[GitHub/mt8127/android_kernel_alcatel_ttab.git] / drivers / net / virtio_net.c

/* A simple network driver using virtio.
 *
 * Copyright 2007 Rusty Russell <rusty@rustcorp.com.au> IBM Corporation
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */
//#define DEBUG
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/module.h>
#include <linux/virtio.h>
#include <linux/virtio_net.h>
#include <linux/scatterlist.h>

static int napi_weight = 128;
module_param(napi_weight, int, 0444);

static int csum = 1, gso = 1;
module_param(csum, bool, 0444);
module_param(gso, bool, 0444);

/* FIXME: MTU in config. */
#define MAX_PACKET_LEN (ETH_HLEN+ETH_DATA_LEN)

struct virtnet_info
{
	struct virtio_device *vdev;
	struct virtqueue *rvq, *svq;
	struct net_device *dev;
	struct napi_struct napi;

	/* The skb we couldn't send because buffers were full. */
	struct sk_buff *last_xmit_skb;

	/* Number of input buffers, and max we've ever had. */
	unsigned int num, max;

	/* For cleaning up after transmission. */
	struct tasklet_struct tasklet;

	/* Receive & send queues. */
	struct sk_buff_head recv;
	struct sk_buff_head send;
};

static inline struct virtio_net_hdr *skb_vnet_hdr(struct sk_buff *skb)
{
	return (struct virtio_net_hdr *)skb->cb;
}

static inline void vnet_hdr_to_sg(struct scatterlist *sg, struct sk_buff *skb)
{
	sg_init_one(sg, skb_vnet_hdr(skb), sizeof(struct virtio_net_hdr));
}

static void skb_xmit_done(struct virtqueue *svq)
{
	struct virtnet_info *vi = svq->vdev->priv;

	/* Suppress further interrupts. */
	svq->vq_ops->disable_cb(svq);

	/* We were waiting for more output buffers. */
	netif_wake_queue(vi->dev);

	/* Make sure we re-xmit last_xmit_skb: if there are no more packets
	 * queued, start_xmit won't be called. */
	tasklet_schedule(&vi->tasklet);
}

static void receive_skb(struct net_device *dev, struct sk_buff *skb,
			unsigned len)
{
	struct virtio_net_hdr *hdr = skb_vnet_hdr(skb);

	if (unlikely(len < sizeof(struct virtio_net_hdr) + ETH_HLEN)) {
		pr_debug("%s: short packet %i\n", dev->name, len);
		dev->stats.rx_length_errors++;
		goto drop;
	}
	len -= sizeof(struct virtio_net_hdr);
	BUG_ON(len > MAX_PACKET_LEN);

	skb_trim(skb, len);

	dev->stats.rx_bytes += skb->len;
	dev->stats.rx_packets++;

	if (hdr->flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) {
		pr_debug("Needs csum!\n");
		if (!skb_partial_csum_set(skb,hdr->csum_start,hdr->csum_offset))
			goto frame_err;
	}

	skb->protocol = eth_type_trans(skb, dev);
	pr_debug("Receiving skb proto 0x%04x len %i type %i\n",
		 ntohs(skb->protocol), skb->len, skb->pkt_type);

	if (hdr->gso_type != VIRTIO_NET_HDR_GSO_NONE) {
		pr_debug("GSO!\n");
		switch (hdr->gso_type & ~VIRTIO_NET_HDR_GSO_ECN) {
		case VIRTIO_NET_HDR_GSO_TCPV4:
			skb_shinfo(skb)->gso_type = SKB_GSO_TCPV4;
			break;
		case VIRTIO_NET_HDR_GSO_UDP:
			skb_shinfo(skb)->gso_type = SKB_GSO_UDP;
			break;
		case VIRTIO_NET_HDR_GSO_TCPV6:
			skb_shinfo(skb)->gso_type = SKB_GSO_TCPV6;
			break;
		default:
			if (net_ratelimit())
				printk(KERN_WARNING "%s: bad gso type %u.\n",
				       dev->name, hdr->gso_type);
			goto frame_err;
		}

		if (hdr->gso_type & VIRTIO_NET_HDR_GSO_ECN)
			skb_shinfo(skb)->gso_type |= SKB_GSO_TCP_ECN;

		skb_shinfo(skb)->gso_size = hdr->gso_size;
		if (skb_shinfo(skb)->gso_size == 0) {
			if (net_ratelimit())
				printk(KERN_WARNING "%s: zero gso size.\n",
				       dev->name);
			goto frame_err;
		}

		/* Header must be checked, and gso_segs computed. */
		skb_shinfo(skb)->gso_type |= SKB_GSO_DODGY;
		skb_shinfo(skb)->gso_segs = 0;
	}

	netif_receive_skb(skb);
	return;

frame_err:
	dev->stats.rx_frame_errors++;
drop:
	dev_kfree_skb(skb);
}

static void try_fill_recv(struct virtnet_info *vi)
{
	struct sk_buff *skb;
	struct scatterlist sg[2+MAX_SKB_FRAGS];
	int num, err;

	sg_init_table(sg, 2+MAX_SKB_FRAGS);
	for (;;) {
		skb = netdev_alloc_skb(vi->dev, MAX_PACKET_LEN);
		if (unlikely(!skb))
			break;

		skb_put(skb, MAX_PACKET_LEN);
		vnet_hdr_to_sg(sg, skb);
		num = skb_to_sgvec(skb, sg+1, 0, skb->len) + 1;
		skb_queue_head(&vi->recv, skb);

