[mirror_ubuntu-artful-kernel.git] / net / ipv4 / netfilter / nf_conntrack_l3proto_ipv4.c

/* (C) 1999-2001 Paul `Rusty' Russell
 * (C) 2002-2004 Netfilter Core Team <coreteam@netfilter.org>
 *
 * 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.
 */

#include <linux/types.h>
#include <linux/ip.h>
#include <linux/netfilter.h>
#include <linux/module.h>
#include <linux/skbuff.h>
#include <linux/icmp.h>
#include <linux/sysctl.h>
#include <net/route.h>
#include <net/ip.h>

#include <linux/netfilter_ipv4.h>
#include <net/netfilter/nf_conntrack.h>
#include <net/netfilter/nf_conntrack_helper.h>
#include <net/netfilter/nf_conntrack_l4proto.h>
#include <net/netfilter/nf_conntrack_l3proto.h>
#include <net/netfilter/nf_conntrack_core.h>
#include <net/netfilter/ipv4/nf_conntrack_ipv4.h>

#if 0
#define DEBUGP printk
#else
#define DEBUGP(format, args...)
#endif

static int ipv4_pkt_to_tuple(const struct sk_buff *skb, unsigned int nhoff,
			     struct nf_conntrack_tuple *tuple)
{
	__be32 _addrs[2], *ap;
	ap = skb_header_pointer(skb, nhoff + offsetof(struct iphdr, saddr),
				sizeof(u_int32_t) * 2, _addrs);
	if (ap == NULL)
		return 0;

	tuple->src.u3.ip = ap[0];
	tuple->dst.u3.ip = ap[1];

	return 1;
}

static int ipv4_invert_tuple(struct nf_conntrack_tuple *tuple,
			   const struct nf_conntrack_tuple *orig)
{
	tuple->src.u3.ip = orig->dst.u3.ip;
	tuple->dst.u3.ip = orig->src.u3.ip;

	return 1;
}

static int ipv4_print_tuple(struct seq_file *s,
			    const struct nf_conntrack_tuple *tuple)
{
	return seq_printf(s, "src=%u.%u.%u.%u dst=%u.%u.%u.%u ",
			  NIPQUAD(tuple->src.u3.ip),
			  NIPQUAD(tuple->dst.u3.ip));
}

static int ipv4_print_conntrack(struct seq_file *s,
				const struct nf_conn *conntrack)
{
	return 0;
}

/* Returns new sk_buff, or NULL */
static struct sk_buff *
nf_ct_ipv4_gather_frags(struct sk_buff *skb, u_int32_t user)
{
	skb_orphan(skb);

	local_bh_disable();
	skb = ip_defrag(skb, user);
	local_bh_enable();

	if (skb)
		ip_send_check(ip_hdr(skb));

	return skb;
}

static int
ipv4_prepare(struct sk_buff **pskb, unsigned int hooknum, unsigned int *dataoff,
	     u_int8_t *protonum)
{
	/* Never happen */
	if (ip_hdr(*pskb)->frag_off & htons(IP_OFFSET)) {
		if (net_ratelimit()) {
			printk(KERN_ERR "ipv4_prepare: Frag of proto %u (hook=%u)\n",
			ip_hdr(*pskb)->protocol, hooknum);
		}
		return -NF_DROP;
	}

	*dataoff = skb_network_offset(*pskb) + ip_hdrlen(*pskb);
	*protonum = ip_hdr(*pskb)->protocol;

	return NF_ACCEPT;
}

static unsigned int ipv4_confirm(unsigned int hooknum,
				 struct sk_buff **pskb,
				 const struct net_device *in,
				 const struct net_device *out,
				 int (*okfn)(struct sk_buff *))
{
	/* We've seen it coming out the other side: confirm it */
	return nf_conntrack_confirm(pskb);
}

static unsigned int ipv4_conntrack_help(unsigned int hooknum,
				      struct sk_buff **pskb,
				      const struct net_device *in,
				      const struct net_device *out,
				      int (*okfn)(struct sk_buff *))
{
	struct nf_conn *ct;
	enum ip_conntrack_info ctinfo;
	struct nf_conn_help *help;
	struct nf_conntrack_helper *helper;

	/* This is where we call the helper: as the packet goes out. */
	ct = nf_ct_get(*pskb, &ctinfo);
	if (!ct || ctinfo == IP_CT_RELATED + IP_CT_IS_REPLY)
		return NF_ACCEPT;

	help = nfct_help(ct);
	if (!help)
		return NF_ACCEPT;
	/* rcu_read_lock()ed by nf_hook_slow */
	helper = rcu_dereference(help->helper);
	if (!helper)
		return NF_ACCEPT;
	return helper->help(pskb, skb_network_offset(*pskb) + ip_hdrlen(*pskb),
			    ct, ctinfo);
}

static unsigned int ipv4_conntrack_defrag(unsigned int hooknum,
					  struct sk_buff **pskb,
					  const struct net_device *in,
					  const struct net_device *out,
					  int (*okfn)(struct sk_buff *))
{
	/* Previously seen (loopback)?  Ignore.  Do this before
	   fragment check. */
	if ((*pskb)->nfct)
		return NF_ACCEPT;

