[linux-block.git] / drivers / net / xen-netback / interface.c

/*
 * Network-device interface management.
 *
 * Copyright (c) 2004-2005, Keir Fraser
 *
 * 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 "common.h"

#include <linux/kthread.h>
#include <linux/ethtool.h>
#include <linux/rtnetlink.h>
#include <linux/if_vlan.h>

#include <xen/events.h>
#include <asm/xen/hypercall.h>

#define XENVIF_QUEUE_LENGTH 32
#define XENVIF_NAPI_WEIGHT  64

int xenvif_schedulable(struct xenvif *vif)
{
	return netif_running(vif->dev) && netif_carrier_ok(vif->dev);
}

static int xenvif_rx_schedulable(struct xenvif *vif)
{
	return xenvif_schedulable(vif) && !xenvif_rx_ring_full(vif);
}

static irqreturn_t xenvif_tx_interrupt(int irq, void *dev_id)
{
	struct xenvif *vif = dev_id;

	if (RING_HAS_UNCONSUMED_REQUESTS(&vif->tx))
		napi_schedule(&vif->napi);

	return IRQ_HANDLED;
}

static int xenvif_poll(struct napi_struct *napi, int budget)
{
	struct xenvif *vif = container_of(napi, struct xenvif, napi);
	int work_done;

	work_done = xenvif_tx_action(vif, budget);

	if (work_done < budget) {
		int more_to_do = 0;
		unsigned long flags;

		/* It is necessary to disable IRQ before calling
		 * RING_HAS_UNCONSUMED_REQUESTS. Otherwise we might
		 * lose event from the frontend.
		 *
		 * Consider:
		 *   RING_HAS_UNCONSUMED_REQUESTS
		 *   <frontend generates event to trigger napi_schedule>
		 *   __napi_complete
		 *
		 * This handler is still in scheduled state so the
		 * event has no effect at all. After __napi_complete
		 * this handler is descheduled and cannot get
		 * scheduled again. We lose event in this case and the ring
		 * will be completely stalled.
		 */

		local_irq_save(flags);

		RING_FINAL_CHECK_FOR_REQUESTS(&vif->tx, more_to_do);
		if (!more_to_do)
			__napi_complete(napi);

		local_irq_restore(flags);
	}

	return work_done;
}

static irqreturn_t xenvif_rx_interrupt(int irq, void *dev_id)
{
	struct xenvif *vif = dev_id;

	if (xenvif_rx_schedulable(vif))
		netif_wake_queue(vif->dev);

	return IRQ_HANDLED;
}

static irqreturn_t xenvif_interrupt(int irq, void *dev_id)
{
	xenvif_tx_interrupt(irq, dev_id);
	xenvif_rx_interrupt(irq, dev_id);

	return IRQ_HANDLED;
}

static int xenvif_start_xmit(struct sk_buff *skb, struct net_device *dev)
{
	struct xenvif *vif = netdev_priv(dev);

	BUG_ON(skb->dev != dev);

	/* Drop the packet if vif is not ready */
	if (vif->task == NULL)
		goto drop;

	/* Drop the packet if the target domain has no receive buffers. */
	if (!xenvif_rx_schedulable(vif))
		goto drop;

	/* Reserve ring slots for the worst-case number of fragments. */
	vif->rx_req_cons_peek += xenvif_count_skb_slots(vif, skb);

	if (vif->can_queue && xenvif_must_stop_queue(vif))
		netif_stop_queue(dev);

	xenvif_queue_tx_skb(vif, skb);

	return NETDEV_TX_OK;

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

void xenvif_notify_tx_completion(struct xenvif *vif)
{
	if (netif_queue_stopped(vif->dev) && xenvif_rx_schedulable(vif))
		netif_wake_queue(vif->dev);
}

static struct net_device_stats *xenvif_get_stats(struct net_device *dev)
{
	struct xenvif *vif = netdev_priv(dev);
	return &vif->dev->stats;
}

static void xenvif_up(struct xenvif *vif)
{
	napi_enable(&vif->napi);
	enable_irq(vif->tx_irq);
	if (vif->tx_irq != vif->rx_irq)
		enable_irq(vif->rx_irq);
	xenvif_check_rx_xenvif(vif);
}

static void xenvif_down(struct xenvif *vif)
{
	napi_disable(&vif->napi);
	disable_irq(vif->tx_irq);
	if (vif->tx_irq != vif->rx_irq)
		disable_irq(vif->rx_irq);
	del_timer_sync(&vif->credit_timeout);
}

