From: Ingo Molnar Date: Tue, 29 Nov 2011 20:18:22 -0500 Subject: [PATCH 118/342] tasklet: Prevent tasklets from going into infinite spin in RT Origin: https://git.kernel.org/cgit/linux/kernel/git/rt/linux-stable-rt.git/commit?id=e11d4ceb7976f9f4bb42ed76c972b28cc32c295f When CONFIG_PREEMPT_RT_FULL is enabled, tasklets run as threads, and spinlocks turn are mutexes. But this can cause issues with tasks disabling tasklets. A tasklet runs under ksoftirqd, and if a tasklets are disabled with tasklet_disable(), the tasklet count is increased. When a tasklet runs, it checks this counter and if it is set, it adds itself back on the softirq queue and returns. The problem arises in RT because ksoftirq will see that a softirq is ready to run (the tasklet softirq just re-armed itself), and will not sleep, but instead run the softirqs again. The tasklet softirq will still see that the count is non-zero and will not execute the tasklet and requeue itself on the softirq again, which will cause ksoftirqd to run it again and again and again. It gets worse because ksoftirqd runs as a real-time thread. If it preempted the task that disabled tasklets, and that task has migration disabled, or can't run for other reasons, the tasklet softirq will never run because the count will never be zero, and ksoftirqd will go into an infinite loop. As an RT task, it this becomes a big problem. This is a hack solution to have tasklet_disable stop tasklets, and when a tasklet runs, instead of requeueing the tasklet softirqd it delays it. When tasklet_enable() is called, and tasklets are waiting, then the tasklet_enable() will kick the tasklets to continue. This prevents the lock up from ksoftirq going into an infinite loop. [ rostedt@goodmis.org: ported to 3.0-rt ] Signed-off-by: Ingo Molnar Signed-off-by: Steven Rostedt Signed-off-by: Thomas Gleixner --- include/linux/interrupt.h | 33 +++++----- kernel/softirq.c | 126 ++++++++++++++++++++++++++++++++------ 2 files changed, 125 insertions(+), 34 deletions(-) diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h index 315f852b4981..35fa335c475b 100644 --- a/include/linux/interrupt.h +++ b/include/linux/interrupt.h @@ -536,8 +536,9 @@ static inline struct task_struct *this_cpu_ksoftirqd(void) to be executed on some cpu at least once after this. * If the tasklet is already scheduled, but its execution is still not started, it will be executed only once. - * If this tasklet is already running on another CPU (or schedule is called - from tasklet itself), it is rescheduled for later. + * If this tasklet is already running on another CPU, it is rescheduled + for later. + * Schedule must not be called from the tasklet itself (a lockup occurs) * Tasklet is strictly serialized wrt itself, but not wrt another tasklets. If client needs some intertask synchronization, he makes it with spinlocks. @@ -562,27 +563,36 @@ struct tasklet_struct name = { NULL, 0, ATOMIC_INIT(1), func, data } enum { TASKLET_STATE_SCHED, /* Tasklet is scheduled for execution */ - TASKLET_STATE_RUN /* Tasklet is running (SMP only) */ + TASKLET_STATE_RUN, /* Tasklet is running (SMP only) */ + TASKLET_STATE_PENDING /* Tasklet is pending */ }; -#ifdef CONFIG_SMP +#define TASKLET_STATEF_SCHED (1 << TASKLET_STATE_SCHED) +#define TASKLET_STATEF_RUN (1 << TASKLET_STATE_RUN) +#define TASKLET_STATEF_PENDING (1 << TASKLET_STATE_PENDING) + +#if defined(CONFIG_SMP) || defined(CONFIG_PREEMPT_RT_FULL) static inline int tasklet_trylock(struct tasklet_struct *t) { return !test_and_set_bit(TASKLET_STATE_RUN, &(t)->state); } +static inline int tasklet_tryunlock(struct tasklet_struct *t) +{ + return cmpxchg(&t->state, TASKLET_STATEF_RUN, 0) == TASKLET_STATEF_RUN; +} + static inline void tasklet_unlock(struct tasklet_struct *t) { smp_mb__before_atomic(); clear_bit(TASKLET_STATE_RUN, &(t)->state); } -static inline void tasklet_unlock_wait(struct tasklet_struct *t) -{ - while (test_bit(TASKLET_STATE_RUN, &(t)->state)) { barrier(); } -} +extern void tasklet_unlock_wait(struct tasklet_struct *t); + #else #define tasklet_trylock(t) 1 +#define tasklet_tryunlock(t) 1 #define tasklet_unlock_wait(t) do { } while (0) #define tasklet_unlock(t) do { } while (0) #endif @@ -616,12 +626,7 @@ static inline void tasklet_disable(struct tasklet_struct *t) smp_mb(); } -static inline void tasklet_enable(struct tasklet_struct *t) -{ - smp_mb__before_atomic(); - atomic_dec(&t->count); -} - +extern void tasklet_enable(struct tasklet_struct *t); extern void