Hi,
What’s the conclusion here? What I glean from this long thread is that dot1q 
based IP subinterfaces do not work with virtio - true? If so, what’s the fix?
VPP-507 is asking for setting MAC on IP sub-interface, but that’s a separate 
item, as it’s not mandatory for IP routing operation.

-Maciek

On 10 Mar 2017, at 10:38, Matej Klotton -X (mklotton - PANTHEON TECHNOLOGIES at 
Cisco) <mklot...@cisco.com<mailto:mklot...@cisco.com>> wrote:

Hi John,

So I add a sub-interface to a BD 0, the vlan frame is received, but it is also 
received and processed frames with not local MAC address.
vpp# show interface address
GigabitEthernet0/4/0 (up):
GigabitEthernet0/4/0.10 (up):
GigabitEthernet0/4/0.1 (dn):
  l2 bridge bd_id 0 shg 0
GigabitEthernet0/5/0 (up):
  192.168.1.1/24
GigabitEthernet0/6/0 (dn):
GigabitEthernet0/7/0 (dn):
local0 (dn):

I added neighbor entry to arp table and sent packet
sendp(iface='ens6', x=Ether(src='02:00:00:00:00:02', 
dst='02:00:00:00:00:01')/IP(src='10.0.0.1', dst='192.168.1.2'))


00:25:12:769772: dpdk-input
  GigabitEthernet0/4/0 rx queue 0
  buffer 0x4ddc: current data 0, length 34, free-list 0, totlen-nifb 0, trace 
0x1
  PKT MBUF: port 0, nb_segs 1, pkt_len 34
    buf_len 2176, data_len 34, ol_flags 0x0, data_off 128, phys_addr 0x54133600
    packet_type 0x0
  IP4: 02:00:00:00:00:02 -> 02:00:00:00:00:01
  IP6_HOP_BY_HOP_OPTIONS: 10.0.0.1 -> 192.168.1.2
    tos 0x00, ttl 64, length 20, checksum 0xaf3e
    fragment id 0x0001
00:25:12:769783: ethernet-input
  IP4: 02:00:00:00:00:02 -> 02:00:00:00:00:01
00:25:12:769790: ip4-input
  IP6_HOP_BY_HOP_OPTIONS: 10.0.0.1 -> 192.168.1.2
    tos 0x00, ttl 64, length 20, checksum 0xaf3e
    fragment id 0x0001
00:25:12:769792: ip4-lookup
  fib 0 dpo-idx 2 flow hash: 0x00000000
  IP6_HOP_BY_HOP_OPTIONS: 10.0.0.1 -> 192.168.1.2
    tos 0x00, ttl 64, length 20, checksum 0xaf3e
    fragment id 0x0001
00:25:12:769798: ip4-rewrite
  tx_sw_if_index 2 dpo-idx 2 : ipv4 via 192.168.1.2 GigabitEthernet0/5/0: IP4: 
fa:16:3e:16:91:49 -> 02:11:22:33:44:02 flow hash: 0x00000000
  IP4: fa:16:3e:16:91:49 -> 02:11:22:33:44:02
  IP6_HOP_BY_HOP_OPTIONS: 10.0.0.1 -> 192.168.1.2
    tos 0x00, ttl 63, length 20, checksum 0xb03e
    fragment id 0x0001
00:25:12:769800: GigabitEthernet0/5/0-output
  GigabitEthernet0/5/0
  IP4: fa:16:3e:16:91:49 -> 02:11:22:33:44:02
  IP6_HOP_BY_HOP_OPTIONS: 10.0.0.1 -> 192.168.1.2
    tos 0x00, ttl 63, length 20, checksum 0xb03e
    fragment id 0x0001
00:25:12:769804: GigabitEthernet0/5/0-tx
  GigabitEthernet0/5/0 tx queue 0
  buffer 0x4ddc: current data 0, length 34, free-list 0, totlen-nifb 0, trace 
0x1
  IP4: fa:16:3e:16:91:49 -> 02:11:22:33:44:02
  IP6_HOP_BY_HOP_OPTIONS: 10.0.0.1 -> 192.168.1.2
    tos 0x00, ttl 63, length 20, checksum 0xb03e
    fragment id 0x0001

The L3 interface without an IP address processes and forwards packets. Isn’t it 
a security issue?

I was unable to set a MAC address to sub-if.

On 
http://dpdk.org/doc/guides/nics/virtio.html#features-and-limitations-of-virtio-pmd
It says Features of mac/vlan filter are supported, so shouldn’t be driver 
configured in create_vlan_subif command?
Or could we set vlan filter off/on in startup.conf similarly as with 
vlan-offload?

Thanks,
  Matej.


_______________________________________________
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev
  • Re: [vpp-dev]... Dave Barach (dbarach)
  • Re: [vpp-dev]... Matej Klotton -X (mklotton - PANTHEON TECHNOLOGIES at Cisco)
    • Re: [vpp... Maciek Konstantynowicz (mkonstan)
    • Re: [vpp... Matej Klotton -X (mklotton - PANTHEON TECHNOLOGIES at Cisco)
      • Re: ... John Lo (loj)
        • ... John Lo (loj)
          • ... Matej Klotton -X (mklotton - PANTHEON TECHNOLOGIES at Cisco)
            • ... John Lo (loj)
              • ... Matej Klotton -X (mklotton - PANTHEON TECHNOLOGIES at Cisco)
              • ... Maciek Konstantynowicz (mkonstan)
              • ... Matej Klotton -X (mklotton - PANTHEON TECHNOLOGIES at Cisco)

Reply via email to