[pulseaudio-commits] [Git][pulseaudio/pulseaudio][master] bluetooth: Fix crash in setup_stream()

Tanu Kaskinen gitlab at gitlab.freedesktop.org
Tue Jul 2 05:40:04 UTC 2019



Tanu Kaskinen pushed to branch master at PulseAudio / pulseaudio


Commits:
661b13d5 by Frédéric Danis at 2019-07-02T05:22:19Z
bluetooth: Fix crash in setup_stream()

setup_stream() crashes when calling set_nonblock() with an invalid
stream_fd.

On a new call, the ofono backend gets notified of a new connection.
The ofono backend sets the transport state to playing, and that triggers
a profile change, which sets up the stream for the first time.
Then module-bluetooth-policy sets up the loopbacks. The loopbacks get
fully initialized before the crash.

After module-bluetooth-policy has done its things, the execution
continues in the transport state change hook. The next hook user is
module-bluez5-device, whose handle_transport_state_change() function
gets called. It will then set up the stream again even though it's
already set up. I'm not sure if that's a some kind of a bug.
setup_stream() can handle the case where it's unnecessarily called,
though, so this second setup is not a big problem.

The crash happens, because the connection died due to POLLHUP in the IO
thread before the second setup_stream() call.

- - - - -


1 changed file:

- src/modules/bluetooth/module-bluez5-device.c


Changes:

=====================================
src/modules/bluetooth/module-bluez5-device.c
=====================================
@@ -753,6 +753,8 @@ static void setup_stream(struct userdata *u) {
     struct pollfd *pollfd;
     int one;
 
+    pa_assert(u->stream_fd >= 0);
+
     /* return if stream is already set up */
     if (u->stream_setup_done)
         return;
@@ -829,7 +831,17 @@ static int source_process_msg(pa_msgobject *o, int code, void *data, int64_t off
         }
 
         case PA_SOURCE_MESSAGE_SETUP_STREAM:
-            setup_stream(u);
+            /* Skip stream setup if stream_fd has been invalidated.
+               This can occur if the stream has already been set up and
+               then immediately received POLLHUP. If the stream has
+               already been set up earlier, then this setup_stream()
+               call is redundant anyway, but currently the code
+               is such that this kind of unnecessary setup_stream()
+               calls can happen. */
+            if (u->stream_fd < 0)
+                pa_log_debug("Skip source stream setup while closing");
+            else
+                setup_stream(u);
             return 0;
 
     }
@@ -1007,7 +1019,17 @@ static int sink_process_msg(pa_msgobject *o, int code, void *data, int64_t offse
         }
 
         case PA_SINK_MESSAGE_SETUP_STREAM:
-            setup_stream(u);
+            /* Skip stream setup if stream_fd has been invalidated.
+               This can occur if the stream has already been set up and
+               then immediately received POLLHUP. If the stream has
+               already been set up earlier, then this setup_stream()
+               call is redundant anyway, but currently the code
+               is such that this kind of unnecessary setup_stream()
+               calls can happen. */
+            if (u->stream_fd < 0)
+                pa_log_debug("Skip sink stream setup while closing");
+            else
+                setup_stream(u);
             return 0;
     }
 



View it on GitLab: https://gitlab.freedesktop.org/pulseaudio/pulseaudio/commit/661b13d50d5af5c14f80e8678aaf942fbc5628ae

-- 
View it on GitLab: https://gitlab.freedesktop.org/pulseaudio/pulseaudio/commit/661b13d50d5af5c14f80e8678aaf942fbc5628ae
You're receiving this email because of your account on gitlab.freedesktop.org.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/pulseaudio-commits/attachments/20190702/ce427857/attachment-0001.html>


More information about the pulseaudio-commits mailing list