#1104168 callaudiod: segfault on trying to call someone via gnome-calls

Package:
callaudiod
Source:
callaudiod
Description:
Call audio routing daemon
Submitter:
Benedikt Wildenhain
Date:
2025-05-26 09:25:01 UTC
Severity:
normal
#1104168#5
Date:
2025-04-26 15:10:43 UTC
From:
To:
Dear Maintainer,

*** Reporter, please consider answering these questions, where appropriate ***

   * What led up to the situation?
I tried to call a phone via a local asterisk server (using SRTP and
SIPS). I was using gnome-calls,
commit 523ec151d13df47757f30426a3eed67aba306505 from Tue Feb 11 22:13:45
2025 -0500 (repo https://gitlab.gnome.org/GNOME/calls).


   * What was the outcome of this action?

Starting gnome-calls I get "sres: /etc/resolv.conf: unknown option" (by
try-and-error this can be fixed by replacing "options edns0 trust-ad"
via "options edns0").

Even after that I get an info box at gnome-calls telling me "$user@$servername: DNS error", when trying to call some phone.

Furthermore I get the following from the kernel:

callaudiod[43102]: segfault at 0 ip 00005624740ecaac sp 00007fffef5aedc0 error 4 in callaudiod[9aac,5624740e8000+6000] likely on CPU 4 (core 0, socket 0)
Code: 00 75 6b 8b 45 60 c7 45 64 00 00 00 00 83 f8 01 0f 84 10 01 00 00 31 d2 3d ff 00 00 00 75 3b 48 8b 83 88 01 00 00 48 8b 75 48 <48> 8b 38 e8 2c c0 ff ff 48 8b 7d 18 85 c0 0f 84 31 01 00 00 c7 45
A

   * What outcome did you expect instead?

Some more useful error message or a successfull call.

*** End of the template - remove these template lines ***

#1104168#10
Date:
2025-05-26 09:17:16 UTC
From:
To:
This line points to a crash in following line [1], with maybe variable info or self being a NULL pointer:

   407                 if (g_strcmp0(info->active_port->name, self->earpiece_port) == 0) {

https://sources.debian.org/src/callaudiod/0.1.10-1/src/cad-pulse.c/#L407


This upstream bug mentions the same source code line:

https://gitlab.com/mobian1/callaudiod/-/issues/36



Adding a complete backtrace would probably be helpful.
This can usually be found in the journal after the
package systemd-coredump is installed, if possible.

https://wiki.debian.org/HowToGetABacktrace#Core_dump


Kind regards,
Bernhard


[1] https://wiki.debian.org/InterpretingKernelOutputAtProcessCrash