_sched_domain(struct sched_domain_topology_level *tl, sd->level = child->level + 1; sched_domain_level_max = max(sched_domain_level_max, sd->level); child->parent = sd; + sd->child = child; } - sd->child = child; set_domain_attribute(sd, attr); return sd; -- cgit v1.2.3 From 27723a68caf05381b0b0bc6e127da2c9e7bcb775 Mon Sep 17 00:00:00 2001 From: Viresh Kumar Date: Mon, 10 Jun 2013 16:27:20 +0530 Subject: sched: Create for_each_sd_topology() For loop for traversing sched_domain_topology was used at multiple placed in core.c. This patch removes code redundancy by creating for_each_sd_topology(). Signed-off-by: Viresh Kumar Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/e0e04542f54e9464bd9da54f5ccfe62ec6c4c0bc.1370861520.git.viresh.kumar@linaro.org Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 88c2c0ee5a52..547b7d3ff893 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5565,6 +5565,9 @@ static struct sched_domain_topology_level default_topology[] = { static struct sched_domain_topology_level *sched_domain_topology = default_topology; +#define for_each_sd_topology(tl) \ + for (tl = sched_domain_topology; tl->init; tl++) + #ifdef CONFIG_NUMA static int sched_domains_numa_levels; @@ -5862,7 +5865,7 @@ static int __sdt_alloc(const struct cpumask *cpu_map) struct sched_domain_topology_level *tl; int j; - for (tl = sched_domain_topology; tl->init; tl++) { + for_each_sd_topology(tl) { struct sd_data *sdd = &tl->data; sdd->sd = alloc_percpu(struct sched_domain *); @@ -5915,7 +5918,7 @@ static void __sdt_free(const struct cpumask *cpu_map) struct sched_domain_topology_level *tl; int j; - for (tl = sched_domain_topology; tl->init; tl++) { + for_each_sd_topology(tl) { struct sd_data *sdd = &tl->data; for_each_cpu(j, cpu_map) { @@ -5983,7 +5986,7 @@ static int build_sched_domains(const struct cpumask *cpu_map, struct sched_domain_topology_level *tl; sd = NULL; - for (tl = sched_domain_topology; tl->init; tl++) { + for_each_sd_topology(tl) { sd = build_sched_domain(tl, cpu_map, attr, sd, i); if (tl == sched_domain_topology) *per_cpu_ptr(d.sd, i) = sd; -- cgit v1.2.3 From 0936629f01bb1b11772db8c36be421365238cbec Mon Sep 17 00:00:00 2001 From: Viresh Kumar Date: Tue, 11 Jun 2013 16:32:43 +0530 Subject: sched: Use cached value of span instead of calling sched_domain_span() In the beginning of build_sched_groups() we called sched_domain_span() and cached its return value in span. Few statements later we are calling it again to get the same pointer. Lets use the cached value instead as it hasn't changed in between. Signed-off-by: Viresh Kumar Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/834ecd507071ad88aff039352dbc7e063dd996a7.1370948150.git.viresh.kumar@linaro.org Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 547b7d3ff893..3388387e1330 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5347,7 +5347,7 @@ build_sched_groups(struct sched_domain *sd, int cpu) get_group(cpu, sdd, &sd->groups); atomic_inc(&sd->groups->ref); - if (cpu != cpumask_first(sched_domain_span(sd))) + if (cpu != cpumask_first(span)) return 0; lockdep_assert_held(&sched_domains_mutex); -- cgit v1.2.3 From cd08e9234c987766ad077bba80eb5a07d0855525 Mon Sep 17 00:00:00 2001 From: Viresh Kumar Date: Tue, 11 Jun 2013 16:32:44 +0530 Subject: sched: Fix memory leakage in build_sched_groups() In build_sched_groups() we don't need to call get_group() for cpus which are already covered in previous iterations. Calling get_group() would mark the group used and eventually leak it since we wouldn't connect it and not find it again to free it. This will happen only in cases where sg->cpumask contained more than one cpu (For any topology level). This patch would free sg's memory for all cpus leaving the group leader as the group isn't marked used now. Signed-off-by: Viresh Kumar Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/7a61e955abdcbb1dfa9fe493f11a5ec53a11ddd3.1370948150.git.viresh.kumar@linaro.org Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 3388387e1330..014c97f00732 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5357,12 +5357,12 @@ build_sched_groups(struct sched_domain *sd, int cpu) for_each_cpu(i, span) { struct sched_group *sg; - int group = get_group(i, sdd, &sg); - int j; + int group, j; if (cpumask_test_cpu(i, covered)) continue; + group = get_group(i, sdd, &sg); cpumask_clear(sched_group_cpus(sg)); sg->sgp->power = 0; cpumask_setall(sched_group_mask(sg)); -- cgit v1.2.3 From 94c95ba69f31e435416988ddb223c92e5b0e9e83 Mon Sep 17 00:00:00 2001 From: Viresh Kumar Date: Tue, 11 Jun 2013 16:32:45 +0530 Subject: sched: Remove WARN_ON(!sd) from init_sched_groups_power() sd can't be NULL in init_sched_groups_power() and so checking it for NULL isn't useful. In case it is required, then also we need to rearrange the code a bit as we already accessed invalid pointer sd to get sg: sg = sd->groups. Signed-off-by: Viresh Kumar Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/2bbe633cd74b431c05253a8ce61fdfd5066a531b.1370948150.git.viresh.kumar@linaro.org Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 014c97f00732..21b1403a10a2 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5400,7 +5400,7 @@ static void init_sched_groups_power(int cpu, struct sched_domain *sd) { struct sched_group *sg = sd->groups; - WARN_ON(!sd || !sg); + WARN_ON(!sg); do { sg->group_weight = cpumask_weight(sched_group_cpus(sg)); -- cgit v1.2.3 From be7002e6c613d22976f2b8d4bae6121a5fc0433a Mon Sep 17 00:00:00 2001 From: Joe Perches Date: Wed, 12 Jun 2013 11:55:36 -0700 Subject: sched: Don't mix use of typedef ctl_table and struct ctl_table Just use struct ctl_table. Signed-off-by: Joe Perches Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1371063336.2069.22.camel@joe-AO722 Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 21b1403a10a2..ceeaf0f45be0 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -4533,7 +4533,7 @@ sd_alloc_ctl_domain_table(struct sched_domain *sd) return table; } -static ctl_table *sd_alloc_ctl_cpu_table(int cpu) +static struct ctl_table *sd_alloc_ctl_cpu_table(int cpu) { struct ctl_table *entry, *table; struct sched_domain *sd; -- cgit v1.2.3 From e12d0271774fea9fddf1e2a7952a0bffb2ee8e8b Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 10 May 2013 17:12:28 -0400 Subject: nohz: Warn if the machine can not perform nohz_full If the user configures NO_HZ_FULL and defines nohz_full=XXX on the kernel command line, or enables NO_HZ_FULL_ALL, but nohz fails due to the machine having a unstable clock, warn about it. We do not want users thinking that they are getting the benefit of nohz when their machine can not support it. Signed-off-by: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Andrew Morton Cc: Thomas Gleixner Cc: H. Peter Anvin Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Signed-off-by: Frederic Weisbecker --- kernel/time/tick-sched.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'kernel') diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index f4208138fbf4..d87d22cb9bf2 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -178,6 +178,11 @@ static bool can_stop_full_tick(void) */ if (!sched_clock_stable) { trace_tick_stop(0, "unstable sched clock\n"); + /* + * Don't allow the user to think they can get + * full NO_HZ with this machine. + */ + WARN_ONCE(1, "NO_HZ FULL will not work with unstable sched clock"); return false; } #endif -- cgit v1.2.3 From b8900bc0217fac8e68085997bee2f05e6db931a2 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 6 Jun 2013 15:42:53 +0200 Subject: watchdog: Register / unregister watchdog kthreads on sysctl control The user activation/deactivation of the watchdog through boot parameters or systcl is currently implemented with a dance involving kthreads parking and unparking methods: the threads are unconditionally registered on boot and they park as soon as the user want the watchdog to be disabled. This method involves a few noisy details to handle though: the watchdog kthreads may be unparked anytime due to hotplug operations, after which the watchdog internals have to decide to park again if it is user-disabled. As a result the setup() and unpark() methods need to be able to request a reparking. This is not currently supported in the kthread infrastructure so this piece of the watchdog code only works halfway. Besides, unparking/reparking the watchdog kthreads consume unnecessary cputime on hotplug operations when those could be simply ignored in the first place. As suggested by Srivatsa, let's instead only register the watchdog threads when they are needed. This way we don't need to think about hotplug operations and we don't burden the CPU onlining when the watchdog is simply disabled. Suggested-by: Srivatsa S. Bhat Signed-off-by: Frederic Weisbecker Cc: Srivatsa S. Bhat Cc: Anish Singh Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Don Zickus --- kernel/watchdog.c | 87 ++++++++++++++++++++++++++++++------------------------- 1 file changed, 47 insertions(+), 40 deletions(-) (limited to 'kernel') diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 05039e348f07..52c9a9b91bdd 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -31,7 +31,7 @@ int watchdog_enabled = 1; int __read_mostly watchdog_thresh = 10; -static int __read_mostly watchdog_disabled; +static int __read_mostly watchdog_disabled = 1; static u64 __read_mostly sample_period; static DEFINE_PER_CPU(unsigned long, watchdog_touch_ts); @@ -347,11 +347,6 @@ static void watchdog_enable(unsigned int cpu) hrtimer_init(hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); hrtimer->function = watchdog_timer_fn; - if (!watchdog_enabled) { - kthread_park(current); - return; - } - /* Enable the perf event */ watchdog_nmi_enable(cpu); @@ -374,6 +369,11 @@ static void watchdog_disable(unsigned int cpu) watchdog_nmi_disable(cpu); } +static void watchdog_cleanup(unsigned int cpu, bool online) +{ + watchdog_disable(cpu); +} + static int watchdog_should_run(unsigned int cpu) { return __this_cpu_read(hrtimer_interrupts) != @@ -475,28 +475,40 @@ static int watchdog_nmi_enable(unsigned int cpu) { return 0; } static void watchdog_nmi_disable(unsigned int cpu) { return; } #endif /* CONFIG_HARDLOCKUP_DETECTOR */ -/* prepare/enable/disable routines */ -/* sysctl functions */ -#ifdef CONFIG_SYSCTL -static void watchdog_enable_all_cpus(void) +static struct smp_hotplug_thread watchdog_threads = { + .store = &softlockup_watchdog, + .thread_should_run = watchdog_should_run, + .thread_fn = watchdog, + .thread_comm = "watchdog/%u", + .setup = watchdog_enable, + .cleanup = watchdog_cleanup, + .park = watchdog_disable, + .unpark = watchdog_enable, +}; + +static int watchdog_enable_all_cpus(void) { - unsigned int cpu; + int err = 0; if (watchdog_disabled) { - watchdog_disabled = 0; - for_each_online_cpu(cpu) - kthread_unpark(per_cpu(softlockup_watchdog, cpu)); + err = smpboot_register_percpu_thread(&watchdog_threads); + if (err) + pr_err("Failed to create watchdog threads, disabled\n"); + else + watchdog_disabled = 0; } + + return err; } +/* prepare/enable/disable routines */ +/* sysctl functions */ +#ifdef CONFIG_SYSCTL static void watchdog_disable_all_cpus(void) { - unsigned int cpu; - if (!watchdog_disabled) { watchdog_disabled = 1; - for_each_online_cpu(cpu) - kthread_park(per_cpu(softlockup_watchdog, cpu)); + smpboot_unregister_percpu_thread(&watchdog_threads); } } @@ -507,14 +519,14 @@ static void watchdog_disable_all_cpus(void) int proc_dowatchdog(struct ctl_table *table, int write, void __user *buffer, size_t *lenp, loff_t *ppos) { - int ret; + int err, old_thresh, old_enabled; - if (watchdog_disabled < 0) - return -ENODEV; + old_thresh = ACCESS_ONCE(watchdog_thresh); + old_enabled = ACCESS_ONCE(watchdog_enabled); - ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos); - if (ret || !write) - return ret; + err = proc_dointvec_minmax(table, write, buffer, lenp, ppos); + if (err || !write) + return err; set_sample_period(); /* @@ -523,29 +535,24 @@ int proc_dowatchdog(struct ctl_table *table, int write, * watchdog_*_all_cpus() function takes care of this. */ if (watchdog_enabled && watchdog_thresh) - watchdog_enable_all_cpus(); + err = watchdog_enable_all_cpus(); else watchdog_disable_all_cpus(); - return ret; + /* Restore old values on failure */ + if (err) { + watchdog_thresh = old_thresh; + watchdog_enabled = old_enabled; + } + + return err; } #endif /* CONFIG_SYSCTL */ -static struct smp_hotplug_thread watchdog_threads = { - .store = &softlockup_watchdog, - .thread_should_run = watchdog_should_run, - .thread_fn = watchdog, - .thread_comm = "watchdog/%u", - .setup = watchdog_enable, - .park = watchdog_disable, - .unpark = watchdog_enable, -}; - void __init lockup_detector_init(void) { set_sample_period(); - if (smpboot_register_percpu_thread(&watchdog_threads)) { - pr_err("Failed to create watchdog threads, disabled\n"); - watchdog_disabled = -ENODEV; - } + + if (watchdog_enabled) + watchdog_enable_all_cpus(); } -- cgit v1.2.3 From 1a891cf19cdfb645827969cc6aeaeebdefeb87b2 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 12 Jun 2013 13:16:25 -0400 Subject: tracing: Add binary '&' filter for events There are some cases when filtering on a set flag of a field of a tracepoint is useful. But currently the only filtering commands for numbered fields is ==, !=, <, <=, >, >=. This does not help when you just want to trace if a specific flag is set. For example: > # sudo trace-cmd record -e brcmfmac:brcmf_dbg -f 'level & 0x40000' > disable all > enable brcmfmac:brcmf_dbg > path = /sys/kernel/debug/tracing/events/brcmfmac/brcmf_dbg/enable > (level & 0x40000) > ^ > parse_error: Invalid operator > When trying to trace brcmf_dbg when level has its 1 << 18 bit set, the filter fails to perform. By allowing a binary '&' operation, this gives the user the ability to test a bit. Note, a binary '|' is not added, as it doesn't make sense as fields must be compared to constants (for now), and ORing a constant will always return true. Link: http://lkml.kernel.org/r/1371057385.9844.261.camel@gandalf.local.home Suggested-by: Arend van Spriel Tested-by: Arend van Spriel Signed-off-by: Steven Rostedt --- kernel/trace/trace_events_filter.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'kernel') diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c index e1b653f7e1ca..0d883dc057d6 100644 --- a/kernel/trace/trace_events_filter.c +++ b/kernel/trace/trace_events_filter.c @@ -44,6 +44,7 @@ enum filter_op_ids OP_LE, OP_GT, OP_GE, + OP_BAND, OP_NONE, OP_OPEN_PAREN, }; @@ -54,6 +55,7 @@ struct filter_op { int precedence; }; +/* Order must be the same as enum filter_op_ids above */ static struct filter_op filter_ops[] = { { OP_OR, "||", 1 }, { OP_AND, "&&", 2 }, @@ -64,6 +66,7 @@ static struct filter_op filter_ops[] = { { OP_LE, "<=", 5 }, { OP_GT, ">", 5 }, { OP_GE, ">=", 5 }, + { OP_BAND, "&", 6 }, { OP_NONE, "OP_NONE", 0 }, { OP_OPEN_PAREN, "(", 0 }, }; @@ -156,6 +159,9 @@ static int filter_pred_##type(struct filter_pred *pred, void *event) \ case OP_GE: \ match = (*addr >= val); \ break; \ + case OP_BAND: \ + match = (*addr & val); \ + break; \ default: \ break; \ } \ -- cgit v1.2.3 From de7edd31457b626e54a0b2a7e8ff4d65492f01ad Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Fri, 14 Jun 2013 16:21:43 -0400 Subject: tracing: Disable tracing on warning Add a traceoff_on_warning option in both the kernel command line as well as a sysctl option. When set, any WARN*() function that is hit will cause the tracing_on variable to be cleared, which disables writing to the ring buffer. This is useful especially when tracing a bug with function tracing. When a warning is hit, the print caused by the warning can flood the trace with the functions that producing the output for the warning. This can make the resulting trace useless by either hiding where the bug happened, or worse, by overflowing the buffer and losing the trace of the bug totally. Acked-by: Peter Zijlstra Signed-off-by: Steven Rostedt --- kernel/panic.c | 3 +++ kernel/sysctl.c | 7 +++++++ kernel/trace/trace.c | 17 +++++++++++++++++ 3 files changed, 27 insertions(+) (limited to 'kernel') diff --git a/kernel/panic.c b/kernel/panic.c index 167ec097ce8b..4cea6cc628ab 100644 --- a/kernel/panic.c +++ b/kernel/panic.c @@ -15,6 +15,7 @@ #include #include #include +#include #include #include #include @@ -399,6 +400,8 @@ struct slowpath_args { static void warn_slowpath_common(const char *file, int line, void *caller, unsigned taint, struct slowpath_args *args) { + disable_trace_on_warning(); + printk(KERN_WARNING "------------[ cut here ]------------\n"); printk(KERN_WARNING "WARNING: at %s:%d %pS()\n", file, line, caller); diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 9edcf456e0fc..5b0f18c12800 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -600,6 +600,13 @@ static struct ctl_table kern_table[] = { .mode = 0644, .proc_handler = proc_dointvec, }, + { + .procname = "traceoff_on_warning", + .data = &__disable_trace_on_warning, + .maxlen = sizeof(__disable_trace_on_warning), + .mode = 0644, + .proc_handler = proc_dointvec, + }, #endif #ifdef CONFIG_MODULES { diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 5f4a09c12e0b..c4c9296b1916 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -115,6 +115,9 @@ cpumask_var_t __read_mostly tracing_buffer_mask; enum ftrace_dump_mode ftrace_dump_on_oops; +/* When set, tracing will stop when a WARN*() is hit */ +int __disable_trace_on_warning; + static int tracing_set_tracer(const char *buf); #define MAX_TRACER_SIZE 100 @@ -149,6 +152,13 @@ static int __init set_ftrace_dump_on_oops(char *str) } __setup("ftrace_dump_on_oops", set_ftrace_dump_on_oops); +static int __init stop_trace_on_warning(char *str) +{ + __disable_trace_on_warning = 1; + return 1; +} +__setup("traceoff_on_warning=", stop_trace_on_warning); + static int __init boot_alloc_snapshot(char *str) { allocate_snapshot = true; @@ -170,6 +180,7 @@ static int __init set_trace_boot_options(char *str) } __setup("trace_options=", set_trace_boot_options); + unsigned long long ns2usecs(cycle_t nsec) { nsec += 500; @@ -562,6 +573,12 @@ void tracing_off(void) } EXPORT_SYMBOL_GPL(tracing_off); +void disable_trace_on_warning(void) +{ + if (__disable_trace_on_warning) + tracing_off(); +} + /** * tracing_is_on - show state of ring buffers enabled */ -- cgit v1.2.3 From 195a84d91e92ee3fe571a2086a6db7e17bf5bc7c Mon Sep 17 00:00:00 2001 From: "zhangwei(Jovi)" Date: Fri, 14 Jun 2013 10:10:38 +0800 Subject: tracing/kprobes: Remove unnecessary checking of trace_probe_is_enabled Since tp->flags assignment was moved into function enable_trace_probe(), there is no need to use trace_probe_is_enabled to check flags in the same function. Remove the unnecessary checking. Link: http://lkml.kernel.org/r/51BA7B9E.3040807@huawei.com Acked-by: Masami Hiramatsu Cc: Frederic Weisbecker Cc: Oleg Nesterov Cc: Srikar Dronamraju Signed-off-by: zhangwei(Jovi) Signed-off-by: Steven Rostedt --- kernel/trace/trace_kprobe.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 9f46e98ba8f2..f2374172ba7b 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -240,8 +240,7 @@ enable_trace_probe(struct trace_probe *tp, struct ftrace_event_file *file) } else tp->flags |= TP_FLAG_PROFILE; - if (trace_probe_is_enabled(tp) && trace_probe_is_registered(tp) && - !trace_probe_has_gone(tp)) { + if (trace_probe_is_registered(tp) && !trace_probe_has_gone(tp)) { if (trace_probe_is_return(tp)) ret = enable_kretprobe(&tp->rp); else -- cgit v1.2.3 From 52d85d763086594f139bf7d3a5641abeb91d9f57 Mon Sep 17 00:00:00 2001 From: Juri Lelli Date: Wed, 12 Jun 2013 12:03:18 +0200 Subject: ftrace: Fix stddev calculation in function profiler MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When FUNCTION_GRAPH_TRACER is enabled, ftrace can profile kernel functions and print basic statistics about them. Unfortunately, running stddev calculation is wrong. This patch corrects it implementing Welford’s method: s^2 = 1 / (n * (n-1)) * (n * \Sum (x_i)^2 - (\Sum x_i)^2) . Link: http://lkml.kernel.org/r/1371031398-24048-1-git-send-email-juri.lelli@gmail.com Cc: Frederic Weisbecker Cc: Ingo Molnar Signed-off-by: Juri Lelli Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 800a8a2fbddb..26e19105cdcc 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -641,12 +641,18 @@ static int function_stat_show(struct seq_file *m, void *v) if (rec->counter <= 1) stddev = 0; else { - stddev = rec->time_squared - rec->counter * avg * avg; + /* + * Apply Welford's method: + * s^2 = 1 / (n * (n-1)) * (n * \Sum (x_i)^2 - (\Sum x_i)^2) + */ + stddev = rec->counter * rec->time_squared - + rec->time * rec->time; + /* * Divide only 1000 for ns^2 -> us^2 conversion. * trace_print_graph_duration will divide 1000 again. */ - do_div(stddev, (rec->counter - 1) * 1000); + do_div(stddev, rec->counter * (rec->counter - 1) * 1000); } trace_seq_init(&s); -- cgit v1.2.3 From 3c00ea82c724fab0b98f15428a804cb45eb9ad38 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 19 May 2013 20:45:15 +0200 Subject: watchdog: Rename confusing state variable We have two very conflicting state variable names in the watchdog: * watchdog_enabled: This one reflects the user interface. It's set to 1 by default and can be overriden with boot options or sysctl/procfs interface. * watchdog_disabled: This is the internal toggle state that tells if watchdog threads, timers and NMI events are currently running or not. This state mostly depends on the user settings. It's a convenient state latch. Now we really need to find clearer names because those are just too confusing to encourage deep review. watchdog_enabled now becomes watchdog_user_enabled to reflect its purpose as an interface. watchdog_disabled becomes watchdog_running to suggest its role as a pure internal state. Signed-off-by: Frederic Weisbecker Cc: Srivatsa S. Bhat Cc: Anish Singh Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Don Zickus --- kernel/sysctl.c | 4 ++-- kernel/watchdog.c | 30 +++++++++++++++--------------- 2 files changed, 17 insertions(+), 17 deletions(-) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 9edcf456e0fc..b0805652c4ff 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -801,7 +801,7 @@ static struct ctl_table kern_table[] = { #if defined(CONFIG_LOCKUP_DETECTOR) { .procname = "watchdog", - .data = &watchdog_enabled, + .data = &watchdog_user_enabled, .maxlen = sizeof (int), .mode = 0644, .proc_handler = proc_dowatchdog, @@ -828,7 +828,7 @@ static struct ctl_table kern_table[] = { }, { .procname = "nmi_watchdog", - .data = &watchdog_enabled, + .data = &watchdog_user_enabled, .maxlen = sizeof (int), .mode = 0644, .proc_handler = proc_dowatchdog, diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 52c9a9b91bdd..51c4f34d258e 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -29,9 +29,9 @@ #include #include -int watchdog_enabled = 1; +int watchdog_user_enabled = 1; int __read_mostly watchdog_thresh = 10; -static int __read_mostly watchdog_disabled = 1; +static int __read_mostly watchdog_running; static u64 __read_mostly sample_period; static DEFINE_PER_CPU(unsigned long, watchdog_touch_ts); @@ -63,7 +63,7 @@ static int __init hardlockup_panic_setup(char *str) else if (!strncmp(str, "nopanic", 7)) hardlockup_panic = 0; else if (!strncmp(str, "0", 1)) - watchdog_enabled = 0; + watchdog_user_enabled = 0; return 1; } __setup("nmi_watchdog=", hardlockup_panic_setup); @@ -82,7 +82,7 @@ __setup("softlockup_panic=", softlockup_panic_setup); static int __init nowatchdog_setup(char *str) { - watchdog_enabled = 0; + watchdog_user_enabled = 0; return 1; } __setup("nowatchdog", nowatchdog_setup); @@ -90,7 +90,7 @@ __setup("nowatchdog", nowatchdog_setup); /* deprecated */ static int __init nosoftlockup_setup(char *str) { - watchdog_enabled = 0; + watchdog_user_enabled = 0; return 1; } __setup("nosoftlockup", nosoftlockup_setup); @@ -158,7 +158,7 @@ void touch_all_softlockup_watchdogs(void) #ifdef CONFIG_HARDLOCKUP_DETECTOR void touch_nmi_watchdog(void) { - if (watchdog_enabled) { + if (watchdog_user_enabled) { unsigned cpu; for_each_present_cpu(cpu) { @@ -490,12 +490,12 @@ static int watchdog_enable_all_cpus(void) { int err = 0; - if (watchdog_disabled) { + if (!watchdog_running) { err = smpboot_register_percpu_thread(&watchdog_threads); if (err) pr_err("Failed to create watchdog threads, disabled\n"); else - watchdog_disabled = 0; + watchdog_running = 1; } return err; @@ -506,8 +506,8 @@ static int watchdog_enable_all_cpus(void) #ifdef CONFIG_SYSCTL static void watchdog_disable_all_cpus(void) { - if (!watchdog_disabled) { - watchdog_disabled = 1; + if (watchdog_running) { + watchdog_running = 0; smpboot_unregister_percpu_thread(&watchdog_threads); } } @@ -522,7 +522,7 @@ int proc_dowatchdog(struct ctl_table *table, int write, int err, old_thresh, old_enabled; old_thresh = ACCESS_ONCE(watchdog_thresh); - old_enabled = ACCESS_ONCE(watchdog_enabled); + old_enabled = ACCESS_ONCE(watchdog_user_enabled); err = proc_dointvec_minmax(table, write, buffer, lenp, ppos); if (err || !write) @@ -531,10 +531,10 @@ int proc_dowatchdog(struct ctl_table *table, int write, set_sample_period(); /* * Watchdog threads shouldn't be enabled if they are - * disabled. The 'watchdog_disabled' variable check in + * disabled. The 'watchdog_running' variable check in * watchdog_*_all_cpus() function takes care of this. */ - if (watchdog_enabled && watchdog_thresh) + if (watchdog_user_enabled && watchdog_thresh) err = watchdog_enable_all_cpus(); else watchdog_disable_all_cpus(); @@ -542,7 +542,7 @@ int proc_dowatchdog(struct ctl_table *table, int write, /* Restore old values on failure */ if (err) { watchdog_thresh = old_thresh; - watchdog_enabled = old_enabled; + watchdog_user_enabled = old_enabled; } return err; @@ -553,6 +553,6 @@ void __init lockup_detector_init(void) { set_sample_period(); - if (watchdog_enabled) + if (watchdog_user_enabled) watchdog_enable_all_cpus(); } -- cgit v1.2.3 From 940be35ac0139530d7554aa2352a8388e3d4adca Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 7 Jun 2013 13:35:42 +0200 Subject: watchdog: Boot-disable by default on full dynticks When the watchdog runs, it prevents the full dynticks CPUs from stopping their tick because the hard lockup detector uses perf events internally, which in turn rely on the periodic tick. Since this is a rather confusing behaviour that is not easy to track down and identify for those who want to test CONFIG_NO_HZ_FULL, let's default disable the watchdog on boot time when full dynticks is enabled. The user can still enable it later on runtime using proc or sysctl. Reported-by: Steven Rostedt Suggested-by: Peter Zijlstra Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Andrew Morton Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Li Zhong Cc: Don Zickus Cc: Srivatsa S. Bhat Cc: Anish Singh --- kernel/watchdog.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'kernel') diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 51c4f34d258e..1241d8c91d5e 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -553,6 +553,14 @@ void __init lockup_detector_init(void) { set_sample_period(); +#ifdef CONFIG_NO_HZ_FULL + if (watchdog_user_enabled) { + watchdog_user_enabled = 0; + pr_warning("Disabled lockup detectors by default for full dynticks\n"); + pr_warning("You can reactivate it with 'sysctl -w kernel.watchdog=1'\n"); + } +#endif + if (watchdog_user_enabled) watchdog_enable_all_cpus(); } -- cgit v1.2.3 From 5b8621a68fdcd2baf1d3b413726f913a5254d46a Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 8 Jun 2013 13:47:31 +0200 Subject: nohz: Remove obsolete check for full dynticks CPUs to be RCU nocbs Building full dynticks now implies that all CPUs are forced into RCU nocb mode through CONFIG_RCU_NOCB_CPU_ALL. The dynamic check has become useless. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Andrew Morton Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Li Zhong Cc: Borislav Petkov --- kernel/time/tick-sched.c | 10 ---------- 1 file changed, 10 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index d87d22cb9bf2..b15750139260 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -351,16 +351,6 @@ void __init tick_nohz_init(void) } cpu_notifier(tick_nohz_cpu_down_callback, 0); - - /* Make sure full dynticks CPU are also RCU nocbs */ - for_each_cpu(cpu, nohz_full_mask) { - if (!rcu_is_nocb_cpu(cpu)) { - pr_warning("NO_HZ: CPU %d is not RCU nocb: " - "cleared from nohz_full range", cpu); - cpumask_clear_cpu(cpu, nohz_full_mask); - } - } - cpulist_scnprintf(nohz_full_buf, sizeof(nohz_full_buf), nohz_full_mask); pr_info("NO_HZ: Full dynticks CPUs: %s.\n", nohz_full_buf); } -- cgit v1.2.3 From 8b4d801b2b123b6c09742f861fe44a8527b84d47 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Thu, 20 Jun 2013 17:50:06 +0200 Subject: hw_breakpoint: Fix cpu check in task_bp_pinned(cpu) trinity fuzzer triggered WARN_ONCE("Can't find any breakpoint slot") in arch_install_hw_breakpoint() but the problem is not arch-specific. The problem is, task_bp_pinned(cpu) checks "cpu == iter->cpu" but this doesn't account the "all cpus" events with iter->cpu < 0. This means that, say, register_user_hw_breakpoint(tsk) can happily create the arbitrary number > HBP_NUM of breakpoints which can not be activated. toggle_bp_task_slot() is equally wrong by the same reason and nr_task_bp_pinned[] can have negative entries. Simple test: # perl -e 'sleep 1 while 1' & # perf record -e mem:0x10,mem:0x10,mem:0x10,mem:0x10,mem:0x10 -p `pidof perl` Before this patch this triggers the same problem/WARN_ON(), after the patch it correctly fails with -ENOSPC. Reported-by: Vince Weaver Signed-off-by: Oleg Nesterov Acked-by: Frederic Weisbecker Cc: Link: http://lkml.kernel.org/r/20130620155006.GA6324@redhat.com Signed-off-by: Ingo Molnar --- kernel/events/hw_breakpoint.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/events/hw_breakpoint.c b/kernel/events/hw_breakpoint.c index a64f8aeb5c1f..a853deabe6cf 100644 --- a/kernel/events/hw_breakpoint.c +++ b/kernel/events/hw_breakpoint.c @@ -120,7 +120,7 @@ static int task_bp_pinned(int cpu, struct perf_event *bp, enum bp_type_idx type) list_for_each_entry(iter, &bp_task_head, hw.bp_list) { if (iter->hw.bp_target == tsk && find_slot_idx(iter) == type && - cpu == iter->cpu) + (iter->cpu < 0 || cpu == iter->cpu)) count += hw_breakpoint_weight(iter); } -- cgit v1.2.3 From c790b0ad23f427c7522ffed264706238c57c007e Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Thu, 20 Jun 2013 17:50:09 +0200 Subject: hw_breakpoint: Use cpu_possible_mask in {reserve,release}_bp_slot() fetch_bp_busy_slots() and toggle_bp_slot() use for_each_online_cpu(), this is obviously wrong wrt cpu_up() or cpu_down(), we can over/under account the per-cpu numbers. For example: # echo 0 >> /sys/devices/system/cpu/cpu1/online # perf record -e mem:0x10 -p 1 & # echo 1 >> /sys/devices/system/cpu/cpu1/online # perf record -e mem:0x10,mem:0x10,mem:0x10,mem:0x10 -C1 -a & # taskset -p 0x2 1 triggers the same WARN_ONCE("Can't find any breakpoint slot") in arch_install_hw_breakpoint(). Reported-by: Vince Weaver Signed-off-by: Oleg Nesterov Acked-by: Frederic Weisbecker Cc: Link: http://lkml.kernel.org/r/20130620155009.GA6327@redhat.com Signed-off-by: Ingo Molnar --- kernel/events/hw_breakpoint.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/events/hw_breakpoint.c b/kernel/events/hw_breakpoint.c index a853deabe6cf..20185ea64aa6 100644 --- a/kernel/events/hw_breakpoint.c +++ b/kernel/events/hw_breakpoint.c @@ -149,7 +149,7 @@ fetch_bp_busy_slots(struct bp_busy_slots *slots, struct perf_event *bp, return; } - for_each_online_cpu(cpu) { + for_each_possible_cpu(cpu) { unsigned int nr; nr = per_cpu(nr_cpu_bp_pinned[type], cpu); @@ -235,7 +235,7 @@ toggle_bp_slot(struct perf_event *bp, bool enable, enum bp_type_idx type, if (cpu >= 0) { toggle_bp_task_slot(bp, cpu, enable, type, weight); } else { - for_each_online_cpu(cpu) + for_each_possible_cpu(cpu) toggle_bp_task_slot(bp, cpu, enable, type, weight); } -- cgit v1.2.3 From e1ebe86203e6532eb5a0ae8f26ccae47aca548ae Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Thu, 20 Jun 2013 17:50:11 +0200 Subject: hw_breakpoint: Simplify list/idx mess in toggle_bp_slot() paths The enable/disable logic in toggle_bp_slot() is not symmetrical and imho very confusing. "old_count" in toggle_bp_task_slot() is actually new_count because this bp was already removed from the list. Change toggle_bp_slot() to always call list_add/list_del after toggle_bp_task_slot(). This way old_idx is task_bp_pinned() and this entry should be decremented, new_idx is +/-weight and we need to increment this element. The code/logic looks obvious. Reported-by: Vince Weaver Signed-off-by: Oleg Nesterov Acked-by: Frederic Weisbecker Link: http://lkml.kernel.org/r/20130620155011.GA6330@redhat.com Signed-off-by: Ingo Molnar --- kernel/events/hw_breakpoint.c | 40 ++++++++++++++++------------------------ 1 file changed, 16 insertions(+), 24 deletions(-) (limited to 'kernel') diff --git a/kernel/events/hw_breakpoint.c b/kernel/events/hw_breakpoint.c index ef8ebe560949..dee0148dcf54 100644 --- a/kernel/events/hw_breakpoint.c +++ b/kernel/events/hw_breakpoint.c @@ -185,26 +185,20 @@ fetch_this_slot(struct bp_busy_slots *slots, int weight) static void toggle_bp_task_slot(struct perf_event *bp, int cpu, bool enable, enum bp_type_idx type, int weight) { - unsigned int *tsk_pinned; - int old_count = 0; - int old_idx = 0; - int idx = 0; - - old_count = task_bp_pinned(cpu, bp, type); - old_idx = old_count - 1; - idx = old_idx + weight; - - /* tsk_pinned[n] is the number of tasks having n breakpoints */ - tsk_pinned = per_cpu(nr_task_bp_pinned[type], cpu); - if (enable) { - tsk_pinned[idx]++; - if (old_count > 0) - tsk_pinned[old_idx]--; - } else { - tsk_pinned[idx]--; - if (old_count > 0) - tsk_pinned[old_idx]++; - } + /* tsk_pinned[n-1] is the number of tasks having n>0 breakpoints */ + unsigned int *tsk_pinned = per_cpu(nr_task_bp_pinned[type], cpu); + int old_idx, new_idx; + + old_idx = task_bp_pinned(cpu, bp, type) - 1; + if (enable) + new_idx = old_idx + weight; + else + new_idx = old_idx - weight; + + if (old_idx >= 0) + tsk_pinned[old_idx]--; + if (new_idx >= 0) + tsk_pinned[new_idx]++; } /* @@ -228,10 +222,6 @@ toggle_bp_slot(struct perf_event *bp, bool enable, enum bp_type_idx type, } /* Pinned counter task profiling */ - - if (!enable) - list_del(&bp->hw.bp_list); - if (cpu >= 0) { toggle_bp_task_slot(bp, cpu, enable, type, weight); } else { @@ -241,6 +231,8 @@ toggle_bp_slot(struct perf_event *bp, bool enable, enum bp_type_idx type, if (enable) list_add_tail(&bp->hw.bp_list, &bp_task_head); + else + list_del(&bp->hw.bp_list); } /* -- cgit v1.2.3 From 7ab71f3244e9f970c29566c5a67e13d1fa38c387 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Thu, 20 Jun 2013 17:50:13 +0200 Subject: hw_breakpoint: Simplify the "weight" usage in toggle_bp_slot() paths Change toggle_bp_slot() to make "weight" negative if !enable. This way we can always use "+ weight" without additional "if (enable)" check and toggle_bp_task_slot() no longer needs this arg. Reported-by: Vince Weaver Signed-off-by: Oleg Nesterov Acked-by: Frederic Weisbecker Link: http://lkml.kernel.org/r/20130620155013.GA6337@redhat.com Signed-off-by: Ingo Molnar --- kernel/events/hw_breakpoint.c | 20 ++++++++------------ 1 file changed, 8 insertions(+), 12 deletions(-) (limited to 'kernel') diff --git a/kernel/events/hw_breakpoint.c b/kernel/events/hw_breakpoint.c index dee0148dcf54..5cd4f6d9652c 100644 --- a/kernel/events/hw_breakpoint.c +++ b/kernel/events/hw_breakpoint.c @@ -182,7 +182,7 @@ fetch_this_slot(struct bp_busy_slots *slots, int weight) /* * Add a pinned breakpoint for the given task in our constraint table */ -static void toggle_bp_task_slot(struct perf_event *bp, int cpu, bool enable, +static void toggle_bp_task_slot(struct perf_event *bp, int cpu, enum bp_type_idx type, int weight) { /* tsk_pinned[n-1] is the number of tasks having n>0 breakpoints */ @@ -190,10 +190,7 @@ static void toggle_bp_task_slot(struct perf_event *bp, int cpu, bool enable, int old_idx, new_idx; old_idx = task_bp_pinned(cpu, bp, type) - 1; - if (enable) - new_idx = old_idx + weight; - else - new_idx = old_idx - weight; + new_idx = old_idx + weight; if (old_idx >= 0) tsk_pinned[old_idx]--; @@ -211,22 +208,21 @@ toggle_bp_slot(struct perf_event *bp, bool enable, enum bp_type_idx type, int cpu = bp->cpu; struct task_struct *tsk = bp->hw.bp_target; + if (!enable) + weight = -weight; + /* Pinned counter cpu profiling */ if (!tsk) { - - if (enable) - per_cpu(nr_cpu_bp_pinned[type], bp->cpu) += weight; - else - per_cpu(nr_cpu_bp_pinned[type], bp->cpu) -= weight; + per_cpu(nr_cpu_bp_pinned[type], cpu) += weight; return; } /* Pinned counter task profiling */ if (cpu >= 0) { - toggle_bp_task_slot(bp, cpu, enable, type, weight); + toggle_bp_task_slot(bp, cpu, type, weight); } else { for_each_possible_cpu(cpu) - toggle_bp_task_slot(bp, cpu, enable, type, weight); + toggle_bp_task_slot(bp, cpu, type, weight); } if (enable) -- cgit v1.2.3 From 1c10adbb929936316f71df089ace699fce037e24 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Thu, 20 Jun 2013 17:50:15 +0200 Subject: hw_breakpoint: Introduce cpumask_of_bp() Add the trivial helper which simply returns cpumask_of() or cpu_possible_mask depending on bp->cpu. Change fetch_bp_busy_slots() and toggle_bp_slot() to always do for_each_cpu(cpumask_of_bp) to simplify the code and avoid the code duplication. Reported-by: Vince Weaver Signed-off-by: Oleg Nesterov Acked-by: Frederic Weisbecker Link: http://lkml.kernel.org/r/20130620155015.GA6340@redhat.com Signed-off-by: Ingo Molnar --- kernel/events/hw_breakpoint.c | 43 +++++++++++++++++-------------------------- 1 file changed, 17 insertions(+), 26 deletions(-) (limited to 'kernel') diff --git a/kernel/events/hw_breakpoint.c b/kernel/events/hw_breakpoint.c index 5cd4f6d9652c..9c71445328af 100644 --- a/kernel/events/hw_breakpoint.c +++ b/kernel/events/hw_breakpoint.c @@ -127,6 +127,13 @@ static int task_bp_pinned(int cpu, struct perf_event *bp, enum bp_type_idx type) return count; } +static const struct cpumask *cpumask_of_bp(struct perf_event *bp) +{ + if (bp->cpu >= 0) + return cpumask_of(bp->cpu); + return cpu_possible_mask; +} + /* * Report the number of pinned/un-pinned breakpoints we have in * a given cpu (cpu > -1) or in all of them (cpu = -1). @@ -135,25 +142,13 @@ static void fetch_bp_busy_slots(struct bp_busy_slots *slots, struct perf_event *bp, enum bp_type_idx type) { - int cpu = bp->cpu; - struct task_struct *tsk = bp->hw.bp_target; - - if (cpu >= 0) { - slots->pinned = per_cpu(nr_cpu_bp_pinned[type], cpu); - if (!tsk) - slots->pinned += max_task_bp_pinned(cpu, type); - else - slots->pinned += task_bp_pinned(cpu, bp, type); - slots->flexible = per_cpu(nr_bp_flexible[type], cpu); - - return; - } + const struct cpumask *cpumask = cpumask_of_bp(bp); + int cpu; - for_each_possible_cpu(cpu) { - unsigned int nr; + for_each_cpu(cpu, cpumask) { + unsigned int nr = per_cpu(nr_cpu_bp_pinned[type], cpu); - nr = per_cpu(nr_cpu_bp_pinned[type], cpu); - if (!tsk) + if (!bp->hw.bp_target) nr += max_task_bp_pinned(cpu, type); else nr += task_bp_pinned(cpu, bp, type); @@ -205,25 +200,21 @@ static void toggle_bp_slot(struct perf_event *bp, bool enable, enum bp_type_idx type, int weight) { - int cpu = bp->cpu; - struct task_struct *tsk = bp->hw.bp_target; + const struct cpumask *cpumask = cpumask_of_bp(bp); + int cpu; if (!enable) weight = -weight; /* Pinned counter cpu profiling */ - if (!tsk) { - per_cpu(nr_cpu_bp_pinned[type], cpu) += weight; + if (!bp->hw.bp_target) { + per_cpu(nr_cpu_bp_pinned[type], bp->cpu) += weight; return; } /* Pinned counter task profiling */ - if (cpu >= 0) { + for_each_cpu(cpu, cpumask) toggle_bp_task_slot(bp, cpu, type, weight); - } else { - for_each_possible_cpu(cpu) - toggle_bp_task_slot(bp, cpu, type, weight); - } if (enable) list_add_tail(&bp->hw.bp_list, &bp_task_head); -- cgit v1.2.3 From e12cbc10cb27fcbe51b5f68e2015138dc451a2eb Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Thu, 20 Jun 2013 17:50:18 +0200 Subject: hw_breakpoint: Simplify *register_wide_hw_breakpoint() 1. register_wide_hw_breakpoint() can use unregister_ if failure, no need to duplicate the code. 2. "struct perf_event **pevent" adds the unnecesary lever of indirection and complication, use per_cpu(*cpu_events, cpu). Reported-by: Vince Weaver Signed-off-by: Oleg Nesterov Acked-by: Frederic Weisbecker Link: http://lkml.kernel.org/r/20130620155018.GA6347@redhat.com Signed-off-by: Ingo Molnar --- kernel/events/hw_breakpoint.c | 34 +++++++++++----------------------- 1 file changed, 11 insertions(+), 23 deletions(-) (limited to 'kernel') diff --git a/kernel/events/hw_breakpoint.c b/kernel/events/hw_breakpoint.c index 9c71445328af..38418f786f36 100644 --- a/kernel/events/hw_breakpoint.c +++ b/kernel/events/hw_breakpoint.c @@ -497,8 +497,8 @@ register_wide_hw_breakpoint(struct perf_event_attr *attr, perf_overflow_handler_t triggered, void *context) { - struct perf_event * __percpu *cpu_events, **pevent, *bp; - long err; + struct perf_event * __percpu *cpu_events, *bp; + long err = 0; int cpu; cpu_events = alloc_percpu(typeof(*cpu_events)); @@ -507,31 +507,21 @@ register_wide_hw_breakpoint(struct perf_event_attr *attr, get_online_cpus(); for_each_online_cpu(cpu) { - pevent = per_cpu_ptr(cpu_events, cpu); bp = perf_event_create_kernel_counter(attr, cpu, NULL, triggered, context); - - *pevent = bp; - if (IS_ERR(bp)) { err = PTR_ERR(bp); - goto fail; + break; } - } - put_online_cpus(); - - return cpu_events; -fail: - for_each_online_cpu(cpu) { - pevent = per_cpu_ptr(cpu_events, cpu); - if (IS_ERR(*pevent)) - break; - unregister_hw_breakpoint(*pevent); + per_cpu(*cpu_events, cpu) = bp; } put_online_cpus(); - free_percpu(cpu_events); + if (likely(!err)) + return cpu_events; + + unregister_wide_hw_breakpoint(cpu_events); return (void __percpu __force *)ERR_PTR(err); } EXPORT_SYMBOL_GPL(register_wide_hw_breakpoint); @@ -543,12 +533,10 @@ EXPORT_SYMBOL_GPL(register_wide_hw_breakpoint); void unregister_wide_hw_breakpoint(struct perf_event * __percpu *cpu_events) { int cpu; - struct perf_event **pevent; - for_each_possible_cpu(cpu) { - pevent = per_cpu_ptr(cpu_events, cpu); - unregister_hw_breakpoint(*pevent); - } + for_each_possible_cpu(cpu) + unregister_hw_breakpoint(per_cpu(*cpu_events, cpu)); + free_percpu(cpu_events); } EXPORT_SYMBOL_GPL(unregister_wide_hw_breakpoint); -- cgit v1.2.3 From bde96030f438b5eb6fb74f3bdd06d9f68bb3ba00 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Thu, 20 Jun 2013 17:50:20 +0200 Subject: hw_breakpoint: Introduce "struct bp_cpuinfo" This patch simply moves all per-cpu variables into the new single per-cpu "struct bp_cpuinfo". To me this looks more logical and clean, but this can also simplify the further potential changes. In particular, I do not think this memory should be per-cpu, it is never used "locally". After this change it is trivial to turn it into, say, bootmem[nr_cpu_ids]. Reported-by: Vince Weaver Signed-off-by: Oleg Nesterov Acked-by: Frederic Weisbecker Link: http://lkml.kernel.org/r/20130620155020.GA6350@redhat.com Signed-off-by: Ingo Molnar --- kernel/events/hw_breakpoint.c | 69 ++++++++++++++++++++++--------------------- 1 file changed, 36 insertions(+), 33 deletions(-) (limited to 'kernel') diff --git a/kernel/events/hw_breakpoint.c b/kernel/events/hw_breakpoint.c index 38418f786f36..1559fb0b9296 100644 --- a/kernel/events/hw_breakpoint.c +++ b/kernel/events/hw_breakpoint.c @@ -46,23 +46,26 @@ #include #include - - /* * Constraints data */ +struct bp_cpuinfo { + /* Number of pinned cpu breakpoints in a cpu */ + unsigned int cpu_pinned; + /* tsk_pinned[n] is the number of tasks having n+1 breakpoints */ + unsigned int *tsk_pinned; + /* Number of non-pinned cpu/task breakpoints in a cpu */ + unsigned int flexible; /* XXX: placeholder, see fetch_this_slot() */ +}; -/* Number of pinned cpu breakpoints in a cpu */ -static DEFINE_PER_CPU(unsigned int, nr_cpu_bp_pinned[TYPE_MAX]); - -/* Number of pinned task breakpoints in a cpu */ -static DEFINE_PER_CPU(unsigned int *, nr_task_bp_pinned[TYPE_MAX]); - -/* Number of non-pinned cpu/task breakpoints in a cpu */ -static DEFINE_PER_CPU(unsigned int, nr_bp_flexible[TYPE_MAX]); - +static DEFINE_PER_CPU(struct bp_cpuinfo, bp_cpuinfo[TYPE_MAX]); static int nr_slots[TYPE_MAX]; +static struct bp_cpuinfo *get_bp_info(int cpu, enum bp_type_idx type) +{ + return per_cpu_ptr(bp_cpuinfo + type, cpu); +} + /* Keep track of the breakpoints attached to tasks */ static LIST_HEAD(bp_task_head); @@ -96,8 +99,8 @@ static inline enum bp_type_idx find_slot_idx(struct perf_event *bp) */ static unsigned int max_task_bp_pinned(int cpu, enum bp_type_idx type) { + unsigned int *tsk_pinned = get_bp_info(cpu, type)->tsk_pinned; int i; - unsigned int *tsk_pinned = per_cpu(nr_task_bp_pinned[type], cpu); for (i = nr_slots[type] - 1; i >= 0; i--) { if (tsk_pinned[i] > 0) @@ -146,8 +149,10 @@ fetch_bp_busy_slots(struct bp_busy_slots *slots, struct perf_event *bp, int cpu; for_each_cpu(cpu, cpumask) { - unsigned int nr = per_cpu(nr_cpu_bp_pinned[type], cpu); + struct bp_cpuinfo *info = get_bp_info(cpu, type); + int nr; + nr = info->cpu_pinned; if (!bp->hw.bp_target) nr += max_task_bp_pinned(cpu, type); else @@ -156,8 +161,7 @@ fetch_bp_busy_slots(struct bp_busy_slots *slots, struct perf_event *bp, if (nr > slots->pinned) slots->pinned = nr; - nr = per_cpu(nr_bp_flexible[type], cpu); - + nr = info->flexible; if (nr > slots->flexible) slots->flexible = nr; } @@ -180,8 +184,7 @@ fetch_this_slot(struct bp_busy_slots *slots, int weight) static void toggle_bp_task_slot(struct perf_event *bp, int cpu, enum bp_type_idx type, int weight) { - /* tsk_pinned[n-1] is the number of tasks having n>0 breakpoints */ - unsigned int *tsk_pinned = per_cpu(nr_task_bp_pinned[type], cpu); + unsigned int *tsk_pinned = get_bp_info(cpu, type)->tsk_pinned; int old_idx, new_idx; old_idx = task_bp_pinned(cpu, bp, type) - 1; @@ -208,7 +211,7 @@ toggle_bp_slot(struct perf_event *bp, bool enable, enum bp_type_idx type, /* Pinned counter cpu profiling */ if (!bp->hw.bp_target) { - per_cpu(nr_cpu_bp_pinned[type], bp->cpu) += weight; + get_bp_info(bp->cpu, type)->cpu_pinned += weight; return; } @@ -240,8 +243,8 @@ __weak void arch_unregister_hw_breakpoint(struct perf_event *bp) * * - If attached to a single cpu, check: * - * (per_cpu(nr_bp_flexible, cpu) || (per_cpu(nr_cpu_bp_pinned, cpu) - * + max(per_cpu(nr_task_bp_pinned, cpu)))) < HBP_NUM + * (per_cpu(info->flexible, cpu) || (per_cpu(info->cpu_pinned, cpu) + * + max(per_cpu(info->tsk_pinned, cpu)))) < HBP_NUM * * -> If there are already non-pinned counters in this cpu, it means * there is already a free slot for them. @@ -251,8 +254,8 @@ __weak void arch_unregister_hw_breakpoint(struct perf_event *bp) * * - If attached to every cpus, check: * - * (per_cpu(nr_bp_flexible, *) || (max(per_cpu(nr_cpu_bp_pinned, *)) - * + max(per_cpu(nr_task_bp_pinned, *)))) < HBP_NUM + * (per_cpu(info->flexible, *) || (max(per_cpu(info->cpu_pinned, *)) + * + max(per_cpu(info->tsk_pinned, *)))) < HBP_NUM * * -> This is roughly the same, except we check the number of per cpu * bp for every cpu and we keep the max one. Same for the per tasks @@ -263,16 +266,16 @@ __weak void arch_unregister_hw_breakpoint(struct perf_event *bp) * * - If attached to a single cpu, check: * - * ((per_cpu(nr_bp_flexible, cpu) > 1) + per_cpu(nr_cpu_bp_pinned, cpu) - * + max(per_cpu(nr_task_bp_pinned, cpu))) < HBP_NUM + * ((per_cpu(info->flexible, cpu) > 1) + per_cpu(info->cpu_pinned, cpu) + * + max(per_cpu(info->tsk_pinned, cpu))) < HBP_NUM * - * -> Same checks as before. But now the nr_bp_flexible, if any, must keep + * -> Same checks as before. But now the info->flexible, if any, must keep * one register at least (or they will never be fed). * * - If attached to every cpus, check: * - * ((per_cpu(nr_bp_flexible, *) > 1) + max(per_cpu(nr_cpu_bp_pinned, *)) - * + max(per_cpu(nr_task_bp_pinned, *))) < HBP_NUM + * ((per_cpu(info->flexible, *) > 1) + max(per_cpu(info->cpu_pinned, *)) + * + max(per_cpu(info->tsk_pinned, *))) < HBP_NUM */ static int __reserve_bp_slot(struct perf_event *bp) { @@ -622,7 +625,6 @@ static struct pmu perf_breakpoint = { int __init init_hw_breakpoint(void) { - unsigned int **task_bp_pinned; int cpu, err_cpu; int i; @@ -631,10 +633,11 @@ int __init init_hw_breakpoint(void) for_each_possible_cpu(cpu) { for (i = 0; i < TYPE_MAX; i++) { - task_bp_pinned = &per_cpu(nr_task_bp_pinned[i], cpu); - *task_bp_pinned = kzalloc(sizeof(int) * nr_slots[i], - GFP_KERNEL); - if (!*task_bp_pinned) + struct bp_cpuinfo *info = get_bp_info(cpu, i); + + info->tsk_pinned = kcalloc(nr_slots[i], sizeof(int), + GFP_KERNEL); + if (!info->tsk_pinned) goto err_alloc; } } @@ -648,7 +651,7 @@ int __init init_hw_breakpoint(void) err_alloc: for_each_possible_cpu(err_cpu) { for (i = 0; i < TYPE_MAX; i++) - kfree(per_cpu(nr_task_bp_pinned[i], err_cpu)); + kfree(get_bp_info(err_cpu, i)->tsk_pinned); if (err_cpu == cpu) break; } -- cgit v1.2.3 From d24c2a4f919d17bd1ae4f4010a38ab07ece99cf7 Mon Sep 17 00:00:00 2001 From: Sahara Date: Thu, 20 Jun 2013 11:33:57 +0900 Subject: PM / QoS: correct the valid range of pm_qos_class The valid start index for pm_qos_array is not 0, but PM_QOS_CPU_DMA_LATENCY. There is a null_pm_qos at index 0 of pm_qos_array. However, null_pm_qos is not created as misc device so that inclusion of 0 index for checking pm_qos_class especially for file operations is not proper here. [rjw: Changelog, a bit] Signed-off-by: Sahara Signed-off-by: Rafael J. Wysocki --- kernel/power/qos.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/power/qos.c b/kernel/power/qos.c index 587dddeebf15..f2f5f6e22a3c 100644 --- a/kernel/power/qos.c +++ b/kernel/power/qos.c @@ -477,7 +477,7 @@ static int find_pm_qos_object_by_minor(int minor) { int pm_qos_class; - for (pm_qos_class = 0; + for (pm_qos_class = PM_QOS_CPU_DMA_LATENCY; pm_qos_class < PM_QOS_NUM_CLASSES; pm_qos_class++) { if (minor == pm_qos_array[pm_qos_class]->pm_qos_power_miscdev.minor) @@ -491,7 +491,7 @@ static int pm_qos_power_open(struct inode *inode, struct file *filp) long pm_qos_class; pm_qos_class = find_pm_qos_object_by_minor(iminor(inode)); - if (pm_qos_class >= 0) { + if (pm_qos_class >= PM_QOS_CPU_DMA_LATENCY) { struct pm_qos_request *req = kzalloc(sizeof(*req), GFP_KERNEL); if (!req) return -ENOMEM; @@ -584,7 +584,7 @@ static int __init pm_qos_power_init(void) BUILD_BUG_ON(ARRAY_SIZE(pm_qos_array) != PM_QOS_NUM_CLASSES); - for (i = 1; i < PM_QOS_NUM_CLASSES; i++) { + for (i = PM_QOS_CPU_DMA_LATENCY; i < PM_QOS_NUM_CLASSES; i++) { ret = register_pm_qos_misc(pm_qos_array[i]); if (ret < 0) { printk(KERN_ERR "pm_qos_param: %s setup failed\n", -- cgit v1.2.3 From bb177fedd348c92c2bea6adc9a2163ebff15272e Mon Sep 17 00:00:00 2001 From: Julius Werner Date: Wed, 12 Jun 2013 12:55:22 -0700 Subject: PM / Sleep: Print last wakeup source on failed wakeup_count write Commit a938da06 introduced a useful little log message to tell users/debuggers which wakeup source aborted a suspend. However, this message is only printed if the abort happens during the in-kernel suspend path (after writing /sys/power/state). The full specification of the /sys/power/wakeup_count facility allows user-space power managers to double-check if wakeups have already happened before it actually tries to suspend (e.g. while it was running user-space pre-suspend hooks), by writing the last known wakeup_count value to /sys/power/wakeup_count. This patch changes the sysfs handler for that node to also print said log message if that write fails, so that we can figure out the offending wakeup source for both kinds of suspend aborts. Signed-off-by: Julius Werner Signed-off-by: Rafael J. Wysocki --- kernel/power/main.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'kernel') diff --git a/kernel/power/main.c b/kernel/power/main.c index d77663bfedeb..0828070d38b4 100644 --- a/kernel/power/main.c +++ b/kernel/power/main.c @@ -424,6 +424,8 @@ static ssize_t wakeup_count_store(struct kobject *kobj, if (sscanf(buf, "%u", &val) == 1) { if (pm_save_wakeup_count(val)) error = n; + else + pm_print_active_wakeup_sources(); } out: -- cgit v1.2.3 From ea8deb8dfa6b0e8d1b3d1051585706739b46656c Mon Sep 17 00:00:00 2001 From: Daniel Lezcano Date: Mon, 17 Jun 2013 18:15:35 +0200 Subject: tick: Fix tick_broadcast_pending_mask not cleared The recent modification in the cpuidle framework consolidated the timer broadcast code across the different drivers by setting a new flag in the idle state. It tells the cpuidle core code to enter/exit the broadcast mode for the cpu when entering a deep idle state. The broadcast timer enter/exit is no longer handled by the back-end driver. This change made the local interrupt to be enabled *before* calling CLOCK_EVENT_NOTIFY_EXIT. On a tegra114, a four cores system, when the flag has been introduced in the driver, the following warning appeared: WARNING: at kernel/time/tick-broadcast.c:578 tick_broadcast_oneshot_control CPU: 2 PID: 0 Comm: swapper/2 Not tainted 3.10.0-rc3-next-20130529+ #15 [] (tick_broadcast_oneshot_control+0x1a4/0x1d0) from [] (tick_notify+0x240/0x40c) [] (tick_notify+0x240/0x40c) from [] (notifier_call_chain+0x44/0x84) [] (notifier_call_chain+0x44/0x84) from [] (raw_notifier_call_chain+0x18/0x20) [] (raw_notifier_call_chain+0x18/0x20) from [] (clockevents_notify+0x28/0x170) [] (clockevents_notify+0x28/0x170) from [] (cpuidle_idle_call+0x11c/0x168) [] (cpuidle_idle_call+0x11c/0x168) from [] (arch_cpu_idle+0x8/0x38) [] (arch_cpu_idle+0x8/0x38) from [] (cpu_startup_entry+0x60/0x134) [] (cpu_startup_entry+0x60/0x134) from [<804fe9a4>] (0x804fe9a4) I don't have the hardware, so I wasn't able to reproduce the warning but after looking a while at the code, I deduced the following: 1. the CPU2 enters a deep idle state and sets the broadcast timer 2. the timer expires, the tick_handle_oneshot_broadcast function is called, setting the tick_broadcast_pending_mask and waking up the idle cpu CPU2 3. the CPU2 exits idle handles the interrupt and then invokes tick_broadcast_oneshot_control with CLOCK_EVENT_NOTIFY_EXIT which runs the following code: [...] if (dev->next_event.tv64 == KTIME_MAX) goto out; if (cpumask_test_and_clear_cpu(cpu, tick_broadcast_pending_mask)) goto out; [...] So if there is no next event scheduled for CPU2, we fulfil the first condition and jump out without clearing the tick_broadcast_pending_mask. 4. CPU2 goes to deep idle again and calls tick_broadcast_oneshot_control with CLOCK_NOTIFY_EVENT_ENTER but with the tick_broadcast_pending_mask set for CPU2, triggering the warning. The issue only surfaced due to the modifications of the cpuidle framework, which resulted in interrupts being enabled before the call to the clockevents code. If the call happens before interrupts have been enabled, the warning cannot trigger, because there is still the event pending which caused the broadcast timer expiry. Move the check for the next event below the check for the pending bit, so the pending bit gets cleared whether an event is scheduled on the cpu or not. [ tglx: Massaged changelog ] Signed-off-by: Daniel Lezcano Reported-and-tested-by: Joseph Lo Cc: Stephen Warren Cc: linux-arm-kernel@lists.infradead.org Cc: linaro-kernel@lists.linaro.org Link: http://lkml.kernel.org/r/1371485735-31249-1-git-send-email-daniel.lezcano@linaro.org Signed-off-by: Thomas Gleixner --- kernel/time/tick-broadcast.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-broadcast.c b/kernel/time/tick-broadcast.c index b4c245580b79..20d6fba70652 100644 --- a/kernel/time/tick-broadcast.c +++ b/kernel/time/tick-broadcast.c @@ -599,8 +599,6 @@ void tick_broadcast_oneshot_control(unsigned long reason) } else { if (cpumask_test_and_clear_cpu(cpu, tick_broadcast_oneshot_mask)) { clockevents_set_mode(dev, CLOCK_EVT_MODE_ONESHOT); - if (dev->next_event.tv64 == KTIME_MAX) - goto out; /* * The cpu which was handling the broadcast * timer marked this cpu in the broadcast @@ -614,6 +612,11 @@ void tick_broadcast_oneshot_control(unsigned long reason) tick_broadcast_pending_mask)) goto out; + /* + * Bail out if there is no next event. + */ + if (dev->next_event.tv64 == KTIME_MAX) + goto out; /* * If the pending bit is not set, then we are * either the CPU handling the broadcast -- cgit v1.2.3 From 14c63f17b1fde5a575a28e96547a22b451c71fb5 Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 21 Jun 2013 08:51:36 -0700 Subject: perf: Drop sample rate when sampling is too slow This patch keeps track of how long perf's NMI handler is taking, and also calculates how many samples perf can take a second. If the sample length times the expected max number of samples exceeds a configurable threshold, it drops the sample rate. This way, we don't have a runaway sampling process eating up the CPU. This patch can tend to drop the sample rate down to level where perf doesn't work very well. *BUT* the alternative is that my system hangs because it spends all of its time handling NMIs. I'll take a busted performance tool over an entire system that's busted and undebuggable any day. BTW, my suspicion is that there's still an underlying bug here. Using the HPET instead of the TSC is definitely a contributing factor, but I suspect there are some other things going on. But, I can't go dig down on a bug like that with my machine hanging all the time. Signed-off-by: Dave Hansen Acked-by: Peter Zijlstra Cc: paulus@samba.org Cc: acme@ghostprotocols.net Cc: Dave Hansen [ Prettified it a bit. ] Signed-off-by: Ingo Molnar --- kernel/events/core.c | 92 +++++++++++++++++++++++++++++++++++++++++++++++++--- kernel/sysctl.c | 9 +++++ 2 files changed, 97 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 9c8920783317..1db3af933704 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -165,10 +165,26 @@ int sysctl_perf_event_mlock __read_mostly = 512 + (PAGE_SIZE / 1024); /* 'free' /* * max perf event sample rate */ -#define DEFAULT_MAX_SAMPLE_RATE 100000 -int sysctl_perf_event_sample_rate __read_mostly = DEFAULT_MAX_SAMPLE_RATE; -static int max_samples_per_tick __read_mostly = - DIV_ROUND_UP(DEFAULT_MAX_SAMPLE_RATE, HZ); +#define DEFAULT_MAX_SAMPLE_RATE 100000 +#define DEFAULT_SAMPLE_PERIOD_NS (NSEC_PER_SEC / DEFAULT_MAX_SAMPLE_RATE) +#define DEFAULT_CPU_TIME_MAX_PERCENT 25 + +int sysctl_perf_event_sample_rate __read_mostly = DEFAULT_MAX_SAMPLE_RATE; + +static int max_samples_per_tick __read_mostly = DIV_ROUND_UP(DEFAULT_MAX_SAMPLE_RATE, HZ); +static int perf_sample_period_ns __read_mostly = DEFAULT_SAMPLE_PERIOD_NS; + +static atomic_t perf_sample_allowed_ns __read_mostly = + ATOMIC_INIT( DEFAULT_SAMPLE_PERIOD_NS * DEFAULT_CPU_TIME_MAX_PERCENT / 100); + +void update_perf_cpu_limits(void) +{ + u64 tmp = perf_sample_period_ns; + + tmp *= sysctl_perf_cpu_time_max_percent; + tmp = do_div(tmp, 100); + atomic_set(&perf_sample_allowed_ns, tmp); +} static int perf_rotate_context(struct perf_cpu_context *cpuctx); @@ -182,10 +198,78 @@ int perf_proc_update_handler(struct ctl_table *table, int write, return ret; max_samples_per_tick = DIV_ROUND_UP(sysctl_perf_event_sample_rate, HZ); + perf_sample_period_ns = NSEC_PER_SEC / sysctl_perf_event_sample_rate; + update_perf_cpu_limits(); return 0; } +int sysctl_perf_cpu_time_max_percent __read_mostly = DEFAULT_CPU_TIME_MAX_PERCENT; + +int perf_cpu_time_max_percent_handler(struct ctl_table *table, int write, + void __user *buffer, size_t *lenp, + loff_t *ppos) +{ + int ret = proc_dointvec(table, write, buffer, lenp, ppos); + + if (ret || !write) + return ret; + + update_perf_cpu_limits(); + + return 0; +} + +/* + * perf samples are done in some very critical code paths (NMIs). + * If they take too much CPU time, the system can lock up and not + * get any real work done. This will drop the sample rate when + * we detect that events are taking too long. + */ +#define NR_ACCUMULATED_SAMPLES 128 +DEFINE_PER_CPU(u64, running_sample_length); + +void perf_sample_event_took(u64 sample_len_ns) +{ + u64 avg_local_sample_len; + u64 local_samples_len = __get_cpu_var(running_sample_length); + + if (atomic_read(&perf_sample_allowed_ns) == 0) + return; + + /* decay the counter by 1 average sample */ + local_samples_len = __get_cpu_var(running_sample_length); + local_samples_len -= local_samples_len/NR_ACCUMULATED_SAMPLES; + local_samples_len += sample_len_ns; + __get_cpu_var(running_sample_length) = local_samples_len; + + /* + * note: this will be biased artifically low until we have + * seen NR_ACCUMULATED_SAMPLES. Doing it this way keeps us + * from having to maintain a count. + */ + avg_local_sample_len = local_samples_len/NR_ACCUMULATED_SAMPLES; + + if (avg_local_sample_len <= atomic_read(&perf_sample_allowed_ns)) + return; + + if (max_samples_per_tick <= 1) + return; + + max_samples_per_tick = DIV_ROUND_UP(max_samples_per_tick, 2); + sysctl_perf_event_sample_rate = max_samples_per_tick * HZ; + perf_sample_period_ns = NSEC_PER_SEC / sysctl_perf_event_sample_rate; + + printk_ratelimited(KERN_WARNING + "perf samples too long (%lld > %d), lowering " + "kernel.perf_event_max_sample_rate to %d\n", + avg_local_sample_len, + atomic_read(&perf_sample_allowed_ns), + sysctl_perf_event_sample_rate); + + update_perf_cpu_limits(); +} + static atomic64_t perf_event_id; static void cpu_ctx_sched_out(struct perf_cpu_context *cpuctx, diff --git a/kernel/sysctl.c b/kernel/sysctl.c index b0a1f99907f3..4ce13c3cedb9 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -1043,6 +1043,15 @@ static struct ctl_table kern_table[] = { .mode = 0644, .proc_handler = perf_proc_update_handler, }, + { + .procname = "perf_cpu_time_max_percent", + .data = &sysctl_perf_cpu_time_max_percent, + .maxlen = sizeof(sysctl_perf_cpu_time_max_percent), + .mode = 0644, + .proc_handler = perf_cpu_time_max_percent_handler, + .extra1 = &zero, + .extra2 = &one_hundred, + }, #endif #ifdef CONFIG_KMEMCHECK { -- cgit v1.2.3 From 247e9ee034b0448a585afa16e292cbb9dc0aef68 Mon Sep 17 00:00:00 2001 From: Sahara Date: Fri, 21 Jun 2013 11:12:28 +0900 Subject: PM / QoS: Add pm_qos_update_target/flags tracepoints This patch adds tracepoints to pm_qos_update_target and pm_qos_update_flags. It's useful for checking pm qos action, previous value and current value. Signed-off-by: Sahara Signed-off-by: Rafael J. Wysocki --- kernel/power/qos.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'kernel') diff --git a/kernel/power/qos.c b/kernel/power/qos.c index f2f5f6e22a3c..4fb8d1427938 100644 --- a/kernel/power/qos.c +++ b/kernel/power/qos.c @@ -44,6 +44,7 @@ #include #include +#include /* * locking rule: all changes to constraints or notifiers lists @@ -202,6 +203,7 @@ int pm_qos_update_target(struct pm_qos_constraints *c, struct plist_node *node, spin_unlock_irqrestore(&pm_qos_lock, flags); + trace_pm_qos_update_target(action, prev_value, curr_value); if (prev_value != curr_value) { blocking_notifier_call_chain(c->notifiers, (unsigned long)curr_value, @@ -272,6 +274,7 @@ bool pm_qos_update_flags(struct pm_qos_flags *pqf, spin_unlock_irqrestore(&pm_qos_lock, irqflags); + trace_pm_qos_update_flags(action, prev_value, curr_value); return prev_value != curr_value; } -- cgit v1.2.3 From ae8822b842e229fa4459fca2d979b630d812311d Mon Sep 17 00:00:00 2001 From: Sahara Date: Fri, 21 Jun 2013 11:12:29 +0900 Subject: PM / QoS: Add pm_qos_request tracepoints Adds tracepoints to pm_qos_add_request, pm_qos_update_request, pm_qos_remove_request, and pm_qos_update_request_timeout. It's useful for checking pm_qos_class, value, and timeout_us. Signed-off-by: Sahara Signed-off-by: Rafael J. Wysocki --- kernel/power/qos.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'kernel') diff --git a/kernel/power/qos.c b/kernel/power/qos.c index 4fb8d1427938..06fe28589e9c 100644 --- a/kernel/power/qos.c +++ b/kernel/power/qos.c @@ -336,6 +336,7 @@ void pm_qos_add_request(struct pm_qos_request *req, } req->pm_qos_class = pm_qos_class; INIT_DELAYED_WORK(&req->work, pm_qos_work_fn); + trace_pm_qos_add_request(pm_qos_class, value); pm_qos_update_target(pm_qos_array[pm_qos_class]->constraints, &req->node, PM_QOS_ADD_REQ, value); } @@ -364,6 +365,7 @@ void pm_qos_update_request(struct pm_qos_request *req, cancel_delayed_work_sync(&req->work); + trace_pm_qos_update_request(req->pm_qos_class, new_value); if (new_value != req->node.prio) pm_qos_update_target( pm_qos_array[req->pm_qos_class]->constraints, @@ -390,6 +392,8 @@ void pm_qos_update_request_timeout(struct pm_qos_request *req, s32 new_value, cancel_delayed_work_sync(&req->work); + trace_pm_qos_update_request_timeout(req->pm_qos_class, + new_value, timeout_us); if (new_value != req->node.prio) pm_qos_update_target( pm_qos_array[req->pm_qos_class]->constraints, @@ -419,6 +423,7 @@ void pm_qos_remove_request(struct pm_qos_request *req) cancel_delayed_work_sync(&req->work); + trace_pm_qos_remove_request(req->pm_qos_class, PM_QOS_DEFAULT_VALUE); pm_qos_update_target(pm_qos_array[req->pm_qos_class]->constraints, &req->node, PM_QOS_REMOVE_REQ, PM_QOS_DEFAULT_VALUE); -- cgit v1.2.3 From ddaf144c61da45ae5c49ae38556c3ac4524f9318 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Mon, 10 Jun 2013 01:06:02 +0100 Subject: irqdomain: Refactor irq_domain_associate_many() Originally, irq_domain_associate_many() was designed to unwind the mapped irqs on a failure of any individual association. However, that proved to be a problem with certain IRQ controllers. Some of them only support a subset of irqs, and will fail when attempting to map a reserved IRQ. In those cases we want to map as many IRQs as possible, so instead it is better for irq_domain_associate_many() to make a best-effort attempt to map irqs, but not fail if any or all of them don't succeed. If a caller really cares about how many irqs got associated, then it should instead go back and check that all of the irqs is cares about were mapped. The original design open-coded the individual association code into the body of irq_domain_associate_many(), but with no longer needing to unwind associations, the code becomes simpler to split out irq_domain_associate() to contain the bulk of the logic, and irq_domain_associate_many() to be a simple loop wrapper. This patch also adds a new error check to the associate path to make sure it isn't called for an irq larger than the controller can handle, and adds locking so that the irq_domain_mutex is held while setting up a new association. v3: Fixup missing change to irq_domain_add_tree() v2: Fixup x86 warning. irq_domain_associate_many() no longer returns an error code, but reports errors to the printk log directly. In the majority of cases we don't actually want to fail if there is a problem, but rather log it and still try to boot the system. Signed-off-by: Grant Likely irqdomain: Fix flubbed irq_domain_associate_many refactoring commit d39046ec72, "irqdomain: Refactor irq_domain_associate_many()" was missing the following hunk which causes a boot failure on anything using irq_domain_add_tree() to allocate an irq domain. Signed-off-by: Grant Likely Cc: Michael Neuling Cc: Benjamin Herrenschmidt , Cc: Thomas Gleixner , Cc: Stephen Rothwell --- kernel/irq/irqdomain.c | 185 ++++++++++++++++++++++++------------------------- 1 file changed, 91 insertions(+), 94 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c index 280b8047d8db..80e92492c77b 100644 --- a/kernel/irq/irqdomain.c +++ b/kernel/irq/irqdomain.c @@ -35,8 +35,8 @@ static struct irq_domain *irq_default_domain; * register allocated irq_domain with irq_domain_register(). Returns pointer * to IRQ domain, or NULL on failure. */ -struct irq_domain *__irq_domain_add(struct device_node *of_node, - int size, int direct_max, +struct irq_domain *__irq_domain_add(struct device_node *of_node, int size, + irq_hw_number_t hwirq_max, int direct_max, const struct irq_domain_ops *ops, void *host_data) { @@ -52,6 +52,7 @@ struct irq_domain *__irq_domain_add(struct device_node *of_node, domain->ops = ops; domain->host_data = host_data; domain->of_node = of_node_get(of_node); + domain->hwirq_max = hwirq_max; domain->revmap_size = size; domain->revmap_direct_max_irq = direct_max; @@ -126,7 +127,7 @@ struct irq_domain *irq_domain_add_simple(struct device_node *of_node, { struct irq_domain *domain; - domain = __irq_domain_add(of_node, size, 0, ops, host_data); + domain = __irq_domain_add(of_node, size, size, 0, ops, host_data); if (!domain) return NULL; @@ -139,7 +140,7 @@ struct irq_domain *irq_domain_add_simple(struct device_node *of_node, pr_info("Cannot allocate irq_descs @ IRQ%d, assuming pre-allocated\n", first_irq); } - WARN_ON(irq_domain_associate_many(domain, first_irq, 0, size)); + irq_domain_associate_many(domain, first_irq, 0, size); } return domain; @@ -170,11 +171,12 @@ struct irq_domain *irq_domain_add_legacy(struct device_node *of_node, { struct irq_domain *domain; - domain = __irq_domain_add(of_node, first_hwirq + size, 0, ops, host_data); + domain = __irq_domain_add(of_node, first_hwirq + size, + first_hwirq + size, 0, ops, host_data); if (!domain) return NULL; - WARN_ON(irq_domain_associate_many(domain, first_irq, first_hwirq, size)); + irq_domain_associate_many(domain, first_irq, first_hwirq, size); return domain; } @@ -228,109 +230,109 @@ void irq_set_default_host(struct irq_domain *domain) } EXPORT_SYMBOL_GPL(irq_set_default_host); -static void irq_domain_disassociate_many(struct irq_domain *domain, - unsigned int irq_base, int count) +static void irq_domain_disassociate(struct irq_domain *domain, unsigned int irq) { - /* - * disassociate in reverse order; - * not strictly necessary, but nice for unwinding - */ - while (count--) { - int irq = irq_base + count; - struct irq_data *irq_data = irq_get_irq_data(irq); - irq_hw_number_t hwirq; + struct irq_data *irq_data = irq_get_irq_data(irq); + irq_hw_number_t hwirq; - if (WARN_ON(!irq_data || irq_data->domain != domain)) - continue; + if (WARN(!irq_data || irq_data->domain != domain, + "virq%i doesn't exist; cannot disassociate\n", irq)) + return; - hwirq = irq_data->hwirq; - irq_set_status_flags(irq, IRQ_NOREQUEST); + hwirq = irq_data->hwirq; + irq_set_status_flags(irq, IRQ_NOREQUEST); - /* remove chip and handler */ - irq_set_chip_and_handler(irq, NULL, NULL); + /* remove chip and handler */ + irq_set_chip_and_handler(irq, NULL, NULL); - /* Make sure it's completed */ - synchronize_irq(irq); + /* Make sure it's completed */ + synchronize_irq(irq); - /* Tell the PIC about it */ - if (domain->ops->unmap) - domain->ops->unmap(domain, irq); - smp_mb(); + /* Tell the PIC about it */ + if (domain->ops->unmap) + domain->ops->unmap(domain, irq); + smp_mb(); - irq_data->domain = NULL; - irq_data->hwirq = 0; + irq_data->domain = NULL; + irq_data->hwirq = 0; - /* Clear reverse map for this hwirq */ - if (hwirq < domain->revmap_size) { - domain->linear_revmap[hwirq] = 0; - } else { - mutex_lock(&revmap_trees_mutex); - radix_tree_delete(&domain->revmap_tree, hwirq); - mutex_unlock(&revmap_trees_mutex); - } + /* Clear reverse map for this hwirq */ + if (hwirq < domain->revmap_size) { + domain->linear_revmap[hwirq] = 0; + } else { + mutex_lock(&revmap_trees_mutex); + radix_tree_delete(&domain->revmap_tree, hwirq); + mutex_unlock(&revmap_trees_mutex); } } -int irq_domain_associate_many(struct irq_domain *domain, unsigned int irq_base, - irq_hw_number_t hwirq_base, int count) +int irq_domain_associate(struct irq_domain *domain, unsigned int virq, + irq_hw_number_t hwirq) { - unsigned int virq = irq_base; - irq_hw_number_t hwirq = hwirq_base; - int i, ret; + struct irq_data *irq_data = irq_get_irq_data(virq); + int ret; - pr_debug("%s(%s, irqbase=%i, hwbase=%i, count=%i)\n", __func__, - of_node_full_name(domain->of_node), irq_base, (int)hwirq_base, count); + if (WARN(hwirq >= domain->hwirq_max, + "error: hwirq 0x%x is too large for %s\n", (int)hwirq, domain->name)) + return -EINVAL; + if (WARN(!irq_data, "error: virq%i is not allocated", virq)) + return -EINVAL; + if (WARN(irq_data->domain, "error: virq%i is already associated", virq)) + return -EINVAL; - for (i = 0; i < count; i++) { - struct irq_data *irq_data = irq_get_irq_data(virq + i); - - if (WARN(!irq_data, "error: irq_desc not allocated; " - "irq=%i hwirq=0x%x\n", virq + i, (int)hwirq + i)) - return -EINVAL; - if (WARN(irq_data->domain, "error: irq_desc already associated; " - "irq=%i hwirq=0x%x\n", virq + i, (int)hwirq + i)) - return -EINVAL; - }; - - for (i = 0; i < count; i++, virq++, hwirq++) { - struct irq_data *irq_data = irq_get_irq_data(virq); - - irq_data->hwirq = hwirq; - irq_data->domain = domain; - if (domain->ops->map) { - ret = domain->ops->map(domain, virq, hwirq); - if (ret != 0) { - /* - * If map() returns -EPERM, this interrupt is protected - * by the firmware or some other service and shall not - * be mapped. Don't bother telling the user about it. - */ - if (ret != -EPERM) { - pr_info("%s didn't like hwirq-0x%lx to VIRQ%i mapping (rc=%d)\n", - domain->name, hwirq, virq, ret); - } - irq_data->domain = NULL; - irq_data->hwirq = 0; - continue; + mutex_lock(&irq_domain_mutex); + irq_data->hwirq = hwirq; + irq_data->domain = domain; + if (domain->ops->map) { + ret = domain->ops->map(domain, virq, hwirq); + if (ret != 0) { + /* + * If map() returns -EPERM, this interrupt is protected + * by the firmware or some other service and shall not + * be mapped. Don't bother telling the user about it. + */ + if (ret != -EPERM) { + pr_info("%s didn't like hwirq-0x%lx to VIRQ%i mapping (rc=%d)\n", + domain->name, hwirq, virq, ret); } - /* If not already assigned, give the domain the chip's name */ - if (!domain->name && irq_data->chip) - domain->name = irq_data->chip->name; + irq_data->domain = NULL; + irq_data->hwirq = 0; + mutex_unlock(&irq_domain_mutex); + return ret; } - if (hwirq < domain->revmap_size) { - domain->linear_revmap[hwirq] = virq; - } else { - mutex_lock(&revmap_trees_mutex); - radix_tree_insert(&domain->revmap_tree, hwirq, irq_data); - mutex_unlock(&revmap_trees_mutex); - } + /* If not already assigned, give the domain the chip's name */ + if (!domain->name && irq_data->chip) + domain->name = irq_data->chip->name; + } - irq_clear_status_flags(virq, IRQ_NOREQUEST); + if (hwirq < domain->revmap_size) { + domain->linear_revmap[hwirq] = virq; + } else { + mutex_lock(&revmap_trees_mutex); + radix_tree_insert(&domain->revmap_tree, hwirq, irq_data); + mutex_unlock(&revmap_trees_mutex); } + mutex_unlock(&irq_domain_mutex); + + irq_clear_status_flags(virq, IRQ_NOREQUEST); return 0; } +EXPORT_SYMBOL_GPL(irq_domain_associate); + +void irq_domain_associate_many(struct irq_domain *domain, unsigned int irq_base, + irq_hw_number_t hwirq_base, int count) +{ + int i; + + pr_debug("%s(%s, irqbase=%i, hwbase=%i, count=%i)\n", __func__, + of_node_full_name(domain->of_node), irq_base, (int)hwirq_base, count); + + for (i = 0; i < count; i++) { + irq_domain_associate(domain, irq_base + i, hwirq_base + i); + } +} EXPORT_SYMBOL_GPL(irq_domain_associate_many); /** @@ -460,12 +462,7 @@ int irq_create_strict_mappings(struct irq_domain *domain, unsigned int irq_base, if (unlikely(ret < 0)) return ret; - ret = irq_domain_associate_many(domain, irq_base, hwirq_base, count); - if (unlikely(ret < 0)) { - irq_free_descs(irq_base, count); - return ret; - } - + irq_domain_associate_many(domain, irq_base, hwirq_base, count); return 0; } EXPORT_SYMBOL_GPL(irq_create_strict_mappings); @@ -535,7 +532,7 @@ void irq_dispose_mapping(unsigned int virq) if (WARN_ON(domain == NULL)) return; - irq_domain_disassociate_many(domain, virq, 1); + irq_domain_disassociate(domain, virq); irq_free_desc(virq); } EXPORT_SYMBOL_GPL(irq_dispose_mapping); -- cgit v1.2.3 From 56a3d5ac774d054ece9373277a861338a468a294 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Mon, 10 Jun 2013 01:09:33 +0100 Subject: irqdomain: remove irq_domain_generate_simple() Nobody calls it; remove the function Signed-off-by: Grant Likely --- kernel/irq/irqdomain.c | 15 --------------- 1 file changed, 15 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c index 80e92492c77b..e47b35671384 100644 --- a/kernel/irq/irqdomain.c +++ b/kernel/irq/irqdomain.c @@ -741,18 +741,3 @@ const struct irq_domain_ops irq_domain_simple_ops = { .xlate = irq_domain_xlate_onetwocell, }; EXPORT_SYMBOL_GPL(irq_domain_simple_ops); - -#ifdef CONFIG_OF_IRQ -void irq_domain_generate_simple(const struct of_device_id *match, - u64 phys_base, unsigned int irq_start) -{ - struct device_node *node; - pr_debug("looking for phys_base=%llx, irq_start=%i\n", - (unsigned long long) phys_base, (int) irq_start); - node = of_find_matching_node_by_address(NULL, match, phys_base); - if (node) - irq_domain_add_legacy(node, 32, irq_start, 0, - &irq_domain_simple_ops, NULL); -} -EXPORT_SYMBOL_GPL(irq_domain_generate_simple); -#endif -- cgit v1.2.3 From d3dcb436f61593843af178d4a520c8c43c04d3fc Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Mon, 10 Jun 2013 12:19:17 +0100 Subject: irqdomain: make irq_linear_revmap() a fast path again Over the years, irq_linear_revmap() gained tests and checks to make sure callers were using it safely, which while important, also make it less of a fast path. After the irqdomain refactoring done recently, it is now possible to make irq_linear_revmap() a fast path again. This patch moves irq_linear_revmap() to the header file and makes it a static inline so that interrupt controller drivers using a linear mapping can decode the virq from a hwirq in just a couple of instructions. Signed-off-by: Grant Likely --- kernel/irq/irqdomain.c | 34 ++++++++-------------------------- 1 file changed, 8 insertions(+), 26 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c index e47b35671384..836a0f7ec2a9 100644 --- a/kernel/irq/irqdomain.c +++ b/kernel/irq/irqdomain.c @@ -559,35 +559,17 @@ unsigned int irq_find_mapping(struct irq_domain *domain, return hwirq; } - return irq_linear_revmap(domain, hwirq); + /* Check if the hwirq is in the linear revmap. */ + if (hwirq < domain->revmap_size) + return domain->linear_revmap[hwirq]; + + rcu_read_lock(); + data = radix_tree_lookup(&domain->revmap_tree, hwirq); + rcu_read_unlock(); + return data ? data->irq : 0; } EXPORT_SYMBOL_GPL(irq_find_mapping); -/** - * irq_linear_revmap() - Find a linux irq from a hw irq number. - * @domain: domain owning this hardware interrupt - * @hwirq: hardware irq number in that domain space - * - * This is a fast path that can be called directly by irq controller code to - * save a handful of instructions. - */ -unsigned int irq_linear_revmap(struct irq_domain *domain, - irq_hw_number_t hwirq) -{ - struct irq_data *data; - - /* Check revmap bounds; complain if exceeded */ - if (hwirq >= domain->revmap_size) { - rcu_read_lock(); - data = radix_tree_lookup(&domain->revmap_tree, hwirq); - rcu_read_unlock(); - return data ? data->irq : 0; - } - - return domain->linear_revmap[hwirq]; -} -EXPORT_SYMBOL_GPL(irq_linear_revmap); - #ifdef CONFIG_IRQ_DOMAIN_DEBUG static int virq_debug_show(struct seq_file *m, void *private) { -- cgit v1.2.3 From c12d2f42a96d72cffa4d9335ca455a2243333c79 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Thu, 26 Jan 2012 16:29:19 -0700 Subject: irqdomain: Include hwirq number in /proc/interrupts Add the hardware interrupt number to the output of /proc/interrupts. It is often important to have access to the hardware interrupt number because it identifies exactly how an interrupt signal is wired up to the interrupt controller. This is especially important when using irq_domains since irq numbers get dynamically allocated in that case, and have no relation to the actual hardware number. Note: This output is currently conditional on whether or not the irq_domain pointer is set; however hwirq could still be used without irq_domain. It may be worthwhile to always output the hwirq number regardless of the domain pointer. Signed-off-by: Grant Likely Tested-by: Olof Johansson Cc: Ben Herrenschmidt Cc: Thomas Gleixner --- kernel/irq/proc.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'kernel') diff --git a/kernel/irq/proc.c b/kernel/irq/proc.c index 19ed5c425c3b..36f6ee181b0c 100644 --- a/kernel/irq/proc.c +++ b/kernel/irq/proc.c @@ -462,6 +462,8 @@ int show_interrupts(struct seq_file *p, void *v) } else { seq_printf(p, " %8s", "None"); } + if (desc->irq_data.domain) + seq_printf(p, " %*d", prec, (int) desc->irq_data.hwirq); #ifdef CONFIG_GENERIC_IRQ_SHOW_LEVEL seq_printf(p, " %-8s", irqd_is_level_type(&desc->irq_data) ? "Level" : "Edge"); #endif -- cgit v1.2.3 From 798f0fd188be3656991c8745104b5ee045769a5f Mon Sep 17 00:00:00 2001 From: Kefeng Wang Date: Thu, 6 Jun 2013 19:20:27 +0800 Subject: irq: fix checkpatch error ERROR: space required before the open parenthesis '(' WARNING: Prefer pr_warn(... to pr_warning(... Just fix above 2 issue. Signed-off-by: Kefeng Wang Signed-off-by: Grant Likely --- kernel/irq/irqdomain.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c index 836a0f7ec2a9..13f265430c25 100644 --- a/kernel/irq/irqdomain.c +++ b/kernel/irq/irqdomain.c @@ -396,9 +396,7 @@ unsigned int irq_create_mapping(struct irq_domain *domain, if (domain == NULL) domain = irq_default_domain; if (domain == NULL) { - pr_warning("irq_create_mapping called for" - " NULL domain, hwirq=%lx\n", hwirq); - WARN_ON(1); + WARN(1, "%s(, %lx) called with NULL domain\n", __func__, hwirq); return 0; } pr_debug("-> using domain @%p\n", domain); @@ -489,8 +487,8 @@ unsigned int irq_create_of_mapping(struct device_node *controller, if (intsize > 0) return intspec[0]; #endif - pr_warning("no irq domain found for %s !\n", - of_node_full_name(controller)); + pr_warn("no irq domain found for %s !\n", + of_node_full_name(controller)); return 0; } -- cgit v1.2.3 From 6fff8314046276331314ae32cea34c6d11c440d2 Mon Sep 17 00:00:00 2001 From: James Hogan Date: Tue, 18 Jun 2013 15:08:33 +0100 Subject: genirq: Irqchip: document gcflags arg of irq_alloc_domain_generic_chips Commit 088f40b7b027dad6519712ff224a5798dd62a204 ("genirq: Generic chip: Add linear irq domain support") missed kerneldoc for the gcflags argument of irq_alloc_domain_generic_chips(). Add it now. Signed-off-by: James Hogan Acked-by: Grant Likely Link: http://lkml.kernel.org/r/1371564513-4327-1-git-send-email-james.hogan@imgtec.com Signed-off-by: Thomas Gleixner --- kernel/irq/generic-chip.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/irq/generic-chip.c b/kernel/irq/generic-chip.c index 95575d8d5392..a746a8f54dae 100644 --- a/kernel/irq/generic-chip.c +++ b/kernel/irq/generic-chip.c @@ -254,6 +254,7 @@ irq_gc_init_mask_cache(struct irq_chip_generic *gc, enum irq_gc_flags flags) * @handler: Default flow handler associated with these chips * @clr: IRQ_* bits to clear in the mapping function * @set: IRQ_* bits to set in the mapping function + * @gcflags: Generic chip specific setup flags */ int irq_alloc_domain_generic_chips(struct irq_domain *d, int irqs_per_chip, int num_ct, const char *name, -- cgit v1.2.3 From 70e5975d3a04be5479a28eec4a2fb10f98ad2785 Mon Sep 17 00:00:00 2001 From: Stephen Boyd Date: Thu, 13 Jun 2013 11:39:50 -0700 Subject: clockevents: Prefer CPU local devices over global devices On an SMP system with only one global clockevent and a dummy clockevent per CPU we run into problems. We want the dummy clockevents to be registered as the per CPU tick devices, but we can only achieve that if we register the dummy clockevents before the global clockevent or if we artificially inflate the rating of the dummy clockevents to be higher than the rating of the global clockevent. Failure to do so leads to boot hangs when the dummy timers are registered on all other CPUs besides the CPU that accepted the global clockevent as its tick device and there is no broadcast timer to poke the dummy devices. If we're registering multiple clockevents and one clockevent is global and the other is local to a particular CPU we should choose to use the local clockevent regardless of the rating of the device. This way, if the clockevent is a dummy it will take the tick device duty as long as there isn't a higher rated tick device and any global clockevent will be bumped out into broadcast mode, fixing the problem described above. Reported-and-tested-by: Mark Rutland Signed-off-by: Stephen Boyd Tested-by: soren.brinkmann@xilinx.com Cc: John Stultz Cc: Daniel Lezcano Cc: linux-arm-kernel@lists.infradead.org Cc: John Stultz Link: http://lkml.kernel.org/r/20130613183950.GA32061@codeaurora.org Signed-off-by: Thomas Gleixner --- kernel/time/tick-common.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-common.c b/kernel/time/tick-common.c index 5edfb4806032..edd45f64162f 100644 --- a/kernel/time/tick-common.c +++ b/kernel/time/tick-common.c @@ -243,8 +243,13 @@ static bool tick_check_preferred(struct clock_event_device *curdev, return false; } - /* Use the higher rated one */ - return !curdev || newdev->rating > curdev->rating; + /* + * Use the higher rated one, but prefer a CPU local device with a lower + * rating than a non-CPU local device + */ + return !curdev || + newdev->rating > curdev->rating || + !cpumask_equal(curdev->cpumask, newdev->cpumask); } /* -- cgit v1.2.3 From 9871bf9550d25e488cd2f0ce958d3f59f17fa720 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 24 Jun 2013 15:21:47 -0700 Subject: cgroup: prefix global variables with "cgroup_" Global variable names in kernel/cgroup.c are asking for trouble - subsys, roots, rootnode and so on. Rename them to have "cgroup_" prefix. * s/subsys/cgroup_subsys/ * s/rootnode/cgroup_dummy_root/ * s/dummytop/cgroup_cummy_top/ * s/roots/cgroup_roots/ * s/root_count/cgroup_root_count/ This patch is purely cosmetic. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 153 ++++++++++++++++++++++++++++---------------------------- 1 file changed, 77 insertions(+), 76 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 1051c1f69674..8f296b83b6a3 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -96,16 +96,19 @@ static DEFINE_MUTEX(cgroup_root_mutex); */ #define SUBSYS(_x) [_x ## _subsys_id] = &_x ## _subsys, #define IS_SUBSYS_ENABLED(option) IS_BUILTIN(option) -static struct cgroup_subsys *subsys[CGROUP_SUBSYS_COUNT] = { +static struct cgroup_subsys *cgroup_subsys[CGROUP_SUBSYS_COUNT] = { #include }; /* - * The "rootnode" hierarchy is the "dummy hierarchy", reserved for the - * subsystems that are otherwise unattached - it never has more than a - * single cgroup, and all tasks are part of that cgroup. + * The dummy hierarchy, reserved for the subsystems that are otherwise + * unattached - it never has more than a single cgroup, and all tasks are + * part of that cgroup. */ -static struct cgroupfs_root rootnode; +static struct cgroupfs_root cgroup_dummy_root; + +/* dummy_top is a shorthand for the dummy hierarchy's top cgroup */ +static struct cgroup * const cgroup_dummy_top = &cgroup_dummy_root.top_cgroup; /* * cgroupfs file entry, pointed to from leaf dentry->d_fsdata. @@ -183,8 +186,8 @@ struct cgroup_event { /* The list of hierarchy roots */ -static LIST_HEAD(roots); -static int root_count; +static LIST_HEAD(cgroup_roots); +static int cgroup_root_count; /* * Hierarchy ID allocation and mapping. It follows the same exclusion @@ -193,9 +196,6 @@ static int root_count; */ static DEFINE_IDR(cgroup_hierarchy_idr); -/* dummytop is a shorthand for the dummy hierarchy's top cgroup */ -#define dummytop (&rootnode.top_cgroup) - static struct cgroup_name root_cgroup_name = { .name = "/" }; /* @@ -268,7 +268,7 @@ list_for_each_entry(_ss, &_root->subsys_list, sibling) /* for_each_active_root() allows you to iterate across the active hierarchies */ #define for_each_active_root(_root) \ -list_for_each_entry(_root, &roots, root_list) +list_for_each_entry(_root, &cgroup_roots, root_list) static inline struct cgroup *__d_cgrp(struct dentry *dentry) { @@ -650,7 +650,7 @@ static struct css_set *find_css_set(struct css_set *old_cset, return NULL; /* Allocate all the cgrp_cset_link objects that we'll need */ - if (allocate_cgrp_cset_links(root_count, &tmp_links) < 0) { + if (allocate_cgrp_cset_links(cgroup_root_count, &tmp_links) < 0) { kfree(cset); return NULL; } @@ -1000,7 +1000,7 @@ static int rebind_subsystems(struct cgroupfs_root *root, /* Check that any added subsystems are currently free */ for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { unsigned long bit = 1UL << i; - struct cgroup_subsys *ss = subsys[i]; + struct cgroup_subsys *ss = cgroup_subsys[i]; if (!(bit & added_mask)) continue; /* @@ -1009,7 +1009,7 @@ static int rebind_subsystems(struct cgroupfs_root *root, * ensure that subsystems won't disappear once selected. */ BUG_ON(ss == NULL); - if (ss->root != &rootnode) { + if (ss->root != &cgroup_dummy_root) { /* Subsystem isn't free */ return -EBUSY; } @@ -1024,15 +1024,15 @@ static int rebind_subsystems(struct cgroupfs_root *root, /* Process each subsystem */ for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = subsys[i]; + struct cgroup_subsys *ss = cgroup_subsys[i]; unsigned long bit = 1UL << i; if (bit & added_mask) { /* We're binding this subsystem to this hierarchy */ BUG_ON(ss == NULL); BUG_ON(cgrp->subsys[i]); - BUG_ON(!dummytop->subsys[i]); - BUG_ON(dummytop->subsys[i]->cgroup != dummytop); - cgrp->subsys[i] = dummytop->subsys[i]; + BUG_ON(!cgroup_dummy_top->subsys[i]); + BUG_ON(cgroup_dummy_top->subsys[i]->cgroup != cgroup_dummy_top); + cgrp->subsys[i] = cgroup_dummy_top->subsys[i]; cgrp->subsys[i]->cgroup = cgrp; list_move(&ss->sibling, &root->subsys_list); ss->root = root; @@ -1042,14 +1042,14 @@ static int rebind_subsystems(struct cgroupfs_root *root, } else if (bit & removed_mask) { /* We're removing this subsystem */ BUG_ON(ss == NULL); - BUG_ON(cgrp->subsys[i] != dummytop->subsys[i]); + BUG_ON(cgrp->subsys[i] != cgroup_dummy_top->subsys[i]); BUG_ON(cgrp->subsys[i]->cgroup != cgrp); if (ss->bind) - ss->bind(dummytop); - dummytop->subsys[i]->cgroup = dummytop; + ss->bind(cgroup_dummy_top); + cgroup_dummy_top->subsys[i]->cgroup = cgroup_dummy_top; cgrp->subsys[i] = NULL; - subsys[i]->root = &rootnode; - list_move(&ss->sibling, &rootnode.subsys_list); + cgroup_subsys[i]->root = &cgroup_dummy_root; + list_move(&ss->sibling, &cgroup_dummy_root.subsys_list); /* subsystem is now free - drop reference on module */ module_put(ss->module); } else if (bit & final_subsys_mask) { @@ -1112,10 +1112,10 @@ struct cgroup_sb_opts { }; /* - * Convert a hierarchy specifier into a bitmask of subsystems and flags. Call - * with cgroup_mutex held to protect the subsys[] array. This function takes - * refcounts on subsystems to be used, unless it returns error, in which case - * no refcounts are taken. + * Convert a hierarchy specifier into a bitmask of subsystems and + * flags. Call with cgroup_mutex held to protect the cgroup_subsys[] + * array. This function takes refcounts on subsystems to be used, unless it + * returns error, in which case no refcounts are taken. */ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) { @@ -1201,7 +1201,7 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) } for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = subsys[i]; + struct cgroup_subsys *ss = cgroup_subsys[i]; if (ss == NULL) continue; if (strcmp(token, ss->name)) @@ -1228,7 +1228,7 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) */ if (all_ss || (!one_ss && !opts->none && !opts->name)) { for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = subsys[i]; + struct cgroup_subsys *ss = cgroup_subsys[i]; if (ss == NULL) continue; if (ss->disabled) @@ -1284,7 +1284,7 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) if (!(bit & opts->subsys_mask)) continue; - if (!try_module_get(subsys[i]->module)) { + if (!try_module_get(cgroup_subsys[i]->module)) { module_pin_failed = true; break; } @@ -1301,7 +1301,7 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) if (!(bit & opts->subsys_mask)) continue; - module_put(subsys[i]->module); + module_put(cgroup_subsys[i]->module); } return -ENOENT; } @@ -1317,7 +1317,7 @@ static void drop_parsed_module_refcounts(unsigned long subsys_mask) if (!(bit & subsys_mask)) continue; - module_put(subsys[i]->module); + module_put(cgroup_subsys[i]->module); } } @@ -1648,8 +1648,8 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, /* EBUSY should be the only error here */ BUG_ON(ret); - list_add(&root->root_list, &roots); - root_count++; + list_add(&root->root_list, &cgroup_roots); + cgroup_root_count++; sb->s_root->d_fsdata = root_cgrp; root->top_cgroup.dentry = sb->s_root; @@ -1746,7 +1746,7 @@ static void cgroup_kill_sb(struct super_block *sb) { if (!list_empty(&root->root_list)) { list_del(&root->root_list); - root_count--; + cgroup_root_count--; } cgroup_exit_root_id(root); @@ -2807,7 +2807,7 @@ static void cgroup_cfts_commit(struct cgroup_subsys *ss, u64 update_before; /* %NULL @cfts indicates abort and don't bother if @ss isn't attached */ - if (!cfts || ss->root == &rootnode || + if (!cfts || ss->root == &cgroup_dummy_root || !atomic_inc_not_zero(&sb->s_active)) { mutex_unlock(&cgroup_mutex); return; @@ -4186,7 +4186,7 @@ static void init_cgroup_css(struct cgroup_subsys_state *css, css->cgroup = cgrp; css->flags = 0; css->id = NULL; - if (cgrp == dummytop) + if (cgrp == cgroup_dummy_top) css->flags |= CSS_ROOT; BUG_ON(cgrp->subsys[ss->subsys_id]); cgrp->subsys[ss->subsys_id] = css; @@ -4615,12 +4615,12 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss) cgroup_init_cftsets(ss); /* Create the top cgroup state for this subsystem */ - list_add(&ss->sibling, &rootnode.subsys_list); - ss->root = &rootnode; - css = ss->css_alloc(dummytop); + list_add(&ss->sibling, &cgroup_dummy_root.subsys_list); + ss->root = &cgroup_dummy_root; + css = ss->css_alloc(cgroup_dummy_top); /* We don't handle early failures gracefully */ BUG_ON(IS_ERR(css)); - init_cgroup_css(css, ss, dummytop); + init_cgroup_css(css, ss, cgroup_dummy_top); /* Update the init_css_set to contain a subsys * pointer to this state - since the subsystem is @@ -4635,7 +4635,7 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss) * need to invoke fork callbacks here. */ BUG_ON(!list_empty(&init_task.tasks)); - BUG_ON(online_css(ss, dummytop)); + BUG_ON(online_css(ss, cgroup_dummy_top)); mutex_unlock(&cgroup_mutex); @@ -4681,7 +4681,7 @@ int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss) */ if (ss->module == NULL) { /* a sanity check */ - BUG_ON(subsys[ss->subsys_id] != ss); + BUG_ON(cgroup_subsys[ss->subsys_id] != ss); return 0; } @@ -4689,26 +4689,26 @@ int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss) cgroup_init_cftsets(ss); mutex_lock(&cgroup_mutex); - subsys[ss->subsys_id] = ss; + cgroup_subsys[ss->subsys_id] = ss; /* * no ss->css_alloc seems to need anything important in the ss - * struct, so this can happen first (i.e. before the rootnode + * struct, so this can happen first (i.e. before the dummy root * attachment). */ - css = ss->css_alloc(dummytop); + css = ss->css_alloc(cgroup_dummy_top); if (IS_ERR(css)) { - /* failure case - need to deassign the subsys[] slot. */ - subsys[ss->subsys_id] = NULL; + /* failure case - need to deassign the cgroup_subsys[] slot. */ + cgroup_subsys[ss->subsys_id] = NULL; mutex_unlock(&cgroup_mutex); return PTR_ERR(css); } - list_add(&ss->sibling, &rootnode.subsys_list); - ss->root = &rootnode; + list_add(&ss->sibling, &cgroup_dummy_root.subsys_list); + ss->root = &cgroup_dummy_root; /* our new subsystem will be attached to the dummy hierarchy. */ - init_cgroup_css(css, ss, dummytop); + init_cgroup_css(css, ss, cgroup_dummy_top); /* init_idr must be after init_cgroup_css because it sets css->id. */ if (ss->use_id) { ret = cgroup_init_idr(ss, css); @@ -4739,7 +4739,7 @@ int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss) } write_unlock(&css_set_lock); - ret = online_css(ss, dummytop); + ret = online_css(ss, cgroup_dummy_top); if (ret) goto err_unload; @@ -4774,27 +4774,28 @@ void cgroup_unload_subsys(struct cgroup_subsys *ss) * try_module_get in parse_cgroupfs_options should ensure that it * doesn't start being used while we're killing it off. */ - BUG_ON(ss->root != &rootnode); + BUG_ON(ss->root != &cgroup_dummy_root); mutex_lock(&cgroup_mutex); - offline_css(ss, dummytop); + offline_css(ss, cgroup_dummy_top); if (ss->use_id) idr_destroy(&ss->idr); /* deassign the subsys_id */ - subsys[ss->subsys_id] = NULL; + cgroup_subsys[ss->subsys_id] = NULL; - /* remove subsystem from rootnode's list of subsystems */ + /* remove subsystem from the dummy root's list of subsystems */ list_del_init(&ss->sibling); /* - * disentangle the css from all css_sets attached to the dummytop. as - * in loading, we need to pay our respects to the hashtable gods. + * disentangle the css from all css_sets attached to the dummy + * top. as in loading, we need to pay our respects to the hashtable + * gods. */ write_lock(&css_set_lock); - list_for_each_entry(link, &dummytop->cset_links, cset_link) { + list_for_each_entry(link, &cgroup_dummy_top->cset_links, cset_link) { struct css_set *cset = link->cset; unsigned long key; @@ -4806,13 +4807,13 @@ void cgroup_unload_subsys(struct cgroup_subsys *ss) write_unlock(&css_set_lock); /* - * remove subsystem's css from the dummytop and free it - need to - * free before marking as null because ss->css_free needs the - * cgrp->subsys pointer to find their state. note that this also - * takes care of freeing the css_id. + * remove subsystem's css from the cgroup_dummy_top and free it - + * need to free before marking as null because ss->css_free needs + * the cgrp->subsys pointer to find their state. note that this + * also takes care of freeing the css_id. */ - ss->css_free(dummytop); - dummytop->subsys[ss->subsys_id] = NULL; + ss->css_free(cgroup_dummy_top); + cgroup_dummy_top->subsys[ss->subsys_id] = NULL; mutex_unlock(&cgroup_mutex); } @@ -4832,17 +4833,17 @@ int __init cgroup_init_early(void) INIT_LIST_HEAD(&init_css_set.tasks); INIT_HLIST_NODE(&init_css_set.hlist); css_set_count = 1; - init_cgroup_root(&rootnode); - root_count = 1; + init_cgroup_root(&cgroup_dummy_root); + cgroup_root_count = 1; init_task.cgroups = &init_css_set; init_cgrp_cset_link.cset = &init_css_set; - init_cgrp_cset_link.cgrp = dummytop; - list_add(&init_cgrp_cset_link.cset_link, &rootnode.top_cgroup.cset_links); + init_cgrp_cset_link.cgrp = cgroup_dummy_top; + list_add(&init_cgrp_cset_link.cset_link, &cgroup_dummy_top->cset_links); list_add(&init_cgrp_cset_link.cgrp_link, &init_css_set.cgrp_links); for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = subsys[i]; + struct cgroup_subsys *ss = cgroup_subsys[i]; /* at bootup time, we don't worry about modular subsystems */ if (!ss || ss->module) @@ -4881,7 +4882,7 @@ int __init cgroup_init(void) return err; for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = subsys[i]; + struct cgroup_subsys *ss = cgroup_subsys[i]; /* at bootup time, we don't worry about modular subsystems */ if (!ss || ss->module) @@ -4900,7 +4901,7 @@ int __init cgroup_init(void) mutex_lock(&cgroup_mutex); mutex_lock(&cgroup_root_mutex); - BUG_ON(cgroup_init_root_id(&rootnode)); + BUG_ON(cgroup_init_root_id(&cgroup_dummy_root)); mutex_unlock(&cgroup_root_mutex); mutex_unlock(&cgroup_mutex); @@ -5004,7 +5005,7 @@ static int proc_cgroupstats_show(struct seq_file *m, void *v) */ mutex_lock(&cgroup_mutex); for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = subsys[i]; + struct cgroup_subsys *ss = cgroup_subsys[i]; if (ss == NULL) continue; seq_printf(m, "%s\t%d\t%d\t%d\n", @@ -5101,7 +5102,7 @@ void cgroup_post_fork(struct task_struct *child) * can't touch that. */ for (i = 0; i < CGROUP_BUILTIN_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = subsys[i]; + struct cgroup_subsys *ss = cgroup_subsys[i]; if (ss->fork) ss->fork(child); @@ -5172,7 +5173,7 @@ void cgroup_exit(struct task_struct *tsk, int run_callbacks) * subsystems, see cgroup_post_fork() for details. */ for (i = 0; i < CGROUP_BUILTIN_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = subsys[i]; + struct cgroup_subsys *ss = cgroup_subsys[i]; if (ss->exit) { struct cgroup *old_cgrp = @@ -5291,7 +5292,7 @@ static int __init cgroup_disable(char *str) if (!*token) continue; for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = subsys[i]; + struct cgroup_subsys *ss = cgroup_subsys[i]; /* * cgroup_disable, being at boot time, can't -- cgit v1.2.3 From a8a648c4acee2095262f7fa65b0d8a68a03c32e4 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 24 Jun 2013 15:21:47 -0700 Subject: cgroup: remove cgroup->actual_subsys_mask cgroup curiously has two subsystem masks, ->subsys_mask and ->actual_subsys_mask. The latter only exists because the new target subsys_mask is passed into rebind_subsystems() via @root>subsys_mask. rebind_subsystems() needs to know what the current mask is to decide how to reach the target mask so ->actual_subsys_mask is used as the temp location to remember the current state. Adding a temporary field to a permanent data structure is rather silly and can be misleading. Update rebind_subsystems() to take @added_mask and @removed_mask instead and remove @root->actual_subsys_mask. This patch shouldn't introduce any behavior changes. v2: Comment and description updated as suggested by Li. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 22 ++++++++++++---------- 1 file changed, 12 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 8f296b83b6a3..67fc953c816a 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -986,17 +986,14 @@ static void cgroup_d_remove_dir(struct dentry *dentry) * returns an error, no reference counts are touched. */ static int rebind_subsystems(struct cgroupfs_root *root, - unsigned long final_subsys_mask) + unsigned long added_mask, unsigned removed_mask) { - unsigned long added_mask, removed_mask; struct cgroup *cgrp = &root->top_cgroup; int i; BUG_ON(!mutex_is_locked(&cgroup_mutex)); BUG_ON(!mutex_is_locked(&cgroup_root_mutex)); - removed_mask = root->actual_subsys_mask & ~final_subsys_mask; - added_mask = final_subsys_mask & ~root->actual_subsys_mask; /* Check that any added subsystems are currently free */ for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { unsigned long bit = 1UL << i; @@ -1032,27 +1029,33 @@ static int rebind_subsystems(struct cgroupfs_root *root, BUG_ON(cgrp->subsys[i]); BUG_ON(!cgroup_dummy_top->subsys[i]); BUG_ON(cgroup_dummy_top->subsys[i]->cgroup != cgroup_dummy_top); + cgrp->subsys[i] = cgroup_dummy_top->subsys[i]; cgrp->subsys[i]->cgroup = cgrp; list_move(&ss->sibling, &root->subsys_list); ss->root = root; if (ss->bind) ss->bind(cgrp); + /* refcount was already taken, and we're keeping it */ + root->subsys_mask |= bit; } else if (bit & removed_mask) { /* We're removing this subsystem */ BUG_ON(ss == NULL); BUG_ON(cgrp->subsys[i] != cgroup_dummy_top->subsys[i]); BUG_ON(cgrp->subsys[i]->cgroup != cgrp); + if (ss->bind) ss->bind(cgroup_dummy_top); cgroup_dummy_top->subsys[i]->cgroup = cgroup_dummy_top; cgrp->subsys[i] = NULL; cgroup_subsys[i]->root = &cgroup_dummy_root; list_move(&ss->sibling, &cgroup_dummy_root.subsys_list); + /* subsystem is now free - drop reference on module */ module_put(ss->module); - } else if (bit & final_subsys_mask) { + root->subsys_mask &= ~bit; + } else if (bit & root->subsys_mask) { /* Subsystem state should already exist */ BUG_ON(ss == NULL); BUG_ON(!cgrp->subsys[i]); @@ -1069,7 +1072,6 @@ static int rebind_subsystems(struct cgroupfs_root *root, BUG_ON(cgrp->subsys[i]); } } - root->subsys_mask = root->actual_subsys_mask = final_subsys_mask; return 0; } @@ -1343,7 +1345,7 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) if (ret) goto out_unlock; - if (opts.subsys_mask != root->actual_subsys_mask || opts.release_agent) + if (opts.subsys_mask != root->subsys_mask || opts.release_agent) pr_warning("cgroup: option changes via remount are deprecated (pid=%d comm=%s)\n", task_tgid_nr(current), current->comm); @@ -1365,7 +1367,7 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) */ cgroup_clear_directory(cgrp->dentry, false, removed_mask); - ret = rebind_subsystems(root, opts.subsys_mask); + ret = rebind_subsystems(root, added_mask, removed_mask); if (ret) { /* rebind_subsystems failed, re-populate the removed files */ cgroup_populate_dir(cgrp, false, removed_mask); @@ -1634,7 +1636,7 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, if (ret) goto unlock_drop; - ret = rebind_subsystems(root, root->subsys_mask); + ret = rebind_subsystems(root, root->subsys_mask, 0); if (ret == -EBUSY) { free_cgrp_cset_links(&tmp_links); goto unlock_drop; @@ -1727,7 +1729,7 @@ static void cgroup_kill_sb(struct super_block *sb) { mutex_lock(&cgroup_root_mutex); /* Rebind all subsystems back to the default hierarchy */ - ret = rebind_subsystems(root, 0); + ret = rebind_subsystems(root, 0, root->subsys_mask); /* Shouldn't be able to fail ... */ BUG_ON(ret); -- cgit v1.2.3 From b326f9d0dbd066b0aafbe88e6011a680a36de6e8 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 24 Jun 2013 15:21:48 -0700 Subject: cgroup: clean up find_css_set() and friends find_css_set() passes uninitialized on-stack template[] array to find_existing_css_set() which sets the entries for all subsystems. Passing around an uninitialized array is a bit icky and we want to introduce an iterator which only iterates loaded subsystems. Let's initialize it on definition. While at it, also make the following cosmetic cleanups. * Convert to proper /** comments. * Reorder variable declarations. * Replace comment on synchronization with lockdep_assert_held(). This patch doesn't make any functional differences. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 38 +++++++++++++++++--------------------- 1 file changed, 17 insertions(+), 21 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 67fc953c816a..c8d3175c429c 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -434,7 +434,7 @@ static inline void put_css_set_taskexit(struct css_set *cset) __put_css_set(cset, 1); } -/* +/** * compare_css_sets - helper function for find_existing_css_set(). * @cset: candidate css_set being tested * @old_cset: existing css_set for a task @@ -506,27 +506,20 @@ static bool compare_css_sets(struct css_set *cset, return true; } -/* - * find_existing_css_set() is a helper for - * find_css_set(), and checks to see whether an existing - * css_set is suitable. - * - * oldcg: the cgroup group that we're using before the cgroup - * transition - * - * cgrp: the cgroup that we're moving into - * - * template: location in which to build the desired set of subsystem - * state objects for the new cgroup group +/** + * find_existing_css_set - init css array and find the matching css_set + * @old_cset: the css_set that we're using before the cgroup transition + * @cgrp: the cgroup that we're moving into + * @template: out param for the new set of csses, should be clear on entry */ static struct css_set *find_existing_css_set(struct css_set *old_cset, struct cgroup *cgrp, struct cgroup_subsys_state *template[]) { - int i; struct cgroupfs_root *root = cgrp->root; struct css_set *cset; unsigned long key; + int i; /* * Build the set of subsystem state objects that we want to see in the @@ -618,22 +611,25 @@ static void link_css_set(struct list_head *tmp_links, struct css_set *cset, list_add_tail(&link->cgrp_link, &cset->cgrp_links); } -/* - * find_css_set() takes an existing cgroup group and a - * cgroup object, and returns a css_set object that's - * equivalent to the old group, but with the given cgroup - * substituted into the appropriate hierarchy. Must be called with - * cgroup_mutex held +/** + * find_css_set - return a new css_set with one cgroup updated + * @old_cset: the baseline css_set + * @cgrp: the cgroup to be updated + * + * Return a new css_set that's equivalent to @old_cset, but with @cgrp + * substituted into the appropriate hierarchy. */ static struct css_set *find_css_set(struct css_set *old_cset, struct cgroup *cgrp) { + struct cgroup_subsys_state *template[CGROUP_SUBSYS_COUNT] = { }; struct css_set *cset; - struct cgroup_subsys_state *template[CGROUP_SUBSYS_COUNT]; struct list_head tmp_links; struct cgrp_cset_link *link; unsigned long key; + lockdep_assert_held(&cgroup_mutex); + /* First see if we already have a cgroup group that matches * the desired set */ read_lock(&css_set_lock); -- cgit v1.2.3 From 5549c497913ad860d3dff4386c6423268bb85693 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 24 Jun 2013 15:21:48 -0700 Subject: cgroup: s/for_each_subsys()/for_each_root_subsys()/ for_each_subsys() walks over subsystems attached to a hierarchy and we're gonna add iterators which walk over all available subsystems. Rename for_each_subsys() to for_each_root_subsys() so that it's more appropriately named and for_each_subsys() can be used to iterate all subsystems. While at it, remove unnecessary underbar prefix from macro arguments, put them inside parentheses, and adjust indentation for the two for_each_*() macros. This patch is purely cosmetic. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 47 ++++++++++++++++++++++------------------------- 1 file changed, 22 insertions(+), 25 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index c8d3175c429c..605cb13a1574 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -259,16 +259,13 @@ static int notify_on_release(const struct cgroup *cgrp) return test_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags); } -/* - * for_each_subsys() allows you to iterate on each subsystem attached to - * an active hierarchy - */ -#define for_each_subsys(_root, _ss) \ -list_for_each_entry(_ss, &_root->subsys_list, sibling) +/* iterate each subsystem attached to a hierarchy */ +#define for_each_root_subsys(root, ss) \ + list_for_each_entry((ss), &(root)->subsys_list, sibling) -/* for_each_active_root() allows you to iterate across the active hierarchies */ -#define for_each_active_root(_root) \ -list_for_each_entry(_root, &cgroup_roots, root_list) +/* iterate across the active hierarchies */ +#define for_each_active_root(root) \ + list_for_each_entry((root), &cgroup_roots, root_list) static inline struct cgroup *__d_cgrp(struct dentry *dentry) { @@ -828,7 +825,7 @@ static void cgroup_free_fn(struct work_struct *work) /* * Release the subsystem state objects. */ - for_each_subsys(cgrp->root, ss) + for_each_root_subsys(cgrp->root, ss) ss->css_free(cgrp); cgrp->root->number_of_cgroups--; @@ -944,7 +941,7 @@ static void cgroup_clear_directory(struct dentry *dir, bool base_files, struct cgroup *cgrp = __d_cgrp(dir); struct cgroup_subsys *ss; - for_each_subsys(cgrp->root, ss) { + for_each_root_subsys(cgrp->root, ss) { struct cftype_set *set; if (!test_bit(ss->subsys_id, &subsys_mask)) continue; @@ -1078,7 +1075,7 @@ static int cgroup_show_options(struct seq_file *seq, struct dentry *dentry) struct cgroup_subsys *ss; mutex_lock(&cgroup_root_mutex); - for_each_subsys(root, ss) + for_each_root_subsys(root, ss) seq_printf(seq, ",%s", ss->name); if (root->flags & CGRP_ROOT_SANE_BEHAVIOR) seq_puts(seq, ",sane_behavior"); @@ -2054,7 +2051,7 @@ static int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk, /* * step 1: check that we can legitimately attach to the cgroup. */ - for_each_subsys(root, ss) { + for_each_root_subsys(root, ss) { if (ss->can_attach) { retval = ss->can_attach(cgrp, &tset); if (retval) { @@ -2091,7 +2088,7 @@ static int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk, /* * step 4: do subsystem attach callbacks. */ - for_each_subsys(root, ss) { + for_each_root_subsys(root, ss) { if (ss->attach) ss->attach(cgrp, &tset); } @@ -2111,7 +2108,7 @@ out_put_css_set_refs: } out_cancel_attach: if (retval) { - for_each_subsys(root, ss) { + for_each_root_subsys(root, ss) { if (ss == failed_ss) break; if (ss->cancel_attach) @@ -4137,7 +4134,7 @@ static int cgroup_populate_dir(struct cgroup *cgrp, bool base_files, } /* process cftsets of each subsystem */ - for_each_subsys(cgrp->root, ss) { + for_each_root_subsys(cgrp->root, ss) { struct cftype_set *set; if (!test_bit(ss->subsys_id, &subsys_mask)) continue; @@ -4147,7 +4144,7 @@ static int cgroup_populate_dir(struct cgroup *cgrp, bool base_files, } /* This cgroup is ready now */ - for_each_subsys(cgrp->root, ss) { + for_each_root_subsys(cgrp->root, ss) { struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; /* * Update id->css pointer and make this css visible from @@ -4294,7 +4291,7 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, if (test_bit(CGRP_CPUSET_CLONE_CHILDREN, &parent->flags)) set_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags); - for_each_subsys(root, ss) { + for_each_root_subsys(root, ss) { struct cgroup_subsys_state *css; css = ss->css_alloc(cgrp); @@ -4333,14 +4330,14 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, root->number_of_cgroups++; /* each css holds a ref to the cgroup's dentry */ - for_each_subsys(root, ss) + for_each_root_subsys(root, ss) dget(dentry); /* hold a ref to the parent's dentry */ dget(parent->dentry); /* creation succeeded, notify subsystems */ - for_each_subsys(root, ss) { + for_each_root_subsys(root, ss) { err = online_css(ss, cgrp); if (err) goto err_destroy; @@ -4365,7 +4362,7 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, return 0; err_free_all: - for_each_subsys(root, ss) { + for_each_root_subsys(root, ss) { struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; if (css) { @@ -4478,7 +4475,7 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) * be killed. */ atomic_set(&cgrp->css_kill_cnt, 1); - for_each_subsys(cgrp->root, ss) { + for_each_root_subsys(cgrp->root, ss) { struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; /* @@ -4552,7 +4549,7 @@ static void cgroup_offline_fn(struct work_struct *work) * css_tryget() is guaranteed to fail now. Tell subsystems to * initate destruction. */ - for_each_subsys(cgrp->root, ss) + for_each_root_subsys(cgrp->root, ss) offline_css(ss, cgrp); /* @@ -4562,7 +4559,7 @@ static void cgroup_offline_fn(struct work_struct *work) * whenever that may be, the extra dentry ref is put so that dentry * destruction happens only after all css's are released. */ - for_each_subsys(cgrp->root, ss) + for_each_root_subsys(cgrp->root, ss) css_put(cgrp->subsys[ss->subsys_id]); /* delete this cgroup from parent->children */ @@ -4967,7 +4964,7 @@ int proc_cgroup_show(struct seq_file *m, void *v) int count = 0; seq_printf(m, "%d:", root->hierarchy_id); - for_each_subsys(root, ss) + for_each_root_subsys(root, ss) seq_printf(m, "%s%s", count++ ? "," : "", ss->name); if (strlen(root->name)) seq_printf(m, "%sname=%s", count ? "," : "", -- cgit v1.2.3 From fbab62c5cd57a6acd9ed80903532c86897d2d560 Mon Sep 17 00:00:00 2001 From: Javier Martinez Canillas Date: Fri, 14 Jun 2013 18:40:49 +0200 Subject: irqdomain: Use irq_get_trigger_type() to get IRQ flags Use irq_get_trigger_type() to get the IRQ trigger type flags instead calling irqd_get_trigger_type(irq_desc_get_irq_data(virq)) Signed-off-by: Javier Martinez Canillas Acked-by: Grant Likely Cc: Linus Walleij Cc: Samuel Ortiz Cc: Jason Cooper Cc: Andrew Lunn Cc: Russell King Cc: Ralf Baechle Cc: Benjamin Herrenschmidt Cc: linux-arm-kernel@lists.infradead.org Cc: linux-mips@linux-mips.org Link: http://lkml.kernel.org/r/1371228049-27080-8-git-send-email-javier.martinez@collabora.co.uk Signed-off-by: Thomas Gleixner --- kernel/irq/irqdomain.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c index 1db9e70f5488..489921e6242a 100644 --- a/kernel/irq/irqdomain.c +++ b/kernel/irq/irqdomain.c @@ -687,7 +687,7 @@ unsigned int irq_create_of_mapping(struct device_node *controller, /* Set type if specified and different than the current one */ if (type != IRQ_TYPE_NONE && - type != (irqd_get_trigger_type(irq_get_irq_data(virq)))) + type != irq_get_trigger_type(virq)) irq_set_irq_type(virq, type); return virq; } -- cgit v1.2.3 From 82fe9b0da0d50e2795a49c268676fd132cbc3eea Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 25 Jun 2013 11:53:37 -0700 Subject: cgroup: move init_css_set initialization inside cgroup_mutex cgroup_init() was doing init_css_set initialization outside cgroup_mutex, which is fine but we want to add lockdep annotation on subsystem iterations and cgroup_init() will trigger it spuriously. Move init_css_set initialization inside cgroup_mutex. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 605cb13a1574..3409698bd9fd 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4888,14 +4888,14 @@ int __init cgroup_init(void) cgroup_init_idr(ss, init_css_set.subsys[ss->subsys_id]); } - /* Add init_css_set to the hash table */ - key = css_set_hash(init_css_set.subsys); - hash_add(css_set_table, &init_css_set.hlist, key); - /* allocate id for the dummy hierarchy */ mutex_lock(&cgroup_mutex); mutex_lock(&cgroup_root_mutex); + /* Add init_css_set to the hash table */ + key = css_set_hash(init_css_set.subsys); + hash_add(css_set_table, &init_css_set.hlist, key); + BUG_ON(cgroup_init_root_id(&cgroup_dummy_root)); mutex_unlock(&cgroup_root_mutex); -- cgit v1.2.3 From 30159ec7a9db7f3c91e2b27e66389c49302efd5c Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 25 Jun 2013 11:53:37 -0700 Subject: cgroup: implement for_each_[builtin_]subsys() There are quite a few places where all loaded [builtin] subsys are iterated. Implement for_each_[builtin_]subsys() and replace manual iterations with those to simplify those places a bit. The new iterators automatically skip NULL subsystems. This shouldn't cause any functional difference. Iteration loops which scan all subsystems and then skipping modular ones explicitly are converted to use for_each_builtin_subsys(). While at it, reorder variable declarations and adjust whitespaces a bit in the affected functions. v2: Add lockdep_assert_held() in for_each_subsys() and add comments about synchronization as suggested by Li. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 147 +++++++++++++++++++++++++++----------------------------- 1 file changed, 71 insertions(+), 76 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 3409698bd9fd..cef688128fb8 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -259,6 +259,31 @@ static int notify_on_release(const struct cgroup *cgrp) return test_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags); } +/** + * for_each_subsys - iterate all loaded cgroup subsystems + * @ss: the iteration cursor + * @i: the index of @ss, CGROUP_SUBSYS_COUNT after reaching the end + * + * Should be called under cgroup_mutex. + */ +#define for_each_subsys(ss, i) \ + for ((i) = 0; (i) < CGROUP_SUBSYS_COUNT; (i)++) \ + if (({ lockdep_assert_held(&cgroup_mutex); \ + !((ss) = cgroup_subsys[i]); })) { } \ + else + +/** + * for_each_builtin_subsys - iterate all built-in cgroup subsystems + * @ss: the iteration cursor + * @i: the index of @ss, CGROUP_BUILTIN_SUBSYS_COUNT after reaching the end + * + * Bulit-in subsystems are always present and iteration itself doesn't + * require any synchronization. + */ +#define for_each_builtin_subsys(ss, i) \ + for ((i) = 0; (i) < CGROUP_BUILTIN_SUBSYS_COUNT && \ + (((ss) = cgroup_subsys[i]) || true); (i)++) + /* iterate each subsystem attached to a hierarchy */ #define for_each_root_subsys(root, ss) \ list_for_each_entry((ss), &(root)->subsys_list, sibling) @@ -356,10 +381,11 @@ static DEFINE_HASHTABLE(css_set_table, CSS_SET_HASH_BITS); static unsigned long css_set_hash(struct cgroup_subsys_state *css[]) { - int i; unsigned long key = 0UL; + struct cgroup_subsys *ss; + int i; - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) + for_each_subsys(ss, i) key += (unsigned long)css[i]; key = (key >> 16) ^ key; @@ -514,6 +540,7 @@ static struct css_set *find_existing_css_set(struct css_set *old_cset, struct cgroup_subsys_state *template[]) { struct cgroupfs_root *root = cgrp->root; + struct cgroup_subsys *ss; struct css_set *cset; unsigned long key; int i; @@ -523,7 +550,7 @@ static struct css_set *find_existing_css_set(struct css_set *old_cset, * new css_set. while subsystems can change globally, the entries here * won't change, so no need for locking. */ - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { + for_each_subsys(ss, i) { if (root->subsys_mask & (1UL << i)) { /* Subsystem is in this hierarchy. So we want * the subsystem state from the new @@ -982,23 +1009,19 @@ static int rebind_subsystems(struct cgroupfs_root *root, unsigned long added_mask, unsigned removed_mask) { struct cgroup *cgrp = &root->top_cgroup; + struct cgroup_subsys *ss; int i; BUG_ON(!mutex_is_locked(&cgroup_mutex)); BUG_ON(!mutex_is_locked(&cgroup_root_mutex)); /* Check that any added subsystems are currently free */ - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { + for_each_subsys(ss, i) { unsigned long bit = 1UL << i; - struct cgroup_subsys *ss = cgroup_subsys[i]; + if (!(bit & added_mask)) continue; - /* - * Nobody should tell us to do a subsys that doesn't exist: - * parse_cgroupfs_options should catch that case and refcounts - * ensure that subsystems won't disappear once selected. - */ - BUG_ON(ss == NULL); + if (ss->root != &cgroup_dummy_root) { /* Subsystem isn't free */ return -EBUSY; @@ -1013,12 +1036,11 @@ static int rebind_subsystems(struct cgroupfs_root *root, return -EBUSY; /* Process each subsystem */ - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = cgroup_subsys[i]; + for_each_subsys(ss, i) { unsigned long bit = 1UL << i; + if (bit & added_mask) { /* We're binding this subsystem to this hierarchy */ - BUG_ON(ss == NULL); BUG_ON(cgrp->subsys[i]); BUG_ON(!cgroup_dummy_top->subsys[i]); BUG_ON(cgroup_dummy_top->subsys[i]->cgroup != cgroup_dummy_top); @@ -1034,7 +1056,6 @@ static int rebind_subsystems(struct cgroupfs_root *root, root->subsys_mask |= bit; } else if (bit & removed_mask) { /* We're removing this subsystem */ - BUG_ON(ss == NULL); BUG_ON(cgrp->subsys[i] != cgroup_dummy_top->subsys[i]); BUG_ON(cgrp->subsys[i]->cgroup != cgrp); @@ -1050,7 +1071,6 @@ static int rebind_subsystems(struct cgroupfs_root *root, root->subsys_mask &= ~bit; } else if (bit & root->subsys_mask) { /* Subsystem state should already exist */ - BUG_ON(ss == NULL); BUG_ON(!cgrp->subsys[i]); /* * a refcount was taken, but we already had one, so @@ -1117,8 +1137,9 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) char *token, *o = data; bool all_ss = false, one_ss = false; unsigned long mask = (unsigned long)-1; - int i; bool module_pin_failed = false; + struct cgroup_subsys *ss; + int i; BUG_ON(!mutex_is_locked(&cgroup_mutex)); @@ -1195,10 +1216,7 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) continue; } - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = cgroup_subsys[i]; - if (ss == NULL) - continue; + for_each_subsys(ss, i) { if (strcmp(token, ss->name)) continue; if (ss->disabled) @@ -1221,16 +1239,10 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) * otherwise if 'none', 'name=' and a subsystem name options * were not specified, let's default to 'all' */ - if (all_ss || (!one_ss && !opts->none && !opts->name)) { - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = cgroup_subsys[i]; - if (ss == NULL) - continue; - if (ss->disabled) - continue; - set_bit(i, &opts->subsys_mask); - } - } + if (all_ss || (!one_ss && !opts->none && !opts->name)) + for_each_subsys(ss, i) + if (!ss->disabled) + set_bit(i, &opts->subsys_mask); /* Consistency checks */ @@ -1274,10 +1286,8 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) * take duplicate reference counts on a subsystem that's already used, * but rebind_subsystems handles this case. */ - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - unsigned long bit = 1UL << i; - - if (!(bit & opts->subsys_mask)) + for_each_subsys(ss, i) { + if (!(opts->subsys_mask & (1UL << i))) continue; if (!try_module_get(cgroup_subsys[i]->module)) { module_pin_failed = true; @@ -1306,11 +1316,11 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) static void drop_parsed_module_refcounts(unsigned long subsys_mask) { + struct cgroup_subsys *ss; int i; - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - unsigned long bit = 1UL << i; - if (!(bit & subsys_mask)) + for_each_subsys(ss, i) { + if (!(subsys_mask & (1UL << i))) continue; module_put(cgroup_subsys[i]->module); } @@ -4822,7 +4832,9 @@ EXPORT_SYMBOL_GPL(cgroup_unload_subsys); */ int __init cgroup_init_early(void) { + struct cgroup_subsys *ss; int i; + atomic_set(&init_css_set.refcount, 1); INIT_LIST_HEAD(&init_css_set.cgrp_links); INIT_LIST_HEAD(&init_css_set.tasks); @@ -4837,13 +4849,8 @@ int __init cgroup_init_early(void) list_add(&init_cgrp_cset_link.cset_link, &cgroup_dummy_top->cset_links); list_add(&init_cgrp_cset_link.cgrp_link, &init_css_set.cgrp_links); - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = cgroup_subsys[i]; - - /* at bootup time, we don't worry about modular subsystems */ - if (!ss || ss->module) - continue; - + /* at bootup time, we don't worry about modular subsystems */ + for_each_builtin_subsys(ss, i) { BUG_ON(!ss->name); BUG_ON(strlen(ss->name) > MAX_CGROUP_TYPE_NAMELEN); BUG_ON(!ss->css_alloc); @@ -4868,20 +4875,15 @@ int __init cgroup_init_early(void) */ int __init cgroup_init(void) { - int err; - int i; + struct cgroup_subsys *ss; unsigned long key; + int i, err; err = bdi_init(&cgroup_backing_dev_info); if (err) return err; - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = cgroup_subsys[i]; - - /* at bootup time, we don't worry about modular subsystems */ - if (!ss || ss->module) - continue; + for_each_builtin_subsys(ss, i) { if (!ss->early_init) cgroup_init_subsys(ss); if (ss->use_id) @@ -4990,6 +4992,7 @@ out: /* Display information about each subsystem and each hierarchy */ static int proc_cgroupstats_show(struct seq_file *m, void *v) { + struct cgroup_subsys *ss; int i; seq_puts(m, "#subsys_name\thierarchy\tnum_cgroups\tenabled\n"); @@ -4999,14 +5002,12 @@ static int proc_cgroupstats_show(struct seq_file *m, void *v) * subsys/hierarchy state. */ mutex_lock(&cgroup_mutex); - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = cgroup_subsys[i]; - if (ss == NULL) - continue; + + for_each_subsys(ss, i) seq_printf(m, "%s\t%d\t%d\t%d\n", ss->name, ss->root->hierarchy_id, ss->root->number_of_cgroups, !ss->disabled); - } + mutex_unlock(&cgroup_mutex); return 0; } @@ -5060,6 +5061,7 @@ void cgroup_fork(struct task_struct *child) */ void cgroup_post_fork(struct task_struct *child) { + struct cgroup_subsys *ss; int i; /* @@ -5096,12 +5098,9 @@ void cgroup_post_fork(struct task_struct *child) * of the array can be freed at module unload, so we * can't touch that. */ - for (i = 0; i < CGROUP_BUILTIN_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = cgroup_subsys[i]; - + for_each_builtin_subsys(ss, i) if (ss->fork) ss->fork(child); - } } } @@ -5142,6 +5141,7 @@ void cgroup_post_fork(struct task_struct *child) */ void cgroup_exit(struct task_struct *tsk, int run_callbacks) { + struct cgroup_subsys *ss; struct css_set *cset; int i; @@ -5167,13 +5167,12 @@ void cgroup_exit(struct task_struct *tsk, int run_callbacks) * fork/exit callbacks are supported only for builtin * subsystems, see cgroup_post_fork() for details. */ - for (i = 0; i < CGROUP_BUILTIN_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = cgroup_subsys[i]; - + for_each_builtin_subsys(ss, i) { if (ss->exit) { struct cgroup *old_cgrp = rcu_dereference_raw(cset->subsys[i])->cgroup; struct cgroup *cgrp = task_cgroup(tsk, i); + ss->exit(cgrp, old_cgrp, tsk); } } @@ -5280,23 +5279,19 @@ static void cgroup_release_agent(struct work_struct *work) static int __init cgroup_disable(char *str) { - int i; + struct cgroup_subsys *ss; char *token; + int i; while ((token = strsep(&str, ",")) != NULL) { if (!*token) continue; - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = cgroup_subsys[i]; - - /* - * cgroup_disable, being at boot time, can't - * know about module subsystems, so we don't - * worry about them. - */ - if (!ss || ss->module) - continue; + /* + * cgroup_disable, being at boot time, can't know about + * module subsystems, so we don't worry about them. + */ + for_each_builtin_subsys(ss, i) { if (!strcmp(token, ss->name)) { ss->disabled = 1; printk(KERN_INFO "Disabling %s control group" -- cgit v1.2.3 From fc76df706123602214da494ba98bccea83e2cfff Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 25 Jun 2013 11:53:37 -0700 Subject: cgroup: reserve ID 0 for dummy_root and 1 for unified hierarchy Before 1a57423166 ("cgroup: make hierarchy_id use cyclic idr"), hierarchy IDs were allocated from 0. As the dummy hierarchy was always the one first initialized, it got assigned 0 and all other hierarchies from 1. The patch accidentally changed the minimum useable ID to 2. Let's restore ID 0 for dummy_root and while at it reserve 1 for unified hierarchy. Signed-off-by: Tejun Heo Acked-by: Li Zefan Cc: stable@vger.kernel.org --- kernel/cgroup.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index cef688128fb8..f9c99abc38ab 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1425,14 +1425,15 @@ static void init_cgroup_root(struct cgroupfs_root *root) init_cgroup_housekeeping(cgrp); } -static int cgroup_init_root_id(struct cgroupfs_root *root) +static int cgroup_init_root_id(struct cgroupfs_root *root, int start, int end) { int id; lockdep_assert_held(&cgroup_mutex); lockdep_assert_held(&cgroup_root_mutex); - id = idr_alloc_cyclic(&cgroup_hierarchy_idr, root, 2, 0, GFP_KERNEL); + id = idr_alloc_cyclic(&cgroup_hierarchy_idr, root, start, end, + GFP_KERNEL); if (id < 0) return id; @@ -1635,7 +1636,8 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, if (ret) goto unlock_drop; - ret = cgroup_init_root_id(root); + /* ID 0 is reserved for dummy root, 1 for unified hierarchy */ + ret = cgroup_init_root_id(root, 2, 0); if (ret) goto unlock_drop; @@ -4898,7 +4900,7 @@ int __init cgroup_init(void) key = css_set_hash(init_css_set.subsys); hash_add(css_set_table, &init_css_set.hlist, key); - BUG_ON(cgroup_init_root_id(&cgroup_dummy_root)); + BUG_ON(cgroup_init_root_id(&cgroup_dummy_root, 0, 1)); mutex_unlock(&cgroup_root_mutex); mutex_unlock(&cgroup_mutex); -- cgit v1.2.3 From 13d60f4b6ab5b702dc8d2ee20999f98a93728aec Mon Sep 17 00:00:00 2001 From: Zhang Yi Date: Tue, 25 Jun 2013 21:19:31 +0800 Subject: futex: Take hugepages into account when generating futex_key The futex_keys of process shared futexes are generated from the page offset, the mapping host and the mapping index of the futex user space address. This should result in an unique identifier for each futex. Though this is not true when futexes are located in different subpages of an hugepage. The reason is, that the mapping index for all those futexes evaluates to the index of the base page of the hugetlbfs mapping. So a futex at offset 0 of the hugepage mapping and another one at offset PAGE_SIZE of the same hugepage mapping have identical futex_keys. This happens because the futex code blindly uses page->index. Steps to reproduce the bug: 1. Map a file from hugetlbfs. Initialize pthread_mutex1 at offset 0 and pthread_mutex2 at offset PAGE_SIZE of the hugetlbfs mapping. The mutexes must be initialized as PTHREAD_PROCESS_SHARED because PTHREAD_PROCESS_PRIVATE mutexes are not affected by this issue as their keys solely depend on the user space address. 2. Lock mutex1 and mutex2 3. Create thread1 and in the thread function lock mutex1, which results in thread1 blocking on the locked mutex1. 4. Create thread2 and in the thread function lock mutex2, which results in thread2 blocking on the locked mutex2. 5. Unlock mutex2. Despite the fact that mutex2 got unlocked, thread2 still blocks on mutex2 because the futex_key points to mutex1. To solve this issue we need to take the normal page index of the page which contains the futex into account, if the futex is in an hugetlbfs mapping. In other words, we calculate the normal page mapping index of the subpage in the hugetlbfs mapping. Mappings which are not based on hugetlbfs are not affected and still use page->index. Thanks to Mel Gorman who provided a patch for adding proper evaluation functions to the hugetlbfs code to avoid exposing hugetlbfs specific details to the futex code. [ tglx: Massaged changelog ] Signed-off-by: Zhang Yi Reviewed-by: Jiang Biao Tested-by: Ma Chenggong Reviewed-by: 'Mel Gorman' Acked-by: 'Darren Hart' Cc: 'Peter Zijlstra' Cc: stable@vger.kernel.org Link: http://lkml.kernel.org/r/000101ce71a6%24a83c5880%24f8b50980%24@com Signed-off-by: Thomas Gleixner --- kernel/futex.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/futex.c b/kernel/futex.c index b26dcfc02c94..49dacfb45745 100644 --- a/kernel/futex.c +++ b/kernel/futex.c @@ -61,6 +61,7 @@ #include #include #include +#include #include @@ -365,7 +366,7 @@ again: } else { key->both.offset |= FUT_OFF_INODE; /* inode-based key */ key->shared.inode = page_head->mapping->host; - key->shared.pgoff = page_head->index; + key->shared.pgoff = basepage_index(page); } get_futex_key_refs(key); -- cgit v1.2.3 From 88c8004fd3a5fdd2378069de86b90b21110d33a4 Mon Sep 17 00:00:00 2001 From: Colin Cross Date: Wed, 1 May 2013 18:35:05 -0700 Subject: futex: Use freezable blocking call Avoid waking up every thread sleeping in a futex_wait call during suspend and resume by calling a freezable blocking call. Previous patches modified the freezer to avoid sending wakeups to threads that are blocked in freezable blocking calls. This call was selected to be converted to a freezable call because it doesn't hold any locks or release any resources when interrupted that might be needed by another freezing task or a kernel driver during suspend, and is a common site where idle userspace tasks are blocked. Signed-off-by: Colin Cross Cc: Rafael J. Wysocki Cc: arve@android.com Cc: Tejun Heo Cc: Oleg Nesterov Cc: Darren Hart Cc: Randy Dunlap Cc: Al Viro Link: http://lkml.kernel.org/r/1367458508-9133-8-git-send-email-ccross@android.com Signed-off-by: Thomas Gleixner --- kernel/futex.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/futex.c b/kernel/futex.c index 49dacfb45745..c3a1a55a5214 100644 --- a/kernel/futex.c +++ b/kernel/futex.c @@ -62,6 +62,7 @@ #include #include #include +#include #include @@ -1808,7 +1809,7 @@ static void futex_wait_queue_me(struct futex_hash_bucket *hb, struct futex_q *q, * is no timeout, or if it has yet to expire. */ if (!timeout || timeout->task) - schedule(); + freezable_schedule(); } __set_current_state(TASK_RUNNING); } -- cgit v1.2.3 From d0667186eb0eab78dcca9f75af6ed03873ca8d9f Mon Sep 17 00:00:00 2001 From: JunweiZhang Date: Wed, 26 Jun 2013 16:40:05 +0800 Subject: kernel: remove unnecessary head file ip_vs.h is not necessary for sysctl_binary.c. prepare for the next patch to avoid compile issue. Signed-off-by: JunweiZhang Signed-off-by: Nicolas Dichtel Reviewed-by: Julian Anastasov Signed-off-by: Simon Horman --- kernel/sysctl_binary.c | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sysctl_binary.c b/kernel/sysctl_binary.c index aea4a9ea6fc8..b609213ca9a2 100644 --- a/kernel/sysctl_binary.c +++ b/kernel/sysctl_binary.c @@ -3,7 +3,6 @@ #include "../fs/xfs/xfs_sysctl.h" #include #include -#include #include #include #include -- cgit v1.2.3 From a41b56efa70e060f650aeb54740aaf52044a1ead Mon Sep 17 00:00:00 2001 From: Maarten Lankhorst Date: Thu, 20 Jun 2013 13:31:05 +0200 Subject: arch: Make __mutex_fastpath_lock_retval return whether fastpath succeeded or not This will allow me to call functions that have multiple arguments if fastpath fails. This is required to support ticket mutexes, because they need to be able to pass an extra argument to the fail function. Originally I duplicated the functions, by adding __mutex_fastpath_lock_retval_arg. This ended up being just a duplication of the existing function, so a way to test if fastpath was called ended up being better. This also cleaned up the reservation mutex patch some by being able to call an atomic_set instead of atomic_xchg, and making it easier to detect if the wrong unlock function was previously used. Signed-off-by: Maarten Lankhorst Acked-by: Peter Zijlstra Cc: dri-devel@lists.freedesktop.org Cc: linaro-mm-sig@lists.linaro.org Cc: robclark@gmail.com Cc: rostedt@goodmis.org Cc: daniel@ffwll.ch Cc: Linus Torvalds Cc: Andrew Morton Cc: Thomas Gleixner Link: http://lkml.kernel.org/r/20130620113105.4001.83929.stgit@patser Signed-off-by: Ingo Molnar --- kernel/mutex.c | 32 ++++++++++++++------------------ 1 file changed, 14 insertions(+), 18 deletions(-) (limited to 'kernel') diff --git a/kernel/mutex.c b/kernel/mutex.c index ad53a664f113..42f8dda2467b 100644 --- a/kernel/mutex.c +++ b/kernel/mutex.c @@ -494,10 +494,10 @@ __mutex_unlock_slowpath(atomic_t *lock_count) * mutex_lock_interruptible() and mutex_trylock(). */ static noinline int __sched -__mutex_lock_killable_slowpath(atomic_t *lock_count); +__mutex_lock_killable_slowpath(struct mutex *lock); static noinline int __sched -__mutex_lock_interruptible_slowpath(atomic_t *lock_count); +__mutex_lock_interruptible_slowpath(struct mutex *lock); /** * mutex_lock_interruptible - acquire the mutex, interruptible @@ -515,12 +515,12 @@ int __sched mutex_lock_interruptible(struct mutex *lock) int ret; might_sleep(); - ret = __mutex_fastpath_lock_retval - (&lock->count, __mutex_lock_interruptible_slowpath); - if (!ret) + ret = __mutex_fastpath_lock_retval(&lock->count); + if (likely(!ret)) { mutex_set_owner(lock); - - return ret; + return 0; + } else + return __mutex_lock_interruptible_slowpath(lock); } EXPORT_SYMBOL(mutex_lock_interruptible); @@ -530,12 +530,12 @@ int __sched mutex_lock_killable(struct mutex *lock) int ret; might_sleep(); - ret = __mutex_fastpath_lock_retval - (&lock->count, __mutex_lock_killable_slowpath); - if (!ret) + ret = __mutex_fastpath_lock_retval(&lock->count); + if (likely(!ret)) { mutex_set_owner(lock); - - return ret; + return 0; + } else + return __mutex_lock_killable_slowpath(lock); } EXPORT_SYMBOL(mutex_lock_killable); @@ -548,18 +548,14 @@ __mutex_lock_slowpath(atomic_t *lock_count) } static noinline int __sched -__mutex_lock_killable_slowpath(atomic_t *lock_count) +__mutex_lock_killable_slowpath(struct mutex *lock) { - struct mutex *lock = container_of(lock_count, struct mutex, count); - return __mutex_lock_common(lock, TASK_KILLABLE, 0, NULL, _RET_IP_); } static noinline int __sched -__mutex_lock_interruptible_slowpath(atomic_t *lock_count) +__mutex_lock_interruptible_slowpath(struct mutex *lock) { - struct mutex *lock = container_of(lock_count, struct mutex, count); - return __mutex_lock_common(lock, TASK_INTERRUPTIBLE, 0, NULL, _RET_IP_); } #endif -- cgit v1.2.3 From 040a0a37100563754bb1fee6ff6427420bcfa609 Mon Sep 17 00:00:00 2001 From: Maarten Lankhorst Date: Mon, 24 Jun 2013 10:30:04 +0200 Subject: mutex: Add support for wound/wait style locks Wound/wait mutexes are used when other multiple lock acquisitions of a similar type can be done in an arbitrary order. The deadlock handling used here is called wait/wound in the RDBMS literature: The older tasks waits until it can acquire the contended lock. The younger tasks needs to back off and drop all the locks it is currently holding, i.e. the younger task is wounded. For full documentation please read Documentation/ww-mutex-design.txt. References: https://lwn.net/Articles/548909/ Signed-off-by: Maarten Lankhorst Acked-by: Daniel Vetter Acked-by: Rob Clark Acked-by: Peter Zijlstra Cc: dri-devel@lists.freedesktop.org Cc: linaro-mm-sig@lists.linaro.org Cc: rostedt@goodmis.org Cc: daniel@ffwll.ch Cc: Linus Torvalds Cc: Andrew Morton Cc: Thomas Gleixner Link: http://lkml.kernel.org/r/51C8038C.9000106@canonical.com Signed-off-by: Ingo Molnar --- kernel/mutex.c | 318 ++++++++++++++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 302 insertions(+), 16 deletions(-) (limited to 'kernel') diff --git a/kernel/mutex.c b/kernel/mutex.c index 42f8dda2467b..fc801aafe8fd 100644 --- a/kernel/mutex.c +++ b/kernel/mutex.c @@ -254,16 +254,165 @@ void __sched mutex_unlock(struct mutex *lock) EXPORT_SYMBOL(mutex_unlock); +/** + * ww_mutex_unlock - release the w/w mutex + * @lock: the mutex to be released + * + * Unlock a mutex that has been locked by this task previously with any of the + * ww_mutex_lock* functions (with or without an acquire context). It is + * forbidden to release the locks after releasing the acquire context. + * + * This function must not be used in interrupt context. Unlocking + * of a unlocked mutex is not allowed. + */ +void __sched ww_mutex_unlock(struct ww_mutex *lock) +{ + /* + * The unlocking fastpath is the 0->1 transition from 'locked' + * into 'unlocked' state: + */ + if (lock->ctx) { +#ifdef CONFIG_DEBUG_MUTEXES + DEBUG_LOCKS_WARN_ON(!lock->ctx->acquired); +#endif + if (lock->ctx->acquired > 0) + lock->ctx->acquired--; + lock->ctx = NULL; + } + +#ifndef CONFIG_DEBUG_MUTEXES + /* + * When debugging is enabled we must not clear the owner before time, + * the slow path will always be taken, and that clears the owner field + * after verifying that it was indeed current. + */ + mutex_clear_owner(&lock->base); +#endif + __mutex_fastpath_unlock(&lock->base.count, __mutex_unlock_slowpath); +} +EXPORT_SYMBOL(ww_mutex_unlock); + +static inline int __sched +__mutex_lock_check_stamp(struct mutex *lock, struct ww_acquire_ctx *ctx) +{ + struct ww_mutex *ww = container_of(lock, struct ww_mutex, base); + struct ww_acquire_ctx *hold_ctx = ACCESS_ONCE(ww->ctx); + + if (!hold_ctx) + return 0; + + if (unlikely(ctx == hold_ctx)) + return -EALREADY; + + if (ctx->stamp - hold_ctx->stamp <= LONG_MAX && + (ctx->stamp != hold_ctx->stamp || ctx > hold_ctx)) { +#ifdef CONFIG_DEBUG_MUTEXES + DEBUG_LOCKS_WARN_ON(ctx->contending_lock); + ctx->contending_lock = ww; +#endif + return -EDEADLK; + } + + return 0; +} + +static __always_inline void ww_mutex_lock_acquired(struct ww_mutex *ww, + struct ww_acquire_ctx *ww_ctx) +{ +#ifdef CONFIG_DEBUG_MUTEXES + /* + * If this WARN_ON triggers, you used ww_mutex_lock to acquire, + * but released with a normal mutex_unlock in this call. + * + * This should never happen, always use ww_mutex_unlock. + */ + DEBUG_LOCKS_WARN_ON(ww->ctx); + + /* + * Not quite done after calling ww_acquire_done() ? + */ + DEBUG_LOCKS_WARN_ON(ww_ctx->done_acquire); + + if (ww_ctx->contending_lock) { + /* + * After -EDEADLK you tried to + * acquire a different ww_mutex? Bad! + */ + DEBUG_LOCKS_WARN_ON(ww_ctx->contending_lock != ww); + + /* + * You called ww_mutex_lock after receiving -EDEADLK, + * but 'forgot' to unlock everything else first? + */ + DEBUG_LOCKS_WARN_ON(ww_ctx->acquired > 0); + ww_ctx->contending_lock = NULL; + } + + /* + * Naughty, using a different class will lead to undefined behavior! + */ + DEBUG_LOCKS_WARN_ON(ww_ctx->ww_class != ww->ww_class); +#endif + ww_ctx->acquired++; +} + +/* + * after acquiring lock with fastpath or when we lost out in contested + * slowpath, set ctx and wake up any waiters so they can recheck. + * + * This function is never called when CONFIG_DEBUG_LOCK_ALLOC is set, + * as the fastpath and opportunistic spinning are disabled in that case. + */ +static __always_inline void +ww_mutex_set_context_fastpath(struct ww_mutex *lock, + struct ww_acquire_ctx *ctx) +{ + unsigned long flags; + struct mutex_waiter *cur; + + ww_mutex_lock_acquired(lock, ctx); + + lock->ctx = ctx; + + /* + * The lock->ctx update should be visible on all cores before + * the atomic read is done, otherwise contended waiters might be + * missed. The contended waiters will either see ww_ctx == NULL + * and keep spinning, or it will acquire wait_lock, add itself + * to waiter list and sleep. + */ + smp_mb(); /* ^^^ */ + + /* + * Check if lock is contended, if not there is nobody to wake up + */ + if (likely(atomic_read(&lock->base.count) == 0)) + return; + + /* + * Uh oh, we raced in fastpath, wake up everyone in this case, + * so they can see the new lock->ctx. + */ + spin_lock_mutex(&lock->base.wait_lock, flags); + list_for_each_entry(cur, &lock->base.wait_list, list) { + debug_mutex_wake_waiter(&lock->base, cur); + wake_up_process(cur->task); + } + spin_unlock_mutex(&lock->base.wait_lock, flags); +} + /* * Lock a mutex (possibly interruptible), slowpath: */ -static inline int __sched +static __always_inline int __sched __mutex_lock_common(struct mutex *lock, long state, unsigned int subclass, - struct lockdep_map *nest_lock, unsigned long ip) + struct lockdep_map *nest_lock, unsigned long ip, + struct ww_acquire_ctx *ww_ctx) { struct task_struct *task = current; struct mutex_waiter waiter; unsigned long flags; + int ret; preempt_disable(); mutex_acquire_nest(&lock->dep_map, subclass, 0, nest_lock, ip); @@ -298,6 +447,22 @@ __mutex_lock_common(struct mutex *lock, long state, unsigned int subclass, struct task_struct *owner; struct mspin_node node; + if (!__builtin_constant_p(ww_ctx == NULL) && ww_ctx->acquired > 0) { + struct ww_mutex *ww; + + ww = container_of(lock, struct ww_mutex, base); + /* + * If ww->ctx is set the contents are undefined, only + * by acquiring wait_lock there is a guarantee that + * they are not invalid when reading. + * + * As such, when deadlock detection needs to be + * performed the optimistic spinning cannot be done. + */ + if (ACCESS_ONCE(ww->ctx)) + break; + } + /* * If there's an owner, wait for it to either * release the lock or go to sleep. @@ -312,6 +477,13 @@ __mutex_lock_common(struct mutex *lock, long state, unsigned int subclass, if ((atomic_read(&lock->count) == 1) && (atomic_cmpxchg(&lock->count, 1, 0) == 1)) { lock_acquired(&lock->dep_map, ip); + if (!__builtin_constant_p(ww_ctx == NULL)) { + struct ww_mutex *ww; + ww = container_of(lock, struct ww_mutex, base); + + ww_mutex_set_context_fastpath(ww, ww_ctx); + } + mutex_set_owner(lock); mspin_unlock(MLOCK(lock), &node); preempt_enable(); @@ -371,15 +543,16 @@ slowpath: * TASK_UNINTERRUPTIBLE case.) */ if (unlikely(signal_pending_state(state, task))) { - mutex_remove_waiter(lock, &waiter, - task_thread_info(task)); - mutex_release(&lock->dep_map, 1, ip); - spin_unlock_mutex(&lock->wait_lock, flags); + ret = -EINTR; + goto err; + } - debug_mutex_free_waiter(&waiter); - preempt_enable(); - return -EINTR; + if (!__builtin_constant_p(ww_ctx == NULL) && ww_ctx->acquired > 0) { + ret = __mutex_lock_check_stamp(lock, ww_ctx); + if (ret) + goto err; } + __set_task_state(task, state); /* didn't get the lock, go to sleep: */ @@ -394,6 +567,30 @@ done: mutex_remove_waiter(lock, &waiter, current_thread_info()); mutex_set_owner(lock); + if (!__builtin_constant_p(ww_ctx == NULL)) { + struct ww_mutex *ww = container_of(lock, + struct ww_mutex, + base); + struct mutex_waiter *cur; + + /* + * This branch gets optimized out for the common case, + * and is only important for ww_mutex_lock. + */ + + ww_mutex_lock_acquired(ww, ww_ctx); + ww->ctx = ww_ctx; + + /* + * Give any possible sleeping processes the chance to wake up, + * so they can recheck if they have to back off. + */ + list_for_each_entry(cur, &lock->wait_list, list) { + debug_mutex_wake_waiter(lock, cur); + wake_up_process(cur->task); + } + } + /* set it to 0 if there are no waiters left: */ if (likely(list_empty(&lock->wait_list))) atomic_set(&lock->count, 0); @@ -404,6 +601,14 @@ done: preempt_enable(); return 0; + +err: + mutex_remove_waiter(lock, &waiter, task_thread_info(task)); + spin_unlock_mutex(&lock->wait_lock, flags); + debug_mutex_free_waiter(&waiter); + mutex_release(&lock->dep_map, 1, ip); + preempt_enable(); + return ret; } #ifdef CONFIG_DEBUG_LOCK_ALLOC @@ -411,7 +616,8 @@ void __sched mutex_lock_nested(struct mutex *lock, unsigned int subclass) { might_sleep(); - __mutex_lock_common(lock, TASK_UNINTERRUPTIBLE, subclass, NULL, _RET_IP_); + __mutex_lock_common(lock, TASK_UNINTERRUPTIBLE, + subclass, NULL, _RET_IP_, NULL); } EXPORT_SYMBOL_GPL(mutex_lock_nested); @@ -420,7 +626,8 @@ void __sched _mutex_lock_nest_lock(struct mutex *lock, struct lockdep_map *nest) { might_sleep(); - __mutex_lock_common(lock, TASK_UNINTERRUPTIBLE, 0, nest, _RET_IP_); + __mutex_lock_common(lock, TASK_UNINTERRUPTIBLE, + 0, nest, _RET_IP_, NULL); } EXPORT_SYMBOL_GPL(_mutex_lock_nest_lock); @@ -429,7 +636,8 @@ int __sched mutex_lock_killable_nested(struct mutex *lock, unsigned int subclass) { might_sleep(); - return __mutex_lock_common(lock, TASK_KILLABLE, subclass, NULL, _RET_IP_); + return __mutex_lock_common(lock, TASK_KILLABLE, + subclass, NULL, _RET_IP_, NULL); } EXPORT_SYMBOL_GPL(mutex_lock_killable_nested); @@ -438,10 +646,30 @@ mutex_lock_interruptible_nested(struct mutex *lock, unsigned int subclass) { might_sleep(); return __mutex_lock_common(lock, TASK_INTERRUPTIBLE, - subclass, NULL, _RET_IP_); + subclass, NULL, _RET_IP_, NULL); } EXPORT_SYMBOL_GPL(mutex_lock_interruptible_nested); + + +int __sched +__ww_mutex_lock(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) +{ + might_sleep(); + return __mutex_lock_common(&lock->base, TASK_UNINTERRUPTIBLE, + 0, &ctx->dep_map, _RET_IP_, ctx); +} +EXPORT_SYMBOL_GPL(__ww_mutex_lock); + +int __sched +__ww_mutex_lock_interruptible(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) +{ + might_sleep(); + return __mutex_lock_common(&lock->base, TASK_INTERRUPTIBLE, + 0, &ctx->dep_map, _RET_IP_, ctx); +} +EXPORT_SYMBOL_GPL(__ww_mutex_lock_interruptible); + #endif /* @@ -544,20 +772,39 @@ __mutex_lock_slowpath(atomic_t *lock_count) { struct mutex *lock = container_of(lock_count, struct mutex, count); - __mutex_lock_common(lock, TASK_UNINTERRUPTIBLE, 0, NULL, _RET_IP_); + __mutex_lock_common(lock, TASK_UNINTERRUPTIBLE, 0, + NULL, _RET_IP_, NULL); } static noinline int __sched __mutex_lock_killable_slowpath(struct mutex *lock) { - return __mutex_lock_common(lock, TASK_KILLABLE, 0, NULL, _RET_IP_); + return __mutex_lock_common(lock, TASK_KILLABLE, 0, + NULL, _RET_IP_, NULL); } static noinline int __sched __mutex_lock_interruptible_slowpath(struct mutex *lock) { - return __mutex_lock_common(lock, TASK_INTERRUPTIBLE, 0, NULL, _RET_IP_); + return __mutex_lock_common(lock, TASK_INTERRUPTIBLE, 0, + NULL, _RET_IP_, NULL); +} + +static noinline int __sched +__ww_mutex_lock_slowpath(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) +{ + return __mutex_lock_common(&lock->base, TASK_UNINTERRUPTIBLE, 0, + NULL, _RET_IP_, ctx); } + +static noinline int __sched +__ww_mutex_lock_interruptible_slowpath(struct ww_mutex *lock, + struct ww_acquire_ctx *ctx) +{ + return __mutex_lock_common(&lock->base, TASK_INTERRUPTIBLE, 0, + NULL, _RET_IP_, ctx); +} + #endif /* @@ -613,6 +860,45 @@ int __sched mutex_trylock(struct mutex *lock) } EXPORT_SYMBOL(mutex_trylock); +#ifndef CONFIG_DEBUG_LOCK_ALLOC +int __sched +__ww_mutex_lock(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) +{ + int ret; + + might_sleep(); + + ret = __mutex_fastpath_lock_retval(&lock->base.count); + + if (likely(!ret)) { + ww_mutex_set_context_fastpath(lock, ctx); + mutex_set_owner(&lock->base); + } else + ret = __ww_mutex_lock_slowpath(lock, ctx); + return ret; +} +EXPORT_SYMBOL(__ww_mutex_lock); + +int __sched +__ww_mutex_lock_interruptible(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) +{ + int ret; + + might_sleep(); + + ret = __mutex_fastpath_lock_retval(&lock->base.count); + + if (likely(!ret)) { + ww_mutex_set_context_fastpath(lock, ctx); + mutex_set_owner(&lock->base); + } else + ret = __ww_mutex_lock_interruptible_slowpath(lock, ctx); + return ret; +} +EXPORT_SYMBOL(__ww_mutex_lock_interruptible); + +#endif + /** * atomic_dec_and_mutex_lock - return holding mutex if we dec to 0 * @cnt: the atomic which we are to dec -- cgit v1.2.3 From 230100276955529d5a7c69207421756b9a61a8e5 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Thu, 20 Jun 2013 13:31:17 +0200 Subject: mutex: Add w/w mutex slowpath debugging Injects EDEADLK conditions at pseudo-random interval, with exponential backoff up to UINT_MAX (to ensure that every lock operation still completes in a reasonable time). This way we can test the wound slowpath even for ww mutex users where contention is never expected, and the ww deadlock avoidance algorithm is only needed for correctness against malicious userspace. An example would be protecting kernel modesetting properties, which thanks to single-threaded X isn't really expected to contend, ever. I've looked into using the CONFIG_FAULT_INJECTION infrastructure, but decided against it for two reasons: - EDEADLK handling is mandatory for ww mutex users and should never affect the outcome of a syscall. This is in contrast to -ENOMEM injection. So fine configurability isn't required. - The fault injection framework only allows to set a simple probability for failure. Now the probability that a ww mutex acquire stage with N locks will never complete (due to too many injected EDEADLK backoffs) is zero. But the expected number of ww_mutex_lock operations for the completely uncontended case would be O(exp(N)). The per-acuiqire ctx exponential backoff solution choosen here only results in O(log N) overhead due to injection and so O(log N * N) lock operations. This way we can fail with high probability (and so have good test coverage even for fancy backoff and lock acquisition paths) without running into patalogical cases. Note that EDEADLK will only ever be injected when we managed to acquire the lock. This prevents any behaviour changes for users which rely on the EALREADY semantics. Signed-off-by: Daniel Vetter Signed-off-by: Maarten Lankhorst Acked-by: Peter Zijlstra Cc: dri-devel@lists.freedesktop.org Cc: linaro-mm-sig@lists.linaro.org Cc: rostedt@goodmis.org Cc: daniel@ffwll.ch Cc: Linus Torvalds Cc: Andrew Morton Cc: Thomas Gleixner Link: http://lkml.kernel.org/r/20130620113117.4001.21681.stgit@patser Signed-off-by: Ingo Molnar --- kernel/mutex.c | 44 +++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 41 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/mutex.c b/kernel/mutex.c index fc801aafe8fd..e581ada5faf4 100644 --- a/kernel/mutex.c +++ b/kernel/mutex.c @@ -651,22 +651,60 @@ mutex_lock_interruptible_nested(struct mutex *lock, unsigned int subclass) EXPORT_SYMBOL_GPL(mutex_lock_interruptible_nested); +static inline int +ww_mutex_deadlock_injection(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) +{ +#ifdef CONFIG_DEBUG_WW_MUTEX_SLOWPATH + unsigned tmp; + + if (ctx->deadlock_inject_countdown-- == 0) { + tmp = ctx->deadlock_inject_interval; + if (tmp > UINT_MAX/4) + tmp = UINT_MAX; + else + tmp = tmp*2 + tmp + tmp/2; + + ctx->deadlock_inject_interval = tmp; + ctx->deadlock_inject_countdown = tmp; + ctx->contending_lock = lock; + + ww_mutex_unlock(lock); + + return -EDEADLK; + } +#endif + + return 0; +} int __sched __ww_mutex_lock(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) { + int ret; + might_sleep(); - return __mutex_lock_common(&lock->base, TASK_UNINTERRUPTIBLE, + ret = __mutex_lock_common(&lock->base, TASK_UNINTERRUPTIBLE, 0, &ctx->dep_map, _RET_IP_, ctx); + if (!ret && ctx->acquired > 0) + return ww_mutex_deadlock_injection(lock, ctx); + + return ret; } EXPORT_SYMBOL_GPL(__ww_mutex_lock); int __sched __ww_mutex_lock_interruptible(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) { + int ret; + might_sleep(); - return __mutex_lock_common(&lock->base, TASK_INTERRUPTIBLE, - 0, &ctx->dep_map, _RET_IP_, ctx); + ret = __mutex_lock_common(&lock->base, TASK_INTERRUPTIBLE, + 0, &ctx->dep_map, _RET_IP_, ctx); + + if (!ret && ctx->acquired > 0) + return ww_mutex_deadlock_injection(lock, ctx); + + return ret; } EXPORT_SYMBOL_GPL(__ww_mutex_lock_interruptible); -- cgit v1.2.3 From 1672d040709b789671c0502e7aac9d632c2f9175 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 25 Jun 2013 18:04:54 -0700 Subject: cgroup: fix cgroupfs_root early destruction path cgroupfs_root used to have ->actual_subsys_mask in addition to ->subsys_mask. a8a648c4ac ("cgroup: remove cgroup->actual_subsys_mask") removed it noting that the subsys_mask is essentially temporary and doesn't belong in cgroupfs_root; however, the patch made it impossible to tell whether a cgroupfs_root actually has the subsystems bound or just have the bits set leading to the following BUG when trying to mount with subsystems which are already mounted elsewhere. kernel BUG at kernel/cgroup.c:1038! invalid opcode: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC ... CPU: 1 PID: 7973 Comm: mount Tainted: G W 3.10.0-rc7-next-20130625-sasha-00011-g1c1dc0e #1105 task: ffff880fc0ae8000 ti: ffff880fc0b9a000 task.ti: ffff880fc0b9a000 RIP: 0010:[] [] rebind_subsystems+0x409/0x5f0 ... Call Trace: [] cgroup_kill_sb+0xff/0x210 [] deactivate_locked_super+0x4f/0x90 [] cgroup_mount+0x673/0x6e0 [] cpuset_mount+0xd9/0x110 [] mount_fs+0xb0/0x2d0 [] vfs_kern_mount+0xbd/0x180 [] do_new_mount+0x145/0x2c0 [] do_mount+0x356/0x3c0 [] SyS_mount+0xfd/0x140 [] tracesys+0xdd/0xe2 We still want rebind_subsystems() to take added/removed masks, so let's fix it by marking whether a cgroupfs_root has finished binding or not. Also, document what's going on around ->subsys_mask initialization so that similar mistakes aren't repeated. Signed-off-by: Tejun Heo Reported-by: Sasha Levin Acked-by: Li Zefan --- kernel/cgroup.c | 22 +++++++++++++++++++--- 1 file changed, 19 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index f9c99abc38ab..e801ecfa36ef 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1086,6 +1086,12 @@ static int rebind_subsystems(struct cgroupfs_root *root, } } + /* + * Mark @root has finished binding subsystems. @root->subsys_mask + * now matches the bound subsystems. + */ + root->flags |= CGRP_ROOT_SUBSYS_BOUND; + return 0; } @@ -1485,6 +1491,14 @@ static struct cgroupfs_root *cgroup_root_from_opts(struct cgroup_sb_opts *opts) init_cgroup_root(root); + /* + * We need to set @root->subsys_mask now so that @root can be + * matched by cgroup_test_super() before it finishes + * initialization; otherwise, competing mounts with the same + * options may try to bind the same subsystems instead of waiting + * for the first one leading to unexpected mount errors. + * SUBSYS_BOUND will be set once actual binding is complete. + */ root->subsys_mask = opts->subsys_mask; root->flags = opts->flags; ida_init(&root->cgroup_ida); @@ -1734,9 +1748,11 @@ static void cgroup_kill_sb(struct super_block *sb) { mutex_lock(&cgroup_root_mutex); /* Rebind all subsystems back to the default hierarchy */ - ret = rebind_subsystems(root, 0, root->subsys_mask); - /* Shouldn't be able to fail ... */ - BUG_ON(ret); + if (root->flags & CGRP_ROOT_SUBSYS_BOUND) { + ret = rebind_subsystems(root, 0, root->subsys_mask); + /* Shouldn't be able to fail ... */ + BUG_ON(ret); + } /* * Release all the links from cset_links to this hierarchy's -- cgit v1.2.3 From eb178d063324d9c30f673db3877b892a48ade21e Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 25 Jun 2013 18:05:21 -0700 Subject: cgroup: grab cgroup_mutex in drop_parsed_module_refcounts() This isn't strictly necessary as all subsystems specified in @subsys_mask are guaranteed to be pinned; however, it does spuriously trigger lockdep warning. Let's grab cgroup_mutex around it. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index e801ecfa36ef..2d3a132e881d 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1325,11 +1325,11 @@ static void drop_parsed_module_refcounts(unsigned long subsys_mask) struct cgroup_subsys *ss; int i; - for_each_subsys(ss, i) { - if (!(subsys_mask & (1UL << i))) - continue; - module_put(cgroup_subsys[i]->module); - } + mutex_lock(&cgroup_mutex); + for_each_subsys(ss, i) + if (subsys_mask & (1UL << i)) + module_put(cgroup_subsys[i]->module); + mutex_unlock(&cgroup_mutex); } static int cgroup_remount(struct super_block *sb, int *flags, char *data) -- cgit v1.2.3 From a8ad805cfde00be8fe3b3dae8890996dbeb91e2c Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 21 Jun 2013 15:52:04 -0700 Subject: cgroup: fix RCU accesses around task->cgroups There are several places in kernel/cgroup.c where task->cgroups is accessed and modified without going through proper RCU accessors. None is broken as they're all lock protected accesses; however, this still triggers sparse RCU address space warnings. * Consistently use task_css_set() for task->cgroups dereferencing. * Use RCU_INIT_POINTER() to clear task->cgroups to &init_css_set on exit. * Remove unnecessary rcu_dereference_raw() from cset->subsys[] dereference in cgroup_exit(). Signed-off-by: Tejun Heo Reported-by: Fengguang Wu Acked-by: Li Zefan --- kernel/cgroup.c | 24 +++++++++++++----------- 1 file changed, 13 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 2d3a132e881d..ee9f0c1c8bff 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -724,7 +724,7 @@ static struct cgroup *task_cgroup_from_root(struct task_struct *task, * task can't change groups, so the only thing that can happen * is that it exits and its css is set back to init_css_set. */ - cset = task->cgroups; + cset = task_css_set(task); if (cset == &init_css_set) { res = &root->top_cgroup; } else { @@ -1971,7 +1971,7 @@ static void cgroup_task_migrate(struct cgroup *old_cgrp, * css_set to init_css_set and dropping the old one. */ WARN_ON_ONCE(tsk->flags & PF_EXITING); - old_cset = tsk->cgroups; + old_cset = task_css_set(tsk); task_lock(tsk); rcu_assign_pointer(tsk->cgroups, new_cset); @@ -2094,8 +2094,11 @@ static int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk, * we use find_css_set, which allocates a new one if necessary. */ for (i = 0; i < group_size; i++) { + struct css_set *old_cset; + tc = flex_array_get(group, i); - tc->cg = find_css_set(tc->task->cgroups, cgrp); + old_cset = task_css_set(tc->task); + tc->cg = find_css_set(old_cset, cgrp); if (!tc->cg) { retval = -ENOMEM; goto out_put_css_set_refs; @@ -3012,7 +3015,7 @@ static void cgroup_enable_task_cg_lists(void) * entry won't be deleted though the process has exited. */ if (!(p->flags & PF_EXITING) && list_empty(&p->cg_list)) - list_add(&p->cg_list, &p->cgroups->tasks); + list_add(&p->cg_list, &task_css_set(p)->tasks); task_unlock(p); } while_each_thread(g, p); read_unlock(&tasklist_lock); @@ -5061,8 +5064,8 @@ static const struct file_operations proc_cgroupstats_operations = { void cgroup_fork(struct task_struct *child) { task_lock(current); + get_css_set(task_css_set(current)); child->cgroups = current->cgroups; - get_css_set(child->cgroups); task_unlock(current); INIT_LIST_HEAD(&child->cg_list); } @@ -5097,7 +5100,7 @@ void cgroup_post_fork(struct task_struct *child) write_lock(&css_set_lock); task_lock(child); if (list_empty(&child->cg_list)) - list_add(&child->cg_list, &child->cgroups->tasks); + list_add(&child->cg_list, &task_css_set(child)->tasks); task_unlock(child); write_unlock(&css_set_lock); } @@ -5177,8 +5180,8 @@ void cgroup_exit(struct task_struct *tsk, int run_callbacks) /* Reassign the task to the init_css_set. */ task_lock(tsk); - cset = tsk->cgroups; - tsk->cgroups = &init_css_set; + cset = task_css_set(tsk); + RCU_INIT_POINTER(tsk->cgroups, &init_css_set); if (run_callbacks && need_forkexit_callback) { /* @@ -5187,8 +5190,7 @@ void cgroup_exit(struct task_struct *tsk, int run_callbacks) */ for_each_builtin_subsys(ss, i) { if (ss->exit) { - struct cgroup *old_cgrp = - rcu_dereference_raw(cset->subsys[i])->cgroup; + struct cgroup *old_cgrp = cset->subsys[i]->cgroup; struct cgroup *cgrp = task_cgroup(tsk, i); ss->exit(cgrp, old_cgrp, tsk); @@ -5555,7 +5557,7 @@ static u64 current_css_set_refcount_read(struct cgroup *cgrp, u64 count; rcu_read_lock(); - count = atomic_read(¤t->cgroups->refcount); + count = atomic_read(&task_css_set(current)->refcount); rcu_read_unlock(); return count; } -- cgit v1.2.3 From a4ea1cc90604df08d471ae84eb9627319d10c844 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 21 Jun 2013 15:52:33 -0700 Subject: cgroup: always use RCU accessors for protected accesses kernel/cgroup.c still has places where a RCU pointer is set and accessed directly without going through RCU_INIT_POINTER() or rcu_dereference_protected(). They're all properly protected accesses so nothing is broken but it leads to spurious sparse RCU address space warnings. Substitute direct accesses with RCU_INIT_POINTER() and rcu_dereference_protected(). Note that %true is specified as the extra condition for all derference updates. This isn't ideal as all it does is suppressing warning without actually policing synchronization rules; however, most are scheduled to be removed pretty soon along with css_id itself, so no reason to be more elaborate. Combined with the previous changes, this removes all RCU related sparse warnings from cgroup. Signed-off-by: Tejun Heo Reported-by: Fengguang Wu Acked-by; Li Zefan --- kernel/cgroup.c | 21 ++++++++++++--------- 1 file changed, 12 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index ee9f0c1c8bff..4ed86773fff7 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1427,7 +1427,7 @@ static void init_cgroup_root(struct cgroupfs_root *root) INIT_LIST_HEAD(&root->root_list); root->number_of_cgroups = 1; cgrp->root = root; - cgrp->name = &root_cgroup_name; + RCU_INIT_POINTER(cgrp->name, &root_cgroup_name); init_cgroup_housekeeping(cgrp); } @@ -2558,7 +2558,7 @@ static int cgroup_rename(struct inode *old_dir, struct dentry *old_dentry, return ret; } - old_name = cgrp->name; + old_name = rcu_dereference_protected(cgrp->name, true); rcu_assign_pointer(cgrp->name, name); kfree_rcu(old_name, rcu_head); @@ -4177,13 +4177,15 @@ static int cgroup_populate_dir(struct cgroup *cgrp, bool base_files, /* This cgroup is ready now */ for_each_root_subsys(cgrp->root, ss) { struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + struct css_id *id = rcu_dereference_protected(css->id, true); + /* * Update id->css pointer and make this css visible from * CSS ID functions. This pointer will be dereferened * from RCU-read-side without locks. */ - if (css->id) - rcu_assign_pointer(css->id->css, css); + if (id) + rcu_assign_pointer(id->css, css); } return 0; @@ -4863,7 +4865,7 @@ int __init cgroup_init_early(void) css_set_count = 1; init_cgroup_root(&cgroup_dummy_root); cgroup_root_count = 1; - init_task.cgroups = &init_css_set; + RCU_INIT_POINTER(init_task.cgroups, &init_css_set); init_cgrp_cset_link.cset = &init_css_set; init_cgrp_cset_link.cgrp = cgroup_dummy_top; @@ -5380,7 +5382,8 @@ bool css_is_ancestor(struct cgroup_subsys_state *child, void free_css_id(struct cgroup_subsys *ss, struct cgroup_subsys_state *css) { - struct css_id *id = css->id; + struct css_id *id = rcu_dereference_protected(css->id, true); + /* When this is called before css_id initialization, id can be NULL */ if (!id) return; @@ -5446,8 +5449,8 @@ static int __init_or_module cgroup_init_idr(struct cgroup_subsys *ss, return PTR_ERR(newid); newid->stack[0] = newid->id; - newid->css = rootcss; - rootcss->id = newid; + RCU_INIT_POINTER(newid->css, rootcss); + RCU_INIT_POINTER(rootcss->id, newid); return 0; } @@ -5461,7 +5464,7 @@ static int alloc_css_id(struct cgroup_subsys *ss, struct cgroup *parent, subsys_id = ss->subsys_id; parent_css = parent->subsys[subsys_id]; child_css = child->subsys[subsys_id]; - parent_id = parent_css->id; + parent_id = rcu_dereference_protected(parent_css->id, true); depth = parent_id->depth + 1; child_id = get_new_cssid(ss, depth); -- cgit v1.2.3 From 141965c7494d984b2bf24efd361a3125278869c6 Mon Sep 17 00:00:00 2001 From: Alex Shi Date: Wed, 26 Jun 2013 13:05:39 +0800 Subject: Revert "sched: Introduce temporary FAIR_GROUP_SCHED dependency for load-tracking" Remove CONFIG_FAIR_GROUP_SCHED that covers the runnable info, then we can use runnable load variables. Also remove 2 CONFIG_FAIR_GROUP_SCHED setting which is not in reverted patch(introduced in 9ee474f), but also need to revert. Signed-off-by: Alex Shi Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/51CA76A3.3050207@intel.com Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 7 +------ kernel/sched/fair.c | 17 ++++------------- kernel/sched/sched.h | 19 ++----------------- 3 files changed, 7 insertions(+), 36 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index ceeaf0f45be0..0241b1b55a04 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -1611,12 +1611,7 @@ static void __sched_fork(struct task_struct *p) p->se.vruntime = 0; INIT_LIST_HEAD(&p->se.group_node); -/* - * Load-tracking only depends on SMP, FAIR_GROUP_SCHED dependency below may be - * removed when useful for applications beyond shares distribution (e.g. - * load-balance). - */ -#if defined(CONFIG_SMP) && defined(CONFIG_FAIR_GROUP_SCHED) +#ifdef CONFIG_SMP p->se.avg.runnable_avg_period = 0; p->se.avg.runnable_avg_sum = 0; #endif diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index c0ac2c3b56e1..36eadaaa4e5b 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -1128,8 +1128,7 @@ static inline void update_cfs_shares(struct cfs_rq *cfs_rq) } #endif /* CONFIG_FAIR_GROUP_SCHED */ -/* Only depends on SMP, FAIR_GROUP_SCHED may be removed when useful in lb */ -#if defined(CONFIG_SMP) && defined(CONFIG_FAIR_GROUP_SCHED) +#ifdef CONFIG_SMP /* * We choose a half-life close to 1 scheduling period. * Note: The tables below are dependent on this value. @@ -3430,12 +3429,6 @@ unlock: return new_cpu; } -/* - * Load-tracking only depends on SMP, FAIR_GROUP_SCHED dependency below may be - * removed when useful for applications beyond shares distribution (e.g. - * load-balance). - */ -#ifdef CONFIG_FAIR_GROUP_SCHED /* * Called immediately before a task is migrated to a new cpu; task_cpu(p) and * cfs_rq_of(p) references at time of call are still valid and identify the @@ -3459,7 +3452,6 @@ migrate_task_rq_fair(struct task_struct *p, int next_cpu) atomic64_add(se->avg.load_avg_contrib, &cfs_rq->removed_load); } } -#endif #endif /* CONFIG_SMP */ static unsigned long @@ -5861,7 +5853,7 @@ static void switched_from_fair(struct rq *rq, struct task_struct *p) se->vruntime -= cfs_rq->min_vruntime; } -#if defined(CONFIG_FAIR_GROUP_SCHED) && defined(CONFIG_SMP) +#ifdef CONFIG_SMP /* * Remove our load from contribution when we leave sched_fair * and ensure we don't carry in an old decay_count if we @@ -5920,7 +5912,7 @@ void init_cfs_rq(struct cfs_rq *cfs_rq) #ifndef CONFIG_64BIT cfs_rq->min_vruntime_copy = cfs_rq->min_vruntime; #endif -#if defined(CONFIG_FAIR_GROUP_SCHED) && defined(CONFIG_SMP) +#ifdef CONFIG_SMP atomic64_set(&cfs_rq->decay_counter, 1); atomic64_set(&cfs_rq->removed_load, 0); #endif @@ -6162,9 +6154,8 @@ const struct sched_class fair_sched_class = { #ifdef CONFIG_SMP .select_task_rq = select_task_rq_fair, -#ifdef CONFIG_FAIR_GROUP_SCHED .migrate_task_rq = migrate_task_rq_fair, -#endif + .rq_online = rq_online_fair, .rq_offline = rq_offline_fair, diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index 029601a61587..77ce668ba302 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -269,12 +269,6 @@ struct cfs_rq { #endif #ifdef CONFIG_SMP -/* - * Load-tracking only depends on SMP, FAIR_GROUP_SCHED dependency below may be - * removed when useful for applications beyond shares distribution (e.g. - * load-balance). - */ -#ifdef CONFIG_FAIR_GROUP_SCHED /* * CFS Load tracking * Under CFS, load is tracked on a per-entity basis and aggregated up. @@ -284,9 +278,9 @@ struct cfs_rq { u64 runnable_load_avg, blocked_load_avg; atomic64_t decay_counter, removed_load; u64 last_decay; -#endif /* CONFIG_FAIR_GROUP_SCHED */ -/* These always depend on CONFIG_FAIR_GROUP_SCHED */ + #ifdef CONFIG_FAIR_GROUP_SCHED + /* Required to track per-cpu representation of a task_group */ u32 tg_runnable_contrib; u64 tg_load_contrib; #endif /* CONFIG_FAIR_GROUP_SCHED */ @@ -1027,17 +1021,8 @@ extern void update_group_power(struct sched_domain *sd, int cpu); extern void trigger_load_balance(struct rq *rq, int cpu); extern void idle_balance(int this_cpu, struct rq *this_rq); -/* - * Only depends on SMP, FAIR_GROUP_SCHED may be removed when runnable_avg - * becomes useful in lb - */ -#if defined(CONFIG_FAIR_GROUP_SCHED) extern void idle_enter_fair(struct rq *this_rq); extern void idle_exit_fair(struct rq *this_rq); -#else -static inline void idle_enter_fair(struct rq *this_rq) {} -static inline void idle_exit_fair(struct rq *this_rq) {} -#endif #else /* CONFIG_SMP */ -- cgit v1.2.3 From fa6bddeb14d59d701f846b174b59c9982e926e66 Mon Sep 17 00:00:00 2001 From: Alex Shi Date: Thu, 20 Jun 2013 10:18:46 +0800 Subject: sched: Move a few runnable tg variables into CONFIG_SMP The following 2 variables are only used under CONFIG_SMP, so its better to move their definiation into CONFIG_SMP too. atomic64_t load_avg; atomic_t runnable_avg; Signed-off-by: Alex Shi Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1371694737-29336-3-git-send-email-alex.shi@intel.com Signed-off-by: Ingo Molnar --- kernel/sched/sched.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'kernel') diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index 77ce668ba302..31d25f80a7c6 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -149,9 +149,11 @@ struct task_group { unsigned long shares; atomic_t load_weight; +#ifdef CONFIG_SMP atomic64_t load_avg; atomic_t runnable_avg; #endif +#endif #ifdef CONFIG_RT_GROUP_SCHED struct sched_rt_entity **rt_se; -- cgit v1.2.3 From a75cdaa915e42ef0e6f38dc7f2a6a1deca91d648 Mon Sep 17 00:00:00 2001 From: Alex Shi Date: Thu, 20 Jun 2013 10:18:47 +0800 Subject: sched: Set an initial value of runnable avg for new forked task We need to initialize the se.avg.{decay_count, load_avg_contrib} for a new forked task. Otherwise random values of above variables cause a mess when a new task is enqueued: enqueue_task_fair enqueue_entity enqueue_entity_load_avg and make fork balancing imbalance due to incorrect load_avg_contrib. Further more, Morten Rasmussen notice some tasks were not launched at once after created. So Paul and Peter suggest giving a start value for new task runnable avg time same as sched_slice(). PeterZ said: > So the 'problem' is that our running avg is a 'floating' average; ie. it > decays with time. Now we have to guess about the future of our newly > spawned task -- something that is nigh impossible seeing these CPU > vendors keep refusing to implement the crystal ball instruction. > > So there's two asymptotic cases we want to deal well with; 1) the case > where the newly spawned program will be 'nearly' idle for its lifetime; > and 2) the case where its cpu-bound. > > Since we have to guess, we'll go for worst case and assume its > cpu-bound; now we don't want to make the avg so heavy adjusting to the > near-idle case takes forever. We want to be able to quickly adjust and > lower our running avg. > > Now we also don't want to make our avg too light, such that it gets > decremented just for the new task not having had a chance to run yet -- > even if when it would run, it would be more cpu-bound than not. > > So what we do is we make the initial avg of the same duration as that we > guess it takes to run each task on the system at least once -- aka > sched_slice(). > > Of course we can defeat this with wakeup/fork bombs, but in the 'normal' > case it should be good enough. Paul also contributed most of the code comments in this commit. Signed-off-by: Alex Shi Reviewed-by: Gu Zheng Reviewed-by: Paul Turner [peterz; added explanation of sched_slice() usage] Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1371694737-29336-4-git-send-email-alex.shi@intel.com Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 6 ++---- kernel/sched/fair.c | 24 ++++++++++++++++++++++++ kernel/sched/sched.h | 2 ++ 3 files changed, 28 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 0241b1b55a04..729e7fc7634b 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -1611,10 +1611,6 @@ static void __sched_fork(struct task_struct *p) p->se.vruntime = 0; INIT_LIST_HEAD(&p->se.group_node); -#ifdef CONFIG_SMP - p->se.avg.runnable_avg_period = 0; - p->se.avg.runnable_avg_sum = 0; -#endif #ifdef CONFIG_SCHEDSTATS memset(&p->se.statistics, 0, sizeof(p->se.statistics)); #endif @@ -1758,6 +1754,8 @@ void wake_up_new_task(struct task_struct *p) set_task_cpu(p, select_task_rq(p, SD_BALANCE_FORK, 0)); #endif + /* Initialize new task's runnable average */ + init_task_runnable_average(p); rq = __task_rq_lock(p); activate_task(rq, p, 0); p->on_rq = 1; diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 36eadaaa4e5b..e1602a0fdbf8 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -680,6 +680,26 @@ static u64 sched_vslice(struct cfs_rq *cfs_rq, struct sched_entity *se) return calc_delta_fair(sched_slice(cfs_rq, se), se); } +#ifdef CONFIG_SMP +static inline void __update_task_entity_contrib(struct sched_entity *se); + +/* Give new task start runnable values to heavy its load in infant time */ +void init_task_runnable_average(struct task_struct *p) +{ + u32 slice; + + p->se.avg.decay_count = 0; + slice = sched_slice(task_cfs_rq(p), &p->se) >> 10; + p->se.avg.runnable_avg_sum = slice; + p->se.avg.runnable_avg_period = slice; + __update_task_entity_contrib(&p->se); +} +#else +void init_task_runnable_average(struct task_struct *p) +{ +} +#endif + /* * Update the current task's runtime statistics. Skip current tasks that * are not in our scheduling class. @@ -1527,6 +1547,10 @@ static inline void enqueue_entity_load_avg(struct cfs_rq *cfs_rq, * We track migrations using entity decay_count <= 0, on a wake-up * migration we use a negative decay count to track the remote decays * accumulated while sleeping. + * + * Newly forked tasks are enqueued with se->avg.decay_count == 0, they + * are seen by enqueue_entity_load_avg() as a migration with an already + * constructed load_avg_contrib. */ if (unlikely(se->avg.decay_count <= 0)) { se->avg.last_runnable_update = rq_clock_task(rq_of(cfs_rq)); diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index 31d25f80a7c6..9c65d46504b1 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -1048,6 +1048,8 @@ extern void init_rt_bandwidth(struct rt_bandwidth *rt_b, u64 period, u64 runtime extern void update_idle_cpu_load(struct rq *this_rq); +extern void init_task_runnable_average(struct task_struct *p); + #ifdef CONFIG_PARAVIRT static inline u64 steal_ticks(u64 steal) { -- cgit v1.2.3 From 282cf499f03ec1754b6c8c945c9674b02631fb0f Mon Sep 17 00:00:00 2001 From: Alex Shi Date: Thu, 20 Jun 2013 10:18:48 +0800 Subject: sched: Fix sleep time double accounting in enqueue entity The woken migrated task will __synchronize_entity_decay(se); in migrate_task_rq_fair, then it needs to set `se->avg.last_runnable_update -= (-se->avg.decay_count) << 20' before update_entity_load_avg, in order to avoid sleep time is updated twice for se.avg.load_avg_contrib in both __syncchronize and update_entity_load_avg. However if the sleeping task is woken up from the same cpu, it miss the last_runnable_update before update_entity_load_avg(se, 0, 1), then the sleep time was used twice in both functions. So we need to remove the double sleep time accounting. Paul also contributed some code comments in this commit. Signed-off-by: Alex Shi Reviewed-by: Paul Turner Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1371694737-29336-5-git-send-email-alex.shi@intel.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index e1602a0fdbf8..9bbc303598ea 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -1571,7 +1571,13 @@ static inline void enqueue_entity_load_avg(struct cfs_rq *cfs_rq, } wakeup = 0; } else { - __synchronize_entity_decay(se); + /* + * Task re-woke on same cpu (or else migrate_task_rq_fair() + * would have made count negative); we must be careful to avoid + * double-accounting blocked time after synchronizing decays. + */ + se->avg.last_runnable_update += __synchronize_entity_decay(se) + << 20; } /* migrated tasks did not contribute to our blocked load */ -- cgit v1.2.3 From 83dfd5235ebd66c284b97befe6eabff7132333e6 Mon Sep 17 00:00:00 2001 From: Alex Shi Date: Thu, 20 Jun 2013 10:18:49 +0800 Subject: sched: Update cpu load after task_tick To get the latest runnable info, we need do this cpuload update after task_tick. Signed-off-by: Alex Shi Reviewed-by: Paul Turner Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1371694737-29336-6-git-send-email-alex.shi@intel.com Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 729e7fc7634b..08746cc12370 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -2165,8 +2165,8 @@ void scheduler_tick(void) raw_spin_lock(&rq->lock); update_rq_clock(rq); - update_cpu_load_active(rq); curr->sched_class->task_tick(rq, curr, 0); + update_cpu_load_active(rq); raw_spin_unlock(&rq->lock); perf_event_task_tick(); -- cgit v1.2.3 From b92486cbf2aa230d00f160664858495c81d2b37b Mon Sep 17 00:00:00 2001 From: Alex Shi Date: Thu, 20 Jun 2013 10:18:50 +0800 Subject: sched: Compute runnable load avg in cpu_load and cpu_avg_load_per_task They are the base values in load balance, update them with rq runnable load average, then the load balance will consider runnable load avg naturally. We also try to include the blocked_load_avg as cpu load in balancing, but that cause kbuild performance drop 6% on every Intel machine, and aim7/oltp drop on some of 4 CPU sockets machines. Or only add blocked_load_avg into get_rq_runable_load, hackbench still drop a little on NHM EX. Signed-off-by: Alex Shi Reviewed-by: Gu Zheng Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1371694737-29336-7-git-send-email-alex.shi@intel.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 5 +++-- kernel/sched/proc.c | 17 +++++++++++++++-- 2 files changed, 18 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 9bbc303598ea..e6d82cae4910 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -2963,7 +2963,7 @@ static void dequeue_task_fair(struct rq *rq, struct task_struct *p, int flags) /* Used instead of source_load when we know the type == 0 */ static unsigned long weighted_cpuload(const int cpu) { - return cpu_rq(cpu)->load.weight; + return cpu_rq(cpu)->cfs.runnable_load_avg; } /* @@ -3008,9 +3008,10 @@ static unsigned long cpu_avg_load_per_task(int cpu) { struct rq *rq = cpu_rq(cpu); unsigned long nr_running = ACCESS_ONCE(rq->nr_running); + unsigned long load_avg = rq->cfs.runnable_load_avg; if (nr_running) - return rq->load.weight / nr_running; + return load_avg / nr_running; return 0; } diff --git a/kernel/sched/proc.c b/kernel/sched/proc.c index bb3a6a0b8623..ce5cd4892e43 100644 --- a/kernel/sched/proc.c +++ b/kernel/sched/proc.c @@ -501,6 +501,18 @@ static void __update_cpu_load(struct rq *this_rq, unsigned long this_load, sched_avg_update(this_rq); } +#ifdef CONFIG_SMP +unsigned long get_rq_runnable_load(struct rq *rq) +{ + return rq->cfs.runnable_load_avg; +} +#else +unsigned long get_rq_runnable_load(struct rq *rq) +{ + return rq->load.weight; +} +#endif + #ifdef CONFIG_NO_HZ_COMMON /* * There is no sane way to deal with nohz on smp when using jiffies because the @@ -522,7 +534,7 @@ static void __update_cpu_load(struct rq *this_rq, unsigned long this_load, void update_idle_cpu_load(struct rq *this_rq) { unsigned long curr_jiffies = ACCESS_ONCE(jiffies); - unsigned long load = this_rq->load.weight; + unsigned long load = get_rq_runnable_load(this_rq); unsigned long pending_updates; /* @@ -568,11 +580,12 @@ void update_cpu_load_nohz(void) */ void update_cpu_load_active(struct rq *this_rq) { + unsigned long load = get_rq_runnable_load(this_rq); /* * See the mess around update_idle_cpu_load() / update_cpu_load_nohz(). */ this_rq->last_load_update_tick = jiffies; - __update_cpu_load(this_rq, this_rq->load.weight, 1); + __update_cpu_load(this_rq, load, 1); calc_load_account_active(this_rq); } -- cgit v1.2.3 From a003a25b227d59ded9197ced109517f037d01c27 Mon Sep 17 00:00:00 2001 From: Alex Shi Date: Thu, 20 Jun 2013 10:18:51 +0800 Subject: sched: Consider runnable load average in move_tasks() Aside from using runnable load average in background, move_tasks is also the key function in load balance. We need consider the runnable load average in it in order to make it an apple to apple load comparison. Morten had caught a div u64 bug on ARM, thanks! Thanks-to: Morten Rasmussen Signed-off-by: Alex Shi Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1371694737-29336-8-git-send-email-alex.shi@intel.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index e6d82cae4910..7948bb825985 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -4179,11 +4179,14 @@ static int tg_load_down(struct task_group *tg, void *data) long cpu = (long)data; if (!tg->parent) { - load = cpu_rq(cpu)->load.weight; + load = cpu_rq(cpu)->avg.load_avg_contrib; } else { + unsigned long tmp_rla; + tmp_rla = tg->parent->cfs_rq[cpu]->runnable_load_avg + 1; + load = tg->parent->cfs_rq[cpu]->h_load; - load *= tg->se[cpu]->load.weight; - load /= tg->parent->cfs_rq[cpu]->load.weight + 1; + load *= tg->se[cpu]->avg.load_avg_contrib; + load /= tmp_rla; } tg->cfs_rq[cpu]->h_load = load; @@ -4209,12 +4212,9 @@ static void update_h_load(long cpu) static unsigned long task_h_load(struct task_struct *p) { struct cfs_rq *cfs_rq = task_cfs_rq(p); - unsigned long load; - - load = p->se.load.weight; - load = div_u64(load * cfs_rq->h_load, cfs_rq->load.weight + 1); - return load; + return div64_ul(p->se.avg.load_avg_contrib * cfs_rq->h_load, + cfs_rq->runnable_load_avg + 1); } #else static inline void update_blocked_averages(int cpu) @@ -4227,7 +4227,7 @@ static inline void update_h_load(long cpu) static unsigned long task_h_load(struct task_struct *p) { - return p->se.load.weight; + return p->se.avg.load_avg_contrib; } #endif -- cgit v1.2.3 From 72a4cf20cb71a327c636c7042fdacc25abffc87c Mon Sep 17 00:00:00 2001 From: Alex Shi Date: Thu, 20 Jun 2013 10:18:53 +0800 Subject: sched: Change cfs_rq load avg to unsigned long Since the 'u64 runnable_load_avg, blocked_load_avg' in cfs_rq struct are smaller than 'unsigned long' cfs_rq->load.weight. We don't need u64 vaiables to describe them. unsigned long is more efficient and convenience. Signed-off-by: Alex Shi Reviewed-by: Paul Turner Tested-by: Vincent Guittot Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1371694737-29336-10-git-send-email-alex.shi@intel.com Signed-off-by: Ingo Molnar --- kernel/sched/debug.c | 4 ++-- kernel/sched/fair.c | 7 ++----- kernel/sched/sched.h | 2 +- 3 files changed, 5 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c index 75024a673520..160afdc5cdff 100644 --- a/kernel/sched/debug.c +++ b/kernel/sched/debug.c @@ -211,9 +211,9 @@ void print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq) SEQ_printf(m, " .%-30s: %ld\n", "load", cfs_rq->load.weight); #ifdef CONFIG_FAIR_GROUP_SCHED #ifdef CONFIG_SMP - SEQ_printf(m, " .%-30s: %lld\n", "runnable_load_avg", + SEQ_printf(m, " .%-30s: %ld\n", "runnable_load_avg", cfs_rq->runnable_load_avg); - SEQ_printf(m, " .%-30s: %lld\n", "blocked_load_avg", + SEQ_printf(m, " .%-30s: %ld\n", "blocked_load_avg", cfs_rq->blocked_load_avg); SEQ_printf(m, " .%-30s: %lld\n", "tg_load_avg", (unsigned long long)atomic64_read(&cfs_rq->tg->load_avg)); diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 7948bb825985..f19772de1b1c 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -4181,12 +4181,9 @@ static int tg_load_down(struct task_group *tg, void *data) if (!tg->parent) { load = cpu_rq(cpu)->avg.load_avg_contrib; } else { - unsigned long tmp_rla; - tmp_rla = tg->parent->cfs_rq[cpu]->runnable_load_avg + 1; - load = tg->parent->cfs_rq[cpu]->h_load; - load *= tg->se[cpu]->avg.load_avg_contrib; - load /= tmp_rla; + load = div64_ul(load * tg->se[cpu]->avg.load_avg_contrib, + tg->parent->cfs_rq[cpu]->runnable_load_avg + 1); } tg->cfs_rq[cpu]->h_load = load; diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index 9c65d46504b1..9eb12d9edd35 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -277,7 +277,7 @@ struct cfs_rq { * This allows for the description of both thread and group usage (in * the FAIR_GROUP_SCHED case). */ - u64 runnable_load_avg, blocked_load_avg; + unsigned long runnable_load_avg, blocked_load_avg; atomic64_t decay_counter, removed_load; u64 last_decay; -- cgit v1.2.3 From bf5b986ed4d20428eeec3df4a03dbfebb9b6538c Mon Sep 17 00:00:00 2001 From: Alex Shi Date: Thu, 20 Jun 2013 10:18:54 +0800 Subject: sched/tg: Use 'unsigned long' for load variable in task group Since tg->load_avg is smaller than tg->load_weight, we don't need a atomic64_t variable for load_avg in 32 bit machine. The same reason for cfs_rq->tg_load_contrib. The atomic_long_t/unsigned long variable type are more efficient and convenience for them. Signed-off-by: Alex Shi Tested-by: Vincent Guittot Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1371694737-29336-11-git-send-email-alex.shi@intel.com Signed-off-by: Ingo Molnar --- kernel/sched/debug.c | 6 +++--- kernel/sched/fair.c | 12 ++++++------ kernel/sched/sched.h | 4 ++-- 3 files changed, 11 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c index 160afdc5cdff..d803989defc0 100644 --- a/kernel/sched/debug.c +++ b/kernel/sched/debug.c @@ -215,9 +215,9 @@ void print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq) cfs_rq->runnable_load_avg); SEQ_printf(m, " .%-30s: %ld\n", "blocked_load_avg", cfs_rq->blocked_load_avg); - SEQ_printf(m, " .%-30s: %lld\n", "tg_load_avg", - (unsigned long long)atomic64_read(&cfs_rq->tg->load_avg)); - SEQ_printf(m, " .%-30s: %lld\n", "tg_load_contrib", + SEQ_printf(m, " .%-30s: %ld\n", "tg_load_avg", + atomic_long_read(&cfs_rq->tg->load_avg)); + SEQ_printf(m, " .%-30s: %ld\n", "tg_load_contrib", cfs_rq->tg_load_contrib); SEQ_printf(m, " .%-30s: %d\n", "tg_runnable_contrib", cfs_rq->tg_runnable_contrib); diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index f19772de1b1c..30ccc37112d0 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -1075,7 +1075,7 @@ static inline long calc_tg_weight(struct task_group *tg, struct cfs_rq *cfs_rq) * to gain a more accurate current total weight. See * update_cfs_rq_load_contribution(). */ - tg_weight = atomic64_read(&tg->load_avg); + tg_weight = atomic_long_read(&tg->load_avg); tg_weight -= cfs_rq->tg_load_contrib; tg_weight += cfs_rq->load.weight; @@ -1356,13 +1356,13 @@ static inline void __update_cfs_rq_tg_load_contrib(struct cfs_rq *cfs_rq, int force_update) { struct task_group *tg = cfs_rq->tg; - s64 tg_contrib; + long tg_contrib; tg_contrib = cfs_rq->runnable_load_avg + cfs_rq->blocked_load_avg; tg_contrib -= cfs_rq->tg_load_contrib; - if (force_update || abs64(tg_contrib) > cfs_rq->tg_load_contrib / 8) { - atomic64_add(tg_contrib, &tg->load_avg); + if (force_update || abs(tg_contrib) > cfs_rq->tg_load_contrib / 8) { + atomic_long_add(tg_contrib, &tg->load_avg); cfs_rq->tg_load_contrib += tg_contrib; } } @@ -1397,8 +1397,8 @@ static inline void __update_group_entity_contrib(struct sched_entity *se) u64 contrib; contrib = cfs_rq->tg_load_contrib * tg->shares; - se->avg.load_avg_contrib = div64_u64(contrib, - atomic64_read(&tg->load_avg) + 1); + se->avg.load_avg_contrib = div_u64(contrib, + atomic_long_read(&tg->load_avg) + 1); /* * For group entities we need to compute a correction term in the case diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index 9eb12d9edd35..5585eb25e9a3 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -150,7 +150,7 @@ struct task_group { atomic_t load_weight; #ifdef CONFIG_SMP - atomic64_t load_avg; + atomic_long_t load_avg; atomic_t runnable_avg; #endif #endif @@ -284,7 +284,7 @@ struct cfs_rq { #ifdef CONFIG_FAIR_GROUP_SCHED /* Required to track per-cpu representation of a task_group */ u32 tg_runnable_contrib; - u64 tg_load_contrib; + unsigned long tg_load_contrib; #endif /* CONFIG_FAIR_GROUP_SCHED */ /* -- cgit v1.2.3 From 2509940fd71c2e2915a05052bbdbf2d478364184 Mon Sep 17 00:00:00 2001 From: Alex Shi Date: Thu, 20 Jun 2013 10:18:55 +0800 Subject: sched/cfs_rq: Change atomic64_t removed_load to atomic_long_t Similar to runnable_load_avg, blocked_load_avg variable, long type is enough for removed_load in 64 bit or 32 bit machine. Then we avoid the expensive atomic64 operations on 32 bit machine. Signed-off-by: Alex Shi Reviewed-by: Paul Turner Tested-by: Vincent Guittot Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1371694737-29336-12-git-send-email-alex.shi@intel.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 10 ++++++---- kernel/sched/sched.h | 3 ++- 2 files changed, 8 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 30ccc37112d0..b43474a964c2 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -1517,8 +1517,9 @@ static void update_cfs_rq_blocked_load(struct cfs_rq *cfs_rq, int force_update) if (!decays && !force_update) return; - if (atomic64_read(&cfs_rq->removed_load)) { - u64 removed_load = atomic64_xchg(&cfs_rq->removed_load, 0); + if (atomic_long_read(&cfs_rq->removed_load)) { + unsigned long removed_load; + removed_load = atomic_long_xchg(&cfs_rq->removed_load, 0); subtract_blocked_load_contrib(cfs_rq, removed_load); } @@ -3480,7 +3481,8 @@ migrate_task_rq_fair(struct task_struct *p, int next_cpu) */ if (se->avg.decay_count) { se->avg.decay_count = -__synchronize_entity_decay(se); - atomic64_add(se->avg.load_avg_contrib, &cfs_rq->removed_load); + atomic_long_add(se->avg.load_avg_contrib, + &cfs_rq->removed_load); } } #endif /* CONFIG_SMP */ @@ -5942,7 +5944,7 @@ void init_cfs_rq(struct cfs_rq *cfs_rq) #endif #ifdef CONFIG_SMP atomic64_set(&cfs_rq->decay_counter, 1); - atomic64_set(&cfs_rq->removed_load, 0); + atomic_long_set(&cfs_rq->removed_load, 0); #endif } diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index 5585eb25e9a3..705991906fbe 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -278,8 +278,9 @@ struct cfs_rq { * the FAIR_GROUP_SCHED case). */ unsigned long runnable_load_avg, blocked_load_avg; - atomic64_t decay_counter, removed_load; + atomic64_t decay_counter; u64 last_decay; + atomic_long_t removed_load; #ifdef CONFIG_FAIR_GROUP_SCHED /* Required to track per-cpu representation of a task_group */ -- cgit v1.2.3 From a9cef46a10cc1b84bf2cdf4060766d858c0439d8 Mon Sep 17 00:00:00 2001 From: Alex Shi Date: Thu, 20 Jun 2013 10:18:56 +0800 Subject: sched/tg: Remove tg.load_weight Since no one use it. Signed-off-by: Alex Shi Reviewed-by: Paul Turner Tested-by: Vincent Guittot Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1371694737-29336-13-git-send-email-alex.shi@intel.com Signed-off-by: Ingo Molnar --- kernel/sched/sched.h | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index 705991906fbe..ef0a7b2439dd 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -148,7 +148,6 @@ struct task_group { struct cfs_rq **cfs_rq; unsigned long shares; - atomic_t load_weight; #ifdef CONFIG_SMP atomic_long_t load_avg; atomic_t runnable_avg; -- cgit v1.2.3 From a9dc5d0e33c677619e4b97a38c23db1a42857905 Mon Sep 17 00:00:00 2001 From: Alex Shi Date: Thu, 20 Jun 2013 10:18:57 +0800 Subject: sched: Change get_rq_runnable_load() to static and inline Based-on-patch-by: Fengguang Wu Signed-off-by: Alex Shi Tested-by: Vincent Guittot Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1371694737-29336-14-git-send-email-alex.shi@intel.com Signed-off-by: Ingo Molnar --- kernel/sched/proc.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/proc.c b/kernel/sched/proc.c index ce5cd4892e43..16f5a30f9c88 100644 --- a/kernel/sched/proc.c +++ b/kernel/sched/proc.c @@ -502,12 +502,12 @@ static void __update_cpu_load(struct rq *this_rq, unsigned long this_load, } #ifdef CONFIG_SMP -unsigned long get_rq_runnable_load(struct rq *rq) +static inline unsigned long get_rq_runnable_load(struct rq *rq) { return rq->cfs.runnable_load_avg; } #else -unsigned long get_rq_runnable_load(struct rq *rq) +static inline unsigned long get_rq_runnable_load(struct rq *rq) { return rq->load.weight; } -- cgit v1.2.3 From 939fd731eb88a0cdd9058d0b0143563172a217d7 Mon Sep 17 00:00:00 2001 From: Kamalesh Babulal Date: Tue, 25 Jun 2013 13:33:36 +0530 Subject: sched/debug: Add load-tracking statistics to task At present we print per-entity load-tracking statistics for cfs_rq of cgroups/runqueues. Given that per task statistics is maintained, it can be used to know the contribution made by the task to its parenting cfs_rq level. This patch adds per-task load-tracking statistics to /proc//sched. Signed-off-by: Kamalesh Babulal Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/20130625080336.GA20175@linux.vnet.ibm.com Signed-off-by: Ingo Molnar --- kernel/sched/debug.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'kernel') diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c index d803989defc0..626320985366 100644 --- a/kernel/sched/debug.c +++ b/kernel/sched/debug.c @@ -566,6 +566,12 @@ void proc_sched_show_task(struct task_struct *p, struct seq_file *m) "nr_involuntary_switches", (long long)p->nivcsw); P(se.load.weight); +#if defined(CONFIG_SMP) && defined(CONFIG_FAIR_GROUP_SCHED) + P(se.avg.runnable_avg_sum); + P(se.avg.runnable_avg_period); + P(se.avg.load_avg_contrib); + P(se.avg.decay_count); +#endif P(policy); P(prio); #undef PN -- cgit v1.2.3 From 0fc576d592bd137437fdeb059738b789e642b744 Mon Sep 17 00:00:00 2001 From: Kamalesh Babulal Date: Thu, 27 Jun 2013 11:24:18 +0530 Subject: sched/fair: Fix typo describing flags in enqueue_entity Fix spelling of 'calling' in description of se flags in enqueue_entity(). Signed-off-by: Kamalesh Babulal Cc: peterz@infradead.org Link: http://lkml.kernel.org/r/20130627055418.GA18582@linux.vnet.ibm.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index b43474a964c2..f77f9c527449 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -1760,7 +1760,7 @@ enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags) { /* * Update the normalized vruntime before updating min_vruntime - * through callig update_curr(). + * through calling update_curr(). */ if (!(flags & ENQUEUE_WAKEUP) || (flags & ENQUEUE_WAKING)) se->vruntime += cfs_rq->min_vruntime; -- cgit v1.2.3 From 9dceefe483d7640ba0bbf3e53d1db880e7469aba Mon Sep 17 00:00:00 2001 From: Shuah Khan Date: Wed, 26 Jun 2013 16:27:35 -0600 Subject: PM / Sleep: Warn about system time after resume with pm_trace pm_trace uses the system's Real Time Clock (RTC) to save the magic number. The reason for this is that the RTC is the only reliably available piece of hardware during resume operations where a value can be set that will survive a reboot. Consequence is that after a resume (even if it is successful) your system clock will have a value corresponding to the magic number instead of the correct date/time! It is therefore advisable to use a program like ntp-date or rdate to reset the correct date/time from an external time source when using this trace option. There is no run-time message to warn users of the consequences of enabling pm_trace. Adding a warning message to pm_trace_store() will serve as a reminder to users to set the system date and time after resume. Signed-off-by: Shuah Khan Acked-by: Pavel Machek Signed-off-by: Rafael J. Wysocki --- kernel/power/main.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'kernel') diff --git a/kernel/power/main.c b/kernel/power/main.c index 0828070d38b4..1d1bf630e6e9 100644 --- a/kernel/power/main.c +++ b/kernel/power/main.c @@ -530,6 +530,10 @@ pm_trace_store(struct kobject *kobj, struct kobj_attribute *attr, if (sscanf(buf, "%d", &val) == 1) { pm_trace_enabled = !!val; + if (pm_trace_enabled) { + pr_warn("PM: Enabling pm_trace changes system date and time during resume.\n" + "PM: Correct system time has to be restored manually after resume.\n"); + } return n; } return -EINVAL; -- cgit v1.2.3 From e2bd416f6246d11be29999c177d2534943a5c2df Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 27 Jun 2013 19:37:23 -0700 Subject: cgroup: fix deadlock on cgroup_mutex via drop_parsed_module_refcounts() eb178d06332 ("cgroup: grab cgroup_mutex in drop_parsed_module_refcounts()") made drop_parsed_module_refcounts() grab cgroup_mutex to make lockdep assertion in for_each_subsys() happy. Unfortunately, cgroup_remount() calls the function while holding cgroup_mutex in its failure path leading to the following deadlock. # mount -t cgroup -o remount,memory,blkio cgroup blkio cgroup: option changes via remount are deprecated (pid=525 comm=mount) ============================================= [ INFO: possible recursive locking detected ] 3.10.0-rc4-work+ #1 Not tainted --------------------------------------------- mount/525 is trying to acquire lock: (cgroup_mutex){+.+.+.}, at: [] drop_parsed_module_refcounts+0x21/0xb0 but task is already holding lock: (cgroup_mutex){+.+.+.}, at: [] cgroup_remount+0x51/0x200 other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock(cgroup_mutex); lock(cgroup_mutex); *** DEADLOCK *** May be due to missing lock nesting notation 4 locks held by mount/525: #0: (&type->s_umount_key#30){+.+...}, at: [] do_mount+0x2bd/0xa30 #1: (&sb->s_type->i_mutex_key#9){+.+.+.}, at: [] cgroup_remount+0x43/0x200 #2: (cgroup_mutex){+.+.+.}, at: [] cgroup_remount+0x51/0x200 #3: (cgroup_root_mutex){+.+.+.}, at: [] cgroup_remount+0x5f/0x200 stack backtrace: CPU: 2 PID: 525 Comm: mount Not tainted 3.10.0-rc4-work+ #1 Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011 ffffffff829651f0 ffff88000ec2fc28 ffffffff81c24bb1 ffff88000ec2fce8 ffffffff810f420d 0000000000000006 0000000000000001 0000000000000056 ffff8800153b4640 ffff880000000000 ffffffff81c2e468 ffff8800153b4640 Call Trace: [] dump_stack+0x19/0x1b [] __lock_acquire+0x15dd/0x1e60 [] lock_acquire+0x9c/0x1f0 [] mutex_lock_nested+0x65/0x410 [] drop_parsed_module_refcounts+0x21/0xb0 [] cgroup_remount+0x1ae/0x200 [] do_remount_sb+0x82/0x190 [] do_mount+0x5f1/0xa30 [] SyS_mount+0x83/0xc0 [] system_call_fastpath+0x16/0x1b Fix it by moving the drop_parsed_module_refcounts() invocation outside cgroup_mutex. Signed-off-by: Tejun Heo --- kernel/cgroup.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 4ed86773fff7..1b7b567208cd 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1365,7 +1365,6 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) if (opts.flags != root->flags || (opts.name && strcmp(opts.name, root->name))) { ret = -EINVAL; - drop_parsed_module_refcounts(opts.subsys_mask); goto out_unlock; } @@ -1380,7 +1379,6 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) if (ret) { /* rebind_subsystems failed, re-populate the removed files */ cgroup_populate_dir(cgrp, false, removed_mask); - drop_parsed_module_refcounts(opts.subsys_mask); goto out_unlock; } @@ -1395,6 +1393,8 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) mutex_unlock(&cgroup_root_mutex); mutex_unlock(&cgroup_mutex); mutex_unlock(&cgrp->dentry->d_inode->i_mutex); + if (ret) + drop_parsed_module_refcounts(opts.subsys_mask); return ret; } -- cgit v1.2.3 From 0ce6cba35777cf96a54ce0d5856dc962566b8717 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 27 Jun 2013 19:37:26 -0700 Subject: cgroup: CGRP_ROOT_SUBSYS_BOUND should be ignored when comparing mount options 1672d04070 ("cgroup: fix cgroupfs_root early destruction path") introduced CGRP_ROOT_SUBSYS_BOUND which is used to mark completion of subsys binding on a new root; however, this broke remounts. cgroup_remount() doesn't allow changing root options via remount and CGRP_ROOT_SUBSYS_BOUND, which is set on all fully initialized roots, makes the function reject all remounts. Fix it by putting the options part in the lower 16 bits of root->flags and masking the comparions. While at it, make cgroup_remount() emit an error message explaining why it's rejecting a remount request, so that it's less of a mystery. Signed-off-by: Tejun Heo --- kernel/cgroup.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 1b7b567208cd..5a2fcf5bcc4a 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1362,8 +1362,11 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) removed_mask = root->subsys_mask & ~opts.subsys_mask; /* Don't allow flags or name to change at remount */ - if (opts.flags != root->flags || + if (((opts.flags ^ root->flags) & CGRP_ROOT_OPTION_MASK) || (opts.name && strcmp(opts.name, root->name))) { + pr_err("cgroup: option or name mismatch, new: 0x%lx \"%s\", old: 0x%lx \"%s\"\n", + opts.flags & CGRP_ROOT_OPTION_MASK, opts.name ?: "", + root->flags & CGRP_ROOT_OPTION_MASK, root->name); ret = -EINVAL; goto out_unlock; } -- cgit v1.2.3 From add332a1523a09cf6d429933f1e2fb4ccdfe6479 Mon Sep 17 00:00:00 2001 From: Kamalesh Babulal Date: Thu, 27 Jun 2013 22:20:05 +0530 Subject: sched/debug: Fix formatting of /proc//sched This patch alters format string's width, to align all statistics at par with the longest struct sched_statistic member name under /proc//sched. Signed-off-by: Kamalesh Babulal Cc: peterz@infradead.org Link: http://lkml.kernel.org/r/20130627165005.GA15583@linux.vnet.ibm.com Signed-off-by: Ingo Molnar --- kernel/sched/debug.c | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c index 626320985366..159561415d13 100644 --- a/kernel/sched/debug.c +++ b/kernel/sched/debug.c @@ -493,15 +493,16 @@ void proc_sched_show_task(struct task_struct *p, struct seq_file *m) SEQ_printf(m, "%s (%d, #threads: %d)\n", p->comm, p->pid, get_nr_threads(p)); SEQ_printf(m, - "---------------------------------------------------------\n"); + "---------------------------------------------------------" + "----------\n"); #define __P(F) \ - SEQ_printf(m, "%-35s:%21Ld\n", #F, (long long)F) + SEQ_printf(m, "%-45s:%21Ld\n", #F, (long long)F) #define P(F) \ - SEQ_printf(m, "%-35s:%21Ld\n", #F, (long long)p->F) + SEQ_printf(m, "%-45s:%21Ld\n", #F, (long long)p->F) #define __PN(F) \ - SEQ_printf(m, "%-35s:%14Ld.%06ld\n", #F, SPLIT_NS((long long)F)) + SEQ_printf(m, "%-45s:%14Ld.%06ld\n", #F, SPLIT_NS((long long)F)) #define PN(F) \ - SEQ_printf(m, "%-35s:%14Ld.%06ld\n", #F, SPLIT_NS((long long)p->F)) + SEQ_printf(m, "%-45s:%14Ld.