xen-netfront.c

/*
 * Virtual network driver for conversing with remote driver backends.
 *
 * Copyright (c) 2002-2005, K A Fraser
 * Copyright (c) 2005, XenSource Ltd
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License version 2
 * as published by the Free Software Foundation; or, when distributed
 * separately from the Linux kernel or incorporated into other
 * software packages, subject to the following license:
 *
 * Permission is hereby granted, free of charge, to any person obtaining a copy
 * of this source file (the "Software"), to deal in the Software without
 * restriction, including without limitation the rights to use, copy, modify,
 * merge, publish, distribute, sublicense, and/or sell copies of the Software,
 * and to permit persons to whom the Software is furnished to do so, subject to
 * the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in
 * all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
 * IN THE SOFTWARE.
 */

#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/skbuff.h>
#include <linux/ethtool.h>
#include <linux/if_ether.h>
#include <linux/tcp.h>
#include <linux/udp.h>
#include <linux/moduleparam.h>
#include <linux/mm.h>
#include <linux/slab.h>
#include <net/ip.h>

#include <xen/xen.h>
#include <xen/xenbus.h>
#include <xen/events.h>
#include <xen/page.h>
#include <xen/grant_table.h>

#include <xen/interface/io/netif.h>
#include <xen/interface/memory.h>
#include <xen/interface/grant_table.h>

static const struct ethtool_ops xennet_ethtool_ops;

struct netfront_cb {
	struct page *page;
	unsigned offset;
};

#define NETFRONT_SKB_CB(skb)	((struct netfront_cb *)((skb)->cb))

#define RX_COPY_THRESHOLD 256

#define GRANT_INVALID_REF	0

#define NET_TX_RING_SIZE __CONST_RING_SIZE(xen_netif_tx, PAGE_SIZE)
#define NET_RX_RING_SIZE __CONST_RING_SIZE(xen_netif_rx, PAGE_SIZE)
#define TX_MAX_TARGET min_t(int, NET_RX_RING_SIZE, 256)

struct netfront_info {
	struct list_head list;
	struct net_device *netdev;

	struct napi_struct napi;

	unsigned int evtchn;
	struct xenbus_device *xbdev;

	spinlock_t   tx_lock;
	struct xen_netif_tx_front_ring tx;
	int tx_ring_ref;

	/*
	 * {tx,rx}_skbs store outstanding skbuffs. Free tx_skb entries
	 * are linked from tx_skb_freelist through skb_entry.link.
	 *
	 *  NB. Freelist index entries are always going to be less than
	 *  PAGE_OFFSET, whereas pointers to skbs will always be equal or
	 *  greater than PAGE_OFFSET: we use this property to distinguish
	 *  them.
	 */
	union skb_entry {
		struct sk_buff *skb;
		unsigned long link;
	} tx_skbs[NET_TX_RING_SIZE];
	grant_ref_t gref_tx_head;
	grant_ref_t grant_tx_ref[NET_TX_RING_SIZE];
	unsigned tx_skb_freelist;

	spinlock_t   rx_lock ____cacheline_aligned_in_smp;
	struct xen_netif_rx_front_ring rx;
	int rx_ring_ref;

	/* Receive-ring batched refills. */
#define RX_MIN_TARGET 8
#define RX_DFL_MIN_TARGET 64
#define RX_MAX_TARGET min_t(int, NET_RX_RING_SIZE, 256)
	unsigned rx_min_target, rx_max_target, rx_target;
	struct sk_buff_head rx_batch;

	struct timer_list rx_refill_timer;

	struct sk_buff *rx_skbs[NET_RX_RING_SIZE];
	grant_ref_t gref_rx_head;
	grant_ref_t grant_rx_ref[NET_RX_RING_SIZE];

	unsigned long rx_pfn_array[NET_RX_RING_SIZE];
	struct multicall_entry rx_mcl[NET_RX_RING_SIZE+1];
	struct mmu_update rx_mmu[NET_RX_RING_SIZE];

