tcp.c 82.6 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7
/*
 * 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).
 *
8
 * Authors:	Ross Biro
L
Linus Torvalds 已提交
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247
 *		Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *		Mark Evans, <evansmp@uhura.aston.ac.uk>
 *		Corey Minyard <wf-rch!minyard@relay.EU.net>
 *		Florian La Roche, <flla@stud.uni-sb.de>
 *		Charles Hedrick, <hedrick@klinzhai.rutgers.edu>
 *		Linus Torvalds, <torvalds@cs.helsinki.fi>
 *		Alan Cox, <gw4pts@gw4pts.ampr.org>
 *		Matthew Dillon, <dillon@apollo.west.oic.com>
 *		Arnt Gulbrandsen, <agulbra@nvg.unit.no>
 *		Jorge Cwik, <jorge@laser.satlink.net>
 *
 * Fixes:
 *		Alan Cox	:	Numerous verify_area() calls
 *		Alan Cox	:	Set the ACK bit on a reset
 *		Alan Cox	:	Stopped it crashing if it closed while
 *					sk->inuse=1 and was trying to connect
 *					(tcp_err()).
 *		Alan Cox	:	All icmp error handling was broken
 *					pointers passed where wrong and the
 *					socket was looked up backwards. Nobody
 *					tested any icmp error code obviously.
 *		Alan Cox	:	tcp_err() now handled properly. It
 *					wakes people on errors. poll
 *					behaves and the icmp error race
 *					has gone by moving it into sock.c
 *		Alan Cox	:	tcp_send_reset() fixed to work for
 *					everything not just packets for
 *					unknown sockets.
 *		Alan Cox	:	tcp option processing.
 *		Alan Cox	:	Reset tweaked (still not 100%) [Had
 *					syn rule wrong]
 *		Herp Rosmanith  :	More reset fixes
 *		Alan Cox	:	No longer acks invalid rst frames.
 *					Acking any kind of RST is right out.
 *		Alan Cox	:	Sets an ignore me flag on an rst
 *					receive otherwise odd bits of prattle
 *					escape still
 *		Alan Cox	:	Fixed another acking RST frame bug.
 *					Should stop LAN workplace lockups.
 *		Alan Cox	: 	Some tidyups using the new skb list
 *					facilities
 *		Alan Cox	:	sk->keepopen now seems to work
 *		Alan Cox	:	Pulls options out correctly on accepts
 *		Alan Cox	:	Fixed assorted sk->rqueue->next errors
 *		Alan Cox	:	PSH doesn't end a TCP read. Switched a
 *					bit to skb ops.
 *		Alan Cox	:	Tidied tcp_data to avoid a potential
 *					nasty.
 *		Alan Cox	:	Added some better commenting, as the
 *					tcp is hard to follow
 *		Alan Cox	:	Removed incorrect check for 20 * psh
 *	Michael O'Reilly	:	ack < copied bug fix.
 *	Johannes Stille		:	Misc tcp fixes (not all in yet).
 *		Alan Cox	:	FIN with no memory -> CRASH
 *		Alan Cox	:	Added socket option proto entries.
 *					Also added awareness of them to accept.
 *		Alan Cox	:	Added TCP options (SOL_TCP)
 *		Alan Cox	:	Switched wakeup calls to callbacks,
 *					so the kernel can layer network
 *					sockets.
 *		Alan Cox	:	Use ip_tos/ip_ttl settings.
 *		Alan Cox	:	Handle FIN (more) properly (we hope).
 *		Alan Cox	:	RST frames sent on unsynchronised
 *					state ack error.
 *		Alan Cox	:	Put in missing check for SYN bit.
 *		Alan Cox	:	Added tcp_select_window() aka NET2E
 *					window non shrink trick.
 *		Alan Cox	:	Added a couple of small NET2E timer
 *					fixes
 *		Charles Hedrick :	TCP fixes
 *		Toomas Tamm	:	TCP window fixes
 *		Alan Cox	:	Small URG fix to rlogin ^C ack fight
 *		Charles Hedrick	:	Rewrote most of it to actually work
 *		Linus		:	Rewrote tcp_read() and URG handling
 *					completely
 *		Gerhard Koerting:	Fixed some missing timer handling
 *		Matthew Dillon  :	Reworked TCP machine states as per RFC
 *		Gerhard Koerting:	PC/TCP workarounds
 *		Adam Caldwell	:	Assorted timer/timing errors
 *		Matthew Dillon	:	Fixed another RST bug
 *		Alan Cox	:	Move to kernel side addressing changes.
 *		Alan Cox	:	Beginning work on TCP fastpathing
 *					(not yet usable)
 *		Arnt Gulbrandsen:	Turbocharged tcp_check() routine.
 *		Alan Cox	:	TCP fast path debugging
 *		Alan Cox	:	Window clamping
 *		Michael Riepe	:	Bug in tcp_check()
 *		Matt Dillon	:	More TCP improvements and RST bug fixes
 *		Matt Dillon	:	Yet more small nasties remove from the
 *					TCP code (Be very nice to this man if
 *					tcp finally works 100%) 8)
 *		Alan Cox	:	BSD accept semantics.
 *		Alan Cox	:	Reset on closedown bug.
 *	Peter De Schrijver	:	ENOTCONN check missing in tcp_sendto().
 *		Michael Pall	:	Handle poll() after URG properly in
 *					all cases.
 *		Michael Pall	:	Undo the last fix in tcp_read_urg()
 *					(multi URG PUSH broke rlogin).
 *		Michael Pall	:	Fix the multi URG PUSH problem in
 *					tcp_readable(), poll() after URG
 *					works now.
 *		Michael Pall	:	recv(...,MSG_OOB) never blocks in the
 *					BSD api.
 *		Alan Cox	:	Changed the semantics of sk->socket to
 *					fix a race and a signal problem with
 *					accept() and async I/O.
 *		Alan Cox	:	Relaxed the rules on tcp_sendto().
 *		Yury Shevchuk	:	Really fixed accept() blocking problem.
 *		Craig I. Hagan  :	Allow for BSD compatible TIME_WAIT for
 *					clients/servers which listen in on
 *					fixed ports.
 *		Alan Cox	:	Cleaned the above up and shrank it to
