static inline void debug_hrtimer_deactivate(struct hrtimer *timer) { }
#endif
-/*
- * Check, whether the timer is on the callback pending list
- */
-static inline int hrtimer_cb_pending(const struct hrtimer *timer)
-{
- return timer->state & HRTIMER_STATE_PENDING;
-}
-
-/*
- * Remove a timer from the callback pending list
- */
-static inline void hrtimer_remove_cb_pending(struct hrtimer *timer)
-{
- list_del_init(&timer->cb_entry);
-}
-
/* High resolution timer related functions */
#ifdef CONFIG_HIGH_RES_TIMERS
{
}
+static void __run_hrtimer(struct hrtimer *timer);
+
/*
* When High resolution timers are active, try to reprogram. Note, that in case
* the state has HRTIMER_STATE_CALLBACK set, no reprogramming and no expiry
struct hrtimer_clock_base *base)
{
if (base->cpu_base->hres_active && hrtimer_reprogram(timer, base)) {
-
- /* Timer is expired, act upon the callback mode */
- switch(timer->cb_mode) {
- case HRTIMER_CB_IRQSAFE_PERCPU:
- case HRTIMER_CB_IRQSAFE_UNLOCKED:
- /*
- * This is solely for the sched tick emulation with
- * dynamic tick support to ensure that we do not
- * restart the tick right on the edge and end up with
- * the tick timer in the softirq ! The calling site
- * takes care of this. Also used for hrtimer sleeper !
- */
- debug_hrtimer_deactivate(timer);
- return 1;
- case HRTIMER_CB_SOFTIRQ:
- /*
- * Move everything else into the softirq pending list !
- */
- list_add_tail(&timer->cb_entry,
- &base->cpu_base->cb_pending);
- timer->state = HRTIMER_STATE_PENDING;
- return 1;
- default:
- BUG();
- }
+ /*
+ * XXX: recursion check?
+ * hrtimer_forward() should round up with timer granularity
+ * so that we never get into inf recursion here,
+ * it doesn't do that though
+ */
+ __run_hrtimer(timer);
+ return 1;
}
return 0;
}
return 1;
}
-static inline void hrtimer_raise_softirq(void)
-{
- raise_softirq(HRTIMER_SOFTIRQ);
-}
-
#else
static inline int hrtimer_hres_active(void) { return 0; }
{
return 0;
}
-static inline void hrtimer_raise_softirq(void) { }
#endif /* CONFIG_HIGH_RES_TIMERS */
struct hrtimer_clock_base *base,
unsigned long newstate, int reprogram)
{
- /* High res. callback list. NOP for !HIGHRES */
- if (hrtimer_cb_pending(timer))
- hrtimer_remove_cb_pending(timer);
- else {
+ if (timer->state & HRTIMER_STATE_ENQUEUED) {
/*
* Remove the timer from the rbtree and replace the
* first entry pointer if necessary.
{
struct hrtimer_clock_base *base, *new_base;
unsigned long flags;
- int ret, raise;
+ int ret;
base = lock_hrtimer_base(timer, &flags);
enqueue_hrtimer(timer, new_base,
new_base->cpu_base == &__get_cpu_var(hrtimer_bases));
- /*
- * The timer may be expired and moved to the cb_pending
- * list. We can not raise the softirq with base lock held due
- * to a possible deadlock with runqueue lock.
- */
- raise = timer->state == HRTIMER_STATE_PENDING;
-
- /*
- * We use preempt_disable to prevent this task from migrating after
- * setting up the softirq and raising it. Otherwise, if me migrate
- * we will raise the softirq on the wrong CPU.
- */
- preempt_disable();
-
unlock_hrtimer_base(timer, &flags);
- if (raise)
- hrtimer_raise_softirq();
- preempt_enable();
-
return ret;
}
EXPORT_SYMBOL_GPL(hrtimer_start_range_ns);
}
EXPORT_SYMBOL_GPL(hrtimer_get_res);
-static void run_hrtimer_pending(struct hrtimer_cpu_base *cpu_base)
-{
- spin_lock_irq(&cpu_base->lock);
-
- while (!list_empty(&cpu_base->cb_pending)) {
- enum hrtimer_restart (*fn)(struct hrtimer *);
- struct hrtimer *timer;
- int restart;
- int emulate_hardirq_ctx = 0;
-
- timer = list_entry(cpu_base->cb_pending.next,
- struct hrtimer, cb_entry);
-
- debug_hrtimer_deactivate(timer);
- timer_stats_account_hrtimer(timer);
-
- fn = timer->function;
- /*
- * A timer might have been added to the cb_pending list
- * when it was migrated during a cpu-offline operation.