		err = vi->rvq->vq_ops->add_buf(vi->rvq, sg, 0, num, skb);
		if (err) {
			skb_unlink(skb, &vi->recv);
			kfree_skb(skb);
			break;
		}
		vi->num++;
	}
	if (unlikely(vi->num > vi->max))
		vi->max = vi->num;
	vi->rvq->vq_ops->kick(vi->rvq);
}

static void skb_recv_done(struct virtqueue *rvq)
{
	struct virtnet_info *vi = rvq->vdev->priv;
	/* Schedule NAPI, Suppress further interrupts if successful. */
	if (netif_rx_schedule_prep(vi->dev, &vi->napi)) {
		rvq->vq_ops->disable_cb(rvq);
		__netif_rx_schedule(vi->dev, &vi->napi);
	}
}

static int virtnet_poll(struct napi_struct *napi, int budget)
{
	struct virtnet_info *vi = container_of(napi, struct virtnet_info, napi);
	struct sk_buff *skb = NULL;
	unsigned int len, received = 0;

again:
	while (received < budget &&
	       (skb = vi->rvq->vq_ops->get_buf(vi->rvq, &len)) != NULL) {
		__skb_unlink(skb, &vi->recv);
		receive_skb(vi->dev, skb, len);
		vi->num--;
		received++;
	}

	/* FIXME: If we oom and completely run out of inbufs, we need
	 * to start a timer trying to fill more. */
	if (vi->num < vi->max / 2)
		try_fill_recv(vi);

	/* Out of packets? */
	if (received < budget) {
		netif_rx_complete(vi->dev, napi);
		if (unlikely(!vi->rvq->vq_ops->enable_cb(vi->rvq))
		    && napi_schedule_prep(napi)) {
			vi->rvq->vq_ops->disable_cb(vi->rvq);
			__netif_rx_schedule(vi->dev, napi);
			goto again;
		}
	}

	return received;
}

static void free_old_xmit_skbs(struct virtnet_info *vi)
{
	struct sk_buff *skb;
	unsigned int len;

	while ((skb = vi->svq->vq_ops->get_buf(vi->svq, &len)) != NULL) {
		pr_debug("Sent skb %p\n", skb);
		__skb_unlink(skb, &vi->send);
		vi->dev->stats.tx_bytes += skb->len;
		vi->dev->stats.tx_packets++;
		kfree_skb(skb);
	}
}

static int xmit_skb(struct virtnet_info *vi, struct sk_buff *skb)
{
	int num;
	struct scatterlist sg[2+MAX_SKB_FRAGS];
	struct virtio_net_hdr *hdr;
	const unsigned char *dest = ((struct ethhdr *)skb->data)->h_dest;

	sg_init_table(sg, 2+MAX_SKB_FRAGS);

	pr_debug("%s: xmit %p " MAC_FMT "\n", vi->dev->name, skb,
		 dest[0], dest[1], dest[2],
		 dest[3], dest[4], dest[5]);

	/* Encode metadata header at front. */
	hdr = skb_vnet_hdr(skb);
	if (skb->ip_summed == CHECKSUM_PARTIAL) {
		hdr->flags = VIRTIO_NET_HDR_F_NEEDS_CSUM;
		hdr->csum_start = skb->csum_start - skb_headroom(skb);
		hdr->csum_offset = skb->csum_offset;
	} else {
		hdr->flags = 0;
		hdr->csum_offset = hdr->csum_start = 0;
	}

	if (skb_is_gso(skb)) {
		hdr->hdr_len = skb_transport_header(skb) - skb->data;
		hdr->gso_size = skb_shinfo(skb)->gso_size;
		if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV4)
			hdr->gso_type = VIRTIO_NET_HDR_GSO_TCPV4;
		else if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6)
			hdr->gso_type = VIRTIO_NET_HDR_GSO_TCPV6;
		else if (skb_shinfo(skb)->gso_type & SKB_GSO_UDP)
			hdr->gso_type = VIRTIO_NET_HDR_GSO_UDP;
		else
			BUG();
		if (skb_shinfo(skb)->gso_type & SKB_GSO_TCP_ECN)
			hdr->gso_type |= VIRTIO_NET_HDR_GSO_ECN;
	} else {
		hdr->gso_type = VIRTIO_NET_HDR_GSO_NONE;
		hdr->gso_size = hdr->hdr_len = 0;
	}

	vnet_hdr_to_sg(sg, skb);
	num = skb_to_sgvec(skb, sg+1, 0, skb->len) + 1;

	return vi->svq->vq_ops->add_buf(vi->svq, sg, num, 0, skb);
}

static void xmit_tasklet(unsigned long data)
{
	struct virtnet_info *vi = (void *)data;

	netif_tx_lock_bh(vi->dev);
	if (vi->last_xmit_skb && xmit_skb(vi, vi->last_xmit_skb) == 0) {
		vi->svq->vq_ops->kick(vi->svq);
		vi->last_xmit_skb = NULL;
	}
	netif_tx_unlock_bh(vi->dev);
}

static int start_xmit(struct sk_buff *skb, struct net_device *dev)
{
	struct virtnet_info *vi = netdev_priv(dev);

again:
	/* Free up any pending old buffers before queueing new ones. */
	free_old_xmit_skbs(vi);

	/* If we has a buffer left over from last time, send it now. */
	if (unlikely(vi->last_xmit_skb)) {
		if (xmit_skb(vi, vi->last_xmit_skb) != 0) {
			/* Drop this skb: we only queue one. */
			vi->dev->stats.tx_dropped++;
			kfree_skb(skb);
			skb = NULL;
			goto stop_queue;
		}
		vi->last_xmit_skb = NULL;
	}

