Re: pthread coldfire thread lockup under stress
Ed Slas dixit:
>coldfire runs for hours, then a thread will get ‘stuck’ in the pthread library
May this be related to these problems? Threads at:
http://lists.debian.org/debian-68k/2012/07/msg00000.html
and
http://lists.debian.org/debian-68k/2012/08/msg00019.html
In those cases, I think there are two threads, and one is waiting
for the other, hence 0% CPU in the end. I suspected that the code
changing the wait condition never gets run. (I even considered
adding some yield-like call to the atomic routines, but have not
done more yet.)
>I can inspect the contents
>of the mutex that was being locked with gdb, and its owner is 0! So it looks
>like its waiting for a mutex that is already unlocked.
How does one do that? I’ve got almost no idea about parallel
programming.
bye,
//mirabilos
--
“Having a smoking section in a restaurant is like having
a peeing section in a swimming pool.”
-- Edward Burr
Reply to: