From cf417141cbb3a4ceb5cca15b2c1f099bd0a6603c Mon Sep 17 00:00:00 2001 From: Max Krasnyansky Date: Mon, 11 Aug 2008 14:33:53 -0700 Subject: sched, cpuset: rework sched domains and CPU hotplug handling (v4) This is an updated version of my previous cpuset patch on top of the latest mainline git. The patch fixes CPU hotplug handling issues in the current cpusets code. Namely circular locking in rebuild_sched_domains() and unsafe access to the cpu_online_map in the cpuset cpu hotplug handler. This version includes changes suggested by Paul Jackson (naming, comments, style, etc). I also got rid of the separate workqueue thread because it is now safe to call get_online_cpus() from workqueue callbacks. Here are some more details: rebuild_sched_domains() is the only way to rebuild sched domains correctly based on the current cpuset settings. What this means is that we need to be able to call it from different contexts, like cpu hotplug for example. Also latest scheduler code in -tip now calls rebuild_sched_domains() directly from functions like arch_reinit_sched_domains(). In order to support that properly we need to rework cpuset locking rules to avoid circular dependencies, which is what this patch does. New lock nesting rules are explained in the comments. We can now safely call rebuild_sched_domains() from virtually any context. The only requirement is that it needs to be called under get_online_cpus(). This allows cpu hotplug handlers and the scheduler to call rebuild_sched_domains() directly. The rest of the cpuset code now offloads sched domains rebuilds to a workqueue (async_rebuild_sched_domains()). This version of the patch addresses comments from the previous review. I fixed all miss-formated comments and trailing spaces. I also factored out the code that builds domain masks and split up CPU and memory hotplug handling. This was needed to simplify locking, to avoid unsafe access to the cpu_online_map from mem hotplug handler, and in general to make things cleaner. The patch passes moderate testing (building kernel with -j 16, creating & removing domains and bringing cpus off/online at the same time) on the quad-core2 based machine. It passes lockdep checks, even with preemptable RCU enabled. This time I also tested in with suspend/resume path and everything is working as expected. Signed-off-by: Max Krasnyansky Acked-by: Paul Jackson Cc: menage@google.com Cc: a.p.zijlstra@chello.nl Cc: vegard.nossum@gmail.com Signed-off-by: Ingo Molnar --- kernel/cpuset.c | 312 +++++++++++++++++++++++++++++++++----------------------- 1 file changed, 182 insertions(+), 130 deletions(-) (limited to 'kernel') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index d5ab79cf516d..f227bc172690 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -14,6 +14,8 @@ * 2003-10-22 Updates by Stephen Hemminger. * 2004 May-July Rework by Paul Jackson. * 2006 Rework by Paul Menage to use generic cgroups + * 2008 Rework of the scheduler domains and CPU hotplug handling + * by Max Krasnyansky * * This file is subject to the terms and conditions of the GNU General Public * License. See the file COPYING in the main directory of the Linux @@ -236,9 +238,11 @@ static struct cpuset top_cpuset = { static DEFINE_MUTEX(callback_mutex); -/* This is ugly, but preserves the userspace API for existing cpuset +/* + * This is ugly, but preserves the userspace API for existing cpuset * users. If someone tries to mount the "cpuset" filesystem, we - * silently switch it to mount "cgroup" instead */ + * silently switch it to mount "cgroup" instead + */ static int cpuset_get_sb(struct file_system_type *fs_type, int flags, const char *unused_dev_name, void *data, struct vfsmount *mnt) @@ -473,10 +477,9 @@ static int validate_change(const struct cpuset *cur, const struct cpuset *trial) } /* - * Helper routine for rebuild_sched_domains(). + * Helper routine for generate_sched_domains(). * Do cpusets a, b have overlapping cpus_allowed masks? */ - static int cpusets_overlap(struct cpuset *a, struct cpuset *b) { return cpus_intersects(a->cpus_allowed, b->cpus_allowed); @@ -518,26 +521,15 @@ update_domain_attr_tree(struct sched_domain_attr *dattr, struct cpuset *c) } /* - * rebuild_sched_domains() - * - * This routine will be called to rebuild the scheduler's dynamic - * sched domains: - * - if the flag 'sched_load_balance' of any cpuset with non-empty - * 'cpus' changes, - * - or if the 'cpus' allowed changes in any cpuset which has that - * flag enabled, - * - or if the 'sched_relax_domain_level' of any cpuset which has - * that flag enabled and with non-empty 'cpus' changes, - * - or if any cpuset with non-empty 'cpus' is removed, - * - or if a cpu gets offlined. - * - * This routine builds a partial partition of the systems CPUs - * (the set of non-overlappping cpumask_t's in the array 'part' - * below), and passes that partial partition to the kernel/sched.c - * partition_sched_domains() routine, which will rebuild the - * schedulers load balancing domains (sched domains) as specified - * by that partial partition. A 'partial partition' is a set of - * non-overlapping subsets whose union is a subset of that set. + * generate_sched_domains() + * + * 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 + * partition_sched_domains() routine, which will rebuild the scheduler's + * load balancing domains (sched domains) as specified by that partial + * partition. * * See "What is sched_load_balance" in Documentation/cpusets.txt * for a background explanation of this. @@ -547,13 +539,7 @@ update_domain_attr_tree(struct sched_domain_attr *dattr, struct cpuset *c) * domains when operating in the severe memory shortage situations * that could cause allocation failures below. * - * Call with cgroup_mutex held. May take callback_mutex during - * call due to the kfifo_alloc() and kmalloc() calls. May nest - * a call to the get_online_cpus()/put_online_cpus() pair. - * Must not be called holding callback_mutex, because we must not - * call get_online_cpus() while holding callback_mutex. Elsewhere - * the kernel nests callback_mutex inside get_online_cpus() calls. - * So the reverse nesting would risk an ABBA deadlock. + * Must be called with cgroup_lock held. * * The three key local variables below are: * q - a linked-list queue of cpuset pointers, used to implement a @@ -588,10 +574,10 @@ update_domain_attr_tree(struct sched_domain_attr *dattr, struct cpuset *c) * element of the partition (one sched domain) to be passed to * partition_sched_domains(). */ - -void rebuild_sched_domains(void) +static int generate_sched_domains(cpumask_t **domains, + struct sched_domain_attr **attributes) { - LIST_HEAD(q); /* queue of cpusets to be scanned*/ + LIST_HEAD(q); /* queue of cpusets to be scanned */ struct cpuset *cp; /* scans q */ struct cpuset **csa; /* array of all cpuset ptrs */ int csn; /* how many cpuset ptrs in csa so far */ @@ -601,23 +587,26 @@ void rebuild_sched_domains(void) int ndoms; /* number of sched domains in result */ int nslot; /* next empty doms[] cpumask_t slot */ - csa = NULL; + ndoms = 0; doms = NULL; dattr = NULL; + csa = NULL; /* Special case for the 99% of systems with one, full, sched domain */ if (is_sched_load_balance(&top_cpuset)) { - ndoms = 1; doms = kmalloc(sizeof(cpumask_t), GFP_KERNEL); if (!doms) - goto rebuild; + goto done; + dattr = kmalloc(sizeof(struct sched_domain_attr), GFP_KERNEL); if (dattr) { *dattr = SD_ATTR_INIT; update_domain_attr_tree(dattr, &top_cpuset); } *doms = top_cpuset.cpus_allowed; - goto rebuild; + + ndoms = 1; + goto done; } csa = kmalloc(number_of_cpusets * sizeof(cp), GFP_KERNEL); @@ -680,61 +669,141 @@ restart: } } - /* Convert to */ + /* + * Now we know how many domains to create. + * Convert to and populate cpu masks. + */ doms = kmalloc(ndoms * sizeof(cpumask_t), GFP_KERNEL); - if (!doms) - goto rebuild; + if (!doms) { + ndoms = 0; + goto done; + } + + /* + * The rest of the code, including the scheduler, can deal with + * dattr==NULL case. No need to abort if alloc fails. + */ dattr = kmalloc(ndoms * sizeof(struct sched_domain_attr), GFP_KERNEL); for (nslot = 0, i = 0; i < csn; i++) { struct cpuset *a = csa[i]; + cpumask_t *dp; int apn = a->pn; - if (apn >= 0) { - cpumask_t *dp = doms + nslot; - - if (nslot == ndoms) { - static int warnings = 10; - if (warnings) { - printk(KERN_WARNING - "rebuild_sched_domains confused:" - " nslot %d, ndoms %d, csn %d, i %d," - " apn %d\n", - nslot, ndoms, csn, i, apn); - warnings--; - } - continue; + if (apn < 0) { + /* Skip completed partitions */ + continue; + } + + dp = doms + nslot; + + if (nslot == ndoms) { + static int warnings = 10; + if (warnings) { + printk(KERN_WARNING + "rebuild_sched_domains confused:" + " nslot %d, ndoms %d, csn %d, i %d," + " apn %d\n", + nslot, ndoms, csn, i, apn); + warnings--; } + continue; + } - cpus_clear(*dp); - if (dattr) - *(dattr + nslot) = SD_ATTR_INIT; - for (j = i; j < csn; j++) { - struct cpuset *b = csa[j]; - - if (apn == b->pn) { - cpus_or(*dp, *dp, b->cpus_allowed); - b->pn = -1; - if (dattr) - update_domain_attr_tree(dattr - + nslot, b); - } + cpus_clear(*dp); + if (dattr) + *(dattr + nslot) = SD_ATTR_INIT; + for (j = i; j < csn; j++) { + struct cpuset *b = csa[j]; + + if (apn == b->pn) { + cpus_or(*dp, *dp, b->cpus_allowed); + if (dattr) + update_domain_attr_tree(dattr + nslot, b); + + /* Done with this partition */ + b->pn = -1; } - nslot++; } + nslot++; } BUG_ON(nslot != ndoms); -rebuild: - /* Have scheduler rebuild sched domains */ +done: + kfree(csa); + + *domains = doms; + *attributes = dattr; + return ndoms; +} + +/* + * Rebuild scheduler domains. + * + * Call with neither cgroup_mutex held nor within get_online_cpus(). + * Takes both cgroup_mutex and get_online_cpus(). + * + * Cannot be directly called from cpuset code handling changes + * to the cpuset pseudo-filesystem, because it cannot be called + * from code that already holds cgroup_mutex. + */ +static void do_rebuild_sched_domains(struct work_struct *unused) +{ + struct sched_domain_attr *attr; + cpumask_t *doms; + int ndoms; + get_online_cpus(); - partition_sched_domains(ndoms, doms, dattr); + + /* Generate domain masks and attrs */ + cgroup_lock(); + ndoms = generate_sched_domains(&doms, &attr); + cgroup_unlock(); + + /* Have scheduler rebuild the domains */ + partition_sched_domains(ndoms, doms, attr); + put_online_cpus(); +} -done: - kfree(csa); - /* Don't kfree(doms) -- partition_sched_domains() does that. */ - /* Don't kfree(dattr) -- partition_sched_domains() does that. */ +static DECLARE_WORK(rebuild_sched_domains_work, do_rebuild_sched_domains); + +/* + * Rebuild scheduler domains, asynchronously via workqueue. + * + * If the flag 'sched_load_balance' of any cpuset with non-empty + * 'cpus' changes, or if the 'cpus' allowed changes in any cpuset + * which has that flag enabled, or if any cpuset with a non-empty + * 'cpus' is removed, then call this routine to rebuild the + * scheduler's dynamic sched domains. + * + * The rebuild_sched_domains() and partition_sched_domains() + * routines must nest cgroup_lock() inside get_online_cpus(), + * but such cpuset changes as these must nest that locking the + * other way, holding cgroup_lock() for much of the code. + * + * So in order to avoid an ABBA deadlock, the cpuset code handling + * these user changes delegates the actual sched domain rebuilding + * to a separate workqueue thread, which ends up processing the + * above do_rebuild_sched_domains() function. + */ +static void async_rebuild_sched_domains(void) +{ + schedule_work(&rebuild_sched_domains_work); +} + +/* + * Accomplishes the same scheduler domain rebuild as the above + * async_rebuild_sched_domains(), however it directly calls the + * rebuild routine synchronously rather than calling it via an + * asynchronous work thread. + * + * This can only be called from code that is not holding + * cgroup_mutex (not nested in a cgroup_lock() call.) + */ +void rebuild_sched_domains(void) +{ + do_rebuild_sched_domains(NULL); } /** @@ -863,7 +932,7 @@ static int update_cpumask(struct cpuset *cs, const char *buf) return retval; if (is_load_balanced) - rebuild_sched_domains(); + async_rebuild_sched_domains(); return 0; } @@ -1090,7 +1159,7 @@ static int update_relax_domain_level(struct cpuset *cs, s64 val) if (val != cs->relax_domain_level) { cs->relax_domain_level = val; if (!cpus_empty(cs->cpus_allowed) && is_sched_load_balance(cs)) - rebuild_sched_domains(); + async_rebuild_sched_domains(); } return 0; @@ -1131,7 +1200,7 @@ static int update_flag(cpuset_flagbits_t