[net-next-2.6.git] / net / ipv4 / syncookies.c

/*
 *  Syncookies implementation for the Linux kernel
 *
 *  Copyright (C) 1997 Andi Kleen
 *  Based on ideas by D.J.Bernstein and Eric Schenk.
 *
 *	This program is free software; you can redistribute it and/or
 *      modify it under the terms of the GNU General Public License
 *      as published by the Free Software Foundation; either version
 *      2 of the License, or (at your option) any later version.
 */

#include <linux/tcp.h>
#include <linux/slab.h>
#include <linux/random.h>
#include <linux/cryptohash.h>
#include <linux/kernel.h>
#include <net/tcp.h>
#include <net/route.h>

/* Timestamps: lowest 9 bits store TCP options */
#define TSBITS 9
#define TSMASK (((__u32)1 << TSBITS) - 1)

extern int sysctl_tcp_syncookies;

__u32 syncookie_secret[2][16-4+SHA_DIGEST_WORDS];
EXPORT_SYMBOL(syncookie_secret);

static __init int init_syncookies(void)
{
	get_random_bytes(syncookie_secret, sizeof(syncookie_secret));
	return 0;
}
__initcall(init_syncookies);

#define COOKIEBITS 24	/* Upper bits store count */
#define COOKIEMASK (((__u32)1 << COOKIEBITS) - 1)

static DEFINE_PER_CPU(__u32 [16 + 5 + SHA_WORKSPACE_WORDS],
		      ipv4_cookie_scratch);

static u32 cookie_hash(__be32 saddr, __be32 daddr, __be16 sport, __be16 dport,
		       u32 count, int c)
{
	__u32 *tmp = __get_cpu_var(ipv4_cookie_scratch);

	memcpy(tmp + 4, syncookie_secret[c], sizeof(syncookie_secret[c]));
	tmp[0] = (__force u32)saddr;
	tmp[1] = (__force u32)daddr;
	tmp[2] = ((__force u32)sport << 16) + (__force u32)dport;
	tmp[3] = count;
	sha_transform(tmp + 16, (__u8 *)tmp, tmp + 16 + 5);

	return tmp[17];
}


/*
 * when syncookies are in effect and tcp timestamps are enabled we encode
 * tcp options in the lowest 9 bits of the timestamp value that will be
 * sent in the syn-ack.
 * Since subsequent timestamps use the normal tcp_time_stamp value, we
 * must make sure that the resulting initial timestamp is <= tcp_time_stamp.
 */
__u32 cookie_init_timestamp(struct request_sock *req)
{
	struct inet_request_sock *ireq;
	u32 ts, ts_now = tcp_time_stamp;
	u32 options = 0;

	ireq = inet_rsk(req);
	if (ireq->wscale_ok) {
		options = ireq->snd_wscale;
		options |= ireq->rcv_wscale << 4;
	}
	options |= ireq->sack_ok << 8;

	ts = ts_now & ~TSMASK;
	ts |= options;
	if (ts > ts_now) {
		ts >>= TSBITS;
		ts--;
		ts <<= TSBITS;
		ts |= options;
	}
	return ts;
}


static __u32 secure_tcp_syn_cookie(__be32 saddr, __be32 daddr, __be16 sport,
				   __be16 dport, __u32 sseq, __u32 count,
				   __u32 data)
{
	/*
	 * Compute the secure sequence number.
	 * The output should be:
	 *   HASH(sec1,saddr,sport,daddr,dport,sec1) + sseq + (count * 2^24)
	 *      + (HASH(sec2,saddr,sport,daddr,dport,count,sec2) % 2^24).
	 * Where sseq is their sequence number and count increases every
	 * minute by 1.
	 * As an extra hack, we add a small "data" value that encodes the
	 * MSS into the second hash value.
	 */

	return (cookie_hash(saddr, daddr, sport, dport, 0, 0) +
		sseq + (count << COOKIEBITS) +
		((cookie_hash(saddr, daddr, sport, dport, count, 1) + data)
		 & COOKIEMASK));
}

