tcp_ipv4.c 59.2 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
/*
 * 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).
 *
 *		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
37
 *					request_sock handling and moved
L
Linus Torvalds 已提交
38 39
 *					most of it into the af independent code.
 *					Added tail drop and some other bugfixes.
S
Stephen Hemminger 已提交
40
 *					Added new listen semantics.
L
Linus Torvalds 已提交
41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62
 *		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>

63
#include <net/net_namespace.h>
L
Linus Torvalds 已提交
64
#include <net/icmp.h>
65
#include <net/inet_hashtables.h>
L
Linus Torvalds 已提交
66
#include <net/tcp.h>
67
#include <net/transp_v6.h>
L
Linus Torvalds 已提交
68 69
#include <net/ipv6.h>
#include <net/inet_common.h>
70
#include <net/timewait_sock.h>
L
Linus Torvalds 已提交
71
#include <net/xfrm.h>
C
Chris Leech 已提交
72
#include <net/netdma.h>
L
Linus Torvalds 已提交
73 74 75 76 77 78 79

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

80 81 82
#include <linux/crypto.h>
#include <linux/scatterlist.h>

83 84
int sysctl_tcp_tw_reuse __read_mostly;
int sysctl_tcp_low_latency __read_mostly;
L
Linus Torvalds 已提交
85 86


87
#ifdef CONFIG_TCP_MD5SIG
88 89
static struct tcp_md5sig_key *tcp_v4_md5_do_lookup(struct sock *sk,
						   __be32 addr);
90
static int tcp_v4_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key,
91
				   __be32 saddr, __be32 daddr,
92
				   struct tcphdr *th, unsigned int tcplen);
93 94 95 96 97 98
#else
static inline
struct tcp_md5sig_key *tcp_v4_md5_do_lookup(struct sock *sk, __be32 addr)
{
	return NULL;
}
99 100
#endif

101
struct inet_hashinfo __cacheline_aligned tcp_hashinfo = {
102 103 104
	.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
105 106
};

107
static inline __u32 tcp_v4_init_sequence(struct sk_buff *skb)
L
Linus Torvalds 已提交
108
{
109 110
	return secure_tcp_sequence_number(ip_hdr(skb)->daddr,
					  ip_hdr(skb)->saddr,
111 112
					  tcp_hdr(skb)->dest,
					  tcp_hdr(skb)->source);
L
Linus Torvalds 已提交
113 114
}

115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
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 &&
133
			     get_seconds() - tcptw->tw_ts_recent_stamp > 1))) {
134 135 136 137 138 139 140 141 142 143 144 145 146 147
		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
148 149 150 151 152 153 154
/* 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;
155
	__be32 daddr, nexthop;
L
Linus Torvalds 已提交
156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174
	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,
175
			       inet->sport, usin->sin_port, sk, 1);
176 177
	if (tmp < 0) {
		if (tmp == -ENETUNREACH)
178
			IP_INC_STATS_BH(sock_net(sk), IPSTATS_MIB_OUTNOROUTES);
L
Linus Torvalds 已提交
179
		return tmp;
180
	}
L
Linus Torvalds 已提交
181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200

	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;
	}

201
	if (tcp_death_row.sysctl_tw_recycle &&
L
Linus Torvalds 已提交
202 203
	    !tp->rx_opt.ts_recent_stamp && rt->rt_dst == daddr) {
		struct inet_peer *peer = rt_get_peer(rt);
204 205 206 207 208
		/*
		 * 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
209
		 */
210
		if (peer != NULL &&
211
		    peer->tcp_ts_stamp + TCP_PAWS_MSL >= get_seconds()) {
L
Linus Torvalds 已提交
212 213 214 215 216 217 218 219
			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;

220
	inet_csk(sk)->icsk_ext_hdr_len = 0;
L
Linus Torvalds 已提交
221
	if (inet->opt)
222
		inet_csk(sk)->icsk_ext_hdr_len = inet->opt->optlen;
L
Linus Torvalds 已提交
223 224 225 226 227 228 229 230 231

	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);
232
	err = inet_hash_connect(&tcp_death_row, sk);
L
Linus Torvalds 已提交
233 234 235
	if (err)
		goto failure;

236 237
	err = ip_route_newports(&rt, IPPROTO_TCP,
				inet->sport, inet->dport, sk);
L
Linus Torvalds 已提交
238 239 240 241
	if (err)
		goto failure;

	/* OK, now commit destination to socket.  */
242
	sk->sk_gso_type = SKB_GSO_TCPV4;
243
	sk_setup_caps(sk, &rt->u.dst);
L
Linus Torvalds 已提交
244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260

	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:
261 262 263 264
	/*
	 * This unhashes the socket and releases the local port,
	 * if necessary.
	 */
L
Linus Torvalds 已提交
265 266 267 268 269 270 271 272 273 274
	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
275
static void do_pmtu_discovery(struct sock *sk, struct iphdr *iph, u32 mtu)
L
Linus Torvalds 已提交
276 277 278 279 280 281 282 283 284 285 286 287 288 289
{
	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.
290
	 * There is a small race when the user changes this flag in the
L
Linus Torvalds 已提交
291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306
	 * 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 &&
307
	    inet_csk(sk)->icsk_pmtu_cookie > mtu) {
L
Linus Torvalds 已提交
308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340
		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;
341 342
	const int type = icmp_hdr(skb)->type;
	const int code = icmp_hdr(skb)->code;
L
Linus Torvalds 已提交
343 344 345
	struct sock *sk;
	__u32 seq;
	int err;
346
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
347 348

	if (skb->len < (iph->ihl << 2) + 8) {
349
		ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS);
L
Linus Torvalds 已提交
350 351 352
		return;
	}

353
	sk = inet_lookup(net, &tcp_hashinfo, iph->daddr, th->dest,
354
			iph->saddr, th->source, inet_iif(skb));
L
Linus Torvalds 已提交
355
	if (!sk) {
356
		ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS);
L
Linus Torvalds 已提交
357 358 359
		return;
	}
	if (sk->sk_state == TCP_TIME_WAIT) {
360
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377
		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)) {
378
		NET_INC_STATS_BH(LINUX_MIB_OUTOFWINDOWICMPS);
L
Linus Torvalds 已提交
379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408
		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) {
409
		struct request_sock *req, **prev;
L
Linus Torvalds 已提交
410 411 412 413
	case TCP_LISTEN:
		if (sock_owned_by_user(sk))
			goto out;

414 415
		req = inet_csk_search_req(sk, &prev, th->dest,
					  iph->daddr, iph->saddr);
L
Linus Torvalds 已提交
416 417 418 419 420 421 422 423
		if (!req)
			goto out;

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

424
		if (seq != tcp_rsk(req)->snt_isn) {
L
Linus Torvalds 已提交
425 426 427 428 429 430 431 432 433 434
			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().
		 */
435
		inet_csk_reqsk_queue_drop(sk, req, prev);
L
Linus Torvalds 已提交
436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483
		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. */
484
void tcp_v4_send_check(struct sock *sk, int len, struct sk_buff *skb)
L
Linus Torvalds 已提交
485 486
{
	struct inet_sock *inet = inet_sk(sk);
487
	struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
488

489
	if (skb->ip_summed == CHECKSUM_PARTIAL) {
490 491
		th->check = ~tcp_v4_check(len, inet->saddr,
					  inet->daddr, 0);
492
		skb->csum_start = skb_transport_header(skb) - skb->head;
A
Al Viro 已提交
493
		skb->csum_offset = offsetof(struct tcphdr, check);
L
Linus Torvalds 已提交
494
	} else {
495
		th->check = tcp_v4_check(len, inet->saddr, inet->daddr,
L
Linus Torvalds 已提交
496 497 498 499 500 501
					 csum_partial((char *)th,
						      th->doff << 2,
						      skb->csum));
	}
}

502 503
int tcp_v4_gso_send_check(struct sk_buff *skb)
{
504
	const struct iphdr *iph;
505 506 507 508 509
	struct tcphdr *th;