bit, struct cpuset *cs, mutex_unlock(&callback_mutex); if (cpus_nonempty && balance_flag_changed) - rebuild_sched_domains(); + async_rebuild_sched_domains(); return 0; } @@ -1492,6 +1561,9 @@ static u64 cpuset_read_u64(struct cgroup *cont, struct cftype *cft) default: BUG(); } + + /* Unreachable but makes gcc happy */ + return 0; } static s64 cpuset_read_s64(struct cgroup *cont, struct cftype *cft) @@ -1504,6 +1576,9 @@ static s64 cpuset_read_s64(struct cgroup *cont, struct cftype *cft) default: BUG(); } + + /* Unrechable but makes gcc happy */ + return 0; } @@ -1692,15 +1767,9 @@ static struct cgroup_subsys_state *cpuset_create( } /* - * Locking note on the strange update_flag() call below: - * * 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(). The get_online_cpus() - * call in rebuild_sched_domains() must not be made while holding - * callback_mutex. Elsewhere the kernel nests callback_mutex inside - * get_online_cpus() calls. So the reverse nesting would risk an - * ABBA deadlock. + * will call async_rebuild_sched_domains(). */ static void cpuset_destroy(struct cgroup_subsys *ss, struct cgroup *cont) @@ -1719,7 +1788,7 @@ static void cpuset_destroy(struct cgroup_subsys *ss, struct cgroup *cont) struct cgroup_subsys cpuset_subsys = { .name = "cpuset", .create = cpuset_create, - .destroy = cpuset_destroy, + .destroy = cpuset_destroy, .can_attach = cpuset_can_attach, .attach = cpuset_attach, .populate = cpuset_populate, @@ -1811,7 +1880,7 @@ static void move_member_tasks_to_cpuset(struct cpuset *from, struct cpuset *to) } /* - * If common_cpu_mem_hotplug_unplug(), below, unplugs any CPUs + * If CPU and/or memory hotplug handlers, below, unplug any CPUs * or memory nodes, we need to walk over the cpuset hierarchy, * removing that CPU or node from all cpusets. If this removes the * last CPU or node from a cpuset, then move the tasks in the empty @@ -1902,35 +1971,6 @@ static void scan_for_empty_cpusets(const struct cpuset *root) } } -/* - * The cpus_allowed and mems_allowed nodemasks in the top_cpuset track - * cpu_online_map and node_states[N_HIGH_MEMORY]. Force the top cpuset to - * track what's online after any CPU or memory node hotplug or unplug event. - * - * Since there are two callers of this routine, one for CPU hotplug - * events and one for memory node hotplug events, we could have coded - * two separate routines here. We code it as a single common routine - * in order to minimize text size. - */ - -static void common_cpu_mem_hotplug_unplug(int rebuild_sd) -{ - cgroup_lock(); - - top_cpuset.cpus_allowed = cpu_online_map; - top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY]; - scan_for_empty_cpusets(&top_cpuset); - - /* - * Scheduler destroys domains on hotplug events. - * Rebuild them based on the current settings. - */ - if (rebuild_sd) - rebuild_sched_domains(); - - cgroup_unlock(); -} - /* * The top_cpuset tracks what CPUs and Memory Nodes are online, * period. This is necessary in order to make cpusets transparent @@ -1939,40 +1979,52 @@ static void common_cpu_mem_hotplug_unplug(int rebuild_sd) * * This routine ensures that top_cpuset.cpus_allowed tracks * cpu_online_map on each CPU hotplug (cpuhp) event. + * + * Called within get_online_cpus(). Needs to call cgroup_lock() + * before calling generate_sched_domains(). */ - -static int cpuset_handle_cpuhp(struct notifier_block *unused_nb, +static int cpuset_track_online_cpus(struct notifier_block *unused_nb, unsigned long phase, void *unused_cpu) { + struct sched_domain_attr *attr; + cpumask_t *doms; + int ndoms; + switch (phase) { - case CPU_UP_CANCELED: - case CPU_UP_CANCELED_FROZEN: - case CPU_DOWN_FAILED: - case CPU_DOWN_FAILED_FROZEN: case CPU_ONLINE: case CPU_ONLINE_FROZEN: case CPU_DEAD: case CPU_DEAD_FROZEN: - common_cpu_mem_hotplug_unplug(1); break; + default: return NOTIFY_DONE; } + cgroup_lock(); + top_cpuset.cpus_allowed = cpu_online_map; + scan_for_empty_cpusets(&top_cpuset); + ndoms = generate_sched_domains(&doms, &attr); + cgroup_unlock(); + + /* Have scheduler rebuild the domains */ + partition_sched_domains(ndoms, doms, attr); + return NOTIFY_OK; } #ifdef CONFIG_MEMORY_HOTPLUG /* * Keep top_cpuset.mems_allowed tracking node_states[N_HIGH_MEMORY]. - * Call this routine anytime after you change - * node_states[N_HIGH_MEMORY]. - * See also the previous routine cpuset_handle_cpuhp(). + * Call this routine anytime after node_states[N_HIGH_MEMORY] changes. + * See also the previous routine cpuset_track_online_cpus(). */ - void cpuset_track_online_nodes(void) { - common_cpu_mem_hotplug_unplug(0); + cgroup_lock(); + top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY]; + scan_for_empty_cpusets(&top_cpuset); + cgroup_unlock(); } #endif @@ -1987,7 +2039,7 @@ void __init cpuset_init_smp(void) top_cpuset.cpus_allowed = cpu_online_map; top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY]; - hotcpu_notifier(cpuset_handle_cpuhp, 0); + hotcpu_notifier(cpuset_track_online_cpus, 0); } /** -- cgit v1.2.3 From 6951b12a0fe7fc64789f2c4d62d2a304e93836a8 Mon Sep 17 00:00:00 2001 From: Dmitry Baryshkov Date: Mon, 18 Aug 2008 04:26:37 +0400 Subject: lockdep: fix spurious 'inconsistent lock state' warning Since f82b217e3513fe3af342c0f3ee1494e86250c21c lockdep can output spurious warnings related to hwirqs due to hardirq_off shrinkage from int to bit-sized flag. Guard it with double negation to fix the warning. Signed-off-by: Dmitry Baryshkov Acked-by: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/lockdep.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/lockdep.c b/kernel/lockdep.c index 77fa776a2da8..3bfb1877a003 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -2582,7 +2582,7 @@ static int __lock_acquire(struct lockdep_map *lock, unsigned int subclass, hlock->trylock = trylock; hlock->read = read; hlock->check = check; - hlock->hardirqs_off = hardirqs_off; + hlock->hardirqs_off = !!hardirqs_off; #ifdef CONFIG_LOCK_STAT hlock->waittime_stamp = 0; hlock->holdtime_stamp = sched_clock(); -- cgit v1.2.3 From 1b04624f93bb1c4f9495b8476d1dd0200af019e2 Mon Sep 17 00:00:00 2001 From: Roland McGrath Date: Tue, 19 Aug 2008 20:37:07 -0700 Subject: tracehook: fix SA_NOCLDWAIT I outwitted myself again in commit 2b2a1ff64afbadac842bbc58c5166962cf4f7664, and broke the SA_NOCLDWAIT behavior so it leaks zombies. This fixes it. Reported-by: Andi Kleen Signed-off-by: Roland McGrath --- kernel/signal.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index c539f60c6f41..e661b01d340f 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -1338,6 +1338,7 @@ int do_notify_parent(struct task_struct *tsk, int sig) struct siginfo info; unsigned long flags; struct sighand_struct *psig; + int ret = sig; BUG_ON(sig == -1); @@ -1402,7 +1403,7 @@ int do_notify_parent(struct task_struct *tsk, int sig) * is implementation-defined: we do (if you don't want * it, just use SIG_IGN instead). */ - tsk->exit_signal = -1; + ret = tsk->exit_signal = -1; if (psig->action[SIGCHLD-1].sa.sa_handler == SIG_IGN) sig = -1; } @@ -1411,7 +1412,7 @@ int do_notify_parent(struct task_struct *tsk, int sig) __wake_up_parent(tsk, tsk->parent); spin_unlock_irqrestore(&psig->siglock, flags); - return sig; + return ret; } static void do_notify_parent_cldstop(struct task_struct *tsk, int why) -- cgit v1.2.3 From 2d70b68d42b5196a48ccb639e3797f097ef5bea3 Mon Sep 17 00:00:00 2001 From: Ken Chen Date: Wed, 20 Aug 2008 14:09:17 -0700 Subject: fix setpriority(PRIO_PGRP) thread iterator breakage When user calls sys_setpriority(PRIO_PGRP ...) on a NPTL style multi-LWP process, only the task leader of the process is affected, all other sibling LWP threads didn't receive the setting. The problem was that the iterator used in sys_setpriority() only iteartes over one task for each process, ignoring all other sibling thread. Introduce a new macro do_each_pid_thread / while_each_pid_thread to walk each thread of a process. Convert 4 call sites in {set/get}priority and ioprio_{set/get}. Signed-off-by: Ken Chen Cc: Oleg Nesterov Cc: Roland McGrath Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Jens Axboe Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sys.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/sys.c b/kernel/sys.c index 3dacb00a7f76..038a7bc0901d 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -169,9 +169,9 @@ asmlinkage long sys_setpriority(int which, int who, int niceval) pgrp = find_vpid(who); else pgrp = task_pgrp(current); - do_each_pid_task(pgrp, PIDTYPE_PGID, p) { + do_each_pid_thread(pgrp, PIDTYPE_PGID, p) { error = set_one_prio(p, niceval, error); - } while_each_pid_task(pgrp, PIDTYPE_PGID, p); + } while_each_pid_thread(pgrp, PIDTYPE_PGID, p); break; case PRIO_USER: user = current->user; @@ -229,11 +229,11 @@ asmlinkage long sys_getpriority(int which, int who) pgrp = find_vpid(who); else pgrp = task_pgrp(current); - do_each_pid_task(pgrp, PIDTYPE_PGID, p) { + do_each_pid_thread(pgrp, PIDTYPE_PGID, p) { niceval = 20 - task_nice(p); if (niceval > retval) retval = niceval; - } while_each_pid_task(pgrp, PIDTYPE_PGID, p); + } while_each_pid_thread(pgrp, PIDTYPE_PGID, p); break; case PRIO_USER: user = current->user; -- cgit v1.2.3 From efc2dead2c82cae31943828f6d977c483942b0eb Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 20 Aug 2008 12:44:55 +0200 Subject: sched: enable LB_BIAS by default Yanmin reported a significant regression on his 16-core machine due to: commit 93b75217df39e6d75889cc6f8050343286aff4a5 Author: Peter Zijlstra Date: Fri Jun 27 13:41:33 2008 +0200 Flip back to the old behaviour. Reported-by: "Zhang, Yanmin" Signed-off-by: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/sched_features.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched_features.h b/kernel/sched_features.h index 862b06bd560a..9353ca78154e 100644 --- a/kernel/sched_features.h +++ b/kernel/sched_features.h @@ -8,6 +8,6 @@ SCHED_FEAT(SYNC_WAKEUPS, 1) SCHED_FEAT(HRTICK, 1) SCHED_FEAT(DOUBLE_TICK, 0) SCHED_FEAT(ASYM_GRAN, 1) -SCHED_FEAT(LB_BIAS, 0) +SCHED_FEAT(LB_BIAS, 1) SCHED_FEAT(LB_WAKEUP_UPDATE, 1) SCHED_FEAT(ASYM_EFF_LOAD, 1) -- cgit v1.2.3 From 01dcb0443ed89eccf26c2b43f1ea13b368ae740d Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Wed, 20 Aug 2008 16:35:19 -0700 Subject: rcu: fix synchronize_rcu() so that kernel-doc works Fix RCU's synchronize_rcu() so that it looks like a C function, enabling it to be recognized as a function with kernel-doc annotation. Warning(linux-2.6.26-git11//kernel/rcupdate.c:81): No description found for parameter 'synchronize_rcu' Warning(linux-2.6.26-git11//kernel/rcupdate.c:81): No description found for parameter 'call_rcu' [akpm@linux-foundation.org: fix comment] Signed-off-by: Randy Dunlap Signed-off-by: Andrew Morton Signed-off-by: Ingo Molnar --- kernel/rcupdate.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c index f14f372cf6f5..467d5940f624 100644 --- a/kernel/rcupdate.c +++ b/kernel/rcupdate.c @@ -77,6 +77,7 @@ void wakeme_after_rcu(struct rcu_head *head) * sections are delimited by rcu_read_lock() and rcu_read_unlock(), * and may be nested. */ +void synchronize_rcu(void); /* Makes kernel-doc tools happy */ synchronize_rcu_xxx(synchronize_rcu, call_rcu) EXPORT_SYMBOL_GPL(synchronize_rcu); -- cgit v1.2.3 From 3c4fbe5e01d7e5309be5045e7ae0db20a049e6dc Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Wed, 20 Aug 2008 16:37:38 -0700 Subject: nohz: fix wrong event handler after online an offlined cpu On the tickless system(CONFIG_NO_HZ=y and CONFIG_HIGH_RES_TIMERS=n), after I made an offlined cpu online, I found this cpu's event handler was tick_handle_periodic, not tick_nohz_handler. After debuging, I found this bug was caused by the wrong tick mode. the tick mode is not changed to NOHZ_MODE_INACTIVE when the cpu is offline. This patch fixes this bug. Signed-off-by: Miao Xie Signed-off-by: Andrew Morton Signed-off-by: Ingo Molnar --- kernel/time/tick-sched.