	/* Gather fragments. */
	if (ip_hdr(*pskb)->frag_off & htons(IP_MF | IP_OFFSET)) {
		*pskb = nf_ct_ipv4_gather_frags(*pskb,
						hooknum == NF_IP_PRE_ROUTING ?
						IP_DEFRAG_CONNTRACK_IN :
						IP_DEFRAG_CONNTRACK_OUT);
		if (!*pskb)
			return NF_STOLEN;
	}
	return NF_ACCEPT;
}

static unsigned int ipv4_conntrack_in(unsigned int hooknum,
				      struct sk_buff **pskb,
				      const struct net_device *in,
				      const struct net_device *out,
				      int (*okfn)(struct sk_buff *))
{
	return nf_conntrack_in(PF_INET, hooknum, pskb);
}

static unsigned int ipv4_conntrack_local(unsigned int hooknum,
					 struct sk_buff **pskb,
					 const struct net_device *in,
					 const struct net_device *out,
					 int (*okfn)(struct sk_buff *))
{
	/* root is playing with raw sockets. */
	if ((*pskb)->len < sizeof(struct iphdr)
	    || ip_hdrlen(*pskb) < sizeof(struct iphdr)) {
		if (net_ratelimit())
			printk("ipt_hook: happy cracking.\n");
		return NF_ACCEPT;
	}
	return nf_conntrack_in(PF_INET, hooknum, pskb);
}

/* Connection tracking may drop packets, but never alters them, so
   make it the first hook. */
static struct nf_hook_ops ipv4_conntrack_ops[] = {
	{
		.hook		= ipv4_conntrack_defrag,
		.owner		= THIS_MODULE,
		.pf		= PF_INET,
		.hooknum	= NF_IP_PRE_ROUTING,
		.priority	= NF_IP_PRI_CONNTRACK_DEFRAG,
	},
	{
		.hook		= ipv4_conntrack_in,
		.owner		= THIS_MODULE,
		.pf		= PF_INET,
		.hooknum	= NF_IP_PRE_ROUTING,
		.priority	= NF_IP_PRI_CONNTRACK,
	},
	{
		.hook           = ipv4_conntrack_defrag,
		.owner          = THIS_MODULE,
		.pf             = PF_INET,
		.hooknum        = NF_IP_LOCAL_OUT,
		.priority       = NF_IP_PRI_CONNTRACK_DEFRAG,
	},
	{
		.hook		= ipv4_conntrack_local,
		.owner		= THIS_MODULE,
		.pf		= PF_INET,
		.hooknum	= NF_IP_LOCAL_OUT,
		.priority	= NF_IP_PRI_CONNTRACK,
	},
	{
		.hook		= ipv4_conntrack_help,
		.owner		= THIS_MODULE,
		.pf		= PF_INET,
		.hooknum	= NF_IP_POST_ROUTING,
		.priority	= NF_IP_PRI_CONNTRACK_HELPER,
	},
	{
		.hook		= ipv4_conntrack_help,
		.owner		= THIS_MODULE,
		.pf		= PF_INET,
		.hooknum	= NF_IP_LOCAL_IN,
		.priority	= NF_IP_PRI_CONNTRACK_HELPER,
	},
	{
		.hook		= ipv4_confirm,
		.owner		= THIS_MODULE,
		.pf		= PF_INET,
		.hooknum	= NF_IP_POST_ROUTING,
		.priority	= NF_IP_PRI_CONNTRACK_CONFIRM,
	},
	{
		.hook		= ipv4_confirm,
		.owner		= THIS_MODULE,
		.pf		= PF_INET,
		.hooknum	= NF_IP_LOCAL_IN,
		.priority	= NF_IP_PRI_CONNTRACK_CONFIRM,
	},
};

#if defined(CONFIG_SYSCTL) && defined(CONFIG_NF_CONNTRACK_PROC_COMPAT)
static int log_invalid_proto_min = 0;
static int log_invalid_proto_max = 255;

static ctl_table ip_ct_sysctl_table[] = {
	{
		.ctl_name	= NET_IPV4_NF_CONNTRACK_MAX,
		.procname	= "ip_conntrack_max",
		.data		= &nf_conntrack_max,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= &proc_dointvec,
	},
	{
		.ctl_name	= NET_IPV4_NF_CONNTRACK_COUNT,
		.procname	= "ip_conntrack_count",
		.data		= &nf_conntrack_count,
		.maxlen		= sizeof(int),
		.mode		= 0444,
		.proc_handler	= &proc_dointvec,
	},
	{
		.ctl_name	= NET_IPV4_NF_CONNTRACK_BUCKETS,
		.procname	= "ip_conntrack_buckets",
		.data		= &nf_conntrack_htable_size,
		.maxlen		= sizeof(unsigned int),
		.mode		= 0444,
		.proc_handler	= &proc_dointvec,
	},
	{
		.ctl_name	= NET_IPV4_NF_CONNTRACK_CHECKSUM,
		.procname	= "ip_conntrack_checksum",
		.data		= &nf_conntrack_checksum,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= &proc_dointvec,
	},
	{
		.ctl_name	= NET_IPV4_NF_CONNTRACK_LOG_INVALID,
		.procname	= "ip_conntrack_log_invalid",
		.data		= &nf_ct_log_invalid,
		.maxlen		= sizeof(unsigned int),
		.mode		= 0644,
		.proc_handler	= &proc_dointvec_minmax,
		.strategy	= &sysctl_intvec,
		.extra1		= &log_invalid_proto_min,
		.extra2		= &log_invalid_proto_max,
	},
	{
		.ctl_name	= 0
	}
};
#endif /* CONFIG_SYSCTL && CONFIG_NF_CONNTRACK_PROC_COMPAT */

/* Fast function for those who don't want to parse /proc (and I don't
   blame them). */
/* Reversing the socket's dst/src point of view gives us the reply
   mapping. */
static int
getorigdst(struct sock *sk, int optval, void __user *user, int *len)
{
	struct inet_sock *inet = inet_sk(sk);
	struct nf_conntrack_tuple_hash *h;
	struct nf_conntrack_tuple tuple;

