[deliverable/linux.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/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;
}

/* 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)
{
	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 (rep_len != match_len) {
		set_bit(IPS_SEQ_ADJUST_BIT, &ct->status);
		adjust_tcp_sequence(ntohl(tcph->seq),
				    (int)rep_len - (int)match_len,
				    ct, ctinfo);
		nf_conntrack_event_cache(IPCT_NATSEQADJ, ct);
	}
	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>
	11	#include <linux/kmod.h>
	12	#include <linux/types.h>
	13	#include <linux/timer.h>
	14	#include <linux/skbuff.h>
	15	#include <linux/tcp.h>
	16	#include <linux/udp.h>
	17	#include <net/checksum.h>
	18	#include <net/tcp.h>
1668e010	19	#include <net/route.h>
5b1158e9 JK	20
	21	#include <linux/netfilter_ipv4.h>
	22	#include <net/netfilter/nf_conntrack.h>
	23	#include <net/netfilter/nf_conntrack_helper.h>
13eae15a	24	#include <net/netfilter/nf_conntrack_ecache.h>
5b1158e9 JK	25	#include <net/netfilter/nf_conntrack_expect.h>
	26	#include <net/netfilter/nf_nat.h>
	27	#include <net/netfilter/nf_nat_protocol.h>
	28	#include <net/netfilter/nf_nat_core.h>
	29	#include <net/netfilter/nf_nat_helper.h>
	30
0d53778e PM	31	#define DUMP_OFFSET(x) \
	32	pr_debug("offset_before=%d, offset_after=%d, correction_pos=%u\n", \
	33	x->offset_before, x->offset_after, x->correction_pos);
5b1158e9 JK	34
	35	static DEFINE_SPINLOCK(nf_nat_seqofs_lock);
	36
	37	/* Setup TCP sequence correction given this change at this sequence */
	38	static inline void
	39	adjust_tcp_sequence(u32 seq,
	40	int sizediff,
	41	struct nf_conn *ct,
	42	enum ip_conntrack_info ctinfo)
	43	{
76ac8940	44	enum ip_conntrack_dir dir = CTINFO2DIR(ctinfo);
5b1158e9	45	struct nf_conn_nat *nat = nfct_nat(ct);
76ac8940	46	struct nf_nat_seq *this_way = &nat->seq[dir];
5b1158e9	47
76ac8940 HE	48	pr_debug("adjust_tcp_sequence: seq = %u, sizediff = %d\n",
76ac8940 HE	49	seq, sizediff);
5b1158e9	50
76ac8940	51	pr_debug("adjust_tcp_sequence: Seq_offset before: ");
5b1158e9 JK	52	DUMP_OFFSET(this_way);
	53
	54	spin_lock_bh(&nf_nat_seqofs_lock);
	55
	56	/* SYN adjust. If it's uninitialized, or this is after last
	57	* correction, record it: we don't handle more than one
	58	* adjustment in the window, but do deal with common case of a
	59	* retransmit */
	60	if (this_way->offset_before == this_way->offset_after \|\|
	61	before(this_way->correction_pos, seq)) {
76ac8940 HE	62	this_way->correction_pos = seq;
	63	this_way->offset_before = this_way->offset_after;
	64	this_way->offset_after += sizediff;
5b1158e9 JK	65	}
	66	spin_unlock_bh(&nf_nat_seqofs_lock);
	67
76ac8940	68	pr_debug("adjust_tcp_sequence: Seq_offset after: ");
5b1158e9 JK	69	DUMP_OFFSET(this_way);
	70	}
	71
f9dd09c7 JK	72	/* Get the offset value, for conntrack */
	73	s16 nf_nat_get_offset(const struct nf_conn *ct,
	74	enum ip_conntrack_dir dir,
	75	u32 seq)
	76	{
	77	struct nf_conn_nat *nat = nfct_nat(ct);
	78	struct nf_nat_seq *this_way;
	79	s16 offset;
	80
	81	if (!nat)
	82	return 0;
	83
	84	this_way = &nat->seq[dir];
	85	spin_lock_bh(&nf_nat_seqofs_lock);
	86	offset = after(seq, this_way->correction_pos)
