本文将结合qemu与linux源码，解析vhost中ioeventfd与irqfd相关内容。

prerequisite

overview

ioeventfd与kick绑定，irqfd与中断绑定

ioeventfd:

  • qemu利用KVM_IOEVENTFD ioctl，将ioeventfd与guest kick寄存器的地址(pio/mmio地址)和vq index的值绑定，传给kvm
    • 当kvm检测到guest往kick寄存器写入vq index后，写eventfd通知vhost
  • qemu利用VHOST_SET_VRING_KICK ioctl，将ioeventfd传给vhost，vhost就会poll ioeventfd的写
    • 当vhost poll到ioeventfd的写后，就会开始从avai ring中拉取请求，处理完io请求后，更新used ring，最后给guest注入中断(需要借助于irqfd)

irqfd:

  • qemu利用VHOST_SET_VRING_CALL ioctl，将irqfd传给vhost
    • vhost在更新完used ring后，写eventfd通知kvm注入中断
  • qemu利用KVM_IRQFD ioctl，将irqfd与vq的中断绑定，传给kvm，kvm就会poll irqfd的写
    • 当kvm poll到irqfd的写后，就会根据中断路由信息，给guest注入中断

ioeventfd

qemu侧ioeventfd的关联

1
2
3
4
5
6
7
8
9
10
11
12
virtio_bus_start_ioeventfd
└── virtio_device_start_ioeventfd_impl[start_ioeventfd]
    ├── event_notifier_set(&vq->host_notifier)
    └── memory_region_transaction_commit
        └── address_space_update_ioeventfds
            └── address_space_add_del_ioeventfds
                ├── kvm_io_ioeventfd_add[eventfd_add] //pio
                │   └── kvm_set_ioeventfd_pio
                │       └── kvm_vm_ioctl(kvm_state, KVM_IOEVENTFD, &kick)
                └── kvm_mem_ioeventfd_add[eventfd_add] //mmio
                    └── kvm_set_ioeventfd_mmio
                        └── kvm_vm_ioctl(kvm_state, KVM_IOEVENTFD, &iofd)
1
2
3
4
vhost_virtqueue_start
├── event_notifier_get_fd(virtio_queue_get_host_notifier(vvq))
└── vhost_kernel_set_vring_kick[vhost_set_vring_kick]
    └── vhost_kernel_call(dev, VHOST_SET_VRING_KICK, file)
1
2
3
4
5
6
struct VirtQueue
{
        ...
        EventNotifier host_notifier;
        ...
}

由上述代码可知，qemu侧通过host_notifier的ioeventfd，将vhost与kvm关联了起来；

  • vhost负责poll ioeventfd
  • kvm负责写ioeventfd来通知vhost guest的kick操作

kvm侧ioeventfd处理

参考Dive into ioeventfd(KVM side) mechanism即可。

vhost侧ioeventfd处理

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
long vhost_vring_ioctl(struct vhost_dev *d, unsigned int ioctl, void __user *argp)
{
        ...
	switch (ioctl) {
        ...
	case VHOST_SET_VRING_KICK:
		if (copy_from_user(&f, argp, sizeof f)) {
			r = -EFAULT;
			break;
		}
		eventfp = f.fd == VHOST_FILE_UNBIND ? NULL : eventfd_fget(f.fd);
		if (IS_ERR(eventfp)) {
			r = PTR_ERR(eventfp);
			break;
		}
		if (eventfp != vq->kick) {
			pollstop = (filep = vq->kick) != NULL;
			pollstart = (vq->kick = eventfp) != NULL;
		} else
			filep = eventfp;
		break;
        ...
        }
        ...
        if (pollstart && vq->handle_kick)
        r = vhost_poll_start(&vq->poll, vq->kick);
        ...
}

qemu利用VHOST_SET_VRING_KICK ioctl，将ioeventfd传给vhost，然后vhost就开始poll ioeventfd(vhost_poll_start)。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
/* Start polling a file. We add ourselves to file's wait queue. The caller must
 * keep a reference to a file until after vhost_poll_stop is called. */
int vhost_poll_start(struct vhost_poll *poll, struct file *file)
{
	__poll_t mask;

	if (poll->wqh)
		return 0;

	mask = vfs_poll(file, &poll->table); //调用callback vhost_poll_func
	if (mask)
		vhost_poll_wakeup(&poll->wait, 0, 0, poll_to_key(mask));
	if (mask & EPOLLERR) {
		vhost_poll_stop(poll);
		return -EINVAL;
	}

	return 0;
}
EXPORT_SYMBOL_GPL(vhost_poll_start);

