epoll: fix possible lost wakeup on epoll_ctl() path
This fixes possible lost wakeup introduced by commit a218cc491420.
Originally modifications to ep->wq were serialized by ep->wq.lock, but
in commit a218cc491420 ("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():
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: a218cc491420 ("epoll: use rwlock in order to reduce ep_poll_callback() contention")
Reported-by: Max Neunhoeffer <firstname.lastname@example.org>
Signed-off-by: Roman Penyaev <email@example.com>
Signed-off-by: Andrew Morton <firstname.lastname@example.org>
Tested-by: Max Neunhoeffer <email@example.com>
Cc: Jakub Kicinski <firstname.lastname@example.org>
Cc: Christopher Kohlhoff <email@example.com>
Cc: Davidlohr Bueso <firstname.lastname@example.org>
Cc: Jason Baron <email@example.com>
Cc: Jes Sorensen <firstname.lastname@example.org>
Cc: <email@example.com> [5.1+]
Bisected-by: Max Neunhoeffer <firstname.lastname@example.org>
Signed-off-by: Linus Torvalds <email@example.com>
1 file changed