[Qemu-devel] [RFC 2/2] KVM: s390: add floating irq controller

[Jens Freimann <jfrei@linux.vnet.ibm.com>]

This patch adds a floating irq controller as a kvm_device.
It will be necesary for migration of floating interrupts as well
as for hardening the reset code by allowing user space to explicitly
remove all pending floating interrupts.

Signed-off-by: Jens Freimann <jfrei@linux.vnet.ibm.com>

---
 arch/s390/include/uapi/asm/kvm.h |   5 +
 arch/s390/kvm/interrupt.c        | 192 +++++++++++++++++++++++++++++++--------
 arch/s390/kvm/kvm-s390.c         |   1 +
 include/linux/kvm_host.h         |   1 +
 include/uapi/linux/kvm.h         |   1 +
 virt/kvm/kvm_main.c              |   3 +
 6 files changed, 163 insertions(+), 40 deletions(-)

diff --git a/arch/s390/include/uapi/asm/kvm.h b/arch/s390/include/uapi/asm/kvm.h
index d25da59..33d52b8 100644
--- a/arch/s390/include/uapi/asm/kvm.h
+++ b/arch/s390/include/uapi/asm/kvm.h
@@ -16,6 +16,11 @@
 
 #define __KVM_S390
 
+/* Device control API: s390-specific devices */
+#define KVM_DEV_FLIC_DEQUEUE 1
+#define KVM_DEV_FLIC_ENQUEUE 2
+#define KVM_DEV_FLIC_CLEAR_IRQS 3
+
 /* for KVM_GET_REGS and KVM_SET_REGS */
 struct kvm_regs {
 	/* general purpose regs for s390 */
diff --git a/arch/s390/kvm/interrupt.c b/arch/s390/kvm/interrupt.c
index 25cf71d..065a402 100644
--- a/arch/s390/kvm/interrupt.c
+++ b/arch/s390/kvm/interrupt.c
@@ -656,15 +656,57 @@ struct kvm_s390_interrupt_info *kvm_s390_get_io_int(struct kvm *kvm,
 	return inti;
 }
 
-int kvm_s390_inject_vm(struct kvm *kvm,
-		       struct kvm_s390_interrupt *s390int)
+static void __inject_vm(struct kvm *kvm, struct kvm_s390_interrupt_info *inti)
 {
 	struct kvm_s390_local_interrupt *li;
 	struct kvm_s390_float_interrupt *fi;
-	struct kvm_s390_interrupt_info *inti, *iter;
-	struct kvm_s390_irq *irq;
+	struct kvm_s390_interrupt_info *iter;
 	int sigcpu;
 
+	mutex_lock(&kvm->lock);
+	fi = &kvm->arch.float_int;
+	spin_lock(&fi->lock);
+	if (!is_ioint(inti->irq.type))
+		list_add_tail(&inti->list, &fi->list);
+	else {
+		u64 isc_bits = int_word_to_isc_bits(inti->irq.io.io_int_word);
+
+		/* Keep I/O interrupts sorted in isc order. */
+		list_for_each_entry(iter, &fi->list, list) {
+			if (!is_ioint(iter->irq.type))
+				continue;
+			if (int_word_to_isc_bits(iter->irq.io.io_int_word)
+					<= isc_bits)
+				continue;
+			break;
+		}
+		list_add_tail(&inti->list, &iter->list);
+	}
+	atomic_set(&fi->active, 1);
+	sigcpu = find_first_bit(fi->idle_mask, KVM_MAX_VCPUS);
+	if (sigcpu == KVM_MAX_VCPUS) {
+		do {
+			sigcpu = fi->next_rr_cpu++;
+			if (sigcpu == KVM_MAX_VCPUS)
+				sigcpu = fi->next_rr_cpu = 0;
+		} while (fi->local_int[sigcpu] == NULL);
+	}
+	li = fi->local_int[sigcpu];
+	spin_lock_bh(&li->lock);
+	atomic_set_mask(CPUSTAT_EXT_INT, li->cpuflags);
+	if (waitqueue_active(li->wq))
+		wake_up_interruptible(li->wq);
+	spin_unlock_bh(&li->lock);
+	spin_unlock(&fi->lock);
+	mutex_unlock(&kvm->lock);
+}
+
+int kvm_s390_inject_vm(struct kvm *kvm,
+		       struct kvm_s390_interrupt *s390int)
+{
+	struct kvm_s390_interrupt_info *inti;
+	struct kvm_s390_irq *irq;
+
 	inti = kzalloc(sizeof(*inti), GFP_KERNEL);
 	if (!inti)
 		return -ENOMEM;
@@ -712,42 +754,7 @@ int kvm_s390_inject_vm(struct kvm *kvm,
 	}
 	trace_kvm_s390_inject_vm(s390int->type, s390int->parm, s390int->parm64, 2);
 