static int xenvif_open(struct net_device *dev)
{
	struct xenvif *vif = netdev_priv(dev);
	if (netif_carrier_ok(dev))
		xenvif_up(vif);
	netif_start_queue(dev);
	return 0;
}

static int xenvif_close(struct net_device *dev)
{
	struct xenvif *vif = netdev_priv(dev);
	if (netif_carrier_ok(dev))
		xenvif_down(vif);
	netif_stop_queue(dev);
	return 0;
}

static int xenvif_change_mtu(struct net_device *dev, int mtu)
{
	struct xenvif *vif = netdev_priv(dev);
	int max = vif->can_sg ? 65535 - VLAN_ETH_HLEN : ETH_DATA_LEN;

	if (mtu > max)
		return -EINVAL;
	dev->mtu = mtu;
	return 0;
}

static netdev_features_t xenvif_fix_features(struct net_device *dev,
	netdev_features_t features)
{
	struct xenvif *vif = netdev_priv(dev);

	if (!vif->can_sg)
		features &= ~NETIF_F_SG;
	if (!vif->gso && !vif->gso_prefix)
		features &= ~NETIF_F_TSO;
	if (!vif->csum)
		features &= ~NETIF_F_IP_CSUM;

	return features;
}

static const struct xenvif_stat {
	char name[ETH_GSTRING_LEN];
	u16 offset;
} xenvif_stats[] = {
	{
		"rx_gso_checksum_fixup",
		offsetof(struct xenvif, rx_gso_checksum_fixup)
	},
};

static int xenvif_get_sset_count(struct net_device *dev, int string_set)
{
	switch (string_set) {
	case ETH_SS_STATS:
		return ARRAY_SIZE(xenvif_stats);
	default:
		return -EINVAL;
	}
}

static void xenvif_get_ethtool_stats(struct net_device *dev,
				     struct ethtool_stats *stats, u64 * data)
{
	void *vif = netdev_priv(dev);
	int i;

	for (i = 0; i < ARRAY_SIZE(xenvif_stats); i++)
		data[i] = *(unsigned long *)(vif + xenvif_stats[i].offset);
}

static void xenvif_get_strings(struct net_device *dev, u32 stringset, u8 * data)
{
	int i;

	switch (stringset) {
	case ETH_SS_STATS:
		for (i = 0; i < ARRAY_SIZE(xenvif_stats); i++)
			memcpy(data + i * ETH_GSTRING_LEN,
			       xenvif_stats[i].name, ETH_GSTRING_LEN);
		break;
	}
}

static const struct ethtool_ops xenvif_ethtool_ops = {
	.get_link	= ethtool_op_get_link,

	.get_sset_count = xenvif_get_sset_count,
	.get_ethtool_stats = xenvif_get_ethtool_stats,
	.get_strings = xenvif_get_strings,
};

static const struct net_device_ops xenvif_netdev_ops = {
	.ndo_start_xmit	= xenvif_start_xmit,
	.ndo_get_stats	= xenvif_get_stats,
	.ndo_open	= xenvif_open,
	.ndo_stop	= xenvif_close,
	.ndo_change_mtu	= xenvif_change_mtu,
	.ndo_fix_features = xenvif_fix_features,
	.ndo_set_mac_address = eth_mac_addr,
	.ndo_validate_addr   = eth_validate_addr,
};

struct xenvif *xenvif_alloc(struct device *parent, domid_t domid,
			    unsigned int handle)
{
	int err;
	struct net_device *dev;
	struct xenvif *vif;
	char name[IFNAMSIZ] = {};
	int i;

	snprintf(name, IFNAMSIZ - 1, "vif%u.%u", domid, handle);
	dev = alloc_netdev(sizeof(struct xenvif), name, ether_setup);
	if (dev == NULL) {
		pr_warn("Could not allocate netdev for %s\n", name);
		return ERR_PTR(-ENOMEM);
	}

