Hello, draft-eubanks-chimento-6man suggests that 0 checksum should be (conditionnaly) allowed for UDP/IPv6 packets.
First, I do not understand the rationale. As the draft notes, UDP-Lite already allows not to checksum the payload of an IPv6 datagram (inconditionnaly). Legacy NATs are the usual case against UDP-Lite. But that concerns only (outer) IPv4 not IPv6. Second, I believe the proposed change is harmful. Some widely deployed IPv6 implementations do follow RFC2460 §8.2 to the letter. Namely they do "discard UDP packets containing a zero checksum, and (...) log the error". I am not very keen on the idea that we would update the spec in a way that will cause plenty of dummy errors to be logged by "legacy" IPv6 stacks. FYI, at least the Linux kernel falls in this case. -- Rémi Denis-Courmont -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
