Re: [TLS] Signature Algorithms

[Hubert Kario <[e-mail]>]

On Sunday 15 March 2015 03:40:38 Mehner, Carl wrote:
> >I believe we are going to remove this entire field and its associated
> >code points because it is redundant with the signature algorithms.
> 
> sounds good.
> 
> Speaking of Signature Algorithms, what do you think of this?
> 
> 
> Since root certificates are trusted by their identity rather than the
> signature I think that amendments to the text in the Server Certificates
> and Certificate Request sections are necessary.
> 
> As we move into a world that lacks trusted SHA-1 signatures, a change in the
> text would be necessary in order for clients that drop SHA-1 from the
> supported hash algorithms to continue to connect to servers that send a
> certificate_list that includes roots signed with SHA-1.

server that provides a self signed certificate is misconfigured in the first 
place...

a change that would reflect better the real world deployments would be to 
change the MUST to SHOULD - significant portion of servers ignore 
signature_algorithms when selecting server certificate (i.e. send sha-1 signed 
certs even when client didn't advertise SHA-1 in sig_algs)

and its the client's duty anyway to decide if it trusts a chain or not, the 
server should just provide one that is most likely to be trusted, not second-
guess client's intentions
-- 
Regards,
Hubert Kario
Quality Engineer, QE BaseOS Security team
Web: www.cz.redhat.com
Red Hat Czech s.r.o., Purkyňova 99/71, 612 45, Brno, Czech Republic