tcp_ipv4.c 61.2 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38
/*
 * INET		An implementation of the TCP/IP protocol suite for the LINUX
 *		operating system.  INET is implemented using the  BSD Socket
 *		interface as the means of communication with the user level.
 *
 *		Implementation of the Transmission Control Protocol(TCP).
 *
 * Version:	$Id: tcp_ipv4.c,v 1.240 2002/02/01 22:01:04 davem Exp $
 *
 *		IPv4 specific functions
 *
 *
 *		code split from:
 *		linux/ipv4/tcp.c
 *		linux/ipv4/tcp_input.c
 *		linux/ipv4/tcp_output.c
 *
 *		See tcp.c for author information
 *
 *	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.
 */

/*
 * Changes:
 *		David S. Miller	:	New socket lookup architecture.
 *					This code is dedicated to John Dyson.
 *		David S. Miller :	Change semantics of established hash,
 *					half is devoted to TIME_WAIT sockets
 *					and the rest go in the other half.
 *		Andi Kleen :		Add support for syncookies and fixed
 *					some bugs: ip options weren't passed to
 *					the TCP layer, missed a check for an
 *					ACK bit.
 *		Andi Kleen :		Implemented fast path mtu discovery.
 *	     				Fixed many serious bugs in the
39
 *					request_sock handling and moved
L
Linus Torvalds 已提交
40 41
 *					most of it into the af independent code.
 *					Added tail drop and some other bugfixes.
S
Stephen Hemminger 已提交
42
 *					Added new listen semantics.
L
Linus Torvalds 已提交
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64
 *		Mike McLagan	:	Routing by source
 *	Juan Jose Ciarlante:		ip_dynaddr bits
 *		Andi Kleen:		various fixes.
 *	Vitaly E. Lavrov	:	Transparent proxy revived after year
 *					coma.
 *	Andi Kleen		:	Fix new listen.
 *	Andi Kleen		:	Fix accept error reporting.
 *	YOSHIFUJI Hideaki @USAGI and:	Support IPV6_V6ONLY socket option, which
 *	Alexey Kuznetsov		allow both IPv4 and IPv6 sockets to bind
 *					a single port at the same time.
 */


#include <linux/types.h>
#include <linux/fcntl.h>
#include <linux/module.h>
#include <linux/random.h>
#include <linux/cache.h>
#include <linux/jhash.h>
#include <linux/init.h>
#include <linux/times.h>

65
#include <net/net_namespace.h>
L
Linus Torvalds 已提交
66
#include <net/icmp.h>
67
#include <net/inet_hashtables.h>
L
Linus Torvalds 已提交
68
#include <net/tcp.h>
69
#include <net/transp_v6.h>
L
Linus Torvalds 已提交
70 71
#include <net/ipv6.h>
#include <net/inet_common.h>
72
#include <net/timewait_sock.h>
L
Linus Torvalds 已提交
73
#include <net/xfrm.h>
C
Chris Leech 已提交
74
#include <net/netdma.h>
L
Linus Torvalds 已提交
75 76 77 78 79 80 81

#include <linux/inet.h>
#include <linux/ipv6.h>
#include <linux/stddef.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>

82 83 84
#include <linux/crypto.h>
#include <linux/scatterlist.h>

85 86
int sysctl_tcp_tw_reuse __read_mostly;
int sysctl_tcp_low_latency __read_mostly;
L
Linus Torvalds 已提交
87 88 89 90 91

/* Check TCP sequence numbers in ICMP packets. */
#define ICMP_MIN_LENGTH 8

/* Socket used for sending RSTs */
92
static struct socket *tcp_socket __read_mostly;
L
Linus Torvalds 已提交
93

94
void tcp_v4_send_check(struct sock *sk, int len, struct sk_buff *skb);
L
Linus Torvalds 已提交
95

96
#ifdef CONFIG_TCP_MD5SIG
97 98
static struct tcp_md5sig_key *tcp_v4_md5_do_lookup(struct sock *sk,
						   __be32 addr);
99
static int tcp_v4_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key,
100 101 102
				   __be32 saddr, __be32 daddr,
				   struct tcphdr *th, int protocol,
				   int tcplen);
103 104
#endif

105
struct inet_hashinfo __cacheline_aligned tcp_hashinfo = {
106 107 108
	.lhash_lock  = __RW_LOCK_UNLOCKED(tcp_hashinfo.lhash_lock),
	.lhash_users = ATOMIC_INIT(0),
	.lhash_wait  = __WAIT_QUEUE_HEAD_INITIALIZER(tcp_hashinfo.lhash_wait),
L
Linus Torvalds 已提交
109 110
};

111 112
static int tcp_v4_get_port(struct sock *sk, unsigned short snum)
{
113 114
	return inet_csk_get_port(&tcp_hashinfo, sk, snum,
				 inet_csk_bind_conflict);
115 116
}

L
Linus Torvalds 已提交
117 118
static void tcp_v4_hash(struct sock *sk)
{
119
	inet_hash(&tcp_hashinfo, sk);
L
Linus Torvalds 已提交
120 121 122 123
}

void tcp_unhash(struct sock *sk)
{
124
	inet_unhash(&tcp_hashinfo, sk);
L
Linus Torvalds 已提交
125 126
}

127
static inline __u32 tcp_v4_init_sequence(struct sk_buff *skb)
L
Linus Torvalds 已提交
128
{
129 130
	return secure_tcp_sequence_number(ip_hdr(skb)->daddr,
					  ip_hdr(skb)->saddr,
131 132
					  tcp_hdr(skb)->dest,
					  tcp_hdr(skb)->source);
L
Linus Torvalds 已提交
133 134
}

135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152
int tcp_twsk_unique(struct sock *sk, struct sock *sktw, void *twp)
{
	const struct tcp_timewait_sock *tcptw = tcp_twsk(sktw);
	struct tcp_sock *tp = tcp_sk(sk);

	/* With PAWS, it is safe from the viewpoint
	   of data integrity. Even without PAWS it is safe provided sequence
	   spaces do not overlap i.e. at data rates <= 80Mbit/sec.

	   Actually, the idea is close to VJ's one, only timestamp cache is
	   held not per host, but per port pair and TW bucket is used as state
	   holder.

	   If TW bucket has been already destroyed we fall back to VJ's scheme
	   and use initial timestamp retrieved from peer table.
	 */
	if (tcptw->tw_ts_recent_stamp &&
	    (twp == NULL || (sysctl_tcp_tw_reuse &&
153
			     get_seconds() - tcptw->tw_ts_recent_stamp > 1))) {
154 155 156 157 158 159 160 161 162 163 164 165 166 167
		tp->write_seq = tcptw->tw_snd_nxt + 65535 + 2;
		if (tp->write_seq == 0)
			tp->write_seq = 1;
		tp->rx_opt.ts_recent	   = tcptw->tw_ts_recent;
		tp->rx_opt.ts_recent_stamp = tcptw->tw_ts_recent_stamp;
		sock_hold(sktw);
		return 1;
	}

	return 0;
}

EXPORT_SYMBOL_GPL(tcp_twsk_unique);

L
Linus Torvalds 已提交
168 169 170 171 172 173 174
/* This will initiate an outgoing connection. */
int tcp_v4_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)
{
	struct inet_sock *inet = inet_sk(sk);
	struct tcp_sock *tp = tcp_sk(sk);
	struct sockaddr_in *usin = (struct sockaddr_in *)uaddr;
	struct rtable *rt;
175
	__be32 daddr, nexthop;
L
Linus Torvalds 已提交
176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194
	int tmp;
	int err;

	if (addr_len < sizeof(struct sockaddr_in))
		return -EINVAL;

	if (usin->sin_family != AF_INET)
		return -EAFNOSUPPORT;

	nexthop = daddr = usin->sin_addr.s_addr;
	if (inet->opt && inet->opt->srr) {
		if (!daddr)
			return -EINVAL;
		nexthop = inet->opt->faddr;
	}

	tmp = ip_route_connect(&rt, nexthop, inet->saddr,
			       RT_CONN_FLAGS(sk), sk->sk_bound_dev_if,
			       IPPROTO_TCP,
195
			       inet->sport, usin->sin_port, sk, 1);
196 197 198
	if (tmp < 0) {
		if (tmp == -ENETUNREACH)
			IP_INC_STATS_BH(IPSTATS_MIB_OUTNOROUTES);
L
Linus Torvalds 已提交
199
		return tmp;
200
	}
L
Linus Torvalds 已提交
201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220

	if (rt->rt_flags & (RTCF_MULTICAST | RTCF_BROADCAST)) {
		ip_rt_put(rt);
		return -ENETUNREACH;
	}

	if (!inet->opt || !inet->opt->srr)
		daddr = rt->rt_dst;

	if (!inet->saddr)
		inet->saddr = rt->rt_src;
	inet->rcv_saddr = inet->saddr;

	if (tp->rx_opt.ts_recent_stamp && inet->daddr != daddr) {
		/* Reset inherited state */
		tp->rx_opt.ts_recent	   = 0;
		tp->rx_opt.ts_recent_stamp = 0;
		tp->write_seq		   = 0;
	}

221
	if (tcp_death_row.sysctl_tw_recycle &&
L
Linus Torvalds 已提交
222 223
	    !tp->rx_opt.ts_recent_stamp && rt->rt_dst == daddr) {
		struct inet_peer *peer = rt_get_peer(rt);
224 225 226 227 228
		/*
		 * VJ's idea. We save last timestamp seen from
		 * the destination in peer table, when entering state
		 * TIME-WAIT * and initialize rx_opt.ts_recent from it,
		 * when trying new connection.
L
Linus Torvalds 已提交
229
		 */
230
		if (peer != NULL &&
231
		    peer->tcp_ts_stamp + TCP_PAWS_MSL >= get_seconds()) {
L
Linus Torvalds 已提交
232 233 234 235 236 237 238 239
			tp->rx_opt.ts_recent_stamp = peer->tcp_ts_stamp;
			tp->rx_opt.ts_recent = peer->tcp_ts;
		}
	}

	inet->dport = usin->sin_port;
	inet->daddr = daddr;

240
	inet_csk(sk)->icsk_ext_hdr_len = 0;
L
Linus Torvalds 已提交
241
	if (inet->opt)
242
		inet_csk(sk)->icsk_ext_hdr_len = inet->opt->optlen;
L
Linus Torvalds 已提交
243 244 245 246 247 248 249 250 251

	tp->rx_opt.mss_clamp = 536;

	/* Socket identity is still unknown (sport may be zero).
	 * However we set state to SYN-SENT and not releasing socket
	 * lock select source port, enter ourselves into the hash tables and
	 * complete initialization after this.
	 */
	tcp_set_state(sk, TCP_SYN_SENT);
252
	err = inet_hash_connect(&tcp_death_row, sk);
L
Linus Torvalds 已提交
253 254 255
	if (err)
		goto failure;

256 257
	err = ip_route_newports(&rt, IPPROTO_TCP,
				inet->sport, inet->dport, sk);
L
Linus Torvalds 已提交
258 259 260 261
	if (err)
		goto failure;

	/* OK, now commit destination to socket.  */
262
	sk->sk_gso_type = SKB_GSO_TCPV4;
263
	sk_setup_caps(sk, &rt->u.dst);
L
Linus Torvalds 已提交
264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280

	if (!tp->write_seq)
		tp->write_seq = secure_tcp_sequence_number(inet->saddr,
							   inet->daddr,
							   inet->sport,
							   usin->sin_port);

	inet->id = tp->write_seq ^ jiffies;

	err = tcp_connect(sk);
	rt = NULL;
	if (err)
		goto failure;

	return 0;

failure:
281 282 283 284
	/*
	 * This unhashes the socket and releases the local port,
	 * if necessary.
	 */
L
Linus Torvalds 已提交
285 286 287 288 289 290 291 292 293 294
	tcp_set_state(sk, TCP_CLOSE);
	ip_rt_put(rt);
	sk->sk_route_caps = 0;
	inet->dport = 0;
	return err;
}

/*
 * This routine does path mtu discovery as defined in RFC1191.
 */
S
Stephen Hemminger 已提交
295
static void do_pmtu_discovery(struct sock *sk, struct iphdr *iph, u32 mtu)
L
Linus Torvalds 已提交
296 297 298 299 300 301 302 303 304 305 306 307 308 309
{
	struct dst_entry *dst;
	struct inet_sock *inet = inet_sk(sk);

	/* We are not interested in TCP_LISTEN and open_requests (SYN-ACKs
	 * send out by Linux are always <576bytes so they should go through
	 * unfragmented).
	 */
	if (sk->sk_state == TCP_LISTEN)
		return;

