[mirror_ubuntu-jammy-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.
 *
 * 16 Dec 2003: Yasuyuki Kozakai @USAGI <yasuyuki.kozakai@toshiba.co.jp>
 *	- move L3 protocol dependent part to this file.
 * 23 Mar 2004: Yasuyuki Kozakai @USAGI <yasuyuki.kozakai@toshiba.co.jp>
 *	- add get_features() to support various size of conntrack
 *	  structures.
 *
 * Derived from net/ipv4/netfilter/ip_conntrack_standalone.c
 */

#include <linux/config.h>
#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_protocol.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

DECLARE_PER_CPU(struct nf_conntrack_stat, nf_conntrack_stat);

static int ipv4_pkt_to_tuple(const struct sk_buff *skb, unsigned int nhoff,
			     struct nf_conntrack_tuple *tuple)
{
	u_int32_t _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(skb->nh.iph);

        return skb;
}

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

	*dataoff = (*pskb)->nh.raw - (*pskb)->data + (*pskb)->nh.iph->ihl*4;
	*protonum = (*pskb)->nh.iph->protocol;

	return NF_ACCEPT;
}

int nat_module_is_loaded = 0;
static u_int32_t ipv4_get_features(const struct nf_conntrack_tuple *tuple)
{
	if (nat_module_is_loaded)
		return NF_CT_F_NAT;

	return NF_CT_F_BASIC;
}

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;

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

	help = nfct_help(ct);
	if (!help || !help->helper)
		return NF_ACCEPT;

	return help->helper->help(pskb,
			       (*pskb)->nh.raw - (*pskb)->data
					       + (*pskb)->nh.iph->ihl*4,
			       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 *))
{
#if !defined(CONFIG_IP_NF_NAT) && !defined(CONFIG_IP_NF_NAT_MODULE)
	/* Previously seen (loopback)?  Ignore.  Do this before
	   fragment check. */
	if ((*pskb)->nfct)
		return NF_ACCEPT;
#endif

	/* Gather fragments. */
	if ((*pskb)->nh.iph->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)
	    || (*pskb)->nh.iph->ihl * 4 < 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,
	},
};

#ifdef CONFIG_SYSCTL
/* From nf_conntrack_proto_icmp.c */
extern unsigned int nf_ct_icmp_timeout;
static struct ctl_table_header *nf_ct_ipv4_sysctl_header;

static ctl_table nf_ct_sysctl_table[] = {
	{
		.ctl_name	= NET_NF_CONNTRACK_ICMP_TIMEOUT,
		.procname	= "nf_conntrack_icmp_timeout",
		.data		= &nf_ct_icmp_timeout,
		.maxlen		= sizeof(unsigned int),
		.mode		= 0644,
		.proc_handler	= &proc_dointvec_jiffies,
	},
        { .ctl_name = 0 }
};

static ctl_table nf_ct_netfilter_table[] = {
	{
		.ctl_name       = NET_NETFILTER,
		.procname       = "netfilter",
		.mode           = 0555,
		.child          = nf_ct_sysctl_table,
	},
	{ .ctl_name = 0 }
};

static ctl_table nf_ct_net_table[] = {
	{
		.ctl_name       = CTL_NET,
		.procname       = "net",
		.mode           = 0555,
		.child          = nf_ct_netfilter_table,
	},
	{ .ctl_name = 0 }
};
#endif

/* 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;

		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 =
		*(u_int32_t *)NFA_DATA(tb[CTA_IP_V4_SRC-1]);
	t->dst.u3.ip =
		*(u_int32_t *)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,
	.get_features	 = ipv4_get_features,
#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
	.me		 = THIS_MODULE,
};

extern struct nf_conntrack_protocol nf_conntrack_protocol_tcp4;
extern struct nf_conntrack_protocol nf_conntrack_protocol_udp4;
extern struct nf_conntrack_protocol nf_conntrack_protocol_icmp;
static int init_or_cleanup(int init)
{
	int ret = 0;

	if (!init) goto cleanup;

