> -----Original Message----- > From: Ilya Maximets <i.maxim...@ovn.org> > Sent: Monday 14 March 2022 17:09 > To: Phelan, Michael <michael.phe...@intel.com>; Amber, Kumar > <kumar.am...@intel.com>; ovs-dev <ovs-dev@openvswitch.org> > Cc: i.maxim...@ovn.org; Stokes, Ian <ian.sto...@intel.com>; Ferriter, Cian > <cian.ferri...@intel.com>; Aaron Conole <acon...@redhat.com>; Kalahasthi, > Suneetha <suneetha.kalahas...@intel.com> > Subject: Re: [ovs-build] |fail| pw1600225 [ovs-dev, branch-2.15] dpdk: Use > DPDK 20.11.4 release > > On 3/14/22 17:21, Phelan, Michael wrote: > > > >> -----Original Message----- > >> From: Amber, Kumar <kumar.am...@intel.com> > >> Sent: Monday 14 March 2022 13:59 > >> To: Ilya Maximets <i.maxim...@ovn.org>; Phelan, Michael > >> <michael.phe...@intel.com>; ovs-dev <ovs-dev@openvswitch.org> > >> Cc: Stokes, Ian <ian.sto...@intel.com>; Ferriter, Cian > >> <cian.ferri...@intel.com>; Aaron Conole <acon...@redhat.com>; > >> Kalahasthi, Suneetha <suneetha.kalahas...@intel.com> > >> Subject: RE: [ovs-build] |fail| pw1600225 [ovs-dev, branch-2.15] > >> dpdk: Use DPDK 20.11.4 release > >> > >> Hi Ilya, Michael, > >> > >> Please find my reply inline. > >> > >>> -----Original Message----- > >>> From: Ilya Maximets <i.maxim...@ovn.org> > >>> Sent: Monday, March 14, 2022 5:44 PM > >>> To: Phelan, Michael <michael.phe...@intel.com>; ovs-dev <ovs- > >>> d...@openvswitch.org> > >>> Cc: i.maxim...@ovn.org; Stokes, Ian <ian.sto...@intel.com>; > >>> Ferriter, Cian <cian.ferri...@intel.com>; Aaron Conole > >>> <acon...@redhat.com>; Kalahasthi, Suneetha > >>> <suneetha.kalahas...@intel.com>; Amber, Kumar > >>> <kumar.am...@intel.com> > >>> Subject: Re: [ovs-build] |fail| pw1600225 [ovs-dev, branch-2.15] dpdk: > >>> Use DPDK 20.11.4 release > >>> > >>> On 3/14/22 12:29, Phelan, Michael wrote: > >>>>> -----Original Message----- > >>>>> From: Ilya Maximets <i.maxim...@ovn.org> > >>>>> Sent: Friday 11 March 2022 20:38 > >>>>> To: ovs-dev <ovs-dev@openvswitch.org>; Phelan, Michael > >>>>> <michael.phe...@intel.com> > >>>>> Cc: i.maxim...@ovn.org; Stokes, Ian <ian.sto...@intel.com>; > >>>>> Ferriter, Cian <cian.ferri...@intel.com>; Aaron Conole > >>>>> <acon...@redhat.com>; Kalahasthi, Suneetha > >>>>> <suneetha.kalahas...@intel.com>; Amber, Kumar > >>>>> <kumar.am...@intel.com> > >>>>> Subject: Re: [ovs-build] |fail| pw1600225 [ovs-dev, branch-2.15] > >>>>> dpdk: Use DPDK 20.11.4 release > >>>>> > >>>>>> Test-Label: intel-ovs-compilation > >>>>>> Test-Status: fail > >>>>>> http://patchwork.ozlabs.org/api/patches/1600225/ > >>>>>> > >>>>>> AVX-512_compilation: failed > >>>>>> DPLCS Test: fail > >>>>>> DPIF Test: fail > >>>>>> MFEX Test: fail > >>>>>> Errors in DPCLS test: > >>>>>>> Good hash: 121684652 len: 72 Test hash:0 len:0 > >>>>>>> 2022-03-02T16:05:44.579Z|00034|dpif_netdev_extract(pmd- > >>>>> c21/id:101)|ERR|Autovalidation for avx512_ipv4_udp failed in pkt > >>>>> 15, disabling. > >>>>>>> 2022-03-02T16:05:44.579Z|00035|dpif_netdev_extract(pmd- > >>>>> c21/id:101)|ERR|Autovalidation failure details: > >>>>>>> MFEX autovalidator pkt 15 > >>>>>>> Good hash: 121684652 len: 72 Test hash:0 len:0 > >>>>>>> 2022-03-02T16:05:44.579Z|00036|dpif_netdev_extract(pmd- > >>>>> c21/id:101)|ERR|Autovalidation for avx512_ipv4_udp failed in pkt > >>>>> 19, disabling. > >>>>>>> 2022-03-02T16:05:44.579Z|00037|dpif_netdev_extract(pmd- > >>>>> c21/id:101)|ERR|Autovalidation failure details: > >>>>>>> MFEX autovalidator pkt 19 > >>>>>>> Good hash: 121684652 len: 72 Test hash:0 len:0 > >>>>>>> 2022-03-02T16:05:44.579Z|00038|dpif_netdev_extract(pmd- > >>>>> c21/id:101)|ERR|Autovalidation for avx512_ipv4_udp failed in pkt > >>>>> 21, disabling. > >>>>>>> 2022-03-02T16:05:44.579Z|00039|dpif_netdev_extract(pmd- > >>>>> c21/id:101)|ERR|Autovalidation failure details: > >>>>>>> MFEX autovalidator pkt 21 > >>>>>>> Good hash: 121684652 len: 72 Test hash:0 len:0 > >>>>>>> 2022-03-02T16:05:44.579Z|00040|dpif_netdev_extract(pmd- > >>>>> c21/id:101)|ERR|Autovalidation for avx512_ipv4_udp failed in pkt > >>>>> 25, disabling. > >>>>>>> 2022-03-02T16:05:44.579Z|00041|dpif_netdev_extract(pmd- > >>>>> c21/id:101)|ERR|Autovalidation failure details: > >>>>>>> MFEX autovalidator pkt 25 > >>>>>>> Good hash: 121684652 len: 72 Test hash:0 len:0 > >>>>>>> 2022-03-02T16:05:44.579Z|00042|dpif_netdev_extract(pmd- > >>>>> c21/id:101)|ERR|Autovalidation for avx512_ipv4_udp failed in pkt > >>>>> 27, disabling. > >>>>>>> 2022-03-02T16:05:44.579Z|00043|dpif_netdev_extract(pmd- > >>>>> c21/id:101)|ERR|Autovalidation