tcp_ipv4.c 60.5 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 91
static int tcp_v4_md5_hash_hdr(char *md5_hash, struct tcp_md5sig_key *key,
			       __be32 daddr, __be32 saddr, struct tcphdr *th);
92 93 94 95 96 97
#else
static inline
struct tcp_md5sig_key *tcp_v4_md5_do_lookup(struct sock *sk, __be32 addr)
{
	return NULL;
}
98 99
#endif

100
struct inet_hashinfo tcp_hashinfo;
L
Linus Torvalds 已提交
101

102
static inline __u32 tcp_v4_init_sequence(struct sk_buff *skb)
L
Linus Torvalds 已提交
103
{
104 105
	return secure_tcp_sequence_number(ip_hdr(skb)->daddr,
					  ip_hdr(skb)->saddr,
106 107
					  tcp_hdr(skb)->dest,
					  tcp_hdr(skb)->source);
L
Linus Torvalds 已提交
108 109
}

110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127
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 &&
128
			     get_seconds() - tcptw->tw_ts_recent_stamp > 1))) {
129 130 131 132 133 134 135 136 137 138 139 140 141 142
		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
143 144 145 146 147 148 149
/* 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;
150
	__be32 daddr, nexthop;
L
Linus Torvalds 已提交
151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169
	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,
170
			       inet->sport, usin->sin_port, sk, 1);
171 172
	if (tmp < 0) {
		if (tmp == -ENETUNREACH)
173
			IP_INC_STATS_BH(sock_net(sk), IPSTATS_MIB_OUTNOROUTES);
L
Linus Torvalds 已提交
174
		return tmp;
175
	}
L
Linus Torvalds 已提交
176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195

	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;
	}

196
	if (tcp_death_row.sysctl_tw_recycle &&
L
Linus Torvalds 已提交
197 198
	    !tp->rx_opt.ts_recent_stamp && rt->rt_dst == daddr) {
		struct inet_peer *peer = rt_get_peer(rt);
199 200 201 202 203
		/*
		 * 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
204
		 */
205
		if (peer != NULL &&
206
		    peer->tcp_ts_stamp + TCP_PAWS_MSL >= get_seconds()) {
L
Linus Torvalds 已提交
207 208 209 210 211 212 213 214
			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;

215
	inet_csk(sk)->icsk_ext_hdr_len = 0;
L
Linus Torvalds 已提交
216
	if (inet->opt)
217
		inet_csk(sk)->icsk_ext_hdr_len = inet->opt->optlen;
L
Linus Torvalds 已提交
218 219 220 221 222 223 224 225 226

	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);
227
	err = inet_hash_connect(&tcp_death_row, sk);
L
Linus Torvalds 已提交
228 229 230
	if (err)
		goto failure;

231 232
	err = ip_route_newports(&rt, IPPROTO_TCP,
				inet->sport, inet->dport, sk);
L
Linus Torvalds 已提交
233 234 235 236
	if (err)
		goto failure;

	/* OK, now commit destination to socket.  */
237
	sk->sk_gso_type = SKB_GSO_TCPV4;
238
	sk_setup_caps(sk, &rt->u.dst);
L
Linus Torvalds 已提交
239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255

	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:
256 257 258 259
	/*
	 * This unhashes the socket and releases the local port,
	 * if necessary.
	 */
L
Linus Torvalds 已提交
260 261 262 263 264 265 266 267 268 269
	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
270
static void do_pmtu_discovery(struct sock *sk, struct iphdr *iph, u32 mtu)
L
Linus Torvalds 已提交
271 272 273 274 275 276 277 278 279 280 281 282 283 284
{
	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.
285
	 * There is a small race when the user changes this flag in the
L
Linus Torvalds 已提交
286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301
	 * 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 &&
302
	    inet_csk(sk)->icsk_pmtu_cookie > mtu) {
L
Linus Torvalds 已提交
303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335
		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;
336 337
	const int type = icmp_hdr(skb)->type;
	const int code = icmp_hdr(skb)->code;
L
Linus Torvalds 已提交
338 339 340
	struct sock *sk;
	__u32 seq;
	int err;
341
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
342 343

	if (skb->len < (iph->ihl << 2) + 8) {
344
		ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS);
L
Linus Torvalds 已提交
345 346 347
		return;
	}

348
	sk = inet_lookup(net, &tcp_hashinfo, iph->daddr, th->dest,
349
			iph->saddr, th->source, inet_iif(skb));
L
Linus Torvalds 已提交
350
	if (!sk) {
351
		ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS);
L
Linus Torvalds 已提交
352 353 354
		return;
	}
	if (sk->sk_state == TCP_TIME_WAIT) {
355
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
356 357 358 359 360 361 362 363
		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))
364
		NET_INC_STATS_BH(net, LINUX_MIB_LOCKDROPPEDICMPS);
L
Linus Torvalds 已提交
365 366 367 368 369 370 371 372

	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)) {
373
		NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS);
L
Linus Torvalds 已提交
374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403
		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) {
404
		struct request_sock *req, **prev;
L
Linus Torvalds 已提交
405 406 407 408
	case TCP_LISTEN:
		if (sock_owned_by_user(sk))
			goto out;

409 410
		req = inet_csk_search_req(sk, &prev, th->dest,
					  iph->daddr, iph->saddr);
L
Linus Torvalds 已提交
411 412 413 414 415 416
		if (!req)
			goto out;

		/* ICMPs are not backlogged, hence we cannot get
		   an established socket here.
		 */
417
		WARN_ON(req->sk);
L
Linus Torvalds 已提交
418

419
		if (seq != tcp_rsk(req)->snt_isn) {
420
			NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS);
L
Linus Torvalds 已提交
421 422 423 424 425 426 427 428 429
			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().
		 */
430
		inet_csk_reqsk_queue_drop(sk, req, prev);
L
Linus Torvalds 已提交
431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478
		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. */
479
void tcp_v4_send_check(struct sock *sk, int len, struct sk_buff *skb)
L
Linus Torvalds 已提交
480 481
{
	struct inet_sock *inet = inet_sk(sk);
482
	struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
483

484
	if (skb->ip_summed == CHECKSUM_PARTIAL) {
485 486
		th->check = ~tcp_v4_check(len, inet->saddr,
					  inet->daddr, 0);
487
		skb->csum_start = skb_transport_header(skb) - skb->head;
A
Al Viro 已提交
488
		skb->csum_offset = offsetof(struct tcphdr, check);
L
Linus Torvalds 已提交
489
	} else {
490
		th->check = tcp_v4_check(len, inet->saddr, inet->daddr,
491
					 csum_partial(th,
L
Linus Torvalds 已提交
492 493 494 495 496
						      th->doff << 2,
						      skb->csum));
	}
}

497 498
int tcp_v4_gso_send_check(struct sk_buff *skb)
{
499
	const struct iphdr *iph;
500 501 502 503 504
	struct tcphdr *th;

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

505
	iph = ip_hdr(skb);
506
	th = tcp_hdr(skb);
507 508

	th->check = 0;
509
	th->check = ~tcp_v4_check(skb->len, iph->saddr, iph->daddr, 0);
510
	skb->csum_start = skb_transport_header(skb) - skb->head;
A
Al Viro 已提交
511
	skb->csum_offset = offsetof(struct tcphdr, check);
512
	skb->ip_summed = CHECKSUM_PARTIAL;
513 514 515
	return 0;
}

L
Linus Torvalds 已提交
516 517 518 519 520 521 522 523 524 525 526 527 528
/*
 *	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.
 */

529
static void tcp_v4_send_reset(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
530
{
531
	struct tcphdr *th = tcp_hdr(skb);
532 533 534
	struct {
		struct tcphdr th;
#ifdef CONFIG_TCP_MD5SIG
535
		__be32 opt[(TCPOLEN_MD5SIG_ALIGNED >> 2)];
536 537
#endif
	} rep;
L
Linus Torvalds 已提交
538
	struct ip_reply_arg arg;
539 540 541
#ifdef CONFIG_TCP_MD5SIG
	struct tcp_md5sig_key *key;
#endif
542
	struct net *net;
L
Linus Torvalds 已提交
543 544 545 546 547

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

548
	if (skb->rtable->rt_type != RTN_LOCAL)
L
Linus Torvalds 已提交
549 550 551
		return;

	/* Swap the send and the receive. */
552 553 554 555 556
	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
557 558

	if (th->ack) {
559
		rep.th.seq = th->ack_seq;
L
Linus Torvalds 已提交
560
	} else {
561 562 563
		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
564 565
	}