 *					a sensible code size.
 *		Alan Cox	:	Self connect lockup fix.
 *		Alan Cox	:	No connect to multicast.
 *		Ross Biro	:	Close unaccepted children on master
 *					socket close.
 *		Alan Cox	:	Reset tracing code.
 *		Alan Cox	:	Spurious resets on shutdown.
 *		Alan Cox	:	Giant 15 minute/60 second timer error
 *		Alan Cox	:	Small whoops in polling before an
 *					accept.
 *		Alan Cox	:	Kept the state trace facility since
 *					it's handy for debugging.
 *		Alan Cox	:	More reset handler fixes.
 *		Alan Cox	:	Started rewriting the code based on
 *					the RFC's for other useful protocol
 *					references see: Comer, KA9Q NOS, and
 *					for a reference on the difference
 *					between specifications and how BSD
 *					works see the 4.4lite source.
 *		A.N.Kuznetsov	:	Don't time wait on completion of tidy
 *					close.
 *		Linus Torvalds	:	Fin/Shutdown & copied_seq changes.
 *		Linus Torvalds	:	Fixed BSD port reuse to work first syn
 *		Alan Cox	:	Reimplemented timers as per the RFC
 *					and using multiple timers for sanity.
 *		Alan Cox	:	Small bug fixes, and a lot of new
 *					comments.
 *		Alan Cox	:	Fixed dual reader crash by locking
 *					the buffers (much like datagram.c)
 *		Alan Cox	:	Fixed stuck sockets in probe. A probe
 *					now gets fed up of retrying without
 *					(even a no space) answer.
 *		Alan Cox	:	Extracted closing code better
 *		Alan Cox	:	Fixed the closing state machine to
 *					resemble the RFC.
 *		Alan Cox	:	More 'per spec' fixes.
 *		Jorge Cwik	:	Even faster checksumming.
 *		Alan Cox	:	tcp_data() doesn't ack illegal PSH
 *					only frames. At least one pc tcp stack
 *					generates them.
 *		Alan Cox	:	Cache last socket.
 *		Alan Cox	:	Per route irtt.
 *		Matt Day	:	poll()->select() match BSD precisely on error
 *		Alan Cox	:	New buffers
 *		Marc Tamsky	:	Various sk->prot->retransmits and
 *					sk->retransmits misupdating fixed.
 *					Fixed tcp_write_timeout: stuck close,
 *					and TCP syn retries gets used now.
 *		Mark Yarvis	:	In tcp_read_wakeup(), don't send an
 *					ack if state is TCP_CLOSED.
 *		Alan Cox	:	Look up device on a retransmit - routes may
 *					change. Doesn't yet cope with MSS shrink right
 *					but it's a start!
 *		Marc Tamsky	:	Closing in closing fixes.
 *		Mike Shaver	:	RFC1122 verifications.
 *		Alan Cox	:	rcv_saddr errors.
 *		Alan Cox	:	Block double connect().
 *		Alan Cox	:	Small hooks for enSKIP.
 *		Alexey Kuznetsov:	Path MTU discovery.
 *		Alan Cox	:	Support soft errors.
 *		Alan Cox	:	Fix MTU discovery pathological case
 *					when the remote claims no mtu!
 *		Marc Tamsky	:	TCP_CLOSE fix.
 *		Colin (G3TNE)	:	Send a reset on syn ack replies in
 *					window but wrong (fixes NT lpd problems)
 *		Pedro Roque	:	Better TCP window handling, delayed ack.
 *		Joerg Reuter	:	No modification of locked buffers in
 *					tcp_do_retransmit()
 *		Eric Schenk	:	Changed receiver side silly window
 *					avoidance algorithm to BSD style
 *					algorithm. This doubles throughput
 *					against machines running Solaris,
 *					and seems to result in general
 *					improvement.
 *	Stefan Magdalinski	:	adjusted tcp_readable() to fix FIONREAD
 *	Willy Konynenberg	:	Transparent proxying support.
 *	Mike McLagan		:	Routing by source
 *		Keith Owens	:	Do proper merging with partial SKB's in
 *					tcp_do_sendmsg to avoid burstiness.
 *		Eric Schenk	:	Fix fast close down bug with
 *					shutdown() followed by close().
 *		Andi Kleen 	:	Make poll agree with SIGIO
 *	Salvatore Sanfilippo	:	Support SO_LINGER with linger == 1 and
 *					lingertime == 0 (RFC 793 ABORT Call)
 *	Hirokazu Takahashi	:	Use copy_from_user() instead of
 *					csum_and_copy_from_user() if possible.
 *
 *		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.
 *
 * Description of States:
 *
 *	TCP_SYN_SENT		sent a connection request, waiting for ack
 *
 *	TCP_SYN_RECV		received a connection request, sent ack,
 *				waiting for final ack in three-way handshake.
 *
 *	TCP_ESTABLISHED		connection established
 *
 *	TCP_FIN_WAIT1		our side has shutdown, waiting to complete
 *				transmission of remaining buffered data
 *
 *	TCP_FIN_WAIT2		all buffered data sent, waiting for remote
 *				to shutdown
 *
 *	TCP_CLOSING		both sides have shutdown but we still have
 *				data we have to finish sending
 *
 *	TCP_TIME_WAIT		timeout to catch resent junk before entering
 *				closed, can only be entered from FIN_WAIT2
 *				or CLOSING.  Required because the other end
 *				may not have gotten our last ACK causing it
 *				to retransmit the data packet (which we ignore)
 *
 *	TCP_CLOSE_WAIT		remote side has shutdown and is waiting for
 *				us to finish writing our data and to shutdown
 *				(we have to close() to move on to LAST_ACK)
 *
 *	TCP_LAST_ACK		out side has shutdown after remote has
 *				shutdown.  There may still be data in our
 *				buffer that we have to finish sending
 *
 *	TCP_CLOSE		socket is finished
 */

