Are you sure that kernel is not de-scheduling VPP and running something else on
the same core?
rx-miss happens typically when vpp is not able to service rx queue (either
because it is too busy or because kernel is not giving him opportunity to do
the job).
In this case I bet on second...
> On 29 Nov 2017, at 10:35, Abhilash Lakshmeshwar <l.s.abhil...@gmail.com>
> wrote:
>
> Hello,
>
> I am using Intel X710/XL710 (10G) , running with VPP 17.07 with line rate of
> 1G 1500 bytes of packet size,using 1G huge page. i am observing Rx misses on
> the interfaces.
>
> Initially i tried with default configuration with single core with uio_driver
> as igb_uio and uio_pci_generic. I even tried with couple of CPU parameters
> like skip-cores, increasing Rx queues for interface, but there is no
> improvement.
>
> vpp# show int TenGigabitEthernet1/0/0
> Name Idx State Counter
> Count
> TenGigabitEthernet1/0/0 1 up rx packets
> 1100478
> rx
> bytes 1646315088
>
> rx-miss 2842
>
> below is my startup config.
>
> unix {
> nodaemon
> log /tmp/vpp.log
> full-coredump
> }
>
> api-trace {
> on
> }
>
> api-segment {
> gid vpp
> }
>
> dpdk {
> dev 0000:01:00.0
> dev 0000:01:00.1
> dev 0000:01:00.2
> dev 0000:01:00.3
>
> uio-driver uio_pci_generic
> num-mbufs 16384
> }
>
> heapsize 3G
>
> Could you please let me know if i am missing anything ?
>
> Thanks,
> Abhilash
> _______________________________________________
> vpp-dev mailing list
> vpp-dev@lists.fd.io
> https://lists.fd.io/mailman/listinfo/vpp-dev
_______________________________________________
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev
