[deliverable/linux.git] / net / netfilter / xt_TPROXY.c

/*
 * Transparent proxy support for Linux/iptables
 *
 * Copyright (c) 2006-2010 BalaBit IT Ltd.
 * Author: Balazs Scheidler, Krisztian Kovacs
 *
 * 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.
 *
 */
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include <linux/module.h>
#include <linux/skbuff.h>
#include <linux/ip.h>
#include <net/checksum.h>
#include <net/udp.h>
#include <net/tcp.h>
#include <net/inet_sock.h>
#include <net/inet_hashtables.h>
#include <linux/inetdevice.h>
#include <linux/netfilter/x_tables.h>
#include <linux/netfilter_ipv4/ip_tables.h>

#include <net/netfilter/ipv4/nf_defrag_ipv4.h>

#if IS_ENABLED(CONFIG_IP6_NF_IPTABLES)
#define XT_TPROXY_HAVE_IPV6 1
#include <net/if_inet6.h>
#include <net/addrconf.h>
#include <net/inet6_hashtables.h>
#include <linux/netfilter_ipv6/ip6_tables.h>
#include <net/netfilter/ipv6/nf_defrag_ipv6.h>
#endif

#include <linux/netfilter/xt_TPROXY.h>

enum nf_tproxy_lookup_t {
	 NFT_LOOKUP_LISTENER,
	 NFT_LOOKUP_ESTABLISHED,
};

static bool tproxy_sk_is_transparent(struct sock *sk)
{
	switch (sk->sk_state) {
	case TCP_TIME_WAIT:
		if (inet_twsk(sk)->tw_transparent)
			return true;
		break;
	case TCP_NEW_SYN_RECV:
		if (inet_rsk(inet_reqsk(sk))->no_srccheck)
			return true;
		break;
	default:
		if (inet_sk(sk)->transparent)
			return true;
	}

	sock_gen_put(sk);
	return false;
}

static inline __be32
tproxy_laddr4(struct sk_buff *skb, __be32 user_laddr, __be32 daddr)
{
	struct in_device *indev;
	__be32 laddr;

	if (user_laddr)
		return user_laddr;

	laddr = 0;
	rcu_read_lock();
	indev = __in_dev_get_rcu(skb->dev);
	for_primary_ifa(indev) {
		laddr = ifa->ifa_local;
		break;
	} endfor_ifa(indev);
	rcu_read_unlock();

	return laddr ? laddr : daddr;
}

/*
 * This is used when the user wants to intercept a connection matching
 * an explicit iptables rule. In this case the sockets are assumed
 * matching in preference order:
 *
 *   - match: if there's a fully established connection matching the
 *     _packet_ tuple, it is returned, assuming the redirection
 *     already took place and we process a packet belonging to an
 *     established connection
 *
 *   - match: if there's a listening socket matching the redirection
 *     (e.g. on-port & on-ip of the connection), it is returned,
 *     regardless if it was bound to 0.0.0.0 or an explicit
 *     address. The reasoning is that if there's an explicit rule, it
 *     does not really matter if the listener is bound to an interface
 *     or to 0. The user already stated that he wants redirection
 *     (since he added the rule).
 *
 * Please note that there's an overlap between what a TPROXY target
 * and a socket match will match. Normally if you have both rules the
 * "socket" match will be the first one, effectively all packets
 * belonging to established connections going through that one.
 */
static inline struct sock *
nf_tproxy_get_sock_v4(struct net *net, const u8 protocol,
		      const __be32 saddr, const __be32 daddr,
		      const __be16 sport, const __be16 dport,
		      const struct net_device *in,
		      const enum nf_tproxy_lookup_t lookup_type)
{
	struct sock *sk;

	switch (protocol) {
	case IPPROTO_TCP:
		switch (lookup_type) {
		case NFT_LOOKUP_LISTENER:
			sk = inet_lookup_listener(net, &tcp_hashinfo,
						    saddr, sport,
						    daddr, dport,
						    in->ifindex);

			/* NOTE: we return listeners even if bound to
			 * 0.0.0.0, those are filtered out in
			 * xt_socket, since xt_TPROXY needs 0 bound
			 * listeners too
			 */
			break;
		case NFT_LOOKUP_ESTABLISHED:
			sk = inet_lookup_established(net, &tcp_hashinfo,
						    saddr, sport, daddr, dport,
						    in->ifindex);
			break;
		default:
			BUG();
		}
		break;
	case IPPROTO_UDP:
		sk = udp4_lib_lookup(net, saddr, sport, daddr, dport,
				     in->ifindex);
		if (sk) {
			int connected = (sk->sk_state == TCP_ESTABLISHED);
			int wildcard = (inet_sk(sk)->inet_rcv_saddr == 0);

			/* NOTE: we return listeners even if bound to
			 * 0.0.0.0, those are filtered out in
			 * xt_socket, since xt_TPROXY needs 0 bound
			 * listeners too
			 */
			if ((lookup_type == NFT_LOOKUP_ESTABLISHED && (!connected || wildcard)) ||
			    (lookup_type == NFT_LOOKUP_LISTENER && connected)) {
				sock_put(sk);
				sk = NULL;
			}
		}
		break;
	default:
		WARN_ON(1);
		sk = NULL;
	}

	pr_debug("tproxy socket lookup: proto %u %08x:%u -> %08x:%u, lookup type: %d, sock %p\n",
		 protocol, ntohl(saddr), ntohs(sport), ntohl(daddr), ntohs(dport), lookup_type, sk);

	return sk;
}

#ifdef XT_TPROXY_HAVE_IPV6
static inline struct sock *
nf_tproxy_get_sock_v6(struct net *net, const u8 protocol,
		      const struct in6_addr *saddr, const struct in6_addr *daddr,
		      const __be16 sport, const __be16 dport,
		      const struct net_device *in,
		      const enum nf_tproxy_lookup_t lookup_type)
{
	struct sock *sk;

	switch (protocol) {
	case IPPROTO_TCP:
		switch (lookup_type) {
		case NFT_LOOKUP_LISTENER:
			sk = inet6_lookup_listener(net, &tcp_hashinfo,
						   saddr, sport,
						   daddr, ntohs(dport),
						   in->ifindex);