	/* We don't check in the destentry if pmtu discovery is forbidden
	 * on this route. We just assume that no packet_to_big packets
	 * are send back when pmtu discovery is not active.
310
	 * There is a small race when the user changes this flag in the
L
Linus Torvalds 已提交
311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326
	 * route, but I think that's acceptable.
	 */
	if ((dst = __sk_dst_check(sk, 0)) == NULL)
		return;

	dst->ops->update_pmtu(dst, mtu);

	/* Something is about to be wrong... Remember soft error
	 * for the case, if this connection will not able to recover.
	 */
	if (mtu < dst_mtu(dst) && ip_dont_fragment(sk, dst))
		sk->sk_err_soft = EMSGSIZE;

	mtu = dst_mtu(dst);

	if (inet->pmtudisc != IP_PMTUDISC_DONT &&
327
	    inet_csk(sk)->icsk_pmtu_cookie > mtu) {
L
Linus Torvalds 已提交
328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360
		tcp_sync_mss(sk, mtu);

		/* Resend the TCP packet because it's
		 * clear that the old packet has been
		 * dropped. This is the new "fast" path mtu
		 * discovery.
		 */
		tcp_simple_retransmit(sk);
	} /* else let the usual retransmit timer handle it */
}

/*
 * This routine is called by the ICMP module when it gets some
 * sort of error condition.  If err < 0 then the socket should
 * be closed and the error returned to the user.  If err > 0
 * it's just the icmp type << 8 | icmp code.  After adjustment
 * header points to the first 8 bytes of the tcp header.  We need
 * to find the appropriate port.
 *
 * The locking strategy used here is very "optimistic". When
 * someone else accesses the socket the ICMP is just dropped
 * and for some paths there is no check at all.
 * A more general error queue to queue errors for later handling
 * is probably better.
 *
 */

void tcp_v4_err(struct sk_buff *skb, u32 info)
{
	struct iphdr *iph = (struct iphdr *)skb->data;
	struct tcphdr *th = (struct tcphdr *)(skb->data + (iph->ihl << 2));
	struct tcp_sock *tp;
	struct inet_sock *inet;
361 362
	const int type = icmp_hdr(skb)->type;
	const int code = icmp_hdr(skb)->code;
L
Linus Torvalds 已提交
363 364 365 366 367 368 369 370 371
	struct sock *sk;
	__u32 seq;
	int err;

	if (skb->len < (iph->ihl << 2) + 8) {
		ICMP_INC_STATS_BH(ICMP_MIB_INERRORS);
		return;
	}

372
	sk = inet_lookup(&tcp_hashinfo, iph->daddr, th->dest, iph->saddr,
373
			 th->source, inet_iif(skb));
L
Linus Torvalds 已提交
374 375 376 377 378
	if (!sk) {
		ICMP_INC_STATS_BH(ICMP_MIB_INERRORS);
		return;
	}
	if (sk->sk_state == TCP_TIME_WAIT) {
379
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396
		return;
	}

	bh_lock_sock(sk);
	/* If too many ICMPs get dropped on busy
	 * servers this needs to be solved differently.
	 */
	if (sock_owned_by_user(sk))
		NET_INC_STATS_BH(LINUX_MIB_LOCKDROPPEDICMPS);

	if (sk->sk_state == TCP_CLOSE)
		goto out;

	tp = tcp_sk(sk);
	seq = ntohl(th->seq);
	if (sk->sk_state != TCP_LISTEN &&
	    !between(seq, tp->snd_una, tp->snd_nxt)) {
397
		NET_INC_STATS_BH(LINUX_MIB_OUTOFWINDOWICMPS);
L
Linus Torvalds 已提交
398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427
		goto out;
	}

	switch (type) {
	case ICMP_SOURCE_QUENCH:
		/* Just silently ignore these. */
		goto out;
	case ICMP_PARAMETERPROB:
		err = EPROTO;
		break;
	case ICMP_DEST_UNREACH:
		if (code > NR_ICMP_UNREACH)
			goto out;

		if (code == ICMP_FRAG_NEEDED) { /* PMTU discovery (RFC1191) */
			if (!sock_owned_by_user(sk))
				do_pmtu_discovery(sk, iph, info);
			goto out;
		}

		err = icmp_err_convert[code].errno;
		break;
	case ICMP_TIME_EXCEEDED:
		err = EHOSTUNREACH;
		break;
	default:
		goto out;
	}

	switch (sk->sk_state) {
428
		struct request_sock *req, **prev;
L
Linus Torvalds 已提交
429 430 431 432
	case TCP_LISTEN:
		if (sock_owned_by_user(sk))
			goto out;

433 434
		req = inet_csk_search_req(sk, &prev, th->dest,
					  iph->daddr, iph->saddr);
L
Linus Torvalds 已提交
435 436 437 438 439 440 441 442
		if (!req)
			goto out;

		/* ICMPs are not backlogged, hence we cannot get
		   an established socket here.
		 */
		BUG_TRAP(!req->sk);

443
		if (seq != tcp_rsk(req)->snt_isn) {
L
Linus Torvalds 已提交
444 445 446 447 448 449 450 451 452 453
			NET_INC_STATS_BH(LINUX_MIB_OUTOFWINDOWICMPS);
			goto out;
		}

		/*
		 * Still in SYN_RECV, just remove it silently.
		 * There is no good way to pass the error to the newly
		 * created socket, and POSIX does not want network
		 * errors returned from accept().
		 */
454
		inet_csk_reqsk_queue_drop(sk, req, prev);
L
Linus Torvalds 已提交
455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502
		goto out;

	case TCP_SYN_SENT:
	case TCP_SYN_RECV:  /* Cannot happen.
			       It can f.e. if SYNs crossed.
			     */
		if (!sock_owned_by_user(sk)) {
			sk->sk_err = err;

			sk->sk_error_report(sk);

			tcp_done(sk);
		} else {
			sk->sk_err_soft = err;
		}
		goto out;
	}

	/* If we've already connected we will keep trying
	 * until we time out, or the user gives up.
	 *
	 * rfc1122 4.2.3.9 allows to consider as hard errors
	 * only PROTO_UNREACH and PORT_UNREACH (well, FRAG_FAILED too,
	 * but it is obsoleted by pmtu discovery).
	 *
	 * Note, that in modern internet, where routing is unreliable
	 * and in each dark corner broken firewalls sit, sending random
	 * errors ordered by their masters even this two messages finally lose
	 * their original sense (even Linux sends invalid PORT_UNREACHs)
	 *
	 * Now we are in compliance with RFCs.
	 *							--ANK (980905)
	 */

	inet = inet_sk(sk);
	if (!sock_owned_by_user(sk) && inet->recverr) {
		sk->sk_err = err;
		sk->sk_error_report(sk);
	} else	{ /* Only an error on timeout */
		sk->sk_err_soft = err;
	}

out:
	bh_unlock_sock(sk);
	sock_put(sk);
}

/* This routine computes an IPv4 TCP checksum. */
503
void tcp_v4_send_check(struct sock *sk, int len, struct sk_buff *skb)
L
Linus Torvalds 已提交
504 505
{
	struct inet_sock *inet = inet_sk(sk);
506
	struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
507

508
	if (skb->ip_summed == CHECKSUM_PARTIAL) {
509 510
		th->check = ~tcp_v4_check(len, inet->saddr,
					  inet->daddr, 0);
511
		skb->csum_start = skb_transport_header(skb) - skb->head;
A
Al Viro 已提交
512
		skb->csum_offset = offsetof(struct tcphdr, check);
L
Linus Torvalds 已提交
513
	} else {
514
		th->check = tcp_v4_check(len, inet->saddr, inet->daddr,
L
Linus Torvalds 已提交
515 516 517 518 519 520
					 csum_partial((char *)th,
						      th->doff << 2,
						      skb->csum));
	}
}

521 522
int tcp_v4_gso_send_check(struct sk_buff *skb)
{
523
	const struct iphdr *iph;
524 525 526 527 528
	struct tcphdr *th;

	if (!pskb_may_pull(skb, sizeof(*th)))
		return -EINVAL;

529
	iph = ip_hdr(skb);
530
	th = tcp_hdr(skb);
531 532

	th->check = 0;
533
	th->check = ~tcp_v4_check(skb->len, iph->saddr, iph->daddr, 0);
534
	skb->csum_start = skb_transport_header(skb) - skb->head;
A
Al Viro 已提交
535
	skb->csum_offset = offsetof(struct tcphdr, check);
536
	skb->ip_summed = CHECKSUM_PARTIAL;
537 538 539
	return 0;
}

L
Linus Torvalds 已提交
540 541 542 543 544 545 546 547 548 549 550 551 552
/*
 *	This routine will send an RST to the other tcp.
 *
 *	Someone asks: why I NEVER use socket parameters (TOS, TTL etc.)
 *		      for reset.
 *	Answer: if a packet caused RST, it is not for a socket
 *		existing in our system, if it is matched to a socket,
 *		it is just duplicate segment or bug in other side's TCP.
 *		So that we build reply only basing on parameters
 *		arrived with segment.
 *	Exception: precedence violation. We do not implement it in any case.
 */

553
static void tcp_v4_send_reset(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
554
{
555
	struct tcphdr *th = tcp_hdr(skb);
556 557 558
	struct {
		struct tcphdr th;
#ifdef CONFIG_TCP_MD5SIG
559
		__be32 opt[(TCPOLEN_MD5SIG_ALIGNED >> 2)];
560 561
#endif
	} rep;
L
Linus Torvalds 已提交
562
	struct ip_reply_arg arg;
563 564 565
#ifdef CONFIG_TCP_MD5SIG
	struct tcp_md5sig_key *key;
#endif
L
Linus Torvalds 已提交
566 567 568 569 570 571 572 573 574

	/* Never send a reset in response to a reset. */
	if (th->rst)
		return;

	if (((struct rtable *)skb->dst)->rt_type != RTN_LOCAL)
		return;

	/* Swap the send and the receive. */
575 576 577 578 579
	memset(&rep, 0, sizeof(rep));
	rep.th.dest   = th->source;
	rep.th.source = th->dest;
	rep.th.doff   = sizeof(struct tcphdr) / 4;
	rep.th.rst    = 1;
L
Linus Torvalds 已提交
580 581

	if (th->ack) {
582
		rep.th.seq = th->ack_seq;
L
Linus Torvalds 已提交
583
	} else {
584 585 586
		rep.th.ack = 1;
		rep.th.ack_seq = htonl(ntohl(th->seq) + th->syn + th->fin +
				       skb->len - (th->doff << 2));
L
Linus Torvalds 已提交
587 588
	}

589
	memset(&arg, 0, sizeof(arg));
590 591 592 593
	arg.iov[0].iov_base = (unsigned char *)&rep;
	arg.iov[0].iov_len  = sizeof(rep.th);

#ifdef CONFIG_TCP_MD5SIG
594
	key = sk ? tcp_v4_md5_do_lookup(sk, ip_hdr(skb)->daddr) : NULL;
595 596 597 598 599 600 601 602 603 604 605
	if (key) {
		rep.opt[0] = htonl((TCPOPT_NOP << 24) |
				   (TCPOPT_NOP << 16) |
				   (TCPOPT_MD5SIG << 8) |
				   TCPOLEN_MD5SIG);
		/* Update length and the length the header thinks exists */
		arg.iov[0].iov_len += TCPOLEN_MD5SIG_ALIGNED;
		rep.th.doff = arg.iov[0].iov_len / 4;

		tcp_v4_do_calc_md5_hash((__u8 *)&rep.opt[1],
					key,
606 607
					ip_hdr(skb)->daddr,
					ip_hdr(skb)->saddr,
608 609 610 611
					&rep.th, IPPROTO_TCP,
					arg.iov[0].iov_len);
	}
#endif
612 613
	arg.csum = csum_tcpudp_nofold(ip_hdr(skb)->daddr,
				      ip_hdr(skb)->saddr, /* XXX */
L
Linus Torvalds 已提交
614 615 616
				      sizeof(struct tcphdr), IPPROTO_TCP, 0);
	arg.csumoffset = offsetof(struct tcphdr, check) / 2;

617
	ip_send_reply(tcp_socket->sk, skb, &arg, arg.iov[0].iov_len);
L
Linus Torvalds 已提交
618 619 620 621 622 623 624 625 626

	TCP_INC_STATS_BH(TCP_MIB_OUTSEGS);
	TCP_INC_STATS_BH(TCP_MIB_OUTRSTS);
}

/* The code following below sending ACKs in SYN-RECV and TIME-WAIT states
   outside socket context is ugly, certainly. What can I do?
 */

