[deliverable/linux.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;

	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_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 (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;
	u64 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);
	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);
	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);

	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->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);

	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
	47	atomic_t queued_replies;
	48
	49	/* The following fields are protected by rx_lock. vqs[VSOCK_VQ_RX]
	50	* must be accessed with rx_lock held.
	51	*/
	52	struct mutex rx_lock;
	53	int rx_buf_nr;
	54	int rx_buf_max_nr;
	55
	56	/* The following fields are protected by event_lock.
	57	* vqs[VSOCK_VQ_EVENT] must be accessed with event_lock held.
	58	*/
	59	struct mutex event_lock;
	60	struct virtio_vsock_event event_list[8];
	61
	62	u32 guest_cid;
	63	};
	64
65	static struct virtio_vsock *virtio_vsock_get(void)
66	{
67	return the_virtio_vsock;
68	}
69
70	static u32 virtio_transport_get_local_cid(void)
71	{
72	struct virtio_vsock *vsock = virtio_vsock_get();
73
74	return vsock->guest_cid;
75	}
76
77	static void
78	virtio_transport_send_pkt_work(struct work_struct *work)
79	{
80	struct virtio_vsock *vsock =
81	container_of(work, struct virtio_vsock, send_pkt_work);
82	struct virtqueue *vq;
83	bool added = false;
84	bool restart_rx = false;
85
86	mutex_lock(&vsock->tx_lock);
87
88	vq = vsock->vqs[VSOCK_VQ_TX];
89
0ea9e1d3 AH	90	for (;;) {
	91	struct virtio_vsock_pkt *pkt;
	92	struct scatterlist hdr, buf, *sgs[2];
	93	int ret, in_sg = 0, out_sg = 0;
	94	bool reply;
	95
	96	spin_lock_bh(&vsock->send_pkt_list_lock);
	97	if (list_empty(&vsock->send_pkt_list)) {
	98	spin_unlock_bh(&vsock->send_pkt_list_lock);
0ea9e1d3 AH	99	break;
	100	}
	101
	102	pkt = list_first_entry(&vsock->send_pkt_list,
	103	struct virtio_vsock_pkt, list);
	104	list_del_init(&pkt->list);
	105	spin_unlock_bh(&vsock->send_pkt_list_lock);
	106
	107	reply = pkt->reply;
	108
	109	sg_init_one(&hdr, &pkt->hdr, sizeof(pkt->hdr));
	110	sgs[out_sg++] = &hdr;
	111	if (pkt->buf) {
	112	sg_init_one(&buf, pkt->buf, pkt->len);
	113	sgs[out_sg++] = &buf;
	114	}
	115
	116	ret = virtqueue_add_sgs(vq, sgs, out_sg, in_sg, pkt, GFP_KERNEL);
21bc54fc GG	117	/* Usually this means that there is no more space available in
	118	* the vq
	119	*/
0ea9e1d3 AH	120	if (ret < 0) {
	121	spin_lock_bh(&vsock->send_pkt_list_lock);
	122	list_add(&pkt->list, &vsock->send_pkt_list);
	123	spin_unlock_bh(&vsock->send_pkt_list_lock);
0ea9e1d3 AH	124	break;
	125	}
	126
	127	if (reply) {
	128	struct virtqueue *rx_vq = vsock->vqs[VSOCK_VQ_RX];
	129	int val;
	130
	131	val = atomic_dec_return(&vsock->queued_replies);
	132
	133	/* Do we now have resources to resume rx processing? */
	134	if (val + 1 == virtqueue_get_vring_size(rx_vq))
	135	restart_rx = true;
	136	}
	137
	138	added = true;
	139	}
	140
	141	if (added)
	142	virtqueue_kick(vq);
	143
	144	mutex_unlock(&vsock->tx_lock);
	145
	146	if (restart_rx)
	147	queue_work(virtio_vsock_workqueue, &vsock->rx_work);