	SET_NETDEV_DEV(dev, parent);

	vif = netdev_priv(dev);
	vif->domid  = domid;
	vif->handle = handle;
	vif->can_sg = 1;
	vif->csum = 1;
	vif->dev = dev;

	vif->credit_bytes = vif->remaining_credit = ~0UL;
	vif->credit_usec  = 0UL;
	init_timer(&vif->credit_timeout);
	/* Initialize 'expires' now: it's used to track the credit window. */
	vif->credit_timeout.expires = jiffies;

	dev->netdev_ops	= &xenvif_netdev_ops;
	dev->hw_features = NETIF_F_SG | NETIF_F_IP_CSUM | NETIF_F_TSO;
	dev->features = dev->hw_features;
	SET_ETHTOOL_OPS(dev, &xenvif_ethtool_ops);

	dev->tx_queue_len = XENVIF_QUEUE_LENGTH;

	skb_queue_head_init(&vif->rx_queue);
	skb_queue_head_init(&vif->tx_queue);

	vif->pending_cons = 0;
	vif->pending_prod = MAX_PENDING_REQS;
	for (i = 0; i < MAX_PENDING_REQS; i++)
		vif->pending_ring[i] = i;
	for (i = 0; i < MAX_PENDING_REQS; i++)
		vif->mmap_pages[i] = NULL;

	/*
	 * Initialise a dummy MAC address. We choose the numerically
	 * largest non-broadcast address to prevent the address getting
	 * stolen by an Ethernet bridge for STP purposes.
	 * (FE:FF:FF:FF:FF:FF)
	 */
	memset(dev->dev_addr, 0xFF, ETH_ALEN);
	dev->dev_addr[0] &= ~0x01;

	netif_napi_add(dev, &vif->napi, xenvif_poll, XENVIF_NAPI_WEIGHT);

	netif_carrier_off(dev);

	err = register_netdev(dev);
	if (err) {
		netdev_warn(dev, "Could not register device: err=%d\n", err);
		free_netdev(dev);
		return ERR_PTR(err);
	}

	netdev_dbg(dev, "Successfully created xenvif\n");
	return vif;
}

int xenvif_connect(struct xenvif *vif, unsigned long tx_ring_ref,
		   unsigned long rx_ring_ref, unsigned int tx_evtchn,
		   unsigned int rx_evtchn)
{
	int err = -ENOMEM;

	/* Already connected through? */
	if (vif->tx_irq)
		return 0;

	__module_get(THIS_MODULE);

	err = xenvif_map_frontend_rings(vif, tx_ring_ref, rx_ring_ref);
	if (err < 0)
		goto err;

	if (tx_evtchn == rx_evtchn) {
		/* feature-split-event-channels == 0 */
		err = bind_interdomain_evtchn_to_irqhandler(
			vif->domid, tx_evtchn, xenvif_interrupt, 0,
			vif->dev->name, vif);
		if (err < 0)
			goto err_unmap;
		vif->tx_irq = vif->rx_irq = err;
		disable_irq(vif->tx_irq);
	} else {
		/* feature-split-event-channels == 1 */
		snprintf(vif->tx_irq_name, sizeof(vif->tx_irq_name),
			 "%s-tx", vif->dev->name);
		err = bind_interdomain_evtchn_to_irqhandler(
			vif->domid, tx_evtchn, xenvif_tx_interrupt, 0,
			vif->tx_irq_name, vif);
		if (err < 0)
			goto err_unmap;
		vif->tx_irq = err;
		disable_irq(vif->tx_irq);

		snprintf(vif->rx_irq_name, sizeof(vif->rx_irq_name),
			 "%s-rx", vif->dev->name);
		err = bind_interdomain_evtchn_to_irqhandler(
			vif->domid, rx_evtchn, xenvif_rx_interrupt, 0,
			vif->rx_irq_name, vif);
		if (err < 0)
			goto err_tx_unbind;
		vif->rx_irq = err;
		disable_irq(vif->rx_irq);
	}

	init_waitqueue_head(&vif->wq);
	vif->task = kthread_create(xenvif_kthread,
				   (void *)vif, "%s", vif->dev->name);
	if (IS_ERR(vif->task)) {
		pr_warn("Could not allocate kthread for %s\n", vif->dev->name);
		err = PTR_ERR(vif->task);
		goto err_rx_unbind;
	}

	rtnl_lock();
	if (!vif->can_sg && vif->dev->mtu > ETH_DATA_LEN)
		dev_set_mtu(vif->dev, ETH_DATA_LEN);
	netdev_update_features(vif->dev);
	netif_carrier_on(vif->dev);
	if (netif_running(vif->dev))
		xenvif_up(vif);
	rtnl_unlock();

	wake_up_process(vif->task);

	return 0;

err_rx_unbind:
	unbind_from_irqhandler(vif->rx_irq, vif);
	vif->rx_irq = 0;
err_tx_unbind:
	unbind_from_irqhandler(vif->tx_irq, vif);
	vif->tx_irq = 0;
err_unmap:
	xenvif_unmap_frontend_rings(vif);
err:
	module_put(THIS_MODULE);
	return err;
}

