Hi Mohammed,
Are you sure packets are sent to the correct destination mac address? Looking
at DPDK stats, it reports a lot of 'unicast packets' but only ~1800 'good
packets' which in turn are all delivered to VPP.
The fact that xconnect works also hints toward a mac destination issue, as
xconnect put the interfaces in promiscuous mode.
You can check by setting the interface manually in promiscuous mode and look at
a packet trace:
vpp# set interface promiscuous on HundredGigabitEthernet86/0/0
vpp# clear trace
vpp# trace add dpdk-input 10
vpp# show trace
ben
> -----Original Message-----
> From: vpp-dev@lists.fd.io <vpp-dev@lists.fd.io> On Behalf Of Mohammed
> Alshohayeb
> Sent: mardi 19 mai 2020 03:32
> To: vpp-dev@lists.fd.io
> Subject: [vpp-dev] DPDK packets received by NIC but not delivered to
> engine
>
> I am having an issue where I see packets in the #show hardware-interfaces
> but only a very small fraction is deliver to the vlib engine
>
> Here are the things I've tried
>
>
> * Using different packet generators (pktgen/trex/tcpreplay)
> * Using variety of physical servers
> * All versions running from 19.01 to 20.01
> * Tried multiple NICs (Mellanox ConnectX5) and Chelsio T6 (cxgb)
> * Made sure checksums are ok since some NICs drop bad frames in the
> pmd
>
>
> The vpp.conf is straight forward
>
> unix {
> nodaemon
> log /var/log/vpp/vpp.log
> cli-listen /run/vpp/cli.sock
> interactive
> }
> dpdk {
> dev 0000:86:00.0
> dev 0000:86:00.1
> }
>
> Notes
> - When connecting the two interfaces via xconnect things work well
> - Tried using the macswap plugin and enabling it but exhibited the same
> very slow behaviour
>
> Here is the show interface counters after pushing ~1000000 packets
>
>
>
> vpp# sh hardware-interfaces
>
> Name Idx Link Hardware
>
> HundredGigabitEthernet86/0/0 1 up HundredGigabitEthernet86/0/0
>
> Link speed: 100 Gbps
>
> Ethernet address ec:0d:9a:cd:94:8a
>
> Mellanox ConnectX-4 Family
>
> carrier up full duplex mtu 9206
>
> flags: admin-up pmd maybe-multiseg rx-ip4-cksum
>
> rx: queues 1 (max 65535), desc 1024 (min 0 max 65535 align 1)
>
> tx: queues 1 (max 65535), desc 1024 (min 0 max 65535 align 1)
>
> pci: device 15b3:1017 subsystem 15b3:0007 address 0000:86:00.00 numa 1
>
> module: unknown
>
> max rx packet len: 65536
>
> promiscuous: unicast off all-multicast on
>
> vlan offload: strip off filter off qinq off
>
> rx offload avail: vlan-strip ipv4-cksum udp-cksum tcp-cksum vlan-
> filter
>
> jumbo-frame scatter timestamp keep-crc
>
> rx offload active: ipv4-cksum jumbo-frame scatter
>
> tx offload avail: vlan-insert ipv4-cksum udp-cksum tcp-cksum tcp-tso
>
> outer-ipv4-cksum vxlan-tnl-tso gre-tnl-tso multi-
> segs
>
> udp-tnl-tso ip-tnl-tso
>
> tx offload active: multi-segs
>
> rss avail: ipv4 ipv4-frag ipv4-tcp ipv4-udp ipv4-other ipv6
> ipv6-frag
>
> ipv6-tcp ipv6-udp ipv6-other ipv6-tcp-ex ipv6-udp-
> ex
>
> ipv6-ex ipv6-tcp-ex ipv6-udp-ex
>
> rss active: none
>
> tx burst function: mlx5_tx_burst_vec
>
> rx burst function: mlx5_rx_burst
>
>
>
> rx frames ok 1847
>
> rx bytes ok 465972
>
> extended stats:
>
> rx good packets 1847
>
> rx good bytes 465972
>
> rx q0packets 1847
>
> rx q0bytes 465972
>
> rx port unicast packets 100006
>
> rx port unicast bytes 1034369007
>
> rx port multicast packets 1838
>
> rx port multicast bytes 462894
>
> rx port broadcast packets 9
>
> rx port broadcast bytes 3078
>
> rx packets phy 1000042
>
> rx bytes phy 1038835147
>
>
>
>
>
>
>
> vpp# sh int
>
> Name Idx State MTU (L3/IP4/IP6/MPLS)
> Counter Count
>
> HundredGigabitEthernet86/0/0 1 up 9000/0/0/0 rx
> packets 1856
>
> rx
> bytes 467272
>
> drops
> 1856
>
> ip4
> 1816
>
> ip6
> 26
>
> local0 0 down 0/0/0/0
>
> vpp#
>
> You can see the packets are received in the unicast but for some reason
> they are not being forwarded to further
>
> Is there something obvious I am missing?
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#16450): https://lists.fd.io/g/vpp-dev/message/16450
Mute This Topic: https://lists.fd.io/mt/74316945/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-
