timer: create timer_quiesce_cpu() to isolate CPU from timers

To isolate CPUs (isolate from timers) from sysfs using cpusets, we need some
support from the timer core. i.e. A routine timer_quiesce_cpu() which would
migrates away all the unpinned timers, but shouldn't touch the pinned ones.

This patch creates this routine.

Change-Id: I8624e0659b86b7b8fa425a3fafdb0784fe005124
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
[forward port to 3.18]
Signed-off-by: Santosh Shukla <santosh.shukla@linaro.org>
[ohaugan@codeaurora.org: Port to 4.4. Fixes for compilation error]
Git-commit: 313910b70ea0c73f8789d9189c11e1f339080646
Git-repo: git://git.linaro.org/people/mike.holmes/santosh.shukla/lng-isol.git
Signed-off-by: Olav Haugan <ohaugan@codeaurora.org>

1 files changed, 35 insertions, 21 deletions

diff --git a/kernel/time/timer.c b/kernel/time/timer.c
index be750f6b2a68..067174a4dde3 100644
--- a/kernel/time/timer.c
+++ b/kernel/time/timer.c
@@ -1620,44 +1620,49 @@ signed long __sched schedule_timeout_uninterruptible(signed long timeout)
 }
 EXPORT_SYMBOL(schedule_timeout_uninterruptible);
 
-#ifdef CONFIG_HOTPLUG_CPU
-static void migrate_timer_list(struct tvec_base *new_base, struct hlist_head *head)
+#if defined(CONFIG_HOTPLUG_CPU)
+static void migrate_timer_list(struct tvec_base *new_base,
+			       struct hlist_head *head, bool remove_pinned)
 {
 	struct timer_list *timer;
 	int cpu = new_base->cpu;
+	struct hlist_node *n;
+	int is_pinned;
 
-	while (!hlist_empty(head)) {
-		timer = hlist_entry(head->first, struct timer_list, entry);
-		/* We ignore the accounting on the dying cpu */
-		detach_timer(timer, false);
+	hlist_for_each_entry_safe(timer, n, head, entry) {
+		is_pinned = timer->flags & TIMER_PINNED_ON_CPU;
+		if (!remove_pinned && is_pinned)
+			continue;
+
+		detach_if_pending(timer, get_timer_base(timer->flags), false);
 		timer->flags = (timer->flags & ~TIMER_BASEMASK) | cpu;
 		internal_add_timer(new_base, timer);
 	}
 }
 
-static void migrate_timers(int cpu, bool wait)
+static void __migrate_timers(int cpu, bool wait, bool remove_pinned)
 {
 	struct tvec_base *old_base;
 	struct tvec_base *new_base;
+	unsigned long flags;
 	int i;
 
-	BUG_ON(cpu_online(cpu));
 	old_base = per_cpu_ptr(&tvec_bases, cpu);
 	new_base = get_cpu_ptr(&tvec_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_irqsave(&new_base->lock, flags);
 	spin_lock_nested(&old_base->lock, SINGLE_DEPTH_NESTING);
 
 	if (wait) {
 		/* Ensure timers are done running before continuing */
 		while (old_base->running_timer) {
 			spin_unlock(&old_base->lock);
-			spin_unlock_irq(&new_base->lock);
+			spin_unlock_irqrestore(&new_base->lock, flags);
 			cpu_relax();
-			spin_lock_irq(&new_base->lock);
+			spin_lock_irqsave(&new_base->lock, flags);
 			spin_lock_nested(&old_base->lock, SINGLE_DEPTH_NESTING);
 		}
 	} else {
@@ -1665,29 +1670,38 @@ static void migrate_timers(int cpu, bool wait)
 	}
 
 	for (i = 0; i < TVR_SIZE; i++)
-		migrate_timer_list(new_base, old_base->tv1.vec + i);
+		migrate_timer_list(new_base, old_base->tv1.vec + i,
+				   remove_pinned);
 	for (i = 0; i < TVN_SIZE; i++) {
-		migrate_timer_list(new_base, old_base->tv2.vec + i);
-		migrate_timer_list(new_base, old_base->tv3.vec + i);
-		migrate_timer_list(new_base, old_base->tv4.vec + i);
-		migrate_timer_list(new_base, old_base->tv5.vec + i);
+		migrate_timer_list(new_base, old_base->tv2.vec + i,
+				remove_pinned);
+		migrate_timer_list(new_base, old_base->tv3.vec + i,
+				remove_pinned);
+		migrate_timer_list(new_base, old_base->tv4.vec + i,
+				remove_pinned);
+		migrate_timer_list(new_base, old_base->tv5.vec + i,
+				remove_pinned);
 	}
 
-	old_base->active_timers = 0;
-	old_base->all_timers = 0;
-
 	spin_unlock(&old_base->lock);
-	spin_unlock_irq(&new_base->lock);
+	spin_unlock_irqrestore(&new_base->lock, flags);
 	put_cpu_ptr(&tvec_bases);
 }
 
+/* Migrate timers from 'cpu' to this_cpu */
+static void migrate_timers(int cpu)
+{
+	BUG_ON(cpu_online(cpu));
+	__migrate_timers(cpu, false, true);
+}
+
 static int timer_cpu_notify(struct notifier_block *self,
 				unsigned long action, void *hcpu)
 {
 	switch (action) {
 	case CPU_DEAD:
 	case CPU_DEAD_FROZEN:
-		migrate_timers((long)hcpu, false);
+		migrate_timers((long)hcpu);
 		break;
 	default:
 		break;