Re: attaching device to PV guest broken by your rework of libxl's PCI handling?

[Jason Andryuk <jandryuk@xxxxxxxxx>]

On Thu, Dec 9, 2021 at 8:08 AM Durrant, Paul <xadimgnik@xxxxxxxxx> wrote:
>
> On 09/12/2021 04:58, Jason Andryuk wrote:
> >
> > My attempt at a fix was this:
> > https://lore.kernel.org/xen-devel/20210812005700.3159-1-jandryuk@xxxxxxxxx/
> >
> > It was in terms of PCI & stubdom startup , but that is the same as PV
> > hotplug.  There were questions about further re-work which went
> > unanswered, but my patch at least restores the old working state.
> >
>
> I guess the question is do we need to wait for 'connected' in any case
> other than initial creation?

It's the other way around.  When running (not initial creation -
!starting) and there are already entries (num_devs != NULL), we have
to wait for Connected to avoid interfering with Reconfiguring.

Regards,
Jason