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index f5da526424a9..7a46bde78c66 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -643,17 +643,21 @@ void tick_setup_sched_timer(void) ts->nohz_mode = NOHZ_MODE_HIGHRES; #endif } +#endif /* HIGH_RES_TIMERS */ +#if defined CONFIG_NO_HZ || defined CONFIG_HIGH_RES_TIMERS void tick_cancel_sched_timer(int cpu) { struct tick_sched *ts = &per_cpu(tick_cpu_sched, cpu); +# ifdef CONFIG_HIGH_RES_TIMERS if (ts->sched_timer.base) hrtimer_cancel(&ts->sched_timer); +# endif ts->nohz_mode = NOHZ_MODE_INACTIVE; } -#endif /* HIGH_RES_TIMERS */ +#endif /** * Async notification about clocksource changes -- cgit v1.2.3 From 7a8fc9b248e77a4eab0613acf30a6811799786b3 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Sun, 17 Aug 2008 17:36:59 +0300 Subject: removed unused #include 's This patch lets the files using linux/version.h match the files that #include it. Signed-off-by: Adrian Bunk Signed-off-by: Linus Torvalds --- kernel/nsproxy.c | 1 - kernel/power/swap.c | 1 - kernel/user_namespace.c | 1 - kernel/utsname.c | 1 - kernel/utsname_sysctl.c | 1 - 5 files changed, 5 deletions(-) (limited to 'kernel') diff --git a/kernel/nsproxy.c b/kernel/nsproxy.c index 21575fc46d05..1d3ef29a2583 100644 --- a/kernel/nsproxy.c +++ b/kernel/nsproxy.c @@ -14,7 +14,6 @@ */ #include -#include #include #include #include diff --git a/kernel/power/swap.c b/kernel/power/swap.c index a0abf9a463f9..80ccac849e46 100644 --- a/kernel/power/swap.c +++ b/kernel/power/swap.c @@ -14,7 +14,6 @@ #include #include #include -#include #include #include #include diff --git a/kernel/user_namespace.c b/kernel/user_namespace.c index a9ab0596de44..532858fa5b88 100644 --- a/kernel/user_namespace.c +++ b/kernel/user_namespace.c @@ -6,7 +6,6 @@ */ #include -#include #include #include #include diff --git a/kernel/utsname.c b/kernel/utsname.c index 64d398f12444..815237a55af8 100644 --- a/kernel/utsname.c +++ b/kernel/utsname.c @@ -12,7 +12,6 @@ #include #include #include -#include #include #include diff --git a/kernel/utsname_sysctl.c b/kernel/utsname_sysctl.c index fe3a56c2256d..4ab9659d269e 100644 --- a/kernel/utsname_sysctl.c +++ b/kernel/utsname_sysctl.c @@ -12,7 +12,6 @@ #include #include #include -#include #include static void *get_uts(ctl_table *table, int write) -- cgit v1.2.3 From 354879bb977e06695993435745f06a0f6d39ce2b Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 25 Aug 2008 17:15:34 +0200 Subject: sched_clock: fix cpu_clock() This patch fixes 3 issues: a) it removes the dependency on jiffies, because jiffies are incremented by a single CPU, and the tick is not synchronized between CPUs. Therefore relying on it to calculate a window to clip whacky TSC values doesn't work as it can drift around. So instead use [GTOD, GTOD+TICK_NSEC) as the window. b) __update_sched_clock() did (roughly speaking): delta = sched_clock() - scd->tick_raw; clock += delta; Which gives exponential growth, instead of linear. c) allows the sched_clock_cpu() value to warp the u64 without breaking. the results are more reliable sched_clock() deltas: before after sched_clock cpu_clock: 15750 51312 51488 cpu_clock: 59719 51052 50947 cpu_clock: 15879 51249 51061 cpu_clock: 1 50933 51198 cpu_clock: 1 50931 51039 cpu_clock: 1 51093 50981 cpu_clock: 1 51043 51040 cpu_clock: 1 50959 50938 cpu_clock: 1 50981 51011 cpu_clock: 1 51364 51212 cpu_clock: 1 51219 51273 cpu_clock: 1 51389 51048 cpu_clock: 1 51285 51611 cpu_clock: 1 50964 51137 cpu_clock: 1 50973 50968 cpu_clock: 1 50967 50972 cpu_clock: 1 58910 58485 cpu_clock: 1 51082 51025 cpu_clock: 1 50957 50958 cpu_clock: 1 50958 50957 cpu_clock: 1006128 51128 50971 cpu_clock: 1 51107 51155 cpu_clock: 1 51371 51081 cpu_clock: 1 51104 51365 cpu_clock: 1 51363 51309 cpu_clock: 1 51107 51160 cpu_clock: 1 51139 51100 cpu_clock: 1 51216 51136 cpu_clock: 1 51207 51215 cpu_clock: 1 51087 51263 cpu_clock: 1 51249 51177 cpu_clock: 1 51519 51412 cpu_clock: 1 51416 51255 cpu_clock: 1 51591 51594 cpu_clock: 1 50966 51374 cpu_clock: 1 50966 50966 cpu_clock: 1 51291 50948 cpu_clock: 1 50973 50867 cpu_clock: 1 50970 50970 cpu_clock: 998306 50970 50971 cpu_clock: 1 50971 50970 cpu_clock: 1 50970 50970 cpu_clock: 1 50971 50971 cpu_clock: 1 50970 50970 cpu_clock: 1 51351 50970 cpu_clock: 1 50970 51352 cpu_clock: 1 50971 50970 cpu_clock: 1 50970 50970 cpu_clock: 1 51321 50971 cpu_clock: 1 50974 51324 Signed-off-by: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/sched_clock.c | 84 +++++++++++++++++++++------------------------------- 1 file changed, 34 insertions(+), 50 deletions(-) (limited to 'kernel') diff --git a/kernel/sched_clock.c b/kernel/sched_clock.c index 204991a0bfa7..e8ab096ddfe3 100644 --- a/kernel/sched_clock.c +++ b/kernel/sched_clock.c @@ -12,19 +12,17 @@ * * Create a semi stable clock from a mixture of other events, including: * - gtod - * - jiffies * - sched_clock() * - explicit idle events * * We use gtod as base and the unstable clock deltas. The deltas are filtered, - * making it monotonic and keeping it within an expected window. This window - * is set up using jiffies. + * making it monotonic and keeping it within an expected window. * * Furthermore, explicit sleep and wakeup hooks allow us to account for time * that is otherwise invisible (TSC gets stopped). * * The clock: sched_clock_cpu() is monotonic per cpu, and should be somewhat - * consistent between cpus (never more than 1 jiffies difference). + * consistent between cpus (never more than 2 jiffies difference). */ #include #include @@ -54,7 +52,6 @@ struct sched_clock_data { */ raw_spinlock_t lock; - unsigned long tick_jiffies; u64 tick_raw; u64 tick_gtod; u64 clock; @@ -75,14 +72,12 @@ static inline struct sched_clock_data *cpu_sdc(int cpu) void sched_clock_init(void) { u64 ktime_now = ktime_to_ns(ktime_get()); - unsigned long now_jiffies = jiffies; int cpu; for_each_possible_cpu(cpu) { struct sched_clock_data *scd = cpu_sdc(cpu); scd->lock = (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED; - scd->tick_jiffies = now_jiffies; scd->tick_raw = 0; scd->tick_gtod = ktime_now; scd->clock = ktime_now; @@ -91,47 +86,52 @@ void sched_clock_init(void) sched_clock_running = 1; } +/* + * min,max except they take wrapping into account + */ + +static inline u64 wrap_min(u64 x, u64 y) +{ + return (s64)(x - y) < 0 ? x : y; +} + +static inline u64 wrap_max(u64 x, u64 y) +{ + return (s64)(x - y) > 0 ? x : y; +} + /* * update the percpu scd from the raw @now value * * - filter out backward motion - * - use jiffies to generate a min,max window to clip the raw values + * - use the GTOD tick value to create a window to filter crazy TSC values */ static u64 __update_sched_clock(struct sched_clock_data *scd, u64 now) { - unsigned long now_jiffies = jiffies; - long delta_jiffies = now_jiffies - scd->tick_jiffies; - u64 clock = scd->clock; - u64 min_clock, max_clock; s64 delta = now - scd->tick_raw; + u64 clock, min_clock, max_clock; WARN_ON_ONCE(!irqs_disabled()); - min_clock = scd->tick_gtod + delta_jiffies * TICK_NSEC; - if (unlikely(delta < 0)) { - clock++; - goto out; - } + if (unlikely(delta < 0)) + delta = 0; - max_clock = min_clock + TICK_NSEC; + /* + * scd->clock = clamp(scd->tick_gtod + delta, + * max(scd->tick_gtod, scd->clock), + * scd->tick_gtod + TICK_NSEC); + */ - if (unlikely(clock + delta > max_clock)) { - if (clock < max_clock) - clock = max_clock; - else - clock++; - } else { - clock += delta; - } + clock = scd->tick_gtod + delta; + min_clock = wrap_max(scd->tick_gtod, scd->clock); + max_clock = scd->tick_gtod + TICK_NSEC; - out: - if (unlikely(clock < min_clock)) - clock = min_clock; + clock = wrap_max(clock, min_clock); + clock = wrap_min(clock, max_clock); - scd->tick_jiffies = now_jiffies; scd->clock = clock; - return clock; + return scd->clock; } static void lock_double_clock(struct sched_clock_data *data1, @@ -171,7 +171,7 @@ u64 sched_clock_cpu(int cpu) * larger time as the latest time for both * runqueues. (this creates monotonic movement) */ - if (likely(remote_clock < this_clock)) { + if (likely((s64)(remote_clock - this_clock) < 0)) { clock = this_clock; scd->clock = clock; } else { @@ -207,14 +207,9 @@ void sched_clock_tick(void) now = sched_clock(); __raw_spin_lock(&scd->lock); - __update_sched_clock(scd, now); - /* - * update tick_gtod after __update_sched_clock() because that will - * already observe 1 new jiffy; adding a new tick_gtod to that would - * increase the clock 2 jiffies. - */ scd->tick_raw = now; scd->tick_gtod = now_gtod; + __update_sched_clock(scd, now); __raw_spin_unlock(&scd->lock); } @@ -232,18 +227,7 @@ EXPORT_SYMBOL_GPL(sched_clock_idle_sleep_event); */ void sched_clock_idle_wakeup_event(u64 delta_ns) { - struct sched_clock_data *scd = this_scd(); - - /* - * Override the previous timestamp and ignore all - * sched_clock() deltas that occured while we idled, - * and use the PM-provided delta_ns to advance the - * rq clock: - */ - __raw_spin_lock(&scd->lock); - scd->clock += delta_ns; - __raw_spin_unlock(&scd->lock); - + sched_clock_tick(); touch_softlockup_watchdog(); } EXPORT_SYMBOL_GPL(sched_clock_idle_wakeup_event); -- cgit v1.2.3 From ffb4ba76a25ab6c9deeec33e4f58395586ca747c Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Mon, 25 Aug 2008 11:10:26 -0700 Subject: [module] Don't let gcc inline load_module() 'load_module()' is a complex function that contains all the ELF section logic, and inlining it is utterly insane. But gcc will do it, simply because there is only one call-site. As a result, all the stack space that is allocated for all the work to load the module will still be active when we actually call the module init sequence, and the deep call chain makes stack overflows happen. And stack overflows are really hard to debug, because they not only corrupt random pages below the stack, but also corrupt the thread_info structure that is allocated under the stack. In this case, Alan Brunelle reported some crazy oopses at bootup, after loading the processor module that ends up doing complex ACPI stuff and has quite a deep callchain. This should fix it, and is the sane thing to do regardless. Cc: Alan D. Brunelle Cc: Arjan van de Ven Cc: Rusty Russell Signed-off-by: Linus Torvalds --- 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 08864d257eb0..9db11911e04b 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -1799,7 +1799,7 @@ static void *module_alloc_update_bounds(unsigned long size) /* Allocate and load the module: note that size of section 0 is always zero, and we rely on this for optional sections. */ -static struct module *load_module(void __user *umod, +static noinline struct module *load_module(void __user *umod, unsigned long len, const char __user *uargs) { -- cgit v1.2.3 From f73be6dedf4fa058ce80846dae604b08fa805ca1 Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Mon, 25 Aug 2008 17:07:14 -0700 Subject: smp: have smp_call_function_single() detect invalid CPUs Have smp_call_function_single() return invalid CPU indicies and return -ENXIO. This function is already executed inside a get_cpu()..put_cpu() which locks out CPU removal, so rather than having the higher layers doing another layer of locking to guard against unplugged CPUs do the test here. Signed-off-by: H. Peter Anvin --- kernel/smp.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/smp.c b/kernel/smp.c index 782e2b93e465..f362a8553777 100644 --- a/kernel/smp.c +++ b/kernel/smp.c @@ -210,8 +210,10 @@ int smp_call_function_single(int cpu, void (*func) (void *info), void *info, { struct call_single_data d; unsigned long flags; - /* prevent preemption and reschedule on another processor */ + /* prevent preemption and reschedule on another processor, + as well as CPU removal */ int me = get_cpu(); + int err = 0; /* Can deadlock when called with interrupts disabled */ WARN_ON(irqs_disabled()); @@ -220,7 +222,7 @@ int smp_call_function_single(int cpu, void (*func) (void *info), void *info, local_irq_save(flags); func(info); local_irq_restore(flags); - } else { + } else if ((unsigned)cpu < NR_CPUS && cpu_online(cpu)) { struct call_single_data *data = NULL; if (!wait) { @@ -236,10 +238,12 @@ int smp_call_function_single(int cpu, void (*func) (void *info), void *info, data->func = func; data->info = info; generic_exec_single(cpu, data); + } else { + err = -ENXIO; /* CPU not online */ } put_cpu(); - return 0; + return err; } EXPORT_SYMBOL(smp_call_function_single); -- cgit v1.2.3 From 2189459d25a47401c69a17794c9d390c890351f9 Mon Sep 17 00:00:00 2001 From: Joe Korty Date: Mon, 25 Aug 2008 17:15:33 -0400 Subject: lockstat: fix numerical output rounding error Fix rounding error in /proc/lock_stat numerical output. On occasion the two digit fractional part contains the three digit value '100'. This is due to a bug in the rounding algorithm which pushes values in the range '95..99' to '100' rather than to '00' + an increment to the integer part. For example, - 123456.100 old display + 123457.00 new display Signed-off-by: Ingo Molnar --- kernel/lockdep_proc.