[ovs-dev] how a trunk port treats untagged packets
Ben Pfaff
blp at ovn.org
Sat Dec 24 05:17:42 UTC 2016
On Sat, Dec 24, 2016 at 11:50:00AM +0800, 贺斌 wrote:
> hello all,
> I'm a ovs beginer. Recently, I test on ovs trunk port:
> 1. I created a ovs bridge named br0, and then add two trunk ports on it, named if0, if1. And the trunk is the same, 44,55,66.
> 2. Then I put intreface if0 to Linux name space ns0, and if1 to ns1.
> 3. Associate a ip 110.0.0.2/24 on if0 in ns0 and 110.0.0.2/24 on if1 in ns1.
> 4. The I execute 'ping 110.0.0.3' from ns0, But I can't get any packet captured on default intterface br0 or if1 in ns1 using 'tcpdump'. But when I use comman 'ovs-ofctl dump-flows br0', I find that every time after I excute ping commands, there is an increase of 3 for 'n_packets' in the flow. I guest it's arp packets. But I can't capture it from any where.
>
>
> I'm sure all the interface is up, and the route table is corret.
>
> So, I wonder that, when an untagged packet passed to the trunk port directly, how the ovs brigde treat it? Drop it directly or add a default VLAN id then forwarding it?
Untagged packets are in VLAN 0.
More information about the dev
mailing list