248 249
#define pr_fmt(fmt) "TCP: " fmt

250
#include <linux/kernel.h>
L
Linus Torvalds 已提交
251 252 253 254 255 256
#include <linux/module.h>
#include <linux/types.h>
#include <linux/fcntl.h>
#include <linux/poll.h>
#include <linux/init.h>
#include <linux/fs.h>
J
Jens Axboe 已提交
257
#include <linux/skbuff.h>
258
#include <linux/scatterlist.h>
J
Jens Axboe 已提交
259 260 261
#include <linux/splice.h>
#include <linux/net.h>
#include <linux/socket.h>
L
Linus Torvalds 已提交
262 263
#include <linux/random.h>
#include <linux/bootmem.h>
264 265
#include <linux/highmem.h>
#include <linux/swap.h>
266
#include <linux/cache.h>
H
Herbert Xu 已提交
267
#include <linux/err.h>
268
#include <linux/crypto.h>
269
#include <linux/time.h>
270
#include <linux/slab.h>
L
Linus Torvalds 已提交
271 272

#include <net/icmp.h>
273
#include <net/inet_common.h>
L
Linus Torvalds 已提交
274 275 276
#include <net/tcp.h>
#include <net/xfrm.h>
#include <net/ip.h>
C
Chris Leech 已提交
277
#include <net/netdma.h>
J
Jens Axboe 已提交
278
#include <net/sock.h>
L
Linus Torvalds 已提交
279 280 281

#include <asm/uaccess.h>
#include <asm/ioctls.h>
282
#include <net/busy_poll.h>
L
Linus Torvalds 已提交
283

284
int sysctl_tcp_fin_timeout __read_mostly = TCP_FIN_TIMEOUT;
L
Linus Torvalds 已提交
285

E
Eric Dumazet 已提交
286 287
int sysctl_tcp_min_tso_segs __read_mostly = 2;

E
Eric Dumazet 已提交
288 289
int sysctl_tcp_autocorking __read_mostly = 1;

290
struct percpu_counter tcp_orphan_count;
291 292
EXPORT_SYMBOL_GPL(tcp_orphan_count);

293
long sysctl_tcp_mem[3] __read_mostly;
294 295
int sysctl_tcp_wmem[3] __read_mostly;
int sysctl_tcp_rmem[3] __read_mostly;
L
Linus Torvalds 已提交
296

297
EXPORT_SYMBOL(sysctl_tcp_mem);
L
Linus Torvalds 已提交
298 299 300
EXPORT_SYMBOL(sysctl_tcp_rmem);
EXPORT_SYMBOL(sysctl_tcp_wmem);