	/* Put new one in send queue and do transmit */
	if (likely(skb)) {
		__skb_queue_head(&vi->send, skb);
		if (xmit_skb(vi, skb) != 0) {
			vi->last_xmit_skb = skb;
			skb = NULL;
			goto stop_queue;
		}
	}
done:
	vi->svq->vq_ops->kick(vi->svq);
	return NETDEV_TX_OK;

stop_queue:
	pr_debug("%s: virtio not prepared to send\n", dev->name);
	netif_stop_queue(dev);

	/* Activate callback for using skbs: if this returns false it
	 * means some were used in the meantime. */
	if (unlikely(!vi->svq->vq_ops->enable_cb(vi->svq))) {
		vi->svq->vq_ops->disable_cb(vi->svq);
		netif_start_queue(dev);
		goto again;
	}
	goto done;
}

#ifdef CONFIG_NET_POLL_CONTROLLER
static void virtnet_netpoll(struct net_device *dev)
{
	struct virtnet_info *vi = netdev_priv(dev);

	napi_schedule(&vi->napi);
}
#endif

static int virtnet_open(struct net_device *dev)
{
	struct virtnet_info *vi = netdev_priv(dev);

	napi_enable(&vi->napi);

	/* If all buffers were filled by other side before we napi_enabled, we
	 * won't get another interrupt, so process any outstanding packets
	 * now.  virtnet_poll wants re-enable the queue, so we disable here.
	 * We synchronize against interrupts via NAPI_STATE_SCHED */
	if (netif_rx_schedule_prep(dev, &vi->napi)) {
		vi->rvq->vq_ops->disable_cb(vi->rvq);
		__netif_rx_schedule(dev, &vi->napi);
	}
	return 0;
}

static int virtnet_close(struct net_device *dev)
{
	struct virtnet_info *vi = netdev_priv(dev);

	napi_disable(&vi->napi);

	return 0;
}

static int virtnet_probe(struct virtio_device *vdev)
{
	int err;
	struct net_device *dev;
	struct virtnet_info *vi;

	/* Allocate ourselves a network device with room for our info */
	dev = alloc_etherdev(sizeof(struct virtnet_info));
	if (!dev)
		return -ENOMEM;

	/* Set up network device as normal. */
	dev->open = virtnet_open;
	dev->stop = virtnet_close;
	dev->hard_start_xmit = start_xmit;
	dev->features = NETIF_F_HIGHDMA;
#ifdef CONFIG_NET_POLL_CONTROLLER
	dev->poll_controller = virtnet_netpoll;
#endif
	SET_NETDEV_DEV(dev, &vdev->dev);

	/* Do we support "hardware" checksums? */
	if (csum && virtio_has_feature(vdev, VIRTIO_NET_F_CSUM)) {
		/* This opens up the world of extra features. */
		dev->features |= NETIF_F_HW_CSUM|NETIF_F_SG|NETIF_F_FRAGLIST;
		if (gso && virtio_has_feature(vdev, VIRTIO_NET_F_GSO)) {
			dev->features |= NETIF_F_TSO | NETIF_F_UFO
				| NETIF_F_TSO_ECN | NETIF_F_TSO6;
		}
		/* Individual feature bits: what can host handle? */
		if (gso && virtio_has_feature(vdev, VIRTIO_NET_F_HOST_TSO4))
			dev->features |= NETIF_F_TSO;
		if (gso && virtio_has_feature(vdev, VIRTIO_NET_F_HOST_TSO6))
			dev->features |= NETIF_F_TSO6;
		if (gso && virtio_has_feature(vdev, VIRTIO_NET_F_HOST_ECN))
			dev->features |= NETIF_F_TSO_ECN;
		if (gso && virtio_has_feature(vdev, VIRTIO_NET_F_HOST_UFO))
			dev->features |= NETIF_F_UFO;
	}

	/* Configuration may specify what MAC to use.  Otherwise random. */
	if (virtio_has_feature(vdev, VIRTIO_NET_F_MAC)) {
		vdev->config->get(vdev,
				  offsetof(struct virtio_net_config, mac),
				  dev->dev_addr, dev->addr_len);
	} else
		random_ether_addr(dev->dev_addr);

	/* Set up our device-specific information */
	vi = netdev_priv(dev);
	netif_napi_add(dev, &vi->napi, virtnet_poll, napi_weight);
	vi->dev = dev;
	vi->vdev = vdev;
	vdev->priv = vi;

	/* We expect two virtqueues, receive then send. */
	vi->rvq = vdev->config->find_vq(vdev, 0, skb_recv_done);
	if (IS_ERR(vi->rvq)) {
		err = PTR_ERR(vi->rvq);
		goto free;
	}

	vi->svq = vdev->config->find_vq(vdev, 1, skb_xmit_done);
	if (IS_ERR(vi->svq)) {
		err = PTR_ERR(vi->svq);
		goto free_recv;
	}

	/* Initialize our empty receive and send queues. */
	skb_queue_head_init(&vi->recv);
	skb_queue_head_init(&vi->send);

	tasklet_init(&vi->tasklet, xmit_tasklet, (unsigned long)vi);

	err = register_netdev(dev);
	if (err) {
		pr_debug("virtio_net: registering device failed\n");
		goto free_send;
	}

	/* Last of all, set up some receive buffers. */
	try_fill_recv(vi);

