Just like for VLPIs, it is beneficial to avoid trapping on WFI when the vcpu is using the GICv4.1 SGIs.
Add such a check to vcpu_clear_wfx_traps(). Signed-off-by: Marc Zyngier <[email protected]> Reviewed-by: Zenghui Yu <[email protected]> Reviewed-by: Eric Auger <[email protected]> Link: https://lore.kernel.org/r/[email protected] --- arch/arm64/include/asm/kvm_emulate.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/arch/arm64/include/asm/kvm_emulate.h b/arch/arm64/include/asm/kvm_emulate.h index f658dda12364..a30b4eec7cb4 100644 --- a/arch/arm64/include/asm/kvm_emulate.h +++ b/arch/arm64/include/asm/kvm_emulate.h @@ -89,7 +89,8 @@ static inline unsigned long *vcpu_hcr(struct kvm_vcpu *vcpu) static inline void vcpu_clear_wfx_traps(struct kvm_vcpu *vcpu) { vcpu->arch.hcr_el2 &= ~HCR_TWE; - if (atomic_read(&vcpu->arch.vgic_cpu.vgic_v3.its_vpe.vlpi_count)) + if (atomic_read(&vcpu->arch.vgic_cpu.vgic_v3.its_vpe.vlpi_count) || + vcpu->kvm->arch.vgic.nassgireq) vcpu->arch.hcr_el2 &= ~HCR_TWI; else vcpu->arch.hcr_el2 |= HCR_TWI; -- 2.20.1 _______________________________________________ kvmarm mailing list [email protected] https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