566
	memset(&arg, 0, sizeof(arg));
567 568 569 570
	arg.iov[0].iov_base = (unsigned char *)&rep;
	arg.iov[0].iov_len  = sizeof(rep.th);

#ifdef CONFIG_TCP_MD5SIG
571
	key = sk ? tcp_v4_md5_do_lookup(sk, ip_hdr(skb)->daddr) : NULL;
572 573 574 575 576 577 578 579 580
	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;

581
		tcp_v4_md5_hash_hdr((__u8 *) &rep.opt[1],
582 583
				     key, ip_hdr(skb)->saddr,
				     ip_hdr(skb)->daddr, &rep.th);
584 585
	}
#endif
586 587
	arg.csum = csum_tcpudp_nofold(ip_hdr(skb)->daddr,
				      ip_hdr(skb)->saddr, /* XXX */
588
				      arg.iov[0].iov_len, IPPROTO_TCP, 0);
L
Linus Torvalds 已提交
589
	arg.csumoffset = offsetof(struct tcphdr, check) / 2;
590
	arg.flags = (sk && inet_sk(sk)->transparent) ? IP_REPLY_ARG_NOSRCCHECK : 0;
L
Linus Torvalds 已提交
591

592 593
	net = dev_net(skb->dst->dev);
	ip_send_reply(net->ipv4.tcp_sock, skb,
594
		      &arg, arg.iov[0].iov_len);
L
Linus Torvalds 已提交
595

596 597
	TCP_INC_STATS_BH(net, TCP_MIB_OUTSEGS);
	TCP_INC_STATS_BH(net, TCP_MIB_OUTRSTS);
L
Linus Torvalds 已提交
598 599 600 601 602 603
}

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

604 605
static void tcp_v4_send_ack(struct sk_buff *skb, u32 seq, u32 ack,
			    u32 win, u32 ts, int oif,
606 607
			    struct tcp_md5sig_key *key,
			    int reply_flags)
L
Linus Torvalds 已提交
608
{
609
	struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
610 611
	struct {
		struct tcphdr th;
612
		__be32 opt[(TCPOLEN_TSTAMP_ALIGNED >> 2)
613
#ifdef CONFIG_TCP_MD5SIG
614
			   + (TCPOLEN_MD5SIG_ALIGNED >> 2)
615 616
#endif
			];
L
Linus Torvalds 已提交
617 618
	} rep;
	struct ip_reply_arg arg;
619
	struct net *net = dev_net(skb->dst->dev);
L
Linus Torvalds 已提交
620 621

	memset(&rep.th, 0, sizeof(struct tcphdr));
622
	memset(&arg, 0, sizeof(arg));
L
Linus Torvalds 已提交
623 624 625 626

	arg.iov[0].iov_base = (unsigned char *)&rep;
	arg.iov[0].iov_len  = sizeof(rep.th);
	if (ts) {
627 628 629 630 631
		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);
632
		arg.iov[0].iov_len += TCPOLEN_TSTAMP_ALIGNED;
L
Linus Torvalds 已提交
633 634 635 636 637 638 639 640 641 642 643
	}

	/* 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);

644 645 646 647 648 649 650 651 652 653 654
#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;

655
		tcp_v4_md5_hash_hdr((__u8 *) &rep.opt[offset],
656 657
				    key, ip_hdr(skb)->saddr,
				    ip_hdr(skb)->daddr, &rep.th);
658 659
	}
#endif
660
	arg.flags = reply_flags;
661 662
	arg.csum = csum_tcpudp_nofold(ip_hdr(skb)->daddr,
				      ip_hdr(skb)->saddr, /* XXX */
L
Linus Torvalds 已提交
663 664
				      arg.iov[0].iov_len, IPPROTO_TCP, 0);
	arg.csumoffset = offsetof(struct tcphdr, check) / 2;
665 666
	if (oif)
		arg.bound_dev_if = oif;
L
Linus Torvalds 已提交
667

668
	ip_send_reply(net->ipv4.tcp_sock, skb,
669
		      &arg, arg.iov[0].iov_len);
L
Linus Torvalds 已提交
670

671
	TCP_INC_STATS_BH(net, TCP_MIB_OUTSEGS);
L
Linus Torvalds 已提交
672 673 674 675
}

static void tcp_v4_timewait_ack(struct sock *sk, struct sk_buff *skb)
{
676
	struct inet_timewait_sock *tw = inet_twsk(sk);
677
	struct tcp_timewait_sock *tcptw = tcp_twsk(sk);
L
Linus Torvalds 已提交
678

679
	tcp_v4_send_ack(skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt,
680
			tcptw->tw_rcv_wnd >> tw->tw_rcv_wscale,
681 682
			tcptw->tw_ts_recent,
			tw->tw_bound_dev_if,
683 684
			tcp_twsk_md5_key(tcptw),
			tw->tw_transparent ? IP_REPLY_ARG_NOSRCCHECK : 0
685
			);
L
Linus Torvalds 已提交
686

687
	inet_twsk_put(tw);
L
Linus Torvalds 已提交
688 689
}

690
static void tcp_v4_reqsk_send_ack(struct sock *sk, struct sk_buff *skb,
691
				  struct request_sock *req)
L
Linus Torvalds 已提交
692
{
693
	tcp_v4_send_ack(skb, tcp_rsk(req)->snt_isn + 1,
694
			tcp_rsk(req)->rcv_isn + 1, req->rcv_wnd,
695 696
			req->ts_recent,
			0,
697 698
			tcp_v4_md5_do_lookup(sk, ip_hdr(skb)->daddr),
			inet_rsk(req)->no_srccheck ? IP_REPLY_ARG_NOSRCCHECK : 0);
L
Linus Torvalds 已提交
699 700 701
}

/*
702
 *	Send a SYN-ACK after having received a SYN.
703
 *	This still operates on a request_sock only, not on a big
L
Linus Torvalds 已提交
704 705
 *	socket.
 */
706 707
static int __tcp_v4_send_synack(struct sock *sk, struct request_sock *req,
				struct dst_entry *dst)
L
Linus Torvalds 已提交
708
{
709
	const struct inet_request_sock *ireq = inet_rsk(req);
L
Linus Torvalds 已提交
710 711 712 713
	int err = -1;
	struct sk_buff * skb;

	/* First, grab a route. */
714
	if (!dst && (dst = inet_csk_route_req(sk, req)) == NULL)
715
		return -1;
L
Linus Torvalds 已提交
716 717 718 719

	skb = tcp_make_synack(sk, dst, req);

	if (skb) {
720
		struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
721

722
		th->check = tcp_v4_check(skb->len,
723 724
					 ireq->loc_addr,
					 ireq->rmt_addr,
725
					 csum_partial(th, skb->len,
L
Linus Torvalds 已提交
726 727
						      skb->csum));

728 729 730
		err = ip_build_and_send_pkt(skb, sk, ireq->loc_addr,
					    ireq->rmt_addr,
					    ireq->opt);
731
		err = net_xmit_eval(err);
L
Linus Torvalds 已提交
732 733 734 735 736 737
	}

	dst_release(dst);
	return err;
}

738 739 740 741 742
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
743
/*
744
 *	IPv4 request_sock destructor.
L
Linus Torvalds 已提交
745
 */
746
static void tcp_v4_reqsk_destructor(struct request_sock *req)
L
Linus Torvalds 已提交
747
{
J
Jesper Juhl 已提交
748
	kfree(inet_rsk(req)->opt);
L
Linus Torvalds 已提交
749 750
}

751
#ifdef CONFIG_SYN_COOKIES
S
Stephen Hemminger 已提交
752
static void syn_flood_warning(struct sk_buff *skb)
L
Linus Torvalds 已提交
753 754 755 756 757 758 759
{
	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",
760
		       ntohs(tcp_hdr(skb)->dest));
L
Linus Torvalds 已提交
761 762
	}
}
763
#endif
L
Linus Torvalds 已提交
764 765

/*
766
 * Save and compile IPv4 options into the request_sock if needed.
L
Linus Torvalds 已提交
767
 */
S
Stephen Hemminger 已提交
768 769
static struct ip_options *tcp_v4_save_options(struct sock *sk,
					      struct sk_buff *skb)
L
Linus Torvalds 已提交
770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786
{
	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;
}

787 788 789 790 791 792 793 794
#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.  */
795 796
static struct tcp_md5sig_key *
			tcp_v4_md5_do_lookup(struct sock *sk, __be32 addr)
797 798 799 800 801 802 803 804
{
	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)
805
			return &tp->md5sig_info->keys4[i].base;
806 807 808 809 810 811 812 813 814 815 816 817
	}
	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
818 819
static struct tcp_md5sig_key *tcp_v4_reqsk_md5_lookup(struct sock *sk,
						      struct request_sock *req)
820 821 822 823 824 825 826 827 828
{
	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 */
829
	struct tcp_md5sig_key *key;
830 831 832
	struct tcp_sock *tp = tcp_sk(sk);
	struct tcp4_md5sig_key *keys;

