On Mon, Aug 11, 2008 at 02:14:46PM +0100, saqmaster wrote:
| > And here's the gist of your problem. Your vic0 interface doesn't have
| > a link-local address.
| >
| > That's very interesting - did you somehow configure this yourself ?
| > Generally speaking, when an interface is up and your kernel supports
| > IPv6, the interface gets a link-local address (see lo0 and gif0 for
| > examples, they're the fe80::/64 addresses).
| 
| Actually, there was one before.. 'someone said' I don't need it, so I
| removed it from the interface to see if it made any difference. Alas,
| it did not. So to clarify, with the link-local address in there before
| (fe80xxx), the problem was exactly the same. Well, I say exactly - I
| don't have before/after verbose tcpdumps, but the error messages
| presented were identical.

Like it or not - they're required by many parts of the stack, most
notably neighbor solicitation (of which RS/RA is a part). Don't remove
them, you'll break stuff.

Try again with link-local addresses on your interfaces, no rtadvd.conf
file and tcpdump to see what's going on. Post back if you have more
details.

BTW, I think the behaviour you've found warrants a bug report (use
sendbug(1)). From rfc2461, section 4.2 :

        Source Address
                MUST be the link-local address assigned to the
                interface from which this message is sent.

If rtadvd sends RA's with non-link-local src addresses, that is a bug
and it should be fixed.

| I'll have a play around with it, I don't really want to go to physical
| for this, so i'm entertaining ideas :)

Good luck ;)

Paul 'WEiRD' de Weerd

-- 
>++++++++[<++++++++++>-]<+++++++.>+++[<------>-]<.>+++[<+
+++++++++++>-]<.>++[<------------>-]<+.--------------.[-]
                 http://www.weirdnet.nl/                 

Reply via email to