On 15 Dec 2025, at 12:58, Toke Høiland-Jørgensen wrote:

> Eelco Chaudron <[email protected]> writes:
>
>> On 11 Dec 2025, at 12:50, Toke Høiland-Jørgensen wrote:
>>
>>> The openvswitch teardown code will immediately call
>>> ovs_netdev_detach_dev() in response to a NETDEV_UNREGISTER notification.
>>> It will then start the dp_notify_work workqueue, which will later end up
>>> calling the vport destroy() callback. This callback takes the RTNL to do
>>> another ovs_netdev_detach_port(), which in this case is unnecessary.
>>> This causes extra pressure on the RTNL, in some cases leading to
>>> "unregister_netdevice: waiting for XX to become free" warnings on
>>> teardown.
>>>
>>> We can straight-forwardly avoid the extra RTNL lock acquisition by
>>> checking the device flags before taking the lock, and skip the locking
>>> altogether if the IFF_OVS_DATAPATH flag has already been unset.
>>>
>>> Fixes: b07c26511e94 ("openvswitch: fix vport-netdev unregister")
>>> Tested-by: Adrian Moreno <[email protected]>
>>> Signed-off-by: Toke Høiland-Jørgensen <[email protected]>
>>
>> Guess the change looks good, but I’m waiting for some feedback from
>> Adrian to see if this change makes sense.
>
> OK.
>
>> Any luck reproducing the issue it’s supposed to fix?
>
> We got a report from the customer that originally reported it (who had
> their own reproducer) that this patch fixes their issue to the point
> where they can now delete ~2000 pods/node without triggering the
> unregister_netdevice warning at all (where before it triggered at around
> ~500 pod deletions). So that's encouraging :)

That’s good news; just wanted to make sure we are not chasing a red herring :)

Acked-by: Eelco Chaudron [email protected]

_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Reply via email to