[PATCH v13 13/48] kvm: arm64: Don't expose unsupported capabilities for realm guests

[Suzuki K Poulose]

On 18/03/2026 15:53, Steven Price wrote:
> From: Suzuki K Poulose <suzuki.poulose at arm.com>
> 
> RMM v1.0 provides no mechanism for the host to perform debug operations
> on the guest. So limit the extensions that are visible to an allowlist
> so that only those capabilities we can support are advertised.
> 
> Signed-off-by: Suzuki K Poulose <suzuki.poulose at arm.com>
> Signed-off-by: Steven Price <steven.price at arm.com>
> ---
> Changes since v10:
>   * Add a kvm_realm_ext_allowed() function which limits which extensions
>     are exposed to an allowlist. This removes the need for special casing
>     various extensions.
> Changes since v7:
>   * Remove the helper functions and inline the kvm_is_realm() check with
>     a ternary operator.
>   * Rewrite the commit message to explain this patch.
> ---
>   arch/arm64/kvm/arm.c | 22 ++++++++++++++++++++++
>   1 file changed, 22 insertions(+)
> 
> diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c
> index 9b17bdfaf0c2..ddbf080e4f55 100644
> --- a/arch/arm64/kvm/arm.c
> +++ b/arch/arm64/kvm/arm.c
> @@ -357,6 +357,25 @@ static bool kvm_has_full_ptr_auth(void)
>   		(apa + api + apa3) == 1);
>   }
>   
> +static bool kvm_realm_ext_allowed(long ext)
> +{
> +	switch (ext) {
> +	case KVM_CAP_IRQCHIP:
> +	case KVM_CAP_ARM_PSCI:
> +	case KVM_CAP_ARM_PSCI_0_2:
> +	case KVM_CAP_NR_VCPUS:
> +	case KVM_CAP_MAX_VCPUS:
> +	case KVM_CAP_MAX_VCPU_ID:
> +	case KVM_CAP_MSI_DEVID:
> +	case KVM_CAP_ARM_VM_IPA_SIZE:
> +	case KVM_CAP_ARM_PTRAUTH_ADDRESS:
> +	case KVM_CAP_ARM_PTRAUTH_GENERIC:
> +	case KVM_CAP_ARM_RMI:
> +		return true;
> +	}
> +	return false;
> +}
> +
>   int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
>   {
>   	int r;
> @@ -364,6 +383,9 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
>   	if (is_protected_kvm_enabled() && !kvm_pkvm_ext_allowed(kvm, ext))
>   		return 0;
>   
> +	if (kvm && kvm_is_realm(kvm) && !kvm_realm_ext_allowed(ext))
> +		return 0;
> +

We need a similar check in in kvm_vm_ioctl_enable_cap() to prevent 
enabling the filtered caps ? Otherwise looks good to me.

Suzuki

>   	switch (ext) {
>   	case KVM_CAP_IRQCHIP:
>   		r = vgic_present;