Re: [TLS] Pull request for session hash

[[e-mail] (Martin Rex)]

Eric Rescorla wrote:
> 
> - IMPORTANT: Because the initial handshake messages are encrypted
>   under separate keys from the application data traffic, it seems like we
>   need to have 2 sets of CCS messages or none. In anticipation of removing
>   renegotiation (see PR https://github.com/tlswg/tls13-spec/pull/88).
>   Based on conversations with Alfredo and Martin, it seemed easier
>   to remove them. Warning: analysis needed here.

Huh?

>From an implementation standpoint CCS being a seperate content type
record that is clearly visible to the record layer is one of the better
ideas of the original SSLv3 protocol.

It reduces the necessity for synchonizing the record layer and
the handshake state machine, and allows having the record layer
concatenate consecutive handshake messages into a single TLS record
automatically on the sender side.
On the recipient side, it allows the record layer to preprocess a batch
of incoming TLS records without having to wait for the handshake message
state machine (some TLS stacks create a seperate TLS record for each
handshake message).

-Martin