Re: Breaking out of chroot
From: Tim Haynes (usenet-20030826_at_stirfried.vegetable.org.uk)
Date: Tue, 26 Aug 2003 00:22:47 +0100
Floyd Davidson <firstname.lastname@example.org> writes:
> Exploits using resources that you aren't making available do not result
> in security breaches.
Yes, but it leaves one gap to consider: say, for example, I run bind9 -t
-u, which effects a chroot and setuid() to someone unprivileged; now, if I
see a snort signature for a known bind exploit, how do I test the integrity
of the box? Worse, if I see something successfully breaking into bind9, how
do I guarantee the *rest* of the box wasn't breached outside the jail, ie
that someone didn't break the jail?
ISTM this would be a valuable skill to enhance, anyway.
-- Bagpuss gave a big yawn, |email@example.com and settled down to sleep. |http://spodzone.org.uk/