- * Emulate hardirq context for such timers.
- */
- if (timer->cb_mode == HRTIMER_CB_IRQSAFE_PERCPU ||
- timer->cb_mode == HRTIMER_CB_IRQSAFE_UNLOCKED)
- emulate_hardirq_ctx = 1;
-
- __remove_hrtimer(timer, timer->base, HRTIMER_STATE_CALLBACK, 0);
- spin_unlock_irq(&cpu_base->lock);
-
- if (unlikely(emulate_hardirq_ctx)) {
- local_irq_disable();
- restart = fn(timer);
- local_irq_enable();
- } else
- restart = fn(timer);
-
- spin_lock_irq(&cpu_base->lock);
-
- timer->state &= ~HRTIMER_STATE_CALLBACK;
- if (restart == HRTIMER_RESTART) {
- BUG_ON(hrtimer_active(timer));
- /*
- * Enqueue the timer, allow reprogramming of the event
- * device
- */
- enqueue_hrtimer(timer, timer->base, 1);
- } else if (hrtimer_active(timer)) {
- /*
- * If the timer was rearmed on another CPU, reprogram
- * the event device.
- */
- struct hrtimer_clock_base *base = timer->base;
-
- if (base->first == &timer->node &&
- hrtimer_reprogram(timer, base)) {
- /*
- * Timer is expired. Thus move it from tree to
- * pending list again.
- */
- __remove_hrtimer(timer, base,
- HRTIMER_STATE_PENDING, 0);
- list_add_tail(&timer->cb_entry,
- &base->cpu_base->cb_pending);
- }
- }
- }
- spin_unlock_irq(&cpu_base->lock);
-}
-
static void __run_hrtimer(struct hrtimer *timer)
{
struct hrtimer_clock_base *base = timer->base;
enum hrtimer_restart (*fn)(struct hrtimer *);
int restart;
+ WARN_ON(!irqs_disabled());
+
debug_hrtimer_deactivate(timer);
__remove_hrtimer(timer, base, HRTIMER_STATE_CALLBACK, 0);
timer_stats_account_hrtimer(timer);
-
fn = timer->function;
- if (timer->cb_mode == HRTIMER_CB_IRQSAFE_PERCPU ||
- timer->cb_mode == HRTIMER_CB_IRQSAFE_UNLOCKED) {
- /*
- * Used for scheduler timers, avoid lock inversion with
- * rq->lock and tasklist_lock.
- *
- * These timers are required to deal with enqueue expiry
- * themselves and are not allowed to migrate.
- */
- spin_unlock(&cpu_base->lock);
- restart = fn(timer);
- spin_lock(&cpu_base->lock);
- } else
- restart = fn(timer);
+
+ /*
+ * Because we run timers from hardirq context, there is no chance
+ * they get migrated to another cpu, therefore its safe to unlock
+ * the timer base.
+ */
+ spin_unlock(&cpu_base->lock);
+ restart = fn(timer);
+ spin_lock(&cpu_base->lock);
/*
* Note: We clear the CALLBACK bit after enqueue_hrtimer to avoid
#ifdef CONFIG_HIGH_RES_TIMERS
+static int force_clock_reprogram;
+
+/*
+ * After 5 iteration's attempts, we consider that hrtimer_interrupt()
+ * is hanging, which could happen with something that slows the interrupt
+ * such as the tracing. Then we force the clock reprogramming for each future
+ * hrtimer interrupts to avoid infinite loops and use the min_delta_ns
+ * threshold that we will overwrite.
+ * The next tick event will be scheduled to 3 times we currently spend on
+ * hrtimer_interrupt(). This gives a good compromise, the cpus will spend
+ * 1/4 of their time to process the hrtimer interrupts. This is enough to
+ * let it running without serious starvation.
