Re: [PATCH v20 07/14] KVM: arm64: Forward GCS exceptions to nested guests

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Incorrect context check for forwarding GCS exceptions to a nested hypervisor.
--

commit 2265e268ea1606e4aa21cb8062727c3b949089d6
Author: Mark Brown <broonie@kernel.org>

KVM: arm64: Forward GCS exceptions to nested guests

For nested guests where HFGITR_EL2.nGCSSTR_EL1 is clear or when there are
L2 GCS data check exceptions we need to forward the exception to the guest.
Add handling to do so.

> diff --git a/arch/arm64/kvm/handle_exit.c b/arch/arm64/kvm/handle_exit.c
> index db37678dcb05c..a8198e96fcb41 100644
> --- a/arch/arm64/kvm/handle_exit.c
> +++ b/arch/arm64/kvm/handle_exit.c

[ ... ]

> +static int kvm_handle_gcs(struct kvm_vcpu *vcpu)
>  {
> -	/* We don't expect GCS, so treat it with contempt */
> -	if (kvm_has_feat(vcpu->kvm, ID_AA64PFR1_EL1, GCS, IMP))
> -		WARN_ON_ONCE(1);
> +	if (!kvm_has_gcs(vcpu->kvm)) {
> +		kvm_inject_undefined(vcpu);
> +		return 1;
> +	}
>  
> +	if (vcpu_has_nv(vcpu)) {
> +		kvm_inject_nested_sync(vcpu, kvm_vcpu_get_esr(vcpu));
> +		return 1;
> +	}

[Severity: Medium]
Is it correct to only check vcpu_has_nv() here without verifying if the vCPU
is currently executing in a nested context?

If a hardware or hypervisor bug causes a GCS exception to trap to the host
while running L1's EL2, vcpu_has_nv() will be true, but it shouldn't be
forwarded to the guest. Instead of falling back to the WARN_ON_ONCE() and
undefined exception injection at the end of kvm_handle_gcs(), this would
silently emulate an exception to L1's EL2, potentially corrupting its state.

Should this condition also check is_nested_ctxt(vcpu) to ensure we are
actually running a nested guest before forwarding the exception?

> +
> +	WARN_ON_ONCE(1);
>  	kvm_inject_undefined(vcpu);
>  	return 1;
>  }

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260901-arm64-gcs-v20-0-f31750bdfadb@kernel.org?part=7