Openwall Project   /home  Owl  JtR  Pro  crypt  pam_passwdqc  tcb  phpass  scanlogd  popa3d  msulogin  /  Linux  BIND  /  advisories  presentations  /  services  donations  /  wordlists  passwords  /  NEWS  community  lists  Wiki  CVSweb  mirrors  signatures
bringing security into open environments
 
Password Recovery Resources on the Net
[<prev] [next>] [<thread-prev] [thread-next>] [month] [year] [list]
Date: Tue, 17 Mar 2009 15:51:09 -0400 (EDT)
From: "Steven M. Christey" <coley@...us.mitre.org>
To: oss-security@...ts.openwall.com
Subject: Re: CVE-2009-0876 (VirtualBox) references


On Tue, 17 Mar 2009, Nico Golde wrote:

> Any reason the CVE description says "Unspecified
> vulnerability...via unknown vectors"?

This was based on the original Sun alert.  I cleaned up the description
yesterday, and the CVE web site was updated today.

> Looking at the Gentoo bug report[0] it seems obvious to me
> that this is caused by insecurely loading shared libraries
> so you can inject your own shared lib code.
>
> [0] https://bugs.gentoo.org/show_bug.cgi?id=260331#c0

It wasn't particularly obvious to me.  I may be getting hung up on the use
of hardlinks.

Is the problem that the executable includes a "." in its library path
(presumably DT_RPATH), and that path isn't cleansed until later during
program execution?  If it's just that, then the use of a hardlink doesn't
seem to be essential - the attacker could run the program from their own
directory.  Or, is it that the executable eventually removes "." from its
path, but not before some libraries have already been loaded?

- Steve

Please check out the Open Source Software Security Wiki, which is counterpart to this mailing list.

Powered by Openwall GNU/*/Linux - Powered by OpenVZ