|Main Archive Page > Month Archives > oss-security archives|
On Mon, 06 Sep 2010 20:19:52 +0200 Florian Weimer wrote:
> > 1, Network Security Services (NSS) handled wildcard (*) character
> > in the Common Name field of a x509v3 digital certificate.
> > If an attacker is able to get a carefully-crafted certificate,
> > signed by a Certificate Authority trusted by Firefox, the
> > attacker could use the certificate during the man-in-the-middle
> > attack and potentially confuse Firefox into accepting it by
> > mistake. Different vulnerability than CVE-2009-2408.
> > References:
> >  http://www.westpoint.ltd.uk/advisories/wp-10-0001.txt
> >  http://bugs.gentoo.org/show_bug.cgi?id=335731
> Is this really a _security_ bug? The CN was not validated by the CA,
> so it's the CA's fault (which you have to trust, but still).
If we look back to CVE-2009-2408, similar bug was fixed at that time
without being handled as security flaw. NSS used to allow '*' wildcard
to match more than one host name label, or even the whole name. As
noted elsewhere in this thread, such wildcard handling can be found
elsewhere too, with fixes handled as security enhancements, rather than
security flaw fixes.
-- Tomas Hoger / Red Hat Security Response Team