/*
 * This retrieves the small "data" value from the syncookie.
 * If the syncookie is bad, the data returned will be out of
 * range.  This must be checked by the caller.
 *
 * The count value used to generate the cookie must be within
 * "maxdiff" if the current (passed-in) "count".  The return value
 * is (__u32)-1 if this test fails.
 */
static __u32 check_tcp_syn_cookie(__u32 cookie, __be32 saddr, __be32 daddr,
				  __be16 sport, __be16 dport, __u32 sseq,
				  __u32 count, __u32 maxdiff)
{
	__u32 diff;

	/* Strip away the layers from the cookie */
	cookie -= cookie_hash(saddr, daddr, sport, dport, 0, 0) + sseq;

	/* Cookie is now reduced to (count * 2^24) ^ (hash % 2^24) */
	diff = (count - (cookie >> COOKIEBITS)) & ((__u32) - 1 >> COOKIEBITS);
	if (diff >= maxdiff)
		return (__u32)-1;

	return (cookie -
		cookie_hash(saddr, daddr, sport, dport, count - diff, 1))
		& COOKIEMASK;	/* Leaving the data behind */
}

/*
 * This table has to be sorted and terminated with (__u16)-1.
 * XXX generate a better table.
 * Unresolved Issues: HIPPI with a 64k MSS is not well supported.
 */
static __u16 const msstab[] = {
	64 - 1,
	256 - 1,
	512 - 1,
	536 - 1,
	1024 - 1,
	1440 - 1,
	1460 - 1,
	4312 - 1,
	(__u16)-1
};
/* The number doesn't include the -1 terminator */
#define NUM_MSS (ARRAY_SIZE(msstab) - 1)

/*
 * Generate a syncookie.  mssp points to the mss, which is returned
 * rounded down to the value encoded in the cookie.
 */
__u32 cookie_v4_init_sequence(struct sock *sk, struct sk_buff *skb, __u16 *mssp)
{
	const struct iphdr *iph = ip_hdr(skb);
	const struct tcphdr *th = tcp_hdr(skb);
	int mssind;
	const __u16 mss = *mssp;

	tcp_synq_overflow(sk);

	/* XXX sort msstab[] by probability?  Binary search? */
	for (mssind = 0; mss > msstab[mssind + 1]; mssind++)
		;
	*mssp = msstab[mssind] + 1;

	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_SYNCOOKIESSENT);

	return secure_tcp_syn_cookie(iph->saddr, iph->daddr,
				     th->source, th->dest, ntohl(th->seq),
				     jiffies / (HZ * 60), mssind);
}

/*
 * This (misnamed) value is the age of syncookie which is permitted.
 * Its ideal value should be dependent on TCP_TIMEOUT_INIT and
 * sysctl_tcp_retries1. It's a rather complicated formula (exponential
 * backoff) to compute at runtime so it's currently hardcoded here.
 */
#define COUNTER_TRIES 4
/*
 * Check if a ack sequence number is a valid syncookie.
 * Return the decoded mss if it is, or 0 if not.
 */
static inline int cookie_check(struct sk_buff *skb, __u32 cookie)
{
	const struct iphdr *iph = ip_hdr(skb);
	const struct tcphdr *th = tcp_hdr(skb);
	__u32 seq = ntohl(th->seq) - 1;
	__u32 mssind = check_tcp_syn_cookie(cookie, iph->saddr, iph->daddr,
					    th->source, th->dest, seq,
					    jiffies / (HZ * 60),
					    COUNTER_TRIES);

	return mssind < NUM_MSS ? msstab[mssind] + 1 : 0;
}

static inline struct sock *get_cookie_sock(struct sock *sk, struct sk_buff *skb,
					   struct request_sock *req,
					   struct dst_entry *dst)
{
	struct inet_connection_sock *icsk = inet_csk(sk);
	struct sock *child;

