[mirror_ubuntu-artful-kernel.git] / net / vmw_vsock / virtio_transport.c

/*
 * virtio transport for vsock
 *
 * Copyright (C) 2013-2015 Red Hat, Inc.
 * Author: Asias He <asias@redhat.com>
 *         Stefan Hajnoczi <stefanha@redhat.com>
 *
 * Some of the code is take from Gerd Hoffmann <kraxel@redhat.com>'s
 * early virtio-vsock proof-of-concept bits.
 *
 * This work is licensed under the terms of the GNU GPL, version 2.
 */
#include <linux/spinlock.h>
#include <linux/module.h>
#include <linux/list.h>
#include <linux/atomic.h>
#include <linux/virtio.h>
#include <linux/virtio_ids.h>
#include <linux/virtio_config.h>
#include <linux/virtio_vsock.h>
#include <net/sock.h>
#include <linux/mutex.h>
#include <net/af_vsock.h>

static struct workqueue_struct *virtio_vsock_workqueue;
static struct virtio_vsock *the_virtio_vsock;
static DEFINE_MUTEX(the_virtio_vsock_mutex); /* protects the_virtio_vsock */

struct virtio_vsock {
	struct virtio_device *vdev;
	struct virtqueue *vqs[VSOCK_VQ_MAX];

	/* Virtqueue processing is deferred to a workqueue */
	struct work_struct tx_work;
	struct work_struct rx_work;
	struct work_struct event_work;

	/* The following fields are protected by tx_lock.  vqs[VSOCK_VQ_TX]
	 * must be accessed with tx_lock held.
	 */
	struct mutex tx_lock;

	struct work_struct send_pkt_work;
	spinlock_t send_pkt_list_lock;
	struct list_head send_pkt_list;

	struct work_struct loopback_work;
	spinlock_t loopback_list_lock; /* protects loopback_list */
	struct list_head loopback_list;

	atomic_t queued_replies;

	/* The following fields are protected by rx_lock.  vqs[VSOCK_VQ_RX]
	 * must be accessed with rx_lock held.
	 */
	struct mutex rx_lock;
	int rx_buf_nr;
	int rx_buf_max_nr;

	/* The following fields are protected by event_lock.
	 * vqs[VSOCK_VQ_EVENT] must be accessed with event_lock held.
	 */
	struct mutex event_lock;
	struct virtio_vsock_event event_list[8];

	u32 guest_cid;
};

static struct virtio_vsock *virtio_vsock_get(void)
{
	return the_virtio_vsock;
}

static u32 virtio_transport_get_local_cid(void)
{
	struct virtio_vsock *vsock = virtio_vsock_get();

	return vsock->guest_cid;
}

static void virtio_transport_loopback_work(struct work_struct *work)
{
	struct virtio_vsock *vsock =
		container_of(work, struct virtio_vsock, loopback_work);
	LIST_HEAD(pkts);

	spin_lock_bh(&vsock->loopback_list_lock);
	list_splice_init(&vsock->loopback_list, &pkts);
	spin_unlock_bh(&vsock->loopback_list_lock);

	mutex_lock(&vsock->rx_lock);
	while (!list_empty(&pkts)) {
		struct virtio_vsock_pkt *pkt;

		pkt = list_first_entry(&pkts, struct virtio_vsock_pkt, list);
		list_del_init(&pkt->list);

		virtio_transport_recv_pkt(pkt);
	}
	mutex_unlock(&vsock->rx_lock);
}

static int virtio_transport_send_pkt_loopback(struct virtio_vsock *vsock,
					      struct virtio_vsock_pkt *pkt)
{
	int len = pkt->len;

	spin_lock_bh(&vsock->loopback_list_lock);
	list_add_tail(&pkt->list, &vsock->loopback_list);
	spin_unlock_bh(&vsock->loopback_list_lock);

	queue_work(virtio_vsock_workqueue, &vsock->loopback_work);

	return len;
}

static void
virtio_transport_send_pkt_work(struct work_struct *work)
{
	struct virtio_vsock *vsock =
		container_of(work, struct virtio_vsock, send_pkt_work);
	struct virtqueue *vq;
	bool added = false;
	bool restart_rx = false;

	mutex_lock(&vsock->tx_lock);

	vq = vsock->vqs[VSOCK_VQ_TX];

	for (;;) {
		struct virtio_vsock_pkt *pkt;
		struct scatterlist hdr, buf, *sgs[2];
		int ret, in_sg = 0, out_sg = 0;
		bool reply;

		spin_lock_bh(&vsock->send_pkt_list_lock);
		if (list_empty(&vsock->send_pkt_list)) {
			spin_unlock_bh(&vsock->send_pkt_list_lock);
			break;
		}

		pkt = list_first_entry(&vsock->send_pkt_list,
				       struct virtio_vsock_pkt, list);
		list_del_init(&pkt->list);
		spin_unlock_bh(&vsock->send_pkt_list_lock);

		reply = pkt->reply;

		sg_init_one(&hdr, &pkt->hdr, sizeof(pkt->hdr));
		sgs[out_sg++] = &hdr;
		if (pkt->buf) {
			sg_init_one(&buf, pkt->buf, pkt->len);
			sgs[out_sg++] = &buf;
		}

		ret = virtqueue_add_sgs(vq, sgs, out_sg, in_sg, pkt, GFP_KERNEL);
		/* Usually this means that there is no more space available in
		 * the vq
		 */
		if (ret < 0) {
			spin_lock_bh(&vsock->send_pkt_list_lock);
			list_add(&pkt->list, &vsock->send_pkt_list);
			spin_unlock_bh(&vsock->send_pkt_list_lock);
			break;
		}

		if (reply) {
			struct virtqueue *rx_vq = vsock->vqs[VSOCK_VQ_RX];
			int val;

			val = atomic_dec_return(&vsock->queued_replies);

