1 files changed, 115 insertions, 77 deletions

diff --git a/kernel/hrtimer.c b/kernel/hrtimer.c
index bda9cb92427..f394d2a42ca 100644
--- a/kernel/hrtimer.c
+++ b/kernel/hrtimer.c
@@ -32,7 +32,6 @@
  */
 
 #include <linux/cpu.h>
-#include <linux/irq.h>
 #include <linux/module.h>
 #include <linux/percpu.h>
 #include <linux/hrtimer.h>
@@ -502,6 +501,13 @@ static void hrtimer_force_reprogram(struct hrtimer_cpu_base *cpu_base)
 			continue;
 		timer = rb_entry(base->first, struct hrtimer, node);
 		expires = ktime_sub(hrtimer_get_expires(timer), base->offset);
+		/*
+		 * clock_was_set() has changed base->offset so the
+		 * result might be negative. Fix it up to prevent a
+		 * false positive in clockevents_program_event()
+		 */
+		if (expires.tv64 < 0)
+			expires.tv64 = 0;
 		if (expires.tv64 < cpu_base->expires_next.tv64)
 			cpu_base->expires_next = expires;
 	}
@@ -615,7 +621,9 @@ void clock_was_set(void)
  */
 void hres_timers_resume(void)
 {
-	/* Retrigger the CPU local events: */
+	WARN_ONCE(!irqs_disabled(),
+		  KERN_INFO "hres_timers_resume() called with IRQs enabled!");
+
 	retrigger_next_event(NULL);
 }
 
@@ -635,7 +643,6 @@ static inline void hrtimer_init_timer_hres(struct hrtimer *timer)
 {
 }
 
-static void __run_hrtimer(struct hrtimer *timer);
 
 /*
  * When High resolution timers are active, try to reprogram. Note, that in case
@@ -647,13 +654,9 @@ static inline int hrtimer_enqueue_reprogram(struct hrtimer *timer,
 					    struct hrtimer_clock_base *base)
 {
 	if (base->cpu_base->hres_active && hrtimer_reprogram(timer, base)) {
-		/*
-		 * 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);
+		spin_unlock(&base->cpu_base->lock);
+		raise_softirq_irqoff(HRTIMER_SOFTIRQ);
+		spin_lock(&base->cpu_base->lock);
 		return 1;
 	}
 	return 0;
@@ -706,11 +709,6 @@ static inline int hrtimer_enqueue_reprogram(struct hrtimer *timer,
 }
 static inline void hrtimer_init_hres(struct hrtimer_cpu_base *base) { }
 static inline void hrtimer_init_timer_hres(struct hrtimer *timer) { }
-static inline int hrtimer_reprogram(struct hrtimer *timer,
-				    struct hrtimer_clock_base *base)
-{
-	return 0;
-}
 
 #endif /* CONFIG_HIGH_RES_TIMERS */
 
@@ -781,9 +779,11 @@ EXPORT_SYMBOL_GPL(hrtimer_forward);
  *
  * The timer is inserted in expiry order. Insertion into the
  * red black tree is O(log(n)). Must hold the base lock.
+ *
+ * Returns 1 when the new timer is the leftmost timer in the tree.
  */
-static void enqueue_hrtimer(struct hrtimer *timer,
-			    struct hrtimer_clock_base *base, int reprogram)
+static int enqueue_hrtimer(struct hrtimer *timer,
+			   struct hrtimer_clock_base *base)
 {
 	struct rb_node **link = &base->active.rb_node;
 	struct rb_node *parent = NULL;
@@ -815,20 +815,8 @@ static void enqueue_hrtimer(struct hrtimer *timer,
 	 * Insert the timer to the rbtree and check whether it
 	 * replaces the first pending timer
 	 */
-	if (leftmost) {
-		/*
-		 * Reprogram the clock event device. When the timer is already
-		 * expired hrtimer_enqueue_reprogram has either called the
-		 * callback or added it to the pending list and raised the
-		 * softirq.
-		 *
-		 * This is a NOP for !HIGHRES
-		 */
-		if (reprogram && hrtimer_enqueue_reprogram(timer, base))
-			return;
-
+	if (leftmost)
 		base->first = &timer->node;
-	}
 
 	rb_link_node(&timer->node, parent, link);
 	rb_insert_color(&timer->node, &base->active);
@@ -837,6 +825,8 @@ static void enqueue_hrtimer(struct hrtimer *timer,
 	 * state of a possibly running callback.
 	 */
 	timer->state |= HRTIMER_STATE_ENQUEUED;
+
+	return leftmost;
 }
 