	/* Statistics */
	unsigned long rx_gso_checksum_fixup;
};

struct netfront_rx_info {
	struct xen_netif_rx_response rx;
	struct xen_netif_extra_info extras[XEN_NETIF_EXTRA_TYPE_MAX - 1];
};

static void skb_entry_set_link(union skb_entry *list, unsigned short id)
{
	list->link = id;
}

static int skb_entry_is_link(const union skb_entry *list)
{
	BUILD_BUG_ON(sizeof(list->skb) != sizeof(list->link));
	return (unsigned long)list->skb < PAGE_OFFSET;
}

/*
 * Access macros for acquiring freeing slots in tx_skbs[].
 */

static void add_id_to_freelist(unsigned *head, union skb_entry *list,
			       unsigned short id)
{
	skb_entry_set_link(&list[id], *head);
	*head = id;
}

static unsigned short get_id_from_freelist(unsigned *head,
					   union skb_entry *list)
{
	unsigned int id = *head;
	*head = list[id].link;
	return id;
}

static int xennet_rxidx(RING_IDX idx)
{
	return idx & (NET_RX_RING_SIZE - 1);
}

static struct sk_buff *xennet_get_rx_skb(struct netfront_info *np,
					 RING_IDX ri)
{
	int i = xennet_rxidx(ri);
	struct sk_buff *skb = np->rx_skbs[i];
	np->rx_skbs[i] = NULL;
	return skb;
}

static grant_ref_t xennet_get_rx_ref(struct netfront_info *np,
					    RING_IDX ri)
{
	int i = xennet_rxidx(ri);
	grant_ref_t ref = np->grant_rx_ref[i];
	np->grant_rx_ref[i] = GRANT_INVALID_REF;
	return ref;
}

#ifdef CONFIG_SYSFS
static int xennet_sysfs_addif(struct net_device *netdev);
static void xennet_sysfs_delif(struct net_device *netdev);
#else /* !CONFIG_SYSFS */
#define xennet_sysfs_addif(dev) (0)
#define xennet_sysfs_delif(dev) do { } while (0)
#endif

static int xennet_can_sg(struct net_device *dev)
{
	return dev->features & NETIF_F_SG;
}


static void rx_refill_timeout(unsigned long data)
{
	struct net_device *dev = (struct net_device *)data;
	struct netfront_info *np = netdev_priv(dev);
	napi_schedule(&np->napi);
}

static int netfront_tx_slot_available(struct netfront_info *np)
{
	return (np->tx.req_prod_pvt - np->tx.rsp_cons) <
		(TX_MAX_TARGET - MAX_SKB_FRAGS - 2);
}

static void xennet_maybe_wake_tx(struct net_device *dev)
{
	struct netfront_info *np = netdev_priv(dev);

	if (unlikely(netif_queue_stopped(dev)) &&
	    netfront_tx_slot_available(np) &&
	    likely(netif_running(dev)))
		netif_wake_queue(dev);
}

static void xennet_alloc_rx_buffers(struct net_device *dev)
{
	unsigned short id;
	struct netfront_info *np = netdev_priv(dev);
	struct sk_buff *skb;
	struct page *page;
	int i, batch_target, notify;
	RING_IDX req_prod = np->rx.req_prod_pvt;
	grant_ref_t ref;
	unsigned long pfn;
	void *vaddr;
	struct xen_netif_rx_request *req;

	if (unlikely(!netif_carrier_ok(dev)))
		return;

	/*
	 * Allocate skbuffs greedily, even though we batch updates to the
	 * receive ring. This creates a less bursty demand on the memory
	 * allocator, so should reduce the chance of failed allocation requests
	 * both for ourself and for other kernel subsystems.
	 */
	batch_target = np->rx_target - (req_prod - np->rx.rsp_cons);
	for (i = skb_queue_len(&np->rx_batch); i < batch_target; i++) {
		skb = __netdev_alloc_skb(dev, RX_COPY_THRESHOLD + NET_IP_ALIGN,
					 GFP_ATOMIC | __GFP_NOWARN);
		if (unlikely(!skb))
			goto no_skb;

		/* Align ip header to a 16 bytes boundary */
		skb_reserve(skb, NET_IP_ALIGN);

		page = alloc_page(GFP_ATOMIC | __GFP_NOWARN);
		if (!page) {
			kfree_skb(skb);
no_skb:
			/* Any skbuffs queued for refill? Force them out. */
			if (i != 0)
				goto refill;
			/* Could not allocate any skbuffs. Try again later. */
			mod_timer(&np->rx_refill_timer,
				  jiffies + (HZ/10));
			break;
		}

		skb_shinfo(skb)->frags[0].page = page;
		skb_shinfo(skb)->nr_frags = 1;
		__skb_queue_tail(&np->rx_batch, skb);
	}