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

	ret = nf_conntrack_protocol_register(&nf_conntrack_protocol_tcp4);
	if (ret < 0) {
		printk("nf_conntrack_ipv4: can't register tcp.\n");
		goto cleanup_sockopt;
	}

	ret = nf_conntrack_protocol_register(&nf_conntrack_protocol_udp4);
	if (ret < 0) {
		printk("nf_conntrack_ipv4: can't register udp.\n");
		goto cleanup_tcp;
	}

	ret = nf_conntrack_protocol_register(&nf_conntrack_protocol_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;
	}
#ifdef CONFIG_SYSCTL
	nf_ct_ipv4_sysctl_header = register_sysctl_table(nf_ct_net_table, 0);
	if (nf_ct_ipv4_sysctl_header == NULL) {
		printk("nf_conntrack: can't register to sysctl.\n");
		ret = -ENOMEM;
		goto cleanup_hooks;
	}
#endif
	return ret;

 cleanup:
	synchronize_net();
#ifdef CONFIG_SYSCTL
 	unregister_sysctl_table(nf_ct_ipv4_sysctl_header);
 cleanup_hooks:
#endif
	nf_unregister_hooks(ipv4_conntrack_ops, ARRAY_SIZE(ipv4_conntrack_ops));
 cleanup_ipv4:
	nf_conntrack_l3proto_unregister(&nf_conntrack_l3proto_ipv4);
 cleanup_icmp:
	nf_conntrack_protocol_unregister(&nf_conntrack_protocol_icmp);
 cleanup_udp:
	nf_conntrack_protocol_unregister(&nf_conntrack_protocol_udp4);
 cleanup_tcp:
	nf_conntrack_protocol_unregister(&nf_conntrack_protocol_tcp4);
 cleanup_sockopt:
	nf_unregister_sockopt(&so_getorigdst);
 cleanup_nothing:
	return ret;
}

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

static int __init nf_conntrack_l3proto_ipv4_init(void)
{
	need_conntrack();
	return init_or_cleanup(1);
}

static void __exit nf_conntrack_l3proto_ipv4_fini(void)
{
	init_or_cleanup(0);
}

module_init(nf_conntrack_l3proto_ipv4_init);
module_exit(nf_conntrack_l3proto_ipv4_fini);