	NF_CT_TUPLE_U_BLANK(&tuple);
	tuple.src.u3.ip = inet->rcv_saddr;
	tuple.src.u.tcp.port = inet->sport;
	tuple.dst.u3.ip = inet->daddr;
	tuple.dst.u.tcp.port = inet->dport;
	tuple.src.l3num = PF_INET;
	tuple.dst.protonum = IPPROTO_TCP;

	/* We only do TCP at the moment: is there a better way? */
	if (strcmp(sk->sk_prot->name, "TCP")) {
		DEBUGP("SO_ORIGINAL_DST: Not a TCP socket\n");
		return -ENOPROTOOPT;
	}

	if ((unsigned int) *len < sizeof(struct sockaddr_in)) {
		DEBUGP("SO_ORIGINAL_DST: len %u not %u\n",
		       *len, sizeof(struct sockaddr_in));
		return -EINVAL;
	}

	h = nf_conntrack_find_get(&tuple, NULL);
	if (h) {
		struct sockaddr_in sin;
		struct nf_conn *ct = nf_ct_tuplehash_to_ctrack(h);

		sin.sin_family = AF_INET;
		sin.sin_port = ct->tuplehash[IP_CT_DIR_ORIGINAL]
			.tuple.dst.u.tcp.port;
		sin.sin_addr.s_addr = ct->tuplehash[IP_CT_DIR_ORIGINAL]
			.tuple.dst.u3.ip;
		memset(sin.sin_zero, 0, sizeof(sin.sin_zero));

		DEBUGP("SO_ORIGINAL_DST: %u.%u.%u.%u %u\n",
		       NIPQUAD(sin.sin_addr.s_addr), ntohs(sin.sin_port));
		nf_ct_put(ct);
		if (copy_to_user(user, &sin, sizeof(sin)) != 0)
			return -EFAULT;
		else
			return 0;
	}
	DEBUGP("SO_ORIGINAL_DST: Can't find %u.%u.%u.%u/%u-%u.%u.%u.%u/%u.\n",
	       NIPQUAD(tuple.src.u3.ip), ntohs(tuple.src.u.tcp.port),
	       NIPQUAD(tuple.dst.u3.ip), ntohs(tuple.dst.u.tcp.port));
	return -ENOENT;
}

#if defined(CONFIG_NF_CT_NETLINK) || defined(CONFIG_NF_CT_NETLINK_MODULE)

#include <linux/netfilter/nfnetlink.h>
#include <linux/netfilter/nfnetlink_conntrack.h>

static int ipv4_tuple_to_nfattr(struct sk_buff *skb,
				const struct nf_conntrack_tuple *tuple)
{
	NFA_PUT(skb, CTA_IP_V4_SRC, sizeof(u_int32_t),
		&tuple->src.u3.ip);
	NFA_PUT(skb, CTA_IP_V4_DST, sizeof(u_int32_t),
		&tuple->dst.u3.ip);
	return 0;

nfattr_failure:
	return -1;
}

static const size_t cta_min_ip[CTA_IP_MAX] = {
	[CTA_IP_V4_SRC-1]       = sizeof(u_int32_t),
	[CTA_IP_V4_DST-1]       = sizeof(u_int32_t),
};

static int ipv4_nfattr_to_tuple(struct nfattr *tb[],
				struct nf_conntrack_tuple *t)
{
	if (!tb[CTA_IP_V4_SRC-1] || !tb[CTA_IP_V4_DST-1])
		return -EINVAL;

	if (nfattr_bad_size(tb, CTA_IP_MAX, cta_min_ip))
		return -EINVAL;

	t->src.u3.ip = *(__be32 *)NFA_DATA(tb[CTA_IP_V4_SRC-1]);
	t->dst.u3.ip = *(__be32 *)NFA_DATA(tb[CTA_IP_V4_DST-1]);

	return 0;
}
#endif

static struct nf_sockopt_ops so_getorigdst = {
	.pf		= PF_INET,
	.get_optmin	= SO_ORIGINAL_DST,
	.get_optmax	= SO_ORIGINAL_DST+1,
	.get		= &getorigdst,
};

struct nf_conntrack_l3proto nf_conntrack_l3proto_ipv4 = {
	.l3proto	 = PF_INET,
	.name		 = "ipv4",
	.pkt_to_tuple	 = ipv4_pkt_to_tuple,
	.invert_tuple	 = ipv4_invert_tuple,
	.print_tuple	 = ipv4_print_tuple,
	.print_conntrack = ipv4_print_conntrack,
	.prepare	 = ipv4_prepare,
#if defined(CONFIG_NF_CT_NETLINK) || defined(CONFIG_NF_CT_NETLINK_MODULE)
	.tuple_to_nfattr = ipv4_tuple_to_nfattr,
	.nfattr_to_tuple = ipv4_nfattr_to_tuple,
#endif
#if defined(CONFIG_SYSCTL) && defined(CONFIG_NF_CONNTRACK_PROC_COMPAT)
	.ctl_table_path  = nf_net_ipv4_netfilter_sysctl_path,
	.ctl_table	 = ip_ct_sysctl_table,
#endif
	.me		 = THIS_MODULE,
};