	/* Is the batch large enough to be worthwhile? */
	if (i < (np->rx_target/2)) {
		if (req_prod > np->rx.sring->req_prod)
			goto push;
		return;
	}

	/* Adjust our fill target if we risked running out of buffers. */
	if (((req_prod - np->rx.sring->rsp_prod) < (np->rx_target / 4)) &&
	    ((np->rx_target *= 2) > np->rx_max_target))
		np->rx_target = np->rx_max_target;

 refill:
	for (i = 0; ; i++) {
		skb = __skb_dequeue(&np->rx_batch);
		if (skb == NULL)
			break;

		skb->dev = dev;

		id = xennet_rxidx(req_prod + i);

		BUG_ON(np->rx_skbs[id]);
		np->rx_skbs[id] = skb;

		ref = gnttab_claim_grant_reference(&np->gref_rx_head);
		BUG_ON((signed short)ref < 0);
		np->grant_rx_ref[id] = ref;

		pfn = page_to_pfn(skb_shinfo(skb)->frags[0].page);
		vaddr = page_address(skb_shinfo(skb)->frags[0].page);

		req = RING_GET_REQUEST(&np->rx, req_prod + i);
		gnttab_grant_foreign_access_ref(ref,
						np->xbdev->otherend_id,
						pfn_to_mfn(pfn),
						0);

		req->id = id;
		req->gref = ref;
	}

	wmb();		/* barrier so backend seens requests */

	/* Above is a suitable barrier to ensure backend will see requests. */
	np->rx.req_prod_pvt = req_prod + i;
 push:
	RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&np->rx, notify);
	if (notify)
		notify_remote_via_irq(np->netdev->irq);
}

static int xennet_open(struct net_device *dev)
{
	struct netfront_info *np = netdev_priv(dev);

	napi_enable(&np->napi);

	spin_lock_bh(&np->rx_lock);
	if (netif_carrier_ok(dev)) {
		xennet_alloc_rx_buffers(dev);
		np->rx.sring->rsp_event = np->rx.rsp_cons + 1;
		if (RING_HAS_UNCONSUMED_RESPONSES(&np->rx))
			napi_schedule(&np->napi);
	}
	spin_unlock_bh(&np->rx_lock);

	netif_start_queue(dev);

	return 0;
}

static void xennet_tx_buf_gc(struct net_device *dev)
{
	RING_IDX cons, prod;
	unsigned short id;
	struct netfront_info *np = netdev_priv(dev);
	struct sk_buff *skb;

	BUG_ON(!netif_carrier_ok(dev));

	do {
		prod = np->tx.sring->rsp_prod;
		rmb(); /* Ensure we see responses up to 'rp'. */

		for (cons = np->tx.rsp_cons; cons != prod; cons++) {
			struct xen_netif_tx_response *txrsp;

			txrsp = RING_GET_RESPONSE(&np->tx, cons);
			if (txrsp->status == XEN_NETIF_RSP_NULL)
				continue;

			id  = txrsp->id;
			skb = np->tx_skbs[id].skb;
			if (unlikely(gnttab_query_foreign_access(
				np->grant_tx_ref[id]) != 0)) {
				printk(KERN_ALERT "xennet_tx_buf_gc: warning "
				       "-- grant still in use by backend "
				       "domain.\n");
				BUG();
			}
			gnttab_end_foreign_access_ref(
				np->grant_tx_ref[id], GNTMAP_readonly);
			gnttab_release_grant_reference(
				&np->gref_tx_head, np->grant_tx_ref[id]);
			np->grant_tx_ref[id] = GRANT_INVALID_REF;
			add_id_to_freelist(&np->tx_skb_freelist, np->tx_skbs, id);
			dev_kfree_skb_irq(skb);
		}

		np->tx.rsp_cons = prod;