	148	}
	149
	150	static int
	151	virtio_transport_send_pkt(struct virtio_vsock_pkt *pkt)
	152	{
	153	struct virtio_vsock *vsock;
	154	int len = pkt->len;
	155
	156	vsock = virtio_vsock_get();
	157	if (!vsock) {
	158	virtio_transport_free_pkt(pkt);
	159	return -ENODEV;
	160	}
	161
	162	if (pkt->reply)
	163	atomic_inc(&vsock->queued_replies);
	164
	165	spin_lock_bh(&vsock->send_pkt_list_lock);
	166	list_add_tail(&pkt->list, &vsock->send_pkt_list);
	167	spin_unlock_bh(&vsock->send_pkt_list_lock);
	168
	169	queue_work(virtio_vsock_workqueue, &vsock->send_pkt_work);
	170	return len;
	171	}
	172
	173	static void virtio_vsock_rx_fill(struct virtio_vsock *vsock)
	174	{
	175	int buf_len = VIRTIO_VSOCK_DEFAULT_RX_BUF_SIZE;
	176	struct virtio_vsock_pkt *pkt;
	177	struct scatterlist hdr, buf, *sgs[2];
	178	struct virtqueue *vq;
	179	int ret;
	180
	181	vq = vsock->vqs[VSOCK_VQ_RX];
	182
	183	do {
	184	pkt = kzalloc(sizeof(*pkt), GFP_KERNEL);
	185	if (!pkt)
	186	break;
	187
188	pkt->buf = kmalloc(buf_len, GFP_KERNEL);
189	if (!pkt->buf) {
190	virtio_transport_free_pkt(pkt);
191	break;
192	}
193
194	pkt->len = buf_len;
195
196	sg_init_one(&hdr, &pkt->hdr, sizeof(pkt->hdr));
197	sgs[0] = &hdr;
198
199	sg_init_one(&buf, pkt->buf, buf_len);
200	sgs[1] = &buf;
201	ret = virtqueue_add_sgs(vq, sgs, 0, 2, pkt, GFP_KERNEL);
202	if (ret) {
203	virtio_transport_free_pkt(pkt);
204	break;
205	}
206	vsock->rx_buf_nr++;
207	} while (vq->num_free);
208	if (vsock->rx_buf_nr > vsock->rx_buf_max_nr)
209	vsock->rx_buf_max_nr = vsock->rx_buf_nr;
210	virtqueue_kick(vq);
211	}
212
213	static void virtio_transport_tx_work(struct work_struct *work)
214	{
215	struct virtio_vsock *vsock =
216	container_of(work, struct virtio_vsock, tx_work);
217	struct virtqueue *vq;
218	bool added = false;
219
220	vq = vsock->vqs[VSOCK_VQ_TX];
221	mutex_lock(&vsock->tx_lock);
222	do {
223	struct virtio_vsock_pkt *pkt;
224	unsigned int len;
225
226	virtqueue_disable_cb(vq);
227	while ((pkt = virtqueue_get_buf(vq, &len)) != NULL) {
228	virtio_transport_free_pkt(pkt);
229	added = true;
230	}
231	} while (!virtqueue_enable_cb(vq));
232	mutex_unlock(&vsock->tx_lock);
233
234	if (added)
235	queue_work(virtio_vsock_workqueue, &vsock->send_pkt_work);
236	}
237
238	/* Is there space left for replies to rx packets? */
239	static bool virtio_transport_more_replies(struct virtio_vsock *vsock)
240	{
241	struct virtqueue *vq = vsock->vqs[VSOCK_VQ_RX];
242	int val;
243
244	smp_rmb(); /* paired with atomic_inc() and atomic_dec_return() */
245	val = atomic_read(&vsock->queued_replies);
246
247	return val < virtqueue_get_vring_size(vq);
248	}
249
250	static void virtio_transport_rx_work(struct work_struct *work)
251	{
252	struct virtio_vsock *vsock =
253	container_of(work, struct virtio_vsock, rx_work);
254	struct virtqueue *vq;
255
256	vq = vsock->vqs[VSOCK_VQ_RX];
257
258	mutex_lock(&vsock->rx_lock);
259
260	do {
261	virtqueue_disable_cb(vq);
262	for (;;) {
263	struct virtio_vsock_pkt *pkt;
264	unsigned int len;
265
266	if (!virtio_transport_more_replies(vsock)) {
267	/* Stop rx until the device processes already
268	* pending replies. Leave rx virtqueue
269	* callbacks disabled.
