Re: [dhcwg] WGLC: draft-ietf-dhc-dhcpv6-client-link-layer-addr-opt-01

[Chuck Anderson <[e-mail]>]

On Thu, Aug 16, 2012 at 05:43:12PM +0000, Ted Lemon wrote:
> Chuck, I think I can read your response as saying that you are happy with the draft as it is, and while you would not oppose having the draft say how the DHCP server might acquire the link local address for a packet that arrived without being relayed, you don't think it's necessary or desirable, correct?   (I ask only so that we will know what to do later on when we evaluate consensus).

I don't think it is necessary to update this draft to specify how a
DHCP server might acquire the link layer address from a non-relayed
packet, and I'm somewhat opposed to updating this particular draft for
that behavior now only because it would add more delay.

However, I do recognize that there may be benefit to updating section
4 to clarify what LL address is to be put into the RELAY-FORW option,
but I'm open to others' opinions on this.  Do others think this text
is clear enough, or do we need to clarify what goes into the "client
link-layer address option"?  Right now it just doesn't say what goes
in there:

"4.  DHCPv6 Relay Agent Behavior

   DHCPv6 Relay agents which are directly connected to clients/hosts MAY
   include client link-layer address option in relayed DHCPv6 (RELAY-
   FORW) message.  The DHCPv6 Relay agent behaviour can depend on
   configuration that decides whether Client Link-layer Address option
   needs to be processed and included.

   In Relay chaining scenarios, any other relay agent other than first
   hop DHCPv6 Relay agent or DHCPv6 LDRA [RFC6221] MUST not add this
   option."