void xenvif_carrier_off(struct xenvif *vif)
{
	struct net_device *dev = vif->dev;

	rtnl_lock();
	netif_carrier_off(dev); /* discard queued packets */
	if (netif_running(dev))
		xenvif_down(vif);
	rtnl_unlock();
}

void xenvif_disconnect(struct xenvif *vif)
{
	/* Disconnect funtion might get called by generic framework
	 * even before vif connects, so we need to check if we really
	 * need to do a module_put.
	 */
	int need_module_put = 0;

	if (netif_carrier_ok(vif->dev))
		xenvif_carrier_off(vif);

	if (vif->tx_irq) {
		if (vif->tx_irq == vif->rx_irq)
			unbind_from_irqhandler(vif->tx_irq, vif);
		else {
			unbind_from_irqhandler(vif->tx_irq, vif);
			unbind_from_irqhandler(vif->rx_irq, vif);
		}
		/* vif->irq is valid, we had a module_get in
		 * xenvif_connect.
		 */
		need_module_put = 1;
	}

	if (vif->task)
		kthread_stop(vif->task);

	netif_napi_del(&vif->napi);

	unregister_netdev(vif->dev);

	xenvif_unmap_frontend_rings(vif);

	free_netdev(vif->dev);

	if (need_module_put)
		module_put(THIS_MODULE);
}
Commit	Line	Data
f942dc25 IC	1	/*
	2	* Network-device interface management.
	3	*
	4	* Copyright (c) 2004-2005, Keir Fraser
	5	*
	6	* This program is free software; you can redistribute it and/or
	7	* modify it under the terms of the GNU General Public License version 2
	8	* as published by the Free Software Foundation; or, when distributed
	9	* separately from the Linux kernel or incorporated into other
	10	* software packages, subject to the following license:
	11	*
	12	* Permission is hereby granted, free of charge, to any person obtaining a copy
	13	* of this source file (the "Software"), to deal in the Software without
	14	* restriction, including without limitation the rights to use, copy, modify,
	15	* merge, publish, distribute, sublicense, and/or sell copies of the Software,
	16	* and to permit persons to whom the Software is furnished to do so, subject to
	17	* the following conditions:
	18	*
	19	* The above copyright notice and this permission notice shall be included in
	20	* all copies or substantial portions of the Software.
	21	*
	22	* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
	23	* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
	24	* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
	25	* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
	26	* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
	27	* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
	28	* IN THE SOFTWARE.
	29	*/
	30
	31	#include "common.h"
	32
b3f980bd	33	#include <linux/kthread.h>
f942dc25 IC	34	#include <linux/ethtool.h>
	35	#include <linux/rtnetlink.h>
	36	#include <linux/if_vlan.h>
	37
	38	#include <xen/events.h>
	39	#include <asm/xen/hypercall.h>
	40
	41	#define XENVIF_QUEUE_LENGTH 32
b3f980bd	42	#define XENVIF_NAPI_WEIGHT 64
f942dc25 IC	43
	44	int xenvif_schedulable(struct xenvif *vif)
	45	{
	46	return netif_running(vif->dev) && netif_carrier_ok(vif->dev);
	47	}
	48
	49	static int xenvif_rx_schedulable(struct xenvif *vif)
	50	{
7376419a	51	return xenvif_schedulable(vif) && !xenvif_rx_ring_full(vif);
f942dc25 IC	52	}
f942dc25 IC	53
e1f00a69	54	static irqreturn_t xenvif_tx_interrupt(int irq, void *dev_id)
f942dc25 IC	55	{
	56	struct xenvif *vif = dev_id;
