[linux.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 <linux/export.h>
#include <net/tcp.h>
#include <net/route.h>

/* Timestamps: lowest bits store TCP options */
#define TSBITS 6
#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 lower 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);

	options = ireq->wscale_ok ? ireq->snd_wscale : 0xf;
	options |= ireq->sack_ok << 4;
	options |= ireq->ecn_ok << 5;

	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 */
}

/*
 * MSS Values are taken from the 2009 paper
 * 'Measuring TCP Maximum Segment Size' by S. Alcock and R. Nelson:
 *  - values 1440 to 1460 accounted for 80% of observed mss values
 *  - values outside the 536-1460 range are rare (<0.2%).
 *
 * Table must be sorted.
 */
static __u16 const msstab[] = {
	64,
	512,
	536,
	1024,
	1440,
	1460,
	4312,
	8960,
};

/*
 * 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);

	for (mssind = ARRAY_SIZE(msstab) - 1; mssind ; mssind--)
		if (mss >= msstab[mssind])
			break;
	*mssp = msstab[mssind];

	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 < ARRAY_SIZE(msstab) ? msstab[mssind] : 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 store snd_wscale.
 * next 2 bits indicate SACK and ECN support.
 *
 * return false if we decode an option that should not be.
 */
bool cookie_check_timestamp(struct tcp_options_received *tcp_opt, bool *ecn_ok)
{
	/* echoed timestamp, lowest bits contain options */
	u32 options = tcp_opt->rcv_tsecr & TSMASK;

	if (!tcp_opt->saw_tstamp)  {
		tcp_clear_options(tcp_opt);
		return true;
	}

	if (!sysctl_tcp_timestamps)
		return false;

	tcp_opt->sack_ok = (options & (1 << 4)) ? TCP_SACK_SEEN : 0;
	*ecn_ok = (options >> 5) & 1;
	if (*ecn_ok && !sysctl_tcp_ecn)
		return false;

	if (tcp_opt->sack_ok && !sysctl_tcp_sack)
		return false;

	if ((options & 0xf) == 0xf)
		return true; /* no window scaling */

	tcp_opt->wscale_ok = 1;
	tcp_opt->snd_wscale = options & 0xf;
	return sysctl_tcp_window_scaling != 0;
}
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;
	const 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;
	bool ecn_ok = false;
	struct flowi4 fl4;

	if (!sysctl_tcp_syncookies || !th->ack || th->rst)
		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, NULL);

	if (!cookie_check_timestamp(&tcp_opt, &ecn_ok))
		goto out;

	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		= ecn_ok;
	ireq->snd_wscale	= tcp_opt.snd_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;
	treq->snt_synack	= tcp_opt.saw_tstamp ? tcp_opt.rcv_tsecr : 0;
	treq->listener		= NULL;

	/* 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_rcu) + opt->optlen;

		ireq->opt = kmalloc(opt_size, GFP_ATOMIC);
		if (ireq->opt != NULL && ip_options_echo(&ireq->opt->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->num_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.
	 */
	flowi4_init_output(&fl4, 0, sk->sk_mark, RT_CONN_FLAGS(sk),
			   RT_SCOPE_UNIVERSE, IPPROTO_TCP,
			   inet_sk_flowi_flags(sk),
			   (opt && opt->srr) ? opt->faddr : ireq->rmt_addr,
			   ireq->loc_addr, th->source, th->dest);
	security_req_classify_flow(req, flowi4_to_flowi(&fl4));
	rt = ip_route_output_key(sock_net(sk), &fl4);
	if (IS_ERR(rt)) {
		reqsk_free(req);
		goto out;
	}