270	*/
271	goto out;
272	}
273
274	pkt = virtqueue_get_buf(vq, &len);
275	if (!pkt) {
276	break;
277	}
278
279	vsock->rx_buf_nr--;
280
281	/* Drop short/long packets */
282	if (unlikely(len < sizeof(pkt->hdr) \|\|
283	len > sizeof(pkt->hdr) + pkt->len)) {
284	virtio_transport_free_pkt(pkt);
285	continue;
286	}
287
288	pkt->len = len - sizeof(pkt->hdr);
289	virtio_transport_recv_pkt(pkt);
290	}
291	} while (!virtqueue_enable_cb(vq));
292
293	out:
294	if (vsock->rx_buf_nr < vsock->rx_buf_max_nr / 2)
295	virtio_vsock_rx_fill(vsock);
296	mutex_unlock(&vsock->rx_lock);
297	}
298
299	/* event_lock must be held */
300	static int virtio_vsock_event_fill_one(struct virtio_vsock *vsock,
301	struct virtio_vsock_event *event)
302	{
303	struct scatterlist sg;
304	struct virtqueue *vq;
305
306	vq = vsock->vqs[VSOCK_VQ_EVENT];
307
308	sg_init_one(&sg, event, sizeof(*event));
309
310	return virtqueue_add_inbuf(vq, &sg, 1, event, GFP_KERNEL);
311	}
312
313	/* event_lock must be held */
314	static void virtio_vsock_event_fill(struct virtio_vsock *vsock)
315	{
316	size_t i;
317
318	for (i = 0; i < ARRAY_SIZE(vsock->event_list); i++) {
319	struct virtio_vsock_event *event = &vsock->event_list[i];
320
321	virtio_vsock_event_fill_one(vsock, event);
322	}
323
324	virtqueue_kick(vsock->vqs[VSOCK_VQ_EVENT]);
325	}
326
327	static void virtio_vsock_reset_sock(struct sock *sk)
328	{
329	lock_sock(sk);
330	sk->sk_state = SS_UNCONNECTED;
331	sk->sk_err = ECONNRESET;
332	sk->sk_error_report(sk);
333	release_sock(sk);
334	}
335
336	static void virtio_vsock_update_guest_cid(struct virtio_vsock *vsock)
337	{
338	struct virtio_device *vdev = vsock->vdev;
339	u64 guest_cid;
340
341	vdev->config->get(vdev, offsetof(struct virtio_vsock_config, guest_cid),
342	&guest_cid, sizeof(guest_cid));
343	vsock->guest_cid = le64_to_cpu(guest_cid);
344	}
345
346	/* event_lock must be held */
347	static void virtio_vsock_event_handle(struct virtio_vsock *vsock,
348	struct virtio_vsock_event *event)
349	{
350	switch (le32_to_cpu(event->id)) {
351	case VIRTIO_VSOCK_EVENT_TRANSPORT_RESET:
352	virtio_vsock_update_guest_cid(vsock);
353	vsock_for_each_connected_socket(virtio_vsock_reset_sock);
354	break;
355	}
356	}
357
358	static void virtio_transport_event_work(struct work_struct *work)
359	{
360	struct virtio_vsock *vsock =
361	container_of(work, struct virtio_vsock, event_work);
362	struct virtqueue *vq;
363
364	vq = vsock->vqs[VSOCK_VQ_EVENT];
365
366	mutex_lock(&vsock->event_lock);
367
368	do {
369	struct virtio_vsock_event *event;
370	unsigned int len;
371
372	virtqueue_disable_cb(vq);
373	while ((event = virtqueue_get_buf(vq, &len)) != NULL) {
374	if (len == sizeof(*event))
375	virtio_vsock_event_handle(vsock, event);
376
377	virtio_vsock_event_fill_one(vsock, event);
378	}
379	} while (!virtqueue_enable_cb(vq));
380
381	virtqueue_kick(vsock->vqs[VSOCK_VQ_EVENT]);
382
383	mutex_unlock(&vsock->event_lock);
384	}
385
386	static void virtio_vsock_event_done(struct virtqueue *vq)
387	{
388	struct virtio_vsock *vsock = vq->vdev->priv;
389
390	if (!vsock)
391	return;
392	queue_work(virtio_vsock_workqueue, &vsock->event_work);
393	}
394
395	static void virtio_vsock_tx_done(struct virtqueue *vq)
396	{
397	struct virtio_vsock *vsock = vq->vdev->priv;
398
399	if (!vsock)
400	return;
401	queue_work(virtio_vsock_workqueue, &vsock->tx_work);