			/* NOTE: we return listeners even if bound to
			 * 0.0.0.0, those are filtered out in
			 * xt_socket, since xt_TPROXY needs 0 bound
			 * listeners too
			 */
			break;
		case NFT_LOOKUP_ESTABLISHED:
			sk = __inet6_lookup_established(net, &tcp_hashinfo,
							saddr, sport, daddr, ntohs(dport),
							in->ifindex);
			break;
		default:
			BUG();
		}
		break;
	case IPPROTO_UDP:
		sk = udp6_lib_lookup(net, saddr, sport, daddr, dport,
				     in->ifindex);
		if (sk) {
			int connected = (sk->sk_state == TCP_ESTABLISHED);
			int wildcard = ipv6_addr_any(&sk->sk_v6_rcv_saddr);

			/* NOTE: we return listeners even if bound to
			 * 0.0.0.0, those are filtered out in
			 * xt_socket, since xt_TPROXY needs 0 bound
			 * listeners too
			 */
			if ((lookup_type == NFT_LOOKUP_ESTABLISHED && (!connected || wildcard)) ||
			    (lookup_type == NFT_LOOKUP_LISTENER && connected)) {
				sock_put(sk);
				sk = NULL;
			}
		}
		break;
	default:
		WARN_ON(1);
		sk = NULL;
	}

	pr_debug("tproxy socket lookup: proto %u %pI6:%u -> %pI6:%u, lookup type: %d, sock %p\n",
		 protocol, saddr, ntohs(sport), daddr, ntohs(dport), lookup_type, sk);

	return sk;
}
#endif

/**
 * tproxy_handle_time_wait4 - handle IPv4 TCP TIME_WAIT reopen redirections
 * @skb:	The skb being processed.
 * @laddr:	IPv4 address to redirect to or zero.
 * @lport:	TCP port to redirect to or zero.
 * @sk:		The TIME_WAIT TCP socket found by the lookup.
 *
 * We have to handle SYN packets arriving to TIME_WAIT sockets
 * differently: instead of reopening the connection we should rather
 * redirect the new connection to the proxy if there's a listener
 * socket present.
 *
 * tproxy_handle_time_wait4() consumes the socket reference passed in.
 *
 * Returns the listener socket if there's one, the TIME_WAIT socket if
 * no such listener is found, or NULL if the TCP header is incomplete.
 */
static struct sock *
tproxy_handle_time_wait4(struct net *net, struct sk_buff *skb,
			 __be32 laddr, __be16 lport, struct sock *sk)
{
	const struct iphdr *iph = ip_hdr(skb);
	struct tcphdr _hdr, *hp;

	hp = skb_header_pointer(skb, ip_hdrlen(skb), sizeof(_hdr), &_hdr);
	if (hp == NULL) {
		inet_twsk_put(inet_twsk(sk));
		return NULL;
	}

	if (hp->syn && !hp->rst && !hp->ack && !hp->fin) {
		/* SYN to a TIME_WAIT socket, we'd rather redirect it
		 * to a listener socket if there's one */
		struct sock *sk2;

		sk2 = nf_tproxy_get_sock_v4(net, iph->protocol,
					    iph->saddr, laddr ? laddr : iph->daddr,
					    hp->source, lport ? lport : hp->dest,
					    skb->dev, NFT_LOOKUP_LISTENER);
		if (sk2) {
			inet_twsk_deschedule_put(inet_twsk(sk));
			sk = sk2;
		}
	}

	return sk;
}

/* assign a socket to the skb -- consumes sk */
static void
nf_tproxy_assign_sock(struct sk_buff *skb, struct sock *sk)
{
	skb_orphan(skb);
	skb->sk = sk;
	skb->destructor = sock_edemux;
}

static unsigned int
tproxy_tg4(struct net *net, struct sk_buff *skb, __be32 laddr, __be16 lport,
	   u_int32_t mark_mask, u_int32_t mark_value)
{
	const struct iphdr *iph = ip_hdr(skb);
	struct udphdr _hdr, *hp;
	struct sock *sk;

	hp = skb_header_pointer(skb, ip_hdrlen(skb), sizeof(_hdr), &_hdr);
	if (hp == NULL)
		return NF_DROP;

	/* check if there's an ongoing connection on the packet
	 * addresses, this happens if the redirect already happened
	 * and the current packet belongs to an already established
	 * connection */
	sk = nf_tproxy_get_sock_v4(net, iph->protocol,
				   iph->saddr, iph->daddr,
				   hp->source, hp->dest,
				   skb->dev, NFT_LOOKUP_ESTABLISHED);

	laddr = tproxy_laddr4(skb, laddr, iph->daddr);
	if (!lport)
		lport = hp->dest;

	/* UDP has no TCP_TIME_WAIT state, so we never enter here */
	if (sk && sk->sk_state == TCP_TIME_WAIT)
		/* reopening a TIME_WAIT connection needs special handling */
		sk = tproxy_handle_time_wait4(net, skb, laddr, lport, sk);
	else if (!sk)
		/* no, there's no established connection, check if
		 * there's a listener on the redirected addr/port */
		sk = nf_tproxy_get_sock_v4(net, iph->protocol,
					   iph->saddr, laddr,
					   hp->source, lport,
					   skb->dev, NFT_LOOKUP_LISTENER);

	/* NOTE: assign_sock consumes our sk reference */
	if (sk && tproxy_sk_is_transparent(sk)) {
		/* This should be in a separate target, but we don't do multiple
		   targets on the same rule yet */
		skb->mark = (skb->mark & ~mark_mask) ^ mark_value;

		pr_debug("redirecting: proto %hhu %pI4:%hu -> %pI4:%hu, mark: %x\n",
			 iph->protocol, &iph->daddr, ntohs(hp->dest),
			 &laddr, ntohs(lport), skb->mark);

		nf_tproxy_assign_sock(skb, sk);
		return NF_ACCEPT;
	}

	pr_debug("no socket, dropping: proto %hhu %pI4:%hu -> %pI4:%hu, mark: %x\n",
		 iph->protocol, &iph->saddr, ntohs(hp->source),
		 &iph->daddr, ntohs(hp->dest), skb->mark);
	return NF_DROP;
}

static unsigned int
tproxy_tg4_v0(struct sk_buff *skb, const struct xt_action_param *par)
{
	const struct xt_tproxy_target_info *tgi = par->targinfo;

	return tproxy_tg4(par->net, skb, tgi->laddr, tgi->lport, tgi->mark_mask, tgi->mark_value);
}

static unsigned int
tproxy_tg4_v1(struct sk_buff *skb, const struct xt_action_param *par)
{
	const struct xt_tproxy_target_info_v1 *tgi = par->targinfo;

	return tproxy_tg4(par->net, skb, tgi->laddr.ip, tgi->lport, tgi->mark_mask, tgi->mark_value);
}

