In your previous mail you wrote: > On 04/04/2013 18:51, Hannes Frederic Sowa wrote: > > Hello! > > > > What is the proposed action if the hardware address of an interface is > > changed regarding link-local addresses while the interface is up? I see > > a few possibilities here but have not yet found an answer in the rfcs: > > > > a) generate a new link-local interface address in addition to the old one > > b) drop the old one and generate a new one > > c) don't change any addresses (this is e.g. current behavior in linux) > > > > I would be happy if you could give me some pointers. > > Does it matter? It seems that all three will work properly.
=> I fully agree: RFCs are only about link-local interface ID generation and as soon as it works no detail really matter. > c) will surely only apply until the interface is restarted; then you > will get b) automatically. => note the "restart" is OS dependent. On OSs where the IID is generated just after board initialization (BSDs for instance) it gives c) which is IMHO a good solution (simple and, stable). Regards [email protected] PS: in conclusion this is "implementation dependent" and there is no (not yet?) a good reason to change this. -------------------------------------------------------------------- IETF IPv6 working group mailing list [email protected] Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