E
Eric Dumazet 已提交
301
atomic_long_t tcp_memory_allocated;	/* Current allocated memory. */
L
Linus Torvalds 已提交
302
EXPORT_SYMBOL(tcp_memory_allocated);
303 304 305 306 307

/*
 * Current number of TCP sockets.
 */
struct percpu_counter tcp_sockets_allocated;
L
Linus Torvalds 已提交
308 309
EXPORT_SYMBOL(tcp_sockets_allocated);

J
Jens Axboe 已提交
310 311 312 313 314 315 316 317 318
/*
 * TCP splice context
 */
struct tcp_splice_state {
	struct pipe_inode_info *pipe;
	size_t len;
	unsigned int flags;
};

L
Linus Torvalds 已提交
319 320 321
/*
 * Pressure flag: try to collapse.
 * Technical note: it is used by multiple contexts non atomically.
322
 * All the __sk_mem_schedule() is of this nature: accounting
L
Linus Torvalds 已提交
323 324
 * is strict, actions are advisory and have some latency.
 */
325
int tcp_memory_pressure __read_mostly;
L
Linus Torvalds 已提交
326 327
EXPORT_SYMBOL(tcp_memory_pressure);

328
void tcp_enter_memory_pressure(struct sock *sk)
L
Linus Torvalds 已提交
329 330
{
	if (!tcp_memory_pressure) {
P
Pavel Emelyanov 已提交
331
		NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMEMORYPRESSURES);
L
Linus Torvalds 已提交
332 333 334 335 336
		tcp_memory_pressure = 1;
	}
}
EXPORT_SYMBOL(tcp_enter_memory_pressure);

337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373
/* Convert seconds to retransmits based on initial and max timeout */
static u8 secs_to_retrans(int seconds, int timeout, int rto_max)
{
	u8 res = 0;

	if (seconds > 0) {
		int period = timeout;

		res = 1;
		while (seconds > period && res < 255) {
			res++;
			timeout <<= 1;
			if (timeout > rto_max)
				timeout = rto_max;
			period += timeout;
		}
	}
	return res;
}

/* Convert retransmits to seconds based on initial and max timeout */
static int retrans_to_secs(u8 retrans, int timeout, int rto_max)
{
	int period = 0;

	if (retrans > 0) {
		period = timeout;
		while (--retrans) {
			timeout <<= 1;
			if (timeout > rto_max)
				timeout = rto_max;
			period += timeout;
		}
	}
	return period;
}

374 375 376 377 378 379 380 381 382 383 384 385 386
/* Address-family independent initialization for a tcp_sock.
 *
 * NOTE: A lot of things set to zero explicitly by call to
 *       sk_alloc() so need not be done here.
 */
void tcp_init_sock(struct sock *sk)
{
	struct inet_connection_sock *icsk = inet_csk(sk);
	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);
E
Eric Dumazet 已提交
387
	INIT_LIST_HEAD(&tp->tsq_node);
388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406

	icsk->icsk_rto = TCP_TIMEOUT_INIT;
	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 = TCP_INIT_CWND;

	/* See draft-stevens-tcpca-spec-01 for discussion of the
	 * initialization of these values.
	 */
	tp->snd_ssthresh = TCP_INFINITE_SSTHRESH;
	tp->snd_cwnd_clamp = ~0;
	tp->mss_cache = TCP_MSS_DEFAULT;

	tp->reordering = sysctl_tcp_reordering;
Y
Yuchung Cheng 已提交
407
	tcp_enable_early_retrans(tp);
408 409
	icsk->icsk_ca_ops = &tcp_init_congestion_ops;

410 411
	tp->tsoffset = 0;

412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428
	sk->sk_state = TCP_CLOSE;

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

	icsk->icsk_sync_mss = tcp_sync_mss;

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

	local_bh_disable();
	sock_update_memcg(sk);
	sk_sockets_allocated_inc(sk);
	local_bh_enable();
}
EXPORT_SYMBOL(tcp_init_sock);

L
Linus Torvalds 已提交
429 430 431 432 433 434 435 436 437 438 439
/*
 *	Wait for a TCP event.
 *
 *	Note that we don't need to lock the socket, as the upper poll layers
 *	take care of normal races (between the test and the event) and we don't
 *	go look at any of the socket buffers directly.
 */
unsigned int tcp_poll(struct file *file, struct socket *sock, poll_table *wait)
{
	unsigned int mask;
	struct sock *sk = sock->sk;
440
	const struct tcp_sock *tp = tcp_sk(sk);
L
Linus Torvalds 已提交
441

442 443
	sock_rps_record_flow(sk);

E
Eric Dumazet 已提交
444
	sock_poll_wait(file, sk_sleep(sk), wait);
L
Linus Torvalds 已提交
445
	if (sk->sk_state == TCP_LISTEN)
446
		return inet_csk_listen_poll(sk);
L
Linus Torvalds 已提交
447 448

