From f319da0c6894fcf55e21320e40506418a2aad629 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 20 Aug 2012 11:26:57 +0200 Subject: sched: Fix load avg vs cpu-hotplug Rabik and Paul reported two different issues related to the same few lines of code. Rabik's issue is that the nr_uninterruptible migration code is wrong in that he sees artifacts due to this (Rabik please do expand in more detail). Paul's issue is that this code as it stands relies on us using stop_machine() for unplug, we all would like to remove this assumption so that eventually we can remove this stop_machine() usage altogether. The only reason we'd have to migrate nr_uninterruptible is so that we could use for_each_online_cpu() loops in favour of for_each_possible_cpu() loops, however since nr_uninterruptible() is the only such loop and its using possible lets not bother at all. The problem Rabik sees is (probably) caused by the fact that by migrating nr_uninterruptible we screw rq->calc_load_active for both rqs involved. So don't bother with fancy migration schemes (meaning we now have to keep using for_each_possible_cpu()) and instead fold any nr_active delta after we migrate all tasks away to make sure we don't have any skewed nr_active accounting. Reported-by: Rakib Mullick Reported-by: Paul E. McKenney Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1345454817.23018.27.camel@twins Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 31 ++++++++++--------------------- 1 file changed, 10 insertions(+), 21 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index fbf1fd098dc6..207a81c769d4 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5304,27 +5304,17 @@ void idle_task_exit(void) } /* - * While a dead CPU has no uninterruptible tasks queued at this point, - * it might still have a nonzero ->nr_uninterruptible counter, because - * for performance reasons the counter is not stricly tracking tasks to - * their home CPUs. So we just add the counter to another CPU's counter, - * to keep the global sum constant after CPU-down: - */ -static void migrate_nr_uninterruptible(struct rq *rq_src) -{ - struct rq *rq_dest = cpu_rq(cpumask_any(cpu_active_mask)); - - rq_dest->nr_uninterruptible += rq_src->nr_uninterruptible; - rq_src->nr_uninterruptible = 0; -} - -/* - * remove the tasks which were accounted by rq from calc_load_tasks. + * Since this CPU is going 'away' for a while, fold any nr_active delta + * we might have. Assumes we're called after migrate_tasks() so that the + * nr_active count is stable. + * + * Also see the comment "Global load-average calculations". */ -static void calc_global_load_remove(struct rq *rq) +static void calc_load_migrate(struct rq *rq) { - atomic_long_sub(rq->calc_load_active, &calc_load_tasks); - rq->calc_load_active = 0; + long delta = calc_load_fold_active(rq); + if (delta) + atomic_long_add(delta, &calc_load_tasks); } /* @@ -5618,8 +5608,7 @@ migration_call(struct notifier_block *nfb, unsigned long action, void *hcpu) BUG_ON(rq->nr_running != 1); /* the migration thread */ raw_spin_unlock_irqrestore(&rq->lock, flags); - migrate_nr_uninterruptible(rq); - calc_global_load_remove(rq); + calc_load_migrate(rq); break; #endif } -- cgit v1.2.3 From 749c8814f08f12baa4a9c2812a7c6ede7d69507d Mon Sep 17 00:00:00 2001 From: Charles Wang Date: Mon, 20 Aug 2012 16:02:33 +0800 Subject: sched: Add missing call to calc_load_exit_idle() Azat Khuzhin reported high loadavg in Linux v3.6 After checking the upstream scheduler code, I found Peter's commit: 5167e8d5417b sched/nohz: Rewrite and fix load-avg computation -- again not fully applied, missing the call to calc_load_exit_idle(). After that idle exit in sampling window will always be calculated to non-idle, and the load will be higher than normal. This patch adds the missing call to calc_load_exit_idle(). Signed-off-by: Charles Wang Cc: stable@kernel.org Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1345449754-27130-1-git-send-email-muming.wq@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 024540f97f74..3a9e5d5c1091 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -573,6 +573,7 @@ static void tick_nohz_restart_sched_tick(struct tick_sched *ts, ktime_t now) tick_do_update_jiffies64(now); update_cpu_load_nohz(); + calc_load_exit_idle(); touch_softlockup_watchdog(); /* * Cancel the scheduled timer and restore