Two main functions (vgic_get_pending_irq() and vgic_clear_pending_irq())
are responsible for ACK and EOI respectively. Maintenance code path
refactored in order to accommodate these changes. The main problem is that
__kvm_vgic_sync_hwstate() is called before I/O emulation takes place,
therefore vgic_process_maintenance() cannot be reused.
Signed-off-by: Pavel Fedin <p.fedin@xxxxxxxxxxx>
---
include/kvm/arm_vgic.h | 3 +
virt/kvm/arm/vgic.c | 149 ++++++++++++++++++++++++++++++++++++++-----------
virt/kvm/arm/vgic.h | 3 +
3 files changed, 121 insertions(+), 34 deletions(-)
diff --git a/include/kvm/arm_vgic.h b/include/kvm/arm_vgic.h
index 133ea00..31cda96 100644
--- a/include/kvm/arm_vgic.h
+++ b/include/kvm/arm_vgic.h
@@ -160,6 +160,9 @@ struct vgic_dist {
bool in_kernel;
bool ready;
+ /* Use software emulation of CPU interface */
+ bool sw_cpuif;
+
/* vGIC model the kernel emulates for the guest (GICv2 or GICv3) */
u32 vgic_model;
diff --git a/virt/kvm/arm/vgic.c b/virt/kvm/arm/vgic.c
index 78fb820..aec6063 100644
--- a/virt/kvm/arm/vgic.c
+++ b/virt/kvm/arm/vgic.c
@@ -1104,6 +1104,8 @@ static void vgic_retire_disabled_irqs(struct kvm_vcpu *vcpu)
static void vgic_queue_irq_to_lr(struct kvm_vcpu *vcpu, int irq,
int lr_nr, struct vgic_lr vlr)
{
+ struct vgic_dist *dist = &vcpu->kvm->arch.vgic;
+
if (vgic_irq_is_active(vcpu, irq)) {
vlr.state |= LR_STATE_ACTIVE;
kvm_debug("Set active, clear distributor: 0x%x\n", vlr.state);
@@ -1119,6 +1121,9 @@ static void vgic_queue_irq_to_lr(struct kvm_vcpu *vcpu, int irq,
vgic_set_lr(vcpu, lr_nr, vlr);
vgic_sync_lr_elrsr(vcpu, lr_nr, vlr);
+
+ if (dist->sw_cpuif && (vlr.state & LR_STATE_PENDING))
+ vcpu_set_hcr(vcpu, vcpu_get_hcr(vcpu) | HCR_VI);
}
/*
@@ -1171,6 +1176,115 @@ bool vgic_queue_irq(struct kvm_vcpu *vcpu, u8 sgi_source_id, int irq)
return true;
}
+int vgic_get_pending_irq(struct kvm_vcpu *vcpu)
+{
+ struct vgic_cpu *vgic_cpu = &vcpu->arch.vgic_cpu;
+ struct vgic_lr vlr;
+ int lr;
+
+ for_each_set_bit(lr, vgic_cpu->lr_used, vgic->nr_lr) {
+ vlr = vgic_get_lr(vcpu, lr);
+
+ if (vlr.state & LR_STATE_PENDING) {
+ vlr.state &= ~LR_STATE_PENDING;
+ vlr.state |= LR_STATE_ACTIVE;
+ vgic_set_lr(vcpu, lr, vlr);
+
+ return vlr.irq;
+ }
+ }
+ vcpu_set_hcr(vcpu, vcpu_get_hcr(vcpu) & ~HCR_VI);
+ return 1023;
+}
+
+static int vgic_eoi_level_irq(struct kvm_vcpu *vcpu, int irq)
+{
+ struct vgic_dist *dist = &vcpu->kvm->arch.vgic;
+ int level_pending;
+
+ vgic_irq_clear_queued(vcpu, irq);
+
+ /*
+ * If the IRQ was EOIed it was also ACKed and we we
+ * therefore assume we can clear the soft pending
+ * state (should it had been set) for this interrupt.
+ *
+ * Note: if the IRQ soft pending state was set after
+ * the IRQ was acked, it actually shouldn't be
+ * cleared, but we have no way of knowing that unless
+ * we start trapping ACKs when the soft-pending state
+ * is set.
+ */
+ vgic_dist_irq_clear_soft_pend(vcpu, irq);
+
+ /*
+ * kvm_notify_acked_irq calls kvm_set_irq()
+ * to reset the IRQ level. Need to release the
+ * lock for kvm_set_irq to grab it.
