Re: [PATCH v7 12/41] KVM: arm64: gic-v5: Sanitize ID_AA64PFR2_EL1.GCIE

[Marc Zyngier <maz@kernel.org>]

On Tue, 24 Mar 2026 14:53:58 +0000,
Mark Brown <broonie@kernel.org> wrote:
> 
> [1  <text/plain; us-ascii (7bit)>]
> On Tue, Mar 24, 2026 at 02:48:04PM +0000, Marc Zyngier wrote:
> > Mark Brown <broonie@kernel.org> wrote:
> 
> > > We're also seeing an issue with this in the aarch32_id_regs test:
> 
> ...
> 
> > > which for some reason only manifests on TX2 of the platforms I've
> > > seen.
> 
> > Is your TX2 the only machine you have that is AArch64 only at all ELs?
> 
> Yes, that should be the case.

The hack below fixes it on my favourite Qualcomm box.

	M.

diff --git a/arch/arm64/kvm/vgic/vgic-init.c b/arch/arm64/kvm/vgic/vgic-init.c
index fd872079f2a24..7212cb3766d3d 100644
--- a/arch/arm64/kvm/vgic/vgic-init.c
+++ b/arch/arm64/kvm/vgic/vgic-init.c
@@ -696,7 +696,8 @@ void kvm_vgic_finalize_idregs(struct kvm *kvm)
 		break;
 	case KVM_DEV_TYPE_ARM_VGIC_V3:
 		aa64pfr0 |= SYS_FIELD_PREP_ENUM(ID_AA64PFR0_EL1, GIC, IMP);
-		pfr1 |= SYS_FIELD_PREP_ENUM(ID_PFR1_EL1, GIC, GICv3);
+		if (kvm_supports_32bit_el0())
+			pfr1 |= SYS_FIELD_PREP_ENUM(ID_PFR1_EL1, GIC, GICv3);
 		break;
 	case KVM_DEV_TYPE_ARM_VGIC_V5:
 		aa64pfr2 |= SYS_FIELD_PREP_ENUM(ID_AA64PFR2_EL1, GCIE, IMP);

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