the tick -- cgit v1.2.3 From a4c96ae319b8047f62dedbe1eac79e321c185749 Mon Sep 17 00:00:00 2001 From: Peter Boonstoppel Date: Thu, 9 Aug 2012 15:34:47 -0700 Subject: sched: Unthrottle rt runqueues in __disable_runtime() migrate_tasks() uses _pick_next_task_rt() to get tasks from the real-time runqueues to be migrated. When rt_rq is throttled _pick_next_task_rt() won't return anything, in which case migrate_tasks() can't move all threads over and gets stuck in an infinite loop. Instead unthrottle rt runqueues before migrating tasks. Additionally: move unthrottle_offline_cfs_rqs() to rq_offline_fair() Signed-off-by: Peter Boonstoppel Signed-off-by: Peter Zijlstra Cc: Paul Turner Link: http://lkml.kernel.org/r/5FBF8E85CA34454794F0F7ECBA79798F379D3648B7@HQMAIL04.nvidia.com Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 3 --- kernel/sched/fair.c | 7 +++++-- kernel/sched/rt.c | 1 + kernel/sched/sched.h | 1 - 4 files changed, 6 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 207a81c769d4..a4ea245f3d85 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5342,9 +5342,6 @@ static void migrate_tasks(unsigned int dead_cpu) */ rq->stop = NULL; - /* Ensure any throttled groups are reachable by pick_next_task */ - unthrottle_offline_cfs_rqs(rq); - for ( ; ; ) { /* * There's this thread running, bail when that's the only diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index c219bf8d704c..86ad83c45dae 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -2052,7 +2052,7 @@ static void destroy_cfs_bandwidth(struct cfs_bandwidth *cfs_b) hrtimer_cancel(&cfs_b->slack_timer); } -void unthrottle_offline_cfs_rqs(struct rq *rq) +static void unthrottle_offline_cfs_rqs(struct rq *rq) { struct cfs_rq *cfs_rq; @@ -2106,7 +2106,7 @@ static inline struct cfs_bandwidth *tg_cfs_bandwidth(struct task_group *tg) return NULL; } static inline void destroy_cfs_bandwidth(struct cfs_bandwidth *cfs_b) {} -void unthrottle_offline_cfs_rqs(struct rq *rq) {} +static inline void unthrottle_offline_cfs_rqs(struct rq *rq) {} #endif /* CONFIG_CFS_BANDWIDTH */ @@ -4956,6 +4956,9 @@ static void rq_online_fair(struct rq *rq) static void rq_offline_fair(struct rq *rq) { update_sysctl(); + + /* Ensure any throttled groups are reachable by pick_next_task */ + unthrottle_offline_cfs_rqs(rq); } #endif /* CONFIG_SMP */ diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c index 944cb68420e9..e0b7ba9c040f 100644 --- a/kernel/sched/rt.c +++ b/kernel/sched/rt.c @@ -691,6 +691,7 @@ balanced: * runtime - in which case borrowing doesn't make sense. */ rt_rq->rt_runtime = RUNTIME_INF; + rt_rq->rt_throttled = 0; raw_spin_unlock(&rt_rq->rt_runtime_lock); raw_spin_unlock(&rt_b->rt_runtime_lock); } diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index f6714d009e77..0848fa36c383 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -1144,7 +1144,6 @@ extern void print_rt_stats(struct seq_file *m, int cpu); extern void init_cfs_rq(struct cfs_rq *cfs_rq); extern void init_rt_rq(struct rt_rq *rt_rq, struct rq *rq); -extern void unthrottle_offline_cfs_rqs(struct rq *rq); extern void account_cfs_bandwidth_used(int enabled, int was_enabled); -- cgit v1.2.3 From 9450d57eab5cad36774c297da123062744472588 Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Sat, 18 Aug 2012 17:45:08 -0700 Subject: sched: Fix kernel-doc warnings in kernel/sched/fair.c Fix two kernel-doc warnings in kernel/sched/fair.c: Warning(kernel/sched/fair.c:3660): Excess function parameter 'cpus' description in 'update_sg_lb_stats' Warning(kernel/sched/fair.c:3806): Excess function parameter 'cpus' description in 'update_sd_lb_stats' Signed-off-by: Randy Dunlap Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/50303714.3090204@xenotime.net Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 86ad83c45dae..42d9df6a5ca4 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -3658,7 +3658,6 @@ 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. - * @cpus: Set of cpus considered for load balancing. * @balance: Should we balance. * @sgs: variable to hold the statistics for this group. */ @@ -3805,7 +3804,6 @@ static bool update_sd_pick_busiest(struct lb_env *env, /** * update_sd_lb_stats - Update sched_domain's statistics for load balancing. * @env: The load balancing environment. - * @cpus: Set of cpus considered for load balancing. * @balance: Should