		/*
		 * Set a new event, then check for race with update of tx_cons.
		 * Note that it is essential to schedule a callback, no matter
		 * how few buffers are pending. Even if there is space in the
		 * transmit ring, higher layers may be blocked because too much
		 * data is outstanding: in such cases notification from Xen is
		 * likely to be the only kick that we'll get.
		 */
		np->tx.sring->rsp_event =
			prod + ((np->tx.sring->req_prod - prod) >> 1) + 1;
		mb();		/* update shared area */
	} while ((cons == prod) && (prod != np->tx.sring->rsp_prod));

	xennet_maybe_wake_tx(dev);
}

static void xennet_make_frags(struct sk_buff *skb, struct net_device *dev,
			      struct xen_netif_tx_request *tx)
{
	struct netfront_info *np = netdev_priv(dev);
	char *data = skb->data;
	unsigned long mfn;
	RING_IDX prod = np->tx.req_prod_pvt;
	int frags = skb_shinfo(skb)->nr_frags;
	unsigned int offset = offset_in_page(data);
	unsigned int len = skb_headlen(skb);
	unsigned int id;
	grant_ref_t ref;
	int i;

	/* While the header overlaps a page boundary (including being
	   larger than a page), split it it into page-sized chunks. */
	while (len > PAGE_SIZE - offset) {
		tx->size = PAGE_SIZE - offset;
		tx->flags |= XEN_NETTXF_more_data;
		len -= tx->size;
		data += tx->size;
		offset = 0;

		id = get_id_from_freelist(&np->tx_skb_freelist, np->tx_skbs);
		np->tx_skbs[id].skb = skb_get(skb);
		tx = RING_GET_REQUEST(&np->tx, prod++);
		tx->id = id;
		ref = gnttab_claim_grant_reference(&np->gref_tx_head);
		BUG_ON((signed short)ref < 0);

		mfn = virt_to_mfn(data);
		gnttab_grant_foreign_access_ref(ref, np->xbdev->otherend_id,
						mfn, GNTMAP_readonly);

		tx->gref = np->grant_tx_ref[id] = ref;
		tx->offset = offset;
		tx->size = len;
		tx->flags = 0;
	}

	/* Grant backend access to each skb fragment page. */
	for (i = 0; i < frags; i++) {
		skb_frag_t *frag = skb_shinfo(skb)->frags + i;

		tx->flags |= XEN_NETTXF_more_data;

		id = get_id_from_freelist(&np->tx_skb_freelist, np->tx_skbs);
		np->tx_skbs[id].skb = skb_get(skb);
		tx = RING_GET_REQUEST(&np->tx, prod++);
		tx->id = id;
		ref = gnttab_claim_grant_reference(&np->gref_tx_head);
		BUG_ON((signed short)ref < 0);

		mfn = pfn_to_mfn(page_to_pfn(frag->page));
		gnttab_grant_foreign_access_ref(ref, np->xbdev->otherend_id,
						mfn, GNTMAP_readonly);

		tx->gref = np->grant_tx_ref[id] = ref;
		tx->offset = frag->page_offset;
		tx->size = frag->size;
		tx->flags = 0;
	}

	np->tx.req_prod_pvt = prod;
}

static int xennet_start_xmit(struct sk_buff *skb, struct net_device *dev)
{
	unsigned short id;
	struct netfront_info *np = netdev_priv(dev);
	struct xen_netif_tx_request *tx;
	struct xen_netif_extra_info *extra;
	char *data = skb->data;
	RING_IDX i;
	grant_ref_t ref;
	unsigned long mfn;
	int notify;
	int frags = skb_shinfo(skb)->nr_frags;
	unsigned int offset = offset_in_page(data);
	unsigned int len = skb_headlen(skb);

	frags += DIV_ROUND_UP(offset + len, PAGE_SIZE);
	if (unlikely(frags > MAX_SKB_FRAGS + 1)) {
		printk(KERN_ALERT "xennet: skb rides the rocket: %d frags\n",
		       frags);
		dump_stack();
		goto drop;
	}

	spin_lock_irq(&np->tx_lock);

	if (unlikely(!netif_carrier_ok(dev) ||
		     (frags > 1 && !xennet_can_sg(dev)) ||
		     netif_needs_gso(skb, netif_skb_features(skb)))) {
		spin_unlock_irq(&np->tx_lock);
		goto drop;
	}

