[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/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 <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)
{
	int dir;
	struct nf_nat_seq *this_way, *other_way;
	struct nf_conn_nat *nat = nfct_nat(ct);

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

	dir = CTINFO2DIR(ctinfo);

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

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

/* 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;
	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);
		/* Tell TCP window tracking about seq change */
		nf_conntrack_tcp_update(skb, ip_hdrlen(skb),
					ct, CTINFO2DIR(ctinfo));

		nf_conntrack_event_cache(IPCT_NATSEQADJ, skb);
	}
	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;
	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;
	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))
		newseq = htonl(ntohl(tcph->seq) + this_way->offset_after);
	else
		newseq = htonl(ntohl(tcph->seq) + this_way->offset_before);

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

	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;

	if (!nf_nat_sack_adjust(skb, tcph, ct, ctinfo))
		return 0;

	nf_conntrack_tcp_update(skb, ip_hdrlen(skb), ct, dir);

	return 1;
}
EXPORT_SYMBOL(nf_nat_seq_adjust);

/* 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>
	19
	20	#include <linux/netfilter_ipv4.h>
	21	#include <net/netfilter/nf_conntrack.h>
	22	#include <net/netfilter/nf_conntrack_helper.h>
13eae15a	23	#include <net/netfilter/nf_conntrack_ecache.h>
5b1158e9 JK	24	#include <net/netfilter/nf_conntrack_expect.h>
	25	#include <net/netfilter/nf_nat.h>
	26	#include <net/netfilter/nf_nat_protocol.h>
	27	#include <net/netfilter/nf_nat_core.h>
	28	#include <net/netfilter/nf_nat_helper.h>
	29
0d53778e PM	30	#define DUMP_OFFSET(x) \
	31	pr_debug("offset_before=%d, offset_after=%d, correction_pos=%u\n", \
	32	x->offset_before, x->offset_after, x->correction_pos);
5b1158e9 JK	33
	34	static DEFINE_SPINLOCK(nf_nat_seqofs_lock);
	35
	36	/* Setup TCP sequence correction given this change at this sequence */
	37	static inline void
	38	adjust_tcp_sequence(u32 seq,
	39	int sizediff,
	40	struct nf_conn *ct,
	41	enum ip_conntrack_info ctinfo)
	42	{
	43	int dir;
	44	struct nf_nat_seq this_way, other_way;
	45	struct nf_conn_nat *nat = nfct_nat(ct);
	46
969d7108	47	pr_debug("adjust_tcp_sequence: seq = %u, sizediff = %d\n", seq, seq);
5b1158e9 JK	48
	49	dir = CTINFO2DIR(ctinfo);
	50
b6b84d4a YK	51	this_way = &nat->seq[dir];
b6b84d4a YK	52	other_way = &nat->seq[!dir];
5b1158e9	53
0d53778e	54	pr_debug("nf_nat_resize_packet: Seq_offset before: ");
5b1158e9 JK	55	DUMP_OFFSET(this_way);
	56
	57	spin_lock_bh(&nf_nat_seqofs_lock);
	58
	59	/* SYN adjust. If it's uninitialized, or this is after last
	60	* correction, record it: we don't handle more than one
	61	* adjustment in the window, but do deal with common case of a
	62	* retransmit */
	63	if (this_way->offset_before == this_way->offset_after \|\|
	64	before(this_way->correction_pos, seq)) {
	65	this_way->correction_pos = seq;
	66	this_way->offset_before = this_way->offset_after;
	67	this_way->offset_after += sizediff;
	68	}
	69	spin_unlock_bh(&nf_nat_seqofs_lock);