	57
b3f980bd WL	58	if (RING_HAS_UNCONSUMED_REQUESTS(&vif->tx))
b3f980bd WL	59	napi_schedule(&vif->napi);
f942dc25	60
e1f00a69 WL	61	return IRQ_HANDLED;
	62	}
	63
b3f980bd WL	64	static int xenvif_poll(struct napi_struct *napi, int budget)
	65	{
	66	struct xenvif *vif = container_of(napi, struct xenvif, napi);
	67	int work_done;
	68
7376419a	69	work_done = xenvif_tx_action(vif, budget);
b3f980bd WL	70
	71	if (work_done < budget) {
	72	int more_to_do = 0;
	73	unsigned long flags;
	74
	75	/* It is necessary to disable IRQ before calling
	76	* RING_HAS_UNCONSUMED_REQUESTS. Otherwise we might
	77	* lose event from the frontend.
	78	*
	79	* Consider:
	80	* RING_HAS_UNCONSUMED_REQUESTS
	81	* <frontend generates event to trigger napi_schedule>
	82	* __napi_complete
	83	*
	84	* This handler is still in scheduled state so the
	85	* event has no effect at all. After __napi_complete
	86	* this handler is descheduled and cannot get
	87	* scheduled again. We lose event in this case and the ring
	88	* will be completely stalled.
	89	*/
	90
	91	local_irq_save(flags);
	92
	93	RING_FINAL_CHECK_FOR_REQUESTS(&vif->tx, more_to_do);
	94	if (!more_to_do)
	95	__napi_complete(napi);
	96
	97	local_irq_restore(flags);
	98	}
	99
	100	return work_done;
	101	}
	102
e1f00a69 WL	103	static irqreturn_t xenvif_rx_interrupt(int irq, void *dev_id)
	104	{
	105	struct xenvif *vif = dev_id;
	106
f942dc25 IC	107	if (xenvif_rx_schedulable(vif))
	108	netif_wake_queue(vif->dev);
	109
	110	return IRQ_HANDLED;
	111	}
	112
e1f00a69 WL	113	static irqreturn_t xenvif_interrupt(int irq, void *dev_id)
	114	{
	115	xenvif_tx_interrupt(irq, dev_id);
	116	xenvif_rx_interrupt(irq, dev_id);
	117
	118	return IRQ_HANDLED;
	119	}
	120
f942dc25 IC	121	static int xenvif_start_xmit(struct sk_buff skb, struct net_device dev)
	122	{
	123	struct xenvif *vif = netdev_priv(dev);
	124
	125	BUG_ON(skb->dev != dev);
	126
b3f980bd WL	127	/* Drop the packet if vif is not ready */
b3f980bd WL	128	if (vif->task == NULL)
f942dc25 IC	129	goto drop;
	130
	131	/* Drop the packet if the target domain has no receive buffers. */
	132	if (!xenvif_rx_schedulable(vif))
	133	goto drop;
	134
	135	/* Reserve ring slots for the worst-case number of fragments. */
7376419a	136	vif->rx_req_cons_peek += xenvif_count_skb_slots(vif, skb);
f942dc25	137
7376419a	138	if (vif->can_queue && xenvif_must_stop_queue(vif))
f942dc25 IC	139	netif_stop_queue(dev);
f942dc25 IC	140
7376419a	141	xenvif_queue_tx_skb(vif, skb);
f942dc25 IC	142
	143	return NETDEV_TX_OK;
	144
	145	drop:
	146	vif->dev->stats.tx_dropped++;
	147	dev_kfree_skb(skb);
	148	return NETDEV_TX_OK;
	149	}
	150
f942dc25 IC	151	void xenvif_notify_tx_completion(struct xenvif *vif)
	152	{
	153	if (netif_queue_stopped(vif->dev) && xenvif_rx_schedulable(vif))
	154	netif_wake_queue(vif->dev);
	155	}
	156
	157	static struct net_device_stats xenvif_get_stats(struct net_device dev)
	158	{
	159	struct xenvif *vif = netdev_priv(dev);
	160	return &vif->dev->stats;
	161	}
	162
	163	static void xenvif_up(struct xenvif *vif)
	164	{
b3f980bd	165	napi_enable(&vif->napi);
e1f00a69 WL	166	enable_irq(vif->tx_irq);
	167	if (vif->tx_irq != vif->rx_irq)
	168	enable_irq(vif->rx_irq);
7376419a	169	xenvif_check_rx_xenvif(vif);
f942dc25 IC	170	}
