[linux-2.6-block.git] / drivers / net / loopback.c

/*
 * INET		An implementation of the TCP/IP protocol suite for the LINUX
 *		operating system.  INET is implemented using the  BSD Socket
 *		interface as the means of communication with the user level.
 *
 *		Pseudo-driver for the loopback interface.
 *
 * Version:	@(#)loopback.c	1.0.4b	08/16/93
 *
 * Authors:	Ross Biro
 *		Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *		Donald Becker, <becker@scyld.com>
 *
 *		Alan Cox	:	Fixed oddments for NET3.014
 *		Alan Cox	:	Rejig for NET3.029 snap #3
 *		Alan Cox	: 	Fixed NET3.029 bugs and sped up
 *		Larry McVoy	:	Tiny tweak to double performance
 *		Alan Cox	:	Backed out LMV's tweak - the linux mm
 *					can't take it...
 *              Michael Griffith:       Don't bother computing the checksums
 *                                      on packets received on the loopback
 *                                      interface.
 *		Alexey Kuznetsov:	Potential hang under some extreme
 *					cases removed.
 *
 *		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.
 */
#include <linux/kernel.h>
#include <linux/jiffies.h>
#include <linux/module.h>
#include <linux/interrupt.h>
#include <linux/fs.h>
#include <linux/types.h>
#include <linux/string.h>
#include <linux/socket.h>
#include <linux/errno.h>
#include <linux/fcntl.h>
#include <linux/in.h>
#include <linux/init.h>

#include <asm/system.h>
#include <asm/uaccess.h>
#include <asm/io.h>

#include <linux/inet.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/skbuff.h>
#include <linux/ethtool.h>
#include <net/sock.h>
#include <net/checksum.h>
#include <linux/if_ether.h>	/* For the statistics structure. */
#include <linux/if_arp.h>	/* For ARPHRD_ETHER */
#include <linux/ip.h>
#include <linux/tcp.h>
#include <linux/percpu.h>
#include <net/net_namespace.h>

struct pcpu_lstats {
	unsigned long packets;
	unsigned long bytes;
};

/* KISS: just allocate small chunks and copy bits.
 *
 * So, in fact, this is documentation, explaining what we expect
 * of largesending device modulo TCP checksum, which is ignored for loopback.
 */

#ifdef LOOPBACK_TSO
static void emulate_large_send_offload(struct sk_buff *skb)
{
	struct iphdr *iph = ip_hdr(skb);
	struct tcphdr *th = (struct tcphdr *)(skb_network_header(skb) +
					      (iph->ihl * 4));
	unsigned int doffset = (iph->ihl + th->doff) * 4;
	unsigned int mtu = skb_shinfo(skb)->gso_size + doffset;
	unsigned int offset = 0;
	u32 seq = ntohl(th->seq);
	u16 id  = ntohs(iph->id);

	while (offset + doffset < skb->len) {
		unsigned int frag_size = min(mtu, skb->len - offset) - doffset;
		struct sk_buff *nskb = alloc_skb(mtu + 32, GFP_ATOMIC);

		if (!nskb)
			break;
		skb_reserve(nskb, 32);
		skb_set_mac_header(nskb, -ETH_HLEN);
		skb_reset_network_header(nskb);
		iph = ip_hdr(nskb);
		skb_copy_to_linear_data(nskb, skb_network_header(skb),
					doffset);
		if (skb_copy_bits(skb,
				  doffset + offset,
				  nskb->data + doffset,
				  frag_size))
			BUG();
		skb_put(nskb, doffset + frag_size);
		nskb->ip_summed = CHECKSUM_UNNECESSARY;
		nskb->dev = skb->dev;
		nskb->priority = skb->priority;
		nskb->protocol = skb->protocol;
		nskb->dst = dst_clone(skb->dst);
		memcpy(nskb->cb, skb->cb, sizeof(skb->cb));
		nskb->pkt_type = skb->pkt_type;

		th = (struct tcphdr *)(skb_network_header(nskb) + iph->ihl * 4);
		iph->tot_len = htons(frag_size + doffset);
		iph->id = htons(id);
		iph->check = 0;
		iph->check = ip_fast_csum((unsigned char *) iph, iph->ihl);
		th->seq = htonl(seq);
		if (offset + doffset + frag_size < skb->len)
			th->fin = th->psh = 0;
		netif_rx(nskb);
		offset += frag_size;
		seq += frag_size;
		id++;
	}

	dev_kfree_skb(skb);
}
#endif /* LOOPBACK_TSO */

/*
 * The higher levels take care of making this non-reentrant (it's
 * called with bh's disabled).
 */
static int loopback_xmit(struct sk_buff *skb, struct net_device *dev)
{
	struct pcpu_lstats *pcpu_lstats, *lb_stats;

	skb_orphan(skb);

	skb->protocol = eth_type_trans(skb,dev);
#ifndef LOOPBACK_MUST_CHECKSUM
	skb->ip_summed = CHECKSUM_UNNECESSARY;
#endif

