[asterisk-dev] [Code Review] chan_sip crash in interpret_t38_parameters()

Kevin Fleming kpfleming at digium.com
Thu May 20 16:04:33 CDT 2010


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviewboard.asterisk.org/r/671/#review2048
-----------------------------------------------------------


I think it would be much more logical to just clear the SIP_PAGE2_T38SUPPORT flag on the sip_pvt if creation of the udptl structure fails. There's no value in leaving that flag turned on if we can't actually do T.38 for that dialog.

- Kevin


On 2010-05-20 15:22:53, David Vossel wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviewboard.asterisk.org/r/671/
> -----------------------------------------------------------
> 
> (Updated 2010-05-20 15:22:53)
> 
> 
> Review request for Asterisk Developers.
> 
> 
> Summary
> -------
> 
> This is a really simple patch, I just wanted to verify this with someone who has more knowledge of our SIP T38 implementation than me.
> 
> There is a crash occurring in interpret_t38_parameters() because we are attempting to access a NULL p->udptl struct.  It turns out it is possible for p->udptl to be NULL when the SIP_PAGE2_T38SUPPORT flag is set.  This is because no action is taken if p->udptl creation fails during the creation of a new dialog.  Is ignoring this failure done on purpose because there is no appropriate action to take? If so this patch makes sense to me because we can't couple T38 support with the creation of a p->udptl struct, else I have no idea what kind of action should be taken here.
> 
> 
> This addresses bug 17264.
>     https://issues.asterisk.org/view.php?id=17264
> 
> 
> Diffs
> -----
> 
>   /trunk/channels/chan_sip.c 264538 
> 
> Diff: https://reviewboard.asterisk.org/r/671/diff
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> David
> 
>




More information about the asterisk-dev mailing list