Re: [TLS] Closing some open comments on draft-ietf-tls-renegotiation

[Marsh Ray <[e-mail]>]

Dr Stephen Henson wrote:
> Eric Rescorla wrote:
>> I'll also make whatever small editorial changes I see. If I've
>> missed something important that people think there is consensus on,
>> please let me know. There are a lot of messages, so I may well
>> have.
>> 
> 
> In 6.1:
> 
> If clients wish to ensure that such attacks are impossible, they MUST
>  terminate the connection immediately upon failure to receive the 
> extension without completing the handshake.
> 
> Should we be more specific about what "terminate the connection"
> means? Does that mean drop the connection immediately without sending
> anything else or by sending a fatal alert?

The spec says "Upon transmission or receipt of a fatal alert message,
both parties immediately close the connection."

> If a fatal alert which one? Presumably handshake_failure?

How about 'unexpected_message'? That's what I see when mixing
application data into a renegotiation handshake that wasn't expecting it.

RFC 4346 defines it as:
> unexpected_message An inappropriate message was received.  This alert
> is always fatal and should never be observed in communication between
> proper implementations.


> I'm thinking here in terms of sending something useful which would
> most likely appear in server logs to help diagnose the cause: i.e.
> unpatched server which strict clients refuse to connect to.

That would seem to set off the appropriate alarm bells.

- Marsh