#ifdef LOOPBACK_TSO
	if (skb_is_gso(skb)) {
		BUG_ON(skb->protocol != htons(ETH_P_IP));
		BUG_ON(ip_hdr(skb)->protocol != IPPROTO_TCP);

		emulate_large_send_offload(skb);
		return 0;
	}
#endif
	dev->last_rx = jiffies;

	/* it's OK to use per_cpu_ptr() because BHs are off */
	pcpu_lstats = netdev_priv(dev);
	lb_stats = per_cpu_ptr(pcpu_lstats, smp_processor_id());
	lb_stats->bytes += skb->len;
	lb_stats->packets++;

	netif_rx(skb);

	return 0;
}

static struct net_device_stats *get_stats(struct net_device *dev)
{
	const struct pcpu_lstats *pcpu_lstats;
	struct net_device_stats *stats = &dev->stats;
	unsigned long bytes = 0;
	unsigned long packets = 0;
	int i;

	pcpu_lstats = netdev_priv(dev);
	for_each_possible_cpu(i) {
		const struct pcpu_lstats *lb_stats;

		lb_stats = per_cpu_ptr(pcpu_lstats, i);
		bytes   += lb_stats->bytes;
		packets += lb_stats->packets;
	}
	stats->rx_packets = packets;
	stats->tx_packets = packets;
	stats->rx_bytes = bytes;
	stats->tx_bytes = bytes;
	return stats;
}

static u32 always_on(struct net_device *dev)
{
	return 1;
}

static const struct ethtool_ops loopback_ethtool_ops = {
	.get_link		= always_on,
	.set_tso		= ethtool_op_set_tso,
	.get_tx_csum		= always_on,
	.get_sg			= always_on,
	.get_rx_csum		= always_on,
};

static int loopback_dev_init(struct net_device *dev)
{
	struct pcpu_lstats *lstats;

	lstats = alloc_percpu(struct pcpu_lstats);
	if (!lstats)
		return -ENOMEM;

	dev->priv = lstats;
	return 0;
}

static void loopback_dev_free(struct net_device *dev)
{
	struct pcpu_lstats *lstats = netdev_priv(dev);

	free_percpu(lstats);
	free_netdev(dev);
}

/*
 * The loopback device is special. There is only one instance
 * per network namespace.
 */
static void loopback_setup(struct net_device *dev)
{
	dev->get_stats		= &get_stats;
	dev->mtu		= (16 * 1024) + 20 + 20 + 12;
	dev->hard_start_xmit	= loopback_xmit;
	dev->hard_header_len	= ETH_HLEN;	/* 14	*/
	dev->addr_len		= ETH_ALEN;	/* 6	*/
	dev->tx_queue_len	= 0;
	dev->type		= ARPHRD_LOOPBACK;	/* 0x0001*/
	dev->flags		= IFF_LOOPBACK;
	dev->features 		= NETIF_F_SG | NETIF_F_FRAGLIST
#ifdef LOOPBACK_TSO
		| NETIF_F_TSO
#endif
		| NETIF_F_NO_CSUM
		| NETIF_F_HIGHDMA
		| NETIF_F_LLTX
		| NETIF_F_NETNS_LOCAL;
	dev->ethtool_ops	= &loopback_ethtool_ops;
	dev->header_ops		= &eth_header_ops;
	dev->init = loopback_dev_init;
	dev->destructor = loopback_dev_free;
}

/* Setup and register the loopback device. */
static __net_init int loopback_net_init(struct net *net)
{
	struct net_device *dev;
	int err;

	err = -ENOMEM;
	dev = alloc_netdev(0, "lo", loopback_setup);
	if (!dev)
		goto out;

	dev_net_set(dev, net);
	err = register_netdev(dev);
	if (err)
		goto out_free_netdev;

	net->loopback_dev = dev;
	return 0;


out_free_netdev:
	free_netdev(dev);
out:
	if (net == &init_net)
		panic("loopback: Failed to register netdevice: %d\n", err);
	return err;
}

static __net_exit void loopback_net_exit(struct net *net)
{
	struct net_device *dev = net->loopback_dev;

	unregister_netdev(dev);
}

static struct pernet_operations __net_initdata loopback_net_ops = {
       .init = loopback_net_init,
       .exit = loopback_net_exit,
};

static int __init loopback_init(void)
{
	return register_pernet_device(&loopback_net_ops);
}