	171
	172	static void xenvif_down(struct xenvif *vif)
	173	{
b3f980bd	174	napi_disable(&vif->napi);
e1f00a69 WL	175	disable_irq(vif->tx_irq);
	176	if (vif->tx_irq != vif->rx_irq)
	177	disable_irq(vif->rx_irq);
3e55f8b3	178	del_timer_sync(&vif->credit_timeout);
f942dc25 IC	179	}
	180
	181	static int xenvif_open(struct net_device *dev)
	182	{
	183	struct xenvif *vif = netdev_priv(dev);
	184	if (netif_carrier_ok(dev))
	185	xenvif_up(vif);
	186	netif_start_queue(dev);
	187	return 0;
	188	}
	189
	190	static int xenvif_close(struct net_device *dev)
	191	{
	192	struct xenvif *vif = netdev_priv(dev);
	193	if (netif_carrier_ok(dev))
	194	xenvif_down(vif);
	195	netif_stop_queue(dev);
	196	return 0;
	197	}
	198
	199	static int xenvif_change_mtu(struct net_device *dev, int mtu)
	200	{
	201	struct xenvif *vif = netdev_priv(dev);
	202	int max = vif->can_sg ? 65535 - VLAN_ETH_HLEN : ETH_DATA_LEN;
	203
	204	if (mtu > max)
	205	return -EINVAL;
	206	dev->mtu = mtu;
	207	return 0;
	208	}
	209
c8f44aff MM	210	static netdev_features_t xenvif_fix_features(struct net_device *dev,
c8f44aff MM	211	netdev_features_t features)
f942dc25 IC	212	{
f942dc25 IC	213	struct xenvif *vif = netdev_priv(dev);
f942dc25	214
47103041 MM	215	if (!vif->can_sg)
	216	features &= ~NETIF_F_SG;
	217	if (!vif->gso && !vif->gso_prefix)
	218	features &= ~NETIF_F_TSO;
	219	if (!vif->csum)
	220	features &= ~NETIF_F_IP_CSUM;
f942dc25	221
47103041	222	return features;
f942dc25 IC	223	}
	224
	225	static const struct xenvif_stat {
	226	char name[ETH_GSTRING_LEN];
	227	u16 offset;
	228	} xenvif_stats[] = {
	229	{
	230	"rx_gso_checksum_fixup",
	231	offsetof(struct xenvif, rx_gso_checksum_fixup)
	232	},
	233	};
	234
	235	static int xenvif_get_sset_count(struct net_device *dev, int string_set)
	236	{
	237	switch (string_set) {
	238	case ETH_SS_STATS:
	239	return ARRAY_SIZE(xenvif_stats);
	240	default:
	241	return -EINVAL;
	242	}
	243	}
	244
	245	static void xenvif_get_ethtool_stats(struct net_device *dev,
	246	struct ethtool_stats stats, u64 data)
	247	{
	248	void *vif = netdev_priv(dev);
	249	int i;
	250
	251	for (i = 0; i < ARRAY_SIZE(xenvif_stats); i++)
	252	data[i] = (unsigned long )(vif + xenvif_stats[i].offset);
	253	}
	254
	255	static void xenvif_get_strings(struct net_device dev, u32 stringset, u8 data)
	256	{
	257	int i;
	258
	259	switch (stringset) {
	260	case ETH_SS_STATS:
	261	for (i = 0; i < ARRAY_SIZE(xenvif_stats); i++)
	262	memcpy(data + i * ETH_GSTRING_LEN,
	263	xenvif_stats[i].name, ETH_GSTRING_LEN);
	264	break;
	265	}
	266	}
	267
813abbba	268	static const struct ethtool_ops xenvif_ethtool_ops = {
f942dc25 IC	269	.get_link = ethtool_op_get_link,
	270
	271	.get_sset_count = xenvif_get_sset_count,
	272	.get_ethtool_stats = xenvif_get_ethtool_stats,
	273	.get_strings = xenvif_get_strings,
	274	};
	275
813abbba	276	static const struct net_device_ops xenvif_netdev_ops = {
f942dc25 IC	277	.ndo_start_xmit = xenvif_start_xmit,
	278	.ndo_get_stats = xenvif_get_stats,
	279	.ndo_open = xenvif_open,
	280	.ndo_stop = xenvif_close,
	281	.ndo_change_mtu = xenvif_change_mtu,
47103041	282	.ndo_fix_features = xenvif_fix_features,
4a633a60 MW	283	.ndo_set_mac_address = eth_mac_addr,
4a633a60 MW	284	.ndo_validate_addr = eth_validate_addr,