%06ld\n", #F, SPLIT_NS((long long)p->F)) PN(se.exec_start); PN(se.vruntime); @@ -560,9 +561,9 @@ void proc_sched_show_task(struct task_struct *p, struct seq_file *m) } #endif __P(nr_switches); - SEQ_printf(m, "%-35s:%21Ld\n", + SEQ_printf(m, "%-45s:%21Ld\n", "nr_voluntary_switches", (long long)p->nvcsw); - SEQ_printf(m, "%-35s:%21Ld\n", + SEQ_printf(m, "%-45s:%21Ld\n", "nr_involuntary_switches", (long long)p->nivcsw); P(se.load.weight); @@ -585,7 +586,7 @@ void proc_sched_show_task(struct task_struct *p, struct seq_file *m) t0 = cpu_clock(this_cpu); t1 = cpu_clock(this_cpu); - SEQ_printf(m, "%-35s:%21Ld\n", + SEQ_printf(m, "%-45s:%21Ld\n", "clock-delta", (long long)(t1-t0)); } } -- cgit v1.2.3 From 2779db8d37d4b542d9ca2575f5f178dbeaca6c86 Mon Sep 17 00:00:00 2001 From: Ben Hutchings Date: Fri, 28 Jun 2013 02:40:30 +0100 Subject: genirq: Fix can_request_irq() for IRQs without an action Commit 02725e7471b8 ('genirq: Use irq_get/put functions'), inadvertently changed can_request_irq() to return 0 for IRQs that have no action. This causes pcibios_lookup_irq() to select only IRQs that already have an action with IRQF_SHARED set, or to fail if there are none. Change can_request_irq() to return 1 for IRQs that have no action (if the first two conditions are met). Reported-by: Bjarni Ingi Gislason Tested-by: Bjarni Ingi Gislason (against 3.2) Signed-off-by: Ben Hutchings Cc: 709647@bugs.debian.org Cc: stable@vger.kernel.org # 2.6.39+ Link: http://bugs.debian.org/709647 Link: http://lkml.kernel.org/r/1372383630.23847.40.camel@deadeye.wl.decadent.org.uk Signed-off-by: Thomas Gleixner --- kernel/irq/manage.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index e16caa81f887..514bcfd855a8 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -555,9 +555,9 @@ int can_request_irq(unsigned int irq, unsigned long irqflags) return 0; if (irq_settings_can_request(desc)) { - if (desc->action) - if (irqflags & desc->action->flags & IRQF_SHARED) - canrequest =1; + if (!desc->action || + irqflags & desc->action->flags & IRQF_SHARED) + canrequest = 1; } irq_put_desc_unlock(desc, flags); return canrequest; -- cgit v1.2.3 From d55f0cc4c9a70e3105f1e813ab5f221a65ac2ec3 Mon Sep 17 00:00:00 2001 From: Fabio Estevam Date: Fri, 28 Jun 2013 00:23:09 -0300 Subject: genirq: generic-chip: Export some irq_gc_ functions When building imx_v6_v7_defconfig with imx-drm drivers selected as modules, we get the following build errors: ERROR: "irq_gc_mask_clr_bit" [drivers/staging/imx-drm/ipu-v3/imx-ipu-v3.ko] undefined! ERROR: "irq_gc_mask_set_bit" [drivers/staging/imx-drm/ipu-v3/imx-ipu-v3.ko] undefined! ERROR: "irq_gc_ack_set_bit" [drivers/staging/imx-drm/ipu-v3/imx-ipu-v3.ko] undefined! Export the required functions to avoid this problem. Signed-off-by: Fabio Estevam Cc: shawn.guo@linaro.org Cc: kernel@pengutronix.de Link: http://lkml.kernel.org/r/1372389789-7048-1-git-send-email-festevam@gmail.com Signed-off-by: Thomas Gleixner --- kernel/irq/generic-chip.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'kernel') diff --git a/kernel/irq/generic-chip.c b/kernel/irq/generic-chip.c index a746a8f54dae..76ea748324f5 100644 --- a/kernel/irq/generic-chip.c +++ b/kernel/irq/generic-chip.c @@ -62,6 +62,7 @@ void irq_gc_mask_set_bit(struct irq_data *d) irq_reg_writel(*ct->mask_cache, gc->reg_base + ct->regs.mask); irq_gc_unlock(gc); } +EXPORT_SYMBOL_GPL(irq_gc_mask_set_bit); /** * irq_gc_mask_set_mask_bit - Mask chip via clearing bit in mask register @@ -81,6 +82,7 @@ void irq_gc_mask_clr_bit(struct irq_data *d) irq_reg_writel(*ct->mask_cache, gc->reg_base + ct->regs.mask); irq_gc_unlock(gc); } +EXPORT_SYMBOL_GPL(irq_gc_mask_clr_bit); /** * irq_gc_unmask_enable_reg - Unmask chip via enable register @@ -115,6 +117,7 @@ void irq_gc_ack_set_bit(struct irq_data *d) irq_reg_writel(mask, gc->reg_base + ct->regs.ack); irq_gc_unlock(gc); } +EXPORT_SYMBOL_GPL(irq_gc_ack_set_bit); /** * irq_gc_ack_clr_bit - Ack pending interrupt via clearing bit -- cgit v1.2.3 From ccc414f83914178c7ab04ac4d4f0331fe4c37231 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 28 Jun 2013 11:45:15 +0200 Subject: genirq: Add the generic chip to the genirq docbook Signed-off-by: Thomas Gleixner Cc: Randy Dunlap --- kernel/irq/generic-chip.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/generic-chip.c b/kernel/irq/generic-chip.c index 76ea748324f5..1c39eccc1eaf 100644 --- a/kernel/irq/generic-chip.c +++ b/kernel/irq/generic-chip.c @@ -45,7 +45,7 @@ void irq_gc_mask_disable_reg(struct irq_data *d) } /** - * irq_gc_mask_set_mask_bit - Mask chip via setting bit in mask register + * irq_gc_mask_set_bit - Mask chip via setting bit in mask register * @d: irq_data * * Chip has a single mask register. Values of this register are cached @@ -65,7 +65,7 @@ void irq_gc_mask_set_bit(struct irq_data *d) EXPORT_SYMBOL_GPL(irq_gc_mask_set_bit); /** - * irq_gc_mask_set_mask_bit - Mask chip via clearing bit in mask register + * irq_gc_mask_clr_bit - Mask chip via clearing bit in mask register * @d: irq_data * * Chip has a single mask register. Values of this register are cached @@ -167,7 +167,8 @@ void irq_gc_eoi(struct irq_data *d) /** * irq_gc_set_wake - Set/clr wake bit for an interrupt - * @d: irq_data + * @d: irq_data + * @on: Indicates whether the wake bit should be set or cleared * * For chips where the wake from suspend functionality is not * configured in a separate register and the wakeup active state is @@ -339,7 +340,7 @@ EXPORT_SYMBOL_GPL(irq_get_domain_generic_chip); */ static struct lock_class_key irq_nested_lock_class; -/** +/* * irq_map_generic_chip - Map a generic chip for an irq domain */ static int irq_map_generic_chip(struct irq_domain *d, unsigned int virq, @@ -454,7 +455,7 @@ EXPORT_SYMBOL_GPL(irq_setup_generic_chip); /** * irq_setup_alt_chip - Switch to alternative chip * @d: irq_data for this interrupt - * @type Flow type to be initialized + * @type: Flow type to be initialized * * Only to be called from chip->irq_set_type() callbacks. */ -- cgit v1.2.3 From 333bb864f192015a53b5060b829089decd0220ef Mon Sep 17 00:00:00 2001 From: Alex Shi Date: Fri, 28 Jun 2013 19:10:35 +0800 Subject: sched/debug: Remove CONFIG_FAIR_GROUP_SCHED mask Now that we are using runnable load avg in sched balance, we don't need to keep it under CONFIG_FAIR_GROUP_SCHED. Also align the code style to #ifdef instead of #if defined() and reorder the tg output info. Signed-off-by: Alex Shi Cc: pjt@google.com Cc: kamalesh@linux.vnet.ibm.com Cc: peterz@infradead.org Link: http://lkml.kernel.org/r/1372417835-4698-1-git-send-email-alex.shi@intel.com Signed-off-by: Ingo Molnar --- kernel/sched/debug.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c index 159561415d13..e076bddd4c66 100644 --- a/kernel/sched/debug.c +++ b/kernel/sched/debug.c @@ -209,22 +209,24 @@ void print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq) cfs_rq->nr_spread_over); SEQ_printf(m, " .%-30s: %d\n", "nr_running", cfs_rq->nr_running); SEQ_printf(m, " .%-30s: %ld\n", "load", cfs_rq->load.weight); -#ifdef CONFIG_FAIR_GROUP_SCHED #ifdef CONFIG_SMP SEQ_printf(m, " .%-30s: %ld\n", "runnable_load_avg", cfs_rq->runnable_load_avg); SEQ_printf(m, " .%-30s: %ld\n", "blocked_load_avg", cfs_rq->blocked_load_avg); - SEQ_printf(m, " .%-30s: %ld\n", "tg_load_avg", - atomic_long_read(&cfs_rq->tg->load_avg)); +#ifdef CONFIG_FAIR_GROUP_SCHED SEQ_printf(m, " .%-30s: %ld\n", "tg_load_contrib", cfs_rq->tg_load_contrib); SEQ_printf(m, " .%-30s: %d\n", "tg_runnable_contrib", cfs_rq->tg_runnable_contrib); + SEQ_printf(m, " .%-30s: %ld\n", "tg_load_avg", + atomic_long_read(&cfs_rq->tg->load_avg)); SEQ_printf(m, " .%-30s: %d\n", "tg->runnable_avg", atomic_read(&cfs_rq->tg->runnable_avg)); #endif +#endif +#ifdef CONFIG_FAIR_GROUP_SCHED print_cfs_group_stats(m, cpu, cfs_rq->tg); #endif } @@ -567,7 +569,7 @@ void proc_sched_show_task(struct task_struct *p, struct seq_file *m) "nr_involuntary_switches", (long long)p->nivcsw); P(se.load.weight); -#if defined(CONFIG_SMP) && defined(CONFIG_FAIR_GROUP_SCHED) +#ifdef CONFIG_SMP P(se.avg.runnable_avg_sum); P(se.avg.runnable_avg_period); P(se.avg.load_avg_contrib); -- cgit v1.2.3 From d2e08473f2488d53a71c2f53455f934ec6c44c53 Mon Sep 17 00:00:00 2001 From: Davidlohr Bueso Date: Tue, 30 Apr 2013 11:46:09 -0700 Subject: softirq: Use _RET_IP_ Use the already defined macro to pass the function return address. Signed-off-by: Davidlohr Bueso Cc: Frederic Weisbecker Link: http://lkml.kernel.org/r/1367347569.1784.3.camel@buesod1.americas.hpqcorp.net Signed-off-by: Thomas Gleixner --- kernel/softirq.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/softirq.c b/kernel/softirq.c index b5197dcb0dad..a5f88362589b 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -127,8 +127,7 @@ static inline void __local_bh_disable(unsigned long ip, unsigned int cnt) void local_bh_disable(void) { - __local_bh_disable((unsigned long)__builtin_return_address(0), - SOFTIRQ_DISABLE_OFFSET); + __local_bh_disable(_RET_IP_, SOFTIRQ_DISABLE_OFFSET); } EXPORT_SYMBOL(local_bh_disable); @@ -139,7 +138,7 @@ static void __local_bh_enable(unsigned int cnt) WARN_ON_ONCE(!irqs_disabled()); if (softirq_count() == cnt) - trace_softirqs_on((unsigned long)__builtin_return_address(0)); + trace_softirqs_on(_RET_IP_); sub_preempt_count(cnt); } @@ -184,7 +183,7 @@ static inline void _local_bh_enable_ip(unsigned long ip) void local_bh_enable(void) { - _local_bh_enable_ip((unsigned long)__builtin_return_address(0)); + _local_bh_enable_ip(_RET_IP_); } EXPORT_SYMBOL(local_bh_enable); @@ -223,8 +222,7 @@ asmlinkage void __do_softirq(void) pending = local_softirq_pending(); account_irq_enter_time(current); - __local_bh_disable((unsigned long)__builtin_return_address(0), - SOFTIRQ_OFFSET); + __local_bh_disable(_RET_IP_, SOFTIRQ_OFFSET); lockdep_softirq_enter(); cpu = smp_processor_id(); -- cgit v1.2.3 From 9e04d3804d3ac97d8c03a41d78d0f0674b5d01e1 Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Tue, 21 May 2013 20:43:50 +0200 Subject: timer: Fix jiffies wrap behavior of round_jiffies_common() Direct compare of jiffies related values does not work in the wrap around case. Replace it with time_is_after_jiffies(). Signed-off-by: Bart Van Assche Cc: Arjan van de Ven Cc: Stephen Rothwell Link: http://lkml.kernel.org/r/519BC066.5080600@acm.org Cc: stable@vger.kernel.org Signed-off-by: Thomas Gleixner --- kernel/timer.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/timer.c b/kernel/timer.c index 15ffdb3f1948..15bc1b41021d 100644 --- a/kernel/timer.c +++ b/kernel/timer.c @@ -149,9 +149,11 @@ static unsigned long round_jiffies_common(unsigned long j, int cpu, /* now that we have rounded, subtract the extra skew again */ j -= cpu * 3; - if (j <= jiffies) /* rounding ate our timeout entirely; */ - return original; - return j; + /* + * Make sure j is still in the future. Otherwise return the + * unmodified value. + */ + return time_is_after_jiffies(j) ? j : original; } /** -- cgit v1.2.3 From 7c4c3a0f18ba57ea2a2985034532303d2929902a Mon Sep 17 00:00:00 2001 From: David Vrabel Date: Thu, 27 Jun 2013 11:35:44 +0100 Subject: hrtimers: Support resuming with two or more CPUs online (but stopped) hrtimers_resume() only reprograms the timers for the current CPU as it assumes that all other CPUs are offline at this point in the resume process. If other CPUs are online then their timers will not be corrected and they may fire at the wrong time. When running as a Xen guest, this assumption is not true. Non-boot CPUs are only stopped with IRQs disabled instead of offlining them. This is a performance optimization as disabling the CPUs would add an unacceptable amount of additional downtime during a live migration (> 200 ms for a 4 VCPU guest). hrtimers_resume() cannot call on_each_cpu(retrigger_next_event,...) as the other CPUs will be stopped with IRQs disabled. Instead, defer the call to the next softirq. [ tglx: Separated the xen change out ] Signed-off-by: David Vrabel Cc: Konrad Rzeszutek Wilk Cc: John Stultz Cc: Link: http://lkml.kernel.org/r/1372329348-20841-2-git-send-email-david.vrabel@citrix.com Signed-off-by: Thomas Gleixner --- kernel/hrtimer.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/hrtimer.c b/kernel/hrtimer.c index fd4b13b131f8..e86827e94c9a 100644 --- a/kernel/hrtimer.c +++ b/kernel/hrtimer.c @@ -773,15 +773,24 @@ void clock_was_set(void) /* * During resume we might have to reprogram the high resolution timer - * interrupt (on the local CPU): + * interrupt on all online CPUs. However, all other CPUs will be + * stopped with IRQs interrupts disabled so the clock_was_set() call + * must be deferred to the softirq. + * + * The one-shot timer has already been programmed to fire immediately + * (see tick_resume_oneshot()) and this interrupt will trigger the + * softirq to run early enough to correctly reprogram the timers on + * all CPUs. */ void hrtimers_resume(void) { + struct hrtimer_cpu_base *cpu_base = &__get_cpu_var(hrtimer_bases); + WARN_ONCE(!irqs_disabled(), KERN_INFO "hrtimers_resume() called with IRQs enabled!"); - retrigger_next_event(NULL); - timerfd_clock_was_set(); + cpu_base->clock_was_set = 1; + __raise_softirq_irqoff(HRTIMER_SOFTIRQ); } static inline void timer_stats_hrtimer_set_start_info(struct hrtimer *timer) -- cgit v1.2.3 From 04397fe94ad65289884b9862b6a0c722ececaadf Mon Sep 17 00:00:00 2001 From: David Vrabel Date: Thu, 27 Jun 2013 11:35:45 +0100 Subject: timekeeping: Pass flags instead of multiple bools to timekeeping_update() Instead of passing multiple bools to timekeeping_updated(), define flags and use a single 'action' parameter. It is then more obvious what each timekeeping_update() call does. Signed-off-by: David Vrabel Cc: Konrad Rzeszutek Wilk Cc: John Stultz Cc: Link: http://lkml.kernel.org/r/1372329348-20841-3-git-send-email-david.vrabel@citrix.com Signed-off-by: Thomas Gleixner --- kernel/time/timekeeping.c | 21 ++++++++++++--------- 1 file changed, 12 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index 838fc0777b68..d8b23a929e66 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -27,6 +27,9 @@ #include "ntp_internal.h" #include "timekeeping_internal.h" +#define TK_CLEAR_NTP (1 << 0) +#define TK_MIRROR (1 << 1) + static struct timekeeper timekeeper; static DEFINE_RAW_SPINLOCK(timekeeper_lock); static seqcount_t timekeeper_seq; @@ -242,16 +245,16 @@ int pvclock_gtod_unregister_notifier(struct notifier_block *nb) EXPORT_SYMBOL_GPL(pvclock_gtod_unregister_notifier); /* must hold timekeeper_lock */ -static void timekeeping_update(struct timekeeper *tk, bool clearntp, bool mirror) +static void timekeeping_update(struct timekeeper *tk, unsigned int action) { - if (clearntp) { + if (action & TK_CLEAR_NTP) { tk->ntp_error = 0; ntp_clear(); } update_vsyscall(tk); update_pvclock_gtod(tk); - if (mirror) + if (action & TK_MIRROR) memcpy(&shadow_timekeeper, &timekeeper, sizeof(timekeeper)); } @@ -509,7 +512,7 @@ int do_settimeofday(const struct timespec *tv) tk_set_xtime(tk, tv); - timekeeping_update(tk, true, true); + timekeeping_update(tk, TK_CLEAR_NTP | TK_MIRROR); write_seqcount_end(&timekeeper_seq); raw_spin_unlock_irqrestore(&timekeeper_lock, flags); @@ -553,7 +556,7 @@ int timekeeping_inject_offset(struct timespec *ts) tk_set_wall_to_mono(tk, timespec_sub(tk->wall_to_monotonic, *ts)); error: /* even if we error out, we forwarded the time, so call update */ - timekeeping_update(tk, true, true); + timekeeping_update(tk, TK_CLEAR_NTP | TK_MIRROR); write_seqcount_end(&timekeeper_seq); raw_spin_unlock_irqrestore(&timekeeper_lock, flags); @@ -643,7 +646,7 @@ static int change_clocksource(void *data) module_put(new->owner); } } - timekeeping_update(tk, true, true); + timekeeping_update(tk, TK_CLEAR_NTP | TK_MIRROR); write_seqcount_end(&timekeeper_seq); raw_spin_unlock_irqrestore(&timekeeper_lock, flags); @@ -884,7 +887,7 @@ void timekeeping_inject_sleeptime(struct timespec *delta) __timekeeping_inject_sleeptime(tk, delta); - timekeeping_update(tk, true, true); + timekeeping_update(tk, TK_CLEAR_NTP | TK_MIRROR); write_seqcount_end(&timekeeper_seq); raw_spin_unlock_irqrestore(&timekeeper_lock, flags); @@ -966,7 +969,7 @@ static void timekeeping_resume(void) tk->cycle_last = clock->cycle_last = cycle_now; tk->ntp_error = 0; timekeeping_suspended = 0; - timekeeping_update(tk, false, true); + timekeeping_update(tk, TK_MIRROR); write_seqcount_end(&timekeeper_seq); raw_spin_unlock_irqrestore(&timekeeper_lock, flags); @@ -1419,7 +1422,7 @@ static void update_wall_time(void) * updating. */ memcpy(real_tk, tk, sizeof(*tk)); - timekeeping_update(real_tk, false, false); + timekeeping_update(real_tk, 0); write_seqcount_end(&timekeeper_seq); out: raw_spin_unlock_irqrestore(&timekeeper_lock, flags); -- cgit v1.2.3 From 780427f0e113b4c77dfff4d258c05a902cdb0eb9 Mon Sep 17 00:00:00 2001 From: David Vrabel Date: Thu, 27 Jun 2013 11:35:46 +0100 Subject: timekeeping: Indicate that clock was set in the pvclock gtod notifier If the clock was set (stepped), set the action parameter to functions in the pvclock gtod notifier chain to non-zero. This allows the callee to only do work if the clock was stepped. This will be used on Xen as the synchronization of the Xen wallclock to the control domain's (dom0) system time will be done with this notifier and updating on every timer tick is unnecessary and too expensive. Signed-off-by: David Vrabel Cc: Konrad Rzeszutek Wilk Cc: John Stultz Cc: Link: http://lkml.kernel.org/r/1372329348-20841-4-git-send-email-david.vrabel@citrix.com Signed-off-by: Thomas Gleixner --- kernel/time/timekeeping.c | 30 ++++++++++++++++++------------ 1 file changed, 18 insertions(+), 12 deletions(-) (limited to 'kernel') diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index d8b23a929e66..846d0a1f235e 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -29,6 +29,7 @@ #define TK_CLEAR_NTP (1 << 0) #define TK_MIRROR (1 << 1) +#define TK_CLOCK_WAS_SET (1 << 2) static struct timekeeper timekeeper; static DEFINE_RAW_SPINLOCK(timekeeper_lock); @@ -204,9 +205,9 @@ static inline s64 timekeeping_get_ns_raw(struct timekeeper *tk) static RAW_NOTIFIER_HEAD(pvclock_gtod_chain); -static void update_pvclock_gtod(struct timekeeper *tk) +static void update_pvclock_gtod(struct timekeeper *tk, bool was_set) { - raw_notifier_call_chain(&pvclock_gtod_chain, 0, tk); + raw_notifier_call_chain(&pvclock_gtod_chain, was_set, tk); } /** @@ -220,7 +221,7 @@ int pvclock_gtod_register_notifier(struct notifier_block *nb) raw_spin_lock_irqsave(&timekeeper_lock, flags); ret = raw_notifier_chain_register(&pvclock_gtod_chain, nb); - update_pvclock_gtod(tk); + update_pvclock_gtod(tk, true); raw_spin_unlock_irqrestore(&timekeeper_lock, flags); return ret; @@ -252,7 +253,7 @@ static void timekeeping_update(struct timekeeper *tk, unsigned int action) ntp_clear(); } update_vsyscall(tk); - update_pvclock_gtod(tk); + update_pvclock_gtod(tk, action & TK_CLOCK_WAS_SET); if (action & TK_MIRROR) memcpy(&shadow_timekeeper, &timekeeper, sizeof(timekeeper)); @@ -512,7 +513,7 @@ int do_settimeofday(const struct timespec *tv) tk_set_xtime(tk, tv); - timekeeping_update(tk, TK_CLEAR_NTP | TK_MIRROR); + timekeeping_update(tk, TK_CLEAR_NTP | TK_MIRROR | TK_CLOCK_WAS_SET); write_seqcount_end(&timekeeper_seq); raw_spin_unlock_irqrestore(&timekeeper_lock, flags); @@ -556,7 +557,7 @@ int timekeeping_inject_offset(struct timespec *ts) tk_set_wall_to_mono(tk, timespec_sub(tk->wall_to_monotonic, *ts)); error: /* even if we error out, we forwarded the time, so call update */ - timekeeping_update(tk, TK_CLEAR_NTP | TK_MIRROR); + timekeeping_update(tk, TK_CLEAR_NTP | TK_MIRROR | TK_CLOCK_WAS_SET); write_seqcount_end(&timekeeper_seq); raw_spin_unlock_irqrestore(&timekeeper_lock, flags); @@ -646,7 +647,7 @@ static int change_clocksource(void *data) module_put(new->owner); } } - timekeeping_update(tk, TK_CLEAR_NTP | TK_MIRROR); + timekeeping_update(tk, TK_CLEAR_NTP | TK_MIRROR | TK_CLOCK_WAS_SET); write_seqcount_end(&timekeeper_seq); raw_spin_unlock_irqrestore(&timekeeper_lock, flags); @@ -887,7 +888,7 @@ void timekeeping_inject_sleeptime(struct timespec *delta) __timekeeping_inject_sleeptime(tk, delta); - timekeeping_update(tk, TK_CLEAR_NTP | TK_MIRROR); + timekeeping_update(tk, TK_CLEAR_NTP | TK_MIRROR | TK_CLOCK_WAS_SET); write_seqcount_end(&timekeeper_seq); raw_spin_unlock_irqrestore(&timekeeper_lock, flags); @@ -969,7 +970,7 @@ static void timekeeping_resume(void) tk->cycle_last = clock->cycle_last = cycle_now; tk->ntp_error = 0; timekeeping_suspended = 0; - timekeeping_update(tk, TK_MIRROR); + timekeeping_update(tk, TK_MIRROR | TK_CLOCK_WAS_SET); write_seqcount_end(&timekeeper_seq); raw_spin_unlock_irqrestore(&timekeeper_lock, flags); @@ -1243,9 +1244,10 @@ out_adjust: * It also calls into the NTP code to handle leapsecond processing. * */ -static inline void accumulate_nsecs_to_secs(struct timekeeper *tk) +static inline unsigned int accumulate_nsecs_to_secs(struct timekeeper *tk) { u64 nsecps = (u64)NSEC_PER_SEC << tk->shift; + unsigned int action = 0; while (tk->xtime_nsec >= nsecps) { int leap; @@ -1268,8 +1270,10 @@ static inline void accumulate_nsecs_to_secs(struct timekeeper *tk) __timekeeping_set_tai_offset(tk, tk->tai_offset - leap); clock_was_set_delayed(); + action = TK_CLOCK_WAS_SET; } } + return action; } /** @@ -1354,6 +1358,7 @@ static void update_wall_time(void) struct timekeeper *tk = &shadow_timekeeper; cycle_t offset; int shift = 0, maxshift; + unsigned int action; unsigned long flags; raw_spin_lock_irqsave(&timekeeper_lock, flags); @@ -1406,7 +1411,7 @@ static void update_wall_time(void) * Finally, make sure that after the rounding * xtime_nsec isn't larger than NSEC_PER_SEC */ - accumulate_nsecs_to_secs(tk); + action = accumulate_nsecs_to_secs(tk); write_seqcount_begin(&timekeeper_seq); /* Update clock->cycle_last with the new value */ @@ -1422,7 +1427,7 @@ static void update_wall_time(void) * updating. */ memcpy(real_tk, tk, sizeof(*tk)); - timekeeping_update(real_tk, 0); + timekeeping_update(real_tk, action); write_seqcount_end(&timekeeper_seq); out: raw_spin_unlock_irqrestore(&timekeeper_lock, flags); @@ -1684,6 +1689,7 @@ int do_adjtimex(struct timex *txc) if (tai != orig_tai) { __timekeeping_set_tai_offset(tk, tai); + update_pvclock_gtod(tk, true); clock_was_set_delayed(); } write_seqcount_end(&timekeeper_seq); -- cgit v1.2.3 From 706b23bde27a391f0974df2a8351661770fa2e07 Mon Sep 17 00:00:00 2001 From: Mathieu Desnoyers Date: Fri, 28 Jun 2013 09:49:46 -0400 Subject: Fix: kernel/ptrace.c: ptrace_peek_siginfo() missing __put_user() validation This __put_user() could be used by unprivileged processes to write into kernel memory. The issue here is that even if copy_siginfo_to_user() fails, the error code is not checked before __put_user() is executed. Luckily, ptrace_peek_siginfo() has been added within the 3.10-rc cycle, so it has not hit a stable release yet. Signed-off-by: Mathieu Desnoyers Acked-by: Oleg Nesterov Cc: Andrey Vagin Cc: Roland McGrath Cc: Paul McKenney Cc: David Howells Cc: Dave Jones Cc: Pavel Emelyanov Cc: Pedro Alves Cc: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/ptrace.c | 20 +++++++++++--------- 1 file changed, 11 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/ptrace.c b/kernel/ptrace.c index aed981a3f69c..335a7ae697f5 100644 --- a/kernel/ptrace.c +++ b/kernel/ptrace.c @@ -665,20 +665,22 @@ static int ptrace_peek_siginfo(struct task_struct *child, if (unlikely(is_compat_task())) { compat_siginfo_t __user *uinfo = compat_ptr(data); - ret = copy_siginfo_to_user32(uinfo, &info); - ret |= __put_user(info.si_code, &uinfo->si_code); + if (copy_siginfo_to_user32(uinfo, &info) || + __put_user(info.si_code, &uinfo->si_code)) { + ret = -EFAULT; + break; + } + } else #endif { siginfo_t __user *uinfo = (siginfo_t __user *) data; - ret = copy_siginfo_to_user(uinfo, &info); - ret |= __put_user(info.si_code, &uinfo->si_code); - } - - if (ret) { - ret = -EFAULT; - break; + if (copy_siginfo_to_user(uinfo, &info) || + __put_user(info.si_code, &uinfo->si_code)) { + ret = -EFAULT; + break; + } } data += sizeof(siginfo_t); -- cgit v1.2.3 From c7ba8287cd11f2fc9e2feee9e1fac34b7293658f Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Sat, 29 Jun 2013 14:06:10 -0700 Subject: cgroup: CGRP_ROOT_SUBSYS_BOUND should also be ignored when mounting an existing hierarchy 0ce6cba357 ("cgroup: CGRP_ROOT_SUBSYS_BOUND should be ignored when comparing mount options") only updated the remount path but CGRP_ROOT_SUBSYS_BOUND should also be ignored when comparing options while mounting an existing hierarchy. As option mismatch triggers a warning but doesn't fail the mount without sane_behavior, this only triggers a spurious warning message. Fix it by only comparing CGRP_ROOT_OPTION_MASK bits when comparing new and existing root options. Signed-off-by: Tejun Heo --- kernel/cgroup.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 5a2fcf5bcc4a..e5583d10a325 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1703,7 +1703,7 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, */ cgroup_free_root(opts.new_root); - if (root->flags != opts.flags) { + if ((root->flags ^ opts.flags) & CGRP_ROOT_OPTION_MASK) { if ((root->flags | opts.flags) & CGRP_ROOT_SANE_BEHAVIOR) { pr_err("cgroup: sane_behavior: new mount options should match the existing superblock\n"); ret = -EINVAL; -- cgit v1.2.3 From 6e94a780374ed31b280f939d4757e8d7858dff16 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 27 Jun 2013 10:58:31 -0400 Subject: tracing: Failed to create system directory Running the following: # cd /sys/kernel/debug/tracing # echo p:i do_sys_open > kprobe_events # echo p:j schedule >> kprobe_events # cat kprobe_events p:kprobes/i do_sys_open p:kprobes/j schedule # echo p:i do_sys_open >> kprobe_events # cat kprobe_events p:kprobes/j schedule p:kprobes/i do_sys_open # ls /sys/kernel/debug/tracing/events/kprobes/ enable filter j Notice that the 'i' is missing from the kprobes directory. The console produces: "Failed to create system directory kprobes" This is because kprobes passes in a allocated name for the system and the ftrace event subsystem saves off that name instead of creating a duplicate for it. But the kprobes may free the system name making the pointer to it invalid. This bug was introduced by 92edca073c37 "tracing: Use direct field, type and system names" which switched from using kstrdup() on the system name in favor of just keeping apointer to it, as the internal ftrace event system names are static and exist for the life of the computer being booted. Instead of reverting back to duplicating system names again, we can use core_kernel_data() to determine if the passed in name was allocated or static. Then use the MSB of the ref_count to be a flag to keep track if the name was allocated or not. Then we can still save from having to duplicate strings that will always exist, but still copy the ones that may be freed. Cc: stable@vger.kernel.org # 3.10 Reported-by: "zhangwei(Jovi)" Reported-by: Masami Hiramatsu Tested-by: Masami Hiramatsu Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 41 +++++++++++++++++++++++++++++++++++------ 1 file changed, 35 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index f57b01574a30..903a0bf2685e 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -41,6 +41,23 @@ static LIST_HEAD(ftrace_common_fields); static struct kmem_cache *field_cachep; static struct kmem_cache *file_cachep; +#define SYSTEM_FL_FREE_NAME (1 << 31) + +static inline int system_refcount(struct event_subsystem *system) +{ + return system->ref_count & ~SYSTEM_FL_FREE_NAME; +} + +static int system_refcount_inc(struct event_subsystem *system) +{ + return (system->ref_count++) & ~SYSTEM_FL_FREE_NAME; +} + +static int system_refcount_dec(struct event_subsystem *system) +{ + return (--system->ref_count) & ~SYSTEM_FL_FREE_NAME; +} + /* Double loops, do not use break, only goto's work */ #define do_for_each_event_file(tr, file) \ list_for_each_entry(tr, &ftrace_trace_arrays, list) { \ @@ -344,8 +361,8 @@ static void __put_system(struct event_subsystem *system) { struct event_filter *filter = system->filter; - WARN_ON_ONCE(system->ref_count == 0); - if (--system->ref_count) + WARN_ON_ONCE(system_refcount(system) == 0); + if (system_refcount_dec(system)) return; list_del(&system->list); @@ -354,13 +371,15 @@ static void __put_system(struct event_subsystem *system) kfree(filter->filter_string); kfree(filter); } + if (system->ref_count & SYSTEM_FL_FREE_NAME) + kfree(system->name); kfree(system); } static void __get_system(struct event_subsystem *system) { - WARN_ON_ONCE(system->ref_count == 0); - system->ref_count++; + WARN_ON_ONCE(system_refcount(system) == 0); + system_refcount_inc(system); } static void __get_system_dir(struct ftrace_subsystem_dir *dir) @@ -374,7 +393,7 @@ static void __put_system_dir(struct ftrace_subsystem_dir *dir) { WARN_ON_ONCE(dir->ref_count == 0); /* If the subsystem is about to be freed, the dir must be too */ - WARN_ON_ONCE(dir->subsystem->ref_count == 1 && dir->ref_count != 1); + WARN_ON_ONCE(system_refcount(dir->subsystem) == 1 && dir->ref_count != 1); __put_system(dir->subsystem); if (!--dir->ref_count) @@ -1274,7 +1293,15 @@ create_new_subsystem(const char *name) return NULL; system->ref_count = 1; - system->name = name; + + /* Only allocate if dynamic (kprobes and modules) */ + if (!core_kernel_data((unsigned long)name)) { + system->ref_count |= SYSTEM_FL_FREE_NAME; + system->name = kstrdup(name, GFP_KERNEL); + if (!system->name) + goto out_free; + } else + system->name = name; system->filter = NULL; @@ -1287,6 +1314,8 @@ create_new_subsystem(const char *name) return system; out_free: + if (system->ref_count & SYSTEM_FL_FREE_NAME) + kfree(system->name); kfree(system); return NULL; } -- cgit v1.2.3 From 288e984e622336bab8bc3dfdf2f190816362d9a1 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Thu, 20 Jun 2013 19:38:06 +0200 Subject: tracing/kprobes: Avoid perf_trace_buf_*() if ->perf_events is empty perf_trace_buf_prepare() + perf_trace_buf_submit() make no sense if this task/CPU has no active counters. Change kprobe_perf_func() and kretprobe_perf_func() to check call->perf_events beforehand and return if this list is empty. For example, "perf record -e some_probe -p1". Only /sbin/init will report, all other threads which hit the same probe will do perf_trace_buf_prepare/perf_trace_buf_submit just to realize that nobody wants perf_swevent_event(). Link: http://lkml.kernel.org/r/20130620173806.GA13151@redhat.com Acked-by: Masami Hiramatsu Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_kprobe.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index f2374172ba7b..c35bebe53ffe 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -1156,6 +1156,10 @@ kprobe_perf_func(struct trace_probe *tp, struct pt_regs *regs) int size, __size, dsize; int rctx; + head = this_cpu_ptr(call->perf_events); + if (hlist_empty(head)) + return; + dsize = __get_data_size(tp, regs); __size = sizeof(*entry) + tp->size + dsize; size = ALIGN(__size + sizeof(u32), sizeof(u64)); @@ -1171,8 +1175,6 @@ kprobe_perf_func(struct trace_probe *tp, struct pt_regs *regs) entry->ip = (unsigned long)tp->rp.kp.addr; memset(&entry[1], 0, dsize); store_trace_args(sizeof(*entry), tp, regs, (u8 *)&entry[1], dsize); - - head = this_cpu_ptr(call->perf_events); perf_trace_buf_submit(entry, size, rctx, entry->ip, 1, regs, head, NULL); } @@ -1188,6 +1190,10 @@ kretprobe_perf_func(struct trace_probe *tp, struct kretprobe_instance *ri, int size, __size, dsize; int rctx; + head = this_cpu_ptr(call->perf_events); + if (hlist_empty(head)) + return; + dsize = __get_data_size(tp, regs); __size = sizeof(*entry) + tp->size + dsize; size = ALIGN(__size + sizeof(u32), sizeof(u64)); @@ -1203,8 +1209,6 @@ kretprobe_perf_func(struct trace_probe *tp, struct kretprobe_instance *ri, entry->func = (unsigned long)tp->rp.kp.addr; entry->ret_ip = (unsigned long)ri->ret_addr; store_trace_args(sizeof(*entry), tp, regs, (u8 *)&entry[1], dsize); - - head = this_cpu_ptr(call->perf_events); perf_trace_buf_submit(entry, size, rctx, entry->ret_ip, 1, regs, head, NULL); } -- cgit v1.2.3 From 3fe3d6193e7cd7b4dd2bde10772f048bdefea4ee Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Thu, 20 Jun 2013 19:38:09 +0200 Subject: tracing/kprobes: Kill probe_enable_lock enable_trace_probe() and disable_trace_probe() should not worry about serialization, the caller (perf_trace_init or __ftrace_set_clr_event) holds event_mutex. They are also called by kprobe_trace_self_tests_init(), but this __init function can't race with itself or trace_events.c And note that this code depended on event_mutex even before 41a7dd420c which introduced probe_enable_lock. In fact it assumes that the caller kprobe_register() can never race with itself. Otherwise, say, tp->flags manipulations are racy. Link: http://lkml.kernel.org/r/20130620173809.GA13158@redhat.com Acked-by: Masami Hiramatsu Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_kprobe.c | 43 ++++++++++++++++++++----------------------- 1 file changed, 20 insertions(+), 23 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index c35bebe53ffe..282f86cfd304 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -183,16 +183,15 @@ static struct trace_probe *find_trace_probe(const char *event, return NULL; } +/* + * This and enable_trace_probe/disable_trace_probe rely on event_mutex + * held by the caller, __ftrace_set_clr_event(). + */ static int trace_probe_nr_files(struct trace_probe *tp) { - struct ftrace_event_file **file; + struct ftrace_event_file **file = rcu_dereference_raw(tp->files); int ret = 0; - /* - * Since all tp->files updater is protected by probe_enable_lock, - * we don't need to lock an rcu_read_lock. - */ - file = rcu_dereference_raw(tp->files); if (file) while (*(file++)) ret++; @@ -200,8 +199,6 @@ static int trace_probe_nr_files(struct trace_probe *tp) return ret; } -static DEFINE_MUTEX(probe_enable_lock); - /* * Enable trace_probe * if the file is NULL, enable "perf" handler, or enable "trace" handler. @@ -211,8 +208,6 @@ enable_trace_probe(struct trace_probe *tp, struct ftrace_event_file *file) { int ret = 0; - mutex_lock(&probe_enable_lock); - if (file) { struct ftrace_event_file **new, **old; int n = trace_probe_nr_files(tp); @@ -223,7 +218,7 @@ enable_trace_probe(struct trace_probe *tp, struct ftrace_event_file *file) GFP_KERNEL); if (!new) { ret = -ENOMEM; - goto out_unlock; + goto out; } memcpy(new, old, n * sizeof(struct ftrace_event_file *)); new[n] = file; @@ -246,10 +241,7 @@ enable_trace_probe(struct trace_probe *tp, struct ftrace_event_file *file) else ret = enable_kprobe(&tp->rp.kp); } - - out_unlock: - mutex_unlock(&probe_enable_lock); - + out: return ret; } @@ -282,8 +274,6 @@ disable_trace_probe(struct trace_probe *tp, struct ftrace_event_file *file) { int ret = 0; - mutex_lock(&probe_enable_lock); - if (file) { struct ftrace_event_file **new, **old; int n = trace_probe_nr_files(tp); @@ -292,7 +282,7 @@ disable_trace_probe(struct trace_probe *tp, struct ftrace_event_file *file) old = rcu_dereference_raw(tp->files); if (n == 0 || trace_probe_file_index(tp, file) < 0) { ret = -EINVAL; - goto out_unlock; + goto out; } if (n == 1) { /* Remove the last file */ @@ -303,7 +293,7 @@ disable_trace_probe(struct trace_probe *tp, struct ftrace_event_file *file) GFP_KERNEL); if (!new) { ret = -ENOMEM; - goto out_unlock; + goto out; } /* This copy & check loop copies the NULL stopper too */ @@ -326,10 +316,7 @@ disable_trace_probe(struct trace_probe *tp, struct ftrace_event_file *file) else disable_kprobe(&tp->rp.kp); } - - out_unlock: - mutex_unlock(&probe_enable_lock); - + out: return ret; } @@ -1214,6 +1201,12 @@ kretprobe_perf_func(struct trace_probe *tp, struct kretprobe_instance *ri, } #endif /* CONFIG_PERF_EVENTS */ +/* + * called by perf_trace_init() or __ftrace_set_clr_event() under event_mutex. + * + * kprobe_trace_self_tests_init() does enable_trace_probe/disable_trace_probe + * lockless, but we can't race with this __init function. + */ static __kprobes int kprobe_register(struct ftrace_event_call *event, enum trace_reg type, void *data) @@ -1379,6 +1372,10 @@ find_trace_probe_file(struct trace_probe *tp, struct trace_array *tr) return NULL; } +/* + * Nobody but us can call enable_trace_probe/disable_trace_probe at this + * stage, we can do this lockless. + */ static __init int kprobe_trace_self_tests_init(void) { int ret, warn = 0; -- cgit v1.2.3 From a439059610ecd257dba29a612729132e470d118f Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Sat, 29 Jun 2013 00:08:04 -0500 Subject: tracing: Simplify code for showing of soft disabled flag Rather than enumerating each permutation, build the enable state string up from the combination of states. This also allows for the simpler addition of more states. Link: http://lkml.kernel.org/r/9aff5af6dee2f5a40ca30df41c39d5f33e998d7a.1372479499.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 903a0bf2685e..7ee08b95c384 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -638,17 +638,17 @@ event_enable_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { struct ftrace_event_file *file = filp->private_data; - char *buf; + char buf[4] = "0"; - if (file->flags & FTRACE_EVENT_FL_ENABLED) { - if (file->flags & FTRACE_EVENT_FL_SOFT_DISABLED) - buf = "0*\n"; - else if (file->flags & FTRACE_EVENT_FL_SOFT_MODE) - buf = "1*\n"; - else - buf = "1\n"; - } else - buf = "0\n"; + if (file->flags & FTRACE_EVENT_FL_ENABLED && + !(file->flags & FTRACE_EVENT_FL_SOFT_DISABLED)) + strcpy(buf, "1"); + + if (file->flags & FTRACE_EVENT_FL_SOFT_DISABLED || + file->flags & FTRACE_EVENT_FL_SOFT_MODE) + strcat(buf, "*"); + + strcat(buf, "\n"); return simple_read_from_buffer(ubuf, cnt, ppos, buf, strlen(buf)); } -- cgit v1.2.3 From 3baa5e4cf224b8a55220cc841bb475e164b84ceb Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Sat, 29 Jun 2013 00:08:07 -0500 Subject: tracing: Fix disabling of soft disable The comment on the soft disable 'disable' case of __ftrace_event_enable_disable() states that the soft disable bit should be cleared in that case, but currently only the soft mode bit is actually cleared. This essentially leaves the standard non-soft-enable enable/disable paths as the only way to clear the soft disable flag, but the soft disable bit should also be cleared when removing a trigger with '!'. Also, the SOFT_DISABLED bit should never be set if SOFT_MODE is cleared. This fixes the above discrepancies. Link: http://lkml.kernel.org/r/b9c68dd50bc07019e6c67d3f9b29be4ef1b2badb.1372479499.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 7ee08b95c384..5892470bc2ee 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -291,9 +291,11 @@ static int __ftrace_event_enable_disable(struct ftrace_event_file *file, } call->class->reg(call, TRACE_REG_UNREGISTER, file); } - /* If in SOFT_MODE, just set the SOFT_DISABLE_BIT */ + /* If in SOFT_MODE, just set the SOFT_DISABLE_BIT, else clear it */ if (file->flags & FTRACE_EVENT_FL_SOFT_MODE) set_bit(FTRACE_EVENT_FL_SOFT_DISABLED_BIT, &file->flags); + else + clear_bit(FTRACE_EVENT_FL_SOFT_DISABLED_BIT, &file->flags); break; case 1: /* -- cgit v1.2.3 From b04d52e368e2cf526abb2bab61f304eaea126af2 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Thu, 20 Jun 2013 19:38:14 +0200 Subject: tracing/kprobes: Turn trace_probe->files into list_head I think that "ftrace_event_file *trace_probe[]" complicates the code for no reason, turn it into list_head to simplify the code. enable_trace_probe() no longer needs synchronize_sched(). This needs the extra sizeof(list_head) memory for every attached ftrace_event_file, hopefully not a problem in this case. Link: http://lkml.kernel.org/r/20130620173814.GA13165@redhat.com Acked-by: Masami Hiramatsu Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_kprobe.c | 138 ++++++++++++-------------------------------- 1 file changed, 37 insertions(+), 101 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 282f86cfd304..405b5b0f903e 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -35,12 +35,17 @@ struct trace_probe { const char *symbol; /* symbol name */ struct ftrace_event_class class; struct ftrace_event_call call; - struct ftrace_event_file * __rcu *files; + struct list_head files; ssize_t size; /* trace entry size */ unsigned int nr_args; struct probe_arg args[]; }; +struct event_file_link { + struct ftrace_event_file *file; + struct list_head list; +}; + #define SIZEOF_TRACE_PROBE(n) \ (offsetof(struct trace_probe, args) + \ (sizeof(struct probe_arg) * (n))) @@ -150,6 +155,7 @@ static struct trace_probe *alloc_trace_probe(const char *group, goto error; INIT_LIST_HEAD(&tp->list); + INIT_LIST_HEAD(&tp->files); return tp; error: kfree(tp->call.name); @@ -183,22 +189,6 @@ static struct trace_probe *find_trace_probe(const char *event, return NULL; } -/* - * This and enable_trace_probe/disable_trace_probe rely on event_mutex - * held by the caller, __ftrace_set_clr_event(). - */ -static int trace_probe_nr_files(struct trace_probe *tp) -{ - struct ftrace_event_file **file = rcu_dereference_raw(tp->files); - int ret = 0; - - if (file) - while (*(file++)) - ret++; - - return ret; -} - /* * Enable trace_probe * if the file is NULL, enable "perf" handler, or enable "trace" handler. @@ -209,29 +199,18 @@ enable_trace_probe(struct trace_probe *tp, struct ftrace_event_file *file) int ret = 0; if (file) { - struct ftrace_event_file **new, **old; - int n = trace_probe_nr_files(tp); - - old = rcu_dereference_raw(tp->files); - /* 1 is for new one and 1 is for stopper */ - new = kzalloc((n + 2) * sizeof(struct ftrace_event_file *), - GFP_KERNEL); - if (!new) { + struct event_file_link *link; + + link = kmalloc(sizeof(*link), GFP_KERNEL); + if (!link) { ret = -ENOMEM; goto out; } - memcpy(new, old, n * sizeof(struct ftrace_event_file *)); - new[n] = file; - /* The last one keeps a NULL */ - rcu_assign_pointer(tp->files, new); - tp->flags |= TP_FLAG_TRACE; + link->file = file; + list_add_tail_rcu(&link->list, &tp->files); - if (old) { - /* Make sure the probe is done with old files */ - synchronize_sched(); - kfree(old); - } + tp->flags |= TP_FLAG_TRACE; } else tp->flags |= TP_FLAG_PROFILE; @@ -245,24 +224,16 @@ enable_trace_probe(struct trace_probe *tp, struct ftrace_event_file *file) return ret; } -static int -trace_probe_file_index(struct trace_probe *tp, struct ftrace_event_file *file) +static struct event_file_link * +find_event_file_link(struct trace_probe *tp, struct ftrace_event_file *file) { - struct ftrace_event_file **files; - int i; + struct event_file_link *link; - /* - * Since all tp->files updater is protected by probe_enable_lock, - * we don't need to lock an rcu_read_lock. - */ - files = rcu_dereference_raw(tp->files); - if (files) { - for (i = 0; files[i]; i++) - if (files[i] == file) - return i; - } + list_for_each_entry(link, &tp->files, list) + if (link->file == file) + return link; - return -1; + return NULL; } /* @@ -275,38 +246,23 @@ disable_trace_probe(struct trace_probe *tp, struct ftrace_event_file *file) int ret = 0; if (file) { - struct ftrace_event_file **new, **old; - int n = trace_probe_nr_files(tp); - int i, j; + struct event_file_link *link; - old = rcu_dereference_raw(tp->files); - if (n == 0 || trace_probe_file_index(tp, file) < 0) { + link = find_event_file_link(tp, file); + if (!link) { ret = -EINVAL; goto out; } - if (n == 1) { /* Remove the last file */ - tp->flags &= ~TP_FLAG_TRACE; - new = NULL; - } else { - new = kzalloc(n * sizeof(struct ftrace_event_file *), - GFP_KERNEL); - if (!new) { - ret = -ENOMEM; - goto out; - } - - /* This copy & check loop copies the NULL stopper too */ - for (i = 0, j = 0; j < n && i < n + 1; i++) - if (old[i] != file) - new[j++] = old[i]; - } + list_del_rcu(&link->list); + /* synchronize with kprobe_trace_func/kretprobe_trace_func */ + synchronize_sched(); + kfree(link); - rcu_assign_pointer(tp->files, new); + if (!list_empty(&tp->files)) + goto out; - /* Make sure the probe is done with old files */ - synchronize_sched(); - kfree(old); + tp->flags &= ~TP_FLAG_TRACE; } else tp->flags &= ~TP_FLAG_PROFILE; @@ -871,20 +827,10 @@ __kprobe_trace_func(struct trace_probe *tp, struct pt_regs *regs, static __kprobes void kprobe_trace_func(struct trace_probe *tp, struct pt_regs *regs) { - /* - * Note: preempt is already disabled around the kprobe handler. - * However, we still need an smp_read_barrier_depends() corresponding - * to smp_wmb() in rcu_assign_pointer() to access the pointer. - */ - struct ftrace_event_file **file = rcu_dereference_raw(tp->files); - - if (unlikely(!file)) - return; + struct event_file_link *link; - while (*file) { - __kprobe_trace_func(tp, regs, *file); - file++; - } + list_for_each_entry_rcu(link, &tp->files, list) + __kprobe_trace_func(tp, regs, link->file); } /* Kretprobe handler */ @@ -931,20 +877,10 @@ static __kprobes void kretprobe_trace_func(struct trace_probe *tp, struct kretprobe_instance *ri, struct pt_regs *regs) { - /* - * Note: preempt is already disabled around the kprobe handler. - * However, we still need an smp_read_barrier_depends() corresponding - * to smp_wmb() in rcu_assign_pointer() to access the pointer. - */ - struct ftrace_event_file **file = rcu_dereference_raw(tp->files); - - if (unlikely(!file)) - return; + struct event_file_link *link; - while (*file) { - __kretprobe_trace_func(tp, ri, regs, *file); - file++; - } + list_for_each_entry_rcu(link, &tp->files, list) + __kretprobe_trace_func(tp, ri, regs, link->file); } /* Event entry printers */ -- cgit v1.2.3 From 10246fa35d4ffdfe472185d4cbf9c2dfd9a9f023 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Mon, 1 Jul 2013 15:58:24 -0400 Subject: tracing: Use flag buffer_disabled for irqsoff tracer If the ring buffer is disabled and the irqsoff tracer records a trace it will clear out its buffer and lose the data it had previously recorded. Currently there's a callback when writing to the tracing_of file, but if tracing is disabled via the function tracer trigger, it will not inform the irqsoff tracer to stop recording. By using the "mirror" flag (buffer_disabled) in the trace_array, that keeps track of the status of the trace_array's buffer, it gives the irqsoff tracer a fast way to know if it should record a new trace or not. The flag may be a little behind the real state of the buffer, but it should not affect the trace too much. It's more important for the irqsoff tracer to be fast. Reported-by: Dave Jones Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 101 ++++++++++++++++++++++++++++++------------- kernel/trace/trace_irqsoff.c | 4 +- 2 files changed, 72 insertions(+), 33 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index c4c9296b1916..0dc50711d656 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -226,9 +226,24 @@ cycle_t ftrace_now(int cpu) return ts; } +/** + * tracing_is_enabled - Show if global_trace has been disabled + * + * Shows if the global trace has been enabled or not. It uses the + * mirror flag "buffer_disabled" to be used in fast paths such as for + * the irqsoff tracer. But it may be inaccurate due to races. If you + * need to know the accurate state, use tracing_is_on() which is a little + * slower, but accurate. + */ int tracing_is_enabled(void) { - return tracing_is_on(); + /* + * For quick access (irqsoff uses this in fast path), just + * return the mirror variable of the state of the ring buffer. + * It's a little racy, but we don't really care. + */ + smp_rmb(); + return !global_trace.buffer_disabled; } /* @@ -341,6 +356,23 @@ unsigned long trace_flags = TRACE_ITER_PRINT_PARENT | TRACE_ITER_PRINTK | TRACE_ITER_GRAPH_TIME | TRACE_ITER_RECORD_CMD | TRACE_ITER_OVERWRITE | TRACE_ITER_IRQ_INFO | TRACE_ITER_MARKERS | TRACE_ITER_FUNCTION; +void tracer_tracing_on(struct trace_array *tr) +{ + if (tr->trace_buffer.buffer) + ring_buffer_record_on(tr->trace_buffer.buffer); + /* + * This flag is looked at when buffers haven't been allocated + * yet, or by some tracers (like irqsoff), that just want to + * know if the ring buffer has been disabled, but it can handle + * races of where it gets disabled but we still do a record. + * As the check is in the fast path of the tracers, it is more + * important to be fast than accurate. + */ + tr->buffer_disabled = 0; + /* Make the flag seen by readers */ + smp_wmb(); +} + /** * tracing_on - enable tracing buffers * @@ -349,15 +381,7 @@ unsigned long trace_flags = TRACE_ITER_PRINT_PARENT | TRACE_ITER_PRINTK | */ void tracing_on(void) { - if (global_trace.trace_buffer.buffer) - ring_buffer_record_on(global_trace.trace_buffer.buffer); - /* - * This flag is only looked at when buffers haven't been - * allocated yet. We don't really care about the race - * between setting this flag and actually turning - * on the buffer. - */ - global_trace.buffer_disabled = 0; + tracer_tracing_on(&global_trace); } EXPORT_SYMBOL_GPL(tracing_on); @@ -551,6 +575,23 @@ void tracing_snapshot_alloc(void) EXPORT_SYMBOL_GPL(tracing_snapshot_alloc); #endif /* CONFIG_TRACER_SNAPSHOT */ +void tracer_tracing_off(struct trace_array *tr) +{ + if (tr->trace_buffer.buffer) + ring_buffer_record_off(tr->trace_buffer.buffer); + /* + * This flag is looked at when buffers haven't been allocated + * yet, or by some tracers (like irqsoff), that just want to + * know if the ring buffer has been disabled, but it can handle + * races of where it gets disabled but we still do a record. + * As the check is in the fast path of the tracers, it is more + * important to be fast than accurate. + */ + tr->buffer_disabled = 1; + /* Make the flag seen by readers */ + smp_wmb(); +} + /** * tracing_off - turn off tracing buffers * @@ -561,15 +602,7 @@ EXPORT_SYMBOL_GPL(tracing_snapshot_alloc); */ void tracing_off(void) { - if (global_trace.trace_buffer.buffer) - ring_buffer_record_off(global_trace.trace_buffer.buffer); - /* - * This flag is only looked at when buffers haven't been - * allocated yet. We don't really care about the race - * between setting this flag and actually turning - * on the buffer. - */ - global_trace.buffer_disabled = 1; + tracer_tracing_off(&global_trace); } EXPORT_SYMBOL_GPL(tracing_off); @@ -579,14 +612,25 @@ void disable_trace_on_warning(void) tracing_off(); } +/** + * tracer_tracing_is_on - show real state of ring buffer enabled + * @tr : the trace array to know if ring buffer is enabled + * + * Shows real state of the ring buffer if it is enabled or not. + */ +int tracer_tracing_is_on(struct trace_array *tr) +{ + if (tr->trace_buffer.buffer) + return ring_buffer_record_is_on(tr->trace_buffer.buffer); + return !tr->buffer_disabled; +} + /** * tracing_is_on - show state of ring buffers enabled */ int tracing_is_on(void) { - if (global_trace.trace_buffer.buffer) - return ring_buffer_record_is_on(global_trace.trace_buffer.buffer); - return !global_trace.buffer_disabled; + return tracer_tracing_is_on(&global_trace); } EXPORT_SYMBOL_GPL(tracing_is_on); @@ -3958,7 +4002,7 @@ static int tracing_wait_pipe(struct file *filp) * * iter->pos will be 0 if we haven't read anything. */ - if (!tracing_is_enabled() && iter->pos) + if (!tracing_is_on() && iter->pos) break; } @@ -5631,15 +5675,10 @@ rb_simple_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { struct trace_array *tr = filp->private_data; - struct ring_buffer *buffer = tr->trace_buffer.buffer; char buf[64]; int r; - if (buffer) - r = ring_buffer_record_is_on(buffer); - else - r = 0; - + r = tracer_tracing_is_on(tr); r = sprintf(buf, "%d\n", r); return simple_read_from_buffer(ubuf, cnt, ppos, buf, r); @@ -5661,11 +5700,11 @@ rb_simple_write(struct file *filp, const char __user *ubuf, if (buffer) { mutex_lock(&trace_types_lock); if (val) { - ring_buffer_record_on(buffer); + tracer_tracing_on(tr); if (tr->current_trace->start) tr->current_trace->start(tr); } else { - ring_buffer_record_off(buffer); + tracer_tracing_off(tr); if (tr->current_trace->stop) tr->current_trace->stop(tr); } diff --git a/kernel/trace/trace_irqsoff.c b/kernel/trace/trace_irqsoff.c index b19d065a28cb..2aefbee93a6d 100644 --- a/kernel/trace/trace_irqsoff.c +++ b/kernel/trace/trace_irqsoff.c @@ -373,7 +373,7 @@ start_critical_timing(unsigned long ip, unsigned long parent_ip) struct trace_array_cpu *data; unsigned long flags; - if (likely(!tracer_enabled)) + if (!tracer_enabled || !tracing_is_enabled()) return; cpu = raw_smp_processor_id(); @@ -416,7 +416,7 @@ stop_critical_timing(unsigned long ip, unsigned long parent_ip) else return; - if (!tracer_enabled) + if (!tracer_enabled || !tracing_is_enabled()) return; data = per_cpu_ptr(tr->trace_buffer.data, cpu); -- cgit v1.2.3 From cf6735a4b103b801753748531e3658cdc8cafa5e Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Thu, 20 Jun 2013 19:38:11 +0200 Subject: tracing/kprobes: Don't pass addr=ip to perf_trace_buf_submit() kprobe_perf_func() and kretprobe_perf_func() pass addr=ip to perf_trace_buf_submit() for no reason. This sets perf_sample_data->addr for PERF_SAMPLE_ADDR, we already have perf_sample_data->ip initialized if PERF_SAMPLE_IP. Link: http://lkml.kernel.org/r/20130620173811.GA13161@redhat.com Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_kprobe.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 405b5b0f903e..7ed6976493c8 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -1098,8 +1098,7 @@ kprobe_perf_func(struct trace_probe *tp, struct pt_regs *regs) entry->ip = (unsigned long)tp->rp.kp.addr; memset(&entry[1], 0, dsize); store_trace_args(sizeof(*entry), tp, regs, (u8 *)&entry[1], dsize); - perf_trace_buf_submit(entry, size, rctx, - entry->ip, 1, regs, head, NULL); + perf_trace_buf_submit(entry, size, rctx, 0, 1, regs, head, NULL); } /* Kretprobe profile handler */ @@ -1132,8 +1131,7 @@ kretprobe_perf_func(struct trace_probe *tp, struct kretprobe_instance *ri, entry->func = (unsigned long)tp->rp.kp.addr; entry->ret_ip = (unsigned long)ri->ret_addr; store_trace_args(sizeof(*entry), tp, regs, (u8 *)&entry[1], dsize); - perf_trace_buf_submit(entry, size, rctx, - entry->ret_ip, 1, regs, head, NULL); + perf_trace_buf_submit(entry, size, rctx, 0, 1, regs, head, NULL); } #endif /* CONFIG_PERF_EVENTS */ -- cgit v1.2.3 From f1ed7c741fcd0c3d7d318e7c19813d89934b9296 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Thu, 27 Jun 2013 22:18:06 -0400 Subject: ftrace: Do not run selftest if command line parameter is set If the kernel command line ftrace filter parameters are set (ftrace_filter or ftrace_notrace), force the function self test to pass, with a warning why it was forced. If the user adds a filter to the kernel command line, it is assumed that they know what they are doing, and the self test should just not run instead of failing (which disables function tracing) or clearing the filter, as that will probably annoy the user. If the user wants the selftest to run, the message will tell them why it did not. Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 5 +++++ kernel/trace/trace.h | 1 + kernel/trace/trace_selftest.c | 18 ++++++++++++++++-- 3 files changed, 22 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 26e19105cdcc..67708f46baae 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -3537,8 +3537,12 @@ EXPORT_SYMBOL_GPL(ftrace_set_global_notrace); static char ftrace_notrace_buf[FTRACE_FILTER_SIZE] __initdata; static char ftrace_filter_buf[FTRACE_FILTER_SIZE] __initdata; +/* Used by function selftest to not test if filter is set */ +bool ftrace_filter_param __initdata; + static int __init set_ftrace_notrace(char *str) { + ftrace_filter_param = true; strlcpy(ftrace_notrace_buf, str, FTRACE_FILTER_SIZE); return 1; } @@ -3546,6 +3550,7 @@ __setup("ftrace_notrace=", set_ftrace_notrace); static int __init set_ftrace_filter(char *str) { + ftrace_filter_param = true; strlcpy(ftrace_filter_buf, str, FTRACE_FILTER_SIZE); return 1; } diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 711ca7d3e7f1..a88939e666b7 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -776,6 +776,7 @@ print_graph_function_flags(struct trace_iterator *iter, u32 flags) extern struct list_head ftrace_pids; #ifdef CONFIG_FUNCTION_TRACER +extern bool ftrace_filter_param __initdata; static inline int ftrace_trace_task(struct task_struct *task) { if (list_empty(&ftrace_pids)) diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c index 2901e3b88590..a7329b7902f8 100644 --- a/kernel/trace/trace_selftest.c +++ b/kernel/trace/trace_selftest.c @@ -640,13 +640,20 @@ out: * Enable ftrace, sleep 1/10 second, and then read the trace * buffer to see if all is in order. */ -int +__init int trace_selftest_startup_function(struct tracer *trace, struct trace_array *tr) { int save_ftrace_enabled = ftrace_enabled; unsigned long count; int ret; +#ifdef CONFIG_DYNAMIC_FTRACE + if (ftrace_filter_param) { + printk(KERN_CONT " ... kernel command line filter set: force PASS ... "); + return 0; + } +#endif + /* make sure msleep has been recorded */ msleep(1); @@ -727,13 +734,20 @@ static int trace_graph_entry_watchdog(struct ftrace_graph_ent *trace) * Pretty much the same than for the function tracer from which the selftest * has been borrowed. */ -int +__init int trace_selftest_startup_function_graph(struct tracer *trace, struct trace_array *tr) { int ret; unsigned long count; +#ifdef CONFIG_DYNAMIC_FTRACE + if (ftrace_filter_param) { + printk(KERN_CONT " ... kernel command line filter set: force PASS ... "); + return 0; + } +#endif + /* * Simulate the init() callback but we attach a watchdog callback * to detect and recover from possible hangs -- cgit v1.2.3 From 2d71619c59fac95a5415a326162fa046161b938c Mon Sep 17 00:00:00 2001 From: Alexander Z Lam Date: Mon, 1 Jul 2013 15:31:24 -0700 Subject: tracing: Make trace_marker use the correct per-instance buffer The trace_marker file was present for each new instance created, but it added the trace mark to the global trace buffer instead of to the instance's buffer. Link: http://lkml.kernel.org/r/1372717885-4543-2-git-send-email-azl@google.com Cc: David Sharp Cc: Vaibhav Nagarnaik Cc: Alexander Z Lam Cc: stable@vger.kernel.org # 3.10 Signed-off-by: Alexander Z Lam Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 0dc50711d656..e04e7119633d 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -4391,6 +4391,7 @@ tracing_mark_write(struct file *filp, const char __user *ubuf, size_t cnt, loff_t *fpos) { unsigned long addr = (unsigned long)ubuf; + struct trace_array *tr = filp->private_data; struct ring_buffer_event *event; struct ring_buffer *buffer; struct print_entry *entry; @@ -4450,7 +4451,7 @@ tracing_mark_write(struct file *filp, const char __user *ubuf, local_save_flags(irq_flags); size = sizeof(*entry) + cnt + 2; /* possible \n added */ - buffer = global_trace.trace_buffer.buffer; + buffer = tr->trace_buffer.buffer; event = trace_buffer_lock_reserve(buffer, TRACE_PRINT, size, irq_flags, preempt_count()); if (!event) { -- cgit v1.2.3 From a82274151af2b075163e3c42c828529dee311487 Mon Sep 17 00:00:00 2001 From: Alexander Z Lam Date: Mon, 1 Jul 2013 19:37:54 -0700 Subject: tracing: Protect ftrace_trace_arrays list in trace_events.c There are multiple places where the ftrace_trace_arrays list is accessed in trace_events.c without the trace_types_lock held. Link: http://lkml.kernel.org/r/1372732674-22726-1-git-send-email-azl@google.com Cc: Vaibhav Nagarnaik Cc: David Sharp Cc: Alexander Z Lam Cc: stable@vger.kernel.org # 3.10 Signed-off-by: Alexander Z Lam Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 2 +- kernel/trace/trace.h | 2 ++ kernel/trace/trace_events.c | 11 ++++++++++- 3 files changed, 13 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index e04e7119633d..e36da7ff59bf 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -266,7 +266,7 @@ static struct tracer *trace_types __read_mostly; /* * trace_types_lock is used to protect the trace_types list. */ -static DEFINE_MUTEX(trace_types_lock); +DEFINE_MUTEX(trace_types_lock); /* * serialize the access of the ring buffer diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index a88939e666b7..2c3cba59552d 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -224,6 +224,8 @@ enum { extern struct list_head ftrace_trace_arrays; +extern struct mutex trace_types_lock; + /* * The global tracer (top) should be the first trace array added, * but we check the flag anyway. diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 5892470bc2ee..35c6f23c71b2 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -1008,6 +1008,7 @@ static int subsystem_open(struct inode *inode, struct file *filp) int ret; /* Make sure the system still exists */ + mutex_lock(&trace_types_lock); mutex_lock(&event_mutex); list_for_each_entry(tr, &ftrace_trace_arrays, list) { list_for_each_entry(dir, &tr->systems, list) { @@ -1023,6 +1024,7 @@ static int subsystem_open(struct inode *inode, struct file *filp) } exit_loop: mutex_unlock(&event_mutex); + mutex_unlock(&trace_types_lock); if (!system) return -ENODEV; @@ -1617,6 +1619,7 @@ static void __add_event_to_tracers(struct ftrace_event_call *call, int trace_add_event_call(struct ftrace_event_call *call) { int ret; + mutex_lock(&trace_types_lock); mutex_lock(&event_mutex); ret = __register_event(call, NULL); @@ -1624,11 +1627,13 @@ int trace_add_event_call(struct ftrace_event_call *call) __add_event_to_tracers(call, NULL); mutex_unlock(&event_mutex); + mutex_unlock(&trace_types_lock); return ret; } /* - * Must be called under locking both of event_mutex and trace_event_sem. + * Must be called under locking of trace_types_lock, event_mutex and + * trace_event_sem. */ static void __trace_remove_event_call(struct ftrace_event_call *call) { @@ -1640,11 +1645,13 @@ static void __trace_remove_event_call(struct ftrace_event_call *call) /* Remove an event_call */ void trace_remove_event_call(struct ftrace_event_call *call) { + mutex_lock(&trace_types_lock); mutex_lock(&event_mutex); down_write(&trace_event_sem); __trace_remove_event_call(call); up_write(&trace_event_sem); mutex_unlock(&event_mutex); + mutex_unlock(&trace_types_lock); } #define for_each_event(event, start, end) \ @@ -1788,6 +1795,7 @@ static int trace_module_notify(struct notifier_block *self, { struct module *mod = data; + mutex_lock(&trace_types_lock); mutex_lock(&event_mutex); switch (val) { case MODULE_STATE_COMING: @@ -1798,6 +1806,7 @@ static int trace_module_notify(struct notifier_block *self, break; } mutex_unlock(&event_mutex); + mutex_unlock(&trace_types_lock); return 0; } -- cgit v1.2.3 From 4f6de4d51f4a3ab06a85e91e708cc89a513ef30c Mon Sep 17 00:00:00 2001 From: Mathias Krause Date: Tue, 2 Jul 2013 15:35:11 +0930 Subject: module: don't modify argument of module_kallsyms_lookup_name() If we pass a pointer to a const string in the form "module:symbol" module_kallsyms_lookup_name() will try to split the string at the colon, i.e., will try to modify r/o data. That will, in fact, fail on a kernel with enabled CONFIG_DEBUG_RODATA. Avoid modifying the passed string in module_kallsyms_lookup_name(), modify find_module_all() instead to pass it the module name length. Signed-off-by: Mathias Krause Signed-off-by: Rusty Russell --- kernel/module.c | 15 +++++++-------- 1 file changed, 7 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/module.c b/kernel/module.c index b049939177f6..a1951aba7a03 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -455,7 +455,7 @@ const struct kernel_symbol *find_symbol(const char *name, EXPORT_SYMBOL_GPL(find_symbol); /* Search for module by name: must hold module_mutex. */ -static struct module *find_module_all(const char *name, +static struct module *find_module_all(const char *name, size_t len, bool even_unformed) { struct module *mod; @@ -463,7 +463,7 @@ static struct module *find_module_all(const char *name, list_for_each_entry(mod, &modules, list) { if (!even_unformed && mod->state == MODULE_STATE_UNFORMED) continue; - if (strcmp(mod->name, name) == 0) + if (strlen(mod->name) == len && !memcmp(mod->name, name, len)) return mod; } return NULL; @@ -471,7 +471,7 @@ static struct module *find_module_all(const char *name, struct module *find_module(const char *name) { - return find_module_all(name, false); + return find_module_all(name, strlen(name), false); } EXPORT_SYMBOL_GPL(find_module); @@ -3027,7 +3027,7 @@ static bool finished_loading(const char *name) bool ret; mutex_lock(&module_mutex); - mod = find_module_all(name, true); + mod = find_module_all(name, strlen(name), true); ret = !mod || mod->state == MODULE_STATE_LIVE || mod->state == MODULE_STATE_GOING; mutex_unlock(&module_mutex); @@ -3165,7 +3165,8 @@ static int add_unformed_module(struct module *mod) again: mutex_lock(&module_mutex); - if ((old = find_module_all(mod->name, true)) != NULL) { + old = find_module_all(mod->name, strlen(mod->name), true); + if (old != NULL) { if (old->state == MODULE_STATE_COMING || old->state == MODULE_STATE_UNFORMED) { /* Wait in case it fails to load. */ @@ -3576,10 +3577,8 @@ unsigned long module_kallsyms_lookup_name(const char *name) /* Don't lock: we're in enough trouble already. */ preempt_disable(); if ((colon = strchr(name, ':')) != NULL) { - *colon = '\0'; - if ((mod = find_module(name)) != NULL) + if ((mod = find_module_all(name, colon - name, false)) != NULL) ret = mod_find_symname(mod, colon+1); - *colon = ':'; } else { list_for_each_entry_rcu(mod, &modules, list) { if (mod->state == MODULE_STATE_UNFORMED) -- cgit v1.2.3 From b634d130e46a093ddf716ae9cf1bfa258ede36cf Mon Sep 17 00:00:00 2001 From: Jean Delvare Date: Tue, 2 Jul 2013 15:35:11 +0930 Subject: There is no /sys/parameters There is no such path as /sys/parameters, module parameters live in /sys/module/*/parameters. Signed-off-by: Jean Delvare Cc: Rusty Russell Signed-off-by: Rusty Russell --- kernel/params.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/params.c b/kernel/params.c index 53b958fcd639..440e65d1a544 100644 --- a/kernel/params.c +++ b/kernel/params.c @@ -787,7 +787,7 @@ static void __init kernel_add_sysfs_param(const char *name, } /* - * param_sysfs_builtin - add contents in /sys/parameters for built-in modules + * param_sysfs_builtin - add sysfs parameters for built-in modules * * Add module_parameters to sysfs for "modules" built into the kernel. * -- cgit v1.2.3 From 54041d8a73337411b485ff76957fb106cb5d40d0 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Tue, 2 Jul 2013 15:35:12 +0930 Subject: modules: don't fail to load on unknown parameters. Although parameters are supposed to be part of the kernel API, experimental parameters are often removed. In addition, downgrading a kernel might cause previously-working modules to fail to load. On balance, it's probably better to warn, and load the module anyway. This may let through a typo, but at least the logs will show it. Reported-by: Andy Lutomirski Signed-off-by: Rusty Russell --- kernel/module.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/module.c b/kernel/module.c index a1951aba7a03..5184877ce98a 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -3212,6 +3212,17 @@ out: return err; } +static int unknown_module_param_cb(char *param, char *val, const char *modname) +{ + /* Check for magic 'dyndbg' arg */ + int ret = ddebug_dyndbg_module_param_cb(param, val, modname); + if (ret != 0) { + printk(KERN_WARNING "%s: unknown parameter '%s' ignored\n", + modname, param); + } + return 0; +} + /* Allocate and load the module: note that size of section 0 is always zero, and we rely on this for optional sections. */ static int load_module(struct load_info *info, const char __user *uargs, @@ -3298,7 +3309,7 @@ static int load_module(struct load_info *info, const char __user *uargs, /* Module is ready to execute: parsing args may do that. */ err = parse_args(mod->name, mod->args, mod->kp, mod->num_kp, - -32768, 32767, &ddebug_dyndbg_module_param_cb); + -32768, 32767, unknown_module_param_cb); if (err < 0) goto bug_cleanup; -- cgit v1.2.3 From c9b5a266b103af873abb9ac03bc3d067702c8f4b Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 26 Jun 2013 12:17:32 +0200 Subject: tick: Make oneshot broadcast robust vs. CPU offlining In periodic mode we remove offline cpus from the broadcast propagation mask. In oneshot mode we fail to do so. This was not a problem so far, but the recent changes to the broadcast propagation introduced a constellation which can result in a NULL pointer dereference. What happens is: CPU0 CPU1 idle() arch_idle() tick_broadcast_oneshot_control(OFF); set cpu1 in tick_broadcast_force_mask if (cpu_offline()) arch_cpu_dead() cpu_dead_cleanup(cpu1) cpu1 tickdevice pointer = NULL broadcast interrupt dereference cpu1 tickdevice pointer -> OOPS We dereference the pointer because cpu1 is still set in tick_broadcast_force_mask and tick_do_broadcast() expects a valid cpumask and therefor lacks any further checks. Remove the cpu from the tick_broadcast_force_mask before we set the tick device pointer to NULL. Also add a sanity check to the oneshot broadcast function, so we can detect such issues w/o crashing the machine. Reported-by: Prarit Bhargava Cc: athorlton@sgi.com Cc: CAI Qian Link: http://lkml.kernel.org/r/alpine.DEB.2.02.1306261303260.4013@ionos.tec.linutronix.de Signed-off-by: Thomas Gleixner --- kernel/time/tick-broadcast.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-broadcast.c b/kernel/time/tick-broadcast.c index d067c01586f5..4790037163f6 100644 --- a/kernel/time/tick-broadcast.c +++ b/kernel/time/tick-broadcast.c @@ -532,6 +532,13 @@ again: cpumask_or(tmpmask, tmpmask, tick_broadcast_force_mask); cpumask_clear(tick_broadcast_force_mask); + /* + * Sanity check. Catch the case where we try to broadcast to + * offline cpus. + */ + if (WARN_ON_ONCE(!cpumask_subset(tmpmask, cpu_online_mask))) + cpumask_and(tmpmask, tmpmask, cpu_online_mask); + /* * Wakeup the cpus which have an expired event. */ @@ -773,10 +780,12 @@ void tick_shutdown_broadcast_oneshot(unsigned int *cpup) raw_spin_lock_irqsave(&tick_broadcast_lock, flags); /* - * Clear the broadcast mask flag for the dead cpu, but do not - * stop the broadcast device! + * Clear the broadcast masks for the dead cpu, but do not stop + * the broadcast device! */ cpumask_clear_cpu(cpu, tick_broadcast_oneshot_mask); + cpumask_clear_cpu(cpu, tick_broadcast_pending_mask); + cpumask_clear_cpu(cpu, tick_broadcast_force_mask); raw_spin_unlock_irqrestore(&tick_broadcast_lock, flags); } -- cgit v1.2.3 From 1f73a9806bdd07a5106409bbcab3884078bd34fe Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 1 Jul 2013 22:14:10 +0200 Subject: tick: Prevent uncontrolled switch to oneshot mode When the system switches from periodic to oneshot mode, the broadcast logic causes a possibility that a CPU which has not yet switched to oneshot mode puts its own clock event device into oneshot mode without updating the state and the timer handler. CPU0 CPU1 per cpu tickdev is in periodic mode and switched to broadcast Switch to oneshot mode tick_broadcast_switch_to_oneshot() cpumask_copy(tick_oneshot_broacast_mask, tick_broadcast_mask); broadcast device mode = oneshot Timer interrupt irq_enter() tick_check_oneshot_broadcast() dev->set_mode(ONESHOT); tick_handle_periodic() if (dev->mode == ONESHOT) dev->next_event += period; FAIL. We fail, because dev->next_event contains KTIME_MAX, if the device was in periodic mode before the uncontrolled switch to oneshot happened. We must copy the broadcast bits over to the oneshot mask, because otherwise a CPU which relies on the broadcast would not been woken up anymore after the broadcast device switched to oneshot mode. So we need to verify in tick_check_oneshot_broadcast() whether the CPU has already switched to oneshot mode. If not, leave the device untouched and let the CPU switch controlled into oneshot mode. This is a long standing bug, which was never noticed, because the main user of the broadcast x86 cannot run into that scenario, AFAICT. The nonarchitected timer mess of ARM creates a gazillion of differently broken abominations which trigger the shortcomings of that broadcast code, which better had never been necessary in the first place. Reported-and-tested-by: Stehle Vincent-B46079 Reviewed-by: Stephen Boyd Cc: John Stultz , Cc: Mark Rutland Link: http://lkml.kernel.org/r/alpine.DEB.2.02.1307012153060.4013@ionos.tec.linutronix.de Cc: stable@vger.kernel.org Signed-off-by: Thomas Gleixner --- kernel/time/tick-broadcast.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/time/tick-broadcast.c b/kernel/time/tick-broadcast.c index 4790037163f6..248f80dba746 100644 --- a/kernel/time/tick-broadcast.c +++ b/kernel/time/tick-broadcast.c @@ -492,7 +492,15 @@ void tick_check_oneshot_broadcast(int cpu) if (cpumask_test_cpu(cpu, tick_broadcast_oneshot_mask)) { struct tick_device *td = &per_cpu(tick_cpu_device, cpu); - clockevents_set_mode(td->evtdev, CLOCK_EVT_MODE_ONESHOT); + /* + * We might be in the middle of switching over from + * periodic to oneshot. If the CPU has not yet + * switched over, leave the device alone. + */ + if (td->mode == TICKDEV_MODE_ONESHOT) { + clockevents_set_mode(td->evtdev, + CLOCK_EVT_MODE_ONESHOT); + } } } -- cgit v1.2.3 From 07bd1172902e782f288e4d44b1fde7dec0f08b6f Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Mon, 1 Jul 2013 22:14:10 +0200 Subject: tick: Sanitize broadcast control logic The recent implementation of a generic dummy timer resulted in a different registration order of per cpu local timers which made the broadcast control logic go belly up. If the dummy timer is the first clock event device which is registered for a CPU, then it is installed, the broadcast timer is initialized and the CPU is marked as broadcast target. If a real clock event device is installed after that, we can fail to take the CPU out of the broadcast mask. In the worst case we end up with two periodic timer events firing for the same CPU. One from the per cpu hardware device and one from the broadcast. Now the problem is that we have no way to distinguish whether the system is in a state which makes broadcasting necessary or the broadcast bit was set due to the nonfunctional dummy timer installment. To solve this we need to keep track of the system state seperately and provide a more detailed decision logic whether we keep the CPU in broadcast mode or not. The old decision logic only clears the broadcast mode, if the newly installed clock event device is not affected by power states. The new logic clears the broadcast mode if one of the following is true: - The new device is not affected by power states. - The system is not in a power state affected mode - The system has switched to oneshot mode. The oneshot broadcast is controlled from the deep idle state. The CPU is not in idle at this point, so it's safe to remove it from the mask. If we clear the broadcast bit for the CPU when a new device is installed, we also shutdown the broadcast device when this was the last CPU in the broadcast mask. If the broadcast bit is kept, then we leave the new device in shutdown state and rely on the broadcast to deliver the timer interrupts via the broadcast ipis. Reported-and-tested-by: Stehle Vincent-B46079 Reviewed-by: Stephen Boyd Cc: John Stultz , Cc: Mark Rutland Link: http://lkml.kernel.org/r/alpine.DEB.2.02.1307012153060.4013@ionos.tec.linutronix.de Cc: stable@vger.kernel.org Signed-off-by: Thomas Gleixner --- kernel/time/tick-broadcast.c | 70 +++++++++++++++++++++++++++++++++++++------- kernel/time/tick-common.c | 3 +- 2 files changed, 61 insertions(+), 12 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-broadcast.c b/kernel/time/tick-broadcast.c index 248f80dba746..4430fa695b48 100644 --- a/kernel/time/tick-broadcast.c +++ b/kernel/time/tick-broadcast.c @@ -30,6 +30,7 @@ static struct tick_device tick_broadcast_device; static cpumask_var_t tick_broadcast_mask; +static cpumask_var_t tick_broadcast_on; static cpumask_var_t tmpmask; static DEFINE_RAW_SPINLOCK(tick_broadcast_lock); static int tick_broadcast_force; @@ -140,8 +141,9 @@ static void tick_device_setup_broadcast_func(struct clock_event_device *dev) */ int tick_device_uses_broadcast(struct clock_event_device *dev, int cpu) { + struct clock_event_device *bc = tick_broadcast_device.evtdev; unsigned long flags; - int ret = 0; + int ret; raw_spin_lock_irqsave(&tick_broadcast_lock, flags); @@ -155,20 +157,59 @@ int tick_device_uses_broadcast(struct clock_event_device *dev, int cpu) dev->event_handler = tick_handle_periodic; tick_device_setup_broadcast_func(dev); cpumask_set_cpu(cpu, tick_broadcast_mask); - tick_broadcast_start_periodic(tick_broadcast_device.evtdev); + tick_broadcast_start_periodic(bc); ret = 1; } else { /* - * When the new device is not affected by the stop - * feature and the cpu is marked in the broadcast mask - * then clear the broadcast bit. + * Clear the broadcast bit for this cpu if the + * device is not power state affected. */ - if (!(dev->features & CLOCK_EVT_FEAT_C3STOP)) { - int cpu = smp_processor_id(); + if (!(dev->features & CLOCK_EVT_FEAT_C3STOP)) cpumask_clear_cpu(cpu, tick_broadcast_mask); - tick_broadcast_clear_oneshot(cpu); - } else { + else tick_device_setup_broadcast_func(dev); + + /* + * Clear the broadcast bit if the CPU is not in + * periodic broadcast on state. + */ + if (!cpumask_test_cpu(cpu, tick_broadcast_on)) + cpumask_clear_cpu(cpu, tick_broadcast_mask); + + switch (tick_broadcast_device.mode) { + case TICKDEV_MODE_ONESHOT: + /* + * If the system is in oneshot mode we can + * unconditionally clear the oneshot mask bit, + * because the CPU is running and therefore + * not in an idle state which causes the power + * state affected device to stop. Let the + * caller initialize the device. + */ + tick_broadcast_clear_oneshot(cpu); + ret = 0; + break; + + case TICKDEV_MODE_PERIODIC: + /* + * If the system is in periodic mode, check + * whether the broadcast device can be + * switched off now. + */ + if (cpumask_empty(tick_broadcast_mask) && bc) + clockevents_shutdown(bc); + /* + * If we kept the cpu in the broadcast mask, + * tell the caller to leave the per cpu device + * in shutdown state. The periodic interrupt + * is delivered by the broadcast device. + */ + ret = cpumask_test_cpu(cpu, tick_broadcast_mask); + break; + default: + /* Nothing to do */ + ret = 0; + break; } } raw_spin_unlock_irqrestore(&tick_broadcast_lock, flags); @@ -298,6 +339,7 @@ static void tick_do_broadcast_on_off(unsigned long *reason) switch (*reason) { case CLOCK_EVT_NOTIFY_BROADCAST_ON: case CLOCK_EVT_NOTIFY_BROADCAST_FORCE: + cpumask_set_cpu(cpu, tick_broadcast_on); if (!cpumask_test_and_set_cpu(cpu, tick_broadcast_mask)) { if (tick_broadcast_device.mode == TICKDEV_MODE_PERIODIC) @@ -307,8 +349,12 @@ static void tick_do_broadcast_on_off(unsigned long *reason) tick_broadcast_force = 1; break; case CLOCK_EVT_NOTIFY_BROADCAST_OFF: - if (!tick_broadcast_force && - cpumask_test_and_clear_cpu(cpu, tick_broadcast_mask)) { + if (tick_broadcast_force) + break; + cpumask_clear_cpu(cpu, tick_broadcast_on); + if (!tick_device_is_functional(dev)) + break; + if (cpumask_test_and_clear_cpu(cpu, tick_broadcast_mask)) { if (tick_broadcast_device.mode == TICKDEV_MODE_PERIODIC) tick_setup_periodic(dev, 0); @@ -366,6 +412,7 @@ void tick_shutdown_broadcast(unsigned int *cpup) bc = tick_broadcast_device.evtdev; cpumask_clear_cpu(cpu, tick_broadcast_mask); + cpumask_clear_cpu(cpu, tick_broadcast_on); if (tick_broadcast_device.mode == TICKDEV_MODE_PERIODIC) { if (bc && cpumask_empty(tick_broadcast_mask)) @@ -821,6 +868,7 @@ bool tick_broadcast_oneshot_available(void) void __init tick_broadcast_init(void) { zalloc_cpumask_var(&tick_broadcast_mask, GFP_NOWAIT); + zalloc_cpumask_var(&tick_broadcast_on, GFP_NOWAIT); zalloc_cpumask_var(&tmpmask, GFP_NOWAIT); #ifdef CONFIG_TICK_ONESHOT zalloc_cpumask_var(&tick_broadcast_oneshot_mask, GFP_NOWAIT); diff --git a/kernel/time/tick-common.c b/kernel/time/tick-common.c index edd45f64162f..64522ecdfe0e 100644 --- a/kernel/time/tick-common.c +++ b/kernel/time/tick-common.c @@ -194,7 +194,8 @@ static void tick_setup_device(struct tick_device *td, * When global broadcasting is active, check if the current * device is registered as a placeholder for broadcast mode. * This allows us to handle this x86 misfeature in a generic - * way. + * way. This function also returns !=0 when we keep the + * current active broadcast state for this CPU. */ if (tick_device_uses_broadcast(newdev, cpu)) return; -- cgit v1.2.3 From ff451961a8b2a17667a7bfa39c86fb9b351445db Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Mon, 1 Jul 2013 22:50:29 -0400 Subject: tracing: Add trace_array_get/put() to handle instance refs better Commit a695cb58162 "tracing: Prevent deleting instances when they are being read" tried to fix a race between deleting a trace instance and reading contents of a trace file. But it wasn't good enough. The following could crash the kernel: # cd /sys/kernel/debug/tracing/instances # ( while :; do mkdir foo; rmdir foo; done ) & # ( while :; do cat foo/trace &> /dev/null; done ) & Luckily this can only be done by root user, but it should be fixed regardless. The problem is that a delete of the file can happen after the reader starts to open the file but before it grabs the trace_types_mutex. The solution is to validate the trace array before using it. If the trace array does not exist in the list of trace arrays, then it returns -ENODEV. There's a possibility that a trace_array could be deleted and a new one created and the open would open its file instead. But that is very minor as it will just return the data of the new trace array, it may confuse the user but it will not crash the system. As this can only be done by root anyway, the race will only occur if root is deleting what its trying to read at the same time. Cc: stable@vger.kernel.org # 3.10 Reported-by: Alexander Lam Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 83 ++++++++++++++++++++++++++++++++++++++++------------ 1 file changed, 65 insertions(+), 18 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index e36da7ff59bf..6be9df1aa513 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -204,6 +204,37 @@ static struct trace_array global_trace; LIST_HEAD(ftrace_trace_arrays); +int trace_array_get(struct trace_array *this_tr) +{ + struct trace_array *tr; + int ret = -ENODEV; + + mutex_lock(&trace_types_lock); + list_for_each_entry(tr, &ftrace_trace_arrays, list) { + if (tr == this_tr) { + tr->ref++; + ret = 0; + break; + } + } + mutex_unlock(&trace_types_lock); + + return ret; +} + +static void __trace_array_put(struct trace_array *this_tr) +{ + WARN_ON(!this_tr->ref); + this_tr->ref--; +} + +void trace_array_put(struct trace_array *this_tr) +{ + mutex_lock(&trace_types_lock); + __trace_array_put(this_tr); + mutex_unlock(&trace_types_lock); +} + int filter_current_check_discard(struct ring_buffer *buffer, struct ftrace_event_call *call, void *rec, struct ring_buffer_event *event) @@ -2831,10 +2862,9 @@ static const struct seq_operations tracer_seq_ops = { }; static struct trace_iterator * -__tracing_open(struct inode *inode, struct file *file, bool snapshot) +__tracing_open(struct trace_array *tr, struct trace_cpu *tc, + struct inode *inode, struct file *file, bool snapshot) { - struct trace_cpu *tc = inode->i_private; - struct trace_array *tr = tc->tr; struct trace_iterator *iter; int cpu; @@ -2913,8 +2943,6 @@ __tracing_open(struct inode *inode, struct file *file, bool snapshot) tracing_iter_reset(iter, cpu); } - tr->ref++; - mutex_unlock(&trace_types_lock); return iter; @@ -2944,17 +2972,20 @@ static int tracing_release(struct inode *inode, struct file *file) struct trace_array *tr; int cpu; - if (!(file->f_mode & FMODE_READ)) + /* Writes do not use seq_file, need to grab tr from inode */ + if (!(file->f_mode & FMODE_READ)) { + struct trace_cpu *tc = inode->i_private; + + trace_array_put(tc->tr); return 0; + } iter = m->private; tr = iter->tr; + trace_array_put(tr); mutex_lock(&trace_types_lock); - WARN_ON(!tr->ref); - tr->ref--; - for_each_tracing_cpu(cpu) { if (iter->buffer_iter[cpu]) ring_buffer_read_finish(iter->buffer_iter[cpu]); @@ -2973,20 +3004,23 @@ static int tracing_release(struct inode *inode, struct file *file) kfree(iter->trace); kfree(iter->buffer_iter); seq_release_private(inode, file); + return 0; } static int tracing_open(struct inode *inode, struct file *file) { + struct trace_cpu *tc = inode->i_private; + struct trace_array *tr = tc->tr; struct trace_iterator *iter; int ret = 0; + if (trace_array_get(tr) < 0) + return -ENODEV; + /* If this file was open for write, then erase contents */ if ((file->f_mode & FMODE_WRITE) && (file->f_flags & O_TRUNC)) { - struct trace_cpu *tc = inode->i_private; - struct trace_array *tr = tc->tr; - if (tc->cpu == RING_BUFFER_ALL_CPUS) tracing_reset_online_cpus(&tr->trace_buffer); else @@ -2994,12 +3028,16 @@ static int tracing_open(struct inode *inode, struct file *file) } if (file->f_mode & FMODE_READ) { - iter = __tracing_open(inode, file, false); + iter = __tracing_open(tr, tc, inode, file, false); if (IS_ERR(iter)) ret = PTR_ERR(iter); else if (trace_flags & TRACE_ITER_LATENCY_FMT) iter->iter_flags |= TRACE_FILE_LAT_FMT; } + + if (ret < 0) + trace_array_put(tr); + return ret; } @@ -4575,12 +4613,16 @@ struct ftrace_buffer_info { static int tracing_snapshot_open(struct inode *inode, struct file *file) { struct trace_cpu *tc = inode->i_private; + struct trace_array *tr = tc->tr; struct trace_iterator *iter; struct seq_file *m; int ret = 0; + if (trace_array_get(tr) < 0) + return -ENODEV; + if (file->f_mode & FMODE_READ) { - iter = __tracing_open(inode, file, true); + iter = __tracing_open(tr, tc, inode, file, true); if (IS_ERR(iter)) ret = PTR_ERR(iter); } else { @@ -4593,13 +4635,16 @@ static int tracing_snapshot_open(struct inode *inode, struct file *file) kfree(m); return -ENOMEM; } - iter->tr = tc->tr; + iter->tr = tr; iter->trace_buffer = &tc->tr->max_buffer; iter->cpu_file = tc->cpu; m->private = iter; file->private_data = m; } + if (ret < 0) + trace_array_put(tr); + return ret; } @@ -4680,9 +4725,12 @@ out: static int tracing_snapshot_release(struct inode *inode, struct file *file) { struct seq_file *m = file->private_data; + int ret; + + ret = tracing_release(inode, file); if (file->f_mode & FMODE_READ) - return tracing_release(inode, file); + return ret; /* If write only, the seq_file is just a stub */ if (m) @@ -4927,8 +4975,7 @@ static int tracing_buffers_release(struct inode *inode, struct file *file) mutex_lock(&trace_types_lock); - WARN_ON(!iter->tr->ref); - iter->tr->ref--; + __trace_array_put(iter->tr); if (info->spare) ring_buffer_free_read_page(iter->trace_buffer->buffer, info->spare); -- cgit v1.2.3 From 7b85af63034818e43aee6c1d7bf1c7c6796a9073 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Mon, 1 Jul 2013 23:34:22 -0400 Subject: tracing: Get trace_array ref counts when accessing trace files When a trace file is opened that may access a trace array, it must increment its ref count to prevent it from being deleted. Cc: stable@vger.kernel.org # 3.10 Reported-by: Alexander Lam Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 121 +++++++++++++++++++++++++++++++++++++++++++++++---- 1 file changed, 112 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 6be9df1aa513..6d9bd9b43e43 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2965,6 +2965,43 @@ int tracing_open_generic(struct inode *inode, struct file *filp) return 0; } +/* + * Open and update trace_array ref count. + * Must have the current trace_array passed to it. + */ +int tracing_open_generic_tr(struct inode *inode, struct file *filp) +{ + struct trace_array *tr = inode->i_private; + + if (tracing_disabled) + return -ENODEV; + + if (trace_array_get(tr) < 0) + return -ENODEV; + + filp->private_data = inode->i_private; + + return 0; + +} + +int tracing_open_generic_tc(struct inode *inode, struct file *filp) +{ + struct trace_cpu *tc = inode->i_private; + struct trace_array *tr = tc->tr; + + if (tracing_disabled) + return -ENODEV; + + if (trace_array_get(tr) < 0) + return -ENODEV; + + filp->private_data = inode->i_private; + + return 0; + +} + static int tracing_release(struct inode *inode, struct file *file) { struct seq_file *m = file->private_data; @@ -3008,6 +3045,32 @@ static int tracing_release(struct inode *inode, struct file *file) return 0; } +static int tracing_release_generic_tr(struct inode *inode, struct file *file) +{ + struct trace_array *tr = inode->i_private; + + trace_array_put(tr); + return 0; +} + +static int tracing_release_generic_tc(struct inode *inode, struct file *file) +{ + struct trace_cpu *tc = inode->i_private; + struct trace_array *tr = tc->tr; + + trace_array_put(tr); + return 0; +} + +static int tracing_single_release_tr(struct inode *inode, struct file *file) +{ + struct trace_array *tr = inode->i_private; + + trace_array_put(tr); + + return single_release(inode, file); +} + static int tracing_open(struct inode *inode, struct file *file) { struct trace_cpu *tc = inode->i_private; @@ -3394,9 +3457,14 @@ tracing_trace_options_write(struct file *filp, const char __user *ubuf, static int tracing_trace_options_open(struct inode *inode, struct file *file) { + struct trace_array *tr = inode->i_private; + if (tracing_disabled) return -ENODEV; + if (trace_array_get(tr) < 0) + return -ENODEV; + return single_open(file, tracing_trace_options_show, inode->i_private); } @@ -3404,7 +3472,7 @@ static const struct file_operations tracing_iter_fops = { .open = tracing_trace_options_open, .read = seq_read, .llseek = seq_lseek, - .release = single_release, + .release = tracing_single_release_tr, .write = tracing_trace_options_write, }; @@ -3892,6 +3960,9 @@ static int tracing_open_pipe(struct inode *inode, struct file *filp) if (tracing_disabled) return -ENODEV; + if (trace_array_get(tr) < 0) + return -ENODEV; + mutex_lock(&trace_types_lock); /* create a buffer to store the information to pass to userspace */ @@ -3944,6 +4015,7 @@ out: fail: kfree(iter->trace); kfree(iter); + __trace_array_put(tr); mutex_unlock(&trace_types_lock); return ret; } @@ -3951,6 +4023,8 @@ fail: static int tracing_release_pipe(struct inode *inode, struct file *file) { struct trace_iterator *iter = file->private_data; + struct trace_cpu *tc = inode->i_private; + struct trace_array *tr = tc->tr; mutex_lock(&trace_types_lock); @@ -3964,6 +4038,8 @@ static int tracing_release_pipe(struct inode *inode, struct file *file) kfree(iter->trace); kfree(iter); + trace_array_put(tr); + return 0; } @@ -4421,6 +4497,8 @@ tracing_free_buffer_release(struct inode *inode, struct file *filp) /* resize the ring buffer to 0 */ tracing_resize_ring_buffer(tr, 0, RING_BUFFER_ALL_CPUS); + trace_array_put(tr); + return 0; } @@ -4597,10 +4675,20 @@ static ssize_t tracing_clock_write(struct file *filp, const char __user *ubuf, static int tracing_clock_open(struct inode *inode, struct file *file) { + struct trace_array *tr = inode->i_private; + int ret; + if (tracing_disabled) return -ENODEV; - return single_open(file, tracing_clock_show, inode->i_private); + if (trace_array_get(tr)) + return -ENODEV; + + ret = single_open(file, tracing_clock_show, inode->i_private); + if (ret < 0) + trace_array_put(tr); + + return ret; } struct ftrace_buffer_info { @@ -4796,34 +4884,38 @@ static const struct file_operations tracing_pipe_fops = { }; static const struct file_operations tracing_entries_fops = { - .open = tracing_open_generic, + .open = tracing_open_generic_tc, .read = tracing_entries_read, .write = tracing_entries_write, .llseek = generic_file_llseek, + .release = tracing_release_generic_tc, }; static const struct file_operations tracing_total_entries_fops = { - .open = tracing_open_generic, + .open = tracing_open_generic_tr, .read = tracing_total_entries_read, .llseek = generic_file_llseek, + .release = tracing_release_generic_tr, }; static const struct file_operations tracing_free_buffer_fops = { + .open = tracing_open_generic_tr, .write = tracing_free_buffer_write, .release = tracing_free_buffer_release, }; static const struct file_operations tracing_mark_fops = { - .open = tracing_open_generic, + .open = tracing_open_generic_tr, .write = tracing_mark_write, .llseek = generic_file_llseek, + .release = tracing_release_generic_tr, }; static const struct file_operations trace_clock_fops = { .open = tracing_clock_open, .read = seq_read, .llseek = seq_lseek, - .release = single_release, + .release = tracing_single_release_tr, .write = tracing_clock_write, }; @@ -4851,13 +4943,19 @@ static int tracing_buffers_open(struct inode *inode, struct file *filp) struct trace_cpu *tc = inode->i_private; struct trace_array *tr = tc->tr; struct ftrace_buffer_info *info; + int ret; if (tracing_disabled) return -ENODEV; + if (trace_array_get(tr) < 0) + return -ENODEV; + info = kzalloc(sizeof(*info), GFP_KERNEL); - if (!info) + if (!info) { + trace_array_put(tr); return -ENOMEM; + } mutex_lock(&trace_types_lock); @@ -4875,7 +4973,11 @@ static int tracing_buffers_open(struct inode *inode, struct file *filp) mutex_unlock(&trace_types_lock); - return nonseekable_open(inode, filp); + ret = nonseekable_open(inode, filp); + if (ret < 0) + trace_array_put(tr); + + return ret; } static unsigned int @@ -5765,9 +5867,10 @@ rb_simple_write(struct file *filp, const char __user *ubuf, } static const struct file_operations rb_simple_fops = { - .open = tracing_open_generic, + .open = tracing_open_generic_tr, .read = rb_simple_read, .write = rb_simple_write, + .release = tracing_release_generic_tr, .llseek = default_llseek, }; -- cgit v1.2.3 From 8e2e2fa47129532a30cff6c25a47078dc97d9260 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Tue, 2 Jul 2013 15:30:53 -0400 Subject: tracing: Add trace_array_get/put() to event handling Commit a695cb58162 "tracing: Prevent deleting instances when they are being read" tried to fix a race between deleting a trace instance and reading contents of a trace file. But it wasn't good enough. The following could crash the kernel: # cd /sys/kernel/debug/tracing/instances # ( while :; do mkdir foo; rmdir foo; done ) & # ( while :; do echo 1 > foo/events/sched/sched_switch 2> /dev/null; done ) & Luckily this can only be done by root user, but it should be fixed regardless. The problem is that a delete of the file can happen after the write to the event is opened, but before the enabling happens. The solution is to make sure the trace_array is available before succeeding in opening for write, and incerment the ref counter while opened. Now the instance can be deleted when the events are writing to the buffer, but the deletion of the instance will disable all events before the instance is actually deleted. Cc: stable@vger.kernel.org # 3.10 Reported-by: Alexander Lam Signed-off-by: Steven Rostedt --- kernel/trace/trace.h | 3 +++ kernel/trace/trace_events.c | 55 +++++++++++++++++++++++++++++++++++++++++---- 2 files changed, 54 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 2c3cba59552d..c7fbf93f1b7c 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -226,6 +226,9 @@ extern struct list_head ftrace_trace_arrays; extern struct mutex trace_types_lock; +extern int trace_array_get(struct trace_array *tr); +extern void trace_array_put(struct trace_array *tr); + /* * The global tracer (top) should be the first trace array added, * but we check the flag anyway. diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 35c6f23c71b2..920e08fb53b3 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -409,6 +409,35 @@ static void put_system(struct ftrace_subsystem_dir *dir) mutex_unlock(&event_mutex); } +/* + * Open and update trace_array ref count. + * Must have the current trace_array passed to it. + */ +static int tracing_open_generic_file(struct inode *inode, struct file *filp) +{ + struct ftrace_event_file *file = inode->i_private; + struct trace_array *tr = file->tr; + int ret; + + if (trace_array_get(tr) < 0) + return -ENODEV; + + ret = tracing_open_generic(inode, filp); + if (ret < 0) + trace_array_put(tr); + return ret; +} + +static int tracing_release_generic_file(struct inode *inode, struct file *filp) +{ + struct ftrace_event_file *file = inode->i_private; + struct trace_array *tr = file->tr; + + trace_array_put(tr); + + return 0; +} + /* * __ftrace_set_clr_event(NULL, NULL, NULL, set) will set/unset all events. */ @@ -1032,9 +1061,17 @@ static int subsystem_open(struct inode *inode, struct file *filp) /* Some versions of gcc think dir can be uninitialized here */ WARN_ON(!dir); + /* Still need to increment the ref count of the system */ + if (trace_array_get(tr) < 0) { + put_system(dir); + return -ENODEV; + } + ret = tracing_open_generic(inode, filp); - if (ret < 0) + if (ret < 0) { + trace_array_put(tr); put_system(dir); + } return ret; } @@ -1045,16 +1082,23 @@ static int system_tr_open(struct inode *inode, struct file *filp) struct trace_array *tr = inode->i_private; int ret; + if (trace_array_get(tr) < 0) + return -ENODEV; + /* Make a temporary dir that has no system but points to tr */ dir = kzalloc(sizeof(*dir), GFP_KERNEL); - if (!dir) + if (!dir) { + trace_array_put(tr); return -ENOMEM; + } dir->tr = tr; ret = tracing_open_generic(inode, filp); - if (ret < 0) + if (ret < 0) { + trace_array_put(tr); kfree(dir); + } filp->private_data = dir; @@ -1065,6 +1109,8 @@ static int subsystem_release(struct inode *inode, struct file *file) { struct ftrace_subsystem_dir *dir = file->private_data; + trace_array_put(dir->tr); + /* * If dir->subsystem is NULL, then this is a temporary * descriptor that was made for a trace_array to enable @@ -1192,9 +1238,10 @@ static const struct file_operations ftrace_set_event_fops = { }; static const struct file_operations ftrace_enable_fops = { - .open = tracing_open_generic, + .open = tracing_open_generic_file, .read = event_enable_read, .write = event_enable_write, + .release = tracing_release_generic_file, .llseek = default_llseek, }; -- cgit v1.2.3 From 2a6c24afab70dbcfee49f4c76e1511eec1a3298b Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Tue, 2 Jul 2013 14:48:23 -0400 Subject: tracing: Fix race between deleting buffer and setting events While analyzing the code, I discovered that there's a potential race between deleting a trace instance and setting events. There are a few races that can occur if events are being traced as the buffer is being deleted. Mostly the problem comes with freeing the descriptor used by the trace event callback. To prevent problems like this, the events are disabled before the buffer is deleted. The problem with the current solution is that the event_mutex is let go between disabling the events and freeing the files, which means that the events could be enabled again while the freeing takes place. Cc: stable@vger.kernel.org # 3.10 Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 23 +++++++++++++++++------ 1 file changed, 17 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 920e08fb53b3..7d854290bf81 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -441,14 +441,14 @@ static int tracing_release_generic_file(struct inode *inode, struct file *filp) /* * __ftrace_set_clr_event(NULL, NULL, NULL, set) will set/unset all events. */ -static int __ftrace_set_clr_event(struct trace_array *tr, const char *match, - const char *sub, const char *event, int set) +static int +__ftrace_set_clr_event_nolock(struct trace_array *tr, const char *match, + const char *sub, const char *event, int set) { struct ftrace_event_file *file; struct ftrace_event_call *call; int ret = -EINVAL; - mutex_lock(&event_mutex); list_for_each_entry(file, &tr->events, list) { call = file->event_call; @@ -474,6 +474,17 @@ static int __ftrace_set_clr_event(struct trace_array *tr, const char *match, ret = 0; } + + return ret; +} + +static int __ftrace_set_clr_event(struct trace_array *tr, const char *match, + const char *sub, const char *event, int set) +{ + int ret; + + mutex_lock(&event_mutex); + ret = __ftrace_set_clr_event_nolock(tr, match, sub, event, set); mutex_unlock(&event_mutex); return ret; @@ -2408,11 +2419,11 @@ early_event_add_tracer(struct dentry *parent, struct trace_array *tr) int event_trace_del_tracer(struct trace_array *tr) { - /* Disable any running events */ - __ftrace_set_clr_event(tr, NULL, NULL, NULL, 0); - mutex_lock(&event_mutex); + /* Disable any running events */ + __ftrace_set_clr_event_nolock(tr, NULL, NULL, NULL, 0); + down_write(&trace_event_sem); __trace_remove_event_dirs(tr); debugfs_remove_recursive(tr->event_dir); -- cgit v1.2.3 From fa44063f9ef163c3a4c8d8c0465bb8a056b42035 Mon Sep 17 00:00:00 2001 From: "zhangwei(Jovi)" Date: Thu, 13 Jun 2013 14:21:51 +0800 Subject: uprobes: Fix return value in error handling path When wrong argument is passed into uprobe_events it does not return an error: [root@jovi tracing]# echo 'p:myprobe /bin/bash' > uprobe_events [root@jovi tracing]# The proper response is: [root@jovi tracing]# echo 'p:myprobe /bin/bash' > uprobe_events -bash: echo: write error: Invalid argument Link: http://lkml.kernel.org/r/51B964FF.5000106@huawei.com Cc: Frederic Weisbecker Cc: Cc: stable@vger.kernel.org # 3.5+ Signed-off-by: zhangwei(Jovi) Signed-off-by: Steven Rostedt --- kernel/trace/trace_uprobe.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/trace_uprobe.c b/kernel/trace/trace_uprobe.c index 32494fb0ee64..d5d0cd368a56 100644 --- a/kernel/trace/trace_uprobe.c +++ b/kernel/trace/trace_uprobe.c @@ -283,8 +283,10 @@ static int create_trace_uprobe(int argc, char **argv) return -EINVAL; } arg = strchr(argv[1], ':'); - if (!arg) + if (!arg) { + ret = -EINVAL; goto fail_address_parse; + } *arg++ = '\0'; filename = argv[1]; -- cgit v1.2.3 From 11034ae9c20f4057a6127fc965906417978e69b2 Mon Sep 17 00:00:00 2001 From: "zhangwei(Jovi)" Date: Wed, 10 Apr 2013 11:26:23 +0800 Subject: tracing: Fix irqs-off tag display in syscall tracing All syscall tracing irqs-off tags are wrong, the syscall enter entry doesn't disable irqs. [root@jovi tracing]#echo "syscalls:sys_enter_open" > set_event [root@jovi tracing]# cat trace # tracer: nop # # entries-in-buffer/entries-written: 13/13 #P:2 # # _-----=> irqs-off # / _----=> need-resched # | / _---=> hardirq/softirq # || / _--=> preempt-depth # ||| / delay # TASK-PID CPU# |||| TIMESTAMP FUNCTION # | | | |||| | | irqbalance-513 [000] d... 56115.496766: sys_open(filename: 804e1a6, flags: 0, mode: 1b6) irqbalance-513 [000] d... 56115.497008: sys_open(filename: 804e1bb, flags: 0, mode: 1b6) sendmail-771 [000] d... 56115.827982: sys_open(filename: b770e6d1, flags: 0, mode: 1b6) The reason is syscall tracing doesn't record irq_flags into buffer. The proper display is: [root@jovi tracing]#echo "syscalls:sys_enter_open" > set_event [root@jovi tracing]# cat trace # tracer: nop # # entries-in-buffer/entries-written: 14/14 #P:2 # # _-----=> irqs-off # / _----=> need-resched # | / _---=> hardirq/softirq # || / _--=> preempt-depth # ||| / delay # TASK-PID CPU# |||| TIMESTAMP FUNCTION # | | | |||| | | irqbalance-514 [001] .... 46.213921: sys_open(filename: 804e1a6, flags: 0, mode: 1b6) irqbalance-514 [001] .... 46.214160: sys_open(filename: 804e1bb, flags: 0, mode: 1b6) <...>-920 [001] .... 47.307260: sys_open(filename: 4e82a0c5, flags: 80000, mode: 0) Link: http://lkml.kernel.org/r/1365564393-10972-3-git-send-email-jovi.zhangwei@huawei.com Cc: stable@vger.kernel.org # 2.6.35 Signed-off-by: zhangwei(Jovi) Signed-off-by: Steven Rostedt --- kernel/trace/trace_syscalls.c | 21 +++++++++++++++++---- 1 file changed, 17 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 8f2ac73c7a5f..322e16461072 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -306,6 +306,8 @@ static void ftrace_syscall_enter(void *data, struct pt_regs *regs, long id) struct syscall_metadata *sys_data; struct ring_buffer_event *event; struct ring_buffer *buffer; + unsigned long irq_flags; + int pc; int syscall_nr; int size; @@ -321,9 +323,12 @@ static void ftrace_syscall_enter(void *data, struct pt_regs *regs, long id) size = sizeof(*entry) + sizeof(unsigned long) * sys_data->nb_args; + local_save_flags(irq_flags); + pc = preempt_count(); + buffer = tr->trace_buffer.buffer; event = trace_buffer_lock_reserve(buffer, - sys_data->enter_event->event.type, size, 0, 0); + sys_data->enter_event->event.type, size, irq_flags, pc); if (!event) return; @@ -333,7 +338,8 @@ static void ftrace_syscall_enter(void *data, struct pt_regs *regs, long id) if (!filter_current_check_discard(buffer, sys_data->enter_event, entry, event)) - trace_current_buffer_unlock_commit(buffer, event, 0, 0); + trace_current_buffer_unlock_commit(buffer, event, + irq_flags, pc); } static void ftrace_syscall_exit(void *data, struct pt_regs *regs, long ret) @@ -343,6 +349,8 @@ static void ftrace_syscall_exit(void *data, struct pt_regs *regs, long ret) struct syscall_metadata *sys_data; struct ring_buffer_event *event; struct ring_buffer *buffer; + unsigned long irq_flags; + int pc; int syscall_nr; syscall_nr = trace_get_syscall_nr(current, regs); @@ -355,9 +363,13 @@ static void ftrace_syscall_exit(void *data, struct pt_regs *regs, long ret) if (!sys_data) return; + local_save_flags(irq_flags); + pc = preempt_count(); + buffer = tr->trace_buffer.buffer; event = trace_buffer_lock_reserve(buffer, - sys_data->exit_event->event.type, sizeof(*entry), 0, 0); + sys_data->exit_event->event.type, sizeof(*entry), + irq_flags, pc); if (!event) return; @@ -367,7 +379,8 @@ static void ftrace_syscall_exit(void *data, struct pt_regs *regs, long ret) if (!filter_current_check_discard(buffer, sys_data->exit_event, entry, event)) - trace_current_buffer_unlock_commit(buffer, event, 0, 0); + trace_current_buffer_unlock_commit(buffer, event, + irq_flags, pc); } static int reg_event_syscall_enter(struct ftrace_event_file *file, -- cgit v1.2.3 From 5280bcef91e706770cc1706eb97353e3513322b9 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Tue, 2 Jul 2013 19:59:57 -0400 Subject: tracing: Make tracer_tracing_{off,on,is_on}() static I have patches that will use tracer_tracing_on/off/is_on() in other files, but as they are not ready to be merged yet, and Fengguang Wu's sparse scripts pointed out that these functions were not declared anywhere, I'll make them static for now. When these functions are required to be used elsewhere, I'll remove the static then. Reported-by: kbuild test robot Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 6d9bd9b43e43..48aceb8a0328 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -387,7 +387,7 @@ unsigned long trace_flags = TRACE_ITER_PRINT_PARENT | TRACE_ITER_PRINTK | TRACE_ITER_GRAPH_TIME | TRACE_ITER_RECORD_CMD | TRACE_ITER_OVERWRITE | TRACE_ITER_IRQ_INFO | TRACE_ITER_MARKERS | TRACE_ITER_FUNCTION; -void tracer_tracing_on(struct trace_array *tr) +static void tracer_tracing_on(struct trace_array *tr) { if (tr->trace_buffer.buffer) ring_buffer_record_on(tr->trace_buffer.buffer); @@ -606,7 +606,7 @@ void tracing_snapshot_alloc(void) EXPORT_SYMBOL_GPL(tracing_snapshot_alloc); #endif /* CONFIG_TRACER_SNAPSHOT */ -void tracer_tracing_off(struct trace_array *tr) +static void tracer_tracing_off(struct trace_array *tr) { if (tr->trace_buffer.buffer) ring_buffer_record_off(tr->trace_buffer.buffer); @@ -649,7 +649,7 @@ void disable_trace_on_warning(void) * * Shows real state of the ring buffer if it is enabled or not. */ -int tracer_tracing_is_on(struct trace_array *tr) +static int tracer_tracing_is_on(struct trace_array *tr) { if (tr->trace_buffer.buffer) return ring_buffer_record_is_on(tr->trace_buffer.buffer); -- cgit v1.2.3 From 4480361c3c592fcbce3ef74e030719f0715e3a7e Mon Sep 17 00:00:00 2001 From: "zhangwei(Jovi)" Date: Wed, 10 Apr 2013 11:26:28 +0800 Subject: tracing: Remove TRACE_EVENT_TYPE enum definition TRACE_EVENT_TYPE enum is not used at present, remove it. Link: http://lkml.kernel.org/r/1365564393-10972-8-git-send-email-jovi.zhangwei@huawei.com Signed-off-by: zhangwei(Jovi) Signed-off-by: Steven Rostedt --- kernel/trace/trace.h | 6 ------ 1 file changed, 6 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index c7fbf93f1b7c..1cbba04976b4 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -907,12 +907,6 @@ static inline void trace_branch_disable(void) /* set ring buffers to default size if not already done so */ int tracing_update_buffers(void); -/* trace event type bit fields, not numeric */ -enum { - TRACE_EVENT_TYPE_PRINTF = 1, - TRACE_EVENT_TYPE_RAW = 2, -}; - struct ftrace_event_field { struct list_head link; const char *name; -- cgit v1.2.3 From 8de1eb02778b64f8b292db531cf39a429f84315f Mon Sep 17 00:00:00 2001 From: "zhangwei(Jovi)" Date: Wed, 10 Apr 2013 11:26:30 +0800 Subject: tracing: Remove ftrace() function The only caller of function ftrace(...) was removed a long time ago, so remove the function body as well. Link: http://lkml.kernel.org/r/1365564393-10972-10-git-send-email-jovi.zhangwei@huawei.com Signed-off-by: zhangwei(Jovi) Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 9 --------- kernel/trace/trace.h | 5 ----- 2 files changed, 14 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 48aceb8a0328..f6fed9e51c64 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1637,15 +1637,6 @@ trace_function(struct trace_array *tr, __buffer_unlock_commit(buffer, event); } -void -ftrace(struct trace_array *tr, struct trace_array_cpu *data, - unsigned long ip, unsigned long parent_ip, unsigned long flags, - int pc) -{ - if (likely(!atomic_read(&data->disabled))) - trace_function(tr, ip, parent_ip, flags, pc); -} - #ifdef CONFIG_STACKTRACE #define FTRACE_STACK_MAX_ENTRIES (PAGE_SIZE / sizeof(unsigned long)) diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 1cbba04976b4..a4ed382dea2f 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -559,11 +559,6 @@ void tracing_iter_reset(struct trace_iterator *iter, int cpu); void poll_wait_pipe(struct trace_iterator *iter); -void ftrace(struct trace_array *tr, - struct trace_array_cpu *data, - unsigned long ip, - unsigned long parent_ip, - unsigned long flags, int pc); void tracing_sched_switch_trace(struct trace_array *tr, struct task_struct *prev, struct task_struct *next, -- cgit v1.2.3 From dcc302232c1f9b3ca16f6b8ee190eb0b1a8a0da3 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Tue, 2 Jul 2013 20:30:52 -0400 Subject: tracing: Make tracing_open_generic_{tr,tc}() static I have patches that will use tracing_open_generic_tr/tc() in other files, but as they are not ready to be merged yet, and Fengguang Wu's sparse scripts pointed out that these functions were not declared anywhere, I'll make them static for now. When these functions are required to be used elsewhere, I'll remove the static then. Reported-by: kbuild test robot Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index f6fed9e51c64..dc473b51415f 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2960,7 +2960,7 @@ int tracing_open_generic(struct inode *inode, struct file *filp) * Open and update trace_array ref count. * Must have the current trace_array passed to it. */ -int tracing_open_generic_tr(struct inode *inode, struct file *filp) +static int tracing_open_generic_tr(struct inode *inode, struct file *filp) { struct trace_array *tr = inode->i_private; @@ -2976,7 +2976,7 @@ int tracing_open_generic_tr(struct inode *inode, struct file *filp) } -int tracing_open_generic_tc(struct inode *inode, struct file *filp) +static int tracing_open_generic_tc(struct inode *inode, struct file *filp) { struct trace_cpu *tc = inode->i_private; struct trace_array *tr = tc->tr; -- cgit v1.2.3 From 8d8022e8aba85192e937f1f0f7450e256d66ae5c Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Wed, 3 Jul 2013 10:06:28 +0930 Subject: module: do percpu allocation after uniqueness check. No, really! v3.8-rc1-5-g1fb9341 was supposed to stop parallel kvm loads exhausting percpu memory on large machines: Now we have a new state MODULE_STATE_UNFORMED, we can insert the module into the list (and thus guarantee its uniqueness) before we allocate the per-cpu region. In my defence, it didn't actually say the patch did this. Just that we "can". This patch actually *does* it. Signed-off-by: Rusty Russell Tested-by: Jim Hull Cc: stable@kernel.org # 3.8 --- kernel/module.c | 34 ++++++++++++++++++---------------- 1 file changed, 18 insertions(+), 16 deletions(-) (limited to 'kernel') diff --git a/kernel/module.c b/kernel/module.c index 5184877ce98a..d1a161be7b04 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -2940,7 +2940,6 @@ static struct module *layout_and_allocate(struct load_info *info, int flags) { /* Module within temporary copy. */ struct module *mod; - Elf_Shdr *pcpusec; int err; mod = setup_load_info(info, flags); @@ -2955,17 +2954,10 @@ static struct module *layout_and_allocate(struct load_info *info, int flags) err = module_frob_arch_sections(info->hdr, info->sechdrs, info->secstrings, mod); if (err < 0) - goto out; + return ERR_PTR(err); - pcpusec = &info->sechdrs[info->index.pcpu]; - if (pcpusec->sh_size) { - /* We have a special allocation for this section. */ - err = percpu_modalloc(mod, - pcpusec->sh_size, pcpusec->sh_addralign); - if (err) - goto out; - pcpusec->sh_flags &= ~(unsigned long)SHF_ALLOC; - } + /* We will do a special allocation for per-cpu sections later. */ + info->sechdrs[info->index.pcpu].sh_flags &= ~(unsigned long)SHF_ALLOC; /* Determine total sizes, and put offsets in sh_entsize. For now this is done generically; there doesn't appear to be any @@ -2976,17 +2968,22 @@ static struct module *layout_and_allocate(struct load_info *info, int flags) /* Allocate and move to the final place */ err = move_module(mod, info); if (err) - goto free_percpu; + return ERR_PTR(err); /* Module has been copied to its final place now: return it. */ mod = (void *)info->sechdrs[info->index.mod].sh_addr; kmemleak_load_module(mod, info); return mod; +} -free_percpu: - percpu_modfree(mod); -out: - return ERR_PTR(err); +static int alloc_module_percpu(struct module *mod, struct load_info *info) +{ + Elf_Shdr *pcpusec = &info->sechdrs[info->index.pcpu]; + if (!pcpusec->sh_size) + return 0; + + /* We have a special allocation for this section. */ + return percpu_modalloc(mod, pcpusec->sh_size, pcpusec->sh_addralign); } /* mod is no longer valid after this! */ @@ -3262,6 +3259,11 @@ static int load_module(struct load_info *info, const char __user *uargs, } #endif + /* To avoid stressing percpu allocator, do this once we're unique. */ + err = alloc_module_percpu(mod, info); + if (err) + goto unlink_mod; + /* Now module is in final location, initialize linked lists, etc. */ err = module_unload_init(mod); if (err) -- cgit v1.2.3 From 9eb76d7797b892a1dad4f2efb6f786681306dd13 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Wed, 3 Jul 2013 10:06:29 +0930 Subject: module: cleanup call chain. Fold alloc_module_percpu into percpu_modalloc(). Signed-off-by: Rusty Russell --- kernel/module.c | 35 ++++++++++++++++------------------- 1 file changed, 16 insertions(+), 19 deletions(-) (limited to 'kernel') diff --git a/kernel/module.c b/kernel/module.c index d1a161be7b04..c18107942ac2 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -482,23 +482,28 @@ static inline void __percpu *mod_percpu(struct module *mod) return mod->percpu; } -static int percpu_modalloc(struct module *mod, - unsigned long size, unsigned long align) +static int percpu_modalloc(struct module *mod, struct load_info *info) { + Elf_Shdr *pcpusec = &info->sechdrs[info->index.pcpu]; + unsigned long align = pcpusec->sh_addralign; + + if (!pcpusec->sh_size) + return 0; + if (align > PAGE_SIZE) { printk(KERN_WARNING "%s: per-cpu alignment %li > %li\n", mod->name, align, PAGE_SIZE); align = PAGE_SIZE; } - mod->percpu = __alloc_reserved_percpu(size, align); + mod->percpu = __alloc_reserved_percpu(pcpusec->sh_size, align); if (!mod->percpu) { printk(KERN_WARNING "%s: Could not allocate %lu bytes percpu data\n", - mod->name, size); + mod->name, (unsigned long)pcpusec->sh_size); return -ENOMEM; } - mod->percpu_size = size; + mod->percpu_size = pcpusec->sh_size; return 0; } @@ -563,10 +568,12 @@ static inline void __percpu *mod_percpu(struct module *mod) { return NULL; } -static inline int percpu_modalloc(struct module *mod, - unsigned long size, unsigned long align) +static int percpu_modalloc(struct module *mod, struct load_info *info) { - return -ENOMEM; + /* UP modules shouldn't have this section: ENOMEM isn't quite right */ + if (info->sechdrs[info->index.pcpu].sh_size != 0) + return -ENOMEM; + return 0; } static inline void percpu_modfree(struct module *mod) { @@ -2976,16 +2983,6 @@ static struct module *layout_and_allocate(struct load_info *info, int flags) return mod; } -static int alloc_module_percpu(struct module *mod, struct load_info *info) -{ - Elf_Shdr *pcpusec = &info->sechdrs[info->index.pcpu]; - if (!pcpusec->sh_size) - return 0; - - /* We have a special allocation for this section. */ - return percpu_modalloc(mod, pcpusec->sh_size, pcpusec->sh_addralign); -} - /* mod is no longer valid after this! */ static void module_deallocate(struct module *mod, struct load_info *info) { @@ -3260,7 +3257,7 @@ static int load_module(struct load_info *info, const char __user *uargs, #endif /* To avoid stressing percpu allocator, do this once we're unique. */ - err = alloc_module_percpu(mod, info); + err = percpu_modalloc(mod, info); if (err) goto unlink_mod; -- cgit v1.2.3 From 55ccb616a6e42052edb37e9c4f82cf8854a59429 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 28 Jun 2013 00:06:42 +0000 Subject: posix_cpu_timer: consolidate expiry time type The posix cpu timer expiry time is stored in a union of two types: a 64 bits field if we rely on scheduler precise accounting, or a cputime_t if we rely on jiffies. This results in quite some duplicate code and special cases to handle the two types. Just unify this into a single 64 bits field. cputime_t can always fit into it. Signed-off-by: Frederic Weisbecker Cc: Stanislaw Gruszka Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Ingo Molnar Cc: Oleg Nesterov Cc: KOSAKI Motohiro Cc: Olivier Langlois Signed-off-by: Andrew Morton --- kernel/posix-cpu-timers.c | 266 ++++++++++++++++++---------------------------- 1 file changed, 106 insertions(+), 160 deletions(-) (limited to 'kernel') diff --git a/kernel/posix-cpu-timers.c b/kernel/posix-cpu-timers.c index 42670e9b44e0..c3c4ea1225a4 100644 --- a/kernel/posix-cpu-timers.c +++ b/kernel/posix-cpu-timers.c @@ -51,59 +51,28 @@ static int check_clock(const clockid_t which_clock) return error; } -static inline union cpu_time_count +static inline unsigned long long timespec_to_sample(const clockid_t which_clock, const struct timespec *tp) { - union cpu_time_count ret; - ret.sched = 0; /* high half always zero when .cpu used */ + unsigned long long ret; + + ret = 0; /* high half always zero when .cpu used */ if (CPUCLOCK_WHICH(which_clock) == CPUCLOCK_SCHED) { - ret.sched = (unsigned long long)tp->tv_sec * NSEC_PER_SEC + tp->tv_nsec; + ret = (unsigned long long)tp->tv_sec * NSEC_PER_SEC + tp->tv_nsec; } else { - ret.cpu = timespec_to_cputime(tp); + ret = cputime_to_expires(timespec_to_cputime(tp)); } return ret; } static void sample_to_timespec(const clockid_t which_clock, - union cpu_time_count cpu, + unsigned long long expires, struct timespec *tp) { if (CPUCLOCK_WHICH(which_clock) == CPUCLOCK_SCHED) - *tp = ns_to_timespec(cpu.sched); + *tp = ns_to_timespec(expires); else - cputime_to_timespec(cpu.cpu, tp); -} - -static inline int cpu_time_before(const clockid_t which_clock, - union cpu_time_count now, - union cpu_time_count then) -{ - if (CPUCLOCK_WHICH(which_clock) == CPUCLOCK_SCHED) { - return now.sched < then.sched; - } else { - return now.cpu < then.cpu; - } -} -static inline void cpu_time_add(const clockid_t which_clock, - union cpu_time_count *acc, - union cpu_time_count val) -{ - if (CPUCLOCK_WHICH(which_clock) == CPUCLOCK_SCHED) { - acc->sched += val.sched; - } else { - acc->cpu += val.cpu; - } -} -static inline union cpu_time_count cpu_time_sub(const clockid_t which_clock, - union cpu_time_count a, - union cpu_time_count b) -{ - if (CPUCLOCK_WHICH(which_clock) == CPUCLOCK_SCHED) { - a.sched -= b.sched; - } else { - a.cpu -= b.cpu; - } - return a; + cputime_to_timespec((__force cputime_t)expires, tp); } /* @@ -111,47 +80,31 @@ static inline union cpu_time_count cpu_time_sub(const clockid_t which_clock, * given the current clock sample. */ static void bump_cpu_timer(struct k_itimer *timer, - union cpu_time_count now) + unsigned long long now) { int i; + unsigned long long delta, incr; - if (timer->it.cpu.incr.sched == 0) + if (timer->it.cpu.incr == 0) return; - if (CPUCLOCK_WHICH(timer->it_clock) == CPUCLOCK_SCHED) { - unsigned long long delta, incr; + if (now < timer->it.cpu.expires) + return; - if (now.sched < timer->it.cpu.expires.sched) - return; - incr = timer->it.cpu.incr.sched; - delta = now.sched + incr - timer->it.cpu.expires.sched; - /* Don't use (incr*2 < delta), incr*2 might overflow. */ - for (i = 0; incr < delta - incr; i++) - incr = incr << 1; - for (; i >= 0; incr >>= 1, i--) { - if (delta < incr) - continue; - timer->it.cpu.expires.sched += incr; - timer->it_overrun += 1 << i; - delta -= incr; - } - } else { - cputime_t delta, incr; + incr = timer->it.cpu.incr; + delta = now + incr - timer->it.cpu.expires; - if (now.cpu < timer->it.cpu.expires.cpu) - return; - incr = timer->it.cpu.incr.cpu; - delta = now.cpu + incr - timer->it.cpu.expires.cpu; - /* Don't use (incr*2 < delta), incr*2 might overflow. */ - for (i = 0; incr < delta - incr; i++) - incr += incr; - for (; i >= 0; incr = incr >> 1, i--) { - if (delta < incr) - continue; - timer->it.cpu.expires.cpu += incr; - timer->it_overrun += 1 << i; - delta -= incr; - } + /* Don't use (incr*2 < delta), incr*2 might overflow. */ + for (i = 0; incr < delta - incr; i++) + incr = incr << 1; + + for (; i >= 0; incr >>= 1, i--) { + if (delta < incr) + continue; + + timer->it.cpu.expires += incr; + timer->it_overrun += 1 << i; + delta -= incr; } } @@ -170,21 +123,21 @@ static inline int task_cputime_zero(const struct task_cputime *cputime) return 0; } -static inline cputime_t prof_ticks(struct task_struct *p) +static inline unsigned long long prof_ticks(struct task_struct *p) { cputime_t utime, stime; task_cputime(p, &utime, &stime); - return utime + stime; + return cputime_to_expires(utime + stime); } -static inline cputime_t virt_ticks(struct task_struct *p) +static inline unsigned long long virt_ticks(struct task_struct *p) { cputime_t utime; task_cputime(p, &utime, NULL); - return utime; + return cputime_to_expires(utime); } static int @@ -225,19 +178,19 @@ posix_cpu_clock_set(const clockid_t which_clock, const struct timespec *tp) * Sample a per-thread clock for the given task. */ static int cpu_clock_sample(const clockid_t which_clock, struct task_struct *p, - union cpu_time_count *cpu) + unsigned long long *sample) { switch (CPUCLOCK_WHICH(which_clock)) { default: return -EINVAL; case CPUCLOCK_PROF: - cpu->cpu = prof_ticks(p); + *sample = prof_ticks(p); break; case CPUCLOCK_VIRT: - cpu->cpu = virt_ticks(p); + *sample = virt_ticks(p); break; case CPUCLOCK_SCHED: - cpu->sched = task_sched_runtime(p); + *sample = task_sched_runtime(p); break; } return 0; @@ -284,7 +237,7 @@ void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times) */ static int cpu_clock_sample_group(const clockid_t which_clock, struct task_struct *p, - union cpu_time_count *cpu) + unsigned long long *sample) { struct task_cputime cputime; @@ -293,15 +246,15 @@ static int cpu_clock_sample_group(const clockid_t which_clock, return -EINVAL; case CPUCLOCK_PROF: thread_group_cputime(p, &cputime); - cpu->cpu = cputime.utime + cputime.stime; + *sample = cputime_to_expires(cputime.utime + cputime.stime); break; case CPUCLOCK_VIRT: thread_group_cputime(p, &cputime); - cpu->cpu = cputime.utime; + *sample = cputime_to_expires(cputime.utime); break; case CPUCLOCK_SCHED: thread_group_cputime(p, &cputime); - cpu->sched = cputime.sum_exec_runtime; + *sample = cputime.sum_exec_runtime; break; } return 0; @@ -312,7 +265,7 @@ static int posix_cpu_clock_get(const clockid_t which_clock, struct timespec *tp) { const pid_t pid = CPUCLOCK_PID(which_clock); int error = -EINVAL; - union cpu_time_count rtn; + unsigned long long rtn; if (pid == 0) { /* @@ -461,30 +414,30 @@ static void cleanup_timers(struct list_head *head, list_for_each_entry_safe(timer, next, head, entry) { list_del_init(&timer->entry); - if (timer->expires.cpu < ptime) { - timer->expires.cpu = 0; + if (timer->expires < cputime_to_expires(ptime)) { + timer->expires = 0; } else { - timer->expires.cpu -= ptime; + timer->expires -= cputime_to_expires(ptime); } } ++head; list_for_each_entry_safe(timer, next, head, entry) { list_del_init(&timer->entry); - if (timer->expires.cpu < utime) { - timer->expires.cpu = 0; + if (timer->expires < cputime_to_expires(utime)) { + timer->expires = 0; } else { - timer->expires.cpu -= utime; + timer->expires -= cputime_to_expires(utime); } } ++head; list_for_each_entry_safe(timer, next, head, entry) { list_del_init(&timer->entry); - if (timer->expires.sched < sum_exec_runtime) { - timer->expires.sched = 0; + if (timer->expires < sum_exec_runtime) { + timer->expires = 0; } else { - timer->expires.sched -= sum_exec_runtime; + timer->expires -= sum_exec_runtime; } } } @@ -516,7 +469,7 @@ void posix_cpu_timers_exit_group(struct task_struct *tsk) tsk->se.sum_exec_runtime + sig->sum_sched_runtime); } -static void clear_dead_task(struct k_itimer *timer, union cpu_time_count now) +static void clear_dead_task(struct k_itimer *timer, unsigned long long now) { /* * That's all for this thread or process. @@ -524,9 +477,7 @@ static void clear_dead_task(struct k_itimer *timer, union cpu_time_count now) */ put_task_struct(timer->it.cpu.task); timer->it.cpu.task = NULL; - timer->it.cpu.expires = cpu_time_sub(timer->it_clock, - timer->it.cpu.expires, - now); + timer->it.cpu.expires -= now; } static inline int expires_gt(cputime_t expires, cputime_t new_exp) @@ -558,14 +509,14 @@ static void arm_timer(struct k_itimer *timer) listpos = head; list_for_each_entry(next, head, entry) { - if (cpu_time_before(timer->it_clock, nt->expires, next->expires)) + if (nt->expires < next->expires) break; listpos = &next->entry; } list_add(&nt->entry, listpos); if (listpos == head) { - union cpu_time_count *exp = &nt->expires; + unsigned long long exp = nt->expires; /* * We are the new earliest-expiring POSIX 1.b timer, hence @@ -576,17 +527,17 @@ static void arm_timer(struct k_itimer *timer) switch (CPUCLOCK_WHICH(timer->it_clock)) { case CPUCLOCK_PROF: - if (expires_gt(cputime_expires->prof_exp, exp->cpu)) - cputime_expires->prof_exp = exp->cpu; + if (expires_gt(cputime_expires->prof_exp, expires_to_cputime(exp))) + cputime_expires->prof_exp = expires_to_cputime(exp); break; case CPUCLOCK_VIRT: - if (expires_gt(cputime_expires->virt_exp, exp->cpu)) - cputime_expires->virt_exp = exp->cpu; + if (expires_gt(cputime_expires->virt_exp, expires_to_cputime(exp))) + cputime_expires->virt_exp = expires_to_cputime(exp); break; case CPUCLOCK_SCHED: if (cputime_expires->sched_exp == 0 || - cputime_expires->sched_exp > exp->sched) - cputime_expires->sched_exp = exp->sched; + cputime_expires->sched_exp > exp) + cputime_expires->sched_exp = exp; break; } } @@ -601,20 +552,20 @@ static void cpu_timer_fire(struct k_itimer *timer) /* * User don't want any signal. */ - timer->it.cpu.expires.sched = 0; + timer->it.cpu.expires = 0; } else if (unlikely(timer->sigq == NULL)) { /* * This a special case for clock_nanosleep, * not a normal timer from sys_timer_create. */ wake_up_process(timer->it_process); - timer->it.cpu.expires.sched = 0; - } else if (timer->it.cpu.incr.sched == 0) { + timer->it.cpu.expires = 0; + } else if (timer->it.cpu.incr == 0) { /* * One-shot timer. Clear it as soon as it's fired. */ posix_timer_event(timer, 0); - timer->it.cpu.expires.sched = 0; + timer->it.cpu.expires = 0; } else if (posix_timer_event(timer, ++timer->it_requeue_pending)) { /* * The signal did not get queued because the signal @@ -632,7 +583,7 @@ static void cpu_timer_fire(struct k_itimer *timer) */ static int cpu_timer_sample_group(const clockid_t which_clock, struct task_struct *p, - union cpu_time_count *cpu) + unsigned long long *sample) { struct task_cputime cputime; @@ -641,13 +592,13 @@ static int cpu_timer_sample_group(const clockid_t which_clock, default: return -EINVAL; case CPUCLOCK_PROF: - cpu->cpu = cputime.utime + cputime.stime; + *sample = cputime_to_expires(cputime.utime + cputime.stime); break; case CPUCLOCK_VIRT: - cpu->cpu = cputime.utime; + *sample = cputime_to_expires(cputime.utime); break; case CPUCLOCK_SCHED: - cpu->sched = cputime.sum_exec_runtime + task_delta_exec(p); + *sample = cputime.sum_exec_runtime + task_delta_exec(p); break; } return 0; @@ -694,7 +645,7 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int flags, struct itimerspec *new, struct itimerspec *old) { struct task_struct *p = timer->it.cpu.task; - union cpu_time_count old_expires, new_expires, old_incr, val; + unsigned long long old_expires, new_expires, old_incr, val; int ret; if (unlikely(p == NULL)) { @@ -749,7 +700,7 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int flags, } if (old) { - if (old_expires.sched == 0) { + if (old_expires == 0) { old->it_value.tv_sec = 0; old->it_value.tv_nsec = 0; } else { @@ -764,11 +715,8 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int flags, * new setting. */ bump_cpu_timer(timer, val); - if (cpu_time_before(timer->it_clock, val, - timer->it.cpu.expires)) { - old_expires = cpu_time_sub( - timer->it_clock, - timer->it.cpu.expires, val); + if (val < timer->it.cpu.expires) { + old_expires = timer->it.cpu.expires - val; sample_to_timespec(timer->it_clock, old_expires, &old->it_value); @@ -791,8 +739,8 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int flags, goto out; } - if (new_expires.sched != 0 && !(flags & TIMER_ABSTIME)) { - cpu_time_add(timer->it_clock, &new_expires, val); + if (new_expires != 0 && !(flags & TIMER_ABSTIME)) { + new_expires += val; } /* @@ -801,8 +749,7 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int flags, * arm the timer (we'll just fake it for timer_gettime). */ timer->it.cpu.expires = new_expires; - if (new_expires.sched != 0 && - cpu_time_before(timer->it_clock, val, new_expires)) { + if (new_expires != 0 && val < new_expires) { arm_timer(timer); } @@ -826,8 +773,7 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int flags, timer->it_overrun_last = 0; timer->it_overrun = -1; - if (new_expires.sched != 0 && - !cpu_time_before(timer->it_clock, val, new_expires)) { + if (new_expires != 0 && !(val < new_expires)) { /* * The designated time already passed, so we notify * immediately, even if the thread never runs to @@ -849,7 +795,7 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int flags, static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec *itp) { - union cpu_time_count now; + unsigned long long now; struct task_struct *p = timer->it.cpu.task; int clear_dead; @@ -859,7 +805,7 @@ static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec *itp) sample_to_timespec(timer->it_clock, timer->it.cpu.incr, &itp->it_interval); - if (timer->it.cpu.expires.sched == 0) { /* Timer not armed at all. */ + if (timer->it.cpu.expires == 0) { /* Timer not armed at all. */ itp->it_value.tv_sec = itp->it_value.tv_nsec = 0; return; } @@ -891,7 +837,7 @@ static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec *itp) */ put_task_struct(p); timer->it.cpu.task = NULL; - timer->it.cpu.expires.sched = 0; + timer->it.cpu.expires = 0; read_unlock(&tasklist_lock); goto dead; } else { @@ -912,10 +858,9 @@ static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec *itp) goto dead; } - if (cpu_time_before(timer->it_clock, now, timer->it.cpu.expires)) { + if (now < timer->it.cpu.expires) { sample_to_timespec(timer->it_clock, - cpu_time_sub(timer->it_clock, - timer->it.cpu.expires, now), + timer->it.cpu.expires - now, &itp->it_value); } else { /* @@ -946,8 +891,8 @@ static void check_thread_timers(struct task_struct *tsk, struct cpu_timer_list *t = list_first_entry(timers, struct cpu_timer_list, entry); - if (!--maxfire || prof_ticks(tsk) < t->expires.cpu) { - tsk->cputime_expires.prof_exp = t->expires.cpu; + if (!--maxfire || prof_ticks(tsk) < t->expires) { + tsk->cputime_expires.prof_exp = expires_to_cputime(t->expires); break; } t->firing = 1; @@ -961,8 +906,8 @@ static void check_thread_timers(struct task_struct *tsk, struct cpu_timer_list *t = list_first_entry(timers, struct cpu_timer_list, entry); - if (!--maxfire || virt_ticks(tsk) < t->expires.cpu) { - tsk->cputime_expires.virt_exp = t->expires.cpu; + if (!--maxfire || virt_ticks(tsk) < t->expires) { + tsk->cputime_expires.virt_exp = expires_to_cputime(t->expires); break; } t->firing = 1; @@ -976,8 +921,8 @@ static void check_thread_timers(struct task_struct *tsk, struct cpu_timer_list *t = list_first_entry(timers, struct cpu_timer_list, entry); - if (!--maxfire || tsk->se.sum_exec_runtime < t->expires.sched) { - tsk->cputime_expires.sched_exp = t->expires.sched; + if (!--maxfire || tsk->se.sum_exec_runtime < t->expires) { + tsk->cputime_expires.sched_exp = t->expires; break; } t->firing = 1; @@ -1030,7 +975,8 @@ static void stop_process_timers(struct signal_struct *sig) static u32 onecputick; static void check_cpu_itimer(struct task_struct *tsk, struct cpu_itimer *it, - cputime_t *expires, cputime_t cur_time, int signo) + unsigned long long *expires, + unsigned long long cur_time, int signo) { if (!it->expires) return; @@ -1068,7 +1014,7 @@ static void check_process_timers(struct task_struct *tsk, { int maxfire; struct signal_struct *const sig = tsk->signal; - cputime_t utime, ptime, virt_expires, prof_expires; + unsigned long long utime, ptime, virt_expires, prof_expires; unsigned long long sum_sched_runtime, sched_expires; struct list_head *timers = sig->cpu_timers; struct task_cputime cputime; @@ -1078,8 +1024,8 @@ static void check_process_timers(struct task_struct *tsk, * Collect the current process totals. */ thread_group_cputimer(tsk, &cputime); - utime = cputime.utime; - ptime = utime + cputime.stime; + utime = cputime_to_expires(cputime.utime); + ptime = utime + cputime_to_expires(cputime.stime); sum_sched_runtime = cputime.sum_exec_runtime; maxfire = 20; prof_expires = 0; @@ -1087,8 +1033,8 @@ static void check_process_timers(struct task_struct *tsk, struct cpu_timer_list *tl = list_first_entry(timers, struct cpu_timer_list, entry); - if (!--maxfire || ptime < tl->expires.cpu) { - prof_expires = tl->expires.cpu; + if (!--maxfire || ptime < tl->expires) { + prof_expires = tl->expires; break; } tl->firing = 1; @@ -1102,8 +1048,8 @@ static void check_process_timers(struct task_struct *tsk, struct cpu_timer_list *tl = list_first_entry(timers, struct cpu_timer_list, entry); - if (!--maxfire || utime < tl->expires.cpu) { - virt_expires = tl->expires.cpu; + if (!--maxfire || utime < tl->expires) { + virt_expires = tl->expires; break; } tl->firing = 1; @@ -1117,8 +1063,8 @@ static void check_process_timers(struct task_struct *tsk, struct cpu_timer_list *tl = list_first_entry(timers, struct cpu_timer_list, entry); - if (!--maxfire || sum_sched_runtime < tl->expires.sched) { - sched_expires = tl->expires.sched; + if (!--maxfire || sum_sched_runtime < tl->expires) { + sched_expires = tl->expires; break; } tl->firing = 1; @@ -1162,8 +1108,8 @@ static void check_process_timers(struct task_struct *tsk, } } - sig->cputime_expires.prof_exp = prof_expires; - sig->cputime_expires.virt_exp = virt_expires; + sig->cputime_expires.prof_exp = expires_to_cputime(prof_expires); + sig->cputime_expires.virt_exp = expires_to_cputime(virt_expires); sig->cputime_expires.sched_exp = sched_expires; if (task_cputime_zero(&sig->cputime_expires)) stop_process_timers(sig); @@ -1176,7 +1122,7 @@ static void check_process_timers(struct task_struct *tsk, void posix_cpu_timer_schedule(struct k_itimer *timer) { struct task_struct *p = timer->it.cpu.task; - union cpu_time_count now; + unsigned long long now; if (unlikely(p == NULL)) /* @@ -1205,7 +1151,7 @@ void posix_cpu_timer_schedule(struct k_itimer *timer) */ put_task_struct(p); timer->it.cpu.task = p = NULL; - timer->it.cpu.expires.sched = 0; + timer->it.cpu.expires = 0; goto out_unlock; } else if (unlikely(p->exit_state) && thread_group_empty(p)) { /* @@ -1387,7 +1333,7 @@ void run_posix_cpu_timers(struct task_struct *tsk) void set_process_cpu_timer(struct task_struct *tsk, unsigned int clock_idx, cputime_t *newval, cputime_t *oldval) { - union cpu_time_count now; + unsigned long long now; BUG_ON(clock_idx == CPUCLOCK_SCHED); cpu_timer_sample_group(clock_idx, tsk, &now); @@ -1399,17 +1345,17 @@ void set_process_cpu_timer(struct task_struct *tsk, unsigned int clock_idx, * it to be absolute. */ if (*oldval) { - if (*oldval <= now.cpu) { + if (*oldval <= now) { /* Just about to fire. */ *oldval = cputime_one_jiffy; } else { - *oldval -= now.cpu; + *oldval -= now; } } if (!*newval) goto out; - *newval += now.cpu; + *newval += now; } /* @@ -1459,7 +1405,7 @@ static int do_cpu_nanosleep(const clockid_t which_clock, int flags, } while (!signal_pending(current)) { - if (timer.it.cpu.expires.sched == 0) { + if (timer.it.cpu.expires == 0) { /* * Our timer fired and was reset, below * deletion can not fail. -- cgit v1.2.3 From 1a7fa510b38e518d11365883934f1afa41625424 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 28 Jun 2013 00:06:42 +0000 Subject: posix_cpu_timers: consolidate timer list cleanups Cleaning up the posix cpu timers on task exit shares some common code among timer list types, most notably the list traversal and expiry time update. Unify this in a common helper. Signed-off-by: Frederic Weisbecker Cc: Stanislaw Gruszka Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Ingo Molnar Cc: Oleg Nesterov Cc: KOSAKI Motohiro Cc: Olivier Langlois Signed-off-by: Andrew Morton --- kernel/posix-cpu-timers.c | 48 +++++++++++++++++++---------------------------- 1 file changed, 19 insertions(+), 29 deletions(-) (limited to 'kernel') diff --git a/kernel/posix-cpu-timers.c b/kernel/posix-cpu-timers.c index c3c4ea1225a4..b1450cee6d6d 100644 --- a/kernel/posix-cpu-timers.c +++ b/kernel/posix-cpu-timers.c @@ -399,6 +399,21 @@ static int posix_cpu_timer_del(struct k_itimer *timer) return ret; } +static void cleanup_timers_list(struct list_head *head, + unsigned long long curr) +{ + struct cpu_timer_list *timer, *next; + + list_for_each_entry_safe(timer, next, head, entry) { + list_del_init(&timer->entry); + if (timer->expires < curr) { + timer->expires = 0; + } else { + timer->expires -= curr; + } + } +} + /* * Clean out CPU timers still ticking when a thread exited. The task * pointer is cleared, and the expiry time is replaced with the residual @@ -409,37 +424,12 @@ static void cleanup_timers(struct list_head *head, cputime_t utime, cputime_t stime, unsigned long long sum_exec_runtime) { - struct cpu_timer_list *timer, *next; - cputime_t ptime = utime + stime; - list_for_each_entry_safe(timer, next, head, entry) { - list_del_init(&timer->entry); - if (timer->expires < cputime_to_expires(ptime)) { - timer->expires = 0; - } else { - timer->expires -= cputime_to_expires(ptime); - } - } - - ++head; - list_for_each_entry_safe(timer, next, head, entry) { - list_del_init(&timer->entry); - if (timer->expires < cputime_to_expires(utime)) { - timer->expires = 0; - } else { - timer->expires -= cputime_to_expires(utime); - } - } + cputime_t ptime = utime + stime; - ++head; - list_for_each_entry_safe(timer, next, head, entry) { - list_del_init(&timer->entry); - if (timer->expires < sum_exec_runtime) { - timer->expires = 0; - } else { - timer->expires -= sum_exec_runtime; - } - } + cleanup_timers_list(head, cputime_to_expires(ptime)); + cleanup_timers_list(++head, cputime_to_expires(utime)); + cleanup_timers_list(++head, sum_exec_runtime); } /* -- cgit v1.2.3 From 2473f3e7a97ce8bc0fe7596cdb361b21221418eb Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 28 Jun 2013 00:06:43 +0000 Subject: posix_cpu_timers: consolidate expired timers check Consolidate the common code amongst per thread and per process timers list on tick time. List traversal, expiry check and subsequent updates can be shared in a common helper. Signed-off-by: Frederic Weisbecker Cc: Stanislaw Gruszka Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Ingo Molnar Cc: Oleg Nesterov Cc: KOSAKI Motohiro Cc: Olivier Langlois Signed-off-by: Andrew Morton --- kernel/posix-cpu-timers.c | 118 +++++++++++++--------------------------------- 1 file changed, 33 insertions(+), 85 deletions(-) (limited to 'kernel') diff --git a/kernel/posix-cpu-timers.c b/kernel/posix-cpu-timers.c index b1450cee6d6d..92a4fbf44f86 100644 --- a/kernel/posix-cpu-timers.c +++ b/kernel/posix-cpu-timers.c @@ -862,6 +862,28 @@ static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec *itp) } } +static unsigned long long +check_timers_list(struct list_head *timers, + struct list_head *firing, + unsigned long long curr) +{ + int maxfire = 20; + + while (!list_empty(timers)) { + struct cpu_timer_list *t; + + t = list_first_entry(timers, struct cpu_timer_list, entry); + + if (!--maxfire || curr < t->expires) + return t->expires; + + t->firing = 1; + list_move_tail(&t->entry, firing); + } + + return 0; +} + /* * Check for any per-thread CPU timers that have fired and move them off * the tsk->cpu_timers[N] list onto the firing list. Here we update the @@ -870,54 +892,20 @@ static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec *itp) static void check_thread_timers(struct task_struct *tsk, struct list_head *firing) { - int maxfire; struct list_head *timers = tsk->cpu_timers; struct signal_struct *const sig = tsk->signal; + struct task_cputime *tsk_expires = &tsk->cputime_expires; + unsigned long long expires; unsigned long soft; - maxfire = 20; - tsk->cputime_expires.prof_exp = 0; - while (!list_empty(timers)) { - struct cpu_timer_list *t = list_first_entry(timers, - struct cpu_timer_list, - entry); - if (!--maxfire || prof_ticks(tsk) < t->expires) { - tsk->cputime_expires.prof_exp = expires_to_cputime(t->expires); - break; - } - t->firing = 1; - list_move_tail(&t->entry, firing); - } + expires = check_timers_list(timers, firing, prof_ticks(tsk)); + tsk_expires->prof_exp = expires_to_cputime(expires); - ++timers; - maxfire = 20; - tsk->cputime_expires.virt_exp = 0; - while (!list_empty(timers)) { - struct cpu_timer_list *t = list_first_entry(timers, - struct cpu_timer_list, - entry); - if (!--maxfire || virt_ticks(tsk) < t->expires) { - tsk->cputime_expires.virt_exp = expires_to_cputime(t->expires); - break; - } - t->firing = 1; - list_move_tail(&t->entry, firing); - } + expires = check_timers_list(++timers, firing, virt_ticks(tsk)); + tsk_expires->virt_exp = expires_to_cputime(expires); - ++timers; - maxfire = 20; - tsk->cputime_expires.sched_exp = 0; - while (!list_empty(timers)) { - struct cpu_timer_list *t = list_first_entry(timers, - struct cpu_timer_list, - entry); - if (!--maxfire || tsk->se.sum_exec_runtime < t->expires) { - tsk->cputime_expires.sched_exp = t->expires; - break; - } - t->firing = 1; - list_move_tail(&t->entry, firing); - } + tsk_expires->sched_exp = check_timers_list(++timers, firing, + tsk->se.sum_exec_runtime); /* * Check for the special case thread timers. @@ -1002,7 +990,6 @@ static void check_cpu_itimer(struct task_struct *tsk, struct cpu_itimer *it, static void check_process_timers(struct task_struct *tsk, struct list_head *firing) { - int maxfire; struct signal_struct *const sig = tsk->signal; unsigned long long utime, ptime, virt_expires, prof_expires; unsigned long long sum_sched_runtime, sched_expires; @@ -1017,49 +1004,10 @@ static void check_process_timers(struct task_struct *tsk, utime = cputime_to_expires(cputime.utime); ptime = utime + cputime_to_expires(cputime.stime); sum_sched_runtime = cputime.sum_exec_runtime; - maxfire = 20; - prof_expires = 0; - while (!list_empty(timers)) { - struct cpu_timer_list *tl = list_first_entry(timers, - struct cpu_timer_list, - entry); - if (!--maxfire || ptime < tl->expires) { - prof_expires = tl->expires; - break; - } - tl->firing = 1; - list_move_tail(&tl->entry, firing); - } - ++timers; - maxfire = 20; - virt_expires = 0; - while (!list_empty(timers)) { - struct cpu_timer_list *tl = list_first_entry(timers, - struct cpu_timer_list, - entry); - if (!--maxfire || utime < tl->expires) { - virt_expires = tl->expires; - break; - } - tl->firing = 1; - list_move_tail(&tl->entry, firing); - } - - ++timers; - maxfire = 20; - sched_expires = 0; - while (!list_empty(timers)) { - struct cpu_timer_list *tl = list_first_entry(timers, - struct cpu_timer_list, - entry); - if (!--maxfire || sum_sched_runtime < tl->expires) { - sched_expires = tl->expires; - break; - } - tl->firing = 1; - list_move_tail(&tl->entry, firing); - } + prof_expires = check_timers_list(timers, firing, ptime); + virt_expires = check_timers_list(++timers, firing, utime); + sched_expires = check_timers_list(++timers, firing, sum_sched_runtime); /* * Check for the special case process timers. -- cgit v1.2.3 From 76cdcdd979ce00f5037804d73da583fb488ec1b2 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 28 Jun 2013 00:06:43 +0000 Subject: posix-timers: correctly get dying task time sample in posix_cpu_timer_schedule() In order to re-arm a timer after it fired, we take a sample of the current process or thread cputime. If the task is dying though, we don't arm anything but we cache the remaining timer expiration delta for further reads. Something similar is performed in posix_cpu_timer_get() but here we forget to take the process wide cputime sample before caching it. As a result we are storing random stack content, leading every further reads of that timer to return junk values. Fix this by taking the appropriate sample in the case of process wide timers. This probably doesn't matter much in practice because, at this stage, the thread is the last one in the group and we reached exit_notify(). This implies that we called exit_itimers() and there should be no more timers to handle for that task. So this is likely dead code anyway but let's fix the current logic and the warning that came along: kernel/posix-cpu-timers.c: In function 'posix_cpu_timer_schedule': kernel/posix-cpu-timers.c:1127: warning: 'now' may be used uninitialized in this function Then we can start to think further about cleaning up that code. Reported-by: Andrew Morton Reported-by: Chen Gang Signed-off-by: Frederic Weisbecker Cc: Stanislaw Gruszka Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Ingo Molnar Cc: Oleg Nesterov Cc: Chen Gang Cc: KOSAKI Motohiro Cc: Olivier Langlois Signed-off-by: Andrew Morton --- kernel/posix-cpu-timers.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/posix-cpu-timers.c b/kernel/posix-cpu-timers.c index 92a4fbf44f86..4ebd8ad07c66 100644 --- a/kernel/posix-cpu-timers.c +++ b/kernel/posix-cpu-timers.c @@ -1097,6 +1097,7 @@ void posix_cpu_timer_schedule(struct k_itimer *timer) * not yet reaped. Take this opportunity to * drop our task ref. */ + cpu_timer_sample_group(timer->it_clock, p, &now); clear_dead_task(timer, now); goto out_unlock; } -- cgit v1.2.3 From a0b2062b0904ef07944c4a6e4d0f88ee44f1e9f2 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 28 Jun 2013 00:06:43 +0000 Subject: posix_timers: fix racy timer delta caching on task exit When a task exits, we perform a caching of the remaining cputime delta before expiring of its timers. This is done from the following places: * When the task is reaped. We iterate through its list of posix cpu timers and store the remaining timer delta to the timer struct instead of the absolute value. (See posix_cpu_timers_exit() / posix_cpu_timers_exit_group() ) * When we call posix_cpu_timer_get() or posix_cpu_timer_schedule(). If the timer's task is considered dying when watched from these places, the same conversion from absolute to relative expiry time is performed. Then the given task's reference is released. (See clear_dead_task() ). The relevance of this caching is questionable but this is another and deeper debate. The big issue here is that these two sources of caching don't mix up very well together. More specifically, the caching can easily be done twice, resulting in a wrong delta as it gets spuriously substracted a second time by the elapsed clock. This can happen in the following scenario: 1) The task exits and gets reaped: we call posix_cpu_timers_exit() and the absolute timer expiry values are converted to a relative delta. 2) timer_gettime() -> posix_cpu_timer_get() is called and relies on clear_dead_task() because tsk->exit_state == EXIT_DEAD. The delta gets substracted again by the elapsed clock and we return a wrong result. To fix this, just remove the caching done on task reaping time. It doesn't bring much value on its own. The caching done from posix_cpu_timer_get/schedule is enough. And it would also be hard to get it really right: we could make it put and clear the target task in the timer struct so that readers know if they are dealing with a relative cached of absolute value. But it would be racy. The only safe way to do it would be to lock the itimer->it_lock so that we know nobody reads the cputime expiry value while we modify it and its target task reference. Doing so would involve some funny workarounds to avoid circular lock against the sighand lock. There is just no reason to maintain this. The user visible effect of this patch can be observed by running the following code: it creates a subthread that launches a posix cputimer which expires after 10 seconds. But then the subthread only busy loops for 2 seconds and exits. The parent reaps the subthread and read the timer value. Its expected value should the be the initial timer's expiration value minus the cputime elapsed in the subthread. Roughly 10 - 2 = 8 seconds: #include #include #include #include #include static timer_t id; static struct itimerspec val = { .it_value.tv_sec = 10, }, new; static void *thread(void *unused) { int err; struct timeval start, end, diff; timer_create(CLOCK_THREAD_CPUTIME_ID, NULL, &id); if (err < 0) { perror("Can't create timer\n"); return NULL; } /* Arm 10 sec timer */ err = timer_settime(id, 0, &val, NULL); if (err < 0) { perror("Can't set timer\n"); return NULL; } /* Exit after 2 seconds of execution */ gettimeofday(&start, NULL); do { gettimeofday(&end, NULL); timersub(&end, &start, &diff); } while (diff.tv_sec < 2); return NULL; } int main(int argc, char **argv) { pthread_t pthread; int err; err = pthread_create(&pthread, NULL, thread, NULL); if (err) { perror("Can't create thread\n"); return -1; } pthread_join(pthread, NULL); /* Just wait a little bit to make sure the child got reaped */ sleep(1); err = timer_gettime(id, &new); if (err) perror("Can't get timer value\n"); printf("%d %ld\n", new.it_value.tv_sec, new.it_value.tv_nsec); return 0; } Before the patch: $ ./posix_cpu_timers 6 2278074 After the patch: $ ./posix_cpu_timers 8 1158766 Before the patch, the elapsed time got two more seconds spuriously accounted. Signed-off-by: Frederic Weisbecker Cc: Stanislaw Gruszka Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Ingo Molnar Cc: Oleg Nesterov Cc: KOSAKI Motohiro Cc: Olivier Langlois Signed-off-by: Andrew Morton --- kernel/posix-cpu-timers.c | 22 +++++++++++----------- 1 file changed, 11 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/posix-cpu-timers.c b/kernel/posix-cpu-timers.c index 4ebd8ad07c66..c7f31aa272f7 100644 --- a/kernel/posix-cpu-timers.c +++ b/kernel/posix-cpu-timers.c @@ -404,14 +404,8 @@ static void cleanup_timers_list(struct list_head *head, { struct cpu_timer_list *timer, *next; - list_for_each_entry_safe(timer, next, head, entry) { + list_for_each_entry_safe(timer, next, head, entry) list_del_init(&timer->entry); - if (timer->expires < curr) { - timer->expires = 0; - } else { - timer->expires -= curr; - } - } } /* @@ -459,15 +453,21 @@ void posix_cpu_timers_exit_group(struct task_struct *tsk) tsk->se.sum_exec_runtime + sig->sum_sched_runtime); } -static void clear_dead_task(struct k_itimer *timer, unsigned long long now) +static void clear_dead_task(struct k_itimer *itimer, unsigned long long now) { + struct cpu_timer_list *timer = &itimer->it.cpu; + /* * That's all for this thread or process. * We leave our residual in expires to be reported. */ - put_task_struct(timer->it.cpu.task); - timer->it.cpu.task = NULL; - timer->it.cpu.expires -= now; + put_task_struct(timer->task); + timer->task = NULL; + if (timer->expires < now) { + timer->expires = 0; + } else { + timer->expires -= now; + } } static inline int expires_gt(cputime_t expires, cputime_t new_exp) -- cgit v1.2.3 From 0ed5fd138539940a493dc69359cb2f49de70ad89 Mon Sep 17 00:00:00 2001 From: Jiang Liu Date: Wed, 3 Jul 2013 15:03:43 -0700 Subject: mm: use totalram_pages instead of num_physpages at runtime The global variable num_physpages is scheduled to be removed, so use totalram_pages instead of num_physpages at runtime. Signed-off-by: Jiang Liu Cc: Miklos Szeredi Cc: "David S. Miller" Cc: Alexey Kuznetsov Cc: James Morris Cc: Hideaki YOSHIFUJI Cc: Patrick McHardy Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/power/snapshot.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c index 0de28576807d..8b5d1cd933f4 100644 --- a/kernel/power/snapshot.c +++ b/kernel/power/snapshot.c @@ -1651,7 +1651,7 @@ unsigned long snapshot_get_image_size(void) static int init_header(struct swsusp_info *info) { memset(info, 0, sizeof(struct swsusp_info)); - info->num_physpages = num_physpages; + info->num_physpages = get_num_physpages(); info->image_pages = nr_copy_pages; info->pages = snapshot_get_image_size(); info->size = info->pages; @@ -1795,7 +1795,7 @@ static int check_header(struct swsusp_info *info) char *reason; reason = check_image_kernel(info); - if (!reason && info->num_physpages != num_physpages) + if (!reason && info->num_physpages != get_num_physpages()) reason = "memory size"; if (reason) { printk(KERN_ERR "PM: Image mismatch: %s\n", reason); -- cgit v1.2.3 From f170168b9a0b61ea1e647b082b38f605f1d3de3e Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Wed, 3 Jul 2013 15:04:58 -0700 Subject: drivers: avoid parsing names as kthread_run() format strings Calling kthread_run with a single name parameter causes it to be handled as a format string. Many callers are passing potentially dynamic string content, so use "%s" in those cases to avoid any potential accidents. Signed-off-by: Kees Cook Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/rcutree.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index cf3adc6fe001..e08abb9461ac 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -3026,7 +3026,7 @@ static int __init rcu_spawn_gp_kthread(void) struct task_struct *t; for_each_rcu_flavor(rsp) { - t = kthread_run(rcu_gp_kthread, rsp, rsp->name); + t = kthread_run(rcu_gp_kthread, rsp, "%s", rsp->name); BUG_ON(IS_ERR(t)); rnp = rcu_get_root(rsp); raw_spin_lock_irqsave(&rnp->lock, flags); -- cgit v1.2.3 From 7ec75e1ca1bd35872a5c7b33da4b05395bc74364 Mon Sep 17 00:00:00 2001 From: liguang Date: Wed, 3 Jul 2013 15:05:00 -0700 Subject: kernel/sys.c: sys_reboot(): fix malformed panic message If LINUX_REBOOT_CMD_HALT for reboot failed, the message "cannot halt" will stay on the same line with the next message, so append a '\n'. Signed-off-by: liguang Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sys.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sys.c b/kernel/sys.c index 2bbd9a73b54c..c1da757a97b0 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -511,7 +511,7 @@ SYSCALL_DEFINE4(reboot, int, magic1, int, magic2, unsigned int, cmd, case LINUX_REBOOT_CMD_HALT: kernel_halt(); do_exit(0); - panic("cannot halt"); + panic("cannot halt.\n"); case LINUX_REBOOT_CMD_POWER_OFF: kernel_power_off(); -- cgit v1.2.3 From 45c64940c8bb64a042464ecec89d95eb4cce9b07 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Wed, 3 Jul 2013 15:05:01 -0700 Subject: kernel/sys.c:do_sysinfo(): use get_monotonic_boottime() Change do_sysinfo() to use get_monotonic_boottime() instead of do_posix_clock_monotonic_gettime() + monotonic_to_bootbased(). Signed-off-by: Oleg Nesterov Cc: "Eric W. Biederman" Acked-by: John Stultz Cc: Tomas Janousek Cc: Tomas Smetana Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sys.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sys.c b/kernel/sys.c index c1da757a97b0..7bf50dcc6d53 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -2355,8 +2355,7 @@ static int do_sysinfo(struct sysinfo *info) memset(info, 0, sizeof(struct sysinfo)); - ktime_get_ts(&tp); - monotonic_to_bootbased(&tp); + get_monotonic_boottime(&tp); info->uptime = tp.tv_sec + (tp.tv_nsec ? 1 : 0); get_avenrun(info->loads, 0, SI_LOAD_SHIFT - FSHIFT); -- cgit v1.2.3 From 10fb46d5f79147620d0afda7d3d51302a1e38191 Mon Sep 17 00:00:00 2001 From: Mathias Krause Date: Wed, 3 Jul 2013 15:05:39 -0700 Subject: kprobes: handle empty/invalid input to debugfs "enabled" file When writing invalid input to 'debug/kprobes/enabled' it'll silently be ignored. Even worse, when writing an empty string to this file, the outcome is purely random as the switch statement will make its decision based on the value of an uninitialized stack variable. Fix this by handling invalid/empty input as error returning -EINVAL. Signed-off-by: Mathias Krause Cc: Ananth N Mavinakayanahalli Cc: Anil S Keshavamurthy Cc: "David S. Miller" Cc: Masami Hiramatsu Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/kprobes.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'kernel') diff --git a/kernel/kprobes.c b/kernel/kprobes.c index bddf3b201a48..6e33498d665c 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -2332,6 +2332,7 @@ static ssize_t write_enabled_file_bool(struct file *file, if (copy_from_user(buf, user_buf, buf_size)) return -EFAULT; + buf[buf_size] = '\0'; switch (buf[0]) { case 'y': case 'Y': @@ -2343,6 +2344,8 @@ static ssize_t write_enabled_file_bool(struct file *file, case '0': disarm_all_kprobes(); break; + default: + return -EINVAL; } return count; -- cgit v1.2.3 From 29000caecbe87b6b66f144f72111f0d02fbbf0c1 Mon Sep 17 00:00:00 2001 From: Andrey Vagin Date: Wed, 3 Jul 2013 15:08:12 -0700 Subject: ptrace: add ability to get/set signal-blocked mask crtools uses a parasite code for dumping processes. The parasite code is injected into a process with help PTRACE_SEIZE. Currently crtools blocks signals from a parasite code. If a process has pending signals, crtools wait while a process handles these signals. This method is not suitable for stopped tasks. A stopped task can have a few pending signals, when we will try to execute a parasite code, we will need to drop SIGSTOP, but all other signals must remain pending, because a state of processes must not be changed during checkpointing. This patch adds two ptrace commands to set/get signal-blocked mask. I think gdb can use this commands too. [akpm@linux-foundation.org: be consistent with brace layout] Signed-off-by: Andrey Vagin Reviewed-by: Oleg Nesterov Cc: Roland McGrath Cc: Michael Kerrisk Cc: Pavel Emelyanov Cc: Cyrill Gorcunov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/ptrace.c | 44 ++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 42 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/ptrace.c b/kernel/ptrace.c index 335a7ae697f5..ba5e6cea181a 100644 --- a/kernel/ptrace.c +++ b/kernel/ptrace.c @@ -844,6 +844,47 @@ int ptrace_request(struct task_struct *child, long request, ret = ptrace_setsiginfo(child, &siginfo); break; + case PTRACE_GETSIGMASK: + if (addr != sizeof(sigset_t)) { + ret = -EINVAL; + break; + } + + if (copy_to_user(datavp, &child->blocked, sizeof(sigset_t))) + ret = -EFAULT; + else + ret = 0; + + break; + + case PTRACE_SETSIGMASK: { + sigset_t new_set; + + if (addr != sizeof(sigset_t)) { + ret = -EINVAL; + break; + } + + if (copy_from_user(&new_set, datavp, sizeof(sigset_t))) { + ret = -EFAULT; + break; + } + + sigdelsetmask(&new_set, sigmask(SIGKILL)|sigmask(SIGSTOP)); + + /* + * Every thread does recalc_sigpending() after resume, so + * retarget_shared_pending() and recalc_sigpending() are not + * called here. + */ + spin_lock_irq(&child->sighand->siglock); + child->blocked = new_set; + spin_unlock_irq(&child->sighand->siglock); + + ret = 0; + break; + } + case PTRACE_INTERRUPT: /* * Stop tracee without any side-effect on signal or job @@ -948,8 +989,7 @@ int ptrace_request(struct task_struct *child, long request, #ifdef CONFIG_HAVE_ARCH_TRACEHOOK case PTRACE_GETREGSET: - case PTRACE_SETREGSET: - { + case PTRACE_SETREGSET: { struct iovec kiov; struct iovec __user *uiov = datavp; -- cgit v1.2.3 From 7f57cfa4e2aa29fabe69e41529fd26578adc9b58 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Wed, 3 Jul 2013 15:08:15 -0700 Subject: usermodehelper: kill the sub_info->path[0] check call_usermodehelper_exec() does nothing but returns success if path[0] == 0. The only user which needs this strange feature is request_module(), it can check modprobe_path[0] itself like other users do if they want to detect the "disabled by admin" case. Kill it. Not only it looks strange, it can confuse other callers. And this allows us to revert 264b83c0 ("usermodehelper: check subprocess_info->path != NULL"), do_execve(NULL) is safe. Signed-off-by: Oleg Nesterov Acked-by: Rusty Russell Cc: Lucas De Marchi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/kmod.c | 11 +++-------- 1 file changed, 3 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/kmod.c b/kernel/kmod.c index 8241906c4b61..fb326365b694 100644 --- a/kernel/kmod.c +++ b/kernel/kmod.c @@ -147,6 +147,9 @@ int __request_module(bool wait, const char *fmt, ...) */ WARN_ON_ONCE(wait && current_is_async()); + if (!modprobe_path[0]) + return 0; + va_start(args, fmt); ret = vsnprintf(module_name, MODULE_NAME_LEN, fmt, args); va_end(args); @@ -569,14 +572,6 @@ int call_usermodehelper_exec(struct subprocess_info *sub_info, int wait) int retval = 0; helper_lock(); - if (!sub_info->path) { - retval = -EINVAL; - goto out; - } - - if (sub_info->path[0] == '\0') - goto out; - if (!khelper_wq || usermodehelper_disabled) { retval = -EBUSY; goto out; -- cgit v1.2.3 From 81dabb464139324c005159f5afba377104d8828d Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Wed, 3 Jul 2013 15:08:26 -0700 Subject: exit.c: unexport __set_special_pids() Move __set_special_pids() from exit.c to sys.c close to its single caller and make it static. And rename it to set_special_pids(), another helper with this name has gone away. Signed-off-by: Oleg Nesterov Cc: "Eric W. Biederman" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 11 ----------- kernel/sys.c | 13 ++++++++++++- 2 files changed, 12 insertions(+), 12 deletions(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index 7bb73f9d09db..3a77cd9390a1 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -312,17 +312,6 @@ kill_orphaned_pgrp(struct task_struct *tsk, struct task_struct *parent) } } -void __set_special_pids(struct pid *pid) -{ - struct task_struct *curr = current->group_leader; - - if (task_session(curr) != pid) - change_pid(curr, PIDTYPE_SID, pid); - - if (task_pgrp(curr) != pid) - change_pid(curr, PIDTYPE_PGID, pid); -} - /* * Let kernel threads use this to say that they allow a certain signal. * Must not be used if kthread was cloned with CLONE_SIGHAND. diff --git a/kernel/sys.c b/kernel/sys.c index 7bf50dcc6d53..071de900c824 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -1309,6 +1309,17 @@ out: return retval; } +static void set_special_pids(struct pid *pid) +{ + struct task_struct *curr = current->group_leader; + + if (task_session(curr) != pid) + change_pid(curr, PIDTYPE_SID, pid); + + if (task_pgrp(curr) != pid) + change_pid(curr, PIDTYPE_PGID, pid); +} + SYSCALL_DEFINE0(setsid) { struct task_struct *group_leader = current->group_leader; @@ -1328,7 +1339,7 @@ SYSCALL_DEFINE0(setsid) goto out; group_leader->signal->leader = 1; - __set_special_pids(sid); + set_special_pids(sid); proc_clear_tty(group_leader); -- cgit v1.2.3 From b57922b6c76c3ee401bb32fd3f298409dd6e6a53 Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Wed, 3 Jul 2013 15:08:29 -0700 Subject: fork: reorder permissions when violating number of processes limits When a task is attempting to violate the RLIMIT_NPROC limit we have a check to see if the task is sufficiently priviledged. The check first looks at CAP_SYS_ADMIN, then CAP_SYS_RESOURCE, then if the task is uid=0. A result is that tasks which are allowed by the uid=0 check are first checked against the security subsystem. This results in the security subsystem auditting a denial for sys_admin and sys_resource and then the task passing the uid=0 check. This patch rearranges the code to first check uid=0, since if we pass that we shouldn't hit the security system at all. We then check sys_resource, since it is the smallest capability which will solve the problem. Lastly we check the fallback everything cap_sysadmin. We don't want to give this capability many places since it is so powerful. This will eliminate many of the false positive/needless denial messages we get when a root task tries to violate the nproc limit. (note that kthreads count against root, so on a sufficiently large machine we can actually get past the default limits before any userspace tasks are launched.) Signed-off-by: Eric Paris Cc: Al Viro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index 987b28a1f01b..09dbda38a54b 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1199,8 +1199,8 @@ static struct task_struct *copy_process(unsigned long clone_flags, retval = -EAGAIN; if (atomic_read(&p->real_cred->user->processes) >= task_rlimit(p, RLIMIT_NPROC)) { - if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_RESOURCE) && - p->real_cred->user != INIT_USER) + if (p->real_cred->user != INIT_USER && + !capable(CAP_SYS_RESOURCE) && !capable(CAP_SYS_ADMIN)) goto bad_fork_free; } current->flags &= ~PF_NPROC_EXCEEDED; -- cgit v1.2.3 From 80628ca06c5d42929de6bc22c0a41589a834d151 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Wed, 3 Jul 2013 15:08:30 -0700 Subject: kernel/fork.c:copy_process(): unify CLONE_THREAD-or-thread_group_leader code Cleanup and preparation for the next changes. Move the "if (clone_flags & CLONE_THREAD)" code down under "if (likely(p->pid))" and turn it into into the "else" branch. This makes the process/thread initialization more symmetrical and removes one check. Signed-off-by: Oleg Nesterov Cc: "Eric W. Biederman" Cc: Michal Hocko Cc: Pavel Emelyanov Cc: Sergey Dyasly Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 15 +++++++-------- 1 file changed, 7 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index 09dbda38a54b..417cb864e20c 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1446,14 +1446,6 @@ static struct task_struct *copy_process(unsigned long clone_flags, goto bad_fork_free_pid; } - if (clone_flags & CLONE_THREAD) { - current->signal->nr_threads++; - atomic_inc(¤t->signal->live); - atomic_inc(¤t->signal->sigcnt); - p->group_leader = current->group_leader; - list_add_tail_rcu(&p->thread_group, &p->group_leader->thread_group); - } - if (likely(p->pid)) { ptrace_init_task(p, (clone_flags & CLONE_PTRACE) || trace); @@ -1470,6 +1462,13 @@ static struct task_struct *copy_process(unsigned long clone_flags, list_add_tail(&p->sibling, &p->real_parent->children); list_add_tail_rcu(&p->tasks, &init_task.tasks); __this_cpu_inc(process_counts); + } else { + current->signal->nr_threads++; + atomic_inc(¤t->signal->live); + atomic_inc(¤t->signal->sigcnt); + p->group_leader = current->group_leader; + list_add_tail_rcu(&p->thread_group, + &p->group_leader->thread_group); } attach_pid(p, PIDTYPE_PID, pid); nr_threads++; -- cgit v1.2.3 From 8190773985141f063e1d6dc10200527c655abfb5 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Wed, 3 Jul 2013 15:08:31 -0700 Subject: kernel/fork.c:copy_process(): don't add the uninitialized child to thread/task/pid lists copy_process() adds the new child to thread_group/init_task.tasks list and then does attach_pid(child, PIDTYPE_PID). This means that the lockless next_thread() or next_task() can see this thread with the wrong pid. Say, "ls /proc/pid/task" can list the same inode twice. We could move attach_pid(child, PIDTYPE_PID) up, but in this case find_task_by_vpid() can find the new thread before it was fully initialized. And this is already true for PIDTYPE_PGID/PIDTYPE_SID, With this patch copy_process() initializes child->pids[*].pid first, then calls attach_pid() to insert the task into the pid->tasks list. attach_pid() no longer need the "struct pid*" argument, it is always called after pid_link->pid was already set. Signed-off-by: Oleg Nesterov Cc: "Eric W. Biederman" Cc: Michal Hocko Cc: Pavel Emelyanov Cc: Sergey Dyasly Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 16 +++++++++++++--- kernel/pid.c | 12 ++++-------- 2 files changed, 17 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index 417cb864e20c..7d6962fb6156 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1121,6 +1121,12 @@ static void posix_cpu_timers_init(struct task_struct *tsk) INIT_LIST_HEAD(&tsk->cpu_timers[2]); } +static inline void +init_task_pid(struct task_struct *task, enum pid_type type, struct pid *pid) +{ + task->pids[type].pid = pid; +} + /* * This creates a new process as a copy of the old one, * but does not actually start it yet. @@ -1449,7 +1455,11 @@ static struct task_struct *copy_process(unsigned long clone_flags, if (likely(p->pid)) { ptrace_init_task(p, (clone_flags & CLONE_PTRACE) || trace); + init_task_pid(p, PIDTYPE_PID, pid); if (thread_group_leader(p)) { + init_task_pid(p, PIDTYPE_PGID, task_pgrp(current)); + init_task_pid(p, PIDTYPE_SID, task_session(current)); + if (is_child_reaper(pid)) { ns_of_pid(pid)->child_reaper = p; p->signal->flags |= SIGNAL_UNKILLABLE; @@ -1457,10 +1467,10 @@ static struct task_struct *copy_process(unsigned long clone_flags, p->signal->leader_pid = pid; p->signal->tty = tty_kref_get(current->signal->tty); - attach_pid(p, PIDTYPE_PGID, task_pgrp(current)); - attach_pid(p, PIDTYPE_SID, task_session(current)); list_add_tail(&p->sibling, &p->real_parent->children); list_add_tail_rcu(&p->tasks, &init_task.tasks); + attach_pid(p, PIDTYPE_PGID); + attach_pid(p, PIDTYPE_SID); __this_cpu_inc(process_counts); } else { current->signal->nr_threads++; @@ -1470,7 +1480,7 @@ static struct task_struct *copy_process(unsigned long clone_flags, list_add_tail_rcu(&p->thread_group, &p->group_leader->thread_group); } - attach_pid(p, PIDTYPE_PID, pid); + attach_pid(p, PIDTYPE_PID); nr_threads++; } diff --git a/kernel/pid.c b/kernel/pid.c index 0db3e791a06d..61980cefb1f5 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -373,14 +373,10 @@ EXPORT_SYMBOL_GPL(find_vpid); /* * attach_pid() must be called with the tasklist_lock write-held. */ -void attach_pid(struct task_struct *task, enum pid_type type, - struct pid *pid) +void attach_pid(struct task_struct *task, enum pid_type type) { - struct pid_link *link; - - link = &task->pids[type]; - link->pid = pid; - hlist_add_head_rcu(&link->node, &pid->tasks[type]); + struct pid_link *link = &task->pids[type]; + hlist_add_head_rcu(&link->node, &link->pid->tasks[type]); } static void __change_pid(struct task_struct *task, enum pid_type type, @@ -412,7 +408,7 @@ void change_pid(struct task_struct *task, enum pid_type type, struct pid *pid) { __change_pid(task, type, pid); - attach_pid(task, type, pid); + attach_pid(task, type); } /* transfer_pid is an optimization of attach_pid(new), detach_pid(old) */ -- cgit v1.2.3 From 18c830df771f2ba8b4699fea9af1492275ae627b Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Wed, 3 Jul 2013 15:08:32 -0700 Subject: kernel/fork.c:copy_process(): consolidate the lockless CLONE_THREAD checks copy_process() does a lot of "chaotic" initializations and checks CLONE_THREAD twice before it takes tasklist. In particular it sets "p->group_leader = p" and then changes it again under tasklist if !thread_group_leader(p). This looks a bit confusing, lets create a single "if (CLONE_THREAD)" block which initializes ->exit_signal, ->group_leader, and ->tgid. Signed-off-by: Oleg Nesterov Cc: "Eric W. Biederman" Cc: Michal Hocko Cc: Pavel Emelyanov Cc: Sergey Dyasly Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 33 ++++++++++++++++----------------- 1 file changed, 16 insertions(+), 17 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index 7d6962fb6156..6e6a1c11b3e5 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1360,11 +1360,6 @@ static struct task_struct *copy_process(unsigned long clone_flags, goto bad_fork_cleanup_io; } - p->pid = pid_nr(pid); - p->tgid = p->pid; - if (clone_flags & CLONE_THREAD) - p->tgid = current->tgid; - p->set_child_tid = (clone_flags & CLONE_CHILD_SETTID) ? child_tidptr : NULL; /* * Clear TID on mm_release()? @@ -1400,12 +1395,19 @@ static struct task_struct *copy_process(unsigned long clone_flags, clear_all_latency_tracing(p); /* ok, now we should be set up.. */ - if (clone_flags & CLONE_THREAD) + p->pid = pid_nr(pid); + if (clone_flags & CLONE_THREAD) { p->exit_signal = -1; - else if (clone_flags & CLONE_PARENT) - p->exit_signal = current->group_leader->exit_signal; - else - p->exit_signal = (clone_flags & CSIGNAL); + p->group_leader = current->group_leader; + p->tgid = current->tgid; + } else { + if (clone_flags & CLONE_PARENT) + p->exit_signal = current->group_leader->exit_signal; + else + p->exit_signal = (clone_flags & CSIGNAL); + p->group_leader = p; + p->tgid = p->pid; + } p->pdeath_signal = 0; p->exit_state = 0; @@ -1414,15 +1416,13 @@ static struct task_struct *copy_process(unsigned long clone_flags, p->nr_dirtied_pause = 128 >> (PAGE_SHIFT - 10); p->dirty_paused_when = 0; - /* - * Ok, make it visible to the rest of the system. - * We dont wake it up yet. - */ - p->group_leader = p; INIT_LIST_HEAD(&p->thread_group); p->task_works = NULL; - /* Need tasklist lock for parent etc handling! */ + /* + * Make it visible to the rest of the system, but dont wake it up yet. + * Need tasklist lock for parent etc handling! + */ write_lock_irq(&tasklist_lock); /* CLONE_PARENT re-uses the old parent */ @@ -1476,7 +1476,6 @@ static struct task_struct *copy_process(unsigned long clone_flags, current->signal->nr_threads++; atomic_inc(¤t->signal->live); atomic_inc(¤t->signal->sigcnt); - p->group_leader = current->group_leader; list_add_tail_rcu(&p->thread_group, &p->group_leader->thread_group); } -- cgit v1.2.3 From 8f75af44eed0c81f818b5b345023cebfe5209400 Mon Sep 17 00:00:00 2001 From: "Raphael S. Carvalho" Date: Wed, 3 Jul 2013 15:09:02 -0700 Subject: kernel/pid.c: move statement Move statement to static initilization of init_pid_ns. Signed-off-by: Raphael S. Carvalho Cc: "Eric W. Biederman" Acked-by: Serge Hallyn Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/pid.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/pid.c b/kernel/pid.c index 61980cefb1f5..66505c1dfc51 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -75,6 +75,7 @@ struct pid_namespace init_pid_ns = { [ 0 ... PIDMAP_ENTRIES-1] = { ATOMIC_INIT(BITS_PER_PAGE), NULL } }, .last_pid = 0, + .nr_hashed = PIDNS_HASH_ADDING, .level = 0, .child_reaper = &init_task, .user_ns = &init_user_ns, @@ -590,7 +591,6 @@ void __init pidmap_init(void) /* Reserve PID 0. We never call free_pidmap(0) */ set_bit(0, init_pid_ns.pidmap[0].page); atomic_dec(&init_pid_ns.pidmap[0].nr_free); - init_pid_ns.nr_hashed = PIDNS_HASH_ADDING; init_pid_ns.pid_cachep = KMEM_CACHE(pid, SLAB_HWCACHE_ALIGN | SLAB_PANIC); -- cgit v1.2.3 From 0786f7b225ba1edd801dc4bfbf6191d058b943a2 Mon Sep 17 00:00:00 2001 From: Kevin Hao Date: Wed, 3 Jul 2013 15:09:16 -0700 Subject: kernel/resource.c: remove the unneeded assignment in function __find_resource This line was introduced by fcb11918 ("resources: add arch hook for preventing allocation in reserved areas"). But the struct tmp was already assigned to *new in the above line, so this seems superfluous. Just remove it. Signed-off-by: Kevin Hao Cc: Bjorn Helgaas Cc: Jesse Barnes Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/resource.c | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/resource.c b/kernel/resource.c index 77bf11a86c7d..3f285dce9347 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -449,7 +449,6 @@ static int __find_resource(struct resource *root, struct resource *old, struct resource *this = root->child; struct resource tmp = *new, avail, alloc; - tmp.flags = new->flags; tmp.start = root->start; /* * Skip past an allocated resource that starts at 0, since the assignment -- cgit v1.2.3 From fa18f7bde3ad4568d1d343b60d963bfbd8dc3991 Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Sun, 26 May 2013 17:35:41 -0400 Subject: posix-cpu-timers: don't account cpu timer after stopped thread runtime accounting When tsk->signal->cputimer->running is 1, signal->cputimer (i.e. per process timer account) and tsk->sum_sched_runtime (i.e. per thread timer account) increase at the same pace because update_curr() increases both accounting. However, there is one exception. When thread exiting, __exit_signal() turns over task's sum_shced_runtime to sig->sum_sched_runtime, but it doesn't stop signal->cputimer accounting. This inconsistency makes POSIX timer wake up too early. This patch fixes it. Original-patch-by: Olivier Langlois Cc: Thomas Gleixner Cc: Ingo Molnar Cc: Peter Zijlstra Acked-by: Peter Zijlstra Signed-off-by: Olivier Langlois Signed-off-by: KOSAKI Motohiro Signed-off-by: Frederic Weisbecker --- kernel/sched/stats.h | 39 ++++++++++++++++++++++++++++++++++++--- 1 file changed, 36 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/stats.h b/kernel/sched/stats.h index 2ef90a51ec5e..71bac979d5ee 100644 --- a/kernel/sched/stats.h +++ b/kernel/sched/stats.h @@ -161,6 +161,39 @@ sched_info_switch(struct task_struct *prev, struct task_struct *next) * on CONFIG_SCHEDSTATS. */ +/** + * cputimer_running - return true if cputimer is running + * + * @tsk: Pointer to target task. + */ +static inline bool cputimer_running(struct task_struct *tsk) + +{ + struct thread_group_cputimer *cputimer = &tsk->signal->cputimer; + + if (!cputimer->running) + return false; + + /* + * After we flush the task's sum_exec_runtime to sig->sum_sched_runtime + * in __exit_signal(), we won't account to the signal struct further + * cputime consumed by that task, even though the task can still be + * ticking after __exit_signal(). + * + * In order to keep a consistent behaviour between thread group cputime + * and thread group cputimer accounting, lets also ignore the cputime + * elapsing after __exit_signal() in any thread group timer running. + * + * This makes sure that POSIX CPU clocks and timers are synchronized, so + * that a POSIX CPU timer won't expire while the corresponding POSIX CPU + * clock delta is behind the expiring timer value. + */ + if (unlikely(!tsk->sighand)) + return false; + + return true; +} + /** * account_group_user_time - Maintain utime for a thread group. * @@ -176,7 +209,7 @@ static inline void account_group_user_time(struct task_struct *tsk, { struct thread_group_cputimer *cputimer = &tsk->signal->cputimer; - if (!cputimer->running) + if (!cputimer_running(tsk)) return; raw_spin_lock(&cputimer->lock); @@ -199,7 +232,7 @@ static inline void account_group_system_time(struct task_struct *tsk, { struct thread_group_cputimer *cputimer = &tsk->signal->cputimer; - if (!cputimer->running) + if (!cputimer_running(tsk)) return; raw_spin_lock(&cputimer->lock); @@ -222,7 +255,7 @@ static inline void account_group_exec_runtime(struct task_struct *tsk, { struct thread_group_cputimer *cputimer = &tsk->signal->cputimer; - if (!cputimer->running) + if (!cputimer_running(tsk)) return; raw_spin_lock(&cputimer->lock); -- cgit v1.2.3 From e5302920da9ef23f9d19d4e9ac85704cc25bee7a Mon Sep 17 00:00:00 2001 From: Stephane Eranian Date: Fri, 5 Jul 2013 00:30:11 +0200 Subject: perf: Fix interrupt handler timing harness This patch fixes a serious bug in: 14c63f17b1fd perf: Drop sample rate when sampling is too slow There was an misunderstanding on the API of the do_div() macro. It returns the remainder of the division and this was not what the function expected leading to disabling the interrupt latency watchdog. This patch also remove a duplicate assignment in perf_sample_event_took(). Signed-off-by: Stephane Eranian Cc: peterz@infradead.org Cc: dave.hansen@linux.intel.com Cc: ak@linux.intel.com Cc: jolsa@redhat.com Link: http://lkml.kernel.org/r/20130704223010.GA30625@quad Signed-off-by: Ingo Molnar --- kernel/events/core.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 1db3af933704..1833bc5a84a7 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -182,7 +182,7 @@ void update_perf_cpu_limits(void) u64 tmp = perf_sample_period_ns; tmp *= sysctl_perf_cpu_time_max_percent; - tmp = do_div(tmp, 100); + do_div(tmp, 100); atomic_set(&perf_sample_allowed_ns, tmp); } @@ -232,7 +232,7 @@ DEFINE_PER_CPU(u64, running_sample_length); void perf_sample_event_took(u64 sample_len_ns) { u64 avg_local_sample_len; - u64 local_samples_len = __get_cpu_var(running_sample_length); + u64 local_samples_len; if (atomic_read(&perf_sample_allowed_ns) == 0) return; -- cgit v1.2.3 From 332962f2c88868ed3cdab466870baaa34dd58612 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 4 Jul 2013 22:46:45 +0200 Subject: clocksource: Reselect clocksource when watchdog validated high-res capability Up to commit 5d33b883a (clocksource: Always verify highres capability) we had no sanity check when selecting a clocksource, which prevented that a non highres capable clocksource is used when the system already switched to highres/nohz mode. The new sanity check works as Alex and Tim found out. It prevents the TSC from being used. This happens because on x86 the boot process looks like this: tsc_start_freqency_validation(TSC); clocksource_register(HPET); clocksource_done_booting(); clocksource_select() Selects HPET which is valid for high-res switch_to_highres(); clocksource_register(TSC); TSC is not selected, because it is not yet flagged as VALID_HIGH_RES clocksource_watchdog() Validates TSC for highres, but that does not make TSC the current clocksource. Before the sanity check was added, we installed TSC unvalidated which worked most of the time. If the TSC was really detected as unstable, then the unstable logic removed it and installed HPET again. The sanity check is correct and needed. So the watchdog needs to kick a reselection of the clocksource, when it qualifies TSC as a valid high res clocksource. To solve this, we mark the clocksource which got the flag CLOCK_SOURCE_VALID_FOR_HRES set by the watchdog with an new flag CLOCK_SOURCE_RESELECT and trigger the watchdog thread. The watchdog thread evaluates the flag and invokes clocksource_select() when set. To avoid that the clocksource_done_booting() code, which is about to install the first real clocksource anyway, needs to go through clocksource_select and tick_oneshot_notify() pointlessly, split out the clocksource_watchdog_kthread() list walk code and invoke the select/notify only when called from clocksource_watchdog_kthread(). So clocksource_done_booting() can utilize the same splitout code without the select/notify invocation and the clocksource_mutex unlock/relock dance. Reported-and-tested-by: Alex Shi Cc: Hans Peter Anvin Cc: Tim Chen Cc: Andi Kleen Tested-by: Peter Zijlstra Cc: Ingo Molnar Cc: Davidlohr Bueso Cc: John Stultz Link: http://lkml.kernel.org/r/alpine.DEB.2.02.1307042239150.11637@ionos.tec.linutronix.de Signed-off-by: Thomas Gleixner --- kernel/time/clocksource.c | 57 ++++++++++++++++++++++++++++++++++------------- 1 file changed, 42 insertions(+), 15 deletions(-) (limited to 'kernel') diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c index e713ef7d19a7..50a8736757f3 100644 --- a/kernel/time/clocksource.c +++ b/kernel/time/clocksource.c @@ -181,6 +181,7 @@ static int finished_booting; #ifdef CONFIG_CLOCKSOURCE_WATCHDOG static void clocksource_watchdog_work(struct work_struct *work); +static void clocksource_select(void); static LIST_HEAD(watchdog_list); static struct clocksource *watchdog; @@ -301,13 +302,30 @@ static void clocksource_watchdog(unsigned long data) if (!(cs->flags & CLOCK_SOURCE_VALID_FOR_HRES) && (cs->flags & CLOCK_SOURCE_IS_CONTINUOUS) && (watchdog->flags & CLOCK_SOURCE_IS_CONTINUOUS)) { + /* Mark it valid for high-res. */ cs->flags |= CLOCK_SOURCE_VALID_FOR_HRES; + + /* + * clocksource_done_booting() will sort it if + * finished_booting is not set yet. + */ + if (!finished_booting) + continue; + /* - * We just marked the clocksource as highres-capable, - * notify the rest of the system as well so that we - * transition into high-res mode: + * If this is not the current clocksource let + * the watchdog thread reselect it. Due to the + * change to high res this clocksource might + * be preferred now. If it is the current + * clocksource let the tick code know about + * that change. */ - tick_clock_notify(); + if (cs != curr_clocksource) { + cs->flags |= CLOCK_SOURCE_RESELECT; + schedule_work(&watchdog_work); + } else { + tick_clock_notify(); + } } } @@ -404,19 +422,25 @@ static void clocksource_dequeue_watchdog(struct clocksource *cs) spin_unlock_irqrestore(&watchdog_lock, flags); } -static int clocksource_watchdog_kthread(void *data) +static int __clocksource_watchdog_kthread(void) { struct clocksource *cs, *tmp; unsigned long flags; LIST_HEAD(unstable); + int select = 0; - mutex_lock(&clocksource_mutex); spin_lock_irqsave(&watchdog_lock, flags); - list_for_each_entry_safe(cs, tmp, &watchdog_list, wd_list) + list_for_each_entry_safe(cs, tmp, &watchdog_list, wd_list) { if (cs->flags & CLOCK_SOURCE_UNSTABLE) { list_del_init(&cs->wd_list); list_add(&cs->wd_list, &unstable); + select = 1; + } + if (cs->flags & CLOCK_SOURCE_RESELECT) { + cs->flags &= ~CLOCK_SOURCE_RESELECT; + select = 1; } + } /* Check if the watchdog timer needs to be stopped. */ clocksource_stop_watchdog(); spin_unlock_irqrestore(&watchdog_lock, flags); @@ -426,6 +450,14 @@ static int clocksource_watchdog_kthread(void *data) list_del_init(&cs->wd_list); __clocksource_change_rating(cs, 0); } + return select; +} + +static int clocksource_watchdog_kthread(void *data) +{ + mutex_lock(&clocksource_mutex); + if (__clocksource_watchdog_kthread()) + clocksource_select(); mutex_unlock(&clocksource_mutex); return 0; } @@ -445,7 +477,7 @@ static void clocksource_enqueue_watchdog(struct clocksource *cs) static inline void clocksource_dequeue_watchdog(struct clocksource *cs) { } static inline void clocksource_resume_watchdog(void) { } -static inline int clocksource_watchdog_kthread(void *data) { return 0; } +static inline int __clocksource_watchdog_kthread(void) { return 0; } static bool clocksource_is_watchdog(struct clocksource *cs) { return false; } #endif /* CONFIG_CLOCKSOURCE_WATCHDOG */ @@ -647,16 +679,11 @@ static int __init clocksource_done_booting(void) { mutex_lock(&clocksource_mutex); curr_clocksource = clocksource_default_clock(); - mutex_unlock(&clocksource_mutex); - finished_booting = 1; - /* * Run the watchdog first to eliminate unstable clock sources */ - clocksource_watchdog_kthread(NULL); - - mutex_lock(&clocksource_mutex); + __clocksource_watchdog_kthread(); clocksource_select(); mutex_unlock(&clocksource_mutex); return 0; @@ -789,7 +816,6 @@ static void __clocksource_change_rating(struct clocksource *cs, int rating) list_del(&cs->list); cs->rating = rating; clocksource_enqueue(cs); - clocksource_select(); } /** @@ -801,6 +827,7 @@ void clocksource_change_rating(struct clocksource *cs, int rating) { mutex_lock(&clocksource_mutex); __clocksource_change_rating(cs, rating); + clocksource_select(); mutex_unlock(&clocksource_mutex); } EXPORT_SYMBOL(clocksource_change_rating); -- cgit v1.2.3 From 002fca5df168922103a2bb52748f9984e6de80b2 Mon Sep 17 00:00:00 2001 From: Axel Lin Date: Fri, 5 Jul 2013 17:13:12 +0800 Subject: genirq: generic chip: Use DIV_ROUND_UP to calculate numchips The number of interrupts in a domain may be not divisible by the number of interrupts each chip handles. Integer division may truncate the result, thus use DIV_ROUND_UP to count numchips. Seems all users of irq_alloc_domain_generic_chips() in current code do not have this issue. I just found the issue while reading the code. Signed-off-by: Axel Lin Cc: Grant Likely Cc: Tony Lindgren Cc: Arnd Bergmann Link: http://lkml.kernel.org/r/1373015592.18252.2.camel@phoenix Signed-off-by: Thomas Gleixner --- kernel/irq/generic-chip.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/irq/generic-chip.c b/kernel/irq/generic-chip.c index 1c39eccc1eaf..2f274f30b7e2 100644 --- a/kernel/irq/generic-chip.c +++ b/kernel/irq/generic-chip.c @@ -278,7 +278,7 @@ int irq_alloc_domain_generic_chips(struct irq_domain *d, int irqs_per_chip, if (d->revmap_type != IRQ_DOMAIN_MAP_LINEAR) return -EINVAL; - numchips = d->revmap_data.linear.size / irqs_per_chip; + numchips = DIV_ROUND_UP(d->revmap_data.linear.size, irqs_per_chip); if (!numchips) return -EINVAL; -- cgit v1.2.3 From 5ec2481b7b47a4005bb446d176e5d0257400c77d Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 5 Jul 2013 12:09:18 +0200 Subject: hrtimers: Move SMP function call to thread context smp_call_function_* must not be called from softirq context. But clock_was_set() which calls on_each_cpu() is called from softirq context to implement a delayed clock_was_set() for the timer interrupt handler. Though that almost never gets invoked. A recent change in the resume code uses the softirq based delayed clock_was_set to support Xens resume mechanism. linux-next contains a new warning which warns if smp_call_function_* is called from softirq context which gets triggered by that Xen change. Fix this by moving the delayed clock_was_set() call to a work context. Reported-and-tested-by: Artem Savkov Reported-by: Sasha Levin Cc: David Vrabel Cc: Ingo Molnar Cc: H. Peter Anvin , Cc: Konrad Wilk Cc: John Stultz Cc: xen-devel@lists.xen.org Cc: stable@vger.kernel.org Signed-off-by: Thomas Gleixner --- kernel/hrtimer.c | 37 +++++++++++++++---------------------- 1 file changed, 15 insertions(+), 22 deletions(-) (limited to 'kernel') diff --git a/kernel/hrtimer.c b/kernel/hrtimer.c index e86827e94c9a..b9b9420a1297 100644 --- a/kernel/hrtimer.c +++ b/kernel/hrtimer.c @@ -721,17 +721,20 @@ static int hrtimer_switch_to_hres(void) return 1; } +static void clock_was_set_work(struct work_struct *work) +{ + clock_was_set(); +} + +static DECLARE_WORK(hrtimer_work, clock_was_set_work); + /* - * Called from timekeeping code to reprogramm the hrtimer interrupt - * device. If called from the timer interrupt context we defer it to - * softirq context. + * Called from timekeeping and resume code to reprogramm the hrtimer + * interrupt device on all cpus. */ void clock_was_set_delayed(void) { - struct hrtimer_cpu_base *cpu_base = &__get_cpu_var(hrtimer_bases); - - cpu_base->clock_was_set = 1; - __raise_softirq_irqoff(HRTIMER_SOFTIRQ); + schedule_work(&hrtimer_work); } #else @@ -775,12 +778,7 @@ void clock_was_set(void) * During resume we might have to reprogram the high resolution timer * interrupt on all online CPUs. However, all other CPUs will be * stopped with IRQs interrupts disabled so the clock_was_set() call - * must be deferred to the softirq. - * - * The one-shot timer has already been programmed to fire immediately - * (see tick_resume_oneshot()) and this interrupt will trigger the - * softirq to run early enough to correctly reprogram the timers on - * all CPUs. + * must be deferred. */ void hrtimers_resume(void) { @@ -789,8 +787,10 @@ void hrtimers_resume(void) WARN_ONCE(!irqs_disabled(), KERN_INFO "hrtimers_resume() called with IRQs enabled!"); - cpu_base->clock_was_set = 1; - __raise_softirq_irqoff(HRTIMER_SOFTIRQ); + /* Retrigger on the local CPU */ + retrigger_next_event(NULL); + /* And schedule a retrigger for all others */ + clock_was_set_delayed(); } static inline void timer_stats_hrtimer_set_start_info(struct hrtimer *timer) @@ -1441,13 +1441,6 @@ void hrtimer_peek_ahead_timers(void) static void run_hrtimer_softirq(struct softirq_action *h) { - struct hrtimer_cpu_base *cpu_base = &__get_cpu_var(hrtimer_bases); - - if (cpu_base->clock_was_set) { - cpu_base->clock_was_set = 0; - clock_was_set(); - } - hrtimer_peek_ahead_timers(); } -- cgit v1.2.3 From 73b0cd674ccc64c921e25bd7154f26d342116539 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Sat, 6 Jul 2013 10:34:00 +0200 Subject: hrtimer: Remove unused variable Sigh, should have noticed myself. Reported-by: fengguang.wu@intel.com Signed-off-by: Thomas Gleixner --- kernel/hrtimer.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'kernel') diff --git a/kernel/hrtimer.c b/kernel/hrtimer.c index b9b9420a1297..3a951d8d5770 100644 --- a/kernel/hrtimer.c +++ b/kernel/hrtimer.c @@ -782,8 +782,6 @@ void clock_was_set(void) */ void hrtimers_resume(void) { - struct hrtimer_cpu_base *cpu_base = &__get_cpu_var(hrtimer_bases); - WARN_ONCE(!irqs_disabled(), KERN_INFO "hrtimers_resume() called with IRQs enabled!"); -- cgit v1.2.3 From 79f6530cb59e2a0af6953742a33cc29e98ca631c Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Mon, 8 Jul 2013 15:59:36 -0700 Subject: audit: fix mq_open and mq_unlink to add the MQ root as a hidden parent audit_names record The old audit PATH records for mq_open looked like this: type=PATH msg=audit(1366282323.982:869): item=1 name=(null) inode=6777 dev=00:0c mode=041777 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282323.982:869): item=0 name="test_mq" inode=26732 dev=00:0c mode=0100700 ouid=0 ogid=0 rdev=00:00 obj=staff_u:object_r:user_tmpfs_t:s15:c0.c1023 ...with the audit related changes that went into 3.7, they now look like this: type=PATH msg=audit(1366282236.776:3606): item=2 name=(null) inode=66655 dev=00:0c mode=0100700 ouid=0 ogid=0 rdev=00:00 obj=staff_u:object_r:user_tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282236.776:3606): item=1 name=(null) inode=6926 dev=00:0c mode=041777 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282236.776:3606): item=0 name="test_mq" Both of these look wrong to me. As Steve Grubb pointed out: "What we need is 1 PATH record that identifies the MQ. The other PATH records probably should not be there." Fix it to record the mq root as a parent, and flag it such that it should be hidden from view when the names are logged, since the root of the mq filesystem isn't terribly interesting. With this change, we get a single PATH record that looks more like this: type=PATH msg=audit(1368021604.836:484): item=0 name="test_mq" inode=16914 dev=00:0c mode=0100644 ouid=0 ogid=0 rdev=00:00 obj=unconfined_u:object_r:user_tmpfs_t:s0 In order to do this, a new audit_inode_parent_hidden() function is added. If we do it this way, then we avoid having the existing callers of audit_inode needing to do any sort of flag conversion if auditing is inactive. Signed-off-by: Jeff Layton Reported-by: Jiri Jaburek Cc: Steve Grubb Cc: Eric Paris Cc: Al Viro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/audit.h | 1 + kernel/auditsc.c | 12 +++++++++--- 2 files changed, 10 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/audit.h b/kernel/audit.h index 1c95131ef760..123c9b7c3979 100644 --- a/kernel/audit.h +++ b/kernel/audit.h @@ -85,6 +85,7 @@ struct audit_names { struct filename *name; int name_len; /* number of chars to log */ + bool hidden; /* don't log this record */ bool name_put; /* call __putname()? */ unsigned long ino; diff --git a/kernel/auditsc.c b/kernel/auditsc.c index 3c8a601324a2..9845cb32b60a 100644 --- a/kernel/auditsc.c +++ b/kernel/auditsc.c @@ -1399,8 +1399,11 @@ static void audit_log_exit(struct audit_context *context, struct task_struct *ts } i = 0; - list_for_each_entry(n, &context->names_list, list) + list_for_each_entry(n, &context->names_list, list) { + if (n->hidden) + continue; audit_log_name(context, n, NULL, i++, &call_panic); + } /* Send end of event record to help user space know we are finished */ ab = audit_log_start(context, GFP_KERNEL, AUDIT_EOE); @@ -1769,14 +1772,15 @@ void audit_putname(struct filename *name) * __audit_inode - store the inode and device from a lookup * @name: name being audited * @dentry: dentry being audited - * @parent: does this dentry represent the parent? + * @flags: attributes for this particular entry */ void __audit_inode(struct filename *name, const struct dentry *dentry, - unsigned int parent) + unsigned int flags) { struct audit_context *context = current->audit_context; const struct inode *inode = dentry->d_inode; struct audit_names *n; + bool parent = flags & AUDIT_INODE_PARENT; if (!context->in_syscall) return; @@ -1831,6 +1835,8 @@ out: if (parent) { n->name_len = n->name ? parent_len(n->name->name) : AUDIT_NAME_FULL; n->type = AUDIT_TYPE_PARENT; + if (flags & AUDIT_INODE_HIDDEN) + n->hidden = true; } else { n->name_len = AUDIT_NAME_FULL; n->type = AUDIT_TYPE_NORMAL; -- cgit v1.2.3 From 6beb8a23b50d38a003e80c5f16b50c56e8ae3387 Mon Sep 17 00:00:00 2001 From: "Raphael S. Carvalho" Date: Mon, 8 Jul 2013 15:59:37 -0700 Subject: kernel/auditfilter.c: fixing build warning kernel/auditfilter.c:426: warning: this decimal constant is unsigned only in ISO C90 Signed-off-by: Raphael S. Carvalho Cc: Eric Paris Cc: Al Viro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/auditfilter.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/auditfilter.c b/kernel/auditfilter.c index 6bd4a90d1991..0ee9eff866d6 100644 --- a/kernel/auditfilter.c +++ b/kernel/auditfilter.c @@ -423,7 +423,7 @@ static struct audit_entry *audit_data_to_entry(struct audit_rule_data *data, f->lsm_rule = NULL; /* Support legacy tests for a valid loginuid */ - if ((f->type == AUDIT_LOGINUID) && (f->val == 4294967295)) { + if ((f->type == AUDIT_LOGINUID) && (f->val == 4294967295U)) { f->type = AUDIT_LOGINUID_SET; f->val = 0; } -- cgit v1.2.3 From 2f992ee85aaa7dfd2bda43efe4493af1e108d054 Mon Sep 17 00:00:00 2001 From: Chen Gang Date: Mon, 8 Jul 2013 15:59:38 -0700 Subject: kernel/auditfilter.c: fix leak in audit_add_rule() error path If both 'tree' and 'watch' are valid we must call audit_put_tree(), just like the preceding code within audit_add_rule(). Signed-off-by: Chen Gang Cc: Al Viro Cc: Eric Paris Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/auditfilter.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'kernel') diff --git a/kernel/auditfilter.c b/kernel/auditfilter.c index 0ee9eff866d6..3d15c66b7f0b 100644 --- a/kernel/auditfilter.c +++ b/kernel/auditfilter.c @@ -865,6 +865,12 @@ static inline int audit_add_rule(struct audit_entry *entry) err = audit_add_watch(&entry->rule, &list); if (err) { mutex_unlock(&audit_filter_mutex); + /* + * normally audit_add_tree_rule() will free it + * on failure + */ + if (tree) + audit_put_tree(tree); goto error; } } -- cgit v1.2.3 From b9ce54c9f59894e787e3067d2f758c297fcd6fd0 Mon Sep 17 00:00:00 2001 From: Michal Simek Date: Mon, 8 Jul 2013 15:59:39 -0700 Subject: audit: Fix decimal constant description Use proper decimal type for comparison with u32. Compilation warning was introduced by 780a7654 ("audit: Make testing for a valid loginuid explicit.") kernel/auditfilter.c: In function 'audit_data_to_entry': kernel/auditfilter.c:426:3: warning: this decimal constant is unsigned only in ISO C90 [enabled by default] if ((f->type == AUDIT_LOGINUID) && (f->val == 4294967295)) { Signed-off-by: Michal Simek Cc: Al Viro Cc: Eric Paris Acked-by: Geert Uytterhoeven Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/auditfilter.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/auditfilter.c b/kernel/auditfilter.c index 3d15c66b7f0b..f7aee8be7fb2 100644 --- a/kernel/auditfilter.c +++ b/kernel/auditfilter.c @@ -423,7 +423,7 @@ static struct audit_entry *audit_data_to_entry(struct audit_rule_data *data, f->lsm_rule = NULL; /* Support legacy tests for a valid loginuid */ - if ((f->type == AUDIT_LOGINUID) && (f->val == 4294967295U)) { + if ((f->type == AUDIT_LOGINUID) && (f->val == ~0U)) { f->type = AUDIT_LOGINUID_SET; f->val = 0; } -- cgit v1.2.3 From dcb6b45254e2281b6f99ea7f2d51343954aa3ba8 Mon Sep 17 00:00:00 2001 From: Alex Thorlton Date: Mon, 8 Jul 2013 16:00:42 -0700 Subject: panic: add cpu/pid to warn_slowpath_common in WARNING printk()s Add the cpu/pid that called WARN() so that the stack traces can be matched up with the WARNING messages. [akpm@linux-foundation.org: remove stray quote] Signed-off-by: Alex Thorlton Reviewed-by: Robin Holt Cc: Stephen Boyd Cc: Vikram Mulukutla Cc: Rusty Russell Cc: Tejun Heo Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/panic.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/panic.c b/kernel/panic.c index 167ec097ce8b..97712319f128 100644 --- a/kernel/panic.c +++ b/kernel/panic.c @@ -399,8 +399,9 @@ struct slowpath_args { static void warn_slowpath_common(const char *file, int line, void *caller, unsigned taint, struct slowpath_args *args) { - printk(KERN_WARNING "------------[ cut here ]------------\n"); - printk(KERN_WARNING "WARNING: at %s:%d %pS()\n", file, line, caller); + pr_warn("------------[ cut here ]------------\n"); + pr_warn("WARNING: CPU: %d PID: %d at %s:%d %pS()\n", + raw_smp_processor_id(), current->pid, file, line, caller); if (args) vprintk(args->fmt, args->args); -- cgit v1.2.3 From 7c8df28633bf0b7eb253f866029be0ac59ddb062 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Mon, 8 Jul 2013 16:00:54 -0700 Subject: ptrace: revert "Prepare to fix racy accesses on task breakpoints" This reverts commit bf26c018490c ("Prepare to fix racy accesses on task breakpoints"). The patch was fine but we can no longer race with SIGKILL after commit 9899d11f6544 ("ptrace: ensure arch_ptrace/ptrace_request can never race with SIGKILL"), the __TASK_TRACED tracee can't be woken up and ->ptrace_bps[] can't go away. Now that ptrace_get_breakpoints/ptrace_put_breakpoints have no callers, we can kill them and remove task->ptrace_bp_refcnt. Signed-off-by: Oleg Nesterov Acked-by: Frederic Weisbecker Acked-by: Michael Neuling Cc: Benjamin Herrenschmidt Cc: Ingo Molnar Cc: Jan Kratochvil Cc: Paul Mackerras Cc: Paul Mundt Cc: Will Deacon Cc: Prasad Cc: Russell King Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 2 +- kernel/ptrace.c | 16 ---------------- 2 files changed, 1 insertion(+), 17 deletions(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index fafe75d9e6f6..a949819055d5 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -808,7 +808,7 @@ void do_exit(long code) /* * FIXME: do that only when needed, using sched_exit tracepoint */ - ptrace_put_breakpoints(tsk); + flush_ptrace_hw_breakpoint(tsk); exit_notify(tsk, group_dead); #ifdef CONFIG_NUMA diff --git a/kernel/ptrace.c b/kernel/ptrace.c index ba5e6cea181a..a146ee327f6a 100644 --- a/kernel/ptrace.c +++ b/kernel/ptrace.c @@ -1221,19 +1221,3 @@ asmlinkage long compat_sys_ptrace(compat_long_t request, compat_long_t pid, return ret; } #endif /* CONFIG_COMPAT */ - -#ifdef CONFIG_HAVE_HW_BREAKPOINT -int ptrace_get_breakpoints(struct task_struct *tsk) -{ - if (atomic_inc_not_zero(&tsk->ptrace_bp_refcnt)) - return 0; - - return -1; -} - -void ptrace_put_breakpoints(struct task_struct *tsk) -{ - if (atomic_dec_and_test(&tsk->ptrace_bp_refcnt)) - flush_ptrace_hw_breakpoint(tsk); -} -#endif /* CONFIG_HAVE_HW_BREAKPOINT */ -- cgit v1.2.3 From fab840fc2d542fabcab903db8e03589a6702ba5f Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Mon, 8 Jul 2013 16:01:05 -0700 Subject: ptrace: PTRACE_DETACH should do flush_ptrace_hw_breakpoint(child) Change ptrace_detach() to call flush_ptrace_hw_breakpoint(child). This frees the slots for non-ptrace PERF_TYPE_BREAKPOINT users, and this ensures that the tracee won't be killed by SIGTRAP triggered by the active breakpoints. Test-case: unsigned long encode_dr7(int drnum, int enable, unsigned int type, unsigned int len) { unsigned long dr7; dr7 = ((len | type) & 0xf) << (DR_CONTROL_SHIFT + drnum * DR_CONTROL_SIZE); if (enable) dr7 |= (DR_GLOBAL_ENABLE << (drnum * DR_ENABLE_SIZE)); return dr7; } int write_dr(int pid, int dr, unsigned long val) { return ptrace(PTRACE_POKEUSER, pid, offsetof (struct user, u_debugreg[dr]), val); } void func(void) { } int main(void) { int pid, stat; unsigned long dr7; pid = fork(); if (!pid) { assert(ptrace(PTRACE_TRACEME, 0,0,0) == 0); kill(getpid(), SIGHUP); func(); return 0x13; } assert(pid == waitpid(-1, &stat, 0)); assert(WSTOPSIG(stat) == SIGHUP); assert(write_dr(pid, 0, (long)func) == 0); dr7 = encode_dr7(0, 1, DR_RW_EXECUTE, DR_LEN_1); assert(write_dr(pid, 7, dr7) == 0); assert(ptrace(PTRACE_DETACH, pid, 0,0) == 0); assert(pid == waitpid(-1, &stat, 0)); assert(stat == 0x1300); return 0; } Before this patch the child is killed after PTRACE_DETACH. Signed-off-by: Oleg Nesterov Acked-by: Frederic Weisbecker Cc: Benjamin Herrenschmidt Cc: Ingo Molnar Cc: Jan Kratochvil Cc: Michael Neuling Cc: Paul Mackerras Cc: Paul Mundt Cc: Will Deacon Cc: Prasad Cc: Russell King Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/ptrace.