vdr-sxfe crashed with SIGSEGV in dbus_g_proxy_call()

[wilderigel]

Binary package hint: vdr-plugin-xineliboutput

stefan@homer:~$ vdr-sxfe
vdr-sxfe 1.0.0rc2 (build with xine-lib 1.1.10, using xine-lib 1.1.10)

VDR server not given, searching ...
Found VDR server: host 192.168.1.2, port 37890
Segmentation fault (core dumped)

ProblemType: Crash
Architecture: amd64
Date: Sat Mar 1 21:24:34 2008
DistroRelease: Ubuntu 8.04
ExecutablePath: /usr/bin/vdr-sxfe
NonfreeKernelModules: nvidia
Package: xineliboutput-sxfe 1.0.0~rc2-13
PackageArchitecture: amd64
ProcCmdline: vdr-sxfe
ProcEnviron:
 SHELL=/bin/bash
 PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games
 LANG=de_DE.UTF-8
Signal: 11
SourcePackage: vdr-plugin-xineliboutput
StacktraceTop:
 ?? () from /usr/lib/libdbus-glib-1.so.2
 ?? () from /usr/lib/libdbus-glib-1.so.2
 dbus_g_proxy_call ()
 ?? ()
 ?? ()
Title: vdr-sxfe crashed with SIGSEGV in dbus_g_proxy_call()
Uname: Linux 2.6.24-10-generic x86_64
UserGroups: adm admin audio cdrom dialout dip floppy lpadmin plugdev video

[Apport retracing service (apport)]

StacktraceTop:_dbus_gvalue_store (value=0x7fff256261d0, storage=0x7f0d00000000)
dbus_g_proxy_end_call_internal (proxy=0x639000, call_id=<value optimized out>,
dbus_g_proxy_call (proxy=0x639000, method=0x4135c2 "Inhibit",
gnome_screensaver_control (enable=0) at gnome_screensaver.c:87
sxfe_display_open (this_gen=0x6165e0, width=<value optimized out>, height=576,

[wilderigel (wilderigel)]

vdr-sxfe crashs only as user, as root it works fine.

[Martti Kuparinen (mk1970)]

For me (Ubuntu 8.04) running this with sudo does not help...

vdr-sxfe: pulsecore/mutex-posix.c:98: pa_mutex_unlock: Assertion `pthread_mutex_unlock(&m->mutex) == 0' failed.
Aborted (core dumped)

[JStk (janvanstiphout)]

Got the same problem here, segmentation fault when running as user, when running as root, it runs fine.

[Robert Schneider (robert-schneider)]

I am seeing a segmentation fault whenever I press a key on the keyboard, but that occurs no matter whether I am a user or root. Anybody else seeing this? Is this another bug that should be filed?

[JStk (janvanstiphout)]

Some environment variable seems to have something to do with it.
When using sudo su <user> -c "vdr-sxfe xvdr://localhost" is does work.

I then tried to unset all variables with
`export | grep -v "declare -x PATH" | sed 's/declare -x/unset/g' | awk -F\= '{ print $1 }'`
in the BASH shell, and then vdr-sxfe works! The sound does disappear some times though, but I don't really think that that's related to it.

BTW. It occurs for me as soon as I start vdr-sxfe, not when I press a key (however I don't press any keys, because I use LIRC)

[Robert Schneider (robert-schneider)]

I have done some further tests yesterday. As of my 04/28 posting, I only have segmentation faults whenever I press a key. I can have the frontend running fine as long as I don't touch the keyboard.

So again my question, am I seeing a different bug than what this one relates to?

However, I also noticed that the sound goes away every now and then. It is neither reproducable nor does it appear at a certain time. Switching channels within VDR brings the sound back. Though only for some time.