/* Loopback is special. It should be initialized before any other network
 * device and network subsystem.
 */
fs_initcall(loopback_init);
Commit	Line	Data
1da177e4 LT	1	/*
	2	* INET An implementation of the TCP/IP protocol suite for the LINUX
	3	* operating system. INET is implemented using the BSD Socket
	4	* interface as the means of communication with the user level.
	5	*
	6	* Pseudo-driver for the loopback interface.
	7	*
	8	* Version: @(#)loopback.c 1.0.4b 08/16/93
	9	*
02c30a84	10	* Authors: Ross Biro
1da177e4 LT	11	* Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
	12	* Donald Becker, <becker@scyld.com>
	13	*
	14	* Alan Cox : Fixed oddments for NET3.014
	15	* Alan Cox : Rejig for NET3.029 snap #3
	16	* Alan Cox : Fixed NET3.029 bugs and sped up
	17	* Larry McVoy : Tiny tweak to double performance
	18	* Alan Cox : Backed out LMV's tweak - the linux mm
	19	* can't take it...
	20	* Michael Griffith: Don't bother computing the checksums
	21	* on packets received on the loopback
	22	* interface.
	23	* Alexey Kuznetsov: Potential hang under some extreme
	24	* cases removed.
	25	*
	26	* This program is free software; you can redistribute it and/or
	27	* modify it under the terms of the GNU General Public License
	28	* as published by the Free Software Foundation; either version
	29	* 2 of the License, or (at your option) any later version.
	30	*/
	31	#include <linux/kernel.h>
	32	#include <linux/jiffies.h>
	33	#include <linux/module.h>
	34	#include <linux/interrupt.h>
	35	#include <linux/fs.h>
	36	#include <linux/types.h>
	37	#include <linux/string.h>
	38	#include <linux/socket.h>
	39	#include <linux/errno.h>
	40	#include <linux/fcntl.h>
	41	#include <linux/in.h>
	42	#include <linux/init.h>
	43
	44	#include <asm/system.h>
	45	#include <asm/uaccess.h>
	46	#include <asm/io.h>
	47
	48	#include <linux/inet.h>
	49	#include <linux/netdevice.h>
	50	#include <linux/etherdevice.h>
	51	#include <linux/skbuff.h>
	52	#include <linux/ethtool.h>
	53	#include <net/sock.h>
	54	#include <net/checksum.h>
	55	#include <linux/if_ether.h> /* For the statistics structure. */
	56	#include <linux/if_arp.h> /* For ARPHRD_ETHER */
	57	#include <linux/ip.h>
	58	#include <linux/tcp.h>
	59	#include <linux/percpu.h>
2774c7ab	60	#include <net/net_namespace.h>
1da177e4	61
5175c378 ED	62	struct pcpu_lstats {
	63	unsigned long packets;
	64	unsigned long bytes;
	65	};
1da177e4	66
1da177e4 LT	67	/* KISS: just allocate small chunks and copy bits.
	68	*
	69	* So, in fact, this is documentation, explaining what we expect
	70	* of largesending device modulo TCP checksum, which is ignored for loopback.
	71	*/
	72
d2ae1d2f	73	#ifdef LOOPBACK_TSO
1da177e4 LT	74	static void emulate_large_send_offload(struct sk_buff *skb)
1da177e4 LT	75	{
eddc9ec5	76	struct iphdr *iph = ip_hdr(skb);
d56f90a7 ACM	77	struct tcphdr th = (struct tcphdr )(skb_network_header(skb) +
d56f90a7 ACM	78	(iph->ihl * 4));
1da177e4	79	unsigned int doffset = (iph->ihl + th->doff) * 4;
7967168c	80	unsigned int mtu = skb_shinfo(skb)->gso_size + doffset;
1da177e4 LT	81	unsigned int offset = 0;
	82	u32 seq = ntohl(th->seq);
	83	u16 id = ntohs(iph->id);
	84
	85	while (offset + doffset < skb->len) {
	86	unsigned int frag_size = min(mtu, skb->len - offset) - doffset;