833
	key = tcp_v4_md5_do_lookup(sk, addr);
834 835
	if (key) {
		/* Pre-existing entry - just update that one. */
836 837 838
		kfree(key->key);
		key->key = newkey;
		key->keylen = newkeylen;
839
	} else {
840 841
		struct tcp_md5sig_info *md5sig;

842
		if (!tp->md5sig_info) {
843 844
			tp->md5sig_info = kzalloc(sizeof(*tp->md5sig_info),
						  GFP_ATOMIC);
845 846 847 848
			if (!tp->md5sig_info) {
				kfree(newkey);
				return -ENOMEM;
			}
849
			sk->sk_route_caps &= ~NETIF_F_GSO_MASK;
850 851 852 853 854
		}
		if (tcp_alloc_md5sig_pool() == NULL) {
			kfree(newkey);
			return -ENOMEM;
		}
855 856 857 858
		md5sig = tp->md5sig_info;

		if (md5sig->alloced4 == md5sig->entries4) {
			keys = kmalloc((sizeof(*keys) *
859
					(md5sig->entries4 + 1)), GFP_ATOMIC);
860 861 862 863 864 865
			if (!keys) {
				kfree(newkey);
				tcp_free_md5sig_pool();
				return -ENOMEM;
			}

866 867 868
			if (md5sig->entries4)
				memcpy(keys, md5sig->keys4,
				       sizeof(*keys) * md5sig->entries4);
869 870

			/* Free old key list, and reference new one */
871
			kfree(md5sig->keys4);
872 873
			md5sig->keys4 = keys;
			md5sig->alloced4++;
874
		}
875
		md5sig->entries4++;
876 877 878
		md5sig->keys4[md5sig->entries4 - 1].addr        = addr;
		md5sig->keys4[md5sig->entries4 - 1].base.key    = newkey;
		md5sig->keys4[md5sig->entries4 - 1].base.keylen = newkeylen;
879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899
	}
	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 */
900
			kfree(tp->md5sig_info->keys4[i].base.key);
901 902 903 904 905
			tp->md5sig_info->entries4--;

			if (tp->md5sig_info->entries4 == 0) {
				kfree(tp->md5sig_info->keys4);
				tp->md5sig_info->keys4 = NULL;
906
				tp->md5sig_info->alloced4 = 0;
907
			} else if (tp->md5sig_info->entries4 != i) {
908
				/* Need to do some manipulation */
909 910 911 912
				memmove(&tp->md5sig_info->keys4[i],
					&tp->md5sig_info->keys4[i+1],
					(tp->md5sig_info->entries4 - i) *
					 sizeof(struct tcp4_md5sig_key));
913 914 915 916 917 918 919 920 921 922
			}
			tcp_free_md5sig_pool();
			return 0;
		}
	}
	return -ENOENT;
}

EXPORT_SYMBOL(tcp_v4_md5_do_del);

923
static void tcp_v4_clear_md5_list(struct sock *sk)
924 925 926 927 928 929 930 931 932 933
{
	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++)
934
			kfree(tp->md5sig_info->keys4[i].base.key);
935 936 937 938 939 940 941 942 943 944
		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;
	}
}

945 946
static int tcp_v4_parse_md5_keys(struct sock *sk, char __user *optval,
				 int optlen)
947 948 949 950 951 952 953 954
{
	struct tcp_md5sig cmd;
	struct sockaddr_in *sin = (struct sockaddr_in *)&cmd.tcpm_addr;
	u8 *newkey;

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

955
	if (copy_from_user(&cmd, optval, sizeof(cmd)))
956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971
		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);
972
		struct tcp_md5sig_info *p = kzalloc(sizeof(*p), GFP_KERNEL);
973 974 975 976 977

		if (!p)
			return -EINVAL;

		tp->md5sig_info = p;
978
		sk->sk_route_caps &= ~NETIF_F_GSO_MASK;
979 980
	}

981
	newkey = kmemdup(cmd.tcpm_key, cmd.tcpm_keylen, GFP_KERNEL);
982 983 984 985 986 987
	if (!newkey)
		return -ENOMEM;
	return tcp_v4_md5_do_add(sk, sin->sin_addr.s_addr,
				 newkey, cmd.tcpm_keylen);
}

988 989
static int tcp_v4_md5_hash_pseudoheader(struct tcp_md5sig_pool *hp,
					__be32 daddr, __be32 saddr, int nbytes)
990 991
{
	struct tcp4_pseudohdr *bp;
992
	struct scatterlist sg;
993 994 995 996

	bp = &hp->md5_blk.ip4;

	/*
997
	 * 1. the TCP pseudo-header (in the order: source IP address,
998 999 1000 1001 1002 1003
	 * destination IP address, zero-padded protocol number, and
	 * segment length)
	 */
	bp->saddr = saddr;
	bp->daddr = daddr;
	bp->pad = 0;
1004
	bp->protocol = IPPROTO_TCP;
1005
	bp->len = cpu_to_be16(nbytes);
1006

1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030
	sg_init_one(&sg, bp, sizeof(*bp));
	return crypto_hash_update(&hp->md5_desc, &sg, sizeof(*bp));
}

static int tcp_v4_md5_hash_hdr(char *md5_hash, struct tcp_md5sig_key *key,
			       __be32 daddr, __be32 saddr, struct tcphdr *th)
{
	struct tcp_md5sig_pool *hp;
	struct hash_desc *desc;

	hp = tcp_get_md5sig_pool();
	if (!hp)
		goto clear_hash_noput;
	desc = &hp->md5_desc;

	if (crypto_hash_init(desc))
		goto clear_hash;
	if (tcp_v4_md5_hash_pseudoheader(hp, daddr, saddr, th->doff << 2))
		goto clear_hash;
	if (tcp_md5_hash_header(hp, th))
		goto clear_hash;
	if (tcp_md5_hash_key(hp, key))
		goto clear_hash;
	if (crypto_hash_final(desc, md5_hash))
1031 1032 1033 1034
		goto clear_hash;

	tcp_put_md5sig_pool();
	return 0;
1035

1036 1037 1038 1039
clear_hash:
	tcp_put_md5sig_pool();
clear_hash_noput:
	memset(md5_hash, 0, 16);
1040
	return 1;
1041 1042
}

1043 1044 1045
int tcp_v4_md5_hash_skb(char *md5_hash, struct tcp_md5sig_key *key,
			struct sock *sk, struct request_sock *req,
			struct sk_buff *skb)
1046
{
1047 1048 1049
	struct tcp_md5sig_pool *hp;
	struct hash_desc *desc;
	struct tcphdr *th = tcp_hdr(skb);
1050 1051 1052 1053 1054
	__be32 saddr, daddr;

	if (sk) {
		saddr = inet_sk(sk)->saddr;
		daddr = inet_sk(sk)->daddr;
1055 1056 1057
	} else if (req) {
		saddr = inet_rsk(req)->loc_addr;
		daddr = inet_rsk(req)->rmt_addr;
1058
	} else {
1059 1060 1061
		const struct iphdr *iph = ip_hdr(skb);
		saddr = iph->saddr;
		daddr = iph->daddr;
1062
	}
1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090

	hp = tcp_get_md5sig_pool();
	if (!hp)
		goto clear_hash_noput;
	desc = &hp->md5_desc;

	if (crypto_hash_init(desc))
		goto clear_hash;

	if (tcp_v4_md5_hash_pseudoheader(hp, daddr, saddr, skb->len))
		goto clear_hash;
	if (tcp_md5_hash_header(hp, th))
		goto clear_hash;
	if (tcp_md5_hash_skb_data(hp, skb, th->doff << 2))
		goto clear_hash;
	if (tcp_md5_hash_key(hp, key))
		goto clear_hash;
	if (crypto_hash_final(desc, md5_hash))
		goto clear_hash;

	tcp_put_md5sig_pool();
	return 0;

clear_hash:
	tcp_put_md5sig_pool();
clear_hash_noput:
	memset(md5_hash, 0, 16);
	return 1;
1091 1092
}

1093
EXPORT_SYMBOL(tcp_v4_md5_hash_skb);
1094

1095
static int tcp_v4_inbound_md5_hash(struct sock *sk, struct sk_buff *skb)
1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106
{
	/*
	 * 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;
1107
	const struct iphdr *iph = ip_hdr(skb);
1108
	struct tcphdr *th = tcp_hdr(skb);
1109 1110 1111 1112
	int genhash;
	unsigned char newhash[16];

	hash_expected = tcp_v4_md5_do_lookup(sk, iph->saddr);
1113
	hash_location = tcp_parse_md5sig_option(th);
1114 1115 1116 1117 1118 1119

