[mirror_ubuntu-bionic-kernel.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);

static void nf_nat_csum(struct sk_buff *skb, const struct iphdr *iph, void *data,
			int datalen, __sum16 *check, int oldlen)
{
	struct rtable *rt = skb_rtable(skb);

	if (skb->ip_summed != CHECKSUM_PARTIAL) {
		if (!(rt->rt_flags & RTCF_LOCAL) &&
		    (!skb->dev || 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 = (void *)check - data;
			*check = ~csum_tcpudp_magic(iph->saddr, iph->daddr,
						    datalen, iph->protocol, 0);
		} else {
			*check = 0;
			*check = csum_tcpudp_magic(iph->saddr, iph->daddr,
						   datalen, iph->protocol,
						   csum_partial(data, datalen,
								0));
			if (iph->protocol == IPPROTO_UDP && !*check)
				*check = CSUM_MANGLED_0;
		}
	} else
		inet_proto_csum_replace2(check, skb,
					 htons(oldlen), htons(datalen), 1);
}

/* 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 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;
	nf_nat_csum(skb, iph, tcph, datalen, &tcph->check, oldlen);

	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 iphdr *iph;
	struct udphdr *udph;
	int datalen, oldlen;

	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;

	nf_nat_csum(skb, iph, udph, datalen, &udph->check, oldlen);

	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_ipv4_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 = NF_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, NF_NAT_MANIP_SRC);