failure details: > >>>>>>> MFEX autovalidator pkt 27 > >>>>>>> Good hash: 121684652 len: 72 Test hash:0 len:0 > >>>>>>> 2022-03-02T16:05:44.579Z|00044|dpif_netdev_extract(pmd- > >>>>> c21/id:101)|ERR|Autovalidation for avx512_ipv4_udp failed in pkt > >>>>> 31, disabling. > >>>>>>> 2022-03-02T16:05:44.579Z|00045|dpif_netdev_extract(pmd- > >>>>> c21/id:101)|ERR|Autovalidation failure details: > >>>>>>> MFEX autovalidator pkt 31 > >>>>> > >>>>> Hey, Michael, others. > >>>>> > >>>>> Above errors are not part of the current OVS code, nor the patch > >>>>> this report is for. But I see that code in the other patch that > >>>>> has no test report from the > >>>>> intel-ovs-compilation: > >>>>> > >>>>> > >> https://patchwork.ozlabs.org/project/openvswitch/patch/202202251156 > >>>>> 54 > >>>>> .5 > >>>>> 21213-3-kumar.am...@intel.com/ > >>>>> > >>>>> Seems like report was sent to a wrong patch. The patch number at > >>>>> the top of the mail is correct though, so I'm not sure what is going on. > >>>>> > >>>>> Best regards, Ilya Maximets. > >>>> > >>>> Hi Ilya, > >>>> The patch was tagged with branch-2.15 so the CI tested the patch > >>>> with OVS > >>> 2.15 and DPDK 20.11 so perhaps that is why the unusual errors occurred? > >>> > >>> There is no code that could print such messages on any upstream > >>> branch. So, I see only two options: > >>> > >>> - Report sent to a wrong patch > >>> or > >>> - The patch for testing was applied to non-upstream version of OVS. > >>> > >>> Maybe I'm missing something, but I don't see other option on how > >>> the 'Good hash: 121684652 len: 72 Test hash:0 len:0' can be > >>> printed with upstream OVS + DPDK 20.11.4 patch. > >>> > > > > After looking through the build logs I can say for certain that the report > > was > not sent for the wrong patch. Also, the Jenkins job resets the repository > before > applying patches so I don't see how it could've been applied to a non-upstream > version unless I had the wrong branch. Am I correct in saying that the branch > "branch-2.15" is the correct branch for OVS 2.15? > > Yes, "branch-2.15" is a correct branch. > And you can see for yourself that the following grep emits no results: > > git grep 'Good hash:' origin/branch-2.15 > > > > > I did notice that the command used to call the unit tests was incorrect for > > the > older OVS versions but I couldn't recreate the error that was reported in the > mail. The mail reported that the 6th DPDK unit test had failed however with > the > current up to date branch-2.15 there seems to be only 5 tests which are > available. The command used was "make check-dpdk 'TESTSUITEFLAGS=1-6 8'" > however when I try that command now it returns an "invalid test group: 6" > error > as there is no sixth test whereas before it generated the log file containing > the > hash errors in tests/system-dpdk-testsuite.dir/6. > > I guess, that's another evidence that the test was performed with some > incorrect base code. > > I'm not sure if we can get to the bottom of this right now, but let's just > keep an > eye on testing to track the issue down in case it happens again. Thanks for > checking.
Sure Ilya, no problem. I'll keep an eye out for any patches tagged for branch-2.15. Thanks for raising the issue. > > > > >>>> > >>>> Amber's patch was tested and results reported to the build mailing > >>>> list, > >>> however the result was not picked up by patchwork as it was sent > >>> before I had corrected how failures were reported. > >>>> > >>>> Kind regards, > >>>> Michael. > >> > >> This confusion is basically caused because of a patch I sent with > >> IPv6 + Hashing optimizations that’s verifies the validity of hashing as > >> well. > >> The patch next in series adds the hashing functionality so this > >> problem is there. > >> > >> Patch which will add the logs: > >> http://patchwork.ozlabs.org/project/openvswitch/patch/20220225115654. > >> 52 > >> 1213-3-kumar.am...@intel.com/ > > > > Thanks for clarifying Amber, I'm not sure how this could have affected the > testing of Suneetha's patch though as the repository would've been reset after > testing your patch and there was other patches tested between testing your > patch and testing Suneetha's which didn’t encounter any issues. > > > > > >> > >> Hope this clarifies everything. > >> Regards > >> Amber _______________________________________________ dev mailing list d...@openvswitch.org https://mail.openvswitch.org/mailman/listinfo/ovs-dev
