[Serdev] Does Nathelper crash SER?

[Jan Janak]

Thanks for the report. Could you please generate a coredump and send it
to serhelp at lists.iptel.org along with compiled sources which created the
coredump (including debugging symbols) ?

  Jan.

On 23-12 10:54, Johan B wrote:
> Hi,
> 
> I have stumbled on some kind of problem which I believe is related to the 
> Nathelper module. SER crashes when forwarding a request that has had the 
> contact header rewritten twice. I can reproduce the error on both the 
> 0.8.12 release and the latest cvs tarball.
> 
> I use Nathelper in combination with the Mangler module, which also rewrites 
> the contact header. A simple test case is to have a route block like this:
> 
> fix_nated_contact();
> encode_contact(?prefix?, ?<ip addr>?);
> 
> rewritehostport(?<ip addr>?);
> t_relay();
> 
> 
> SER crashes when rendering the final request header, going through the lump 
> lists.
> 
> The reason I suspect Nathelper to do something wrong is that you may have 
> two encode_contact() calls in your route block (not using Nathelper at 
> all), but you will crash when having two fix_nated_contact() calls (and not 
> using Mangler). Naturally, having two identical rewrites is not of 
> practical interest but a way to illustrate the problem.
> 
> I am not clever enough to fully understand why and where we get the memory 
> corruption that results in a seg fault. I?m running SER on a Linux box 
> (kernel 2.4.20) and using WM 5.0 as test client sending REGISTERs to the 
> proxy.
> 
> Hopefully some one else can tell what?s going on here and how the crash can 
> be solved, or at least point me in the right direction.
> 
> 
> Happy holidays
> Regards,
> Johan
> 
> _________________________________________________________________
> Express yourself with cool new emoticons http://www.msn.co.uk/specials/myemo
> 
> _______________________________________________
> Serdev mailing list
> serdev at lists.iptel.org
> http://lists.iptel.org/mailman/listinfo/serdev