qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PATCH V3 8/8] hw/mdio: Use bitbang core for smc91c111


From: Grant Likely
Subject: Re: [Qemu-devel] [PATCH V3 8/8] hw/mdio: Use bitbang core for smc91c111 network device
Date: Sun, 03 Feb 2013 21:30:54 +0000

On Sat, 2 Feb 2013 23:51:42 +0000, Peter Maydell <address@hidden> wrote:
> On 2 February 2013 23:40, Grant Likely <address@hidden> wrote:
> >  static const VMStateDescription vmstate_smc91c111 = {
> > @@ -71,6 +76,8 @@ static const VMStateDescription vmstate_smc91c111 = {
> >          VMSTATE_BUFFER_UNSAFE(data, smc91c111_state, 0, NUM_PACKETS * 
> > 2048),
> >          VMSTATE_UINT8(int_level, smc91c111_state),
> >          VMSTATE_UINT8(int_mask, smc91c111_state),
> > +        VMSTATE_MDIO(mdio_bus, smc91c111_state),
> > +        VMSTATE_MDIO_PHY(phy, smc91c111_state),
> >          VMSTATE_END_OF_LIST()
> >      }
> 
> If you're adding vmstate fields to an existing structure
> you need to either:
>  (a) increment the .version_id field, and set .minimum_version_id
> field to the same value [and accept that old-to-new migration
> won't be possible, which is OK in this particular case as it's
> only used by ARM boards and I'm happy that we don't currently
> support cross version migration on ARM]
>  (b) increment .version_id only, mark the new vmstate fields
> as only-from-version-N, and cope with what you get if an
> incoming migration hasn't got the fields.
> 
> For a complicated thing like the phy I would suggest course
> (a) as the simplest approach.

Okay, I'll do that and submit a fixed up patch. Thanks.

g.




reply via email to

[Prev in Thread] Current Thread [Next in Thread]