EXPORT_SYMBOL(nf_ct_ipv4_gather_frags);
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.
	7	*
	8	* 16 Dec 2003: Yasuyuki Kozakai @USAGI <yasuyuki.kozakai@toshiba.co.jp>
	9	* - move L3 protocol dependent part to this file.
	10	* 23 Mar 2004: Yasuyuki Kozakai @USAGI <yasuyuki.kozakai@toshiba.co.jp>
	11	* - add get_features() to support various size of conntrack
	12	* structures.
	13	*
	14	* Derived from net/ipv4/netfilter/ip_conntrack_standalone.c
	15	*/
	16
	17	#include <linux/config.h>
	18	#include <linux/types.h>
	19	#include <linux/ip.h>
	20	#include <linux/netfilter.h>
	21	#include <linux/module.h>
	22	#include <linux/skbuff.h>
	23	#include <linux/icmp.h>
	24	#include <linux/sysctl.h>
0ae2cfe7	25	#include <net/route.h>
9fb9cbb1 YK	26	#include <net/ip.h>
	27
	28	#include <linux/netfilter_ipv4.h>
	29	#include <net/netfilter/nf_conntrack.h>
	30	#include <net/netfilter/nf_conntrack_helper.h>
	31	#include <net/netfilter/nf_conntrack_protocol.h>
	32	#include <net/netfilter/nf_conntrack_l3proto.h>
	33	#include <net/netfilter/nf_conntrack_core.h>
	34	#include <net/netfilter/ipv4/nf_conntrack_ipv4.h>
	35
	36	#if 0
	37	#define DEBUGP printk
	38	#else
	39	#define DEBUGP(format, args...)
	40	#endif
	41
	42	DECLARE_PER_CPU(struct nf_conntrack_stat, nf_conntrack_stat);
	43
	44	static int ipv4_pkt_to_tuple(const struct sk_buff *skb, unsigned int nhoff,
	45	struct nf_conntrack_tuple *tuple)
	46	{
	47	u_int32_t _addrs[2], *ap;
	48	ap = skb_header_pointer(skb, nhoff + offsetof(struct iphdr, saddr),
	49	sizeof(u_int32_t) * 2, _addrs);
	50	if (ap == NULL)
	51	return 0;
	52
	53	tuple->src.u3.ip = ap[0];
	54	tuple->dst.u3.ip = ap[1];
	55
	56	return 1;
	57	}
	58
	59	static int ipv4_invert_tuple(struct nf_conntrack_tuple *tuple,
	60	const struct nf_conntrack_tuple *orig)
	61	{
	62	tuple->src.u3.ip = orig->dst.u3.ip;
	63	tuple->dst.u3.ip = orig->src.u3.ip;
	64
	65	return 1;
	66	}
	67
	68	static int ipv4_print_tuple(struct seq_file *s,
	69	const struct nf_conntrack_tuple *tuple)
	70	{
	71	return seq_printf(s, "src=%u.%u.%u.%u dst=%u.%u.%u.%u ",
	72	NIPQUAD(tuple->src.u3.ip),
	73	NIPQUAD(tuple->dst.u3.ip));
	74	}
	75
	76	static int ipv4_print_conntrack(struct seq_file *s,
	77	const struct nf_conn *conntrack)
	78	{
	79	return 0;
	80	}
	81
	82	/* Returns new sk_buff, or NULL */
	83	static struct sk_buff *
	84	nf_ct_ipv4_gather_frags(struct sk_buff *skb, u_int32_t user)
	85	{
	86	skb_orphan(skb);
	87
	88	local_bh_disable();
	89	skb = ip_defrag(skb, user);
90	local_bh_enable();
91
92	if (skb)
93	ip_send_check(skb->nh.iph);
94
95	return skb;
96	}
97
98	static int
99	ipv4_prepare(struct sk_buff *pskb, unsigned int hooknum, unsigned int dataoff,
100	u_int8_t *protonum)
101	{
102	/* Never happen */
103	if ((*pskb)->nh.iph->frag_off & htons(IP_OFFSET)) {
104	if (net_ratelimit()) {
105	printk(KERN_ERR "ipv4_prepare: Frag of proto %u (hook=%u)\n",
106	(*pskb)->nh.iph->protocol, hooknum);
107	}
108	return -NF_DROP;
109	}
110
111	dataoff = (pskb)->nh.raw - (pskb)->data + (pskb)->nh.iph->ihl*4;
112	protonum = (pskb)->nh.iph->protocol;
113