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

	if (hash_expected && !hash_location) {
1120
		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPMD5NOTFOUND);
1121 1122 1123 1124
		return 1;
	}

	if (!hash_expected && hash_location) {
1125
		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPMD5UNEXPECTED);
1126 1127 1128 1129 1130 1131
		return 1;
	}

	/* Okay, so this is hash_expected and hash_location -
	 * so we need to calculate the checksum.
	 */
1132 1133 1134
	genhash = tcp_v4_md5_hash_skb(newhash,
				      hash_expected,
				      NULL, NULL, skb);
1135 1136 1137

	if (genhash || memcmp(hash_location, newhash, 16) != 0) {
		if (net_ratelimit()) {
1138 1139 1140
			printk(KERN_INFO "MD5 Hash failed for (%pI4, %d)->(%pI4, %d)%s\n",
			       &iph->saddr, ntohs(th->source),
			       &iph->daddr, ntohs(th->dest),
1141 1142 1143 1144 1145 1146 1147 1148 1149
			       genhash ? " tcp_v4_calc_md5_hash failed" : "");
		}
		return 1;
	}
	return 0;
}

#endif

1150
struct request_sock_ops tcp_request_sock_ops __read_mostly = {
L
Linus Torvalds 已提交
1151
	.family		=	PF_INET,
1152
	.obj_size	=	sizeof(struct tcp_request_sock),
L
Linus Torvalds 已提交
1153
	.rtx_syn_ack	=	tcp_v4_send_synack,
1154 1155
	.send_ack	=	tcp_v4_reqsk_send_ack,
	.destructor	=	tcp_v4_reqsk_destructor,
L
Linus Torvalds 已提交
1156 1157 1158
	.send_reset	=	tcp_v4_send_reset,
};

1159
#ifdef CONFIG_TCP_MD5SIG
1160
static struct tcp_request_sock_ops tcp_request_sock_ipv4_ops = {
1161 1162
	.md5_lookup	=	tcp_v4_reqsk_md5_lookup,
};
1163
#endif
1164

1165 1166 1167
static struct timewait_sock_ops tcp_timewait_sock_ops = {
	.twsk_obj_size	= sizeof(struct tcp_timewait_sock),
	.twsk_unique	= tcp_twsk_unique,
1168
	.twsk_destructor= tcp_twsk_destructor,
1169 1170
};

L
Linus Torvalds 已提交
1171 1172
int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
{
1173
	struct inet_request_sock *ireq;
L
Linus Torvalds 已提交
1174
	struct tcp_options_received tmp_opt;
1175
	struct request_sock *req;
1176 1177
	__be32 saddr = ip_hdr(skb)->saddr;
	__be32 daddr = ip_hdr(skb)->daddr;
L
Linus Torvalds 已提交
1178 1179 1180 1181 1182 1183 1184 1185 1186
	__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 */
1187
	if (skb->rtable->rt_flags & (RTCF_BROADCAST | RTCF_MULTICAST))
L
Linus Torvalds 已提交
1188 1189 1190 1191 1192 1193
		goto drop;

	/* TW buckets are converted to open requests without
	 * limitations, they conserve resources and peer is
	 * evidently real one.
	 */
1194
	if (inet_csk_reqsk_queue_is_full(sk) && !isn) {
L
Linus Torvalds 已提交
1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207
#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.
	 */
1208
	if (sk_acceptq_is_full(sk) && inet_csk_reqsk_queue_young(sk) > 1)
L
Linus Torvalds 已提交
1209 1210
		goto drop;

1211
	req = inet_reqsk_alloc(&tcp_request_sock_ops);
L
Linus Torvalds 已提交
1212 1213 1214
	if (!req)
		goto drop;

1215 1216 1217 1218
#ifdef CONFIG_TCP_MD5SIG
	tcp_rsk(req)->af_specific = &tcp_request_sock_ipv4_ops;
#endif

L
Linus Torvalds 已提交
1219 1220 1221 1222 1223 1224
	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);

1225
	if (want_cookie && !tmp_opt.saw_tstamp)
L
Linus Torvalds 已提交
1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240
		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);

1241 1242 1243
	if (security_inet_conn_request(sk, skb, req))
		goto drop_and_free;

1244 1245 1246
	ireq = inet_rsk(req);
	ireq->loc_addr = daddr;
	ireq->rmt_addr = saddr;
1247
	ireq->no_srccheck = inet_sk(sk)->transparent;
1248
	ireq->opt = tcp_v4_save_options(sk, skb);
L
Linus Torvalds 已提交
1249
	if (!want_cookie)
1250
		TCP_ECN_create_request(req, tcp_hdr(skb));
L
Linus Torvalds 已提交
1251 1252 1253 1254

	if (want_cookie) {
#ifdef CONFIG_SYN_COOKIES
		syn_flood_warning(skb);
1255
		req->cookie_ts = tmp_opt.tstamp_ok;
L
Linus Torvalds 已提交
1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270
#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 &&
1271
		    tcp_death_row.sysctl_tw_recycle &&
1272
		    (dst = inet_csk_route_req(sk, req)) != NULL &&
L
Linus Torvalds 已提交
1273 1274
		    (peer = rt_get_peer((struct rtable *)dst)) != NULL &&
		    peer->v4daddr == saddr) {
1275
			if (get_seconds() < peer->tcp_ts_stamp + TCP_PAWS_MSL &&
L
Linus Torvalds 已提交
1276 1277
			    (s32)(peer->tcp_ts - req->ts_recent) >
							TCP_PAWS_WINDOW) {
1278
				NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_PAWSPASSIVEREJECTED);
1279
				goto drop_and_release;
L
Linus Torvalds 已提交
1280 1281 1282 1283
			}
		}
		/* Kill the following clause, if you dislike this way. */
		else if (!sysctl_tcp_syncookies &&
1284
			 (sysctl_max_syn_backlog - inet_csk_reqsk_queue_len(sk) <
L
Linus Torvalds 已提交
1285 1286 1287 1288 1289 1290 1291 1292 1293 1294
			  (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.
			 */
1295 1296
			LIMIT_NETDEBUG(KERN_DEBUG "TCP: drop open request from %pI4/%u\n",
				       &saddr, ntohs(tcp_hdr(skb)->source));
1297
			goto drop_and_release;
L
Linus Torvalds 已提交
1298 1299
		}

1300
		isn = tcp_v4_init_sequence(skb);
L
Linus Torvalds 已提交
1301
	}
1302
	tcp_rsk(req)->snt_isn = isn;
L
Linus Torvalds 已提交
1303

1304
	if (__tcp_v4_send_synack(sk, req, dst) || want_cookie)
L
Linus Torvalds 已提交
1305 1306
		goto drop_and_free;

1307
	inet_csk_reqsk_queue_hash_add(sk, req, TCP_TIMEOUT_INIT);
L
Linus Torvalds 已提交
1308 1309
	return 0;

1310 1311
drop_and_release:
	dst_release(dst);
L
Linus Torvalds 已提交
1312
drop_and_free:
1313
	reqsk_free(req);
L
Linus Torvalds 已提交
1314 1315 1316 1317 1318 1319 1320 1321 1322 1323
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,
1324
				  struct request_sock *req,
L
Linus Torvalds 已提交
1325 1326
				  struct dst_entry *dst)
{
1327
	struct inet_request_sock *ireq;
L
Linus Torvalds 已提交
1328 1329 1330
	struct inet_sock *newinet;
	struct tcp_sock *newtp;
	struct sock *newsk;
1331 1332 1333
#ifdef CONFIG_TCP_MD5SIG
	struct tcp_md5sig_key *key;
#endif
L
Linus Torvalds 已提交
1334 1335 1336 1337

	if (sk_acceptq_is_full(sk))
		goto exit_overflow;

1338
	if (!dst && (dst = inet_csk_route_req(sk, req)) == NULL)
L
Linus Torvalds 已提交
1339 1340 1341 1342 1343 1344
		goto exit;

