[Openswan Users] 2.6.24rc1 (klips) segfault on client ip change

[Sven Schiwek]

On Oct 28, 2009, at 7:48 AM, David McCullough wrote:
> Jivin Sven Schiwek lays it down ...
>> Hi,
>>
>> with Openswan 2.6.24rc1 (klips) on Kernel 2.6.30.9 the pluto process
>> dies if the client ip has changed (not NATed) with this syslog  
>> message:
>>
>> ----8<----
[...]
>> ---->8----
>>
>> From the ongoing discussion "L2TP/IPSEC response unencrypted (was
>> openswan-2.6.24rc1 NATed MacOS Kernel crash)" I have installed the
>> first patch from David "natt-oa.patch" (don't know if this is  
>> relevant).
>
> Unfortunately it's not.  This was a klips kernel oops,  not a pluto  
> crash.
> Perhaps someone can run you through getting a stack trace out of  
> pluto ?
> I can't say it's something I do enough to know the best way on a  
> desktop
> system ;-)



Hi David,

sorry, it was my mistake. The Linksys router is a Road Warrior and for  
this client-type I need the option "dpdaction=clear" and not  
"dpdaction=restart" (according to the manpage).

But I believe it's not normal that a wrong config option segfault the  
pluto process. If I should still analyze this problem so that some  
developer can for example add a warning message I need some debug  
assistance, because I never debugged such a large program.

Please let me know what to do next. Thanks,
Sven