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

510
	iph = ip_hdr(skb);
511
	th = tcp_hdr(skb);
512 513

	th->check = 0;
514
	th->check = ~tcp_v4_check(skb->len, iph->saddr, iph->daddr, 0);
515
	skb->csum_start = skb_transport_header(skb) - skb->head;
A
Al Viro 已提交
516
	skb->csum_offset = offsetof(struct tcphdr, check);
517
	skb->ip_summed = CHECKSUM_PARTIAL;
518 519 520
	return 0;
}

L
Linus Torvalds 已提交
521 522 523 524 525 526 527 528 529 530 531 532 533
/*
 *	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.
 */

534
static void tcp_v4_send_reset(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
535
{
536
	struct tcphdr *th = tcp_hdr(skb);
537 538 539
	struct {
		struct tcphdr th;
#ifdef CONFIG_TCP_MD5SIG
540
		__be32 opt[(TCPOLEN_MD5SIG_ALIGNED >> 2)];
541 542
#endif
	} rep;
L
Linus Torvalds 已提交
543
	struct ip_reply_arg arg;
544 545 546
#ifdef CONFIG_TCP_MD5SIG
	struct tcp_md5sig_key *key;
#endif
547
	struct net *net;
L
Linus Torvalds 已提交
548 549 550 551 552

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

553
	if (skb->rtable->rt_type != RTN_LOCAL)
L
Linus Torvalds 已提交
554 555 556
		return;

	/* Swap the send and the receive. */
557 558 559 560 561
	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
562 563

	if (th->ack) {
564
		rep.th.seq = th->ack_seq;
L
Linus Torvalds 已提交
565
	} else {
566 567 568
		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
569 570
	}

571
	memset(&arg, 0, sizeof(arg));
572 573 574 575
	arg.iov[0].iov_base = (unsigned char *)&rep;
	arg.iov[0].iov_len  = sizeof(rep.th);

#ifdef CONFIG_TCP_MD5SIG
576
	key = sk ? tcp_v4_md5_do_lookup(sk, ip_hdr(skb)->daddr) : NULL;
577 578 579 580 581 582 583 584 585 586 587
	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,
588 589
					ip_hdr(skb)->daddr,
					ip_hdr(skb)->saddr,
590
					&rep.th, arg.iov[0].iov_len);
591 592
	}
#endif
593 594
	arg.csum = csum_tcpudp_nofold(ip_hdr(skb)->daddr,
				      ip_hdr(skb)->saddr, /* XXX */
L
Linus Torvalds 已提交
595 596 597
				      sizeof(struct tcphdr), IPPROTO_TCP, 0);
	arg.csumoffset = offsetof(struct tcphdr, check) / 2;

598 599
	net = dev_net(skb->dst->dev);
	ip_send_reply(net->ipv4.tcp_sock, skb,
600
		      &arg, arg.iov[0].iov_len);
L
Linus Torvalds 已提交
601

602 603
	TCP_INC_STATS_BH(net, TCP_MIB_OUTSEGS);
	TCP_INC_STATS_BH(net, TCP_MIB_OUTRSTS);
L
Linus Torvalds 已提交
604 605 606 607 608 609
}

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

610 611 612
static void tcp_v4_send_ack(struct sk_buff *skb, u32 seq, u32 ack,
			    u32 win, u32 ts, int oif,
			    struct tcp_md5sig_key *key)
L
Linus Torvalds 已提交
613
{
614
	struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
615 616
	struct {
		struct tcphdr th;
617
		__be32 opt[(TCPOLEN_TSTAMP_ALIGNED >> 2)
618
#ifdef CONFIG_TCP_MD5SIG
619
			   + (TCPOLEN_MD5SIG_ALIGNED >> 2)
620 621
#endif
			];
L
Linus Torvalds 已提交
622 623
	} rep;
	struct ip_reply_arg arg;
624
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
625 626

	memset(&rep.th, 0, sizeof(struct tcphdr));
627
	memset(&arg, 0, sizeof(arg));
L
Linus Torvalds 已提交
628 629 630 631

	arg.iov[0].iov_base = (unsigned char *)&rep;
	arg.iov[0].iov_len  = sizeof(rep.th);
	if (ts) {
632 633 634 635 636
		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);
637
		arg.iov[0].iov_len += TCPOLEN_TSTAMP_ALIGNED;
L
Linus Torvalds 已提交
638 639 640 641 642 643 644 645 646 647 648
	}

	/* 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);

649 650 651 652 653 654 655 656 657 658 659 660 661
#ifdef CONFIG_TCP_MD5SIG
	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,
662 663
					ip_hdr(skb)->daddr,
					ip_hdr(skb)->saddr,
664
					&rep.th, arg.iov[0].iov_len);
665 666
	}
#endif
667 668
	arg.csum = csum_tcpudp_nofold(ip_hdr(skb)->daddr,
				      ip_hdr(skb)->saddr, /* XXX */
L
Linus Torvalds 已提交
669 670
				      arg.iov[0].iov_len, IPPROTO_TCP, 0);
	arg.csumoffset = offsetof(struct tcphdr, check) / 2;
671 672
	if (oif)
		arg.bound_dev_if = oif;
L
Linus Torvalds 已提交
673

674
	ip_send_reply(net->ipv4.tcp_sock, skb,
675
		      &arg, arg.iov[0].iov_len);
L
Linus Torvalds 已提交
676

677
	TCP_INC_STATS_BH(net, TCP_MIB_OUTSEGS);
L
Linus Torvalds 已提交
678 679 680 681
}

static void tcp_v4_timewait_ack(struct sock *sk, struct sk_buff *skb)
{
682
	struct inet_timewait_sock *tw = inet_twsk(sk);
683
	struct tcp_timewait_sock *tcptw = tcp_twsk(sk);
L
Linus Torvalds 已提交
684

685
	tcp_v4_send_ack(skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt,
686
			tcptw->tw_rcv_wnd >> tw->tw_rcv_wscale,
687 688 689 690
			tcptw->tw_ts_recent,
			tw->tw_bound_dev_if,
			tcp_twsk_md5_key(tcptw)
			);
L
Linus Torvalds 已提交
691

692
	inet_twsk_put(tw);
L
Linus Torvalds 已提交
693 694
}

695 696
static void tcp_v4_reqsk_send_ack(struct sk_buff *skb,
				  struct request_sock *req)
L
Linus Torvalds 已提交
697
{
698
	tcp_v4_send_ack(skb, tcp_rsk(req)->snt_isn + 1,
699
			tcp_rsk(req)->rcv_isn + 1, req->rcv_wnd,
700 701 702
			req->ts_recent,
			0,
			tcp_v4_md5_do_lookup(skb->sk, ip_hdr(skb)->daddr));
L
Linus Torvalds 已提交
703 704 705
}

/*
706
 *	Send a SYN-ACK after having received a SYN.
707
 *	This still operates on a request_sock only, not on a big
L
Linus Torvalds 已提交
708 709
 *	socket.
 */
710 711
static int __tcp_v4_send_synack(struct sock *sk, struct request_sock *req,
				struct dst_entry *dst)
L
Linus Torvalds 已提交
712
{
713
	const struct inet_request_sock *ireq = inet_rsk(req);
L
Linus Torvalds 已提交
714 715 716 717
	int err = -1;
	struct sk_buff * skb;

	/* First, grab a route. */
718
	if (!dst && (dst = inet_csk_route_req(sk, req)) == NULL)
719
		return -1;
L
Linus Torvalds 已提交
720 721 722 723

	skb = tcp_make_synack(sk, dst, req);

	if (skb) {
724
		struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
725

726
		th->check = tcp_v4_check(skb->len,
727 728
					 ireq->loc_addr,
					 ireq->rmt_addr,
L
Linus Torvalds 已提交
729 730 731
					 csum_partial((char *)th, skb->len,
						      skb->csum));

732 733 734
		err = ip_build_and_send_pkt(skb, sk, ireq->loc_addr,
					    ireq->rmt_addr,
					    ireq->opt);
735
		err = net_xmit_eval(err);
L
Linus Torvalds 已提交
736 737 738 739 740 741
	}

	dst_release(dst);
	return err;
}

742 743 744 745 746
static int tcp_v4_send_synack(struct sock *sk, struct request_sock *req)
{
	return __tcp_v4_send_synack(sk, req, NULL);
}

