Commit fa2563e41c3d6d6e8af437643981ed28ae0cb56d

Authored by Thomas Tuttle
Committed by Linus Torvalds
1 parent df4e33ad24

workqueue: lock cwq access in drain_workqueue

Take cwq->gcwq->lock to avoid racing between drain_workqueue checking to
make sure the workqueues are empty and cwq_dec_nr_in_flight decrementing
and then incrementing nr_active when it activates a delayed work.

We discovered this when a corner case in one of our drivers resulted in
us trying to destroy a workqueue in which the remaining work would
always requeue itself again in the same workqueue.  We would hit this
race condition and trip the BUG_ON on workqueue.c:3080.

Signed-off-by: Thomas Tuttle <ttuttle@chromium.org>
Acked-by: Tejun Heo <tj@kernel.org>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Showing 1 changed file with 6 additions and 1 deletions Side-by-side Diff

... ... @@ -2412,8 +2412,13 @@
2412 2412  
2413 2413 for_each_cwq_cpu(cpu, wq) {
2414 2414 struct cpu_workqueue_struct *cwq = get_cwq(cpu, wq);
  2415 + bool drained;
2415 2416  
2416   - if (!cwq->nr_active && list_empty(&cwq->delayed_works))
  2417 + spin_lock_irq(&cwq->gcwq->lock);
  2418 + drained = !cwq->nr_active && list_empty(&cwq->delayed_works);
  2419 + spin_unlock_irq(&cwq->gcwq->lock);
  2420 +
  2421 + if (drained)
2417 2422 continue;
2418 2423  
2419 2424 if (++flush_cnt == 10 ||