Bug #1223
segfault
100%
Description
CRASH: Signal: 11 in PRG: /usr/bin/tvheadend (3.1.670.g9ed26.dirty) [9816688a8b2d8d8a7e0e3f38ff0fc25507c82008] CWD: /home/hts CRASH: Fault address 0x41 (Address not mapped) CRASH: Loaded libraries: /usr/lib/x86_64-linux-gnu/libssl.so.1.0.0 /usr/lib/x86_64-linux-gnu/libcrypto.so.1.0.0 /lib/x86_64-linux-gnu/libz.so.1 /usr/lib/x86_64-linux-gnu/libavahi-common.so.3 /usr/lib/x86_64-linux-gnu/libavahi-client.so.3 /lib/x86_64-linux-gnu/librt.so.1 /lib/x86_64-linux-gnu/libdl.so.2 /lib/x86_64-linux-gnu/libpthread.so.0 /lib/x86_64-linux-gnu/libc.so.6 /lib/x86_64-linux-gnu/libdbus-1.so.3 /lib64/ld-linux-x86-64.so.2 CRASH: Register dump [23]: 0000000000000000 0000000000000000 00007ffff68c64a0 a3d70a3d70a3d70b 00007fffe40fe550 000000000133ee48 0000000000939bf0 000000000133ee48 0000000000469111 0000000000000041 00007fffe40fec90 0000000000469111 0000000000000003 0000000000000011 0000000000000001 00007fffefffd668 00007ffff68c5d7a 0000000000010287 0000000000000033 0000000000000004 000000000000000e fffffffe7ffbfa07 0000000000000041 CRASH: STACKTRACE CRASH: /usr/src/tvheadend/src/trap.c:139 0x4249a6 CRASH: ??:0 0x7ffff6b4c030 CRASH: ??:0 0x7ffff68c5d7a CRASH: /usr/src/tvheadend/src/lang_str.c:33 0x4291b0 CRASH: /usr/src/tvheadend/src/epg.c:297 0x40af37 CRASH: /usr/src/tvheadend/src/epg.c:309 0x40afe6 CRASH: /usr/src/tvheadend/src/epggrab/module/eit.c:554 0x42e6ff CRASH: /usr/src/tvheadend/src/epggrab/module/eit.c:665 0x42f12c CRASH: /usr/src/tvheadend/src/dvb/dvb_tables.c:173 0x44459a CRASH: ??:0 0x7ffff6b43b50 CRASH: clone+0x6d (/lib/x86_64-linux-gnu/libc.so.6) Segmentation fault
Attached is gdb output.
Compiled from git #9ed26c1a54 -- https://github.com/tvheadend/tvheadend/tree/9ed26c1a54a173104e2b2c714b75dc7c21036183
Debian Wheezy
Files
Subtasks
History
Updated by
Updated by Steve Allison about 12 years ago
Updated by Just EIT.
.hts/tvheadend/epggrab/config
{
"mod_enabled": {
"uk_freeview": 1,
"eit": 1
}
}
Updated by Adam Sutton about 12 years ago
And second question was this an update from an existing install? I.e. would if have had an existing EPG database?
Some question hanging over whether there is a cockup in recent fixes that mean it crashes with older DBs. Though I'm reviewing the code to see what I might have done.
Updated by Steve Allison about 12 years ago
Fresh install. I removed /home/hts/.hts prior to dpkg -i
Updated by Adam Sutton about 12 years ago
- Status changed from New to Need feedback
Steve,
If you fancy helping to track this down do you think you could pull the latest master and rebuild with some of my additional debug enabled:
make distclean
./configure
CFLAGS=-DEPG_TRACE make
And then run TVH to capture all output to file and ideally get a gdb stack trace to go with it.
Sorry to ask for so much, but the info you've given so far has been pretty detailed but unfortunately I'm struggling to find the root of the problem.
Adam
P.S.
The extra debug will add a LOT of output so might be best to zip it and mail to me directly.
Updated by Steve Allison about 12 years ago
Don't apologies, its the least I can do in return for your hard work.
It hasn't crashed again since, so I'll stop, rebuild with that configure and install.
Is it okay for me to keep the config?
Is logging to syslog enough?
Updated by Adam Sutton about 12 years ago
I'd redirect to file if I were you. It's a LOT of a logging and some syslog implementations get overloaded and stop writing to file from processes that spam the system logs.
Adn don't enable debug logging in the web UI, you'll most likely crash the browser 
Adam
Updated by Adam Sutton about 12 years ago
- Status changed from Need feedback to Fixed
Applied in changeset commit:c3fe6fae734ecef6f7cdecd135f888a803e81105.
Updated by Steve Allison about 12 years ago
I can kill this now? I think I will run out of space over night, haha.
-rw-r--r-- 1 hts hts 789M Sep 16 23:42 tvheadend-gdb.log
Updated by Adam Sutton about 12 years ago
Yeah that's nice and big still not quite on a par with Andy B's 2G log he mailed me
But yeah I think this should now be fixed, certainly it was a bad bug and would have led to my own crash and could easily explain most of the others.
Update to latest code and run that (without EPG_TRACE) and see what happens.
Adam