Thanks Benoit! I will try the above mentioned steps.
I am not sure why it works fine with 2Rx and 2Tx queue configuration
GigabitEthernet13/0/0 1 up GigabitEthernet13/0/0
Link speed: 10 Gbps
Ethernet address 00:50:56:9b:f5:c5
VMware VMXNET3
carrier up full duplex mtu 9206
flags: admin-up pmd maybe-multiseg rx-ip4-cksum
Devargs:
rx: queues 2 (max 16), desc 1024 (min 128 max 4096 align 1)
tx: queues 2 (max 8), desc 1024 (min 512 max 4096 align 1)
pci: device 15ad:07b0 subsystem 15ad:07b0 address 0000:13:00.00 numa 0
max rx packet len: 16384
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 tcp-lro
vlan-filter jumbo-frame scatter
rx offload active: ipv4-cksum jumbo-frame scatter
tx offload avail: vlan-insert ipv4-cksum udp-cksum tcp-cksum tcp-tso
multi-segs
tx offload active: multi-segs
rss avail: ipv4-tcp ipv4-udp ipv4 ipv6-tcp ipv6-udp ipv6
rss active: none
tx burst function: vmxnet3_xmit_pkts
rx burst function: vmxnet3_recv_pkts
Thanks,
Chetan Bhasin
On Mon, Jan 13, 2020 at 9:17 PM Benoit Ganne (bganne) <[email protected]>
wrote:
> Hmm,
>
> - I suppose you run VPP as root and not in a container
> - if you use CentOS/RHEL can you check disabling SELinux ('setenforce 0')
> - can you share the output of Linux dmesg and VPP 'show pci'
>
> Best
> ben
>
> > -----Original Message-----
> > From: chetan bhasin <[email protected]>
> > Sent: lundi 13 janvier 2020 15:51
> > To: Benoit Ganne (bganne) <[email protected]>
> > Cc: vpp-dev <[email protected]>
> > Subject: Re: [vpp-dev] (Vpp 19,08)Facing issue with vmxnet3 with 1rx/tx
> > queue
> >
> > Hi Benoit,
> >
> > Thanks for your prompt response.
> >
> > We are migrating from vpp 18.01 to vpp.19.08 , that's why we want least
> > modification in our build system and we want to use DPDK as we were using
> > earlier
> >
> > .
> > DBGvpp# show log
> > 2020/01/13 14:44:42:014 notice dhcp/client plugin initialized
> > 2020/01/13 14:44:42:051 warn dpdk EAL init args: -c 14 -n
> > 4 --in-memory --log-level debug --file-prefix vpp -w 0000:1b:00.0 -w
> > 0000:13:00.0 --master-lcore 4
> > 2020/01/13 14:44:42:603 notice dpdk DPDK drivers found 2
> > ports...
> > 2020/01/13 14:44:42:622 notice dpdk EAL: Detected 6
> lcore(s)
> > 2020/01/13 14:44:42:622 notice dpdk EAL: Detected 1 NUMA
> > nodes
> >
> > 2020/01/13 14:44:42:623 notice dpdk EAL: PCI device
> > 0000:13:00.0 on NUMA socket -1
> > 2020/01/13 14:44:42:623 notice dpdk EAL: Invalid NUMA
> > socket, default to 0
> > 2020/01/13 14:44:42:623 notice dpdk EAL: probe driver:
> > 15ad:7b0 net_vmxnet3
> > 2020/01/13 14:44:42:623 notice dpdk EAL: using IOMMU type
> > 8 (No-IOMMU)
> > 2020/01/13 14:44:42:623 notice dpdk EAL: Ignore mapping IO
> > port bar(3)
> > 2020/01/13 14:44:42:623 notice dpdk EAL: PCI device
> > 0000:1b:00.0 on NUMA socket -1
> > 2020/01/13 14:44:42:623 notice dpdk EAL: Invalid NUMA
> > socket, default to 0
> > 2020/01/13 14:44:42:623 notice dpdk EAL: probe driver:
> > 15ad:7b0 net_vmxnet3
> > 2020/01/13 14:44:42:623 notice dpdk EAL: Ignore mapping IO
> > port bar(3)
> > 2020/01/13 14:45:02:475 err dpdk Interface
> > GigabitEthernet13/0/0 error 1: Operation not permitted
> > 2020/01/13 14:45:02:475 notice dpdk
> > vmxnet3_v4_rss_configure(): Set RSS fields (v4) failed: 1
> > 2020/01/13 14:45:02:475 notice dpdk vmxnet3_dev_start():
> > Failed to configure v4 RSS
> >
> >
> >
> > Thanks,
> > Chetan Bhasin
> >
> > On Mon, Jan 13, 2020 at 7:58 PM Benoit Ganne (bganne) <[email protected]
> > <mailto:[email protected]> > wrote:
> >
> >
> > Hi Chetan,
> >
> > Any reason for not using VPP built-in vmxnet3 driver instead of
> > DPDK? That should give you better performance and would be easier for us
> > to debug. See https://docs.fd.io/vpp/20.01/d2/d1a/vmxnet3_doc.html
> >
> > Otherwise, can you share 'show logging' output?
