siegert at sfu.ca
Fri Sep 19 15:08:26 EST 2003
On Thu, Sep 18, 2003 at 01:01:23PM +1000, Darren Tucker wrote:
> Martin Siegert wrote:
> > the following problem occurs on Solaris 2.6. openssh-3.7p1 and openssh-3.7.1p1
> > both show the same behaviour.
> I've had a closer look at the debugging here (pretty good info, BTW).
> Your gdb+backtrace doesn't capture the problem, however, since the
> backtrace is from the privileged process and the SEGV appears to be
> occurring in the unprivileged child.
> Can you try:
> 1) Reproducing the problem with "UsePrivilegeSeparation=no".
> If it happens with privsep=no, use gdb to get a backtrace and post it.
Yes, it happens with privsep=no.
> 2) If it doesn't happend with privsep, you need to try and debug the
> child, which can be tricky. I suggest setting a breakpoint for sshd.c:650
> (just before the fork), then set "set follow-fork child", then continue.
> Hopefully this will catch it so you can do a backtrace.
It seems to happen in the child that is forked in pthread_create in
auth-pam.c although I failed to used "set follow-fork child" - gdb
followed the parent nevertheless. I inserted a sleep(20) and attached
a second gdb to the child.
> I also suggest that if you haven't already, open a bug at
> bugzilla.mindrot.org (check for dupes first) as this looks like it might
> take a bit of work and it's easier to track if it's in bugzilla.
Did that: bug #687 with attachment #431.
Manager, Research Services
WestGrid Site Manager
Academic Computing Services phone: (604) 291-4691
Simon Fraser University fax: (604) 291-4242
Burnaby, British Columbia email: siegert at sfu.ca
Canada V5A 1S6
More information about the openssh-unix-dev