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

1345
	newsk->sk_gso_type = SKB_GSO_TCPV4;
1346
	sk_setup_caps(newsk, dst);
L
Linus Torvalds 已提交
1347 1348 1349

	newtp		      = tcp_sk(newsk);
	newinet		      = inet_sk(newsk);
1350 1351 1352 1353 1354 1355
	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;
1356
	newinet->mc_index     = inet_iif(skb);
1357
	newinet->mc_ttl	      = ip_hdr(skb)->ttl;
1358
	inet_csk(newsk)->icsk_ext_hdr_len = 0;
L
Linus Torvalds 已提交
1359
	if (newinet->opt)
1360
		inet_csk(newsk)->icsk_ext_hdr_len = newinet->opt->optlen;
L
Linus Torvalds 已提交
1361 1362
	newinet->id = newtp->write_seq ^ jiffies;

J
John Heffner 已提交
1363
	tcp_mtup_init(newsk);
L
Linus Torvalds 已提交
1364 1365
	tcp_sync_mss(newsk, dst_mtu(dst));
	newtp->advmss = dst_metric(dst, RTAX_ADVMSS);
1366 1367 1368 1369
	if (tcp_sk(sk)->rx_opt.user_mss &&
	    tcp_sk(sk)->rx_opt.user_mss < newtp->advmss)
		newtp->advmss = tcp_sk(sk)->rx_opt.user_mss;

L
Linus Torvalds 已提交
1370 1371
	tcp_initialize_rcv_mss(newsk);

1372 1373 1374 1375 1376 1377 1378 1379 1380
#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.
		 */
1381 1382
		char *newkey = kmemdup(key->key, key->keylen, GFP_ATOMIC);
		if (newkey != NULL)
1383 1384
			tcp_v4_md5_do_add(newsk, inet_sk(sk)->daddr,
					  newkey, key->keylen);
1385
		newsk->sk_route_caps &= ~NETIF_F_GSO_MASK;
1386 1387 1388
	}
#endif

1389 1390
	__inet_hash_nolisten(newsk);
	__inet_inherit_port(sk, newsk);
L
Linus Torvalds 已提交
1391 1392 1393 1394

	return newsk;

exit_overflow:
1395
	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENOVERFLOWS);
L
Linus Torvalds 已提交
1396
exit:
1397
	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS);
L
Linus Torvalds 已提交
1398 1399 1400 1401 1402 1403
	dst_release(dst);
	return NULL;
}

static struct sock *tcp_v4_hnd_req(struct sock *sk, struct sk_buff *skb)
{
1404
	struct tcphdr *th = tcp_hdr(skb);
1405
	const struct iphdr *iph = ip_hdr(skb);
L
Linus Torvalds 已提交
1406
	struct sock *nsk;
1407
	struct request_sock **prev;
L
Linus Torvalds 已提交
1408
	/* Find possible connection requests. */
1409 1410
	struct request_sock *req = inet_csk_search_req(sk, &prev, th->source,
						       iph->saddr, iph->daddr);
L
Linus Torvalds 已提交
1411 1412 1413
	if (req)
		return tcp_check_req(sk, skb, req, prev);

1414
	nsk = inet_lookup_established(sock_net(sk), &tcp_hashinfo, iph->saddr,
1415
			th->source, iph->daddr, th->dest, inet_iif(skb));
L
Linus Torvalds 已提交
1416 1417 1418 1419 1420 1421

	if (nsk) {
		if (nsk->sk_state != TCP_TIME_WAIT) {
			bh_lock_sock(nsk);
			return nsk;
		}
1422
		inet_twsk_put(inet_twsk(nsk));
L
Linus Torvalds 已提交
1423 1424 1425 1426 1427 1428 1429 1430 1431 1432
		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;
}

1433
static __sum16 tcp_v4_checksum_init(struct sk_buff *skb)
L
Linus Torvalds 已提交
1434
{
1435 1436
	const struct iphdr *iph = ip_hdr(skb);

1437
	if (skb->ip_summed == CHECKSUM_COMPLETE) {
1438 1439
		if (!tcp_v4_check(skb->len, iph->saddr,
				  iph->daddr, skb->csum)) {
1440
			skb->ip_summed = CHECKSUM_UNNECESSARY;
L
Linus Torvalds 已提交
1441
			return 0;
1442
		}
L
Linus Torvalds 已提交
1443
	}
1444

1445
	skb->csum = csum_tcpudp_nofold(iph->saddr, iph->daddr,
1446 1447
				       skb->len, IPPROTO_TCP, 0);

L
Linus Torvalds 已提交
1448
	if (skb->len <= 76) {
1449
		return __skb_checksum_complete(skb);
L
Linus Torvalds 已提交
1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464
	}
	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)
{
1465 1466 1467 1468 1469 1470 1471 1472
	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
	 */
1473
	if (tcp_v4_inbound_md5_hash(sk, skb))
1474 1475 1476
		goto discard;
#endif

L
Linus Torvalds 已提交
1477 1478
	if (sk->sk_state == TCP_ESTABLISHED) { /* Fast path */
		TCP_CHECK_TIMER(sk);
1479
		if (tcp_rcv_established(sk, skb, tcp_hdr(skb), skb->len)) {
1480
			rsk = sk;
L
Linus Torvalds 已提交
1481
			goto reset;
1482
		}
L
Linus Torvalds 已提交
1483 1484 1485 1486
		TCP_CHECK_TIMER(sk);
		return 0;
	}

1487
	if (skb->len < tcp_hdrlen(skb) || tcp_checksum_complete(skb))
L
Linus Torvalds 已提交
1488 1489 1490 1491 1492 1493 1494 1495
		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) {
1496 1497
			if (tcp_child_process(sk, nsk, skb)) {
				rsk = nsk;
L
Linus Torvalds 已提交
1498
				goto reset;
1499
			}
L
Linus Torvalds 已提交
1500 1501 1502 1503 1504
			return 0;
		}
	}

	TCP_CHECK_TIMER(sk);
1505
	if (tcp_rcv_state_process(sk, skb, tcp_hdr(skb), skb->len)) {
1506
		rsk = sk;
L
Linus Torvalds 已提交
1507
		goto reset;
1508
	}
L
Linus Torvalds 已提交
1509 1510 1511 1512
	TCP_CHECK_TIMER(sk);
	return 0;

reset:
1513
	tcp_v4_send_reset(rsk, skb);
L
Linus Torvalds 已提交
1514 1515 1516 1517 1518 1519 1520 1521 1522 1523
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:
1524
	TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_INERRS);
L
Linus Torvalds 已提交
1525 1526 1527 1528 1529 1530 1531 1532 1533
	goto discard;
}

/*
 *	From tcp_input.c
 */

int tcp_v4_rcv(struct sk_buff *skb)
{
1534
	const struct iphdr *iph;
L
Linus Torvalds 已提交
1535 1536 1537
	struct tcphdr *th;
	struct sock *sk;
	int ret;
1538
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
1539 1540 1541 1542 1543

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

	/* Count it even if it's bad */
1544
	TCP_INC_STATS_BH(net, TCP_MIB_INSEGS);
L
Linus Torvalds 已提交
1545 1546 1547 1548

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

1549
	th = tcp_hdr(skb);
L
Linus Torvalds 已提交
1550 1551 1552 1553 1554 1555 1556 1557

	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
1558
	 * provided case of th->doff==0 is eliminated.
L
Linus Torvalds 已提交
1559
	 * So, we defer the checks. */
1560
	if (!skb_csum_unnecessary(skb) && tcp_v4_checksum_init(skb))
L
Linus Torvalds 已提交
1561 1562
		goto bad_packet;

1563
	th = tcp_hdr(skb);
1564
	iph = ip_hdr(skb);
L
Linus Torvalds 已提交
1565 1566 1567 1568 1569
	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;
1570
	TCP_SKB_CB(skb)->flags	 = iph->tos;
L
Linus Torvalds 已提交
1571 1572
	TCP_SKB_CB(skb)->sacked	 = 0;

1573
	sk = __inet_lookup_skb(&tcp_hashinfo, skb, th->source, th->dest);
L
Linus Torvalds 已提交
1574 1575 1576 1577 1578 1579 1580 1581 1582
	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;
1583
	nf_reset(skb);
L
Linus Torvalds 已提交
1584

1585
	if (sk_filter(sk, skb))
L
Linus Torvalds 已提交
1586 1587 1588 1589
		goto discard_and_relse;

	skb->dev = NULL;