	/* If we didn't even get one input buffer, we're useless. */
	if (vi->num == 0) {
		err = -ENOMEM;
		goto unregister;
	}

	pr_debug("virtnet: registered device %s\n", dev->name);
	return 0;

unregister:
	unregister_netdev(dev);
free_send:
	vdev->config->del_vq(vi->svq);
free_recv:
	vdev->config->del_vq(vi->rvq);
free:
	free_netdev(dev);
	return err;
}

static void virtnet_remove(struct virtio_device *vdev)
{
	struct virtnet_info *vi = vdev->priv;
	struct sk_buff *skb;

	/* Stop all the virtqueues. */
	vdev->config->reset(vdev);

	/* Free our skbs in send and recv queues, if any. */
	while ((skb = __skb_dequeue(&vi->recv)) != NULL) {
		kfree_skb(skb);
		vi->num--;
	}
	__skb_queue_purge(&vi->send);

	BUG_ON(vi->num != 0);

	vdev->config->del_vq(vi->svq);
	vdev->config->del_vq(vi->rvq);
	unregister_netdev(vi->dev);
	free_netdev(vi->dev);
}

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

static unsigned int features[] = {
	VIRTIO_NET_F_CSUM, VIRTIO_NET_F_GSO, VIRTIO_NET_F_MAC,
	VIRTIO_NET_F_HOST_TSO4, VIRTIO_NET_F_HOST_UFO, VIRTIO_NET_F_HOST_TSO6,
	VIRTIO_NET_F_HOST_ECN,
};

static struct virtio_driver virtio_net = {
	.feature_table = features,
	.feature_table_size = ARRAY_SIZE(features),
	.driver.name =	KBUILD_MODNAME,
	.driver.owner =	THIS_MODULE,
	.id_table =	id_table,
	.probe =	virtnet_probe,
	.remove =	__devexit_p(virtnet_remove),
};

static int __init init(void)
{
	return register_virtio_driver(&virtio_net);
}

static void __exit fini(void)
{
	unregister_virtio_driver(&virtio_net);
}
module_init(init);
module_exit(fini);