L
Linus Torvalds 已提交
747
/*
748
 *	IPv4 request_sock destructor.
L
Linus Torvalds 已提交
749
 */
750
static void tcp_v4_reqsk_destructor(struct request_sock *req)
L
Linus Torvalds 已提交
751
{
J
Jesper Juhl 已提交
752
	kfree(inet_rsk(req)->opt);
L
Linus Torvalds 已提交
753 754
}

755
#ifdef CONFIG_SYN_COOKIES
S
Stephen Hemminger 已提交
756
static void syn_flood_warning(struct sk_buff *skb)
L
Linus Torvalds 已提交
757 758 759 760 761 762 763
{
	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",
764
		       ntohs(tcp_hdr(skb)->dest));
L
Linus Torvalds 已提交
765 766
	}
}
767
#endif
L
Linus Torvalds 已提交
768 769

/*
770
 * Save and compile IPv4 options into the request_sock if needed.
L
Linus Torvalds 已提交
771
 */
S
Stephen Hemminger 已提交
772 773
static struct ip_options *tcp_v4_save_options(struct sock *sk,
					      struct sk_buff *skb)
L
Linus Torvalds 已提交
774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790
{
	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;
}

791 792 793 794 795 796 797 798
#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.  */
799 800
static struct tcp_md5sig_key *
			tcp_v4_md5_do_lookup(struct sock *sk, __be32 addr)
801 802 803 804 805 806 807 808
{
	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)
809
			return &tp->md5sig_info->keys4[i].base;
810 811 812 813 814 815 816 817 818 819 820 821
	}
	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
822 823
static struct tcp_md5sig_key *tcp_v4_reqsk_md5_lookup(struct sock *sk,
						      struct request_sock *req)
824 825 826 827 828 829 830 831 832
{
	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 */
833
	struct tcp_md5sig_key *key;
834 835 836
	struct tcp_sock *tp = tcp_sk(sk);
	struct tcp4_md5sig_key *keys;

837
	key = tcp_v4_md5_do_lookup(sk, addr);
838 839
	if (key) {
		/* Pre-existing entry - just update that one. */
840 841 842
		kfree(key->key);
		key->key = newkey;
		key->keylen = newkeylen;
843
	} else {
844 845
		struct tcp_md5sig_info *md5sig;

846
		if (!tp->md5sig_info) {
847 848
			tp->md5sig_info = kzalloc(sizeof(*tp->md5sig_info),
						  GFP_ATOMIC);
849 850 851 852
			if (!tp->md5sig_info) {
				kfree(newkey);
				return -ENOMEM;
			}
853
			sk->sk_route_caps &= ~NETIF_F_GSO_MASK;
854 855 856 857 858
		}
		if (tcp_alloc_md5sig_pool() == NULL) {
			kfree(newkey);
			return -ENOMEM;
		}
859 860 861 862
		md5sig = tp->md5sig_info;

		if (md5sig->alloced4 == md5sig->entries4) {
			keys = kmalloc((sizeof(*keys) *
863
					(md5sig->entries4 + 1)), GFP_ATOMIC);
864 865 866 867 868 869
			if (!keys) {
				kfree(newkey);
				tcp_free_md5sig_pool();
				return -ENOMEM;
			}

870 871 872
			if (md5sig->entries4)
				memcpy(keys, md5sig->keys4,
				       sizeof(*keys) * md5sig->entries4);
873 874

			/* Free old key list, and reference new one */
875
			kfree(md5sig->keys4);
876 877
			md5sig->keys4 = keys;
			md5sig->alloced4++;
878
		}
879
		md5sig->entries4++;
880 881 882
		md5sig->keys4[md5sig->entries4 - 1].addr        = addr;
		md5sig->keys4[md5sig->entries4 - 1].base.key    = newkey;
		md5sig->keys4[md5sig->entries4 - 1].base.keylen = newkeylen;
883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903
	}
	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 */
904
			kfree(tp->md5sig_info->keys4[i].base.key);
905 906 907 908 909
			tp->md5sig_info->entries4--;

			if (tp->md5sig_info->entries4 == 0) {
				kfree(tp->md5sig_info->keys4);
				tp->md5sig_info->keys4 = NULL;
910
				tp->md5sig_info->alloced4 = 0;
911
			} else if (tp->md5sig_info->entries4 != i) {
912
				/* Need to do some manipulation */
913 914 915 916
				memmove(&tp->md5sig_info->keys4[i],
					&tp->md5sig_info->keys4[i+1],
					(tp->md5sig_info->entries4 - i) *
					 sizeof(struct tcp4_md5sig_key));
917 918 919 920 921 922 923 924 925 926
			}
			tcp_free_md5sig_pool();
			return 0;
		}
	}
	return -ENOENT;
}

EXPORT_SYMBOL(tcp_v4_md5_do_del);

927
static void tcp_v4_clear_md5_list(struct sock *sk)
928 929 930 931 932 933 934 935 936 937
{
	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++)
938
			kfree(tp->md5sig_info->keys4[i].base.key);
939 940 941 942 943 944 945 946 947 948
		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;
	}
}

949 950
static int tcp_v4_parse_md5_keys(struct sock *sk, char __user *optval,
				 int optlen)
951 952 953 954 955 956 957 958
{
	struct tcp_md5sig cmd;
	struct sockaddr_in *sin = (struct sockaddr_in *)&cmd.tcpm_addr;
	u8 *newkey;

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

959
	if (copy_from_user(&cmd, optval, sizeof(cmd)))
960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975
		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);
976
		struct tcp_md5sig_info *p = kzalloc(sizeof(*p), GFP_KERNEL);
977 978 979 980 981

		if (!p)
			return -EINVAL;

		tp->md5sig_info = p;
982
		sk->sk_route_caps &= ~NETIF_F_GSO_MASK;
983 984
	}

985
	newkey = kmemdup(cmd.tcpm_key, cmd.tcpm_keylen, GFP_KERNEL);
986 987 988 989 990 991 992 993
	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,
994
				   struct tcphdr *th,
995
				   unsigned int tcplen)
996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012
{
	struct tcp_md5sig_pool *hp;
	struct tcp4_pseudohdr *bp;
	int err;

	/*
	 * 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;

	/*
1013
	 * The TCP pseudo-header (in the order: source IP address,
1014 1015 1016 1017 1018 1019
	 * destination IP address, zero-padded protocol number, and
	 * segment length)
	 */
	bp->saddr = saddr;
	bp->daddr = daddr;
	bp->pad = 0;
1020
	bp->protocol = IPPROTO_TCP;
1021
	bp->len = htons(tcplen);
1022

1023 1024
	err = tcp_calc_md5_hash(md5_hash, key, sizeof(*bp),
				th, tcplen, hp);
1025 1026 1027
	if (err)
		goto clear_hash;

1028
	/* Free up the crypto pool */
1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042
	tcp_put_md5sig_pool();
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,
1043
			 struct tcphdr *th,
1044
			 unsigned int tcplen)
1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058
{
	__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,
1059
				       th, tcplen);
1060 1061 1062 1063
}

EXPORT_SYMBOL(tcp_v4_calc_md5_hash);

1064
static int tcp_v4_inbound_md5_hash(struct sock *sk, struct sk_buff *skb)
1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075
{
	/*
	 * 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;
1076
	const struct iphdr *iph = ip_hdr(skb);
1077
	struct tcphdr *th = tcp_hdr(skb);
1078 1079 1080 1081
	int genhash;
	unsigned char newhash[16];

	hash_expected = tcp_v4_md5_do_lookup(sk, iph->saddr);
1082
	hash_location = tcp_parse_md5sig_option(th);
1083 1084 1085 1086 1087 1088

	/* We've parsed the options - do we have a hash? */
	if (!hash_expected && !hash_location)
		return 0;

	if (hash_expected && !hash_location) {
1089
		LIMIT_NETDEBUG(KERN_INFO "MD5 Hash expected but NOT found "
1090
			       "(" NIPQUAD_FMT ", %d)->(" NIPQUAD_FMT ", %d)\n",
1091 1092
			       NIPQUAD(iph->saddr), ntohs(th->source),
			       NIPQUAD(iph->daddr), ntohs(th->dest));
1093 1094 1095 1096
		return 1;
	}

	if (!hash_expected && hash_location) {
1097
		LIMIT_NETDEBUG(KERN_INFO "MD5 Hash NOT expected but found "
1098
			       "(" NIPQUAD_FMT ", %d)->(" NIPQUAD_FMT ", %d)\n",
1099 1100
			       NIPQUAD(iph->saddr), ntohs(th->source),
			       NIPQUAD(iph->daddr), ntohs(th->dest));
1101 1102 1103 1104 1105 1106 1107 1108 1109
		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,
1110
					  th, skb->len);
1111 1112 1113 1114 1115