	70
0d53778e	71	pr_debug("nf_nat_resize_packet: Seq_offset after: ");
5b1158e9 JK	72	DUMP_OFFSET(this_way);
	73	}
	74
	75	/* Frobs data inside this packet, which is linear. */
	76	static void mangle_contents(struct sk_buff *skb,
	77	unsigned int dataoff,
	78	unsigned int match_offset,
	79	unsigned int match_len,
	80	const char *rep_buffer,
	81	unsigned int rep_len)
	82	{
	83	unsigned char *data;
	84
	85	BUG_ON(skb_is_nonlinear(skb));
eddc9ec5	86	data = skb_network_header(skb) + dataoff;
5b1158e9 JK	87
	88	/* move post-replacement */
	89	memmove(data + match_offset + rep_len,
	90	data + match_offset + match_len,
27a884dc ACM	91	skb->tail - (skb->network_header + dataoff +
27a884dc ACM	92	match_offset + match_len));
5b1158e9 JK	93
	94	/* insert data from buffer */
	95	memcpy(data + match_offset, rep_buffer, rep_len);
	96
	97	/* update skb info */
	98	if (rep_len > match_len) {
0d53778e PM	99	pr_debug("nf_nat_mangle_packet: Extending packet by "
0d53778e PM	100	"%u from %u bytes\n", rep_len - match_len, skb->len);
5b1158e9 JK	101	skb_put(skb, rep_len - match_len);
5b1158e9 JK	102	} else {
0d53778e PM	103	pr_debug("nf_nat_mangle_packet: Shrinking packet from "
0d53778e PM	104	"%u from %u bytes\n", match_len - rep_len, skb->len);
5b1158e9 JK	105	__skb_trim(skb, skb->len + rep_len - match_len);
	106	}
	107
	108	/* fix IP hdr checksum information */
eddc9ec5 ACM	109	ip_hdr(skb)->tot_len = htons(skb->len);
eddc9ec5 ACM	110	ip_send_check(ip_hdr(skb));
5b1158e9 JK	111	}
	112
	113	/* Unusual, but possible case. */
3db05fea	114	static int enlarge_skb(struct sk_buff *skb, unsigned int extra)
5b1158e9	115	{
3db05fea	116	if (skb->len + extra > 65535)
5b1158e9 JK	117	return 0;
5b1158e9 JK	118
3db05fea	119	if (pskb_expand_head(skb, 0, extra - skb_tailroom(skb), GFP_ATOMIC))
5b1158e9 JK	120	return 0;
5b1158e9 JK	121
5b1158e9 JK	122	return 1;
	123	}
	124
	125	/* Generic function for mangling variable-length address changes inside
	126	* NATed TCP connections (like the PORT XXX,XXX,XXX,XXX,XXX,XXX
	127	* command in FTP).
	128	*
	129	* Takes care about all the nasty sequence number changes, checksumming,
	130	* skb enlargement, ...
	131	*
	132	* */
	133	int
3db05fea	134	nf_nat_mangle_tcp_packet(struct sk_buff *skb,
5b1158e9 JK	135	struct nf_conn *ct,
	136	enum ip_conntrack_info ctinfo,
	137	unsigned int match_offset,
	138	unsigned int match_len,
	139	const char *rep_buffer,
	140	unsigned int rep_len)
	141	{
ee6b9673	142	struct rtable *rt = skb->rtable;
5b1158e9 JK	143	struct iphdr *iph;
	144	struct tcphdr *tcph;
	145	int oldlen, datalen;
	146
3db05fea	147	if (!skb_make_writable(skb, skb->len))
5b1158e9 JK	148	return 0;
	149
	150	if (rep_len > match_len &&
3db05fea HX	151	rep_len - match_len > skb_tailroom(skb) &&
3db05fea HX	152	!enlarge_skb(skb, rep_len - match_len))
5b1158e9 JK	153	return 0;
5b1158e9 JK	154
3db05fea	155	SKB_LINEAR_ASSERT(skb);
5b1158e9	156
3db05fea	157	iph = ip_hdr(skb);
5b1158e9 JK	158	tcph = (void )iph + iph->ihl4;
5b1158e9 JK	159
3db05fea HX	160	oldlen = skb->len - iph->ihl*4;
3db05fea HX	161	mangle_contents(skb, iph->ihl4 + tcph->doff4,
5b1158e9 JK	162	match_offset, match_len, rep_buffer, rep_len);
5b1158e9 JK	163
3db05fea HX	164	datalen = skb->len - iph->ihl*4;
