From b8f8c3cf0a4ac0632ec3f0e15e9dc0c29de917af Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 18 Jul 2008 17:27:28 +0200 Subject: nohz: prevent tick stop outside of the idle loop Jack Ren and Eric Miao tracked down the following long standing problem in the NOHZ code: scheduler switch to idle task enable interrupts Window starts here ----> interrupt happens (does not set NEED_RESCHED) irq_exit() stops the tick ----> interrupt happens (does set NEED_RESCHED) return from schedule() cpu_idle(): preempt_disable(); Window ends here The interrupts can happen at any point inside the race window. The first interrupt stops the tick, the second one causes the scheduler to rerun and switch away from idle again and we end up with the tick disabled. The fact that it needs two interrupts where the first one does not set NEED_RESCHED and the second one does made the bug obscure and extremly hard to reproduce and analyse. Kudos to Jack and Eric. Solution: Limit the NOHZ functionality to the idle loop to make sure that we can not run into such a situation ever again. cpu_idle() { preempt_disable(); while(1) { tick_nohz_stop_sched_tick(1); <- tell NOHZ code that we are in the idle loop while (!need_resched()) halt(); tick_nohz_restart_sched_tick(); <- disables NOHZ mode preempt_enable_no_resched(); schedule(); preempt_disable(); } } In hindsight we should have done this forever, but ... /me grabs a large brown paperbag. Debugged-by: Jack Ren , Debugged-by: eric miao Signed-off-by: Thomas Gleixner --- kernel/softirq.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/softirq.c') diff --git a/kernel/softirq.c b/kernel/softirq.c index 36e061740047..05f248039d77 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -312,7 +312,7 @@ void irq_exit(void) #ifdef CONFIG_NO_HZ /* Make sure that timer wheel updates are propagated */ if (!in_interrupt() && idle_cpu(smp_processor_id()) && !need_resched()) - tick_nohz_stop_sched_tick(); + tick_nohz_stop_sched_tick(0); rcu_irq_exit(); #endif preempt_enable_no_resched(); -- cgit v1.2.3 From 7babe8db99d305340cf4828ce1f5a1481d5622ef Mon Sep 17 00:00:00 2001 From: Eduard - Gabriel Munteanu Date: Fri, 25 Jul 2008 19:45:11 -0700 Subject: Full conversion to early_initcall() interface, remove old interface A previous patch added the early_initcall(), to allow a cleaner hooking of pre-SMP initcalls. Now we remove the older interface, converting all existing users to the new one. [akpm@linux-foundation.org: cleanups] [akpm@linux-foundation.org: build fix] [kosaki.motohiro@jp.fujitsu.com: warning fix] [kosaki.motohiro@jp.fujitsu.com: warning fix] Signed-off-by: Eduard - Gabriel Munteanu Cc: Tom Zanussi Signed-off-by: KOSAKI Motohiro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/sched.h | 9 --------- include/linux/smp.h | 5 ----- init/main.c | 23 +---------------------- kernel/sched.c | 5 ++++- kernel/smp.c | 4 +++- kernel/softirq.c | 3 ++- kernel/softlockup.c | 25 ++++++++++++++++++++++--- 7 files changed, 32 insertions(+), 42 deletions(-) (limited to 'kernel/softirq.c') diff --git a/include/linux/sched.h b/include/linux/sched.h index 3260a5c42b91..adb8077dc463 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -292,7 +292,6 @@ extern void sched_show_task(struct task_struct *p); #ifdef CONFIG_DETECT_SOFTLOCKUP extern void softlockup_tick(void); -extern void spawn_softlockup_task(void); extern void touch_softlockup_watchdog(void); extern void touch_all_softlockup_watchdogs(void); extern unsigned int softlockup_panic; @@ -2222,14 +2221,6 @@ static inline void inc_syscw(struct task_struct *tsk) } #endif -#ifdef CONFIG_SMP -void migration_init(void); -#else -static inline void migration_init(void) -{ -} -#endif - #ifndef TASK_SIZE_OF #define TASK_SIZE_OF(tsk) TASK_SIZE #endif diff --git a/include/linux/smp.h b/include/linux/smp.h index 48262f86c969..66484d4a8459 100644 --- a/include/linux/smp.h +++ b/include/linux/smp.h @@ -74,15 +74,10 @@ void __smp_call_function_single(int cpuid, struct call_single_data *data); #ifdef CONFIG_USE_GENERIC_SMP_HELPERS void