Define what should happen if packets arrive outside Missing Packet Threshold

[michael-eriksson]

The document does not specify what should happen if a packet arrives more out of order than Missing Packet Threshold. An immediate ack is the most reasonable action, but it should probably be explicitly specified.

See also the quicdev Slack message: https://quicdev.slack.com/archives/CTDAH4H71/p1725959171986379

[mirjak]

Not sure I understand this comment. the draft says:

An endpoint that receives an ACK_FREQUENCY frame with a non-zero Reordering Threshold value SHOULD send an immediate ACK when the gap between the smallest Unreported Missing packet and the Largest Unacked is greater than or equal to the Reordering Threshold value.

This say "larger or equal" which should include this case, no?

[ianswett]

Unfortunately, I can't see the slack history anymore, but I agree with @mirjak that the current text seems to cover this. We could say it elsewhere if it's helpful, but I'm not sure where?

[michael-eriksson]

No, the text quoted from the draft above does not cover the case that I'm after.

Let's use Table 1 of draft-10 as an example. What should happen if packet 2 is received after packet 5 has been received (but before packet 8 is received)? An ack for packet 5 has already been sent, to indicate that packet 2 is missing. Packet 2 then arrives so late that it's outside the Reordering Threshold.

[mirjak]

This is covered in RFC9000 (and not changed):

In order to assist loss detection at the sender, an endpoint SHOULD generate and send an ACK frame without delay when it receives an ack-eliciting packet either:

* when the received packet has a packet number less than another ack-eliciting packet that has been received, or

[mirjak]

Okay, looking at the draft, this is actually not very clear. My understanding was always that we keep the behavior as specified in RFC9000 when the PN is less than an already received one and we only change the behavior in case of gaps. If so, we should maybe clarify this. @ian can you confirm?