Re: nss RSA forgery (CVE-2014-1568)

[Yves-Alexis Perez <corsac () debian org>]

On Thu, Sep 25, 2014 at 12:10:36AM +0200, Marcus Meissner wrote:
> On Thu, Sep 25, 2014 at 12:03:21AM +0200, Hanno Böck wrote:
> > One serious vuln per day isn't enough, so nss decided to bring us
> > another one.
> >
> > Mozilla reports this:
> > https://www.mozilla.org/security/announce/2014/mfsa2014-73.html
> > Bugtracker entry still private, so hard to judge about details.
> > Interesting: Two independent discoveries (we had the same with
> > heartbleed and I couldn't believe this was coincidence).
> >
> > This is what mcaffee has to say:
> > http://blogs.mcafee.com/executive-perspectives/need-know-berserk-mozilla
> >
> > They say its related to BER/ASN1-parsing, but adam langley disagrees:
> > https://twitter.com/agl__/status/514881918110683136
> >
> >
> > And it seems cyassl had something similar, also found by intel:
> > http://www.yassl.com/yaSSL/Blog/Entries/2014/9/12_CyaSSL_3.2.0_Released.html

So, according to:

http://www.intelsecurity.com/advanced-threat-research/#
http://tif.mcafee.com/threats/12817

that's indeed the same (or really close) vulnerability. Can a CVE be
assigned?
> > No real details yet and information seems confusing.
>
> I also see gnutls posted a new release some days, with this entry without CVE
>
> http://article.gmane.org/gmane.comp.encryption.gpg.gnutls.devel/7663
> "** libgnutls: Fixed issue with certificates being sanitized by gnutls prior
> to signature verification. That resulted to certain non-DER compliant modifications
> of valid certificates, being corrected by libtasn1's parser and restructured as
> the original. Issue found and reported by Antti Karjalainen and Matti Kamunen from
> Codenomicon."
>
> which smells similar.
When trying to test NSS updates I tried with epiphany browser, which
uses gnutls (through libsoup and glib-networking), and it seemed indeed
affected. It's apparently fixed in the 3.3 branch at 3.3.8:

http://article.gmane.org/gmane.network.gnutls.general/3593

Can a CVE be assigned too?

Regards,
-- 
Yves-Alexis Perez

Attachment: signature.asc
Description: Digital signature