Merge tag 'kvmarm-6.8' of git://git.kernel.org/pub/scm/linux/kernel/git/kvmarm/kvmarm into HEAD

KVM/arm64 updates for Linux 6.8

- LPA2 support, adding 52bit IPA/PA capability for 4kB and 16kB
  base granule sizes. Branch shared with the arm64 tree.

- Large Fine-Grained Trap rework, bringing some sanity to the
  feature, although there is more to come. This comes with
  a prefix branch shared with the arm64 tree.

- Some additional Nested Virtualization groundwork, mostly
  introducing the NV2 VNCR support and retargetting the NV
  support to that version of the architecture.

- A small set of vgic fixes and associated cleanups.

3 files changed, 53 insertions, 81 deletions

diff --git a/arch/arm64/kvm/vgic/vgic-its.c b/arch/arm64/kvm/vgic/vgic-its.c
index 2dad2d095160..e2764d0ffa9f 100644
--- a/arch/arm64/kvm/vgic/vgic-its.c
+++ b/arch/arm64/kvm/vgic/vgic-its.c
@@ -590,7 +590,11 @@ static struct vgic_irq *vgic_its_check_cache(struct kvm *kvm, phys_addr_t db,
 	unsigned long flags;
 
 	raw_spin_lock_irqsave(&dist->lpi_list_lock, flags);
+
 	irq = __vgic_its_check_cache(dist, db, devid, eventid);
+	if (irq)
+		vgic_get_irq_kref(irq);
+
 	raw_spin_unlock_irqrestore(&dist->lpi_list_lock, flags);
 
 	return irq;
@@ -769,6 +773,7 @@ int vgic_its_inject_cached_translation(struct kvm *kvm, struct kvm_msi *msi)
 	raw_spin_lock_irqsave(&irq->irq_lock, flags);
 	irq->pending_latch = true;
 	vgic_queue_irq_unlock(kvm, irq, flags);
+	vgic_put_irq(kvm, irq);
 
 	return 0;
 }
diff --git a/arch/arm64/kvm/vgic/vgic-mmio-v3.c b/arch/arm64/kvm/vgic/vgic-mmio-v3.c
index a764b0ab8bf9..c15ee1df036a 100644
--- a/arch/arm64/kvm/vgic/vgic-mmio-v3.c
+++ b/arch/arm64/kvm/vgic/vgic-mmio-v3.c
@@ -357,31 +357,13 @@ static int vgic_v3_uaccess_write_pending(struct kvm_vcpu *vcpu,
 					 gpa_t addr, unsigned int len,
 					 unsigned long val)
 {
-	u32 intid = VGIC_ADDR_TO_INTID(addr, 1);
-	int i;
-	unsigned long flags;
-
-	for (i = 0; i < len * 8; i++) {
-		struct vgic_irq *irq = vgic_get_irq(vcpu->kvm, vcpu, intid + i);
-
-		raw_spin_lock_irqsave(&irq->irq_lock, flags);
-		if (test_bit(i, &val)) {
-			/*
-			 * pending_latch is set irrespective of irq type
-			 * (level or edge) to avoid dependency that VM should
-			 * restore irq config before pending info.
-			 */
-			irq->pending_latch = true;
-			vgic_queue_irq_unlock(vcpu->kvm, irq, flags);
-		} else {
-			irq->pending_latch = false;
-			raw_spin_unlock_irqrestore(&irq->irq_lock, flags);
-		}
+	int ret;
 
-		vgic_put_irq(vcpu->kvm, irq);
-	}
+	ret = vgic_uaccess_write_spending(vcpu, addr, len, val);
+	if (ret)
+		return ret;
 