tasklet_kill(struct tasklet_struct *t); extern void tasklet_kill_immediate(struct tasklet_struct *t, unsigned int cpu); extern void tasklet_init(struct tasklet_struct *t, diff --git a/kernel/softirq.c b/kernel/softirq.c index 6f584861d329..1d3a482246cc 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -21,6 +21,7 @@ #include #include #include +#include #include #include #include @@ -475,11 +476,38 @@ static void __tasklet_schedule_common(struct tasklet_struct *t, unsigned long flags; local_irq_save(flags); + if (!tasklet_trylock(t)) { + local_irq_restore(flags); + return; + } + head = this_cpu_ptr(headp); - t->next = NULL; - *head->tail = t; - head->tail = &(t->next); - raise_softirq_irqoff(softirq_nr); +again: + /* We may have been preempted before tasklet_trylock + * and __tasklet_action may have already run. + * So double check the sched bit while the takslet + * is locked before adding it to the list. + */ + if (test_bit(TASKLET_STATE_SCHED, &t->state)) { + t->next = NULL; + *head->tail = t; + head->tail = &(t->next); + raise_softirq_irqoff(softirq_nr); + tasklet_unlock(t); + } else { + /* This is subtle. If we hit the corner case above + * It is possible that we get preempted right here, + * and another task has successfully called + * tasklet_schedule(), then this function, and + * failed on the trylock. Thus we must be sure + * before releasing the tasklet lock, that the + * SCHED_BIT is clear. Otherwise the tasklet + * may get its SCHED_BIT set, but not added to the + * list + */ + if (!tasklet_tryunlock(t)) + goto again; + } local_irq_restore(flags); } @@ -497,11 +525,21 @@ void __tasklet_hi_schedule(struct tasklet_struct *t) } EXPORT_SYMBOL(__tasklet_hi_schedule); +void tasklet_enable(struct tasklet_struct *t) +{ + if (!atomic_dec_and_test(&t->count)) + return; + if (test_and_clear_bit(TASKLET_STATE_PENDING, &t->state)) + tasklet_schedule(t); +} +EXPORT_SYMBOL(tasklet_enable); + static void tasklet_action_common(struct softirq_action *a, struct tasklet_head *tl_head, unsigned int softirq_nr) { struct tasklet_struct *list; + int loops = 1000000; local_irq_disable(); list = tl_head->head; @@ -513,25 +551,56 @@ static void tasklet_action_common(struct softirq_action *a, struct tasklet_struct *t = list; list = list->next; + /* + * Should always succeed - after a tasklist got on the + * list (after getting the SCHED bit set from 0 to 1), + * nothing but the tasklet softirq it got queued to can + * lock it: + */ + if (!tasklet_trylock(t)) { + WARN_ON(1); + continue; + } - if (tasklet_trylock(t)) { - if (!atomic_read(&t->count)) { - if (!test_and_clear_bit(TASKLET_STATE_SCHED, - &t->state)) - BUG(); - t->func(t->data); + t->next = NULL; + + if (unlikely(atomic_read(&t->count))) { +out_disabled: + /* implicit unlock: */ + wmb(); + t->state = TASKLET_STATEF_PENDING; + continue; + } + /* + * After this point on the tasklet might be rescheduled + * on another CPU, but it can only be added to another + * CPU's tasklet list if we unlock the tasklet (which we + * dont do yet). + */ + if (!test_and_clear_bit(TASKLET_STATE_SCHED, &t->state)) + WARN_ON(1); +again: + t->func(t->data); + + while (!tasklet_tryunlock(t)) { + /* + * If it got disabled meanwhile, bail out: + */ + if (atomic_read(&t->count)) + goto out_disabled; + /* + * If it got scheduled meanwhile, re-execute + * the tasklet function: + */ + if (test_and_clear_bit(TASKLET_STATE_SCHED, &t->state)) + goto again; + if (!--loops) { + printk("hm, tasklet state: %08lx\n", t->state); + WARN_ON(1); tasklet_unlock(t); - continue; + break; } - tasklet_unlock(t); } - - local_irq_disable(); - t->next = NULL; - *tl_head->tail = t; - tl_head->tail = &t->next; - __raise_softirq_irqoff(softirq_nr); - local_irq_enable(); } } @@ -563,7 +632,7 @@ void tasklet_kill(struct tasklet_struct *t) while (test_and_set_bit(TASKLET_STATE_SCHED, &t->state)) { do { - yield(); + msleep(1); } while (test_bit(TASKLET_STATE_SCHED, &t->state)); } tasklet_unlock_wait(t); @@ -637,6 +706,23 @@ void __init softirq_init(void) open_softirq(HI_SOFTIRQ, tasklet_hi_action); } +#if defined(CONFIG_SMP) || defined(CONFIG_PREEMPT_RT_FULL) +void tasklet_unlock_wait(struct tasklet_struct *t) +{ + while (test_bit(TASKLET_STATE_RUN, &(t)->state)) { + /* + * Hack for now to avoid this busy-loop: + */ +#ifdef CONFIG_PREEMPT_RT_FULL + msleep(1); +#else + barrier(); +#endif + } +} +EXPORT_SYMBOL(tasklet_unlock_wait); +#endif + static int ksoftirqd_should_run(unsigned int cpu) { return local_softirq_pending();