[mythtv] Need help catching a segfault in mythbackend?

[Mitko Haralanov]

Hi, due to some issues that I had with the stable release of myth
(0.20.1), I am currently running the rpms from ATrpm's bleeding repo.
Everything is working great, except that mythbackend is segfaulting
from time to time.

The segfaults happen under conditions that make it hard for me to catch
so I am asking for some help on the subject. Here is the situation:

I have my box setup to auto-shutdown and wake-up using nvram-wakeup.
Usually, the box wakes up sometime in the afternoon and that is when
the segfault happens. After it wakes up, the backend is running for a
random amount of time (I say random because sometimes the backend works
for an hour, sometimes four) and then it crashes with messages similar
to this:

May 19 19:19:19 mythtv kernel: mythfrontend[4318]: segfault at
0000000000676e6d rip 0000003b4f66eb92 rsp 000000004467a120 error 4

The interesting part is that when I come home and find the backend
gone, I restart it and then I continue to use the machine the entire
night. I watch recordings (both HD and SD), I watch live tv (both HD
and SD), and the backend records without a crash.

Due to the fact that I am not at home when the machine wakes up and
seem to crash, I don't have the chance to catch the crash. I have tried
to set the core ulimit to "unlimited" in /etc/init.d/mythbackend, so it
generates a core but it's not happening.

I am looking for idea of how I can get a backtrace or a core (or both)
that I can submit to the developers so, hopefully, the bug is fixed.

Thanks in advance!

-- 
Mitko Haralanov
voidtrance at comcast dot net
http://voidtrance.home.comcast.net
==========================================