]> www.infradead.org Git - users/jedix/linux-maple.git/commitdiff
futex: Acknowledge a new waiter in counter before plist
authorDavidlohr Bueso <dave@stgolabs.net>
Thu, 21 Apr 2016 03:09:24 +0000 (20:09 -0700)
committerDhaval Giani <dhaval.giani@oracle.com>
Fri, 20 Jan 2017 22:22:01 +0000 (17:22 -0500)
Orabug: 25308063

[ Upstream commit fe1bce9e2107ba3a8faffe572483b6974201a0e6 ]

Otherwise an incoming waker on the dest hash bucket can miss
the waiter adding itself to the plist during the lockless
check optimization (small window but still the correct way
of doing this); similarly to the decrement counterpart.

Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: bigeasy@linutronix.de
Cc: dvhart@infradead.org
Cc: stable@kernel.org
Link: http://lkml.kernel.org/r/1461208164-29150-1-git-send-email-dave@stgolabs.net
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
(cherry picked from commit 0ee1b6dbc57dc78541df83157e1251b983c51485)
Signed-off-by: Dhaval Giani <dhaval.giani@oracle.com>
kernel/futex.c

index 46b168e19c984ddeed3a0691dbee08bcede6f557..2214b70f191047b6962a1951a48be872028ac274 100644 (file)
@@ -1381,8 +1381,8 @@ void requeue_futex(struct futex_q *q, struct futex_hash_bucket *hb1,
        if (likely(&hb1->chain != &hb2->chain)) {
                plist_del(&q->list, &hb1->chain);
                hb_waiters_dec(hb1);
-               plist_add(&q->list, &hb2->chain);
                hb_waiters_inc(hb2);
+               plist_add(&q->list, &hb2->chain);
                q->lock_ptr = &hb2->lock;
        }
        get_futex_key_refs(key2);