Re: [PATCH v1] s390/virtio_ccw: don't allocate/assign airqs for non-existing queues

[David Hildenbrand]

On 07.04.25 10:34, Michael S. Tsirkin wrote:
> On Mon, Apr 07, 2025 at 10:17:10AM +0200, David Hildenbrand wrote:
> > On 07.04.25 09:52, Michael S. Tsirkin wrote:
> > > On Fri, Apr 04, 2025 at 05:39:10PM +0200, Halil Pasic wrote:
> > > > > Not perfect, but AFAIKS, not horrible.
> > > >
> > > > It is like it is. QEMU does queue exist if the corresponding feature
> > > > is offered by the device, and that is what we have to live with.
> > >
> > > I don't think we can live with this properly though.
> > > It means a guest that does not know about some features
> > > does not know where to find things.
> >
> > Please describe a real scenario, I'm missing the point.
>
>
> OK so.
>
> Device has VIRTIO_BALLOON_F_FREE_PAGE_HINT and VIRTIO_BALLOON_F_REPORTING
> Driver only knows about VIRTIO_BALLOON_F_REPORTING so
> it does not know what does VIRTIO_BALLOON_F_FREE_PAGE_HINT do.
> How does it know which vq to use for reporting?
> It will try to use the free page hint one.

"only knows" -- VIRTIO_BALLOON_F_FREE_PAGE_HINT was proposed + specified 
in the spec.

So I think this is not a very good example.

> > Whoever adds new feat_X *must be aware* about all previous features,
> > otherwise we'd be reusing feature bits and everything falls to pieces.
>
>
> The knowledge is supposed be limited to which feature bit to use.

I think we also have to know which virtqueue bits can be used, right?

I mean, I agree that it's all nasty ...

> > > So now, I am inclined to add linux code to work with current qemu and
> > > with spec compliant one, and add qemu code to work with current linux
> > > and spec compliant one.
> > >
> > > Document the bug in the spec, maybe, in a non conformance section.
> >
> > I'm afraid this results in a lot of churn without really making things
> > better.
>
> > IMHO, documenting things how they actually behave, and maybe moving towards
> > fixed queue indexes for new features is the low hanging fruit.
>
> I worry about how to we ensure that?
> If old code is messed up people will just keep propagating that.
> I would like to fix old code so that new code is correct.
>
> > As raised, it's not just qemu+linux, it's *at least* also cloud-hypervisor.
> >
> > --
> > Cheers,
> >
> > David / dhildenb
>
> There's a slippery slope here in that people will come to us
> with buggy devices and ask to change the spec.

I yet have to fully digest cross-vm: 
https://github.com/google/crosvm/blob/main/devices/src/virtio/balloon.rs

and how it would free-page-reporting work with upstream Linux ... :(

--
Cheers,

David / dhildenb