114	return NF_ACCEPT;
115	}
116
117	int nat_module_is_loaded = 0;
118	static u_int32_t ipv4_get_features(const struct nf_conntrack_tuple *tuple)
119	{
120	if (nat_module_is_loaded)
121	return NF_CT_F_NAT;
122
123	return NF_CT_F_BASIC;
124	}
125
126	static unsigned int ipv4_confirm(unsigned int hooknum,
127	struct sk_buff **pskb,
128	const struct net_device *in,
129	const struct net_device *out,
130	int (okfn)(struct sk_buff ))
131	{
132	/* We've seen it coming out the other side: confirm it */
133	return nf_conntrack_confirm(pskb);
134	}
135
136	static unsigned int ipv4_conntrack_help(unsigned int hooknum,
137	struct sk_buff **pskb,
138	const struct net_device *in,
139	const struct net_device *out,
140	int (okfn)(struct sk_buff ))
141	{
142	struct nf_conn *ct;
143	enum ip_conntrack_info ctinfo;
dc808fe2	144	struct nf_conn_help *help;
9fb9cbb1 YK	145
	146	/* This is where we call the helper: as the packet goes out. */
	147	ct = nf_ct_get(*pskb, &ctinfo);
dc808fe2 HW	148	if (!ct)
	149	return NF_ACCEPT;
	150
	151	help = nfct_help(ct);
	152	if (!help \|\| !help->helper)
	153	return NF_ACCEPT;
	154
	155	return help->helper->help(pskb,
	156	(pskb)->nh.raw - (pskb)->data
	157	+ (pskb)->nh.iph->ihl4,
	158	ct, ctinfo);
9fb9cbb1 YK	159	}
	160
	161	static unsigned int ipv4_conntrack_defrag(unsigned int hooknum,
	162	struct sk_buff **pskb,
	163	const struct net_device *in,
	164	const struct net_device *out,
	165	int (okfn)(struct sk_buff ))
	166	{
	167	#if !defined(CONFIG_IP_NF_NAT) && !defined(CONFIG_IP_NF_NAT_MODULE)
	168	/* Previously seen (loopback)? Ignore. Do this before
	169	fragment check. */
	170	if ((*pskb)->nfct)
	171	return NF_ACCEPT;
	172	#endif
	173
	174	/* Gather fragments. */
	175	if ((*pskb)->nh.iph->frag_off & htons(IP_MF\|IP_OFFSET)) {
	176	pskb = nf_ct_ipv4_gather_frags(pskb,
	177	hooknum == NF_IP_PRE_ROUTING ?
	178	IP_DEFRAG_CONNTRACK_IN :
	179	IP_DEFRAG_CONNTRACK_OUT);
	180	if (!*pskb)
	181	return NF_STOLEN;
	182	}
	183	return NF_ACCEPT;
	184	}
	185
9fb9cbb1 YK	186	static unsigned int ipv4_conntrack_in(unsigned int hooknum,
	187	struct sk_buff **pskb,
	188	const struct net_device *in,
	189	const struct net_device *out,
	190	int (okfn)(struct sk_buff ))
	191	{
	192	return nf_conntrack_in(PF_INET, hooknum, pskb);
	193	}
	194
	195	static unsigned int ipv4_conntrack_local(unsigned int hooknum,
	196	struct sk_buff **pskb,
	197	const struct net_device *in,
	198	const struct net_device *out,
	199	int (okfn)(struct sk_buff ))
	200	{
	201	/* root is playing with raw sockets. */
	202	if ((*pskb)->len < sizeof(struct iphdr)
	203	\|\| (pskb)->nh.iph->ihl 4 < sizeof(struct iphdr)) {
	204	if (net_ratelimit())
	205	printk("ipt_hook: happy cracking.\n");
	206	return NF_ACCEPT;
	207	}
	208	return nf_conntrack_in(PF_INET, hooknum, pskb);
	209	}
	210
	211	/* Connection tracking may drop packets, but never alters them, so
	212	make it the first hook. */
964ddaa1 PM	213	static struct nf_hook_ops ipv4_conntrack_ops[] = {
	214	{
	215	.hook = ipv4_conntrack_defrag,
	216	.owner = THIS_MODULE,
	217	.pf = PF_INET,
