> -----Original Message----- > From: Suresh Krishnan [mailto:[EMAIL PROTECTED] > Sent: Thursday, March 20, 2008 4:30 PM > To: Markku Savela
> > And, if you hit unknown header, there is *NO WAY* to skip > over it. You > > have no idea whether it is an extension header (following > the standard > > format), or something totally different. > > And hence this text in the future work section of the draft about > remaining issues to be resolved. > > o Unknown extension headers cannot be differentiated from unknown > upper layer protocols > > We had discussed this in the 6man meeting and there is no > consensus on > the way forward. A common extension header format is ONLY ONE STEP in > the direction of the solution. It is certainly not the > complete answer > and there is no claim in the draft that says otherwise. I think the wording I quoted previously from RFC 2460 is intended to keep the receiver of the packet, the one identified in the IP DA, from skipping over unknown EHs. Which seems like the right thing to do. An unknown EH at the receiver should generate and ICMP error, I think. If this draft wants to bypass that rule: If new extension headers are defined and the intermediate node is not aware of them, the intermediate node cannot proceed further in the header chain since it does not know where the unknown header ends and the next header begins. The main issue is that the extension header format is not standard and hence it is not possible to skip past the unknown header. This document intends to define a standard format for IPv6 extension headers. perhaps it makes sense to specify clearly that this skipping option ONLY applies to intermediate nodes? It doesn't seem prudent to leave that unspecified. Bert -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------