当vhost poll到ioeventfd写后，就会触发vhost_poll_wakeup回调。

1
2
3
4
5
6
vhost_poll_wakeup
└── vhost_poll_queue
    └── vhost_vq_work_queue
        └── vhost_worker_queue
            ├── llist_add(&work->node, &worker->work_list)
            └── vhost_task_wake(worker->vtsk)

worker->vtsk又会如何操作呢？且看worker->vtsk的初始化情况以及worker->vtsk的执行函数吧。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
static struct vhost_worker *vhost_worker_create(struct vhost_dev *dev)
{
	struct vhost_worker *worker;
	struct vhost_task *vtsk;
	char name[TASK_COMM_LEN];
	int ret;
	u32 id;

	worker = kzalloc(sizeof(*worker), GFP_KERNEL_ACCOUNT);
	if (!worker)
		return NULL;

	snprintf(name, sizeof(name), "vhost-%d", current->pid);

	vtsk = vhost_task_create(vhost_worker, worker, name);
	if (!vtsk)
		goto free_worker;

	mutex_init(&worker->mutex);
	init_llist_head(&worker->work_list);
	worker->kcov_handle = kcov_common_handle();
	worker->vtsk = vtsk;
	...
}

static bool vhost_worker(void *data)
{
	struct vhost_worker *worker = data;
	struct vhost_work *work, *work_next;
	struct llist_node *node;

	node = llist_del_all(&worker->work_list);
	if (node) {
		__set_current_state(TASK_RUNNING);

		node = llist_reverse_order(node);
		/* make sure flag is seen after deletion */
		smp_wmb();
		llist_for_each_entry_safe(work, work_next, node, node) {
			clear_bit(VHOST_WORK_QUEUED, &work->flags);
			kcov_remote_start_common(worker->kcov_handle);
			work->fn(work); //vq->handle_kick
			kcov_remote_stop();
			cond_resched();
		}
	}

	return !!node;
}

在初始化过程中，vhost会创建一个名为vhost-$pid的内核线程，其中$pid是QEMU进程的pid。该线程被称为“vhost工作线程”。
vhost工作线程的运行函数为vhost_worker，而vhost_worker就会触发vq->handle_kick的回调。

那么vhost_poll与vq->handle_kick等又是如何初始化的呢？

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
void vhost_dev_init(struct vhost_dev *dev,
		    struct vhost_virtqueue **vqs, int nvqs,
		    int iov_limit, int weight, int byte_weight,
		    bool use_worker,
		    int (*msg_handler)(struct vhost_dev *dev, u32 asid,
				       struct vhost_iotlb_msg *msg))
{
	struct vhost_virtqueue *vq;
	int i;

	dev->vqs = vqs;
	dev->nvqs = nvqs;
	mutex_init(&dev->mutex);
	dev->log_ctx = NULL;
	dev->umem = NULL;
	dev->iotlb = NULL;
	dev->mm = NULL;
	dev->iov_limit = iov_limit;
	dev->weight = weight;
	dev->byte_weight = byte_weight;
	dev->use_worker = use_worker;
	dev->msg_handler = msg_handler;
	init_waitqueue_head(&dev->wait);
	INIT_LIST_HEAD(&dev->read_list);
	INIT_LIST_HEAD(&dev->pending_list);
	spin_lock_init(&dev->iotlb_lock);
	xa_init_flags(&dev->worker_xa, XA_FLAGS_ALLOC);

	for (i = 0; i < dev->nvqs; ++i) {
		vq = dev->vqs[i];
		vq->log = NULL;
		vq->indirect = NULL;
		vq->heads = NULL;
		vq->dev = dev;
		mutex_init(&vq->mutex);
		vhost_vq_reset(dev, vq);
		if (vq->handle_kick)
			vhost_poll_init(&vq->poll, vq->handle_kick,
					EPOLLIN, dev, vq);
	}
}

