Hi Hesham, Greg.
Soliman, Hesham wrote:Christian, Thanks for the detailed description. I think Nick brought this up some time ago too. My suggestion is that upon reception of an RS with no SLLAO the router checks if an entry already exists, if none exists then it creates one and puts it in STALE. [...]
Greg Daley wrote:
Putting things in STALE doesn't work unless there's a link-layer address known ( and there's none in the received RS).
Greg is correct. When a node has a packet for a neighbor for which the NC entry is STALE, it does send the packet (trial and error), puts the entry into DELAY, and waits a while for upper-layer reachability confirmation. If the node receives reachability confirmation, then the entry goes back to REACHABLE. Otherwise it goes to PROBE, and the active part of NUD begins (i.e., the node starts transmitting NS's for the neighbor).
As far as I see it, there is currently no state defined in RFC 2461[bis], except INCOMPLETE, that is appropriate for a NC entry without link-layer address. And INCOMPLETE has the additional semantics that address resolution is in progress.
I guess this is why FreeBSD introduces a new state, NOSTATE. It does not do immediate address resolution on an entry in this state. It doesn't need to, because Rtadvd (on FreeBSD) sends multicast RA's in all cases except for ISATAP interfaces. When a packet is eventually to be transmitted for an entry in NOSTATE, that entry just moves to INCOMPLETE and a NS is sent.
The question is whether keeping entries in an additional state makes a lot of sense. The approach is more conservate, though, than doing address resolution rightway (which is how things are done in Linux according to Greg).
If an RS contains a TSLLAO [1], the router does not have to immediately initiate address resolution (i.e., be conservative), but can still send a unicast RA.
Soliman, Hesham wrote:[...] If an entry already exists with a LLA then it responds with a (two options):
- unicast RA unless a multicast RA was already scheduled.
- A multicast RA.
I think the second option might be better to allow for ODAD to work.
Hesham, why would a multicast RA be required for ODAD? An optimistic node can always send a RS from the unspecified source address to have the router multicast the RA.
Unicast RA's could be advantageous on link layers with acknowledgements, like IEEE 802.11, where they are realiably transmitted.
- Christian
[1] draft-daley-ipv6-tsllao-01.txt
-- Christian Vogt, Institute of Telematics, University of Karlsruhe www.tm.uka.de/~chvogt/pubkey/
"No great genius has ever existed without some touch of madness." (Aristotle)
-------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