	/* Socket is not locked. We are protected from async events
449 450
	 * by poll logic and correct handling of state changes
	 * made by other threads is impossible in any case.
L
Linus Torvalds 已提交
451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473
	 */

	mask = 0;

	/*
	 * POLLHUP is certainly not done right. But poll() doesn't
	 * have a notion of HUP in just one direction, and for a
	 * socket the read side is more interesting.
	 *
	 * Some poll() documentation says that POLLHUP is incompatible
	 * with the POLLOUT/POLLWR flags, so somebody should check this
	 * all. But careful, it tends to be safer to return too many
	 * bits than too few, and you can easily break real applications
	 * if you don't tell them that something has hung up!
	 *
	 * Check-me.
	 *
	 * Check number 1. POLLHUP is _UNMASKABLE_ event (see UNIX98 and
	 * our fs/select.c). It means that after we received EOF,
	 * poll always returns immediately, making impossible poll() on write()
	 * in state CLOSE_WAIT. One solution is evident --- to set POLLHUP
	 * if and only if shutdown has been made in both directions.
	 * Actually, it is interesting to look how Solaris and DUX
474
	 * solve this dilemma. I would prefer, if POLLHUP were maskable,
L
Linus Torvalds 已提交
475 476
	 * then we could set it on SND_SHUTDOWN. BTW examples given
	 * in Stevens' books assume exactly this behaviour, it explains
477
	 * why POLLHUP is incompatible with POLLOUT.	--ANK
L
Linus Torvalds 已提交
478 479 480 481 482 483 484
	 *
	 * NOTE. Check for TCP_CLOSE is added. The goal is to prevent
	 * blocking on fresh not-connected or disconnected socket. --ANK
	 */
	if (sk->sk_shutdown == SHUTDOWN_MASK || sk->sk_state == TCP_CLOSE)
		mask |= POLLHUP;
	if (sk->sk_shutdown & RCV_SHUTDOWN)
485
		mask |= POLLIN | POLLRDNORM | POLLRDHUP;
L
Linus Torvalds 已提交
486

487 488 489
	/* Connected or passive Fast Open socket? */
	if (sk->sk_state != TCP_SYN_SENT &&
	    (sk->sk_state != TCP_SYN_RECV || tp->fastopen_rsk != NULL)) {
490 491 492 493 494
		int target = sock_rcvlowat(sk, 0, INT_MAX);

		if (tp->urg_seq == tp->copied_seq &&
		    !sock_flag(sk, SOCK_URGINLINE) &&
		    tp->urg_data)
495
			target++;
496

L
Linus Torvalds 已提交
497 498 499
		/* Potential race condition. If read of tp below will
		 * escape above sk->sk_state, we can be illegally awaken
		 * in SYN_* states. */
500
		if (tp->rcv_nxt - tp->copied_seq >= target)
L
Linus Torvalds 已提交
501 502 503
			mask |= POLLIN | POLLRDNORM;

		if (!(sk->sk_shutdown & SEND_SHUTDOWN)) {
504
			if (sk_stream_is_writeable(sk)) {
L
Linus Torvalds 已提交
505 506 507 508 509 510 511 512 513 514
				mask |= POLLOUT | POLLWRNORM;
			} else {  /* send SIGIO later */
				set_bit(SOCK_ASYNC_NOSPACE,
					&sk->sk_socket->flags);
				set_bit(SOCK_NOSPACE, &sk->sk_socket->flags);

				/* Race breaker. If space is freed after
				 * wspace test but before the flags are set,
				 * IO signal will be lost.
				 */
515
				if (sk_stream_is_writeable(sk))
L
Linus Torvalds 已提交
516 517
					mask |= POLLOUT | POLLWRNORM;
			}
518 519
		} else
			mask |= POLLOUT | POLLWRNORM;
L
Linus Torvalds 已提交
520 521 522 523

		if (tp->urg_data & TCP_URG_VALID)
			mask |= POLLPRI;
	}
T
Tom Marshall 已提交
524 525 526 527 528
	/* This barrier is coupled with smp_wmb() in tcp_reset() */
	smp_rmb();
	if (sk->sk_err)
		mask |= POLLERR;

L
Linus Torvalds 已提交
529 530
	return mask;
}
E
Eric Dumazet 已提交
531
EXPORT_SYMBOL(tcp_poll);
L
Linus Torvalds 已提交
532 533 534 535 536

int tcp_ioctl(struct sock *sk, int cmd, unsigned long arg)
{
	struct tcp_sock *tp = tcp_sk(sk);
	int answ;
E
Eric Dumazet 已提交
537
	bool slow;
L
Linus Torvalds 已提交
538 539 540 541 542 543

	switch (cmd) {
	case SIOCINQ:
		if (sk->sk_state == TCP_LISTEN)
			return -EINVAL;

