[Openswan dev] Re: Pluto segfault debugging information

mcr at xelerance.com mcr at xelerance.com
Sat Feb 12 11:11:11 CET 2005


>>>>> "Robert" == Robert Hardy <rhardy at webcon.ca> writes:
    Robert> I'm Delemas on #openswan. We were recently discussing the
    Robert> odd segfaults of pluto I had been seeing recently. Please
    Robert> find attached, as requested on #openswan, a debug log, a
    Robert> core file and a pluto binary.

    Robert> Pluto segfaulted and dumped core on Jan 31 at 16:27:42. I
    Robert> wasn't sure how much logging info you needed, so I've

  I examined the core dump.

  My impression is that a connection and/or state was deleted while it
was being processed for a cryptographic operation. 

  So, when the cryptographic operation completed, it tried to reference
the connection and state => died.
  The asynchronous crypto operations need to be tolerant of having the
rug pulled out from under them.

  I wrote some code to delete the continuation if the state gets
deleted. I'm not sure how I'll test it yet.  I will be putting this code
through regression testing on Sunday. 
  This patch is against 2.3.1dr2.

- -- 
] Michael Richardson          Xelerance Corporation, Ottawa, ON |  firewalls  [
] mcr @ xelerance.com           Now doing IPsec training, see   |net architect[
] http://www.sandelman.ca/mcr/    www.xelerance.com/training/   |device driver[
] panic("Just another Debian GNU/Linux using, kernel hacking, security guy"); [

Version: GnuPG v1.2.2 (GNU/Linux)
Comment: Finger me for keys


-------------- next part --------------
A non-text attachment was scrubbed...
Name: state.patch
Type: text/x-c
Size: 3323 bytes
Desc: delete_continuation.patch
Url : http://lists.openswan.org/pipermail/dev/attachments/20050212/7ebfbb81/state.bin

More information about the Dev mailing list