	/* For DST manip, map port here to where it's expected. */
	range.flags = (NF_NAT_RANGE_MAP_IPS | NF_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, NF_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
b71d1d42	156	static void nf_nat_csum(struct sk_buff skb, const struct iphdr iph, void *data,
0abee526 CG	157	int datalen, __sum16 *check, int oldlen)
	158	{
	159	struct rtable *rt = skb_rtable(skb);
	160
	161	if (skb->ip_summed != CHECKSUM_PARTIAL) {
	162	if (!(rt->rt_flags & RTCF_LOCAL) &&
d9be76f3	163	(!skb->dev \|\| skb->dev->features & NETIF_F_V4_CSUM)) {
0abee526 CG	164	skb->ip_summed = CHECKSUM_PARTIAL;
	165	skb->csum_start = skb_headroom(skb) +
	166	skb_network_offset(skb) +
	167	iph->ihl * 4;
	168	skb->csum_offset = (void *)check - data;
	169	*check = ~csum_tcpudp_magic(iph->saddr, iph->daddr,
	170	datalen, iph->protocol, 0);
	171	} else {
	172	*check = 0;
	173	*check = csum_tcpudp_magic(iph->saddr, iph->daddr,
	174	datalen, iph->protocol,
	175	csum_partial(data, datalen,
	176	0));
	177	if (iph->protocol == IPPROTO_UDP && !*check)
	178	*check = CSUM_MANGLED_0;
	179	}
	180	} else
	181	inet_proto_csum_replace2(check, skb,
	182	htons(oldlen), htons(datalen), 1);
	183	}
	184
5b1158e9 JK	185	/* Generic function for mangling variable-length address changes inside
	186	* NATed TCP connections (like the PORT XXX,XXX,XXX,XXX,XXX,XXX
	187	* command in FTP).
	188	*
	189	* Takes care about all the nasty sequence number changes, checksumming,
	190	* skb enlargement, ...
	191	*
	192	* */
010c0b9f PM	193	int __nf_nat_mangle_tcp_packet(struct sk_buff *skb,
	194	struct nf_conn *ct,
	195	enum ip_conntrack_info ctinfo,
	196	unsigned int match_offset,
	197	unsigned int match_len,
	198	const char *rep_buffer,
	199	unsigned int rep_len, bool adjust)
5b1158e9 JK	200	{
	201	struct iphdr *iph;
	202	struct tcphdr *tcph;
	203	int oldlen, datalen;
	204
3db05fea	205	if (!skb_make_writable(skb, skb->len))
5b1158e9 JK	206	return 0;
	207
	208	if (rep_len > match_len &&
3db05fea HX	209	rep_len - match_len > skb_tailroom(skb) &&
3db05fea HX	210	!enlarge_skb(skb, rep_len - match_len))
5b1158e9 JK	211	return 0;
5b1158e9 JK	212
3db05fea	213	SKB_LINEAR_ASSERT(skb);
5b1158e9	214
3db05fea	215	iph = ip_hdr(skb);
5b1158e9 JK	216	tcph = (void )iph + iph->ihl4;
5b1158e9 JK	217
3db05fea HX	218	oldlen = skb->len - iph->ihl*4;
3db05fea HX	219	mangle_contents(skb, iph->ihl4 + tcph->doff4,
5b1158e9 JK	220	match_offset, match_len, rep_buffer, rep_len);
5b1158e9 JK	221
3db05fea	222	datalen = skb->len - iph->ihl*4;
0abee526	223	nf_nat_csum(skb, iph, tcph, datalen, &tcph->check, oldlen);
5b1158e9	224
010c0b9f PM	225	if (adjust && rep_len != match_len)
	226	nf_nat_set_seq_adjust(ct, ctinfo, tcph->seq,
	227	(int)rep_len - (int)match_len);
	228
5b1158e9 JK	229	return 1;
5b1158e9 JK	230	}
010c0b9f	231	EXPORT_SYMBOL(__nf_nat_mangle_tcp_packet);
5b1158e9 JK	232
	233	/* Generic function for mangling variable-length address changes inside
	234	* NATed UDP connections (like the CONNECT DATA XXXXX MESG XXXXX INDEX XXXXX
	235	* command in the Amanda protocol)
	236	*
	237	* Takes care about all the nasty sequence number changes, checksumming,
	238	* skb enlargement, ...
	239	*
	240	* XXX - This function could be merged with nf_nat_mangle_tcp_packet which
	241	* should be fairly easy to do.
	242	*/
	243	int
3db05fea	244	nf_nat_mangle_udp_packet(struct sk_buff *skb,
5b1158e9 JK	245	struct nf_conn *ct,