	218	.hooknum = NF_IP_PRE_ROUTING,
	219	.priority = NF_IP_PRI_CONNTRACK_DEFRAG,
	220	},
	221	{
	222	.hook = ipv4_conntrack_in,
	223	.owner = THIS_MODULE,
	224	.pf = PF_INET,
	225	.hooknum = NF_IP_PRE_ROUTING,
	226	.priority = NF_IP_PRI_CONNTRACK,
	227	},
	228	{
	229	.hook = ipv4_conntrack_defrag,
	230	.owner = THIS_MODULE,
	231	.pf = PF_INET,
	232	.hooknum = NF_IP_LOCAL_OUT,
	233	.priority = NF_IP_PRI_CONNTRACK_DEFRAG,
	234	},
	235	{
	236	.hook = ipv4_conntrack_local,
	237	.owner = THIS_MODULE,
	238	.pf = PF_INET,
	239	.hooknum = NF_IP_LOCAL_OUT,
	240	.priority = NF_IP_PRI_CONNTRACK,
	241	},
	242	{
	243	.hook = ipv4_conntrack_help,
	244	.owner = THIS_MODULE,
	245	.pf = PF_INET,
	246	.hooknum = NF_IP_POST_ROUTING,
	247	.priority = NF_IP_PRI_CONNTRACK_HELPER,
	248	},
	249	{
	250	.hook = ipv4_conntrack_help,
	251	.owner = THIS_MODULE,
	252	.pf = PF_INET,
	253	.hooknum = NF_IP_LOCAL_IN,
	254	.priority = NF_IP_PRI_CONNTRACK_HELPER,
	255	},
	256	{
	257	.hook = ipv4_confirm,
	258	.owner = THIS_MODULE,
	259	.pf = PF_INET,
	260	.hooknum = NF_IP_POST_ROUTING,
	261	.priority = NF_IP_PRI_CONNTRACK_CONFIRM,
	262	},
	263	{
	264	.hook = ipv4_confirm,
	265	.owner = THIS_MODULE,
	266	.pf = PF_INET,
	267	.hooknum = NF_IP_LOCAL_IN,
	268	.priority = NF_IP_PRI_CONNTRACK_CONFIRM,
	269	},
9fb9cbb1 YK	270	};
	271
	272	#ifdef CONFIG_SYSCTL
	273	/* From nf_conntrack_proto_icmp.c */
babbdb1a	274	extern unsigned int nf_ct_icmp_timeout;
9fb9cbb1 YK	275	static struct ctl_table_header *nf_ct_ipv4_sysctl_header;
	276
	277	static ctl_table nf_ct_sysctl_table[] = {
	278	{
	279	.ctl_name = NET_NF_CONNTRACK_ICMP_TIMEOUT,
	280	.procname = "nf_conntrack_icmp_timeout",
	281	.data = &nf_ct_icmp_timeout,
	282	.maxlen = sizeof(unsigned int),
	283	.mode = 0644,
	284	.proc_handler = &proc_dointvec_jiffies,
	285	},
	286	{ .ctl_name = 0 }
	287	};
	288
	289	static ctl_table nf_ct_netfilter_table[] = {
	290	{
	291	.ctl_name = NET_NETFILTER,
	292	.procname = "netfilter",
	293	.mode = 0555,
	294	.child = nf_ct_sysctl_table,
	295	},
	296	{ .ctl_name = 0 }
	297	};
	298
	299	static ctl_table nf_ct_net_table[] = {
	300	{
	301	.ctl_name = CTL_NET,
	302	.procname = "net",
	303	.mode = 0555,
	304	.child = nf_ct_netfilter_table,
	305	},
	306	{ .ctl_name = 0 }
	307	};
	308	#endif
	309
	310	/* Fast function for those who don't want to parse /proc (and I don't
	311	blame them). */
	312	/* Reversing the socket's dst/src point of view gives us the reply
	313	mapping. */
	314	static int
	315	getorigdst(struct sock sk, int optval, void __user user, int *len)
	316	{
	317	struct inet_sock *inet = inet_sk(sk);
	318	struct nf_conntrack_tuple_hash *h;
	319	struct nf_conntrack_tuple tuple;
	320
	321	NF_CT_TUPLE_U_BLANK(&tuple);
	322	tuple.src.u3.ip = inet->rcv_saddr;
	323	tuple.src.u.tcp.port = inet->sport;
	324	tuple.dst.u3.ip = inet->daddr;
	325	tuple.dst.u.tcp.port = inet->dport;
	326	tuple.src.l3num = PF_INET;
	327	tuple.dst.protonum = IPPROTO_TCP;
	328
	329	/* We only do TCP at the moment: is there a better way? */
	330	if (strcmp(sk->sk_prot->name, "TCP")) {
	331	DEBUGP("SO_ORIGINAL_DST: Not a TCP socket\n");
