<!DOCTYPE html>
<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    <div class="moz-cite-prefix">Hi Adam,<br>
      <br>
      On 11/16/23 23:03, Adam Hassick wrote:<br>
    </div>
    <blockquote type="cite"
cite="mid:CAC-YWqjqOoRc51ksTGh3+yYHGYfBQWSmrs4E-ES4etDVKS3jyg@mail.gmail.com">
      <meta http-equiv="content-type" content="text/html; charset=UTF-8">
      <div dir="ltr">
        <div>Hi Andrew,<br>
          <br>
        </div>
        <blockquote class="gmail_quote"
style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">If
          you use copy of dpdk-ethdev-ts has
          398e272495143884274f5a53c6fe0cc16df41052, you don't need to
          pass --trc-tag=<span>pci-8086-1572 any more since
            corresponding changeset updates expectations to have the
            same for </span><span>pci-8086-1583.</span></blockquote>
        <div><br>
        </div>
        <div>I'll try this for the next run.<br>
          <br>
          <blockquote class="gmail_quote"
style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Sorry,
            but I've failed to find what's wrong there.</blockquote>
          <div><br>
          </div>
          <div>That if statement works if using the traditional
            single-bracket conditional, or it needs to be rewritten as
            "[[ -z "${test_log}" ]] || [[ ! -r "${test_log}" ]]". The
            latter is the change I made, but both work.<br>
          </div>
        </div>
      </div>
    </blockquote>
    <br>
    Thanks a lot. Hopefully fixed.<br>
    <br>
    <blockquote type="cite"
cite="mid:CAC-YWqjqOoRc51ksTGh3+yYHGYfBQWSmrs4E-ES4etDVKS3jyg@mail.gmail.com">
      <div dir="ltr">
        <div>
          <div><br>
            <blockquote class="gmail_quote"
style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">As
              far as I can see LLDP packets spoil testing results:<br>
              <a
href="https://ts-factory.io/bublik/v2/log/362398?focusId=362760&mode=treeAndinfoAndlog&experimental=true&lineNumber=1_63"
                target="_blank" moz-do-not-send="true">https://ts-factory.io/bublik/v2/log/362398?focusId=362760&mode=treeAndinfoAndlog&experimental=true&lineNumber=1_63</a><br>
              <br>
              As far as I can see main prologue disables FW LLDP on
              Tester<br>
              <a
href="https://ts-factory.io/bublik/v2/log/362398?focusId=362400&mode=treeAndinfoAndlog&experimental=true&lineNumber=1_80"
                target="_blank" moz-do-not-send="true">https://ts-factory.io/bublik/v2/log/362398?focusId=362400&mode=treeAndinfoAndlog&experimental=true&lineNumber=1_80</a><br>
              but I guess it could be still enabled on DUT side and DPDK
              do not provide means to disable it as far as I know. I
              vaguely remember that Intel provides FW configuration
              tools which can do it.<br>
              It is interesting since DPDK gets unexpected LLDP packets
              but may be packets sent by FW go via loopback and visible
              to PF as well.<br>
              Other possible source of LLDP packet is a switch if NICs
              are connected via switch. If so, LLDP should be disabled
              on corresponding switch ports.<br>
              <br>
              As far as I can see fixing the problem should make results
              much closer. However, I already see some differences in
              behaviour which should be simply fixed in TRC. For
              example, X710 gets 9 packets less than configuration
              number of Rx descriptors, but XL710 gets 10 packets less.</blockquote>
            <div><br>
            </div>
            <div>I have the "disable-fw-lldp" private flag set on both
              of the XL710 ports on the DUT machine. Very strange how
              there are still LLDP packets appearing in there.<br>
            </div>
          </div>
        </div>
      </div>
    </blockquote>
    <br>
    Me too. Corresponding packet has source MAC from Peer/Tester machine
    NIC.<br>
    It is really strange since prologue disabled LLDP there as well.
    I'll try to play with it locally more, but have no good ideas in
    fact.<br>
    <br>
    <blockquote type="cite"
cite="mid:CAC-YWqjqOoRc51ksTGh3+yYHGYfBQWSmrs4E-ES4etDVKS3jyg@mail.gmail.com">
      <div dir="ltr">
        <div>
          <div>
            <div>These systems are not connected to any switch, so maybe
              a service on the DUT itself is sending them. I'm not sure
              how that could be happening though, because I don't have
              the LLDP daemon installed on either system.<br>
              <br>
              <blockquote class="gmail_quote"
style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Also
                I see that performance tests are not run because of
                failed prologue:<br>
                <a
href="https://ts-factory.io/bublik/v2/log/362398?focusId=369564&mode=treeAndinfoAndlog&experimental=true"
                  target="_blank" moz-do-not-send="true">https://ts-factory.io/bublik/v2/log/362398?focusId=369564&mode=treeAndinfoAndlog&experimental=true</a><br>
                I'll investigate it, but I guess the source of
                difference is that we always run tests on single
                interface. Just add -p0 (--cfg=<span>iol-dts-xl710-p0</span>)
                to your configuration name. You don't need to change
                ts-rigs for it since the suffix is handled by generic
                code. It simply comments the second instance and forces
                take the first interface only into account. Initially it
                was introduced to run independent tests on different
                ports to be able to share configuration, but I guess
                right now it has limitations for some packages like
                representors which require entire NIC.</blockquote>
              <div><br>
              </div>
              <div>I can try that and will see if it works. <br>
              </div>
            </div>
          </div>
        </div>
      </div>
    </blockquote>
    <br>
    This problem is fixed in fresh TE and dpdk-ethdev-ts published on
    GitHub.<br>
    <br>
    Regards,<br>
    Andrew.<br>
    <br>
    <blockquote type="cite"
cite="mid:CAC-YWqjqOoRc51ksTGh3+yYHGYfBQWSmrs4E-ES4etDVKS3jyg@mail.gmail.com">
      <div dir="ltr">
        <div>
          <div>
            <div>
              <div><br>
              </div>
              <div>Thanks,<br>