627 628
static void tcp_v4_send_ack(struct tcp_timewait_sock *twsk,
			    struct sk_buff *skb, u32 seq, u32 ack,
L
Linus Torvalds 已提交
629 630
			    u32 win, u32 ts)
{
631
	struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
632 633
	struct {
		struct tcphdr th;
634
		__be32 opt[(TCPOLEN_TSTAMP_ALIGNED >> 2)
635
#ifdef CONFIG_TCP_MD5SIG
636
			   + (TCPOLEN_MD5SIG_ALIGNED >> 2)
637 638
#endif
			];
L
Linus Torvalds 已提交
639 640
	} rep;
	struct ip_reply_arg arg;
641 642 643 644
#ifdef CONFIG_TCP_MD5SIG
	struct tcp_md5sig_key *key;
	struct tcp_md5sig_key tw_key;
#endif
L
Linus Torvalds 已提交
645 646

	memset(&rep.th, 0, sizeof(struct tcphdr));
647
	memset(&arg, 0, sizeof(arg));
L
Linus Torvalds 已提交
648 649 650 651

	arg.iov[0].iov_base = (unsigned char *)&rep;
	arg.iov[0].iov_len  = sizeof(rep.th);
	if (ts) {
652 653 654 655 656
		rep.opt[0] = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) |
				   (TCPOPT_TIMESTAMP << 8) |
				   TCPOLEN_TIMESTAMP);
		rep.opt[1] = htonl(tcp_time_stamp);
		rep.opt[2] = htonl(ts);
657
		arg.iov[0].iov_len += TCPOLEN_TSTAMP_ALIGNED;
L
Linus Torvalds 已提交
658 659 660 661 662 663 664 665 666 667 668
	}

	/* Swap the send and the receive. */
	rep.th.dest    = th->source;
	rep.th.source  = th->dest;
	rep.th.doff    = arg.iov[0].iov_len / 4;
	rep.th.seq     = htonl(seq);
	rep.th.ack_seq = htonl(ack);
	rep.th.ack     = 1;
	rep.th.window  = htons(win);

669 670 671 672 673 674 675 676 677 678
#ifdef CONFIG_TCP_MD5SIG
	/*
	 * The SKB holds an imcoming packet, but may not have a valid ->sk
	 * pointer. This is especially the case when we're dealing with a
	 * TIME_WAIT ack, because the sk structure is long gone, and only
	 * the tcp_timewait_sock remains. So the md5 key is stashed in that
	 * structure, and we use it in preference.  I believe that (twsk ||
	 * skb->sk) holds true, but we program defensively.
	 */
	if (!twsk && skb->sk) {
679
		key = tcp_v4_md5_do_lookup(skb->sk, ip_hdr(skb)->daddr);
680 681 682 683
	} else if (twsk && twsk->tw_md5_keylen) {
		tw_key.key = twsk->tw_md5_key;
		tw_key.keylen = twsk->tw_md5_keylen;
		key = &tw_key;
684
	} else
685 686 687 688 689 690 691 692 693 694 695 696 697 698
		key = NULL;

	if (key) {
		int offset = (ts) ? 3 : 0;

		rep.opt[offset++] = htonl((TCPOPT_NOP << 24) |
					  (TCPOPT_NOP << 16) |
					  (TCPOPT_MD5SIG << 8) |
					  TCPOLEN_MD5SIG);
		arg.iov[0].iov_len += TCPOLEN_MD5SIG_ALIGNED;
		rep.th.doff = arg.iov[0].iov_len/4;

		tcp_v4_do_calc_md5_hash((__u8 *)&rep.opt[offset],
					key,
699 700
					ip_hdr(skb)->daddr,
					ip_hdr(skb)->saddr,
701 702 703 704
					&rep.th, IPPROTO_TCP,
					arg.iov[0].iov_len);
	}
#endif
705 706
	arg.csum = csum_tcpudp_nofold(ip_hdr(skb)->daddr,
				      ip_hdr(skb)->saddr, /* XXX */
L
Linus Torvalds 已提交
707 708
				      arg.iov[0].iov_len, IPPROTO_TCP, 0);
	arg.csumoffset = offsetof(struct tcphdr, check) / 2;
709 710
	if (twsk)
		arg.bound_dev_if = twsk->tw_sk.tw_bound_dev_if;
L
Linus Torvalds 已提交
711 712 713 714 715 716 717 718

	ip_send_reply(tcp_socket->sk, skb, &arg, arg.iov[0].iov_len);

	TCP_INC_STATS_BH(TCP_MIB_OUTSEGS);
}

static void tcp_v4_timewait_ack(struct sock *sk, struct sk_buff *skb)
{
719
	struct inet_timewait_sock *tw = inet_twsk(sk);
720
	struct tcp_timewait_sock *tcptw = tcp_twsk(sk);
L
Linus Torvalds 已提交
721

722
	tcp_v4_send_ack(tcptw, skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt,
723 724
			tcptw->tw_rcv_wnd >> tw->tw_rcv_wscale,
			tcptw->tw_ts_recent);
L
Linus Torvalds 已提交
725

726
	inet_twsk_put(tw);
L
Linus Torvalds 已提交
727 728
}

729 730
static void tcp_v4_reqsk_send_ack(struct sk_buff *skb,
				  struct request_sock *req)
L
Linus Torvalds 已提交
731
{
732 733
	tcp_v4_send_ack(NULL, skb, tcp_rsk(req)->snt_isn + 1,
			tcp_rsk(req)->rcv_isn + 1, req->rcv_wnd,
L
Linus Torvalds 已提交
734 735 736 737 738
			req->ts_recent);
}

/*
 *	Send a SYN-ACK after having received an ACK.
739
 *	This still operates on a request_sock only, not on a big
L
Linus Torvalds 已提交
740 741
 *	socket.
 */
742
static int tcp_v4_send_synack(struct sock *sk, struct request_sock *req,
L
Linus Torvalds 已提交
743 744
			      struct dst_entry *dst)
{
745
	const struct inet_request_sock *ireq = inet_rsk(req);
L
Linus Torvalds 已提交
746 747 748 749
	int err = -1;
	struct sk_buff * skb;

	/* First, grab a route. */
750
	if (!dst && (dst = inet_csk_route_req(sk, req)) == NULL)
L
Linus Torvalds 已提交
751 752 753 754 755
		goto out;

	skb = tcp_make_synack(sk, dst, req);

	if (skb) {
756
		struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
757

758
		th->check = tcp_v4_check(skb->len,
759 760
					 ireq->loc_addr,
					 ireq->rmt_addr,
L
Linus Torvalds 已提交
761 762 763
					 csum_partial((char *)th, skb->len,
						      skb->csum));

764 765 766
		err = ip_build_and_send_pkt(skb, sk, ireq->loc_addr,
					    ireq->rmt_addr,
					    ireq->opt);
767
		err = net_xmit_eval(err);
L
Linus Torvalds 已提交
768 769 770 771 772 773 774 775
	}

out:
	dst_release(dst);
	return err;
}

/*
776
 *	IPv4 request_sock destructor.
L
Linus Torvalds 已提交
777
 */
778
static void tcp_v4_reqsk_destructor(struct request_sock *req)
L
Linus Torvalds 已提交
779
{
J
Jesper Juhl 已提交
780
	kfree(inet_rsk(req)->opt);
L
Linus Torvalds 已提交
781 782
}

783
#ifdef CONFIG_SYN_COOKIES
S
Stephen Hemminger 已提交
784
static void syn_flood_warning(struct sk_buff *skb)
L
Linus Torvalds 已提交
785 786 787 788 789 790 791
{
	static unsigned long warntime;

	if (time_after(jiffies, (warntime + HZ * 60))) {
		warntime = jiffies;
		printk(KERN_INFO
		       "possible SYN flooding on port %d. Sending cookies.\n",
792
		       ntohs(tcp_hdr(skb)->dest));
L
Linus Torvalds 已提交
793 794
	}
}
795
#endif
L
Linus Torvalds 已提交
796 797

/*
798
 * Save and compile IPv4 options into the request_sock if needed.
L
Linus Torvalds 已提交
799
 */
S
Stephen Hemminger 已提交
800 801
static struct ip_options *tcp_v4_save_options(struct sock *sk,
					      struct sk_buff *skb)
L
Linus Torvalds 已提交
802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818
{
	struct ip_options *opt = &(IPCB(skb)->opt);
	struct ip_options *dopt = NULL;

	if (opt && opt->optlen) {
		int opt_size = optlength(opt);
		dopt = kmalloc(opt_size, GFP_ATOMIC);
		if (dopt) {
			if (ip_options_echo(dopt, skb)) {
				kfree(dopt);
				dopt = NULL;
			}
		}
	}
	return dopt;
}

819 820 821 822 823 824 825 826
#ifdef CONFIG_TCP_MD5SIG
/*
 * RFC2385 MD5 checksumming requires a mapping of
 * IP address->MD5 Key.
 * We need to maintain these in the sk structure.
 */

/* Find the Key structure for an address.  */
827 828
static struct tcp_md5sig_key *
			tcp_v4_md5_do_lookup(struct sock *sk, __be32 addr)
829 830 831 832 833 834 835 836
{
	struct tcp_sock *tp = tcp_sk(sk);
	int i;

	if (!tp->md5sig_info || !tp->md5sig_info->entries4)
		return NULL;
	for (i = 0; i < tp->md5sig_info->entries4; i++) {
		if (tp->md5sig_info->keys4[i].addr == addr)
837
			return &tp->md5sig_info->keys4[i].base;
838 839 840 841 842 843 844 845 846 847 848 849
	}
	return NULL;
}

struct tcp_md5sig_key *tcp_v4_md5_lookup(struct sock *sk,
					 struct sock *addr_sk)
{
	return tcp_v4_md5_do_lookup(sk, inet_sk(addr_sk)->daddr);
}

EXPORT_SYMBOL(tcp_v4_md5_lookup);

A
Adrian Bunk 已提交
850 851
static struct tcp_md5sig_key *tcp_v4_reqsk_md5_lookup(struct sock *sk,
						      struct request_sock *req)
852 853 854 855 856 857 858 859 860
{
	return tcp_v4_md5_do_lookup(sk, inet_rsk(req)->rmt_addr);
}

/* This can be called on a newly created socket, from other files */
int tcp_v4_md5_do_add(struct sock *sk, __be32 addr,
		      u8 *newkey, u8 newkeylen)
{
	/* Add Key to the list */
861
	struct tcp_md5sig_key *key;
862 863 864
	struct tcp_sock *tp = tcp_sk(sk);
	struct tcp4_md5sig_key *keys;

865
	key = tcp_v4_md5_do_lookup(sk, addr);
866 867
	if (key) {
		/* Pre-existing entry - just update that one. */
868 869 870
		kfree(key->key);
		key->key = newkey;
		key->keylen = newkeylen;
871
	} else {
872 873
		struct tcp_md5sig_info *md5sig;

874
		if (!tp->md5sig_info) {
875 876
			tp->md5sig_info = kzalloc(sizeof(*tp->md5sig_info),
						  GFP_ATOMIC);
877 878 879 880
			if (!tp->md5sig_info) {
				kfree(newkey);
				return -ENOMEM;
			}
881
			sk->sk_route_caps &= ~NETIF_F_GSO_MASK;
882 883 884 885 886
		}
		if (tcp_alloc_md5sig_pool() == NULL) {
			kfree(newkey);
			return -ENOMEM;
		}
887 888 889 890
		md5sig = tp->md5sig_info;

		if (md5sig->alloced4 == md5sig->entries4) {
			keys = kmalloc((sizeof(*keys) *
891
					(md5sig->entries4 + 1)), GFP_ATOMIC);
892 893 894 895 896 897
			if (!keys) {
				kfree(newkey);
				tcp_free_md5sig_pool();
				return -ENOMEM;
			}

898 899 900
			if (md5sig->entries4)
				memcpy(keys, md5sig->keys4,
				       sizeof(*keys) * md5sig->entries4);
901 902

			/* Free old key list, and reference new one */
903 904 905 906
			if (md5sig->keys4)
				kfree(md5sig->keys4);
			md5sig->keys4 = keys;
			md5sig->alloced4++;
907
		}
908
		md5sig->entries4++;
909 910 911
		md5sig->keys4[md5sig->entries4 - 1].addr        = addr;
		md5sig->keys4[md5sig->entries4 - 1].base.key    = newkey;
		md5sig->keys4[md5sig->entries4 - 1].base.keylen = newkeylen;
912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932
	}
	return 0;
}

