Created attachment 71420 [details] Log of pulseaudio crash I can reproduce this issue consistently. The setup is as follows: - This is a Car PC setup - Latest Arch Linux - Bluez 4.101-1 configured as an A2DP gateway - PulseAudio 2.99.3 When I connect my A2DP source device (smartphone) to the PC, pulseaudio crashes (or exits, I haven't been able to find a core or how to enable core dumping). I then ran pulseaudio with logging enabled: pulseaudio -vvvvvvvv --log-time=1 > pulse.log 2>&1 & I then proceeded to reproduce the problem. The log is attached. Like I said I can reproduce the problem, and I am able to build my own pulse audio and try it out. Let me know if there is any further debug info that you'd like. Thanks!
This has been reported also by others, but this hasn't been investigated yet. Since this appears to be a quite common problem, this should probably be regarded as a release blocker. The crash appears to be happen somewhere deep in libc when loading module-bluetooth-device. There shouldn't be anything special about loading module-bluetooth-device, but all the reports have been about that particular module anyway. Maybe module-bluetooth-discover overwrites some random memory, breaking the module loading procedure? Could you get a backtrace with gdb? Instructions: $ gdb pulseaudio [gdb starts and you get to the gdb prompt] (gdb) run -vvv [reproduce crash] (gdb) thread apply all bt Also, what is the exact version of your libc?
From the logs, I see: Inconsistency detected by ld.so: dl-open.c: 666: _dl_open: Assertion `_dl_debug_initialize (0, args.nsid)->r_state == RT_CONSISTENT' failed! Is this some sort of standard distribution? Could you also provide the output of ldd /usr/lib/pulse-2.99/modules/module-bluetooth-policy.so ?
JFYI, the reason I asked those questions is: https://jira.skype.com/browse/SCL-510 http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=543448
Created attachment 71423 [details] gdb output It appears that the process doesn't crash, but it cleanly exits... As you can see from the attachment, there was no crash and thus no core or backtrace... :(
I'm running on Arch Linux (www.archlinux.org). It's the latest version and I just set up this computer last night. Here's my version of glibc according to my distro's package manager: [alex@carpc ~]$ pacman -Q glibc glibc 2.16.0-5 And here's the output of ldd on module-bluetooth-policy.so: [alex@carpc ~]$ ldd /usr/lib/pulse-2.99/modules/module-bluetooth-policy.so linux-vdso.so.1 (0x00007fffd15ff000) libpulsecore-2.99.so => /usr/lib/libpulsecore-2.99.so (0x00007f14ba556000) libpulse.so.0 => /usr/lib/libpulse.so.0 (0x00007f14ba30c000) libpulsecommon-2.99.so => /usr/lib/pulseaudio/libpulsecommon-2.99.so (0x00007f14ba0a4000) libpthread.so.0 => /usr/lib/libpthread.so.0 (0x00007f14b9e87000) libc.so.6 => /usr/lib/libc.so.6 (0x00007f14b9ae0000) libltdl.so.7 => /usr/lib/libltdl.so.7 (0x00007f14b98d6000) libsamplerate.so.0 => /usr/lib/libsamplerate.so.0 (0x00007f14b9569000) libspeexdsp.so.1 => /usr/lib/libspeexdsp.so.1 (0x00007f14b9357000) libX11-xcb.so.1 => /usr/lib/libX11-xcb.so.1 (0x00007f14b9155000) libX11.so.6 => /usr/lib/libX11.so.6 (0x00007f14b8e1b000) libsndfile.so.1 => /usr/lib/libsndfile.so.1 (0x00007f14b8bb5000) liborc-0.4.so.0 => /usr/lib/liborc-0.4.so.0 (0x00007f14b893a000) libdbus-1.so.3 => /usr/lib/libdbus-1.so.3 (0x00007f14b86f3000) libtdb.so.1 => /usr/lib/libtdb.so.1 (0x00007f14b84e0000) libm.so.6 => /usr/lib/libm.so.6 (0x00007f14b81e6000) /usr/lib/ld-linux-x86-64.so.2 (0x00007f14baa17000) libjson.so.0 => /usr/lib/libjson.so.0 (0x00007f14b7fdd000) libxcb.so.1 => /usr/lib/libxcb.so.1 (0x00007f14b7dbe000) libasyncns.so.0 => /usr/lib/libasyncns.so.0 (0x00007f14b7bb8000) librt.so.1 => /usr/lib/librt.so.1 (0x00007f14b79af000) libdl.so.2 => /usr/lib/libdl.so.2 (0x00007f14b77ab000) libFLAC.so.8 => /usr/lib/libFLAC.so.8 (0x00007f14b755f000) libvorbisenc.so.2 => /usr/lib/libvorbisenc.so.2 (0x00007f14b7090000) libvorbis.so.0 => /usr/lib/libvorbis.so.0 (0x00007f14b6e63000) libogg.so.0 => /usr/lib/libogg.so.0 (0x00007f14b6c5c000) libXau.so.6 => /usr/lib/libXau.so.6 (0x00007f14b6a58000) libXdmcp.so.6 => /usr/lib/libXdmcp.so.6 (0x00007f14b6852000) libnsl.so.1 => /usr/lib/libnsl.so.1 (0x00007f14b6639000) libresolv.so.2 => /usr/lib/libresolv.so.2 (0x00007f14b6422000) [alex@carpc ~]$
Thanks for the output. Could you also provide the output of ldd /usr/lib32/libpulse* ?
