Re: [PATCH v2 47/65] target/hppa: Remove TARGET_REGISTER_BITS

[Richard Henderson]

On 10/20/23 14:31, Philippe Mathieu-Daudé wrote:
> > diff --git a/target/hppa/machine.c b/target/hppa/machine.c
> > index 0c0bba68c0..ab34b72910 100644
> > --- a/target/hppa/machine.c
> > +++ b/target/hppa/machine.c
> > @@ -21,21 +21,12 @@
> >   #include "cpu.h"
> >   #include "migration/cpu.h"
> > -#if TARGET_REGISTER_BITS == 64
> >   #define qemu_put_betr   qemu_put_be64
> >   #define qemu_get_betr   qemu_get_be64
> >   #define VMSTATE_UINTTR_V(_f, _s, _v) \
> >       VMSTATE_UINT64_V(_f, _s, _v)
> >   #define VMSTATE_UINTTR_ARRAY_V(_f, _s, _n, _v) \
> >       VMSTATE_UINT64_ARRAY_V(_f, _s, _n, _v)
>
> Total 6 uses, let's use in place, removing the
> definitions.

I had meant to go back and remove these, thanks for the reminder.
I didn't realize there were so few uses.  I will just fold this in.

> IIUC for TARGET_REGISTER_BITS == 32 we need:
>
> -- >8 --
>   static const VMStateDescription vmstate_env = {
>       .name = "env",
> -    .version_id = 1,
> -    .minimum_version_id = 1,
> +    .version_id = 2,
> +    .minimum_version_id = 2,
>       .fields = vmstate_env_fields,
>   };

You're right -- a version bump is required.  I will simply do this unconditionally, as the 
effort of back-compat is not warranted for this target.

I also need to handle the TLB format change for pa2.0.
(There's an existing comment about that!)


r~