EXPORT_SYMBOL(tcp_v4_md5_do_add);

static int tcp_v4_md5_add_func(struct sock *sk, struct sock *addr_sk,
			       u8 *newkey, u8 newkeylen)
{
	return tcp_v4_md5_do_add(sk, inet_sk(addr_sk)->daddr,
				 newkey, newkeylen);
}

int tcp_v4_md5_do_del(struct sock *sk, __be32 addr)
{
	struct tcp_sock *tp = tcp_sk(sk);
	int i;

	for (i = 0; i < tp->md5sig_info->entries4; i++) {
		if (tp->md5sig_info->keys4[i].addr == addr) {
			/* Free the key */
933
			kfree(tp->md5sig_info->keys4[i].base.key);
934 935 936 937 938
			tp->md5sig_info->entries4--;

			if (tp->md5sig_info->entries4 == 0) {
				kfree(tp->md5sig_info->keys4);
				tp->md5sig_info->keys4 = NULL;
939
				tp->md5sig_info->alloced4 = 0;
940
			} else if (tp->md5sig_info->entries4 != i) {
941
				/* Need to do some manipulation */
942 943 944
				memcpy(&tp->md5sig_info->keys4[i],
				       &tp->md5sig_info->keys4[i+1],
				       (tp->md5sig_info->entries4 - i) *
945
					sizeof(struct tcp4_md5sig_key));
946 947 948 949 950 951 952 953 954 955
			}
			tcp_free_md5sig_pool();
			return 0;
		}
	}
	return -ENOENT;
}

EXPORT_SYMBOL(tcp_v4_md5_do_del);

956
static void tcp_v4_clear_md5_list(struct sock *sk)
957 958 959 960 961 962 963 964 965 966
{
	struct tcp_sock *tp = tcp_sk(sk);

	/* Free each key, then the set of key keys,
	 * the crypto element, and then decrement our
	 * hold on the last resort crypto.
	 */
	if (tp->md5sig_info->entries4) {
		int i;
		for (i = 0; i < tp->md5sig_info->entries4; i++)
967
			kfree(tp->md5sig_info->keys4[i].base.key);
968 969 970 971 972 973 974 975 976 977
		tp->md5sig_info->entries4 = 0;
		tcp_free_md5sig_pool();
	}
	if (tp->md5sig_info->keys4) {
		kfree(tp->md5sig_info->keys4);
		tp->md5sig_info->keys4 = NULL;
		tp->md5sig_info->alloced4  = 0;
	}
}

978 979
static int tcp_v4_parse_md5_keys(struct sock *sk, char __user *optval,
				 int optlen)
980 981 982 983 984 985 986 987
{
	struct tcp_md5sig cmd;
	struct sockaddr_in *sin = (struct sockaddr_in *)&cmd.tcpm_addr;
	u8 *newkey;

	if (optlen < sizeof(cmd))
		return -EINVAL;

988
	if (copy_from_user(&cmd, optval, sizeof(cmd)))
989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004
		return -EFAULT;

	if (sin->sin_family != AF_INET)
		return -EINVAL;

	if (!cmd.tcpm_key || !cmd.tcpm_keylen) {
		if (!tcp_sk(sk)->md5sig_info)
			return -ENOENT;
		return tcp_v4_md5_do_del(sk, sin->sin_addr.s_addr);
	}

	if (cmd.tcpm_keylen > TCP_MD5SIG_MAXKEYLEN)
		return -EINVAL;

	if (!tcp_sk(sk)->md5sig_info) {
		struct tcp_sock *tp = tcp_sk(sk);
1005
		struct tcp_md5sig_info *p = kzalloc(sizeof(*p), GFP_KERNEL);
1006 1007 1008 1009 1010

		if (!p)
			return -EINVAL;

		tp->md5sig_info = p;
1011
		sk->sk_route_caps &= ~NETIF_F_GSO_MASK;
1012 1013
	}

1014
	newkey = kmemdup(cmd.tcpm_key, cmd.tcpm_keylen, GFP_KERNEL);
1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028
	if (!newkey)
		return -ENOMEM;
	return tcp_v4_md5_do_add(sk, sin->sin_addr.s_addr,
				 newkey, cmd.tcpm_keylen);
}

static int tcp_v4_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key,
				   __be32 saddr, __be32 daddr,
				   struct tcphdr *th, int protocol,
				   int tcplen)
{
	struct scatterlist sg[4];
	__u16 data_len;
	int block = 0;
1029
	__sum16 old_checksum;
1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057
	struct tcp_md5sig_pool *hp;
	struct tcp4_pseudohdr *bp;
	struct hash_desc *desc;
	int err;
	unsigned int nbytes = 0;

	/*
	 * Okay, so RFC2385 is turned on for this connection,
	 * so we need to generate the MD5 hash for the packet now.
	 */

	hp = tcp_get_md5sig_pool();
	if (!hp)
		goto clear_hash_noput;

	bp = &hp->md5_blk.ip4;
	desc = &hp->md5_desc;

	/*
	 * 1. the TCP pseudo-header (in the order: source IP address,
	 * destination IP address, zero-padded protocol number, and
	 * segment length)
	 */
	bp->saddr = saddr;
	bp->daddr = daddr;
	bp->pad = 0;
	bp->protocol = protocol;
	bp->len = htons(tcplen);
1058 1059 1060

	sg_init_table(sg, 4);

1061 1062 1063 1064 1065 1066 1067 1068 1069 1070
	sg_set_buf(&sg[block++], bp, sizeof(*bp));
	nbytes += sizeof(*bp);

	/* 2. the TCP header, excluding options, and assuming a
	 * checksum of zero/
	 */
	old_checksum = th->check;
	th->check = 0;
	sg_set_buf(&sg[block++], th, sizeof(struct tcphdr));
	nbytes += sizeof(struct tcphdr);
1071

1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085
	/* 3. the TCP segment data (if any) */
	data_len = tcplen - (th->doff << 2);
	if (data_len > 0) {
		unsigned char *data = (unsigned char *)th + (th->doff << 2);
		sg_set_buf(&sg[block++], data, data_len);
		nbytes += data_len;
	}

	/* 4. an independently-specified key or password, known to both
	 * TCPs and presumably connection-specific
	 */
	sg_set_buf(&sg[block++], key->key, key->keylen);
	nbytes += key->keylen;

J
Jens Axboe 已提交
1086
	sg_mark_end(&sg[block - 1]);
1087

1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136
	/* Now store the Hash into the packet */
	err = crypto_hash_init(desc);
	if (err)
		goto clear_hash;
	err = crypto_hash_update(desc, sg, nbytes);
	if (err)
		goto clear_hash;
	err = crypto_hash_final(desc, md5_hash);
	if (err)
		goto clear_hash;

	/* Reset header, and free up the crypto */
	tcp_put_md5sig_pool();
	th->check = old_checksum;

out:
	return 0;
clear_hash:
	tcp_put_md5sig_pool();
clear_hash_noput:
	memset(md5_hash, 0, 16);
	goto out;
}

int tcp_v4_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key,
			 struct sock *sk,
			 struct dst_entry *dst,
			 struct request_sock *req,
			 struct tcphdr *th, int protocol,
			 int tcplen)
{
	__be32 saddr, daddr;

	if (sk) {
		saddr = inet_sk(sk)->saddr;
		daddr = inet_sk(sk)->daddr;
	} else {
		struct rtable *rt = (struct rtable *)dst;
		BUG_ON(!rt);
		saddr = rt->rt_src;
		daddr = rt->rt_dst;
	}
	return tcp_v4_do_calc_md5_hash(md5_hash, key,
				       saddr, daddr,
				       th, protocol, tcplen);
}

EXPORT_SYMBOL(tcp_v4_calc_md5_hash);

1137
static int tcp_v4_inbound_md5_hash(struct sock *sk, struct sk_buff *skb)
1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148
{
	/*
	 * This gets called for each TCP segment that arrives
	 * so we want to be efficient.
	 * We have 3 drop cases:
	 * o No MD5 hash and one expected.
	 * o MD5 hash and we're not expecting one.
	 * o MD5 hash and its wrong.
	 */
	__u8 *hash_location = NULL;
	struct tcp_md5sig_key *hash_expected;
1149
	const struct iphdr *iph = ip_hdr(skb);
1150
	struct tcphdr *th = tcp_hdr(skb);
1151
	int length = (th->doff << 2) - sizeof(struct tcphdr);
1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201
	int genhash;
	unsigned char *ptr;
	unsigned char newhash[16];

	hash_expected = tcp_v4_md5_do_lookup(sk, iph->saddr);

	/*
	 * If the TCP option length is less than the TCP_MD5SIG
	 * option length, then we can shortcut
	 */
	if (length < TCPOLEN_MD5SIG) {
		if (hash_expected)
			return 1;
		else
			return 0;
	}

	/* Okay, we can't shortcut - we have to grub through the options */
	ptr = (unsigned char *)(th + 1);
	while (length > 0) {
		int opcode = *ptr++;
		int opsize;

		switch (opcode) {
		case TCPOPT_EOL:
			goto done_opts;
		case TCPOPT_NOP:
			length--;
			continue;
		default:
			opsize = *ptr++;
			if (opsize < 2)
				goto done_opts;
			if (opsize > length)
				goto done_opts;

			if (opcode == TCPOPT_MD5SIG) {
				hash_location = ptr;
				goto done_opts;
			}
		}
		ptr += opsize-2;
		length -= opsize;
	}
done_opts:
	/* We've parsed the options - do we have a hash? */
	if (!hash_expected && !hash_location)
		return 0;

	if (hash_expected && !hash_location) {
1202
		LIMIT_NETDEBUG(KERN_INFO "MD5 Hash expected but NOT found "
1203
			       "(" NIPQUAD_FMT ", %d)->(" NIPQUAD_FMT ", %d)\n",
1204 1205
			       NIPQUAD(iph->saddr), ntohs(th->source),
			       NIPQUAD(iph->daddr), ntohs(th->dest));
1206 1207 1208 1209
		return 1;
	}

	if (!hash_expected && hash_location) {
1210
		LIMIT_NETDEBUG(KERN_INFO "MD5 Hash NOT expected but found "
1211
			       "(" NIPQUAD_FMT ", %d)->(" NIPQUAD_FMT ", %d)\n",
1212 1213
			       NIPQUAD(iph->saddr), ntohs(th->source),
			       NIPQUAD(iph->daddr), ntohs(th->dest));
1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229
		return 1;
	}

	/* Okay, so this is hash_expected and hash_location -
	 * so we need to calculate the checksum.
	 */
	genhash = tcp_v4_do_calc_md5_hash(newhash,
					  hash_expected,
					  iph->saddr, iph->daddr,
					  th, sk->sk_protocol,
					  skb->len);

	if (genhash || memcmp(hash_location, newhash, 16) != 0) {
		if (net_ratelimit()) {
			printk(KERN_INFO "MD5 Hash failed for "
			       "(" NIPQUAD_FMT ", %d)->(" NIPQUAD_FMT ", %d)%s\n",
1230 1231
			       NIPQUAD(iph->saddr), ntohs(th->source),
			       NIPQUAD(iph->daddr), ntohs(th->dest),
1232 1233 1234 1235 1236 1237 1238 1239 1240
			       genhash ? " tcp_v4_calc_md5_hash failed" : "");
		}
		return 1;
	}
	return 0;
}

#endif

1241
struct request_sock_ops tcp_request_sock_ops __read_mostly = {
L
Linus Torvalds 已提交
1242
	.family		=	PF_INET,
1243
	.obj_size	=	sizeof(struct tcp_request_sock),
L
Linus Torvalds 已提交
1244
	.rtx_syn_ack	=	tcp_v4_send_synack,
1245 1246
	.send_ack	=	tcp_v4_reqsk_send_ack,
	.destructor	=	tcp_v4_reqsk_destructor,
L
Linus Torvalds 已提交
1247 1248 1249
	.send_reset	=	tcp_v4_send_reset,
};

1250
#ifdef CONFIG_TCP_MD5SIG
1251
static struct tcp_request_sock_ops tcp_request_sock_ipv4_ops = {
1252 1253
	.md5_lookup	=	tcp_v4_reqsk_md5_lookup,
};
1254
#endif
1255

1256 1257 1258
static struct timewait_sock_ops tcp_timewait_sock_ops = {
	.twsk_obj_size	= sizeof(struct tcp_timewait_sock),
	.twsk_unique	= tcp_twsk_unique,
1259
	.twsk_destructor= tcp_twsk_destructor,
1260 1261
};

L
Linus Torvalds 已提交
1262 1263
int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
{
1264
	struct inet_request_sock *ireq;
L
Linus Torvalds 已提交
1265
	struct tcp_options_received tmp_opt;
1266
	struct request_sock *req;
1267 1268
	__be32 saddr = ip_hdr(skb)->saddr;
	__be32 daddr = ip_hdr(skb)->daddr;
L
Linus Torvalds 已提交
1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285
	__u32 isn = TCP_SKB_CB(skb)->when;
	struct dst_entry *dst = NULL;
#ifdef CONFIG_SYN_COOKIES
	int want_cookie = 0;
#else
#define want_cookie 0 /* Argh, why doesn't gcc optimize this :( */
#endif

	/* Never answer to SYNs send to broadcast or multicast */
	if (((struct rtable *)skb->dst)->rt_flags &
	    (RTCF_BROADCAST | RTCF_MULTICAST))
		goto drop;

