On Sun, Jun 24, 2012 at 1:49 AM, 闫兴安 <yanxingan...@126.com> wrote:
> I am reading the gre tunnel related codes in ovs.
> The fuction "ovs_tnl_send" encapsulates packets coming from vports, then
> sends the packets out from a NIC based on a host route entry.
> I was puzzled on the paragraph:
> ********
> cache_vport = ovs_internal_dev_get_vport(rt_dst(rt).dev);
> skb->protocol = htons(ETH_P_IP);
> iph = ip_hdr(skb);
> iph->tot_len = htons(skb->len - skb_network_offset(skb));
> ip_send_check(iph);
> if (cache_vport) {
> if (unlikely(compute_ip_summed(skb, true))) {
> kfree_skb(skb);
> goto next;
> }
> OVS_CB(skb)->flow = cache->flow;
> ovs_vport_receive(cache_vport, skb);--->send to datapath again?
> sent_len += orig_len;
> } else {------>send to NIC
> int xmit_err;
> skb->dev = rt_dst(rt).dev;
> xmit_err = dev_queue_xmit(skb);
> if (likely(net_xmit_eval(xmit_err) == 0))
> sent_len += orig_len;
> }
> **********
> The question is when the rt_dst(rt).dev is a internal dev? Why sending the
> packet to the datapath again?
If your physical interface is itself bridged then the routing table
may point to an OVS device.
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
