From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:40781) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1gTnA8-0005ab-Lm for qemu-devel@nongnu.org; Mon, 03 Dec 2018 07:17:21 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1gTnA4-0000nA-Qv for qemu-devel@nongnu.org; Mon, 03 Dec 2018 07:17:20 -0500 Received: from mx1.redhat.com ([209.132.183.28]:52366) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1gTnA2-0000lf-Rb for qemu-devel@nongnu.org; Mon, 03 Dec 2018 07:17:16 -0500 Received: from smtp.corp.redhat.com (int-mx04.intmail.prod.int.phx2.redhat.com [10.5.11.14]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 2C87483F40 for ; Mon, 3 Dec 2018 12:17:14 +0000 (UTC) From: Markus Armbruster References: <20181029125733.14597-1-marcandre.lureau@redhat.com> <20181029125733.14597-7-marcandre.lureau@redhat.com> <87h8fv6jwy.fsf@dusky.pond.sub.org> Date: Mon, 03 Dec 2018 13:17:12 +0100 In-Reply-To: (=?utf-8?Q?=22Marc-Andr=C3=A9?= Lureau"'s message of "Mon, 3 Dec 2018 14:02:59 +0400") Message-ID: <87k1kq4xfb.fsf@dusky.pond.sub.org> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH v2 6/6] monitor: avoid potential dead-lock when cleaning up List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: =?utf-8?Q?Marc-Andr=C3=A9?= Lureau Cc: "Bonzini, Paolo" , qemu-devel , Peter Xu , "Dr. David Alan Gilbert" Marc-Andr=C3=A9 Lureau writes: > On Mon, Dec 3, 2018 at 1:26 PM Markus Armbruster wrot= e: >> >> Marc-Andr=C3=A9 Lureau writes: >> >> > When a monitor is connected to a Spice chardev, the monitor cleanup >> > can dead-lock: >> > >> > #0 0x00007f43446637fd in __lll_lock_wait () at /lib64/libpthread.so.0 >> > #1 0x00007f434465ccf4 in pthread_mutex_lock () at /lib64/libpthread.= so.0 >> > #2 0x0000556dd79f22ba in qemu_mutex_lock_impl (mutex=3D0x556dd81c922= 0 , file=3D0x556dd7ae3648 "/home/elmarco/src/qq/monitor.c", l= ine=3D645) at /home/elmarco/src/qq/util/qemu-thread-posix.c:66 >> > #3 0x0000556dd7431bd5 in monitor_qapi_event_queue (event=3DQAPI_EVEN= T_SPICE_DISCONNECTED, qdict=3D0x556dd9abc850, errp=3D0x7fffb7bbddd8) at /ho= me/elmarco/src/qq/monitor.c:645 >> > #4 0x0000556dd79d476b in qapi_event_send_spice_disconnected (server= =3D0x556dd98ee760, client=3D0x556ddaaa8560, errp=3D0x556dd82180d0 ) at qapi/qapi-events-ui.c:149 >> > #5 0x0000556dd7870fc1 in channel_event (event=3D3, info=3D0x556ddad1= b590) at /home/elmarco/src/qq/ui/spice-core.c:235 >> > #6 0x00007f434560a6bb in reds_handle_channel_event (reds=3D, event=3D3, info=3D0x556ddad1b590) at reds.c:316 >> > #7 0x00007f43455f393b in main_dispatcher_self_handle_channel_event (= info=3D0x556ddad1b590, event=3D3, self=3D0x556dd9a7d8c0) at main-dispatcher= .c:197 >> > #8 0x00007f43455f393b in main_dispatcher_channel_event (self=3D0x556= dd9a7d8c0, event=3Devent@entry=3D3, info=3D0x556ddad1b590) at main-dispatch= er.c:197 >> > #9 0x00007f4345612833 in red_stream_push_channel_event (s=3Ds@entry= =3D0x556ddae2ef40, event=3Devent@entry=3D3) at red-stream.c:414 >> > #10 0x00007f434561286b in red_stream_free (s=3D0x556ddae2ef40) at red= -stream.c:388 >> > #11 0x00007f43455f9ddc in red_channel_client_finalize (object=3D0x556= dd9bb21a0) at red-channel-client.c:347 >> > #12 0x00007f434b5f9fb9 in g_object_unref () at /lib64/libgobject-2.0.