Commit d436edfa authored by Roman Penyaev's avatar Roman Penyaev Committed by Greg Kroah-Hartman

epoll: fix possible lost wakeup on epoll_ctl() path

commit 1b53734b upstream.

This fixes possible lost wakeup introduced by commit a218cc49.
Originally modifications to ep->wq were serialized by ep->wq.lock, but
in commit a218cc49 ("epoll: use rwlock in order to reduce
ep_poll_callback() contention") a new rw lock was introduced in order to
relax fd event path, i.e. callers of ep_poll_callback() function.

After the change ep_modify and ep_insert (both are called on epoll_ctl()
path) were switched to ep->lock, but ep_poll (epoll_wait) was using
ep->wq.lock on wqueue list modification.

The bug doesn't lead to any wqueue list corruptions, because wake up
path and list modifications were serialized by ep->wq.lock internally,
but actual waitqueue_active() check prior wake_up() call can be
reordered with modifications of ep ready list, thus wake up can be lost.

And yes, can be healed by explicit smp_mb():

  list_add_tail(&epi->rdlink, &ep->rdllist);
  smp_mb();
  if (waitqueue_active(&ep->wq))
	wake_up(&ep->wp);

But let's make it simple, thus current patch replaces ep->wq.lock with
the ep->lock for wqueue modifications, thus wake up path always observes
activeness of the wqueue correcty.

Fixes: a218cc49 ("epoll: use rwlock in order to reduce ep_poll_callback() contention")
Reported-by: default avatarMax Neunhoeffer <[email protected]>
Signed-off-by: default avatarRoman Penyaev <[email protected]>
Signed-off-by: default avatarAndrew Morton <[email protected]>
Tested-by: default avatarMax Neunhoeffer <[email protected]>
Cc: Jakub Kicinski <[email protected]>
Cc: Christopher Kohlhoff <[email protected]>
Cc: Davidlohr Bueso <[email protected]>
Cc: Jason Baron <[email protected]>
Cc: Jes Sorensen <[email protected]>
Cc: <[email protected]>	[5.1+]
Link: http://lkml.kernel.org/r/[email protected]
References: https://bugzilla.kernel.org/show_bug.cgi?id=205933Bisected-by: default avatarMax Neunhoeffer <[email protected]>
Signed-off-by: default avatarLinus Torvalds <[email protected]>
Signed-off-by: default avatarGreg Kroah-Hartman <[email protected]>
parent ba741d9e
......@@ -1860,9 +1860,9 @@ static int ep_poll(struct eventpoll *ep, struct epoll_event __user *events,
waiter = true;
init_waitqueue_entry(&wait, current);
spin_lock_irq(&ep->wq.lock);
write_lock_irq(&ep->lock);
__add_wait_queue_exclusive(&ep->wq, &wait);
spin_unlock_irq(&ep->wq.lock);
write_unlock_irq(&ep->lock);
}
for (;;) {
......@@ -1910,9 +1910,9 @@ static int ep_poll(struct eventpoll *ep, struct epoll_event __user *events,
goto fetch_events;
if (waiter) {
spin_lock_irq(&ep->wq.lock);
write_lock_irq(&ep->lock);
__remove_wait_queue(&ep->wq, &wait);
spin_unlock_irq(&ep->wq.lock);
write_unlock_irq(&ep->lock);
}
return res;
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment