[9667] in bugtraq
Re: [HERT] Advisory #002 Buffer overflow in lsof
daemon@ATHENA.MIT.EDU (Robert Watson)
Fri Feb 19 16:53:40 1999
Date: Fri, 19 Feb 1999 13:35:09 -0500
Reply-To: Robert Watson <robert+freebsd@cyrus.watson.org>
From: Robert Watson <robert@CYRUS.WATSON.ORG>
X-To: Mariusz Marcinkiewicz <many@ENSI.NET>
To: BUGTRAQ@NETSPACE.ORG
In-Reply-To: <99Feb19.015748met.11650@fw.ibs.net.pl>
On Fri, 19 Feb 1999, Mariusz Marcinkiewicz wrote:
> On Thu, 18 Feb 1999, Don Lewis wrote:
>
> > ... or are there systems that give group kmem write privileges? If so,
> > I'd say that's a security hole.
>
> Yes, you are right... but... I saw that hole after installing new linx and
> checked it's security. First I was suprised but not for a long time.
> In a few mins I noticed all linux versions are chown .kmem; chmod g+s
> lsof... on linux /dev/kmem is +w for gid kmem, on bsd too (probably, I
Sorry, no go. FreeBSD 2.2-STABLE and 4.0-CURRENT, the two versions I
have sitting around, have the following permissions on /dev/kmem:
crw-r----- 1 root kmem 2, 1 Mar 7 1998 /dev/kmem
Please verify claims such as these before posting them. FreeBSD also does
not ship with lsof installed, although if the package/port is installed,
it will be installed sgid kmem.
> didn't checked that), so... all of std. distributions are vuln. without
> ONE! the slackware, IMHO, it's the most secure distribution [ :))) i know:
> slackware doesn't has lsof;))) but by tahat way that distr. is secure ;P ]
Write access to /dev/kmem would certainly be fairly nasty--you might as
well give the user kernel privileges outright (to be distinguished, btw,
from root access, due to securelevels). As has been pointed out, even
read access to /dev/kmem can be sufficient to gain root access, it just
requires a little more skill and patience, however. I do not know of a
way, however, to gain kernel privileges using only read access to
/dev/kmem, when securelevels are in use (correctly).
I haven't given that particular problem much thought, and discouraging
kmem access both in and out of securelevel is probably a good idea for
other reasons (such as IPsec keying material privacy, passwords, etc).
Robert N Watson
robert@fledge.watson.org http://www.watson.org/~robert/
PGP key fingerprint: 03 01 DD 8E 15 67 48 73 25 6D 10 FC EC 68 C1 1C
Carnegie Mellon University http://www.cmu.edu/
TIS Labs at Network Associates, Inc. http://www.tis.com/
SafePort Network Services http://www.safeport.com/
