[asterisk-dev] [Code Review] Regression: chan_iax2 registration issues.

[Simon Perreault]

> On 2010-11-01 10:14:48, Simon Perreault wrote:
> > branches/1.8/channels/chan_iax2.c, line 11195
> > <https://reviewboard.asterisk.org/r/993/diff/1/?file=12796#file12796line11195>
> >
> >     Why is this needed?
> 
> pabelanger wrote:
>     I found if you register and unregister an IAX2 client, you would get the following error:
>     [Oct 24 19:07:32] ERROR[1403] netsock2.c: getnameinfo(): ai_family not supported
>     
>     Adding the sin_family resolved it, if the location of the code is incorrect, where would you suggest?

Makes sense. If you were running in debug mode you would probably have seen this from _ast_sockaddr_from_sin():

                ast_log(__LOG_DEBUG, file, line, func, "Address family is not AF_INET\n");


- Simon


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviewboard.asterisk.org/r/993/#review2867
-----------------------------------------------------------


On 2010-11-01 08:00:20, pabelanger wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviewboard.asterisk.org/r/993/
> -----------------------------------------------------------
> 
> (Updated 2010-11-01 08:00:20)
> 
> 
> Review request for Asterisk Developers.
> 
> 
> Summary
> -------
> 
> A few people on -users[1] reported an issue with IAX2 registrations.  Looking at the code, I managed to find an issue when we convert between sockaddr_in and ast_sockaddr (Bad address cast to IPv4, etc).
> 
> [1] http://lists.digium.com/pipermail/asterisk-users/2010-October/255251.html
> 
> 
> This addresses bug 18202.
>     https://issues.asterisk.org/view.php?id=18202
> 
> 
> Diffs
> -----
> 
>   branches/1.8/channels/chan_iax2.c 293157 
> 
> Diff: https://reviewboard.asterisk.org/r/993/diff
> 
> 
> Testing
> -------
> 
> Was able to reproduce and fix on local dev box, plus reporter reported patch fixed the issue.
> 
> 
> Thanks,
> 
> pabelanger
> 
>