+ */
+ spin_unlock(&dist->lock);
+
+ kvm_notify_acked_irq(vcpu->kvm, 0, irq - VGIC_NR_PRIVATE_IRQS);
+ spin_lock(&dist->lock);
+
+ /* Any additional pending interrupt? */
+ level_pending = vgic_dist_irq_get_level(vcpu, irq);
+ if (level_pending) {
+ vgic_cpu_irq_set(vcpu, irq);
+ } else {
+ vgic_dist_irq_clear_pending(vcpu, irq);
+ vgic_cpu_irq_clear(vcpu, irq);
+ }
+
+ return level_pending;
+}
+
+void vgic_clear_pending_irq(struct kvm_vcpu *vcpu, int irq)
+{
+ struct vgic_dist *dist = &vcpu->kvm->arch.vgic;
+ struct vgic_cpu *vgic_cpu = &vcpu->arch.vgic_cpu;
+ int level_pending = 0;
+ struct vgic_lr vlr;
+ int lr;
+
+ if (unlikely(irq >= dist->nr_irqs)) {
+ kvm_info("EOI on invalid IRQ %d from guest\n", irq);
+ return;
+ }
+
+ lr = vgic_cpu->vgic_irq_lr_map[irq];
+ if (unlikely(lr == LR_EMPTY)) {
+ kvm_info("EOI on non-pending IRQ %d from guest\n", irq);
+ return;
+ }
+
+ vlr = vgic_get_lr(vcpu, lr);
+
+ if (vlr.state & LR_EOI_INT)
+ level_pending = vgic_eoi_level_irq(vcpu, vlr.irq);
+
+ vlr.state = 0;
+ vgic_set_lr(vcpu, lr, vlr);
+ vgic_sync_lr_elrsr(vcpu, lr, vlr);
+
+ clear_bit(lr, vgic_cpu->lr_used);
+ vgic_cpu->vgic_irq_lr_map[vlr.irq] = LR_EMPTY;
+
+ /*
+ * The following is the same as end of __kvm_vgic_sync_hwstate,
+ * but with conditions inverted for performance optimization.
+ */
+ if (!level_pending) {
+ u64 elrsr = vgic_get_elrsr(vcpu);
+ unsigned long *elrsr_ptr = u64_to_bitmask(&elrsr);
+ int pending = find_first_zero_bit(elrsr_ptr, vgic->nr_lr);
+
+ if (pending >= vgic->nr_lr)
+ return;
+ }
+ set_bit(vcpu->vcpu_id, dist->irq_pending_on_cpu);
+}
+
static bool vgic_queue_hwirq(struct kvm_vcpu *vcpu, int irq)
{
if (!vgic_can_sample_irq(vcpu, irq))
@@ -1261,7 +1375,6 @@ 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;
- struct kvm *kvm = vcpu->kvm;
kvm_debug("STATUS = %08x\n", status);
@@ -1279,43 +1392,11 @@ static bool vgic_process_maintenance(struct kvm_vcpu *vcpu)
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_set_lr(vcpu, lr, vlr);
- /*
- * If the IRQ was EOIed it was also ACKed and we we
- * therefore assume we can clear the soft pending
- * state (should it had been set) for this interrupt.
- *
- * Note: if the IRQ soft pending state was set after
- * the IRQ was acked, it actually shouldn't be
- * cleared, but we have no way of knowing that unless
- * we start trapping ACKs when the soft-pending state
- * is set.
- */
- vgic_dist_irq_clear_soft_pend(vcpu, vlr.irq);
-
- /*
- * kvm_notify_acked_irq calls kvm_set_irq()
- * to reset the IRQ level. Need to release the
- * lock for kvm_set_irq to grab it.
- */
- spin_unlock(&dist->lock);
-
- kvm_notify_acked_irq(kvm, 0,
- vlr.irq - VGIC_NR_PRIVATE_IRQS);
- spin_lock(&dist->lock);
-
- /* Any additional pending interrupt? */
- if (vgic_dist_irq_get_level(vcpu, vlr.irq)) {
- vgic_cpu_irq_set(vcpu, vlr.irq);
- level_pending = true;
- } else {
- vgic_dist_irq_clear_pending(vcpu, vlr.irq);
- vgic_cpu_irq_clear(vcpu, vlr.irq);
- }
+ level_pending |= vgic_eoi_level_irq(vcpu, vlr.irq);
spin_unlock(&dist->lock);
diff --git a/virt/kvm/arm/vgic.h b/virt/kvm/arm/vgic.h
index 0df74cb..75ecfef 100644
--- a/virt/kvm/arm/vgic.h
+++ b/virt/kvm/arm/vgic.h
@@ -59,6 +59,9 @@ void vgic_set_vmcr(struct kvm_vcpu *vcpu, struct vgic_vmcr *vmcr);
bool vgic_queue_irq(struct kvm_vcpu *vcpu, u8 sgi_source_id, int irq);
void vgic_unqueue_irqs(struct kvm_vcpu *vcpu);
+int vgic_get_pending_irq(struct kvm_vcpu *vcpu);
+void vgic_clear_pending_irq(struct kvm_vcpu *vcpu, int irq);
+
struct kvm_exit_mmio {
phys_addr_t phys_addr;
void *data;
--
2.1.4
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
