[asterisk-dev] [Code Review] Add error checking to DEADLOCK_AVOIDANCE macro

Tilghman Lesher tlesher at digium.com
Fri Jul 2 16:20:49 CDT 2010



> On 2010-07-01 16:07:06, David Vossel wrote:
> > /branches/1.4/channels/chan_local.c, lines 186-197
> > <https://reviewboard.asterisk.org/r/751/diff/1/?file=11145#file11145line186>
> >
> >     This is dealing with some crazy scary stuff.  Are we seeing this actually happen?!
> >     
> >     Isn't other being returned locked here on error?
> 
> David Vossel wrote:
>     Ah, never mind, this is within the loop attempting to lock other.

We are not actually seeing this happen, but if it ever did, then we'd want to complain loudly, so that it could be caught early.


- Tilghman


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviewboard.asterisk.org/r/751/#review2322
-----------------------------------------------------------


On 2010-06-28 13:36:49, Tilghman Lesher wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviewboard.asterisk.org/r/751/
> -----------------------------------------------------------
> 
> (Updated 2010-06-28 13:36:49)
> 
> 
> Review request for Asterisk Developers.
> 
> 
> Summary
> -------
> 
> If DEADLOCK_AVOIDANCE is called in a situation where the lock is not held, a deadlock could result.  While this is a programming error, we should at the very least send up a warning about the problem.
> 
> 
> This addresses bug 17407.
>     https://issues.asterisk.org/view.php?id=17407
> 
> 
> Diffs
> -----
> 
>   /branches/1.4/channels/chan_agent.c 266141 
>   /branches/1.4/channels/chan_dahdi.c 266141 
>   /branches/1.4/channels/chan_h323.c 266141 
>   /branches/1.4/channels/chan_local.c 266141 
>   /branches/1.4/include/asterisk/lock.h 266141 
> 
> Diff: https://reviewboard.asterisk.org/r/751/diff
> 
> 
> Testing
> -------
> 
> No functional changes.  Runs better according to reporter.
> 
> 
> Thanks,
> 
> Tilghman
> 
>




More information about the asterisk-dev mailing list