Re: [PATCH v3 15/23] xen/riscv: introduce (de)initialization helpers for vINTC

[Oleksii Kurochko <oleksii.kurochko@xxxxxxxxx>]

On 6/22/26 5:01 PM, Jan Beulich wrote:
> On 17.06.2026 13:17, Oleksii Kurochko wrote:
> > @@ -83,3 +84,42 @@ int __init make_intc_domU_node(struct kernel_info *kinfo)
> >   
> >       return vintc->init_ops->make_domu_dt_node(kinfo);
> >   }
> > +
> > +/*
> > + * Guests are given a virtual interrupt controller that mirrors the host
> > + * hardware: an AIA-capable host yields a virtual AIA for the guest, and
> > + * so on for any future controller types.
> > + */
> This comment is worded as if it should be like this, when (as indicated before
> at the very least for migration) there can (and eventually should) be
> alternatives.
Agreed. I think we can simply drop the comment. The function name is 
descriptive enough on its own.
> > +int domain_vintc_init(struct domain *d)
> > +{
> > +    int ret = -EOPNOTSUPP;
> > +    const enum intc_version ver = intc_hw_ops->info->hw_version;
> > +
> > +    switch ( ver )
> > +    {
> > +    case INTC_APLIC:
> "version" to me means something different. Do you perhaps mean "variant" or
> "flavor" or some such? (I realize this also affects pre-existing code.)
It is more about a variant than a version.

Would you be okay with me renaming intc_version and ->hw_version in this 
patch, or would you prefer that I do the renaming in a separate 
preparatory patch?
Thanks.

~ Oleksii