E
Eric Dumazet 已提交
544
		slow = lock_sock_fast(sk);
L
Linus Torvalds 已提交
545 546 547 548 549 550
		if ((1 << sk->sk_state) & (TCPF_SYN_SENT | TCPF_SYN_RECV))
			answ = 0;
		else if (sock_flag(sk, SOCK_URGINLINE) ||
			 !tp->urg_data ||
			 before(tp->urg_seq, tp->copied_seq) ||
			 !before(tp->urg_seq, tp->rcv_nxt)) {
551

L
Linus Torvalds 已提交
552 553
			answ = tp->rcv_nxt - tp->copied_seq;

E
Eric Dumazet 已提交
554 555 556
			/* Subtract 1, if FIN was received */
			if (answ && sock_flag(sk, SOCK_DONE))
				answ--;
L
Linus Torvalds 已提交
557 558
		} else
			answ = tp->urg_seq - tp->copied_seq;
E
Eric Dumazet 已提交
559
		unlock_sock_fast(sk, slow);
L
Linus Torvalds 已提交
560 561 562 563 564 565 566 567 568 569 570 571 572
		break;
	case SIOCATMARK:
		answ = tp->urg_data && tp->urg_seq == tp->copied_seq;
		break;
	case SIOCOUTQ:
		if (sk->sk_state == TCP_LISTEN)
			return -EINVAL;

		if ((1 << sk->sk_state) & (TCPF_SYN_SENT | TCPF_SYN_RECV))
			answ = 0;
		else
			answ = tp->write_seq - tp->snd_una;
		break;
573 574 575 576 577 578 579 580 581
	case SIOCOUTQNSD:
		if (sk->sk_state == TCP_LISTEN)
			return -EINVAL;

		if ((1 << sk->sk_state) & (TCPF_SYN_SENT | TCPF_SYN_RECV))
			answ = 0;
		else
			answ = tp->write_seq - tp->snd_nxt;
		break;
L
Linus Torvalds 已提交
582 583
	default:
		return -ENOIOCTLCMD;
584
	}
L
Linus Torvalds 已提交
585 586 587

	return put_user(answ, (int __user *)arg);
}
E
Eric Dumazet 已提交
588
EXPORT_SYMBOL(tcp_ioctl);
L
Linus Torvalds 已提交
589 590 591

static inline void tcp_mark_push(struct tcp_sock *tp, struct sk_buff *skb)
{
E
Eric Dumazet 已提交
592
	TCP_SKB_CB(skb)->tcp_flags |= TCPHDR_PSH;
L
Linus Torvalds 已提交
593 594 595
	tp->pushed_seq = tp->write_seq;
}

E
Eric Dumazet 已提交
596
static inline bool forced_push(const struct tcp_sock *tp)
L
Linus Torvalds 已提交
597 598 599 600
{
	return after(tp->write_seq, tp->pushed_seq + (tp->max_window >> 1));
}

601
static inline void skb_entail(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
602
{
603
	struct tcp_sock *tp = tcp_sk(sk);
604 605 606 607
	struct tcp_skb_cb *tcb = TCP_SKB_CB(skb);

	skb->csum    = 0;
	tcb->seq     = tcb->end_seq = tp->write_seq;
E
Eric Dumazet 已提交
608
	tcb->tcp_flags = TCPHDR_ACK;
609
	tcb->sacked  = 0;
L
Linus Torvalds 已提交
610
	skb_header_release(skb);
611
	tcp_add_write_queue_tail(sk, skb);
612 613
	sk->sk_wmem_queued += skb->truesize;
	sk_mem_charge(sk, skb->truesize);
614
	if (tp->nonagle & TCP_NAGLE_PUSH)
615
		tp->nonagle &= ~TCP_NAGLE_PUSH;
L
Linus Torvalds 已提交
616 617
}

618
static inline void tcp_mark_urg(struct tcp_sock *tp, int flags)
L
Linus Torvalds 已提交
619
{
I
Ilpo Järvinen 已提交
620
	if (flags & MSG_OOB)
L
Linus Torvalds 已提交
621 622 623
		tp->snd_up = tp->write_seq;
}

E
Eric Dumazet 已提交
624 625 626 627 628 629 630 631 632 633 634
/* If a not yet filled skb is pushed, do not send it if
 * we have packets in Qdisc or NIC queues :
 * Because TX completion will happen shortly, it gives a chance
 * to coalesce future sendmsg() payload into this skb, without
 * need for a timer, and with no latency trade off.
 * As packets containing data payload have a bigger truesize
 * than pure acks (dataless) packets, the last check prevents
 * autocorking if we only have an ACK in Qdisc/NIC queues.
 */
static bool tcp_should_autocork(struct sock *sk, struct sk_buff *skb,
				int size_goal)
L
Linus Torvalds 已提交
635
{
E
Eric Dumazet 已提交
636 637 638 639 640 641 642 643 644 645
	return skb->len < size_goal &&
	       sysctl_tcp_autocorking &&
	       atomic_read(&sk->sk_wmem_alloc) > skb->truesize;
}

static void tcp_push(struct sock *sk, int flags, int mss_now,
		     int nonagle, int size_goal)
{
	struct tcp_sock *tp = tcp_sk(sk);
	struct sk_buff *skb;
646