	87	? this_way->offset_after : this_way->offset_before;
	88	spin_unlock_bh(&nf_nat_seqofs_lock);
	89
	90	return offset;
	91	}
	92	EXPORT_SYMBOL_GPL(nf_nat_get_offset);
	93
5b1158e9 JK	94	/* Frobs data inside this packet, which is linear. */
	95	static void mangle_contents(struct sk_buff *skb,
	96	unsigned int dataoff,
	97	unsigned int match_offset,
	98	unsigned int match_len,
	99	const char *rep_buffer,
	100	unsigned int rep_len)
	101	{
	102	unsigned char *data;
	103
	104	BUG_ON(skb_is_nonlinear(skb));
eddc9ec5	105	data = skb_network_header(skb) + dataoff;
5b1158e9 JK	106
	107	/* move post-replacement */
	108	memmove(data + match_offset + rep_len,
	109	data + match_offset + match_len,
27a884dc ACM	110	skb->tail - (skb->network_header + dataoff +
27a884dc ACM	111	match_offset + match_len));
5b1158e9 JK	112
	113	/* insert data from buffer */
	114	memcpy(data + match_offset, rep_buffer, rep_len);
	115
	116	/* update skb info */
	117	if (rep_len > match_len) {
0d53778e PM	118	pr_debug("nf_nat_mangle_packet: Extending packet by "
0d53778e PM	119	"%u from %u bytes\n", rep_len - match_len, skb->len);
5b1158e9 JK	120	skb_put(skb, rep_len - match_len);
5b1158e9 JK	121	} else {
0d53778e PM	122	pr_debug("nf_nat_mangle_packet: Shrinking packet from "
0d53778e PM	123	"%u from %u bytes\n", match_len - rep_len, skb->len);
5b1158e9 JK	124	__skb_trim(skb, skb->len + rep_len - match_len);
	125	}
	126
	127	/* fix IP hdr checksum information */
eddc9ec5 ACM	128	ip_hdr(skb)->tot_len = htons(skb->len);
eddc9ec5 ACM	129	ip_send_check(ip_hdr(skb));
5b1158e9 JK	130	}
	131
	132	/* Unusual, but possible case. */
3db05fea	133	static int enlarge_skb(struct sk_buff *skb, unsigned int extra)
5b1158e9	134	{
3db05fea	135	if (skb->len + extra > 65535)
5b1158e9 JK	136	return 0;
5b1158e9 JK	137
3db05fea	138	if (pskb_expand_head(skb, 0, extra - skb_tailroom(skb), GFP_ATOMIC))
5b1158e9 JK	139	return 0;
5b1158e9 JK	140
5b1158e9 JK	141	return 1;
	142	}
	143
	144	/* Generic function for mangling variable-length address changes inside
	145	* NATed TCP connections (like the PORT XXX,XXX,XXX,XXX,XXX,XXX
	146	* command in FTP).
	147	*
	148	* Takes care about all the nasty sequence number changes, checksumming,
	149	* skb enlargement, ...
	150	*
	151	* */
	152	int
3db05fea	153	nf_nat_mangle_tcp_packet(struct sk_buff *skb,
5b1158e9 JK	154	struct nf_conn *ct,
	155	enum ip_conntrack_info ctinfo,
	156	unsigned int match_offset,
	157	unsigned int match_len,
	158	const char *rep_buffer,
	159	unsigned int rep_len)
	160	{
511c3f92	161	struct rtable *rt = skb_rtable(skb);
5b1158e9 JK	162	struct iphdr *iph;
	163	struct tcphdr *tcph;
	164	int oldlen, datalen;
	165
3db05fea	166	if (!skb_make_writable(skb, skb->len))
5b1158e9 JK	167	return 0;
	168
	169	if (rep_len > match_len &&
3db05fea HX	170	rep_len - match_len > skb_tailroom(skb) &&
3db05fea HX	171	!enlarge_skb(skb, rep_len - match_len))
5b1158e9 JK	172	return 0;
5b1158e9 JK	173
3db05fea	174	SKB_LINEAR_ASSERT(skb);
5b1158e9	175
3db05fea	176	iph = ip_hdr(skb);
5b1158e9 JK	177	tcph = (void )iph + iph->ihl4;
5b1158e9 JK	178
3db05fea HX	179	oldlen = skb->len - iph->ihl*4;
3db05fea HX	180	mangle_contents(skb, iph->ihl4 + tcph->doff4,
