[net-next-2.6.git] / net / ipv4 / netfilter / nf_nat_helper.c

/* ip_nat_helper.c - generic support functions for NAT helpers
 *
 * (C) 2000-2002 Harald Welte <laforge@netfilter.org>
 * (C) 2003-2006 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/module.h>
#include <linux/gfp.h>
#include <linux/kmod.h>
#include <linux/types.h>
#include <linux/timer.h>
#include <linux/skbuff.h>
#include <linux/tcp.h>
#include <linux/udp.h>
#include <net/checksum.h>
#include <net/tcp.h>
#include <net/route.h>

#include <linux/netfilter_ipv4.h>
#include <net/netfilter/nf_conntrack.h>
#include <net/netfilter/nf_conntrack_helper.h>
#include <net/netfilter/nf_conntrack_ecache.h>
#include <net/netfilter/nf_conntrack_expect.h>
#include <net/netfilter/nf_nat.h>
#include <net/netfilter/nf_nat_protocol.h>
#include <net/netfilter/nf_nat_core.h>
#include <net/netfilter/nf_nat_helper.h>

#define DUMP_OFFSET(x) \
	pr_debug("offset_before=%d, offset_after=%d, correction_pos=%u\n", \
		 x->offset_before, x->offset_after, x->correction_pos);

static DEFINE_SPINLOCK(nf_nat_seqofs_lock);

/* Setup TCP sequence correction given this change at this sequence */
static inline void
adjust_tcp_sequence(u32 seq,
		    int sizediff,
		    struct nf_conn *ct,
		    enum ip_conntrack_info ctinfo)
{
	enum ip_conntrack_dir dir = CTINFO2DIR(ctinfo);
	struct nf_conn_nat *nat = nfct_nat(ct);
	struct nf_nat_seq *this_way = &nat->seq[dir];

	pr_debug("adjust_tcp_sequence: seq = %u, sizediff = %d\n",
		 seq, sizediff);

	pr_debug("adjust_tcp_sequence: Seq_offset before: ");
	DUMP_OFFSET(this_way);

	spin_lock_bh(&nf_nat_seqofs_lock);

	/* SYN adjust. If it's uninitialized, or this is after last
	 * correction, record it: we don't handle more than one
	 * adjustment in the window, but do deal with common case of a
	 * retransmit */
	if (this_way->offset_before == this_way->offset_after ||
	    before(this_way->correction_pos, seq)) {
		this_way->correction_pos = seq;
		this_way->offset_before = this_way->offset_after;
		this_way->offset_after += sizediff;
	}
	spin_unlock_bh(&nf_nat_seqofs_lock);

	pr_debug("adjust_tcp_sequence: Seq_offset after: ");
	DUMP_OFFSET(this_way);
}

/* Get the offset value, for conntrack */
s16 nf_nat_get_offset(const struct nf_conn *ct,
		      enum ip_conntrack_dir dir,
		      u32 seq)
{
	struct nf_conn_nat *nat = nfct_nat(ct);
	struct nf_nat_seq *this_way;
	s16 offset;

	if (!nat)
		return 0;

	this_way = &nat->seq[dir];
	spin_lock_bh(&nf_nat_seqofs_lock);
	offset = after(seq, this_way->correction_pos)
		 ? this_way->offset_after : this_way->offset_before;
	spin_unlock_bh(&nf_nat_seqofs_lock);

	return offset;
}
EXPORT_SYMBOL_GPL(nf_nat_get_offset);

/* Frobs data inside this packet, which is linear. */
static void mangle_contents(struct sk_buff *skb,
			    unsigned int dataoff,
			    unsigned int match_offset,
			    unsigned int match_len,
			    const char *rep_buffer,
			    unsigned int rep_len)
{
	unsigned char *data;

	BUG_ON(skb_is_nonlinear(skb));
	data = skb_network_header(skb) + dataoff;

	/* move post-replacement */
	memmove(data + match_offset + rep_len,
		data + match_offset + match_len,
		skb->tail - (skb->network_header + dataoff +
			     match_offset + match_len));

	/* insert data from buffer */
	memcpy(data + match_offset, rep_buffer, rep_len);

	/* update skb info */
	if (rep_len > match_len) {
		pr_debug("nf_nat_mangle_packet: Extending packet by "
			 "%u from %u bytes\n", rep_len - match_len, skb->len);
		skb_put(skb, rep_len - match_len);
	} else {
		pr_debug("nf_nat_mangle_packet: Shrinking packet from "
			 "%u from %u bytes\n", match_len - rep_len, skb->len);
		__skb_trim(skb, skb->len + rep_len - match_len);
	}

	/* fix IP hdr checksum information */
	ip_hdr(skb)->tot_len = htons(skb->len);
	ip_send_check(ip_hdr(skb));
}