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

	ireq->rcv_wscale  = rcv_wscale;

	ret = get_cookie_sock(sk, skb, req, &rt->dst);
	/* ip_queue_xmit() depends on our flow being setup
	 * Normal sockets get it right from inet_csk_route_child_sock()
	 */
	if (ret)
		inet_sk(ret)->cork.fl.u.ip4 = fl4;
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>
bc3b2d7f	18	#include <linux/export.h>
1da177e4	19	#include <net/tcp.h>
86b08d86	20	#include <net/route.h>
1da177e4	21
734f614b	22	/* Timestamps: lowest bits store TCP options */
172d69e6	23	#define TSBITS 6
4dfc2817 FW	24	#define TSMASK (((__u32)1 << TSBITS) - 1)
4dfc2817 FW	25
1da177e4 LT	26	extern int sysctl_tcp_syncookies;
1da177e4 LT	27
2051f11f	28	__u32 syncookie_secret[2][16-4+SHA_DIGEST_WORDS];
c6aefafb	29	EXPORT_SYMBOL(syncookie_secret);
1da177e4 LT	30
	31	static __init int init_syncookies(void)
	32	{
	33	get_random_bytes(syncookie_secret, sizeof(syncookie_secret));
	34	return 0;
	35	}
c6aefafb	36	__initcall(init_syncookies);
1da177e4 LT	37
	38	#define COOKIEBITS 24 /* Upper bits store count */
	39	#define COOKIEMASK (((__u32)1 << COOKIEBITS) - 1)
	40
245b2e70 TH	41	static DEFINE_PER_CPU(__u32 [16 + 5 + SHA_WORKSPACE_WORDS],
245b2e70 TH	42	ipv4_cookie_scratch);
11baab7a	43
714e85be	44	static u32 cookie_hash(__be32 saddr, __be32 daddr, __be16 sport, __be16 dport,
1da177e4 LT	45	u32 count, int c)
1da177e4 LT	46	{
245b2e70	47	__u32 *tmp = __get_cpu_var(ipv4_cookie_scratch);
1da177e4	48
2051f11f	49	memcpy(tmp + 4, syncookie_secret[c], sizeof(syncookie_secret[c]));
714e85be AV	50	tmp[0] = (__force u32)saddr;
	51	tmp[1] = (__force u32)daddr;
	52	tmp[2] = ((__force u32)sport << 16) + (__force u32)dport;
1da177e4 LT	53	tmp[3] = count;
	54	sha_transform(tmp + 16, (__u8 *)tmp, tmp + 16 + 5);
	55
	56	return tmp[17];
	57	}
	58
4dfc2817 FW	59
	60	/*
	61	* when syncookies are in effect and tcp timestamps are enabled we encode
734f614b	62	* tcp options in the lower bits of the timestamp value that will be
4dfc2817 FW	63	* sent in the syn-ack.
	64	* Since subsequent timestamps use the normal tcp_time_stamp value, we
	65	* must make sure that the resulting initial timestamp is <= tcp_time_stamp.
	66	*/
	67	__u32 cookie_init_timestamp(struct request_sock *req)
	68	{
	69	struct inet_request_sock *ireq;
	70	u32 ts, ts_now = tcp_time_stamp;
	71	u32 options = 0;
	72
	73	ireq = inet_rsk(req);
734f614b FW	74
	75	options = ireq->wscale_ok ? ireq->snd_wscale : 0xf;
	76	options \|= ireq->sack_ok << 4;
172d69e6	77	options \|= ireq->ecn_ok << 5;
4dfc2817 FW	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	/*
5918e2fb FW	141	* MSS Values are taken from the 2009 paper
	142	* 'Measuring TCP Maximum Segment Size' by S. Alcock and R. Nelson:
	143	* - values 1440 to 1460 accounted for 80% of observed mss values
	144	* - values outside the 536-1460 range are rare (<0.2%).
	145	*
	146	* Table must be sorted.
1da177e4 LT	147	*/
1da177e4 LT	148	static __u16 const msstab[] = {
5918e2fb FW	149	64,
	150	512,
	151	536,
	152	1024,
	153	1440,
	154	1460,
	155	4312,
	156	8960,
1da177e4	157	};
