I think the upstream kernel has a similar issue with tunneling. I'm preparing a patch.
thanks, On Tue, Jun 25, 2013 at 05:11:47PM +0900, Isaku Yamahata wrote: > Reset skb->pkt_type when sending tunneled packets. > Packet can have pkt_type of PACKET_OTHERHOST when coming to ovs bridge. > If such packets are handled by loopback device, the packet is dropped > by ip_rcv(). > On the other hand when flow rule isn't installed in kernel datapath, > OVS_ACTION_ATTR_OUTPUT is used to send the packet. In that case, packet > isn't dropped. > > netns A | root netns | netns B > veth<->veth=ovs bridge=gre<-loop back->gre=ovs bridge=veth<->veth > > arp packet -> > pkt_type > BROADCAST------------>ip_rcv()------------------------> > > <- arp reply > pkt_type > > rule exists > ip_rcv()<-----------------OTHERHOST > drop > > <--------------------ip_rcv()<--- PACKET_HOST rule doesn't exists > pass ^ | > | |upcall > | V > OVS_ACTION_ATTR_OUTPUT > ovs-switchd > > Signed-off-by: Isaku Yamahata <yamah...@valinux.co.jp> > --- > datapath/tunnel.c | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/datapath/tunnel.c b/datapath/tunnel.c > index 8c93e18..1975c7b 100644 > --- a/datapath/tunnel.c > +++ b/datapath/tunnel.c > @@ -275,6 +275,7 @@ int ovs_tnl_send(struct vport *vport, struct sk_buff *skb, > skb_dst_set(skb, dst_clone(&rt_dst(rt))); > else > skb_dst_set(skb, &rt_dst(rt)); > + skb->pkt_type = PACKET_HOST; > > /* Push Tunnel header. */ > build_header(vport, skb, tunnel_hlen); > -- > 1.7.10.4 > > _______________________________________________ > dev mailing list > dev@openvswitch.org > http://openvswitch.org/mailman/listinfo/dev > -- yamahata _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