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/lockdep_proc.c b/kernel/lockdep_proc.c index 4b194d34d77f..20dbcbf9c7dd 100644 --- a/kernel/lockdep_proc.c +++ b/kernel/lockdep_proc.c @@ -472,8 +472,9 @@ static void snprint_time(char *buf, size_t bufsiz, s64 nr) { unsigned long rem; + nr += 5; /* for display rounding */ rem = do_div(nr, 1000); /* XXX: do_div_signed */ - snprintf(buf, bufsiz, "%lld.%02d", (long long)nr, ((int)rem+5)/10); + snprintf(buf, bufsiz, "%lld.%02d", (long long)nr, (int)rem/10); } static void seq_time(struct seq_file *m, s64 time) -- cgit v1.2.3 From 04148b73b89d49fe0fe201bcee395e51f7d637ce Mon Sep 17 00:00:00 2001 From: Joe Korty Date: Mon, 25 Aug 2008 17:16:23 -0400 Subject: lockstat: repair erronous contention statistics Fix bad contention counting in /proc/lock_stat. /proc/lockstat tries to gather per-ip contention statistics per-lock. This was failing due to a garbage per-ip index selector being used. Signed-off-by: Ingo Molnar --- kernel/lockdep.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/lockdep.c b/kernel/lockdep.c index 3bfb1877a003..b5db51d2803d 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -3029,7 +3029,7 @@ found_it: stats = get_lock_stats(hlock_class(hlock)); if (point < ARRAY_SIZE(stats->contention_point)) - stats->contention_point[i]++; + stats->contention_point[point]++; if (lock->cpu != smp_processor_id()) stats->bounces[bounce_contended + !!hlock->read]++; put_lock_stats(stats); -- cgit v1.2.3 From 74870172824a78640ec4f03058d9bd35dfa08618 Mon Sep 17 00:00:00 2001 From: Zhu Yi Date: Wed, 27 Aug 2008 14:33:00 +0800 Subject: lockdep: fix invalid list_del_rcu in zap_class The problem is found during iwlagn driver testing on v2.6.27-rc4-176-gb8e6c91 kernel, but it turns out to be a lockdep bug. In our testing, we frequently load and unload the iwlagn driver (>50 times). Then the MAX_STACK_TRACE_ENTRIES is reached (expected behaviour?). The error message with the call trace is as below. BUG: MAX_STACK_TRACE_ENTRIES too low! turning off the locking correctness validator. Pid: 4895, comm: iwlagn Not tainted 2.6.27-rc4 #13 Call Trace: [] save_stack_trace+0x22/0x3e [] save_trace+0x8b/0x91 [] mark_lock+0x1b0/0x8fa [] __lock_acquire+0x5b9/0x716 [] ieee80211_sta_work+0x0/0x6ea [mac80211] [] lock_acquire+0x52/0x6b [] run_workqueue+0x97/0x1ed [] run_workqueue+0xe7/0x1ed [] run_workqueue+0x97/0x1ed [] worker_thread+0xd8/0xe3 [] autoremove_wake_function+0x0/0x2e [] worker_thread+0x0/0xe3 [] kthread+0x47/0x73 [] trace_hardirqs_on_thunk+0x3a/0x3f [] child_rip+0xa/0x11 [] restore_args+0x0/0x30 [] finish_task_switch+0x0/0xcc [] kthread+0x0/0x73 [] child_rip+0x0/0x11 Although the above is harmless, when the ilwagn module is removed later lockdep will trigger a kernel oops as below. BUG: unable to handle kernel NULL pointer dereference at 0000000000000008 IP: [] zap_class+0x24/0x82 PGD 73128067 PUD 7448c067 PMD 0 Oops: 0002 [1] SMP CPU 0 Modules linked in: rfcomm l2cap bluetooth autofs4 sunrpc nf_conntrack_ipv6 xt_state nf_conntrack xt_tcpudp ip6t_ipv6header ip6t_REJECT ip6table_filter ip6_tables x_tables ipv6 cpufreq_ondemand acpi_cpufreq dm_mirror dm_log dm_multipath dm_mod snd_hda_intel sr_mod snd_seq_dummy snd_seq_oss snd_seq_midi_event battery snd_seq snd_seq_device cdrom button snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd_page_alloc e1000e snd_hwdep sg iTCO_wdt iTCO_vendor_support ac pcspkr i2c_i801 i2c_core snd soundcore video output ata_piix ata_generic libata sd_mod scsi_mod ext3 jbd mbcache uhci_hcd ohci_hcd ehci_hcd [last unloaded: mac80211] Pid: 4941, comm: modprobe Not tainted 2.6.27-rc4 #10 RIP: 0010:[] [] zap_class+0x24/0x82 RSP: 0000:ffff88007bcb3eb0 EFLAGS: 00010046 RAX: 0000000000068ee8 RBX: ffffffff8192a0a0 RCX: 0000000000000000 RDX: 0000000000000000 RSI: 0000000000001dfb RDI: ffffffff816e70b0 RBP: ffffffffa00cd000 R08: ffffffff816818f8 R09: ffff88007c923558 R10: ffffe20002ad2408 R11: ffffffff811028ec R12: ffffffff8192a0a0 R13: 000000000002bd90 R14: 0000000000000000 R15: 0000000000000296 FS: 00007f9d1cee56f0(0000) GS:ffffffff814a58c0(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b CR2: 0000000000000008 CR3: 0000000073047000 CR4: 00000000000006e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process modprobe (pid: 4941, threadinfo ffff88007bcb2000, task ffff8800758d1fc0) Stack: ffffffff81057376 0000000000000000 ffffffffa00f7b00 0000000000000000 0000000000000080 0000000000618278 00007fff24f16720 0000000000000000 ffffffff8105d37a ffffffffa00f7b00 ffffffff8105d591 313132303863616d Call Trace: [] ? lockdep_free_key_range+0x61/0xf5 [] ? free_module+0xd4/0xe4 [] ? sys_delete_module+0x1de/0x1f9 [] ? audit_syscall_entry+0x12d/0x160 [] ? system_call_fastpath+0x16/0x1b Code: b2 00 01 00 00 00 c3 31 f6 49 c7 c0 10 8a 61 81 eb 32 49 39 38 75 26 48 98 48 6b c0 38 48 8b 90 08 8a 61 81 48 8b 88 00 8a 61 81 <48> 89 51 08 48 89 0a 48 c7 80 08 8a 61 81 00 02 20 00 48 ff c6 RIP [] zap_class+0x24/0x82 RSP CR2: 0000000000000008 ---[ end trace a1297e0c4abb0f2e ]--- The root cause for this oops is in add_lock_to_list() when save_trace() fails due to MAX_STACK_TRACE_ENTRIES is reached, entry->class is assigned but entry is never added into any lock list. This makes the list_del_rcu() in zap_class() oops later when the module is unloaded. This patch fixes the problem by assigning entry->class after save_trace() returns success. Signed-off-by: Zhu Yi Signed-off-by: Ingo Molnar --- kernel/lockdep.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/lockdep.c b/kernel/lockdep.c index b5db51d2803d..dbda475b13bd 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -875,11 +875,11 @@ static int add_lock_to_list(struct lock_class *class, struct lock_class *this, if (!entry) return 0; - entry->class = this; - entry->distance = distance; if (!save_trace(&entry->trace)) return 0; + entry->class = this; + entry->distance = distance; /* * Since we never remove from the dependency list, the list can * be walked lockless by other CPUs, it's only allocation -- cgit v1.2.3 From 2633f0e57b1127f4060d70bf460140dc9bb19386 Mon Sep 17 00:00:00 2001 From: Steve VanDeBogart Date: Tue, 26 Aug 2008 15:14:36 -0700 Subject: exit signals: use of uninitialized field notify_count task->signal->notify_count is only initialized if task->signal->group_exit_task is not NULL. Reorder a conditional so that uninitialised memory is not used. Found by Valgrind. Signed-off-by: Steve VanDeBogart Signed-off-by: Ingo Molnar --- kernel/exit.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index 38ec40630149..75c647387639 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -918,8 +918,8 @@ static void exit_notify(struct task_struct *tsk, int group_dead) /* mt-exec, de_thread() is waiting for us */ if (thread_group_leader(tsk) && - tsk->signal->notify_count < 0 && - tsk->signal->group_exit_task) + tsk->signal->group_exit_task && + tsk->signal->notify_count < 0) wake_up_process(tsk->signal->group_exit_task); write_unlock_irq(&tasklist_lock); -- cgit v1.2.3 From f42ac38c59e0a03d6da0c24a63fb211393f484b0 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 27 Aug 2008 09:14:40 -0400 Subject: ftrace: disable tracing for suspend to ram I've been painstakingly debugging the issue with suspend to ram and ftraced. The 2.6.28 code does not have this issue, but since the mcount recording is not going to be in 27, this must be solved for the ftrace daemon version. The resume from suspend to ram would reboot because it was triple faulting. Debugging further, I found that calling the mcount function itself was not an issue, but it would fault when it incremented preempt_count. preempt_count is on the tasks info structure that is on the low memory address of the task's stack. For some reason, it could not write to it. Resuming out of suspend to ram does quite a lot of funny tricks to get to work, so it is not surprising at all that simply doing a preempt_disable() would cause a fault. Thanks to Rafael for suggesting to add a "while (1);" to find the place in resuming that is causing the fault. I would place the loop somewhere in the code, compile and reboot and see if it would either reboot (hit the fault) or simply hang (hit the loop). Doing this over and over again, I narrowed it down that it was happening in enable_nonboot_cpus. At this point, I found that it is easier to simply disable tracing around the suspend code, instead of searching for the particular function that can not handle doing a preempt_disable. This patch disables the tracer as it suspends and reenables it on resume. I tested this patch on my Laptop, and it can resume fine with the patch. Signed-off-by: Steven Rostedt Acked-by: Rafael J. Wysocki Signed-off-by: Linus Torvalds --- kernel/power/main.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/power/main.c b/kernel/power/main.c index 0b7476f5d2a6..540b16b68565 100644 --- a/kernel/power/main.c +++ b/kernel/power/main.c @@ -21,6 +21,7 @@ #include #include #include +#include #include "power.h" @@ -310,7 +311,7 @@ static int suspend_enter(suspend_state_t state) */ int suspend_devices_and_enter(suspend_state_t state) { - int error; + int error, ftrace_save; if (!suspend_ops) return -ENOSYS; @@ -321,6 +322,7 @@ int suspend_devices_and_enter(suspend_state_t state) goto Close; } suspend_console(); + ftrace_save = __ftrace_enabled_save(); suspend_test_start(); error = device_suspend(PMSG_SUSPEND); if (error) { @@ -352,6 +354,7 @@ int suspend_devices_and_enter(suspend_state_t state) suspend_test_start(); device_resume(PMSG_RESUME); suspend_test_finish("resume devices"); + __ftrace_enabled_restore(ftrace_save); resume_console(); Close: if (suspend_ops->end) -- cgit v1.2.3 From f3ade837808121ff8bab9c56725f4fe40ec85a56 Mon Sep 17 00:00:00 2001 From: John Blackwood Date: Tue, 26 Aug 2008 15:09:43 -0400 Subject: sched: fix sched_rt_rq_enqueue() resched idle When sysctl_sched_rt_runtime is set to something other than -1 and the CONFIG_RT_GROUP_SCHED kernel parameter is NOT enabled, we get into a state where we see one or more CPUs idling forvever even though there are real-time tasks in their rt runqueue that are able to run (no longer throttled). The sequence is: - A real-time task is running when the timer sets the rt runqueue to throttled, and the rt task is resched_task()ed and switched out, and idle is switched in since there are no non-rt tasks to run on that cpu. - Eventually the do_sched_rt_period_timer() runs and un-throttles the rt runqueue, but we just exit the timer interrupt and go back to executing the idle task in the idle loop forever. If we change the sched_rt_rq_enqueue() routine to use some of the code from the CONFIG_RT_GROUP_SCHED enabled version of this same routine and resched_task() the currently executing task (idle in our case) if it is a lower priority task than the higher rt task in the now un-throttled runqueue, the problem is no longer observed. Signed-off-by: John Blackwood Signed-off-by: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/sched_rt.