we balance. * @sds: variable to hold the statistics for this sched_domain. */ -- cgit v1.2.3 From a6fa941d94b411bbd2b6421ffbde6db3c93e65ab Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 20 Aug 2012 14:59:25 +0100 Subject: perf_event: Switch to internal refcount, fix race with close() Don't mess with file refcounts (or keep a reference to file, for that matter) in perf_event. Use explicit refcount of its own instead. Deal with the race between the final reference to event going away and new children getting created for it by use of atomic_long_inc_not_zero() in inherit_event(); just have the latter free what it had allocated and return NULL, that works out just fine (children of siblings of something doomed are created as singletons, same as if the child of leader had been created and immediately killed). Signed-off-by: Al Viro Cc: stable@kernel.org Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/20120820135925.GG23464@ZenIV.linux.org.uk Signed-off-by: Ingo Molnar --- kernel/events/core.c | 62 ++++++++++++++++++++++++++++------------------------ 1 file changed, 33 insertions(+), 29 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index b7935fcec7d9..efef4282b8e8 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -2935,12 +2935,12 @@ EXPORT_SYMBOL_GPL(perf_event_release_kernel); /* * Called when the last reference to the file is gone. */ -static int perf_release(struct inode *inode, struct file *file) +static void put_event(struct perf_event *event) { - struct perf_event *event = file->private_data; struct task_struct *owner; - file->private_data = NULL; + if (!atomic_long_dec_and_test(&event->refcount)) + return; rcu_read_lock(); owner = ACCESS_ONCE(event->owner); @@ -2975,7 +2975,13 @@ static int perf_release(struct inode *inode, struct file *file) put_task_struct(owner); } - return perf_event_release_kernel(event); + perf_event_release_kernel(event); +} + +static int perf_release(struct inode *inode, struct file *file) +{ + put_event(file->private_data); + return 0; } u64 perf_event_read_value(struct perf_event *event, u64 *enabled, u64 *running) @@ -3227,7 +3233,7 @@ unlock: static const struct file_operations perf_fops; -static struct perf_event *perf_fget_light(int fd, int *fput_needed) +static struct file *perf_fget_light(int fd, int *fput_needed) { struct file *file; @@ -3241,7 +3247,7 @@ static struct perf_event *perf_fget_light(int fd, int *fput_needed) return ERR_PTR(-EBADF); } - return file->private_data; + return file; } static int perf_event_set_output(struct perf_event *event, @@ -3273,19 +3279,21 @@ static long perf_ioctl(struct file *file, unsigned int cmd, unsigned long arg) case PERF_EVENT_IOC_SET_OUTPUT: { + struct file *output_file = NULL; struct perf_event *output_event = NULL; int fput_needed = 0; int ret; if (arg != -1) { - output_event = perf_fget_light(arg, &fput_needed); - if (IS_ERR(output_event)) - return PTR_ERR(output_event); + output_file = perf_fget_light(arg, &fput_needed); + if (IS_ERR(output_file)) + return PTR_ERR(output_file); + output_event = output_file->private_data; } ret = perf_event_set_output(event, output_event); if (output_event) - fput_light(output_event->filp, fput_needed); + fput_light(output_file, fput_needed); return ret; } @@ -5950,6 +5958,7 @@ perf_event_alloc(struct perf_event_attr *attr, int cpu, mutex_init(&event->mmap_mutex); + atomic_long_set(&event->refcount, 1); event->cpu = cpu; event->attr = *attr; event->group_leader = group_leader; @@ -6260,12 +6269,12 @@ SYSCALL_DEFINE5(perf_event_open, return event_fd; if (group_fd != -1) { - group_leader = perf_fget_light(group_fd, &fput_needed); - if (IS_ERR(group_leader)) { - err = PTR_ERR(group_leader); + group_file = perf_fget_light(group_fd, &fput_needed); + if (IS_ERR(group_file)) { + err = PTR_ERR(group_file); goto err_fd; } - group_file = group_leader->filp; + group_leader = group_file->private_data; if (flags & PERF_FLAG_FD_OUTPUT) output_event = group_leader; if (flags & PERF_FLAG_FD_NO_GROUP) @@ -6402,7 +6411,6 @@ SYSCALL_DEFINE5(perf_event_open, put_ctx(gctx); } - event->filp = event_file; WARN_ON_ONCE(ctx->parent_ctx); mutex_lock(&ctx->mutex); @@ -6496,7 +6504,6 @@ perf_event_create_kernel_counter(struct perf_event_attr *attr, int cpu, goto err_free; } - event->filp = NULL; WARN_ON_ONCE(ctx->parent_ctx); mutex_lock(&ctx->mutex); perf_install_in_context(ctx, event, cpu); @@ -6578,7 +6585,7 @@ static void sync_child_event(struct perf_event *child_event, * Release the parent event, if this was the last * reference to it. */ - fput(parent_event->filp); + put_event(parent_event); } static void @@ -6654,9 +6661,8 @@ static void perf_event_exit_task_context(struct task_struct *child, int ctxn) * * __perf_event_exit_task() * sync_child_event() - * fput(parent_event->filp) - * perf_release() - * mutex_lock(&ctx->mutex) + * put_event() + * mutex_lock(&ctx->mutex) * * But since its the parent context it won't be the same instance. */ @@ -6724,7 +6730,7 @@ static void perf_free_event(struct perf_event *event, list_del_init(&event->child_list); mutex_unlock(&parent->child_mutex); - fput(parent->filp); + put_event(parent); perf_group_detach(event); list_del_event(event, ctx); @@ -6804,6 +6810,12 @@ inherit_event(struct perf_event *parent_event, NULL, NULL); if (IS_ERR(child_event)) return child_event; + + if (!atomic_long_inc_not_zero(&parent_event->refcount)) { + free_event(child_event); + return NULL; + } + get_ctx(child_ctx); /* @@ -6844,14 +6856,6 @@ inherit_event(struct perf_event *parent_event, add_event_to_ctx(child_event, child_ctx); raw_spin_unlock_irqrestore(&child_ctx->lock, flags); - /* - * Get a reference to the parent filp - we will fput it - * when the child event exits. This is safe to do because - * we are in the parent and we know that the filp still - * exists and has a nonzero count: - */ - atomic_long_inc(&parent_event->filp->f_count); - /* * Link this into the parent event's child list */ -- cgit v1.2.3 From 500ad2d8b01390c98bc6dce068bccfa9534b8212 Mon Sep 17 00:00:00 2001 From: "K.Prasad" Date: Thu, 2 Aug 2012 13:46:35 +0530 Subject: perf/hwpb: Invoke __perf_event_disable() if interrupts are already disabled While debugging a warning message on PowerPC while using hardware breakpoints, it was discovered that when perf_event_disable is invoked through hw_breakpoint_handler function with interrupts disabled, a subsequent IPI in the code path would trigger a WARN_ON_ONCE message in smp_call_function_single function. This patch calls __perf_event_disable() when interrupts are already disabled, instead of perf_event_disable(). Reported-by: Edjunior Barbosa Machado Signed-off-by: K.Prasad [naveen.n.rao@linux.vnet.ibm.com: v3: Check to make sure we target current task] Signed-off-by: Naveen N. Rao Acked-by: Frederic Weisbecker Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/20120802081635.5811.17737.stgit@localhost.localdomain [ Fixed build error on MIPS. ] Signed-off-by: Ingo Molnar --- kernel/events/core.c | 2 +- kernel/events/hw_breakpoint.c | 11 ++++++++++- 2 files changed, 11 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index efef4282b8e8..7fee567153f0 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -1253,7 +1253,7 @@ retry: /* * Cross CPU call to disable a performance event */ -static int __perf_event_disable(void *info) +int __perf_event_disable(void *info) { struct perf_event *event = info; struct perf_event_context *ctx = event->ctx; diff --git a/kernel/events/hw_breakpoint.c b/kernel/events/hw_breakpoint.c index bb38c4d3ee12..9a7b487c6fe2 100644 --- a/kernel/events/hw_breakpoint.c +++ b/kernel/events/hw_breakpoint.c @@ -453,7 +453,16 @@ int modify_user_hw_breakpoint(struct perf_event *bp, struct perf_event_attr *att int old_type = bp->attr.bp_type; int err = 0; - perf_event_disable(bp); + /* + * modify_user_hw_breakpoint can be invoked with IRQs disabled and hence it + * will not be possible to raise IPIs that invoke __perf_event_disable. + * So call the function directly after making sure we are targeting the + * current task. + */ + if (irqs_disabled() && bp->ctx && bp->ctx->task == current) + __perf_event_disable(bp); + else + perf_event_disable(bp); bp->attr.bp_addr = attr->bp_addr; bp->attr.bp_type = attr->bp_type; -- cgit v1.2.3 From 96e65306b81351b656835c15931d1d237b252f27 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Sun, 2 Sep 2012 00:28:19 +0800 Subject: workqueue: UNBOUND -> REBIND morphing in rebind_workers() should be atomic The compiler may compile the following code into TWO write/modify instructions. worker->flags &= ~WORKER_UNBOUND; worker->flags |= WORKER_REBIND; so the other CPU may temporarily see worker->flags which doesn't have either WORKER_UNBOUND or WORKER_REBIND set and perform local wakeup prematurely. Fix it by using single explicit assignment via ACCESS_ONCE(). Because idle workers have another WORKER_NOT_RUNNING flag, this bug doesn't exist for them; however, update it to use the same pattern for consistency. tj: Applied the change to idle workers too and updated comments and patch description a bit. Signed-off-by: Lai Jiangshan Signed-off-by: Tejun Heo Cc: stable@vger.kernel.org --- kernel/workqueue.c | 17 +++++++++++------ 1 file changed, 11 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 692d97628a10..c462cd60c374 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -1396,12 +1396,15 @@ retry: /* set REBIND and kick idle ones, we'll wait for these later */ for_each_worker_pool(pool, gcwq) { list_for_each_entry(worker, &pool->idle_list, entry) { + unsigned long worker_flags = worker->flags; + if (worker->flags & WORKER_REBIND) continue; - /* morph UNBOUND to REBIND */ - worker->flags &= ~WORKER_UNBOUND; - worker->flags |= WORKER_REBIND; + /* morph UNBOUND to REBIND atomically */ + worker_flags &= ~WORKER_UNBOUND; + worker_flags |= WORKER_REBIND; + ACCESS_ONCE(worker->flags) = worker_flags; idle_rebind.cnt++; worker->idle_rebind = &idle_rebind; @@ -1434,10 +1437,12 @@ retry: /* rebind busy workers */ for_each_busy_worker(worker, i, pos, gcwq) { struct work_struct *rebind_work = &worker->rebind_work; + unsigned long worker_flags = worker->flags; - /* morph UNBOUND to REBIND */ - worker->flags &= ~WORKER_UNBOUND; - worker->flags |= WORKER_REBIND; + /* morph UNBOUND to REBIND atomically */ + worker_flags &= ~WORKER_UNBOUND; + worker_flags |= WORKER_REBIND; + ACCESS_ONCE(worker->flags) = worker_flags; if (test_and_set_bit(WORK_STRUCT_PENDING_BIT, work_data_bits(rebind_work))) -- cgit v1.2.3 From 90beca5de591e12482a812f23a7f10690962ed4a Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 4 Sep 2012 23:12:33 -0700 Subject: workqueue: move WORKER_REBIND clearing in rebind_workers() to the end of the function This doesn't make any functional difference and is purely to help the next patch to be simpler. Signed-off-by: Tejun Heo Cc: Lai Jiangshan --- kernel/workqueue.c | 26 +++++++++++++------------- 1 file changed, 13 insertions(+), 13 deletions(-) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index c462cd60c374..d79a18d0c42e 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -1422,19 +1422,7 @@ retry: goto retry; } - /* - * All idle workers are rebound and waiting for %WORKER_REBIND to - * be cleared inside idle_worker_rebind(). Clear and release. - * Clearing %WORKER_REBIND from this foreign context is safe - * because these workers are still guaranteed to be idle. - */ - for_each_worker_pool(pool, gcwq) - list_for_each_entry(worker, &pool->idle_list, entry) - worker->flags &= ~WORKER_REBIND; - - wake_up_all(&gcwq->rebind_hold); - - /* rebind busy workers */ + /* all idle workers are rebound, rebind busy workers */ for_each_busy_worker(worker, i, pos, gcwq) { struct work_struct *rebind_work = &worker->rebind_work; unsigned long worker_flags = worker->flags; @@ -1454,6 +1442,18 @@ retry: worker->scheduled.next, work_color_to_flags(WORK_NO_COLOR)); } + + /* + * All idle workers are rebound and waiting for %WORKER_REBIND to + * be cleared inside idle_worker_rebind(). Clear and release. + * Clearing %WORKER_REBIND from this foreign context is safe + * because these workers are still guaranteed to be idle. + */ + for_each_worker_pool(pool, gcwq) + list_for_each_entry(worker, &pool->idle_list, entry) + worker->flags &= ~WORKER_REBIND; + + wake_up_all(&gcwq->rebind_hold); } static struct worker *alloc_worker(void) -- cgit v1.2.3 From ec58815ab0409a921a7c9744eb4ca44866b14d71 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 4 Sep 2012 23:16:32 -0700 Subject: workqueue: fix possible deadlock in idle worker rebinding Currently, rebind_workers() and idle_worker_rebind() are two-way interlocked. rebind_workers() waits for idle workers to finish rebinding and rebound idle workers wait for rebind_workers() to finish rebinding busy workers before proceeding. Unfortunately, this isn't enough. The second wait from idle workers is implemented as follows. wait_event(gcwq->rebind_hold, !