E
Eric Dumazet 已提交
647 648
	if (!tcp_send_head(sk))
		return;
649

E
Eric Dumazet 已提交
650 651 652 653 654 655 656 657 658 659 660 661 662 663
	skb = tcp_write_queue_tail(sk);
	if (!(flags & MSG_MORE) || forced_push(tp))
		tcp_mark_push(tp, skb);

	tcp_mark_urg(tp, flags);

	if (tcp_should_autocork(sk, skb, size_goal)) {

		/* avoid atomic op if TSQ_THROTTLED bit is already set */
		if (!test_bit(TSQ_THROTTLED, &tp->tsq_flags)) {
			NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPAUTOCORKING);
			set_bit(TSQ_THROTTLED, &tp->tsq_flags);
		}
		return;
L
Linus Torvalds 已提交
664
	}
E
Eric Dumazet 已提交
665 666 667 668 669

	if (flags & MSG_MORE)
		nonagle = TCP_NAGLE_CORK;

	__tcp_push_pending_frames(sk, mss_now, nonagle);
L
Linus Torvalds 已提交
670 671
}

672 673
static int tcp_splice_data_recv(read_descriptor_t *rd_desc, struct sk_buff *skb,
				unsigned int offset, size_t len)
J
Jens Axboe 已提交
674 675
{
	struct tcp_splice_state *tss = rd_desc->arg.data;
676
	int ret;
J
Jens Axboe 已提交
677

678 679
	ret = skb_splice_bits(skb, offset, tss->pipe, min(rd_desc->count, len),
			      tss->flags);
680 681 682
	if (ret > 0)
		rd_desc->count -= ret;
	return ret;
J
Jens Axboe 已提交
683 684 685 686 687 688 689
}

static int __tcp_splice_read(struct sock *sk, struct tcp_splice_state *tss)
{
	/* Store TCP splice context information in read_descriptor_t. */
	read_descriptor_t rd_desc = {
		.arg.data = tss,
690
		.count	  = tss->len,
J
Jens Axboe 已提交
691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721
	};

	return tcp_read_sock(sk, &rd_desc, tcp_splice_data_recv);
}

/**
 *  tcp_splice_read - splice data from TCP socket to a pipe
 * @sock:	socket to splice from
 * @ppos:	position (not valid)
 * @pipe:	pipe to splice to
 * @len:	number of bytes to splice
 * @flags:	splice modifier flags
 *
 * Description:
 *    Will read pages from given socket and fill them into a pipe.
 *
 **/
ssize_t tcp_splice_read(struct socket *sock, loff_t *ppos,
			struct pipe_inode_info *pipe, size_t len,
			unsigned int flags)
{
	struct sock *sk = sock->sk;
	struct tcp_splice_state tss = {
		.pipe = pipe,
		.len = len,
		.flags = flags,
	};
	long timeo;
	ssize_t spliced;
	int ret;

722
	sock_rps_record_flow(sk);
J
Jens Axboe 已提交
723 724 725 726 727 728 729 730 731 732
	/*
	 * We can't seek on a socket input
	 */
	if (unlikely(*ppos))
		return -ESPIPE;

	ret = spliced = 0;

	lock_sock(sk);

733
	timeo = sock_rcvtimeo(sk, sock->file->f_flags & O_NONBLOCK);
J
Jens Axboe 已提交
734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771
	while (tss.len) {
		ret = __tcp_splice_read(sk, &tss);
		if (ret < 0)
			break;
		else if (!ret) {
			if (spliced)
				break;
			if (sock_flag(sk, SOCK_DONE))
				break;
			if (sk->sk_err) {
				ret = sock_error(sk);
				break;
			}
			if (sk->sk_shutdown & RCV_SHUTDOWN)
				break;
			if (sk->sk_state == TCP_CLOSE) {
				/*
				 * This occurs when user tries to read
				 * from never connected socket.
				 */
				if (!sock_flag(sk, SOCK_DONE))
					ret = -ENOTCONN;
				break;
			}
			if (!timeo) {
				ret = -EAGAIN;
				break;
			}
			sk_wait_data(sk, &timeo);
			if (signal_pending(current)) {
				ret = sock_intr_errno(timeo);
				break;
			}
			continue;
		}
		tss.len -= ret;
		spliced += ret;

