Stack overflow in RFC 8203 BGP admin. shutdown comm. handling since 7ff34ca2

[Daniel McCarney]

> If you could, i would be glad.

Done. I'll update this thread when MITRE replies.

Thanks again Ondrej,

On Sun, Sep 8, 2019 at 9:56 PM Ondrej Zajicek <santiago at crfreenet.org> wrote:
>
> On Sun, Sep 08, 2019 at 05:54:35PM -0400, Daniel McCarney wrote:
> > Hi Ondrej,
> >
> > Thanks for the quick response.
> >
> > > Unfortunately it has been included in released versions 1.6.7 and 2.0.5.
> >
> > Bummer, apologies for missing that. Do you want to request a CVE or should I?
>
> If you could, i would be glad.
>
>
> > While I believe 7ff34ca2 introduced the ability to overflow a stack buffer it
> > seems to me the original RFC 8203 support hasn't been correctly verifying
> > shutdown communication `msg_len` since support was added in BIRD 2 versions >=
> > 2.0.0 and BIRD 1 versions >= 1.6.4. Details to follow.
>
> I think that the incorrect check in the original code also allows this
> stack overflow, as a properly packed 255B message would trigger the first
> condition but not the second, so would be accepted.
>
> Therefore, the stack overflow could happen on BIRD 1 versions >= 1.6.4
> and BIRD 2 versions >= 2.0.0.
>
>
> The bugfix patches are:
> 1657c41c96b3c07d9265b07dd4912033ead4124b (1.6.x)
> 8388f5a7e14108a1458fea35bfbb5a453e2c563c (2.0.x)
>
> --
> Elen sila lumenn' omentielvo
>
> Ondrej 'Santiago' Zajicek (email: santiago at crfreenet.org)
> OpenPGP encrypted e-mails preferred (KeyID 0x11DEADC3, wwwkeys.pgp.net)
> "To err is human -- to blame it on a computer is even more so."