Re: [PATCH] timer: add_timer_on() should perform proper migration

[Thomas Gleixner]

Tejun,

On Wed, 4 Nov 2015, Tejun Heo wrote:

> Regardless of the previous CPU a timer was on, add_timer_on()
> currently simply sets timer->flags to the new CPU.  As the caller must
> be seeing the timer as idle, this is locally fine, but the timer
> leaving the old base while unlocked can lead to race conditions as
> follows.

nice detective work. This has been there forever. I really wonder why
nobody ever triggered this before.

@stable: 

The patch does only apply to kernels >= 4.2. Backport for older
kernels is below.

Thanks,

	tglx

----------->

--- a/kernel/time/timer.c
+++ b/kernel/time/timer.c
@@ -966,13 +966,26 @@ EXPORT_SYMBOL(add_timer);
  */
 void add_timer_on(struct timer_list *timer, int cpu)
 {
-	struct tvec_base *base = per_cpu(tvec_bases, cpu);
+	struct tvec_base *new_base = per_cpu(tvec_bases, cpu);
+	struct tvec_base *base;
 	unsigned long flags;
 
 	timer_stats_timer_set_start_info(timer);
 	BUG_ON(timer_pending(timer) || !timer->function);
-	spin_lock_irqsave(&base->lock, flags);
-	timer_set_base(timer, base);
+
+	/*
+	 * If @timer was on a different CPU, it should be migrated with the
+	 * old base locked to prevent other operations proceeding with the
+	 * wrong base locked.  See lock_timer_base().
+	 */
+	base = lock_timer_base(timer, &flags);
+	if (base != new_base) {
+		timer_set_base(timer, NULL);
+		spin_unlock(&base->lock);
+		base = new_base;
+		spin_lock(&base->lock);
+		timer_set_base(timer, base);
+	}
 	debug_activate(timer, timer->expires);
 	internal_add_timer(base, timer);
 	spin_unlock_irqrestore(&base->lock, flags);


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/