generic_smp_call_function_single_interrupt(void); void generic_smp_call_function_interrupt(void); -void init_call_single_data(void); void ipi_call_lock(void); void ipi_call_unlock(void); void ipi_call_lock_irq(void); void ipi_call_unlock_irq(void); -#else -static inline void init_call_single_data(void) -{ -} #endif /* diff --git a/init/main.c b/init/main.c index b6fec08dbbef..20fdc9884b77 100644 --- a/init/main.c +++ b/init/main.c @@ -774,16 +774,7 @@ static void __init do_basic_setup(void) do_initcalls(); } -static int __initdata nosoftlockup; - -static int __init nosoftlockup_setup(char *str) -{ - nosoftlockup = 1; - return 1; -} -__setup("nosoftlockup", nosoftlockup_setup); - -static void __init __do_pre_smp_initcalls(void) +static void __init do_pre_smp_initcalls(void) { initcall_t *call; @@ -791,17 +782,6 @@ static void __init __do_pre_smp_initcalls(void) do_one_initcall(*call); } -static void __init do_pre_smp_initcalls(void) -{ - extern int spawn_ksoftirqd(void); - - init_call_single_data(); - migration_init(); - spawn_ksoftirqd(); - if (!nosoftlockup) - spawn_softlockup_task(); -} - static void run_init_process(char *init_filename) { argv_init[0] = init_filename; @@ -873,7 +853,6 @@ static int __init kernel_init(void * unused) smp_prepare_cpus(setup_max_cpus); - __do_pre_smp_initcalls(); do_pre_smp_initcalls(); smp_init(); diff --git a/kernel/sched.c b/kernel/sched.c index 0047bd9b96aa..fde1a1026359 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -6389,7 +6389,7 @@ static struct notifier_block __cpuinitdata migration_notifier = { .priority = 10 }; -void __init migration_init(void) +static int __init migration_init(void) { void *cpu = (void *)(long)smp_processor_id(); int err; @@ -6399,7 +6399,10 @@ void __init migration_init(void) BUG_ON(err == NOTIFY_BAD); migration_call(&migration_notifier, CPU_ONLINE, cpu); register_cpu_notifier(&migration_notifier); + + return err; } +early_initcall(migration_init); #endif #ifdef CONFIG_SMP diff --git a/kernel/smp.c b/kernel/smp.c index 462c785ca1ee..96fc7c0edc59 100644 --- a/kernel/smp.c +++ b/kernel/smp.c @@ -33,7 +33,7 @@ struct call_single_queue { spinlock_t lock; }; -void __cpuinit init_call_single_data(void) +static int __cpuinit init_call_single_data(void) { int i; @@ -43,7 +43,9 @@ void __cpuinit init_call_single_data(void) spin_lock_init(&q->lock); INIT_LIST_HEAD(&q->list); } + return 0; } +early_initcall(init_call_single_data); static void csd_flag_wait(struct call_single_data *data) { diff --git a/kernel/softirq.c b/kernel/softirq.c index f6b03d56c2bf..c506f266a6b9 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -630,7 +630,7 @@ static struct notifier_block __cpuinitdata cpu_nfb = { .notifier_call = cpu_callback }; -__init int spawn_ksoftirqd(void) +static __init int spawn_ksoftirqd(void) { void *cpu = (void *)(long)smp_processor_id(); int err = cpu_callback(&cpu_nfb, CPU_UP_PREPARE, cpu); @@ -640,6 +640,7 @@ __init int spawn_ksoftirqd(void) register_cpu_notifier(&cpu_nfb); return 0; } +early_initcall(spawn_ksoftirqd); #ifdef CONFIG_SMP /* diff --git a/kernel/softlockup.c b/kernel/softlockup.c index 7bd8d1aadd5d..b75b492fbfcf 100644 --- a/kernel/softlockup.c +++ b/kernel/softlockup.c @@ -338,14 +338,33 @@ static struct notifier_block __cpuinitdata cpu_nfb = { .notifier_call = cpu_callback }; -__init void spawn_softlockup_task(void) +static int __initdata nosoftlockup; + +static int __init nosoftlockup_setup(char *str) +{ + nosoftlockup = 1; + return 1; +} +__setup("nosoftlockup", nosoftlockup_setup); + +static int __init spawn_softlockup_task(void) { void *cpu = (void *)(long)smp_processor_id(); - int err = cpu_callback(&cpu_nfb, CPU_UP_PREPARE, cpu); + int err; - BUG_ON(err == NOTIFY_BAD); + if (nosoftlockup) + return 0; + + err = cpu_callback(&cpu_nfb, CPU_UP_PREPARE, cpu); + if (err == NOTIFY_BAD) { + BUG(); + return 1; + } cpu_callback(&cpu_nfb, CPU_ONLINE, cpu); register_cpu_notifier(&cpu_nfb); atomic_notifier_chain_register(&panic_notifier_list, &panic_block); + + return 0; } +early_initcall(spawn_softlockup_task); -- cgit v1.2.3