			/* Do we now have resources to resume rx processing? */
			if (val + 1 == virtqueue_get_vring_size(rx_vq))
				restart_rx = true;
		}

		added = true;
	}

	if (added)
		virtqueue_kick(vq);

	mutex_unlock(&vsock->tx_lock);

	if (restart_rx)
		queue_work(virtio_vsock_workqueue, &vsock->rx_work);
}

static int
virtio_transport_send_pkt(struct virtio_vsock_pkt *pkt)
{
	struct virtio_vsock *vsock;
	int len = pkt->len;

	vsock = virtio_vsock_get();
	if (!vsock) {
		virtio_transport_free_pkt(pkt);
		return -ENODEV;
	}

	if (le32_to_cpu(pkt->hdr.dst_cid) == vsock->guest_cid)
		return virtio_transport_send_pkt_loopback(vsock, pkt);

	if (pkt->reply)
		atomic_inc(&vsock->queued_replies);

	spin_lock_bh(&vsock->send_pkt_list_lock);
	list_add_tail(&pkt->list, &vsock->send_pkt_list);
	spin_unlock_bh(&vsock->send_pkt_list_lock);

	queue_work(virtio_vsock_workqueue, &vsock->send_pkt_work);
	return len;
}

static void virtio_vsock_rx_fill(struct virtio_vsock *vsock)
{
	int buf_len = VIRTIO_VSOCK_DEFAULT_RX_BUF_SIZE;
	struct virtio_vsock_pkt *pkt;
	struct scatterlist hdr, buf, *sgs[2];
	struct virtqueue *vq;
	int ret;

	vq = vsock->vqs[VSOCK_VQ_RX];

	do {
		pkt = kzalloc(sizeof(*pkt), GFP_KERNEL);
		if (!pkt)
			break;

		pkt->buf = kmalloc(buf_len, GFP_KERNEL);
		if (!pkt->buf) {
			virtio_transport_free_pkt(pkt);
			break;
		}

		pkt->len = buf_len;

		sg_init_one(&hdr, &pkt->hdr, sizeof(pkt->hdr));
		sgs[0] = &hdr;

		sg_init_one(&buf, pkt->buf, buf_len);
		sgs[1] = &buf;
		ret = virtqueue_add_sgs(vq, sgs, 0, 2, pkt, GFP_KERNEL);
		if (ret) {
			virtio_transport_free_pkt(pkt);
			break;
		}
		vsock->rx_buf_nr++;
	} while (vq->num_free);
	if (vsock->rx_buf_nr > vsock->rx_buf_max_nr)
		vsock->rx_buf_max_nr = vsock->rx_buf_nr;
	virtqueue_kick(vq);
}

static void virtio_transport_tx_work(struct work_struct *work)
{
	struct virtio_vsock *vsock =
		container_of(work, struct virtio_vsock, tx_work);
	struct virtqueue *vq;
	bool added = false;

	vq = vsock->vqs[VSOCK_VQ_TX];
	mutex_lock(&vsock->tx_lock);
	do {
		struct virtio_vsock_pkt *pkt;
		unsigned int len;

		virtqueue_disable_cb(vq);
		while ((pkt = virtqueue_get_buf(vq, &len)) != NULL) {
			virtio_transport_free_pkt(pkt);
			added = true;
		}
	} while (!virtqueue_enable_cb(vq));
	mutex_unlock(&vsock->tx_lock);

	if (added)
		queue_work(virtio_vsock_workqueue, &vsock->send_pkt_work);
}

/* Is there space left for replies to rx packets? */
static bool virtio_transport_more_replies(struct virtio_vsock *vsock)
{
	struct virtqueue *vq = vsock->vqs[VSOCK_VQ_RX];
	int val;

	smp_rmb(); /* paired with atomic_inc() and atomic_dec_return() */
	val = atomic_read(&vsock->queued_replies);

	return val < virtqueue_get_vring_size(vq);
}

static void virtio_transport_rx_work(struct work_struct *work)
{
	struct virtio_vsock *vsock =
		container_of(work, struct virtio_vsock, rx_work);
	struct virtqueue *vq;

	vq = vsock->vqs[VSOCK_VQ_RX];

	mutex_lock(&vsock->rx_lock);

	do {
		virtqueue_disable_cb(vq);
		for (;;) {
			struct virtio_vsock_pkt *pkt;
			unsigned int len;

			if (!virtio_transport_more_replies(vsock)) {
				/* Stop rx until the device processes already
				 * pending replies.  Leave rx virtqueue
				 * callbacks disabled.
				 */
				goto out;
			}

			pkt = virtqueue_get_buf(vq, &len);
			if (!pkt) {
				break;
			}

			vsock->rx_buf_nr--;

			/* Drop short/long packets */
			if (unlikely(len < sizeof(pkt->hdr) ||
				     len > sizeof(pkt->hdr) + pkt->len)) {
				virtio_transport_free_pkt(pkt);
				continue;
			}

			pkt->len = len - sizeof(pkt->hdr);
			virtio_transport_recv_pkt(pkt);
		}
	} while (!virtqueue_enable_cb(vq));

out:
	if (vsock->rx_buf_nr < vsock->rx_buf_max_nr / 2)
		virtio_vsock_rx_fill(vsock);
	mutex_unlock(&vsock->rx_lock);
}

/* event_lock must be held */
static int virtio_vsock_event_fill_one(struct virtio_vsock *vsock,
				       struct virtio_vsock_event *event)
{
	struct scatterlist sg;
	struct virtqueue *vq;

	vq = vsock->vqs[VSOCK_VQ_EVENT];

	sg_init_one(&sg, event, sizeof(*event));

	return virtqueue_add_inbuf(vq, &sg, 1, event, GFP_KERNEL);
}

/* event_lock must be held */
static void virtio_vsock_event_fill(struct virtio_vsock *vsock)
{
	size_t i;

	for (i = 0; i < ARRAY_SIZE(vsock->event_list); i++) {
		struct virtio_vsock_event *event = &vsock->event_list[i];

		virtio_vsock_event_fill_one(vsock, event);
	}

	virtqueue_kick(vsock->vqs[VSOCK_VQ_EVENT]);
}

static void virtio_vsock_reset_sock(struct sock *sk)
{
	lock_sock(sk);
	sk->sk_state = SS_UNCONNECTED;
	sk->sk_err = ECONNRESET;
	sk->sk_error_report(sk);
	release_sock(sk);
}

static void virtio_vsock_update_guest_cid(struct virtio_vsock *vsock)
{
	struct virtio_device *vdev = vsock->vdev;
	__le64 guest_cid;

	vdev->config->get(vdev, offsetof(struct virtio_vsock_config, guest_cid),
			  &guest_cid, sizeof(guest_cid));
	vsock->guest_cid = le64_to_cpu(guest_cid);
}

