static bool vgic_process_maintenance(struct kvm_vcpu *vcpu)
 {
        u32 status = vgic_get_interrupt_status(vcpu);
+       struct vgic_dist *dist = &vcpu->kvm->arch.vgic;
        bool level_pending = false;
 
        kvm_debug("STATUS = %08x\n", status);
                        struct vgic_lr vlr = vgic_get_lr(vcpu, lr);
                        WARN_ON(vgic_irq_is_edge(vcpu, vlr.irq));
 
+                       spin_lock(&dist->lock);
                        vgic_irq_clear_queued(vcpu, vlr.irq);
                        WARN_ON(vlr.state & LR_STATE_MASK);
                        vlr.state = 0;
                                vgic_cpu_irq_clear(vcpu, vlr.irq);
                        }
 
+                       spin_unlock(&dist->lock);
+
                        /*
                         * Despite being EOIed, the LR may not have
                         * been marked as empty.
        return level_pending;
 }
 
-/*
- * Sync back the VGIC state after a guest run. The distributor lock is
- * needed so we don't get preempted in the middle of the state processing.
- */
+/* Sync back the VGIC state after a guest run */
 static void __kvm_vgic_sync_hwstate(struct kvm_vcpu *vcpu)
 {
        struct vgic_cpu *vgic_cpu = &vcpu->arch.vgic_cpu;
 
 void kvm_vgic_sync_hwstate(struct kvm_vcpu *vcpu)
 {
-       struct vgic_dist *dist = &vcpu->kvm->arch.vgic;
-
        if (!irqchip_in_kernel(vcpu->kvm))
                return;
 
-       spin_lock(&dist->lock);
        __kvm_vgic_sync_hwstate(vcpu);
-       spin_unlock(&dist->lock);
 }
 
 int kvm_vgic_vcpu_pending_irq(struct kvm_vcpu *vcpu)