5b1158e9 JK	181	match_offset, match_len, rep_buffer, rep_len);
5b1158e9 JK	182
3db05fea HX	183	datalen = skb->len - iph->ihl*4;
3db05fea HX	184	if (skb->ip_summed != CHECKSUM_PARTIAL) {
fe6092ea	185	if (!(rt->rt_flags & RTCF_LOCAL) &&
3db05fea HX	186	skb->dev->features & NETIF_F_V4_CSUM) {
	187	skb->ip_summed = CHECKSUM_PARTIAL;
	188	skb->csum_start = skb_headroom(skb) +
	189	skb_network_offset(skb) +
	190	iph->ihl * 4;
	191	skb->csum_offset = offsetof(struct tcphdr, check);
fe6092ea PM	192	tcph->check = ~tcp_v4_check(datalen,
	193	iph->saddr, iph->daddr, 0);
	194	} else {
	195	tcph->check = 0;
	196	tcph->check = tcp_v4_check(datalen,
	197	iph->saddr, iph->daddr,
a47362a2	198	csum_partial(tcph,
fe6092ea PM	199	datalen, 0));
fe6092ea PM	200	}
5b1158e9	201	} else
be0ea7d5 PM	202	inet_proto_csum_replace2(&tcph->check, skb,
be0ea7d5 PM	203	htons(oldlen), htons(datalen), 1);
5b1158e9 JK	204
	205	if (rep_len != match_len) {
	206	set_bit(IPS_SEQ_ADJUST_BIT, &ct->status);
	207	adjust_tcp_sequence(ntohl(tcph->seq),
	208	(int)rep_len - (int)match_len,
	209	ct, ctinfo);
a71996fc	210	nf_conntrack_event_cache(IPCT_NATSEQADJ, ct);
5b1158e9 JK	211	}
	212	return 1;
	213	}
	214	EXPORT_SYMBOL(nf_nat_mangle_tcp_packet);
	215
	216	/* Generic function for mangling variable-length address changes inside
	217	* NATed UDP connections (like the CONNECT DATA XXXXX MESG XXXXX INDEX XXXXX
	218	* command in the Amanda protocol)
	219	*
	220	* Takes care about all the nasty sequence number changes, checksumming,
	221	* skb enlargement, ...
	222	*
	223	* XXX - This function could be merged with nf_nat_mangle_tcp_packet which
	224	* should be fairly easy to do.
	225	*/
	226	int
3db05fea	227	nf_nat_mangle_udp_packet(struct sk_buff *skb,
5b1158e9 JK	228	struct nf_conn *ct,
	229	enum ip_conntrack_info ctinfo,
	230	unsigned int match_offset,
	231	unsigned int match_len,
	232	const char *rep_buffer,
	233	unsigned int rep_len)
	234	{
511c3f92	235	struct rtable *rt = skb_rtable(skb);
5b1158e9 JK	236	struct iphdr *iph;
	237	struct udphdr *udph;
	238	int datalen, oldlen;
	239
	240	/* UDP helpers might accidentally mangle the wrong packet */
3db05fea HX	241	iph = ip_hdr(skb);
3db05fea HX	242	if (skb->len < iph->ihl4 + sizeof(udph) +
e905a9ed	243	match_offset + match_len)
5b1158e9 JK	244	return 0;
5b1158e9 JK	245
3db05fea	246	if (!skb_make_writable(skb, skb->len))
5b1158e9 JK	247	return 0;
	248
	249	if (rep_len > match_len &&
3db05fea HX	250	rep_len - match_len > skb_tailroom(skb) &&
3db05fea HX	251	!enlarge_skb(skb, rep_len - match_len))
5b1158e9 JK	252	return 0;
5b1158e9 JK	253
3db05fea	254	iph = ip_hdr(skb);
5b1158e9 JK	255	udph = (void )iph + iph->ihl4;
5b1158e9 JK	256
3db05fea HX	257	oldlen = skb->len - iph->ihl*4;
3db05fea HX	258	mangle_contents(skb, iph->ihl4 + sizeof(udph),
5b1158e9 JK	259	match_offset, match_len, rep_buffer, rep_len);
	260
	261	/* update the length of the UDP packet */
3db05fea	262	datalen = skb->len - iph->ihl*4;
5b1158e9 JK	263	udph->len = htons(datalen);
	264
	265	/* fix udp checksum if udp checksum was previously calculated */
3db05fea	266	if (!udph->check && skb->ip_summed != CHECKSUM_PARTIAL)
5b1158e9 JK	267	return 1;
5b1158e9 JK	268
3db05fea	269	if (skb->ip_summed != CHECKSUM_PARTIAL) {