f942dc25 IC	285	};
	286
	287	struct xenvif xenvif_alloc(struct device parent, domid_t domid,
	288	unsigned int handle)
	289	{
	290	int err;
	291	struct net_device *dev;
	292	struct xenvif *vif;
	293	char name[IFNAMSIZ] = {};
b3f980bd	294	int i;
f942dc25 IC	295
	296	snprintf(name, IFNAMSIZ - 1, "vif%u.%u", domid, handle);
	297	dev = alloc_netdev(sizeof(struct xenvif), name, ether_setup);
	298	if (dev == NULL) {
b3f980bd	299	pr_warn("Could not allocate netdev for %s\n", name);
f942dc25 IC	300	return ERR_PTR(-ENOMEM);
	301	}
	302
	303	SET_NETDEV_DEV(dev, parent);
	304
	305	vif = netdev_priv(dev);
	306	vif->domid = domid;
	307	vif->handle = handle;
f942dc25 IC	308	vif->can_sg = 1;
f942dc25 IC	309	vif->csum = 1;
f942dc25	310	vif->dev = dev;
f942dc25 IC	311
	312	vif->credit_bytes = vif->remaining_credit = ~0UL;
	313	vif->credit_usec = 0UL;
	314	init_timer(&vif->credit_timeout);
	315	/* Initialize 'expires' now: it's used to track the credit window. */
	316	vif->credit_timeout.expires = jiffies;
	317
	318	dev->netdev_ops = &xenvif_netdev_ops;
47103041 MM	319	dev->hw_features = NETIF_F_SG \| NETIF_F_IP_CSUM \| NETIF_F_TSO;
47103041 MM	320	dev->features = dev->hw_features;
f942dc25 IC	321	SET_ETHTOOL_OPS(dev, &xenvif_ethtool_ops);
	322
	323	dev->tx_queue_len = XENVIF_QUEUE_LENGTH;
	324
b3f980bd WL	325	skb_queue_head_init(&vif->rx_queue);
	326	skb_queue_head_init(&vif->tx_queue);
	327
	328	vif->pending_cons = 0;
	329	vif->pending_prod = MAX_PENDING_REQS;
	330	for (i = 0; i < MAX_PENDING_REQS; i++)
	331	vif->pending_ring[i] = i;
	332	for (i = 0; i < MAX_PENDING_REQS; i++)
	333	vif->mmap_pages[i] = NULL;
	334
f942dc25 IC	335	/*
	336	* Initialise a dummy MAC address. We choose the numerically
	337	* largest non-broadcast address to prevent the address getting
	338	* stolen by an Ethernet bridge for STP purposes.
	339	* (FE:FF:FF:FF:FF:FF)
	340	*/
	341	memset(dev->dev_addr, 0xFF, ETH_ALEN);
	342	dev->dev_addr[0] &= ~0x01;
	343
b3f980bd WL	344	netif_napi_add(dev, &vif->napi, xenvif_poll, XENVIF_NAPI_WEIGHT);
b3f980bd WL	345
f942dc25 IC	346	netif_carrier_off(dev);
	347
	348	err = register_netdev(dev);
	349	if (err) {
	350	netdev_warn(dev, "Could not register device: err=%d\n", err);
	351	free_netdev(dev);
	352	return ERR_PTR(err);
	353	}
	354
	355	netdev_dbg(dev, "Successfully created xenvif\n");
	356	return vif;
	357	}
	358
	359	int xenvif_connect(struct xenvif *vif, unsigned long tx_ring_ref,
e1f00a69 WL	360	unsigned long rx_ring_ref, unsigned int tx_evtchn,
e1f00a69 WL	361	unsigned int rx_evtchn)
f942dc25 IC	362	{
	363	int err = -ENOMEM;
	364
	365	/* Already connected through? */
e1f00a69	366	if (vif->tx_irq)
f942dc25 IC	367	return 0;
f942dc25 IC	368
b103f358 WL	369	__module_get(THIS_MODULE);
b103f358 WL	370
7376419a	371	err = xenvif_map_frontend_rings(vif, tx_ring_ref, rx_ring_ref);
f942dc25 IC	372	if (err < 0)
	373	goto err;
	374
e1f00a69 WL	375	if (tx_evtchn == rx_evtchn) {
	376	/* feature-split-event-channels == 0 */
	377	err = bind_interdomain_evtchn_to_irqhandler(
	378	vif->domid, tx_evtchn, xenvif_interrupt, 0,
	379	vif->dev->name, vif);
	380	if (err < 0)
	381	goto err_unmap;
	382	vif->tx_irq = vif->rx_irq = err;
	383	disable_irq(vif->tx_irq);
	384	} else {
