From 2a01bb3885c9145dbb7583d5aa5f5d5504f6f46f Mon Sep 17 00:00:00 2001 From: Kyle McMartin Date: Wed, 11 Apr 2012 08:15:29 -0400 Subject: panic: Make panic_on_oops configurable Several distros set this by default by patching panic_on_oops. It seems to fit with the BOOTPARAM_{HARD,SOFT}_PANIC options though, so let's add a Kconfig entry and reduce some more upstream delta. Signed-off-by: Kyle McMartin Cc: Andrew Morton Cc: Linus Torvalds Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/20120411121529.GH26688@redacted.bos.redhat.com Signed-off-by: Ingo Molnar --- kernel/panic.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/panic.c b/kernel/panic.c index 8ed89a175d79..b6215b7ce99d 100644 --- a/kernel/panic.c +++ b/kernel/panic.c @@ -27,7 +27,7 @@ #define PANIC_TIMER_STEP 100 #define PANIC_BLINK_SPD 18 -int panic_on_oops; +int panic_on_oops = CONFIG_PANIC_ON_OOPS_VALUE; static unsigned long tainted_mask; static int pause_on_oops; static int pause_on_oops_flag; -- cgit v1.2.3 From 62be73eafaa045d3233337303fb140f7f8a61135 Mon Sep 17 00:00:00 2001 From: Seiji Aguchi Date: Tue, 15 May 2012 17:35:09 -0400 Subject: kdump: Execute kmsg_dump(KMSG_DUMP_PANIC) after smp_send_stop() This patch moves kmsg_dump(KMSG_DUMP_PANIC) below smp_send_stop(), to serialize the crash-logging process via smp_send_stop() and to thus retrieve a more stable crash image of all CPUs stopped. Signed-off-by: Seiji Aguchi Acked-by: Don Zickus Cc: dle-develop@lists.sourceforge.net Cc: Satoru Moriya Cc: Tony Luck Cc: a.p.zijlstra@chello.nl Link: http://lkml.kernel.org/r/5C4C569E8A4B9B42A84A977CF070A35B2E4D7A5CE2@USINDEVS01.corp.hds.com Signed-off-by: Ingo Molnar --- kernel/panic.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/panic.c b/kernel/panic.c index b6215b7ce99d..d2a5f4ecc6dd 100644 --- a/kernel/panic.c +++ b/kernel/panic.c @@ -108,8 +108,6 @@ void panic(const char *fmt, ...) */ crash_kexec(NULL); - kmsg_dump(KMSG_DUMP_PANIC); - /* * Note smp_send_stop is the usual smp shutdown function, which * unfortunately means it may not be hardened to work in a panic @@ -117,6 +115,8 @@ void panic(const char *fmt, ...) */ smp_send_stop(); + kmsg_dump(KMSG_DUMP_PANIC); + atomic_notifier_call_chain(&panic_notifier_list, 0, buf); bust_spinlocks(0); -- cgit v1.2.3 From 23812b9d9e497580d38c62ebdc6f308733b0a32a Mon Sep 17 00:00:00 2001 From: Ning Jiang Date: Tue, 22 May 2012 00:19:20 +0800 Subject: genirq: Add IRQS_PENDING for nested and simple irq Every interrupt which is an active wakeup source needs the ability to abort suspend if there is a pending irq. Right now only edge and level irqs can do that. | +---------+ | INTC | +---------+ | GPIO_IRQ +------------+ | gpio-exp | +------------+ | | GPIO0_IRQ GPIO1_IRQ In the above diagram, gpio expander has irq number GPIO_IRQ, it is connected with two sub GPIO pins, GPIO0 and GPIO1. During suspend, we set IRQF_NO_SUSPEND for GPIO_IRQ so that gpio expander driver can handle the sub irq GPIO0_IRQ and GPIO1_IRQ, and these two irqs themselves can further be handled by simple or nested irq in some drivers(typically gpio and mfd driver). If they are used as wakeup sources during suspend, we want them to be able to abort suspend too. Setting IRQS_PENDING flag in handle_nested_irq() and handle_simple_irq() when the irq is disabled allows check_wakeup_irqs() to identify such irqs as source for aborting suspend. Signed-off-by: Ning Jiang Cc: rjw@sisk.pl Link: http://lkml.kernel.org/r/CAH3Oq6T905%2B3fkF43NAMMFvJvq7dsk_so6T2vQ8ZJrA5xiU3YA@mail.gmail.com Signed-off-by: Thomas Gleixner --- kernel/irq/chip.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index fc275e4f629b..eebd6d5cfb44 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -275,8 +275,10 @@ void handle_nested_irq(unsigned int irq) kstat_incr_irqs_this_cpu(irq, desc); action = desc->action; - if (unlikely(!action || irqd_irq_disabled(&desc->irq_data))) + if (unlikely(!action || irqd_irq_disabled(&desc->irq_data))) { + desc->istate |= IRQS_PENDING; goto out_unlock; + } irqd_set(&desc->irq_data, IRQD_IRQ_INPROGRESS); raw_spin_unlock_irq(&desc->lock); @@ -324,8 +326,10 @@ handle_simple_irq(unsigned int irq, struct irq_desc *desc) desc->istate &= ~(IRQS_REPLAY | IRQS_WAITING); kstat_incr_irqs_this_cpu(irq, desc); - if (unlikely(!desc->action || irqd_irq_disabled(&desc->irq_data))) + if (unlikely(!desc->action || irqd_irq_disabled(&desc->irq_data))) { + desc->istate |= IRQS_PENDING; goto out_unlock; + } handle_irq_event(desc); -- cgit v1.2.3 From 818b0f3bfb236ae66cac3ff38e86b9e47f24b7aa Mon Sep 17 00:00:00 2001 From: Jiang Liu Date: Fri, 30 Mar 2012 23:11:34 +0800 Subject: genirq: Introduce irq_do_set_affinity() to reduce duplicated code All invocations of chip->irq_set_affinity() are doing the same return value checks. Let them all use a common function. [ tglx: removed the silly likely while at it ] Signed-off-by: Jiang Liu Cc: Jiang Liu Cc: Keping Chen Link: http://lkml.kernel.org/r/1333120296-13563-3-git-send-email-jiang.liu@huawei.com Signed-off-by: Thomas Gleixner --- kernel/irq/internals.h | 3 +++ kernel/irq/manage.c | 39 ++++++++++++++++++++++----------------- kernel/irq/migration.c | 13 ++----------- 3 files changed, 27 insertions(+), 28 deletions(-) (limited to 'kernel') diff --git a/kernel/irq/internals.h b/kernel/irq/internals.h index 8e5c56b3b7d9..001fa5bab490 100644 --- a/kernel/irq/internals.h +++ b/kernel/irq/internals.h @@ -101,6 +101,9 @@ extern int irq_select_affinity_usr(unsigned int irq, struct cpumask *mask); extern void irq_set_thread_affinity(struct irq_desc *desc); +extern int irq_do_set_affinity(struct irq_data *data, + const struct cpumask *dest, bool force); + /* Inline functions for support of irq chips on slow busses */ static inline void chip_bus_lock(struct irq_desc *desc) { diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index bb32326afe87..a1b903380bcf 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -139,6 +139,25 @@ static inline void irq_get_pending(struct cpumask *mask, struct irq_desc *desc) { } #endif +int irq_do_set_affinity(struct irq_data *data, const struct cpumask *mask, + bool force) +{ + struct irq_desc *desc = irq_data_to_desc(data); + struct irq_chip *chip = irq_data_get_irq_chip(data); + int ret; + + ret = chip->irq_set_affinity(data, mask, false); + switch (ret) { + case IRQ_SET_MASK_OK: + cpumask_copy(data->affinity, mask); + case IRQ_SET_MASK_OK_NOCOPY: + irq_set_thread_affinity(desc); + ret = 0; + } + + return ret; +} + int __irq_set_affinity_locked(struct irq_data *data, const struct cpumask *mask) { struct irq_chip *chip = irq_data_get_irq_chip(data); @@ -149,14 +168,7 @@ int __irq_set_affinity_locked(struct irq_data *data, const struct cpumask *mask) return -EINVAL; if (irq_can_move_pcntxt(data)) { - ret = chip->irq_set_affinity(data, mask, false); - switch (ret) { - case IRQ_SET_MASK_OK: - cpumask_copy(data->affinity, mask); - case IRQ_SET_MASK_OK_NOCOPY: - irq_set_thread_affinity(desc); - ret = 0; - } + ret = irq_do_set_affinity(data, mask, false); } else { irqd_set_move_pending(data); irq_copy_pending(desc, mask); @@ -280,9 +292,8 @@ EXPORT_SYMBOL_GPL(irq_set_affinity_notifier); static int setup_affinity(unsigned int irq, struct irq_desc *desc, struct cpumask *mask) { - struct irq_chip *chip = irq_desc_get_chip(desc); struct cpumask *set = irq_default_affinity; - int ret, node = desc->irq_data.node; + int node = desc->irq_data.node; /* Excludes PER_CPU and NO_BALANCE interrupts */ if (!irq_can_set_affinity(irq)) @@ -308,13 +319,7 @@ setup_affinity(unsigned int irq, struct irq_desc *desc, struct cpumask *mask) if (cpumask_intersects(mask, nodemask)) cpumask_and(mask, mask, nodemask); } - ret = chip->irq_set_affinity(&desc->irq_data, mask, false); - switch (ret) { - case IRQ_SET_MASK_OK: - cpumask_copy(desc->irq_data.affinity, mask); - case IRQ_SET_MASK_OK_NOCOPY: - irq_set_thread_affinity(desc); - } + irq_do_set_affinity(&desc->irq_data, mask, false); return 0; } #else diff --git a/kernel/irq/migration.c b/kernel/irq/migration.c index c3c89751b327..ca3f4aaff707 100644 --- a/kernel/irq/migration.c +++ b/kernel/irq/migration.c @@ -42,17 +42,8 @@ void irq_move_masked_irq(struct irq_data *idata) * For correct operation this depends on the caller * masking the irqs. */ - if (likely(cpumask_any_and(desc->pending_mask, cpu_online_mask) - < nr_cpu_ids)) { - int ret = chip->irq_set_affinity(&desc->irq_data, - desc->pending_mask, false); - switch (ret) { - case IRQ_SET_MASK_OK: - cpumask_copy(desc->irq_data.affinity, desc->pending_mask); - case IRQ_SET_MASK_OK_NOCOPY: - irq_set_thread_affinity(desc); - } - } + if (cpumask_any_and(desc->pending_mask, cpu_online_mask) < nr_cpu_ids) + irq_do_set_affinity(&desc->irq_data, desc->pending_mask, false); cpumask_clear(desc->pending_mask); } -- cgit v1.2.3 From ee74d13229fb606353ff56f4927fa93b37e95bbe Mon Sep 17 00:00:00 2001 From: "Srivatsa S. Bhat" Date: Thu, 24 May 2012 20:40:55 +0530 Subject: smpboot, idle: Optimize calls to smp_processor_id() in idle_threads_init() While trying to initialize idle threads for all cpus, idle_threads_init() calls smp_processor_id() in a loop, which is unnecessary. The intent is to initialize idle threads for all non-boot cpus. So just use a variable to note the boot cpu and use it in the loop. Signed-off-by: Srivatsa S. Bhat Cc: suresh.b.siddha@intel.com Cc: venki@google.com Cc: nikunj@linux.vnet.ibm.com Link: http://lkml.kernel.org/r/20120524151055.2549.64309.stgit@srivatsabhat.in.ibm.com Signed-off-by: Thomas Gleixner --- kernel/smpboot.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/smpboot.c b/kernel/smpboot.c index e1a797e028a3..0f2162f808a7 100644 --- a/kernel/smpboot.c +++ b/kernel/smpboot.c @@ -52,10 +52,12 @@ static inline void idle_init(unsigned int cpu) */ void __init idle_threads_init(void) { - unsigned int cpu; + unsigned int cpu, boot_cpu; + + boot_cpu = smp_processor_id(); for_each_possible_cpu(cpu) { - if (cpu != smp_processor_id()) + if (cpu != boot_cpu) idle_init(cpu); } } -- cgit v1.2.3 From 4a70d2d9909b43ed88043b98cabe2c7fbd563021 Mon Sep 17 00:00:00 2001 From: "Srivatsa S. Bhat" Date: Thu, 24 May 2012 20:41:00 +0530 Subject: smpboot, idle: Fix comment mismatch over idle_threads_init() The comment over idle_threads_init() really talks about the functionality of idle_init(). Move that comment to idle_init(), and add a suitable comment over idle_threads_init(). Signed-off-by: Srivatsa S. Bhat Cc: suresh.b.siddha@intel.com Cc: venki@google.com Cc: nikunj@linux.vnet.ibm.com Link: http://lkml.kernel.org/r/20120524151100.2549.66501.stgit@srivatsabhat.in.ibm.com Signed-off-by: Thomas Gleixner --- kernel/smpboot.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/smpboot.c b/kernel/smpboot.c index 0f2162f808a7..98f60c5caa1b 100644 --- a/kernel/smpboot.c +++ b/kernel/smpboot.c @@ -31,6 +31,12 @@ void __init idle_thread_set_boot_cpu(void) per_cpu(idle_threads, smp_processor_id()) = current; } +/** + * idle_init - Initialize the idle thread for a cpu + * @cpu: The cpu for which the idle thread should be initialized + * + * Creates the thread if it does not exist. + */ static inline void idle_init(unsigned int cpu) { struct task_struct *tsk = per_cpu(idle_threads, cpu); @@ -45,10 +51,7 @@ static inline void idle_init(unsigned int cpu) } /** - * idle_thread_init - Initialize the idle thread for a cpu - * @cpu: The cpu for which the idle thread should be initialized - * - * Creates the thread if it does not exist. + * idle_threads_init - Initialize idle threads for all cpus */ void __init idle_threads_init(void) { -- cgit v1.2.3 From 5307c9556bc17e3cd26d4e94fc3b2565921834de Mon Sep 17 00:00:00 2001 From: Mike Galbraith Date: Tue, 8 May 2012 12:20:58 +0200 Subject: tick: Add tick skew boot option Let the user decide whether power consumption or jitter is the more important consideration for their machines. Quoting removal commit af5ab277ded04bd9bc6b048c5a2f0e7d70ef0867: "Historically, Linux has tried to make the regular timer tick on the various CPUs not happen at the same time, to avoid contention on xtime_lock. Nowadays, with the tickless kernel, this contention no longer happens since time keeping and updating are done differently. In addition, this skew is actually hurting power consumption in a measurable way on many-core systems." Problems: - Contrary to the above, systems do encounter contention on both xtime_lock and RCU structure locks when the tick is synchronized. - Moderate sized RT systems suffer intolerable jitter due to the tick being synchronized. - SGI reports the same for their large systems. - Fully utilized systems reap no power saving benefit from skew removal, but do suffer from resulting induced lock contention. - 0209f649 rcu: limit rcu_node leaf-level fanout This patch was born to combat lock contention which testing showed to have been _induced by_ skew removal. Skew the tick, contention disappeared virtually completely. Signed-off-by: Mike Galbraith Link: http://lkml.kernel.org/r/1336472458.21924.78.camel@marge.simpson.net Signed-off-by: Thomas Gleixner --- kernel/time/tick-sched.c | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) (limited to 'kernel') diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 6a3a5b9ff561..4eddbb5ea9c5 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -814,6 +814,8 @@ static enum hrtimer_restart tick_sched_timer(struct hrtimer *timer) return HRTIMER_RESTART; } +static int sched_skew_tick; + /** * tick_setup_sched_timer - setup the tick emulation timer */ @@ -831,6 +833,14 @@ void tick_setup_sched_timer(void) /* Get the next period (per cpu) */ hrtimer_set_expires(&ts->sched_timer, tick_init_jiffy_update()); + /* Offset the tick to avert xtime_lock contention. */ + if (sched_skew_tick) { + u64 offset = ktime_to_ns(tick_period) >> 1; + do_div(offset, num_possible_cpus()); + offset *= smp_processor_id(); + hrtimer_add_expires_ns(&ts->sched_timer, offset); + } + for (;;) { hrtimer_forward(&ts->sched_timer, now, tick_period); hrtimer_start_expires(&ts->sched_timer, @@ -910,3 +920,11 @@ int tick_check_oneshot_change(int allow_nohz) tick_nohz_switch_to_nohz(); return 0; } + +static int __init skew_tick(char *str) +{ + get_option(&str, &sched_skew_tick); + + return 0; +} +early_param("skew_tick", skew_tick); -- cgit v1.2.3 From e5400321a6f15ce0fe77c8455954f213ef7dcc54 Mon Sep 17 00:00:00 2001 From: Magnus Damm Date: Wed, 9 May 2012 23:39:34 +0900 Subject: clockevents: Make clockevents_config() a global symbol Make clockevents_config() into a global symbol to allow it to be used by compiled-in clockevent drivers. This is needed by drivers that want to update the timer frequency after registration time. Signed-off-by: Magnus Damm Tested-by: Simon Horman Cc: arnd@arndb.de Cc: johnstul@us.ibm.com Cc: rjw@sisk.pl Cc: lethal@linux-sh.org Cc: gregkh@linuxfoundation.org Cc: olof@lixom.net Cc: Magnus Damm Link: http://lkml.kernel.org/r/20120509143934.27521.46553.sendpatchset@w520 Signed-off-by: Thomas Gleixner --- kernel/time/clockevents.