(worker->flags & WORKER_REBIND)); rebind_workers() clears WORKER_REBIND, wakes up the idle workers and then returns. If CPU hotplug cycle happens again before one of the idle workers finishes the above wait_event(), rebind_workers() will repeat the first part of the handshake - set WORKER_REBIND again and wait for the idle worker to finish rebinding - and this leads to deadlock because the idle worker would be waiting for WORKER_REBIND to clear. This is fixed by adding another interlocking step at the end - rebind_workers() now waits for all the idle workers to finish the above WORKER_REBIND wait before returning. This ensures that all rebinding steps are complete on all idle workers before the next hotplug cycle can happen. This problem was diagnosed by Lai Jiangshan who also posted a patch to fix the issue, upon which this patch is based. This is the minimal fix and further patches are scheduled for the next merge window to simplify the CPU hotplug path. Signed-off-by: Tejun Heo Original-patch-by: Lai Jiangshan LKML-Reference: <1346516916-1991-3-git-send-email-laijs@cn.fujitsu.com> --- kernel/workqueue.c | 29 +++++++++++++++++++++++++++-- 1 file changed, 27 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index d79a18d0c42e..dc7b8458e275 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -1326,6 +1326,15 @@ static void idle_worker_rebind(struct worker *worker) /* we did our part, wait for rebind_workers() to finish up */ wait_event(gcwq->rebind_hold, !(worker->flags & WORKER_REBIND)); + + /* + * rebind_workers() shouldn't finish until all workers passed the + * above WORKER_REBIND wait. Tell it when done. + */ + spin_lock_irq(&worker->pool->gcwq->lock); + if (!--worker->idle_rebind->cnt) + complete(&worker->idle_rebind->done); + spin_unlock_irq(&worker->pool->gcwq->lock); } /* @@ -1448,12 +1457,28 @@ retry: * be cleared inside idle_worker_rebind(). Clear and release. * Clearing %WORKER_REBIND from this foreign context is safe * because these workers are still guaranteed to be idle. + * + * We need to make sure all idle workers passed WORKER_REBIND wait + * in idle_worker_rebind() before returning; otherwise, workers can + * get stuck at the wait if hotplug cycle repeats. */ - for_each_worker_pool(pool, gcwq) - list_for_each_entry(worker, &pool->idle_list, entry) + idle_rebind.cnt = 1; + INIT_COMPLETION(idle_rebind.done); + + for_each_worker_pool(pool, gcwq) { + list_for_each_entry(worker, &pool->idle_list, entry) { worker->flags &= ~WORKER_REBIND; + idle_rebind.cnt++; + } + } wake_up_all(&gcwq->rebind_hold); + + if (--idle_rebind.cnt) { + spin_unlock_irq(&gcwq->lock); + wait_for_completion(&idle_rebind.done); + spin_lock_irq(&gcwq->lock); + } } static struct worker *alloc_worker(void) -- cgit v1.2.3 From 552a37e9360a293cd20e7f8ff1fb326a244c5f1e Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Mon, 10 Sep 2012 10:03:33 -0700 Subject: workqueue: restore POOL_MANAGING_WORKERS This patch restores POOL_MANAGING_WORKERS which was replaced by pool->manager_mutex by 6037315269 "workqueue: use mutex for global_cwq manager exclusion". There's a subtle idle worker depletion bug across CPU hotplug events and we need to distinguish an actual manager and CPU hotplug preventing management. POOL_MANAGING_WORKERS will be used for the former and manager_mutex the later. This patch just lays POOL_MANAGING_WORKERS on top of the existing manager_mutex and doesn't introduce any synchronization changes. The next patch will update it. Note that this patch fixes a non-critical anomaly where too_many_workers() may return %true spuriously while CPU hotplug is in progress. While the issue could schedule idle timer spuriously, it didn't trigger any actual misbehavior. tj: Rewrote patch description. Signed-off-by: Lai Jiangshan Signed-off-by: Tejun Heo --- kernel/workqueue.