-	return 0;
+	return vgic_uaccess_write_cpending(vcpu, addr, len, ~val);
 }
 
 /* We want to avoid outer shareable. */
diff --git a/arch/arm64/kvm/vgic/vgic-mmio.c b/arch/arm64/kvm/vgic/vgic-mmio.c
index ff558c05e990..cf76523a2194 100644
--- a/arch/arm64/kvm/vgic/vgic-mmio.c
+++ b/arch/arm64/kvm/vgic/vgic-mmio.c
@@ -301,9 +301,8 @@ static bool is_vgic_v2_sgi(struct kvm_vcpu *vcpu, struct vgic_irq *irq)
 		vcpu->kvm->arch.vgic.vgic_model == KVM_DEV_TYPE_ARM_VGIC_V2);
 }
 
-void vgic_mmio_write_spending(struct kvm_vcpu *vcpu,
-			      gpa_t addr, unsigned int len,
-			      unsigned long val)
+static void __set_pending(struct kvm_vcpu *vcpu, gpa_t addr, unsigned int len,
+			  unsigned long val, bool is_user)
 {
 	u32 intid = VGIC_ADDR_TO_INTID(addr, 1);
 	int i;
@@ -312,14 +311,22 @@ void vgic_mmio_write_spending(struct kvm_vcpu *vcpu,
 	for_each_set_bit(i, &val, len * 8) {
 		struct vgic_irq *irq = vgic_get_irq(vcpu->kvm, vcpu, intid + i);
 
-		/* GICD_ISPENDR0 SGI bits are WI */
-		if (is_vgic_v2_sgi(vcpu, irq)) {
+		/* GICD_ISPENDR0 SGI bits are WI when written from the guest. */
+		if (is_vgic_v2_sgi(vcpu, irq) && !is_user) {
 			vgic_put_irq(vcpu->kvm, irq);
 			continue;
 		}
 
 		raw_spin_lock_irqsave(&irq->irq_lock, flags);
 
+		/*
+		 * GICv2 SGIs are terribly broken. We can't restore
+		 * the source of the interrupt, so just pick the vcpu
+		 * itself as the source...
+		 */
+		if (is_vgic_v2_sgi(vcpu, irq))
+			irq->source |= BIT(vcpu->vcpu_id);
+
 		if (irq->hw && vgic_irq_is_sgi(irq->intid)) {
 			/* HW SGI? Ask the GIC to inject it */
 			int err;
@@ -335,7 +342,7 @@ void vgic_mmio_write_spending(struct kvm_vcpu *vcpu,
 		}
 
 		irq->pending_latch = true;
-		if (irq->hw)
+		if (irq->hw && !is_user)
 			vgic_irq_set_phys_active(irq, true);
 
 		vgic_queue_irq_unlock(vcpu->kvm, irq, flags);
@@ -343,33 +350,18 @@ void vgic_mmio_write_spending(struct kvm_vcpu *vcpu,
 	}
 }
 
+void vgic_mmio_write_spending(struct kvm_vcpu *vcpu,
+			      gpa_t addr, unsigned int len,
+			      unsigned long val)
+{
+	__set_pending(vcpu, addr, len, val, false);
+}
+
 int vgic_uaccess_write_spending(struct kvm_vcpu *vcpu,
 				gpa_t addr, unsigned int len,
 				unsigned long val)
 {
-	u32 intid = VGIC_ADDR_TO_INTID(addr, 1);
-	int i;
-	unsigned long flags;
-
-	for_each_set_bit(i, &val, len * 8) {
-		struct vgic_irq *irq = vgic_get_irq(vcpu->kvm, vcpu, intid + i);
-
-		raw_spin_lock_irqsave(&irq->irq_lock, flags);
-		irq->pending_latch = true;
-
-		/*
-		 * GICv2 SGIs are terribly broken. We can't restore
-		 * the source of the interrupt, so just pick the vcpu
-		 * itself as the source...
-		 */
-		if (is_vgic_v2_sgi(vcpu, irq))
-			irq->source |= BIT(vcpu->vcpu_id);
-
-		vgic_queue_irq_unlock(vcpu->kvm, irq, flags);
-
-		vgic_put_irq(vcpu->kvm, irq);
-	}
-
+	__set_pending(vcpu, addr, len, val, true);
 	return 0;
 }
 
