Re: [PATCH v3] KVM: arm64: selftests: Test VNCR translation invalidation

[Itaru Kitayama <itaru.kitayama@fujitsu.com>]

On Mon, Oct 26, 2026 at 08:30:07AM +0000, Marc Zyngier wrote:
> On Mon, 26 Oct 2026 03:10:10 +0000,
> Itaru Kitayama <itaru.kitayama@fujitsu.com> wrote:
> > 
> > Add a standalone KVM selftest that checks VNCR accesses after L1
> > updates its stage-1 mapping and issues a TLBI. This guards against
> > stale VNCR translations, as KVM caches L1's translation of
> > VNCR_EL2 and must drop it when L1 invalidates the mapping in
> > kvm_handle_s1e2_tlbi().
> > 
> > Keep VNCR_EL2 unchanged while remapping its backing page A to B then
> > again to A. At each step, L2 reads and writes TPIDR_EL2, and L1
> > verifies that the accesses reach the selected page while leaving the
> > other page unchanged.
> 
> If you were tracking cross vcpu invalidation and exercising all the
> architecture corner cases (such as permission changes and co), that
> would be helpful.
> 
> But as it is, it doesn't help much as a validation tool. Its scope is
> way too narrow.

Thank you for your comments. I will try to incorporate the suggested
test scenarios.

Thanks,
Itaru.

> 
> 	M.
> 
> -- 
> Without deviation from the norm, progress is not possible.
>