[quicwg/base-drafts] Decoding Largest Acked and Reliability (#731)

[Patrick McManus <[e-mail]>]

In an ACK frame largest acked is "A variable-sized unsigned value". The text doesn't give a lot of guidance on exactly how its determined (other than its size), but presumably it uses the algorithm from 5.8.. except 5.8 clearly doesn't exactly apply because the packet number defined by 5.8 is "used as part of a cryptographic nonce for packet encryption. " which largest acked is not.

In any event, assuming a similar algorithm we would decode by " finding the packet number value that is closest to the next expected packet.".. which in the case of largest ack might be "largest previously received largest ack + 1".

So far, that's editorial. But, as is the case with packet envelope numbers, its certainly possible to mis-decode the number.. especially when using small encodings and reorderings occur. (and this is especially true with acknowledgment information which is actually retransmitted, unlike the packet envelope numbers themselves).

If you mis-interpret an envelope packet number its no big deal - the crypto will fail and you'll drop the packet. But if you mis-interpret a largest acked field then you mark as acked the wrong data. Perhaps data that never arrived and will now never be re transmitted. And you don't have a reliable protocol anymore. :(

oops?

-- 
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/731