1590
	bh_lock_sock_nested(sk);
L
Linus Torvalds 已提交
1591 1592
	ret = 0;
	if (!sock_owned_by_user(sk)) {
C
Chris Leech 已提交
1593 1594 1595 1596 1597
#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
1598
			ret = tcp_v4_do_rcv(sk, skb);
C
Chris Leech 已提交
1599 1600 1601 1602 1603 1604
		else
#endif
		{
			if (!tcp_prequeue(sk, skb))
			ret = tcp_v4_do_rcv(sk, skb);
		}
L
Linus Torvalds 已提交
1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618
	} 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:
1619
		TCP_INC_STATS_BH(net, TCP_MIB_INERRS);
L
Linus Torvalds 已提交
1620
	} else {
1621
		tcp_v4_send_reset(NULL, skb);
L
Linus Torvalds 已提交
1622 1623 1624 1625 1626
	}

discard_it:
	/* Discard frame. */
	kfree_skb(skb);
1627
	return 0;
L
Linus Torvalds 已提交
1628 1629 1630 1631 1632 1633 1634

discard_and_relse:
	sock_put(sk);
	goto discard_it;

do_time_wait:
	if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) {
1635
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1636 1637 1638 1639
		goto discard_it;
	}

	if (skb->len < (th->doff << 2) || tcp_checksum_complete(skb)) {
1640
		TCP_INC_STATS_BH(net, TCP_MIB_INERRS);
1641
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1642 1643
		goto discard_it;
	}
1644
	switch (tcp_timewait_state_process(inet_twsk(sk), skb, th)) {
L
Linus Torvalds 已提交
1645
	case TCP_TW_SYN: {
1646
		struct sock *sk2 = inet_lookup_listener(dev_net(skb->dev),
1647
							&tcp_hashinfo,
1648
							iph->daddr, th->dest,
1649
							inet_iif(skb));
L
Linus Torvalds 已提交
1650
		if (sk2) {
1651 1652
			inet_twsk_deschedule(inet_twsk(sk), &tcp_death_row);
			inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692
			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 ||
1693
		    (peer->tcp_ts_stamp + TCP_PAWS_MSL < get_seconds() &&
L
Linus Torvalds 已提交
1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705
		     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;
}

1706
int tcp_v4_tw_remember_stamp(struct inet_timewait_sock *tw)
L
Linus Torvalds 已提交
1707
{
1708
	struct inet_peer *peer = inet_getpeer(tw->tw_daddr, 1);
L
Linus Torvalds 已提交
1709 1710

	if (peer) {
1711 1712 1713
		const struct tcp_timewait_sock *tcptw = tcp_twsk((struct sock *)tw);

		if ((s32)(peer->tcp_ts - tcptw->tw_ts_recent) <= 0 ||
1714
		    (peer->tcp_ts_stamp + TCP_PAWS_MSL < get_seconds() &&
1715 1716 1717
		     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
1718 1719 1720 1721 1722 1723 1724 1725
		}
		inet_putpeer(peer);
		return 1;
	}

	return 0;
}

1726
struct inet_connection_sock_af_ops ipv4_specific = {
1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737
	.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),
1738
	.bind_conflict	   = inet_csk_bind_conflict,
1739
#ifdef CONFIG_COMPAT
1740 1741
	.compat_setsockopt = compat_ip_setsockopt,
	.compat_getsockopt = compat_ip_getsockopt,
1742
#endif
L
Linus Torvalds 已提交
1743 1744
};

1745
#ifdef CONFIG_TCP_MD5SIG
1746
static struct tcp_sock_af_ops tcp_sock_ipv4_specific = {
1747
	.md5_lookup		= tcp_v4_md5_lookup,
1748
	.calc_md5_hash		= tcp_v4_md5_hash_skb,
1749 1750 1751
	.md5_add		= tcp_v4_md5_add_func,
	.md5_parse		= tcp_v4_parse_md5_keys,
};
1752
#endif
1753

L
Linus Torvalds 已提交
1754 1755 1756 1757 1758
/* 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)
{
1759
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
1760 1761 1762 1763 1764 1765
	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);

1766
	icsk->icsk_rto = TCP_TIMEOUT_INIT;
L
Linus Torvalds 已提交
1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780
	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;
1781
	tp->mss_cache = 536;
L
Linus Torvalds 已提交
1782 1783

	tp->reordering = sysctl_tcp_reordering;
1784
	icsk->icsk_ca_ops = &tcp_init_congestion_ops;
L
Linus Torvalds 已提交
1785 1786 1787 1788 1789 1790

	sk->sk_state = TCP_CLOSE;

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

1791
	icsk->icsk_af_ops = &ipv4_specific;
1792
	icsk->icsk_sync_mss = tcp_sync_mss;
1793 1794 1795
#ifdef CONFIG_TCP_MD5SIG
	tp->af_specific = &tcp_sock_ipv4_specific;
#endif
L
Linus Torvalds 已提交
1796 1797 1798 1799 1800 1801 1802 1803 1804

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

	atomic_inc(&tcp_sockets_allocated);

	return 0;
}

1805
void tcp_v4_destroy_sock(struct sock *sk)
L
Linus Torvalds 已提交
1806 1807 1808 1809 1810
{
	struct tcp_sock *tp = tcp_sk(sk);

	tcp_clear_xmit_timers(sk);

1811
	tcp_cleanup_congestion_control(sk);
1812

L
Linus Torvalds 已提交
1813
	/* Cleanup up the write buffer. */
1814
	tcp_write_queue_purge(sk);
L
Linus Torvalds 已提交
1815 1816

	/* Cleans up our, hopefully empty, out_of_order_queue. */
1817
	__skb_queue_purge(&tp->out_of_order_queue);
L
Linus Torvalds 已提交
1818

1819 1820 1821 1822 1823 1824 1825 1826 1827
#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
1828 1829
#ifdef CONFIG_NET_DMA
	/* Cleans up our sk_async_wait_queue */
1830
	__skb_queue_purge(&sk->sk_async_wait_queue);
C
Chris Leech 已提交
1831 1832
#endif

L
Linus Torvalds 已提交
1833 1834 1835 1836
	/* Clean prequeue, it must be empty really */
	__skb_queue_purge(&tp->ucopy.prequeue);

	/* Clean up a referenced TCP bind bucket. */
1837
	if (inet_csk(sk)->icsk_bind_hash)
1838
		inet_put_port(sk);
L
Linus Torvalds 已提交
1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855

	/*
	 * 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. */

1856
static inline struct inet_timewait_sock *tw_head(struct hlist_nulls_head *head)
L
Linus Torvalds 已提交
1857
{
1858
	return hlist_nulls_empty(head) ? NULL :
1859
		list_entry(head->first, struct inet_timewait_sock, tw_node);
L
Linus Torvalds 已提交
1860 1861
}

1862
static inline struct inet_timewait_sock *tw_next(struct inet_timewait_sock *tw)
L
Linus Torvalds 已提交
1863
{
1864 1865
	return !is_a_nulls(tw->tw_node.next) ?
		hlist_nulls_entry(tw->tw_node.next, typeof(*tw), tw_node) : NULL;
L
Linus Torvalds 已提交
1866 1867 1868 1869
}

static void *listening_get_next(struct seq_file *seq, void *cur)
{
1870
	struct inet_connection_sock *icsk;
L
Linus Torvalds 已提交
1871 1872
	struct hlist_node *node;
	struct sock *sk = cur;
1873
	struct inet_listen_hashbucket *ilb;
J
Jianjun Kong 已提交
1874
	struct tcp_iter_state *st = seq->private;
1875
	struct net *net = seq_file_net(seq);
L
Linus Torvalds 已提交
1876 1877 1878

	if (!sk) {
		st->bucket = 0;
1879 1880 1881
		ilb = &tcp_hashinfo.listening_hash[0];
		spin_lock_bh(&ilb->lock);
		sk = sk_head(&ilb->head);
L
Linus Torvalds 已提交
1882 1883
		goto get_sk;
	}
1884
	ilb = &tcp_hashinfo.listening_hash[st->bucket];
L
Linus Torvalds 已提交
1885 1886 1887
	++st->num;

	if (st->state == TCP_SEQ_STATE_OPENREQ) {
1888
		struct request_sock *req = cur;
L
Linus Torvalds 已提交
1889