/* event_lock must be held */
static void virtio_vsock_event_handle(struct virtio_vsock *vsock,
				      struct virtio_vsock_event *event)
{
	switch (le32_to_cpu(event->id)) {
	case VIRTIO_VSOCK_EVENT_TRANSPORT_RESET:
		virtio_vsock_update_guest_cid(vsock);
		vsock_for_each_connected_socket(virtio_vsock_reset_sock);
		break;
	}
}

static void virtio_transport_event_work(struct work_struct *work)
{
	struct virtio_vsock *vsock =
		container_of(work, struct virtio_vsock, event_work);
	struct virtqueue *vq;

	vq = vsock->vqs[VSOCK_VQ_EVENT];

	mutex_lock(&vsock->event_lock);

	do {
		struct virtio_vsock_event *event;
		unsigned int len;

		virtqueue_disable_cb(vq);
		while ((event = virtqueue_get_buf(vq, &len)) != NULL) {
			if (len == sizeof(*event))
				virtio_vsock_event_handle(vsock, event);

			virtio_vsock_event_fill_one(vsock, event);
		}
	} while (!virtqueue_enable_cb(vq));

	virtqueue_kick(vsock->vqs[VSOCK_VQ_EVENT]);

	mutex_unlock(&vsock->event_lock);
}

static void virtio_vsock_event_done(struct virtqueue *vq)
{
	struct virtio_vsock *vsock = vq->vdev->priv;

	if (!vsock)
		return;
	queue_work(virtio_vsock_workqueue, &vsock->event_work);
}

static void virtio_vsock_tx_done(struct virtqueue *vq)
{
	struct virtio_vsock *vsock = vq->vdev->priv;

	if (!vsock)
		return;
	queue_work(virtio_vsock_workqueue, &vsock->tx_work);
}

static void virtio_vsock_rx_done(struct virtqueue *vq)
{
	struct virtio_vsock *vsock = vq->vdev->priv;

	if (!vsock)
		return;
	queue_work(virtio_vsock_workqueue, &vsock->rx_work);
}

static struct virtio_transport virtio_transport = {
	.transport = {
		.get_local_cid            = virtio_transport_get_local_cid,

		.init                     = virtio_transport_do_socket_init,
		.destruct                 = virtio_transport_destruct,
		.release                  = virtio_transport_release,
		.connect                  = virtio_transport_connect,
		.shutdown                 = virtio_transport_shutdown,

		.dgram_bind               = virtio_transport_dgram_bind,
		.dgram_dequeue            = virtio_transport_dgram_dequeue,
		.dgram_enqueue            = virtio_transport_dgram_enqueue,
		.dgram_allow              = virtio_transport_dgram_allow,

		.stream_dequeue           = virtio_transport_stream_dequeue,
		.stream_enqueue           = virtio_transport_stream_enqueue,
		.stream_has_data          = virtio_transport_stream_has_data,
		.stream_has_space         = virtio_transport_stream_has_space,
		.stream_rcvhiwat          = virtio_transport_stream_rcvhiwat,
		.stream_is_active         = virtio_transport_stream_is_active,
		.stream_allow             = virtio_transport_stream_allow,

		.notify_poll_in           = virtio_transport_notify_poll_in,
		.notify_poll_out          = virtio_transport_notify_poll_out,
		.notify_recv_init         = virtio_transport_notify_recv_init,
		.notify_recv_pre_block    = virtio_transport_notify_recv_pre_block,
		.notify_recv_pre_dequeue  = virtio_transport_notify_recv_pre_dequeue,
		.notify_recv_post_dequeue = virtio_transport_notify_recv_post_dequeue,
		.notify_send_init         = virtio_transport_notify_send_init,
		.notify_send_pre_block    = virtio_transport_notify_send_pre_block,
		.notify_send_pre_enqueue  = virtio_transport_notify_send_pre_enqueue,
		.notify_send_post_enqueue = virtio_transport_notify_send_post_enqueue,

		.set_buffer_size          = virtio_transport_set_buffer_size,
		.set_min_buffer_size      = virtio_transport_set_min_buffer_size,
		.set_max_buffer_size      = virtio_transport_set_max_buffer_size,
		.get_buffer_size          = virtio_transport_get_buffer_size,
		.get_min_buffer_size      = virtio_transport_get_min_buffer_size,
		.get_max_buffer_size      = virtio_transport_get_max_buffer_size,
	},

	.send_pkt = virtio_transport_send_pkt,
};

static int virtio_vsock_probe(struct virtio_device *vdev)
{
	vq_callback_t *callbacks[] = {
		virtio_vsock_rx_done,
		virtio_vsock_tx_done,
		virtio_vsock_event_done,
	};
	static const char * const names[] = {
		"rx",
		"tx",
		"event",
	};
	struct virtio_vsock *vsock = NULL;
	int ret;

	ret = mutex_lock_interruptible(&the_virtio_vsock_mutex);
	if (ret)
		return ret;

	/* Only one virtio-vsock device per guest is supported */
	if (the_virtio_vsock) {
		ret = -EBUSY;
		goto out;
	}

	vsock = kzalloc(sizeof(*vsock), GFP_KERNEL);
	if (!vsock) {
		ret = -ENOMEM;
		goto out;
	}

	vsock->vdev = vdev;

	ret = vsock->vdev->config->find_vqs(vsock->vdev, VSOCK_VQ_MAX,
					    vsock->vqs, callbacks, names,
					    NULL);
	if (ret < 0)
		goto out;

	virtio_vsock_update_guest_cid(vsock);

	ret = vsock_core_init(&virtio_transport.transport);
	if (ret < 0)
		goto out_vqs;

	vsock->rx_buf_nr = 0;
	vsock->rx_buf_max_nr = 0;
	atomic_set(&vsock->queued_replies, 0);

	vdev->priv = vsock;
	the_virtio_vsock = vsock;
	mutex_init(&vsock->tx_lock);
	mutex_init(&vsock->rx_lock);
	mutex_init(&vsock->event_lock);
	spin_lock_init(&vsock->send_pkt_list_lock);
	INIT_LIST_HEAD(&vsock->send_pkt_list);
	spin_lock_init(&vsock->loopback_list_lock);
	INIT_LIST_HEAD(&vsock->loopback_list);
	INIT_WORK(&vsock->rx_work, virtio_transport_rx_work);
	INIT_WORK(&vsock->tx_work, virtio_transport_tx_work);
	INIT_WORK(&vsock->event_work, virtio_transport_event_work);
	INIT_WORK(&vsock->send_pkt_work, virtio_transport_send_pkt_work);
	INIT_WORK(&vsock->loopback_work, virtio_transport_loopback_work);