	child = icsk->icsk_af_ops->syn_recv_sock(sk, skb, req, dst);
	if (child)
		inet_csk_reqsk_queue_add(sk, req, child);
	else
		reqsk_free(req);

	return child;
}


/*
 * when syncookies are in effect and tcp timestamps are enabled we stored
 * additional tcp options in the timestamp.
 * This extracts these options from the timestamp echo.
 *
 * The lowest 4 bits are for snd_wscale
 * The next 4 lsb are for rcv_wscale
 * The next lsb is for sack_ok
 */
void cookie_check_timestamp(struct tcp_options_received *tcp_opt)
{
	/* echoed timestamp, 9 lowest bits contain options */
	u32 options = tcp_opt->rcv_tsecr & TSMASK;

	tcp_opt->snd_wscale = options & 0xf;
	options >>= 4;
	tcp_opt->rcv_wscale = options & 0xf;

	tcp_opt->sack_ok = (options >> 4) & 0x1;

	if (tcp_opt->sack_ok)
		tcp_sack_reset(tcp_opt);

	if (tcp_opt->snd_wscale || tcp_opt->rcv_wscale)
		tcp_opt->wscale_ok = 1;
}
EXPORT_SYMBOL(cookie_check_timestamp);

struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb,
			     struct ip_options *opt)
{
	struct tcp_options_received tcp_opt;
	u8 *hash_location;
	struct inet_request_sock *ireq;
	struct tcp_request_sock *treq;
	struct tcp_sock *tp = tcp_sk(sk);
	const struct tcphdr *th = tcp_hdr(skb);
	__u32 cookie = ntohl(th->ack_seq) - 1;
	struct sock *ret = sk;
	struct request_sock *req;
	int mss;
	struct rtable *rt;
	__u8 rcv_wscale;

	if (!sysctl_tcp_syncookies || !th->ack)
		goto out;

	if (tcp_synq_no_recent_overflow(sk) ||
	    (mss = cookie_check(skb, cookie)) == 0) {
		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_SYNCOOKIESFAILED);
		goto out;
	}

	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_SYNCOOKIESRECV);

	/* check for timestamp cookie support */
	memset(&tcp_opt, 0, sizeof(tcp_opt));
	tcp_parse_options(skb, &tcp_opt, &hash_location, 0);

	if (tcp_opt.saw_tstamp)
		cookie_check_timestamp(&tcp_opt);

	ret = NULL;
	req = inet_reqsk_alloc(&tcp_request_sock_ops); /* for safety */
	if (!req)
		goto out;

	ireq = inet_rsk(req);
	treq = tcp_rsk(req);
	treq->rcv_isn		= ntohl(th->seq) - 1;
	treq->snt_isn		= cookie;
	req->mss		= mss;
	ireq->loc_port		= th->dest;
	ireq->rmt_port		= th->source;
	ireq->loc_addr		= ip_hdr(skb)->daddr;
	ireq->rmt_addr		= ip_hdr(skb)->saddr;
	ireq->ecn_ok		= 0;
	ireq->snd_wscale	= tcp_opt.snd_wscale;
	ireq->rcv_wscale	= tcp_opt.rcv_wscale;
	ireq->sack_ok		= tcp_opt.sack_ok;
	ireq->wscale_ok		= tcp_opt.wscale_ok;
	ireq->tstamp_ok		= tcp_opt.saw_tstamp;
	req->ts_recent		= tcp_opt.saw_tstamp ? tcp_opt.rcv_tsval : 0;

	/* We throwed the options of the initial SYN away, so we hope
	 * the ACK carries the same options again (see RFC1122 4.2.3.8)
	 */
	if (opt && opt->optlen) {
		int opt_size = sizeof(struct ip_options) + opt->optlen;

		ireq->opt = kmalloc(opt_size, GFP_ATOMIC);
		if (ireq->opt != NULL && ip_options_echo(ireq->opt, skb)) {
			kfree(ireq->opt);
			ireq->opt = NULL;
		}
	}

	if (security_inet_conn_request(sk, skb, req)) {
		reqsk_free(req);
		goto out;
	}

	req->expires	= 0UL;
	req->retrans	= 0;

