Manuel Bouyer <bouyer%antioche.eu.org@localhost> writes: > On Wed, May 21, 2014 at 09:54:22AM +0900, Masao Uebayashi wrote: >> What is missing to catch these errors by tests using RUMP? > > If rump's network interface uses if_ethersubr.c (looks like it does), > most network test would also check this, I guess (as long as the rump > kernels do have DIAGNOSTIOC). > > For this specific failure, I'm not sure what's needed exactly. > Probably a UDP6 socket, maybe with a special binding (like multicast). I think this issue, and certainly the one I saw before, would need a rump kernel with a multicast-capable interface and something that does a v6 join group and leave group.
Description: PGP signature