	87	struct sk_buff *nskb = alloc_skb(mtu + 32, GFP_ATOMIC);
	88
	89	if (!nskb)
	90	break;
	91	skb_reserve(nskb, 32);
48d49d0c	92	skb_set_mac_header(nskb, -ETH_HLEN);
c1d2bbe1	93	skb_reset_network_header(nskb);
eddc9ec5	94	iph = ip_hdr(nskb);
27d7ff46 ACM	95	skb_copy_to_linear_data(nskb, skb_network_header(skb),
27d7ff46 ACM	96	doffset);
1da177e4 LT	97	if (skb_copy_bits(skb,
	98	doffset + offset,
	99	nskb->data + doffset,
	100	frag_size))
	101	BUG();
	102	skb_put(nskb, doffset + frag_size);
	103	nskb->ip_summed = CHECKSUM_UNNECESSARY;
	104	nskb->dev = skb->dev;
	105	nskb->priority = skb->priority;
	106	nskb->protocol = skb->protocol;
	107	nskb->dst = dst_clone(skb->dst);
	108	memcpy(nskb->cb, skb->cb, sizeof(skb->cb));
	109	nskb->pkt_type = skb->pkt_type;
	110
d56f90a7	111	th = (struct tcphdr )(skb_network_header(nskb) + iph->ihl 4);
1da177e4 LT	112	iph->tot_len = htons(frag_size + doffset);
	113	iph->id = htons(id);
	114	iph->check = 0;
	115	iph->check = ip_fast_csum((unsigned char *) iph, iph->ihl);
	116	th->seq = htonl(seq);
	117	if (offset + doffset + frag_size < skb->len)
	118	th->fin = th->psh = 0;
	119	netif_rx(nskb);
	120	offset += frag_size;
	121	seq += frag_size;
	122	id++;
	123	}
	124
	125	dev_kfree_skb(skb);
	126	}
d2ae1d2f	127	#endif /* LOOPBACK_TSO */
1da177e4 LT	128
	129	/*
	130	* The higher levels take care of making this non-reentrant (it's
	131	* called with bh's disabled).
	132	*/
	133	static int loopback_xmit(struct sk_buff skb, struct net_device dev)
	134	{
5f6d88b9	135	struct pcpu_lstats pcpu_lstats, lb_stats;
1da177e4 LT	136
	137	skb_orphan(skb);
	138
0e920bfb	139	skb->protocol = eth_type_trans(skb,dev);
1da177e4 LT	140	#ifndef LOOPBACK_MUST_CHECKSUM
	141	skb->ip_summed = CHECKSUM_UNNECESSARY;
	142	#endif
	143
d2ae1d2f	144	#ifdef LOOPBACK_TSO
89114afd	145	if (skb_is_gso(skb)) {
1da177e4	146	BUG_ON(skb->protocol != htons(ETH_P_IP));
eddc9ec5	147	BUG_ON(ip_hdr(skb)->protocol != IPPROTO_TCP);
1da177e4 LT	148
	149	emulate_large_send_offload(skb);
	150	return 0;
	151	}
d2ae1d2f	152	#endif
1da177e4 LT	153	dev->last_rx = jiffies;
1da177e4 LT	154
9e0db4b1	155	/* it's OK to use per_cpu_ptr() because BHs are off */
5f6d88b9 EB	156	pcpu_lstats = netdev_priv(dev);
5f6d88b9 EB	157	lb_stats = per_cpu_ptr(pcpu_lstats, smp_processor_id());
5175c378 ED	158	lb_stats->bytes += skb->len;
5175c378 ED	159	lb_stats->packets++;
1da177e4 LT	160
	161	netif_rx(skb);
	162
58f53974	163	return 0;
1da177e4 LT	164	}
	165
	166	static struct net_device_stats get_stats(struct net_device dev)
	167	{
5f6d88b9	168	const struct pcpu_lstats *pcpu_lstats;
33036807	169	struct net_device_stats *stats = &dev->stats;
5175c378 ED	170	unsigned long bytes = 0;
5175c378 ED	171	unsigned long packets = 0;
1da177e4 LT	172	int i;
1da177e4 LT	173
5f6d88b9	174	pcpu_lstats = netdev_priv(dev);
0fed4846	175	for_each_possible_cpu(i) {
5175c378	176	const struct pcpu_lstats *lb_stats;
1da177e4	177
5f6d88b9	178	lb_stats = per_cpu_ptr(pcpu_lstats, i);
5175c378 ED	179	bytes += lb_stats->bytes;
5175c378 ED	180	packets += lb_stats->packets;
1da177e4	181	}
5175c378 ED	182	stats->rx_packets = packets;
	183	stats->tx_packets = packets;
	184	stats->rx_bytes = bytes;
	185	stats->tx_bytes = bytes;
1da177e4 LT	186	return stats;
	187	}
	188
7fa6b066	189	static u32 always_on(struct net_device *dev)