fe6092ea	270	if (!(rt->rt_flags & RTCF_LOCAL) &&
3db05fea HX	271	skb->dev->features & NETIF_F_V4_CSUM) {
	272	skb->ip_summed = CHECKSUM_PARTIAL;
	273	skb->csum_start = skb_headroom(skb) +
	274	skb_network_offset(skb) +
	275	iph->ihl * 4;
	276	skb->csum_offset = offsetof(struct udphdr, check);
fe6092ea PM	277	udph->check = ~csum_tcpudp_magic(iph->saddr, iph->daddr,
	278	datalen, IPPROTO_UDP,
	279	0);
	280	} else {
	281	udph->check = 0;
	282	udph->check = csum_tcpudp_magic(iph->saddr, iph->daddr,
	283	datalen, IPPROTO_UDP,
a47362a2	284	csum_partial(udph,
fe6092ea PM	285	datalen, 0));
	286	if (!udph->check)
	287	udph->check = CSUM_MANGLED_0;
	288	}
5b1158e9	289	} else
be0ea7d5 PM	290	inet_proto_csum_replace2(&udph->check, skb,
be0ea7d5 PM	291	htons(oldlen), htons(datalen), 1);
5b1158e9 JK	292
	293	return 1;
	294	}
	295	EXPORT_SYMBOL(nf_nat_mangle_udp_packet);
	296
	297	/* Adjust one found SACK option including checksum correction */
	298	static void
	299	sack_adjust(struct sk_buff *skb,
	300	struct tcphdr *tcph,
	301	unsigned int sackoff,
	302	unsigned int sackend,
	303	struct nf_nat_seq *natseq)
	304	{
	305	while (sackoff < sackend) {
	306	struct tcp_sack_block_wire *sack;
	307	__be32 new_start_seq, new_end_seq;
	308
	309	sack = (void *)skb->data + sackoff;
	310	if (after(ntohl(sack->start_seq) - natseq->offset_before,
	311	natseq->correction_pos))
	312	new_start_seq = htonl(ntohl(sack->start_seq)
	313	- natseq->offset_after);
	314	else
	315	new_start_seq = htonl(ntohl(sack->start_seq)
	316	- natseq->offset_before);
	317
	318	if (after(ntohl(sack->end_seq) - natseq->offset_before,
	319	natseq->correction_pos))
	320	new_end_seq = htonl(ntohl(sack->end_seq)
	321	- natseq->offset_after);
	322	else
	323	new_end_seq = htonl(ntohl(sack->end_seq)
	324	- natseq->offset_before);
	325
0d53778e PM	326	pr_debug("sack_adjust: start_seq: %d->%d, end_seq: %d->%d\n",
	327	ntohl(sack->start_seq), new_start_seq,
	328	ntohl(sack->end_seq), new_end_seq);
5b1158e9	329
be0ea7d5 PM	330	inet_proto_csum_replace4(&tcph->check, skb,
	331	sack->start_seq, new_start_seq, 0);
	332	inet_proto_csum_replace4(&tcph->check, skb,
	333	sack->end_seq, new_end_seq, 0);
5b1158e9 JK	334	sack->start_seq = new_start_seq;
	335	sack->end_seq = new_end_seq;
	336	sackoff += sizeof(*sack);
	337	}
	338	}
	339
	340	/* TCP SACK sequence number adjustment */
	341	static inline unsigned int
3db05fea	342	nf_nat_sack_adjust(struct sk_buff *skb,
5b1158e9 JK	343	struct tcphdr *tcph,
	344	struct nf_conn *ct,
	345	enum ip_conntrack_info ctinfo)
	346	{
	347	unsigned int dir, optoff, optend;
	348	struct nf_conn_nat *nat = nfct_nat(ct);
	349
3db05fea HX	350	optoff = ip_hdrlen(skb) + sizeof(struct tcphdr);
3db05fea HX	351	optend = ip_hdrlen(skb) + tcph->doff * 4;
5b1158e9	352
3db05fea	353	if (!skb_make_writable(skb, optend))
5b1158e9 JK	354	return 0;
	355
	356	dir = CTINFO2DIR(ctinfo);
	357
	358	while (optoff < optend) {
	359	/* Usually: option, length. */
3db05fea	360	unsigned char *op = skb->data + optoff;
5b1158e9 JK	361
	362	switch (op[0]) {
	363	case TCPOPT_EOL:
	364	return 1;
	365	case TCPOPT_NOP:
	366	optoff++;
	367	continue;
	368	default:
	369	/* no partial options */
