Re: [Qemu-devel] [PATCH 00/15] qdev: make reset semantics more clear and consistent, reset qbuses under virtio devices

[Michael S. Tsirkin]

On Tue, Dec 18, 2012 at 09:35:02AM +0100, Paolo Bonzini wrote:
> Il 18/12/2012 08:27, Paolo Bonzini ha scritto:
> > Il 17/12/2012 22:43, Michael S. Tsirkin ha scritto:
> >> On Mon, Dec 17, 2012 at 05:24:35PM +0100, Paolo Bonzini wrote:
> >>> After discussion with mst on the topic of resetting virtio devices,
> >>> here is a series that hopefully clarifies the semantics of bus and
> >>> device resets.
> >>>
> >>> After this series, there are two kinds of resets:
> >>
> >> So just to clarify, what I proposed was this
> >> (on top of my type safety patch). Then
> >> all transports can call virtio_config_reset
> >> when appropriate (e.g. when PA is set to 0).
> >>
> >> Signed-off-by: Michael S. Tsirkin <address@hidden>
> >>
> >> diff --git a/hw/virtio.c b/hw/virtio.c
> >> index f40a8c5..e65d7c8 100644
> >> --- a/hw/virtio.c
> >> +++ b/hw/virtio.c
> >> @@ -554,6 +554,14 @@ void virtio_reset(void *opaque)
> >>      }
> >>  }
> >>  
> >> +/* Device-specific reset through virtio config space.
> >> + * Reset virtio config and backend child devices if any.
> >> + */
> >> +void virtio_config_reset(VirtIODevice *vdev)
> >> +{
> >> +    qdev_reset_all(vdev->binding_opaque);
> >> +}
> > 
> > Yes, I had understood.  As I said, this is the wrong direction.
> > Resetting happens from vdev->binding_opaque, it can just do
> > qdev_reset_all(myself).

It can but it's the wrong thing for transport to know about.
Let PCI worry about PCI things. This is not
a transport specific thing so belongs in virtio.c

> ... besides, this only works if the reset callback of
> vdev->binding_opaque remembers to call virtio_reset (in the s390
> bindings, it doesn't and this series fixes it).

That's a separate bug I think.

>  So IMO it is not only
> useless, it is also misleading.
> 
> Paolo