	/*
	 * We need to lookup the route here to get at the correct
	 * window size. We should better make sure that the window size
	 * hasn't changed since we received the original syn, but I see
	 * no easy way to do this.
	 */
	{
		struct flowi fl = { .mark = sk->sk_mark,
				    .nl_u = { .ip4_u =
					      { .daddr = ((opt && opt->srr) ?
							  opt->faddr :
							  ireq->rmt_addr),
						.saddr = ireq->loc_addr,
						.tos = RT_CONN_FLAGS(sk) } },
				    .proto = IPPROTO_TCP,
				    .flags = inet_sk_flowi_flags(sk),
				    .uli_u = { .ports =
					       { .sport = th->dest,
						 .dport = th->source } } };
		security_req_classify_flow(req, &fl);
		if (ip_route_output_key(&init_net, &rt, &fl)) {
			reqsk_free(req);
			goto out;
		}
	}

	/* Try to redo what tcp_v4_send_synack did. */
	req->window_clamp = tp->window_clamp ? :dst_metric(&rt->u.dst, RTAX_WINDOW);

	tcp_select_initial_window(tcp_full_space(sk), req->mss,
				  &req->rcv_wnd, &req->window_clamp,
				  ireq->wscale_ok, &rcv_wscale,
				  dst_metric(&rt->u.dst, RTAX_INITRWND));

	ireq->rcv_wscale  = rcv_wscale;