	332	return -ENOPROTOOPT;
	333	}
	334
	335	if ((unsigned int) *len < sizeof(struct sockaddr_in)) {
	336	DEBUGP("SO_ORIGINAL_DST: len %u not %u\n",
	337	*len, sizeof(struct sockaddr_in));
	338	return -EINVAL;
339	}
340
341	h = nf_conntrack_find_get(&tuple, NULL);
342	if (h) {
343	struct sockaddr_in sin;
344	struct nf_conn *ct = nf_ct_tuplehash_to_ctrack(h);
345
346	sin.sin_family = AF_INET;
347	sin.sin_port = ct->tuplehash[IP_CT_DIR_ORIGINAL]
348	.tuple.dst.u.tcp.port;
349	sin.sin_addr.s_addr = ct->tuplehash[IP_CT_DIR_ORIGINAL]
350	.tuple.dst.u3.ip;
351
352	DEBUGP("SO_ORIGINAL_DST: %u.%u.%u.%u %u\n",
353	NIPQUAD(sin.sin_addr.s_addr), ntohs(sin.sin_port));
354	nf_ct_put(ct);
355	if (copy_to_user(user, &sin, sizeof(sin)) != 0)
356	return -EFAULT;
357	else
358	return 0;
359	}
360	DEBUGP("SO_ORIGINAL_DST: Can't find %u.%u.%u.%u/%u-%u.%u.%u.%u/%u.\n",
361	NIPQUAD(tuple.src.u3.ip), ntohs(tuple.src.u.tcp.port),
362	NIPQUAD(tuple.dst.u3.ip), ntohs(tuple.dst.u.tcp.port));
363	return -ENOENT;
364	}
365
c1d10adb PNA	366	#if defined(CONFIG_NF_CT_NETLINK) \|\| \
	367	defined(CONFIG_NF_CT_NETLINK_MODULE)
	368
	369	#include <linux/netfilter/nfnetlink.h>
	370	#include <linux/netfilter/nfnetlink_conntrack.h>
	371
	372	static int ipv4_tuple_to_nfattr(struct sk_buff *skb,
	373	const struct nf_conntrack_tuple *tuple)
	374	{
	375	NFA_PUT(skb, CTA_IP_V4_SRC, sizeof(u_int32_t),
	376	&tuple->src.u3.ip);
	377	NFA_PUT(skb, CTA_IP_V4_DST, sizeof(u_int32_t),
	378	&tuple->dst.u3.ip);
	379	return 0;
	380
	381	nfattr_failure:
	382	return -1;
	383	}
	384
	385	static const size_t cta_min_ip[CTA_IP_MAX] = {
	386	[CTA_IP_V4_SRC-1] = sizeof(u_int32_t),
	387	[CTA_IP_V4_DST-1] = sizeof(u_int32_t),
	388	};
	389
	390	static int ipv4_nfattr_to_tuple(struct nfattr *tb[],
	391	struct nf_conntrack_tuple *t)
	392	{
	393	if (!tb[CTA_IP_V4_SRC-1] \|\| !tb[CTA_IP_V4_DST-1])
	394	return -EINVAL;
	395
	396	if (nfattr_bad_size(tb, CTA_IP_MAX, cta_min_ip))
	397	return -EINVAL;
	398
	399	t->src.u3.ip =
	400	(u_int32_t )NFA_DATA(tb[CTA_IP_V4_SRC-1]);
	401	t->dst.u3.ip =
	402	(u_int32_t )NFA_DATA(tb[CTA_IP_V4_DST-1]);
	403
	404	return 0;
	405	}
	406	#endif
	407
9fb9cbb1 YK	408	static struct nf_sockopt_ops so_getorigdst = {
	409	.pf = PF_INET,
	410	.get_optmin = SO_ORIGINAL_DST,
	411	.get_optmax = SO_ORIGINAL_DST+1,
	412	.get = &getorigdst,
	413	};
	414
	415	struct nf_conntrack_l3proto nf_conntrack_l3proto_ipv4 = {
	416	.l3proto = PF_INET,
	417	.name = "ipv4",
	418	.pkt_to_tuple = ipv4_pkt_to_tuple,
	419	.invert_tuple = ipv4_invert_tuple,
	420	.print_tuple = ipv4_print_tuple,
	421	.print_conntrack = ipv4_print_conntrack,
	422	.prepare = ipv4_prepare,
	423	.get_features = ipv4_get_features,
c1d10adb PNA	424	#if defined(CONFIG_NF_CT_NETLINK) \|\| \
	425	defined(CONFIG_NF_CT_NETLINK_MODULE)
	426	.tuple_to_nfattr = ipv4_tuple_to_nfattr,
	427	.nfattr_to_tuple = ipv4_nfattr_to_tuple,
	428	#endif
9fb9cbb1 YK	429	.me = THIS_MODULE,
	430	};
	431
	432	extern struct nf_conntrack_protocol nf_conntrack_protocol_tcp4;
	433	extern struct nf_conntrack_protocol nf_conntrack_protocol_udp4;
