Re: [quicwg/base-drafts] unrecoverable loss pattern leads to deadlock (#2863)

[martinduke <[e-mail]>]

I remember thinking RETIRE_KEYS was extraneous given all the implicit
signals. But it seems that screwy-but-legal behavior seems to make those
implicit signals risky. So let's stop being cute and just use the signal.

On Fri, Oct 18, 2019, 4:37 PM Kazuho Oku <[e-mail]> wrote:

> @RyanAtGoogle <https://github.com/RyanAtGoogle>
>
> it sounds like you are recommending something like the previous
> RETIRE_KEYS proposal?
>
> That might be the case (honestly I do not remember the details of the
> RETIRES_KEYS proposal), though I think I'm inclined to suggesting using a
> continuous signal (i.e., a Key Update or burning an unused header bit). The
> reasons being i) using a continuous signal allows us to abandon the
> Handshake keys earlier ii) we do not need to define a new signal that needs
> to be retransmitted in the transport draft, or rely on a particular
> behavior defined in the recovery draft that is not generally considered as
> necessary for a recovery logic (i.e. PTO "MUST" trigger a PING when there
> is no data to send).
>
> —
> You are receiving this because you were assigned.
> Reply to this email directly, view it on GitHub
> <https://github.com/quicwg/base-drafts/issues/2863?email_source=notifications&email_token=AF2EYEJOXBO3V7VFTTX5MJDQPJCFDA5CNFSM4H4LZFRKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEBWONZQ#issuecomment-544007910>,
> or unsubscribe
> <https://github.com/notifications/unsubscribe-auth/AF2EYEOL7XJAQAYKIY6MYLLQPJCFDANCNFSM4H4LZFRA>
> .
>


-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/quicwg/base-drafts/issues/2863#issuecomment-544019361