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/time/clockevents.c b/kernel/time/clockevents.c index 9cd928f7a7c6..7e1ce012a851 100644 --- a/kernel/time/clockevents.c +++ b/kernel/time/clockevents.c @@ -297,8 +297,7 @@ void clockevents_register_device(struct clock_event_device *dev) } EXPORT_SYMBOL_GPL(clockevents_register_device); -static void clockevents_config(struct clock_event_device *dev, - u32 freq) +void clockevents_config(struct clock_event_device *dev, u32 freq) { u64 sec; -- cgit v1.2.3 From 62cf20b32aee4ae889a2eb40fd41c0eab73de970 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 25 May 2012 14:08:57 +0200 Subject: tick: Move skew_tick option into the HIGH_RES_TIMER section commit 5307c95 (tick: Add tick skew boot option) broke the !CONFIG_HIGH_RES_TIMERS build. Move the boot option parsing into the CONFIG_HIGH_RES_TIMERS section. Reported-by: Ingo Molnar Signed-off-by: Thomas Gleixner Cc: Mike Galbraith --- kernel/time/tick-sched.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 4eddbb5ea9c5..efd386667536 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -816,6 +816,14 @@ static enum hrtimer_restart tick_sched_timer(struct hrtimer *timer) static int sched_skew_tick; +static int __init skew_tick(char *str) +{ + get_option(&str, &sched_skew_tick); + + return 0; +} +early_param("skew_tick", skew_tick); + /** * tick_setup_sched_timer - setup the tick emulation timer */ @@ -920,11 +928,3 @@ int tick_check_oneshot_change(int allow_nohz) tick_nohz_switch_to_nohz(); return 0; } - -static int __init skew_tick(char *str) -{ - get_option(&str, &sched_skew_tick); - - return 0; -} -early_param("skew_tick", skew_tick); -- cgit v1.2.3 From fa980ca87d15bb8a1317853f257a505990f3ffde Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 24 May 2012 08:24:39 -0700 Subject: cgroup: superblock can't be released with active dentries 48ddbe1946 "cgroup: make css->refcnt clearing on cgroup removal optional" allowed a css to linger after the associated cgroup is removed. As a css holds a reference on the cgroup's dentry, it means that cgroup dentries may linger for a while. cgroup_create() does grab an active reference on the superblock to prevent it from going away while there are !root cgroups; however, the reference is put from cgroup_diput() which is invoked on cgroup removal, so cgroup dentries which are removed but persisting due to lingering csses already have released their superblock active refs allowing superblock to be killed while those dentries are around. Given the right condition, this makes cgroup_kill_sb() call kill_litter_super() with dentries with non-zero d_count leading to BUG() in shrink_dcache_for_umount_subtree(). Fix it by adding cgroup_dops->d_release() operation and moving deactivate_super() to it. cgroup_diput() now marks dentry->d_fsdata with itself if superblock should be deactivated and cgroup_d_release() deactivates the superblock on dentry release. Signed-off-by: Tejun Heo Reported-by: Sasha Levin Tested-by: Sasha Levin LKML-Reference: Acked-by: Li Zefan --- kernel/cgroup.c | 17 ++++++++++++++--- 1 file changed, 14 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/cgroup.c b/kernel/cgroup.c index ad8eae5bb801..e887b55f1f29 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -896,10 +896,13 @@ static void cgroup_diput(struct dentry *dentry, struct inode *inode) mutex_unlock(&cgroup_mutex); /* - * Drop the active superblock reference that we took when we - * created the cgroup + * We want to drop the active superblock reference from the + * cgroup creation after all the dentry refs are gone - + * kill_sb gets mighty unhappy otherwise. Mark + * dentry->d_fsdata with cgroup_diput() to tell + * cgroup_d_release() to call deactivate_super(). */ - deactivate_super(cgrp->root->sb); + dentry->d_fsdata = cgroup_diput; /* * if we're getting rid of the cgroup, refcount should ensure @@ -925,6 +928,13 @@ static int cgroup_delete(const struct dentry *d) return 1; } +static void cgroup_d_release(struct dentry *dentry) +{ + /* did cgroup_diput() tell me to deactivate super? */ + if (dentry->d_fsdata == cgroup_diput) + deactivate_super(dentry->d_sb); +} + static void remove_dir(struct dentry *d) { struct dentry *parent = dget(d->d_parent); @@ -1532,6 +1542,7 @@ static int cgroup_get_rootdir(struct super_block *sb) static const struct dentry_operations cgroup_dops = { .d_iput = cgroup_diput, .d_delete = cgroup_delete, + .d_release = cgroup_d_release, }; struct inode *inode = -- cgit v1.2.3 From 5aaa0b7a2ed5b12692c9ffb5222182bd558d3146 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 17 May 2012 17:15:29 +0200 Subject: sched/nohz: Fix rq->cpu_load calculations some more Follow up on commit 556061b00 ("sched/nohz: Fix rq->cpu_load[] calculations") since while that fixed the busy case it regressed the mostly idle case. Add a callback from the nohz exit to also age the rq->cpu_load[] array. This closes the hole where either there was no nohz load balance pass during the nohz, or there was a 'significant' amount of idle time between the last nohz balance and the nohz exit. So we'll update unconditionally from the tick to not insert any accidental 0 load periods while busy, and we try and catch up from nohz idle balance and nohz exit. Both these are still prone to missing a jiffy, but that has always been the case. Signed-off-by: Peter Zijlstra Cc: pjt@google.com Cc: Venkatesh Pallipadi Link: http://lkml.kernel.org/n/tip-kt0trz0apodbf84ucjfdbr1a@git.kernel.org Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 53 +++++++++++++++++++++++++++++++++++++++--------- kernel/time/tick-sched.c | 1 + 2 files changed, 44 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 39eb6011bc38..75844a8f9aeb 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -2517,25 +2517,32 @@ static void __update_cpu_load(struct rq *this_rq, unsigned long this_load, sched_avg_update(this_rq); } +#ifdef CONFIG_NO_HZ +/* + * There is no sane way to deal with nohz on smp when using jiffies because the + * cpu doing the jiffies update might drift wrt the cpu doing the jiffy reading + * causing off-by-one errors in observed deltas; {0,2} instead of {1,1}. + * + * Therefore we cannot use the delta approach from the regular tick since that + * would seriously skew the load calculation. However we'll make do for those + * updates happening while idle (nohz_idle_balance) or coming out of idle + * (tick_nohz_idle_exit). + * + * This means we might still be one tick off for nohz periods. + */ + /* * Called from nohz_idle_balance() to update the load ratings before doing the * idle balance. */ void update_idle_cpu_load(struct rq *this_rq) { - unsigned long curr_jiffies = jiffies; + unsigned long curr_jiffies = ACCESS_ONCE(jiffies); unsigned long load = this_rq->load.weight; unsigned long pending_updates; /* - * Bloody broken means of dealing with nohz, but better than nothing.. - * jiffies is updated by one cpu, another cpu can drift wrt the jiffy - * update and see 0 difference the one time and 2 the next, even though - * we ticked at roughtly the same rate. - * - * Hence we only use this from nohz_idle_balance() and skip this - * nonsense when called from the scheduler_tick() since that's - * guaranteed a stable rate. + * bail if there's load or we're actually up-to-date. */ if (load || curr_jiffies == this_rq->last_load_update_tick) return; @@ -2546,13 +2553,39 @@ void update_idle_cpu_load(struct rq *this_rq) __update_cpu_load(this_rq, load, pending_updates); } +/* + * Called from tick_nohz_idle_exit() -- try and fix up the ticks we missed. + */ +void update_cpu_load_nohz(void) +{ + struct rq *this_rq = this_rq(); + unsigned long curr_jiffies = ACCESS_ONCE(jiffies); + unsigned long pending_updates; + + if (curr_jiffies == this_rq->last_load_update_tick) + return; + + raw_spin_lock(&this_rq->lock); + pending_updates = curr_jiffies - this_rq->last_load_update_tick; + if (pending_updates) { + this_rq->last_load_update_tick = curr_jiffies; + /* + * We were idle, this means load 0, the current load might be + * !0 due to remote wakeups and the sort. + */ + __update_cpu_load(this_rq, 0, pending_updates); + } + raw_spin_unlock(&this_rq->lock); +} +#endif /* CONFIG_NO_HZ */ + /* * Called from scheduler_tick() */ static void update_cpu_load_active(struct rq *this_rq) { /* - * See the mess in update_idle_cpu_load(). + * See the mess around update_idle_cpu_load() / update_cpu_load_nohz(). */ this_rq->last_load_update_tick = jiffies; __update_cpu_load(this_rq, this_rq->load.weight, 1); diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 6a3a5b9ff561..0c927cd85345 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -576,6 +576,7 @@ void tick_nohz_idle_exit(void) /* Update jiffies first */ select_nohz_load_balancer(0); tick_do_update_jiffies64(now); + update_cpu_load_nohz(); #ifndef CONFIG_VIRT_CPU_ACCOUNTING /* -- cgit v1.2.3 From 2ea45800d8e1c3c51c45a233d6bd6289a297a386 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 25 May 2012 09:26:43 +0200 Subject: sched: Don't try allocating memory from offline nodes Allocators don't appreciate it when you try and allocate memory from offline nodes. Reported-and-tested-by: Tony Luck Reported-and-tested-by: Anton Blanchard Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-epfc1io9whb7o22bcujf31vn@git.kernel.org Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 75844a8f9aeb..55733616baaa 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -6436,7 +6436,7 @@ static void sched_init_numa(void) return; for (j = 0; j < nr_node_ids; j++) { - struct cpumask *mask = kzalloc_node(cpumask_size(), GFP_KERNEL, j); + struct cpumask *mask = kzalloc(cpumask_size(), GFP_KERNEL); if (!mask) return; -- cgit v1.2.3 From 74a5ce20e6eeeb3751340b390e7ac1d1d07bbf55 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 23 May 2012 18:00:43 +0200 Subject: sched: Fix SD_OVERLAP SD_OVERLAP exists to allow overlapping groups, overlapping groups appear in NUMA topologies that aren't fully connected. The typical result of not fully connected NUMA is that each cpu (or rather node) will have different spans for a particular distance. However due to how sched domains are traversed -- only the first cpu in the mask goes one level up -- the next level only cares about the spans of the cpus that went up. Due to this two things were observed to be broken: - build_overlap_sched_groups() -- since its possible the cpu we're building the groups for exists in multiple (or all) groups, the selection criteria of the first group didn't ensure there was a cpu for which is was true that cpumask_first(span) == cpu. Thus load- balancing would terminate. - update_group_power() -- assumed that the cpu span of the first group of the domain was covered by all groups of the child domain. The above explains why this isn't true, so deal with it. Signed-off-by: Peter Zijlstra Cc: David Rientjes Link: http://lkml.kernel.org/r/1337788843.9783.14.camel@laptop Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 7 +++++-- kernel/sched/fair.c | 25 ++++++++++++++++++++----- 2 files changed, 25 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 55733616baaa..3a69374fb427 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -6030,11 +6030,14 @@ build_overlap_sched_groups(struct sched_domain *sd, int cpu) cpumask_or(covered, covered, sg_span); - sg->sgp = *per_cpu_ptr(sdd->sgp, cpumask_first(sg_span)); + sg->sgp = *per_cpu_ptr(sdd->sgp, i); atomic_inc(&sg->sgp->ref); - if (cpumask_test_cpu(cpu, sg_span)) + if ((!groups && cpumask_test_cpu(cpu, sg_span)) || + cpumask_first(sg_span) == cpu) { + WARN_ON_ONCE(!cpumask_test_cpu(cpu, sg_span)); groups = sg; + } if (!first) first = sg; diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 940e6d17cf96..f0380d4987b3 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -3574,11 +3574,26 @@ void update_group_power(struct sched_domain *sd, int cpu) power = 0; - group = child->groups; - do { - power += group->sgp->power; - group = group->next; - } while (group != child->groups); + if (child->flags & SD_OVERLAP) { + /* + * SD_OVERLAP domains cannot assume that child groups + * span the current group. + */ + + for_each_cpu(cpu, sched_group_cpus(sdg)) + power += power_of(cpu); + } else { + /* + * !SD_OVERLAP domains can assume that child groups + * span the current group. + */ + + group = child->groups; + do { + power += group->sgp->power; + group = group->next; + } while (group != child->groups); + } sdg->sgp->power = power; } -- cgit v1.2.3 From b654f7de41b0e3903ee2b51d3b8db77fe52ce728 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 22 May 2012 14:04:28 +0200 Subject: sched: Make sure to not re-read variables after validation We could re-read rq->rt_avg after we validated it was smaller than total, invalidating the check and resulting in an unintended negative. Signed-off-by: Peter Zijlstra Cc: David Rientjes Link: http://lkml.kernel.org/r/1337688268.9698.29.camel@twins Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 15 +++++++++++---- 1 file changed, 11 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index f0380d4987b3..2b449a762074 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -3503,15 +3503,22 @@ unsigned long __weak arch_scale_smt_power(struct sched_domain *sd, int cpu) unsigned long scale_rt_power(int cpu) { struct rq *rq = cpu_rq(cpu); - u64 total, available; + u64 total, available, age_stamp, avg; - total = sched_avg_period() + (rq->clock - rq->age_stamp); + /* + * Since we're reading these variables without serialization make sure + * we read them once before doing sanity checks on them. + */ + age_stamp = ACCESS_ONCE(rq->age_stamp); + avg = ACCESS_ONCE(rq->rt_avg); + + total = sched_avg_period() + (rq->clock - age_stamp); - if (unlikely(total < rq->rt_avg)) { + if (unlikely(total < avg)) { /* Ensures that power won't end up being negative */ available = 0; } else { - available = total - rq->rt_avg; + available = total - avg; } if (unlikely((s64)total < SCHED_POWER_SCALE)) -- cgit v1.2.3 From 29baa7478ba47d746e3625c91d3b2afbf46b4312 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 23 Apr 2012 12:11:21 +0200 Subject: sched: Move nr_cpus_allowed out of 'struct sched_rt_entity' Since nr_cpus_allowed is used outside of sched/rt.c and wants to be used outside of there more, move it to a more natural site. Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-kr61f02y9brwzkh6x53pdptm@git.kernel.org Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 2 +- kernel/sched/fair.c | 2 +- kernel/sched/rt.c | 36 +++++++++++++++++++++--------------- 3 files changed, 23 insertions(+), 17 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 3a69374fb427..70cc36a6073f 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5015,7 +5015,7 @@ void do_set_cpus_allowed(struct task_struct *p, const struct cpumask *new_mask) p->sched_class->set_cpus_allowed(p, new_mask); cpumask_copy(&p->cpus_allowed, new_mask); - p->rt.nr_cpus_allowed = cpumask_weight(new_mask); + p->nr_cpus_allowed = cpumask_weight(new_mask); } /* diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 2b449a762074..b2a2d236f27b 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -2703,7 +2703,7 @@ select_task_rq_fair(struct task_struct *p, int sd_flag, int wake_flags) int want_sd = 1; int sync = wake_flags & WF_SYNC; - if (p->rt.nr_cpus_allowed == 1) + if (p->nr_cpus_allowed == 1) return prev_cpu; if (sd_flag & SD_BALANCE_WAKE) { diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c index c5565c3c515f..295da737b6fe 100644 --- a/kernel/sched/rt.c +++ b/kernel/sched/rt.c @@ -274,13 +274,16 @@ static void update_rt_migration(struct rt_rq *rt_rq) static void inc_rt_migration(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) { + struct task_struct *p; + if (!rt_entity_is_task(rt_se)) return; + p = rt_task_of(rt_se); rt_rq = &rq_of_rt_rq(rt_rq)->rt; rt_rq->rt_nr_total++; - if (rt_se->nr_cpus_allowed > 1) + if (p->nr_cpus_allowed > 1) rt_rq->rt_nr_migratory++; update_rt_migration(rt_rq); @@ -288,13 +291,16 @@ static void inc_rt_migration(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) static void dec_rt_migration(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) { + struct task_struct *p; + if (!rt_entity_is_task(rt_se)) return; + p = rt_task_of(rt_se); rt_rq = &rq_of_rt_rq(rt_rq)->rt; rt_rq->rt_nr_total--; - if (rt_se->nr_cpus_allowed > 1) + if (p->nr_cpus_allowed > 1) rt_rq->rt_nr_migratory--; update_rt_migration(rt_rq); @@ -1161,7 +1167,7 @@ enqueue_task_rt(struct rq *rq, struct task_struct *p, int flags) enqueue_rt_entity(rt_se, flags & ENQUEUE_HEAD); - if (!task_current(rq, p) && p->rt.nr_cpus_allowed > 1) + if (!task_current(rq, p) && p->nr_cpus_allowed > 1) enqueue_pushable_task(rq, p); inc_nr_running(rq); @@ -1225,7 +1231,7 @@ select_task_rq_rt(struct task_struct *p, int sd_flag, int flags) cpu = task_cpu(p); - if (p->rt.nr_cpus_allowed == 1) + if (p->nr_cpus_allowed == 1) goto out; /* For anything but wake ups, just return the task_cpu */ @@ -1260,9 +1266,9 @@ select_task_rq_rt(struct task_struct *p, int sd_flag, int flags) * will have to sort it out. */ if (curr && unlikely(rt_task(curr)) && - (curr->rt.nr_cpus_allowed < 2 || + (curr->nr_cpus_allowed < 2 || curr->prio <= p->prio) && - (p->rt.nr_cpus_allowed > 1)) { + (p->nr_cpus_allowed > 1)) { int target = find_lowest_rq(p); if (target != -1) @@ -1276,10 +1282,10 @@ out: static void check_preempt_equal_prio(struct rq *rq, struct task_struct *p) { - if (rq->curr->rt.nr_cpus_allowed == 1) + if (rq->curr->nr_cpus_allowed == 1) return; - if (p->rt.nr_cpus_allowed != 1 + if (p->nr_cpus_allowed != 1 && cpupri_find(&rq->rd->cpupri, p, NULL)) return; @@ -1395,7 +1401,7 @@ static void put_prev_task_rt(struct rq *rq, struct task_struct *p) * The previous task needs to be made eligible for pushing * if it is still active */ - if (on_rt_rq(&p->rt) && p->rt.nr_cpus_allowed > 1) + if (on_rt_rq(&p->rt) && p->nr_cpus_allowed > 1) enqueue_pushable_task(rq, p); } @@ -1408,7 +1414,7 @@ static int pick_rt_task(struct rq *rq, struct task_struct *p, int cpu) { if (!task_running(rq, p) && (cpu < 0 || cpumask_test_cpu(cpu, tsk_cpus_allowed(p))) && - (p->rt.nr_cpus_allowed > 1)) + (p->nr_cpus_allowed > 1)) return 1; return 0; } @@ -1464,7 +1470,7 @@ static int find_lowest_rq(struct task_struct *task) if (unlikely(!lowest_mask)) return -1; - if (task->rt.nr_cpus_allowed == 1) + if (task->nr_cpus_allowed == 1) return -1; /* No other targets possible */ if (!cpupri_find(&task_rq(task)->rd->cpupri, task, lowest_mask)) @@ -1586,7 +1592,7 @@ static struct task_struct *pick_next_pushable_task(struct rq *rq) BUG_ON(rq->cpu != task_cpu(p)); BUG_ON(task_current(rq, p)); - BUG_ON(p->rt.nr_cpus_allowed <= 1); + BUG_ON(p->nr_cpus_allowed <= 1); BUG_ON(!p->on_rq); BUG_ON(!rt_task(p)); @@ -1793,9 +1799,9 @@ static void task_woken_rt(struct rq *rq, struct task_struct *p) if (!task_running(rq, p) && !test_tsk_need_resched(rq->curr) && has_pushable_tasks(rq) && - p->rt.nr_cpus_allowed > 1 && + p->nr_cpus_allowed > 1 && rt_task(rq->curr) && - (rq->curr->rt.nr_cpus_allowed < 2 || + (rq->curr->nr_cpus_allowed < 2 || rq->curr->prio <= p->prio)) push_rt_tasks(rq); } @@ -1817,7 +1823,7 @@ static void set_cpus_allowed_rt(struct task_struct *p, * Only update if the process changes its state from whether it * can migrate or not. */ - if ((p->rt.nr_cpus_allowed > 1) == (weight > 1)) + if ((p->nr_cpus_allowed > 1) == (weight > 1)) return; rq = task_rq(p); -- cgit v1.2.3 From 454c79999f7eaedcdf4c15c449e43902980cbdf5 Mon Sep 17 00:00:00 2001 From: Colin Cross Date: Wed, 16 May 2012 21:34:23 -0700 Subject: sched/rt: Fix SCHED_RR across cgroups task_tick_rt() has an optimization to only reschedule SCHED_RR tasks if they were the only element on their rq. However, with cgroups a SCHED_RR task could be the only element on its per-cgroup rq but still be competing with other SCHED_RR tasks in its parent's cgroup. In this case, the SCHED_RR task in the child cgroup would never yield at the end of its timeslice. If the child cgroup rt_runtime_us was the same as the parent cgroup rt_runtime_us, the task in the parent cgroup would starve completely. Modify task_tick_rt() to check that the task is the only task on its rq, and that the each of the scheduling entities of its ancestors is also the only entity on its rq. Signed-off-by: Colin Cross Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1337229266-15798-1-git-send-email-ccross@android.com Signed-off-by: Ingo Molnar --- kernel/sched/rt.c | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c index 295da737b6fe..2a4e8dffbd6b 100644 --- a/kernel/sched/rt.c +++ b/kernel/sched/rt.c @@ -1985,6 +1985,8 @@ static void watchdog(struct rq *rq, struct task_struct *p) static void task_tick_rt(struct rq *rq, struct task_struct *p, int queued) { + struct sched_rt_entity *rt_se = &p->rt; + update_curr_rt(rq); watchdog(rq, p); @@ -2002,12 +2004,15 @@ static void task_tick_rt(struct rq *rq, struct task_struct *p, int queued) p->rt.time_slice = RR_TIMESLICE; /* - * Requeue to the end of queue if we are not the only element - * on the queue: + * Requeue to the end of queue if we (and all of our ancestors) are the + * only element on the queue */ - if (p->rt.run_list.prev != p->rt.run_list.next) { - requeue_task_rt(rq, p, 0); - set_tsk_need_resched(p); + for_each_sched_rt_entity(rt_se) { + if (rt_se->run_list.prev != rt_se->run_list.next) { + requeue_task_rt(rq, p, 0); + set_tsk_need_resched(p); + return; + } } } -- cgit v1.2.3 From 1292531f6f27af909e713671dd9cc3bcab8114b7 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Fri, 25 May 2012 15:41:54 +0900 Subject: sched: Make sched_feat_names const The strings sched_feat_names are never changed. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/4FBF29B2.9030904@ct.jp.nec.com Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 70cc36a6073f..c1679a098fc7 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -142,7 +142,7 @@ const_debug unsigned int sysctl_sched_features = #define SCHED_FEAT(name, enabled) \ #name , -static __read_mostly char *sched_feat_names[] = { +static const char * const sched_feat_names[] = { #include "features.h" NULL }; -- cgit v1.2.3 From 7997a456ef841bb78eb6f881d7cc2c17c2f9b35e Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Fri, 25 May 2012 15:42:47 +0900 Subject: sched: Remove the last NULL entry from sched_feat_names No need to have the last NULL entry. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/4FBF29E7.5020805@ct.jp.nec.com Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index c1679a098fc7..94d598ac5e64 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -144,7 +144,6 @@ const_debug unsigned int sysctl_sched_features = static const char * const sched_feat_names[] = { #include "features.h" - NULL }; #undef SCHED_FEAT -- cgit v1.2.3 From 6a4c96eef42f835734a82c6b512abf9881b7c55d Mon Sep 17 00:00:00 2001 From: Kamalesh Babulal Date: Wed, 23 May 2012 14:44:11 +0530 Subject: sched: Remove NULL assignment of dattr_cur Remove explicit NULL assignment of static pointer dattr_cur from init_sched_domains(). Signed-off-by: Kamalesh Babulal Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/20120523091411.GG5005@linux.vnet.ibm.com Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 94d598ac5e64..c46958e26121 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -6726,7 +6726,6 @@ static int init_sched_domains(const struct cpumask *cpu_map) if (!doms_cur) doms_cur = &fallback_doms; cpumask_andnot(doms_cur[0], cpu_map, cpu_isolated_map); - dattr_cur = NULL; err = build_sched_domains(doms_cur[0], NULL); register_sched_domain_sysctl(); -- cgit v1.2.3 From cb7225feec627e91d598198996429e9ee6804f8d Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 31 May 2012 14:51:44 +0900 Subject: perf: Remove duplicate invocation on perf_event_for_each The @func callback was invoked twice for group leader when perf_event_for_each() called. It seems the commit 75f937f24bd9 ("perf_counter: Fix ctx->mutex vs counter ->mutex inversion") made the mistake during the change. Signed-off-by: Namhyung Kim Acked-by: Peter Zijlstra Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1338443506-25009-1-git-send-email-namhyung.kim@lge.com Signed-off-by: Arnaldo Carvalho de Melo --- kernel/events/core.c | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/events/core.c b/kernel/events/core.c index 5b06cbbf6931..f85c0154b333 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -3181,7 +3181,6 @@ static void perf_event_for_each(struct perf_event *event, event = event->group_leader; perf_event_for_each_child(event, func); - func(event); list_for_each_entry(sibling, &event->sibling_list, group_entry) perf_event_for_each_child(sibling, func); mutex_unlock(&ctx->mutex); -- cgit v1.2.3 From fad0c66c4bb836d57a5f125ecd38bed653ca863a Mon Sep 17 00:00:00 2001 From: John Stultz Date: Wed, 30 May 2012 10:54:57 -0700 Subject: timekeeping: Fix CLOCK_MONOTONIC inconsistency during leapsecond Commit 6b43ae8a61 (ntp: Fix leap-second hrtimer livelock) broke the leapsecond update of CLOCK_MONOTONIC. The missing leapsecond update to wall_to_monotonic causes discontinuities in CLOCK_MONOTONIC. Adjust wall_to_monotonic when NTP inserted a leapsecond. Reported-by: Richard Cochran Signed-off-by: John Stultz Tested-by: Richard Cochran Cc: stable@kernel.org Link: http://lkml.kernel.org/r/1338400497-12420-1-git-send-email-john.stultz@linaro.org Signed-off-by: Thomas Gleixner --- kernel/time/timekeeping.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'kernel') diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index 6e46cacf5969..6f46a00a1e8a 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -962,6 +962,7 @@ static cycle_t logarithmic_accumulation(cycle_t offset, int shift) timekeeper.xtime.tv_sec++; leap = second_overflow(timekeeper.xtime.tv_sec); timekeeper.xtime.tv_sec += leap; + timekeeper.wall_to_monotonic.tv_sec -= leap; } /* Accumulate raw time */ @@ -1077,6 +1078,7 @@ static void update_wall_time(void) timekeeper.xtime.tv_sec++; leap = second_overflow(timekeeper.xtime.tv_sec); timekeeper.xtime.tv_sec += leap; + timekeeper.wall_to_monotonic.tv_sec -= leap; } timekeeping_update(false); -- cgit v1.2.3 From 10717dcde10d09f9fcee53a12a4236af1a82b484 Mon Sep 17 00:00:00 2001 From: Alex Shi Date: Wed, 6 Jun 2012 14:52:51 +0800 Subject: sched/numa: Load balance between remote nodes Commit cb83b629b ("sched/numa: Rewrite the CONFIG_NUMA sched domain support") removed the NODE sched domain and started checking if the node distance in SLIT table is farther than REMOTE_DISTANCE, if so, it will lose the load balance chance at exec/fork/wake_affine points. But actually, even the node distance is farther than REMOTE_DISTANCE. Modern CPUs also has QPI like connections, which ensures that memory access is not too slow between nodes. So the above change in behavior on NUMA machine causes a performance regression on various benchmarks: hackbench, tbench, netperf, oltp, etc. This patch will recover the scheduler behavior to old mode on all my Intel platforms: NHM EP/EX, WSM EP, SNB EP/EP4S, and thus fixes the perfromance regressions. (all of them just have 2 kinds distance, 10, 21) Signed-off-by: Alex Shi Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1338965571-9812-1-git-send-email-alex.shi@intel.com Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index c46958e26121..6546083af3e0 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -6321,7 +6321,7 @@ static int sched_domains_curr_level; static inline int sd_local_flags(int level) { - if (sched_domains_numa_distance[level] > REMOTE_DISTANCE) + if (sched_domains_numa_distance[level] > RECLAIM_DISTANCE) return 0; return SD_BALANCE_EXEC | SD_BALANCE_FORK | SD_WAKE_AFFINE; -- cgit v1.2.3 From 7f1b43936f0ecad14770634c021cf4a929aec74d Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 17 May 2012 21:19:46 +0200 Subject: sched/rt: Fix lockdep annotation within find_lock_lowest_rq() Roland Dreier reported spurious, hard to trigger lockdep warnings within the scheduler - without any real lockup. This bit gives us the right clue: > [89945.640512] [] double_lock_balance+0x5a/0x90 > [89945.640568] [] push_rt_task+0xc6/0x290 if you look at that code you'll find the double_lock_balance() in question is the one in find_lock_lowest_rq() [yay for inlining]. Now find_lock_lowest_rq() has a bug.. it fails to use double_unlock_balance() in one exit path, if this results in a retry in push_rt_task() we'll call double_lock_balance() again, at which point we'll run into said lockdep confusion. Reported-by: Roland Dreier Acked-by: Steven Rostedt Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1337282386.4281.77.camel@twins Signed-off-by: Ingo Molnar --- kernel/sched/rt.