	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",
1116 1117
			       NIPQUAD(iph->saddr), ntohs(th->source),
			       NIPQUAD(iph->daddr), ntohs(th->dest),
1118 1119 1120 1121 1122 1123 1124 1125 1126
			       genhash ? " tcp_v4_calc_md5_hash failed" : "");
		}
		return 1;
	}
	return 0;
}

#endif

1127
struct request_sock_ops tcp_request_sock_ops __read_mostly = {
L
Linus Torvalds 已提交
1128
	.family		=	PF_INET,
1129
	.obj_size	=	sizeof(struct tcp_request_sock),
L
Linus Torvalds 已提交
1130
	.rtx_syn_ack	=	tcp_v4_send_synack,
1131 1132
	.send_ack	=	tcp_v4_reqsk_send_ack,
	.destructor	=	tcp_v4_reqsk_destructor,
L
Linus Torvalds 已提交
1133 1134 1135
	.send_reset	=	tcp_v4_send_reset,
};

1136
#ifdef CONFIG_TCP_MD5SIG
1137
static struct tcp_request_sock_ops tcp_request_sock_ipv4_ops = {
1138 1139
	.md5_lookup	=	tcp_v4_reqsk_md5_lookup,
};
1140
#endif
1141

1142 1143 1144
static struct timewait_sock_ops tcp_timewait_sock_ops = {
	.twsk_obj_size	= sizeof(struct tcp_timewait_sock),
	.twsk_unique	= tcp_twsk_unique,
1145
	.twsk_destructor= tcp_twsk_destructor,
1146 1147
};

L
Linus Torvalds 已提交
1148 1149
int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
{
1150
	struct inet_request_sock *ireq;
L
Linus Torvalds 已提交
1151
	struct tcp_options_received tmp_opt;
1152
	struct request_sock *req;
1153 1154
	__be32 saddr = ip_hdr(skb)->saddr;
	__be32 daddr = ip_hdr(skb)->daddr;
L
Linus Torvalds 已提交
1155 1156 1157 1158 1159 1160 1161 1162 1163
	__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 */
1164
	if (skb->rtable->rt_flags & (RTCF_BROADCAST | RTCF_MULTICAST))
L
Linus Torvalds 已提交
1165 1166 1167 1168 1169 1170
		goto drop;

	/* TW buckets are converted to open requests without
	 * limitations, they conserve resources and peer is
	 * evidently real one.
	 */
1171
	if (inet_csk_reqsk_queue_is_full(sk) && !isn) {
L
Linus Torvalds 已提交
1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184
#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.
	 */
1185
	if (sk_acceptq_is_full(sk) && inet_csk_reqsk_queue_young(sk) > 1)
L
Linus Torvalds 已提交
1186 1187
		goto drop;

1188
	req = inet_reqsk_alloc(&tcp_request_sock_ops);
L
Linus Torvalds 已提交
1189 1190 1191
	if (!req)
		goto drop;

1192 1193 1194 1195
#ifdef CONFIG_TCP_MD5SIG
	tcp_rsk(req)->af_specific = &tcp_request_sock_ipv4_ops;
#endif

L
Linus Torvalds 已提交
1196 1197 1198 1199 1200 1201
	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);

1202
	if (want_cookie && !tmp_opt.saw_tstamp)
L
Linus Torvalds 已提交
1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217
		tcp_clear_options(&tmp_opt);

	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);

1218 1219 1220
	if (security_inet_conn_request(sk, skb, req))
		goto drop_and_free;

1221 1222 1223 1224
	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
1225
	if (!want_cookie)
1226
		TCP_ECN_create_request(req, tcp_hdr(skb));
L
Linus Torvalds 已提交
1227 1228 1229 1230

	if (want_cookie) {
#ifdef CONFIG_SYN_COOKIES
		syn_flood_warning(skb);
1231
		req->cookie_ts = tmp_opt.tstamp_ok;
L
Linus Torvalds 已提交
1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246
#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 &&
1247
		    tcp_death_row.sysctl_tw_recycle &&
1248
		    (dst = inet_csk_route_req(sk, req)) != NULL &&
L
Linus Torvalds 已提交
1249 1250
		    (peer = rt_get_peer((struct rtable *)dst)) != NULL &&
		    peer->v4daddr == saddr) {
1251
			if (get_seconds() < peer->tcp_ts_stamp + TCP_PAWS_MSL &&
L
Linus Torvalds 已提交
1252 1253 1254
			    (s32)(peer->tcp_ts - req->ts_recent) >
							TCP_PAWS_WINDOW) {
				NET_INC_STATS_BH(LINUX_MIB_PAWSPASSIVEREJECTED);
1255
				goto drop_and_release;
L
Linus Torvalds 已提交
1256 1257 1258 1259
			}
		}
		/* Kill the following clause, if you dislike this way. */
		else if (!sysctl_tcp_syncookies &&
1260
			 (sysctl_max_syn_backlog - inet_csk_reqsk_queue_len(sk) <
L
Linus Torvalds 已提交
1261 1262 1263 1264 1265 1266 1267 1268 1269 1270
			  (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.
			 */
1271
			LIMIT_NETDEBUG(KERN_DEBUG "TCP: drop open "
1272
				       "request from " NIPQUAD_FMT "/%u\n",
1273
				       NIPQUAD(saddr),
1274
				       ntohs(tcp_hdr(skb)->source));
1275
			goto drop_and_release;
L
Linus Torvalds 已提交
1276 1277
		}

1278
		isn = tcp_v4_init_sequence(skb);
L
Linus Torvalds 已提交
1279
	}
1280
	tcp_rsk(req)->snt_isn = isn;
L
Linus Torvalds 已提交
1281

1282
	if (__tcp_v4_send_synack(sk, req, dst) || want_cookie)
L
Linus Torvalds 已提交
1283 1284
		goto drop_and_free;

1285
	inet_csk_reqsk_queue_hash_add(sk, req, TCP_TIMEOUT_INIT);
L
Linus Torvalds 已提交
1286 1287
	return 0;

1288 1289
drop_and_release:
	dst_release(dst);
L
Linus Torvalds 已提交
1290
drop_and_free:
1291
	reqsk_free(req);
L
Linus Torvalds 已提交
1292 1293 1294 1295 1296 1297 1298 1299 1300 1301
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,
1302
				  struct request_sock *req,
L
Linus Torvalds 已提交
1303 1304
				  struct dst_entry *dst)
{
1305
	struct inet_request_sock *ireq;
L
Linus Torvalds 已提交
1306 1307 1308
	struct inet_sock *newinet;
	struct tcp_sock *newtp;
	struct sock *newsk;
1309 1310 1311
#ifdef CONFIG_TCP_MD5SIG
	struct tcp_md5sig_key *key;
#endif
L
Linus Torvalds 已提交
1312 1313 1314 1315

	if (sk_acceptq_is_full(sk))
		goto exit_overflow;

1316
	if (!dst && (dst = inet_csk_route_req(sk, req)) == NULL)
L
Linus Torvalds 已提交
1317 1318 1319 1320 1321 1322
		goto exit;

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

1323
	newsk->sk_gso_type = SKB_GSO_TCPV4;
1324
	sk_setup_caps(newsk, dst);
L
Linus Torvalds 已提交
1325 1326 1327

	newtp		      = tcp_sk(newsk);
	newinet		      = inet_sk(newsk);
1328 1329 1330 1331 1332 1333
	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;
1334
	newinet->mc_index     = inet_iif(skb);
1335
	newinet->mc_ttl	      = ip_hdr(skb)->ttl;
1336
	inet_csk(newsk)->icsk_ext_hdr_len = 0;
L
Linus Torvalds 已提交
1337
	if (newinet->opt)
1338
		inet_csk(newsk)->icsk_ext_hdr_len = newinet->opt->optlen;
L
Linus Torvalds 已提交
1339 1340
	newinet->id = newtp->write_seq ^ jiffies;