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/ptrace.c b/kernel/ptrace.c index a146ee327f6a..4041f5747e73 100644 --- a/kernel/ptrace.c +++ b/kernel/ptrace.c @@ -469,6 +469,7 @@ static int ptrace_detach(struct task_struct *child, unsigned int data) /* Architecture-specific hardware disable .. */ ptrace_disable(child); clear_tsk_thread_flag(child, TIF_SYSCALL_TRACE); + flush_ptrace_hw_breakpoint(child); write_lock_irq(&tasklist_lock); /* -- cgit v1.2.3 From 0efbee70890c992f31a7b294ac654ff6c62d51c5 Mon Sep 17 00:00:00 2001 From: Robin Holt Date: Mon, 8 Jul 2013 16:01:31 -0700 Subject: reboot: remove -stable friendly PF_THREAD_BOUND define Remove the prior patch's #define for easier backporting to the stable releases. Signed-off-by: Robin Holt Cc: H. Peter Anvin Cc: Russ Anderson Cc: Robin Holt Cc: Russell King Cc: Guan Xuetao Cc: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sys.c | 5 ----- 1 file changed, 5 deletions(-) (limited to 'kernel') diff --git a/kernel/sys.c b/kernel/sys.c index 071de900c824..b882440bd0c0 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -362,11 +362,6 @@ int unregister_reboot_notifier(struct notifier_block *nb) } EXPORT_SYMBOL(unregister_reboot_notifier); -/* Add backwards compatibility for stable trees. */ -#ifndef PF_NO_SETAFFINITY -#define PF_NO_SETAFFINITY PF_THREAD_BOUND -#endif - static void migrate_to_reboot_cpu(void) { /* The boot cpu is always logical cpu 0 */ -- cgit v1.2.3 From 15d94b82565ebfb0cf27830b96e6cf5ed2d12a9a Mon Sep 17 00:00:00 2001 From: Robin Holt Date: Mon, 8 Jul 2013 16:01:32 -0700 Subject: reboot: move shutdown/reboot related functions to kernel/reboot.c This patch is preparatory. It moves reboot related syscall, etc functions from kernel/sys.c to kernel/reboot.c. Signed-off-by: Robin Holt Cc: H. Peter Anvin Cc: Russ Anderson Cc: Robin Holt Cc: Russell King Cc: Guan Xuetao Cc: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/Makefile | 2 +- kernel/reboot.c | 346 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ kernel/sys.c | 331 ----------------------------------------------------- 3 files changed, 347 insertions(+), 332 deletions(-) create mode 100644 kernel/reboot.c (limited to 'kernel') diff --git a/kernel/Makefile b/kernel/Makefile index 271fd3119af9..470839d1a30e 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -9,7 +9,7 @@ obj-y = fork.o exec_domain.o panic.o printk.o \ rcupdate.o extable.o params.o posix-timers.o \ kthread.o wait.o sys_ni.o posix-cpu-timers.o mutex.o \ hrtimer.o rwsem.o nsproxy.o srcu.o semaphore.o \ - notifier.o ksysfs.o cred.o \ + notifier.o ksysfs.o cred.o reboot.o \ async.o range.o groups.o lglock.o smpboot.o ifdef CONFIG_FUNCTION_TRACER diff --git a/kernel/reboot.c b/kernel/reboot.c new file mode 100644 index 000000000000..37d2636a65c2 --- /dev/null +++ b/kernel/reboot.c @@ -0,0 +1,346 @@ +/* + * linux/kernel/reboot.c + * + * Copyright (C) 2013 Linus Torvalds + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include + +/* + * this indicates whether you can reboot with ctrl-alt-del: the default is yes + */ + +int C_A_D = 1; +struct pid *cad_pid; +EXPORT_SYMBOL(cad_pid); + +/* + * If set, this is used for preparing the system to power off. + */ + +void (*pm_power_off_prepare)(void); + +/** + * emergency_restart - reboot the system + * + * Without shutting down any hardware or taking any locks + * reboot the system. This is called when we know we are in + * trouble so this is our best effort to reboot. This is + * safe to call in interrupt context. + */ +void emergency_restart(void) +{ + kmsg_dump(KMSG_DUMP_EMERG); + machine_emergency_restart(); +} +EXPORT_SYMBOL_GPL(emergency_restart); + +void kernel_restart_prepare(char *cmd) +{ + blocking_notifier_call_chain(&reboot_notifier_list, SYS_RESTART, cmd); + system_state = SYSTEM_RESTART; + usermodehelper_disable(); + device_shutdown(); +} + +/** + * register_reboot_notifier - Register function to be called at reboot time + * @nb: Info about notifier function to be called + * + * Registers a function with the list of functions + * to be called at reboot time. + * + * Currently always returns zero, as blocking_notifier_chain_register() + * always returns zero. + */ +int register_reboot_notifier(struct notifier_block *nb) +{ + return blocking_notifier_chain_register(&reboot_notifier_list, nb); +} +EXPORT_SYMBOL(register_reboot_notifier); + +/** + * unregister_reboot_notifier - Unregister previously registered reboot notifier + * @nb: Hook to be unregistered + * + * Unregisters a previously registered reboot + * notifier function. + * + * Returns zero on success, or %-ENOENT on failure. + */ +int unregister_reboot_notifier(struct notifier_block *nb) +{ + return blocking_notifier_chain_unregister(&reboot_notifier_list, nb); +} +EXPORT_SYMBOL(unregister_reboot_notifier); + +static void migrate_to_reboot_cpu(void) +{ + /* The boot cpu is always logical cpu 0 */ + int cpu = 0; + + cpu_hotplug_disable(); + + /* Make certain the cpu I'm about to reboot on is online */ + if (!cpu_online(cpu)) + cpu = cpumask_first(cpu_online_mask); + + /* Prevent races with other tasks migrating this task */ + current->flags |= PF_NO_SETAFFINITY; + + /* Make certain I only run on the appropriate processor */ + set_cpus_allowed_ptr(current, cpumask_of(cpu)); +} + +/** + * kernel_restart - reboot the system + * @cmd: pointer to buffer containing command to execute for restart + * or %NULL + * + * Shutdown everything and perform a clean reboot. + * This is not safe to call in interrupt context. + */ +void kernel_restart(char *cmd) +{ + kernel_restart_prepare(cmd); + migrate_to_reboot_cpu(); + syscore_shutdown(); + if (!cmd) + printk(KERN_EMERG "Restarting system.\n"); + else + printk(KERN_EMERG "Restarting system with command '%s'.\n", cmd); + kmsg_dump(KMSG_DUMP_RESTART); + machine_restart(cmd); +} +EXPORT_SYMBOL_GPL(kernel_restart); + +static void kernel_shutdown_prepare(enum system_states state) +{ + blocking_notifier_call_chain(&reboot_notifier_list, + (state == SYSTEM_HALT)?SYS_HALT:SYS_POWER_OFF, NULL); + system_state = state; + usermodehelper_disable(); + device_shutdown(); +} +/** + * kernel_halt - halt the system + * + * Shutdown everything and perform a clean system halt. + */ +void kernel_halt(void) +{ + kernel_shutdown_prepare(SYSTEM_HALT); + migrate_to_reboot_cpu(); + syscore_shutdown(); + printk(KERN_EMERG "System halted.\n"); + kmsg_dump(KMSG_DUMP_HALT); + machine_halt(); +} + +EXPORT_SYMBOL_GPL(kernel_halt); + +/** + * kernel_power_off - power_off the system + * + * Shutdown everything and perform a clean system power_off. + */ +void kernel_power_off(void) +{ + kernel_shutdown_prepare(SYSTEM_POWER_OFF); + if (pm_power_off_prepare) + pm_power_off_prepare(); + migrate_to_reboot_cpu(); + syscore_shutdown(); + printk(KERN_EMERG "Power down.\n"); + kmsg_dump(KMSG_DUMP_POWEROFF); + machine_power_off(); +} +EXPORT_SYMBOL_GPL(kernel_power_off); + +static DEFINE_MUTEX(reboot_mutex); + +/* + * Reboot system call: for obvious reasons only root may call it, + * and even root needs to set up some magic numbers in the registers + * so that some mistake won't make this reboot the whole machine. + * You can also set the meaning of the ctrl-alt-del-key here. + * + * reboot doesn't sync: do that yourself before calling this. + */ +SYSCALL_DEFINE4(reboot, int, magic1, int, magic2, unsigned int, cmd, + void __user *, arg) +{ + struct pid_namespace *pid_ns = task_active_pid_ns(current); + char buffer[256]; + int ret = 0; + + /* We only trust the superuser with rebooting the system. */ + if (!ns_capable(pid_ns->user_ns, CAP_SYS_BOOT)) + return -EPERM; + + /* For safety, we require "magic" arguments. */ + if (magic1 != LINUX_REBOOT_MAGIC1 || + (magic2 != LINUX_REBOOT_MAGIC2 && + magic2 != LINUX_REBOOT_MAGIC2A && + magic2 != LINUX_REBOOT_MAGIC2B && + magic2 != LINUX_REBOOT_MAGIC2C)) + return -EINVAL; + + /* + * If pid namespaces are enabled and the current task is in a child + * pid_namespace, the command is handled by reboot_pid_ns() which will + * call do_exit(). + */ + ret = reboot_pid_ns(pid_ns, cmd); + if (ret) + return ret; + + /* Instead of trying to make the power_off code look like + * halt when pm_power_off is not set do it the easy way. + */ + if ((cmd == LINUX_REBOOT_CMD_POWER_OFF) && !pm_power_off) + cmd = LINUX_REBOOT_CMD_HALT; + + mutex_lock(&reboot_mutex); + switch (cmd) { + case LINUX_REBOOT_CMD_RESTART: + kernel_restart(NULL); + break; + + case LINUX_REBOOT_CMD_CAD_ON: + C_A_D = 1; + break; + + case LINUX_REBOOT_CMD_CAD_OFF: + C_A_D = 0; + break; + + case LINUX_REBOOT_CMD_HALT: + kernel_halt(); + do_exit(0); + panic("cannot halt"); + + case LINUX_REBOOT_CMD_POWER_OFF: + kernel_power_off(); + do_exit(0); + break; + + case LINUX_REBOOT_CMD_RESTART2: + if (strncpy_from_user(&buffer[0], arg, sizeof(buffer) - 1) < 0) { + ret = -EFAULT; + break; + } + buffer[sizeof(buffer) - 1] = '\0'; + + kernel_restart(buffer); + break; + +#ifdef CONFIG_KEXEC + case LINUX_REBOOT_CMD_KEXEC: + ret = kernel_kexec(); + break; +#endif + +#ifdef CONFIG_HIBERNATION + case LINUX_REBOOT_CMD_SW_SUSPEND: + ret = hibernate(); + break; +#endif + + default: + ret = -EINVAL; + break; + } + mutex_unlock(&reboot_mutex); + return ret; +} + +static void deferred_cad(struct work_struct *dummy) +{ + kernel_restart(NULL); +} + +/* + * This function gets called by ctrl-alt-del - ie the keyboard interrupt. + * As it's called within an interrupt, it may NOT sync: the only choice + * is whether to reboot at once, or just ignore the ctrl-alt-del. + */ +void ctrl_alt_del(void) +{ + static DECLARE_WORK(cad_work, deferred_cad); + + if (C_A_D) + schedule_work(&cad_work); + else + kill_cad_pid(SIGINT, 1); +} + +char poweroff_cmd[POWEROFF_CMD_PATH_LEN] = "/sbin/poweroff"; + +static int __orderly_poweroff(bool force) +{ + char **argv; + static char *envp[] = { + "HOME=/", + "PATH=/sbin:/bin:/usr/sbin:/usr/bin", + NULL + }; + int ret; + + argv = argv_split(GFP_KERNEL, poweroff_cmd, NULL); + if (argv) { + ret = call_usermodehelper(argv[0], argv, envp, UMH_WAIT_EXEC); + argv_free(argv); + } else { + printk(KERN_WARNING "%s failed to allocate memory for \"%s\"\n", + __func__, poweroff_cmd); + ret = -ENOMEM; + } + + if (ret && force) { + printk(KERN_WARNING "Failed to start orderly shutdown: " + "forcing the issue\n"); + /* + * I guess this should try to kick off some daemon to sync and + * poweroff asap. Or not even bother syncing if we're doing an + * emergency shutdown? + */ + emergency_sync(); + kernel_power_off(); + } + + return ret; +} + +static bool poweroff_force; + +static void poweroff_work_func(struct work_struct *work) +{ + __orderly_poweroff(poweroff_force); +} + +static DECLARE_WORK(poweroff_work, poweroff_work_func); + +/** + * orderly_poweroff - Trigger an orderly system poweroff + * @force: force poweroff if command execution fails + * + * This may be called from any context to trigger a system shutdown. + * If the orderly shutdown fails, it will force an immediate shutdown. + */ +int orderly_poweroff(bool force) +{ + if (force) /* do not override the pending "true" */ + poweroff_force = true; + schedule_work(&poweroff_work); + return 0; +} +EXPORT_SYMBOL_GPL(orderly_poweroff); diff --git a/kernel/sys.c b/kernel/sys.c index b882440bd0c0..771129b299f8 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -115,20 +115,6 @@ int fs_overflowgid = DEFAULT_FS_OVERFLOWUID; EXPORT_SYMBOL(fs_overflowuid); EXPORT_SYMBOL(fs_overflowgid); -/* - * this indicates whether you can reboot with ctrl-alt-del: the default is yes - */ - -int C_A_D = 1; -struct pid *cad_pid; -EXPORT_SYMBOL(cad_pid); - -/* - * If set, this is used for preparing the system to power off. - */ - -void (*pm_power_off_prepare)(void); - /* * Returns true if current's euid is same as p's uid or euid, * or has CAP_SYS_NICE to p's user_ns. @@ -308,261 +294,6 @@ out_unlock: return retval; } -/** - * emergency_restart - reboot the system - * - * Without shutting down any hardware or taking any locks - * reboot the system. This is called when we know we are in - * trouble so this is our best effort to reboot. This is - * safe to call in interrupt context. - */ -void emergency_restart(void) -{ - kmsg_dump(KMSG_DUMP_EMERG); - machine_emergency_restart(); -} -EXPORT_SYMBOL_GPL(emergency_restart); - -void kernel_restart_prepare(char *cmd) -{ - blocking_notifier_call_chain(&reboot_notifier_list, SYS_RESTART, cmd); - system_state = SYSTEM_RESTART; - usermodehelper_disable(); - device_shutdown(); -} - -/** - * register_reboot_notifier - Register function to be called at reboot time - * @nb: Info about notifier function to be called - * - * Registers a function with the list of functions - * to be called at reboot time. - * - * Currently always returns zero, as blocking_notifier_chain_register() - * always returns zero. - */ -int register_reboot_notifier(struct notifier_block *nb) -{ - return blocking_notifier_chain_register(&reboot_notifier_list, nb); -} -EXPORT_SYMBOL(register_reboot_notifier); - -/** - * unregister_reboot_notifier - Unregister previously registered reboot notifier - * @nb: Hook to be unregistered - * - * Unregisters a previously registered reboot - * notifier function. - * - * Returns zero on success, or %-ENOENT on failure. - */ -int unregister_reboot_notifier(struct notifier_block *nb) -{ - return blocking_notifier_chain_unregister(&reboot_notifier_list, nb); -} -EXPORT_SYMBOL(unregister_reboot_notifier); - -static void migrate_to_reboot_cpu(void) -{ - /* The boot cpu is always logical cpu 0 */ - int cpu = 0; - - cpu_hotplug_disable(); - - /* Make certain the cpu I'm about to reboot on is online */ - if (!cpu_online(cpu)) - cpu = cpumask_first(cpu_online_mask); - - /* Prevent races with other tasks migrating this task */ - current->flags |= PF_NO_SETAFFINITY; - - /* Make certain I only run on the appropriate processor */ - set_cpus_allowed_ptr(current, cpumask_of(cpu)); -} - -/** - * kernel_restart - reboot the system - * @cmd: pointer to buffer containing command to execute for restart - * or %NULL - * - * Shutdown everything and perform a clean reboot. - * This is not safe to call in interrupt context. - */ -void kernel_restart(char *cmd) -{ - kernel_restart_prepare(cmd); - migrate_to_reboot_cpu(); - syscore_shutdown(); - if (!cmd) - printk(KERN_EMERG "Restarting system.\n"); - else - printk(KERN_EMERG "Restarting system with command '%s'.\n", cmd); - kmsg_dump(KMSG_DUMP_RESTART); - machine_restart(cmd); -} -EXPORT_SYMBOL_GPL(kernel_restart); - -static void kernel_shutdown_prepare(enum system_states state) -{ - blocking_notifier_call_chain(&reboot_notifier_list, - (state == SYSTEM_HALT)?SYS_HALT:SYS_POWER_OFF, NULL); - system_state = state; - usermodehelper_disable(); - device_shutdown(); -} -/** - * kernel_halt - halt the system - * - * Shutdown everything and perform a clean system halt. - */ -void kernel_halt(void) -{ - kernel_shutdown_prepare(SYSTEM_HALT); - migrate_to_reboot_cpu(); - syscore_shutdown(); - printk(KERN_EMERG "System halted.\n"); - kmsg_dump(KMSG_DUMP_HALT); - machine_halt(); -} - -EXPORT_SYMBOL_GPL(kernel_halt); - -/** - * kernel_power_off - power_off the system - * - * Shutdown everything and perform a clean system power_off. - */ -void kernel_power_off(void) -{ - kernel_shutdown_prepare(SYSTEM_POWER_OFF); - if (pm_power_off_prepare) - pm_power_off_prepare(); - migrate_to_reboot_cpu(); - syscore_shutdown(); - printk(KERN_EMERG "Power down.\n"); - kmsg_dump(KMSG_DUMP_POWEROFF); - machine_power_off(); -} -EXPORT_SYMBOL_GPL(kernel_power_off); - -static DEFINE_MUTEX(reboot_mutex); - -/* - * Reboot system call: for obvious reasons only root may call it, - * and even root needs to set up some magic numbers in the registers - * so that some mistake won't make this reboot the whole machine. - * You can also set the meaning of the ctrl-alt-del-key here. - * - * reboot doesn't sync: do that yourself before calling this. - */ -SYSCALL_DEFINE4(reboot, int, magic1, int, magic2, unsigned int, cmd, - void __user *, arg) -{ - struct pid_namespace *pid_ns = task_active_pid_ns(current); - char buffer[256]; - int ret = 0; - - /* We only trust the superuser with rebooting the system. */ - if (!ns_capable(pid_ns->user_ns, CAP_SYS_BOOT)) - return -EPERM; - - /* For safety, we require "magic" arguments. */ - if (magic1 != LINUX_REBOOT_MAGIC1 || - (magic2 != LINUX_REBOOT_MAGIC2 && - magic2 != LINUX_REBOOT_MAGIC2A && - magic2 != LINUX_REBOOT_MAGIC2B && - magic2 != LINUX_REBOOT_MAGIC2C)) - return -EINVAL; - - /* - * If pid namespaces are enabled and the current task is in a child - * pid_namespace, the command is handled by reboot_pid_ns() which will - * call do_exit(). - */ - ret = reboot_pid_ns(pid_ns, cmd); - if (ret) - return ret; - - /* Instead of trying to make the power_off code look like - * halt when pm_power_off is not set do it the easy way. - */ - if ((cmd == LINUX_REBOOT_CMD_POWER_OFF) && !pm_power_off) - cmd = LINUX_REBOOT_CMD_HALT; - - mutex_lock(&reboot_mutex); - switch (cmd) { - case LINUX_REBOOT_CMD_RESTART: - kernel_restart(NULL); - break; - - case LINUX_REBOOT_CMD_CAD_ON: - C_A_D = 1; - break; - - case LINUX_REBOOT_CMD_CAD_OFF: - C_A_D = 0; - break; - - case LINUX_REBOOT_CMD_HALT: - kernel_halt(); - do_exit(0); - panic("cannot halt.\n"); - - case LINUX_REBOOT_CMD_POWER_OFF: - kernel_power_off(); - do_exit(0); - break; - - case LINUX_REBOOT_CMD_RESTART2: - if (strncpy_from_user(&buffer[0], arg, sizeof(buffer) - 1) < 0) { - ret = -EFAULT; - break; - } - buffer[sizeof(buffer) - 1] = '\0'; - - kernel_restart(buffer); - break; - -#ifdef CONFIG_KEXEC - case LINUX_REBOOT_CMD_KEXEC: - ret = kernel_kexec(); - break; -#endif - -#ifdef CONFIG_HIBERNATION - case LINUX_REBOOT_CMD_SW_SUSPEND: - ret = hibernate(); - break; -#endif - - default: - ret = -EINVAL; - break; - } - mutex_unlock(&reboot_mutex); - return ret; -} - -static void deferred_cad(struct work_struct *dummy) -{ - kernel_restart(NULL); -} - -/* - * This function gets called by ctrl-alt-del - ie the keyboard interrupt. - * As it's called within an interrupt, it may NOT sync: the only choice - * is whether to reboot at once, or just ignore the ctrl-alt-del. - */ -void ctrl_alt_del(void) -{ - static DECLARE_WORK(cad_work, deferred_cad); - - if (C_A_D) - schedule_work(&cad_work); - else - kill_cad_pid(SIGINT, 1); -} - /* * Unprivileged users may change the real gid to the effective gid * or vice versa. (BSD-style) @@ -2287,68 +2018,6 @@ SYSCALL_DEFINE3(getcpu, unsigned __user *, cpup, unsigned __user *, nodep, return err ? -EFAULT : 0; } -char poweroff_cmd[POWEROFF_CMD_PATH_LEN] = "/sbin/poweroff"; - -static int __orderly_poweroff(bool force) -{ - char **argv; - static char *envp[] = { - "HOME=/", - "PATH=/sbin:/bin:/usr/sbin:/usr/bin", - NULL - }; - int ret; - - argv = argv_split(GFP_KERNEL, poweroff_cmd, NULL); - if (argv) { - ret = call_usermodehelper(argv[0], argv, envp, UMH_WAIT_EXEC); - argv_free(argv); - } else { - printk(KERN_WARNING "%s failed to allocate memory for \"%s\"\n", - __func__, poweroff_cmd); - ret = -ENOMEM; - } - - if (ret && force) { - printk(KERN_WARNING "Failed to start orderly shutdown: " - "forcing the issue\n"); - /* - * I guess this should try to kick off some daemon to sync and - * poweroff asap. Or not even bother syncing if we're doing an - * emergency shutdown? - */ - emergency_sync(); - kernel_power_off(); - } - - return ret; -} - -static bool poweroff_force; - -static void poweroff_work_func(struct work_struct *work) -{ - __orderly_poweroff(poweroff_force); -} - -static DECLARE_WORK(poweroff_work, poweroff_work_func); - -/** - * orderly_poweroff - Trigger an orderly system poweroff - * @force: force poweroff if command execution fails - * - * This may be called from any context to trigger a system shutdown. - * If the orderly shutdown fails, it will force an immediate shutdown. - */ -int orderly_poweroff(bool force) -{ - if (force) /* do not override the pending "true" */ - poweroff_force = true; - schedule_work(&poweroff_work); - return 0; -} -EXPORT_SYMBOL_GPL(orderly_poweroff); - /** * do_sysinfo - fill in sysinfo struct * @info: pointer to buffer to fill -- cgit v1.2.3 From 972ee83df88a7fd84c228a31b4f9611299898984 Mon Sep 17 00:00:00 2001 From: Robin Holt Date: Mon, 8 Jul 2013 16:01:34 -0700 Subject: reboot: checkpatch.pl the new kernel/reboot.c file Get the new file to pass scripts/checkpatch.pl Signed-off-by: Robin Holt Cc: H. Peter Anvin Cc: Russ Anderson Cc: Robin Holt Cc: Russell King Cc: Guan Xuetao Cc: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/reboot.c | 27 +++++++++++++-------------- 1 file changed, 13 insertions(+), 14 deletions(-) (limited to 'kernel') diff --git a/kernel/reboot.c b/kernel/reboot.c index 37d2636a65c2..abb6a0483716 100644 --- a/kernel/reboot.c +++ b/kernel/reboot.c @@ -4,6 +4,8 @@ * Copyright (C) 2013 Linus Torvalds */ +#define pr_fmt(fmt) "reboot: " fmt + #include #include #include @@ -114,9 +116,9 @@ void kernel_restart(char *cmd) migrate_to_reboot_cpu(); syscore_shutdown(); if (!cmd) - printk(KERN_EMERG "Restarting system.\n"); + pr_emerg("Restarting system\n"); else - printk(KERN_EMERG "Restarting system with command '%s'.\n", cmd); + pr_emerg("Restarting system with command '%s'\n", cmd); kmsg_dump(KMSG_DUMP_RESTART); machine_restart(cmd); } @@ -125,7 +127,7 @@ EXPORT_SYMBOL_GPL(kernel_restart); static void kernel_shutdown_prepare(enum system_states state) { blocking_notifier_call_chain(&reboot_notifier_list, - (state == SYSTEM_HALT)?SYS_HALT:SYS_POWER_OFF, NULL); + (state == SYSTEM_HALT) ? SYS_HALT : SYS_POWER_OFF, NULL); system_state = state; usermodehelper_disable(); device_shutdown(); @@ -140,11 +142,10 @@ void kernel_halt(void) kernel_shutdown_prepare(SYSTEM_HALT); migrate_to_reboot_cpu(); syscore_shutdown(); - printk(KERN_EMERG "System halted.\n"); + pr_emerg("System halted\n"); kmsg_dump(KMSG_DUMP_HALT); machine_halt(); } - EXPORT_SYMBOL_GPL(kernel_halt); /** @@ -159,7 +160,7 @@ void kernel_power_off(void) pm_power_off_prepare(); migrate_to_reboot_cpu(); syscore_shutdown(); - printk(KERN_EMERG "Power down.\n"); + pr_emerg("Power down\n"); kmsg_dump(KMSG_DUMP_POWEROFF); machine_power_off(); } @@ -188,10 +189,10 @@ SYSCALL_DEFINE4(reboot, int, magic1, int, magic2, unsigned int, cmd, /* For safety, we require "magic" arguments. */ if (magic1 != LINUX_REBOOT_MAGIC1 || - (magic2 != LINUX_REBOOT_MAGIC2 && - magic2 != LINUX_REBOOT_MAGIC2A && + (magic2 != LINUX_REBOOT_MAGIC2 && + magic2 != LINUX_REBOOT_MAGIC2A && magic2 != LINUX_REBOOT_MAGIC2B && - magic2 != LINUX_REBOOT_MAGIC2C)) + magic2 != LINUX_REBOOT_MAGIC2C)) return -EINVAL; /* @@ -234,7 +235,8 @@ SYSCALL_DEFINE4(reboot, int, magic1, int, magic2, unsigned int, cmd, break; case LINUX_REBOOT_CMD_RESTART2: - if (strncpy_from_user(&buffer[0], arg, sizeof(buffer) - 1) < 0) { + ret = strncpy_from_user(&buffer[0], arg, sizeof(buffer) - 1); + if (ret < 0) { ret = -EFAULT; break; } @@ -300,14 +302,11 @@ static int __orderly_poweroff(bool force) ret = call_usermodehelper(argv[0], argv, envp, UMH_WAIT_EXEC); argv_free(argv); } else { - printk(KERN_WARNING "%s failed to allocate memory for \"%s\"\n", - __func__, poweroff_cmd); ret = -ENOMEM; } if (ret && force) { - printk(KERN_WARNING "Failed to start orderly shutdown: " - "forcing the issue\n"); + pr_warn("Failed to start orderly shutdown: forcing the issue\n"); /* * I guess this should try to kick off some daemon to sync and * poweroff asap. Or not even bother syncing if we're doing an -- cgit v1.2.3 From 1b3a5d02ee070c8f9943333b9b6370f486601e0f Mon Sep 17 00:00:00 2001 From: Robin Holt Date: Mon, 8 Jul 2013 16:01:42 -0700 Subject: reboot: move arch/x86 reboot= handling to generic kernel Merge together the unicore32, arm, and x86 reboot= command line parameter handling. Signed-off-by: Robin Holt Cc: H. Peter Anvin Cc: Russell King Cc: Guan Xuetao Cc: Russ Anderson Cc: Robin Holt Acked-by: Ingo Molnar Acked-by: Guan Xuetao Acked-by: Russell King Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/reboot.c | 76 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 75 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/reboot.c b/kernel/reboot.c index abb6a0483716..269ed9384cc4 100644 --- a/kernel/reboot.c +++ b/kernel/reboot.c @@ -6,6 +6,7 @@ #define pr_fmt(fmt) "reboot: " fmt +#include #include #include #include @@ -24,6 +25,18 @@ int C_A_D = 1; struct pid *cad_pid; EXPORT_SYMBOL(cad_pid); +#if defined(CONFIG_ARM) || defined(CONFIG_UNICORE32) +#define DEFAULT_REBOOT_MODE = REBOOT_HARD +#else +#define DEFAULT_REBOOT_MODE +#endif +enum reboot_mode reboot_mode DEFAULT_REBOOT_MODE; + +int reboot_default; +int reboot_cpu; +enum reboot_type reboot_type = BOOT_ACPI; +int reboot_force; + /* * If set, this is used for preparing the system to power off. */ @@ -87,7 +100,7 @@ EXPORT_SYMBOL(unregister_reboot_notifier); static void migrate_to_reboot_cpu(void) { /* The boot cpu is always logical cpu 0 */ - int cpu = 0; + int cpu = reboot_cpu; cpu_hotplug_disable(); @@ -343,3 +356,64 @@ int orderly_poweroff(bool force) return 0; } EXPORT_SYMBOL_GPL(orderly_poweroff); + +static int __init reboot_setup(char *str) +{ + for (;;) { + /* + * Having anything passed on the command line via + * reboot= will cause us to disable DMI checking + * below. + */ + reboot_default = 0; + + switch (*str) { + case 'w': + reboot_mode = REBOOT_WARM; + break; + + case 'c': + reboot_mode = REBOOT_COLD; + break; + + case 'h': + reboot_mode = REBOOT_HARD; + break; + + case 's': + if (isdigit(*(str+1))) + reboot_cpu = simple_strtoul(str+1, NULL, 0); + else if (str[1] == 'm' && str[2] == 'p' && + isdigit(*(str+3))) + reboot_cpu = simple_strtoul(str+3, NULL, 0); + else + reboot_mode = REBOOT_SOFT; + break; + + case 'g': + reboot_mode = REBOOT_GPIO; + break; + + case 'b': + case 'a': + case 'k': + case 't': + case 'e': + case 'p': + reboot_type = *str; + break; + + case 'f': + reboot_force = 1; + break; + } + + str = strchr(str, ','); + if (str) + str++; + else + break; + } + return 1; +} +__setup("reboot=", reboot_setup); -- cgit v1.2.3 From 98d1e64f95b177d0f14efbdf695a1b28e1428035 Mon Sep 17 00:00:00 2001 From: Michel Lespinasse Date: Wed, 10 Jul 2013 16:05:12 -0700 Subject: mm: remove free_area_cache Since all architectures have been converted to use vm_unmapped_area(), there is no remaining use for the free_area_cache. Signed-off-by: Michel Lespinasse Acked-by: Rik van Riel Cc: "James E.J. Bottomley" Cc: "Luck, Tony" Cc: Benjamin Herrenschmidt Cc: David Howells Cc: Helge Deller Cc: Ivan Kokshaysky Cc: Matt Turner Cc: Paul Mackerras Cc: Richard Henderson Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 4 ---- 1 file changed, 4 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index 6e6a1c11b3e5..66635c80a813 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -365,8 +365,6 @@ static int dup_mmap(struct mm_struct *mm, struct mm_struct *oldmm) mm->locked_vm = 0; mm->mmap = NULL; mm->mmap_cache = NULL; - mm->free_area_cache = oldmm->mmap_base; - mm->cached_hole_size = ~0UL; mm->map_count = 0; cpumask_clear(mm_cpumask(mm)); mm->mm_rb = RB_ROOT; @@ -540,8 +538,6 @@ static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p) mm->nr_ptes = 0; memset(&mm->rss_stat, 0, sizeof(mm->rss_stat)); spin_lock_init(&mm->page_table_lock); - mm->free_area_cache = TASK_UNMAPPED_BASE; - mm->cached_hole_size = ~0UL; mm_init_aio(mm); mm_init_owner(mm, p); -- cgit v1.2.3 From 734df5ab549ca44f40de0f07af1c8803856dfb18 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Tue, 9 Jul 2013 17:44:10 +0200 Subject: perf: Clone child context from parent context pmu Currently when the child context for inherited events is created, it's based on the pmu object of the first event of the parent context. This is wrong for the following scenario: - HW context having HW and SW event - HW event got removed (closed) - SW event stays in HW context as the only event and its pmu is used to clone the child context The issue starts when the cpu context object is touched based on the pmu context object (__get_cpu_context). In this case the HW context will work with SW cpu context ending up with following WARN below. Fixing this by using parent context pmu object to clone from child context. Addresses the following warning reported by Vince Weaver: [ 2716.472065] ------------[ cut here ]------------ [ 2716.476035] WARNING: at kernel/events/core.c:2122 task_ctx_sched_out+0x3c/0x) [ 2716.476035] Modules linked in: nfsd auth_rpcgss oid_registry nfs_acl nfs locn [ 2716.476035] CPU: 0 PID: 3164 Comm: perf_fuzzer Not tainted 3.10.0-rc4 #2 [ 2716.476035] Hardware name: AOpen DE7000/nMCP7ALPx-DE R1.06 Oct.19.2012, BI2 [ 2716.476035] 0000000000000000 ffffffff8102e215 0000000000000000 ffff88011fc18 [ 2716.476035] ffff8801175557f0 0000000000000000 ffff880119fda88c ffffffff810ad [ 2716.476035] ffff880119fda880 ffffffff810af02a 0000000000000009 ffff880117550 [ 2716.476035] Call Trace: [ 2716.476035] [] ? warn_slowpath_common+0x5b/0x70 [ 2716.476035] [] ? task_ctx_sched_out+0x3c/0x5f [ 2716.476035] [] ? perf_event_exit_task+0xbf/0x194 [ 2716.476035] [] ? do_exit+0x3e7/0x90c [ 2716.476035] [] ? __do_fault+0x359/0x394 [ 2716.476035] [] ? do_group_exit+0x66/0x98 [ 2716.476035] [] ? get_signal_to_deliver+0x479/0x4ad [ 2716.476035] [] ? __perf_event_task_sched_out+0x230/0x2d1 [ 2716.476035] [] ? do_signal+0x3c/0x432 [ 2716.476035] [] ? ctx_sched_in+0x43/0x141 [ 2716.476035] [] ? perf_event_context_sched_in+0x7a/0x90 [ 2716.476035] [] ? __perf_event_task_sched_in+0x31/0x118 [ 2716.476035] [] ? mmdrop+0xd/0x1c [ 2716.476035] [] ? finish_task_switch+0x7d/0xa6 [ 2716.476035] [] ? do_notify_resume+0x20/0x5d [ 2716.476035] [] ? retint_signal+0x3d/0x78 [ 2716.476035] ---[ end trace 827178d8a5966c3d ]--- Reported-by: Vince Weaver Signed-off-by: Jiri Olsa Cc: Corey Ashford Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1373384651-6109-1-git-send-email-jolsa@redhat.com Signed-off-by: Ingo Molnar --- kernel/events/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 1833bc5a84a7..1d1f030e2f1e 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -7465,7 +7465,7 @@ inherit_task_group(struct perf_event *event, struct task_struct *parent, * child. */ - child_ctx = alloc_perf_context(event->pmu, child); + child_ctx = alloc_perf_context(parent_ctx->pmu, child); if (!child_ctx) return -ENOMEM; -- cgit v1.2.3 From 06f417968beac6e6b614e17b37d347aa6a6b1d30 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Tue, 9 Jul 2013 17:44:11 +0200 Subject: perf: Remove WARN_ON_ONCE() check in __perf_event_enable() for valid scenario The '!ctx->is_active' check has a valid scenario, so there's no need for the warning. The reason is that there's a time window between the 'ctx->is_active' check in the perf_event_enable() function and the __perf_event_enable() function having: - IRQs on - ctx->lock unlocked where the task could be killed and 'ctx' deactivated by perf_event_exit_task(), ending up with the warning below. So remove the WARN_ON_ONCE() check and add comments to explain it all. This addresses the following warning reported by Vince Weaver: [ 324.983534] ------------[ cut here ]------------ [ 324.984420] WARNING: at kernel/events/core.c:1953 __perf_event_enable+0x187/0x190() [ 324.984420] Modules linked in: [ 324.984420] CPU: 19 PID: 2715 Comm: nmi_bug_snb Not tainted 3.10.0+ #246 [ 324.984420] Hardware name: Supermicro X8DTN/X8DTN, BIOS 4.6.3 01/08/2010 [ 324.984420] 0000000000000009 ffff88043fce3ec8 ffffffff8160ea0b ffff88043fce3f00 [ 324.984420] ffffffff81080ff0 ffff8802314fdc00 ffff880231a8f800 ffff88043fcf7860 [ 324.984420] 0000000000000286 ffff880231a8f800 ffff88043fce3f10 ffffffff8108103a [ 324.984420] Call Trace: [ 324.984420] [] dump_stack+0x19/0x1b [ 324.984420] [] warn_slowpath_common+0x70/0xa0 [ 324.984420] [] warn_slowpath_null+0x1a/0x20 [ 324.984420] [] __perf_event_enable+0x187/0x190 [ 324.984420] [] remote_function+0x40/0x50 [ 324.984420] [] generic_smp_call_function_single_interrupt+0xbe/0x130 [ 324.984420] [] smp_call_function_single_interrupt+0x27/0x40 [ 324.984420] [] call_function_single_interrupt+0x6f/0x80 [ 324.984420] [] ? _raw_spin_unlock_irqrestore+0x41/0x70 [ 324.984420] [] perf_event_exit_task+0x14d/0x210 [ 324.984420] [] ? switch_task_namespaces+0x24/0x60 [ 324.984420] [] do_exit+0x2b6/0xa40 [ 324.984420] [] ? _raw_spin_unlock_irq+0x2c/0x30 [ 324.984420] [] do_group_exit+0x49/0xc0 [ 324.984420] [] get_signal_to_deliver+0x254/0x620 [ 324.984420] [] do_signal+0x57/0x5a0 [ 324.984420] [] ? __do_page_fault+0x2a4/0x4e0 [ 324.984420] [] ? retint_restore_args+0xe/0xe [ 324.984420] [] ? retint_signal+0x11/0x84 [ 324.984420] [] do_notify_resume+0x65/0x80 [ 324.984420] [] retint_signal+0x46/0x84 [ 324.984420] ---[ end trace 442ec2f04db3771a ]--- Reported-by: Vince Weaver Signed-off-by: Jiri Olsa Suggested-by: Peter Zijlstra Cc: Corey Ashford Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1373384651-6109-2-git-send-email-jolsa@redhat.com Signed-off-by: Ingo Molnar --- kernel/events/core.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 1d1f030e2f1e..ef5e7cc686e3 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -1950,7 +1950,16 @@ static int __perf_event_enable(void *info) struct perf_cpu_context *cpuctx = __get_cpu_context(ctx); int err; - if (WARN_ON_ONCE(!ctx->is_active)) + /* + * There's a time window between 'ctx->is_active' check + * in perf_event_enable function and this place having: + * - IRQs on + * - ctx->lock unlocked + * + * where the task could be killed and 'ctx' deactivated + * by perf_event_exit_task. + */ + if (!ctx->is_active) return -EINVAL; raw_spin_lock(&ctx->lock); -- cgit v1.2.3 From 058ebd0eba3aff16b144eabf4510ed9510e1416e Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 12 Jul 2013 11:08:33 +0200 Subject: perf: Fix perf_lock_task_context() vs RCU Jiri managed to trigger this warning: [] ====================================================== [] [ INFO: possible circular locking dependency detected ] [] 3.10.0+ #228 Tainted: G W [] ------------------------------------------------------- [] p/6613 is trying to acquire lock: [] (rcu_node_0){..-...}, at: [] rcu_read_unlock_special+0xa7/0x250 [] [] but task is already holding lock: [] (&ctx->lock){-.-...}, at: [] perf_lock_task_context+0xd9/0x2c0 [] [] which lock already depends on the new lock. [] [] the existing dependency chain (in reverse order) is: [] [] -> #4 (&ctx->lock){-.-...}: [] -> #3 (&rq->lock){-.-.-.}: [] -> #2 (&p->pi_lock){-.-.-.}: [] -> #1 (&rnp->nocb_gp_wq[1]){......}: [] -> #0 (rcu_node_0){..-...}: Paul was quick to explain that due to preemptible RCU we cannot call rcu_read_unlock() while holding scheduler (or nested) locks when part of the read side critical section was preemptible. Therefore solve it by making the entire RCU read side non-preemptible. Also pull out the retry from under the non-preempt to play nice with RT. Reported-by: Jiri Olsa Helped-out-by: Paul E. McKenney Cc: Signed-off-by: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/events/core.c | 15 ++++++++++++++- 1 file changed, 14 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index ef5e7cc686e3..eba8fb5834ae 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -947,8 +947,18 @@ perf_lock_task_context(struct task_struct *task, int ctxn, unsigned long *flags) { struct perf_event_context *ctx; - rcu_read_lock(); retry: + /* + * One of the few rules of preemptible RCU is that one cannot do + * rcu_read_unlock() while holding a scheduler (or nested) lock when + * part of the read side critical section was preemptible -- see + * rcu_read_unlock_special(). + * + * Since ctx->lock nests under rq->lock we must ensure the entire read + * side critical section is non-preemptible. + */ + preempt_disable(); + rcu_read_lock(); ctx = rcu_dereference(task->perf_event_ctxp[ctxn]); if (ctx) { /* @@ -964,6 +974,8 @@ retry: raw_spin_lock_irqsave(&ctx->lock, *flags); if (ctx != rcu_dereference(task->perf_event_ctxp[ctxn])) { raw_spin_unlock_irqrestore(&ctx->lock, *flags); + rcu_read_unlock(); + preempt_enable(); goto retry; } @@ -973,6 +985,7 @@ retry: } } rcu_read_unlock(); + preempt_enable(); return ctx; } -- cgit v1.2.3 From 1b375dc30710180c4b88cc59caba6e3481ec5c8b Mon Sep 17 00:00:00 2001 From: Maarten Lankhorst Date: Fri, 5 Jul 2013 09:29:32 +0200 Subject: mutex: Move ww_mutex definitions to ww_mutex.h Move the definitions for wound/wait mutexes out to a separate header, ww_mutex.h. This reduces clutter in mutex.h, and increases readability. Suggested-by: Linus Torvalds Signed-off-by: Maarten Lankhorst Acked-by: Peter Zijlstra Acked-by: Rik van Riel Acked-by: Maarten Lankhorst Cc: Dave Airlie Link: http://lkml.kernel.org/r/51D675DC.3000907@canonical.com [ Tidied up the code a bit. ] Signed-off-by: Ingo Molnar --- kernel/mutex.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/mutex.c b/kernel/mutex.c index e581ada5faf4..ff05f4bd86eb 100644 --- a/kernel/mutex.c +++ b/kernel/mutex.c @@ -18,6 +18,7 @@ * Also see Documentation/mutex-design.txt. */ #include +#include #include #include #include -- cgit v1.2.3 From a272dcca1802a7e265a56e60b0d0a6715b0a8ac2 Mon Sep 17 00:00:00 2001 From: Stephen Boyd Date: Thu, 11 Jul 2013 07:00:59 -0700 Subject: tick: broadcast: Check broadcast mode on CPU hotplug On ARM systems the dummy clockevent is registered with the cpu hotplug notifier chain before any other per-cpu clockevent. This has the side-effect of causing the dummy clockevent to be registered first in every hotplug sequence. Because the dummy is first, we'll try to turn the broadcast source on but the code in tick_device_uses_broadcast() assumes the broadcast source is in periodic mode and calls tick_broadcast_start_periodic() unconditionally. On boot this isn't a problem because we typically haven't switched into oneshot mode yet (if at all). During hotplug, if the broadcast source isn't in periodic mode we'll replace the broadcast oneshot handler with the broadcast periodic handler and start emulating oneshot mode when we shouldn't. Due to the way the broadcast oneshot handler programs the next_event it's possible for it to contain KTIME_MAX and cause us to hang the system when the periodic handler tries to program the next tick. Fix this by using the appropriate function to start the broadcast source. Reported-by: Stephen Warren Tested-by: Stephen Warren Signed-off-by: Stephen Boyd Cc: Mark Rutland Cc: Marc Zyngier Cc: ARM kernel mailing list Cc: John Stultz Cc: Joseph Lo Link: http://lkml.kernel.org/r/20130711140059.GA27430@codeaurora.org Signed-off-by: Thomas Gleixner --- kernel/time/tick-broadcast.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/time/tick-broadcast.c b/kernel/time/tick-broadcast.c index 6d3f91631de6..218bcb565fed 100644 --- a/kernel/time/tick-broadcast.c +++ b/kernel/time/tick-broadcast.c @@ -157,7 +157,10 @@ int tick_device_uses_broadcast(struct clock_event_device *dev, int cpu) dev->event_handler = tick_handle_periodic; tick_device_setup_broadcast_func(dev); cpumask_set_cpu(cpu, tick_broadcast_mask); - tick_broadcast_start_periodic(bc); + if (tick_broadcast_device.mode == TICKDEV_MODE_PERIODIC) + tick_broadcast_start_periodic(bc); + else + tick_broadcast_setup_oneshot(bc); ret = 1; } else { /* -- cgit v1.2.3 From 67516844625f45f0ce148a01c27bf41f591872b2 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Tue, 9 Jul 2013 18:56:31 +0200 Subject: perf: Remove the 'match' callback for auxiliary events processing It gives the following benefits: - only one function pointer is passed along the way - the 'match' function is called within output function and could be inlined by the compiler Suggested-by: Peter Zijlstra Signed-off-by: Jiri Olsa Cc: Corey Ashford Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1373388991-9711-1-git-send-email-jolsa@redhat.com Signed-off-by: Ingo Molnar --- kernel/events/core.c | 79 ++++++++++++++++++++++++++-------------------------- 1 file changed, 39 insertions(+), 40 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index eba8fb5834ae..708ab70ca442 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -4680,12 +4680,10 @@ perf_event_read_event(struct perf_event *event, perf_output_end(&handle); } -typedef int (perf_event_aux_match_cb)(struct perf_event *event, void *data); typedef void (perf_event_aux_output_cb)(struct perf_event *event, void *data); static void perf_event_aux_ctx(struct perf_event_context *ctx, - perf_event_aux_match_cb match, perf_event_aux_output_cb output, void *data) { @@ -4696,15 +4694,12 @@ perf_event_aux_ctx(struct perf_event_context *ctx, continue; if (!event_filter_match(event)) continue; - if (match(event, data)) - output(event, data); + output(event, data); } } static void -perf_event_aux(perf_event_aux_match_cb match, - perf_event_aux_output_cb output, - void *data, +perf_event_aux(perf_event_aux_output_cb output, void *data, struct perf_event_context *task_ctx) { struct perf_cpu_context *cpuctx; @@ -4717,7 +4712,7 @@ perf_event_aux(perf_event_aux_match_cb match, cpuctx = get_cpu_ptr(pmu->pmu_cpu_context); if (cpuctx->unique_pmu != pmu) goto next; - perf_event_aux_ctx(&cpuctx->ctx, match, output, data); + perf_event_aux_ctx(&cpuctx->ctx, output, data); if (task_ctx) goto next; ctxn = pmu->task_ctx_nr; @@ -4725,14 +4720,14 @@ perf_event_aux(perf_event_aux_match_cb match, goto next; ctx = rcu_dereference(current->perf_event_ctxp[ctxn]); if (ctx) - perf_event_aux_ctx(ctx, match, output, data); + perf_event_aux_ctx(ctx, output, data); next: put_cpu_ptr(pmu->pmu_cpu_context); } if (task_ctx) { preempt_disable(); - perf_event_aux_ctx(task_ctx, match, output, data); + perf_event_aux_ctx(task_ctx, output, data); preempt_enable(); } rcu_read_unlock(); @@ -4759,6 +4754,12 @@ struct perf_task_event { } event_id; }; +static int perf_event_task_match(struct perf_event *event) +{ + return event->attr.comm || event->attr.mmap || + event->attr.mmap_data || event->attr.task; +} + static void perf_event_task_output(struct perf_event *event, void *data) { @@ -4768,6 +4769,9 @@ static void perf_event_task_output(struct perf_event *event, struct task_struct *task = task_event->task; int ret, size = task_event->event_id.header.size; + if (!perf_event_task_match(event)) + return; + perf_event_header__init_id(&task_event->event_id.header, &sample, event); ret = perf_output_begin(&handle, event, @@ -4790,13 +4794,6 @@ out: task_event->event_id.header.size = size; } -static int perf_event_task_match(struct perf_event *event, - void *data __maybe_unused) -{ - return event->attr.comm || event->attr.mmap || - event->attr.mmap_data || event->attr.task; -} - static void perf_event_task(struct task_struct *task, struct perf_event_context *task_ctx, int new) @@ -4825,8 +4822,7 @@ static void perf_event_task(struct task_struct *task, }, }; - perf_event_aux(perf_event_task_match, - perf_event_task_output, + perf_event_aux(perf_event_task_output, &task_event, task_ctx); } @@ -4853,6 +4849,11 @@ struct perf_comm_event { } event_id; }; +static int perf_event_comm_match(struct perf_event *event) +{ + return event->attr.comm; +} + static void perf_event_comm_output(struct perf_event *event, void *data) { @@ -4862,6 +4863,9 @@ static void perf_event_comm_output(struct perf_event *event, int size = comm_event->event_id.header.size; int ret; + if (!perf_event_comm_match(event)) + return; + perf_event_header__init_id(&comm_event->event_id.header, &sample, event); ret = perf_output_begin(&handle, event, comm_event->event_id.header.size); @@ -4883,12 +4887,6 @@ out: comm_event->event_id.header.size = size; } -static int perf_event_comm_match(struct perf_event *event, - void *data __maybe_unused) -{ - return event->attr.comm; -} - static void perf_event_comm_event(struct perf_comm_event *comm_event) { char comm[TASK_COMM_LEN]; @@ -4903,8 +4901,7 @@ static void perf_event_comm_event(struct perf_comm_event *comm_event) comm_event->event_id.header.size = sizeof(comm_event->event_id) + size; - perf_event_aux(perf_event_comm_match, - perf_event_comm_output, + perf_event_aux(perf_event_comm_output, comm_event, NULL); } @@ -4967,6 +4964,17 @@ struct perf_mmap_event { } event_id; }; +static int perf_event_mmap_match(struct perf_event *event, + void *data) +{ + struct perf_mmap_event *mmap_event = data; + struct vm_area_struct *vma = mmap_event->vma; + int executable = vma->vm_flags & VM_EXEC; + + return (!executable && event->attr.mmap_data) || + (executable && event->attr.mmap); +} + static void perf_event_mmap_output(struct perf_event *event, void *data) { @@ -4976,6 +4984,9 @@ static void perf_event_mmap_output(struct perf_event *event, int size = mmap_event->event_id.header.size; int ret; + if (!perf_event_mmap_match(event, data)) + return; + perf_event_header__init_id(&mmap_event->event_id.header, &sample, event); ret = perf_output_begin(&handle, event, mmap_event->event_id.header.size); @@ -4996,17 +5007,6 @@ out: mmap_event->event_id.header.size = size; } -static int perf_event_mmap_match(struct perf_event *event, - void *data) -{ - struct perf_mmap_event *mmap_event = data; - struct vm_area_struct *vma = mmap_event->vma; - int executable = vma->vm_flags & VM_EXEC; - - return (!executable && event->attr.mmap_data) || - (executable && event->attr.mmap); -} - static void perf_event_mmap_event(struct perf_mmap_event *mmap_event) { struct vm_area_struct *vma = mmap_event->vma; @@ -5070,8 +5070,7 @@ got_name: mmap_event->event_id.header.size = sizeof(mmap_event->event_id) + size; - perf_event_aux(perf_event_mmap_match, - perf_event_mmap_output, + perf_event_aux(perf_event_mmap_output, mmap_event, NULL); -- cgit v1.2.3 From c4be9cb4f19cbd534a6c4c334cd48d8bb483e17a Mon Sep 17 00:00:00 2001 From: Michel Lespinasse Date: Mon, 8 Jul 2013 14:23:51 -0700 Subject: lglock: Update lockdep annotations to report recursive local locks Oleg Nesterov recently noticed that the lockdep annotations in lglock.c are not sufficient to detect some obvious deadlocks, such as lg_local_lock(LOCK) + lg_local_lock(LOCK) or spin_lock(X) + lg_local_lock(Y) vs lg_local_lock(Y) + spin_lock(X). Both issues are easily fixed by indicating to lockdep that lglock's local locks are not recursive. We shouldn't use the rwlock acquire/release functions here, as lglock doesn't share the same semantics. Instead we can base our lockdep annotations on the lock_acquire_shared (for local lglock) and lock_acquire_exclusive (for global lglock) helpers. I am not proposing new lglock specific helpers as I don't see the point of the existing second level of helpers :) Noticed-by: Oleg Nesterov Signed-off-by: Michel Lespinasse Cc: Lai Jiangshan Cc: "Srivatsa S. Bhat" Cc: Rusty Russell Cc: Andi Kleen Cc: "Paul E. McKenney" Cc: Steven Rostedt Signed-off-by: Andrew Morton Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/20130708212352.1769031C15E@corp2gmr1-1.hot.corp.google.com Signed-off-by: Ingo Molnar --- kernel/lglock.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/lglock.c b/kernel/lglock.c index 6535a667a5a7..86ae2aebf004 100644 --- a/kernel/lglock.c +++ b/kernel/lglock.c @@ -21,7 +21,7 @@ void lg_local_lock(struct lglock *lg) arch_spinlock_t *lock; preempt_disable(); - rwlock_acquire_read(&lg->lock_dep_map, 0, 0, _RET_IP_); + lock_acquire_shared(&lg->lock_dep_map, 0, 0, NULL, _RET_IP_); lock = this_cpu_ptr(lg->lock); arch_spin_lock(lock); } @@ -31,7 +31,7 @@ void lg_local_unlock(struct lglock *lg) { arch_spinlock_t *lock; - rwlock_release(&lg->lock_dep_map, 1, _RET_IP_); + lock_release(&lg->lock_dep_map, 1, _RET_IP_); lock = this_cpu_ptr(lg->lock); arch_spin_unlock(lock); preempt_enable(); @@ -43,7 +43,7 @@ void lg_local_lock_cpu(struct lglock *lg, int cpu) arch_spinlock_t *lock; preempt_disable(); - rwlock_acquire_read(&lg->lock_dep_map, 0, 0, _RET_IP_); + lock_acquire_shared(&lg->lock_dep_map, 0, 0, NULL, _RET_IP_); lock = per_cpu_ptr(lg->lock, cpu); arch_spin_lock(lock); } @@ -53,7 +53,7 @@ void lg_local_unlock_cpu(struct lglock *lg, int cpu) { arch_spinlock_t *lock; - rwlock_release(&lg->lock_dep_map, 1, _RET_IP_); + lock_release(&lg->lock_dep_map, 1, _RET_IP_); lock = per_cpu_ptr(lg->lock, cpu); arch_spin_unlock(lock); preempt_enable(); @@ -65,7 +65,7 @@ void lg_global_lock(struct lglock *lg) int i; preempt_disable(); - rwlock_acquire(&lg->lock_dep_map, 0, 0, _RET_IP_); + lock_acquire_exclusive(&lg->lock_dep_map, 0, 0, NULL, _RET_IP_); for_each_possible_cpu(i) { arch_spinlock_t *lock; lock = per_cpu_ptr(lg->lock, i); @@ -78,7 +78,7 @@ void lg_global_unlock(struct lglock *lg) { int i; - rwlock_release(&lg->lock_dep_map, 1, _RET_IP_); + lock_release(&lg->lock_dep_map, 1, _RET_IP_); for_each_possible_cpu(i) { arch_spinlock_t *lock; lock = per_cpu_ptr(lg->lock, i); -- cgit v1.2.3 From 971ee28cbd1ccd87b3164facd9359a534c1d2892 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 28 Jun 2013 11:18:53 +0200 Subject: sched: Fix HRTICK David reported that the HRTICK sched feature was borken; which was enough motivation for me to finally fix it ;-) We should not allow hrtimer code to do softirq wakeups while holding scheduler locks. The hrtimer code only needs this when we accidentally try to program an expired time. We don't much care about those anyway since we have the regular tick to fall back to. Reported-by: David Ahern Tested-by: David Ahern Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/20130628091853.GE29209@dyad.programming.kicks-ass.net Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 20 +++++++++++--------- 1 file changed, 11 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 9b1f2e533b95..0d8eb4525e76 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -370,13 +370,6 @@ static struct rq *this_rq_lock(void) #ifdef CONFIG_SCHED_HRTICK /* * Use HR-timers to deliver accurate preemption points. - * - * Its all a bit involved since we cannot program an hrt while holding the - * rq->lock. So what we do is store a state in in rq->hrtick_* and ask for a - * reschedule event. - * - * When we get rescheduled we reprogram the hrtick_timer outside of the - * rq->lock. */ static void hrtick_clear(struct rq *rq) @@ -404,6 +397,15 @@ static enum hrtimer_restart hrtick(struct hrtimer *timer) } #ifdef CONFIG_SMP + +static int __hrtick_restart(struct rq *rq) +{ + struct hrtimer *timer = &rq->hrtick_timer; + ktime_t time = hrtimer_get_softexpires(timer); + + return __hrtimer_start_range_ns(timer, time, 0, HRTIMER_MODE_ABS_PINNED, 0); +} + /* * called from hardirq (IPI) context */ @@ -412,7 +414,7 @@ static void __hrtick_start(void *arg) struct rq *rq = arg; raw_spin_lock(&rq->lock); - hrtimer_restart(&rq->hrtick_timer); + __hrtick_restart(rq); rq->hrtick_csd_pending = 0; raw_spin_unlock(&rq->lock); } @@ -430,7 +432,7 @@ void hrtick_start(struct rq *rq, u64 delay) hrtimer_set_expires(timer, time); if (rq == this_rq()) { - hrtimer_restart(timer); + __hrtick_restart(rq); } else if (!rq->hrtick_csd_pending) { __smp_call_function_single(cpu_of(rq), &rq->hrtick_csd, 0); rq->hrtick_csd_pending = 1; -- cgit v1.2.3 From cedce3e730833d26a37826a96e1905b6ef387df9 Mon Sep 17 00:00:00 2001 From: Kirill Tkhai Date: Thu, 4 Jul 2013 22:48:20 +0400 Subject: sched/__wake_up_sync_key(): Fix nr_exclusive tasks which lead to WF_SYNC clearing Only one task can replace the waker. Signed-off-by: Kirill Tkhai CC: Steven Rostedt Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/512421372963700@web25f.yandex.ru Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 0d8eb4525e76..f73787159188 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -2660,7 +2660,7 @@ void __wake_up_sync_key(wait_queue_head_t *q, unsigned int mode, if (unlikely(!q)) return; - if (unlikely(!nr_exclusive)) + if (unlikely(nr_exclusive != 1)) wake_flags = 0; spin_lock_irqsave(&q->lock, flags); -- cgit v1.2.3 From 8f89140ae41ccd9c63344e6823faa862aa7435e3 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 28 Jun 2013 16:24:10 -0700 Subject: cgroup: minor updates around cgroup_clear_directory() * Rename it to cgroup_clear_dir() and make it take the pointer to the target cgroup instead of the the dentry. This makes the function consistent with its counterpart - cgroup_populate_dir(). * Move cgroup_clear_directory() invocation from cgroup_d_remove_dir() to cgroup_remount() so that the function doesn't have to determine the cgroup pointer back from the dentry. cgroup_d_remove_dir() now only deals with vfs, which is slightly cleaner. This patch doesn't introduce any functional differences. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 19 ++++++++----------- 1 file changed, 8 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index e5583d10a325..09bfa870e698 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -957,15 +957,14 @@ static void cgroup_rm_file(struct cgroup *cgrp, const struct cftype *cft) } /** - * cgroup_clear_directory - selective removal of base and subsystem files - * @dir: directory containing the files + * cgroup_clear_dir - selective removal of base and subsystem files + * @cgrp: target cgroup * @base_files: true if the base files should be removed * @subsys_mask: mask of the subsystem ids whose files should be removed */ -static void cgroup_clear_directory(struct dentry *dir, bool base_files, - unsigned long subsys_mask) +static void cgroup_clear_dir(struct cgroup *cgrp, bool base_files, + unsigned long subsys_mask) { - struct cgroup *cgrp = __d_cgrp(dir); struct cgroup_subsys *ss; for_each_root_subsys(cgrp->root, ss) { @@ -987,9 +986,6 @@ static void cgroup_clear_directory(struct dentry *dir, bool base_files, static void cgroup_d_remove_dir(struct dentry *dentry) { struct dentry *parent; - struct cgroupfs_root *root = dentry->d_sb->s_fs_info; - - cgroup_clear_directory(dentry, true, root->subsys_mask); parent = dentry->d_parent; spin_lock(&parent->d_lock); @@ -1376,7 +1372,7 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) * this before rebind_subsystems, since rebind_subsystems may * change this hierarchy's subsys_list. */ - cgroup_clear_directory(cgrp->dentry, false, removed_mask); + cgroup_clear_dir(cgrp, false, removed_mask); ret = rebind_subsystems(root, added_mask, removed_mask); if (ret) { @@ -4541,9 +4537,10 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) raw_spin_unlock(&release_list_lock); /* - * Remove @cgrp directory. The removal puts the base ref but we - * aren't quite done with @cgrp yet, so hold onto it. + * Clear and remove @cgrp directory. The removal puts the base ref + * but we aren't quite done with @cgrp yet, so hold onto it. */ + cgroup_clear_dir(cgrp, true, cgrp->root->subsys_mask); dget(d); cgroup_d_remove_dir(d); -- cgit v1.2.3 From b1f28d3109349899e87377e89f9d8ab5bc95ec57 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 28 Jun 2013 16:24:10 -0700 Subject: cgroup: fix error path of cgroup_addrm_files() cgroup_addrm_files() mishandled error return value from cgroup_add_file() and returns error iff the last file fails to create. As we're in the process of cleaning up file add/rm error handling and will reliably propagate file creation failures, there's no point in keeping adding files after a failure. Replace the broken error collection logic with immediate error return. While at it, add lockdep assertions and function comment. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 28 ++++++++++++++++++++++------ 1 file changed, 22 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 09bfa870e698..9b16d75bec63 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -2780,11 +2780,26 @@ out: return error; } +/** + * cgroup_addrm_files - add or remove files to a cgroup directory + * @cgrp: the target cgroup + * @subsys: the subsystem of files to be added + * @cfts: array of cftypes to be added + * @is_add: whether to add or remove + * + * Depending on @is_add, add or remove files defined by @cfts on @cgrp. + * All @cfts should belong to @subsys. For removals, this function never + * fails. If addition fails, this function doesn't remove files already + * added. The caller is responsible for cleaning up. + */ static int cgroup_addrm_files(struct cgroup *cgrp, struct cgroup_subsys *subsys, struct cftype cfts[], bool is_add) { struct cftype *cft; - int err, ret = 0; + int ret; + + lockdep_assert_held(&cgrp->dentry->d_inode->i_mutex); + lockdep_assert_held(&cgroup_mutex); for (cft = cfts; cft->name[0] != '\0'; cft++) { /* does cft->flags tell us to skip this file on @cgrp? */ @@ -2796,16 +2811,17 @@ static int cgroup_addrm_files(struct cgroup *cgrp, struct cgroup_subsys *subsys, continue; if (is_add) { - err = cgroup_add_file(cgrp, subsys, cft); - if (err) + ret = cgroup_add_file(cgrp, subsys, cft); + if (ret) { pr_warn("cgroup_addrm_files: failed to add %s, err=%d\n", - cft->name, err); - ret = err; + cft->name, ret); + return ret; + } } else { cgroup_rm_file(cgrp, cft); } } - return ret; + return 0; } static void cgroup_cfts_prepare(void) -- cgit v1.2.3 From 9ccece80ae19ed42439fc0ced76858f189cd41e8 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 28 Jun 2013 16:24:11 -0700 Subject: cgroup: fix cgroup_add_cftypes() error handling cgroup_add_cftypes() uses cgroup_cfts_commit() to actually create the files; however, both functions ignore actual file creation errors and just assume success. This can lead to, for example, blkio hierarchy with some of the cgroups with only subset of interface files populated after cfq-iosched is loaded under heavy memory pressure, which is nasty. This patch updates cgroup_cfts_commit() and cgroup_add_cftypes() to guarantee that all files are created on success and no file is created on failure. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 26 ++++++++++++++++++-------- 1 file changed, 18 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 9b16d75bec63..36c0ccc921f4 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -2836,8 +2836,8 @@ static void cgroup_cfts_prepare(void) mutex_lock(&cgroup_mutex); } -static void cgroup_cfts_commit(struct cgroup_subsys *ss, - struct cftype *cfts, bool is_add) +static int cgroup_cfts_commit(struct cgroup_subsys *ss, + struct cftype *cfts, bool is_add) __releases(&cgroup_mutex) { LIST_HEAD(pending); @@ -2846,12 +2846,13 @@ static void cgroup_cfts_commit(struct cgroup_subsys *ss, struct dentry *prev = NULL; struct inode *inode; u64 update_before; + int ret = 0; /* %NULL @cfts indicates abort and don't bother if @ss isn't attached */ if (!cfts || ss->root == &cgroup_dummy_root || !atomic_inc_not_zero(&sb->s_active)) { mutex_unlock(&cgroup_mutex); - return; + return 0; } /* @@ -2867,10 +2868,13 @@ static void cgroup_cfts_commit(struct cgroup_subsys *ss, inode = root->dentry->d_inode; mutex_lock(&inode->i_mutex); mutex_lock(&cgroup_mutex); - cgroup_addrm_files(root, ss, cfts, is_add); + ret = cgroup_addrm_files(root, ss, cfts, is_add); mutex_unlock(&cgroup_mutex); mutex_unlock(&inode->i_mutex); + if (ret) + goto out_deact; + /* add/rm files for all cgroups created before */ rcu_read_lock(); cgroup_for_each_descendant_pre(cgrp, root) { @@ -2887,15 +2891,19 @@ static void cgroup_cfts_commit(struct cgroup_subsys *ss, mutex_lock(&inode->i_mutex); mutex_lock(&cgroup_mutex); if (cgrp->serial_nr < update_before && !cgroup_is_dead(cgrp)) - cgroup_addrm_files(cgrp, ss, cfts, is_add); + ret = cgroup_addrm_files(cgrp, ss, cfts, is_add); mutex_unlock(&cgroup_mutex); mutex_unlock(&inode->i_mutex); rcu_read_lock(); + if (ret) + break; } rcu_read_unlock(); dput(prev); +out_deact: deactivate_super(sb); + return ret; } /** @@ -2915,6 +2923,7 @@ static void cgroup_cfts_commit(struct cgroup_subsys *ss, int cgroup_add_cftypes(struct cgroup_subsys *ss, struct cftype *cfts) { struct cftype_set *set; + int ret; set = kzalloc(sizeof(*set), GFP_KERNEL); if (!set) @@ -2923,9 +2932,10 @@ int cgroup_add_cftypes(struct cgroup_subsys *ss, struct cftype *cfts) cgroup_cfts_prepare(); set->cfts = cfts; list_add_tail(&set->node, &ss->cftsets); - cgroup_cfts_commit(ss, cfts, true); - - return 0; + ret = cgroup_cfts_commit(ss, cfts, true); + if (ret) + cgroup_rm_cftypes(ss, cfts); + return ret; } EXPORT_SYMBOL_GPL(cgroup_add_cftypes); -- cgit v1.2.3 From 628f7cd47ab758cae0353d1a6decf3d1459dca24 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 28 Jun 2013 16:24:11 -0700 Subject: cgroup: separate out cgroup_base_files[] handling out of cgroup_populate/clear_dir() cgroup_populate/clear_dir() currently take @base_files and adds and removes, respectively, cgroup_base_files[] to the directory. File additions and removals are being reorganized for proper error handling and more dynamic handling for the unified hierarchy, and mixing base and subsys file handling into the same functions gets a bit confusing. This patch moves base file handling out of cgroup_populate/clear_dir() into their users - cgroup_mount(), cgroup_create() and cgroup_destroy_locked(). Note that this changes the behavior of base file removal. If @base_files is %true, cgroup_clear_dir() used to delete files regardless of cftype until there's no files left. Now, only files with matching cfts are removed. As files can only be created by the base or registered cftypes, this shouldn't result in any behavior difference. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 46 +++++++++++++++++++--------------------------- 1 file changed, 19 insertions(+), 27 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 36c0ccc921f4..9835a097f3c0 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -215,6 +215,8 @@ static u64 cgroup_serial_nr_next = 1; */ static int need_forkexit_callback __read_mostly; +static struct cftype cgroup_base_files[]; + static void cgroup_offline_fn(struct work_struct *work); static int cgroup_destroy_locked(struct cgroup *cgrp); static int cgroup_addrm_files(struct cgroup *cgrp, struct cgroup_subsys *subsys, @@ -804,8 +806,7 @@ static struct cgroup *task_cgroup_from_root(struct task_struct *task, static int cgroup_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode); static struct dentry *cgroup_lookup(struct inode *, struct dentry *, unsigned int); static int cgroup_rmdir(struct inode *unused_dir, struct dentry *dentry); -static int cgroup_populate_dir(struct cgroup *cgrp, bool base_files, - unsigned long subsys_mask); +static int cgroup_populate_dir(struct cgroup *cgrp, unsigned long subsys_mask); static const struct inode_operations cgroup_dir_inode_operations; static const struct file_operations proc_cgroupstats_operations; @@ -957,13 +958,11 @@ static void cgroup_rm_file(struct cgroup *cgrp, const struct cftype *cft) } /** - * cgroup_clear_dir - selective removal of base and subsystem files + * cgroup_clear_dir - remove subsys files in a cgroup directory * @cgrp: target cgroup - * @base_files: true if the base files should be removed * @subsys_mask: mask of the subsystem ids whose files should be removed */ -static void cgroup_clear_dir(struct cgroup *cgrp, bool base_files, - unsigned long subsys_mask) +static void cgroup_clear_dir(struct cgroup *cgrp, unsigned long subsys_mask) { struct cgroup_subsys *ss; @@ -974,10 +973,6 @@ static void cgroup_clear_dir(struct cgroup *cgrp, bool base_files, list_for_each_entry(set, &ss->cftsets, node) cgroup_addrm_files(cgrp, NULL, set->cfts, false); } - if (base_files) { - while (!list_empty(&cgrp->files)) - cgroup_rm_file(cgrp, NULL); - } } /* @@ -1372,17 +1367,17 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) * this before rebind_subsystems, since rebind_subsystems may * change this hierarchy's subsys_list. */ - cgroup_clear_dir(cgrp, false, removed_mask); + cgroup_clear_dir(cgrp, removed_mask); ret = rebind_subsystems(root, added_mask, removed_mask); if (ret) { /* rebind_subsystems failed, re-populate the removed files */ - cgroup_populate_dir(cgrp, false, removed_mask); + cgroup_populate_dir(cgrp, removed_mask); goto out_unlock; } /* re-populate subsystem files */ - cgroup_populate_dir(cgrp, false, added_mask); + cgroup_populate_dir(cgrp, added_mask); if (opts.release_agent) strcpy(root->release_agent_path, opts.release_agent); @@ -1687,7 +1682,8 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, BUG_ON(root->number_of_cgroups != 1); cred = override_creds(&init_cred); - cgroup_populate_dir(root_cgrp, true, root->subsys_mask); + cgroup_addrm_files(root_cgrp, NULL, cgroup_base_files, true); + cgroup_populate_dir(root_cgrp, root->subsys_mask); revert_creds(cred); mutex_unlock(&cgroup_root_mutex); mutex_unlock(&cgroup_mutex); @@ -4172,23 +4168,14 @@ static struct cftype cgroup_base_files[] = { }; /** - * cgroup_populate_dir - selectively creation of files in a directory + * cgroup_populate_dir - create subsys files in a cgroup directory * @cgrp: target cgroup - * @base_files: true if the base files should be added * @subsys_mask: mask of the subsystem ids whose files should be added */ -static int cgroup_populate_dir(struct cgroup *cgrp, bool base_files, - unsigned long subsys_mask) +static int cgroup_populate_dir(struct cgroup *cgrp, unsigned long subsys_mask) { - int err; struct cgroup_subsys *ss; - if (base_files) { - err = cgroup_addrm_files(cgrp, NULL, cgroup_base_files, true); - if (err < 0) - return err; - } - /* process cftsets of each subsystem */ for_each_root_subsys(cgrp->root, ss) { struct cftype_set *set; @@ -4410,7 +4397,11 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, } } - err = cgroup_populate_dir(cgrp, true, root->subsys_mask); + err = cgroup_addrm_files(cgrp, NULL, cgroup_base_files, true); + if (err) + goto err_destroy; + + err = cgroup_populate_dir(cgrp, root->subsys_mask); if (err) goto err_destroy; @@ -4566,7 +4557,8 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) * Clear and remove @cgrp directory. The removal puts the base ref * but we aren't quite done with @cgrp yet, so hold onto it. */ - cgroup_clear_dir(cgrp, true, cgrp->root->subsys_mask); + cgroup_clear_dir(cgrp, cgrp->root->subsys_mask); + cgroup_addrm_files(cgrp, NULL, cgroup_base_files, false); dget(d); cgroup_d_remove_dir(d); -- cgit v1.2.3 From bee550994f6b0c1179bd3ccea58dc5c2c4ccf842 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 28 Jun 2013 16:24:11 -0700 Subject: cgroup: update error handling in cgroup_populate_dir() cgroup_populate_dir() didn't use to check whether the actual file creations were successful and could return success with only subset of the requested files created, which is nasty. This patch udpates cgroup_populate_dir() so that it either succeeds with all files or fails with no file. v2: The original patch also converted for_each_root_subsys() usages to for_each_subsys() without explaining why. That part has been moved to a separate patch. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 9835a097f3c0..6b7324431b99 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4171,10 +4171,13 @@ static struct cftype cgroup_base_files[] = { * cgroup_populate_dir - create subsys files in a cgroup directory * @cgrp: target cgroup * @subsys_mask: mask of the subsystem ids whose files should be added + * + * On failure, no file is added. */ static int cgroup_populate_dir(struct cgroup *cgrp, unsigned long subsys_mask) { struct cgroup_subsys *ss; + int ret = 0; /* process cftsets of each subsystem */ for_each_root_subsys(cgrp->root, ss) { @@ -4182,8 +4185,11 @@ static int cgroup_populate_dir(struct cgroup *cgrp, unsigned long subsys_mask) if (!test_bit(ss->subsys_id, &subsys_mask)) continue; - list_for_each_entry(set, &ss->cftsets, node) - cgroup_addrm_files(cgrp, ss, set->cfts, true); + list_for_each_entry(set, &ss->cftsets, node) { + ret = cgroup_addrm_files(cgrp, ss, set->cfts, true); + if (ret < 0) + goto err; + } } /* This cgroup is ready now */ @@ -4201,6 +4207,9 @@ static int cgroup_populate_dir(struct cgroup *cgrp, unsigned long subsys_mask) } return 0; +err: + cgroup_clear_dir(cgrp, subsys_mask); + return ret; } static void css_dput_fn(struct work_struct *work) -- cgit v1.2.3 From b420ba7db15659253d4f286a0ba479d336371999 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 12 Jul 2013 12:34:02 -0700 Subject: cgroup: use for_each_subsys() instead of for_each_root_subsys() in cgroup_populate/clear_dir() rebind_subsystems() will be updated to handle file creations and removals with proper error handling and to do that will need to perform file operations before actually adding the subsystem to the hierarchy. To enable such usage, update cgroup_populate/clear_dir() to use for_each_subsys() instead of for_each_root_subsys() so that they operate on all subsystems specified by @subsys_mask whether that subsystem is currently bound to the hierarchy or not. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 6b7324431b99..8f70dc0c0c79 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -965,10 +965,12 @@ static void cgroup_rm_file(struct cgroup *cgrp, const struct cftype *cft) static void cgroup_clear_dir(struct cgroup *cgrp, unsigned long subsys_mask) { struct cgroup_subsys *ss; + int i; - for_each_root_subsys(cgrp->root, ss) { + for_each_subsys(ss, i) { struct cftype_set *set; - if (!test_bit(ss->subsys_id, &subsys_mask)) + + if (!test_bit(i, &subsys_mask)) continue; list_for_each_entry(set, &ss->cftsets, node) cgroup_addrm_files(cgrp, NULL, set->cfts, false); @@ -4177,12 +4179,13 @@ static struct cftype cgroup_base_files[] = { static int cgroup_populate_dir(struct cgroup *cgrp, unsigned long subsys_mask) { struct cgroup_subsys *ss; - int ret = 0; + int i, ret = 0; /* process cftsets of each subsystem */ - for_each_root_subsys(cgrp->root, ss) { + for_each_subsys(ss, i) { struct cftype_set *set; - if (!test_bit(ss->subsys_id, &subsys_mask)) + + if (!test_bit(i, &subsys_mask)) continue; list_for_each_entry(set, &ss->cftsets, node) { -- cgit v1.2.3 From 3126121fb30941552b1a806c7c2e686bde57e270 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 28 Jun 2013 17:07:30 -0700 Subject: cgroup: make rebind_subsystems() handle file additions and removals with proper error handling Currently, creating and removing cgroup files in the root directory are handled separately from the actual subsystem binding and unbinding which happens in rebind_subsystems(). Also, rebind_subsystems() users aren't handling file creation errors properly. Let's integrate top_cgroup file handling into rebind_subsystems() so that it's simpler to use and everyone handles file creation errors correctly. * On a successful return, rebind_subsystems() is guaranteed to have created all files of the new subsystems and deleted the ones belonging to the removed subsystems. After a failure, no file is created or removed. * cgroup_remount() no longer needs to make explicit populate/clear calls as it's all handled by rebind_subsystems(), and it gets proper error handling automatically. * cgroup_mount() has been updated such that the root dentry and cgroup are linked before rebind_subsystems(). Also, the init_cred dancing and base file handling are moved right above rebind_subsystems() call and proper error handling for the base files is added. While at it, add a comment explaining what's going on with the cred thing. * cgroup_kill_sb() calls rebind_subsystems() to unbind all subsystems which now implies removing all subsystem files which requires the directory's i_mutex. Grab it. This means that files on the root cgroup are removed earlier - they used to be deleted from generic super_block cleanup from vfs. This doesn't lead to any functional difference and it's cleaner to do the clean up explicitly for all files. Combined with the previous changes, this makes all cgroup file creation errors handled correctly. v2: Added comment on init_cred. v3: Li spotted that cgroup_mount() wasn't freeing tmp_links after base file addition failure. Fix it by adding free_tmp_links error handling label. v4: v3 introduced build bugs which got noticed by Fengguang's awesome kbuild test robot. Fixed, and shame on me. Signed-off-by: Tejun Heo Acked-by: Li Zefan Cc: Fengguang Wu --- kernel/cgroup.c | 73 ++++++++++++++++++++++++++++++++------------------------- 1 file changed, 41 insertions(+), 32 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 8f70dc0c0c79..4ec8d2da94d1 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1003,7 +1003,7 @@ static int rebind_subsystems(struct cgroupfs_root *root, { struct cgroup *cgrp = &root->top_cgroup; struct cgroup_subsys *ss; - int i; + int i, ret; BUG_ON(!mutex_is_locked(&cgroup_mutex)); BUG_ON(!mutex_is_locked(&cgroup_root_mutex)); @@ -1028,7 +1028,16 @@ static int rebind_subsystems(struct cgroupfs_root *root, if (root->number_of_cgroups > 1) return -EBUSY; - /* Process each subsystem */ + ret = cgroup_populate_dir(cgrp, added_mask); + if (ret) + return ret; + + /* + * Nothing can fail from this point on. Remove files for the + * removed subsystems and rebind each subsystem. + */ + cgroup_clear_dir(cgrp, removed_mask); + for_each_subsys(ss, i) { unsigned long bit = 1UL << i; @@ -1364,22 +1373,9 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) goto out_unlock; } - /* - * Clear out the files of subsystems that should be removed, do - * this before rebind_subsystems, since rebind_subsystems may - * change this hierarchy's subsys_list. - */ - cgroup_clear_dir(cgrp, removed_mask); - ret = rebind_subsystems(root, added_mask, removed_mask); - if (ret) { - /* rebind_subsystems failed, re-populate the removed files */ - cgroup_populate_dir(cgrp, removed_mask); + if (ret) goto out_unlock; - } - - /* re-populate subsystem files */ - cgroup_populate_dir(cgrp, added_mask); if (opts.release_agent) strcpy(root->release_agent_path, opts.release_agent); @@ -1578,7 +1574,9 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, int ret = 0; struct super_block *sb; struct cgroupfs_root *new_root; + struct list_head tmp_links; struct inode *inode; + const struct cred *cred; /* First find the desired set of subsystems */ mutex_lock(&cgroup_mutex); @@ -1610,10 +1608,8 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, BUG_ON(!root); if (root == opts.new_root) { /* We used the new root structure, so this is a new hierarchy */ - struct list_head tmp_links; struct cgroup *root_cgrp = &root->top_cgroup; struct cgroupfs_root *existing_root; - const struct cred *cred; int i; struct css_set *cset; @@ -1651,26 +1647,37 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, if (ret) goto unlock_drop; + sb->s_root->d_fsdata = root_cgrp; + root_cgrp->dentry = sb->s_root; + + /* + * We're inside get_sb() and will call lookup_one_len() to + * create the root files, which doesn't work if SELinux is + * in use. The following cred dancing somehow works around + * it. See 2ce9738ba ("cgroupfs: use init_cred when + * populating new cgroupfs mount") for more details. + */ + cred = override_creds(&init_cred); + + ret = cgroup_addrm_files(root_cgrp, NULL, cgroup_base_files, true); + if (ret) + goto rm_base_files; + ret = rebind_subsystems(root, root->subsys_mask, 0); - if (ret == -EBUSY) { - free_cgrp_cset_links(&tmp_links); - goto unlock_drop; - } + if (ret) + goto rm_base_files; + + revert_creds(cred); + /* * There must be no failure case after here, since rebinding * takes care of subsystems' refcounts, which are explicitly * dropped in the failure exit path. */ - /* EBUSY should be the only error here */ - BUG_ON(ret); - list_add(&root->root_list, &cgroup_roots); cgroup_root_count++; - sb->s_root->d_fsdata = root_cgrp; - root->top_cgroup.dentry = sb->s_root; - /* Link the top cgroup in this hierarchy into all * the css_set objects */ write_lock(&css_set_lock); @@ -1683,10 +1690,6 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, BUG_ON(!list_empty(&root_cgrp->children)); BUG_ON(root->number_of_cgroups != 1); - cred = override_creds(&init_cred); - cgroup_addrm_files(root_cgrp, NULL, cgroup_base_files, true); - cgroup_populate_dir(root_cgrp, root->subsys_mask); - revert_creds(cred); mutex_unlock(&cgroup_root_mutex); mutex_unlock(&cgroup_mutex); mutex_unlock(&inode->i_mutex); @@ -1715,6 +1718,10 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, kfree(opts.name); return dget(sb->s_root); + rm_base_files: + free_cgrp_cset_links(&tmp_links); + cgroup_addrm_files(&root->top_cgroup, NULL, cgroup_base_files, false); + revert_creds(cred); unlock_drop: cgroup_exit_root_id(root); mutex_unlock(&cgroup_root_mutex); @@ -1741,6 +1748,7 @@ static void cgroup_kill_sb(struct super_block *sb) { BUG_ON(root->number_of_cgroups != 1); BUG_ON(!list_empty(&cgrp->children)); + mutex_lock(&cgrp->dentry->d_inode->i_mutex); mutex_lock(&cgroup_mutex); mutex_lock(&cgroup_root_mutex); @@ -1773,6 +1781,7 @@ static void cgroup_kill_sb(struct super_block *sb) { mutex_unlock(&cgroup_root_mutex); mutex_unlock(&cgroup_mutex); + mutex_unlock(&cgrp->dentry->d_inode->i_mutex); simple_xattrs_free(&cgrp->xattrs); -- cgit v1.2.3 From f172e67cf9d842bc646d0f66792e38435a334b1e Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 28 Jun 2013 17:07:30 -0700 Subject: cgroup: move number_of_cgroups test out of rebind_subsystems() into cgroup_remount() rebind_subsystems() currently fails if the hierarchy has any !root cgroups; however, on the planned unified hierarchy, rebind_subsystems() will be used while populated. Move the test to cgroup_remount(), which is the only place the test is necessary anyway. As it's impossible for the other two callers of rebind_subsystems() to have populated hierarchy, this doesn't make any behavior changes. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 4ec8d2da94d1..c108d3d1ea30 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1021,13 +1021,6 @@ static int rebind_subsystems(struct cgroupfs_root *root, } } - /* Currently we don't handle adding/removing subsystems when - * any child cgroups exist. This is theoretically supportable - * but involves complex error handling, so it's being left until - * later */ - if (root->number_of_cgroups > 1) - return -EBUSY; - ret = cgroup_populate_dir(cgrp, added_mask); if (ret) return ret; @@ -1373,6 +1366,12 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) goto out_unlock; } + /* remounting is not allowed for populated hierarchies */ + if (root->number_of_cgroups > 1) { + ret = -EBUSY; + goto out_unlock; + } + ret = rebind_subsystems(root, added_mask, removed_mask); if (ret) goto out_unlock; -- cgit v1.2.3 From 913ffdb54366f94eec65c656cae8c6e00e1ab1b0 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 11 Jul 2013 16:34:48 -0700 Subject: cgroup: replace task_cgroup_path_from_hierarchy() with task_cgroup_path() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit task_cgroup_path_from_hierarchy() was added for the planned new users and none of the currently planned users wants to know about multiple hierarchies. This patch drops the multiple hierarchy part and makes it always return the path in the first non-dummy hierarchy. As unified hierarchy will always have id 1, this is guaranteed to return the path for the unified hierarchy if mounted; otherwise, it will return the path from the hierarchy which happens to occupy the lowest hierarchy id, which will usually be the first hierarchy mounted after boot. Signed-off-by: Tejun Heo Acked-by: Li Zefan Cc: Lennart Poettering Cc: Kay Sievers Cc: Jan Kaluža --- kernel/cgroup.c | 31 +++++++++++++++++++------------ 1 file changed, 19 insertions(+), 12 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index e5583d10a325..afb8d53ca6c7 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1846,36 +1846,43 @@ out: EXPORT_SYMBOL_GPL(cgroup_path); /** - * task_cgroup_path_from_hierarchy - cgroup path of a task on a hierarchy + * task_cgroup_path - cgroup path of a task in the first cgroup hierarchy * @task: target task - * @hierarchy_id: the hierarchy to look up @task's cgroup from * @buf: the buffer to write the path into * @buflen: the length of the buffer * - * Determine @task's cgroup on the hierarchy specified by @hierarchy_id and - * copy its path into @buf. This function grabs cgroup_mutex and shouldn't - * be used inside locks used by cgroup controller callbacks. + * Determine @task's cgroup on the first (the one with the lowest non-zero + * hierarchy_id) cgroup hierarchy and copy its path into @buf. This + * function grabs cgroup_mutex and shouldn't be used inside locks used by + * cgroup controller callbacks. + * + * Returns 0 on success, fails with -%ENAMETOOLONG if @buflen is too short. */ -int task_cgroup_path_from_hierarchy(struct task_struct *task, int hierarchy_id, - char *buf, size_t buflen) +int task_cgroup_path(struct task_struct *task, char *buf, size_t buflen) { struct cgroupfs_root *root; - struct cgroup *cgrp = NULL; - int ret = -ENOENT; + struct cgroup *cgrp; + int hierarchy_id = 1, ret = 0; + + if (buflen < 2) + return -ENAMETOOLONG; mutex_lock(&cgroup_mutex); - root = idr_find(&cgroup_hierarchy_idr, hierarchy_id); + root = idr_get_next(&cgroup_hierarchy_idr, &hierarchy_id); + if (root) { cgrp = task_cgroup_from_root(task, root); ret = cgroup_path(cgrp, buf, buflen); + } else { + /* if no hierarchy exists, everyone is in "/" */ + memcpy(buf, "/", 2); } mutex_unlock(&cgroup_mutex); - return ret; } -EXPORT_SYMBOL_GPL(task_cgroup_path_from_hierarchy); +EXPORT_SYMBOL_GPL(task_cgroup_path); /* * Control Group taskset -- cgit v1.2.3 From 786e1448d9c5d2a469bcc9d2aecacd418ee1aca0 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 14 Jul 2013 17:50:23 +0400 Subject: cgroup: we can use simple_lookup() now Signed-off-by: Al Viro --- kernel/cgroup.c | 11 +---------- 1 file changed, 1 insertion(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index e5583d10a325..0e0b20b8c5db 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -802,7 +802,6 @@ static struct cgroup *task_cgroup_from_root(struct task_struct *task, */ static int cgroup_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode); -static struct dentry *cgroup_lookup(struct inode *, struct dentry *, unsigned int); static int cgroup_rmdir(struct inode *unused_dir, struct dentry *dentry); static int cgroup_populate_dir(struct cgroup *cgrp, bool base_files, unsigned long subsys_mask); @@ -2642,7 +2641,7 @@ static const struct inode_operations cgroup_file_inode_operations = { }; static const struct inode_operations cgroup_dir_inode_operations = { - .lookup = cgroup_lookup, + .lookup = simple_lookup, .mkdir = cgroup_mkdir, .rmdir = cgroup_rmdir, .rename = cgroup_rename, @@ -2652,14 +2651,6 @@ static const struct inode_operations cgroup_dir_inode_operations = { .removexattr = cgroup_removexattr, }; -static struct dentry *cgroup_lookup(struct inode *dir, struct dentry *dentry, unsigned int flags) -{ - if (dentry->d_name.len > NAME_MAX) - return ERR_PTR(-ENAMETOOLONG); - d_add(dentry, NULL); - return NULL; -} - /* * Check if a file is a control file */ -- cgit v1.2.3 From e5248a111bf4048a9f3fab1a9c94c4630a10592a Mon Sep 17 00:00:00 2001 From: Liu ShuoX Date: Thu, 11 Jul 2013 16:03:45 +0800 Subject: PM / Sleep: avoid 'autosleep' in shutdown progress Prevent automatic system suspend from happening during system shutdown by making try_to_suspend() check system_state and return immediately if it is not SYSTEM_RUNNING. This prevents the following breakage from happening (scenario from Zhang Yanmin): Kernel starts shutdown and calls all device driver's shutdown callback. When a driver's shutdown is called, the last wakelock is released and suspend-to-ram starts. However, as some driver's shut down callbacks already shut down devices and disabled runtime pm, the suspend-to-ram calls driver's suspend callback without noticing that device is already off and causes crash. [rjw: Changelog] Signed-off-by: Liu ShuoX Cc: 3.5+ Signed-off-by: Rafael J. Wysocki --- kernel/power/autosleep.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/power/autosleep.c b/kernel/power/autosleep.c index c6422ffeda9a..9012ecf7b814 100644 --- a/kernel/power/autosleep.c +++ b/kernel/power/autosleep.c @@ -32,7 +32,8 @@ static void try_to_suspend(struct work_struct *work) mutex_lock(&autosleep_lock); - if (!pm_save_wakeup_count(initial_count)) { + if (!pm_save_wakeup_count(initial_count) || + system_state != SYSTEM_RUNNING) { mutex_unlock(&autosleep_lock); goto out; } -- cgit v1.2.3 From 49fb4c6290c70c418a5c25eee996d6b55ea132d6 Mon Sep 17 00:00:00 2001 From: Paul Gortmaker Date: Wed, 19 Jun 2013 14:52:21 -0400 Subject: rcu: delete __cpuinit usage from all rcu files The __cpuinit type of throwaway sections might have made sense some time ago when RAM was more constrained, but now the savings do not offset the cost and complications. For example, the fix in commit 5e427ec2d0 ("x86: Fix bit corruption at CPU resume time") is a good example of the nasty type of bugs that can be created with improper use of the various __init prefixes. After a discussion on LKML[1] it was decided that cpuinit should go the way of devinit and be phased out. Once all the users are gone, we can then finally remove the macros themselves from linux/init.h. This removes all the drivers/rcu uses of the __cpuinit macros from all C files. [1] https://lkml.org/lkml/2013/5/20/589 Cc: "Paul E. McKenney" Cc: Josh Triplett Cc: Dipankar Sarma Reviewed-by: Josh Triplett Signed-off-by: Paul Gortmaker --- kernel/rcutorture.c | 6 +++--- kernel/rcutree.c | 6 +++--- kernel/rcutree.h | 4 ++-- kernel/rcutree_plugin.h | 6 +++--- 4 files changed, 11 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c index b1fa5510388d..f4871e52c546 100644 --- a/kernel/rcutorture.c +++ b/kernel/rcutorture.c @@ -1476,7 +1476,7 @@ rcu_torture_shutdown(void *arg) * Execute random CPU-hotplug operations at the interval specified * by the onoff_interval. */ -static int __cpuinit +static int rcu_torture_onoff(void *arg) { int cpu; @@ -1558,7 +1558,7 @@ rcu_torture_onoff(void *arg) return 0; } -static int __cpuinit +static int rcu_torture_onoff_init(void) { int ret; @@ -1601,7 +1601,7 @@ static void rcu_torture_onoff_cleanup(void) * CPU-stall kthread. It waits as specified by stall_cpu_holdoff, then * induces a CPU stall for the time specified by stall_cpu. */ -static int __cpuinit rcu_torture_stall(void *args) +static int rcu_torture_stall(void *args) { unsigned long stop_at; diff --git a/kernel/rcutree.c b/kernel/rcutree.c index e08abb9461ac..068de3a93606 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -2910,7 +2910,7 @@ rcu_boot_init_percpu_data(int cpu, struct rcu_state *rsp) * can accept some slop in the rsp->completed access due to the fact * that this CPU cannot possibly have any RCU callbacks in flight yet. */ -static void __cpuinit +static void rcu_init_percpu_data(int cpu, struct rcu_state *rsp, int preemptible) { unsigned long flags; @@ -2962,7 +2962,7 @@ rcu_init_percpu_data(int cpu, struct rcu_state *rsp, int preemptible) mutex_unlock(&rsp->onoff_mutex); } -static void __cpuinit rcu_prepare_cpu(int cpu) +static void rcu_prepare_cpu(int cpu) { struct rcu_state *rsp; @@ -2974,7 +2974,7 @@ static void __cpuinit rcu_prepare_cpu(int cpu) /* * Handle CPU online/offline notification events. */ -static int __cpuinit rcu_cpu_notify(struct notifier_block *self, +static int rcu_cpu_notify(struct notifier_block *self, unsigned long action, void *hcpu) { long cpu = (long)hcpu; diff --git a/kernel/rcutree.h b/kernel/rcutree.h index 4a39d364493c..b3832581043c 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -521,10 +521,10 @@ static void invoke_rcu_callbacks_kthread(void); static bool rcu_is_callbacks_kthread(void); #ifdef CONFIG_RCU_BOOST static void rcu_preempt_do_callbacks(void); -static int __cpuinit rcu_spawn_one_boost_kthread(struct rcu_state *rsp, +static int rcu_spawn_one_boost_kthread(struct rcu_state *rsp, struct rcu_node *rnp); #endif /* #ifdef CONFIG_RCU_BOOST */ -static void __cpuinit rcu_prepare_kthreads(int cpu); +static void rcu_prepare_kthreads(int cpu); static void rcu_cleanup_after_idle(int cpu); static void rcu_prepare_for_idle(int cpu); static void rcu_idle_count_callbacks_posted(void); diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 63098a59216e..769e12e3151b 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -1352,7 +1352,7 @@ static void rcu_preempt_boost_start_gp(struct rcu_node *rnp) * already exist. We only create this kthread for preemptible RCU. * Returns zero if all is well, a negated errno otherwise. */ -static int __cpuinit rcu_spawn_one_boost_kthread(struct rcu_state *rsp, +static int rcu_spawn_one_boost_kthread(struct rcu_state *rsp, struct rcu_node *rnp) { int rnp_index = rnp - &rsp->node[0]; @@ -1507,7 +1507,7 @@ static int __init rcu_spawn_kthreads(void) } early_initcall(rcu_spawn_kthreads); -static void __cpuinit rcu_prepare_kthreads(int cpu) +static void rcu_prepare_kthreads(int cpu) { struct rcu_data *rdp = per_cpu_ptr(rcu_state->rda, cpu); struct rcu_node *rnp = rdp->mynode; @@ -1549,7 +1549,7 @@ static int __init rcu_scheduler_really_started(void) } early_initcall(rcu_scheduler_really_started); -static void __cpuinit rcu_prepare_kthreads(int cpu) +static void rcu_prepare_kthreads(int cpu) { } -- cgit v1.2.3 From 0db0628d90125193280eabb501c94feaf48fa9ab Mon Sep 17 00:00:00 2001 From: Paul Gortmaker Date: Wed, 19 Jun 2013 14:53:51 -0400 Subject: kernel: delete __cpuinit usage from all core kernel files The __cpuinit type of throwaway sections might have made sense some time ago when RAM was more constrained, but now the savings do not offset the cost and complications. For example, the fix in commit 5e427ec2d0 ("x86: Fix bit corruption at CPU resume time") is a good example of the nasty type of bugs that can be created with improper use of the various __init prefixes. After a discussion on LKML[1] it was decided that cpuinit should go the way of devinit and be phased out. Once all the users are gone, we can then finally remove the macros themselves from linux/init.h. This removes all the uses of the __cpuinit macros from C files in the core kernel directories (kernel, init, lib, mm, and include) that don't really have a specific maintainer. [1] https://lkml.org/lkml/2013/5/20/589 Signed-off-by: Paul Gortmaker --- kernel/cpu.c | 6 +++--- kernel/events/core.c | 4 ++-- kernel/fork.c | 2 +- kernel/hrtimer.c | 6 +++--- kernel/printk.c | 2 +- kernel/profile.c | 2 +- kernel/relay.c | 2 +- kernel/sched/core.c | 12 ++++++------ kernel/sched/fair.c | 2 +- kernel/smp.c | 2 +- kernel/smpboot.c | 2 +- kernel/softirq.c | 8 ++++---- kernel/time/tick-sched.c | 2 +- kernel/timer.c | 10 +++++----- kernel/workqueue.c | 4 ++-- 15 files changed, 33 insertions(+), 33 deletions(-) (limited to 'kernel') diff --git a/kernel/cpu.c b/kernel/cpu.c index 198a38883e64..b2b227b82123 100644 --- a/kernel/cpu.c +++ b/kernel/cpu.c @@ -366,7 +366,7 @@ EXPORT_SYMBOL(cpu_down); #endif /*CONFIG_HOTPLUG_CPU*/ /* Requires cpu_add_remove_lock to be held */ -static int __cpuinit _cpu_up(unsigned int cpu, int tasks_frozen) +static int _cpu_up(unsigned int cpu, int tasks_frozen) { int ret, nr_calls = 0; void *hcpu = (void *)(long)cpu; @@ -419,7 +419,7 @@ out: return ret; } -int __cpuinit cpu_up(unsigned int cpu) +int cpu_up(unsigned int cpu) { int err = 0; @@ -618,7 +618,7 @@ core_initcall(cpu_hotplug_pm_sync_init); * It must be called by the arch code on the new cpu, before the new cpu * enables interrupts and before the "boot" cpu returns from __cpu_up(). */ -void __cpuinit notify_cpu_starting(unsigned int cpu) +void notify_cpu_starting(unsigned int cpu) { unsigned long val = CPU_STARTING; diff --git a/kernel/events/core.c b/kernel/events/core.c index eba8fb5834ae..f3e9dce39bc9 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -7630,7 +7630,7 @@ static void __init perf_event_init_all_cpus(void) } } -static void __cpuinit perf_event_init_cpu(int cpu) +static void perf_event_init_cpu(int cpu) { struct swevent_htable *swhash = &per_cpu(swevent_htable, cpu); @@ -7719,7 +7719,7 @@ static struct notifier_block perf_reboot_notifier = { .priority = INT_MIN, }; -static int __cpuinit +static int perf_cpu_notify(struct notifier_block *self, unsigned long action, void *hcpu) { unsigned int cpu = (long)hcpu; diff --git a/kernel/fork.c b/kernel/fork.c index 66635c80a813..403d2bb8a968 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1546,7 +1546,7 @@ static inline void init_idle_pids(struct pid_link *links) } } -struct task_struct * __cpuinit fork_idle(int cpu) +struct task_struct *fork_idle(int cpu) { struct task_struct *task; task = copy_process(CLONE_VM, 0, 0, NULL, &init_struct_pid, 0); diff --git a/kernel/hrtimer.c b/kernel/hrtimer.c index f0f4fe29cd21..383319bae3f7 100644 --- a/kernel/hrtimer.c +++ b/kernel/hrtimer.c @@ -1659,7 +1659,7 @@ SYSCALL_DEFINE2(nanosleep, struct timespec __user *, rqtp, /* * Functions related to boot-time initialization: */ -static void __cpuinit init_hrtimers_cpu(int cpu) +static void init_hrtimers_cpu(int cpu) { struct hrtimer_cpu_base *cpu_base = &per_cpu(hrtimer_bases, cpu); int i; @@ -1740,7 +1740,7 @@ static void migrate_hrtimers(int scpu) #endif /* CONFIG_HOTPLUG_CPU */ -static int __cpuinit hrtimer_cpu_notify(struct notifier_block *self, +static int hrtimer_cpu_notify(struct notifier_block *self, unsigned long action, void *hcpu) { int scpu = (long)hcpu; @@ -1773,7 +1773,7 @@ static int __cpuinit hrtimer_cpu_notify(struct notifier_block *self, return NOTIFY_OK; } -static struct notifier_block __cpuinitdata hrtimers_nb = { +static struct notifier_block hrtimers_nb = { .notifier_call = hrtimer_cpu_notify, }; diff --git a/kernel/printk.c b/kernel/printk.c index d37d45c90ae6..69b0890ed7e5 100644 --- a/kernel/printk.c +++ b/kernel/printk.c @@ -1921,7 +1921,7 @@ void resume_console(void) * called when a new CPU comes online (or fails to come up), and ensures * that any such output gets printed. */ -static int __cpuinit console_cpu_notify(struct notifier_block *self, +static int console_cpu_notify(struct notifier_block *self, unsigned long action, void *hcpu) { switch (action) { diff --git a/kernel/profile.c b/kernel/profile.c index 0bf400737660..6631e1ef55ab 100644 --- a/kernel/profile.c +++ b/kernel/profile.c @@ -331,7 +331,7 @@ out: put_cpu(); } -static int __cpuinit profile_cpu_callback(struct notifier_block *info, +static int profile_cpu_callback(struct notifier_block *info, unsigned long action, void *__cpu) { int node, cpu = (unsigned long)__cpu; diff --git a/kernel/relay.c b/kernel/relay.c index b91488ba2e5a..5001c9887db1 100644 --- a/kernel/relay.c +++ b/kernel/relay.c @@ -516,7 +516,7 @@ static void setup_callbacks(struct rchan *chan, * * Returns the success/failure of the operation. (%NOTIFY_OK, %NOTIFY_BAD) */ -static int __cpuinit relay_hotcpu_callback(struct notifier_block *nb, +static int relay_hotcpu_callback(struct notifier_block *nb, unsigned long action, void *hcpu) { diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 0d8eb4525e76..b7c32cb7bfeb 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -4133,7 +4133,7 @@ void show_state_filter(unsigned long state_filter) debug_show_all_locks(); } -void __cpuinit init_idle_bootup_task(struct task_struct *idle) +void init_idle_bootup_task(struct task_struct *idle) { idle->sched_class = &idle_sched_class; } @@ -4146,7 +4146,7 @@ void __cpuinit init_idle_bootup_task(struct task_struct *idle) * NOTE: this function does not set the idle thread's NEED_RESCHED * flag, to make booting more robust. */ -void __cpuinit init_idle(struct task_struct *idle, int cpu) +void init_idle(struct task_struct *idle, int cpu) { struct rq *rq = cpu_rq(cpu); unsigned long flags; @@ -4630,7 +4630,7 @@ static void set_rq_offline(struct rq *rq) * migration_call - callback that gets triggered when a CPU is added. * Here we can start up the necessary migration thread for the new CPU. */ -static int __cpuinit +static int migration_call(struct notifier_block *nfb, unsigned long action, void *hcpu) { int cpu = (long)hcpu; @@ -4684,12 +4684,12 @@ migration_call(struct notifier_block *nfb, unsigned long action, void *hcpu) * happens before everything else. This has to be lower priority than * the notifier in the perf_event subsystem, though. */ -static struct notifier_block __cpuinitdata migration_notifier = { +static struct notifier_block migration_notifier = { .notifier_call = migration_call, .priority = CPU_PRI_MIGRATION, }; -static int __cpuinit sched_cpu_active(struct notifier_block *nfb, +static int sched_cpu_active(struct notifier_block *nfb, unsigned long action, void *hcpu) { switch (action & ~CPU_TASKS_FROZEN) { @@ -4702,7 +4702,7 @@ static int __cpuinit sched_cpu_active(struct notifier_block *nfb, } } -static int __cpuinit sched_cpu_inactive(struct notifier_block *nfb, +static int sched_cpu_inactive(struct notifier_block *nfb, unsigned long action, void *hcpu) { switch (action & ~CPU_TASKS_FROZEN) { diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index f77f9c527449..bb456f44b7b1 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -5506,7 +5506,7 @@ void nohz_balance_enter_idle(int cpu) set_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu)); } -static int __cpuinit sched_ilb_notifier(struct notifier_block *nfb, +static int sched_ilb_notifier(struct notifier_block *nfb, unsigned long action, void *hcpu) { switch (action & ~CPU_TASKS_FROZEN) { diff --git a/kernel/smp.c b/kernel/smp.c index 4dba0f7b72ad..fe9f773d7114 100644 --- a/kernel/smp.c +++ b/kernel/smp.c @@ -73,7 +73,7 @@ hotplug_cfd(struct notifier_block *nfb, unsigned long action, void *hcpu) return NOTIFY_OK; } -static struct notifier_block __cpuinitdata hotplug_cfd_notifier = { +static struct notifier_block hotplug_cfd_notifier = { .notifier_call = hotplug_cfd, }; diff --git a/kernel/smpboot.c b/kernel/smpboot.c index 02fc5c933673..eb89e1807408 100644 --- a/kernel/smpboot.c +++ b/kernel/smpboot.c @@ -24,7 +24,7 @@ */ static DEFINE_PER_CPU(struct task_struct *, idle_threads); -struct task_struct * __cpuinit idle_thread_get(unsigned int cpu) +struct task_struct *idle_thread_get(unsigned int cpu) { struct task_struct *tsk = per_cpu(idle_threads, cpu); diff --git a/kernel/softirq.c b/kernel/softirq.c index ca25e6e704a2..be3d3514c325 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -699,7 +699,7 @@ void send_remote_softirq(struct call_single_data *cp, int cpu, int softirq) } EXPORT_SYMBOL(send_remote_softirq); -static int __cpuinit remote_softirq_cpu_notify(struct notifier_block *self, +static int remote_softirq_cpu_notify(struct notifier_block *self, unsigned long action, void *hcpu) { /* @@ -728,7 +728,7 @@ static int __cpuinit remote_softirq_cpu_notify(struct notifier_block *self, return NOTIFY_OK; } -static struct notifier_block __cpuinitdata remote_softirq_cpu_notifier = { +static struct notifier_block remote_softirq_cpu_notifier = { .notifier_call = remote_softirq_cpu_notify, }; @@ -830,7 +830,7 @@ static void takeover_tasklets(unsigned int cpu) } #endif /* CONFIG_HOTPLUG_CPU */ -static int __cpuinit cpu_callback(struct notifier_block *nfb, +static int cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu) { @@ -845,7 +845,7 @@ static int __cpuinit cpu_callback(struct notifier_block *nfb, return NOTIFY_OK; } -static struct notifier_block __cpuinitdata cpu_nfb = { +static struct notifier_block cpu_nfb = { .notifier_call = cpu_callback }; diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 69601726a745..e80183f4a6c4 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -298,7 +298,7 @@ static int __init tick_nohz_full_setup(char *str) } __setup("nohz_full=", tick_nohz_full_setup); -static int __cpuinit tick_nohz_cpu_down_callback(struct notifier_block *nfb, +static int tick_nohz_cpu_down_callback(struct notifier_block *nfb, unsigned long action, void *hcpu) { diff --git a/kernel/timer.c b/kernel/timer.c index 15bc1b41021d..4296d13db3d1 100644 --- a/kernel/timer.c +++ b/kernel/timer.c @@ -1505,11 +1505,11 @@ signed long __sched schedule_timeout_uninterruptible(signed long timeout) } EXPORT_SYMBOL(schedule_timeout_uninterruptible); -static int __cpuinit init_timers_cpu(int cpu) +static int init_timers_cpu(int cpu) { int j; struct tvec_base *base; - static char __cpuinitdata tvec_base_done[NR_CPUS]; + static char tvec_base_done[NR_CPUS]; if (!tvec_base_done[cpu]) { static char boot_done; @@ -1577,7 +1577,7 @@ static void migrate_timer_list(struct tvec_base *new_base, struct list_head *hea } } -static void __cpuinit migrate_timers(int cpu) +static void migrate_timers(int cpu) { struct tvec_base *old_base; struct tvec_base *new_base; @@ -1610,7 +1610,7 @@ static void __cpuinit migrate_timers(int cpu) } #endif /* CONFIG_HOTPLUG_CPU */ -static int __cpuinit timer_cpu_notify(struct notifier_block *self, +static int timer_cpu_notify(struct notifier_block *self, unsigned long action, void *hcpu) { long cpu = (long)hcpu; @@ -1635,7 +1635,7 @@ static int __cpuinit timer_cpu_notify(struct notifier_block *self, return NOTIFY_OK; } -static struct notifier_block __cpuinitdata timers_nb = { +static struct notifier_block timers_nb = { .notifier_call = timer_cpu_notify, }; diff --git a/kernel/workqueue.c b/kernel/workqueue.c index f02c4a4a0c3c..0b72e816b8d0 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -4644,7 +4644,7 @@ static void restore_unbound_workers_cpumask(struct worker_pool *pool, int cpu) * Workqueues should be brought up before normal priority CPU notifiers. * This will be registered high priority CPU notifier. */ -static int __cpuinit workqueue_cpu_up_callback(struct notifier_block *nfb, +static int workqueue_cpu_up_callback(struct notifier_block *nfb, unsigned long action, void *hcpu) { @@ -4697,7 +4697,7 @@ static int __cpuinit workqueue_cpu_up_callback(struct notifier_block *nfb, * Workqueues should be brought down after normal priority CPU notifiers. * This will be registered as low priority CPU notifier. */ -static int __cpuinit workqueue_cpu_down_callback(struct notifier_block *nfb, +static int workqueue_cpu_down_callback(struct notifier_block *nfb, unsigned long action, void *hcpu) { -- cgit v1.2.3 From 991821c86c2fb6cc4104ce679247864dbc070a83 Mon Sep 17 00:00:00 2001 From: "zhangwei(Jovi)" Date: Mon, 15 Jul 2013 16:32:34 +0800 Subject: tracing: Use correct config guard CONFIG_STACK_TRACER We should use CONFIG_STACK_TRACER to guard readme text of stack tracer related file, not CONFIG_STACKTRACE. Link: http://lkml.kernel.org/r/51E3B3A2.8080609@huawei.com Signed-off-by: zhangwei(Jovi) Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 0cd500bffd9b..25b91afc29e0 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -3537,14 +3537,14 @@ static const char readme_msg[] = "\n snapshot\t\t- Like 'trace' but shows the content of the static snapshot buffer\n" "\t\t\t Read the contents for more information\n" #endif -#ifdef CONFIG_STACKTRACE +#ifdef CONFIG_STACK_TRACER " stack_trace\t\t- Shows the max stack trace when active\n" " stack_max_size\t- Shows current max stack size that was traced\n" "\t\t\t Write into this file to reset the max size (trigger a new trace)\n" #ifdef CONFIG_DYNAMIC_FTRACE " stack_trace_filter\t- Like set_ftrace_filter but limits what stack_trace traces\n" #endif -#endif /* CONFIG_STACKTRACE */ +#endif /* CONFIG_STACK_TRACER */ ; static ssize_t -- cgit v1.2.3 From 1d5be6b287c8efc879fbe578e2b7bc8f7a38f313 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 12 Jul 2013 13:38:17 -0700 Subject: cgroup: move module ref handling into rebind_subsystems() Module ref handling in cgroup is rather weird. parse_cgroupfs_options() grabs all the modules for the specified subsystems. A module ref is kept if the specified subsystem is newly bound to the hierarchy. If not, or the operation fails, the refs are dropped. This scatters module ref handling across multiple functions making it difficult to track. It also make the function nasty to use for dynamic subsystem binding which is necessary for the planned unified hierarchy. There's nothing which requires the subsystem modules to be pinned between parse_cgroupfs_options() and rebind_subsystems() in both mount and remount paths. parse_cgroupfs_options() can just parse and rebind_subsystems() can handle pinning the subsystems that it wants to bind, which is a natural part of its task - binding - anyway. Move module ref handling into rebind_subsystems() which makes the code a lot simpler - modules are gotten iff it's gonna be bound and put iff unbound or binding fails. v2: Li pointed out that if a controller module is unloaded between parsing and binding, rebind_subsystems() won't notice the missing controller as it only iterates through existing controllers. Fix it by updating rebind_subsystems() to compare @added_mask to @pinned and fail with -ENOENT if they don't match. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 93 +++++++++++++++++---------------------------------------- 1 file changed, 28 insertions(+), 65 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index c108d3d1ea30..2a8cf1a7d2f4 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1003,6 +1003,7 @@ static int rebind_subsystems(struct cgroupfs_root *root, { struct cgroup *cgrp = &root->top_cgroup; struct cgroup_subsys *ss; + unsigned long pinned = 0; int i, ret; BUG_ON(!mutex_is_locked(&cgroup_mutex)); @@ -1010,20 +1011,32 @@ static int rebind_subsystems(struct cgroupfs_root *root, /* Check that any added subsystems are currently free */ for_each_subsys(ss, i) { - unsigned long bit = 1UL << i; - - if (!(bit & added_mask)) + if (!(added_mask & (1 << i))) continue; + /* is the subsystem mounted elsewhere? */ if (ss->root != &cgroup_dummy_root) { - /* Subsystem isn't free */ - return -EBUSY; + ret = -EBUSY; + goto out_put; + } + + /* pin the module */ + if (!try_module_get(ss->module)) { + ret = -ENOENT; + goto out_put; } + pinned |= 1 << i; + } + + /* subsys could be missing if unloaded between parsing and here */ + if (added_mask != pinned) { + ret = -ENOENT; + goto out_put; } ret = cgroup_populate_dir(cgrp, added_mask); if (ret) - return ret; + goto out_put; /* * Nothing can fail from this point on. Remove files for the @@ -1067,11 +1080,6 @@ static int rebind_subsystems(struct cgroupfs_root *root, } else if (bit & root->subsys_mask) { /* Subsystem state should already exist */ BUG_ON(!cgrp->subsys[i]); - /* - * a refcount was taken, but we already had one, so - * drop the extra reference. - */ - module_put(ss->module); #ifdef CONFIG_MODULE_UNLOAD BUG_ON(ss->module && !module_refcount(ss->module)); #endif @@ -1088,6 +1096,12 @@ static int rebind_subsystems(struct cgroupfs_root *root, root->flags |= CGRP_ROOT_SUBSYS_BOUND; return 0; + +out_put: + for_each_subsys(ss, i) + if (pinned & (1 << i)) + module_put(ss->module); + return ret; } static int cgroup_show_options(struct seq_file *seq, struct dentry *dentry) @@ -1138,7 +1152,6 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) char *token, *o = data; bool all_ss = false, one_ss = false; unsigned long mask = (unsigned long)-1; - bool module_pin_failed = false; struct cgroup_subsys *ss; int i; @@ -1281,52 +1294,9 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) if (!opts->subsys_mask && !opts->name) return -EINVAL; - /* - * Grab references on all the modules we'll need, so the subsystems - * don't dance around before rebind_subsystems attaches them. This may - * take duplicate reference counts on a subsystem that's already used, - * but rebind_subsystems handles this case. - */ - for_each_subsys(ss, i) { - if (!(opts->subsys_mask & (1UL << i))) - continue; - if (!try_module_get(cgroup_subsys[i]->module)) { - module_pin_failed = true; - break; - } - } - if (module_pin_failed) { - /* - * oops, one of the modules was going away. this means that we - * raced with a module_delete call, and to the user this is - * essentially a "subsystem doesn't exist" case. - */ - for (i--; i >= 0; i--) { - /* drop refcounts only on the ones we took */ - unsigned long bit = 1UL << i; - - if (!(bit & opts->subsys_mask)) - continue; - module_put(cgroup_subsys[i]->module); - } - return -ENOENT; - } - return 0; } -static void drop_parsed_module_refcounts(unsigned long subsys_mask) -{ - struct cgroup_subsys *ss; - int i; - - mutex_lock(&cgroup_mutex); - for_each_subsys(ss, i) - if (subsys_mask & (1UL << i)) - module_put(cgroup_subsys[i]->module); - mutex_unlock(&cgroup_mutex); -} - static int cgroup_remount(struct super_block *sb, int *flags, char *data) { int ret = 0; @@ -1384,8 +1354,6 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) mutex_unlock(&cgroup_root_mutex); mutex_unlock(&cgroup_mutex); mutex_unlock(&cgrp->dentry->d_inode->i_mutex); - if (ret) - drop_parsed_module_refcounts(opts.subsys_mask); return ret; } @@ -1591,7 +1559,7 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, new_root = cgroup_root_from_opts(&opts); if (IS_ERR(new_root)) { ret = PTR_ERR(new_root); - goto drop_modules; + goto out_err; } opts.new_root = new_root; @@ -1600,7 +1568,7 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, if (IS_ERR(sb)) { ret = PTR_ERR(sb); cgroup_free_root(opts.new_root); - goto drop_modules; + goto out_err; } root = sb->s_fs_info; @@ -1708,9 +1676,6 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, pr_warning("cgroup: new mount options do not match the existing superblock, will be ignored\n"); } } - - /* no subsys rebinding, so refcounts don't change */ - drop_parsed_module_refcounts(opts.subsys_mask); } kfree(opts.release_agent); @@ -1728,8 +1693,6 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, mutex_unlock(&inode->i_mutex); drop_new_super: deactivate_locked_super(sb); - drop_modules: - drop_parsed_module_refcounts(opts.subsys_mask); out_err: kfree(opts.release_agent); kfree(opts.name); @@ -4837,7 +4800,7 @@ void cgroup_unload_subsys(struct cgroup_subsys *ss) /* * we shouldn't be called if the subsystem is in use, and the use of - * try_module_get in parse_cgroupfs_options should ensure that it + * try_module_get() in rebind_subsystems() should ensure that it * doesn't start being used while we're killing it off. */ BUG_ON(ss->root != &cgroup_dummy_root); -- cgit v1.2.3 From a698b4488ab98deef6c3beeba3e27fea17650132 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 28 Jun 2013 21:08:27 -0700 Subject: cgroup: remove gratuituous BUG_ON()s from rebind_subsystems() rebind_subsystems() performs santiy checks even on subsystems which aren't specified to be added or removed and the checks aren't all that useful given that these are in a very cold path while the violations they check would trip up in much hotter paths. Let's remove these from rebind_subsystems(). Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 9 --------- 1 file changed, 9 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 2a8cf1a7d2f4..345fac8e4fba 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1077,15 +1077,6 @@ static int rebind_subsystems(struct cgroupfs_root *root, /* subsystem is now free - drop reference on module */ module_put(ss->module); root->subsys_mask &= ~bit; - } else if (bit & root->subsys_mask) { - /* Subsystem state should already exist */ - BUG_ON(!cgrp->subsys[i]); -#ifdef CONFIG_MODULE_UNLOAD - BUG_ON(ss->module && !module_refcount(ss->module)); -#endif - } else { - /* Subsystem state shouldn't exist */ - BUG_ON(cgrp->subsys[i]); } } -- cgit v1.2.3 From b9b3259746d77f4fcb786e2a43c25bcc40773755 Mon Sep 17 00:00:00 2001 From: Greg Kroah-Hartman Date: Sun, 14 Jul 2013 16:05:51 -0700 Subject: sysfs.h: add __ATTR_RW() macro A number of parts of the kernel created their own version of this, might as well have the sysfs core provide it instead. Reviewed-by: Guenter Roeck Tested-by: Guenter Roeck Signed-off-by: Greg Kroah-Hartman --- kernel/events/core.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index eba8fb5834ae..dd9878029d1f 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -6234,8 +6234,6 @@ perf_event_mux_interval_ms_store(struct device *dev, return count; } -#define __ATTR_RW(attr) __ATTR(attr, 0644, attr##_show, attr##_store) - static struct device_attribute pmu_dev_attrs[] = { __ATTR_RO(type), __ATTR_RW(perf_event_mux_interval_ms), -- cgit v1.2.3 From fd4363fff3d96795d3feb1b3fb48ce590f186bdd Mon Sep 17 00:00:00 2001 From: Jiri Kosina Date: Fri, 12 Jul 2013 11:21:48 +0200 Subject: x86: Introduce int3 (breakpoint)-based instruction patching Introduce a method for run-time instruction patching on a live SMP kernel based on int3 breakpoint, completely avoiding the need for stop_machine(). The way this is achieved: - add a int3 trap to the address that will be patched - sync cores - update all but the first byte of the patched range - sync cores - replace the first byte (int3) by the first byte of replacing opcode - sync cores According to http://lkml.indiana.edu/hypermail/linux/kernel/1001.1/01530.html synchronization after replacing "all but first" instructions should not be necessary (on Intel hardware), as the syncing after the subsequent patching of the first byte provides enough safety. But there's not only Intel HW out there, and we'd rather be on a safe side. If any CPU instruction execution would collide with the patching, it'd be trapped by the int3 breakpoint and redirected to the provided "handler" (which would typically mean just skipping over the patched region, acting as "nop" has been there, in case we are doing nop -> jump and jump -> nop transitions). Ftrace has been using this very technique since 08d636b ("ftrace/x86: Have arch x86_64 use breakpoints instead of stop machine") for ages already, and jump labels are another obvious potential user of this. Based on activities of Masami Hiramatsu a few years ago. Reviewed-by: Steven Rostedt Reviewed-by: Masami Hiramatsu Signed-off-by: Jiri Kosina Link: http://lkml.kernel.org/r/alpine.LNX.2.00.1307121102440.29788@pobox.suse.cz Signed-off-by: H. Peter Anvin --- kernel/kprobes.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/kprobes.c b/kernel/kprobes.c index 6e33498d665c..b58b490fa439 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -1709,7 +1709,7 @@ EXPORT_SYMBOL_GPL(unregister_kprobes); static struct notifier_block kprobe_exceptions_nb = { .notifier_call = kprobe_exceptions_notify, - .priority = 0x7fffffff /* we need to be notified first */ + .priority = 0x7ffffff0 /* High priority, but not first. */ }; unsigned long __weak arch_deref_entry_point(void *entry) -- cgit v1.2.3 From e69f61862ab833e9b8d3c15b6ce07fd69f3bfecc Mon Sep 17 00:00:00 2001 From: Yacine Belkadi Date: Fri, 12 Jul 2013 20:45:47 +0200 Subject: sched: Fix some kernel-doc warnings When building the htmldocs (in verbose mode), scripts/kernel-doc reports the follwing type of warnings: Warning(kernel/sched/core.c:936): No description found for return value of 'task_curr' ... Fix those by: - adding the missing descriptions - using "Return" sections for the descriptions Signed-off-by: Yacine Belkadi Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1373654747-2389-1-git-send-email-yacine.belkadi.1@gmail.com [ While at it, fix the cpupri_set() explanation. ] Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 82 ++++++++++++++++++++++++++++++++++++++------------- kernel/sched/cpupri.c | 4 +-- kernel/sched/fair.c | 9 ++++-- 3 files changed, 70 insertions(+), 25 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 0d8eb4525e76..4c3967f91e20 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -933,6 +933,8 @@ static int effective_prio(struct task_struct *p) /** * task_curr - is this task currently executing on a CPU? * @p: the task in question. + * + * Return: 1 if the task is currently executing. 0 otherwise. */ inline int task_curr(const struct task_struct *p) { @@ -1482,7 +1484,7 @@ static void ttwu_queue(struct task_struct *p, int cpu) * the simpler "current->state = TASK_RUNNING" to mark yourself * runnable without the overhead of this. * - * Returns %true if @p was woken up, %false if it was already running + * Return: %true if @p was woken up, %false if it was already running. * or @state didn't match @p's state. */ static int @@ -1577,8 +1579,9 @@ out: * @p: The process to be woken up. * * Attempt to wake up the nominated process and move it to the set of runnable - * processes. Returns 1 if the process was woken up, 0 if it was already - * running. + * processes. + * + * Return: 1 if the process was woken up, 0 if it was already running. * * It may be assumed that this function implies a write memory barrier before * changing the task state if and only if any tasks are woken up. @@ -2191,6 +2194,8 @@ void scheduler_tick(void) * This makes sure that uptime, CFS vruntime, load * balancing, etc... continue to move forward, even * with a very low granularity. + * + * Return: Maximum deferment in nanoseconds. */ u64 scheduler_tick_max_deferment(void) { @@ -2796,8 +2801,8 @@ EXPORT_SYMBOL(wait_for_completion); * specified timeout to expire. The timeout is in jiffies. It is not * interruptible. * - * The return value is 0 if timed out, and positive (at least 1, or number of - * jiffies left till timeout) if completed. + * Return: 0 if timed out, and positive (at least 1, or number of jiffies left + * till timeout) if completed. */ unsigned long __sched wait_for_completion_timeout(struct completion *x, unsigned long timeout) @@ -2829,8 +2834,8 @@ EXPORT_SYMBOL(wait_for_completion_io); * specified timeout to expire. The timeout is in jiffies. It is not * interruptible. The caller is accounted as waiting for IO. * - * The return value is 0 if timed out, and positive (at least 1, or number of - * jiffies left till timeout) if completed. + * Return: 0 if timed out, and positive (at least 1, or number of jiffies left + * till timeout) if completed. */ unsigned long __sched wait_for_completion_io_timeout(struct completion *x, unsigned long timeout) @@ -2846,7 +2851,7 @@ EXPORT_SYMBOL(wait_for_completion_io_timeout); * This waits for completion of a specific task to be signaled. It is * interruptible. * - * The return value is -ERESTARTSYS if interrupted, 0 if completed. + * Return: -ERESTARTSYS if interrupted, 0 if completed. */ int __sched wait_for_completion_interruptible(struct completion *x) { @@ -2865,8 +2870,8 @@ EXPORT_SYMBOL(wait_for_completion_interruptible); * This waits for either a completion of a specific task to be signaled or for a * specified timeout to expire. It is interruptible. The timeout is in jiffies. * - * The return value is -ERESTARTSYS if interrupted, 0 if timed out, - * positive (at least 1, or number of jiffies left till timeout) if completed. + * Return: -ERESTARTSYS if interrupted, 0 if timed out, positive (at least 1, + * or number of jiffies left till timeout) if completed. */ long __sched wait_for_completion_interruptible_timeout(struct completion *x, @@ -2883,7 +2888,7 @@ EXPORT_SYMBOL(wait_for_completion_interruptible_timeout); * This waits to be signaled for completion of a specific task. It can be * interrupted by a kill signal. * - * The return value is -ERESTARTSYS if interrupted, 0 if completed. + * Return: -ERESTARTSYS if interrupted, 0 if completed. */ int __sched wait_for_completion_killable(struct completion *x) { @@ -2903,8 +2908,8 @@ EXPORT_SYMBOL(wait_for_completion_killable); * signaled or for a specified timeout to expire. It can be * interrupted by a kill signal. The timeout is in jiffies. * - * The return value is -ERESTARTSYS if interrupted, 0 if timed out, - * positive (at least 1, or number of jiffies left till timeout) if completed. + * Return: -ERESTARTSYS if interrupted, 0 if timed out, positive (at least 1, + * or number of jiffies left till timeout) if completed. */ long __sched wait_for_completion_killable_timeout(struct completion *x, @@ -2918,7 +2923,7 @@ EXPORT_SYMBOL(wait_for_completion_killable_timeout); * try_wait_for_completion - try to decrement a completion without blocking * @x: completion structure * - * Returns: 0 if a decrement cannot be done without blocking + * Return: 0 if a decrement cannot be done without blocking * 1 if a decrement succeeded. * * If a completion is being used as a counting completion, @@ -2945,7 +2950,7 @@ EXPORT_SYMBOL(try_wait_for_completion); * completion_done - Test to see if a completion has any waiters * @x: completion structure * - * Returns: 0 if there are waiters (wait_for_completion() in progress) + * Return: 0 if there are waiters (wait_for_completion() in progress) * 1 if there are no waiters. * */ @@ -3182,7 +3187,7 @@ SYSCALL_DEFINE1(nice, int, increment) * task_prio - return the priority value of a given task. * @p: the task in question. * - * This is the priority value as seen by users in /proc. + * Return: The priority value as seen by users in /proc. * RT tasks are offset by -200. Normal tasks are centered * around 0, value goes from -16 to +15. */ @@ -3194,6 +3199,8 @@ int task_prio(const struct task_struct *p) /** * task_nice - return the nice value of a given task. * @p: the task in question. + * + * Return: The nice value [ -20 ... 0 ... 19 ]. */ int task_nice(const struct task_struct *p) { @@ -3204,6 +3211,8 @@ EXPORT_SYMBOL(task_nice); /** * idle_cpu - is a given cpu idle currently? * @cpu: the processor in question. + * + * Return: 1 if the CPU is currently idle. 0 otherwise. */ int idle_cpu(int cpu) { @@ -3226,6 +3235,8 @@ int idle_cpu(int cpu) /** * idle_task - return the idle task for a given cpu. * @cpu: the processor in question. + * + * Return: The idle task for the cpu @cpu. */ struct task_struct *idle_task(int cpu) { @@ -3235,6 +3246,8 @@ struct task_struct *idle_task(int cpu) /** * find_process_by_pid - find a process with a matching PID value. * @pid: the pid in question. + * + * The task of @pid, if found. %NULL otherwise. */ static struct task_struct *find_process_by_pid(pid_t pid) { @@ -3432,6 +3445,8 @@ recheck: * @policy: new policy. * @param: structure containing the new RT priority. * + * Return: 0 on success. An error code otherwise. + * * NOTE that the task may be already dead. */ int sched_setscheduler(struct task_struct *p, int policy, @@ -3451,6 +3466,8 @@ EXPORT_SYMBOL_GPL(sched_setscheduler); * current context has permission. For example, this is needed in * stop_machine(): we create temporary high priority worker threads, * but our caller might not have that capability. + * + * Return: 0 on success. An error code otherwise. */ int sched_setscheduler_nocheck(struct task_struct *p, int policy, const struct sched_param *param) @@ -3485,6 +3502,8 @@ do_sched_setscheduler(pid_t pid, int policy, struct sched_param __user *param) * @pid: the pid in question. * @policy: new policy. * @param: structure containing the new RT priority. + * + * Return: 0 on success. An error code otherwise. */ SYSCALL_DEFINE3(sched_setscheduler, pid_t, pid, int, policy, struct sched_param __user *, param) @@ -3500,6 +3519,8 @@ SYSCALL_DEFINE3(sched_setscheduler, pid_t, pid, int, policy, * sys_sched_setparam - set/change the RT priority of a thread * @pid: the pid in question. * @param: structure containing the new RT priority. + * + * Return: 0 on success. An error code otherwise. */ SYSCALL_DEFINE2(sched_setparam, pid_t, pid, struct sched_param __user *, param) { @@ -3509,6 +3530,9 @@ SYSCALL_DEFINE2(sched_setparam, pid_t, pid, struct sched_param __user *, param) /** * sys_sched_getscheduler - get the policy (scheduling class) of a thread * @pid: the pid in question. + * + * Return: On success, the policy of the thread. Otherwise, a negative error + * code. */ SYSCALL_DEFINE1(sched_getscheduler, pid_t, pid) { @@ -3535,6 +3559,9 @@ SYSCALL_DEFINE1(sched_getscheduler, pid_t, pid) * sys_sched_getparam - get the RT priority of a thread * @pid: the pid in question. * @param: structure containing the RT priority. + * + * Return: On success, 0 and the RT priority is in @param. Otherwise, an error + * code. */ SYSCALL_DEFINE2(sched_getparam, pid_t, pid, struct sched_param __user *, param) { @@ -3659,6 +3686,8 @@ static int get_user_cpu_mask(unsigned long __user *user_mask_ptr, unsigned len, * @pid: pid of the process * @len: length in bytes of the bitmask pointed to by user_mask_ptr * @user_mask_ptr: user-space pointer to the new cpu mask + * + * Return: 0 on success. An error code otherwise. */ SYSCALL_DEFINE3(sched_setaffinity, pid_t, pid, unsigned int, len, unsigned long __user *, user_mask_ptr) @@ -3710,6 +3739,8 @@ out_unlock: * @pid: pid of the process * @len: length in bytes of the bitmask pointed to by user_mask_ptr * @user_mask_ptr: user-space pointer to hold the current cpu mask + * + * Return: 0 on success. An error code otherwise. */ SYSCALL_DEFINE3(sched_getaffinity, pid_t, pid, unsigned int, len, unsigned long __user *, user_mask_ptr) @@ -3744,6 +3775,8 @@ SYSCALL_DEFINE3(sched_getaffinity, pid_t, pid, unsigned int, len, * * This function yields the current CPU to other tasks. If there are no * other threads running on this CPU then this function will return. + * + * Return: 0. */ SYSCALL_DEFINE0(sched_yield) { @@ -3869,7 +3902,7 @@ EXPORT_SYMBOL(yield); * It's the caller's job to ensure that the target task struct * can't go away on us before we can do any checks. * - * Returns: + * Return: * true (>0) if we indeed boosted the target task. * false (0) if we failed to boost the target. * -ESRCH if there's no task to yield to. @@ -3972,8 +4005,9 @@ long __sched io_schedule_timeout(long timeout) * sys_sched_get_priority_max - return maximum RT priority. * @policy: scheduling class. * - * this syscall returns the maximum rt_priority that can be used - * by a given scheduling class. + * Return: On success, this syscall returns the maximum + * rt_priority that can be used by a given scheduling class. + * On failure, a negative error code is returned. */ SYSCALL_DEFINE1(sched_get_priority_max, int, policy) { @@ -3997,8 +4031,9 @@ SYSCALL_DEFINE1(sched_get_priority_max, int, policy) * sys_sched_get_priority_min - return minimum RT priority. * @policy: scheduling class. * - * this syscall returns the minimum rt_priority that can be used - * by a given scheduling class. + * Return: On success, this syscall returns the minimum + * rt_priority that can be used by a given scheduling class. + * On failure, a negative error code is returned. */ SYSCALL_DEFINE1(sched_get_priority_min, int, policy) { @@ -4024,6 +4059,9 @@ SYSCALL_DEFINE1(sched_get_priority_min, int, policy) * * this syscall writes the default timeslice value of a given process * into the user-space timespec buffer. A value of '0' means infinity. + * + * Return: On success, 0 and the timeslice is in @interval. Otherwise, + * an error code. */ SYSCALL_DEFINE2(sched_rr_get_interval, pid_t, pid, struct timespec __user *, interval) @@ -6632,6 +6670,8 @@ void normalize_rt_tasks(void) * @cpu: the processor in question. * * ONLY VALID WHEN THE WHOLE SYSTEM IS STOPPED! + * + * Return: The current task for @cpu. */ struct task_struct *curr_task(int cpu) { diff --git a/kernel/sched/cpupri.c b/kernel/sched/cpupri.c index 1095e878a46f..8b836b376d91 100644 --- a/kernel/sched/cpupri.c +++ b/kernel/sched/cpupri.c @@ -62,7 +62,7 @@ static int convert_prio(int prio) * any discrepancies created by racing against the uncertainty of the current * priority configuration. * - * Returns: (int)bool - CPUs were found + * Return: (int)bool - CPUs were found */ int cpupri_find(struct cpupri *cp, struct task_struct *p, struct cpumask *lowest_mask) @@ -203,7 +203,7 @@ void cpupri_set(struct cpupri *cp, int cpu, int newpri) * cpupri_init - initialize the cpupri structure * @cp: The cpupri context * - * Returns: -ENOMEM if memory fails. + * Return: -ENOMEM on memory allocation failure. */ int cpupri_init(struct cpupri *cp) { diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index f77f9c527449..98d135584b4b 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -4280,6 +4280,8 @@ struct sg_lb_stats { * get_sd_load_idx - Obtain the load index for a given sched domain. * @sd: The sched_domain whose load_idx is to be obtained. * @idle: The Idle status of the CPU for whose sd load_icx is obtained. + * + * Return: The load index. */ static inline int get_sd_load_idx(struct sched_domain *sd, enum cpu_idle_type idle) @@ -4574,6 +4576,9 @@ static inline void update_sg_lb_stats(struct lb_env *env, * * Determine if @sg is a busier group than the previously selected * busiest group. + * + * Return: %true if @sg is a busier group than the previously selected + * busiest group. %false otherwise. */ static bool update_sd_pick_busiest(struct lb_env *env, struct sd_lb_stats *sds, @@ -4691,7 +4696,7 @@ static inline void update_sd_lb_stats(struct lb_env *env, * assuming lower CPU number will be equivalent to lower a SMT thread * number. * - * Returns 1 when packing is required and a task should be moved to + * Return: 1 when packing is required and a task should be moved to * this CPU. The amount of the imbalance is returned in *imbalance. * * @env: The load balancing environment. @@ -4869,7 +4874,7 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s * @balance: Pointer to a variable indicating if this_cpu * is the appropriate cpu to perform load balancing at this_level. * - * Returns: - the busiest group if imbalance exists. + * Return: - The busiest group if imbalance exists. * - If no imbalance and user has opted for power-savings balance, * return the least loaded group whose CPUs can be * put to idle by rebalancing its tasks onto our group. -- cgit v1.2.3 From e04c5d76b0cfb66cadd900cf147526f2271884b8 Mon Sep 17 00:00:00 2001 From: Marcelo Tosatti Date: Wed, 10 Jul 2013 22:21:57 -0300 Subject: remove sched notifier for cross-cpu migrations Linux as a guest on KVM hypervisor, the only user of the pvclock vsyscall interface, does not require notification on task migration because: 1. cpu ID number maps 1:1 to per-CPU pvclock time info. 2. per-CPU pvclock time info is updated if the underlying CPU changes. 3. that version is increased whenever underlying CPU changes. Which is sufficient to guarantee nanoseconds counter is calculated properly. Signed-off-by: Marcelo Tosatti Acked-by: Peter Zijlstra Signed-off-by: Gleb Natapov --- kernel/sched/core.c | 15 --------------- 1 file changed, 15 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 0d8eb4525e76..0efd2eefb027 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -976,13 +976,6 @@ void check_preempt_curr(struct rq *rq, struct task_struct *p, int flags) rq->skip_clock_update = 1; } -static ATOMIC_NOTIFIER_HEAD(task_migration_notifier); - -void register_task_migration_notifier(struct notifier_block *n) -{ - atomic_notifier_chain_register(&task_migration_notifier, n); -} - #ifdef CONFIG_SMP void set_task_cpu(struct task_struct *p, unsigned int new_cpu) { @@ -1013,18 +1006,10 @@ void set_task_cpu(struct task_struct *p, unsigned int new_cpu) trace_sched_migrate_task(p, new_cpu); if (task_cpu(p) != new_cpu) { - struct task_migration_notifier tmn; - if (p->sched_class->migrate_task_rq) p->sched_class->migrate_task_rq(p, new_cpu); p->se.nr_migrations++; perf_sw_event(PERF_COUNT_SW_CPU_MIGRATIONS, 1, NULL, 0); - - tmn.task = p; - tmn.from_cpu = task_cpu(p); - tmn.to_cpu = new_cpu; - - atomic_notifier_call_chain(&task_migration_notifier, 0, &tmn); } __set_task_cpu(p, new_cpu); -- cgit v1.2.3 From 146c3442f2dd0f50d9431aea5d0d10dfd97c9999 Mon Sep 17 00:00:00 2001 From: "zhangwei(Jovi)" Date: Mon, 15 Jul 2013 16:32:44 +0800 Subject: tracing: Use trace_seq_puts()/trace_seq_putc() where possible For string without format specifiers, use trace_seq_puts() or trace_seq_putc(). Link: http://lkml.kernel.org/r/51E3B3AC.1000605@huawei.com Signed-off-by: zhangwei(Jovi) [ fixed a trace_seq_putc(s, " ") to trace_seq_putc(s, ' ') ] Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 10 +++---- kernel/trace/trace_events_filter.c | 4 +-- kernel/trace/trace_functions_graph.c | 52 ++++++++++++++++++------------------ kernel/trace/trace_mmiotrace.c | 8 +++--- kernel/trace/trace_output.c | 14 +++++----- kernel/trace/trace_syscalls.c | 2 +- 6 files changed, 45 insertions(+), 45 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index e444ff88f0a4..eef2e566b2e7 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -36,11 +36,11 @@ int ring_buffer_print_entry_header(struct trace_seq *s) { int ret; - ret = trace_seq_printf(s, "# compressed entry header\n"); - ret = trace_seq_printf(s, "\ttype_len : 5 bits\n"); - ret = trace_seq_printf(s, "\ttime_delta : 27 bits\n"); - ret = trace_seq_printf(s, "\tarray : 32 bits\n"); - ret = trace_seq_printf(s, "\n"); + ret = trace_seq_puts(s, "# compressed entry header\n"); + ret = trace_seq_puts(s, "\ttype_len : 5 bits\n"); + ret = trace_seq_puts(s, "\ttime_delta : 27 bits\n"); + ret = trace_seq_puts(s, "\tarray : 32 bits\n"); + ret = trace_seq_putc(s, '\n'); ret = trace_seq_printf(s, "\tpadding : type == %d\n", RINGBUF_TYPE_PADDING); ret = trace_seq_printf(s, "\ttime_extend : type == %d\n", diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c index 0d883dc057d6..0c7b75a8acc8 100644 --- a/kernel/trace/trace_events_filter.c +++ b/kernel/trace/trace_events_filter.c @@ -646,7 +646,7 @@ void print_event_filter(struct ftrace_event_call *call, struct trace_seq *s) if (filter && filter->filter_string) trace_seq_printf(s, "%s\n", filter->filter_string); else - trace_seq_printf(s, "none\n"); + trace_seq_puts(s, "none\n"); mutex_unlock(&event_mutex); } @@ -660,7 +660,7 @@ void print_subsystem_event_filter(struct event_subsystem *system, if (filter && filter->filter_string) trace_seq_printf(s, "%s\n", filter->filter_string); else - trace_seq_printf(s, DEFAULT_SYS_FILTER_MESSAGE "\n"); + trace_seq_puts(s, DEFAULT_SYS_FILTER_MESSAGE "\n"); mutex_unlock(&event_mutex); } diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 8388bc99f2ee..d56ae9bae00b 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -446,7 +446,7 @@ print_graph_proc(struct trace_seq *s, pid_t pid) /* First spaces to align center */ for (i = 0; i < spaces / 2; i++) { - ret = trace_seq_printf(s, " "); + ret = trace_seq_putc(s, ' '); if (!ret) return TRACE_TYPE_PARTIAL_LINE; } @@ -457,7 +457,7 @@ print_graph_proc(struct trace_seq *s, pid_t pid) /* Last spaces to align center */ for (i = 0; i < spaces - (spaces / 2); i++) { - ret = trace_seq_printf(s, " "); + ret = trace_seq_putc(s, ' '); if (!ret) return TRACE_TYPE_PARTIAL_LINE; } @@ -503,7 +503,7 @@ verif_pid(struct trace_seq *s, pid_t pid, int cpu, struct fgraph_data *data) ------------------------------------------ */ - ret = trace_seq_printf(s, + ret = trace_seq_puts(s, " ------------------------------------------\n"); if (!ret) return TRACE_TYPE_PARTIAL_LINE; @@ -516,7 +516,7 @@ verif_pid(struct trace_seq *s, pid_t pid, int cpu, struct fgraph_data *data) if (ret == TRACE_TYPE_PARTIAL_LINE) return TRACE_TYPE_PARTIAL_LINE; - ret = trace_seq_printf(s, " => "); + ret = trace_seq_puts(s, " => "); if (!ret) return TRACE_TYPE_PARTIAL_LINE; @@ -524,7 +524,7 @@ verif_pid(struct trace_seq *s, pid_t pid, int cpu, struct fgraph_data *data) if (ret == TRACE_TYPE_PARTIAL_LINE) return TRACE_TYPE_PARTIAL_LINE; - ret = trace_seq_printf(s, + ret = trace_seq_puts(s, "\n ------------------------------------------\n\n"); if (!ret) return TRACE_TYPE_PARTIAL_LINE; @@ -645,7 +645,7 @@ print_graph_irq(struct trace_iterator *iter, unsigned long addr, ret = print_graph_proc(s, pid); if (ret == TRACE_TYPE_PARTIAL_LINE) return TRACE_TYPE_PARTIAL_LINE; - ret = trace_seq_printf(s, " | "); + ret = trace_seq_puts(s, " | "); if (!ret) return TRACE_TYPE_PARTIAL_LINE; } @@ -657,9 +657,9 @@ print_graph_irq(struct trace_iterator *iter, unsigned long addr, return ret; if (type == TRACE_GRAPH_ENT) - ret = trace_seq_printf(s, "==========>"); + ret = trace_seq_puts(s, "==========>"); else - ret = trace_seq_printf(s, "<=========="); + ret = trace_seq_puts(s, "<=========="); if (!ret) return TRACE_TYPE_PARTIAL_LINE; @@ -668,7 +668,7 @@ print_graph_irq(struct trace_iterator *iter, unsigned long addr, if (ret != TRACE_TYPE_HANDLED) return ret; - ret = trace_seq_printf(s, "\n"); + ret = trace_seq_putc(s, '\n'); if (!ret) return TRACE_TYPE_PARTIAL_LINE; @@ -705,13 +705,13 @@ trace_print_graph_duration(unsigned long long duration, struct trace_seq *s) len += strlen(nsecs_str); } - ret = trace_seq_printf(s, " us "); + ret = trace_seq_puts(s, " us "); if (!ret) return TRACE_TYPE_PARTIAL_LINE; /* Print remaining spaces to fit the row's width */ for (i = len; i < 7; i++) { - ret = trace_seq_printf(s, " "); + ret = trace_seq_putc(s, ' '); if (!ret) return TRACE_TYPE_PARTIAL_LINE; } @@ -731,13 +731,13 @@ print_graph_duration(unsigned long long duration, struct trace_seq *s, /* No real adata, just filling the column with spaces */ switch (duration) { case DURATION_FILL_FULL: - ret = trace_seq_printf(s, " | "); + ret = trace_seq_puts(s, " | "); return ret ? TRACE_TYPE_HANDLED : TRACE_TYPE_PARTIAL_LINE; case DURATION_FILL_START: - ret = trace_seq_printf(s, " "); + ret = trace_seq_puts(s, " "); return ret ? TRACE_TYPE_HANDLED : TRACE_TYPE_PARTIAL_LINE; case DURATION_FILL_END: - ret = trace_seq_printf(s, " |"); + ret = trace_seq_puts(s, " |"); return ret ? TRACE_TYPE_HANDLED : TRACE_TYPE_PARTIAL_LINE; } @@ -745,10 +745,10 @@ print_graph_duration(unsigned long long duration, struct trace_seq *s, if (flags & TRACE_GRAPH_PRINT_OVERHEAD) { /* Duration exceeded 100 msecs */ if (duration > 100000ULL) - ret = trace_seq_printf(s, "! "); + ret = trace_seq_puts(s, "! "); /* Duration exceeded 10 msecs */ else if (duration > 10000ULL) - ret = trace_seq_printf(s, "+ "); + ret = trace_seq_puts(s, "+ "); } /* @@ -757,7 +757,7 @@ print_graph_duration(unsigned long long duration, struct trace_seq *s, * to fill out the space. */ if (ret == -1) - ret = trace_seq_printf(s, " "); + ret = trace_seq_puts(s, " "); /* Catching here any failure happenned above */ if (!ret) @@ -767,7 +767,7 @@ print_graph_duration(unsigned long long duration, struct trace_seq *s, if (ret != TRACE_TYPE_HANDLED) return ret; - ret = trace_seq_printf(s, "| "); + ret = trace_seq_puts(s, "| "); if (!ret) return TRACE_TYPE_PARTIAL_LINE; @@ -817,7 +817,7 @@ print_graph_entry_leaf(struct trace_iterator *iter, /* Function */ for (i = 0; i < call->depth * TRACE_GRAPH_INDENT; i++) { - ret = trace_seq_printf(s, " "); + ret = trace_seq_putc(s, ' '); if (!ret) return TRACE_TYPE_PARTIAL_LINE; } @@ -858,7 +858,7 @@ print_graph_entry_nested(struct trace_iterator *iter, /* Function */ for (i = 0; i < call->depth * TRACE_GRAPH_INDENT; i++) { - ret = trace_seq_printf(s, " "); + ret = trace_seq_putc(s, ' '); if (!ret) return TRACE_TYPE_PARTIAL_LINE; } @@ -917,7 +917,7 @@ print_graph_prologue(struct trace_iterator *iter, struct trace_seq *s, if (ret == TRACE_TYPE_PARTIAL_LINE) return TRACE_TYPE_PARTIAL_LINE; - ret = trace_seq_printf(s, " | "); + ret = trace_seq_puts(s, " | "); if (!ret) return TRACE_TYPE_PARTIAL_LINE; } @@ -1117,7 +1117,7 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, /* Closing brace */ for (i = 0; i < trace->depth * TRACE_GRAPH_INDENT; i++) { - ret = trace_seq_printf(s, " "); + ret = trace_seq_putc(s, ' '); if (!ret) return TRACE_TYPE_PARTIAL_LINE; } @@ -1129,7 +1129,7 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, * belongs to, write out the function name. */ if (func_match) { - ret = trace_seq_printf(s, "}\n"); + ret = trace_seq_puts(s, "}\n"); if (!ret) return TRACE_TYPE_PARTIAL_LINE; } else { @@ -1179,13 +1179,13 @@ print_graph_comment(struct trace_seq *s, struct trace_entry *ent, /* Indentation */ if (depth > 0) for (i = 0; i < (depth + 1) * TRACE_GRAPH_INDENT; i++) { - ret = trace_seq_printf(s, " "); + ret = trace_seq_putc(s, ' '); if (!ret) return TRACE_TYPE_PARTIAL_LINE; } /* The comment */ - ret = trace_seq_printf(s, "/* "); + ret = trace_seq_puts(s, "/* "); if (!ret) return TRACE_TYPE_PARTIAL_LINE; @@ -1216,7 +1216,7 @@ print_graph_comment(struct trace_seq *s, struct trace_entry *ent, s->len--; } - ret = trace_seq_printf(s, " */\n"); + ret = trace_seq_puts(s, " */\n"); if (!ret) return TRACE_TYPE_PARTIAL_LINE; diff --git a/kernel/trace/trace_mmiotrace.c b/kernel/trace/trace_mmiotrace.c index a5e8f4878bfa..b3dcfb2f0fef 100644 --- a/kernel/trace/trace_mmiotrace.c +++ b/kernel/trace/trace_mmiotrace.c @@ -90,7 +90,7 @@ static int mmio_print_pcidev(struct trace_seq *s, const struct pci_dev *dev) if (drv) ret += trace_seq_printf(s, " %s\n", drv->name); else - ret += trace_seq_printf(s, " \n"); + ret += trace_seq_puts(s, " \n"); return ret; } @@ -107,7 +107,7 @@ static void mmio_pipe_open(struct trace_iterator *iter) struct header_iter *hiter; struct trace_seq *s = &iter->seq; - trace_seq_printf(s, "VERSION 20070824\n"); + trace_seq_puts(s, "VERSION 20070824\n"); hiter = kzalloc(sizeof(*hiter), GFP_KERNEL); if (!hiter) @@ -209,7 +209,7 @@ static enum print_line_t mmio_print_rw(struct trace_iterator *iter) (rw->value >> 0) & 0xff, rw->pc, 0); break; default: - ret = trace_seq_printf(s, "rw what?\n"); + ret = trace_seq_puts(s, "rw what?\n"); break; } if (ret) @@ -245,7 +245,7 @@ static enum print_line_t mmio_print_map(struct trace_iterator *iter) secs, usec_rem, m->map_id, 0UL, 0); break; default: - ret = trace_seq_printf(s, "map what?\n"); + ret = trace_seq_puts(s, "map what?\n"); break; } if (ret) diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index bb922d9ee51b..34e7cbac0c9c 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -78,7 +78,7 @@ enum print_line_t trace_print_printk_msg_only(struct trace_iterator *iter) trace_assign_type(field, entry); - ret = trace_seq_printf(s, "%s", field->buf); + ret = trace_seq_puts(s, field->buf); if (!ret) return TRACE_TYPE_PARTIAL_LINE; @@ -558,14 +558,14 @@ seq_print_userip_objs(const struct userstack_entry *entry, struct trace_seq *s, if (ret) ret = trace_seq_puts(s, "??"); if (ret) - ret = trace_seq_puts(s, "\n"); + ret = trace_seq_putc(s, '\n'); continue; } if (!ret) break; if (ret) ret = seq_print_user_ip(s, mm, ip, sym_flags); - ret = trace_seq_puts(s, "\n"); + ret = trace_seq_putc(s, '\n'); } if (mm) @@ -579,7 +579,7 @@ seq_print_ip_sym(struct trace_seq *s, unsigned long ip, unsigned long sym_flags) int ret; if (!ip) - return trace_seq_printf(s, "0"); + return trace_seq_putc(s, '0'); if (sym_flags & TRACE_ITER_SYM_OFFSET) ret = seq_print_sym_offset(s, "%s", ip); @@ -964,14 +964,14 @@ static enum print_line_t trace_fn_trace(struct trace_iterator *iter, int flags, goto partial; if ((flags & TRACE_ITER_PRINT_PARENT) && field->parent_ip) { - if (!trace_seq_printf(s, " <-")) + if (!trace_seq_puts(s, " <-")) goto partial; if (!seq_print_ip_sym(s, field->parent_ip, flags)) goto partial; } - if (!trace_seq_printf(s, "\n")) + if (!trace_seq_putc(s, '\n')) goto partial; return TRACE_TYPE_HANDLED; @@ -1210,7 +1210,7 @@ static enum print_line_t trace_stack_print(struct trace_iterator *iter, if (!seq_print_ip_sym(s, *p, flags)) goto partial; - if (!trace_seq_puts(s, "\n")) + if (!trace_seq_putc(s, '\n')) goto partial; } diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 322e16461072..061156215721 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -175,7 +175,7 @@ print_syscall_exit(struct trace_iterator *iter, int flags, entry = syscall_nr_to_meta(syscall); if (!entry) { - trace_seq_printf(s, "\n"); + trace_seq_putc(s, '\n'); return TRACE_TYPE_HANDLED; } -- cgit v1.2.3 From d611851b421731e2afd9cb956daae001af57a423 Mon Sep 17 00:00:00 2001 From: "zhangwei(Jovi)" Date: Mon, 15 Jul 2013 16:32:50 +0800 Subject: tracing: Typo fix on ring buffer comments There have some mismatch between comments with real function name, update it. This patch also add some missed function arguments description. Link: http://lkml.kernel.org/r/51E3B3B2.4080307@huawei.com Signed-off-by: zhangwei(Jovi) Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 16 +++++++++------- 1 file changed, 9 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index eef2e566b2e7..cc2f66f68dc5 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -1066,7 +1066,7 @@ static int rb_check_list(struct ring_buffer_per_cpu *cpu_buffer, } /** - * check_pages - integrity check of buffer pages + * rb_check_pages - integrity check of buffer pages * @cpu_buffer: CPU buffer with pages to test * * As a safety measure we check to make sure the data pages have not @@ -1258,7 +1258,7 @@ static int rb_cpu_notify(struct notifier_block *self, #endif /** - * ring_buffer_alloc - allocate a new ring_buffer + * __ring_buffer_alloc - allocate a new ring_buffer * @size: the size in bytes per cpu that is needed. * @flags: attributes to set for the ring buffer. * @@ -1607,6 +1607,7 @@ static void update_pages_handler(struct work_struct *work) * ring_buffer_resize - resize the ring buffer * @buffer: the buffer to resize. * @size: the new size. + * @cpu_id: the cpu buffer to resize * * Minimum size is 2 * BUF_PAGE_SIZE. * @@ -3956,11 +3957,11 @@ EXPORT_SYMBOL_GPL(ring_buffer_consume); * expected. * * After a sequence of ring_buffer_read_prepare calls, the user is - * expected to make at least one call to ring_buffer_prepare_sync. + * expected to make at least one call to ring_buffer_read_prepare_sync. * Afterwards, ring_buffer_read_start is invoked to get things going * for real. * - * This overall must be paired with ring_buffer_finish. + * This overall must be paired with ring_buffer_read_finish. */ struct ring_buffer_iter * ring_buffer_read_prepare(struct ring_buffer *buffer, int cpu) @@ -4009,7 +4010,7 @@ EXPORT_SYMBOL_GPL(ring_buffer_read_prepare_sync); * an intervening ring_buffer_read_prepare_sync must have been * performed. * - * Must be paired with ring_buffer_finish. + * Must be paired with ring_buffer_read_finish. */ void ring_buffer_read_start(struct ring_buffer_iter *iter) @@ -4031,7 +4032,7 @@ ring_buffer_read_start(struct ring_buffer_iter *iter) EXPORT_SYMBOL_GPL(ring_buffer_read_start); /** - * ring_buffer_finish - finish reading the iterator of the buffer + * ring_buffer_read_finish - finish reading the iterator of the buffer * @iter: The iterator retrieved by ring_buffer_start * * This re-enables the recording to the buffer, and frees the @@ -4346,6 +4347,7 @@ EXPORT_SYMBOL_GPL(ring_buffer_swap_cpu); /** * ring_buffer_alloc_read_page - allocate a page to read from buffer * @buffer: the buffer to allocate for. + * @cpu: the cpu buffer to allocate. * * This function is used in conjunction with ring_buffer_read_page. * When reading a full page from the ring buffer, these functions @@ -4403,7 +4405,7 @@ EXPORT_SYMBOL_GPL(ring_buffer_free_read_page); * to swap with a page in the ring buffer. * * for example: - * rpage = ring_buffer_alloc_read_page(buffer); + * rpage = ring_buffer_alloc_read_page(buffer, cpu); * if (!rpage) * return error; * ret = ring_buffer_read_page(buffer, &rpage, len, cpu, 0); -- cgit v1.2.3 From b8ebfd3f7113b63dda93d76bfec638c00e6bd514 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Mon, 17 Jun 2013 19:02:04 +0200 Subject: tracing/function: Avoid perf_trace_buf_*() if event_function.perf_events is empty perf_trace_buf_prepare() + perf_trace_buf_submit(head, task => NULL) make no sense if hlist_empty(head). Change perf_ftrace_function_call() to check event_function.perf_events beforehand. Link: http://lkml.kernel.org/r/20130617170204.GA19803@redhat.com Acked-by: Peter Zijlstra Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_event_perf.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_event_perf.c b/kernel/trace/trace_event_perf.c index 84b1e045faba..12df5573086e 100644 --- a/kernel/trace/trace_event_perf.c +++ b/kernel/trace/trace_event_perf.c @@ -266,6 +266,10 @@ perf_ftrace_function_call(unsigned long ip, unsigned long parent_ip, struct pt_regs regs; int rctx; + head = this_cpu_ptr(event_function.perf_events); + if (hlist_empty(head)) + return; + #define ENTRY_SIZE (ALIGN(sizeof(struct ftrace_entry) + sizeof(u32), \ sizeof(u64)) - sizeof(u32)) @@ -279,8 +283,6 @@ perf_ftrace_function_call(unsigned long ip, unsigned long parent_ip, entry->ip = ip; entry->parent_ip = parent_ip; - - head = this_cpu_ptr(event_function.perf_events); perf_trace_buf_submit(entry, ENTRY_SIZE, rctx, 0, 1, ®s, head, NULL); -- cgit v1.2.3 From 421c7860c6e1989da3962fafdd6699316c9f8e20 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Mon, 17 Jun 2013 19:02:07 +0200 Subject: tracing/syscall: Avoid perf_trace_buf_*() if sys_data->perf_events is empty perf_trace_buf_prepare() + perf_trace_buf_submit(head, task => NULL) make no sense if hlist_empty(head). Change perf_syscall_enter/exit() to check sys_data->{enter,exit}_event->perf_events beforehand. Link: http://lkml.kernel.org/r/20130617170207.GA19806@redhat.com Acked-by: Peter Zijlstra Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_syscalls.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 061156215721..ac0085777fbd 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -566,6 +566,10 @@ static void perf_syscall_enter(void *ignore, struct pt_regs *regs, long id) if (!sys_data) return; + head = this_cpu_ptr(sys_data->enter_event->perf_events); + if (hlist_empty(head)) + return; + /* get the size after alignment with the u32 buffer size field */ size = sizeof(unsigned long) * sys_data->nb_args + sizeof(*rec); size = ALIGN(size + sizeof(u32), sizeof(u64)); @@ -583,8 +587,6 @@ static void perf_syscall_enter(void *ignore, struct pt_regs *regs, long id) rec->nr = syscall_nr; syscall_get_arguments(current, regs, 0, sys_data->nb_args, (unsigned long *)&rec->args); - - head = this_cpu_ptr(sys_data->enter_event->perf_events); perf_trace_buf_submit(rec, size, rctx, 0, 1, regs, head, NULL); } @@ -642,6 +644,10 @@ static void perf_syscall_exit(void *ignore, struct pt_regs *regs, long ret) if (!sys_data) return; + head = this_cpu_ptr(sys_data->exit_event->perf_events); + if (hlist_empty(head)) + return; + /* We can probably do that at build time */ size = ALIGN(sizeof(*rec) + sizeof(u32), sizeof(u64)); size -= sizeof(u32); @@ -661,8 +667,6 @@ static void perf_syscall_exit(void *ignore, struct pt_regs *regs, long ret) rec->nr = syscall_nr; rec->ret = syscall_get_return_value(current, regs); - - head = this_cpu_ptr(sys_data->exit_event->perf_events); perf_trace_buf_submit(rec, size, rctx, 0, 1, regs, head, NULL); } -- cgit v1.2.3 From cd92bf61d6d70bd3eb33b46d600e3f3eb9c5778a Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Mon, 17 Jun 2013 19:02:11 +0200 Subject: tracing/perf: Move the PERF_MAX_TRACE_SIZE check into perf_trace_buf_prepare() Every perf_trace_buf_prepare() caller does WARN_ONCE(size > PERF_MAX_TRACE_SIZE, message) and "message" is almost the same. Shift this WARN_ONCE() into perf_trace_buf_prepare(). This changes the meaning of _ONCE, but I think this is fine. - 4947014 2932448 10104832 17984294 1126b26 vmlinux + 4948422 2932448 10104832 17985702 11270a6 vmlinux on my build. Link: http://lkml.kernel.org/r/20130617170211.GA19813@redhat.com Acked-by: Peter Zijlstra Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_event_perf.c | 4 ++++ kernel/trace/trace_kprobe.c | 6 ------ kernel/trace/trace_syscalls.c | 12 ------------ kernel/trace/trace_uprobe.c | 2 -- 4 files changed, 4 insertions(+), 20 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_event_perf.c b/kernel/trace/trace_event_perf.c index 12df5573086e..80c36bcf66e8 100644 --- a/kernel/trace/trace_event_perf.c +++ b/kernel/trace/trace_event_perf.c @@ -236,6 +236,10 @@ __kprobes void *perf_trace_buf_prepare(int size, unsigned short type, BUILD_BUG_ON(PERF_MAX_TRACE_SIZE % sizeof(unsigned long)); + if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE, + "perf buffer not large enough")) + return NULL; + pc = preempt_count(); *rctxp = perf_swevent_get_recursion_context(); diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 7ed6976493c8..ae6ce835b023 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -1087,9 +1087,6 @@ kprobe_perf_func(struct trace_probe *tp, struct pt_regs *regs) __size = sizeof(*entry) + tp->size + dsize; size = ALIGN(__size + sizeof(u32), sizeof(u64)); size -= sizeof(u32); - if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE, - "profile buffer not large enough")) - return; entry = perf_trace_buf_prepare(size, call->event.type, regs, &rctx); if (!entry) @@ -1120,9 +1117,6 @@ kretprobe_perf_func(struct trace_probe *tp, struct kretprobe_instance *ri, __size = sizeof(*entry) + tp->size + dsize; size = ALIGN(__size + sizeof(u32), sizeof(u64)); size -= sizeof(u32); - if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE, - "profile buffer not large enough")) - return; entry = perf_trace_buf_prepare(size, call->event.type, regs, &rctx); if (!entry) diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index ac0085777fbd..8fd03657bc7d 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -575,10 +575,6 @@ static void perf_syscall_enter(void *ignore, struct pt_regs *regs, long id) size = ALIGN(size + sizeof(u32), sizeof(u64)); size -= sizeof(u32); - if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE, - "perf buffer not large enough")) - return; - rec = (struct syscall_trace_enter *)perf_trace_buf_prepare(size, sys_data->enter_event->event.type, regs, &rctx); if (!rec) @@ -652,14 +648,6 @@ static void perf_syscall_exit(void *ignore, struct pt_regs *regs, long ret) size = ALIGN(sizeof(*rec) + sizeof(u32), sizeof(u64)); size -= sizeof(u32); - /* - * Impossible, but be paranoid with the future - * How to put this check outside runtime? - */ - if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE, - "exit event has grown above perf buffer size")) - return; - rec = (struct syscall_trace_exit *)perf_trace_buf_prepare(size, sys_data->exit_event->event.type, regs, &rctx); if (!rec) diff --git a/kernel/trace/trace_uprobe.c b/kernel/trace/trace_uprobe.c index d5d0cd368a56..a23d2d71188e 100644 --- a/kernel/trace/trace_uprobe.c +++ b/kernel/trace/trace_uprobe.c @@ -818,8 +818,6 @@ static void uprobe_perf_print(struct trace_uprobe *tu, size = SIZEOF_TRACE_ENTRY(is_ret_probe(tu)); size = ALIGN(size + tu->size + sizeof(u32), sizeof(u64)) - sizeof(u32); - if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE, "profile buffer not large enough")) - return; preempt_disable(); head = this_cpu_ptr(call->perf_events); -- cgit v1.2.3 From a232e270dcb55a70ad3241bc6fc160fd9b5c9e6c Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 9 Jul 2013 18:35:26 +0900 Subject: tracing/kprobe: Wait for disabling all running kprobe handlers Wait for disabling all running kprobe handlers when a kprobe event is disabled, since the caller, trace_remove_event_call() supposes that a removing event is disabled completely by disabling the event. With this change, ftrace can ensure that there is no running event handlers after disabling it. Link: http://lkml.kernel.org/r/20130709093526.20138.93100.stgit@mhiramat-M0-7522 Signed-off-by: Masami Hiramatsu Signed-off-by: Steven Rostedt --- kernel/trace/trace_kprobe.c | 23 +++++++++++++++++------ 1 file changed, 17 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index ae6ce835b023..3811487e7a7a 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -243,11 +243,11 @@ find_event_file_link(struct trace_probe *tp, struct ftrace_event_file *file) static int disable_trace_probe(struct trace_probe *tp, struct ftrace_event_file *file) { + struct event_file_link *link = NULL; + int wait = 0; int ret = 0; if (file) { - struct event_file_link *link; - link = find_event_file_link(tp, file); if (!link) { ret = -EINVAL; @@ -255,10 +255,7 @@ disable_trace_probe(struct trace_probe *tp, struct ftrace_event_file *file) } list_del_rcu(&link->list); - /* synchronize with kprobe_trace_func/kretprobe_trace_func */ - synchronize_sched(); - kfree(link); - + wait = 1; if (!list_empty(&tp->files)) goto out; @@ -271,8 +268,22 @@ disable_trace_probe(struct trace_probe *tp, struct ftrace_event_file *file) disable_kretprobe(&tp->rp); else disable_kprobe(&tp->rp.kp); + wait = 1; } out: + if (wait) { + /* + * Synchronize with kprobe_trace_func/kretprobe_trace_func + * to ensure disabled (all running handlers are finished). + * This is not only for kfree(), but also the caller, + * trace_remove_event_call() supposes it for releasing + * event_call related objects, which will be accessed in + * the kprobe_trace_func/kretprobe_trace_func. + */ + synchronize_sched(); + kfree(link); /* Ignored if link == NULL */ + } + return ret; } -- cgit v1.2.3 From 609e85a70bcd0eedf4ec60639dbcfb1ab011e054 Mon Sep 17 00:00:00 2001 From: Alexander Z Lam Date: Wed, 10 Jul 2013 17:34:34 -0700 Subject: tracing: Fix error handling to ensure instances can always be removed Remove debugfs directories for tracing instances during creation if an error occurs causing the trace_array for that instance to not be added to ftrace_trace_arrays. If the directory continues to exist after the error, it cannot be removed because the respective trace_array is not in ftrace_trace_arrays. Link: http://lkml.kernel.org/r/1373502874-1706-2-git-send-email-azl@google.com Cc: stable@vger.kernel.org # 3.10 Cc: Vaibhav Nagarnaik Cc: David Sharp Cc: Alexander Z Lam Signed-off-by: Alexander Z Lam Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 25b91afc29e0..7c3da7bca05b 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -5973,8 +5973,10 @@ static int new_instance_create(const char *name) goto out_free_tr; ret = event_trace_add_tracer(tr->dir, tr); - if (ret) + if (ret) { + debugfs_remove_recursive(tr->dir); goto out_free_tr; + } init_tracer_debugfs(tr, tr->dir); -- cgit v1.2.3 From f77d09a384676bde6445413949d9d2c508ff3e62 Mon Sep 17 00:00:00 2001 From: Alexander Z Lam Date: Thu, 18 Jul 2013 11:18:44 -0700 Subject: tracing: Miscellaneous fixes for trace_array ref counting Some error paths did not handle ref counting properly, and some trace files need ref counting. Link: http://lkml.kernel.org/r/1374171524-11948-1-git-send-email-azl@google.com Cc: stable@vger.kernel.org # 3.10 Cc: Vaibhav Nagarnaik Cc: David Sharp Cc: Alexander Z Lam Signed-off-by: Alexander Z Lam Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 24 ++++++++++++++++++------ kernel/trace/trace_events.c | 21 +++++++++++++++++++-- 2 files changed, 37 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 7c3da7bca05b..7d9ceab42564 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -3008,7 +3008,6 @@ static int tracing_release(struct inode *inode, struct file *file) iter = m->private; tr = iter->tr; - trace_array_put(tr); mutex_lock(&trace_types_lock); @@ -3023,6 +3022,9 @@ static int tracing_release(struct inode *inode, struct file *file) if (!iter->snapshot) /* reenable tracing if it was previously enabled */ tracing_start_tr(tr); + + __trace_array_put(tr); + mutex_unlock(&trace_types_lock); mutex_destroy(&iter->mutex); @@ -3447,6 +3449,7 @@ tracing_trace_options_write(struct file *filp, const char __user *ubuf, static int tracing_trace_options_open(struct inode *inode, struct file *file) { struct trace_array *tr = inode->i_private; + int ret; if (tracing_disabled) return -ENODEV; @@ -3454,7 +3457,11 @@ static int tracing_trace_options_open(struct inode *inode, struct file *file) if (trace_array_get(tr) < 0) return -ENODEV; - return single_open(file, tracing_trace_options_show, inode->i_private); + ret = single_open(file, tracing_trace_options_show, inode->i_private); + if (ret < 0) + trace_array_put(tr); + + return ret; } static const struct file_operations tracing_iter_fops = { @@ -3958,6 +3965,7 @@ static int tracing_open_pipe(struct inode *inode, struct file *filp) iter = kzalloc(sizeof(*iter), GFP_KERNEL); if (!iter) { ret = -ENOMEM; + __trace_array_put(tr); goto out; } @@ -4704,21 +4712,24 @@ static int tracing_snapshot_open(struct inode *inode, struct file *file) ret = PTR_ERR(iter); } else { /* Writes still need the seq_file to hold the private data */ + ret = -ENOMEM; m = kzalloc(sizeof(*m), GFP_KERNEL); if (!m) - return -ENOMEM; + goto out; iter = kzalloc(sizeof(*iter), GFP_KERNEL); if (!iter) { kfree(m); - return -ENOMEM; + goto out; } + ret = 0; + iter->tr = tr; iter->trace_buffer = &tc->tr->max_buffer; iter->cpu_file = tc->cpu; m->private = iter; file->private_data = m; } - +out: if (ret < 0) trace_array_put(tr); @@ -5328,9 +5339,10 @@ tracing_stats_read(struct file *filp, char __user *ubuf, } static const struct file_operations tracing_stats_fops = { - .open = tracing_open_generic, + .open = tracing_open_generic_tc, .read = tracing_stats_read, .llseek = generic_file_llseek, + .release = tracing_release_generic_tc, }; #ifdef CONFIG_DYNAMIC_FTRACE diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 7d854290bf81..7a75cb22eab7 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -1218,6 +1218,7 @@ show_header(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) static int ftrace_event_avail_open(struct inode *inode, struct file *file); static int ftrace_event_set_open(struct inode *inode, struct file *file); +static int ftrace_event_release(struct inode *inode, struct file *file); static const struct seq_operations show_event_seq_ops = { .start = t_start, @@ -1245,7 +1246,7 @@ static const struct file_operations ftrace_set_event_fops = { .read = seq_read, .write = ftrace_event_write, .llseek = seq_lseek, - .release = seq_release, + .release = ftrace_event_release, }; static const struct file_operations ftrace_enable_fops = { @@ -1323,6 +1324,15 @@ ftrace_event_open(struct inode *inode, struct file *file, return ret; } +static int ftrace_event_release(struct inode *inode, struct file *file) +{ + struct trace_array *tr = inode->i_private; + + trace_array_put(tr); + + return seq_release(inode, file); +} + static int ftrace_event_avail_open(struct inode *inode, struct file *file) { @@ -1336,12 +1346,19 @@ ftrace_event_set_open(struct inode *inode, struct file *file) { const struct seq_operations *seq_ops = &show_set_event_seq_ops; struct trace_array *tr = inode->i_private; + int ret; + + if (trace_array_get(tr) < 0) + return -ENODEV; if ((file->f_mode & FMODE_WRITE) && (file->f_flags & O_TRUNC)) ftrace_clear_events(tr); - return ftrace_event_open(inode, file, seq_ops); + ret = ftrace_event_open(inode, file, seq_ops); + if (ret < 0) + trace_array_put(tr); + return ret; } static struct event_subsystem * -- cgit v1.2.3 From 8f768993394a8c0d3801033c11fd86ce8c88dcac Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Thu, 18 Jul 2013 14:41:51 -0400 Subject: tracing: Add ref_data to function and fgraph tracer structs The selftest for function and function graph tracers are defined as __init, as they are only executed at boot up. The "tracer" structs that are associated to those tracers are not setup as __init as they are used after boot. To stop mismatch warnings, those structures need to be annotated with __ref_data. Currently, the tracer structures are defined to __read_mostly, as they do not really change. But in the future they should be converted to consts, but that will take a little work because they have a "next" pointer that gets updated when they are registered. That will have to wait till the next major release. Link: http://lkml.kernel.org/r/1373596735.17876.84.camel@gandalf.local.home Reported-by: kbuild test robot Reported-by: Chen Gang Signed-off-by: Steven Rostedt --- kernel/trace/trace.h | 9 +++++++++ kernel/trace/trace_functions.c | 2 +- kernel/trace/trace_functions_graph.c | 2 +- 3 files changed, 11 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 4a4f6e1828b6..57b7bb0d39b7 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -680,6 +680,15 @@ extern int trace_selftest_startup_sched_switch(struct tracer *trace, struct trace_array *tr); extern int trace_selftest_startup_branch(struct tracer *trace, struct trace_array *tr); +/* + * Tracer data references selftest functions that only occur + * on boot up. These can be __init functions. Thus, when selftests + * are enabled, then the tracers need to reference __init functions. + */ +#define __tracer_data __refdata +#else +/* Tracers are seldom changed. Optimize when selftests are disabled. */ +#define __tracer_data __read_mostly #endif /* CONFIG_FTRACE_STARTUP_TEST */ extern void *head_page(struct trace_array_cpu *data); diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c index b863f93b30f3..38fe1483c508 100644 --- a/kernel/trace/trace_functions.c +++ b/kernel/trace/trace_functions.c @@ -199,7 +199,7 @@ static int func_set_flag(u32 old_flags, u32 bit, int set) return 0; } -static struct tracer function_trace __read_mostly = +static struct tracer function_trace __tracer_data = { .name = "function", .init = function_trace_init, diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index d56ae9bae00b..b5c09242683d 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -1448,7 +1448,7 @@ static struct trace_event graph_trace_ret_event = { .funcs = &graph_functions }; -static struct tracer graph_trace __read_mostly = { +static struct tracer graph_trace __tracer_data = { .name = "function_graph", .open = graph_trace_open, .pipe_open = graph_trace_open, -- cgit v1.2.3 From 7710b639953b791610f0022a7d52d9801c93b969 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Thu, 18 Jul 2013 20:47:10 +0200 Subject: tracing: Simplify the iteration logic in f_start/f_next f_next() looks overcomplicated, and it is not strictly correct even if this doesn't matter. Say, FORMAT_FIELD_SEPERATOR should not return NULL (means EOF) if trace_get_fields() returns an empty list, we should simply advance to FORMAT_PRINTFMT as we do when we find the end of list. 1. Change f_next() to return "struct list_head *" rather than "ftrace_event_field *", and change f_show() to do list_entry(). This simplifies the code a bit, only f_show() needs to know about ftrace_event_field, and f_next() can play with ->prev directly 2. Change f_next() to not play with ->prev / return inside the switch() statement. It can simply set node = head/common_head, the prev-or-advance-to-the-next-magic below does all work. While at it. f_start() looks overcomplicated too. I don't think *pos == 0 makes sense as a separate case, just change this code to do "while" instead of "do/while". The patch also moves f_start() down, close to f_stop(). This is purely cosmetic, just to make the locking added by the next patch more clear/visible. Link: http://lkml.kernel.org/r/20130718184710.GA4783@redhat.com Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 60 +++++++++++++++++---------------------------- 1 file changed, 22 insertions(+), 38 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 7a75cb22eab7..76defd91f9b4 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -826,59 +826,33 @@ enum { static void *f_next(struct seq_file *m, void *v, loff_t *pos) { struct ftrace_event_call *call = m->private; - struct ftrace_event_field *field; struct list_head *common_head = &ftrace_common_fields; struct list_head *head = trace_get_fields(call); + struct list_head *node = v; (*pos)++; switch ((unsigned long)v) { case FORMAT_HEADER: - if (unlikely(list_empty(common_head))) - return NULL; - - field = list_entry(common_head->prev, - struct ftrace_event_field, link); - return field; + node = common_head; + break; case FORMAT_FIELD_SEPERATOR: - if (unlikely(list_empty(head))) - return NULL; - - field = list_entry(head->prev, struct ftrace_event_field, link); - return field; + node = head; + break; case FORMAT_PRINTFMT: /* all done */ return NULL; } - field = v; - if (field->link.prev == common_head) + node = node->prev; + if (node == common_head) return (void *)FORMAT_FIELD_SEPERATOR; - else if (field->link.prev == head) + else if (node == head) return (void *)FORMAT_PRINTFMT; - - field = list_entry(field->link.prev, struct ftrace_event_field, link); - - return field; -} - -static void *f_start(struct seq_file *m, loff_t *pos) -{ - loff_t l = 0; - void *p; - - /* Start by showing the header */ - if (!*pos) - return (void *)FORMAT_HEADER; - - p = (void *)FORMAT_HEADER; - do { - p = f_next(m, p, &l); - } while (p && l < *pos); - - return p; + else + return node; } static int f_show(struct seq_file *m, void *v) @@ -904,8 +878,7 @@ static int f_show(struct seq_file *m, void *v) return 0; } - field = v; - + field = list_entry(v, struct ftrace_event_field, link); /* * Smartly shows the array type(except dynamic array). * Normal: @@ -932,6 +905,17 @@ static int f_show(struct seq_file *m, void *v) return 0; } +static void *f_start(struct seq_file *m, loff_t *pos) +{ + void *p = (void *)FORMAT_HEADER; + loff_t l = 0; + + while (l < *pos && p) + p = f_next(m, p, &l); + + return p; +} + static void f_stop(struct seq_file *m, void *p) { } -- cgit v1.2.3 From cd458ba9d5a5592d37b5145e560071e91ea762ac Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Thu, 18 Jul 2013 20:47:12 +0200 Subject: tracing: Do not (ab)use trace_seq in event_id_read() event_id_read() has no reason to kmalloc "struct trace_seq" (more than PAGE_SIZE!), it can use a small buffer instead. Note: "if (*ppos) return 0" looks strange and even wrong, simple_read_from_buffer() handles ppos != 0 case corrrectly. And it seems that almost every user of trace_seq in this file should be converted too. Unless you use seq_open(), trace_seq buys nothing compared to the raw buffer, but it needs a bit more memory and code. Link: http://lkml.kernel.org/r/20130718184712.GA4786@redhat.com Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 17 ++++------------- 1 file changed, 4 insertions(+), 13 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 76defd91f9b4..898f868833f2 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -947,23 +947,14 @@ static ssize_t event_id_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { struct ftrace_event_call *call = filp->private_data; - struct trace_seq *s; - int r; + char buf[32]; + int len; if (*ppos) return 0; - s = kmalloc(sizeof(*s), GFP_KERNEL); - if (!s) - return -ENOMEM; - - trace_seq_init(s); - trace_seq_printf(s, "%d\n", call->event.type); - - r = simple_read_from_buffer(ubuf, cnt, ppos, - s->buffer, s->len); - kfree(s); - return r; + len = sprintf(buf, "%d\n", call->event.type); + return simple_read_from_buffer(ubuf, cnt, ppos, buf, len); } static ssize_t -- cgit v1.2.3 From a644a7e9587802eabb2e229177606f6a74a60fc1 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 19 Jul 2013 16:20:36 +0200 Subject: tracing: Kill trace_array->waiter Trivial. trace_array->waiter has no users since 6eaaa5d5 "tracing/core: use appropriate waiting on trace_pipe". Link: http://lkml.kernel.org/r/20130719142036.GA1594@redhat.com Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace.h | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 57b7bb0d39b7..e7d643b8a907 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -214,7 +214,6 @@ struct trace_array { struct dentry *event_dir; struct list_head systems; struct list_head events; - struct task_struct *waiter; int ref; }; -- cgit v1.2.3 From e70e78e3c83b536730e31231dd9b979768d8df3c Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 19 Jul 2013 17:36:44 +0200 Subject: tracing: Kill the unbalanced tr->ref++ in tracing_buffers_open() tracing_buffers_open() does trace_array_get() and then it wrongly inrcements tr->ref again under trace_types_lock. This means that every caller leaks trace_array: # cd /sys/kernel/debug/tracing/ # mkdir instances/X # true < instances/X/per_cpu/cpu0/trace_pipe_raw # rmdir instances/X rmdir: failed to remove `instances/X': Device or resource busy Link: http://lkml.kernel.org/r/20130719153644.GA18899@redhat.com Cc: Ingo Molnar Cc: Frederic Weisbecker Cc: Masami Hiramatsu Cc: stable@vger.kernel.org # 3.10 Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 7d9ceab42564..3f2477713aca 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -4959,8 +4959,6 @@ static int tracing_buffers_open(struct inode *inode, struct file *filp) mutex_lock(&trace_types_lock); - tr->ref++; - info->iter.tr = tr; info->iter.cpu_file = tc->cpu; info->iter.trace = tr->current_trace; -- cgit v1.2.3 From 87e3c8ae1c8676b9dd56b56456dafa14a4bacf97 Mon Sep 17 00:00:00 2001 From: Kirill Tkhai Date: Sun, 21 Jul 2013 04:32:07 +0400 Subject: sched/fair: Cleanup: remove duplicate variable declaration cfs_rq is declared twice, fix it. Also use 'se' instead of '&p->se'. Signed-off-by: Kirill Tkhai Cc: Peter Zijlstra Cc: Steven Rostedt Link: http://lkml.kernel.org/r/169201374366727@web6d.yandex.ru Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index bb456f44b7b1..ab599781129d 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -5889,11 +5889,9 @@ static void switched_from_fair(struct rq *rq, struct task_struct *p) * and ensure we don't carry in an old decay_count if we * switch back. */ - if (p->se.avg.decay_count) { - struct cfs_rq *cfs_rq = cfs_rq_of(&p->se); - __synchronize_entity_decay(&p->se); - subtract_blocked_load_contrib(cfs_rq, - p->se.avg.load_avg_contrib); + if (se->avg.decay_count) { + __synchronize_entity_decay(se); + subtract_blocked_load_contrib(cfs_rq, se->avg.load_avg_contrib); } #endif } -- cgit v1.2.3 From 1e40c2edef2537f87f94d0baf80aeaeb7d51cc23 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 19 Jul 2013 20:31:01 +0200 Subject: mutex: Fix/document access-once assumption in mutex_can_spin_on_owner() mutex_can_spin_on_owner() is technically broken in that it would in theory allow the compiler to load lock->owner twice, seeing a pointer first time and a NULL pointer the second time. Linus pointed out that a compiler has to be seriously broken to not compile this correctly - but nevertheless this change is correct as it will better document the implementation. Signed-off-by: Peter Zijlstra Acked-by: Davidlohr Bueso Acked-by: Waiman Long Acked-by: Linus Torvalds Acked-by: Thomas Gleixner Acked-by: Rik van Riel Cc: Paul E. McKenney Cc: David Howells Link: http://lkml.kernel.org/r/20130719183101.GA20909@twins.programming.kicks-ass.net Signed-off-by: Ingo Molnar --- kernel/mutex.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/mutex.c b/kernel/mutex.c index ff05f4bd86eb..7ff48c55a98b 100644 --- a/kernel/mutex.c +++ b/kernel/mutex.c @@ -209,11 +209,13 @@ int mutex_spin_on_owner(struct mutex *lock, struct task_struct *owner) */ static inline int mutex_can_spin_on_owner(struct mutex *lock) { + struct task_struct *owner; int retval = 1; rcu_read_lock(); - if (lock->owner) - retval = lock->owner->on_cpu; + owner = ACCESS_ONCE(lock->owner); + if (owner) + retval = owner->on_cpu; rcu_read_unlock(); /* * if lock->owner is not set, the mutex owner may have just acquired -- cgit v1.2.3 From 53c035204253efe373d9ff166fae6147e8c693b6 Mon Sep 17 00:00:00 2001 From: Baruch Siach Date: Wed, 17 Jul 2013 12:46:53 +0300 Subject: sched_clock: Fix integer overflow The expression '(1 << 32)' happens to evaluate as 0 on ARM, but it evaluates as 1 on xtensa and x86_64. This zeros sched_clock_mask, and breaks sched_clock(). Set the type of 1 to 'unsigned long long' to get the value we need. Reported-by: Max Filippov Tested-by: Max Filippov Acked-by: Russell King Signed-off-by: Baruch Siach Signed-off-by: John Stultz --- kernel/time/sched_clock.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/time/sched_clock.c b/kernel/time/sched_clock.c index a326f27d7f09..0b479a6a22bb 100644 --- a/kernel/time/sched_clock.c +++ b/kernel/time/sched_clock.c @@ -121,7 +121,7 @@ void __init setup_sched_clock(u32 (*read)(void), int bits, unsigned long rate) BUG_ON(bits > 32); WARN_ON(!irqs_disabled()); read_sched_clock = read; - sched_clock_mask = (1 << bits) - 1; + sched_clock_mask = (1ULL << bits) - 1; cd.rate = rate; /* calculate the mult/shift to convert counter ticks to ns. */ -- cgit v1.2.3 From 17f41571bb2c4a398785452ac2718a6c5d77180e Mon Sep 17 00:00:00 2001 From: Jiri Kosina Date: Tue, 23 Jul 2013 10:09:28 +0200 Subject: kprobes/x86: Call out into INT3 handler directly instead of using notifier In fd4363fff3d96 ("x86: Introduce int3 (breakpoint)-based instruction patching"), the mechanism that was introduced for notifying alternatives code from int3 exception handler that and exception occured was die_notifier. This is however problematic, as early code might be using jump labels even before the notifier registration has been performed, which will then lead to an oops due to unhandled exception. One of such occurences has been encountered by Fengguang: int3: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC Modules linked in: CPU: 1 PID: 0 Comm: swapper/1 Not tainted 3.11.0-rc1-01429-g04bf576 #8 task: ffff88000da1b040 ti: ffff88000da1c000 task.ti: ffff88000da1c000 RIP: 0010:[] [] ttwu_do_wakeup+0x28/0x225 RSP: 0000:ffff88000dd03f10 EFLAGS: 00000006 RAX: 0000000000000000 RBX: ffff88000dd12940 RCX: ffffffff81769c40 RDX: 0000000000000002 RSI: 0000000000000000 RDI: 0000000000000001 RBP: ffff88000dd03f28 R08: ffffffff8176a8c0 R09: 0000000000000002 R10: ffffffff810ff484 R11: ffff88000dd129e8 R12: ffff88000dbc90c0 R13: ffff88000dbc90c0 R14: ffff88000da1dfd8 R15: ffff88000da1dfd8 FS: 0000000000000000(0000) GS:ffff88000dd00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b CR2: 00000000ffffffff CR3: 0000000001c88000 CR4: 00000000000006e0 Stack: ffff88000dd12940 ffff88000dbc90c0 ffff88000da1dfd8 ffff88000dd03f48 ffffffff81109e2b ffff88000dd12940 0000000000000000 ffff88000dd03f68 ffffffff81109e9e 0000000000000000 0000000000012940 ffff88000dd03f98 Call Trace: [] ttwu_do_activate.constprop.56+0x6d/0x79 [] sched_ttwu_pending+0x67/0x84 [] scheduler_ipi+0x15a/0x2b0 [] smp_reschedule_interrupt+0x38/0x41 [] reschedule_interrupt+0x6d/0x80 [] ? __atomic_notifier_call_chain+0x5/0xc1 [] ? native_safe_halt+0xd/0x16 [] default_idle+0x147/0x282 [] arch_cpu_idle+0x3d/0x5d [] cpu_idle_loop+0x46d/0x5db [] cpu_startup_entry+0x84/0x84 [] start_secondary+0x3c8/0x3d5 [...] Fix this by directly calling poke_int3_handler() from the int3 exception handler (analogically to what ftrace has been doing already), instead of relying on notifier, registration of which might not have yet been finalized by the time of the first trap. Reported-and-tested-by: Fengguang Wu Signed-off-by: Jiri Kosina Acked-by: Masami Hiramatsu Cc: H. Peter Anvin Cc: Fengguang Wu Cc: Steven Rostedt Link: http://lkml.kernel.org/r/alpine.LNX.2.00.1307231007490.14024@pobox.suse.cz Signed-off-by: Ingo Molnar --- kernel/kprobes.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/kprobes.c b/kernel/kprobes.c index b58b490fa439..6e33498d665c 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -1709,7 +1709,7 @@ EXPORT_SYMBOL_GPL(unregister_kprobes); static struct notifier_block kprobe_exceptions_nb = { .notifier_call = kprobe_exceptions_notify, - .priority = 0x7ffffff0 /* High priority, but not first. */ + .priority = 0x7fffffff /* we need to be notified first */ }; unsigned long __weak arch_deref_entry_point(void *entry) -- cgit v1.2.3 From ec83f425dbca47e19c6737e8e7db0d0924a5de1b Mon Sep 17 00:00:00 2001 From: Davidlohr Bueso Date: Fri, 28 Jun 2013 13:13:18 -0700 Subject: mutex: Do not unnecessarily deal with waiters Upon entering the slowpath, we immediately attempt to acquire the lock by checking if it is already unlocked. If we are lucky enough that this is the case, then we don't need to deal with any waiter related logic. Furthermore any checks for an empty wait_list are unnecessary as we already know that count is non-negative and hence no one is waiting for the lock. Move the count check and xchg calls to be done before any waiters are setup - including waiter debugging. Upon failure to acquire the lock, the xchg sets the counter to 0, instead of -1 as it was originally. This can be done here since we set it back to -1 right at the beginning of the loop so other waiters are woken up when the lock is released. When tested on a 8-socket (80 core) system against a vanilla 3.10-rc1 kernel, this patch provides some small performance benefits (+2-6%). While it could be considered in the noise level, the average percentages were stable across multiple runs and no performance regressions were seen. Two big winners, for small amounts of users (10-100), were the short and compute workloads had a +19.36% and +%15.76% in jobs per minute. Also change some break statements to 'goto slowpath', which IMO makes a little more intuitive to read. Signed-off-by: Davidlohr Bueso Acked-by: Rik van Riel Acked-by: Maarten Lankhorst Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1372450398.2106.1.camel@buesod1.americas.hpqcorp.net Signed-off-by: Ingo Molnar --- kernel/mutex.c | 41 ++++++++++++++++++----------------------- 1 file changed, 18 insertions(+), 23 deletions(-) (limited to 'kernel') diff --git a/kernel/mutex.c b/kernel/mutex.c index 7ff48c55a98b..386ad5da47a5 100644 --- a/kernel/mutex.c +++ b/kernel/mutex.c @@ -463,7 +463,7 @@ __mutex_lock_common(struct mutex *lock, long state, unsigned int subclass, * performed the optimistic spinning cannot be done. */ if (ACCESS_ONCE(ww->ctx)) - break; + goto slowpath; } /* @@ -474,7 +474,7 @@ __mutex_lock_common(struct mutex *lock, long state, unsigned int subclass, owner = ACCESS_ONCE(lock->owner); if (owner && !mutex_spin_on_owner(lock, owner)) { mspin_unlock(MLOCK(lock), &node); - break; + goto slowpath; } if ((atomic_read(&lock->count) == 1) && @@ -489,8 +489,7 @@ __mutex_lock_common(struct mutex *lock, long state, unsigned int subclass, mutex_set_owner(lock); mspin_unlock(MLOCK(lock), &node); - preempt_enable(); - return 0; + goto done; } mspin_unlock(MLOCK(lock), &node); @@ -501,7 +500,7 @@ __mutex_lock_common(struct mutex *lock, long state, unsigned int subclass, * the owner complete. */ if (!owner && (need_resched() || rt_task(task))) - break; + goto slowpath; /* * The cpu_relax() call is a compiler barrier which forces @@ -515,6 +514,10 @@ slowpath: #endif spin_lock_mutex(&lock->wait_lock, flags); + /* once more, can we acquire the lock? */ + if (MUTEX_SHOW_NO_WAITER(lock) && (atomic_xchg(&lock->count, 0) == 1)) + goto skip_wait; + debug_mutex_lock_common(lock, &waiter); debug_mutex_add_waiter(lock, &waiter, task_thread_info(task)); @@ -522,9 +525,6 @@ slowpath: list_add_tail(&waiter.list, &lock->wait_list); waiter.task = task; - if (MUTEX_SHOW_NO_WAITER(lock) && (atomic_xchg(&lock->count, -1) == 1)) - goto done; - lock_contended(&lock->dep_map, ip); for (;;) { @@ -538,7 +538,7 @@ slowpath: * other waiters: */ if (MUTEX_SHOW_NO_WAITER(lock) && - (atomic_xchg(&lock->count, -1) == 1)) + (atomic_xchg(&lock->count, -1) == 1)) break; /* @@ -563,24 +563,25 @@ slowpath: schedule_preempt_disabled(); spin_lock_mutex(&lock->wait_lock, flags); } + mutex_remove_waiter(lock, &waiter, current_thread_info()); + /* set it to 0 if there are no waiters left: */ + if (likely(list_empty(&lock->wait_list))) + atomic_set(&lock->count, 0); + debug_mutex_free_waiter(&waiter); -done: +skip_wait: + /* got the lock - cleanup and rejoice! */ lock_acquired(&lock->dep_map, ip); - /* got the lock - rejoice! */ - mutex_remove_waiter(lock, &waiter, current_thread_info()); mutex_set_owner(lock); if (!__builtin_constant_p(ww_ctx == NULL)) { - struct ww_mutex *ww = container_of(lock, - struct ww_mutex, - base); + struct ww_mutex *ww = container_of(lock, struct ww_mutex, base); struct mutex_waiter *cur; /* * This branch gets optimized out for the common case, * and is only important for ww_mutex_lock. */ - ww_mutex_lock_acquired(ww, ww_ctx); ww->ctx = ww_ctx; @@ -594,15 +595,9 @@ done: } } - /* set it to 0 if there are no waiters left: */ - if (likely(list_empty(&lock->wait_list))) - atomic_set(&lock->count, 0); - spin_unlock_mutex(&lock->wait_lock, flags); - - debug_mutex_free_waiter(&waiter); +done: preempt_enable(); - return 0; err: -- cgit v1.2.3 From a5cdd40c9877e9aba704c020fd65d26b5cfecf18 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 16 Jul 2013 17:09:07 +0200 Subject: perf: Update perf_event_type documentation Due to a discussion with Adrian I had a good look at the perf_event_type record layout and found the documentation to be somewhat unclear. Cc: Adrian Hunter Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/20130716150907.GL23818@dyad.programming.kicks-ass.net Signed-off-by: Ingo Molnar --- kernel/events/core.c | 31 ++++++++++++++++--------------- 1 file changed, 16 insertions(+), 15 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 5e2bce90b477..127411400116 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -4462,20 +4462,6 @@ void perf_output_sample(struct perf_output_handle *handle, } } - if (!event->attr.watermark) { - int wakeup_events = event->attr.wakeup_events; - - if (wakeup_events) { - struct ring_buffer *rb = handle->rb; - int events = local_inc_return(&rb->events); - - if (events >= wakeup_events) { - local_sub(wakeup_events, &rb->events); - local_inc(&rb->wakeup); - } - } - } - if (sample_type & PERF_SAMPLE_BRANCH_STACK) { if (data->br_stack) { size_t size; @@ -4511,16 +4497,31 @@ void perf_output_sample(struct perf_output_handle *handle, } } - if (sample_type & PERF_SAMPLE_STACK_USER) + if (sample_type & PERF_SAMPLE_STACK_USER) { perf_output_sample_ustack(handle, data->stack_user_size, data->regs_user.regs); + } if (sample_type & PERF_SAMPLE_WEIGHT) perf_output_put(handle, data->weight); if (sample_type & PERF_SAMPLE_DATA_SRC) perf_output_put(handle, data->data_src.val); + + if (!event->attr.watermark) { + int wakeup_events = event->attr.wakeup_events; + + if (wakeup_events) { + struct ring_buffer *rb = handle->rb; + int events = local_inc_return(&rb->events); + + if (events >= wakeup_events) { + local_sub(wakeup_events, &rb->events); + local_inc(&rb->wakeup); + } + } + } } void perf_prepare_sample(struct perf_event_header *header, -- cgit v1.2.3 From 685207963be973fbb73550db6edaf920a283e1a7 Mon Sep 17 00:00:00 2001 From: Vladimir Davydov Date: Mon, 15 Jul 2013 17:49:19 +0400 Subject: sched: Move h_load calculation to task_h_load() The bad thing about update_h_load(), which computes hierarchical load factor for task groups, is that it is called for each task group in the system before every load balancer run, and since rebalance can be triggered very often, this function can eat really a lot of cpu time if there are many cpu cgroups in the system. Although the situation was improved significantly by commit a35b646 ('sched, cgroup: Reduce rq->lock hold times for large cgroup hierarchies'), the problem still can arise under some kinds of loads, e.g. when cpus are switching from idle to busy and back very frequently. For instance, when I start 1000 of processes that wake up every millisecond on my 8 cpus host, 'top' and 'perf top' show: Cpu(s): 17.8%us, 24.3%sy, 0.0%ni, 57.9%id, 0.0%wa, 0.0%hi, 0.0%si Events: 243K cycles 7.57% [kernel] [k] __schedule 7.08% [kernel] [k] timerqueue_add 6.13% libc-2.12.so [.] usleep Then if I create 10000 *idle* cpu cgroups (no processes in them), cpu usage increases significantly although the 'wakers' are still executing in the root cpu cgroup: Cpu(s): 19.1%us, 48.7%sy, 0.0%ni, 31.6%id, 0.0%wa, 0.0%hi, 0.7%si Events: 230K cycles 24.56% [kernel] [k] tg_load_down 5.76% [kernel] [k] __schedule This happens because this particular kind of load triggers 'new idle' rebalance very frequently, which requires calling update_h_load(), which, in turn, calls tg_load_down() for every *idle* cpu cgroup even though it is absolutely useless, because idle cpu cgroups have no tasks to pull. This patch tries to improve the situation by making h_load calculation proceed only when h_load is really necessary. To achieve this, it substitutes update_h_load() with update_cfs_rq_h_load(), which computes h_load only for a given cfs_rq and all its ascendants, and makes the load balancer call this function whenever it considers if a task should be pulled, i.e. it moves h_load calculations directly to task_h_load(). For h_load of the same cfs_rq not to be updated multiple times (in case several tasks in the same cgroup are considered during the same balance run), the patch keeps the time of the last h_load update for each cfs_rq and breaks calculation when it finds h_load to be uptodate. The benefit of it is that h_load is computed only for those cfs_rq's, which really need it, in particular all idle task groups are skipped. Although this, in fact, moves h_load calculation under rq lock, it should not affect latency much, because the amount of work done under rq lock while trying to pull tasks is limited by sched_nr_migrate. After the patch applied with the setup described above (1000 wakers in the root cgroup and 10000 idle cgroups), I get: Cpu(s): 16.9%us, 24.8%sy, 0.0%ni, 58.4%id, 0.0%wa, 0.0%hi, 0.0%si Events: 242K cycles 7.57% [kernel] [k] __schedule 6.70% [kernel] [k] timerqueue_add 5.93% libc-2.12.so [.] usleep Signed-off-by: Vladimir Davydov Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1373896159-1278-1-git-send-email-vdavydov@parallels.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 58 ++++++++++++++++++++++++---------------------------- kernel/sched/sched.h | 7 +++---- 2 files changed, 30 insertions(+), 35 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index bb456f44b7b1..765d87acdf05 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -4171,47 +4171,48 @@ static void update_blocked_averages(int cpu) } /* - * Compute the cpu's hierarchical load factor for each task group. + * Compute the hierarchical load factor for cfs_rq and all its ascendants. * This needs to be done in a top-down fashion because the load of a child * group is a fraction of its parents load. */ -static int tg_load_down(struct task_group *tg, void *data) +static void update_cfs_rq_h_load(struct cfs_rq *cfs_rq) { - unsigned long load; - long cpu = (long)data; - - if (!tg->parent) { - load = cpu_rq(cpu)->avg.load_avg_contrib; - } else { - load = tg->parent->cfs_rq[cpu]->h_load; - load = div64_ul(load * tg->se[cpu]->avg.load_avg_contrib, - tg->parent->cfs_rq[cpu]->runnable_load_avg + 1); - } - - tg->cfs_rq[cpu]->h_load = load; - - return 0; -} - -static void update_h_load(long cpu) -{ - struct rq *rq = cpu_rq(cpu); + struct rq *rq = rq_of(cfs_rq); + struct sched_entity *se = cfs_rq->tg->se[cpu_of(rq)]; unsigned long now = jiffies; + unsigned long load; - if (rq->h_load_throttle == now) + if (cfs_rq->last_h_load_update == now) return; - rq->h_load_throttle = now; + cfs_rq->h_load_next = NULL; + for_each_sched_entity(se) { + cfs_rq = cfs_rq_of(se); + cfs_rq->h_load_next = se; + if (cfs_rq->last_h_load_update == now) + break; + } - rcu_read_lock(); - walk_tg_tree(tg_load_down, tg_nop, (void *)cpu); - rcu_read_unlock(); + if (!se) { + cfs_rq->h_load = rq->avg.load_avg_contrib; + cfs_rq->last_h_load_update = now; + } + + while ((se = cfs_rq->h_load_next) != NULL) { + load = cfs_rq->h_load; + load = div64_ul(load * se->avg.load_avg_contrib, + cfs_rq->runnable_load_avg + 1); + cfs_rq = group_cfs_rq(se); + cfs_rq->h_load = load; + cfs_rq->last_h_load_update = now; + } } static unsigned long task_h_load(struct task_struct *p) { struct cfs_rq *cfs_rq = task_cfs_rq(p); + update_cfs_rq_h_load(cfs_rq); return div64_ul(p->se.avg.load_avg_contrib * cfs_rq->h_load, cfs_rq->runnable_load_avg + 1); } @@ -4220,10 +4221,6 @@ static inline void update_blocked_averages(int cpu) { } -static inline void update_h_load(long cpu) -{ -} - static unsigned long task_h_load(struct task_struct *p) { return p->se.avg.load_avg_contrib; @@ -5108,7 +5105,6 @@ redo: env.src_rq = busiest; env.loop_max = min(sysctl_sched_nr_migrate, busiest->nr_running); - update_h_load(env.src_cpu); more_balance: local_irq_save(flags); double_rq_lock(env.dst_rq, busiest); diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index ef0a7b2439dd..5e129efb84ce 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -285,7 +285,6 @@ struct cfs_rq { /* Required to track per-cpu representation of a task_group */ u32 tg_runnable_contrib; unsigned long tg_load_contrib; -#endif /* CONFIG_FAIR_GROUP_SCHED */ /* * h_load = weight * f(tg) @@ -294,6 +293,9 @@ struct cfs_rq { * this group. */ unsigned long h_load; + u64 last_h_load_update; + struct sched_entity *h_load_next; +#endif /* CONFIG_FAIR_GROUP_SCHED */ #endif /* CONFIG_SMP */ #ifdef CONFIG_FAIR_GROUP_SCHED @@ -429,9 +431,6 @@ struct rq { #ifdef CONFIG_FAIR_GROUP_SCHED /* list of leaf cfs_rq on this cpu: */ struct list_head leaf_cfs_rq_list; -#ifdef CONFIG_SMP - unsigned long h_load_throttle; -#endif /* CONFIG_SMP */ #endif /* CONFIG_FAIR_GROUP_SCHED */ #ifdef CONFIG_RT_GROUP_SCHED -- cgit v1.2.3 From 62470419e993f8d9d93db0effd3af4296ecb79a5 Mon Sep 17 00:00:00 2001 From: Michael Wang Date: Thu, 4 Jul 2013 12:55:51 +0800 Subject: sched: Implement smarter wake-affine logic The wake-affine scheduler feature is currently always trying to pull the wakee close to the waker. In theory this should be beneficial if the waker's CPU caches hot data for the wakee, and it's also beneficial in the extreme ping-pong high context switch rate case. Testing shows it can benefit hackbench up to 15%. However, the feature is somewhat blind, from which some workloads such as pgbench suffer. It's also time-consuming algorithmically. Testing shows it can damage pgbench up to 50% - far more than the benefit it brings in the best case. So wake-affine should be smarter and it should realize when to stop its thankless effort at trying to find a suitable CPU to wake on. This patch introduces 'wakee_flips', which will be increased each time the task flips (switches) its wakee target. So a high 'wakee_flips' value means the task has more than one wakee, and the bigger the number, the higher the wakeup frequency. Now when making the decision on whether to pull or not, pay attention to the wakee with a high 'wakee_flips', pulling such a task may benefit the wakee. Also imply that the waker will face cruel competition later, it could be very cruel or very fast depends on the story behind 'wakee_flips', waker therefore suffers. Furthermore, if waker also has a high 'wakee_flips', that implies that multiple tasks rely on it, then waker's higher latency will damage all of them, so pulling wakee seems to be a bad deal. Thus, when 'waker->wakee_flips / wakee->wakee_flips' becomes higher and higher, the cost of pulling seems to be worse and worse. The patch therefore helps the wake-affine feature to stop its pulling work when: wakee->wakee_flips > factor && waker->wakee_flips > (factor * wakee->wakee_flips) The 'factor' here is the number of CPUs in the current CPU's NUMA node, so a bigger node will lead to more pulling since the trial becomes more severe. After applying the patch, pgbench shows up to 40% improvements and no regressions. Tested with 12 cpu x86 server and tip 3.10.0-rc7. The percentages in the final column highlight the areas with the biggest wins, all other areas improved as well: pgbench base smart | db_size | clients | tps | | tps | +---------+---------+-------+ +-------+ | 22 MB | 1 | 10598 | | 10796 | | 22 MB | 2 | 21257 | | 21336 | | 22 MB | 4 | 41386 | | 41622 | | 22 MB | 8 | 51253 | | 57932 | | 22 MB | 12 | 48570 | | 54000 | | 22 MB | 16 | 46748 | | 55982 | +19.75% | 22 MB | 24 | 44346 | | 55847 | +25.93% | 22 MB | 32 | 43460 | | 54614 | +25.66% | 7484 MB | 1 | 8951 | | 9193 | | 7484 MB | 2 | 19233 | | 19240 | | 7484 MB | 4 | 37239 | | 37302 | | 7484 MB | 8 | 46087 | | 50018 | | 7484 MB | 12 | 42054 | | 48763 | | 7484 MB | 16 | 40765 | | 51633 | +26.66% | 7484 MB | 24 | 37651 | | 52377 | +39.11% | 7484 MB | 32 | 37056 | | 51108 | +37.92% | 15 GB | 1 | 8845 | | 9104 | | 15 GB | 2 | 19094 | | 19162 | | 15 GB | 4 | 36979 | | 36983 | | 15 GB | 8 | 46087 | | 49977 | | 15 GB | 12 | 41901 | | 48591 | | 15 GB | 16 | 40147 | | 50651 | +26.16% | 15 GB | 24 | 37250 | | 52365 | +40.58% | 15 GB | 32 | 36470 | | 50015 | +37.14% Signed-off-by: Michael Wang Cc: Mike Galbraith Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/51D50057.9000809@linux.vnet.ibm.com [ Improved the changelog. ] Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 47 +++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 47 insertions(+) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 765d87acdf05..860063a8c849 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -3017,6 +3017,23 @@ static unsigned long cpu_avg_load_per_task(int cpu) return 0; } +static void record_wakee(struct task_struct *p) +{ + /* + * Rough decay (wiping) for cost saving, don't worry + * about the boundary, really active task won't care + * about the loss. + */ + if (jiffies > current->wakee_flip_decay_ts + HZ) { + current->wakee_flips = 0; + current->wakee_flip_decay_ts = jiffies; + } + + if (current->last_wakee != p) { + current->last_wakee = p; + current->wakee_flips++; + } +} static void task_waking_fair(struct task_struct *p) { @@ -3037,6 +3054,7 @@ static void task_waking_fair(struct task_struct *p) #endif se->vruntime -= min_vruntime; + record_wakee(p); } #ifdef CONFIG_FAIR_GROUP_SCHED @@ -3155,6 +3173,28 @@ static inline unsigned long effective_load(struct task_group *tg, int cpu, #endif +static int wake_wide(struct task_struct *p) +{ + int factor = nr_cpus_node(cpu_to_node(smp_processor_id())); + + /* + * Yeah, it's the switching-frequency, could means many wakee or + * rapidly switch, use factor here will just help to automatically + * adjust the loose-degree, so bigger node will lead to more pull. + */ + if (p->wakee_flips > factor) { + /* + * wakee is somewhat hot, it needs certain amount of cpu + * resource, so if waker is far more hot, prefer to leave + * it alone. + */ + if (current->wakee_flips > (factor * p->wakee_flips)) + return 1; + } + + return 0; +} + static int wake_affine(struct sched_domain *sd, struct task_struct *p, int sync) { s64 this_load, load; @@ -3164,6 +3204,13 @@ static int wake_affine(struct sched_domain *sd, struct task_struct *p, int sync) unsigned long weight; int balanced; + /* + * If we wake multiple tasks be careful to not bounce + * ourselves around too much. + */ + if (wake_wide(p)) + return 0; + idx = sd->wake_idx; this_cpu = smp_processor_id(); prev_cpu = task_cpu(p); -- cgit v1.2.3 From 7d9ffa8961482232d964173cccba6e14d2d543b2 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 4 Jul 2013 12:56:46 +0800 Subject: sched: Micro-optimize the smart wake-affine logic Smart wake-affine is using node-size as the factor currently, but the overhead of the mask operation is high. Thus, this patch introduce the 'sd_llc_size' percpu variable, which will record the highest cache-share domain size, and make it to be the new factor, in order to reduce the overhead and make it more reasonable. Tested-by: Davidlohr Bueso Tested-by: Michael Wang Signed-off-by: Peter Zijlstra Acked-by: Michael Wang Cc: Mike Galbraith Link: http://lkml.kernel.org/r/51D5008E.6030102@linux.vnet.ibm.com [ Tidied up the changelog. ] Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 7 ++++++- kernel/sched/fair.c | 2 +- kernel/sched/sched.h | 1 + 3 files changed, 8 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index b7c32cb7bfeb..6df0fbe53767 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5083,18 +5083,23 @@ static void destroy_sched_domains(struct sched_domain *sd, int cpu) * two cpus are in the same cache domain, see cpus_share_cache(). */ DEFINE_PER_CPU(struct sched_domain *, sd_llc); +DEFINE_PER_CPU(int, sd_llc_size); DEFINE_PER_CPU(int, sd_llc_id); static void update_top_cache_domain(int cpu) { struct sched_domain *sd; int id = cpu; + int size = 1; sd = highest_flag_domain(cpu, SD_SHARE_PKG_RESOURCES); - if (sd) + if (sd) { id = cpumask_first(sched_domain_span(sd)); + size = cpumask_weight(sched_domain_span(sd)); + } rcu_assign_pointer(per_cpu(sd_llc, cpu), sd); + per_cpu(sd_llc_size, cpu) = size; per_cpu(sd_llc_id, cpu) = id; } diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 860063a8c849..f237437446e5 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -3175,7 +3175,7 @@ static inline unsigned long effective_load(struct task_group *tg, int cpu, static int wake_wide(struct task_struct *p) { - int factor = nr_cpus_node(cpu_to_node(smp_processor_id())); + int factor = this_cpu_read(sd_llc_size); /* * Yeah, it's the switching-frequency, could means many wakee or diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index 5e129efb84ce..4c1cb8029feb 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -594,6 +594,7 @@ static inline struct sched_domain *highest_flag_domain(int cpu, int flag) } DECLARE_PER_CPU(struct sched_domain *, sd_llc); +DECLARE_PER_CPU(int, sd_llc_size); DECLARE_PER_CPU(int, sd_llc_id); struct sched_group_power { -- cgit v1.2.3 From 42577ca8c3616baaafdd8f167b2e1fb959026081 Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 23 Jul 2013 16:49:24 +0100 Subject: Fix __wait_on_atomic_t() to call the action func if the counter != 0 Fix __wait_on_atomic_t() so that it calls the action func if the counter != 0 rather than if the counter is 0 so as to be analogous to __wait_on_bit(). Thanks to Yacine who found this by visual inspection. This will affect FS-Cache in that it will could fail to sleep correctly when trying to clean up after a netfs cookie is withdrawn. Reported-by: Yacine Belkadi Signed-off-by: David Howells Reviewed-by: Jeff Layton cc: Milosz Tanski Signed-off-by: Linus Torvalds --- kernel/wait.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/wait.c b/kernel/wait.c index ce0daa320a26..dec68bd4e9d8 100644 --- a/kernel/wait.c +++ b/kernel/wait.c @@ -333,7 +333,8 @@ int __wait_on_atomic_t(wait_queue_head_t *wq, struct wait_bit_queue *q, prepare_to_wait(wq, &q->wait, mode); val = q->key.flags; if (atomic_read(val) == 0) - ret = (*action)(val); + break; + ret = (*action)(val); } while (!ret && atomic_read(val) != 0); finish_wait(wq, &q->wait); return ret; -- cgit v1.2.3 From 649e9c70da6bfbeb563193a35d3424a5aa7c0d38 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 23 Jul 2013 17:25:54 +0200 Subject: tracing: Introduce trace_create_cpu_file() and tracing_get_cpu() Every "file_operations" used by tracing_init_debugfs_percpu is buggy. f_op->open/etc does: 1. struct trace_cpu *tc = inode->i_private; struct trace_array *tr = tc->tr; 2. trace_array_get(tr) or fail; 3. do_something(tc); But tc (and tr) can be already freed before trace_array_get() is called. And it doesn't matter whether this file is per-cpu or it was created by init_tracer_debugfs(), free_percpu() or kfree() are equally bad. Note that even 1. is not safe, the freed memory can be unmapped. But even if it was safe trace_array_get() can wrongly succeed if we also race with the next new_instance_create() which can re-allocate the same tr, or tc was overwritten and ->tr points to the valid tr. In this case 3. uses the freed/reused memory. Add the new trivial helper, trace_create_cpu_file() which simply calls trace_create_file() and encodes "cpu" in "struct inode". Another helper, tracing_get_cpu() will be used to read cpu_nr-or-RING_BUFFER_ALL_CPUS. The patch abuses ->i_cdev to encode the number, it is never used unless the file is S_ISCHR(). But we could use something else, say, i_bytes or even ->d_fsdata. In any case this hack is hidden inside these 2 helpers, it would be trivial to change them if needed. This patch only changes tracing_init_debugfs_percpu() to use the new trace_create_cpu_file(), the next patches will change file_operations. Note: tracing_get_cpu(inode) is always safe but you can't trust the result unless trace_array_get() was called, without trace_types_lock which acts as a barrier it can wrongly return RING_BUFFER_ALL_CPUS. Link: http://lkml.kernel.org/r/20130723152554.GA23710@redhat.com Cc: Al Viro Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 50 ++++++++++++++++++++++++++++++++++++-------------- 1 file changed, 36 insertions(+), 14 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 3f2477713aca..cfff63c2148a 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2843,6 +2843,17 @@ static int s_show(struct seq_file *m, void *v) return 0; } +/* + * Should be used after trace_array_get(), trace_types_lock + * ensures that i_cdev was already initialized. + */ +static inline int tracing_get_cpu(struct inode *inode) +{ + if (inode->i_cdev) /* See trace_create_cpu_file() */ + return (long)inode->i_cdev - 1; + return RING_BUFFER_ALL_CPUS; +} + static const struct seq_operations tracer_seq_ops = { .start = s_start, .next = s_next, @@ -5529,6 +5540,17 @@ static struct dentry *tracing_dentry_percpu(struct trace_array *tr, int cpu) return tr->percpu_dir; } +static struct dentry * +trace_create_cpu_file(const char *name, umode_t mode, struct dentry *parent, + void *data, long cpu, const struct file_operations *fops) +{ + struct dentry *ret = trace_create_file(name, mode, parent, data, fops); + + if (ret) /* See tracing_get_cpu() */ + ret->d_inode->i_cdev = (void *)(cpu + 1); + return ret; +} + static void tracing_init_debugfs_percpu(struct trace_array *tr, long cpu) { @@ -5548,28 +5570,28 @@ tracing_init_debugfs_percpu(struct trace_array *tr, long cpu) } /* per cpu trace_pipe */ - trace_create_file("trace_pipe", 0444, d_cpu, - (void *)&data->trace_cpu, &tracing_pipe_fops); + trace_create_cpu_file("trace_pipe", 0444, d_cpu, + &data->trace_cpu, cpu, &tracing_pipe_fops); /* per cpu trace */ - trace_create_file("trace", 0644, d_cpu, - (void *)&data->trace_cpu, &tracing_fops); + trace_create_cpu_file("trace", 0644, d_cpu, + &data->trace_cpu, cpu, &tracing_fops); - trace_create_file("trace_pipe_raw", 0444, d_cpu, - (void *)&data->trace_cpu, &tracing_buffers_fops); + trace_create_cpu_file("trace_pipe_raw", 0444, d_cpu, + &data->trace_cpu, cpu, &tracing_buffers_fops); - trace_create_file("stats", 0444, d_cpu, - (void *)&data->trace_cpu, &tracing_stats_fops); + trace_create_cpu_file("stats", 0444, d_cpu, + &data->trace_cpu, cpu, &tracing_stats_fops); - trace_create_file("buffer_size_kb", 0444, d_cpu, - (void *)&data->trace_cpu, &tracing_entries_fops); + trace_create_cpu_file("buffer_size_kb", 0444, d_cpu, + &data->trace_cpu, cpu, &tracing_entries_fops); #ifdef CONFIG_TRACER_SNAPSHOT - trace_create_file("snapshot", 0644, d_cpu, - (void *)&data->trace_cpu, &snapshot_fops); + trace_create_cpu_file("snapshot", 0644, d_cpu, + &data->trace_cpu, cpu, &snapshot_fops); - trace_create_file("snapshot_raw", 0444, d_cpu, - (void *)&data->trace_cpu, &snapshot_raw_fops); + trace_create_cpu_file("snapshot_raw", 0444, d_cpu, + &data->trace_cpu, cpu, &snapshot_raw_fops); #endif } -- cgit v1.2.3 From 15544209cb0b5312e5220a9337a1fe61d1a1f2d9 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 23 Jul 2013 17:25:57 +0200 Subject: tracing: Change tracing_pipe_fops() to rely on tracing_get_cpu() tracing_open_pipe() is racy, the memory inode->i_private points to can be already freed. Change debugfs_create_file("trace_pipe", data) callers to to pass "data = tr", tracing_open_pipe() can use tracing_get_cpu(). Link: http://lkml.kernel.org/r/20130723152557.GA23717@redhat.com Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 16 +++++++--------- 1 file changed, 7 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index cfff63c2148a..51a99ef2a6e5 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -3959,8 +3959,7 @@ tracing_max_lat_write(struct file *filp, const char __user *ubuf, static int tracing_open_pipe(struct inode *inode, struct file *filp) { - struct trace_cpu *tc = inode->i_private; - struct trace_array *tr = tc->tr; + struct trace_array *tr = inode->i_private; struct trace_iterator *iter; int ret = 0; @@ -4006,9 +4005,9 @@ static int tracing_open_pipe(struct inode *inode, struct file *filp) if (trace_clocks[tr->clock_id].in_ns) iter->iter_flags |= TRACE_FILE_TIME_IN_NS; - iter->cpu_file = tc->cpu; - iter->tr = tc->tr; - iter->trace_buffer = &tc->tr->trace_buffer; + iter->tr = tr; + iter->trace_buffer = &tr->trace_buffer; + iter->cpu_file = tracing_get_cpu(inode); mutex_init(&iter->mutex); filp->private_data = iter; @@ -4031,8 +4030,7 @@ fail: static int tracing_release_pipe(struct inode *inode, struct file *file) { struct trace_iterator *iter = file->private_data; - struct trace_cpu *tc = inode->i_private; - struct trace_array *tr = tc->tr; + struct trace_array *tr = inode->i_private; mutex_lock(&trace_types_lock); @@ -5571,7 +5569,7 @@ tracing_init_debugfs_percpu(struct trace_array *tr, long cpu) /* per cpu trace_pipe */ trace_create_cpu_file("trace_pipe", 0444, d_cpu, - &data->trace_cpu, cpu, &tracing_pipe_fops); + tr, cpu, &tracing_pipe_fops); /* per cpu trace */ trace_create_cpu_file("trace", 0644, d_cpu, @@ -6157,7 +6155,7 @@ init_tracer_debugfs(struct trace_array *tr, struct dentry *d_tracer) (void *)&tr->trace_cpu, &tracing_fops); trace_create_file("trace_pipe", 0444, d_tracer, - (void *)&tr->trace_cpu, &tracing_pipe_fops); + tr, &tracing_pipe_fops); trace_create_file("buffer_size_kb", 0644, d_tracer, (void *)&tr->trace_cpu, &tracing_entries_fops); -- cgit v1.2.3 From 46ef2be0d1d5ccea0c41bb606143586daadd537c Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 23 Jul 2013 17:26:00 +0200 Subject: tracing: Change tracing_buffers_fops to rely on tracing_get_cpu() tracing_buffers_open() is racy, the memory inode->i_private points to can be already freed. Change debugfs_create_file("trace_pipe_raw", data) caller to pass "data = tr", tracing_buffers_open() can use tracing_get_cpu(). Change debugfs_create_file("snapshot_raw_fops", data) caller too, this file uses tracing_buffers_open/release. Link: http://lkml.kernel.org/r/20130723152600.GA23720@redhat.com Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 51a99ef2a6e5..30c058a56ffb 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -4949,8 +4949,7 @@ static const struct file_operations snapshot_raw_fops = { static int tracing_buffers_open(struct inode *inode, struct file *filp) { - struct trace_cpu *tc = inode->i_private; - struct trace_array *tr = tc->tr; + struct trace_array *tr = inode->i_private; struct ftrace_buffer_info *info; int ret; @@ -4969,7 +4968,7 @@ static int tracing_buffers_open(struct inode *inode, struct file *filp) mutex_lock(&trace_types_lock); info->iter.tr = tr; - info->iter.cpu_file = tc->cpu; + info->iter.cpu_file = tracing_get_cpu(inode); info->iter.trace = tr->current_trace; info->iter.trace_buffer = &tr->trace_buffer; info->spare = NULL; @@ -5576,7 +5575,7 @@ tracing_init_debugfs_percpu(struct trace_array *tr, long cpu) &data->trace_cpu, cpu, &tracing_fops); trace_create_cpu_file("trace_pipe_raw", 0444, d_cpu, - &data->trace_cpu, cpu, &tracing_buffers_fops); + tr, cpu, &tracing_buffers_fops); trace_create_cpu_file("stats", 0444, d_cpu, &data->trace_cpu, cpu, &tracing_stats_fops); @@ -5589,7 +5588,7 @@ tracing_init_debugfs_percpu(struct trace_array *tr, long cpu) &data->trace_cpu, cpu, &snapshot_fops); trace_create_cpu_file("snapshot_raw", 0444, d_cpu, - &data->trace_cpu, cpu, &snapshot_raw_fops); + tr, cpu, &snapshot_raw_fops); #endif } -- cgit v1.2.3 From 4d3435b8a4c3357695e09c5e7a3bf73a19fca5b0 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 23 Jul 2013 17:26:03 +0200 Subject: tracing: Change tracing_stats_fops to rely on tracing_get_cpu() tracing_open_generic_tc() is racy, the memory inode->i_private points to can be already freed. 