[20662] in bugtraq

home help back first fref pref prev next nref lref last post

MUAs that delete spoolfiles (was Solaris /usr/bin/mailx exploit (SPARC))

daemon@ATHENA.MIT.EDU (Rich Lafferty)
Wed May 16 08:56:43 2001

Date: Tue, 15 May 2001 17:00:43 -0400
From: Rich Lafferty <rich@alcor.concordia.ca>
To: bugtraq@securityfocus.com
Message-ID: <20010515170043.C28724@alcor.concordia.ca>
Mail-Followup-To: bugtraq@securityfocus.com
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <uag0e692im.fsf@sapphire.noc.gxn.net>; from andrew.hilborne@uk.xo.com on Tue, May 15, 2001 at 02:15:45PM +0100

On Tue, May 15, 2001 at 02:15:45PM +0100, Andrew Hilborne (andrew.hilborne@uk.xo.com) wrote:
> > 
> > (At least not if you /var/mail directory has the standard 1777 permissions)
> > 
> > By forcing a file permission of 600 on mailboxes, group mail should not
> > gain you anything.
> 
> Just how do you force 0600 on mailboxes which don't exist (many MUAs remove
> empty mailboxes?)

If that's true, then even *without* this particular bug in Solaris,
there's an icky denial of service attack waiting to happen. Sticky
mailspools are awfully common these days, and all that stops Bob from
doing

  touch /var/spool/mail/alice

and causing the MTA to refuse to deliver is that Alice's mailbox
should never *not* be there in the first place. 

Which MUAs behave in the way you describe?
 
> Since you cannot easily do this, at the very least a malicious user should be
> able to steal other users' mail. I think.

If they can, then *that's* a flaw in the MTA, which should never
deliver into something that isn't owned by the recipient.

  -Rich 

-- 
------------------------------ Rich Lafferty ---------------------------
 Sysadmin/Programmer, Instructional and Information Technology Services
   Concordia University, Montreal, QC                 (514) 848-7625
------------------------- rich@alcor.concordia.ca ----------------------

home help back first fref pref prev next nref lref last post