J
John Heffner 已提交
1341
	tcp_mtup_init(newsk);
L
Linus Torvalds 已提交
1342 1343 1344 1345
	tcp_sync_mss(newsk, dst_mtu(dst));
	newtp->advmss = dst_metric(dst, RTAX_ADVMSS);
	tcp_initialize_rcv_mss(newsk);

1346 1347 1348 1349 1350 1351 1352 1353 1354
#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.
		 */
1355 1356
		char *newkey = kmemdup(key->key, key->keylen, GFP_ATOMIC);
		if (newkey != NULL)
1357 1358 1359 1360 1361
			tcp_v4_md5_do_add(newsk, inet_sk(sk)->daddr,
					  newkey, key->keylen);
	}
#endif

1362 1363
	__inet_hash_nolisten(newsk);
	__inet_inherit_port(sk, newsk);
L
Linus Torvalds 已提交
1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376

	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)
{
1377
	struct tcphdr *th = tcp_hdr(skb);
1378
	const struct iphdr *iph = ip_hdr(skb);
L
Linus Torvalds 已提交
1379
	struct sock *nsk;
1380
	struct request_sock **prev;
L
Linus Torvalds 已提交
1381
	/* Find possible connection requests. */
1382 1383
	struct request_sock *req = inet_csk_search_req(sk, &prev, th->source,
						       iph->saddr, iph->daddr);
L
Linus Torvalds 已提交
1384 1385 1386
	if (req)
		return tcp_check_req(sk, skb, req, prev);

1387
	nsk = inet_lookup_established(sock_net(sk), &tcp_hashinfo, iph->saddr,
1388
			th->source, iph->daddr, th->dest, inet_iif(skb));
L
Linus Torvalds 已提交
1389 1390 1391 1392 1393 1394

	if (nsk) {
		if (nsk->sk_state != TCP_TIME_WAIT) {
			bh_lock_sock(nsk);
			return nsk;
		}
1395
		inet_twsk_put(inet_twsk(nsk));
L
Linus Torvalds 已提交
1396 1397 1398 1399 1400 1401 1402 1403 1404 1405
		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;
}

1406
static __sum16 tcp_v4_checksum_init(struct sk_buff *skb)
L
Linus Torvalds 已提交
1407
{
1408 1409
	const struct iphdr *iph = ip_hdr(skb);

1410
	if (skb->ip_summed == CHECKSUM_COMPLETE) {
1411 1412
		if (!tcp_v4_check(skb->len, iph->saddr,
				  iph->daddr, skb->csum)) {
1413
			skb->ip_summed = CHECKSUM_UNNECESSARY;
L
Linus Torvalds 已提交
1414
			return 0;
1415
		}
L
Linus Torvalds 已提交
1416
	}
1417

1418
	skb->csum = csum_tcpudp_nofold(iph->saddr, iph->daddr,
1419 1420
				       skb->len, IPPROTO_TCP, 0);

L
Linus Torvalds 已提交
1421
	if (skb->len <= 76) {
1422
		return __skb_checksum_complete(skb);
L
Linus Torvalds 已提交
1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437
	}
	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)
{
1438 1439 1440 1441 1442 1443 1444 1445
	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
	 */
1446
	if (tcp_v4_inbound_md5_hash(sk, skb))
1447 1448 1449
		goto discard;
#endif

L
Linus Torvalds 已提交
1450 1451
	if (sk->sk_state == TCP_ESTABLISHED) { /* Fast path */
		TCP_CHECK_TIMER(sk);
1452
		if (tcp_rcv_established(sk, skb, tcp_hdr(skb), skb->len)) {
1453
			rsk = sk;
L
Linus Torvalds 已提交
1454
			goto reset;
1455
		}
L
Linus Torvalds 已提交
1456 1457 1458 1459
		TCP_CHECK_TIMER(sk);
		return 0;
	}

1460
	if (skb->len < tcp_hdrlen(skb) || tcp_checksum_complete(skb))
L
Linus Torvalds 已提交
1461 1462 1463 1464 1465 1466 1467 1468
		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) {
1469 1470
			if (tcp_child_process(sk, nsk, skb)) {
				rsk = nsk;
L
Linus Torvalds 已提交
1471
				goto reset;
1472
			}
L
Linus Torvalds 已提交
1473 1474 1475 1476 1477
			return 0;
		}
	}

	TCP_CHECK_TIMER(sk);
1478
	if (tcp_rcv_state_process(sk, skb, tcp_hdr(skb), skb->len)) {
1479
		rsk = sk;
L
Linus Torvalds 已提交
1480
		goto reset;
1481
	}
L
Linus Torvalds 已提交
1482 1483 1484 1485
	TCP_CHECK_TIMER(sk);
	return 0;

reset:
1486
	tcp_v4_send_reset(rsk, skb);
L
Linus Torvalds 已提交
1487 1488 1489 1490 1491 1492 1493 1494 1495 1496
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:
1497
	TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_INERRS);
L
Linus Torvalds 已提交
1498 1499 1500 1501 1502 1503 1504 1505 1506
	goto discard;
}

/*
 *	From tcp_input.c
 */

int tcp_v4_rcv(struct sk_buff *skb)
{
1507
	const struct iphdr *iph;
L
Linus Torvalds 已提交
1508 1509 1510
	struct tcphdr *th;
	struct sock *sk;
	int ret;
1511
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
1512 1513 1514 1515 1516

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

	/* Count it even if it's bad */
1517
	TCP_INC_STATS_BH(net, TCP_MIB_INSEGS);
L
Linus Torvalds 已提交
1518 1519 1520 1521

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

1522
	th = tcp_hdr(skb);
L
Linus Torvalds 已提交
1523 1524 1525 1526 1527 1528 1529 1530

	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
1531
	 * provided case of th->doff==0 is eliminated.
L
Linus Torvalds 已提交
1532
	 * So, we defer the checks. */
1533
	if (!skb_csum_unnecessary(skb) && tcp_v4_checksum_init(skb))
L
Linus Torvalds 已提交
1534 1535
		goto bad_packet;

1536
	th = tcp_hdr(skb);
1537
	iph = ip_hdr(skb);
L
Linus Torvalds 已提交
1538 1539 1540 1541 1542
	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;
1543
	TCP_SKB_CB(skb)->flags	 = iph->tos;
L
Linus Torvalds 已提交
1544 1545
	TCP_SKB_CB(skb)->sacked	 = 0;

1546
	sk = __inet_lookup(net, &tcp_hashinfo, iph->saddr,
1547
			th->source, iph->daddr, th->dest, inet_iif(skb));
L
Linus Torvalds 已提交
1548 1549 1550 1551 1552 1553 1554 1555 1556
	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;
1557
	nf_reset(skb);
L
Linus Torvalds 已提交
1558

1559
	if (sk_filter(sk, skb))
L
Linus Torvalds 已提交
1560 1561 1562 1563
		goto discard_and_relse;

	skb->dev = NULL;

