[Devel] [PATCH] ms/sched: Fix hrtimer_cancel()/rq->lock deadlock

[Kirill Tkhai]

ms commit 927b54fccbf0 by Ben Segall <bsegall at google.com>

__start_cfs_bandwidth calls hrtimer_cancel while holding rq->lock,
waiting for the hrtimer to finish. However, if sched_cfs_period_timer
runs for another loop iteration, the hrtimer can attempt to take
rq->lock, resulting in deadlock.

Fix this by ensuring that cfs_b->timer_active is cleared only if the
_latest_ call to do_sched_cfs_period_timer is returning as idle. Then
__start_cfs_bandwidth can just call hrtimer_try_to_cancel and wait for
that to succeed or timer_active == 1.

Signed-off-by: Ben Segall <bsegall at google.com>
Signed-off-by: Peter Zijlstra <peterz at infradead.org>
Cc: pjt at google.com
Link: http://lkml.kernel.org/r/20131016181622.22647.16643.stgit@sword-of-the-dawn.mtv.corp.google.com
Signed-off-by: Ingo Molnar <mingo at kernel.org>
Signed-off-by: Kirill Tkhai <ktkhai at virtuozzo.com>
---
 kernel/sched/fair.c |   15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 96fe1d5..e9dbb90 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -3690,6 +3690,13 @@ static int do_sched_cfs_period_timer(struct cfs_bandwidth *cfs_b, int overrun)
 	if (idle)
 		goto out_unlock;
 
+	/*
+	 * if we have relooped after returning idle once, we need to update our
+	 * status as actually running, so that other cpus doing
+	 * __start_cfs_bandwidth will stop trying to cancel us.
+	 */
+	cfs_b->timer_active = 1;
+
 	__refill_cfs_bandwidth_runtime(cfs_b);
 
 	if (!throttled) {
@@ -3986,11 +3993,11 @@ void __start_cfs_bandwidth(struct cfs_bandwidth *cfs_b)
 	 * (timer_active==0 becomes visible before the hrtimer call-back
 	 * terminates).  In either case we ensure that it's re-programmed
 	 */
-	while (unlikely(hrtimer_active(&cfs_b->period_timer))) {
+	while (unlikely(hrtimer_active(&cfs_b->period_timer)) &&
+	       hrtimer_try_to_cancel(&cfs_b->period_timer) < 0) {
+		/* bounce the lock to allow do_sched_cfs_period_timer to run */
 		raw_spin_unlock(&cfs_b->lock);
-		/* ensure cfs_b->lock is available while we wait */
-		hrtimer_cancel(&cfs_b->period_timer);
-
+		cpu_relax();
 		raw_spin_lock(&cfs_b->lock);
 		/* if someone else restarted the timer then we're done */
 		if (cfs_b->timer_active)