-	mutex_lock(&kvm->lock);
-	fi = &kvm->arch.float_int;
-	spin_lock(&fi->lock);
-	if (!is_ioint(inti->irq.type))
-		list_add_tail(&inti->list, &fi->list);
-	else {
-		u64 isc_bits = int_word_to_isc_bits(inti->irq.io.io_int_word);
-
-		/* Keep I/O interrupts sorted in isc order. */
-		list_for_each_entry(iter, &fi->list, list) {
-			if (!is_ioint(iter->irq.type))
-				continue;
-			if (int_word_to_isc_bits(iter->irq.io.io_int_word)
-			    <= isc_bits)
-				continue;
-			break;
-		}
-		list_add_tail(&inti->list, &iter->list);
-	}
-	atomic_set(&fi->active, 1);
-	sigcpu = find_first_bit(fi->idle_mask, KVM_MAX_VCPUS);
-	if (sigcpu == KVM_MAX_VCPUS) {
-		do {
-			sigcpu = fi->next_rr_cpu++;
-			if (sigcpu == KVM_MAX_VCPUS)
-				sigcpu = fi->next_rr_cpu = 0;
-		} while (fi->local_int[sigcpu] == NULL);
-	}
-	li = fi->local_int[sigcpu];
-	spin_lock_bh(&li->lock);
-	atomic_set_mask(CPUSTAT_EXT_INT, li->cpuflags);
-	if (waitqueue_active(li->wq))
-		wake_up_interruptible(li->wq);
-	spin_unlock_bh(&li->lock);
-	spin_unlock(&fi->lock);
-	mutex_unlock(&kvm->lock);
+	__inject_vm(kvm, inti);
 	return 0;
 }
 