1564
	bh_lock_sock_nested(sk);
L
Linus Torvalds 已提交
1565 1566
	ret = 0;
	if (!sock_owned_by_user(sk)) {
C
Chris Leech 已提交
1567 1568 1569 1570 1571
#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
1572
			ret = tcp_v4_do_rcv(sk, skb);
C
Chris Leech 已提交
1573 1574 1575 1576 1577 1578
		else
#endif
		{
			if (!tcp_prequeue(sk, skb))
			ret = tcp_v4_do_rcv(sk, skb);
		}
L
Linus Torvalds 已提交
1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592
	} 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:
1593
		TCP_INC_STATS_BH(net, TCP_MIB_INERRS);
L
Linus Torvalds 已提交
1594
	} else {
1595
		tcp_v4_send_reset(NULL, skb);
L
Linus Torvalds 已提交
1596 1597 1598 1599 1600
	}

discard_it:
	/* Discard frame. */
	kfree_skb(skb);
1601
	return 0;
L
Linus Torvalds 已提交
1602 1603 1604 1605 1606 1607 1608

discard_and_relse:
	sock_put(sk);
	goto discard_it;

do_time_wait:
	if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) {
1609
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1610 1611 1612 1613
		goto discard_it;
	}

	if (skb->len < (th->doff << 2) || tcp_checksum_complete(skb)) {
1614
		TCP_INC_STATS_BH(net, TCP_MIB_INERRS);
1615
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1616 1617
		goto discard_it;
	}
1618
	switch (tcp_timewait_state_process(inet_twsk(sk), skb, th)) {
L
Linus Torvalds 已提交
1619
	case TCP_TW_SYN: {
1620
		struct sock *sk2 = inet_lookup_listener(dev_net(skb->dev),
1621
							&tcp_hashinfo,
1622
							iph->daddr, th->dest,
1623
							inet_iif(skb));
L
Linus Torvalds 已提交
1624
		if (sk2) {
1625 1626
			inet_twsk_deschedule(inet_twsk(sk), &tcp_death_row);
			inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666
			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 ||
1667
		    (peer->tcp_ts_stamp + TCP_PAWS_MSL < get_seconds() &&
L
Linus Torvalds 已提交
1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679
		     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;
}

1680
int tcp_v4_tw_remember_stamp(struct inet_timewait_sock *tw)
L
Linus Torvalds 已提交
1681
{
1682
	struct inet_peer *peer = inet_getpeer(tw->tw_daddr, 1);
L
Linus Torvalds 已提交
1683 1684

	if (peer) {
1685 1686 1687
		const struct tcp_timewait_sock *tcptw = tcp_twsk((struct sock *)tw);

		if ((s32)(peer->tcp_ts - tcptw->tw_ts_recent) <= 0 ||
1688
		    (peer->tcp_ts_stamp + TCP_PAWS_MSL < get_seconds() &&
1689 1690 1691
		     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
1692 1693 1694 1695 1696 1697 1698 1699
		}
		inet_putpeer(peer);
		return 1;
	}

	return 0;
}

1700
struct inet_connection_sock_af_ops ipv4_specific = {
1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711
	.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),
1712
	.bind_conflict	   = inet_csk_bind_conflict,
1713
#ifdef CONFIG_COMPAT
1714 1715
	.compat_setsockopt = compat_ip_setsockopt,
	.compat_getsockopt = compat_ip_getsockopt,
1716
#endif
L
Linus Torvalds 已提交
1717 1718
};

1719
#ifdef CONFIG_TCP_MD5SIG
1720
static struct tcp_sock_af_ops tcp_sock_ipv4_specific = {
1721 1722 1723 1724 1725
	.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,
};
1726
#endif
1727

L
Linus Torvalds 已提交
1728 1729 1730 1731 1732
/* 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)
{
1733
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
1734 1735 1736 1737 1738 1739
	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);

1740
	icsk->icsk_rto = TCP_TIMEOUT_INIT;
L
Linus Torvalds 已提交
1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754
	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;
1755
	tp->mss_cache = 536;
L
Linus Torvalds 已提交
1756 1757

	tp->reordering = sysctl_tcp_reordering;
1758
	icsk->icsk_ca_ops = &tcp_init_congestion_ops;
L
Linus Torvalds 已提交
1759 1760 1761 1762 1763 1764

	sk->sk_state = TCP_CLOSE;

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

1765
	icsk->icsk_af_ops = &ipv4_specific;
1766
	icsk->icsk_sync_mss = tcp_sync_mss;
1767 1768 1769
#ifdef CONFIG_TCP_MD5SIG
	tp->af_specific = &tcp_sock_ipv4_specific;
#endif
L
Linus Torvalds 已提交
1770 1771 1772 1773 1774 1775 1776 1777 1778

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

	atomic_inc(&tcp_sockets_allocated);

	return 0;
}

1779
void tcp_v4_destroy_sock(struct sock *sk)
L
Linus Torvalds 已提交
1780 1781 1782 1783 1784
{
	struct tcp_sock *tp = tcp_sk(sk);

	tcp_clear_xmit_timers(sk);

1785
	tcp_cleanup_congestion_control(sk);
1786

L
Linus Torvalds 已提交
1787
	/* Cleanup up the write buffer. */
1788
	tcp_write_queue_purge(sk);
L
Linus Torvalds 已提交
1789 1790

	/* Cleans up our, hopefully empty, out_of_order_queue. */
1791
	__skb_queue_purge(&tp->out_of_order_queue);
L
Linus Torvalds 已提交
1792

1793 1794 1795 1796 1797 1798 1799 1800 1801
#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
1802 1803
#ifdef CONFIG_NET_DMA
	/* Cleans up our sk_async_wait_queue */
1804
	__skb_queue_purge(&sk->sk_async_wait_queue);
C
Chris Leech 已提交
1805 1806
#endif

L
Linus Torvalds 已提交
1807 1808 1809 1810
	/* Clean prequeue, it must be empty really */
	__skb_queue_purge(&tp->ucopy.prequeue);

	/* Clean up a referenced TCP bind bucket. */
1811
	if (inet_csk(sk)->icsk_bind_hash)
1812
		inet_put_port(sk);
L
Linus Torvalds 已提交
1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829

	/*
	 * 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);
}

EXPORT_SYMBOL(tcp_v4_destroy_sock);

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

1830
static inline struct inet_timewait_sock *tw_head(struct hlist_head *head)
L
Linus Torvalds 已提交
1831 1832
{
	return hlist_empty(head) ? NULL :
1833
		list_entry(head->first, struct inet_timewait_sock, tw_node);
L
Linus Torvalds 已提交
1834 1835
}

1836
static inline struct inet_timewait_sock *tw_next(struct inet_timewait_sock *tw)
L
Linus Torvalds 已提交
1837 1838 1839 1840 1841 1842 1843
{
	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)
{
1844
	struct inet_connection_sock *icsk;
L
Linus Torvalds 已提交
1845 1846 1847
	struct hlist_node *node;
	struct sock *sk = cur;
	struct tcp_iter_state* st = seq->private;
1848
	struct net *net = seq_file_net(seq);
L
Linus Torvalds 已提交
1849 1850 1851

	if (!sk) {
		st->bucket = 0;
1852
		sk = sk_head(&tcp_hashinfo.listening_hash[0]);
L
Linus Torvalds 已提交
1853 1854 1855 1856 1857 1858
		goto get_sk;
	}

	++st->num;

	if (st->state == TCP_SEQ_STATE_OPENREQ) {
1859
		struct request_sock *req = cur;
L
Linus Torvalds 已提交
1860

1861
		icsk = inet_csk(st->syn_wait_sk);
L
Linus Torvalds 已提交
1862 1863 1864
		req = req->dl_next;
		while (1) {
			while (req) {
1865
				if (req->rsk_ops->family == st->family &&
1866
				    net_eq(sock_net(req->sk), net)) {
L
Linus Torvalds 已提交
1867 1868 1869 1870 1871
					cur = req;
					goto out;
				}
				req = req->dl_next;
			}
1872
			if (++st->sbucket >= icsk->icsk_accept_queue.listen_opt->nr_table_entries)
L
Linus Torvalds 已提交
1873 1874
				break;
get_req:
1875
			req = icsk->icsk_accept_queue.listen_opt->syn_table[st->sbucket];
L
Linus Torvalds 已提交
1876 1877 1878
		}
		sk	  = sk_next(st->syn_wait_sk);
		st->state = TCP_SEQ_STATE_LISTENING;
1879
		read_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
L
Linus Torvalds 已提交
1880
	} else {
1881
		icsk = inet_csk(sk);
1882 1883
		read_lock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
		if (reqsk_queue_len(&icsk->icsk_accept_queue))
L
Linus Torvalds 已提交
1884
			goto start_req;
1885
		read_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
L
Linus Torvalds 已提交
1886 1887 1888 1889
		sk = sk_next(sk);
	}
get_sk:
	sk_for_each_from(sk, node) {
1890
		if (sk->sk_family == st->family && net_eq(sock_net(sk), net)) {
L
Linus Torvalds 已提交
1891 1892 1893
			cur = sk;
			goto out;
		}
1894
		icsk = inet_csk(sk);
1895 1896
		read_lock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
		if (reqsk_queue_len(&icsk->icsk_accept_queue)) {
L
Linus Torvalds 已提交
1897 1898 1899 1900 1901 1902 1903
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;
		}
1904
		read_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
L
Linus Torvalds 已提交
1905
	}
1906
	if (++st->bucket < INET_LHTABLE_SIZE) {
1907
		sk = sk_head(&tcp_hashinfo.listening_hash[st->bucket]);
L
Linus Torvalds 已提交
1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928
		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;
1929
	struct net *net = seq_file_net(seq);
L
Linus Torvalds 已提交
1930 1931
	void *rc = NULL;