	mutex_lock(&vsock->rx_lock);
	virtio_vsock_rx_fill(vsock);
	mutex_unlock(&vsock->rx_lock);

	mutex_lock(&vsock->event_lock);
	virtio_vsock_event_fill(vsock);
	mutex_unlock(&vsock->event_lock);

	mutex_unlock(&the_virtio_vsock_mutex);
	return 0;

out_vqs:
	vsock->vdev->config->del_vqs(vsock->vdev);
out:
	kfree(vsock);
	mutex_unlock(&the_virtio_vsock_mutex);
	return ret;
}

static void virtio_vsock_remove(struct virtio_device *vdev)
{
	struct virtio_vsock *vsock = vdev->priv;
	struct virtio_vsock_pkt *pkt;

	flush_work(&vsock->loopback_work);
	flush_work(&vsock->rx_work);
	flush_work(&vsock->tx_work);
	flush_work(&vsock->event_work);
	flush_work(&vsock->send_pkt_work);

	vdev->config->reset(vdev);

	mutex_lock(&vsock->rx_lock);
	while ((pkt = virtqueue_detach_unused_buf(vsock->vqs[VSOCK_VQ_RX])))
		virtio_transport_free_pkt(pkt);
	mutex_unlock(&vsock->rx_lock);

	mutex_lock(&vsock->tx_lock);
	while ((pkt = virtqueue_detach_unused_buf(vsock->vqs[VSOCK_VQ_TX])))
		virtio_transport_free_pkt(pkt);
	mutex_unlock(&vsock->tx_lock);

	spin_lock_bh(&vsock->send_pkt_list_lock);
	while (!list_empty(&vsock->send_pkt_list)) {
		pkt = list_first_entry(&vsock->send_pkt_list,
				       struct virtio_vsock_pkt, list);
		list_del(&pkt->list);
		virtio_transport_free_pkt(pkt);
	}
	spin_unlock_bh(&vsock->send_pkt_list_lock);

	spin_lock_bh(&vsock->loopback_list_lock);
	while (!list_empty(&vsock->loopback_list)) {
		pkt = list_first_entry(&vsock->loopback_list,
				       struct virtio_vsock_pkt, list);
		list_del(&pkt->list);
		virtio_transport_free_pkt(pkt);
	}
	spin_unlock_bh(&vsock->loopback_list_lock);

	mutex_lock(&the_virtio_vsock_mutex);
	the_virtio_vsock = NULL;
	vsock_core_exit();
	mutex_unlock(&the_virtio_vsock_mutex);

	vdev->config->del_vqs(vdev);

	kfree(vsock);
}

static struct virtio_device_id id_table[] = {
	{ VIRTIO_ID_VSOCK, VIRTIO_DEV_ANY_ID },
	{ 0 },
};

static unsigned int features[] = {
};

static struct virtio_driver virtio_vsock_driver = {
	.feature_table = features,
	.feature_table_size = ARRAY_SIZE(features),
	.driver.name = KBUILD_MODNAME,
	.driver.owner = THIS_MODULE,
	.id_table = id_table,
	.probe = virtio_vsock_probe,
	.remove = virtio_vsock_remove,
};

static int __init virtio_vsock_init(void)
{
	int ret;

	virtio_vsock_workqueue = alloc_workqueue("virtio_vsock", 0, 0);
	if (!virtio_vsock_workqueue)
		return -ENOMEM;
	ret = register_virtio_driver(&virtio_vsock_driver);
	if (ret)
		destroy_workqueue(virtio_vsock_workqueue);
	return ret;
}

static void __exit virtio_vsock_exit(void)
{
	unregister_virtio_driver(&virtio_vsock_driver);
	destroy_workqueue(virtio_vsock_workqueue);
}