@@ -394,9 +386,9 @@ static void vgic_hw_irq_cpending(struct kvm_vcpu *vcpu, struct vgic_irq *irq)
 		vgic_irq_set_phys_active(irq, false);
 }
 
-void vgic_mmio_write_cpending(struct kvm_vcpu *vcpu,
-			      gpa_t addr, unsigned int len,
-			      unsigned long val)
+static void __clear_pending(struct kvm_vcpu *vcpu,
+			    gpa_t addr, unsigned int len,
+			    unsigned long val, bool is_user)
 {
 	u32 intid = VGIC_ADDR_TO_INTID(addr, 1);
 	int i;
@@ -405,14 +397,22 @@ void vgic_mmio_write_cpending(struct kvm_vcpu *vcpu,
 	for_each_set_bit(i, &val, len * 8) {
 		struct vgic_irq *irq = vgic_get_irq(vcpu->kvm, vcpu, intid + i);
 
-		/* GICD_ICPENDR0 SGI bits are WI */
-		if (is_vgic_v2_sgi(vcpu, irq)) {
+		/* GICD_ICPENDR0 SGI bits are WI when written from the guest. */
+		if (is_vgic_v2_sgi(vcpu, irq) && !is_user) {
 			vgic_put_irq(vcpu->kvm, irq);
 			continue;
 		}
 
 		raw_spin_lock_irqsave(&irq->irq_lock, flags);
 
+		/*
+		 * More fun with GICv2 SGIs! If we're clearing one of them
+		 * from userspace, which source vcpu to clear? Let's not
+		 * even think of it, and blow the whole set.
+		 */
+		if (is_vgic_v2_sgi(vcpu, irq))
+			irq->source = 0;
+
 		if (irq->hw && vgic_irq_is_sgi(irq->intid)) {
 			/* HW SGI? Ask the GIC to clear its pending bit */
 			int err;
@@ -427,7 +427,7 @@ void vgic_mmio_write_cpending(struct kvm_vcpu *vcpu,
 			continue;
 		}
 
-		if (irq->hw)
+		if (irq->hw && !is_user)
 			vgic_hw_irq_cpending(vcpu, irq);
 		else
 			irq->pending_latch = false;
@@ -437,33 +437,18 @@ void vgic_mmio_write_cpending(struct kvm_vcpu *vcpu,
 	}
 }
 
+void vgic_mmio_write_cpending(struct kvm_vcpu *vcpu,
+			      gpa_t addr, unsigned int len,
+			      unsigned long val)
+{
+	__clear_pending(vcpu, addr, len, val, false);
+}
+
 int vgic_uaccess_write_cpending(struct kvm_vcpu *vcpu,
 				gpa_t addr, unsigned int len,
 				unsigned long val)
 {
-	u32 intid = VGIC_ADDR_TO_INTID(addr, 1);
-	int i;
-	unsigned long flags;
-
-	for_each_set_bit(i, &val, len * 8) {
-		struct vgic_irq *irq = vgic_get_irq(vcpu->kvm, vcpu, intid + i);
-
-		raw_spin_lock_irqsave(&irq->irq_lock, flags);
-		/*
-		 * More fun with GICv2 SGIs! If we're clearing one of them
-		 * from userspace, which source vcpu to clear? Let's not
-		 * even think of it, and blow the whole set.
-		 */
-		if (is_vgic_v2_sgi(vcpu, irq))
-			irq->source = 0;
-
-		irq->pending_latch = false;
-
-		raw_spin_unlock_irqrestore(&irq->irq_lock, flags);
-
-		vgic_put_irq(vcpu->kvm, irq);
-	}
-
+	__clear_pending(vcpu, addr, len, val, true);
 	return 0;
 }