	ret = get_cookie_sock(sk, skb, req, &rt->u.dst);
out:	return ret;
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* Syncookies implementation for the Linux kernel
	3	*
	4	* Copyright (C) 1997 Andi Kleen
e905a9ed	5	* Based on ideas by D.J.Bernstein and Eric Schenk.
1da177e4 LT	6	*
	7	* This program is free software; you can redistribute it and/or
	8	* modify it under the terms of the GNU General Public License
	9	* as published by the Free Software Foundation; either version
	10	* 2 of the License, or (at your option) any later version.
1da177e4 LT	11	*/
	12
	13	#include <linux/tcp.h>
	14	#include <linux/slab.h>
	15	#include <linux/random.h>
	16	#include <linux/cryptohash.h>
	17	#include <linux/kernel.h>
	18	#include <net/tcp.h>
86b08d86	19	#include <net/route.h>
1da177e4	20
4dfc2817 FW	21	/* Timestamps: lowest 9 bits store TCP options */
	22	#define TSBITS 9
	23	#define TSMASK (((__u32)1 << TSBITS) - 1)
	24
1da177e4 LT	25	extern int sysctl_tcp_syncookies;
1da177e4 LT	26
2051f11f	27	__u32 syncookie_secret[2][16-4+SHA_DIGEST_WORDS];
c6aefafb	28	EXPORT_SYMBOL(syncookie_secret);
1da177e4 LT	29
	30	static __init int init_syncookies(void)
	31	{
	32	get_random_bytes(syncookie_secret, sizeof(syncookie_secret));
	33	return 0;
	34	}
c6aefafb	35	__initcall(init_syncookies);
1da177e4 LT	36
	37	#define COOKIEBITS 24 /* Upper bits store count */
	38	#define COOKIEMASK (((__u32)1 << COOKIEBITS) - 1)
	39
245b2e70 TH	40	static DEFINE_PER_CPU(__u32 [16 + 5 + SHA_WORKSPACE_WORDS],
245b2e70 TH	41	ipv4_cookie_scratch);
11baab7a	42
714e85be	43	static u32 cookie_hash(__be32 saddr, __be32 daddr, __be16 sport, __be16 dport,
1da177e4 LT	44	u32 count, int c)
1da177e4 LT	45	{
245b2e70	46	__u32 *tmp = __get_cpu_var(ipv4_cookie_scratch);
1da177e4	47
2051f11f	48	memcpy(tmp + 4, syncookie_secret[c], sizeof(syncookie_secret[c]));
714e85be AV	49	tmp[0] = (__force u32)saddr;
	50	tmp[1] = (__force u32)daddr;
	51	tmp[2] = ((__force u32)sport << 16) + (__force u32)dport;
1da177e4 LT	52	tmp[3] = count;
	53	sha_transform(tmp + 16, (__u8 *)tmp, tmp + 16 + 5);
	54
	55	return tmp[17];
	56	}
	57
4dfc2817 FW	58
	59	/*
	60	* when syncookies are in effect and tcp timestamps are enabled we encode
	61	* tcp options in the lowest 9 bits of the timestamp value that will be
	62	* sent in the syn-ack.
	63	* Since subsequent timestamps use the normal tcp_time_stamp value, we
	64	* must make sure that the resulting initial timestamp is <= tcp_time_stamp.
	65	*/
	66	__u32 cookie_init_timestamp(struct request_sock *req)
	67	{
	68	struct inet_request_sock *ireq;
	69	u32 ts, ts_now = tcp_time_stamp;
	70	u32 options = 0;
	71
	72	ireq = inet_rsk(req);
	73	if (ireq->wscale_ok) {
	74	options = ireq->snd_wscale;
	75	options \|= ireq->rcv_wscale << 4;
	76	}
	77	options \|= ireq->sack_ok << 8;
	78
	79	ts = ts_now & ~TSMASK;
	80	ts \|= options;
	81	if (ts > ts_now) {
	82	ts >>= TSBITS;
	83	ts--;
	84	ts <<= TSBITS;
	85	ts \|= options;
	86	}
	87	return ts;
	88	}
	89
	90
714e85be AV	91	static __u32 secure_tcp_syn_cookie(__be32 saddr, __be32 daddr, __be16 sport,
714e85be AV	92	__be16 dport, __u32 sseq, __u32 count,
1da177e4 LT	93	__u32 data)
	94	{
	95	/*
	96	* Compute the secure sequence number.
	97	* The output should be:
e905a9ed	98	* HASH(sec1,saddr,sport,daddr,dport,sec1) + sseq + (count * 2^24)
1da177e4 LT	99	* + (HASH(sec2,saddr,sport,daddr,dport,count,sec2) % 2^24).
	100	* Where sseq is their sequence number and count increases every
	101	* minute by 1.
	102	* As an extra hack, we add a small "data" value that encodes the
	103	* MSS into the second hash value.
	104	*/
	105
	106	return (cookie_hash(saddr, daddr, sport, dport, 0, 0) +
	107	sseq + (count << COOKIEBITS) +
	108	((cookie_hash(saddr, daddr, sport, dport, count, 1) + data)
	109	& COOKIEMASK));
	110	}
	111
	112	/*
	113	* This retrieves the small "data" value from the syncookie.
	114	* If the syncookie is bad, the data returned will be out of
	115	* range. This must be checked by the caller.
	116	*
	117	* The count value used to generate the cookie must be within
	118	* "maxdiff" if the current (passed-in) "count". The return value
	119	* is (__u32)-1 if this test fails.
	120	*/
714e85be AV	121	static __u32 check_tcp_syn_cookie(__u32 cookie, __be32 saddr, __be32 daddr,
714e85be AV	122	__be16 sport, __be16 dport, __u32 sseq,
1da177e4 LT	123	__u32 count, __u32 maxdiff)
	124	{
	125	__u32 diff;
	126
	127	/* Strip away the layers from the cookie */
	128	cookie -= cookie_hash(saddr, daddr, sport, dport, 0, 0) + sseq;
	129
	130	/* Cookie is now reduced to (count * 2^24) ^ (hash % 2^24) */
	131	diff = (count - (cookie >> COOKIEBITS)) & ((__u32) - 1 >> COOKIEBITS);
	132	if (diff >= maxdiff)
	133	return (__u32)-1;
	134
	135	return (cookie -
	136	cookie_hash(saddr, daddr, sport, dport, count - diff, 1))
	137	& COOKIEMASK; /* Leaving the data behind */
	138	}
	139
e905a9ed	140	/*
1da177e4 LT	141	* This table has to be sorted and terminated with (__u16)-1.
	142	* XXX generate a better table.
	143	* Unresolved Issues: HIPPI with a 64k MSS is not well supported.
	144	*/
	145	static __u16 const msstab[] = {
	146	64 - 1,
e905a9ed	147	256 - 1,
1da177e4 LT	148	512 - 1,
1da177e4 LT	149	536 - 1,
e905a9ed	150	1024 - 1,
1da177e4 LT	151	1440 - 1,
	152	1460 - 1,
	153	4312 - 1,
	154	(__u16)-1
	155	};
	156	/* The number doesn't include the -1 terminator */
	157	#define NUM_MSS (ARRAY_SIZE(msstab) - 1)
	158
	159	/*
	160	* Generate a syncookie. mssp points to the mss, which is returned
	161	* rounded down to the value encoded in the cookie.
	162	*/
	163	__u32 cookie_v4_init_sequence(struct sock sk, struct sk_buff skb, __u16 *mssp)
	164	{
aa8223c7 ACM	165	const struct iphdr *iph = ip_hdr(skb);
aa8223c7 ACM	166	const struct tcphdr *th = tcp_hdr(skb);
1da177e4 LT	167	int mssind;
	168	const __u16 mss = *mssp;
	169
a0f82f64	170	tcp_synq_overflow(sk);
1da177e4 LT	171
	172	/* XXX sort msstab[] by probability? Binary search? */
	173	for (mssind = 0; mss > msstab[mssind + 1]; mssind++)
	174	;
	175	*mssp = msstab[mssind] + 1;
	176
de0744af	177	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_SYNCOOKIESSENT);
1da177e4	178
aa8223c7 ACM	179	return secure_tcp_syn_cookie(iph->saddr, iph->daddr,
aa8223c7 ACM	180	th->source, th->dest, ntohl(th->seq),
1da177e4 LT	181	jiffies / (HZ * 60), mssind);
	182	}
	183
e905a9ed	184	/*
1da177e4 LT	185	* This (misnamed) value is the age of syncookie which is permitted.
	186	* Its ideal value should be dependent on TCP_TIMEOUT_INIT and