1da177e4 LT	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	171
5918e2fb FW	172	for (mssind = ARRAY_SIZE(msstab) - 1; mssind ; mssind--)
	173	if (mss >= msstab[mssind])
	174	break;
	175	*mssp = msstab[mssind];
1da177e4	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	204
5918e2fb	205	return mssind < ARRAY_SIZE(msstab) ? msstab[mssind] : 0;
1da177e4 LT	206	}
1da177e4 LT	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	*
734f614b	230	* The lowest 4 bits store snd_wscale.
172d69e6	231	* next 2 bits indicate SACK and ECN support.
8c763681 FW	232	*
8c763681 FW	233	* return false if we decode an option that should not be.
4dfc2817	234	*/
172d69e6	235	bool cookie_check_timestamp(struct tcp_options_received tcp_opt, bool ecn_ok)
4dfc2817	236	{
734f614b	237	/* echoed timestamp, lowest bits contain options */
4dfc2817 FW	238	u32 options = tcp_opt->rcv_tsecr & TSMASK;
4dfc2817 FW	239
8c763681 FW	240	if (!tcp_opt->saw_tstamp) {
	241	tcp_clear_options(tcp_opt);
	242	return true;
	243	}
	244
	245	if (!sysctl_tcp_timestamps)
	246	return false;
	247
ab56222a	248	tcp_opt->sack_ok = (options & (1 << 4)) ? TCP_SACK_SEEN : 0;
172d69e6 FW	249	*ecn_ok = (options >> 5) & 1;
	250	if (*ecn_ok && !sysctl_tcp_ecn)
	251	return false;
4dfc2817	252
8c763681 FW	253	if (tcp_opt->sack_ok && !sysctl_tcp_sack)
8c763681 FW	254	return false;
4dfc2817	255
734f614b FW	256	if ((options & 0xf) == 0xf)
	257	return true; /* no window scaling */
	258
	259	tcp_opt->wscale_ok = 1;
	260	tcp_opt->snd_wscale = options & 0xf;
	261	return sysctl_tcp_window_scaling != 0;
4dfc2817 FW	262	}
	263	EXPORT_SYMBOL(cookie_check_timestamp);
	264
1da177e4 LT	265	struct sock cookie_v4_check(struct sock sk, struct sk_buff *skb,
	266	struct ip_options *opt)
	267	{
4957faad	268	struct tcp_options_received tcp_opt;
cf533ea5	269	const u8 *hash_location;
2e6599cb ACM	270	struct inet_request_sock *ireq;
2e6599cb ACM	271	struct tcp_request_sock *treq;
1da177e4	272	struct tcp_sock *tp = tcp_sk(sk);
aa8223c7 ACM	273	const struct tcphdr *th = tcp_hdr(skb);
aa8223c7 ACM	274	__u32 cookie = ntohl(th->ack_seq) - 1;
1da177e4	275	struct sock *ret = sk;
e905a9ed YH	276	struct request_sock *req;
	277	int mss;
	278	struct rtable *rt;
1da177e4	279	__u8 rcv_wscale;
f0e3d068	280	bool ecn_ok = false;
dfd25fff	281	struct flowi4 fl4;
1da177e4	282
af9b4738	283	if (!sysctl_tcp_syncookies \|\| !th->ack \|\| th->rst)
1da177e4 LT	284	goto out;
1da177e4 LT	285
a0f82f64	286	if (tcp_synq_no_recent_overflow(sk) \|\|
1da177e4	287	(mss = cookie_check(skb, cookie)) == 0) {
de0744af	288	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_SYNCOOKIESFAILED);
1da177e4 LT	289	goto out;
	290	}
	291
de0744af	292	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_SYNCOOKIESRECV);
1da177e4	293
bb5b7c11 DM	294	/* check for timestamp cookie support */
bb5b7c11 DM	295	memset(&tcp_opt, 0, sizeof(tcp_opt));
2100c8d2	296	tcp_parse_options(skb, &tcp_opt, &hash_location, 0, NULL);
bb5b7c11	297
172d69e6	298	if (!cookie_check_timestamp(&tcp_opt, &ecn_ok))
8c763681	299	goto out;
bb5b7c11	300
1da177e4	301	ret = NULL;
