Re: [PATCH v15 15/37] KVM: arm64: CCA: Handle RMI_EXIT_RIPAS_CHANGE

[Suzuki K Poulose]

Hi Fuad

Thanks for your feedback.

On 04/08/2026 13:20, Fuad Tabba wrote:
> Hi Suzuki,
>
> + Quentin (who's on leave, but for when he gets back)
>
> On Mon, 3 Aug 2026 at 16:30, Suzuki K Poulose <suzuki.poulose@xxxxxxx> wrote:
> ...
> > > I.e don't hack into the API. Use the information you already have to
> > > decide whether you need to use the existing API at all.
> > >
> > > If that cannot be done for whatever reason, then you need to indirect
> > > the call with a set of operations that leave the existing code
> > > untouched.
> >
> > Please let us know if you think we need to explore this ^ ?
>
> FWIW, pKVM already has this kind of indirection: in protected mode,
> the kvm_pgtable_stage2_* calls in mmu.c dispatch to the
> pkvm_pgtable_stage2_* wrappers through KVM_PGT_FN(), and the callers
> don't branch. It is global today because pKVM owns the whole host, but
> a per-VM version would let the CCA backend handle the filter and the
> IPA aliasing internally, keeping __unmap_stage2_range()'s signature
> intact. That should work for pKVM as well.

Sure, will explore this option. For the record, the patch here only 
passes down the "filter" parameter to __unmap_stage2_range() [ though
in a different name, only_shared ] and we still handle the aliasing 
inside the CCA code. Let me explore on this front.

Suzuki



> Cheers,
> /fuad