MODULE_DEVICE_TABLE(virtio, id_table);
MODULE_DESCRIPTION("Virtio network driver");
MODULE_LICENSE("GPL");
Commit	Line	Data
296f96fc RR	1	/* A simple network driver using virtio.
	2	*
	3	* Copyright 2007 Rusty Russell <rusty@rustcorp.com.au> IBM Corporation
	4	*
	5	* This program is free software; you can redistribute it and/or modify
	6	* it under the terms of the GNU General Public License as published by
	7	* the Free Software Foundation; either version 2 of the License, or
	8	* (at your option) any later version.
	9	*
	10	* This program is distributed in the hope that it will be useful,
	11	* but WITHOUT ANY WARRANTY; without even the implied warranty of
	12	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	13	* GNU General Public License for more details.
	14	*
	15	* You should have received a copy of the GNU General Public License
	16	* along with this program; if not, write to the Free Software
	17	* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
	18	*/
	19	//#define DEBUG
	20	#include <linux/netdevice.h>
	21	#include <linux/etherdevice.h>
	22	#include <linux/module.h>
	23	#include <linux/virtio.h>
	24	#include <linux/virtio_net.h>
	25	#include <linux/scatterlist.h>
	26
6c0cd7c0 DL	27	static int napi_weight = 128;
	28	module_param(napi_weight, int, 0444);
	29
34a48579 RR	30	static int csum = 1, gso = 1;
	31	module_param(csum, bool, 0444);
	32	module_param(gso, bool, 0444);
	33
296f96fc RR	34	/* FIXME: MTU in config. */
	35	#define MAX_PACKET_LEN (ETH_HLEN+ETH_DATA_LEN)
	36
	37	struct virtnet_info
	38	{
	39	struct virtio_device *vdev;
	40	struct virtqueue rvq, svq;
	41	struct net_device *dev;
	42	struct napi_struct napi;
	43
99ffc696 RR	44	/* The skb we couldn't send because buffers were full. */
	45	struct sk_buff *last_xmit_skb;
	46
296f96fc RR	47	/* Number of input buffers, and max we've ever had. */
	48	unsigned int num, max;
	49
11a3a154 RR	50	/* For cleaning up after transmission. */
	51	struct tasklet_struct tasklet;
	52
296f96fc RR	53	/* Receive & send queues. */
	54	struct sk_buff_head recv;
	55	struct sk_buff_head send;
	56	};
	57
	58	static inline struct virtio_net_hdr skb_vnet_hdr(struct sk_buff skb)
	59	{
	60	return (struct virtio_net_hdr *)skb->cb;
	61	}
	62
	63	static inline void vnet_hdr_to_sg(struct scatterlist sg, struct sk_buff skb)
	64	{
	65	sg_init_one(sg, skb_vnet_hdr(skb), sizeof(struct virtio_net_hdr));
	66	}
	67
2cb9c6ba	68	static void skb_xmit_done(struct virtqueue *svq)
296f96fc	69	{
2cb9c6ba	70	struct virtnet_info *vi = svq->vdev->priv;
296f96fc	71
2cb9c6ba RR	72	/* Suppress further interrupts. */
2cb9c6ba RR	73	svq->vq_ops->disable_cb(svq);
11a3a154	74
2cb9c6ba	75	/* We were waiting for more output buffers. */
296f96fc	76	netif_wake_queue(vi->dev);
11a3a154 RR	77
	78	/* Make sure we re-xmit last_xmit_skb: if there are no more packets
	79	* queued, start_xmit won't be called. */
	80	tasklet_schedule(&vi->tasklet);
296f96fc RR	81	}
	82
	83	static void receive_skb(struct net_device dev, struct sk_buff skb,
	84	unsigned len)
	85	{
	86	struct virtio_net_hdr *hdr = skb_vnet_hdr(skb);
	87
	88	if (unlikely(len < sizeof(struct virtio_net_hdr) + ETH_HLEN)) {
	89	pr_debug("%s: short packet %i\n", dev->name, len);
	90	dev->stats.rx_length_errors++;
	91	goto drop;
	92	}
	93	len -= sizeof(struct virtio_net_hdr);
	94	BUG_ON(len > MAX_PACKET_LEN);
	95
	96	skb_trim(skb, len);
23cde76d	97
296f96fc RR	98	dev->stats.rx_bytes += skb->len;
	99	dev->stats.rx_packets++;
	100
	101	if (hdr->flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) {
	102	pr_debug("Needs csum!\n");
f35d9d8a	103	if (!skb_partial_csum_set(skb,hdr->csum_start,hdr->csum_offset))
296f96fc	104	goto frame_err;
296f96fc RR	105	}
296f96fc RR	106
23cde76d MM	107	skb->protocol = eth_type_trans(skb, dev);
	108	pr_debug("Receiving skb proto 0x%04x len %i type %i\n",
	109	ntohs(skb->protocol), skb->len, skb->pkt_type);
	110
296f96fc RR	111	if (hdr->gso_type != VIRTIO_NET_HDR_GSO_NONE) {
296f96fc RR	112	pr_debug("GSO!\n");
34a48579	113	switch (hdr->gso_type & ~VIRTIO_NET_HDR_GSO_ECN) {
296f96fc RR	114	case VIRTIO_NET_HDR_GSO_TCPV4:
	115	skb_shinfo(skb)->gso_type = SKB_GSO_TCPV4;
	116	break;
296f96fc RR	117	case VIRTIO_NET_HDR_GSO_UDP:
	118	skb_shinfo(skb)->gso_type = SKB_GSO_UDP;
	119	break;
	120	case VIRTIO_NET_HDR_GSO_TCPV6:
	121	skb_shinfo(skb)->gso_type = SKB_GSO_TCPV6;
	122	break;
	123	default:
	124	if (net_ratelimit())
	125	printk(KERN_WARNING "%s: bad gso type %u.\n",
	126	dev->name, hdr->gso_type);
	127	goto frame_err;
	128	}
	129