1890
		icsk = inet_csk(st->syn_wait_sk);
L
Linus Torvalds 已提交
1891 1892 1893
		req = req->dl_next;
		while (1) {
			while (req) {
1894
				if (req->rsk_ops->family == st->family) {
L
Linus Torvalds 已提交
1895 1896 1897 1898 1899
					cur = req;
					goto out;
				}
				req = req->dl_next;
			}
1900
			if (++st->sbucket >= icsk->icsk_accept_queue.listen_opt->nr_table_entries)
L
Linus Torvalds 已提交
1901 1902
				break;
get_req:
1903
			req = icsk->icsk_accept_queue.listen_opt->syn_table[st->sbucket];
L
Linus Torvalds 已提交
1904 1905 1906
		}
		sk	  = sk_next(st->syn_wait_sk);
		st->state = TCP_SEQ_STATE_LISTENING;
1907
		read_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
L
Linus Torvalds 已提交
1908
	} else {
1909
		icsk = inet_csk(sk);
1910 1911
		read_lock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
		if (reqsk_queue_len(&icsk->icsk_accept_queue))
L
Linus Torvalds 已提交
1912
			goto start_req;
1913
		read_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
L
Linus Torvalds 已提交
1914 1915 1916 1917
		sk = sk_next(sk);
	}
get_sk:
	sk_for_each_from(sk, node) {
1918
		if (sk->sk_family == st->family && net_eq(sock_net(sk), net)) {
L
Linus Torvalds 已提交
1919 1920 1921
			cur = sk;
			goto out;
		}
1922
		icsk = inet_csk(sk);
1923 1924
		read_lock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
		if (reqsk_queue_len(&icsk->icsk_accept_queue)) {
L
Linus Torvalds 已提交
1925 1926 1927 1928 1929 1930 1931
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;
		}
1932
		read_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
L
Linus Torvalds 已提交
1933
	}
1934
	spin_unlock_bh(&ilb->lock);
1935
	if (++st->bucket < INET_LHTABLE_SIZE) {
1936 1937 1938
		ilb = &tcp_hashinfo.listening_hash[st->bucket];
		spin_lock_bh(&ilb->lock);
		sk = sk_head(&ilb->head);
L
Linus Torvalds 已提交
1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956
		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;
}

1957 1958
static inline int empty_bucket(struct tcp_iter_state *st)
{
1959 1960
	return hlist_nulls_empty(&tcp_hashinfo.ehash[st->bucket].chain) &&
		hlist_nulls_empty(&tcp_hashinfo.ehash[st->bucket].twchain);
1961 1962
}

L
Linus Torvalds 已提交
1963 1964
static void *established_get_first(struct seq_file *seq)
{
J
Jianjun Kong 已提交
1965
	struct tcp_iter_state *st = seq->private;
1966
	struct net *net = seq_file_net(seq);
L
Linus Torvalds 已提交
1967 1968
	void *rc = NULL;

1969
	for (st->bucket = 0; st->bucket < tcp_hashinfo.ehash_size; ++st->bucket) {
L
Linus Torvalds 已提交
1970
		struct sock *sk;
1971
		struct hlist_nulls_node *node;
1972
		struct inet_timewait_sock *tw;
1973
		spinlock_t *lock = inet_ehash_lockp(&tcp_hashinfo, st->bucket);
L
Linus Torvalds 已提交
1974

1975 1976 1977 1978
		/* Lockless fast path for the common case of empty buckets */
		if (empty_bucket(st))
			continue;

1979
		spin_lock_bh(lock);
1980
		sk_nulls_for_each(sk, node, &tcp_hashinfo.ehash[st->bucket].chain) {
1981
			if (sk->sk_family != st->family ||
1982
			    !net_eq(sock_net(sk), net)) {
L
Linus Torvalds 已提交
1983 1984 1985 1986 1987 1988
				continue;
			}
			rc = sk;
			goto out;
		}
		st->state = TCP_SEQ_STATE_TIME_WAIT;
1989
		inet_twsk_for_each(tw, node,
1990
				   &tcp_hashinfo.ehash[st->bucket].twchain) {
1991
			if (tw->tw_family != st->family ||
1992
			    !net_eq(twsk_net(tw), net)) {
L
Linus Torvalds 已提交
1993 1994 1995 1996 1997
				continue;
			}
			rc = tw;
			goto out;
		}
1998
		spin_unlock_bh(lock);
L
Linus Torvalds 已提交
1999 2000 2001 2002 2003 2004 2005 2006 2007
		st->state = TCP_SEQ_STATE_ESTABLISHED;
	}
out:
	return rc;
}

static void *established_get_next(struct seq_file *seq, void *cur)
{
	struct sock *sk = cur;
2008
	struct inet_timewait_sock *tw;
2009
	struct hlist_nulls_node *node;
J
Jianjun Kong 已提交
2010
	struct tcp_iter_state *st = seq->private;
2011
	struct net *net = seq_file_net(seq);
L
Linus Torvalds 已提交
2012 2013 2014 2015 2016 2017 2018

	++st->num;

	if (st->state == TCP_SEQ_STATE_TIME_WAIT) {
		tw = cur;
		tw = tw_next(tw);
get_tw:
2019
		while (tw && (tw->tw_family != st->family || !net_eq(twsk_net(tw), net))) {
L
Linus Torvalds 已提交
2020 2021 2022 2023 2024 2025
			tw = tw_next(tw);
		}
		if (tw) {
			cur = tw;
			goto out;
		}
2026
		spin_unlock_bh(inet_ehash_lockp(&tcp_hashinfo, st->bucket));
L
Linus Torvalds 已提交
2027 2028
		st->state = TCP_SEQ_STATE_ESTABLISHED;

2029 2030 2031 2032 2033 2034 2035
		/* Look for next non empty bucket */
		while (++st->bucket < tcp_hashinfo.ehash_size &&
				empty_bucket(st))
			;
		if (st->bucket >= tcp_hashinfo.ehash_size)
			return NULL;

2036
		spin_lock_bh(inet_ehash_lockp(&tcp_hashinfo, st->bucket));
2037
		sk = sk_nulls_head(&tcp_hashinfo.ehash[st->bucket].chain);
L
Linus Torvalds 已提交
2038
	} else
2039
		sk = sk_nulls_next(sk);
L
Linus Torvalds 已提交
2040

2041
	sk_nulls_for_each_from(sk, node) {
2042
		if (sk->sk_family == st->family && net_eq(sock_net(sk), net))
L
Linus Torvalds 已提交
2043 2044 2045 2046
			goto found;
	}

	st->state = TCP_SEQ_STATE_TIME_WAIT;
2047
	tw = tw_head(&tcp_hashinfo.ehash[st->bucket].twchain);
L
Linus Torvalds 已提交
2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061
	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;
2062
	}
L
Linus Torvalds 已提交
2063 2064 2065 2066 2067 2068
	return rc;
}

static void *tcp_get_idx(struct seq_file *seq, loff_t pos)
{
	void *rc;
J
Jianjun Kong 已提交
2069
	struct tcp_iter_state *st = seq->private;
L
Linus Torvalds 已提交
2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083

	st->state = TCP_SEQ_STATE_LISTENING;
	rc	  = listening_get_idx(seq, &pos);

	if (!rc) {
		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)
{
J
Jianjun Kong 已提交
2084
	struct tcp_iter_state *st = seq->private;
L
Linus Torvalds 已提交
2085 2086 2087 2088 2089 2090 2091 2092
	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;
J
Jianjun Kong 已提交
2093
	struct tcp_iter_state *st;
L
Linus Torvalds 已提交
2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121

	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) {
			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)
{
J
Jianjun Kong 已提交
2122
	struct tcp_iter_state *st = seq->private;
L
Linus Torvalds 已提交
2123 2124 2125 2126

	switch (st->state) {
	case TCP_SEQ_STATE_OPENREQ:
		if (v) {
2127 2128
			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
2129 2130 2131
		}
	case TCP_SEQ_STATE_LISTENING:
		if (v != SEQ_START_TOKEN)
2132
			spin_unlock_bh(&tcp_hashinfo.listening_hash[st->bucket].lock);
L
Linus Torvalds 已提交
2133 2134 2135 2136
		break;
	case TCP_SEQ_STATE_TIME_WAIT:
	case TCP_SEQ_STATE_ESTABLISHED:
		if (v)
2137
			spin_unlock_bh(inet_ehash_lockp(&tcp_hashinfo, st->bucket));
L
Linus Torvalds 已提交
2138 2139 2140 2141 2142 2143 2144 2145
		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;
2146
	int err;
L
Linus Torvalds 已提交
2147

2148 2149 2150 2151
	err = seq_open_net(inode, file, &afinfo->seq_ops,
			  sizeof(struct tcp_iter_state));
	if (err < 0)
		return err;
2152

2153
	s = ((struct seq_file *)file->private_data)->private;
L
Linus Torvalds 已提交
2154
	s->family		= afinfo->family;
2155 2156 2157
	return 0;
}

2158
int tcp_proc_register(struct net *net, struct tcp_seq_afinfo *afinfo)
L
Linus Torvalds 已提交
2159 2160 2161 2162
{
	int rc = 0;
	struct proc_dir_entry *p;

2163 2164 2165 2166
	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;
2167

2168 2169 2170 2171
	afinfo->seq_ops.start		= tcp_seq_start;
	afinfo->seq_ops.next		= tcp_seq_next;
	afinfo->seq_ops.stop		= tcp_seq_stop;

