epoll: call final ep_events_available() check under the lock

There is a possible race when ep_scan_ready_list() leaves ->rdllist and
->obflist empty for a short period of time although some events are
pending.  It is quite likely that ep_events_available() observes empty
lists and goes to sleep.

Since commit 339ddb53d373 ("fs/epoll: remove unnecessary wakeups of
nested epoll") we are conservative in wakeups (there is only one place
for wakeup and this is ep_poll_callback()), thus ep_events_available()
must always observe correct state of two lists.

The easiest and correct way is to do the final check under the lock.
This does not impact the performance, since lock is taken anyway for
adding a wait entry to the wait queue.

The discussion of the problem can be found here:

   https://lore.kernel.org/linux-fsdevel/a2f22c3c-c25a-4bda-8339-a7bdaf17849e@akamai.com/

In this patch barrierless __set_current_state() is used.  This is safe
since waitqueue_active() is called under the same lock on wakeup side.

Short-circuit for fatal signals (i.e.  fatal_signal_pending() check) is
moved to the line just before actual events harvesting routine.  This is
fully compliant to what is said in the comment of the patch where the
actual fatal_signal_pending() check was added: c257a340ede0 ("fs, epoll:
short circuit fetching events if thread has been killed").

Fixes: 339ddb53d373 ("fs/epoll: remove unnecessary wakeups of nested epoll")
Reported-by: Jason Baron <jbaron@akamai.com>
Reported-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Roman Penyaev <rpenyaev@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Jason Baron <jbaron@akamai.com>
Cc: Khazhismel Kumykov <khazhy@google.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: <stable@vger.kernel.org>
Link: http://lkml.kernel.org/r/20200505145609.1865152-1-rpenyaev@suse.de
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

1 files changed, 28 insertions, 20 deletions

diff --git a/fs/eventpoll.c b/fs/eventpoll.c
index aba03ee749f8..12eebcdea9c8 100644
--- a/fs/eventpoll.c
+++ b/fs/eventpoll.c
@@ -1879,34 +1879,33 @@ fetch_events:
 		 * event delivery.
 		 */
 		init_wait(&wait);
-		write_lock_irq(&ep->lock);
-		__add_wait_queue_exclusive(&ep->wq, &wait);
-		write_unlock_irq(&ep->lock);
 
+		write_lock_irq(&ep->lock);
 		/*
-		 * We don't want to sleep if the ep_poll_callback() sends us
-		 * a wakeup in between. That's why we set the task state
-		 * to TASK_INTERRUPTIBLE before doing the checks.
+		 * Barrierless variant, waitqueue_active() is called under
+		 * the same lock on wakeup ep_poll_callback() side, so it
+		 * is safe to avoid an explicit barrier.
 		 */
-		set_current_state(TASK_INTERRUPTIBLE);
+		__set_current_state(TASK_INTERRUPTIBLE);
+
 		/*
-		 * Always short-circuit for fatal signals to allow
-		 * threads to make a timely exit without the chance of
-		 * finding more events available and fetching
-		 * repeatedly.
+		 * Do the final check under the lock. ep_scan_ready_list()
+		 * plays with two lists (->rdllist and ->ovflist) and there
+		 * is always a race when both lists are empty for short
+		 * period of time although events are pending, so lock is
+		 * important.
 		 */
-		if (fatal_signal_pending(current)) {
-			res = -EINTR;
-			break;
+		eavail = ep_events_available(ep);
+		if (!eavail) {
+			if (signal_pending(current))
+				res = -EINTR;
+			else
+				__add_wait_queue_exclusive(&ep->wq, &wait);
 		}
+		write_unlock_irq(&ep->lock);
 
-		eavail = ep_events_available(ep);
-		if (eavail)
-			break;
-		if (signal_pending(current)) {
-			res = -EINTR;
+		if (eavail || res)
 			break;
-		}
 
 		if (!schedule_hrtimeout_range(to, slack, HRTIMER_MODE_ABS)) {
 			timed_out = 1;
@@ -1927,6 +1926,15 @@ fetch_events:
 	}
 
 send_events:
+	if (fatal_signal_pending(current)) {
+		/*
+		 * Always short-circuit for fatal signals to allow
+		 * threads to make a timely exit without the chance of
+		 * finding more events available and fetching
+		 * repeatedly.
+		 */
+		res = -EINTR;
+	}
 	/*
 	 * Try to transfer events to user space. In case we get 0 events and
 	 * there's still timeout left over, we go trying again in search of