1 files changed, 266 insertions, 2 deletions

diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 98cd916448a8..485bcdc16552 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -51,6 +51,12 @@
 #include "coalesced_mmio.h"
 #endif
 
+#ifdef KVM_CAP_DEVICE_ASSIGNMENT
+#include <linux/pci.h>
+#include <linux/interrupt.h>
+#include "irq.h"
+#endif
+
 MODULE_AUTHOR("Qumranet");
 MODULE_LICENSE("GPL");
 
@@ -71,6 +77,240 @@ static long kvm_vcpu_ioctl(struct file *file, unsigned int ioctl,
 
 bool kvm_rebooting;
 
+#ifdef KVM_CAP_DEVICE_ASSIGNMENT
+static struct kvm_assigned_dev_kernel *kvm_find_assigned_dev(struct list_head *head,
+                                                      int assigned_dev_id)
+{
+        struct list_head *ptr;
+        struct kvm_assigned_dev_kernel *match;
+
+        list_for_each(ptr, head) {
+                match = list_entry(ptr, struct kvm_assigned_dev_kernel, list);
+                if (match->assigned_dev_id == assigned_dev_id)
+                        return match;
+        }
+        return NULL;
+}
+
+static void kvm_assigned_dev_interrupt_work_handler(struct work_struct *work)
+{
+        struct kvm_assigned_dev_kernel *assigned_dev;
+
+        assigned_dev = container_of(work, struct kvm_assigned_dev_kernel,
+                                    interrupt_work);
+
+        /* This is taken to safely inject irq inside the guest. When
+         * the interrupt injection (or the ioapic code) uses a
+         * finer-grained lock, update this
+         */
+        mutex_lock(&assigned_dev->kvm->lock);
+        kvm_set_irq(assigned_dev->kvm,
+                    assigned_dev->guest_irq, 1);
+        mutex_unlock(&assigned_dev->kvm->lock);
+        kvm_put_kvm(assigned_dev->kvm);
+}
+
+/* FIXME: Implement the OR logic needed to make shared interrupts on
+ * this line behave properly
+ */
+static irqreturn_t kvm_assigned_dev_intr(int irq, void *dev_id)
+{
+        struct kvm_assigned_dev_kernel *assigned_dev =
+                (struct kvm_assigned_dev_kernel *) dev_id;
+
+        kvm_get_kvm(assigned_dev->kvm);
+        schedule_work(&assigned_dev->interrupt_work);
+        disable_irq_nosync(irq);
+        return IRQ_HANDLED;
+}
+
+/* Ack the irq line for an assigned device */
+static void kvm_assigned_dev_ack_irq(struct kvm_irq_ack_notifier *kian)
+{
+        struct kvm_assigned_dev_kernel *dev;
+
+        if (kian->gsi == -1)
+                return;
+
+        dev = container_of(kian, struct kvm_assigned_dev_kernel,
+                           ack_notifier);
+        kvm_set_irq(dev->kvm, dev->guest_irq, 0);
+        enable_irq(dev->host_irq);
+}
+
+static void kvm_free_assigned_device(struct kvm *kvm,
+                                     struct kvm_assigned_dev_kernel
+                                     *assigned_dev)
+{
+        if (irqchip_in_kernel(kvm) && assigned_dev->irq_requested)
+                free_irq(assigned_dev->host_irq, (void *)assigned_dev);
+
+        kvm_unregister_irq_ack_notifier(kvm, &assigned_dev->ack_notifier);
+
+        if (cancel_work_sync(&assigned_dev->interrupt_work))
+                /* We had pending work. That means we will have to take
+                 * care of kvm_put_kvm.
+                 */
+                kvm_put_kvm(kvm);
+
+        pci_release_regions(assigned_dev->dev);
+        pci_disable_device(assigned_dev->dev);
+        pci_dev_put(assigned_dev->dev);
+
+        list_del(&assigned_dev->list);
+        kfree(assigned_dev);
+}
+
+void kvm_free_all_assigned_devices(struct kvm *kvm)
+{
+        struct list_head *ptr, *ptr2;
+        struct kvm_assigned_dev_kernel *assigned_dev;
+
+        list_for_each_safe(ptr, ptr2, &kvm->arch.assigned_dev_head) {
+                assigned_dev = list_entry(ptr,
+                                          struct kvm_assigned_dev_kernel,
+                                          list);
+
+                kvm_free_assigned_device(kvm, assigned_dev);
+        }
+}
+
+static int kvm_vm_ioctl_assign_irq(struct kvm *kvm,
+                                   struct kvm_assigned_irq
+                                   *assigned_irq)
+{
+        int r = 0;
+        struct kvm_assigned_dev_kernel *match;
+
+        mutex_lock(&kvm->lock);
+
+        match = kvm_find_assigned_dev(&kvm->arch.assigned_dev_head,
+                                      assigned_irq->assigned_dev_id);
+        if (!match) {
+                mutex_unlock(&kvm->lock);
+                return -EINVAL;
+        }
+
+        if (match->irq_requested) {
+                match->guest_irq = assigned_irq->guest_irq;
+                match->ack_notifier.gsi = assigned_irq->guest_irq;
+                mutex_unlock(&kvm->lock);
+                return 0;
+        }
+
+        INIT_WORK(&match->interrupt_work,
+                  kvm_assigned_dev_interrupt_work_handler);
+
+        if (irqchip_in_kernel(kvm)) {
+                if (!capable(CAP_SYS_RAWIO)) {
+                        r = -EPERM;
+                        goto out_release;
+                }
+
+                if (assigned_irq->host_irq)
+                        match->host_irq = assigned_irq->host_irq;
+                else
+                        match->host_irq = match->dev->irq;
+                match->guest_irq = assigned_irq->guest_irq;
+                match->ack_notifier.gsi = assigned_irq->guest_irq;
+                match->ack_notifier.irq_acked = kvm_assigned_dev_ack_irq;
+                kvm_register_irq_ack_notifier(kvm, &match->ack_notifier);
+
+                /* Even though this is PCI, we don't want to use shared
+                 * interrupts. Sharing host devices with guest-assigned devices
+                 * on the same interrupt line is not a happy situation: there
+                 * are going to be long delays in accepting, acking, etc.
+                 */
+                if (request_irq(match->host_irq, kvm_assigned_dev_intr, 0,
+                                "kvm_assigned_device", (void *)match)) {
+                        r = -EIO;
+                        goto out_release;
+                }
+        }
+
+        match->irq_requested = true;
+        mutex_unlock(&kvm->lock);
+        return r;
+out_release:
+        mutex_unlock(&kvm->lock);
+        kvm_free_assigned_device(kvm, match);
+        return r;
+}
+
+static int kvm_vm_ioctl_assign_device(struct kvm *kvm,
+                                      struct kvm_assigned_pci_dev *assigned_dev)
+{
+        int r = 0;
+        struct kvm_assigned_dev_kernel *match;
+        struct pci_dev *dev;
+
+        mutex_lock(&kvm->lock);
+
+        match = kvm_find_assigned_dev(&kvm->arch.assigned_dev_head,
+                                      assigned_dev->assigned_dev_id);
+        if (match) {
+                /* device already assigned */
+                r = -EINVAL;
+                goto out;
+        }
+
+        match = kzalloc(sizeof(struct kvm_assigned_dev_kernel), GFP_KERNEL);
+        if (match == NULL) {
+                printk(KERN_INFO "%s: Couldn't allocate memory\n",
+                       __func__);
+                r = -ENOMEM;
+                goto out;
+        }
+        dev = pci_get_bus_and_slot(assigned_dev->busnr,
+                                   assigned_dev->devfn);
+        if (!dev) {
+                printk(KERN_INFO "%s: host device not found\n", __func__);
+                r = -EINVAL;
+                goto out_free;
+        }
+        if (pci_enable_device(dev)) {
+                printk(KERN_INFO "%s: Could not enable PCI device\n", __func__);
+                r = -EBUSY;
+                goto out_put;
+        }
+        r = pci_request_regions(dev, "kvm_assigned_device");
+        if (r) {
+                printk(KERN_INFO "%s: Could not get access to device regions\n",
+                       __func__);
+                goto out_disable;
+        }
+        match->assigned_dev_id = assigned_dev->assigned_dev_id;
+        match->host_busnr = assigned_dev->busnr;
+        match->host_devfn = assigned_dev->devfn;
+        match->dev = dev;
+
+        match->kvm = kvm;
+
+        list_add(&match->list, &kvm->arch.assigned_dev_head);
+
+        if (assigned_dev->flags & KVM_DEV_ASSIGN_ENABLE_IOMMU) {
+                r = kvm_iommu_map_guest(kvm, match);
+                if (r)
+                        goto out_list_del;
+        }
+
+out:
+        mutex_unlock(&kvm->lock);
+        return r;
+out_list_del:
+        list_del(&match->list);
+        pci_release_regions(dev);
+out_disable:
+        pci_disable_device(dev);
+out_put:
+        pci_dev_put(dev);
+out_free:
+        kfree(match);
+        mutex_unlock(&kvm->lock);
+        return r;
+}
+#endif
+
 static inline int valid_vcpu(int n)
 {
         return likely(n >= 0 && n < KVM_MAX_VCPUS);
@@ -578,12 +818,12 @@ int __kvm_set_memory_region(struct kvm *kvm,
         }
 