	/* TW buckets are converted to open requests without
	 * limitations, they conserve resources and peer is
	 * evidently real one.
	 */
1286
	if (inet_csk_reqsk_queue_is_full(sk) && !isn) {
L
Linus Torvalds 已提交
1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299
#ifdef CONFIG_SYN_COOKIES
		if (sysctl_tcp_syncookies) {
			want_cookie = 1;
		} else
#endif
		goto drop;
	}

	/* Accept backlog is full. If we have already queued enough
	 * of warm entries in syn queue, drop request. It is better than
	 * clogging syn queue with openreqs with exponentially increasing
	 * timeout.
	 */
1300
	if (sk_acceptq_is_full(sk) && inet_csk_reqsk_queue_young(sk) > 1)
L
Linus Torvalds 已提交
1301 1302
		goto drop;

1303
	req = reqsk_alloc(&tcp_request_sock_ops);
L
Linus Torvalds 已提交
1304 1305 1306
	if (!req)
		goto drop;

1307 1308 1309 1310
#ifdef CONFIG_TCP_MD5SIG
	tcp_rsk(req)->af_specific = &tcp_request_sock_ipv4_ops;
#endif

L
Linus Torvalds 已提交
1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334
	tcp_clear_options(&tmp_opt);
	tmp_opt.mss_clamp = 536;
	tmp_opt.user_mss  = tcp_sk(sk)->rx_opt.user_mss;

	tcp_parse_options(skb, &tmp_opt, 0);

	if (want_cookie) {
		tcp_clear_options(&tmp_opt);
		tmp_opt.saw_tstamp = 0;
	}

	if (tmp_opt.saw_tstamp && !tmp_opt.rcv_tsval) {
		/* Some OSes (unknown ones, but I see them on web server, which
		 * contains information interesting only for windows'
		 * users) do not send their stamp in SYN. It is easy case.
		 * We simply do not advertise TS support.
		 */
		tmp_opt.saw_tstamp = 0;
		tmp_opt.tstamp_ok  = 0;
	}
	tmp_opt.tstamp_ok = tmp_opt.saw_tstamp;

	tcp_openreq_init(req, &tmp_opt, skb);

1335 1336 1337
	if (security_inet_conn_request(sk, skb, req))
		goto drop_and_free;

1338 1339 1340 1341
	ireq = inet_rsk(req);
	ireq->loc_addr = daddr;
	ireq->rmt_addr = saddr;
	ireq->opt = tcp_v4_save_options(sk, skb);
L
Linus Torvalds 已提交
1342
	if (!want_cookie)
1343
		TCP_ECN_create_request(req, tcp_hdr(skb));
L
Linus Torvalds 已提交
1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362

	if (want_cookie) {
#ifdef CONFIG_SYN_COOKIES
		syn_flood_warning(skb);
#endif
		isn = cookie_v4_init_sequence(sk, skb, &req->mss);
	} else if (!isn) {
		struct inet_peer *peer = NULL;

		/* VJ's idea. We save last timestamp seen
		 * from the destination in peer table, when entering
		 * state TIME-WAIT, and check against it before
		 * accepting new connection request.
		 *
		 * If "isn" is not zero, this request hit alive
		 * timewait bucket, so that all the necessary checks
		 * are made in the function processing timewait state.
		 */
		if (tmp_opt.saw_tstamp &&
1363
		    tcp_death_row.sysctl_tw_recycle &&
1364
		    (dst = inet_csk_route_req(sk, req)) != NULL &&
L
Linus Torvalds 已提交
1365 1366
		    (peer = rt_get_peer((struct rtable *)dst)) != NULL &&
		    peer->v4daddr == saddr) {
1367
			if (get_seconds() < peer->tcp_ts_stamp + TCP_PAWS_MSL &&
L
Linus Torvalds 已提交
1368 1369 1370 1371 1372 1373 1374 1375 1376
			    (s32)(peer->tcp_ts - req->ts_recent) >
							TCP_PAWS_WINDOW) {
				NET_INC_STATS_BH(LINUX_MIB_PAWSPASSIVEREJECTED);
				dst_release(dst);
				goto drop_and_free;
			}
		}
		/* Kill the following clause, if you dislike this way. */
		else if (!sysctl_tcp_syncookies &&
1377
			 (sysctl_max_syn_backlog - inet_csk_reqsk_queue_len(sk) <
L
Linus Torvalds 已提交
1378 1379 1380 1381 1382 1383 1384 1385 1386 1387
			  (sysctl_max_syn_backlog >> 2)) &&
			 (!peer || !peer->tcp_ts_stamp) &&
			 (!dst || !dst_metric(dst, RTAX_RTT))) {
			/* Without syncookies last quarter of
			 * backlog is filled with destinations,
			 * proven to be alive.
			 * It means that we continue to communicate
			 * to destinations, already remembered
			 * to the moment of synflood.
			 */
1388 1389 1390
			LIMIT_NETDEBUG(KERN_DEBUG "TCP: drop open "
				       "request from %u.%u.%u.%u/%u\n",
				       NIPQUAD(saddr),
1391
				       ntohs(tcp_hdr(skb)->source));
L
Linus Torvalds 已提交
1392 1393 1394 1395
			dst_release(dst);
			goto drop_and_free;
		}

1396
		isn = tcp_v4_init_sequence(skb);
L
Linus Torvalds 已提交
1397
	}
1398
	tcp_rsk(req)->snt_isn = isn;
L
Linus Torvalds 已提交
1399 1400 1401 1402 1403

	if (tcp_v4_send_synack(sk, req, dst))
		goto drop_and_free;

	if (want_cookie) {
1404
		reqsk_free(req);
L
Linus Torvalds 已提交
1405
	} else {
1406
		inet_csk_reqsk_queue_hash_add(sk, req, TCP_TIMEOUT_INIT);
L
Linus Torvalds 已提交
1407 1408 1409 1410
	}
	return 0;

drop_and_free:
1411
	reqsk_free(req);
L
Linus Torvalds 已提交
1412 1413 1414 1415 1416 1417 1418 1419 1420 1421
drop:
	return 0;
}


/*
 * The three way handshake has completed - we got a valid synack -
 * now create the new socket.
 */
struct sock *tcp_v4_syn_recv_sock(struct sock *sk, struct sk_buff *skb,
1422
				  struct request_sock *req,
L
Linus Torvalds 已提交
1423 1424
				  struct dst_entry *dst)
{
1425
	struct inet_request_sock *ireq;
L
Linus Torvalds 已提交
1426 1427 1428
	struct inet_sock *newinet;
	struct tcp_sock *newtp;
	struct sock *newsk;
1429 1430 1431
#ifdef CONFIG_TCP_MD5SIG
	struct tcp_md5sig_key *key;
#endif
L
Linus Torvalds 已提交
1432 1433 1434 1435

	if (sk_acceptq_is_full(sk))
		goto exit_overflow;

1436
	if (!dst && (dst = inet_csk_route_req(sk, req)) == NULL)
L
Linus Torvalds 已提交
1437 1438 1439 1440 1441 1442
		goto exit;

	newsk = tcp_create_openreq_child(sk, req, skb);
	if (!newsk)
		goto exit;

1443
	newsk->sk_gso_type = SKB_GSO_TCPV4;
1444
	sk_setup_caps(newsk, dst);
L
Linus Torvalds 已提交
1445 1446 1447

	newtp		      = tcp_sk(newsk);
	newinet		      = inet_sk(newsk);
1448 1449 1450 1451 1452 1453
	ireq		      = inet_rsk(req);
	newinet->daddr	      = ireq->rmt_addr;
	newinet->rcv_saddr    = ireq->loc_addr;
	newinet->saddr	      = ireq->loc_addr;
	newinet->opt	      = ireq->opt;
	ireq->opt	      = NULL;
1454
	newinet->mc_index     = inet_iif(skb);
1455
	newinet->mc_ttl	      = ip_hdr(skb)->ttl;
1456
	inet_csk(newsk)->icsk_ext_hdr_len = 0;
L
Linus Torvalds 已提交
1457
	if (newinet->opt)
1458
		inet_csk(newsk)->icsk_ext_hdr_len = newinet->opt->optlen;
L
Linus Torvalds 已提交
1459 1460
	newinet->id = newtp->write_seq ^ jiffies;

J
John Heffner 已提交
1461
	tcp_mtup_init(newsk);
L
Linus Torvalds 已提交
1462 1463 1464 1465
	tcp_sync_mss(newsk, dst_mtu(dst));
	newtp->advmss = dst_metric(dst, RTAX_ADVMSS);
	tcp_initialize_rcv_mss(newsk);

1466 1467 1468 1469 1470 1471 1472 1473 1474
#ifdef CONFIG_TCP_MD5SIG
	/* Copy over the MD5 key from the original socket */
	if ((key = tcp_v4_md5_do_lookup(sk, newinet->daddr)) != NULL) {
		/*
		 * We're using one, so create a matching key
		 * on the newsk structure. If we fail to get
		 * memory, then we end up not copying the key
		 * across. Shucks.
		 */
1475 1476
		char *newkey = kmemdup(key->key, key->keylen, GFP_ATOMIC);
		if (newkey != NULL)
1477 1478 1479 1480 1481
			tcp_v4_md5_do_add(newsk, inet_sk(sk)->daddr,
					  newkey, key->keylen);
	}
#endif

1482
	__inet_hash(&tcp_hashinfo, newsk, 0);
1483
	__inet_inherit_port(&tcp_hashinfo, sk, newsk);
L
Linus Torvalds 已提交
1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496

	return newsk;

exit_overflow:
	NET_INC_STATS_BH(LINUX_MIB_LISTENOVERFLOWS);
exit:
	NET_INC_STATS_BH(LINUX_MIB_LISTENDROPS);
	dst_release(dst);
	return NULL;
}

static struct sock *tcp_v4_hnd_req(struct sock *sk, struct sk_buff *skb)
{
1497
	struct tcphdr *th = tcp_hdr(skb);
1498
	const struct iphdr *iph = ip_hdr(skb);
L
Linus Torvalds 已提交
1499
	struct sock *nsk;
1500
	struct request_sock **prev;
L
Linus Torvalds 已提交
1501
	/* Find possible connection requests. */
1502 1503
	struct request_sock *req = inet_csk_search_req(sk, &prev, th->source,
						       iph->saddr, iph->daddr);
L
Linus Torvalds 已提交
1504 1505 1506
	if (req)
		return tcp_check_req(sk, skb, req, prev);

1507 1508
	nsk = inet_lookup_established(&tcp_hashinfo, iph->saddr, th->source,
				      iph->daddr, th->dest, inet_iif(skb));
L
Linus Torvalds 已提交
1509 1510 1511 1512 1513 1514

	if (nsk) {
		if (nsk->sk_state != TCP_TIME_WAIT) {
			bh_lock_sock(nsk);
			return nsk;
		}
1515
		inet_twsk_put(inet_twsk(nsk));
L
Linus Torvalds 已提交
1516 1517 1518 1519 1520 1521 1522 1523 1524 1525
		return NULL;
	}

#ifdef CONFIG_SYN_COOKIES
	if (!th->rst && !th->syn && th->ack)
		sk = cookie_v4_check(sk, skb, &(IPCB(skb)->opt));
#endif
	return sk;
}

1526
static __sum16 tcp_v4_checksum_init(struct sk_buff *skb)
L
Linus Torvalds 已提交
1527
{
1528 1529
	const struct iphdr *iph = ip_hdr(skb);

1530
	if (skb->ip_summed == CHECKSUM_COMPLETE) {
1531 1532
		if (!tcp_v4_check(skb->len, iph->saddr,
				  iph->daddr, skb->csum)) {
1533
			skb->ip_summed = CHECKSUM_UNNECESSARY;
L
Linus Torvalds 已提交
1534
			return 0;
1535
		}
L
Linus Torvalds 已提交
1536
	}
1537

1538
	skb->csum = csum_tcpudp_nofold(iph->saddr, iph->daddr,
1539 1540
				       skb->len, IPPROTO_TCP, 0);

L
Linus Torvalds 已提交
1541
	if (skb->len <= 76) {
1542
		return __skb_checksum_complete(skb);
L
Linus Torvalds 已提交
1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557
	}
	return 0;
}


/* The socket must have it's spinlock held when we get
 * here.
 *
 * We have a potential double-lock case here, so even when
 * doing backlog processing we use the BH locking scheme.
 * This is because we cannot sleep with the original spinlock
 * held.
 */
int tcp_v4_do_rcv(struct sock *sk, struct sk_buff *skb)
{
1558 1559 1560 1561 1562 1563 1564 1565
	struct sock *rsk;
#ifdef CONFIG_TCP_MD5SIG
	/*
	 * We really want to reject the packet as early as possible
	 * if:
	 *  o We're expecting an MD5'd packet and this is no MD5 tcp option
	 *  o There is an MD5 option and we're not expecting one
	 */
1566
	if (tcp_v4_inbound_md5_hash(sk, skb))
1567 1568 1569
		goto discard;
#endif

