On Sun, May 17, 2015, at 18:06, Dan Lukes wrote:> On 05/18/15 00:00, Mark Felder:
> >> If TLS 1.0 is considered severe security issue AND system
utilities are
> >> using it, why there is no Security Advisory describing this system
> >> vulnerability ?
> >>
> >
> > It's not a vulnerability in software, it's weakness in the
protocol
> > design.
>
> Like protocol protocol downgrade triggered by MITM attack flaw or
> protocol design flaw in session renegotiation support. The first one
> addressed in FreeBSD-SA-14:23.openssl, the second one in
> FreeBSD-SA-09:15.ssl
>
> So the "is it protocol flaw or implementation bug" seems not to
be true
> major criteria.
>
> OK, I wish I got best answer to my question possible. I'm not going to
> discuss SA issuing policy in this thread.
>
FreeBSD-SA-14:23: primarily backported a new feature (TLS_FALLBACK_SCSV)
to help prevent those with stronger crypto from being forced to
downgrade to weak crypto via a MITM attack
FreeBSD-SA-09:15: fixes some bugs dealing with potential MITM attacks
Neither of these directly address a broken protocol, such as warning all
users that "using SSL 3.0 or TLS 1.0 is dangerous"
I mean, should we have an SA because our libc supports strcpy and people
can use that and create severe vulnerabilities? Or the fact that there
is no firewall enabled by default, so you should probably enable one?
That seems a bit extreme. You could write a whole book and still not
cover all of these topics :-)
Hope that helps