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'kernel') diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index 998ba54b4543..07d9b3307907 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -199,6 +199,8 @@ static inline struct rt_rq *group_rt_rq(struct sched_rt_entity *rt_se) static inline void sched_rt_rq_enqueue(struct rt_rq *rt_rq) { + if (rt_rq->rt_nr_running) + resched_task(rq_of_rt_rq(rt_rq)->curr); } static inline void sched_rt_rq_dequeue(struct rt_rq *rt_rq) -- cgit v1.2.3 From cc2991cf15ae92fa30b3ea9f56a8a5a337bd33c7 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 19 Aug 2008 12:33:03 +0200 Subject: sched: rt-bandwidth accounting fix It fixes an accounting bug where we would continue accumulating runtime even though the bandwidth control is disabled. This would lead to very long throttle periods once bandwidth control gets turned on again. Signed-off-by: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/sched_rt.c | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index 07d9b3307907..552310798dad 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -440,9 +440,6 @@ static int sched_rt_runtime_exceeded(struct rt_rq *rt_rq) { u64 runtime = sched_rt_runtime(rt_rq); - if (runtime == RUNTIME_INF) - return 0; - if (rt_rq->rt_throttled) return rt_rq_throttled(rt_rq); @@ -493,9 +490,11 @@ static void update_curr_rt(struct rq *rq) rt_rq = rt_rq_of_se(rt_se); spin_lock(&rt_rq->rt_runtime_lock); - rt_rq->rt_time += delta_exec; - if (sched_rt_runtime_exceeded(rt_rq)) - resched_task(curr); + if (sched_rt_runtime(rt_rq) != RUNTIME_INF) { + rt_rq->rt_time += delta_exec; + if (sched_rt_runtime_exceeded(rt_rq)) + resched_task(curr); + } spin_unlock(&rt_rq->rt_runtime_lock); } } -- cgit v1.2.3 From 41108eb10142e0552f2de1e4c0675b108c5f018f Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Thu, 28 Aug 2008 14:39:12 +0200 Subject: ftrace: disable tracing for hibernation In accordance with commit f42ac38c59e0a03d6da0c24a63fb211393f484b0 ("ftrace: disable tracing for suspend to ram"), disable tracing around the suspend code in hibernation code paths. Signed-off-by: Rafael J. Wysocki Acked-by: Steven Rostedt Signed-off-by: Linus Torvalds --- kernel/power/disk.c | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/power/disk.c b/kernel/power/disk.c index f011e0870b52..bbd85c60f741 100644 --- a/kernel/power/disk.c +++ b/kernel/power/disk.c @@ -21,6 +21,7 @@ #include #include #include +#include #include "power.h" @@ -255,7 +256,7 @@ static int create_image(int platform_mode) int hibernation_snapshot(int platform_mode) { - int error; + int error, ftrace_save; /* Free memory before shutting down devices. */ error = swsusp_shrink_memory(); @@ -267,6 +268,7 @@ int hibernation_snapshot(int platform_mode) goto Close; suspend_console(); + ftrace_save = __ftrace_enabled_save(); error = device_suspend(PMSG_FREEZE); if (error) goto Recover_platform; @@ -296,6 +298,7 @@ int hibernation_snapshot(int platform_mode) Resume_devices: device_resume(in_suspend ? (error ? PMSG_RECOVER : PMSG_THAW) : PMSG_RESTORE); + __ftrace_enabled_restore(ftrace_save); resume_console(); Close: platform_end(platform_mode); @@ -366,10 +369,11 @@ static int resume_target_kernel(void) int hibernation_restore(int platform_mode) { - int error; + int error, ftrace_save; pm_prepare_console(); suspend_console(); + ftrace_save = __ftrace_enabled_save(); error = device_suspend(PMSG_QUIESCE); if (error) goto Finish; @@ -384,6 +388,7 @@ int hibernation_restore(int platform_mode) platform_restore_cleanup(platform_mode); device_resume(PMSG_RECOVER); Finish: + __ftrace_enabled_restore(ftrace_save); resume_console(); pm_restore_console(); return error; @@ -396,7 +401,7 @@ int hibernation_restore(int platform_mode) int hibernation_platform_enter(void) { - int error; + int error, ftrace_save; if (!hibernation_ops) return -ENOSYS; @@ -411,6 +416,7 @@ int hibernation_platform_enter(void) goto Close; suspend_console(); + ftrace_save = __ftrace_enabled_save(); error = device_suspend(PMSG_HIBERNATE); if (error) { if (hibernation_ops->recover) @@ -445,6 +451,7 @@ int hibernation_platform_enter(void) hibernation_ops->finish(); Resume_devices: device_resume(PMSG_RESTORE); + __ftrace_enabled_restore(ftrace_save); resume_console(); Close: hibernation_ops->end(); -- cgit v1.2.3 From 316d9679f33caf7e683471647d1472bfe133d858 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Fri, 29 Aug 2008 20:06:23 +0200 Subject: Don't trigger softlockup detector on network fs blocked tasks Pulling the ethernet cable on a 2.6.27-rc system with NFS mounts currently leads to an ongoing flood of soft lockup detector backtraces for all tasks blocked on the NFS mounts when the hickup takes longer than 120s. I don't think NFS problems should be all that noisy. Luckily there's a reasonably easy way to distingush this case. Don't report task softlockup warnings for tasks in TASK_KILLABLE state, which is used by the network file systems. I believe this patch is a 2.6.27 candidate. Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- kernel/softlockup.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'kernel') diff --git a/kernel/softlockup.c b/kernel/softlockup.c index b75b492fbfcf..1a07f8ca4b92 100644 --- a/kernel/softlockup.c +++ b/kernel/softlockup.c @@ -180,6 +180,10 @@ static void check_hung_task(struct task_struct *t, unsigned long now) if (t->flags & PF_FROZEN) return; + /* Don't check for tasks waiting on network file systems like NFS */ + if (t->state & TASK_KILLABLE) + return; + if (switch_count != t->last_switch_count || !t->last_switch_timestamp) { t->last_switch_count = switch_count; t->last_switch_timestamp = now; -- cgit v1.2.3 From bef69ea0dcce574a425feb0a5aa4c63dd108b9a6 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Fri, 29 Aug 2008 20:18:31 -0700 Subject: Resource handling: add 'insert_resource_expand_to_fit()' function Not used anywhere yet, but this complements the existing plain 'insert_resource()' functionality with a version that can expand the resource we are adding in order to fix up any conflicts it has with existing resources. Signed-off-by: Linus Torvalds --- kernel/resource.c | 88 +++++++++++++++++++++++++++++++++++++++---------------- 1 file changed, 63 insertions(+), 25 deletions(-) (limited to 'kernel') diff --git a/kernel/resource.c b/kernel/resource.c index f5b518eabefe..cf0a178c7513 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -362,35 +362,21 @@ int allocate_resource(struct resource *root, struct resource *new, EXPORT_SYMBOL(allocate_resource); -/** - * insert_resource - Inserts a resource in the resource tree - * @parent: parent of the new resource - * @new: new resource to insert - * - * Returns 0 on success, -EBUSY if the resource can't be inserted. - * - * This function is equivalent to request_resource when no conflict - * happens. If a conflict happens, and the conflicting resources - * entirely fit within the range of the new resource, then the new - * resource is inserted and the conflicting resources become children of - * the new resource. +/* + * Insert a resource into the resource tree. If successful, return NULL, + * otherwise return the conflicting resource (compare to __request_resource()) */ -int insert_resource(struct resource *parent, struct resource *new) +static struct resource * __insert_resource(struct resource *parent, struct resource *new) { - int result; struct resource *first, *next; - write_lock(&resource_lock); - for (;; parent = first) { - result = 0; first = __request_resource(parent, new); if (!first) - goto out; + return first; - result = -EBUSY; if (first == parent) - goto out; + return first; if ((first->start > new->start) || (first->end < new->end)) break; @@ -401,15 +387,13 @@ int insert_resource(struct resource *parent, struct resource *new) for (next = first; ; next = next->sibling) { /* Partial overlap? Bad, and unfixable */ if (next->start < new->start || next->end > new->end) - goto out; + return next; if (!next->sibling) break; if (next->sibling->start > new->end) break; } - result = 0; - new->parent = parent; new->sibling = next->sibling; new->child = first; @@ -426,10 +410,64 @@ int insert_resource(struct resource *parent, struct resource *new) next = next->sibling; next->sibling = new; } + return NULL; +} - out: +/** + * insert_resource - Inserts a resource in the resource tree + * @parent: parent of the new resource + * @new: new resource to insert + * + * Returns 0 on success, -EBUSY if the resource can't be inserted. + * + * This function is equivalent to request_resource when no conflict + * happens. If a conflict happens, and the conflicting resources + * entirely fit within the range of the new resource, then the new + * resource is inserted and the conflicting resources become children of + * the new resource. + */ +int insert_resource(struct resource *parent, struct resource *new) +{ + struct resource *conflict; + + write_lock(&resource_lock); + conflict = __insert_resource(parent, new); + write_unlock(&resource_lock); + return conflict ? -EBUSY : 0; +} + +/** + * insert_resource_expand_to_fit - Insert a resource into the resource tree + * @parent: parent of the new resource + * @new: new resource to insert + * + * Insert a resource into the resource tree, possibly expanding it in order + * to make it encompass any conflicting resources. + */ +void insert_resource_expand_to_fit(struct resource *root, struct resource *new) +{ + if (new->parent) + return; + + write_lock(&resource_lock); + for (;;) { + struct resource *conflict; + + conflict = __insert_resource(root, new); + if (!conflict) + break; + if (conflict == root) + break; + + /* Ok, expand resource to cover the conflict, then try again .. */ + if (conflict->start < new->start) + new->start = conflict->start; + if (conflict->end > new->end) + new->end = conflict->end; + + printk("Expanded resource %s due to conflict with %s\n", new->name, conflict->name); + } write_unlock(&resource_lock); - return result; } /** -- cgit v1.2.3 From c4bacefb7aaf49da11a695f29d85d40909f17693 Mon Sep 17 00:00:00 2001 From: Cordelia Date: Mon, 18 Aug 2008 09:45:51 -0700 Subject: [PATCH] audit: Moved variable declaration to beginning of function got rid of compilation warning: ISO C90 forbids mixed declarations and code Signed-off-by: Cordelia Sam Signed-off-by: Al Viro --- kernel/auditsc.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/auditsc.c b/kernel/auditsc.c index 972f8e61d36a..59cedfb040e7 100644 --- a/kernel/auditsc.c +++ b/kernel/auditsc.c @@ -243,10 +243,11 @@ static inline int open_arg(int flags, int mask) static int audit_match_perm(struct audit_context *ctx, int mask) { + unsigned n; if (unlikely(!ctx)) return 0; - unsigned n = ctx->major; + n = ctx->major; switch (audit_classify_syscall(ctx->arch, n)) { case 0: /* native */ if ((mask & AUDIT_PERM_WRITE) && -- cgit v1.2.3 From 6781f4ae30bbb8ebf31187b3c9304be16966f5a0 Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Sun, 31 Aug 2008 20:31:55 -0700 Subject: kernel/resource.c: fix new kernel-doc warning Fix kernel-doc warning for new function: Warning(linux-2.6.27-rc5-git2//kernel/resource.c:448): No description found for parameter 'root' Signed-off-by: Randy Dunlap Signed-off-by: Linus Torvalds --- kernel/resource.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/resource.c b/kernel/resource.c index cf0a178c7513..03d796c1b2e9 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -438,7 +438,7 @@ int insert_resource(struct resource *parent, struct resource *new) /** * insert_resource_expand_to_fit - Insert a resource into the resource tree - * @parent: parent of the new resource + * @root: root resource descriptor * @new: new resource to insert * * Insert a resource into the resource tree, possibly expanding it in order -- cgit v1.2.3 From cbaed698f37494b30b2449b51c728ae48630cb2b Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Sat, 30 Aug 2008 21:08:40 +0400 Subject: softlockup: minor cleanup, don't check task->state twice The recent commit 16d9679f33caf7e683471647d1472bfe133d858 changed check_hung_task() to filter out the TASK_KILLABLE tasks. We can move this check to the caller which has to test t->state anyway. Signed-off-by: Oleg Nesterov Acked-by: Andi Kleen Signed-off-by: Linus Torvalds --- kernel/softlockup.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/softlockup.c b/kernel/softlockup.c index 1a07f8ca4b92..cb838ee93a82 100644 --- a/kernel/softlockup.c +++ b/kernel/softlockup.c @@ -180,10 +180,6 @@ static void check_hung_task(struct task_struct *t, unsigned long now) if (t->flags & PF_FROZEN) return; - /* Don't check for tasks waiting on network file systems like NFS */ - if (t->state & TASK_KILLABLE) - return; - if (switch_count != t->last_switch_count || !t->last_switch_timestamp) { t->last_switch_count = switch_count; t->last_switch_timestamp = now; @@ -237,7 +233,8 @@ static void check_hung_uninterruptible_tasks(int this_cpu) do_each_thread(g, t) { if (!--max_count) goto unlock; - if (t->state & TASK_UNINTERRUPTIBLE) + /* use "==" to skip the TASK_KILLABLE tasks waiting on NFS */ + if (t->state == TASK_UNINTERRUPTIBLE) check_hung_task(t, now); } while_each_thread(g, t); unlock: -- cgit v1.2.3 From add0d4dfd660e9e4fd0af3eac3cad23583c9558f Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 2 Sep 2008 14:35:48 -0700 Subject: pid_ns: zap_pid_ns_processes: fix the ->child_reaper changing zap_pid_ns_processes() sets pid_ns->child_reaper = NULL, this is wrong. Yes, we have already killed all tasks in this namespace, and sys_wait4() doesn't see any child. But this doesn't mean ->children list is empty, we may have EXIT_DEAD tasks which are not visible to do_wait(). In that case the subsequent forget_original_parent() will crash the kernel because it will try to re-parent these tasks to the NULL reaper. Even if there are no childs, it is not good that forget_original_parent() uses reaper == NULL. Change the code to set ->child_reaper = init_pid_ns.child_reaper instead. We could use pid_ns->parent->child_reaper as well, I think this does not really matter. These EXIT_DEAD tasks are not visible to the new ->parent after re-parenting, they will silently do release_task() eventually. Note that we must change ->child_reaper, otherwise forget_original_parent() will use reaper == father, and in that case we will hit the (correct) BUG_ON(!list_empty(&father->children)). Signed-off-by: Oleg Nesterov Acked-by: Serge Hallyn Acked-by: Sukadev Bhattiprolu Acked-by: Pavel Emelyanov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/pid_namespace.