= so.0 >> > #13 0x00007f43455fc212 in red_channel_client_push (rcc=3D0x556dd9bb21= a0) at red-channel-client.c:1341 >> > #14 0x0000556dd76081ba in spice_port_set_fe_open (chr=3D0x556dd9925e2= 0, fe_open=3D0) at /home/elmarco/src/qq/chardev/spice.c:241 >> > #15 0x0000556dd796d74a in qemu_chr_fe_set_open (be=3D0x556dd9a37c00, = fe_open=3D0) at /home/elmarco/src/qq/chardev/char-fe.c:340 >> > #16 0x0000556dd796d4d9 in qemu_chr_fe_set_handlers (b=3D0x556dd9a37c0= 0, fd_can_read=3D0x0, fd_read=3D0x0, fd_event=3D0x0, be_change=3D0x0, opaqu= e=3D0x0, context=3D0x0, set_open=3Dtrue) at /home/elmarco/src/qq/chardev/ch= ar-fe.c:280 >> > #17 0x0000556dd796d359 in qemu_chr_fe_deinit (b=3D0x556dd9a37c00, del= =3Dfalse) at /home/elmarco/src/qq/chardev/char-fe.c:233 >> > #18 0x0000556dd7432240 in monitor_data_destroy (mon=3D0x556dd9a37c00)= at /home/elmarco/src/qq/monitor.c:786 >> > #19 0x0000556dd743b968 in monitor_cleanup () at /home/elmarco/src/qq/= monitor.c:4683 >> > #20 0x0000556dd75ce776 in main (argc=3D3, argv=3D0x7fffb7bbe458, envp= =3D0x7fffb7bbe478) at /home/elmarco/src/qq/vl.c:4660 >> > >> > Because spice code tries to emit a "disconnected" signal on the >> > monitors. Fix this dead-lock by releasing the monitor lock for >> > flush/destroy. >> > >> > Signed-off-by: Marc-Andr=C3=A9 Lureau >> > --- >> > monitor.c | 2 ++ >> > 1 file changed, 2 insertions(+) >> > >> > diff --git a/monitor.c b/monitor.c >> > index 7fe89daa87..b55e604a98 100644 >> > --- a/monitor.c >> > +++ b/monitor.c >> > @@ -4643,8 +4643,10 @@ void monitor_cleanup(void) >> > monitor_destroyed =3D true; >> > QTAILQ_FOREACH_SAFE(mon, &mon_list, entry, next) { >> > QTAILQ_REMOVE(&mon_list, mon, entry); >> > + qemu_mutex_unlock(&monitor_lock); >> > monitor_flush(mon); >> > monitor_data_destroy(mon); >> > + qemu_mutex_lock(&monitor_lock); >> > g_free(mon); >> > } >> > qemu_mutex_unlock(&monitor_lock); >> >> I think a comment hinting at the reason for relinquishing the lock would >> be in order. Perhaps >> >> /* Permit QAPI event emission from character frontend release= */ >> qemu_mutex_unlock(&monitor_lock); >> >> We need to demonstrate calling monitor_flush() and >> monitor_data_destroy() without holding @monitor_lock is safe. >> >> @monitor_lock's comment states it "protects mon_list, >> monitor_qapi_event_state." Looks plausible from how it's used. >> >> As far as I can tell, monitor_flush() and monitor_data_destroy() don't >> access mon_list and monitor_qapi_event_state. >> >> monitor_cleanup()'s loop itself is safe because it uses >> QTAILQ_FOREACH_SAFE(), unlike similar loops elsewhere. >> >> I think we're good. I'd like you to work this argument into the commit >> message. > > What about adding: > > monitor_lock protects mon_list, monitor_qapi_event_state and > monitor_destroyed. monitor_flush() and monitor_data_destroy() don't > access any of those variables. > > monitor_cleanup()'s loop is safe because it uses > QTAILQ_FOREACH_SAFE(), and no further monitor can be added after > calling monitor_cleanup() thanks to monitor_destroyed check in > monitor_list_append(). Works for me. Thanks!