L
Linus Torvalds 已提交
1570 1571
	if (sk->sk_state == TCP_ESTABLISHED) { /* Fast path */
		TCP_CHECK_TIMER(sk);
1572
		if (tcp_rcv_established(sk, skb, tcp_hdr(skb), skb->len)) {
1573
			rsk = sk;
L
Linus Torvalds 已提交
1574
			goto reset;
1575
		}
L
Linus Torvalds 已提交
1576 1577 1578 1579
		TCP_CHECK_TIMER(sk);
		return 0;
	}

1580
	if (skb->len < tcp_hdrlen(skb) || tcp_checksum_complete(skb))
L
Linus Torvalds 已提交
1581 1582 1583 1584 1585 1586 1587 1588
		goto csum_err;

	if (sk->sk_state == TCP_LISTEN) {
		struct sock *nsk = tcp_v4_hnd_req(sk, skb);
		if (!nsk)
			goto discard;

		if (nsk != sk) {
1589 1590
			if (tcp_child_process(sk, nsk, skb)) {
				rsk = nsk;
L
Linus Torvalds 已提交
1591
				goto reset;
1592
			}
L
Linus Torvalds 已提交
1593 1594 1595 1596 1597
			return 0;
		}
	}

	TCP_CHECK_TIMER(sk);
1598
	if (tcp_rcv_state_process(sk, skb, tcp_hdr(skb), skb->len)) {
1599
		rsk = sk;
L
Linus Torvalds 已提交
1600
		goto reset;
1601
	}
L
Linus Torvalds 已提交
1602 1603 1604 1605
	TCP_CHECK_TIMER(sk);
	return 0;

reset:
1606
	tcp_v4_send_reset(rsk, skb);
L
Linus Torvalds 已提交
1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626
discard:
	kfree_skb(skb);
	/* Be careful here. If this function gets more complicated and
	 * gcc suffers from register pressure on the x86, sk (in %ebx)
	 * might be destroyed here. This current version compiles correctly,
	 * but you have been warned.
	 */
	return 0;

csum_err:
	TCP_INC_STATS_BH(TCP_MIB_INERRS);
	goto discard;
}

/*
 *	From tcp_input.c
 */

int tcp_v4_rcv(struct sk_buff *skb)
{
1627
	const struct iphdr *iph;
L
Linus Torvalds 已提交
1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640
	struct tcphdr *th;
	struct sock *sk;
	int ret;

	if (skb->pkt_type != PACKET_HOST)
		goto discard_it;

	/* Count it even if it's bad */
	TCP_INC_STATS_BH(TCP_MIB_INSEGS);

	if (!pskb_may_pull(skb, sizeof(struct tcphdr)))
		goto discard_it;

1641
	th = tcp_hdr(skb);
L
Linus Torvalds 已提交
1642 1643 1644 1645 1646 1647 1648 1649

	if (th->doff < sizeof(struct tcphdr) / 4)
		goto bad_packet;
	if (!pskb_may_pull(skb, th->doff * 4))
		goto discard_it;

	/* An explanation is required here, I think.
	 * Packet length and doff are validated by header prediction,
S
Stephen Hemminger 已提交
1650
	 * provided case of th->doff==0 is eliminated.
L
Linus Torvalds 已提交
1651
	 * So, we defer the checks. */
1652
	if (!skb_csum_unnecessary(skb) && tcp_v4_checksum_init(skb))
L
Linus Torvalds 已提交
1653 1654
		goto bad_packet;

1655
	th = tcp_hdr(skb);
1656
	iph = ip_hdr(skb);
L
Linus Torvalds 已提交
1657 1658 1659 1660 1661
	TCP_SKB_CB(skb)->seq = ntohl(th->seq);
	TCP_SKB_CB(skb)->end_seq = (TCP_SKB_CB(skb)->seq + th->syn + th->fin +
				    skb->len - th->doff * 4);
	TCP_SKB_CB(skb)->ack_seq = ntohl(th->ack_seq);
	TCP_SKB_CB(skb)->when	 = 0;
1662
	TCP_SKB_CB(skb)->flags	 = iph->tos;
L
Linus Torvalds 已提交
1663 1664
	TCP_SKB_CB(skb)->sacked	 = 0;

1665 1666
	sk = __inet_lookup(&tcp_hashinfo, iph->saddr, th->source,
			   iph->daddr, th->dest, inet_iif(skb));
L
Linus Torvalds 已提交
1667 1668 1669 1670 1671 1672 1673 1674 1675
	if (!sk)
		goto no_tcp_socket;

process:
	if (sk->sk_state == TCP_TIME_WAIT)
		goto do_time_wait;

	if (!xfrm4_policy_check(sk, XFRM_POLICY_IN, skb))
		goto discard_and_relse;
1676
	nf_reset(skb);
L
Linus Torvalds 已提交
1677

1678
	if (sk_filter(sk, skb))
L
Linus Torvalds 已提交
1679 1680 1681 1682
		goto discard_and_relse;

	skb->dev = NULL;

1683
	bh_lock_sock_nested(sk);
L
Linus Torvalds 已提交
1684 1685
	ret = 0;
	if (!sock_owned_by_user(sk)) {
C
Chris Leech 已提交
1686 1687 1688 1689 1690
#ifdef CONFIG_NET_DMA
		struct tcp_sock *tp = tcp_sk(sk);
		if (!tp->ucopy.dma_chan && tp->ucopy.pinned_list)
			tp->ucopy.dma_chan = get_softnet_dma();
		if (tp->ucopy.dma_chan)
L
Linus Torvalds 已提交
1691
			ret = tcp_v4_do_rcv(sk, skb);
C
Chris Leech 已提交
1692 1693 1694 1695 1696 1697
		else
#endif
		{
			if (!tcp_prequeue(sk, skb))
			ret = tcp_v4_do_rcv(sk, skb);
		}
L
Linus Torvalds 已提交
1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713
	} else
		sk_add_backlog(sk, skb);
	bh_unlock_sock(sk);

	sock_put(sk);

	return ret;

no_tcp_socket:
	if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb))
		goto discard_it;

	if (skb->len < (th->doff << 2) || tcp_checksum_complete(skb)) {
bad_packet:
		TCP_INC_STATS_BH(TCP_MIB_INERRS);
	} else {
1714
		tcp_v4_send_reset(NULL, skb);
L
Linus Torvalds 已提交
1715 1716 1717 1718 1719
	}

discard_it:
	/* Discard frame. */
	kfree_skb(skb);
1720
	return 0;
L
Linus Torvalds 已提交
1721 1722 1723 1724 1725 1726 1727

discard_and_relse:
	sock_put(sk);
	goto discard_it;

do_time_wait:
	if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) {
1728
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1729 1730 1731 1732 1733
		goto discard_it;
	}

	if (skb->len < (th->doff << 2) || tcp_checksum_complete(skb)) {
		TCP_INC_STATS_BH(TCP_MIB_INERRS);
1734
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1735 1736
		goto discard_it;
	}
1737
	switch (tcp_timewait_state_process(inet_twsk(sk), skb, th)) {
L
Linus Torvalds 已提交
1738
	case TCP_TW_SYN: {
1739
		struct sock *sk2 = inet_lookup_listener(&tcp_hashinfo,
1740
							iph->daddr, th->dest,
1741
							inet_iif(skb));
L
Linus Torvalds 已提交
1742
		if (sk2) {
1743 1744
			inet_twsk_deschedule(inet_twsk(sk), &tcp_death_row);
			inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784
			sk = sk2;
			goto process;
		}
		/* Fall through to ACK */
	}
	case TCP_TW_ACK:
		tcp_v4_timewait_ack(sk, skb);
		break;
	case TCP_TW_RST:
		goto no_tcp_socket;
	case TCP_TW_SUCCESS:;
	}
	goto discard_it;
}

/* VJ's idea. Save last timestamp seen from this destination
 * and hold it at least for normal timewait interval to use for duplicate
 * segment detection in subsequent connections, before they enter synchronized
 * state.
 */

int tcp_v4_remember_stamp(struct sock *sk)
{
	struct inet_sock *inet = inet_sk(sk);
	struct tcp_sock *tp = tcp_sk(sk);
	struct rtable *rt = (struct rtable *)__sk_dst_get(sk);
	struct inet_peer *peer = NULL;
	int release_it = 0;

	if (!rt || rt->rt_dst != inet->daddr) {
		peer = inet_getpeer(inet->daddr, 1);
		release_it = 1;
	} else {
		if (!rt->peer)
			rt_bind_peer(rt, 1);
		peer = rt->peer;
	}

	if (peer) {
		if ((s32)(peer->tcp_ts - tp->rx_opt.ts_recent) <= 0 ||
1785
		    (peer->tcp_ts_stamp + TCP_PAWS_MSL < get_seconds() &&
L
Linus Torvalds 已提交
1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797
		     peer->tcp_ts_stamp <= tp->rx_opt.ts_recent_stamp)) {
			peer->tcp_ts_stamp = tp->rx_opt.ts_recent_stamp;
			peer->tcp_ts = tp->rx_opt.ts_recent;
		}
		if (release_it)
			inet_putpeer(peer);
		return 1;
	}

	return 0;
}

1798
int tcp_v4_tw_remember_stamp(struct inet_timewait_sock *tw)
L
Linus Torvalds 已提交
1799
{
1800
	struct inet_peer *peer = inet_getpeer(tw->tw_daddr, 1);
L
Linus Torvalds 已提交
1801 1802

	if (peer) {
1803 1804 1805
		const struct tcp_timewait_sock *tcptw = tcp_twsk((struct sock *)tw);

		if ((s32)(peer->tcp_ts - tcptw->tw_ts_recent) <= 0 ||
1806
		    (peer->tcp_ts_stamp + TCP_PAWS_MSL < get_seconds() &&
1807 1808 1809
		     peer->tcp_ts_stamp <= tcptw->tw_ts_recent_stamp)) {
			peer->tcp_ts_stamp = tcptw->tw_ts_recent_stamp;
			peer->tcp_ts	   = tcptw->tw_ts_recent;
L
Linus Torvalds 已提交
1810 1811 1812 1813 1814 1815 1816 1817
		}
		inet_putpeer(peer);
		return 1;
	}

	return 0;
}

1818
struct inet_connection_sock_af_ops ipv4_specific = {
1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829
	.queue_xmit	   = ip_queue_xmit,
	.send_check	   = tcp_v4_send_check,
	.rebuild_header	   = inet_sk_rebuild_header,
	.conn_request	   = tcp_v4_conn_request,
	.syn_recv_sock	   = tcp_v4_syn_recv_sock,
	.remember_stamp	   = tcp_v4_remember_stamp,
	.net_header_len	   = sizeof(struct iphdr),
	.setsockopt	   = ip_setsockopt,
	.getsockopt	   = ip_getsockopt,
	.addr2sockaddr	   = inet_csk_addr2sockaddr,
	.sockaddr_len	   = sizeof(struct sockaddr_in),
1830
#ifdef CONFIG_COMPAT
1831 1832
	.compat_setsockopt = compat_ip_setsockopt,
	.compat_getsockopt = compat_ip_getsockopt,
1833
#endif
L
Linus Torvalds 已提交
1834 1835
};

1836
#ifdef CONFIG_TCP_MD5SIG
1837
static struct tcp_sock_af_ops tcp_sock_ipv4_specific = {
1838 1839 1840 1841 1842
	.md5_lookup		= tcp_v4_md5_lookup,
	.calc_md5_hash		= tcp_v4_calc_md5_hash,
	.md5_add		= tcp_v4_md5_add_func,
	.md5_parse		= tcp_v4_parse_md5_keys,
};
1843
#endif
1844

L
Linus Torvalds 已提交
1845 1846 1847 1848 1849
/* NOTE: A lot of things set to zero explicitly by call to
 *       sk_alloc() so need not be done here.
 */
static int tcp_v4_init_sock(struct sock *sk)
{
1850
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
1851 1852 1853 1854 1855 1856
	struct tcp_sock *tp = tcp_sk(sk);

	skb_queue_head_init(&tp->out_of_order_queue);
	tcp_init_xmit_timers(sk);
	tcp_prequeue_init(tp);

1857
	icsk->icsk_rto = TCP_TIMEOUT_INIT;
L
Linus Torvalds 已提交
1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871
	tp->mdev = TCP_TIMEOUT_INIT;

