On Tue, May 20, 2014 at 07:26:31PM -0300, Flavio Leitner wrote: > On Tue, May 20, 2014 at 02:38:08PM -0700, Ben Pfaff wrote: > > On Fri, Apr 11, 2014 at 06:34:09PM -0300, Flavio Leitner wrote: > > > This patch adds generic IPv4 library code to deal with > > > igmp snooping that is implemented in follow-up patches. > > > > > > Signed-off-by: Cong Wang <amw...@redhat.com> > > > Signed-off-by: Daniel Borkmann <dbork...@redhat.com> > > > Acked-by: Thomas Graf <tg...@redhat.com> > > > Signed-off-by: Flavio Leitner <f...@redhat.com>
... > OK, I can fix those issues and send a v2. I will wait for > feedbacks on the remaining patches to address all them together. That makes sense. > > How much of a common case do you expect multicast lookups to be? Do > > you think it will be worth trying to RCUify this, to avoid taking a > > lock on the read path? > > Enough to worth to do that. There was no RCU when this work has > started and rwlock is more simple to implement and review. rwlock is definitely easier to review. It's fine to get this in before worrying about RCU, by the way; I just wanted you to start thinking about it. > Anyway, it's on my ToDo list, along with adding support for IGMPv3 and IPv6. OK, that answers another question ;-) _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