#ifdef XT_TPROXY_HAVE_IPV6

static inline const struct in6_addr *
tproxy_laddr6(struct sk_buff *skb, const struct in6_addr *user_laddr,
	      const struct in6_addr *daddr)
{
	struct inet6_dev *indev;
	struct inet6_ifaddr *ifa;
	struct in6_addr *laddr;

	if (!ipv6_addr_any(user_laddr))
		return user_laddr;
	laddr = NULL;

	rcu_read_lock();
	indev = __in6_dev_get(skb->dev);
	if (indev)
		list_for_each_entry(ifa, &indev->addr_list, if_list) {
			if (ifa->flags & (IFA_F_TENTATIVE | IFA_F_DEPRECATED))
				continue;

			laddr = &ifa->addr;
			break;
		}
	rcu_read_unlock();

	return laddr ? laddr : daddr;
}

/**
 * tproxy_handle_time_wait6 - handle IPv6 TCP TIME_WAIT reopen redirections
 * @skb:	The skb being processed.
 * @tproto:	Transport protocol.
 * @thoff:	Transport protocol header offset.
 * @par:	Iptables target parameters.
 * @sk:		The TIME_WAIT TCP socket found by the lookup.
 *
 * We have to handle SYN packets arriving to TIME_WAIT sockets
 * differently: instead of reopening the connection we should rather
 * redirect the new connection to the proxy if there's a listener
 * socket present.
 *
 * tproxy_handle_time_wait6() consumes the socket reference passed in.
 *
 * Returns the listener socket if there's one, the TIME_WAIT socket if
 * no such listener is found, or NULL if the TCP header is incomplete.
 */
static struct sock *
tproxy_handle_time_wait6(struct sk_buff *skb, int tproto, int thoff,
			 const struct xt_action_param *par,
			 struct sock *sk)
{
	const struct ipv6hdr *iph = ipv6_hdr(skb);
	struct tcphdr _hdr, *hp;
	const struct xt_tproxy_target_info_v1 *tgi = par->targinfo;

	hp = skb_header_pointer(skb, thoff, sizeof(_hdr), &_hdr);
	if (hp == NULL) {
		inet_twsk_put(inet_twsk(sk));
		return NULL;
	}

	if (hp->syn && !hp->rst && !hp->ack && !hp->fin) {
		/* SYN to a TIME_WAIT socket, we'd rather redirect it
		 * to a listener socket if there's one */
		struct sock *sk2;

		sk2 = nf_tproxy_get_sock_v6(par->net, tproto,
					    &iph->saddr,
					    tproxy_laddr6(skb, &tgi->laddr.in6, &iph->daddr),
					    hp->source,
					    tgi->lport ? tgi->lport : hp->dest,
					    skb->dev, NFT_LOOKUP_LISTENER);
		if (sk2) {
			inet_twsk_deschedule_put(inet_twsk(sk));
			sk = sk2;
		}
	}

	return sk;
}

static unsigned int
tproxy_tg6_v1(struct sk_buff *skb, const struct xt_action_param *par)
{
	const struct ipv6hdr *iph = ipv6_hdr(skb);
	const struct xt_tproxy_target_info_v1 *tgi = par->targinfo;
	struct udphdr _hdr, *hp;
	struct sock *sk;
	const struct in6_addr *laddr;
	__be16 lport;
	int thoff = 0;
	int tproto;

	tproto = ipv6_find_hdr(skb, &thoff, -1, NULL, NULL);
	if (tproto < 0) {
		pr_debug("unable to find transport header in IPv6 packet, dropping\n");
		return NF_DROP;
	}

	hp = skb_header_pointer(skb, thoff, sizeof(_hdr), &_hdr);
	if (hp == NULL) {
		pr_debug("unable to grab transport header contents in IPv6 packet, dropping\n");
		return NF_DROP;
	}

	/* check if there's an ongoing connection on the packet
	 * addresses, this happens if the redirect already happened
	 * and the current packet belongs to an already established
	 * connection */
	sk = nf_tproxy_get_sock_v6(par->net, tproto,
				   &iph->saddr, &iph->daddr,
				   hp->source, hp->dest,
				   par->in, NFT_LOOKUP_ESTABLISHED);

	laddr = tproxy_laddr6(skb, &tgi->laddr.in6, &iph->daddr);
	lport = tgi->lport ? tgi->lport : hp->dest;

	/* UDP has no TCP_TIME_WAIT state, so we never enter here */
	if (sk && sk->sk_state == TCP_TIME_WAIT)
		/* reopening a TIME_WAIT connection needs special handling */
		sk = tproxy_handle_time_wait6(skb, tproto, thoff, par, sk);
	else if (!sk)
		/* no there's no established connection, check if
		 * there's a listener on the redirected addr/port */
		sk = nf_tproxy_get_sock_v6(par->net, tproto,
					   &iph->saddr, laddr,
					   hp->source, lport,
					   par->in, NFT_LOOKUP_LISTENER);