/* Unusual, but possible case. */
static int enlarge_skb(struct sk_buff *skb, unsigned int extra)
{
	if (skb->len + extra > 65535)
		return 0;

	if (pskb_expand_head(skb, 0, extra - skb_tailroom(skb), GFP_ATOMIC))
		return 0;

	return 1;
}

void nf_nat_set_seq_adjust(struct nf_conn *ct, enum ip_conntrack_info ctinfo,
			   __be32 seq, s16 off)
{
	if (!off)
		return;
	set_bit(IPS_SEQ_ADJUST_BIT, &ct->status);
	adjust_tcp_sequence(ntohl(seq), off, ct, ctinfo);
	nf_conntrack_event_cache(IPCT_NATSEQADJ, ct);
}
EXPORT_SYMBOL_GPL(nf_nat_set_seq_adjust);

/* Generic function for mangling variable-length address changes inside
 * NATed TCP connections (like the PORT XXX,XXX,XXX,XXX,XXX,XXX
 * command in FTP).
 *
 * Takes care about all the nasty sequence number changes, checksumming,
 * skb enlargement, ...
 *
 * */
int __nf_nat_mangle_tcp_packet(struct sk_buff *skb,
			       struct nf_conn *ct,
			       enum ip_conntrack_info ctinfo,
			       unsigned int match_offset,
			       unsigned int match_len,
			       const char *rep_buffer,
			       unsigned int rep_len, bool adjust)
{
	struct rtable *rt = skb_rtable(skb);
	struct iphdr *iph;
	struct tcphdr *tcph;
	int oldlen, datalen;

	if (!skb_make_writable(skb, skb->len))
		return 0;

	if (rep_len > match_len &&
	    rep_len - match_len > skb_tailroom(skb) &&
	    !enlarge_skb(skb, rep_len - match_len))
		return 0;

	SKB_LINEAR_ASSERT(skb);

	iph = ip_hdr(skb);
	tcph = (void *)iph + iph->ihl*4;

	oldlen = skb->len - iph->ihl*4;
	mangle_contents(skb, iph->ihl*4 + tcph->doff*4,
			match_offset, match_len, rep_buffer, rep_len);

	datalen = skb->len - iph->ihl*4;
	if (skb->ip_summed != CHECKSUM_PARTIAL) {
		if (!(rt->rt_flags & RTCF_LOCAL) &&
		    skb->dev->features & NETIF_F_V4_CSUM) {
			skb->ip_summed = CHECKSUM_PARTIAL;
			skb->csum_start = skb_headroom(skb) +
					  skb_network_offset(skb) +
					  iph->ihl * 4;
			skb->csum_offset = offsetof(struct tcphdr, check);
			tcph->check = ~tcp_v4_check(datalen,
						    iph->saddr, iph->daddr, 0);
		} else {
			tcph->check = 0;
			tcph->check = tcp_v4_check(datalen,
						   iph->saddr, iph->daddr,
						   csum_partial(tcph,
								datalen, 0));
		}
	} else
		inet_proto_csum_replace2(&tcph->check, skb,
					 htons(oldlen), htons(datalen), 1);

	if (adjust && rep_len != match_len)
		nf_nat_set_seq_adjust(ct, ctinfo, tcph->seq,
				      (int)rep_len - (int)match_len);

	return 1;
}
EXPORT_SYMBOL(__nf_nat_mangle_tcp_packet);

/* Generic function for mangling variable-length address changes inside
 * NATed UDP connections (like the CONNECT DATA XXXXX MESG XXXXX INDEX XXXXX
 * command in the Amanda protocol)
 *
 * Takes care about all the nasty sequence number changes, checksumming,
 * skb enlargement, ...
 *
 * XXX - This function could be merged with nf_nat_mangle_tcp_packet which
 *       should be fairly easy to do.
 */
int
nf_nat_mangle_udp_packet(struct sk_buff *skb,
			 struct nf_conn *ct,
			 enum ip_conntrack_info ctinfo,
			 unsigned int match_offset,
			 unsigned int match_len,
			 const char *rep_buffer,
			 unsigned int rep_len)
{
	struct rtable *rt = skb_rtable(skb);
	struct iphdr *iph;
	struct udphdr *udph;
	int datalen, oldlen;

	/* UDP helpers might accidentally mangle the wrong packet */
	iph = ip_hdr(skb);
	if (skb->len < iph->ihl*4 + sizeof(*udph) +
			       match_offset + match_len)
		return 0;

	if (!skb_make_writable(skb, skb->len))
		return 0;

	if (rep_len > match_len &&
	    rep_len - match_len > skb_tailroom(skb) &&
	    !enlarge_skb(skb, rep_len - match_len))
		return 0;

	iph = ip_hdr(skb);
	udph = (void *)iph + iph->ihl*4;

	oldlen = skb->len - iph->ihl*4;
	mangle_contents(skb, iph->ihl*4 + sizeof(*udph),
			match_offset, match_len, rep_buffer, rep_len);

	/* update the length of the UDP packet */
	datalen = skb->len - iph->ihl*4;
	udph->len = htons(datalen);

