On 5/9/22 10:47, Eelco Chaudron wrote:
>
>
> On 29 Apr 2022, at 19:40, Aaron Conole wrote:
>
>> This reverts commit c645550bb249 ("odp-util: Always report
>> ODP_FIT_TOO_LITTLE for IGMP.")
>>
>> Always forcing a slow path action can result in some over-broad
>> flows which swallow all traffic and force them to userspace, as reported
>> in the thread at
>> https://mail.openvswitch.org/pipermail/ovs-dev/2021-September/387706.html
>> and at
>> https://mail.openvswitch.org/pipermail/ovs-dev/2021-September/387689.html
>>
>> Revert the ODP_FIT_TOO_LITTLE return for IGMP specifically.
>> Additionally, remove the userspace wc mask to prevent revalidator from
>> cycling flows.
>>
>> Fixes: c645550bb249 ("odp-util: Always report ODP_FIT_TOO_LITTLE for IGMP.")
>> Signed-off-by: Aaron Conole <[email protected]>
>> Acked-by: Eelco Chaudron <[email protected]>
>> ---
>> v1->v2: Add unit test for coverage of mcast traffic when not doing snooping
>> v2->v3: Add a unit test for the non-dummy datapaths as well to demonstrate
>> the
>> matching issue.
>> v3->v4: Fix up the unit tests to be consistent when reporting dp flows.
>> Split
>> IGMP unit test into two different tests. Use a multicast eth addr
>> for
>> the FLOOD under NORMAL action case.
>> v4->v5: Address Ilya's comments.
>
> V5 changes look good to me, and tests are passing…
>
> Acked-by: Eelco Chaudron <[email protected]>
>
Applied and backported down to 2.13. On 2.13 and 2.14 backported without
tests, because userspace conntrack is behaving differently there, so tests
do not work.
Best regards, Ilya Maximets.
_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