+ */
+
+static inline void
+hrtimer_interrupt_hanging(struct clock_event_device *dev,
+ ktime_t try_time)
+{
+ force_clock_reprogram = 1;
+ dev->min_delta_ns = (unsigned long)try_time.tv64 * 3;
+ printk(KERN_WARNING "hrtimer: interrupt too slow, "
+ "forcing clock min delta to %lu ns\n", dev->min_delta_ns);
+}
/*
* High resolution timer interrupt
* Called with interrupts disabled
struct hrtimer_cpu_base *cpu_base = &__get_cpu_var(hrtimer_bases);
struct hrtimer_clock_base *base;
ktime_t expires_next, now;
- int i, raise = 0;
+ int nr_retries = 0;
+ int i;
BUG_ON(!cpu_base->hres_active);
cpu_base->nr_events++;
dev->next_event.tv64 = KTIME_MAX;
retry:
+ /* 5 retries is enough to notice a hang */
+ if (!(++nr_retries % 5))
+ hrtimer_interrupt_hanging(dev, ktime_sub(ktime_get(), now));
+
now = ktime_get();
expires_next.tv64 = KTIME_MAX;
break;
}
- /* Move softirq callbacks to the pending list */
- if (timer->cb_mode == HRTIMER_CB_SOFTIRQ) {
- __remove_hrtimer(timer, base,
- HRTIMER_STATE_PENDING, 0);
- list_add_tail(&timer->cb_entry,
- &base->cpu_base->cb_pending);
- raise = 1;
- continue;
- }
-
__run_hrtimer(timer);
}
spin_unlock(&cpu_base->lock);
/* Reprogramming necessary ? */
if (expires_next.tv64 != KTIME_MAX) {
- if (tick_program_event(expires_next, 0))
+ if (tick_program_event(expires_next, force_clock_reprogram))
goto retry;
}
-
- /* Raise softirq ? */
- if (raise)
- raise_softirq(HRTIMER_SOFTIRQ);
}
/**
local_irq_restore(flags);
}
-static void run_hrtimer_softirq(struct softirq_action *h)
-{
- run_hrtimer_pending(&__get_cpu_var(hrtimer_bases));
-}
-
#endif /* CONFIG_HIGH_RES_TIMERS */
/*
*/
void hrtimer_run_pending(void)
{
- struct hrtimer_cpu_base *cpu_base = &__get_cpu_var(hrtimer_bases);
-
if (hrtimer_hres_active())
return;
*/
if (tick_check_oneshot_change(!hrtimer_is_hres_enabled()))
hrtimer_switch_to_hres();
-
- run_hrtimer_pending(cpu_base);
}
/*
hrtimer_get_expires_tv64(timer))
break;
- if (timer->cb_mode == HRTIMER_CB_SOFTIRQ) {
- __remove_hrtimer(timer, base,
- HRTIMER_STATE_PENDING, 0);
- list_add_tail(&timer->cb_entry,
- &base->cpu_base->cb_pending);
- continue;
- }
-
__run_hrtimer(timer);
}
spin_unlock(&cpu_base->lock);
{
sl->timer.function = hrtimer_wakeup;
sl->task = task;
-#ifdef CONFIG_HIGH_RES_TIMERS
- sl->timer.cb_mode = HRTIMER_CB_IRQSAFE_UNLOCKED;
-#endif
}
static int __sched do_nanosleep(struct hrtimer_sleeper *t, enum hrtimer_mode mode)
for (i = 0; i < HRTIMER_MAX_CLOCK_BASES; i++)
cpu_base->clock_base[i].cpu_base = cpu_base;
- INIT_LIST_HEAD(&cpu_base->cb_pending);
hrtimer_init_hres(cpu_base);
}
#ifdef CONFIG_HOTPLUG_CPU
-static int migrate_hrtimer_list(struct hrtimer_clock_base *old_base,
- struct hrtimer_clock_base *new_base, int dcpu)
+static void migrate_hrtimer_list(struct hrtimer_clock_base *old_base,
+ struct hrtimer_clock_base *new_base)
{
struct hrtimer *timer;
struct rb_node *node;
- int raise = 0;
while ((node = rb_first(&old_base->active))) {
timer = rb_entry(node, struct hrtimer, node);
BUG_ON(hrtimer_callback_running(timer));
debug_hrtimer_deactivate(timer);
- /*
- * Should not happen. Per CPU timers should be
- * canceled _before_ the migration code is called
- */
- if (timer->cb_mode == HRTIMER_CB_IRQSAFE_PERCPU) {
- __remove_hrtimer(timer, old_base,
- HRTIMER_STATE_INACTIVE, 0);
- WARN(1, "hrtimer (%p %p)active but cpu %d dead\n",
- timer, timer->function, dcpu);
- continue;
- }
-
/*
* Mark it as STATE_MIGRATE not INACTIVE otherwise the
* timer could be seen as !active and just vanish away
__remove_hrtimer(timer, old_base, HRTIMER_STATE_MIGRATE, 0);
timer->base = new_base;
/*
- * Enqueue the timer. Allow reprogramming of the event device
+ * Enqueue the timers on the new cpu, but do not reprogram
+ * the timer as that would enable a deadlock between
+ * hrtimer_enqueue_reprogramm() running the timer and us still
+ * holding a nested base lock.
