Re: [Qemu-devel] [PATCH 5/5] migration: Route more error paths

[Dr. David Alan Gilbert]

* Cornelia Huck (address@hidden) wrote:
> On Tue, 19 Sep 2017 19:00:38 +0100
> "Dr. David Alan Gilbert (git)" <address@hidden> wrote:
> 
> > From: "Dr. David Alan Gilbert" <address@hidden>
> > 
> > vmstate_save is called in a few places, and vmstate_save_state is
> > called in lots of places.
> > 
> > Route error returns from the easier cases back up;  there are lots
> > of more complex cases where there own error paths need fixing.
> > 
> > Signed-off-by: Dr. David Alan Gilbert <address@hidden>
> > ---
> >  hw/display/virtio-gpu.c    |  4 +---
> >  hw/virtio/virtio.c         | 13 +++++++------
> >  include/hw/virtio/virtio.h |  2 +-
> >  migration/vmstate-types.c  | 11 ++++++++---
> >  tests/test-vmstate.c       |  6 +++---
> >  5 files changed, 20 insertions(+), 16 deletions(-)
> > 
> 
> > diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
> > index 464947f76d..860333788b 100644
> > --- a/hw/virtio/virtio.c
> > +++ b/hw/virtio/virtio.c
> > @@ -1899,7 +1899,7 @@ static const VMStateDescription vmstate_virtio = {
> >      }
> >  };
> >  
> > -void virtio_save(VirtIODevice *vdev, QEMUFile *f)
> > +int virtio_save(VirtIODevice *vdev, QEMUFile *f)
> >  {
> >      BusState *qbus = qdev_get_parent_bus(DEVICE(vdev));
> >      VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(qbus);
> 
> Would it make sense to touch up the save_config callback as well? Else
> virtio_save() looks a bit lopsided.
> 
> [For virtio-ccw, the callback can simply pass on any return code from
> vmstate_save_state(). For virtio-pci, we can probably make
> pci_device_save() restore the interrupt state in any case.]

What I'd really love to do with save_config is turn it into a
VMStateDescription* - I think I've figured out how to
do that for ccw and mmio - but the pci version is a much harder
mess of casts and stuff.

Dave

--
Dr. David Alan Gilbert / address@hidden / Manchester, UK