Re: [xine-devel] deadlocks

[Barry S. <bar...@on...>]

[Opss forgot to CC the list the first time I replied]

Maximilian Schwerin wrote:
> Hi,
>
> some users of my xine-lib frontend oxine (oxine.sf.net) are experiencing
> deadlocks that I am unable to reproduce on the machines I have available for
> testing purposes.
>
> My best guess at the moment is that either this is an issue with multicore CPUs
> (I've only been able to test this in a virtual machine on a quad core) or that
> the deadlock is against xine-lib.
>
> As xine-lib has it's own mutexes and I guess you guys have some experience with
> avoiding deadlocks I wanted to ask if there are things I should not do or if you
> have any other idea or hint that might help me.
>   

Find out where you are dead locking. Use gdb to show what all the 
threads are doing is
a start.

gdb> thread apply all bt

 From that you can make progress in isolating the deadlock.


The version of xine lib we use is a year old and has no known locking 
bugs. We contributed
back all the locking fixes at that time for the bugs we found. We run on 
Core 2 Duo without
problems.

Barry