	/* fix udp checksum if udp checksum was previously calculated */
	if (!udph->check && skb->ip_summed != CHECKSUM_PARTIAL)
		return 1;

	if (skb->ip_summed != CHECKSUM_PARTIAL) {
		if (!(rt->rt_flags & RTCF_LOCAL) &&
		    skb->dev->features & NETIF_F_V4_CSUM) {
			skb->ip_summed = CHECKSUM_PARTIAL;
			skb->csum_start = skb_headroom(skb) +
					  skb_network_offset(skb) +
					  iph->ihl * 4;
			skb->csum_offset = offsetof(struct udphdr, check);
			udph->check = ~csum_tcpudp_magic(iph->saddr, iph->daddr,
							 datalen, IPPROTO_UDP,
							 0);
		} else {
			udph->check = 0;
			udph->check = csum_tcpudp_magic(iph->saddr, iph->daddr,
							datalen, IPPROTO_UDP,
							csum_partial(udph,
								     datalen, 0));
			if (!udph->check)
				udph->check = CSUM_MANGLED_0;
		}
	} else
		inet_proto_csum_replace2(&udph->check, skb,
					 htons(oldlen), htons(datalen), 1);

	return 1;
}
EXPORT_SYMBOL(nf_nat_mangle_udp_packet);

/* Adjust one found SACK option including checksum correction */
static void
sack_adjust(struct sk_buff *skb,
	    struct tcphdr *tcph,
	    unsigned int sackoff,
	    unsigned int sackend,
	    struct nf_nat_seq *natseq)
{
	while (sackoff < sackend) {
		struct tcp_sack_block_wire *sack;
		__be32 new_start_seq, new_end_seq;

		sack = (void *)skb->data + sackoff;
		if (after(ntohl(sack->start_seq) - natseq->offset_before,
			  natseq->correction_pos))
			new_start_seq = htonl(ntohl(sack->start_seq)
					- natseq->offset_after);
		else
			new_start_seq = htonl(ntohl(sack->start_seq)
					- natseq->offset_before);

		if (after(ntohl(sack->end_seq) - natseq->offset_before,
			  natseq->correction_pos))
			new_end_seq = htonl(ntohl(sack->end_seq)
				      - natseq->offset_after);
		else
			new_end_seq = htonl(ntohl(sack->end_seq)
				      - natseq->offset_before);

		pr_debug("sack_adjust: start_seq: %d->%d, end_seq: %d->%d\n",
			 ntohl(sack->start_seq), new_start_seq,
			 ntohl(sack->end_seq), new_end_seq);

		inet_proto_csum_replace4(&tcph->check, skb,
					 sack->start_seq, new_start_seq, 0);
		inet_proto_csum_replace4(&tcph->check, skb,
					 sack->end_seq, new_end_seq, 0);
		sack->start_seq = new_start_seq;
		sack->end_seq = new_end_seq;
		sackoff += sizeof(*sack);
	}
}

/* TCP SACK sequence number adjustment */
static inline unsigned int
nf_nat_sack_adjust(struct sk_buff *skb,
		   struct tcphdr *tcph,
		   struct nf_conn *ct,
		   enum ip_conntrack_info ctinfo)
{
	unsigned int dir, optoff, optend;
	struct nf_conn_nat *nat = nfct_nat(ct);

	optoff = ip_hdrlen(skb) + sizeof(struct tcphdr);
	optend = ip_hdrlen(skb) + tcph->doff * 4;

	if (!skb_make_writable(skb, optend))
		return 0;

	dir = CTINFO2DIR(ctinfo);

	while (optoff < optend) {
		/* Usually: option, length. */
		unsigned char *op = skb->data + optoff;

		switch (op[0]) {
		case TCPOPT_EOL:
			return 1;
		case TCPOPT_NOP:
			optoff++;
			continue;
		default:
			/* no partial options */
			if (optoff + 1 == optend ||
			    optoff + op[1] > optend ||
			    op[1] < 2)
				return 0;
			if (op[0] == TCPOPT_SACK &&
			    op[1] >= 2+TCPOLEN_SACK_PERBLOCK &&
			    ((op[1] - 2) % TCPOLEN_SACK_PERBLOCK) == 0)
				sack_adjust(skb, tcph, optoff+2,
					    optoff+op[1], &nat->seq[!dir]);
			optoff += op[1];
		}
	}
	return 1;
}

/* TCP sequence number adjustment.  Returns 1 on success, 0 on failure */
int
nf_nat_seq_adjust(struct sk_buff *skb,
		  struct nf_conn *ct,
		  enum ip_conntrack_info ctinfo)
{
	struct tcphdr *tcph;
	int dir;
	__be32 newseq, newack;
	s16 seqoff, ackoff;
	struct nf_conn_nat *nat = nfct_nat(ct);
	struct nf_nat_seq *this_way, *other_way;

	dir = CTINFO2DIR(ctinfo);

	this_way = &nat->seq[dir];
	other_way = &nat->seq[!dir];

	if (!skb_make_writable(skb, ip_hdrlen(skb) + sizeof(*tcph)))
		return 0;

	tcph = (void *)skb->data + ip_hdrlen(skb);
	if (after(ntohl(tcph->seq), this_way->correction_pos))
		seqoff = this_way->offset_after;
	else
		seqoff = this_way->offset_before;