1932
	for (st->bucket = 0; st->bucket < tcp_hashinfo.ehash_size; ++st->bucket) {
L
Linus Torvalds 已提交
1933 1934
		struct sock *sk;
		struct hlist_node *node;
1935
		struct inet_timewait_sock *tw;
1936
		rwlock_t *lock = inet_ehash_lockp(&tcp_hashinfo, st->bucket);
L
Linus Torvalds 已提交
1937

1938
		read_lock_bh(lock);
1939
		sk_for_each(sk, node, &tcp_hashinfo.ehash[st->bucket].chain) {
1940
			if (sk->sk_family != st->family ||
1941
			    !net_eq(sock_net(sk), net)) {
L
Linus Torvalds 已提交
1942 1943 1944 1945 1946 1947
				continue;
			}
			rc = sk;
			goto out;
		}
		st->state = TCP_SEQ_STATE_TIME_WAIT;
1948
		inet_twsk_for_each(tw, node,
1949
				   &tcp_hashinfo.ehash[st->bucket].twchain) {
1950
			if (tw->tw_family != st->family ||
1951
			    !net_eq(twsk_net(tw), net)) {
L
Linus Torvalds 已提交
1952 1953 1954 1955 1956
				continue;
			}
			rc = tw;
			goto out;
		}
1957
		read_unlock_bh(lock);
L
Linus Torvalds 已提交
1958 1959 1960 1961 1962 1963 1964 1965 1966
		st->state = TCP_SEQ_STATE_ESTABLISHED;
	}
out:
	return rc;
}

static void *established_get_next(struct seq_file *seq, void *cur)
{
	struct sock *sk = cur;
1967
	struct inet_timewait_sock *tw;
L
Linus Torvalds 已提交
1968 1969
	struct hlist_node *node;
	struct tcp_iter_state* st = seq->private;
1970
	struct net *net = seq_file_net(seq);
L
Linus Torvalds 已提交
1971 1972 1973 1974 1975 1976 1977

	++st->num;

	if (st->state == TCP_SEQ_STATE_TIME_WAIT) {
		tw = cur;
		tw = tw_next(tw);
get_tw:
1978
		while (tw && (tw->tw_family != st->family || !net_eq(twsk_net(tw), net))) {
L
Linus Torvalds 已提交
1979 1980 1981 1982 1983 1984
			tw = tw_next(tw);
		}
		if (tw) {
			cur = tw;
			goto out;
		}
1985
		read_unlock_bh(inet_ehash_lockp(&tcp_hashinfo, st->bucket));
L
Linus Torvalds 已提交
1986 1987
		st->state = TCP_SEQ_STATE_ESTABLISHED;

1988
		if (++st->bucket < tcp_hashinfo.ehash_size) {
1989
			read_lock_bh(inet_ehash_lockp(&tcp_hashinfo, st->bucket));
1990
			sk = sk_head(&tcp_hashinfo.ehash[st->bucket].chain);
L
Linus Torvalds 已提交
1991 1992 1993 1994 1995 1996 1997 1998
		} else {
			cur = NULL;
			goto out;
		}
	} else
		sk = sk_next(sk);

	sk_for_each_from(sk, node) {
1999
		if (sk->sk_family == st->family && net_eq(sock_net(sk), net))
L
Linus Torvalds 已提交
2000 2001 2002 2003
			goto found;
	}

	st->state = TCP_SEQ_STATE_TIME_WAIT;
2004
	tw = tw_head(&tcp_hashinfo.ehash[st->bucket].twchain);
L
Linus Torvalds 已提交
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
	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;
2019
	}
L
Linus Torvalds 已提交
2020 2021 2022 2023 2024 2025 2026 2027
	return rc;
}

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

2028
	inet_listen_lock(&tcp_hashinfo);
L
Linus Torvalds 已提交
2029 2030 2031 2032
	st->state = TCP_SEQ_STATE_LISTENING;
	rc	  = listening_get_idx(seq, &pos);

	if (!rc) {
2033
		inet_listen_unlock(&tcp_hashinfo);
L
Linus Torvalds 已提交
2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064
		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) {
2065
			inet_listen_unlock(&tcp_hashinfo);
L
Linus Torvalds 已提交
2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086
			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) {
2087 2088
			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
2089 2090 2091
		}
	case TCP_SEQ_STATE_LISTENING:
		if (v != SEQ_START_TOKEN)
2092
			inet_listen_unlock(&tcp_hashinfo);
L
Linus Torvalds 已提交
2093 2094 2095 2096
		break;
	case TCP_SEQ_STATE_TIME_WAIT:
	case TCP_SEQ_STATE_ESTABLISHED:
		if (v)
2097
			read_unlock_bh(inet_ehash_lockp(&tcp_hashinfo, st->bucket));
L
Linus Torvalds 已提交
2098 2099 2100 2101 2102 2103 2104 2105
		break;
	}
}

static int tcp_seq_open(struct inode *inode, struct file *file)
{
	struct tcp_seq_afinfo *afinfo = PDE(inode)->data;
	struct tcp_iter_state *s;
2106
	int err;
L
Linus Torvalds 已提交
2107

2108 2109 2110 2111
	err = seq_open_net(inode, file, &afinfo->seq_ops,
			  sizeof(struct tcp_iter_state));
	if (err < 0)
		return err;
2112

2113
	s = ((struct seq_file *)file->private_data)->private;
L
Linus Torvalds 已提交
2114
	s->family		= afinfo->family;
2115 2116 2117
	return 0;
}

2118
int tcp_proc_register(struct net *net, struct tcp_seq_afinfo *afinfo)
L
Linus Torvalds 已提交
2119 2120 2121 2122
{
	int rc = 0;
	struct proc_dir_entry *p;

2123 2124 2125 2126
	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_net;
2127

2128 2129 2130 2131
	afinfo->seq_ops.start		= tcp_seq_start;
	afinfo->seq_ops.next		= tcp_seq_next;
	afinfo->seq_ops.stop		= tcp_seq_stop;

2132 2133 2134
	p = proc_create_data(afinfo->name, S_IRUGO, net->proc_net,
			     &afinfo->seq_fops, afinfo);
	if (!p)
L
Linus Torvalds 已提交
2135 2136 2137 2138
		rc = -ENOMEM;
	return rc;
}

2139
void tcp_proc_unregister(struct net *net, struct tcp_seq_afinfo *afinfo)
L
Linus Torvalds 已提交
2140
{
2141
	proc_net_remove(net, afinfo->name);
L
Linus Torvalds 已提交
2142 2143
}

2144
static void get_openreq4(struct sock *sk, struct request_sock *req,
2145
			 struct seq_file *f, int i, int uid, int *len)
L
Linus Torvalds 已提交
2146
{
2147
	const struct inet_request_sock *ireq = inet_rsk(req);
L
Linus Torvalds 已提交
2148 2149
	int ttd = req->expires - jiffies;

2150 2151
	seq_printf(f, "%4d: %08X:%04X %08X:%04X"
		" %02X %08X:%08X %02X:%08lX %08X %5d %8d %u %d %p%n",
L
Linus Torvalds 已提交
2152
		i,
2153
		ireq->loc_addr,
L
Linus Torvalds 已提交
2154
		ntohs(inet_sk(sk)->sport),
2155 2156
		ireq->rmt_addr,
		ntohs(ireq->rmt_port),
L
Linus Torvalds 已提交
2157 2158 2159 2160 2161 2162 2163 2164 2165
		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),
2166 2167
		req,
		len);
L
Linus Torvalds 已提交
2168 2169
}