1. Change one of its users, tracing_stats_fops, to use tracing_*_generic_tr() instead. 2. Change trace_create_cpu_file("stats", data) to pass "data = tr". 3. Change tracing_stats_read() to use tracing_get_cpu(). Link: http://lkml.kernel.org/r/20130723152603.GA23727@redhat.com Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 30c058a56ffb..e29dc8f69aac 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2982,7 +2982,6 @@ static int tracing_open_generic_tr(struct inode *inode, struct file *filp) filp->private_data = inode->i_private; return 0; - } static int tracing_open_generic_tc(struct inode *inode, struct file *filp) @@ -5285,14 +5284,14 @@ static ssize_t tracing_stats_read(struct file *filp, char __user *ubuf, size_t count, loff_t *ppos) { - struct trace_cpu *tc = filp->private_data; - struct trace_array *tr = tc->tr; + struct inode *inode = file_inode(filp); + struct trace_array *tr = inode->i_private; struct trace_buffer *trace_buf = &tr->trace_buffer; + int cpu = tracing_get_cpu(inode); struct trace_seq *s; unsigned long cnt; unsigned long long t; unsigned long usec_rem; - int cpu = tc->cpu; s = kmalloc(sizeof(*s), GFP_KERNEL); if (!s) @@ -5345,10 +5344,10 @@ tracing_stats_read(struct file *filp, char __user *ubuf, } static const struct file_operations tracing_stats_fops = { - .open = tracing_open_generic_tc, + .open = tracing_open_generic_tr, .read = tracing_stats_read, .llseek = generic_file_llseek, - .release = tracing_release_generic_tc, + .release = tracing_release_generic_tr, }; #ifdef CONFIG_DYNAMIC_FTRACE @@ -5578,7 +5577,7 @@ tracing_init_debugfs_percpu(struct trace_array *tr, long cpu) tr, cpu, &tracing_buffers_fops); trace_create_cpu_file("stats", 0444, d_cpu, - &data->trace_cpu, cpu, &tracing_stats_fops); + tr, cpu, &tracing_stats_fops); trace_create_cpu_file("buffer_size_kb", 0444, d_cpu, &data->trace_cpu, cpu, &tracing_entries_fops); -- cgit v1.2.3 From 0bc392ee46d0fd8e6b678457ef71f074f19a03c5 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 23 Jul 2013 17:26:06 +0200 Subject: tracing: Change tracing_entries_fops to rely on tracing_get_cpu() tracing_open_generic_tc() is racy, the memory inode->i_private points to can be already freed. 1. Change its last user, tracing_entries_fops, to use tracing_*_generic_tr() instead. 2. Change debugfs_create_file("buffer_size_kb", data) callers to pass "data = tr". 3. Change tracing_entries_read() and tracing_entries_write() to use tracing_get_cpu(). 4. Kill the no longer used tracing_open_generic_tc() and tracing_release_generic_tc(). Link: http://lkml.kernel.org/r/20130723152606.GA23730@redhat.com Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 49 ++++++++++++------------------------------------- 1 file changed, 12 insertions(+), 37 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index e29dc8f69aac..68b46851666f 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2984,23 +2984,6 @@ static int tracing_open_generic_tr(struct inode *inode, struct file *filp) return 0; } -static int tracing_open_generic_tc(struct inode *inode, struct file *filp) -{ - struct trace_cpu *tc = inode->i_private; - struct trace_array *tr = tc->tr; - - if (tracing_disabled) - return -ENODEV; - - if (trace_array_get(tr) < 0) - return -ENODEV; - - filp->private_data = inode->i_private; - - return 0; - -} - static int tracing_release(struct inode *inode, struct file *file) { struct seq_file *m = file->private_data; @@ -3054,15 +3037,6 @@ static int tracing_release_generic_tr(struct inode *inode, struct file *file) return 0; } -static int tracing_release_generic_tc(struct inode *inode, struct file *file) -{ - struct trace_cpu *tc = inode->i_private; - struct trace_array *tr = tc->tr; - - trace_array_put(tr); - return 0; -} - static int tracing_single_release_tr(struct inode *inode, struct file *file) { struct trace_array *tr = inode->i_private; @@ -4382,15 +4356,16 @@ static ssize_t tracing_entries_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { - struct trace_cpu *tc = filp->private_data; - struct trace_array *tr = tc->tr; + struct inode *inode = file_inode(filp); + struct trace_array *tr = inode->i_private; + int cpu = tracing_get_cpu(inode); char buf[64]; int r = 0; ssize_t ret; mutex_lock(&trace_types_lock); - if (tc->cpu == RING_BUFFER_ALL_CPUS) { + if (cpu == RING_BUFFER_ALL_CPUS) { int cpu, buf_size_same; unsigned long size; @@ -4417,7 +4392,7 @@ tracing_entries_read(struct file *filp, char __user *ubuf, } else r = sprintf(buf, "X\n"); } else - r = sprintf(buf, "%lu\n", per_cpu_ptr(tr->trace_buffer.data, tc->cpu)->entries >> 10); + r = sprintf(buf, "%lu\n", per_cpu_ptr(tr->trace_buffer.data, cpu)->entries >> 10); mutex_unlock(&trace_types_lock); @@ -4429,7 +4404,8 @@ static ssize_t tracing_entries_write(struct file *filp, const char __user *ubuf, size_t cnt, loff_t *ppos) { - struct trace_cpu *tc = filp->private_data; + struct inode *inode = file_inode(filp); + struct trace_array *tr = inode->i_private; unsigned long val; int ret; @@ -4443,8 +4419,7 @@ tracing_entries_write(struct file *filp, const char __user *ubuf, /* value is in KB */ val <<= 10; - - ret = tracing_resize_ring_buffer(tc->tr, val, tc->cpu); + ret = tracing_resize_ring_buffer(tr, val, tracing_get_cpu(inode)); if (ret < 0) return ret; @@ -4892,11 +4867,11 @@ static const struct file_operations tracing_pipe_fops = { }; static const struct file_operations tracing_entries_fops = { - .open = tracing_open_generic_tc, + .open = tracing_open_generic_tr, .read = tracing_entries_read, .write = tracing_entries_write, .llseek = generic_file_llseek, - .release = tracing_release_generic_tc, + .release = tracing_release_generic_tr, }; static const struct file_operations tracing_total_entries_fops = { @@ -5580,7 +5555,7 @@ tracing_init_debugfs_percpu(struct trace_array *tr, long cpu) tr, cpu, &tracing_stats_fops); trace_create_cpu_file("buffer_size_kb", 0444, d_cpu, - &data->trace_cpu, cpu, &tracing_entries_fops); + tr, cpu, &tracing_entries_fops); #ifdef CONFIG_TRACER_SNAPSHOT trace_create_cpu_file("snapshot", 0644, d_cpu, @@ -6156,7 +6131,7 @@ init_tracer_debugfs(struct trace_array *tr, struct dentry *d_tracer) tr, &tracing_pipe_fops); trace_create_file("buffer_size_kb", 0644, d_tracer, - (void *)&tr->trace_cpu, &tracing_entries_fops); + tr, &tracing_entries_fops); trace_create_file("buffer_total_size_kb", 0444, d_tracer, tr, &tracing_total_entries_fops); -- cgit v1.2.3 From 6484c71cbc170634fa131b6d022d86d61686b88b Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 23 Jul 2013 17:26:10 +0200 Subject: tracing: Change tracing_fops/snapshot_fops to rely on tracing_get_cpu() tracing_open() and tracing_snapshot_open() are racy, the memory inode->i_private points to can be already freed. Convert these last users of "inode->i_private == trace_cpu" to use "i_private = trace_array" and rely on tracing_get_cpu(). v2: incorporate the fix from Steven, tracing_release() must not blindly dereference file->private_data unless we know that the file was opened for reading. Link: http://lkml.kernel.org/r/20130723152610.GA23737@redhat.com Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 50 ++++++++++++++++++++++---------------------------- 1 file changed, 22 insertions(+), 28 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 68b46851666f..dd7780ddde08 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2862,9 +2862,9 @@ static const struct seq_operations tracer_seq_ops = { }; static struct trace_iterator * -__tracing_open(struct trace_array *tr, struct trace_cpu *tc, - struct inode *inode, struct file *file, bool snapshot) +__tracing_open(struct inode *inode, struct file *file, bool snapshot) { + struct trace_array *tr = inode->i_private; struct trace_iterator *iter; int cpu; @@ -2905,8 +2905,8 @@ __tracing_open(struct trace_array *tr, struct trace_cpu *tc, iter->trace_buffer = &tr->trace_buffer; iter->snapshot = snapshot; iter->pos = -1; + iter->cpu_file = tracing_get_cpu(inode); mutex_init(&iter->mutex); - iter->cpu_file = tc->cpu; /* Notify the tracer early; before we stop tracing. */ if (iter->trace && iter->trace->open) @@ -2986,22 +2986,18 @@ static int tracing_open_generic_tr(struct inode *inode, struct file *filp) static int tracing_release(struct inode *inode, struct file *file) { + struct trace_array *tr = inode->i_private; struct seq_file *m = file->private_data; struct trace_iterator *iter; - struct trace_array *tr; int cpu; - /* Writes do not use seq_file, need to grab tr from inode */ if (!(file->f_mode & FMODE_READ)) { - struct trace_cpu *tc = inode->i_private; - - trace_array_put(tc->tr); + trace_array_put(tr); return 0; } + /* Writes do not use seq_file */ iter = m->private; - tr = iter->tr; - mutex_lock(&trace_types_lock); for_each_tracing_cpu(cpu) { @@ -3048,8 +3044,7 @@ static int tracing_single_release_tr(struct inode *inode, struct file *file) static int tracing_open(struct inode *inode, struct file *file) { - struct trace_cpu *tc = inode->i_private; - struct trace_array *tr = tc->tr; + struct trace_array *tr = inode->i_private; struct trace_iterator *iter; int ret = 0; @@ -3057,16 +3052,17 @@ static int tracing_open(struct inode *inode, struct file *file) return -ENODEV; /* If this file was open for write, then erase contents */ - if ((file->f_mode & FMODE_WRITE) && - (file->f_flags & O_TRUNC)) { - if (tc->cpu == RING_BUFFER_ALL_CPUS) + if ((file->f_mode & FMODE_WRITE) && (file->f_flags & O_TRUNC)) { + int cpu = tracing_get_cpu(inode); + + if (cpu == RING_BUFFER_ALL_CPUS) tracing_reset_online_cpus(&tr->trace_buffer); else - tracing_reset(&tr->trace_buffer, tc->cpu); + tracing_reset(&tr->trace_buffer, cpu); } if (file->f_mode & FMODE_READ) { - iter = __tracing_open(tr, tc, inode, file, false); + iter = __tracing_open(inode, file, false); if (IS_ERR(iter)) ret = PTR_ERR(iter); else if (trace_flags & TRACE_ITER_LATENCY_FMT) @@ -4680,8 +4676,7 @@ struct ftrace_buffer_info { #ifdef CONFIG_TRACER_SNAPSHOT static int tracing_snapshot_open(struct inode *inode, struct file *file) { - struct trace_cpu *tc = inode->i_private; - struct trace_array *tr = tc->tr; + struct trace_array *tr = inode->i_private; struct trace_iterator *iter; struct seq_file *m; int ret = 0; @@ -4690,7 +4685,7 @@ static int tracing_snapshot_open(struct inode *inode, struct file *file) return -ENODEV; if (file->f_mode & FMODE_READ) { - iter = __tracing_open(tr, tc, inode, file, true); + iter = __tracing_open(inode, file, true); if (IS_ERR(iter)) ret = PTR_ERR(iter); } else { @@ -4707,8 +4702,8 @@ static int tracing_snapshot_open(struct inode *inode, struct file *file) ret = 0; iter->tr = tr; - iter->trace_buffer = &tc->tr->max_buffer; - iter->cpu_file = tc->cpu; + iter->trace_buffer = &tr->max_buffer; + iter->cpu_file = tracing_get_cpu(inode); m->private = iter; file->private_data = m; } @@ -5525,7 +5520,6 @@ trace_create_cpu_file(const char *name, umode_t mode, struct dentry *parent, static void tracing_init_debugfs_percpu(struct trace_array *tr, long cpu) { - struct trace_array_cpu *data = per_cpu_ptr(tr->trace_buffer.data, cpu); struct dentry *d_percpu = tracing_dentry_percpu(tr, cpu); struct dentry *d_cpu; char cpu_dir[30]; /* 30 characters should be more than enough */ @@ -5546,7 +5540,7 @@ tracing_init_debugfs_percpu(struct trace_array *tr, long cpu) /* per cpu trace */ trace_create_cpu_file("trace", 0644, d_cpu, - &data->trace_cpu, cpu, &tracing_fops); + tr, cpu, &tracing_fops); trace_create_cpu_file("trace_pipe_raw", 0444, d_cpu, tr, cpu, &tracing_buffers_fops); @@ -5559,7 +5553,7 @@ tracing_init_debugfs_percpu(struct trace_array *tr, long cpu) #ifdef CONFIG_TRACER_SNAPSHOT trace_create_cpu_file("snapshot", 0644, d_cpu, - &data->trace_cpu, cpu, &snapshot_fops); + tr, cpu, &snapshot_fops); trace_create_cpu_file("snapshot_raw", 0444, d_cpu, tr, cpu, &snapshot_raw_fops); @@ -6125,7 +6119,7 @@ init_tracer_debugfs(struct trace_array *tr, struct dentry *d_tracer) tr, &tracing_iter_fops); trace_create_file("trace", 0644, d_tracer, - (void *)&tr->trace_cpu, &tracing_fops); + tr, &tracing_fops); trace_create_file("trace_pipe", 0444, d_tracer, tr, &tracing_pipe_fops); @@ -6146,11 +6140,11 @@ init_tracer_debugfs(struct trace_array *tr, struct dentry *d_tracer) &trace_clock_fops); trace_create_file("tracing_on", 0644, d_tracer, - tr, &rb_simple_fops); + tr, &rb_simple_fops); #ifdef CONFIG_TRACER_SNAPSHOT trace_create_file("snapshot", 0644, d_tracer, - (void *)&tr->trace_cpu, &snapshot_fops); + tr, &snapshot_fops); #endif for_each_tracing_cpu(cpu) -- cgit v1.2.3 From 9c01fe4593db123c5a72dc36f0400f776e92c954 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 23 Jul 2013 17:26:13 +0200 Subject: tracing: Kill trace_cpu struct/members After the previous changes trace_array_cpu->trace_cpu and trace_array->trace_cpu becomes write-only. Remove these members and kill "struct trace_cpu" as well. As a side effect this also removes memset(per_cpu_memory, 0). It was not needed, alloc_percpu() returns zero-filled memory. Link: http://lkml.kernel.org/r/20130723152613.GA23741@redhat.com Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 21 --------------------- kernel/trace/trace.h | 8 -------- 2 files changed, 29 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index dd7780ddde08..69cba470ea96 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -5865,17 +5865,6 @@ struct dentry *trace_instance_dir; static void init_tracer_debugfs(struct trace_array *tr, struct dentry *d_tracer); -static void init_trace_buffers(struct trace_array *tr, struct trace_buffer *buf) -{ - int cpu; - - for_each_tracing_cpu(cpu) { - memset(per_cpu_ptr(buf->data, cpu), 0, sizeof(struct trace_array_cpu)); - per_cpu_ptr(buf->data, cpu)->trace_cpu.cpu = cpu; - per_cpu_ptr(buf->data, cpu)->trace_cpu.tr = tr; - } -} - static int allocate_trace_buffer(struct trace_array *tr, struct trace_buffer *buf, int size) { @@ -5893,8 +5882,6 @@ allocate_trace_buffer(struct trace_array *tr, struct trace_buffer *buf, int size return -ENOMEM; } - init_trace_buffers(tr, buf); - /* Allocate the first page for all buffers */ set_buffer_entries(&tr->trace_buffer, ring_buffer_size(tr->trace_buffer.buffer, 0)); @@ -5961,10 +5948,6 @@ static int new_instance_create(const char *name) if (allocate_trace_buffers(tr, trace_buf_size) < 0) goto out_free_tr; - /* Holder for file callbacks */ - tr->trace_cpu.cpu = RING_BUFFER_ALL_CPUS; - tr->trace_cpu.tr = tr; - tr->dir = debugfs_create_dir(name, trace_instance_dir); if (!tr->dir) goto out_free_tr; @@ -6438,10 +6421,6 @@ __init static int tracer_alloc_buffers(void) global_trace.flags = TRACE_ARRAY_FL_GLOBAL; - /* Holder for file callbacks */ - global_trace.trace_cpu.cpu = RING_BUFFER_ALL_CPUS; - global_trace.trace_cpu.tr = &global_trace; - INIT_LIST_HEAD(&global_trace.systems); INIT_LIST_HEAD(&global_trace.events); list_add(&global_trace.list, &ftrace_trace_arrays); diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index e7d643b8a907..afaae41b0a02 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -130,19 +130,12 @@ enum trace_flag_type { struct trace_array; -struct trace_cpu { - struct trace_array *tr; - struct dentry *dir; - int cpu; -}; - /* * The CPU trace array - it consists of thousands of trace entries * plus some other descriptor data: (for example which task started * the trace, etc.) */ struct trace_array_cpu { - struct trace_cpu trace_cpu; atomic_t disabled; void *buffer_page; /* ring buffer spare */ @@ -196,7 +189,6 @@ struct trace_array { bool allocated_snapshot; #endif int buffer_disabled; - struct trace_cpu trace_cpu; /* place holder */ #ifdef CONFIG_FTRACE_SYSCALLS int sys_refcount_enter; int sys_refcount_exit; -- cgit v1.2.3 From 195a8afc7ac962f8da795549fe38e825f1372b0d Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Tue, 23 Jul 2013 22:06:15 -0400 Subject: ftrace: Add check for NULL regs if ops has SAVE_REGS set If a ftrace ops is registered with the SAVE_REGS flag set, and there's already a ops registered to one of its functions but without the SAVE_REGS flag, there's a small race window where the SAVE_REGS ops gets added to the list of callbacks to call for that function before the callback trampoline gets set to save the regs. The problem is, the function is not currently saving regs, which opens a small race window where the ops that is expecting regs to be passed to it, wont. This can cause a crash if the callback were to reference the regs, as the SAVE_REGS guarantees that regs will be set. To fix this, we add a check in the loop case where it checks if the ops has the SAVE_REGS flag set, and if so, it will ignore it if regs is not set. Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 18 ++++++++++++++---- 1 file changed, 14 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 67708f46baae..8ce9eefc5bb4 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1441,12 +1441,22 @@ ftrace_hash_move(struct ftrace_ops *ops, int enable, * the hashes are freed with call_rcu_sched(). */ static int -ftrace_ops_test(struct ftrace_ops *ops, unsigned long ip) +ftrace_ops_test(struct ftrace_ops *ops, unsigned long ip, void *regs) { struct ftrace_hash *filter_hash; struct ftrace_hash *notrace_hash; int ret; +#ifdef CONFIG_DYNAMIC_FTRACE_WITH_REGS + /* + * There's a small race when adding ops that the ftrace handler + * that wants regs, may be called without them. We can not + * allow that handler to be called if regs is NULL. + */ + if (regs == NULL && (ops->flags & FTRACE_OPS_FL_SAVE_REGS)) + return 0; +#endif + filter_hash = rcu_dereference_raw_notrace(ops->filter_hash); notrace_hash = rcu_dereference_raw_notrace(ops->notrace_hash); @@ -4218,7 +4228,7 @@ static inline void ftrace_startup_enable(int command) { } # define ftrace_shutdown_sysctl() do { } while (0) static inline int -ftrace_ops_test(struct ftrace_ops *ops, unsigned long ip) +ftrace_ops_test(struct ftrace_ops *ops, unsigned long ip, void *regs) { return 1; } @@ -4241,7 +4251,7 @@ ftrace_ops_control_func(unsigned long ip, unsigned long parent_ip, do_for_each_ftrace_op(op, ftrace_control_list) { if (!(op->flags & FTRACE_OPS_FL_STUB) && !ftrace_function_local_disabled(op) && - ftrace_ops_test(op, ip)) + ftrace_ops_test(op, ip, regs)) op->func(ip, parent_ip, op, regs); } while_for_each_ftrace_op(op); trace_recursion_clear(TRACE_CONTROL_BIT); @@ -4274,7 +4284,7 @@ __ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip, */ preempt_disable_notrace(); do_for_each_ftrace_op(op, ftrace_ops_list) { - if (ftrace_ops_test(op, ip)) + if (ftrace_ops_test(op, ip, regs)) op->func(ip, parent_ip, op, regs); } while_for_each_ftrace_op(op); preempt_enable_notrace(); -- cgit v1.2.3 From c2fda509667b0fda4372a237f5a59ea4570b1627 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Wed, 24 Jul 2013 18:31:42 +0800 Subject: workqueue: allow work_on_cpu() to be called recursively If the @fn call work_on_cpu() again, the lockdep will complain: > [ INFO: possible recursive locking detected ] > 3.11.0-rc1-lockdep-fix-a #6 Not tainted > --------------------------------------------- > kworker/0:1/142 is trying to acquire lock: > ((&wfc.work)){+.+.+.}, at: [] flush_work+0x0/0xb0 > > but task is already holding lock: > ((&wfc.work)){+.+.+.}, at: [] process_one_work+0x169/0x610 > > other info that might help us debug this: > Possible unsafe locking scenario: > > CPU0 > ---- > lock((&wfc.work)); > lock((&wfc.work)); > > *** DEADLOCK *** It is false-positive lockdep report. In this sutiation, the two "wfc"s of the two work_on_cpu() are different, they are both on stack. flush_work() can't be deadlock. To fix this, we need to avoid the lockdep checking in this case, thus we instroduce a internal __flush_work() which skip the lockdep. tj: Minor comment adjustment. Signed-off-by: Lai Jiangshan Reported-by: "Srivatsa S. Bhat" Reported-by: Alexander Duyck Signed-off-by: Tejun Heo --- kernel/workqueue.c | 32 ++++++++++++++++++++++---------- 1 file changed, 22 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index f02c4a4a0c3c..55f5f0afcd0d 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -2817,6 +2817,19 @@ already_gone: return false; } +static bool __flush_work(struct work_struct *work) +{ + struct wq_barrier barr; + + if (start_flush_work(work, &barr)) { + wait_for_completion(&barr.done); + destroy_work_on_stack(&barr.work); + return true; + } else { + return false; + } +} + /** * flush_work - wait for a work to finish executing the last queueing instance * @work: the work to flush @@ -2830,18 +2843,10 @@ already_gone: */ bool flush_work(struct work_struct *work) { - struct wq_barrier barr; - lock_map_acquire(&work->lockdep_map); lock_map_release(&work->lockdep_map); - if (start_flush_work(work, &barr)) { - wait_for_completion(&barr.done); - destroy_work_on_stack(&barr.work); - return true; - } else { - return false; - } + return __flush_work(work); } EXPORT_SYMBOL_GPL(flush_work); @@ -4756,7 +4761,14 @@ long work_on_cpu(int cpu, long (*fn)(void *), void *arg) INIT_WORK_ONSTACK(&wfc.work, work_for_cpu_fn); schedule_work_on(cpu, &wfc.work); - flush_work(&wfc.work); + + /* + * The work item is on-stack and can't lead to deadlock through + * flushing. Use __flush_work() to avoid spurious lockdep warnings + * when work_on_cpu()s are nested. + */ + __flush_work(&wfc.work); + return wfc.ret; } EXPORT_SYMBOL_GPL(work_on_cpu); -- cgit v1.2.3 From 543487c7a2670bb0d96c00673a44b74360e3b6c1 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 16 Jul 2013 10:22:12 -0400 Subject: nohz: Do not warn about unstable tsc unless user uses nohz_full If the user enables CONFIG_NO_HZ_FULL and runs the kernel on a machine with an unstable TSC, it will produce a WARN_ON dump as well as taint the kernel. This is a bit extreme for a kernel that just enables a feature but doesn't use it. The warning should only happen if the user tries to use the feature by either adding nohz_full to the kernel command line, or by enabling CONFIG_NO_HZ_FULL_ALL that makes nohz used on all CPUs at boot up. Note, this second feature should not (yet) be used by distros or anyone that doesn't care if NO_HZ is used or not. Signed-off-by: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman Signed-off-by: Frederic Weisbecker --- kernel/time/tick-sched.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index e80183f4a6c4..1102534a1a57 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -182,7 +182,8 @@ static bool can_stop_full_tick(void) * Don't allow the user to think they can get * full NO_HZ with this machine. */ - WARN_ONCE(1, "NO_HZ FULL will not work with unstable sched clock"); + WARN_ONCE(have_nohz_full_mask, + "NO_HZ FULL will not work with unstable sched clock"); return false; } #endif -- cgit v1.2.3 From ca06416b2b4fa562cd3c3f9eb4198c3b2a983342 Mon Sep 17 00:00:00 2001 From: Li Zhong Date: Tue, 16 Jul 2013 12:18:47 +0800 Subject: nohz: fix compile warning in tick_nohz_init() cpu is not used after commit 5b8621a68fdcd2baf1d3b413726f913a5254d46a Signed-off-by: Li Zhong Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman Signed-off-by: Frederic Weisbecker --- kernel/time/tick-sched.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 1102534a1a57..9563c744dad2 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -344,8 +344,6 @@ static int tick_nohz_init_all(void) void __init tick_nohz_init(void) { - int cpu; - if (!have_nohz_full_mask) { if (tick_nohz_init_all() < 0) return; -- cgit v1.2.3 From 2b4883972271f8d61de67aa365ade89dfff69db1 Mon Sep 17 00:00:00 2001 From: Davidlohr Bueso Date: Wed, 24 Jul 2013 11:25:17 -0700 Subject: mutex: Avoid label warning when !CONFIG_MUTEX_SPIN_ON_OWNER Fengguang reported the following warning when optimistic spinning is disabled (ie: make allnoconfig): kernel/mutex.c:599:1: warning: label 'done' defined but not used Remove the 'done' label altogether. Reported-by: Fengguang Wu Signed-off-by: Davidlohr Bueso Signed-off-by: Ingo Molnar --- kernel/mutex.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/mutex.c b/kernel/mutex.c index 386ad5da47a5..98164a55a4dc 100644 --- a/kernel/mutex.c +++ b/kernel/mutex.c @@ -489,7 +489,8 @@ __mutex_lock_common(struct mutex *lock, long state, unsigned int subclass, mutex_set_owner(lock); mspin_unlock(MLOCK(lock), &node); - goto done; + preempt_enable(); + return 0; } mspin_unlock(MLOCK(lock), &node); @@ -596,7 +597,6 @@ skip_wait: } spin_unlock_mutex(&lock->wait_lock, flags); -done: preempt_enable(); return 0; -- cgit v1.2.3 From 3831261eb08557a1390b29c1038a0217232d8fdb Mon Sep 17 00:00:00 2001 From: "Brandt, Todd E" Date: Thu, 11 Jul 2013 07:44:35 +0000 Subject: PM / Sleep: increase ftrace coverage in suspend/resume Change where ftrace is disabled and re-enabled during system suspend/resume to allow tracing of device driver pm callbacks. Ftrace will now be turned off when suspend reaches disable_nonboot_cpus() instead of at the very beginning of system suspend. Ftrace was disabled during suspend/resume back in 2008 by Steven Rostedt as he discovered there was a conflict in the enable_nonboot_cpus() call (see commit f42ac38 "ftrace: disable tracing for suspend to ram"). This change preserves his fix by disabling ftrace, but only at the function where it is known to cause problems. The new change allows tracing of the device level code for better debug. [rjw: Changelog] Signed-off-by: Todd Brandt Signed-off-by: Rafael J. Wysocki --- kernel/power/suspend.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/power/suspend.c b/kernel/power/suspend.c index ece04223bb1e..62ee437b5c7e 100644 --- a/kernel/power/suspend.c +++ b/kernel/power/suspend.c @@ -210,6 +210,7 @@ static int suspend_enter(suspend_state_t state, bool *wakeup) goto Platform_wake; } + ftrace_stop(); error = disable_nonboot_cpus(); if (error || suspend_test(TEST_CPUS)) goto Enable_cpus; @@ -232,6 +233,7 @@ static int suspend_enter(suspend_state_t state, bool *wakeup) Enable_cpus: enable_nonboot_cpus(); + ftrace_start(); Platform_wake: if (need_suspend_ops(state) && suspend_ops->wake) @@ -265,7 +267,6 @@ int suspend_devices_and_enter(suspend_state_t state) goto Close; } suspend_console(); - ftrace_stop(); suspend_test_start(); error = dpm_suspend_start(PMSG_SUSPEND); if (error) { @@ -285,7 +286,6 @@ int suspend_devices_and_enter(suspend_state_t state) suspend_test_start(); dpm_resume_end(PMSG_RESUME); suspend_test_finish("resume devices"); - ftrace_start(); resume_console(); Close: if (need_suspend_ops(state) && suspend_ops->end) -- cgit v1.2.3 From 09d8091c024ec88d1541d93eb8ddb2bd5cf10c39 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Tue, 23 Jul 2013 22:21:59 -0400 Subject: tracing: Remove locking trace_types_lock from tracing_reset_all_online_cpus() Commit a82274151af "tracing: Protect ftrace_trace_arrays list in trace_events.c" added taking the trace_types_lock mutex in trace_events.c as there were several locations that needed it for protection. Unfortunately, it also encapsulated a call to tracing_reset_all_online_cpus() which also takes the trace_types_lock, causing a deadlock. This happens when a module has tracepoints and has been traced. When the module is removed, the trace events module notifier will grab the trace_types_lock, do a bunch of clean ups, and also clears the buffer by calling tracing_reset_all_online_cpus. This doesn't happen often which explains why it wasn't caught right away. Commit a82274151af was marked for stable, which means this must be sent to stable too. Link: http://lkml.kernel.org/r/51EEC646.7070306@broadcom.com Reported-by: Arend van Spril Tested-by: Arend van Spriel Cc: Alexander Z Lam Cc: Vaibhav Nagarnaik Cc: David Sharp Cc: stable@vger.kernel.org # 3.10 Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 69cba470ea96..882ec1dd1515 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1224,18 +1224,17 @@ void tracing_reset_current(int cpu) tracing_reset(&global_trace.trace_buffer, cpu); } +/* Must have trace_types_lock held */ void tracing_reset_all_online_cpus(void) { struct trace_array *tr; - mutex_lock(&trace_types_lock); list_for_each_entry(tr, &ftrace_trace_arrays, list) { tracing_reset_online_cpus(&tr->trace_buffer); #ifdef CONFIG_TRACER_MAX_TRACE tracing_reset_online_cpus(&tr->max_buffer); #endif } - mutex_unlock(&trace_types_lock); } #define SAVED_CMDLINES 128 -- cgit v1.2.3 From 102c9323c35a83789ad5ebd3c45fa8fb389add88 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Fri, 12 Jul 2013 17:07:27 -0400 Subject: tracing: Add __tracepoint_string() to export string pointers There are several tracepoints (mostly in RCU), that reference a string pointer and uses the print format of "%s" to display the string that exists in the kernel, instead of copying the actual string to the ring buffer (saves time and ring buffer space). But this has an issue with userspace tools that read the binary buffers that has the address of the string but has no access to what the string itself is. The end result is just output that looks like: rcu_dyntick: ffffffff818adeaa 1 0 rcu_dyntick: ffffffff818adeb5 0 140000000000000 rcu_dyntick: ffffffff818adeb5 0 140000000000000 rcu_utilization: ffffffff8184333b rcu_utilization: ffffffff8184333b The above is pretty useless when read by the userspace tools. Ideally we would want something that looks like this: rcu_dyntick: Start 1 0 rcu_dyntick: End 0 140000000000000 rcu_dyntick: Start 140000000000000 0 rcu_callback: rcu_preempt rhp=0xffff880037aff710 func=put_cred_rcu 0/4 rcu_callback: rcu_preempt rhp=0xffff880078961980 func=file_free_rcu 0/5 rcu_dyntick: End 0 1 The trace_printk() which also only stores the address of the string format instead of recording the string into the buffer itself, exports the mapping of kernel addresses to format strings via the printk_format file in the debugfs tracing directory. The tracepoint strings can use this same method and output the format to the same file and the userspace tools will be able to decipher the address without any modification. The tracepoint strings need its own section to save the strings because the trace_printk section will cause the trace_printk() buffers to be allocated if anything exists within the section. trace_printk() is only used for debugging and should never exist in the kernel, we can not use the trace_printk sections. Add a new tracepoint_str section that will also be examined by the output of the printk_format file. Cc: Paul E. McKenney Signed-off-by: Steven Rostedt --- kernel/trace/trace.h | 3 +++ kernel/trace/trace_printk.c | 19 +++++++++++++++++++ 2 files changed, 22 insertions(+) (limited to 'kernel') diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 4a4f6e1828b6..ba321f12df8c 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -1022,6 +1022,9 @@ extern struct list_head ftrace_events; extern const char *__start___trace_bprintk_fmt[]; extern const char *__stop___trace_bprintk_fmt[]; +extern const char *__start___tracepoint_str[]; +extern const char *__stop___tracepoint_str[]; + void trace_printk_init_buffers(void); void trace_printk_start_comm(void); int trace_keep_overwrite(struct tracer *tracer, u32 mask, int set); diff --git a/kernel/trace/trace_printk.c b/kernel/trace/trace_printk.c index a9077c1b4ad3..2900817ba65c 100644 --- a/kernel/trace/trace_printk.c +++ b/kernel/trace/trace_printk.c @@ -244,12 +244,31 @@ static const char **find_next(void *v, loff_t *pos) { const char **fmt = v; int start_index; + int last_index; start_index = __stop___trace_bprintk_fmt - __start___trace_bprintk_fmt; if (*pos < start_index) return __start___trace_bprintk_fmt + *pos; + /* + * The __tracepoint_str section is treated the same as the + * __trace_printk_fmt section. The difference is that the + * __trace_printk_fmt section should only be used by trace_printk() + * in a debugging environment, as if anything exists in that section + * the trace_prink() helper buffers are allocated, which would just + * waste space in a production environment. + * + * The __tracepoint_str sections on the other hand are used by + * tracepoints which need to map pointers to their strings to + * the ASCII text for userspace. + */ + last_index = start_index; + start_index = __stop___tracepoint_str - __start___tracepoint_str; + + if (*pos < last_index + start_index) + return __start___tracepoint_str + (*pos - last_index); + return find_next_mod_format(start_index, v, fmt, pos); } -- cgit v1.2.3 From d738ce8fdc05ebf5b1475f8ae26d908c8c50970b Mon Sep 17 00:00:00 2001 From: Francesco Fusco Date: Wed, 24 Jul 2013 10:39:07 +0200 Subject: sysctl: range checking in do_proc_dointvec_ms_jiffies_conv When (integer) sysctl values are expressed in ms and have to be represented internally as jiffies. The msecs_to_jiffies function returns an unsigned long, which gets assigned to the integer. This patch prevents the value to be assigned if bigger than INT_MAX, done in a similar way as in cba9f3 ("Range checking in do_proc_dointvec_(userhz_)jiffies_conv"). Signed-off-by: Francesco Fusco CC: Andrew Morton CC: linux-kernel@vger.kernel.org Signed-off-by: David S. Miller --- kernel/sysctl.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index ac09d98490aa..07f6fc468e17 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -2346,7 +2346,11 @@ static int do_proc_dointvec_ms_jiffies_conv(bool *negp, unsigned long *lvalp, int write, void *data) { if (write) { - *valp = msecs_to_jiffies(*negp ? -*lvalp : *lvalp); + unsigned long jif = msecs_to_jiffies(*negp ? -*lvalp : *lvalp); + + if (jif > INT_MAX) + return 1; + *valp = (int)jif; } else { int val = *valp; unsigned long lval; -- cgit v1.2.3 From 148519120c6d1f19ad53349683aeae9f228b0b8d Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Sat, 27 Jul 2013 01:41:34 +0200 Subject: Revert "cpuidle: Quickly notice prediction failure for repeat mode" Revert commit 69a37bea (cpuidle: Quickly notice prediction failure for repeat mode), because it has been identified as the source of a significant performance regression in v3.8 and later as explained by Jeremy Eder: We believe we've identified a particular commit to the cpuidle code that seems to be impacting performance of variety of workloads. The simplest way to reproduce is using netperf TCP_RR test, so we're using that, on a pair of Sandy Bridge based servers. We also have data from a large database setup where performance is also measurably/positively impacted, though that test data isn't easily share-able. Included below are test results from 3 test kernels: kernel reverts ----------------------------------------------------------- 1) vanilla upstream (no reverts) 2) perfteam2 reverts e11538d1f03914eb92af5a1a378375c05ae8520c 3) test reverts 69a37beabf1f0a6705c08e879bdd5d82ff6486c4 e11538d1f03914eb92af5a1a378375c05ae8520c In summary, netperf TCP_RR numbers improve by approximately 4% after reverting 69a37beabf1f0a6705c08e879bdd5d82ff6486c4. When 69a37beabf1f0a6705c08e879bdd5d82ff6486c4 is included, C0 residency never seems to get above 40%. Taking that patch out gets C0 near 100% quite often, and performance increases. The below data are histograms representing the %c0 residency @ 1-second sample rates (using turbostat), while under netperf test. - If you look at the first 4 histograms, you can see %c0 residency almost entirely in the 30,40% bin. - The last pair, which reverts 69a37beabf1f0a6705c08e879bdd5d82ff6486c4, shows %c0 in the 80,90,100% bins. Below each kernel name are netperf TCP_RR trans/s numbers for the particular kernel that can be disclosed publicly, comparing the 3 test kernels. We ran a 4th test with the vanilla kernel where we've also set /dev/cpu_dma_latency=0 to show overall impact boosting single-threaded TCP_RR performance over 11% above baseline. 3.10-rc2 vanilla RX + c0 lock (/dev/cpu_dma_latency=0): TCP_RR trans/s 54323.78 ----------------------------------------------------------- 3.10-rc2 vanilla RX (no reverts) TCP_RR trans/s 48192.47 Receiver %c0 0.0000 - 10.0000 [ 1]: * 10.0000 - 20.0000 [ 0]: 20.0000 - 30.0000 [ 0]: 30.0000 - 40.0000 [ 59]: *********************************************************** 40.0000 - 50.0000 [ 1]: * 50.0000 - 60.0000 [ 0]: 60.0000 - 70.0000 [ 0]: 70.0000 - 80.0000 [ 0]: 80.0000 - 90.0000 [ 0]: 90.0000 - 100.0000 [ 0]: Sender %c0 0.0000 - 10.0000 [ 1]: * 10.0000 - 20.0000 [ 0]: 20.0000 - 30.0000 [ 0]: 30.0000 - 40.0000 [ 11]: *********** 40.0000 - 50.0000 [ 49]: ************************************************* 50.0000 - 60.0000 [ 0]: 60.0000 - 70.0000 [ 0]: 70.0000 - 80.0000 [ 0]: 80.0000 - 90.0000 [ 0]: 90.0000 - 100.0000 [ 0]: ----------------------------------------------------------- 3.10-rc2 perfteam2 RX (reverts commit e11538d1f03914eb92af5a1a378375c05ae8520c) TCP_RR trans/s 49698.69 Receiver %c0 0.0000 - 10.0000 [ 1]: * 10.0000 - 20.0000 [ 1]: * 20.0000 - 30.0000 [ 0]: 30.0000 - 40.0000 [ 59]: *********************************************************** 40.0000 - 50.0000 [ 0]: 50.0000 - 60.0000 [ 0]: 60.0000 - 70.0000 [ 0]: 70.0000 - 80.0000 [ 0]: 80.0000 - 90.0000 [ 0]: 90.0000 - 100.0000 [ 0]: Sender %c0 0.0000 - 10.0000 [ 1]: * 10.0000 - 20.0000 [ 0]: 20.0000 - 30.0000 [ 0]: 30.0000 - 40.0000 [ 2]: ** 40.0000 - 50.0000 [ 58]: ********************************************************** 50.0000 - 60.0000 [ 0]: 60.0000 - 70.0000 [ 0]: 70.0000 - 80.0000 [ 0]: 80.0000 - 90.0000 [ 0]: 90.0000 - 100.0000 [ 0]: ----------------------------------------------------------- 3.10-rc2 test RX (reverts 69a37beabf1f0a6705c08e879bdd5d82ff6486c4 and e11538d1f03914eb92af5a1a378375c05ae8520c) TCP_RR trans/s 47766.95 Receiver %c0 0.0000 - 10.0000 [ 1]: * 10.0000 - 20.0000 [ 1]: * 20.0000 - 30.0000 [ 0]: 30.0000 - 40.0000 [ 27]: *************************** 40.0000 - 50.0000 [ 2]: ** 50.0000 - 60.0000 [ 0]: 60.0000 - 70.0000 [ 2]: ** 70.0000 - 80.0000 [ 0]: 80.0000 - 90.0000 [ 0]: 90.0000 - 100.0000 [ 28]: **************************** Sender: 0.0000 - 10.0000 [ 1]: * 10.0000 - 20.0000 [ 0]: 20.0000 - 30.0000 [ 0]: 30.0000 - 40.0000 [ 11]: *********** 40.0000 - 50.0000 [ 0]: 50.0000 - 60.0000 [ 1]: * 60.0000 - 70.0000 [ 0]: 70.0000 - 80.0000 [ 3]: *** 80.0000 - 90.0000 [ 7]: ******* 90.0000 - 100.0000 [ 38]: ************************************** These results demonstrate gaining back the tendency of the CPU to stay in more responsive, performant C-states (and thus yield measurably better performance), by reverting commit 69a37beabf1f0a6705c08e879bdd5d82ff6486c4. Requested-by: Jeremy Eder Tested-by: Len Brown Cc: 3.8+ Signed-off-by: Rafael J. Wysocki --- kernel/time/tick-sched.c | 9 ++------- 1 file changed, 2 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index e80183f4a6c4..e77edc97e036 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -827,13 +827,10 @@ void tick_nohz_irq_exit(void) { struct tick_sched *ts = &__get_cpu_var(tick_cpu_sched); - if (ts->inidle) { - /* Cancel the timer because CPU already waken up from the C-states*/ - menu_hrtimer_cancel(); + if (ts->inidle) __tick_nohz_idle_enter(ts); - } else { + else tick_nohz_full_stop_tick(ts); - } } /** @@ -931,8 +928,6 @@ void tick_nohz_idle_exit(void) ts->inidle = 0; - /* Cancel the timer because CPU already waken up from the C-states*/ - menu_hrtimer_cancel(); if (ts->idle_active || ts->tick_stopped) now = ktime_get(); -- cgit v1.2.3 From 9ad9d25a1ec82d6e52d687348e8cdd4942e7d393 Mon Sep 17 00:00:00 2001 From: Zhao Hongjiang Date: Sat, 27 Jul 2013 11:56:49 +0800 Subject: cpuset: get rid of the useless forward declaration of cpuset get rid of the useless forward declaration of the struct cpuset cause the below define it. Signed-off-by: Zhao Hongjiang Acked-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cpuset.c | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index e5657788fedd..2ddd9b93feaa 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -70,7 +70,6 @@ int number_of_cpusets __read_mostly; /* Forward declare cgroup structures */ struct cgroup_subsys cpuset_subsys; -struct cpuset; /* See "Frequency meter" comments, below. */ -- cgit v1.2.3 From 0b9e6965add0701e5cbf56d5bab6d9181e948359 Mon Sep 17 00:00:00 2001 From: Zhao Hongjiang Date: Sat, 27 Jul 2013 11:56:53 +0800 Subject: cpuset: relocate a misplaced comment Comment for cpuset_css_offline() was on top of cpuset_css_free(). Move it. Signed-off-by: Zhao Hongjiang Acked-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cpuset.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 2ddd9b93feaa..703bfd5a32a9 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -2020,6 +2020,12 @@ out_unlock: return 0; } +/* + * If the cpuset being removed has its flag 'sched_load_balance' + * enabled, then simulate turning sched_load_balance off, which + * will call rebuild_sched_domains_locked(). + */ + static void cpuset_css_offline(struct cgroup *cgrp) { struct cpuset *cs = cgroup_cs(cgrp); @@ -2035,12 +2041,6 @@ static void cpuset_css_offline(struct cgroup *cgrp) mutex_unlock(&cpuset_mutex); } -/* - * If the cpuset being removed has its flag 'sched_load_balance' - * enabled, then simulate turning sched_load_balance off, which - * will call rebuild_sched_domains_locked(). - */ - static void cpuset_css_free(struct cgroup *cgrp) { struct cpuset *cs = cgroup_cs(cgrp); -- cgit v1.2.3 From e66c33d579ea566d10e8c8695a7168aae3e02992 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Fri, 12 Jul 2013 16:50:28 -0400 Subject: rcu: Add const annotation to char * for RCU tracepoints and functions All the RCU tracepoints and functions that reference char pointers do so with just 'char *' even though they do not modify the contents of the string itself. This will cause warnings if a const char * is used in one of these functions. The RCU tracepoints store the pointer to the string to refer back to them when the trace output is displayed. As this can be minutes, hours or even days later, those strings had better be constant. This change also opens the door to allow the RCU tracepoint strings and their addresses to be exported so that userspace tracing tools can translate the contents of the pointers of the RCU tracepoints. Signed-off-by: Steven Rostedt --- kernel/rcu.h | 2 +- kernel/rcupdate.c | 2 +- kernel/rcutiny.c | 2 +- kernel/rcutiny_plugin.h | 2 +- kernel/rcutorture.c | 8 ++++---- kernel/rcutree.c | 4 ++-- kernel/rcutree.h | 2 +- 7 files changed, 11 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/rcu.h b/kernel/rcu.h index 7f8e7590e3e5..0a90ccc65bfb 100644 --- a/kernel/rcu.h +++ b/kernel/rcu.h @@ -94,7 +94,7 @@ static inline void debug_rcu_head_unqueue(struct rcu_head *head) extern void kfree(const void *); -static inline bool __rcu_reclaim(char *rn, struct rcu_head *head) +static inline bool __rcu_reclaim(const char *rn, struct rcu_head *head) { unsigned long offset = (unsigned long)head->func; diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c index cce6ba8bbace..14994d4e1a54 100644 --- a/kernel/rcupdate.c +++ b/kernel/rcupdate.c @@ -377,7 +377,7 @@ EXPORT_SYMBOL_GPL(rcuhead_debug_descr); #endif /* #ifdef CONFIG_DEBUG_OBJECTS_RCU_HEAD */ #if defined(CONFIG_TREE_RCU) || defined(CONFIG_TREE_PREEMPT_RCU) || defined(CONFIG_RCU_TRACE) -void do_trace_rcu_torture_read(char *rcutorturename, struct rcu_head *rhp, +void do_trace_rcu_torture_read(const char *rcutorturename, struct rcu_head *rhp, unsigned long secs, unsigned long c_old, unsigned long c) { diff --git a/kernel/rcutiny.c b/kernel/rcutiny.c index aa344111de3e..9ed6075dc562 100644 --- a/kernel/rcutiny.c +++ b/kernel/rcutiny.c @@ -264,7 +264,7 @@ void rcu_check_callbacks(int cpu, int user) */ static void __rcu_process_callbacks(struct rcu_ctrlblk *rcp) { - char *rn = NULL; + const char *rn = NULL; struct rcu_head *next, *list; unsigned long flags; RCU_TRACE(int cb_count = 0); diff --git a/kernel/rcutiny_plugin.h b/kernel/rcutiny_plugin.h index 0cd385acccfa..280d06cae352 100644 --- a/kernel/rcutiny_plugin.h +++ b/kernel/rcutiny_plugin.h @@ -36,7 +36,7 @@ struct rcu_ctrlblk { RCU_TRACE(unsigned long gp_start); /* Start time for stalls. */ RCU_TRACE(unsigned long ticks_this_gp); /* Statistic for stalls. */ RCU_TRACE(unsigned long jiffies_stall); /* Jiffies at next stall. */ - RCU_TRACE(char *name); /* Name of RCU type. */ + RCU_TRACE(const char *name); /* Name of RCU type. */ }; /* Definition for rcupdate control block. */ diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c index f4871e52c546..3d936f0fbcd8 100644 --- a/kernel/rcutorture.c +++ b/kernel/rcutorture.c @@ -267,7 +267,7 @@ rcutorture_shutdown_notify(struct notifier_block *unused1, * Absorb kthreads into a kernel function that won't return, so that * they won't ever access module text or data again. */ -static void rcutorture_shutdown_absorb(char *title) +static void rcutorture_shutdown_absorb(const char *title) { if (ACCESS_ONCE(fullstop) == FULLSTOP_SHUTDOWN) { pr_notice( @@ -337,7 +337,7 @@ rcu_random(struct rcu_random_state *rrsp) } static void -rcu_stutter_wait(char *title) +rcu_stutter_wait(const char *title) { while (stutter_pause_test || !rcutorture_runnable) { if (rcutorture_runnable) @@ -366,7 +366,7 @@ struct rcu_torture_ops { int (*stats)(char *page); int irq_capable; int can_boost; - char *name; + const char *name; }; static struct rcu_torture_ops *cur_ops; @@ -1364,7 +1364,7 @@ rcu_torture_stutter(void *arg) } static inline void -rcu_torture_print_module_parms(struct rcu_torture_ops *cur_ops, char *tag) +rcu_torture_print_module_parms(struct rcu_torture_ops *cur_ops, const char *tag) { pr_alert("%s" TORTURE_FLAG "--- %s: nreaders=%d nfakewriters=%d " diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 068de3a93606..30201494560b 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1032,7 +1032,7 @@ static unsigned long rcu_cbs_completed(struct rcu_state *rsp, * rcu_nocb_wait_gp(). */ static void trace_rcu_future_gp(struct rcu_node *rnp, struct rcu_data *rdp, - unsigned long c, char *s) + unsigned long c, const char *s) { trace_rcu_future_grace_period(rdp->rsp->name, rnp->gpnum, rnp->completed, c, rnp->level, @@ -2720,7 +2720,7 @@ static int rcu_cpu_has_callbacks(int cpu, bool *all_lazy) * Helper function for _rcu_barrier() tracing. If tracing is disabled, * the compiler is expected to optimize this away. */ -static void _rcu_barrier_trace(struct rcu_state *rsp, char *s, +static void _rcu_barrier_trace(struct rcu_state *rsp, const char *s, int cpu, unsigned long done) { trace_rcu_barrier(rsp->name, s, cpu, diff --git a/kernel/rcutree.h b/kernel/rcutree.h index b3832581043c..cbdeac6cea9e 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -445,7 +445,7 @@ struct rcu_state { /* for CPU stalls. */ unsigned long gp_max; /* Maximum GP duration in */ /* jiffies. */ - char *name; /* Name of structure. */ + const char *name; /* Name of structure. */ char abbr; /* Abbreviated name. */ struct list_head flavors; /* List of RCU flavors. */ struct irq_work wakeup_work; /* Postponed wakeups */ -- cgit v1.2.3 From a41bfeb2f8ed59410be7ca0f8fbc6138a758b746 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Fri, 12 Jul 2013 17:00:28 -0400 Subject: rcu: Simplify RCU_STATE_INITIALIZER() macro The RCU_STATE_INITIALIZER() macro is used only in the rcutree.c file as well as the rcutree_plugin.h file. It is passed as a rvalue to a variable of a similar name. A per_cpu variable is also created with a similar name as well. The uses of RCU_STATE_INITIALIZER() can be simplified to remove some of the duplicate code that is done. Currently the three users of this macro has this format: struct rcu_state rcu_sched_state = RCU_STATE_INITIALIZER(rcu_sched, call_rcu_sched); DEFINE_PER_CPU(struct rcu_data, rcu_sched_data); Notice that "rcu_sched" is called three times. This is the same with the other two users. This can be condensed to just: RCU_STATE_INITIALIZER(rcu_sched, call_rcu_sched); by moving the rest into the macro itself. This also opens the door to allow the RCU tracepoint strings and their addresses to be exported so that userspace tracing tools can translate the contents of the pointers of the RCU tracepoints. The change will allow for helper code to be placed in the RCU_STATE_INITIALIZER() macro to export the name that is used. Signed-off-by: Steven Rostedt --- kernel/rcutree.c | 14 ++++++-------- kernel/rcutree_plugin.h | 4 +--- 2 files changed, 7 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 30201494560b..97994a329d80 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -64,7 +64,8 @@ static struct lock_class_key rcu_node_class[RCU_NUM_LVLS]; static struct lock_class_key rcu_fqs_class[RCU_NUM_LVLS]; -#define RCU_STATE_INITIALIZER(sname, sabbr, cr) { \ +#define RCU_STATE_INITIALIZER(sname, sabbr, cr) \ +struct rcu_state sname##_state = { \ .level = { &sname##_state.node[0] }, \ .call = cr, \ .fqs_state = RCU_GP_IDLE, \ @@ -77,14 +78,11 @@ static struct lock_class_key rcu_fqs_class[RCU_NUM_LVLS]; .onoff_mutex = __MUTEX_INITIALIZER(sname##_state.onoff_mutex), \ .name = #sname, \ .abbr = sabbr, \ -} - -struct rcu_state rcu_sched_state = - RCU_STATE_INITIALIZER(rcu_sched, 's', call_rcu_sched); -DEFINE_PER_CPU(struct rcu_data, rcu_sched_data); +}; \ +DEFINE_PER_CPU(struct rcu_data, sname##_data) -struct rcu_state rcu_bh_state = RCU_STATE_INITIALIZER(rcu_bh, 'b', call_rcu_bh); -DEFINE_PER_CPU(struct rcu_data, rcu_bh_data); +RCU_STATE_INITIALIZER(rcu_sched, 's', call_rcu_sched); +RCU_STATE_INITIALIZER(rcu_bh, 'b', call_rcu_bh); static struct rcu_state *rcu_state; LIST_HEAD(rcu_struct_flavors); diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 769e12e3151b..6976a7dde874 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -110,9 +110,7 @@ static void __init rcu_bootup_announce_oddness(void) #ifdef CONFIG_TREE_PREEMPT_RCU -struct rcu_state rcu_preempt_state = - RCU_STATE_INITIALIZER(rcu_preempt, 'p', call_rcu); -DEFINE_PER_CPU(struct rcu_data, rcu_preempt_data); +RCU_STATE_INITIALIZER(rcu_preempt, 'p', call_rcu); static struct rcu_state *rcu_state = &rcu_preempt_state; static int rcu_preempted_readers_exp(struct rcu_node *rnp); -- cgit v1.2.3 From f7f7bac9cb1c50783f15937a11743655a5756a36 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Fri, 12 Jul 2013 17:18:47 -0400 Subject: rcu: Have the RCU tracepoints use the tracepoint_string infrastructure Currently, RCU tracepoints save only a pointer to strings in the ring buffer. When displayed via the /sys/kernel/debug/tracing/trace file they are referenced like the printf "%s" that looks at the address in the ring buffer and prints out the string it points too. This requires that the strings are constant and persistent in the kernel. The problem with this is for tools like trace-cmd and perf that read the binary data from the buffers but have no access to the kernel memory to find out what string is represented by the address in the buffer. By using the tracepoint_string infrastructure, the RCU tracepoint strings can be exported such that userspace tools can map the addresses to the strings. # cat /sys/kernel/debug/tracing/printk_formats 0xffffffff81a4a0e8 : "rcu_preempt" 0xffffffff81a4a0f4 : "rcu_bh" 0xffffffff81a4a100 : "rcu_sched" 0xffffffff818437a0 : "cpuqs" 0xffffffff818437a6 : "rcu_sched" 0xffffffff818437a0 : "cpuqs" 0xffffffff818437b0 : "rcu_bh" 0xffffffff818437b7 : "Start context switch" 0xffffffff818437cc : "End context switch" 0xffffffff818437a0 : "cpuqs" [...] Now userspaces tools can display: rcu_utilization: Start context switch rcu_dyntick: Start 1 0 rcu_utilization: End context switch rcu_batch_start: rcu_preempt CBs=0/5 bl=10 rcu_dyntick: End 0 140000000000000 rcu_invoke_callback: rcu_preempt rhp=0xffff880071c0d600 func=proc_i_callback rcu_invoke_callback: rcu_preempt rhp=0xffff880077b5b230 func=__d_free rcu_dyntick: Start 140000000000000 0 rcu_invoke_callback: rcu_preempt rhp=0xffff880077563980 func=file_free_rcu rcu_batch_end: rcu_preempt CBs-invoked=3 idle=>c<>c<>c<>c< rcu_utilization: End RCU core rcu_grace_period: rcu_preempt 9741 start rcu_dyntick: Start 1 0 rcu_dyntick: End 0 140000000000000 rcu_dyntick: Start 140000000000000 0 Instead of: rcu_utilization: ffffffff81843110 rcu_future_grace_period: ffffffff81842f1d 9939 9939 9940 0 0 3 ffffffff81842f32 rcu_batch_start: ffffffff81842f1d CBs=0/4 bl=10 rcu_future_grace_period: ffffffff81842f1d 9939 9939 9940 0 0 3 ffffffff81842f3c rcu_grace_period: ffffffff81842f1d 9939 ffffffff81842f80 rcu_invoke_callback: ffffffff81842f1d rhp=0xffff88007888aac0 func=file_free_rcu rcu_grace_period: ffffffff81842f1d 9939 ffffffff81842f95 rcu_invoke_callback: ffffffff81842f1d rhp=0xffff88006aeb4600 func=proc_i_callback rcu_future_grace_period: ffffffff81842f1d 9939 9939 9940 0 0 3 ffffffff81842f32 rcu_future_grace_period: ffffffff81842f1d 9939 9939 9940 0 0 3 ffffffff81842f3c rcu_invoke_callback: ffffffff81842f1d rhp=0xffff880071cb9fc0 func=__d_free rcu_grace_period: ffffffff81842f1d 9939 ffffffff81842f80 rcu_invoke_callback: ffffffff81842f1d rhp=0xffff88007888ae80 func=file_free_rcu rcu_batch_end: ffffffff81842f1d CBs-invoked=4 idle=>c<>c<>c<>c< rcu_utilization: ffffffff8184311f Signed-off-by: Steven Rostedt --- kernel/rcutree.c | 87 ++++++++++++++++++++++++++++++------------------- kernel/rcutree_plugin.h | 32 +++++++++--------- 2 files changed, 69 insertions(+), 50 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 97994a329d80..338f1d1c1c66 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -53,18 +53,36 @@ #include #include #include +#include #include "rcutree.h" #include #include "rcu.h" +/* + * Strings used in tracepoints need to be exported via the + * tracing system such that tools like perf and trace-cmd can + * translate the string address pointers to actual text. + */ +#define TPS(x) tracepoint_string(x) + /* Data structures. */ static struct lock_class_key rcu_node_class[RCU_NUM_LVLS]; static struct lock_class_key rcu_fqs_class[RCU_NUM_LVLS]; +/* + * In order to export the rcu_state name to the tracing tools, it + * needs to be added in the __tracepoint_string section. + * This requires defining a separate variable tp__varname + * that points to the string being used, and this will allow + * the tracing userspace tools to be able to decipher the string + * address to the matching string. + */ #define RCU_STATE_INITIALIZER(sname, sabbr, cr) \ +static char sname##_varname[] = #sname; \ +static const char *tp_##sname##_varname __used __tracepoint_string = sname##_varname; \ struct rcu_state sname##_state = { \ .level = { &sname##_state.node[0] }, \ .call = cr, \ @@ -76,7 +94,7 @@ struct rcu_state sname##_state = { \ .orphan_donetail = &sname##_state.orphan_donelist, \ .barrier_mutex = __MUTEX_INITIALIZER(sname##_state.barrier_mutex), \ .onoff_mutex = __MUTEX_INITIALIZER(sname##_state.onoff_mutex), \ - .name = #sname, \ + .name = sname##_varname, \ .abbr = sabbr, \ }; \ DEFINE_PER_CPU(struct rcu_data, sname##_data) @@ -176,7 +194,7 @@ void rcu_sched_qs(int cpu) struct rcu_data *rdp = &per_cpu(rcu_sched_data, cpu); if (rdp->passed_quiesce == 0) - trace_rcu_grace_period("rcu_sched", rdp->gpnum, "cpuqs"); + trace_rcu_grace_period(TPS("rcu_sched"), rdp->gpnum, TPS("cpuqs")); rdp->passed_quiesce = 1; } @@ -185,7 +203,7 @@ void rcu_bh_qs(int cpu) struct rcu_data *rdp = &per_cpu(rcu_bh_data, cpu); if (rdp->passed_quiesce == 0) - trace_rcu_grace_period("rcu_bh", rdp->gpnum, "cpuqs"); + trace_rcu_grace_period(TPS("rcu_bh"), rdp->gpnum, TPS("cpuqs")); rdp->passed_quiesce = 1; } @@ -196,10 +214,10 @@ void rcu_bh_qs(int cpu) */ void rcu_note_context_switch(int cpu) { - trace_rcu_utilization("Start context switch"); + trace_rcu_utilization(TPS("Start context switch")); rcu_sched_qs(cpu); rcu_preempt_note_context_switch(cpu); - trace_rcu_utilization("End context switch"); + trace_rcu_utilization(TPS("End context switch")); } EXPORT_SYMBOL_GPL(rcu_note_context_switch); @@ -343,11 +361,11 @@ static struct rcu_node *rcu_get_root(struct rcu_state *rsp) static void rcu_eqs_enter_common(struct rcu_dynticks *rdtp, long long oldval, bool user) { - trace_rcu_dyntick("Start", oldval, rdtp->dynticks_nesting); + trace_rcu_dyntick(TPS("Start"), oldval, rdtp->dynticks_nesting); if (!user && !is_idle_task(current)) { struct task_struct *idle = idle_task(smp_processor_id()); - trace_rcu_dyntick("Error on entry: not idle task", oldval, 0); + trace_rcu_dyntick(TPS("Error on entry: not idle task"), oldval, 0); ftrace_dump(DUMP_ORIG); WARN_ONCE(1, "Current pid: %d comm: %s / Idle pid: %d comm: %s", current->pid, current->comm, @@ -477,7 +495,7 @@ void rcu_irq_exit(void) rdtp->dynticks_nesting--; WARN_ON_ONCE(rdtp->dynticks_nesting < 0); if (rdtp->dynticks_nesting) - trace_rcu_dyntick("--=", oldval, rdtp->dynticks_nesting); + trace_rcu_dyntick(TPS("--="), oldval, rdtp->dynticks_nesting); else rcu_eqs_enter_common(rdtp, oldval, true); local_irq_restore(flags); @@ -499,11 +517,11 @@ static void rcu_eqs_exit_common(struct rcu_dynticks *rdtp, long long oldval, smp_mb__after_atomic_inc(); /* See above. */ WARN_ON_ONCE(!(atomic_read(&rdtp->dynticks) & 0x1)); rcu_cleanup_after_idle(smp_processor_id()); - trace_rcu_dyntick("End", oldval, rdtp->dynticks_nesting); + trace_rcu_dyntick(TPS("End"), oldval, rdtp->dynticks_nesting); if (!user && !is_idle_task(current)) { struct task_struct *idle = idle_task(smp_processor_id()); - trace_rcu_dyntick("Error on exit: not idle task", + trace_rcu_dyntick(TPS("Error on exit: not idle task"), oldval, rdtp->dynticks_nesting); ftrace_dump(DUMP_ORIG); WARN_ONCE(1, "Current pid: %d comm: %s / Idle pid: %d comm: %s", @@ -618,7 +636,7 @@ void rcu_irq_enter(void) rdtp->dynticks_nesting++; WARN_ON_ONCE(rdtp->dynticks_nesting == 0); if (oldval) - trace_rcu_dyntick("++=", oldval, rdtp->dynticks_nesting); + trace_rcu_dyntick(TPS("++="), oldval, rdtp->dynticks_nesting); else rcu_eqs_exit_common(rdtp, oldval, true); local_irq_restore(flags); @@ -773,7 +791,7 @@ static int rcu_implicit_dynticks_qs(struct rcu_data *rdp) * of the current RCU grace period. */ if ((curr & 0x1) == 0 || UINT_CMP_GE(curr, snap + 2)) { - trace_rcu_fqs(rdp->rsp->name, rdp->gpnum, rdp->cpu, "dti"); + trace_rcu_fqs(rdp->rsp->name, rdp->gpnum, rdp->cpu, TPS("dti")); rdp->dynticks_fqs++; return 1; } @@ -793,7 +811,7 @@ static int rcu_implicit_dynticks_qs(struct rcu_data *rdp) return 0; /* Grace period is not old enough. */ barrier(); if (cpu_is_offline(rdp->cpu)) { - trace_rcu_fqs(rdp->rsp->name, rdp->gpnum, rdp->cpu, "ofl"); + trace_rcu_fqs(rdp->rsp->name, rdp->gpnum, rdp->cpu, TPS("ofl")); rdp->offline_fqs++; return 1; } @@ -1056,9 +1074,9 @@ rcu_start_future_gp(struct rcu_node *rnp, struct rcu_data *rdp) * grace period is already marked as needed, return to the caller. */ c = rcu_cbs_completed(rdp->rsp, rnp); - trace_rcu_future_gp(rnp, rdp, c, "Startleaf"); + trace_rcu_future_gp(rnp, rdp, c, TPS("Startleaf")); if (rnp->need_future_gp[c & 0x1]) { - trace_rcu_future_gp(rnp, rdp, c, "Prestartleaf"); + trace_rcu_future_gp(rnp, rdp, c, TPS("Prestartleaf")); return c; } @@ -1072,7 +1090,7 @@ rcu_start_future_gp(struct rcu_node *rnp, struct rcu_data *rdp) if (rnp->gpnum != rnp->completed || ACCESS_ONCE(rnp->gpnum) != ACCESS_ONCE(rnp->completed)) { rnp->need_future_gp[c & 0x1]++; - trace_rcu_future_gp(rnp, rdp, c, "Startedleaf"); + trace_rcu_future_gp(rnp, rdp, c, TPS("Startedleaf")); return c; } @@ -1100,7 +1118,7 @@ rcu_start_future_gp(struct rcu_node *rnp, struct rcu_data *rdp) * recorded, trace and leave. */ if (rnp_root->need_future_gp[c & 0x1]) { - trace_rcu_future_gp(rnp, rdp, c, "Prestartedroot"); + trace_rcu_future_gp(rnp, rdp, c, TPS("Prestartedroot")); goto unlock_out; } @@ -1109,9 +1127,9 @@ rcu_start_future_gp(struct rcu_node *rnp, struct rcu_data *rdp) /* If a grace period is not already in progress, start one. */ if (rnp_root->gpnum != rnp_root->completed) { - trace_rcu_future_gp(rnp, rdp, c, "Startedleafroot"); + trace_rcu_future_gp(rnp, rdp, c, TPS("Startedleafroot")); } else { - trace_rcu_future_gp(rnp, rdp, c, "Startedroot"); + trace_rcu_future_gp(rnp, rdp, c, TPS("Startedroot")); rcu_start_gp_advanced(rdp->rsp, rnp_root, rdp); } unlock_out: @@ -1135,7 +1153,8 @@ static int rcu_future_gp_cleanup(struct rcu_state *rsp, struct rcu_node *rnp) rcu_nocb_gp_cleanup(rsp, rnp); rnp->need_future_gp[c & 0x1] = 0; needmore = rnp->need_future_gp[(c + 1) & 0x1]; - trace_rcu_future_gp(rnp, rdp, c, needmore ? "CleanupMore" : "Cleanup"); + trace_rcu_future_gp(rnp, rdp, c, + needmore ? TPS("CleanupMore") : TPS("Cleanup")); return needmore; } @@ -1203,9 +1222,9 @@ static void rcu_accelerate_cbs(struct rcu_state *rsp, struct rcu_node *rnp, /* Trace depending on how much we were able to accelerate. */ if (!*rdp->nxttail[RCU_WAIT_TAIL]) - trace_rcu_grace_period(rsp->name, rdp->gpnum, "AccWaitCB"); + trace_rcu_grace_period(rsp->name, rdp->gpnum, TPS("AccWaitCB")); else - trace_rcu_grace_period(rsp->name, rdp->gpnum, "AccReadyCB"); + trace_rcu_grace_period(rsp->name, rdp->gpnum, TPS("AccReadyCB")); } /* @@ -1271,7 +1290,7 @@ static void __note_gp_changes(struct rcu_state *rsp, struct rcu_node *rnp, struc /* Remember that we saw this grace-period completion. */ rdp->completed = rnp->completed; - trace_rcu_grace_period(rsp->name, rdp->gpnum, "cpuend"); + trace_rcu_grace_period(rsp->name, rdp->gpnum, TPS("cpuend")); } if (rdp->gpnum != rnp->gpnum) { @@ -1281,7 +1300,7 @@ static void __note_gp_changes(struct rcu_state *rsp, struct rcu_node *rnp, struc * go looking for one. */ rdp->gpnum = rnp->gpnum; - trace_rcu_grace_period(rsp->name, rdp->gpnum, "cpustart"); + trace_rcu_grace_period(rsp->name, rdp->gpnum, TPS("cpustart")); rdp->passed_quiesce = 0; rdp->qs_pending = !!(rnp->qsmask & rdp->grpmask); zero_cpu_stall_ticks(rdp); @@ -1324,7 +1343,7 @@ static int rcu_gp_init(struct rcu_state *rsp) /* Advance to a new grace period and initialize state. */ rsp->gpnum++; - trace_rcu_grace_period(rsp->name, rsp->gpnum, "start"); + trace_rcu_grace_period(rsp->name, rsp->gpnum, TPS("start")); record_gp_stall_check_time(rsp); raw_spin_unlock_irq(&rnp->lock); @@ -1446,7 +1465,7 @@ static void rcu_gp_cleanup(struct rcu_state *rsp) rcu_nocb_gp_set(rnp, nocb); rsp->completed = rsp->gpnum; /* Declare grace period done. */ - trace_rcu_grace_period(rsp->name, rsp->completed, "end"); + trace_rcu_grace_period(rsp->name, rsp->completed, TPS("end")); rsp->fqs_state = RCU_GP_IDLE; rdp = this_cpu_ptr(rsp->rda); rcu_advance_cbs(rsp, rnp, rdp); /* Reduce false positives below. */ @@ -1855,7 +1874,7 @@ static void rcu_cleanup_dying_cpu(struct rcu_state *rsp) RCU_TRACE(mask = rdp->grpmask); trace_rcu_grace_period(rsp->name, rnp->gpnum + 1 - !!(rnp->qsmask & mask), - "cpuofl"); + TPS("cpuofl")); } /* @@ -2042,7 +2061,7 @@ static void rcu_do_batch(struct rcu_state *rsp, struct rcu_data *rdp) */ void rcu_check_callbacks(int cpu, int user) { - trace_rcu_utilization("Start scheduler-tick"); + trace_rcu_utilization(TPS("Start scheduler-tick")); increment_cpu_stall_ticks(); if (user || rcu_is_cpu_rrupt_from_idle()) { @@ -2075,7 +2094,7 @@ void rcu_check_callbacks(int cpu, int user) rcu_preempt_check_callbacks(cpu); if (rcu_pending(cpu)) invoke_rcu_core(); - trace_rcu_utilization("End scheduler-tick"); + trace_rcu_utilization(TPS("End scheduler-tick")); } /* @@ -2206,10 +2225,10 @@ static void rcu_process_callbacks(struct softirq_action *unused) if (cpu_is_offline(smp_processor_id())) return; - trace_rcu_utilization("Start RCU core"); + trace_rcu_utilization(TPS("Start RCU core")); for_each_rcu_flavor(rsp) __rcu_process_callbacks(rsp); - trace_rcu_utilization("End RCU core"); + trace_rcu_utilization(TPS("End RCU core")); } /* @@ -2950,7 +2969,7 @@ rcu_init_percpu_data(int cpu, struct rcu_state *rsp, int preemptible) rdp->completed = rnp->completed; rdp->passed_quiesce = 0; rdp->qs_pending = 0; - trace_rcu_grace_period(rsp->name, rdp->gpnum, "cpuonl"); + trace_rcu_grace_period(rsp->name, rdp->gpnum, TPS("cpuonl")); } raw_spin_unlock(&rnp->lock); /* irqs already disabled. */ rnp = rnp->parent; @@ -2980,7 +2999,7 @@ static int rcu_cpu_notify(struct notifier_block *self, struct rcu_node *rnp = rdp->mynode; struct rcu_state *rsp; - trace_rcu_utilization("Start CPU hotplug"); + trace_rcu_utilization(TPS("Start CPU hotplug")); switch (action) { case CPU_UP_PREPARE: case CPU_UP_PREPARE_FROZEN: @@ -3009,7 +3028,7 @@ static int rcu_cpu_notify(struct notifier_block *self, default: break; } - trace_rcu_utilization("End CPU hotplug"); + trace_rcu_utilization(TPS("End CPU hotplug")); return NOTIFY_OK; } diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 6976a7dde874..dff86f53ee09 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -167,7 +167,7 @@ static void rcu_preempt_qs(int cpu) struct rcu_data *rdp = &per_cpu(rcu_preempt_data, cpu); if (rdp->passed_quiesce == 0) - trace_rcu_grace_period("rcu_preempt", rdp->gpnum, "cpuqs"); + trace_rcu_grace_period(TPS("rcu_preempt"), rdp->gpnum, TPS("cpuqs")); rdp->passed_quiesce = 1; current->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_NEED_QS; } @@ -386,7 +386,7 @@ void rcu_read_unlock_special(struct task_struct *t) np = rcu_next_node_entry(t, rnp); list_del_init(&t->rcu_node_entry); t->rcu_blocked_node = NULL; - trace_rcu_unlock_preempted_task("rcu_preempt", + trace_rcu_unlock_preempted_task(TPS("rcu_preempt"), rnp->gpnum, t->pid); if (&t->rcu_node_entry == rnp->gp_tasks) rnp->gp_tasks = np; @@ -410,7 +410,7 @@ void rcu_read_unlock_special(struct task_struct *t) */ empty_exp_now = !rcu_preempted_readers_exp(rnp); if (!empty && !rcu_preempt_blocked_readers_cgp(rnp)) { - trace_rcu_quiescent_state_report("preempt_rcu", + trace_rcu_quiescent_state_report(TPS("preempt_rcu"), rnp->gpnum, 0, rnp->qsmask, rnp->level, @@ -1248,12 +1248,12 @@ static int rcu_boost_kthread(void *arg) int spincnt = 0; int more2boost; - trace_rcu_utilization("Start boost kthread@init"); + trace_rcu_utilization(TPS("Start boost kthread@init")); for (;;) { rnp->boost_kthread_status = RCU_KTHREAD_WAITING; - trace_rcu_utilization("End boost kthread@rcu_wait"); + trace_rcu_utilization(TPS("End boost kthread@rcu_wait")); rcu_wait(rnp->boost_tasks || rnp->exp_tasks); - trace_rcu_utilization("Start boost kthread@rcu_wait"); + trace_rcu_utilization(TPS("Start boost kthread@rcu_wait")); rnp->boost_kthread_status = RCU_KTHREAD_RUNNING; more2boost = rcu_boost(rnp); if (more2boost) @@ -1262,14 +1262,14 @@ static int rcu_boost_kthread(void *arg) spincnt = 0; if (spincnt > 10) { rnp->boost_kthread_status = RCU_KTHREAD_YIELDING; - trace_rcu_utilization("End boost kthread@rcu_yield"); + trace_rcu_utilization(TPS("End boost kthread@rcu_yield")); schedule_timeout_interruptible(2); - trace_rcu_utilization("Start boost kthread@rcu_yield"); + trace_rcu_utilization(TPS("Start boost kthread@rcu_yield")); spincnt = 0; } } /* NOTREACHED */ - trace_rcu_utilization("End boost kthread@notreached"); + trace_rcu_utilization(TPS("End boost kthread@notreached")); return 0; } @@ -1417,7 +1417,7 @@ static void rcu_cpu_kthread(unsigned int cpu) int spincnt; for (spincnt = 0; spincnt < 10; spincnt++) { - trace_rcu_utilization("Start CPU kthread@rcu_wait"); + trace_rcu_utilization(TPS("Start CPU kthread@rcu_wait")); local_bh_disable(); *statusp = RCU_KTHREAD_RUNNING; this_cpu_inc(rcu_cpu_kthread_loops); @@ -1429,15 +1429,15 @@ static void rcu_cpu_kthread(unsigned int cpu) rcu_kthread_do_work(); local_bh_enable(); if (*workp == 0) { - trace_rcu_utilization("End CPU kthread@rcu_wait"); + trace_rcu_utilization(TPS("End CPU kthread@rcu_wait")); *statusp = RCU_KTHREAD_WAITING; return; } } *statusp = RCU_KTHREAD_YIELDING; - trace_rcu_utilization("Start CPU kthread@rcu_yield"); + trace_rcu_utilization(TPS("Start CPU kthread@rcu_yield")); schedule_timeout_interruptible(2); - trace_rcu_utilization("End CPU kthread@rcu_yield"); + trace_rcu_utilization(TPS("End CPU kthread@rcu_yield")); *statusp = RCU_KTHREAD_WAITING; } @@ -2200,7 +2200,7 @@ static void rcu_nocb_wait_gp(struct rcu_data *rdp) * Wait for the grace period. Do so interruptibly to avoid messing * up the load average. */ - trace_rcu_future_gp(rnp, rdp, c, "StartWait"); + trace_rcu_future_gp(rnp, rdp, c, TPS("StartWait")); for (;;) { wait_event_interruptible( rnp->nocb_gp_wq[c & 0x1], @@ -2208,9 +2208,9 @@ static void rcu_nocb_wait_gp(struct rcu_data *rdp) if (likely(d)) break; flush_signals(current); - trace_rcu_future_gp(rnp, rdp, c, "ResumeWait"); + trace_rcu_future_gp(rnp, rdp, c, TPS("ResumeWait")); } - trace_rcu_future_gp(rnp, rdp, c, "EndWait"); + trace_rcu_future_gp(rnp, rdp, c, TPS("EndWait")); smp_mb(); /* Ensure that CB invocation happens after GP end. */ } -- cgit v1.2.3 From 1a11126bcb7c93c289bf3218fa546fd3b0c0df8b Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 26 Jul 2013 19:25:32 +0200 Subject: tracing: Turn event/id->i_private into call->event.type event_id_read() is racy, ftrace_event_call can be already freed by trace_remove_event_call() callers. Change event_create_dir() to pass "data = call->event.type", this is all event_id_read() needs. ftrace_event_id_fops no longer needs tracing_open_generic(). We add the new helper, event_file_data(), to read ->i_private, it will have more users. Note: currently ACCESS_ONCE() and "id != 0" check are not needed, but we are going to change event_remove/rmdir to clear ->i_private. Link: http://lkml.kernel.org/r/20130726172532.GA3605@redhat.com Reviewed-by: Masami Hiramatsu Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 18 +++++++++++++----- 1 file changed, 13 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 898f868833f2..c2d13c528c3c 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -409,6 +409,11 @@ static void put_system(struct ftrace_subsystem_dir *dir) mutex_unlock(&event_mutex); } +static void *event_file_data(struct file *filp) +{ + return ACCESS_ONCE(file_inode(filp)->i_private); +} + /* * Open and update trace_array ref count. * Must have the current trace_array passed to it. @@ -946,14 +951,18 @@ static int trace_format_open(struct inode *inode, struct file *file) static ssize_t event_id_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { - struct ftrace_event_call *call = filp->private_data; + int id = (long)event_file_data(filp); char buf[32]; int len; if (*ppos) return 0; - len = sprintf(buf, "%d\n", call->event.type); + if (unlikely(!id)) + return -ENODEV; + + len = sprintf(buf, "%d\n", id); + return simple_read_from_buffer(ubuf, cnt, ppos, buf, len); } @@ -1240,7 +1249,6 @@ static const struct file_operations ftrace_event_format_fops = { }; static const struct file_operations ftrace_event_id_fops = { - .open = tracing_open_generic, .read = event_id_read, .llseek = default_llseek, }; @@ -1488,8 +1496,8 @@ event_create_dir(struct dentry *parent, #ifdef CONFIG_PERF_EVENTS if (call->event.type && call->class->reg) - trace_create_file("id", 0444, file->dir, call, - id); + trace_create_file("id", 0444, file->dir, + (void *)(long)call->event.type, id); #endif /* -- cgit v1.2.3 From bc6f6b08dee5645770efb4b76186ded313f23752 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 26 Jul 2013 19:25:36 +0200 Subject: tracing: Change event_enable/disable_read() to verify i_private != NULL tracing_open_generic_file() is racy, ftrace_event_file can be already freed by rmdir or trace_remove_event_call(). Change event_enable_read() and event_disable_read() to read and verify "file = i_private" under event_mutex. This fixes nothing, but now we can change debugfs_remove("enable") callers to nullify ->i_private and fix the the problem. Link: http://lkml.kernel.org/r/20130726172536.GA3612@redhat.com Reviewed-by: Masami Hiramatsu Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 30 ++++++++++++++++++++---------- 1 file changed, 20 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index c2d13c528c3c..3dfa8419d0dc 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -684,15 +684,25 @@ static ssize_t event_enable_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { - struct ftrace_event_file *file = filp->private_data; + struct ftrace_event_file *file; + unsigned long flags; char buf[4] = "0"; - if (file->flags & FTRACE_EVENT_FL_ENABLED && - !(file->flags & FTRACE_EVENT_FL_SOFT_DISABLED)) + mutex_lock(&event_mutex); + file = event_file_data(filp); + if (likely(file)) + flags = file->flags; + mutex_unlock(&event_mutex); + + if (!file) + return -ENODEV; + + if (flags & FTRACE_EVENT_FL_ENABLED && + !(flags & FTRACE_EVENT_FL_SOFT_DISABLED)) strcpy(buf, "1"); - if (file->flags & FTRACE_EVENT_FL_SOFT_DISABLED || - file->flags & FTRACE_EVENT_FL_SOFT_MODE) + if (flags & FTRACE_EVENT_FL_SOFT_DISABLED || + flags & FTRACE_EVENT_FL_SOFT_MODE) strcat(buf, "*"); strcat(buf, "\n"); @@ -704,13 +714,10 @@ static ssize_t event_enable_write(struct file *filp, const char __user *ubuf, size_t cnt, loff_t *ppos) { - struct ftrace_event_file *file = filp->private_data; + struct ftrace_event_file *file; unsigned long val; int ret; - if (!file) - return -EINVAL; - ret = kstrtoul_from_user(ubuf, cnt, 10, &val); if (ret) return ret; @@ -722,8 +729,11 @@ event_enable_write(struct file *filp, const char __user *ubuf, size_t cnt, switch (val) { case 0: case 1: + ret = -ENODEV; mutex_lock(&event_mutex); - ret = ftrace_event_enable_disable(file, val); + file = event_file_data(filp); + if (likely(file)) + ret = ftrace_event_enable_disable(file, val); mutex_unlock(&event_mutex); break; -- cgit v1.2.3 From e2912b091c26b8ea95e5e00a43a7ac620f6c94a6 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 26 Jul 2013 19:25:40 +0200 Subject: tracing: Change event_filter_read/write to verify i_private != NULL event_filter_read/write() are racy, ftrace_event_call can be already freed by trace_remove_event_call() callers. 1. Shift mutex_lock(event_mutex) from print/apply_event_filter to the callers. 2. Change the callers, event_filter_read() and event_filter_write() to read i_private under this mutex and abort if it is NULL. This fixes nothing, but now we can change debugfs_remove("filter") callers to nullify ->i_private and fix the the problem. Link: http://lkml.kernel.org/r/20130726172540.GA3619@redhat.com Reviewed-by: Masami Hiramatsu Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 26 +++++++++++++++++++------- kernel/trace/trace_events_filter.c | 17 ++++++----------- 2 files changed, 25 insertions(+), 18 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 3dfa8419d0dc..1d7b6d03cd51 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -980,21 +980,28 @@ static ssize_t event_filter_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { - struct ftrace_event_call *call = filp->private_data; + struct ftrace_event_call *call; struct trace_seq *s; - int r; + int r = -ENODEV; if (*ppos) return 0; s = kmalloc(sizeof(*s), GFP_KERNEL); + if (!s) return -ENOMEM; trace_seq_init(s); - print_event_filter(call, s); - r = simple_read_from_buffer(ubuf, cnt, ppos, s->buffer, s->len); + mutex_lock(&event_mutex); + call = event_file_data(filp); + if (call) + print_event_filter(call, s); + mutex_unlock(&event_mutex); + + if (call) + r = simple_read_from_buffer(ubuf, cnt, ppos, s->buffer, s->len); kfree(s); @@ -1005,9 +1012,9 @@ static ssize_t event_filter_write(struct file *filp, const char __user *ubuf, size_t cnt, loff_t *ppos) { - struct ftrace_event_call *call = filp->private_data; + struct ftrace_event_call *call; char *buf; - int err; + int err = -ENODEV; if (cnt >= PAGE_SIZE) return -EINVAL; @@ -1022,7 +1029,12 @@ event_filter_write(struct file *filp, const char __user *ubuf, size_t cnt, } buf[cnt] = '\0'; - err = apply_event_filter(call, buf); + mutex_lock(&event_mutex); + call = event_file_data(filp); + if (call) + err = apply_event_filter(call, buf); + mutex_unlock(&event_mutex); + free_page((unsigned long) buf); if (err < 0) return err; diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c index 0c7b75a8acc8..97daa8cf958d 100644 --- a/kernel/trace/trace_events_filter.c +++ b/kernel/trace/trace_events_filter.c @@ -637,17 +637,15 @@ static void append_filter_err(struct filter_parse_state *ps, free_page((unsigned long) buf); } +/* caller must hold event_mutex */ void print_event_filter(struct ftrace_event_call *call, struct trace_seq *s) { - struct event_filter *filter; + struct event_filter *filter = call->filter; - mutex_lock(&event_mutex); - filter = call->filter; if (filter && filter->filter_string) trace_seq_printf(s, "%s\n", filter->filter_string); else trace_seq_puts(s, "none\n"); - mutex_unlock(&event_mutex); } void print_subsystem_event_filter(struct event_subsystem *system, @@ -1841,23 +1839,22 @@ static int create_system_filter(struct event_subsystem *system, return err; } +/* caller must hold event_mutex */ int apply_event_filter(struct ftrace_event_call *call, char *filter_string) { struct event_filter *filter; - int err = 0; - - mutex_lock(&event_mutex); + int err; if (!strcmp(strstrip(filter_string), "0")) { filter_disable(call); filter = call->filter; if (!filter) - goto out_unlock; + return 0; RCU_INIT_POINTER(call->filter, NULL); /* Make sure the filter is not being used */ synchronize_sched(); __free_filter(filter); - goto out_unlock; + return 0; } err = create_filter(call, filter_string, true, &filter); @@ -1884,8 +1881,6 @@ int apply_event_filter(struct ftrace_event_call *call, char *filter_string) __free_filter(tmp); } } -out_unlock: - mutex_unlock(&event_mutex); return err; } -- cgit v1.2.3 From c5a44a1200c6eda2202434f25325e8ad19533fca Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 26 Jul 2013 19:25:43 +0200 Subject: tracing: Change f_start() to take event_mutex and verify i_private != NULL trace_format_open() and trace_format_seq_ops are racy, nothing protects ftrace_event_call from trace_remove_event_call(). Change f_start() to take event_mutex and verify i_private != NULL, change f_stop() to drop this lock. This fixes nothing, but now we can change debugfs_remove("format") callers to nullify ->i_private and fix the the problem. Note: the usage of event_mutex is sub-optimal but simple, we can change this later. Link: http://lkml.kernel.org/r/20130726172543.GA3622@redhat.com Reviewed-by: Masami Hiramatsu Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 1d7b6d03cd51..50dc8b2e5435 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -840,7 +840,7 @@ enum { static void *f_next(struct seq_file *m, void *v, loff_t *pos) { - struct ftrace_event_call *call = m->private; + struct ftrace_event_call *call = event_file_data(m->private); struct list_head *common_head = &ftrace_common_fields; struct list_head *head = trace_get_fields(call); struct list_head *node = v; @@ -872,7 +872,7 @@ static void *f_next(struct seq_file *m, void *v, loff_t *pos) static int f_show(struct seq_file *m, void *v) { - struct ftrace_event_call *call = m->private; + struct ftrace_event_call *call = event_file_data(m->private); struct ftrace_event_field *field; const char *array_descriptor; @@ -925,6 +925,11 @@ static void *f_start(struct seq_file *m, loff_t *pos) void *p = (void *)FORMAT_HEADER; loff_t l = 0; + /* ->stop() is called even if ->start() fails */ + mutex_lock(&event_mutex); + if (!event_file_data(m->private)) + return ERR_PTR(-ENODEV); + while (l < *pos && p) p = f_next(m, p, &l); @@ -933,6 +938,7 @@ static void *f_start(struct seq_file *m, loff_t *pos) static void f_stop(struct seq_file *m, void *p) { + mutex_unlock(&event_mutex); } static const struct seq_operations trace_format_seq_ops = { @@ -944,7 +950,6 @@ static const struct seq_operations trace_format_seq_ops = { static int trace_format_open(struct inode *inode, struct file *file) { - struct ftrace_event_call *call = inode->i_private; struct seq_file *m; int ret; @@ -953,7 +958,7 @@ static int trace_format_open(struct inode *inode, struct file *file) return ret; m = file->private_data; - m->private = call; + m->private = file; return 0; } -- cgit v1.2.3 From f6a84bdc75b5c11621dec58db73fe102cbaf40cc Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 26 Jul 2013 19:25:47 +0200 Subject: tracing: Introduce remove_event_file_dir() Preparation for the next patch. Extract the common code from remove_event_from_tracers() and __trace_remove_event_dirs() into the new helper, remove_event_file_dir(). The patch looks more complicated than it actually is, it also moves remove_subsystem() up to avoid the forward declaration. Link: http://lkml.kernel.org/r/20130726172547.GA3629@redhat.com Reviewed-by: Masami Hiramatsu Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 47 ++++++++++++++++++++++----------------------- 1 file changed, 23 insertions(+), 24 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 50dc8b2e5435..05d647ecd01a 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -409,11 +409,31 @@ static void put_system(struct ftrace_subsystem_dir *dir) mutex_unlock(&event_mutex); } +static void remove_subsystem(struct ftrace_subsystem_dir *dir) +{ + if (!dir) + return; + + if (!--dir->nr_events) { + debugfs_remove_recursive(dir->entry); + list_del(&dir->list); + __put_system_dir(dir); + } +} + static void *event_file_data(struct file *filp) { return ACCESS_ONCE(file_inode(filp)->i_private); } +static void remove_event_file_dir(struct ftrace_event_file *file) +{ + list_del(&file->list); + debugfs_remove_recursive(file->dir); + remove_subsystem(file->system); + kmem_cache_free(file_cachep, file); +} + /* * Open and update trace_array ref count. * Must have the current trace_array passed to it. @@ -1549,33 +1569,16 @@ event_create_dir(struct dentry *parent, return 0; } -static void remove_subsystem(struct ftrace_subsystem_dir *dir) -{ - if (!dir) - return; - - if (!--dir->nr_events) { - debugfs_remove_recursive(dir->entry); - list_del(&dir->list); - __put_system_dir(dir); - } -} - static void remove_event_from_tracers(struct ftrace_event_call *call) { struct ftrace_event_file *file; struct trace_array *tr; do_for_each_event_file_safe(tr, file) { - if (file->event_call != call) continue; - list_del(&file->list); - debugfs_remove_recursive(file->dir); - remove_subsystem(file->system); - kmem_cache_free(file_cachep, file); - + remove_event_file_dir(file); /* * The do_for_each_event_file_safe() is * a double loop. After finding the call for this @@ -2305,12 +2308,8 @@ __trace_remove_event_dirs(struct trace_array *tr) { struct ftrace_event_file *file, *next; - list_for_each_entry_safe(file, next, &tr->events, list) { - list_del(&file->list); - debugfs_remove_recursive(file->dir); - remove_subsystem(file->system); - kmem_cache_free(file_cachep, file); - } + list_for_each_entry_safe(file, next, &tr->events, list) + remove_event_file_dir(file); } static void -- cgit v1.2.3 From bf682c3159c4d298d1126a56793ed3f5e80395f7 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Sun, 28 Jul 2013 20:35:27 +0200 Subject: tracing: Change remove_event_file_dir() to clear "d_subdirs"->i_private Change remove_event_file_dir() to clear ->i_private for every file we are going to remove. We need to check file->dir != NULL because event_create_dir() can fail. debugfs_remove_recursive(NULL) is fine but the patch moves it under the same check anyway for readability. spin_lock(d_lock) and "d_inode != NULL" check are not needed afaics, but I do not understand this code enough. tracing_open_generic_file() and tracing_release_generic_file() can go away, ftrace_enable_fops and ftrace_event_filter_fops() use tracing_open_generic() but only to check tracing_disabled. This fixes all races with event_remove() or instance_delete(). f_op->read/write/whatever can never use the freed file/call, all event/* files were changed to check and use ->i_private under event_mutex. Note: this doesn't not fix other problems, event_remove() can destroy the active ftrace_event_call, we need more changes but those changes are completely orthogonal. Link: http://lkml.kernel.org/r/20130728183527.GB16723@redhat.com Reviewed-by: Masami Hiramatsu Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 47 +++++++++++++++------------------------------ 1 file changed, 15 insertions(+), 32 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 05d647ecd01a..a67c913e2f9f 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -428,41 +428,25 @@ static void *event_file_data(struct file *filp) static void remove_event_file_dir(struct ftrace_event_file *file) { + struct dentry *dir = file->dir; + struct dentry *child; + + if (dir) { + spin_lock(&dir->d_lock); /* probably unneeded */ + list_for_each_entry(child, &dir->d_subdirs, d_u.d_child) { + if (child->d_inode) /* probably unneeded */ + child->d_inode->i_private = NULL; + } + spin_unlock(&dir->d_lock); + + debugfs_remove_recursive(dir); + } + list_del(&file->list); - debugfs_remove_recursive(file->dir); remove_subsystem(file->system); kmem_cache_free(file_cachep, file); } -/* - * Open and update trace_array ref count. - * Must have the current trace_array passed to it. - */ -static int tracing_open_generic_file(struct inode *inode, struct file *filp) -{ - struct ftrace_event_file *file = inode->i_private; - struct trace_array *tr = file->tr; - int ret; - - if (trace_array_get(tr) < 0) - return -ENODEV; - - ret = tracing_open_generic(inode, filp); - if (ret < 0) - trace_array_put(tr); - return ret; -} - -static int tracing_release_generic_file(struct inode *inode, struct file *filp) -{ - struct ftrace_event_file *file = inode->i_private; - struct trace_array *tr = file->tr; - - trace_array_put(tr); - - return 0; -} - /* * __ftrace_set_clr_event(NULL, NULL, NULL, set) will set/unset all events. */ @@ -1281,10 +1265,9 @@ static const struct file_operations ftrace_set_event_fops = { }; static const struct file_operations ftrace_enable_fops = { - .open = tracing_open_generic_file, + .open = tracing_open_generic, .read = event_enable_read, .write = event_enable_write, - .release = tracing_release_generic_file, .llseek = default_llseek, }; -- cgit v1.2.3 From 1c80c43290ee576afe8d39ecc905fa3958a5858c Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Thu, 25 Jul 2013 20:22:00 -0400 Subject: ftrace: Consolidate some duplicate code for updating ftrace ops When ftrace ops modifies the functions that it will trace, the update to the function mcount callers may need to be modified. Consolidate the two places that do the checks to see if an update is required with a wrapper function for those checks. Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 16 ++++++++++------ 1 file changed, 10 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 8ce9eefc5bb4..92d3334de0c3 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -3384,6 +3384,12 @@ ftrace_match_addr(struct ftrace_hash *hash, unsigned long ip, int remove) return add_hash_entry(hash, ip); } +static void ftrace_ops_update_code(struct ftrace_ops *ops) +{ + if (ops->flags & FTRACE_OPS_FL_ENABLED && ftrace_enabled) + ftrace_run_update_code(FTRACE_UPDATE_CALLS); +} + static int ftrace_set_hash(struct ftrace_ops *ops, unsigned char *buf, int len, unsigned long ip, int remove, int reset, int enable) @@ -3426,9 +3432,8 @@ ftrace_set_hash(struct ftrace_ops *ops, unsigned char *buf, int len, mutex_lock(&ftrace_lock); ret = ftrace_hash_move(ops, enable, orig_hash, hash); - if (!ret && ops->flags & FTRACE_OPS_FL_ENABLED - && ftrace_enabled) - ftrace_run_update_code(FTRACE_UPDATE_CALLS); + if (!ret) + ftrace_ops_update_code(ops); mutex_unlock(&ftrace_lock); @@ -3655,9 +3660,8 @@ int ftrace_regex_release(struct inode *inode, struct file *file) mutex_lock(&ftrace_lock); ret = ftrace_hash_move(iter->ops, filter_hash, orig_hash, iter->hash); - if (!ret && (iter->ops->flags & FTRACE_OPS_FL_ENABLED) - && ftrace_enabled) - ftrace_run_update_code(FTRACE_UPDATE_CALLS); + if (!ret) + ftrace_ops_update_code(iter->ops); mutex_unlock(&ftrace_lock); } -- cgit v1.2.3 From 2b44c4db2e2f1765d35163a861d301038e0c8a75 Mon Sep 17 00:00:00 2001 From: Colin Cross Date: Wed, 24 Jul 2013 17:41:33 -0700 Subject: freezer: set PF_SUSPEND_TASK flag on tasks that call freeze_processes Calling freeze_processes sets a global flag that will cause any process that calls try_to_freeze to enter the refrigerator. It skips sending a signal to the current task, but if the current task ever hits try_to_freeze, all threads will be frozen and the system will deadlock. Set a new flag, PF_SUSPEND_TASK, on the task that calls freeze_processes. The flag notifies the freezer that the thread is involved in suspend and should not be frozen. Also add a WARN_ON in thaw_processes if the caller does not have the PF_SUSPEND_TASK flag set to catch if a different task calls thaw_processes than the one that called freeze_processes, leaving a task with PF_SUSPEND_TASK permanently set on it. Threads that spawn off a task with PF_SUSPEND_TASK set (which swsusp does) will also have PF_SUSPEND_TASK set, preventing them from freezing while they are helping with suspend, but they need to be dead by the time suspend is triggered, otherwise they may run when userspace is expected to be frozen. Add a WARN_ON in thaw_processes if more than one thread has the PF_SUSPEND_TASK flag set. Reported-and-tested-by: Michael Leun Signed-off-by: Colin Cross Signed-off-by: Rafael J. Wysocki --- kernel/freezer.c | 2 +- kernel/power/process.c | 11 +++++++++++ 2 files changed, 12 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/freezer.c b/kernel/freezer.c index 8b2afc1c9df0..b462fa197517 100644 --- a/kernel/freezer.c +++ b/kernel/freezer.c @@ -33,7 +33,7 @@ static DEFINE_SPINLOCK(freezer_lock); */ bool freezing_slow_path(struct task_struct *p) { - if (p->flags & PF_NOFREEZE) + if (p->flags & (PF_NOFREEZE | PF_SUSPEND_TASK)) return false; if (pm_nosig_freezing || cgroup_freezing(p)) diff --git a/kernel/power/process.c b/kernel/power/process.c index fc0df8486449..06ec8869dbf1 100644 --- a/kernel/power/process.c +++ b/kernel/power/process.c @@ -109,6 +109,8 @@ static int try_to_freeze_tasks(bool user_only) /** * freeze_processes - Signal user space processes to enter the refrigerator. + * The current thread will not be frozen. The same process that calls + * freeze_processes must later call thaw_processes. * * On success, returns 0. On failure, -errno and system is fully thawed. */ @@ -120,6 +122,9 @@ int freeze_processes(void) if (error) return error; + /* Make sure this task doesn't get frozen */ + current->flags |= PF_SUSPEND_TASK; + if (!pm_freezing) atomic_inc(&system_freezing_cnt); @@ -168,6 +173,7 @@ int freeze_kernel_threads(void) void thaw_processes(void) { struct task_struct *g, *p; + struct task_struct *curr = current; if (pm_freezing) atomic_dec(&system_freezing_cnt); @@ -182,10 +188,15 @@ void thaw_processes(void) read_lock(&tasklist_lock); do_each_thread(g, p) { + /* No other threads should have PF_SUSPEND_TASK set */ + WARN_ON((p != curr) && (p->flags & PF_SUSPEND_TASK)); __thaw_task(p); } while_each_thread(g, p); read_unlock(&tasklist_lock); + WARN_ON(!(curr->flags & PF_SUSPEND_TASK)); + curr->flags &= ~PF_SUSPEND_TASK; + usermodehelper_enable(); schedule(); -- cgit v1.2.3 From db446a08c23d5475e6b08c87acca79ebb20f283c Mon Sep 17 00:00:00 2001 From: Benjamin LaHaise Date: Tue, 30 Jul 2013 12:54:40 -0400 Subject: aio: convert the ioctx list to table lookup v3 On Wed, Jun 12, 2013 at 11:14:40AM -0700, Kent Overstreet wrote: > On Mon, Apr 15, 2013 at 02:40:55PM +0300, Octavian Purdila wrote: > > When using a large number of threads performing AIO operations the > > IOCTX list may get a significant number of entries which will cause > > significant overhead. For example, when running this fio script: > > > > rw=randrw; size=256k ;directory=/mnt/fio; ioengine=libaio; iodepth=1 > > blocksize=1024; numjobs=512; thread; loops=100 > > > > on an EXT2 filesystem mounted on top of a ramdisk we can observe up to > > 30% CPU time spent by lookup_ioctx: > > > > 32.51% [guest.kernel] [g] lookup_ioctx > > 9.19% [guest.kernel] [g] __lock_acquire.isra.28 > > 4.40% [guest.kernel] [g] lock_release > > 4.19% [guest.kernel] [g] sched_clock_local > > 3.86% [guest.kernel] [g] local_clock > > 3.68% [guest.kernel] [g] native_sched_clock > > 3.08% [guest.kernel] [g] sched_clock_cpu > > 2.64% [guest.kernel] [g] lock_release_holdtime.part.11 > > 2.60% [guest.kernel] [g] memcpy > > 2.33% [guest.kernel] [g] lock_acquired > > 2.25% [guest.kernel] [g] lock_acquire > > 1.84% [guest.kernel] [g] do_io_submit > > > > This patchs converts the ioctx list to a radix tree. For a performance > > comparison the above FIO script was run on a 2 sockets 8 core > > machine. This are the results (average and %rsd of 10 runs) for the > > original list based implementation and for the radix tree based > > implementation: > > > > cores 1 2 4 8 16 32 > > list 109376 ms 69119 ms 35682 ms 22671 ms 19724 ms 16408 ms > > %rsd 0.69% 1.15% 1.17% 1.21% 1.71% 1.43% > > radix 73651 ms 41748 ms 23028 ms 16766 ms 15232 ms 13787 ms > > %rsd 1.19% 0.98% 0.69% 1.13% 0.72% 0.75% > > % of radix > > relative 66.12% 65.59% 66.63% 72.31% 77.26% 83.66% > > to list > > > > To consider the impact of the patch on the typical case of having > > only one ctx per process the following FIO script was run: > > > > rw=randrw; size=100m ;directory=/mnt/fio; ioengine=libaio; iodepth=1 > > blocksize=1024; numjobs=1; thread; loops=100 > > > > on the same system and the results are the following: > > > > list 58892 ms > > %rsd 0.91% > > radix 59404 ms > > %rsd 0.81% > > % of radix > > relative 100.87% > > to list > > So, I was just doing some benchmarking/profiling to get ready to send > out the aio patches I've got for 3.11 - and it looks like your patch is > causing a ~1.5% throughput regression in my testing :/ ... I've got an alternate approach for fixing this wart in lookup_ioctx()... Instead of using an rbtree, just use the reserved id in the ring buffer header to index an array pointing the ioctx. It's not finished yet, and it needs to be tidied up, but is most of the way there. -ben -- "Thought is the essence of where you are now." -- kmo> And, a rework of Ben's code, but this was entirely his idea kmo> -Kent bcrl> And fix the code to use the right mm_struct in kill_ioctx(), actually free memory. Signed-off-by: Benjamin LaHaise --- kernel/fork.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index 66635c80a813..db5f541c5488 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -522,7 +522,7 @@ static void mm_init_aio(struct mm_struct *mm) { #ifdef CONFIG_AIO spin_lock_init(&mm->ioctx_lock); - INIT_HLIST_HEAD(&mm->ioctx_list); + mm->ioctx_table = NULL; #endif } -- cgit v1.2.3 From bf0bd948d1682e3996adc093b43021ed391983e6 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 26 Jul 2013 23:48:42 +0200 Subject: sched: Ensure update_cfs_shares() is called for parents of continuously-running tasks We typically update a task_group's shares within the dequeue/enqueue path. However, continuously running tasks sharing a CPU are not subject to these updates as they are only put/picked. Unfortunately, when we reverted f269ae046 (in 17bc14b7), we lost the augmenting periodic update that was supposed to account for this; resulting in a potential loss of fairness. To fix this, re-introduce the explicit update in update_cfs_rq_blocked_load() [called via entity_tick()]. Reported-by: Max Hailperin Signed-off-by: Peter Zijlstra Reviewed-by: Paul Turner Link: http://lkml.kernel.org/n/tip-9545m3apw5d93ubyrotrj31y@git.kernel.org Cc: Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 98d135584b4b..06db94bf47a8 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -2032,6 +2032,7 @@ entity_tick(struct cfs_rq *cfs_rq, struct sched_entity *curr, int queued) */ update_entity_load_avg(curr, 1); update_cfs_rq_blocked_load(cfs_rq, 1); + update_cfs_shares(cfs_rq); #ifdef CONFIG_SCHED_HRTICK /* -- cgit v1.2.3 From 85f4896123d0299128f2c95cc40f3b8b01d4b0f6 Mon Sep 17 00:00:00 2001 From: Maarten Lankhorst Date: Tue, 30 Jul 2013 10:13:41 +0200 Subject: mutex: Fix w/w mutex deadlock injection The check needs to be for > 1, because ctx->acquired is already incremented. This will prevent ww_mutex_lock_slow from returning -EDEADLK and not locking the mutex. It caused a lot of false gpu lockups on radeon with CONFIG_DEBUG_WW_MUTEX_SLOWPATH=y because a function that shouldn't be able to return -EDEADLK did. Signed-off-by: Maarten Lankhorst Signed-off-by: Peter Zijlstra Cc: Alex Deucher Cc: Linus Torvalds Cc: Andrew Morton Cc: Thomas Gleixner Link: http://lkml.kernel.org/r/51F775B5.201@canonical.com Signed-off-by: Ingo Molnar --- kernel/mutex.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/mutex.c b/kernel/mutex.c index ff05f4bd86eb..a52ee7bb830d 100644 --- a/kernel/mutex.c +++ b/kernel/mutex.c @@ -686,7 +686,7 @@ __ww_mutex_lock(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) might_sleep(); ret = __mutex_lock_common(&lock->base, TASK_UNINTERRUPTIBLE, 0, &ctx->dep_map, _RET_IP_, ctx); - if (!ret && ctx->acquired > 0) + if (!ret && ctx->acquired > 1) return ww_mutex_deadlock_injection(lock, ctx); return ret; @@ -702,7 +702,7 @@ __ww_mutex_lock_interruptible(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) ret = __mutex_lock_common(&lock->base, TASK_INTERRUPTIBLE, 0, &ctx->dep_map, _RET_IP_, ctx); - if (!ret && ctx->acquired > 0) + if (!ret && ctx->acquired > 1) return ww_mutex_deadlock_injection(lock, ctx); return ret; -- cgit v1.2.3 From 46591962cb5bfd2bfb0baf42497119c816503598 Mon Sep 17 00:00:00 2001 From: Xie XiuQi Date: Tue, 30 Jul 2013 11:06:09 +0800 Subject: generic-ipi: Kill unnecessary variable - csd_flags After commit 8969a5ede0f9e17da4b943712429aef2c9bcd82b ("generic-ipi: remove kmalloc()"), wait = 0 can be guaranteed, and all callsites of generic_exec_single() do an unconditional csd_lock() now. So csd_flags is unnecessary now. Remove it. Signed-off-by: Xie XiuQi Signed-off-by: Peter Zijlstra Cc: Oleg Nesterov Cc: Linus Torvalds Cc: Nick Piggin Cc: Jens Axboe Cc: "Paul E. McKenney" Cc: Rusty Russell Link: http://lkml.kernel.org/r/51F72DA1.7010401@huawei.com Signed-off-by: Ingo Molnar --- kernel/smp.c | 14 +------------- 1 file changed, 1 insertion(+), 13 deletions(-) (limited to 'kernel') diff --git a/kernel/smp.c b/kernel/smp.c index fe9f773d7114..7332697cd184 100644 --- a/kernel/smp.c +++ b/kernel/smp.c @@ -186,25 +186,13 @@ void generic_smp_call_function_single_interrupt(void) while (!list_empty(&list)) { struct call_single_data *csd; - unsigned int csd_flags; csd = list_entry(list.next, struct call_single_data, list); list_del(&csd->list); - /* - * 'csd' can be invalid after this call if flags == 0 - * (when called through generic_exec_single()), - * so save them away before making the call: - */ - csd_flags = csd->flags; - csd->func(csd->info); - /* - * Unlocked CSDs are valid through generic_exec_single(): - */ - if (csd_flags & CSD_FLAG_LOCK) - csd_unlock(csd); + csd_unlock(csd); } } -- cgit v1.2.3 From 6050cb0b0b366092d1383bc23d7b16cd26db00f0 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 23 Jul 2013 02:30:59 +0200 Subject: perf: Fix branch stack refcount leak on callchain init failure On callchain buffers allocation failure, free_event() is called and all the accounting performed in perf_event_alloc() for that event is cancelled. But if the event has branch stack sampling, it is unaccounted as well from the branch stack sampling events refcounts. This is a bug because this accounting is performed after the callchain buffer allocation. As a result, the branch stack sampling events refcount can become negative. To fix this, move the branch stack event accounting before the callchain buffer allocation. Reported-by: Peter Zijlstra Signed-off-by: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Arnaldo Carvalho de Melo Cc: Stephane Eranian Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1374539466-4799-2-git-send-email-fweisbec@gmail.com Signed-off-by: Ingo Molnar --- kernel/events/core.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 127411400116..f35aa7e69e2d 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -6567,6 +6567,12 @@ done: atomic_inc(&nr_comm_events); if (event->attr.task) atomic_inc(&nr_task_events); + if (has_branch_stack(event)) { + static_key_slow_inc(&perf_sched_events.key); + if (!(event->attach_state & PERF_ATTACH_TASK)) + atomic_inc(&per_cpu(perf_branch_stack_events, + event->cpu)); + } if (event->attr.sample_type & PERF_SAMPLE_CALLCHAIN) { err = get_callchain_buffers(); if (err) { @@ -6574,12 +6580,6 @@ done: return ERR_PTR(err); } } - if (has_branch_stack(event)) { - static_key_slow_inc(&perf_sched_events.key); - if (!(event->attach_state & PERF_ATTACH_TASK)) - atomic_inc(&per_cpu(perf_branch_stack_events, - event->cpu)); - } } return event; -- cgit v1.2.3 From 90983b16078ab0fdc58f0dab3e8e3da79c9579a2 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 23 Jul 2013 02:31:00 +0200 Subject: perf: Sanitize get_callchain_buffer() In case of allocation failure, get_callchain_buffer() keeps the refcount incremented for the current event. As a result, when get_callchain_buffers() returns an error, we must cleanup what it did by cancelling its last refcount with a call to put_callchain_buffers(). This is a hack in order to be able to call free_event() after that failure. The original purpose of that was to simplify the failure path. But this error handling is actually counter intuitive, ugly and not very easy to follow because one expect to see the resources used to perform a service to be cleaned by the callee if case of failure, not by the caller. So lets clean this up by cancelling the refcount from get_callchain_buffer() in case of failure. And correctly free the event accordingly in perf_event_alloc(). Signed-off-by: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Arnaldo Carvalho de Melo Cc: Stephane Eranian Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1374539466-4799-3-git-send-email-fweisbec@gmail.com Signed-off-by: Ingo Molnar --- kernel/events/callchain.c | 2 ++ kernel/events/core.c | 41 +++++++++++++++++++++-------------------- 2 files changed, 23 insertions(+), 20 deletions(-) (limited to 'kernel') diff --git a/kernel/events/callchain.c b/kernel/events/callchain.c index c77206184b8b..76a8bc5f6265 100644 --- a/kernel/events/callchain.c +++ b/kernel/events/callchain.c @@ -117,6 +117,8 @@ int get_callchain_buffers(void) err = alloc_callchain_buffers(); exit: mutex_unlock(&callchain_mutex); + if (err) + atomic_dec(&nr_callchain_events); return err; } diff --git a/kernel/events/core.c b/kernel/events/core.c index f35aa7e69e2d..3b998626b7a0 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -6457,7 +6457,7 @@ perf_event_alloc(struct perf_event_attr *attr, int cpu, struct pmu *pmu; struct perf_event *event; struct hw_perf_event *hwc; - long err; + long err = -EINVAL; if ((unsigned)cpu >= nr_cpu_ids) { if (!task || cpu != -1) @@ -6540,25 +6540,23 @@ perf_event_alloc(struct perf_event_attr *attr, int cpu, * we currently do not support PERF_FORMAT_GROUP on inherited events */ if (attr->inherit && (attr->read_format & PERF_FORMAT_GROUP)) - goto done; + goto err_ns; pmu = perf_init_event(event); - -done: - err = 0; if (!pmu) - err = -EINVAL; - else if (IS_ERR(pmu)) + goto err_ns; + else if (IS_ERR(pmu)) { err = PTR_ERR(pmu); - - if (err) { - if (event->ns) - put_pid_ns(event->ns); - kfree(event); - return ERR_PTR(err); + goto err_ns; } if (!event->parent) { + if (event->attr.sample_type & PERF_SAMPLE_CALLCHAIN) { + err = get_callchain_buffers(); + if (err) + goto err_pmu; + } + if (event->attach_state & PERF_ATTACH_TASK) static_key_slow_inc(&perf_sched_events.key); if (event->attr.mmap || event->attr.mmap_data) @@ -6573,16 +6571,19 @@ done: atomic_inc(&per_cpu(perf_branch_stack_events, event->cpu)); } - if (event->attr.sample_type & PERF_SAMPLE_CALLCHAIN) { - err = get_callchain_buffers(); - if (err) { - free_event(event); - return ERR_PTR(err); - } - } } return event; + +err_pmu: + if (event->destroy) + event->destroy(event); +err_ns: + if (event->ns) + put_pid_ns(event->ns); + kfree(event); + + return ERR_PTR(err); } static int perf_copy_attr(struct perf_event_attr __user *uattr, -- cgit v1.2.3 From 766d6c076928191d75ad5b0d0f58f52b1e7682d8 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 23 Jul 2013 02:31:01 +0200 Subject: perf: Factor out event accounting code to account_event()/__free_event() Gather all the event accounting code to a single place, once all the prerequisites are completed. This simplifies the refcounting. Original-patch-by: Peter Zijlstra Signed-off-by: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Arnaldo Carvalho de Melo Cc: Stephane Eranian Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1374539466-4799-4-git-send-email-fweisbec@gmail.com Signed-off-by: Ingo Molnar --- kernel/events/core.c | 79 +++++++++++++++++++++++++++++++--------------------- 1 file changed, 47 insertions(+), 32 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 3b998626b7a0..158fd5789e58 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -3128,6 +3128,21 @@ static void free_event_rcu(struct rcu_head *head) static void ring_buffer_put(struct ring_buffer *rb); static void ring_buffer_detach(struct perf_event *event, struct ring_buffer *rb); +static void __free_event(struct perf_event *event) +{ + if (!event->parent) { + if (event->attr.sample_type & PERF_SAMPLE_CALLCHAIN) + put_callchain_buffers(); + } + + if (event->destroy) + event->destroy(event); + + if (event->ctx) + put_ctx(event->ctx); + + call_rcu(&event->rcu_head, free_event_rcu); +} static void free_event(struct perf_event *event) { irq_work_sync(&event->pending); @@ -3141,8 +3156,6 @@ static void free_event(struct perf_event *event) atomic_dec(&nr_comm_events); if (event->attr.task) atomic_dec(&nr_task_events); - if (event->attr.sample_type & PERF_SAMPLE_CALLCHAIN) - put_callchain_buffers(); if (is_cgroup_event(event)) { atomic_dec(&per_cpu(perf_cgroup_events, event->cpu)); static_key_slow_dec_deferred(&perf_sched_events); @@ -3180,13 +3193,8 @@ static void free_event(struct perf_event *event) if (is_cgroup_event(event)) perf_detach_cgroup(event); - if (event->destroy) - event->destroy(event); - - if (event->ctx) - put_ctx(event->ctx); - call_rcu(&event->rcu_head, free_event_rcu); + __free_event(event); } int perf_event_release_kernel(struct perf_event *event) @@ -6443,6 +6451,29 @@ unlock: return pmu; } +static void account_event(struct perf_event *event) +{ + if (event->attach_state & PERF_ATTACH_TASK) + static_key_slow_inc(&perf_sched_events.key); + if (event->attr.mmap || event->attr.mmap_data) + atomic_inc(&nr_mmap_events); + if (event->attr.comm) + atomic_inc(&nr_comm_events); + if (event->attr.task) + atomic_inc(&nr_task_events); + if (has_branch_stack(event)) { + static_key_slow_inc(&perf_sched_events.key); + if (!(event->attach_state & PERF_ATTACH_TASK)) + atomic_inc(&per_cpu(perf_branch_stack_events, + event->cpu)); + } + + if (is_cgroup_event(event)) { + atomic_inc(&per_cpu(perf_cgroup_events, event->cpu)); + static_key_slow_inc(&perf_sched_events.key); + } +} + /* * Allocate and initialize a event structure */ @@ -6556,21 +6587,6 @@ perf_event_alloc(struct perf_event_attr *attr, int cpu, if (err) goto err_pmu; } - - if (event->attach_state & PERF_ATTACH_TASK) - static_key_slow_inc(&perf_sched_events.key); - if (event->attr.mmap || event->attr.mmap_data) - atomic_inc(&nr_mmap_events); - if (event->attr.comm) - atomic_inc(&nr_comm_events); - if (event->attr.task) - atomic_inc(&nr_task_events); - if (has_branch_stack(event)) { - static_key_slow_inc(&perf_sched_events.key); - if (!(event->attach_state & PERF_ATTACH_TASK)) - atomic_inc(&per_cpu(perf_branch_stack_events, - event->cpu)); - } } return event; @@ -6865,17 +6881,14 @@ SYSCALL_DEFINE5(perf_event_open, if (flags & PERF_FLAG_PID_CGROUP) { err = perf_cgroup_connect(pid, event, &attr, group_leader); - if (err) - goto err_alloc; - /* - * one more event: - * - that has cgroup constraint on event->cpu - * - that may need work on context switch - */ - atomic_inc(&per_cpu(perf_cgroup_events, event->cpu)); - static_key_slow_inc(&perf_sched_events.key); + if (err) { + __free_event(event); + goto err_task; + } } + account_event(event); + /* * Special case software events and allow them to be part of * any hardware group. @@ -7071,6 +7084,8 @@ perf_event_create_kernel_counter(struct perf_event_attr *attr, int cpu, goto err; } + account_event(event); + ctx = find_get_context(event->pmu, task, cpu); if (IS_ERR(ctx)) { err = PTR_ERR(ctx); -- cgit v1.2.3 From 4beb31f3657348a8b702dd014d01c520e522012f Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 23 Jul 2013 02:31:02 +0200 Subject: perf: Split the per-cpu accounting part of the event accounting code This way we can use the per-cpu handling seperately. This is going to be used by to fix the event migration code accounting. Original-patch-by: Peter Zijlstra Signed-off-by: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Arnaldo Carvalho de Melo Cc: Stephane Eranian Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1374539466-4799-5-git-send-email-fweisbec@gmail.com Signed-off-by: Ingo Molnar --- kernel/events/core.c | 87 +++++++++++++++++++++++++++++++++------------------- 1 file changed, 55 insertions(+), 32 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 158fd5789e58..3a4b73aebc42 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -3128,6 +3128,40 @@ static void free_event_rcu(struct rcu_head *head) static void ring_buffer_put(struct ring_buffer *rb); static void ring_buffer_detach(struct perf_event *event, struct ring_buffer *rb); +static void unaccount_event_cpu(struct perf_event *event, int cpu) +{ + if (event->parent) + return; + + if (has_branch_stack(event)) { + if (!(event->attach_state & PERF_ATTACH_TASK)) + atomic_dec(&per_cpu(perf_branch_stack_events, cpu)); + } + if (is_cgroup_event(event)) + atomic_dec(&per_cpu(perf_cgroup_events, cpu)); +} + +static void unaccount_event(struct perf_event *event) +{ + if (event->parent) + return; + + if (event->attach_state & PERF_ATTACH_TASK) + static_key_slow_dec_deferred(&perf_sched_events); + if (event->attr.mmap || event->attr.mmap_data) + atomic_dec(&nr_mmap_events); + if (event->attr.comm) + atomic_dec(&nr_comm_events); + if (event->attr.task) + atomic_dec(&nr_task_events); + if (is_cgroup_event(event)) + static_key_slow_dec_deferred(&perf_sched_events); + if (has_branch_stack(event)) + static_key_slow_dec_deferred(&perf_sched_events); + + unaccount_event_cpu(event, event->cpu); +} + static void __free_event(struct perf_event *event) { if (!event->parent) { @@ -3147,29 +3181,7 @@ static void free_event(struct perf_event *event) { irq_work_sync(&event->pending); - if (!event->parent) { - if (event->attach_state & PERF_ATTACH_TASK) - static_key_slow_dec_deferred(&perf_sched_events); - if (event->attr.mmap || event->attr.mmap_data) - atomic_dec(&nr_mmap_events); - if (event->attr.comm) - atomic_dec(&nr_comm_events); - if (event->attr.task) - atomic_dec(&nr_task_events); - if (is_cgroup_event(event)) { - atomic_dec(&per_cpu(perf_cgroup_events, event->cpu)); - static_key_slow_dec_deferred(&perf_sched_events); - } - - if (has_branch_stack(event)) { - static_key_slow_dec_deferred(&perf_sched_events); - /* is system-wide event */ - if (!(event->attach_state & PERF_ATTACH_TASK)) { - atomic_dec(&per_cpu(perf_branch_stack_events, - event->cpu)); - } - } - } + unaccount_event(event); if (event->rb) { struct ring_buffer *rb; @@ -6451,8 +6463,24 @@ unlock: return pmu; } +static void account_event_cpu(struct perf_event *event, int cpu) +{ + if (event->parent) + return; + + if (has_branch_stack(event)) { + if (!(event->attach_state & PERF_ATTACH_TASK)) + atomic_inc(&per_cpu(perf_branch_stack_events, cpu)); + } + if (is_cgroup_event(event)) + atomic_inc(&per_cpu(perf_cgroup_events, cpu)); +} + static void account_event(struct perf_event *event) { + if (event->parent) + return; + if (event->attach_state & PERF_ATTACH_TASK) static_key_slow_inc(&perf_sched_events.key); if (event->attr.mmap || event->attr.mmap_data) @@ -6461,17 +6489,12 @@ static void account_event(struct perf_event *event) atomic_inc(&nr_comm_events); if (event->attr.task) atomic_inc(&nr_task_events); - if (has_branch_stack(event)) { + if (has_branch_stack(event)) static_key_slow_inc(&perf_sched_events.key); - if (!(event->attach_state & PERF_ATTACH_TASK)) - atomic_inc(&per_cpu(perf_branch_stack_events, - event->cpu)); - } - - if (is_cgroup_event(event)) { - atomic_inc(&per_cpu(perf_cgroup_events, event->cpu)); + if (is_cgroup_event(event)) static_key_slow_inc(&perf_sched_events.key); - } + + account_event_cpu(event, event->cpu); } /* -- cgit v1.2.3 From 9a545de019b536771feefb76f85e5038b65c2190 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 23 Jul 2013 02:31:03 +0200 Subject: perf: Migrate per cpu event accounting When an event is migrated, move the event per-cpu accounting accordingly so that branch stack and cgroup events work correctly on the new CPU. Original-patch-by: Peter Zijlstra Signed-off-by: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Arnaldo Carvalho de Melo Cc: Stephane Eranian Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1374539466-4799-6-git-send-email-fweisbec@gmail.com Signed-off-by: Ingo Molnar --- kernel/events/core.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 3a4b73aebc42..63bdec9fdd21 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -7145,6 +7145,7 @@ void perf_pmu_migrate_context(struct pmu *pmu, int src_cpu, int dst_cpu) list_for_each_entry_safe(event, tmp, &src_ctx->event_list, event_entry) { perf_remove_from_context(event); + unaccount_event_cpu(event, src_cpu); put_ctx(src_ctx); list_add(&event->event_entry, &events); } @@ -7157,6 +7158,7 @@ void perf_pmu_migrate_context(struct pmu *pmu, int src_cpu, int dst_cpu) list_del(&event->event_entry); if (event->state >= PERF_EVENT_STATE_OFF) event->state = PERF_EVENT_STATE_INACTIVE; + account_event_cpu(event, dst_cpu); perf_install_in_context(dst_ctx, event, dst_cpu); get_ctx(dst_ctx); } -- cgit v1.2.3 From ba8a75c16e292c0a3a87406a77508cbbc6cf4ee2 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 23 Jul 2013 02:31:04 +0200 Subject: perf: Account freq events per cpu This is going to be used by the full dynticks subsystem as a finer-grained information to know when to keep and when to stop the tick. Original-patch-by: Peter Zijlstra Signed-off-by: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Arnaldo Carvalho de Melo Cc: Stephane Eranian Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1374539466-4799-7-git-send-email-fweisbec@gmail.com Signed-off-by: Ingo Molnar --- kernel/events/core.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 63bdec9fdd21..3fe385aa93e6 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -141,6 +141,7 @@ enum event_type_t { struct static_key_deferred perf_sched_events __read_mostly; static DEFINE_PER_CPU(atomic_t, perf_cgroup_events); static DEFINE_PER_CPU(atomic_t, perf_branch_stack_events); +static DEFINE_PER_CPU(atomic_t, perf_freq_events); static atomic_t nr_mmap_events __read_mostly; static atomic_t nr_comm_events __read_mostly; @@ -3139,6 +3140,9 @@ static void unaccount_event_cpu(struct perf_event *event, int cpu) } if (is_cgroup_event(event)) atomic_dec(&per_cpu(perf_cgroup_events, cpu)); + + if (event->attr.freq) + atomic_dec(&per_cpu(perf_freq_events, cpu)); } static void unaccount_event(struct perf_event *event) @@ -6474,6 +6478,9 @@ static void account_event_cpu(struct perf_event *event, int cpu) } if (is_cgroup_event(event)) atomic_inc(&per_cpu(perf_cgroup_events, cpu)); + + if (event->attr.freq) + atomic_inc(&per_cpu(perf_freq_events, cpu)); } static void account_event(struct perf_event *event) -- cgit v1.2.3 From d84153d6c96f61aa06429586284639f32debf03e Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 23 Jul 2013 02:31:05 +0200 Subject: perf: Implement finer grained full dynticks kick Currently the full dynticks subsystem keep the tick alive as long as there are perf events running. This prevents the tick from being stopped as long as features such that the lockup detectors are running. As a temporary fix, the lockup detector is disabled by default when full dynticks is built but this is not a long term viable solution. To fix this, only keep the tick alive when an event configured with a frequency rather than a period is running on the CPU, or when an event throttles on the CPU. These are the only purposes of the perf tick, especially now that the rotation of flexible events is handled from a seperate hrtimer. The tick can be shutdown the rest of the time. Original-patch-by: Peter Zijlstra Signed-off-by: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Arnaldo Carvalho de Melo Cc: Stephane Eranian Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1374539466-4799-8-git-send-email-fweisbec@gmail.com Signed-off-by: Ingo Molnar --- kernel/events/core.c | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 3fe385aa93e6..916cf1f593b4 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -870,12 +870,8 @@ static void perf_pmu_rotate_start(struct pmu *pmu) WARN_ON(!irqs_disabled()); - if (list_empty(&cpuctx->rotation_list)) { - int was_empty = list_empty(head); + if (list_empty(&cpuctx->rotation_list)) list_add(&cpuctx->rotation_list, head); - if (was_empty) - tick_nohz_full_kick(); - } } static void get_ctx(struct perf_event_context *ctx) @@ -1875,6 +1871,9 @@ static int __perf_install_in_context(void *info) perf_pmu_enable(cpuctx->ctx.pmu); perf_ctx_unlock(cpuctx, task_ctx); + if (atomic_read(&__get_cpu_var(perf_freq_events))) + tick_nohz_full_kick(); + return 0; } @@ -2812,10 +2811,11 @@ done: #ifdef CONFIG_NO_HZ_FULL bool perf_event_can_stop_tick(void) { - if (list_empty(&__get_cpu_var(rotation_list))) - return true; - else + if (atomic_read(&__get_cpu_var(perf_freq_events)) || + __this_cpu_read(perf_throttled_count)) return false; + else + return true; } #endif @@ -5202,6 +5202,7 @@ static int __perf_event_overflow(struct perf_event *event, __this_cpu_inc(perf_throttled_count); hwc->interrupts = MAX_INTERRUPTS; perf_log_throttle(event, 0); + tick_nohz_full_kick(); ret = 1; } } -- cgit v1.2.3 From 93786a5f6aeb9c032c1c240246c5aabcf457b38f Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 23 Jul 2013 02:31:06 +0200 Subject: watchdog: Make it work under full dynticks A perf event can be used without forcing the tick to stay alive if it doesn't use a frequency but a sample period and if it doesn't throttle (raise storm of events). Since the lockup detector neither use a perf event frequency nor should ever throttle due to its high period, it can now run concurrently with the full dynticks feature. So remove the hack that disabled the watchdog. Signed-off-by: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Arnaldo Carvalho de Melo Cc: Stephane Eranian Cc: Don Zickus Cc: Srivatsa S. Bhat Cc: Anish Singh Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1374539466-4799-9-git-send-email-fweisbec@gmail.com Signed-off-by: Ingo Molnar --- kernel/watchdog.c | 8 -------- 1 file changed, 8 deletions(-) (limited to 'kernel') diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 1241d8c91d5e..51c4f34d258e 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -553,14 +553,6 @@ void __init lockup_detector_init(void) { set_sample_period(); -#ifdef CONFIG_NO_HZ_FULL - if (watchdog_user_enabled) { - watchdog_user_enabled = 0; - pr_warning("Disabled lockup detectors by default for full dynticks\n"); - pr_warning("You can reactivate it with 'sysctl -w kernel.watchdog=1'\n"); - } -#endif - if (watchdog_user_enabled) watchdog_enable_all_cpus(); } -- cgit v1.2.3 From 8c4f3c3fa9681dc549cd35419b259496082fef8b Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Tue, 30 Jul 2013 00:04:32 -0400 Subject: ftrace: Check module functions being traced on reload MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit There's been a nasty bug that would show up and not give much info. The bug displayed the following warning: WARNING: at kernel/trace/ftrace.c:1529 __ftrace_hash_rec_update+0x1e3/0x230() Pid: 20903, comm: bash Tainted: G O 3.6.11+ #38405.trunk Call Trace: [] warn_slowpath_common+0x7f/0xc0 [] warn_slowpath_null+0x1a/0x20 [] __ftrace_hash_rec_update+0x1e3/0x230 [] ftrace_hash_move+0x28/0x1d0 [] ? kfree+0x2c/0x110 [] ftrace_regex_release+0x8e/0x150 [] __fput+0xae/0x220 [] ____fput+0xe/0x10 [] task_work_run+0x72/0x90 [] do_notify_resume+0x6c/0xc0 [] ? trace_hardirqs_on_thunk+0x3a/0x3c [] int_signal+0x12/0x17 ---[ end trace 793179526ee09b2c ]--- It was finally narrowed down to unloading a module that was being traced. It was actually more than that. When functions are being traced, there's a table of all functions that have a ref count of the number of active tracers attached to that function. When a function trace callback is registered to a function, the function's record ref count is incremented. When it is unregistered, the function's record ref count is decremented. If an inconsistency is detected (ref count goes below zero) the above warning is shown and the function tracing is permanently disabled until reboot. The ftrace callback ops holds a hash of functions that it filters on (and/or filters off). If the hash is empty, the default means to filter all functions (for the filter_hash) or to disable no functions (for the notrace_hash). When a module is unloaded, it frees the function records that represent the module functions. These records exist on their own pages, that is function records for one module will not exist on the same page as function records for other modules or even the core kernel. Now when a module unloads, the records that represents its functions are freed. When the module is loaded again, the records are recreated with a default ref count of zero (unless there's a callback that traces all functions, then they will also be traced, and the ref count will be incremented). The problem is that if an ftrace callback hash includes functions of the module being unloaded, those hash entries will not be removed. If the module is reloaded in the same location, the hash entries still point to the functions of the module but the module's ref counts do not reflect that. With the help of Steve and Joern, we found a reproducer: Using uinput module and uinput_release function. cd /sys/kernel/debug/tracing modprobe uinput echo uinput_release > set_ftrace_filter echo function > current_tracer rmmod uinput modprobe uinput # check /proc/modules to see if loaded in same addr, otherwise try again echo nop > current_tracer [BOOM] The above loads the uinput module, which creates a table of functions that can be traced within the module. We add uinput_release to the filter_hash to trace just that function. Enable function tracincg, which increments the ref count of the record associated to uinput_release. Remove uinput, which frees the records including the one that represents uinput_release. Load the uinput module again (and make sure it's at the same address). This recreates the function records all with a ref count of zero, including uinput_release. Disable function tracing, which will decrement the ref count for uinput_release which is now zero because of the module removal and reload, and we have a mismatch (below zero ref count). The solution is to check all currently tracing ftrace callbacks to see if any are tracing any of the module's functions when a module is loaded (it already does that with callbacks that trace all functions). If a callback happens to have a module function being traced, it increments that records ref count and starts tracing that function. There may be a strange side effect with this, where tracing module functions on unload and then reloading a new module may have that new module's functions being traced. This may be something that confuses the user, but it's not a big deal. Another approach is to disable all callback hashes on module unload, but this leaves some ftrace callbacks that may not be registered, but can still have hashes tracing the module's function where ftrace doesn't know about it. That situation can cause the same bug. This solution solves that case too. Another benefit of this solution, is it is possible to trace a module's function on unload and load. Link: http://lkml.kernel.org/r/20130705142629.GA325@redhat.com Reported-by: Jörn Engel Reported-by: Dave Jones Reported-by: Steve Hodgson Tested-by: Steve Hodgson Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 71 ++++++++++++++++++++++++++++++++++++++++++++------- 1 file changed, 62 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 92d3334de0c3..a6d098c6df3f 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -2169,12 +2169,57 @@ static cycle_t ftrace_update_time; static unsigned long ftrace_update_cnt; unsigned long ftrace_update_tot_cnt; -static int ops_traces_mod(struct ftrace_ops *ops) +static inline int ops_traces_mod(struct ftrace_ops *ops) { - struct ftrace_hash *hash; + /* + * Filter_hash being empty will default to trace module. + * But notrace hash requires a test of individual module functions. + */ + return ftrace_hash_empty(ops->filter_hash) && + ftrace_hash_empty(ops->notrace_hash); +} + +/* + * Check if the current ops references the record. + * + * If the ops traces all functions, then it was already accounted for. + * If the ops does not trace the current record function, skip it. + * If the ops ignores the function via notrace filter, skip it. + */ +static inline bool +ops_references_rec(struct ftrace_ops *ops, struct dyn_ftrace *rec) +{ + /* If ops isn't enabled, ignore it */ + if (!(ops->flags & FTRACE_OPS_FL_ENABLED)) + return 0; + + /* If ops traces all mods, we already accounted for it */ + if (ops_traces_mod(ops)) + return 0; + + /* The function must be in the filter */ + if (!ftrace_hash_empty(ops->filter_hash) && + !ftrace_lookup_ip(ops->filter_hash, rec->ip)) + return 0; - hash = ops->filter_hash; - return ftrace_hash_empty(hash); + /* If in notrace hash, we ignore it too */ + if (ftrace_lookup_ip(ops->notrace_hash, rec->ip)) + return 0; + + return 1; +} + +static int referenced_filters(struct dyn_ftrace *rec) +{ + struct ftrace_ops *ops; + int cnt = 0; + + for (ops = ftrace_ops_list; ops != &ftrace_list_end; ops = ops->next) { + if (ops_references_rec(ops, rec)) + cnt++; + } + + return cnt; } static int ftrace_update_code(struct module *mod) @@ -2183,6 +2228,7 @@ static int ftrace_update_code(struct module *mod) struct dyn_ftrace *p; cycle_t start, stop; unsigned long ref = 0; + bool test = false; int i; /* @@ -2196,9 +2242,12 @@ static int ftrace_update_code(struct module *mod) for (ops = ftrace_ops_list; ops != &ftrace_list_end; ops = ops->next) { - if (ops->flags & FTRACE_OPS_FL_ENABLED && - ops_traces_mod(ops)) - ref++; + if (ops->flags & FTRACE_OPS_FL_ENABLED) { + if (ops_traces_mod(ops)) + ref++; + else + test = true; + } } } @@ -2208,12 +2257,16 @@ static int ftrace_update_code(struct module *mod) for (pg = ftrace_new_pgs; pg; pg = pg->next) { for (i = 0; i < pg->index; i++) { + int cnt = ref; + /* If something went wrong, bail without enabling anything */ if (unlikely(ftrace_disabled)) return -1; p = &pg->records[i]; - p->flags = ref; + if (test) + cnt += referenced_filters(p); + p->flags = cnt; /* * Do the initial record conversion from mcount jump @@ -2233,7 +2286,7 @@ static int ftrace_update_code(struct module *mod) * conversion puts the module to the correct state, thus * passing the ftrace_make_call check. */ - if (ftrace_start_up && ref) { + if (ftrace_start_up && cnt) { int failed = __ftrace_replace_code(p, 1); if (failed) ftrace_bug(failed, p->ip); -- cgit v1.2.3 From da0a12caffad2eeadea429f83818408e7b77379a Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 31 Jul 2013 16:16:28 +0800 Subject: cgroup: fix a leak when percpu_ref_init() fails ss->css_free() is not called when perfcpu_ref_init() fails. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cgroup.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index afb8d53ca6c7..468e410f9e61 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4344,8 +4344,10 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, } err = percpu_ref_init(&css->refcnt, css_release); - if (err) + if (err) { + ss->css_free(cgrp); goto err_free_all; + } init_cgroup_css(css, ss, cgrp); -- cgit v1.2.3 From 2a4ac63333584b2791986cf2270f5ba9a4b97606 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 31 Jul 2013 16:16:40 +0800 Subject: cgroup: remove sparse tags from offline_css() This should have been removed in commit d7eeac1913ff ("cgroup: hold cgroup_mutex before calling css_offline"). While at it, update the comments. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cgroup.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 345fac8e4fba..41b559f51502 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4214,7 +4214,7 @@ static void init_cgroup_css(struct cgroup_subsys_state *css, INIT_WORK(&css->dput_work, css_dput_fn); } -/* invoke ->post_create() on a new CSS and mark it online if successful */ +/* invoke ->css_online() on a new CSS and mark it online if successful */ static int online_css(struct cgroup_subsys *ss, struct cgroup *cgrp) { int ret = 0; @@ -4228,9 +4228,8 @@ static int online_css(struct cgroup_subsys *ss, struct cgroup *cgrp) return ret; } -/* if the CSS is online, invoke ->pre_destory() on it and mark it offline */ +/* if the CSS is online, invoke ->css_offline() on it and mark it offline */ static void offline_css(struct cgroup_subsys *ss, struct cgroup *cgrp) - __releases(&cgroup_mutex) __acquires(&cgroup_mutex) { struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; -- cgit v1.2.3 From e0798ce27346edb8aa369b5b39af5a47fdf2b25c Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 31 Jul 2013 17:36:25 +0800 Subject: cgroup: remove struct cgroup_seqfile_state We can use struct cfent instead. v2: - remove cgroup_seqfile_release(). Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cgroup.c | 45 +++++++++++++-------------------------------- 1 file changed, 13 insertions(+), 32 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 41b559f51502..ed2104304833 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -2397,11 +2397,6 @@ static ssize_t cgroup_file_read(struct file *file, char __user *buf, * supports string->u64 maps, but can be extended in future. */ -struct cgroup_seqfile_state { - struct cftype *cft; - struct cgroup *cgroup; -}; - static int cgroup_map_add(struct cgroup_map_cb *cb, const char *key, u64 value) { struct seq_file *sf = cb->state; @@ -2410,59 +2405,45 @@ static int cgroup_map_add(struct cgroup_map_cb *cb, const char *key, u64 value) static int cgroup_seqfile_show(struct seq_file *m, void *arg) { - struct cgroup_seqfile_state *state = m->private; - struct cftype *cft = state->cft; + struct cfent *cfe = m->private; + struct cftype *cft = cfe->type; + struct cgroup *cgrp = __d_cgrp(cfe->dentry->d_parent); + if (cft->read_map) { struct cgroup_map_cb cb = { .fill = cgroup_map_add, .state = m, }; - return cft->read_map(state->cgroup, cft, &cb); + return cft->read_map(cgrp, cft, &cb); } - return cft->read_seq_string(state->cgroup, cft, m); -} - -static int cgroup_seqfile_release(struct inode *inode, struct file *file) -{ - struct seq_file *seq = file->private_data; - kfree(seq->private); - return single_release(inode, file); + return cft->read_seq_string(cgrp, cft, m); } static const struct file_operations cgroup_seqfile_operations = { .read = seq_read, .write = cgroup_file_write, .llseek = seq_lseek, - .release = cgroup_seqfile_release, + .release = single_release, }; static int cgroup_file_open(struct inode *inode, struct file *file) { int err; + struct cfent *cfe; struct cftype *cft; err = generic_file_open(inode, file); if (err) return err; - cft = __d_cft(file->f_dentry); + cfe = __d_cfe(file->f_dentry); + cft = cfe->type; if (cft->read_map || cft->read_seq_string) { - struct cgroup_seqfile_state *state; - - state = kzalloc(sizeof(*state), GFP_USER); - if (!state) - return -ENOMEM; - - state->cft = cft; - state->cgroup = __d_cgrp(file->f_dentry->d_parent); file->f_op = &cgroup_seqfile_operations; - err = single_open(file, cgroup_seqfile_show, state); - if (err < 0) - kfree(state); - } else if (cft->open) + err = single_open(file, cgroup_seqfile_show, cfe); + } else if (cft->open) { err = cft->open(inode, file); - else - err = 0; + } return err; } -- cgit v1.2.3 From 6f4b7e632d78c2d91502211c430722cc66428492 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 31 Jul 2013 16:18:36 +0800 Subject: cgroup: more naming cleanups Constantly use @cset for css_set variables and use @cgrp as cgroup variables. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cgroup.c | 26 +++++++++++++------------- kernel/cpuset.c | 16 ++++++++-------- 2 files changed, 21 insertions(+), 21 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index ed2104304833..9577bebe2546 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -466,7 +466,7 @@ static inline void put_css_set_taskexit(struct css_set *cset) * @new_cgrp: cgroup that's being entered by the task * @template: desired set of css pointers in css_set (pre-calculated) * - * Returns true if "cg" matches "old_cg" except for the hierarchy + * Returns true if "cset" matches "old_cset" except for the hierarchy * which "new_cgrp" belongs to, for which it should match "new_cgrp". */ static bool compare_css_sets(struct css_set *cset, @@ -1839,7 +1839,7 @@ EXPORT_SYMBOL_GPL(task_cgroup_path_from_hierarchy); struct task_and_cgroup { struct task_struct *task; struct cgroup *cgrp; - struct css_set *cg; + struct css_set *cset; }; struct cgroup_taskset { @@ -2057,8 +2057,8 @@ static int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk, tc = flex_array_get(group, i); old_cset = task_css_set(tc->task); - tc->cg = find_css_set(old_cset, cgrp); - if (!tc->cg) { + tc->cset = find_css_set(old_cset, cgrp); + if (!tc->cset) { retval = -ENOMEM; goto out_put_css_set_refs; } @@ -2071,7 +2071,7 @@ static int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk, */ for (i = 0; i < group_size; i++) { tc = flex_array_get(group, i); - cgroup_task_migrate(tc->cgrp, tc->task, tc->cg); + cgroup_task_migrate(tc->cgrp, tc->task, tc->cset); } /* nothing is sensitive to fork() after this point. */ @@ -2091,9 +2091,9 @@ out_put_css_set_refs: if (retval) { for (i = 0; i < group_size; i++) { tc = flex_array_get(group, i); - if (!tc->cg) + if (!tc->cset) break; - put_css_set(tc->cg); + put_css_set(tc->cset); } } out_cancel_attach: @@ -2203,9 +2203,9 @@ int cgroup_attach_task_all(struct task_struct *from, struct task_struct *tsk) mutex_lock(&cgroup_mutex); for_each_active_root(root) { - struct cgroup *from_cg = task_cgroup_from_root(from, root); + struct cgroup *from_cgrp = task_cgroup_from_root(from, root); - retval = cgroup_attach_task(from_cg, tsk, false); + retval = cgroup_attach_task(from_cgrp, tsk, false); if (retval) break; } @@ -3305,8 +3305,8 @@ int cgroup_scan_tasks(struct cgroup_scanner *scan) * guarantees forward progress and that we don't miss any tasks. */ heap->size = 0; - cgroup_iter_start(scan->cg, &it); - while ((p = cgroup_iter_next(scan->cg, &it))) { + cgroup_iter_start(scan->cgrp, &it); + while ((p = cgroup_iter_next(scan->cgrp, &it))) { /* * Only affect tasks that qualify per the caller's callback, * if he provided one @@ -3339,7 +3339,7 @@ int cgroup_scan_tasks(struct cgroup_scanner *scan) * the heap and wasn't inserted */ } - cgroup_iter_end(scan->cg, &it); + cgroup_iter_end(scan->cgrp, &it); if (heap->size) { for (i = 0; i < heap->size; i++) { @@ -3385,7 +3385,7 @@ int cgroup_transfer_tasks(struct cgroup *to, struct cgroup *from) { struct cgroup_scanner scan; - scan.cg = from; + scan.cgrp = from; scan.test_task = NULL; /* select all tasks in cgroup */ scan.process_task = cgroup_transfer_one_task; scan.heap = NULL; diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 703bfd5a32a9..1b9c31549797 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -845,7 +845,7 @@ static void cpuset_change_cpumask(struct task_struct *tsk, { struct cpuset *cpus_cs; - cpus_cs = effective_cpumask_cpuset(cgroup_cs(scan->cg)); + cpus_cs = effective_cpumask_cpuset(cgroup_cs(scan->cgrp)); set_cpus_allowed_ptr(tsk, cpus_cs->cpus_allowed); } @@ -866,7 +866,7 @@ static void update_tasks_cpumask(struct cpuset *cs, struct ptr_heap *heap) { struct cgroup_scanner scan; - scan.cg = cs->css.cgroup; + scan.cgrp = cs->css.cgroup; scan.test_task = NULL; scan.process_task = cpuset_change_cpumask; scan.heap = heap; @@ -1062,7 +1062,7 @@ static void cpuset_change_task_nodemask(struct task_struct *tsk, static void cpuset_change_nodemask(struct task_struct *p, struct cgroup_scanner *scan) { - struct cpuset *cs = cgroup_cs(scan->cg); + struct cpuset *cs = cgroup_cs(scan->cgrp); struct mm_struct *mm; int migrate; nodemask_t *newmems = scan->data; @@ -1102,7 +1102,7 @@ static void update_tasks_nodemask(struct cpuset *cs, struct ptr_heap *heap) guarantee_online_mems(mems_cs, &newmems); - scan.cg = cs->css.cgroup; + scan.cgrp = cs->css.cgroup; scan.test_task = NULL; scan.process_task = cpuset_change_nodemask; scan.heap = heap; @@ -1275,7 +1275,7 @@ static int update_relax_domain_level(struct cpuset *cs, s64 val) static void cpuset_change_flag(struct task_struct *tsk, struct cgroup_scanner *scan) { - cpuset_update_task_spread_flag(cgroup_cs(scan->cg), tsk); + cpuset_update_task_spread_flag(cgroup_cs(scan->cgrp), tsk); } /* @@ -1295,7 +1295,7 @@ static void update_tasks_flags(struct cpuset *cs, struct ptr_heap *heap) { struct cgroup_scanner scan; - scan.cg = cs->css.cgroup; + scan.cgrp = cs->css.cgroup; scan.test_task = NULL; scan.process_task = cpuset_change_flag; scan.heap = heap; @@ -1971,7 +1971,7 @@ static int cpuset_css_online(struct cgroup *cgrp) struct cpuset *cs = cgroup_cs(cgrp); struct cpuset *parent = parent_cs(cs); struct cpuset *tmp_cs; - struct cgroup *pos_cg; + struct cgroup *pos_cgrp; if (!parent) return 0; @@ -2003,7 +2003,7 @@ static int cpuset_css_online(struct cgroup *cgrp) * (and likewise for mems) to the new cgroup. */ rcu_read_lock(); - cpuset_for_each_child(tmp_cs, pos_cg, parent) { + cpuset_for_each_child(tmp_cs, pos_cgrp, parent) { if (is_mem_exclusive(tmp_cs) || is_cpu_exclusive(tmp_cs)) { rcu_read_unlock(); goto out_unlock; -- cgit v1.2.3 From 4e96ee8e981b5140a2bcc5fff0d5c0eef39a62ee Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 31 Jul 2013 09:50:50 +0800 Subject: cgroup: convert cgroup_ida to cgroup_idr This enables us to lookup a cgroup by its id. v4: - add a comment for idr_remove() in cgroup_offline_fn(). v3: - on success, idr_alloc() returns the id but not 0, so fix the BUG_ON() in cgroup_init(). - pass the right value to idr_alloc() so that the id for dummy cgroup is 0. Signed-off-by: Li Zefan Reviewed-by: Michal Hocko Signed-off-by: Tejun Heo --- kernel/cgroup.c | 33 +++++++++++++++++++++++++++------ 1 file changed, 27 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 9577bebe2546..3f6593333525 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -866,8 +866,6 @@ static void cgroup_free_fn(struct work_struct *work) */ dput(cgrp->parent->dentry); - ida_simple_remove(&cgrp->root->cgroup_ida, cgrp->id); - /* * Drop the active superblock reference that we took when we * created the cgroup. This will free cgrp->root, if we are @@ -1379,6 +1377,7 @@ static void init_cgroup_root(struct cgroupfs_root *root) cgrp->root = root; RCU_INIT_POINTER(cgrp->name, &root_cgroup_name); init_cgroup_housekeeping(cgrp); + idr_init(&root->cgroup_idr); } static int cgroup_init_root_id(struct cgroupfs_root *root, int start, int end) @@ -1451,7 +1450,6 @@ static struct cgroupfs_root *cgroup_root_from_opts(struct cgroup_sb_opts *opts) */ root->subsys_mask = opts->subsys_mask; root->flags = opts->flags; - ida_init(&root->cgroup_ida); if (opts->release_agent) strcpy(root->release_agent_path, opts->release_agent); if (opts->name) @@ -1467,7 +1465,7 @@ static void cgroup_free_root(struct cgroupfs_root *root) /* hierarhcy ID shoulid already have been released */ WARN_ON_ONCE(root->hierarchy_id); - ida_destroy(&root->cgroup_ida); + idr_destroy(&root->cgroup_idr); kfree(root); } } @@ -1582,6 +1580,11 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, mutex_lock(&cgroup_mutex); mutex_lock(&cgroup_root_mutex); + root_cgrp->id = idr_alloc(&root->cgroup_idr, root_cgrp, + 0, 1, GFP_KERNEL); + if (root_cgrp->id < 0) + goto unlock_drop; + /* Check for name clashes with existing mounts */ ret = -EBUSY; if (strlen(root->name)) @@ -4253,7 +4256,11 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, goto err_free_cgrp; rcu_assign_pointer(cgrp->name, name); - cgrp->id = ida_simple_get(&root->cgroup_ida, 1, 0, GFP_KERNEL); + /* + * Temporarily set the pointer to NULL, so idr_find() won't return + * a half-baked cgroup. + */ + cgrp->id = idr_alloc(&root->cgroup_idr, NULL, 1, 0, GFP_KERNEL); if (cgrp->id < 0) goto err_free_name; @@ -4351,6 +4358,8 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, } } + idr_replace(&root->cgroup_idr, cgrp, cgrp->id); + err = cgroup_addrm_files(cgrp, NULL, cgroup_base_files, true); if (err) goto err_destroy; @@ -4377,7 +4386,7 @@ err_free_all: /* Release the reference count that we took on the superblock */ deactivate_super(sb); err_free_id: - ida_simple_remove(&root->cgroup_ida, cgrp->id); + idr_remove(&root->cgroup_idr, cgrp->id); err_free_name: kfree(rcu_dereference_raw(cgrp->name)); err_free_cgrp: @@ -4570,6 +4579,14 @@ static void cgroup_offline_fn(struct work_struct *work) /* delete this cgroup from parent->children */ list_del_rcu(&cgrp->sibling); + /* + * We should remove the cgroup object from idr before its grace + * period starts, so we won't be looking up a cgroup while the + * cgroup is being freed. + */ + idr_remove(&cgrp->root->cgroup_idr, cgrp->id); + cgrp->id = -1; + dput(d); set_bit(CGRP_RELEASABLE, &parent->flags); @@ -4895,6 +4912,10 @@ int __init cgroup_init(void) BUG_ON(cgroup_init_root_id(&cgroup_dummy_root, 0, 1)); + err = idr_alloc(&cgroup_dummy_root.cgroup_idr, cgroup_dummy_top, + 0, 1, GFP_KERNEL); + BUG_ON(err < 0); + mutex_unlock(&cgroup_root_mutex); mutex_unlock(&cgroup_mutex); -- cgit v1.2.3 From 2816c551c796ec14620325b2c9ed75b9979d3125 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Mon, 29 Jul 2013 19:50:33 +0200 Subject: tracing: trace_remove_event_call() should fail if call/file is in use Change trace_remove_event_call(call) to return the error if this call is active. This is what the callers assume but can't verify outside of the tracing locks. Both trace_kprobe.c/trace_uprobe.c need the additional changes, unregister_trace_probe() should abort if trace_remove_event_call() fails. The caller is going to free this call/file so we must ensure that nobody can use them after trace_remove_event_call() succeeds. debugfs should be fine after the previous changes and event_remove() does TRACE_REG_UNREGISTER, but still there are 2 reasons why we need the additional checks: - There could be a perf_event(s) attached to this tp_event, so the patch checks ->perf_refcount. - TRACE_REG_UNREGISTER can be suppressed by FTRACE_EVENT_FL_SOFT_MODE, so we simply check FTRACE_EVENT_FL_ENABLED protected by event_mutex. Link: http://lkml.kernel.org/r/20130729175033.GB26284@redhat.com Reviewed-by: Masami Hiramatsu Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 35 +++++++++++++++++++++++++++++++++-- 1 file changed, 33 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index a67c913e2f9f..ec04836273c0 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -1713,16 +1713,47 @@ static void __trace_remove_event_call(struct ftrace_event_call *call) destroy_preds(call); } +static int probe_remove_event_call(struct ftrace_event_call *call) +{ + struct trace_array *tr; + struct ftrace_event_file *file; + +#ifdef CONFIG_PERF_EVENTS + if (call->perf_refcount) + return -EBUSY; +#endif + do_for_each_event_file(tr, file) { + if (file->event_call != call) + continue; + /* + * We can't rely on ftrace_event_enable_disable(enable => 0) + * we are going to do, FTRACE_EVENT_FL_SOFT_MODE can suppress + * TRACE_REG_UNREGISTER. + */ + if (file->flags & FTRACE_EVENT_FL_ENABLED) + return -EBUSY; + break; + } while_for_each_event_file(); + + __trace_remove_event_call(call); + + return 0; +} + /* Remove an event_call */ -void trace_remove_event_call(struct ftrace_event_call *call) +int trace_remove_event_call(struct ftrace_event_call *call) { + int ret; + mutex_lock(&trace_types_lock); mutex_lock(&event_mutex); down_write(&trace_event_sem); - __trace_remove_event_call(call); + ret = probe_remove_event_call(call); up_write(&trace_event_sem); mutex_unlock(&event_mutex); mutex_unlock(&trace_types_lock); + + return ret; } #define for_each_event(event, start, end) \ -- cgit v1.2.3 From 2ba64035d0ca966fd189bc3e0826343fc81bf482 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Wed, 31 Jul 2013 13:16:22 -0400 Subject: tracing: Add comment to describe special break case in probe_remove_event_call() The "break" used in the do_for_each_event_file() is used as an optimization as the loop is really a double loop. The loop searches all event files for each trace_array. There's only one matching event file per trace_array and after we find the event file for the trace_array, the break is used to jump to the next trace_array and start the search there. As this is not a standard way of using "break" in C code, it requires a comment right before the break to let people know what is going on. Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index ec04836273c0..29a7ebcfb426 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -1732,6 +1732,12 @@ static int probe_remove_event_call(struct ftrace_event_call *call) */ if (file->flags & FTRACE_EVENT_FL_ENABLED) return -EBUSY; + /* + * The do_for_each_event_file_safe() is + * a double loop. After finding the call for this + * trace_array, we use break to jump to the next + * trace_array. + */ break; } while_for_each_event_file(); -- cgit v1.2.3 From 10e84b97ed799be404836dc7f71ab47d4571265a Mon Sep 17 00:00:00 2001 From: Dave Kleikamp Date: Wed, 31 Jul 2013 13:53:35 -0700 Subject: mm: sched: numa: fix NUMA balancing when !SCHED_DEBUG Commit 3105b86a9fee ("mm: sched: numa: Control enabling and disabling of NUMA balancing if !SCHED_DEBUG") defined numabalancing_enabled to control the enabling and disabling of automatic NUMA balancing, but it is never used. I believe the intention was to use this in place of sched_feat_numa(NUMA). Currently, if SCHED_DEBUG is not defined, sched_feat_numa(NUMA) will never be changed from the initial "false". Signed-off-by: Dave Kleikamp Acked-by: Mel Gorman Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sched/fair.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index bb456f44b7b1..9565645e3202 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -851,7 +851,7 @@ void task_numa_fault(int node, int pages, bool migrated) { struct task_struct *p = current; - if (!sched_feat_numa(NUMA)) + if (!numabalancing_enabled) return; /* FIXME: Allocate task-specific structure for placement policy here */ @@ -5786,7 +5786,7 @@ static void task_tick_fair(struct rq *rq, struct task_struct *curr, int queued) entity_tick(cfs_rq, se, queued); } - if (sched_feat_numa(NUMA)) + if (numabalancing_enabled) task_tick_numa(rq, curr); update_rq_runnable_avg(rq, 1); -- cgit v1.2.3 From b9ee979e9d770dc10f94936ef6ff9efddc23c911 Mon Sep 17 00:00:00 2001 From: Joe Perches Date: Wed, 31 Jul 2013 13:53:42 -0700 Subject: printk: move to separate directory for easier modification Make it easier to break up printk into bite-sized chunks. Remove printk path/filename from comment. Signed-off-by: Joe Perches Cc: Samuel Thibault Cc: Ming Lei Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/Makefile | 3 +- kernel/printk.c | 2924 ------------------------------------------------ kernel/printk/Makefile | 1 + kernel/printk/printk.c | 2924 ++++++++++++++++++++++++++++++++++++++++++++++++ 4 files changed, 2927 insertions(+), 2925 deletions(-) delete mode 100644 kernel/printk.c create mode 100644 kernel/printk/Makefile create mode 100644 kernel/printk/printk.c (limited to 'kernel') diff --git a/kernel/Makefile b/kernel/Makefile index 470839d1a30e..35ef1185e359 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -2,7 +2,7 @@ # Makefile for the linux kernel. # -obj-y = fork.o exec_domain.o panic.o printk.o \ +obj-y = fork.o exec_domain.o panic.o \ cpu.o exit.o itimer.o time.o softirq.o resource.o \ sysctl.o sysctl_binary.o capability.o ptrace.o timer.o user.o \ signal.o sys.o kmod.o workqueue.o pid.o task_work.o \ @@ -24,6 +24,7 @@ endif obj-y += sched/ obj-y += power/ +obj-y += printk/ obj-y += cpu/ obj-$(CONFIG_CHECKPOINT_RESTORE) += kcmp.o diff --git a/kernel/printk.c b/kernel/printk.c deleted file mode 100644 index 69b0890ed7e5..000000000000 --- a/kernel/printk.c +++ /dev/null @@ -1,2924 +0,0 @@ -/* - * linux/kernel/printk.c - * - * Copyright (C) 1991, 1992 Linus Torvalds - * - * Modified to make sys_syslog() more flexible: added commands to - * return the last 4k of kernel messages, regardless of whether - * they've been read or not. Added option to suppress kernel printk's - * to the console. Added hook for sending the console messages - * elsewhere, in preparation for a serial line console (someday). - * Ted Ts'o, 2/11/93. - * Modified for sysctl support, 1/8/97, Chris Horn. - * Fixed SMP synchronization, 08/08/99, Manfred Spraul - * manfred@colorfullife.com - * Rewrote bits to get rid of console_lock - * 01Mar01 Andrew Morton - */ - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include /* For in_interrupt() */ -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include - -#define CREATE_TRACE_POINTS -#include - -/* printk's without a loglevel use this.. */ -#define DEFAULT_MESSAGE_LOGLEVEL CONFIG_DEFAULT_MESSAGE_LOGLEVEL - -/* We show everything that is MORE important than this.. */ -#define MINIMUM_CONSOLE_LOGLEVEL 1 /* Minimum loglevel we let people use */ -#define DEFAULT_CONSOLE_LOGLEVEL 7 /* anything MORE serious than KERN_DEBUG */ - -int console_printk[4] = { - DEFAULT_CONSOLE_LOGLEVEL, /* console_loglevel */ - DEFAULT_MESSAGE_LOGLEVEL, /* default_message_loglevel */ - MINIMUM_CONSOLE_LOGLEVEL, /* minimum_console_loglevel */ - DEFAULT_CONSOLE_LOGLEVEL, /* default_console_loglevel */ -}; - -/* - * Low level drivers may need that to know if they can schedule in - * their unblank() callback or not. So let's export it. - */ -int oops_in_progress; -EXPORT_SYMBOL(oops_in_progress); - -/* - * console_sem protects the console_drivers list, and also - * provides serialisation for access to the entire console - * driver system. - */ -static DEFINE_SEMAPHORE(console_sem); -struct console *console_drivers; -EXPORT_SYMBOL_GPL(console_drivers); - -#ifdef CONFIG_LOCKDEP -static struct lockdep_map console_lock_dep_map = { - .name = "console_lock" -}; -#endif - -/* - * This is used for debugging the mess that is the VT code by - * keeping track if we have the console semaphore held. It's - * definitely not the perfect debug tool (we don't know if _WE_ - * hold it are racing, but it helps tracking those weird code - * path in the console code where we end up in places I want - * locked without the console sempahore held - */ -static int console_locked, console_suspended; - -/* - * If exclusive_console is non-NULL then only this console is to be printed to. - */ -static struct console *exclusive_console; - -/* - * Array of consoles built from command line options (console=) - */ -struct console_cmdline -{ - char name[8]; /* Name of the driver */ - int index; /* Minor dev. to use */ - char *options; /* Options for the driver */ -#ifdef CONFIG_A11Y_BRAILLE_CONSOLE - char *brl_options; /* Options for braille driver */ -#endif -}; - -#define MAX_CMDLINECONSOLES 8 - -static struct console_cmdline console_cmdline[MAX_CMDLINECONSOLES]; -static int selected_console = -1; -static int preferred_console = -1; -int console_set_on_cmdline; -EXPORT_SYMBOL(console_set_on_cmdline); - -/* Flag: console code may call schedule() */ -static int console_may_schedule; - -/* - * The printk log buffer consists of a chain of concatenated variable - * length records. Every record starts with a record header, containing - * the overall length of the record. - * - * The heads to the first and last entry in the buffer, as well as the - * sequence numbers of these both entries are maintained when messages - * are stored.. - * - * If the heads indicate available messages, the length in the header - * tells the start next message. A length == 0 for the next message - * indicates a wrap-around to the beginning of the buffer. - * - * Every record carries the monotonic timestamp in microseconds, as well as - * the standard userspace syslog level and syslog facility. The usual - * kernel messages use LOG_KERN; userspace-injected messages always carry - * a matching syslog facility, by default LOG_USER. The origin of every - * message can be reliably determined that way. - * - * The human readable log message directly follows the message header. The - * length of the message text is stored in the header, the stored message - * is not terminated. - * - * Optionally, a message can carry a dictionary of properties (key/value pairs), - * to provide userspace with a machine-readable message context. - * - * Examples for well-defined, commonly used property names are: - * DEVICE=b12:8 device identifier - * b12:8 block dev_t - * c127:3 char dev_t - * n8 netdev ifindex - * +sound:card0 subsystem:devname - * SUBSYSTEM=pci driver-core subsystem name - * - * Valid characters in property names are [a-zA-Z0-9.-_]. The plain text value - * follows directly after a '=' character. Every property is terminated by - * a '\0' character. The last property is not terminated. - * - * Example of a message structure: - * 0000 ff 8f 00 00 00 00 00 00 monotonic time in nsec - * 0008 34 00 record is 52 bytes long - * 000a 0b 00 text is 11 bytes long - * 000c 1f 00 dictionary is 23 bytes long - * 000e 03 00 LOG_KERN (facility) LOG_ERR (level) - * 0010 69 74 27 73 20 61 20 6c "it's a l" - * 69 6e 65 "ine" - * 001b 44 45 56 49 43 "DEVIC" - * 45 3d 62 38 3a 32 00 44 "E=b8:2\0D" - * 52 49 56 45 52 3d 62 75 "RIVER=bu" - * 67 "g" - * 0032 00 00 00 padding to next message header - * - * The 'struct log' buffer header must never be directly exported to - * userspace, it is a kernel-private implementation detail that might - * need to be changed in the future, when the requirements change. - * - * /dev/kmsg exports the structured data in the following line format: - * "level,sequnum,timestamp;\n" - * - * The optional key/value pairs are attached as continuation lines starting - * with a space character and terminated by a newline. All possible - * non-prinatable characters are escaped in the "\xff" notation. - * - * Users of the export format should ignore possible additional values - * separated by ',', and find the message after the ';' character. - */ - -enum log_flags { - LOG_NOCONS = 1, /* already flushed, do not print to console */ - LOG_NEWLINE = 2, /* text ended with a newline */ - LOG_PREFIX = 4, /* text started with a prefix */ - LOG_CONT = 8, /* text is a fragment of a continuation line */ -}; - -struct log { - u64 ts_nsec; /* timestamp in nanoseconds */ - u16 len; /* length of entire record */ - u16 text_len; /* length of text buffer */ - u16 dict_len; /* length of dictionary buffer */ - u8 facility; /* syslog facility */ - u8 flags:5; /* internal record flags */ - u8 level:3; /* syslog level */ -}; - -/* - * The logbuf_lock protects kmsg buffer, indices, counters. It is also - * used in interesting ways to provide interlocking in console_unlock(); - */ -static DEFINE_RAW_SPINLOCK(logbuf_lock); - -#ifdef CONFIG_PRINTK -DECLARE_WAIT_QUEUE_HEAD(log_wait); -/* the next printk record to read by syslog(READ) or /proc/kmsg */ -static u64 syslog_seq; -static u32 syslog_idx; -static enum log_flags syslog_prev; -static size_t syslog_partial; - -/* index and sequence number of the first record stored in the buffer */ -static u64 log_first_seq; -static u32 log_first_idx; - -/* index and sequence number of the next record to store in the buffer */ -static u64 log_next_seq; -static u32 log_next_idx; - -/* the next printk record to write to the console */ -static u64 console_seq; -static u32 console_idx; -static enum log_flags console_prev; - -/* the next printk record to read after the last 'clear' command */ -static u64 clear_seq; -static u32 clear_idx; - -#define PREFIX_MAX 32 -#define LOG_LINE_MAX 1024 - PREFIX_MAX - -/* record buffer */ -#if defined(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS) -#define LOG_ALIGN 4 -#else -#define LOG_ALIGN __alignof__(struct log) -#endif -#define __LOG_BUF_LEN (1 << CONFIG_LOG_BUF_SHIFT) -static char __log_buf[__LOG_BUF_LEN] __aligned(LOG_ALIGN); -static char *log_buf = __log_buf; -static u32 log_buf_len = __LOG_BUF_LEN; - -/* cpu currently holding logbuf_lock */ -static volatile unsigned int logbuf_cpu = UINT_MAX; - -/* human readable text of the record */ -static char *log_text(const struct log *msg) -{ - return (char *)msg + sizeof(struct log); -} - -/* optional key/value pair dictionary attached to the record */ -static char *log_dict(const struct log *msg) -{ - return (char *)msg + sizeof(struct log) + msg->text_len; -} - -/* get record by index; idx must point to valid msg */ -static struct log *log_from_idx(u32 idx) -{ - struct log *msg = (struct log *)(log_buf + idx); - - /* - * A length == 0 record is the end of buffer marker. Wrap around and - * read the message at the start of the buffer. - */ - if (!msg->len) - return (struct log *)log_buf; - return msg; -} - -/* get next record; idx must point to valid msg */ -static u32 log_next(u32 idx) -{ - struct log *msg = (struct log *)(log_buf + idx); - - /* length == 0 indicates the end of the buffer; wrap */ - /* - * A length == 0 record is the end of buffer marker. Wrap around and - * read the message at the start of the buffer as *this* one, and - * return the one after that. - */ - if (!msg->len) { - msg = (struct log *)log_buf; - return msg->len; - } - return idx + msg->len; -} - -/* insert record into the buffer, discard old ones, update heads */ -static void log_store(int facility, int level, - enum log_flags flags, u64 ts_nsec, - const char *dict, u16 dict_len, - const char *text, u16 text_len) -{ - struct log *msg; - u32 size, pad_len; - - /* number of '\0' padding bytes to next message */ - size = sizeof(struct log) + text_len + dict_len; - pad_len = (-size) & (LOG_ALIGN - 1); - size += pad_len; - - while (log_first_seq < log_next_seq) { - u32 free; - - if (log_next_idx > log_first_idx) - free = max(log_buf_len - log_next_idx, log_first_idx); - else - free = log_first_idx - log_next_idx; - - if (free > size + sizeof(struct log)) - break; - - /* drop old messages until we have enough contiuous space */ - log_first_idx = log_next(log_first_idx); - log_first_seq++; - } - - if (log_next_idx + size + sizeof(struct log) >= log_buf_len) { - /* - * This message + an additional empty header does not fit - * at the end of the buffer. Add an empty header with len == 0 - * to signify a wrap around. - */ - memset(log_buf + log_next_idx, 0, sizeof(struct log)); - log_next_idx = 0; - } - - /* fill message */ - msg = (struct log *)(log_buf + log_next_idx); - memcpy(log_text(msg), text, text_len); - msg->text_len = text_len; - memcpy(log_dict(msg), dict, dict_len); - msg->dict_len = dict_len; - msg->facility = facility; - msg->level = level & 7; - msg->flags = flags & 0x1f; - if (ts_nsec > 0) - msg->ts_nsec = ts_nsec; - else - msg->ts_nsec = local_clock(); - memset(log_dict(msg) + dict_len, 0, pad_len); - msg->len = sizeof(struct log) + text_len + dict_len + pad_len; - - /* insert message */ - log_next_idx += msg->len; - log_next_seq++; -} - -#ifdef CONFIG_SECURITY_DMESG_RESTRICT -int dmesg_restrict = 1; -#else -int dmesg_restrict; -#endif - -static int syslog_action_restricted(int type) -{ - if (dmesg_restrict) - return 1; - /* - * Unless restricted, we allow "read all" and "get buffer size" - * for everybody. - */ - return type != SYSLOG_ACTION_READ_ALL && - type != SYSLOG_ACTION_SIZE_BUFFER; -} - -static int check_syslog_permissions(int type, bool from_file) -{ - /* - * If this is from /proc/kmsg and we've already opened it, then we've - * already done the capabilities checks at open time. - */ - if (from_file && type != SYSLOG_ACTION_OPEN) - return 0; - - if (syslog_action_restricted(type)) { - if (capable(CAP_SYSLOG)) - return 0; - /* - * For historical reasons, accept CAP_SYS_ADMIN too, with - * a warning. - */ - if (capable(CAP_SYS_ADMIN)) { - pr_warn_once("%s (%d): Attempt to access syslog with " - "CAP_SYS_ADMIN but no CAP_SYSLOG " - "(deprecated).\n", - current->comm, task_pid_nr(current)); - return 0; - } - return -EPERM; - } - return security_syslog(type); -} - - -/* /dev/kmsg - userspace message inject/listen interface */ -struct devkmsg_user { - u64 seq; - u32 idx; - enum log_flags prev; - struct mutex lock; - char buf[8192]; -}; - -static ssize_t devkmsg_writev(struct kiocb *iocb, const struct iovec *iv, - unsigned long count, loff_t pos) -{ - char *buf, *line; - int i; - int level = default_message_loglevel; - int facility = 1; /* LOG_USER */ - size_t len = iov_length(iv, count); - ssize_t ret = len; - - if (len > LOG_LINE_MAX) - return -EINVAL; - buf = kmalloc(len+1, GFP_KERNEL); - if (buf == NULL) - return -ENOMEM; - - line = buf; - for (i = 0; i < count; i++) { - if (copy_from_user(line, iv[i].iov_base, iv[i].iov_len)) { - ret = -EFAULT; - goto out; - } - line += iv[i].iov_len; - } - - /* - * Extract and skip the syslog prefix <[0-9]*>. Coming from userspace - * the decimal value represents 32bit, the lower 3 bit are the log - * level, the rest are the log facility. - * - * If no prefix or no userspace facility is specified, we - * enforce LOG_USER, to be able to reliably distinguish - * kernel-generated messages from userspace-injected ones. - */ - line = buf; - if (line[0] == '<') { - char *endp = NULL; - - i = simple_strtoul(line+1, &endp, 10); - if (endp && endp[0] == '>') { - level = i & 7; - if (i >> 3) - facility = i >> 3; - endp++; - len -= endp - line; - line = endp; - } - } - line[len] = '\0'; - - printk_emit(facility, level, NULL, 0, "%s", line); -out: - kfree(buf); - return ret; -} - -static ssize_t devkmsg_read(struct file *file, char __user *buf, - size_t count, loff_t *ppos) -{ - struct devkmsg_user *user = file->private_data; - struct log *msg; - u64 ts_usec; - size_t i; - char cont = '-'; - size_t len; - ssize_t ret; - - if (!user) - return -EBADF; - - ret = mutex_lock_interruptible(&user->lock); - if (ret) - return ret; - raw_spin_lock_irq(&logbuf_lock); - while (user->seq == log_next_seq) { - if (file->f_flags & O_NONBLOCK) { - ret = -EAGAIN; - raw_spin_unlock_irq(&logbuf_lock); - goto out; - } - - raw_spin_unlock_irq(&logbuf_lock); - ret = wait_event_interruptible(log_wait, - user->seq != log_next_seq); - if (ret) - goto out; - raw_spin_lock_irq(&logbuf_lock); - } - - if (user->seq < log_first_seq) { - /* our last seen message is gone, return error and reset */ - user->idx = log_first_idx; - user->seq = log_first_seq; - ret = -EPIPE; - raw_spin_unlock_irq(&logbuf_lock); - goto out; - } - - msg = log_from_idx(user->idx); - ts_usec = msg->ts_nsec; - do_div(ts_usec, 1000); - - /* - * If we couldn't merge continuation line fragments during the print, - * export the stored flags to allow an optional external merge of the - * records. Merging the records isn't always neccessarily correct, like - * when we hit a race during printing. In most cases though, it produces - * better readable output. 'c' in the record flags mark the first - * fragment of a line, '+' the following. - */ - if (msg->flags & LOG_CONT && !(user->prev & LOG_CONT)) - cont = 'c'; - else if ((msg->flags & LOG_CONT) || - ((user->prev & LOG_CONT) && !(msg->flags & LOG_PREFIX))) - cont = '+'; - - len = sprintf(user->buf, "%u,%llu,%llu,%c;", - (msg->facility << 3) | msg->level, - user->seq, ts_usec, cont); - user->prev = msg->flags; - - /* escape non-printable characters */ - for (i = 0; i < msg->text_len; i++) { - unsigned char c = log_text(msg)[i]; - - if (c < ' ' || c >= 127 || c == '\\') - len += sprintf(user->buf + len, "\\x%02x", c); - else - user->buf[len++] = c; - } - user->buf[len++] = '\n'; - - if (msg->dict_len) { - bool line = true; - - for (i = 0; i < msg->dict_len; i++) { - unsigned char c = log_dict(msg)[i]; - - if (line) { - user->buf[len++] = ' '; - line = false; - } - - if (c == '\0') { - user->buf[len++] = '\n'; - line = true; - continue; - } - - if (c < ' ' || c >= 127 || c == '\\') { - len += sprintf(user->buf + len, "\\x%02x", c); - continue; - } - - user->buf[len++] = c; - } - user->buf[len++] = '\n'; - } - - user->idx = log_next(user->idx); - user->seq++; - raw_spin_unlock_irq(&logbuf_lock); - - if (len > count) { - ret = -EINVAL; - goto out; - } - - if (copy_to_user(buf, user->buf, len)) { - ret = -EFAULT; - goto out; - } - ret = len; -out: - mutex_unlock(&user->lock); - return ret; -} - -static loff_t devkmsg_llseek(struct file *file, loff_t offset, int whence) -{ - struct devkmsg_user *user = file->private_data; - loff_t ret = 0; - - if (!user) - return -EBADF; - if (offset) - return -ESPIPE; - - raw_spin_lock_irq(&logbuf_lock); - switch (whence) { - case SEEK_SET: - /* the first record */ - user->idx = log_first_idx; - user->seq = log_first_seq; - break; - case SEEK_DATA: - /* - * The first record after the last SYSLOG_ACTION_CLEAR, - * like issued by 'dmesg -c'. Reading /dev/kmsg itself - * changes no global state, and does not clear anything. - */ - user->idx = clear_idx; - user->seq = clear_seq; - break; - case SEEK_END: - /* after the last record */ - user->idx = log_next_idx; - user->seq = log_next_seq; - break; - default: - ret = -EINVAL; - } - raw_spin_unlock_irq(&logbuf_lock); - return ret; -} - -static unsigned int devkmsg_poll(struct file *file, poll_table *wait) -{ - struct devkmsg_user *user = file->private_data; - int ret = 0; - - if (!user) - return POLLERR|POLLNVAL; - - poll_wait(file, &log_wait, wait); - - raw_spin_lock_irq(&logbuf_lock); - if (user->seq < log_next_seq) { - /* return error when data has vanished underneath us */ - if (user->seq < log_first_seq) - ret = POLLIN|POLLRDNORM|POLLERR|POLLPRI; - else - ret = POLLIN|POLLRDNORM; - } - raw_spin_unlock_irq(&logbuf_lock); - - return ret; -} - -static int devkmsg_open(struct inode *inode, struct file *file) -{ - struct devkmsg_user *user; - int err; - - /* write-only does not need any file context */ - if ((file->f_flags & O_ACCMODE) == O_WRONLY) - return 0; - - err = check_syslog_permissions(SYSLOG_ACTION_READ_ALL, - SYSLOG_FROM_READER); - if (err) - return err; - - user = kmalloc(sizeof(struct devkmsg_user), GFP_KERNEL); - if (!user) - return -ENOMEM; - - mutex_init(&user->lock); - - raw_spin_lock_irq(&logbuf_lock); - user->idx = log_first_idx; - user->seq = log_first_seq; - raw_spin_unlock_irq(&logbuf_lock); - - file->private_data = user; - return 0; -} - -static int devkmsg_release(struct inode *inode, struct file *file) -{ - struct devkmsg_user *user = file->private_data; - - if (!user) - return 0; - - mutex_destroy(&user->lock); - kfree(user); - return 0; -} - -const struct file_operations kmsg_fops = { - .open = devkmsg_open, - .read = devkmsg_read, - .aio_write = devkmsg_writev, - .llseek = devkmsg_llseek, - .poll = devkmsg_poll, - .release = devkmsg_release, -}; - -#ifdef CONFIG_KEXEC -/* - * This appends the listed symbols to /proc/vmcoreinfo - * - * /proc/vmcoreinfo is used by various utiilties, like crash and makedumpfile to - * obtain access to symbols that are otherwise very difficult to locate. These - * symbols are specifically used so that utilities can access and extract the - * dmesg log from a vmcore file after a crash. - */ -void log_buf_kexec_setup(void) -{ - VMCOREINFO_SYMBOL(log_buf); - VMCOREINFO_SYMBOL(log_buf_len); - VMCOREINFO_SYMBOL(log_first_idx); - VMCOREINFO_SYMBOL(log_next_idx); - /* - * Export struct log size and field offsets. User space tools can - * parse it and detect any changes to structure down the line. - */ - VMCOREINFO_STRUCT_SIZE(log); - VMCOREINFO_OFFSET(log, ts_nsec); - VMCOREINFO_OFFSET(log, len); - VMCOREINFO_OFFSET(log, text_len); - VMCOREINFO_OFFSET(log, dict_len); -} -#endif - -/* requested log_buf_len from kernel cmdline */ -static unsigned long __initdata new_log_buf_len; - -/* save requested log_buf_len since it's too early to process it */ -static int __init log_buf_len_setup(char *str) -{ - unsigned size = memparse(str, &str); - - if (size) - size = roundup_pow_of_two(size); - if (size > log_buf_len) - new_log_buf_len = size; - - return 0; -} -early_param("log_buf_len", log_buf_len_setup); - -void __init setup_log_buf(int early) -{ - unsigned long flags; - char *new_log_buf; - int free; - - if (!new_log_buf_len) - return; - - if (early) { - unsigned long mem; - - mem = memblock_alloc(new_log_buf_len, PAGE_SIZE); - if (!mem) - return; - new_log_buf = __va(mem); - } else { - new_log_buf = alloc_bootmem_nopanic(new_log_buf_len); - } - - if (unlikely(!new_log_buf)) { - pr_err("log_buf_len: %ld bytes not available\n", - new_log_buf_len); - return; - } - - raw_spin_lock_irqsave(&logbuf_lock, flags); - log_buf_len = new_log_buf_len; - log_buf = new_log_buf; - new_log_buf_len = 0; - free = __LOG_BUF_LEN - log_next_idx; - memcpy(log_buf, __log_buf, __LOG_BUF_LEN); - raw_spin_unlock_irqrestore(&logbuf_lock, flags); - - pr_info("log_buf_len: %d\n", log_buf_len); - pr_info("early log buf free: %d(%d%%)\n", - free, (free * 100) / __LOG_BUF_LEN); -} - -static bool __read_mostly ignore_loglevel; - -static int __init ignore_loglevel_setup(char *str) -{ - ignore_loglevel = 1; - printk(KERN_INFO "debug: ignoring loglevel setting.\n"); - - return 0; -} - -early_param("ignore_loglevel", ignore_loglevel_setup); -module_param(ignore_loglevel, bool, S_IRUGO | S_IWUSR); -MODULE_PARM_DESC(ignore_loglevel, "ignore loglevel setting, to" - "print all kernel messages to the console."); - -#ifdef CONFIG_BOOT_PRINTK_DELAY - -static int boot_delay; /* msecs delay after each printk during bootup */ -static unsigned long long loops_per_msec; /* based on boot_delay */ - -static int __init boot_delay_setup(char *str) -{ - unsigned long lpj; - - lpj = preset_lpj ? preset_lpj : 1000000; /* some guess */ - loops_per_msec = (unsigned long long)lpj / 1000 * HZ; - - get_option(&str, &boot_delay); - if (boot_delay > 10 * 1000) - boot_delay = 0; - - pr_debug("boot_delay: %u, preset_lpj: %ld, lpj: %lu, " - "HZ: %d, loops_per_msec: %llu\n", - boot_delay, preset_lpj, lpj, HZ, loops_per_msec); - return 1; -} -__setup("boot_delay=", boot_delay_setup); - -static void boot_delay_msec(int level) -{ - unsigned long long k; - unsigned long timeout; - - if ((boot_delay == 0 || system_state != SYSTEM_BOOTING) - || (level >= console_loglevel && !ignore_loglevel)) { - return; - } - - k = (unsigned long long)loops_per_msec * boot_delay; - - timeout = jiffies + msecs_to_jiffies(boot_delay); - while (k) { - k--; - cpu_relax(); - /* - * use (volatile) jiffies to prevent - * compiler reduction; loop termination via jiffies - * is secondary and may or may not happen. - */ - if (time_after(jiffies, timeout)) - break; - touch_nmi_watchdog(); - } -} -#else -static inline void boot_delay_msec(int level) -{ -} -#endif - -#if defined(CONFIG_PRINTK_TIME) -static bool printk_time = 1; -#else -static bool printk_time; -#endif -module_param_named(time, printk_time, bool, S_IRUGO | S_IWUSR); - -static size_t print_time(u64 ts, char *buf) -{ - unsigned long rem_nsec; - - if (!printk_time) - return 0; - - rem_nsec = do_div(ts, 1000000000); - - if (!buf) - return snprintf(NULL, 0, "[%5lu.000000] ", (unsigned long)ts); - - return sprintf(buf, "[%5lu.%06lu] ", - (unsigned long)ts, rem_nsec / 1000); -} - -static size_t print_prefix(const struct log *msg, bool syslog, char *buf) -{ - size_t len = 0; - unsigned int prefix = (msg->facility << 3) | msg->level; - - if (syslog) { - if (buf) { - len += sprintf(buf, "<%u>", prefix); - } else { - len += 3; - if (prefix > 999) - len += 3; - else if (prefix > 99) - len += 2; - else if (prefix > 9) - len++; - } - } - - len += print_time(msg->ts_nsec, buf ? buf + len : NULL); - return len; -} - -static size_t msg_print_text(const struct log *msg, enum log_flags prev, - bool syslog, char *buf, size_t size) -{ - const char *text = log_text(msg); - size_t text_size = msg->text_len; - bool prefix = true; - bool newline = true; - size_t len = 0; - - if ((prev & LOG_CONT) && !(msg->flags & LOG_PREFIX)) - prefix = false; - - if (msg->flags & LOG_CONT) { - if ((prev & LOG_CONT) && !(prev & LOG_NEWLINE)) - prefix = false; - - if (!(msg->flags & LOG_NEWLINE)) - newline = false; - } - - do { - const char *next = memchr(text, '\n', text_size); - size_t text_len; - - if (next) { - text_len = next - text; - next++; - text_size -= next - text; - } else { - text_len = text_size; - } - - if (buf) { - if (print_prefix(msg, syslog, NULL) + - text_len + 1 >= size - len) - break; - - if (prefix) - len += print_prefix(msg, syslog, buf + len); - memcpy(buf + len, text, text_len); - len += text_len; - if (next || newline) - buf[len++] = '\n'; - } else { - /* SYSLOG_ACTION_* buffer size only calculation */ - if (prefix) - len += print_prefix(msg, syslog, NULL); - len += text_len; - if (next || newline) - len++; - } - - prefix = true; - text = next; - } while (text); - - return len; -} - -static int syslog_print(char __user *buf, int size) -{ - char *text; - struct log *msg; - int len = 0; - - text = kmalloc(LOG_LINE_MAX + PREFIX_MAX, GFP_KERNEL); - if (!text) - return -ENOMEM; - - while (size > 0) { - size_t n; - size_t skip; - - raw_spin_lock_irq(&logbuf_lock); - if (syslog_seq < log_first_seq) { - /* messages are gone, move to first one */ - syslog_seq = log_first_seq; - syslog_idx = log_first_idx; - syslog_prev = 0; - syslog_partial = 0; - } - if (syslog_seq == log_next_seq) { - raw_spin_unlock_irq(&logbuf_lock); - break; - } - - skip = syslog_partial; - msg = log_from_idx(syslog_idx); - n = msg_print_text(msg, syslog_prev, true, text, - LOG_LINE_MAX + PREFIX_MAX); - if (n - syslog_partial <= size) { - /* message fits into buffer, move forward */ - syslog_idx = log_next(syslog_idx); - syslog_seq++; - syslog_prev = msg->flags; - n -= syslog_partial; - syslog_partial = 0; - } else if (!len){ - /* partial read(), remember position */ - n = size; - syslog_partial += n; - } else - n = 0; - raw_spin_unlock_irq(&logbuf_lock); - - if (!n) - break; - - if (copy_to_user(buf, text + skip, n)) { - if (!len) - len = -EFAULT; - break; - } - - len += n; - size -= n; - buf += n; - } - - kfree(text); - return len; -} - -static int syslog_print_all(char __user *buf, int size, bool clear) -{ - char *text; - int len = 0; - - text = kmalloc(LOG_LINE_MAX + PREFIX_MAX, GFP_KERNEL); - if (!text) - return -ENOMEM; - - raw_spin_lock_irq(&logbuf_lock); - if (buf) { - u64 next_seq; - u64 seq; - u32 idx; - enum log_flags prev; - - if (clear_seq < log_first_seq) { - /* messages are gone, move to first available one */ - clear_seq = log_first_seq; - clear_idx = log_first_idx; - } - - /* - * Find first record that fits, including all following records, - * into the user-provided buffer for this dump. - */ - seq = clear_seq; - idx = clear_idx; - prev = 0; - while (seq < log_next_seq) { - struct log *msg = log_from_idx(idx); - - len += msg_print_text(msg, prev, true, NULL, 0); - prev = msg->flags; - idx = log_next(idx); - seq++; - } - - /* move first record forward until length fits into the buffer */ - seq = clear_seq; - idx = clear_idx; - prev = 0; - while (len > size && seq < log_next_seq) { - struct log *msg = log_from_idx(idx); - - len -= msg_print_text(msg, prev, true, NULL, 0); - prev = msg->flags; - idx = log_next(idx); - seq++; - } - - /* last message fitting into this dump */ - next_seq = log_next_seq; - - len = 0; - prev = 0; - while (len >= 0 && seq < next_seq) { - struct log *msg = log_from_idx(idx); - int textlen; - - textlen = msg_print_text(msg, prev, true, text, - LOG_LINE_MAX + PREFIX_MAX); - if (textlen < 0) { - len = textlen; - break; - } - idx = log_next(idx); - seq++; - prev = msg->flags; - - raw_spin_unlock_irq(&logbuf_lock); - if (copy_to_user(buf + len, text, textlen)) - len = -EFAULT; - else - len += textlen; - raw_spin_lock_irq(&logbuf_lock); - - if (seq < log_first_seq) { - /* messages are gone, move to next one */ - seq = log_first_seq; - idx = log_first_idx; - prev = 0; - } - } - } - - if (clear) { - clear_seq = log_next_seq; - clear_idx = log_next_idx; - } - raw_spin_unlock_irq(&logbuf_lock); - - kfree(text); - return len; -} - -int do_syslog(int type, char __user *buf, int len, bool from_file) -{ - bool clear = false; - static int saved_console_loglevel = -1; - int error; - - error = check_syslog_permissions(type, from_file); - if (error) - goto out; - - error = security_syslog(type); - if (error) - return error; - - switch (type) { - case SYSLOG_ACTION_CLOSE: /* Close log */ - break; - case SYSLOG_ACTION_OPEN: /* Open log */ - break; - case SYSLOG_ACTION_READ: /* Read from log */ - error = -EINVAL; - if (!buf || len < 0) - goto out; - error = 0; - if (!len) - goto out; - if (!access_ok(VERIFY_WRITE, buf, len)) { - error = -EFAULT; - goto out; - } - error = wait_event_interruptible(log_wait, - syslog_seq != log_next_seq); - if (error) - goto out; - error = syslog_print(buf, len); - break; - /* Read/clear last kernel messages */ - case SYSLOG_ACTION_READ_CLEAR: - clear = true; - /* FALL THRU */ - /* Read last kernel messages */ - case SYSLOG_ACTION_READ_ALL: - error = -EINVAL; - if (!buf || len < 0) - goto out; - error = 0; - if (!len) - goto out; - if (!access_ok(VERIFY_WRITE, buf, len)) { - error = -EFAULT; - goto out; - } - error = syslog_print_all(buf, len, clear); - break; - /* Clear ring buffer */ - case SYSLOG_ACTION_CLEAR: - syslog_print_all(NULL, 0, true); - break; - /* Disable logging to console */ - case SYSLOG_ACTION_CONSOLE_OFF: - if (saved_console_loglevel == -1) - saved_console_loglevel = console_loglevel; - console_loglevel = minimum_console_loglevel; - break; - /* Enable logging to console */ - case SYSLOG_ACTION_CONSOLE_ON: - if (saved_console_loglevel != -1) { - console_loglevel = saved_console_loglevel; - saved_console_loglevel = -1; - } - break; - /* Set level of messages printed to console */ - case SYSLOG_ACTION_CONSOLE_LEVEL: - error = -EINVAL; - if (len < 1 || len > 8) - goto out; - if (len < minimum_console_loglevel) - len = minimum_console_loglevel; - console_loglevel = len; - /* Implicitly re-enable logging to console */ - saved_console_loglevel = -1; - error = 0; - break; - /* Number of chars in the log buffer */ - case SYSLOG_ACTION_SIZE_UNREAD: - raw_spin_lock_irq(&logbuf_lock); - if (syslog_seq < log_first_seq) { - /* messages are gone, move to first one */ - syslog_seq = log_first_seq; - syslog_idx = log_first_idx; - syslog_prev = 0; - syslog_partial = 0; - } - if (from_file) { - /* - * Short-cut for poll(/"proc/kmsg") which simply checks - * for pending data, not the size; return the count of - * records, not the length. - */ - error = log_next_idx - syslog_idx; - } else { - u64 seq = syslog_seq; - u32 idx = syslog_idx; - enum log_flags prev = syslog_prev; - - error = 0; - while (seq < log_next_seq) { - struct log *msg = log_from_idx(idx); - - error += msg_print_text(msg, prev, true, NULL, 0); - idx = log_next(idx); - seq++; - prev = msg->flags; - } - error -= syslog_partial; - } - raw_spin_unlock_irq(&logbuf_lock); - break; - /* Size of the log buffer */ - case SYSLOG_ACTION_SIZE_BUFFER: - error = log_buf_len; - break; - default: - error = -EINVAL; - break; - } -out: - return error; -} - -SYSCALL_DEFINE3(syslog, int, type, char __user *, buf, int, len) -{ - return do_syslog(type, buf, len, SYSLOG_FROM_READER); -} - -/* - * Call the console drivers, asking them to write out - * log_buf[start] to log_buf[end - 1]. - * The console_lock must be held. - */ -static void call_console_drivers(int level, const char *text, size_t len) -{ - struct console *con; - - trace_console(text, len); - - if (level >= console_loglevel && !ignore_loglevel) - return; - if (!console_drivers) - return; - - for_each_console(con) { - if (exclusive_console && con != exclusive_console) - continue; - if (!(con->flags & CON_ENABLED)) - continue; - if (!con->write) - continue; - if (!cpu_online(smp_processor_id()) && - !(con->flags & CON_ANYTIME)) - continue; - con->write(con, text, len); - } -} - -/* - * Zap console related locks when oopsing. Only zap at most once - * every 10 seconds, to leave time for slow consoles to print a - * full oops. - */ -static void zap_locks(void) -{ - static unsigned long oops_timestamp; - - if (time_after_eq(jiffies, oops_timestamp) && - !time_after(jiffies, oops_timestamp + 30 * HZ)) - return; - - oops_timestamp = jiffies; - - debug_locks_off(); - /* If a crash is occurring, make sure we can't deadlock */ - raw_spin_lock_init(&logbuf_lock); - /* And make sure that we print immediately */ - sema_init(&console_sem, 1); -} - -/* Check if we have any console registered that can be called early in boot. */ -static int have_callable_console(void) -{ - struct console *con; - - for_each_console(con) - if (con->flags & CON_ANYTIME) - return 1; - - return 0; -} - -/* - * Can we actually use the console at this time on this cpu? - * - * Console drivers may assume that per-cpu resources have - * been allocated. So unless they're explicitly marked as - * being able to cope (CON_ANYTIME) don't call them until - * this CPU is officially up. - */ -static inline int can_use_console(unsigned int cpu) -{ - return cpu_online(cpu) || have_callable_console(); -} - -/* - * Try to get console ownership to actually show the kernel - * messages from a 'printk'. Return true (and with the - * console_lock held, and 'console_locked' set) if it - * is successful, false otherwise. - * - * This gets called with the 'logbuf_lock' spinlock held and - * interrupts disabled. It should return with 'lockbuf_lock' - * released but interrupts still disabled. - */ -static int console_trylock_for_printk(unsigned int cpu) - __releases(&logbuf_lock) -{ - int retval = 0, wake = 0; - - if (console_trylock()) { - retval = 1; - - /* - * If we can't use the console, we need to release - * the console semaphore by hand to avoid flushing - * the buffer. We need to hold the console semaphore - * in order to do this test safely. - */ - if (!can_use_console(cpu)) { - console_locked = 0; - wake = 1; - retval = 0; - } - } - logbuf_cpu = UINT_MAX; - raw_spin_unlock(&logbuf_lock); - if (wake) - up(&console_sem); - return retval; -} - -int printk_delay_msec __read_mostly; - -static inline void printk_delay(void) -{ - if (unlikely(printk_delay_msec)) { - int m = printk_delay_msec; - - while (m--) { - mdelay(1); - touch_nmi_watchdog(); - } - } -} - -/* - * Continuation lines are buffered, and not committed to the record buffer - * until the line is complete, or a race forces it. The line fragments - * though, are printed immediately to the consoles to ensure everything has - * reached the console in case of a kernel crash. - */ -static struct cont { - char buf[LOG_LINE_MAX]; - size_t len; /* length == 0 means unused buffer */ - size_t cons; /* bytes written to console */ - struct task_struct *owner; /* task of first print*/ - u64 ts_nsec; /* time of first print */ - u8 level; /* log level of first message */ - u8 facility; /* log level of first message */ - enum log_flags flags; /* prefix, newline flags */ - bool flushed:1; /* buffer sealed and committed */ -} cont; - -static void cont_flush(enum log_flags flags) -{ - if (cont.flushed) - return; - if (cont.len == 0) - return; - - if (cont.cons) { - /* - * If a fragment of this line was directly flushed to the - * console; wait for the console to pick up the rest of the - * line. LOG_NOCONS suppresses a duplicated output. - */ - log_store(cont.facility, cont.level, flags | LOG_NOCONS, - cont.ts_nsec, NULL, 0, cont.buf, cont.len); - cont.flags = flags; - cont.flushed = true; - } else { - /* - * If no fragment of this line ever reached the console, - * just submit it to the store and free the buffer. - */ - log_store(cont.facility, cont.level, flags, 0, - NULL, 0, cont.buf, cont.len); - cont.len = 0; - } -} - -static bool cont_add(int facility, int level, const char *text, size_t len) -{ - if (cont.len && cont.flushed) - return false; - - if (cont.len + len > sizeof(cont.buf)) { - /* the line gets too long, split it up in separate records */ - cont_flush(LOG_CONT); - return false; - } - - if (!cont.len) { - cont.facility = facility; - cont.level = level; - cont.owner = current; - cont.ts_nsec = local_clock(); - cont.flags = 0; - cont.cons = 0; - cont.flushed = false; - } - - memcpy(cont.buf + cont.len, text, len); - cont.len += len; - - if (cont.len > (sizeof(cont.buf) * 80) / 100) - cont_flush(LOG_CONT); - - return true; -} - -static size_t cont_print_text(char *text, size_t size) -{ - size_t textlen = 0; - size_t len; - - if (cont.cons == 0 && (console_prev & LOG_NEWLINE)) { - textlen += print_time(cont.ts_nsec, text); - size -= textlen; - } - - len = cont.len - cont.cons; - if (len > 0) { - if (len+1 > size) - len = size-1; - memcpy(text + textlen, cont.buf + cont.cons, len); - textlen += len; - cont.cons = cont.len; - } - - if (cont.flushed) { - if (cont.flags & LOG_NEWLINE) - text[textlen++] = '\n'; - /* got everything, release buffer */ - cont.len = 0; - } - return textlen; -} - -asmlinkage int vprintk_emit(int facility, int level, - const char *dict, size_t dictlen, - const char *fmt, va_list args) -{ - static int recursion_bug; - static char textbuf[LOG_LINE_MAX]; - char *text = textbuf; - size_t text_len; - enum log_flags lflags = 0; - unsigned long flags; - int this_cpu; - int printed_len = 0; - - boot_delay_msec(level); - printk_delay(); - - /* This stops the holder of console_sem just where we want him */ - local_irq_save(flags); - this_cpu = smp_processor_id(); - - /* - * Ouch, printk recursed into itself! - */ - if (unlikely(logbuf_cpu == this_cpu)) { - /* - * If a crash is occurring during printk() on this CPU, - * then try to get the crash message out but make sure - * we can't deadlock. Otherwise just return to avoid the - * recursion and return - but flag the recursion so that - * it can be printed at the next appropriate moment: - */ - if (!oops_in_progress && !lockdep_recursing(current)) { - recursion_bug = 1; - goto out_restore_irqs; - } - zap_locks(); - } - - lockdep_off(); - raw_spin_lock(&logbuf_lock); - logbuf_cpu = this_cpu; - - if (recursion_bug) { - static const char recursion_msg[] = - "BUG: recent printk recursion!"; - - recursion_bug = 0; - printed_len += strlen(recursion_msg); - /* emit KERN_CRIT message */ - log_store(0, 2, LOG_PREFIX|LOG_NEWLINE, 0, - NULL, 0, recursion_msg, printed_len); - } - - /* - * The printf needs to come first; we need the syslog - * prefix which might be passed-in as a parameter. - */ - text_len = vscnprintf(text, sizeof(textbuf), fmt, args); - - /* mark and strip a trailing newline */ - if (text_len && text[text_len-1] == '\n') { - text_len--; - lflags |= LOG_NEWLINE; - } - - /* strip kernel syslog prefix and extract log level or control flags */ - if (facility == 0) { - int kern_level = printk_get_level(text); - - if (kern_level) { - const char *end_of_header = printk_skip_level(text); - switch (kern_level) { - case '0' ... '7': - if (level == -1) - level = kern_level - '0'; - case 'd': /* KERN_DEFAULT */ - lflags |= LOG_PREFIX; - case 'c': /* KERN_CONT */ - break; - } - text_len -= end_of_header - text; - text = (char *)end_of_header; - } - } - - if (level == -1) - level = default_message_loglevel; - - if (dict) - lflags |= LOG_PREFIX|LOG_NEWLINE; - - if (!(lflags & LOG_NEWLINE)) { - /* - * Flush the conflicting buffer. An earlier newline was missing, - * or another task also prints continuation lines. - */ - if (cont.len && (lflags & LOG_PREFIX || cont.owner != current)) - cont_flush(LOG_NEWLINE); - - /* buffer line if possible, otherwise store it right away */ - if (!cont_add(facility, level, text, text_len)) - log_store(facility, level, lflags | LOG_CONT, 0, - dict, dictlen, text, text_len); - } else { - bool stored = false; - - /* - * If an earlier newline was missing and it was the same task, - * either merge it with the current buffer and flush, or if - * there was a race with interrupts (prefix == true) then just - * flush it out and store this line separately. - */ - if (cont.len && cont.owner == current) { - if (!(lflags & LOG_PREFIX)) - stored = cont_add(facility, level, text, text_len); - cont_flush(LOG_NEWLINE); - } - - if (!stored) - log_store(facility, level, lflags, 0, - dict, dictlen, text, text_len); - } - printed_len += text_len; - - /* - * Try to acquire and then immediately release the console semaphore. - * The release will print out buffers and wake up /dev/kmsg and syslog() - * users. - * - * The console_trylock_for_printk() function will release 'logbuf_lock' - * regardless of whether it actually gets the console semaphore or not. - */ - if (console_trylock_for_printk(this_cpu)) - console_unlock(); - - lockdep_on(); -out_restore_irqs: - local_irq_restore(flags); - - return printed_len; -} -EXPORT_SYMBOL(vprintk_emit); - -asmlinkage int vprintk(const char *fmt, va_list args) -{ - return vprintk_emit(0, -1, NULL, 0, fmt, args); -} -EXPORT_SYMBOL(vprintk); - -asmlinkage int printk_emit(int facility, int level, - const char *dict, size_t dictlen, - const char *fmt, ...) -{ - va_list args; - int r; - - va_start(args, fmt); - r = vprintk_emit(facility, level, dict, dictlen, fmt, args); - va_end(args); - - return r; -} -EXPORT_SYMBOL(printk_emit); - -/** - * printk - print a kernel message - * @fmt: format string - * - * This is printk(). It can be called from any context. We want it to work. - * - * We try to grab the console_lock. If we succeed, it's easy - we log the - * output and call the console drivers. If we fail to get the semaphore, we - * place the output into the log buffer and return. The current holder of - * the console_sem will notice the new output in console_unlock(); and will - * send it to the consoles before releasing the lock. - * - * One effect of this deferred printing is that code which calls printk() and - * then changes console_loglevel may break. This is because console_loglevel - * is inspected when the actual printing occurs. - * - * See also: - * printf(3) - * - * See the vsnprintf() documentation for format string extensions over C99. - */ -asmlinkage int printk(const char *fmt, ...) -{ - va_list args; - int r; - -#ifdef CONFIG_KGDB_KDB - if (unlikely(kdb_trap_printk)) { - va_start(args, fmt); - r = vkdb_printf(fmt, args); - va_end(args); - return r; - } -#endif - va_start(args, fmt); - r = vprintk_emit(0, -1, NULL, 0, fmt, args); - va_end(args); - - return r; -} -EXPORT_SYMBOL(printk); - -#else /* CONFIG_PRINTK */ - -#define LOG_LINE_MAX 0 -#define PREFIX_MAX 0 -#define LOG_LINE_MAX 0 -static u64 syslog_seq; -static u32 syslog_idx; -static u64 console_seq; -static u32 console_idx; -static enum log_flags syslog_prev; -static u64 log_first_seq; -static u32 log_first_idx; -static u64 log_next_seq; -static enum log_flags console_prev; -static struct cont { - size_t len; - size_t cons; - u8 level; - bool flushed:1; -} cont; -static struct log *log_from_idx(u32 idx) { return NULL; } -static u32 log_next(u32 idx) { return 0; } -static void call_console_drivers(int level, const char *text, size_t len) {} -static size_t msg_print_text(const struct log *msg, enum log_flags prev, - bool syslog, char *buf, size_t size) { return 0; } -static size_t cont_print_text(char *text, size_t size) { return 0; } - -#endif /* CONFIG_PRINTK */ - -#ifdef CONFIG_EARLY_PRINTK -struct console *early_console; - -void early_vprintk(const char *fmt, va_list ap) -{ - if (early_console) { - char buf[512]; - int n = vscnprintf(buf, sizeof(buf), fmt, ap); - - early_console->write(early_console, buf, n); - } -} - -asmlinkage void early_printk(const char *fmt, ...) -{ - va_list ap; - - va_start(ap, fmt); - early_vprintk(fmt, ap); - va_end(ap); -} -#endif - -static int __add_preferred_console(char *name, int idx, char *options, - char *brl_options) -{ - struct console_cmdline *c; - int i; - - /* - * See if this tty is not yet registered, and - * if we have a slot free. - */ - for (i = 0; i < MAX_CMDLINECONSOLES && console_cmdline[i].name[0]; i++) - if (strcmp(console_cmdline[i].name, name) == 0 && - console_cmdline[i].index == idx) { - if (!brl_options) - selected_console = i; - return 0; - } - if (i == MAX_CMDLINECONSOLES) - return -E2BIG; - if (!brl_options) - selected_console = i; - c = &console_cmdline[i]; - strlcpy(c->name, name, sizeof(c->name)); - c->options = options; -#ifdef CONFIG_A11Y_BRAILLE_CONSOLE - c->brl_options = brl_options; -#endif - c->index = idx; - return 0; -} -/* - * Set up a list of consoles. Called from init/main.c - */ -static int __init console_setup(char *str) -{ - char buf[sizeof(console_cmdline[0].name) + 4]; /* 4 for index */ - char *s, *options, *brl_options = NULL; - int idx; - -#ifdef CONFIG_A11Y_BRAILLE_CONSOLE - if (!memcmp(str, "brl,", 4)) { - brl_options = ""; - str += 4; - } else if (!memcmp(str, "brl=", 4)) { - brl_options = str + 4; - str = strchr(brl_options, ','); - if (!str) { - printk(KERN_ERR "need port name after brl=\n"); - return 1; - } - *(str++) = 0; - } -#endif - - /* - * Decode str into name, index, options. - */ - if (str[0] >= '0' && str[0] <= '9') { - strcpy(buf, "ttyS"); - strncpy(buf + 4, str, sizeof(buf) - 5); - } else { - strncpy(buf, str, sizeof(buf) - 1); - } - buf[sizeof(buf) - 1] = 0; - if ((options = strchr(str, ',')) != NULL) - *(options++) = 0; -#ifdef __sparc__ - if (!strcmp(str, "ttya")) - strcpy(buf, "ttyS0"); - if (!strcmp(str, "ttyb")) - strcpy(buf, "ttyS1"); -#endif - for (s = buf; *s; s++) - if ((*s >= '0' && *s <= '9') || *s == ',') - break; - idx = simple_strtoul(s, NULL, 10); - *s = 0; - - __add_preferred_console(buf, idx, options, brl_options); - console_set_on_cmdline = 1; - return 1; -} -__setup("console=", console_setup); - -/** - * add_preferred_console - add a device to the list of preferred consoles. - * @name: device name - * @idx: device index - * @options: options for this console - * - * The last preferred console added will be used for kernel messages - * and stdin/out/err for init. Normally this is used by console_setup - * above to handle user-supplied console arguments; however it can also - * be used by arch-specific code either to override the user or more - * commonly to provide a default console (ie from PROM variables) when - * the user has not supplied one. - */ -int add_preferred_console(char *name, int idx, char *options) -{ - return __add_preferred_console(name, idx, options, NULL); -} - -int update_console_cmdline(char *name, int idx, char *name_new, int idx_new, char *options) -{ - struct console_cmdline *c; - int i; - - for (i = 0; i < MAX_CMDLINECONSOLES && console_cmdline[i].name[0]; i++) - if (strcmp(console_cmdline[i].name, name) == 0 && - console_cmdline[i].index == idx) { - c = &console_cmdline[i]; - strlcpy(c->name, name_new, sizeof(c->name)); - c->name[sizeof(c->name) - 1] = 0; - c->options = options; - c->index = idx_new; - return i; - } - /* not found */ - return -1; -} - -bool console_suspend_enabled = 1; -EXPORT_SYMBOL(console_suspend_enabled); - -static int __init console_suspend_disable(char *str) -{ - console_suspend_enabled = 0; - return 1; -} -__setup("no_console_suspend", console_suspend_disable); -module_param_named(console_suspend, console_suspend_enabled, - bool, S_IRUGO | S_IWUSR); -MODULE_PARM_DESC(console_suspend, "suspend console during suspend" - " and hibernate operations"); - -/** - * suspend_console - suspend the console subsystem - * - * This disables printk() while we go into suspend states - */ -void suspend_console(void) -{ - if (!console_suspend_enabled) - return; - printk("Suspending console(s) (use no_console_suspend to debug)\n"); - console_lock(); - console_suspended = 1; - up(&console_sem); -} - -void resume_console(void) -{ - if (!console_suspend_enabled) - return; - down(&console_sem); - console_suspended = 0; - console_unlock(); -} - -/** - * console_cpu_notify - print deferred console messages after CPU hotplug - * @self: notifier struct - * @action: CPU hotplug event - * @hcpu: unused - * - * If printk() is called from a CPU that is not online yet, the messages - * will be spooled but will not show up on the console. This function is - * called when a new CPU comes online (or fails to come up), and ensures - * that any such output gets printed. - */ -static int console_cpu_notify(struct notifier_block *self, - unsigned long action, void *hcpu) -{ - switch (action) { - case CPU_ONLINE: - case CPU_DEAD: - case CPU_DOWN_FAILED: - case CPU_UP_CANCELED: - console_lock(); - console_unlock(); - } - return NOTIFY_OK; -} - -/** - * console_lock - lock the console system for exclusive use. - * - * Acquires a lock which guarantees that the caller has - * exclusive access to the console system and the console_drivers list. - * - * Can sleep, returns nothing. - */ -void console_lock(void) -{ - might_sleep(); - - down(&console_sem); - if (console_suspended) - return; - console_locked = 1; - console_may_schedule = 1; - mutex_acquire(&console_lock_dep_map, 0, 0, _RET_IP_); -} -EXPORT_SYMBOL(console_lock); - -/** - * console_trylock - try to lock the console system for exclusive use. - * - * Tried to acquire a lock which guarantees that the caller has - * exclusive access to the console system and the console_drivers list. - * - * returns 1 on success, and 0 on failure to acquire the lock. - */ -int console_trylock(void) -{ - if (down_trylock(&console_sem)) - return 0; - if (console_suspended) { - up(&console_sem); - return 0; - } - console_locked = 1; - console_may_schedule = 0; - mutex_acquire(&console_lock_dep_map, 0, 1, _RET_IP_); - return 1; -} -EXPORT_SYMBOL(console_trylock); - -int is_console_locked(void) -{ - return console_locked; -} - -static void console_cont_flush(char *text, size_t size) -{ - unsigned long flags; - size_t len; - - raw_spin_lock_irqsave(&logbuf_lock, flags); - - if (!cont.len) - goto out; - - /* - * We still queue earlier records, likely because the console was - * busy. The earlier ones need to be printed before this one, we - * did not flush any fragment so far, so just let it queue up. - */ - if (console_seq < log_next_seq && !cont.cons) - goto out; - - len = cont_print_text(text, size); - raw_spin_unlock(&logbuf_lock); - stop_critical_timings(); - call_console_drivers(cont.level, text, len); - start_critical_timings(); - local_irq_restore(flags); - return; -out: - raw_spin_unlock_irqrestore(&logbuf_lock, flags); -} - -/** - * console_unlock - unlock the console system - * - * Releases the console_lock which the caller holds on the console system - * and the console driver list. - * - * While the console_lock was held, console output may have been buffered - * by printk(). If this is the case, console_unlock(); emits - * the output prior to releasing the lock. - * - * If there is output waiting, we wake /dev/kmsg and syslog() users. - * - * console_unlock(); may be called from any context. - */ -void console_unlock(void) -{ - static char text[LOG_LINE_MAX + PREFIX_MAX]; - static u64 seen_seq; - unsigned long flags; - bool wake_klogd = false; - bool retry; - - if (console_suspended) { - up(&console_sem); - return; - } - - console_may_schedule = 0; - - /* flush buffered message fragment immediately to console */ - console_cont_flush(text, sizeof(text)); -again: - for (;;) { - struct log *msg; - size_t len; - int level; - - raw_spin_lock_irqsave(&logbuf_lock, flags); - if (seen_seq != log_next_seq) { - wake_klogd = true; - seen_seq = log_next_seq; - } - - if (console_seq < log_first_seq) { - /* messages are gone, move to first one */ - console_seq = log_first_seq; - console_idx = log_first_idx; - console_prev = 0; - } -skip: - if (console_seq == log_next_seq) - break; - - msg = log_from_idx(console_idx); - if (msg->flags & LOG_NOCONS) { - /* - * Skip record we have buffered and already printed - * directly to the console when we received it. - */ - console_idx = log_next(console_idx); - console_seq++; - /* - * We will get here again when we register a new - * CON_PRINTBUFFER console. Clear the flag so we - * will properly dump everything later. - */ - msg->flags &= ~LOG_NOCONS; - console_prev = msg->flags; - goto skip; - } - - level = msg->level; - len = msg_print_text(msg, console_prev, false, - text, sizeof(text)); - console_idx = log_next(console_idx); - console_seq++; - console_prev = msg->flags; - raw_spin_unlock(&logbuf_lock); - - stop_critical_timings(); /* don't trace print latency */ - call_console_drivers(level, text, len); - start_critical_timings(); - local_irq_restore(flags); - } - console_locked = 0; - mutex_release(&console_lock_dep_map, 1, _RET_IP_); - - /* Release the exclusive_console once it is used */ - if (unlikely(exclusive_console)) - exclusive_console = NULL; - - raw_spin_unlock(&logbuf_lock); - - up(&console_sem); - - /* - * Someone could have filled up the buffer again, so re-check if there's - * something to flush. In case we cannot trylock the console_sem again, - * there's a new owner and the console_unlock() from them will do the - * flush, no worries. - */ - raw_spin_lock(&logbuf_lock); - retry = console_seq != log_next_seq; - raw_spin_unlock_irqrestore(&logbuf_lock, flags); - - if (retry && console_trylock()) - goto again; - - if (wake_klogd) - wake_up_klogd(); -} -EXPORT_SYMBOL(console_unlock); - -/** - * console_conditional_schedule - yield the CPU if required - * - * If the console code is currently allowed to sleep, and - * if this CPU should yield the CPU to another task, do - * so here. - * - * Must be called within console_lock();. - */ -void __sched console_conditional_schedule(void) -{ - if (console_may_schedule) - cond_resched(); -} -EXPORT_SYMBOL(console_conditional_schedule); - -void console_unblank(void) -{ - struct console *c; - - /* - * console_unblank can no longer be called in interrupt context unless - * oops_in_progress is set to 1.. - */ - if (oops_in_progress) { - if (down_trylock(&console_sem) != 0) - return; - } else - console_lock(); - - console_locked = 1; - console_may_schedule = 0; - for_each_console(c) - if ((c->flags & CON_ENABLED) && c->unblank) - c->unblank(); - console_unlock(); -} - -/* - * Return the console tty driver structure and its associated index - */ -struct tty_driver *console_device(int *index) -{ - struct console *c; - struct tty_driver *driver = NULL; - - console_lock(); - for_each_console(c) { - if (!c->device) - continue; - driver = c->device(c, index); - if (driver) - break; - } - console_unlock(); - return driver; -} - -/* - * Prevent further output on the passed console device so that (for example) - * serial drivers can disable console output before suspending a port, and can - * re-enable output afterwards. - */ -void console_stop(struct console *console) -{ - console_lock(); - console->flags &= ~CON_ENABLED; - console_unlock(); -} -EXPORT_SYMBOL(console_stop); - -void console_start(struct console *console) -{ - console_lock(); - console->flags |= CON_ENABLED; - console_unlock(); -} -EXPORT_SYMBOL(console_start); - -static int __read_mostly keep_bootcon; - -static int __init keep_bootcon_setup(char *str) -{ - keep_bootcon = 1; - printk(KERN_INFO "debug: skip boot console de-registration.\n"); - - return 0; -} - -early_param("keep_bootcon", keep_bootcon_setup); - -/* - * The console driver calls this routine during kernel initialization - * to register the console printing procedure with printk() and to - * print any messages that were printed by the kernel before the - * console driver was initialized. - * - * This can happen pretty early during the boot process (because of - * early_printk) - sometimes before setup_arch() completes - be careful - * of what kernel features are used - they may not be initialised yet. - * - * There are two types of consoles - bootconsoles (early_printk) and - * "real" consoles (everything which is not a bootconsole) which are - * handled differently. - * - Any number of bootconsoles can be registered at any time. - * - As soon as a "real" console is registered, all bootconsoles - * will be unregistered automatically. - * - Once a "real" console is registered, any attempt to register a - * bootconsoles will be rejected - */ -void register_console(struct console *newcon) -{ - int i; - unsigned long flags; - struct console *bcon = NULL; - - /* - * before we register a new CON_BOOT console, make sure we don't - * already have a valid console - */ - if (console_drivers && newcon->flags & CON_BOOT) { - /* find the last or real console */ - for_each_console(bcon) { - if (!