[Openswan dev] pluto crash (DoS)

[D. Hugh Redelmeier]

| From: Rene Mayrhofer <rene.mayrhofer at gibraltar.at>

|  It's about the crash I 
| reported earlier, e.g.
| 
| http://lists.openswan.org/pipermail/dev/2005-January/000757.html

This message appears to be the end of a thread.  That thread isn't
clear to me.  Your message started it, and the archive has no content
for that message.

Paul's message suggests reading a users list message with no
explanation.  The all I get out of it is that 2.3.0 should be ignored.

Michael's message isn't clear without reading your (missing) message.

| http://lists.openswan.org/pipermail/dev/2005-April/000844.html
| and the following messages.

This thread seems more useful.  I even vaguely remember it.

I'm not working on openswan, so I'm going to be lazy and leave you to
do some leg work.

Can you predictably crash this setup?  That would be great.

Are you getting a core dump?  If not, please do.

Are you running a pluto compiled with -g (to keep symbols so that gdb
can reasonably interpret the core file)?  If not, please do.

When you get a core file, how did it crash?  If it was the same
assertion failure, could you get gdb to print out the values of
st->st_new_iv_len and e->enc_blocksize?  Might as well print out *e
and *st in case they have other useful information.  It is also useful
to see the output of the gdb command "where".