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/pid_namespace.c b/kernel/pid_namespace.c index ea567b78d1aa..598f1eec9826 100644 --- a/kernel/pid_namespace.c +++ b/kernel/pid_namespace.c @@ -179,9 +179,12 @@ void zap_pid_ns_processes(struct pid_namespace *pid_ns) rc = sys_wait4(-1, NULL, __WALL, NULL); } while (rc != -ECHILD); - - /* Child reaper for the pid namespace is going away */ - pid_ns->child_reaper = NULL; + /* + * We can not clear ->child_reaper or leave it alone. + * There may by stealth EXIT_DEAD tasks on ->children, + * forget_original_parent() must move them somewhere. + */ + pid_ns->child_reaper = init_pid_ns.child_reaper; acct_exit_ns(pid_ns); return; } -- cgit v1.2.3 From 950bbabb5a804690a0201190de5c22837f72f83f Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 2 Sep 2008 14:35:49 -0700 Subject: pid_ns: (BUG 11391) change ->child_reaper when init->group_leader exits We don't change pid_ns->child_reaper when the main thread of the subnamespace init exits. As Robert Rex pointed out this is wrong. Yes, the re-parenting itself works correctly, but if the reparented task exits it needs ->parent->nsproxy->pid_ns in do_notify_parent(), and if the main thread is zombie its ->nsproxy was already cleared by exit_task_namespaces(). Introduce the new function, find_new_reaper(), which finds the new ->parent for the re-parenting and changes ->child_reaper if needed. Kill the now unneeded exit_child_reaper(). Also move the changing of ->child_reaper from zap_pid_ns_processes() to find_new_reaper(), this consolidates the games with ->child_reaper and makes it stable under tasklist_lock. Addresses http://bugzilla.kernel.org/show_bug.cgi?id=11391 Reported-by: Robert Rex Signed-off-by: Oleg Nesterov Acked-by: Serge Hallyn Acked-by: Pavel Emelyanov Acked-by: Sukadev Bhattiprolu Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 78 ++++++++++++++++++++++---------------------------- kernel/pid_namespace.c | 6 ---- 2 files changed, 34 insertions(+), 50 deletions(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index 75c647387639..25ed2ad986df 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -831,26 +831,50 @@ static void reparent_thread(struct task_struct *p, struct task_struct *father) * the child reaper process (ie "init") in our pid * space. */ +static struct task_struct *find_new_reaper(struct task_struct *father) +{ + struct pid_namespace *pid_ns = task_active_pid_ns(father); + struct task_struct *thread; + + thread = father; + while_each_thread(father, thread) { + if (thread->flags & PF_EXITING) + continue; + if (unlikely(pid_ns->child_reaper == father)) + pid_ns->child_reaper = thread; + return thread; + } + + if (unlikely(pid_ns->child_reaper == father)) { + write_unlock_irq(&tasklist_lock); + if (unlikely(pid_ns == &init_pid_ns)) + panic("Attempted to kill init!"); + + zap_pid_ns_processes(pid_ns); + write_lock_irq(&tasklist_lock); + /* + * We can not clear ->child_reaper or leave it alone. + * There may by stealth EXIT_DEAD tasks on ->children, + * forget_original_parent() must move them somewhere. + */ + pid_ns->child_reaper = init_pid_ns.child_reaper; + } + + return pid_ns->child_reaper; +} + static void forget_original_parent(struct task_struct *father) { - struct task_struct *p, *n, *reaper = father; + struct task_struct *p, *n, *reaper; LIST_HEAD(ptrace_dead); write_lock_irq(&tasklist_lock); - + reaper = find_new_reaper(father); /* * First clean up ptrace if we were using it. */ ptrace_exit(father, &ptrace_dead); - do { - reaper = next_thread(reaper); - if (reaper == father) { - reaper = task_child_reaper(father); - break; - } - } while (reaper->flags & PF_EXITING); - list_for_each_entry_safe(p, n, &father->children, sibling) { p->real_parent = reaper; if (p->parent == father) { @@ -959,39 +983,6 @@ static void check_stack_usage(void) static inline void check_stack_usage(void) {} #endif -static inline void exit_child_reaper(struct task_struct *tsk) -{ - if (likely(tsk->group_leader != task_child_reaper(tsk))) - return; - - if (tsk->nsproxy->pid_ns == &init_pid_ns) - panic("Attempted to kill init!"); - - /* - * @tsk is the last thread in the 'cgroup-init' and is exiting. - * Terminate all remaining processes in the namespace and reap them - * before exiting @tsk. - * - * Note that @tsk (last thread of cgroup-init) may not necessarily - * be the child-reaper (i.e main thread of cgroup-init) of the - * namespace i.e the child_reaper may have already exited. - * - * Even after a child_reaper exits, we let it inherit orphaned children, - * because, pid_ns->child_reaper remains valid as long as there is - * at least one living sub-thread in the cgroup init. - - * This living sub-thread of the cgroup-init will be notified when - * a child inherited by the 'child-reaper' exits (do_notify_parent() - * uses __group_send_sig_info()). Further, when reaping child processes, - * do_wait() iterates over children of all living sub threads. - - * i.e even though 'child_reaper' thread is listed as the parent of the - * orphaned children, any living sub-thread in the cgroup-init can - * perform the role of the child_reaper. - */ - zap_pid_ns_processes(tsk->nsproxy->pid_ns); -} - NORET_TYPE void do_exit(long code) { struct task_struct *tsk = current; @@ -1051,7 +1042,6 @@ NORET_TYPE void do_exit(long code) } group_dead = atomic_dec_and_test(&tsk->signal->live); if (group_dead) { - exit_child_reaper(tsk); hrtimer_cancel(&tsk->signal->real_timer); exit_itimers(tsk->signal); } diff --git a/kernel/pid_namespace.c b/kernel/pid_namespace.c index 598f1eec9826..fab8ea86fac3 100644 --- a/kernel/pid_namespace.c +++ b/kernel/pid_namespace.c @@ -179,12 +179,6 @@ void zap_pid_ns_processes(struct pid_namespace *pid_ns) rc = sys_wait4(-1, NULL, __WALL, NULL); } while (rc != -ECHILD); - /* - * We can not clear ->child_reaper or leave it alone. - * There may by stealth EXIT_DEAD tasks on ->children, - * forget_original_parent() must move them somewhere. - */ - pid_ns->child_reaper = init_pid_ns.child_reaper; acct_exit_ns(pid_ns); return; } -- cgit v1.2.3 From 9d3593574702ae1899e23a1535da1ac71f928042 Mon Sep 17 00:00:00 2001 From: John Kacur Date: Tue, 2 Sep 2008 14:36:13 -0700 Subject: pm_qos_requirement might sleep Make PM_QOS and CPU_IDLE play nicer when run with the RT-Preempt kernel. The purpose of the patch is to remove the spin_lock around the read in the function pm_qos_requirement - since spinlocks can sleep in -rt and this function is called from idle. CPU_IDLE polls the target_value's of some of the pm_qos parameters from the idle loop causing sleeping locking warnings. Changing the target_value to an atomic avoids this issue. Remove the spinlock in pm_qos_requirement by making target_value an atomic type. Signed-off-by: mark gross Signed-off-by: John Kacur Cc: Steven Rostedt Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/pm_qos_params.c | 25 +++++++++---------------- 1 file changed, 9 insertions(+), 16 deletions(-) (limited to 'kernel') diff --git a/kernel/pm_qos_params.c b/kernel/pm_qos_params.c index da9c2dda6a4e..dfdec524d1b7 100644 --- a/kernel/pm_qos_params.c +++ b/kernel/pm_qos_params.c @@ -43,7 +43,7 @@ #include /* - * locking rule: all changes to target_value or requirements or notifiers lists + * locking rule: all changes to requirements or notifiers lists * or pm_qos_object list and pm_qos_objects need to happen with pm_qos_lock * held, taken with _irqsave. One lock to rule them all */ @@ -66,7 +66,7 @@ struct pm_qos_object { struct miscdevice pm_qos_power_miscdev; char *name; s32 default_value; - s32 target_value; + atomic_t target_value; s32 (*comparitor)(s32, s32); }; @@ -77,7 +77,7 @@ static struct pm_qos_object cpu_dma_pm_qos = { .notifiers = &cpu_dma_lat_notifier, .name = "cpu_dma_latency", .default_value = 2000 * USEC_PER_SEC, - .target_value = 2000 * USEC_PER_SEC, + .target_value = ATOMIC_INIT(2000 * USEC_PER_SEC), .comparitor = min_compare }; @@ -87,7 +87,7 @@ static struct pm_qos_object network_lat_pm_qos = { .notifiers = &network_lat_notifier, .name = "network_latency", .default_value = 2000 * USEC_PER_SEC, - .target_value = 2000 * USEC_PER_SEC, + .target_value = ATOMIC_INIT(2000 * USEC_PER_SEC), .comparitor = min_compare }; @@ -99,7 +99,7 @@ static struct pm_qos_object network_throughput_pm_qos = { .notifiers = &network_throughput_notifier, .name = "network_throughput", .default_value = 0, - .target_value = 0, + .target_value = ATOMIC_INIT(0), .comparitor = max_compare }; @@ -150,11 +150,11 @@ static void update_target(int target) extreme_value = pm_qos_array[target]->comparitor( extreme_value, node->value); } - if (pm_qos_array[target]->target_value != extreme_value) { + if (atomic_read(&pm_qos_array[target]->target_value) != extreme_value) { call_notifier = 1; - pm_qos_array[target]->target_value = extreme_value; + atomic_set(&pm_qos_array[target]->target_value, extreme_value); pr_debug(KERN_ERR "new target for qos %d is %d\n", target, - pm_qos_array[target]->target_value); + atomic_read(&pm_qos_array[target]->target_value)); } spin_unlock_irqrestore(&pm_qos_lock, flags); @@ -193,14 +193,7 @@ static int find_pm_qos_object_by_minor(int minor) */ int pm_qos_requirement(int pm_qos_class) { - int ret_val; - unsigned long flags; - - spin_lock_irqsave(&pm_qos_lock, flags); - ret_val = pm_qos_array[pm_qos_class]->target_value; - spin_unlock_irqrestore(&pm_qos_lock, flags); - - return ret_val; + return atomic_read(&pm_qos_array[pm_qos_class]->target_value); } EXPORT_SYMBOL_GPL(pm_qos_requirement); -- cgit v1.2.3 From b380b0d4f7dffcc235c0facefa537d4655619101 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Thu, 4 Sep 2008 17:05:57 +0100 Subject: forgotten refcount on sysctl root table We should've set refcount on the root sysctl table; otherwise we'll blow up the first time we get down to zero dynamically registered sysctl tables. Signed-off-by: Al Viro Tested-by: James Bottomley Signed-off-by: Linus Torvalds --- kernel/sysctl.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/sysctl.c b/kernel/sysctl.c index fe4713347275..50ec0886fa3d 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -159,6 +159,7 @@ static int proc_dointvec_taint(struct ctl_table *table, int write, struct file * static struct ctl_table root_table[]; static struct ctl_table_root sysctl_table_root; static struct ctl_table_header root_table_header = { + .count = 1, .ctl_table = root_table, .ctl_entry = LIST_HEAD_INIT(sysctl_table_root.default_set.list), .root = &sysctl_table_root, -- cgit v1.2.3 From 7c1e76897492d92b6a1c2d6892494d39ded9680c Mon Sep 17 00:00:00 2001 From: Venkatesh Pallipadi Date: Wed, 3 Sep 2008 21:36:50 +0000 Subject: clockevents: prevent clockevent event_handler ending up handler_noop There is a ordering related problem with clockevents code, due to which clockevents_register_device() called after tickless/highres switch will not work. The new clockevent ends up with clockevents_handle_noop as event handler, resulting in no timer activity. The problematic path seems to be * old device already has hrtimer_interrupt as the event_handler * new clockevent device registers with a higher rating * tick_check_new_device() is called * clockevents_exchange_device() gets called * old->event_handler is set to clockevents_handle_noop * tick_setup_device() is called for the new device * which sets new->event_handler using the old->event_handler which is noop. Change the ordering so that new device inherits the proper handler. This does not have any issue in normal case as most likely all the clockevent devices are setup before the highres switch. But, can potentially be affecting some corner case where HPET force detect happens after the highres switch. This was a problem with HPET in MSI mode code that we have been experimenting with. Signed-off-by: Venkatesh Pallipadi Signed-off-by: Shaohua Li Signed-off-by: Thomas Gleixner Signed-off-by: Ingo Molnar --- kernel/time/clockevents.c | 3 +-- kernel/time/tick-common.c | 1 + 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/time/clockevents.c b/kernel/time/clockevents.c index 3d1e3e1a1971..1876b526c778 100644 --- a/kernel/time/clockevents.c +++ b/kernel/time/clockevents.c @@ -177,7 +177,7 @@ void clockevents_register_device(struct clock_event_device *dev) /* * Noop handler when we shut down an event device */ -static void clockevents_handle_noop(struct clock_event_device *dev) +void clockevents_handle_noop(struct clock_event_device *dev) { } @@ -199,7 +199,6 @@ void clockevents_exchange_device(struct clock_event_device *old, * released list and do a notify add later. */ if (old) { - old->event_handler = clockevents_handle_noop; clockevents_set_mode(old, CLOCK_EVT_MODE_UNUSED); list_del(&old->list); list_add(&old->list, &clockevents_released); diff --git a/kernel/time/tick-common.c b/kernel/time/tick-common.c index 80c4336f4188..c4777193d567 100644 --- a/kernel/time/tick-common.c +++ b/kernel/time/tick-common.c @@ -161,6 +161,7 @@ static void tick_setup_device(struct tick_device *td, } else { handler = td->evtdev->event_handler; next_event = td->evtdev->next_event; + td->evtdev->event_handler = clockevents_handle_noop; } td->evtdev = newdev; -- cgit v1.2.3 From d4496b39559c6d43f83e4c08b899984f8b8089b5 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 3 Sep 2008 21:36:57 +0000 Subject: clockevents: prevent endless loop in periodic broadcast handler The reprogramming of the periodic broadcast handler was broken, when the first programming returned -ETIME. The clockevents code stores the new expiry value in the clock events device next_event field only when the programming time has not been elapsed yet. The loop in question calculates the new expiry value from the next_event value and therefor never increases. Signed-off-by: Thomas Gleixner Signed-off-by: Ingo Molnar --- kernel/time/tick-broadcast.