Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 13 Mar 2013 21:56:44 +0100
From: Salvatore Bonaccorso <carnil@...ian.org>
To: oss-security@...ts.openwall.com, kseifried@...hat.com
Cc: Russ Allbery <rra@...nford.edu>,
	Remi Gacogne <rgacogne-bugs@...edump.fr>
Subject: Re: Reverse lookup issue in Net::Server

Hi

On Mon, Mar 11, 2013 at 08:42:44PM -0600, Kurt Seifried wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On 03/04/2013 12:36 PM, Russ Allbery wrote:
> > Remi Gacogne <rgacogne-bugs@...edump.fr> writes:
> > 
> >> I think there is a security issue in the way the access control
> >> feature of Net::Server
> >> (http://search.cpan.org/perldoc?Net%3A%3AServer) works. 
> >> Net::Server is used by various projects including Munin, Postgrey
> >> and SQLgrey.
> > 
> >> The issue lies in the fact that the allow / deny access control
> >> does not perform a valid DNS check when given a hostname
> >> parameter and the 'reverse_lookups' option is enabled.  The
> >> current code only checks that the incoming connection source IP
> >> address has a reverse DNS matching the given hostname, but does
> >> not check that the hostname resolves back to this source IP
> >> address (see how the $prop->{'peerhost'} property is set in
> >> get_client_info(), lib/Net/Server.pm:553, then used in
> >> allow_deny(), lib/Net/Server.pm:597).  As it is trivial for an
> >> attacker to be able to set his own source IP's reverse DNS, the
> >> current check is not safe (this probably matches CWE-807:
> >> Reliance on Untrusted Inputs in a Security Decision).
> > 
> > This is a very weak security measure, but yes, the need to check
> > the reverse DNS results with a forward DNS query to make the
> > security check at all useful has been well-known going all the way
> > back to the days when TCP wrappers was the UNIX firewalling system
> > of choice.  I remember discussion of this in security contexts in
> > 1994, and I'm sure it was an old discussion even then.
> 
> Yup. Please use CVE-2013-1841 for this issue.

Thank you Kurt for assinging the CVE. Upstream already answered on
this on the request tracker[1] and Paul mentions they will add an
option to do the forward lookup.

 [1]: https://rt.cpan.org/Ticket/Display.html?id=83909

Regards,
Salvatore

Powered by blists - more mailing lists

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

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.