Re: CVE Request: libgcrypt hardening for RSA-CRT leak

[Florian Weimer <fweimer () redhat com>]

On 09/08/2015 12:05 PM, Marcus Meissner wrote:
> Hi,
>
> Redhat has published a paper on RSA-CRT keyleakage.
>
> https://securityblog.redhat.com/2015/09/02/factoring-rsa-keys-with-tls-perfect-forward-secrecy/
>
> There was a CVE assigned for this issue CVE-2015-5738, but the software scope of this assigned is not clear.
>
> libgcrypt has published a hardening fix for the same issue.
> https://lists.gnupg.org/pipermail/gnupg-announce/2015q3/000370.html
> http://git.gnupg.org/cgi-bin/gitweb.cgi?p=libgcrypt.git;a=commit;h=b85c8d6645039fc9d403791750510e439731d479
>
> Should it get a new CVE?

For context, Oracle has assigned CVE-2015-0478 for the missing hardening
in the default JCE implementation:

https://bugzilla.redhat.com/show_bug.cgi?id=1210355
http://www.oracle.com/technetwork/topics/security/cpuapr2015-2365600.html#AppendixJAVA
https://access.redhat.com/security/cve/CVE-2015-0478

This case is similar to libgcrypt, I believe: no key leaks have been
attributed to this implementation, the change is purely hardening in
this sense (and I would not have assigned a CVE ID to this).

-- 
Florian Weimer / Red Hat Product Security