From 6ee9f8df3bfbb811526cff3313da5758b1277bc6 Mon Sep 17 00:00:00 2001 From: Willy Tarreau Date: Mon, 26 Aug 2019 10:55:52 +0200 Subject: [PATCH] BUG/MEDIUM: listener/threads: fix an AB/BA locking issue in delete_listener() The delete_listener() function takes the listener's lock before taking the proto_lock, which is contrary to what other functions do, possibly causing an AB/BA deadlock. In practice the two only places where both are taken are during protocol_enable_all() and delete_listener(), the former being used during startup and the latter during stop. In practice during reload floods, it is technically possible for a thread to be initializing the listeners while another one is stopping. While this is too hard to trigger on 2.0 and above due to the synchronization of all threads during startup, it's reasonably easy to do in 1.9 by having hundreds of listeners, starting 64 threads and flooding them with reloads like this : $ while usleep 50000; do killall -USR2 haproxy; done Usually in less than a minute, all threads will be deadlocked. The fix consists in always taking the proto_lock before the listener lock. It seems to be the only place where these two locks were reversed. This fix needs to be backported to 2.0, 1.9, and 1.8. --- src/listener.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/src/listener.c b/src/listener.c index b5fe2ac21..54c099602 100644 --- a/src/listener.c +++ b/src/listener.c @@ -595,17 +595,17 @@ int create_listeners(struct bind_conf *bc, const struct sockaddr_storage *ss, */ void delete_listener(struct listener *listener) { + HA_SPIN_LOCK(PROTO_LOCK, &proto_lock); HA_SPIN_LOCK(LISTENER_LOCK, &listener->lock); if (listener->state == LI_ASSIGNED) { listener->state = LI_INIT; - HA_SPIN_LOCK(PROTO_LOCK, &proto_lock); LIST_DEL(&listener->proto_list); listener->proto->nb_listeners--; - HA_SPIN_UNLOCK(PROTO_LOCK, &proto_lock); _HA_ATOMIC_SUB(&jobs, 1); _HA_ATOMIC_SUB(&listeners, 1); } HA_SPIN_UNLOCK(LISTENER_LOCK, &listener->lock); + HA_SPIN_UNLOCK(PROTO_LOCK, &proto_lock); } /* Returns a suitable value for a listener's backlog. It uses the listener's,