	if (after(ntohl(tcph->ack_seq) - other_way->offset_before,
		  other_way->correction_pos))
		ackoff = other_way->offset_after;
	else
		ackoff = other_way->offset_before;

	newseq = htonl(ntohl(tcph->seq) + seqoff);
	newack = htonl(ntohl(tcph->ack_seq) - ackoff);

	inet_proto_csum_replace4(&tcph->check, skb, tcph->seq, newseq, 0);
	inet_proto_csum_replace4(&tcph->check, skb, tcph->ack_seq, newack, 0);

	pr_debug("Adjusting sequence number from %u->%u, ack from %u->%u\n",
		 ntohl(tcph->seq), ntohl(newseq), ntohl(tcph->ack_seq),
		 ntohl(newack));

	tcph->seq = newseq;
	tcph->ack_seq = newack;

	return nf_nat_sack_adjust(skb, tcph, ct, ctinfo);
}

/* Setup NAT on this expected conntrack so it follows master. */
/* If we fail to get a free NAT slot, we'll get dropped on confirm */
void nf_nat_follow_master(struct nf_conn *ct,
			  struct nf_conntrack_expect *exp)
{
	struct nf_nat_range range;

	/* This must be a fresh one. */
	BUG_ON(ct->status & IPS_NAT_DONE_MASK);

	/* Change src to where master sends to */
	range.flags = IP_NAT_RANGE_MAP_IPS;
	range.min_ip = range.max_ip
		= ct->master->tuplehash[!exp->dir].tuple.dst.u3.ip;
	nf_nat_setup_info(ct, &range, IP_NAT_MANIP_SRC);

