Oh. I hadn't realized that this allowed greasing handshake & 0-RTT.
I think that's a great feature, but I might suggest mentioning it in as
many words to avoid any confusion.

David

On Thu, May 12, 2022 at 4:33 PM Martin Thomson <[email protected]> wrote:

> Hey folks,
>
> David just provided some feedback on this that suggested a more
> comprehensive rewrite.  In case you looked already, you might want to take
> another peek.  One consequence of that change is that clearing the bit is
> now also possible on 0-RTT and Handshake packets, which might have been how
> people implemented this (that's what I did).
>
> Cheers,
> Martin
>
> On Thu, May 12, 2022, at 16:58, Zaheduzzaman Sarker wrote:
> > Hello all,
> >
> > Martin and me has chatted about the 7 day rule and it appeared that not
> > following the rule would actually lead to connection failure. Hence,
> > this actually should be a MUST. See the pull request
> > https://github.com/quicwg/quic-bit-grease/pull/24.
> >
> > Please reflect on this change by the end of next week (20th May, 2022)
> > along with any thoughts regarding the 7 (day) number. Unless any
> > critical issue found, after that I will consider the AD review issues
> > are resolved and move the doc to IETF LC.
> >
> > //Zahed
> >
> >
> >> On 28 Apr 2022, at 04:33, Martin Thomson <[email protected]> wrote:
> >>
> >> Thanks Zahed.
> >>
> >> I've added references as you suggest and reworded some of the intro
> (fewer words!)
> >>
> >> You asked about the 7 day thing, which is almost entirely arbitrary.
> QUIC doesn't time limit NEW_TOKEN in any way, so this time is only
> necessary to avoid unbounded use of the mechanism (which might prevent a
> server from ever disabling greasing).  It's aligned with TLS requirements
> for session tickets, because that is likely natural for client
> implementations, but it is still ultimately an arbitrary value.  We can say
> that much, but I don't think we need to.
> >>
> >> On Wed, Apr 27, 2022, at 20:09, Zaheduzzaman Sarker wrote:
> >>> Hi,
> >>>
> >>> Thanks for the short and nice draft.
> >>>
> >>> I have done my AD review. I haven’t noticed any technical issues so
> >>> far. However, I have some editorial comments and I have created issues
> >>> for those (
> https://protect2.fireeye.com/v1/url?k=31323334-501d5122-313273af-454445555731-bfd564275429d53b&q=1&e=47a1b565-597e-492d-a7c0-c2291d74fae4&u=https%3A%2F%2Fgithub.com%2Fquicwg%2Fquic-bit-grease%2Fissues).
>
> >>>
> >>> //Zahed
> >>> Attachments:
> >>> * smime.p7s
> >
> > Attachments:
> > * smime.p7s
>
>

Reply via email to