	187	* sysctl_tcp_retries1. It's a rather complicated formula (exponential
	188	* backoff) to compute at runtime so it's currently hardcoded here.
	189	*/
	190	#define COUNTER_TRIES 4
e905a9ed YH	191	/*
e905a9ed YH	192	* Check if a ack sequence number is a valid syncookie.
1da177e4 LT	193	* Return the decoded mss if it is, or 0 if not.
	194	*/
	195	static inline int cookie_check(struct sk_buff *skb, __u32 cookie)
	196	{
aa8223c7 ACM	197	const struct iphdr *iph = ip_hdr(skb);
	198	const struct tcphdr *th = tcp_hdr(skb);
	199	__u32 seq = ntohl(th->seq) - 1;
	200	__u32 mssind = check_tcp_syn_cookie(cookie, iph->saddr, iph->daddr,
	201	th->source, th->dest, seq,
	202	jiffies / (HZ * 60),
	203	COUNTER_TRIES);
1da177e4 LT	204
	205	return mssind < NUM_MSS ? msstab[mssind] + 1 : 0;
	206	}
	207
1da177e4	208	static inline struct sock get_cookie_sock(struct sock sk, struct sk_buff *skb,
60236fdd	209	struct request_sock *req,
1da177e4 LT	210	struct dst_entry *dst)
1da177e4 LT	211	{
8292a17a	212	struct inet_connection_sock *icsk = inet_csk(sk);
1da177e4 LT	213	struct sock *child;
1da177e4 LT	214
8292a17a	215	child = icsk->icsk_af_ops->syn_recv_sock(sk, skb, req, dst);
1da177e4	216	if (child)
463c84b9	217	inet_csk_reqsk_queue_add(sk, req, child);
1da177e4	218	else
60236fdd	219	reqsk_free(req);
1da177e4 LT	220
	221	return child;
	222	}
	223
4dfc2817 FW	224
	225	/*
	226	* when syncookies are in effect and tcp timestamps are enabled we stored
	227	* additional tcp options in the timestamp.
	228	* This extracts these options from the timestamp echo.
	229	*
	230	* The lowest 4 bits are for snd_wscale
	231	* The next 4 lsb are for rcv_wscale
	232	* The next lsb is for sack_ok
	233	*/
	234	void cookie_check_timestamp(struct tcp_options_received *tcp_opt)
	235	{
	236	/* echoed timestamp, 9 lowest bits contain options */
	237	u32 options = tcp_opt->rcv_tsecr & TSMASK;
	238
	239	tcp_opt->snd_wscale = options & 0xf;
	240	options >>= 4;
	241	tcp_opt->rcv_wscale = options & 0xf;
	242
	243	tcp_opt->sack_ok = (options >> 4) & 0x1;
	244
	245	if (tcp_opt->sack_ok)
	246	tcp_sack_reset(tcp_opt);
	247
	248	if (tcp_opt->snd_wscale \|\| tcp_opt->rcv_wscale)
	249	tcp_opt->wscale_ok = 1;
	250	}
	251	EXPORT_SYMBOL(cookie_check_timestamp);
	252
1da177e4 LT	253	struct sock cookie_v4_check(struct sock sk, struct sk_buff *skb,
	254	struct ip_options *opt)
	255	{
4957faad WAS	256	struct tcp_options_received tcp_opt;
4957faad WAS	257	u8 *hash_location;
2e6599cb ACM	258	struct inet_request_sock *ireq;
2e6599cb ACM	259	struct tcp_request_sock *treq;
1da177e4	260	struct tcp_sock *tp = tcp_sk(sk);
aa8223c7 ACM	261	const struct tcphdr *th = tcp_hdr(skb);
aa8223c7 ACM	262	__u32 cookie = ntohl(th->ack_seq) - 1;
1da177e4	263	struct sock *ret = sk;
e905a9ed YH	264	struct request_sock *req;
	265	int mss;
	266	struct rtable *rt;
1da177e4 LT	267	__u8 rcv_wscale;
1da177e4 LT	268
aa8223c7	269	if (!sysctl_tcp_syncookies \|\| !th->ack)
1da177e4 LT	270	goto out;
1da177e4 LT	271
a0f82f64	272	if (tcp_synq_no_recent_overflow(sk) \|\|
1da177e4	273	(mss = cookie_check(skb, cookie)) == 0) {
de0744af	274	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_SYNCOOKIESFAILED);
1da177e4 LT	275	goto out;
	276	}
	277
de0744af	278	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_SYNCOOKIESRECV);