2172 2173 2174
	p = proc_create_data(afinfo->name, S_IRUGO, net->proc_net,
			     &afinfo->seq_fops, afinfo);
	if (!p)
L
Linus Torvalds 已提交
2175 2176 2177 2178
		rc = -ENOMEM;
	return rc;
}

2179
void tcp_proc_unregister(struct net *net, struct tcp_seq_afinfo *afinfo)
L
Linus Torvalds 已提交
2180
{
2181
	proc_net_remove(net, afinfo->name);
L
Linus Torvalds 已提交
2182 2183
}

2184
static void get_openreq4(struct sock *sk, struct request_sock *req,
2185
			 struct seq_file *f, int i, int uid, int *len)
L
Linus Torvalds 已提交
2186
{
2187
	const struct inet_request_sock *ireq = inet_rsk(req);
L
Linus Torvalds 已提交
2188 2189
	int ttd = req->expires - jiffies;

2190 2191
	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
2192
		i,
2193
		ireq->loc_addr,
L
Linus Torvalds 已提交
2194
		ntohs(inet_sk(sk)->sport),
2195 2196
		ireq->rmt_addr,
		ntohs(ireq->rmt_port),
L
Linus Torvalds 已提交
2197 2198 2199 2200 2201 2202 2203 2204 2205
		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),
2206 2207
		req,
		len);
L
Linus Torvalds 已提交
2208 2209
}

2210
static void get_tcp4_sock(struct sock *sk, struct seq_file *f, int i, int *len)
L
Linus Torvalds 已提交
2211 2212 2213
{
	int timer_active;
	unsigned long timer_expires;
2214 2215 2216
	struct tcp_sock *tp = tcp_sk(sk);
	const struct inet_connection_sock *icsk = inet_csk(sk);
	struct inet_sock *inet = inet_sk(sk);
2217 2218
	__be32 dest = inet->daddr;
	__be32 src = inet->rcv_saddr;
L
Linus Torvalds 已提交
2219 2220 2221
	__u16 destp = ntohs(inet->dport);
	__u16 srcp = ntohs(inet->sport);

2222
	if (icsk->icsk_pending == ICSK_TIME_RETRANS) {
L
Linus Torvalds 已提交
2223
		timer_active	= 1;
2224 2225
		timer_expires	= icsk->icsk_timeout;
	} else if (icsk->icsk_pending == ICSK_TIME_PROBE0) {
L
Linus Torvalds 已提交
2226
		timer_active	= 4;
2227
		timer_expires	= icsk->icsk_timeout;
2228
	} else if (timer_pending(&sk->sk_timer)) {
L
Linus Torvalds 已提交
2229
		timer_active	= 2;
2230
		timer_expires	= sk->sk_timer.expires;
L
Linus Torvalds 已提交
2231 2232 2233 2234 2235
	} else {
		timer_active	= 0;
		timer_expires = jiffies;
	}

2236
	seq_printf(f, "%4d: %08X:%04X %08X:%04X %02X %08X:%08X %02X:%08lX "
2237
			"%08X %5d %8d %lu %d %p %lu %lu %u %u %d%n",
2238
		i, src, srcp, dest, destp, sk->sk_state,
2239
		tp->write_seq - tp->snd_una,
2240
		sk->sk_state == TCP_LISTEN ? sk->sk_ack_backlog :
2241
					     (tp->rcv_nxt - tp->copied_seq),
L
Linus Torvalds 已提交
2242 2243
		timer_active,
		jiffies_to_clock_t(timer_expires - jiffies),
2244
		icsk->icsk_retransmits,
2245
		sock_i_uid(sk),
2246
		icsk->icsk_probes_out,
2247 2248
		sock_i_ino(sk),
		atomic_read(&sk->sk_refcnt), sk,
2249 2250
		jiffies_to_clock_t(icsk->icsk_rto),
		jiffies_to_clock_t(icsk->icsk_ack.ato),
2251
		(icsk->icsk_ack.quick << 1) | icsk->icsk_ack.pingpong,
L
Linus Torvalds 已提交
2252
		tp->snd_cwnd,
2253 2254
		tp->snd_ssthresh >= 0xFFFF ? -1 : tp->snd_ssthresh,
		len);
L
Linus Torvalds 已提交
2255 2256
}

2257
static void get_timewait4_sock(struct inet_timewait_sock *tw,
2258
			       struct seq_file *f, int i, int *len)
L
Linus Torvalds 已提交
2259
{
2260
	__be32 dest, src;
L
Linus Torvalds 已提交
2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271
	__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);

2272 2273
	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
2274 2275
		i, src, srcp, dest, destp, tw->tw_substate, 0, 0,
		3, jiffies_to_clock_t(ttd), 0, 0, 0, 0,
2276
		atomic_read(&tw->tw_refcnt), tw, len);
L
Linus Torvalds 已提交
2277 2278 2279 2280 2281 2282
}

#define TMPSZ 150

static int tcp4_seq_show(struct seq_file *seq, void *v)
{
J
Jianjun Kong 已提交
2283
	struct tcp_iter_state *st;
2284
	int len;
L
Linus Torvalds 已提交
2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297

	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:
2298
		get_tcp4_sock(v, seq, st->num, &len);
L
Linus Torvalds 已提交
2299 2300
		break;
	case TCP_SEQ_STATE_OPENREQ:
2301
		get_openreq4(st->syn_wait_sk, v, seq, st->num, st->uid, &len);
L
Linus Torvalds 已提交
2302 2303
		break;
	case TCP_SEQ_STATE_TIME_WAIT:
2304
		get_timewait4_sock(v, seq, st->num, &len);
L
Linus Torvalds 已提交
2305 2306
		break;
	}
2307
	seq_printf(seq, "%*s\n", TMPSZ - 1 - len, "");
L
Linus Torvalds 已提交
2308 2309 2310 2311 2312 2313 2314
out:
	return 0;
}

static struct tcp_seq_afinfo tcp4_seq_afinfo = {
	.name		= "tcp",
	.family		= AF_INET,
2315 2316 2317
	.seq_fops	= {
		.owner		= THIS_MODULE,
	},
2318 2319 2320
	.seq_ops	= {
		.show		= tcp4_seq_show,
	},
L
Linus Torvalds 已提交
2321 2322
};

2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337
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
2338 2339
int __init tcp4_proc_init(void)
{
2340
	return register_pernet_subsys(&tcp4_net_ops);
L
Linus Torvalds 已提交
2341 2342 2343 2344
}

void tcp4_proc_exit(void)
{
2345
	unregister_pernet_subsys(&tcp4_net_ops);
L
Linus Torvalds 已提交
2346 2347 2348 2349 2350 2351 2352 2353 2354
}
#endif /* CONFIG_PROC_FS */

struct proto tcp_prot = {
	.name			= "TCP",
	.owner			= THIS_MODULE,
	.close			= tcp_close,
	.connect		= tcp_v4_connect,
	.disconnect		= tcp_disconnect,
2355
	.accept			= inet_csk_accept,
L
Linus Torvalds 已提交
2356 2357 2358 2359 2360 2361 2362 2363
	.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,
2364 2365 2366
	.hash			= inet_hash,
	.unhash			= inet_unhash,
	.get_port		= inet_csk_get_port,
L
Linus Torvalds 已提交
2367 2368
	.enter_memory_pressure	= tcp_enter_memory_pressure,
	.sockets_allocated	= &tcp_sockets_allocated,
2369
	.orphan_count		= &tcp_orphan_count,
L
Linus Torvalds 已提交
2370 2371 2372 2373 2374 2375 2376
	.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),
2377
	.slab_flags		= SLAB_DESTROY_BY_RCU,
2378
	.twsk_prot		= &tcp_timewait_sock_ops,
2379
	.rsk_prot		= &tcp_request_sock_ops,
2380
	.h.hashinfo		= &tcp_hashinfo,
2381 2382 2383 2384
#ifdef CONFIG_COMPAT
	.compat_setsockopt	= compat_tcp_setsockopt,
	.compat_getsockopt	= compat_tcp_getsockopt,
#endif
L
Linus Torvalds 已提交
2385 2386
};

2387 2388 2389 2390 2391 2392 2393 2394 2395 2396

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);
2397
	inet_twsk_purge(net, &tcp_hashinfo, &tcp_death_row, AF_INET);
2398 2399 2400 2401 2402 2403 2404
}

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

2405
void __init tcp_v4_init(void)
L
Linus Torvalds 已提交
2406
{
2407
	inet_hashinfo_init(&tcp_hashinfo);
2408
	if (register_pernet_device(&tcp_sk_ops))
L
Linus Torvalds 已提交
2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427
		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);