On Mon, May 17, 2021 at 10:51:38PM +0100, Michael Brown wrote:
> On 17/05/2021 22:43, Marek Marczykowski-Górecki wrote:
> > In any case, the issue is not calling the hotplug script, responsible
> > for configuring newly created vif interface. Not kernel waiting for it.
> > So, I think both commits should still be reverted.
> 
> Did you also test the ability for a domU to have the netfront driver
> reloaded?  (That should be roughly equivalent to my original test scenario
> comprising the iPXE netfront driver used to boot a kernel that then loads
> the Linux netfront driver.)

Yes, with both commits reverted, it just works. Without the need to do
anything at the backend side.

-- 
Best Regards,
Marek Marczykowski-Górecki
Invisible Things Lab