MODULE_ALIAS("nf_conntrack-" __stringify(AF_INET));
MODULE_ALIAS("ip_conntrack");
MODULE_LICENSE("GPL");

static int __init nf_conntrack_l3proto_ipv4_init(void)
{
	int ret = 0;

	need_conntrack();

	ret = nf_register_sockopt(&so_getorigdst);
	if (ret < 0) {
		printk(KERN_ERR "Unable to register netfilter socket option\n");
		return ret;
	}

	ret = nf_conntrack_l4proto_register(&nf_conntrack_l4proto_tcp4);
	if (ret < 0) {
		printk("nf_conntrack_ipv4: can't register tcp.\n");
		goto cleanup_sockopt;
	}

	ret = nf_conntrack_l4proto_register(&nf_conntrack_l4proto_udp4);
	if (ret < 0) {
		printk("nf_conntrack_ipv4: can't register udp.\n");
		goto cleanup_tcp;
	}

	ret = nf_conntrack_l4proto_register(&nf_conntrack_l4proto_icmp);
	if (ret < 0) {
		printk("nf_conntrack_ipv4: can't register icmp.\n");
		goto cleanup_udp;
	}

	ret = nf_conntrack_l3proto_register(&nf_conntrack_l3proto_ipv4);
	if (ret < 0) {
		printk("nf_conntrack_ipv4: can't register ipv4\n");
		goto cleanup_icmp;
	}

	ret = nf_register_hooks(ipv4_conntrack_ops,
				ARRAY_SIZE(ipv4_conntrack_ops));
	if (ret < 0) {
		printk("nf_conntrack_ipv4: can't register hooks.\n");
		goto cleanup_ipv4;
	}
#if defined(CONFIG_PROC_FS) && defined(CONFIG_NF_CONNTRACK_PROC_COMPAT)
	ret = nf_conntrack_ipv4_compat_init();
	if (ret < 0)
		goto cleanup_hooks;
#endif
	return ret;
#if defined(CONFIG_PROC_FS) && defined(CONFIG_NF_CONNTRACK_PROC_COMPAT)
 cleanup_hooks:
	nf_unregister_hooks(ipv4_conntrack_ops, ARRAY_SIZE(ipv4_conntrack_ops));
#endif
 cleanup_ipv4:
	nf_conntrack_l3proto_unregister(&nf_conntrack_l3proto_ipv4);
 cleanup_icmp:
	nf_conntrack_l4proto_unregister(&nf_conntrack_l4proto_icmp);
 cleanup_udp:
	nf_conntrack_l4proto_unregister(&nf_conntrack_l4proto_udp4);
 cleanup_tcp:
	nf_conntrack_l4proto_unregister(&nf_conntrack_l4proto_tcp4);
 cleanup_sockopt:
	nf_unregister_sockopt(&so_getorigdst);
	return ret;
}

static void __exit nf_conntrack_l3proto_ipv4_fini(void)
{
	synchronize_net();
#if defined(CONFIG_PROC_FS) && defined(CONFIG_NF_CONNTRACK_PROC_COMPAT)
	nf_conntrack_ipv4_compat_fini();
#endif
	nf_unregister_hooks(ipv4_conntrack_ops, ARRAY_SIZE(ipv4_conntrack_ops));
	nf_conntrack_l3proto_unregister(&nf_conntrack_l3proto_ipv4);
	nf_conntrack_l4proto_unregister(&nf_conntrack_l4proto_icmp);
	nf_conntrack_l4proto_unregister(&nf_conntrack_l4proto_udp4);
	nf_conntrack_l4proto_unregister(&nf_conntrack_l4proto_tcp4);
	nf_unregister_sockopt(&so_getorigdst);
}

