DPDK usage discussions
 help / color / mirror / Atom feed
From: Suchetha p <suchetha2005@gmail.com>
To: users@dpdk.org
Subject: [dpdk-users] Fragmentation ping test failing for large packets - DPDK 18.08
Date: Fri, 3 Jul 2020 08:13:39 +0530	[thread overview]
Message-ID: <CAOGjjAMzYGU4vOe4Fa0hCL-sXk+N8cXaZ+no46QjkH3ijUeAtg@mail.gmail.com> (raw)

Hi,




Our KVM based VMs on HP Gen10 Servers are running on RHEL 7.6 and having
Intel X710 NIC's.
The VMs are also running on RHEL 7.6 and the DPDK version used in the guest
application is 18.08.
Intel DPDK VF driver vfio-pci is compiled and loaded currently in our VM.





When performing the below fragmentation test it’s working only for max
packet size 1258. For 1500 bytes test, ping is failing in our DPDK based
application.



   1.

   Start ping with 1200bytes--successful
   2.

   Start ping with 1250bytes-successful
   3.

   Start ping with 1300bytes - not successful
   4.

   Start ping with 1500bytes---not successful


   In a successful case (using 1250 bytes ping),  we can see that an
   unfragmented packet entered in gateway –( DPDK based application)
   and gateway forwarding those packets towards destination. In response we
   can see the destination responded with ping reply and reply entering into
   the gateway.



   But in an unsuccessful case (using 1500 bytes ping), we can see 2
   fragmented packets (T-PDU 1284 and 48) entered in the gateway. But DPDK
   based application is forwarding only one packet (T-PDU 48) towards the
   destination and not forwarding other packets. Hence there is no response
   for ping requests. And the ping test is failing. Seems our application is
   dropping fragmented ping echo request packet internally as I didn’t see any
   dropped packet count



   This seems to be an issue with  fragmentation/re-assembly.

   As per the wireshark trace I am getting the below error " IPV4 total
   length exceeds packet length".



   Please find the ethtool settings on the Host machine.



   tcp-segmentation-offload: on
           tx-tcp-segmentation: on
           tx-tcp-ecn-segmentation: on
           tx-tcp6-segmentation: on
           tx-tcp-mangleid-segmentation: off
   udp-fragmentation-offload: off [fixed]
   generic-segmentation-offload: on
   generic-receive-offload: on



   I see that udp-fragmentation offload is off.
   Please let me know if I need to turn on any segmentation parameters with
   respect to ethtool settings to make the fragmentation ping test successful.



   Thanks,

   Suchetha

                 reply	other threads:[~2020-07-03  2:43 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=CAOGjjAMzYGU4vOe4Fa0hCL-sXk+N8cXaZ+no46QjkH3ijUeAtg@mail.gmail.com \
    --to=suchetha2005@gmail.com \
    --cc=users@dpdk.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body. 
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).