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index dc7b8458e275..383548ed0b54 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -66,6 +66,7 @@ enum { /* pool flags */ POOL_MANAGE_WORKERS = 1 << 0, /* need to manage workers */ + POOL_MANAGING_WORKERS = 1 << 1, /* managing workers */ /* worker flags */ WORKER_STARTED = 1 << 0, /* started */ @@ -652,7 +653,7 @@ static bool need_to_manage_workers(struct worker_pool *pool) /* Do we have too many workers and should some go away? */ static bool too_many_workers(struct worker_pool *pool) { - bool managing = mutex_is_locked(&pool->manager_mutex); + bool managing = pool->flags & POOL_MANAGING_WORKERS; int nr_idle = pool->nr_idle + managing; /* manager is considered idle */ int nr_busy = pool->nr_workers - nr_idle; @@ -1827,6 +1828,7 @@ static bool manage_workers(struct worker *worker) if (!mutex_trylock(&pool->manager_mutex)) return ret; + pool->flags |= POOL_MANAGING_WORKERS; pool->flags &= ~POOL_MANAGE_WORKERS; /* @@ -1836,6 +1838,7 @@ static bool manage_workers(struct worker *worker) ret |= maybe_destroy_workers(pool); ret |= maybe_create_worker(pool); + pool->flags &= ~POOL_MANAGING_WORKERS; mutex_unlock(&pool->manager_mutex); return ret; } -- cgit v1.2.3 From ee378aa49b594da9bda6a2c768cc5b2ad585f911 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Mon, 10 Sep 2012 10:03:44 -0700 Subject: workqueue: fix possible idle worker depletion across CPU hotplug To simplify both normal and CPU hotplug paths, worker management is prevented while CPU hoplug is in progress. This is achieved by CPU hotplug holding the same exclusion mechanism used by workers to ensure there's only one manager per pool. If someone else seems to be performing the manager role, workers proceed to execute work items. CPU hotplug using the same mechanism can lead to idle worker depletion because all workers could proceed to execute work items while CPU hotplug is in progress and CPU hotplug itself wouldn't actually perform the worker management duty - it doesn't guarantee that there's an idle worker left when it releases management. This idle worker depletion, under extreme circumstances, can break forward-progress guarantee and thus lead to deadlock. This patch fixes the bug by using separate mechanisms for manager exclusion among workers and hotplug exclusion. For manager exclusion, POOL_MANAGING_WORKERS which was restored by the previous patch is used. pool->manager_mutex is now only used for exclusion between the elected manager and CPU hotplug. The elected manager won't proceed without holding pool->manager_mutex. This ensures that the worker which won the manager position can't skip managing while CPU hotplug is in progress. It will block on manager_mutex and perform management after CPU hotplug is complete. Note that hotplug may happen while waiting for manager_mutex. A manager isn't either on idle or busy list and thus the hoplug code can't unbind/rebind it. Make the manager handle its own un/rebinding. tj: Updated comment and description. Signed-off-by: Lai Jiangshan Signed-off-by: Tejun Heo --- kernel/workqueue.c | 37 ++++++++++++++++++++++++++++++++++++- 1 file changed, 36 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 383548ed0b54..1e1373bcb3e3 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -1825,10 +1825,45 @@ static bool manage_workers(struct worker *worker) struct worker_pool *pool = worker->pool; bool ret = false; - if (!mutex_trylock(&pool->manager_mutex)) + if (pool->flags & POOL_MANAGING_WORKERS) return ret; pool->flags |= POOL_MANAGING_WORKERS; + + /* + * To simplify both worker management and CPU hotplug, hold off + * management while hotplug is in progress. CPU hotplug path can't + * grab %POOL_MANAGING_WORKERS to achieve this because that can + * lead to idle worker depletion (all become busy thinking someone + * else is managing) which in turn can result in deadlock under + * extreme circumstances. Use @pool->manager_mutex to synchronize + * manager against CPU hotplug. + * + * manager_mutex would always be free unless CPU hotplug is in + * progress. trylock first without dropping @gcwq->lock. + */ + if (unlikely(!mutex_trylock(&pool->manager_mutex))) { + spin_unlock_irq(&pool->gcwq->lock); + mutex_lock(&pool->manager_mutex); + /* + * CPU hotplug could have happened while we were waiting + * for manager_mutex. Hotplug itself can't handle us + * because manager isn't either on idle or busy list, and + * @gcwq's state and ours could have deviated. + * + * As hotplug is now excluded via manager_mutex, we can + * simply try to bind. It will succeed or fail depending + * on @gcwq's current state. Try it and adjust + * %WORKER_UNBOUND accordingly. + */ + if (worker_maybe_bind_and_lock(worker)) + worker->flags &= ~WORKER_UNBOUND; + else + worker->flags |= WORKER_UNBOUND; + + ret = true; + } + pool->flags &= ~POOL_MANAGE_WORKERS; /* -- cgit v1.2.3 From 37407ea7f93864c2cfc03edf8f37872ec539ea2b Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sun, 16 