module_init(virtio_vsock_init);
module_exit(virtio_vsock_exit);
MODULE_LICENSE("GPL v2");
MODULE_AUTHOR("Asias He");
MODULE_DESCRIPTION("virtio transport for vsock");
MODULE_DEVICE_TABLE(virtio, id_table);
Commit	Line	Data
0ea9e1d3 AH	1	/*
	2	* virtio transport for vsock
	3	*
	4	* Copyright (C) 2013-2015 Red Hat, Inc.
	5	* Author: Asias He <asias@redhat.com>
	6	* Stefan Hajnoczi <stefanha@redhat.com>
	7	*
	8	* Some of the code is take from Gerd Hoffmann <kraxel@redhat.com>'s
	9	* early virtio-vsock proof-of-concept bits.
	10	*
	11	* This work is licensed under the terms of the GNU GPL, version 2.
	12	*/
	13	#include <linux/spinlock.h>
	14	#include <linux/module.h>
	15	#include <linux/list.h>
	16	#include <linux/atomic.h>
	17	#include <linux/virtio.h>
	18	#include <linux/virtio_ids.h>
	19	#include <linux/virtio_config.h>
	20	#include <linux/virtio_vsock.h>
	21	#include <net/sock.h>
	22	#include <linux/mutex.h>
	23	#include <net/af_vsock.h>
	24
	25	static struct workqueue_struct *virtio_vsock_workqueue;
	26	static struct virtio_vsock *the_virtio_vsock;
	27	static DEFINE_MUTEX(the_virtio_vsock_mutex); /* protects the_virtio_vsock */
	28
	29	struct virtio_vsock {
	30	struct virtio_device *vdev;
	31	struct virtqueue *vqs[VSOCK_VQ_MAX];
	32
	33	/* Virtqueue processing is deferred to a workqueue */
	34	struct work_struct tx_work;
	35	struct work_struct rx_work;
	36	struct work_struct event_work;
	37
	38	/* The following fields are protected by tx_lock. vqs[VSOCK_VQ_TX]
	39	* must be accessed with tx_lock held.
	40	*/
	41	struct mutex tx_lock;
	42
	43	struct work_struct send_pkt_work;
	44	spinlock_t send_pkt_list_lock;
	45	struct list_head send_pkt_list;
	46
b9116823 SH	47	struct work_struct loopback_work;
	48	spinlock_t loopback_list_lock; /* protects loopback_list */
	49	struct list_head loopback_list;
	50
0ea9e1d3 AH	51	atomic_t queued_replies;
	52
	53	/* The following fields are protected by rx_lock. vqs[VSOCK_VQ_RX]
	54	* must be accessed with rx_lock held.
	55	*/
	56	struct mutex rx_lock;
	57	int rx_buf_nr;
	58	int rx_buf_max_nr;
	59
	60	/* The following fields are protected by event_lock.
	61	* vqs[VSOCK_VQ_EVENT] must be accessed with event_lock held.
	62	*/
	63	struct mutex event_lock;
	64	struct virtio_vsock_event event_list[8];
	65
	66	u32 guest_cid;
	67	};
	68
	69	static struct virtio_vsock *virtio_vsock_get(void)
	70	{
	71	return the_virtio_vsock;
	72	}
	73
	74	static u32 virtio_transport_get_local_cid(void)
	75	{
	76	struct virtio_vsock *vsock = virtio_vsock_get();
	77
	78	return vsock->guest_cid;
	79	}
	80
b9116823 SH	81	static void virtio_transport_loopback_work(struct work_struct *work)
	82	{
	83	struct virtio_vsock *vsock =
	84	container_of(work, struct virtio_vsock, loopback_work);
	85	LIST_HEAD(pkts);
	86
	87	spin_lock_bh(&vsock->loopback_list_lock);
	88	list_splice_init(&vsock->loopback_list, &pkts);
	89	spin_unlock_bh(&vsock->loopback_list_lock);
	90
	91	mutex_lock(&vsock->rx_lock);
	92	while (!list_empty(&pkts)) {
	93	struct virtio_vsock_pkt *pkt;
	94
	95	pkt = list_first_entry(&pkts, struct virtio_vsock_pkt, list);
	96	list_del_init(&pkt->list);
	97
	98	virtio_transport_recv_pkt(pkt);
	99	}
	100	mutex_unlock(&vsock->rx_lock);
	101	}
	102
	103	static int virtio_transport_send_pkt_loopback(struct virtio_vsock *vsock,
	104	struct virtio_vsock_pkt *pkt)
	105	{
	106	int len = pkt->len;
	107
	108	spin_lock_bh(&vsock->loopback_list_lock);
	109	list_add_tail(&pkt->list, &vsock->loopback_list);
	110	spin_unlock_bh(&vsock->loopback_list_lock);
	111
	112	queue_work(virtio_vsock_workqueue, &vsock->loopback_work);
	113
	114	return len;
	115	}
	116
0ea9e1d3 AH	117	static void
	118	virtio_transport_send_pkt_work(struct work_struct *work)
	119	{
	120	struct virtio_vsock *vsock =
	121	container_of(work, struct virtio_vsock, send_pkt_work);
	122	struct virtqueue *vq;
	123	bool added = false;
	124	bool restart_rx = false;
	125
	126	mutex_lock(&vsock->tx_lock);
	127
	128	vq = vsock->vqs[VSOCK_VQ_TX];
	129
0ea9e1d3 AH	130	for (;;) {
	131	struct virtio_vsock_pkt *pkt;
	132	struct scatterlist hdr, buf, *sgs[2];
	133	int ret, in_sg = 0, out_sg = 0;
	134	bool reply;
	135
	136	spin_lock_bh(&vsock->send_pkt_list_lock);
	137	if (list_empty(&vsock->send_pkt_list)) {
	138	spin_unlock_bh(&vsock->send_pkt_list_lock);
0ea9e1d3 AH	139	break;
	140	}
	141
	142	pkt = list_first_entry(&vsock->send_pkt_list,
	143	struct virtio_vsock_pkt, list);
	144	list_del_init(&pkt->list);
	145	spin_unlock_bh(&vsock->send_pkt_list_lock);
	146
	147	reply = pkt->reply;
	148
	149	sg_init_one(&hdr, &pkt->hdr, sizeof(pkt->hdr));
	150	sgs[out_sg++] = &hdr;
	151	if (pkt->buf) {
	152	sg_init_one(&buf, pkt->buf, pkt->len);
	153	sgs[out_sg++] = &buf;
	154	}
	155
	156	ret = virtqueue_add_sgs(vq, sgs, out_sg, in_sg, pkt, GFP_KERNEL);
21bc54fc GG	157	/* Usually this means that there is no more space available in
	158	* the vq
	159	*/
0ea9e1d3 AH	160	if (ret < 0) {
	161	spin_lock_bh(&vsock->send_pkt_list_lock);
