Re: [PATCH v21 10/23] KVM: arm64: Add VM specific callback for S2 MMU operations

[Suzuki K Poulose]

On 03/10/2026 10:01, Marc Zyngier wrote:
> On Thu, 01 Oct 2026 22:06:50 +0100,
> Suzuki K Poulose <suzuki.poulose@xxxxxxx> wrote:
> > Add VM type specific S2 MMU operation backends which can be initialized per
> > VM flavor, to keep the handling cleaner.
> >
> > Reviewed-by: Jonathan Cameron <jonathan.cameron@xxxxxxxxxxxxxxxx>
> > Tested-by: Gavin Shan <gshan@xxxxxxxxxx>
> > Signed-off-by: Suzuki K Poulose <suzuki.poulose@xxxxxxx>
> > ---
> > Changes since v19:
> >   - Add a blank line in kvm_arch_flush_remote_tlbs()


> >   struct kvm_s2_mmu {
> >   	struct kvm_vmid vmid;
> >   
> > @@ -332,6 +345,8 @@ struct kvm_arch {
> >   	 */
> >   	u64 fgu[__NR_FGT_GROUP_IDS__];
> >   
> > +	const struct kvm_vm_s2_ops *vm_s2_ops;
> > +
>
> If you're touching this patch, can you please move this pointer next
> to the vcpu_ops pointer?

Sure, I will do that.

> > +
> > +static const struct kvm_vm_s2_ops protected_pkvm_vm_s2_ops = {
> > +	.vm_flush_remote_tlbs		= pkvm_flush_remote_tlbs,
> > +	.vm_flush_remote_tlbs_range	= pkvm_flush_remote_tlbs_range,
> > +	/*
> > +	 * Not supported for Protected VMs under pKVM
> > +	 * .vm_age_gfn
> > +	 * .vm_test_age_gfn
> > +	 * .vm_stage2_unmap_range
> > +	 */
>
> I really think we should have *something* here that returns
> "unsupported", and avoid NULL-checks in the dispatchers.

Ack

> > +};
> > +
> > +static const struct kvm_vm_s2_ops pkvm_vm_s2_ops = {
> > +	.vm_flush_remote_tlbs		= pkvm_flush_remote_tlbs,
> > +	.vm_flush_remote_tlbs_range	= pkvm_flush_remote_tlbs_range,
> > +	.vm_age_gfn			= kvm_vm_age_gfn,
> > +	.vm_test_age_gfn		= kvm_vm_test_age_gfn,
> > +	.vm_stage2_unmap_range		= kvm_vm_stage2_unmap_range,
>
> and since we have this: can we get rid of the KVM_PGT_FN() hack?

Ack, I will give it a go.


> > +#define KVM_VM_S2_OPS(flavor, ops)		\
> > +		[flavor] = ops
> > +static const struct kvm_vm_s2_ops *arm64_vm_s2_ops[] = {
> > +	KVM_VM_S2_OPS(VM_VHE, &kvm_default_vm_s2_ops),
> > +	KVM_VM_S2_OPS(VM_NVHE, &kvm_default_vm_s2_ops),
> > +	KVM_VM_S2_OPS(VM_PKVM, &pkvm_vm_s2_ops),
> > +	KVM_VM_S2_OPS(VM_PROTECTED_PKVM, &protected_pkvm_vm_s2_ops),
>
> nit: move the '&' into the macro.

Ack.

Suzuki