Sep 2012 12:29:43 -0700 Subject: Revert "sched: Improve scalability via 'CPU buddies', which withstand random perturbations" This reverts commit 970e178985cadbca660feb02f4d2ee3a09f7fdda. Nikolay Ulyanitsky reported thatthe 3.6-rc5 kernel has a 15-20% performance drop on PostgreSQL 9.2 on his machine (running "pgbench"). Borislav Petkov was able to reproduce this, and bisected it to this commit 970e178985ca ("sched: Improve scalability via 'CPU buddies' ...") apparently because the new single-idle-buddy model simply doesn't find idle CPU's to reschedule on aggressively enough. Mike Galbraith suspects that it is likely due to the user-mode spinlocks in PostgreSQL not reacting well to preemption, but we don't really know the details - I'll just revert the commit for now. There are hopefully other approaches to improve scheduler scalability without it causing these kinds of downsides. Reported-by: Nikolay Ulyanitsky Bisected-by: Borislav Petkov Acked-by: Mike Galbraith Cc: Andrew Morton Cc: Thomas Gleixner Cc: Ingo Molnar Signed-off-by: Linus Torvalds --- kernel/sched/core.c | 39 +-------------------------------------- kernel/sched/fair.c | 28 +++++++++++++++++++++------- 2 files changed, 22 insertions(+), 45 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index a4ea245f3d85..649c9f876cb1 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -6014,11 +6014,6 @@ static void destroy_sched_domains(struct sched_domain *sd, int cpu) * SD_SHARE_PKG_RESOURCE set (Last Level Cache Domain) for this * allows us to avoid some pointer chasing select_idle_sibling(). * - * Iterate domains and sched_groups downward, assigning CPUs to be - * select_idle_sibling() hw buddy. Cross-wiring hw makes bouncing - * due to random perturbation self canceling, ie sw buddies pull - * their counterpart to their CPU's hw counterpart. - * * Also keep a unique ID per domain (we use the first cpu number in * the cpumask of the domain), this allows us to quickly tell if * two cpus are in the same cache domain, see cpus_share_cache(). @@ -6032,40 +6027,8 @@ static void update_top_cache_domain(int cpu) int id = cpu; sd = highest_flag_domain(cpu, SD_SHARE_PKG_RESOURCES); - if (sd) { - struct sched_domain *tmp = sd; - struct sched_group *sg, *prev; - bool right; - - /* - * Traverse to first CPU in group, and count hops - * to cpu from there, switching direction on each - * hop, never ever pointing the last CPU rightward. - */ - do { - id = cpumask_first(sched_domain_span(tmp)); - prev = sg = tmp->groups; - right = 1; - - while (cpumask_first(sched_group_cpus(sg)) != id) - sg = sg->next; - - while (!cpumask_test_cpu(cpu, sched_group_cpus(sg))) { - prev = sg; - sg = sg->next; - right = !right; - } - - /* A CPU went down, never point back to domain start. */ - if (right && cpumask_first(sched_group_cpus(sg->next)) == id) - right = false; - - sg = right ? sg->next : prev; - tmp->idle_buddy = cpumask_first(sched_group_cpus(sg)); - } while ((tmp = tmp->child)); - + if (sd) id = cpumask_first(sched_domain_span(sd)); - } rcu_assign_pointer(per_cpu(sd_llc, cpu), sd); per_cpu(sd_llc_id, cpu) = id; diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 42d9df6a5ca4..96e2b18b6283 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -2637,6 +2637,8 @@ static int select_idle_sibling(struct task_struct *p, int target) int cpu = smp_processor_id(); int prev_cpu = task_cpu(p); struct sched_domain *sd; + struct sched_group *sg; + int i; /* * If the task is going to be woken-up on this cpu and if it is @@ -2653,17 +2655,29 @@ static int select_idle_sibling(struct task_struct *p, int target) return prev_cpu; /* - * Otherwise, check assigned siblings to find an elegible idle cpu. + * Otherwise, iterate the domains and find an elegible idle cpu. */ sd = rcu_dereference(per_cpu(sd_llc, target)); - for_each_lower_domain(sd) { - if (!cpumask_test_cpu(sd->idle_buddy, tsk_cpus_allowed(p))) - continue; - if (idle_cpu(sd->idle_buddy)) - return sd->idle_buddy; - } + sg = sd->groups; + do { + if (!cpumask_intersects(sched_group_cpus(sg), + tsk_cpus_allowed(p))) + goto next; + for_each_cpu(i, sched_group_cpus(sg)) { + if (!idle_cpu(i)) + goto next; + } + + target = cpumask_first_and(sched_group_cpus(sg), + tsk_cpus_allowed(p)); + goto done; +next: + sg = sg->next; + } while (sg != sd->groups); + } +done: return target; } -- cgit v1.2.3