On Mon, May 06, 2013 at 09:37:11AM +0200, Paolo Bonzini wrote:
> Il 03/05/2013 03:38, David Gibson ha scritto:
> > This patch adds the necessary support for saving the state of the PAPR VIO
> > virtual SCSI device. This turns out to be trivial, because the generiC
> > SCSI code already quiesces the attached virtual SCSI bus.
> >
> > Signed-off-by: David Gibson <da...@gibson.dropbear.id.au>
> > ---
> > hw/scsi/spapr_vscsi.c | 28 ++++++++++++++++++++++++++++
> > 1 file changed, 28 insertions(+)
> >
> > diff --git a/hw/scsi/spapr_vscsi.c b/hw/scsi/spapr_vscsi.c
> > index 3d322d5..f416871 100644
> > --- a/hw/scsi/spapr_vscsi.c
> > +++ b/hw/scsi/spapr_vscsi.c
> > @@ -954,6 +954,33 @@ static Property spapr_vscsi_properties[] = {
> > DEFINE_PROP_END_OF_LIST(),
> > };
> >
> > +static void spapr_vscsi_pre_save(void *opaque)
> > +{
> > + VSCSIState *s = opaque;
> > + int i;
> > +
> > + /* Can't save active requests, apparently the general SCSI code
> > + * quiesces the queue for us on vmsave */
> > + for (i = 0; i < VSCSI_REQ_LIMIT; i++) {
> > + assert(!s->reqs[i].active);
> > + }
> > +}
>
> This is only true when the rerror and werror options have the values
> "ignore" or "report". See virtio-scsi for an example of how to save the
> requests using the save_request and load_request callbacks in
> SCSIBusInfo.
Ah, bother. Unfortunately the save request is quite a lot more
complicated for vscsi, since we have a lot more private data, and I'm
not sure which bits can be reconstructed from other information. I'll
see what I can come up with.
What guarantees _does_ the scsi layer give about the lifecycle state
of the requests when we savevm?
--
David Gibson | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
http://www.ozlabs.org/~dgibson
signature.asc
Description: Digital signature