34a48579 RR	130	if (hdr->gso_type & VIRTIO_NET_HDR_GSO_ECN)
	131	skb_shinfo(skb)->gso_type \|= SKB_GSO_TCP_ECN;
	132
296f96fc RR	133	skb_shinfo(skb)->gso_size = hdr->gso_size;
	134	if (skb_shinfo(skb)->gso_size == 0) {
	135	if (net_ratelimit())
	136	printk(KERN_WARNING "%s: zero gso size.\n",
	137	dev->name);
	138	goto frame_err;
	139	}
	140
	141	/* Header must be checked, and gso_segs computed. */
	142	skb_shinfo(skb)->gso_type \|= SKB_GSO_DODGY;
	143	skb_shinfo(skb)->gso_segs = 0;
	144	}
	145
	146	netif_receive_skb(skb);
	147	return;
	148
	149	frame_err:
	150	dev->stats.rx_frame_errors++;
	151	drop:
	152	dev_kfree_skb(skb);
	153	}
	154
	155	static void try_fill_recv(struct virtnet_info *vi)
	156	{
	157	struct sk_buff *skb;
05271685	158	struct scatterlist sg[2+MAX_SKB_FRAGS];
296f96fc RR	159	int num, err;
296f96fc RR	160
05271685	161	sg_init_table(sg, 2+MAX_SKB_FRAGS);
296f96fc RR	162	for (;;) {
	163	skb = netdev_alloc_skb(vi->dev, MAX_PACKET_LEN);
	164	if (unlikely(!skb))
	165	break;
	166
	167	skb_put(skb, MAX_PACKET_LEN);
	168	vnet_hdr_to_sg(sg, skb);
	169	num = skb_to_sgvec(skb, sg+1, 0, skb->len) + 1;
	170	skb_queue_head(&vi->recv, skb);
	171
	172	err = vi->rvq->vq_ops->add_buf(vi->rvq, sg, 0, num, skb);
	173	if (err) {
	174	skb_unlink(skb, &vi->recv);
	175	kfree_skb(skb);
	176	break;
	177	}
	178	vi->num++;
	179	}
	180	if (unlikely(vi->num > vi->max))
	181	vi->max = vi->num;
	182	vi->rvq->vq_ops->kick(vi->rvq);
	183	}
	184
18445c4d	185	static void skb_recv_done(struct virtqueue *rvq)
296f96fc RR	186	{
296f96fc RR	187	struct virtnet_info *vi = rvq->vdev->priv;
18445c4d RR	188	/* Schedule NAPI, Suppress further interrupts if successful. */
	189	if (netif_rx_schedule_prep(vi->dev, &vi->napi)) {
	190	rvq->vq_ops->disable_cb(rvq);
	191	__netif_rx_schedule(vi->dev, &vi->napi);
	192	}
296f96fc RR	193	}
	194
	195	static int virtnet_poll(struct napi_struct *napi, int budget)
	196	{
	197	struct virtnet_info *vi = container_of(napi, struct virtnet_info, napi);
	198	struct sk_buff *skb = NULL;
	199	unsigned int len, received = 0;
	200
	201	again:
	202	while (received < budget &&
	203	(skb = vi->rvq->vq_ops->get_buf(vi->rvq, &len)) != NULL) {
	204	__skb_unlink(skb, &vi->recv);
	205	receive_skb(vi->dev, skb, len);
	206	vi->num--;
	207	received++;
	208	}
	209
	210	/* FIXME: If we oom and completely run out of inbufs, we need
	211	* to start a timer trying to fill more. */
	212	if (vi->num < vi->max / 2)
	213	try_fill_recv(vi);
	214
8329d98e RR	215	/* Out of packets? */
8329d98e RR	216	if (received < budget) {
296f96fc	217	netif_rx_complete(vi->dev, napi);
18445c4d	218	if (unlikely(!vi->rvq->vq_ops->enable_cb(vi->rvq))
4265f161 CB	219	&& napi_schedule_prep(napi)) {
	220	vi->rvq->vq_ops->disable_cb(vi->rvq);
	221	__netif_rx_schedule(vi->dev, napi);
296f96fc	222	goto again;
4265f161	223	}
296f96fc RR	224	}
	225
	226	return received;
	227	}
	228
	229	static void free_old_xmit_skbs(struct virtnet_info *vi)
	230	{
	231	struct sk_buff *skb;
	232	unsigned int len;
	233
	234	while ((skb = vi->svq->vq_ops->get_buf(vi->svq, &len)) != NULL) {
	235	pr_debug("Sent skb %p\n", skb);
	236	__skb_unlink(skb, &vi->send);
655aa31f	237	vi->dev->stats.tx_bytes += skb->len;
296f96fc RR	238	vi->dev->stats.tx_packets++;
	239	kfree_skb(skb);
	240	}
	241	}
	242
99ffc696	243	static int xmit_skb(struct virtnet_info vi, struct sk_buff skb)
296f96fc	244	{
99ffc696	245	int num;
05271685	246	struct scatterlist sg[2+MAX_SKB_FRAGS];
296f96fc RR	247	struct virtio_net_hdr *hdr;
296f96fc RR	248	const unsigned char dest = ((struct ethhdr )skb->data)->h_dest;
296f96fc	249
05271685	250	sg_init_table(sg, 2+MAX_SKB_FRAGS);
4d125de3	251
99ffc696	252	pr_debug("%s: xmit %p " MAC_FMT "\n", vi->dev->name, skb,
21f644f3 DM	253	dest[0], dest[1], dest[2],
21f644f3 DM	254	dest[3], dest[4], dest[5]);
296f96fc	255
296f96fc RR	256	/* Encode metadata header at front. */
	257	hdr = skb_vnet_hdr(skb);
	258	if (skb->ip_summed == CHECKSUM_PARTIAL) {
	259	hdr->flags = VIRTIO_NET_HDR_F_NEEDS_CSUM;
	260	hdr->csum_start = skb->csum_start - skb_headroom(skb);
	261	hdr->csum_offset = skb->csum_offset;
	262	} else {
	263	hdr->flags = 0;
	264	hdr->csum_offset = hdr->csum_start = 0;
	265	}
	266
	267	if (skb_is_gso(skb)) {
50c8ea80	268	hdr->hdr_len = skb_transport_header(skb) - skb->data;
296f96fc	269	hdr->gso_size = skb_shinfo(skb)->gso_size;
34a48579	270	if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV4)
296f96fc RR	271	hdr->gso_type = VIRTIO_NET_HDR_GSO_TCPV4;
	272	else if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6)
