Clarify coalesced packet processing #1844
Conversation
draft-ietf-quic-transport.md
Outdated
| packets were received out-of-order in separate datagrams. | ||
| they were received as the payload of separate UDP datagrams. For example, if | ||
| one or more packets in a datagram cannot be decrypted (the keys are not yet | ||
| available, decryption failure, unknown type), the receiver MAY either discard or |
There was a problem hiding this comment.
", decryption fails, or the packet is of an unknown type"
There was a problem hiding this comment.
Thanks. Applied the suggestion.
draft-ietf-quic-transport.md
Outdated
| they were received as the payload of separate UDP datagrams. For example, if | ||
| one or more packets in a datagram cannot be decrypted (the keys are not yet | ||
| available, decryption failure, unknown type), the receiver MAY either discard or | ||
| buffer for packet for later processing and MUST attempt to process the remaining |
|
Long headers are useful for PMTUD to validate (and route) PTB ICMP messages. So we discussed in Seattle the idea of prepending long header packets to PMTUD Probe packets. |
draft-ietf-quic-transport.md
Outdated
| packets MAY either be discarded or buffered for later processing, just as if the | ||
| packets were received out-of-order in separate datagrams. | ||
| they were received as the payload of separate UDP datagrams. For example, if | ||
| decryption fails or the packet is of an unknown type, the receiver MAY either |
There was a problem hiding this comment.
This seems to lose the case this text was originally added to address, which is that you don't have the keys to process this packet type yet. While these are non-limiting examples, that's probably a case worth calling out explicitly.
There was a problem hiding this comment.
Makes sense. Updated the PR to mention this case specifically.
| packets were received out-of-order in separate datagrams. | ||
| they were received as the payload of separate UDP datagrams. For example, if | ||
| decryption fails (the keys are not yet available or any other reason) or the | ||
| packet is of an unknown type, the receiver MAY either discard or buffer the |
There was a problem hiding this comment.
"the packet" does not make sense in this context. How about "If one or more packets in a datagram cannot be processed yet, because of decryption failure or any other reason, the receiver MAY either discard or buffer the packets for later processing. The receiver MUST still attempt to process the remaining packets in the UDP datagram."
There was a problem hiding this comment.
That sounds good. Want to port that over to #1857?
There was a problem hiding this comment.
I would omit "yet". "Yet" implies knowing that it will be possible to decrypt the packet in the future. It is hard to know the future.
|
Merged by hand. |
Addressing #1840