	162	list_add(&pkt->list, &vsock->send_pkt_list);
	163	spin_unlock_bh(&vsock->send_pkt_list_lock);
0ea9e1d3 AH	164	break;
	165	}
	166
	167	if (reply) {
	168	struct virtqueue *rx_vq = vsock->vqs[VSOCK_VQ_RX];
	169	int val;
	170
	171	val = atomic_dec_return(&vsock->queued_replies);
	172
	173	/* Do we now have resources to resume rx processing? */
	174	if (val + 1 == virtqueue_get_vring_size(rx_vq))
	175	restart_rx = true;
	176	}
	177
	178	added = true;
	179	}
	180
	181	if (added)
	182	virtqueue_kick(vq);
	183
	184	mutex_unlock(&vsock->tx_lock);
	185
	186	if (restart_rx)
	187	queue_work(virtio_vsock_workqueue, &vsock->rx_work);
	188	}
	189
	190	static int
	191	virtio_transport_send_pkt(struct virtio_vsock_pkt *pkt)
	192	{
	193	struct virtio_vsock *vsock;
	194	int len = pkt->len;
	195
	196	vsock = virtio_vsock_get();
	197	if (!vsock) {
	198	virtio_transport_free_pkt(pkt);
	199	return -ENODEV;
	200	}
	201
b9116823 SH	202	if (le32_to_cpu(pkt->hdr.dst_cid) == vsock->guest_cid)
	203	return virtio_transport_send_pkt_loopback(vsock, pkt);
	204
0ea9e1d3 AH	205	if (pkt->reply)
	206	atomic_inc(&vsock->queued_replies);
	207
	208	spin_lock_bh(&vsock->send_pkt_list_lock);
	209	list_add_tail(&pkt->list, &vsock->send_pkt_list);
	210	spin_unlock_bh(&vsock->send_pkt_list_lock);
	211
	212	queue_work(virtio_vsock_workqueue, &vsock->send_pkt_work);
	213	return len;
	214	}
	215
	216	static void virtio_vsock_rx_fill(struct virtio_vsock *vsock)
	217	{
	218	int buf_len = VIRTIO_VSOCK_DEFAULT_RX_BUF_SIZE;
	219	struct virtio_vsock_pkt *pkt;
	220	struct scatterlist hdr, buf, *sgs[2];
	221	struct virtqueue *vq;
	222	int ret;
	223
	224	vq = vsock->vqs[VSOCK_VQ_RX];
	225
	226	do {
	227	pkt = kzalloc(sizeof(*pkt), GFP_KERNEL);
	228	if (!pkt)
	229	break;
	230
	231	pkt->buf = kmalloc(buf_len, GFP_KERNEL);
	232	if (!pkt->buf) {
	233	virtio_transport_free_pkt(pkt);
	234	break;
	235	}
	236
	237	pkt->len = buf_len;
	238
	239	sg_init_one(&hdr, &pkt->hdr, sizeof(pkt->hdr));
	240	sgs[0] = &hdr;
	241
	242	sg_init_one(&buf, pkt->buf, buf_len);
	243	sgs[1] = &buf;
	244	ret = virtqueue_add_sgs(vq, sgs, 0, 2, pkt, GFP_KERNEL);
	245	if (ret) {
	246	virtio_transport_free_pkt(pkt);
	247	break;
	248	}
	249	vsock->rx_buf_nr++;
	250	} while (vq->num_free);
	251	if (vsock->rx_buf_nr > vsock->rx_buf_max_nr)
	252	vsock->rx_buf_max_nr = vsock->rx_buf_nr;
	253	virtqueue_kick(vq);
	254	}
	255
	256	static void virtio_transport_tx_work(struct work_struct *work)
	257	{
	258	struct virtio_vsock *vsock =
	259	container_of(work, struct virtio_vsock, tx_work);
	260	struct virtqueue *vq;
	261	bool added = false;
	262
	263	vq = vsock->vqs[VSOCK_VQ_TX];
	264	mutex_lock(&vsock->tx_lock);
	265	do {
	266	struct virtio_vsock_pkt *pkt;
	267	unsigned int len;
	268
269	virtqueue_disable_cb(vq);
270	while ((pkt = virtqueue_get_buf(vq, &len)) != NULL) {
271	virtio_transport_free_pkt(pkt);
272	added = true;
273	}
274	} while (!virtqueue_enable_cb(vq));
275	mutex_unlock(&vsock->tx_lock);
276
277	if (added)
278	queue_work(virtio_vsock_workqueue, &vsock->send_pkt_work);
279	}
280
281	/* Is there space left for replies to rx packets? */
282	static bool virtio_transport_more_replies(struct virtio_vsock *vsock)
283	{
284	struct virtqueue *vq = vsock->vqs[VSOCK_VQ_RX];
285	int val;
286
287	smp_rmb(); /* paired with atomic_inc() and atomic_dec_return() */
288	val = atomic_read(&vsock->queued_replies);
289
290	return val < virtqueue_get_vring_size(vq);
291	}
292
293	static void virtio_transport_rx_work(struct work_struct *work)
294	{
295	struct virtio_vsock *vsock =
296	container_of(work, struct virtio_vsock, rx_work);
297	struct virtqueue *vq;
298
299	vq = vsock->vqs[VSOCK_VQ_RX];
300
301	mutex_lock(&vsock->rx_lock);
302
303	do {
304	virtqueue_disable_cb(vq);
305	for (;;) {
306	struct virtio_vsock_pkt *pkt;
307	unsigned int len;
308
309	if (!virtio_transport_more_replies(vsock)) {
310	/* Stop rx until the device processes already
311	* pending replies. Leave rx virtqueue
312	* callbacks disabled.
313	*/
314	goto out;
315	}
316
317	pkt = virtqueue_get_buf(vq, &len);
318	if (!pkt) {
319	break;
320	}
321
322	vsock->rx_buf_nr--;
323
324	/* Drop short/long packets */
325	if (unlikely(len < sizeof(pkt->hdr) \|\|
326	len > sizeof(pkt->hdr) + pkt->len)) {
327	virtio_transport_free_pkt(pkt);
328	continue;
329	}
330
331	pkt->len = len - sizeof(pkt->hdr);
332	virtio_transport_recv_pkt(pkt);
333	}
334	} while (!virtqueue_enable_cb(vq));
335
336	out:
337	if (vsock->rx_buf_nr < vsock->rx_buf_max_nr / 2)
338	virtio_vsock_rx_fill(vsock);
339	mutex_unlock(&vsock->rx_lock);
340	}
341
342	/* event_lock must be held */
343	static int virtio_vsock_event_fill_one(struct virtio_vsock *vsock,
344	struct virtio_vsock_event *event)
345	{
346	struct scatterlist sg;
347	struct virtqueue *vq;
348
349	vq = vsock->vqs[VSOCK_VQ_EVENT];
350
351	sg_init_one(&sg, event, sizeof(*event));
352
353	return virtqueue_add_inbuf(vq, &sg, 1, event, GFP_KERNEL);
354	}
355
356	/* event_lock must be held */