1da177e4	279
bb5b7c11 DM	280	/* check for timestamp cookie support */
	281	memset(&tcp_opt, 0, sizeof(tcp_opt));
	282	tcp_parse_options(skb, &tcp_opt, &hash_location, 0);
	283
	284	if (tcp_opt.saw_tstamp)
	285	cookie_check_timestamp(&tcp_opt);
	286
1da177e4	287	ret = NULL;
ce4a7d0d	288	req = inet_reqsk_alloc(&tcp_request_sock_ops); /* for safety */
1da177e4 LT	289	if (!req)
	290	goto out;
	291
2e6599cb ACM	292	ireq = inet_rsk(req);
2e6599cb ACM	293	treq = tcp_rsk(req);
aa8223c7	294	treq->rcv_isn = ntohl(th->seq) - 1;
e905a9ed	295	treq->snt_isn = cookie;
1da177e4	296	req->mss = mss;
a3116ac5	297	ireq->loc_port = th->dest;
aa8223c7	298	ireq->rmt_port = th->source;
eddc9ec5 ACM	299	ireq->loc_addr = ip_hdr(skb)->daddr;
eddc9ec5 ACM	300	ireq->rmt_addr = ip_hdr(skb)->saddr;
16df845f	301	ireq->ecn_ok = 0;
bb5b7c11 DM	302	ireq->snd_wscale = tcp_opt.snd_wscale;
	303	ireq->rcv_wscale = tcp_opt.rcv_wscale;
	304	ireq->sack_ok = tcp_opt.sack_ok;
	305	ireq->wscale_ok = tcp_opt.wscale_ok;
	306	ireq->tstamp_ok = tcp_opt.saw_tstamp;
	307	req->ts_recent = tcp_opt.saw_tstamp ? tcp_opt.rcv_tsval : 0;
1da177e4 LT	308
	309	/* We throwed the options of the initial SYN away, so we hope
	310	* the ACK carries the same options again (see RFC1122 4.2.3.8)
	311	*/
	312	if (opt && opt->optlen) {
	313	int opt_size = sizeof(struct ip_options) + opt->optlen;
	314
2e6599cb ACM	315	ireq->opt = kmalloc(opt_size, GFP_ATOMIC);
	316	if (ireq->opt != NULL && ip_options_echo(ireq->opt, skb)) {
	317	kfree(ireq->opt);
	318	ireq->opt = NULL;
1da177e4 LT	319	}
	320	}
	321
284904aa PM	322	if (security_inet_conn_request(sk, skb, req)) {
	323	reqsk_free(req);
	324	goto out;
	325	}
	326
e905a9ed YH	327	req->expires = 0UL;
	328	req->retrans = 0;
	329
1da177e4 LT	330	/*
	331	* We need to lookup the route here to get at the correct
	332	* window size. We should better make sure that the window size
	333	* hasn't changed since we received the original syn, but I see
e905a9ed	334	* no easy way to do this.
1da177e4 LT	335	*/
1da177e4 LT	336	{
ffce9082 AE	337	struct flowi fl = { .mark = sk->sk_mark,
ffce9082 AE	338	.nl_u = { .ip4_u =
1da177e4 LT	339	{ .daddr = ((opt && opt->srr) ?
1da177e4 LT	340	opt->faddr :
2e6599cb ACM	341	ireq->rmt_addr),
2e6599cb ACM	342	.saddr = ireq->loc_addr,
1da177e4 LT	343	.tos = RT_CONN_FLAGS(sk) } },
1da177e4 LT	344	.proto = IPPROTO_TCP,
86b08d86	345	.flags = inet_sk_flowi_flags(sk),
1da177e4	346	.uli_u = { .ports =
aa8223c7 ACM	347	{ .sport = th->dest,
aa8223c7 ACM	348	.dport = th->source } } };
4237c75c	349	security_req_classify_flow(req, &fl);
f206351a	350	if (ip_route_output_key(&init_net, &rt, &fl)) {
60236fdd	351	reqsk_free(req);
e905a9ed	352	goto out;
1da177e4 LT	353	}
	354	}
	355
	356	/* Try to redo what tcp_v4_send_synack did. */
4dfc2817 FW	357	req->window_clamp = tp->window_clamp ? :dst_metric(&rt->u.dst, RTAX_WINDOW);
4dfc2817 FW	358
1da177e4	359	tcp_select_initial_window(tcp_full_space(sk), req->mss,
e905a9ed	360	&req->rcv_wnd, &req->window_clamp,
31d12926	361	ireq->wscale_ok, &rcv_wscale,
31d12926	362	dst_metric(&rt->u.dst, RTAX_INITRWND));
4dfc2817	363
e905a9ed	364	ireq->rcv_wscale = rcv_wscale;
1da177e4 LT	365
	366	ret = get_cookie_sock(sk, skb, req, &rt->u.dst);
	367	out: return ret;
	368	}