[quicwg/base-drafts] reordered NEW_CONNECTION_ID frames with retired sequence numbers shouldn't be used (#3194)

Marten Seemann <notifications@github.com> Wed, 06 November 2019 10:32 UTC

Date: Wed, 06 Nov 2019 02:32:57 -0800
From: Marten Seemann <notifications@github.com>
Reply-To: quicwg/base-drafts <reply+AFTOJK5B6VQ3LVCJZP4UXLV3Z7J5TEVBNHHB5ZPCDM@reply.github.com>
To: quicwg/base-drafts <base-drafts@noreply.github.com>
Cc: Subscribed <subscribed@noreply.github.com>
Message-ID: <quicwg/base-drafts/issues/3194@github.com>
Subject: [quicwg/base-drafts] reordered NEW_CONNECTION_ID frames with retired sequence numbers shouldn't be used (#3194)
Mime-Version: 1.0
Content-Type: multipart/alternative; boundary="--==_mimepart_5dc2a159df4b_3c3a3fdc4f8cd9609724"; charset="UTF-8"
Content-Transfer-Encoding: 7bit
Precedence: list
Archived-At: <https://mailarchive.ietf.org/arch/msg/quic-issues/ga4qe5VO1tNXOZGOtFN8uxhTRdo>

Let's say an endpoint receives a NEW_CONNECTION_ID frame with Sequence Number: X, Retire Prior To: M. At a later moment, it receives a (reordered) NEW_CONNECTION_ID frame with Sequence Number: N (where N < M).

As the spec is currently written, it would be perfectly legal to use CID N at this moment. Obviously, this doesn't make sense, since the receiving peer is about to or has already stopped accepting packets with using CIDs < M. When receiving such a reordered NEW_CONNECTION_ID frame, the reasonable thing to do is to immediately send the corresponding RETIRE_CONNECTION_ID frame for N.

-- 
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/3194

[quicwg/base-drafts] reordered NEW_CONNECTION_ID … Marten Seemann
Re: [quicwg/base-drafts] reordered NEW_CONNECTION… Eric Kinnear
Re: [quicwg/base-drafts] reordered NEW_CONNECTION… Martin Thomson
Re: [quicwg/base-drafts] reordered NEW_CONNECTION… Martin Thomson