module_init(nf_conntrack_l3proto_ipv4_init);
module_exit(nf_conntrack_l3proto_ipv4_fini);
Commit	Line	Data
9fb9cbb1 YK	1	/* (C) 1999-2001 Paul `Rusty' Russell
	2	* (C) 2002-2004 Netfilter Core Team <coreteam@netfilter.org>
	3	*
	4	* This program is free software; you can redistribute it and/or modify
	5	* it under the terms of the GNU General Public License version 2 as
	6	* published by the Free Software Foundation.
9fb9cbb1 YK	7	*/
9fb9cbb1 YK	8
9fb9cbb1 YK	9	#include <linux/types.h>
	10	#include <linux/ip.h>
	11	#include <linux/netfilter.h>
	12	#include <linux/module.h>
	13	#include <linux/skbuff.h>
	14	#include <linux/icmp.h>
	15	#include <linux/sysctl.h>
0ae2cfe7	16	#include <net/route.h>
9fb9cbb1 YK	17	#include <net/ip.h>
	18
	19	#include <linux/netfilter_ipv4.h>
	20	#include <net/netfilter/nf_conntrack.h>
	21	#include <net/netfilter/nf_conntrack_helper.h>
605dcad6	22	#include <net/netfilter/nf_conntrack_l4proto.h>
9fb9cbb1 YK	23	#include <net/netfilter/nf_conntrack_l3proto.h>
	24	#include <net/netfilter/nf_conntrack_core.h>
	25	#include <net/netfilter/ipv4/nf_conntrack_ipv4.h>
	26
	27	#if 0
	28	#define DEBUGP printk
	29	#else
	30	#define DEBUGP(format, args...)
	31	#endif
	32
9fb9cbb1 YK	33	static int ipv4_pkt_to_tuple(const struct sk_buff *skb, unsigned int nhoff,
	34	struct nf_conntrack_tuple *tuple)
	35	{
bff9a89b	36	__be32 _addrs[2], *ap;
9fb9cbb1 YK	37	ap = skb_header_pointer(skb, nhoff + offsetof(struct iphdr, saddr),
	38	sizeof(u_int32_t) * 2, _addrs);
	39	if (ap == NULL)
	40	return 0;
	41
	42	tuple->src.u3.ip = ap[0];
	43	tuple->dst.u3.ip = ap[1];
	44
	45	return 1;
	46	}
	47
	48	static int ipv4_invert_tuple(struct nf_conntrack_tuple *tuple,
	49	const struct nf_conntrack_tuple *orig)
	50	{
	51	tuple->src.u3.ip = orig->dst.u3.ip;
	52	tuple->dst.u3.ip = orig->src.u3.ip;
	53
	54	return 1;
	55	}
	56
	57	static int ipv4_print_tuple(struct seq_file *s,
	58	const struct nf_conntrack_tuple *tuple)
	59	{
	60	return seq_printf(s, "src=%u.%u.%u.%u dst=%u.%u.%u.%u ",
e905a9ed	61	NIPQUAD(tuple->src.u3.ip),
9fb9cbb1 YK	62	NIPQUAD(tuple->dst.u3.ip));
	63	}
	64
	65	static int ipv4_print_conntrack(struct seq_file *s,
	66	const struct nf_conn *conntrack)
	67	{
	68	return 0;
	69	}
	70
	71	/* Returns new sk_buff, or NULL */
	72	static struct sk_buff *
	73	nf_ct_ipv4_gather_frags(struct sk_buff *skb, u_int32_t user)
	74	{
	75	skb_orphan(skb);
	76
e905a9ed YH	77	local_bh_disable();
	78	skb = ip_defrag(skb, user);
	79	local_bh_enable();
9fb9cbb1	80
e905a9ed	81	if (skb)
eddc9ec5	82	ip_send_check(ip_hdr(skb));
9fb9cbb1	83
e905a9ed	84	return skb;
9fb9cbb1 YK	85	}
	86
	87	static int
	88	ipv4_prepare(struct sk_buff *pskb, unsigned int hooknum, unsigned int dataoff,
	89	u_int8_t *protonum)
	90	{
	91	/* Never happen */
eddc9ec5	92	if (ip_hdr(*pskb)->frag_off & htons(IP_OFFSET)) {
9fb9cbb1 YK	93	if (net_ratelimit()) {
9fb9cbb1 YK	94	printk(KERN_ERR "ipv4_prepare: Frag of proto %u (hook=%u)\n",
eddc9ec5	95	ip_hdr(*pskb)->protocol, hooknum);
9fb9cbb1 YK	96	}
	97	return -NF_DROP;
	98	}
	99
c9bdd4b5	100	dataoff = skb_network_offset(pskb) + ip_hdrlen(*pskb);
eddc9ec5	101	protonum = ip_hdr(pskb)->protocol;
9fb9cbb1 YK	102
	103	return NF_ACCEPT;
	104	}
	105
9fb9cbb1 YK	106	static unsigned int ipv4_confirm(unsigned int hooknum,
	107	struct sk_buff **pskb,
	108	const struct net_device *in,
	109	const struct net_device *out,
	110	int (okfn)(struct sk_buff ))
	111	{
	112	/* We've seen it coming out the other side: confirm it */
	113	return nf_conntrack_confirm(pskb);
	114	}
	115
	116	static unsigned int ipv4_conntrack_help(unsigned int hooknum,
	117	struct sk_buff **pskb,
	118	const struct net_device *in,
	119	const struct net_device *out,
	120	int (okfn)(struct sk_buff ))
	121	{
	122	struct nf_conn *ct;
	123	enum ip_conntrack_info ctinfo;
dc808fe2	124	struct nf_conn_help *help;
3c158f7f	125	struct nf_conntrack_helper *helper;
9fb9cbb1 YK	126
	127	/* This is where we call the helper: as the packet goes out. */
	128	ct = nf_ct_get(*pskb, &ctinfo);
6442f1cf	129	if (!ct \|\| ctinfo == IP_CT_RELATED + IP_CT_IS_REPLY)
dc808fe2 HW	130	return NF_ACCEPT;
	131
	132	help = nfct_help(ct);
3c158f7f	133	if (!help)