	/* NOTE: assign_sock consumes our sk reference */
	if (sk && tproxy_sk_is_transparent(sk)) {
		/* This should be in a separate target, but we don't do multiple
		   targets on the same rule yet */
		skb->mark = (skb->mark & ~tgi->mark_mask) ^ tgi->mark_value;

		pr_debug("redirecting: proto %hhu %pI6:%hu -> %pI6:%hu, mark: %x\n",
			 tproto, &iph->saddr, ntohs(hp->source),
			 laddr, ntohs(lport), skb->mark);

		nf_tproxy_assign_sock(skb, sk);
		return NF_ACCEPT;
	}

	pr_debug("no socket, dropping: proto %hhu %pI6:%hu -> %pI6:%hu, mark: %x\n",
		 tproto, &iph->saddr, ntohs(hp->source),
		 &iph->daddr, ntohs(hp->dest), skb->mark);

	return NF_DROP;
}

static int tproxy_tg6_check(const struct xt_tgchk_param *par)
{
	const struct ip6t_ip6 *i = par->entryinfo;

	if ((i->proto == IPPROTO_TCP || i->proto == IPPROTO_UDP) &&
	    !(i->invflags & IP6T_INV_PROTO))
		return 0;

	pr_info("Can be used only in combination with "
		"either -p tcp or -p udp\n");
	return -EINVAL;
}
#endif

static int tproxy_tg4_check(const struct xt_tgchk_param *par)
{
	const struct ipt_ip *i = par->entryinfo;

	if ((i->proto == IPPROTO_TCP || i->proto == IPPROTO_UDP)
	    && !(i->invflags & IPT_INV_PROTO))
		return 0;

	pr_info("Can be used only in combination with "
		"either -p tcp or -p udp\n");
	return -EINVAL;
}

static struct xt_target tproxy_tg_reg[] __read_mostly = {
	{
		.name		= "TPROXY",
		.family		= NFPROTO_IPV4,
		.table		= "mangle",
		.target		= tproxy_tg4_v0,
		.revision	= 0,
		.targetsize	= sizeof(struct xt_tproxy_target_info),
		.checkentry	= tproxy_tg4_check,
		.hooks		= 1 << NF_INET_PRE_ROUTING,
		.me		= THIS_MODULE,
	},
	{
		.name		= "TPROXY",
		.family		= NFPROTO_IPV4,
		.table		= "mangle",
		.target		= tproxy_tg4_v1,
		.revision	= 1,
		.targetsize	= sizeof(struct xt_tproxy_target_info_v1),
		.checkentry	= tproxy_tg4_check,
		.hooks		= 1 << NF_INET_PRE_ROUTING,
		.me		= THIS_MODULE,
	},
#ifdef XT_TPROXY_HAVE_IPV6
	{
		.name		= "TPROXY",
		.family		= NFPROTO_IPV6,
		.table		= "mangle",
		.target		= tproxy_tg6_v1,
		.revision	= 1,
		.targetsize	= sizeof(struct xt_tproxy_target_info_v1),
		.checkentry	= tproxy_tg6_check,
		.hooks		= 1 << NF_INET_PRE_ROUTING,
		.me		= THIS_MODULE,
	},
#endif

};

static int __init tproxy_tg_init(void)
{
	nf_defrag_ipv4_enable();
#ifdef XT_TPROXY_HAVE_IPV6
	nf_defrag_ipv6_enable();
#endif

	return xt_register_targets(tproxy_tg_reg, ARRAY_SIZE(tproxy_tg_reg));
}

static void __exit tproxy_tg_exit(void)
{
	xt_unregister_targets(tproxy_tg_reg, ARRAY_SIZE(tproxy_tg_reg));
}