Found the problem! First of all a big thank you for the hint of using ldd. I was totally looking the wrong way at this problem. Turns out I was missing the sbc package, and module-bluetooth-device.so was referencing that library. Thanks again! You may mark this issue as resolved/closed. FYI if this ever shows up again it's most likely a missing library. Here's the output of ldd on the "culprit": [alex@carpc ~]$ ldd /usr/lib/pulse-2.99/modules/module-bluetooth-device.so linux-vdso.so.1 (0x00007fffbdbff000) libsbc.so.1 => not found libbluetooth-util.so => /usr/lib/pulse-2.99/modules/libbluetooth-util.so (0x00007f72afa0f000) libpulsecore-2.99.so => /usr/lib/libpulsecore-2.99.so (0x00007f72af77b000) libpulse.so.0 => /usr/lib/libpulse.so.0 (0x00007f72af530000) libpulsecommon-2.99.so => /usr/lib/pulseaudio/libpulsecommon-2.99.so (0x00007f72af2c8000) libdbus-1.so.3 => /usr/lib/libdbus-1.so.3 (0x00007f72af082000) libpthread.so.0 => /usr/lib/libpthread.so.0 (0x00007f72aee65000) libc.so.6 => /usr/lib/libc.so.6 (0x00007f72aeabe000) libltdl.so.7 => /usr/lib/libltdl.so.7 (0x00007f72ae8b4000) libsamplerate.so.0 => /usr/lib/libsamplerate.so.0 (0x00007f72ae547000) libspeexdsp.so.1 => /usr/lib/libspeexdsp.so.1 (0x00007f72ae335000) libX11-xcb.so.1 => /usr/lib/libX11-xcb.so.1 (0x00007f72ae133000) libX11.so.6 => /usr/lib/libX11.so.6 (0x00007f72addf9000) libsndfile.so.1 => /usr/lib/libsndfile.so.1 (0x00007f72adb93000) liborc-0.4.so.0 => /usr/lib/liborc-0.4.so.0 (0x00007f72ad918000) libtdb.so.1 => /usr/lib/libtdb.so.1 (0x00007f72ad704000) libm.so.6 => /usr/lib/libm.so.6 (0x00007f72ad40a000) /usr/lib/ld-linux-x86-64.so.2 (0x00007f72afe55000) libjson.so.0 => /usr/lib/libjson.so.0 (0x00007f72ad202000) libxcb.so.1 => /usr/lib/libxcb.so.1 (0x00007f72acfe2000) libasyncns.so.0 => /usr/lib/libasyncns.so.0 (0x00007f72acddc000) librt.so.1 => /usr/lib/librt.so.1 (0x00007f72acbd4000) libdl.so.2 => /usr/lib/libdl.so.2 (0x00007f72ac9cf000) libFLAC.so.8 => /usr/lib/libFLAC.so.8 (0x00007f72ac784000) libvorbisenc.so.2 => /usr/lib/libvorbisenc.so.2 (0x00007f72ac2b5000) libvorbis.so.0 => /usr/lib/libvorbis.so.0 (0x00007f72ac087000) libogg.so.0 => /usr/lib/libogg.so.0 (0x00007f72abe81000) libXau.so.6 => /usr/lib/libXau.so.6 (0x00007f72abc7d000) libXdmcp.so.6 => /usr/lib/libXdmcp.so.6 (0x00007f72aba76000) libnsl.so.1 => /usr/lib/libnsl.so.1 (0x00007f72ab85e000) libresolv.so.2 => /usr/lib/libresolv.so.2 (0x00007f72ab647000)
So yeah, it all appears to work: after installing the sbc package I now have A2DP streaming in my car :D
Excellent, glad it's working now! :)
I've also pinged the Arch maintainer about this to double-check whether there's a packaging issue there.
Thanks! I was going to investigate a package dependency issue and report, but I wouldn't have gotten to it until after Christmas and the sooner they know about it the better :)
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.