	i = np->tx.req_prod_pvt;

	id = get_id_from_freelist(&np->tx_skb_freelist, np->tx_skbs);
	np->tx_skbs[id].skb = skb;

	tx = RING_GET_REQUEST(&np->tx, i);

	tx->id   = id;
	ref = gnttab_claim_grant_reference(&np->gref_tx_head);
	BUG_ON((signed short)ref < 0);
	mfn = virt_to_mfn(data);
	gnttab_grant_foreign_access_ref(
		ref, np->xbdev->otherend_id, mfn, GNTMAP_readonly);
	tx->gref = np->grant_tx_ref[id] = ref;
	tx->offset = offset;
	tx->size = len;
	extra = NULL;

	tx->flags = 0;
	if (skb->ip_summed == CHECKSUM_PARTIAL)
		/* local packet? */
		tx->flags |= XEN_NETTXF_csum_blank | XEN_NETTXF_data_validated;
	else if (skb->ip_summed == CHECKSUM_UNNECESSARY)
		/* remote but checksummed. */
		tx->flags |= XEN_NETTXF_data_validated;

	if (skb_shinfo(skb)->gso_size) {
		struct xen_netif_extra_info *gso;

		gso = (struct xen_netif_extra_info *)
			RING_GET_REQUEST(&np->tx, ++i);

		if (extra)
			extra->flags |= XEN_NETIF_EXTRA_FLAG_MORE;
		else
			tx->flags |= XEN_NETTXF_extra_info;

		gso->u.gso.size = skb_shinfo(skb)->gso_size;
		gso->u.gso.type = XEN_NETIF_GSO_TYPE_TCPV4;
		gso->u.gso.pad = 0;
		gso->u.gso.features = 0;

		gso->type = XEN_NETIF_EXTRA_TYPE_GSO;
		gso->flags = 0;
		extra = gso;
	}

	np->tx.req_prod_pvt = i + 1;

	xennet_make_frags(skb, dev, tx);
	tx->size = skb->len;

	RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&np->tx, notify);
	if (notify)
		notify_remote_via_irq(np->netdev->irq);

	dev->stats.tx_bytes += skb->len;
	dev->stats.tx_packets++;

	/* Note: It is not safe to access skb after xennet_tx_buf_gc()! */
	xennet_tx_buf_gc(dev);

	if (!netfront_tx_slot_available(np))
		netif_stop_queue(dev);

	spin_unlock_irq(&np->tx_lock);

	return NETDEV_TX_OK;

 drop:
	dev->stats.tx_dropped++;
	dev_kfree_skb(skb);
	return NETDEV_TX_OK;
}

static int xennet_close(struct net_device *dev)
{
	struct netfront_info *np = netdev_priv(dev);
	netif_stop_queue(np->netdev);
	napi_disable(&np->napi);
	return 0;
}

static void xennet_move_rx_slot(struct netfront_info *np, struct sk_buff *skb,
				grant_ref_t ref)
{
	int new = xennet_rxidx(np->rx.req_prod_pvt);

	BUG_ON(np->rx_skbs[new]);
	np->rx_skbs[new] = skb;
	np->grant_rx_ref[new] = ref;
	RING_GET_REQUEST(&np->rx, np->rx.req_prod_pvt)->id = new;
	RING_GET_REQUEST(&np->rx, np->rx.req_prod_pvt)->gref = ref;
	np->rx.req_prod_pvt++;
}

static int xennet_get_extras(struct netfront_info *np,
			     struct xen_netif_extra_info *extras,
			     RING_IDX rp)

{
	struct xen_netif_extra_info *extra;
	struct device *dev = &np->netdev->dev;
	RING_IDX cons = np->rx.rsp_cons;
	int err = 0;

	do {
		struct sk_buff *skb;
		grant_ref_t ref;

		if (unlikely(cons + 1 == rp)) {
			if (net_ratelimit())
				dev_warn(dev, "Missing extra info\n");
			err = -EBADR;
			break;
		}

		extra = (struct xen_netif_extra_info *)
			RING_GET_RESPONSE(&np->rx, ++cons);