357	static void virtio_vsock_event_fill(struct virtio_vsock *vsock)
358	{
359	size_t i;
360
361	for (i = 0; i < ARRAY_SIZE(vsock->event_list); i++) {
362	struct virtio_vsock_event *event = &vsock->event_list[i];
363
364	virtio_vsock_event_fill_one(vsock, event);
365	}
366
367	virtqueue_kick(vsock->vqs[VSOCK_VQ_EVENT]);
368	}
369
370	static void virtio_vsock_reset_sock(struct sock *sk)
371	{
372	lock_sock(sk);
373	sk->sk_state = SS_UNCONNECTED;
374	sk->sk_err = ECONNRESET;
375	sk->sk_error_report(sk);
376	release_sock(sk);
377	}
378
379	static void virtio_vsock_update_guest_cid(struct virtio_vsock *vsock)
380	{
381	struct virtio_device *vdev = vsock->vdev;
6c7efafd	382	__le64 guest_cid;
0ea9e1d3 AH	383
	384	vdev->config->get(vdev, offsetof(struct virtio_vsock_config, guest_cid),
	385	&guest_cid, sizeof(guest_cid));
	386	vsock->guest_cid = le64_to_cpu(guest_cid);
	387	}
	388
	389	/* event_lock must be held */
	390	static void virtio_vsock_event_handle(struct virtio_vsock *vsock,
	391	struct virtio_vsock_event *event)
	392	{
	393	switch (le32_to_cpu(event->id)) {
	394	case VIRTIO_VSOCK_EVENT_TRANSPORT_RESET:
	395	virtio_vsock_update_guest_cid(vsock);
	396	vsock_for_each_connected_socket(virtio_vsock_reset_sock);
	397	break;
	398	}
	399	}
	400
	401	static void virtio_transport_event_work(struct work_struct *work)
	402	{
	403	struct virtio_vsock *vsock =
	404	container_of(work, struct virtio_vsock, event_work);
	405	struct virtqueue *vq;
	406
	407	vq = vsock->vqs[VSOCK_VQ_EVENT];
	408
	409	mutex_lock(&vsock->event_lock);
	410
	411	do {
	412	struct virtio_vsock_event *event;
	413	unsigned int len;
	414
	415	virtqueue_disable_cb(vq);
	416	while ((event = virtqueue_get_buf(vq, &len)) != NULL) {
	417	if (len == sizeof(*event))
	418	virtio_vsock_event_handle(vsock, event);
	419
	420	virtio_vsock_event_fill_one(vsock, event);
	421	}
	422	} while (!virtqueue_enable_cb(vq));
	423
	424	virtqueue_kick(vsock->vqs[VSOCK_VQ_EVENT]);
	425
	426	mutex_unlock(&vsock->event_lock);
	427	}
	428
	429	static void virtio_vsock_event_done(struct virtqueue *vq)
	430	{
	431	struct virtio_vsock *vsock = vq->vdev->priv;
	432
	433	if (!vsock)
	434	return;
	435	queue_work(virtio_vsock_workqueue, &vsock->event_work);
	436	}
	437
	438	static void virtio_vsock_tx_done(struct virtqueue *vq)
	439	{
	440	struct virtio_vsock *vsock = vq->vdev->priv;
	441
	442	if (!vsock)
	443	return;
	444	queue_work(virtio_vsock_workqueue, &vsock->tx_work);
	445	}
	446
447	static void virtio_vsock_rx_done(struct virtqueue *vq)
448	{
449	struct virtio_vsock *vsock = vq->vdev->priv;
450
451	if (!vsock)
452	return;
453	queue_work(virtio_vsock_workqueue, &vsock->rx_work);
454	}
455
456	static struct virtio_transport virtio_transport = {
457	.transport = {
458	.get_local_cid = virtio_transport_get_local_cid,
459
460	.init = virtio_transport_do_socket_init,
461	.destruct = virtio_transport_destruct,
462	.release = virtio_transport_release,
463	.connect = virtio_transport_connect,
464	.shutdown = virtio_transport_shutdown,
465
466	.dgram_bind = virtio_transport_dgram_bind,
467	.dgram_dequeue = virtio_transport_dgram_dequeue,
468	.dgram_enqueue = virtio_transport_dgram_enqueue,
469	.dgram_allow = virtio_transport_dgram_allow,
470
471	.stream_dequeue = virtio_transport_stream_dequeue,
472	.stream_enqueue = virtio_transport_stream_enqueue,
473	.stream_has_data = virtio_transport_stream_has_data,
474	.stream_has_space = virtio_transport_stream_has_space,
475	.stream_rcvhiwat = virtio_transport_stream_rcvhiwat,
476	.stream_is_active = virtio_transport_stream_is_active,
477	.stream_allow = virtio_transport_stream_allow,
478
479	.notify_poll_in = virtio_transport_notify_poll_in,
480	.notify_poll_out = virtio_transport_notify_poll_out,
481	.notify_recv_init = virtio_transport_notify_recv_init,
482	.notify_recv_pre_block = virtio_transport_notify_recv_pre_block,
483	.notify_recv_pre_dequeue = virtio_transport_notify_recv_pre_dequeue,
484	.notify_recv_post_dequeue = virtio_transport_notify_recv_post_dequeue,
485	.notify_send_init = virtio_transport_notify_send_init,
486	.notify_send_pre_block = virtio_transport_notify_send_pre_block,
487	.notify_send_pre_enqueue = virtio_transport_notify_send_pre_enqueue,
488	.notify_send_post_enqueue = virtio_transport_notify_send_post_enqueue,
489
490	.set_buffer_size = virtio_transport_set_buffer_size,
491	.set_min_buffer_size = virtio_transport_set_min_buffer_size,
492	.set_max_buffer_size = virtio_transport_set_max_buffer_size,
493	.get_buffer_size = virtio_transport_get_buffer_size,
494	.get_min_buffer_size = virtio_transport_get_min_buffer_size,
495	.get_max_buffer_size = virtio_transport_get_max_buffer_size,
496	},
497
498	.send_pkt = virtio_transport_send_pkt,
499	};
500
501	static int virtio_vsock_probe(struct virtio_device *vdev)
502	{
503	vq_callback_t *callbacks[] = {
504	virtio_vsock_rx_done,
505	virtio_vsock_tx_done,
506	virtio_vsock_event_done,
507	};
508	static const char * const names[] = {