3db05fea HX	165	if (skb->ip_summed != CHECKSUM_PARTIAL) {
fe6092ea	166	if (!(rt->rt_flags & RTCF_LOCAL) &&
3db05fea HX	167	skb->dev->features & NETIF_F_V4_CSUM) {
	168	skb->ip_summed = CHECKSUM_PARTIAL;
	169	skb->csum_start = skb_headroom(skb) +
	170	skb_network_offset(skb) +
	171	iph->ihl * 4;
	172	skb->csum_offset = offsetof(struct tcphdr, check);
fe6092ea PM	173	tcph->check = ~tcp_v4_check(datalen,
	174	iph->saddr, iph->daddr, 0);
	175	} else {
	176	tcph->check = 0;
	177	tcph->check = tcp_v4_check(datalen,
	178	iph->saddr, iph->daddr,
a47362a2	179	csum_partial(tcph,
fe6092ea PM	180	datalen, 0));
fe6092ea PM	181	}
5b1158e9	182	} else
be0ea7d5 PM	183	inet_proto_csum_replace2(&tcph->check, skb,
be0ea7d5 PM	184	htons(oldlen), htons(datalen), 1);
5b1158e9 JK	185
	186	if (rep_len != match_len) {
	187	set_bit(IPS_SEQ_ADJUST_BIT, &ct->status);
	188	adjust_tcp_sequence(ntohl(tcph->seq),
	189	(int)rep_len - (int)match_len,
	190	ct, ctinfo);
	191	/* Tell TCP window tracking about seq change */
3db05fea	192	nf_conntrack_tcp_update(skb, ip_hdrlen(skb),
5b1158e9	193	ct, CTINFO2DIR(ctinfo));
13eae15a PNA	194
13eae15a PNA	195	nf_conntrack_event_cache(IPCT_NATSEQADJ, skb);
5b1158e9 JK	196	}
	197	return 1;
	198	}
	199	EXPORT_SYMBOL(nf_nat_mangle_tcp_packet);
	200
	201	/* Generic function for mangling variable-length address changes inside
	202	* NATed UDP connections (like the CONNECT DATA XXXXX MESG XXXXX INDEX XXXXX
	203	* command in the Amanda protocol)
	204	*
	205	* Takes care about all the nasty sequence number changes, checksumming,
	206	* skb enlargement, ...
	207	*
	208	* XXX - This function could be merged with nf_nat_mangle_tcp_packet which
	209	* should be fairly easy to do.
	210	*/
	211	int
3db05fea	212	nf_nat_mangle_udp_packet(struct sk_buff *skb,
5b1158e9 JK	213	struct nf_conn *ct,
	214	enum ip_conntrack_info ctinfo,
	215	unsigned int match_offset,
	216	unsigned int match_len,
	217	const char *rep_buffer,
	218	unsigned int rep_len)
	219	{
ee6b9673	220	struct rtable *rt = skb->rtable;
5b1158e9 JK	221	struct iphdr *iph;
	222	struct udphdr *udph;
	223	int datalen, oldlen;
	224
	225	/* UDP helpers might accidentally mangle the wrong packet */
3db05fea HX	226	iph = ip_hdr(skb);
3db05fea HX	227	if (skb->len < iph->ihl4 + sizeof(udph) +
e905a9ed	228	match_offset + match_len)
5b1158e9 JK	229	return 0;
5b1158e9 JK	230
3db05fea	231	if (!skb_make_writable(skb, skb->len))
5b1158e9 JK	232	return 0;
	233
	234	if (rep_len > match_len &&
3db05fea HX	235	rep_len - match_len > skb_tailroom(skb) &&
3db05fea HX	236	!enlarge_skb(skb, rep_len - match_len))
5b1158e9 JK	237	return 0;
5b1158e9 JK	238
3db05fea	239	iph = ip_hdr(skb);
5b1158e9 JK	240	udph = (void )iph + iph->ihl4;
5b1158e9 JK	241
3db05fea HX	242	oldlen = skb->len - iph->ihl*4;
3db05fea HX	243	mangle_contents(skb, iph->ihl4 + sizeof(udph),
5b1158e9 JK	244	match_offset, match_len, rep_buffer, rep_len);
	245
	246	/* update the length of the UDP packet */
3db05fea	247	datalen = skb->len - iph->ihl*4;
5b1158e9 JK	248	udph->len = htons(datalen);
	249
	250	/* fix udp checksum if udp checksum was previously calculated */
3db05fea	251	if (!udph->check && skb->ip_summed != CHECKSUM_PARTIAL)
5b1158e9 JK	252	return 1;
5b1158e9 JK	253
