[mythtv] Hunting down deadlock (ticket #9261)

Juhani Rautiainen jrauti at iki.fi
Sun Dec 5 19:45:31 UTC 2010


I've been suffering intermitting backend deadlocks since last summer. I 
finally took gdb backtrace of the problem and it seems that it same as 
ticket #9261. Ticket is low priority so I decided to try to find out 
what's going on myself. My C++ is little rusty but I probably figured out 
sequence of events but I can't see what other thread could lock 
EITScanner::lock. Waiting for that lock causes two other threads to lock. 
So what I'm asking what can I do with gdb when it this happens next time? 
Is there way to see from gdb what holds that lock? Any pointers are 
greatly appreciated. Forgot to mention I'm using 0.24-fixes branch.

Juhani Rautiainen				jrauti at iki.fi

More information about the mythtv-dev mailing list