+ *
+ * Instead we tickle the hrtimer interrupt after the migration
+ * is done, which will run all expired timers and re-programm
+ * the timer device.
*/
- enqueue_hrtimer(timer, new_base, 1);
+ enqueue_hrtimer(timer, new_base, 0);
-#ifdef CONFIG_HIGH_RES_TIMERS
- /*
- * Happens with high res enabled when the timer was
- * already expired and the callback mode is
- * HRTIMER_CB_IRQSAFE_UNLOCKED (hrtimer_sleeper). The
- * enqueue code does not move them to the soft irq
- * pending list for performance/latency reasons, but
- * in the migration state, we need to do that
- * otherwise we end up with a stale timer.
- */
- if (timer->state == HRTIMER_STATE_MIGRATE) {
- timer->state = HRTIMER_STATE_PENDING;
- list_add_tail(&timer->cb_entry,
- &new_base->cpu_base->cb_pending);
- raise = 1;
- }
-#endif
/* Clear the migration state bit */
timer->state &= ~HRTIMER_STATE_MIGRATE;
}
- return raise;
-}
-
-#ifdef CONFIG_HIGH_RES_TIMERS
-static int migrate_hrtimer_pending(struct hrtimer_cpu_base *old_base,
- struct hrtimer_cpu_base *new_base)
-{
- struct hrtimer *timer;
- int raise = 0;
-
- while (!list_empty(&old_base->cb_pending)) {
- timer = list_entry(old_base->cb_pending.next,
- struct hrtimer, cb_entry);
-
- __remove_hrtimer(timer, timer->base, HRTIMER_STATE_PENDING, 0);
- timer->base = &new_base->clock_base[timer->base->index];
- list_add_tail(&timer->cb_entry, &new_base->cb_pending);
- raise = 1;
- }
- return raise;
-}
-#else
-static int migrate_hrtimer_pending(struct hrtimer_cpu_base *old_base,
- struct hrtimer_cpu_base *new_base)
-{
- return 0;
}
-#endif
-static void migrate_hrtimers(int cpu)
+static int migrate_hrtimers(int scpu)
{
struct hrtimer_cpu_base *old_base, *new_base;
- int i, raise = 0;
+ int dcpu, i;
- BUG_ON(cpu_online(cpu));
- old_base = &per_cpu(hrtimer_bases, cpu);
+ BUG_ON(cpu_online(scpu));
+ old_base = &per_cpu(hrtimer_bases, scpu);
new_base = &get_cpu_var(hrtimer_bases);
- tick_cancel_sched_timer(cpu);
+ dcpu = smp_processor_id();
+
+ tick_cancel_sched_timer(scpu);
/*
* The caller is globally serialized and nobody else
* takes two locks at once, deadlock is not possible.
spin_lock_nested(&old_base->lock, SINGLE_DEPTH_NESTING);
for (i = 0; i < HRTIMER_MAX_CLOCK_BASES; i++) {
- if (migrate_hrtimer_list(&old_base->clock_base[i],
- &new_base->clock_base[i], cpu))
- raise = 1;
+ migrate_hrtimer_list(&old_base->clock_base[i],
+ &new_base->clock_base[i]);
}
- if (migrate_hrtimer_pending(old_base, new_base))
- raise = 1;
-
spin_unlock(&old_base->lock);
spin_unlock_irq(&new_base->lock);
put_cpu_var(hrtimer_bases);
- if (raise)
- hrtimer_raise_softirq();
+ return dcpu;
}
+
+static void tickle_timers(void *arg)
+{
+ hrtimer_peek_ahead_timers();
+}
+
#endif /* CONFIG_HOTPLUG_CPU */
static int __cpuinit hrtimer_cpu_notify(struct notifier_block *self,
unsigned long action, void *hcpu)
{
- unsigned int cpu = (long)hcpu;
+ int scpu = (long)hcpu;
switch (action) {
case CPU_UP_PREPARE:
case CPU_UP_PREPARE_FROZEN:
- init_hrtimers_cpu(cpu);
+ init_hrtimers_cpu(scpu);
break;
#ifdef CONFIG_HOTPLUG_CPU
case CPU_DEAD:
case CPU_DEAD_FROZEN:
- clockevents_notify(CLOCK_EVT_NOTIFY_CPU_DEAD, &cpu);
- migrate_hrtimers(cpu);
+ {
+ int dcpu;
+
+ clockevents_notify(CLOCK_EVT_NOTIFY_CPU_DEAD, &scpu);
+ dcpu = migrate_hrtimers(scpu);
+ smp_call_function_single(dcpu, tickle_timers, NULL, 0);
break;
+ }
#endif
default:
hrtimer_cpu_notify(&hrtimers_nb, (unsigned long)CPU_UP_PREPARE,
(void *)(long)smp_processor_id());
register_cpu_notifier(&hrtimers_nb);
-#ifdef CONFIG_HIGH_RES_TIMERS
- open_softirq(HRTIMER_SOFTIRQ, run_hrtimer_softirq);
-#endif
}
/**
* must supply functions here, even if the function just returns
* ENOSYS. The standard POSIX timer management code assumes the
* following: 1.) The k_itimer struct (sched.h) is used for the
- * timer. 2.) The list, it_lock, it_clock, it_id and it_process
+ * timer. 2.) The list, it_lock, it_clock, it_id and it_pid
* fields are not modified by timer code.
