Erik, What you explained below is not very obvious from reading the paragraph. Since none of the authors of RFC 4861 commented on our drafts since June 2007 and counting, I decided to ping the Chairs and AD with the same paragraph. In that private email exchange, one person also did not understand this paragraph as you have explained. Further the new rephrased sentence that you propose is already along the lines of what we have said in tighter text in our on-and-off-link draft - for example see the last paragraph of section 2.1 of our on-and-off-link draft.
In 4861, in a lot of places there is text that says, "send all traffic to default router" when the text should say "send all non-link-local traffic to default router". If folks read our on-and-off-link draft, there are a lot of fixes like this over 4861 that we had done. Further, DHCPv6 in RFC 3315, and ND RFCs in 4861 and 4862 are collectively very lengthy documents. There is no document that defines how are DHCPv6 and ND in a IPv6 router and DHCPv6 server environment are supposed to behave or what is the boundary between DHCPv6 and the IPv6 router sending RA. We have summarized how DHCPv6 and ND interface with each other in our drafts. Such IPv6 info is useful to developers coming from an IPv4 world. On a separate note, folks also need to reply to the mailer if our ND implementation drafts can be accepted as a work item in 6man. Thanks. Hemant -----Original Message----- From: Erik Nordmark [mailto:[EMAIL PROTECTED] Sent: Wednesday, December 05, 2007 4:28 PM To: Hemant Singh (shemant) Cc: Suresh Krishnan; [EMAIL PROTECTED]; IETF IPv6 Mailing List Subject: Re: Here is the reference to 6.3.4 text that is ambigious text Hemant Singh (shemant) wrote: > Erik, > > I appreciate the quick reply. OK, here is an explanation with > non-colored text. The text snipped below for 6.3.4 is as follows to > please note first the text in double quotes: > > [Note, however, that a Prefix Information option with the on-link flag > set to zero conveys no information concerning on-link determination > and "MUST NOT be interpreted to mean that addresses covered by the > prefix are off-link."] So far we are talking about the semantics of a single prefix information option. > After the above text in the para the para says the following a little > later: > > [The default behavior (see Section 5.2) when sending a packet to an > address for which no information is known about the on-link status of > the address is to forward the packet to a default router;] And this talks about a particular destination address, looking at all of the prefix list. Thus you could rephrase the above sentence as When sending a packet to a destination and there is no matching prefix list entry, and no matching redirect entry, the packet is sent to a default router. > I am saying the para immediately above contradicts the text in quotes > because the quoted text says off link MUST NOT be interpreted but then > later the same section says "send data to default router". I interpret > "sending data to default router" as signaling off-link behavior. They are not in conflict, since the first is about a particular prefix information option, and the second is about the system as a whole. > As for the R bit, what if the router has implemented ONLY RFC 4861. I > don't want to bring RFC 4775 into the discussion just yet. I was merely trying to explain the motivation for L=0 being a no-op. Erik -------------------------------------------------------------------- IETF IPv6 working group mailing list [email protected] Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
