[asterisk-dev] [svn-commits] tilghman: branch 1.4 r120001 - /branches/1.4/channels/chan_iax2.c

[Russell Bryant]

On Jun 3, 2008, at 4:42 PM, Tilghman Lesher wrote:
>> What deadlock avoidance are you referring to here?  It looks to me  
>> like
>> the real problem is that the peer->callno could change at any time,
>> because the peer is not locked at this point.  However, with the  
>> change
>> you have added, it's not needed anymore.  I just wanted to point out
>> where the problem came from.
>
> The deadlock avoidance occurs within iax2_destroy(callno)

The deadlock avoidance in iax2_destroy() is between the ast_channel  
lock and the iax2 pvt struct lock.  The peer is not involved there.

--
Russell Bryant
Senior Software Engineer
Open Source Team Lead
Digium, Inc.