*
* At this time all functions EXCEPT clock_nanosleep can be
int posix_timer_event(struct k_itimer *timr, int si_private)
{
- int shared, ret;
+ struct task_struct *task;
+ int shared, ret = -1;
/*
* FIXME: if ->sigq is queued we can race with
* dequeue_signal()->do_schedule_next_timer().
*/
timr->sigq->info.si_sys_private = si_private;
- shared = !(timr->it_sigev_notify & SIGEV_THREAD_ID);
- ret = send_sigqueue(timr->sigq, timr->it_process, shared);
+ rcu_read_lock();
+ task = pid_task(timr->it_pid, PIDTYPE_PID);
+ if (task) {
+ shared = !(timr->it_sigev_notify & SIGEV_THREAD_ID);
+ ret = send_sigqueue(timr->sigq, task, shared);
+ }
+ rcu_read_unlock();
/* If we failed to send the signal the timer stops. */
return ret > 0;
}
return ret;
}
-static struct task_struct * good_sigevent(sigevent_t * event)
+static struct pid *good_sigevent(sigevent_t * event)
{
struct task_struct *rtn = current->group_leader;
((event->sigev_signo <= 0) || (event->sigev_signo > SIGRTMAX)))
return NULL;
- return rtn;
+ return task_pid(rtn);
}
void register_posix_clock(const clockid_t clock_id, struct k_clock *new_clock)
idr_remove(&posix_timers_id, tmr->it_id);
spin_unlock_irqrestore(&idr_lock, flags);
}
+ put_pid(tmr->it_pid);
sigqueue_free(tmr->sigq);
kmem_cache_free(posix_timers_cache, tmr);
}
{
struct k_itimer *new_timer;
int error, new_timer_id;
- struct task_struct *process;
sigevent_t event;
int it_id_set = IT_ID_NOT_SET;
goto out;
}
rcu_read_lock();
- process = good_sigevent(&event);
- if (process)
- get_task_struct(process);
+ new_timer->it_pid = get_pid(good_sigevent(&event));
rcu_read_unlock();
- if (!process) {
+ if (!new_timer->it_pid) {
error = -EINVAL;
goto out;
}
event.sigev_notify = SIGEV_SIGNAL;
event.sigev_signo = SIGALRM;
event.sigev_value.sival_int = new_timer->it_id;
- process = current->group_leader;
- get_task_struct(process);
+ new_timer->it_pid = get_pid(task_tgid(current));
}
new_timer->it_sigev_notify = event.sigev_notify;
new_timer->sigq->info.si_code = SI_TIMER;
spin_lock_irq(¤t->sighand->siglock);
- new_timer->it_process = process;
+ new_timer->it_signal = current->signal;
list_add(&new_timer->list, ¤t->signal->posix_timers);
spin_unlock_irq(¤t->sighand->siglock);
timr = idr_find(&posix_timers_id, (int)timer_id);
if (timr) {
spin_lock(&timr->it_lock);
- if (timr->it_process &&
- same_thread_group(timr->it_process, current)) {
+ if (timr->it_signal == current->signal) {
spin_unlock(&idr_lock);
return timr;
}
* This keeps any tasks waiting on the spin lock from thinking
* they got something (see the lock code above).
*/
- put_task_struct(timer->it_process);
- timer->it_process = NULL;
+ timer->it_signal = NULL;
unlock_timer(timer, flags);
release_posix_timer(timer, IT_ID_SET);
* This keeps any tasks waiting on the spin lock from thinking
* they got something (see the lock code above).
*/
- put_task_struct(timer->it_process);
- timer->it_process = NULL;
+ timer->it_signal = NULL;
unlock_timer(timer, flags);
release_posix_timer(timer, IT_ID_SET);