module_init(tproxy_tg_init);
module_exit(tproxy_tg_exit);
MODULE_LICENSE("GPL");
MODULE_AUTHOR("Balazs Scheidler, Krisztian Kovacs");
MODULE_DESCRIPTION("Netfilter transparent proxy (TPROXY) target module.");
MODULE_ALIAS("ipt_TPROXY");
MODULE_ALIAS("ip6t_TPROXY");
Commit	Line	Data
e8439270 KK	1	/*
	2	* Transparent proxy support for Linux/iptables
	3	*
6ad78893	4	* Copyright (c) 2006-2010 BalaBit IT Ltd.
e8439270 KK	5	* Author: Balazs Scheidler, Krisztian Kovacs
	6	*
	7	* This program is free software; you can redistribute it and/or modify
	8	* it under the terms of the GNU General Public License version 2 as
	9	* published by the Free Software Foundation.
	10	*
	11	*/
ff67e4e4	12	#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
e8439270 KK	13	#include <linux/module.h>
	14	#include <linux/skbuff.h>
	15	#include <linux/ip.h>
	16	#include <net/checksum.h>
	17	#include <net/udp.h>
93742cf8	18	#include <net/tcp.h>
e8439270	19	#include <net/inet_sock.h>
93742cf8	20	#include <net/inet_hashtables.h>
cc6eb433	21	#include <linux/inetdevice.h>
e8439270 KK	22	#include <linux/netfilter/x_tables.h>
e8439270 KK	23	#include <linux/netfilter_ipv4/ip_tables.h>
e8439270 KK	24
e8439270 KK	25	#include <net/netfilter/ipv4/nf_defrag_ipv4.h>
f6318e55	26
c0cd1156	27	#if IS_ENABLED(CONFIG_IP6_NF_IPTABLES)
f6318e55	28	#define XT_TPROXY_HAVE_IPV6 1
cc6eb433 BS	29	#include <net/if_inet6.h>
cc6eb433 BS	30	#include <net/addrconf.h>
93742cf8	31	#include <net/inet6_hashtables.h>
cc6eb433	32	#include <linux/netfilter_ipv6/ip6_tables.h>
6ad78893	33	#include <net/netfilter/ipv6/nf_defrag_ipv6.h>
cc6eb433 BS	34	#endif
cc6eb433 BS	35
cc6eb433 BS	36	#include <linux/netfilter/xt_TPROXY.h>
cc6eb433 BS	37
93742cf8 FW	38	enum nf_tproxy_lookup_t {
	39	NFT_LOOKUP_LISTENER,
	40	NFT_LOOKUP_ESTABLISHED,
	41	};
	42
d503b30b FW	43	static bool tproxy_sk_is_transparent(struct sock *sk)
d503b30b FW	44	{
8b580147 ED	45	switch (sk->sk_state) {
8b580147 ED	46	case TCP_TIME_WAIT:
d503b30b FW	47	if (inet_twsk(sk)->tw_transparent)
d503b30b FW	48	return true;
8b580147 ED	49	break;
	50	case TCP_NEW_SYN_RECV:
	51	if (inet_rsk(inet_reqsk(sk))->no_srccheck)
	52	return true;
	53	break;
	54	default:
	55	if (inet_sk(sk)->transparent)
	56	return true;
d503b30b	57	}
8b580147 ED	58
8b580147 ED	59	sock_gen_put(sk);
d503b30b FW	60	return false;
	61	}
	62
cc6eb433 BS	63	static inline __be32
	64	tproxy_laddr4(struct sk_buff *skb, __be32 user_laddr, __be32 daddr)
	65	{
	66	struct in_device *indev;
	67	__be32 laddr;
	68
	69	if (user_laddr)
	70	return user_laddr;
	71
	72	laddr = 0;
	73	rcu_read_lock();
	74	indev = __in_dev_get_rcu(skb->dev);
	75	for_primary_ifa(indev) {
	76	laddr = ifa->ifa_local;
	77	break;
	78	} endfor_ifa(indev);
	79	rcu_read_unlock();
	80
	81	return laddr ? laddr : daddr;
	82	}
e8439270	83
93742cf8 FW	84	/*
	85	* This is used when the user wants to intercept a connection matching
	86	* an explicit iptables rule. In this case the sockets are assumed
	87	* matching in preference order:
	88	*
	89	* - match: if there's a fully established connection matching the
	90	* _packet_ tuple, it is returned, assuming the redirection
	91	* already took place and we process a packet belonging to an
	92	* established connection
	93	*
	94	* - match: if there's a listening socket matching the redirection
	95	* (e.g. on-port & on-ip of the connection), it is returned,
	96	* regardless if it was bound to 0.0.0.0 or an explicit
	97	* address. The reasoning is that if there's an explicit rule, it
	98	* does not really matter if the listener is bound to an interface
	99	* or to 0. The user already stated that he wants redirection
	100	* (since he added the rule).
	101	*
	102	* Please note that there's an overlap between what a TPROXY target
	103	* and a socket match will match. Normally if you have both rules the
	104	* "socket" match will be the first one, effectively all packets
	105	* belonging to established connections going through that one.
	106	*/
	107	static inline struct sock *
	108	nf_tproxy_get_sock_v4(struct net *net, const u8 protocol,
	109	const __be32 saddr, const __be32 daddr,
	110	const __be16 sport, const __be16 dport,
	111	const struct net_device *in,
	112	const enum nf_tproxy_lookup_t lookup_type)
	113	{
	114	struct sock *sk;
	115
	116	switch (protocol) {
	117	case IPPROTO_TCP:
	118	switch (lookup_type) {
	119	case NFT_LOOKUP_LISTENER:
	120	sk = inet_lookup_listener(net, &tcp_hashinfo,
	121	saddr, sport,
	122	daddr, dport,
	123	in->ifindex);
	124
	125	/* NOTE: we return listeners even if bound to
	126	* 0.0.0.0, those are filtered out in
	127	* xt_socket, since xt_TPROXY needs 0 bound
	128	* listeners too
	129	*/
	130	break;
	131	case NFT_LOOKUP_ESTABLISHED:
	132	sk = inet_lookup_established(net, &tcp_hashinfo,
	133	saddr, sport, daddr, dport,
	134	in->ifindex);
	135	break;
	136	default:
	137	BUG();
	138	}
	139	break;
	140	case IPPROTO_UDP:
	141	sk = udp4_lib_lookup(net, saddr, sport, daddr, dport,
	142	in->ifindex);
	143	if (sk) {
	144	int connected = (sk->sk_state == TCP_ESTABLISHED);
	145	int wildcard = (inet_sk(sk)->inet_rcv_saddr == 0);
	146
	147	/* NOTE: we return listeners even if bound to
148	* 0.0.0.0, those are filtered out in