402	}
403
404	static void virtio_vsock_rx_done(struct virtqueue *vq)
405	{
406	struct virtio_vsock *vsock = vq->vdev->priv;
407
408	if (!vsock)
409	return;
410	queue_work(virtio_vsock_workqueue, &vsock->rx_work);
411	}
412
413	static struct virtio_transport virtio_transport = {
414	.transport = {
415	.get_local_cid = virtio_transport_get_local_cid,
416
417	.init = virtio_transport_do_socket_init,
418	.destruct = virtio_transport_destruct,
419	.release = virtio_transport_release,
420	.connect = virtio_transport_connect,
421	.shutdown = virtio_transport_shutdown,
422
423	.dgram_bind = virtio_transport_dgram_bind,
424	.dgram_dequeue = virtio_transport_dgram_dequeue,
425	.dgram_enqueue = virtio_transport_dgram_enqueue,
426	.dgram_allow = virtio_transport_dgram_allow,
427
428	.stream_dequeue = virtio_transport_stream_dequeue,
429	.stream_enqueue = virtio_transport_stream_enqueue,
430	.stream_has_data = virtio_transport_stream_has_data,
431	.stream_has_space = virtio_transport_stream_has_space,
432	.stream_rcvhiwat = virtio_transport_stream_rcvhiwat,
433	.stream_is_active = virtio_transport_stream_is_active,
434	.stream_allow = virtio_transport_stream_allow,
435
436	.notify_poll_in = virtio_transport_notify_poll_in,
437	.notify_poll_out = virtio_transport_notify_poll_out,
438	.notify_recv_init = virtio_transport_notify_recv_init,
439	.notify_recv_pre_block = virtio_transport_notify_recv_pre_block,
440	.notify_recv_pre_dequeue = virtio_transport_notify_recv_pre_dequeue,
441	.notify_recv_post_dequeue = virtio_transport_notify_recv_post_dequeue,
442	.notify_send_init = virtio_transport_notify_send_init,
443	.notify_send_pre_block = virtio_transport_notify_send_pre_block,
444	.notify_send_pre_enqueue = virtio_transport_notify_send_pre_enqueue,
445	.notify_send_post_enqueue = virtio_transport_notify_send_post_enqueue,
446
447	.set_buffer_size = virtio_transport_set_buffer_size,
448	.set_min_buffer_size = virtio_transport_set_min_buffer_size,
449	.set_max_buffer_size = virtio_transport_set_max_buffer_size,
450	.get_buffer_size = virtio_transport_get_buffer_size,
451	.get_min_buffer_size = virtio_transport_get_min_buffer_size,
452	.get_max_buffer_size = virtio_transport_get_max_buffer_size,
453	},
454
455	.send_pkt = virtio_transport_send_pkt,
456	};
457
458	static int virtio_vsock_probe(struct virtio_device *vdev)
459	{
460	vq_callback_t *callbacks[] = {
461	virtio_vsock_rx_done,
462	virtio_vsock_tx_done,
463	virtio_vsock_event_done,
464	};
465	static const char * const names[] = {
466	"rx",
467	"tx",
468	"event",
469	};
470	struct virtio_vsock *vsock = NULL;
471	int ret;
472
473	ret = mutex_lock_interruptible(&the_virtio_vsock_mutex);
474	if (ret)
475	return ret;
476
477	/* Only one virtio-vsock device per guest is supported */
478	if (the_virtio_vsock) {
479	ret = -EBUSY;
480	goto out;
481	}
482
483	vsock = kzalloc(sizeof(*vsock), GFP_KERNEL);
484	if (!vsock) {
485	ret = -ENOMEM;
486	goto out;
487	}
488
489	vsock->vdev = vdev;
490
491	ret = vsock->vdev->config->find_vqs(vsock->vdev, VSOCK_VQ_MAX,
492	vsock->vqs, callbacks, names);
493	if (ret < 0)
494	goto out;
495
496	virtio_vsock_update_guest_cid(vsock);
497
498	ret = vsock_core_init(&virtio_transport.transport);
499	if (ret < 0)
500	goto out_vqs;
501
502	vsock->rx_buf_nr = 0;
503	vsock->rx_buf_max_nr = 0;
504	atomic_set(&vsock->queued_replies, 0);
