workqueue: convert BUG_ON()s in __queue_delayed_work() to WARN_ON_ONCE()s
authorTejun Heo <tj@kernel.org>
Tue, 4 Dec 2012 15:40:39 +0000 (07:40 -0800)
committerTejun Heo <tj@kernel.org>
Tue, 4 Dec 2012 15:58:47 +0000 (07:58 -0800)
8852aac25e ("workqueue: mod_delayed_work_on() shouldn't queue timer on
0 delay") unexpectedly uncovered a very nasty abuse of delayed_work in
megaraid - it allocated work_struct, casted it to delayed_work and
then pass that into queue_delayed_work().

Previously, this was okay because 0 @delay short-circuited to
queue_work() before doing anything with delayed_work.  8852aac25e
moved 0 @delay test into __queue_delayed_work() after sanity check on
delayed_work making megaraid trigger BUG_ON().

Although megaraid is already fixed by c1d390d8e6 ("megaraid: fix
BUG_ON() from incorrect use of delayed work"), this patch converts
BUG_ON()s in __queue_delayed_work() to WARN_ON_ONCE()s so that such
abusers, if there are more, trigger warning but don't crash the
machine.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Xiaotian Feng <xtfeng@gmail.com>
kernel/workqueue.c

index 084aa47bac820498f932ae6ff54158414e48e889..1dae900df798d1c4e1be722567fd2906041f08f5 100644 (file)
@@ -1361,8 +1361,8 @@ static void __queue_delayed_work(int cpu, struct workqueue_struct *wq,
 
        WARN_ON_ONCE(timer->function != delayed_work_timer_fn ||
                     timer->data != (unsigned long)dwork);
-       BUG_ON(timer_pending(timer));
-       BUG_ON(!list_empty(&work->entry));
+       WARN_ON_ONCE(timer_pending(timer));
+       WARN_ON_ONCE(!list_empty(&work->entry));
 
        /*
         * If @delay is 0, queue @dwork->work immediately.  This is for