              </div>
              <div>Adam<br>
              </div>
            </div>
          </div>
        </div>
      </div>
      <br>
      <div class="gmail_quote">
        <div dir="ltr" class="gmail_attr">On Wed, Nov 8, 2023 at 2:20 AM
          Andrew Rybchenko <<a
            href="mailto:andrew.rybchenko@oktetlabs.ru" target="_blank"
            moz-do-not-send="true" class="moz-txt-link-freetext">andrew.rybchenko@oktetlabs.ru</a>>
          wrote:<br>
        </div>
        <blockquote class="gmail_quote"
style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
          <div>
            <div>Hi Adam,<br>
              <br>
              On 11/7/23 23:30, Adam Hassick wrote:<br>
            </div>
            <blockquote type="cite">
              <div dir="ltr">
                <div>
                  <div>
                    <div>
                      <div>
                        <div>Hi Andrew,<br>
                          <br>
                        </div>
                        The runner machine was missing a dependency for
                        one of the scripts, "pixz". After installing
                        that, it appears to have worked. I can see the
                        results listed on the ts-factory Bublik
                        instance.<br>
                      </div>
                    </div>
                  </div>
                </div>
              </div>
            </blockquote>
            <br>
            If you use copy of dpdk-ethdev-ts has
            398e272495143884274f5a53c6fe0cc16df41052, you don't need to
            pass --trc-tag=<span>pci-8086-1572 any more since
              corresponding changeset updates expectations to have the
              same for </span><span>pci-8086-1583.</span><br>
            <br>
            <blockquote type="cite">
              <div dir="ltr">
                <div>
                  <div>
                    <div>In the latest revision of ts-rigs, there
                      appears to be a syntax error at line 42 within the
                      script located at
                      "ts-rigs/scripts/publish_logs/prj/ts-factory/publish",
                      within the if condition. I fixed it locally to get
                      it to run.<br>
                    </div>
                  </div>
                </div>
              </div>
            </blockquote>
            <br>
            Sorry, but I've failed to find what's wrong there.<br>
            <br>
            <blockquote type="cite">
              <div dir="ltr">
                <div>
                  <div>Taking a quick look at a comparison against your
                    most recent X710 run, it looks like we're NOK on
                    around ~400 more test cases. By percentage of tests,
                    we're 1% off, however, it looks like whole subsets
                    of the test suite that contain low numbers of tests
                    are failing. I wonder if this is due to differences
                    between the Intel X710 and XL710 or issues in our
                    dev testbed.<br>
                  </div>
                </div>
              </div>
            </blockquote>
            <br>
            As far as I can see LLDP packets spoil testing results:<br>
            <a
href="https://ts-factory.io/bublik/v2/log/362398?focusId=362760&mode=treeAndinfoAndlog&experimental=true&lineNumber=1_63"
              target="_blank" moz-do-not-send="true">https://ts-factory.io/bublik/v2/log/362398?focusId=362760&mode=treeAndinfoAndlog&experimental=true&lineNumber=1_63</a><br>
            <br>
            As far as I can see main prologue disables FW LLDP on Tester<br>
            <a
href="https://ts-factory.io/bublik/v2/log/362398?focusId=362400&mode=treeAndinfoAndlog&experimental=true&lineNumber=1_80"
              target="_blank" moz-do-not-send="true">https://ts-factory.io/bublik/v2/log/362398?focusId=362400&mode=treeAndinfoAndlog&experimental=true&lineNumber=1_80</a><br>
            but I guess it could be still enabled on DUT side and DPDK
            do not provide means to disable it as far as I know. I
            vaguely remember that Intel provides FW configuration tools
            which can do it.<br>
            It is interesting since DPDK gets unexpected LLDP packets
            but may be packets sent by FW go via loopback and visible to
            PF as well.<br>
            Other possible source of LLDP packet is a switch if NICs are
            connected via switch. If so, LLDP should be disabled on
            corresponding switch ports.<br>
            <br>
            As far as I can see fixing the problem should make results
            much closer. However, I already see some differences in
            behaviour which should be simply fixed in TRC. For example,
            X710 gets 9 packets less than configuration number of Rx
            descriptors, but XL710 gets 10 packets less.<br>
            <br>
            Also I see that performance tests are not run because of
            failed prologue:<br>
            <a
href="https://ts-factory.io/bublik/v2/log/362398?focusId=369564&mode=treeAndinfoAndlog&experimental=true"
              target="_blank" moz-do-not-send="true">https://ts-factory.io/bublik/v2/log/362398?focusId=369564&mode=treeAndinfoAndlog&experimental=true</a><br>
            I'll investigate it, but I guess the source of difference is
            that we always run tests on single interface. Just add -p0
            (--cfg=<span>iol-dts-xl710-p0</span>) to your configuration
            name. You don't need to change ts-rigs for it since the
            suffix is handled by generic code. It simply comments the
            second instance and forces take the first interface only
            into account. Initially it was introduced to run independent
            tests on different ports to be able to share configuration,
            but I guess right now it has limitations for some packages
            like representors which require entire NIC.<br>
            <br>
            Regards,<br>
            Andrew.<br>
            <br>
            <blockquote type="cite">
              <div dir="ltr">
                <div>Thanks,<br>
                </div>
                <div>Adam</div>
              </div>
            </blockquote>
            <br>
            (dropped history, to keep mail size small) </div>
        </blockquote>
      </div>
    </blockquote>
    <br>
  </body>
</html>