Re: CVE for crypto_get_random() from libsrtp

[Adam Maris <amaris () redhat com>]

Hello,

The weakest method it provides uses no encryption at all, just HMAC-SHA1 
with 80 bit authentication tag:
http://srtp.sourcearchive.com/documentation/1.4.2.dfsg/group__SRTP_g94d0056e812802ac2920aa474bc5b59b.html

Unless CVE is assigned, we don't plan to ship any patch at the moment.

Regards,

On 01/08/15 11:31, Michael Samuel wrote:
> Hi,
>
> I can't see any reference to it using 80 bits of random data - it looks
> like it's AES-CTR mode.  Do you have further information on that?
>
> That being said, I can see quite a few ways it can go wrong - it's doesn't
> appear thread-safe for a start.  Is it worth taking a closer look or are
> you planning on shipping the patch anyway?
>
> Regards,
>    Michael
>
> On 31 July 2015 at 22:47, Adam Maris <amaris () redhat com> wrote:
>
> > Hello,
> >
> > I've got question whether this bug (
> > https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=793971) is CVE-worthy?
> > Could it be classified as CWE-330: Use of Insufficiently Random Values?
> >
> > According to the SRTP documentation (
> > http://srtp.sourcearchive.com/documentation/1.4.2.dfsg/group__SRTP_g1d4c228c6a58096dfab3cefbabd66f17.html),
> > it provides 80 bits of random data, which is quite a borderline.
> >
> > Thanks.
> >
> > --
> > Adam Maris / Red Hat Product Security

--
Adam Maris / Red Hat Product Security