1da177e4 LT	190	{
	191	return 1;
	192	}
	193
7282d491	194	static const struct ethtool_ops loopback_ethtool_ops = {
7fa6b066	195	.get_link = always_on,
1da177e4	196	.set_tso = ethtool_op_set_tso,
7fa6b066 SH	197	.get_tx_csum = always_on,
	198	.get_sg = always_on,
	199	.get_rx_csum = always_on,
1da177e4 LT	200	};
1da177e4 LT	201
5f6d88b9 EB	202	static int loopback_dev_init(struct net_device *dev)
	203	{
	204	struct pcpu_lstats *lstats;
	205
	206	lstats = alloc_percpu(struct pcpu_lstats);
	207	if (!lstats)
	208	return -ENOMEM;
	209
	210	dev->priv = lstats;
	211	return 0;
	212	}
	213
	214	static void loopback_dev_free(struct net_device *dev)
	215	{
	216	struct pcpu_lstats *lstats = netdev_priv(dev);
	217
	218	free_percpu(lstats);
	219	free_netdev(dev);
	220	}
	221
7fa6b066	222	/*
9e0db4b1 EB	223	* The loopback device is special. There is only one instance
9e0db4b1 EB	224	* per network namespace.
7fa6b066	225	*/
854d8363 DL	226	static void loopback_setup(struct net_device *dev)
	227	{
	228	dev->get_stats = &get_stats;
	229	dev->mtu = (16 * 1024) + 20 + 20 + 12;
	230	dev->hard_start_xmit = loopback_xmit;
854d8363 DL	231	dev->hard_header_len = ETH_HLEN; /* 14 */
	232	dev->addr_len = ETH_ALEN; /* 6 */
	233	dev->tx_queue_len = 0;
	234	dev->type = ARPHRD_LOOPBACK; /* 0x0001*/
854d8363 DL	235	dev->flags = IFF_LOOPBACK;
854d8363 DL	236	dev->features = NETIF_F_SG \| NETIF_F_FRAGLIST
d2ae1d2f	237	#ifdef LOOPBACK_TSO
854d8363	238	\| NETIF_F_TSO
d2ae1d2f	239	#endif
854d8363 DL	240	\| NETIF_F_NO_CSUM
	241	\| NETIF_F_HIGHDMA
	242	\| NETIF_F_LLTX
2d2c54e3	243	\| NETIF_F_NETNS_LOCAL;
854d8363	244	dev->ethtool_ops = &loopback_ethtool_ops;
3b04ddde	245	dev->header_ops = &eth_header_ops;
5f6d88b9 EB	246	dev->init = loopback_dev_init;
5f6d88b9 EB	247	dev->destructor = loopback_dev_free;
854d8363	248	}
de3cb747	249
22783649	250	/* Setup and register the loopback device. */
4665079c	251	static __net_init int loopback_net_init(struct net *net)
1da177e4	252	{
854d8363 DL	253	struct net_device *dev;
	254	int err;
	255
	256	err = -ENOMEM;
	257	dev = alloc_netdev(0, "lo", loopback_setup);
	258	if (!dev)
	259	goto out;
aeed9e82	260
c346dca1	261	dev_net_set(dev, net);
854d8363	262	err = register_netdev(dev);
aeed9e82	263	if (err)
854d8363	264	goto out_free_netdev;
aeed9e82	265
2774c7ab	266	net->loopback_dev = dev;
9d6dda32	267	return 0;
854d8363	268
1da177e4	269
854d8363 DL	270	out_free_netdev:
854d8363 DL	271	free_netdev(dev);
9d6dda32 PE	272	out:
	273	if (net == &init_net)
	274	panic("loopback: Failed to register netdevice: %d\n", err);
	275	return err;
854d8363 DL	276	}
854d8363 DL	277
4665079c	278	static __net_exit void loopback_net_exit(struct net *net)
2774c7ab EB	279	{
	280	struct net_device *dev = net->loopback_dev;
	281
	282	unregister_netdev(dev);
	283	}
	284
022cbae6	285	static struct pernet_operations __net_initdata loopback_net_ops = {
2774c7ab EB	286	.init = loopback_net_init,
	287	.exit = loopback_net_exit,
	288	};
	289
	290	static int __init loopback_init(void)
	291	{
	292	return register_pernet_device(&loopback_net_ops);
	293	}
60903f2c	294
070ac3a2 DL	295	/* Loopback is special. It should be initialized before any other network
	296	* device and network subsystem.
	297	*/
2774c7ab	298	fs_initcall(loopback_init);