@@ -835,3 +842,108 @@ int kvm_s390_inject_vcpu(struct kvm_vcpu *vcpu,
 	mutex_unlock(&vcpu->kvm->lock);
 	return 0;
 }
+
+static void clear_floating_interrupts(struct kvm *kvm)
+{
+	struct kvm_s390_float_interrupt *fi = &kvm->arch.float_int;
+	struct kvm_s390_interrupt_info  *n, *inti = NULL;
+
+	if (atomic_read(&fi->active)) {
+		spin_lock_bh(&fi->lock);
+		list_for_each_entry_safe(inti, n, &fi->list, list) {
+			list_del(&inti->list);
+			kfree(inti);
+		}
+		atomic_set(&fi->active, 0);
+		spin_unlock_bh(&fi->lock);
+	}
+}
+
+
+static int dequeue_floating_irq(struct kvm *kvm, __u64 addr)
+{
+	struct kvm_s390_interrupt_info *inti;
+	struct kvm_s390_float_interrupt *fi;
+	int r = 0;
+
+	fi = &kvm->arch.float_int;
+	mutex_lock(&kvm->lock);
+	spin_lock(&fi->lock);
+
+	if (!list_empty(&fi->list)) {
+		inti = list_first_entry(&fi->list,
+				struct kvm_s390_interrupt_info, list);
+		list_del(&inti->list);
+		if (copy_to_user((void *) addr, &inti->irq, sizeof(inti->irq)))
+			r = -EFAULT;
+	} else
+		r = -ENODATA;
+
+	spin_unlock(&fi->lock);
+	mutex_unlock(&kvm->lock);
+
+	return r;
+}
+
+static int flic_get_attr(struct kvm_device *dev, struct kvm_device_attr *attr)
+{
+	int r = 0;
+
+	switch (attr->group) {
+	case KVM_DEV_FLIC_DEQUEUE:
+		r = dequeue_floating_irq(dev->kvm, attr->addr);
+		break;
+	default:
+		r = -EINVAL;
+	}
+
+	return r;
+}
+
+static int flic_set_attr(struct kvm_device *dev, struct kvm_device_attr *attr)
+{
+	int r;
+
+	switch (attr->group) {
+	case KVM_DEV_FLIC_ENQUEUE: {
+		struct kvm_s390_irq *s390irq;
+		struct kvm_s390_interrupt_info *inti;
+		inti = kzalloc(sizeof(*inti), GFP_KERNEL);
+		if (!inti)
+			return -ENOMEM;
+		s390irq = kzalloc(sizeof(*s390irq), GFP_KERNEL);
+		if (!s390irq)
+			return -ENOMEM;
+		if (copy_from_user(s390irq, (u64 __user *)attr->addr,
+				  sizeof(s390irq)))
+			return -EFAULT;
+		inti->irq = *s390irq;
+		__inject_vm(dev->kvm, inti);
+		}
+		break;
+	case KVM_DEV_FLIC_CLEAR_IRQS:
+		clear_floating_interrupts(dev->kvm);
+	default:
+		r = -EINVAL;
+	}
+
+	return r;
+}
+
+static int flic_create(struct kvm_device *dev, u32 type)
+{
+	return 0;
+}
+
+static void flic_destroy(struct kvm_device *dev)
+{
+}
+
+/* s390 floating irq controller (flic) */
+struct kvm_device_ops kvm_flic_ops = {
+	.name = "kvm-flic",
+	.get_attr = flic_get_attr,
+	.set_attr = flic_set_attr,
+	.create = flic_create,
+	.destroy = flic_destroy,
+};
diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
index ba694d2..b85a3f0 100644
--- a/arch/s390/kvm/kvm-s390.c
+++ b/arch/s390/kvm/kvm-s390.c
@@ -150,6 +150,7 @@ int kvm_dev_ioctl_check_extension(long ext)
 	case KVM_CAP_ENABLE_CAP:
 	case KVM_CAP_S390_CSS_SUPPORT:
 	case KVM_CAP_IOEVENTFD:
+	case KVM_CAP_DEVICE_CTRL:
 		r = 1;
 		break;
 	case KVM_CAP_NR_VCPUS:
diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index a63d83e..fa517dd 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -1052,6 +1052,7 @@ struct kvm_device *kvm_device_from_filp(struct file *filp);
 
 extern struct kvm_device_ops kvm_mpic_ops;
 extern struct kvm_device_ops kvm_xics_ops;
+extern struct kvm_device_ops kvm_flic_ops;
 
 #ifdef CONFIG_HAVE_KVM_CPU_RELAX_INTERCEPT
 
diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h
index d6a1584..c5aa7ce 100644
--- a/include/uapi/linux/kvm.h
+++ b/include/uapi/linux/kvm.h
@@ -904,6 +904,7 @@ struct kvm_device_attr {
 #define KVM_DEV_TYPE_FSL_MPIC_20	1
 #define KVM_DEV_TYPE_FSL_MPIC_42	2
 #define KVM_DEV_TYPE_XICS		3
+#define KVM_DEV_TYPE_FLIC		4
 
 /*
  * ioctls for VM fds
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 1580dd4..cb017eb 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -2282,6 +2282,9 @@ static int kvm_ioctl_create_device(struct kvm *kvm,
 		ops = &kvm_xics_ops;
 		break;
 #endif
+	case KVM_DEV_TYPE_FLIC:
+		ops = &kvm_flic_ops;
+		break;
 	default:
 		return -ENODEV;
 	}
-- 
1.8.0.1