3db05fea	254	if (skb->ip_summed != CHECKSUM_PARTIAL) {
fe6092ea	255	if (!(rt->rt_flags & RTCF_LOCAL) &&
3db05fea HX	256	skb->dev->features & NETIF_F_V4_CSUM) {
	257	skb->ip_summed = CHECKSUM_PARTIAL;
	258	skb->csum_start = skb_headroom(skb) +
	259	skb_network_offset(skb) +
	260	iph->ihl * 4;
	261	skb->csum_offset = offsetof(struct udphdr, check);
fe6092ea PM	262	udph->check = ~csum_tcpudp_magic(iph->saddr, iph->daddr,
	263	datalen, IPPROTO_UDP,
	264	0);
	265	} else {
	266	udph->check = 0;
	267	udph->check = csum_tcpudp_magic(iph->saddr, iph->daddr,
	268	datalen, IPPROTO_UDP,
a47362a2	269	csum_partial(udph,
fe6092ea PM	270	datalen, 0));
	271	if (!udph->check)
	272	udph->check = CSUM_MANGLED_0;
	273	}
5b1158e9	274	} else
be0ea7d5 PM	275	inet_proto_csum_replace2(&udph->check, skb,
be0ea7d5 PM	276	htons(oldlen), htons(datalen), 1);
5b1158e9 JK	277
	278	return 1;
	279	}
	280	EXPORT_SYMBOL(nf_nat_mangle_udp_packet);
	281
	282	/* Adjust one found SACK option including checksum correction */
	283	static void
	284	sack_adjust(struct sk_buff *skb,
	285	struct tcphdr *tcph,
	286	unsigned int sackoff,
	287	unsigned int sackend,
	288	struct nf_nat_seq *natseq)
	289	{
	290	while (sackoff < sackend) {
	291	struct tcp_sack_block_wire *sack;
	292	__be32 new_start_seq, new_end_seq;
	293
	294	sack = (void *)skb->data + sackoff;
	295	if (after(ntohl(sack->start_seq) - natseq->offset_before,
	296	natseq->correction_pos))
	297	new_start_seq = htonl(ntohl(sack->start_seq)
	298	- natseq->offset_after);
	299	else
	300	new_start_seq = htonl(ntohl(sack->start_seq)
	301	- natseq->offset_before);
	302
	303	if (after(ntohl(sack->end_seq) - natseq->offset_before,
	304	natseq->correction_pos))
	305	new_end_seq = htonl(ntohl(sack->end_seq)
	306	- natseq->offset_after);
	307	else
	308	new_end_seq = htonl(ntohl(sack->end_seq)
	309	- natseq->offset_before);
	310
0d53778e PM	311	pr_debug("sack_adjust: start_seq: %d->%d, end_seq: %d->%d\n",
	312	ntohl(sack->start_seq), new_start_seq,
	313	ntohl(sack->end_seq), new_end_seq);
5b1158e9	314
be0ea7d5 PM	315	inet_proto_csum_replace4(&tcph->check, skb,
	316	sack->start_seq, new_start_seq, 0);
	317	inet_proto_csum_replace4(&tcph->check, skb,
	318	sack->end_seq, new_end_seq, 0);
5b1158e9 JK	319	sack->start_seq = new_start_seq;
	320	sack->end_seq = new_end_seq;
	321	sackoff += sizeof(*sack);
	322	}
	323	}
	324
	325	/* TCP SACK sequence number adjustment */
	326	static inline unsigned int
3db05fea	327	nf_nat_sack_adjust(struct sk_buff *skb,
5b1158e9 JK	328	struct tcphdr *tcph,
	329	struct nf_conn *ct,
	330	enum ip_conntrack_info ctinfo)
	331	{
	332	unsigned int dir, optoff, optend;
	333	struct nf_conn_nat *nat = nfct_nat(ct);
	334
3db05fea HX	335	optoff = ip_hdrlen(skb) + sizeof(struct tcphdr);
3db05fea HX	336	optend = ip_hdrlen(skb) + tcph->doff * 4;
5b1158e9	337
3db05fea	338	if (!skb_make_writable(skb, optend))
5b1158e9 JK	339	return 0;
	340
	341	dir = CTINFO2DIR(ctinfo);
	342
	343	while (optoff < optend) {
	344	/* Usually: option, length. */
3db05fea	345	unsigned char *op = skb->data + optoff;
5b1158e9 JK	346
	347	switch (op[0]) {
	348	case TCPOPT_EOL:
	349	return 1;
	350	case TCPOPT_NOP:
	351	optoff++;
	352	continue;
	353	default:
	354	/* no partial options */