		if (unlikely(!extra->type ||
			     extra->type >= XEN_NETIF_EXTRA_TYPE_MAX)) {
			if (net_ratelimit())
				dev_warn(dev, "Invalid extra type: %d\n",
					extra->type);
			err = -EINVAL;
		} else {
			memcpy(&extras[extra->type - 1], extra,
			       sizeof(*extra));
		}

		skb = xennet_get_rx_skb(np, cons);
		ref = xennet_get_rx_ref(np, cons);
		xennet_move_rx_slot(np, skb, ref);
	} while (extra->flags & XEN_NETIF_EXTRA_FLAG_MORE);

	np->rx.rsp_cons = cons;
	return err;
}

static int xennet_get_responses(struct netfront_info *np,
				struct netfront_rx_info *rinfo, RING_IDX rp,
				struct sk_buff_head *list)
{
	struct xen_netif_rx_response *rx = &rinfo->rx;
	struct xen_netif_extra_info *extras = rinfo->extras;
	struct device *dev = &np->netdev->dev;
	RING_IDX cons = np->rx.rsp_cons;
	struct sk_buff *skb = xennet_get_rx_skb(np, cons);
	grant_ref_t ref = xennet_get_rx_ref(np, cons);
	int max = MAX_SKB_FRAGS + (rx->status <= RX_COPY_THRESHOLD);
	int frags = 1;
	int err = 0;
	unsigned long ret;

	if (rx->flags & XEN_NETRXF_extra_info) {
		err = xennet_get_extras(np, extras, rp);
		cons = np->rx.rsp_cons;
	}

	for (;;) {
		if (unlikely(rx->status < 0 ||
			     rx->offset + rx->status > PAGE_SIZE)) {
			if (net_ratelimit())
				dev_warn(dev, "rx->offset: %x, size: %u\n",
					 rx->offset, rx->status);
			xennet_move_rx_slot(np, skb, ref);
			err = -EINVAL;
			goto next;
		}

		/*
		 * This definitely indicates a bug, either in this driver or in
		 * the backend driver. In future this should flag the bad
		 * situation to the system controller to reboot the backed.
		 */
		if (ref == GRANT_INVALID_REF) {
			if (net_ratelimit())
				dev_warn(dev, "Bad rx response id %d.\n",
					 rx->id);
			err = -EINVAL;
			goto next;
		}

		ret = gnttab_end_foreign_access_ref(ref, 0);
		BUG_ON(!ret);

		gnttab_release_grant_reference(&np->gref_rx_head, ref);

		__skb_queue_tail(list, skb);

next:
		if (!(rx->flags & XEN_NETRXF_more_data))
			break;

		if (cons + frags == rp) {
			if (net_ratelimit())
				dev_warn(dev, "Need more frags\n");
			err = -ENOENT;
			break;
		}

		rx = RING_GET_RESPONSE(&np->rx, cons + frags);
		skb = xennet_get_rx_skb(np, cons + frags);
		ref = xennet_get_rx_ref(np, cons + frags);
		frags++;
	}

	if (unlikely(frags > max)) {
		if (net_ratelimit())
			dev_warn(dev, "Too many frags\n");
		err = -E2BIG;
	}

	if (unlikely(err))
		np->rx.rsp_cons = cons + frags;

	return err;
}

static int xennet_set_skb_gso(struct sk_buff *skb,
			      struct xen_netif_extra_info *gso)
{
	if (!gso->u.gso.size) {
		if (net_ratelimit())
			printk(KERN_WARNING "GSO size must not be zero.\n");
		return -EINVAL;
	}

	/* Currently only TCPv4 S.O. is supported. */
	if (gso->u.gso.type != XEN_NETIF_GSO_TYPE_TCPV4) {
		if (net_ratelimit())
			printk(KERN_WARNING "Bad GSO type %d.\n", gso->u.gso.type);
		return -EINVAL;
	}

	skb_shinfo(skb)->gso_size = gso->u.gso.size;
	skb_shinfo(skb)->gso_type = SKB_GSO_TCPV4;

