[PATCH 3/4] KVM: arm64: vgic-v3: Only pause the targeted vCPU when disabling LPIs

[Oliver Upton <oupton@kernel.org>]

KVM pauses the VM any time LPIs are disabled at a redistributor, as
retiring pending LPIs was a bit dicey since it involved manipulating the
AP list of another vCPU while it was running. Of course, this is a very
large hammer and we could do slightly better.

Only pause the targeted vCPU when disabling LPIs, which is almost always
going to be done locally unless your guest is up to no good.

Signed-off-by: Oliver Upton <oupton@kernel.org>
---
 arch/arm64/kvm/vgic/vgic-mmio-v3.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/arch/arm64/kvm/vgic/vgic-mmio-v3.c b/arch/arm64/kvm/vgic/vgic-mmio-v3.c
index e5fd7a0002b5..2dd8d83fc93d 100644
--- a/arch/arm64/kvm/vgic/vgic-mmio-v3.c
+++ b/arch/arm64/kvm/vgic/vgic-mmio-v3.c
@@ -302,16 +302,16 @@ static void vgic_mmio_write_v3r_ctlr(struct kvm_vcpu *vcpu,
 		/*
 		 * Yes, disabling LPIs is painful, since it can be done from
 		 * a *remote* vcpu! So let's not take any chance, and make
-		 * sure that everybody has written their LRs back to the irq
-		 * structures, and release any reference they would have.
+		 * sure that the target vCPU has written their LRs back to the
+		 * irq structures, and release any reference they would have.
 		 *
 		 * If it hurts, don't do it.
 		 */
-		kvm_arm_halt_guest(vcpu->kvm);
+		kvm_pause_vcpu(vcpu);
 		vgic_flush_pending_lpis(vcpu);
 		vgic_its_invalidate_all_caches(vcpu->kvm);
 		atomic_set_release(&vgic_cpu->ctlr, 0);
-		kvm_arm_resume_guest(vcpu->kvm);
+		kvm_resume_vcpu(vcpu);
 	} else {
 		ctlr = atomic_cmpxchg_acquire(&vgic_cpu->ctlr, 0,
 					      GICR_CTLR_ENABLE_LPIS);
-- 
2.47.3