 /*
@@ -913,7 +903,7 @@ hrtimer_start_range_ns(struct hrtimer *timer, ktime_t tim, unsigned long delta_n
 {
 	struct hrtimer_clock_base *base, *new_base;
 	unsigned long flags;
-	int ret;
+	int ret, leftmost;
 
 	base = lock_hrtimer_base(timer, &flags);
 
@@ -941,12 +931,16 @@ hrtimer_start_range_ns(struct hrtimer *timer, ktime_t tim, unsigned long delta_n
 
 	timer_stats_hrtimer_set_start_info(timer);
 
+	leftmost = enqueue_hrtimer(timer, new_base);
+
 	/*
 	 * Only allow reprogramming if the new base is on this CPU.
 	 * (it might still be on another CPU if the timer was pending)
+	 *
+	 * XXX send_remote_softirq() ?
 	 */
-	enqueue_hrtimer(timer, new_base,
-			new_base->cpu_base == &__get_cpu_var(hrtimer_bases));
+	if (leftmost && new_base->cpu_base == &__get_cpu_var(hrtimer_bases))
+		hrtimer_enqueue_reprogram(timer, new_base);
 
 	unlock_hrtimer_base(timer, &flags);
 
@@ -1158,19 +1152,42 @@ static void __run_hrtimer(struct hrtimer *timer)
 	spin_lock(&cpu_base->lock);
 
 	/*
-	 * Note: We clear the CALLBACK bit after enqueue_hrtimer to avoid
-	 * reprogramming of the event hardware. This happens at the end of this
-	 * function anyway.
+	 * Note: We clear the CALLBACK bit after enqueue_hrtimer and
+	 * we do not reprogramm the event hardware. Happens either in
+	 * hrtimer_start_range_ns() or in hrtimer_interrupt()
 	 */
 	if (restart != HRTIMER_NORESTART) {
 		BUG_ON(timer->state != HRTIMER_STATE_CALLBACK);
-		enqueue_hrtimer(timer, base, 0);
+		enqueue_hrtimer(timer, base);
 	}
 	timer->state &= ~HRTIMER_STATE_CALLBACK;
 }
 
 #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
@@ -1180,6 +1197,7 @@ void hrtimer_interrupt(struct clock_event_device *dev)
 	struct hrtimer_cpu_base *cpu_base = &__get_cpu_var(hrtimer_bases);
 	struct hrtimer_clock_base *base;
 	ktime_t expires_next, now;
+	int nr_retries = 0;
 	int i;
 
 	BUG_ON(!cpu_base->hres_active);
@@ -1187,6 +1205,10 @@ void hrtimer_interrupt(struct clock_event_device *dev)
 	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;
@@ -1239,11 +1261,27 @@ void hrtimer_interrupt(struct clock_event_device *dev)
 
 	/* 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;
 	}
 }
 
+/*
+ * local version of hrtimer_peek_ahead_timers() called with interrupts
+ * disabled.
+ */
+static void __hrtimer_peek_ahead_timers(void)
+{
+	struct tick_device *td;
+
+	if (!hrtimer_hres_active())
+		return;
+
+	td = &__get_cpu_var(tick_cpu_device);
+	if (td && td->evtdev)
+		hrtimer_interrupt(td->evtdev);
+}
+
 /**
  * hrtimer_peek_ahead_timers -- run soft-expired timers now
  *
@@ -1255,20 +1293,23 @@ void hrtimer_interrupt(struct clock_event_device *dev)
  */
 void hrtimer_peek_ahead_timers(void)
 {
-	struct tick_device *td;
 	unsigned long flags;
 
-	if (!hrtimer_hres_active())
-		return;
-
 	local_irq_save(flags);
-	td = &__get_cpu_var(tick_cpu_device);
-	if (td && td->evtdev)
-		hrtimer_interrupt(td->evtdev);
+	__hrtimer_peek_ahead_timers();
 	local_irq_restore(flags);
 }
 
