From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([209.51.188.92]:51333) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1gt4aP-0006ku-JG for qemu-devel@nongnu.org; Mon, 11 Feb 2019 00:56:59 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1gt4aO-0005RU-IM for qemu-devel@nongnu.org; Mon, 11 Feb 2019 00:56:57 -0500 From: Stefan Hajnoczi Date: Mon, 11 Feb 2019 13:50:32 +0800 Message-Id: <20190211055040.13528-2-stefanha@redhat.com> In-Reply-To: <20190211055040.13528-1-stefanha@redhat.com> References: <20190211055040.13528-1-stefanha@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Subject: [Qemu-devel] [PULL 1/9] iothread: fix iothread hang when stop too soon List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Cc: Marcel Apfelbaum , Peter Maydell , Thomas Huth , Stefan Hajnoczi , Eduardo Habkost , Laurent Vivier , Max Reitz , "Michael S. Tsirkin" , Paolo Bonzini , qemu-block@nongnu.org, Kevin Wolf , Peter Xu , "Dr . David Alan Gilbert" , =?UTF-8?q?Luk=C3=A1=C5=A1=20Doktor?= , Markus Armbruster , Eric Blake From: Peter Xu Lukas reported an hard to reproduce QMP iothread hang on s390 that QEMU might hang at pthread_join() of the QMP monitor iothread before quitting: Thread 1 #0 0x000003ffad10932c in pthread_join #1 0x0000000109e95750 in qemu_thread_join at /home/thuth/devel/qemu/util/qemu-thread-posix.c:570 #2 0x0000000109c95a1c in iothread_stop #3 0x0000000109bb0874 in monitor_cleanup #4 0x0000000109b55042 in main While the iothread is still in the main loop: Thread 4 #0 0x000003ffad0010e4 in ?? #1 0x000003ffad553958 in g_main_context_iterate.isra.19 #2 0x000003ffad553d90 in g_main_loop_run #3 0x0000000109c9585a in iothread_run at /home/thuth/devel/qemu/iothread.c:74 #4 0x0000000109e94752 in qemu_thread_start at /home/thuth/devel/qemu/util/qemu-thread-posix.c:502 #5 0x000003ffad10825a in start_thread #6 0x000003ffad00dcf2 in thread_start IMHO it's because there's a race between the main thread and iothread when stopping the thread in following sequence: main thread iothread =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D =3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D aio_poll() iothread_get_g_main_context set iothread->worker_context iothread_stop schedule iothread_stop_bh execute iothread_stop_bh [1] set iothread->running=3Dfalse (since main_loop=3D=3DNULL so skip to quit main loop. Note: although main_loop is NULL but worker_context is not!) atomic_read(&iothread->worker_conte= xt) [2] create main_loop object g_main_loop_run() [3] pthread_join() [4] We can see that when execute iothread_stop_bh() at [1] it's possible that main_loop is still NULL because it's only created until the first check of the worker_context later at [2]. Then the iothread will hang in the main loop [3] and it'll starve the main thread too [4]. Here the simple solution should be that we check again the "running" variable before check against worker_context. CC: Thomas Huth CC: Dr. David Alan Gilbert CC: Stefan Hajnoczi CC: Luk=C3=A1=C5=A1 Doktor CC: Markus Armbruster CC: Eric Blake CC: Paolo Bonzini Reported-by: Luk=C3=A1=C5=A1 Doktor Signed-off-by: Peter Xu Tested-by: Thomas Huth Message-id: 20190129051432.22023-1-peterx@redhat.com Signed-off-by: Stefan Hajnoczi --- iothread.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/iothread.c b/iothread.c index 2fb1cdf55d..e615b7ae52 100644 --- a/iothread.c +++ b/iothread.c @@ -63,7 +63,11 @@ static void *iothread_run(void *opaque) while (iothread->running) { aio_poll(iothread->ctx, true); =20 - if (atomic_read(&iothread->worker_context)) { + /* + * We must check the running state again in case it was + * changed in previous aio_poll() + */ + if (iothread->running && atomic_read(&iothread->worker_context))= { GMainLoop *loop; =20 g_main_context_push_thread_default(iothread->worker_context)= ; --=20 2.20.1