dc808fe2	134	return NF_ACCEPT;
3c158f7f PM	135	/* rcu_read_lock()ed by nf_hook_slow */
	136	helper = rcu_dereference(help->helper);
	137	if (!helper)
	138	return NF_ACCEPT;
	139	return helper->help(pskb, skb_network_offset(pskb) + ip_hdrlen(pskb),
	140	ct, ctinfo);
9fb9cbb1 YK	141	}
	142
	143	static unsigned int ipv4_conntrack_defrag(unsigned int hooknum,
	144	struct sk_buff **pskb,
	145	const struct net_device *in,
	146	const struct net_device *out,
	147	int (okfn)(struct sk_buff ))
	148	{
9fb9cbb1 YK	149	/* Previously seen (loopback)? Ignore. Do this before
	150	fragment check. */
	151	if ((*pskb)->nfct)
	152	return NF_ACCEPT;
9fb9cbb1 YK	153
9fb9cbb1 YK	154	/* Gather fragments. */
eddc9ec5	155	if (ip_hdr(*pskb)->frag_off & htons(IP_MF \| IP_OFFSET)) {
9fb9cbb1 YK	156	pskb = nf_ct_ipv4_gather_frags(pskb,
	157	hooknum == NF_IP_PRE_ROUTING ?
	158	IP_DEFRAG_CONNTRACK_IN :
	159	IP_DEFRAG_CONNTRACK_OUT);
	160	if (!*pskb)
	161	return NF_STOLEN;
	162	}
	163	return NF_ACCEPT;
	164	}
	165
9fb9cbb1 YK	166	static unsigned int ipv4_conntrack_in(unsigned int hooknum,
	167	struct sk_buff **pskb,
	168	const struct net_device *in,
	169	const struct net_device *out,
	170	int (okfn)(struct sk_buff ))
	171	{
	172	return nf_conntrack_in(PF_INET, hooknum, pskb);
	173	}
	174
	175	static unsigned int ipv4_conntrack_local(unsigned int hooknum,
e905a9ed YH	176	struct sk_buff **pskb,
	177	const struct net_device *in,
	178	const struct net_device *out,
	179	int (okfn)(struct sk_buff ))
9fb9cbb1 YK	180	{
	181	/* root is playing with raw sockets. */
	182	if ((*pskb)->len < sizeof(struct iphdr)
c9bdd4b5	183	\|\| ip_hdrlen(*pskb) < sizeof(struct iphdr)) {
9fb9cbb1 YK	184	if (net_ratelimit())
	185	printk("ipt_hook: happy cracking.\n");
	186	return NF_ACCEPT;
	187	}
	188	return nf_conntrack_in(PF_INET, hooknum, pskb);
	189	}
	190
	191	/* Connection tracking may drop packets, but never alters them, so
	192	make it the first hook. */
964ddaa1 PM	193	static struct nf_hook_ops ipv4_conntrack_ops[] = {
	194	{
	195	.hook = ipv4_conntrack_defrag,
	196	.owner = THIS_MODULE,
	197	.pf = PF_INET,
	198	.hooknum = NF_IP_PRE_ROUTING,
	199	.priority = NF_IP_PRI_CONNTRACK_DEFRAG,
	200	},
	201	{
	202	.hook = ipv4_conntrack_in,
	203	.owner = THIS_MODULE,
	204	.pf = PF_INET,
	205	.hooknum = NF_IP_PRE_ROUTING,
	206	.priority = NF_IP_PRI_CONNTRACK,
	207	},
	208	{
	209	.hook = ipv4_conntrack_defrag,
	210	.owner = THIS_MODULE,
	211	.pf = PF_INET,
	212	.hooknum = NF_IP_LOCAL_OUT,
	213	.priority = NF_IP_PRI_CONNTRACK_DEFRAG,
	214	},
	215	{
	216	.hook = ipv4_conntrack_local,
	217	.owner = THIS_MODULE,
	218	.pf = PF_INET,
	219	.hooknum = NF_IP_LOCAL_OUT,
	220	.priority = NF_IP_PRI_CONNTRACK,
	221	},
	222	{
	223	.hook = ipv4_conntrack_help,
	224	.owner = THIS_MODULE,
	225	.pf = PF_INET,
	226	.hooknum = NF_IP_POST_ROUTING,
	227	.priority = NF_IP_PRI_CONNTRACK_HELPER,
	228	},
	229	{
	230	.hook = ipv4_conntrack_help,
	231	.owner = THIS_MODULE,
	232	.pf = PF_INET,
	233	.hooknum = NF_IP_LOCAL_IN,
	234	.priority = NF_IP_PRI_CONNTRACK_HELPER,
	235	},
	236	{
	237	.hook = ipv4_confirm,
	238	.owner = THIS_MODULE,
	239	.pf = PF_INET,
	240	.hooknum = NF_IP_POST_ROUTING,
	241	.priority = NF_IP_PRI_CONNTRACK_CONFIRM,
	242	},
	243	{
	244	.hook = ipv4_confirm,
	245	.owner = THIS_MODULE,
	246	.pf = PF_INET,
	247	.hooknum = NF_IP_LOCAL_IN,
	248	.priority = NF_IP_PRI_CONNTRACK_CONFIRM,
	249	},
9fb9cbb1 YK	250	};
9fb9cbb1 YK	251
a999e683 PM	252	#if defined(CONFIG_SYSCTL) && defined(CONFIG_NF_CONNTRACK_PROC_COMPAT)
	253	static int log_invalid_proto_min = 0;
	254	static int log_invalid_proto_max = 255;
	255
	256	static ctl_table ip_ct_sysctl_table[] = {
	257	{
	258	.ctl_name = NET_IPV4_NF_CONNTRACK_MAX,
	259	.procname = "ip_conntrack_max",
	260	.data = &nf_conntrack_max,
	261	.maxlen = sizeof(int),
	262	.mode = 0644,
	263	.proc_handler = &proc_dointvec,
	264	},
	265	{
	266	.ctl_name = NET_IPV4_NF_CONNTRACK_COUNT,
	267	.procname = "ip_conntrack_count",
	268	.data = &nf_conntrack_count,
	269	.maxlen = sizeof(int),
	270	.mode = 0444,
	271	.proc_handler = &proc_dointvec,
	272	},
	273	{
	274	.ctl_name = NET_IPV4_NF_CONNTRACK_BUCKETS,