-#endif	/* CONFIG_HIGH_RES_TIMERS */
+static void run_hrtimer_softirq(struct softirq_action *h)
+{
+	hrtimer_peek_ahead_timers();
+}
+
+#else /* CONFIG_HIGH_RES_TIMERS */
+
+static inline void __hrtimer_peek_ahead_timers(void) { }
+
+#endif	/* !CONFIG_HIGH_RES_TIMERS */
 
 /*
  * Called from timer softirq every jiffy, expire hrtimers:
@@ -1463,8 +1504,8 @@ out:
 	return ret;
 }
 
-asmlinkage long
-sys_nanosleep(struct timespec __user *rqtp, struct timespec __user *rmtp)
+SYSCALL_DEFINE2(nanosleep, struct timespec __user *, rqtp,
+		struct timespec __user *, rmtp)
 {
 	struct timespec tu;
 
@@ -1514,39 +1555,36 @@ static void migrate_hrtimer_list(struct hrtimer_clock_base *old_base,
 		__remove_hrtimer(timer, old_base, HRTIMER_STATE_MIGRATE, 0);
 		timer->base = new_base;
 		/*
-		 * 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 the timers on the new cpu. This does not
+		 * reprogram the event device in case the timer
+		 * expires before the earliest on this CPU, but we run
+		 * hrtimer_interrupt after we migrated everything to
+		 * sort out already expired timers and reprogram the
+		 * event device.
 		 */
-		enqueue_hrtimer(timer, new_base, 0);
+		enqueue_hrtimer(timer, new_base);
 
 		/* Clear the migration state bit */
 		timer->state &= ~HRTIMER_STATE_MIGRATE;
 	}
 }
 
-static int migrate_hrtimers(int scpu)
+static void migrate_hrtimers(int scpu)
 {
 	struct hrtimer_cpu_base *old_base, *new_base;
-	int dcpu, i;
+	int i;
 
 	BUG_ON(cpu_online(scpu));
-	old_base = &per_cpu(hrtimer_bases, scpu);
-	new_base = &get_cpu_var(hrtimer_bases);
-
-	dcpu = smp_processor_id();
-
 	tick_cancel_sched_timer(scpu);
+
+	local_irq_disable();
+	old_base = &per_cpu(hrtimer_bases, scpu);
+	new_base = &__get_cpu_var(hrtimer_bases);
 	/*
 	 * The caller is globally serialized and nobody else
 	 * takes two locks at once, deadlock is not possible.
 	 */
-	spin_lock_irq(&new_base->lock);
+	spin_lock(&new_base->lock);
 	spin_lock_nested(&old_base->lock, SINGLE_DEPTH_NESTING);
 
 	for (i = 0; i < HRTIMER_MAX_CLOCK_BASES; i++) {
@@ -1555,15 +1593,11 @@ static int migrate_hrtimers(int scpu)
 	}
 
 	spin_unlock(&old_base->lock);
-	spin_unlock_irq(&new_base->lock);
-	put_cpu_var(hrtimer_bases);
+	spin_unlock(&new_base->lock);
 
-	return dcpu;
-}
-
-static void tickle_timers(void *arg)
-{
-	hrtimer_peek_ahead_timers();
+	/* Check, if we got expired work to do */
+	__hrtimer_peek_ahead_timers();
+	local_irq_enable();
 }
 
 #endif /* CONFIG_HOTPLUG_CPU */
@@ -1581,14 +1615,15 @@ static int __cpuinit hrtimer_cpu_notify(struct notifier_block *self,
 		break;
 
 #ifdef CONFIG_HOTPLUG_CPU
+	case CPU_DYING:
+	case CPU_DYING_FROZEN:
+		clockevents_notify(CLOCK_EVT_NOTIFY_CPU_DYING, &scpu);
+		break;
 	case CPU_DEAD:
 	case CPU_DEAD_FROZEN:
 	{
-		int dcpu;
-
 		clockevents_notify(CLOCK_EVT_NOTIFY_CPU_DEAD, &scpu);
-		dcpu = migrate_hrtimers(scpu);
-		smp_call_function_single(dcpu, tickle_timers, NULL, 0);
+		migrate_hrtimers(scpu);
 		break;
 	}
 #endif
@@ -1609,6 +1644,9 @@ void __init hrtimers_init(void)
 	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
 }
 
 /**