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-broadcast.c b/kernel/time/tick-broadcast.c index 31463d370b94..3044a88357fa 100644 --- a/kernel/time/tick-broadcast.c +++ b/kernel/time/tick-broadcast.c @@ -175,6 +175,8 @@ static void tick_do_periodic_broadcast(void) */ static void tick_handle_periodic_broadcast(struct clock_event_device *dev) { + ktime_t next; + tick_do_periodic_broadcast(); /* @@ -185,10 +187,13 @@ static void tick_handle_periodic_broadcast(struct clock_event_device *dev) /* * Setup the next period for devices, which do not have - * periodic mode: + * periodic mode. We read dev->next_event first and add to it + * when the event alrady expired. clockevents_program_event() + * sets dev->next_event only when the event is really + * programmed to the device. */ - for (;;) { - ktime_t next = ktime_add(dev->next_event, tick_period); + for (next = dev->next_event; ;) { + next = ktime_add(next, tick_period); if (!clockevents_program_event(dev, next, ktime_get())) return; -- cgit v1.2.3 From 7205656ab48da29a95d7f55e43a81db755d3cb3a Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 3 Sep 2008 21:37:03 +0000 Subject: clockevents: enforce reprogram in oneshot setup In tick_oneshot_setup we program the device to the given next_event, but we do not check the return value. We need to make sure that the device is programmed enforced so the interrupt handler engine starts working. Split out the reprogramming function from tick_program_event() and call it with the device, which was handed in to tick_setup_oneshot(). Set the force argument, so the devices is firing an interrupt. Signed-off-by: Thomas Gleixner Signed-off-by: Ingo Molnar --- kernel/time/tick-oneshot.c | 18 ++++++++++++++---- 1 file changed, 14 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-oneshot.c b/kernel/time/tick-oneshot.c index 450c04935b66..06595c64b0c9 100644 --- a/kernel/time/tick-oneshot.c +++ b/kernel/time/tick-oneshot.c @@ -23,11 +23,11 @@ #include "tick-internal.h" /** - * tick_program_event + * tick_program_event internal worker function */ -int tick_program_event(ktime_t expires, int force) +static int __tick_program_event(struct clock_event_device *dev, + ktime_t expires, int force) { - struct clock_event_device *dev = __get_cpu_var(tick_cpu_device).evtdev; ktime_t now = ktime_get(); while (1) { @@ -40,6 +40,16 @@ int tick_program_event(ktime_t expires, int force) } } +/** + * tick_program_event + */ +int tick_program_event(ktime_t expires, int force) +{ + struct clock_event_device *dev = __get_cpu_var(tick_cpu_device).evtdev; + + return __tick_program_event(dev, expires, force); +} + /** * tick_resume_onshot - resume oneshot mode */ @@ -61,7 +71,7 @@ void tick_setup_oneshot(struct clock_event_device *newdev, { newdev->event_handler = handler; clockevents_set_mode(newdev, CLOCK_EVT_MODE_ONESHOT); - clockevents_program_event(newdev, next_event, ktime_get()); + __tick_program_event(newdev, next_event, 1); } /** -- cgit v1.2.3 From 9c17bcda991000351cb2373f78be7e4b1c44caa3 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 3 Sep 2008 21:37:08 +0000 Subject: clockevents: prevent multiple init/shutdown While chasing the C1E/HPET bugreports I went through the clock events code inch by inch and found that the broadcast device can be initialized and shutdown multiple times. Multiple shutdowns are not critical, but useless waste of time. Multiple initializations are simply broken. Another CPU might have the device in use already after the first initialization and the second init could just render it unusable again. Signed-off-by: Thomas Gleixner Signed-off-by: Ingo Molnar --- kernel/time/tick-broadcast.c | 20 +++++++++++++------- 1 file changed, 13 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-broadcast.c b/kernel/time/tick-broadcast.c index 3044a88357fa..5744f40b2697 100644 --- a/kernel/time/tick-broadcast.c +++ b/kernel/time/tick-broadcast.c @@ -210,7 +210,7 @@ static void tick_do_broadcast_on_off(void *why) struct clock_event_device *bc, *dev; struct tick_device *td; unsigned long flags, *reason = why; - int cpu; + int cpu, bc_stopped; spin_lock_irqsave(&tick_broadcast_lock, flags); @@ -228,6 +228,8 @@ static void tick_do_broadcast_on_off(void *why) if (!tick_device_is_functional(dev)) goto out; + bc_stopped = cpus_empty(tick_broadcast_mask); + switch (*reason) { case CLOCK_EVT_NOTIFY_BROADCAST_ON: case CLOCK_EVT_NOTIFY_BROADCAST_FORCE: @@ -250,9 +252,10 @@ static void tick_do_broadcast_on_off(void *why) break; } - if (cpus_empty(tick_broadcast_mask)) - clockevents_set_mode(bc, CLOCK_EVT_MODE_SHUTDOWN); - else { + if (cpus_empty(tick_broadcast_mask)) { + if (!bc_stopped) + clockevents_set_mode(bc, CLOCK_EVT_MODE_SHUTDOWN); + } else if (bc_stopped) { if (tick_broadcast_device.mode == TICKDEV_MODE_PERIODIC) tick_broadcast_start_periodic(bc); else @@ -501,9 +504,12 @@ static void tick_broadcast_clear_oneshot(int cpu) */ void tick_broadcast_setup_oneshot(struct clock_event_device *bc) { - bc->event_handler = tick_handle_oneshot_broadcast; - clockevents_set_mode(bc, CLOCK_EVT_MODE_ONESHOT); - bc->next_event.tv64 = KTIME_MAX; + /* Set it up only once ! */ + if (bc->event_handler != tick_handle_oneshot_broadcast) { + bc->event_handler = tick_handle_oneshot_broadcast; + clockevents_set_mode(bc, CLOCK_EVT_MODE_ONESHOT); + bc->next_event.tv64 = KTIME_MAX; + } } /* -- cgit v1.2.3 From 1fb9b7d29d8e85ba3196eaa7ab871bf76fc98d36 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 3 Sep 2008 21:37:14 +0000 Subject: clockevents: prevent endless loop lockup The C1E/HPET bug reports on AMDX2/RS690 systems where tracked down to a too small value of the HPET minumum delta for programming an event. The clockevents code needs to enforce an interrupt event on the clock event device in some cases. The enforcement code was stupid and naive, as it just added the minimum delta to the current time and tried to reprogram the device. When the minimum delta is too small, then this loops forever. Add a sanity check. Allow reprogramming to fail 3 times, then print a warning and double the minimum delta value to make sure, that this does not happen again. Use the same function for both tick-oneshot and tick-broadcast code. Signed-off-by: Thomas Gleixner Signed-off-by: Ingo Molnar --- kernel/time/tick-broadcast.c | 12 ++---------- kernel/time/tick-internal.h | 2 ++ kernel/time/tick-oneshot.c | 36 ++++++++++++++++++++++++++++++------ 3 files changed, 34 insertions(+), 16 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-broadcast.c b/kernel/time/tick-broadcast.c index 5744f40b2697..2bc1f046151c 100644 --- a/kernel/time/tick-broadcast.c +++ b/kernel/time/tick-broadcast.c @@ -372,16 +372,8 @@ cpumask_t *tick_get_broadcast_oneshot_mask(void) static int tick_broadcast_set_event(ktime_t expires, int force) { struct clock_event_device *bc = tick_broadcast_device.evtdev; - ktime_t now = ktime_get(); - int res; - - for(;;) { - res = clockevents_program_event(bc, expires, now); - if (!res || !force) - return res; - now = ktime_get(); - expires = ktime_add(now, ktime_set(0, bc->min_delta_ns)); - } + + return tick_dev_program_event(bc, expires, force); } int tick_resume_broadcast_oneshot(struct clock_event_device *bc) diff --git a/kernel/time/tick-internal.h b/kernel/time/tick-internal.h index f13f2b7f4fd4..0ffc2918ea6f 100644 --- a/kernel/time/tick-internal.h +++ b/kernel/time/tick-internal.h @@ -17,6 +17,8 @@ extern void tick_handle_periodic(struct clock_event_device *dev); extern void tick_setup_oneshot(struct clock_event_device *newdev, void (*handler)(struct clock_event_device *), ktime_t nextevt); +extern int tick_dev_program_event(struct clock_event_device *dev, + ktime_t expires, int force); extern int tick_program_event(ktime_t expires, int force); extern void tick_oneshot_notify(void); extern int tick_switch_to_oneshot(void (*handler)(struct clock_event_device *)); diff --git a/kernel/time/tick-oneshot.c b/kernel/time/tick-oneshot.c index 06595c64b0c9..2e35501e61dd 100644 --- a/kernel/time/tick-oneshot.c +++ b/kernel/time/tick-oneshot.c @@ -25,18 +25,42 @@ /** * tick_program_event internal worker function */ -static int __tick_program_event(struct clock_event_device *dev, - ktime_t expires, int force) +int tick_dev_program_event(struct clock_event_device *dev, ktime_t expires, + int force) { ktime_t now = ktime_get(); + int i; - while (1) { + for (i = 0;;) { int ret = clockevents_program_event(dev, expires, now); if (!ret || !force) return ret; + + /* + * We tried 2 times to program the device with the given + * min_delta_ns. If that's not working then we double it + * and emit a warning. + */ + if (++i > 2) { + printk(KERN_WARNING "CE: __tick_program_event of %s is " + "stuck %llx %llx\n", dev->name ? dev->name : "?", + now.tv64, expires.tv64); + printk(KERN_WARNING + "CE: increasing min_delta_ns %ld to %ld nsec\n", + dev->min_delta_ns, dev->min_delta_ns << 1); + WARN_ON(1); + + /* Double the min. delta and try again */ + if (!dev->min_delta_ns) + dev->min_delta_ns = 5000; + else + dev->min_delta_ns <<= 1; + i = 0; + } + now = ktime_get(); - expires = ktime_add(now, ktime_set(0, dev->min_delta_ns)); + expires = ktime_add_ns(now, dev->min_delta_ns); } } @@ -47,7 +71,7 @@ int tick_program_event(ktime_t expires, int force) { struct clock_event_device *dev = __get_cpu_var(tick_cpu_device).evtdev; - return __tick_program_event(dev, expires, force); + return tick_dev_program_event(dev, expires, force); } /** @@ -71,7 +95,7 @@ void tick_setup_oneshot(struct clock_event_device *newdev, { newdev->event_handler = handler; clockevents_set_mode(newdev, CLOCK_EVT_MODE_ONESHOT); - __tick_program_event(newdev, next_event, 1); + tick_dev_program_event(newdev, next_event, 1); } /** -- cgit v1.2.3 From 56c7426b3951e4f35a71d695f1c982989399d6fd Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 1 Sep 2008 16:44:23 +0200 Subject: sched_clock: fix NOHZ interaction If HLT stops the TSC, we'll fail to account idle time, thereby inflating the actual process times. Fix this by re-calibrating the clock against GTOD when leaving nohz mode. Signed-off-by: Peter Zijlstra Tested-by: Avi Kivity Signed-off-by: Ingo Molnar --- kernel/time/tick-sched.