(bcon->flags & CON_BOOT)) { - printk(KERN_INFO "Too late to register bootconsole %s%d\n", - newcon->name, newcon->index); - return; - } - } - } - - if (console_drivers && console_drivers->flags & CON_BOOT) - bcon = console_drivers; - - if (preferred_console < 0 || bcon || !console_drivers) - preferred_console = selected_console; - - if (newcon->early_setup) - newcon->early_setup(); - - /* - * See if we want to use this console driver. If we - * didn't select a console we take the first one - * that registers here. - */ - if (preferred_console < 0) { - if (newcon->index < 0) - newcon->index = 0; - if (newcon->setup == NULL || - newcon->setup(newcon, NULL) == 0) { - newcon->flags |= CON_ENABLED; - if (newcon->device) { - newcon->flags |= CON_CONSDEV; - preferred_console = 0; - } - } - } - - /* - * See if this console matches one we selected on - * the command line. - */ - for (i = 0; i < MAX_CMDLINECONSOLES && console_cmdline[i].name[0]; - i++) { - if (strcmp(console_cmdline[i].name, newcon->name) != 0) - continue; - if (newcon->index >= 0 && - newcon->index != console_cmdline[i].index) - continue; - if (newcon->index < 0) - newcon->index = console_cmdline[i].index; -#ifdef CONFIG_A11Y_BRAILLE_CONSOLE - if (console_cmdline[i].brl_options) { - newcon->flags |= CON_BRL; - braille_register_console(newcon, - console_cmdline[i].index, - console_cmdline[i].options, - console_cmdline[i].brl_options); - return; - } -#endif - if (newcon->setup && - newcon->setup(newcon, console_cmdline[i].options) != 0) - break; - newcon->flags |= CON_ENABLED; - newcon->index = console_cmdline[i].index; - if (i == selected_console) { - newcon->flags |= CON_CONSDEV; - preferred_console = selected_console; - } - break; - } - - if (!(newcon->flags & CON_ENABLED)) - return; - - /* - * If we have a bootconsole, and are switching to a real console, - * don't print everything out again, since when the boot console, and - * the real console are the same physical device, it's annoying to - * see the beginning boot messages twice - */ - if (bcon && ((newcon->flags & (CON_CONSDEV | CON_BOOT)) == CON_CONSDEV)) - newcon->flags &= ~CON_PRINTBUFFER; - - /* - * Put this console in the list - keep the - * preferred driver at the head of the list. - */ - console_lock(); - if ((newcon->flags & CON_CONSDEV) || console_drivers == NULL) { - newcon->next = console_drivers; - console_drivers = newcon; - if (newcon->next) - newcon->next->flags &= ~CON_CONSDEV; - } else { - newcon->next = console_drivers->next; - console_drivers->next = newcon; - } - if (newcon->flags & CON_PRINTBUFFER) { - /* - * console_unlock(); will print out the buffered messages - * for us. - */ - raw_spin_lock_irqsave(&logbuf_lock, flags); - console_seq = syslog_seq; - console_idx = syslog_idx; - console_prev = syslog_prev; - raw_spin_unlock_irqrestore(&logbuf_lock, flags); - /* - * We're about to replay the log buffer. Only do this to the - * just-registered console to avoid excessive message spam to - * the already-registered consoles. - */ - exclusive_console = newcon; - } - console_unlock(); - console_sysfs_notify(); - - /* - * By unregistering the bootconsoles after we enable the real console - * we get the "console xxx enabled" message on all the consoles - - * boot consoles, real consoles, etc - this is to ensure that end - * users know there might be something in the kernel's log buffer that - * went to the bootconsole (that they do not see on the real console) - */ - if (bcon && - ((newcon->flags & (CON_CONSDEV | CON_BOOT)) == CON_CONSDEV) && - !keep_bootcon) { - /* we need to iterate through twice, to make sure we print - * everything out, before we unregister the console(s) - */ - printk(KERN_INFO "console [%s%d] enabled, bootconsole disabled\n", - newcon->name, newcon->index); - for_each_console(bcon) - if (bcon->flags & CON_BOOT) - unregister_console(bcon); - } else { - printk(KERN_INFO "%sconsole [%s%d] enabled\n", - (newcon->flags & CON_BOOT) ? "boot" : "" , - newcon->name, newcon->index); - } -} -EXPORT_SYMBOL(register_console); - -int unregister_console(struct console *console) -{ - struct console *a, *b; - int res = 1; - -#ifdef CONFIG_A11Y_BRAILLE_CONSOLE - if (console->flags & CON_BRL) - return braille_unregister_console(console); -#endif - - console_lock(); - if (console_drivers == console) { - console_drivers=console->next; - res = 0; - } else if (console_drivers) { - for (a=console_drivers->next, b=console_drivers ; - a; b=a, a=b->next) { - if (a == console) { - b->next = a->next; - res = 0; - break; - } - } - } - - /* - * If this isn't the last console and it has CON_CONSDEV set, we - * need to set it on the next preferred console. - */ - if (console_drivers != NULL && console->flags & CON_CONSDEV) - console_drivers->flags |= CON_CONSDEV; - - console_unlock(); - console_sysfs_notify(); - return res; -} -EXPORT_SYMBOL(unregister_console); - -static int __init printk_late_init(void) -{ - struct console *con; - - for_each_console(con) { - if (!keep_bootcon && con->flags & CON_BOOT) { - printk(KERN_INFO "turn off boot console %s%d\n", - con->name, con->index); - unregister_console(con); - } - } - hotcpu_notifier(console_cpu_notify, 0); - return 0; -} -late_initcall(printk_late_init); - -#if defined CONFIG_PRINTK -/* - * Delayed printk version, for scheduler-internal messages: - */ -#define PRINTK_BUF_SIZE 512 - -#define PRINTK_PENDING_WAKEUP 0x01 -#define PRINTK_PENDING_SCHED 0x02 - -static DEFINE_PER_CPU(int, printk_pending); -static DEFINE_PER_CPU(char [PRINTK_BUF_SIZE], printk_sched_buf); - -static void wake_up_klogd_work_func(struct irq_work *irq_work) -{ - int pending = __this_cpu_xchg(printk_pending, 0); - - if (pending & PRINTK_PENDING_SCHED) { - char *buf = __get_cpu_var(printk_sched_buf); - printk(KERN_WARNING "[sched_delayed] %s", buf); - } - - if (pending & PRINTK_PENDING_WAKEUP) - wake_up_interruptible(&log_wait); -} - -static DEFINE_PER_CPU(struct irq_work, wake_up_klogd_work) = { - .func = wake_up_klogd_work_func, - .flags = IRQ_WORK_LAZY, -}; - -void wake_up_klogd(void) -{ - preempt_disable(); - if (waitqueue_active(&log_wait)) { - this_cpu_or(printk_pending, PRINTK_PENDING_WAKEUP); - irq_work_queue(&__get_cpu_var(wake_up_klogd_work)); - } - preempt_enable(); -} - -int printk_sched(const char *fmt, ...) -{ - unsigned long flags; - va_list args; - char *buf; - int r; - - local_irq_save(flags); - buf = __get_cpu_var(printk_sched_buf); - - va_start(args, fmt); - r = vsnprintf(buf, PRINTK_BUF_SIZE, fmt, args); - va_end(args); - - __this_cpu_or(printk_pending, PRINTK_PENDING_SCHED); - irq_work_queue(&__get_cpu_var(wake_up_klogd_work)); - local_irq_restore(flags); - - return r; -} - -/* - * printk rate limiting, lifted from the networking subsystem. - * - * This enforces a rate limit: not more than 10 kernel messages - * every 5s to make a denial-of-service attack impossible. - */ -DEFINE_RATELIMIT_STATE(printk_ratelimit_state, 5 * HZ, 10); - -int __printk_ratelimit(const char *func) -{ - return ___ratelimit(&printk_ratelimit_state, func); -} -EXPORT_SYMBOL(__printk_ratelimit); - -/** - * printk_timed_ratelimit - caller-controlled printk ratelimiting - * @caller_jiffies: pointer to caller's state - * @interval_msecs: minimum interval between prints - * - * printk_timed_ratelimit() returns true if more than @interval_msecs - * milliseconds have elapsed since the last time printk_timed_ratelimit() - * returned true. - */ -bool printk_timed_ratelimit(unsigned long *caller_jiffies, - unsigned int interval_msecs) -{ - if (*caller_jiffies == 0 - || !time_in_range(jiffies, *caller_jiffies, - *caller_jiffies - + msecs_to_jiffies(interval_msecs))) { - *caller_jiffies = jiffies; - return true; - } - return false; -} -EXPORT_SYMBOL(printk_timed_ratelimit); - -static DEFINE_SPINLOCK(dump_list_lock); -static LIST_HEAD(dump_list); - -/** - * kmsg_dump_register - register a kernel log dumper. - * @dumper: pointer to the kmsg_dumper structure - * - * Adds a kernel log dumper to the system. The dump callback in the - * structure will be called when the kernel oopses or panics and must be - * set. Returns zero on success and %-EINVAL or %-EBUSY otherwise. - */ -int kmsg_dump_register(struct kmsg_dumper *dumper) -{ - unsigned long flags; - int err = -EBUSY; - - /* The dump callback needs to be set */ - if (!dumper->dump) - return -EINVAL; - - spin_lock_irqsave(&dump_list_lock, flags); - /* Don't allow registering multiple times */ - if (!dumper->registered) { - dumper->registered = 1; - list_add_tail_rcu(&dumper->list, &dump_list); - err = 0; - } - spin_unlock_irqrestore(&dump_list_lock, flags); - - return err; -} -EXPORT_SYMBOL_GPL(kmsg_dump_register); - -/** - * kmsg_dump_unregister - unregister a kmsg dumper. - * @dumper: pointer to the kmsg_dumper structure - * - * Removes a dump device from the system. Returns zero on success and - * %-EINVAL otherwise. - */ -int kmsg_dump_unregister(struct kmsg_dumper *dumper) -{ - unsigned long flags; - int err = -EINVAL; - - spin_lock_irqsave(&dump_list_lock, flags); - if (dumper->registered) { - dumper->registered = 0; - list_del_rcu(&dumper->list); - err = 0; - } - spin_unlock_irqrestore(&dump_list_lock, flags); - synchronize_rcu(); - - return err; -} -EXPORT_SYMBOL_GPL(kmsg_dump_unregister); - -static bool always_kmsg_dump; -module_param_named(always_kmsg_dump, always_kmsg_dump, bool, S_IRUGO | S_IWUSR); - -/** - * kmsg_dump - dump kernel log to kernel message dumpers. - * @reason: the reason (oops, panic etc) for dumping - * - * Call each of the registered dumper's dump() callback, which can - * retrieve the kmsg records with kmsg_dump_get_line() or - * kmsg_dump_get_buffer(). - */ -void kmsg_dump(enum kmsg_dump_reason reason) -{ - struct kmsg_dumper *dumper; - unsigned long flags; - - if ((reason > KMSG_DUMP_OOPS) && !always_kmsg_dump) - return; - - rcu_read_lock(); - list_for_each_entry_rcu(dumper, &dump_list, list) { - if (dumper->max_reason && reason > dumper->max_reason) - continue; - - /* initialize iterator with data about the stored records */ - dumper->active = true; - - raw_spin_lock_irqsave(&logbuf_lock, flags); - dumper->cur_seq = clear_seq; - dumper->cur_idx = clear_idx; - dumper->next_seq = log_next_seq; - dumper->next_idx = log_next_idx; - raw_spin_unlock_irqrestore(&logbuf_lock, flags); - - /* invoke dumper which will iterate over records */ - dumper->dump(dumper, reason); - - /* reset iterator */ - dumper->active = false; - } - rcu_read_unlock(); -} - -/** - * kmsg_dump_get_line_nolock - retrieve one kmsg log line (unlocked version) - * @dumper: registered kmsg dumper - * @syslog: include the "<4>" prefixes - * @line: buffer to copy the line to - * @size: maximum size of the buffer - * @len: length of line placed into buffer - * - * Start at the beginning of the kmsg buffer, with the oldest kmsg - * record, and copy one record into the provided buffer. - * - * Consecutive calls will return the next available record moving - * towards the end of the buffer with the youngest messages. - * - * A return value of FALSE indicates that there are no more records to - * read. - * - * The function is similar to kmsg_dump_get_line(), but grabs no locks. - */ -bool kmsg_dump_get_line_nolock(struct kmsg_dumper *dumper, bool syslog, - char *line, size_t size, size_t *len) -{ - struct log *msg; - size_t l = 0; - bool ret = false; - - if (!dumper->active) - goto out; - - if (dumper->cur_seq < log_first_seq) { - /* messages are gone, move to first available one */ - dumper->cur_seq = log_first_seq; - dumper->cur_idx = log_first_idx; - } - - /* last entry */ - if (dumper->cur_seq >= log_next_seq) - goto out; - - msg = log_from_idx(dumper->cur_idx); - l = msg_print_text(msg, 0, syslog, line, size); - - dumper->cur_idx = log_next(dumper->cur_idx); - dumper->cur_seq++; - ret = true; -out: - if (len) - *len = l; - return ret; -} - -/** - * kmsg_dump_get_line - retrieve one kmsg log line - * @dumper: registered kmsg dumper - * @syslog: include the "<4>" prefixes - * @line: buffer to copy the line to - * @size: maximum size of the buffer - * @len: length of line placed into buffer - * - * Start at the beginning of the kmsg buffer, with the oldest kmsg - * record, and copy one record into the provided buffer. - * - * Consecutive calls will return the next available record moving - * towards the end of the buffer with the youngest messages. - * - * A return value of FALSE indicates that there are no more records to - * read. - */ -bool kmsg_dump_get_line(struct kmsg_dumper *dumper, bool syslog, - char *line, size_t size, size_t *len) -{ - unsigned long flags; - bool ret; - - raw_spin_lock_irqsave(&logbuf_lock, flags); - ret = kmsg_dump_get_line_nolock(dumper, syslog, line, size, len); - raw_spin_unlock_irqrestore(&logbuf_lock, flags); - - return ret; -} -EXPORT_SYMBOL_GPL(kmsg_dump_get_line); - -/** - * kmsg_dump_get_buffer - copy kmsg log lines - * @dumper: registered kmsg dumper - * @syslog: include the "<4>" prefixes - * @buf: buffer to copy the line to - * @size: maximum size of the buffer - * @len: length of line placed into buffer - * - * Start at the end of the kmsg buffer and fill the provided buffer - * with as many of the the *youngest* kmsg records that fit into it. - * If the buffer is large enough, all available kmsg records will be - * copied with a single call. - * - * Consecutive calls will fill the buffer with the next block of - * available older records, not including the earlier retrieved ones. - * - * A return value of FALSE indicates that there are no more records to - * read. - */ -bool kmsg_dump_get_buffer(struct kmsg_dumper *dumper, bool syslog, - char *buf, size_t size, size_t *len) -{ - unsigned long flags; - u64 seq; - u32 idx; - u64 next_seq; - u32 next_idx; - enum log_flags prev; - size_t l = 0; - bool ret = false; - - if (!dumper->active) - goto out; - - raw_spin_lock_irqsave(&logbuf_lock, flags); - if (dumper->cur_seq < log_first_seq) { - /* messages are gone, move to first available one */ - dumper->cur_seq = log_first_seq; - dumper->cur_idx = log_first_idx; - } - - /* last entry */ - if (dumper->cur_seq >= dumper->next_seq) { - raw_spin_unlock_irqrestore(&logbuf_lock, flags); - goto out; - } - - /* calculate length of entire buffer */ - seq = dumper->cur_seq; - idx = dumper->cur_idx; - prev = 0; - while (seq < dumper->next_seq) { - struct log *msg = log_from_idx(idx); - - l += msg_print_text(msg, prev, true, NULL, 0); - idx = log_next(idx); - seq++; - prev = msg->flags; - } - - /* move first record forward until length fits into the buffer */ - seq = dumper->cur_seq; - idx = dumper->cur_idx; - prev = 0; - while (l > size && seq < dumper->next_seq) { - struct log *msg = log_from_idx(idx); - - l -= msg_print_text(msg, prev, true, NULL, 0); - idx = log_next(idx); - seq++; - prev = msg->flags; - } - - /* last message in next interation */ - next_seq = seq; - next_idx = idx; - - l = 0; - prev = 0; - while (seq < dumper->next_seq) { - struct log *msg = log_from_idx(idx); - - l += msg_print_text(msg, prev, syslog, buf + l, size - l); - idx = log_next(idx); - seq++; - prev = msg->flags; - } - - dumper->next_seq = next_seq; - dumper->next_idx = next_idx; - ret = true; - raw_spin_unlock_irqrestore(&logbuf_lock, flags); -out: - if (len) - *len = l; - return ret; -} -EXPORT_SYMBOL_GPL(kmsg_dump_get_buffer); - -/** - * kmsg_dump_rewind_nolock - reset the interator (unlocked version) - * @dumper: registered kmsg dumper - * - * Reset the dumper's iterator so that kmsg_dump_get_line() and - * kmsg_dump_get_buffer() can be called again and used multiple - * times within the same dumper.dump() callback. - * - * The function is similar to kmsg_dump_rewind(), but grabs no locks. - */ -void kmsg_dump_rewind_nolock(struct kmsg_dumper *dumper) -{ - dumper->cur_seq = clear_seq; - dumper->cur_idx = clear_idx; - dumper->next_seq = log_next_seq; - dumper->next_idx = log_next_idx; -} - -/** - * kmsg_dump_rewind - reset the interator - * @dumper: registered kmsg dumper - * - * Reset the dumper's iterator so that kmsg_dump_get_line() and - * kmsg_dump_get_buffer() can be called again and used multiple - * times within the same dumper.dump() callback. - */ -void kmsg_dump_rewind(struct kmsg_dumper *dumper) -{ - unsigned long flags; - - raw_spin_lock_irqsave(&logbuf_lock, flags); - kmsg_dump_rewind_nolock(dumper); - raw_spin_unlock_irqrestore(&logbuf_lock, flags); -} -EXPORT_SYMBOL_GPL(kmsg_dump_rewind); - -static char dump_stack_arch_desc_str[128]; - -/** - * dump_stack_set_arch_desc - set arch-specific str to show with task dumps - * @fmt: printf-style format string - * @...: arguments for the format string - * - * The configured string will be printed right after utsname during task - * dumps. Usually used to add arch-specific system identifiers. If an - * arch wants to make use of such an ID string, it should initialize this - * as soon as possible during boot. - */ -void __init dump_stack_set_arch_desc(const char *fmt, ...) -{ - va_list args; - - va_start(args, fmt); - vsnprintf(dump_stack_arch_desc_str, sizeof(dump_stack_arch_desc_str), - fmt, args); - va_end(args); -} - -/** - * dump_stack_print_info - print generic debug info for dump_stack() - * @log_lvl: log level - * - * Arch-specific dump_stack() implementations can use this function to - * print out the same debug information as the generic dump_stack(). - */ -void dump_stack_print_info(const char *log_lvl) -{ - printk("%sCPU: %d PID: %d Comm: %.20s %s %s %.*s\n", - log_lvl, raw_smp_processor_id(), current->pid, current->comm, - print_tainted(), init_utsname()->release, - (int)strcspn(init_utsname()->version, " "), - init_utsname()->version); - - if (dump_stack_arch_desc_str[0] != '\0') - printk("%sHardware name: %s\n", - log_lvl, dump_stack_arch_desc_str); - - print_worker_info(log_lvl, current); -} - -/** - * show_regs_print_info - print generic debug info for show_regs() - * @log_lvl: log level - * - * show_regs() implementations can use this function to print out generic - * debug information. - */ -void show_regs_print_info(const char *log_lvl) -{ - dump_stack_print_info(log_lvl); - - printk("%stask: %p ti: %p task.ti: %p\n", - log_lvl, current, current_thread_info(), - task_thread_info(current)); -} - -#endif diff --git a/kernel/printk/Makefile b/kernel/printk/Makefile new file mode 100644 index 000000000000..36d306d9273c --- /dev/null +++ b/kernel/printk/Makefile @@ -0,0 +1 @@ +obj-y = printk.o diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c new file mode 100644 index 000000000000..69b0890ed7e5 --- /dev/null +++ b/kernel/printk/printk.c @@ -0,0 +1,2924 @@ +/* + * linux/kernel/printk.c + * + * Copyright (C) 1991, 1992 Linus Torvalds + * + * Modified to make sys_syslog() more flexible: added commands to + * return the last 4k of kernel messages, regardless of whether + * they've been read or not. Added option to suppress kernel printk's + * to the console. Added hook for sending the console messages + * elsewhere, in preparation for a serial line console (someday). + * Ted Ts'o, 2/11/93. + * Modified for sysctl support, 1/8/97, Chris Horn. + * Fixed SMP synchronization, 08/08/99, Manfred Spraul + * manfred@colorfullife.com + * Rewrote bits to get rid of console_lock + * 01Mar01 Andrew Morton + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include /* For in_interrupt() */ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include + +#define CREATE_TRACE_POINTS +#include + +/* printk's without a loglevel use this.. */ +#define DEFAULT_MESSAGE_LOGLEVEL CONFIG_DEFAULT_MESSAGE_LOGLEVEL + +/* We show everything that is MORE important than this.. */ +#define MINIMUM_CONSOLE_LOGLEVEL 1 /* Minimum loglevel we let people use */ +#define DEFAULT_CONSOLE_LOGLEVEL 7 /* anything MORE serious than KERN_DEBUG */ + +int console_printk[4] = { + DEFAULT_CONSOLE_LOGLEVEL, /* console_loglevel */ + DEFAULT_MESSAGE_LOGLEVEL, /* default_message_loglevel */ + MINIMUM_CONSOLE_LOGLEVEL, /* minimum_console_loglevel */ + DEFAULT_CONSOLE_LOGLEVEL, /* default_console_loglevel */ +}; + +/* + * Low level drivers may need that to know if they can schedule in + * their unblank() callback or not. So let's export it. + */ +int oops_in_progress; +EXPORT_SYMBOL(oops_in_progress); + +/* + * console_sem protects the console_drivers list, and also + * provides serialisation for access to the entire console + * driver system. + */ +static DEFINE_SEMAPHORE(console_sem); +struct console *console_drivers; +EXPORT_SYMBOL_GPL(console_drivers); + +#ifdef CONFIG_LOCKDEP +static struct lockdep_map console_lock_dep_map = { + .name = "console_lock" +}; +#endif + +/* + * This is used for debugging the mess that is the VT code by + * keeping track if we have the console semaphore held. It's + * definitely not the perfect debug tool (we don't know if _WE_ + * hold it are racing, but it helps tracking those weird code + * path in the console code where we end up in places I want + * locked without the console sempahore held + */ +static int console_locked, console_suspended; + +/* + * If exclusive_console is non-NULL then only this console is to be printed to. + */ +static struct console *exclusive_console; + +/* + * Array of consoles built from command line options (console=) + */ +struct console_cmdline +{ + char name[8]; /* Name of the driver */ + int index; /* Minor dev. to use */ + char *options; /* Options for the driver */ +#ifdef CONFIG_A11Y_BRAILLE_CONSOLE + char *brl_options; /* Options for braille driver */ +#endif +}; + +#define MAX_CMDLINECONSOLES 8 + +static struct console_cmdline console_cmdline[MAX_CMDLINECONSOLES]; +static int selected_console = -1; +static int preferred_console = -1; +int console_set_on_cmdline; +EXPORT_SYMBOL(console_set_on_cmdline); + +/* Flag: console code may call schedule() */ +static int console_may_schedule; + +/* + * The printk log buffer consists of a chain of concatenated variable + * length records. Every record starts with a record header, containing + * the overall length of the record. + * + * The heads to the first and last entry in the buffer, as well as the + * sequence numbers of these both entries are maintained when messages + * are stored.. + * + * If the heads indicate available messages, the length in the header + * tells the start next message. A length == 0 for the next message + * indicates a wrap-around to the beginning of the buffer. + * + * Every record carries the monotonic timestamp in microseconds, as well as + * the standard userspace syslog level and syslog facility. The usual + * kernel messages use LOG_KERN; userspace-injected messages always carry + * a matching syslog facility, by default LOG_USER. The origin of every + * message can be reliably determined that way. + * + * The human readable log message directly follows the message header. The + * length of the message text is stored in the header, the stored message + * is not terminated. + * + * Optionally, a message can carry a dictionary of properties (key/value pairs), + * to provide userspace with a machine-readable message context. + * + * Examples for well-defined, commonly used property names are: + * DEVICE=b12:8 device identifier + * b12:8 block dev_t + * c127:3 char dev_t + * n8 netdev ifindex + * +sound:card0 subsystem:devname + * SUBSYSTEM=pci driver-core subsystem name + * + * Valid characters in property names are [a-zA-Z0-9.-_]. The plain text value + * follows directly after a '=' character. Every property is terminated by + * a '\0' character. The last property is not terminated. + * + * Example of a message structure: + * 0000 ff 8f 00 00 00 00 00 00 monotonic time in nsec + * 0008 34 00 record is 52 bytes long + * 000a 0b 00 text is 11 bytes long + * 000c 1f 00 dictionary is 23 bytes long + * 000e 03 00 LOG_KERN (facility) LOG_ERR (level) + * 0010 69 74 27 73 20 61 20 6c "it's a l" + * 69 6e 65 "ine" + * 001b 44 45 56 49 43 "DEVIC" + * 45 3d 62 38 3a 32 00 44 "E=b8:2\0D" + * 52 49 56 45 52 3d 62 75 "RIVER=bu" + * 67 "g" + * 0032 00 00 00 padding to next message header + * + * The 'struct log' buffer header must never be directly exported to + * userspace, it is a kernel-private implementation detail that might + * need to be changed in the future, when the requirements change. + * + * /dev/kmsg exports the structured data in the following line format: + * "level,sequnum,timestamp;\n" + * + * The optional key/value pairs are attached as continuation lines starting + * with a space character and terminated by a newline. All possible + * non-prinatable characters are escaped in the "\xff" notation. + * + * Users of the export format should ignore possible additional values + * separated by ',', and find the message after the ';' character. + */ + +enum log_flags { + LOG_NOCONS = 1, /* already flushed, do not print to console */ + LOG_NEWLINE = 2, /* text ended with a newline */ + LOG_PREFIX = 4, /* text started with a prefix */ + LOG_CONT = 8, /* text is a fragment of a continuation line */ +}; + +struct log { + u64 ts_nsec; /* timestamp in nanoseconds */ + u16 len; /* length of entire record */ + u16 text_len; /* length of text buffer */ + u16 dict_len; /* length of dictionary buffer */ + u8 facility; /* syslog facility */ + u8 flags:5; /* internal record flags */ + u8 level:3; /* syslog level */ +}; + +/* + * The logbuf_lock protects kmsg buffer, indices, counters. It is also + * used in interesting ways to provide interlocking in console_unlock(); + */ +static DEFINE_RAW_SPINLOCK(logbuf_lock); + +#ifdef CONFIG_PRINTK +DECLARE_WAIT_QUEUE_HEAD(log_wait); +/* the next printk record to read by syslog(READ) or /proc/kmsg */ +static u64 syslog_seq; +static u32 syslog_idx; +static enum log_flags syslog_prev; +static size_t syslog_partial; + +/* index and sequence number of the first record stored in the buffer */ +static u64 log_first_seq; +static u32 log_first_idx; + +/* index and sequence number of the next record to store in the buffer */ +static u64 log_next_seq; +static u32 log_next_idx; + +/* the next printk record to write to the console */ +static u64 console_seq; +static u32 console_idx; +static enum log_flags console_prev; + +/* the next printk record to read after the last 'clear' command */ +static u64 clear_seq; +static u32 clear_idx; + +#define PREFIX_MAX 32 +#define LOG_LINE_MAX 1024 - PREFIX_MAX + +/* record buffer */ +#if defined(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS) +#define LOG_ALIGN 4 +#else +#define LOG_ALIGN __alignof__(struct log) +#endif +#define __LOG_BUF_LEN (1 << CONFIG_LOG_BUF_SHIFT) +static char __log_buf[__LOG_BUF_LEN] __aligned(LOG_ALIGN); +static char *log_buf = __log_buf; +static u32 log_buf_len = __LOG_BUF_LEN; + +/* cpu currently holding logbuf_lock */ +static volatile unsigned int logbuf_cpu = UINT_MAX; + +/* human readable text of the record */ +static char *log_text(const struct log *msg) +{ + return (char *)msg + sizeof(struct log); +} + +/* optional key/value pair dictionary attached to the record */ +static char *log_dict(const struct log *msg) +{ + return (char *)msg + sizeof(struct log) + msg->text_len; +} + +/* get record by index; idx must point to valid msg */ +static struct log *log_from_idx(u32 idx) +{ + struct log *msg = (struct log *)(log_buf + idx); + + /* + * A length == 0 record is the end of buffer marker. Wrap around and + * read the message at the start of the buffer. + */ + if (!msg->len) + return (struct log *)log_buf; + return msg; +} + +/* get next record; idx must point to valid msg */ +static u32 log_next(u32 idx) +{ + struct log *msg = (struct log *)(log_buf + idx); + + /* length == 0 indicates the end of the buffer; wrap */ + /* + * A length == 0 record is the end of buffer marker. Wrap around and + * read the message at the start of the buffer as *this* one, and + * return the one after that. + */ + if (!msg->len) { + msg = (struct log *)log_buf; + return msg->len; + } + return idx + msg->len; +} + +/* insert record into the buffer, discard old ones, update heads */ +static void log_store(int facility, int level, + enum log_flags flags, u64 ts_nsec, + const char *dict, u16 dict_len, + const char *text, u16 text_len) +{ + struct log *msg; + u32 size, pad_len; + + /* number of '\0' padding bytes to next message */ + size = sizeof(struct log) + text_len + dict_len; + pad_len = (-size) & (LOG_ALIGN - 1); + size += pad_len; + + while (log_first_seq < log_next_seq) { + u32 free; + + if (log_next_idx > log_first_idx) + free = max(log_buf_len - log_next_idx, log_first_idx); + else + free = log_first_idx - log_next_idx; + + if (free > size + sizeof(struct log)) + break; + + /* drop old messages until we have enough contiuous space */ + log_first_idx = log_next(log_first_idx); + log_first_seq++; + } + + if (log_next_idx + size + sizeof(struct log) >= log_buf_len) { + /* + * This message + an additional empty header does not fit + * at the end of the buffer. Add an empty header with len == 0 + * to signify a wrap around. + */ + memset(log_buf + log_next_idx, 0, sizeof(struct log)); + log_next_idx = 0; + } + + /* fill message */ + msg = (struct log *)(log_buf + log_next_idx); + memcpy(log_text(msg), text, text_len); + msg->text_len = text_len; + memcpy(log_dict(msg), dict, dict_len); + msg->dict_len = dict_len; + msg->facility = facility; + msg->level = level & 7; + msg->flags = flags & 0x1f; + if (ts_nsec > 0) + msg->ts_nsec = ts_nsec; + else + msg->ts_nsec = local_clock(); + memset(log_dict(msg) + dict_len, 0, pad_len); + msg->len = sizeof(struct log) + text_len + dict_len + pad_len; + + /* insert message */ + log_next_idx += msg->len; + log_next_seq++; +} + +#ifdef CONFIG_SECURITY_DMESG_RESTRICT +int dmesg_restrict = 1; +#else +int dmesg_restrict; +#endif + +static int syslog_action_restricted(int type) +{ + if (dmesg_restrict) + return 1; + /* + * Unless restricted, we allow "read all" and "get buffer size" + * for everybody. + */ + return type != SYSLOG_ACTION_READ_ALL && + type != SYSLOG_ACTION_SIZE_BUFFER; +} + +static int check_syslog_permissions(int type, bool from_file) +{ + /* + * If this is from /proc/kmsg and we've already opened it, then we've + * already done the capabilities checks at open time. + */ + if (from_file && type != SYSLOG_ACTION_OPEN) + return 0; + + if (syslog_action_restricted(type)) { + if (capable(CAP_SYSLOG)) + return 0; + /* + * For historical reasons, accept CAP_SYS_ADMIN too, with + * a warning. + */ + if (capable(CAP_SYS_ADMIN)) { + pr_warn_once("%s (%d): Attempt to access syslog with " + "CAP_SYS_ADMIN but no CAP_SYSLOG " + "(deprecated).\n", + current->comm, task_pid_nr(current)); + return 0; + } + return -EPERM; + } + return security_syslog(type); +} + + +/* /dev/kmsg - userspace message inject/listen interface */ +struct devkmsg_user { + u64 seq; + u32 idx; + enum log_flags prev; + struct mutex lock; + char buf[8192]; +}; + +static ssize_t devkmsg_writev(struct kiocb *iocb, const struct iovec *iv, + unsigned long count, loff_t pos) +{ + char *buf, *line; + int i; + int level = default_message_loglevel; + int facility = 1; /* LOG_USER */ + size_t len = iov_length(iv, count); + ssize_t ret = len; + + if (len > LOG_LINE_MAX) + return -EINVAL; + buf = kmalloc(len+1, GFP_KERNEL); + if (buf == NULL) + return -ENOMEM; + + line = buf; + for (i = 0; i < count; i++) { + if (copy_from_user(line, iv[i].iov_base, iv[i].iov_len)) { + ret = -EFAULT; + goto out; + } + line += iv[i].iov_len; + } + + /* + * Extract and skip the syslog prefix <[0-9]*>. Coming from userspace + * the decimal value represents 32bit, the lower 3 bit are the log + * level, the rest are the log facility. + * + * If no prefix or no userspace facility is specified, we + * enforce LOG_USER, to be able to reliably distinguish + * kernel-generated messages from userspace-injected ones. + */ + line = buf; + if (line[0] == '<') { + char *endp = NULL; + + i = simple_strtoul(line+1, &endp, 10); + if (endp && endp[0] == '>') { + level = i & 7; + if (i >> 3) + facility = i >> 3; + endp++; + len -= endp - line; + line = endp; + } + } + line[len] = '\0'; + + printk_emit(facility, level, NULL, 0, "%s", line); +out: + kfree(buf); + return ret; +} + +static ssize_t devkmsg_read(struct file *file, char __user *buf, + size_t count, loff_t *ppos) +{ + struct devkmsg_user *user = file->private_data; + struct log *msg; + u64 ts_usec; + size_t i; + char cont = '-'; + size_t len; + ssize_t ret; + + if (!user) + return -EBADF; + + ret = mutex_lock_interruptible(&user->lock); + if (ret) + return ret; + raw_spin_lock_irq(&logbuf_lock); + while (user->seq == log_next_seq) { + if (file->f_flags & O_NONBLOCK) { + ret = -EAGAIN; + raw_spin_unlock_irq(&logbuf_lock); + goto out; + } + + raw_spin_unlock_irq(&logbuf_lock); + ret = wait_event_interruptible(log_wait, + user->seq != log_next_seq); + if (ret) + goto out; + raw_spin_lock_irq(&logbuf_lock); + } + + if (user->seq < log_first_seq) { + /* our last seen message is gone, return error and reset */ + user->idx = log_first_idx; + user->seq = log_first_seq; + ret = -EPIPE; + raw_spin_unlock_irq(&logbuf_lock); + goto out; + } + + msg = log_from_idx(user->idx); + ts_usec = msg->ts_nsec; + do_div(ts_usec, 1000); + + /* + * If we couldn't merge continuation line fragments during the print, + * export the stored flags to allow an optional external merge of the + * records. Merging the records isn't always neccessarily correct, like + * when we hit a race during printing. In most cases though, it produces + * better readable output. 'c' in the record flags mark the first + * fragment of a line, '+' the following. + */ + if (msg->flags & LOG_CONT && !(user->prev & LOG_CONT)) + cont = 'c'; + else if ((msg->flags & LOG_CONT) || + ((user->prev & LOG_CONT) && !(msg->flags & LOG_PREFIX))) + cont = '+'; + + len = sprintf(user->buf, "%u,%llu,%llu,%c;", + (msg->facility << 3) | msg->level, + user->seq, ts_usec, cont); + user->prev = msg->flags; + + /* escape non-printable characters */ + for (i = 0; i < msg->text_len; i++) { + unsigned char c = log_text(msg)[i]; + + if (c < ' ' || c >= 127 || c == '\\') + len += sprintf(user->buf + len, "\\x%02x", c); + else + user->buf[len++] = c; + } + user->buf[len++] = '\n'; + + if (msg->dict_len) { + bool line = true; + + for (i = 0; i < msg->dict_len; i++) { + unsigned char c = log_dict(msg)[i]; + + if (line) { + user->buf[len++] = ' '; + line = false; + } + + if (c == '\0') { + user->buf[len++] = '\n'; + line = true; + continue; + } + + if (c < ' ' || c >= 127 || c == '\\') { + len += sprintf(user->buf + len, "\\x%02x", c); + continue; + } + + user->buf[len++] = c; + } + user->buf[len++] = '\n'; + } + + user->idx = log_next(user->idx); + user->seq++; + raw_spin_unlock_irq(&logbuf_lock); + + if (len > count) { + ret = -EINVAL; + goto out; + } + + if (copy_to_user(buf, user->buf, len)) { + ret = -EFAULT; + goto out; + } + ret = len; +out: + mutex_unlock(&user->lock); + return ret; +} + +static loff_t devkmsg_llseek(struct file *file, loff_t offset, int whence) +{ + struct devkmsg_user *user = file->private_data; + loff_t ret = 0; + + if (!user) + return -EBADF; + if (offset) + return -ESPIPE; + + raw_spin_lock_irq(&logbuf_lock); + switch (whence) { + case SEEK_SET: + /* the first record */ + user->idx = log_first_idx; + user->seq = log_first_seq; + break; + case SEEK_DATA: + /* + * The first record after the last SYSLOG_ACTION_CLEAR, + * like issued by 'dmesg -c'. Reading /dev/kmsg itself + * changes no global state, and does not clear anything. + */ + user->idx = clear_idx; + user->seq = clear_seq; + break; + case SEEK_END: + /* after the last record */ + user->idx = log_next_idx; + user->seq = log_next_seq; + break; + default: + ret = -EINVAL; + } + raw_spin_unlock_irq(&logbuf_lock); + return ret; +} + +static unsigned int devkmsg_poll(struct file *file, poll_table *wait) +{ + struct devkmsg_user *user = file->private_data; + int ret = 0; + + if (!user) + return POLLERR|POLLNVAL; + + poll_wait(file, &log_wait, wait); + + raw_spin_lock_irq(&logbuf_lock); + if (user->seq < log_next_seq) { + /* return error when data has vanished underneath us */ + if (user->seq < log_first_seq) + ret = POLLIN|POLLRDNORM|POLLERR|POLLPRI; + else + ret = POLLIN|POLLRDNORM; + } + raw_spin_unlock_irq(&logbuf_lock); + + return ret; +} + +static int devkmsg_open(struct inode *inode, struct file *file) +{ + struct devkmsg_user *user; + int err; + + /* write-only does not need any file context */ + if ((file->f_flags & O_ACCMODE) == O_WRONLY) + return 0; + + err = check_syslog_permissions(SYSLOG_ACTION_READ_ALL, + SYSLOG_FROM_READER); + if (err) + return err; + + user = kmalloc(sizeof(struct devkmsg_user), GFP_KERNEL); + if (!user) + return -ENOMEM; + + mutex_init(&user->lock); + + raw_spin_lock_irq(&logbuf_lock); + user->idx = log_first_idx; + user->seq = log_first_seq; + raw_spin_unlock_irq(&logbuf_lock); + + file->private_data = user; + return 0; +} + +static int devkmsg_release(struct inode *inode, struct file *file) +{ + struct devkmsg_user *user = file->private_data; + + if (!user) + return 0; + + mutex_destroy(&user->lock); + kfree(user); + return 0; +} + +const struct file_operations kmsg_fops = { + .open = devkmsg_open, + .read = devkmsg_read, + .aio_write = devkmsg_writev, + .llseek = devkmsg_llseek, + .poll = devkmsg_poll, + .release = devkmsg_release, +}; + +#ifdef CONFIG_KEXEC +/* + * This appends the listed symbols to /proc/vmcoreinfo + * + * /proc/vmcoreinfo is used by various utiilties, like crash and makedumpfile to + * obtain access to symbols that are otherwise very difficult to locate. These + * symbols are specifically used so that utilities can access and extract the + * dmesg log from a vmcore file after a crash. + */ +void log_buf_kexec_setup(void) +{ + VMCOREINFO_SYMBOL(log_buf); + VMCOREINFO_SYMBOL(log_buf_len); + VMCOREINFO_SYMBOL(log_first_idx); + VMCOREINFO_SYMBOL(log_next_idx); + /* + * Export struct log size and field offsets. User space tools can + * parse it and detect any changes to structure down the line. + */ + VMCOREINFO_STRUCT_SIZE(log); + VMCOREINFO_OFFSET(log, ts_nsec); + VMCOREINFO_OFFSET(log, len); + VMCOREINFO_OFFSET(log, text_len); + VMCOREINFO_OFFSET(log, dict_len); +} +#endif + +/* requested log_buf_len from kernel cmdline */ +static unsigned long __initdata new_log_buf_len; + +/* save requested log_buf_len since it's too early to process it */ +static int __init log_buf_len_setup(char *str) +{ + unsigned size = memparse(str, &str); + + if (size) + size = roundup_pow_of_two(size); + if (size > log_buf_len) + new_log_buf_len = size; + + return 0; +} +early_param("log_buf_len", log_buf_len_setup); + +void __init setup_log_buf(int early) +{ + unsigned long flags; + char *new_log_buf; + int free; + + if (!new_log_buf_len) + return; + + if (early) { + unsigned long mem; + + mem = memblock_alloc(new_log_buf_len, PAGE_SIZE); + if (!mem) + return; + new_log_buf = __va(mem); + } else { + new_log_buf = alloc_bootmem_nopanic(new_log_buf_len); + } + + if (unlikely(!new_log_buf)) { + pr_err("log_buf_len: %ld bytes not available\n", + new_log_buf_len); + return; + } + + raw_spin_lock_irqsave(&logbuf_lock, flags); + log_buf_len = new_log_buf_len; + log_buf = new_log_buf; + new_log_buf_len = 0; + free = __LOG_BUF_LEN - log_next_idx; + memcpy(log_buf, __log_buf, __LOG_BUF_LEN); + raw_spin_unlock_irqrestore(&logbuf_lock, flags); + + pr_info("log_buf_len: %d\n", log_buf_len); + pr_info("early log buf free: %d(%d%%)\n", + free, (free * 100) / __LOG_BUF_LEN); +} + +static bool __read_mostly ignore_loglevel; + +static int __init ignore_loglevel_setup(char *str) +{ + ignore_loglevel = 1; + printk(KERN_INFO "debug: ignoring loglevel setting.\n"); + + return 0; +} + +early_param("ignore_loglevel", ignore_loglevel_setup); +module_param(ignore_loglevel, bool, S_IRUGO | S_IWUSR); +MODULE_PARM_DESC(ignore_loglevel, "ignore loglevel setting, to" + "print all kernel messages to the console."); + +#ifdef CONFIG_BOOT_PRINTK_DELAY + +static int boot_delay; /* msecs delay after each printk during bootup */ +static unsigned long long loops_per_msec; /* based on boot_delay */ + +static int __init boot_delay_setup(char *str) +{ + unsigned long lpj; + + lpj = preset_lpj ? preset_lpj : 1000000; /* some guess */ + loops_per_msec = (unsigned long long)lpj / 1000 * HZ; + + get_option(&str, &boot_delay); + if (boot_delay > 10 * 1000) + boot_delay = 0; + + pr_debug("boot_delay: %u, preset_lpj: %ld, lpj: %lu, " + "HZ: %d, loops_per_msec: %llu\n", + boot_delay, preset_lpj, lpj, HZ, loops_per_msec); + return 1; +} +__setup("boot_delay=", boot_delay_setup); + +static void boot_delay_msec(int level) +{ + unsigned long long k; + unsigned long timeout; + + if ((boot_delay == 0 || system_state != SYSTEM_BOOTING) + || (level >= console_loglevel && !ignore_loglevel)) { + return; + } + + k = (unsigned long long)loops_per_msec * boot_delay; + + timeout = jiffies + msecs_to_jiffies(boot_delay); + while (k) { + k--; + cpu_relax(); + /* + * use (volatile) jiffies to prevent + * compiler reduction; loop termination via jiffies + * is secondary and may or may not happen. + */ + if (time_after(jiffies, timeout)) + break; + touch_nmi_watchdog(); + } +} +#else +static inline void boot_delay_msec(int level) +{ +} +#endif + +#if defined(CONFIG_PRINTK_TIME) +static bool printk_time = 1; +#else +static bool printk_time; +#endif +module_param_named(time, printk_time, bool, S_IRUGO | S_IWUSR); + +static size_t print_time(u64 ts, char *buf) +{ + unsigned long rem_nsec; + + if (!printk_time) + return 0; + + rem_nsec = do_div(ts, 1000000000); + + if (!buf) + return snprintf(NULL, 0, "[%5lu.000000] ", (unsigned long)ts); + + return sprintf(buf, "[%5lu.%06lu] ", + (unsigned long)ts, rem_nsec / 1000); +} + +static size_t print_prefix(const struct log *msg, bool syslog, char *buf) +{ + size_t len = 0; + unsigned int prefix = (msg->facility << 3) | msg->level; + + if (syslog) { + if (buf) { + len += sprintf(buf, "<%u>", prefix); + } else { + len += 3; + if (prefix > 999) + len += 3; + else if (prefix > 99) + len += 2; + else if (prefix > 9) + len++; + } + } + + len += print_time(msg->ts_nsec, buf ? buf + len : NULL); + return len; +} + +static size_t msg_print_text(const struct log *msg, enum log_flags prev, + bool syslog, char *buf, size_t size) +{ + const char *text = log_text(msg); + size_t text_size = msg->text_len; + bool prefix = true; + bool newline = true; + size_t len = 0; + + if ((prev & LOG_CONT) && !(msg->flags & LOG_PREFIX)) + prefix = false; + + if (msg->flags & LOG_CONT) { + if ((prev & LOG_CONT) && !(prev & LOG_NEWLINE)) + prefix = false; + + if (!(msg->flags & LOG_NEWLINE)) + newline = false; + } + + do { + const char *next = memchr(text, '\n', text_size); + size_t text_len; + + if (next) { + text_len = next - text; + next++; + text_size -= next - text; + } else { + text_len = text_size; + } + + if (buf) { + if (print_prefix(msg, syslog, NULL) + + text_len + 1 >= size - len) + break; + + if (prefix) + len += print_prefix(msg, syslog, buf + len); + memcpy(buf + len, text, text_len); + len += text_len; + if (next || newline) + buf[len++] = '\n'; + } else { + /* SYSLOG_ACTION_* buffer size only calculation */ + if (prefix) + len += print_prefix(msg, syslog, NULL); + len += text_len; + if (next || newline) + len++; + } + + prefix = true; + text = next; + } while (text); + + return len; +} + +static int syslog_print(char __user *buf, int size) +{ + char *text; + struct log *msg; + int len = 0; + + text = kmalloc(LOG_LINE_MAX + PREFIX_MAX, GFP_KERNEL); + if (!text) + return -ENOMEM; + + while (size > 0) { + size_t n; + size_t skip; + + raw_spin_lock_irq(&logbuf_lock); + if (syslog_seq < log_first_seq) { + /* messages are gone, move to first one */ + syslog_seq = log_first_seq; + syslog_idx = log_first_idx; + syslog_prev = 0; + syslog_partial = 0; + } + if (syslog_seq == log_next_seq) { + raw_spin_unlock_irq(&logbuf_lock); + break; + } + + skip = syslog_partial; + msg = log_from_idx(syslog_idx); + n = msg_print_text(msg, syslog_prev, true, text, + LOG_LINE_MAX + PREFIX_MAX); + if (n - syslog_partial <= size) { + /* message fits into buffer, move forward */ + syslog_idx = log_next(syslog_idx); + syslog_seq++; + syslog_prev = msg->flags; + n -= syslog_partial; + syslog_partial = 0; + } else if (!len){ + /* partial read(), remember position */ + n = size; + syslog_partial += n; + } else + n = 0; + raw_spin_unlock_irq(&logbuf_lock); + + if (!n) + break; + + if (copy_to_user(buf, text + skip, n)) { + if (!len) + len = -EFAULT; + break; + } + + len += n; + size -= n; + buf += n; + } + + kfree(text); + return len; +} + +static int syslog_print_all(char __user *buf, int size, bool clear) +{ + char *text; + int len = 0; + + text = kmalloc(LOG_LINE_MAX + PREFIX_MAX, GFP_KERNEL); + if (!text) + return -ENOMEM; + + raw_spin_lock_irq(&logbuf_lock); + if (buf) { + u64 next_seq; + u64 seq; + u32 idx; + enum log_flags prev; + + if (clear_seq < log_first_seq) { + /* messages are gone, move to first available one */ + clear_seq = log_first_seq; + clear_idx = log_first_idx; + } + + /* + * Find first record that fits, including all following records, + * into the user-provided buffer for this dump. + */ + seq = clear_seq; + idx = clear_idx; + prev = 0; + while (seq < log_next_seq) { + struct log *msg = log_from_idx(idx); + + len += msg_print_text(msg, prev, true, NULL, 0); + prev = msg->flags; + idx = log_next(idx); + seq++; + } + + /* move first record forward until length fits into the buffer */ + seq = clear_seq; + idx = clear_idx; + prev = 0; + while (len > size && seq < log_next_seq) { + struct log *msg = log_from_idx(idx); + + len -= msg_print_text(msg, prev, true, NULL, 0); + prev = msg->flags; + idx = log_next(idx); + seq++; + } + + /* last message fitting into this dump */ + next_seq = log_next_seq; + + len = 0; + prev = 0; + while (len >= 0 && seq < next_seq) { + struct log *msg = log_from_idx(idx); + int textlen; + + textlen = msg_print_text(msg, prev, true, text, + LOG_LINE_MAX + PREFIX_MAX); + if (textlen < 0) { + len = textlen; + break; + } + idx = log_next(idx); + seq++; + prev = msg->flags; + + raw_spin_unlock_irq(&logbuf_lock); + if (copy_to_user(buf + len, text, textlen)) + len = -EFAULT; + else + len += textlen; + raw_spin_lock_irq(&logbuf_lock); + + if (seq < log_first_seq) { + /* messages are gone, move to next one */ + seq = log_first_seq; + idx = log_first_idx; + prev = 0; + } + } + } + + if (clear) { + clear_seq = log_next_seq; + clear_idx = log_next_idx; + } + raw_spin_unlock_irq(&logbuf_lock); + + kfree(text); + return len; +} + +int do_syslog(int type, char __user *buf, int len, bool from_file) +{ + bool clear = false; + static int saved_console_loglevel = -1; + int error; + + error = check_syslog_permissions(type, from_file); + if (error) + goto out; + + error = security_syslog(type); + if (error) + return error; + + switch (type) { + case SYSLOG_ACTION_CLOSE: /* Close log */ + break; + case SYSLOG_ACTION_OPEN: /* Open log */ + break; + case SYSLOG_ACTION_READ: /* Read from log */ + error = -EINVAL; + if (!buf || len < 0) + goto out; + error = 0; + if (!len) + goto out; + if (!access_ok(VERIFY_WRITE, buf, len)) { + error = -EFAULT; + goto out; + } + error = wait_event_interruptible(log_wait, + syslog_seq != log_next_seq); + if (error) + goto out; + error = syslog_print(buf, len); + break; + /* Read/clear last kernel messages */ + case SYSLOG_ACTION_READ_CLEAR: + clear = true; + /* FALL THRU */ + /* Read last kernel messages */ + case SYSLOG_ACTION_READ_ALL: + error = -EINVAL; + if (!buf || len < 0) + goto out; + error = 0; + if (!len) + goto out; + if (!access_ok(VERIFY_WRITE, buf, len)) { + error = -EFAULT; + goto out; + } + error = syslog_print_all(buf, len, clear); + break; + /* Clear ring buffer */ + case SYSLOG_ACTION_CLEAR: + syslog_print_all(NULL, 0, true); + break; + /* Disable logging to console */ + case SYSLOG_ACTION_CONSOLE_OFF: + if (saved_console_loglevel == -1) + saved_console_loglevel = console_loglevel; + console_loglevel = minimum_console_loglevel; + break; + /* Enable logging to console */ + case SYSLOG_ACTION_CONSOLE_ON: + if (saved_console_loglevel != -1) { + console_loglevel = saved_console_loglevel; + saved_console_loglevel = -1; + } + break; + /* Set level of messages printed to console */ + case SYSLOG_ACTION_CONSOLE_LEVEL: + error = -EINVAL; + if (len < 1 || len > 8) + goto out; + if (len < minimum_console_loglevel) + len = minimum_console_loglevel; + console_loglevel = len; + /* Implicitly re-enable logging to console */ + saved_console_loglevel = -1; + error = 0; + break; + /* Number of chars in the log buffer */ + case SYSLOG_ACTION_SIZE_UNREAD: + raw_spin_lock_irq(&logbuf_lock); + if (syslog_seq < log_first_seq) { + /* messages are gone, move to first one */ + syslog_seq = log_first_seq; + syslog_idx = log_first_idx; + syslog_prev = 0; + syslog_partial = 0; + } + if (from_file) { + /* + * Short-cut for poll(/"proc/kmsg") which simply checks + * for pending data, not the size; return the count of + * records, not the length. + */ + error = log_next_idx - syslog_idx; + } else { + u64 seq = syslog_seq; + u32 idx = syslog_idx; + enum log_flags prev = syslog_prev; + + error = 0; + while (seq < log_next_seq) { + struct log *msg = log_from_idx(idx); + + error += msg_print_text(msg, prev, true, NULL, 0); + idx = log_next(idx); + seq++; + prev = msg->flags; + } + error -= syslog_partial; + } + raw_spin_unlock_irq(&logbuf_lock); + break; + /* Size of the log buffer */ + case SYSLOG_ACTION_SIZE_BUFFER: + error = log_buf_len; + break; + default: + error = -EINVAL; + break; + } +out: + return error; +} + +SYSCALL_DEFINE3(syslog, int, type, char __user *, buf, int, len) +{ + return do_syslog(type, buf, len, SYSLOG_FROM_READER); +} + +/* + * Call the console drivers, asking them to write out + * log_buf[start] to log_buf[end - 1]. + * The console_lock must be held. + */ +static void call_console_drivers(int level, const char *text, size_t len) +{ + struct console *con; + + trace_console(text, len); + + if (level >= console_loglevel && !ignore_loglevel) + return; + if (!console_drivers) + return; + + for_each_console(con) { + if (exclusive_console && con != exclusive_console) + continue; + if (!(con->flags & CON_ENABLED)) + continue; + if (!con->write) + continue; + if (!cpu_online(smp_processor_id()) && + !(con->flags & CON_ANYTIME)) + continue; + con->write(con, text, len); + } +} + +/* + * Zap console related locks when oopsing. Only zap at most once + * every 10 seconds, to leave time for slow consoles to print a + * full oops. + */ +static void zap_locks(void) +{ + static unsigned long oops_timestamp; + + if (time_after_eq(jiffies, oops_timestamp) && + !time_after(jiffies, oops_timestamp + 30 * HZ)) + return; + + oops_timestamp = jiffies; + + debug_locks_off(); + /* If a crash is occurring, make sure we can't deadlock */ + raw_spin_lock_init(&logbuf_lock); + /* And make sure that we print immediately */ + sema_init(&console_sem, 1); +} + +/* Check if we have any console registered that can be called early in boot. */ +static int have_callable_console(void) +{ + struct console *con; + + for_each_console(con) + if (con->flags & CON_ANYTIME) + return 1; + + return 0; +} + +/* + * Can we actually use the console at this time on this cpu? + * + * Console drivers may assume that per-cpu resources have + * been allocated. So unless they're explicitly marked as + * being able to cope (CON_ANYTIME) don't call them until + * this CPU is officially up. + */ +static inline int can_use_console(unsigned int cpu) +{ + return cpu_online(cpu) || have_callable_console(); +} + +/* + * Try to get console ownership to actually show the kernel + * messages from a 'printk'. Return true (and with the + * console_lock held, and 'console_locked' set) if it + * is successful, false otherwise. + * + * This gets called with the 'logbuf_lock' spinlock held and + * interrupts disabled. It should return with 'lockbuf_lock' + * released but interrupts still disabled. + */ +static int console_trylock_for_printk(unsigned int cpu) + __releases(&logbuf_lock) +{ + int retval = 0, wake = 0; + + if (console_trylock()) { + retval = 1; + + /* + * If we can't use the console, we need to release + * the console semaphore by hand to avoid flushing + * the buffer. We need to hold the console semaphore + * in order to do this test safely. + */ + if (!can_use_console(cpu)) { + console_locked = 0; + wake = 1; + retval = 0; + } + } + logbuf_cpu = UINT_MAX; + raw_spin_unlock(&logbuf_lock); + if (wake) + up(&console_sem); + return retval; +} + +int printk_delay_msec __read_mostly; + +static inline void printk_delay(void) +{ + if (unlikely(printk_delay_msec)) { + int m = printk_delay_msec; + + while (m--) { + mdelay(1); + touch_nmi_watchdog(); + } + } +} + +/* + * Continuation lines are buffered, and not committed to the record buffer + * until the line is complete, or a race forces it. The line fragments + * though, are printed immediately to the consoles to ensure everything has + * reached the console in case of a kernel crash. + */ +static struct cont { + char buf[LOG_LINE_MAX]; + size_t len; /* length == 0 means unused buffer */ + size_t cons; /* bytes written to console */ + struct task_struct *owner; /* task of first print*/ + u64 ts_nsec; /* time of first print */ + u8 level; /* log level of first message */ + u8 facility; /* log level of first message */ + enum log_flags flags; /* prefix, newline flags */ + bool flushed:1; /* buffer sealed and committed */ +} cont; + +static void cont_flush(enum log_flags flags) +{ + if (cont.flushed) + return; + if (cont.len == 0) + return; + + if (cont.cons) { + /* + * If a fragment of this line was directly flushed to the + * console; wait for the console to pick up the rest of the + * line. LOG_NOCONS suppresses a duplicated output. + */ + log_store(cont.facility, cont.level, flags | LOG_NOCONS, + cont.ts_nsec, NULL, 0, cont.buf, cont.len); + cont.flags = flags; + cont.flushed = true; + } else { + /* + * If no fragment of this line ever reached the console, + * just submit it to the store and free the buffer. + */ + log_store(cont.facility, cont.level, flags, 0, + NULL, 0, cont.buf, cont.len); + cont.len = 0; + } +} + +static bool cont_add(int facility, int level, const char *text, size_t len) +{ + if (cont.len && cont.flushed) + return false; + + if (cont.len + len > sizeof(cont.buf)) { + /* the line gets too long, split it up in separate records */ + cont_flush(LOG_CONT); + return false; + } + + if (!cont.len) { + cont.facility = facility; + cont.level = level; + cont.owner = current; + cont.ts_nsec = local_clock(); + cont.flags = 0; + cont.cons = 0; + cont.flushed = false; + } + + memcpy(cont.buf + cont.len, text, len); + cont.len += len; + + if (cont.len > (sizeof(cont.buf) * 80) / 100) + cont_flush(LOG_CONT); + + return true; +} + +static size_t cont_print_text(char *text, size_t size) +{ + size_t textlen = 0; + size_t len; + + if (cont.cons == 0 && (console_prev & LOG_NEWLINE)) { + textlen += print_time(cont.ts_nsec, text); + size -= textlen; + } + + len = cont.len - cont.cons; + if (len > 0) { + if (len+1 > size) + len = size-1; + memcpy(text + textlen, cont.buf + cont.cons, len); + textlen += len; + cont.cons = cont.len; + } + + if (cont.flushed) { + if (cont.flags & LOG_NEWLINE) + text[textlen++] = '\n'; + /* got everything, release buffer */ + cont.len = 0; + } + return textlen; +} + +asmlinkage int vprintk_emit(int facility, int level, + const char *dict, size_t dictlen, + const char *fmt, va_list args) +{ + static int recursion_bug; + static char textbuf[LOG_LINE_MAX]; + char *text = textbuf; + size_t text_len; + enum log_flags lflags = 0; + unsigned long flags; + int this_cpu; + int printed_len = 0; + + boot_delay_msec(level); + printk_delay(); + + /* This stops the holder of console_sem just where we want him */ + local_irq_save(flags); + this_cpu = smp_processor_id(); + + /* + * Ouch, printk recursed into itself! + */ + if (unlikely(logbuf_cpu == this_cpu)) { + /* + * If a crash is occurring during printk() on this CPU, + * then try to get the crash message out but make sure + * we can't deadlock. Otherwise just return to avoid the + * recursion and return - but flag the recursion so that + * it can be printed at the next appropriate moment: + */ + if (!oops_in_progress && !lockdep_recursing(current)) { + recursion_bug = 1; + goto out_restore_irqs; + } + zap_locks(); + } + + lockdep_off(); + raw_spin_lock(&logbuf_lock); + logbuf_cpu = this_cpu; + + if (recursion_bug) { + static const char recursion_msg[] = + "BUG: recent printk recursion!"; + + recursion_bug = 0; + printed_len += strlen(recursion_msg); + /* emit KERN_CRIT message */ + log_store(0, 2, LOG_PREFIX|LOG_NEWLINE, 0, + NULL, 0, recursion_msg, printed_len); + } + + /* + * The printf needs to come first; we need the syslog + * prefix which might be passed-in as a parameter. + */ + text_len = vscnprintf(text, sizeof(textbuf), fmt, args); + + /* mark and strip a trailing newline */ + if (text_len && text[text_len-1] == '\n') { + text_len--; + lflags |= LOG_NEWLINE; + } + + /* strip kernel syslog prefix and extract log level or control flags */ + if (facility == 0) { + int kern_level = printk_get_level(text); + + if (kern_level) { + const char *end_of_header = printk_skip_level(text); + switch (kern_level) { + case '0' ... '7': + if (level == -1) + level = kern_level - '0'; + case 'd': /* KERN_DEFAULT */ + lflags |= LOG_PREFIX; + case 'c': /* KERN_CONT */ + break; + } + text_len -= end_of_header - text; + text = (char *)end_of_header; + } + } + + if (level == -1) + level = default_message_loglevel; + + if (dict) + lflags |= LOG_PREFIX|LOG_NEWLINE; + + if (!(lflags & LOG_NEWLINE)) { + /* + * Flush the conflicting buffer. An earlier newline was missing, + * or another task also prints continuation lines. + */ + if (cont.len && (lflags & LOG_PREFIX || cont.owner != current)) + cont_flush(LOG_NEWLINE); + + /* buffer line if possible, otherwise store it right away */ + if (!cont_add(facility, level, text, text_len)) + log_store(facility, level, lflags | LOG_CONT, 0, + dict, dictlen, text, text_len); + } else { + bool stored = false; + + /* + * If an earlier newline was missing and it was the same task, + * either merge it with the current buffer and flush, or if + * there was a race with interrupts (prefix == true) then just + * flush it out and store this line separately. + */ + if (cont.len && cont.owner == current) { + if (!(lflags & LOG_PREFIX)) + stored = cont_add(facility, level, text, text_len); + cont_flush(LOG_NEWLINE); + } + + if (!stored) + log_store(facility, level, lflags, 0, + dict, dictlen, text, text_len); + } + printed_len += text_len; + + /* + * Try to acquire and then immediately release the console semaphore. + * The release will print out buffers and wake up /dev/kmsg and syslog() + * users. + * + * The console_trylock_for_printk() function will release 'logbuf_lock' + * regardless of whether it actually gets the console semaphore or not. + */ + if (console_trylock_for_printk(this_cpu)) + console_unlock(); + + lockdep_on(); +out_restore_irqs: + local_irq_restore(flags); + + return printed_len; +} +EXPORT_SYMBOL(vprintk_emit); + +asmlinkage int vprintk(const char *fmt, va_list args) +{ + return vprintk_emit(0, -1, NULL, 0, fmt, args); +} +EXPORT_SYMBOL(vprintk); + +asmlinkage int printk_emit(int facility, int level, + const char *dict, size_t dictlen, + const char *fmt, ...) +{ + va_list args; + int r; + + va_start(args, fmt); + r = vprintk_emit(facility, level, dict, dictlen, fmt, args); + va_end(args); + + return r; +} +EXPORT_SYMBOL(printk_emit); + +/** + * printk - print a kernel message + * @fmt: format string + * + * This is printk(). It can be called from any context. We want it to work. + * + * We try to grab the console_lock. If we succeed, it's easy - we log the + * output and call the console drivers. If we fail to get the semaphore, we + * place the output into the log buffer and return. The current holder of + * the console_sem will notice the new output in console_unlock(); and will + * send it to the consoles before releasing the lock. + * + * One effect of this deferred printing is that code which calls printk() and + * then changes console_loglevel may break. This is because console_loglevel + * is inspected when the actual printing occurs. + * + * See also: + * printf(3) + * + * See the vsnprintf() documentation for format string extensions over C99. + */ +asmlinkage int printk(const char *fmt, ...) +{ + va_list args; + int r; + +#ifdef CONFIG_KGDB_KDB + if (unlikely(kdb_trap_printk)) { + va_start(args, fmt); + r = vkdb_printf(fmt, args); + va_end(args); + return r; + } +#endif + va_start(args, fmt); + r = vprintk_emit(0, -1, NULL, 0, fmt, args); + va_end(args); + + return r; +} +EXPORT_SYMBOL(printk); + +#else /* CONFIG_PRINTK */ + +#define LOG_LINE_MAX 0 +#define PREFIX_MAX 0 +#define LOG_LINE_MAX 0 +static u64 syslog_seq; +static u32 syslog_idx; +static u64 console_seq; +static u32 console_idx; +static enum log_flags syslog_prev; +static u64 log_first_seq; +static u32 log_first_idx; +static u64 log_next_seq; +static enum log_flags console_prev; +static struct cont { + size_t len; + size_t cons; + u8 level; + bool flushed:1; +} cont; +static struct log *log_from_idx(u32 idx) { return NULL; } +static u32 log_next(u32 idx) { return 0; } +static void call_console_drivers(int level, const char *text, size_t len) {} +static size_t msg_print_text(const struct log *msg, enum log_flags prev, + bool syslog, char *buf, size_t size) { return 0; } +static size_t cont_print_text(char *text, size_t size) { return 0; } + +#endif /* CONFIG_PRINTK */ + +#ifdef CONFIG_EARLY_PRINTK +struct console *early_console; + +void early_vprintk(const char *fmt, va_list ap) +{ + if (early_console) { + char buf[512]; + int n = vscnprintf(buf, sizeof(buf), fmt, ap); + + early_console->write(early_console, buf, n); + } +} + +asmlinkage void early_printk(const char *fmt, ...) +{ + va_list ap; + + va_start(ap, fmt); + early_vprintk(fmt, ap); + va_end(ap); +} +#endif + +static int __add_preferred_console(char *name, int idx, char *options, + char *brl_options) +{ + struct console_cmdline *c; + int i; + + /* + * See if this tty is not yet registered, and + * if we have a slot free. + */ + for (i = 0; i < MAX_CMDLINECONSOLES && console_cmdline[i].name[0]; i++) + if (strcmp(console_cmdline[i].name, name) == 0 && + console_cmdline[i].index == idx) { + if (!brl_options) + selected_console = i; + return 0; + } + if (i == MAX_CMDLINECONSOLES) + return -E2BIG; + if (!brl_options) + selected_console = i; + c = &console_cmdline[i]; + strlcpy(c->name, name, sizeof(c->name)); + c->options = options; +#ifdef CONFIG_A11Y_BRAILLE_CONSOLE + c->brl_options = brl_options; +#endif + c->index = idx; + return 0; +} +/* + * Set up a list of consoles. Called from init/main.c + */ +static int __init console_setup(char *str) +{ + char buf[sizeof(console_cmdline[0].name) + 4]; /* 4 for index */ + char *s, *options, *brl_options = NULL; + int idx; + +#ifdef CONFIG_A11Y_BRAILLE_CONSOLE + if (!memcmp(str, "brl,", 4)) { + brl_options = ""; + str += 4; + } else if (!memcmp(str, "brl=", 4)) { + brl_options = str + 4; + str = strchr(brl_options, ','); + if (!str) { + printk(KERN_ERR "need port name after brl=\n"); + return 1; + } + *(str++) = 0; + } +#endif + + /* + * Decode str into name, index, options. + */ + if (str[0] >= '0' && str[0] <= '9') { + strcpy(buf, "ttyS"); + strncpy(buf + 4, str, sizeof(buf) - 5); + } else { + strncpy(buf, str, sizeof(buf) - 1); + } + buf[sizeof(buf) - 1] = 0; + if ((options = strchr(str, ',')) != NULL) + *(options++) = 0; +#ifdef __sparc__ + if (!strcmp(str, "ttya")) + strcpy(buf, "ttyS0"); + if (!strcmp(str, "ttyb")) + strcpy(buf, "ttyS1"); +#endif + for (s = buf; *s; s++) + if ((*s >= '0' && *s <= '9') || *s == ',') + break; + idx = simple_strtoul(s, NULL, 10); + *s = 0; + + __add_preferred_console(buf, idx, options, brl_options); + console_set_on_cmdline = 1; + return 1; +} +__setup("console=", console_setup); + +/** + * add_preferred_console - add a device to the list of preferred consoles. + * @name: device name + * @idx: device index + * @options: options for this console + * + * The last preferred console added will be used for kernel messages + * and stdin/out/err for init. Normally this is used by console_setup + * above to handle user-supplied console arguments; however it can also + * be used by arch-specific code either to override the user or more + * commonly to provide a default console (ie from PROM variables) when + * the user has not supplied one. + */ +int add_preferred_console(char *name, int idx, char *options) +{ + return __add_preferred_console(name, idx, options, NULL); +} + +int update_console_cmdline(char *name, int idx, char *name_new, int idx_new, char *options) +{ + struct console_cmdline *c; + int i; + + for (i = 0; i < MAX_CMDLINECONSOLES && console_cmdline[i].name[0]; i++) + if (strcmp(console_cmdline[i].name, name) == 0 && + console_cmdline[i].index == idx) { + c = &console_cmdline[i]; + strlcpy(c->name, name_new, sizeof(c->name)); + c->name[sizeof(c->name) - 1] = 0; + c->options = options; + c->index = idx_new; + return i; + } + /* not found */ + return -1; +} + +bool console_suspend_enabled = 1; +EXPORT_SYMBOL(console_suspend_enabled); + +static int __init console_suspend_disable(char *str) +{ + console_suspend_enabled = 0; + return 1; +} +__setup("no_console_suspend", console_suspend_disable); +module_param_named(console_suspend, console_suspend_enabled, + bool, S_IRUGO | S_IWUSR); +MODULE_PARM_DESC(console_suspend, "suspend console during suspend" + " and hibernate operations"); + +/** + * suspend_console - suspend the console subsystem + * + * This disables printk() while we go into suspend states + */ +void suspend_console(void) +{ + if (!console_suspend_enabled) + return; + printk("Suspending console(s) (use no_console_suspend to debug)\n"); + console_lock(); + console_suspended = 1; + up(&console_sem); +} + +void resume_console(void) +{ + if (!console_suspend_enabled) + return; + down(&console_sem); + console_suspended = 0; + console_unlock(); +} + +/** + * console_cpu_notify - print deferred console messages after CPU hotplug + * @self: notifier struct + * @action: CPU hotplug event + * @hcpu: unused + * + * If printk() is called from a CPU that is not online yet, the messages + * will be spooled but will not show up on the console. This function is + * called when a new CPU comes online (or fails to come up), and ensures + * that any such output gets printed. + */ +static int console_cpu_notify(struct notifier_block *self, + unsigned long action, void *hcpu) +{ + switch (action) { + case CPU_ONLINE: + case CPU_DEAD: + case CPU_DOWN_FAILED: + case CPU_UP_CANCELED: + console_lock(); + console_unlock(); + } + return NOTIFY_OK; +} + +/** + * console_lock - lock the console system for exclusive use. + * + * Acquires a lock which guarantees that the caller has + * exclusive access to the console system and the console_drivers list. + * + * Can sleep, returns nothing. + */ +void console_lock(void) +{ + might_sleep(); + + down(&console_sem); + if (console_suspended) + return; + console_locked = 1; + console_may_schedule = 1; + mutex_acquire(&console_lock_dep_map, 0, 0, _RET_IP_); +} +EXPORT_SYMBOL(console_lock); + +/** + * console_trylock - try to lock the console system for exclusive use. + * + * Tried to acquire a lock which guarantees that the caller has + * exclusive access to the console system and the console_drivers list. + * + * returns 1 on success, and 0 on failure to acquire the lock. + */ +int console_trylock(void) +{ + if (down_trylock(&console_sem)) + return 0; + if (console_suspended) { + up(&console_sem); + return 0; + } + console_locked = 1; + console_may_schedule = 0; + mutex_acquire(&console_lock_dep_map, 0, 1, _RET_IP_); + return 1; +} +EXPORT_SYMBOL(console_trylock); + +int is_console_locked(void) +{ + return console_locked; +} + +static void console_cont_flush(char *text, size_t size) +{ + unsigned long flags; + size_t len; + + raw_spin_lock_irqsave(&logbuf_lock, flags); + + if (!cont.len) + goto out; + + /* + * We still queue earlier records, likely because the console was + * busy. The earlier ones need to be printed before this one, we + * did not flush any fragment so far, so just let it queue up. + */ + if (console_seq < log_next_seq && !cont.cons) + goto out; + + len = cont_print_text(text, size); + raw_spin_unlock(&logbuf_lock); + stop_critical_timings(); + call_console_drivers(cont.level, text, len); + start_critical_timings(); + local_irq_restore(flags); + return; +out: + raw_spin_unlock_irqrestore(&logbuf_lock, flags); +} + +/** + * console_unlock - unlock the console system + * + * Releases the console_lock which the caller holds on the console system + * and the console driver list. + * + * While the console_lock was held, console output may have been buffered + * by printk(). If this is the case, console_unlock(); emits + * the output prior to releasing the lock. + * + * If there is output waiting, we wake /dev/kmsg and syslog() users. + * + * console_unlock(); may be called from any context. + */ +void console_unlock(void) +{ + static char text[LOG_LINE_MAX + PREFIX_MAX]; + static u64 seen_seq; + unsigned long flags; + bool wake_klogd = false; + bool retry; + + if (console_suspended) { + up(&console_sem); + return; + } + + console_may_schedule = 0; + + /* flush buffered message fragment immediately to console */ + console_cont_flush(text, sizeof(text)); +again: + for (;;) { + struct log *msg; + size_t len; + int level; + + raw_spin_lock_irqsave(&logbuf_lock, flags); + if (seen_seq != log_next_seq) { + wake_klogd = true; + seen_seq = log_next_seq; + } + + if (console_seq < log_first_seq) { + /* messages are gone, move to first one */ + console_seq = log_first_seq; + console_idx = log_first_idx; + console_prev = 0; + } +skip: + if (console_seq == log_next_seq) + break; + + msg = log_from_idx(console_idx); + if (msg->flags & LOG_NOCONS) { + /* + * Skip record we have buffered and already printed + * directly to the console when we received it. + */ + console_idx = log_next(console_idx); + console_seq++; + /* + * We will get here again when we register a new + * CON_PRINTBUFFER console. Clear the flag so we + * will properly dump everything later. + */ + msg->flags &= ~LOG_NOCONS; + console_prev = msg->flags; + goto skip; + } + + level = msg->level; + len = msg_print_text(msg, console_prev, false, + text, sizeof(text)); + console_idx = log_next(console_idx); + console_seq++; + console_prev = msg->flags; + raw_spin_unlock(&logbuf_lock); + + stop_critical_timings(); /* don't trace print latency */ + call_console_drivers(level, text, len); + start_critical_timings(); + local_irq_restore(flags); + } + console_locked = 0; + mutex_release(&console_lock_dep_map, 1, _RET_IP_); + + /* Release the exclusive_console once it is used */ + if (unlikely(exclusive_console)) + exclusive_console = NULL; + + raw_spin_unlock(&logbuf_lock); + + up(&console_sem); + + /* + * Someone could have filled up the buffer again, so re-check if there's + * something to flush. In case we cannot trylock the console_sem again, + * there's a new owner and the console_unlock() from them will do the + * flush, no worries. + */ + raw_spin_lock(&logbuf_lock); + retry = console_seq != log_next_seq; + raw_spin_unlock_irqrestore(&logbuf_lock, flags); + + if (retry && console_trylock()) + goto again; + + if (wake_klogd) + wake_up_klogd(); +} +EXPORT_SYMBOL(console_unlock); + +/** + * console_conditional_schedule - yield the CPU if required + * + * If the console code is currently allowed to sleep, and + * if this CPU should yield the CPU to another task, do + * so here. + * + * Must be called within console_lock();. + */ +void __sched console_conditional_schedule(void) +{ + if (console_may_schedule) + cond_resched(); +} +EXPORT_SYMBOL(console_conditional_schedule); + +void console_unblank(void) +{ + struct console *c; + + /* + * console_unblank can no longer be called in interrupt context unless + * oops_in_progress is set to 1.. + */ + if (oops_in_progress) { + if (down_trylock(&console_sem) != 0) + return; + } else + console_lock(); + + console_locked = 1; + console_may_schedule = 0; + for_each_console(c) + if ((c->flags & CON_ENABLED) && c->unblank) + c->unblank(); + console_unlock(); +} + +/* + * Return the console tty driver structure and its associated index + */ +struct tty_driver *console_device(int *index) +{ + struct console *c; + struct tty_driver *driver = NULL; + + console_lock(); + for_each_console(c) { + if (!c->device) + continue; + driver = c->device(c, index); + if (driver) + break; + } + console_unlock(); + return driver; +} + +/* + * Prevent further output on the passed console device so that (for example) + * serial drivers can disable console output before suspending a port, and can + * re-enable output afterwards. + */ +void console_stop(struct console *console) +{ + console_lock(); + console->flags &= ~CON_ENABLED; + console_unlock(); +} +EXPORT_SYMBOL(console_stop); + +void console_start(struct console *console) +{ + console_lock(); + console->flags |= CON_ENABLED; + console_unlock(); +} +EXPORT_SYMBOL(console_start); + +static int __read_mostly keep_bootcon; + +static int __init keep_bootcon_setup(char *str) +{ + keep_bootcon = 1; + printk(KERN_INFO "debug: skip boot console de-registration.\n"); + + return 0; +} + +early_param("keep_bootcon", keep_bootcon_setup); + +/* + * The console driver calls this routine during kernel initialization + * to register the console printing procedure with printk() and to + * print any messages that were printed by the kernel before the + * console driver was initialized. + * + * This can happen pretty early during the boot process (because of + * early_printk) - sometimes before setup_arch() completes - be careful + * of what kernel features are used - they may not be initialised yet. + * + * There are two types of consoles - bootconsoles (early_printk) and + * "real" consoles (everything which is not a bootconsole) which are + * handled differently. + * - Any number of bootconsoles can be registered at any time. + * - As soon as a "real" console is registered, all bootconsoles + * will be unregistered automatically. + * - Once a "real" console is registered, any attempt to register a + * bootconsoles will be rejected + */ +void register_console(struct console *newcon) +{ + int i; + unsigned long flags; + struct console *bcon = NULL; + + /* + * before we register a new CON_BOOT console, make sure we don't + * already have a valid console + */ + if (console_drivers && newcon->flags & CON_BOOT) { + /* find the last or real console */ + for_each_console(bcon) { + if (!(bcon->flags & CON_BOOT)) { + printk(KERN_INFO "Too late to register bootconsole %s%d\n", + newcon->name, newcon->index); + return; + } + } + } + + if (console_drivers && console_drivers->flags & CON_BOOT) + bcon = console_drivers; + + if (preferred_console < 0 || bcon || !console_drivers) + preferred_console = selected_console; + + if (newcon->early_setup) + newcon->early_setup(); + + /* + * See if we want to use this console driver. If we + * didn't select a console we take the first one + * that registers here. + */ + if (preferred_console < 0) { + if (newcon->index < 0) + newcon->index = 0; + if (newcon->setup == NULL || + newcon->setup(newcon, NULL) == 0) { + newcon->flags |= CON_ENABLED; + if (newcon->device) { + newcon->flags |= CON_CONSDEV; + preferred_console = 0; + } + } + } + + /* + * See if this console matches one we selected on + * the command line. + */ + for (i = 0; i < MAX_CMDLINECONSOLES && console_cmdline[i].name[0]; + i++) { + if (strcmp(console_cmdline[i].name, newcon->name) != 0) + continue; + if (newcon->index >= 0 && + newcon->index != console_cmdline[i].index) + continue; + if (newcon->index < 0) + newcon->index = console_cmdline[i].index; +#ifdef CONFIG_A11Y_BRAILLE_CONSOLE + if (console_cmdline[i].brl_options) { + newcon->flags |= CON_BRL; + braille_register_console(newcon, + console_cmdline[i].index, + console_cmdline[i].options, + console_cmdline[i].brl_options); + return; + } +#endif + if (newcon->setup && + newcon->setup(newcon, console_cmdline[i].options) != 0) + break; + newcon->flags |= CON_ENABLED; + newcon->index = console_cmdline[i].index; + if (i == selected_console) { + newcon->flags |= CON_CONSDEV; + preferred_console = selected_console; + } + break; + } + + if (!(newcon->flags & CON_ENABLED)) + return; + + /* + * If we have a bootconsole, and are switching to a real console, + * don't print everything out again, since when the boot console, and + * the real console are the same physical device, it's annoying to + * see the beginning boot messages twice + */ + if (bcon && ((newcon->flags & (CON_CONSDEV | CON_BOOT)) == CON_CONSDEV)) + newcon->flags &= ~CON_PRINTBUFFER; + + /* + * Put this console in the list - keep the + * preferred driver at the head of the list. + */ + console_lock(); + if ((newcon->flags & CON_CONSDEV) || console_drivers == NULL) { + newcon->next = console_drivers; + console_drivers = newcon; + if (newcon->next) + newcon->next->flags &= ~CON_CONSDEV; + } else { + newcon->next = console_drivers->next; + console_drivers->next = newcon; + } + if (newcon->flags & CON_PRINTBUFFER) { + /* + * console_unlock(); will print out the buffered messages + * for us. + */ + raw_spin_lock_irqsave(&logbuf_lock, flags); + console_seq = syslog_seq; + console_idx = syslog_idx; + console_prev = syslog_prev; + raw_spin_unlock_irqrestore(&logbuf_lock, flags); + /* + * We're about to replay the log buffer. Only do this to the + * just-registered console to avoid excessive message spam to + * the already-registered consoles. + */ + exclusive_console = newcon; + } + console_unlock(); + console_sysfs_notify(); + + /* + * By unregistering the bootconsoles after we enable the real console + * we get the "console xxx enabled" message on all the consoles - + * boot consoles, real consoles, etc - this is to ensure that end + * users know there might be something in the kernel's log buffer that + * went to the bootconsole (that they do not see on the real console) + */ + if (bcon && + ((newcon->flags & (CON_CONSDEV | CON_BOOT)) == CON_CONSDEV) && + !keep_bootcon) { + /* we need to iterate through twice, to make sure we print + * everything out, before we unregister the console(s) + */ + printk(KERN_INFO "console [%s%d] enabled, bootconsole disabled\n", + newcon->name, newcon->index); + for_each_console(bcon) + if (bcon->flags & CON_BOOT) + unregister_console(bcon); + } else { + printk(KERN_INFO "%sconsole [%s%d] enabled\n", + (newcon->flags & CON_BOOT) ? "boot" : "" , + newcon->name, newcon->index); + } +} +EXPORT_SYMBOL(register_console); + +int unregister_console(struct console *console) +{ + struct console *a, *b; + int res = 1; + +#ifdef CONFIG_A11Y_BRAILLE_CONSOLE + if (console->flags & CON_BRL) + return braille_unregister_console(console); +#endif + + console_lock(); + if (console_drivers == console) { + console_drivers=console->next; + res = 0; + } else if (console_drivers) { + for (a=console_drivers->next, b=console_drivers ; + a; b=a, a=b->next) { + if (a == console) { + b->next = a->next; + res = 0; + break; + } + } + } + + /* + * If this isn't the last console and it has CON_CONSDEV set, we + * need to set it on the next preferred console. + */ + if (console_drivers != NULL && console->flags & CON_CONSDEV) + console_drivers->flags |= CON_CONSDEV; + + console_unlock(); + console_sysfs_notify(); + return res; +} +EXPORT_SYMBOL(unregister_console); + +static int __init printk_late_init(void) +{ + struct console *con; + + for_each_console(con) { + if (!keep_bootcon && con->flags & CON_BOOT) { + printk(KERN_INFO "turn off boot console %s%d\n", + con->name, con->index); + unregister_console(con); + } + } + hotcpu_notifier(console_cpu_notify, 0); + return 0; +} +late_initcall(printk_late_init); + +#if defined CONFIG_PRINTK +/* + * Delayed printk version, for scheduler-internal messages: + */ +#define PRINTK_BUF_SIZE 512 + +#define PRINTK_PENDING_WAKEUP 0x01 +#define PRINTK_PENDING_SCHED 0x02 + +static DEFINE_PER_CPU(int, printk_pending); +static DEFINE_PER_CPU(char [PRINTK_BUF_SIZE], printk_sched_buf); + +static void wake_up_klogd_work_func(struct irq_work *irq_work) +{ + int pending = __this_cpu_xchg(printk_pending, 0); + + if (pending & PRINTK_PENDING_SCHED) { + char *buf = __get_cpu_var(printk_sched_buf); + printk(KERN_WARNING "[sched_delayed] %s", buf); + } + + if (pending & PRINTK_PENDING_WAKEUP) + wake_up_interruptible(&log_wait); +} + +static DEFINE_PER_CPU(struct irq_work, wake_up_klogd_work) = { + .func = wake_up_klogd_work_func, + .flags = IRQ_WORK_LAZY, +}; + +void wake_up_klogd(void) +{ + preempt_disable(); + if (waitqueue_active(&log_wait)) { + this_cpu_or(printk_pending, PRINTK_PENDING_WAKEUP); + irq_work_queue(&__get_cpu_var(wake_up_klogd_work)); + } + preempt_enable(); +} + +int printk_sched(const char *fmt, ...) +{ + unsigned long flags; + va_list args; + char *buf; + int r; + + local_irq_save(flags); + buf = __get_cpu_var(printk_sched_buf); + + va_start(args, fmt); + r = vsnprintf(buf, PRINTK_BUF_SIZE, fmt, args); + va_end(args); + + __this_cpu_or(printk_pending, PRINTK_PENDING_SCHED); + irq_work_queue(&__get_cpu_var(wake_up_klogd_work)); + local_irq_restore(flags); + + return r; +} + +/* + * printk rate limiting, lifted from the networking subsystem. + * + * This enforces a rate limit: not more than 10 kernel messages + * every 5s to make a denial-of-service attack impossible. + */ +DEFINE_RATELIMIT_STATE(printk_ratelimit_state, 5 * HZ, 10); + +int __printk_ratelimit(const char *func) +{ + return ___ratelimit(&printk_ratelimit_state, func); +} +EXPORT_SYMBOL(__printk_ratelimit); + +/** + * printk_timed_ratelimit - caller-controlled printk ratelimiting + * @caller_jiffies: pointer to caller's state + * @interval_msecs: minimum interval between prints + * + * printk_timed_ratelimit() returns true if more than @interval_msecs + * milliseconds have elapsed since the last time printk_timed_ratelimit() + * returned true. + */ +bool printk_timed_ratelimit(unsigned long *caller_jiffies, + unsigned int interval_msecs) +{ + if (*caller_jiffies == 0 + || !time_in_range(jiffies, *caller_jiffies, + *caller_jiffies + + msecs_to_jiffies(interval_msecs))) { + *caller_jiffies = jiffies; + return true; + } + return false; +} +EXPORT_SYMBOL(printk_timed_ratelimit); + +static DEFINE_SPINLOCK(dump_list_lock); +static LIST_HEAD(dump_list); + +/** + * kmsg_dump_register - register a kernel log dumper. + * @dumper: pointer to the kmsg_dumper structure + * + * Adds a kernel log dumper to the system. The dump callback in the + * structure will be called when the kernel oopses or panics and must be + * set. Returns zero on success and %-EINVAL or %-EBUSY otherwise. + */ +int kmsg_dump_register(struct kmsg_dumper *dumper) +{ + unsigned long flags; + int err = -EBUSY; + + /* The dump callback needs to be set */ + if (!dumper->dump) + return -EINVAL; + + spin_lock_irqsave(&dump_list_lock, flags); + /* Don't allow registering multiple times */ + if (!dumper->registered) { + dumper->registered = 1; + list_add_tail_rcu(&dumper->list, &dump_list); + err = 0; + } + spin_unlock_irqrestore(&dump_list_lock, flags); + + return err; +} +EXPORT_SYMBOL_GPL(kmsg_dump_register); + +/** + * kmsg_dump_unregister - unregister a kmsg dumper. + * @dumper: pointer to the kmsg_dumper structure + * + * Removes a dump device from the system. Returns zero on success and + * %-EINVAL otherwise. + */ +int kmsg_dump_unregister(struct kmsg_dumper *dumper) +{ + unsigned long flags; + int err = -EINVAL; + + spin_lock_irqsave(&dump_list_lock, flags); + if (dumper->registered) { + dumper->registered = 0; + list_del_rcu(&dumper->list); + err = 0; + } + spin_unlock_irqrestore(&dump_list_lock, flags); + synchronize_rcu(); + + return err; +} +EXPORT_SYMBOL_GPL(kmsg_dump_unregister); + +static bool always_kmsg_dump; +module_param_named(always_kmsg_dump, always_kmsg_dump, bool, S_IRUGO | S_IWUSR); + +/** + * kmsg_dump - dump kernel log to kernel message dumpers. + * @reason: the reason (oops, panic etc) for dumping + * + * Call each of the registered dumper's dump() callback, which can + * retrieve the kmsg records with kmsg_dump_get_line() or + * kmsg_dump_get_buffer(). + */ +void kmsg_dump(enum kmsg_dump_reason reason) +{ + struct kmsg_dumper *dumper; + unsigned long flags; + + if ((reason > KMSG_DUMP_OOPS) && !always_kmsg_dump) + return; + + rcu_read_lock(); + list_for_each_entry_rcu(dumper, &dump_list, list) { + if (dumper->max_reason && reason > dumper->max_reason) + continue; + + /* initialize iterator with data about the stored records */ + dumper->active = true; + + raw_spin_lock_irqsave(&logbuf_lock, flags); + dumper->cur_seq = clear_seq; + dumper->cur_idx = clear_idx; + dumper->next_seq = log_next_seq; + dumper->next_idx = log_next_idx; + raw_spin_unlock_irqrestore(&logbuf_lock, flags); + + /* invoke dumper which will iterate over records */ + dumper->dump(dumper, reason); + + /* reset iterator */ + dumper->active = false; + } + rcu_read_unlock(); +} + +/** + * kmsg_dump_get_line_nolock - retrieve one kmsg log line (unlocked version) + * @dumper: registered kmsg dumper + * @syslog: include the "<4>" prefixes + * @line: buffer to copy the line to + * @size: maximum size of the buffer + * @len: length of line placed into buffer + * + * Start at the beginning of the kmsg buffer, with the oldest kmsg + * record, and copy one record into the provided buffer. + * + * Consecutive calls will return the next available record moving + * towards the end of the buffer with the youngest messages. + * + * A return value of FALSE indicates that there are no more records to + * read. + * + * The function is similar to kmsg_dump_get_line(), but grabs no locks. + */ +bool kmsg_dump_get_line_nolock(struct kmsg_dumper *dumper, bool syslog, + char *line, size_t size, size_t *len) +{ + struct log *msg; + size_t l = 0; + bool ret = false; + + if (!dumper->active) + goto out; + + if (dumper->cur_seq < log_first_seq) { + /* messages are gone, move to first available one */ + dumper->cur_seq = log_first_seq; + dumper->cur_idx = log_first_idx; + } + + /* last entry */ + if (dumper->cur_seq >= log_next_seq) + goto out; + + msg = log_from_idx(dumper->cur_idx); + l = msg_print_text(msg, 0, syslog, line, size); + + dumper->cur_idx = log_next(dumper->cur_idx); + dumper->cur_seq++; + ret = true; +out: + if (len) + *len = l; + return ret; +} + +/** + * kmsg_dump_get_line - retrieve one kmsg log line + * @dumper: registered kmsg dumper + * @syslog: include the "<4>" prefixes + * @line: buffer to copy the line to + * @size: maximum size of the buffer + * @len: length of line placed into buffer + * + * Start at the beginning of the kmsg buffer, with the oldest kmsg + * record, and copy one record into the provided buffer. + * + * Consecutive calls will return the next available record moving + * towards the end of the buffer with the youngest messages. + * + * A return value of FALSE indicates that there are no more records to + * read. + */ +bool kmsg_dump_get_line(struct kmsg_dumper *dumper, bool syslog, + char *line, size_t size, size_t *len) +{ + unsigned long flags; + bool ret; + + raw_spin_lock_irqsave(&logbuf_lock, flags); + ret = kmsg_dump_get_line_nolock(dumper, syslog, line, size, len); + raw_spin_unlock_irqrestore(&logbuf_lock, flags); + + return ret; +} +EXPORT_SYMBOL_GPL(kmsg_dump_get_line); + +/** + * kmsg_dump_get_buffer - copy kmsg log lines + * @dumper: registered kmsg dumper + * @syslog: include the "<4>" prefixes + * @buf: buffer to copy the line to + * @size: maximum size of the buffer + * @len: length of line placed into buffer + * + * Start at the end of the kmsg buffer and fill the provided buffer + * with as many of the the *youngest* kmsg records that fit into it. + * If the buffer is large enough, all available kmsg records will be + * copied with a single call. + * + * Consecutive calls will fill the buffer with the next block of + * available older records, not including the earlier retrieved ones. + * + * A return value of FALSE indicates that there are no more records to + * read. + */ +bool kmsg_dump_get_buffer(struct kmsg_dumper *dumper, bool syslog, + char *buf, size_t size, size_t *len) +{ + unsigned long flags; + u64 seq; + u32 idx; + u64 next_seq; + u32 next_idx; + enum log_flags prev; + size_t l = 0; + bool ret = false; + + if (!dumper->active) + goto out; + + raw_spin_lock_irqsave(&logbuf_lock, flags); + if (dumper->cur_seq < log_first_seq) { + /* messages are gone, move to first available one */ + dumper->cur_seq = log_first_seq; + dumper->cur_idx = log_first_idx; + } + + /* last entry */ + if (dumper->cur_seq >= dumper->next_seq) { + raw_spin_unlock_irqrestore(&logbuf_lock, flags); + goto out; + } + + /* calculate length of entire buffer */ + seq = dumper->cur_seq; + idx = dumper->cur_idx; + prev = 0; + while (seq < dumper->next_seq) { + struct log *msg = log_from_idx(idx); + + l += msg_print_text(msg, prev, true, NULL, 0); + idx = log_next(idx); + seq++; + prev = msg->flags; + } + + /* move first record forward until length fits into the buffer */ + seq = dumper->cur_seq; + idx = dumper->cur_idx; + prev = 0; + while (l > size && seq < dumper->next_seq) { + struct log *msg = log_from_idx(idx); + + l -= msg_print_text(msg, prev, true, NULL, 0); + idx = log_next(idx); + seq++; + prev = msg->flags; + } + + /* last message in next interation */ + next_seq = seq; + next_idx = idx; + + l = 0; + prev = 0; + while (seq < dumper->next_seq) { + struct log *msg = log_from_idx(idx); + + l += msg_print_text(msg, prev, syslog, buf + l, size - l); + idx = log_next(idx); + seq++; + prev = msg->flags; + } + + dumper->next_seq = next_seq; + dumper->next_idx = next_idx; + ret = true; + raw_spin_unlock_irqrestore(&logbuf_lock, flags); +out: + if (len) + *len = l; + return ret; +} +EXPORT_SYMBOL_GPL(kmsg_dump_get_buffer); + +/** + * kmsg_dump_rewind_nolock - reset the interator (unlocked version) + * @dumper: registered kmsg dumper + * + * Reset the dumper's iterator so that kmsg_dump_get_line() and + * kmsg_dump_get_buffer() can be called again and used multiple + * times within the same dumper.dump() callback. + * + * The function is similar to kmsg_dump_rewind(), but grabs no locks. + */ +void kmsg_dump_rewind_nolock(struct kmsg_dumper *dumper) +{ + dumper->cur_seq = clear_seq; + dumper->cur_idx = clear_idx; + dumper->next_seq = log_next_seq; + dumper->next_idx = log_next_idx; +} + +/** + * kmsg_dump_rewind - reset the interator + * @dumper: registered kmsg dumper + * + * Reset the dumper's iterator so that kmsg_dump_get_line() and + * kmsg_dump_get_buffer() can be called again and used multiple + * times within the same dumper.dump() callback. + */ +void kmsg_dump_rewind(struct kmsg_dumper *dumper) +{ + unsigned long flags; + + raw_spin_lock_irqsave(&logbuf_lock, flags); + kmsg_dump_rewind_nolock(dumper); + raw_spin_unlock_irqrestore(&logbuf_lock, flags); +} +EXPORT_SYMBOL_GPL(kmsg_dump_rewind); + +static char dump_stack_arch_desc_str[128]; + +/** + * dump_stack_set_arch_desc - set arch-specific str to show with task dumps + * @fmt: printf-style format string + * @...: arguments for the format string + * + * The configured string will be printed right after utsname during task + * dumps. Usually used to add arch-specific system identifiers. If an + * arch wants to make use of such an ID string, it should initialize this + * as soon as possible during boot. + */ +void __init dump_stack_set_arch_desc(const char *fmt, ...) +{ + va_list args; + + va_start(args, fmt); + vsnprintf(dump_stack_arch_desc_str, sizeof(dump_stack_arch_desc_str), + fmt, args); + va_end(args); +} + +/** + * dump_stack_print_info - print generic debug info for dump_stack() + * @log_lvl: log level + * + * Arch-specific dump_stack() implementations can use this function to + * print out the same debug information as the generic dump_stack(). + */ +void dump_stack_print_info(const char *log_lvl) +{ + printk("%sCPU: %d PID: %d Comm: %.20s %s %s %.*s\n", + log_lvl, raw_smp_processor_id(), current->pid, current->comm, + print_tainted(), init_utsname()->release, + (int)strcspn(init_utsname()->version, " "), + init_utsname()->version); + + if (dump_stack_arch_desc_str[0] != '\0') + printk("%sHardware name: %s\n", + log_lvl, dump_stack_arch_desc_str); + + print_worker_info(log_lvl, current); +} + +/** + * show_regs_print_info - print generic debug info for show_regs() + * @log_lvl: log level + * + * show_regs() implementations can use this function to print out generic + * debug information. + */ +void show_regs_print_info(const char *log_lvl) +{ + dump_stack_print_info(log_lvl); + + printk("%stask: %p ti: %p task.ti: %p\n", + log_lvl, current, current_thread_info(), + task_thread_info(current)); +} + +#endif -- cgit v1.2.3 From d197c43d04decb6b1298fa3ef26ea04a9ca7c977 Mon Sep 17 00:00:00 2001 From: Joe Perches Date: Wed, 31 Jul 2013 13:53:44 -0700 Subject: printk: add console_cmdline.h Add an include file for the console_cmdline struct so that the braille console driver can be separated. Signed-off-by: Joe Perches Cc: Samuel Thibault Cc: Ming Lei Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/printk/console_cmdline.h | 14 ++++++++++++++ kernel/printk/printk.c | 12 +++--------- 2 files changed, 17 insertions(+), 9 deletions(-) create mode 100644 kernel/printk/console_cmdline.h (limited to 'kernel') diff --git a/kernel/printk/console_cmdline.h b/kernel/printk/console_cmdline.h new file mode 100644 index 000000000000..cbd69d842341 --- /dev/null +++ b/kernel/printk/console_cmdline.h @@ -0,0 +1,14 @@ +#ifndef _CONSOLE_CMDLINE_H +#define _CONSOLE_CMDLINE_H + +struct console_cmdline +{ + char name[8]; /* Name of the driver */ + int index; /* Minor dev. to use */ + char *options; /* Options for the driver */ +#ifdef CONFIG_A11Y_BRAILLE_CONSOLE + char *brl_options; /* Options for braille driver */ +#endif +}; + +#endif diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c index 69b0890ed7e5..4da2b2c7f67d 100644 --- a/kernel/printk/printk.c +++ b/kernel/printk/printk.c @@ -51,6 +51,8 @@ #define CREATE_TRACE_POINTS #include +#include "console_cmdline.h" + /* printk's without a loglevel use this.. */ #define DEFAULT_MESSAGE_LOGLEVEL CONFIG_DEFAULT_MESSAGE_LOGLEVEL @@ -105,19 +107,11 @@ static struct console *exclusive_console; /* * Array of consoles built from command line options (console=) */ -struct console_cmdline -{ - char name[8]; /* Name of the driver */ - int index; /* Minor dev. to use */ - char *options; /* Options for the driver */ -#ifdef CONFIG_A11Y_BRAILLE_CONSOLE - char *brl_options; /* Options for braille driver */ -#endif -}; #define MAX_CMDLINECONSOLES 8 static struct console_cmdline console_cmdline[MAX_CMDLINECONSOLES]; + static int selected_console = -1; static int preferred_console = -1; int console_set_on_cmdline; -- cgit v1.2.3 From bbeddf52adc1b4207674ab88686cbbe58c24f721 Mon Sep 17 00:00:00 2001 From: Joe Perches Date: Wed, 31 Jul 2013 13:53:45 -0700 Subject: printk: move braille console support into separate braille.[ch] files Create files with prototypes and static inlines for braille support. Make braille_console functions return 1 on success. Corrected CONFIG_A11Y_BRAILLE_CONSOLE=n _braille_console_setup return value to NULL. Signed-off-by: Joe Perches Reviewed-by: Samuel Thibault Cc: Ming Lei Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/printk/Makefile | 1 + kernel/printk/braille.c | 48 ++++++++++++++++++++++++++++++++++++++++++++++++ kernel/printk/braille.h | 48 ++++++++++++++++++++++++++++++++++++++++++++++++ kernel/printk/printk.c | 44 +++++++++++++------------------------------- 4 files changed, 110 insertions(+), 31 deletions(-) create mode 100644 kernel/printk/braille.c create mode 100644 kernel/printk/braille.h (limited to 'kernel') diff --git a/kernel/printk/Makefile b/kernel/printk/Makefile index 36d306d9273c..85405bdcf2b3 100644 --- a/kernel/printk/Makefile +++ b/kernel/printk/Makefile @@ -1 +1,2 @@ obj-y = printk.o +obj-$(CONFIG_A11Y_BRAILLE_CONSOLE) += braille.o diff --git a/kernel/printk/braille.c b/kernel/printk/braille.c new file mode 100644 index 000000000000..b51087fb9ace --- /dev/null +++ b/kernel/printk/braille.c @@ -0,0 +1,48 @@ +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt + +#include +#include +#include + +#include "console_cmdline.h" +#include "braille.h" + +char *_braille_console_setup(char **str, char **brl_options) +{ + if (!memcmp(*str, "brl,", 4)) { + *brl_options = ""; + *str += 4; + } else if (!memcmp(str, "brl=", 4)) { + *brl_options = *str + 4; + *str = strchr(*brl_options, ','); + if (!*str) + pr_err("need port name after brl=\n"); + else + *((*str)++) = 0; + } + + return *str; +} + +int +_braille_register_console(struct console *console, struct console_cmdline *c) +{ + int rtn = 0; + + if (c->brl_options) { + console->flags |= CON_BRL; + rtn = braille_register_console(console, c->index, c->options, + c->brl_options); + } + + return rtn; +} + +int +_braille_unregister_console(struct console *console) +{ + if (console->flags & CON_BRL) + return braille_unregister_console(console); + + return 0; +} diff --git a/kernel/printk/braille.h b/kernel/printk/braille.h new file mode 100644 index 000000000000..769d771145c8 --- /dev/null +++ b/kernel/printk/braille.h @@ -0,0 +1,48 @@ +#ifndef _PRINTK_BRAILLE_H +#define _PRINTK_BRAILLE_H + +#ifdef CONFIG_A11Y_BRAILLE_CONSOLE + +static inline void +braille_set_options(struct console_cmdline *c, char *brl_options) +{ + c->brl_options = brl_options; +} + +char * +_braille_console_setup(char **str, char **brl_options); + +int +_braille_register_console(struct console *console, struct console_cmdline *c); + +int +_braille_unregister_console(struct console *console); + +#else + +static inline void +braille_set_options(struct console_cmdline *c, char *brl_options) +{ +} + +static inline char * +_braille_console_setup(char **str, char **brl_options) +{ + return NULL; +} + +static inline int +_braille_register_console(struct console *console, struct console_cmdline *c) +{ + return 0; +} + +static inline int +_braille_unregister_console(struct console *console) +{ + return 0; +} + +#endif + +#endif diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c index 4da2b2c7f67d..5a022e0c654c 100644 --- a/kernel/printk/printk.c +++ b/kernel/printk/printk.c @@ -52,6 +52,7 @@ #include #include "console_cmdline.h" +#include "braille.h" /* printk's without a loglevel use this.. */ #define DEFAULT_MESSAGE_LOGLEVEL CONFIG_DEFAULT_MESSAGE_LOGLEVEL @@ -1769,9 +1770,8 @@ static int __add_preferred_console(char *name, int idx, char *options, c = &console_cmdline[i]; strlcpy(c->name, name, sizeof(c->name)); c->options = options; -#ifdef CONFIG_A11Y_BRAILLE_CONSOLE - c->brl_options = brl_options; -#endif + braille_set_options(c, brl_options); + c->index = idx; return 0; } @@ -1784,20 +1784,8 @@ static int __init console_setup(char *str) char *s, *options, *brl_options = NULL; int idx; -#ifdef CONFIG_A11Y_BRAILLE_CONSOLE - if (!memcmp(str, "brl,", 4)) { - brl_options = ""; - str += 4; - } else if (!memcmp(str, "brl=", 4)) { - brl_options = str + 4; - str = strchr(brl_options, ','); - if (!str) { - printk(KERN_ERR "need port name after brl=\n"); - return 1; - } - *(str++) = 0; - } -#endif + if (_braille_console_setup(&str, &brl_options)) + return 1; /* * Decode str into name, index, options. @@ -2291,16 +2279,10 @@ void register_console(struct console *newcon) continue; if (newcon->index < 0) newcon->index = console_cmdline[i].index; -#ifdef CONFIG_A11Y_BRAILLE_CONSOLE - if (console_cmdline[i].brl_options) { - newcon->flags |= CON_BRL; - braille_register_console(newcon, - console_cmdline[i].index, - console_cmdline[i].options, - console_cmdline[i].brl_options); + + if (_braille_register_console(newcon, &console_cmdline[i])) return; - } -#endif + if (newcon->setup && newcon->setup(newcon, console_cmdline[i].options) != 0) break; @@ -2388,13 +2370,13 @@ EXPORT_SYMBOL(register_console); int unregister_console(struct console *console) { struct console *a, *b; - int res = 1; + int res; -#ifdef CONFIG_A11Y_BRAILLE_CONSOLE - if (console->flags & CON_BRL) - return braille_unregister_console(console); -#endif + res = _braille_unregister_console(console); + if (res) + return res; + res = 1; console_lock(); if (console_drivers == console) { console_drivers=console->next; -- cgit v1.2.3 From 23475408c618ecd5b44b7e069fd65ec73d17d9f0 Mon Sep 17 00:00:00 2001 From: Joe Perches Date: Wed, 31 Jul 2013 13:53:46 -0700 Subject: printk: use pointer for console_cmdline indexing Make the code a bit more compact by always using a pointer for the active console_cmdline. Move overly indented code to correct indent level. Signed-off-by: Joe Perches Cc: Samuel Thibault Cc: Ming Lei Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/printk/printk.c | 49 ++++++++++++++++++++++++++----------------------- 1 file changed, 26 insertions(+), 23 deletions(-) (limited to 'kernel') diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c index 5a022e0c654c..8f1fb50aa3ce 100644 --- a/kernel/printk/printk.c +++ b/kernel/printk/printk.c @@ -1756,18 +1756,19 @@ static int __add_preferred_console(char *name, int idx, char *options, * See if this tty is not yet registered, and * if we have a slot free. */ - for (i = 0; i < MAX_CMDLINECONSOLES && console_cmdline[i].name[0]; i++) - if (strcmp(console_cmdline[i].name, name) == 0 && - console_cmdline[i].index == idx) { - if (!brl_options) - selected_console = i; - return 0; + for (i = 0, c = console_cmdline; + i < MAX_CMDLINECONSOLES && c->name[0]; + i++, c++) { + if (strcmp(c->name, name) == 0 && c->index == idx) { + if (!brl_options) + selected_console = i; + return 0; } + } if (i == MAX_CMDLINECONSOLES) return -E2BIG; if (!brl_options) selected_console = i; - c = &console_cmdline[i]; strlcpy(c->name, name, sizeof(c->name)); c->options = options; braille_set_options(c, brl_options); @@ -1840,15 +1841,15 @@ int update_console_cmdline(char *name, int idx, char *name_new, int idx_new, cha struct console_cmdline *c; int i; - for (i = 0; i < MAX_CMDLINECONSOLES && console_cmdline[i].name[0]; i++) - if (strcmp(console_cmdline[i].name, name) == 0 && - console_cmdline[i].index == idx) { - c = &console_cmdline[i]; - strlcpy(c->name, name_new, sizeof(c->name)); - c->name[sizeof(c->name) - 1] = 0; - c->options = options; - c->index = idx_new; - return i; + for (i = 0, c = console_cmdline; + i < MAX_CMDLINECONSOLES && c->name[0]; + i++, c++) + if (strcmp(c->name, name) == 0 && c->index == idx) { + strlcpy(c->name, name_new, sizeof(c->name)); + c->name[sizeof(c->name) - 1] = 0; + c->options = options; + c->index = idx_new; + return i; } /* not found */ return -1; @@ -2223,6 +2224,7 @@ void register_console(struct console *newcon) int i; unsigned long flags; struct console *bcon = NULL; + struct console_cmdline *c; /* * before we register a new CON_BOOT console, make sure we don't @@ -2270,24 +2272,25 @@ void register_console(struct console *newcon) * See if this console matches one we selected on * the command line. */ - for (i = 0; i < MAX_CMDLINECONSOLES && console_cmdline[i].name[0]; - i++) { - if (strcmp(console_cmdline[i].name, newcon->name) != 0) + for (i = 0, c = console_cmdline; + i < MAX_CMDLINECONSOLES && c->name[0]; + i++, c++) { + if (strcmp(c->name, newcon->name) != 0) continue; if (newcon->index >= 0 && - newcon->index != console_cmdline[i].index) + newcon->index != c->index) continue; if (newcon->index < 0) - newcon->index = console_cmdline[i].index; + newcon->index = c->index; - if (_braille_register_console(newcon, &console_cmdline[i])) + if (_braille_register_console(newcon, c)) return; if (newcon->setup && newcon->setup(newcon, console_cmdline[i].options) != 0) break; newcon->flags |= CON_ENABLED; - newcon->index = console_cmdline[i].index; + newcon->index = c->index; if (i == selected_console) { newcon->flags |= CON_CONSDEV; preferred_console = selected_console; -- cgit v1.2.3 From 62e32ac3505a0cab1c5ef8ea2c0eab3b26ed855f Mon Sep 17 00:00:00 2001 From: Joe Perches Date: Wed, 31 Jul 2013 13:53:47 -0700 Subject: printk: rename struct log to struct printk_log Rename the struct to enable moving portions of printk.c to separate files. The rename changes output of /proc/vmcoreinfo. Signed-off-by: Joe Perches Cc: Samuel Thibault Cc: Ming Lei Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/printk/printk.c | 80 +++++++++++++++++++++++++------------------------- 1 file changed, 40 insertions(+), 40 deletions(-) (limited to 'kernel') diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c index 8f1fb50aa3ce..5b5a7080e2a5 100644 --- a/kernel/printk/printk.c +++ b/kernel/printk/printk.c @@ -173,7 +173,7 @@ static int console_may_schedule; * 67 "g" * 0032 00 00 00 padding to next message header * - * The 'struct log' buffer header must never be directly exported to + * The 'struct printk_log' buffer header must never be directly exported to * userspace, it is a kernel-private implementation detail that might * need to be changed in the future, when the requirements change. * @@ -195,7 +195,7 @@ enum log_flags { LOG_CONT = 8, /* text is a fragment of a continuation line */ }; -struct log { +struct printk_log { u64 ts_nsec; /* timestamp in nanoseconds */ u16 len; /* length of entire record */ u16 text_len; /* length of text buffer */ @@ -243,7 +243,7 @@ static u32 clear_idx; #if defined(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS) #define LOG_ALIGN 4 #else -#define LOG_ALIGN __alignof__(struct log) +#define LOG_ALIGN __alignof__(struct printk_log) #endif #define __LOG_BUF_LEN (1 << CONFIG_LOG_BUF_SHIFT) static char __log_buf[__LOG_BUF_LEN] __aligned(LOG_ALIGN); @@ -254,35 +254,35 @@ static u32 log_buf_len = __LOG_BUF_LEN; static volatile unsigned int logbuf_cpu = UINT_MAX; /* human readable text of the record */ -static char *log_text(const struct log *msg) +static char *log_text(const struct printk_log *msg) { - return (char *)msg + sizeof(struct log); + return (char *)msg + sizeof(struct printk_log); } /* optional key/value pair dictionary attached to the record */ -static char *log_dict(const struct log *msg) +static char *log_dict(const struct printk_log *msg) { - return (char *)msg + sizeof(struct log) + msg->text_len; + return (char *)msg + sizeof(struct printk_log) + msg->text_len; } /* get record by index; idx must point to valid msg */ -static struct log *log_from_idx(u32 idx) +static struct printk_log *log_from_idx(u32 idx) { - struct log *msg = (struct log *)(log_buf + idx); + struct printk_log *msg = (struct printk_log *)(log_buf + idx); /* * A length == 0 record is the end of buffer marker. Wrap around and * read the message at the start of the buffer. */ if (!msg->len) - return (struct log *)log_buf; + return (struct printk_log *)log_buf; return msg; } /* get next record; idx must point to valid msg */ static u32 log_next(u32 idx) { - struct log *msg = (struct log *)(log_buf + idx); + struct printk_log *msg = (struct printk_log *)(log_buf + idx); /* length == 0 indicates the end of the buffer; wrap */ /* @@ -291,7 +291,7 @@ static u32 log_next(u32 idx) * return the one after that. */ if (!msg->len) { - msg = (struct log *)log_buf; + msg = (struct printk_log *)log_buf; return msg->len; } return idx + msg->len; @@ -303,11 +303,11 @@ static void log_store(int facility, int level, const char *dict, u16 dict_len, const char *text, u16 text_len) { - struct log *msg; + struct printk_log *msg; u32 size, pad_len; /* number of '\0' padding bytes to next message */ - size = sizeof(struct log) + text_len + dict_len; + size = sizeof(struct printk_log) + text_len + dict_len; pad_len = (-size) & (LOG_ALIGN - 1); size += pad_len; @@ -319,7 +319,7 @@ static void log_store(int facility, int level, else free = log_first_idx - log_next_idx; - if (free > size + sizeof(struct log)) + if (free > size + sizeof(struct printk_log)) break; /* drop old messages until we have enough contiuous space */ @@ -327,18 +327,18 @@ static void log_store(int facility, int level, log_first_seq++; } - if (log_next_idx + size + sizeof(struct log) >= log_buf_len) { + if (log_next_idx + size + sizeof(struct printk_log) >= log_buf_len) { /* * This message + an additional empty header does not fit * at the end of the buffer. Add an empty header with len == 0 * to signify a wrap around. */ - memset(log_buf + log_next_idx, 0, sizeof(struct log)); + memset(log_buf + log_next_idx, 0, sizeof(struct printk_log)); log_next_idx = 0; } /* fill message */ - msg = (struct log *)(log_buf + log_next_idx); + msg = (struct printk_log *)(log_buf + log_next_idx); memcpy(log_text(msg), text, text_len); msg->text_len = text_len; memcpy(log_dict(msg), dict, dict_len); @@ -351,7 +351,7 @@ static void log_store(int facility, int level, else msg->ts_nsec = local_clock(); memset(log_dict(msg) + dict_len, 0, pad_len); - msg->len = sizeof(struct log) + text_len + dict_len + pad_len; + msg->len = sizeof(struct printk_log) + text_len + dict_len + pad_len; /* insert message */ log_next_idx += msg->len; @@ -474,7 +474,7 @@ static ssize_t devkmsg_read(struct file *file, char __user *buf, size_t count, loff_t *ppos) { struct devkmsg_user *user = file->private_data; - struct log *msg; + struct printk_log *msg; u64 ts_usec; size_t i; char cont = '-'; @@ -719,14 +719,14 @@ void log_buf_kexec_setup(void) VMCOREINFO_SYMBOL(log_first_idx); VMCOREINFO_SYMBOL(log_next_idx); /* - * Export struct log size and field offsets. User space tools can + * Export struct printk_log size and field offsets. User space tools can * parse it and detect any changes to structure down the line. */ - VMCOREINFO_STRUCT_SIZE(log); - VMCOREINFO_OFFSET(log, ts_nsec); - VMCOREINFO_OFFSET(log, len); - VMCOREINFO_OFFSET(log, text_len); - VMCOREINFO_OFFSET(log, dict_len); + VMCOREINFO_STRUCT_SIZE(printk_log); + VMCOREINFO_OFFSET(printk_log, ts_nsec); + VMCOREINFO_OFFSET(printk_log, len); + VMCOREINFO_OFFSET(printk_log, text_len); + VMCOREINFO_OFFSET(printk_log, dict_len); } #endif @@ -879,7 +879,7 @@ static size_t print_time(u64 ts, char *buf) (unsigned long)ts, rem_nsec / 1000); } -static size_t print_prefix(const struct log *msg, bool syslog, char *buf) +static size_t print_prefix(const struct printk_log *msg, bool syslog, char *buf) { size_t len = 0; unsigned int prefix = (msg->facility << 3) | msg->level; @@ -902,7 +902,7 @@ static size_t print_prefix(const struct log *msg, bool syslog, char *buf) return len; } -static size_t msg_print_text(const struct log *msg, enum log_flags prev, +static size_t msg_print_text(const struct printk_log *msg, enum log_flags prev, bool syslog, char *buf, size_t size) { const char *text = log_text(msg); @@ -964,7 +964,7 @@ static size_t msg_print_text(const struct log *msg, enum log_flags prev, static int syslog_print(char __user *buf, int size) { char *text; - struct log *msg; + struct printk_log *msg; int len = 0; text = kmalloc(LOG_LINE_MAX + PREFIX_MAX, GFP_KERNEL); @@ -1055,7 +1055,7 @@ static int syslog_print_all(char __user *buf, int size, bool clear) idx = clear_idx; prev = 0; while (seq < log_next_seq) { - struct log *msg = log_from_idx(idx); + struct printk_log *msg = log_from_idx(idx); len += msg_print_text(msg, prev, true, NULL, 0); prev = msg->flags; @@ -1068,7 +1068,7 @@ static int syslog_print_all(char __user *buf, int size, bool clear) idx = clear_idx; prev = 0; while (len > size && seq < log_next_seq) { - struct log *msg = log_from_idx(idx); + struct printk_log *msg = log_from_idx(idx); len -= msg_print_text(msg, prev, true, NULL, 0); prev = msg->flags; @@ -1082,7 +1082,7 @@ static int syslog_print_all(char __user *buf, int size, bool clear) len = 0; prev = 0; while (len >= 0 && seq < next_seq) { - struct log *msg = log_from_idx(idx); + struct printk_log *msg = log_from_idx(idx); int textlen; textlen = msg_print_text(msg, prev, true, text, @@ -1228,7 +1228,7 @@ int do_syslog(int type, char __user *buf, int len, bool from_file) error = 0; while (seq < log_next_seq) { - struct log *msg = log_from_idx(idx); + struct printk_log *msg = log_from_idx(idx); error += msg_print_text(msg, prev, true, NULL, 0); idx = log_next(idx); @@ -1714,10 +1714,10 @@ static struct cont { u8 level; bool flushed:1; } cont; -static struct log *log_from_idx(u32 idx) { return NULL; } +static struct printk_log *log_from_idx(u32 idx) { return NULL; } static u32 log_next(u32 idx) { return 0; } static void call_console_drivers(int level, const char *text, size_t len) {} -static size_t msg_print_text(const struct log *msg, enum log_flags prev, +static size_t msg_print_text(const struct printk_log *msg, enum log_flags prev, bool syslog, char *buf, size_t size) { return 0; } static size_t cont_print_text(char *text, size_t size) { return 0; } @@ -2029,7 +2029,7 @@ void console_unlock(void) console_cont_flush(text, sizeof(text)); again: for (;;) { - struct log *msg; + struct printk_log *msg; size_t len; int level; @@ -2645,7 +2645,7 @@ void kmsg_dump(enum kmsg_dump_reason reason) bool kmsg_dump_get_line_nolock(struct kmsg_dumper *dumper, bool syslog, char *line, size_t size, size_t *len) { - struct log *msg; + struct printk_log *msg; size_t l = 0; bool ret = false; @@ -2757,7 +2757,7 @@ bool kmsg_dump_get_buffer(struct kmsg_dumper *dumper, bool syslog, idx = dumper->cur_idx; prev = 0; while (seq < dumper->next_seq) { - struct log *msg = log_from_idx(idx); + struct printk_log *msg = log_from_idx(idx); l += msg_print_text(msg, prev, true, NULL, 0); idx = log_next(idx); @@ -2770,7 +2770,7 @@ bool kmsg_dump_get_buffer(struct kmsg_dumper *dumper, bool syslog, idx = dumper->cur_idx; prev = 0; while (l > size && seq < dumper->next_seq) { - struct log *msg = log_from_idx(idx); + struct printk_log *msg = log_from_idx(idx); l -= msg_print_text(msg, prev, true, NULL, 0); idx = log_next(idx); @@ -2785,7 +2785,7 @@ bool kmsg_dump_get_buffer(struct kmsg_dumper *dumper, bool syslog, l = 0; prev = 0; while (seq < dumper->next_seq) { - struct log *msg = log_from_idx(idx); + struct printk_log *msg = log_from_idx(idx); l += msg_print_text(msg, prev, syslog, buf + l, size - l); idx = log_next(idx); -- cgit v1.2.3 From 40c32592668b727cbfcf7b1c0567f581bd62a5e4 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Wed, 3 Jul 2013 23:33:50 -0400 Subject: tracing/kprobes: Fail to unregister if probe event files are in use When a probe is being removed, it cleans up the event files that correspond to the probe. But there is a race between writing to one of these files and deleting the probe. This is especially true for the "enable" file. CPU 0 CPU 1 ----- ----- fd = open("enable",O_WRONLY); probes_open() release_all_trace_probes() unregister_trace_probe() if (trace_probe_is_enabled(tp)) return -EBUSY write(fd, "1", 1) __ftrace_set_clr_event() call->class->reg() (kprobe_register) enable_trace_probe(tp) __unregister_trace_probe(tp); list_del(&tp->list) unregister_probe_event(tp) <-- fails! free_trace_probe(tp) write(fd, "0", 1) __ftrace_set_clr_event() call->class->unreg (kprobe_register) disable_trace_probe(tp) <-- BOOM! A test program was written that used two threads to simulate the above scenario adding a nanosleep() interval to change the timings and after several thousand runs, it was able to trigger this bug and crash: BUG: unable to handle kernel paging request at 00000005000000f9 IP: [] probes_open+0x3b/0xa7 PGD 7808a067 PUD 0 Oops: 0000 [#1] PREEMPT SMP Dumping ftrace buffer: --------------------------------- Modules linked in: ipt_MASQUERADE sunrpc ip6t_REJECT nf_conntrack_ipv6 CPU: 1 PID: 2070 Comm: test-kprobe-rem Not tainted 3.11.0-rc3-test+ #47 Hardware name: To Be Filled By O.E.M. To Be Filled By O.E.M./To be filled by O.E.M., BIOS SDBLI944.86P 05/08/2007 task: ffff880077756440 ti: ffff880076e52000 task.ti: ffff880076e52000 RIP: 0010:[] [] probes_open+0x3b/0xa7 RSP: 0018:ffff880076e53c38 EFLAGS: 00010203 RAX: 0000000500000001 RBX: ffff88007844f440 RCX: 0000000000000003 RDX: 0000000000000003 RSI: 0000000000000003 RDI: ffff880076e52000 RBP: ffff880076e53c58 R08: ffff880076e53bd8 R09: 0000000000000000 R10: ffff880077756440 R11: 0000000000000006 R12: ffffffff810dee35 R13: ffff880079250418 R14: 0000000000000000 R15: ffff88007844f450 FS: 00007f87a276f700(0000) GS:ffff88007d480000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b CR2: 00000005000000f9 CR3: 0000000077262000 CR4: 00000000000007e0 Stack: ffff880076e53c58 ffffffff81219ea0 ffff88007844f440 ffffffff810dee35 ffff880076e53ca8 ffffffff81130f78 ffff8800772986c0 ffff8800796f93a0 ffffffff81d1b5d8 ffff880076e53e04 0000000000000000 ffff88007844f440 Call Trace: [] ? security_file_open+0x2c/0x30 [] ? unregister_trace_probe+0x4b/0x4b [] do_dentry_open+0x162/0x226 [] finish_open+0x46/0x54 [] do_last+0x7f6/0x996 [] ? inode_permission+0x42/0x44 [] path_openat+0x232/0x496 [] do_filp_open+0x3a/0x8a [] ? __alloc_fd+0x168/0x17a [] do_sys_open+0x70/0x102 [] ? trace_hardirqs_on_caller+0x160/0x197 [] SyS_open+0x1e/0x20 [] system_call_fastpath+0x16/0x1b Code: e5 41 54 53 48 89 f3 48 83 ec 10 48 23 56 78 48 39 c2 75 6c 31 f6 48 c7 RIP [] probes_open+0x3b/0xa7 RSP CR2: 00000005000000f9 ---[ end trace 35f17d68fc569897 ]--- The unregister_trace_probe() must be done first, and if it fails it must fail the removal of the kprobe. Several changes have already been made by Oleg Nesterov and Masami Hiramatsu to allow moving the unregister_probe_event() before the removal of the probe and exit the function if it fails. This prevents the tp structure from being used after it is freed. Link: http://lkml.kernel.org/r/20130704034038.819592356@goodmis.org Acked-by: Masami Hiramatsu Signed-off-by: Steven Rostedt --- kernel/trace/trace_kprobe.c | 21 +++++++++++++++------ 1 file changed, 15 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 3811487e7a7a..243f6834d026 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -95,7 +95,7 @@ static __kprobes bool trace_probe_is_on_module(struct trace_probe *tp) } static int register_probe_event(struct trace_probe *tp); -static void unregister_probe_event(struct trace_probe *tp); +static int unregister_probe_event(struct trace_probe *tp); static DEFINE_MUTEX(probe_lock); static LIST_HEAD(probe_list); @@ -351,9 +351,12 @@ static int unregister_trace_probe(struct trace_probe *tp) if (trace_probe_is_enabled(tp)) return -EBUSY; + /* Will fail if probe is being used by ftrace or perf */ + if (unregister_probe_event(tp)) + return -EBUSY; + __unregister_trace_probe(tp); list_del(&tp->list); - unregister_probe_event(tp); return 0; } @@ -632,7 +635,9 @@ static int release_all_trace_probes(void) /* TODO: Use batch unregistration */ while (!list_empty(&probe_list)) { tp = list_entry(probe_list.next, struct trace_probe, list); - unregister_trace_probe(tp); + ret = unregister_trace_probe(tp); + if (ret) + goto end; free_trace_probe(tp); } @@ -1247,11 +1252,15 @@ static int register_probe_event(struct trace_probe *tp) return ret; } -static void unregister_probe_event(struct trace_probe *tp) +static int unregister_probe_event(struct trace_probe *tp) { + int ret; + /* tp->event is unregistered in trace_remove_event_call() */ - trace_remove_event_call(&tp->call); - kfree(tp->call.print_fmt); + ret = trace_remove_event_call(&tp->call); + if (!ret) + kfree(tp->call.print_fmt); + return ret; } /* Make a debugfs interface for controlling probe points */ -- cgit v1.2.3 From 2865a8fb44cc32420407362cbda80c10fa09c6b2 Mon Sep 17 00:00:00 2001 From: Shaohua Li Date: Thu, 1 Aug 2013 09:56:36 +0800 Subject: workqueue: copy workqueue_attrs with all fields $echo '0' > /sys/bus/workqueue/devices/xxx/numa $cat /sys/bus/workqueue/devices/xxx/numa I got 1. It should be 0, the reason is copy_workqueue_attrs() called in apply_workqueue_attrs() doesn't copy no_numa field. Fix it by making copy_workqueue_attrs() copy ->no_numa too. This would also make get_unbound_pool() set a pool's ->no_numa attribute according to the workqueue attributes used when the pool was created. While harmelss, as ->no_numa isn't a pool attribute, this is a bit confusing. Clear it explicitly. tj: Updated description and comments a bit. Signed-off-by: Shaohua Li Signed-off-by: Tejun Heo Cc: stable@vger.kernel.org --- kernel/workqueue.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 55f5f0afcd0d..726adc84b3ca 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -3416,6 +3416,12 @@ static void copy_workqueue_attrs(struct workqueue_attrs *to, { to->nice = from->nice; cpumask_copy(to->cpumask, from->cpumask); + /* + * Unlike hash and equality test, this function doesn't ignore + * ->no_numa as it is used for both pool and wq attrs. Instead, + * get_unbound_pool() explicitly clears ->no_numa after copying. + */ + to->no_numa = from->no_numa; } /* hash value of the content of @attr */ @@ -3583,6 +3589,12 @@ static struct worker_pool *get_unbound_pool(const struct workqueue_attrs *attrs) lockdep_set_subclass(&pool->lock, 1); /* see put_pwq() */ copy_workqueue_attrs(pool->attrs, attrs); + /* + * no_numa isn't a worker_pool attribute, always clear it. See + * 'struct workqueue_attrs' comments for detail. + */ + pool->attrs->no_numa = false; + /* if cpumask is contained inside a NUMA node, we belong to that node */ if (wq_numa_enabled) { for_each_node(node) { -- cgit v1.2.3 From 876ede8b2b9880615be0de3ec7b8afd0a1786e76 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Thu, 1 Aug 2013 09:51:47 +0800 Subject: cgroup: restructure the failure path in cgroup_write_event_control() It uses a single label and checks the validity of each pointer. This is err-prone, and actually we had a bug because one of the check was insufficient. Use multi lables as we do in other places. v2: - drop initializations of local variables. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cgroup.c | 39 ++++++++++++++++++--------------------- 1 file changed, 18 insertions(+), 21 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 3f6593333525..9f6dab22289e 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -3934,11 +3934,11 @@ static void cgroup_event_ptable_queue_proc(struct file *file, static int cgroup_write_event_control(struct cgroup *cgrp, struct cftype *cft, const char *buffer) { - struct cgroup_event *event = NULL; + struct cgroup_event *event; struct cgroup *cgrp_cfile; unsigned int efd, cfd; - struct file *efile = NULL; - struct file *cfile = NULL; + struct file *efile; + struct file *cfile; char *endp; int ret; @@ -3964,31 +3964,31 @@ static int cgroup_write_event_control(struct cgroup *cgrp, struct cftype *cft, efile = eventfd_fget(efd); if (IS_ERR(efile)) { ret = PTR_ERR(efile); - goto fail; + goto out_kfree; } event->eventfd = eventfd_ctx_fileget(efile); if (IS_ERR(event->eventfd)) { ret = PTR_ERR(event->eventfd); - goto fail; + goto out_put_efile; } cfile = fget(cfd); if (!cfile) { ret = -EBADF; - goto fail; + goto out_put_eventfd; } /* the process need read permission on control file */ /* AV: shouldn't we check that it's been opened for read instead? */ ret = inode_permission(file_inode(cfile), MAY_READ); if (ret < 0) - goto fail; + goto out_put_cfile; event->cft = __file_cft(cfile); if (IS_ERR(event->cft)) { ret = PTR_ERR(event->cft); - goto fail; + goto out_put_cfile; } /* @@ -3998,18 +3998,18 @@ static int cgroup_write_event_control(struct cgroup *cgrp, struct cftype *cft, cgrp_cfile = __d_cgrp(cfile->f_dentry->d_parent); if (cgrp_cfile != cgrp) { ret = -EINVAL; - goto fail; + goto out_put_cfile; } if (!event->cft->register_event || !event->cft->unregister_event) { ret = -EINVAL; - goto fail; + goto out_put_cfile; } ret = event->cft->register_event(cgrp, event->cft, event->eventfd, buffer); if (ret) - goto fail; + goto out_put_cfile; efile->f_op->poll(efile, &event->pt); @@ -4029,16 +4029,13 @@ static int cgroup_write_event_control(struct cgroup *cgrp, struct cftype *cft, return 0; -fail: - if (cfile) - fput(cfile); - - if (event && event->eventfd && !IS_ERR(event->eventfd)) - eventfd_ctx_put(event->eventfd); - - if (!IS_ERR_OR_NULL(efile)) - fput(efile); - +out_put_cfile: + fput(cfile); +out_put_eventfd: + eventfd_ctx_put(event->eventfd); +out_put_efile: + fput(efile); +out_kfree: kfree(event); return ret; -- cgit v1.2.3 From b395890a092d8ecbe54f005179e3dec4b6bf752a Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Thu, 1 Aug 2013 09:52:15 +0800 Subject: cgroup: rename cgroup_pidlist->mutex It's a rw_semaphore not a mutex. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cgroup.c | 22 +++++++++++----------- 1 file changed, 11 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 9f6dab22289e..9420662df87e 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -3436,7 +3436,7 @@ struct cgroup_pidlist { /* pointer to the cgroup we belong to, for list removal purposes */ struct cgroup *owner; /* protects the other fields */ - struct rw_semaphore mutex; + struct rw_semaphore rwsem; }; /* @@ -3509,7 +3509,7 @@ static struct cgroup_pidlist *cgroup_pidlist_find(struct cgroup *cgrp, struct pid_namespace *ns = task_active_pid_ns(current); /* - * We can't drop the pidlist_mutex before taking the l->mutex in case + * We can't drop the pidlist_mutex before taking the l->rwsem in case * the last ref-holder is trying to remove l from the list at the same * time. Holding the pidlist_mutex precludes somebody taking whichever * list we find out from under us - compare release_pid_array(). @@ -3518,7 +3518,7 @@ static struct cgroup_pidlist *cgroup_pidlist_find(struct cgroup *cgrp, list_for_each_entry(l, &cgrp->pidlists, links) { if (l->key.type == type && l->key.ns == ns) { /* make sure l doesn't vanish out from under us */ - down_write(&l->mutex); + down_write(&l->rwsem); mutex_unlock(&cgrp->pidlist_mutex); return l; } @@ -3529,8 +3529,8 @@ static struct cgroup_pidlist *cgroup_pidlist_find(struct cgroup *cgrp, mutex_unlock(&cgrp->pidlist_mutex); return l; } - init_rwsem(&l->mutex); - down_write(&l->mutex); + init_rwsem(&l->rwsem); + down_write(&l->rwsem); l->key.type = type; l->key.ns = get_pid_ns(ns); l->owner = cgrp; @@ -3591,7 +3591,7 @@ static int pidlist_array_load(struct cgroup *cgrp, enum cgroup_filetype type, l->list = array; l->length = length; l->use_count++; - up_write(&l->mutex); + up_write(&l->rwsem); *lp = l; return 0; } @@ -3669,7 +3669,7 @@ static void *cgroup_pidlist_start(struct seq_file *s, loff_t *pos) int index = 0, pid = *pos; int *iter; - down_read(&l->mutex); + down_read(&l->rwsem); if (pid) { int end = l->length; @@ -3696,7 +3696,7 @@ static void *cgroup_pidlist_start(struct seq_file *s, loff_t *pos) static void cgroup_pidlist_stop(struct seq_file *s, void *v) { struct cgroup_pidlist *l = s->private; - up_read(&l->mutex); + up_read(&l->rwsem); } static void *cgroup_pidlist_next(struct seq_file *s, void *v, loff_t *pos) @@ -3742,7 +3742,7 @@ static void cgroup_release_pid_array(struct cgroup_pidlist *l) * pidlist_mutex, we have to take pidlist_mutex first. */ mutex_lock(&l->owner->pidlist_mutex); - down_write(&l->mutex); + down_write(&l->rwsem); BUG_ON(!l->use_count); if (!--l->use_count) { /* we're the last user if refcount is 0; remove and free */ @@ -3750,12 +3750,12 @@ static void cgroup_release_pid_array(struct cgroup_pidlist *l) mutex_unlock(&l->owner->pidlist_mutex); pidlist_free(l->list); put_pid_ns(l->key.ns); - up_write(&l->mutex); + up_write(&l->rwsem); kfree(l); return; } mutex_unlock(&l->owner->pidlist_mutex); - up_write(&l->mutex); + up_write(&l->rwsem); } static int cgroup_pidlist_release(struct inode *inode, struct file *file) -- cgit v1.2.3 From c6c2401d8bbaf9edc189b4c35a8cb2780b8b988e Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Wed, 3 Jul 2013 23:33:51 -0400 Subject: tracing/uprobes: Fail to unregister if probe event files are in use Uprobes suffer the same problem that kprobes have. There's a race between writing to the "enable" file and removing the probe. The probe checks for it being in use and if it is not, goes about deleting the probe and the event that represents it. But the problem with that is, after it checks if it is in use it can be enabled, and the deletion of the event (access to the probe) will fail, as it is in use. But the uprobe will still be deleted. This is a problem as the event can reference the uprobe that was deleted. The fix is to remove the event first, and check to make sure the event removal succeeds. Then it is safe to remove the probe. When the event exists, either ftrace or perf can enable the probe and prevent the event from being removed. Link: http://lkml.kernel.org/r/20130704034038.991525256@goodmis.org Acked-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_uprobe.c | 51 +++++++++++++++++++++++++++++++++------------ 1 file changed, 38 insertions(+), 13 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_uprobe.c b/kernel/trace/trace_uprobe.c index a23d2d71188e..272261b5f94f 100644 --- a/kernel/trace/trace_uprobe.c +++ b/kernel/trace/trace_uprobe.c @@ -70,7 +70,7 @@ struct trace_uprobe { (sizeof(struct probe_arg) * (n))) static int register_uprobe_event(struct trace_uprobe *tu); -static void unregister_uprobe_event(struct trace_uprobe *tu); +static int unregister_uprobe_event(struct trace_uprobe *tu); static DEFINE_MUTEX(uprobe_lock); static LIST_HEAD(uprobe_list); @@ -164,11 +164,17 @@ static struct trace_uprobe *find_probe_event(const char *event, const char *grou } /* Unregister a trace_uprobe and probe_event: call with locking uprobe_lock */ -static void unregister_trace_uprobe(struct trace_uprobe *tu) +static int unregister_trace_uprobe(struct trace_uprobe *tu) { + int ret; + + ret = unregister_uprobe_event(tu); + if (ret) + return ret; + list_del(&tu->list); - unregister_uprobe_event(tu); free_trace_uprobe(tu); + return 0; } /* Register a trace_uprobe and probe_event */ @@ -181,9 +187,12 @@ static int register_trace_uprobe(struct trace_uprobe *tu) /* register as an event */ old_tp = find_probe_event(tu->call.name, tu->call.class->system); - if (old_tp) + if (old_tp) { /* delete old event */ - unregister_trace_uprobe(old_tp); + ret = unregister_trace_uprobe(old_tp); + if (ret) + goto end; + } ret = register_uprobe_event(tu); if (ret) { @@ -256,6 +265,8 @@ static int create_trace_uprobe(int argc, char **argv) group = UPROBE_EVENT_SYSTEM; if (is_delete) { + int ret; + if (!event) { pr_info("Delete command needs an event name.\n"); return -EINVAL; @@ -269,9 +280,9 @@ static int create_trace_uprobe(int argc, char **argv) return -ENOENT; } /* delete an event */ - unregister_trace_uprobe(tu); + ret = unregister_trace_uprobe(tu); mutex_unlock(&uprobe_lock); - return 0; + return ret; } if (argc < 2) { @@ -408,16 +419,20 @@ fail_address_parse: return ret; } -static void cleanup_all_probes(void) +static int cleanup_all_probes(void) { struct trace_uprobe *tu; + int ret = 0; mutex_lock(&uprobe_lock); while (!list_empty(&uprobe_list)) { tu = list_entry(uprobe_list.next, struct trace_uprobe, list); - unregister_trace_uprobe(tu); + ret = unregister_trace_uprobe(tu); + if (ret) + break; } mutex_unlock(&uprobe_lock); + return ret; } /* Probes listing interfaces */ @@ -462,8 +477,13 @@ static const struct seq_operations probes_seq_op = { static int probes_open(struct inode *inode, struct file *file) { - if ((file->f_mode & FMODE_WRITE) && (file->f_flags & O_TRUNC)) - cleanup_all_probes(); + int ret; + + if ((file->f_mode & FMODE_WRITE) && (file->f_flags & O_TRUNC)) { + ret = cleanup_all_probes(); + if (ret) + return ret; + } return seq_open(file, &probes_seq_op); } @@ -968,12 +988,17 @@ static int register_uprobe_event(struct trace_uprobe *tu) return ret; } -static void unregister_uprobe_event(struct trace_uprobe *tu) +static int unregister_uprobe_event(struct trace_uprobe *tu) { + int ret; + /* tu->event is unregistered in trace_remove_event_call() */ - trace_remove_event_call(&tu->call); + ret = trace_remove_event_call(&tu->call); + if (ret) + return ret; kfree(tu->call.print_fmt); tu->call.print_fmt = NULL; + return 0; } /* Make a trace interface for controling probe points */ -- cgit v1.2.3 From 41e85ce8220c6e5fdef706fda6696cd291115b63 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Thu, 1 Aug 2013 18:59:41 +0200 Subject: hung_task debugging: Print more info when reporting the problem printk(KERN_ERR) from check_hung_task() likely means we have a bug, but unlike BUG_ON()/WARN_ON ()it doesn't show the kernel version, this complicates the bug-reports investigation. Add the additional pr_err() to print tainted/release/version like dump_stack_print_info() does, the output becomes: INFO: task perl:504 blocked for more than 2 seconds. Not tainted 3.11.0-rc1-10367-g136bb46-dirty #1763 "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. ... While at it, turn the old printk's into pr_err(). Signed-off-by: Oleg Nesterov Cc: ahecox@redhat.com Cc: Christopher Williams Cc: dwysocha@redhat.com Cc: gavin@redhat.com Cc: Mandeep Singh Baines Cc: nshi@redhat.com Cc: Linus Torvalds Cc: Andrew Morton Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/20130801165941.GA17544@redhat.com Signed-off-by: Ingo Molnar --- kernel/hung_task.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/hung_task.c b/kernel/hung_task.c index 6df614912b9d..3e97fb126e6b 100644 --- a/kernel/hung_task.c +++ b/kernel/hung_task.c @@ -15,6 +15,7 @@ #include #include #include +#include /* * The number of tasks checked: @@ -99,10 +100,14 @@ static void check_hung_task(struct task_struct *t, unsigned long timeout) * Ok, the task did not get scheduled for more than 2 minutes, * complain: */ - printk(KERN_ERR "INFO: task %s:%d blocked for more than " - "%ld seconds.\n", t->comm, t->pid, timeout); - printk(KERN_ERR "\"echo 0 > /proc/sys/kernel/hung_task_timeout_secs\"" - " disables this message.\n"); + pr_err("INFO: task %s:%d blocked for more than %ld seconds.\n", + t->comm, t->pid, timeout); + pr_err(" %s %s %.*s\n", + print_tainted(), init_utsname()->release, + (int)strcspn(init_utsname()->version, " "), + init_utsname()->version); + pr_err("\"echo 0 > /proc/sys/kernel/hung_task_timeout_secs\"" + " disables this message.\n"); sched_show_task(t); debug_show_held_locks(t); -- cgit v1.2.3 From ed5467da0e369e65b247b99eb6403cb79172bcda Mon Sep 17 00:00:00 2001 From: Andrew Vagin Date: Fri, 2 Aug 2013 21:16:43 +0400 Subject: tracing: Fix fields of struct trace_iterator that are zeroed by mistake tracing_read_pipe zeros all fields bellow "seq". The declaration contains a comment about that, but it doesn't help. The first field is "snapshot", it's true when current open file is snapshot. Looks obvious, that it should not be zeroed. The second field is "started". It was converted from cpumask_t to cpumask_var_t (v2.6.28-4983-g4462344), in other words it was converted from cpumask to pointer on cpumask. Currently the reference on "started" memory is lost after the first read from tracing_read_pipe and a proper object will never be freed. The "started" is never dereferenced for trace_pipe, because trace_pipe can't have the TRACE_FILE_ANNOTATE options. Link: http://lkml.kernel.org/r/1375463803-3085183-1-git-send-email-avagin@openvz.org Cc: stable@vger.kernel.org # 2.6.30 Signed-off-by: Andrew Vagin Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 882ec1dd1515..f5b35a5e852f 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -4151,6 +4151,7 @@ waitagain: memset(&iter->seq, 0, sizeof(struct trace_iterator) - offsetof(struct trace_iterator, seq)); + cpumask_clear(iter->started); iter->pos = -1; trace_event_read_lock(); -- cgit v1.2.3 From 711e124379e0f889e40e2f01d7f5d61936d3cd23 Mon Sep 17 00:00:00 2001 From: Alexander Z Lam Date: Fri, 2 Aug 2013 18:36:15 -0700 Subject: tracing: Make TRACE_ITER_STOP_ON_FREE stop the correct buffer Releasing the free_buffer file in an instance causes the global buffer to be stopped when TRACE_ITER_STOP_ON_FREE is enabled. Operate on the correct buffer. Link: http://lkml.kernel.org/r/1375493777-17261-1-git-send-email-azl@google.com Cc: Vaibhav Nagarnaik Cc: David Sharp Cc: Alexander Z Lam Cc: stable@vger.kernel.org # 3.10 Signed-off-by: Alexander Z Lam Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index f5b35a5e852f..531c9e69d0b3 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -4469,7 +4469,7 @@ tracing_free_buffer_release(struct inode *inode, struct file *filp) /* disable tracing ? */ if (trace_flags & TRACE_ITER_STOP_ON_FREE) - tracing_off(); + tracer_tracing_off(tr); /* resize the ring buffer to 0 */ tracing_resize_ring_buffer(tr, 0, RING_BUFFER_ALL_CPUS); -- cgit v1.2.3 From 9457158bbc0ee04ecef76862d73eecd8076e9c7b Mon Sep 17 00:00:00 2001 From: Alexander Z Lam Date: Fri, 2 Aug 2013 18:36:16 -0700 Subject: tracing: Fix reset of time stamps during trace_clock changes Fixed two issues with changing the timestamp clock with trace_clock: - The global buffer was reset on instance clock changes. Change this to pass the correct per-instance buffer - ftrace_now() is used to set buf->time_start in tracing_reset_online_cpus(). This was incorrect because ftrace_now() used the global buffer's clock to return the current time. Change this to use buffer_ftrace_now() which returns the current time for the correct per-instance buffer. Also removed tracing_reset_current() because it is not used anywhere Link: http://lkml.kernel.org/r/1375493777-17261-2-git-send-email-azl@google.com Cc: Vaibhav Nagarnaik Cc: David Sharp Cc: Alexander Z Lam Cc: stable@vger.kernel.org # 3.10 Signed-off-by: Alexander Z Lam Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 24 ++++++++++++------------ 1 file changed, 12 insertions(+), 12 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 531c9e69d0b3..496f94d57698 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -243,20 +243,25 @@ int filter_current_check_discard(struct ring_buffer *buffer, } EXPORT_SYMBOL_GPL(filter_current_check_discard); -cycle_t ftrace_now(int cpu) +cycle_t buffer_ftrace_now(struct trace_buffer *buf, int cpu) { u64 ts; /* Early boot up does not have a buffer yet */ - if (!global_trace.trace_buffer.buffer) + if (!buf->buffer) return trace_clock_local(); - ts = ring_buffer_time_stamp(global_trace.trace_buffer.buffer, cpu); - ring_buffer_normalize_time_stamp(global_trace.trace_buffer.buffer, cpu, &ts); + ts = ring_buffer_time_stamp(buf->buffer, cpu); + ring_buffer_normalize_time_stamp(buf->buffer, cpu, &ts); return ts; } +cycle_t ftrace_now(int cpu) +{ + return buffer_ftrace_now(&global_trace.trace_buffer, cpu); +} + /** * tracing_is_enabled - Show if global_trace has been disabled * @@ -1211,7 +1216,7 @@ void tracing_reset_online_cpus(struct trace_buffer *buf) /* Make sure all commits have finished */ synchronize_sched(); - buf->time_start = ftrace_now(buf->cpu); + buf->time_start = buffer_ftrace_now(buf, buf->cpu); for_each_online_cpu(cpu) ring_buffer_reset_cpu(buffer, cpu); @@ -1219,11 +1224,6 @@ void tracing_reset_online_cpus(struct trace_buffer *buf) ring_buffer_record_enable(buffer); } -void tracing_reset_current(int cpu) -{ - tracing_reset(&global_trace.trace_buffer, cpu); -} - /* Must have trace_types_lock held */ void tracing_reset_all_online_cpus(void) { @@ -4634,12 +4634,12 @@ static ssize_t tracing_clock_write(struct file *filp, const char __user *ubuf, * New clock may not be consistent with the previous clock. * Reset the buffer so that it doesn't have incomparable timestamps. */ - tracing_reset_online_cpus(&global_trace.trace_buffer); + tracing_reset_online_cpus(&tr->trace_buffer); #ifdef CONFIG_TRACER_MAX_TRACE if (tr->flags & TRACE_ARRAY_FL_GLOBAL && tr->max_buffer.buffer) ring_buffer_set_clock(tr->max_buffer.buffer, trace_clocks[i].func); - tracing_reset_online_cpus(&global_trace.max_buffer); + tracing_reset_online_cpus(&tr->max_buffer); #endif mutex_unlock(&trace_types_lock); -- cgit v1.2.3 From 16cf48a6d3e8f9ebe3c3231c12cbe4b0c4ed4d24 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Andreas=20Bie=C3=9Fmann?= Date: Fri, 2 Aug 2013 12:23:34 +0200 Subject: register_console: prevent adding the same console twice MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This patch guards the console_drivers list to be corrupted. The for_each_console() macro insist on a strictly forward list ended by NULL: con0->next->con1->next->NULL Without this patch it may happen easily to destroy this list for example by adding 'earlyprintk' twice, especially on embedded devices where the early console is often a single static instance. This will result in the following list: con0->next->con0 This in turn will result in an endless loop in console_unlock() later on by printing the first __log_buf line endlessly. Signed-off-by: Andreas Bießmann Cc: Kay Sievers Cc: Ben Hutchings Signed-off-by: Greg Kroah-Hartman --- kernel/printk/printk.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'kernel') diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c index 5b5a7080e2a5..b4e8500afdb3 100644 --- a/kernel/printk/printk.c +++ b/kernel/printk/printk.c @@ -2226,6 +2226,13 @@ void register_console(struct console *newcon) struct console *bcon = NULL; struct console_cmdline *c; + if (console_drivers) + for_each_console(bcon) + if (WARN(bcon == newcon, + "console '%s%d' already registered\n", + bcon->name, bcon->index)) + return; + /* * before we register a new CON_BOOT console, make sure we don't * already have a valid console -- cgit v1.2.3 From 6160968cee8b90a5dd95318d716e31d7775c4ef3 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 6 Aug 2013 19:38:55 +0200 Subject: userns: unshare_userns(&cred) should not populate cred on failure unshare_userns(new_cred) does *new_cred = prepare_creds() before create_user_ns() which can fail. However, the caller expects that it doesn't need to take care of new_cred if unshare_userns() fails. We could change the single caller, sys_unshare(), but I think it would be more clean to avoid the side effects on failure, so with this patch unshare_userns() does put_cred() itself and initializes *new_cred only if create_user_ns() succeeeds. Cc: stable@vger.kernel.org Signed-off-by: Oleg Nesterov Reviewed-by: Andy Lutomirski Signed-off-by: Linus Torvalds --- kernel/user_namespace.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/user_namespace.c b/kernel/user_namespace.c index d8c30db06c5b..6e50a44610ee 100644 --- a/kernel/user_namespace.c +++ b/kernel/user_namespace.c @@ -105,16 +105,21 @@ int create_user_ns(struct cred *new) int unshare_userns(unsigned long unshare_flags, struct cred **new_cred) { struct cred *cred; + int err = -ENOMEM; if (!(unshare_flags & CLONE_NEWUSER)) return 0; cred = prepare_creds(); - if (!cred) - return -ENOMEM; + if (cred) { + err = create_user_ns(cred); + if (err) + put_cred(cred); + else + *new_cred = cred; + } - *new_cred = cred; - return create_user_ns(cred); + return err; } void free_user_ns(struct user_namespace *ns) -- cgit v1.2.3 From 35114fcbe0b9b0fa3f6653a2a8e4c6b8a9f8cc2d Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 6 Aug 2013 17:43:37 +0200 Subject: Revert "ptrace: PTRACE_DETACH should do flush_ptrace_hw_breakpoint(child)" This reverts commit fab840fc2d542fabcab903db8e03589a6702ba5f. This commit even has the test-case to prove that the tracee can be killed by SIGTRAP if the debugger does not remove the breakpoints before PTRACE_DETACH. However, this is exactly what wineserver deliberately does, set_thread_context() calls PTRACE_ATTACH + PTRACE_DETACH just for PTRACE_POKEUSER(DR*) in between. So we should revert this fix and document that PTRACE_DETACH should keep the breakpoints. Reported-by: Felipe Contreras Signed-off-by: Oleg Nesterov Signed-off-by: Linus Torvalds --- kernel/ptrace.c | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/ptrace.c b/kernel/ptrace.c index 4041f5747e73..a146ee327f6a 100644 --- a/kernel/ptrace.c +++ b/kernel/ptrace.c @@ -469,7 +469,6 @@ static int ptrace_detach(struct task_struct *child, unsigned int data) /* Architecture-specific hardware disable .. */ ptrace_disable(child); clear_tsk_thread_flag(child, TIF_SYSCALL_TRACE); - flush_ptrace_hw_breakpoint(child); write_lock_irq(&tasklist_lock); /* -- cgit v1.2.3 From 2cfe6c4ac7ee0193780d655c5dea5a73acae1f46 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 5 Aug 2013 22:55:28 -0400 Subject: printk: Fix return of braille_register_console() Some of my configs I test with have CONFIG_A11Y_BRAILLE_CONSOLE set. When I started testing against v3.11-rc4 my console went bonkers. Using ktest to bisect the issue, it came down to: commit bbeddf52a "printk: move braille console support into separate braille.[ch] files" Looking into the patch I found the problem. It's with the return of braille_register_console(). As anything other than NULL is considered a failure. But for those of us that have CONFIG_A11Y_BRAILLE_CONSOLE set but do not define a "brl" or "brl=" on the command line, we still may want a console that those with sight can still use. Return NULL (success) if "brl" or "brl=" is not on the console line. Signed-off-by: Steven Rostedt Acked-by: Joe Perches Cc: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/printk/braille.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/printk/braille.c b/kernel/printk/braille.c index b51087fb9ace..276762f3a460 100644 --- a/kernel/printk/braille.c +++ b/kernel/printk/braille.c @@ -19,7 +19,8 @@ char *_braille_console_setup(char **str, char **brl_options) pr_err("need port name after brl=\n"); else *((*str)++) = 0; - } + } else + return NULL; return *str; } -- cgit v1.2.3 From d6efc2f7240b4e55590df69d74f33fdb72ce934a Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Mon, 5 Aug 2013 15:02:49 -0700 Subject: x86, asmlinkage, power: Make various symbols used by the suspend asm code visible Signed-off-by: Andi Kleen Link: http://lkml.kernel.org/r/1375740170-7446-16-git-send-email-andi@firstfloor.org Signed-off-by: H. Peter Anvin --- kernel/power/hibernate.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/power/hibernate.c b/kernel/power/hibernate.c index b26f5f1e773e..3085e62a80a5 100644 --- a/kernel/power/hibernate.c +++ b/kernel/power/hibernate.c @@ -39,7 +39,7 @@ static int resume_delay; static char resume_file[256] = CONFIG_PM_STD_PARTITION; dev_t swsusp_resume_device; sector_t swsusp_resume_block; -int in_suspend __nosavedata; +__visible int in_suspend __nosavedata; enum { HIBERNATION_INVALID, -- cgit v1.2.3 From cf4957f17f2a89984915ea808876d9c82225b862 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Wed, 24 Oct 2012 13:37:58 +0200 Subject: perf: Add PERF_EVENT_IOC_ID ioctl to return event ID The only way to get the event ID is by reading the event fd, followed by parsing the ID value out of the returned data. While this is ok for current read format used by perf tool, it is not ok when we use PERF_FORMAT_GROUP format. With this format the data are returned for the whole group and there's no way to find out what ID belongs to our fd (if we are not group leader event). Adding a simple ioctl that returns event primary ID for given fd. Signed-off-by: Jiri Olsa Acked-by: Namhyung Kim Acked-by: Peter Zijlstra Cc: Corey Ashford Cc: Frederic Weisbecker Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-v1bn5cto707jn0bon34afqr1@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- kernel/events/core.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 916cf1f593b4..5200b608b481 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -3568,6 +3568,15 @@ static long perf_ioctl(struct file *file, unsigned int cmd, unsigned long arg) case PERF_EVENT_IOC_PERIOD: return perf_event_period(event, (u64 __user *)arg); + case PERF_EVENT_IOC_ID: + { + u64 id = primary_event_id(event); + + if (copy_to_user((void __user *)arg, &id, sizeof(id))) + return -EFAULT; + return 0; + } + case PERF_EVENT_IOC_SET_OUTPUT: { int ret; -- cgit v1.2.3 From 6f5ab0019fd328b50a8488c9e5193fc1dbd8d6ed Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Mon, 15 Oct 2012 20:13:45 +0200 Subject: perf: Do not get values from disabled counters in group format read It's possible some of the counters in the group could be disabled when sampling member of the event group is reading the rest via PERF_SAMPLE_READ sample type processing. Disabled counters could then produce wrong numbers. Fixing that by reading only enabled counters for PERF_SAMPLE_READ sample type processing. Signed-off-by: Jiri Olsa Acked-by: Namhyung Kim Acked-by: Peter Zijlstra Cc: Corey Ashford Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-wwkjb0bbcuslnz0klrmqi26r@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- kernel/events/core.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 5200b608b481..e82e70025d42 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -4388,7 +4388,8 @@ static void perf_output_read_group(struct perf_output_handle *handle, list_for_each_entry(sub, &leader->sibling_list, group_entry) { n = 0; - if (sub != event) + if ((sub != event) && + (sub->state == PERF_EVENT_STATE_ACTIVE)) sub->pmu->read(sub); values[n++] = perf_event_count(sub); -- cgit v1.2.3 From 8742f229b635bf1c1c84a3dfe5e47c814c20b5c8 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Thu, 8 Aug 2013 18:55:32 +0200 Subject: userns: limit the maximum depth of user_namespace->parent chain Ensure that user_namespace->parent chain can't grow too much. Currently we use the hardroded 32 as limit. Reported-by: Andy Lutomirski Signed-off-by: Oleg Nesterov Signed-off-by: Linus Torvalds --- kernel/user_namespace.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'kernel') diff --git a/kernel/user_namespace.c b/kernel/user_namespace.c index 6e50a44610ee..9064b919a406 100644 --- a/kernel/user_namespace.c +++ b/kernel/user_namespace.c @@ -62,6 +62,9 @@ int create_user_ns(struct cred *new) kgid_t group = new->egid; int ret; + if (parent_ns->level > 32) + return -EUSERS; + /* * Verify that we can not violate the policy of which files * may be accessed that is specified by the root directory, @@ -92,6 +95,7 @@ int create_user_ns(struct cred *new) atomic_set(&ns->count, 1); /* Leave the new->user_ns reference with the new user namespace. */ ns->parent = parent_ns; + ns->level = parent_ns->level + 1; ns->owner = owner; ns->group = group; -- cgit v1.2.3 From 8af01f56a03e9cbd91a55d688fce1315021efba8 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:22 -0400 Subject: cgroup: s/cgroup_subsys_state/cgroup_css/ s/task_subsys_state/task_css/ The names of the two struct cgroup_subsys_state accessors - cgroup_subsys_state() and task_subsys_state() - are somewhat awkward. The former clashes with the type name and the latter doesn't even indicate it's somehow related to cgroup. We're about to revamp large portion of cgroup API, so, let's rename them so that they're less awkward. Most per-controller usages of the accessors are localized in accessor wrappers and given the amount of scheduled changes, this isn't gonna add any noticeable headache. Rename cgroup_subsys_state() to cgroup_css() and task_subsys_state() to task_css(). This patch is pure rename. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 2 +- kernel/cgroup_freezer.c | 4 ++-- kernel/cpuset.c | 6 +++--- kernel/events/core.c | 6 +++--- kernel/sched/core.c | 4 ++-- kernel/sched/cpuacct.c | 4 ++-- kernel/sched/sched.h | 6 +++--- 7 files changed, 16 insertions(+), 16 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index ae4c46834633..0b3caa3220cb 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -81,7 +81,7 @@ */ #ifdef CONFIG_PROVE_RCU DEFINE_MUTEX(cgroup_mutex); -EXPORT_SYMBOL_GPL(cgroup_mutex); /* only for task_subsys_state_check() */ +EXPORT_SYMBOL_GPL(cgroup_mutex); /* only for lockdep */ #else static DEFINE_MUTEX(cgroup_mutex); #endif diff --git a/kernel/cgroup_freezer.c b/kernel/cgroup_freezer.c index 75dda1ea5026..9d3f61566fec 100644 --- a/kernel/cgroup_freezer.c +++ b/kernel/cgroup_freezer.c @@ -47,13 +47,13 @@ struct freezer { static inline struct freezer *cgroup_freezer(struct cgroup *cgroup) { - return container_of(cgroup_subsys_state(cgroup, freezer_subsys_id), + return container_of(cgroup_css(cgroup, freezer_subsys_id), struct freezer, css); } static inline struct freezer *task_freezer(struct task_struct *task) { - return container_of(task_subsys_state(task, freezer_subsys_id), + return container_of(task_css(task, freezer_subsys_id), struct freezer, css); } diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 1b9c31549797..be4512ba2c0c 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -117,14 +117,14 @@ struct cpuset { /* Retrieve the cpuset for a cgroup */ static inline struct cpuset *cgroup_cs(struct cgroup *cgrp) { - return container_of(cgroup_subsys_state(cgrp, cpuset_subsys_id), + return container_of(cgroup_css(cgrp, cpuset_subsys_id), struct cpuset, css); } /* Retrieve the cpuset for a task */ static inline struct cpuset *task_cs(struct task_struct *task) { - return container_of(task_subsys_state(task, cpuset_subsys_id), + return container_of(task_css(task, cpuset_subsys_id), struct cpuset, css); } @@ -2724,7 +2724,7 @@ int proc_cpuset_show(struct seq_file *m, void *unused_v) goto out_free; rcu_read_lock(); - css = task_subsys_state(tsk, cpuset_subsys_id); + css = task_css(tsk, cpuset_subsys_id); retval = cgroup_path(css->cgroup, buf, PAGE_SIZE); rcu_read_unlock(); if (retval < 0) diff --git a/kernel/events/core.c b/kernel/events/core.c index 1833bc5a84a7..414c61f4d776 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -340,8 +340,8 @@ struct perf_cgroup { static inline struct perf_cgroup * perf_cgroup_from_task(struct task_struct *task) { - return container_of(task_subsys_state(task, perf_subsys_id), - struct perf_cgroup, css); + return container_of(task_css(task, perf_subsys_id), + struct perf_cgroup, css); } static inline bool @@ -7798,7 +7798,7 @@ static struct cgroup_subsys_state *perf_cgroup_css_alloc(struct cgroup *cont) static void perf_cgroup_css_free(struct cgroup *cont) { struct perf_cgroup *jc; - jc = container_of(cgroup_subsys_state(cont, perf_subsys_id), + jc = container_of(cgroup_css(cont, perf_subsys_id), struct perf_cgroup, css); free_percpu(jc->info); kfree(jc); diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 9b1f2e533b95..323d907eac1a 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -6761,7 +6761,7 @@ void sched_move_task(struct task_struct *tsk) if (unlikely(running)) tsk->sched_class->put_prev_task(rq, tsk); - tg = container_of(task_subsys_state_check(tsk, cpu_cgroup_subsys_id, + tg = container_of(task_css_check(tsk, cpu_cgroup_subsys_id, lockdep_is_held(&tsk->sighand->siglock)), struct task_group, css); tg = autogroup_task_group(tsk, tg); @@ -7086,7 +7086,7 @@ int sched_rt_handler(struct ctl_table *table, int write, /* return corresponding task_group object of a cgroup */ static inline struct task_group *cgroup_tg(struct cgroup *cgrp) { - return container_of(cgroup_subsys_state(cgrp, cpu_cgroup_subsys_id), + return container_of(cgroup_css(cgrp, cpu_cgroup_subsys_id), struct task_group, css); } diff --git a/kernel/sched/cpuacct.c b/kernel/sched/cpuacct.c index dbb7e2cd95eb..4a210faaab77 100644 --- a/kernel/sched/cpuacct.c +++ b/kernel/sched/cpuacct.c @@ -36,14 +36,14 @@ struct cpuacct { /* return cpu accounting group corresponding to this container */ static inline struct cpuacct *cgroup_ca(struct cgroup *cgrp) { - return container_of(cgroup_subsys_state(cgrp, cpuacct_subsys_id), + return container_of(cgroup_css(cgrp, cpuacct_subsys_id), struct cpuacct, css); } /* return cpu accounting group to which this task belongs */ static inline struct cpuacct *task_ca(struct task_struct *tsk) { - return container_of(task_subsys_state(tsk, cpuacct_subsys_id), + return container_of(task_css(tsk, cpuacct_subsys_id), struct cpuacct, css); } diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index ef0a7b2439dd..471a56db05ea 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -665,9 +665,9 @@ extern int group_balance_cpu(struct sched_group *sg); /* * Return the group to which this tasks belongs. * - * We cannot use task_subsys_state() and friends because the cgroup - * subsystem changes that value before the cgroup_subsys::attach() method - * is called, therefore we cannot pin it and might observe the wrong value. + * We cannot use task_css() and friends because the cgroup subsystem + * changes that value before the cgroup_subsys::attach() method is called, + * therefore we cannot pin it and might observe the wrong value. * * The same is true for autogroup's p->signal->autogroup->tg, the autogroup * core changes this before calling sched_move_task(). -- cgit v1.2.3 From c9710d8018273b0740e0794858f1961fcea5e61a Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:22 -0400 Subject: cpuset: drop "const" qualifiers from struct cpuset instances cpuset uses "const" qualifiers on struct cpuset in some functions; however, it doesn't work well when a value derived from returned const pointer has to be passed to an accessor. It's C after all. Drop the "const" qualifiers except for the trivially leaf ones. This patch doesn't make any functional changes. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cpuset.c | 17 ++++++++--------- 1 file changed, 8 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index be4512ba2c0c..f7371341d42a 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -128,7 +128,7 @@ static inline struct cpuset *task_cs(struct task_struct *task) struct cpuset, css); } -static inline struct cpuset *parent_cs(const struct cpuset *cs) +static inline struct cpuset *parent_cs(struct cpuset *cs) { struct cgroup *pcgrp = cs->css.cgroup->parent; @@ -319,8 +319,7 @@ static struct file_system_type cpuset_fs_type = { * * Call with callback_mutex held. */ -static void guarantee_online_cpus(const struct cpuset *cs, - struct cpumask *pmask) +static void guarantee_online_cpus(struct cpuset *cs, struct cpumask *pmask) { while (!cpumask_intersects(cs->cpus_allowed, cpu_online_mask)) cs = parent_cs(cs); @@ -338,7 +337,7 @@ static void guarantee_online_cpus(const struct cpuset *cs, * * Call with callback_mutex held. */ -static void guarantee_online_mems(const struct cpuset *cs, nodemask_t *pmask) +static void guarantee_online_mems(struct cpuset *cs, nodemask_t *pmask) { while (!nodes_intersects(cs->mems_allowed, node_states[N_MEMORY])) cs = parent_cs(cs); @@ -383,7 +382,7 @@ static int is_cpuset_subset(const struct cpuset *p, const struct cpuset *q) * alloc_trial_cpuset - allocate a trial cpuset * @cs: the cpuset that the trial cpuset duplicates */ -static struct cpuset *alloc_trial_cpuset(const struct cpuset *cs) +static struct cpuset *alloc_trial_cpuset(struct cpuset *cs) { struct cpuset *trial; @@ -430,7 +429,7 @@ static void free_trial_cpuset(struct cpuset *trial) * Return 0 if valid, -errno if not. */ -static int validate_change(const struct cpuset *cur, const struct cpuset *trial) +static int validate_change(struct cpuset *cur, struct cpuset *trial) { struct cgroup *cgrp; struct cpuset *c, *par; @@ -2343,7 +2342,7 @@ void cpuset_cpus_allowed(struct task_struct *tsk, struct cpumask *pmask) void cpuset_cpus_allowed_fallback(struct task_struct *tsk) { - const struct cpuset *cpus_cs; + struct cpuset *cpus_cs; rcu_read_lock(); cpus_cs = effective_cpumask_cpuset(task_cs(tsk)); @@ -2416,7 +2415,7 @@ int cpuset_nodemask_valid_mems_allowed(nodemask_t *nodemask) * callback_mutex. If no ancestor is mem_exclusive or mem_hardwall * (an unusual configuration), then returns the root cpuset. */ -static const struct cpuset *nearest_hardwall_ancestor(const struct cpuset *cs) +static struct cpuset *nearest_hardwall_ancestor(struct cpuset *cs) { while (!(is_mem_exclusive(cs) || is_mem_hardwall(cs)) && parent_cs(cs)) cs = parent_cs(cs); @@ -2486,7 +2485,7 @@ static const struct cpuset *nearest_hardwall_ancestor(const struct cpuset *cs) */ int __cpuset_node_allowed_softwall(int node, gfp_t gfp_mask) { - const struct cpuset *cs; /* current cpuset ancestors */ + struct cpuset *cs; /* current cpuset ancestors */ int allowed; /* is allocation in zone z allowed? */ if (in_interrupt() || (gfp_mask & __GFP_THISNODE)) -- cgit v1.2.3 From 72c97e54e0f043d33b246d7460ae0a36c4b8c643 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:22 -0400 Subject: cgroup: add subsystem pointer to cgroup_subsys_state Currently, given a cgroup_subsys_state, there's no way to find out which subsystem the css is for, which we'll need to convert the cgroup controller API to primarily use @css instead of @cgroup. This patch adds cgroup_subsys_state->ss which points to the subsystem the @css belongs to. While at it, remove the comment about accessing @css->cgroup to determine the hierarchy. cgroup core will provide API to traverse hierarchy of css'es and we don't want subsystems to directly walk cgroup hierarchies anymore. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 0b3caa3220cb..4234428f1014 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4186,6 +4186,7 @@ static void init_cgroup_css(struct cgroup_subsys_state *css, struct cgroup *cgrp) { css->cgroup = cgrp; + css->ss = ss; css->flags = 0; css->id = NULL; if (cgrp == cgroup_dummy_top) -- cgit v1.2.3 From a7c6d554aa01236ac2a9f851ab0f75704f76dfa2 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:23 -0400 Subject: cgroup: add/update accessors which obtain subsys specific data from css css (cgroup_subsys_state) is usually embedded in a subsys specific data structure. Subsystems either use container_of() directly to cast from css to such data structure or has an accessor function wrapping such cast. As cgroup as whole is moving towards using css as the main interface handle, add and update such accessors to ease dealing with css's. All accessors explicitly handle NULL input and return NULL in those cases. While this looks like an extra branch in the code, as all controllers specific data structures have css as the first field, the casting doesn't involve any offsetting and the compiler can trivially optimize out the branch. * blkio, freezer, cpuset, cpu, cpuacct and net_cls didn't have such accessor. Added. * memory, hugetlb and devices already had one but didn't explicitly handle NULL input. Updated. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup_freezer.c | 11 +++++++---- kernel/cpuset.c | 11 +++++++---- kernel/sched/core.c | 8 ++++++-- kernel/sched/cpuacct.c | 11 +++++++---- 4 files changed, 27 insertions(+), 14 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup_freezer.c b/kernel/cgroup_freezer.c index 9d3f61566fec..1db686e47a22 100644 --- a/kernel/cgroup_freezer.c +++ b/kernel/cgroup_freezer.c @@ -45,16 +45,19 @@ struct freezer { spinlock_t lock; }; +static inline struct freezer *css_freezer(struct cgroup_subsys_state *css) +{ + return css ? container_of(css, struct freezer, css) : NULL; +} + static inline struct freezer *cgroup_freezer(struct cgroup *cgroup) { - return container_of(cgroup_css(cgroup, freezer_subsys_id), - struct freezer, css); + return css_freezer(cgroup_css(cgroup, freezer_subsys_id)); } static inline struct freezer *task_freezer(struct task_struct *task) { - return container_of(task_css(task, freezer_subsys_id), - struct freezer, css); + return css_freezer(task_css(task, freezer_subsys_id)); } static struct freezer *parent_freezer(struct freezer *freezer) diff --git a/kernel/cpuset.c b/kernel/cpuset.c index f7371341d42a..6e9cbdde25bd 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -114,18 +114,21 @@ struct cpuset { int relax_domain_level; }; +static inline struct cpuset *css_cs(struct cgroup_subsys_state *css) +{ + return css ? container_of(css, struct cpuset, css) : NULL; +} + /* Retrieve the cpuset for a cgroup */ static inline struct cpuset *cgroup_cs(struct cgroup *cgrp) { - return container_of(cgroup_css(cgrp, cpuset_subsys_id), - struct cpuset, css); + return css_cs(cgroup_css(cgrp, cpuset_subsys_id)); } /* Retrieve the cpuset for a task */ static inline struct cpuset *task_cs(struct task_struct *task) { - return container_of(task_css(task, cpuset_subsys_id), - struct cpuset, css); + return css_cs(task_css(task, cpuset_subsys_id)); } static inline struct cpuset *parent_cs(struct cpuset *cs) diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 323d907eac1a..5bccb0277129 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -7083,11 +7083,15 @@ int sched_rt_handler(struct ctl_table *table, int write, #ifdef CONFIG_CGROUP_SCHED +static inline struct task_group *css_tg(struct cgroup_subsys_state *css) +{ + return css ? container_of(css, struct task_group, css) : NULL; +} + /* return corresponding task_group object of a cgroup */ static inline struct task_group *cgroup_tg(struct cgroup *cgrp) { - return container_of(cgroup_css(cgrp, cpu_cgroup_subsys_id), - struct task_group, css); + return css_tg(cgroup_css(cgrp, cpu_cgroup_subsys_id)); } static struct cgroup_subsys_state *cpu_cgroup_css_alloc(struct cgroup *cgrp) diff --git a/kernel/sched/cpuacct.c b/kernel/sched/cpuacct.c index 4a210faaab77..8ccfa10cc89f 100644 --- a/kernel/sched/cpuacct.c +++ b/kernel/sched/cpuacct.c @@ -33,18 +33,21 @@ struct cpuacct { struct kernel_cpustat __percpu *cpustat; }; +static inline struct cpuacct *css_ca(struct cgroup_subsys_state *css) +{ + return css ? container_of(css, struct cpuacct, css) : NULL; +} + /* return cpu accounting group corresponding to this container */ static inline struct cpuacct *cgroup_ca(struct cgroup *cgrp) { - return container_of(cgroup_css(cgrp, cpuacct_subsys_id), - struct cpuacct, css); + return css_ca(cgroup_css(cgrp, cpuacct_subsys_id)); } /* return cpu accounting group to which this task belongs */ static inline struct cpuacct *task_ca(struct task_struct *tsk) { - return container_of(task_css(tsk, cpuacct_subsys_id), - struct cpuacct, css); + return css_ca(task_css(tsk, cpuacct_subsys_id)); } static inline struct cpuacct *__parent_ca(struct cpuacct *ca) -- cgit v1.2.3 From 6387698699afd72d6304566fb6ccf84bffe07c56 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:23 -0400 Subject: cgroup: add css_parent() Currently, controllers have to explicitly follow the cgroup hierarchy to find the parent of a given css. cgroup is moving towards using cgroup_subsys_state as the main controller interface construct, so let's provide a way to climb the hierarchy using just csses. This patch implements css_parent() which, given a css, returns its parent. The function is guarnateed to valid non-NULL parent css as long as the target css is not at the top of the hierarchy. freezer, cpuset, cpu, cpuacct, hugetlb, memory, net_cls and devices are converted to use css_parent() instead of accessing cgroup->parent directly. * __parent_ca() is dropped from cpuacct and its usage is replaced with parent_ca(). The only difference between the two was NULL test on cgroup->parent which is now embedded in css_parent() making the distinction moot. Note that eventually a css->parent field will be added to css and the NULL check in css_parent() will go away. This patch shouldn't cause any behavior differences. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup_freezer.c | 8 ++------ kernel/cpuset.c | 6 +----- kernel/sched/core.c | 9 +++------ kernel/sched/cpuacct.c | 11 ++--------- 4 files changed, 8 insertions(+), 26 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup_freezer.c b/kernel/cgroup_freezer.c index 1db686e47a22..657a73cd44c4 100644 --- a/kernel/cgroup_freezer.c +++ b/kernel/cgroup_freezer.c @@ -62,11 +62,7 @@ static inline struct freezer *task_freezer(struct task_struct *task) static struct freezer *parent_freezer(struct freezer *freezer) { - struct cgroup *pcg = freezer->css.cgroup->parent; - - if (pcg) - return cgroup_freezer(pcg); - return NULL; + return css_freezer(css_parent(&freezer->css)); } bool cgroup_freezing(struct task_struct *task) @@ -234,7 +230,7 @@ static void freezer_fork(struct task_struct *task) * The root cgroup is non-freezable, so we can skip the * following check. */ - if (!freezer->css.cgroup->parent) + if (!parent_freezer(freezer)) goto out; spin_lock_irq(&freezer->lock); diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 6e9cbdde25bd..259a4af37e69 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -133,11 +133,7 @@ static inline struct cpuset *task_cs(struct task_struct *task) static inline struct cpuset *parent_cs(struct cpuset *cs) { - struct cgroup *pcgrp = cs->css.cgroup->parent; - - if (pcgrp) - return cgroup_cs(pcgrp); - return NULL; + return css_cs(css_parent(&cs->css)); } #ifdef CONFIG_NUMA diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 5bccb0277129..7a10742b389a 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -7114,13 +7114,10 @@ static struct cgroup_subsys_state *cpu_cgroup_css_alloc(struct cgroup *cgrp) static int cpu_cgroup_css_online(struct cgroup *cgrp) { struct task_group *tg = cgroup_tg(cgrp); - struct task_group *parent; + struct task_group *parent = css_tg(css_parent(&tg->css)); - if (!cgrp->parent) - return 0; - - parent = cgroup_tg(cgrp->parent); - sched_online_group(tg, parent); + if (parent) + sched_online_group(tg, parent); return 0; } diff --git a/kernel/sched/cpuacct.c b/kernel/sched/cpuacct.c index 8ccfa10cc89f..f6926a149a71 100644 --- a/kernel/sched/cpuacct.c +++ b/kernel/sched/cpuacct.c @@ -50,16 +50,9 @@ static inline struct cpuacct *task_ca(struct task_struct *tsk) return css_ca(task_css(tsk, cpuacct_subsys_id)); } -static inline struct cpuacct *__parent_ca(struct cpuacct *ca) -{ - return cgroup_ca(ca->css.cgroup->parent); -} - static inline struct cpuacct *parent_ca(struct cpuacct *ca) { - if (!ca->css.cgroup->parent) - return NULL; - return cgroup_ca(ca->css.cgroup->parent); + return css_ca(css_parent(&ca->css)); } static DEFINE_PER_CPU(u64, root_cpuacct_cpuusage); @@ -284,7 +277,7 @@ void cpuacct_account_field(struct task_struct *p, int index, u64 val) while (ca != &root_cpuacct) { kcpustat = this_cpu_ptr(ca->cpustat); kcpustat->cpustat[index] += val; - ca = __parent_ca(ca); + ca = parent_ca(ca); } rcu_read_unlock(); } -- cgit v1.2.3 From eb95419b023abacb415e2a18fea899023ce7624d Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:23 -0400 Subject: cgroup: pass around cgroup_subsys_state instead of cgroup in subsystem methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup * in subsystem implementations for the following reasons. * With unified hierarchy, subsystems will be dynamically bound and unbound from cgroups and thus css's (cgroup_subsys_state) may be created and destroyed dynamically over the lifetime of a cgroup, which is different from the current state where all css's are allocated and destroyed together with the associated cgroup. This in turn means that cgroup_css() should be synchronized and may return NULL, making it more cumbersome to use. * Differing levels of per-subsystem granularity in the unified hierarchy means that the task and descendant iterators should behave differently depending on the specific subsystem the iteration is being performed for. * In majority of the cases, subsystems only care about its part in the cgroup hierarchy - ie. the hierarchy of css's. Subsystem methods often obtain the matching css pointer from the cgroup and don't bother with the cgroup pointer itself. Passing around css fits much better. This patch converts all cgroup_subsys methods to take @css instead of @cgroup. The conversions are mostly straight-forward. A few noteworthy changes are * ->css_alloc() now takes css of the parent cgroup rather than the pointer to the new cgroup as the css for the new cgroup doesn't exist yet. Knowing the parent css is enough for all the existing subsystems. * In kernel/cgroup.c::offline_css(), unnecessary open coded css dereference is replaced with local variable access. This patch shouldn't cause any behavior differences. v2: Unnecessary explicit cgrp->subsys[] deref in css_online() replaced with local variable @css as suggested by Li Zefan. Rebased on top of new for-3.12 which includes for-3.11-fixes so that ->css_free() invocation added by da0a12caff ("cgroup: fix a leak when percpu_ref_init() fails") is converted too. Suggested by Li Zefan. Signed-off-by: Tejun Heo Acked-by: Li Zefan Acked-by: Michal Hocko Acked-by: Vivek Goyal Acked-by: Aristeu Rozanski Acked-by: Daniel Wagner Cc: Peter Zijlstra Cc: Ingo Molnar Cc: Johannes Weiner Cc: Balbir Singh Cc: Matt Helsley Cc: Jens Axboe Cc: Steven Rostedt --- kernel/cgroup.c | 57 +++++++++++++++++++++++++++++-------------------- kernel/cgroup_freezer.c | 40 +++++++++++++++++----------------- kernel/cpuset.c | 39 +++++++++++++++++---------------- kernel/events/core.c | 18 +++++++++------- kernel/sched/core.c | 39 ++++++++++++++++----------------- kernel/sched/cpuacct.c | 9 ++++---- 6 files changed, 111 insertions(+), 91 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 4234428f1014..271d9a5cde5f 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -853,8 +853,11 @@ static void cgroup_free_fn(struct work_struct *work) /* * Release the subsystem state objects. */ - for_each_root_subsys(cgrp->root, ss) - ss->css_free(cgrp); + for_each_root_subsys(cgrp->root, ss) { + struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + + ss->css_free(css); + } cgrp->root->number_of_cgroups--; mutex_unlock(&cgroup_mutex); @@ -1056,7 +1059,7 @@ static int rebind_subsystems(struct cgroupfs_root *root, list_move(&ss->sibling, &root->subsys_list); ss->root = root; if (ss->bind) - ss->bind(cgrp); + ss->bind(cgrp->subsys[i]); /* refcount was already taken, and we're keeping it */ root->subsys_mask |= bit; @@ -1066,7 +1069,7 @@ static int rebind_subsystems(struct cgroupfs_root *root, BUG_ON(cgrp->subsys[i]->cgroup != cgrp); if (ss->bind) - ss->bind(cgroup_dummy_top); + ss->bind(cgroup_dummy_top->subsys[i]); cgroup_dummy_top->subsys[i]->cgroup = cgroup_dummy_top; cgrp->subsys[i] = NULL; cgroup_subsys[i]->root = &cgroup_dummy_root; @@ -2049,8 +2052,10 @@ static int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk, * step 1: check that we can legitimately attach to the cgroup. */ for_each_root_subsys(root, ss) { + struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + if (ss->can_attach) { - retval = ss->can_attach(cgrp, &tset); + retval = ss->can_attach(css, &tset); if (retval) { failed_ss = ss; goto out_cancel_attach; @@ -2089,8 +2094,10 @@ static int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk, * step 4: do subsystem attach callbacks. */ for_each_root_subsys(root, ss) { + struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + if (ss->attach) - ss->attach(cgrp, &tset); + ss->attach(css, &tset); } /* @@ -2109,10 +2116,12 @@ out_put_css_set_refs: out_cancel_attach: if (retval) { for_each_root_subsys(root, ss) { + struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + if (ss == failed_ss) break; if (ss->cancel_attach) - ss->cancel_attach(cgrp, &tset); + ss->cancel_attach(css, &tset); } } out_free_group_list: @@ -4206,14 +4215,15 @@ static void init_cgroup_css(struct cgroup_subsys_state *css, /* invoke ->css_online() on a new CSS and mark it online if successful */ static int online_css(struct cgroup_subsys *ss, struct cgroup *cgrp) { + struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; int ret = 0; lockdep_assert_held(&cgroup_mutex); if (ss->css_online) - ret = ss->css_online(cgrp); + ret = ss->css_online(css); if (!ret) - cgrp->subsys[ss->subsys_id]->flags |= CSS_ONLINE; + css->flags |= CSS_ONLINE; return ret; } @@ -4228,9 +4238,9 @@ static void offline_css(struct cgroup_subsys *ss, struct cgroup *cgrp) return; if (ss->css_offline) - ss->css_offline(cgrp); + ss->css_offline(css); - cgrp->subsys[ss->subsys_id]->flags &= ~CSS_ONLINE; + css->flags &= ~CSS_ONLINE; } /* @@ -4305,7 +4315,7 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, for_each_root_subsys(root, ss) { struct cgroup_subsys_state *css; - css = ss->css_alloc(cgrp); + css = ss->css_alloc(parent->subsys[ss->subsys_id]); if (IS_ERR(css)) { err = PTR_ERR(css); goto err_free_all; @@ -4313,7 +4323,7 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, err = percpu_ref_init(&css->refcnt, css_release); if (err) { - ss->css_free(cgrp); + ss->css_free(css); goto err_free_all; } @@ -4386,7 +4396,7 @@ err_free_all: if (css) { percpu_ref_cancel_init(&css->refcnt); - ss->css_free(cgrp); + ss->css_free(css); } } mutex_unlock(&cgroup_mutex); @@ -4641,7 +4651,7 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss) /* Create the top cgroup state for this subsystem */ list_add(&ss->sibling, &cgroup_dummy_root.subsys_list); ss->root = &cgroup_dummy_root; - css = ss->css_alloc(cgroup_dummy_top); + css = ss->css_alloc(cgroup_dummy_top->subsys[ss->subsys_id]); /* We don't handle early failures gracefully */ BUG_ON(IS_ERR(css)); init_cgroup_css(css, ss, cgroup_dummy_top); @@ -4720,7 +4730,7 @@ int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss) * struct, so this can happen first (i.e. before the dummy root * attachment). */ - css = ss->css_alloc(cgroup_dummy_top); + css = ss->css_alloc(cgroup_dummy_top->subsys[ss->subsys_id]); if (IS_ERR(css)) { /* failure case - need to deassign the cgroup_subsys[] slot. */ cgroup_subsys[ss->subsys_id] = NULL; @@ -4836,7 +4846,7 @@ void cgroup_unload_subsys(struct cgroup_subsys *ss) * the cgrp->subsys pointer to find their state. note that this * also takes care of freeing the css_id. */ - ss->css_free(cgroup_dummy_top); + ss->css_free(cgroup_dummy_top->subsys[ss->subsys_id]); cgroup_dummy_top->subsys[ss->subsys_id] = NULL; mutex_unlock(&cgroup_mutex); @@ -5192,10 +5202,10 @@ void cgroup_exit(struct task_struct *tsk, int run_callbacks) */ for_each_builtin_subsys(ss, i) { if (ss->exit) { - struct cgroup *old_cgrp = cset->subsys[i]->cgroup; - struct cgroup *cgrp = task_cgroup(tsk, i); + struct cgroup_subsys_state *old_css = cset->subsys[i]; + struct cgroup_subsys_state *css = task_css(tsk, i); - ss->exit(cgrp, old_cgrp, tsk); + ss->exit(css, old_css, tsk); } } } @@ -5529,7 +5539,8 @@ struct cgroup_subsys_state *cgroup_css_from_dir(struct file *f, int id) } #ifdef CONFIG_CGROUP_DEBUG -static struct cgroup_subsys_state *debug_css_alloc(struct cgroup *cgrp) +static struct cgroup_subsys_state * +debug_css_alloc(struct cgroup_subsys_state *parent_css) { struct cgroup_subsys_state *css = kzalloc(sizeof(*css), GFP_KERNEL); @@ -5539,9 +5550,9 @@ static struct cgroup_subsys_state *debug_css_alloc(struct cgroup *cgrp) return css; } -static void debug_css_free(struct cgroup *cgrp) +static void debug_css_free(struct cgroup_subsys_state *css) { - kfree(cgrp->subsys[debug_subsys_id]); + kfree(css); } static u64 debug_taskcount_read(struct cgroup *cgrp, struct cftype *cft) diff --git a/kernel/cgroup_freezer.c b/kernel/cgroup_freezer.c index 657a73cd44c4..f03a85719c3c 100644 --- a/kernel/cgroup_freezer.c +++ b/kernel/cgroup_freezer.c @@ -91,7 +91,8 @@ static const char *freezer_state_strs(unsigned int state) struct cgroup_subsys freezer_subsys; -static struct cgroup_subsys_state *freezer_css_alloc(struct cgroup *cgroup) +static struct cgroup_subsys_state * +freezer_css_alloc(struct cgroup_subsys_state *parent_css) { struct freezer *freezer; @@ -104,16 +105,16 @@ static struct cgroup_subsys_state *freezer_css_alloc(struct cgroup *cgroup) } /** - * freezer_css_online - commit creation of a freezer cgroup - * @cgroup: cgroup being created + * freezer_css_online - commit creation of a freezer css + * @css: css being created * - * We're committing to creation of @cgroup. Mark it online and inherit + * We're committing to creation of @css. Mark it online and inherit * parent's freezing state while holding both parent's and our * freezer->lock. */ -static int freezer_css_online(struct cgroup *cgroup) +static int freezer_css_online(struct cgroup_subsys_state *css) { - struct freezer *freezer = cgroup_freezer(cgroup); + struct freezer *freezer = css_freezer(css); struct freezer *parent = parent_freezer(freezer); /* @@ -140,15 +141,15 @@ static int freezer_css_online(struct cgroup *cgroup) } /** - * freezer_css_offline - initiate destruction of @cgroup - * @cgroup: cgroup being destroyed + * freezer_css_offline - initiate destruction of a freezer css + * @css: css being destroyed * - * @cgroup is going away. Mark it dead and decrement system_freezing_count - * if it was holding one. + * @css is going away. Mark it dead and decrement system_freezing_count if + * it was holding one. */ -static void freezer_css_offline(struct cgroup *cgroup) +static void freezer_css_offline(struct cgroup_subsys_state *css) { - struct freezer *freezer = cgroup_freezer(cgroup); + struct freezer *freezer = css_freezer(css); spin_lock_irq(&freezer->lock); @@ -160,9 +161,9 @@ static void freezer_css_offline(struct cgroup *cgroup) spin_unlock_irq(&freezer->lock); } -static void freezer_css_free(struct cgroup *cgroup) +static void freezer_css_free(struct cgroup_subsys_state *css) { - kfree(cgroup_freezer(cgroup)); + kfree(css_freezer(css)); } /* @@ -174,25 +175,26 @@ static void freezer_css_free(struct cgroup *cgroup) * @freezer->lock. freezer_attach() makes the new tasks conform to the * current state and all following state changes can see the new tasks. */ -static void freezer_attach(struct cgroup *new_cgrp, struct cgroup_taskset *tset) +static void freezer_attach(struct cgroup_subsys_state *new_css, + struct cgroup_taskset *tset) { - struct freezer *freezer = cgroup_freezer(new_cgrp); + struct freezer *freezer = css_freezer(new_css); struct task_struct *task; bool clear_frozen = false; spin_lock_irq(&freezer->lock); /* - * Make the new tasks conform to the current state of @new_cgrp. + * Make the new tasks conform to the current state of @new_css. * For simplicity, when migrating any task to a FROZEN cgroup, we * revert it to FREEZING and let update_if_frozen() determine the * correct state later. * - * Tasks in @tset are on @new_cgrp but may not conform to its + * Tasks in @tset are on @new_css but may not conform to its * current state before executing the following - !frozen tasks may * be visible in a FROZEN cgroup and frozen tasks in a THAWED one. */ - cgroup_taskset_for_each(task, new_cgrp, tset) { + cgroup_taskset_for_each(task, new_css->cgroup, tset) { if (!(freezer->state & CGROUP_FREEZING)) { __thaw_task(task); } else { diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 259a4af37e69..8ce3fdc3dfcc 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -1455,9 +1455,10 @@ static int fmeter_getrate(struct fmeter *fmp) } /* Called by cgroups to determine if a cpuset is usable; cpuset_mutex held */ -static int cpuset_can_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) +static int cpuset_can_attach(struct cgroup_subsys_state *css, + struct cgroup_taskset *tset) { - struct cpuset *cs = cgroup_cs(cgrp); + struct cpuset *cs = css_cs(css); struct task_struct *task; int ret; @@ -1468,11 +1469,11 @@ static int cpuset_can_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) * flag is set. */ ret = -ENOSPC; - if (!cgroup_sane_behavior(cgrp) && + if (!cgroup_sane_behavior(css->cgroup) && (cpumask_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed))) goto out_unlock; - cgroup_taskset_for_each(task, cgrp, tset) { + cgroup_taskset_for_each(task, css->cgroup, tset) { /* * Kthreads which disallow setaffinity shouldn't be moved * to a new cpuset; we don't want to change their cpu @@ -1501,11 +1502,11 @@ out_unlock: return ret; } -static void cpuset_cancel_attach(struct cgroup *cgrp, +static void cpuset_cancel_attach(struct cgroup_subsys_state *css, struct cgroup_taskset *tset) { mutex_lock(&cpuset_mutex); - cgroup_cs(cgrp)->attach_in_progress--; + css_cs(css)->attach_in_progress--; mutex_unlock(&cpuset_mutex); } @@ -1516,7 +1517,8 @@ static void cpuset_cancel_attach(struct cgroup *cgrp, */ static cpumask_var_t cpus_attach; -static void cpuset_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) +static void cpuset_attach(struct cgroup_subsys_state *css, + struct cgroup_taskset *tset) { /* static buf protected by cpuset_mutex */ static nodemask_t cpuset_attach_nodemask_to; @@ -1524,7 +1526,7 @@ static void cpuset_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) struct task_struct *task; struct task_struct *leader = cgroup_taskset_first(tset); struct cgroup *oldcgrp = cgroup_taskset_cur_cgroup(tset); - struct cpuset *cs = cgroup_cs(cgrp); + struct cpuset *cs = css_cs(css); struct cpuset *oldcs = cgroup_cs(oldcgrp); struct cpuset *cpus_cs = effective_cpumask_cpuset(cs); struct cpuset *mems_cs = effective_nodemask_cpuset(cs); @@ -1539,7 +1541,7 @@ static void cpuset_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) guarantee_online_mems(mems_cs, &cpuset_attach_nodemask_to); - cgroup_taskset_for_each(task, cgrp, tset) { + cgroup_taskset_for_each(task, css->cgroup, tset) { /* * can_attach beforehand should guarantee that this doesn't * fail. TODO: have a better way to handle failure here @@ -1940,11 +1942,12 @@ static struct cftype files[] = { * cgrp: control group that the new cpuset will be part of */ -static struct cgroup_subsys_state *cpuset_css_alloc(struct cgroup *cgrp) +static struct cgroup_subsys_state * +cpuset_css_alloc(struct cgroup_subsys_state *parent_css) { struct cpuset *cs; - if (!cgrp->parent) + if (!parent_css) return &top_cpuset.css; cs = kzalloc(sizeof(*cs), GFP_KERNEL); @@ -1964,9 +1967,9 @@ static struct cgroup_subsys_state *cpuset_css_alloc(struct cgroup *cgrp) return &cs->css; } -static int cpuset_css_online(struct cgroup *cgrp) +static int cpuset_css_online(struct cgroup_subsys_state *css) { - struct cpuset *cs = cgroup_cs(cgrp); + struct cpuset *cs = css_cs(css); struct cpuset *parent = parent_cs(cs); struct cpuset *tmp_cs; struct cgroup *pos_cgrp; @@ -1984,7 +1987,7 @@ static int cpuset_css_online(struct cgroup *cgrp) number_of_cpusets++; - if (!test_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags)) + if (!test_bit(CGRP_CPUSET_CLONE_CHILDREN, &css->cgroup->flags)) goto out_unlock; /* @@ -2024,9 +2027,9 @@ out_unlock: * will call rebuild_sched_domains_locked(). */ -static void cpuset_css_offline(struct cgroup *cgrp) +static void cpuset_css_offline(struct cgroup_subsys_state *css) { - struct cpuset *cs = cgroup_cs(cgrp); + struct cpuset *cs = css_cs(css); mutex_lock(&cpuset_mutex); @@ -2039,9 +2042,9 @@ static void cpuset_css_offline(struct cgroup *cgrp) mutex_unlock(&cpuset_mutex); } -static void cpuset_css_free(struct cgroup *cgrp) +static void cpuset_css_free(struct cgroup_subsys_state *css) { - struct cpuset *cs = cgroup_cs(cgrp); + struct cpuset *cs = css_cs(css); free_cpumask_var(cs->cpus_allowed); kfree(cs); diff --git a/kernel/events/core.c b/kernel/events/core.c index 414c61f4d776..9705a0ed1dce 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -7778,7 +7778,8 @@ unlock: device_initcall(perf_event_sysfs_init); #ifdef CONFIG_CGROUP_PERF -static struct cgroup_subsys_state *perf_cgroup_css_alloc(struct cgroup *cont) +static struct cgroup_subsys_state * +perf_cgroup_css_alloc(struct cgroup_subsys_state *parent_css) { struct perf_cgroup *jc; @@ -7795,11 +7796,10 @@ static struct cgroup_subsys_state *perf_cgroup_css_alloc(struct cgroup *cont) return &jc->css; } -static void perf_cgroup_css_free(struct cgroup *cont) +static void perf_cgroup_css_free(struct cgroup_subsys_state *css) { - struct perf_cgroup *jc; - jc = container_of(cgroup_css(cont, perf_subsys_id), - struct perf_cgroup, css); + struct perf_cgroup *jc = container_of(css, struct perf_cgroup, css); + free_percpu(jc->info); kfree(jc); } @@ -7811,15 +7811,17 @@ static int __perf_cgroup_move(void *info) return 0; } -static void perf_cgroup_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) +static void perf_cgroup_attach(struct cgroup_subsys_state *css, + struct cgroup_taskset *tset) { struct task_struct *task; - cgroup_taskset_for_each(task, cgrp, tset) + cgroup_taskset_for_each(task, css->cgroup, tset) task_function_call(task, __perf_cgroup_move, task); } -static void perf_cgroup_exit(struct cgroup *cgrp, struct cgroup *old_cgrp, +static void perf_cgroup_exit(struct cgroup_subsys_state *css, + struct cgroup_subsys_state *old_css, struct task_struct *task) { /* diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 7a10742b389a..622b7efc5ade 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -7094,16 +7094,17 @@ static inline struct task_group *cgroup_tg(struct cgroup *cgrp) return css_tg(cgroup_css(cgrp, cpu_cgroup_subsys_id)); } -static struct cgroup_subsys_state *cpu_cgroup_css_alloc(struct cgroup *cgrp) +static struct cgroup_subsys_state * +cpu_cgroup_css_alloc(struct cgroup_subsys_state *parent_css) { - struct task_group *tg, *parent; + struct task_group *parent = css_tg(parent_css); + struct task_group *tg; - if (!cgrp->parent) { + if (!parent) { /* This is early initialization for the top cgroup */ return &root_task_group.css; } - parent = cgroup_tg(cgrp->parent); tg = sched_create_group(parent); if (IS_ERR(tg)) return ERR_PTR(-ENOMEM); @@ -7111,38 +7112,38 @@ static struct cgroup_subsys_state *cpu_cgroup_css_alloc(struct cgroup *cgrp) return &tg->css; } -static int cpu_cgroup_css_online(struct cgroup *cgrp) +static int cpu_cgroup_css_online(struct cgroup_subsys_state *css) { - struct task_group *tg = cgroup_tg(cgrp); - struct task_group *parent = css_tg(css_parent(&tg->css)); + struct task_group *tg = css_tg(css); + struct task_group *parent = css_tg(css_parent(css)); if (parent) sched_online_group(tg, parent); return 0; } -static void cpu_cgroup_css_free(struct cgroup *cgrp) +static void cpu_cgroup_css_free(struct cgroup_subsys_state *css) { - struct task_group *tg = cgroup_tg(cgrp); + struct task_group *tg = css_tg(css); sched_destroy_group(tg); } -static void cpu_cgroup_css_offline(struct cgroup *cgrp) +static void cpu_cgroup_css_offline(struct cgroup_subsys_state *css) { - struct task_group *tg = cgroup_tg(cgrp); + struct task_group *tg = css_tg(css); sched_offline_group(tg); } -static int cpu_cgroup_can_attach(struct cgroup *cgrp, +static int cpu_cgroup_can_attach(struct cgroup_subsys_state *css, struct cgroup_taskset *tset) { struct task_struct *task; - cgroup_taskset_for_each(task, cgrp, tset) { + cgroup_taskset_for_each(task, css->cgroup, tset) { #ifdef CONFIG_RT_GROUP_SCHED - if (!sched_rt_can_attach(cgroup_tg(cgrp), task)) + if (!sched_rt_can_attach(css_tg(css), task)) return -EINVAL; #else /* We don't support RT-tasks being in separate groups */ @@ -7153,18 +7154,18 @@ static int cpu_cgroup_can_attach(struct cgroup *cgrp, return 0; } -static void cpu_cgroup_attach(struct cgroup *cgrp, +static void cpu_cgroup_attach(struct cgroup_subsys_state *css, struct cgroup_taskset *tset) { struct task_struct *task; - cgroup_taskset_for_each(task, cgrp, tset) + cgroup_taskset_for_each(task, css->cgroup, tset) sched_move_task(task); } -static void -cpu_cgroup_exit(struct cgroup *cgrp, struct cgroup *old_cgrp, - struct task_struct *task) +static void cpu_cgroup_exit(struct cgroup_subsys_state *css, + struct cgroup_subsys_state *old_css, + struct task_struct *task) { /* * cgroup_exit() is called in the copy_process() failure path. diff --git a/kernel/sched/cpuacct.c b/kernel/sched/cpuacct.c index f6926a149a71..1b784d9b3630 100644 --- a/kernel/sched/cpuacct.c +++ b/kernel/sched/cpuacct.c @@ -62,11 +62,12 @@ static struct cpuacct root_cpuacct = { }; /* create a new cpu accounting group */ -static struct cgroup_subsys_state *cpuacct_css_alloc(struct cgroup *cgrp) +static struct cgroup_subsys_state * +cpuacct_css_alloc(struct cgroup_subsys_state *parent_css) { struct cpuacct *ca; - if (!cgrp->parent) + if (!parent_css) return &root_cpuacct.css; ca = kzalloc(sizeof(*ca), GFP_KERNEL); @@ -92,9 +93,9 @@ out: } /* destroy an existing cpu accounting group */ -static void cpuacct_css_free(struct cgroup *cgrp) +static void cpuacct_css_free(struct cgroup_subsys_state *css) { - struct cpuacct *ca = cgroup_ca(cgrp); + struct cpuacct *ca = css_ca(css); free_percpu(ca->cpustat); free_percpu(ca->cpuusage); -- cgit v1.2.3 From 2bb566cb68dfafad328af666ebadf0e49accd6ca Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:23 -0400 Subject: cgroup: add subsys backlink pointer to cftype cgroup is transitioning to using css (cgroup_subsys_state) instead of cgroup as the primary subsystem handle. The cgroupfs file interface will be converted to use css's which requires finding out the subsystem from cftype so that the matching css can be determined from the cgroup. This patch adds cftype->ss which points to the subsystem the file belongs to. The field is initialized while a cftype is being registered. This makes it unnecessary to explicitly specify the subsystem for other cftype handling functions. @ss argument dropped from various cftype handling functions. This patch shouldn't introduce any behavior differences. Signed-off-by: Tejun Heo Acked-by: Li Zefan Acked-by: Vivek Goyal Cc: Jens Axboe --- kernel/cgroup.c | 78 +++++++++++++++++++++++++++++++-------------------------- 1 file changed, 43 insertions(+), 35 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 271d9a5cde5f..c4bc8dac3b1d 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -219,8 +219,8 @@ static struct cftype cgroup_base_files[]; static void cgroup_offline_fn(struct work_struct *work); static int cgroup_destroy_locked(struct cgroup *cgrp); -static int cgroup_addrm_files(struct cgroup *cgrp, struct cgroup_subsys *subsys, - struct cftype cfts[], bool is_add); +static int cgroup_addrm_files(struct cgroup *cgrp, struct cftype cfts[], + bool is_add); /* convenient tests for these bits */ static inline bool cgroup_is_dead(const struct cgroup *cgrp) @@ -974,7 +974,7 @@ static void cgroup_clear_dir(struct cgroup *cgrp, unsigned long subsys_mask) if (!test_bit(i, &subsys_mask)) continue; list_for_each_entry(set, &ss->cftsets, node) - cgroup_addrm_files(cgrp, NULL, set->cfts, false); + cgroup_addrm_files(cgrp, set->cfts, false); } } @@ -1623,7 +1623,7 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, */ cred = override_creds(&init_cred); - ret = cgroup_addrm_files(root_cgrp, NULL, cgroup_base_files, true); + ret = cgroup_addrm_files(root_cgrp, cgroup_base_files, true); if (ret) goto rm_base_files; @@ -1681,7 +1681,7 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, rm_base_files: free_cgrp_cset_links(&tmp_links); - cgroup_addrm_files(&root->top_cgroup, NULL, cgroup_base_files, false); + cgroup_addrm_files(&root->top_cgroup, cgroup_base_files, false); revert_creds(cred); unlock_drop: cgroup_exit_root_id(root); @@ -2694,8 +2694,7 @@ static umode_t cgroup_file_mode(const struct cftype *cft) return mode; } -static int cgroup_add_file(struct cgroup *cgrp, struct cgroup_subsys *subsys, - struct cftype *cft) +static int cgroup_add_file(struct cgroup *cgrp, struct cftype *cft) { struct dentry *dir = cgrp->dentry; struct cgroup *parent = __d_cgrp(dir); @@ -2705,8 +2704,8 @@ static int cgroup_add_file(struct cgroup *cgrp, struct cgroup_subsys *subsys, umode_t mode; char name[MAX_CGROUP_TYPE_NAMELEN + MAX_CFTYPE_NAME + 2] = { 0 }; - if (subsys && !(cgrp->root->flags & CGRP_ROOT_NOPREFIX)) { - strcpy(name, subsys->name); + if (cft->ss && !(cgrp->root->flags & CGRP_ROOT_NOPREFIX)) { + strcpy(name, cft->ss->name); strcat(name, "."); } strcat(name, cft->name); @@ -2743,17 +2742,16 @@ out: /** * cgroup_addrm_files - add or remove files to a cgroup directory * @cgrp: the target cgroup - * @subsys: the subsystem of files to be added * @cfts: array of cftypes to be added * @is_add: whether to add or remove * * Depending on @is_add, add or remove files defined by @cfts on @cgrp. - * All @cfts should belong to @subsys. For removals, this function never - * fails. If addition fails, this function doesn't remove files already - * added. The caller is responsible for cleaning up. + * For removals, this function never fails. If addition fails, this + * function doesn't remove files already added. The caller is responsible + * for cleaning up. */ -static int cgroup_addrm_files(struct cgroup *cgrp, struct cgroup_subsys *subsys, - struct cftype cfts[], bool is_add) +static int cgroup_addrm_files(struct cgroup *cgrp, struct cftype cfts[], + bool is_add) { struct cftype *cft; int ret; @@ -2771,7 +2769,7 @@ static int cgroup_addrm_files(struct cgroup *cgrp, struct cgroup_subsys *subsys, continue; if (is_add) { - ret = cgroup_add_file(cgrp, subsys, cft); + ret = cgroup_add_file(cgrp, cft); if (ret) { pr_warn("cgroup_addrm_files: failed to add %s, err=%d\n", cft->name, ret); @@ -2796,11 +2794,11 @@ static void cgroup_cfts_prepare(void) mutex_lock(&cgroup_mutex); } -static int cgroup_cfts_commit(struct cgroup_subsys *ss, - struct cftype *cfts, bool is_add) +static int cgroup_cfts_commit(struct cftype *cfts, bool is_add) __releases(&cgroup_mutex) { LIST_HEAD(pending); + struct cgroup_subsys *ss = cfts[0].ss; struct cgroup *cgrp, *root = &ss->root->top_cgroup; struct super_block *sb = ss->root->sb; struct dentry *prev = NULL; @@ -2828,7 +2826,7 @@ static int cgroup_cfts_commit(struct cgroup_subsys *ss, inode = root->dentry->d_inode; mutex_lock(&inode->i_mutex); mutex_lock(&cgroup_mutex); - ret = cgroup_addrm_files(root, ss, cfts, is_add); + ret = cgroup_addrm_files(root, cfts, is_add); mutex_unlock(&cgroup_mutex); mutex_unlock(&inode->i_mutex); @@ -2851,7 +2849,7 @@ static int cgroup_cfts_commit(struct cgroup_subsys *ss, mutex_lock(&inode->i_mutex); mutex_lock(&cgroup_mutex); if (cgrp->serial_nr < update_before && !cgroup_is_dead(cgrp)) - ret = cgroup_addrm_files(cgrp, ss, cfts, is_add); + ret = cgroup_addrm_files(cgrp, cfts, is_add); mutex_unlock(&cgroup_mutex); mutex_unlock(&inode->i_mutex); @@ -2883,51 +2881,56 @@ out_deact: int cgroup_add_cftypes(struct cgroup_subsys *ss, struct cftype *cfts) { struct cftype_set *set; + struct cftype *cft; int ret; set = kzalloc(sizeof(*set), GFP_KERNEL); if (!set) return -ENOMEM; + for (cft = cfts; cft->name[0] != '\0'; cft++) + cft->ss = ss; + cgroup_cfts_prepare(); set->cfts = cfts; list_add_tail(&set->node, &ss->cftsets); - ret = cgroup_cfts_commit(ss, cfts, true); + ret = cgroup_cfts_commit(cfts, true); if (ret) - cgroup_rm_cftypes(ss, cfts); + cgroup_rm_cftypes(cfts); return ret; } EXPORT_SYMBOL_GPL(cgroup_add_cftypes); /** * cgroup_rm_cftypes - remove an array of cftypes from a subsystem - * @ss: target cgroup subsystem * @cfts: zero-length name terminated array of cftypes * - * Unregister @cfts from @ss. Files described by @cfts are removed from - * all existing cgroups to which @ss is attached and all future cgroups - * won't have them either. This function can be called anytime whether @ss - * is attached or not. + * Unregister @cfts. Files described by @cfts are removed from all + * existing cgroups and all future cgroups won't have them either. This + * function can be called anytime whether @cfts' subsys is attached or not. * * Returns 0 on successful unregistration, -ENOENT if @cfts is not - * registered with @ss. + * registered. */ -int cgroup_rm_cftypes(struct cgroup_subsys *ss, struct cftype *cfts) +int cgroup_rm_cftypes(struct cftype *cfts) { struct cftype_set *set; + if (!cfts || !cfts[0].ss) + return -ENOENT; + cgroup_cfts_prepare(); - list_for_each_entry(set, &ss->cftsets, node) { + list_for_each_entry(set, &cfts[0].ss->cftsets, node) { if (set->cfts == cfts) { list_del(&set->node); kfree(set); - cgroup_cfts_commit(ss, cfts, false); + cgroup_cfts_commit(cfts, false); return 0; } } - cgroup_cfts_commit(ss, NULL, false); + cgroup_cfts_commit(NULL, false); return -ENOENT; } @@ -4148,7 +4151,7 @@ static int cgroup_populate_dir(struct cgroup *cgrp, unsigned long subsys_mask) continue; list_for_each_entry(set, &ss->cftsets, node) { - ret = cgroup_addrm_files(cgrp, ss, set->cfts, true); + ret = cgroup_addrm_files(cgrp, set->cfts, true); if (ret < 0) goto err; } @@ -4377,7 +4380,7 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, idr_replace(&root->cgroup_idr, cgrp, cgrp->id); - err = cgroup_addrm_files(cgrp, NULL, cgroup_base_files, true); + err = cgroup_addrm_files(cgrp, cgroup_base_files, true); if (err) goto err_destroy; @@ -4538,7 +4541,7 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) * but we aren't quite done with @cgrp yet, so hold onto it. */ cgroup_clear_dir(cgrp, cgrp->root->subsys_mask); - cgroup_addrm_files(cgrp, NULL, cgroup_base_files, false); + cgroup_addrm_files(cgrp, cgroup_base_files, false); dget(d); cgroup_d_remove_dir(d); @@ -4632,6 +4635,11 @@ static void __init_or_module cgroup_init_cftsets(struct cgroup_subsys *ss) * deregistration. */ if (ss->base_cftypes) { + struct cftype *cft; + + for (cft = ss->base_cftypes; cft->name[0] != '\0'; cft++) + cft->ss = ss; + ss->base_cftset.cfts = ss->base_cftypes; list_add_tail(&ss->base_cftset.node, &ss->cftsets); } -- cgit v1.2.3 From f7d58818ba4249f04a83b73aaac135640050bb4f Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:23 -0400 Subject: cgroup: pin cgroup_subsys_state when opening a cgroupfs file Previously, each file read/write operation relied on the inode reference count pinning the cgroup and simply checked whether the cgroup was marked dead before proceeding to invoke the per-subsystem callback. This was rather silly as it didn't have any synchronization or css pinning around the check and the cgroup may be removed and all css refs drained between the DEAD check and actual method invocation. This patch pins the css between open() and release() so that it is guaranteed to be alive for all file operations and remove the silly DEAD checks from cgroup_file_read/write(). Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 43 ++++++++++++++++++++++++++++++++----------- 1 file changed, 32 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index c4bc8dac3b1d..583f8f66a7e1 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -2277,6 +2277,17 @@ static int cgroup_sane_behavior_show(struct cgroup *cgrp, struct cftype *cft, return 0; } +/* return the css for the given cgroup file */ +static struct cgroup_subsys_state *cgroup_file_css(struct cfent *cfe) +{ + struct cftype *cft = cfe->type; + struct cgroup *cgrp = __d_cgrp(cfe->dentry->d_parent); + + if (cft->ss) + return cgrp->subsys[cft->ss->subsys_id]; + return NULL; +} + /* A buffer size big enough for numbers or short strings */ #define CGROUP_LOCAL_BUFFER_SIZE 64 @@ -2354,8 +2365,6 @@ static ssize_t cgroup_file_write(struct file *file, const char __user *buf, struct cftype *cft = __d_cft(file->f_dentry); struct cgroup *cgrp = __d_cgrp(file->f_dentry->d_parent); - if (cgroup_is_dead(cgrp)) - return -ENODEV; if (cft->write) return cft->write(cgrp, cft, file, buf, nbytes, ppos); if (cft->write_u64 || cft->write_s64) @@ -2399,9 +2408,6 @@ static ssize_t cgroup_file_read(struct file *file, char __user *buf, struct cftype *cft = __d_cft(file->f_dentry); struct cgroup *cgrp = __d_cgrp(file->f_dentry->d_parent); - if (cgroup_is_dead(cgrp)) - return -ENODEV; - if (cft->read) return cft->read(cgrp, cft, file, buf, nbytes, ppos); if (cft->read_u64) @@ -2447,15 +2453,22 @@ static const struct file_operations cgroup_seqfile_operations = { static int cgroup_file_open(struct inode *inode, struct file *file) { + struct cfent *cfe = __d_cfe(file->f_dentry); + struct cftype *cft = __d_cft(file->f_dentry); + struct cgroup_subsys_state *css = cgroup_file_css(cfe); int err; - struct cfent *cfe; - struct cftype *cft; err = generic_file_open(inode, file); if (err) return err; - cfe = __d_cfe(file->f_dentry); - cft = cfe->type; + + /* + * If the file belongs to a subsystem, pin the css. Will be + * unpinned either on open failure or release. This ensures that + * @css stays alive for all file operations. + */ + if (css && !css_tryget(css)) + return -ENODEV; if (cft->read_map || cft->read_seq_string) { file->f_op = &cgroup_seqfile_operations; @@ -2464,15 +2477,23 @@ static int cgroup_file_open(struct inode *inode, struct file *file) err = cft->open(inode, file); } + if (css && err) + css_put(css); return err; } static int cgroup_file_release(struct inode *inode, struct file *file) { + struct cfent *cfe = __d_cfe(file->f_dentry); struct cftype *cft = __d_cft(file->f_dentry); + struct cgroup_subsys_state *css = cgroup_file_css(cfe); + int ret = 0; + if (cft->release) - return cft->release(inode, file); - return 0; + ret = cft->release(inode, file); + if (css) + css_put(css); + return ret; } /* -- cgit v1.2.3 From 67f4c36f83455b253445b2cb28ac9a2c4f85d99a Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:24 -0400 Subject: cgroup: add cgroup->dummy_css cgroup subsystem API is being converted to use css (cgroup_subsys_state) as the main handle, which makes things a bit awkward for subsystem agnostic core features - the "cgroup.*" interface files and various iterations - a bit awkward as they don't have a css to use. This patch adds cgroup->dummy_css which has NULL ->ss and whose only role is pointing back to the cgroup. This will be used to support subsystem agnostic features on the coming css based API. css_parent() is updated to handle dummy_css's. Note that css will soon grow its own ->parent field and css_parent() will be made trivial. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 583f8f66a7e1..c049992f1ffa 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1365,6 +1365,7 @@ static void init_cgroup_housekeeping(struct cgroup *cgrp) INIT_LIST_HEAD(&cgrp->release_list); INIT_LIST_HEAD(&cgrp->pidlists); mutex_init(&cgrp->pidlist_mutex); + cgrp->dummy_css.cgroup = cgrp; INIT_LIST_HEAD(&cgrp->event_list); spin_lock_init(&cgrp->event_list_lock); simple_xattrs_init(&cgrp->xattrs); @@ -2285,7 +2286,7 @@ static struct cgroup_subsys_state *cgroup_file_css(struct cfent *cfe) if (cft->ss) return cgrp->subsys[cft->ss->subsys_id]; - return NULL; + return &cgrp->dummy_css; } /* A buffer size big enough for numbers or short strings */ @@ -2467,7 +2468,7 @@ static int cgroup_file_open(struct inode *inode, struct file *file) * unpinned either on open failure or release. This ensures that * @css stays alive for all file operations. */ - if (css && !css_tryget(css)) + if (css->ss && !css_tryget(css)) return -ENODEV; if (cft->read_map || cft->read_seq_string) { @@ -2477,7 +2478,7 @@ static int cgroup_file_open(struct inode *inode, struct file *file) err = cft->open(inode, file); } - if (css && err) + if (css->ss && err) css_put(css); return err; } @@ -2491,7 +2492,7 @@ static int cgroup_file_release(struct inode *inode, struct file *file) if (cft->release) ret = cft->release(inode, file); - if (css) + if (css->ss) css_put(css); return ret; } -- cgit v1.2.3 From 182446d087906de40e514573a92a97b203695f71 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:24 -0400 Subject: cgroup: pass around cgroup_subsys_state instead of cgroup in file methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup. Please see the previous commit which converts the subsystem methods for rationale. This patch converts all cftype file operations to take @css instead of @cgroup. cftypes for the cgroup core files don't have their subsytem pointer set. These will automatically use the dummy_css added by the previous patch and can be converted the same way. Most subsystem conversions are straight forwards but there are some interesting ones. * freezer: update_if_frozen() is also converted to take @css instead of @cgroup for consistency. This will make the code look simpler too once iterators are converted to use css. * memory/vmpressure: mem_cgroup_from_css() needs to be exported to vmpressure while mem_cgroup_from_cont() can be made static. Updated accordingly. * cpu: cgroup_tg() doesn't have any user left. Removed. * cpuacct: cgroup_ca() doesn't have any user left. Removed. * hugetlb: hugetlb_cgroup_form_cgroup() doesn't have any user left. Removed. * net_cls: cgrp_cls_state() doesn't have any user left. Removed. Signed-off-by: Tejun Heo Acked-by: Li Zefan Acked-by: Michal Hocko Acked-by: Vivek Goyal Acked-by: Aristeu Rozanski Acked-by: Daniel Wagner Cc: Peter Zijlstra Cc: Ingo Molnar Cc: Johannes Weiner Cc: Balbir Singh Cc: Matt Helsley Cc: Jens Axboe Cc: Steven Rostedt --- kernel/cgroup.c | 162 +++++++++++++++++++++++++----------------------- kernel/cgroup_freezer.c | 40 ++++++------ kernel/cpuset.c | 35 ++++++----- kernel/sched/core.c | 65 ++++++++++--------- kernel/sched/cpuacct.c | 28 ++++----- 5 files changed, 165 insertions(+), 165 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index c049992f1ffa..6ee469837fda 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -2235,34 +2235,38 @@ int cgroup_attach_task_all(struct task_struct *from, struct task_struct *tsk) } EXPORT_SYMBOL_GPL(cgroup_attach_task_all); -static int cgroup_tasks_write(struct cgroup *cgrp, struct cftype *cft, u64 pid) +static int cgroup_tasks_write(struct cgroup_subsys_state *css, + struct cftype *cft, u64 pid) { - return attach_task_by_pid(cgrp, pid, false); + return attach_task_by_pid(css->cgroup, pid, false); } -static int cgroup_procs_write(struct cgroup *cgrp, struct cftype *cft, u64 tgid) +static int cgroup_procs_write(struct cgroup_subsys_state *css, + struct cftype *cft, u64 tgid) { - return attach_task_by_pid(cgrp, tgid, true); + return attach_task_by_pid(css->cgroup, tgid, true); } -static int cgroup_release_agent_write(struct cgroup *cgrp, struct cftype *cft, - const char *buffer) +static int cgroup_release_agent_write(struct cgroup_subsys_state *css, + struct cftype *cft, const char *buffer) { - BUILD_BUG_ON(sizeof(cgrp->root->release_agent_path) < PATH_MAX); + BUILD_BUG_ON(sizeof(css->cgroup->root->release_agent_path) < PATH_MAX); if (strlen(buffer) >= PATH_MAX) return -EINVAL; - if (!cgroup_lock_live_group(cgrp)) + if (!cgroup_lock_live_group(css->cgroup)) return -ENODEV; mutex_lock(&cgroup_root_mutex); - strcpy(cgrp->root->release_agent_path, buffer); + strcpy(css->cgroup->root->release_agent_path, buffer); mutex_unlock(&cgroup_root_mutex); mutex_unlock(&cgroup_mutex); return 0; } -static int cgroup_release_agent_show(struct cgroup *cgrp, struct cftype *cft, - struct seq_file *seq) +static int cgroup_release_agent_show(struct cgroup_subsys_state *css, + struct cftype *cft, struct seq_file *seq) { + struct cgroup *cgrp = css->cgroup; + if (!cgroup_lock_live_group(cgrp)) return -ENODEV; seq_puts(seq, cgrp->root->release_agent_path); @@ -2271,10 +2275,10 @@ static int cgroup_release_agent_show(struct cgroup *cgrp, struct cftype *cft, return 0; } -static int cgroup_sane_behavior_show(struct cgroup *cgrp, struct cftype *cft, - struct seq_file *seq) +static int cgroup_sane_behavior_show(struct cgroup_subsys_state *css, + struct cftype *cft, struct seq_file *seq) { - seq_printf(seq, "%d\n", cgroup_sane_behavior(cgrp)); + seq_printf(seq, "%d\n", cgroup_sane_behavior(css->cgroup)); return 0; } @@ -2292,10 +2296,10 @@ static struct cgroup_subsys_state *cgroup_file_css(struct cfent *cfe) /* A buffer size big enough for numbers or short strings */ #define CGROUP_LOCAL_BUFFER_SIZE 64 -static ssize_t cgroup_write_X64(struct cgroup *cgrp, struct cftype *cft, - struct file *file, - const char __user *userbuf, - size_t nbytes, loff_t *unused_ppos) +static ssize_t cgroup_write_X64(struct cgroup_subsys_state *css, + struct cftype *cft, struct file *file, + const char __user *userbuf, size_t nbytes, + loff_t *unused_ppos) { char buffer[CGROUP_LOCAL_BUFFER_SIZE]; int retval = 0; @@ -2313,22 +2317,22 @@ static ssize_t cgroup_write_X64(struct cgroup *cgrp, struct cftype *cft, u64 val = simple_strtoull(strstrip(buffer), &end, 0); if (*end) return -EINVAL; - retval = cft->write_u64(cgrp, cft, val); + retval = cft->write_u64(css, cft, val); } else { s64 val = simple_strtoll(strstrip(buffer), &end, 0); if (*end) return -EINVAL; - retval = cft->write_s64(cgrp, cft, val); + retval = cft->write_s64(css, cft, val); } if (!retval) retval = nbytes; return retval; } -static ssize_t cgroup_write_string(struct cgroup *cgrp, struct cftype *cft, - struct file *file, - const char __user *userbuf, - size_t nbytes, loff_t *unused_ppos) +static ssize_t cgroup_write_string(struct cgroup_subsys_state *css, + struct cftype *cft, struct file *file, + const char __user *userbuf, size_t nbytes, + loff_t *unused_ppos) { char local_buffer[CGROUP_LOCAL_BUFFER_SIZE]; int retval = 0; @@ -2351,7 +2355,7 @@ static ssize_t cgroup_write_string(struct cgroup *cgrp, struct cftype *cft, } buffer[nbytes] = 0; /* nul-terminate */ - retval = cft->write_string(cgrp, cft, strstrip(buffer)); + retval = cft->write_string(css, cft, strstrip(buffer)); if (!retval) retval = nbytes; out: @@ -2361,60 +2365,60 @@ out: } static ssize_t cgroup_file_write(struct file *file, const char __user *buf, - size_t nbytes, loff_t *ppos) + size_t nbytes, loff_t *ppos) { + struct cfent *cfe = __d_cfe(file->f_dentry); struct cftype *cft = __d_cft(file->f_dentry); - struct cgroup *cgrp = __d_cgrp(file->f_dentry->d_parent); + struct cgroup_subsys_state *css = cgroup_file_css(cfe); if (cft->write) - return cft->write(cgrp, cft, file, buf, nbytes, ppos); + return cft->write(css, cft, file, buf, nbytes, ppos); if (cft->write_u64 || cft->write_s64) - return cgroup_write_X64(cgrp, cft, file, buf, nbytes, ppos); + return cgroup_write_X64(css, cft, file, buf, nbytes, ppos); if (cft->write_string) - return cgroup_write_string(cgrp, cft, file, buf, nbytes, ppos); + return cgroup_write_string(css, cft, file, buf, nbytes, ppos); if (cft->trigger) { - int ret = cft->trigger(cgrp, (unsigned int)cft->private); + int ret = cft->trigger(css, (unsigned int)cft->private); return ret ? ret : nbytes; } return -EINVAL; } -static ssize_t cgroup_read_u64(struct cgroup *cgrp, struct cftype *cft, - struct file *file, - char __user *buf, size_t nbytes, - loff_t *ppos) +static ssize_t cgroup_read_u64(struct cgroup_subsys_state *css, + struct cftype *cft, struct file *file, + char __user *buf, size_t nbytes, loff_t *ppos) { char tmp[CGROUP_LOCAL_BUFFER_SIZE]; - u64 val = cft->read_u64(cgrp, cft); + u64 val = cft->read_u64(css, cft); int len = sprintf(tmp, "%llu\n", (unsigned long long) val); return simple_read_from_buffer(buf, nbytes, ppos, tmp, len); } -static ssize_t cgroup_read_s64(struct cgroup *cgrp, struct cftype *cft, - struct file *file, - char __user *buf, size_t nbytes, - loff_t *ppos) +static ssize_t cgroup_read_s64(struct cgroup_subsys_state *css, + struct cftype *cft, struct file *file, + char __user *buf, size_t nbytes, loff_t *ppos) { char tmp[CGROUP_LOCAL_BUFFER_SIZE]; - s64 val = cft->read_s64(cgrp, cft); + s64 val = cft->read_s64(css, cft); int len = sprintf(tmp, "%lld\n", (long long) val); return simple_read_from_buffer(buf, nbytes, ppos, tmp, len); } static ssize_t cgroup_file_read(struct file *file, char __user *buf, - size_t nbytes, loff_t *ppos) + size_t nbytes, loff_t *ppos) { + struct cfent *cfe = __d_cfe(file->f_dentry); struct cftype *cft = __d_cft(file->f_dentry); - struct cgroup *cgrp = __d_cgrp(file->f_dentry->d_parent); + struct cgroup_subsys_state *css = cgroup_file_css(cfe); if (cft->read) - return cft->read(cgrp, cft, file, buf, nbytes, ppos); + return cft->read(css, cft, file, buf, nbytes, ppos); if (cft->read_u64) - return cgroup_read_u64(cgrp, cft, file, buf, nbytes, ppos); + return cgroup_read_u64(css, cft, file, buf, nbytes, ppos); if (cft->read_s64) - return cgroup_read_s64(cgrp, cft, file, buf, nbytes, ppos); + return cgroup_read_s64(css, cft, file, buf, nbytes, ppos); return -EINVAL; } @@ -2433,16 +2437,16 @@ static int cgroup_seqfile_show(struct seq_file *m, void *arg) { struct cfent *cfe = m->private; struct cftype *cft = cfe->type; - struct cgroup *cgrp = __d_cgrp(cfe->dentry->d_parent); + struct cgroup_subsys_state *css = cgroup_file_css(cfe); if (cft->read_map) { struct cgroup_map_cb cb = { .fill = cgroup_map_add, .state = m, }; - return cft->read_map(cgrp, cft, &cb); + return cft->read_map(css, cft, &cb); } - return cft->read_seq_string(cgrp, cft, m); + return cft->read_seq_string(css, cft, m); } static const struct file_operations cgroup_seqfile_operations = { @@ -3860,21 +3864,20 @@ static int cgroup_procs_open(struct inode *unused, struct file *file) return cgroup_pidlist_open(file, CGROUP_FILE_PROCS); } -static u64 cgroup_read_notify_on_release(struct cgroup *cgrp, - struct cftype *cft) +static u64 cgroup_read_notify_on_release(struct cgroup_subsys_state *css, + struct cftype *cft) { - return notify_on_release(cgrp); + return notify_on_release(css->cgroup); } -static int cgroup_write_notify_on_release(struct cgroup *cgrp, - struct cftype *cft, - u64 val) +static int cgroup_write_notify_on_release(struct cgroup_subsys_state *css, + struct cftype *cft, u64 val) { - clear_bit(CGRP_RELEASABLE, &cgrp->flags); + clear_bit(CGRP_RELEASABLE, &css->cgroup->flags); if (val) - set_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags); + set_bit(CGRP_NOTIFY_ON_RELEASE, &css->cgroup->flags); else - clear_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags); + clear_bit(CGRP_NOTIFY_ON_RELEASE, &css->cgroup->flags); return 0; } @@ -3972,9 +3975,10 @@ static void cgroup_event_ptable_queue_proc(struct file *file, * Input must be in format ' '. * Interpretation of args is defined by control file implementation. */ -static int cgroup_write_event_control(struct cgroup *cgrp, struct cftype *cft, - const char *buffer) +static int cgroup_write_event_control(struct cgroup_subsys_state *css, + struct cftype *cft, const char *buffer) { + struct cgroup *cgrp = css->cgroup; struct cgroup_event *event; struct cgroup *cgrp_cfile; unsigned int efd, cfd; @@ -4082,20 +4086,19 @@ out_kfree: return ret; } -static u64 cgroup_clone_children_read(struct cgroup *cgrp, - struct cftype *cft) +static u64 cgroup_clone_children_read(struct cgroup_subsys_state *css, + struct cftype *cft) { - return test_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags); + return test_bit(CGRP_CPUSET_CLONE_CHILDREN, &css->cgroup->flags); } -static int cgroup_clone_children_write(struct cgroup *cgrp, - struct cftype *cft, - u64 val) +static int cgroup_clone_children_write(struct cgroup_subsys_state *css, + struct cftype *cft, u64 val) { if (val) - set_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags); + set_bit(CGRP_CPUSET_CLONE_CHILDREN, &css->cgroup->flags); else - clear_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags); + clear_bit(CGRP_CPUSET_CLONE_CHILDREN, &css->cgroup->flags); return 0; } @@ -5585,17 +5588,19 @@ static void debug_css_free(struct cgroup_subsys_state *css) kfree(css); } -static u64 debug_taskcount_read(struct cgroup *cgrp, struct cftype *cft) +static u64 debug_taskcount_read(struct cgroup_subsys_state *css, + struct cftype *cft) { - return cgroup_task_count(cgrp); + return cgroup_task_count(css->cgroup); } -static u64 current_css_set_read(struct cgroup *cgrp, struct cftype *cft) +static u64 current_css_set_read(struct cgroup_subsys_state *css, + struct cftype *cft) { return (u64)(unsigned long)current->cgroups; } -static u64 current_css_set_refcount_read(struct cgroup *cgrp, +static u64 current_css_set_refcount_read(struct cgroup_subsys_state *css, struct cftype *cft) { u64 count; @@ -5606,7 +5611,7 @@ static u64 current_css_set_refcount_read(struct cgroup *cgrp, return count; } -static int current_css_set_cg_links_read(struct cgroup *cgrp, +static int current_css_set_cg_links_read(struct cgroup_subsys_state *css, struct cftype *cft, struct seq_file *seq) { @@ -5633,14 +5638,13 @@ static int current_css_set_cg_links_read(struct cgroup *cgrp, } #define MAX_TASKS_SHOWN_PER_CSS 25 -static int cgroup_css_links_read(struct cgroup *cgrp, - struct cftype *cft, - struct seq_file *seq) +static int cgroup_css_links_read(struct cgroup_subsys_state *css, + struct cftype *cft, struct seq_file *seq) { struct cgrp_cset_link *link; read_lock(&css_set_lock); - list_for_each_entry(link, &cgrp->cset_links, cset_link) { + list_for_each_entry(link, &css->cgroup->cset_links, cset_link) { struct css_set *cset = link->cset; struct task_struct *task; int count = 0; @@ -5659,9 +5663,9 @@ static int cgroup_css_links_read(struct cgroup *cgrp, return 0; } -static u64 releasable_read(struct cgroup *cgrp, struct cftype *cft) +static u64 releasable_read(struct cgroup_subsys_state *css, struct cftype *cft) { - return test_bit(CGRP_RELEASABLE, &cgrp->flags); + return test_bit(CGRP_RELEASABLE, &css->cgroup->flags); } static struct cftype debug_files[] = { diff --git a/kernel/cgroup_freezer.c b/kernel/cgroup_freezer.c index f03a85719c3c..19613ba51444 100644 --- a/kernel/cgroup_freezer.c +++ b/kernel/cgroup_freezer.c @@ -245,7 +245,7 @@ out: /** * update_if_frozen - update whether a cgroup finished freezing - * @cgroup: cgroup of interest + * @css: css of interest * * Once FREEZING is initiated, transition to FROZEN is lazily updated by * calling this function. If the current state is FREEZING but not FROZEN, @@ -256,12 +256,12 @@ out: * update_if_frozen() on all descendants prior to invoking this function. * * Task states and freezer state might disagree while tasks are being - * migrated into or out of @cgroup, so we can't verify task states against + * migrated into or out of @css, so we can't verify task states against * @freezer state here. See freezer_attach() for details. */ -static void update_if_frozen(struct cgroup *cgroup) +static void update_if_frozen(struct cgroup_subsys_state *css) { - struct freezer *freezer = cgroup_freezer(cgroup); + struct freezer *freezer = css_freezer(css); struct cgroup *pos; struct cgroup_iter it; struct task_struct *task; @@ -275,7 +275,7 @@ static void update_if_frozen(struct cgroup *cgroup) goto out_unlock; /* are all (live) children frozen? */ - cgroup_for_each_child(pos, cgroup) { + cgroup_for_each_child(pos, css->cgroup) { struct freezer *child = cgroup_freezer(pos); if ((child->state & CGROUP_FREEZER_ONLINE) && @@ -284,9 +284,9 @@ static void update_if_frozen(struct cgroup *cgroup) } /* are all tasks frozen? */ - cgroup_iter_start(cgroup, &it); + cgroup_iter_start(css->cgroup, &it); - while ((task = cgroup_iter_next(cgroup, &it))) { + while ((task = cgroup_iter_next(css->cgroup, &it))) { if (freezing(task)) { /* * freezer_should_skip() indicates that the task @@ -301,12 +301,12 @@ static void update_if_frozen(struct cgroup *cgroup) freezer->state |= CGROUP_FROZEN; out_iter_end: - cgroup_iter_end(cgroup, &it); + cgroup_iter_end(css->cgroup, &it); out_unlock: spin_unlock_irq(&freezer->lock); } -static int freezer_read(struct cgroup *cgroup, struct cftype *cft, +static int freezer_read(struct cgroup_subsys_state *css, struct cftype *cft, struct seq_file *m) { struct cgroup *pos; @@ -314,13 +314,13 @@ static int freezer_read(struct cgroup *cgroup, struct cftype *cft, rcu_read_lock(); /* update states bottom-up */ - cgroup_for_each_descendant_post(pos, cgroup) - update_if_frozen(pos); - update_if_frozen(cgroup); + cgroup_for_each_descendant_post(pos, css->cgroup) + update_if_frozen(cgroup_css(pos, freezer_subsys_id)); + update_if_frozen(css); rcu_read_unlock(); - seq_puts(m, freezer_state_strs(cgroup_freezer(cgroup)->state)); + seq_puts(m, freezer_state_strs(css_freezer(css)->state)); seq_putc(m, '\n'); return 0; } @@ -426,7 +426,7 @@ static void freezer_change_state(struct freezer *freezer, bool freeze) rcu_read_unlock(); } -static int freezer_write(struct cgroup *cgroup, struct cftype *cft, +static int freezer_write(struct cgroup_subsys_state *css, struct cftype *cft, const char *buffer) { bool freeze; @@ -438,20 +438,22 @@ static int freezer_write(struct cgroup *cgroup, struct cftype *cft, else return -EINVAL; - freezer_change_state(cgroup_freezer(cgroup), freeze); + freezer_change_state(css_freezer(css), freeze); return 0; } -static u64 freezer_self_freezing_read(struct cgroup *cgroup, struct cftype *cft) +static u64 freezer_self_freezing_read(struct cgroup_subsys_state *css, + struct cftype *cft) { - struct freezer *freezer = cgroup_freezer(cgroup); + struct freezer *freezer = css_freezer(css); return (bool)(freezer->state & CGROUP_FREEZING_SELF); } -static u64 freezer_parent_freezing_read(struct cgroup *cgroup, struct cftype *cft) +static u64 freezer_parent_freezing_read(struct cgroup_subsys_state *css, + struct cftype *cft) { - struct freezer *freezer = cgroup_freezer(cgroup); + struct freezer *freezer = css_freezer(css); return (bool)(freezer->state & CGROUP_FREEZING_PARENT); } diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 8ce3fdc3dfcc..89b76e1d3aa1 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -1603,9 +1603,10 @@ typedef enum { FILE_SPREAD_SLAB, } cpuset_filetype_t; -static int cpuset_write_u64(struct cgroup *cgrp, struct cftype *cft, u64 val) +static int cpuset_write_u64(struct cgroup_subsys_state *css, struct cftype *cft, + u64 val) { - struct cpuset *cs = cgroup_cs(cgrp); + struct cpuset *cs = css_cs(css); cpuset_filetype_t type = cft->private; int retval = -ENODEV; @@ -1650,9 +1651,10 @@ out_unlock: return retval; } -static int cpuset_write_s64(struct cgroup *cgrp, struct cftype *cft, s64 val) +static int cpuset_write_s64(struct cgroup_subsys_state *css, struct cftype *cft, + s64 val) { - struct cpuset *cs = cgroup_cs(cgrp); + struct cpuset *cs = css_cs(css); cpuset_filetype_t type = cft->private; int retval = -ENODEV; @@ -1676,10 +1678,10 @@ out_unlock: /* * Common handling for a write to a "cpus" or "mems" file. */ -static int cpuset_write_resmask(struct cgroup *cgrp, struct cftype *cft, - const char *buf) +static int cpuset_write_resmask(struct cgroup_subsys_state *css, + struct cftype *cft, const char *buf) { - struct cpuset *cs = cgroup_cs(cgrp); + struct cpuset *cs = css_cs(css); struct cpuset *trialcs; int retval = -ENODEV; @@ -1758,13 +1760,12 @@ static size_t cpuset_sprintf_memlist(char *page, struct cpuset *cs) return count; } -static ssize_t cpuset_common_file_read(struct cgroup *cgrp, - struct cftype *cft, - struct file *file, - char __user *buf, - size_t nbytes, loff_t *ppos) +static ssize_t cpuset_common_file_read(struct cgroup_subsys_state *css, + struct cftype *cft, struct file *file, + char __user *buf, size_t nbytes, + loff_t *ppos) { - struct cpuset *cs = cgroup_cs(cgrp); + struct cpuset *cs = css_cs(css); cpuset_filetype_t type = cft->private; char *page; ssize_t retval = 0; @@ -1794,9 +1795,9 @@ out: return retval; } -static u64 cpuset_read_u64(struct cgroup *cgrp, struct cftype *cft) +static u64 cpuset_read_u64(struct cgroup_subsys_state *css, struct cftype *cft) { - struct cpuset *cs = cgroup_cs(cgrp); + struct cpuset *cs = css_cs(css); cpuset_filetype_t type = cft->private; switch (type) { case FILE_CPU_EXCLUSIVE: @@ -1825,9 +1826,9 @@ static u64 cpuset_read_u64(struct cgroup *cgrp, struct cftype *cft) return 0; } -static s64 cpuset_read_s64(struct cgroup *cgrp, struct cftype *cft) +static s64 cpuset_read_s64(struct cgroup_subsys_state *css, struct cftype *cft) { - struct cpuset *cs = cgroup_cs(cgrp); + struct cpuset *cs = css_cs(css); cpuset_filetype_t type = cft->private; switch (type) { case FILE_SCHED_RELAX_DOMAIN_LEVEL: diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 622b7efc5ade..cc9a49266382 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -7088,12 +7088,6 @@ static inline struct task_group *css_tg(struct cgroup_subsys_state *css) return css ? container_of(css, struct task_group, css) : NULL; } -/* return corresponding task_group object of a cgroup */ -static inline struct task_group *cgroup_tg(struct cgroup *cgrp) -{ - return css_tg(cgroup_css(cgrp, cpu_cgroup_subsys_id)); -} - static struct cgroup_subsys_state * cpu_cgroup_css_alloc(struct cgroup_subsys_state *parent_css) { @@ -7179,15 +7173,16 @@ static void cpu_cgroup_exit(struct cgroup_subsys_state *css, } #ifdef CONFIG_FAIR_GROUP_SCHED -static int cpu_shares_write_u64(struct cgroup *cgrp, struct cftype *cftype, - u64 shareval) +static int cpu_shares_write_u64(struct cgroup_subsys_state *css, + struct cftype *cftype, u64 shareval) { - return sched_group_set_shares(cgroup_tg(cgrp), scale_load(shareval)); + return sched_group_set_shares(css_tg(css), scale_load(shareval)); } -static u64 cpu_shares_read_u64(struct cgroup *cgrp, struct cftype *cft) +static u64 cpu_shares_read_u64(struct cgroup_subsys_state *css, + struct cftype *cft) { - struct task_group *tg = cgroup_tg(cgrp); + struct task_group *tg = css_tg(css); return (u64) scale_load_down(tg->shares); } @@ -7309,26 +7304,28 @@ long tg_get_cfs_period(struct task_group *tg) return cfs_period_us; } -static s64 cpu_cfs_quota_read_s64(struct cgroup *cgrp, struct cftype *cft) +static s64 cpu_cfs_quota_read_s64(struct cgroup_subsys_state *css, + struct cftype *cft) { - return tg_get_cfs_quota(cgroup_tg(cgrp)); + return tg_get_cfs_quota(css_tg(css)); } -static int cpu_cfs_quota_write_s64(struct cgroup *cgrp, struct cftype *cftype, - s64 cfs_quota_us) +static int cpu_cfs_quota_write_s64(struct cgroup_subsys_state *css, + struct cftype *cftype, s64 cfs_quota_us) { - return tg_set_cfs_quota(cgroup_tg(cgrp), cfs_quota_us); + return tg_set_cfs_quota(css_tg(css), cfs_quota_us); } -static u64 cpu_cfs_period_read_u64(struct cgroup *cgrp, struct cftype *cft) +static u64 cpu_cfs_period_read_u64(struct cgroup_subsys_state *css, + struct cftype *cft) { - return tg_get_cfs_period(cgroup_tg(cgrp)); + return tg_get_cfs_period(css_tg(css)); } -static int cpu_cfs_period_write_u64(struct cgroup *cgrp, struct cftype *cftype, - u64 cfs_period_us) +static int cpu_cfs_period_write_u64(struct cgroup_subsys_state *css, + struct cftype *cftype, u64 cfs_period_us) { - return tg_set_cfs_period(cgroup_tg(cgrp), cfs_period_us); + return tg_set_cfs_period(css_tg(css), cfs_period_us); } struct cfs_schedulable_data { @@ -7409,10 +7406,10 @@ static int __cfs_schedulable(struct task_group *tg, u64 period, u64 quota) return ret; } -static int cpu_stats_show(struct cgroup *cgrp, struct cftype *cft, +static int cpu_stats_show(struct cgroup_subsys_state *css, struct cftype *cft, struct cgroup_map_cb *cb) { - struct task_group *tg = cgroup_tg(cgrp); + struct task_group *tg = css_tg(css); struct cfs_bandwidth *cfs_b = &tg->cfs_bandwidth; cb->fill(cb, "nr_periods", cfs_b->nr_periods); @@ -7425,26 +7422,28 @@ static int cpu_stats_show(struct cgroup *cgrp, struct cftype *cft, #endif /* CONFIG_FAIR_GROUP_SCHED */ #ifdef CONFIG_RT_GROUP_SCHED -static int cpu_rt_runtime_write(struct cgroup *cgrp, struct cftype *cft, - s64 val) +static int cpu_rt_runtime_write(struct cgroup_subsys_state *css, + struct cftype *cft, s64 val) { - return sched_group_set_rt_runtime(cgroup_tg(cgrp), val); + return sched_group_set_rt_runtime(css_tg(css), val); } -static s64 cpu_rt_runtime_read(struct cgroup *cgrp, struct cftype *cft) +static s64 cpu_rt_runtime_read(struct cgroup_subsys_state *css, + struct cftype *cft) { - return sched_group_rt_runtime(cgroup_tg(cgrp)); + return sched_group_rt_runtime(css_tg(css)); } -static int cpu_rt_period_write_uint(struct cgroup *cgrp, struct cftype *cftype, - u64 rt_period_us) +static int cpu_rt_period_write_uint(struct cgroup_subsys_state *css, + struct cftype *cftype, u64 rt_period_us) { - return sched_group_set_rt_period(cgroup_tg(cgrp), rt_period_us); + return sched_group_set_rt_period(css_tg(css), rt_period_us); } -static u64 cpu_rt_period_read_uint(struct cgroup *cgrp, struct cftype *cft) +static u64 cpu_rt_period_read_uint(struct cgroup_subsys_state *css, + struct cftype *cft) { - return sched_group_rt_period(cgroup_tg(cgrp)); + return sched_group_rt_period(css_tg(css)); } #endif /* CONFIG_RT_GROUP_SCHED */ diff --git a/kernel/sched/cpuacct.c b/kernel/sched/cpuacct.c index 1b784d9b3630..f64722ff0299 100644 --- a/kernel/sched/cpuacct.c +++ b/kernel/sched/cpuacct.c @@ -38,12 +38,6 @@ static inline struct cpuacct *css_ca(struct cgroup_subsys_state *css) return css ? container_of(css, struct cpuacct, css) : NULL; } -/* return cpu accounting group corresponding to this container */ -static inline struct cpuacct *cgroup_ca(struct cgroup *cgrp) -{ - return css_ca(cgroup_css(cgrp, cpuacct_subsys_id)); -} - /* return cpu accounting group to which this task belongs */ static inline struct cpuacct *task_ca(struct task_struct *tsk) { @@ -138,9 +132,9 @@ static void cpuacct_cpuusage_write(struct cpuacct *ca, int cpu, u64 val) } /* return total cpu usage (in nanoseconds) of a group */ -static u64 cpuusage_read(struct cgroup *cgrp, struct cftype *cft) +static u64 cpuusage_read(struct cgroup_subsys_state *css, struct cftype *cft) { - struct cpuacct *ca = cgroup_ca(cgrp); + struct cpuacct *ca = css_ca(css); u64 totalcpuusage = 0; int i; @@ -150,10 +144,10 @@ static u64 cpuusage_read(struct cgroup *cgrp, struct cftype *cft) return totalcpuusage; } -static int cpuusage_write(struct cgroup *cgrp, struct cftype *cftype, - u64 reset) +static int cpuusage_write(struct cgroup_subsys_state *css, struct cftype *cft, + u64 reset) { - struct cpuacct *ca = cgroup_ca(cgrp); + struct cpuacct *ca = css_ca(css); int err = 0; int i; @@ -169,10 +163,10 @@ out: return err; } -static int cpuacct_percpu_seq_read(struct cgroup *cgroup, struct cftype *cft, - struct seq_file *m) +static int cpuacct_percpu_seq_read(struct cgroup_subsys_state *css, + struct cftype *cft, struct seq_file *m) { - struct cpuacct *ca = cgroup_ca(cgroup); + struct cpuacct *ca = css_ca(css); u64 percpu; int i; @@ -189,10 +183,10 @@ static const char * const cpuacct_stat_desc[] = { [CPUACCT_STAT_SYSTEM] = "system", }; -static int cpuacct_stats_show(struct cgroup *cgrp, struct cftype *cft, - struct cgroup_map_cb *cb) +static int cpuacct_stats_show(struct cgroup_subsys_state *css, + struct cftype *cft, struct cgroup_map_cb *cb) { - struct cpuacct *ca = cgroup_ca(cgrp); + struct cpuacct *ca = css_ca(css); int cpu; s64 val = 0; -- cgit v1.2.3 From 3b287a505ef4024634beb12a93773254909d5dae Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:24 -0400 Subject: cgroup: convert cgroup_next_sibling() to cgroup_next_child() cgroup is transitioning to using css (cgroup_subsys_state) as the main subsys interface handle instead of cgroup and the iterators will be updated to use css too. The iterators need to walk the cgroup hierarchy and return the css's matching the origin css, which is a bit cumbersome to open code. This patch converts cgroup_next_sibling() to cgroup_next_child() so that it can handle all steps of direct child iteration. This will be used to update iterators to take @css instead of @cgrp. In addition to the new iteration init handling, cgroup_next_child() is restructured so that the different branches share the end of iteration condition check. This patch doesn't change any behavior. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 59 +++++++++++++++++++++++++++++---------------------------- 1 file changed, 30 insertions(+), 29 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 6ee469837fda..dd55244952bd 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -3037,15 +3037,16 @@ static void cgroup_enable_task_cg_lists(void) } /** - * cgroup_next_sibling - find the next sibling of a given cgroup - * @pos: the current cgroup + * cgroup_next_child - find the next child of a given cgroup + * @pos: the current position (%NULL to initiate traversal) + * @cgrp: cgroup whose descendants to walk * - * This function returns the next sibling of @pos and should be called - * under RCU read lock. The only requirement is that @pos is accessible. - * The next sibling is guaranteed to be returned regardless of @pos's - * state. + * This function returns the next child of @cgrp and should be called under + * RCU read lock. The only requirement is that @cgrp and @pos are + * accessible. The next sibling is guaranteed to be returned regardless of + * their states. */ -struct cgroup *cgroup_next_sibling(struct cgroup *pos) +struct cgroup *cgroup_next_child(struct cgroup *pos, struct cgroup *cgrp) { struct cgroup *next; @@ -3061,30 +3062,30 @@ struct cgroup *cgroup_next_sibling(struct cgroup *pos) * safe to dereference from this RCU critical section. If * ->sibling.next is inaccessible, cgroup_is_dead() is guaranteed * to be visible as %true here. + * + * If @pos is dead, its next pointer can't be dereferenced; + * however, as each cgroup is given a monotonically increasing + * unique serial number and always appended to the sibling list, + * the next one can be found by walking the parent's children until + * we see a cgroup with higher serial number than @pos's. While + * this path can be slower, it's taken only when either the current + * cgroup is removed or iteration and removal race. */ - if (likely(!cgroup_is_dead(pos))) { + if (!pos) { + next = list_entry_rcu(cgrp->children.next, struct cgroup, sibling); + } else if (likely(!cgroup_is_dead(pos))) { next = list_entry_rcu(pos->sibling.next, struct cgroup, sibling); - if (&next->sibling != &pos->parent->children) - return next; - return NULL; + } else { + list_for_each_entry_rcu(next, &cgrp->children, sibling) + if (next->serial_nr > pos->serial_nr) + break; } - /* - * Can't dereference the next pointer. Each cgroup is given a - * monotonically increasing unique serial number and always - * appended to the sibling list, so the next one can be found by - * walking the parent's children until we see a cgroup with higher - * serial number than @pos's. - * - * While this path can be slow, it's taken only when either the - * current cgroup is removed or iteration and removal race. - */ - list_for_each_entry_rcu(next, &pos->parent->children, sibling) - if (next->serial_nr > pos->serial_nr) - return next; + if (&next->sibling != &cgrp->children) + return next; return NULL; } -EXPORT_SYMBOL_GPL(cgroup_next_sibling); +EXPORT_SYMBOL_GPL(cgroup_next_child); /** * cgroup_next_descendant_pre - find the next descendant for pre-order walk @@ -3117,7 +3118,7 @@ struct cgroup *cgroup_next_descendant_pre(struct cgroup *pos, /* no child, visit my or the closest ancestor's next sibling */ while (pos != cgroup) { - next = cgroup_next_sibling(pos); + next = cgroup_next_child(pos, pos->parent); if (next) return next; pos = pos->parent; @@ -3198,7 +3199,7 @@ struct cgroup *cgroup_next_descendant_post(struct cgroup *pos, } /* if there's an unvisited sibling, visit its leftmost descendant */ - next = cgroup_next_sibling(pos); + next = cgroup_next_child(pos, pos->parent); if (next) return cgroup_leftmost_descendant(next); @@ -4549,9 +4550,9 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) /* * Mark @cgrp dead. This prevents further task migration and child * creation by disabling cgroup_lock_live_group(). Note that - * CGRP_DEAD assertion is depended upon by cgroup_next_sibling() to + * CGRP_DEAD assertion is depended upon by cgroup_next_child() to * resume iteration after dropping RCU read lock. See - * cgroup_next_sibling() for details. + * cgroup_next_child() for details. */ set_bit(CGRP_DEAD, &cgrp->flags); -- cgit v1.2.3 From f48e3924dca268c677c4e338e5d91ad9e6fe6b9e Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:24 -0400 Subject: cgroup: always use cgroup_next_child() to walk the children list There are several places where the children list is accessed directly. This patch converts those places to use cgroup_next_child(). This will help updating the hierarchy iterators to use @css instead of @cgrp. While cgroup_next_child() can be heavy in pathological cases - e.g. a lot of dead children, this shouldn't cause any noticeable behavior differences. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index dd55244952bd..2b7354faaca7 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -3112,7 +3112,7 @@ struct cgroup *cgroup_next_descendant_pre(struct cgroup *pos, pos = cgroup; /* visit the first child if exists */ - next = list_first_or_null_rcu(&pos->children, struct cgroup, sibling); + next = cgroup_next_child(NULL, pos); if (next) return next; @@ -3151,7 +3151,7 @@ struct cgroup *cgroup_rightmost_descendant(struct cgroup *pos) last = pos; /* ->prev isn't RCU safe, walk ->next till the end */ pos = NULL; - list_for_each_entry_rcu(tmp, &last->children, sibling) + cgroup_for_each_child(tmp, last) pos = tmp; } while (pos); @@ -3165,8 +3165,7 @@ static struct cgroup *cgroup_leftmost_descendant(struct cgroup *pos) do { last = pos; - pos = list_first_or_null_rcu(&pos->children, struct cgroup, - sibling); + pos = cgroup_next_child(NULL, pos); } while (pos); return last; -- cgit v1.2.3 From 492eb21b98f88e411a8bb43d6edcd7d7022add10 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:25 -0400 Subject: cgroup: make hierarchy iterators deal with cgroup_subsys_state instead of cgroup cgroup is currently in the process of transitioning to using css (cgroup_subsys_state) as the primary handle instead of cgroup in subsystem API. For hierarchy iterators, this is beneficial because * In most cases, css is the only thing subsystems care about anyway. * On the planned unified hierarchy, iterations for different subsystems will need to skip over different subtrees of the hierarchy depending on which subsystems are enabled on each cgroup. Passing around css makes it unnecessary to explicitly specify the subsystem in question as css is intersection between cgroup and subsystem * For the planned unified hierarchy, css's would need to be created and destroyed dynamically independent from cgroup hierarchy. Having cgroup core manage css iteration makes enforcing deref rules a lot easier. Most subsystem conversions are straight-forward. Noteworthy changes are * blkio: cgroup_to_blkcg() is no longer used. Removed. * freezer: cgroup_freezer() is no longer used. Removed. * devices: cgroup_to_devcgroup() is no longer used. Removed. Signed-off-by: Tejun Heo Acked-by: Li Zefan Acked-by: Michal Hocko Acked-by: Vivek Goyal Acked-by: Aristeu Rozanski Cc: Johannes Weiner Cc: Balbir Singh Cc: Matt Helsley Cc: Jens Axboe --- kernel/cgroup.c | 131 +++++++++++++++++++++++++++--------------------- kernel/cgroup_freezer.c | 25 ++++----- kernel/cpuset.c | 58 ++++++++++----------- 3 files changed, 112 insertions(+), 102 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 2b7354faaca7..91eac33fac86 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -2814,8 +2814,8 @@ static void cgroup_cfts_prepare(void) /* * Thanks to the entanglement with vfs inode locking, we can't walk * the existing cgroups under cgroup_mutex and create files. - * Instead, we use cgroup_for_each_descendant_pre() and drop RCU - * read lock before calling cgroup_addrm_files(). + * Instead, we use css_for_each_descendant_pre() and drop RCU read + * lock before calling cgroup_addrm_files(). */ mutex_lock(&cgroup_mutex); } @@ -2825,10 +2825,11 @@ static int cgroup_cfts_commit(struct cftype *cfts, bool is_add) { LIST_HEAD(pending); struct cgroup_subsys *ss = cfts[0].ss; - struct cgroup *cgrp, *root = &ss->root->top_cgroup; + struct cgroup *root = &ss->root->top_cgroup; struct super_block *sb = ss->root->sb; struct dentry *prev = NULL; struct inode *inode; + struct cgroup_subsys_state *css; u64 update_before; int ret = 0; @@ -2861,7 +2862,9 @@ static int cgroup_cfts_commit(struct cftype *cfts, bool is_add) /* add/rm files for all cgroups created before */ rcu_read_lock(); - cgroup_for_each_descendant_pre(cgrp, root) { + css_for_each_descendant_pre(css, cgroup_css(root, ss->subsys_id)) { + struct cgroup *cgrp = css->cgroup; + if (cgroup_is_dead(cgrp)) continue; @@ -3037,17 +3040,21 @@ static void cgroup_enable_task_cg_lists(void) } /** - * cgroup_next_child - find the next child of a given cgroup - * @pos: the current position (%NULL to initiate traversal) - * @cgrp: cgroup whose descendants to walk + * css_next_child - find the next child of a given css + * @pos_css: the current position (%NULL to initiate traversal) + * @parent_css: css whose children to walk * - * This function returns the next child of @cgrp and should be called under - * RCU read lock. The only requirement is that @cgrp and @pos are - * accessible. The next sibling is guaranteed to be returned regardless of - * their states. + * This function returns the next child of @parent_css and should be called + * under RCU read lock. The only requirement is that @parent_css and + * @pos_css are accessible. The next sibling is guaranteed to be returned + * regardless of their states. */ -struct cgroup *cgroup_next_child(struct cgroup *pos, struct cgroup *cgrp) +struct cgroup_subsys_state * +css_next_child(struct cgroup_subsys_state *pos_css, + struct cgroup_subsys_state *parent_css) { + struct cgroup *pos = pos_css ? pos_css->cgroup : NULL; + struct cgroup *cgrp = parent_css->cgroup; struct cgroup *next; WARN_ON_ONCE(!rcu_read_lock_held()); @@ -3081,59 +3088,64 @@ struct cgroup *cgroup_next_child(struct cgroup *pos, struct cgroup *cgrp) break; } - if (&next->sibling != &cgrp->children) - return next; - return NULL; + if (&next->sibling == &cgrp->children) + return NULL; + + if (parent_css->ss) + return cgroup_css(next, parent_css->ss->subsys_id); + else + return &next->dummy_css; } -EXPORT_SYMBOL_GPL(cgroup_next_child); +EXPORT_SYMBOL_GPL(css_next_child); /** - * cgroup_next_descendant_pre - find the next descendant for pre-order walk + * css_next_descendant_pre - find the next descendant for pre-order walk * @pos: the current position (%NULL to initiate traversal) - * @cgroup: cgroup whose descendants to walk + * @root: css whose descendants to walk * - * To be used by cgroup_for_each_descendant_pre(). Find the next - * descendant to visit for pre-order traversal of @cgroup's descendants. + * To be used by css_for_each_descendant_pre(). Find the next descendant + * to visit for pre-order traversal of @root's descendants. * * While this function requires RCU read locking, it doesn't require the * whole traversal to be contained in a single RCU critical section. This * function will return the correct next descendant as long as both @pos - * and @cgroup are accessible and @pos is a descendant of @cgroup. + * and @root are accessible and @pos is a descendant of @root. */ -struct cgroup *cgroup_next_descendant_pre(struct cgroup *pos, - struct cgroup *cgroup) +struct cgroup_subsys_state * +css_next_descendant_pre(struct cgroup_subsys_state *pos, + struct cgroup_subsys_state *root) { - struct cgroup *next; + struct cgroup_subsys_state *next; WARN_ON_ONCE(!rcu_read_lock_held()); - /* if first iteration, pretend we just visited @cgroup */ + /* if first iteration, pretend we just visited @root */ if (!pos) - pos = cgroup; + pos = root; /* visit the first child if exists */ - next = cgroup_next_child(NULL, pos); + next = css_next_child(NULL, pos); if (next) return next; /* no child, visit my or the closest ancestor's next sibling */ - while (pos != cgroup) { - next = cgroup_next_child(pos, pos->parent); + while (pos != root) { + next = css_next_child(pos, css_parent(pos)); if (next) return next; - pos = pos->parent; + pos = css_parent(pos); } return NULL; } -EXPORT_SYMBOL_GPL(cgroup_next_descendant_pre); +EXPORT_SYMBOL_GPL(css_next_descendant_pre); /** - * cgroup_rightmost_descendant - return the rightmost descendant of a cgroup - * @pos: cgroup of interest + * css_rightmost_descendant - return the rightmost descendant of a css + * @pos: css of interest * - * Return the rightmost descendant of @pos. If there's no descendant, - * @pos is returned. This can be used during pre-order traversal to skip + * Return the rightmost descendant of @pos. If there's no descendant, @pos + * is returned. This can be used during pre-order traversal to skip * subtree of @pos. * * While this function requires RCU read locking, it doesn't require the @@ -3141,9 +3153,10 @@ EXPORT_SYMBOL_GPL(cgroup_next_descendant_pre); * function will return the correct rightmost descendant as long as @pos is * accessible. */ -struct cgroup *cgroup_rightmost_descendant(struct cgroup *pos) +struct cgroup_subsys_state * +css_rightmost_descendant(struct cgroup_subsys_state *pos) { - struct cgroup *last, *tmp; + struct cgroup_subsys_state *last, *tmp; WARN_ON_ONCE(!rcu_read_lock_held()); @@ -3151,62 +3164,64 @@ struct cgroup *cgroup_rightmost_descendant(struct cgroup *pos) last = pos; /* ->prev isn't RCU safe, walk ->next till the end */ pos = NULL; - cgroup_for_each_child(tmp, last) + css_for_each_child(tmp, last) pos = tmp; } while (pos); return last; } -EXPORT_SYMBOL_GPL(cgroup_rightmost_descendant); +EXPORT_SYMBOL_GPL(css_rightmost_descendant); -static struct cgroup *cgroup_leftmost_descendant(struct cgroup *pos) +static struct cgroup_subsys_state * +css_leftmost_descendant(struct cgroup_subsys_state *pos) { - struct cgroup *last; + struct cgroup_subsys_state *last; do { last = pos; - pos = cgroup_next_child(NULL, pos); + pos = css_next_child(NULL, pos); } while (pos); return last; } /** - * cgroup_next_descendant_post - find the next descendant for post-order walk + * css_next_descendant_post - find the next descendant for post-order walk * @pos: the current position (%NULL to initiate traversal) - * @cgroup: cgroup whose descendants to walk + * @root: css whose descendants to walk * - * To be used by cgroup_for_each_descendant_post(). Find the next - * descendant to visit for post-order traversal of @cgroup's descendants. + * To be used by css_for_each_descendant_post(). Find the next descendant + * to visit for post-order traversal of @root's descendants. * * While this function requires RCU read locking, it doesn't require the * whole traversal to be contained in a single RCU critical section. This * function will return the correct next descendant as long as both @pos * and @cgroup are accessible and @pos is a descendant of @cgroup. */ -struct cgroup *cgroup_next_descendant_post(struct cgroup *pos, - struct cgroup *cgroup) +struct cgroup_subsys_state * +css_next_descendant_post(struct cgroup_subsys_state *pos, + struct cgroup_subsys_state *root) { - struct cgroup *next; + struct cgroup_subsys_state *next; WARN_ON_ONCE(!rcu_read_lock_held()); /* if first iteration, visit the leftmost descendant */ if (!pos) { - next = cgroup_leftmost_descendant(cgroup); - return next != cgroup ? next : NULL; + next = css_leftmost_descendant(root); + return next != root ? next : NULL; } /* if there's an unvisited sibling, visit its leftmost descendant */ - next = cgroup_next_child(pos, pos->parent); + next = css_next_child(pos, css_parent(pos)); if (next) - return cgroup_leftmost_descendant(next); + return css_leftmost_descendant(next); /* no sibling left, visit parent */ - next = pos->parent; - return next != cgroup ? next : NULL; + next = css_parent(pos); + return next != root ? next : NULL; } -EXPORT_SYMBOL_GPL(cgroup_next_descendant_post); +EXPORT_SYMBOL_GPL(css_next_descendant_post); void cgroup_iter_start(struct cgroup *cgrp, struct cgroup_iter *it) __acquires(css_set_lock) @@ -4549,9 +4564,9 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) /* * Mark @cgrp dead. This prevents further task migration and child * creation by disabling cgroup_lock_live_group(). Note that - * CGRP_DEAD assertion is depended upon by cgroup_next_child() to + * CGRP_DEAD assertion is depended upon by css_next_child() to * resume iteration after dropping RCU read lock. See - * cgroup_next_child() for details. + * css_next_child() for details. */ set_bit(CGRP_DEAD, &cgrp->flags); diff --git a/kernel/cgroup_freezer.c b/kernel/cgroup_freezer.c index 19613ba51444..98ca48d9ceb4 100644 --- a/kernel/cgroup_freezer.c +++ b/kernel/cgroup_freezer.c @@ -50,11 +50,6 @@ static inline struct freezer *css_freezer(struct cgroup_subsys_state *css) return css ? container_of(css, struct freezer, css) : NULL; } -static inline struct freezer *cgroup_freezer(struct cgroup *cgroup) -{ - return css_freezer(cgroup_css(cgroup, freezer_subsys_id)); -} - static inline struct freezer *task_freezer(struct task_struct *task) { return css_freezer(task_css(task, freezer_subsys_id)); @@ -120,7 +115,7 @@ static int freezer_css_online(struct cgroup_subsys_state *css) /* * The following double locking and freezing state inheritance * guarantee that @cgroup can never escape ancestors' freezing - * states. See cgroup_for_each_descendant_pre() for details. + * states. See css_for_each_descendant_pre() for details. */ if (parent) spin_lock_irq(&parent->lock); @@ -262,7 +257,7 @@ out: static void update_if_frozen(struct cgroup_subsys_state *css) { struct freezer *freezer = css_freezer(css); - struct cgroup *pos; + struct cgroup_subsys_state *pos; struct cgroup_iter it; struct task_struct *task; @@ -275,8 +270,8 @@ static void update_if_frozen(struct cgroup_subsys_state *css) goto out_unlock; /* are all (live) children frozen? */ - cgroup_for_each_child(pos, css->cgroup) { - struct freezer *child = cgroup_freezer(pos); + css_for_each_child(pos, css) { + struct freezer *child = css_freezer(pos); if ((child->state & CGROUP_FREEZER_ONLINE) && !(child->state & CGROUP_FROZEN)) @@ -309,13 +304,13 @@ out_unlock: static int freezer_read(struct cgroup_subsys_state *css, struct cftype *cft, struct seq_file *m) { - struct cgroup *pos; + struct cgroup_subsys_state *pos; rcu_read_lock(); /* update states bottom-up */ - cgroup_for_each_descendant_post(pos, css->cgroup) - update_if_frozen(cgroup_css(pos, freezer_subsys_id)); + css_for_each_descendant_post(pos, css) + update_if_frozen(pos); update_if_frozen(css); rcu_read_unlock(); @@ -396,7 +391,7 @@ static void freezer_apply_state(struct freezer *freezer, bool freeze, */ static void freezer_change_state(struct freezer *freezer, bool freeze) { - struct cgroup *pos; + struct cgroup_subsys_state *pos; /* update @freezer */ spin_lock_irq(&freezer->lock); @@ -409,8 +404,8 @@ static void freezer_change_state(struct freezer *freezer, bool freeze) * CGROUP_FREEZING_PARENT. */ rcu_read_lock(); - cgroup_for_each_descendant_pre(pos, freezer->css.cgroup) { - struct freezer *pos_f = cgroup_freezer(pos); + css_for_each_descendant_pre(pos, &freezer->css) { + struct freezer *pos_f = css_freezer(pos); struct freezer *parent = parent_freezer(pos_f); /* diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 89b76e1d3aa1..be4f5036ea5e 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -210,29 +210,29 @@ static struct cpuset top_cpuset = { /** * cpuset_for_each_child - traverse online children of a cpuset * @child_cs: loop cursor pointing to the current child - * @pos_cgrp: used for iteration + * @pos_css: used for iteration * @parent_cs: target cpuset to walk children of * * Walk @child_cs through the online children of @parent_cs. Must be used * with RCU read locked. */ -#define cpuset_for_each_child(child_cs, pos_cgrp, parent_cs) \ - cgroup_for_each_child((pos_cgrp), (parent_cs)->css.cgroup) \ - if (is_cpuset_online(((child_cs) = cgroup_cs((pos_cgrp))))) +#define cpuset_for_each_child(child_cs, pos_css, parent_cs) \ + css_for_each_child((pos_css), &(parent_cs)->css) \ + if (is_cpuset_online(((child_cs) = css_cs((pos_css))))) /** * cpuset_for_each_descendant_pre - pre-order walk of a cpuset's descendants * @des_cs: loop cursor pointing to the current descendant - * @pos_cgrp: used for iteration + * @pos_css: used for iteration * @root_cs: target cpuset to walk ancestor of * * Walk @des_cs through the online descendants of @root_cs. Must be used - * with RCU read locked. The caller may modify @pos_cgrp by calling - * cgroup_rightmost_descendant() to skip subtree. + * with RCU read locked. The caller may modify @pos_css by calling + * css_rightmost_descendant() to skip subtree. */ -#define cpuset_for_each_descendant_pre(des_cs, pos_cgrp, root_cs) \ - cgroup_for_each_descendant_pre((pos_cgrp), (root_cs)->css.cgroup) \ - if (is_cpuset_online(((des_cs) = cgroup_cs((pos_cgrp))))) +#define cpuset_for_each_descendant_pre(des_cs, pos_css, root_cs) \ + css_for_each_descendant_pre((pos_css), &(root_cs)->css) \ + if (is_cpuset_online(((des_cs) = css_cs((pos_css))))) /* * There are two global mutexes guarding cpuset structures - cpuset_mutex @@ -430,7 +430,7 @@ static void free_trial_cpuset(struct cpuset *trial) static int validate_change(struct cpuset *cur, struct cpuset *trial) { - struct cgroup *cgrp; + struct cgroup_subsys_state *css; struct cpuset *c, *par; int ret; @@ -438,7 +438,7 @@ static int validate_change(struct cpuset *cur, struct cpuset *trial) /* Each of our child cpusets must be a subset of us */ ret = -EBUSY; - cpuset_for_each_child(c, cgrp, cur) + cpuset_for_each_child(c, css, cur) if (!is_cpuset_subset(c, trial)) goto out; @@ -459,7 +459,7 @@ static int validate_change(struct cpuset *cur, struct cpuset *trial) * overlap */ ret = -EINVAL; - cpuset_for_each_child(c, cgrp, par) { + cpuset_for_each_child(c, css, par) { if ((is_cpu_exclusive(trial) || is_cpu_exclusive(c)) && c != cur && cpumask_intersects(trial->cpus_allowed, c->cpus_allowed)) @@ -508,13 +508,13 @@ static void update_domain_attr_tree(struct sched_domain_attr *dattr, struct cpuset *root_cs) { struct cpuset *cp; - struct cgroup *pos_cgrp; + struct cgroup_subsys_state *pos_css; rcu_read_lock(); - cpuset_for_each_descendant_pre(cp, pos_cgrp, root_cs) { + cpuset_for_each_descendant_pre(cp, pos_css, root_cs) { /* skip the whole subtree if @cp doesn't have any CPU */ if (cpumask_empty(cp->cpus_allowed)) { - pos_cgrp = cgroup_rightmost_descendant(pos_cgrp); + pos_css = css_rightmost_descendant(pos_css); continue; } @@ -589,7 +589,7 @@ static int generate_sched_domains(cpumask_var_t **domains, struct sched_domain_attr *dattr; /* attributes for custom domains */ int ndoms = 0; /* number of sched domains in result */ int nslot; /* next empty doms[] struct cpumask slot */ - struct cgroup *pos_cgrp; + struct cgroup_subsys_state *pos_css; doms = NULL; dattr = NULL; @@ -618,7 +618,7 @@ static int generate_sched_domains(cpumask_var_t **domains, csn = 0; rcu_read_lock(); - cpuset_for_each_descendant_pre(cp, pos_cgrp, &top_cpuset) { + cpuset_for_each_descendant_pre(cp, pos_css, &top_cpuset) { /* * Continue traversing beyond @cp iff @cp has some CPUs and * isn't load balancing. The former is obvious. The @@ -635,7 +635,7 @@ static int generate_sched_domains(cpumask_var_t **domains, csa[csn++] = cp; /* skip @cp's subtree */ - pos_cgrp = cgroup_rightmost_descendant(pos_cgrp); + pos_css = css_rightmost_descendant(pos_css); } rcu_read_unlock(); @@ -886,16 +886,16 @@ static void update_tasks_cpumask_hier(struct cpuset *root_cs, bool update_root, struct ptr_heap *heap) { struct cpuset *cp; - struct cgroup *pos_cgrp; + struct cgroup_subsys_state *pos_css; if (update_root) update_tasks_cpumask(root_cs, heap); rcu_read_lock(); - cpuset_for_each_descendant_pre(cp, pos_cgrp, root_cs) { + cpuset_for_each_descendant_pre(cp, pos_css, root_cs) { /* skip the whole subtree if @cp have some CPU */ if (!cpumask_empty(cp->cpus_allowed)) { - pos_cgrp = cgroup_rightmost_descendant(pos_cgrp); + pos_css = css_rightmost_descendant(pos_css); continue; } if (!css_tryget(&cp->css)) @@ -1143,16 +1143,16 @@ static void update_tasks_nodemask_hier(struct cpuset *root_cs, bool update_root, struct ptr_heap *heap) { struct cpuset *cp; - struct cgroup *pos_cgrp; + struct cgroup_subsys_state *pos_css; if (update_root) update_tasks_nodemask(root_cs, heap); rcu_read_lock(); - cpuset_for_each_descendant_pre(cp, pos_cgrp, root_cs) { + cpuset_for_each_descendant_pre(cp, pos_css, root_cs) { /* skip the whole subtree if @cp have some CPU */ if (!nodes_empty(cp->mems_allowed)) { - pos_cgrp = cgroup_rightmost_descendant(pos_cgrp); + pos_css = css_rightmost_descendant(pos_css); continue; } if (!css_tryget(&cp->css)) @@ -1973,7 +1973,7 @@ static int cpuset_css_online(struct cgroup_subsys_state *css) struct cpuset *cs = css_cs(css); struct cpuset *parent = parent_cs(cs); struct cpuset *tmp_cs; - struct cgroup *pos_cgrp; + struct cgroup_subsys_state *pos_css; if (!parent) return 0; @@ -2005,7 +2005,7 @@ static int cpuset_css_online(struct cgroup_subsys_state *css) * (and likewise for mems) to the new cgroup. */ rcu_read_lock(); - cpuset_for_each_child(tmp_cs, pos_cgrp, parent) { + cpuset_for_each_child(tmp_cs, pos_css, parent) { if (is_mem_exclusive(tmp_cs) || is_cpu_exclusive(tmp_cs)) { rcu_read_unlock(); goto out_unlock; @@ -2252,10 +2252,10 @@ static void cpuset_hotplug_workfn(struct work_struct *work) /* if cpus or mems changed, we need to propagate to descendants */ if (cpus_updated || mems_updated) { struct cpuset *cs; - struct cgroup *pos_cgrp; + struct cgroup_subsys_state *pos_css; rcu_read_lock(); - cpuset_for_each_descendant_pre(cs, pos_cgrp, &top_cpuset) { + cpuset_for_each_descendant_pre(cs, pos_css, &top_cpuset) { if (!css_tryget(&cs->css)) continue; rcu_read_unlock(); -- cgit v1.2.3 From d515876e9d951d8cf7fc7c90db2967664bdc89ee Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:26 -0400 Subject: cgroup: relocate cgroup_advance_iter() For some reason, cgroup_advance_iter() is standing lonely all away from its iter comrades. Relocate it. This is cosmetic. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 48 ++++++++++++++++++++++++------------------------ 1 file changed, 24 insertions(+), 24 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 91eac33fac86..d56d9363d4b3 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -2981,30 +2981,6 @@ int cgroup_task_count(const struct cgroup *cgrp) return count; } -/* - * Advance a list_head iterator. The iterator should be positioned at - * the start of a css_set - */ -static void cgroup_advance_iter(struct cgroup *cgrp, struct cgroup_iter *it) -{ - struct list_head *l = it->cset_link; - struct cgrp_cset_link *link; - struct css_set *cset; - - /* Advance to the next non-empty css_set */ - do { - l = l->next; - if (l == &cgrp->cset_links) { - it->cset_link = NULL; - return; - } - link = list_entry(l, struct cgrp_cset_link, cset_link); - cset = link->cset; - } while (list_empty(&cset->tasks)); - it->cset_link = l; - it->task = cset->tasks.next; -} - /* * To reduce the fork() overhead for systems that are not actually * using their cgroups capability, we don't maintain the lists running @@ -3223,6 +3199,30 @@ css_next_descendant_post(struct cgroup_subsys_state *pos, } EXPORT_SYMBOL_GPL(css_next_descendant_post); +/* + * Advance a list_head iterator. The iterator should be positioned at + * the start of a css_set + */ +static void cgroup_advance_iter(struct cgroup *cgrp, struct cgroup_iter *it) +{ + struct list_head *l = it->cset_link; + struct cgrp_cset_link *link; + struct css_set *cset; + + /* Advance to the next non-empty css_set */ + do { + l = l->next; + if (l == &cgrp->cset_links) { + it->cset_link = NULL; + return; + } + link = list_entry(l, struct cgrp_cset_link, cset_link); + cset = link->cset; + } while (list_empty(&cset->tasks)); + it->cset_link = l; + it->task = cset->tasks.next; +} + void cgroup_iter_start(struct cgroup *cgrp, struct cgroup_iter *it) __acquires(css_set_lock) { -- cgit v1.2.3 From 0942eeeef68f9493c1bcb1a52baf612b73fcf9fb Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:26 -0400 Subject: cgroup: rename cgroup_iter to cgroup_task_iter cgroup now has multiple iterators and it's quite confusing to have something which walks over tasks of a single cgroup named cgroup_iter. Let's rename it to cgroup_task_iter. While at it, reformat / update comments and replace the overview comment above the interface function decls with proper function comments. Such overview can be useful but function comments should be more than enough here. This is pure rename and doesn't introduce any functional changes. Signed-off-by: Tejun Heo Acked-by: Li Zefan Acked-by: Michal Hocko Cc: Matt Helsley Cc: Johannes Weiner Cc: Balbir Singh --- kernel/cgroup.c | 114 ++++++++++++++++++++++++++++++++---------------- kernel/cgroup_freezer.c | 24 +++++----- 2 files changed, 89 insertions(+), 49 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index d56d9363d4b3..15c93f9c9e57 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -367,9 +367,11 @@ static struct cgrp_cset_link init_cgrp_cset_link; static int cgroup_init_idr(struct cgroup_subsys *ss, struct cgroup_subsys_state *css); -/* css_set_lock protects the list of css_set objects, and the - * chain of tasks off each css_set. Nests outside task->alloc_lock - * due to cgroup_iter_start() */ +/* + * css_set_lock protects the list of css_set objects, and the chain of + * tasks off each css_set. Nests outside task->alloc_lock due to + * cgroup_task_iter_start(). + */ static DEFINE_RWLOCK(css_set_lock); static int css_set_count; @@ -394,10 +396,12 @@ static unsigned long css_set_hash(struct cgroup_subsys_state *css[]) return key; } -/* We don't maintain the lists running through each css_set to its - * task until after the first call to cgroup_iter_start(). This - * reduces the fork()/exit() overhead for people who have cgroups - * compiled into their kernel but not actually in use */ +/* + * We don't maintain the lists running through each css_set to its task + * until after the first call to cgroup_task_iter_start(). This reduces + * the fork()/exit() overhead for people who have cgroups compiled into + * their kernel but not actually in use. + */ static int use_task_css_set_links __read_mostly; static void __put_css_set(struct css_set *cset, int taskexit) @@ -2982,10 +2986,10 @@ int cgroup_task_count(const struct cgroup *cgrp) } /* - * To reduce the fork() overhead for systems that are not actually - * using their cgroups capability, we don't maintain the lists running - * through each css_set to its tasks until we see the list actually - * used - in other words after the first call to cgroup_iter_start(). + * To reduce the fork() overhead for systems that are not actually using + * their cgroups capability, we don't maintain the lists running through + * each css_set to its tasks until we see the list actually used - in other + * words after the first call to cgroup_task_iter_start(). */ static void cgroup_enable_task_cg_lists(void) { @@ -3199,11 +3203,15 @@ css_next_descendant_post(struct cgroup_subsys_state *pos, } EXPORT_SYMBOL_GPL(css_next_descendant_post); -/* - * Advance a list_head iterator. The iterator should be positioned at - * the start of a css_set +/** + * cgroup_advance_task_iter - advance a task itererator to the next css_set + * @cgrp: the cgroup to walk tasks of + * @it: the iterator to advance + * + * Advance @it to the next css_set to walk. */ -static void cgroup_advance_iter(struct cgroup *cgrp, struct cgroup_iter *it) +static void cgroup_advance_task_iter(struct cgroup *cgrp, + struct cgroup_task_iter *it) { struct list_head *l = it->cset_link; struct cgrp_cset_link *link; @@ -3223,7 +3231,21 @@ static void cgroup_advance_iter(struct cgroup *cgrp, struct cgroup_iter *it) it->task = cset->tasks.next; } -void cgroup_iter_start(struct cgroup *cgrp, struct cgroup_iter *it) +/** + * cgroup_task_iter_start - initiate task iteration + * @cgrp: the cgroup to walk tasks of + * @it: the task iterator to use + * + * Initiate iteration through the tasks of @cgrp. The caller can call + * cgroup_task_iter_next() to walk through the tasks until the function + * returns NULL. On completion of iteration, cgroup_task_iter_end() must + * be called. + * + * Note that this function acquires a lock which is released when the + * iteration finishes. The caller can't sleep while iteration is in + * progress. + */ +void cgroup_task_iter_start(struct cgroup *cgrp, struct cgroup_task_iter *it) __acquires(css_set_lock) { /* @@ -3236,11 +3258,20 @@ void cgroup_iter_start(struct cgroup *cgrp, struct cgroup_iter *it) read_lock(&css_set_lock); it->cset_link = &cgrp->cset_links; - cgroup_advance_iter(cgrp, it); + cgroup_advance_task_iter(cgrp, it); } -struct task_struct *cgroup_iter_next(struct cgroup *cgrp, - struct cgroup_iter *it) +/** + * cgroup_task_iter_next - return the next task for the iterator + * @cgrp: the cgroup to walk tasks of + * @it: the task iterator being iterated + * + * The "next" function for task iteration. @it should have been + * initialized via cgroup_task_iter_start(). Returns NULL when the + * iteration reaches the end. + */ +struct task_struct *cgroup_task_iter_next(struct cgroup *cgrp, + struct cgroup_task_iter *it) { struct task_struct *res; struct list_head *l = it->task; @@ -3254,16 +3285,25 @@ struct task_struct *cgroup_iter_next(struct cgroup *cgrp, l = l->next; link = list_entry(it->cset_link, struct cgrp_cset_link, cset_link); if (l == &link->cset->tasks) { - /* We reached the end of this task list - move on to - * the next cg_cgroup_link */ - cgroup_advance_iter(cgrp, it); + /* + * We reached the end of this task list - move on to the + * next cgrp_cset_link. + */ + cgroup_advance_task_iter(cgrp, it); } else { it->task = l; } return res; } -void cgroup_iter_end(struct cgroup *cgrp, struct cgroup_iter *it) +/** + * cgroup_task_iter_end - finish task iteration + * @cgrp: the cgroup to walk tasks of + * @it: the task iterator to finish + * + * Finish task iteration started by cgroup_task_iter_start(). + */ +void cgroup_task_iter_end(struct cgroup *cgrp, struct cgroup_task_iter *it) __releases(css_set_lock) { read_unlock(&css_set_lock); @@ -3312,7 +3352,7 @@ static inline int started_after(void *p1, void *p2) * Iterate through all the tasks in a cgroup, calling test_task() for each, * and if it returns true, call process_task() for it also. * The test_task pointer may be NULL, meaning always true (select all tasks). - * Effectively duplicates cgroup_iter_{start,next,end}() + * Effectively duplicates cgroup_task_iter_{start,next,end}() * but does not lock css_set_lock for the call to process_task(). * The struct cgroup_scanner may be embedded in any structure of the caller's * creation. @@ -3333,7 +3373,7 @@ static inline int started_after(void *p1, void *p2) int cgroup_scan_tasks(struct cgroup_scanner *scan) { int retval, i; - struct cgroup_iter it; + struct cgroup_task_iter it; struct task_struct *p, *dropped; /* Never dereference latest_task, since it's not refcounted */ struct task_struct *latest_task = NULL; @@ -3368,8 +3408,8 @@ int cgroup_scan_tasks(struct cgroup_scanner *scan) * guarantees forward progress and that we don't miss any tasks. */ heap->size = 0; - cgroup_iter_start(scan->cgrp, &it); - while ((p = cgroup_iter_next(scan->cgrp, &it))) { + cgroup_task_iter_start(scan->cgrp, &it); + while ((p = cgroup_task_iter_next(scan->cgrp, &it))) { /* * Only affect tasks that qualify per the caller's callback, * if he provided one @@ -3402,7 +3442,7 @@ int cgroup_scan_tasks(struct cgroup_scanner *scan) * the heap and wasn't inserted */ } - cgroup_iter_end(scan->cgrp, &it); + cgroup_task_iter_end(scan->cgrp, &it); if (heap->size) { for (i = 0; i < heap->size; i++) { @@ -3608,7 +3648,7 @@ static int pidlist_array_load(struct cgroup *cgrp, enum cgroup_filetype type, pid_t *array; int length; int pid, n = 0; /* used for populating the array */ - struct cgroup_iter it; + struct cgroup_task_iter it; struct task_struct *tsk; struct cgroup_pidlist *l; @@ -3623,8 +3663,8 @@ static int pidlist_array_load(struct cgroup *cgrp, enum cgroup_filetype type, if (!array) return -ENOMEM; /* now, populate the array */ - cgroup_iter_start(cgrp, &it); - while ((tsk = cgroup_iter_next(cgrp, &it))) { + cgroup_task_iter_start(cgrp, &it); + while ((tsk = cgroup_task_iter_next(cgrp, &it))) { if (unlikely(n == length)) break; /* get tgid or pid for procs or tasks file respectively */ @@ -3635,7 +3675,7 @@ static int pidlist_array_load(struct cgroup *cgrp, enum cgroup_filetype type, if (pid > 0) /* make sure to only use valid results */ array[n++] = pid; } - cgroup_iter_end(cgrp, &it); + cgroup_task_iter_end(cgrp, &it); length = n; /* now sort & (if procs) strip out duplicates */ sort(array, length, sizeof(pid_t), cmppid, NULL); @@ -3669,7 +3709,7 @@ int cgroupstats_build(struct cgroupstats *stats, struct dentry *dentry) { int ret = -EINVAL; struct cgroup *cgrp; - struct cgroup_iter it; + struct cgroup_task_iter it; struct task_struct *tsk; /* @@ -3683,8 +3723,8 @@ int cgroupstats_build(struct cgroupstats *stats, struct dentry *dentry) ret = 0; cgrp = dentry->d_fsdata; - cgroup_iter_start(cgrp, &it); - while ((tsk = cgroup_iter_next(cgrp, &it))) { + cgroup_task_iter_start(cgrp, &it); + while ((tsk = cgroup_task_iter_next(cgrp, &it))) { switch (tsk->state) { case TASK_RUNNING: stats->nr_running++; @@ -3704,7 +3744,7 @@ int cgroupstats_build(struct cgroupstats *stats, struct dentry *dentry) break; } } - cgroup_iter_end(cgrp, &it); + cgroup_task_iter_end(cgrp, &it); err: return ret; @@ -5137,7 +5177,7 @@ void cgroup_fork(struct task_struct *child) * Adds the task to the list running through its css_set if necessary and * call the subsystem fork() callbacks. Has to be after the task is * visible on the task list in case we race with the first call to - * cgroup_iter_start() - to guarantee that the new task ends up on its + * cgroup_task_iter_start() - to guarantee that the new task ends up on its * list. */ void cgroup_post_fork(struct task_struct *child) diff --git a/kernel/cgroup_freezer.c b/kernel/cgroup_freezer.c index 98ca48d9ceb4..c9177f8fc661 100644 --- a/kernel/cgroup_freezer.c +++ b/kernel/cgroup_freezer.c @@ -258,7 +258,7 @@ static void update_if_frozen(struct cgroup_subsys_state *css) { struct freezer *freezer = css_freezer(css); struct cgroup_subsys_state *pos; - struct cgroup_iter it; + struct cgroup_task_iter it; struct task_struct *task; WARN_ON_ONCE(!rcu_read_lock_held()); @@ -279,9 +279,9 @@ static void update_if_frozen(struct cgroup_subsys_state *css) } /* are all tasks frozen? */ - cgroup_iter_start(css->cgroup, &it); + cgroup_task_iter_start(css->cgroup, &it); - while ((task = cgroup_iter_next(css->cgroup, &it))) { + while ((task = cgroup_task_iter_next(css->cgroup, &it))) { if (freezing(task)) { /* * freezer_should_skip() indicates that the task @@ -296,7 +296,7 @@ static void update_if_frozen(struct cgroup_subsys_state *css) freezer->state |= CGROUP_FROZEN; out_iter_end: - cgroup_iter_end(css->cgroup, &it); + cgroup_task_iter_end(css->cgroup, &it); out_unlock: spin_unlock_irq(&freezer->lock); } @@ -323,25 +323,25 @@ static int freezer_read(struct cgroup_subsys_state *css, struct cftype *cft, static void freeze_cgroup(struct freezer *freezer) { struct cgroup *cgroup = freezer->css.cgroup; - struct cgroup_iter it; + struct cgroup_task_iter it; struct task_struct *task; - cgroup_iter_start(cgroup, &it); - while ((task = cgroup_iter_next(cgroup, &it))) + cgroup_task_iter_start(cgroup, &it); + while ((task = cgroup_task_iter_next(cgroup, &it))) freeze_task(task); - cgroup_iter_end(cgroup, &it); + cgroup_task_iter_end(cgroup, &it); } static void unfreeze_cgroup(struct freezer *freezer) { struct cgroup *cgroup = freezer->css.cgroup; - struct cgroup_iter it; + struct cgroup_task_iter it; struct task_struct *task; - cgroup_iter_start(cgroup, &it); - while ((task = cgroup_iter_next(cgroup, &it))) + cgroup_task_iter_start(cgroup, &it); + while ((task = cgroup_task_iter_next(cgroup, &it))) __thaw_task(task); - cgroup_iter_end(cgroup, &it); + cgroup_task_iter_end(cgroup, &it); } /** -- cgit v1.2.3 From c59cd3d840b1b0a8f996cbbd9132128dcaabbeb9 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:26 -0400 Subject: cgroup: make cgroup_task_iter remember the cgroup being iterated Currently all cgroup_task_iter functions require @cgrp to be passed in, which is superflous and increases chance of usage error. Make cgroup_task_iter remember the cgroup being iterated and drop @cgrp argument from next and end functions. This patch doesn't introduce any behavior differences. Signed-off-by: Tejun Heo Acked-by: Li Zefan Acked-by: Michal Hocko Cc: Matt Helsley Cc: Johannes Weiner Cc: Balbir Singh --- kernel/cgroup.c | 32 +++++++++++++++----------------- kernel/cgroup_freezer.c | 12 ++++++------ 2 files changed, 21 insertions(+), 23 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 15c93f9c9e57..abc62ea1303c 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -3205,13 +3205,11 @@ EXPORT_SYMBOL_GPL(css_next_descendant_post); /** * cgroup_advance_task_iter - advance a task itererator to the next css_set - * @cgrp: the cgroup to walk tasks of * @it: the iterator to advance * * Advance @it to the next css_set to walk. */ -static void cgroup_advance_task_iter(struct cgroup *cgrp, - struct cgroup_task_iter *it) +static void cgroup_advance_task_iter(struct cgroup_task_iter *it) { struct list_head *l = it->cset_link; struct cgrp_cset_link *link; @@ -3220,7 +3218,7 @@ static void cgroup_advance_task_iter(struct cgroup *cgrp, /* Advance to the next non-empty css_set */ do { l = l->next; - if (l == &cgrp->cset_links) { + if (l == &it->origin_cgrp->cset_links) { it->cset_link = NULL; return; } @@ -3257,21 +3255,22 @@ void cgroup_task_iter_start(struct cgroup *cgrp, struct cgroup_task_iter *it) cgroup_enable_task_cg_lists(); read_lock(&css_set_lock); + + it->origin_cgrp = cgrp; it->cset_link = &cgrp->cset_links; - cgroup_advance_task_iter(cgrp, it); + + cgroup_advance_task_iter(it); } /** * cgroup_task_iter_next - return the next task for the iterator - * @cgrp: the cgroup to walk tasks of * @it: the task iterator being iterated * * The "next" function for task iteration. @it should have been * initialized via cgroup_task_iter_start(). Returns NULL when the * iteration reaches the end. */ -struct task_struct *cgroup_task_iter_next(struct cgroup *cgrp, - struct cgroup_task_iter *it) +struct task_struct *cgroup_task_iter_next(struct cgroup_task_iter *it) { struct task_struct *res; struct list_head *l = it->task; @@ -3289,7 +3288,7 @@ struct task_struct *cgroup_task_iter_next(struct cgroup *cgrp, * We reached the end of this task list - move on to the * next cgrp_cset_link. */ - cgroup_advance_task_iter(cgrp, it); + cgroup_advance_task_iter(it); } else { it->task = l; } @@ -3298,12 +3297,11 @@ struct task_struct *cgroup_task_iter_next(struct cgroup *cgrp, /** * cgroup_task_iter_end - finish task iteration - * @cgrp: the cgroup to walk tasks of * @it: the task iterator to finish * * Finish task iteration started by cgroup_task_iter_start(). */ -void cgroup_task_iter_end(struct cgroup *cgrp, struct cgroup_task_iter *it) +void cgroup_task_iter_end(struct cgroup_task_iter *it) __releases(css_set_lock) { read_unlock(&css_set_lock); @@ -3409,7 +3407,7 @@ int cgroup_scan_tasks(struct cgroup_scanner *scan) */ heap->size = 0; cgroup_task_iter_start(scan->cgrp, &it); - while ((p = cgroup_task_iter_next(scan->cgrp, &it))) { + while ((p = cgroup_task_iter_next(&it))) { /* * Only affect tasks that qualify per the caller's callback, * if he provided one @@ -3442,7 +3440,7 @@ int cgroup_scan_tasks(struct cgroup_scanner *scan) * the heap and wasn't inserted */ } - cgroup_task_iter_end(scan->cgrp, &it); + cgroup_task_iter_end(&it); if (heap->size) { for (i = 0; i < heap->size; i++) { @@ -3664,7 +3662,7 @@ static int pidlist_array_load(struct cgroup *cgrp, enum cgroup_filetype type, return -ENOMEM; /* now, populate the array */ cgroup_task_iter_start(cgrp, &it); - while ((tsk = cgroup_task_iter_next(cgrp, &it))) { + while ((tsk = cgroup_task_iter_next(&it))) { if (unlikely(n == length)) break; /* get tgid or pid for procs or tasks file respectively */ @@ -3675,7 +3673,7 @@ static int pidlist_array_load(struct cgroup *cgrp, enum cgroup_filetype type, if (pid > 0) /* make sure to only use valid results */ array[n++] = pid; } - cgroup_task_iter_end(cgrp, &it); + cgroup_task_iter_end(&it); length = n; /* now sort & (if procs) strip out duplicates */ sort(array, length, sizeof(pid_t), cmppid, NULL); @@ -3724,7 +3722,7 @@ int cgroupstats_build(struct cgroupstats *stats, struct dentry *dentry) cgrp = dentry->d_fsdata; cgroup_task_iter_start(cgrp, &it); - while ((tsk = cgroup_task_iter_next(cgrp, &it))) { + while ((tsk = cgroup_task_iter_next(&it))) { switch (tsk->state) { case TASK_RUNNING: stats->nr_running++; @@ -3744,7 +3742,7 @@ int cgroupstats_build(struct cgroupstats *stats, struct dentry *dentry) break; } } - cgroup_task_iter_end(cgrp, &it); + cgroup_task_iter_end(&it); err: return ret; diff --git a/kernel/cgroup_freezer.c b/kernel/cgroup_freezer.c index c9177f8fc661..e0ab9bfd679a 100644 --- a/kernel/cgroup_freezer.c +++ b/kernel/cgroup_freezer.c @@ -281,7 +281,7 @@ static void update_if_frozen(struct cgroup_subsys_state *css) /* are all tasks frozen? */ cgroup_task_iter_start(css->cgroup, &it); - while ((task = cgroup_task_iter_next(css->cgroup, &it))) { + while ((task = cgroup_task_iter_next(&it))) { if (freezing(task)) { /* * freezer_should_skip() indicates that the task @@ -296,7 +296,7 @@ static void update_if_frozen(struct cgroup_subsys_state *css) freezer->state |= CGROUP_FROZEN; out_iter_end: - cgroup_task_iter_end(css->cgroup, &it); + cgroup_task_iter_end(&it); out_unlock: spin_unlock_irq(&freezer->lock); } @@ -327,9 +327,9 @@ static void freeze_cgroup(struct freezer *freezer) struct task_struct *task; cgroup_task_iter_start(cgroup, &it); - while ((task = cgroup_task_iter_next(cgroup, &it))) + while ((task = cgroup_task_iter_next(&it))) freeze_task(task); - cgroup_task_iter_end(cgroup, &it); + cgroup_task_iter_end(&it); } static void unfreeze_cgroup(struct freezer *freezer) @@ -339,9 +339,9 @@ static void unfreeze_cgroup(struct freezer *freezer) struct task_struct *task; cgroup_task_iter_start(cgroup, &it); - while ((task = cgroup_task_iter_next(cgroup, &it))) + while ((task = cgroup_task_iter_next(&it))) __thaw_task(task); - cgroup_task_iter_end(cgroup, &it); + cgroup_task_iter_end(&it); } /** -- cgit v1.2.3 From e535837b1dae17b5a2d76ea1bc22ac1a79354624 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:26 -0400 Subject: cgroup: remove struct cgroup_scanner cgroup_scan_tasks() takes a pointer to struct cgroup_scanner as its sole argument and the only function of that struct is packing the arguments of the function call which are consisted of five fields. It's not too unusual to pack parameters into a struct when the number of arguments gets excessive or the whole set needs to be passed around a lot, but neither holds here making it just weird. Drop struct cgroup_scanner and pass the params directly to cgroup_scan_tasks(). Note that struct cpuset_change_nodemask_arg was added to cpuset.c to pass both ->cs and ->newmems pointer to cpuset_change_nodemask() using single data pointer. This doesn't make any functional differences. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 93 ++++++++++++++++++++++++++------------------------------- kernel/cpuset.c | 63 ++++++++++++++++---------------------- 2 files changed, 69 insertions(+), 87 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index abc62ea1303c..7b16ddb2569b 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -3343,32 +3343,37 @@ static inline int started_after(void *p1, void *p2) /** * cgroup_scan_tasks - iterate though all the tasks in a cgroup - * @scan: struct cgroup_scanner containing arguments for the scan + * @cgrp: the cgroup to iterate tasks of + * @test: optional test callback + * @process: process callback + * @data: data passed to @test and @process + * @heap: optional pre-allocated heap used for task iteration * - * Arguments include pointers to callback functions test_task() and - * process_task(). - * Iterate through all the tasks in a cgroup, calling test_task() for each, - * and if it returns true, call process_task() for it also. - * The test_task pointer may be NULL, meaning always true (select all tasks). - * Effectively duplicates cgroup_task_iter_{start,next,end}() - * but does not lock css_set_lock for the call to process_task(). - * The struct cgroup_scanner may be embedded in any structure of the caller's - * creation. - * It is guaranteed that process_task() will act on every task that - * is a member of the cgroup for the duration of this call. This - * function may or may not call process_task() for tasks that exit - * or move to a different cgroup during the call, or are forked or - * move into the cgroup during the call. + * Iterate through all the tasks in a cgroup, calling @test for each, and + * if it returns %true, call @process for it also. * - * Note that test_task() may be called with locks held, and may in some - * situations be called multiple times for the same task, so it should - * be cheap. - * If the heap pointer in the struct cgroup_scanner is non-NULL, a heap has been - * pre-allocated and will be used for heap operations (and its "gt" member will - * be overwritten), else a temporary heap will be used (allocation of which - * may cause this function to fail). + * @test may be NULL, meaning always true (select all tasks), which + * effectively duplicates cgroup_task_iter_{start,next,end}() but does not + * lock css_set_lock for the call to @process. + * + * It is guaranteed that @process will act on every task that is a member + * of @cgrp for the duration of this call. This function may or may not + * call @process for tasks that exit or move to a different cgroup during + * the call, or are forked or move into the cgroup during the call. + * + * Note that @test may be called with locks held, and may in some + * situations be called multiple times for the same task, so it should be + * cheap. + * + * If @heap is non-NULL, a heap has been pre-allocated and will be used for + * heap operations (and its "gt" member will be overwritten), else a + * temporary heap will be used (allocation of which may cause this function + * to fail). */ -int cgroup_scan_tasks(struct cgroup_scanner *scan) +int cgroup_scan_tasks(struct cgroup *cgrp, + bool (*test)(struct task_struct *, void *), + void (*process)(struct task_struct *, void *), + void *data, struct ptr_heap *heap) { int retval, i; struct cgroup_task_iter it; @@ -3376,12 +3381,10 @@ int cgroup_scan_tasks(struct cgroup_scanner *scan) /* Never dereference latest_task, since it's not refcounted */ struct task_struct *latest_task = NULL; struct ptr_heap tmp_heap; - struct ptr_heap *heap; struct timespec latest_time = { 0, 0 }; - if (scan->heap) { + if (heap) { /* The caller supplied our heap and pre-allocated its memory */ - heap = scan->heap; heap->gt = &started_after; } else { /* We need to allocate our own heap memory */ @@ -3394,25 +3397,24 @@ int cgroup_scan_tasks(struct cgroup_scanner *scan) again: /* - * Scan tasks in the cgroup, using the scanner's "test_task" callback - * to determine which are of interest, and using the scanner's - * "process_task" callback to process any of them that need an update. - * Since we don't want to hold any locks during the task updates, - * gather tasks to be processed in a heap structure. - * The heap is sorted by descending task start time. - * If the statically-sized heap fills up, we overflow tasks that - * started later, and in future iterations only consider tasks that - * started after the latest task in the previous pass. This + * Scan tasks in the cgroup, using the @test callback to determine + * which are of interest, and invoking @process callback on the + * ones which need an update. Since we don't want to hold any + * locks during the task updates, gather tasks to be processed in a + * heap structure. The heap is sorted by descending task start + * time. If the statically-sized heap fills up, we overflow tasks + * that started later, and in future iterations only consider tasks + * that started after the latest task in the previous pass. This * guarantees forward progress and that we don't miss any tasks. */ heap->size = 0; - cgroup_task_iter_start(scan->cgrp, &it); + cgroup_task_iter_start(cgrp, &it); while ((p = cgroup_task_iter_next(&it))) { /* * Only affect tasks that qualify per the caller's callback, * if he provided one */ - if (scan->test_task && !scan->test_task(p, scan)) + if (test && !test(p, data)) continue; /* * Only process tasks that started after the last task @@ -3450,7 +3452,7 @@ int cgroup_scan_tasks(struct cgroup_scanner *scan) latest_task = q; } /* Process the task per the caller's callback */ - scan->process_task(q, scan); + process(q, data); put_task_struct(q); } /* @@ -3467,10 +3469,9 @@ int cgroup_scan_tasks(struct cgroup_scanner *scan) return 0; } -static void cgroup_transfer_one_task(struct task_struct *task, - struct cgroup_scanner *scan) +static void cgroup_transfer_one_task(struct task_struct *task, void *data) { - struct cgroup *new_cgroup = scan->data; + struct cgroup *new_cgroup = data; mutex_lock(&cgroup_mutex); cgroup_attach_task(new_cgroup, task, false); @@ -3484,15 +3485,7 @@ static void cgroup_transfer_one_task(struct task_struct *task, */ int cgroup_transfer_tasks(struct cgroup *to, struct cgroup *from) { - struct cgroup_scanner scan; - - scan.cgrp = from; - scan.test_task = NULL; /* select all tasks in cgroup */ - scan.process_task = cgroup_transfer_one_task; - scan.heap = NULL; - scan.data = to; - - return cgroup_scan_tasks(&scan); + return cgroup_scan_tasks(from, NULL, cgroup_transfer_one_task, to, NULL); } /* diff --git a/kernel/cpuset.c b/kernel/cpuset.c index be4f5036ea5e..6fe23f2ac742 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -830,7 +830,7 @@ static struct cpuset *effective_nodemask_cpuset(struct cpuset *cs) /** * cpuset_change_cpumask - make a task's cpus_allowed the same as its cpuset's * @tsk: task to test - * @scan: struct cgroup_scanner containing the cgroup of the task + * @data: cpuset to @tsk belongs to * * Called by cgroup_scan_tasks() for each task in a cgroup whose * cpus_allowed mask needs to be changed. @@ -838,12 +838,11 @@ static struct cpuset *effective_nodemask_cpuset(struct cpuset *cs) * We don't need to re-check for the cgroup/cpuset membership, since we're * holding cpuset_mutex at this point. */ -static void cpuset_change_cpumask(struct task_struct *tsk, - struct cgroup_scanner *scan) +static void cpuset_change_cpumask(struct task_struct *tsk, void *data) { - struct cpuset *cpus_cs; + struct cpuset *cs = data; + struct cpuset *cpus_cs = effective_cpumask_cpuset(cs); - cpus_cs = effective_cpumask_cpuset(cgroup_cs(scan->cgrp)); set_cpus_allowed_ptr(tsk, cpus_cs->cpus_allowed); } @@ -862,13 +861,8 @@ static void cpuset_change_cpumask(struct task_struct *tsk, */ static void update_tasks_cpumask(struct cpuset *cs, struct ptr_heap *heap) { - struct cgroup_scanner scan; - - scan.cgrp = cs->css.cgroup; - scan.test_task = NULL; - scan.process_task = cpuset_change_cpumask; - scan.heap = heap; - cgroup_scan_tasks(&scan); + cgroup_scan_tasks(cs->css.cgroup, NULL, cpuset_change_cpumask, cs, + heap); } /* @@ -1052,20 +1046,24 @@ static void cpuset_change_task_nodemask(struct task_struct *tsk, task_unlock(tsk); } +struct cpuset_change_nodemask_arg { + struct cpuset *cs; + nodemask_t *newmems; +}; + /* * Update task's mems_allowed and rebind its mempolicy and vmas' mempolicy * of it to cpuset's new mems_allowed, and migrate pages to new nodes if * memory_migrate flag is set. Called with cpuset_mutex held. */ -static void cpuset_change_nodemask(struct task_struct *p, - struct cgroup_scanner *scan) +static void cpuset_change_nodemask(struct task_struct *p, void *data) { - struct cpuset *cs = cgroup_cs(scan->cgrp); + struct cpuset_change_nodemask_arg *arg = data; + struct cpuset *cs = arg->cs; struct mm_struct *mm; int migrate; - nodemask_t *newmems = scan->data; - cpuset_change_task_nodemask(p, newmems); + cpuset_change_task_nodemask(p, arg->newmems); mm = get_task_mm(p); if (!mm) @@ -1075,7 +1073,7 @@ static void cpuset_change_nodemask(struct task_struct *p, mpol_rebind_mm(mm, &cs->mems_allowed); if (migrate) - cpuset_migrate_mm(mm, &cs->old_mems_allowed, newmems); + cpuset_migrate_mm(mm, &cs->old_mems_allowed, arg->newmems); mmput(mm); } @@ -1093,19 +1091,14 @@ static void *cpuset_being_rebound; static void update_tasks_nodemask(struct cpuset *cs, struct ptr_heap *heap) { static nodemask_t newmems; /* protected by cpuset_mutex */ - struct cgroup_scanner scan; struct cpuset *mems_cs = effective_nodemask_cpuset(cs); + struct cpuset_change_nodemask_arg arg = { .cs = cs, + .newmems = &newmems }; cpuset_being_rebound = cs; /* causes mpol_dup() rebind */ guarantee_online_mems(mems_cs, &newmems); - scan.cgrp = cs->css.cgroup; - scan.test_task = NULL; - scan.process_task = cpuset_change_nodemask; - scan.heap = heap; - scan.data = &newmems; - /* * The mpol_rebind_mm() call takes mmap_sem, which we couldn't * take while holding tasklist_lock. Forks can happen - the @@ -1116,7 +1109,8 @@ static void update_tasks_nodemask(struct cpuset *cs, struct ptr_heap *heap) * It's ok if we rebind the same mm twice; mpol_rebind_mm() * is idempotent. Also migrate pages in each mm to new nodes. */ - cgroup_scan_tasks(&scan); + cgroup_scan_tasks(cs->css.cgroup, NULL, cpuset_change_nodemask, &arg, + heap); /* * All the tasks' nodemasks have been updated, update @@ -1263,17 +1257,18 @@ static int update_relax_domain_level(struct cpuset *cs, s64 val) /* * cpuset_change_flag - make a task's spread flags the same as its cpuset's * @tsk: task to be updated - * @scan: struct cgroup_scanner containing the cgroup of the task + * @data: cpuset to @tsk belongs to * * Called by cgroup_scan_tasks() for each task in a cgroup. * * We don't need to re-check for the cgroup/cpuset membership, since we're * holding cpuset_mutex at this point. */ -static void cpuset_change_flag(struct task_struct *tsk, - struct cgroup_scanner *scan) +static void cpuset_change_flag(struct task_struct *tsk, void *data) { - cpuset_update_task_spread_flag(cgroup_cs(scan->cgrp), tsk); + struct cpuset *cs = data; + + cpuset_update_task_spread_flag(cs, tsk); } /* @@ -1291,13 +1286,7 @@ static void cpuset_change_flag(struct task_struct *tsk, */ static void update_tasks_flags(struct cpuset *cs, struct ptr_heap *heap) { - struct cgroup_scanner scan; - - scan.cgrp = cs->css.cgroup; - scan.test_task = NULL; - scan.process_task = cpuset_change_flag; - scan.heap = heap; - cgroup_scan_tasks(&scan); + cgroup_scan_tasks(cs->css.cgroup, NULL, cpuset_change_flag, cs, heap); } /* -- cgit v1.2.3 From 72ec7029937f0518eff21b8762743c31591684f5 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:26 -0400 Subject: cgroup: make task iterators deal with cgroup_subsys_state instead of cgroup cgroup is in the process of converting to css (cgroup_subsys_state) from cgroup as the principal subsystem interface handle. This is mostly to prepare for the unified hierarchy support where css's will be created and destroyed dynamically but also helps cleaning up subsystem implementations as css is usually what they are interested in anyway. This patch converts task iterators to deal with css instead of cgroup. Note that under unified hierarchy, different sets of tasks will be considered belonging to a given cgroup depending on the subsystem in question and making the iterators deal with css instead cgroup provides them with enough information about the iteration. While at it, fix several function comment formats in cpuset.c. This patch doesn't introduce any behavior differences. Signed-off-by: Tejun Heo Acked-by: Li Zefan Acked-by: Michal Hocko Cc: Johannes Weiner Cc: Balbir Singh Cc: Matt Helsley --- kernel/cgroup.c | 112 ++++++++++++++++++++++++------------------------ kernel/cgroup_freezer.c | 26 ++++++----- kernel/cpuset.c | 41 ++++++++---------- 3 files changed, 88 insertions(+), 91 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 7b16ddb2569b..8c57301d0561 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -370,7 +370,7 @@ static int cgroup_init_idr(struct cgroup_subsys *ss, /* * css_set_lock protects the list of css_set objects, and the chain of * tasks off each css_set. Nests outside task->alloc_lock due to - * cgroup_task_iter_start(). + * css_task_iter_start(). */ static DEFINE_RWLOCK(css_set_lock); static int css_set_count; @@ -398,9 +398,9 @@ static unsigned long css_set_hash(struct cgroup_subsys_state *css[]) /* * We don't maintain the lists running through each css_set to its task - * until after the first call to cgroup_task_iter_start(). This reduces - * the fork()/exit() overhead for people who have cgroups compiled into - * their kernel but not actually in use. + * until after the first call to css_task_iter_start(). This reduces the + * fork()/exit() overhead for people who have cgroups compiled into their + * kernel but not actually in use. */ static int use_task_css_set_links __read_mostly; @@ -2989,7 +2989,7 @@ int cgroup_task_count(const struct cgroup *cgrp) * To reduce the fork() overhead for systems that are not actually using * their cgroups capability, we don't maintain the lists running through * each css_set to its tasks until we see the list actually used - in other - * words after the first call to cgroup_task_iter_start(). + * words after the first call to css_task_iter_start(). */ static void cgroup_enable_task_cg_lists(void) { @@ -3204,12 +3204,12 @@ css_next_descendant_post(struct cgroup_subsys_state *pos, EXPORT_SYMBOL_GPL(css_next_descendant_post); /** - * cgroup_advance_task_iter - advance a task itererator to the next css_set + * css_advance_task_iter - advance a task itererator to the next css_set * @it: the iterator to advance * * Advance @it to the next css_set to walk. */ -static void cgroup_advance_task_iter(struct cgroup_task_iter *it) +static void css_advance_task_iter(struct css_task_iter *it) { struct list_head *l = it->cset_link; struct cgrp_cset_link *link; @@ -3218,7 +3218,7 @@ static void cgroup_advance_task_iter(struct cgroup_task_iter *it) /* Advance to the next non-empty css_set */ do { l = l->next; - if (l == &it->origin_cgrp->cset_links) { + if (l == &it->origin_css->cgroup->cset_links) { it->cset_link = NULL; return; } @@ -3230,47 +3230,48 @@ static void cgroup_advance_task_iter(struct cgroup_task_iter *it) } /** - * cgroup_task_iter_start - initiate task iteration - * @cgrp: the cgroup to walk tasks of + * css_task_iter_start - initiate task iteration + * @css: the css to walk tasks of * @it: the task iterator to use * - * Initiate iteration through the tasks of @cgrp. The caller can call - * cgroup_task_iter_next() to walk through the tasks until the function - * returns NULL. On completion of iteration, cgroup_task_iter_end() must - * be called. + * Initiate iteration through the tasks of @css. The caller can call + * css_task_iter_next() to walk through the tasks until the function + * returns NULL. On completion of iteration, css_task_iter_end() must be + * called. * * Note that this function acquires a lock which is released when the * iteration finishes. The caller can't sleep while iteration is in * progress. */ -void cgroup_task_iter_start(struct cgroup *cgrp, struct cgroup_task_iter *it) +void css_task_iter_start(struct cgroup_subsys_state *css, + struct css_task_iter *it) __acquires(css_set_lock) { /* - * The first time anyone tries to iterate across a cgroup, - * we need to enable the list linking each css_set to its - * tasks, and fix up all existing tasks. + * The first time anyone tries to iterate across a css, we need to + * enable the list linking each css_set to its tasks, and fix up + * all existing tasks. */ if (!use_task_css_set_links) cgroup_enable_task_cg_lists(); read_lock(&css_set_lock); - it->origin_cgrp = cgrp; - it->cset_link = &cgrp->cset_links; + it->origin_css = css; + it->cset_link = &css->cgroup->cset_links; - cgroup_advance_task_iter(it); + css_advance_task_iter(it); } /** - * cgroup_task_iter_next - return the next task for the iterator + * css_task_iter_next - return the next task for the iterator * @it: the task iterator being iterated * * The "next" function for task iteration. @it should have been - * initialized via cgroup_task_iter_start(). Returns NULL when the - * iteration reaches the end. + * initialized via css_task_iter_start(). Returns NULL when the iteration + * reaches the end. */ -struct task_struct *cgroup_task_iter_next(struct cgroup_task_iter *it) +struct task_struct *css_task_iter_next(struct css_task_iter *it) { struct task_struct *res; struct list_head *l = it->task; @@ -3288,7 +3289,7 @@ struct task_struct *cgroup_task_iter_next(struct cgroup_task_iter *it) * We reached the end of this task list - move on to the * next cgrp_cset_link. */ - cgroup_advance_task_iter(it); + css_advance_task_iter(it); } else { it->task = l; } @@ -3296,12 +3297,12 @@ struct task_struct *cgroup_task_iter_next(struct cgroup_task_iter *it) } /** - * cgroup_task_iter_end - finish task iteration + * css_task_iter_end - finish task iteration * @it: the task iterator to finish * - * Finish task iteration started by cgroup_task_iter_start(). + * Finish task iteration started by css_task_iter_start(). */ -void cgroup_task_iter_end(struct cgroup_task_iter *it) +void css_task_iter_end(struct css_task_iter *it) __releases(css_set_lock) { read_unlock(&css_set_lock); @@ -3342,24 +3343,24 @@ static inline int started_after(void *p1, void *p2) } /** - * cgroup_scan_tasks - iterate though all the tasks in a cgroup - * @cgrp: the cgroup to iterate tasks of + * css_scan_tasks - iterate though all the tasks in a css + * @css: the css to iterate tasks of * @test: optional test callback * @process: process callback * @data: data passed to @test and @process * @heap: optional pre-allocated heap used for task iteration * - * Iterate through all the tasks in a cgroup, calling @test for each, and - * if it returns %true, call @process for it also. + * Iterate through all the tasks in @css, calling @test for each, and if it + * returns %true, call @process for it also. * * @test may be NULL, meaning always true (select all tasks), which - * effectively duplicates cgroup_task_iter_{start,next,end}() but does not + * effectively duplicates css_task_iter_{start,next,end}() but does not * lock css_set_lock for the call to @process. * * It is guaranteed that @process will act on every task that is a member - * of @cgrp for the duration of this call. This function may or may not - * call @process for tasks that exit or move to a different cgroup during - * the call, or are forked or move into the cgroup during the call. + * of @css for the duration of this call. This function may or may not + * call @process for tasks that exit or move to a different css during the + * call, or are forked or move into the css during the call. * * Note that @test may be called with locks held, and may in some * situations be called multiple times for the same task, so it should be @@ -3370,13 +3371,13 @@ static inline int started_after(void *p1, void *p2) * temporary heap will be used (allocation of which may cause this function * to fail). */ -int cgroup_scan_tasks(struct cgroup *cgrp, - bool (*test)(struct task_struct *, void *), - void (*process)(struct task_struct *, void *), - void *data, struct ptr_heap *heap) +int css_scan_tasks(struct cgroup_subsys_state *css, + bool (*test)(struct task_struct *, void *), + void (*process)(struct task_struct *, void *), + void *data, struct ptr_heap *heap) { int retval, i; - struct cgroup_task_iter it; + struct css_task_iter it; struct task_struct *p, *dropped; /* Never dereference latest_task, since it's not refcounted */ struct task_struct *latest_task = NULL; @@ -3397,7 +3398,7 @@ int cgroup_scan_tasks(struct cgroup *cgrp, again: /* - * Scan tasks in the cgroup, using the @test callback to determine + * Scan tasks in the css, using the @test callback to determine * which are of interest, and invoking @process callback on the * ones which need an update. Since we don't want to hold any * locks during the task updates, gather tasks to be processed in a @@ -3408,8 +3409,8 @@ int cgroup_scan_tasks(struct cgroup *cgrp, * guarantees forward progress and that we don't miss any tasks. */ heap->size = 0; - cgroup_task_iter_start(cgrp, &it); - while ((p = cgroup_task_iter_next(&it))) { + css_task_iter_start(css, &it); + while ((p = css_task_iter_next(&it))) { /* * Only affect tasks that qualify per the caller's callback, * if he provided one @@ -3442,7 +3443,7 @@ int cgroup_scan_tasks(struct cgroup *cgrp, * the heap and wasn't inserted */ } - cgroup_task_iter_end(&it); + css_task_iter_end(&it); if (heap->size) { for (i = 0; i < heap->size; i++) { @@ -3485,7 +3486,8 @@ static void cgroup_transfer_one_task(struct task_struct *task, void *data) */ int cgroup_transfer_tasks(struct cgroup *to, struct cgroup *from) { - return cgroup_scan_tasks(from, NULL, cgroup_transfer_one_task, to, NULL); + return css_scan_tasks(&from->dummy_css, NULL, cgroup_transfer_one_task, + to, NULL); } /* @@ -3639,7 +3641,7 @@ static int pidlist_array_load(struct cgroup *cgrp, enum cgroup_filetype type, pid_t *array; int length; int pid, n = 0; /* used for populating the array */ - struct cgroup_task_iter it; + struct css_task_iter it; struct task_struct *tsk; struct cgroup_pidlist *l; @@ -3654,8 +3656,8 @@ static int pidlist_array_load(struct cgroup *cgrp, enum cgroup_filetype type, if (!array) return -ENOMEM; /* now, populate the array */ - cgroup_task_iter_start(cgrp, &it); - while ((tsk = cgroup_task_iter_next(&it))) { + css_task_iter_start(&cgrp->dummy_css, &it); + while ((tsk = css_task_iter_next(&it))) { if (unlikely(n == length)) break; /* get tgid or pid for procs or tasks file respectively */ @@ -3666,7 +3668,7 @@ static int pidlist_array_load(struct cgroup *cgrp, enum cgroup_filetype type, if (pid > 0) /* make sure to only use valid results */ array[n++] = pid; } - cgroup_task_iter_end(&it); + css_task_iter_end(&it); length = n; /* now sort & (if procs) strip out duplicates */ sort(array, length, sizeof(pid_t), cmppid, NULL); @@ -3700,7 +3702,7 @@ int cgroupstats_build(struct cgroupstats *stats, struct dentry *dentry) { int ret = -EINVAL; struct cgroup *cgrp; - struct cgroup_task_iter it; + struct css_task_iter it; struct task_struct *tsk; /* @@ -3714,8 +3716,8 @@ int cgroupstats_build(struct cgroupstats *stats, struct dentry *dentry) ret = 0; cgrp = dentry->d_fsdata; - cgroup_task_iter_start(cgrp, &it); - while ((tsk = cgroup_task_iter_next(&it))) { + css_task_iter_start(&cgrp->dummy_css, &it); + while ((tsk = css_task_iter_next(&it))) { switch (tsk->state) { case TASK_RUNNING: stats->nr_running++; @@ -3735,7 +3737,7 @@ int cgroupstats_build(struct cgroupstats *stats, struct dentry *dentry) break; } } - cgroup_task_iter_end(&it); + css_task_iter_end(&it); err: return ret; diff --git a/kernel/cgroup_freezer.c b/kernel/cgroup_freezer.c index e0ab9bfd679a..5cd2b6d55243 100644 --- a/kernel/cgroup_freezer.c +++ b/kernel/cgroup_freezer.c @@ -258,7 +258,7 @@ static void update_if_frozen(struct cgroup_subsys_state *css) { struct freezer *freezer = css_freezer(css); struct cgroup_subsys_state *pos; - struct cgroup_task_iter it; + struct css_task_iter it; struct task_struct *task; WARN_ON_ONCE(!rcu_read_lock_held()); @@ -279,9 +279,9 @@ static void update_if_frozen(struct cgroup_subsys_state *css) } /* are all tasks frozen? */ - cgroup_task_iter_start(css->cgroup, &it); + css_task_iter_start(css, &it); - while ((task = cgroup_task_iter_next(&it))) { + while ((task = css_task_iter_next(&it))) { if (freezing(task)) { /* * freezer_should_skip() indicates that the task @@ -296,7 +296,7 @@ static void update_if_frozen(struct cgroup_subsys_state *css) freezer->state |= CGROUP_FROZEN; out_iter_end: - cgroup_task_iter_end(&it); + css_task_iter_end(&it); out_unlock: spin_unlock_irq(&freezer->lock); } @@ -322,26 +322,24 @@ static int freezer_read(struct cgroup_subsys_state *css, struct cftype *cft, static void freeze_cgroup(struct freezer *freezer) { - struct cgroup *cgroup = freezer->css.cgroup; - struct cgroup_task_iter it; + struct css_task_iter it; struct task_struct *task; - cgroup_task_iter_start(cgroup, &it); - while ((task = cgroup_task_iter_next(&it))) + css_task_iter_start(&freezer->css, &it); + while ((task = css_task_iter_next(&it))) freeze_task(task); - cgroup_task_iter_end(&it); + css_task_iter_end(&it); } static void unfreeze_cgroup(struct freezer *freezer) { - struct cgroup *cgroup = freezer->css.cgroup; - struct cgroup_task_iter it; + struct css_task_iter it; struct task_struct *task; - cgroup_task_iter_start(cgroup, &it); - while ((task = cgroup_task_iter_next(&it))) + css_task_iter_start(&freezer->css, &it); + while ((task = css_task_iter_next(&it))) __thaw_task(task); - cgroup_task_iter_end(&it); + css_task_iter_end(&it); } /** diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 6fe23f2ac742..39e52175f4af 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -832,8 +832,8 @@ static struct cpuset *effective_nodemask_cpuset(struct cpuset *cs) * @tsk: task to test * @data: cpuset to @tsk belongs to * - * Called by cgroup_scan_tasks() for each task in a cgroup whose - * cpus_allowed mask needs to be changed. + * Called by css_scan_tasks() for each task in a cgroup whose cpus_allowed + * mask needs to be changed. * * We don't need to re-check for the cgroup/cpuset membership, since we're * holding cpuset_mutex at this point. @@ -849,27 +849,26 @@ static void cpuset_change_cpumask(struct task_struct *tsk, void *data) /** * update_tasks_cpumask - Update the cpumasks of tasks in the cpuset. * @cs: the cpuset in which each task's cpus_allowed mask needs to be changed - * @heap: if NULL, defer allocating heap memory to cgroup_scan_tasks() + * @heap: if NULL, defer allocating heap memory to css_scan_tasks() * * Called with cpuset_mutex held * - * The cgroup_scan_tasks() function will scan all the tasks in a cgroup, + * The css_scan_tasks() function will scan all the tasks in a cgroup, * calling callback functions for each. * - * No return value. It's guaranteed that cgroup_scan_tasks() always returns 0 + * No return value. It's guaranteed that css_scan_tasks() always returns 0 * if @heap != NULL. */ static void update_tasks_cpumask(struct cpuset *cs, struct ptr_heap *heap) { - cgroup_scan_tasks(cs->css.cgroup, NULL, cpuset_change_cpumask, cs, - heap); + css_scan_tasks(&cs->css, NULL, cpuset_change_cpumask, cs, heap); } /* * update_tasks_cpumask_hier - Update the cpumasks of tasks in the hierarchy. * @root_cs: the root cpuset of the hierarchy * @update_root: update root cpuset or not? - * @heap: the heap used by cgroup_scan_tasks() + * @heap: the heap used by css_scan_tasks() * * This will update cpumasks of tasks in @root_cs and all other empty cpusets * which take on cpumask of @root_cs. @@ -1082,11 +1081,10 @@ static void *cpuset_being_rebound; /** * update_tasks_nodemask - Update the nodemasks of tasks in the cpuset. * @cs: the cpuset in which each task's mems_allowed mask needs to be changed - * @heap: if NULL, defer allocating heap memory to cgroup_scan_tasks() + * @heap: if NULL, defer allocating heap memory to css_scan_tasks() * - * Called with cpuset_mutex held - * No return value. It's guaranteed that cgroup_scan_tasks() always returns 0 - * if @heap != NULL. + * Called with cpuset_mutex held. No return value. It's guaranteed that + * css_scan_tasks() always returns 0 if @heap != NULL. */ static void update_tasks_nodemask(struct cpuset *cs, struct ptr_heap *heap) { @@ -1109,8 +1107,7 @@ static void update_tasks_nodemask(struct cpuset *cs, struct ptr_heap *heap) * It's ok if we rebind the same mm twice; mpol_rebind_mm() * is idempotent. Also migrate pages in each mm to new nodes. */ - cgroup_scan_tasks(cs->css.cgroup, NULL, cpuset_change_nodemask, &arg, - heap); + css_scan_tasks(&cs->css, NULL, cpuset_change_nodemask, &arg, heap); /* * All the tasks' nodemasks have been updated, update @@ -1126,7 +1123,7 @@ static void update_tasks_nodemask(struct cpuset *cs, struct ptr_heap *heap) * update_tasks_nodemask_hier - Update the nodemasks of tasks in the hierarchy. * @cs: the root cpuset of the hierarchy * @update_root: update the root cpuset or not? - * @heap: the heap used by cgroup_scan_tasks() + * @heap: the heap used by css_scan_tasks() * * This will update nodemasks of tasks in @root_cs and all other empty cpusets * which take on nodemask of @root_cs. @@ -1254,12 +1251,12 @@ static int update_relax_domain_level(struct cpuset *cs, s64 val) return 0; } -/* +/** * cpuset_change_flag - make a task's spread flags the same as its cpuset's * @tsk: task to be updated * @data: cpuset to @tsk belongs to * - * Called by cgroup_scan_tasks() for each task in a cgroup. + * Called by css_scan_tasks() for each task in a cgroup. * * We don't need to re-check for the cgroup/cpuset membership, since we're * holding cpuset_mutex at this point. @@ -1271,22 +1268,22 @@ static void cpuset_change_flag(struct task_struct *tsk, void *data) cpuset_update_task_spread_flag(cs, tsk); } -/* +/** * update_tasks_flags - update the spread flags of tasks in the cpuset. * @cs: the cpuset in which each task's spread flags needs to be changed - * @heap: if NULL, defer allocating heap memory to cgroup_scan_tasks() + * @heap: if NULL, defer allocating heap memory to css_scan_tasks() * * Called with cpuset_mutex held * - * The cgroup_scan_tasks() function will scan all the tasks in a cgroup, + * The css_scan_tasks() function will scan all the tasks in a cgroup, * calling callback functions for each. * - * No return value. It's guaranteed that cgroup_scan_tasks() always returns 0 + * No return value. It's guaranteed that css_scan_tasks() always returns 0 * if @heap != NULL. */ static void update_tasks_flags(struct cpuset *cs, struct ptr_heap *heap) { - cgroup_scan_tasks(cs->css.cgroup, NULL, cpuset_change_flag, cs, heap); + css_scan_tasks(&cs->css, NULL, cpuset_change_flag, cs, heap); } /* -- cgit v1.2.3 From 81eeaf0411204f52af8ef78ff107cfca2fcfec1d Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:26 -0400 Subject: cgroup: make cftype->[un]register_event() deal with cgroup_subsys_state instead of cgroup cgroup is in the process of converting to css (cgroup_subsys_state) from cgroup as the principal subsystem interface handle. This is mostly to prepare for the unified hierarchy support where css's will be created and destroyed dynamically but also helps cleaning up subsystem implementations as css is usually what they are interested in anyway. cftype->[un]register_event() is among the remaining couple interfaces which still use struct cgroup. Convert it to cgroup_subsys_state. The conversion is mostly mechanical and removes the last users of mem_cgroup_from_cont() and cg_to_vmpressure(), which are removed. v2: indentation update as suggested by Li Zefan. Signed-off-by: Tejun Heo Acked-by: Li Zefan Acked-by: Michal Hocko Cc: Johannes Weiner Cc: Balbir Singh --- kernel/cgroup.c | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 8c57301d0561..a71f2e0f9711 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -159,9 +159,9 @@ struct css_id { */ struct cgroup_event { /* - * Cgroup which the event belongs to. + * css which the event belongs to. */ - struct cgroup *cgrp; + struct cgroup_subsys_state *css; /* * Control file which the event associated. */ @@ -3955,11 +3955,12 @@ static void cgroup_event_remove(struct work_struct *work) { struct cgroup_event *event = container_of(work, struct cgroup_event, remove); - struct cgroup *cgrp = event->cgrp; + struct cgroup_subsys_state *css = event->css; + struct cgroup *cgrp = css->cgroup; remove_wait_queue(event->wqh, &event->wait); - event->cft->unregister_event(cgrp, event->cft, event->eventfd); + event->cft->unregister_event(css, event->cft, event->eventfd); /* Notify userspace the event is going away. */ eventfd_signal(event->eventfd, 1); @@ -3979,7 +3980,7 @@ static int cgroup_event_wake(wait_queue_t *wait, unsigned mode, { struct cgroup_event *event = container_of(wait, struct cgroup_event, wait); - struct cgroup *cgrp = event->cgrp; + struct cgroup *cgrp = event->css->cgroup; unsigned long flags = (unsigned long)key; if (flags & POLLHUP) { @@ -4048,7 +4049,7 @@ static int cgroup_write_event_control(struct cgroup_subsys_state *css, event = kzalloc(sizeof(*event), GFP_KERNEL); if (!event) return -ENOMEM; - event->cgrp = cgrp; + event->css = css; INIT_LIST_HEAD(&event->list); init_poll_funcptr(&event->pt, cgroup_event_ptable_queue_proc); init_waitqueue_func_entry(&event->wait, cgroup_event_wake); @@ -4099,7 +4100,7 @@ static int cgroup_write_event_control(struct cgroup_subsys_state *css, goto out_put_cfile; } - ret = event->cft->register_event(cgrp, event->cft, + ret = event->cft->register_event(css, event->cft, event->eventfd, buffer); if (ret) goto out_put_cfile; -- cgit v1.2.3 From d99c8727e7bbc01b70e2c57e6127bfab26b868fd Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:27 -0400 Subject: cgroup: make cgroup_taskset deal with cgroup_subsys_state instead of cgroup cgroup is in the process of converting to css (cgroup_subsys_state) from cgroup as the principal subsystem interface handle. This is mostly to prepare for the unified hierarchy support where css's will be created and destroyed dynamically but also helps cleaning up subsystem implementations as css is usually what they are interested in anyway. cgroup_taskset which is used by the subsystem attach methods is the last cgroup subsystem API which isn't using css as the handle. Update cgroup_taskset_cur_cgroup() to cgroup_taskset_cur_css() and cgroup_taskset_for_each() to take @skip_css instead of @skip_cgrp. The conversions are pretty mechanical. One exception is cpuset::cgroup_cs(), which lost its last user and got removed. This patch shouldn't introduce any functional changes. Signed-off-by: Tejun Heo Acked-by: Li Zefan Acked-by: Daniel Wagner Cc: Ingo Molnar Cc: Matt Helsley Cc: Steven Rostedt --- kernel/cgroup.c | 16 +++++++++------- kernel/cgroup_freezer.c | 2 +- kernel/cpuset.c | 15 +++++---------- kernel/events/core.c | 2 +- kernel/sched/core.c | 4 ++-- 5 files changed, 18 insertions(+), 21 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index a71f2e0f9711..e5bfb2a81dcb 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1907,18 +1907,20 @@ struct task_struct *cgroup_taskset_next(struct cgroup_taskset *tset) EXPORT_SYMBOL_GPL(cgroup_taskset_next); /** - * cgroup_taskset_cur_cgroup - return the matching cgroup for the current task + * cgroup_taskset_cur_css - return the matching css for the current task * @tset: taskset of interest + * @subsys_id: the ID of the target subsystem * - * Return the cgroup for the current (last returned) task of @tset. This - * function must be preceded by either cgroup_taskset_first() or - * cgroup_taskset_next(). + * Return the css for the current (last returned) task of @tset for + * subsystem specified by @subsys_id. This function must be preceded by + * either cgroup_taskset_first() or cgroup_taskset_next(). */ -struct cgroup *cgroup_taskset_cur_cgroup(struct cgroup_taskset *tset) +struct cgroup_subsys_state *cgroup_taskset_cur_css(struct cgroup_taskset *tset, + int subsys_id) { - return tset->cur_cgrp; + return cgroup_css(tset->cur_cgrp, subsys_id); } -EXPORT_SYMBOL_GPL(cgroup_taskset_cur_cgroup); +EXPORT_SYMBOL_GPL(cgroup_taskset_cur_css); /** * cgroup_taskset_size - return the number of tasks in taskset diff --git a/kernel/cgroup_freezer.c b/kernel/cgroup_freezer.c index 5cd2b6d55243..224da9aa27f5 100644 --- a/kernel/cgroup_freezer.c +++ b/kernel/cgroup_freezer.c @@ -189,7 +189,7 @@ static void freezer_attach(struct cgroup_subsys_state *new_css, * current state before executing the following - !frozen tasks may * be visible in a FROZEN cgroup and frozen tasks in a THAWED one. */ - cgroup_taskset_for_each(task, new_css->cgroup, tset) { + cgroup_taskset_for_each(task, new_css, tset) { if (!(freezer->state & CGROUP_FREEZING)) { __thaw_task(task); } else { diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 39e52175f4af..bf69717325b4 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -119,12 +119,6 @@ static inline struct cpuset *css_cs(struct cgroup_subsys_state *css) return css ? container_of(css, struct cpuset, css) : NULL; } -/* Retrieve the cpuset for a cgroup */ -static inline struct cpuset *cgroup_cs(struct cgroup *cgrp) -{ - return css_cs(cgroup_css(cgrp, cpuset_subsys_id)); -} - /* Retrieve the cpuset for a task */ static inline struct cpuset *task_cs(struct task_struct *task) { @@ -1459,7 +1453,7 @@ static int cpuset_can_attach(struct cgroup_subsys_state *css, (cpumask_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed))) goto out_unlock; - cgroup_taskset_for_each(task, css->cgroup, tset) { + cgroup_taskset_for_each(task, css, tset) { /* * Kthreads which disallow setaffinity shouldn't be moved * to a new cpuset; we don't want to change their cpu @@ -1511,9 +1505,10 @@ static void cpuset_attach(struct cgroup_subsys_state *css, struct mm_struct *mm; struct task_struct *task; struct task_struct *leader = cgroup_taskset_first(tset); - struct cgroup *oldcgrp = cgroup_taskset_cur_cgroup(tset); + struct cgroup_subsys_state *oldcss = cgroup_taskset_cur_css(tset, + cpuset_subsys_id); struct cpuset *cs = css_cs(css); - struct cpuset *oldcs = cgroup_cs(oldcgrp); + struct cpuset *oldcs = css_cs(oldcss); struct cpuset *cpus_cs = effective_cpumask_cpuset(cs); struct cpuset *mems_cs = effective_nodemask_cpuset(cs); @@ -1527,7 +1522,7 @@ static void cpuset_attach(struct cgroup_subsys_state *css, guarantee_online_mems(mems_cs, &cpuset_attach_nodemask_to); - cgroup_taskset_for_each(task, css->cgroup, tset) { + cgroup_taskset_for_each(task, css, tset) { /* * can_attach beforehand should guarantee that this doesn't * fail. TODO: have a better way to handle failure here diff --git a/kernel/events/core.c b/kernel/events/core.c index 9705a0ed1dce..c199c4f24910 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -7816,7 +7816,7 @@ static void perf_cgroup_attach(struct cgroup_subsys_state *css, { struct task_struct *task; - cgroup_taskset_for_each(task, css->cgroup, tset) + cgroup_taskset_for_each(task, css, tset) task_function_call(task, __perf_cgroup_move, task); } diff --git a/kernel/sched/core.c b/kernel/sched/core.c index cc9a49266382..a7122d5b8310 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -7135,7 +7135,7 @@ static int cpu_cgroup_can_attach(struct cgroup_subsys_state *css, { struct task_struct *task; - cgroup_taskset_for_each(task, css->cgroup, tset) { + cgroup_taskset_for_each(task, css, tset) { #ifdef CONFIG_RT_GROUP_SCHED if (!sched_rt_can_attach(css_tg(css), task)) return -EINVAL; @@ -7153,7 +7153,7 @@ static void cpu_cgroup_attach(struct cgroup_subsys_state *css, { struct task_struct *task; - cgroup_taskset_for_each(task, css->cgroup, tset) + cgroup_taskset_for_each(task, css, tset) sched_move_task(task); } -- cgit v1.2.3 From 95109b627ba6a043c181fa5fa45d1c754dd44fbc Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:27 -0400 Subject: cgroup: unexport cgroup_css() cgroup_css() no longer has any user left outside cgroup.c proper and we don't want subsystems to grow new usages of the function. cgroup core should always provide the css to use to the subsystems, which will make dynamic creation and destruction of css's across the lifetime of a cgroup much more manageable than exposing the cgroup directly to subsystems and let them dereference css's from it. Make cgroup_css() a static function in cgroup.c. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index e5bfb2a81dcb..c02a288a4e3d 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -222,6 +222,19 @@ static int cgroup_destroy_locked(struct cgroup *cgrp); static int cgroup_addrm_files(struct cgroup *cgrp, struct cftype cfts[], bool is_add); +/** + * cgroup_css - obtain a cgroup's css for the specified subsystem + * @cgrp: the cgroup of interest + * @subsys_id: the subsystem of interest + * + * Return @cgrp's css (cgroup_subsys_state) associated with @subsys_id. + */ +static struct cgroup_subsys_state *cgroup_css(struct cgroup *cgrp, + int subsys_id) +{ + return cgrp->subsys[subsys_id]; +} + /* convenient tests for these bits */ static inline bool cgroup_is_dead(const struct cgroup *cgrp) { -- cgit v1.2.3 From bd8815a6d802fc16a7a106e170593aa05dc17e72 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Aug 2013 20:11:27 -0400 Subject: cgroup: make css_for_each_descendant() and friends include the origin css in the iteration Previously, all css descendant iterators didn't include the origin (root of subtree) css in the iteration. The reasons were maintaining consistency with css_for_each_child() and that at the time of introduction more use cases needed skipping the origin anyway; however, given that css_is_descendant() considers self to be a descendant, omitting the origin css has become more confusing and looking at the accumulated use cases rather clearly indicates that including origin would result in simpler code overall. While this is a change which can easily lead to subtle bugs, cgroup API including the iterators has recently gone through major restructuring and no out-of-tree changes will be applicable without adjustments making this a relatively acceptable opportunity for this type of change. The conversions are mostly straight-forward. If the iteration block had explicit origin handling before or after, it's moved inside the iteration. If not, if (pos == origin) continue; is added. Some conversions add extra reference get/put around origin handling by consolidating origin handling and the rest. While the extra ref operations aren't strictly necessary, this shouldn't cause any noticeable difference. Signed-off-by: Tejun Heo Acked-by: Li Zefan Acked-by: Vivek Goyal Acked-by: Aristeu Rozanski Acked-by: Michal Hocko Cc: Jens Axboe Cc: Matt Helsley Cc: Johannes Weiner Cc: Balbir Singh --- kernel/cgroup.c | 29 +++++++++++------------------ kernel/cgroup_freezer.c | 29 ++++++++++++++++------------- kernel/cpuset.c | 42 ++++++++++++++++++++++++++---------------- 3 files changed, 53 insertions(+), 47 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index c02a288a4e3d..52f0498db946 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -2868,17 +2868,6 @@ static int cgroup_cfts_commit(struct cftype *cfts, bool is_add) mutex_unlock(&cgroup_mutex); - /* @root always needs to be updated */ - inode = root->dentry->d_inode; - mutex_lock(&inode->i_mutex); - mutex_lock(&cgroup_mutex); - ret = cgroup_addrm_files(root, cfts, is_add); - mutex_unlock(&cgroup_mutex); - mutex_unlock(&inode->i_mutex); - - if (ret) - goto out_deact; - /* add/rm files for all cgroups created before */ rcu_read_lock(); css_for_each_descendant_pre(css, cgroup_css(root, ss->subsys_id)) { @@ -2907,7 +2896,6 @@ static int cgroup_cfts_commit(struct cftype *cfts, bool is_add) } rcu_read_unlock(); dput(prev); -out_deact: deactivate_super(sb); return ret; } @@ -3099,7 +3087,8 @@ EXPORT_SYMBOL_GPL(css_next_child); * @root: css whose descendants to walk * * To be used by css_for_each_descendant_pre(). Find the next descendant - * to visit for pre-order traversal of @root's descendants. + * to visit for pre-order traversal of @root's descendants. @root is + * included in the iteration and the first node to be visited. * * While this function requires RCU read locking, it doesn't require the * whole traversal to be contained in a single RCU critical section. This @@ -3114,9 +3103,9 @@ css_next_descendant_pre(struct cgroup_subsys_state *pos, WARN_ON_ONCE(!rcu_read_lock_held()); - /* if first iteration, pretend we just visited @root */ + /* if first iteration, visit @root */ if (!pos) - pos = root; + return root; /* visit the first child if exists */ next = css_next_child(NULL, pos); @@ -3186,7 +3175,8 @@ css_leftmost_descendant(struct cgroup_subsys_state *pos) * @root: css whose descendants to walk * * To be used by css_for_each_descendant_post(). Find the next descendant - * to visit for post-order traversal of @root's descendants. + * to visit for post-order traversal of @root's descendants. @root is + * included in the iteration and the last node to be visited. * * While this function requires RCU read locking, it doesn't require the * whole traversal to be contained in a single RCU critical section. This @@ -3207,14 +3197,17 @@ css_next_descendant_post(struct cgroup_subsys_state *pos, return next != root ? next : NULL; } + /* if we visited @root, we're done */ + if (pos == root) + return NULL; + /* if there's an unvisited sibling, visit its leftmost descendant */ next = css_next_child(pos, css_parent(pos)); if (next) return css_leftmost_descendant(next); /* no sibling left, visit parent */ - next = css_parent(pos); - return next != root ? next : NULL; + return css_parent(pos); } EXPORT_SYMBOL_GPL(css_next_descendant_post); diff --git a/kernel/cgroup_freezer.c b/kernel/cgroup_freezer.c index 224da9aa27f5..f0ff64d0ebaa 100644 --- a/kernel/cgroup_freezer.c +++ b/kernel/cgroup_freezer.c @@ -311,7 +311,6 @@ static int freezer_read(struct cgroup_subsys_state *css, struct cftype *cft, /* update states bottom-up */ css_for_each_descendant_post(pos, css) update_if_frozen(pos); - update_if_frozen(css); rcu_read_unlock(); @@ -391,11 +390,6 @@ static void freezer_change_state(struct freezer *freezer, bool freeze) { struct cgroup_subsys_state *pos; - /* update @freezer */ - spin_lock_irq(&freezer->lock); - freezer_apply_state(freezer, freeze, CGROUP_FREEZING_SELF); - spin_unlock_irq(&freezer->lock); - /* * Update all its descendants in pre-order traversal. Each * descendant will try to inherit its parent's FREEZING state as @@ -406,14 +400,23 @@ static void freezer_change_state(struct freezer *freezer, bool freeze) struct freezer *pos_f = css_freezer(pos); struct freezer *parent = parent_freezer(pos_f); - /* - * Our update to @parent->state is already visible which is - * all we need. No need to lock @parent. For more info on - * synchronization, see freezer_post_create(). - */ spin_lock_irq(&pos_f->lock); - freezer_apply_state(pos_f, parent->state & CGROUP_FREEZING, - CGROUP_FREEZING_PARENT); + + if (pos_f == freezer) { + freezer_apply_state(pos_f, freeze, + CGROUP_FREEZING_SELF); + } else { + /* + * Our update to @parent->state is already visible + * which is all we need. No need to lock @parent. + * For more info on synchronization, see + * freezer_post_create(). + */ + freezer_apply_state(pos_f, + parent->state & CGROUP_FREEZING, + CGROUP_FREEZING_PARENT); + } + spin_unlock_irq(&pos_f->lock); } rcu_read_unlock(); diff --git a/kernel/cpuset.c b/kernel/cpuset.c index bf69717325b4..72a0383f382f 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -222,7 +222,8 @@ static struct cpuset top_cpuset = { * * Walk @des_cs through the online descendants of @root_cs. Must be used * with RCU read locked. The caller may modify @pos_css by calling - * css_rightmost_descendant() to skip subtree. + * css_rightmost_descendant() to skip subtree. @root_cs is included in the + * iteration and the first node to be visited. */ #define cpuset_for_each_descendant_pre(des_cs, pos_css, root_cs) \ css_for_each_descendant_pre((pos_css), &(root_cs)->css) \ @@ -506,6 +507,9 @@ static void update_domain_attr_tree(struct sched_domain_attr *dattr, rcu_read_lock(); cpuset_for_each_descendant_pre(cp, pos_css, root_cs) { + if (cp == root_cs) + continue; + /* skip the whole subtree if @cp doesn't have any CPU */ if (cpumask_empty(cp->cpus_allowed)) { pos_css = css_rightmost_descendant(pos_css); @@ -613,6 +617,8 @@ static int generate_sched_domains(cpumask_var_t **domains, rcu_read_lock(); cpuset_for_each_descendant_pre(cp, pos_css, &top_cpuset) { + if (cp == &top_cpuset) + continue; /* * Continue traversing beyond @cp iff @cp has some CPUs and * isn't load balancing. The former is obvious. The @@ -875,15 +881,17 @@ static void update_tasks_cpumask_hier(struct cpuset *root_cs, struct cpuset *cp; struct cgroup_subsys_state *pos_css; - if (update_root) - update_tasks_cpumask(root_cs, heap); - rcu_read_lock(); cpuset_for_each_descendant_pre(cp, pos_css, root_cs) { - /* skip the whole subtree if @cp have some CPU */ - if (!cpumask_empty(cp->cpus_allowed)) { - pos_css = css_rightmost_descendant(pos_css); - continue; + if (cp == root_cs) { + if (!update_root) + continue; + } else { + /* skip the whole subtree if @cp have some CPU */ + if (!cpumask_empty(cp->cpus_allowed)) { + pos_css = css_rightmost_descendant(pos_css); + continue; + } } if (!css_tryget(&cp->css)) continue; @@ -1130,15 +1138,17 @@ static void update_tasks_nodemask_hier(struct cpuset *root_cs, struct cpuset *cp; struct cgroup_subsys_state *pos_css; - if (update_root) - update_tasks_nodemask(root_cs, heap); - rcu_read_lock(); cpuset_for_each_descendant_pre(cp, pos_css, root_cs) { - /* skip the whole subtree if @cp have some CPU */ - if (!nodes_empty(cp->mems_allowed)) { - pos_css = css_rightmost_descendant(pos_css); - continue; + if (cp == root_cs) { + if (!update_root) + continue; + } else { + /* skip the whole subtree if @cp have some CPU */ + if (!nodes_empty(cp->mems_allowed)) { + pos_css = css_rightmost_descendant(pos_css); + continue; + } } if (!css_tryget(&cp->css)) continue; @@ -2237,7 +2247,7 @@ static void cpuset_hotplug_workfn(struct work_struct *work) rcu_read_lock(); cpuset_for_each_descendant_pre(cs, pos_css, &top_cpuset) { - if (!css_tryget(&cs->css)) + if (cs == &top_cpuset || !css_tryget(&cs->css)) continue; rcu_read_unlock(); -- cgit v1.2.3 From 851cf6e7d6366195d4ee033cdc7787df1a649a14 Mon Sep 17 00:00:00 2001 From: Andrew Jones Date: Fri, 9 Aug 2013 19:51:57 +0530 Subject: jump_label: Split jumplabel ratelimit Commit b202952075f62603bea9bfb6ebc6b0420db11949 ("perf, core: Rate limit perf_sched_events jump_label patching") introduced rate limiting for jump label disabling. The changes were made in the jump label code in order to be more widely available and to keep things tidier. This is all fine, except now jump_label.h includes linux/workqueue.h, which makes it impossible to include jump_label.h from anything that workqueue.h needs. For example, it's now impossible to include jump_label.h from asm/spinlock.h, which is done in proposed pv-ticketlock patches. This patch splits out the rate limiting related changes from jump_label.h into a new file, jump_label_ratelimit.h, to resolve the issue. Signed-off-by: Andrew Jones Link: http://lkml.kernel.org/r/1376058122-8248-10-git-send-email-raghavendra.kt@linux.vnet.ibm.com Reviewed-by: Konrad Rzeszutek Wilk Signed-off-by: Raghavendra K T Acked-by: Ingo Molnar Signed-off-by: H. Peter Anvin --- kernel/jump_label.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/jump_label.c b/kernel/jump_label.c index 60f48fa0fd0d..297a9247a3b3 100644 --- a/kernel/jump_label.c +++ b/kernel/jump_label.c @@ -13,6 +13,7 @@ #include #include #include +#include #ifdef HAVE_JUMP_LABEL -- cgit v1.2.3 From fbb00b568bc93073452d2a0f9f06e7c33d16eece Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 19 Jun 2013 23:56:22 +0200 Subject: sched: Consolidate open coded preemptible() checks preempt_schedule() and preempt_schedule_context() open code their preemptability checks. Use the standard API instead for consolidation. Signed-off-by: Frederic Weisbecker Cc: Ingo Molnar Cc: Li Zhong Cc: Paul E. McKenney Cc: Peter Zijlstra Cc: Steven Rostedt Cc: Thomas Gleixner Cc: Borislav Petkov Cc: Alex Shi Cc: Paul Turner Cc: Mike Galbraith Cc: Vincent Guittot --- kernel/context_tracking.c | 3 +-- kernel/sched/core.c | 4 +--- 2 files changed, 2 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/context_tracking.c b/kernel/context_tracking.c index 383f8231e436..942835c12ae5 100644 --- a/kernel/context_tracking.c +++ b/kernel/context_tracking.c @@ -87,10 +87,9 @@ void user_enter(void) */ void __sched notrace preempt_schedule_context(void) { - struct thread_info *ti = current_thread_info(); enum ctx_state prev_ctx; - if (likely(ti->preempt_count || irqs_disabled())) + if (likely(!preemptible())) return; /* diff --git a/kernel/sched/core.c b/kernel/sched/core.c index b7c32cb7bfeb..3fb7acee7326 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -2510,13 +2510,11 @@ void __sched schedule_preempt_disabled(void) */ asmlinkage void __sched notrace preempt_schedule(void) { - struct thread_info *ti = current_thread_info(); - /* * If there is a non-zero preempt_count or interrupts are disabled, * we do not want to preempt the current task. Just return.. */ - if (likely(ti->preempt_count || irqs_disabled())) + if (likely(!preemptible())) return; do { -- cgit v1.2.3 From 2d854e5738cded368a0759f85b1197f5c044513d Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 12 Jul 2013 19:02:30 +0200 Subject: context_tracing: Fix guest accounting with native vtime 1) If context tracking is enabled with native vtime accounting (which combo is useless except for dev testing), we call vtime_guest_enter() and vtime_guest_exit() on host <-> guest switches. But those are stubs in this configurations. As a result, cputime is not correctly flushed on kvm context switches. 2) If context tracking runs but is disabled on some CPUs, those CPUs end up calling __guest_enter/__guest_exit which in turn call vtime_account_system(). We don't want to call this because we run in tick based accounting for these CPUs. Refactor the guest_enter/guest_exit code such that all combinations finally work. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman --- kernel/context_tracking.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/context_tracking.c b/kernel/context_tracking.c index 942835c12ae5..1f47119c5b09 100644 --- a/kernel/context_tracking.c +++ b/kernel/context_tracking.c @@ -141,12 +141,13 @@ void user_exit(void) local_irq_restore(flags); } +#ifdef CONFIG_VIRT_CPU_ACCOUNTING_GEN void guest_enter(void) { if (vtime_accounting_enabled()) vtime_guest_enter(current); else - __guest_enter(); + current->flags |= PF_VCPU; } EXPORT_SYMBOL_GPL(guest_enter); @@ -155,9 +156,10 @@ void guest_exit(void) if (vtime_accounting_enabled()) vtime_guest_exit(current); else - __guest_exit(); + current->flags &= ~PF_VCPU; } EXPORT_SYMBOL_GPL(guest_exit); +#endif /* CONFIG_VIRT_CPU_ACCOUNTING_GEN */ /** -- cgit v1.2.3 From 5b206d48e58204e84d249c4eb18651a1ff7a1274 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 12 Jul 2013 19:05:14 +0200 Subject: vtime: Update a few comments Update a stale comment from the old vtime era and document some locking that might be non obvious. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman --- kernel/sched/cputime.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'kernel') diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c index a7959e05a9d5..223a35efa0a6 100644 --- a/kernel/sched/cputime.c +++ b/kernel/sched/cputime.c @@ -712,6 +712,13 @@ void vtime_user_enter(struct task_struct *tsk) void vtime_guest_enter(struct task_struct *tsk) { + /* + * The flags must be updated under the lock with + * the vtime_snap flush and update. + * That enforces a right ordering and update sequence + * synchronization against the reader (task_gtime()) + * that can thus safely catch up with a tickless delta. + */ write_seqlock(&tsk->vtime_seqlock); __vtime_account_system(tsk); current->flags |= PF_VCPU; -- cgit v1.2.3 From d65ec12127a5b6c6d7f5331c78157dab98a20ff0 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 11 Jul 2013 23:59:33 +0200 Subject: context_tracking: Fix runtime CPU off-case As long as the context tracking is enabled on any CPU, even a single one, all other CPUs need to keep track of their user <-> kernel boundaries cross as well. This is because a task can sleep while servicing an exception that happened in the kernel or in userspace. Then when the task eventually wakes up and return from the exception, the CPU needs to know if we resume in userspace or in the kernel. exception_exit() get this information from exception_enter() that saved the previous state. If the CPU where the exception happened didn't keep track of these informations, exception_exit() doesn't know which state tracking to restore on the CPU where the task got migrated and we may return to userspace with the context tracking subsystem thinking that we are in kernel mode. This can be fixed in the long term if we move our context tracking probes on very low level arch fast path user <-> kernel boundary, although even that is worrisome as an exception can still happen in the few instructions between the probe and the actual iret. Also we are not yet ready to set these probes in the fast path given the potential overhead problem it induces. So let's fix this by always enable context tracking even on CPUs that are not in the full dynticks range. OTOH we can spare the rcu_user_*() and vtime_user_*() calls there because the tick runs on these CPUs and we can handle RCU state machine and cputime accounting through it. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman --- kernel/context_tracking.c | 52 ++++++++++++++++++++++++++++++----------------- 1 file changed, 33 insertions(+), 19 deletions(-) (limited to 'kernel') diff --git a/kernel/context_tracking.c b/kernel/context_tracking.c index 1f47119c5b09..7b095de356c5 100644 --- a/kernel/context_tracking.c +++ b/kernel/context_tracking.c @@ -54,17 +54,31 @@ void user_enter(void) WARN_ON_ONCE(!current->mm); local_irq_save(flags); - if (__this_cpu_read(context_tracking.active) && - __this_cpu_read(context_tracking.state) != IN_USER) { + if ( __this_cpu_read(context_tracking.state) != IN_USER) { + if (__this_cpu_read(context_tracking.active)) { + /* + * At this stage, only low level arch entry code remains and + * then we'll run in userspace. We can assume there won't be + * any RCU read-side critical section until the next call to + * user_exit() or rcu_irq_enter(). Let's remove RCU's dependency + * on the tick. + */ + vtime_user_enter(current); + rcu_user_enter(); + } /* - * At this stage, only low level arch entry code remains and - * then we'll run in userspace. We can assume there won't be - * any RCU read-side critical section until the next call to - * user_exit() or rcu_irq_enter(). Let's remove RCU's dependency - * on the tick. + * Even if context tracking is disabled on this CPU, because it's outside + * the full dynticks mask for example, we still have to keep track of the + * context transitions and states to prevent inconsistency on those of + * other CPUs. + * If a task triggers an exception in userspace, sleep on the exception + * handler and then migrate to another CPU, that new CPU must know where + * the exception returns by the time we call exception_exit(). + * This information can only be provided by the previous CPU when it called + * exception_enter(). + * OTOH we can spare the calls to vtime and RCU when context_tracking.active + * is false because we know that CPU is not tickless. */ - vtime_user_enter(current); - rcu_user_enter(); __this_cpu_write(context_tracking.state, IN_USER); } local_irq_restore(flags); @@ -130,12 +144,14 @@ void user_exit(void) local_irq_save(flags); if (__this_cpu_read(context_tracking.state) == IN_USER) { - /* - * We are going to run code that may use RCU. Inform - * RCU core about that (ie: we may need the tick again). - */ - rcu_user_exit(); - vtime_user_exit(current); + if (__this_cpu_read(context_tracking.active)) { + /* + * We are going to run code that may use RCU. Inform + * RCU core about that (ie: we may need the tick again). + */ + rcu_user_exit(); + vtime_user_exit(current); + } __this_cpu_write(context_tracking.state, IN_KERNEL); } local_irq_restore(flags); @@ -178,8 +194,6 @@ EXPORT_SYMBOL_GPL(guest_exit); void context_tracking_task_switch(struct task_struct *prev, struct task_struct *next) { - if (__this_cpu_read(context_tracking.active)) { - clear_tsk_thread_flag(prev, TIF_NOHZ); - set_tsk_thread_flag(next, TIF_NOHZ); - } + clear_tsk_thread_flag(prev, TIF_NOHZ); + set_tsk_thread_flag(next, TIF_NOHZ); } -- cgit v1.2.3 From 2e70933866ace52091a3c11a5c104c063ab0c445 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 10 Jul 2013 00:55:25 +0200 Subject: nohz: Only enable context tracking on full dynticks CPUs The context tracking subsystem has the ability to selectively enable the tracking on any defined subset of CPU. This means that we can define a CPU range that doesn't run the context tracking and another range that does. Now what we want in practice is to enable the tracking on full dynticks CPUs only. In order to perform this, we just need to pass our full dynticks CPU range selection from the full dynticks subsystem to the context tracking. This way we can spare the overhead of RCU user extended quiescent state and vtime maintainance on the CPUs that are outside the full dynticks range. Just keep in mind the raw context tracking itself is still necessary everywhere. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman --- kernel/context_tracking.c | 5 +++++ kernel/time/tick-sched.c | 6 ++++++ 2 files changed, 11 insertions(+) (limited to 'kernel') diff --git a/kernel/context_tracking.c b/kernel/context_tracking.c index 7b095de356c5..72bcb2570d3e 100644 --- a/kernel/context_tracking.c +++ b/kernel/context_tracking.c @@ -26,6 +26,11 @@ DEFINE_PER_CPU(struct context_tracking, context_tracking) = { #endif }; +void context_tracking_cpu_set(int cpu) +{ + per_cpu(context_tracking.active, cpu) = true; +} + /** * user_enter - Inform the context tracking that the CPU is going to * enter userspace mode. diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 9563c744dad2..91a2528b5f44 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -23,6 +23,7 @@ #include #include #include +#include #include @@ -344,11 +345,16 @@ static int tick_nohz_init_all(void) void __init tick_nohz_init(void) { + int cpu; + if (!have_nohz_full_mask) { if (tick_nohz_init_all() < 0) return; } + for_each_cpu(cpu, nohz_full_mask) + context_tracking_cpu_set(cpu); + cpu_notifier(tick_nohz_cpu_down_callback, 0); cpulist_scnprintf(nohz_full_buf, sizeof(nohz_full_buf), nohz_full_mask); pr_info("NO_HZ: Full dynticks CPUs: %s.\n", nohz_full_buf); -- cgit v1.2.3 From d84d27a491880b9902b45c09be8d9e9464fb9b74 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 24 Jul 2013 21:59:29 +0200 Subject: context_tracking: Remove full dynticks' hacky dependency on wide context tracking Now that the full dynticks subsystem only enables the context tracking on full dynticks CPUs, lets remove the dependency on CONTEXT_TRACKING_FORCE This dependency was a hack to enable the context tracking widely for the full dynticks susbsystem until the latter becomes able to enable it in a more CPU-finegrained fashion. Now CONTEXT_TRACKING_FORCE only stands for testing on archs that work on support for the context tracking while full dynticks can't be used yet due to unmet dependencies. It simulates a system where all CPUs are full dynticks so that RCU user extended quiescent states and dynticks cputime accounting can be tested on the given arch. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman --- kernel/time/Kconfig | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/time/Kconfig b/kernel/time/Kconfig index 70f27e89012b..747bbc70f53b 100644 --- a/kernel/time/Kconfig +++ b/kernel/time/Kconfig @@ -105,7 +105,6 @@ config NO_HZ_FULL select RCU_USER_QS select RCU_NOCB_CPU select VIRT_CPU_ACCOUNTING_GEN - select CONTEXT_TRACKING_FORCE select IRQ_WORK help Adaptively try to shutdown the tick whenever possible, even when -- cgit v1.2.3 From b9d10be7a8e88fdcb12540387c219cdde87b0795 Mon Sep 17 00:00:00 2001 From: Toshi Kani Date: Mon, 12 Aug 2013 09:45:53 -0600 Subject: ACPI / processor: Acquire writer lock to update CPU maps CPU system maps are protected with reader/writer locks. The reader lock, get_online_cpus(), assures that the maps are not updated while holding the lock. The writer lock, cpu_hotplug_begin(), is used to udpate the cpu maps along with cpu_maps_update_begin(). However, the ACPI processor handler updates the cpu maps without holding the the writer lock. acpi_map_lsapic() is called from acpi_processor_hotadd_init() to update cpu_possible_mask and cpu_present_mask. acpi_unmap_lsapic() is called from acpi_processor_remove() to update cpu_possible_mask. Currently, they are either unprotected or protected with the reader lock, which is not correct. For example, the get_online_cpus() below is supposed to assure that cpu_possible_mask is not changed while the code is iterating with for_each_possible_cpu(). get_online_cpus(); for_each_possible_cpu(cpu) { : } put_online_cpus(); However, this lock has no protection with CPU hotplug since the ACPI processor handler does not use the writer lock when it updates cpu_possible_mask. The reader lock does not serialize within the readers. This patch protects them with the writer lock with cpu_hotplug_begin() along with cpu_maps_update_begin(), which must be held before calling cpu_hotplug_begin(). It also protects arch_register_cpu() / arch_unregister_cpu(), which creates / deletes a sysfs cpu device interface. For this purpose it changes cpu_hotplug_begin() and cpu_hotplug_done() to global and exports them in cpu.h. Signed-off-by: Toshi Kani Signed-off-by: Rafael J. Wysocki --- kernel/cpu.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/cpu.c b/kernel/cpu.c index b2b227b82123..d7f07a2da5a6 100644 --- a/kernel/cpu.c +++ b/kernel/cpu.c @@ -113,7 +113,7 @@ EXPORT_SYMBOL_GPL(put_online_cpus); * get_online_cpus() not an api which is called all that often. * */ -static void cpu_hotplug_begin(void) +void cpu_hotplug_begin(void) { cpu_hotplug.active_writer = current; @@ -127,7 +127,7 @@ static void cpu_hotplug_begin(void) } } -static void cpu_hotplug_done(void) +void cpu_hotplug_done(void) { cpu_hotplug.active_writer = NULL; mutex_unlock(&cpu_hotplug.lock); @@ -154,10 +154,7 @@ void cpu_hotplug_enable(void) cpu_maps_update_done(); } -#else /* #if CONFIG_HOTPLUG_CPU */ -static void cpu_hotplug_begin(void) {} -static void cpu_hotplug_done(void) {} -#endif /* #else #if CONFIG_HOTPLUG_CPU */ +#endif /* CONFIG_HOTPLUG_CPU */ /* Need to know about CPUs going up/down? */ int __ref register_cpu_notifier(struct notifier_block *nb) -- cgit v1.2.3 From a903f0865a190f8778c73df1a810ea6e25e5d7cf Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Tue, 13 Aug 2013 10:05:59 +0800 Subject: cpuset: fix the return value of cpuset_write_u64() Writing to this file always returns -ENODEV: # echo 1 > cpuset.memory_pressure_enabled -bash: echo: write error: No such device Signed-off-by: Li Zefan Cc: # 3.9+ Signed-off-by: Tejun Heo --- kernel/cpuset.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index e5657788fedd..010a0083c0ae 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -1608,11 +1608,13 @@ static int cpuset_write_u64(struct cgroup *cgrp, struct cftype *cft, u64 val) { struct cpuset *cs = cgroup_cs(cgrp); cpuset_filetype_t type = cft->private; - int retval = -ENODEV; + int retval = 0; mutex_lock(&cpuset_mutex); - if (!is_cpuset_online(cs)) + if (!is_cpuset_online(cs)) { + retval = -ENODEV; goto out_unlock; + } switch (type) { case FILE_CPU_EXCLUSIVE: -- cgit v1.2.3 From 40e93b39cd5b6a347333a95152ce37deef37bbd0 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 13 Aug 2013 11:01:53 -0400 Subject: cgroup: always use cgroup_css() cgroup_css() is the accessor for cgroup->subsys[] but is not used consistently. cgroup->subsys[] will become RCU protected and cgroup_css() will grow synchronization sanity checks. In preparation, make all cgroup->subsys[] dereferences use cgroup_css() consistently. This patch doesn't introduce any functional difference. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 58 ++++++++++++++++++++++++++++----------------------------- 1 file changed, 29 insertions(+), 29 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 52f0498db946..49ad96ee08e1 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -574,7 +574,7 @@ static struct css_set *find_existing_css_set(struct css_set *old_cset, /* Subsystem is in this hierarchy. So we want * the subsystem state from the new * cgroup */ - template[i] = cgrp->subsys[i]; + template[i] = cgroup_css(cgrp, i); } else { /* Subsystem is not in this hierarchy, so we * don't want to change the subsystem state */ @@ -871,7 +871,7 @@ static void cgroup_free_fn(struct work_struct *work) * Release the subsystem state objects. */ for_each_root_subsys(cgrp->root, ss) { - struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); ss->css_free(css); } @@ -1067,27 +1067,27 @@ static int rebind_subsystems(struct cgroupfs_root *root, if (bit & added_mask) { /* We're binding this subsystem to this hierarchy */ - BUG_ON(cgrp->subsys[i]); - BUG_ON(!cgroup_dummy_top->subsys[i]); - BUG_ON(cgroup_dummy_top->subsys[i]->cgroup != cgroup_dummy_top); + BUG_ON(cgroup_css(cgrp, i)); + BUG_ON(!cgroup_css(cgroup_dummy_top, i)); + BUG_ON(cgroup_css(cgroup_dummy_top, i)->cgroup != cgroup_dummy_top); cgrp->subsys[i] = cgroup_dummy_top->subsys[i]; - cgrp->subsys[i]->cgroup = cgrp; + cgroup_css(cgrp, i)->cgroup = cgrp; list_move(&ss->sibling, &root->subsys_list); ss->root = root; if (ss->bind) - ss->bind(cgrp->subsys[i]); + ss->bind(cgroup_css(cgrp, i)); /* refcount was already taken, and we're keeping it */ root->subsys_mask |= bit; } else if (bit & removed_mask) { /* We're removing this subsystem */ - BUG_ON(cgrp->subsys[i] != cgroup_dummy_top->subsys[i]); - BUG_ON(cgrp->subsys[i]->cgroup != cgrp); + BUG_ON(cgroup_css(cgrp, i) != cgroup_css(cgroup_dummy_top, i)); + BUG_ON(cgroup_css(cgrp, i)->cgroup != cgrp); if (ss->bind) - ss->bind(cgroup_dummy_top->subsys[i]); - cgroup_dummy_top->subsys[i]->cgroup = cgroup_dummy_top; + ss->bind(cgroup_css(cgroup_dummy_top, i)); + cgroup_css(cgroup_dummy_top, i)->cgroup = cgroup_dummy_top; cgrp->subsys[i] = NULL; cgroup_subsys[i]->root = &cgroup_dummy_root; list_move(&ss->sibling, &cgroup_dummy_root.subsys_list); @@ -2072,7 +2072,7 @@ static int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk, * step 1: check that we can legitimately attach to the cgroup. */ for_each_root_subsys(root, ss) { - struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); if (ss->can_attach) { retval = ss->can_attach(css, &tset); @@ -2114,7 +2114,7 @@ static int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk, * step 4: do subsystem attach callbacks. */ for_each_root_subsys(root, ss) { - struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); if (ss->attach) ss->attach(css, &tset); @@ -2136,7 +2136,7 @@ out_put_css_set_refs: out_cancel_attach: if (retval) { for_each_root_subsys(root, ss) { - struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); if (ss == failed_ss) break; @@ -2308,7 +2308,7 @@ static struct cgroup_subsys_state *cgroup_file_css(struct cfent *cfe) struct cgroup *cgrp = __d_cgrp(cfe->dentry->d_parent); if (cft->ss) - return cgrp->subsys[cft->ss->subsys_id]; + return cgroup_css(cgrp, cft->ss->subsys_id); return &cgrp->dummy_css; } @@ -4241,7 +4241,7 @@ static int cgroup_populate_dir(struct cgroup *cgrp, unsigned long subsys_mask) /* This cgroup is ready now */ for_each_root_subsys(cgrp->root, ss) { - struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); struct css_id *id = rcu_dereference_protected(css->id, true); /* @@ -4285,7 +4285,7 @@ static void init_cgroup_css(struct cgroup_subsys_state *css, css->id = NULL; if (cgrp == cgroup_dummy_top) css->flags |= CSS_ROOT; - BUG_ON(cgrp->subsys[ss->subsys_id]); + BUG_ON(cgroup_css(cgrp, ss->subsys_id)); cgrp->subsys[ss->subsys_id] = css; /* @@ -4300,7 +4300,7 @@ static void init_cgroup_css(struct cgroup_subsys_state *css, /* invoke ->css_online() on a new CSS and mark it online if successful */ static int online_css(struct cgroup_subsys *ss, struct cgroup *cgrp) { - struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); int ret = 0; lockdep_assert_held(&cgroup_mutex); @@ -4315,7 +4315,7 @@ static int online_css(struct cgroup_subsys *ss, struct cgroup *cgrp) /* if the CSS is online, invoke ->css_offline() on it and mark it offline */ static void offline_css(struct cgroup_subsys *ss, struct cgroup *cgrp) { - struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); lockdep_assert_held(&cgroup_mutex); @@ -4400,7 +4400,7 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, for_each_root_subsys(root, ss) { struct cgroup_subsys_state *css; - css = ss->css_alloc(parent->subsys[ss->subsys_id]); + css = ss->css_alloc(cgroup_css(parent, ss->subsys_id)); if (IS_ERR(css)) { err = PTR_ERR(css); goto err_free_all; @@ -4477,7 +4477,7 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, err_free_all: for_each_root_subsys(root, ss) { - struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); if (css) { percpu_ref_cancel_init(&css->refcnt); @@ -4590,7 +4590,7 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) */ atomic_set(&cgrp->css_kill_cnt, 1); for_each_root_subsys(cgrp->root, ss) { - struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); /* * Killing would put the base ref, but we need to keep it @@ -4676,7 +4676,7 @@ static void cgroup_offline_fn(struct work_struct *work) * destruction happens only after all css's are released. */ for_each_root_subsys(cgrp->root, ss) - css_put(cgrp->subsys[ss->subsys_id]); + css_put(cgroup_css(cgrp, ss->subsys_id)); /* delete this cgroup from parent->children */ list_del_rcu(&cgrp->sibling); @@ -4741,7 +4741,7 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss) /* Create the top cgroup state for this subsystem */ list_add(&ss->sibling, &cgroup_dummy_root.subsys_list); ss->root = &cgroup_dummy_root; - css = ss->css_alloc(cgroup_dummy_top->subsys[ss->subsys_id]); + css = ss->css_alloc(cgroup_css(cgroup_dummy_top, ss->subsys_id)); /* We don't handle early failures gracefully */ BUG_ON(IS_ERR(css)); init_cgroup_css(css, ss, cgroup_dummy_top); @@ -4820,7 +4820,7 @@ int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss) * struct, so this can happen first (i.e. before the dummy root * attachment). */ - css = ss->css_alloc(cgroup_dummy_top->subsys[ss->subsys_id]); + css = ss->css_alloc(cgroup_css(cgroup_dummy_top, ss->subsys_id)); if (IS_ERR(css)) { /* failure case - need to deassign the cgroup_subsys[] slot. */ cgroup_subsys[ss->subsys_id] = NULL; @@ -4936,7 +4936,7 @@ void cgroup_unload_subsys(struct cgroup_subsys *ss) * the cgrp->subsys pointer to find their state. note that this * also takes care of freeing the css_id. */ - ss->css_free(cgroup_dummy_top->subsys[ss->subsys_id]); + ss->css_free(cgroup_css(cgroup_dummy_top, ss->subsys_id)); cgroup_dummy_top->subsys[ss->subsys_id] = NULL; mutex_unlock(&cgroup_mutex); @@ -5562,8 +5562,8 @@ static int alloc_css_id(struct cgroup_subsys *ss, struct cgroup *parent, struct css_id *child_id, *parent_id; subsys_id = ss->subsys_id; - parent_css = parent->subsys[subsys_id]; - child_css = child->subsys[subsys_id]; + parent_css = cgroup_css(parent, subsys_id); + child_css = cgroup_css(child, subsys_id); parent_id = rcu_dereference_protected(parent_css->id, true); depth = parent_id->depth + 1; @@ -5624,7 +5624,7 @@ struct cgroup_subsys_state *cgroup_css_from_dir(struct file *f, int id) /* get cgroup */ cgrp = __d_cgrp(f->f_dentry); - css = cgrp->subsys[id]; + css = cgroup_css(cgrp, id); return css ? css : ERR_PTR(-ENOENT); } -- cgit v1.2.3 From 35ef10da65d43211f4cd7e7822cbb3becdfc0ae1 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 13 Aug 2013 11:01:54 -0400 Subject: cgroup: rename cgroup_subsys_state->dput_work and its callback function css (cgroup_subsys_state) will become RCU protected and there will be two stages which require punting to work item during release. To prepare for using the work item for multiple times, rename css->dput_work to css->destroy_work and css_dput_fn() to css_free_work_fn() and move work item initialization from css init to right before the actual usage. This reorganization doesn't introduce any behavior change. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 21 ++++++++++----------- 1 file changed, 10 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 49ad96ee08e1..0b280978f097 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4259,10 +4259,10 @@ err: return ret; } -static void css_dput_fn(struct work_struct *work) +static void css_free_work_fn(struct work_struct *work) { struct cgroup_subsys_state *css = - container_of(work, struct cgroup_subsys_state, dput_work); + container_of(work, struct cgroup_subsys_state, destroy_work); cgroup_dput(css->cgroup); } @@ -4272,7 +4272,14 @@ static void css_release(struct percpu_ref *ref) struct cgroup_subsys_state *css = container_of(ref, struct cgroup_subsys_state, refcnt); - schedule_work(&css->dput_work); + /* + * css holds an extra ref to @cgrp->dentry which is put on the last + * css_put(). dput() requires process context, which css_put() may + * be called without. @css->destroy_work will be used to invoke + * dput() asynchronously from css_put(). + */ + INIT_WORK(&css->destroy_work, css_free_work_fn); + schedule_work(&css->destroy_work); } static void init_cgroup_css(struct cgroup_subsys_state *css, @@ -4287,14 +4294,6 @@ static void init_cgroup_css(struct cgroup_subsys_state *css, css->flags |= CSS_ROOT; BUG_ON(cgroup_css(cgrp, ss->subsys_id)); cgrp->subsys[ss->subsys_id] = css; - - /* - * css holds an extra ref to @cgrp->dentry which is put on the last - * css_put(). dput() requires process context, which css_put() may - * be called without. @css->dput_work will be used to invoke - * dput() asynchronously from css_put(). - */ - INIT_WORK(&css->dput_work, css_dput_fn); } /* invoke ->css_online() on a new CSS and mark it online if successful */ -- cgit v1.2.3 From 0ae78e0bf10ac38ab53548e18383afc9997eca22 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 13 Aug 2013 11:01:54 -0400 Subject: cgroup: add cgroup_subsys_state->parent With the planned unified hierarchy, css's (cgroup_subsys_state) will be RCU protected and allowed to be attached and detached dynamically over the course of a cgroup's lifetime. This means that css's will stay accessible after being detached from its cgroup - the matching pointer in cgroup->subsys[] cleared - for ref draining and RCU grace period. cgroup core still wants to guarantee that the parent css is never destroyed before its children and css_parent() always returns the parent regardless of the state of the child css as long as it's accessible. This patch makes css's hold onto their parents and adds css->parent so that the parent css is never detroyed before its children and can be determined without consulting the cgroups. cgroup->dummy_css is also updated to point to the parent dummy_css; however, it doesn't need to worry about object lifetime as the parent cgroup is already pinned by the child. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 18 +++++++++++++++--- 1 file changed, 15 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 0b280978f097..5c6dd7ed26a7 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4264,6 +4264,9 @@ static void css_free_work_fn(struct work_struct *work) struct cgroup_subsys_state *css = container_of(work, struct cgroup_subsys_state, destroy_work); + if (css->parent) + css_put(css->parent); + cgroup_dput(css->cgroup); } @@ -4290,8 +4293,12 @@ static void init_cgroup_css(struct cgroup_subsys_state *css, css->ss = ss; css->flags = 0; css->id = NULL; - if (cgrp == cgroup_dummy_top) + + if (cgrp->parent) + css->parent = cgroup_css(cgrp->parent, ss->subsys_id); + else css->flags |= CSS_ROOT; + BUG_ON(cgroup_css(cgrp, ss->subsys_id)); cgrp->subsys[ss->subsys_id] = css; } @@ -4388,6 +4395,7 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, cgrp->dentry = dentry; cgrp->parent = parent; + cgrp->dummy_css.parent = &parent->dummy_css; cgrp->root = parent->root; if (notify_on_release(parent)) @@ -4436,9 +4444,13 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, list_add_tail_rcu(&cgrp->sibling, &cgrp->parent->children); root->number_of_cgroups++; - /* each css holds a ref to the cgroup's dentry */ - for_each_root_subsys(root, ss) + /* each css holds a ref to the cgroup's dentry and the parent css */ + for_each_root_subsys(root, ss) { + struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); + dget(dentry); + percpu_ref_get(&css->parent->refcnt); + } /* hold a ref to the parent's dentry */ dget(parent->dentry); -- cgit v1.2.3 From b77d7b6088377998ebf65eaea5e51008c2d75e94 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 13 Aug 2013 11:01:54 -0400 Subject: cgroup: cgroup_css_from_dir() now should be called with RCU read locked cgroup->subsys[] will become RCU protected and thus all cgroup_css() usages should either be under RCU read lock or cgroup_mutex. This patch updates cgroup_css_from_dir() which returns the matching cgroup_subsys_state given a directory file and subsys_id so that it requires RCU read lock and updates its sole user perf_cgroup_connect(). Signed-off-by: Tejun Heo Acked-by: Li Zefan Cc: Steven Rostedt Cc: Frederic Weisbecker Cc: Ingo Molnar --- kernel/cgroup.c | 12 ++++++++++-- kernel/events/core.c | 3 +++ 2 files changed, 13 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 5c6dd7ed26a7..cbb6314f1836 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -5616,8 +5616,14 @@ struct cgroup_subsys_state *css_lookup(struct cgroup_subsys *ss, int id) } EXPORT_SYMBOL_GPL(css_lookup); -/* - * get corresponding css from file open on cgroupfs directory +/** + * cgroup_css_from_dir - get corresponding css from file open on cgroup dir + * @f: directory file of interest + * @id: subsystem id of interest + * + * Must be called under RCU read lock. The caller is responsible for + * pinning the returned css if it needs to be accessed outside the RCU + * critical section. */ struct cgroup_subsys_state *cgroup_css_from_dir(struct file *f, int id) { @@ -5625,6 +5631,8 @@ struct cgroup_subsys_state *cgroup_css_from_dir(struct file *f, int id) struct inode *inode; struct cgroup_subsys_state *css; + WARN_ON_ONCE(!rcu_read_lock_held()); + inode = file_inode(f); /* check in cgroup filesystem dir */ if (inode->i_op != &cgroup_dir_inode_operations) diff --git a/kernel/events/core.c b/kernel/events/core.c index c199c4f24910..23261f957713 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -591,6 +591,8 @@ static inline int perf_cgroup_connect(int fd, struct perf_event *event, if (!f.file) return -EBADF; + rcu_read_lock(); + css = cgroup_css_from_dir(f.file, perf_subsys_id); if (IS_ERR(css)) { ret = PTR_ERR(css); @@ -617,6 +619,7 @@ static inline int perf_cgroup_connect(int fd, struct perf_event *event, ret = -EINVAL; } out: + rcu_read_unlock(); fdput(f); return ret; } -- cgit v1.2.3 From 105347ba5da3e87facce2337c50cd5df93cc6bec Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 13 Aug 2013 11:01:55 -0400 Subject: cgroup: make cgroup_file_open() rcu_read_lock() around cgroup_css() and add cfent->css For the planned unified hierarchy, each css (cgroup_subsys_state) will be RCU protected so that it can be created and destroyed individually while allowing RCU accesses, and cgroup_css() will soon require either holding cgroup_mutex or RCU read lock. This patch updates cgroup_file_open() such that it acquires the associated css under rcu_read_lock(). While cgroup_file_css() usages in other file operations are safe due to the reference from open, cgroup_css() wouldn't know that and will still trigger warnings. It'd be cleanest to store the acquired css in file->prvidate_data for further file operations but that's already used by seqfile. This patch instead adds cfent->css to cache the associated css. Note that while this field is initialized during cfe init, it should only be considered valid while the file is open. This patch doesn't change visible behavior. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 48 +++++++++++++++++++++++++++++++----------------- 1 file changed, 31 insertions(+), 17 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index cbb6314f1836..d63beffd41e1 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -117,6 +117,7 @@ struct cfent { struct list_head node; struct dentry *dentry; struct cftype *type; + struct cgroup_subsys_state *css; /* file xattrs */ struct simple_xattrs xattrs; @@ -2301,17 +2302,6 @@ static int cgroup_sane_behavior_show(struct cgroup_subsys_state *css, return 0; } -/* return the css for the given cgroup file */ -static struct cgroup_subsys_state *cgroup_file_css(struct cfent *cfe) -{ - struct cftype *cft = cfe->type; - struct cgroup *cgrp = __d_cgrp(cfe->dentry->d_parent); - - if (cft->ss) - return cgroup_css(cgrp, cft->ss->subsys_id); - return &cgrp->dummy_css; -} - /* A buffer size big enough for numbers or short strings */ #define CGROUP_LOCAL_BUFFER_SIZE 64 @@ -2388,7 +2378,7 @@ static ssize_t cgroup_file_write(struct file *file, const char __user *buf, { struct cfent *cfe = __d_cfe(file->f_dentry); struct cftype *cft = __d_cft(file->f_dentry); - struct cgroup_subsys_state *css = cgroup_file_css(cfe); + struct cgroup_subsys_state *css = cfe->css; if (cft->write) return cft->write(css, cft, file, buf, nbytes, ppos); @@ -2430,7 +2420,7 @@ static ssize_t cgroup_file_read(struct file *file, char __user *buf, { struct cfent *cfe = __d_cfe(file->f_dentry); struct cftype *cft = __d_cft(file->f_dentry); - struct cgroup_subsys_state *css = cgroup_file_css(cfe); + struct cgroup_subsys_state *css = cfe->css; if (cft->read) return cft->read(css, cft, file, buf, nbytes, ppos); @@ -2456,7 +2446,7 @@ static int cgroup_seqfile_show(struct seq_file *m, void *arg) { struct cfent *cfe = m->private; struct cftype *cft = cfe->type; - struct cgroup_subsys_state *css = cgroup_file_css(cfe); + struct cgroup_subsys_state *css = cfe->css; if (cft->read_map) { struct cgroup_map_cb cb = { @@ -2479,7 +2469,8 @@ static int cgroup_file_open(struct inode *inode, struct file *file) { struct cfent *cfe = __d_cfe(file->f_dentry); struct cftype *cft = __d_cft(file->f_dentry); - struct cgroup_subsys_state *css = cgroup_file_css(cfe); + struct cgroup *cgrp = __d_cgrp(cfe->dentry->d_parent); + struct cgroup_subsys_state *css; int err; err = generic_file_open(inode, file); @@ -2491,7 +2482,18 @@ static int cgroup_file_open(struct inode *inode, struct file *file) * unpinned either on open failure or release. This ensures that * @css stays alive for all file operations. */ - if (css->ss && !css_tryget(css)) + rcu_read_lock(); + if (cft->ss) { + css = cgroup_css(cgrp, cft->ss->subsys_id); + if (!css_tryget(css)) + css = NULL; + } else { + css = &cgrp->dummy_css; + } + rcu_read_unlock(); + + /* css should match @cfe->css, see cgroup_add_file() for details */ + if (!css || WARN_ON_ONCE(css != cfe->css)) return -ENODEV; if (cft->read_map || cft->read_seq_string) { @@ -2510,7 +2512,7 @@ static int cgroup_file_release(struct inode *inode, struct file *file) { struct cfent *cfe = __d_cfe(file->f_dentry); struct cftype *cft = __d_cft(file->f_dentry); - struct cgroup_subsys_state *css = cgroup_file_css(cfe); + struct cgroup_subsys_state *css = cfe->css; int ret = 0; if (cft->release) @@ -2772,6 +2774,18 @@ static int cgroup_add_file(struct cgroup *cgrp, struct cftype *cft) dentry->d_fsdata = cfe; simple_xattrs_init(&cfe->xattrs); + /* + * cfe->css is used by read/write/close to determine the associated + * css. file->private_data would be a better place but that's + * already used by seqfile. Note that open will use the usual + * cgroup_css() and css_tryget() to acquire the css and this + * caching doesn't affect css lifetime management. + */ + if (cft->ss) + cfe->css = cgroup_css(cgrp, cft->ss->subsys_id); + else + cfe->css = &cgrp->dummy_css; + mode = cgroup_file_mode(cft); error = cgroup_create_file(dentry, mode | S_IFREG, cgrp->root->sb); if (!error) { -- cgit v1.2.3 From 73e80ed8007fc48a6deeb295ba37159fad274bd2 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 13 Aug 2013 11:01:55 -0400 Subject: cgroup: add __rcu modifier to cgroup->subsys[] For the planned unified hierarchy, each css (cgroup_subsys_state) will be RCU protected so that it can be created and destroyed individually while allowing RCU accesses. Previous changes ensured that all cgroup->subsys[] accesses use the cgroup_css() accessor. This patch adds __rcu modifier to cgroup->subsys[], add matching RCU dereference in cgroup_css() and convert all assignments to either rcu_assign_pointer() or RCU_INIT_POINTER(). This change prepares for the actual RCUfication of css's and doesn't introduce any visible behavior change. The conversion is verified with sparse and all accesses are properly RCU annotated. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 19 ++++++++++++++----- 1 file changed, 14 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index d63beffd41e1..c27101622567 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -229,11 +229,16 @@ static int cgroup_addrm_files(struct cgroup *cgrp, struct cftype cfts[], * @subsys_id: the subsystem of interest * * Return @cgrp's css (cgroup_subsys_state) associated with @subsys_id. + * This function must be called either under cgroup_mutex or + * rcu_read_lock() and the caller is responsible for pinning the returned + * css if it wants to keep accessing it outside the said locks. This + * function may return %NULL if @cgrp doesn't have @subsys_id enabled. */ static struct cgroup_subsys_state *cgroup_css(struct cgroup *cgrp, int subsys_id) { - return cgrp->subsys[subsys_id]; + return rcu_dereference_check(cgrp->subsys[subsys_id], + lockdep_is_held(&cgroup_mutex)); } /* convenient tests for these bits */ @@ -1072,8 +1077,10 @@ static int rebind_subsystems(struct cgroupfs_root *root, BUG_ON(!cgroup_css(cgroup_dummy_top, i)); BUG_ON(cgroup_css(cgroup_dummy_top, i)->cgroup != cgroup_dummy_top); - cgrp->subsys[i] = cgroup_dummy_top->subsys[i]; + rcu_assign_pointer(cgrp->subsys[i], + cgroup_css(cgroup_dummy_top, i)); cgroup_css(cgrp, i)->cgroup = cgrp; + list_move(&ss->sibling, &root->subsys_list); ss->root = root; if (ss->bind) @@ -1088,8 +1095,10 @@ static int rebind_subsystems(struct cgroupfs_root *root, if (ss->bind) ss->bind(cgroup_css(cgroup_dummy_top, i)); + cgroup_css(cgroup_dummy_top, i)->cgroup = cgroup_dummy_top; - cgrp->subsys[i] = NULL; + RCU_INIT_POINTER(cgrp->subsys[i], NULL); + cgroup_subsys[i]->root = &cgroup_dummy_root; list_move(&ss->sibling, &cgroup_dummy_root.subsys_list); @@ -4314,7 +4323,7 @@ static void init_cgroup_css(struct cgroup_subsys_state *css, css->flags |= CSS_ROOT; BUG_ON(cgroup_css(cgrp, ss->subsys_id)); - cgrp->subsys[ss->subsys_id] = css; + rcu_assign_pointer(cgrp->subsys[ss->subsys_id], css); } /* invoke ->css_online() on a new CSS and mark it online if successful */ @@ -4962,7 +4971,7 @@ void cgroup_unload_subsys(struct cgroup_subsys *ss) * also takes care of freeing the css_id. */ ss->css_free(cgroup_css(cgroup_dummy_top, ss->subsys_id)); - cgroup_dummy_top->subsys[ss->subsys_id] = NULL; + RCU_INIT_POINTER(cgroup_dummy_top->subsys[ss->subsys_id], NULL); mutex_unlock(&cgroup_mutex); } -- cgit v1.2.3 From 623f926b050e12b0f5e3a2f4d11c36e4ddd63541 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 13 Aug 2013 11:01:55 -0400 Subject: cgroup: reorganize css init / exit paths css (cgroup_subsys_state) lifetime management is about to be restructured. In prepartion, make the following mostly trivial changes. * init_cgroup_css() is renamed to init_css() so that it's consistent with other css handling functions. * alloc_css_id(), online_css() and offline_css() updated to take @css instead of cgroups and subsys IDs. This patch doesn't make any functional changes. v2: v1 merged two for_each_root_subsys() loops in cgroup_create() but Li Zefan pointed out that it breaks error path. Dropped. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 50 +++++++++++++++++++++++--------------------------- 1 file changed, 23 insertions(+), 27 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index c27101622567..a1ebc445f350 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -838,8 +838,7 @@ static struct backing_dev_info cgroup_backing_dev_info = { .capabilities = BDI_CAP_NO_ACCT_AND_WRITEBACK, }; -static int alloc_css_id(struct cgroup_subsys *ss, - struct cgroup *parent, struct cgroup *child); +static int alloc_css_id(struct cgroup_subsys_state *child_css); static struct inode *cgroup_new_inode(umode_t mode, struct super_block *sb) { @@ -4308,9 +4307,8 @@ static void css_release(struct percpu_ref *ref) schedule_work(&css->destroy_work); } -static void init_cgroup_css(struct cgroup_subsys_state *css, - struct cgroup_subsys *ss, - struct cgroup *cgrp) +static void init_css(struct cgroup_subsys_state *css, struct cgroup_subsys *ss, + struct cgroup *cgrp) { css->cgroup = cgrp; css->ss = ss; @@ -4327,9 +4325,9 @@ static void init_cgroup_css(struct cgroup_subsys_state *css, } /* invoke ->css_online() on a new CSS and mark it online if successful */ -static int online_css(struct cgroup_subsys *ss, struct cgroup *cgrp) +static int online_css(struct cgroup_subsys_state *css) { - struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); + struct cgroup_subsys *ss = css->ss; int ret = 0; lockdep_assert_held(&cgroup_mutex); @@ -4342,9 +4340,9 @@ static int online_css(struct cgroup_subsys *ss, struct cgroup *cgrp) } /* if the CSS is online, invoke ->css_offline() on it and mark it offline */ -static void offline_css(struct cgroup_subsys *ss, struct cgroup *cgrp) +static void offline_css(struct cgroup_subsys_state *css) { - struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); + struct cgroup_subsys *ss = css->ss; lockdep_assert_held(&cgroup_mutex); @@ -4442,10 +4440,10 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, goto err_free_all; } - init_cgroup_css(css, ss, cgrp); + init_css(css, ss, cgrp); if (ss->use_id) { - err = alloc_css_id(ss, parent, cgrp); + err = alloc_css_id(css); if (err) goto err_free_all; } @@ -4480,7 +4478,9 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, /* creation succeeded, notify subsystems */ for_each_root_subsys(root, ss) { - err = online_css(ss, cgrp); + struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); + + err = online_css(css); if (err) goto err_destroy; @@ -4700,7 +4700,7 @@ static void cgroup_offline_fn(struct work_struct *work) * initate destruction. */ for_each_root_subsys(cgrp->root, ss) - offline_css(ss, cgrp); + offline_css(cgroup_css(cgrp, ss->subsys_id)); /* * Put the css refs from cgroup_destroy_locked(). Each css holds @@ -4778,7 +4778,7 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss) css = ss->css_alloc(cgroup_css(cgroup_dummy_top, ss->subsys_id)); /* We don't handle early failures gracefully */ BUG_ON(IS_ERR(css)); - init_cgroup_css(css, ss, cgroup_dummy_top); + init_css(css, ss, cgroup_dummy_top); /* Update the init_css_set to contain a subsys * pointer to this state - since the subsystem is @@ -4793,7 +4793,7 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss) * need to invoke fork callbacks here. */ BUG_ON(!list_empty(&init_task.tasks)); - BUG_ON(online_css(ss, cgroup_dummy_top)); + BUG_ON(online_css(cgroup_css(cgroup_dummy_top, ss->subsys_id))); mutex_unlock(&cgroup_mutex); @@ -4866,8 +4866,8 @@ int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss) ss->root = &cgroup_dummy_root; /* our new subsystem will be attached to the dummy hierarchy. */ - init_cgroup_css(css, ss, cgroup_dummy_top); - /* init_idr must be after init_cgroup_css because it sets css->id. */ + init_css(css, ss, cgroup_dummy_top); + /* init_idr must be after init_css() because it sets css->id. */ if (ss->use_id) { ret = cgroup_init_idr(ss, css); if (ret) @@ -4897,7 +4897,7 @@ int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss) } write_unlock(&css_set_lock); - ret = online_css(ss, cgroup_dummy_top); + ret = online_css(cgroup_css(cgroup_dummy_top, ss->subsys_id)); if (ret) goto err_unload; @@ -4936,7 +4936,7 @@ void cgroup_unload_subsys(struct cgroup_subsys *ss) mutex_lock(&cgroup_mutex); - offline_css(ss, cgroup_dummy_top); + offline_css(cgroup_css(cgroup_dummy_top, ss->subsys_id)); if (ss->use_id) idr_destroy(&ss->idr); @@ -5588,20 +5588,16 @@ static int __init_or_module cgroup_init_idr(struct cgroup_subsys *ss, return 0; } -static int alloc_css_id(struct cgroup_subsys *ss, struct cgroup *parent, - struct cgroup *child) +static int alloc_css_id(struct cgroup_subsys_state *child_css) { - int subsys_id, i, depth = 0; - struct cgroup_subsys_state *parent_css, *child_css; + struct cgroup_subsys_state *parent_css = css_parent(child_css); struct css_id *child_id, *parent_id; + int i, depth; - subsys_id = ss->subsys_id; - parent_css = cgroup_css(parent, subsys_id); - child_css = cgroup_css(child, subsys_id); parent_id = rcu_dereference_protected(parent_css->id, true); depth = parent_id->depth + 1; - child_id = get_new_cssid(ss, depth); + child_id = get_new_cssid(child_css->ss, depth); if (IS_ERR(child_id)) return PTR_ERR(child_id); -- cgit v1.2.3 From e0acd0a68ec7dbf6b7a81a87a867ebd7ac9b76c4 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Mon, 12 Aug 2013 18:14:00 +0200 Subject: sched: fix the theoretical signal_wake_up() vs schedule() race This is only theoretical, but after try_to_wake_up(p) was changed to check p->state under p->pi_lock the code like __set_current_state(TASK_INTERRUPTIBLE); schedule(); can miss a signal. This is the special case of wait-for-condition, it relies on try_to_wake_up/schedule interaction and thus it does not need mb() between __set_current_state() and if(signal_pending). However, this __set_current_state() can move into the critical section protected by rq->lock, now that try_to_wake_up() takes another lock we need to ensure that it can't be reordered with "if (signal_pending(current))" check inside that section. The patch is actually one-liner, it simply adds smp_wmb() before spin_lock_irq(rq->lock). This is what try_to_wake_up() already does by the same reason. We turn this wmb() into the new helper, smp_mb__before_spinlock(), for better documentation and to allow the architectures to change the default implementation. While at it, kill smp_mb__after_lock(), it has no callers. Perhaps we can also add smp_mb__before/after_spinunlock() for prepare_to_wait(). Signed-off-by: Oleg Nesterov Acked-by: Peter Zijlstra Signed-off-by: Linus Torvalds --- kernel/sched/core.c | 14 +++++++++++++- 1 file changed, 13 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index b7c32cb7bfeb..ef51b0ef4bdc 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -1491,7 +1491,13 @@ try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags) unsigned long flags; int cpu, success = 0; - smp_wmb(); + /* + * If we are going to wake up a thread waiting for CONDITION we + * need to ensure that CONDITION=1 done by the caller can not be + * reordered with p->state check below. This pairs with mb() in + * set_current_state() the waiting thread does. + */ + smp_mb__before_spinlock(); raw_spin_lock_irqsave(&p->pi_lock, flags); if (!(p->state & state)) goto out; @@ -2394,6 +2400,12 @@ need_resched: if (sched_feat(HRTICK)) hrtick_clear(rq); + /* + * Make sure that signal_pending_state()->signal_pending() below + * can't be reordered with __set_current_state(TASK_INTERRUPTIBLE) + * done by the caller to avoid the race with signal_wake_up(). + */ + smp_mb__before_spinlock(); raw_spin_lock_irq(&rq->lock); switch_count = &prev->nivcsw; -- cgit v1.2.3 From 40fea92ffb5fa0ef26d10ae0fe5688bc8e61c791 Mon Sep 17 00:00:00 2001 From: Stephen Boyd Date: Tue, 13 Aug 2013 14:12:40 -0700 Subject: PM / QoS: Fix workqueue deadlock when using pm_qos_update_request_timeout() pm_qos_update_request_timeout() updates a qos and then schedules a delayed work item to bring the qos back down to the default after the timeout. When the work item runs, pm_qos_work_fn() will call pm_qos_update_request() and deadlock because it tries to cancel itself via cancel_delayed_work_sync(). Future callers of that qos will also hang waiting to cancel the work that is canceling itself. Let's extract the little bit of code that does the real work of pm_qos_update_request() and call it from the work function so that we don't deadlock. Before ed1ac6e (PM: don't use [delayed_]work_pending()) this didn't happen because the work function wouldn't try to cancel itself. Signed-off-by: Stephen Boyd Reviewed-by: Tejun Heo Cc: 3.9+ # 3.9+ Signed-off-by: Rafael J. Wysocki --- kernel/power/qos.c | 20 +++++++++++++------- 1 file changed, 13 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/power/qos.c b/kernel/power/qos.c index 06fe28589e9c..a394297f8b2f 100644 --- a/kernel/power/qos.c +++ b/kernel/power/qos.c @@ -296,6 +296,17 @@ int pm_qos_request_active(struct pm_qos_request *req) } EXPORT_SYMBOL_GPL(pm_qos_request_active); +static void __pm_qos_update_request(struct pm_qos_request *req, + s32 new_value) +{ + trace_pm_qos_update_request(req->pm_qos_class, new_value); + + if (new_value != req->node.prio) + pm_qos_update_target( + pm_qos_array[req->pm_qos_class]->constraints, + &req->node, PM_QOS_UPDATE_REQ, new_value); +} + /** * pm_qos_work_fn - the timeout handler of pm_qos_update_request_timeout * @work: work struct for the delayed work (timeout) @@ -308,7 +319,7 @@ static void pm_qos_work_fn(struct work_struct *work) struct pm_qos_request, work); - pm_qos_update_request(req, PM_QOS_DEFAULT_VALUE); + __pm_qos_update_request(req, PM_QOS_DEFAULT_VALUE); } /** @@ -364,12 +375,7 @@ void pm_qos_update_request(struct pm_qos_request *req, } cancel_delayed_work_sync(&req->work); - - trace_pm_qos_update_request(req->pm_qos_class, new_value); - if (new_value != req->node.prio) - pm_qos_update_target( - pm_qos_array[req->pm_qos_class]->constraints, - &req->node, PM_QOS_UPDATE_REQ, new_value); + __pm_qos_update_request(req, new_value); } EXPORT_SYMBOL_GPL(pm_qos_update_request); -- cgit v1.2.3 From ae7f164a09408bf21ab3c82a9e80a3ff37aa9e3e Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 13 Aug 2013 20:22:50 -0400 Subject: cgroup: move cgroup->subsys[] assignment to online_css() Currently, css (cgroup_subsys_state) lifetime is tied to that of the associated cgroup. With the planned unified hierarchy, css's will be dynamically created and destroyed within the lifetime of a cgroup. To enable such usages, css's will be individually RCU protected instead of being tied to the cgroup. In preparation, this patch moves cgroup->subsys[] assignment from init_css() to online_css(). As this means that a newly initialized css should be remembered separately and that cgroup_css() returns NULL between init and online, cgroup_create() is updated so that it stores newly created css's in a local array css_ar[] and cgroup_init/load_subsys() are updated to use local variable @css instead of using cgroup_css(). This change also slightly simplifies error path of cgroup_create(). While this patch changes when cgroup->subsys[] is initialized, this change isn't visible to subsystems or userland. v2: This patch wasn't updated accordingly after the previous "cgroup: reorganize css init / exit paths" was updated leading to missing a css_ar[] conversion in cgroup_create() and thus boot failure. Fix it. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 21 +++++++++++---------- 1 file changed, 11 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index a1ebc445f350..b9f736c3b36d 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4321,7 +4321,6 @@ static void init_css(struct cgroup_subsys_state *css, struct cgroup_subsys *ss, css->flags |= CSS_ROOT; BUG_ON(cgroup_css(cgrp, ss->subsys_id)); - rcu_assign_pointer(cgrp->subsys[ss->subsys_id], css); } /* invoke ->css_online() on a new CSS and mark it online if successful */ @@ -4334,8 +4333,10 @@ static int online_css(struct cgroup_subsys_state *css) if (ss->css_online) ret = ss->css_online(css); - if (!ret) + if (!ret) { css->flags |= CSS_ONLINE; + rcu_assign_pointer(css->cgroup->subsys[ss->subsys_id], css); + } return ret; } @@ -4366,6 +4367,7 @@ static void offline_css(struct cgroup_subsys_state *css) static long cgroup_create(struct cgroup *parent, struct dentry *dentry, umode_t mode) { + struct cgroup_subsys_state *css_ar[CGROUP_SUBSYS_COUNT] = { }; struct cgroup *cgrp; struct cgroup_name *name; struct cgroupfs_root *root = parent->root; @@ -4433,12 +4435,11 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, err = PTR_ERR(css); goto err_free_all; } + css_ar[ss->subsys_id] = css; err = percpu_ref_init(&css->refcnt, css_release); - if (err) { - ss->css_free(css); + if (err) goto err_free_all; - } init_css(css, ss, cgrp); @@ -4467,7 +4468,7 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, /* each css holds a ref to the cgroup's dentry and the parent css */ for_each_root_subsys(root, ss) { - struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); + struct cgroup_subsys_state *css = css_ar[ss->subsys_id]; dget(dentry); percpu_ref_get(&css->parent->refcnt); @@ -4478,7 +4479,7 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, /* creation succeeded, notify subsystems */ for_each_root_subsys(root, ss) { - struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); + struct cgroup_subsys_state *css = css_ar[ss->subsys_id]; err = online_css(css); if (err) @@ -4511,7 +4512,7 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, err_free_all: for_each_root_subsys(root, ss) { - struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); + struct cgroup_subsys_state *css = css_ar[ss->subsys_id]; if (css) { percpu_ref_cancel_init(&css->refcnt); @@ -4793,7 +4794,7 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss) * need to invoke fork callbacks here. */ BUG_ON(!list_empty(&init_task.tasks)); - BUG_ON(online_css(cgroup_css(cgroup_dummy_top, ss->subsys_id))); + BUG_ON(online_css(css)); mutex_unlock(&cgroup_mutex); @@ -4897,7 +4898,7 @@ int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss) } write_unlock(&css_set_lock); - ret = online_css(cgroup_css(cgroup_dummy_top, ss->subsys_id)); + ret = online_css(css); if (ret) goto err_unload; -- cgit v1.2.3 From 223dbc38d2a8745a93749dc75ed909e274ce075d Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 13 Aug 2013 20:22:50 -0400 Subject: cgroup: bounce cgroup_subsys_state ref kill confirmation to a work item css (cgroup_subsys_state) offlining, which requires process context, will be moved to ref kill confirmation. In preparation, bounce css_killed handling through css->destroy_work. css_ref_killed_fn() is renamed to css_killed_ref_fn() so that it's consistent with the new css_killed_work_fn(). This patch adds an additional work item bouncing but doesn't change the actual logic. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 21 ++++++++++++++++++--- 1 file changed, 18 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index b9f736c3b36d..398ffbbee32f 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4555,12 +4555,27 @@ static void cgroup_css_killed(struct cgroup *cgrp) schedule_work(&cgrp->destroy_work); } -static void css_ref_killed_fn(struct percpu_ref *ref) +/* + * This is called when the refcnt of a css is confirmed to be killed. + * css_tryget() is now guaranteed to fail. + */ +static void css_killed_work_fn(struct work_struct *work) +{ + struct cgroup_subsys_state *css = + container_of(work, struct cgroup_subsys_state, destroy_work); + struct cgroup *cgrp = css->cgroup; + + cgroup_css_killed(cgrp); +} + +/* css kill confirmation processing requires process context, bounce */ +static void css_killed_ref_fn(struct percpu_ref *ref) { struct cgroup_subsys_state *css = container_of(ref, struct cgroup_subsys_state, refcnt); - cgroup_css_killed(css->cgroup); + INIT_WORK(&css->destroy_work, css_killed_work_fn); + schedule_work(&css->destroy_work); } /** @@ -4634,7 +4649,7 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) percpu_ref_get(&css->refcnt); atomic_inc(&cgrp->css_kill_cnt); - percpu_ref_kill_and_confirm(&css->refcnt, css_ref_killed_fn); + percpu_ref_kill_and_confirm(&css->refcnt, css_killed_ref_fn); } cgroup_css_killed(cgrp); -- cgit v1.2.3 From f20104de55a212a9742d8df1807f1f29dc95b748 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 13 Aug 2013 20:22:50 -0400 Subject: cgroup: replace cgroup->css_kill_cnt with ->nr_css Currently, css (cgroup_subsys_state) lifetime is tied to that of the associated cgroup. With the planned unified hierarchy, css's will be dynamically created and destroyed within the lifetime of a cgroup. To enable such usages, css's will be individually RCU protected instead of being tied to the cgroup. cgroup->css_kill_cnt is used during cgroup destruction to wait for css reference count disable; however, this model doesn't work once css's lifetimes are managed separately from cgroup's. This patch replaces it with cgroup->nr_css which is an cgroup_mutex protected integer counting the number of attached css's. The count is incremented from online_css() and decremented after refcnt kill is confirmed. If the count reaches zero and the cgroup is marked dead, the second stage of cgroup destruction is kicked off. If a cgroup doesn't have any css attached at the time of rmdir, cgroup_destroy_locked() now invokes the second stage directly as no css kill confirmation would happen. cgroup_offline_fn() - the second step of cgroup destruction - is renamed to cgroup_destroy_css_killed() and now expects to be called with cgroup_mutex held. While this patch changes how css destruction is punted to work items, it shouldn't change any visible behavior. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 52 ++++++++++++++++++++++++++++------------------------ 1 file changed, 28 insertions(+), 24 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 398ffbbee32f..174f4c3d72ef 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -218,7 +218,7 @@ static int need_forkexit_callback __read_mostly; static struct cftype cgroup_base_files[]; -static void cgroup_offline_fn(struct work_struct *work); +static void cgroup_destroy_css_killed(struct cgroup *cgrp); static int cgroup_destroy_locked(struct cgroup *cgrp); static int cgroup_addrm_files(struct cgroup *cgrp, struct cftype cfts[], bool is_add); @@ -4335,6 +4335,7 @@ static int online_css(struct cgroup_subsys_state *css) ret = ss->css_online(css); if (!ret) { css->flags |= CSS_ONLINE; + css->cgroup->nr_css++; rcu_assign_pointer(css->cgroup->subsys[ss->subsys_id], css); } return ret; @@ -4545,16 +4546,6 @@ static int cgroup_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode) return cgroup_create(c_parent, dentry, mode | S_IFDIR); } -static void cgroup_css_killed(struct cgroup *cgrp) -{ - if (!atomic_dec_and_test(&cgrp->css_kill_cnt)) - return; - - /* percpu ref's of all css's are killed, kick off the next step */ - INIT_WORK(&cgrp->destroy_work, cgroup_offline_fn); - schedule_work(&cgrp->destroy_work); -} - /* * This is called when the refcnt of a css is confirmed to be killed. * css_tryget() is now guaranteed to fail. @@ -4565,7 +4556,17 @@ static void css_killed_work_fn(struct work_struct *work) container_of(work, struct cgroup_subsys_state, destroy_work); struct cgroup *cgrp = css->cgroup; - cgroup_css_killed(cgrp); + mutex_lock(&cgroup_mutex); + + /* + * If @cgrp is marked dead, it's waiting for refs of all css's to + * be disabled before proceeding to the second phase of cgroup + * destruction. If we are the last one, kick it off. + */ + if (!--cgrp->nr_css && cgroup_is_dead(cgrp)) + cgroup_destroy_css_killed(cgrp); + + mutex_unlock(&cgroup_mutex); } /* css kill confirmation processing requires process context, bounce */ @@ -4634,11 +4635,10 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) * Use percpu_ref_kill_and_confirm() to get notifications as each * css is confirmed to be seen as killed on all CPUs. The * notification callback keeps track of the number of css's to be - * killed and schedules cgroup_offline_fn() to perform the rest of - * destruction once the percpu refs of all css's are confirmed to - * be killed. + * killed and invokes cgroup_destroy_css_killed() to perform the + * rest of destruction once the percpu refs of all css's are + * confirmed to be killed. */ - atomic_set(&cgrp->css_kill_cnt, 1); for_each_root_subsys(cgrp->root, ss) { struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); @@ -4648,10 +4648,8 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) */ percpu_ref_get(&css->refcnt); - atomic_inc(&cgrp->css_kill_cnt); percpu_ref_kill_and_confirm(&css->refcnt, css_killed_ref_fn); } - cgroup_css_killed(cgrp); /* * Mark @cgrp dead. This prevents further task migration and child @@ -4668,6 +4666,15 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) list_del_init(&cgrp->release_list); raw_spin_unlock(&release_list_lock); + /* + * If @cgrp has css's attached, the second stage of cgroup + * destruction is kicked off from css_killed_work_fn() after the + * refs of all attached css's are killed. If @cgrp doesn't have + * any css, we kick it off here. + */ + if (!cgrp->nr_css) + cgroup_destroy_css_killed(cgrp); + /* * Clear and remove @cgrp directory. The removal puts the base ref * but we aren't quite done with @cgrp yet, so hold onto it. @@ -4693,7 +4700,7 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) }; /** - * cgroup_offline_fn - the second step of cgroup destruction + * cgroup_destroy_css_killed - the second step of cgroup destruction * @work: cgroup->destroy_free_work * * This function is invoked from a work item for a cgroup which is being @@ -4702,14 +4709,13 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) * is the second step of destruction described in the comment above * cgroup_destroy_locked(). */ -static void cgroup_offline_fn(struct work_struct *work) +static void cgroup_destroy_css_killed(struct cgroup *cgrp) { - struct cgroup *cgrp = container_of(work, struct cgroup, destroy_work); struct cgroup *parent = cgrp->parent; struct dentry *d = cgrp->dentry; struct cgroup_subsys *ss; - mutex_lock(&cgroup_mutex); + lockdep_assert_held(&cgroup_mutex); /* * css_tryget() is guaranteed to fail now. Tell subsystems to @@ -4743,8 +4749,6 @@ static void cgroup_offline_fn(struct work_struct *work) set_bit(CGRP_RELEASABLE, &parent->flags); check_for_release(parent); - - mutex_unlock(&cgroup_mutex); } static int cgroup_rmdir(struct inode *unused_dir, struct dentry *dentry) -- cgit v1.2.3 From 09a503ea3a816b285b0b402b7f785eaec0c7a7e1 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 13 Aug 2013 20:22:50 -0400 Subject: cgroup: decouple cgroup_subsys_state destruction from cgroup destruction Currently, css (cgroup_subsys_state) lifetime is tied to that of the associated cgroup. css's are created when the associated cgroup is created and destroyed when it gets destroyed. Also, individual css's aren't RCU protected but the whole cgroup is. With the planned unified hierarchy, css's will need to be dynamically created and destroyed within the lifetime of a cgroup. To enable such usages, this patch decouples css destruction from cgroup destruction - offline_css() invocation and the final css_put() are moved from cgroup_destroy_css_killed() to css_killed_work_fn(). Now each css is individually offlined and put as its reference count is killed instead of waiting for all css's attached to the cgroup to finish refcnt killing and then proceeding to offlining and putting them together. While this changes the order of destruction operations, the changes shouldn't be noticeable to cgroup subsystems or userland. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 52 ++++++++++++++++++++++++---------------------------- 1 file changed, 24 insertions(+), 28 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 174f4c3d72ef..3c4c4b01ffe5 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4355,6 +4355,7 @@ static void offline_css(struct cgroup_subsys_state *css) ss->css_offline(css); css->flags &= ~CSS_ONLINE; + css->cgroup->nr_css--; } /* @@ -4558,15 +4559,30 @@ static void css_killed_work_fn(struct work_struct *work) mutex_lock(&cgroup_mutex); + /* + * css_tryget() is guaranteed to fail now. Tell subsystems to + * initate destruction. + */ + offline_css(css); + /* * If @cgrp is marked dead, it's waiting for refs of all css's to * be disabled before proceeding to the second phase of cgroup * destruction. If we are the last one, kick it off. */ - if (!--cgrp->nr_css && cgroup_is_dead(cgrp)) + if (!cgrp->nr_css && cgroup_is_dead(cgrp)) cgroup_destroy_css_killed(cgrp); mutex_unlock(&cgroup_mutex); + + /* + * Put the css refs from kill_css(). Each css holds an extra + * reference to the cgroup's dentry and cgroup removal proceeds + * regardless of css refs. On the last put of each css, whenever + * that may be, the extra dentry ref is put so that dentry + * destruction happens only after all css's are released. + */ + css_put(css); } /* css kill confirmation processing requires process context, bounce */ @@ -4633,11 +4649,10 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) * as killed on all CPUs on return. * * Use percpu_ref_kill_and_confirm() to get notifications as each - * css is confirmed to be seen as killed on all CPUs. The - * notification callback keeps track of the number of css's to be - * killed and invokes cgroup_destroy_css_killed() to perform the - * rest of destruction once the percpu refs of all css's are - * confirmed to be killed. + * css is confirmed to be seen as killed on all CPUs. + * cgroup_destroy_css_killed() will be invoked to perform the rest + * of destruction once the percpu refs of all css's are confirmed + * to be killed. */ for_each_root_subsys(cgrp->root, ss) { struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); @@ -4704,36 +4719,17 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) * @work: cgroup->destroy_free_work * * This function is invoked from a work item for a cgroup which is being - * destroyed after the percpu refcnts of all css's are guaranteed to be - * seen as killed on all CPUs, and performs the rest of destruction. This - * is the second step of destruction described in the comment above - * cgroup_destroy_locked(). + * destroyed after all css's are offlined and performs the rest of + * destruction. This is the second step of destruction described in the + * comment above cgroup_destroy_locked(). */ static void cgroup_destroy_css_killed(struct cgroup *cgrp) { struct cgroup *parent = cgrp->parent; struct dentry *d = cgrp->dentry; - struct cgroup_subsys *ss; lockdep_assert_held(&cgroup_mutex); - /* - * css_tryget() is guaranteed to fail now. Tell subsystems to - * initate destruction. - */ - for_each_root_subsys(cgrp->root, ss) - offline_css(cgroup_css(cgrp, ss->subsys_id)); - - /* - * Put the css refs from cgroup_destroy_locked(). Each css holds - * an extra reference to the cgroup's dentry and cgroup removal - * proceeds regardless of css refs. On the last put of each css, - * whenever that may be, the extra dentry ref is put so that dentry - * destruction happens only after all css's are released. - */ - for_each_root_subsys(cgrp->root, ss) - css_put(cgroup_css(cgrp, ss->subsys_id)); - /* delete this cgroup from parent->children */ list_del_rcu(&cgrp->sibling); -- cgit v1.2.3 From edae0c3358947f8be5ca99f762d89e0c38e1f5d5 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 13 Aug 2013 20:22:51 -0400 Subject: cgroup: factor out kill_css() Factor out css ref killing from cgroup_destroy_locked() into kill_css(). We're gonna add more to the path and the factored out function will eventually be called from other places too. While at it, replace open coded percpu_ref_get() with css_get() for consistency. This shouldn't cause any functional difference as the function is not used for root cgroups. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 58 ++++++++++++++++++++++++++++++++++----------------------- 1 file changed, 35 insertions(+), 23 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 3c4c4b01ffe5..7b7575f3119c 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4595,6 +4595,36 @@ static void css_killed_ref_fn(struct percpu_ref *ref) schedule_work(&css->destroy_work); } +/** + * kill_css - destroy a css + * @css: css to destroy + * + * This function initiates destruction of @css by putting its base + * reference. ->css_offline() will be invoked asynchronously once + * css_tryget() is guaranteed to fail and when the reference count reaches + * zero, @css will be released. + */ +static void kill_css(struct cgroup_subsys_state *css) +{ + /* + * Killing would put the base ref, but we need to keep it alive + * until after ->css_offline(). + */ + css_get(css); + + /* + * cgroup core guarantees that, by the time ->css_offline() is + * invoked, no new css reference will be given out via + * css_tryget(). We can't simply call percpu_ref_kill() and + * proceed to offlining css's because percpu_ref_kill() doesn't + * guarantee that the ref is seen as killed on all CPUs on return. + * + * Use percpu_ref_kill_and_confirm() to get notifications as each + * css is confirmed to be seen as killed on all CPUs. + */ + percpu_ref_kill_and_confirm(&css->refcnt, css_killed_ref_fn); +} + /** * cgroup_destroy_locked - the first stage of cgroup destruction * @cgrp: cgroup to be destroyed @@ -4641,30 +4671,12 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) return -EBUSY; /* - * Block new css_tryget() by killing css refcnts. cgroup core - * guarantees that, by the time ->css_offline() is invoked, no new - * css reference will be given out via css_tryget(). We can't - * simply call percpu_ref_kill() and proceed to offlining css's - * because percpu_ref_kill() doesn't guarantee that the ref is seen - * as killed on all CPUs on return. - * - * Use percpu_ref_kill_and_confirm() to get notifications as each - * css is confirmed to be seen as killed on all CPUs. - * cgroup_destroy_css_killed() will be invoked to perform the rest - * of destruction once the percpu refs of all css's are confirmed - * to be killed. + * Initiate massacre of all css's. cgroup_destroy_css_killed() + * will be invoked to perform the rest of destruction once the + * percpu refs of all css's are confirmed to be killed. */ - for_each_root_subsys(cgrp->root, ss) { - struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); - - /* - * Killing would put the base ref, but we need to keep it - * alive until after ->css_offline. - */ - percpu_ref_get(&css->refcnt); - - percpu_ref_kill_and_confirm(&css->refcnt, css_killed_ref_fn); - } + for_each_root_subsys(cgrp->root, ss) + kill_css(cgroup_css(cgrp, ss->subsys_id)); /* * Mark @cgrp dead. This prevents further task migration and child -- cgit v1.2.3 From 3c14f8b44fafaa60519440bea1591e495b928327 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 13 Aug 2013 20:22:51 -0400 Subject: cgroup: move subsys file removal to kill_css() With the planned unified hierarchy, individual css's will be created and destroyed dynamically across the lifetime of a cgroup. To enable such usages, css destruction is being decoupled from cgroup destruction. This patch moves subsys file removal from cgroup_destroy_locked() to kill_css(). While this changes the order of destruction operations, the changes shouldn't be noticeable to cgroup subsystems or userland. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 16 +++++++++------- 1 file changed, 9 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 7b7575f3119c..3137e38995b0 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4599,13 +4599,15 @@ static void css_killed_ref_fn(struct percpu_ref *ref) * kill_css - destroy a css * @css: css to destroy * - * This function initiates destruction of @css by putting its base - * reference. ->css_offline() will be invoked asynchronously once - * css_tryget() is guaranteed to fail and when the reference count reaches - * zero, @css will be released. + * This function initiates destruction of @css by removing cgroup interface + * files and putting its base reference. ->css_offline() will be invoked + * asynchronously once css_tryget() is guaranteed to fail and when the + * reference count reaches zero, @css will be released. */ static void kill_css(struct cgroup_subsys_state *css) { + cgroup_clear_dir(css->cgroup, 1 << css->ss->subsys_id); + /* * Killing would put the base ref, but we need to keep it alive * until after ->css_offline(). @@ -4703,10 +4705,10 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) cgroup_destroy_css_killed(cgrp); /* - * Clear and remove @cgrp directory. The removal puts the base ref - * but we aren't quite done with @cgrp yet, so hold onto it. + * Clear the base files and remove @cgrp directory. The removal + * puts the base ref but we aren't quite done with @cgrp yet, so + * hold onto it. */ - cgroup_clear_dir(cgrp, cgrp->root->subsys_mask); cgroup_addrm_files(cgrp, cgroup_base_files, false); dget(d); cgroup_d_remove_dir(d); -- cgit v1.2.3 From 0c21ead136a900c36f1ab74fd7d09a306dc31324 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 13 Aug 2013 20:22:51 -0400 Subject: cgroup: RCU protect each cgroup_subsys_state release With the planned unified hierarchy, individual css's will be created and destroyed dynamically across the lifetime of a cgroup. To enable such usages, css destruction is being decoupled from cgroup destruction. Most of the destruction path has been decoupled but the actual free of css still depends on cgroup free path. When all css refs are drained, css_release() kicks off css_free_work_fn() which puts the cgroup. When the cgroup refcnt reaches zero, cgroup_diput() is invoked which in turn schedules RCU free of the cgroup. After a grace period, all css's are freed along with the cgroup itself. This patch moves the RCU grace period and css freeing from cgroup release path to css release path. css_release(), instead of kicking off css_free_work_fn() directly, schedules RCU callback css_free_rcu_fn() which in turn kicks off css_free_work_fn() after a RCU grace period. css_free_work_fn() is updated to free the css directly. The five-way punting - percpu ref kill confirmation, a work item, percpu ref release, RCU grace period, and again a work item - is quite hairy but the work items are there only to provide process context and the actual sequence is kill confirm -> release -> RCU free, which isn't simple but not too crazy. This removes cgroup_css() usage after offline_css() allowing clearing cgroup->subsys[] from offline_css(), which makes it consistent with online_css() and brings it closer to proper lifetime management for individual css's. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 53 +++++++++++++++++++++++++++++++++++++---------------- 1 file changed, 37 insertions(+), 16 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 3137e38995b0..66d01078eebe 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -869,18 +869,8 @@ static struct cgroup_name *cgroup_alloc_name(struct dentry *dentry) static void cgroup_free_fn(struct work_struct *work) { struct cgroup *cgrp = container_of(work, struct cgroup, destroy_work); - struct cgroup_subsys *ss; mutex_lock(&cgroup_mutex); - /* - * Release the subsystem state objects. - */ - for_each_root_subsys(cgrp->root, ss) { - struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); - - ss->css_free(css); - } - cgrp->root->number_of_cgroups--; mutex_unlock(&cgroup_mutex); @@ -4281,32 +4271,62 @@ err: return ret; } +/* + * css destruction is four-stage process. + * + * 1. Destruction starts. Killing of the percpu_ref is initiated. + * Implemented in kill_css(). + * + * 2. When the percpu_ref is confirmed to be visible as killed on all CPUs + * and thus css_tryget() is guaranteed to fail, the css can be offlined + * by invoking offline_css(). After offlining, the base ref is put. + * Implemented in css_killed_work_fn(). + * + * 3. When the percpu_ref reaches zero, the only possible remaining + * accessors are inside RCU read sections. css_release() schedules the + * RCU callback. + * + * 4. After the grace period, the css can be freed. Implemented in + * css_free_work_fn(). + * + * It is actually hairier because both step 2 and 4 require process context + * and thus involve punting to css->destroy_work adding two additional + * steps to the already complex sequence. + */ static void css_free_work_fn(struct work_struct *work) { struct cgroup_subsys_state *css = container_of(work, struct cgroup_subsys_state, destroy_work); + struct cgroup *cgrp = css->cgroup; if (css->parent) css_put(css->parent); - cgroup_dput(css->cgroup); + css->ss->css_free(css); + cgroup_dput(cgrp); } -static void css_release(struct percpu_ref *ref) +static void css_free_rcu_fn(struct rcu_head *rcu_head) { struct cgroup_subsys_state *css = - container_of(ref, struct cgroup_subsys_state, refcnt); + container_of(rcu_head, struct cgroup_subsys_state, rcu_head); /* * css holds an extra ref to @cgrp->dentry which is put on the last - * css_put(). dput() requires process context, which css_put() may - * be called without. @css->destroy_work will be used to invoke - * dput() asynchronously from css_put(). + * css_put(). dput() requires process context which we don't have. */ INIT_WORK(&css->destroy_work, css_free_work_fn); schedule_work(&css->destroy_work); } +static void css_release(struct percpu_ref *ref) +{ + struct cgroup_subsys_state *css = + container_of(ref, struct cgroup_subsys_state, refcnt); + + call_rcu(&css->rcu_head, css_free_rcu_fn); +} + static void init_css(struct cgroup_subsys_state *css, struct cgroup_subsys *ss, struct cgroup *cgrp) { @@ -4356,6 +4376,7 @@ static void offline_css(struct cgroup_subsys_state *css) css->flags &= ~CSS_ONLINE; css->cgroup->nr_css--; + RCU_INIT_POINTER(css->cgroup->subsys[ss->subsys_id], css); } /* -- cgit v1.2.3 From ff58ac0d58d51bffe868b239ed8fce7c4a23c5a9 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Tue, 13 Aug 2013 09:17:33 +0800 Subject: cpuset: remove an unncessary forward declaration Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cpuset.