[Heiko Albrecht (heiko-albrecht)]

same problem here after update from GG to HH:

if I run

$ vdr-sxfe

on same host with vdr, everything is fine. But if I

$ vdr-sxfe --tcp

then, vdr-sxfe crashes with segmentation fault after pressing any key.

heiko@heiko:~$ vdr-sxfe --tcp
vdr-sxfe 1.0.0rc2 (build with xine-lib 1.1.11, using xine-lib 1.1.11)

Protocol: TCP

VDR server not given, searching ...
Found VDR server: host 192.168.1.11, port 37890
[26614] [vdr-fe] Error: The name org.gnome.ScreenSaver was not provided by any .service files
[26614] [vdr-fe] (ERROR (gnome_screensaver.c,113): Resource temporarily unavailable)

Press Esc to exit

[26614] [input_vdr] Connecting (control) to tcp://192.168.1.11:37890 ...
[26614] [input_vdr] Server greeting: VDR-1.6.0 xineliboutput-1.0.0rc2 READY
[26614] [input_vdr] Connected (control) to tcp://192.168.1.11:37890
[26614] [input_vdr] Connecting (data) to tcp://192.168.1.11:37890 ...
[26614] [input_vdr] Data stream connected (TCP)
[26614] [input_vdr] using custom "no signal" image /usr/share/libxine1-xvdr/nosignal.mpg
[26632] [input_vdr] No fe_control function! POST AudioVisualization Off failed.
[26632] [input_vdr] No fe_control function! POST tvtime On method=Linear,cheap_mode=1,pulldown=none,framerate_mode=full,judder_correction=1,use_progressive_frame_flag=1,chroma_filter=0,enable=1 failed.
[26632] [input_vdr] No fe_control function! OVERSCAN 0 failed.
[26632] [input_vdr] No fe_control function! POST upmix Off failed.
[26632] [input_vdr] No fe_control function! POST autocrop Off failed.
[26632] [input_vdr] No fe_control function! POST pp Off failed.
[26632] [input_vdr] No fe_control function! POST unsharp Off failed.
[26632] [input_vdr] No fe_control function! POST denoise3d Off failed.
Segmentation fault
heiko@heiko:~$

[JStk (janvanstiphout)]

Found out what the variable was which I have to unset.
It's DBUS_SESSION_BUS_ADDRESS

Looking at the initial bug report, this seems like the same problem as the one which wilderigel describes (there also dbus is mentioned).
A same similarity is the processor architecture, x86_64. It could be a 64-bit arch. specific problem.

[Robert Schneider (robert-schneider)]

I have tested both, neither removing the environment variable nor running vdr-sxfe as root cures my problem. Whenever I press a key, the thing dies.

[Tony Houghton (h-realh)]

It crashes for me on HH amd64 as soon as I try to start it. I think it's something to do with its interaction with gnome-screensaver, because if I don't have that running vdr-sxfe will run OK (except it's pretty flaky).

[zn!K (markus-endl)]

yesterday i made a upgrade from 7.04 to 8.04 - every went fine. i encountered the same problems described above.
vdr-sxfe ended with a seg fault. unset env var or using root user to start vdr-sxfe are working.

[Robert Schneider (robert-schneider)]

Still, neither running it as root nor unsetting any env var cures my problem. I have opened another bug #232241 as this one here could be a different problem. I have also attached more information to the other bug.

If someone wants to look into that and needs more information ... I do anything to help to fix this bug. It keeps me away from upgrading six machines.

[Hanno (hzulla)]

This bug appears to be related to http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=487410

[Hanno (hzulla)]

And related to this one http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=438116 - the Debian team fixed it there.

[Hanno (hzulla)]

Uhm, darn, ignore my last comment. 438116 is another bug. Sorry for the confusion.

[Brian Murray (brian-murray)]

It looks like the debian bug report has been fixed and that the version of this in Intrepid should include the fix.

[Timo Aaltonen (tjaalton)]

Fixed in intrepid, but will also fix in hardy.

[Timo Aaltonen (tjaalton)]

meh, no reason to pretend that this will be fixed in hardy anymore.