Re: KEYS_READY

[David Schinazi <[e-mail]>]

Hi Martin, I've read the PR (in its latest version) and the comments in my
previous email still hold.
The text I object to from the PR is the following:

    A KEYS_ACTIVE frame sent in 1-RTT packets indicates that Handshake keys
are no
    longer needed.  A client MUST send this frame in its first 1-RTT
packet, and a
    server MUST send this frame in the first packet it sends after
completing the
    handshake.  A server might send 1-RTT packets prior to this; a server
MUST NOT
    process 1-RTT packets until the cryptographic handshake is complete.

This means that the server sends KEYS_ACTIVE(1-RTT) not when the 1-RTT keys
are active,
but instead when the handshake finishes, which is when the server has
received the client finished.
Those are two different moments in time.

My proposal is to have the client decide when it is done with the handshake
keys by having the client
delay sending RETIRE_KEYS(Handshake) instead of the server delaying sending
KEYS_ACTIVE(1-RTT).

David

On Thu, Feb 14, 2019 at 2:30 PM Martin Thomson <[e-mail]> wrote:

> On Fri, Feb 15, 2019, at 09:02, David Schinazi wrote:
> >
> > On Thu, Feb 14, 2019 at 1:02 PM Martin Thomson <[e-mail]>
> wrote:
> > > Please explain how the two differ, because I'm not seeing a difference.
> >
> > Here's the difference between them as I understand it:
> >
> > I'd initially assumed KEYS_READY meant that the keys were installed -
> > meaning that you send KEYS_READY when your keys are installed. But not
> > quite! The server does not send KEYS_READY when the 1-RTT keys are
> > installed, it sends it when it believes the handshake is done. So now
> > you have a notion of keys being ready that is impacted by parts of the
> > protocol further away.
> >
> > RETIRE_KEYS just means that you're done sending with a set of keys - so
> > you send it when you have received ACKs for everything you cared about
> > sending with those keys. I find that a lot more easier to reason about.
>
> Take another look at the PR and see if you can spot the difference.  You
> will find that maybe you are just arguing for a name change here.
>