	275	.procname = "ip_conntrack_buckets",
	276	.data = &nf_conntrack_htable_size,
	277	.maxlen = sizeof(unsigned int),
	278	.mode = 0444,
	279	.proc_handler = &proc_dointvec,
	280	},
	281	{
	282	.ctl_name = NET_IPV4_NF_CONNTRACK_CHECKSUM,
	283	.procname = "ip_conntrack_checksum",
	284	.data = &nf_conntrack_checksum,
	285	.maxlen = sizeof(int),
	286	.mode = 0644,
	287	.proc_handler = &proc_dointvec,
	288	},
	289	{
	290	.ctl_name = NET_IPV4_NF_CONNTRACK_LOG_INVALID,
	291	.procname = "ip_conntrack_log_invalid",
	292	.data = &nf_ct_log_invalid,
	293	.maxlen = sizeof(unsigned int),
	294	.mode = 0644,
	295	.proc_handler = &proc_dointvec_minmax,
	296	.strategy = &sysctl_intvec,
	297	.extra1 = &log_invalid_proto_min,
	298	.extra2 = &log_invalid_proto_max,
	299	},
	300	{
	301	.ctl_name = 0
	302	}
	303	};
	304	#endif /* CONFIG_SYSCTL && CONFIG_NF_CONNTRACK_PROC_COMPAT */
	305
9fb9cbb1 YK	306	/* Fast function for those who don't want to parse /proc (and I don't
	307	blame them). */
	308	/* Reversing the socket's dst/src point of view gives us the reply
	309	mapping. */
	310	static int
	311	getorigdst(struct sock sk, int optval, void __user user, int *len)
	312	{
	313	struct inet_sock *inet = inet_sk(sk);
	314	struct nf_conntrack_tuple_hash *h;
	315	struct nf_conntrack_tuple tuple;
e905a9ed	316
9fb9cbb1 YK	317	NF_CT_TUPLE_U_BLANK(&tuple);
	318	tuple.src.u3.ip = inet->rcv_saddr;
	319	tuple.src.u.tcp.port = inet->sport;
	320	tuple.dst.u3.ip = inet->daddr;
	321	tuple.dst.u.tcp.port = inet->dport;
	322	tuple.src.l3num = PF_INET;
	323	tuple.dst.protonum = IPPROTO_TCP;
	324
	325	/* We only do TCP at the moment: is there a better way? */
	326	if (strcmp(sk->sk_prot->name, "TCP")) {
	327	DEBUGP("SO_ORIGINAL_DST: Not a TCP socket\n");
	328	return -ENOPROTOOPT;
	329	}
	330
	331	if ((unsigned int) *len < sizeof(struct sockaddr_in)) {
	332	DEBUGP("SO_ORIGINAL_DST: len %u not %u\n",
	333	*len, sizeof(struct sockaddr_in));
	334	return -EINVAL;
	335	}
	336
	337	h = nf_conntrack_find_get(&tuple, NULL);
	338	if (h) {
	339	struct sockaddr_in sin;
	340	struct nf_conn *ct = nf_ct_tuplehash_to_ctrack(h);
	341
	342	sin.sin_family = AF_INET;
	343	sin.sin_port = ct->tuplehash[IP_CT_DIR_ORIGINAL]
	344	.tuple.dst.u.tcp.port;
	345	sin.sin_addr.s_addr = ct->tuplehash[IP_CT_DIR_ORIGINAL]
	346	.tuple.dst.u3.ip;
6c813c3f	347	memset(sin.sin_zero, 0, sizeof(sin.sin_zero));
9fb9cbb1 YK	348
	349	DEBUGP("SO_ORIGINAL_DST: %u.%u.%u.%u %u\n",
	350	NIPQUAD(sin.sin_addr.s_addr), ntohs(sin.sin_port));
	351	nf_ct_put(ct);
	352	if (copy_to_user(user, &sin, sizeof(sin)) != 0)
	353	return -EFAULT;
	354	else
	355	return 0;
	356	}
	357	DEBUGP("SO_ORIGINAL_DST: Can't find %u.%u.%u.%u/%u-%u.%u.%u.%u/%u.\n",
	358	NIPQUAD(tuple.src.u3.ip), ntohs(tuple.src.u.tcp.port),
	359	NIPQUAD(tuple.dst.u3.ip), ntohs(tuple.dst.u.tcp.port));
	360	return -ENOENT;
	361	}
	362
e281db5c	363	#if defined(CONFIG_NF_CT_NETLINK) \|\| defined(CONFIG_NF_CT_NETLINK_MODULE)
c1d10adb PNA	364
	365	#include <linux/netfilter/nfnetlink.h>
	366	#include <linux/netfilter/nfnetlink_conntrack.h>
	367
	368	static int ipv4_tuple_to_nfattr(struct sk_buff *skb,
	369	const struct nf_conntrack_tuple *tuple)
	370	{
	371	NFA_PUT(skb, CTA_IP_V4_SRC, sizeof(u_int32_t),
	372	&tuple->src.u3.ip);
	373	NFA_PUT(skb, CTA_IP_V4_DST, sizeof(u_int32_t),
	374	&tuple->dst.u3.ip);
	375	return 0;
	376
	377	nfattr_failure:
	378	return -1;
	379	}
	380
	381	static const size_t cta_min_ip[CTA_IP_MAX] = {
	382	[CTA_IP_V4_SRC-1] = sizeof(u_int32_t),
	383	[CTA_IP_V4_DST-1] = sizeof(u_int32_t),
	384	};
	385
	386	static int ipv4_nfattr_to_tuple(struct nfattr *tb[],
	387	struct nf_conntrack_tuple *t)
	388	{
	389	if (!tb[CTA_IP_V4_SRC-1] \|\| !tb[CTA_IP_V4_DST-1])
	390	return -EINVAL;
	391
	392	if (nfattr_bad_size(tb, CTA_IP_MAX, cta_min_ip))
	393	return -EINVAL;
	394
bff9a89b PM	395	t->src.u3.ip = (__be32 )NFA_DATA(tb[CTA_IP_V4_SRC-1]);
bff9a89b PM	396	t->dst.u3.ip = (__be32 )NFA_DATA(tb[CTA_IP_V4_DST-1]);
c1d10adb PNA	397
	398	return 0;
	399	}
	400	#endif
	401
9fb9cbb1 YK	402	static struct nf_sockopt_ops so_getorigdst = {
	403	.pf = PF_INET,
	404	.get_optmin = SO_ORIGINAL_DST,
	405	.get_optmax = SO_ORIGINAL_DST+1,