> >
> > Ben
> >
> > > -----Original Message-----
> > > From: [email protected] <mailto:[email protected]> <vpp-
> > [email protected] <mailto:[email protected]> > On Behalf Of chetan
> bhasin
> > > Sent: lundi 13 janvier 2020 15:20
> > > To: vpp-dev <[email protected] <mailto:[email protected]> >
> > > Subject: [vpp-dev] (Vpp 19,08)Facing issue with vmxnet3 with
> > 1rx/tx queue
> > >
> > > Hello Everyone,
> > >
> > > I am facing an issue while bringing up vpp with less than 2 rx
> and
> > 2 tx
> > > queue. I am using vpp19.08. I have configured pci's under the
> dpdk
> > section
> > > like below -
> > >
> > > 1)
> > > dpdk {
> > > # dpdk-config
> > > dev default {
> > > num-rx-desc 1024
> > > num-rx-queues 1
> > > num-tx-desc 1024
> > > num-tx-queues 1
> > > # vlan-strip-offload off
> > > }
> > > dev 0000:1b:00.0 {
> > > }
> > > dev 0000:13:00.0 {
> > > }
> > > }
> > >
> > > When I bring pci state to up , it is showing error in "show
> > hardware-
> > > interfaces"
> > >
> > > DBGvpp# set interface state GigabitEthernet13/0/0 up DBGvpp#
> > show
> > > hardware-interfaces
> > > Name Idx Link Hardware
> > > GigabitEthernet13/0/0 1 down
> > GigabitEthernet13/0/0
> > > Link speed: 10 Gbps
> > > Ethernet address 00:50:56:9b:f5:c5
> > > VMware VMXNET3
> > > carrier down
> > > flags: admin-up pmd maybe-multiseg rx-ip4-cksum
> > > Devargs:
> > > rx: queues 1 (max 16), desc 1024 (min 128 max 4096 align 1)
> > > tx: queues 1 (max 8), desc 1024 (min 512 max 4096 align 1)
> > > pci: device 15ad:07b0 subsystem 15ad:07b0 address
> > 0000:13:00.00 numa 0
> > > max rx packet len: 16384
> > > promiscuous: unicast off all-multicast off
> > > vlan offload: strip off filter off qinq off
> > > rx offload avail: vlan-strip ipv4-cksum udp-cksum tcp-cksum
> > tcp-lro
> > > vlan-filter jumbo-frame scatter
> > > rx offload active: ipv4-cksum jumbo-frame scatter
> > > tx offload avail: vlan-insert ipv4-cksum udp-cksum tcp-cksum
> > tcp-tso
> > > multi-segs
> > > tx offload active: multi-segs
> > > rss avail: ipv4-tcp ipv4-udp ipv4 ipv6-tcp ipv6-udp
> > ipv6
> > > rss active: none
> > > tx burst function: vmxnet3_xmit_pkts
> > > rx burst function: vmxnet3_recv_pkts
> > > Errors:
> > >
> > > rte_eth_dev_start[port:0, errno:1]: Operation not permitted
> > >
> > > 2) When bring up system without "dev default " section , still
> > facing the
> > > same issue , this time default [Rx-queue is 1 and tx-queue is 2
> > (main
> > > thread + 1 worker)]
> > >
> > > DBGvpp# show hardware-interfaces
> > > Name Idx Link Hardware
> > > GigabitEthernet13/0/0 1 down
> > GigabitEthernet13/0/0
> > > Link speed: 10 Gbps
> > > Ethernet address 00:50:56:9b:f5:c5
> > > VMware VMXNET3
> > > carrier down
> > > flags: admin-up pmd maybe-multiseg rx-ip4-cksum
> > > Devargs:
> > > rx: queues 1 (max 16), desc 1024 (min 128 max 4096 align 1)
> > > tx: queues 2 (max 8), desc 1024 (min 512 max 4096 align 1)
> > > pci: device 15ad:07b0 subsystem 15ad:07b0 address
> > 0000:13:00.00 numa 0
> > > max rx packet len: 16384
> > > promiscuous: unicast off all-multicast off
> > > vlan offload: strip off filter off qinq off
> > > rx offload avail: vlan-strip ipv4-cksum udp-cksum tcp-cksum
> > tcp-lro
> > > vlan-filter jumbo-frame scatter
> > > rx offload active: ipv4-cksum jumbo-frame scatter
> > > tx offload avail: vlan-insert ipv4-cksum udp-cksum tcp-cksum
> > tcp-tso
> > > multi-segs
> > > tx offload active: multi-segs
> > > rss avail: ipv4-tcp ipv4-udp ipv4 ipv6-tcp ipv6-udp
> > ipv6
> > > rss active: none
> > > tx burst function: vmxnet3_xmit_pkts
> > > rx burst function: vmxnet3_recv_pkts
> > > Errors:
> > > rte_eth_dev_start[port:0, errno:1]: Operation not permitted
> > >
> > >
> > > Thanks,
> > > Chetan Bhasin
> > >
> >
> >
>
>
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#15155): https://lists.fd.io/g/vpp-dev/message/15155
Mute This Topic: https://lists.fd.io/mt/69669298/21656
Group Owner: [email protected]
Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-
