Re: CVE Request (ruby -- DNS spoofing vulnerability in resolv.rb)

[Tomas Hoger <thoger () redhat com>]

On Thu, 4 Sep 2008 12:01:04 -0400 (EDT) "Steven M. Christey"
<coley () linus mitre org> wrote:

> > The transaction IDs are assigned in sequential (n+1 order) and the
> > source ports are always the same.
>
> Use CVE-2008-3905, to be filled in soon.
>
> We're treating this as a distinct issue because this is *REALLY* bad
> randomness within a particular implementation, besides the inherent
> limitation of DNS when source ports are fixed.

Applying this rule, separate id should probably be used for PyDNS [1]
[2] and adns [3] as well, at they both suffer from the similar flaws -
use predictable transactions ids and source port.

PyDNS should be fixed as of upstream version 2.3.2 [4], adns issue is
rather considered a design decision as documented in the INSTALL file
[5].

[1] http://pydns.sourceforge.net/
[2] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=490217
[3] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=492698
[4]
http://packages.debian.org/changelogs/pool/main/p/python-dns/python-dns_2.3.3-1/changelog
[5] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=492698#15

-- 
Tomas Hoger / Red Hat Security Response Team