	406	.get = &getorigdst,
	407	};
	408
	409	struct nf_conntrack_l3proto nf_conntrack_l3proto_ipv4 = {
	410	.l3proto = PF_INET,
	411	.name = "ipv4",
	412	.pkt_to_tuple = ipv4_pkt_to_tuple,
	413	.invert_tuple = ipv4_invert_tuple,
	414	.print_tuple = ipv4_print_tuple,
	415	.print_conntrack = ipv4_print_conntrack,
	416	.prepare = ipv4_prepare,
e281db5c	417	#if defined(CONFIG_NF_CT_NETLINK) \|\| defined(CONFIG_NF_CT_NETLINK_MODULE)
c1d10adb PNA	418	.tuple_to_nfattr = ipv4_tuple_to_nfattr,
c1d10adb PNA	419	.nfattr_to_tuple = ipv4_nfattr_to_tuple,
a999e683 PM	420	#endif
	421	#if defined(CONFIG_SYSCTL) && defined(CONFIG_NF_CONNTRACK_PROC_COMPAT)
	422	.ctl_table_path = nf_net_ipv4_netfilter_sysctl_path,
	423	.ctl_table = ip_ct_sysctl_table,
c1d10adb	424	#endif
9fb9cbb1 YK	425	.me = THIS_MODULE,
	426	};
	427
32292a7f	428	MODULE_ALIAS("nf_conntrack-" __stringify(AF_INET));
d2483dde	429	MODULE_ALIAS("ip_conntrack");
32292a7f PM	430	MODULE_LICENSE("GPL");
	431
	432	static int __init nf_conntrack_l3proto_ipv4_init(void)
9fb9cbb1 YK	433	{
	434	int ret = 0;
	435
32292a7f	436	need_conntrack();
9fb9cbb1 YK	437
	438	ret = nf_register_sockopt(&so_getorigdst);
	439	if (ret < 0) {
	440	printk(KERN_ERR "Unable to register netfilter socket option\n");
32292a7f	441	return ret;
9fb9cbb1 YK	442	}
9fb9cbb1 YK	443
605dcad6	444	ret = nf_conntrack_l4proto_register(&nf_conntrack_l4proto_tcp4);
9fb9cbb1 YK	445	if (ret < 0) {
	446	printk("nf_conntrack_ipv4: can't register tcp.\n");
	447	goto cleanup_sockopt;
	448	}
	449
605dcad6	450	ret = nf_conntrack_l4proto_register(&nf_conntrack_l4proto_udp4);
9fb9cbb1 YK	451	if (ret < 0) {
	452	printk("nf_conntrack_ipv4: can't register udp.\n");
	453	goto cleanup_tcp;
	454	}
	455
605dcad6	456	ret = nf_conntrack_l4proto_register(&nf_conntrack_l4proto_icmp);
9fb9cbb1 YK	457	if (ret < 0) {
	458	printk("nf_conntrack_ipv4: can't register icmp.\n");
	459	goto cleanup_udp;
	460	}
	461
	462	ret = nf_conntrack_l3proto_register(&nf_conntrack_l3proto_ipv4);
	463	if (ret < 0) {
	464	printk("nf_conntrack_ipv4: can't register ipv4\n");
	465	goto cleanup_icmp;
	466	}
	467
964ddaa1 PM	468	ret = nf_register_hooks(ipv4_conntrack_ops,
964ddaa1 PM	469	ARRAY_SIZE(ipv4_conntrack_ops));
9fb9cbb1	470	if (ret < 0) {
964ddaa1	471	printk("nf_conntrack_ipv4: can't register hooks.\n");
9fb9cbb1 YK	472	goto cleanup_ipv4;
9fb9cbb1 YK	473	}
e4bd8bce PM	474	#if defined(CONFIG_PROC_FS) && defined(CONFIG_NF_CONNTRACK_PROC_COMPAT)
	475	ret = nf_conntrack_ipv4_compat_init();
	476	if (ret < 0)
	477	goto cleanup_hooks;
	478	#endif
9fb9cbb1	479	return ret;
e4bd8bce PM	480	#if defined(CONFIG_PROC_FS) && defined(CONFIG_NF_CONNTRACK_PROC_COMPAT)
e4bd8bce PM	481	cleanup_hooks:
e905a9ed	482	nf_unregister_hooks(ipv4_conntrack_ops, ARRAY_SIZE(ipv4_conntrack_ops));
e4bd8bce	483	#endif
9fb9cbb1 YK	484	cleanup_ipv4:
	485	nf_conntrack_l3proto_unregister(&nf_conntrack_l3proto_ipv4);
	486	cleanup_icmp:
605dcad6	487	nf_conntrack_l4proto_unregister(&nf_conntrack_l4proto_icmp);
9fb9cbb1	488	cleanup_udp:
605dcad6	489	nf_conntrack_l4proto_unregister(&nf_conntrack_l4proto_udp4);
9fb9cbb1	490	cleanup_tcp:
605dcad6	491	nf_conntrack_l4proto_unregister(&nf_conntrack_l4proto_tcp4);
9fb9cbb1 YK	492	cleanup_sockopt:
9fb9cbb1 YK	493	nf_unregister_sockopt(&so_getorigdst);
9fb9cbb1 YK	494	return ret;
	495	}
	496
65b4b4e8	497	static void __exit nf_conntrack_l3proto_ipv4_fini(void)
9fb9cbb1	498	{
32292a7f	499	synchronize_net();
e4bd8bce PM	500	#if defined(CONFIG_PROC_FS) && defined(CONFIG_NF_CONNTRACK_PROC_COMPAT)
	501	nf_conntrack_ipv4_compat_fini();
	502	#endif
32292a7f PM	503	nf_unregister_hooks(ipv4_conntrack_ops, ARRAY_SIZE(ipv4_conntrack_ops));
32292a7f PM	504	nf_conntrack_l3proto_unregister(&nf_conntrack_l3proto_ipv4);
605dcad6 MJ	505	nf_conntrack_l4proto_unregister(&nf_conntrack_l4proto_icmp);
	506	nf_conntrack_l4proto_unregister(&nf_conntrack_l4proto_udp4);
	507	nf_conntrack_l4proto_unregister(&nf_conntrack_l4proto_tcp4);
32292a7f	508	nf_unregister_sockopt(&so_getorigdst);
9fb9cbb1 YK	509	}
9fb9cbb1 YK	510
65b4b4e8 AM	511	module_init(nf_conntrack_l3proto_ipv4_init);
65b4b4e8 AM	512	module_exit(nf_conntrack_l3proto_ipv4_fini);