	273	hdr->gso_type = VIRTIO_NET_HDR_GSO_TCPV6;
	274	else if (skb_shinfo(skb)->gso_type & SKB_GSO_UDP)
	275	hdr->gso_type = VIRTIO_NET_HDR_GSO_UDP;
	276	else
	277	BUG();
34a48579 RR	278	if (skb_shinfo(skb)->gso_type & SKB_GSO_TCP_ECN)
34a48579 RR	279	hdr->gso_type \|= VIRTIO_NET_HDR_GSO_ECN;
296f96fc RR	280	} else {
296f96fc RR	281	hdr->gso_type = VIRTIO_NET_HDR_GSO_NONE;
50c8ea80	282	hdr->gso_size = hdr->hdr_len = 0;
296f96fc RR	283	}
	284
	285	vnet_hdr_to_sg(sg, skb);
	286	num = skb_to_sgvec(skb, sg+1, 0, skb->len) + 1;
99ffc696 RR	287
	288	return vi->svq->vq_ops->add_buf(vi->svq, sg, num, 0, skb);
	289	}
	290
11a3a154 RR	291	static void xmit_tasklet(unsigned long data)
	292	{
	293	struct virtnet_info vi = (void )data;
	294
	295	netif_tx_lock_bh(vi->dev);
	296	if (vi->last_xmit_skb && xmit_skb(vi, vi->last_xmit_skb) == 0) {
	297	vi->svq->vq_ops->kick(vi->svq);
	298	vi->last_xmit_skb = NULL;
	299	}
	300	netif_tx_unlock_bh(vi->dev);
	301	}
	302
99ffc696 RR	303	static int start_xmit(struct sk_buff skb, struct net_device dev)
	304	{
	305	struct virtnet_info *vi = netdev_priv(dev);
2cb9c6ba RR	306
	307	again:
	308	/* Free up any pending old buffers before queueing new ones. */
	309	free_old_xmit_skbs(vi);
99ffc696 RR	310
99ffc696 RR	311	/* If we has a buffer left over from last time, send it now. */
7eb2e251	312	if (unlikely(vi->last_xmit_skb)) {
99ffc696 RR	313	if (xmit_skb(vi, vi->last_xmit_skb) != 0) {
	314	/* Drop this skb: we only queue one. */
	315	vi->dev->stats.tx_dropped++;
	316	kfree_skb(skb);
7eb2e251	317	skb = NULL;
99ffc696	318	goto stop_queue;
2cb9c6ba	319	}
99ffc696 RR	320	vi->last_xmit_skb = NULL;
99ffc696 RR	321	}
2cb9c6ba	322
99ffc696	323	/* Put new one in send queue and do transmit */
7eb2e251 RR	324	if (likely(skb)) {
	325	__skb_queue_head(&vi->send, skb);
	326	if (xmit_skb(vi, skb) != 0) {
	327	vi->last_xmit_skb = skb;
	328	skb = NULL;
	329	goto stop_queue;
	330	}
296f96fc	331	}
99ffc696	332	done:
296f96fc	333	vi->svq->vq_ops->kick(vi->svq);
99ffc696 RR	334	return NETDEV_TX_OK;
	335
	336	stop_queue:
	337	pr_debug("%s: virtio not prepared to send\n", dev->name);
	338	netif_stop_queue(dev);
	339
	340	/* Activate callback for using skbs: if this returns false it
	341	* means some were used in the meantime. */
	342	if (unlikely(!vi->svq->vq_ops->enable_cb(vi->svq))) {
	343	vi->svq->vq_ops->disable_cb(vi->svq);
	344	netif_start_queue(dev);
	345	goto again;
	346	}
	347	goto done;
296f96fc RR	348	}
296f96fc RR	349
da74e89d AS	350	#ifdef CONFIG_NET_POLL_CONTROLLER
	351	static void virtnet_netpoll(struct net_device *dev)
	352	{
	353	struct virtnet_info *vi = netdev_priv(dev);
	354
	355	napi_schedule(&vi->napi);
	356	}
	357	#endif
	358
296f96fc RR	359	static int virtnet_open(struct net_device *dev)
	360	{
	361	struct virtnet_info *vi = netdev_priv(dev);
	362
296f96fc	363	napi_enable(&vi->napi);
a48bd8f6 RR	364
	365	/* If all buffers were filled by other side before we napi_enabled, we
	366	* won't get another interrupt, so process any outstanding packets
370076d9 CB	367	* now. virtnet_poll wants re-enable the queue, so we disable here.
	368	* We synchronize against interrupts via NAPI_STATE_SCHED */
	369	if (netif_rx_schedule_prep(dev, &vi->napi)) {
	370	vi->rvq->vq_ops->disable_cb(vi->rvq);
	371	__netif_rx_schedule(dev, &vi->napi);
	372	}
296f96fc RR	373	return 0;
	374	}
	375
	376	static int virtnet_close(struct net_device *dev)
	377	{
	378	struct virtnet_info *vi = netdev_priv(dev);
296f96fc RR	379
	380	napi_disable(&vi->napi);
	381
296f96fc RR	382	return 0;
	383	}
	384
	385	static int virtnet_probe(struct virtio_device *vdev)
	386	{
	387	int err;
296f96fc RR	388	struct net_device *dev;
296f96fc RR	389	struct virtnet_info *vi;
296f96fc RR	390
	391	/* Allocate ourselves a network device with room for our info */
	392	dev = alloc_etherdev(sizeof(struct virtnet_info));
	393	if (!dev)
	394	return -ENOMEM;
	395
	396	/* Set up network device as normal. */
296f96fc RR	397	dev->open = virtnet_open;
	398	dev->stop = virtnet_close;
	399	dev->hard_start_xmit = start_xmit;
	400	dev->features = NETIF_F_HIGHDMA;
da74e89d AS	401	#ifdef CONFIG_NET_POLL_CONTROLLER
	402	dev->poll_controller = virtnet_netpoll;
	403	#endif
296f96fc RR	404	SET_NETDEV_DEV(dev, &vdev->dev);
	405
	406	/* Do we support "hardware" checksums? */
c45a6816	407	if (csum && virtio_has_feature(vdev, VIRTIO_NET_F_CSUM)) {
296f96fc RR	408	/* This opens up the world of extra features. */
296f96fc RR	409	dev->features \|= NETIF_F_HW_CSUM\|NETIF_F_SG\|NETIF_F_FRAGLIST;