vhost设备初始化时，会为每个vq调用vhost_poll_init来初始化vhost_poll与vq->handle_kick等内容。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
/* Init poll structure */
void vhost_poll_init(struct vhost_poll *poll, vhost_work_fn_t fn,
		     __poll_t mask, struct vhost_dev *dev,
		     struct vhost_virtqueue *vq)
{
	init_waitqueue_func_entry(&poll->wait, vhost_poll_wakeup); //设置wait的回调，在vhost poll到ioeventfd的写后，会触发回调
	init_poll_funcptr(&poll->table, vhost_poll_func); //vhost_poll_func在vfs_poll时会触发回调，加入到file's wait queue中
	poll->mask = mask;
	poll->dev = dev;
	poll->wqh = NULL;
	poll->vq = vq;

	vhost_work_init(&poll->work, fn); //fn为vq->handle_kick，初始化为work->fn，在vhost_worker中会回调
}

static void vhost_poll_func(struct file *file, wait_queue_head_t *wqh, //在vfs_poll时会触发回调，加入到file's wait queue中
			    poll_table *pt)
{
	struct vhost_poll *poll;

	poll = container_of(pt, struct vhost_poll, table);
	poll->wqh = wqh;
	add_wait_queue(wqh, &poll->wait);
}
1
2
3
4
5
6
7
8
9
10
/* The virtqueue structure describes a queue attached to a device. */
struct vhost_virtqueue {
	struct vhost_dev *dev;
	struct vhost_worker __rcu *worker;
	...
	struct vhost_poll poll;
	/* The routine to call when the Guest pings us, or timeout. */
	vhost_work_fn_t handle_kick;
	...
}

总结下vhost侧poll ioeventfd的流程:

  1. vhost_poll_func会让vhost poll ioeventfd，加入到file’s wait queue中
  2. kvm写ioeventfd通知vhost
  3. vhost回调vhost_poll_wakeup，将work加入到workqueue中，唤醒vhost工作线程
  4. vhost工作线程回调vq->handle_kick，处理vq中的io请求

irqfd

qemu侧irqfd的关联

1
2
3
4
5
6
7
virtio_pci_set_guest_notifiers
└── kvm_virtio_pci_vector_use
    └── kvm_virtio_pci_irqfd_use
        ├── virtio_queue_get_guest_notifier
        └── kvm_irqchip_add_irqfd_notifier_gsi
            └── kvm_irqchip_assign_irqfd
                └── kvm_vm_ioctl(s, KVM_IRQFD, &irqfd)
1
2
3
4
5
vhost_virtqueue_start
└── vhost_virtqueue_mask
    ├── event_notifier_get_wfd(virtio_queue_get_guest_notifier(vvq))
    └── vhost_kernel_set_vring_call[vhost_set_vring_call]
        └── vhost_kernel_call(dev, VHOST_SET_VRING_CALL, file)
1
2
3
4
5
6
struct VirtQueue
{
        ...
        EventNotifier guest_notifier
        ...
}

由上述代码可知，qemu侧通过guest_notifier的irqfd，将vhost与kvm关联了起来；

  • kvm负责poll irqfd，然后给vm注入中断
  • vhost在更新完used ring后，写irqfd来通知kvm注入中断

kvm侧irqfd处理

参考Dive into irqfd(KVM side) mechanism即可。

vhost侧irqfd处理

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
long vhost_vring_ioctl(struct vhost_dev *d, unsigned int ioctl, void __user *argp)
{
        ...
	switch (ioctl) {
        ...
	case VHOST_SET_VRING_CALL:
		if (copy_from_user(&f, argp, sizeof f)) {
			r = -EFAULT;
			break;
		}
		ctx = f.fd == VHOST_FILE_UNBIND ? NULL : eventfd_ctx_fdget(f.fd);
		if (IS_ERR(ctx)) {
			r = PTR_ERR(ctx);
			break;
		}

		swap(ctx, vq->call_ctx.ctx);
		break;
        ...
        }
        ...
}

qemu调用VHOST_SET_VRING_CALL，将irqfd传递给vhost

1
2
3
4
5
6
7
8
/* This actually signals the guest, using eventfd. */
void vhost_signal(struct vhost_dev *dev, struct vhost_virtqueue *vq)
{
	/* Signal the Guest tell them we used something up. */
	if (vq->call_ctx.ctx && vhost_notify(dev, vq))
		eventfd_signal(vq->call_ctx.ctx, 1);
}
EXPORT_SYMBOL_GPL(vhost_signal);

vhost在处理完io请求，并更新used ring后，调用vhost_signal，触发irqfd的写；kvm poll到后，就会给vm注入中断。