	/* For DST manip, map port here to where it's expected. */
	range.flags = (IP_NAT_RANGE_MAP_IPS | IP_NAT_RANGE_PROTO_SPECIFIED);
	range.min = range.max = exp->saved_proto;
	range.min_ip = range.max_ip
		= ct->master->tuplehash[!exp->dir].tuple.src.u3.ip;
	nf_nat_setup_info(ct, &range, IP_NAT_MANIP_DST);
}
EXPORT_SYMBOL(nf_nat_follow_master);
Commit	Line	Data
5b1158e9 JK	1	/* ip_nat_helper.c - generic support functions for NAT helpers
	2	*
	3	* (C) 2000-2002 Harald Welte <laforge@netfilter.org>
	4	* (C) 2003-2006 Netfilter Core Team <coreteam@netfilter.org>
	5	*
	6	* This program is free software; you can redistribute it and/or modify
	7	* it under the terms of the GNU General Public License version 2 as
	8	* published by the Free Software Foundation.
	9	*/
	10	#include <linux/module.h>
5a0e3ad6	11	#include <linux/gfp.h>
5b1158e9 JK	12	#include <linux/kmod.h>
	13	#include <linux/types.h>
	14	#include <linux/timer.h>
	15	#include <linux/skbuff.h>
	16	#include <linux/tcp.h>
	17	#include <linux/udp.h>
	18	#include <net/checksum.h>
	19	#include <net/tcp.h>
1668e010	20	#include <net/route.h>
5b1158e9 JK	21
	22	#include <linux/netfilter_ipv4.h>
	23	#include <net/netfilter/nf_conntrack.h>
	24	#include <net/netfilter/nf_conntrack_helper.h>
13eae15a	25	#include <net/netfilter/nf_conntrack_ecache.h>
5b1158e9 JK	26	#include <net/netfilter/nf_conntrack_expect.h>
	27	#include <net/netfilter/nf_nat.h>
	28	#include <net/netfilter/nf_nat_protocol.h>
	29	#include <net/netfilter/nf_nat_core.h>
	30	#include <net/netfilter/nf_nat_helper.h>
	31
0d53778e PM	32	#define DUMP_OFFSET(x) \
	33	pr_debug("offset_before=%d, offset_after=%d, correction_pos=%u\n", \
	34	x->offset_before, x->offset_after, x->correction_pos);
5b1158e9 JK	35
	36	static DEFINE_SPINLOCK(nf_nat_seqofs_lock);
	37
	38	/* Setup TCP sequence correction given this change at this sequence */
	39	static inline void
	40	adjust_tcp_sequence(u32 seq,
	41	int sizediff,
	42	struct nf_conn *ct,
	43	enum ip_conntrack_info ctinfo)
	44	{
76ac8940	45	enum ip_conntrack_dir dir = CTINFO2DIR(ctinfo);
5b1158e9	46	struct nf_conn_nat *nat = nfct_nat(ct);
76ac8940	47	struct nf_nat_seq *this_way = &nat->seq[dir];
5b1158e9	48
76ac8940 HE	49	pr_debug("adjust_tcp_sequence: seq = %u, sizediff = %d\n",
76ac8940 HE	50	seq, sizediff);
5b1158e9	51
76ac8940	52	pr_debug("adjust_tcp_sequence: Seq_offset before: ");
5b1158e9 JK	53	DUMP_OFFSET(this_way);
	54
	55	spin_lock_bh(&nf_nat_seqofs_lock);
	56
	57	/* SYN adjust. If it's uninitialized, or this is after last
	58	* correction, record it: we don't handle more than one
	59	* adjustment in the window, but do deal with common case of a
	60	* retransmit */
	61	if (this_way->offset_before == this_way->offset_after \|\|
	62	before(this_way->correction_pos, seq)) {
76ac8940 HE	63	this_way->correction_pos = seq;
	64	this_way->offset_before = this_way->offset_after;
	65	this_way->offset_after += sizediff;
5b1158e9 JK	66	}
	67	spin_unlock_bh(&nf_nat_seqofs_lock);
	68
76ac8940	69	pr_debug("adjust_tcp_sequence: Seq_offset after: ");
5b1158e9 JK	70	DUMP_OFFSET(this_way);
	71	}
	72
f9dd09c7 JK	73	/* Get the offset value, for conntrack */
	74	s16 nf_nat_get_offset(const struct nf_conn *ct,
	75	enum ip_conntrack_dir dir,
	76	u32 seq)
	77	{
	78	struct nf_conn_nat *nat = nfct_nat(ct);
	79	struct nf_nat_seq *this_way;
	80	s16 offset;
	81
	82	if (!nat)
	83	return 0;
	84
	85	this_way = &nat->seq[dir];
	86	spin_lock_bh(&nf_nat_seqofs_lock);
	87	offset = after(seq, this_way->correction_pos)
	88	? this_way->offset_after : this_way->offset_before;
	89	spin_unlock_bh(&nf_nat_seqofs_lock);
	90
	91	return offset;
	92	}
	93	EXPORT_SYMBOL_GPL(nf_nat_get_offset);
	94
5b1158e9 JK	95	/* Frobs data inside this packet, which is linear. */
	96	static void mangle_contents(struct sk_buff *skb,
	97	unsigned int dataoff,
	98	unsigned int match_offset,
	99	unsigned int match_len,
	100	const char *rep_buffer,
	101	unsigned int rep_len)
	102	{
	103	unsigned char *data;
	104
	105	BUG_ON(skb_is_nonlinear(skb));
eddc9ec5	106	data = skb_network_header(skb) + dataoff;
5b1158e9 JK	107
	108	/* move post-replacement */
	109	memmove(data + match_offset + rep_len,
