The Dual-Stack Lite draft doesn't mention a different treatment for packets with the Don't Fragment bit set. RFC 2473 does.
So the procedure in the draft is either incomplete (and should be ammended) or different from RFC 2473 (so it should state how and why it is different). Example: If a packet of 1500 bytes and the Don't Fragment bit set is passed into a DS-Lite tunnel with a Tunnel MTU of 1460 (1500 - 40 bytes for headers) , according to the draft it should be encapsulated, and fragmented in IPv6, since it's the tunnel headers that cause it to be too big. In a RFC 2473 compliant tunnel, such packets would be dropped (they are too large for the Tunnel MTU, and can't be fragmented) and an ICMP Unreachable, Packet Too Big should be sent. 2011/5/9 Jacni Qin <[email protected]> > > > Jacni>: It just says "Fragmentation MUST happen after the encapsulation on > the IPv6 packet" , seems to be not in conflict with RFC2473? > > > Cheers, > Jacni > > >
_______________________________________________ Softwires mailing list [email protected] https://www.ietf.org/mailman/listinfo/softwires