	434	extern struct nf_conntrack_protocol nf_conntrack_protocol_icmp;
	435	static int init_or_cleanup(int init)
	436	{
	437	int ret = 0;
	438
	439	if (!init) goto cleanup;
	440
	441	ret = nf_register_sockopt(&so_getorigdst);
	442	if (ret < 0) {
	443	printk(KERN_ERR "Unable to register netfilter socket option\n");
	444	goto cleanup_nothing;
	445	}
	446
	447	ret = nf_conntrack_protocol_register(&nf_conntrack_protocol_tcp4);
	448	if (ret < 0) {
	449	printk("nf_conntrack_ipv4: can't register tcp.\n");
	450	goto cleanup_sockopt;
	451	}
	452
	453	ret = nf_conntrack_protocol_register(&nf_conntrack_protocol_udp4);
	454	if (ret < 0) {
	455	printk("nf_conntrack_ipv4: can't register udp.\n");
	456	goto cleanup_tcp;
	457	}
	458
	459	ret = nf_conntrack_protocol_register(&nf_conntrack_protocol_icmp);
	460	if (ret < 0) {
	461	printk("nf_conntrack_ipv4: can't register icmp.\n");
	462	goto cleanup_udp;
	463	}
	464
	465	ret = nf_conntrack_l3proto_register(&nf_conntrack_l3proto_ipv4);
	466	if (ret < 0) {
	467	printk("nf_conntrack_ipv4: can't register ipv4\n");
	468	goto cleanup_icmp;
	469	}
	470
964ddaa1 PM	471	ret = nf_register_hooks(ipv4_conntrack_ops,
964ddaa1 PM	472	ARRAY_SIZE(ipv4_conntrack_ops));
9fb9cbb1	473	if (ret < 0) {
964ddaa1	474	printk("nf_conntrack_ipv4: can't register hooks.\n");
9fb9cbb1 YK	475	goto cleanup_ipv4;
9fb9cbb1 YK	476	}
9fb9cbb1 YK	477	#ifdef CONFIG_SYSCTL
	478	nf_ct_ipv4_sysctl_header = register_sysctl_table(nf_ct_net_table, 0);
	479	if (nf_ct_ipv4_sysctl_header == NULL) {
	480	printk("nf_conntrack: can't register to sysctl.\n");
	481	ret = -ENOMEM;
964ddaa1	482	goto cleanup_hooks;
9fb9cbb1 YK	483	}
9fb9cbb1 YK	484	#endif
9fb9cbb1 YK	485	return ret;
	486
	487	cleanup:
	488	synchronize_net();
9fb9cbb1 YK	489	#ifdef CONFIG_SYSCTL
9fb9cbb1 YK	490	unregister_sysctl_table(nf_ct_ipv4_sysctl_header);
964ddaa1	491	cleanup_hooks:
9fb9cbb1	492	#endif
964ddaa1	493	nf_unregister_hooks(ipv4_conntrack_ops, ARRAY_SIZE(ipv4_conntrack_ops));
9fb9cbb1 YK	494	cleanup_ipv4:
	495	nf_conntrack_l3proto_unregister(&nf_conntrack_l3proto_ipv4);
	496	cleanup_icmp:
	497	nf_conntrack_protocol_unregister(&nf_conntrack_protocol_icmp);
	498	cleanup_udp:
	499	nf_conntrack_protocol_unregister(&nf_conntrack_protocol_udp4);
	500	cleanup_tcp:
	501	nf_conntrack_protocol_unregister(&nf_conntrack_protocol_tcp4);
	502	cleanup_sockopt:
	503	nf_unregister_sockopt(&so_getorigdst);
	504	cleanup_nothing:
	505	return ret;
	506	}
	507
b9f78f9f	508	MODULE_ALIAS("nf_conntrack-" __stringify(AF_INET));
9fb9cbb1 YK	509	MODULE_LICENSE("GPL");
9fb9cbb1 YK	510
65b4b4e8	511	static int __init nf_conntrack_l3proto_ipv4_init(void)
9fb9cbb1	512	{
2e4e6a17	513	need_conntrack();
9fb9cbb1 YK	514	return init_or_cleanup(1);
	515	}
	516
65b4b4e8	517	static void __exit nf_conntrack_l3proto_ipv4_fini(void)
9fb9cbb1 YK	518	{
	519	init_or_cleanup(0);
	520	}
	521
65b4b4e8 AM	522	module_init(nf_conntrack_l3proto_ipv4_init);
65b4b4e8 AM	523	module_exit(nf_conntrack_l3proto_ipv4_fini);
9fb9cbb1	524
9fb9cbb1	525	EXPORT_SYMBOL(nf_ct_ipv4_gather_frags);