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'kernel') diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 7a46bde78c66..a87b0468568b 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -162,6 +162,8 @@ void tick_nohz_stop_idle(int cpu) ts->idle_lastupdate = now; ts->idle_sleeptime = ktime_add(ts->idle_sleeptime, delta); ts->idle_active = 0; + + sched_clock_idle_wakeup_event(0); } } @@ -177,6 +179,7 @@ static ktime_t tick_nohz_start_idle(struct tick_sched *ts) } ts->idle_entrytime = now; ts->idle_active = 1; + sched_clock_idle_sleep_event(); return now; } -- cgit v1.2.3 From 49048622eae698e5c4ae61f7e71200f265ccc529 Mon Sep 17 00:00:00 2001 From: Balbir Singh Date: Fri, 5 Sep 2008 18:12:23 +0200 Subject: sched: fix process time monotonicity Spencer reported a problem where utime and stime were going negative despite the fixes in commit b27f03d4bdc145a09fb7b0c0e004b29f1ee555fa. The suspected reason for the problem is that signal_struct maintains it's own utime and stime (of exited tasks), these are not updated using the new task_utime() routine, hence sig->utime can go backwards and cause the same problem to occur (sig->utime, adds tsk->utime and not task_utime()). This patch fixes the problem TODO: using max(task->prev_utime, derived utime) works for now, but a more generic solution is to implement cputime_max() and use the cputime_gt() function for comparison. Reported-by: spencer@bluehost.com Signed-off-by: Balbir Singh Signed-off-by: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/exit.c | 6 +++--- kernel/sched.c | 59 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 62 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index 25ed2ad986df..16395644a98f 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -112,9 +112,9 @@ static void __exit_signal(struct task_struct *tsk) * We won't ever get here for the group leader, since it * will have been the last reference on the signal_struct. */ - sig->utime = cputime_add(sig->utime, tsk->utime); - sig->stime = cputime_add(sig->stime, tsk->stime); - sig->gtime = cputime_add(sig->gtime, tsk->gtime); + sig->utime = cputime_add(sig->utime, task_utime(tsk)); + sig->stime = cputime_add(sig->stime, task_stime(tsk)); + sig->gtime = cputime_add(sig->gtime, task_gtime(tsk)); sig->min_flt += tsk->min_flt; sig->maj_flt += tsk->maj_flt; sig->nvcsw += tsk->nvcsw; diff --git a/kernel/sched.c b/kernel/sched.c index 9a1ddb84e26d..1a5f73c1fcdc 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4178,6 +4178,65 @@ void account_steal_time(struct task_struct *p, cputime_t steal) cpustat->steal = cputime64_add(cpustat->steal, tmp); } +/* + * Use precise platform statistics if available: + */ +#ifdef CONFIG_VIRT_CPU_ACCOUNTING +cputime_t task_utime(struct task_struct *p) +{ + return p->utime; +} + +cputime_t task_stime(struct task_struct *p) +{ + return p->stime; +} +#else +cputime_t task_utime(struct task_struct *p) +{ + clock_t utime = cputime_to_clock_t(p->utime), + total = utime + cputime_to_clock_t(p->stime); + u64 temp; + + /* + * Use CFS's precise accounting: + */ + temp = (u64)nsec_to_clock_t(p->se.sum_exec_runtime); + + if (total) { + temp *= utime; + do_div(temp, total); + } + utime = (clock_t)temp; + + p->prev_utime = max(p->prev_utime, clock_t_to_cputime(utime)); + return p->prev_utime; +} + +cputime_t task_stime(struct task_struct *p) +{ + clock_t stime; + + /* + * Use CFS's precise accounting. (we subtract utime from + * the total, to make sure the total observed by userspace + * grows monotonically - apps rely on that): + */ + stime = nsec_to_clock_t(p->se.sum_exec_runtime) - + cputime_to_clock_t(task_utime(p)); + + if (stime >= 0) + p->prev_stime = max(p->prev_stime, clock_t_to_cputime(stime)); + + return p->prev_stime; +} +#endif + +inline cputime_t task_gtime(struct task_struct *p) +{ + return p->gtime; +} + /* * This function gets called by the timer code, with HZ frequency. * We call it with interrupts disabled. -- cgit v1.2.3 From 7300711e8c6824fcfbd42a126980ff50439d8dd0 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Sat, 6 Sep 2008 03:01:45 +0200 Subject: clockevents: broadcast fixup possible waiters Until the C1E patches arrived there where no users of periodic broadcast before switching to oneshot mode. Now we need to trigger a possible waiter for a periodic broadcast when switching to oneshot mode. Otherwise we can starve them for ever. Signed-off-by: Thomas Gleixner --- kernel/time/tick-broadcast.c | 37 ++++++++++++++++++++++++++++++++++++- 1 file changed, 36 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/time/tick-broadcast.c b/kernel/time/tick-broadcast.c index 2bc1f046151c..2f5a38294bf9 100644 --- a/kernel/time/tick-broadcast.c +++ b/kernel/time/tick-broadcast.c @@ -491,6 +491,18 @@ static void tick_broadcast_clear_oneshot(int cpu) cpu_clear(cpu, tick_broadcast_oneshot_mask); } +static void tick_broadcast_init_next_event(cpumask_t *mask, ktime_t expires) +{ + struct tick_device *td; + int cpu; + + for_each_cpu_mask_nr(cpu, *mask) { + td = &per_cpu(tick_cpu_device, cpu); + if (td->evtdev) + td->evtdev->next_event = expires; + } +} + /** * tick_broadcast_setup_oneshot - setup the broadcast device */ @@ -498,9 +510,32 @@ void tick_broadcast_setup_oneshot(struct clock_event_device *bc) { /* Set it up only once ! */ if (bc->event_handler != tick_handle_oneshot_broadcast) { + int was_periodic = bc->mode == CLOCK_EVT_MODE_PERIODIC; + int cpu = smp_processor_id(); + cpumask_t mask; + bc->event_handler = tick_handle_oneshot_broadcast; clockevents_set_mode(bc, CLOCK_EVT_MODE_ONESHOT); - bc->next_event.tv64 = KTIME_MAX; + + /* Take the do_timer update */ + tick_do_timer_cpu = cpu; + + /* + * We must be careful here. There might be other CPUs + * waiting for periodic broadcast. We need to set the + * oneshot_mask bits for those and program the + * broadcast device to fire. + */ + mask = tick_broadcast_mask; + cpu_clear(cpu, mask); + cpus_or(tick_broadcast_oneshot_mask, + tick_broadcast_oneshot_mask, mask); + + if (was_periodic && !cpus_empty(mask)) { + tick_broadcast_init_next_event(&mask, tick_next_period); + tick_broadcast_set_event(tick_next_period, 1); + } else + bc->next_event.tv64 = KTIME_MAX; } } -- cgit v1.2.3 From 4ff4b9e19a80b73959ebeb28d1df40176686f0a8 Mon Sep 17 00:00:00 2001 From: "Maciej W. Rozycki" Date: Fri, 5 Sep 2008 14:05:31 -0700 Subject: ntp: fix calculation of the next jiffie to trigger RTC sync We have a bug in the calculation of the next jiffie to trigger the RTC synchronisation. The aim here is to run sync_cmos_clock() as close as possible to the middle of a second. Which means we want this function to be called less than or equal to half a jiffie away from when now.tv_nsec equals 5e8 (500000000). If this is not the case for a given call to the function, for this purpose instead of updating the RTC we calculate the offset in nanoseconds to the next point in time where now.tv_nsec will be equal 5e8. The calculated offset is then converted to jiffies as these are the unit used by the timer. Hovewer timespec_to_jiffies() used here uses a ceil()-type rounding mode, where the resulting value is rounded up. As a result the range of now.tv_nsec when the timer will trigger is from 5e8 to 5e8 + TICK_NSEC rather than the desired 5e8 - TICK_NSEC / 2 to 5e8 + TICK_NSEC / 2. As a result if for example sync_cmos_clock() happens to be called at the time when now.tv_nsec is between 5e8 + TICK_NSEC / 2 and 5e8 to 5e8 + TICK_NSEC, it will simply be rescheduled HZ jiffies later, falling in the same range of now.tv_nsec again. Similarly for cases offsetted by an integer multiple of TICK_NSEC. This change addresses the problem by subtracting TICK_NSEC / 2 from the nanosecond offset to the next point in time where now.tv_nsec will be equal 5e8, effectively shifting the following rounding in timespec_to_jiffies() so that it produces a rounded-to-nearest result. Signed-off-by: Maciej W. Rozycki Signed-off-by: Andrew Morton Signed-off-by: Ingo Molnar --- kernel/time/ntp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/time/ntp.c b/kernel/time/ntp.c index 5125ddd8196b..1ad46f3df6e7 100644 --- a/kernel/time/ntp.c +++ b/kernel/time/ntp.c @@ -245,7 +245,7 @@ static void sync_cmos_clock(unsigned long dummy) if (abs(now.tv_nsec - (NSEC_PER_SEC / 2)) <= tick_nsec / 2) fail = update_persistent_clock(now); - next.tv_nsec = (NSEC_PER_SEC / 2) - now.tv_nsec; + next.tv_nsec = (NSEC_PER_SEC / 2) - now.tv_nsec - (TICK_NSEC / 2); if (next.tv_nsec <= 0) next.tv_nsec += NSEC_PER_SEC; -- cgit v1.2.3 From dfb512ec4834116124da61d6c1ee10fd0aa32bd6 Mon Sep 17 00:00:00 2001 From: Max Krasnyansky Date: Fri, 29 Aug 2008 13:11:41 -0700 Subject: sched: arch_reinit_sched_domains() must destroy domains to force rebuild What I realized recently is that calling rebuild_sched_domains() in arch_reinit_sched_domains() by itself is not enough when cpusets are enabled. partition_sched_domains() code is trying to avoid unnecessary domain rebuilds and will not actually rebuild anything if new domain masks match the old ones. What this means is that doing echo 1 > /sys/devices/system/cpu/sched_mc_power_savings on a system with cpusets enabled will not take affect untill something changes in the cpuset setup (ie new sets created or deleted). This patch fixes restore correct behaviour where domains must be rebuilt in order to enable MC powersaving flags. Test on quad-core Core2 box with both CONFIG_CPUSETS and !CONFIG_CPUSETS. Also tested on dual-core Core2 laptop. Lockdep is happy and things are working as expected. Signed-off-by: Max Krasnyansky Tested-by: Vaidyanathan Srinivasan Signed-off-by: Ingo Molnar --- kernel/sched.c | 19 +++++++++++++------ 1 file changed, 13 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/sched.c b/kernel/sched.c index d601fb0406ca..d72ee9a0eacd 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -7589,24 +7589,27 @@ static int dattrs_equal(struct sched_domain_attr *cur, int idx_cur, * and partition_sched_domains() will fallback to the single partition * 'fallback_doms', it also forces the domains to be rebuilt. * + * If doms_new==NULL it will be replaced with cpu_online_map. + * ndoms_new==0 is a special case for destroying existing domains. + * It will not create the default domain. + * * Call with hotplug lock held */ void partition_sched_domains(int ndoms_new, cpumask_t *doms_new, struct sched_domain_attr *dattr_new) { - int i, j; + int i, j, n; mutex_lock(&sched_domains_mutex); /* always unregister in case we don't destroy any domains */ unregister_sched_domain_sysctl(); - if (doms_new == NULL) - ndoms_new = 0; + n = doms_new ? ndoms_new : 0; /* Destroy deleted domains */ for (i = 0; i < ndoms_cur; i++) { - for (j = 0; j < ndoms_new; j++) { + for (j = 0; j < n; j++) { if (cpus_equal(doms_cur[i], doms_new[j]) && dattrs_equal(dattr_cur, i, dattr_new, j)) goto match1; @@ -7619,7 +7622,6 @@ match1: if (doms_new == NULL) { ndoms_cur = 0; - ndoms_new = 1; doms_new = &fallback_doms; cpus_andnot(doms_new[0], cpu_online_map, cpu_isolated_map); dattr_new = NULL; @@ -7656,8 +7658,13 @@ match2: int arch_reinit_sched_domains(void) { get_online_cpus(); + + /* Destroy domains first to force the rebuild */ + partition_sched_domains(0, NULL, NULL); + rebuild_sched_domains(); put_online_cpus(); + return 0; } @@ -7741,7 +7748,7 @@ static int update_sched_domains(struct notifier_block *nfb, case CPU_ONLINE_FROZEN: case CPU_DEAD: case CPU_DEAD_FROZEN: - partition_sched_domains(0, NULL, NULL); + partition_sched_domains(1, NULL, NULL); return NOTIFY_OK; default: -- cgit v1.2.3 From 61c22c34c6f80a8e89cff5ff717627c54cc14fd4 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 9 Sep 2008 21:38:57 +0200 Subject: clockevents: remove WARN_ON which was used to gather information The issue of the endless reprogramming loop due to a too small min_delta_ns was fixed with the previous updates of the clock events code, but we had no information about the spread of this problem. I added a WARN_ON to get automated information via kerneloops.org and to get some direct reports, which allowed me to analyse the affected machines. The WARN_ON has served its purpose and would be annoying for a release kernel. Remove it and just keep the information about the increase of the min_delta_ns value. Signed-off-by: Thomas Gleixner --- kernel/time/tick-oneshot.c | 18 ++++++++---------- 1 file changed, 8 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-oneshot.c b/kernel/time/tick-oneshot.c index 2e35501e61dd..2e8de678e767 100644 --- a/kernel/time/tick-oneshot.c +++ b/kernel/time/tick-oneshot.c @@ -43,19 +43,17 @@ int tick_dev_program_event(struct clock_event_device *dev, ktime_t expires, * and emit a warning. */ if (++i > 2) { - printk(KERN_WARNING "CE: __tick_program_event of %s is " - "stuck %llx %llx\n", dev->name ? dev->name : "?", - now.tv64, expires.tv64); - printk(KERN_WARNING - "CE: increasing min_delta_ns %ld to %ld nsec\n", - dev->min_delta_ns, dev->min_delta_ns << 1); - WARN_ON(1); - - /* Double the min. delta and try again */ + /* Increase the min. delta and try again */ if (!dev->min_delta_ns) dev->min_delta_ns = 5000; else - dev->min_delta_ns <<= 1; + dev->min_delta_ns += dev->min_delta_ns >> 1; + + printk(KERN_WARNING + "CE: %s increasing min_delta_ns to %lu nsec\n", + dev->name ? dev->name : "?", + dev->min_delta_ns << 1); + i = 0; } -- cgit v1.2.3