	/* So many TCP implementations out there (incorrectly) count the
	 * initial SYN frame in their delayed-ACK and congestion control
	 * algorithms that we must have the following bandaid to talk
	 * efficiently to them.  -DaveM
	 */
	tp->snd_cwnd = 2;

	/* See draft-stevens-tcpca-spec-01 for discussion of the
	 * initialization of these values.
	 */
	tp->snd_ssthresh = 0x7fffffff;	/* Infinity */
	tp->snd_cwnd_clamp = ~0;
1872
	tp->mss_cache = 536;
L
Linus Torvalds 已提交
1873 1874

	tp->reordering = sysctl_tcp_reordering;
1875
	icsk->icsk_ca_ops = &tcp_init_congestion_ops;
L
Linus Torvalds 已提交
1876 1877 1878 1879 1880 1881

	sk->sk_state = TCP_CLOSE;

	sk->sk_write_space = sk_stream_write_space;
	sock_set_flag(sk, SOCK_USE_WRITE_QUEUE);

1882
	icsk->icsk_af_ops = &ipv4_specific;
1883
	icsk->icsk_sync_mss = tcp_sync_mss;
1884 1885 1886
#ifdef CONFIG_TCP_MD5SIG
	tp->af_specific = &tcp_sock_ipv4_specific;
#endif
L
Linus Torvalds 已提交
1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901

	sk->sk_sndbuf = sysctl_tcp_wmem[1];
	sk->sk_rcvbuf = sysctl_tcp_rmem[1];

	atomic_inc(&tcp_sockets_allocated);

	return 0;
}

int tcp_v4_destroy_sock(struct sock *sk)
{
	struct tcp_sock *tp = tcp_sk(sk);

	tcp_clear_xmit_timers(sk);

1902
	tcp_cleanup_congestion_control(sk);
1903

L
Linus Torvalds 已提交
1904
	/* Cleanup up the write buffer. */
1905
	tcp_write_queue_purge(sk);
L
Linus Torvalds 已提交
1906 1907

	/* Cleans up our, hopefully empty, out_of_order_queue. */
1908
	__skb_queue_purge(&tp->out_of_order_queue);
L
Linus Torvalds 已提交
1909

1910 1911 1912 1913 1914 1915 1916 1917 1918
#ifdef CONFIG_TCP_MD5SIG
	/* Clean up the MD5 key list, if any */
	if (tp->md5sig_info) {
		tcp_v4_clear_md5_list(sk);
		kfree(tp->md5sig_info);
		tp->md5sig_info = NULL;
	}
#endif

C
Chris Leech 已提交
1919 1920
#ifdef CONFIG_NET_DMA
	/* Cleans up our sk_async_wait_queue */
1921
	__skb_queue_purge(&sk->sk_async_wait_queue);
C
Chris Leech 已提交
1922 1923
#endif

L
Linus Torvalds 已提交
1924 1925 1926 1927
	/* Clean prequeue, it must be empty really */
	__skb_queue_purge(&tp->ucopy.prequeue);

	/* Clean up a referenced TCP bind bucket. */
1928
	if (inet_csk(sk)->icsk_bind_hash)
1929
		inet_put_port(&tcp_hashinfo, sk);
L
Linus Torvalds 已提交
1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948

	/*
	 * If sendmsg cached page exists, toss it.
	 */
	if (sk->sk_sndmsg_page) {
		__free_page(sk->sk_sndmsg_page);
		sk->sk_sndmsg_page = NULL;
	}

	atomic_dec(&tcp_sockets_allocated);

	return 0;
}

EXPORT_SYMBOL(tcp_v4_destroy_sock);

#ifdef CONFIG_PROC_FS
/* Proc filesystem TCP sock list dumping. */

1949
static inline struct inet_timewait_sock *tw_head(struct hlist_head *head)
L
Linus Torvalds 已提交
1950 1951
{
	return hlist_empty(head) ? NULL :
1952
		list_entry(head->first, struct inet_timewait_sock, tw_node);
L
Linus Torvalds 已提交
1953 1954
}

1955
static inline struct inet_timewait_sock *tw_next(struct inet_timewait_sock *tw)
L
Linus Torvalds 已提交
1956 1957 1958 1959 1960 1961 1962
{
	return tw->tw_node.next ?
		hlist_entry(tw->tw_node.next, typeof(*tw), tw_node) : NULL;
}

static void *listening_get_next(struct seq_file *seq, void *cur)
{
1963
	struct inet_connection_sock *icsk;
L
Linus Torvalds 已提交
1964 1965 1966 1967 1968 1969
	struct hlist_node *node;
	struct sock *sk = cur;
	struct tcp_iter_state* st = seq->private;

	if (!sk) {
		st->bucket = 0;
1970
		sk = sk_head(&tcp_hashinfo.listening_hash[0]);
L
Linus Torvalds 已提交
1971 1972 1973 1974 1975 1976
		goto get_sk;
	}

	++st->num;

	if (st->state == TCP_SEQ_STATE_OPENREQ) {
1977
		struct request_sock *req = cur;
L
Linus Torvalds 已提交
1978

1979
		icsk = inet_csk(st->syn_wait_sk);
L
Linus Torvalds 已提交
1980 1981 1982
		req = req->dl_next;
		while (1) {
			while (req) {
1983
				if (req->rsk_ops->family == st->family) {
L
Linus Torvalds 已提交
1984 1985 1986 1987 1988
					cur = req;
					goto out;
				}
				req = req->dl_next;
			}
1989
			if (++st->sbucket >= icsk->icsk_accept_queue.listen_opt->nr_table_entries)
L
Linus Torvalds 已提交
1990 1991
				break;
get_req:
1992
			req = icsk->icsk_accept_queue.listen_opt->syn_table[st->sbucket];
L
Linus Torvalds 已提交
1993 1994 1995
		}
		sk	  = sk_next(st->syn_wait_sk);
		st->state = TCP_SEQ_STATE_LISTENING;
1996
		read_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
L
Linus Torvalds 已提交
1997
	} else {
1998
		icsk = inet_csk(sk);
1999 2000
		read_lock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
		if (reqsk_queue_len(&icsk->icsk_accept_queue))
L
Linus Torvalds 已提交
2001
			goto start_req;
2002
		read_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
L
Linus Torvalds 已提交
2003 2004 2005 2006 2007 2008 2009 2010
		sk = sk_next(sk);
	}
get_sk:
	sk_for_each_from(sk, node) {
		if (sk->sk_family == st->family) {
			cur = sk;
			goto out;
		}
2011
		icsk = inet_csk(sk);
2012 2013
		read_lock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
		if (reqsk_queue_len(&icsk->icsk_accept_queue)) {
L
Linus Torvalds 已提交
2014 2015 2016 2017 2018 2019 2020
start_req:
			st->uid		= sock_i_uid(sk);
			st->syn_wait_sk = sk;
			st->state	= TCP_SEQ_STATE_OPENREQ;
			st->sbucket	= 0;
			goto get_req;
		}
2021
		read_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
L
Linus Torvalds 已提交
2022
	}
2023
	if (++st->bucket < INET_LHTABLE_SIZE) {
2024
		sk = sk_head(&tcp_hashinfo.listening_hash[st->bucket]);
L
Linus Torvalds 已提交
2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047
		goto get_sk;
	}
	cur = NULL;
out:
	return cur;
}

static void *listening_get_idx(struct seq_file *seq, loff_t *pos)
{
	void *rc = listening_get_next(seq, NULL);

	while (rc && *pos) {
		rc = listening_get_next(seq, rc);
		--*pos;
	}
	return rc;
}

static void *established_get_first(struct seq_file *seq)
{
	struct tcp_iter_state* st = seq->private;
	void *rc = NULL;

2048
	for (st->bucket = 0; st->bucket < tcp_hashinfo.ehash_size; ++st->bucket) {
L
Linus Torvalds 已提交
2049 2050
		struct sock *sk;
		struct hlist_node *node;
2051
		struct inet_timewait_sock *tw;
2052
		rwlock_t *lock = inet_ehash_lockp(&tcp_hashinfo, st->bucket);
L
Linus Torvalds 已提交
2053

2054
		read_lock_bh(lock);
2055
		sk_for_each(sk, node, &tcp_hashinfo.ehash[st->bucket].chain) {
L
Linus Torvalds 已提交
2056 2057 2058 2059 2060 2061 2062
			if (sk->sk_family != st->family) {
				continue;
			}
			rc = sk;
			goto out;
		}
		st->state = TCP_SEQ_STATE_TIME_WAIT;
2063
		inet_twsk_for_each(tw, node,
2064
				   &tcp_hashinfo.ehash[st->bucket].twchain) {
L
Linus Torvalds 已提交
2065 2066 2067 2068 2069 2070
			if (tw->tw_family != st->family) {
				continue;
			}
			rc = tw;
			goto out;
		}
2071
		read_unlock_bh(lock);
L
Linus Torvalds 已提交
2072 2073 2074 2075 2076 2077 2078 2079 2080
		st->state = TCP_SEQ_STATE_ESTABLISHED;
	}
out:
	return rc;
}

static void *established_get_next(struct seq_file *seq, void *cur)
{
	struct sock *sk = cur;
2081
	struct inet_timewait_sock *tw;
L
Linus Torvalds 已提交
2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097
	struct hlist_node *node;
	struct tcp_iter_state* st = seq->private;

	++st->num;

	if (st->state == TCP_SEQ_STATE_TIME_WAIT) {
		tw = cur;
		tw = tw_next(tw);
get_tw:
		while (tw && tw->tw_family != st->family) {
			tw = tw_next(tw);
		}
		if (tw) {
			cur = tw;
			goto out;
		}
2098
		read_unlock_bh(inet_ehash_lockp(&tcp_hashinfo, st->bucket));
L
Linus Torvalds 已提交
2099 2100
		st->state = TCP_SEQ_STATE_ESTABLISHED;

2101
		if (++st->bucket < tcp_hashinfo.ehash_size) {
2102
			read_lock_bh(inet_ehash_lockp(&tcp_hashinfo, st->bucket));
2103
			sk = sk_head(&tcp_hashinfo.ehash[st->bucket].chain);
L
Linus Torvalds 已提交
2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116
		} else {
			cur = NULL;
			goto out;
		}
	} else
		sk = sk_next(sk);

	sk_for_each_from(sk, node) {
		if (sk->sk_family == st->family)
			goto found;
	}

	st->state = TCP_SEQ_STATE_TIME_WAIT;
2117
	tw = tw_head(&tcp_hashinfo.ehash[st->bucket].twchain);
L
Linus Torvalds 已提交
2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131
	goto get_tw;
found:
	cur = sk;
out:
	return cur;
}

static void *established_get_idx(struct seq_file *seq, loff_t pos)
{
	void *rc = established_get_first(seq);

	while (rc && pos) {
		rc = established_get_next(seq, rc);
		--pos;
2132
	}
L
Linus Torvalds 已提交
2133 2134 2135 2136 2137 2138 2139 2140
	return rc;
}

static void *tcp_get_idx(struct seq_file *seq, loff_t pos)
{
	void *rc;
	struct tcp_iter_state* st = seq->private;

2141
	inet_listen_lock(&tcp_hashinfo);
L
Linus Torvalds 已提交
2142 2143 2144 2145
	st->state = TCP_SEQ_STATE_LISTENING;
	rc	  = listening_get_idx(seq, &pos);

	if (!rc) {
2146
		inet_listen_unlock(&tcp_hashinfo);
L
Linus Torvalds 已提交
2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177
		st->state = TCP_SEQ_STATE_ESTABLISHED;
		rc	  = established_get_idx(seq, pos);
	}

	return rc;
}

static void *tcp_seq_start(struct seq_file *seq, loff_t *pos)
{
	struct tcp_iter_state* st = seq->private;
	st->state = TCP_SEQ_STATE_LISTENING;
	st->num = 0;
	return *pos ? tcp_get_idx(seq, *pos - 1) : SEQ_START_TOKEN;
}

static void *tcp_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
	void *rc = NULL;
	struct tcp_iter_state* st;

	if (v == SEQ_START_TOKEN) {
		rc = tcp_get_idx(seq, 0);
		goto out;
	}
	st = seq->private;

	switch (st->state) {
	case TCP_SEQ_STATE_OPENREQ:
	case TCP_SEQ_STATE_LISTENING:
		rc = listening_get_next(seq, v);
		if (!rc) {
2178
			inet_listen_unlock(&tcp_hashinfo);
L
Linus Torvalds 已提交
2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199
			st->state = TCP_SEQ_STATE_ESTABLISHED;
			rc	  = established_get_first(seq);
		}
		break;
	case TCP_SEQ_STATE_ESTABLISHED:
	case TCP_SEQ_STATE_TIME_WAIT:
		rc = established_get_next(seq, v);
		break;
	}
out:
	++*pos;
	return rc;
}

static void tcp_seq_stop(struct seq_file *seq, void *v)
{
	struct tcp_iter_state* st = seq->private;

	switch (st->state) {
	case TCP_SEQ_STATE_OPENREQ:
		if (v) {
2200 2201
			struct inet_connection_sock *icsk = inet_csk(st->syn_wait_sk);
			read_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
L
Linus Torvalds 已提交
2202 2203 2204
		}
	case TCP_SEQ_STATE_LISTENING:
		if (v != SEQ_START_TOKEN)
2205
			inet_listen_unlock(&tcp_hashinfo);
L
Linus Torvalds 已提交
2206 2207 2208 2209
		break;
	case TCP_SEQ_STATE_TIME_WAIT:
	case TCP_SEQ_STATE_ESTABLISHED:
		if (v)
2210
			read_unlock_bh(inet_ehash_lockp(&tcp_hashinfo, st->bucket));
L
Linus Torvalds 已提交
2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224
		break;
	}
}

static int tcp_seq_open(struct inode *inode, struct file *file)
{
	struct tcp_seq_afinfo *afinfo = PDE(inode)->data;
	struct seq_file *seq;
	struct tcp_iter_state *s;
	int rc;

	if (unlikely(afinfo == NULL))
		return -EINVAL;

