[PATCH] Implementation of draft-ietf-tcpm-newcwv-06

[Adrian Chadd]

Hi!

This is looking better!

Structurally though, I'd look at all those places where you only
update tp->newcv and instead of passing in the tp, pass in a pointer
to tp->newcv. That way you're keeping a lid on the scope of the helper
functions - they only get access to as much data as they need so you
later won't be tempted to do things like go "oh, I just need to get
access to this one tcpcb field!" and suddenly it's not so well
contained.

As for behavioural - I think you'll have to poke Robert or Lawrence a
little more just to get some feedback from them.

It's good that it's disabled-by-default for now - that lets it get
into -HEAD with a lack of surprise.

Thanks for doing this!



-a