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c index 2a4e8dffbd6b..573e1ca01102 100644 --- a/kernel/sched/rt.c +++ b/kernel/sched/rt.c @@ -1562,7 +1562,7 @@ static struct rq *find_lock_lowest_rq(struct task_struct *task, struct rq *rq) task_running(rq, task) || !task->on_rq)) { - raw_spin_unlock(&lowest_rq->lock); + double_unlock_balance(rq, lowest_rq); lowest_rq = NULL; break; } -- cgit v1.2.3 From c1174876874dcf8986806e4dad3d7d07af20b439 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 31 May 2012 14:47:33 +0200 Subject: sched: Fix domain iteration Weird topologies can lead to asymmetric domain setups. This needs further consideration since these setups are typically non-minimal too. For now, make it work by adding an extra mask selecting which CPUs are allowed to iterate up. The topology that triggered it is the one from David Rientjes: 10 20 20 30 20 10 20 20 20 20 10 20 30 20 20 10 resulting in boxes that wouldn't even boot. Reported-by: David Rientjes Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-3p86l9cuaqnxz7uxsojmz5rm@git.kernel.org Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 64 +++++++++++++++++++++++++++++++++++++++++++++------- kernel/sched/fair.c | 5 ++-- kernel/sched/sched.h | 2 ++ 3 files changed, 61 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 6546083af3e0..781acb91a50a 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5994,6 +5994,44 @@ struct sched_domain_topology_level { struct sd_data data; }; +/* + * Build an iteration mask that can exclude certain CPUs from the upwards + * domain traversal. + * + * Asymmetric node setups can result in situations where the domain tree is of + * unequal depth, make sure to skip domains that already cover the entire + * range. + * + * In that case build_sched_domains() will have terminated the iteration early + * and our sibling sd spans will be empty. Domains should always include the + * cpu they're built on, so check that. + * + */ +static void build_group_mask(struct sched_domain *sd, struct sched_group *sg) +{ + const struct cpumask *span = sched_domain_span(sd); + struct sd_data *sdd = sd->private; + struct sched_domain *sibling; + int i; + + for_each_cpu(i, span) { + sibling = *per_cpu_ptr(sdd->sd, i); + if (!cpumask_test_cpu(i, sched_domain_span(sibling))) + continue; + + cpumask_set_cpu(i, sched_group_mask(sg)); + } +} + +/* + * Return the canonical balance cpu for this group, this is the first cpu + * of this group that's also in the iteration mask. + */ +int group_balance_cpu(struct sched_group *sg) +{ + return cpumask_first_and(sched_group_cpus(sg), sched_group_mask(sg)); +} + static int build_overlap_sched_groups(struct sched_domain *sd, int cpu) { @@ -6012,6 +6050,12 @@ build_overlap_sched_groups(struct sched_domain *sd, int cpu) if (cpumask_test_cpu(i, covered)) continue; + child = *per_cpu_ptr(sdd->sd, i); + + /* See the comment near build_group_mask(). */ + if (!cpumask_test_cpu(i, sched_domain_span(child))) + continue; + sg = kzalloc_node(sizeof(struct sched_group) + cpumask_size(), GFP_KERNEL, cpu_to_node(cpu)); @@ -6019,8 +6063,6 @@ build_overlap_sched_groups(struct sched_domain *sd, int cpu) goto fail; sg_span = sched_group_cpus(sg); - - child = *per_cpu_ptr(sdd->sd, i); if (child->child) { child = child->child; cpumask_copy(sg_span, sched_domain_span(child)); @@ -6030,13 +6072,18 @@ build_overlap_sched_groups(struct sched_domain *sd, int cpu) cpumask_or(covered, covered, sg_span); sg->sgp = *per_cpu_ptr(sdd->sgp, i); - atomic_inc(&sg->sgp->ref); + if (atomic_inc_return(&sg->sgp->ref) == 1) + build_group_mask(sd, sg); + + /* + * Make sure the first group of this domain contains the + * canonical balance cpu. Otherwise the sched_domain iteration + * breaks. See update_sg_lb_stats(). + */ if ((!groups && cpumask_test_cpu(cpu, sg_span)) || - cpumask_first(sg_span) == cpu) { - WARN_ON_ONCE(!cpumask_test_cpu(cpu, sg_span)); + group_balance_cpu(sg) == cpu) groups = sg; - } if (!first) first = sg; @@ -6109,6 +6156,7 @@ build_sched_groups(struct sched_domain *sd, int cpu) cpumask_clear(sched_group_cpus(sg)); sg->sgp->power = 0; + cpumask_setall(sched_group_mask(sg)); for_each_cpu(j, span) { if (get_group(j, sdd, NULL) != group) @@ -6150,7 +6198,7 @@ static void init_sched_groups_power(int cpu, struct sched_domain *sd) sg = sg->next; } while (sg != sd->groups); - if (cpu != group_first_cpu(sg)) + if (cpu != group_balance_cpu(sg)) return; update_group_power(sd, cpu); @@ -6525,7 +6573,7 @@ static int __sdt_alloc(const struct cpumask *cpu_map) *per_cpu_ptr(sdd->sg, j) = sg; - sgp = kzalloc_node(sizeof(struct sched_group_power), + sgp = kzalloc_node(sizeof(struct sched_group_power) + cpumask_size(), GFP_KERNEL, cpu_to_node(j)); if (!sgp) return -ENOMEM; diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index b2a2d236f27b..54cbaa4e7b37 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -3652,7 +3652,7 @@ static inline void update_sg_lb_stats(struct lb_env *env, int i; if (local_group) - balance_cpu = group_first_cpu(group); + balance_cpu = group_balance_cpu(group); /* Tally up the load of all CPUs in the group */ max_cpu_load = 0; @@ -3667,7 +3667,8 @@ static inline void update_sg_lb_stats(struct lb_env *env, /* Bias balancing toward cpus of our domain */ if (local_group) { - if (idle_cpu(i) && !first_idle_cpu) { + if (idle_cpu(i) && !first_idle_cpu && + cpumask_test_cpu(i, sched_group_mask(group))) { first_idle_cpu = 1; balance_cpu = i; } diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index ba9dccfd24ce..6d52cea7f33d 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -526,6 +526,8 @@ static inline struct sched_domain *highest_flag_domain(int cpu, int flag) DECLARE_PER_CPU(struct sched_domain *, sd_llc); DECLARE_PER_CPU(int, sd_llc_id); +extern int group_balance_cpu(struct sched_group *sg); + #endif /* CONFIG_SMP */ #include "stats.h" -- cgit v1.2.3 From c3decf0dfbc95736b7c0ab68fa4e5854c4734da9 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 31 May 2012 12:05:32 +0200 Subject: sched: Always initialize cpu-power Often when we run into mis-shapen topologies the balance iteration fails to update the cpu power properly and we'll end up in /0 traps. Always initialize the cpu-power to a semi-sane value so that we can at least boot the machine, even if the load-balancer might not function correctly. Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-3lbhyj25sr169ha7z3qht5na@git.kernel.org Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 13 ++++++++++++- kernel/sched/fair.c | 2 +- 2 files changed, 13 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 781acb91a50a..725ee7c1c8cf 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5604,7 +5604,12 @@ static int sched_domain_debug_one(struct sched_domain *sd, int cpu, int level, break; } - if (!group->sgp->power) { + /* + * Even though we initialize ->power to something semi-sane, + * we leave power_orig unset. This allows us to detect if + * domain iteration is still funny without causing /0 traps. + */ + if (!group->sgp->power_orig) { printk(KERN_CONT "\n"); printk(KERN_ERR "ERROR: domain->cpu_power not " "set\n"); @@ -6075,6 +6080,12 @@ build_overlap_sched_groups(struct sched_domain *sd, int cpu) if (atomic_inc_return(&sg->sgp->ref) == 1) build_group_mask(sd, sg); + /* + * Initialize sgp->power such that even if we mess up the + * domains and no possible iteration will get us here, we won't + * die on a /0 trap. + */ + sg->sgp->power = SCHED_POWER_SCALE * cpumask_weight(sg_span); /* * Make sure the first group of this domain contains the diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 54cbaa4e7b37..c9fd6d673d05 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -3602,7 +3602,7 @@ void update_group_power(struct sched_domain *sd, int cpu) } while (group != child->groups); } - sdg->sgp->power = power; + sdg->sgp->power_orig = sdg->sgp->power = power; } /* -- cgit v1.2.3 From d039ac60800fe8ed8522ec3b9ca796aaf748c18b Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 31 May 2012 21:20:16 +0200 Subject: sched: Validate assumptions in sched_init_numa() Add some code to validate assumptions we're making and output warnings if they are not. If this trigger we want to know about it. Signed-off-by: Peter Zijlstra Cc: Alex Shi Link: http://lkml.kernel.org/n/tip-6uc3wk5s9udxtdl9cnku0vtt@git.kernel.org Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 99 +++++++++++++++++++++++++++++++++++++++++++---------- 1 file changed, 80 insertions(+), 19 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 725ee7c1c8cf..2bdd17616437 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5556,15 +5556,20 @@ static cpumask_var_t sched_domains_tmpmask; /* sched_domains_mutex */ #ifdef CONFIG_SCHED_DEBUG -static __read_mostly int sched_domain_debug_enabled; +static __read_mostly int sched_debug_enabled; -static int __init sched_domain_debug_setup(char *str) +static int __init sched_debug_setup(char *str) { - sched_domain_debug_enabled = 1; + sched_debug_enabled = 1; return 0; } -early_param("sched_debug", sched_domain_debug_setup); +early_param("sched_debug", sched_debug_setup); + +static inline bool sched_debug(void) +{ + return sched_debug_enabled; +} static int sched_domain_debug_one(struct sched_domain *sd, int cpu, int level, struct cpumask *groupmask) @@ -5657,7 +5662,7 @@ static void sched_domain_debug(struct sched_domain *sd, int cpu) { int level = 0; - if (!sched_domain_debug_enabled) + if (!sched_debug_enabled) return; if (!sd) { @@ -5678,6 +5683,10 @@ static void sched_domain_debug(struct sched_domain *sd, int cpu) } #else /* !CONFIG_SCHED_DEBUG */ # define sched_domain_debug(sd, cpu) do { } while (0) +static inline bool sched_debug(void) +{ + return false; +} #endif /* CONFIG_SCHED_DEBUG */ static int sd_degenerate(struct sched_domain *sd) @@ -6373,7 +6382,6 @@ static struct sched_domain_topology_level *sched_domain_topology = default_topol #ifdef CONFIG_NUMA static int sched_domains_numa_levels; -static int sched_domains_numa_scale; static int *sched_domains_numa_distance; static struct cpumask ***sched_domains_numa_masks; static int sched_domains_curr_level; @@ -6438,6 +6446,42 @@ static const struct cpumask *sd_numa_mask(int cpu) return sched_domains_numa_masks[sched_domains_curr_level][cpu_to_node(cpu)]; } +static void sched_numa_warn(const char *str) +{ + static int done = false; + int i,j; + + if (done) + return; + + done = true; + + printk(KERN_WARNING "ERROR: %s\n\n", str); + + for (i = 0; i < nr_node_ids; i++) { + printk(KERN_WARNING " "); + for (j = 0; j < nr_node_ids; j++) + printk(KERN_CONT "%02d ", node_distance(i,j)); + printk(KERN_CONT "\n"); + } + printk(KERN_WARNING "\n"); +} + +static bool find_numa_distance(int distance) +{ + int i; + + if (distance == node_distance(0, 0)) + return true; + + for (i = 0; i < sched_domains_numa_levels; i++) { + if (sched_domains_numa_distance[i] == distance) + return true; + } + + return false; +} + static void sched_init_numa(void) { int next_distance, curr_distance = node_distance(0, 0); @@ -6445,7 +6489,6 @@ static void sched_init_numa(void) int level = 0; int i, j, k; - sched_domains_numa_scale = curr_distance; sched_domains_numa_distance = kzalloc(sizeof(int) * nr_node_ids, GFP_KERNEL); if (!sched_domains_numa_distance) return; @@ -6456,23 +6499,41 @@ static void sched_init_numa(void) * * Assumes node_distance(0,j) includes all distances in * node_distance(i,j) in order to avoid cubic time. - * - * XXX: could be optimized to O(n log n) by using sort() */ next_distance = curr_distance; for (i = 0; i < nr_node_ids; i++) { for (j = 0; j < nr_node_ids; j++) { - int distance = node_distance(0, j); - if (distance > curr_distance && - (distance < next_distance || - next_distance == curr_distance)) - next_distance = distance; + for (k = 0; k < nr_node_ids; k++) { + int distance = node_distance(i, k); + + if (distance > curr_distance && + (distance < next_distance || + next_distance == curr_distance)) + next_distance = distance; + + /* + * While not a strong assumption it would be nice to know + * about cases where if node A is connected to B, B is not + * equally connected to A. + */ + if (sched_debug() && node_distance(k, i) != distance) + sched_numa_warn("Node-distance not symmetric"); + + if (sched_debug() && i && !find_numa_distance(distance)) + sched_numa_warn("Node-0 not representative"); + } + if (next_distance != curr_distance) { + sched_domains_numa_distance[level++] = next_distance; + sched_domains_numa_levels = level; + curr_distance = next_distance; + } else break; } - if (next_distance != curr_distance) { - sched_domains_numa_distance[level++] = next_distance; - sched_domains_numa_levels = level; - curr_distance = next_distance; - } else break; + + /* + * In case of sched_debug() we verify the above assumption. + */ + if (!sched_debug()) + break; } /* * 'level' contains the number of unique distances, excluding the -- cgit v1.2.3 From a841f8cef4bb124f0f5563314d0beaf2e1249d72 Mon Sep 17 00:00:00 2001 From: Dimitri Sivanich Date: Tue, 5 Jun 2012 13:44:36 -0500 Subject: sched: Fix the relax_domain_level boot parameter It does not get processed because sched_domain_level_max is 0 at the time that setup_relax_domain_level() is run. Simply accept the value as it is, as we don't know the value of sched_domain_level_max until sched domain construction is completed. Fix sched_relax_domain_level in cpuset. The build_sched_domain() routine calls the set_domain_attribute() routine prior to setting the sd->level, however, the set_domain_attribute() routine relies on the sd->level to decide whether idle load balancing will be off/on. Signed-off-by: Dimitri Sivanich Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/20120605184436.GA15668@sgi.com Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 2bdd17616437..d5594a4268d4 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -6268,11 +6268,8 @@ int sched_domain_level_max; static int __init setup_relax_domain_level(char *str) { - unsigned long val; - - val = simple_strtoul(str, NULL, 0); - if (val < sched_domain_level_max) - default_relax_domain_level = val; + if (kstrtoint(str, 0, &default_relax_domain_level)) + pr_warn("Unable to set relax_domain_level\n"); return 1; } @@ -6698,7 +6695,6 @@ struct sched_domain *build_sched_domain(struct sched_domain_topology_level *tl, if (!sd)