Re: [Qemu-ppc] [RFC 8/9] spapr: Improve setting of default machine version

[Thomas Huth]

On 30/11/15 09:51, David Gibson wrote:
> This tweaks the way the default machine version is controlled, so that
> there will be a bit less churn when each new version is introduced.
> 
> Signed-off-by: David Gibson <address@hidden>
> ---
>  hw/ppc/spapr.c | 20 ++++++++++----------
>  1 file changed, 10 insertions(+), 10 deletions(-)
> 
> diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
> index 1a4fe30..3d31fe7 100644
> --- a/hw/ppc/spapr.c
> +++ b/hw/ppc/spapr.c
> @@ -2301,12 +2301,16 @@ static const TypeInfo spapr_machine_info = {
>      },
>  };
>  
> -#define DEFINE_SPAPR_MACHINE(suffix, verstr, instance_compat)        \
> +#define DEFINE_SPAPR_MACHINE(suffix, verstr, instance_compat, dflt)  \
>      static void spapr_machine_##suffix##_class_init(ObjectClass *oc, \
>                                                      void *data)      \
>      {                                                                \
>          MachineClass *mc = MACHINE_CLASS(oc);                        \
>          spapr_machine_##suffix##_class_compat(mc);                   \
> +        if (dflt) {                                               \
> +            mc->alias = "pseries";                                   \
> +            mc->is_default = 1;                                      \
> +        }                                                            \
>      }                                                                \
>      static void spapr_machine_##suffix##_instance_init(Object *obj)  \
>      {                                                                \
> @@ -2338,11 +2342,9 @@ static void 
> spapr_machine_2_5_class_compat(MachineClass *mc)
>  
>      smc->dr_lmb_enabled = true;
>      mc->desc = "pSeries Logical Partition (PAPR compliant)";
> -    mc->alias = "pseries";
> -    mc->is_default = 1;
>  }
>  
> -DEFINE_SPAPR_MACHINE(2_5, "2.5", NULL);
> +DEFINE_SPAPR_MACHINE(2_5, "2.5", NULL, true);

Good idea ... but I wonder whether we could even do it without the
additional macro parameter? For example, we could
 #define SPAPR_DEFAULT_VERSTR "2.5"
and then in the macro, use a
 if (!strcmp(verstr, SPAPR_DEFAULT_VERSTR))
instead of the "if (dflt)".
... or is this too much macro magic with side effects already?

 Thomas