149	* xt_socket, since xt_TPROXY needs 0 bound
150	* listeners too
151	*/
152	if ((lookup_type == NFT_LOOKUP_ESTABLISHED && (!connected \|\| wildcard)) \|\|
153	(lookup_type == NFT_LOOKUP_LISTENER && connected)) {
154	sock_put(sk);
155	sk = NULL;
156	}
157	}
158	break;
159	default:
160	WARN_ON(1);
161	sk = NULL;
162	}
163
164	pr_debug("tproxy socket lookup: proto %u %08x:%u -> %08x:%u, lookup type: %d, sock %p\n",
165	protocol, ntohl(saddr), ntohs(sport), ntohl(daddr), ntohs(dport), lookup_type, sk);
166
167	return sk;
168	}
169
d8b3bfc2	170	#ifdef XT_TPROXY_HAVE_IPV6
93742cf8 FW	171	static inline struct sock *
	172	nf_tproxy_get_sock_v6(struct net *net, const u8 protocol,
	173	const struct in6_addr saddr, const struct in6_addr daddr,
	174	const __be16 sport, const __be16 dport,
	175	const struct net_device *in,
	176	const enum nf_tproxy_lookup_t lookup_type)
	177	{
	178	struct sock *sk;
	179
	180	switch (protocol) {
	181	case IPPROTO_TCP:
	182	switch (lookup_type) {
	183	case NFT_LOOKUP_LISTENER:
	184	sk = inet6_lookup_listener(net, &tcp_hashinfo,
	185	saddr, sport,
	186	daddr, ntohs(dport),
	187	in->ifindex);
	188
	189	/* NOTE: we return listeners even if bound to
	190	* 0.0.0.0, those are filtered out in
	191	* xt_socket, since xt_TPROXY needs 0 bound
	192	* listeners too
	193	*/
	194	break;
	195	case NFT_LOOKUP_ESTABLISHED:
	196	sk = __inet6_lookup_established(net, &tcp_hashinfo,
	197	saddr, sport, daddr, ntohs(dport),
	198	in->ifindex);
	199	break;
	200	default:
	201	BUG();
	202	}
	203	break;
	204	case IPPROTO_UDP:
	205	sk = udp6_lib_lookup(net, saddr, sport, daddr, dport,
	206	in->ifindex);
	207	if (sk) {
	208	int connected = (sk->sk_state == TCP_ESTABLISHED);
efe4208f	209	int wildcard = ipv6_addr_any(&sk->sk_v6_rcv_saddr);
93742cf8 FW	210
	211	/* NOTE: we return listeners even if bound to
	212	* 0.0.0.0, those are filtered out in
	213	* xt_socket, since xt_TPROXY needs 0 bound
	214	* listeners too
	215	*/
	216	if ((lookup_type == NFT_LOOKUP_ESTABLISHED && (!connected \|\| wildcard)) \|\|
	217	(lookup_type == NFT_LOOKUP_LISTENER && connected)) {
	218	sock_put(sk);
	219	sk = NULL;
	220	}
	221	}
	222	break;
	223	default:
	224	WARN_ON(1);
	225	sk = NULL;
	226	}
	227
	228	pr_debug("tproxy socket lookup: proto %u %pI6:%u -> %pI6:%u, lookup type: %d, sock %p\n",
	229	protocol, saddr, ntohs(sport), daddr, ntohs(dport), lookup_type, sk);
	230
	231	return sk;
	232	}
	233	#endif
	234
106e4c26	235	/**
2c53040f	236	* tproxy_handle_time_wait4 - handle IPv4 TCP TIME_WAIT reopen redirections
106e4c26	237	* @skb: The skb being processed.
6ad78893 BS	238	* @laddr: IPv4 address to redirect to or zero.
6ad78893 BS	239	* @lport: TCP port to redirect to or zero.
106e4c26 BS	240	* @sk: The TIME_WAIT TCP socket found by the lookup.
	241	*
	242	* We have to handle SYN packets arriving to TIME_WAIT sockets
	243	* differently: instead of reopening the connection we should rather
	244	* redirect the new connection to the proxy if there's a listener
	245	* socket present.
	246	*
6ad78893	247	* tproxy_handle_time_wait4() consumes the socket reference passed in.
106e4c26 BS	248	*
	249	* Returns the listener socket if there's one, the TIME_WAIT socket if
	250	* no such listener is found, or NULL if the TCP header is incomplete.
	251	*/
	252	static struct sock *
686c9b50 EB	253	tproxy_handle_time_wait4(struct net net, struct sk_buff skb,
686c9b50 EB	254	__be32 laddr, __be16 lport, struct sock *sk)
106e4c26 BS	255	{
106e4c26 BS	256	const struct iphdr *iph = ip_hdr(skb);
106e4c26 BS	257	struct tcphdr _hdr, *hp;
	258
	259	hp = skb_header_pointer(skb, ip_hdrlen(skb), sizeof(_hdr), &_hdr);
	260	if (hp == NULL) {
	261	inet_twsk_put(inet_twsk(sk));
	262	return NULL;
	263	}
	264
	265	if (hp->syn && !hp->rst && !hp->ack && !hp->fin) {
	266	/* SYN to a TIME_WAIT socket, we'd rather redirect it
	267	* to a listener socket if there's one */
	268	struct sock *sk2;
	269
686c9b50	270	sk2 = nf_tproxy_get_sock_v4(net, iph->protocol,
6ad78893 BS	271	iph->saddr, laddr ? laddr : iph->daddr,
	272	hp->source, lport ? lport : hp->dest,
	273	skb->dev, NFT_LOOKUP_LISTENER);
	274	if (sk2) {
dbe7faa4	275	inet_twsk_deschedule_put(inet_twsk(sk));
6ad78893 BS	276	sk = sk2;
	277	}
	278	}
	279
	280	return sk;
	281	}
	282
fd158d79 FW	283	/* assign a socket to the skb -- consumes sk */
	284	static void
	285	nf_tproxy_assign_sock(struct sk_buff skb, struct sock sk)
	286	{
	287	skb_orphan(skb);
	288	skb->sk = sk;
	289	skb->destructor = sock_edemux;
	290	}
	291
e8439270	292	static unsigned int
686c9b50	293	tproxy_tg4(struct net net, struct sk_buff skb, __be32 laddr, __be16 lport,
6ad78893	294	u_int32_t mark_mask, u_int32_t mark_value)
e8439270 KK	295	{
e8439270 KK	296	const struct iphdr *iph = ip_hdr(skb);
e8439270 KK	297	struct udphdr _hdr, *hp;
	298	struct sock *sk;
	299
	300	hp = skb_header_pointer(skb, ip_hdrlen(skb), sizeof(_hdr), &_hdr);
	301	if (hp == NULL)
	302	return NF_DROP;
	303
6ad78893 BS	304	/* check if there's an ongoing connection on the packet
	305	* addresses, this happens if the redirect already happened
	306	* and the current packet belongs to an already established
	307	* connection */
686c9b50	308	sk = nf_tproxy_get_sock_v4(net, iph->protocol,