	110	data + match_offset + match_len,
27a884dc ACM	111	skb->tail - (skb->network_header + dataoff +
27a884dc ACM	112	match_offset + match_len));
5b1158e9 JK	113
	114	/* insert data from buffer */
	115	memcpy(data + match_offset, rep_buffer, rep_len);
	116
	117	/* update skb info */
	118	if (rep_len > match_len) {
0d53778e PM	119	pr_debug("nf_nat_mangle_packet: Extending packet by "
0d53778e PM	120	"%u from %u bytes\n", rep_len - match_len, skb->len);
5b1158e9 JK	121	skb_put(skb, rep_len - match_len);
5b1158e9 JK	122	} else {
0d53778e PM	123	pr_debug("nf_nat_mangle_packet: Shrinking packet from "
0d53778e PM	124	"%u from %u bytes\n", match_len - rep_len, skb->len);
5b1158e9 JK	125	__skb_trim(skb, skb->len + rep_len - match_len);
	126	}
	127
	128	/* fix IP hdr checksum information */
eddc9ec5 ACM	129	ip_hdr(skb)->tot_len = htons(skb->len);
eddc9ec5 ACM	130	ip_send_check(ip_hdr(skb));
5b1158e9 JK	131	}
	132
	133	/* Unusual, but possible case. */
3db05fea	134	static int enlarge_skb(struct sk_buff *skb, unsigned int extra)
5b1158e9	135	{
3db05fea	136	if (skb->len + extra > 65535)
5b1158e9 JK	137	return 0;
5b1158e9 JK	138
3db05fea	139	if (pskb_expand_head(skb, 0, extra - skb_tailroom(skb), GFP_ATOMIC))
5b1158e9 JK	140	return 0;
5b1158e9 JK	141
5b1158e9 JK	142	return 1;
	143	}
	144
010c0b9f PM	145	void nf_nat_set_seq_adjust(struct nf_conn *ct, enum ip_conntrack_info ctinfo,
	146	__be32 seq, s16 off)
	147	{
	148	if (!off)
	149	return;
	150	set_bit(IPS_SEQ_ADJUST_BIT, &ct->status);
	151	adjust_tcp_sequence(ntohl(seq), off, ct, ctinfo);
	152	nf_conntrack_event_cache(IPCT_NATSEQADJ, ct);
	153	}
	154	EXPORT_SYMBOL_GPL(nf_nat_set_seq_adjust);
	155
5b1158e9 JK	156	/* Generic function for mangling variable-length address changes inside
	157	* NATed TCP connections (like the PORT XXX,XXX,XXX,XXX,XXX,XXX
	158	* command in FTP).
	159	*
	160	* Takes care about all the nasty sequence number changes, checksumming,
	161	* skb enlargement, ...
	162	*
	163	* */
010c0b9f PM	164	int __nf_nat_mangle_tcp_packet(struct sk_buff *skb,
	165	struct nf_conn *ct,
	166	enum ip_conntrack_info ctinfo,
	167	unsigned int match_offset,
	168	unsigned int match_len,
	169	const char *rep_buffer,
	170	unsigned int rep_len, bool adjust)
5b1158e9	171	{
511c3f92	172	struct rtable *rt = skb_rtable(skb);
5b1158e9 JK	173	struct iphdr *iph;
	174	struct tcphdr *tcph;
	175	int oldlen, datalen;
	176
3db05fea	177	if (!skb_make_writable(skb, skb->len))
5b1158e9 JK	178	return 0;
	179
	180	if (rep_len > match_len &&
3db05fea HX	181	rep_len - match_len > skb_tailroom(skb) &&
3db05fea HX	182	!enlarge_skb(skb, rep_len - match_len))
5b1158e9 JK	183	return 0;
5b1158e9 JK	184
3db05fea	185	SKB_LINEAR_ASSERT(skb);
5b1158e9	186
3db05fea	187	iph = ip_hdr(skb);
5b1158e9 JK	188	tcph = (void )iph + iph->ihl4;
5b1158e9 JK	189
3db05fea HX	190	oldlen = skb->len - iph->ihl*4;
3db05fea HX	191	mangle_contents(skb, iph->ihl4 + tcph->doff4,
5b1158e9 JK	192	match_offset, match_len, rep_buffer, rep_len);
5b1158e9 JK	193
3db05fea HX	194	datalen = skb->len - iph->ihl*4;
3db05fea HX	195	if (skb->ip_summed != CHECKSUM_PARTIAL) {
fe6092ea	196	if (!(rt->rt_flags & RTCF_LOCAL) &&
3db05fea HX	197	skb->dev->features & NETIF_F_V4_CSUM) {
	198	skb->ip_summed = CHECKSUM_PARTIAL;
	199	skb->csum_start = skb_headroom(skb) +
	200	skb_network_offset(skb) +
	201	iph->ihl * 4;
	202	skb->csum_offset = offsetof(struct tcphdr, check);
fe6092ea PM	203	tcph->check = ~tcp_v4_check(datalen,
	204	iph->saddr, iph->daddr, 0);
	205	} else {
	206	tcph->check = 0;
	207	tcph->check = tcp_v4_check(datalen,
	208	iph->saddr, iph->daddr,
a47362a2	209	csum_partial(tcph,
fe6092ea PM	210	datalen, 0));
fe6092ea PM	211	}
5b1158e9	212	} else
be0ea7d5 PM	213	inet_proto_csum_replace2(&tcph->check, skb,
be0ea7d5 PM	214	htons(oldlen), htons(datalen), 1);
5b1158e9	215
010c0b9f PM	216	if (adjust && rep_len != match_len)
	217	nf_nat_set_seq_adjust(ct, ctinfo, tcph->seq,
	218	(int)rep_len - (int)match_len);
	219
5b1158e9 JK	220	return 1;
5b1158e9 JK	221	}
010c0b9f	222	EXPORT_SYMBOL(__nf_nat_mangle_tcp_packet);
5b1158e9 JK	223
	224	/* Generic function for mangling variable-length address changes inside