2225
	s = kzalloc(sizeof(*s), GFP_KERNEL);
L
Linus Torvalds 已提交
2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257
	if (!s)
		return -ENOMEM;
	s->family		= afinfo->family;
	s->seq_ops.start	= tcp_seq_start;
	s->seq_ops.next		= tcp_seq_next;
	s->seq_ops.show		= afinfo->seq_show;
	s->seq_ops.stop		= tcp_seq_stop;

	rc = seq_open(file, &s->seq_ops);
	if (rc)
		goto out_kfree;
	seq	     = file->private_data;
	seq->private = s;
out:
	return rc;
out_kfree:
	kfree(s);
	goto out;
}

int tcp_proc_register(struct tcp_seq_afinfo *afinfo)
{
	int rc = 0;
	struct proc_dir_entry *p;

	if (!afinfo)
		return -EINVAL;
	afinfo->seq_fops->owner		= afinfo->owner;
	afinfo->seq_fops->open		= tcp_seq_open;
	afinfo->seq_fops->read		= seq_read;
	afinfo->seq_fops->llseek	= seq_lseek;
	afinfo->seq_fops->release	= seq_release_private;
2258

2259
	p = proc_net_fops_create(&init_net, afinfo->name, S_IRUGO, afinfo->seq_fops);
L
Linus Torvalds 已提交
2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270
	if (p)
		p->data = afinfo;
	else
		rc = -ENOMEM;
	return rc;
}

void tcp_proc_unregister(struct tcp_seq_afinfo *afinfo)
{
	if (!afinfo)
		return;
2271
	proc_net_remove(&init_net, afinfo->name);
2272
	memset(afinfo->seq_fops, 0, sizeof(*afinfo->seq_fops));
L
Linus Torvalds 已提交
2273 2274
}

2275
static void get_openreq4(struct sock *sk, struct request_sock *req,
L
Linus Torvalds 已提交
2276 2277
			 char *tmpbuf, int i, int uid)
{
2278
	const struct inet_request_sock *ireq = inet_rsk(req);
L
Linus Torvalds 已提交
2279 2280 2281 2282 2283
	int ttd = req->expires - jiffies;

	sprintf(tmpbuf, "%4d: %08X:%04X %08X:%04X"
		" %02X %08X:%08X %02X:%08lX %08X %5d %8d %u %d %p",
		i,
2284
		ireq->loc_addr,
L
Linus Torvalds 已提交
2285
		ntohs(inet_sk(sk)->sport),
2286 2287
		ireq->rmt_addr,
		ntohs(ireq->rmt_port),
L
Linus Torvalds 已提交
2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299
		TCP_SYN_RECV,
		0, 0, /* could print option size, but that is af dependent. */
		1,    /* timers active (only the expire timer) */
		jiffies_to_clock_t(ttd),
		req->retrans,
		uid,
		0,  /* non standard timer */
		0, /* open_requests have no inode */
		atomic_read(&sk->sk_refcnt),
		req);
}

2300
static void get_tcp4_sock(struct sock *sk, char *tmpbuf, int i)
L
Linus Torvalds 已提交
2301 2302 2303
{
	int timer_active;
	unsigned long timer_expires;
2304 2305 2306
	struct tcp_sock *tp = tcp_sk(sk);
	const struct inet_connection_sock *icsk = inet_csk(sk);
	struct inet_sock *inet = inet_sk(sk);
2307 2308
	__be32 dest = inet->daddr;
	__be32 src = inet->rcv_saddr;
L
Linus Torvalds 已提交
2309 2310 2311
	__u16 destp = ntohs(inet->dport);
	__u16 srcp = ntohs(inet->sport);

2312
	if (icsk->icsk_pending == ICSK_TIME_RETRANS) {
L
Linus Torvalds 已提交
2313
		timer_active	= 1;
2314 2315
		timer_expires	= icsk->icsk_timeout;
	} else if (icsk->icsk_pending == ICSK_TIME_PROBE0) {
L
Linus Torvalds 已提交
2316
		timer_active	= 4;
2317
		timer_expires	= icsk->icsk_timeout;
2318
	} else if (timer_pending(&sk->sk_timer)) {
L
Linus Torvalds 已提交
2319
		timer_active	= 2;
2320
		timer_expires	= sk->sk_timer.expires;
L
Linus Torvalds 已提交
2321 2322 2323 2324 2325 2326 2327
	} else {
		timer_active	= 0;
		timer_expires = jiffies;
	}

	sprintf(tmpbuf, "%4d: %08X:%04X %08X:%04X %02X %08X:%08X %02X:%08lX "
			"%08X %5d %8d %lu %d %p %u %u %u %u %d",
2328
		i, src, srcp, dest, destp, sk->sk_state,
2329
		tp->write_seq - tp->snd_una,
2330
		sk->sk_state == TCP_LISTEN ? sk->sk_ack_backlog :
2331
					     (tp->rcv_nxt - tp->copied_seq),
L
Linus Torvalds 已提交
2332 2333
		timer_active,
		jiffies_to_clock_t(timer_expires - jiffies),
2334
		icsk->icsk_retransmits,
2335
		sock_i_uid(sk),
2336
		icsk->icsk_probes_out,
2337 2338
		sock_i_ino(sk),
		atomic_read(&sk->sk_refcnt), sk,
2339 2340 2341
		icsk->icsk_rto,
		icsk->icsk_ack.ato,
		(icsk->icsk_ack.quick << 1) | icsk->icsk_ack.pingpong,
L
Linus Torvalds 已提交
2342 2343 2344 2345
		tp->snd_cwnd,
		tp->snd_ssthresh >= 0xFFFF ? -1 : tp->snd_ssthresh);
}

2346 2347
static void get_timewait4_sock(struct inet_timewait_sock *tw,
			       char *tmpbuf, int i)
L
Linus Torvalds 已提交
2348
{
2349
	__be32 dest, src;
L
Linus Torvalds 已提交
2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420
	__u16 destp, srcp;
	int ttd = tw->tw_ttd - jiffies;

	if (ttd < 0)
		ttd = 0;

	dest  = tw->tw_daddr;
	src   = tw->tw_rcv_saddr;
	destp = ntohs(tw->tw_dport);
	srcp  = ntohs(tw->tw_sport);

	sprintf(tmpbuf, "%4d: %08X:%04X %08X:%04X"
		" %02X %08X:%08X %02X:%08lX %08X %5d %8d %d %d %p",
		i, src, srcp, dest, destp, tw->tw_substate, 0, 0,
		3, jiffies_to_clock_t(ttd), 0, 0, 0, 0,
		atomic_read(&tw->tw_refcnt), tw);
}

#define TMPSZ 150

static int tcp4_seq_show(struct seq_file *seq, void *v)
{
	struct tcp_iter_state* st;
	char tmpbuf[TMPSZ + 1];

	if (v == SEQ_START_TOKEN) {
		seq_printf(seq, "%-*s\n", TMPSZ - 1,
			   "  sl  local_address rem_address   st tx_queue "
			   "rx_queue tr tm->when retrnsmt   uid  timeout "
			   "inode");
		goto out;
	}
	st = seq->private;

	switch (st->state) {
	case TCP_SEQ_STATE_LISTENING:
	case TCP_SEQ_STATE_ESTABLISHED:
		get_tcp4_sock(v, tmpbuf, st->num);
		break;
	case TCP_SEQ_STATE_OPENREQ:
		get_openreq4(st->syn_wait_sk, v, tmpbuf, st->num, st->uid);
		break;
	case TCP_SEQ_STATE_TIME_WAIT:
		get_timewait4_sock(v, tmpbuf, st->num);
		break;
	}
	seq_printf(seq, "%-*s\n", TMPSZ - 1, tmpbuf);
out:
	return 0;
}

static struct file_operations tcp4_seq_fops;
static struct tcp_seq_afinfo tcp4_seq_afinfo = {
	.owner		= THIS_MODULE,
	.name		= "tcp",
	.family		= AF_INET,
	.seq_show	= tcp4_seq_show,
	.seq_fops	= &tcp4_seq_fops,
};

int __init tcp4_proc_init(void)
{
	return tcp_proc_register(&tcp4_seq_afinfo);
}

void tcp4_proc_exit(void)
{
	tcp_proc_unregister(&tcp4_seq_afinfo);
}
#endif /* CONFIG_PROC_FS */

2421 2422
DEFINE_PROTO_INUSE(tcp)

L
Linus Torvalds 已提交
2423 2424 2425 2426 2427 2428
struct proto tcp_prot = {
	.name			= "TCP",
	.owner			= THIS_MODULE,
	.close			= tcp_close,
	.connect		= tcp_v4_connect,
	.disconnect		= tcp_disconnect,
2429
	.accept			= inet_csk_accept,
L
Linus Torvalds 已提交
2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442
	.ioctl			= tcp_ioctl,
	.init			= tcp_v4_init_sock,
	.destroy		= tcp_v4_destroy_sock,
	.shutdown		= tcp_shutdown,
	.setsockopt		= tcp_setsockopt,
	.getsockopt		= tcp_getsockopt,
	.recvmsg		= tcp_recvmsg,
	.backlog_rcv		= tcp_v4_do_rcv,
	.hash			= tcp_v4_hash,
	.unhash			= tcp_unhash,
	.get_port		= tcp_v4_get_port,
	.enter_memory_pressure	= tcp_enter_memory_pressure,
	.sockets_allocated	= &tcp_sockets_allocated,
2443
	.orphan_count		= &tcp_orphan_count,
L
Linus Torvalds 已提交
2444 2445 2446 2447 2448 2449 2450
	.memory_allocated	= &tcp_memory_allocated,
	.memory_pressure	= &tcp_memory_pressure,
	.sysctl_mem		= sysctl_tcp_mem,
	.sysctl_wmem		= sysctl_tcp_wmem,
	.sysctl_rmem		= sysctl_tcp_rmem,
	.max_header		= MAX_TCP_HEADER,
	.obj_size		= sizeof(struct tcp_sock),
2451
	.twsk_prot		= &tcp_timewait_sock_ops,
2452
	.rsk_prot		= &tcp_request_sock_ops,
2453 2454 2455 2456
#ifdef CONFIG_COMPAT
	.compat_setsockopt	= compat_tcp_setsockopt,
	.compat_getsockopt	= compat_tcp_getsockopt,
#endif
2457
	REF_PROTO_INUSE(tcp)
L
Linus Torvalds 已提交
2458 2459 2460 2461
};

void __init tcp_v4_init(struct net_proto_family *ops)
{
2462 2463
	if (inet_csk_ctl_sock_create(&tcp_socket, PF_INET, SOCK_RAW,
				     IPPROTO_TCP) < 0)
L
Linus Torvalds 已提交
2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483
		panic("Failed to create the TCP control socket.\n");
}

EXPORT_SYMBOL(ipv4_specific);
EXPORT_SYMBOL(tcp_hashinfo);
EXPORT_SYMBOL(tcp_prot);
EXPORT_SYMBOL(tcp_unhash);
EXPORT_SYMBOL(tcp_v4_conn_request);
EXPORT_SYMBOL(tcp_v4_connect);
EXPORT_SYMBOL(tcp_v4_do_rcv);
EXPORT_SYMBOL(tcp_v4_remember_stamp);
EXPORT_SYMBOL(tcp_v4_send_check);
EXPORT_SYMBOL(tcp_v4_syn_recv_sock);

#ifdef CONFIG_PROC_FS
EXPORT_SYMBOL(tcp_proc_register);
EXPORT_SYMBOL(tcp_proc_unregister);
#endif
EXPORT_SYMBOL(sysctl_tcp_low_latency);