At Tue, 5 May 2009 10:11:33 -0400, "Hemant Singh (shemant)" <shem...@cisco.com> wrote:
> Sorry, I don't consider your example as valid because there is at > least one step in your example's sequence of events that is invalid. > That step is mentioned below in quotes > > "- Y sends an NS to P::X without link-layer source address option, with > the source address being P::Y." > > According to RFC 4861, a mcast NS MUST include the src link-layer > option. So if the NS in your example was a mcast NS with no source > link-layer option, then the NS is invalid. Further, if your NS is a > unicast NS which is a NUD message, then the L2 link-layer address of > Y is available to X when X receives the unicast NUD message. It's of course unicast (note the "to P::X"). BTW I don't understand this part: "the L2 link-layer address of Y is available to X when X receives the unicast NUD message." Why is this ensured? For example, X may have just been rebooted and its neighbor cache may be empty. --- JINMEI, Tatuya Internet Systems Consortium, Inc. -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------