	385	/* feature-split-event-channels == 1 */
	386	snprintf(vif->tx_irq_name, sizeof(vif->tx_irq_name),
	387	"%s-tx", vif->dev->name);
	388	err = bind_interdomain_evtchn_to_irqhandler(
	389	vif->domid, tx_evtchn, xenvif_tx_interrupt, 0,
	390	vif->tx_irq_name, vif);
	391	if (err < 0)
	392	goto err_unmap;
	393	vif->tx_irq = err;
	394	disable_irq(vif->tx_irq);
	395
	396	snprintf(vif->rx_irq_name, sizeof(vif->rx_irq_name),
	397	"%s-rx", vif->dev->name);
	398	err = bind_interdomain_evtchn_to_irqhandler(
	399	vif->domid, rx_evtchn, xenvif_rx_interrupt, 0,
	400	vif->rx_irq_name, vif);
	401	if (err < 0)
	402	goto err_tx_unbind;
	403	vif->rx_irq = err;
	404	disable_irq(vif->rx_irq);
	405	}
f942dc25	406
b3f980bd	407	init_waitqueue_head(&vif->wq);
7376419a	408	vif->task = kthread_create(xenvif_kthread,
a9677bc0	409	(void *)vif, "%s", vif->dev->name);
b3f980bd WL	410	if (IS_ERR(vif->task)) {
	411	pr_warn("Could not allocate kthread for %s\n", vif->dev->name);
	412	err = PTR_ERR(vif->task);
	413	goto err_rx_unbind;
	414	}
f942dc25 IC	415
f942dc25 IC	416	rtnl_lock();
47103041 MM	417	if (!vif->can_sg && vif->dev->mtu > ETH_DATA_LEN)
	418	dev_set_mtu(vif->dev, ETH_DATA_LEN);
	419	netdev_update_features(vif->dev);
	420	netif_carrier_on(vif->dev);
d0e5d832 DV	421	if (netif_running(vif->dev))
d0e5d832 DV	422	xenvif_up(vif);
f942dc25 IC	423	rtnl_unlock();
f942dc25 IC	424
b3f980bd WL	425	wake_up_process(vif->task);
b3f980bd WL	426
f942dc25	427	return 0;
b3f980bd WL	428
	429	err_rx_unbind:
	430	unbind_from_irqhandler(vif->rx_irq, vif);
	431	vif->rx_irq = 0;
e1f00a69 WL	432	err_tx_unbind:
	433	unbind_from_irqhandler(vif->tx_irq, vif);
	434	vif->tx_irq = 0;
f942dc25	435	err_unmap:
7376419a	436	xenvif_unmap_frontend_rings(vif);
f942dc25	437	err:
b103f358	438	module_put(THIS_MODULE);
f942dc25 IC	439	return err;
	440	}
	441
48856286	442	void xenvif_carrier_off(struct xenvif *vif)
f942dc25 IC	443	{
f942dc25 IC	444	struct net_device *dev = vif->dev;
48856286 IC	445
	446	rtnl_lock();
	447	netif_carrier_off(dev); /* discard queued packets */
	448	if (netif_running(dev))
	449	xenvif_down(vif);
	450	rtnl_unlock();
48856286 IC	451	}
	452
	453	void xenvif_disconnect(struct xenvif *vif)
	454	{
b103f358 WL	455	/* Disconnect funtion might get called by generic framework
	456	* even before vif connects, so we need to check if we really
	457	* need to do a module_put.
	458	*/
	459	int need_module_put = 0;
	460
48856286 IC	461	if (netif_carrier_ok(vif->dev))
48856286 IC	462	xenvif_carrier_off(vif);
f942dc25	463
e1f00a69 WL	464	if (vif->tx_irq) {
	465	if (vif->tx_irq == vif->rx_irq)
	466	unbind_from_irqhandler(vif->tx_irq, vif);
	467	else {
	468	unbind_from_irqhandler(vif->tx_irq, vif);
	469	unbind_from_irqhandler(vif->rx_irq, vif);
	470	}
b103f358 WL	471	/* vif->irq is valid, we had a module_get in
	472	* xenvif_connect.
	473	*/
	474	need_module_put = 1;
	475	}
f942dc25	476
b3f980bd WL	477	if (vif->task)
	478	kthread_stop(vif->task);
	479
	480	netif_napi_del(&vif->napi);
	481
f942dc25 IC	482	unregister_netdev(vif->dev);
f942dc25 IC	483
7376419a	484	xenvif_unmap_frontend_rings(vif);
f942dc25 IC	485
f942dc25 IC	486	free_netdev(vif->dev);
b103f358 WL	487
	488	if (need_module_put)
	489	module_put(THIS_MODULE);
f942dc25	490	}