__kvm_vgic_sync_from_cpu() doesn't really touch the distributor itself,
as the irq_pending_on_cpu access doesn't really need any locking.
Remove the corresponding spin_lock() access from kvm_vgic_sync_from_cpu().
Signed-off-by: Marc Zyngier <marc.zyngier@xxxxxxx>
---
arch/arm/kvm/vgic.c | 12 ++++++------
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/arch/arm/kvm/vgic.c b/arch/arm/kvm/vgic.c
index b77f8bf..9cac7b2 100644
--- a/arch/arm/kvm/vgic.c
+++ b/arch/arm/kvm/vgic.c
@@ -790,7 +790,9 @@ epilog:
}
/*
- * Sync back the VGIC state after a guest run.
+ * Sync back the VGIC state after a guest run. We do not really touch
+ * the distributor here (the irq_pending_on_cpu bit is safe to set),
+ * so there is no need for taking its lock.
*/
static void __kvm_vgic_sync_from_cpu(struct kvm_vcpu *vcpu)
{
@@ -815,8 +817,10 @@ static void __kvm_vgic_sync_from_cpu(struct kvm_vcpu *vcpu)
/* Check if we still have something up our sleeve... */
pending = find_first_zero_bit((unsigned long *)vgic_cpu->vgic_elrsr,
vgic_cpu->nr_lr);
- if (pending < vgic_cpu->nr_lr)
+ if (pending < vgic_cpu->nr_lr) {
set_bit(vcpu->vcpu_id, &dist->irq_pending_on_cpu);
+ smp_mb();
+ }
}
void kvm_vgic_sync_to_cpu(struct kvm_vcpu *vcpu)
@@ -833,14 +837,10 @@ void kvm_vgic_sync_to_cpu(struct kvm_vcpu *vcpu)
void kvm_vgic_sync_from_cpu(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_from_cpu(vcpu);
- spin_unlock(&dist->lock);
}
int kvm_vgic_vcpu_pending_irq(struct kvm_vcpu *vcpu)
--
1.7.12
_______________________________________________
kvmarm mailing list
kvmarm@xxxxxxxxxxxxxxxxxxxxx
https://lists.cs.columbia.edu/cucslists/listinfo/kvmarm