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

	return 0;
}

static RING_IDX xennet_fill_frags(struct netfront_info *np,
				  struct sk_buff *skb,
				  struct sk_buff_head *list)
{
	struct skb_shared_info *shinfo = skb_shinfo(skb);
	int nr_frags = shinfo->nr_frags;
	RING_IDX cons = np->rx.rsp_cons;
	skb_frag_t *frag = shinfo->frags + nr_frags;
	struct sk_buff *nskb;

	while ((nskb = __skb_dequeue(list))) {
		struct xen_netif_rx_response *rx =
			RING_GET_RESPONSE(&np->rx, ++cons);

		frag->page = skb_shinfo(nskb)->frags[0].page;
		frag->page_offset = rx->offset;
		frag->size = rx->status;

		skb->data_len += rx->status;

		skb_shinfo(nskb)->nr_frags = 0;
		kfree_skb(nskb);

		frag++;
		nr_frags++;
	}

	shinfo->nr_frags = nr_frags;
	return cons;
}

static int checksum_setup(struct net_device *dev, struct sk_buff *skb)
{
	struct iphdr *iph;
	unsigned char *th;
	int err = -EPROTO;
	int recalculate_partial_csum = 0;

	/*
	 * A GSO SKB must be CHECKSUM_PARTIAL. However some buggy
	 * peers can fail to set NETRXF_csum_blank when sending a GSO
	 * frame. In this case force the SKB to CHECKSUM_PARTIAL and
	 * recalculate the partial checksum.
	 */
	if (skb->ip_summed != CHECKSUM_PARTIAL && skb_is_gso(skb)) {
		struct netfront_info *np = netdev_priv(dev);
		np->rx_gso_checksum_fixup++;
		skb->ip_summed = CHECKSUM_PARTIAL;
		recalculate_partial_csum = 1;
	}

	/* A non-CHECKSUM_PARTIAL SKB does not require setup. */
	if (skb->ip_summed != CHECKSUM_PARTIAL)
		return 0;

	if (skb->protocol != htons(ETH_P_IP))
		goto out;

	iph = (void *)skb->data;
	th = skb->data + 4 * iph->ihl;
	if (th >= skb_tail_pointer(skb))
		goto out;

	skb->csum_start = th - skb->head;
	switch (iph->protocol) {
	case IPPROTO_TCP:
		skb->csum_offset = offsetof(struct tcphdr, check);

		if (recalculate_partial_csum) {
			struct tcphdr *tcph = (struct tcphdr *)th;
			tcph->check = ~csum_tcpudp_magic(iph->saddr, iph->daddr,
							 skb->len - iph->ihl*4,
							 IPPROTO_TCP, 0);
		}
		break;
	case IPPROTO_UDP:
		skb->csum_offset = offsetof(struct udphdr, check);

		if (recalculate_partial_csum) {
			struct udphdr *udph = (struct udphdr *)th;
			udph->check = ~csum_tcpudp_magic(iph->saddr, iph->daddr,
							 skb->len - iph->ihl*4,
							 IPPROTO_UDP, 0);
		}
		break;
	default:
		if (net_ratelimit())
			printk(KERN_ERR "Attempting to checksum a non-"
			       "TCP/UDP packet, dropping a protocol"
			       " %d packet", iph->protocol);
		goto out;
	}

	if ((th + skb->csum_offset + 2) > skb_tail_pointer(skb))
		goto out;

	err = 0;

out:
	return err;
}

static int handle_incoming_queue(struct net_device *dev,
				 struct sk_buff_head *rxq)
{
	int packets_dropped = 0;
	struct sk_buff *skb;

	while ((skb = __skb_dequeue(rxq)) != NULL) {
		struct page *page = NETFRONT_SKB_CB(skb)->page;
		void *vaddr = page_address(page);
		unsigned offset = NETFRONT_SKB_CB(skb)->offset;

		memcpy(skb->data, vaddr + offset,
		       skb_headlen(skb));

		if (page != skb_shinfo(skb)->frags[0].page)
			__free_page(page);

		/* Ethernet work: Delayed to here as it peeks the header. */
		skb->protocol = eth_type_trans(skb, dev);

		if (checksum_setup(dev, skb)) {
			kfree_skb(skb);
			packets_dropped++;
			dev->stats.rx_errors++;
			continue;
		}