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 72a0383f382f..95f4b25e1538 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -68,9 +68,6 @@ */ int number_of_cpusets __read_mostly; -/* Forward declare cgroup structures */ -struct cgroup_subsys cpuset_subsys; - /* See "Frequency meter" comments, below. */ struct fmeter { -- cgit v1.2.3 From dfa9771a7c4784bafd0673bc7abcee3813088b77 Mon Sep 17 00:00:00 2001 From: Michal Simek Date: Tue, 13 Aug 2013 16:00:53 -0700 Subject: microblaze: fix clone syscall Fix inadvertent breakage in the clone syscall ABI for Microblaze that was introduced in commit f3268edbe6fe ("microblaze: switch to generic fork/vfork/clone"). The Microblaze syscall ABI for clone takes the parent tid address in the 4th argument; the third argument slot is used for the stack size. The incorrectly-used CLONE_BACKWARDS type assigned parent tid to the 3rd slot. This commit restores the original ABI so that existing userspace libc code will work correctly. All kernel versions from v3.8-rc1 were affected. Signed-off-by: Michal Simek Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index 403d2bb8a968..e23bb19e2a3e 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1679,6 +1679,12 @@ SYSCALL_DEFINE5(clone, unsigned long, newsp, unsigned long, clone_flags, int __user *, parent_tidptr, int __user *, child_tidptr, int, tls_val) +#elif defined(CONFIG_CLONE_BACKWARDS3) +SYSCALL_DEFINE6(clone, unsigned long, clone_flags, unsigned long, newsp, + int, stack_size, + int __user *, parent_tidptr, + int __user *, child_tidptr, + int, tls_val) #else SYSCALL_DEFINE5(clone, unsigned long, clone_flags, unsigned long, newsp, int __user *, parent_tidptr, -- cgit v1.2.3 From 65f382fd0c8fa483713c0971de9f1dfb4cf1ad9c Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 11 Jul 2013 19:12:32 +0200 Subject: context_tracking: Ground setup for static key use Prepare for using a static key in the context tracking subsystem. This will help optimizing the off case on its many users: * user_enter, user_exit, exception_enter, exception_exit, guest_enter, guest_exit, vtime_*() Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman --- kernel/context_tracking.c | 23 +++++++++++++++++------ 1 file changed, 17 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/context_tracking.c b/kernel/context_tracking.c index 72bcb2570d3e..839d377d0da5 100644 --- a/kernel/context_tracking.c +++ b/kernel/context_tracking.c @@ -20,15 +20,16 @@ #include #include -DEFINE_PER_CPU(struct context_tracking, context_tracking) = { -#ifdef CONFIG_CONTEXT_TRACKING_FORCE - .active = true, -#endif -}; +struct static_key context_tracking_enabled = STATIC_KEY_INIT_FALSE; + +DEFINE_PER_CPU(struct context_tracking, context_tracking); void context_tracking_cpu_set(int cpu) { - per_cpu(context_tracking.active, cpu) = true; + if (!per_cpu(context_tracking.active, cpu)) { + per_cpu(context_tracking.active, cpu) = true; + static_key_slow_inc(&context_tracking_enabled); + } } /** @@ -202,3 +203,13 @@ void context_tracking_task_switch(struct task_struct *prev, clear_tsk_thread_flag(prev, TIF_NOHZ); set_tsk_thread_flag(next, TIF_NOHZ); } + +#ifdef CONFIG_CONTEXT_TRACKING_FORCE +void __init context_tracking_init(void) +{ + int cpu; + + for_each_possible_cpu(cpu) + context_tracking_cpu_set(cpu); +} +#endif -- cgit v1.2.3 From ad65782fba507d91a0a98f519b59e79cac1b474c Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 10 Jul 2013 02:44:35 +0200 Subject: context_tracking: Optimize main APIs off case with static key Optimize user and exception entry/exit APIs with static keys. This minimize the overhead for those who enable CONFIG_NO_HZ_FULL without always using it. Having no range passed to nohz_full= should result in the probes to be nopped (at least we hope so...). If this proves not be enough in the long term, we'll need to bring an exception slow path by re-routing the exception handlers. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman --- kernel/context_tracking.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/context_tracking.c b/kernel/context_tracking.c index 839d377d0da5..6e89e094c80e 100644 --- a/kernel/context_tracking.c +++ b/kernel/context_tracking.c @@ -33,15 +33,15 @@ void context_tracking_cpu_set(int cpu) } /** - * user_enter - Inform the context tracking that the CPU is going to - * enter userspace mode. + * context_tracking_user_enter - Inform the context tracking that the CPU is going to + * enter userspace mode. * * This function must be called right before we switch from the kernel * to userspace, when it's guaranteed the remaining kernel instructions * to execute won't use any RCU read side critical section because this * function sets RCU in extended quiescent state. */ -void user_enter(void) +void context_tracking_user_enter(void) { unsigned long flags; @@ -131,8 +131,8 @@ EXPORT_SYMBOL_GPL(preempt_schedule_context); #endif /* CONFIG_PREEMPT */ /** - * user_exit - Inform the context tracking that the CPU is - * exiting userspace mode and entering the kernel. + * context_tracking_user_exit - Inform the context tracking that the CPU is + * exiting userspace mode and entering the kernel. * * This function must be called after we entered the kernel from userspace * before any use of RCU read side critical section. This potentially include @@ -141,7 +141,7 @@ EXPORT_SYMBOL_GPL(preempt_schedule_context); * This call supports re-entrancy. This way it can be called from any exception * handler without needing to know if we came from userspace or not. */ -void user_exit(void) +void context_tracking_user_exit(void) { unsigned long flags; -- cgit v1.2.3 From 48d6a816a8bf36e2a197c322697323003bdc1cfe Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 10 Jul 2013 02:44:35 +0200 Subject: context_tracking: Optimize guest APIs off case with static key Optimize guest entry/exit APIs with static keys. This minimize the overhead for those who enable CONFIG_NO_HZ_FULL without always using it. Having no range passed to nohz_full= should result in the probes overhead to be minimized. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman --- kernel/context_tracking.c | 23 ++--------------------- kernel/sched/cputime.c | 2 ++ 2 files changed, 4 insertions(+), 21 deletions(-) (limited to 'kernel') diff --git a/kernel/context_tracking.c b/kernel/context_tracking.c index 6e89e094c80e..b6a186c4b886 100644 --- a/kernel/context_tracking.c +++ b/kernel/context_tracking.c @@ -21,8 +21,10 @@ #include struct static_key context_tracking_enabled = STATIC_KEY_INIT_FALSE; +EXPORT_SYMBOL_GPL(context_tracking_enabled); DEFINE_PER_CPU(struct context_tracking, context_tracking); +EXPORT_SYMBOL_GPL(context_tracking); void context_tracking_cpu_set(int cpu) { @@ -163,27 +165,6 @@ void context_tracking_user_exit(void) local_irq_restore(flags); } -#ifdef CONFIG_VIRT_CPU_ACCOUNTING_GEN -void guest_enter(void) -{ - if (vtime_accounting_enabled()) - vtime_guest_enter(current); - else - current->flags |= PF_VCPU; -} -EXPORT_SYMBOL_GPL(guest_enter); - -void guest_exit(void) -{ - if (vtime_accounting_enabled()) - vtime_guest_exit(current); - else - current->flags &= ~PF_VCPU; -} -EXPORT_SYMBOL_GPL(guest_exit); -#endif /* CONFIG_VIRT_CPU_ACCOUNTING_GEN */ - - /** * context_tracking_task_switch - context switch the syscall callbacks * @prev: the task that is being switched out diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c index 223a35efa0a6..bb6b29a3067c 100644 --- a/kernel/sched/cputime.c +++ b/kernel/sched/cputime.c @@ -724,6 +724,7 @@ void vtime_guest_enter(struct task_struct *tsk) current->flags |= PF_VCPU; write_sequnlock(&tsk->vtime_seqlock); } +EXPORT_SYMBOL_GPL(vtime_guest_enter); void vtime_guest_exit(struct task_struct *tsk) { @@ -732,6 +733,7 @@ void vtime_guest_exit(struct task_struct *tsk) current->flags &= ~PF_VCPU; write_sequnlock(&tsk->vtime_seqlock); } +EXPORT_SYMBOL_GPL(vtime_guest_exit); void vtime_account_idle(struct task_struct *tsk) { -- cgit v1.2.3 From 73d424f9af7b571276e6284617cb59726d47bf12 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 11 Jul 2013 19:42:13 +0200 Subject: context_tracking: Optimize context switch off case with static keys No need for syscall slowpath if no CPU is full dynticks, rather nop this in this case. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman --- kernel/context_tracking.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/context_tracking.c b/kernel/context_tracking.c index b6a186c4b886..c17822673c39 100644 --- a/kernel/context_tracking.c +++ b/kernel/context_tracking.c @@ -166,7 +166,7 @@ void context_tracking_user_exit(void) } /** - * context_tracking_task_switch - context switch the syscall callbacks + * __context_tracking_task_switch - context switch the syscall callbacks * @prev: the task that is being switched out * @next: the task that is being switched in * @@ -178,8 +178,8 @@ void context_tracking_user_exit(void) * migrate to some CPU that doesn't do the context tracking. As such the TIF * flag may not be desired there. */ -void context_tracking_task_switch(struct task_struct *prev, - struct task_struct *next) +void __context_tracking_task_switch(struct task_struct *prev, + struct task_struct *next) { clear_tsk_thread_flag(prev, TIF_NOHZ); set_tsk_thread_flag(next, TIF_NOHZ); -- cgit v1.2.3 From 1b6a259aa5ab16d8b215bfc19ff7c9ffa8858f10 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 11 Jul 2013 20:27:43 +0200 Subject: context_tracking: User/kernel broundary cross trace events This can be useful to track all kernel/user round trips. And it's also helpful to debug the context tracking subsystem. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman --- kernel/context_tracking.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'kernel') diff --git a/kernel/context_tracking.c b/kernel/context_tracking.c index c17822673c39..247091bf0587 100644 --- a/kernel/context_tracking.c +++ b/kernel/context_tracking.c @@ -20,6 +20,9 @@ #include #include +#define CREATE_TRACE_POINTS +#include + struct static_key context_tracking_enabled = STATIC_KEY_INIT_FALSE; EXPORT_SYMBOL_GPL(context_tracking_enabled); @@ -64,6 +67,7 @@ void context_tracking_user_enter(void) local_irq_save(flags); if ( __this_cpu_read(context_tracking.state) != IN_USER) { if (__this_cpu_read(context_tracking.active)) { + trace_user_enter(0); /* * At this stage, only low level arch entry code remains and * then we'll run in userspace. We can assume there won't be @@ -159,6 +163,7 @@ void context_tracking_user_exit(void) */ rcu_user_exit(); vtime_user_exit(current); + trace_user_exit(0); } __this_cpu_write(context_tracking.state, IN_KERNEL); } -- cgit v1.2.3 From 7621d1f8bcb418e7a7ac583e89e38ec01b7ed182 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 13 Jul 2013 17:07:35 +0200 Subject: vtime: Remove a few unneeded generic vtime state checks Some generic vtime APIs check if the vtime accounting is enabled on the local CPU before doing their work. Some of these are not needed because all their callers already take care of that. Let's remove the checks on these. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman --- kernel/sched/cputime.c | 13 +------------ 1 file changed, 1 insertion(+), 12 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c index bb6b29a3067c..5f273b477764 100644 --- a/kernel/sched/cputime.c +++ b/kernel/sched/cputime.c @@ -664,9 +664,6 @@ static void __vtime_account_system(struct task_struct *tsk) void vtime_account_system(struct task_struct *tsk) { - if (!vtime_accounting_enabled()) - return; - write_seqlock(&tsk->vtime_seqlock); __vtime_account_system(tsk); write_sequnlock(&tsk->vtime_seqlock); @@ -686,12 +683,7 @@ void vtime_account_irq_exit(struct task_struct *tsk) void vtime_account_user(struct task_struct *tsk) { - cputime_t delta_cpu; - - if (!vtime_accounting_enabled()) - return; - - delta_cpu = get_vtime_delta(tsk); + cputime_t delta_cpu = get_vtime_delta(tsk); write_seqlock(&tsk->vtime_seqlock); tsk->vtime_snap_whence = VTIME_SYS; @@ -701,9 +693,6 @@ void vtime_account_user(struct task_struct *tsk) void vtime_user_enter(struct task_struct *tsk) { - if (!vtime_accounting_enabled()) - return; - write_seqlock(&tsk->vtime_seqlock); tsk->vtime_snap_whence = VTIME_USER; __vtime_account_system(tsk); -- cgit v1.2.3 From 54461562c90e0ac104764c5a9de637fd9151a1c1 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 13 Jul 2013 17:10:18 +0200 Subject: vtime: Fix racy cputime delta update get_vtime_delta() must be called under the task vtime_seqlock with the code that does the cputime accounting flush. Otherwise the cputime reader can be fooled and run into a race where it sees the snapshot update but misses the cputime flush. As a result it can report a cputime that is way too short. Fix vtime_account_user() that wasn't complying to that rule. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman --- kernel/sched/cputime.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c index 5f273b477764..b62d5c027c7e 100644 --- a/kernel/sched/cputime.c +++ b/kernel/sched/cputime.c @@ -683,9 +683,10 @@ void vtime_account_irq_exit(struct task_struct *tsk) void vtime_account_user(struct task_struct *tsk) { - cputime_t delta_cpu = get_vtime_delta(tsk); + cputime_t delta_cpu; write_seqlock(&tsk->vtime_seqlock); + delta_cpu = get_vtime_delta(tsk); tsk->vtime_snap_whence = VTIME_SYS; account_user_time(tsk, delta_cpu, cputime_to_scaled(delta_cpu)); write_sequnlock(&tsk->vtime_seqlock); -- cgit v1.2.3 From b04934061330a4a449cfce703c97d887c3e11cd7 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 12 Jul 2013 03:10:15 +0200 Subject: vtime: Optimize full dynticks accounting off case with static keys If no CPU is in the full dynticks range, we can avoid the full dynticks cputime accounting through generic vtime along with its overhead and use the traditional tick based accounting instead. Let's do this and nope the off case with static keys. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman --- kernel/sched/cputime.c | 22 ++++------------------ 1 file changed, 4 insertions(+), 18 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c index b62d5c027c7e..0831b06aab97 100644 --- a/kernel/sched/cputime.c +++ b/kernel/sched/cputime.c @@ -378,11 +378,8 @@ static inline void irqtime_account_process_tick(struct task_struct *p, int user_ #ifdef CONFIG_VIRT_CPU_ACCOUNTING #ifndef __ARCH_HAS_VTIME_TASK_SWITCH -void vtime_task_switch(struct task_struct *prev) +void vtime_common_task_switch(struct task_struct *prev) { - if (!vtime_accounting_enabled()) - return; - if (is_idle_task(prev)) vtime_account_idle(prev); else @@ -404,11 +401,8 @@ void vtime_task_switch(struct task_struct *prev) * vtime_account(). */ #ifndef __ARCH_HAS_VTIME_ACCOUNT -void vtime_account_irq_enter(struct task_struct *tsk) +void vtime_common_account_irq_enter(struct task_struct *tsk) { - if (!vtime_accounting_enabled()) - return; - if (!in_interrupt()) { /* * If we interrupted user, context_tracking_in_user() @@ -428,7 +422,7 @@ void vtime_account_irq_enter(struct task_struct *tsk) } vtime_account_system(tsk); } -EXPORT_SYMBOL_GPL(vtime_account_irq_enter); +EXPORT_SYMBOL_GPL(vtime_common_account_irq_enter); #endif /* __ARCH_HAS_VTIME_ACCOUNT */ #endif /* CONFIG_VIRT_CPU_ACCOUNTING */ @@ -669,11 +663,8 @@ void vtime_account_system(struct task_struct *tsk) write_sequnlock(&tsk->vtime_seqlock); } -void vtime_account_irq_exit(struct task_struct *tsk) +void vtime_gen_account_irq_exit(struct task_struct *tsk) { - if (!vtime_accounting_enabled()) - return; - write_seqlock(&tsk->vtime_seqlock); if (context_tracking_in_user()) tsk->vtime_snap_whence = VTIME_USER; @@ -732,11 +723,6 @@ void vtime_account_idle(struct task_struct *tsk) account_idle_time(delta_cpu); } -bool vtime_accounting_enabled(void) -{ - return context_tracking_active(); -} - void arch_vtime_task_switch(struct task_struct *prev) { write_seqlock(&prev->vtime_seqlock); -- cgit v1.2.3 From b854fafa4e06c50a92e00b39d75ee62083d986d6 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 13 Jul 2013 17:24:20 +0200 Subject: vtime: Always scale generic vtime accounting results The cputime accounting in full dynticks can be a subtle mixup of CPUs using tick based accounting and others using generic vtime. As long as the tick can have a share on producing these stats, we want to scale the result against CFS precise accounting as the tick can miss some task hiding between the periodic interrupt. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman --- kernel/sched/cputime.c | 6 ------ 1 file changed, 6 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c index 0831b06aab97..e9e742ed7280 100644 --- a/kernel/sched/cputime.c +++ b/kernel/sched/cputime.c @@ -553,12 +553,6 @@ static void cputime_adjust(struct task_cputime *curr, { cputime_t rtime, stime, utime, total; - if (vtime_accounting_enabled()) { - *ut = curr->utime; - *st = curr->stime; - return; - } - stime = curr->stime; total = stime + curr->utime; -- cgit v1.2.3 From af2350bd12096dfd04e1090b90bfecea1f75f84e Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Mon, 15 Jul 2013 16:35:55 +0200 Subject: vtime: Always debug check snapshot source _before_ updating it The vtime delta update performed by get_vtime_delta() always check that the source of the snapshot is valid. Meanhile the snapshot updaters that rely on get_vtime_delta() also set the new snapshot origin. But some of them do this right before the call to get_vtime_delta(), making its debug check useless. This is easily fixable by moving the snapshot origin update after the call to get_vtime_delta(). The order doesn't matter there. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman --- kernel/sched/cputime.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c index e9e742ed7280..c1d7493825ae 100644 --- a/kernel/sched/cputime.c +++ b/kernel/sched/cputime.c @@ -660,9 +660,9 @@ void vtime_account_system(struct task_struct *tsk) void vtime_gen_account_irq_exit(struct task_struct *tsk) { write_seqlock(&tsk->vtime_seqlock); + __vtime_account_system(tsk); if (context_tracking_in_user()) tsk->vtime_snap_whence = VTIME_USER; - __vtime_account_system(tsk); write_sequnlock(&tsk->vtime_seqlock); } @@ -680,8 +680,8 @@ void vtime_account_user(struct task_struct *tsk) void vtime_user_enter(struct task_struct *tsk) { write_seqlock(&tsk->vtime_seqlock); - tsk->vtime_snap_whence = VTIME_USER; __vtime_account_system(tsk); + tsk->vtime_snap_whence = VTIME_USER; write_sequnlock(&tsk->vtime_seqlock); } -- cgit v1.2.3 From 73867dcd0792ad14fb31bfe73d09d9a4576f7fc2 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 24 Jul 2013 23:31:00 +0200 Subject: nohz: Rename a few state variables Rename the full dynticks's cpumask and cpumask state variables to some more exportable names. These will be used later from global headers to optimize the main full dynticks APIs in conjunction with static keys. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman --- kernel/time/tick-sched.c | 44 ++++++++++++++++++++++---------------------- 1 file changed, 22 insertions(+), 22 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 91a2528b5f44..b28dee43e644 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -149,8 +149,8 @@ static void tick_sched_handle(struct tick_sched *ts, struct pt_regs *regs) } #ifdef CONFIG_NO_HZ_FULL -static cpumask_var_t nohz_full_mask; -bool have_nohz_full_mask; +static cpumask_var_t tick_nohz_full_mask; +bool tick_nohz_full_running; static bool can_stop_full_tick(void) { @@ -183,7 +183,7 @@ static bool can_stop_full_tick(void) * Don't allow the user to think they can get * full NO_HZ with this machine. */ - WARN_ONCE(have_nohz_full_mask, + WARN_ONCE(tick_nohz_full_running, "NO_HZ FULL will not work with unstable sched clock"); return false; } @@ -240,11 +240,11 @@ static void nohz_full_kick_ipi(void *info) */ void tick_nohz_full_kick_all(void) { - if (!have_nohz_full_mask) + if (!tick_nohz_full_running) return; preempt_disable(); - smp_call_function_many(nohz_full_mask, + smp_call_function_many(tick_nohz_full_mask, nohz_full_kick_ipi, NULL, false); preempt_enable(); } @@ -272,10 +272,10 @@ out: int tick_nohz_full_cpu(int cpu) { - if (!have_nohz_full_mask) + if (!tick_nohz_full_running) return 0; - return cpumask_test_cpu(cpu, nohz_full_mask); + return cpumask_test_cpu(cpu, tick_nohz_full_mask); } /* Parse the boot-time nohz CPU list from the kernel parameters. */ @@ -283,18 +283,18 @@ static int __init tick_nohz_full_setup(char *str) { int cpu; - alloc_bootmem_cpumask_var(&nohz_full_mask); - if (cpulist_parse(str, nohz_full_mask) < 0) { + alloc_bootmem_cpumask_var(&tick_nohz_full_mask); + if (cpulist_parse(str, tick_nohz_full_mask) < 0) { pr_warning("NOHZ: Incorrect nohz_full cpumask\n"); return 1; } cpu = smp_processor_id(); - if (cpumask_test_cpu(cpu, nohz_full_mask)) { + if (cpumask_test_cpu(cpu, tick_nohz_full_mask)) { pr_warning("NO_HZ: Clearing %d from nohz_full range for timekeeping\n", cpu); - cpumask_clear_cpu(cpu, nohz_full_mask); + cpumask_clear_cpu(cpu, tick_nohz_full_mask); } - have_nohz_full_mask = true; + tick_nohz_full_running = true; return 1; } @@ -312,7 +312,7 @@ static int tick_nohz_cpu_down_callback(struct notifier_block *nfb, * If we handle the timekeeping duty for full dynticks CPUs, * we can't safely shutdown that CPU. */ - if (have_nohz_full_mask && tick_do_timer_cpu == cpu) + if (tick_nohz_full_running && tick_do_timer_cpu == cpu) return NOTIFY_BAD; break; } @@ -331,14 +331,14 @@ static int tick_nohz_init_all(void) int err = -1; #ifdef CONFIG_NO_HZ_FULL_ALL - if (!alloc_cpumask_var(&nohz_full_mask, GFP_KERNEL)) { + if (!alloc_cpumask_var(&tick_nohz_full_mask, GFP_KERNEL)) { pr_err("NO_HZ: Can't allocate full dynticks cpumask\n"); return err; } err = 0; - cpumask_setall(nohz_full_mask); - cpumask_clear_cpu(smp_processor_id(), nohz_full_mask); - have_nohz_full_mask = true; + cpumask_setall(tick_nohz_full_mask); + cpumask_clear_cpu(smp_processor_id(), tick_nohz_full_mask); + tick_nohz_full_running = true; #endif return err; } @@ -347,20 +347,20 @@ void __init tick_nohz_init(void) { int cpu; - if (!have_nohz_full_mask) { + if (!tick_nohz_full_running) { if (tick_nohz_init_all() < 0) return; } - for_each_cpu(cpu, nohz_full_mask) + for_each_cpu(cpu, tick_nohz_full_mask) context_tracking_cpu_set(cpu); cpu_notifier(tick_nohz_cpu_down_callback, 0); - cpulist_scnprintf(nohz_full_buf, sizeof(nohz_full_buf), nohz_full_mask); + cpulist_scnprintf(nohz_full_buf, sizeof(nohz_full_buf), tick_nohz_full_mask); pr_info("NO_HZ: Full dynticks CPUs: %s.\n", nohz_full_buf); } #else -#define have_nohz_full_mask (0) +#define tick_nohz_full_running (0) #endif /* @@ -738,7 +738,7 @@ static bool can_stop_idle_tick(int cpu, struct tick_sched *ts) return false; } - if (have_nohz_full_mask) { + if (tick_nohz_full_running) { /* * Keep the tick alive to guarantee timekeeping progression * if there are full dynticks CPUs around -- cgit v1.2.3 From 460775df4680b4593d8449bc171008578625a850 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 24 Jul 2013 23:52:27 +0200 Subject: nohz: Optimize full dynticks state checks with static keys These APIs are frequenctly accessed and priority is given to optimize the full dynticks off-case in order to let distros enable this feature without suffering from significant performance regressions. Let's inline these APIs and optimize them with static keys. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman --- kernel/time/tick-sched.c | 14 ++------------ 1 file changed, 2 insertions(+), 12 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index b28dee43e644..0b7887389bd2 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -149,7 +149,7 @@ static void tick_sched_handle(struct tick_sched *ts, struct pt_regs *regs) } #ifdef CONFIG_NO_HZ_FULL -static cpumask_var_t tick_nohz_full_mask; +cpumask_var_t tick_nohz_full_mask; bool tick_nohz_full_running; static bool can_stop_full_tick(void) @@ -270,14 +270,6 @@ out: local_irq_restore(flags); } -int tick_nohz_full_cpu(int cpu) -{ - if (!tick_nohz_full_running) - return 0; - - return cpumask_test_cpu(cpu, tick_nohz_full_mask); -} - /* Parse the boot-time nohz CPU list from the kernel parameters. */ static int __init tick_nohz_full_setup(char *str) { @@ -359,8 +351,6 @@ void __init tick_nohz_init(void) cpulist_scnprintf(nohz_full_buf, sizeof(nohz_full_buf), tick_nohz_full_mask); pr_info("NO_HZ: Full dynticks CPUs: %s.\n", nohz_full_buf); } -#else -#define tick_nohz_full_running (0) #endif /* @@ -738,7 +728,7 @@ static bool can_stop_idle_tick(int cpu, struct tick_sched *ts) return false; } - if (tick_nohz_full_running) { + if (tick_nohz_full_enabled()) { /* * Keep the tick alive to guarantee timekeeping progression * if there are full dynticks CPUs around -- cgit v1.2.3 From d13508f9440e46dccac6a2dd48d51a73b2207482 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 24 Jul 2013 23:52:27 +0200 Subject: nohz: Optimize full dynticks's sched hooks with static keys Scheduler IPIs and task context switches are serious fast path. Let's try to hide as much as we can the impact of full dynticks APIs' off case that are called on these sites through the use of static keys. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman --- kernel/time/tick-sched.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 0b7887389bd2..0ff6ae710161 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -198,7 +198,7 @@ static void tick_nohz_restart_sched_tick(struct tick_sched *ts, ktime_t now); * Re-evaluate the need for the tick on the current CPU * and restart it if necessary. */ -void tick_nohz_full_check(void) +void __tick_nohz_full_check(void) { struct tick_sched *ts = &__get_cpu_var(tick_cpu_sched); @@ -212,7 +212,7 @@ void tick_nohz_full_check(void) static void nohz_full_kick_work_func(struct irq_work *work) { - tick_nohz_full_check(); + __tick_nohz_full_check(); } static DEFINE_PER_CPU(struct irq_work, nohz_full_kick_work) = { @@ -231,7 +231,7 @@ void tick_nohz_full_kick(void) static void nohz_full_kick_ipi(void *info) { - tick_nohz_full_check(); + __tick_nohz_full_check(); } /* @@ -254,7 +254,7 @@ void tick_nohz_full_kick_all(void) * It might need the tick due to per task/process properties: * perf events, posix cpu timers, ... */ -void tick_nohz_task_switch(struct task_struct *tsk) +void __tick_nohz_task_switch(struct task_struct *tsk) { unsigned long flags; -- cgit v1.2.3 From fd5e2aa8653665ae1cc60f7aca1069abdbcad3f6 Mon Sep 17 00:00:00 2001 From: Dwight Engen Date: Thu, 15 Aug 2013 14:08:00 -0400 Subject: xfs: ioctl check for capabilities in the current user namespace Use inode_capable() to check if SUID|SGID bits should be cleared to match similar check in inode_change_ok(). The check for CAP_LINUX_IMMUTABLE was not modified since all other file systems also check against init_user_ns rather than current_user_ns. Only allow changing of projid from init_user_ns. Reviewed-by: Dave Chinner Reviewed-by: Gao feng Signed-off-by: Dwight Engen Signed-off-by: Ben Myers --- kernel/capability.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/capability.c b/kernel/capability.c index f6c2ce5701e1..a4b67446dc87 100644 --- a/kernel/capability.c +++ b/kernel/capability.c @@ -464,3 +464,4 @@ bool inode_capable(const struct inode *inode, int cap) return ns_capable(ns, cap) && kuid_has_mapping(ns, inode->i_uid); } +EXPORT_SYMBOL(inode_capable); -- cgit v1.2.3 From 930913a31289202d232186b82854b26d7fb7cf4d Mon Sep 17 00:00:00 2001 From: Li Zhong Date: Fri, 16 Aug 2013 17:57:14 +0800 Subject: cgroup: use css_get() in cgroup_create() to check CSS_ROOT It seems that the root css doesn't have refcnt allocated(not needed?), and would cause the booting error attached. This patch tries to use css_get() to not increase the refcnt if parent is root. BUG: unable to handle kernel NULL pointer dereference at (null) IP: [] cgroup_mkdir+0x37c/0x740 PGD 0 Oops: 0002 [#1] Modules linked in: CPU: 0 PID: 1 Comm: systemd Not tainted 3.11.0-rc5-next-20130815+ #1 Hardware name: Bochs Bochs, BIOS Bochs 01/01/2007 task: ffff88007f868000 ti: ffff88007f864000 task.ti: ffff88007f864000 RIP: 0010:[] [] cgroup_mkdir+0x37c/0x740 RSP: 0018:ffff88007f865df8 EFLAGS: 00010246 RAX: 0000000000000000 RBX: ffffffff81a46ee0 RCX: 0000000000000001 RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffffffff81a415c0 RBP: ffff88007f865ec8 R08: 0000000000000001 R09: 0000000000000000 R10: ffff88007ce6d060 R11: 0000000000000000 R12: ffff88007ce6d000 R13: ffff88007ce6d060 R14: ffffffff81a46d80 R15: ffff88007c6e8018 FS: 00007f13dbf6f840(0000) GS:ffffffff81a23000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000000 CR3: 000000007b7e5000 CR4: 00000000000006b0 Stack: ffffffff810b380d 0000000000000002 ffff88007f865e18 ffffffff81167069 ffff88007f865ed8 ffffffff8116a3f5 ffff880037454400 ffff88007c6e8018 ffff88007c6e8028 ffff88007c6e8328 ffff88007c6e8000 ffff88007ce6d000 Call Trace: [] ? cgroup_mkdir+0x3bd/0x740 [] ? lookup_hash+0x19/0x20 [] ? kern_path_create+0x95/0x170 [] vfs_mkdir+0x9e/0xf0 [] SyS_mkdirat+0x60/0xe0 [] SyS_mkdir+0x19/0x20 [] tracesys+0xcf/0xd4 Code: ad 70 ff ff ff 48 89 9d 60 ff ff ff 4d 89 d5 4c 8b bd 68 ff ff ff 4c 8b 65 88 eb 50 0f 1f 00 48 8b 43 18 a8 03 0f 85 6c 03 00 00 00 e8 1d 0a fb ff 85 c0 74 0d 80 3d f0 45 a1 00 00 0f 84 4c RIP [] cgroup_mkdir+0x37c/0x740 RSP CR2: 0000000000000000 ---[ end trace a4b14b49bc46fd60 ]--- Signed-off-by: Li Zhong Acked-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cgroup.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 66d01078eebe..b69b572131e5 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4494,7 +4494,7 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, struct cgroup_subsys_state *css = css_ar[ss->subsys_id]; dget(dentry); - percpu_ref_get(&css->parent->refcnt); + css_get(css->parent); } /* hold a ref to the parent's dentry */ -- cgit v1.2.3 From c8d2d47a9cbb4222ae4e993aa0e3703430c8193c Mon Sep 17 00:00:00 2001 From: Xiaotian Feng Date: Tue, 6 Aug 2013 20:06:42 +0800 Subject: cpumask: Fix cpumask leak in partition_sched_domains() If doms_new is NULL, partition_sched_domains() will reset ndoms_cur to 0, and free old sched domains with free_sched_domains(doms_cur, ndoms_cur). As ndoms_cur is 0, the cpumask will not be freed. Signed-off-by: Xiaotian Feng Cc: Rusty Russell Cc: linux-kernel@vger.kernel.org Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1375790802-11857-1-git-send-email-xtfeng@gmail.com Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index b7415cfdd7de..cf8f100433e0 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -6184,8 +6184,9 @@ match1: ; } + n = ndoms_cur; if (doms_new == NULL) { - ndoms_cur = 0; + n = 0; doms_new = &fallback_doms; cpumask_andnot(doms_new[0], cpu_active_mask, cpu_isolated_map); WARN_ON_ONCE(dattr_new); @@ -6193,7 +6194,7 @@ match1: /* Build new domains */ for (i = 0; i < ndoms_new; i++) { - for (j = 0; j < ndoms_cur && !new_topology; j++) { + for (j = 0; j < n && !new_topology; j++) { if (cpumask_equal(doms_new[i], doms_cur[j]) && dattrs_equal(dattr_new, i, dattr_cur, j)) goto match2; -- cgit v1.2.3 From a4f61cc03e443647211a5ae0ab8f8cda2e9e1043 Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Wed, 7 Aug 2013 15:38:24 +0000 Subject: sched/cputime: Use this_cpu_add() in task_group_account_field() Use of a this_cpu() operation reduces the number of instructions used for accounting (account_user_time()) and frees up some registers. This is in the scheduler tick hotpath. Signed-off-by: Christoph Lameter Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/00000140596dd165-338ff7f5-893b-4fec-b251-aaac5557239e-000000@email.amazonses.com Signed-off-by: Ingo Molnar --- kernel/sched/cputime.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c index a7959e05a9d5..e89ccefef278 100644 --- a/kernel/sched/cputime.c +++ b/kernel/sched/cputime.c @@ -121,7 +121,7 @@ static inline void task_group_account_field(struct task_struct *p, int index, * is the only cgroup, then nothing else should be necessary. * */ - __get_cpu_var(kernel_cpustat).cpustat[index] += tmp; + __this_cpu_add(kernel_cpustat.cpustat[index], tmp); cpuacct_account_field(p, index, tmp); } -- cgit v1.2.3 From c2e7fcf53c3cb02b4ada1c66a9bc8a4d97d58aba Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 2 Aug 2013 18:29:56 +0200 Subject: nohz: Include local CPU in full dynticks global kick tick_nohz_full_kick_all() is useful to notify all full dynticks CPUs that there is a system state change to checkout before re-evaluating the need for the tick. Unfortunately this is implemented using smp_call_function_many() that ignores the local CPU. This CPU also needs to re-evaluate the tick. on_each_cpu_mask() is not useful either because we don't want to re-evaluate the tick state in place but asynchronously from an IPI to avoid messing up with any random locking scenario. So lets call tick_nohz_full_kick() from tick_nohz_full_kick_all() so that the usual irq work takes care of it. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul E. McKenney Cc: Borislav Petkov Cc: Li Zhong Cc: Mike Galbraith Cc: Kevin Hilman Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1375460996-16329-4-git-send-email-fweisbec@gmail.com Signed-off-by: Ingo Molnar --- kernel/time/tick-sched.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index adea6fc3ba2a..3612fc77f834 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -246,6 +246,7 @@ void tick_nohz_full_kick_all(void) preempt_disable(); smp_call_function_many(tick_nohz_full_mask, nohz_full_kick_ipi, NULL, false); + tick_nohz_full_kick(); preempt_enable(); } -- cgit v1.2.3 From fc3b86d673e41ac66b4ba5b75a90c2fcafb90089 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 2 Aug 2013 18:29:54 +0200 Subject: perf: Roll back callchain buffer refcount under the callchain mutex When we fail to allocate the callchain buffers, we roll back the refcount we did and return from get_callchain_buffers(). However we take the refcount and allocate under the callchain lock but the rollback is done outside the lock. As a result, while we roll back, some concurrent callchain user may call get_callchain_buffers(), see the non-zero refcount and give up because the buffers are NULL without itself retrying the allocation. The consequences aren't that bad but that behaviour looks weird enough and it's better to give their chances to the following callchain users where we failed. Reported-by: Jiri Olsa Signed-off-by: Frederic Weisbecker Acked-by: Jiri Olsa Cc: Namhyung Kim Cc: Arnaldo Carvalho de Melo Cc: Stephane Eranian Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1375460996-16329-2-git-send-email-fweisbec@gmail.com Signed-off-by: Ingo Molnar --- kernel/events/callchain.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/events/callchain.c b/kernel/events/callchain.c index 76a8bc5f6265..97b67df8fbfe 100644 --- a/kernel/events/callchain.c +++ b/kernel/events/callchain.c @@ -116,10 +116,11 @@ int get_callchain_buffers(void) err = alloc_callchain_buffers(); exit: - mutex_unlock(&callchain_mutex); if (err) atomic_dec(&nr_callchain_events); + mutex_unlock(&callchain_mutex); + return err; } -- cgit v1.2.3 From 948b26b6ddd08a57cb95ebb0dc96fde2edd5c383 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 2 Aug 2013 18:29:55 +0200 Subject: perf: Account freq events globally Freq events may not always be affine to a particular CPU. As such, account_event_cpu() may crash if we account per cpu a freq event that has event->cpu == -1. To solve this, lets account freq events globally. In practice this doesn't change much the picture because perf tools create per-task perf events with one event per CPU by default. Profiling a single CPU is usually a corner case so there is no much point in optimizing things that way. Reported-by: Jiri Olsa Suggested-by: Peter Zijlstra Signed-off-by: Frederic Weisbecker Tested-by: Jiri Olsa Cc: Namhyung Kim Cc: Arnaldo Carvalho de Melo Cc: Stephane Eranian Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1375460996-16329-3-git-send-email-fweisbec@gmail.com Signed-off-by: Ingo Molnar --- kernel/events/core.c | 19 ++++++++----------- 1 file changed, 8 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index e82e70025d42..2e675e830976 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -141,11 +141,11 @@ enum event_type_t { struct static_key_deferred perf_sched_events __read_mostly; static DEFINE_PER_CPU(atomic_t, perf_cgroup_events); static DEFINE_PER_CPU(atomic_t, perf_branch_stack_events); -static DEFINE_PER_CPU(atomic_t, perf_freq_events); static atomic_t nr_mmap_events __read_mostly; static atomic_t nr_comm_events __read_mostly; static atomic_t nr_task_events __read_mostly; +static atomic_t nr_freq_events __read_mostly; static LIST_HEAD(pmus); static DEFINE_MUTEX(pmus_lock); @@ -1871,9 +1871,6 @@ static int __perf_install_in_context(void *info) perf_pmu_enable(cpuctx->ctx.pmu); perf_ctx_unlock(cpuctx, task_ctx); - if (atomic_read(&__get_cpu_var(perf_freq_events))) - tick_nohz_full_kick(); - return 0; } @@ -2811,7 +2808,7 @@ done: #ifdef CONFIG_NO_HZ_FULL bool perf_event_can_stop_tick(void) { - if (atomic_read(&__get_cpu_var(perf_freq_events)) || + if (atomic_read(&nr_freq_events) || __this_cpu_read(perf_throttled_count)) return false; else @@ -3140,9 +3137,6 @@ static void unaccount_event_cpu(struct perf_event *event, int cpu) } if (is_cgroup_event(event)) atomic_dec(&per_cpu(perf_cgroup_events, cpu)); - - if (event->attr.freq) - atomic_dec(&per_cpu(perf_freq_events, cpu)); } static void unaccount_event(struct perf_event *event) @@ -3158,6 +3152,8 @@ static void unaccount_event(struct perf_event *event) atomic_dec(&nr_comm_events); if (event->attr.task) atomic_dec(&nr_task_events); + if (event->attr.freq) + atomic_dec(&nr_freq_events); if (is_cgroup_event(event)) static_key_slow_dec_deferred(&perf_sched_events); if (has_branch_stack(event)) @@ -6489,9 +6485,6 @@ static void account_event_cpu(struct perf_event *event, int cpu) } if (is_cgroup_event(event)) atomic_inc(&per_cpu(perf_cgroup_events, cpu)); - - if (event->attr.freq) - atomic_inc(&per_cpu(perf_freq_events, cpu)); } static void account_event(struct perf_event *event) @@ -6507,6 +6500,10 @@ static void account_event(struct perf_event *event) atomic_inc(&nr_comm_events); if (event->attr.task) atomic_inc(&nr_task_events); + if (event->attr.freq) { + if (atomic_inc_return(&nr_freq_events) == 1) + tick_nohz_full_kick_all(); + } if (has_branch_stack(event)) static_key_slow_inc(&perf_sched_events.key); if (is_cgroup_event(event)) -- cgit v1.2.3 From 5ec4c599a52362896c3e7c6a31ba6145dca9c6f5 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 2 Aug 2013 21:16:30 +0200 Subject: perf: Do not compute time values unnecessarily We should not be calling calc_timer_values() for events that do not actually have an mmap()'ed userpage. Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/20130802191630.GT27162@twins.programming.kicks-ass.net Signed-off-by: Ingo Molnar --- kernel/events/core.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 2e675e830976..928fae7ca8c7 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -3670,6 +3670,10 @@ void perf_event_update_userpage(struct perf_event *event) u64 enabled, running, now; rcu_read_lock(); + rb = rcu_dereference(event->rb); + if (!rb) + goto unlock; + /* * compute total_time_enabled, total_time_running * based on snapshot values taken when the event @@ -3680,12 +3684,8 @@ void perf_event_update_userpage(struct perf_event *event) * NMI context */ calc_timer_values(event, &now, &enabled, &running); - rb = rcu_dereference(event->rb); - if (!rb) - goto unlock; userpg = rb->user_page; - /* * Disable preemption so as to not let the corresponding user-space * spin too long if we get preempted. -- cgit v1.2.3 From d1d74d14e98a6be740a6f12456c7d9ad47be9c9c Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Mon, 22 Apr 2013 00:12:42 +0200 Subject: rcu: Expedite grace periods during suspend/resume MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit CONFIG_RCU_FAST_NO_HZ can increase grace-period durations by up to a factor of four, which can result in long suspend and resume times. Thus, this commit temporarily switches to expedited grace periods when suspending the box and return to normal settings when resuming. Similar logic is applied to hibernation. Because expedited grace periods are of dubious benefit on very large systems, so this commit restricts their automated use during suspend and resume to systems of 256 or fewer CPUs. (Some day a number of Linux-kernel facilities, including RCU's expedited grace periods, will be more scalable, but I need to see bug reports first.) [ paulmck: This also papers over an audio/irq bug, but hopefully that will be fixed soon. ] Signed-off-by: Borislav Petkov Signed-off-by: Bjørn Mork Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutree.c | 21 +++++++++++++++++++++ 1 file changed, 21 insertions(+) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 338f1d1c1c66..a7bf517b0482 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -54,6 +54,7 @@ #include #include #include +#include #include "rcutree.h" #include @@ -3032,6 +3033,25 @@ static int rcu_cpu_notify(struct notifier_block *self, return NOTIFY_OK; } +static int rcu_pm_notify(struct notifier_block *self, + unsigned long action, void *hcpu) +{ + switch (action) { + case PM_HIBERNATION_PREPARE: + case PM_SUSPEND_PREPARE: + if (nr_cpu_ids <= 256) /* Expediting bad for large systems. */ + rcu_expedited = 1; + break; + case PM_POST_HIBERNATION: + case PM_POST_SUSPEND: + rcu_expedited = 0; + break; + default: + break; + } + return NOTIFY_OK; +} + /* * Spawn the kthread that handles this RCU flavor's grace periods. */ @@ -3273,6 +3293,7 @@ void __init rcu_init(void) * or the scheduler are operational. */ cpu_notifier(rcu_cpu_notify, 0); + pm_notifier(rcu_pm_notify, 0); for_each_online_cpu(cpu) rcu_cpu_notify(NULL, CPU_UP_PREPARE, (void *)(long)cpu); } -- cgit v1.2.3 From 15100df81fcc3109862f7c03266c0abff4262564 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Tue, 23 Apr 2013 11:31:50 -0700 Subject: rcu: Simplify debug-objects fixups The current debug-objects fixups are complex and heavyweight, and the fixups are not complete: Even with the fixups, RCU's callback lists can still be corrupted. This commit therefore strips the fixups down to their minimal form, eliminating two of the three. It would be even better if (for example) call_rcu() simply leaked any problematic callbacks, but for that to happen, the debug-objects system would need to inform its caller of suspicious situations. This is the subject of a later commit in this series. Signed-off-by: Paul E. McKenney Cc: Mathieu Desnoyers Cc: Sedat Dilek Cc: Davidlohr Bueso Cc: Rik van Riel Cc: Thomas Gleixner Cc: Linus Torvalds Tested-by: Sedat Dilek Reviewed-by: Josh Triplett --- kernel/rcupdate.c | 100 ------------------------------------------------------ 1 file changed, 100 deletions(-) (limited to 'kernel') diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c index 14994d4e1a54..33eb4620aa17 100644 --- a/kernel/rcupdate.c +++ b/kernel/rcupdate.c @@ -211,43 +211,6 @@ static inline void debug_rcu_head_free(struct rcu_head *head) debug_object_free(head, &rcuhead_debug_descr); } -/* - * fixup_init is called when: - * - an active object is initialized - */ -static int rcuhead_fixup_init(void *addr, enum debug_obj_state state) -{ - struct rcu_head *head = addr; - - switch (state) { - case ODEBUG_STATE_ACTIVE: - /* - * Ensure that queued callbacks are all executed. - * If we detect that we are nested in a RCU read-side critical - * section, we should simply fail, otherwise we would deadlock. - * In !PREEMPT configurations, there is no way to tell if we are - * in a RCU read-side critical section or not, so we never - * attempt any fixup and just print a warning. - */ -#ifndef CONFIG_PREEMPT - WARN_ON_ONCE(1); - return 0; -#endif - if (rcu_preempt_depth() != 0 || preempt_count() != 0 || - irqs_disabled()) { - WARN_ON_ONCE(1); - return 0; - } - rcu_barrier(); - rcu_barrier_sched(); - rcu_barrier_bh(); - debug_object_init(head, &rcuhead_debug_descr); - return 1; - default: - return 0; - } -} - /* * fixup_activate is called when: * - an active object is activated @@ -268,69 +231,8 @@ static int rcuhead_fixup_activate(void *addr, enum debug_obj_state state) debug_object_init(head, &rcuhead_debug_descr); debug_object_activate(head, &rcuhead_debug_descr); return 0; - - case ODEBUG_STATE_ACTIVE: - /* - * Ensure that queued callbacks are all executed. - * If we detect that we are nested in a RCU read-side critical - * section, we should simply fail, otherwise we would deadlock. - * In !PREEMPT configurations, there is no way to tell if we are - * in a RCU read-side critical section or not, so we never - * attempt any fixup and just print a warning. - */ -#ifndef CONFIG_PREEMPT - WARN_ON_ONCE(1); - return 0; -#endif - if (rcu_preempt_depth() != 0 || preempt_count() != 0 || - irqs_disabled()) { - WARN_ON_ONCE(1); - return 0; - } - rcu_barrier(); - rcu_barrier_sched(); - rcu_barrier_bh(); - debug_object_activate(head, &rcuhead_debug_descr); - return 1; default: - return 0; - } -} - -/* - * fixup_free is called when: - * - an active object is freed - */ -static int rcuhead_fixup_free(void *addr, enum debug_obj_state state) -{ - struct rcu_head *head = addr; - - switch (state) { - case ODEBUG_STATE_ACTIVE: - /* - * Ensure that queued callbacks are all executed. - * If we detect that we are nested in a RCU read-side critical - * section, we should simply fail, otherwise we would deadlock. - * In !PREEMPT configurations, there is no way to tell if we are - * in a RCU read-side critical section or not, so we never - * attempt any fixup and just print a warning. - */ -#ifndef CONFIG_PREEMPT - WARN_ON_ONCE(1); - return 0; -#endif - if (rcu_preempt_depth() != 0 || preempt_count() != 0 || - irqs_disabled()) { - WARN_ON_ONCE(1); - return 0; - } - rcu_barrier(); - rcu_barrier_sched(); - rcu_barrier_bh(); - debug_object_free(head, &rcuhead_debug_descr); return 1; - default: - return 0; } } @@ -369,9 +271,7 @@ EXPORT_SYMBOL_GPL(destroy_rcu_head_on_stack); struct debug_obj_descr rcuhead_debug_descr = { .name = "rcu_head", - .fixup_init = rcuhead_fixup_init, .fixup_activate = rcuhead_fixup_activate, - .fixup_free = rcuhead_fixup_free, }; EXPORT_SYMBOL_GPL(rcuhead_debug_descr); #endif /* #ifdef CONFIG_DEBUG_OBJECTS_RCU_HEAD */ -- cgit v1.2.3 From ae15018456c44b742d352af323e0b89eae4a6383 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Tue, 23 Apr 2013 13:20:57 -0700 Subject: rcu: Make call_rcu() leak callbacks for debug-object errors If someone does a duplicate call_rcu(), the worst thing the second call_rcu() could do would be to actually queue the callback the second time because doing so corrupts whatever list the callback was already queued on. This commit therefore makes __call_rcu() check the new return value from debug-objects and leak the callback upon error. This commit also substitutes rcu_leak_callback() for whatever callback function was previously in place in order to avoid freeing the callback out from under any readers that might still be referencing it. These changes increase the probability that the debug-objects error messages will actually make it somewhere visible. Signed-off-by: Paul E. McKenney Cc: Mathieu Desnoyers Cc: Sedat Dilek Cc: Davidlohr Bueso Cc: Rik van Riel Cc: Thomas Gleixner Cc: Linus Torvalds Tested-by: Sedat Dilek Reviewed-by: Josh Triplett --- kernel/rcu.h | 10 +++++++--- kernel/rcutree.c | 14 +++++++++++++- 2 files changed, 20 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/rcu.h b/kernel/rcu.h index 0a90ccc65bfb..77131966c4ad 100644 --- a/kernel/rcu.h +++ b/kernel/rcu.h @@ -67,12 +67,15 @@ extern struct debug_obj_descr rcuhead_debug_descr; -static inline void debug_rcu_head_queue(struct rcu_head *head) +static inline int debug_rcu_head_queue(struct rcu_head *head) { - debug_object_activate(head, &rcuhead_debug_descr); + int r1; + + r1 = debug_object_activate(head, &rcuhead_debug_descr); debug_object_active_state(head, &rcuhead_debug_descr, STATE_RCU_HEAD_READY, STATE_RCU_HEAD_QUEUED); + return r1; } static inline void debug_rcu_head_unqueue(struct rcu_head *head) @@ -83,8 +86,9 @@ static inline void debug_rcu_head_unqueue(struct rcu_head *head) debug_object_deactivate(head, &rcuhead_debug_descr); } #else /* !CONFIG_DEBUG_OBJECTS_RCU_HEAD */ -static inline void debug_rcu_head_queue(struct rcu_head *head) +static inline int debug_rcu_head_queue(struct rcu_head *head) { + return 0; } static inline void debug_rcu_head_unqueue(struct rcu_head *head) diff --git a/kernel/rcutree.c b/kernel/rcutree.c index a7bf517b0482..91840566e294 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -2304,6 +2304,13 @@ static void __call_rcu_core(struct rcu_state *rsp, struct rcu_data *rdp, } } +/* + * RCU callback function to leak a callback. + */ +static void rcu_leak_callback(struct rcu_head *rhp) +{ +} + /* * Helper function for call_rcu() and friends. The cpu argument will * normally be -1, indicating "currently running CPU". It may specify @@ -2318,7 +2325,12 @@ __call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu), struct rcu_data *rdp; WARN_ON_ONCE((unsigned long)head & 0x3); /* Misaligned rcu_head! */ - debug_rcu_head_queue(head); + if (debug_rcu_head_queue(head)) { + /* Probable double call_rcu(), so leak the callback. */ + ACCESS_ONCE(head->func) = rcu_leak_callback; + WARN_ONCE(1, "__call_rcu(): Leaked duplicate callback\n"); + return; + } head->func = func; head->next = NULL; -- cgit v1.2.3 From 1eafd31c640d6799c63136246a59d608bed93c74 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 20 Jun 2013 13:50:40 -0700 Subject: rcu: Avoid redundant grace-period kthread wakeups When setting up an in-the-future "advanced" grace period, the code needs to wake up the relevant grace-period kthread, which it currently does unconditionally. However, this results in needless wakeups in the case where the advanced grace period is being set up by the grace-period kthread itself, which is a non-uncommon situation. This commit therefore checks to see if the running thread is the grace-period kthread, and avoids doing the irq_work_queue()-mediated wakeup in that case. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutree.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 91840566e294..c6a064abd6a0 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1576,10 +1576,12 @@ rcu_start_gp_advanced(struct rcu_state *rsp, struct rcu_node *rnp, /* * We can't do wakeups while holding the rnp->lock, as that - * could cause possible deadlocks with the rq->lock. Deter - * the wakeup to interrupt context. + * could cause possible deadlocks with the rq->lock. Defer + * the wakeup to interrupt context. And don't bother waking + * up the running kthread. */ - irq_work_queue(&rsp->wakeup_work); + if (current != rsp->gp_kthread) + irq_work_queue(&rsp->wakeup_work); } /* -- cgit v1.2.3 From feed66ed26a53e700ca02ce1744fed7d0c647292 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 9 May 2013 08:55:54 -0700 Subject: rcu: Eliminate unused APIs intended for adaptive ticks The rcu_user_enter_after_irq() and rcu_user_exit_after_irq() functions were intended for use by adaptive ticks, but changes in implementation have rendered them unnecessary. This commit therefore removes them. Reported-by: Frederic Weisbecker Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutree.c | 43 ------------------------------------------- 1 file changed, 43 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 338f1d1c1c66..8807019138c6 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -444,27 +444,6 @@ void rcu_user_enter(void) { rcu_eqs_enter(1); } - -/** - * rcu_user_enter_after_irq - inform RCU that we are going to resume userspace - * after the current irq returns. - * - * This is similar to rcu_user_enter() but in the context of a non-nesting - * irq. After this call, RCU enters into idle mode when the interrupt - * returns. - */ -void rcu_user_enter_after_irq(void) -{ - unsigned long flags; - struct rcu_dynticks *rdtp; - - local_irq_save(flags); - rdtp = &__get_cpu_var(rcu_dynticks); - /* Ensure this irq is interrupting a non-idle RCU state. */ - WARN_ON_ONCE(!(rdtp->dynticks_nesting & DYNTICK_TASK_MASK)); - rdtp->dynticks_nesting = 1; - local_irq_restore(flags); -} #endif /* CONFIG_RCU_USER_QS */ /** @@ -581,28 +560,6 @@ void rcu_user_exit(void) { rcu_eqs_exit(1); } - -/** - * rcu_user_exit_after_irq - inform RCU that we won't resume to userspace - * idle mode after the current non-nesting irq returns. - * - * This is similar to rcu_user_exit() but in the context of an irq. - * This is called when the irq has interrupted a userspace RCU idle mode - * context. When the current non-nesting interrupt returns after this call, - * the CPU won't restore the RCU idle mode. - */ -void rcu_user_exit_after_irq(void) -{ - unsigned long flags; - struct rcu_dynticks *rdtp; - - local_irq_save(flags); - rdtp = &__get_cpu_var(rcu_dynticks); - /* Ensure we are interrupting an RCU idle mode. */ - WARN_ON_ONCE(rdtp->dynticks_nesting & DYNTICK_TASK_NEST_MASK); - rdtp->dynticks_nesting += DYNTICK_TASK_EXIT_IDLE; - local_irq_restore(flags); -} #endif /* CONFIG_RCU_USER_QS */ /** -- cgit v1.2.3 From b44379af1cf40050794832c38ea6a64e07eb5087 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Fri, 21 Jun 2013 11:08:45 -0700 Subject: nohz_full: Add Kconfig parameter for scalable detection of all-idle state At least one CPU must keep the scheduling-clock tick running for timekeeping purposes whenever there is a non-idle CPU. However, with the new nohz_full adaptive-idle machinery, it is difficult to distinguish between all CPUs really being idle as opposed to all non-idle CPUs being in adaptive-ticks mode. This commit therefore adds a Kconfig parameter as a first step towards enabling a scalable detection of full-system idle state. Signed-off-by: Paul E. McKenney Cc: Frederic Weisbecker Cc: Steven Rostedt [ paulmck: Update help text per Frederic Weisbecker. ] Reviewed-by: Josh Triplett --- kernel/time/Kconfig | 23 +++++++++++++++++++++++ 1 file changed, 23 insertions(+) (limited to 'kernel') diff --git a/kernel/time/Kconfig b/kernel/time/Kconfig index 70f27e89012b..c7d2fd67799e 100644 --- a/kernel/time/Kconfig +++ b/kernel/time/Kconfig @@ -134,6 +134,29 @@ config NO_HZ_FULL_ALL Note the boot CPU will still be kept outside the range to handle the timekeeping duty. +config NO_HZ_FULL_SYSIDLE + bool "Detect full-system idle state for full dynticks system" + depends on NO_HZ_FULL + default n + help + At least one CPU must keep the scheduling-clock tick running for + timekeeping purposes whenever there is a non-idle CPU, where + "non-idle" also includes dynticks CPUs as long as they are + running non-idle tasks. Because the underlying adaptive-tick + support cannot distinguish between all CPUs being idle and + all CPUs each running a single task in dynticks mode, the + underlying support simply ensures that there is always a CPU + handling the scheduling-clock tick, whether or not all CPUs + are idle. This Kconfig option enables scalable detection of + the all-CPUs-idle state, thus allowing the scheduling-clock + tick to be disabled when all CPUs are idle. Note that scalable + detection of the all-CPUs-idle state means that larger systems + will be slower to declare the all-CPUs-idle state. + + Say Y if you would like to help debug all-CPUs-idle detection. + + Say N if you are unsure. + config NO_HZ bool "Old Idle dynticks config" depends on !ARCH_USES_GETTIMEOFFSET && GENERIC_CLOCKEVENTS -- cgit v1.2.3 From 2333210b26cf7aaf48d71343029afb860103d9f9 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Fri, 21 Jun 2013 12:34:33 -0700 Subject: nohz_full: Add rcu_dyntick data for scalable detection of all-idle state This commit adds fields to the rcu_dyntick structure that are used to detect idle CPUs. These new fields differ from the existing ones in that the existing ones consider a CPU executing in user mode to be idle, where the new ones consider CPUs executing in user mode to be busy. The handling of these new fields is otherwise quite similar to that for the exiting fields. This commit also adds the initialization required for these fields. So, why is usermode execution treated differently, with RCU considering it a quiescent state equivalent to idle, while in contrast the new full-system idle state detection considers usermode execution to be non-idle? It turns out that although one of RCU's quiescent states is usermode execution, it is not a full-system idle state. This is because the purpose of the full-system idle state is not RCU, but rather determining when accurate timekeeping can safely be disabled. Whenever accurate timekeeping is required in a CONFIG_NO_HZ_FULL kernel, at least one CPU must keep the scheduling-clock tick going. If even one CPU is executing in user mode, accurate timekeeping is requires, particularly for architectures where gettimeofday() and friends do not enter the kernel. Only when all CPUs are really and truly idle can accurate timekeeping be disabled, allowing all CPUs to turn off the scheduling clock interrupt, thus greatly improving energy efficiency. This naturally raises the question "Why is this code in RCU rather than in timekeeping?", and the answer is that RCU has the data and infrastructure to efficiently make this determination. Signed-off-by: Paul E. McKenney Acked-by: Frederic Weisbecker Cc: Steven Rostedt Reviewed-by: Josh Triplett --- kernel/rcutree.c | 5 +++++ kernel/rcutree.h | 9 +++++++++ kernel/rcutree_plugin.h | 19 +++++++++++++++++++ 3 files changed, 33 insertions(+) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 8807019138c6..4f27b85d8c86 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -224,6 +224,10 @@ EXPORT_SYMBOL_GPL(rcu_note_context_switch); DEFINE_PER_CPU(struct rcu_dynticks, rcu_dynticks) = { .dynticks_nesting = DYNTICK_TASK_EXIT_IDLE, .dynticks = ATOMIC_INIT(1), +#ifdef CONFIG_NO_HZ_FULL_SYSIDLE + .dynticks_idle_nesting = DYNTICK_TASK_NEST_VALUE, + .dynticks_idle = ATOMIC_INIT(1), +#endif /* #ifdef CONFIG_NO_HZ_FULL_SYSIDLE */ }; static long blimit = 10; /* Maximum callbacks per rcu_do_batch. */ @@ -2904,6 +2908,7 @@ rcu_init_percpu_data(int cpu, struct rcu_state *rsp, int preemptible) rdp->blimit = blimit; init_callback_list(rdp); /* Re-enable callbacks on this CPU. */ rdp->dynticks->dynticks_nesting = DYNTICK_TASK_EXIT_IDLE; + rcu_sysidle_init_percpu_data(rdp->dynticks); atomic_set(&rdp->dynticks->dynticks, (atomic_read(&rdp->dynticks->dynticks) & ~0x1) + 1); raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ diff --git a/kernel/rcutree.h b/kernel/rcutree.h index cbdeac6cea9e..52d1be108e75 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -88,6 +88,14 @@ struct rcu_dynticks { /* Process level is worth LLONG_MAX/2. */ int dynticks_nmi_nesting; /* Track NMI nesting level. */ atomic_t dynticks; /* Even value for idle, else odd. */ +#ifdef CONFIG_NO_HZ_FULL_SYSIDLE + long long dynticks_idle_nesting; + /* irq/process nesting level from idle. */ + atomic_t dynticks_idle; /* Even value for idle, else odd. */ + /* "Idle" excludes userspace execution. */ + unsigned long dynticks_idle_jiffies; + /* End of last non-NMI non-idle period. */ +#endif /* #ifdef CONFIG_NO_HZ_FULL_SYSIDLE */ #ifdef CONFIG_RCU_FAST_NO_HZ bool all_lazy; /* Are all CPU's CBs lazy? */ unsigned long nonlazy_posted; @@ -545,6 +553,7 @@ static void rcu_boot_init_nocb_percpu_data(struct rcu_data *rdp); static void rcu_spawn_nocb_kthreads(struct rcu_state *rsp); static void rcu_kick_nohz_cpu(int cpu); static bool init_nocb_callback_list(struct rcu_data *rdp); +static void rcu_sysidle_init_percpu_data(struct rcu_dynticks *rdtp); #endif /* #ifndef RCU_TREE_NONCORE */ diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index dff86f53ee09..e5baccbd8038 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -2373,3 +2373,22 @@ static void rcu_kick_nohz_cpu(int cpu) smp_send_reschedule(cpu); #endif /* #ifdef CONFIG_NO_HZ_FULL */ } + + +#ifdef CONFIG_NO_HZ_FULL_SYSIDLE + +/* + * Initialize dynticks sysidle state for CPUs coming online. + */ +static void rcu_sysidle_init_percpu_data(struct rcu_dynticks *rdtp) +{ + rdtp->dynticks_idle_nesting = DYNTICK_TASK_NEST_VALUE; +} + +#else /* #ifdef CONFIG_NO_HZ_FULL_SYSIDLE */ + +static void rcu_sysidle_init_percpu_data(struct rcu_dynticks *rdtp) +{ +} + +#endif /* #else #ifdef CONFIG_NO_HZ_FULL_SYSIDLE */ -- cgit v1.2.3 From eb348b898290da242e46df75ab0b9772003e08b8 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Fri, 21 Jun 2013 13:00:57 -0700 Subject: nohz_full: Add per-CPU idle-state tracking This commit adds the code that updates the rcu_dyntick structure's new fields to track the per-CPU idle state based on interrupts and transitions into and out of the idle loop (NMIs are ignored because NMI handlers cannot cleanly read out the time anyway). This code is similar to the code that maintains RCU's idea of per-CPU idleness, but differs in that RCU treats CPUs running in user mode as idle, where this new code does not. Signed-off-by: Paul E. McKenney Acked-by: Frederic Weisbecker Cc: Steven Rostedt Reviewed-by: Josh Triplett --- kernel/rcutree.c | 4 +++ kernel/rcutree.h | 2 ++ kernel/rcutree_plugin.h | 79 +++++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 85 insertions(+) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 4f27b85d8c86..b0d2cc3ea15a 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -431,6 +431,7 @@ void rcu_idle_enter(void) local_irq_save(flags); rcu_eqs_enter(false); + rcu_sysidle_enter(&__get_cpu_var(rcu_dynticks), 0); local_irq_restore(flags); } EXPORT_SYMBOL_GPL(rcu_idle_enter); @@ -481,6 +482,7 @@ void rcu_irq_exit(void) trace_rcu_dyntick(TPS("--="), oldval, rdtp->dynticks_nesting); else rcu_eqs_enter_common(rdtp, oldval, true); + rcu_sysidle_enter(rdtp, 1); local_irq_restore(flags); } @@ -549,6 +551,7 @@ void rcu_idle_exit(void) local_irq_save(flags); rcu_eqs_exit(false); + rcu_sysidle_exit(&__get_cpu_var(rcu_dynticks), 0); local_irq_restore(flags); } EXPORT_SYMBOL_GPL(rcu_idle_exit); @@ -600,6 +603,7 @@ void rcu_irq_enter(void) trace_rcu_dyntick(TPS("++="), oldval, rdtp->dynticks_nesting); else rcu_eqs_exit_common(rdtp, oldval, true); + rcu_sysidle_exit(rdtp, 1); local_irq_restore(flags); } diff --git a/kernel/rcutree.h b/kernel/rcutree.h index 52d1be108e75..9dd8b177f1ac 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -553,6 +553,8 @@ static void rcu_boot_init_nocb_percpu_data(struct rcu_data *rdp); static void rcu_spawn_nocb_kthreads(struct rcu_state *rsp); static void rcu_kick_nohz_cpu(int cpu); static bool init_nocb_callback_list(struct rcu_data *rdp); +static void rcu_sysidle_enter(struct rcu_dynticks *rdtp, int irq); +static void rcu_sysidle_exit(struct rcu_dynticks *rdtp, int irq); static void rcu_sysidle_init_percpu_data(struct rcu_dynticks *rdtp); #endif /* #ifndef RCU_TREE_NONCORE */ diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index e5baccbd8038..eab81da614b8 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -2377,6 +2377,77 @@ static void rcu_kick_nohz_cpu(int cpu) #ifdef CONFIG_NO_HZ_FULL_SYSIDLE +/* + * Invoked to note exit from irq or task transition to idle. Note that + * usermode execution does -not- count as idle here! After all, we want + * to detect full-system idle states, not RCU quiescent states and grace + * periods. The caller must have disabled interrupts. + */ +static void rcu_sysidle_enter(struct rcu_dynticks *rdtp, int irq) +{ + unsigned long j; + + /* Adjust nesting, check for fully idle. */ + if (irq) { + rdtp->dynticks_idle_nesting--; + WARN_ON_ONCE(rdtp->dynticks_idle_nesting < 0); + if (rdtp->dynticks_idle_nesting != 0) + return; /* Still not fully idle. */ + } else { + if ((rdtp->dynticks_idle_nesting & DYNTICK_TASK_NEST_MASK) == + DYNTICK_TASK_NEST_VALUE) { + rdtp->dynticks_idle_nesting = 0; + } else { + rdtp->dynticks_idle_nesting -= DYNTICK_TASK_NEST_VALUE; + WARN_ON_ONCE(rdtp->dynticks_idle_nesting < 0); + return; /* Still not fully idle. */ + } + } + + /* Record start of fully idle period. */ + j = jiffies; + ACCESS_ONCE(rdtp->dynticks_idle_jiffies) = j; + smp_mb__before_atomic_inc(); + atomic_inc(&rdtp->dynticks_idle); + smp_mb__after_atomic_inc(); + WARN_ON_ONCE(atomic_read(&rdtp->dynticks_idle) & 0x1); +} + +/* + * Invoked to note entry to irq or task transition from idle. Note that + * usermode execution does -not- count as idle here! The caller must + * have disabled interrupts. + */ +static void rcu_sysidle_exit(struct rcu_dynticks *rdtp, int irq) +{ + /* Adjust nesting, check for already non-idle. */ + if (irq) { + rdtp->dynticks_idle_nesting++; + WARN_ON_ONCE(rdtp->dynticks_idle_nesting <= 0); + if (rdtp->dynticks_idle_nesting != 1) + return; /* Already non-idle. */ + } else { + /* + * Allow for irq misnesting. Yes, it really is possible + * to enter an irq handler then never leave it, and maybe + * also vice versa. Handle both possibilities. + */ + if (rdtp->dynticks_idle_nesting & DYNTICK_TASK_NEST_MASK) { + rdtp->dynticks_idle_nesting += DYNTICK_TASK_NEST_VALUE; + WARN_ON_ONCE(rdtp->dynticks_idle_nesting <= 0); + return; /* Already non-idle. */ + } else { + rdtp->dynticks_idle_nesting = DYNTICK_TASK_EXIT_IDLE; + } + } + + /* Record end of idle period. */ + smp_mb__before_atomic_inc(); + atomic_inc(&rdtp->dynticks_idle); + smp_mb__after_atomic_inc(); + WARN_ON_ONCE(!(atomic_read(&rdtp->dynticks_idle) & 0x1)); +} + /* * Initialize dynticks sysidle state for CPUs coming online. */ @@ -2387,6 +2458,14 @@ static void rcu_sysidle_init_percpu_data(struct rcu_dynticks *rdtp) #else /* #ifdef CONFIG_NO_HZ_FULL_SYSIDLE */ +static void rcu_sysidle_enter(struct rcu_dynticks *rdtp, int irq) +{ +} + +static void rcu_sysidle_exit(struct rcu_dynticks *rdtp, int irq) +{ +} + static void rcu_sysidle_init_percpu_data(struct rcu_dynticks *rdtp) { } -- cgit v1.2.3 From d4bd54fbac2ea5c30eb976ca557e905f489d55f4 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Fri, 21 Jun 2013 14:51:40 -0700 Subject: nohz_full: Add full-system idle states and variables This commit adds control variables and states for full-system idle. The system will progress through the states in numerical order when the system is fully idle (other than the timekeeping CPU), and reset down to the initial state if any non-timekeeping CPU goes non-idle. The current state is kept in full_sysidle_state. One flavor of RCU will be in charge of driving the state machine, defined by rcu_sysidle_state. This should be the busiest flavor of RCU. Signed-off-by: Paul E. McKenney Cc: Frederic Weisbecker Cc: Steven Rostedt Reviewed-by: Josh Triplett --- kernel/rcutree_plugin.h | 17 +++++++++++++++++ 1 file changed, 17 insertions(+) (limited to 'kernel') diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index eab81da614b8..a7419ceb19ad 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -2377,6 +2377,23 @@ static void rcu_kick_nohz_cpu(int cpu) #ifdef CONFIG_NO_HZ_FULL_SYSIDLE +/* + * Define RCU flavor that holds sysidle state. This needs to be the + * most active flavor of RCU. + */ +#ifdef CONFIG_PREEMPT_RCU +static struct rcu_state __maybe_unused *rcu_sysidle_state = &rcu_preempt_state; +#else /* #ifdef CONFIG_PREEMPT_RCU */ +static struct rcu_state __maybe_unused *rcu_sysidle_state = &rcu_sched_state; +#endif /* #else #ifdef CONFIG_PREEMPT_RCU */ + +static int __maybe_unused full_sysidle_state; /* Current system-idle state. */ +#define RCU_SYSIDLE_NOT 0 /* Some CPU is not idle. */ +#define RCU_SYSIDLE_SHORT 1 /* All CPUs idle for brief period. */ +#define RCU_SYSIDLE_LONG 2 /* All CPUs idle for long enough. */ +#define RCU_SYSIDLE_FULL 3 /* All CPUs idle, ready for sysidle. */ +#define RCU_SYSIDLE_FULL_NOTED 4 /* Actually entered sysidle state. */ + /* * Invoked to note exit from irq or task transition to idle. Note that * usermode execution does -not- count as idle here! After all, we want -- cgit v1.2.3 From 217af2a2ffbfc1498d1cf3a89fa478b5632df8f7 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Fri, 21 Jun 2013 15:39:06 -0700 Subject: nohz_full: Add full-system-idle arguments to API This commit adds an isidle and jiffies argument to force_qs_rnp(), dyntick_save_progress_counter(), and rcu_implicit_dynticks_qs() to enable RCU's force-quiescent-state process to check for full-system idle. Signed-off-by: Paul E. McKenney Cc: Frederic Weisbecker Cc: Steven Rostedt Cc: Lai Jiangshan [ paulmck: Use true and false for boolean constants per Lai Jiangshan. ] Reviewed-by: Josh Triplett --- kernel/rcutree.c | 25 ++++++++++++++++++------- 1 file changed, 18 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index b0d2cc3ea15a..7b5be56d95ae 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -246,7 +246,10 @@ module_param(jiffies_till_next_fqs, ulong, 0644); static void rcu_start_gp_advanced(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp); -static void force_qs_rnp(struct rcu_state *rsp, int (*f)(struct rcu_data *)); +static void force_qs_rnp(struct rcu_state *rsp, + int (*f)(struct rcu_data *rsp, bool *isidle, + unsigned long *maxj), + bool *isidle, unsigned long *maxj); static void force_quiescent_state(struct rcu_state *rsp); static int rcu_pending(int cpu); @@ -727,7 +730,8 @@ static int rcu_is_cpu_rrupt_from_idle(void) * credit them with an implicit quiescent state. Return 1 if this CPU * is in dynticks idle mode, which is an extended quiescent state. */ -static int dyntick_save_progress_counter(struct rcu_data *rdp) +static int dyntick_save_progress_counter(struct rcu_data *rdp, + bool *isidle, unsigned long *maxj) { rdp->dynticks_snap = atomic_add_return(0, &rdp->dynticks->dynticks); return (rdp->dynticks_snap & 0x1) == 0; @@ -739,7 +743,8 @@ static int dyntick_save_progress_counter(struct rcu_data *rdp) * idle state since the last call to dyntick_save_progress_counter() * for this same CPU, or by virtue of having been offline. */ -static int rcu_implicit_dynticks_qs(struct rcu_data *rdp) +static int rcu_implicit_dynticks_qs(struct rcu_data *rdp, + bool *isidle, unsigned long *maxj) { unsigned int curr; unsigned int snap; @@ -1361,16 +1366,19 @@ static int rcu_gp_init(struct rcu_state *rsp) int rcu_gp_fqs(struct rcu_state *rsp, int fqs_state_in) { int fqs_state = fqs_state_in; + bool isidle = false; + unsigned long maxj; struct rcu_node *rnp = rcu_get_root(rsp); rsp->n_force_qs++; if (fqs_state == RCU_SAVE_DYNTICK) { /* Collect dyntick-idle snapshots. */ - force_qs_rnp(rsp, dyntick_save_progress_counter); + force_qs_rnp(rsp, dyntick_save_progress_counter, + &isidle, &maxj); fqs_state = RCU_FORCE_QS; } else { /* Handle dyntick-idle and offline CPUs. */ - force_qs_rnp(rsp, rcu_implicit_dynticks_qs); + force_qs_rnp(rsp, rcu_implicit_dynticks_qs, &isidle, &maxj); } /* Clear flag to prevent immediate re-entry. */ if (ACCESS_ONCE(rsp->gp_flags) & RCU_GP_FLAG_FQS) { @@ -2069,7 +2077,10 @@ void rcu_check_callbacks(int cpu, int user) * * The caller must have suppressed start of new grace periods. */ -static void force_qs_rnp(struct rcu_state *rsp, int (*f)(struct rcu_data *)) +static void force_qs_rnp(struct rcu_state *rsp, + int (*f)(struct rcu_data *rsp, bool *isidle, + unsigned long *maxj), + bool *isidle, unsigned long *maxj) { unsigned long bit; int cpu; @@ -2093,7 +2104,7 @@ static void force_qs_rnp(struct rcu_state *rsp, int (*f)(struct rcu_data *)) bit = 1; for (; cpu <= rnp->grphi; cpu++, bit <<= 1) { if ((rnp->qsmask & bit) != 0 && - f(per_cpu_ptr(rsp->rda, cpu))) + f(per_cpu_ptr(rsp->rda, cpu), isidle, maxj)) mask |= bit; } if (mask != 0) { -- cgit v1.2.3 From 15e71911fcc655508e02f767a3d9b8b138051d2b Mon Sep 17 00:00:00 2001 From: Xie XiuQi Date: Mon, 29 Jul 2013 11:52:24 +0800 Subject: generic-ipi/locking: Fix misleading smp_call_function_any() description Fix locking description: after commit 8969a5ede0f9e17da4b9437 ("generic-ipi: remove kmalloc()"), wait = 0 can be guaranteed because we don't kmalloc() anymore. Signed-off-by: Xie XiuQi Cc: Sheng Yang Cc: Peter Zijlstra Cc: Jens Axboe Cc: Rusty Russell Link: http://lkml.kernel.org/r/51F5E6F8.1000801@huawei.com Signed-off-by: Ingo Molnar --- kernel/smp.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'kernel') diff --git a/kernel/smp.c b/kernel/smp.c index fe9f773d7114..b1c9034bdfcb 100644 --- a/kernel/smp.c +++ b/kernel/smp.c @@ -278,8 +278,6 @@ EXPORT_SYMBOL(smp_call_function_single); * @wait: If true, wait until function has completed. * * Returns 0 on success, else a negative status code (if no cpus were online). - * Note that @wait will be implicitly turned on in case of allocation failures, - * since we fall back to on-stack allocation. * * Selection preference: * 1) current cpu if in @mask -- cgit v1.2.3 From 1cb650b91ba582f6737457b7d22e368585596d2c Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Mon, 19 Aug 2013 10:05:24 +0800 Subject: cgroup: change cgroup_from_id() to css_from_id() Now we want cgroup core to always provide the css to use to the subsystems, so change this API to css_from_id(). Uninline css_from_id(), because it's getting bigger and cgroup_css() has been unexported. While at it, remove the #ifdef, and shuffle the order of the args. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cgroup.c | 22 ++++++++++++++++++++++ 1 file changed, 22 insertions(+) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index b69b572131e5..ff7d642a070a 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -5717,6 +5717,28 @@ struct cgroup_subsys_state *cgroup_css_from_dir(struct file *f, int id) return css ? css : ERR_PTR(-ENOENT); } +/** + * css_from_id - lookup css by id + * @id: the cgroup id + * @ss: cgroup subsys to be looked into + * + * Returns the css if there's valid one with @id, otherwise returns NULL. + * Should be called under rcu_read_lock(). + */ +struct cgroup_subsys_state *css_from_id(int id, struct cgroup_subsys *ss) +{ + struct cgroup *cgrp; + + rcu_lockdep_assert(rcu_read_lock_held() || + lockdep_is_held(&cgroup_mutex), + "css_from_id() needs proper protection"); + + cgrp = idr_find(&ss->root->cgroup_idr, id); + if (cgrp) + return cgroup_css(cgrp, ss->subsys_id); + return NULL; +} + #ifdef CONFIG_CGROUP_DEBUG static struct cgroup_subsys_state * debug_css_alloc(struct cgroup_subsys_state *parent_css) -- cgit v1.2.3 From 0bfb4aa67cef4982adc70590a31624d7b35a0bda Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 15 Aug 2013 11:42:36 -0400 Subject: cgroup: fix subsystem file accesses on the root cgroup 105347ba5 ("cgroup: make cgroup_file_open() rcu_read_lock() around cgroup_css() and add cfent->css") added cfent->css to cache the associted cgroup_subsys_state across file operations. A cfent is associated with single css throughout its lifetime and the origimal commit initialized the cache pointer during cgroup_add_file() and verified that it matches the actual one in cgroup_file_open(). While this works fine for !root cgroups, it's broken for root cgroups as files in a root cgroup are created before the css's are associated with the cgroup and thus cgroup_css() call in cgroup_add_file() returns NULL associating all cfents in the root cgroup with NULL css. This makes cgroup_file_open() trigger WARN and fail with -ENODEV for all !core subsystem files in the root cgroups. There's no reason to initialize cfent->css separately from cgroup_add_file(). As the association never changes, cgroup_file_open() can set it unconditionally every time and containing the logic in cgroup_file_open() makes more sense anyway as the only reason it's necessary is file->private_data being already occupied. Fix it by setting cfent->css unconditionally from cgroup_file_open(). Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 24 ++++++++++-------------- 1 file changed, 10 insertions(+), 14 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index ff7d642a070a..896e035eb6e4 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -2490,10 +2490,18 @@ static int cgroup_file_open(struct inode *inode, struct file *file) } rcu_read_unlock(); - /* css should match @cfe->css, see cgroup_add_file() for details */ - if (!css || WARN_ON_ONCE(css != cfe->css)) + if (!css) return -ENODEV; + /* + * @cfe->css is used by read/write/close to determine the + * associated css. @file->private_data would be a better place but + * that's already used by seqfile. Multiple accessors may use it + * simultaneously which is okay as the association never changes. + */ + WARN_ON_ONCE(cfe->css && cfe->css != css); + cfe->css = css; + if (cft->read_map || cft->read_seq_string) { file->f_op = &cgroup_seqfile_operations; err = single_open(file, cgroup_seqfile_show, cfe); @@ -2772,18 +2780,6 @@ static int cgroup_add_file(struct cgroup *cgrp, struct cftype *cft) dentry->d_fsdata = cfe; simple_xattrs_init(&cfe->xattrs); - /* - * cfe->css is used by read/write/close to determine the associated - * css. file->private_data would be a better place but that's - * already used by seqfile. Note that open will use the usual - * cgroup_css() and css_tryget() to acquire the css and this - * caching doesn't affect css lifetime management. - */ - if (cft->ss) - cfe->css = cgroup_css(cgrp, cft->ss->subsys_id); - else - cfe->css = &cgrp->dummy_css; - mode = cgroup_file_mode(cft); error = cgroup_create_file(dentry, mode | S_IFREG, cgrp->root->sb); if (!error) { -- cgit v1.2.3 From 6e6eab0efdf48fb2d8d7aee904d7740acb4661c6 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 15 Aug 2013 11:43:15 -0400 Subject: cgroup: fix cgroup_write_event_control() 81eeaf0411 ("cgroup: make cftype->[un]register_event() deal with cgroup_subsys_state inst ead of cgroup") updated the cftype event methods to take @css (cgroup_subsys_state) instead of @cgroup; however, it incorrectly used @css passed to cgroup_write_event_control(), which the dummy_css for the cgroup as the file is a cgroup core file. This leads to oops on event registration. Fix it by using the css matching the event target file. Note that cgroup_write_event_control() now disallows cgroup core files from being event sources. This is for simplicity and doesn't matter as cgroup_event will be moved and made specific to memcg. Signed-off-by: Tejun Heo Acked-by: Li Zefan --- kernel/cgroup.c | 25 +++++++++++++++++++++---- 1 file changed, 21 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 896e035eb6e4..ef43e3f453ef 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4040,10 +4040,10 @@ static void cgroup_event_ptable_queue_proc(struct file *file, * Input must be in format ' '. * Interpretation of args is defined by control file implementation. */ -static int cgroup_write_event_control(struct cgroup_subsys_state *css, +static int cgroup_write_event_control(struct cgroup_subsys_state *dummy_css, struct cftype *cft, const char *buffer) { - struct cgroup *cgrp = css->cgroup; + struct cgroup *cgrp = dummy_css->cgroup; struct cgroup_event *event; struct cgroup *cgrp_cfile; unsigned int efd, cfd; @@ -4065,7 +4065,7 @@ static int cgroup_write_event_control(struct cgroup_subsys_state *css, event = kzalloc(sizeof(*event), GFP_KERNEL); if (!event) return -ENOMEM; - event->css = css; + INIT_LIST_HEAD(&event->list); init_poll_funcptr(&event->pt, cgroup_event_ptable_queue_proc); init_waitqueue_func_entry(&event->wait, cgroup_event_wake); @@ -4101,6 +4101,23 @@ static int cgroup_write_event_control(struct cgroup_subsys_state *css, goto out_put_cfile; } + if (!event->cft->ss) { + ret = -EBADF; + goto out_put_cfile; + } + + /* determine the css of @cfile and associate @event with it */ + rcu_read_lock(); + + ret = -EINVAL; + event->css = cgroup_css(cgrp, event->cft->ss->subsys_id); + if (event->css) + ret = 0; + + rcu_read_unlock(); + if (ret) + goto out_put_cfile; + /* * The file to be monitored must be in the same cgroup as * cgroup.event_control is. @@ -4116,7 +4133,7 @@ static int cgroup_write_event_control(struct cgroup_subsys_state *css, goto out_put_cfile; } - ret = event->cft->register_event(css, event->cft, + ret = event->cft->register_event(event->css, event->cft, event->eventfd, buffer); if (ret) goto out_put_cfile; -- cgit v1.2.3 From 2203547f82b7727e2cd3fee3e56fceae2b2b691c Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Sun, 18 Aug 2013 20:08:07 -0700 Subject: kernel: fix new kernel-doc warning in wait.c Fix new kernel-doc warnings in kernel/wait.c: Warning(kernel/wait.c:374): No description found for parameter 'p' Warning(kernel/wait.c:374): Excess function parameter 'word' description in 'wake_up_atomic_t' Warning(kernel/wait.c:374): Excess function parameter 'bit' description in 'wake_up_atomic_t' Signed-off-by: Randy Dunlap Cc: David Howells Signed-off-by: Linus Torvalds --- kernel/wait.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/wait.c b/kernel/wait.c index dec68bd4e9d8..d550920e040c 100644 --- a/kernel/wait.c +++ b/kernel/wait.c @@ -363,8 +363,7 @@ EXPORT_SYMBOL(out_of_line_wait_on_atomic_t); /** * wake_up_atomic_t - Wake up a waiter on a atomic_t - * @word: The word being waited on, a kernel virtual address - * @bit: The bit of the word being waited on + * @p: The atomic_t being waited on, a kernel virtual address * * Wake up anyone waiting for the atomic_t to go to zero. * -- cgit v1.2.3 From 79ac6834c255d9e3832209f3738d6bff7b87c743 Mon Sep 17 00:00:00 2001 From: Christoph Jaeger Date: Tue, 20 Aug 2013 15:33:18 +0930 Subject: module: fix sprintf format specifier in param_get_byte() In param_get_byte(), to which the macro STANDARD_PARAM_DEF(byte, ...) expands, "%c" is used to print an unsigned char. So it gets printed as a character what is not intended here. Use "%hhu" instead. [Rusty: note drivers which would be effected: drivers/net/wireless/cw1200/main.c drivers/ntb/ntb_transport.c:68 drivers/scsi/lpfc/lpfc_attr.c drivers/usb/atm/speedtch.c drivers/usb/gadget/g_ffs.c ] Acked-by: Jon Mason (for ntb) Acked-by: Michal Nazarewicz (for g_ffs.c) Signed-off-by: Christoph Jaeger Signed-off-by: Rusty Russell --- kernel/params.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/params.c b/kernel/params.c index 440e65d1a544..59f7ac7bec04 100644 --- a/kernel/params.c +++ b/kernel/params.c @@ -252,7 +252,7 @@ int parse_args(const char *doing, EXPORT_SYMBOL(param_ops_##name) -STANDARD_PARAM_DEF(byte, unsigned char, "%c", unsigned long, strict_strtoul); +STANDARD_PARAM_DEF(byte, unsigned char, "%hhu", unsigned long, strict_strtoul); STANDARD_PARAM_DEF(short, short, "%hi", long, strict_strtol); STANDARD_PARAM_DEF(ushort, unsigned short, "%hu", unsigned long, strict_strtoul); STANDARD_PARAM_DEF(int, int, "%i", long, strict_strtol); -- cgit v1.2.3 From ab013c5f60b7ead254863c75b9adc2a47992d01b Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 20 Aug 2013 15:33:19 +0930 Subject: module: Add flag to allow mod params to have no arguments Currently the params.c code allows only two "set" functions to have no arguments. If a parameter does not have an argument, then it looks at the set function and tests if it is either param_set_bool() or param_set_bint(). If it is not one of these functions, then it fails the loading of the module. But there may be module parameters that have different set functions and still allow no arguments. But unless each of these cases adds their function to the if statement, it wont be allowed to have no arguments. This method gets rather messing and does not scale. Instead, introduce a flags field to the kernel_param_ops, where if the flag KERNEL_PARAM_FL_NOARG is set, the parameter will not fail if it does not contain an argument. It will be expected that the corresponding set function can handle a NULL pointer as "val". Signed-off-by: Steven Rostedt Signed-off-by: Rusty Russell --- kernel/params.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/params.c b/kernel/params.c index 59f7ac7bec04..ec4299cfade8 100644 --- a/kernel/params.c +++ b/kernel/params.c @@ -103,8 +103,8 @@ static int parse_one(char *param, || params[i].level > max_level) return 0; /* No one handled NULL, so do it here. */ - if (!val && params[i].ops->set != param_set_bool - && params[i].ops->set != param_set_bint) + if (!val && + !(params[i].ops->flags & KERNEL_PARAM_FL_NOARG)) return -EINVAL; pr_debug("handling %s with %p\n", param, params[i].ops->set); @@ -320,6 +320,7 @@ int param_get_bool(char *buffer, const struct kernel_param *kp) EXPORT_SYMBOL(param_get_bool); struct kernel_param_ops param_ops_bool = { + .flags = KERNEL_PARAM_FL_NOARG, .set = param_set_bool, .get = param_get_bool, }; @@ -370,6 +371,7 @@ int param_set_bint(const char *val, const struct kernel_param *kp) EXPORT_SYMBOL(param_set_bint); struct kernel_param_ops param_ops_bint = { + .flags = KERNEL_PARAM_FL_NOARG, .set = param_set_bint, .get = param_get_int, }; -- cgit v1.2.3 From 0ce814096f388f6801587f01c1c5ee1d04e746b3 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 20 Aug 2013 15:33:19 +0930 Subject: module: Add NOARG flag for ops with param_set_bool_enable_only() set function The ops that uses param_set_bool_enable_only() as its set function can easily handle being used without an argument. There's no reason to fail the loading of the module if it does not have one. Signed-off-by: Steven Rostedt Signed-off-by: Rusty Russell --- kernel/module.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/module.c b/kernel/module.c index 206915830d29..4eb26b6d6547 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -136,6 +136,7 @@ static int param_set_bool_enable_only(const char *val, } static const struct kernel_param_ops param_ops_bool_enable_only = { + .flags = KERNEL_PARAM_FL_NOARG, .set = param_set_bool_enable_only, .get = param_get_bool, }; -- cgit v1.2.3 From cc56ded3fdd365e07e03315379ee6612a68fd817 Mon Sep 17 00:00:00 2001 From: Chen Gang Date: Tue, 20 Aug 2013 15:34:21 +0930 Subject: kernel/module.c: use scnprintf() instead of sprintf() For some strings, they are permitted to be larger than PAGE_SIZE, so need use scnprintf() instead of sprintf(), or it will cause issue. One case is: if a module version is crazy defined (length more than PAGE_SIZE), 'modinfo' command is still OK (print full contents), but for "cat /sys/modules/'modname'/version", will cause issue in kernel. Signed-off-by: Chen Gang Signed-off-by: Rusty Russell --- kernel/module.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/module.c b/kernel/module.c index 4eb26b6d6547..40ee1dc3c3bf 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -604,7 +604,7 @@ static void setup_modinfo_##field(struct module *mod, const char *s) \ static ssize_t show_modinfo_##field(struct module_attribute *mattr, \ struct module_kobject *mk, char *buffer) \ { \ - return sprintf(buffer, "%s\n", mk->mod->field); \ + return scnprintf(buffer, PAGE_SIZE, "%s\n", mk->mod->field); \ } \ static int modinfo_##field##_exists(struct module *mod) \ { \ -- cgit v1.2.3 From f4940ab7c5992d3fabcda039744fb7657749798e Mon Sep 17 00:00:00 2001 From: Chen Gang Date: Tue, 20 Aug 2013 15:35:04 +0930 Subject: kernel/params.c: use scnprintf() instead of sprintf() For some strings (e.g. version string), they are permitted to be larger than PAGE_SIZE (although meaningless), so recommend to use scnprintf() instead of sprintf(). Signed-off-by: Chen Gang Signed-off-by: Rusty Russell --- kernel/params.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/params.c b/kernel/params.c index ec4299cfade8..e5f8f17e57cf 100644 --- a/kernel/params.c +++ b/kernel/params.c @@ -241,7 +241,8 @@ int parse_args(const char *doing, } \ int param_get_##name(char *buffer, const struct kernel_param *kp) \ { \ - return sprintf(buffer, format, *((type *)kp->arg)); \ + return scnprintf(buffer, PAGE_SIZE, format, \ + *((type *)kp->arg)); \ } \ struct kernel_param_ops param_ops_##name = { \ .set = param_set_##name, \ @@ -285,7 +286,7 @@ EXPORT_SYMBOL(param_set_charp); int param_get_charp(char *buffer, const struct kernel_param *kp) { - return sprintf(buffer, "%s", *((char **)kp->arg)); + return scnprintf(buffer, PAGE_SIZE, "%s", *((char **)kp->arg)); } EXPORT_SYMBOL(param_get_charp); @@ -829,7 +830,7 @@ ssize_t __modver_version_show(struct module_attribute *mattr, struct module_version_attribute *vattr = container_of(mattr, struct module_version_attribute, mattr); - return sprintf(buf, "%s\n", vattr->version); + return scnprintf(buf, PAGE_SIZE, "%s\n", vattr->version); } extern const struct module_version_attribute *__start___modver[]; -- cgit v1.2.3 From d185af300fe43c130083851ca918ea2bb9600f0f Mon Sep 17 00:00:00 2001 From: Yacine Belkadi Date: Wed, 31 Jul 2013 14:59:24 -0700 Subject: workqueue: fix some scripts/kernel-doc warnings When building the htmldocs (in verbose mode), scripts/kernel-doc reports the following type of warnings: Warning(kernel/workqueue.c:653): No description found for return value of 'get_work_pool' Fix them by: - Using "Return:" sections to introduce descriptions of return values - Adding some missing descriptions Signed-off-by: Yacine Belkadi Signed-off-by: Jiri Kosina --- kernel/workqueue.c | 107 +++++++++++++++++++++++++++++++++-------------------- 1 file changed, 66 insertions(+), 41 deletions(-) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 0b72e816b8d0..7f01a3eeaf95 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -540,6 +540,8 @@ static int worker_pool_assign_id(struct worker_pool *pool) * This must be called either with pwq_lock held or sched RCU read locked. * If the pwq needs to be used beyond the locking in effect, the caller is * responsible for guaranteeing that the pwq stays online. + * + * Return: The unbound pool_workqueue for @node. */ static struct pool_workqueue *unbound_pwq_by_node(struct workqueue_struct *wq, int node) @@ -638,8 +640,6 @@ static struct pool_workqueue *get_work_pwq(struct work_struct *work) * get_work_pool - return the worker_pool a given work was associated with * @work: the work item of interest * - * Return the worker_pool @work was last associated with. %NULL if none. - * * Pools are created and destroyed under wq_pool_mutex, and allows read * access under sched-RCU read lock. As such, this function should be * called under wq_pool_mutex or with preemption disabled. @@ -648,6 +648,8 @@ static struct pool_workqueue *get_work_pwq(struct work_struct *work) * mentioned locking is in effect. If the returned pool needs to be used * beyond the critical section, the caller is responsible for ensuring the * returned pool is and stays online. + * + * Return: The worker_pool @work was last associated with. %NULL if none. */ static struct worker_pool *get_work_pool(struct work_struct *work) { @@ -671,7 +673,7 @@ static struct worker_pool *get_work_pool(struct work_struct *work) * get_work_pool_id - return the worker pool ID a given work is associated with * @work: the work item of interest * - * Return the worker_pool ID @work was last associated with. + * Return: The worker_pool ID @work was last associated with. * %WORK_OFFQ_POOL_NONE if none. */ static int get_work_pool_id(struct work_struct *work) @@ -830,7 +832,7 @@ void wq_worker_waking_up(struct task_struct *task, int cpu) * CONTEXT: * spin_lock_irq(rq->lock) * - * RETURNS: + * Return: * Worker task on @cpu to wake up, %NULL if none. */ struct task_struct *wq_worker_sleeping(struct task_struct *task, int cpu) @@ -965,8 +967,8 @@ static inline void worker_clr_flags(struct worker *worker, unsigned int flags) * CONTEXT: * spin_lock_irq(pool->lock). * - * RETURNS: - * Pointer to worker which is executing @work if found, NULL + * Return: + * Pointer to worker which is executing @work if found, %NULL * otherwise. */ static struct worker *find_worker_executing_work(struct worker_pool *pool, @@ -1154,14 +1156,16 @@ out_put: * @flags: place to store irq state * * Try to grab PENDING bit of @work. This function can handle @work in any - * stable state - idle, on timer or on worklist. Return values are + * stable state - idle, on timer or on worklist. * + * Return: * 1 if @work was pending and we successfully stole PENDING * 0 if @work was idle and we claimed PENDING * -EAGAIN if PENDING couldn't be grabbed at the moment, safe to busy-retry * -ENOENT if someone else is canceling @work, this state may persist * for arbitrarily long * + * Note: * On >= 0 return, the caller owns @work's PENDING bit. To avoid getting * interrupted while holding PENDING and @work off queue, irq must be * disabled on entry. This, combined with delayed_work->timer being @@ -1403,10 +1407,10 @@ retry: * @wq: workqueue to use * @work: work to queue * - * Returns %false if @work was already on a queue, %true otherwise. - * * We queue the work to a specific CPU, the caller must ensure it * can't go away. + * + * Return: %false if @work was already on a queue, %true otherwise. */ bool queue_work_on(int cpu, struct workqueue_struct *wq, struct work_struct *work) @@ -1476,7 +1480,7 @@ static void __queue_delayed_work(int cpu, struct workqueue_struct *wq, * @dwork: work to queue * @delay: number of jiffies to wait before queueing * - * Returns %false if @work was already on a queue, %true otherwise. If + * Return: %false if @work was already on a queue, %true otherwise. If * @delay is zero and @dwork is idle, it will be scheduled for immediate * execution. */ @@ -1512,7 +1516,7 @@ EXPORT_SYMBOL(queue_delayed_work_on); * zero, @work is guaranteed to be scheduled immediately regardless of its * current state. * - * Returns %false if @dwork was idle and queued, %true if @dwork was + * Return: %false if @dwork was idle and queued, %true if @dwork was * pending and its timer was modified. * * This function is safe to call from any context including IRQ handler. @@ -1627,7 +1631,7 @@ static void worker_leave_idle(struct worker *worker) * Might sleep. Called without any lock but returns with pool->lock * held. * - * RETURNS: + * Return: * %true if the associated pool is online (@worker is successfully * bound), %false if offline. */ @@ -1688,7 +1692,7 @@ static struct worker *alloc_worker(void) * CONTEXT: * Might sleep. Does GFP_KERNEL allocations. * - * RETURNS: + * Return: * Pointer to the newly created worker. */ static struct worker *create_worker(struct worker_pool *pool) @@ -1788,6 +1792,8 @@ static void start_worker(struct worker *worker) * @pool: the target pool * * Grab the managership of @pool and create and start a new worker for it. + * + * Return: 0 on success. A negative error code otherwise. */ static int create_and_start_worker(struct worker_pool *pool) { @@ -1932,7 +1938,7 @@ static void pool_mayday_timeout(unsigned long __pool) * multiple times. Does GFP_KERNEL allocations. Called only from * manager. * - * RETURNS: + * Return: * %false if no action was taken and pool->lock stayed locked, %true * otherwise. */ @@ -1989,7 +1995,7 @@ restart: * spin_lock_irq(pool->lock) which may be released and regrabbed * multiple times. Called only from manager. * - * RETURNS: + * Return: * %false if no action was taken and pool->lock stayed locked, %true * otherwise. */ @@ -2032,7 +2038,7 @@ static bool maybe_destroy_workers(struct worker_pool *pool) * spin_lock_irq(pool->lock) which may be released and regrabbed * multiple times. Does GFP_KERNEL allocations. * - * RETURNS: + * Return: * spin_lock_irq(pool->lock) which may be released and regrabbed * multiple times. Does GFP_KERNEL allocations. */ @@ -2246,6 +2252,8 @@ static void process_scheduled_works(struct worker *worker) * work items regardless of their specific target workqueue. The only * exception is work items which belong to workqueues with a rescuer which * will be explained in rescuer_thread(). + * + * Return: 0 */ static int worker_thread(void *__worker) { @@ -2344,6 +2352,8 @@ sleep: * those works so that forward progress can be guaranteed. * * This should happen rarely. + * + * Return: 0 */ static int rescuer_thread(void *__rescuer) { @@ -2516,7 +2526,7 @@ static void insert_wq_barrier(struct pool_workqueue *pwq, * CONTEXT: * mutex_lock(wq->mutex). * - * RETURNS: + * Return: * %true if @flush_color >= 0 and there's something to flush. %false * otherwise. */ @@ -2824,7 +2834,7 @@ already_gone: * Wait until @work has finished execution. @work is guaranteed to be idle * on return if it hasn't been requeued since flush started. * - * RETURNS: + * Return: * %true if flush_work() waited for the work to finish execution, * %false if it was already idle. */ @@ -2884,7 +2894,7 @@ static bool __cancel_work_timer(struct work_struct *work, bool is_dwork) * The caller must ensure that the workqueue on which @work was last * queued can't be destroyed before this function returns. * - * RETURNS: + * Return: * %true if @work was pending, %false otherwise. */ bool cancel_work_sync(struct work_struct *work) @@ -2901,7 +2911,7 @@ EXPORT_SYMBOL_GPL(cancel_work_sync); * immediate execution. Like flush_work(), this function only * considers the last queueing instance of @dwork. * - * RETURNS: + * Return: * %true if flush_work() waited for the work to finish execution, * %false if it was already idle. */ @@ -2919,11 +2929,15 @@ EXPORT_SYMBOL(flush_delayed_work); * cancel_delayed_work - cancel a delayed work * @dwork: delayed_work to cancel * - * Kill off a pending delayed_work. Returns %true if @dwork was pending - * and canceled; %false if wasn't pending. Note that the work callback - * function may still be running on return, unless it returns %true and the - * work doesn't re-arm itself. Explicitly flush or use - * cancel_delayed_work_sync() to wait on it. + * Kill off a pending delayed_work. + * + * Return: %true if @dwork was pending and canceled; %false if it wasn't + * pending. + * + * Note: + * The work callback function may still be running on return, unless + * it returns %true and the work doesn't re-arm itself. Explicitly flush or + * use cancel_delayed_work_sync() to wait on it. * * This function is safe to call from any context including IRQ handler. */ @@ -2952,7 +2966,7 @@ EXPORT_SYMBOL(cancel_delayed_work); * * This is cancel_work_sync() for delayed works. * - * RETURNS: + * Return: * %true if @dwork was pending, %false otherwise. */ bool cancel_delayed_work_sync(struct delayed_work *dwork) @@ -2969,7 +2983,7 @@ EXPORT_SYMBOL(cancel_delayed_work_sync); * system workqueue and blocks until all CPUs have completed. * schedule_on_each_cpu() is very slow. * - * RETURNS: + * Return: * 0 on success, -errno on failure. */ int schedule_on_each_cpu(work_func_t func) @@ -3037,7 +3051,7 @@ EXPORT_SYMBOL(flush_scheduled_work); * Executes the function immediately if process context is available, * otherwise schedules the function for delayed execution. * - * Returns: 0 - function was executed + * Return: 0 - function was executed * 1 - function was scheduled for execution */ int execute_in_process_context(work_func_t fn, struct execute_work *ew) @@ -3294,7 +3308,7 @@ static void wq_device_release(struct device *dev) * apply_workqueue_attrs() may race against userland updating the * attributes. * - * Returns 0 on success, -errno on failure. + * Return: 0 on success, -errno on failure. */ int workqueue_sysfs_register(struct workqueue_struct *wq) { @@ -3387,7 +3401,9 @@ void free_workqueue_attrs(struct workqueue_attrs *attrs) * @gfp_mask: allocation mask to use * * Allocate a new workqueue_attrs, initialize with default settings and - * return it. Returns NULL on failure. + * return it. + * + * Return: The allocated new workqueue_attr on success. %NULL on failure. */ struct workqueue_attrs *alloc_workqueue_attrs(gfp_t gfp_mask) { @@ -3440,7 +3456,8 @@ static bool wqattrs_equal(const struct workqueue_attrs *a, * @pool: worker_pool to initialize * * Initiailize a newly zalloc'd @pool. It also allocates @pool->attrs. - * Returns 0 on success, -errno on failure. Even on failure, all fields + * + * Return: 0 on success, -errno on failure. Even on failure, all fields * inside @pool proper are initialized and put_unbound_pool() can be called * on @pool safely to release it. */ @@ -3547,9 +3564,12 @@ static void put_unbound_pool(struct worker_pool *pool) * Obtain a worker_pool which has the same attributes as @attrs, bump the * reference count and return it. If there already is a matching * worker_pool, it will be used; otherwise, this function attempts to - * create a new one. On failure, returns NULL. + * create a new one. * * Should be called with wq_pool_mutex held. + * + * Return: On success, a worker_pool with the same attributes as @attrs. + * On failure, %NULL. */ static struct worker_pool *get_unbound_pool(const struct workqueue_attrs *attrs) { @@ -3779,9 +3799,7 @@ static void free_unbound_pwq(struct pool_workqueue *pwq) * * Calculate the cpumask a workqueue with @attrs should use on @node. If * @cpu_going_down is >= 0, that cpu is considered offline during - * calculation. The result is stored in @cpumask. This function returns - * %true if the resulting @cpumask is different from @attrs->cpumask, - * %false if equal. + * calculation. The result is stored in @cpumask. * * If NUMA affinity is not enabled, @attrs->cpumask is always used. If * enabled and @node has online CPUs requested by @attrs, the returned @@ -3790,6 +3808,9 @@ static void free_unbound_pwq(struct pool_workqueue *pwq) * * The caller is responsible for ensuring that the cpumask of @node stays * stable. + * + * Return: %true if the resulting @cpumask is different from @attrs->cpumask, + * %false if equal. */ static bool wq_calc_node_cpumask(const struct workqueue_attrs *attrs, int node, int cpu_going_down, cpumask_t *cpumask) @@ -3843,8 +3864,9 @@ static struct pool_workqueue *numa_pwq_tbl_install(struct workqueue_struct *wq, * items finish. Note that a work item which repeatedly requeues itself * back-to-back will stay on its current pwq. * - * Performs GFP_KERNEL allocations. Returns 0 on success and -errno on - * failure. + * Performs GFP_KERNEL allocations. + * + * Return: 0 on success and -errno on failure. */ int apply_workqueue_attrs(struct workqueue_struct *wq, const struct workqueue_attrs *attrs) @@ -4312,6 +4334,8 @@ EXPORT_SYMBOL_GPL(workqueue_set_max_active); * * Determine whether %current is a workqueue rescuer. Can be used from * work functions to determine whether it's being run off the rescuer task. + * + * Return: %true if %current is a workqueue rescuer. %false otherwise. */ bool current_is_workqueue_rescuer(void) { @@ -4335,7 +4359,7 @@ bool current_is_workqueue_rescuer(void) * workqueue being congested on one CPU doesn't mean the workqueue is also * contested on other CPUs / NUMA nodes. * - * RETURNS: + * Return: * %true if congested, %false otherwise. */ bool workqueue_congested(int cpu, struct workqueue_struct *wq) @@ -4368,7 +4392,7 @@ EXPORT_SYMBOL_GPL(workqueue_congested); * synchronization around this function and the test result is * unreliable and only useful as advisory hints or for debugging. * - * RETURNS: + * Return: * OR'd bitmask of WORK_BUSY_* bits. */ unsigned int work_busy(struct work_struct *work) @@ -4746,9 +4770,10 @@ static void work_for_cpu_fn(struct work_struct *work) * @fn: the function to run * @arg: the function arg * - * This will return the value @fn returns. * It is up to the caller to ensure that the cpu doesn't go offline. * The caller must not hold any locks which would prevent @fn from completing. + * + * Return: The value @fn returns. */ long work_on_cpu(int cpu, long (*fn)(void *), void *arg) { @@ -4813,7 +4838,7 @@ void freeze_workqueues_begin(void) * CONTEXT: * Grabs and releases wq_pool_mutex. * - * RETURNS: + * Return: * %true if some freezable workqueues are still busy. %false if freezing * is complete. */ -- cgit v1.2.3 From d2818df168b2c80c7449e47bd349094c308fa323 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Tue, 23 Apr 2013 17:05:42 -0700 Subject: rcu: Add duplicate-callback tests to rcutorture This commit adds a object_debug option to rcutorture to allow the debug-object-based checks for duplicate call_rcu() invocations to be deterministically tested. Signed-off-by: Paul E. McKenney Cc: Mathieu Desnoyers Cc: Sedat Dilek Cc: Davidlohr Bueso Cc: Rik van Riel Cc: Thomas Gleixner Cc: Linus Torvalds Tested-by: Sedat Dilek [ paulmck: Banish mid-function ifdef, more or less per Josh Triplett. ] Reviewed-by: Josh Triplett [ paulmck: Improve duplicate-callback test, per Lai Jiangshan. ] --- kernel/rcutorture.c | 61 +++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 61 insertions(+) (limited to 'kernel') diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c index 3d936f0fbcd8..c898f14a5b7d 100644 --- a/kernel/rcutorture.c +++ b/kernel/rcutorture.c @@ -66,6 +66,7 @@ static int fqs_duration; /* Duration of bursts (us), 0 to disable. */ static int fqs_holdoff; /* Hold time within burst (us). */ static int fqs_stutter = 3; /* Wait time between bursts (s). */ static int n_barrier_cbs; /* Number of callbacks to test RCU barriers. */ +static int object_debug; /* Test object-debug double call_rcu()?. */ static int onoff_interval; /* Wait time between CPU hotplugs, 0=disable. */ static int onoff_holdoff; /* Seconds after boot before CPU hotplugs. */ static int shutdown_secs; /* Shutdown time (s). <=0 for no shutdown. */ @@ -100,6 +101,8 @@ module_param(fqs_stutter, int, 0444); MODULE_PARM_DESC(fqs_stutter, "Wait time between fqs bursts (s)"); module_param(n_barrier_cbs, int, 0444); MODULE_PARM_DESC(n_barrier_cbs, "# of callbacks/kthreads for barrier testing"); +module_param(object_debug, int, 0444); +MODULE_PARM_DESC(object_debug, "Enable debug-object double call_rcu() testing"); module_param(onoff_interval, int, 0444); MODULE_PARM_DESC(onoff_interval, "Time between CPU hotplugs (s), 0=disable"); module_param(onoff_holdoff, int, 0444); @@ -1934,6 +1937,62 @@ rcu_torture_cleanup(void) rcu_torture_print_module_parms(cur_ops, "End of test: SUCCESS"); } +#ifdef CONFIG_DEBUG_OBJECTS_RCU_HEAD +static void rcu_torture_leak_cb(struct rcu_head *rhp) +{ +} + +static void rcu_torture_err_cb(struct rcu_head *rhp) +{ + /* + * This -might- happen due to race conditions, but is unlikely. + * The scenario that leads to this happening is that the + * first of the pair of duplicate callbacks is queued, + * someone else starts a grace period that includes that + * callback, then the second of the pair must wait for the + * next grace period. Unlikely, but can happen. If it + * does happen, the debug-objects subsystem won't have splatted. + */ + pr_alert("rcutorture: duplicated callback was invoked.\n"); +} +#endif /* #ifdef CONFIG_DEBUG_OBJECTS_RCU_HEAD */ + +/* + * Verify that double-free causes debug-objects to complain, but only + * if CONFIG_DEBUG_OBJECTS_RCU_HEAD=y. Otherwise, say that the test + * cannot be carried out. + */ +static void rcu_test_debug_objects(void) +{ +#ifdef CONFIG_DEBUG_OBJECTS_RCU_HEAD + struct rcu_head rh1; + struct rcu_head rh2; + + init_rcu_head_on_stack(&rh1); + init_rcu_head_on_stack(&rh2); + pr_alert("rcutorture: WARN: Duplicate call_rcu() test starting.\n"); + + /* Try to queue the rh2 pair of callbacks for the same grace period. */ + preempt_disable(); /* Prevent preemption from interrupting test. */ + rcu_read_lock(); /* Make it impossible to finish a grace period. */ + call_rcu(&rh1, rcu_torture_leak_cb); /* Start grace period. */ + local_irq_disable(); /* Make it harder to start a new grace period. */ + call_rcu(&rh2, rcu_torture_leak_cb); + call_rcu(&rh2, rcu_torture_err_cb); /* Duplicate callback. */ + local_irq_enable(); + rcu_read_unlock(); + preempt_enable(); + + /* Wait for them all to get done so we can safely return. */ + rcu_barrier(); + pr_alert("rcutorture: WARN: Duplicate call_rcu() test complete.\n"); + destroy_rcu_head_on_stack(&rh1); + destroy_rcu_head_on_stack(&rh2); +#else /* #ifdef CONFIG_DEBUG_OBJECTS_RCU_HEAD */ + pr_alert("rcutorture: !CONFIG_DEBUG_OBJECTS_RCU_HEAD, not testing duplicate call_rcu()\n"); +#endif /* #else #ifdef CONFIG_DEBUG_OBJECTS_RCU_HEAD */ +} + static int __init rcu_torture_init(void) { @@ -2163,6 +2222,8 @@ rcu_torture_init(void) firsterr = retval; goto unwind; } + if (object_debug) + rcu_test_debug_objects(); rcutorture_record_test_transition(); mutex_unlock(&fullstop_mutex); return 0; -- cgit v1.2.3 From 2ec1f2d98752293f4831ce7d7bdbc3fc36bdd114 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Wed, 12 Jun 2013 15:12:21 -0700 Subject: rcu: Increase rcutorture test coverage Currently, rcutorture has separate torture_types to test synchronous, asynchronous, and expedited grace-period primitives. This has two disadvantages: (1) Three times the number of runs to cover the combinations and (2) Little testing of concurrent combinations of the three options. This commit therefore adds a pair of module parameters that control normal and expedited state, with the default being both types, randomly selected, by the fakewriter processes, thus reducing source-code size and increasing test coverage. In addtion, the writer task switches between asynchronous-normal and expedited grace-period primitives driven by the same pair of module parameters. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutorture.c | 226 +++++++++++++++------------------------------------- 1 file changed, 63 insertions(+), 163 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c index c898f14a5b7d..ddef61871878 100644 --- a/kernel/rcutorture.c +++ b/kernel/rcutorture.c @@ -65,6 +65,8 @@ static int irqreader = 1; /* RCU readers from irq (timers). */ static int fqs_duration; /* Duration of bursts (us), 0 to disable. */ static int fqs_holdoff; /* Hold time within burst (us). */ static int fqs_stutter = 3; /* Wait time between bursts (s). */ +static bool gp_exp; /* Use expedited GP wait primitives. */ +static bool gp_normal; /* Use normal GP wait primitives. */ static int n_barrier_cbs; /* Number of callbacks to test RCU barriers. */ static int object_debug; /* Test object-debug double call_rcu()?. */ static int onoff_interval; /* Wait time between CPU hotplugs, 0=disable. */ @@ -99,6 +101,10 @@ module_param(fqs_holdoff, int, 0444); MODULE_PARM_DESC(fqs_holdoff, "Holdoff time within fqs bursts (us)"); module_param(fqs_stutter, int, 0444); MODULE_PARM_DESC(fqs_stutter, "Wait time between fqs bursts (s)"); +module_param(gp_normal, bool, 0444); +MODULE_PARM_DESC(gp_normal, "Use normal (non-expedited) GP wait primitives"); +module_param(gp_exp, bool, 0444); +MODULE_PARM_DESC(gp_exp, "Use expedited GP wait primitives"); module_param(n_barrier_cbs, int, 0444); MODULE_PARM_DESC(n_barrier_cbs, "# of callbacks/kthreads for barrier testing"); module_param(object_debug, int, 0444); @@ -363,6 +369,7 @@ struct rcu_torture_ops { int (*completed)(void); void (*deferred_free)(struct rcu_torture *p); void (*sync)(void); + void (*exp_sync)(void); void (*call)(struct rcu_head *head, void (*func)(struct rcu_head *rcu)); void (*cb_barrier)(void); void (*fqs)(void); @@ -446,81 +453,27 @@ static void rcu_torture_deferred_free(struct rcu_torture *p) call_rcu(&p->rtort_rcu, rcu_torture_cb); } -static struct rcu_torture_ops rcu_ops = { - .init = NULL, - .readlock = rcu_torture_read_lock, - .read_delay = rcu_read_delay, - .readunlock = rcu_torture_read_unlock, - .completed = rcu_torture_completed, - .deferred_free = rcu_torture_deferred_free, - .sync = synchronize_rcu, - .call = call_rcu, - .cb_barrier = rcu_barrier, - .fqs = rcu_force_quiescent_state, - .stats = NULL, - .irq_capable = 1, - .can_boost = rcu_can_boost(), - .name = "rcu" -}; - -static void rcu_sync_torture_deferred_free(struct rcu_torture *p) -{ - int i; - struct rcu_torture *rp; - struct rcu_torture *rp1; - - cur_ops->sync(); - list_add(&p->rtort_free, &rcu_torture_removed); - list_for_each_entry_safe(rp, rp1, &rcu_torture_removed, rtort_free) { - i = rp->rtort_pipe_count; - if (i > RCU_TORTURE_PIPE_LEN) - i = RCU_TORTURE_PIPE_LEN; - atomic_inc(&rcu_torture_wcount[i]); - if (++rp->rtort_pipe_count >= RCU_TORTURE_PIPE_LEN) { - rp->rtort_mbtest = 0; - list_del(&rp->rtort_free); - rcu_torture_free(rp); - } - } -} - static void rcu_sync_torture_init(void) { INIT_LIST_HEAD(&rcu_torture_removed); } -static struct rcu_torture_ops rcu_sync_ops = { +static struct rcu_torture_ops rcu_ops = { .init = rcu_sync_torture_init, .readlock = rcu_torture_read_lock, .read_delay = rcu_read_delay, .readunlock = rcu_torture_read_unlock, .completed = rcu_torture_completed, - .deferred_free = rcu_sync_torture_deferred_free, + .deferred_free = rcu_torture_deferred_free, .sync = synchronize_rcu, - .call = NULL, - .cb_barrier = NULL, - .fqs = rcu_force_quiescent_state, - .stats = NULL, - .irq_capable = 1, - .can_boost = rcu_can_boost(), - .name = "rcu_sync" -}; - -static struct rcu_torture_ops rcu_expedited_ops = { - .init = rcu_sync_torture_init, - .readlock = rcu_torture_read_lock, - .read_delay = rcu_read_delay, /* just reuse rcu's version. */ - .readunlock = rcu_torture_read_unlock, - .completed = rcu_no_completed, - .deferred_free = rcu_sync_torture_deferred_free, - .sync = synchronize_rcu_expedited, - .call = NULL, - .cb_barrier = NULL, + .exp_sync = synchronize_rcu_expedited, + .call = call_rcu, + .cb_barrier = rcu_barrier, .fqs = rcu_force_quiescent_state, .stats = NULL, .irq_capable = 1, .can_boost = rcu_can_boost(), - .name = "rcu_expedited" + .name = "rcu" }; /* @@ -549,13 +502,14 @@ static void rcu_bh_torture_deferred_free(struct rcu_torture *p) } static struct rcu_torture_ops rcu_bh_ops = { - .init = NULL, + .init = rcu_sync_torture_init, .readlock = rcu_bh_torture_read_lock, .read_delay = rcu_read_delay, /* just reuse rcu's version. */ .readunlock = rcu_bh_torture_read_unlock, .completed = rcu_bh_torture_completed, .deferred_free = rcu_bh_torture_deferred_free, .sync = synchronize_rcu_bh, + .exp_sync = synchronize_rcu_bh_expedited, .call = call_rcu_bh, .cb_barrier = rcu_barrier_bh, .fqs = rcu_bh_force_quiescent_state, @@ -564,38 +518,6 @@ static struct rcu_torture_ops rcu_bh_ops = { .name = "rcu_bh" }; -static struct rcu_torture_ops rcu_bh_sync_ops = { - .init = rcu_sync_torture_init, - .readlock = rcu_bh_torture_read_lock, - .read_delay = rcu_read_delay, /* just reuse rcu's version. */ - .readunlock = rcu_bh_torture_read_unlock, - .completed = rcu_bh_torture_completed, - .deferred_free = rcu_sync_torture_deferred_free, - .sync = synchronize_rcu_bh, - .call = NULL, - .cb_barrier = NULL, - .fqs = rcu_bh_force_quiescent_state, - .stats = NULL, - .irq_capable = 1, - .name = "rcu_bh_sync" -}; - -static struct rcu_torture_ops rcu_bh_expedited_ops = { - .init = rcu_sync_torture_init, - .readlock = rcu_bh_torture_read_lock, - .read_delay = rcu_read_delay, /* just reuse rcu's version. */ - .readunlock = rcu_bh_torture_read_unlock, - .completed = rcu_bh_torture_completed, - .deferred_free = rcu_sync_torture_deferred_free, - .sync = synchronize_rcu_bh_expedited, - .call = NULL, - .cb_barrier = NULL, - .fqs = rcu_bh_force_quiescent_state, - .stats = NULL, - .irq_capable = 1, - .name = "rcu_bh_expedited" -}; - /* * Definitions for srcu torture testing. */ @@ -670,6 +592,11 @@ static int srcu_torture_stats(char *page) return cnt; } +static void srcu_torture_synchronize_expedited(void) +{ + synchronize_srcu_expedited(&srcu_ctl); +} + static struct rcu_torture_ops srcu_ops = { .init = rcu_sync_torture_init, .readlock = srcu_torture_read_lock, @@ -678,45 +605,13 @@ static struct rcu_torture_ops srcu_ops = { .completed = srcu_torture_completed, .deferred_free = srcu_torture_deferred_free, .sync = srcu_torture_synchronize, + .exp_sync = srcu_torture_synchronize_expedited, .call = srcu_torture_call, .cb_barrier = srcu_torture_barrier, .stats = srcu_torture_stats, .name = "srcu" }; -static struct rcu_torture_ops srcu_sync_ops = { - .init = rcu_sync_torture_init, - .readlock = srcu_torture_read_lock, - .read_delay = srcu_read_delay, - .readunlock = srcu_torture_read_unlock, - .completed = srcu_torture_completed, - .deferred_free = rcu_sync_torture_deferred_free, - .sync = srcu_torture_synchronize, - .call = NULL, - .cb_barrier = NULL, - .stats = srcu_torture_stats, - .name = "srcu_sync" -}; - -static void srcu_torture_synchronize_expedited(void) -{ - synchronize_srcu_expedited(&srcu_ctl); -} - -static struct rcu_torture_ops srcu_expedited_ops = { - .init = rcu_sync_torture_init, - .readlock = srcu_torture_read_lock, - .read_delay = srcu_read_delay, - .readunlock = srcu_torture_read_unlock, - .completed = srcu_torture_completed, - .deferred_free = rcu_sync_torture_deferred_free, - .sync = srcu_torture_synchronize_expedited, - .call = NULL, - .cb_barrier = NULL, - .stats = srcu_torture_stats, - .name = "srcu_expedited" -}; - /* * Definitions for sched torture testing. */ @@ -745,6 +640,8 @@ static struct rcu_torture_ops sched_ops = { .completed = rcu_no_completed, .deferred_free = rcu_sched_torture_deferred_free, .sync = synchronize_sched, + .exp_sync = synchronize_sched_expedited, + .call = call_rcu_sched, .cb_barrier = rcu_barrier_sched, .fqs = rcu_sched_force_quiescent_state, .stats = NULL, @@ -752,35 +649,6 @@ static struct rcu_torture_ops sched_ops = { .name = "sched" }; -static struct rcu_torture_ops sched_sync_ops = { - .init = rcu_sync_torture_init, - .readlock = sched_torture_read_lock, - .read_delay = rcu_read_delay, /* just reuse rcu's version. */ - .readunlock = sched_torture_read_unlock, - .completed = rcu_no_completed, - .deferred_free = rcu_sync_torture_deferred_free, - .sync = synchronize_sched, - .cb_barrier = NULL, - .fqs = rcu_sched_force_quiescent_state, - .stats = NULL, - .name = "sched_sync" -}; - -static struct rcu_torture_ops sched_expedited_ops = { - .init = rcu_sync_torture_init, - .readlock = sched_torture_read_lock, - .read_delay = rcu_read_delay, /* just reuse rcu's version. */ - .readunlock = sched_torture_read_unlock, - .completed = rcu_no_completed, - .deferred_free = rcu_sync_torture_deferred_free, - .sync = synchronize_sched_expedited, - .cb_barrier = NULL, - .fqs = rcu_sched_force_quiescent_state, - .stats = NULL, - .irq_capable = 1, - .name = "sched_expedited" -}; - /* * RCU torture priority-boost testing. Runs one real-time thread per * CPU for moderate bursts, repeatedly registering RCU callbacks and @@ -930,9 +798,11 @@ rcu_torture_fqs(void *arg) static int rcu_torture_writer(void *arg) { + bool exp; int i; long oldbatch = rcu_batches_completed(); struct rcu_torture *rp; + struct rcu_torture *rp1; struct rcu_torture *old_rp; static DEFINE_RCU_RANDOM(rand); @@ -957,7 +827,31 @@ rcu_torture_writer(void *arg) i = RCU_TORTURE_PIPE_LEN; atomic_inc(&rcu_torture_wcount[i]); old_rp->rtort_pipe_count++; - cur_ops->deferred_free(old_rp); + if (gp_normal == gp_exp) + exp = !!(rcu_random(&rand) & 0x80); + else + exp = gp_exp; + if (!exp) { + cur_ops->deferred_free(old_rp); + } else { + cur_ops->exp_sync(); + list_add(&old_rp->rtort_free, + &rcu_torture_removed); + list_for_each_entry_safe(rp, rp1, + &rcu_torture_removed, + rtort_free) { + i = rp->rtort_pipe_count; + if (i > RCU_TORTURE_PIPE_LEN) + i = RCU_TORTURE_PIPE_LEN; + atomic_inc(&rcu_torture_wcount[i]); + if (++rp->rtort_pipe_count >= + RCU_TORTURE_PIPE_LEN) { + rp->rtort_mbtest = 0; + list_del(&rp->rtort_free); + rcu_torture_free(rp); + } + } + } } rcutorture_record_progress(++rcu_torture_current_version); oldbatch = cur_ops->completed(); @@ -986,10 +880,18 @@ rcu_torture_fakewriter(void *arg) schedule_timeout_uninterruptible(1 + rcu_random(&rand)%10); udelay(rcu_random(&rand) & 0x3ff); if (cur_ops->cb_barrier != NULL && - rcu_random(&rand) % (nfakewriters * 8) == 0) + rcu_random(&rand) % (nfakewriters * 8) == 0) { cur_ops->cb_barrier(); - else + } else if (gp_normal == gp_exp) { + if (rcu_random(&rand) & 0x80) + cur_ops->sync(); + else + cur_ops->exp_sync(); + } else if (gp_normal) { cur_ops->sync(); + } else { + cur_ops->exp_sync(); + } rcu_stutter_wait("rcu_torture_fakewriter"); } while (!kthread_should_stop() && fullstop == FULLSTOP_DONTSTOP); @@ -2000,11 +1902,9 @@ rcu_torture_init(void) int cpu; int firsterr = 0; int retval; - static struct rcu_torture_ops *torture_ops[] = - { &rcu_ops, &rcu_sync_ops, &rcu_expedited_ops, - &rcu_bh_ops, &rcu_bh_sync_ops, &rcu_bh_expedited_ops, - &srcu_ops, &srcu_sync_ops, &srcu_expedited_ops, - &sched_ops, &sched_sync_ops, &sched_expedited_ops, }; + static struct rcu_torture_ops *torture_ops[] = { + &rcu_ops, &rcu_bh_ops, &srcu_ops, &sched_ops, + }; mutex_lock(&fullstop_mutex); -- cgit v1.2.3 From d10453e9742f4711b004caae7741476073b4f603 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 13 Jun 2013 15:12:24 -0700 Subject: rcu: Sort rcutorture module parameters There are getting to be too many module parameters to permit the current semi-random order, so this patch orders them. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutorture.c | 101 +++++++++++++++++++++++++--------------------------- 1 file changed, 49 insertions(+), 52 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c index ddef61871878..e3a1244eeb56 100644 --- a/kernel/rcutorture.c +++ b/kernel/rcutorture.c @@ -52,81 +52,78 @@ MODULE_LICENSE("GPL"); MODULE_AUTHOR("Paul E. McKenney and Josh Triplett "); -static int nreaders = -1; /* # reader threads, defaults to 2*ncpus */ -static int nfakewriters = 4; /* # fake writer threads */ -static int stat_interval = 60; /* Interval between stats, in seconds. */ - /* Zero means "only at end of test". */ -static bool verbose; /* Print more debug info. */ -static bool test_no_idle_hz = true; - /* Test RCU support for tickless idle CPUs. */ -static int shuffle_interval = 3; /* Interval between shuffles (in sec)*/ -static int stutter = 5; /* Start/stop testing interval (in sec) */ -static int irqreader = 1; /* RCU readers from irq (timers). */ -static int fqs_duration; /* Duration of bursts (us), 0 to disable. */ -static int fqs_holdoff; /* Hold time within burst (us). */ -static int fqs_stutter = 3; /* Wait time between bursts (s). */ -static bool gp_exp; /* Use expedited GP wait primitives. */ -static bool gp_normal; /* Use normal GP wait primitives. */ -static int n_barrier_cbs; /* Number of callbacks to test RCU barriers. */ -static int object_debug; /* Test object-debug double call_rcu()?. */ -static int onoff_interval; /* Wait time between CPU hotplugs, 0=disable. */ -static int onoff_holdoff; /* Seconds after boot before CPU hotplugs. */ -static int shutdown_secs; /* Shutdown time (s). <=0 for no shutdown. */ -static int stall_cpu; /* CPU-stall duration (s). 0 for no stall. */ -static int stall_cpu_holdoff = 10; /* Time to wait until stall (s). */ -static int test_boost = 1; /* Test RCU prio boost: 0=no, 1=maybe, 2=yes. */ -static int test_boost_interval = 7; /* Interval between boost tests, seconds. */ -static int test_boost_duration = 4; /* Duration of each boost test, seconds. */ -static char *torture_type = "rcu"; /* What RCU implementation to torture. */ - -module_param(nreaders, int, 0444); -MODULE_PARM_DESC(nreaders, "Number of RCU reader threads"); -module_param(nfakewriters, int, 0444); -MODULE_PARM_DESC(nfakewriters, "Number of RCU fake writer threads"); -module_param(stat_interval, int, 0644); -MODULE_PARM_DESC(stat_interval, "Number of seconds between stats printk()s"); -module_param(verbose, bool, 0444); -MODULE_PARM_DESC(verbose, "Enable verbose debugging printk()s"); -module_param(test_no_idle_hz, bool, 0444); -MODULE_PARM_DESC(test_no_idle_hz, "Test support for tickless idle CPUs"); -module_param(shuffle_interval, int, 0444); -MODULE_PARM_DESC(shuffle_interval, "Number of seconds between shuffles"); -module_param(stutter, int, 0444); -MODULE_PARM_DESC(stutter, "Number of seconds to run/halt test"); -module_param(irqreader, int, 0444); -MODULE_PARM_DESC(irqreader, "Allow RCU readers from irq handlers"); +static int fqs_duration; module_param(fqs_duration, int, 0444); -MODULE_PARM_DESC(fqs_duration, "Duration of fqs bursts (us)"); +MODULE_PARM_DESC(fqs_duration, "Duration of fqs bursts (us), 0 to disable"); +static int fqs_holdoff; module_param(fqs_holdoff, int, 0444); MODULE_PARM_DESC(fqs_holdoff, "Holdoff time within fqs bursts (us)"); +static int fqs_stutter = 3; module_param(fqs_stutter, int, 0444); MODULE_PARM_DESC(fqs_stutter, "Wait time between fqs bursts (s)"); -module_param(gp_normal, bool, 0444); -MODULE_PARM_DESC(gp_normal, "Use normal (non-expedited) GP wait primitives"); +static bool gp_exp; module_param(gp_exp, bool, 0444); MODULE_PARM_DESC(gp_exp, "Use expedited GP wait primitives"); +static bool gp_normal; +module_param(gp_normal, bool, 0444); +MODULE_PARM_DESC(gp_normal, "Use normal (non-expedited) GP wait primitives"); +static int irqreader = 1; +module_param(irqreader, int, 0444); +MODULE_PARM_DESC(irqreader, "Allow RCU readers from irq handlers"); +static int n_barrier_cbs; module_param(n_barrier_cbs, int, 0444); MODULE_PARM_DESC(n_barrier_cbs, "# of callbacks/kthreads for barrier testing"); +static int nfakewriters = 4; +module_param(nfakewriters, int, 0444); +MODULE_PARM_DESC(nfakewriters, "Number of RCU fake writer threads"); +static int nreaders = -1; +module_param(nreaders, int, 0444); +MODULE_PARM_DESC(nreaders, "Number of RCU reader threads"); +static int object_debug; module_param(object_debug, int, 0444); MODULE_PARM_DESC(object_debug, "Enable debug-object double call_rcu() testing"); -module_param(onoff_interval, int, 0444); -MODULE_PARM_DESC(onoff_interval, "Time between CPU hotplugs (s), 0=disable"); +static int onoff_holdoff; module_param(onoff_holdoff, int, 0444); MODULE_PARM_DESC(onoff_holdoff, "Time after boot before CPU hotplugs (s)"); +static int onoff_interval; +module_param(onoff_interval, int, 0444); +MODULE_PARM_DESC(onoff_interval, "Time between CPU hotplugs (s), 0=disable"); +static int shuffle_interval = 3; +module_param(shuffle_interval, int, 0444); +MODULE_PARM_DESC(shuffle_interval, "Number of seconds between shuffles"); +static int shutdown_secs; module_param(shutdown_secs, int, 0444); -MODULE_PARM_DESC(shutdown_secs, "Shutdown time (s), zero to disable."); +MODULE_PARM_DESC(shutdown_secs, "Shutdown time (s), <= zero to disable."); +static int stall_cpu; module_param(stall_cpu, int, 0444); MODULE_PARM_DESC(stall_cpu, "Stall duration (s), zero to disable."); +static int stall_cpu_holdoff = 10; module_param(stall_cpu_holdoff, int, 0444); MODULE_PARM_DESC(stall_cpu_holdoff, "Time to wait before starting stall (s)."); +static int stat_interval = 60; +module_param(stat_interval, int, 0644); +MODULE_PARM_DESC(stat_interval, "Number of seconds between stats printk()s"); +static int stutter = 5; +module_param(stutter, int, 0444); +MODULE_PARM_DESC(stutter, "Number of seconds to run/halt test"); +static int test_boost = 1; module_param(test_boost, int, 0444); MODULE_PARM_DESC(test_boost, "Test RCU prio boost: 0=no, 1=maybe, 2=yes."); -module_param(test_boost_interval, int, 0444); -MODULE_PARM_DESC(test_boost_interval, "Interval between boost tests, seconds."); +static int test_boost_duration = 4; module_param(test_boost_duration, int, 0444); MODULE_PARM_DESC(test_boost_duration, "Duration of each boost test, seconds."); +static int test_boost_interval = 7; +module_param(test_boost_interval, int, 0444); +MODULE_PARM_DESC(test_boost_interval, "Interval between boost tests, seconds."); +static bool test_no_idle_hz = true; +module_param(test_no_idle_hz, bool, 0444); +MODULE_PARM_DESC(test_no_idle_hz, "Test support for tickless idle CPUs"); +static char *torture_type = "rcu"; module_param(torture_type, charp, 0444); -MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, rcu_bh, srcu)"); +MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, rcu_bh, ...)"); +static bool verbose; +module_param(verbose, bool, 0444); +MODULE_PARM_DESC(verbose, "Enable verbose debugging printk()s"); #define TORTURE_FLAG "-torture:" #define PRINTK_STRING(s) \ -- cgit v1.2.3 From ef47db8e99d53f0da5270accd2ee71bcf9e25f11 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 13 Jun 2013 15:30:00 -0700 Subject: rcu: Remove unused variable from rcu_torture_writer() The oldbatch variable in rcu_torture_writer() is stored to, but never loaded from. This commit therefore removes it. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutorture.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c index e3a1244eeb56..20ce3b6f9aa6 100644 --- a/kernel/rcutorture.c +++ b/kernel/rcutorture.c @@ -797,7 +797,6 @@ rcu_torture_writer(void *arg) { bool exp; int i; - long oldbatch = rcu_batches_completed(); struct rcu_torture *rp; struct rcu_torture *rp1; struct rcu_torture *old_rp; @@ -851,7 +850,6 @@ rcu_torture_writer(void *arg) } } rcutorture_record_progress(++rcu_torture_current_version); - oldbatch = cur_ops->completed(); rcu_stutter_wait("rcu_torture_writer"); } while (!kthread_should_stop() && fullstop == FULLSTOP_DONTSTOP); VERBOSE_PRINTK_STRING("rcu_torture_writer task stopping"); -- cgit v1.2.3 From 7a6a41073c345ff5ef5e81317211481c0da3f7f3 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Fri, 21 Jun 2013 06:24:56 -0700 Subject: rcu: Make rcutorture emit online failures if verbose Although rcutorture counts CPU-hotplug online failures, it does not explicitly record which CPUs were having trouble coming online. This commit therefore emits a console message when online failure occurs. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett --- kernel/rcutorture.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c index 20ce3b6f9aa6..be63101c6175 100644 --- a/kernel/rcutorture.c +++ b/kernel/rcutorture.c @@ -1434,7 +1434,13 @@ rcu_torture_onoff(void *arg) torture_type, cpu); starttime = jiffies; n_online_attempts++; - if (cpu_up(cpu) == 0) { + ret = cpu_up(cpu); + if (ret) { + if (verbose) + pr_alert("%s" TORTURE_FLAG + "rcu_torture_onoff task: online %d failed: errno %d\n", + torture_type, cpu, ret); + } else { if (verbose) pr_alert("%s" TORTURE_FLAG "rcu_torture_onoff task: onlined %d\n", -- cgit v1.2.3 From 458fb381eacdd23366cfa2fbdf5a467848683e3a Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Fri, 26 Jul 2013 20:47:42 -0700 Subject: rcu: Simplify _rcu_barrier() processing This commit drops an unneeded ACCESS_ONCE() and simplifies an "our work is done" check in _rcu_barrier(). This applies feedback from Linus (https://lkml.org/lkml/2013/7/26/777) that he gave to similar code in an unrelated patch. Signed-off-by: Paul E. McKenney Reviewed-by: Josh Triplett [ paulmck: Fix comment to match code, reported by Lai Jiangshan. ] --- kernel/rcutree.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index c6a064abd6a0..a4a04f311cfb 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -2817,9 +2817,20 @@ static void _rcu_barrier(struct rcu_state *rsp) * transition. The "if" expression below therefore rounds the old * value up to the next even number and adds two before comparing. */ - snap_done = ACCESS_ONCE(rsp->n_barrier_done); + snap_done = rsp->n_barrier_done; _rcu_barrier_trace(rsp, "Check", -1, snap_done); - if (ULONG_CMP_GE(snap_done, ((snap + 1) & ~0x1) + 2)) { + + /* + * If the value in snap is odd, we needed to wait for the current + * rcu_barrier() to complete, then wait for the next one, in other + * words, we need the value of snap_done to be three larger than + * the value of snap. On the other hand, if the value in snap is + * even, we only had to wait for the next rcu_barrier() to complete, + * in other words, we need the value of snap_done to be only two + * greater than the value of snap. The "(snap + 3) & ~0x1" computes + * this for us (thank you, Linus!). + */ + if (ULONG_CMP_GE(snap_done, (snap + 3) & ~0x1)) { _rcu_barrier_trace(rsp, "EarlyExit", -1, snap_done); smp_mb(); /* caller's subsequent code after above check. */ mutex_unlock(&rsp->barrier_mutex); -- cgit v1.2.3 From 1c09b195d37fa459844036f429a0f378e70c3db6 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 21 Aug 2013 10:22:28 +0800 Subject: cpuset: fix a regression in validating config change It's not allowed to clear masks of a cpuset if there're tasks in it, but it's broken: # mkdir /cgroup/sub # echo 0 > /cgroup/sub/cpuset.cpus # echo 0 > /cgroup/sub/cpuset.mems # echo $$ > /cgroup/sub/tasks # echo > /cgroup/sub/cpuset.cpus (should fail) This bug was introduced by commit 88fa523bff295f1d60244a54833480b02f775152 ("cpuset: allow to move tasks to empty cpusets"). tj: Dropped temp bool variables and nestes the conditionals directly. Signed-off-by: Li Zefan Signed-off-by: Tejun Heo --- kernel/cpuset.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 010a0083c0ae..ea1966db34f2 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -475,13 +475,17 @@ static int validate_change(const struct cpuset *cur, const struct cpuset *trial) /* * Cpusets with tasks - existing or newly being attached - can't - * have empty cpus_allowed or mems_allowed. + * be changed to have empty cpus_allowed or mems_allowed. */ ret = -ENOSPC; - if ((cgroup_task_count(cur->css.cgroup) || cur->attach_in_progress) && - (cpumask_empty(trial->cpus_allowed) && - nodes_empty(trial->mems_allowed))) - goto out; + if ((cgroup_task_count(cur->css.cgroup) || cur->attach_in_progress)) { + if (!cpumask_empty(cur->cpus_allowed) && + cpumask_empty(trial->cpus_allowed)) + goto out; + if (!nodes_empty(cur->mems_allowed) && + nodes_empty(trial->mems_allowed)) + goto out; + } ret = 0; out: -- cgit v1.2.3 From b11895c45899daff094610f6cdbf7611d74ae2a6 Mon Sep 17 00:00:00 2001 From: Libin Date: Wed, 21 Aug 2013 08:50:39 +0800 Subject: workqueue: Comment correction in file header No functional change. There are two worker pools for each cpu in current implementation (one for normal work items and the other for high priority ones). tj: Whitespace adjustments. Signed-off-by: Libin Signed-off-by: Tejun Heo --- kernel/workqueue.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index f02c4a4a0c3c..eebd9a66c044 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -16,9 +16,10 @@ * * This is the generic async execution mechanism. Work items as are * executed in process context. The worker pool is shared and - * automatically managed. There is one worker pool for each CPU and - * one extra for works which are better served by workers which are - * not bound to any specific CPU. + * automatically managed. There are two worker pools for each CPU (one for + * normal work items and the other for high priority ones) and some extra + * pools for workqueues which are not bound to any specific CPU - the + * number of these backing pools is dynamic. * * Please read Documentation/workqueue.txt for details. */ -- cgit v1.2.3 From 2d498db9814c6f3a79b708c8867c7ffcf7b5e2fc Mon Sep 17 00:00:00 2001 From: Libin Date: Wed, 21 Aug 2013 08:50:40 +0800 Subject: workqueue: Fix manage_workers() RETURNS description No functional change. The comment of function manage_workers() RETURNS description is obvious wrong, same as the CONTEXT. Fix it. Signed-off-by: Libin Signed-off-by: Tejun Heo --- kernel/workqueue.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index eebd9a66c044..10f655ec8de6 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -2034,8 +2034,11 @@ static bool maybe_destroy_workers(struct worker_pool *pool) * multiple times. Does GFP_KERNEL allocations. * * RETURNS: - * spin_lock_irq(pool->lock) which may be released and regrabbed - * multiple times. Does GFP_KERNEL allocations. + * %false if the pool don't need management and the caller can safely start + * processing works, %true indicates that the function released pool->lock + * and reacquired it to perform some management function and that the + * conditions that the caller verified while holding the lock before + * calling the function might no longer be true. */ static bool manage_workers(struct worker *worker) { -- cgit v1.2.3 From 3ddc77f6f4a58ee2e49e0e8c0216105c7f8ddd8c Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 27 Mar 2013 14:15:37 +0800 Subject: tracing/syscalls: Annotate raw_init function with __init init_syscall_trace() can only be called during kernel bootup only, so we can mark it and the functions it calls as __init. Link: http://lkml.kernel.org/r/51528E89.6080508@huawei.com Signed-off-by: Li Zefan Signed-off-by: Steven Rostedt --- kernel/trace/trace_syscalls.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 8fd03657bc7d..559329d9bd2f 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -200,8 +200,8 @@ extern char *__bad_type_size(void); #type, #name, offsetof(typeof(trace), name), \ sizeof(trace.name), is_signed_type(type) -static -int __set_enter_print_fmt(struct syscall_metadata *entry, char *buf, int len) +static int __init +__set_enter_print_fmt(struct syscall_metadata *entry, char *buf, int len) { int i; int pos = 0; @@ -228,7 +228,7 @@ int __set_enter_print_fmt(struct syscall_metadata *entry, char *buf, int len) return pos; } -static int set_syscall_print_fmt(struct ftrace_event_call *call) +static int __init set_syscall_print_fmt(struct ftrace_event_call *call) { char *print_fmt; int len; @@ -253,7 +253,7 @@ static int set_syscall_print_fmt(struct ftrace_event_call *call) return 0; } -static void free_syscall_print_fmt(struct ftrace_event_call *call) +static void __init free_syscall_print_fmt(struct ftrace_event_call *call) { struct syscall_metadata *entry = call->data; @@ -459,7 +459,7 @@ static void unreg_event_syscall_exit(struct ftrace_event_file *file, mutex_unlock(&syscall_trace_lock); } -static int init_syscall_trace(struct ftrace_event_call *call) +static int __init init_syscall_trace(struct ftrace_event_call *call) { int id; int num; -- cgit v1.2.3 From 779c5e379158de3e96112630c543d3c7b37efab9 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Wed, 31 Jul 2013 19:31:32 +0200 Subject: tracing: Kill trace_create_file_ops() and friends trace_create_file_ops() allocates the copy of id/filter/format/enable file_operations to set "f_op->owner = mod" for fops_get(). However after the recent changes there is no reason to prevent rmmod even if one of these files is opened. A file operation can do nothing but fail after remove_event_file_dir() clears ->i_private for every file removed by trace_module_remove_events(). Kill "struct ftrace_module_file_ops" and fix the compilation errors. Link: http://lkml.kernel.org/r/20130731173132.GA31033@redhat.com Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 153 +++----------------------------------------- 1 file changed, 9 insertions(+), 144 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 29a7ebcfb426..2ec82734b8a7 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -1683,8 +1683,7 @@ __trace_early_add_new_event(struct ftrace_event_call *call, } struct ftrace_module_file_ops; -static void __add_event_to_tracers(struct ftrace_event_call *call, - struct ftrace_module_file_ops *file_ops); +static void __add_event_to_tracers(struct ftrace_event_call *call); /* Add an additional event_call dynamically */ int trace_add_event_call(struct ftrace_event_call *call) @@ -1695,7 +1694,7 @@ int trace_add_event_call(struct ftrace_event_call *call) ret = __register_event(call, NULL); if (ret >= 0) - __add_event_to_tracers(call, NULL); + __add_event_to_tracers(call); mutex_unlock(&event_mutex); mutex_unlock(&trace_types_lock); @@ -1769,100 +1768,21 @@ int trace_remove_event_call(struct ftrace_event_call *call) #ifdef CONFIG_MODULES -static LIST_HEAD(ftrace_module_file_list); - -/* - * Modules must own their file_operations to keep up with - * reference counting. - */ -struct ftrace_module_file_ops { - struct list_head list; - struct module *mod; - struct file_operations id; - struct file_operations enable; - struct file_operations format; - struct file_operations filter; -}; - -static struct ftrace_module_file_ops * -find_ftrace_file_ops(struct ftrace_module_file_ops *file_ops, struct module *mod) -{ - /* - * As event_calls are added in groups by module, - * when we find one file_ops, we don't need to search for - * each call in that module, as the rest should be the - * same. Only search for a new one if the last one did - * not match. - */ - if (file_ops && mod == file_ops->mod) - return file_ops; - - list_for_each_entry(file_ops, &ftrace_module_file_list, list) { - if (file_ops->mod == mod) - return file_ops; - } - return NULL; -} - -static struct ftrace_module_file_ops * -trace_create_file_ops(struct module *mod) -{ - struct ftrace_module_file_ops *file_ops; - - /* - * This is a bit of a PITA. To allow for correct reference - * counting, modules must "own" their file_operations. - * To do this, we allocate the file operations that will be - * used in the event directory. - */ - - file_ops = kmalloc(sizeof(*file_ops), GFP_KERNEL); - if (!file_ops) - return NULL; - - file_ops->mod = mod; - - file_ops->id = ftrace_event_id_fops; - file_ops->id.owner = mod; - - file_ops->enable = ftrace_enable_fops; - file_ops->enable.owner = mod; - - file_ops->filter = ftrace_event_filter_fops; - file_ops->filter.owner = mod; - - file_ops->format = ftrace_event_format_fops; - file_ops->format.owner = mod; - - list_add(&file_ops->list, &ftrace_module_file_list); - - return file_ops; -} - static void trace_module_add_events(struct module *mod) { - struct ftrace_module_file_ops *file_ops = NULL; struct ftrace_event_call **call, **start, **end; start = mod->trace_events; end = mod->trace_events + mod->num_trace_events; - if (start == end) - return; - - file_ops = trace_create_file_ops(mod); - if (!file_ops) - return; - for_each_event(call, start, end) { __register_event(*call, mod); - __add_event_to_tracers(*call, file_ops); + __add_event_to_tracers(*call); } } static void trace_module_remove_events(struct module *mod) { - struct ftrace_module_file_ops *file_ops; struct ftrace_event_call *call, *p; bool clear_trace = false; @@ -1874,16 +1794,6 @@ static void trace_module_remove_events(struct module *mod) __trace_remove_event_call(call); } } - - /* Now free the file_operations */ - list_for_each_entry(file_ops, &ftrace_module_file_list, list) { - if (file_ops->mod == mod) - break; - } - if (&file_ops->list != &ftrace_module_file_list) { - list_del(&file_ops->list); - kfree(file_ops); - } up_write(&trace_event_sem); /* @@ -1919,62 +1829,22 @@ static int trace_module_notify(struct notifier_block *self, return 0; } -static int -__trace_add_new_mod_event(struct ftrace_event_call *call, - struct trace_array *tr, - struct ftrace_module_file_ops *file_ops) -{ - return __trace_add_new_event(call, tr, - &file_ops->id, &file_ops->enable, - &file_ops->filter, &file_ops->format); -} - #else -static inline struct ftrace_module_file_ops * -find_ftrace_file_ops(struct ftrace_module_file_ops *file_ops, struct module *mod) -{ - return NULL; -} static inline int trace_module_notify(struct notifier_block *self, unsigned long val, void *data) { return 0; } -static inline int -__trace_add_new_mod_event(struct ftrace_event_call *call, - struct trace_array *tr, - struct ftrace_module_file_ops *file_ops) -{ - return -ENODEV; -} #endif /* CONFIG_MODULES */ /* Create a new event directory structure for a trace directory. */ static void __trace_add_event_dirs(struct trace_array *tr) { - struct ftrace_module_file_ops *file_ops = NULL; struct ftrace_event_call *call; int ret; list_for_each_entry(call, &ftrace_events, list) { - if (call->mod) { - /* - * Directories for events by modules need to - * keep module ref counts when opened (as we don't - * want the module to disappear when reading one - * of these files). The file_ops keep account of - * the module ref count. - */ - file_ops = find_ftrace_file_ops(file_ops, call->mod); - if (!file_ops) - continue; /* Warn? */ - ret = __trace_add_new_mod_event(call, tr, file_ops); - if (ret < 0) - pr_warning("Could not create directory for event %s\n", - call->name); - continue; - } ret = __trace_add_new_event(call, tr, &ftrace_event_id_fops, &ftrace_enable_fops, @@ -2332,21 +2202,16 @@ __trace_remove_event_dirs(struct trace_array *tr) remove_event_file_dir(file); } -static void -__add_event_to_tracers(struct ftrace_event_call *call, - struct ftrace_module_file_ops *file_ops) +static void __add_event_to_tracers(struct ftrace_event_call *call) { struct trace_array *tr; list_for_each_entry(tr, &ftrace_trace_arrays, list) { - if (file_ops) - __trace_add_new_mod_event(call, tr, file_ops); - else - __trace_add_new_event(call, tr, - &ftrace_event_id_fops, - &ftrace_enable_fops, - &ftrace_event_filter_fops, - &ftrace_event_format_fops); + __trace_add_new_event(call, tr, + &ftrace_event_id_fops, + &ftrace_enable_fops, + &ftrace_event_filter_fops, + &ftrace_event_format_fops); } } -- cgit v1.2.3 From 620a30e97febc8332590376c94ed0e9dba522bc8 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Wed, 31 Jul 2013 19:31:35 +0200 Subject: tracing: Don't pass file_operations array to event_create_dir() Now that event_create_dir() and __trace_add_new_event() always use the same file_operations we can kill these arguments and simplify the code. Link: http://lkml.kernel.org/r/20130731173135.GA31040@redhat.com Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 46 ++++++++++++--------------------------------- 1 file changed, 12 insertions(+), 34 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 2ec82734b8a7..4e706a01f1f9 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -1489,12 +1489,7 @@ event_subsystem_dir(struct trace_array *tr, const char *name, } static int -event_create_dir(struct dentry *parent, - struct ftrace_event_file *file, - const struct file_operations *id, - const struct file_operations *enable, - const struct file_operations *filter, - const struct file_operations *format) +event_create_dir(struct dentry *parent, struct ftrace_event_file *file) { struct ftrace_event_call *call = file->event_call; struct trace_array *tr = file->tr; @@ -1522,12 +1517,13 @@ event_create_dir(struct dentry *parent, if (call->class->reg && !(call->flags & TRACE_EVENT_FL_IGNORE_ENABLE)) trace_create_file("enable", 0644, file->dir, file, - enable); + &ftrace_enable_fops); #ifdef CONFIG_PERF_EVENTS if (call->event.type && call->class->reg) trace_create_file("id", 0444, file->dir, - (void *)(long)call->event.type, id); + (void *)(long)call->event.type, + &ftrace_event_id_fops); #endif /* @@ -1544,10 +1540,10 @@ event_create_dir(struct dentry *parent, } } trace_create_file("filter", 0644, file->dir, call, - filter); + &ftrace_event_filter_fops); trace_create_file("format", 0444, file->dir, call, - format); + &ftrace_event_format_fops); return 0; } @@ -1648,12 +1644,7 @@ trace_create_new_event(struct ftrace_event_call *call, /* Add an event to a trace directory */ static int -__trace_add_new_event(struct ftrace_event_call *call, - struct trace_array *tr, - const struct file_operations *id, - const struct file_operations *enable, - const struct file_operations *filter, - const struct file_operations *format) +__trace_add_new_event(struct ftrace_event_call *call, struct trace_array *tr) { struct ftrace_event_file *file; @@ -1661,7 +1652,7 @@ __trace_add_new_event(struct ftrace_event_call *call, if (!file) return -ENOMEM; - return event_create_dir(tr->event_dir, file, id, enable, filter, format); + return event_create_dir(tr->event_dir, file); } /* @@ -1845,11 +1836,7 @@ __trace_add_event_dirs(struct trace_array *tr) int ret; list_for_each_entry(call, &ftrace_events, list) { - ret = __trace_add_new_event(call, tr, - &ftrace_event_id_fops, - &ftrace_enable_fops, - &ftrace_event_filter_fops, - &ftrace_event_format_fops); + ret = __trace_add_new_event(call, tr); if (ret < 0) pr_warning("Could not create directory for event %s\n", call->name); @@ -2157,11 +2144,7 @@ __trace_early_add_event_dirs(struct trace_array *tr) list_for_each_entry(file, &tr->events, list) { - ret = event_create_dir(tr->event_dir, file, - &ftrace_event_id_fops, - &ftrace_enable_fops, - &ftrace_event_filter_fops, - &ftrace_event_format_fops); + ret = event_create_dir(tr->event_dir, file); if (ret < 0) pr_warning("Could not create directory for event %s\n", file->event_call->name); @@ -2206,13 +2189,8 @@ static void __add_event_to_tracers(struct ftrace_event_call *call) { struct trace_array *tr; - list_for_each_entry(tr, &ftrace_trace_arrays, list) { - __trace_add_new_event(call, tr, - &ftrace_event_id_fops, - &ftrace_enable_fops, - &ftrace_event_filter_fops, - &ftrace_event_format_fops); - } + list_for_each_entry(tr, &ftrace_trace_arrays, list) + __trace_add_new_event(call, tr); } static struct notifier_block trace_module_nb = { -- cgit v1.2.3 From 836d481ed7c91152c6144ea3a3363cad3940b3e0 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Wed, 31 Jul 2013 19:31:37 +0200 Subject: tracing: Kill the !CONFIG_MODULES code in trace_events.c Move trace_module_nb under CONFIG_MODULES and kill the dummy trace_module_notify(). Imho it doesn't make sense to define "struct notifier_block" and its .notifier_call just to avoid "ifdef" in event_trace_init(), and all other !CONFIG_MODULES code has already gone away. Link: http://lkml.kernel.org/r/20130731173137.GA31043@redhat.com Signed-off-by: Oleg Nesterov Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 18 ++++++------------ 1 file changed, 6 insertions(+), 12 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 4e706a01f1f9..368a4d50cc30 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -1820,12 +1820,10 @@ static int trace_module_notify(struct notifier_block *self, return 0; } -#else -static inline int trace_module_notify(struct notifier_block *self, - unsigned long val, void *data) -{ - return 0; -} +static struct notifier_block trace_module_nb = { + .notifier_call = trace_module_notify, + .priority = 0, +}; #endif /* CONFIG_MODULES */ /* Create a new event directory structure for a trace directory. */ @@ -2193,11 +2191,6 @@ static void __add_event_to_tracers(struct ftrace_event_call *call) __trace_add_new_event(call, tr); } -static struct notifier_block trace_module_nb = { - .notifier_call = trace_module_notify, - .priority = 0, -}; - extern struct ftrace_event_call *__start_ftrace_events[]; extern struct ftrace_event_call *__stop_ftrace_events[]; @@ -2402,10 +2395,11 @@ static __init int event_trace_init(void) if (ret) return ret; +#ifdef CONFIG_MODULES ret = register_module_notifier(&trace_module_nb); if (ret) pr_warning("Failed to register trace events module notifier\n"); - +#endif return 0; } early_initcall(event_trace_memsetup); -- cgit v1.2.3 From ccfe9e42e451232dd17a230d1b4e979c3d15311e Mon Sep 17 00:00:00 2001 From: Alexander Z Lam Date: Thu, 8 Aug 2013 09:47:45 -0700 Subject: tracing: Make tracing_cpumask available for all instances Allow tracer instances to disable tracing by cpu by moving the static global tracing_cpumask into trace_array. Link: http://lkml.kernel.org/r/921622317f239bfc2283cac2242647801ef584f2.1375980149.git.azl@google.com Cc: Vaibhav Nagarnaik Cc: David Sharp Cc: Alexander Z Lam Signed-off-by: Alexander Z Lam Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 37 ++++++++++++++++++++----------------- kernel/trace/trace.h | 1 + 2 files changed, 21 insertions(+), 17 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 496f94d57698..7974ba20557d 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -3165,11 +3165,6 @@ static const struct file_operations show_traces_fops = { .llseek = seq_lseek, }; -/* - * Only trace on a CPU if the bitmask is set: - */ -static cpumask_var_t tracing_cpumask; - /* * The tracer itself will not take this lock, but still we want * to provide a consistent cpumask to user-space: @@ -3186,11 +3181,12 @@ static ssize_t tracing_cpumask_read(struct file *filp, char __user *ubuf, size_t count, loff_t *ppos) { + struct trace_array *tr = file_inode(filp)->i_private; int len; mutex_lock(&tracing_cpumask_update_lock); - len = cpumask_scnprintf(mask_str, count, tracing_cpumask); + len = cpumask_scnprintf(mask_str, count, tr->tracing_cpumask); if (count - len < 2) { count = -EINVAL; goto out_err; @@ -3208,7 +3204,7 @@ static ssize_t tracing_cpumask_write(struct file *filp, const char __user *ubuf, size_t count, loff_t *ppos) { - struct trace_array *tr = filp->private_data; + struct trace_array *tr = file_inode(filp)->i_private; cpumask_var_t tracing_cpumask_new; int err, cpu; @@ -3228,12 +3224,12 @@ tracing_cpumask_write(struct file *filp, const char __user *ubuf, * Increase/decrease the disabled counter if we are * about to flip a bit in the cpumask: */ - if (cpumask_test_cpu(cpu, tracing_cpumask) && + if (cpumask_test_cpu(cpu, tr->tracing_cpumask) && !cpumask_test_cpu(cpu, tracing_cpumask_new)) { atomic_inc(&per_cpu_ptr(tr->trace_buffer.data, cpu)->disabled); ring_buffer_record_disable_cpu(tr->trace_buffer.buffer, cpu); } - if (!cpumask_test_cpu(cpu, tracing_cpumask) && + if (!cpumask_test_cpu(cpu, tr->tracing_cpumask) && cpumask_test_cpu(cpu, tracing_cpumask_new)) { atomic_dec(&per_cpu_ptr(tr->trace_buffer.data, cpu)->disabled); ring_buffer_record_enable_cpu(tr->trace_buffer.buffer, cpu); @@ -3242,7 +3238,7 @@ tracing_cpumask_write(struct file *filp, const char __user *ubuf, arch_spin_unlock(&ftrace_max_lock); local_irq_enable(); - cpumask_copy(tracing_cpumask, tracing_cpumask_new); + cpumask_copy(tr->tracing_cpumask, tracing_cpumask_new); mutex_unlock(&tracing_cpumask_update_lock); free_cpumask_var(tracing_cpumask_new); @@ -3256,9 +3252,10 @@ err_unlock: } static const struct file_operations tracing_cpumask_fops = { - .open = tracing_open_generic, + .open = tracing_open_generic_tr, .read = tracing_cpumask_read, .write = tracing_cpumask_write, + .release = tracing_release_generic_tr, .llseek = generic_file_llseek, }; @@ -5938,6 +5935,11 @@ static int new_instance_create(const char *name) if (!tr->name) goto out_free_tr; + if (!alloc_cpumask_var(&tr->tracing_cpumask, GFP_KERNEL)) + goto out_free_tr; + + cpumask_copy(tr->tracing_cpumask, cpu_all_mask); + raw_spin_lock_init(&tr->start_lock); tr->current_trace = &nop_trace; @@ -5969,6 +5971,7 @@ static int new_instance_create(const char *name) out_free_tr: if (tr->trace_buffer.buffer) ring_buffer_free(tr->trace_buffer.buffer); + free_cpumask_var(tr->tracing_cpumask); kfree(tr->name); kfree(tr); @@ -6098,6 +6101,9 @@ init_tracer_debugfs(struct trace_array *tr, struct dentry *d_tracer) { int cpu; + trace_create_file("tracing_cpumask", 0644, d_tracer, + tr, &tracing_cpumask_fops); + trace_create_file("trace_options", 0644, d_tracer, tr, &tracing_iter_fops); @@ -6147,9 +6153,6 @@ static __init int tracer_init_debugfs(void) init_tracer_debugfs(&global_trace, d_tracer); - trace_create_file("tracing_cpumask", 0644, d_tracer, - &global_trace, &tracing_cpumask_fops); - trace_create_file("available_tracers", 0444, d_tracer, &global_trace, &show_traces_fops); @@ -6371,7 +6374,7 @@ __init static int tracer_alloc_buffers(void) if (!alloc_cpumask_var(&tracing_buffer_mask, GFP_KERNEL)) goto out; - if (!alloc_cpumask_var(&tracing_cpumask, GFP_KERNEL)) + if (!alloc_cpumask_var(&global_trace.tracing_cpumask, GFP_KERNEL)) goto out_free_buffer_mask; /* Only allocate trace_printk buffers if a trace_printk exists */ @@ -6386,7 +6389,7 @@ __init static int tracer_alloc_buffers(void) ring_buf_size = 1; cpumask_copy(tracing_buffer_mask, cpu_possible_mask); - cpumask_copy(tracing_cpumask, cpu_all_mask); + cpumask_copy(global_trace.tracing_cpumask, cpu_all_mask); raw_spin_lock_init(&global_trace.start_lock); @@ -6441,7 +6444,7 @@ out_free_cpumask: #ifdef CONFIG_TRACER_MAX_TRACE free_percpu(global_trace.max_buffer.data); #endif - free_cpumask_var(tracing_cpumask); + free_cpumask_var(global_trace.tracing_cpumask); out_free_buffer_mask: free_cpumask_var(tracing_buffer_mask); out: diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index afaae41b0a02..502fed770751 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -206,6 +206,7 @@ struct trace_array { struct dentry *event_dir; struct list_head systems; struct list_head events; + cpumask_var_t tracing_cpumask; /* only trace on set CPUs */ int ref; }; -- cgit v1.2.3 From 1a6661dafd2528d03d0eaed898ad596816dfe738 Mon Sep 17 00:00:00 2001 From: Greg Kroah-Hartman Date: Fri, 23 Aug 2013 14:24:41 -0700 Subject: workqueue: convert bus code to use dev_groups The dev_attrs field of struct bus_type is going away soon, dev_groups should be used instead. This converts the workqueue bus code to use the correct field. Acked-by: Tejun Heo Signed-off-by: Greg Kroah-Hartman --- kernel/workqueue.c | 27 +++++++++++++++------------ 1 file changed, 15 insertions(+), 12 deletions(-) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 0b72e816b8d0..d1b5f0662651 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -3081,25 +3081,26 @@ static struct workqueue_struct *dev_to_wq(struct device *dev) return wq_dev->wq; } -static ssize_t wq_per_cpu_show(struct device *dev, - struct device_attribute *attr, char *buf) +static ssize_t per_cpu_show(struct device *dev, struct device_attribute *attr, + char *buf) { struct workqueue_struct *wq = dev_to_wq(dev); return scnprintf(buf, PAGE_SIZE, "%d\n", (bool)!(wq->flags & WQ_UNBOUND)); } +static DEVICE_ATTR_RO(per_cpu); -static ssize_t wq_max_active_show(struct device *dev, - struct device_attribute *attr, char *buf) +static ssize_t max_active_show(struct device *dev, + struct device_attribute *attr, char *buf) { struct workqueue_struct *wq = dev_to_wq(dev); return scnprintf(buf, PAGE_SIZE, "%d\n", wq->saved_max_active); } -static ssize_t wq_max_active_store(struct device *dev, - struct device_attribute *attr, - const char *buf, size_t count) +static ssize_t max_active_store(struct device *dev, + struct device_attribute *attr, const char *buf, + size_t count) { struct workqueue_struct *wq = dev_to_wq(dev); int val; @@ -3110,12 +3111,14 @@ static ssize_t wq_max_active_store(struct device *dev, workqueue_set_max_active(wq, val); return count; } +static DEVICE_ATTR_RW(max_active); -static struct device_attribute wq_sysfs_attrs[] = { - __ATTR(per_cpu, 0444, wq_per_cpu_show, NULL), - __ATTR(max_active, 0644, wq_max_active_show, wq_max_active_store), - __ATTR_NULL, +static struct attribute *wq_sysfs_attrs[] = { + &dev_attr_per_cpu.attr, + &dev_attr_max_active.attr, + NULL, }; +ATTRIBUTE_GROUPS(wq_sysfs); static ssize_t wq_pool_ids_show(struct device *dev, struct device_attribute *attr, char *buf) @@ -3265,7 +3268,7 @@ static struct device_attribute wq_sysfs_unbound_attrs[] = { static struct bus_type wq_subsys = { .name = "workqueue", - .dev_attrs = wq_sysfs_attrs, + .dev_groups = wq_sysfs_groups, }; static int __init wq_sysfs_init(void) -- cgit v1.2.3 From 35cf083619da5677f83e9a8eae813f0b413d7082 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 26 Aug 2013 18:40:56 -0400 Subject: cgroup: rename cgroup_css_from_dir() to css_from_dir() and update its syntax cgroup_css_from_dir() will grow another user. In preparation, make the following changes. * All css functions are prefixed with just "css_", rename it to css_from_dir(). * Take dentry * instead of file * as dentry is what ultimately identifies a cgroup and file may not always be available. Note that the function now checkes whether @dentry->d_inode is NULL as the caller now may specify a negative dentry. * Make it take cgroup_subsys * instead of integer subsys_id. This simplifies the function and allows specifying no subsystem for cgroup->dummy_css. * Make return section a bit less verbose. This patch doesn't introduce any behavior changes. Signed-off-by: Tejun Heo Acked-by: Li Zefan Acked-by: Kirill A. Shutemov Cc: Steven Rostedt Cc: Frederic Weisbecker Cc: Ingo Molnar --- kernel/cgroup.c | 26 ++++++++++---------------- kernel/events/core.c | 2 +- 2 files changed, 11 insertions(+), 17 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index ef43e3f453ef..921b1387c944 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -5700,34 +5700,28 @@ struct cgroup_subsys_state *css_lookup(struct cgroup_subsys *ss, int id) EXPORT_SYMBOL_GPL(css_lookup); /** - * cgroup_css_from_dir - get corresponding css from file open on cgroup dir - * @f: directory file of interest - * @id: subsystem id of interest + * css_from_dir - get corresponding css from the dentry of a cgroup dir + * @dentry: directory dentry of interest + * @ss: subsystem of interest * * Must be called under RCU read lock. The caller is responsible for * pinning the returned css if it needs to be accessed outside the RCU * critical section. */ -struct cgroup_subsys_state *cgroup_css_from_dir(struct file *f, int id) +struct cgroup_subsys_state *css_from_dir(struct dentry *dentry, + struct cgroup_subsys *ss) { struct cgroup *cgrp; - struct inode *inode; - struct cgroup_subsys_state *css; WARN_ON_ONCE(!rcu_read_lock_held()); - inode = file_inode(f); - /* check in cgroup filesystem dir */ - if (inode->i_op != &cgroup_dir_inode_operations) + /* is @dentry a cgroup dir? */ + if (!dentry->d_inode || + dentry->d_inode->i_op != &cgroup_dir_inode_operations) return ERR_PTR(-EBADF); - if (id < 0 || id >= CGROUP_SUBSYS_COUNT) - return ERR_PTR(-EINVAL); - - /* get cgroup */ - cgrp = __d_cgrp(f->f_dentry); - css = cgroup_css(cgrp, id); - return css ? css : ERR_PTR(-ENOENT); + cgrp = __d_cgrp(dentry); + return cgroup_css(cgrp, ss->subsys_id) ?: ERR_PTR(-ENOENT); } /** diff --git a/kernel/events/core.c b/kernel/events/core.c index 23261f957713..b59ab6632f30 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -593,7 +593,7 @@ static inline int perf_cgroup_connect(int fd, struct perf_event *event, rcu_read_lock(); - css = cgroup_css_from_dir(f.file, perf_subsys_id); + css = css_from_dir(f.file->f_dentry, &perf_subsys); if (IS_ERR(css)) { ret = PTR_ERR(css); goto out; -- cgit v1.2.3 From ca8bdcaff0d77990fb69e0f946018c96a70851cc Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 26 Aug 2013 18:40:56 -0400 Subject: cgroup: make cgroup_css() take cgroup_subsys * instead and allow NULL subsys cgroup_css() is no longer used in hot paths. Make it take struct cgroup_subsys * and allow the users to specify NULL subsys to obtain the dummy_css. This removes open-coded NULL subsystem testing in a couple users and generally simplifies the code. After this patch, css_from_dir() also allows NULL @ss and returns the matching dummy_css. This behavior change doesn't affect its only user - perf. Signed-off-by: Tejun Heo Acked-by: Li Zefan Acked-by: Kirill A. Shutemov --- kernel/cgroup.c | 90 +++++++++++++++++++++++++++------------------------------ 1 file changed, 43 insertions(+), 47 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 921b1387c944..7516668d8325 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -226,19 +226,22 @@ static int cgroup_addrm_files(struct cgroup *cgrp, struct cftype cfts[], /** * cgroup_css - obtain a cgroup's css for the specified subsystem * @cgrp: the cgroup of interest - * @subsys_id: the subsystem of interest + * @ss: the subsystem of interest (%NULL returns the dummy_css) * - * Return @cgrp's css (cgroup_subsys_state) associated with @subsys_id. - * This function must be called either under cgroup_mutex or - * rcu_read_lock() and the caller is responsible for pinning the returned - * css if it wants to keep accessing it outside the said locks. This - * function may return %NULL if @cgrp doesn't have @subsys_id enabled. + * Return @cgrp's css (cgroup_subsys_state) associated with @ss. This + * function must be called either under cgroup_mutex or rcu_read_lock() and + * the caller is responsible for pinning the returned css if it wants to + * keep accessing it outside the said locks. This function may return + * %NULL if @cgrp doesn't have @subsys_id enabled. */ static struct cgroup_subsys_state *cgroup_css(struct cgroup *cgrp, - int subsys_id) + struct cgroup_subsys *ss) { - return rcu_dereference_check(cgrp->subsys[subsys_id], - lockdep_is_held(&cgroup_mutex)); + if (ss) + return rcu_dereference_check(cgrp->subsys[ss->subsys_id], + lockdep_is_held(&cgroup_mutex)); + else + return &cgrp->dummy_css; } /* convenient tests for these bits */ @@ -580,7 +583,7 @@ static struct css_set *find_existing_css_set(struct css_set *old_cset, /* Subsystem is in this hierarchy. So we want * the subsystem state from the new * cgroup */ - template[i] = cgroup_css(cgrp, i); + template[i] = cgroup_css(cgrp, ss); } else { /* Subsystem is not in this hierarchy, so we * don't want to change the subsystem state */ @@ -1062,30 +1065,30 @@ static int rebind_subsystems(struct cgroupfs_root *root, if (bit & added_mask) { /* We're binding this subsystem to this hierarchy */ - BUG_ON(cgroup_css(cgrp, i)); - BUG_ON(!cgroup_css(cgroup_dummy_top, i)); - BUG_ON(cgroup_css(cgroup_dummy_top, i)->cgroup != cgroup_dummy_top); + BUG_ON(cgroup_css(cgrp, ss)); + BUG_ON(!cgroup_css(cgroup_dummy_top, ss)); + BUG_ON(cgroup_css(cgroup_dummy_top, ss)->cgroup != cgroup_dummy_top); rcu_assign_pointer(cgrp->subsys[i], - cgroup_css(cgroup_dummy_top, i)); - cgroup_css(cgrp, i)->cgroup = cgrp; + cgroup_css(cgroup_dummy_top, ss)); + cgroup_css(cgrp, ss)->cgroup = cgrp; list_move(&ss->sibling, &root->subsys_list); ss->root = root; if (ss->bind) - ss->bind(cgroup_css(cgrp, i)); + ss->bind(cgroup_css(cgrp, ss)); /* refcount was already taken, and we're keeping it */ root->subsys_mask |= bit; } else if (bit & removed_mask) { /* We're removing this subsystem */ - BUG_ON(cgroup_css(cgrp, i) != cgroup_css(cgroup_dummy_top, i)); - BUG_ON(cgroup_css(cgrp, i)->cgroup != cgrp); + BUG_ON(cgroup_css(cgrp, ss) != cgroup_css(cgroup_dummy_top, ss)); + BUG_ON(cgroup_css(cgrp, ss)->cgroup != cgrp); if (ss->bind) - ss->bind(cgroup_css(cgroup_dummy_top, i)); + ss->bind(cgroup_css(cgroup_dummy_top, ss)); - cgroup_css(cgroup_dummy_top, i)->cgroup = cgroup_dummy_top; + cgroup_css(cgroup_dummy_top, ss)->cgroup = cgroup_dummy_top; RCU_INIT_POINTER(cgrp->subsys[i], NULL); cgroup_subsys[i]->root = &cgroup_dummy_root; @@ -1930,7 +1933,7 @@ EXPORT_SYMBOL_GPL(cgroup_taskset_next); struct cgroup_subsys_state *cgroup_taskset_cur_css(struct cgroup_taskset *tset, int subsys_id) { - return cgroup_css(tset->cur_cgrp, subsys_id); + return cgroup_css(tset->cur_cgrp, cgroup_subsys[subsys_id]); } EXPORT_SYMBOL_GPL(cgroup_taskset_cur_css); @@ -2071,7 +2074,7 @@ static int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk, * step 1: check that we can legitimately attach to the cgroup. */ for_each_root_subsys(root, ss) { - struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); + struct cgroup_subsys_state *css = cgroup_css(cgrp, ss); if (ss->can_attach) { retval = ss->can_attach(css, &tset); @@ -2113,7 +2116,7 @@ static int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk, * step 4: do subsystem attach callbacks. */ for_each_root_subsys(root, ss) { - struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); + struct cgroup_subsys_state *css = cgroup_css(cgrp, ss); if (ss->attach) ss->attach(css, &tset); @@ -2135,7 +2138,7 @@ out_put_css_set_refs: out_cancel_attach: if (retval) { for_each_root_subsys(root, ss) { - struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); + struct cgroup_subsys_state *css = cgroup_css(cgrp, ss); if (ss == failed_ss) break; @@ -2481,13 +2484,9 @@ static int cgroup_file_open(struct inode *inode, struct file *file) * @css stays alive for all file operations. */ rcu_read_lock(); - if (cft->ss) { - css = cgroup_css(cgrp, cft->ss->subsys_id); - if (!css_tryget(css)) - css = NULL; - } else { - css = &cgrp->dummy_css; - } + css = cgroup_css(cgrp, cft->ss); + if (cft->ss && !css_tryget(css)) + css = NULL; rcu_read_unlock(); if (!css) @@ -2878,7 +2877,7 @@ static int cgroup_cfts_commit(struct cftype *cfts, bool is_add) /* add/rm files for all cgroups created before */ rcu_read_lock(); - css_for_each_descendant_pre(css, cgroup_css(root, ss->subsys_id)) { + css_for_each_descendant_pre(css, cgroup_css(root, ss)) { struct cgroup *cgrp = css->cgroup; if (cgroup_is_dead(cgrp)) @@ -3082,10 +3081,7 @@ css_next_child(struct cgroup_subsys_state *pos_css, if (&next->sibling == &cgrp->children) return NULL; - if (parent_css->ss) - return cgroup_css(next, parent_css->ss->subsys_id); - else - return &next->dummy_css; + return cgroup_css(next, parent_css->ss); } EXPORT_SYMBOL_GPL(css_next_child); @@ -4110,7 +4106,7 @@ static int cgroup_write_event_control(struct cgroup_subsys_state *dummy_css, rcu_read_lock(); ret = -EINVAL; - event->css = cgroup_css(cgrp, event->cft->ss->subsys_id); + event->css = cgroup_css(cgrp, event->cft->ss); if (event->css) ret = 0; @@ -4266,7 +4262,7 @@ static int cgroup_populate_dir(struct cgroup *cgrp, unsigned long subsys_mask) /* This cgroup is ready now */ for_each_root_subsys(cgrp->root, ss) { - struct cgroup_subsys_state *css = cgroup_css(cgrp, ss->subsys_id); + struct cgroup_subsys_state *css = cgroup_css(cgrp, ss); struct css_id *id = rcu_dereference_protected(css->id, true); /* @@ -4349,11 +4345,11 @@ static void init_css(struct cgroup_subsys_state *css, struct cgroup_subsys *ss, css->id = NULL; if (cgrp->parent) - css->parent = cgroup_css(cgrp->parent, ss->subsys_id); + css->parent = cgroup_css(cgrp->parent, ss); else css->flags |= CSS_ROOT; - BUG_ON(cgroup_css(cgrp, ss->subsys_id)); + BUG_ON(cgroup_css(cgrp, ss)); } /* invoke ->css_online() on a new CSS and mark it online if successful */ @@ -4466,7 +4462,7 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, for_each_root_subsys(root, ss) { struct cgroup_subsys_state *css; - css = ss->css_alloc(cgroup_css(parent, ss->subsys_id)); + css = ss->css_alloc(cgroup_css(parent, ss)); if (IS_ERR(css)) { err = PTR_ERR(css); goto err_free_all; @@ -4712,7 +4708,7 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) * percpu refs of all css's are confirmed to be killed. */ for_each_root_subsys(cgrp->root, ss) - kill_css(cgroup_css(cgrp, ss->subsys_id)); + kill_css(cgroup_css(cgrp, ss)); /* * Mark @cgrp dead. This prevents further task migration and child @@ -4839,7 +4835,7 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss) /* Create the top cgroup state for this subsystem */ list_add(&ss->sibling, &cgroup_dummy_root.subsys_list); ss->root = &cgroup_dummy_root; - css = ss->css_alloc(cgroup_css(cgroup_dummy_top, ss->subsys_id)); + css = ss->css_alloc(cgroup_css(cgroup_dummy_top, ss)); /* We don't handle early failures gracefully */ BUG_ON(IS_ERR(css)); init_css(css, ss, cgroup_dummy_top); @@ -4918,7 +4914,7 @@ int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss) * struct, so this can happen first (i.e. before the dummy root * attachment). */ - css = ss->css_alloc(cgroup_css(cgroup_dummy_top, ss->subsys_id)); + css = ss->css_alloc(cgroup_css(cgroup_dummy_top, ss)); if (IS_ERR(css)) { /* failure case - need to deassign the cgroup_subsys[] slot. */ cgroup_subsys[ss->subsys_id] = NULL; @@ -5000,7 +4996,7 @@ void cgroup_unload_subsys(struct cgroup_subsys *ss) mutex_lock(&cgroup_mutex); - offline_css(cgroup_css(cgroup_dummy_top, ss->subsys_id)); + offline_css(cgroup_css(cgroup_dummy_top, ss)); if (ss->use_id) idr_destroy(&ss->idr); @@ -5034,7 +5030,7 @@ void cgroup_unload_subsys(struct cgroup_subsys *ss) * the cgrp->subsys pointer to find their state. note that this * also takes care of freeing the css_id. */ - ss->css_free(cgroup_css(cgroup_dummy_top, ss->subsys_id)); + ss->css_free(cgroup_css(cgroup_dummy_top, ss)); RCU_INIT_POINTER(cgroup_dummy_top->subsys[ss->subsys_id], NULL); mutex_unlock(&cgroup_mutex); @@ -5721,7 +5717,7 @@ struct cgroup_subsys_state *css_from_dir(struct dentry *dentry, return ERR_PTR(-EBADF); cgrp = __d_cgrp(dentry); - return cgroup_css(cgrp, ss->subsys_id) ?: ERR_PTR(-ENOENT); + return cgroup_css(cgrp, ss) ?: ERR_PTR(-ENOENT); } /** -- cgit v1.2.3 From 9fa4db334c7d9570aec7a5121e84fae99aae1d04 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 26 Aug 2013 18:40:56 -0400 Subject: cgroup: implement CFTYPE_NO_PREFIX When cgroup files are created, cgroup core automatically prepends the name of the subsystem as prefix. This patch adds CFTYPE_NO_ which disables the automatic prefix. This is to work around historical baggages and shouldn't be used for new files. This will be used to move "cgroup.event_control" from cgroup core to memcg. Signed-off-by: Tejun Heo Acked-by: Li Zefan Acked-by: Kirill A. Shutemov Cc: Glauber Costa --- kernel/cgroup.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 7516668d8325..a41dc87cd07e 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -2756,7 +2756,8 @@ static int cgroup_add_file(struct cgroup *cgrp, struct cftype *cft) umode_t mode; char name[MAX_CGROUP_TYPE_NAMELEN + MAX_CFTYPE_NAME + 2] = { 0 }; - if (cft->ss && !(cgrp->root->flags & CGRP_ROOT_NOPREFIX)) { + if (cft->ss && !(cft->flags & CFTYPE_NO_PREFIX) && + !(cgrp->root->flags & CGRP_ROOT_NOPREFIX)) { strcpy(name, cft->ss->name); strcat(name, "."); } -- cgit v1.2.3 From 7941cb027dccedec3c047271554ddcf4be2e0697 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 26 Aug 2013 18:40:56 -0400 Subject: cgroup: make cgroup_event hold onto cgroup_subsys_state instead of cgroup Currently, each registered cgroup_event holds an extra reference to the cgroup. This is a bit weird as events are subsystem specific and will also be incorrect in the planned unified hierarchy as css (cgroup_subsys_state) may come and go dynamically across the lifetime of a cgroup. Holding onto cgroup won't prevent the target css from going away. Update cgroup_event to hold onto the css the traget file belongs to instead of cgroup. Signed-off-by: Tejun Heo Acked-by: Li Zefan Acked-by: Kirill A. Shutemov --- kernel/cgroup.c | 26 ++++++++++++-------------- 1 file changed, 12 insertions(+), 14 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index a41dc87cd07e..12237a291d88 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -3969,7 +3969,6 @@ static void cgroup_event_remove(struct work_struct *work) struct cgroup_event *event = container_of(work, struct cgroup_event, remove); struct cgroup_subsys_state *css = event->css; - struct cgroup *cgrp = css->cgroup; remove_wait_queue(event->wqh, &event->wait); @@ -3980,7 +3979,7 @@ static void cgroup_event_remove(struct work_struct *work) eventfd_ctx_put(event->eventfd); kfree(event); - cgroup_dput(cgrp); + css_put(css); } /* @@ -4103,12 +4102,16 @@ static int cgroup_write_event_control(struct cgroup_subsys_state *dummy_css, goto out_put_cfile; } - /* determine the css of @cfile and associate @event with it */ + /* + * Determine the css of @cfile and associate @event with it. + * Remaining events are automatically removed on cgroup destruction + * but the removal is asynchronous, so take an extra ref. + */ rcu_read_lock(); ret = -EINVAL; event->css = cgroup_css(cgrp, event->cft->ss); - if (event->css) + if (event->css && css_tryget(event->css)) ret = 0; rcu_read_unlock(); @@ -4122,28 +4125,21 @@ static int cgroup_write_event_control(struct cgroup_subsys_state *dummy_css, cgrp_cfile = __d_cgrp(cfile->f_dentry->d_parent); if (cgrp_cfile != cgrp) { ret = -EINVAL; - goto out_put_cfile; + goto out_put_css; } if (!event->cft->register_event || !event->cft->unregister_event) { ret = -EINVAL; - goto out_put_cfile; + goto out_put_css; } ret = event->cft->register_event(event->css, event->cft, event->eventfd, buffer); if (ret) - goto out_put_cfile; + goto out_put_css; efile->f_op->poll(efile, &event->pt); - /* - * Events should be removed after rmdir of cgroup directory, but before - * destroying subsystem state objects. Let's take reference to cgroup - * directory dentry to do that. - */ - dget(cgrp->dentry); - spin_lock(&cgrp->event_list_lock); list_add(&event->list, &cgrp->event_list); spin_unlock(&cgrp->event_list_lock); @@ -4153,6 +4149,8 @@ static int cgroup_write_event_control(struct cgroup_subsys_state *dummy_css, return 0; +out_put_css: + css_put(event->css); out_put_cfile: fput(cfile); out_put_eventfd: -- cgit v1.2.3 From 7c918cbbd829669bf70ffcc45962d5d992942243 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 26 Aug 2013 18:40:56 -0400 Subject: cgroup: make cgroup_write_event_control() use css_from_dir() instead of __d_cgrp() cgroup_event will be moved to its only user - memcg. Replace __d_cgrp() usage with css_from_dir(), which is already exported. This also simplifies the code a bit. Signed-off-by: Tejun Heo Acked-by: Li Zefan Acked-by: Kirill A. Shutemov --- kernel/cgroup.c | 18 +++++------------- 1 file changed, 5 insertions(+), 13 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 12237a291d88..e76698dd6c08 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4041,7 +4041,7 @@ static int cgroup_write_event_control(struct cgroup_subsys_state *dummy_css, { struct cgroup *cgrp = dummy_css->cgroup; struct cgroup_event *event; - struct cgroup *cgrp_cfile; + struct cgroup_subsys_state *cfile_css; unsigned int efd, cfd; struct file *efile; struct file *cfile; @@ -4103,7 +4103,8 @@ static int cgroup_write_event_control(struct cgroup_subsys_state *dummy_css, } /* - * Determine the css of @cfile and associate @event with it. + * Determine the css of @cfile, verify it belongs to the same + * cgroup as cgroup.event_control, and associate @event with it. * Remaining events are automatically removed on cgroup destruction * but the removal is asynchronous, so take an extra ref. */ @@ -4111,23 +4112,14 @@ static int cgroup_write_event_control(struct cgroup_subsys_state *dummy_css, ret = -EINVAL; event->css = cgroup_css(cgrp, event->cft->ss); - if (event->css && css_tryget(event->css)) + cfile_css = css_from_dir(cfile->f_dentry->d_parent, event->cft->ss); + if (event->css && event->css == cfile_css && css_tryget(event->css)) ret = 0; rcu_read_unlock(); if (ret) goto out_put_cfile; - /* - * The file to be monitored must be in the same cgroup as - * cgroup.event_control is. - */ - cgrp_cfile = __d_cgrp(cfile->f_dentry->d_parent); - if (cgrp_cfile != cgrp) { - ret = -EINVAL; - goto out_put_css; - } - if (!event->cft->register_event || !event->cft->unregister_event) { ret = -EINVAL; goto out_put_css; -- cgit v1.2.3 From 21e851943e31022731cd5fad386ca8fb552dbe64 Mon Sep 17 00:00:00 2001 From: "Raphael S.Carvalho" Date: Wed, 27 Feb 2013 15:32:09 -0300 Subject: kernel/nsproxy.c: Improving a snippet of code. It seems GCC generates a better code in that way, so I changed that statement. Btw, they have the same semantic, so I'm sending this patch due to performance issues. Acked-by: Serge E. Hallyn Signed-off-by: Raphael S.Carvalho Signed-off-by: Eric W. Biederman --- kernel/nsproxy.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/nsproxy.c b/kernel/nsproxy.c index 364ceab15f0c..d9afd256318f 100644 --- a/kernel/nsproxy.c +++ b/kernel/nsproxy.c @@ -148,7 +148,8 @@ int copy_namespaces(unsigned long flags, struct task_struct *tsk) * means share undolist with parent, so we must forbid using * it along with CLONE_NEWIPC. */ - if ((flags & CLONE_NEWIPC) && (flags & CLONE_SYSVSEM)) { + if ((flags & (CLONE_NEWIPC | CLONE_SYSVSEM)) == + (CLONE_NEWIPC | CLONE_SYSVSEM)) { err = -EINVAL; goto out; } -- cgit v1.2.3 From e51db73532955dc5eaba4235e62b74b460709d5b Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Sat, 30 Mar 2013 19:57:41 -0700 Subject: userns: Better restrictions on when proc and sysfs can be mounted Rely on the fact that another flavor of the filesystem is already mounted and do not rely on state in the user namespace. Verify that the mounted filesystem is not covered in any significant way. I would love to verify that the previously mounted filesystem has no mounts on top but there are at least the directories /proc/sys/fs/binfmt_misc and /sys/fs/cgroup/ that exist explicitly for other filesystems to mount on top of. Refactor the test into a function named fs_fully_visible and call that function from the mount routines of proc and sysfs. This makes this test local to the filesystems involved and the results current of when the mounts take place, removing a weird threading of the user namespace, the mount namespace and the filesystems themselves. Signed-off-by: "Eric W. Biederman" --- kernel/user.c | 2 -- kernel/user_namespace.c | 2 -- 2 files changed, 4 deletions(-) (limited to 'kernel') diff --git a/kernel/user.c b/kernel/user.c index 69b4c3d48cde..5bbb91988e69 100644 --- a/kernel/user.c +++ b/kernel/user.c @@ -51,8 +51,6 @@ struct user_namespace init_user_ns = { .owner = GLOBAL_ROOT_UID, .group = GLOBAL_ROOT_GID, .proc_inum = PROC_USER_INIT_INO, - .may_mount_sysfs = true, - .may_mount_proc = true, }; EXPORT_SYMBOL_GPL(init_user_ns); diff --git a/kernel/user_namespace.c b/kernel/user_namespace.c index d8c30db06c5b..d58ad1e7a794 100644 --- a/kernel/user_namespace.c +++ b/kernel/user_namespace.c @@ -97,8 +97,6 @@ int create_user_ns(struct cred *new) set_cred_user_ns(new, ns); - update_mnt_policy(ns); - return 0; } -- cgit v1.2.3 From c2b1df2eb42978073ec27c99cc199d20ae48b849 Mon Sep 17 00:00:00 2001 From: Andy Lutomirski Date: Thu, 22 Aug 2013 11:39:16 -0700 Subject: Rename nsproxy.pid_ns to nsproxy.pid_ns_for_children nsproxy.pid_ns is *not* the task's pid namespace. The name should clarify that. This makes it more obvious that setns on a pid namespace is weird -- it won't change the pid namespace shown in procfs. Signed-off-by: Andy Lutomirski Reviewed-by: "Eric W. Biederman" Signed-off-by: David S. Miller --- kernel/fork.c | 5 +++-- kernel/nsproxy.c | 27 ++++++++++++++------------- kernel/pid_namespace.c | 4 ++-- 3 files changed, 19 insertions(+), 17 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index e23bb19e2a3e..bf46287c91a4 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1177,7 +1177,8 @@ static struct task_struct *copy_process(unsigned long clone_flags, * don't allow the creation of threads. */ if ((clone_flags & (CLONE_VM|CLONE_NEWPID)) && - (task_active_pid_ns(current) != current->nsproxy->pid_ns)) + (task_active_pid_ns(current) != + current->nsproxy->pid_ns_for_children)) return ERR_PTR(-EINVAL); retval = security_task_create(clone_flags); @@ -1351,7 +1352,7 @@ static struct task_struct *copy_process(unsigned long clone_flags, if (pid != &init_struct_pid) { retval = -ENOMEM; - pid = alloc_pid(p->nsproxy->pid_ns); + pid = alloc_pid(p->nsproxy->pid_ns_for_children); if (!pid) goto bad_fork_cleanup_io; } diff --git a/kernel/nsproxy.c b/kernel/nsproxy.c index 364ceab15f0c..997cbb951a3b 100644 --- a/kernel/nsproxy.c +++ b/kernel/nsproxy.c @@ -29,15 +29,15 @@ static struct kmem_cache *nsproxy_cachep; struct nsproxy init_nsproxy = { - .count = ATOMIC_INIT(1), - .uts_ns = &init_uts_ns, + .count = ATOMIC_INIT(1), + .uts_ns = &init_uts_ns, #if defined(CONFIG_POSIX_MQUEUE) || defined(CONFIG_SYSVIPC) - .ipc_ns = &init_ipc_ns, + .ipc_ns = &init_ipc_ns, #endif - .mnt_ns = NULL, - .pid_ns = &init_pid_ns, + .mnt_ns = NULL, + .pid_ns_for_children = &init_pid_ns, #ifdef CONFIG_NET - .net_ns = &init_net, + .net_ns = &init_net, #endif }; @@ -85,9 +85,10 @@ static struct nsproxy *create_new_namespaces(unsigned long flags, goto out_ipc; } - new_nsp->pid_ns = copy_pid_ns(flags, user_ns, tsk->nsproxy->pid_ns); - if (IS_ERR(new_nsp->pid_ns)) { - err = PTR_ERR(new_nsp->pid_ns); + new_nsp->pid_ns_for_children = + copy_pid_ns(flags, user_ns, tsk->nsproxy->pid_ns_for_children); + if (IS_ERR(new_nsp->pid_ns_for_children)) { + err = PTR_ERR(new_nsp->pid_ns_for_children); goto out_pid; } @@ -100,8 +101,8 @@ static struct nsproxy *create_new_namespaces(unsigned long flags, return new_nsp; out_net: - if (new_nsp->pid_ns) - put_pid_ns(new_nsp->pid_ns); + if (new_nsp->pid_ns_for_children) + put_pid_ns(new_nsp->pid_ns_for_children); out_pid: if (new_nsp->ipc_ns) put_ipc_ns(new_nsp->ipc_ns); @@ -174,8 +175,8 @@ void free_nsproxy(struct nsproxy *ns) put_uts_ns(ns->uts_ns); if (ns->ipc_ns) put_ipc_ns(ns->ipc_ns); - if (ns->pid_ns) - put_pid_ns(ns->pid_ns); + if (ns->pid_ns_for_children) + put_pid_ns(ns->pid_ns_for_children); put_net(ns->net_ns); kmem_cache_free(nsproxy_cachep, ns); } diff --git a/kernel/pid_namespace.c b/kernel/pid_namespace.c index 6917e8edb48e..601bb361c235 100644 --- a/kernel/pid_namespace.c +++ b/kernel/pid_namespace.c @@ -349,8 +349,8 @@ static int pidns_install(struct nsproxy *nsproxy, void *ns) if (ancestor != active) return -EINVAL; - put_pid_ns(nsproxy->pid_ns); - nsproxy->pid_ns = get_pid_ns(new); + put_pid_ns(nsproxy->pid_ns_for_children); + nsproxy->pid_ns_for_children = get_pid_ns(new); return 0; } -- cgit v1.2.3 From d1625964da51bda61306ad3ec45307a799c21f08 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 27 Aug 2013 14:27:23 -0400 Subject: cgroup: fix cgroup_css() invocation in css_from_id() ca8bdcaff0 ("cgroup: make cgroup_css() take cgroup_subsys * instead and allow NULL subsys") missed one conversion in css_from_id(), which was newly added. As css_from_id() doesn't have any user yet, this doesn't break anything other than generating a build warning. Convert it. Signed-off-by: Tejun Heo Reported-by: Stephen Rothwell Reported-by: kbuild test robot --- kernel/cgroup.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index e76698dd6c08..b5f4989937f2 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -5729,7 +5729,7 @@ struct cgroup_subsys_state *css_from_id(int id, struct cgroup_subsys *ss) cgrp = idr_find(&ss->root->cgroup_idr, id); if (cgrp) - return cgroup_css(cgrp, ss->subsys_id); + return cgroup_css(cgrp, ss); return NULL; } -- cgit v1.2.3 From 84a78a6504f5c5394a8e558702e5b54131f01d14 Mon Sep 17 00:00:00 2001 From: Nathan Zimmer Date: Wed, 28 Aug 2013 16:35:14 -0700 Subject: timer_list: correct the iterator for timer_list Correct an issue with /proc/timer_list reported by Holger. When reading from the proc file with a sufficiently small buffer, 2k so not really that small, there was one could get hung trying to read the file a chunk at a time. The timer_list_start function failed to account for the possibility that the offset was adjusted outside the timer_list_next. Signed-off-by: Nathan Zimmer Reported-by: Holger Hans Peter Freyther Cc: John Stultz Cc: Thomas Gleixner Cc: Berke Durak Cc: Jeff Layton Tested-by: Al Viro Cc: # 3.10.x Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/time/timer_list.c | 41 ++++++++++++++++++++++++----------------- 1 file changed, 24 insertions(+), 17 deletions(-) (limited to 'kernel') diff --git a/kernel/time/timer_list.c b/kernel/time/timer_list.c index 3bdf28323012..61ed862cdd37 100644 --- a/kernel/time/timer_list.c +++ b/kernel/time/timer_list.c @@ -265,10 +265,9 @@ static inline void timer_list_header(struct seq_file *m, u64 now) static int timer_list_show(struct seq_file *m, void *v) { struct timer_list_iter *iter = v; - u64 now = ktime_to_ns(ktime_get()); if (iter->cpu == -1 && !iter->second_pass) - timer_list_header(m, now); + timer_list_header(m, iter->now); else if (!iter->second_pass) print_cpu(m, iter->cpu, iter->now); #ifdef CONFIG_GENERIC_CLOCKEVENTS @@ -298,33 +297,41 @@ void sysrq_timer_list_show(void) return; } -static void *timer_list_start(struct seq_file *file, loff_t *offset) +static void *move_iter(struct timer_list_iter *iter, loff_t offset) { - struct timer_list_iter *iter = file->private; - - if (!*offset) { - iter->cpu = -1; - iter->now = ktime_to_ns(ktime_get()); - } else if (iter->cpu >= nr_cpu_ids) { + for (; offset; offset--) { + iter->cpu = cpumask_next(iter->cpu, cpu_online_mask); + if (iter->cpu >= nr_cpu_ids) { #ifdef CONFIG_GENERIC_CLOCKEVENTS - if (!iter->second_pass) { - iter->cpu = -1; - iter->second_pass = true; - } else - return NULL; + if (!iter->second_pass) { + iter->cpu = -1; + iter->second_pass = true; + } else + return NULL; #else - return NULL; + return NULL; #endif + } } return iter; } +static void *timer_list_start(struct seq_file *file, loff_t *offset) +{ + struct timer_list_iter *iter = file->private; + + if (!*offset) + iter->now = ktime_to_ns(ktime_get()); + iter->cpu = -1; + iter->second_pass = false; + return move_iter(iter, *offset); +} + static void *timer_list_next(struct seq_file *file, void *v, loff_t *offset) { struct timer_list_iter *iter = file->private; - iter->cpu = cpumask_next(iter->cpu, cpu_online_mask); ++*offset; - return timer_list_start(file, offset); + return move_iter(iter, 1); } static void timer_list_stop(struct seq_file *seq, void *v) -- cgit v1.2.3 From 9c823f9f7e4b392921d0d8b251bec080d58f9077 Mon Sep 17 00:00:00 2001 From: Chen Gang Date: Thu, 22 Aug 2013 06:43:37 +0000 Subject: padata - share code between CPU_ONLINE and CPU_DOWN_FAILED, same to CPU_DOWN_PREPARE and CPU_UP_CANCELED Share code between CPU_ONLINE and CPU_DOWN_FAILED, same to CPU_DOWN_PREPARE and CPU_UP_CANCELED. It will fix 2 bugs: "not check the return value of __padata_remove_cpu() and __padata_add_cpu()". "need add 'break' between CPU_UP_CANCELED and CPU_DOWN_FAILED". Signed-off-by: Chen Gang Acked-by: Steffen Klassert Signed-off-by: Herbert Xu --- kernel/padata.c | 20 ++++---------------- 1 file changed, 4 insertions(+), 16 deletions(-) (limited to 'kernel') diff --git a/kernel/padata.c b/kernel/padata.c index 072f4ee4eb89..2f0037a86289 100644 --- a/kernel/padata.c +++ b/kernel/padata.c @@ -846,6 +846,8 @@ static int padata_cpu_callback(struct notifier_block *nfb, switch (action) { case CPU_ONLINE: case CPU_ONLINE_FROZEN: + case CPU_DOWN_FAILED: + case CPU_DOWN_FAILED_FROZEN: if (!pinst_has_cpu(pinst, cpu)) break; mutex_lock(&pinst->lock); @@ -857,6 +859,8 @@ static int padata_cpu_callback(struct notifier_block *nfb, case CPU_DOWN_PREPARE: case CPU_DOWN_PREPARE_FROZEN: + case CPU_UP_CANCELED: + case CPU_UP_CANCELED_FROZEN: if (!pinst_has_cpu(pinst, cpu)) break; mutex_lock(&pinst->lock); @@ -865,22 +869,6 @@ static int padata_cpu_callback(struct notifier_block *nfb, if (err) return notifier_from_errno(err); break; - - case CPU_UP_CANCELED: - case CPU_UP_CANCELED_FROZEN: - if (!pinst_has_cpu(pinst, cpu)) - break; - mutex_lock(&pinst->lock); - __padata_remove_cpu(pinst, cpu); - mutex_unlock(&pinst->lock); - - case CPU_DOWN_FAILED: - case CPU_DOWN_FAILED_FROZEN: - if (!pinst_has_cpu(pinst, cpu)) - break; - mutex_lock(&pinst->lock); - __padata_add_cpu(pinst, cpu); - mutex_unlock(&pinst->lock); } return NOTIFY_OK; -- cgit v1.2.3 From b8b4a4166e3401b7d8ea9deb8d64d875a468144c Mon Sep 17 00:00:00 2001 From: Richard Weinberger Date: Fri, 23 Aug 2013 13:12:33 +0200 Subject: padata - Register hotcpu notifier after initialization padata_cpu_callback() takes pinst->lock, to avoid taking an uninitialized lock, register the notifier after it's initialization. Signed-off-by: Richard Weinberger Acked-by: Steffen Klassert Signed-off-by: Herbert Xu --- kernel/padata.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/padata.c b/kernel/padata.c index 2f0037a86289..07af2c95dcfe 100644 --- a/kernel/padata.c +++ b/kernel/padata.c @@ -1074,18 +1074,18 @@ struct padata_instance *padata_alloc(struct workqueue_struct *wq, pinst->flags = 0; -#ifdef CONFIG_HOTPLUG_CPU - pinst->cpu_notifier.notifier_call = padata_cpu_callback; - pinst->cpu_notifier.priority = 0; - register_hotcpu_notifier(&pinst->cpu_notifier); -#endif - put_online_cpus(); BLOCKING_INIT_NOTIFIER_HEAD(&pinst->cpumask_change_notifier); kobject_init(&pinst->kobj, &padata_attr_type); mutex_init(&pinst->lock); +#ifdef CONFIG_HOTPLUG_CPU + pinst->cpu_notifier.notifier_call = padata_cpu_callback; + pinst->cpu_notifier.priority = 0; + register_hotcpu_notifier(&pinst->cpu_notifier); +#endif + return pinst; err_free_masks: -- cgit v1.2.3 From b22ce2785d97423846206cceec4efee0c4afd980 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 28 Aug 2013 17:33:37 -0400 Subject: workqueue: cond_resched() after processing each work item If !PREEMPT, a kworker running work items back to back can hog CPU. This becomes dangerous when a self-requeueing work item which is waiting for something to happen races against stop_machine. Such self-requeueing work item would requeue itself indefinitely hogging the kworker and CPU it's running on while stop_machine would wait for that CPU to enter stop_machine while preventing anything else from happening on all other CPUs. The two would deadlock. Jamie Liu reports that this deadlock scenario exists around scsi_requeue_run_queue() and libata port multiplier support, where one port may exclude command processing from other ports. With the right timing, scsi_requeue_run_queue() can end up requeueing itself trying to execute an IO which is asked to be retried while another device has an exclusive access, which in turn can't make forward progress due to stop_machine. Fix it by invoking cond_resched() after executing each work item. Signed-off-by: Tejun Heo Reported-by: Jamie Liu References: http://thread.gmane.org/gmane.linux.kernel/1552567 Cc: stable@vger.kernel.org -- kernel/workqueue.c | 9 +++++++++ 1 file changed, 9 insertions(+) --- kernel/workqueue.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 7f5d4be22034..e93f7b9067d8 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -2201,6 +2201,15 @@ __acquires(&pool->lock) dump_stack(); } + /* + * The following prevents a kworker from hogging CPU on !PREEMPT + * kernels, where a requeueing work item waiting for something to + * happen could deadlock with stop_machine as such work item could + * indefinitely requeue itself while all other CPUs are trapped in + * stop_machine. + */ + cond_resched(); + spin_lock_irq(&pool->lock); /* clear cpu intensive status */ -- cgit v1.2.3 From bb78a92f47696b2da49f2692b6a9fa56d07c444a Mon Sep 17 00:00:00 2001 From: Hugh Dickins Date: Wed, 28 Aug 2013 16:31:23 -0700 Subject: cgroup: fix rmdir EBUSY regression in 3.11 On 3.11-rc we are seeing cgroup directories left behind when they should have been removed. Here's a trivial reproducer: cd /sys/fs/cgroup/memory mkdir parent parent/child; rmdir parent/child parent rmdir: failed to remove `parent': Device or resource busy It's because cgroup_destroy_locked() (step 1 of destruction) leaves cgroup on parent's children list, letting cgroup_offline_fn() (step 2 of destruction) remove it; but step 2 is run by work queue, which may not yet have removed the children when parent destruction checks the list. Fix that by checking through a non-empty list of children: if every one of them has already been marked CGRP_DEAD, then it's safe to proceed: those children are invisible to userspace, and should not obstruct rmdir. (I didn't see any reason to keep the cgrp->children checks under the unrelated css_set_lock, so moved them out.) tj: Flattened nested ifs a bit and updated comment so that it's correct on both for-3.11-fixes and for-3.12. Signed-off-by: Hugh Dickins Signed-off-by: Tejun Heo --- kernel/cgroup.c | 19 ++++++++++++++++++- 1 file changed, 18 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 781845a013ab..e91963302c0d 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4480,6 +4480,7 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) struct dentry *d = cgrp->dentry; struct cgroup_event *event, *tmp; struct cgroup_subsys *ss; + struct cgroup *child; bool empty; lockdep_assert_held(&d->d_inode->i_mutex); @@ -4490,11 +4491,27 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) * @cgrp from being removed while __put_css_set() is in progress. */ read_lock(&css_set_lock); - empty = list_empty(&cgrp->cset_links) && list_empty(&cgrp->children); + empty = list_empty(&cgrp->cset_links); read_unlock(&css_set_lock); if (!empty) return -EBUSY; + /* + * Make sure there's no live children. We can't test ->children + * emptiness as dead children linger on it while being destroyed; + * otherwise, "rmdir parent/child parent" may fail with -EBUSY. + */ + empty = true; + rcu_read_lock(); + list_for_each_entry_rcu(child, &cgrp->children, sibling) { + empty = cgroup_is_dead(child); + if (!empty) + break; + } + rcu_read_unlock(); + if (!empty) + return -EBUSY; + /* * Block new css_tryget() by killing css refcnts. cgroup core * guarantees that, by the time ->css_offline() is invoked, no new -- cgit v1.2.3 From ff3d527cebc1fa3707c617bfe9e74f53fcfb0955 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Tue, 27 Aug 2013 11:23:07 +0300 Subject: perf: make events stream always parsable The event stream is not always parsable because the format of a sample is dependent on the sample_type of the selected event. When there is more than one selected event and the sample_types are not the same then parsing becomes problematic. A sample can be matched to its selected event using the ID that is allocated when the event is opened. Unfortunately, to get the ID from the sample means first parsing it. This patch adds a new sample format bit PERF_SAMPLE_IDENTIFER that puts the ID at a fixed position so that the ID can be retrieved without parsing the sample. For sample events, that is the first position immediately after the header. For non-sample events, that is the last position. In this respect parsing samples requires that the sample_type and ID values are recorded. For example, perf tools records struct perf_event_attr and the IDs within the perf.data file. Those must be read first before it is possible to parse samples found later in the perf.data file. Signed-off-by: Adrian Hunter Tested-by: Stephane Eranian Acked-by: Peter Zijlstra Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mike Galbraith Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1377591794-30553-6-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- kernel/events/core.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 928fae7ca8c7..15d0f2418e54 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -1213,6 +1213,9 @@ static void perf_event__id_header_size(struct perf_event *event) if (sample_type & PERF_SAMPLE_TIME) size += sizeof(data->time); + if (sample_type & PERF_SAMPLE_IDENTIFIER) + size += sizeof(data->id); + if (sample_type & PERF_SAMPLE_ID) size += sizeof(data->id); @@ -4280,7 +4283,7 @@ static void __perf_event_header__init_id(struct perf_event_header *header, if (sample_type & PERF_SAMPLE_TIME) data->time = perf_clock(); - if (sample_type & PERF_SAMPLE_ID) + if (sample_type & (PERF_SAMPLE_ID | PERF_SAMPLE_IDENTIFIER)) data->id = primary_event_id(event); if (sample_type & PERF_SAMPLE_STREAM_ID) @@ -4319,6 +4322,9 @@ static void __perf_event__output_id_sample(struct perf_output_handle *handle, if (sample_type & PERF_SAMPLE_CPU) perf_output_put(handle, data->cpu_entry); + + if (sample_type & PERF_SAMPLE_IDENTIFIER) + perf_output_put(handle, data->id); } void perf_event__output_id_sample(struct perf_event *event, @@ -4432,6 +4438,9 @@ void perf_output_sample(struct perf_output_handle *handle, perf_output_put(handle, *header); + if (sample_type & PERF_SAMPLE_IDENTIFIER) + perf_output_put(handle, data->id); + if (sample_type & PERF_SAMPLE_IP) perf_output_put(handle, data->ip); -- cgit v1.2.3 From a606488513543312805fab2b93070cefe6a3016c Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Thu, 29 Aug 2013 13:56:50 -0700 Subject: pidns: Fix hang in zap_pid_ns_processes by sending a potentially extra wakeup Serge Hallyn writes: > Since commit af4b8a83add95ef40716401395b44a1b579965f4 it's been > possible to get into a situation where a pidns reaper is > , reparented to host pid 1, but never reaped. How to > reproduce this is documented at > > https://bugs.launchpad.net/ubuntu/+source/lxc/+bug/1168526 > (and see > https://bugs.launchpad.net/ubuntu/+source/lxc/+bug/1168526/comments/13) > In short, run repeated starts of a container whose init is > > Process.exit(0); > > sysrq-t when such a task is playing zombie shows: > > [ 131.132978] init x ffff88011fc14580 0 2084 2039 0x00000000 > [ 131.132978] ffff880116e89ea8 0000000000000002 ffff880116e89fd8 0000000000014580 > [ 131.132978] ffff880116e89fd8 0000000000014580 ffff8801172a0000 ffff8801172a0000 > [ 131.132978] ffff8801172a0630 ffff88011729fff0 ffff880116e14650 ffff88011729fff0 > [ 131.132978] Call Trace: > [ 131.132978] [] schedule+0x29/0x70 > [ 131.132978] [] do_exit+0x6e1/0xa40 > [ 131.132978] [] ? signal_wake_up_state+0x1e/0x30 > [ 131.132978] [] do_group_exit+0x3f/0xa0 > [ 131.132978] [] SyS_exit_group+0x14/0x20 > [ 131.132978] [] tracesys+0xe1/0xe6 > > Further debugging showed that every time this happened, zap_pid_ns_processes() > started with nr_hashed being 3, while we were expecting it to drop to 2. > Any time it didn't happen, nr_hashed was 1 or 2. So the reaper was > waiting for nr_hashed to become 2, but free_pid() only wakes the reaper > if nr_hashed hits 1. The issue is that when the task group leader of an init process exits before other tasks of the init process when the init process finally exits it will be a secondary task sleeping in zap_pid_ns_processes and waiting to wake up when the number of hashed pids drops to two. This case waits forever as free_pid only sends a wake up when the number of hashed pids drops to 1. To correct this the simple strategy of sending a possibly unncessary wake up when the number of hashed pids drops to 2 is adopted. Sending one extraneous wake up is relatively harmless, at worst we waste a little cpu time in the rare case when a pid namespace appropaches exiting. We can detect the case when the pid namespace drops to just two pids hashed race free in free_pid. Dereferencing pid_ns->child_reaper with the pidmap_lock held is safe without out the tasklist_lock because it is guaranteed that the detach_pid will be called on the child_reaper before it is freed and detach_pid calls __change_pid which calls free_pid which takes the pidmap_lock. __change_pid only calls free_pid if this is the last use of the pid. For a thread that is not the thread group leader the threads pid will only ever have one user because a threads pid is not allowed to be the pid of a process, of a process group or a session. For a thread that is a thread group leader all of the other threads of that process will be reaped before it is allowed for the thread group leader to be reaped ensuring there will only be one user of the threads pid as a process pid. Furthermore because the thread is the init process of a pid namespace all of the other processes in the pid namespace will have also been already freed leading to the fact that the pid will not be used as a session pid or a process group pid for any other running process. CC: stable@vger.kernel.org Acked-by: Serge Hallyn Tested-by: Serge Hallyn Reported-by: Serge Hallyn Signed-off-by: "Eric W. Biederman" --- kernel/pid.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/pid.c b/kernel/pid.c index 66505c1dfc51..ebe5e80b10f8 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -265,6 +265,7 @@ void free_pid(struct pid *pid) struct pid_namespace *ns = upid->ns; hlist_del_rcu(&upid->pid_chain); switch(--ns->nr_hashed) { + case 2: case 1: /* When all that is left in the pid namespace * is the reaper wake up the reaper. The reaper -- cgit v1.2.3 From dbef0c1c4c5f8ce5d1f5bd8cee092a7afb4ac21b Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Sat, 9 Mar 2013 16:15:23 -0800 Subject: namespaces: Simplify copy_namespaces so it is clear what is going on. Remove the test for the impossible case where tsk->nsproxy == NULL. Fork will never be called with tsk->nsproxy == NULL. Only call get_nsproxy when we don't need to generate a new_nsproxy, and mark the case where we don't generate a new nsproxy as likely. Remove the code to drop an unnecessarily acquired nsproxy value. Acked-by: Serge Hallyn Signed-off-by: "Eric W. Biederman" --- kernel/nsproxy.c | 35 +++++++++++------------------------ 1 file changed, 11 insertions(+), 24 deletions(-) (limited to 'kernel') diff --git a/kernel/nsproxy.c b/kernel/nsproxy.c index d9afd256318f..a1ed01139276 100644 --- a/kernel/nsproxy.c +++ b/kernel/nsproxy.c @@ -125,22 +125,16 @@ int copy_namespaces(unsigned long flags, struct task_struct *tsk) struct nsproxy *old_ns = tsk->nsproxy; struct user_namespace *user_ns = task_cred_xxx(tsk, user_ns); struct nsproxy *new_ns; - int err = 0; - if (!old_ns) + if (likely(!(flags & (CLONE_NEWNS | CLONE_NEWUTS | CLONE_NEWIPC | + CLONE_NEWPID | CLONE_NEWNET)))) { + get_nsproxy(old_ns); return 0; - - get_nsproxy(old_ns); - - if (!(flags & (CLONE_NEWNS | CLONE_NEWUTS | CLONE_NEWIPC | - CLONE_NEWPID | CLONE_NEWNET))) - return 0; - - if (!ns_capable(user_ns, CAP_SYS_ADMIN)) { - err = -EPERM; - goto out; } + if (!ns_capable(user_ns, CAP_SYS_ADMIN)) + return -EPERM; + /* * CLONE_NEWIPC must detach from the undolist: after switching * to a new ipc namespace, the semaphore arrays from the old @@ -149,22 +143,15 @@ int copy_namespaces(unsigned long flags, struct task_struct *tsk) * it along with CLONE_NEWIPC. */ if ((flags & (CLONE_NEWIPC | CLONE_SYSVSEM)) == - (CLONE_NEWIPC | CLONE_SYSVSEM)) { - err = -EINVAL; - goto out; - } + (CLONE_NEWIPC | CLONE_SYSVSEM)) + return -EINVAL; new_ns = create_new_namespaces(flags, tsk, user_ns, tsk->fs); - if (IS_ERR(new_ns)) { - err = PTR_ERR(new_ns); - goto out; - } + if (IS_ERR(new_ns)) + return PTR_ERR(new_ns); tsk->nsproxy = new_ns; - -out: - put_nsproxy(old_ns); - return err; + return 0; } void free_nsproxy(struct nsproxy *ns) -- cgit v1.2.3 From 8fd37a4c9822d58c93f764864582aa13112b1513 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Fri, 30 Aug 2013 14:19:38 +0200 Subject: PM / hibernate: Create memory bitmaps after freezing user space The hibernation core uses special memory bitmaps during image creation and restoration and traditionally those bitmaps are allocated before freezing tasks, because in the past GFP_KERNEL allocations might not work after all tasks had been frozen. However, this is an anachronism, because hibernation_snapshot() now calls hibernate_preallocate_memory() which allocates memory for the image upfront anyway, so the memory bitmaps may be allocated after freezing user space safely. For this reason, move all of the create_basic_memory_bitmaps() calls after freeze_processes() and all of the corresponding free_basic_memory_bitmaps() calls before thaw_processes(). This will allow us to hold device_hotplug_lock around hibernation without the need to worry about freezing issues with user space processes attempting to acquire it via sysfs attributes after the creation of memory bitmaps and before the freezing of tasks. Signed-off-by: Rafael J. Wysocki Acked-by: Toshi Kani --- kernel/power/hibernate.c | 41 +++++++++++++++++++---------------------- kernel/power/user.c | 22 ++++++++++++---------- 2 files changed, 31 insertions(+), 32 deletions(-) (limited to 'kernel') diff --git a/kernel/power/hibernate.c b/kernel/power/hibernate.c index b26f5f1e773e..d4e54053d009 100644 --- a/kernel/power/hibernate.c +++ b/kernel/power/hibernate.c @@ -644,22 +644,22 @@ int hibernate(void) if (error) goto Exit; - /* Allocate memory management structures */ - error = create_basic_memory_bitmaps(); - if (error) - goto Exit; - printk(KERN_INFO "PM: Syncing filesystems ... "); sys_sync(); printk("done.