	355	if (optoff + 1 == optend \|\|
	356	optoff + op[1] > optend \|\|
	357	op[1] < 2)
	358	return 0;
	359	if (op[0] == TCPOPT_SACK &&
	360	op[1] >= 2+TCPOLEN_SACK_PERBLOCK &&
	361	((op[1] - 2) % TCPOLEN_SACK_PERBLOCK) == 0)
3db05fea	362	sack_adjust(skb, tcph, optoff+2,
b6b84d4a	363	optoff+op[1], &nat->seq[!dir]);
5b1158e9 JK	364	optoff += op[1];
	365	}
	366	}
	367	return 1;
	368	}
	369
	370	/* TCP sequence number adjustment. Returns 1 on success, 0 on failure */
	371	int
3db05fea	372	nf_nat_seq_adjust(struct sk_buff *skb,
5b1158e9 JK	373	struct nf_conn *ct,
	374	enum ip_conntrack_info ctinfo)
	375	{
	376	struct tcphdr *tcph;
	377	int dir;
	378	__be32 newseq, newack;
	379	struct nf_conn_nat *nat = nfct_nat(ct);
	380	struct nf_nat_seq this_way, other_way;
	381
	382	dir = CTINFO2DIR(ctinfo);
	383
b6b84d4a YK	384	this_way = &nat->seq[dir];
b6b84d4a YK	385	other_way = &nat->seq[!dir];
5b1158e9	386
3db05fea	387	if (!skb_make_writable(skb, ip_hdrlen(skb) + sizeof(*tcph)))
5b1158e9 JK	388	return 0;
5b1158e9 JK	389
3db05fea	390	tcph = (void *)skb->data + ip_hdrlen(skb);
5b1158e9 JK	391	if (after(ntohl(tcph->seq), this_way->correction_pos))
	392	newseq = htonl(ntohl(tcph->seq) + this_way->offset_after);
	393	else
	394	newseq = htonl(ntohl(tcph->seq) + this_way->offset_before);
	395
	396	if (after(ntohl(tcph->ack_seq) - other_way->offset_before,
	397	other_way->correction_pos))
	398	newack = htonl(ntohl(tcph->ack_seq) - other_way->offset_after);
	399	else
	400	newack = htonl(ntohl(tcph->ack_seq) - other_way->offset_before);
	401
be0ea7d5 PM	402	inet_proto_csum_replace4(&tcph->check, skb, tcph->seq, newseq, 0);
be0ea7d5 PM	403	inet_proto_csum_replace4(&tcph->check, skb, tcph->ack_seq, newack, 0);
5b1158e9	404
0d53778e PM	405	pr_debug("Adjusting sequence number from %u->%u, ack from %u->%u\n",
	406	ntohl(tcph->seq), ntohl(newseq), ntohl(tcph->ack_seq),
	407	ntohl(newack));
5b1158e9 JK	408
	409	tcph->seq = newseq;
	410	tcph->ack_seq = newack;
	411
3db05fea	412	if (!nf_nat_sack_adjust(skb, tcph, ct, ctinfo))
5b1158e9 JK	413	return 0;
5b1158e9 JK	414
3db05fea	415	nf_conntrack_tcp_update(skb, ip_hdrlen(skb), ct, dir);
5b1158e9 JK	416
	417	return 1;
	418	}
	419	EXPORT_SYMBOL(nf_nat_seq_adjust);
	420
	421	/* Setup NAT on this expected conntrack so it follows master. */
	422	/* If we fail to get a free NAT slot, we'll get dropped on confirm */
	423	void nf_nat_follow_master(struct nf_conn *ct,
	424	struct nf_conntrack_expect *exp)
	425	{
	426	struct nf_nat_range range;
	427
	428	/* This must be a fresh one. */
	429	BUG_ON(ct->status & IPS_NAT_DONE_MASK);
	430
	431	/* Change src to where master sends to */
	432	range.flags = IP_NAT_RANGE_MAP_IPS;
	433	range.min_ip = range.max_ip
	434	= ct->master->tuplehash[!exp->dir].tuple.dst.u3.ip;
cc01dcbd	435	nf_nat_setup_info(ct, &range, IP_NAT_MANIP_SRC);
5b1158e9 JK	436
	437	/* For DST manip, map port here to where it's expected. */
	438	range.flags = (IP_NAT_RANGE_MAP_IPS \| IP_NAT_RANGE_PROTO_SPECIFIED);
	439	range.min = range.max = exp->saved_proto;
	440	range.min_ip = range.max_ip
	441	= ct->master->tuplehash[!exp->dir].tuple.src.u3.ip;
cc01dcbd	442	nf_nat_setup_info(ct, &range, IP_NAT_MANIP_DST);
5b1158e9 JK	443	}
5b1158e9 JK	444	EXPORT_SYMBOL(nf_nat_follow_master);