106e4c26 BS	309	iph->saddr, iph->daddr,
106e4c26 BS	310	hp->source, hp->dest,
6ad78893	311	skb->dev, NFT_LOOKUP_ESTABLISHED);
106e4c26	312
cc6eb433 BS	313	laddr = tproxy_laddr4(skb, laddr, iph->daddr);
	314	if (!lport)
	315	lport = hp->dest;
	316
106e4c26 BS	317	/* UDP has no TCP_TIME_WAIT state, so we never enter here */
106e4c26 BS	318	if (sk && sk->sk_state == TCP_TIME_WAIT)
6ad78893	319	/* reopening a TIME_WAIT connection needs special handling */
686c9b50	320	sk = tproxy_handle_time_wait4(net, skb, laddr, lport, sk);
106e4c26	321	else if (!sk)
6ad78893 BS	322	/* no, there's no established connection, check if
6ad78893 BS	323	* there's a listener on the redirected addr/port */
686c9b50	324	sk = nf_tproxy_get_sock_v4(net, iph->protocol,
cc6eb433 BS	325	iph->saddr, laddr,
cc6eb433 BS	326	hp->source, lport,
6ad78893 BS	327	skb->dev, NFT_LOOKUP_LISTENER);
	328
	329	/* NOTE: assign_sock consumes our sk reference */
d503b30b	330	if (sk && tproxy_sk_is_transparent(sk)) {
6ad78893 BS	331	/* This should be in a separate target, but we don't do multiple
	332	targets on the same rule yet */
	333	skb->mark = (skb->mark & ~mark_mask) ^ mark_value;
	334
	335	pr_debug("redirecting: proto %hhu %pI4:%hu -> %pI4:%hu, mark: %x\n",
	336	iph->protocol, &iph->daddr, ntohs(hp->dest),
	337	&laddr, ntohs(lport), skb->mark);
d503b30b FW	338
d503b30b FW	339	nf_tproxy_assign_sock(skb, sk);
6ad78893 BS	340	return NF_ACCEPT;
	341	}
	342
cc6eb433 BS	343	pr_debug("no socket, dropping: proto %hhu %pI4:%hu -> %pI4:%hu, mark: %x\n",
	344	iph->protocol, &iph->saddr, ntohs(hp->source),
	345	&iph->daddr, ntohs(hp->dest), skb->mark);
6ad78893 BS	346	return NF_DROP;
	347	}
	348
	349	static unsigned int
	350	tproxy_tg4_v0(struct sk_buff skb, const struct xt_action_param par)
	351	{
	352	const struct xt_tproxy_target_info *tgi = par->targinfo;
	353
686c9b50	354	return tproxy_tg4(par->net, skb, tgi->laddr, tgi->lport, tgi->mark_mask, tgi->mark_value);
6ad78893 BS	355	}
	356
	357	static unsigned int
	358	tproxy_tg4_v1(struct sk_buff skb, const struct xt_action_param par)
	359	{
	360	const struct xt_tproxy_target_info_v1 *tgi = par->targinfo;
	361
686c9b50	362	return tproxy_tg4(par->net, skb, tgi->laddr.ip, tgi->lport, tgi->mark_mask, tgi->mark_value);
6ad78893 BS	363	}
6ad78893 BS	364
f6318e55	365	#ifdef XT_TPROXY_HAVE_IPV6
cc6eb433 BS	366
	367	static inline const struct in6_addr *
	368	tproxy_laddr6(struct sk_buff skb, const struct in6_addr user_laddr,
	369	const struct in6_addr *daddr)
	370	{
	371	struct inet6_dev *indev;
	372	struct inet6_ifaddr *ifa;
	373	struct in6_addr *laddr;
	374
	375	if (!ipv6_addr_any(user_laddr))
	376	return user_laddr;
	377	laddr = NULL;
	378
	379	rcu_read_lock();
	380	indev = __in6_dev_get(skb->dev);
	381	if (indev)
	382	list_for_each_entry(ifa, &indev->addr_list, if_list) {
	383	if (ifa->flags & (IFA_F_TENTATIVE \| IFA_F_DEPRECATED))
	384	continue;
	385
	386	laddr = &ifa->addr;
	387	break;
	388	}
	389	rcu_read_unlock();
	390
	391	return laddr ? laddr : daddr;
	392	}
	393
	394	/**
2c53040f	395	* tproxy_handle_time_wait6 - handle IPv6 TCP TIME_WAIT reopen redirections
cc6eb433 BS	396	* @skb: The skb being processed.
	397	* @tproto: Transport protocol.
	398	* @thoff: Transport protocol header offset.
	399	* @par: Iptables target parameters.
	400	* @sk: The TIME_WAIT TCP socket found by the lookup.
	401	*
	402	* We have to handle SYN packets arriving to TIME_WAIT sockets
	403	* differently: instead of reopening the connection we should rather
	404	* redirect the new connection to the proxy if there's a listener
	405	* socket present.
	406	*
	407	* tproxy_handle_time_wait6() consumes the socket reference passed in.
	408	*
	409	* Returns the listener socket if there's one, the TIME_WAIT socket if
	410	* no such listener is found, or NULL if the TCP header is incomplete.
	411	*/
	412	static struct sock *
	413	tproxy_handle_time_wait6(struct sk_buff *skb, int tproto, int thoff,
	414	const struct xt_action_param *par,
	415	struct sock *sk)
	416	{
	417	const struct ipv6hdr *iph = ipv6_hdr(skb);
	418	struct tcphdr _hdr, *hp;
	419	const struct xt_tproxy_target_info_v1 *tgi = par->targinfo;
	420
	421	hp = skb_header_pointer(skb, thoff, sizeof(_hdr), &_hdr);
	422	if (hp == NULL) {
	423	inet_twsk_put(inet_twsk(sk));
	424	return NULL;
	425	}
	426
	427	if (hp->syn && !hp->rst && !hp->ack && !hp->fin) {
	428	/* SYN to a TIME_WAIT socket, we'd rather redirect it
	429	* to a listener socket if there's one */
	430	struct sock *sk2;
	431
686c9b50	432	sk2 = nf_tproxy_get_sock_v6(par->net, tproto,
cc6eb433 BS	433	&iph->saddr,
	434	tproxy_laddr6(skb, &tgi->laddr.in6, &iph->daddr),
	435	hp->source,
	436	tgi->lport ? tgi->lport : hp->dest,
	437	skb->dev, NFT_LOOKUP_LISTENER);
	438	if (sk2) {
dbe7faa4	439	inet_twsk_deschedule_put(inet_twsk(sk));
cc6eb433 BS	440	sk = sk2;
	441	}
	442	}
	443
	444	return sk;
	445	}
	446
6ad78893 BS	447	static unsigned int
	448	tproxy_tg6_v1(struct sk_buff skb, const struct xt_action_param par)
	449	{
	450	const struct ipv6hdr *iph = ipv6_hdr(skb);
	451	const struct xt_tproxy_target_info_v1 *tgi = par->targinfo;
	452	struct udphdr _hdr, *hp;
	453	struct sock *sk;
cc6eb433 BS	454	const struct in6_addr *laddr;