509	"rx",
510	"tx",
511	"event",
512	};
513	struct virtio_vsock *vsock = NULL;
514	int ret;
515
516	ret = mutex_lock_interruptible(&the_virtio_vsock_mutex);
517	if (ret)
518	return ret;
519
520	/* Only one virtio-vsock device per guest is supported */
521	if (the_virtio_vsock) {
522	ret = -EBUSY;
523	goto out;
524	}
525
526	vsock = kzalloc(sizeof(*vsock), GFP_KERNEL);
527	if (!vsock) {
528	ret = -ENOMEM;
529	goto out;
530	}
531
532	vsock->vdev = vdev;
533
534	ret = vsock->vdev->config->find_vqs(vsock->vdev, VSOCK_VQ_MAX,
fb5e31d9 CH	535	vsock->vqs, callbacks, names,
fb5e31d9 CH	536	NULL);
0ea9e1d3 AH	537	if (ret < 0)
	538	goto out;
	539
	540	virtio_vsock_update_guest_cid(vsock);
	541
	542	ret = vsock_core_init(&virtio_transport.transport);
	543	if (ret < 0)
	544	goto out_vqs;
	545
	546	vsock->rx_buf_nr = 0;
	547	vsock->rx_buf_max_nr = 0;
	548	atomic_set(&vsock->queued_replies, 0);
	549
	550	vdev->priv = vsock;
	551	the_virtio_vsock = vsock;
	552	mutex_init(&vsock->tx_lock);
	553	mutex_init(&vsock->rx_lock);
	554	mutex_init(&vsock->event_lock);
	555	spin_lock_init(&vsock->send_pkt_list_lock);
	556	INIT_LIST_HEAD(&vsock->send_pkt_list);
b9116823 SH	557	spin_lock_init(&vsock->loopback_list_lock);
b9116823 SH	558	INIT_LIST_HEAD(&vsock->loopback_list);
0ea9e1d3 AH	559	INIT_WORK(&vsock->rx_work, virtio_transport_rx_work);
	560	INIT_WORK(&vsock->tx_work, virtio_transport_tx_work);
	561	INIT_WORK(&vsock->event_work, virtio_transport_event_work);
	562	INIT_WORK(&vsock->send_pkt_work, virtio_transport_send_pkt_work);
b9116823	563	INIT_WORK(&vsock->loopback_work, virtio_transport_loopback_work);
0ea9e1d3 AH	564
	565	mutex_lock(&vsock->rx_lock);
	566	virtio_vsock_rx_fill(vsock);
	567	mutex_unlock(&vsock->rx_lock);
	568
	569	mutex_lock(&vsock->event_lock);
	570	virtio_vsock_event_fill(vsock);
	571	mutex_unlock(&vsock->event_lock);
	572
	573	mutex_unlock(&the_virtio_vsock_mutex);
	574	return 0;
	575
	576	out_vqs:
	577	vsock->vdev->config->del_vqs(vsock->vdev);
	578	out:
	579	kfree(vsock);
	580	mutex_unlock(&the_virtio_vsock_mutex);
	581	return ret;
	582	}
	583
	584	static void virtio_vsock_remove(struct virtio_device *vdev)
	585	{
	586	struct virtio_vsock *vsock = vdev->priv;
	587	struct virtio_vsock_pkt *pkt;
	588
b9116823	589	flush_work(&vsock->loopback_work);
0ea9e1d3 AH	590	flush_work(&vsock->rx_work);
	591	flush_work(&vsock->tx_work);
	592	flush_work(&vsock->event_work);
	593	flush_work(&vsock->send_pkt_work);
	594
	595	vdev->config->reset(vdev);
	596
	597	mutex_lock(&vsock->rx_lock);
	598	while ((pkt = virtqueue_detach_unused_buf(vsock->vqs[VSOCK_VQ_RX])))
	599	virtio_transport_free_pkt(pkt);
	600	mutex_unlock(&vsock->rx_lock);
	601
	602	mutex_lock(&vsock->tx_lock);
	603	while ((pkt = virtqueue_detach_unused_buf(vsock->vqs[VSOCK_VQ_TX])))
	604	virtio_transport_free_pkt(pkt);
	605	mutex_unlock(&vsock->tx_lock);
	606
	607	spin_lock_bh(&vsock->send_pkt_list_lock);
	608	while (!list_empty(&vsock->send_pkt_list)) {
	609	pkt = list_first_entry(&vsock->send_pkt_list,
	610	struct virtio_vsock_pkt, list);
	611	list_del(&pkt->list);
	612	virtio_transport_free_pkt(pkt);
	613	}
	614	spin_unlock_bh(&vsock->send_pkt_list_lock);
	615
b9116823 SH	616	spin_lock_bh(&vsock->loopback_list_lock);
	617	while (!list_empty(&vsock->loopback_list)) {
	618	pkt = list_first_entry(&vsock->loopback_list,
	619	struct virtio_vsock_pkt, list);
	620	list_del(&pkt->list);
	621	virtio_transport_free_pkt(pkt);
	622	}
	623	spin_unlock_bh(&vsock->loopback_list_lock);
	624
0ea9e1d3 AH	625	mutex_lock(&the_virtio_vsock_mutex);
	626	the_virtio_vsock = NULL;
	627	vsock_core_exit();
	628	mutex_unlock(&the_virtio_vsock_mutex);
	629
	630	vdev->config->del_vqs(vdev);
	631
	632	kfree(vsock);
	633	}
	634
	635	static struct virtio_device_id id_table[] = {
	636	{ VIRTIO_ID_VSOCK, VIRTIO_DEV_ANY_ID },
	637	{ 0 },
	638	};
	639
	640	static unsigned int features[] = {
	641	};
	642
	643	static struct virtio_driver virtio_vsock_driver = {
	644	.feature_table = features,
	645	.feature_table_size = ARRAY_SIZE(features),
	646	.driver.name = KBUILD_MODNAME,
	647	.driver.owner = THIS_MODULE,
	648	.id_table = id_table,
	649	.probe = virtio_vsock_probe,
	650	.remove = virtio_vsock_remove,
	651	};
	652
	653	static int __init virtio_vsock_init(void)
	654	{
	655	int ret;
	656
	657	virtio_vsock_workqueue = alloc_workqueue("virtio_vsock", 0, 0);
	658	if (!virtio_vsock_workqueue)
	659	return -ENOMEM;
	660	ret = register_virtio_driver(&virtio_vsock_driver);
	661	if (ret)
	662	destroy_workqueue(virtio_vsock_workqueue);
	663	return ret;
	664	}
	665
	666	static void __exit virtio_vsock_exit(void)
	667	{
	668	unregister_virtio_driver(&virtio_vsock_driver);
	669	destroy_workqueue(virtio_vsock_workqueue);
	670	}
	671
	672	module_init(virtio_vsock_init);
	673	module_exit(virtio_vsock_exit);
	674	MODULE_LICENSE("GPL v2");
	675	MODULE_AUTHOR("Asias He");
	676	MODULE_DESCRIPTION("virtio transport for vsock");
	677	MODULE_DEVICE_TABLE(virtio, id_table);