> On Tue, 14 Dec 2004, Pekka Savola wrote: > I went through one implementation and have a couple of additional > comments wrt suitability for DS/clarify. > > 1) section 6.2.5: when AdvSendAdvertisements changes to FALSE, you > SHOULD send a final RA with zero Router Lifetime. > > At least a couple of implementations do send out final RAs with zero > lifetime when the RA process is killed, but do not have 'state' which > monitors whether AdvSendAdvertisements gets disabled or not on an > interface. (e.g., at HUP signal) > > Are there implementations of this?
The Solaris in.ndpd implements this. > 2) section 6.2.5: if system management disables IP forwarding, > subsequent RAs MUST set the Router Lifetime to zero. > > I've seen no one implementing this (though many implement checks whent > the RA process starts up). This either requires polling forwarding > status constantly, or providing some kind of notifications when IP > forwarding changes from on to off. Implementations? Solaris in.ndpd implements this. > 3) section 6.2.7: RA consistency says like: > - Cur Hop Limit values (except for the unspecified value of zero). > > It is ambiguous what the ()'s means. A couple of possibilities: > a) if "my Cur Hop Limit" is zero, ignore consistency completely > b) if "my Cur Hop Limit" is zero, allow the others to have zero Cur > Hop Limit, otherwise it's inconsistent > c) if "my Cur Hop Limit" is non-zero, allow the others to have same > Cur Hop Limit, otherwise it's inconsistent > d) if "my Cur Hop Limit" is non-zero, allow the others to have same > Cur Hop Limit or zero, otherwise it's inconsistent > > This needs to be clarified. FWIW I think the intent was to say "zero is consistent with a nonp-zero value" but "two different non-zero values are inconsistent". > 4) section 6.2.8: if the link-local address of the router changes, it > should multicast a few RAs from the old address with zero router > lifetime, and a few from the new address. (SHOULD). > > I haven't seen this implemented. Similar reasons as 2). Anyone ? FWIW I don't see this implemented in in.ndpd. If we really want this detailed level of implementation reports, IMHO we need an implementation report form that asks these level of questions so that it can be passed to the different implementors we know of (and advertised to the list.) Erik -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