cc6eb433 BS	455	__be16 lport;
84018f55	456	int thoff = 0;
6ad78893 BS	457	int tproto;
6ad78893 BS	458
84018f55	459	tproto = ipv6_find_hdr(skb, &thoff, -1, NULL, NULL);
6ad78893 BS	460	if (tproto < 0) {
	461	pr_debug("unable to find transport header in IPv6 packet, dropping\n");
	462	return NF_DROP;
	463	}
	464
	465	hp = skb_header_pointer(skb, thoff, sizeof(_hdr), &_hdr);
	466	if (hp == NULL) {
	467	pr_debug("unable to grab transport header contents in IPv6 packet, dropping\n");
	468	return NF_DROP;
	469	}
	470
	471	/* check if there's an ongoing connection on the packet
	472	* addresses, this happens if the redirect already happened
	473	* and the current packet belongs to an already established
	474	* connection */
686c9b50	475	sk = nf_tproxy_get_sock_v6(par->net, tproto,
6ad78893 BS	476	&iph->saddr, &iph->daddr,
	477	hp->source, hp->dest,
	478	par->in, NFT_LOOKUP_ESTABLISHED);
	479
cc6eb433 BS	480	laddr = tproxy_laddr6(skb, &tgi->laddr.in6, &iph->daddr);
	481	lport = tgi->lport ? tgi->lport : hp->dest;
	482
6ad78893 BS	483	/* UDP has no TCP_TIME_WAIT state, so we never enter here */
	484	if (sk && sk->sk_state == TCP_TIME_WAIT)
	485	/* reopening a TIME_WAIT connection needs special handling */
	486	sk = tproxy_handle_time_wait6(skb, tproto, thoff, par, sk);
	487	else if (!sk)
	488	/* no there's no established connection, check if
	489	* there's a listener on the redirected addr/port */
686c9b50	490	sk = nf_tproxy_get_sock_v6(par->net, tproto,
cc6eb433 BS	491	&iph->saddr, laddr,
cc6eb433 BS	492	hp->source, lport,
106e4c26	493	par->in, NFT_LOOKUP_LISTENER);
e8439270 KK	494
e8439270 KK	495	/* NOTE: assign_sock consumes our sk reference */
d503b30b	496	if (sk && tproxy_sk_is_transparent(sk)) {
e8439270 KK	497	/* This should be in a separate target, but we don't do multiple
	498	targets on the same rule yet */
	499	skb->mark = (skb->mark & ~tgi->mark_mask) ^ tgi->mark_value;
	500
6ad78893	501	pr_debug("redirecting: proto %hhu %pI6:%hu -> %pI6:%hu, mark: %x\n",
cc6eb433 BS	502	tproto, &iph->saddr, ntohs(hp->source),
cc6eb433 BS	503	laddr, ntohs(lport), skb->mark);
d503b30b FW	504
d503b30b FW	505	nf_tproxy_assign_sock(skb, sk);
e8439270 KK	506	return NF_ACCEPT;
	507	}
	508
6ad78893	509	pr_debug("no socket, dropping: proto %hhu %pI6:%hu -> %pI6:%hu, mark: %x\n",
cc6eb433 BS	510	tproto, &iph->saddr, ntohs(hp->source),
	511	&iph->daddr, ntohs(hp->dest), skb->mark);
	512
e8439270 KK	513	return NF_DROP;
	514	}
	515
6ad78893 BS	516	static int tproxy_tg6_check(const struct xt_tgchk_param *par)
	517	{
	518	const struct ip6t_ip6 *i = par->entryinfo;
	519
3d8c6dce PNA	520	if ((i->proto == IPPROTO_TCP \|\| i->proto == IPPROTO_UDP) &&
3d8c6dce PNA	521	!(i->invflags & IP6T_INV_PROTO))
6ad78893 BS	522	return 0;
	523
	524	pr_info("Can be used only in combination with "
	525	"either -p tcp or -p udp\n");
	526	return -EINVAL;
	527	}
	528	#endif
	529
	530	static int tproxy_tg4_check(const struct xt_tgchk_param *par)
e8439270	531	{
af5d6dc2	532	const struct ipt_ip *i = par->entryinfo;
e8439270 KK	533
	534	if ((i->proto == IPPROTO_TCP \|\| i->proto == IPPROTO_UDP)
	535	&& !(i->invflags & IPT_INV_PROTO))
d6b00a53	536	return 0;
e8439270	537
ff67e4e4	538	pr_info("Can be used only in combination with "
e8439270	539	"either -p tcp or -p udp\n");
d6b00a53	540	return -EINVAL;
e8439270 KK	541	}
e8439270 KK	542
6ad78893 BS	543	static struct xt_target tproxy_tg_reg[] __read_mostly = {
	544	{
	545	.name = "TPROXY",
	546	.family = NFPROTO_IPV4,
	547	.table = "mangle",
	548	.target = tproxy_tg4_v0,
	549	.revision = 0,
	550	.targetsize = sizeof(struct xt_tproxy_target_info),
	551	.checkentry = tproxy_tg4_check,
	552	.hooks = 1 << NF_INET_PRE_ROUTING,
	553	.me = THIS_MODULE,
	554	},
	555	{
	556	.name = "TPROXY",
	557	.family = NFPROTO_IPV4,
	558	.table = "mangle",
	559	.target = tproxy_tg4_v1,
	560	.revision = 1,
	561	.targetsize = sizeof(struct xt_tproxy_target_info_v1),
	562	.checkentry = tproxy_tg4_check,
	563	.hooks = 1 << NF_INET_PRE_ROUTING,
	564	.me = THIS_MODULE,
	565	},
f6318e55	566	#ifdef XT_TPROXY_HAVE_IPV6
6ad78893 BS	567	{
	568	.name = "TPROXY",
	569	.family = NFPROTO_IPV6,
	570	.table = "mangle",
	571	.target = tproxy_tg6_v1,
	572	.revision = 1,
	573	.targetsize = sizeof(struct xt_tproxy_target_info_v1),
	574	.checkentry = tproxy_tg6_check,
	575	.hooks = 1 << NF_INET_PRE_ROUTING,
	576	.me = THIS_MODULE,
	577	},
	578	#endif
	579
e8439270 KK	580	};
	581
	582	static int __init tproxy_tg_init(void)
	583	{
	584	nf_defrag_ipv4_enable();
f6318e55	585	#ifdef XT_TPROXY_HAVE_IPV6
6ad78893 BS	586	nf_defrag_ipv6_enable();
	587	#endif
	588
	589	return xt_register_targets(tproxy_tg_reg, ARRAY_SIZE(tproxy_tg_reg));
e8439270 KK	590	}
	591
	592	static void __exit tproxy_tg_exit(void)
	593	{
6ad78893	594	xt_unregister_targets(tproxy_tg_reg, ARRAY_SIZE(tproxy_tg_reg));
e8439270 KK	595	}
	596
	597	module_init(tproxy_tg_init);
	598	module_exit(tproxy_tg_exit);
	599	MODULE_LICENSE("GPL");
6ad78893	600	MODULE_AUTHOR("Balazs Scheidler, Krisztian Kovacs");
e8439270 KK	601	MODULE_DESCRIPTION("Netfilter transparent proxy (TPROXY) target module.");
e8439270 KK	602	MODULE_ALIAS("ipt_TPROXY");
6ad78893	603	MODULE_ALIAS("ip6t_TPROXY");