On 10/6/26 9:52 AM, Ales Musil wrote: > On Mon, Oct 5, 2026 at 3:12 PM Dumitru Ceara <[email protected]> wrote: > >> On 10/5/26 3:06 PM, Michal Arbet wrote: >>> Hi Dumitru, >>> >> >> Hi Michal, >> >>> You're welcome :)..I was also checking the same stuff...why our (Ultimum) >>> patches don't pass in GitHub and locally they does... >>> >>> It should be fine to merge this soon, so we can send a several larger >>> changes around BFD... >>> >> >> If no other maintainer beats me to it, I can probably merge it tomorrow. >> But I think you can just go ahead and post the other patches you want >> to post, we'll know this IGMP thing is a known test issue and we can >> ignore it. >> >> Regards, >> Dumitru >> >>> Thank you for your support ! >>> >>> Michal Arbet >>> Openstack Engineer >>> >>> Ultimum Technologies >> <https://www.google.com/maps/search/Ultimum+Technologies?entry=gmail&source=g> >> a.s. >>> Na Poříčí 1047/26, 11000 Praha 1 >>> Czech Republic >>> >>> +420 604 228 897 <+420%20604%20228%20897> >>> [email protected] >>> *https://ultimum.io <https://ultimum.io/>* >>> >>> LinkedIn <https://www.linkedin.com/company/ultimum-technologies> | >> Twitter >>> <https://twitter.com/ultimumtech> | Facebook >>> <https://www.facebook.com/ultimumtechnologies/timeline> >>> >>> On Mon, Oct 5, 2026, 2:56 PM Dumitru Ceara <[email protected]> wrote: >>> >>>> On 10/5/26 2:38 PM, Michal Arbet wrote: >>>>> Scapy 2.8.0 moved the IGMPv3 layers to scapy.layers.igmp and kept >>>>> scapy.contrib.igmpv3 only as a deprecated alias. IGMPv3mr is now >>>>> IGMPv3_MR, a complete Membership Report that carries its own type >>>>> and checksum. The "IGMPv3(type=0x22) / IGMPv3mr(...)" packet used by >>>>> the "Traffic from external port not looped back to localnet" test is >>>>> therefore built with two report headers, and the first one has zero >>>>> group records. pinctrl learns no group from it and the test fails >>>>> waiting for the 239.1.2.3 IGMP_Group row. >>>>> >>>>> No scapy construction produces the same packet with both 2.7 and 2.8, >>>>> so build the report with the send_igmp_v3_report() helper that the >>>>> other IGMP tests already use. >>>>> >>>>> Assisted-by: Claude Opus 5.5, Claude Code >>>>> Signed-off-by: Michal Arbet <[email protected]> >>>>> --- >>>> >>>> Hi Michal, >>>> >>>> Thanks a lot for the fix! >>>> >>>> I was looking at this issue in disbelief this morning trying to >>>> understand why it "works on my machine" and not in GitHub CI. :) >>>> >>>> Assuming ovsrobot reports green CI with this patch: >>>> >>>> Acked-by: Dumitru Ceara <[email protected]> >>>> >>>> Regards, >>>> Dumitru >>>> >>>>> tests/ovn.at | 12 +++++------- >>>>> 1 file changed, 5 insertions(+), 7 deletions(-) >>>>> >>>>> diff --git a/tests/ovn.at b/tests/ovn.at >>>>> index 136b825f9..e0796a76c 100644 >>>>> --- a/tests/ovn.at >>>>> +++ b/tests/ovn.at >>>>> @@ -47594,13 +47594,11 @@ mld_packet=$(fmt_pkt \ >>>>> check as hv1 ovs-appctl netdev-dummy/receive br-phys_n1 $mld_packet >>>>> >>>>> # IGMPv3 report. Same scenario but with IPv4 multicast. >>>>> -igmp_packet=$(fmt_pkt \ >>>>> - "Ether(dst='01:00:5e:00:00:16', src='50:54:00:00:00:01') / \ >>>>> - IP(src='10.0.0.1', dst='224.0.0.22', ttl=1, \ >>>>> - options=IPOption_Router_Alert()) / \ >>>>> - IGMPv3(type=0x22) / \ >>>>> - IGMPv3mr(records=IGMPv3gr(rtype=4, maddr='239.1.2.3'))") >>>>> -check as hv1 ovs-appctl netdev-dummy/receive br-phys_n1 $igmp_packet >>>>> +send_igmp_v3_report br-phys_n1 hv1 \ >>>>> + 505400000001 $(ip_to_hex 10 0 0 1) f9f8 \ >>>>> + $(ip_to_hex 239 1 2 3) 04 e8f9 \ >>>>> + igmp_packet.txt >>>>> +igmp_packet=$(cat igmp_packet.txt) >>>>> >>>>> # Wait for pinctrl to fully process the IGMP/MLD reports and sync them >>>>> # to the SB database. This guarantees that pinctrl has also >> re-injected >>>> >>>> >>> >> >> > Thank you Michal and Dumitru, > > I added the missing "Fixes" tag and applied this to main and 26.09. >
Thanks, Ales, for merging this! I think we're also missing it on 26.03 (LTS) so I cherry picked it there too now. Regards, Dumitru > Regards, > Ales > _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
