Merge branch 'for-3.6-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq

4651afbb · Linus Torvalds <[email protected]> · 2012-09-17 16:05:23 -07:00

Pull another workqueue fix from Tejun Heo:
 "Unfortunately, yet another late fix.  This too is discovered and fixed
  by Lai.  This bug was introduced during this merge window by commit
  25511a477657 ("workqueue: reimplement CPU online rebinding to handle
  idle workers") which started using WORKER_REBIND flag for idle rebind
  too.

  The bug is relatively easy to trigger if the CPU rapidly goes through
  off, on and then off (and stay off).  The fix is on the safer side.
  This hasn't been on linux-next yet but I'm pushing early so that it
  can get more exposure before v3.6 release."

* 'for-3.6-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq:
  workqueue: always clear WORKER_REBIND in busy_worker_rebind_fn()

Commit: 4651afbbae968772efd6dc4ba461cba9b49bb9d8
Parents: 08077ca8 · 960bd11b

Merge commit. Changes below are compared with the first parent: 08077ca8.

Changes

1 files changed · +10 −2

File	Status	Add	Del	Content
kernel/workqueue.c	modified	+10	−2	View file

kernel/workqueue.c

View file

diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index 1e1373b..b80065a 100644
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -1349,8 +1349,16 @@ static void busy_worker_rebind_fn(struct work_struct *work)
 	struct worker *worker = container_of(work, struct worker, rebind_work);
 	struct global_cwq *gcwq = worker->pool->gcwq;
 
-	if (worker_maybe_bind_and_lock(worker))
-		worker_clr_flags(worker, WORKER_REBIND);
+	worker_maybe_bind_and_lock(worker);
+
+	/*
+	 * %WORKER_REBIND must be cleared even if the above binding failed;
+	 * otherwise, we may confuse the next CPU_UP cycle or oops / get
+	 * stuck by calling idle_worker_rebind() prematurely.  If CPU went
+	 * down again inbetween, %WORKER_UNBOUND would be set, so clearing
+	 * %WORKER_REBIND is always safe.
+	 */
+	worker_clr_flags(worker, WORKER_REBIND);
 
 	spin_unlock_irq(&gcwq->lock);
 }