	246	enum ip_conntrack_info ctinfo,
	247	unsigned int match_offset,
	248	unsigned int match_len,
	249	const char *rep_buffer,
	250	unsigned int rep_len)
	251	{
	252	struct iphdr *iph;
	253	struct udphdr *udph;
	254	int datalen, oldlen;
	255
3db05fea	256	if (!skb_make_writable(skb, skb->len))
5b1158e9 JK	257	return 0;
	258
	259	if (rep_len > match_len &&
3db05fea HX	260	rep_len - match_len > skb_tailroom(skb) &&
3db05fea HX	261	!enlarge_skb(skb, rep_len - match_len))
5b1158e9 JK	262	return 0;
5b1158e9 JK	263
3db05fea	264	iph = ip_hdr(skb);
5b1158e9 JK	265	udph = (void )iph + iph->ihl4;
5b1158e9 JK	266
3db05fea HX	267	oldlen = skb->len - iph->ihl*4;
3db05fea HX	268	mangle_contents(skb, iph->ihl4 + sizeof(udph),
5b1158e9 JK	269	match_offset, match_len, rep_buffer, rep_len);
	270
	271	/* update the length of the UDP packet */
3db05fea	272	datalen = skb->len - iph->ihl*4;
5b1158e9 JK	273	udph->len = htons(datalen);
	274
	275	/* fix udp checksum if udp checksum was previously calculated */
3db05fea	276	if (!udph->check && skb->ip_summed != CHECKSUM_PARTIAL)
5b1158e9 JK	277	return 1;
5b1158e9 JK	278
0abee526	279	nf_nat_csum(skb, iph, udph, datalen, &udph->check, oldlen);
5b1158e9 JK	280
	281	return 1;
	282	}
	283	EXPORT_SYMBOL(nf_nat_mangle_udp_packet);
	284
	285	/* Adjust one found SACK option including checksum correction */
	286	static void
	287	sack_adjust(struct sk_buff *skb,
	288	struct tcphdr *tcph,
	289	unsigned int sackoff,
	290	unsigned int sackend,
	291	struct nf_nat_seq *natseq)
	292	{
	293	while (sackoff < sackend) {
	294	struct tcp_sack_block_wire *sack;
	295	__be32 new_start_seq, new_end_seq;
	296
	297	sack = (void *)skb->data + sackoff;
	298	if (after(ntohl(sack->start_seq) - natseq->offset_before,
	299	natseq->correction_pos))
	300	new_start_seq = htonl(ntohl(sack->start_seq)
	301	- natseq->offset_after);
	302	else
	303	new_start_seq = htonl(ntohl(sack->start_seq)
	304	- natseq->offset_before);
	305
	306	if (after(ntohl(sack->end_seq) - natseq->offset_before,
	307	natseq->correction_pos))
	308	new_end_seq = htonl(ntohl(sack->end_seq)
	309	- natseq->offset_after);
	310	else
	311	new_end_seq = htonl(ntohl(sack->end_seq)
	312	- natseq->offset_before);
	313
0d53778e PM	314	pr_debug("sack_adjust: start_seq: %d->%d, end_seq: %d->%d\n",
	315	ntohl(sack->start_seq), new_start_seq,
	316	ntohl(sack->end_seq), new_end_seq);
5b1158e9	317
be0ea7d5 PM	318	inet_proto_csum_replace4(&tcph->check, skb,
	319	sack->start_seq, new_start_seq, 0);
	320	inet_proto_csum_replace4(&tcph->check, skb,
	321	sack->end_seq, new_end_seq, 0);
5b1158e9 JK	322	sack->start_seq = new_start_seq;
	323	sack->end_seq = new_end_seq;
	324	sackoff += sizeof(*sack);
	325	}
	326	}
	327
	328	/* TCP SACK sequence number adjustment */
	329	static inline unsigned int
3db05fea	330	nf_nat_sack_adjust(struct sk_buff *skb,
5b1158e9 JK	331	struct tcphdr *tcph,
	332	struct nf_conn *ct,
	333	enum ip_conntrack_info ctinfo)
	334	{
	335	unsigned int dir, optoff, optend;
	336	struct nf_conn_nat *nat = nfct_nat(ct);
	337
3db05fea HX	338	optoff = ip_hdrlen(skb) + sizeof(struct tcphdr);
3db05fea HX	339	optend = ip_hdrlen(skb) + tcph->doff * 4;
5b1158e9	340
3db05fea	341	if (!skb_make_writable(skb, optend))
5b1158e9 JK	342	return 0;
	343
	344	dir = CTINFO2DIR(ctinfo);
	345
	346	while (optoff < optend) {
	347	/* Usually: option, length. */
3db05fea	348	unsigned char *op = skb->data + optoff;
5b1158e9 JK	349
	350	switch (op[0]) {