         kvm_free_physmem_slot(&old, &new);
-
+#ifdef CONFIG_DMAR
         /* map the pages in iommu page table */
         r = kvm_iommu_map_pages(kvm, base_gfn, npages);
         if (r)
                 goto out;
-
+#endif
         return 0;
 
 out_free:
@@ -1383,6 +1623,30 @@ static long kvm_vm_ioctl(struct file *filp,
                 break;
         }
 #endif
+#ifdef KVM_CAP_DEVICE_ASSIGNMENT
+        case KVM_ASSIGN_PCI_DEVICE: {
+                struct kvm_assigned_pci_dev assigned_dev;
+
+                r = -EFAULT;
+                if (copy_from_user(&assigned_dev, argp, sizeof assigned_dev))
+                        goto out;
+                r = kvm_vm_ioctl_assign_device(kvm, &assigned_dev);
+                if (r)
+                        goto out;
+                break;
+        }
+        case KVM_ASSIGN_IRQ: {
+                struct kvm_assigned_irq assigned_irq;
+
+                r = -EFAULT;
+                if (copy_from_user(&assigned_irq, argp, sizeof assigned_irq))
+                        goto out;
+                r = kvm_vm_ioctl_assign_irq(kvm, &assigned_irq);
+                if (r)
+                        goto out;
+                break;
+        }
+#endif
         default:
                 r = kvm_arch_vm_ioctl(filp, ioctl, arg);
         }