On Thu, Aug 18, 2016 at 05:55:16AM -0700, Eric Rescorla wrote:
> On Thu, Aug 18, 2016 at 5:46 AM, Ilari Liusvaara <[email protected]>
> wrote:
> 
> Agreed. Table fail. Anyway, if it goes anywhere (see below) it should go in
> ServerHello. That was the intent for this draft.

Okay, good to know (/me goes to change it in his code).

> > Then why does that server signature_algorithms exist? I think that
> > special PSK authentication modes should be specified in the PSK
> > extension, so non-PSK clients don't get confused and do inapporiate
> > things with those indications.
> >
> 
> Can you elaborate on what you mean here? If it's what I think you mean
> (namely put it in the server's pre_share_key message) I used to have that
> and davidben convinced me this was better. There's no perfect encoding
> here, AFAICT, so if you want to make an alternate suggestion that we could
> take a look at, that would be great.

The issue I'm thinking about is that if missing signature_algorithms means
no certificate, some dumb TLS implementation (and there are plenty of those)
will heed that even in non-PSK mode...
 
> > David Bejamin already posted a PR about that. Doesn't clearly say
> > that unknown reason handling.
> >
> 
> Yeah, I read the list before the PRs. I'll take a look but if you want to
> comment that would be great.

The only comment is that I would prefer to allow extensions the without
client including them in ClientHello[1] (obviously fatal if unknown one is
seen[2]), instead of special-casing Cookie.



[1] Of course, if the error is related to values in extension, then the
client had to include it in its ClientHello.


[2] Basically, if client is told that its ClientHello is wrong in some
way it does not understand, obviously it can't fix it.



-Ilari

_______________________________________________
TLS mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/tls

Reply via email to