[PATCH v8 11/11] KVM: arm64: Handle protected guests at 32 bits

[Marc Zyngier]

On Sun, 10 Oct 2021 15:56:36 +0100,
Fuad Tabba <tabba at google.com> wrote:
> 
> Protected KVM does not support protected AArch32 guests. However,
> it is possible for the guest to force run AArch32, potentially
> causing problems. Add an extra check so that if the hypervisor
> catches the guest doing that, it can prevent the guest from
> running again by resetting vcpu->arch.target and returning
> ARM_EXCEPTION_IL.
> 
> If this were to happen, The VMM can try and fix it by re-
> initializing the vcpu with KVM_ARM_VCPU_INIT, however, this is
> likely not possible for protected VMs.
> 
> Adapted from commit 22f553842b14 ("KVM: arm64: Handle Asymmetric
> AArch32 systems")
> 
> Signed-off-by: Fuad Tabba <tabba at google.com>
> ---
>  arch/arm64/kvm/hyp/nvhe/switch.c | 34 ++++++++++++++++++++++++++++++++
>  1 file changed, 34 insertions(+)
> 
> diff --git a/arch/arm64/kvm/hyp/nvhe/switch.c b/arch/arm64/kvm/hyp/nvhe/switch.c
> index 2c72c31e516e..f25b6353a598 100644
> --- a/arch/arm64/kvm/hyp/nvhe/switch.c
> +++ b/arch/arm64/kvm/hyp/nvhe/switch.c
> @@ -232,6 +232,37 @@ static const exit_handler_fn *kvm_get_exit_handler_array(struct kvm *kvm)
>  	return hyp_exit_handlers;
>  }
>  
> +/*
> + * Some guests (e.g., protected VMs) are not be allowed to run in AArch32.
> + * The ARMv8 architecture does not give the hypervisor a mechanism to prevent a
> + * guest from dropping to AArch32 EL0 if implemented by the CPU. If the
> + * hypervisor spots a guest in such a state ensure it is handled, and don't
> + * trust the host to spot or fix it.  The check below is based on the one in
> + * kvm_arch_vcpu_ioctl_run().
> + *
> + * Returns false if the guest ran in AArch32 when it shouldn't have, and
> + * thus should exit to the host, or true if a the guest run loop can continue.
> + */
> +static bool handle_aarch32_guest(struct kvm_vcpu *vcpu, u64 *exit_code)
> +{
> +	struct kvm *kvm = kern_hyp_va(vcpu->kvm);
> +
> +	if (kvm_vm_is_protected(kvm) && vcpu_mode_is_32bit(vcpu)) {
> +		/*
> +		 * As we have caught the guest red-handed, decide that it isn't
> +		 * fit for purpose anymore by making the vcpu invalid. The VMM
> +		 * can try and fix it by re-initializing the vcpu with
> +		 * KVM_ARM_VCPU_INIT, however, this is likely not possible for
> +		 * protected VMs.
> +		 */
> +		vcpu->arch.target = -1;
> +		*exit_code = ARM_EXCEPTION_IL;

Aren't we losing a potential SError here, which the original commit
doesn't need to handle? I'd expect something like:

		*exit_code &= BIT(ARM_EXIT_WITH_SERROR_BIT);
		*exit_code |= ARM_EXCEPTION_IL;

> +		return false;
> +	}
> +
> +	return true;
> +}
> +
>  /* Switch to the guest for legacy non-VHE systems */
>  int __kvm_vcpu_run(struct kvm_vcpu *vcpu)
>  {
> @@ -294,6 +325,9 @@ int __kvm_vcpu_run(struct kvm_vcpu *vcpu)
>  		/* Jump in the fire! */
>  		exit_code = __guest_enter(vcpu);
>  
> +		if (unlikely(!handle_aarch32_guest(vcpu, &exit_code)))
> +			break;
> +
>  		/* And we're baaack! */
>  	} while (fixup_guest_exit(vcpu, &exit_code));
>  

Thanks,

	M.

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