ce4a7d0d	302	req = inet_reqsk_alloc(&tcp_request_sock_ops); /* for safety */
1da177e4 LT	303	if (!req)
	304	goto out;
	305
2e6599cb ACM	306	ireq = inet_rsk(req);
2e6599cb ACM	307	treq = tcp_rsk(req);
aa8223c7	308	treq->rcv_isn = ntohl(th->seq) - 1;
e905a9ed	309	treq->snt_isn = cookie;
1da177e4	310	req->mss = mss;
a3116ac5	311	ireq->loc_port = th->dest;
aa8223c7	312	ireq->rmt_port = th->source;
eddc9ec5 ACM	313	ireq->loc_addr = ip_hdr(skb)->daddr;
eddc9ec5 ACM	314	ireq->rmt_addr = ip_hdr(skb)->saddr;
172d69e6	315	ireq->ecn_ok = ecn_ok;
bb5b7c11	316	ireq->snd_wscale = tcp_opt.snd_wscale;
bb5b7c11 DM	317	ireq->sack_ok = tcp_opt.sack_ok;
	318	ireq->wscale_ok = tcp_opt.wscale_ok;
	319	ireq->tstamp_ok = tcp_opt.saw_tstamp;
	320	req->ts_recent = tcp_opt.saw_tstamp ? tcp_opt.rcv_tsval : 0;
9ad7c049	321	treq->snt_synack = tcp_opt.saw_tstamp ? tcp_opt.rcv_tsecr : 0;
8336886f	322	treq->listener = NULL;
1da177e4 LT	323
	324	/* We throwed the options of the initial SYN away, so we hope
	325	* the ACK carries the same options again (see RFC1122 4.2.3.8)
	326	*/
	327	if (opt && opt->optlen) {
f6d8bd05	328	int opt_size = sizeof(struct ip_options_rcu) + opt->optlen;
1da177e4	329
2e6599cb	330	ireq->opt = kmalloc(opt_size, GFP_ATOMIC);
f6d8bd05	331	if (ireq->opt != NULL && ip_options_echo(&ireq->opt->opt, skb)) {
2e6599cb ACM	332	kfree(ireq->opt);
2e6599cb ACM	333	ireq->opt = NULL;
1da177e4 LT	334	}
	335	}
	336
284904aa PM	337	if (security_inet_conn_request(sk, skb, req)) {
	338	reqsk_free(req);
	339	goto out;
	340	}
	341
e905a9ed	342	req->expires = 0UL;
e6c022a4	343	req->num_retrans = 0;
e905a9ed	344
1da177e4 LT	345	/*
	346	* We need to lookup the route here to get at the correct
	347	* window size. We should better make sure that the window size
	348	* hasn't changed since we received the original syn, but I see
e905a9ed	349	* no easy way to do this.
1da177e4	350	*/
dfd25fff ED	351	flowi4_init_output(&fl4, 0, sk->sk_mark, RT_CONN_FLAGS(sk),
	352	RT_SCOPE_UNIVERSE, IPPROTO_TCP,
	353	inet_sk_flowi_flags(sk),
	354	(opt && opt->srr) ? opt->faddr : ireq->rmt_addr,
	355	ireq->loc_addr, th->source, th->dest);
	356	security_req_classify_flow(req, flowi4_to_flowi(&fl4));
	357	rt = ip_route_output_key(sock_net(sk), &fl4);
	358	if (IS_ERR(rt)) {
	359	reqsk_free(req);
	360	goto out;
1da177e4 LT	361	}
	362
	363	/* Try to redo what tcp_v4_send_synack did. */
d8d1f30b	364	req->window_clamp = tp->window_clamp ? :dst_metric(&rt->dst, RTAX_WINDOW);
4dfc2817	365
1da177e4	366	tcp_select_initial_window(tcp_full_space(sk), req->mss,
e905a9ed	367	&req->rcv_wnd, &req->window_clamp,
31d12926	368	ireq->wscale_ok, &rcv_wscale,
d8d1f30b	369	dst_metric(&rt->dst, RTAX_INITRWND));
4dfc2817	370
e905a9ed	371	ireq->rcv_wscale = rcv_wscale;
1da177e4	372
d8d1f30b	373	ret = get_cookie_sock(sk, skb, req, &rt->dst);
dfd25fff ED	374	/* ip_queue_xmit() depends on our flow being setup
	375	* Normal sockets get it right from inet_csk_route_child_sock()
	376	*/
	377	if (ret)
	378	inet_sk(ret)->cork.fl.u.ip4 = fl4;
1da177e4 LT	379	out: return ret;
1da177e4 LT	380	}