c45a6816	410	if (gso && virtio_has_feature(vdev, VIRTIO_NET_F_GSO)) {
34a48579 RR	411	dev->features \|= NETIF_F_TSO \| NETIF_F_UFO
	412	\| NETIF_F_TSO_ECN \| NETIF_F_TSO6;
	413	}
5539ae96	414	/* Individual feature bits: what can host handle? */
c45a6816	415	if (gso && virtio_has_feature(vdev, VIRTIO_NET_F_HOST_TSO4))
5539ae96	416	dev->features \|= NETIF_F_TSO;
c45a6816	417	if (gso && virtio_has_feature(vdev, VIRTIO_NET_F_HOST_TSO6))
5539ae96	418	dev->features \|= NETIF_F_TSO6;
c45a6816	419	if (gso && virtio_has_feature(vdev, VIRTIO_NET_F_HOST_ECN))
5539ae96	420	dev->features \|= NETIF_F_TSO_ECN;
c45a6816	421	if (gso && virtio_has_feature(vdev, VIRTIO_NET_F_HOST_UFO))
5539ae96	422	dev->features \|= NETIF_F_UFO;
296f96fc RR	423	}
	424
	425	/* Configuration may specify what MAC to use. Otherwise random. */
c45a6816	426	if (virtio_has_feature(vdev, VIRTIO_NET_F_MAC)) {
a586d4f6 RR	427	vdev->config->get(vdev,
	428	offsetof(struct virtio_net_config, mac),
	429	dev->dev_addr, dev->addr_len);
296f96fc RR	430	} else
	431	random_ether_addr(dev->dev_addr);
	432
	433	/* Set up our device-specific information */
	434	vi = netdev_priv(dev);
6c0cd7c0	435	netif_napi_add(dev, &vi->napi, virtnet_poll, napi_weight);
296f96fc RR	436	vi->dev = dev;
296f96fc RR	437	vi->vdev = vdev;
d9d5dcc8	438	vdev->priv = vi;
296f96fc RR	439
296f96fc RR	440	/* We expect two virtqueues, receive then send. */
a586d4f6	441	vi->rvq = vdev->config->find_vq(vdev, 0, skb_recv_done);
296f96fc RR	442	if (IS_ERR(vi->rvq)) {
	443	err = PTR_ERR(vi->rvq);
	444	goto free;
	445	}
	446
a586d4f6	447	vi->svq = vdev->config->find_vq(vdev, 1, skb_xmit_done);
296f96fc RR	448	if (IS_ERR(vi->svq)) {
	449	err = PTR_ERR(vi->svq);
	450	goto free_recv;
	451	}
	452
	453	/* Initialize our empty receive and send queues. */
	454	skb_queue_head_init(&vi->recv);
	455	skb_queue_head_init(&vi->send);
	456
11a3a154 RR	457	tasklet_init(&vi->tasklet, xmit_tasklet, (unsigned long)vi);
11a3a154 RR	458
296f96fc RR	459	err = register_netdev(dev);
	460	if (err) {
	461	pr_debug("virtio_net: registering device failed\n");
	462	goto free_send;
	463	}
b3369c1f RR	464
	465	/* Last of all, set up some receive buffers. */
	466	try_fill_recv(vi);
	467
	468	/* If we didn't even get one input buffer, we're useless. */
	469	if (vi->num == 0) {
	470	err = -ENOMEM;
	471	goto unregister;
	472	}
	473
296f96fc	474	pr_debug("virtnet: registered device %s\n", dev->name);
296f96fc RR	475	return 0;
296f96fc RR	476
b3369c1f RR	477	unregister:
b3369c1f RR	478	unregister_netdev(dev);
296f96fc RR	479	free_send:
	480	vdev->config->del_vq(vi->svq);
	481	free_recv:
	482	vdev->config->del_vq(vi->rvq);
	483	free:
	484	free_netdev(dev);
	485	return err;
	486	}
	487
	488	static void virtnet_remove(struct virtio_device *vdev)
	489	{
74b2553f	490	struct virtnet_info *vi = vdev->priv;
b3369c1f RR	491	struct sk_buff *skb;
b3369c1f RR	492
6e5aa7ef RR	493	/* Stop all the virtqueues. */
	494	vdev->config->reset(vdev);
	495
b3369c1f	496	/* Free our skbs in send and recv queues, if any. */
b3369c1f RR	497	while ((skb = __skb_dequeue(&vi->recv)) != NULL) {
	498	kfree_skb(skb);
	499	vi->num--;
	500	}
288369cc	501	__skb_queue_purge(&vi->send);
b3369c1f RR	502
b3369c1f RR	503	BUG_ON(vi->num != 0);
74b2553f RR	504
	505	vdev->config->del_vq(vi->svq);
	506	vdev->config->del_vq(vi->rvq);
	507	unregister_netdev(vi->dev);
	508	free_netdev(vi->dev);
296f96fc RR	509	}
	510
	511	static struct virtio_device_id id_table[] = {
	512	{ VIRTIO_ID_NET, VIRTIO_DEV_ANY_ID },
	513	{ 0 },
	514	};
	515
c45a6816 RR	516	static unsigned int features[] = {
	517	VIRTIO_NET_F_CSUM, VIRTIO_NET_F_GSO, VIRTIO_NET_F_MAC,
	518	VIRTIO_NET_F_HOST_TSO4, VIRTIO_NET_F_HOST_UFO, VIRTIO_NET_F_HOST_TSO6,
	519	VIRTIO_NET_F_HOST_ECN,
	520	};
	521
296f96fc	522	static struct virtio_driver virtio_net = {
c45a6816 RR	523	.feature_table = features,
c45a6816 RR	524	.feature_table_size = ARRAY_SIZE(features),
296f96fc RR	525	.driver.name = KBUILD_MODNAME,
	526	.driver.owner = THIS_MODULE,
	527	.id_table = id_table,
	528	.probe = virtnet_probe,
	529	.remove = __devexit_p(virtnet_remove),
	530	};
	531
	532	static int __init init(void)
	533	{
	534	return register_virtio_driver(&virtio_net);
	535	}
	536
	537	static void __exit fini(void)
	538	{
	539	unregister_virtio_driver(&virtio_net);
	540	}
	541	module_init(init);
	542	module_exit(fini);
	543
	544	MODULE_DEVICE_TABLE(virtio, id_table);
	545	MODULE_DESCRIPTION("Virtio network driver");
	546	MODULE_LICENSE("GPL");