\n"); error = freeze_processes(); if (error) - goto Free_bitmaps; + goto Exit; + + /* Allocate memory management structures */ + error = create_basic_memory_bitmaps(); + if (error) + goto Thaw; error = hibernation_snapshot(hibernation_mode == HIBERNATION_PLATFORM); if (error || freezer_test_done) - goto Thaw; + goto Free_bitmaps; if (in_suspend) { unsigned int flags = 0; @@ -682,14 +682,13 @@ int hibernate(void) pr_debug("PM: Image restored successfully.\n"); } + Free_bitmaps: + free_basic_memory_bitmaps(); Thaw: thaw_processes(); /* Don't bother checking whether freezer_test_done is true */ freezer_test_done = false; - - Free_bitmaps: - free_basic_memory_bitmaps(); Exit: pm_notifier_call_chain(PM_POST_HIBERNATION); pm_restore_console(); @@ -806,21 +805,19 @@ static int software_resume(void) pm_prepare_console(); error = pm_notifier_call_chain(PM_RESTORE_PREPARE); if (error) - goto close_finish; - - error = create_basic_memory_bitmaps(); - if (error) - goto close_finish; + goto Close_Finish; pr_debug("PM: Preparing processes for restore.\n"); error = freeze_processes(); - if (error) { - swsusp_close(FMODE_READ); - goto Done; - } + if (error) + goto Close_Finish; pr_debug("PM: Loading hibernation image.\n"); + error = create_basic_memory_bitmaps(); + if (error) + goto Thaw; + error = swsusp_read(&flags); swsusp_close(FMODE_READ); if (!error) @@ -828,9 +825,9 @@ static int software_resume(void) printk(KERN_ERR "PM: Failed to load hibernation image, recovering.\n"); swsusp_free(); - thaw_processes(); - Done: free_basic_memory_bitmaps(); + Thaw: + thaw_processes(); Finish: pm_notifier_call_chain(PM_POST_RESTORE); pm_restore_console(); @@ -840,7 +837,7 @@ static int software_resume(void) mutex_unlock(&pm_mutex); pr_debug("PM: Hibernation image not present or could not be loaded.\n"); return error; -close_finish: + Close_Finish: swsusp_close(FMODE_READ); goto Finish; } diff --git a/kernel/power/user.c b/kernel/power/user.c index 4ed81e74f86f..63368163e98d 100644 --- a/kernel/power/user.c +++ b/kernel/power/user.c @@ -60,11 +60,6 @@ static int snapshot_open(struct inode *inode, struct file *filp) error = -ENOSYS; goto Unlock; } - if(create_basic_memory_bitmaps()) { - atomic_inc(&snapshot_device_available); - error = -ENOMEM; - goto Unlock; - } nonseekable_open(inode, filp); data = &snapshot_state; filp->private_data = data; @@ -90,10 +85,9 @@ static int snapshot_open(struct inode *inode, struct file *filp) if (error) pm_notifier_call_chain(PM_POST_RESTORE); } - if (error) { - free_basic_memory_bitmaps(); + if (error) atomic_inc(&snapshot_device_available); - } + data->frozen = 0; data->ready = 0; data->platform_support = 0; @@ -111,11 +105,11 @@ static int snapshot_release(struct inode *inode, struct file *filp) lock_system_sleep(); swsusp_free(); - free_basic_memory_bitmaps(); data = filp->private_data; free_all_swap_pages(data->swap); if (data->frozen) { pm_restore_gfp_mask(); + free_basic_memory_bitmaps(); thaw_processes(); } pm_notifier_call_chain(data->mode == O_RDONLY ? @@ -220,14 +214,22 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd, printk("done.\n"); error = freeze_processes(); - if (!error) + if (error) + break; + + error = create_basic_memory_bitmaps(); + if (error) + thaw_processes(); + else data->frozen = 1; + break; case SNAPSHOT_UNFREEZE: if (!data->frozen || data->ready) break; pm_restore_gfp_mask(); + free_basic_memory_bitmaps(); thaw_processes(); data->frozen = 0; break; -- cgit v1.2.3 From 942f40155a743f4204308d62405dacaa4bfadb11 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Fri, 30 Aug 2013 14:19:46 +0200 Subject: PM / hibernate / memory hotplug: Rework mutual exclusion Since all of the memory hotplug operations have to be carried out under device_hotplug_lock, they won't need to acquire pm_mutex if device_hotplug_lock is held around hibernation. For this reason, make the hibernation code acquire device_hotplug_lock after freezing user space processes and release it before thawing them. At the same tim drop the lock_system_sleep() and unlock_system_sleep() calls from lock_memory_hotplug() and unlock_memory_hotplug(), respectively. Signed-off-by: Rafael J. Wysocki Acked-by: Toshi Kani --- kernel/power/hibernate.c | 4 ++++ kernel/power/user.c | 2 ++ 2 files changed, 6 insertions(+) (limited to 'kernel') diff --git a/kernel/power/hibernate.c b/kernel/power/hibernate.c index d4e54053d009..0b78f72ad39d 100644 --- a/kernel/power/hibernate.c +++ b/kernel/power/hibernate.c @@ -652,6 +652,7 @@ int hibernate(void) if (error) goto Exit; + lock_device_hotplug(); /* Allocate memory management structures */ error = create_basic_memory_bitmaps(); if (error) @@ -685,6 +686,7 @@ int hibernate(void) Free_bitmaps: free_basic_memory_bitmaps(); Thaw: + unlock_device_hotplug(); thaw_processes(); /* Don't bother checking whether freezer_test_done is true */ @@ -814,6 +816,7 @@ static int software_resume(void) pr_debug("PM: Loading hibernation image.\n"); + lock_device_hotplug(); error = create_basic_memory_bitmaps(); if (error) goto Thaw; @@ -827,6 +830,7 @@ static int software_resume(void) swsusp_free(); free_basic_memory_bitmaps(); Thaw: + unlock_device_hotplug(); thaw_processes(); Finish: pm_notifier_call_chain(PM_POST_RESTORE); diff --git a/kernel/power/user.c b/kernel/power/user.c index 63368163e98d..72e8f4fd616d 100644 --- a/kernel/power/user.c +++ b/kernel/power/user.c @@ -201,6 +201,7 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd, if (!mutex_trylock(&pm_mutex)) return -EBUSY; + lock_device_hotplug(); data = filp->private_data; switch (cmd) { @@ -373,6 +374,7 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd, } + unlock_device_hotplug(); mutex_unlock(&pm_mutex); return error; -- cgit v1.2.3 From 6e556ce209b09528dbf1931cbfd5d323e1345926 Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Tue, 5 Mar 2013 13:59:48 -0800 Subject: pidns: Don't have unshare(CLONE_NEWPID) imply CLONE_THREAD I goofed when I made unshare(CLONE_NEWPID) only work in a single-threaded process. There is no need for that requirement and in fact I analyzied things right for setns. The hard requirement is for tasks that share a VM to all be in the pid namespace and we properly prevent that in do_fork. Just to be certain I took a look through do_wait and forget_original_parent and there are no cases that make it any harder for children to be in the multiple pid namespaces than it is for children to be in the same pid namespace. I also performed a check to see if there were in uses of task->nsproxy_pid_ns I was not familiar with, but it is only used when allocating a new pid for a new task, and in checks to prevent craziness from happening. Acked-by: Serge Hallyn Signed-off-by: "Eric W. Biederman" --- kernel/fork.c | 5 ----- 1 file changed, 5 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index 66635c80a813..eb45f1d72703 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1817,11 +1817,6 @@ SYSCALL_DEFINE1(unshare, unsigned long, unshare_flags) */ if (unshare_flags & CLONE_NEWUSER) unshare_flags |= CLONE_THREAD | CLONE_FS; - /* - * If unsharing a pid namespace must also unshare the thread. - */ - if (unshare_flags & CLONE_NEWPID) - unshare_flags |= CLONE_THREAD; /* * If unsharing a thread from a thread group, must also unshare vm. */ -- cgit v1.2.3 From c7b96acf1456ef127fef461fcfedb54b81fecfbb Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Wed, 20 Mar 2013 12:49:49 -0700 Subject: userns: Kill nsown_capable it makes the wrong thing easy nsown_capable is a special case of ns_capable essentially for just CAP_SETUID and CAP_SETGID. For the existing users it doesn't noticably simplify things and from the suggested patches I have seen it encourages people to do the wrong thing. So remove nsown_capable. Acked-by: Serge Hallyn Signed-off-by: "Eric W. Biederman" --- kernel/capability.c | 12 ------------ kernel/groups.c | 2 +- kernel/pid_namespace.c | 2 +- kernel/sys.c | 20 ++++++++++---------- kernel/uid16.c | 2 +- kernel/utsname.c | 2 +- 6 files changed, 14 insertions(+), 26 deletions(-) (limited to 'kernel') diff --git a/kernel/capability.c b/kernel/capability.c index f6c2ce5701e1..6fc1c8af44df 100644 --- a/kernel/capability.c +++ b/kernel/capability.c @@ -432,18 +432,6 @@ bool capable(int cap) } EXPORT_SYMBOL(capable); -/** - * nsown_capable - Check superior capability to one's own user_ns - * @cap: The capability in question - * - * Return true if the current task has the given superior capability - * targeted at its own user namespace. - */ -bool nsown_capable(int cap) -{ - return ns_capable(current_user_ns(), cap); -} - /** * inode_capable - Check superior capability over inode * @inode: The inode in question diff --git a/kernel/groups.c b/kernel/groups.c index 6b2588dd04ff..90cf1c38c8ea 100644 --- a/kernel/groups.c +++ b/kernel/groups.c @@ -233,7 +233,7 @@ SYSCALL_DEFINE2(setgroups, int, gidsetsize, gid_t __user *, grouplist) struct group_info *group_info; int retval; - if (!nsown_capable(CAP_SETGID)) + if (!ns_capable(current_user_ns(), CAP_SETGID)) return -EPERM; if ((unsigned)gidsetsize > NGROUPS_MAX) return -EINVAL; diff --git a/kernel/pid_namespace.c b/kernel/pid_namespace.c index 6917e8edb48e..ee1f6bb83d67 100644 --- a/kernel/pid_namespace.c +++ b/kernel/pid_namespace.c @@ -329,7 +329,7 @@ static int pidns_install(struct nsproxy *nsproxy, void *ns) struct pid_namespace *ancestor, *new = ns; if (!ns_capable(new->user_ns, CAP_SYS_ADMIN) || - !nsown_capable(CAP_SYS_ADMIN)) + !ns_capable(current_user_ns(), CAP_SYS_ADMIN)) return -EPERM; /* diff --git a/kernel/sys.c b/kernel/sys.c index 771129b299f8..c18ecca575b4 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -337,7 +337,7 @@ SYSCALL_DEFINE2(setregid, gid_t, rgid, gid_t, egid) if (rgid != (gid_t) -1) { if (gid_eq(old->gid, krgid) || gid_eq(old->egid, krgid) || - nsown_capable(CAP_SETGID)) + ns_capable(old->user_ns, CAP_SETGID)) new->gid = krgid; else goto error; @@ -346,7 +346,7 @@ SYSCALL_DEFINE2(setregid, gid_t, rgid, gid_t, egid) if (gid_eq(old->gid, kegid) || gid_eq(old->egid, kegid) || gid_eq(old->sgid, kegid) || - nsown_capable(CAP_SETGID)) + ns_capable(old->user_ns, CAP_SETGID)) new->egid = kegid; else goto error; @@ -387,7 +387,7 @@ SYSCALL_DEFINE1(setgid, gid_t, gid) old = current_cred(); retval = -EPERM; - if (nsown_capable(CAP_SETGID)) + if (ns_capable(old->user_ns, CAP_SETGID)) new->gid = new->egid = new->sgid = new->fsgid = kgid; else if (gid_eq(kgid, old->gid) || gid_eq(kgid, old->sgid)) new->egid = new->fsgid = kgid; @@ -471,7 +471,7 @@ SYSCALL_DEFINE2(setreuid, uid_t, ruid, uid_t, euid) new->uid = kruid; if (!uid_eq(old->uid, kruid) && !uid_eq(old->euid, kruid) && - !nsown_capable(CAP_SETUID)) + !ns_capable(old->user_ns, CAP_SETUID)) goto error; } @@ -480,7 +480,7 @@ SYSCALL_DEFINE2(setreuid, uid_t, ruid, uid_t, euid) if (!uid_eq(old->uid, keuid) && !uid_eq(old->euid, keuid) && !uid_eq(old->suid, keuid) && - !nsown_capable(CAP_SETUID)) + !ns_capable(old->user_ns, CAP_SETUID)) goto error; } @@ -534,7 +534,7 @@ SYSCALL_DEFINE1(setuid, uid_t, uid) old = current_cred(); retval = -EPERM; - if (nsown_capable(CAP_SETUID)) { + if (ns_capable(old->user_ns, CAP_SETUID)) { new->suid = new->uid = kuid; if (!uid_eq(kuid, old->uid)) { retval = set_user(new); @@ -591,7 +591,7 @@ SYSCALL_DEFINE3(setresuid, uid_t, ruid, uid_t, euid, uid_t, suid) old = current_cred(); retval = -EPERM; - if (!nsown_capable(CAP_SETUID)) { + if (!ns_capable(old->user_ns, CAP_SETUID)) { if (ruid != (uid_t) -1 && !uid_eq(kruid, old->uid) && !uid_eq(kruid, old->euid) && !uid_eq(kruid, old->suid)) goto error; @@ -673,7 +673,7 @@ SYSCALL_DEFINE3(setresgid, gid_t, rgid, gid_t, egid, gid_t, sgid) old = current_cred(); retval = -EPERM; - if (!nsown_capable(CAP_SETGID)) { + if (!ns_capable(old->user_ns, CAP_SETGID)) { if (rgid != (gid_t) -1 && !gid_eq(krgid, old->gid) && !gid_eq(krgid, old->egid) && !gid_eq(krgid, old->sgid)) goto error; @@ -744,7 +744,7 @@ SYSCALL_DEFINE1(setfsuid, uid_t, uid) if (uid_eq(kuid, old->uid) || uid_eq(kuid, old->euid) || uid_eq(kuid, old->suid) || uid_eq(kuid, old->fsuid) || - nsown_capable(CAP_SETUID)) { + ns_capable(old->user_ns, CAP_SETUID)) { if (!uid_eq(kuid, old->fsuid)) { new->fsuid = kuid; if (security_task_fix_setuid(new, old, LSM_SETID_FS) == 0) @@ -783,7 +783,7 @@ SYSCALL_DEFINE1(setfsgid, gid_t, gid) if (gid_eq(kgid, old->gid) || gid_eq(kgid, old->egid) || gid_eq(kgid, old->sgid) || gid_eq(kgid, old->fsgid) || - nsown_capable(CAP_SETGID)) { + ns_capable(old->user_ns, CAP_SETGID)) { if (!gid_eq(kgid, old->fsgid)) { new->fsgid = kgid; goto change_okay; diff --git a/kernel/uid16.c b/kernel/uid16.c index f6c83d7ef000..602e5bbbceff 100644 --- a/kernel/uid16.c +++ b/kernel/uid16.c @@ -176,7 +176,7 @@ SYSCALL_DEFINE2(setgroups16, int, gidsetsize, old_gid_t __user *, grouplist) struct group_info *group_info; int retval; - if (!nsown_capable(CAP_SETGID)) + if (!ns_capable(current_user_ns(), CAP_SETGID)) return -EPERM; if ((unsigned)gidsetsize > NGROUPS_MAX) return -EINVAL; diff --git a/kernel/utsname.c b/kernel/utsname.c index 2fc8576efaa8..fd393124e507 100644 --- a/kernel/utsname.c +++ b/kernel/utsname.c @@ -114,7 +114,7 @@ static int utsns_install(struct nsproxy *nsproxy, void *new) struct uts_namespace *ns = new; if (!ns_capable(ns->user_ns, CAP_SYS_ADMIN) || - !nsown_capable(CAP_SYS_ADMIN)) + !ns_capable(current_user_ns(), CAP_SYS_ADMIN)) return -EPERM; get_uts_ns(ns); -- cgit v1.2.3 From 0edd1b1784cbdad55aca2c1293be018f53c0ab1d Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Fri, 21 Jun 2013 16:37:22 -0700 Subject: nohz_full: Add full-system-idle state machine This commit adds the state machine that takes the per-CPU idle data as input and produces a full-system-idle indication as output. This state machine is driven out of RCU's quiescent-state-forcing mechanism, which invokes rcu_sysidle_check_cpu() to collect per-CPU idle state and then rcu_sysidle_report() to drive the state machine. The full-system-idle state is sampled using rcu_sys_is_idle(), which also drives the state machine if RCU is idle (and does so by forcing RCU to become non-idle). This function returns true if all but the timekeeping CPU (tick_do_timer_cpu) are idle and have been idle long enough to avoid memory contention on the full_sysidle_state state variable. The rcu_sysidle_force_exit() may be called externally to reset the state machine back into non-idle state. For large systems the state machine is driven out of RCU's force-quiescent-state logic, which provides good scalability at the price of millisecond-scale latencies on the transition to full-system-idle state. This is not so good for battery-powered systems, which are usually small enough that they don't need to care about scalability, but which do care deeply about energy efficiency. Small systems therefore drive the state machine directly out of the idle-entry code. The number of CPUs in a "small" system is defined by a new NO_HZ_FULL_SYSIDLE_SMALL Kconfig parameter, which defaults to 8. Note that this is a build-time definition. Signed-off-by: Paul E. McKenney Cc: Frederic Weisbecker Cc: Steven Rostedt Cc: Lai Jiangshan [ paulmck: Use true and false for boolean constants per Lai Jiangshan. ] Reviewed-by: Josh Triplett [ paulmck: Simplify logic and provide better comments for memory barriers, based on review comments and questions by Lai Jiangshan. ] --- kernel/rcutree.c | 16 ++- kernel/rcutree.h | 5 + kernel/rcutree_plugin.h | 296 +++++++++++++++++++++++++++++++++++++++++++++++- kernel/time/Kconfig | 27 +++++ 4 files changed, 337 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 7b5be56d95ae..eca70f4469c1 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -734,6 +734,7 @@ static int dyntick_save_progress_counter(struct rcu_data *rdp, bool *isidle, unsigned long *maxj) { rdp->dynticks_snap = atomic_add_return(0, &rdp->dynticks->dynticks); + rcu_sysidle_check_cpu(rdp, isidle, maxj); return (rdp->dynticks_snap & 0x1) == 0; } @@ -1373,11 +1374,17 @@ int rcu_gp_fqs(struct rcu_state *rsp, int fqs_state_in) rsp->n_force_qs++; if (fqs_state == RCU_SAVE_DYNTICK) { /* Collect dyntick-idle snapshots. */ + if (is_sysidle_rcu_state(rsp)) { + isidle = 1; + maxj = jiffies - ULONG_MAX / 4; + } force_qs_rnp(rsp, dyntick_save_progress_counter, &isidle, &maxj); + rcu_sysidle_report_gp(rsp, isidle, maxj); fqs_state = RCU_FORCE_QS; } else { /* Handle dyntick-idle and offline CPUs. */ + isidle = 0; force_qs_rnp(rsp, rcu_implicit_dynticks_qs, &isidle, &maxj); } /* Clear flag to prevent immediate re-entry. */ @@ -2103,9 +2110,12 @@ static void force_qs_rnp(struct rcu_state *rsp, cpu = rnp->grplo; bit = 1; for (; cpu <= rnp->grphi; cpu++, bit <<= 1) { - if ((rnp->qsmask & bit) != 0 && - f(per_cpu_ptr(rsp->rda, cpu), isidle, maxj)) - mask |= bit; + if ((rnp->qsmask & bit) != 0) { + if ((rnp->qsmaskinit & bit) != 0) + *isidle = 0; + if (f(per_cpu_ptr(rsp->rda, cpu), isidle, maxj)) + mask |= bit; + } } if (mask != 0) { diff --git a/kernel/rcutree.h b/kernel/rcutree.h index 9dd8b177f1ac..6fd3659cf01a 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -555,6 +555,11 @@ static void rcu_kick_nohz_cpu(int cpu); static bool init_nocb_callback_list(struct rcu_data *rdp); static void rcu_sysidle_enter(struct rcu_dynticks *rdtp, int irq); static void rcu_sysidle_exit(struct rcu_dynticks *rdtp, int irq); +static void rcu_sysidle_check_cpu(struct rcu_data *rdp, bool *isidle, + unsigned long *maxj); +static bool is_sysidle_rcu_state(struct rcu_state *rsp); +static void rcu_sysidle_report_gp(struct rcu_state *rsp, int isidle, + unsigned long maxj); static void rcu_sysidle_init_percpu_data(struct rcu_dynticks *rdtp); #endif /* #ifndef RCU_TREE_NONCORE */ diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index a7419ceb19ad..45ebba747af4 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -28,7 +28,7 @@ #include #include #include -#include +#include "time/tick-internal.h" #define RCU_KTHREAD_PRIO 1 @@ -2382,12 +2382,12 @@ static void rcu_kick_nohz_cpu(int cpu) * most active flavor of RCU. */ #ifdef CONFIG_PREEMPT_RCU -static struct rcu_state __maybe_unused *rcu_sysidle_state = &rcu_preempt_state; +static struct rcu_state *rcu_sysidle_state = &rcu_preempt_state; #else /* #ifdef CONFIG_PREEMPT_RCU */ -static struct rcu_state __maybe_unused *rcu_sysidle_state = &rcu_sched_state; +static struct rcu_state *rcu_sysidle_state = &rcu_sched_state; #endif /* #else #ifdef CONFIG_PREEMPT_RCU */ -static int __maybe_unused full_sysidle_state; /* Current system-idle state. */ +static int full_sysidle_state; /* Current system-idle state. */ #define RCU_SYSIDLE_NOT 0 /* Some CPU is not idle. */ #define RCU_SYSIDLE_SHORT 1 /* All CPUs idle for brief period. */ #define RCU_SYSIDLE_LONG 2 /* All CPUs idle for long enough. */ @@ -2430,6 +2430,38 @@ static void rcu_sysidle_enter(struct rcu_dynticks *rdtp, int irq) WARN_ON_ONCE(atomic_read(&rdtp->dynticks_idle) & 0x1); } +/* + * Unconditionally force exit from full system-idle state. This is + * invoked when a normal CPU exits idle, but must be called separately + * for the timekeeping CPU (tick_do_timer_cpu). The reason for this + * is that the timekeeping CPU is permitted to take scheduling-clock + * interrupts while the system is in system-idle state, and of course + * rcu_sysidle_exit() has no way of distinguishing a scheduling-clock + * interrupt from any other type of interrupt. + */ +void rcu_sysidle_force_exit(void) +{ + int oldstate = ACCESS_ONCE(full_sysidle_state); + int newoldstate; + + /* + * Each pass through the following loop attempts to exit full + * system-idle state. If contention proves to be a problem, + * a trylock-based contention tree could be used here. + */ + while (oldstate > RCU_SYSIDLE_SHORT) { + newoldstate = cmpxchg(&full_sysidle_state, + oldstate, RCU_SYSIDLE_NOT); + if (oldstate == newoldstate && + oldstate == RCU_SYSIDLE_FULL_NOTED) { + rcu_kick_nohz_cpu(tick_do_timer_cpu); + return; /* We cleared it, done! */ + } + oldstate = newoldstate; + } + smp_mb(); /* Order initial oldstate fetch vs. later non-idle work. */ +} + /* * Invoked to note entry to irq or task transition from idle. Note that * usermode execution does -not- count as idle here! The caller must @@ -2463,6 +2495,247 @@ static void rcu_sysidle_exit(struct rcu_dynticks *rdtp, int irq) atomic_inc(&rdtp->dynticks_idle); smp_mb__after_atomic_inc(); WARN_ON_ONCE(!(atomic_read(&rdtp->dynticks_idle) & 0x1)); + + /* + * If we are the timekeeping CPU, we are permitted to be non-idle + * during a system-idle state. This must be the case, because + * the timekeeping CPU has to take scheduling-clock interrupts + * during the time that the system is transitioning to full + * system-idle state. This means that the timekeeping CPU must + * invoke rcu_sysidle_force_exit() directly if it does anything + * more than take a scheduling-clock interrupt. + */ + if (smp_processor_id() == tick_do_timer_cpu) + return; + + /* Update system-idle state: We are clearly no longer fully idle! */ + rcu_sysidle_force_exit(); +} + +/* + * Check to see if the current CPU is idle. Note that usermode execution + * does not count as idle. The caller must have disabled interrupts. + */ +static void rcu_sysidle_check_cpu(struct rcu_data *rdp, bool *isidle, + unsigned long *maxj) +{ + int cur; + unsigned long j; + struct rcu_dynticks *rdtp = rdp->dynticks; + + /* + * If some other CPU has already reported non-idle, if this is + * not the flavor of RCU that tracks sysidle state, or if this + * is an offline or the timekeeping CPU, nothing to do. + */ + if (!*isidle || rdp->rsp != rcu_sysidle_state || + cpu_is_offline(rdp->cpu) || rdp->cpu == tick_do_timer_cpu) + return; + /* WARN_ON_ONCE(smp_processor_id() != tick_do_timer_cpu); */ + + /* Pick up current idle and NMI-nesting counter and check. */ + cur = atomic_read(&rdtp->dynticks_idle); + if (cur & 0x1) { + *isidle = false; /* We are not idle! */ + return; + } + smp_mb(); /* Read counters before timestamps. */ + + /* Pick up timestamps. */ + j = ACCESS_ONCE(rdtp->dynticks_idle_jiffies); + /* If this CPU entered idle more recently, update maxj timestamp. */ + if (ULONG_CMP_LT(*maxj, j)) + *maxj = j; +} + +/* + * Is this the flavor of RCU that is handling full-system idle? + */ +static bool is_sysidle_rcu_state(struct rcu_state *rsp) +{ + return rsp == rcu_sysidle_state; +} + +/* + * Return a delay in jiffies based on the number of CPUs, rcu_node + * leaf fanout, and jiffies tick rate. The idea is to allow larger + * systems more time to transition to full-idle state in order to + * avoid the cache thrashing that otherwise occur on the state variable. + * Really small systems (less than a couple of tens of CPUs) should + * instead use a single global atomically incremented counter, and later + * versions of this will automatically reconfigure themselves accordingly. + */ +static unsigned long rcu_sysidle_delay(void) +{ + if (nr_cpu_ids <= CONFIG_NO_HZ_FULL_SYSIDLE_SMALL) + return 0; + return DIV_ROUND_UP(nr_cpu_ids * HZ, rcu_fanout_leaf * 1000); +} + +/* + * Advance the full-system-idle state. This is invoked when all of + * the non-timekeeping CPUs are idle. + */ +static void rcu_sysidle(unsigned long j) +{ + /* Check the current state. */ + switch (ACCESS_ONCE(full_sysidle_state)) { + case RCU_SYSIDLE_NOT: + + /* First time all are idle, so note a short idle period. */ + ACCESS_ONCE(full_sysidle_state) = RCU_SYSIDLE_SHORT; + break; + + case RCU_SYSIDLE_SHORT: + + /* + * Idle for a bit, time to advance to next state? + * cmpxchg failure means race with non-idle, let them win. + */ + if (ULONG_CMP_GE(jiffies, j + rcu_sysidle_delay())) + (void)cmpxchg(&full_sysidle_state, + RCU_SYSIDLE_SHORT, RCU_SYSIDLE_LONG); + break; + + case RCU_SYSIDLE_LONG: + + /* + * Do an additional check pass before advancing to full. + * cmpxchg failure means race with non-idle, let them win. + */ + if (ULONG_CMP_GE(jiffies, j + rcu_sysidle_delay())) + (void)cmpxchg(&full_sysidle_state, + RCU_SYSIDLE_LONG, RCU_SYSIDLE_FULL); + break; + + default: + break; + } +} + +/* + * Found a non-idle non-timekeeping CPU, so kick the system-idle state + * back to the beginning. + */ +static void rcu_sysidle_cancel(void) +{ + smp_mb(); + ACCESS_ONCE(full_sysidle_state) = RCU_SYSIDLE_NOT; +} + +/* + * Update the sysidle state based on the results of a force-quiescent-state + * scan of the CPUs' dyntick-idle state. + */ +static void rcu_sysidle_report(struct rcu_state *rsp, int isidle, + unsigned long maxj, bool gpkt) +{ + if (rsp != rcu_sysidle_state) + return; /* Wrong flavor, ignore. */ + if (gpkt && nr_cpu_ids <= CONFIG_NO_HZ_FULL_SYSIDLE_SMALL) + return; /* Running state machine from timekeeping CPU. */ + if (isidle) + rcu_sysidle(maxj); /* More idle! */ + else + rcu_sysidle_cancel(); /* Idle is over. */ +} + +/* + * Wrapper for rcu_sysidle_report() when called from the grace-period + * kthread's context. + */ +static void rcu_sysidle_report_gp(struct rcu_state *rsp, int isidle, + unsigned long maxj) +{ + rcu_sysidle_report(rsp, isidle, maxj, true); +} + +/* Callback and function for forcing an RCU grace period. */ +struct rcu_sysidle_head { + struct rcu_head rh; + int inuse; +}; + +static void rcu_sysidle_cb(struct rcu_head *rhp) +{ + struct rcu_sysidle_head *rshp; + + /* + * The following memory barrier is needed to replace the + * memory barriers that would normally be in the memory + * allocator. + */ + smp_mb(); /* grace period precedes setting inuse. */ + + rshp = container_of(rhp, struct rcu_sysidle_head, rh); + ACCESS_ONCE(rshp->inuse) = 0; +} + +/* + * Check to see if the system is fully idle, other than the timekeeping CPU. + * The caller must have disabled interrupts. + */ +bool rcu_sys_is_idle(void) +{ + static struct rcu_sysidle_head rsh; + int rss = ACCESS_ONCE(full_sysidle_state); + + if (WARN_ON_ONCE(smp_processor_id() != tick_do_timer_cpu)) + return false; + + /* Handle small-system case by doing a full scan of CPUs. */ + if (nr_cpu_ids <= CONFIG_NO_HZ_FULL_SYSIDLE_SMALL) { + int oldrss = rss - 1; + + /* + * One pass to advance to each state up to _FULL. + * Give up if any pass fails to advance the state. + */ + while (rss < RCU_SYSIDLE_FULL && oldrss < rss) { + int cpu; + bool isidle = true; + unsigned long maxj = jiffies - ULONG_MAX / 4; + struct rcu_data *rdp; + + /* Scan all the CPUs looking for nonidle CPUs. */ + for_each_possible_cpu(cpu) { + rdp = per_cpu_ptr(rcu_sysidle_state->rda, cpu); + rcu_sysidle_check_cpu(rdp, &isidle, &maxj); + if (!isidle) + break; + } + rcu_sysidle_report(rcu_sysidle_state, + isidle, maxj, false); + oldrss = rss; + rss = ACCESS_ONCE(full_sysidle_state); + } + } + + /* If this is the first observation of an idle period, record it. */ + if (rss == RCU_SYSIDLE_FULL) { + rss = cmpxchg(&full_sysidle_state, + RCU_SYSIDLE_FULL, RCU_SYSIDLE_FULL_NOTED); + return rss == RCU_SYSIDLE_FULL; + } + + smp_mb(); /* ensure rss load happens before later caller actions. */ + + /* If already fully idle, tell the caller (in case of races). */ + if (rss == RCU_SYSIDLE_FULL_NOTED) + return true; + + /* + * If we aren't there yet, and a grace period is not in flight, + * initiate a grace period. Either way, tell the caller that + * we are not there yet. We use an xchg() rather than an assignment + * to make up for the memory barriers that would otherwise be + * provided by the memory allocator. + */ + if (nr_cpu_ids > CONFIG_NO_HZ_FULL_SYSIDLE_SMALL && + !rcu_gp_in_progress(rcu_sysidle_state) && + !rsh.inuse && xchg(&rsh.inuse, 1) == 0) + call_rcu(&rsh.rh, rcu_sysidle_cb); + return false; } /* @@ -2483,6 +2756,21 @@ static void rcu_sysidle_exit(struct rcu_dynticks *rdtp, int irq) { } +static void rcu_sysidle_check_cpu(struct rcu_data *rdp, bool *isidle, + unsigned long *maxj) +{ +} + +static bool is_sysidle_rcu_state(struct rcu_state *rsp) +{ + return false; +} + +static void rcu_sysidle_report_gp(struct rcu_state *rsp, int isidle, + unsigned long maxj) +{ +} + static void rcu_sysidle_init_percpu_data(struct rcu_dynticks *rdtp) { } diff --git a/kernel/time/Kconfig b/kernel/time/Kconfig index c7d2fd67799e..3381f098070f 100644 --- a/kernel/time/Kconfig +++ b/kernel/time/Kconfig @@ -157,6 +157,33 @@ config NO_HZ_FULL_SYSIDLE Say N if you are unsure. +config NO_HZ_FULL_SYSIDLE_SMALL + int "Number of CPUs above which large-system approach is used" + depends on NO_HZ_FULL_SYSIDLE + range 1 NR_CPUS + default 8 + help + The full-system idle detection mechanism takes a lazy approach + on large systems, as is required to attain decent scalability. + However, on smaller systems, scalability is not anywhere near as + large a concern as is energy efficiency. The sysidle subsystem + therefore uses a fast but non-scalable algorithm for small + systems and a lazier but scalable algorithm for large systems. + This Kconfig parameter defines the number of CPUs in the largest + system that will be considered to be "small". + + The default value will be fine in most cases. Battery-powered + systems that (1) enable NO_HZ_FULL_SYSIDLE, (2) have larger + numbers of CPUs, and (3) are suffering from battery-lifetime + problems due to long sysidle latencies might wish to experiment + with larger values for this Kconfig parameter. On the other + hand, they might be even better served by disabling NO_HZ_FULL + entirely, given that NO_HZ_FULL is intended for HPC and + real-time workloads that at present do not tend to be run on + battery-powered systems. + + Take the default if you are unsure. + config NO_HZ bool "Old Idle dynticks config" depends on !ARCH_USES_GETTIMEOFFSET && GENERIC_CLOCKEVENTS -- cgit v1.2.3 From eb75767be0e514f97bf1b5cec763696cfc7f7e2a Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Fri, 21 Jun 2013 17:10:40 -0700 Subject: nohz_full: Force RCU's grace-period kthreads onto timekeeping CPU Because RCU's quiescent-state-forcing mechanism is used to drive the full-system-idle state machine, and because this mechanism is executed by RCU's grace-period kthreads, this commit forces these kthreads to run on the timekeeping CPU (tick_do_timer_cpu). To do otherwise would mean that the RCU grace-period kthreads would force the system into non-idle state every time they drove the state machine, which would be just a bit on the futile side. Signed-off-by: Paul E. McKenney Cc: Frederic Weisbecker Cc: Steven Rostedt Cc: Lai Jiangshan Reviewed-by: Josh Triplett --- kernel/rcutree.c | 1 + kernel/rcutree.h | 1 + kernel/rcutree_plugin.h | 21 ++++++++++++++++++++- 3 files changed, 22 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/rcutree.c b/kernel/rcutree.c index eca70f4469c1..64eaafb6c8f7 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1303,6 +1303,7 @@ static int rcu_gp_init(struct rcu_state *rsp) struct rcu_data *rdp; struct rcu_node *rnp = rcu_get_root(rsp); + rcu_bind_gp_kthread(); raw_spin_lock_irq(&rnp->lock); rsp->gp_flags = 0; /* Clear all flags: New grace period. */ diff --git a/kernel/rcutree.h b/kernel/rcutree.h index 6fd3659cf01a..5f97eab602cd 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -560,6 +560,7 @@ static void rcu_sysidle_check_cpu(struct rcu_data *rdp, bool *isidle, static bool is_sysidle_rcu_state(struct rcu_state *rsp); static void rcu_sysidle_report_gp(struct rcu_state *rsp, int isidle, unsigned long maxj); +static void rcu_bind_gp_kthread(void); static void rcu_sysidle_init_percpu_data(struct rcu_dynticks *rdtp); #endif /* #ifndef RCU_TREE_NONCORE */ diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 45ebba747af4..130c97b027f2 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -2531,7 +2531,8 @@ static void rcu_sysidle_check_cpu(struct rcu_data *rdp, bool *isidle, if (!*isidle || rdp->rsp != rcu_sysidle_state || cpu_is_offline(rdp->cpu) || rdp->cpu == tick_do_timer_cpu) return; - /* WARN_ON_ONCE(smp_processor_id() != tick_do_timer_cpu); */ + if (rcu_gp_in_progress(rdp->rsp)) + WARN_ON_ONCE(smp_processor_id() != tick_do_timer_cpu); /* Pick up current idle and NMI-nesting counter and check. */ cur = atomic_read(&rdtp->dynticks_idle); @@ -2556,6 +2557,20 @@ static bool is_sysidle_rcu_state(struct rcu_state *rsp) return rsp == rcu_sysidle_state; } +/* + * Bind the grace-period kthread for the sysidle flavor of RCU to the + * timekeeping CPU. + */ +static void rcu_bind_gp_kthread(void) +{ + int cpu = ACCESS_ONCE(tick_do_timer_cpu); + + if (cpu < 0 || cpu >= nr_cpu_ids) + return; + if (raw_smp_processor_id() != cpu) + set_cpus_allowed_ptr(current, cpumask_of(cpu)); +} + /* * Return a delay in jiffies based on the number of CPUs, rcu_node * leaf fanout, and jiffies tick rate. The idea is to allow larger @@ -2766,6 +2781,10 @@ static bool is_sysidle_rcu_state(struct rcu_state *rsp) return false; } +static void rcu_bind_gp_kthread(void) +{ +} + static void rcu_sysidle_report_gp(struct rcu_state *rsp, int isidle, unsigned long maxj) { -- cgit v1.2.3 From ae23bff1d71f8b416ed740bc458df67355c77c92 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Sat, 24 Aug 2013 16:45:54 +0200 Subject: perf: Prevent race in unthrottling code The current throttling code triggers WARN below via following workload (only hit on AMD machine with 48 CPUs): # while [ 1 ]; do perf record perf bench sched messaging; done WARNING: at arch/x86/kernel/cpu/perf_event.c:1054 x86_pmu_start+0xc6/0x100() SNIP Call Trace: [] dump_stack+0x19/0x1b [] warn_slowpath_common+0x61/0x80 [] warn_slowpath_null+0x1a/0x20 [] x86_pmu_start+0xc6/0x100 [] perf_adjust_freq_unthr_context.part.75+0x182/0x1a0 [] perf_event_task_tick+0xc8/0xf0 [] scheduler_tick+0xd1/0x140 [] update_process_times+0x66/0x80 [] tick_sched_handle.isra.15+0x25/0x60 [] tick_sched_timer+0x41/0x60 [] __run_hrtimer+0x74/0x1d0 [] ? tick_sched_handle.isra.15+0x60/0x60 [] hrtimer_interrupt+0xf7/0x240 [] smp_apic_timer_interrupt+0x69/0x9c [] apic_timer_interrupt+0x6d/0x80 [] ? __perf_event_task_sched_in+0x184/0x1a0 [] ? kfree_skbmem+0x37/0x90 [] ? __slab_free+0x1ac/0x30f [] ? kfree+0xfd/0x130 [] kmem_cache_free+0x1b2/0x1d0 [] kfree_skbmem+0x37/0x90 [] consume_skb+0x34/0x80 [] unix_stream_recvmsg+0x4e7/0x820 [] sock_aio_read.part.7+0x116/0x130 [] ? __perf_sw_event+0x19c/0x1e0 [] sock_aio_read+0x21/0x30 [] do_sync_read+0x80/0xb0 [] vfs_read+0x145/0x170 [] SyS_read+0x49/0xa0 [] ? __audit_syscall_exit+0x1f6/0x2a0 [] system_call_fastpath+0x16/0x1b ---[ end trace 622b7e226c4a766a ]--- The reason is a race in perf_event_task_tick() throttling code. The race flow (simplified code): - perf_throttled_count is per cpu variable and is CPU throttling flag, here starting with 0 - perf_throttled_seq is sequence/domain for allowed count of interrupts within the tick, gets increased each tick on single CPU (CPU bounded event): ... workload perf_event_task_tick: | | T0 inc(perf_throttled_seq) | T1 needs_unthr = xchg(perf_throttled_count, 0) == 0 tick gets interrupted: ... event gets throttled under new seq ... T2 last NMI comes, event is throttled - inc(perf_throttled_count) back to tick: | perf_adjust_freq_unthr_context: | | T3 unthrottling is skiped for event (needs_unthr == 0) | T4 event is stop and started via freq adjustment | tick ends ... workload ... no sample is hit for event ... perf_event_task_tick: | | T5 needs_unthr = xchg(perf_throttled_count, 0) != 0 (from T2) | T6 unthrottling is done on event (interrupts == MAX_INTERRUPTS) | event is already started (from T4) -> WARN Fixing this by not checking needs_unthr again and thus check all events for unthrottling. Signed-off-by: Jiri Olsa Reported-by: Jan Stancek Suggested-by: Peter Zijlstra Cc: Corey Ashford Cc: Frederic Weisbecker Cc: Namhyung Kim Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Andi Kleen Cc: Stephane Eranian Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1377355554-8934-1-git-send-email-jolsa@redhat.com Signed-off-by: Ingo Molnar --- kernel/events/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index f86599e8c123..258eaaffe95a 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -2712,7 +2712,7 @@ static void perf_adjust_freq_unthr_context(struct perf_event_context *ctx, hwc = &event->hw; - if (needs_unthr && hwc->interrupts == MAX_INTERRUPTS) { + if (hwc->interrupts == MAX_INTERRUPTS) { hwc->interrupts = 0; perf_log_throttle(event, 1); event->pmu->start(event, 0); -- cgit v1.2.3 From 95a79b805b935f4a7b685aa8a117d916c638323e Mon Sep 17 00:00:00 2001 From: Joonsoo Kim Date: Tue, 6 Aug 2013 17:36:41 +0900 Subject: sched: Remove one division operation in find_busiest_queue() Remove one division operation in find_busiest_queue() by using crosswise multiplication: wl_i / power_i > wl_j / power_j := wl_i * power_j > wl_j * power_i Signed-off-by: Joonsoo Kim [ Expanded the changelog. ] Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1375778203-31343-2-git-send-email-iamjoonsoo.kim@lge.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index f918635efe09..8aa217f62a9e 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -4968,7 +4968,7 @@ static struct rq *find_busiest_queue(struct lb_env *env, struct sched_group *group) { struct rq *busiest = NULL, *rq; - unsigned long max_load = 0; + unsigned long busiest_load = 0, busiest_power = 1; int i; for_each_cpu(i, sched_group_cpus(group)) { @@ -4998,11 +4998,15 @@ static struct rq *find_busiest_queue(struct lb_env *env, * the weighted_cpuload() scaled with the cpu power, so that * the load can be moved away from the cpu that is potentially * running at a lower capacity. + * + * Thus we're looking for max(wl_i / power_i), crosswise + * multiplication to rid ourselves of the division works out + * to: wl_i * power_j > wl_j * power_i; where j is our + * previous maximum. */ - wl = (wl * SCHED_POWER_SCALE) / power; - - if (wl > max_load) { - max_load = wl; + if (wl * busiest_power > busiest_load * power) { + busiest_load = wl; + busiest_power = power; busiest = rq; } } -- cgit v1.2.3 From 23f0d2093c789e612185180c468fa09063834e87 Mon Sep 17 00:00:00 2001 From: Joonsoo Kim Date: Tue, 6 Aug 2013 17:36:42 +0900 Subject: sched: Factor out code to should_we_balance() Now checking whether this cpu is appropriate to balance or not is embedded into update_sg_lb_stats() and this checking has no direct relationship to this function. There is not enough reason to place this checking at update_sg_lb_stats(), except saving one iteration for sched_group_cpus. In this patch, I factor out this checking to should_we_balance() function. And before doing actual work for load_balancing, check whether this cpu is appropriate to balance via should_we_balance(). If this cpu is not a candidate for balancing, it quit the work immediately. With this change, we can save two memset cost and can expect better compiler optimization. Below is result of this patch. * Vanilla * text data bss dec hex filename 34499 1136 116 35751 8ba7 kernel/sched/fair.o * Patched * text data bss dec hex filename 34243 1136 116 35495 8aa7 kernel/sched/fair.o In addition, rename @balance to @continue_balancing in order to represent its purpose more clearly. Signed-off-by: Joonsoo Kim [ s/should_balance/continue_balancing/g ] Reviewed-by: Paul Turner [ Made style changes and a fix in should_we_balance(). ] Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1375778203-31343-3-git-send-email-iamjoonsoo.kim@lge.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 107 ++++++++++++++++++++++++++-------------------------- 1 file changed, 53 insertions(+), 54 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 8aa217f62a9e..9a6daf86a76a 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -4463,22 +4463,17 @@ fix_small_capacity(struct sched_domain *sd, struct sched_group *group) * @group: sched_group whose statistics are to be updated. * @load_idx: Load index of sched_domain of this_cpu for load calc. * @local_group: Does group contain this_cpu. - * @balance: Should we balance. * @sgs: variable to hold the statistics for this group. */ static inline void update_sg_lb_stats(struct lb_env *env, struct sched_group *group, int load_idx, - int local_group, int *balance, struct sg_lb_stats *sgs) + int local_group, struct sg_lb_stats *sgs) { unsigned long nr_running, max_nr_running, min_nr_running; unsigned long load, max_cpu_load, min_cpu_load; - unsigned int balance_cpu = -1, first_idle_cpu = 0; unsigned long avg_load_per_task = 0; int i; - if (local_group) - balance_cpu = group_balance_cpu(group); - /* Tally up the load of all CPUs in the group */ max_cpu_load = 0; min_cpu_load = ~0UL; @@ -4492,12 +4487,6 @@ static inline void update_sg_lb_stats(struct lb_env *env, /* Bias balancing toward cpus of our domain */ if (local_group) { - if (idle_cpu(i) && !first_idle_cpu && - cpumask_test_cpu(i, sched_group_mask(group))) { - first_idle_cpu = 1; - balance_cpu = i; - } - load = target_load(i, load_idx); } else { load = source_load(i, load_idx); @@ -4519,22 +4508,9 @@ static inline void update_sg_lb_stats(struct lb_env *env, sgs->idle_cpus++; } - /* - * First idle cpu or the first cpu(busiest) in this sched group - * is eligible for doing load balancing at this and above - * domains. In the newly idle case, we will allow all the cpu's - * to do the newly idle load balance. - */ - if (local_group) { - if (env->idle != CPU_NEWLY_IDLE) { - if (balance_cpu != env->dst_cpu) { - *balance = 0; - return; - } - update_group_power(env->sd, env->dst_cpu); - } else if (time_after_eq(jiffies, group->sgp->next_update)) - update_group_power(env->sd, env->dst_cpu); - } + if (local_group && (env->idle != CPU_NEWLY_IDLE || + time_after_eq(jiffies, group->sgp->next_update))) + update_group_power(env->sd, env->dst_cpu); /* Adjust by relative CPU power of the group */ sgs->avg_load = (sgs->group_load*SCHED_POWER_SCALE) / group->sgp->power; @@ -4613,7 +4589,7 @@ static bool update_sd_pick_busiest(struct lb_env *env, * @sds: variable to hold the statistics for this sched_domain. */ static inline void update_sd_lb_stats(struct lb_env *env, - int *balance, struct sd_lb_stats *sds) + struct sd_lb_stats *sds) { struct sched_domain *child = env->sd->child; struct sched_group *sg = env->sd->groups; @@ -4630,10 +4606,7 @@ static inline void update_sd_lb_stats(struct lb_env *env, local_group = cpumask_test_cpu(env->dst_cpu, sched_group_cpus(sg)); memset(&sgs, 0, sizeof(sgs)); - update_sg_lb_stats(env, sg, load_idx, local_group, balance, &sgs); - - if (local_group && !(*balance)) - return; + update_sg_lb_stats(env, sg, load_idx, local_group, &sgs); sds->total_load += sgs.group_load; sds->total_pwr += sg->sgp->power; @@ -4866,8 +4839,6 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s * to restore balance. * * @env: The load balancing environment. - * @balance: Pointer to a variable indicating if this_cpu - * is the appropriate cpu to perform load balancing at this_level. * * Returns: - the busiest group if imbalance exists. * - If no imbalance and user has opted for power-savings balance, @@ -4875,7 +4846,7 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s * put to idle by rebalancing its tasks onto our group. */ static struct sched_group * -find_busiest_group(struct lb_env *env, int *balance) +find_busiest_group(struct lb_env *env) { struct sd_lb_stats sds; @@ -4885,14 +4856,7 @@ find_busiest_group(struct lb_env *env, int *balance) * Compute the various statistics relavent for load balancing at * this level. */ - update_sd_lb_stats(env, balance, &sds); - - /* - * this_cpu is not the appropriate cpu to perform load balancing at - * this level. - */ - if (!(*balance)) - goto ret; + update_sd_lb_stats(env, &sds); if ((env->idle == CPU_IDLE || env->idle == CPU_NEWLY_IDLE) && check_asym_packing(env, &sds)) @@ -4956,7 +4920,6 @@ force_balance: return sds.busiest; out_balanced: -ret: env->imbalance = 0; return NULL; } @@ -5043,13 +5006,47 @@ static int need_active_balance(struct lb_env *env) static int active_load_balance_cpu_stop(void *data); +static int should_we_balance(struct lb_env *env) +{ + struct sched_group *sg = env->sd->groups; + struct cpumask *sg_cpus, *sg_mask; + int cpu, balance_cpu = -1; + + /* + * In the newly idle case, we will allow all the cpu's + * to do the newly idle load balance. + */ + if (env->idle == CPU_NEWLY_IDLE) + return 1; + + sg_cpus = sched_group_cpus(sg); + sg_mask = sched_group_mask(sg); + /* Try to find first idle cpu */ + for_each_cpu_and(cpu, sg_cpus, env->cpus) { + if (!cpumask_test_cpu(cpu, sg_mask) || !idle_cpu(cpu)) + continue; + + balance_cpu = cpu; + break; + } + + if (balance_cpu == -1) + balance_cpu = group_balance_cpu(sg); + + /* + * First idle cpu or the first cpu(busiest) in this sched group + * is eligible for doing load balancing at this and above domains. + */ + return balance_cpu != env->dst_cpu; +} + /* * Check this_cpu to ensure it is balanced within domain. Attempt to move * tasks if there is an imbalance. */ static int load_balance(int this_cpu, struct rq *this_rq, struct sched_domain *sd, enum cpu_idle_type idle, - int *balance) + int *continue_balancing) { int ld_moved, cur_ld_moved, active_balance = 0; struct sched_group *group; @@ -5079,11 +5076,12 @@ static int load_balance(int this_cpu, struct rq *this_rq, schedstat_inc(sd, lb_count[idle]); redo: - group = find_busiest_group(&env, balance); - - if (*balance == 0) + if (!should_we_balance(&env)) { + *continue_balancing = 0; goto out_balanced; + } + group = find_busiest_group(&env); if (!group) { schedstat_inc(sd, lb_nobusyg[idle]); goto out_balanced; @@ -5296,7 +5294,7 @@ void idle_balance(int this_cpu, struct rq *this_rq) rcu_read_lock(); for_each_domain(this_cpu, sd) { unsigned long interval; - int balance = 1; + int continue_balancing = 1; if (!(sd->flags & SD_LOAD_BALANCE)) continue; @@ -5304,7 +5302,8 @@ void idle_balance(int this_cpu, struct rq *this_rq) if (sd->flags & SD_BALANCE_NEWIDLE) { /* If we've pulled tasks over stop searching: */ pulled_task = load_balance(this_cpu, this_rq, - sd, CPU_NEWLY_IDLE, &balance); + sd, CPU_NEWLY_IDLE, + &continue_balancing); } interval = msecs_to_jiffies(sd->balance_interval); @@ -5542,7 +5541,7 @@ void update_max_interval(void) */ static void rebalance_domains(int cpu, enum cpu_idle_type idle) { - int balance = 1; + int continue_balancing = 1; struct rq *rq = cpu_rq(cpu); unsigned long interval; struct sched_domain *sd; @@ -5574,7 +5573,7 @@ static void rebalance_domains(int cpu, enum cpu_idle_type idle) } if (time_after_eq(jiffies, sd->last_balance + interval)) { - if (load_balance(cpu, rq, sd, idle, &balance)) { + if (load_balance(cpu, rq, sd, idle, &continue_balancing)) { /* * The LBF_SOME_PINNED logic could have changed * env->dst_cpu, so we can't know our idle @@ -5597,7 +5596,7 @@ out: * CPU in our sched group which is doing load balancing more * actively. */ - if (!balance) + if (!continue_balancing) break; } rcu_read_unlock(); -- cgit v1.2.3 From 56cf515b4b1567c4e8fa9926175b40c66b9ec472 Mon Sep 17 00:00:00 2001 From: Joonsoo Kim Date: Tue, 6 Aug 2013 17:36:43 +0900 Subject: sched: Clean-up struct sd_lb_stat There is no reason to maintain separate variables for this_group and busiest_group in sd_lb_stat, except saving some space. But this structure is always allocated in stack, so this saving isn't really benificial [peterz: reducing stack space is good; in this case readability increases enough that I think its still beneficial] This patch unify these variables, so IMO, readability may be improved. Signed-off-by: Joonsoo Kim [ Rename this to local -- avoids confusion between this_cpu and the C++ this pointer. ] Reviewed-by: Paul Turner [ Lots of style edits, a few fixes and a rename. ] Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1375778203-31343-4-git-send-email-iamjoonsoo.kim@lge.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 229 ++++++++++++++++++++++++++-------------------------- 1 file changed, 114 insertions(+), 115 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 9a6daf86a76a..2da80a55827b 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -4231,36 +4231,6 @@ static unsigned long task_h_load(struct task_struct *p) #endif /********** Helpers for find_busiest_group ************************/ -/* - * sd_lb_stats - Structure to store the statistics of a sched_domain - * during load balancing. - */ -struct sd_lb_stats { - struct sched_group *busiest; /* Busiest group in this sd */ - struct sched_group *this; /* Local group in this sd */ - unsigned long total_load; /* Total load of all groups in sd */ - unsigned long total_pwr; /* Total power of all groups in sd */ - unsigned long avg_load; /* Average load across all groups in sd */ - - /** Statistics of this group */ - unsigned long this_load; - unsigned long this_load_per_task; - unsigned long this_nr_running; - unsigned long this_has_capacity; - unsigned int this_idle_cpus; - - /* Statistics of the busiest group */ - unsigned int busiest_idle_cpus; - unsigned long max_load; - unsigned long busiest_load_per_task; - unsigned long busiest_nr_running; - unsigned long busiest_group_capacity; - unsigned long busiest_has_capacity; - unsigned int busiest_group_weight; - - int group_imb; /* Is there imbalance in this sd */ -}; - /* * sg_lb_stats - stats of a sched_group required for load_balancing */ @@ -4269,6 +4239,7 @@ struct sg_lb_stats { unsigned long group_load; /* Total load over the CPUs of the group */ unsigned long sum_nr_running; /* Nr tasks running in the group */ unsigned long sum_weighted_load; /* Weighted load of group's tasks */ + unsigned long load_per_task; unsigned long group_capacity; unsigned long idle_cpus; unsigned long group_weight; @@ -4276,6 +4247,21 @@ struct sg_lb_stats { int group_has_capacity; /* Is there extra capacity in the group? */ }; +/* + * sd_lb_stats - Structure to store the statistics of a sched_domain + * during load balancing. + */ +struct sd_lb_stats { + struct sched_group *busiest; /* Busiest group in this sd */ + struct sched_group *local; /* Local group in this sd */ + unsigned long total_load; /* Total load of all groups in sd */ + unsigned long total_pwr; /* Total power of all groups in sd */ + unsigned long avg_load; /* Average load across all groups in sd */ + + struct sg_lb_stats local_stat; /* Statistics of the local group */ + struct sg_lb_stats busiest_stat;/* Statistics of the busiest group */ +}; + /** * get_sd_load_idx - Obtain the load index for a given sched domain. * @sd: The sched_domain whose load_idx is to be obtained. @@ -4490,6 +4476,7 @@ static inline void update_sg_lb_stats(struct lb_env *env, load = target_load(i, load_idx); } else { load = source_load(i, load_idx); + if (load > max_cpu_load) max_cpu_load = load; if (min_cpu_load > load) @@ -4531,10 +4518,12 @@ static inline void update_sg_lb_stats(struct lb_env *env, (max_nr_running - min_nr_running) > 1) sgs->group_imb = 1; - sgs->group_capacity = DIV_ROUND_CLOSEST(group->sgp->power, - SCHED_POWER_SCALE); + sgs->group_capacity = + DIV_ROUND_CLOSEST(group->sgp->power, SCHED_POWER_SCALE); + if (!sgs->group_capacity) sgs->group_capacity = fix_small_capacity(env->sd, group); + sgs->group_weight = group->group_weight; if (sgs->group_capacity > sgs->sum_nr_running) @@ -4556,7 +4545,7 @@ static bool update_sd_pick_busiest(struct lb_env *env, struct sched_group *sg, struct sg_lb_stats *sgs) { - if (sgs->avg_load <= sds->max_load) + if (sgs->avg_load <= sds->busiest_stat.avg_load) return false; if (sgs->sum_nr_running > sgs->group_capacity) @@ -4593,7 +4582,7 @@ static inline void update_sd_lb_stats(struct lb_env *env, { struct sched_domain *child = env->sd->child; struct sched_group *sg = env->sd->groups; - struct sg_lb_stats sgs; + struct sg_lb_stats tmp_sgs; int load_idx, prefer_sibling = 0; if (child && child->flags & SD_PREFER_SIBLING) @@ -4602,14 +4591,17 @@ static inline void update_sd_lb_stats(struct lb_env *env, load_idx = get_sd_load_idx(env->sd, env->idle); do { + struct sg_lb_stats *sgs = &tmp_sgs; int local_group; local_group = cpumask_test_cpu(env->dst_cpu, sched_group_cpus(sg)); - memset(&sgs, 0, sizeof(sgs)); - update_sg_lb_stats(env, sg, load_idx, local_group, &sgs); + if (local_group) { + sds->local = sg; + sgs = &sds->local_stat; + } - sds->total_load += sgs.group_load; - sds->total_pwr += sg->sgp->power; + memset(sgs, 0, sizeof(*sgs)); + update_sg_lb_stats(env, sg, load_idx, local_group, sgs); /* * In case the child domain prefers tasks go to siblings @@ -4621,26 +4613,17 @@ static inline void update_sd_lb_stats(struct lb_env *env, * heaviest group when it is already under-utilized (possible * with a large weight task outweighs the tasks on the system). */ - if (prefer_sibling && !local_group && sds->this_has_capacity) - sgs.group_capacity = min(sgs.group_capacity, 1UL); + if (prefer_sibling && !local_group && + sds->local && sds->local_stat.group_has_capacity) + sgs->group_capacity = min(sgs->group_capacity, 1UL); - if (local_group) { - sds->this_load = sgs.avg_load; - sds->this = sg; - sds->this_nr_running = sgs.sum_nr_running; - sds->this_load_per_task = sgs.sum_weighted_load; - sds->this_has_capacity = sgs.group_has_capacity; - sds->this_idle_cpus = sgs.idle_cpus; - } else if (update_sd_pick_busiest(env, sds, sg, &sgs)) { - sds->max_load = sgs.avg_load; + /* Now, start updating sd_lb_stats */ + sds->total_load += sgs->group_load; + sds->total_pwr += sg->sgp->power; + + if (!local_group && update_sd_pick_busiest(env, sds, sg, sgs)) { sds->busiest = sg; - sds->busiest_nr_running = sgs.sum_nr_running; - sds->busiest_idle_cpus = sgs.idle_cpus; - sds->busiest_group_capacity = sgs.group_capacity; - sds->busiest_load_per_task = sgs.sum_weighted_load; - sds->busiest_has_capacity = sgs.group_has_capacity; - sds->busiest_group_weight = sgs.group_weight; - sds->group_imb = sgs.group_imb; + sds->busiest_stat = *sgs; } sg = sg->next; @@ -4684,8 +4667,8 @@ static int check_asym_packing(struct lb_env *env, struct sd_lb_stats *sds) if (env->dst_cpu > busiest_cpu) return 0; - env->imbalance = DIV_ROUND_CLOSEST( - sds->max_load * sds->busiest->sgp->power, SCHED_POWER_SCALE); + env->imbalance = DIV_ROUND_CLOSEST(sds->busiest_stat.avg_load * + sds->busiest->sgp->power, SCHED_POWER_SCALE); return 1; } @@ -4703,24 +4686,23 @@ void fix_small_imbalance(struct lb_env *env, struct sd_lb_stats *sds) unsigned long tmp, pwr_now = 0, pwr_move = 0; unsigned int imbn = 2; unsigned long scaled_busy_load_per_task; + struct sg_lb_stats *local, *busiest; - if (sds->this_nr_running) { - sds->this_load_per_task /= sds->this_nr_running; - if (sds->busiest_load_per_task > - sds->this_load_per_task) - imbn = 1; - } else { - sds->this_load_per_task = - cpu_avg_load_per_task(env->dst_cpu); - } + local = &sds->local_stat; + busiest = &sds->busiest_stat; - scaled_busy_load_per_task = sds->busiest_load_per_task - * SCHED_POWER_SCALE; - scaled_busy_load_per_task /= sds->busiest->sgp->power; + if (!local->sum_nr_running) + local->load_per_task = cpu_avg_load_per_task(env->dst_cpu); + else if (busiest->load_per_task > local->load_per_task) + imbn = 1; - if (sds->max_load - sds->this_load + scaled_busy_load_per_task >= - (scaled_busy_load_per_task * imbn)) { - env->imbalance = sds->busiest_load_per_task; + scaled_busy_load_per_task = + (busiest->load_per_task * SCHED_POWER_SCALE) / + sds->busiest->sgp->power; + + if (busiest->avg_load - local->avg_load + scaled_busy_load_per_task >= + (scaled_busy_load_per_task * imbn)) { + env->imbalance = busiest->load_per_task; return; } @@ -4731,33 +4713,36 @@ void fix_small_imbalance(struct lb_env *env, struct sd_lb_stats *sds) */ pwr_now += sds->busiest->sgp->power * - min(sds->busiest_load_per_task, sds->max_load); - pwr_now += sds->this->sgp->power * - min(sds->this_load_per_task, sds->this_load); + min(busiest->load_per_task, busiest->avg_load); + pwr_now += sds->local->sgp->power * + min(local->load_per_task, local->avg_load); pwr_now /= SCHED_POWER_SCALE; /* Amount of load we'd subtract */ - tmp = (sds->busiest_load_per_task * SCHED_POWER_SCALE) / + tmp = (busiest->load_per_task * SCHED_POWER_SCALE) / sds->busiest->sgp->power; - if (sds->max_load > tmp) + if (busiest->avg_load > tmp) { pwr_move += sds->busiest->sgp->power * - min(sds->busiest_load_per_task, sds->max_load - tmp); + min(busiest->load_per_task, + busiest->avg_load - tmp); + } /* Amount of load we'd add */ - if (sds->max_load * sds->busiest->sgp->power < - sds->busiest_load_per_task * SCHED_POWER_SCALE) - tmp = (sds->max_load * sds->busiest->sgp->power) / - sds->this->sgp->power; - else - tmp = (sds->busiest_load_per_task * SCHED_POWER_SCALE) / - sds->this->sgp->power; - pwr_move += sds->this->sgp->power * - min(sds->this_load_per_task, sds->this_load + tmp); + if (busiest->avg_load * sds->busiest->sgp->power < + busiest->load_per_task * SCHED_POWER_SCALE) { + tmp = (busiest->avg_load * sds->busiest->sgp->power) / + sds->local->sgp->power; + } else { + tmp = (busiest->load_per_task * SCHED_POWER_SCALE) / + sds->local->sgp->power; + } + pwr_move += sds->local->sgp->power * + min(local->load_per_task, local->avg_load + tmp); pwr_move /= SCHED_POWER_SCALE; /* Move if we gain throughput */ if (pwr_move > pwr_now) - env->imbalance = sds->busiest_load_per_task; + env->imbalance = busiest->load_per_task; } /** @@ -4769,11 +4754,22 @@ void fix_small_imbalance(struct lb_env *env, struct sd_lb_stats *sds) static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *sds) { unsigned long max_pull, load_above_capacity = ~0UL; + struct sg_lb_stats *local, *busiest; + + local = &sds->local_stat; + if (local->sum_nr_running) { + local->load_per_task = + local->sum_weighted_load / local->sum_nr_running; + } - sds->busiest_load_per_task /= sds->busiest_nr_running; - if (sds->group_imb) { - sds->busiest_load_per_task = - min(sds->busiest_load_per_task, sds->avg_load); + busiest = &sds->busiest_stat; + /* busiest must have some tasks */ + busiest->load_per_task = + busiest->sum_weighted_load / busiest->sum_nr_running; + + if (busiest->group_imb) { + busiest->load_per_task = + min(busiest->load_per_task, sds->avg_load); } /* @@ -4781,20 +4777,19 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s * max load less than avg load(as we skip the groups at or below * its cpu_power, while calculating max_load..) */ - if (sds->max_load < sds->avg_load) { + if (busiest->avg_load < sds->avg_load) { env->imbalance = 0; return fix_small_imbalance(env, sds); } - if (!sds->group_imb) { + if (!busiest->group_imb) { /* * Don't want to pull so many tasks that a group would go idle. */ - load_above_capacity = (sds->busiest_nr_running - - sds->busiest_group_capacity); + load_above_capacity = + (busiest->sum_nr_running - busiest->group_capacity); load_above_capacity *= (SCHED_LOAD_SCALE * SCHED_POWER_SCALE); - load_above_capacity /= sds->busiest->sgp->power; } @@ -4808,12 +4803,14 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s * Be careful of negative numbers as they'll appear as very large values * with unsigned longs. */ - max_pull = min(sds->max_load - sds->avg_load, load_above_capacity); + max_pull = min(busiest->avg_load - sds->avg_load, + load_above_capacity); /* How much load to actually move to equalise the imbalance */ - env->imbalance = min(max_pull * sds->busiest->sgp->power, - (sds->avg_load - sds->this_load) * sds->this->sgp->power) - / SCHED_POWER_SCALE; + env->imbalance = min( + max_pull * sds->busiest->sgp->power, + (sds->avg_load - local->avg_load) * sds->local->sgp->power + ) / SCHED_POWER_SCALE; /* * if *imbalance is less than the average load per runnable task @@ -4821,9 +4818,8 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s * a think about bumping its value to force at least one task to be * moved */ - if (env->imbalance < sds->busiest_load_per_task) + if (env->imbalance < busiest->load_per_task) return fix_small_imbalance(env, sds); - } /******* find_busiest_group() helpers end here *********************/ @@ -4845,9 +4841,9 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s * return the least loaded group whose CPUs can be * put to idle by rebalancing its tasks onto our group. */ -static struct sched_group * -find_busiest_group(struct lb_env *env) +static struct sched_group *find_busiest_group(struct lb_env *env) { + struct sg_lb_stats *local, *busiest; struct sd_lb_stats sds; memset(&sds, 0, sizeof(sds)); @@ -4857,13 +4853,15 @@ find_busiest_group(struct lb_env *env) * this level. */ update_sd_lb_stats(env, &sds); + local = &sds.local_stat; + busiest = &sds.busiest_stat; if ((env->idle == CPU_IDLE || env->idle == CPU_NEWLY_IDLE) && check_asym_packing(env, &sds)) return sds.busiest; /* There is no busy sibling group to pull tasks from */ - if (!sds.busiest || sds.busiest_nr_running == 0) + if (!sds.busiest || busiest->sum_nr_running == 0) goto out_balanced; sds.avg_load = (SCHED_POWER_SCALE * sds.total_load) / sds.total_pwr; @@ -4873,26 +4871,26 @@ find_busiest_group(struct lb_env *env) * work because they assumes all things are equal, which typically * isn't true due to cpus_allowed constraints and the like. */ - if (sds.group_imb) + if (busiest->group_imb) goto force_balance; /* SD_BALANCE_NEWIDLE trumps SMP nice when underutilized */ - if (env->idle == CPU_NEWLY_IDLE && sds.this_has_capacity && - !sds.busiest_has_capacity) + if (env->idle == CPU_NEWLY_IDLE && local->group_has_capacity && + !busiest->group_has_capacity) goto force_balance; /* * If the local group is more busy than the selected busiest group * don't try and pull any tasks. */ - if (sds.this_load >= sds.max_load) + if (local->avg_load >= busiest->avg_load) goto out_balanced; /* * Don't pull any tasks if this group is already above the domain * average load. */ - if (sds.this_load >= sds.avg_load) + if (local->avg_load >= sds.avg_load) goto out_balanced; if (env->idle == CPU_IDLE) { @@ -4902,15 +4900,16 @@ find_busiest_group(struct lb_env *env) * there is no imbalance between this and busiest group * wrt to idle cpu's, it is balanced. */ - if ((sds.this_idle_cpus <= sds.busiest_idle_cpus + 1) && - sds.busiest_nr_running <= sds.busiest_group_weight) + if ((local->idle_cpus < busiest->idle_cpus) && + busiest->sum_nr_running <= busiest->group_weight) goto out_balanced; } else { /* * In the CPU_NEWLY_IDLE, CPU_NOT_IDLE cases, use * imbalance_pct to be conservative. */ - if (100 * sds.max_load <= env->sd->imbalance_pct * sds.this_load) + if (100 * busiest->avg_load <= + env->sd->imbalance_pct * local->avg_load) goto out_balanced; } -- cgit v1.2.3 From 147c5fc2bad780d8093b547f2baa204e78107faf Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 19 Aug 2013 15:22:57 +0200 Subject: sched/fair: Shrink sg_lb_stats and play memset games We can shrink sg_lb_stats because rq::nr_running is an unsigned int and cpu numbers are 'int' Before: sgs: /* size: 72, cachelines: 2, members: 10 */ sds: /* size: 184, cachelines: 3, members: 7 */ After: sgs: /* size: 56, cachelines: 1, members: 10 */ sds: /* size: 152, cachelines: 3, members: 7 */ Further we can avoid clearing all of sds since we do a total clear/assignment of sg_stats in update_sg_lb_stats() with exception of busiest_stat.avg_load which is referenced in update_sd_pick_busiest(). Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-0klzmz9okll8wc0nsudguc9p@git.kernel.org Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 33 ++++++++++++++++++++++++++------- 1 file changed, 26 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 2da80a55827b..4c6a8a5a789a 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -4237,12 +4237,12 @@ static unsigned long task_h_load(struct task_struct *p) struct sg_lb_stats { unsigned long avg_load; /*Avg load across the CPUs of the group */ unsigned long group_load; /* Total load over the CPUs of the group */ - unsigned long sum_nr_running; /* Nr tasks running in the group */ unsigned long sum_weighted_load; /* Weighted load of group's tasks */ unsigned long load_per_task; - unsigned long group_capacity; - unsigned long idle_cpus; - unsigned long group_weight; + unsigned int sum_nr_running; /* Nr tasks running in the group */ + unsigned int group_capacity; + unsigned int idle_cpus; + unsigned int group_weight; int group_imb; /* Is there an imbalance in the group ? */ int group_has_capacity; /* Is there extra capacity in the group? */ }; @@ -4258,10 +4258,29 @@ struct sd_lb_stats { unsigned long total_pwr; /* Total power of all groups in sd */ unsigned long avg_load; /* Average load across all groups in sd */ - struct sg_lb_stats local_stat; /* Statistics of the local group */ struct sg_lb_stats busiest_stat;/* Statistics of the busiest group */ + struct sg_lb_stats local_stat; /* Statistics of the local group */ }; +static inline void init_sd_lb_stats(struct sd_lb_stats *sds) +{ + /* + * Skimp on the clearing to avoid duplicate work. We can avoid clearing + * local_stat because update_sg_lb_stats() does a full clear/assignment. + * We must however clear busiest_stat::avg_load because + * update_sd_pick_busiest() reads this before assignment. + */ + *sds = (struct sd_lb_stats){ + .busiest = NULL, + .local = NULL, + .total_load = 0UL, + .total_pwr = 0UL, + .busiest_stat = { + .avg_load = 0UL, + }, + }; +} + /** * get_sd_load_idx - Obtain the load index for a given sched domain. * @sd: The sched_domain whose load_idx is to be obtained. @@ -4615,7 +4634,7 @@ static inline void update_sd_lb_stats(struct lb_env *env, */ if (prefer_sibling && !local_group && sds->local && sds->local_stat.group_has_capacity) - sgs->group_capacity = min(sgs->group_capacity, 1UL); + sgs->group_capacity = min(sgs->group_capacity, 1U); /* Now, start updating sd_lb_stats */ sds->total_load += sgs->group_load; @@ -4846,7 +4865,7 @@ static struct sched_group *find_busiest_group(struct lb_env *env) struct sg_lb_stats *local, *busiest; struct sd_lb_stats sds; - memset(&sds, 0, sizeof(sds)); + init_sd_lb_stats(&sds); /* * Compute the various statistics relavent for load balancing at -- cgit v1.2.3 From 38d0f7708543bcfa03d5ee55e8346f801b4a59c9 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 15 Aug 2013 19:47:56 +0200 Subject: sched/fair: Remove duplicate load_per_task computations Since we already compute (but don't store) the sgs load_per_task value in update_sg_lb_stats() we might as well store it and not re-compute it later on. Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-ym1vmljiwbzgdnnrwp9azftq@git.kernel.org Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 13 ++----------- 1 file changed, 2 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 4c6a8a5a789a..57952198b01e 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -4476,7 +4476,6 @@ static inline void update_sg_lb_stats(struct lb_env *env, { unsigned long nr_running, max_nr_running, min_nr_running; unsigned long load, max_cpu_load, min_cpu_load; - unsigned long avg_load_per_task = 0; int i; /* Tally up the load of all CPUs in the group */ @@ -4531,9 +4530,9 @@ static inline void update_sg_lb_stats(struct lb_env *env, * the hierarchy? */ if (sgs->sum_nr_running) - avg_load_per_task = sgs->sum_weighted_load / sgs->sum_nr_running; + sgs->load_per_task = sgs->sum_weighted_load / sgs->sum_nr_running; - if ((max_cpu_load - min_cpu_load) >= avg_load_per_task && + if ((max_cpu_load - min_cpu_load) >= sgs->load_per_task && (max_nr_running - min_nr_running) > 1) sgs->group_imb = 1; @@ -4776,15 +4775,7 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s struct sg_lb_stats *local, *busiest; local = &sds->local_stat; - if (local->sum_nr_running) { - local->load_per_task = - local->sum_weighted_load / local->sum_nr_running; - } - busiest = &sds->busiest_stat; - /* busiest must have some tasks */ - busiest->load_per_task = - busiest->sum_weighted_load / busiest->sum_nr_running; if (busiest->group_imb) { busiest->load_per_task = -- cgit v1.2.3 From 3ae11c90fd055ba1b1b03a014f851b395bdd26ff Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 15 Aug 2013 20:37:48 +0200 Subject: sched/fair: Make group power more consistent For easier access, less dereferences and more consistent value, store the group power in update_sg_lb_stats() and use it thereafter. The actual value in sched_group::sched_group_power::power can change throughout the load-balance pass if we're unlucky. Reviewed-by: Preeti U Murthy Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-739xxqkyvftrhnh9ncudutc7@git.kernel.org Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 41 ++++++++++++++++++++++------------------- 1 file changed, 22 insertions(+), 19 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 57952198b01e..ccf20e76b6b2 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -4239,6 +4239,7 @@ struct sg_lb_stats { unsigned long group_load; /* Total load over the CPUs of the group */ unsigned long sum_weighted_load; /* Weighted load of group's tasks */ unsigned long load_per_task; + unsigned long group_power; unsigned int sum_nr_running; /* Nr tasks running in the group */ unsigned int group_capacity; unsigned int idle_cpus; @@ -4518,7 +4519,8 @@ static inline void update_sg_lb_stats(struct lb_env *env, update_group_power(env->sd, env->dst_cpu); /* Adjust by relative CPU power of the group */ - sgs->avg_load = (sgs->group_load*SCHED_POWER_SCALE) / group->sgp->power; + sgs->group_power = group->sgp->power; + sgs->avg_load = (sgs->group_load*SCHED_POWER_SCALE) / sgs->group_power; /* * Consider the group unbalanced when the imbalance is larger @@ -4537,7 +4539,7 @@ static inline void update_sg_lb_stats(struct lb_env *env, sgs->group_imb = 1; sgs->group_capacity = - DIV_ROUND_CLOSEST(group->sgp->power, SCHED_POWER_SCALE); + DIV_ROUND_CLOSEST(sgs->group_power, SCHED_POWER_SCALE); if (!sgs->group_capacity) sgs->group_capacity = fix_small_capacity(env->sd, group); @@ -4637,7 +4639,7 @@ static inline void update_sd_lb_stats(struct lb_env *env, /* Now, start updating sd_lb_stats */ sds->total_load += sgs->group_load; - sds->total_pwr += sg->sgp->power; + sds->total_pwr += sgs->group_power; if (!local_group && update_sd_pick_busiest(env, sds, sg, sgs)) { sds->busiest = sg; @@ -4685,8 +4687,9 @@ static int check_asym_packing(struct lb_env *env, struct sd_lb_stats *sds) if (env->dst_cpu > busiest_cpu) return 0; - env->imbalance = DIV_ROUND_CLOSEST(sds->busiest_stat.avg_load * - sds->busiest->sgp->power, SCHED_POWER_SCALE); + env->imbalance = DIV_ROUND_CLOSEST( + sds->busiest_stat.avg_load * sds->busiest_stat.group_power, + SCHED_POWER_SCALE); return 1; } @@ -4716,7 +4719,7 @@ void fix_small_imbalance(struct lb_env *env, struct sd_lb_stats *sds) scaled_busy_load_per_task = (busiest->load_per_task * SCHED_POWER_SCALE) / - sds->busiest->sgp->power; + busiest->group_power; if (busiest->avg_load - local->avg_load + scaled_busy_load_per_task >= (scaled_busy_load_per_task * imbn)) { @@ -4730,32 +4733,32 @@ void fix_small_imbalance(struct lb_env *env, struct sd_lb_stats *sds) * moving them. */ - pwr_now += sds->busiest->sgp->power * + pwr_now += busiest->group_power * min(busiest->load_per_task, busiest->avg_load); - pwr_now += sds->local->sgp->power * + pwr_now += local->group_power * min(local->load_per_task, local->avg_load); pwr_now /= SCHED_POWER_SCALE; /* Amount of load we'd subtract */ tmp = (busiest->load_per_task * SCHED_POWER_SCALE) / - sds->busiest->sgp->power; + busiest->group_power; if (busiest->avg_load > tmp) { - pwr_move += sds->busiest->sgp->power * + pwr_move += busiest->group_power * min(busiest->load_per_task, busiest->avg_load - tmp); } /* Amount of load we'd add */ - if (busiest->avg_load * sds->busiest->sgp->power < + if (busiest->avg_load * busiest->group_power < busiest->load_per_task * SCHED_POWER_SCALE) { - tmp = (busiest->avg_load * sds->busiest->sgp->power) / - sds->local->sgp->power; + tmp = (busiest->avg_load * busiest->group_power) / + local->group_power; } else { tmp = (busiest->load_per_task * SCHED_POWER_SCALE) / - sds->local->sgp->power; + local->group_power; } - pwr_move += sds->local->sgp->power * - min(local->load_per_task, local->avg_load + tmp); + pwr_move += local->group_power * + min(local->load_per_task, local->avg_load + tmp); pwr_move /= SCHED_POWER_SCALE; /* Move if we gain throughput */ @@ -4800,7 +4803,7 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s (busiest->sum_nr_running - busiest->group_capacity); load_above_capacity *= (SCHED_LOAD_SCALE * SCHED_POWER_SCALE); - load_above_capacity /= sds->busiest->sgp->power; + load_above_capacity /= busiest->group_power; } /* @@ -4818,8 +4821,8 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s /* How much load to actually move to equalise the imbalance */ env->imbalance = min( - max_pull * sds->busiest->sgp->power, - (sds->avg_load - local->avg_load) * sds->local->sgp->power + max_pull * busiest->group_power, + (sds->avg_load - local->avg_load) * local->group_power ) / SCHED_POWER_SCALE; /* -- cgit v1.2.3 From 6906a40839198f33dbb56d20e644c01e00663952 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 19 Aug 2013 15:20:21 +0200 Subject: sched/fair: Optimize find_busiest_queue() Use for_each_cpu_and() and thereby avoid computing the capacity for CPUs we know we're not interested in. Reviewed-by: Paul Turner Reviewed-by: Preeti U Murthy Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-lppceyv6kb3a19g8spmrn20b@git.kernel.org Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index ccf20e76b6b2..bedd30b168a5 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -4946,7 +4946,7 @@ static struct rq *find_busiest_queue(struct lb_env *env, unsigned long busiest_load = 0, busiest_power = 1; int i; - for_each_cpu(i, sched_group_cpus(group)) { + for_each_cpu_and(i, sched_group_cpus(group), env->cpus) { unsigned long power = power_of(i); unsigned long capacity = DIV_ROUND_CLOSEST(power, SCHED_POWER_SCALE); @@ -4955,9 +4955,6 @@ static struct rq *find_busiest_queue(struct lb_env *env, if (!capacity) capacity = fix_small_capacity(env->sd, group); - if (!cpumask_test_cpu(i, env->cpus)) - continue; - rq = cpu_rq(i); wl = weighted_cpuload(i); -- cgit v1.2.3 From 30ce5dabc92b5a349a7d9e9cf499494d230e0691 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 15 Aug 2013 20:29:29 +0200 Subject: sched/fair: Rework and comment the group_imb code Rik reported some weirdness due to the group_imb code. As a start to looking at it, clean it up a little and add a few explanatory comments. Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-caeeqttnla4wrrmhp5uf89gp@git.kernel.org Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 123 +++++++++++++++++++++++++++++++++++++--------------- 1 file changed, 89 insertions(+), 34 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index bedd30b168a5..dffb27070ddb 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -4463,6 +4463,81 @@ fix_small_capacity(struct sched_domain *sd, struct sched_group *group) return 0; } +/* + * Group imbalance indicates (and tries to solve) the problem where balancing + * groups is inadequate due to tsk_cpus_allowed() constraints. + * + * Imagine a situation of two groups of 4 cpus each and 4 tasks each with a + * cpumask covering 1 cpu of the first group and 3 cpus of the second group. + * Something like: + * + * { 0 1 2 3 } { 4 5 6 7 } + * * * * * + * + * If we were to balance group-wise we'd place two tasks in the first group and + * two tasks in the second group. Clearly this is undesired as it will overload + * cpu 3 and leave one of the cpus in the second group unused. + * + * The current solution to this issue is detecting the skew in the first group + * by noticing it has a cpu that is overloaded while the remaining cpus are + * idle -- or rather, there's a distinct imbalance in the cpus; see + * sg_imbalanced(). + * + * When this is so detected; this group becomes a candidate for busiest; see + * update_sd_pick_busiest(). And calculcate_imbalance() and + * find_busiest_group() avoid some of the usual balance conditional to allow it + * to create an effective group imbalance. + * + * This is a somewhat tricky proposition since the next run might not find the + * group imbalance and decide the groups need to be balanced again. A most + * subtle and fragile situation. + */ + +struct sg_imb_stats { + unsigned long max_nr_running, min_nr_running; + unsigned long max_cpu_load, min_cpu_load; +}; + +static inline void init_sg_imb_stats(struct sg_imb_stats *sgi) +{ + sgi->max_cpu_load = sgi->max_nr_running = 0UL; + sgi->min_cpu_load = sgi->min_nr_running = ~0UL; +} + +static inline void +update_sg_imb_stats(struct sg_imb_stats *sgi, + unsigned long load, unsigned long nr_running) +{ + if (load > sgi->max_cpu_load) + sgi->max_cpu_load = load; + if (sgi->min_cpu_load > load) + sgi->min_cpu_load = load; + + if (nr_running > sgi->max_nr_running) + sgi->max_nr_running = nr_running; + if (sgi->min_nr_running > nr_running) + sgi->min_nr_running = nr_running; +} + +static inline int +sg_imbalanced(struct sg_lb_stats *sgs, struct sg_imb_stats *sgi) +{ + /* + * Consider the group unbalanced when the imbalance is larger + * than the average weight of a task. + * + * APZ: with cgroup the avg task weight can vary wildly and + * might not be a suitable number - should we keep a + * normalized nr_running number somewhere that negates + * the hierarchy? + */ + if ((sgi->max_cpu_load - sgi->min_cpu_load) >= sgs->load_per_task && + (sgi->max_nr_running - sgi->min_nr_running) > 1) + return 1; + + return 0; +} + /** * update_sg_lb_stats - Update sched_group's statistics for load balancing. * @env: The load balancing environment. @@ -4475,15 +4550,12 @@ static inline void update_sg_lb_stats(struct lb_env *env, struct sched_group *group, int load_idx, int local_group, struct sg_lb_stats *sgs) { - unsigned long nr_running, max_nr_running, min_nr_running; - unsigned long load, max_cpu_load, min_cpu_load; + struct sg_imb_stats sgi; + unsigned long nr_running; + unsigned long load; int i; - /* Tally up the load of all CPUs in the group */ - max_cpu_load = 0; - min_cpu_load = ~0UL; - max_nr_running = 0; - min_nr_running = ~0UL; + init_sg_imb_stats(&sgi); for_each_cpu_and(i, sched_group_cpus(group), env->cpus) { struct rq *rq = cpu_rq(i); @@ -4495,16 +4567,7 @@ static inline void update_sg_lb_stats(struct lb_env *env, load = target_load(i, load_idx); } else { load = source_load(i, load_idx); - - if (load > max_cpu_load) - max_cpu_load = load; - if (min_cpu_load > load) - min_cpu_load = load; - - if (nr_running > max_nr_running) - max_nr_running = nr_running; - if (min_nr_running > nr_running) - min_nr_running = nr_running; + update_sg_imb_stats(&sgi, load, nr_running); } sgs->group_load += load; @@ -4522,21 +4585,10 @@ static inline void update_sg_lb_stats(struct lb_env *env, sgs->group_power = group->sgp->power; sgs->avg_load = (sgs->group_load*SCHED_POWER_SCALE) / sgs->group_power; - /* - * Consider the group unbalanced when the imbalance is larger - * than the average weight of a task. - * - * APZ: with cgroup the avg task weight can vary wildly and - * might not be a suitable number - should we keep a - * normalized nr_running number somewhere that negates - * the hierarchy? - */ if (sgs->sum_nr_running) sgs->load_per_task = sgs->sum_weighted_load / sgs->sum_nr_running; - if ((max_cpu_load - min_cpu_load) >= sgs->load_per_task && - (max_nr_running - min_nr_running) > 1) - sgs->group_imb = 1; + sgs->group_imb = sg_imbalanced(sgs, &sgi); sgs->group_capacity = DIV_ROUND_CLOSEST(sgs->group_power, SCHED_POWER_SCALE); @@ -4781,6 +4833,10 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s busiest = &sds->busiest_stat; if (busiest->group_imb) { + /* + * In the group_imb case we cannot rely on group-wide averages + * to ensure cpu-load equilibrium, look at wider averages. XXX + */ busiest->load_per_task = min(busiest->load_per_task, sds->avg_load); } @@ -4798,6 +4854,8 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s if (!busiest->group_imb) { /* * Don't want to pull so many tasks that a group would go idle. + * Except of course for the group_imb case, since then we might + * have to drop below capacity to reach cpu-load equilibrium. */ load_above_capacity = (busiest->sum_nr_running - busiest->group_capacity); @@ -4813,11 +4871,8 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s * we also don't want to reduce the group load below the group capacity * (so that we can implement power-savings policies etc). Thus we look * for the minimum possible imbalance. - * Be careful of negative numbers as they'll appear as very large values - * with unsigned longs. */ - max_pull = min(busiest->avg_load - sds->avg_load, - load_above_capacity); + max_pull = min(busiest->avg_load - sds->avg_load, load_above_capacity); /* How much load to actually move to equalise the imbalance */ env->imbalance = min( @@ -4881,7 +4936,7 @@ static struct sched_group *find_busiest_group(struct lb_env *env) /* * If the busiest group is imbalanced the below checks don't - * work because they assumes all things are equal, which typically + * work because they assume all things are equal, which typically * isn't true due to cpus_allowed constraints and the like. */ if (busiest->group_imb) -- cgit v1.2.3 From 10866e62e8a6907d9072f10f9a0561db0c0cf50b Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 19 Aug 2013 16:57:04 +0200 Subject: sched/fair: Fix the sd_parent_degenerate() code I found that on my WSM box I had a redundant domain: [ 0.949769] CPU0 attaching sched-domain: [ 0.953765] domain 0: span 0,12 level SIBLING [ 0.958335] groups: 0 (cpu_power = 587) 12 (cpu_power = 588) [ 0.964548] domain 1: span 0-5,12-17 level MC [ 0.969206] groups: 0,12 (cpu_power = 1175) 1,13 (cpu_power = 1176) 2,14 (cpu_power = 1176) 3,15 (cpu_power = 1176) 4,16 (cpu_power = 1176) 5,17 (cpu_power = 1176) [ 0.984993] domain 2: span 0-5,12-17 level CPU [ 0.989822] groups: 0-5,12-17 (cpu_power = 7055) [ 0.995049] domain 3: span 0-23 level NUMA [ 0.999620] groups: 0-5,12-17 (cpu_power = 7055) 6-11,18-23 (cpu_power = 7056) Note how domain 2 has only a single group and spans the same CPUs as domain 1. We should not keep such domains and do in fact have code to prune these. It turns out that the 'new' SD_PREFER_SIBLING flag causes this, it makes sd_parent_degenerate() fail on the CPU domain. We can easily fix this by 'ignoring' the SD_PREFER_SIBLING bit and transfering it to whatever domain ends up covering the span. With this patch the domains now look like this: [ 0.950419] CPU0 attaching sched-domain: [ 0.954454] domain 0: span 0,12 level SIBLING [ 0.959039] groups: 0 (cpu_power = 587) 12 (cpu_power = 588) [ 0.965271] domain 1: span 0-5,12-17 level MC [ 0.969936] groups: 0,12 (cpu_power = 1175) 1,13 (cpu_power = 1176) 2,14 (cpu_power = 1176) 3,15 (cpu_power = 1176) 4,16 (cpu_power = 1176) 5,17 (cpu_power = 1176) [ 0.985737] domain 2: span 0-23 level NUMA [ 0.990231] groups: 0-5,12-17 (cpu_power = 7055) 6-11,18-23 (cpu_power = 7056) Reviewed-by: Paul Turner Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-ys201g4jwukj0h8xcamakxq1@git.kernel.org Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index cf8f100433e0..4da0f4bb2ca4 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -4914,7 +4914,8 @@ sd_parent_degenerate(struct sched_domain *sd, struct sched_domain *parent) SD_BALANCE_FORK | SD_BALANCE_EXEC | SD_SHARE_CPUPOWER | - SD_SHARE_PKG_RESOURCES); + SD_SHARE_PKG_RESOURCES | + SD_PREFER_SIBLING); if (nr_node_ids == 1) pflags &= ~SD_SERIALIZE; } @@ -5118,6 +5119,13 @@ cpu_attach_domain(struct sched_domain *sd, struct root_domain *rd, int cpu) tmp->parent = parent->parent; if (parent->parent) parent->parent->child = tmp; + /* + * Transfer SD_PREFER_SIBLING down in case of a + * degenerate parent; the spans match for this + * so the property transfers. + */ + if (parent->flags & SD_PREFER_SIBLING) + tmp->flags |= SD_PREFER_SIBLING; destroy_sched_domain(parent, cpu); } else tmp = tmp->parent; -- cgit v1.2.3 From 13d7a2410fa637f450a29ecb515ac318ee40c741 Mon Sep 17 00:00:00 2001 From: Stephane Eranian Date: Wed, 21 Aug 2013 12:10:24 +0200 Subject: perf: Add attr->mmap2 attribute to an event Adds a new PERF_RECORD_MMAP2 record type which is essence an expanded version of PERF_RECORD_MMAP. Used to request mmap records with more information about the mapping, including device major, minor and the inode number and generation for mappings associated with files or shared memory segments. Works for code and data (with attr->mmap_data set). Existing PERF_RECORD_MMAP record is unmodified by this patch. Signed-off-by: Stephane Eranian Signed-off-by: Peter Zijlstra Cc: Al Viro Link: http://lkml.kernel.org/r/1377079825-19057-2-git-send-email-eranian@google.com [ Added Al to the Cc:. Are the ino, maj/min exports of vma->vm_file OK? ] Signed-off-by: Ingo Molnar --- kernel/events/core.c | 46 ++++++++++++++++++++++++++++++++++++++++++---- 1 file changed, 42 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 15d0f2418e54..c7ee497c39a7 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -4776,7 +4776,7 @@ next: /* * task tracking -- fork/exit * - * enabled by: attr.comm | attr.mmap | attr.mmap_data | attr.task + * enabled by: attr.comm | attr.mmap | attr.mmap2 | attr.mmap_data | attr.task */ struct perf_task_event { @@ -4796,8 +4796,9 @@ struct perf_task_event { static int perf_event_task_match(struct perf_event *event) { - return event->attr.comm || event->attr.mmap || - event->attr.mmap_data || event->attr.task; + return event->attr.comm || event->attr.mmap || + event->attr.mmap2 || event->attr.mmap_data || + event->attr.task; } static void perf_event_task_output(struct perf_event *event, @@ -4992,6 +4993,9 @@ struct perf_mmap_event { const char *file_name; int file_size; + int maj, min; + u64 ino; + u64 ino_generation; struct { struct perf_event_header header; @@ -5012,7 +5016,7 @@ static int perf_event_mmap_match(struct perf_event *event, int executable = vma->vm_flags & VM_EXEC; return (!executable && event->attr.mmap_data) || - (executable && event->attr.mmap); + (executable && (event->attr.mmap || event->attr.mmap2)); } static void perf_event_mmap_output(struct perf_event *event, @@ -5027,6 +5031,13 @@ static void perf_event_mmap_output(struct perf_event *event, if (!perf_event_mmap_match(event, data)) return; + if (event->attr.mmap2) { + mmap_event->event_id.header.type = PERF_RECORD_MMAP2; + mmap_event->event_id.header.size += sizeof(mmap_event->maj); + mmap_event->event_id.header.size += sizeof(mmap_event->min); + mmap_event->event_id.header.size += sizeof(mmap_event->ino); + } + perf_event_header__init_id(&mmap_event->event_id.header, &sample, event); ret = perf_output_begin(&handle, event, mmap_event->event_id.header.size); @@ -5037,6 +5048,14 @@ static void perf_event_mmap_output(struct perf_event *event, mmap_event->event_id.tid = perf_event_tid(event, current); perf_output_put(&handle, mmap_event->event_id); + + if (event->attr.mmap2) { + perf_output_put(&handle, mmap_event->maj); + perf_output_put(&handle, mmap_event->min); + perf_output_put(&handle, mmap_event->ino); + perf_output_put(&handle, mmap_event->ino_generation); + } + __output_copy(&handle, mmap_event->file_name, mmap_event->file_size); @@ -5051,6 +5070,8 @@ static void perf_event_mmap_event(struct perf_mmap_event *mmap_event) { struct vm_area_struct *vma = mmap_event->vma; struct file *file = vma->vm_file; + int maj = 0, min = 0; + u64 ino = 0, gen = 0; unsigned int size; char tmp[16]; char *buf = NULL; @@ -5059,6 +5080,8 @@ static void perf_event_mmap_event(struct perf_mmap_event *mmap_event) memset(tmp, 0, sizeof(tmp)); if (file) { + struct inode *inode; + dev_t dev; /* * d_path works from the end of the rb backwards, so we * need to add enough zero bytes after the string to handle @@ -5074,6 +5097,13 @@ static void perf_event_mmap_event(struct perf_mmap_event *mmap_event) name = strncpy(tmp, "//toolong", sizeof(tmp)); goto got_name; } + inode = file_inode(vma->vm_file); + dev = inode->i_sb->s_dev; + ino = inode->i_ino; + gen = inode->i_generation; + maj = MAJOR(dev); + min = MINOR(dev); + } else { if (arch_vma_name(mmap_event->vma)) { name = strncpy(tmp, arch_vma_name(mmap_event->vma), @@ -5104,6 +5134,10 @@ got_name: mmap_event->file_name = name; mmap_event->file_size = size; + mmap_event->maj = maj; + mmap_event->min = min; + mmap_event->ino = ino; + mmap_event->ino_generation = gen; if (!(vma->vm_flags & VM_EXEC)) mmap_event->event_id.header.misc |= PERF_RECORD_MISC_MMAP_DATA; @@ -5140,6 +5174,10 @@ void perf_event_mmap(struct vm_area_struct *vma) .len = vma->vm_end - vma->vm_start, .pgoff = (u64)vma->vm_pgoff << PAGE_SHIFT, }, + /* .maj (attr_mmap2 only) */ + /* .min (attr_mmap2 only) */ + /* .ino (attr_mmap2 only) */ + /* .ino_generation (attr_mmap2 only) */ }; perf_event_mmap_event(&mmap_event); -- cgit v1.2.3 From 942e443127e928a5631c3d5102aca8c8b3c2dd98 Mon Sep 17 00:00:00 2001 From: Li Zhong Date: Tue, 3 Sep 2013 16:33:57 +0930 Subject: module: Fix mod->mkobj.kobj potentially freed too early DEBUG_KOBJECT_RELEASE helps to find the issue attached below. After some investigation, it seems the reason is: The mod->mkobj.kobj(ffffffffa01600d0 below) is freed together with mod itself in free_module(). However, its children still hold references to it, as the delay caused by DEBUG_KOBJECT_RELEASE. So when the child(holders below) tries to decrease the reference count to its parent in kobject_del(), BUG happens as it tries to access already freed memory. This patch tries to fix it by waiting for the mod->mkobj.kobj to be really released in the module removing process (and some error code paths). [ 1844.175287] kobject: 'holders' (ffff88007c1f1600): kobject_release, parent ffffffffa01600d0 (delayed) [ 1844.178991] kobject: 'notes' (ffff8800370b2a00): kobject_release, parent ffffffffa01600d0 (delayed) [ 1845.180118] kobject: 'holders' (ffff88007c1f1600): kobject_cleanup, parent ffffffffa01600d0 [ 1845.182130] kobject: 'holders' (ffff88007c1f1600): auto cleanup kobject_del [ 1845.184120] BUG: unable to handle kernel paging request at ffffffffa01601d0 [ 1845.185026] IP: [] kobject_put+0x11/0x60 [ 1845.185026] PGD 1a13067 PUD 1a14063 PMD 7bd30067 PTE 0 [ 1845.185026] Oops: 0000 [#1] PREEMPT [ 1845.185026] Modules linked in: xfs libcrc32c [last unloaded: kprobe_example] [ 1845.185026] CPU: 0 PID: 18 Comm: kworker/0:1 Tainted: G O 3.11.0-rc6-next-20130819+ #1 [ 1845.185026] Hardware name: Bochs Bochs, BIOS Bochs 01/01/2007 [ 1845.185026] Workqueue: events kobject_delayed_cleanup [ 1845.185026] task: ffff88007ca51f00 ti: ffff88007ca5c000 task.ti: ffff88007ca5c000 [ 1845.185026] RIP: 0010:[] [] kobject_put+0x11/0x60 [ 1845.185026] RSP: 0018:ffff88007ca5dd08 EFLAGS: 00010282 [ 1845.185026] RAX: 0000000000002000 RBX: ffffffffa01600d0 RCX: ffffffff8177d638 [ 1845.185026] RDX: ffff88007ca5dc18 RSI: 0000000000000000 RDI: ffffffffa01600d0 [ 1845.185026] RBP: ffff88007ca5dd18 R08: ffffffff824e9810 R09: ffffffffffffffff [ 1845.185026] R10: ffff8800ffffffff R11: dead4ead00000001 R12: ffffffff81a95040 [ 1845.185026] R13: ffff88007b27a960 R14: ffff88007c1f1600 R15: 0000000000000000 [ 1845.185026] FS: 0000000000000000(0000) GS:ffffffff81a23000(0000) knlGS:0000000000000000 [ 1845.185026] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b [ 1845.185026] CR2: ffffffffa01601d0 CR3: 0000000037207000 CR4: 00000000000006b0 [ 1845.185026] Stack: [ 1845.185026] ffff88007c1f1600 ffff88007c1f1600 ffff88007ca5dd38 ffffffff812cdb7e [ 1845.185026] 0000000000000000 ffff88007c1f1640 ffff88007ca5dd68 ffffffff812cdbfe [ 1845.185026] ffff88007c974800 ffff88007c1f1640 ffff88007ff61a00 0000000000000000 [ 1845.185026] Call Trace: [ 1845.185026] [] kobject_del+0x2e/0x40 [ 1845.185026] [] kobject_delayed_cleanup+0x6e/0x1d0 [ 1845.185026] [] process_one_work+0x1e5/0x670 [ 1845.185026] [] ? process_one_work+0x183/0x670 [ 1845.185026] [] worker_thread+0x113/0x370 [ 1845.185026] [] ? rescuer_thread+0x290/0x290 [ 1845.185026] [] kthread+0xda/0xe0 [ 1845.185026] [] ? _raw_spin_unlock_irq+0x30/0x60 [ 1845.185026] [] ? kthread_create_on_node+0x130/0x130 [ 1845.185026] [] ret_from_fork+0x7a/0xb0 [ 1845.185026] [] ? kthread_create_on_node+0x130/0x130 [ 1845.185026] Code: 81 48 c7 c7 28 95 ad 81 31 c0 e8 9b da 01 00 e9 4f ff ff ff 66 0f 1f 44 00 00 55 48 89 e5 53 48 89 fb 48 83 ec 08 48 85 ff 74 1d 87 00 01 00 00 01 74 1e 48 8d 7b 38 83 6b 38 01 0f 94 c0 84 [ 1845.185026] RIP [] kobject_put+0x11/0x60 [ 1845.185026] RSP [ 1845.185026] CR2: ffffffffa01601d0 [ 1845.185026] ---[ end trace 49a70afd109f5653 ]--- Signed-off-by: Li Zhong Acked-by: Greg Kroah-Hartman Signed-off-by: Rusty Russell --- kernel/module.c | 14 +++++++++++--- kernel/params.c | 7 +++++++ 2 files changed, 18 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/module.c b/kernel/module.c index 40ee1dc3c3bf..9f5ddae72f44 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -1612,6 +1612,14 @@ static void module_remove_modinfo_attrs(struct module *mod) kfree(mod->modinfo_attrs); } +static void mod_kobject_put(struct module *mod) +{ + DECLARE_COMPLETION_ONSTACK(c); + mod->mkobj.kobj_completion = &c; + kobject_put(&mod->mkobj.kobj); + wait_for_completion(&c); +} + static int mod_sysfs_init(struct module *mod) { int err; @@ -1639,7 +1647,7 @@ static int mod_sysfs_init(struct module *mod) err = kobject_init_and_add(&mod->mkobj.kobj, &module_ktype, NULL, "%s", mod->name); if (err) - kobject_put(&mod->mkobj.kobj); + mod_kobject_put(mod); /* delay uevent until full sysfs population */ out: @@ -1683,7 +1691,7 @@ out_unreg_param: out_unreg_holders: kobject_put(mod->holders_dir); out_unreg: - kobject_put(&mod->mkobj.kobj); + mod_kobject_put(mod); out: return err; } @@ -1692,7 +1700,7 @@ static void mod_sysfs_fini(struct module *mod) { remove_notes_attrs(mod); remove_sect_attrs(mod); - kobject_put(&mod->mkobj.kobj); + mod_kobject_put(mod); } #else /* !CONFIG_SYSFS */ diff --git a/kernel/params.c b/kernel/params.c index e5f8f17e57cf..501bde4f3bee 100644 --- a/kernel/params.c +++ b/kernel/params.c @@ -915,7 +915,14 @@ static const struct kset_uevent_ops module_uevent_ops = { struct kset *module_kset; int module_sysfs_initialized; +static void module_kobj_release(struct kobject *kobj) +{ + struct module_kobject *mk = to_module_kobject(kobj); + complete(mk->kobj_completion); +} + struct kobj_type module_ktype = { + .release = module_kobj_release, .sysfs_ops = &module_sysfs_ops, }; -- cgit v1.2.3 From 59338f754a55f07857342dbcd81652a4f091d72f Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Sat, 31 Aug 2013 01:04:07 -0400 Subject: ftrace: Fix a slight race in modifying what function callback gets traced There's a slight race when going from a list function to a non list function. That is, when only one callback is registered to the function tracer, it gets called directly by the mcount trampoline. But if this function has filters, it may be called by the wrong functions. As the list ops callback that handles multiple callbacks that are registered to ftrace, it also handles what functions they call. While the transaction is taking place, use the list function always, and after all the updates are finished (only the functions that should be traced are being traced), then we can update the trampoline to call the function directly. Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 17 ++++++++++++++++- 1 file changed, 16 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index a6d098c6df3f..03cf44ac54d3 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1978,12 +1978,27 @@ int __weak ftrace_arch_code_modify_post_process(void) void ftrace_modify_all_code(int command) { + int update = command & FTRACE_UPDATE_TRACE_FUNC; + + /* + * If the ftrace_caller calls a ftrace_ops func directly, + * we need to make sure that it only traces functions it + * expects to trace. When doing the switch of functions, + * we need to update to the ftrace_ops_list_func first + * before the transition between old and new calls are set, + * as the ftrace_ops_list_func will check the ops hashes + * to make sure the ops are having the right functions + * traced. + */ + if (update) + ftrace_update_ftrace_func(ftrace_ops_list_func); + if (command & FTRACE_UPDATE_CALLS) ftrace_replace_code(1); else if (command & FTRACE_DISABLE_CALLS) ftrace_replace_code(0); - if (command & FTRACE_UPDATE_TRACE_FUNC) + if (update && ftrace_trace_function != ftrace_ops_list_func) ftrace_update_ftrace_func(ftrace_trace_function); if (command & FTRACE_START_FUNC_RET) -- cgit v1.2.3 From a2e0578be3652406b2ffd2eeb31cdbdffa325d64 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 30 Aug 2013 12:41:41 -0400 Subject: switch copy_module_from_fd() to fdget Signed-off-by: Al Viro --- kernel/module.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/module.c b/kernel/module.c index 206915830d29..c6756d1c6d73 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -2540,21 +2540,20 @@ static int copy_module_from_user(const void __user *umod, unsigned long len, /* Sets info->hdr and info->len. */ static int copy_module_from_fd(int fd, struct load_info *info) { - struct file *file; + struct fd f = fdget(fd); int err; struct kstat stat; loff_t pos; ssize_t bytes = 0; - file = fget(fd); - if (!file) + if (!f.file) return -ENOEXEC; - err = security_kernel_module_from_file(file); + err = security_kernel_module_from_file(f.file); if (err) goto out; - err = vfs_getattr(&file->f_path, &stat); + err = vfs_getattr(&f.file->f_path, &stat); if (err) goto out; @@ -2577,7 +2576,7 @@ static int copy_module_from_fd(int fd, struct load_info *info) pos = 0; while (pos < stat.size) { - bytes = kernel_read(file, pos, (char *)(info->hdr) + pos, + bytes = kernel_read(f.file, pos, (char *)(info->hdr) + pos, stat.size - pos); if (bytes < 0) { vfree(info->hdr); @@ -2591,7 +2590,7 @@ static int copy_module_from_fd(int fd, struct load_info *info) info->len = pos; out: - fput(file); + fdput(f); return err; } -- cgit v1.2.3 From 5a8e01f8fa51f5cbce8f37acc050eb2319d12956 Mon Sep 17 00:00:00 2001 From: Stanislaw Gruszka Date: Wed, 4 Sep 2013 15:16:03 +0200 Subject: sched/cputime: Do not scale when utime == 0 scale_stime() silently assumes that stime < rtime, otherwise when stime == rtime and both values are big enough (operations on them do not fit in 32 bits), the resulting scaling stime can be bigger than rtime. In consequence utime = rtime - stime results in negative value. User space visible symptoms of the bug are overflowed TIME values on ps/top, for example: $ ps aux | grep rcu root 8 0.0 0.0 0 0 ? S 12:42 0:00 [rcuc/0] root 9 0.0 0.0 0 0 ? S 12:42 0:00 [rcub/0] root 10 62422329 0.0 0 0 ? R 12:42 21114581:37 [rcu_preempt] root 11 0.1 0.0 0 0 ? S 12:42 0:02 [rcuop/0] root 12 62422329 0.0 0 0 ? S 12:42 21114581:35 [rcuop/1] root 10 62422329 0.0 0 0 ? R 12:42 21114581:37 [rcu_preempt] or overflowed utime values read directly from /proc/$PID/stat Reference: https://lkml.org/lkml/2013/8/20/259 Reported-and-tested-by: Sergey Senozhatsky Signed-off-by: Stanislaw Gruszka Cc: stable@vger.kernel.org Cc: Frederic Weisbecker Cc: Peter Zijlstra Cc: Paul E. McKenney Cc: Borislav Petkov Link: http://lkml.kernel.org/r/20130904131602.GC2564@redhat.com Signed-off-by: Ingo Molnar --- kernel/sched/cputime.c | 19 +++++++++++-------- 1 file changed, 11 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c index c1d7493825ae..5b03f5bebabc 100644 --- a/kernel/sched/cputime.c +++ b/kernel/sched/cputime.c @@ -551,10 +551,7 @@ static void cputime_adjust(struct task_cputime *curr, struct cputime *prev, cputime_t *ut, cputime_t *st) { - cputime_t rtime, stime, utime, total; - - stime = curr->stime; - total = stime + curr->utime; + cputime_t rtime, stime, utime; /* * Tick based cputime accounting depend on random scheduling @@ -576,13 +573,19 @@ static void cputime_adjust(struct task_cputime *curr, if (prev->stime + prev->utime >= rtime) goto out; - if (total) { + stime = curr->stime; + utime = curr->utime; + + if (utime == 0) { + stime = rtime; + } else if (stime == 0) { + utime = rtime; + } else { + cputime_t total = stime + utime; + stime = scale_stime((__force u64)stime, (__force u64)rtime, (__force u64)total); utime = rtime - stime; - } else { - stime = rtime; - utime = 0; } /* -- cgit v1.2.3 From a0a5a0561f63905fe94c49bc567615829f42ce1e Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (Red Hat)" Date: Sat, 31 Aug 2013 01:04:07 -0400 Subject: ftrace/rcu: Do not trace debug_lockdep_rcu_enabled() The function debug_lockdep_rcu_enabled() is part of the RCU lockdep debugging, and is called very frequently. I found that if I enable a lot of debugging and run the function graph tracer, this function can cause a live lock of the system. We don't usually trace lockdep infrastructure, no need to trace this either. Reviewed-by: Paul E. McKenney Signed-off-by: Steven Rostedt --- kernel/rcupdate.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c index cce6ba8bbace..4f20c6c6a848 100644 --- a/kernel/rcupdate.c +++ b/kernel/rcupdate.c @@ -122,7 +122,7 @@ struct lockdep_map rcu_sched_lock_map = STATIC_LOCKDEP_MAP_INIT("rcu_read_lock_sched", &rcu_sched_lock_key); EXPORT_SYMBOL_GPL(rcu_sched_lock_map); -int debug_lockdep_rcu_enabled(void) +int notrace debug_lockdep_rcu_enabled(void) { return rcu_scheduler_active && debug_locks && current->lockdep_recursion == 0; -- cgit v1.2.3 From 4e10f3c98888ee88ea2543aa636db6410fa47477 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 30 Aug 2013 12:29:49 -0400 Subject: Kill indirect include of file.h from eventfd.h, use fdget() in cgroup.c kernel/cgroup.c is the only place in the tree that relies on eventfd.h pulling file.h; move that include there. Switch from eventfd_fget()/fput() to fdget()/fdput(), while we are at it - eventfd_ctx_fileget() will fail on non-eventfd descriptors just fine, no need to do that check twice... Signed-off-by: Al Viro --- kernel/cgroup.c | 33 +++++++++++++++++---------------- 1 file changed, 17 insertions(+), 16 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index e0aeb32415ff..2418b6e71a85 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -60,6 +60,7 @@ #include #include /* used in cgroup_attach_task */ #include +#include #include @@ -4034,8 +4035,8 @@ static int cgroup_write_event_control(struct cgroup_subsys_state *dummy_css, struct cgroup_event *event; struct cgroup_subsys_state *cfile_css; unsigned int efd, cfd; - struct file *efile; - struct file *cfile; + struct fd efile; + struct fd cfile; char *endp; int ret; @@ -4058,31 +4059,31 @@ static int cgroup_write_event_control(struct cgroup_subsys_state *dummy_css, init_waitqueue_func_entry(&event->wait, cgroup_event_wake); INIT_WORK(&event->remove, cgroup_event_remove); - efile = eventfd_fget(efd); - if (IS_ERR(efile)) { - ret = PTR_ERR(efile); + efile = fdget(efd); + if (!efile.file) { + ret = -EBADF; goto out_kfree; } - event->eventfd = eventfd_ctx_fileget(efile); + event->eventfd = eventfd_ctx_fileget(efile.file); if (IS_ERR(event->eventfd)) { ret = PTR_ERR(event->eventfd); goto out_put_efile; } - cfile = fget(cfd); - if (!cfile) { + cfile = fdget(cfd); + if (!cfile.file) { ret = -EBADF; goto out_put_eventfd; } /* the process need read permission on control file */ /* AV: shouldn't we check that it's been opened for read instead? */ - ret = inode_permission(file_inode(cfile), MAY_READ); + ret = inode_permission(file_inode(cfile.file), MAY_READ); if (ret < 0) goto out_put_cfile; - event->cft = __file_cft(cfile); + event->cft = __file_cft(cfile.file); if (IS_ERR(event->cft)) { ret = PTR_ERR(event->cft); goto out_put_cfile; @@ -4103,7 +4104,7 @@ static int cgroup_write_event_control(struct cgroup_subsys_state *dummy_css, ret = -EINVAL; event->css = cgroup_css(cgrp, event->cft->ss); - cfile_css = css_from_dir(cfile->f_dentry->d_parent, event->cft->ss); + cfile_css = css_from_dir(cfile.file->f_dentry->d_parent, event->cft->ss); if (event->css && event->css == cfile_css && css_tryget(event->css)) ret = 0; @@ -4121,25 +4122,25 @@ static int cgroup_write_event_control(struct cgroup_subsys_state *dummy_css, if (ret) goto out_put_css; - efile->f_op->poll(efile, &event->pt); + efile.file->f_op->poll(efile.file, &event->pt); spin_lock(&cgrp->event_list_lock); list_add(&event->list, &cgrp->event_list); spin_unlock(&cgrp->event_list_lock); - fput(cfile); - fput(efile); + fdput(cfile); + fdput(efile); return 0; out_put_css: css_put(event->css); out_put_cfile: - fput(cfile); + fdput(cfile); out_put_eventfd: eventfd_ctx_put(event->eventfd); out_put_efile: - fput(efile); + fdput(efile); out_kfree: kfree(event); -- cgit v1.2.3 From b0cff9d88ce2f3030f73138078c5b1019f17e1cc Mon Sep 17 00:00:00 2001 From: Joonsoo Kim Date: Tue, 10 Sep 2013 15:54:49 +0900 Subject: sched: Fix load balancing performance regression in should_we_balance() Commit 23f0d20 ("sched: Factor out code to should_we_balance()") introduces the should_we_balance() function. This function should return 1 if this cpu is appropriate for balancing. But the newly introduced code doesn't do so, it returns 0 instead of 1. This introduces performance regression, reported by Dave Chinner: v4 filesystem v5 filesystem 3.11+xfsdev: 220k files/s 225k files/s 3.12-git 180k files/s 185k files/s 3.12-git-revert 245k files/s 247k files/s You can find more detailed information at: https://lkml.org/lkml/2013/9/10/1 This patch corrects the return value of should_we_balance() function as orignally intended. With this patch, Dave Chinner reports that the regression is gone: v4 filesystem v5 filesystem 3.11+xfsdev: 220k files/s 225k files/s 3.12-git 180k files/s 185k files/s 3.12-git-revert 245k files/s 247k files/s 3.12-git-fix 249k files/s 248k files/s Reported-by: Dave Chinner Tested-by: Dave Chinner Signed-off-by: Joonsoo Kim Cc: Paul Turner Cc: Peter Zijlstra Cc: Dave Chinner Link: http://lkml.kernel.org/r/20130910065448.GA20368@lge.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 7f0a5e6cdae0..9b3fe1cd8f40 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -5151,7 +5151,7 @@ static int should_we_balance(struct lb_env *env) * First idle cpu or the first cpu(busiest) in this sched group * is eligible for doing load balancing at this and above domains. */ - return balance_cpu != env->dst_cpu; + return balance_cpu == env->dst_cpu; } /* -- cgit v1.2.3 From 3942c07ccf98e66b8893f396dca98f5b076f905f Mon Sep 17 00:00:00 2001 From: Glauber Costa Date: Wed, 28 Aug 2013 10:17:53 +1000 Subject: fs: bump inode and dentry counters to long MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This series reworks our current object cache shrinking infrastructure in two main ways: * Noticing that a lot of users copy and paste their own version of LRU lists for objects, we put some effort in providing a generic version. It is modeled after the filesystem users: dentries, inodes, and xfs (for various tasks), but we expect that other users could benefit in the near future with little or no modification. Let us know if you have any issues. * The underlying list_lru being proposed automatically and transparently keeps the elements in per-node lists, and is able to manipulate the node lists individually. Given this infrastructure, we are able to modify the up-to-now hammer called shrink_slab to proceed with node-reclaim instead of always searching memory from all over like it has been doing. Per-node lru lists are also expected to lead to less contention in the lru locks on multi-node scans, since we are now no longer fighting for a global lock. The locks usually disappear from the profilers with this change. Although we have no official benchmarks for this version - be our guest to independently evaluate this - earlier versions of this series were performance tested (details at http://permalink.gmane.org/gmane.linux.kernel.mm/100537) yielding no visible performance regressions while yielding a better qualitative behavior in NUMA machines. With this infrastructure in place, we can use the list_lru entry point to provide memcg isolation and per-memcg targeted reclaim. Historically, those two pieces of work have been posted together. This version presents only the infrastructure work, deferring the memcg work for a later time, so we can focus on getting this part tested. You can see more about the history of such work at http://lwn.net/Articles/552769/ Dave Chinner (18): dcache: convert dentry_stat.nr_unused to per-cpu counters dentry: move to per-sb LRU locks dcache: remove dentries from LRU before putting on dispose list mm: new shrinker API shrinker: convert superblock shrinkers to new API list: add a new LRU list type inode: convert inode lru list to generic lru list code. dcache: convert to use new lru list infrastructure list_lru: per-node list infrastructure shrinker: add node awareness fs: convert inode and dentry shrinking to be node aware xfs: convert buftarg LRU to generic code xfs: rework buffer dispose list tracking xfs: convert dquot cache lru to list_lru fs: convert fs shrinkers to new scan/count API drivers: convert shrinkers to new count/scan API shrinker: convert remaining shrinkers to count/scan API shrinker: Kill old ->shrink API. Glauber Costa (7): fs: bump inode and dentry counters to long super: fix calculation of shrinkable objects for small numbers list_lru: per-node API vmscan: per-node deferred work i915: bail out earlier when shrinker cannot acquire mutex hugepage: convert huge zero page shrinker to new shrinker API list_lru: dynamically adjust node arrays This patch: There are situations in very large machines in which we can have a large quantity of dirty inodes, unused dentries, etc. This is particularly true when umounting a filesystem, where eventually since every live object will eventually be discarded. Dave Chinner reported a problem with this while experimenting with the shrinker revamp patchset. So we believe it is time for a change. This patch just moves int to longs. Machines where it matters should have a big long anyway. Signed-off-by: Glauber Costa Cc: Dave Chinner Cc: "Theodore Ts'o" Cc: Adrian Hunter Cc: Al Viro Cc: Artem Bityutskiy Cc: Arve Hjønnevåg Cc: Carlos Maiolino Cc: Christoph Hellwig Cc: Chuck Lever Cc: Daniel Vetter Cc: Dave Chinner Cc: David Rientjes Cc: Gleb Natapov Cc: Greg Thelen Cc: J. Bruce Fields Cc: Jan Kara Cc: Jerome Glisse Cc: John Stultz Cc: KAMEZAWA Hiroyuki Cc: Kent Overstreet Cc: Kirill A. Shutemov Cc: Marcelo Tosatti Cc: Mel Gorman Cc: Steven Whitehouse Cc: Thomas Hellstrom Cc: Trond Myklebust Signed-off-by: Andrew Morton Signed-off-by: Al Viro --- kernel/sysctl.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 07f6fc468e17..7822cd88a95c 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -1471,14 +1471,14 @@ static struct ctl_table fs_table[] = { { .procname = "inode-nr", .data = &inodes_stat, - .maxlen = 2*sizeof(int), + .maxlen = 2*sizeof(long), .mode = 0444, .proc_handler = proc_nr_inodes, }, { .procname = "inode-state", .data = &inodes_stat, - .maxlen = 7*sizeof(int), + .maxlen = 7*sizeof(long), .mode = 0444, .proc_handler = proc_nr_inodes, }, @@ -1508,7 +1508,7 @@ static struct ctl_table fs_table[] = { { .procname = "dentry-state", .data = &dentry_stat, - .maxlen = 6*sizeof(int), + .maxlen = 6*sizeof(long), .mode = 0444, .proc_handler = proc_nr_dentry, }, -- cgit v1.2.3 From d008d5258e9c1a1b7ee6547b8d444323aef331b3 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 10 Sep 2013 10:24:05 -0300 Subject: perf: Fix up MMAP2 buffer space reservation The ino_generation field was added in the PERF_RECORD_MMAP2 record in the 13d7a24 cset but no space for it was allocated, corrupting the PERF_FORMAT_{TIME,CPU,TID,etc} area (sample_type/sample_id_all), fix it. Detected with one of the regression tests done by 'perf test': [root@sandy ~]# perf test -v 7 7: Validate PERF_RECORD_* events & perf_sample fields : --- start --- 61315294449606 0 PERF_RECORD_SAMPLE 61315294453161 0 PERF_RECORD_SAMPLE 61315294454441 0 PERF_RECORD_SAMPLE 61315294455709 0 PERF_RECORD_SAMPLE 61315295600899 0 PERF_RECORD_COMM: sleep:6500 27917287430500 342521613 PERF_RECORD_MMAP2 6500/6500: [0x400000(0x7000) @ 0 00:1d 311442 9016]: /usr/bin/sleep MMAP2 going backwards in time, prev=61315295600899, curr=27917287430500 MMAP2 with unexpected cpu, expected 0, got 342521613 MMAP2 with unexpected pid, expected 6500, got 1701606191 MMAP2 with unexpected tid, expected 6500, got 28773 27917287430500 342561333 PERF_RECORD_MMAP2 6500/6500: [0x3b7e000000(0x223000) @ 0 00:1d 309186 9016]: /usr/lib64/ld-2.16.so MMAP2 with unexpected cpu, expected 0, got 342561333 MMAP2 with unexpected pid, expected 6500, got 1932408369 MMAP2 with unexpected tid, expected 6500, got 111 27917287430500 342600095 PERF_RECORD_MMAP2 6500/6500: [0x7fffbd7dc000(0x1000) @ 0x7fffbd7dc000 00:00 0 0]: [vdso] MMAP2 with unexpected cpu, expected 0, got 342600095 MMAP2 with unexpected pid, expected 6500, got 1935963739 MMAP2 with unexpected tid, expected 6500, got 23919 27917287430500 342882834 PERF_RECORD_MMAP2 6500/6500: [0x3b7e400000(0x3b8000) @ 0 00:1d 309187 9016]: /usr/lib64/libc-2.16.so MMAP2 with unexpected cpu, expected 0, got 342882834 MMAP2 with unexpected pid, expected 6500, got 909192754 MMAP2 with unexpected tid, expected 6500, got 7303982 61316297195411 0 PERF_RECORD_EXIT(6500:6500):(6500:6500) ---- end ---- Validate PERF_RECORD_* events & perf_sample fields: FAILED! [root@sandy ~]# After this patch: [root@sandy ~]# perf test 7 7: Validate PERF_RECORD_* events & perf_sample fields : Ok [root@sandy ~]# Acked-by: Peter Zijlstra Acked-by: Stephane Eranian Cc: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Mike Galbraith Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/n/tip-heeuv986b8ha7whqg4o3he7c@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- kernel/events/core.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 2207efc941d1..dd236b66ca3a 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -5039,6 +5039,7 @@ static void perf_event_mmap_output(struct perf_event *event, mmap_event->event_id.header.size += sizeof(mmap_event->maj); mmap_event->event_id.header.size += sizeof(mmap_event->min); mmap_event->event_id.header.size += sizeof(mmap_event->ino); + mmap_event->event_id.header.size += sizeof(mmap_event->ino_generation); } perf_event_header__init_id(&mmap_event->event_id.header, &sample, event); -- cgit v1.2.3 From e79f525e99b04390ca4d2366309545a836c03bf1 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Wed, 11 Sep 2013 14:19:38 -0700 Subject: pidns: fix vfork() after unshare(CLONE_NEWPID) Commit 8382fcac1b81 ("pidns: Outlaw thread creation after unshare(CLONE_NEWPID)") nacks CLONE_VM if the forking process unshared pid_ns, this obviously breaks vfork: int main(void) { assert(unshare(CLONE_NEWUSER | CLONE_NEWPID) == 0); assert(vfork() >= 0); _exit(0); return 0; } fails without this patch. Change this check to use CLONE_SIGHAND instead. This also forbids CLONE_THREAD automatically, and this is what the comment implies. We could probably even drop CLONE_SIGHAND and use CLONE_THREAD, but it would be safer to not do this. The current check denies CLONE_SIGHAND implicitely and there is no reason to change this. Eric said "CLONE_SIGHAND is fine. CLONE_THREAD would be even better. Having shared signal handling between two different pid namespaces is the case that we are fundamentally guarding against." Signed-off-by: Oleg Nesterov Reported-by: Colin Walters Acked-by: Andy Lutomirski Reviewed-by: "Eric W. Biederman" Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index c9eaf2013002..3561391ca450 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1173,10 +1173,11 @@ static struct task_struct *copy_process(unsigned long clone_flags, return ERR_PTR(-EINVAL); /* - * If the new process will be in a different pid namespace - * don't allow the creation of threads. + * If the new process will be in a different pid namespace don't + * allow it to share a thread group or signal handlers with the + * forking task. */ - if ((clone_flags & (CLONE_VM|CLONE_NEWPID)) && + if ((clone_flags & (CLONE_SIGHAND | CLONE_NEWPID)) && (task_active_pid_ns(current) != current->nsproxy->pid_ns_for_children)) return ERR_PTR(-EINVAL); -- cgit v1.2.3 From 5167246a8ad617df55717c2d901da5e2aedffcfa Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Wed, 11 Sep 2013 14:19:40 -0700 Subject: pidns: kill the unnecessary CLONE_NEWPID in copy_process() Commit 8382fcac1b81 ("pidns: Outlaw thread creation after unshare(CLONE_NEWPID)") nacks CLONE_NEWPID if the forking process unshared pid_ns. This is correct but unnecessary, copy_pid_ns() does the same check. Remove the CLONE_NEWPID check to cleanup the code and prepare for the next change. Test-case: static int child(void *arg) { return 0; } static char stack[16 * 1024]; int main(void) { pid_t pid; assert(unshare(CLONE_NEWUSER | CLONE_NEWPID) == 0); pid = clone(child, stack + sizeof(stack) / 2, CLONE_NEWPID | SIGCHLD, NULL); assert(pid < 0 && errno == EINVAL); return 0; } clone(CLONE_NEWPID) correctly fails with or without this change. Signed-off-by: Oleg Nesterov Acked-by: Andy Lutomirski Cc: "Eric W. Biederman" Cc: Colin Walters Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index 3561391ca450..68d508f2bfba 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1177,9 +1177,8 @@ static struct task_struct *copy_process(unsigned long clone_flags, * allow it to share a thread group or signal handlers with the * forking task. */ - if ((clone_flags & (CLONE_SIGHAND | CLONE_NEWPID)) && - (task_active_pid_ns(current) != - current->nsproxy->pid_ns_for_children)) + if ((clone_flags & CLONE_SIGHAND) && (task_active_pid_ns(current) != + current->nsproxy->pid_ns_for_children)) return ERR_PTR(-EINVAL); retval = security_task_create(clone_flags); -- cgit v1.2.3 From 40a0d32d1eaffe6aac7324ca92604b6b3977eb0e Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Wed, 11 Sep 2013 14:19:41 -0700 Subject: fork: unify and tighten up CLONE_NEWUSER/CLONE_NEWPID checks do_fork() denies CLONE_THREAD | CLONE_PARENT if NEWUSER | NEWPID. Then later copy_process() denies CLONE_SIGHAND if the new process will be in a different pid namespace (task_active_pid_ns() doesn't match current->nsproxy->pid_ns). This looks confusing and inconsistent. CLONE_NEWPID is very similar to the case when ->pid_ns was already unshared, we want the same restrictions so copy_process() should also nack CLONE_PARENT. And it would be better to deny CLONE_NEWUSER && CLONE_SIGHAND as well just for consistency. Kill the "CLONE_NEWUSER | CLONE_NEWPID" check in do_fork() and change copy_process() to do the same check along with ->pid_ns check we already have. Signed-off-by: Oleg Nesterov Acked-by: Andy Lutomirski Cc: "Eric W. Biederman" Cc: Colin Walters Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 24 +++++++++--------------- 1 file changed, 9 insertions(+), 15 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index 68d508f2bfba..84703db06cf3 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1173,13 +1173,16 @@ static struct task_struct *copy_process(unsigned long clone_flags, return ERR_PTR(-EINVAL); /* - * If the new process will be in a different pid namespace don't - * allow it to share a thread group or signal handlers with the - * forking task. + * If the new process will be in a different pid or user namespace + * do not allow it to share a thread group or signal handlers or + * parent with the forking task. */ - if ((clone_flags & CLONE_SIGHAND) && (task_active_pid_ns(current) != - current->nsproxy->pid_ns_for_children)) - return ERR_PTR(-EINVAL); + if (clone_flags & (CLONE_SIGHAND | CLONE_PARENT)) { + if ((clone_flags & (CLONE_NEWUSER | CLONE_NEWPID)) || + (task_active_pid_ns(current) != + current->nsproxy->pid_ns_for_children)) + return ERR_PTR(-EINVAL); + } retval = security_task_create(clone_flags); if (retval) @@ -1575,15 +1578,6 @@ long do_fork(unsigned long clone_flags, int trace = 0; long nr; - /* - * Do some preliminary argument and permissions checking before we - * actually start allocating stuff - */ - if (clone_flags & (CLONE_NEWUSER | CLONE_NEWPID)) { - if (clone_flags & (CLONE_THREAD|CLONE_PARENT)) - return -EINVAL; - } - /* * Determine whether and which event to report to ptracer. When * called from kernel_thread or CLONE_UNTRACED is explicitly -- cgit v1.2.3 From ef0855d334e1e4af7c3e0c42146a8479ea14a5ab Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Wed, 11 Sep 2013 14:20:14 -0700 Subject: mm: mempolicy: turn vma_set_policy() into vma_dup_policy() Simple cleanup. Every user of vma_set_policy() does the same work, this looks a bit annoying imho. And the new trivial helper which does mpol_dup() + vma_set_policy() to simplify the callers. Signed-off-by: Oleg Nesterov Cc: KOSAKI Motohiro Cc: Mel Gorman Cc: Rik van Riel Cc: Andi Kleen Cc: David Rientjes Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/fork.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index 84703db06cf3..81ccb4f010c2 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -351,7 +351,6 @@ static int dup_mmap(struct mm_struct *mm, struct mm_struct *oldmm) struct rb_node **rb_link, *rb_parent; int retval; unsigned long charge; - struct mempolicy *pol; uprobe_start_dup_mmap(); down_write(&oldmm->mmap_sem); @@ -400,11 +399,9 @@ static int dup_mmap(struct mm_struct *mm, struct mm_struct *oldmm) goto fail_nomem; *tmp = *mpnt; INIT_LIST_HEAD(&tmp->anon_vma_chain); - pol = mpol_dup(vma_policy(mpnt)); - retval = PTR_ERR(pol); - if (IS_ERR(pol)) + retval = vma_dup_policy(mpnt, tmp); + if (retval) goto fail_nomem_policy; - vma_set_policy(tmp, pol); tmp->vm_mm = mm; if (anon_vma_fork(tmp, mpnt)) goto fail_nomem_anon_vma_fork; @@ -472,7 +469,7 @@ out: uprobe_end_dup_mmap(); return retval; fail_nomem_anon_vma_fork: - mpol_put(pol); + mpol_put(vma_policy(tmp)); fail_nomem_policy: kmem_cache_free(vm_area_cachep, tmp); fail_nomem: -- cgit v1.2.3 From c33bc315fd921b1179a1d3df5756e0da6fb73944 Mon Sep 17 00:00:00 2001 From: Xishi Qiu Date: Wed, 11 Sep 2013 14:21:44 -0700 Subject: mm: use zone_end_pfn() instead of zone_start_pfn+spanned_pages Use "zone_end_pfn()" instead of "zone->zone_start_pfn + zone->spanned_pages". Simplify the code, no functional change. [akpm@linux-foundation.org: fix build] Signed-off-by: Xishi Qiu Cc: Cody P Schafer Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/power/snapshot.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c index 349587bb03e1..358a146fd4da 100644 --- a/kernel/power/snapshot.c +++ b/kernel/power/snapshot.c @@ -352,7 +352,7 @@ static int create_mem_extents(struct list_head *list, gfp_t gfp_mask) struct mem_extent *ext, *cur, *aux; zone_start = zone->zone_start_pfn; - zone_end = zone->zone_start_pfn + zone->spanned_pages; + zone_end = zone_end_pfn(zone); list_for_each_entry(ext, list, hook) if (zone_start <= ext->end) @@ -884,7 +884,7 @@ static unsigned int count_highmem_pages(void) continue; mark_free_pages(zone); - max_zone_pfn = zone->zone_start_pfn + zone->spanned_pages; + max_zone_pfn = zone_end_pfn(zone); for (pfn = zone->zone_start_pfn; pfn < max_zone_pfn; pfn++) if (saveable_highmem_page(zone, pfn)) n++; @@ -948,7 +948,7 @@ static unsigned int count_data_pages(void) continue; mark_free_pages(zone); - max_zone_pfn = zone->zone_start_pfn + zone->spanned_pages; + max_zone_pfn = zone_end_pfn(zone); for (pfn = zone->zone_start_pfn; pfn < max_zone_pfn; pfn++) if (saveable_page(zone, pfn)) n++; @@ -1041,7 +1041,7 @@ copy_data_pages(struct memory_bitmap *copy_bm, struct memory_bitmap *orig_bm) unsigned long max_zone_pfn; mark_free_pages(zone); - max_zone_pfn = zone->zone_start_pfn + zone->spanned_pages; + max_zone_pfn = zone_end_pfn(zone); for (pfn = zone->zone_start_pfn; pfn < max_zone_pfn; pfn++) if (page_is_saveable(zone, pfn)) memory_bm_set_bit(orig_bm, pfn); @@ -1093,7 +1093,7 @@ void swsusp_free(void) unsigned long pfn, max_zone_pfn; for_each_populated_zone(zone) { - max_zone_pfn = zone->zone_start_pfn + zone->spanned_pages; + max_zone_pfn = zone_end_pfn(zone); for (pfn = zone->zone_start_pfn; pfn < max_zone_pfn; pfn++) if (pfn_valid(pfn)) { struct page *page = pfn_to_page(pfn); @@ -1755,7 +1755,7 @@ static int mark_unsafe_pages(struct memory_bitmap *bm) /* Clear page flags */ for_each_populated_zone(zone) { - max_zone_pfn = zone->zone_start_pfn + zone->spanned_pages; + max_zone_pfn = zone_end_pfn(zone); for (pfn = zone->zone_start_pfn; pfn < max_zone_pfn; pfn++) if (pfn_valid(pfn)) swsusp_unset_page_free(pfn_to_page(pfn)); -- cgit v1.2.3 From 86cdb465cf3a9d81058b517af05074157fa9dcdd Mon Sep 17 00:00:00 2001 From: Naoya Horiguchi Date: Wed, 11 Sep 2013 14:22:13 -0700 Subject: mm: prepare to remove /proc/sys/vm/hugepages_treat_as_movable Now hugepage migration is enabled, although restricted on pmd-based hugepages for now (due to lack of testing.) So we should allocate migratable hugepages from ZONE_MOVABLE if possible. This patch makes GFP flags in hugepage allocation dependent on migration support, not only the value of hugepages_treat_as_movable. It provides no change on the behavior for architectures which do not support hugepage migration, Signed-off-by: Naoya Horiguchi Acked-by: Andi Kleen Reviewed-by: Wanpeng Li Cc: Hillf Danton Cc: Mel Gorman Cc: Hugh Dickins Cc: KOSAKI Motohiro Cc: Michal Hocko Cc: Rik van Riel Cc: "Aneesh Kumar K.V" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sysctl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 07f6fc468e17..dc69093a8ec4 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -1225,7 +1225,7 @@ static struct ctl_table vm_table[] = { .data = &hugepages_treat_as_movable, .maxlen = sizeof(int), .mode = 0644, - .proc_handler = hugetlb_treat_movable_handler, + .proc_handler = proc_dointvec, }, { .procname = "nr_overcommit_hugepages", -- cgit v1.2.3 From 3ddc5b46a8e90f3c9251338b60191d0a804b0d92 Mon Sep 17 00:00:00 2001 From: Mathieu Desnoyers Date: Wed, 11 Sep 2013 14:23:18 -0700 Subject: kernel-wide: fix missing validations on __get/__put/__copy_to/__copy_from_user() I found the following pattern that leads in to interesting findings: grep -r "ret.*|=.*__put_user" * grep -r "ret.*|=.*__get_user" * grep -r "ret.*|=.*__copy" * The __put_user() calls in compat_ioctl.c, ptrace compat, signal compat, since those appear in compat code, we could probably expect the kernel addresses not to be reachable in the lower 32-bit range, so I think they might not be exploitable. For the "__get_user" cases, I don't think those are exploitable: the worse that can happen is that the kernel will copy kernel memory into in-kernel buffers, and will fail immediately afterward. The alpha csum_partial_copy_from_user() seems to be missing the access_ok() check entirely. The fix is inspired from x86. This could lead to information leak on alpha. I also noticed that many architectures map csum_partial_copy_from_user() to csum_partial_copy_generic(), but I wonder if the latter is performing the access checks on every architectures. Signed-off-by: Mathieu Desnoyers Cc: Richard Henderson Cc: Ivan Kokshaysky Cc: Matt Turner Cc: Jens Axboe Cc: Oleg Nesterov Cc: David Miller Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/signal.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index 50e41075ac77..ded28b91fa53 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -3394,7 +3394,7 @@ COMPAT_SYSCALL_DEFINE4(rt_sigaction, int, sig, new_ka.sa.sa_restorer = compat_ptr(restorer); #endif ret |= copy_from_user(&mask, &act->sa_mask, sizeof(mask)); - ret |= __get_user(new_ka.sa.sa_flags, &act->sa_flags); + ret |= get_user(new_ka.sa.sa_flags, &act->sa_flags); if (ret) return -EFAULT; sigset_from_compat(&new_ka.sa.sa_mask, &mask); @@ -3406,7 +3406,7 @@ COMPAT_SYSCALL_DEFINE4(rt_sigaction, int, sig, ret = put_user(ptr_to_compat(old_ka.sa.sa_handler), &oact->sa_handler); ret |= copy_to_user(&oact->sa_mask, &mask, sizeof(mask)); - ret |= __put_user(old_ka.sa.sa_flags, &oact->sa_flags); + ret |= put_user(old_ka.sa.sa_flags, &oact->sa_flags); #ifdef __ARCH_HAS_SA_RESTORER ret |= put_user(ptr_to_compat(old_ka.sa.sa_restorer), &oact->sa_restorer); -- cgit v1.2.3 From 54a33b1b1470ada14fa2998e8b48ad4a0ef6a916 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Wed, 11 Sep 2013 14:23:19 -0700 Subject: kernel/modsign_pubkey.c: fix init const for module signing code const has to use __initconst, not __initdata Signed-off-by: Andi Kleen Acked-by: David Howells Cc: Rusty Russell Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/modsign_pubkey.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/modsign_pubkey.c b/kernel/modsign_pubkey.c index 2b6e69909c39..7cbd4507a7e6 100644 --- a/kernel/modsign_pubkey.c +++ b/kernel/modsign_pubkey.c @@ -18,14 +18,14 @@ struct key *modsign_keyring; -extern __initdata const u8 modsign_certificate_list[]; -extern __initdata const u8 modsign_certificate_list_end[]; +extern __initconst const u8 modsign_certificate_list[]; +extern __initconst const u8 modsign_certificate_list_end[]; /* * We need to make sure ccache doesn't cache the .o file as it doesn't notice * if modsign.pub changes. */ -static __initdata const char annoy_ccache[] = __TIME__ "foo"; +static __initconst const char annoy_ccache[] = __TIME__ "foo"; /* * Load the compiled-in keys -- cgit v1.2.3 From 60c323699bb308404dcb60e8808531e02651578a Mon Sep 17 00:00:00 2001 From: Chen Gang Date: Wed, 11 Sep 2013 14:23:22 -0700 Subject: kernel/smp.c: free related resources when failure occurs in hotplug_cfd() When failure occurs in hotplug_cfd(), need release related resources, or will cause memory leak. Signed-off-by: Chen Gang Acked-by: Wang YanQing Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/smp.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/smp.c b/kernel/smp.c index 449b707fc20d..3bb6ae533cdf 100644 --- a/kernel/smp.c +++ b/kernel/smp.c @@ -48,10 +48,13 @@ hotplug_cfd(struct notifier_block *nfb, unsigned long action, void *hcpu) cpu_to_node(cpu))) return notifier_from_errno(-ENOMEM); if (!zalloc_cpumask_var_node(&cfd->cpumask_ipi, GFP_KERNEL, - cpu_to_node(cpu))) + cpu_to_node(cpu))) { + free_cpumask_var(cfd->cpumask); return notifier_from_errno(-ENOMEM); + } cfd->csd = alloc_percpu(struct call_single_data); if (!cfd->csd) { + free_cpumask_var(cfd->cpumask_ipi); free_cpumask_var(cfd->cpumask); return notifier_from_errno(-ENOMEM); } -- cgit v1.2.3 From c14c338cb05c700a260480c197cfd6da8f8b7d2e Mon Sep 17 00:00:00 2001 From: Will Deacon Date: Wed, 11 Sep 2013 14:23:23 -0700 Subject: kernel/spinlock.c: add default arch_*_relax definitions for GENERIC_LOCKBREAK When running with GENERIC_LOCKBREAK=y, the locking implementations emit calls to arch_{read,write,spin}_relax when spinning on a contended lock in order to allow architectures to favour the CPU owning the lock if possible. In reality, everybody apart from PowerPC and S390 just does cpu_relax() here, so make that the default behaviour and allow it to be overridden if required. Signed-off-by: Will Deacon Cc: Benjamin Herrenschmidt Cc: Martin Schwidefsky Cc: Thomas Gleixner Cc: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/spinlock.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+) (limited to 'kernel') diff --git a/kernel/spinlock.c b/kernel/spinlock.c index 5cdd8065a3ce..4b082b5cac9e 100644 --- a/kernel/spinlock.c +++ b/kernel/spinlock.c @@ -34,6 +34,20 @@ #else #define raw_read_can_lock(l) read_can_lock(l) #define raw_write_can_lock(l) write_can_lock(l) + +/* + * Some architectures can relax in favour of the CPU owning the lock. + */ +#ifndef arch_read_relax +# define arch_read_relax(l) cpu_relax() +#endif +#ifndef arch_write_relax +# define arch_write_relax(l) cpu_relax() +#endif +#ifndef arch_spin_relax +# define arch_spin_relax(l) cpu_relax() +#endif + /* * We build the __lock_function inlines here. They are too large for * inlining all over the place, but here is only one user per function -- cgit v1.2.3 From fa688207c9db48b64ab6538abc3fcdf26110b9ec Mon Sep 17 00:00:00 2001 From: David Daney Date: Wed, 11 Sep 2013 14:23:24 -0700 Subject: smp: quit unconditionally enabling irq in on_each_cpu_mask and on_each_cpu_cond As in commit f21afc25f9ed ("smp.h: Use local_irq_{save,restore}() in !SMP version of on_each_cpu()"), we don't want to enable irqs if they are not already enabled. There are currently no known problematical callers of these functions, but since it is a known failure pattern, we preemptively fix them. Since they are not trivial functions, make them non-inline by moving them to up.c. This also makes it so we don't have to fix #include dependancies for preempt_{disable,enable}. Signed-off-by: David Daney Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/up.c | 39 +++++++++++++++++++++++++++++++++++++++ 1 file changed, 39 insertions(+) (limited to 'kernel') diff --git a/kernel/up.c b/kernel/up.c index c54c75e9faf7..144e57255234 100644 --- a/kernel/up.c +++ b/kernel/up.c @@ -19,3 +19,42 @@ int smp_call_function_single(int cpu, void (*func) (void *info), void *info, return 0; } EXPORT_SYMBOL(smp_call_function_single); + +/* + * Note we still need to test the mask even for UP + * because we actually can get an empty mask from + * code that on SMP might call us without the local + * CPU in the mask. + */ +void on_each_cpu_mask(const struct cpumask *mask, + smp_call_func_t func, void *info, bool wait) +{ + unsigned long flags; + + if (cpumask_test_cpu(0, mask)) { + local_irq_save(flags); + func(info); + local_irq_restore(flags); + } +} +EXPORT_SYMBOL(on_each_cpu_mask); + +/* + * Preemption is disabled here to make sure the cond_func is called under the + * same condtions in UP and SMP. + */ +void on_each_cpu_cond(bool (*cond_func)(int cpu, void *info), + smp_call_func_t func, void *info, bool wait, + gfp_t gfp_flags) +{ + unsigned long flags; + + preempt_disable(); + if (cond_func(0, info)) { + local_irq_save(flags); + func(info); + local_irq_restore(flags); + } + preempt_enable(); +} +EXPORT_SYMBOL(on_each_cpu_cond); -- cgit v1.2.3 From 081192b25c2d4620b5f5838620624d3daee94b66 Mon Sep 17 00:00:00 2001 From: David Daney Date: Wed, 11 Sep 2013 14:23:25 -0700 Subject: up.c: use local_irq_{save,restore}() in smp_call_function_single. The SMP version of this function doesn't unconditionally enable irqs, so neither should this !SMP version. There are no know problems caused by this, but we make the change for consistency's sake. Signed-off-by: David Daney Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/up.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/up.c b/kernel/up.c index 144e57255234..b1cf036255f3 100644 --- a/kernel/up.c +++ b/kernel/up.c @@ -10,11 +10,13 @@ int smp_call_function_single(int cpu, void (*func) (void *info), void *info, int wait) { + unsigned long flags; + WARN_ON(cpu != 0); - local_irq_disable(); - (func)(info); - local_irq_enable(); + local_irq_save(flags); + func(info); + local_irq_restore(flags); return 0; } -- cgit v1.2.3 From bff2dc42bcafdd75c0296987747f782965d691a0 Mon Sep 17 00:00:00 2001 From: David Daney Date: Wed, 11 Sep 2013 14:23:26 -0700 Subject: smp.h: move !SMP version of on_each_cpu() out-of-line All of the other non-trivial !SMP versions of functions in smp.h are out-of-line in up.c. Move on_each_cpu() there as well. This allows us to get rid of the #include . The drawback is that this makes both the x86_64 and i386 defconfig !SMP kernels about 200 bytes larger each. Signed-off-by: David Daney Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/up.c | 11 +++++++++++ 1 file changed, 11 insertions(+) (limited to 'kernel') diff --git a/kernel/up.c b/kernel/up.c index b1cf036255f3..630d72bf7e41 100644 --- a/kernel/up.c +++ b/kernel/up.c @@ -22,6 +22,17 @@ int smp_call_function_single(int cpu, void (*func) (void *info), void *info, } EXPORT_SYMBOL(smp_call_function_single); +int on_each_cpu(smp_call_func_t func, void *info, int wait) +{ + unsigned long flags; + + local_irq_save(flags); + func(info); + local_irq_restore(flags); + return 0; +} +EXPORT_SYMBOL(on_each_cpu); + /* * Note we still need to test the mask even for UP * because we actually can get an empty mask from -- cgit v1.2.3 From e656a634118285142063527b2cd40c749036de82 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Uwe=20Kleine-K=C3=B6nig?= Date: Wed, 11 Sep 2013 14:23:27 -0700 Subject: extable: skip sorting if the table is empty MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit At least on ARM no-MMU the extable is empty and so there is nothing to sort. So add a check for the table to be empty which effectively only changes that the misleading pr_notice is suppressed. Signed-off-by: Uwe Kleine-König Cc: Ingo Molnar Cc: David Daney Cc: "H. Peter Anvin" Cc: Borislav Petkov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/extable.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/extable.c b/kernel/extable.c index 67460b93b1a1..832cb28105bb 100644 --- a/kernel/extable.c +++ b/kernel/extable.c @@ -41,7 +41,7 @@ u32 __initdata main_extable_sort_needed = 1; /* Sort the kernel's built-in exception table */ void __init sort_main_extable(void) { - if (main_extable_sort_needed) { + if (main_extable_sort_needed && __stop___ex_table > __start___ex_table) { pr_notice("Sorting __ex_table...\n"); sort_extable(__start___ex_table, __stop___ex_table); } -- cgit v1.2.3 From 202da400570d991bacda4a06e878cb901e96a783 Mon Sep 17 00:00:00 2001 From: David Daney Date: Wed, 11 Sep 2013 14:23:29 -0700 Subject: kernel/smp.c: quit unconditionally enabling irqs in on_each_cpu_mask(). As in commit f21afc25f9ed ("smp.h: Use local_irq_{save,restore}() in !SMP version of on_each_cpu()"), we don't want to enable irqs if they are not already enabled. I don't know of any bugs currently caused by this unconditional local_irq_enable(), but I want to use this function in MIPS/OCTEON early boot (when we have early_boot_irqs_disabled). This also makes this function have similar semantics to on_each_cpu() which is good in itself. Signed-off-by: David Daney Cc: Gilad Ben-Yossef Cc: Christoph Lameter Cc: Chris Metcalf Cc: Peter Zijlstra Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/smp.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/smp.c b/kernel/smp.c index 3bb6ae533cdf..0564571dcdf7 100644 --- a/kernel/smp.c +++ b/kernel/smp.c @@ -575,8 +575,10 @@ EXPORT_SYMBOL(on_each_cpu); * * If @wait is true, then returns once @func has returned. * - * You must not call this function with disabled interrupts or - * from a hardware interrupt handler or from a bottom half handler. + * You must not call this function with disabled interrupts or from a + * hardware interrupt handler or from a bottom half handler. The + * exception is that it may be used during early boot while + * early_boot_irqs_disabled is set. */ void on_each_cpu_mask(const struct cpumask *mask, smp_call_func_t func, void *info, bool wait) @@ -585,9 +587,10 @@ void on_each_cpu_mask(const struct cpumask *mask, smp_call_func_t func, smp_call_function_many(mask, func, info, wait); if (cpumask_test_cpu(cpu, mask)) { - local_irq_disable(); + unsigned long flags; + local_irq_save(flags); func(info); - local_irq_enable(); + local_irq_restore(flags); } put_cpu(); } -- cgit v1.2.3 From 205e550a0fb469ae73f91a903f27f4f63e774037 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Wed, 11 Sep 2013 14:23:30 -0700 Subject: task_work: minor cleanups Trivial. Remove the unnecessary "work = NULL" initialization and turn read_barrier_depends() into smp_read_barrier_depends() in task_work_cancel(). Signed-off-by: Oleg Nesterov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/task_work.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/task_work.c b/kernel/task_work.c index 65bd3c92d6f3..6ee09856f725 100644 --- a/kernel/task_work.c +++ b/kernel/task_work.c @@ -25,7 +25,7 @@ struct callback_head * task_work_cancel(struct task_struct *task, task_work_func_t func) { struct callback_head **pprev = &task->task_works; - struct callback_head *work = NULL; + struct callback_head *work; unsigned long flags; /* * If cmpxchg() fails we continue without updating pprev. @@ -35,7 +35,7 @@ task_work_cancel(struct task_struct *task, task_work_func_t func) */ raw_spin_lock_irqsave(&task->pi_lock, flags); while ((work = ACCESS_ONCE(*pprev))) { - read_barrier_depends(); + smp_read_barrier_depends(); if (work->func != func) pprev = &work->next; else if (cmpxchg(pprev, work, work->next) == work) -- cgit v1.2.3 From 892f6668f3a7088c7e30049c3d8e1844531602dc Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Wed, 11 Sep 2013 14:23:31 -0700 Subject: task_work: documentation No functional changes, just comments. Signed-off-by: Oleg Nesterov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/task_work.c | 36 ++++++++++++++++++++++++++++++++++++ 1 file changed, 36 insertions(+) (limited to 'kernel') diff --git a/kernel/task_work.c b/kernel/task_work.c index 6ee09856f725..8727032e3a6f 100644 --- a/kernel/task_work.c +++ b/kernel/task_work.c @@ -4,6 +4,23 @@ static struct callback_head work_exited; /* all we need is ->next == NULL */ +/** + * task_work_add - ask the @task to execute @work->func() + * @task: the task which should run the callback + * @work: the callback to run + * @notify: send the notification if true + * + * Queue @work for task_work_run() below and notify the @task if @notify. + * Fails if the @task is exiting/exited and thus it can't process this @work. + * Otherwise @work->func() will be called when the @task returns from kernel + * mode or exits. + * + * This is like the signal handler which runs in kernel mode, but it doesn't + * try to wake up the @task. + * + * RETURNS: + * 0 if succeeds or -ESRCH. + */ int task_work_add(struct task_struct *task, struct callback_head *work, bool notify) { @@ -21,6 +38,17 @@ task_work_add(struct task_struct *task, struct callback_head *work, bool notify) return 0; } +/** + * task_work_cancel - cancel a pending work added by task_work_add() + * @task: the task which should execute the work + * @func: identifies the work to remove + * + * Find the last queued pending work with ->func == @func and remove + * it from queue. + * + * RETURNS: + * The found work or NULL if not found. + */ struct callback_head * task_work_cancel(struct task_struct *task, task_work_func_t func) { @@ -46,6 +74,14 @@ task_work_cancel(struct task_struct *task, task_work_func_t func) return work; } +/** + * task_work_run - execute the works added by task_work_add() + * + * Flush the pending works. Should be used by the core kernel code. + * Called before the task returns to the user-mode or stops, or when + * it exits. In the latter case task_work_add() can no longer add the + * new work after task_work_run() returns. + */ void task_work_run(void) { struct task_struct *task = current; -- cgit v1.2.3 From c802d64a356b5cf349121ac4c5e005f037ce548d Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Wed, 11 Sep 2013 14:24:11 -0700 Subject: kprobes: unify insn caches The current kpropes insn caches allocate memory areas for insn slots with module_alloc(). The assumption is that the kernel image and module area are both within the same +/- 2GB memory area. This however is not true for s390 where the kernel image resides within the first 2GB (DMA memory area), but the module area is far away in the vmalloc area, usually somewhere close below the 4TB area. For new pc relative instructions s390 needs insn slots that are within +/- 2GB of each area. That way we can patch displacements of pc-relative instructions within the insn slots just like x86 and powerpc. The module area works already with the normal insn slot allocator, however there is currently no way to get insn slots that are within the first 2GB on s390 (aka DMA area). Therefore this patch set modifies the kprobes insn slot cache code in order to allow to specify a custom allocator for the insn slot cache pages. In addition architecure can now have private insn slot caches withhout the need to modify common code. Patch 1 unifies and simplifies the current insn and optinsn caches implementation. This is a preparation which allows to add more insn caches in a simple way. Patch 2 adds the possibility to specify a custom allocator. Patch 3 makes s390 use the new insn slot mechanisms and adds support for pc-relative instructions with long displacements. This patch (of 3): The two insn caches (insn, and optinsn) each have an own mutex and alloc/free functions (get_[opt]insn_slot() / free_[opt]insn_slot()). Since there is the need for yet another insn cache which satifies dma allocations on s390, unify and simplify the current implementation: - Move the per insn cache mutex into struct kprobe_insn_cache. - Move the alloc/free functions to kprobe.h so they are simply wrappers for the generic __get_insn_slot/__free_insn_slot functions. The implementation is done with a DEFINE_INSN_CACHE_OPS() macro which provides the alloc/free functions for each cache if needed. - move the struct kprobe_insn_cache to kprobe.h which allows to generate architecture specific insn slot caches outside of the core kprobes code. Signed-off-by: Heiko Carstens Cc: Masami Hiramatsu Cc: Ananth N Mavinakayanahalli Cc: Ingo Molnar Cc: Martin Schwidefsky Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/kprobes.c | 75 ++++++++++++++++---------------------------------------- 1 file changed, 21 insertions(+), 54 deletions(-) (limited to 'kernel') diff --git a/kernel/kprobes.c b/kernel/kprobes.c index 6e33498d665c..9e4912dc5559 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -121,12 +121,6 @@ struct kprobe_insn_page { (offsetof(struct kprobe_insn_page, slot_used) + \ (sizeof(char) * (slots))) -struct kprobe_insn_cache { - struct list_head pages; /* list of kprobe_insn_page */ - size_t insn_size; /* size of instruction slot */ - int nr_garbage; -}; - static int slots_per_page(struct kprobe_insn_cache *c) { return PAGE_SIZE/(c->insn_size * sizeof(kprobe_opcode_t)); @@ -138,8 +132,8 @@ enum kprobe_slot_state { SLOT_USED = 2, }; -static DEFINE_MUTEX(kprobe_insn_mutex); /* Protects kprobe_insn_slots */ -static struct kprobe_insn_cache kprobe_insn_slots = { +struct kprobe_insn_cache kprobe_insn_slots = { + .mutex = __MUTEX_INITIALIZER(kprobe_insn_slots.mutex), .pages = LIST_HEAD_INIT(kprobe_insn_slots.pages), .insn_size = MAX_INSN_SIZE, .nr_garbage = 0, @@ -150,10 +144,12 @@ static int __kprobes collect_garbage_slots(struct kprobe_insn_cache *c); * __get_insn_slot() - Find a slot on an executable page for an instruction. * We allocate an executable page if there's no room on existing ones. */ -static kprobe_opcode_t __kprobes *__get_insn_slot(struct kprobe_insn_cache *c) +kprobe_opcode_t __kprobes *__get_insn_slot(struct kprobe_insn_cache *c) { struct kprobe_insn_page *kip; + kprobe_opcode_t *slot = NULL; + mutex_lock(&c->mutex); retry: list_for_each_entry(kip, &c->pages, list) { if (kip->nused < slots_per_page(c)) { @@ -162,7 +158,8 @@ static kprobe_opcode_t __kprobes *__get_insn_slot(struct kprobe_insn_cache *c) if (kip->slot_used[i] == SLOT_CLEAN) { kip->slot_used[i] = SLOT_USED; kip->nused++; - return kip->insns + (i * c->insn_size); + slot = kip->insns + (i * c->insn_size); + goto out; } } /* kip->nused is broken. Fix it. */ @@ -178,7 +175,7 @@ static kprobe_opcode_t __kprobes *__get_insn_slot(struct kprobe_insn_cache *c) /* All out of space. Need to allocate a new page. */ kip = kmalloc(KPROBE_INSN_PAGE_SIZE(slots_per_page(c)), GFP_KERNEL); if (!kip) - return NULL; + goto out; /* * Use module_alloc so this page is within +/- 2GB of where the @@ -188,7 +185,7 @@ static kprobe_opcode_t __kprobes *__get_insn_slot(struct kprobe_insn_cache *c) kip->insns = module_alloc(PAGE_SIZE); if (!kip->insns) { kfree(kip); - return NULL; + goto out; } INIT_LIST_HEAD(&kip->list); memset(kip->slot_used, SLOT_CLEAN, slots_per_page(c)); @@ -196,19 +193,10 @@ static kprobe_opcode_t __kprobes *__get_insn_slot(struct kprobe_insn_cache *c) kip->nused = 1; kip->ngarbage = 0; list_add(&kip->list, &c->pages); - return kip->insns; -} - - -kprobe_opcode_t __kprobes *get_insn_slot(void) -{ - kprobe_opcode_t *ret = NULL; - - mutex_lock(&kprobe_insn_mutex); - ret = __get_insn_slot(&kprobe_insn_slots); - mutex_unlock(&kprobe_insn_mutex); - - return ret; + slot = kip->insns; +out: + mutex_unlock(&c->mutex); + return slot; } /* Return 1 if all garbages are collected, otherwise 0. */ @@ -255,11 +243,12 @@ static int __kprobes collect_garbage_slots(struct kprobe_insn_cache *c) return 0; } -static void __kprobes __free_insn_slot(struct kprobe_insn_cache *c, - kprobe_opcode_t *slot, int dirty) +void __kprobes __free_insn_slot(struct kprobe_insn_cache *c, + kprobe_opcode_t *slot, int dirty) { struct kprobe_insn_page *kip; + mutex_lock(&c->mutex); list_for_each_entry(kip, &c->pages, list) { long idx = ((long)slot - (long)kip->insns) / (c->insn_size * sizeof(kprobe_opcode_t)); @@ -272,45 +261,23 @@ static void __kprobes __free_insn_slot(struct kprobe_insn_cache *c, collect_garbage_slots(c); } else collect_one_slot(kip, idx); - return; + goto out; } } /* Could not free this slot. */ WARN_ON(1); +out: + mutex_unlock(&c->mutex); } -void __kprobes free_insn_slot(kprobe_opcode_t * slot, int dirty) -{ - mutex_lock(&kprobe_insn_mutex); - __free_insn_slot(&kprobe_insn_slots, slot, dirty); - mutex_unlock(&kprobe_insn_mutex); -} #ifdef CONFIG_OPTPROBES /* For optimized_kprobe buffer */ -static DEFINE_MUTEX(kprobe_optinsn_mutex); /* Protects kprobe_optinsn_slots */ -static struct kprobe_insn_cache kprobe_optinsn_slots = { +struct kprobe_insn_cache kprobe_optinsn_slots = { + .mutex = __MUTEX_INITIALIZER(kprobe_optinsn_slots.mutex), .pages = LIST_HEAD_INIT(kprobe_optinsn_slots.pages), /* .insn_size is initialized later */ .nr_garbage = 0, }; -/* Get a slot for optimized_kprobe buffer */ -kprobe_opcode_t __kprobes *get_optinsn_slot(void) -{ - kprobe_opcode_t *ret = NULL; - - mutex_lock(&kprobe_optinsn_mutex); - ret = __get_insn_slot(&kprobe_optinsn_slots); - mutex_unlock(&kprobe_optinsn_mutex); - - return ret; -} - -void __kprobes free_optinsn_slot(kprobe_opcode_t * slot, int dirty) -{ - mutex_lock(&kprobe_optinsn_mutex); - __free_insn_slot(&kprobe_optinsn_slots, slot, dirty); - mutex_unlock(&kprobe_optinsn_mutex); -} #endif #endif -- cgit v1.2.3 From af96397de8600232effbff43dc8b4ca20ddc02b1 Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Wed, 11 Sep 2013 14:24:13 -0700 Subject: kprobes: allow to specify custom allocator for insn caches The current two insn slot caches both use module_alloc/module_free to allocate and free insn slot cache pages. For s390 this is not sufficient since there is the need to allocate insn slots that are either within the vmalloc module area or within dma memory. Therefore add a mechanism which allows to specify an own allocator for an own insn slot cache. Signed-off-by: Heiko Carstens Acked-by: Masami Hiramatsu Cc: Ananth N Mavinakayanahalli Cc: Ingo Molnar Cc: Martin Schwidefsky Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/kprobes.c | 20 ++++++++++++++++++-- 1 file changed, 18 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/kprobes.c b/kernel/kprobes.c index 9e4912dc5559..a0d367a49122 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -112,6 +112,7 @@ static struct kprobe_blackpoint kprobe_blacklist[] = { struct kprobe_insn_page { struct list_head list; kprobe_opcode_t *insns; /* Page of instruction slots */ + struct kprobe_insn_cache *cache; int nused; int ngarbage; char slot_used[]; @@ -132,8 +133,20 @@ enum kprobe_slot_state { SLOT_USED = 2, }; +static void *alloc_insn_page(void) +{ + return module_alloc(PAGE_SIZE); +} + +static void free_insn_page(void *page) +{ + module_free(NULL, page); +} + struct kprobe_insn_cache kprobe_insn_slots = { .mutex = __MUTEX_INITIALIZER(kprobe_insn_slots.mutex), + .alloc = alloc_insn_page, + .free = free_insn_page, .pages = LIST_HEAD_INIT(kprobe_insn_slots.pages), .insn_size = MAX_INSN_SIZE, .nr_garbage = 0, @@ -182,7 +195,7 @@ kprobe_opcode_t __kprobes *__get_insn_slot(struct kprobe_insn_cache *c) * kernel image and loaded module images reside. This is required * so x86_64 can correctly handle the %rip-relative fixups. */ - kip->insns = module_alloc(PAGE_SIZE); + kip->insns = c->alloc(); if (!kip->insns) { kfree(kip); goto out; @@ -192,6 +205,7 @@ kprobe_opcode_t __kprobes *__get_insn_slot(struct kprobe_insn_cache *c) kip->slot_used[0] = SLOT_USED; kip->nused = 1; kip->ngarbage = 0; + kip->cache = c; list_add(&kip->list, &c->pages); slot = kip->insns; out: @@ -213,7 +227,7 @@ static int __kprobes collect_one_slot(struct kprobe_insn_page *kip, int idx) */ if (!list_is_singular(&kip->list)) { list_del(&kip->list); - module_free(NULL, kip->insns); + kip->cache->free(kip->insns); kfree(kip); } return 1; @@ -274,6 +288,8 @@ out: /* For optimized_kprobe buffer */ struct kprobe_insn_cache kprobe_optinsn_slots = { .mutex = __MUTEX_INITIALIZER(kprobe_optinsn_slots.mutex), + .alloc = alloc_insn_page, + .free = free_insn_page, .pages = LIST_HEAD_INIT(kprobe_optinsn_slots.pages), /* .insn_size is initialized later */ .nr_garbage = 0, -- cgit v1.2.3 From 73af963f9f3036dffed55c3a2898598186db1045 Mon Sep 17 00:00:00 2001 From: Mark Grondona Date: Wed, 11 Sep 2013 14:24:31 -0700 Subject: __ptrace_may_access() should not deny sub-threads __ptrace_may_access() checks get_dumpable/ptrace_has_cap/etc if task != current, this can can lead to surprising results. For example, a sub-thread can't readlink("/proc/self/exe") if the executable is not readable. setup_new_exec()->would_dump() notices that inode_permission(MAY_READ) fails and then it does set_dumpable(suid_dumpable). After that get_dumpable() fails. (It is not clear why proc_pid_readlink() checks get_dumpable(), perhaps we could add PTRACE_MODE_NODUMPABLE) Change __ptrace_may_access() to use same_thread_group() instead of "task == current". Any security check is pointless when the tasks share the same ->mm. Signed-off-by: Mark Grondona Signed-off-by: Ben Woodard Signed-off-by: Oleg Nesterov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/ptrace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/ptrace.c b/kernel/ptrace.c index a146ee327f6a..dd562e9aa2c8 100644 --- a/kernel/ptrace.c +++ b/kernel/ptrace.c @@ -236,7 +236,7 @@ static int __ptrace_may_access(struct task_struct *task, unsigned int mode) */ int dumpable = 0; /* Don't let security modules deny introspection */ - if (task == current) + if (same_thread_group(task, current)) return 0; rcu_read_lock(); tcred = __task_cred(task); -- cgit v1.2.3 From 80c74f6a40284c5c5d49f3b3289172bbce0b30b8 Mon Sep 17 00:00:00 2001 From: Xishi Qiu Date: Wed, 11 Sep 2013 14:24:47 -0700 Subject: kexec: remove unnecessary return Code can not run here forever, so remove the unnecessary return. Signed-off-by: Xishi Qiu Suggested-by: Zhang Yanfei Reviewed-by: Simon Horman Reviewed-by: Zhang Yanfei Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/kexec.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/kexec.c b/kernel/kexec.c index 59f7b55ba745..2a74f307c5ec 100644 --- a/kernel/kexec.c +++ b/kernel/kexec.c @@ -1474,11 +1474,8 @@ static int __init __parse_crashkernel(char *cmdline, if (first_colon && (!first_space || first_colon < first_space)) return parse_crashkernel_mem(ck_cmdline, system_ram, crash_size, crash_base); - else - return parse_crashkernel_simple(ck_cmdline, crash_size, - crash_base); - return 0; + return parse_crashkernel_simple(ck_cmdline, crash_size, crash_base); } /* -- cgit v1.2.3 From 6723734cdff15211bb78aeea76ca847374bd93ae Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Wed, 11 Sep 2013 14:25:49 -0700 Subject: panic: call panic handlers before kmsg_dump Since the panic handlers may produce additional information (via printk) for the kernel log, it should be reported as part of the panic output saved by kmsg_dump(). Without this re-ordering, nothing that adds information to a panic will show up in pstore's view when kmsg_dump runs, and is therefore not visible to crash reporting tools that examine pstore output. Signed-off-by: Kees Cook Cc: Anton Vorontsov Cc: Colin Cross Acked-by: Tony Luck Cc: Stephen Boyd Cc: Vikram Mulukutla Cc: Peter Zijlstra Cc: Rusty Russell Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/panic.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/panic.c b/kernel/panic.c index 801864600514..b6c482ccc5db 100644 --- a/kernel/panic.c +++ b/kernel/panic.c @@ -123,10 +123,14 @@ void panic(const char *fmt, ...) */ smp_send_stop(); - kmsg_dump(KMSG_DUMP_PANIC); - + /* + * Run any panic handlers, including those that might need to + * add information to the kmsg dump output. + */ atomic_notifier_call_chain(&panic_notifier_list, 0, buf); + kmsg_dump(KMSG_DUMP_PANIC); + bust_spinlocks(0); if (!panic_blink) -- cgit v1.2.3 From 7bd36014460f793c19e7d6c94dab67b0afcfcb7f Mon Sep 17 00:00:00 2001 From: John Stultz Date: Wed, 11 Sep 2013 16:50:56 -0700 Subject: timekeeping: Fix HRTICK related deadlock from ntp lock changes Gerlando Falauto reported that when HRTICK is enabled, it is possible to trigger system deadlocks. These were hard to reproduce, as HRTICK has been broken in the past, but seemed to be connected to the timekeeping_seq lock. Since seqlock/seqcount's aren't supported w/ lockdep, I added some extra spinlock based locking and triggered the following lockdep output: [ 15.849182] ntpd/4062 is trying to acquire lock: [ 15.849765] (&(&pool->lock)->rlock){..-...}, at: [] __queue_work+0x145/0x480 [ 15.850051] [ 15.850051] but task is already holding lock: [ 15.850051] (timekeeper_lock){-.-.-.}, at: [] do_adjtimex+0x7f/0x100 [ 15.850051] Chain exists of: &(&pool->lock)->rlock --> &p->pi_lock --> timekeeper_lock [ 15.850051] Possible unsafe locking scenario: [ 15.850051] [ 15.850051] CPU0 CPU1 [ 15.850051] ---- ---- [ 15.850051] lock(timekeeper_lock); [ 15.850051] lock(&p->pi_lock); [ 15.850051] lock(timekeeper_lock); [ 15.850051] lock(&(&pool->lock)->rlock); [ 15.850051] [ 15.850051] *** DEADLOCK *** The deadlock was introduced by 06c017fdd4dc48451a ("timekeeping: Hold timekeepering locks in do_adjtimex and hardpps") in 3.10 This patch avoids this deadlock, by moving the call to schedule_delayed_work() outside of the timekeeper lock critical section. Reported-by: Gerlando Falauto Tested-by: Lin Ming Signed-off-by: John Stultz Cc: Mathieu Desnoyers Cc: stable #3.11, 3.10 Link: http://lkml.kernel.org/r/1378943457-27314-1-git-send-email-john.stultz@linaro.org Signed-off-by: Ingo Molnar --- kernel/time/ntp.c | 6 ++---- kernel/time/timekeeping.c | 2 ++ 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/time/ntp.c b/kernel/time/ntp.c index 8f5b3b98577b..bb2215174f05 100644 --- a/kernel/time/ntp.c +++ b/kernel/time/ntp.c @@ -516,13 +516,13 @@ static void sync_cmos_clock(struct work_struct *work) schedule_delayed_work(&sync_cmos_work, timespec_to_jiffies(&next)); } -static void notify_cmos_timer(void) +void ntp_notify_cmos_timer(void) { schedule_delayed_work(&sync_cmos_work, 0); } #else -static inline void notify_cmos_timer(void) { } +void ntp_notify_cmos_timer(void) { } #endif @@ -687,8 +687,6 @@ int __do_adjtimex(struct timex *txc, struct timespec *ts, s32 *time_tai) if (!(time_status & STA_NANO)) txc->time.tv_usec /= NSEC_PER_USEC; - notify_cmos_timer(); - return result; } diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index 48b9fffabdc2..947ba25a95a0 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -1703,6 +1703,8 @@ int do_adjtimex(struct timex *txc) write_seqcount_end(&timekeeper_seq); raw_spin_unlock_irqrestore(&timekeeper_lock, flags); + ntp_notify_cmos_timer(); + return ret; } -- cgit v1.2.3 From 878b5a6efd38030c7a90895dc8346e8fb1e09b4c Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Wed, 11 Sep 2013 17:47:26 +0200 Subject: uprobes: Fix utask->depth accounting in handle_trampoline() Currently utask->depth is simply the number of allocated/pending return_instance's in uprobe_task->return_instances list. handle_trampoline() should decrement this counter every time we handle/free an instance, but due to typo it does this only if ->chained == T. This means that in the likely case this counter is never decremented and the probed task can't report more than MAX_URETPROBE_DEPTH events. Reported-by: Mikhail Kulemin Reported-by: Hemant Kumar Shaw Signed-off-by: Oleg Nesterov Acked-by: Anton Arapov Cc: masami.hiramatsu.pt@hitachi.com Cc: srikar@linux.vnet.ibm.com Cc: systemtap@sourceware.org Cc: stable@vger.kernel.org Link: http://lkml.kernel.org/r/20130911154726.GA8093@redhat.com Signed-off-by: Ingo Molnar --- kernel/events/uprobes.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c index f3569747d629..ad8e1bdca70e 100644 --- a/kernel/events/uprobes.c +++ b/kernel/events/uprobes.c @@ -1682,12 +1682,10 @@ static bool handle_trampoline(struct pt_regs *regs) tmp = ri; ri = ri->next; kfree(tmp); + utask->depth--; if (!chained) break; - - utask->depth--; - BUG_ON(!ri); } -- cgit v1.2.3 From 6c9a27f5da9609fca46cb2b183724531b48f71ad Mon Sep 17 00:00:00 2001 From: Daisuke Nishimura Date: Tue, 10 Sep 2013 18:16:36 +0900 Subject: sched/fair: Fix small race where child->se.parent,cfs_rq might point to invalid ones There is a small race between copy_process() and cgroup_attach_task() where child->se.parent,cfs_rq points to invalid (old) ones. parent doing fork() | someone moving the parent to another cgroup -------------------------------+--------------------------------------------- copy_process() + dup_task_struct() -> parent->se is copied to child->se. se.parent,cfs_rq of them point to old ones. cgroup_attach_task() + cgroup_task_migrate() -> parent->cgroup is updated. + cpu_cgroup_attach() + sched_move_task() + task_move_group_fair() +- set_task_rq() -> se.parent,cfs_rq of parent are updated. + cgroup_fork() -> parent->cgroup is copied to child->cgroup. (*1) + sched_fork() + task_fork_fair() -> se.parent,cfs_rq of child are accessed while they point to old ones. (*2) In the worst case, this bug can lead to "use-after-free" and cause a panic, because it's new cgroup's refcount that is incremented at (*1), so the old cgroup(and related data) can be freed before (*2). In fact, a panic caused by this bug was originally caught in RHEL6.4. BUG: unable to handle kernel NULL pointer dereference at (null) IP: [] sched_slice+0x6e/0xa0 [...] Call Trace: [] place_entity+0x75/0xa0 [] task_fork_fair+0xaa/0x160 [] sched_fork+0x6b/0x140 [] copy_process+0x5b2/0x1450 [] ? wake_up_new_task+0xd9/0x130 [] do_fork+0x94/0x460 [] ? sys_wait4+0xae/0x100 [] sys_clone+0x28/0x30 [] stub_clone+0x13/0x20 [] ? system_call_fastpath+0x16/0x1b Signed-off-by: Daisuke Nishimura Signed-off-by: Peter Zijlstra Cc: Link: http://lkml.kernel.org/r/039601ceae06$733d3130$59b79390$@mxp.nes.nec.co.jp Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 9b3fe1cd8f40..11cd13667359 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -5928,11 +5928,15 @@ static void task_fork_fair(struct task_struct *p) cfs_rq = task_cfs_rq(current); curr = cfs_rq->curr; - if (unlikely(task_cpu(p) != this_cpu)) { - rcu_read_lock(); - __set_task_cpu(p, this_cpu); - rcu_read_unlock(); - } + /* + * Not only the cpu but also the task_group of the parent might have + * been changed after parent->se.parent,cfs_rq were copied to + * child->se.parent,cfs_rq. So call __set_task_cpu() to make those + * of child point to valid ones. + */ + rcu_read_lock(); + __set_task_cpu(p, this_cpu); + rcu_read_unlock(); update_curr(cfs_rq); -- cgit v1.2.3 From fc840914e9b07ab4685c195e1e54e58de4f84c03 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 9 Sep 2013 13:01:41 +0200 Subject: sched/debug: Take PID namespace into account Emmanuel reported that /proc/sched_debug didn't report the right PIDs when using namespaces, cure this. Reported-by: Emmanuel Deloget Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/20130909110141.GM31370@twins.programming.kicks-ass.net Signed-off-by: Ingo Molnar --- kernel/sched/debug.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c index e076bddd4c66..196559994f7c 100644 --- a/kernel/sched/debug.c +++ b/kernel/sched/debug.c @@ -124,7 +124,7 @@ print_task(struct seq_file *m, struct rq *rq, struct task_struct *p) SEQ_printf(m, " "); SEQ_printf(m, "%15s %5d %9Ld.%06ld %9Ld %5d ", - p->comm, p->pid, + p->comm, task_pid_nr(p), SPLIT_NS(p->se.vruntime), (long long)(p->nvcsw + p->nivcsw), p->prio); @@ -289,7 +289,7 @@ do { \ P(nr_load_updates); P(nr_uninterruptible); PN(next_balance); - P(curr->pid); + SEQ_printf(m, " .%-30s: %ld\n", "curr->pid", (long)(task_pid_nr(rq->curr))); PN(clock); P(cpu_load[0]); P(cpu_load[1]); @@ -492,7 +492,7 @@ void proc_sched_show_task(struct task_struct *p, struct seq_file *m) { unsigned long nr_switches; - SEQ_printf(m, "%s (%d, #threads: %d)\n", p->comm, p->pid, + SEQ_printf(m, "%s (%d, #threads: %d)\n", p->comm, task_pid_nr(p), get_nr_threads(p)); SEQ_printf(m, "---------------------------------------------------------" -- cgit v1.2.3 From 6de5a8bfcae6e3b427d642eff078d8305b324b52 Mon Sep 17 00:00:00 2001 From: Sha Zhengju Date: Thu, 12 Sep 2013 15:13:47 -0700 Subject: memcg: rename RESOURCE_MAX to RES_COUNTER_MAX RESOURCE_MAX is far too general name, change it to RES_COUNTER_MAX. Signed-off-by: Sha Zhengju Signed-off-by: Qiang Huang Acked-by: Michal Hocko Cc: Daisuke Nishimura Cc: Jeff Liu Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/res_counter.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/res_counter.c b/kernel/res_counter.c index ff55247e7049..3f0417f97e76 100644 --- a/kernel/res_counter.c +++ b/kernel/res_counter.c @@ -17,8 +17,8 @@ void res_counter_init(struct res_counter *counter, struct res_counter *parent) { spin_lock_init(&counter->lock); - counter->limit = RESOURCE_MAX; - counter->soft_limit = RESOURCE_MAX; + counter->limit = RES_COUNTER_MAX; + counter->soft_limit = RES_COUNTER_MAX; counter->parent = parent; } @@ -182,12 +182,12 @@ int res_counter_memparse_write_strategy(const char *buf, { char *end; - /* return RESOURCE_MAX(unlimited) if "-1" is specified */ + /* return RES_COUNTER_MAX(unlimited) if "-1" is specified */ if (*buf == '-') { *res = simple_strtoull(buf + 1, &end, 10); if (*res != 1 || *end != '\0') return -EINVAL; - *res = RESOURCE_MAX; + *res = RES_COUNTER_MAX; return 0; } -- cgit v1.2.3 From 3af3351676c3deecfd632f47719fb0d13a061ba8 Mon Sep 17 00:00:00 2001 From: Sha Zhengju Date: Thu, 12 Sep 2013 15:13:48 -0700 Subject: memcg: avoid overflow caused by PAGE_ALIGN Since PAGE_ALIGN is aligning up(the next page boundary), so after PAGE_ALIGN, the value might be overflow, such as write the MAX value to *.limit_in_bytes. $ cat /cgroup/memory/memory.limit_in_bytes 18446744073709551615 # echo 18446744073709551615 > /cgroup/memory/memory.limit_in_bytes bash: echo: write error: Invalid argument Some user programs might depend on such behaviours(like libcg, we read the value in snapshot, then use the value to reset cgroup later), and that will cause confusion. So we need to fix it. Signed-off-by: Sha Zhengju Signed-off-by: Qiang Huang Acked-by: Michal Hocko Cc: Daisuke Nishimura Cc: Jeff Liu Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/res_counter.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/res_counter.c b/kernel/res_counter.c index 3f0417f97e76..085d3ae478fe 100644 --- a/kernel/res_counter.c +++ b/kernel/res_counter.c @@ -195,6 +195,10 @@ int res_counter_memparse_write_strategy(const char *buf, if (*end != '\0') return -EINVAL; - *res = PAGE_ALIGN(*res); + if (PAGE_ALIGN(*res) >= *res) + *res = PAGE_ALIGN(*res); + else + *res = RES_COUNTER_MAX; + return 0; } -- cgit v1.2.3 From 1a36e59d4833de19120dc7482c61ef69e228c73c Mon Sep 17 00:00:00 2001 From: Sha Zhengju Date: Thu, 12 Sep 2013 15:13:49 -0700 Subject: memcg: reduce function dereference This function dereferences res far too often, so optimize it. Signed-off-by: Sha Zhengju Signed-off-by: Qiang Huang Acked-by: Michal Hocko Cc: Daisuke Nishimura Cc: Jeff Liu Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/res_counter.c | 19 +++++++++++-------- 1 file changed, 11 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/res_counter.c b/kernel/res_counter.c index 085d3ae478fe..4aa8a305aede 100644 --- a/kernel/res_counter.c +++ b/kernel/res_counter.c @@ -178,27 +178,30 @@ u64 res_counter_read_u64(struct res_counter *counter, int member) #endif int res_counter_memparse_write_strategy(const char *buf, - unsigned long long *res) + unsigned long long *resp) { char *end; + unsigned long long res; /* return RES_COUNTER_MAX(unlimited) if "-1" is specified */ if (*buf == '-') { - *res = simple_strtoull(buf + 1, &end, 10); - if (*res != 1 || *end != '\0') + res = simple_strtoull(buf + 1, &end, 10); + if (res != 1 || *end != '\0') return -EINVAL; - *res = RES_COUNTER_MAX; + *resp = RES_COUNTER_MAX; return 0; } - *res = memparse(buf, &end); + res = memparse(buf, &end); if (*end != '\0') return -EINVAL; - if (PAGE_ALIGN(*res) >= *res) - *res = PAGE_ALIGN(*res); + if (PAGE_ALIGN(res) >= res) + res = PAGE_ALIGN(res); else - *res = RES_COUNTER_MAX; + res = RES_COUNTER_MAX; + + *resp = res; return 0; } -- cgit v1.2.3 From 6072ddc8520b86adfac6939ca32fb6e6c4de017a Mon Sep 17 00:00:00 2001 From: Jingoo Han Date: Thu, 12 Sep 2013 15:14:07 -0700 Subject: kernel: replace strict_strto*() with kstrto*() The usage of strict_strto*() is not preferred, because strict_strto*() is obsolete. Thus, kstrto*() should be used. Signed-off-by: Jingoo Han Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/gcov/fs.c | 2 +- kernel/ksysfs.c | 2 +- kernel/params.c | 14 +++++++------- 3 files changed, 9 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/gcov/fs.c b/kernel/gcov/fs.c index 9bd0934f6c33..7a7d2ee96d42 100644 --- a/kernel/gcov/fs.c +++ b/kernel/gcov/fs.c @@ -74,7 +74,7 @@ static int __init gcov_persist_setup(char *str) { unsigned long val; - if (strict_strtoul(str, 0, &val)) { + if (kstrtoul(str, 0, &val)) { pr_warning("invalid gcov_persist parameter '%s'\n", str); return 0; } diff --git a/kernel/ksysfs.c b/kernel/ksysfs.c index 6ada93c23a9a..9659d38e008f 100644 --- a/kernel/ksysfs.c +++ b/kernel/ksysfs.c @@ -113,7 +113,7 @@ static ssize_t kexec_crash_size_store(struct kobject *kobj, unsigned long cnt; int ret; - if (strict_strtoul(buf, 0, &cnt)) + if (kstrtoul(buf, 0, &cnt)) return -EINVAL; ret = crash_shrink_memory(cnt); diff --git a/kernel/params.c b/kernel/params.c index 501bde4f3bee..81c4e78c8f4c 100644 --- a/kernel/params.c +++ b/kernel/params.c @@ -253,13 +253,13 @@ int parse_args(const char *doing, EXPORT_SYMBOL(param_ops_##name) -STANDARD_PARAM_DEF(byte, unsigned char, "%hhu", unsigned long, strict_strtoul); -STANDARD_PARAM_DEF(short, short, "%hi", long, strict_strtol); -STANDARD_PARAM_DEF(ushort, unsigned short, "%hu", unsigned long, strict_strtoul); -STANDARD_PARAM_DEF(int, int, "%i", long, strict_strtol); -STANDARD_PARAM_DEF(uint, unsigned int, "%u", unsigned long, strict_strtoul); -STANDARD_PARAM_DEF(long, long, "%li", long, strict_strtol); -STANDARD_PARAM_DEF(ulong, unsigned long, "%lu", unsigned long, strict_strtoul); +STANDARD_PARAM_DEF(byte, unsigned char, "%hhu", unsigned long, kstrtoul); +STANDARD_PARAM_DEF(short, short, "%hi", long, kstrtoul); +STANDARD_PARAM_DEF(ushort, unsigned short, "%hu", unsigned long, kstrtoul); +STANDARD_PARAM_DEF(int, int, "%i", long, kstrtoul); +STANDARD_PARAM_DEF(uint, unsigned int, "%u", unsigned long, kstrtoul); +STANDARD_PARAM_DEF(long, long, "%li", long, kstrtoul); +STANDARD_PARAM_DEF(ulong, unsigned long, "%lu", unsigned long, kstrtoul); int param_set_charp(const char *val, const struct kernel_param *kp) { -- cgit v1.2.3 From 0244ad004a54e39308d495fee0a2e637f8b5c317 Mon Sep 17 00:00:00 2001 From: Martin Schwidefsky Date: Fri, 30 Aug 2013 09:39:53 +0200 Subject: Remove GENERIC_HARDIRQ config option After the last architecture switched to generic hard irqs the config options HAVE_GENERIC_HARDIRQS & GENERIC_HARDIRQS and the related code for !CONFIG_GENERIC_HARDIRQS can be removed. Signed-off-by: Martin Schwidefsky --- kernel/Makefile | 2 +- kernel/irq/Kconfig | 12 ------------ kernel/softirq.c | 2 -- 3 files changed, 1 insertion(+), 15 deletions(-) (limited to 'kernel') diff --git a/kernel/Makefile b/kernel/Makefile index 35ef1185e359..1ce47553fb02 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -26,6 +26,7 @@ obj-y += sched/ obj-y += power/ obj-y += printk/ obj-y += cpu/ +obj-y += irq/ obj-$(CONFIG_CHECKPOINT_RESTORE) += kcmp.o obj-$(CONFIG_FREEZER) += freezer.o @@ -79,7 +80,6 @@ obj-$(CONFIG_KPROBES) += kprobes.o obj-$(CONFIG_KGDB) += debug/ obj-$(CONFIG_DETECT_HUNG_TASK) += hung_task.o obj-$(CONFIG_LOCKUP_DETECTOR) += watchdog.o -obj-$(CONFIG_GENERIC_HARDIRQS) += irq/ obj-$(CONFIG_SECCOMP) += seccomp.o obj-$(CONFIG_RCU_TORTURE_TEST) += rcutorture.o obj-$(CONFIG_TREE_RCU) += rcutree.o diff --git a/kernel/irq/Kconfig b/kernel/irq/Kconfig index d1a758bc972a..4a1fef09f658 100644 --- a/kernel/irq/Kconfig +++ b/kernel/irq/Kconfig @@ -1,15 +1,4 @@ -# Select this to activate the generic irq options below -config HAVE_GENERIC_HARDIRQS - bool - -if HAVE_GENERIC_HARDIRQS menu "IRQ subsystem" -# -# Interrupt subsystem related configuration options -# -config GENERIC_HARDIRQS - def_bool y - # Options selectable by the architecture code # Make sparse irq Kconfig switch below available @@ -84,4 +73,3 @@ config SPARSE_IRQ If you don't know what to do here, say N. endmenu -endif diff --git a/kernel/softirq.c b/kernel/softirq.c index be3d3514c325..53cc09ceb0b8 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -876,7 +876,6 @@ int __init __weak early_irq_init(void) return 0; } -#ifdef CONFIG_GENERIC_HARDIRQS int __init __weak arch_probe_nr_irqs(void) { return NR_IRQS_LEGACY; @@ -886,4 +885,3 @@ int __init __weak arch_early_irq_init(void) { return 0; } -#endif -- cgit v1.2.3 From 13b62e46d5407c7d619aea1dc9c3e0991b631b57 Mon Sep 17 00:00:00 2001 From: "Michael S. Tsirkin" Date: Mon, 16 Sep 2013 11:30:36 +0300 Subject: sched: Fix comment for sched_info_depart sched_info_depart seems to be only called from sched_info_switch(), so only on involuntary task switch. Fix the comment to match. Signed-off-by: Michael S. Tsirkin Cc: Peter Zijlstra Cc: Frederic Weisbecker Cc: KOSAKI Motohiro Link: http://lkml.kernel.org/r/20130916083036.GA1113@redhat.com Signed-off-by: Ingo Molnar --- kernel/sched/stats.h | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/stats.h b/kernel/sched/stats.h index 5aef494fc8b4..c7edee71bce8 100644 --- a/kernel/sched/stats.h +++ b/kernel/sched/stats.h @@ -104,8 +104,9 @@ static inline void sched_info_queued(struct task_struct *t) } /* - * Called when a process ceases being the active-running process, either - * voluntarily or involuntarily. Now we can calculate how long we ran. + * Called when a process ceases being the active-running process involuntarily + * due, typically, to expiring its time slice (this may also be called when + * switching to the idle task). Now we can calculate how long we ran. * Also, if the process is still in the TASK_RUNNING state, call * sched_info_queued() to mark that it has now again started waiting on * the runqueue. -- cgit v1.2.3 From fa7315871046b9a4c48627905691dbde57e51033 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 19 Sep 2013 10:16:42 +0200 Subject: perf: Fix capabilities bitfield compatibility in 'struct perf_event_mmap_page' Solve the problems around the broken definition of perf_event_mmap_page:: cap_usr_time and cap_usr_rdpmc fields which used to overlap, partially fixed by: 860f085b74e9 ("perf: Fix broken union in 'struct perf_event_mmap_page'") The problem with the fix (merged in v3.12-rc1 and not yet released officially), noticed by Vince Weaver is that the new behavior is not detectable by new user-space, and that due to the reuse of the field names it's easy to mis-compile a binary if old headers are used on a new kernel or new headers are used on an old kernel. To solve all that make this change explicit, detectable and self-contained, by iterating the ABI the following way: - Always clear bit 0, and rename it to usrpage->cap_bit0, to at least not confuse old user-space binaries. RDPMC will be marked as unavailable to old binaries but that's within the ABI, this is a capability bit. - Rename bit 1 to ->cap_bit0_is_deprecated and always set it to 1, so new libraries can reliably detect that bit 0 is deprecated and perma-zero without having to check the kernel version. - Use bits 2, 3, 4 for the newly defined, correct functionality: cap_user_rdpmc : 1, /* The RDPMC instruction can be used to read counts */ cap_user_time : 1, /* The time_* fields are used */ cap_user_time_zero : 1, /* The time_zero field is used */ - Rename all the bitfield names in perf_event.h to be different from the old names, to make sure it's not possible to mis-compile it accidentally with old assumptions. The 'size' field can then be used in the future to add new fields and it will act as a natural ABI version indicator as well. Also adjust tools/perf/ userspace for the new definitions, noticed by Adrian Hunter. Reported-by: Vince Weaver Signed-off-by: Peter Zijlstra Also-Fixed-by: Adrian Hunter Link: http://lkml.kernel.org/n/tip-zr03yxjrpXesOzzupszqglbv@git.kernel.org Signed-off-by: Ingo Molnar --- kernel/events/core.c | 21 +++++++++++++++++++++ 1 file changed, 21 insertions(+) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index dd236b66ca3a..cb4238e85b38 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -3660,6 +3660,26 @@ static void calc_timer_values(struct perf_event *event, *running = ctx_time - event->tstamp_running; } +static void perf_event_init_userpage(struct perf_event *event) +{ + struct perf_event_mmap_page *userpg; + struct ring_buffer *rb; + + rcu_read_lock(); + rb = rcu_dereference(event->rb); + if (!rb) + goto unlock; + + userpg = rb->user_page; + + /* Allow new userspace to detect that bit 0 is deprecated */ + userpg->cap_bit0_is_deprecated = 1; + userpg->size = offsetof(struct perf_event_mmap_page, __reserved); + +unlock: + rcu_read_unlock(); +} + void __weak arch_perf_update_userpage(struct perf_event_mmap_page *userpg, u64 now) { } @@ -4044,6 +4064,7 @@ again: ring_buffer_attach(event, rb); rcu_assign_pointer(event->rb, rb); + perf_event_init_userpage(event); perf_event_update_userpage(event); unlock: -- cgit v1.2.3 From b18855500fc40da050512d9df82d2f1471e59642 Mon Sep 17 00:00:00 2001 From: Vladimir Davydov Date: Sun, 15 Sep 2013 17:49:13 +0400 Subject: sched/balancing: Fix 'local->avg_load > sds->avg_load' case in calculate_imbalance() In busiest->group_imb case we can come to calculate_imbalance() with local->avg_load >= busiest->avg_load >= sds->avg_load. This can result in imbalance overflow, because it is calculated as follows env->imbalance = min( max_pull * busiest->group_power, (sds->avg_load - local->avg_load) * local->group_power) / SCHED_POWER_SCALE; As a result we can end up constantly bouncing tasks from one cpu to another if there are pinned tasks. Fix this by skipping the assignment and assuming imbalance=0 in case local->avg_load > sds->avg_load. [ The bug can be caught by running 2*N cpuhogs pinned to two logical cpus belonging to different cores on an HT-enabled machine with N logical cpus: just look at se.nr_migrations growth. ] Signed-off-by: Vladimir Davydov Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/8f596cc6bc0e5e655119dc892c9bfcad26e971f4.1379252740.git.vdavydov@parallels.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 11cd13667359..0b99aae339cb 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -4896,7 +4896,8 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s * max load less than avg load(as we skip the groups at or below * its cpu_power, while calculating max_load..) */ - if (busiest->avg_load < sds->avg_load) { + if (busiest->avg_load <= sds->avg_load || + local->avg_load >= sds->avg_load) { env->imbalance = 0; return fix_small_imbalance(env, sds); } -- cgit v1.2.3 From 3029ede39373c368f402a76896600d85a4f7121b Mon Sep 17 00:00:00 2001 From: Vladimir Davydov Date: Sun, 15 Sep 2013 17:49:14 +0400 Subject: sched/balancing: Fix 'local->avg_load > busiest->avg_load' case in fix_small_imbalance() In busiest->group_imb case we can come to fix_small_imbalance() with local->avg_load > busiest->avg_load. This can result in wrong imbalance fix-up, because there is the following check there where all the members are unsigned: if (busiest->avg_load - local->avg_load + scaled_busy_load_per_task >= (scaled_busy_load_per_task * imbn)) { env->imbalance = busiest->load_per_task; return; } As a result we can end up constantly bouncing tasks from one cpu to another if there are pinned tasks. Fix it by substituting the subtraction with an equivalent addition in the check. [ The bug can be caught by running 2*N cpuhogs pinned to two logical cpus belonging to different cores on an HT-enabled machine with N logical cpus: just look at se.nr_migrations growth. ] Signed-off-by: Vladimir Davydov Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/ef167822e5c5b2d96cf5b0e3e4f4bdff3f0414a2.1379252740.git.vdavydov@parallels.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 0b99aae339cb..2aedaccebcc8 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -4823,8 +4823,8 @@ void fix_small_imbalance(struct lb_env *env, struct sd_lb_stats *sds) (busiest->load_per_task * SCHED_POWER_SCALE) / busiest->group_power; - if (busiest->avg_load - local->avg_load + scaled_busy_load_per_task >= - (scaled_busy_load_per_task * imbn)) { + if (busiest->avg_load + scaled_busy_load_per_task >= + local->avg_load + (scaled_busy_load_per_task * imbn)) { env->imbalance = busiest->load_per_task; return; } -- cgit v1.2.3 From 7e3115ef5149fc502e3a2e80719dba54a8e7409d Mon Sep 17 00:00:00 2001 From: Vladimir Davydov Date: Sat, 14 Sep 2013 19:39:46 +0400 Subject: sched/balancing: Fix cfs_rq->task_h_load calculation Patch a003a2 (sched: Consider runnable load average in move_tasks()) sets all top-level cfs_rqs' h_load to rq->avg.load_avg_contrib, which is always 0. This mistype leads to all tasks having weight 0 when load balancing in a cpu-cgroup enabled setup. There obviously should be sum of weights of all runnable tasks there instead. Fix it. Signed-off-by: Vladimir Davydov Reviewed-by: Paul Turner Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1379173186-11944-1-git-send-email-vdavydov@parallels.com Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 2aedaccebcc8..7c70201fbc61 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -4242,7 +4242,7 @@ static void update_cfs_rq_h_load(struct cfs_rq *cfs_rq) } if (!se) { - cfs_rq->h_load = rq->avg.load_avg_contrib; + cfs_rq->h_load = cfs_rq->runnable_load_avg; cfs_rq->last_h_load_update = now; } -- cgit v1.2.3 From 359e6fab6600562073162348cd4c18c5958296d8 Mon Sep 17 00:00:00 2001 From: Michal Hocko Date: Tue, 24 Sep 2013 15:27:29 -0700 Subject: watchdog: update watchdog attributes atomically proc_dowatchdog doesn't synchronize multiple callers which might lead to confusion when two parallel callers might confuse watchdog_enable_all_cpus resp watchdog_disable_all_cpus (eg watchdog gets enabled even if watchdog_thresh was set to 0 already). This patch adds a local mutex which synchronizes callers to the sysctl handler. Signed-off-by: Michal Hocko Cc: Frederic Weisbecker Acked-by: Don Zickus Cc: Thomas Gleixner Cc: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/watchdog.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 51c4f34d258e..ced7d0609931 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -520,13 +520,15 @@ int proc_dowatchdog(struct ctl_table *table, int write, void __user *buffer, size_t *lenp, loff_t *ppos) { int err, old_thresh, old_enabled; + static DEFINE_MUTEX(watchdog_proc_mutex); + mutex_lock(&watchdog_proc_mutex); old_thresh = ACCESS_ONCE(watchdog_thresh); old_enabled = ACCESS_ONCE(watchdog_user_enabled); err = proc_dointvec_minmax(table, write, buffer, lenp, ppos); if (err || !write) - return err; + goto out; set_sample_period(); /* @@ -544,7 +546,8 @@ int proc_dowatchdog(struct ctl_table *table, int write, watchdog_thresh = old_thresh; watchdog_user_enabled = old_enabled; } - +out: + mutex_unlock(&watchdog_proc_mutex); return err; } #endif /* CONFIG_SYSCTL */ -- cgit v1.2.3 From 9809b18fcf6b8d8ec4d3643677345907e6b50eca Mon Sep 17 00:00:00 2001 From: Michal Hocko Date: Tue, 24 Sep 2013 15:27:30 -0700 Subject: watchdog: update watchdog_thresh properly watchdog_tresh controls how often nmi perf event counter checks per-cpu hrtimer_interrupts counter and blows up if the counter hasn't changed since the last check. The counter is updated by per-cpu watchdog_hrtimer hrtimer which is scheduled with 2/5 watchdog_thresh period which guarantees that hrtimer is scheduled 2 times per the main period. Both hrtimer and perf event are started together when the watchdog is enabled. So far so good. But... But what happens when watchdog_thresh is updated from sysctl handler? proc_dowatchdog will set a new sampling period and hrtimer callback (watchdog_timer_fn) will use the new value in the next round. The problem, however, is that nobody tells the perf event that the sampling period has changed so it is ticking with the period configured when it has been set up. This might result in an ear ripping dissonance between perf and hrtimer parts if the watchdog_thresh is increased. And even worse it might lead to KABOOM if the watchdog is configured to panic on such a spurious lockup. This patch fixes the issue by updating both nmi perf even counter and hrtimers if the threshold value has changed. The nmi one is disabled and then reinitialized from scratch. This has an unpleasant side effect that the allocation of the new event might fail theoretically so the hard lockup detector would be disabled for such cpus. On the other hand such a memory allocation failure is very unlikely because the original event is deallocated right before. It would be much nicer if we just changed perf event period but there doesn't seem to be any API to do that right now. It is also unfortunate that perf_event_alloc uses GFP_KERNEL allocation unconditionally so we cannot use on_each_cpu() and do the same thing from the per-cpu context. The update from the current CPU should be safe because perf_event_disable removes the event atomically before it clears the per-cpu watchdog_ev so it cannot change anything under running handler feet. The hrtimer is simply restarted (thanks to Don Zickus who has pointed this out) if it is queued because we cannot rely it will fire&adopt to the new sampling period before a new nmi event triggers (when the treshold is decreased). [akpm@linux-foundation.org: the UP version of __smp_call_function_single ended up in the wrong place] Signed-off-by: Michal Hocko Acked-by: Don Zickus Cc: Frederic Weisbecker Cc: Thomas Gleixner Cc: Ingo Molnar Cc: Fabio Estevam Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/watchdog.c | 53 ++++++++++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 50 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/watchdog.c b/kernel/watchdog.c index ced7d0609931..4431610f049a 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -486,7 +486,52 @@ static struct smp_hotplug_thread watchdog_threads = { .unpark = watchdog_enable, }; -static int watchdog_enable_all_cpus(void) +static void restart_watchdog_hrtimer(void *info) +{ + struct hrtimer *hrtimer = &__raw_get_cpu_var(watchdog_hrtimer); + int ret; + + /* + * No need to cancel and restart hrtimer if it is currently executing + * because it will reprogram itself with the new period now. + * We should never see it unqueued here because we are running per-cpu + * with interrupts disabled. + */ + ret = hrtimer_try_to_cancel(hrtimer); + if (ret == 1) + hrtimer_start(hrtimer, ns_to_ktime(sample_period), + HRTIMER_MODE_REL_PINNED); +} + +static void update_timers(int cpu) +{ + struct call_single_data data = {.func = restart_watchdog_hrtimer}; + /* + * Make sure that perf event counter will adopt to a new + * sampling period. Updating the sampling period directly would + * be much nicer but we do not have an API for that now so + * let's use a big hammer. + * Hrtimer will adopt the new period on the next tick but this + * might be late already so we have to restart the timer as well. + */ + watchdog_nmi_disable(cpu); + __smp_call_function_single(cpu, &data, 1); + watchdog_nmi_enable(cpu); +} + +static void update_timers_all_cpus(void) +{ + int cpu; + + get_online_cpus(); + preempt_disable(); + for_each_online_cpu(cpu) + update_timers(cpu); + preempt_enable(); + put_online_cpus(); +} + +static int watchdog_enable_all_cpus(bool sample_period_changed) { int err = 0; @@ -496,6 +541,8 @@ static int watchdog_enable_all_cpus(void) pr_err("Failed to create watchdog threads, disabled\n"); else watchdog_running = 1; + } else if (sample_period_changed) { + update_timers_all_cpus(); } return err; @@ -537,7 +584,7 @@ int proc_dowatchdog(struct ctl_table *table, int write, * watchdog_*_all_cpus() function takes care of this. */ if (watchdog_user_enabled && watchdog_thresh) - err = watchdog_enable_all_cpus(); + err = watchdog_enable_all_cpus(old_thresh != watchdog_thresh); else watchdog_disable_all_cpus(); @@ -557,5 +604,5 @@ void __init lockup_detector_init(void) set_sample_period(); if (watchdog_user_enabled) - watchdog_enable_all_cpus(); + watchdog_enable_all_cpus(false); } -- cgit v1.2.3 From 8ac1c8d5deba65513b6a82c35e89e73996c8e0d6 Mon Sep 17 00:00:00 2001 From: Konstantin Khlebnikov Date: Tue, 24 Sep 2013 15:27:42 -0700 Subject: audit: fix endless wait in audit_log_start() After commit 829199197a43 ("kernel/audit.c: avoid negative sleep durations") audit emitters will block forever if userspace daemon cannot handle backlog. After the timeout the waiting loop turns into busy loop and runs until daemon dies or returns back to work. This is a minimal patch for that bug. Signed-off-by: Konstantin Khlebnikov Cc: Luiz Capitulino Cc: Richard Guy Briggs Cc: Eric Paris Cc: Chuck Anderson Cc: Dan Duval Cc: Dave Kleikamp Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/audit.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/audit.c b/kernel/audit.c index 91e53d04b6a9..7b0e23a740ce 100644 --- a/kernel/audit.c +++ b/kernel/audit.c @@ -1117,9 +1117,10 @@ struct audit_buffer *audit_log_start(struct audit_context *ctx, gfp_t gfp_mask, sleep_time = timeout_start + audit_backlog_wait_time - jiffies; - if ((long)sleep_time > 0) + if ((long)sleep_time > 0) { wait_for_auditd(sleep_time); - continue; + continue; + } } if (audit_rate_check() && printk_ratelimit()) printk(KERN_WARNING -- cgit v1.2.3 From e2f0b88e84eed9cf9797f0a88c8012ee0b885a6d Mon Sep 17 00:00:00 2001 From: Chuansheng Liu Date: Tue, 24 Sep 2013 15:27:43 -0700 Subject: kernel/reboot.c: re-enable the function of variable reboot_default Commit 1b3a5d02ee07 ("reboot: move arch/x86 reboot= handling to generic kernel") did some cleanup for reboot= command line, but it made the reboot_default inoperative. The default value of variable reboot_default should be 1, and if command line reboot= is not set, system will use the default reboot mode. [akpm@linux-foundation.org: fix comment layout] Signed-off-by: Li Fei Signed-off-by: liu chuansheng Acked-by: Robin Holt Cc: [3.11.x] Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/reboot.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/reboot.c b/kernel/reboot.c index 269ed9384cc4..f813b3474646 100644 --- a/kernel/reboot.c +++ b/kernel/reboot.c @@ -32,7 +32,14 @@ EXPORT_SYMBOL(cad_pid); #endif enum reboot_mode reboot_mode DEFAULT_REBOOT_MODE; -int reboot_default; +/* + * This variable is used privately to keep track of whether or not + * reboot_type is still set to its default value (i.e., reboot= hasn't + * been set on the command line). This is needed so that we can + * suppress DMI scanning for reboot quirks. Without it, it's + * impossible to override a faulty reboot quirk without recompiling. + */ +int reboot_default = 1; int reboot_cpu; enum reboot_type reboot_type = BOOT_ACPI; int reboot_force; -- cgit v1.2.3 From 0c06a5d4b13cd66c833805a0d1db76b977944aac Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 10 Sep 2013 00:54:17 +0200 Subject: arm: Fix build error with context tracking calls ad65782fba50 (context_tracking: Optimize main APIs off case with static key) converted context tracking main APIs to inline function and left ARM asm callers behind. This can be easily fixed by making ARM calling the post static keys context tracking function. We just need to replicate the static key checks there. We'll remove these later when ARM will support the context tracking static keys. Reported-by: Guenter Roeck Reported-by: Russell King Signed-off-by: Frederic Weisbecker Tested-by: Kevin Hilman Cc: Nicolas Pitre Cc: Anil Kumar Cc: Tony Lindgren Cc: Benoit Cousson Cc: Guenter Roeck Cc: Russell King Cc: Kevin Hilman --- kernel/context_tracking.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'kernel') diff --git a/kernel/context_tracking.c b/kernel/context_tracking.c index 247091bf0587..859c8dfd78a1 100644 --- a/kernel/context_tracking.c +++ b/kernel/context_tracking.c @@ -50,6 +50,15 @@ void context_tracking_user_enter(void) { unsigned long flags; + /* + * Repeat the user_enter() check here because some archs may be calling + * this from asm and if no CPU needs context tracking, they shouldn't + * go further. Repeat the check here until they support the static key + * check. + */ + if (!static_key_false(&context_tracking_enabled)) + return; + /* * Some contexts may involve an exception occuring in an irq, * leading to that nesting: @@ -151,6 +160,9 @@ void context_tracking_user_exit(void) { unsigned long flags; + if (!static_key_false(&context_tracking_enabled)) + return; + if (in_interrupt()) return; -- cgit v1.2.3 From 3a126f85e015701e56240884f27f97543580d5f7 Mon Sep 17 00:00:00 2001 From: Jean Delvare Date: Fri, 27 Sep 2013 13:17:39 -0700 Subject: kernel/params: fix handling of signed integer types Commit 6072ddc8520b ("kernel: replace strict_strto*() with kstrto*()") broke the handling of signed integer types, fix it. Signed-off-by: Jean Delvare Reported-by: Christian Kujau Tested-by: Christian Kujau Cc: Jingoo Han Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/params.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/params.c b/kernel/params.c index 81c4e78c8f4c..c00d5b502aa4 100644 --- a/kernel/params.c +++ b/kernel/params.c @@ -254,11 +254,11 @@ int parse_args(const char *doing, STANDARD_PARAM_DEF(byte, unsigned char, "%hhu", unsigned long, kstrtoul); -STANDARD_PARAM_DEF(short, short, "%hi", long, kstrtoul); +STANDARD_PARAM_DEF(short, short, "%hi", long, kstrtol); STANDARD_PARAM_DEF(ushort, unsigned short, "%hu", unsigned long, kstrtoul); -STANDARD_PARAM_DEF(int, int, "%i", long, kstrtoul); +STANDARD_PARAM_DEF(int, int, "%i", long, kstrtol); STANDARD_PARAM_DEF(uint, unsigned int, "%u", unsigned long, kstrtoul); -STANDARD_PARAM_DEF(long, long, "%li", long, kstrtoul); +STANDARD_PARAM_DEF(long, long, "%li", long, kstrtol); STANDARD_PARAM_DEF(ulong, unsigned long, "%lu", unsigned long, kstrtoul); int param_set_charp(const char *val, const struct kernel_param *kp) -- cgit v1.2.3 From aab1728915420b5288cd0fc7b5bd320105b48983 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Mon, 30 Sep 2013 19:40:56 +0200 Subject: PM / hibernate: Fix user space driven resume regression Recent commit 8fd37a4 (PM / hibernate: Create memory bitmaps after freezing user space) broke the resume part of the user space driven hibernation (s2disk), because I forgot that the resume utility loaded the image into memory without freezing user space (it still freezes tasks after loading the image). This means that during user space driven resume we need to create the memory bitmaps at the "device open" time rather than at the "freeze tasks" time, so make that happen (that's a special case anyway, so it needs to be treated in a special way). Reported-and-tested-by: Ronald Signed-off-by: Rafael J. Wysocki --- kernel/power/snapshot.c | 5 ++++- kernel/power/user.c | 8 ++++++++ 2 files changed, 12 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c index 358a146fd4da..98c3b34a4cff 100644 --- a/kernel/power/snapshot.c +++ b/kernel/power/snapshot.c @@ -743,7 +743,10 @@ int create_basic_memory_bitmaps(void) struct memory_bitmap *bm1, *bm2; int error = 0; - BUG_ON(forbidden_pages_map || free_pages_map); + if (forbidden_pages_map && free_pages_map) + return 0; + else + BUG_ON(forbidden_pages_map || free_pages_map); bm1 = kzalloc(sizeof(struct memory_bitmap), GFP_KERNEL); if (!bm1) diff --git a/kernel/power/user.c b/kernel/power/user.c index 72e8f4fd616d..957f06164ad1 100644 --- a/kernel/power/user.c +++ b/kernel/power/user.c @@ -39,6 +39,7 @@ static struct snapshot_data { char frozen; char ready; char platform_support; + bool free_bitmaps; } snapshot_state; atomic_t snapshot_device_available = ATOMIC_INIT(1); @@ -82,6 +83,10 @@ static int snapshot_open(struct inode *inode, struct file *filp) data->swap = -1; data->mode = O_WRONLY; error = pm_notifier_call_chain(PM_RESTORE_PREPARE); + if (!error) { + error = create_basic_memory_bitmaps(); + data->free_bitmaps = !error; + } if (error) pm_notifier_call_chain(PM_POST_RESTORE); } @@ -111,6 +116,8 @@ static int snapshot_release(struct inode *inode, struct file *filp) pm_restore_gfp_mask(); free_basic_memory_bitmaps(); thaw_processes(); + } else if (data->free_bitmaps) { + free_basic_memory_bitmaps(); } pm_notifier_call_chain(data->mode == O_RDONLY ? PM_POST_HIBERNATION : PM_POST_RESTORE); @@ -231,6 +238,7 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd, break; pm_restore_gfp_mask(); free_basic_memory_bitmaps(); + data->free_bitmaps = false; thaw_processes(); data->frozen = 0; break; -- cgit v1.2.3 From 4c1c7be95c345cf2ad537a0c48e9aeadc7304527 Mon Sep 17 00:00:00 2001 From: Tetsuo Handa Date: Mon, 30 Sep 2013 13:45:08 -0700 Subject: kernel/kmod.c: check for NULL in call_usermodehelper_exec() If /proc/sys/kernel/core_pattern contains only "|", a NULL pointer dereference happens upon core dump because argv_split("") returns argv[0] == NULL. This bug was once fixed by commit 264b83c07a84 ("usermodehelper: check subprocess_info->path != NULL") but was by error reintroduced by commit 7f57cfa4e2aa ("usermodehelper: kill the sub_info->path[0] check"). This bug seems to exist since 2.6.19 (the version which core dump to pipe was added). Depending on kernel version and config, some side effect might happen immediately after this oops (e.g. kernel panic with 2.6.32-358.18.1.el6). Signed-off-by: Tetsuo Handa Acked-by: Oleg Nesterov Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/kmod.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'kernel') diff --git a/kernel/kmod.c b/kernel/kmod.c index fb326365b694..b086006c59e7 100644 --- a/kernel/kmod.c +++ b/kernel/kmod.c @@ -571,6 +571,10 @@ int call_usermodehelper_exec(struct subprocess_info *sub_info, int wait) DECLARE_COMPLETION_ONSTACK(done); int retval = 0; + if (!sub_info->path) { + call_usermodehelper_freeinfo(sub_info); + return -EINVAL; + } helper_lock(); if (!khelper_wq || usermodehelper_disabled) { retval = -EBUSY; -- cgit v1.2.3 From 314a8ad0f18ac37887896b288939acd8cb17e208 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Mon, 30 Sep 2013 13:45:27 -0700 Subject: pidns: fix free_pid() to handle the first fork failure "case 0" in free_pid() assumes that disable_pid_allocation() should clear PIDNS_HASH_ADDING before the last pid goes away. However this doesn't happen if the first fork() fails to create the child reaper which should call disable_pid_allocation(). Signed-off-by: Oleg Nesterov Reviewed-by: "Eric W. Biederman" Cc: "Serge E. Hallyn" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/pid.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'kernel') diff --git a/kernel/pid.c b/kernel/pid.c index ebe5e80b10f8..9b9a26698144 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -273,6 +273,11 @@ void free_pid(struct pid *pid) */ wake_up_process(ns->child_reaper); break; + case PIDNS_HASH_ADDING: + /* Handle a fork failure of the first process */ + WARN_ON(ns->child_reaper); + ns->nr_hashed = 0; + /* fall through */ case 0: schedule_work(&ns->proc_work); break; -- cgit v1.2.3 From ded797547548a5b8e7b92383a41e4c0e6b0ecb7f Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 24 Sep 2013 00:50:25 +0200 Subject: irq: Force hardirq exit's softirq processing on its own stack The commit facd8b80c67a3cf64a467c4a2ac5fb31f2e6745b ("irq: Sanitize invoke_softirq") converted irq exit calls of do_softirq() to __do_softirq() on all architectures, assuming it was only used there for its irq disablement properties. But as a side effect, the softirqs processed in the end of the hardirq are always called on the inline current stack that is used by irq_exit() instead of the softirq stack provided by the archs that override do_softirq(). The result is mostly safe if the architecture runs irq_exit() on a separate irq stack because then softirqs are processed on that same stack that is near empty at this stage (assuming hardirq aren't nesting). Otherwise irq_exit() runs in the task stack and so does the softirq too. The interrupted call stack can be randomly deep already and the softirq can dig through it even further. To add insult to the injury, this softirq can be interrupted by a new hardirq, maximizing the chances for a stack overrun as reported in powerpc for example: do_IRQ: stack overflow: 1920 CPU: 0 PID: 1602 Comm: qemu-system-ppc Not tainted 3.10.4-300.1.fc19.ppc64p7 #1 Call Trace: [c0000000050a8740] .show_stack+0x130/0x200 (unreliable) [c0000000050a8810] .dump_stack+0x28/0x3c [c0000000050a8880] .do_IRQ+0x2b8/0x2c0 [c0000000050a8930] hardware_interrupt_common+0x154/0x180 --- Exception: 501 at .cp_start_xmit+0x3a4/0x820 [8139cp] LR = .cp_start_xmit+0x390/0x820 [8139cp] [c0000000050a8d40] .dev_hard_start_xmit+0x394/0x640 [c0000000050a8e00] .sch_direct_xmit+0x110/0x260 [c0000000050a8ea0] .dev_queue_xmit+0x260/0x630 [c0000000050a8f40] .br_dev_queue_push_xmit+0xc4/0x130 [bridge] [c0000000050a8fc0] .br_dev_xmit+0x198/0x270 [bridge] [c0000000050a9070] .dev_hard_start_xmit+0x394/0x640 [c0000000050a9130] .dev_queue_xmit+0x428/0x630 [c0000000050a91d0] .ip_finish_output+0x2a4/0x550 [c0000000050a9290] .ip_local_out+0x50/0x70 [c0000000050a9310] .ip_queue_xmit+0x148/0x420 [c0000000050a93b0] .tcp_transmit_skb+0x4e4/0xaf0 [c0000000050a94a0] .__tcp_ack_snd_check+0x7c/0xf0 [c0000000050a9520] .tcp_rcv_established+0x1e8/0x930 [c0000000050a95f0] .tcp_v4_do_rcv+0x21c/0x570 [c0000000050a96c0] .tcp_v4_rcv+0x734/0x930 [c0000000050a97a0] .ip_local_deliver_finish+0x184/0x360 [c0000000050a9840] .ip_rcv_finish+0x148/0x400 [c0000000050a98d0] .__netif_receive_skb_core+0x4f8/0xb00 [c0000000050a99d0] .netif_receive_skb+0x44/0x110 [c0000000050a9a70] .br_handle_frame_finish+0x2bc/0x3f0 [bridge] [c0000000050a9b20] .br_nf_pre_routing_finish+0x2ac/0x420 [bridge] [c0000000050a9bd0] .br_nf_pre_routing+0x4dc/0x7d0 [bridge] [c0000000050a9c70] .nf_iterate+0x114/0x130 [c0000000050a9d30] .nf_hook_slow+0xb4/0x1e0 [c0000000050a9e00] .br_handle_frame+0x290/0x330 [bridge] [c0000000050a9ea0] .__netif_receive_skb_core+0x34c/0xb00 [c0000000050a9fa0] .netif_receive_skb+0x44/0x110 [c0000000050aa040] .napi_gro_receive+0xe8/0x120 [c0000000050aa0c0] .cp_rx_poll+0x31c/0x590 [8139cp] [c0000000050aa1d0] .net_rx_action+0x1dc/0x310 [c0000000050aa2b0] .__do_softirq+0x158/0x330 [c0000000050aa3b0] .irq_exit+0xc8/0x110 [c0000000050aa430] .do_IRQ+0xdc/0x2c0 [c0000000050aa4e0] hardware_interrupt_common+0x154/0x180 --- Exception: 501 at .bad_range+0x1c/0x110 LR = .get_page_from_freelist+0x908/0xbb0 [c0000000050aa7d0] .list_del+0x18/0x50 (unreliable) [c0000000050aa850] .get_page_from_freelist+0x908/0xbb0 [c0000000050aa9e0] .__alloc_pages_nodemask+0x21c/0xae0 [c0000000050aaba0] .alloc_pages_vma+0xd0/0x210 [c0000000050aac60] .handle_pte_fault+0x814/0xb70 [c0000000050aad50] .__get_user_pages+0x1a4/0x640 [c0000000050aae60] .get_user_pages_fast+0xec/0x160 [c0000000050aaf10] .__gfn_to_pfn_memslot+0x3b0/0x430 [kvm] [c0000000050aafd0] .kvmppc_gfn_to_pfn+0x64/0x130 [kvm] [c0000000050ab070] .kvmppc_mmu_map_page+0x94/0x530 [kvm] [c0000000050ab190] .kvmppc_handle_pagefault+0x174/0x610 [kvm] [c0000000050ab270] .kvmppc_handle_exit_pr+0x464/0x9b0 [kvm] [c0000000050ab320] kvm_start_lightweight+0x1ec/0x1fc [kvm] [c0000000050ab4f0] .kvmppc_vcpu_run_pr+0x168/0x3b0 [kvm] [c0000000050ab9c0] .kvmppc_vcpu_run+0xc8/0xf0 [kvm] [c0000000050aba50] .kvm_arch_vcpu_ioctl_run+0x5c/0x1a0 [kvm] [c0000000050abae0] .kvm_vcpu_ioctl+0x478/0x730 [kvm] [c0000000050abc90] .do_vfs_ioctl+0x4ec/0x7c0 [c0000000050abd80] .SyS_ioctl+0xd4/0xf0 [c0000000050abe30] syscall_exit+0x0/0x98 Since this is a regression, this patch proposes a minimalistic and low-risk solution by blindly forcing the hardirq exit processing of softirqs on the softirq stack. This way we should reduce significantly the opportunities for task stack overflow dug by softirqs. Longer term solutions may involve extending the hardirq stack coverage to irq_exit(), etc... Reported-by: Benjamin Herrenschmidt Acked-by: Linus Torvalds Signed-off-by: Frederic Weisbecker Cc: #3.9.. Cc: Benjamin Herrenschmidt Cc: Paul Mackerras Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: H. Peter Anvin Cc: Linus Torvalds Cc: Paul Mackerras Cc: James Hogan Cc: James E.J. Bottomley Cc: Helge Deller Cc: Martin Schwidefsky Cc: Heiko Carstens Cc: David S. Miller Cc: Andrew Morton --- kernel/softirq.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/softirq.c b/kernel/softirq.c index 53cc09ceb0b8..d7d498d8cc4f 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -328,10 +328,19 @@ void irq_enter(void) static inline void invoke_softirq(void) { - if (!force_irqthreads) - __do_softirq(); - else + if (!force_irqthreads) { + /* + * We can safely execute softirq on the current stack if + * it is the irq stack, because it should be near empty + * at this stage. But we have no way to know if the arch + * calls irq_exit() on the irq stack. So call softirq + * in its own stack to prevent from any overrun on top + * of a potentially deep task stack. + */ + do_softirq(); + } else { wakeup_softirqd(); + } } static inline void tick_irq_exit(void) -- cgit v1.2.3