505
506	vdev->priv = vsock;
507	the_virtio_vsock = vsock;
508	mutex_init(&vsock->tx_lock);
509	mutex_init(&vsock->rx_lock);
510	mutex_init(&vsock->event_lock);
511	spin_lock_init(&vsock->send_pkt_list_lock);
512	INIT_LIST_HEAD(&vsock->send_pkt_list);
513	INIT_WORK(&vsock->rx_work, virtio_transport_rx_work);
514	INIT_WORK(&vsock->tx_work, virtio_transport_tx_work);
515	INIT_WORK(&vsock->event_work, virtio_transport_event_work);
516	INIT_WORK(&vsock->send_pkt_work, virtio_transport_send_pkt_work);
517
518	mutex_lock(&vsock->rx_lock);
519	virtio_vsock_rx_fill(vsock);
520	mutex_unlock(&vsock->rx_lock);
521
522	mutex_lock(&vsock->event_lock);
523	virtio_vsock_event_fill(vsock);
524	mutex_unlock(&vsock->event_lock);
525
526	mutex_unlock(&the_virtio_vsock_mutex);
527	return 0;
528
529	out_vqs:
530	vsock->vdev->config->del_vqs(vsock->vdev);
531	out:
532	kfree(vsock);
533	mutex_unlock(&the_virtio_vsock_mutex);
534	return ret;
535	}
536
537	static void virtio_vsock_remove(struct virtio_device *vdev)
538	{
539	struct virtio_vsock *vsock = vdev->priv;
540	struct virtio_vsock_pkt *pkt;
541
542	flush_work(&vsock->rx_work);
543	flush_work(&vsock->tx_work);
544	flush_work(&vsock->event_work);
545	flush_work(&vsock->send_pkt_work);
546
547	vdev->config->reset(vdev);
548
549	mutex_lock(&vsock->rx_lock);
550	while ((pkt = virtqueue_detach_unused_buf(vsock->vqs[VSOCK_VQ_RX])))
551	virtio_transport_free_pkt(pkt);
552	mutex_unlock(&vsock->rx_lock);
553
554	mutex_lock(&vsock->tx_lock);
555	while ((pkt = virtqueue_detach_unused_buf(vsock->vqs[VSOCK_VQ_TX])))
556	virtio_transport_free_pkt(pkt);
557	mutex_unlock(&vsock->tx_lock);
558
559	spin_lock_bh(&vsock->send_pkt_list_lock);
560	while (!list_empty(&vsock->send_pkt_list)) {
561	pkt = list_first_entry(&vsock->send_pkt_list,
562	struct virtio_vsock_pkt, list);
563	list_del(&pkt->list);
564	virtio_transport_free_pkt(pkt);
565	}
566	spin_unlock_bh(&vsock->send_pkt_list_lock);
567
568	mutex_lock(&the_virtio_vsock_mutex);
569	the_virtio_vsock = NULL;
570	vsock_core_exit();
571	mutex_unlock(&the_virtio_vsock_mutex);
572
573	vdev->config->del_vqs(vdev);
574
575	kfree(vsock);
576	}
577
578	static struct virtio_device_id id_table[] = {
579	{ VIRTIO_ID_VSOCK, VIRTIO_DEV_ANY_ID },
580	{ 0 },
581	};
582
583	static unsigned int features[] = {
584	};
585
586	static struct virtio_driver virtio_vsock_driver = {
587	.feature_table = features,
588	.feature_table_size = ARRAY_SIZE(features),
589	.driver.name = KBUILD_MODNAME,
590	.driver.owner = THIS_MODULE,
591	.id_table = id_table,
592	.probe = virtio_vsock_probe,
593	.remove = virtio_vsock_remove,
594	};
595
596	static int __init virtio_vsock_init(void)
597	{
598	int ret;
599
600	virtio_vsock_workqueue = alloc_workqueue("virtio_vsock", 0, 0);
601	if (!virtio_vsock_workqueue)
602	return -ENOMEM;
603	ret = register_virtio_driver(&virtio_vsock_driver);
604	if (ret)
605	destroy_workqueue(virtio_vsock_workqueue);
606	return ret;
607	}
608
609	static void __exit virtio_vsock_exit(void)
610	{
611	unregister_virtio_driver(&virtio_vsock_driver);
612	destroy_workqueue(virtio_vsock_workqueue);
613	}
614
615	module_init(virtio_vsock_init);
616	module_exit(virtio_vsock_exit);
617	MODULE_LICENSE("GPL v2");
618	MODULE_AUTHOR("Asias He");
619	MODULE_DESCRIPTION("virtio transport for vsock");
620	MODULE_DEVICE_TABLE(virtio, id_table);