	225	* NATed UDP connections (like the CONNECT DATA XXXXX MESG XXXXX INDEX XXXXX
	226	* command in the Amanda protocol)
	227	*
	228	* Takes care about all the nasty sequence number changes, checksumming,
	229	* skb enlargement, ...
	230	*
	231	* XXX - This function could be merged with nf_nat_mangle_tcp_packet which
	232	* should be fairly easy to do.
	233	*/
	234	int
3db05fea	235	nf_nat_mangle_udp_packet(struct sk_buff *skb,
5b1158e9 JK	236	struct nf_conn *ct,
	237	enum ip_conntrack_info ctinfo,
	238	unsigned int match_offset,
	239	unsigned int match_len,
	240	const char *rep_buffer,
	241	unsigned int rep_len)
	242	{
511c3f92	243	struct rtable *rt = skb_rtable(skb);
5b1158e9 JK	244	struct iphdr *iph;
	245	struct udphdr *udph;
	246	int datalen, oldlen;
	247
	248	/* UDP helpers might accidentally mangle the wrong packet */
3db05fea HX	249	iph = ip_hdr(skb);
3db05fea HX	250	if (skb->len < iph->ihl4 + sizeof(udph) +
e905a9ed	251	match_offset + match_len)
5b1158e9 JK	252	return 0;
5b1158e9 JK	253
3db05fea	254	if (!skb_make_writable(skb, skb->len))
5b1158e9 JK	255	return 0;
	256
	257	if (rep_len > match_len &&
3db05fea HX	258	rep_len - match_len > skb_tailroom(skb) &&
3db05fea HX	259	!enlarge_skb(skb, rep_len - match_len))
5b1158e9 JK	260	return 0;
5b1158e9 JK	261
3db05fea	262	iph = ip_hdr(skb);
5b1158e9 JK	263	udph = (void )iph + iph->ihl4;
5b1158e9 JK	264
3db05fea HX	265	oldlen = skb->len - iph->ihl*4;
3db05fea HX	266	mangle_contents(skb, iph->ihl4 + sizeof(udph),
5b1158e9 JK	267	match_offset, match_len, rep_buffer, rep_len);
	268
	269	/* update the length of the UDP packet */
3db05fea	270	datalen = skb->len - iph->ihl*4;
5b1158e9 JK	271	udph->len = htons(datalen);
	272
	273	/* fix udp checksum if udp checksum was previously calculated */
3db05fea	274	if (!udph->check && skb->ip_summed != CHECKSUM_PARTIAL)
5b1158e9 JK	275	return 1;
5b1158e9 JK	276
3db05fea	277	if (skb->ip_summed != CHECKSUM_PARTIAL) {
fe6092ea	278	if (!(rt->rt_flags & RTCF_LOCAL) &&
3db05fea HX	279	skb->dev->features & NETIF_F_V4_CSUM) {
	280	skb->ip_summed = CHECKSUM_PARTIAL;
	281	skb->csum_start = skb_headroom(skb) +
	282	skb_network_offset(skb) +
	283	iph->ihl * 4;
	284	skb->csum_offset = offsetof(struct udphdr, check);
fe6092ea PM	285	udph->check = ~csum_tcpudp_magic(iph->saddr, iph->daddr,
	286	datalen, IPPROTO_UDP,
	287	0);
	288	} else {
	289	udph->check = 0;
	290	udph->check = csum_tcpudp_magic(iph->saddr, iph->daddr,
	291	datalen, IPPROTO_UDP,
a47362a2	292	csum_partial(udph,
fe6092ea PM	293	datalen, 0));
	294	if (!udph->check)
	295	udph->check = CSUM_MANGLED_0;
	296	}
5b1158e9	297	} else
be0ea7d5 PM	298	inet_proto_csum_replace2(&udph->check, skb,
be0ea7d5 PM	299	htons(oldlen), htons(datalen), 1);
5b1158e9 JK	300
	301	return 1;
	302	}
	303	EXPORT_SYMBOL(nf_nat_mangle_udp_packet);
	304
	305	/* Adjust one found SACK option including checksum correction */
	306	static void
	307	sack_adjust(struct sk_buff *skb,
	308	struct tcphdr *tcph,
	309	unsigned int sackoff,
	310	unsigned int sackend,
	311	struct nf_nat_seq *natseq)
	312	{
	313	while (sackoff < sackend) {
	314	struct tcp_sack_block_wire *sack;
	315	__be32 new_start_seq, new_end_seq;
	316
	317	sack = (void *)skb->data + sackoff;
	318	if (after(ntohl(sack->start_seq) - natseq->offset_before,
	319	natseq->correction_pos))
	320	new_start_seq = htonl(ntohl(sack->start_seq)
	321	- natseq->offset_after);
	322	else
	323	new_start_seq = htonl(ntohl(sack->start_seq)
	324	- natseq->offset_before);
	325
	326	if (after(ntohl(sack->end_seq) - natseq->offset_before,
	327	natseq->correction_pos))
	328	new_end_seq = htonl(ntohl(sack->end_seq)
	329	- natseq->offset_after);
	330	else
	331	new_end_seq = htonl(ntohl(sack->end_seq)
	332	- natseq->offset_before);
	333
0d53778e PM	334	pr_debug("sack_adjust: start_seq: %d->%d, end_seq: %d->%d\n",
	335	ntohl(sack->start_seq), new_start_seq,
	336	ntohl(sack->end_seq), new_end_seq);
5b1158e9	337
be0ea7d5 PM	338	inet_proto_csum_replace4(&tcph->check, skb,
	339	sack->start_seq, new_start_seq, 0);
	340	inet_proto_csum_replace4(&tcph->check, skb,
	341	sack->end_seq, new_end_seq, 0);
5b1158e9 JK	342	sack->start_seq = new_start_seq;
	343	sack->end_seq = new_end_seq;
	344	sackoff += sizeof(*sack);
	345	}
	346	}
	347
	348	/* TCP SACK sequence number adjustment */
	349	static inline unsigned int
3db05fea	350	nf_nat_sack_adjust(struct sk_buff *skb,
5b1158e9 JK	351	struct tcphdr *tcph,
	352	struct nf_conn *ct,
	353	enum ip_conntrack_info ctinfo)
	354	{
	355	unsigned int dir, optoff, optend;
	356	struct nf_conn_nat *nat = nfct_nat(ct);
	357
3db05fea HX	358	optoff = ip_hdrlen(skb) + sizeof(struct tcphdr);
3db05fea HX	359	optend = ip_hdrlen(skb) + tcph->doff * 4;
5b1158e9	360
3db05fea	361	if (!skb_make_writable(skb, optend))
5b1158e9 JK	362	return 0;
	363
	364	dir = CTINFO2DIR(ctinfo);
	365
	366	while (optoff < optend) {
	367	/* Usually: option, length. */
3db05fea	368	unsigned char *op = skb->data + optoff;
5b1158e9 JK	369
	370	switch (op[0]) {
	371	case TCPOPT_EOL:
	372	return 1;
	373	case TCPOPT_NOP:
	374	optoff++;
	375	continue;
	376	default:
	377	/* no partial options */
	378	if (optoff + 1 == optend \|\|
	379	optoff + op[1] > optend \|\|
	380	op[1] < 2)
	381	return 0;
	382	if (op[0] == TCPOPT_SACK &&
	383	op[1] >= 2+TCPOLEN_SACK_PERBLOCK &&
	384	((op[1] - 2) % TCPOLEN_SACK_PERBLOCK) == 0)
3db05fea	385	sack_adjust(skb, tcph, optoff+2,
b6b84d4a	386	optoff+op[1], &nat->seq[!dir]);
5b1158e9 JK	387	optoff += op[1];
	388	}
	389	}
	390	return 1;
	391	}
	392
	393	/* TCP sequence number adjustment. Returns 1 on success, 0 on failure */
	394	int
3db05fea	395	nf_nat_seq_adjust(struct sk_buff *skb,
5b1158e9 JK	396	struct nf_conn *ct,
	397	enum ip_conntrack_info ctinfo)
	398	{
	399	struct tcphdr *tcph;
	400	int dir;
	401	__be32 newseq, newack;
a3a9f79e	402	s16 seqoff, ackoff;
5b1158e9 JK	403	struct nf_conn_nat *nat = nfct_nat(ct);
	404	struct nf_nat_seq this_way, other_way;
	405
	406	dir = CTINFO2DIR(ctinfo);
	407
b6b84d4a YK	408	this_way = &nat->seq[dir];
b6b84d4a YK	409	other_way = &nat->seq[!dir];
5b1158e9	410
3db05fea	411	if (!skb_make_writable(skb, ip_hdrlen(skb) + sizeof(*tcph)))
5b1158e9 JK	412	return 0;
5b1158e9 JK	413
3db05fea	414	tcph = (void *)skb->data + ip_hdrlen(skb);
5b1158e9	415	if (after(ntohl(tcph->seq), this_way->correction_pos))
a3a9f79e	416	seqoff = this_way->offset_after;
5b1158e9	417	else
a3a9f79e	418	seqoff = this_way->offset_before;
5b1158e9 JK	419
	420	if (after(ntohl(tcph->ack_seq) - other_way->offset_before,
	421	other_way->correction_pos))
a3a9f79e	422	ackoff = other_way->offset_after;
5b1158e9	423	else
a3a9f79e PM	424	ackoff = other_way->offset_before;
	425
	426	newseq = htonl(ntohl(tcph->seq) + seqoff);
	427	newack = htonl(ntohl(tcph->ack_seq) - ackoff);
5b1158e9	428
be0ea7d5 PM	429	inet_proto_csum_replace4(&tcph->check, skb, tcph->seq, newseq, 0);
be0ea7d5 PM	430	inet_proto_csum_replace4(&tcph->check, skb, tcph->ack_seq, newack, 0);
5b1158e9	431
0d53778e PM	432	pr_debug("Adjusting sequence number from %u->%u, ack from %u->%u\n",
	433	ntohl(tcph->seq), ntohl(newseq), ntohl(tcph->ack_seq),
	434	ntohl(newack));
5b1158e9 JK	435
	436	tcph->seq = newseq;
	437	tcph->ack_seq = newack;
	438
f9dd09c7	439	return nf_nat_sack_adjust(skb, tcph, ct, ctinfo);
5b1158e9	440	}
5b1158e9 JK	441
	442	/* Setup NAT on this expected conntrack so it follows master. */
	443	/* If we fail to get a free NAT slot, we'll get dropped on confirm */
	444	void nf_nat_follow_master(struct nf_conn *ct,
	445	struct nf_conntrack_expect *exp)
	446	{
	447	struct nf_nat_range range;
	448
	449	/* This must be a fresh one. */
	450	BUG_ON(ct->status & IPS_NAT_DONE_MASK);
	451
	452	/* Change src to where master sends to */
	453	range.flags = IP_NAT_RANGE_MAP_IPS;
	454	range.min_ip = range.max_ip
	455	= ct->master->tuplehash[!exp->dir].tuple.dst.u3.ip;
cc01dcbd	456	nf_nat_setup_info(ct, &range, IP_NAT_MANIP_SRC);
5b1158e9 JK	457
	458	/* For DST manip, map port here to where it's expected. */
	459	range.flags = (IP_NAT_RANGE_MAP_IPS \| IP_NAT_RANGE_PROTO_SPECIFIED);
	460	range.min = range.max = exp->saved_proto;
	461	range.min_ip = range.max_ip
	462	= ct->master->tuplehash[!exp->dir].tuple.src.u3.ip;
cc01dcbd	463	nf_nat_setup_info(ct, &range, IP_NAT_MANIP_DST);
5b1158e9 JK	464	}
5b1158e9 JK	465	EXPORT_SYMBOL(nf_nat_follow_master);