	351	case TCPOPT_EOL:
	352	return 1;
	353	case TCPOPT_NOP:
	354	optoff++;
	355	continue;
	356	default:
	357	/* no partial options */
	358	if (optoff + 1 == optend \|\|
	359	optoff + op[1] > optend \|\|
	360	op[1] < 2)
	361	return 0;
	362	if (op[0] == TCPOPT_SACK &&
	363	op[1] >= 2+TCPOLEN_SACK_PERBLOCK &&
	364	((op[1] - 2) % TCPOLEN_SACK_PERBLOCK) == 0)
3db05fea	365	sack_adjust(skb, tcph, optoff+2,
b6b84d4a	366	optoff+op[1], &nat->seq[!dir]);
5b1158e9 JK	367	optoff += op[1];
	368	}
	369	}
	370	return 1;
	371	}
	372
	373	/* TCP sequence number adjustment. Returns 1 on success, 0 on failure */
	374	int
3db05fea	375	nf_nat_seq_adjust(struct sk_buff *skb,
5b1158e9 JK	376	struct nf_conn *ct,
	377	enum ip_conntrack_info ctinfo)
	378	{
	379	struct tcphdr *tcph;
	380	int dir;
	381	__be32 newseq, newack;
a3a9f79e	382	s16 seqoff, ackoff;
5b1158e9 JK	383	struct nf_conn_nat *nat = nfct_nat(ct);
	384	struct nf_nat_seq this_way, other_way;
	385
	386	dir = CTINFO2DIR(ctinfo);
	387
b6b84d4a YK	388	this_way = &nat->seq[dir];
b6b84d4a YK	389	other_way = &nat->seq[!dir];
5b1158e9	390
3db05fea	391	if (!skb_make_writable(skb, ip_hdrlen(skb) + sizeof(*tcph)))
5b1158e9 JK	392	return 0;
5b1158e9 JK	393
3db05fea	394	tcph = (void *)skb->data + ip_hdrlen(skb);
5b1158e9	395	if (after(ntohl(tcph->seq), this_way->correction_pos))
a3a9f79e	396	seqoff = this_way->offset_after;
5b1158e9	397	else
a3a9f79e	398	seqoff = this_way->offset_before;
5b1158e9 JK	399
	400	if (after(ntohl(tcph->ack_seq) - other_way->offset_before,
	401	other_way->correction_pos))
a3a9f79e	402	ackoff = other_way->offset_after;
5b1158e9	403	else
a3a9f79e PM	404	ackoff = other_way->offset_before;
	405
	406	newseq = htonl(ntohl(tcph->seq) + seqoff);
	407	newack = htonl(ntohl(tcph->ack_seq) - ackoff);
5b1158e9	408
be0ea7d5 PM	409	inet_proto_csum_replace4(&tcph->check, skb, tcph->seq, newseq, 0);
be0ea7d5 PM	410	inet_proto_csum_replace4(&tcph->check, skb, tcph->ack_seq, newack, 0);
5b1158e9	411
0d53778e PM	412	pr_debug("Adjusting sequence number from %u->%u, ack from %u->%u\n",
	413	ntohl(tcph->seq), ntohl(newseq), ntohl(tcph->ack_seq),
	414	ntohl(newack));
5b1158e9 JK	415
	416	tcph->seq = newseq;
	417	tcph->ack_seq = newack;
	418
f9dd09c7	419	return nf_nat_sack_adjust(skb, tcph, ct, ctinfo);
5b1158e9	420	}
5b1158e9 JK	421
	422	/* Setup NAT on this expected conntrack so it follows master. */
	423	/* If we fail to get a free NAT slot, we'll get dropped on confirm */
	424	void nf_nat_follow_master(struct nf_conn *ct,
	425	struct nf_conntrack_expect *exp)
	426	{
cbc9f2f4	427	struct nf_nat_ipv4_range range;
5b1158e9 JK	428
	429	/* This must be a fresh one. */
	430	BUG_ON(ct->status & IPS_NAT_DONE_MASK);
	431
	432	/* Change src to where master sends to */
cbc9f2f4	433	range.flags = NF_NAT_RANGE_MAP_IPS;
5b1158e9 JK	434	range.min_ip = range.max_ip
5b1158e9 JK	435	= ct->master->tuplehash[!exp->dir].tuple.dst.u3.ip;
cbc9f2f4	436	nf_nat_setup_info(ct, &range, NF_NAT_MANIP_SRC);
5b1158e9 JK	437
5b1158e9 JK	438	/* For DST manip, map port here to where it's expected. */
cbc9f2f4	439	range.flags = (NF_NAT_RANGE_MAP_IPS \| NF_NAT_RANGE_PROTO_SPECIFIED);
5b1158e9 JK	440	range.min = range.max = exp->saved_proto;
	441	range.min_ip = range.max_ip
	442	= ct->master->tuplehash[!exp->dir].tuple.src.u3.ip;
cbc9f2f4	443	nf_nat_setup_info(ct, &range, NF_NAT_MANIP_DST);
5b1158e9 JK	444	}
5b1158e9 JK	445	EXPORT_SYMBOL(nf_nat_follow_master);