On 12/21/2011 06:35 AM, Paolo Bonzini wrote:
On 12/20/2011 09:56 PM, Anthony Liguori wrote:
As always, you can implement that in many ways. However, I think the
point of using Visitors is not to remove QEMUFile.
Yes, it is.
The point of using Visitors is to provide a standard representation of device
state. QEMUFile is but one consumer of that representation, along with any other
migration filter.
Can you do a quick code mock up of how you'd expect this to work? I'm not sure
if I'm just being dense, but I'm having a really hard time understanding what
you're suggesting. How I see this evolving with Mike's series is:
struct DeviceStateClass {
ObjectClass parent;
void (*load_state)(DeviceState *obj, Visitor *v, const char *name,
Error **errp);
void (*save_state)(DeviceState *obj, Visitor *v, const char *name,
Error **errp);
};
The PCIDevice base class would expose:
/* protected */
void pci_load_state(PCIDevice *obj, Visitor *v, const char *name,
Error **errp)
{
visit_start_struct(v, "PCIDevice", name, errp);
visit_type_int8(v, &obj->reg[0], "reg[0]", errp);
..
visit_end_struct(v, errp);
}
Or:
static VMStateDescription pci_device_desc = {
.name = "PCIDevice",
.fields = {
VMSTATE_UINT8(PCIDevice, reg[0]),
...
}
};
void pci_load_state(PCIDevice *obj, Visitor *v, const char *name,
Error **errp)
{
visit_with_vmstate(v, obj, &pci_device_desc, name, errp);
}
A subclass would do:
static void my_device_load_state(DeviceState *obj, Visitor *v, const char *name,
Error **errp)
{
MyDevice *s = MY_DEVICE(obj);
visit_start_struct(v, "MyDevice", name, errp);
pci_load_state(PCI_DEVICE(obj), v, "super", errp);
visit_end_struct(v, errp);
}
static void my_device_class_init(ObjectClass *klass, void *data)
{
DeviceClass *dc = DEVICE_CLASS(klass);
dc->load_state = my_device_load_state;
...
}
There's no reference at all to QEMUFile. The load_state/save_state methods can
be exposed as a "state" property too.
Once the series 2/4 lands and Mike's series, implementing this should be very
straight forward.
Regards,
Anthony Liguori
