0

So, I've a tap interface that's created by kvm which belongs to a VM being managed via libvirt.

The bridge consists of eth0.123 (here eth0.123 is a vlan tagged interface) & the tap device. 

[root@compute1 ~]# brctl show
bridge name     bridge id               STP enabled     interfaces
brq732eb7f9-16          8000.002590c6438e       no              eth0.123
                                                        tap81474f06-29

When I ping 8.8.8.8 from the guest:

  • I can only sniff ICMP request packets on eth0.123. There are no ICMP reply packets to be seen in the tcpdump.

  • On eth0, I can sniff both ICMP request as well as ICMP reply packets.

I'm trying to understand why is that the case. Shouldn't all traffic pass through eth0.123?

Improve this question
3
  • 1
    It should appear on the tagged interface if your ICMP reply came back tagged. Have you verified your vlan routing?
    – Andrew Domaszek
    Dec 2, 2014 at 14:06
  • @AndrewDomaszek How do I verify that? Is it something that's configured on the switch side?
    – Abhijeet Rastogi
    Dec 2, 2014 at 14:10
  • It'd be on the switch.
    – Andrew Domaszek
    Dec 2, 2014 at 15:20

1 Answer 1

Reset to default
1

No, the bridge registers as the receive handler for the eth0.123 interface, so the traffic is received into the bridge and not the VLAN virtual interface.

Once you're down into the physical eth0 interface, this "hook" is gone, so you see the traffic both ways.

Edit: Linux Bridging paper provided by asker's further research:

Improve this answer
3

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .