On Wed, 2014-05-28 at 10:02 -0400, Vlad Yasevich wrote:
...

This is very similar to br_allow_ingress(), so may be you can
re-factor so that we only have 1 such function...

Yes.. indeed, this looks redundant.
What bothered me is br_allowed_ingress() modifies or drops frames but
I'm not expecting it here.
Maybe we can hand a hint to br_allowed_ingress so that it doesn't modify
the skb?

Thanks,
Toshiaki Makita