2170
static void get_tcp4_sock(struct sock *sk, struct seq_file *f, int i, int *len)
L
Linus Torvalds 已提交
2171 2172 2173
{
	int timer_active;
	unsigned long timer_expires;
2174 2175 2176
	struct tcp_sock *tp = tcp_sk(sk);
	const struct inet_connection_sock *icsk = inet_csk(sk);
	struct inet_sock *inet = inet_sk(sk);
2177 2178
	__be32 dest = inet->daddr;
	__be32 src = inet->rcv_saddr;
L
Linus Torvalds 已提交
2179 2180 2181
	__u16 destp = ntohs(inet->dport);
	__u16 srcp = ntohs(inet->sport);

2182
	if (icsk->icsk_pending == ICSK_TIME_RETRANS) {
L
Linus Torvalds 已提交
2183
		timer_active	= 1;
2184 2185
		timer_expires	= icsk->icsk_timeout;
	} else if (icsk->icsk_pending == ICSK_TIME_PROBE0) {
L
Linus Torvalds 已提交
2186
		timer_active	= 4;
2187
		timer_expires	= icsk->icsk_timeout;
2188
	} else if (timer_pending(&sk->sk_timer)) {
L
Linus Torvalds 已提交
2189
		timer_active	= 2;
2190
		timer_expires	= sk->sk_timer.expires;
L
Linus Torvalds 已提交
2191 2192 2193 2194 2195
	} else {
		timer_active	= 0;
		timer_expires = jiffies;
	}

2196
	seq_printf(f, "%4d: %08X:%04X %08X:%04X %02X %08X:%08X %02X:%08lX "
2197
			"%08X %5d %8d %lu %d %p %lu %lu %u %u %d%n",
2198
		i, src, srcp, dest, destp, sk->sk_state,
2199
		tp->write_seq - tp->snd_una,
2200
		sk->sk_state == TCP_LISTEN ? sk->sk_ack_backlog :
2201
					     (tp->rcv_nxt - tp->copied_seq),
L
Linus Torvalds 已提交
2202 2203
		timer_active,
		jiffies_to_clock_t(timer_expires - jiffies),
2204
		icsk->icsk_retransmits,
2205
		sock_i_uid(sk),
2206
		icsk->icsk_probes_out,
2207 2208
		sock_i_ino(sk),
		atomic_read(&sk->sk_refcnt), sk,
2209 2210
		jiffies_to_clock_t(icsk->icsk_rto),
		jiffies_to_clock_t(icsk->icsk_ack.ato),
2211
		(icsk->icsk_ack.quick << 1) | icsk->icsk_ack.pingpong,
L
Linus Torvalds 已提交
2212
		tp->snd_cwnd,
2213 2214
		tp->snd_ssthresh >= 0xFFFF ? -1 : tp->snd_ssthresh,
		len);
L
Linus Torvalds 已提交
2215 2216
}

2217
static void get_timewait4_sock(struct inet_timewait_sock *tw,
2218
			       struct seq_file *f, int i, int *len)
L
Linus Torvalds 已提交
2219
{
2220
	__be32 dest, src;
L
Linus Torvalds 已提交
2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231
	__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);

2232 2233
	seq_printf(f, "%4d: %08X:%04X %08X:%04X"
		" %02X %08X:%08X %02X:%08lX %08X %5d %8d %d %d %p%n",
L
Linus Torvalds 已提交
2234 2235
		i, src, srcp, dest, destp, tw->tw_substate, 0, 0,
		3, jiffies_to_clock_t(ttd), 0, 0, 0, 0,
2236
		atomic_read(&tw->tw_refcnt), tw, len);
L
Linus Torvalds 已提交
2237 2238 2239 2240 2241 2242 2243
}

#define TMPSZ 150

static int tcp4_seq_show(struct seq_file *seq, void *v)
{
	struct tcp_iter_state* st;
2244
	int len;
L
Linus Torvalds 已提交
2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257

	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:
2258
		get_tcp4_sock(v, seq, st->num, &len);
L
Linus Torvalds 已提交
2259 2260
		break;
	case TCP_SEQ_STATE_OPENREQ:
2261
		get_openreq4(st->syn_wait_sk, v, seq, st->num, st->uid, &len);
L
Linus Torvalds 已提交
2262 2263
		break;
	case TCP_SEQ_STATE_TIME_WAIT:
2264
		get_timewait4_sock(v, seq, st->num, &len);
L
Linus Torvalds 已提交
2265 2266
		break;
	}
2267
	seq_printf(seq, "%*s\n", TMPSZ - 1 - len, "");
L
Linus Torvalds 已提交
2268 2269 2270 2271 2272 2273 2274
out:
	return 0;
}

static struct tcp_seq_afinfo tcp4_seq_afinfo = {
	.name		= "tcp",
	.family		= AF_INET,
2275 2276 2277
	.seq_fops	= {
		.owner		= THIS_MODULE,
	},
2278 2279 2280
	.seq_ops	= {
		.show		= tcp4_seq_show,
	},
L
Linus Torvalds 已提交
2281 2282
};

2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297
static int tcp4_proc_init_net(struct net *net)
{
	return tcp_proc_register(net, &tcp4_seq_afinfo);
}

static void tcp4_proc_exit_net(struct net *net)
{
	tcp_proc_unregister(net, &tcp4_seq_afinfo);
}

static struct pernet_operations tcp4_net_ops = {
	.init = tcp4_proc_init_net,
	.exit = tcp4_proc_exit_net,
};

L
Linus Torvalds 已提交
2298 2299
int __init tcp4_proc_init(void)
{
2300
	return register_pernet_subsys(&tcp4_net_ops);
L
Linus Torvalds 已提交
2301 2302 2303 2304
}

void tcp4_proc_exit(void)
{
2305
	unregister_pernet_subsys(&tcp4_net_ops);
L
Linus Torvalds 已提交
2306 2307 2308 2309 2310 2311 2312 2313 2314
}
#endif /* CONFIG_PROC_FS */

struct proto tcp_prot = {
	.name			= "TCP",
	.owner			= THIS_MODULE,
	.close			= tcp_close,
	.connect		= tcp_v4_connect,
	.disconnect		= tcp_disconnect,
2315
	.accept			= inet_csk_accept,
L
Linus Torvalds 已提交
2316 2317 2318 2319 2320 2321 2322 2323
	.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,
2324 2325 2326
	.hash			= inet_hash,
	.unhash			= inet_unhash,
	.get_port		= inet_csk_get_port,
L
Linus Torvalds 已提交
2327 2328
	.enter_memory_pressure	= tcp_enter_memory_pressure,
	.sockets_allocated	= &tcp_sockets_allocated,
2329
	.orphan_count		= &tcp_orphan_count,
L
Linus Torvalds 已提交
2330 2331 2332 2333 2334 2335 2336
	.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),
2337
	.twsk_prot		= &tcp_timewait_sock_ops,
2338
	.rsk_prot		= &tcp_request_sock_ops,
2339
	.h.hashinfo		= &tcp_hashinfo,
2340 2341 2342 2343
#ifdef CONFIG_COMPAT
	.compat_setsockopt	= compat_tcp_setsockopt,
	.compat_getsockopt	= compat_tcp_getsockopt,
#endif
L
Linus Torvalds 已提交
2344 2345
};

2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362

static int __net_init tcp_sk_init(struct net *net)
{
	return inet_ctl_sock_create(&net->ipv4.tcp_sock,
				    PF_INET, SOCK_RAW, IPPROTO_TCP, net);
}

static void __net_exit tcp_sk_exit(struct net *net)
{
	inet_ctl_sock_destroy(net->ipv4.tcp_sock);
}

static struct pernet_operations __net_initdata tcp_sk_ops = {
       .init = tcp_sk_init,
       .exit = tcp_sk_exit,
};

2363
void __init tcp_v4_init(void)
L
Linus Torvalds 已提交
2364
{
2365
	if (register_pernet_device(&tcp_sk_ops))
L
Linus Torvalds 已提交
2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384
		panic("Failed to create the TCP control socket.\n");
}

EXPORT_SYMBOL(ipv4_specific);
EXPORT_SYMBOL(tcp_hashinfo);
EXPORT_SYMBOL(tcp_prot);
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);