Re: [PATCH v1 07/27] KVM: arm64: Provide arm64 KVM API for non-native architectures

[Marc Zyngier <maz@kernel.org>]

On Thu, 02 Apr 2026 05:21:03 +0100,
Steffen Eiden <seiden@linux.ibm.com> wrote:

Drive-by comment as I was idly going through this patch.

[...]

> diff --git a/include/kvm/arm64/kvm_host.h b/include/kvm/arm64/kvm_host.h
> new file mode 100644
> index 000000000000..3a434f47497b
> --- /dev/null
> +++ b/include/kvm/arm64/kvm_host.h

[...]

> +static inline bool kvm_supports_32bit_el0(void)
> +{
> +	return false;
> +}
> +

This looks wrong. The original file still has:

#define kvm_supports_32bit_el0()                                \
        (system_supports_32bit_el0() &&                         \
         !static_branch_unlikely(&arm64_mismatched_32bit_el0))

which should not be tampered with. I guess we are simply lucky that
the preprocessor is braindead enough to make this sort of things go
unnoticed, but I'd expect this predicate to be directly provided by
the s390 code one way or another.

Thanks,

	M.

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