On 12 Feb 2026, at 12:07, David Marchand wrote:
> On Thu, 12 Feb 2026 at 11:29, Eelco Chaudron <[email protected]> wrote:
>>> I don't know how common that is. The con to have it is that having many
>>> knobs, that almost nobody uses is an overhead to maintain etc.
>>>
>>> I just thought that such users have to do some change (either this knob or
>>> the "trick"), they can just do that and keep this old (bad) behavior only
>>> for their scope.
>>
>> Wouldn't it be convenient to have a DPDK option to not probe any bus? This
>> would solve the workarounds cleanly.
>
> Well, that's what I'd like to see too.
>
> But we just bumped DPDK and have to wait until the next LTS for such a
> change to go in DPDK.
> So for now, we must live with existing API.
>
>>
>> Also, will we be creating any problems for existing users? i.e. do all
>> currently supported drivers support hotplug?
>
> Let's imagine we are in the future, and we have a way to disable all
> initial probing DPDK.
>
> As far as the PCI drivers RHEL supports downstream, hotplug is
> supported (and indirectly tested via multiple restarts).
> Some drivers (or buses) may still be buggy, and that would be
> something for DPDK to fix.
> But then the knob makes it possible to save the day.
>
> And there may still be some strange usecases that hit an issue (so the
> option to restore the legacy behavior would also help).
>
>
> You can argue that then we should not take the risk to change the
> default behavior, that's what legacy is always about.
> Though I think this behavior is ugly and should have been questionned
> when adding DPDK support in OVS in the first place :-).
I agree with you, but there are users beyond RHEL. I suppose they can object to
the patch here if needed ;)
//Eelco
_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