	370	if (optoff + 1 == optend \|\|
	371	optoff + op[1] > optend \|\|
	372	op[1] < 2)
	373	return 0;
	374	if (op[0] == TCPOPT_SACK &&
	375	op[1] >= 2+TCPOLEN_SACK_PERBLOCK &&
	376	((op[1] - 2) % TCPOLEN_SACK_PERBLOCK) == 0)
3db05fea	377	sack_adjust(skb, tcph, optoff+2,
b6b84d4a	378	optoff+op[1], &nat->seq[!dir]);
5b1158e9 JK	379	optoff += op[1];
	380	}
	381	}
	382	return 1;
	383	}
	384
	385	/* TCP sequence number adjustment. Returns 1 on success, 0 on failure */
	386	int
3db05fea	387	nf_nat_seq_adjust(struct sk_buff *skb,
5b1158e9 JK	388	struct nf_conn *ct,
	389	enum ip_conntrack_info ctinfo)
	390	{
	391	struct tcphdr *tcph;
	392	int dir;
	393	__be32 newseq, newack;
a3a9f79e	394	s16 seqoff, ackoff;
5b1158e9 JK	395	struct nf_conn_nat *nat = nfct_nat(ct);
	396	struct nf_nat_seq this_way, other_way;
	397
	398	dir = CTINFO2DIR(ctinfo);
	399
b6b84d4a YK	400	this_way = &nat->seq[dir];
b6b84d4a YK	401	other_way = &nat->seq[!dir];
5b1158e9	402
3db05fea	403	if (!skb_make_writable(skb, ip_hdrlen(skb) + sizeof(*tcph)))
5b1158e9 JK	404	return 0;
5b1158e9 JK	405
3db05fea	406	tcph = (void *)skb->data + ip_hdrlen(skb);
5b1158e9	407	if (after(ntohl(tcph->seq), this_way->correction_pos))
a3a9f79e	408	seqoff = this_way->offset_after;
5b1158e9	409	else
a3a9f79e	410	seqoff = this_way->offset_before;
5b1158e9 JK	411
	412	if (after(ntohl(tcph->ack_seq) - other_way->offset_before,
	413	other_way->correction_pos))
a3a9f79e	414	ackoff = other_way->offset_after;
5b1158e9	415	else
a3a9f79e PM	416	ackoff = other_way->offset_before;
	417
	418	newseq = htonl(ntohl(tcph->seq) + seqoff);
	419	newack = htonl(ntohl(tcph->ack_seq) - ackoff);
5b1158e9	420
be0ea7d5 PM	421	inet_proto_csum_replace4(&tcph->check, skb, tcph->seq, newseq, 0);
be0ea7d5 PM	422	inet_proto_csum_replace4(&tcph->check, skb, tcph->ack_seq, newack, 0);
5b1158e9	423
0d53778e PM	424	pr_debug("Adjusting sequence number from %u->%u, ack from %u->%u\n",
	425	ntohl(tcph->seq), ntohl(newseq), ntohl(tcph->ack_seq),
	426	ntohl(newack));
5b1158e9 JK	427
	428	tcph->seq = newseq;
	429	tcph->ack_seq = newack;
	430
f9dd09c7	431	return nf_nat_sack_adjust(skb, tcph, ct, ctinfo);
5b1158e9	432	}
5b1158e9 JK	433
	434	/* Setup NAT on this expected conntrack so it follows master. */
	435	/* If we fail to get a free NAT slot, we'll get dropped on confirm */
	436	void nf_nat_follow_master(struct nf_conn *ct,
	437	struct nf_conntrack_expect *exp)
	438	{
	439	struct nf_nat_range range;
	440
	441	/* This must be a fresh one. */
	442	BUG_ON(ct->status & IPS_NAT_DONE_MASK);
	443
	444	/* Change src to where master sends to */
	445	range.flags = IP_NAT_RANGE_MAP_IPS;
	446	range.min_ip = range.max_ip
	447	= ct->master->tuplehash[!exp->dir].tuple.dst.u3.ip;
cc01dcbd	448	nf_nat_setup_info(ct, &range, IP_NAT_MANIP_SRC);
5b1158e9 JK	449
	450	/* For DST manip, map port here to where it's expected. */
	451	range.flags = (IP_NAT_RANGE_MAP_IPS \| IP_NAT_RANGE_PROTO_SPECIFIED);
	452	range.min = range.max = exp->saved_proto;
	453	range.min_ip = range.max_ip
	454	= ct->master->tuplehash[!exp->dir].tuple.src.u3.ip;
cc01dcbd	455	nf_nat_setup_info(ct, &range, IP_NAT_MANIP_DST);
5b1158e9 JK	456	}
5b1158e9 JK	457	EXPORT_SYMBOL(nf_nat_follow_master);