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

		/* Pass it up. */
		netif_receive_skb(skb);
	}

	return packets_dropped;
}

static int xennet_poll(struct napi_struct *napi, int budget)
{
	struct netfront_info *np = container_of(napi, struct netfront_info, napi);
	struct net_device *dev = np->netdev;
	struct sk_buff *skb;
	struct netfront_rx_info rinfo;
	struct xen_netif_rx_response *rx = &rinfo.rx;
	struct xen_netif_extra_info *extras = rinfo.extras;
	RING_IDX i, rp;
	int work_done;
	struct sk_buff_head rxq;
	struct sk_buff_head errq;
	struct sk_buff_head tmpq;
	unsigned long flags;
	unsigned int len;
	int err;

	spin_lock(&np->rx_lock);

	skb_queue_head_init(&rxq);
	skb_queue_head_init(&errq);
	skb_queue_head_init(&tmpq);

	rp = np->rx.sring->rsp_prod;
	rmb(); /* Ensure we see queued responses up to 'rp'. */

	i = np->rx.rsp_cons;
	work_done = 0;
	while ((i != rp) && (work_done < budget)) {
		memcpy(rx, RING_GET_RESPONSE(&np->rx, i), sizeof(*rx));
		memset(extras, 0, sizeof(rinfo.extras));

		err = xennet_get_responses(np, &rinfo, rp, &tmpq);

		if (unlikely(err)) {
err:
			while ((skb = __skb_dequeue(&tmpq)))
				__skb_queue_tail(&errq, skb);
			dev->stats.rx_errors++;
			i = np->rx.rsp_cons;
			continue;
		}

		skb = __skb_dequeue(&tmpq);

		if (extras[XEN_NETIF_EXTRA_TYPE_GSO - 1].type) {
			struct xen_netif_extra_info *gso;
			gso = &extras[XEN_NETIF_EXTRA_TYPE_GSO - 1];

			if (unlikely(xennet_set_skb_gso(skb, gso))) {
				__skb_queue_head(&tmpq, skb);
				np->rx.rsp_cons += skb_queue_len(&tmpq);
				goto err;
			}
		}

		NETFRONT_SKB_CB(skb)->page = skb_shinfo(skb)->frags[0].page;
		NETFRONT_SKB_CB(skb)->offset = rx->offset;

		len = rx->status;
		if (len > RX_COPY_THRESHOLD)
			len = RX_COPY_THRESHOLD;
		skb_put(skb, len);

		if (rx->status > len) {
			skb_shinfo(skb)->frags[0].page_offset =
				rx->offset + len;
			skb_shinfo(skb)->frags[0].size = rx->status - len;
			skb->data_len = rx->status - len;
		} else {
			skb_shinfo(skb)->frags[0].page = NULL;
			skb_shinfo(skb)->nr_frags = 0;
		}

		i = xennet_fill_frags(np, skb, &tmpq);

		/*
		 * Truesize approximates the size of true data plus
		 * any supervisor overheads. Adding hypervisor
		 * overheads has been shown to significantly reduce
		 * achievable bandwidth with the default receive
		 * buffer size. It is therefore not wise to account
		 * for it here.
		 *
		 * After alloc_skb(RX_COPY_THRESHOLD), truesize is set
		 * to RX_COPY_THRESHOLD + the supervisor
		 * overheads. Here, we add the size of the data pulled
		 * in xennet_fill_frags().
		 *
		 * We also adjust for any unused space in the main
		 * data area by subtracting (RX_COPY_THRESHOLD -
		 * len). This is especially important with drivers
		 * which split incoming packets into header and data,
		 * using only 66 bytes of the main data area (see the
		 * e1000 driver for example.)  On such systems,
		 * without this last adjustement, our achievable
		 * receive throughout using the standard receive
		 * buffer size was cut by 25%(!!!).
		 */
		skb->truesize += skb->data_len - (RX_COPY_THRESHOLD - len);
		skb->len += skb->data_len;

		if (rx->flags & XEN_NETRXF_csum_blank)
			skb->ip_summed = CHECKSUM_PARTIAL;
		else if (rx->flags & XEN_NETRXF_data_validated)
			skb->ip_summed = CHECKSUM_UNNECESSARY;

		__skb_queue_tail(&rxq, skb);

		np->rx.rsp_cons = ++i;
		work_done++;
	}

	__skb_queue_purge(&errq);

	work_done -= handle_incoming_queue(dev, &rxq);