772 773
		if (!timeo)
			break;
J
Jens Axboe 已提交
774 775 776 777
		release_sock(sk);
		lock_sock(sk);

		if (sk->sk_err || sk->sk_state == TCP_CLOSE ||
778
		    (sk->sk_shutdown & RCV_SHUTDOWN) ||
J
Jens Axboe 已提交
779 780 781 782 783 784 785 786 787 788 789
		    signal_pending(current))
			break;
	}

	release_sock(sk);

	if (spliced)
		return spliced;

	return ret;
}
E
Eric Dumazet 已提交
790
EXPORT_SYMBOL(tcp_splice_read);
J
Jens Axboe 已提交
791

792
struct sk_buff *sk_stream_alloc_skb(struct sock *sk, int size, gfp_t gfp)
793 794 795 796 797 798 799 800
{
	struct sk_buff *skb;

	/* The TCP header must be at least 32-bit aligned.  */
	size = ALIGN(size, 4);

	skb = alloc_skb_fclone(size + sk->sk_prot->max_header, gfp);
	if (skb) {
801
		if (sk_wmem_schedule(sk, skb->truesize)) {
802
			skb_reserve(skb, sk->sk_prot->max_header);
803 804 805 806
			/*
			 * Make sure that we have exactly size bytes
			 * available to the caller, no more, no less.
			 */
E
Eric Dumazet 已提交
807
			skb->reserved_tailroom = skb->end - skb->tail - size;
808 809 810 811
			return skb;
		}
		__kfree_skb(skb);
	} else {
812
		sk->sk_prot->enter_memory_pressure(sk);
813 814 815 816 817
		sk_stream_moderate_sndbuf(sk);
	}
	return NULL;
}

I
Ilpo Järvinen 已提交
818 819 820 821
static unsigned int tcp_xmit_size_goal(struct sock *sk, u32 mss_now,
				       int large_allowed)
{
	struct tcp_sock *tp = tcp_sk(sk);
822
	u32 xmit_size_goal, old_size_goal;
I
Ilpo Järvinen 已提交
823 824 825 826

	xmit_size_goal = mss_now;

	if (large_allowed && sk_can_gso(sk)) {
E
Eric Dumazet 已提交
827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845
		u32 gso_size, hlen;

		/* Maybe we should/could use sk->sk_prot->max_header here ? */
		hlen = inet_csk(sk)->icsk_af_ops->net_header_len +
		       inet_csk(sk)->icsk_ext_hdr_len +
		       tp->tcp_header_len;

		/* Goal is to send at least one packet per ms,
		 * not one big TSO packet every 100 ms.
		 * This preserves ACK clocking and is consistent
		 * with tcp_tso_should_defer() heuristic.
		 */
		gso_size = sk->sk_pacing_rate / (2 * MSEC_PER_SEC);
		gso_size = max_t(u32, gso_size,
				 sysctl_tcp_min_tso_segs * mss_now);

		xmit_size_goal = min_t(u32, gso_size,
				       sk->sk_gso_max_size - 1 - hlen);

I
Ilpo Järvinen 已提交
846
		xmit_size_goal = tcp_bound_to_half_wnd(tp, xmit_size_goal);
847 848 849 850 851 852 853 854

		/* We try hard to avoid divides here */
		old_size_goal = tp->xmit_size_goal_segs * mss_now;

		if (likely(old_size_goal <= xmit_size_goal &&
			   old_size_goal + mss_now > xmit_size_goal)) {
			xmit_size_goal = old_size_goal;
		} else {
855 856 857
			tp->xmit_size_goal_segs =
				min_t(u16, xmit_size_goal / mss_now,
				      sk->sk_gso_max_segs);
858 859
			xmit_size_goal = tp->xmit_size_goal_segs * mss_now;
		}
I
Ilpo Järvinen 已提交
860 861
	}

862
	return max(xmit_size_goal, mss_now);
I
Ilpo Järvinen 已提交
863 864 865 866 867 868 869 870 871 872 873 874
}

static int tcp_send_mss(struct sock *sk, int *size_goal, int flags)
{
	int mss_now;

	mss_now = tcp_current_mss(sk);
	*size_goal = tcp_xmit_size_goal(sk, mss_now, !(flags & MSG_OOB));

	return mss_now;
}

875 876
static ssize_t do_tcp_sendpages(struct sock *sk, struct page *page, int offset,
				size_t size, int flags)
L
Linus Torvalds 已提交
877 878
{
	struct tcp_sock *tp = tcp_sk(sk);
879
	int mss_now, size_goal;
L
Linus Torvalds 已提交
880 881 882 883
	int err;
	ssize_t copied;
	long timeo = sock_sndtimeo(sk, flags & MSG_DONTWAIT);

884 885 886 887 888 889
	/* Wait for a connection to finish. One exception is TCP Fast Open
	 * (passive side) where data is allowed to be sent before a connection
	 * is fully established.
	 */
	if (((1 << sk->sk_state) & ~(TCPF_ESTABLISHED | TCPF_CLOSE_WAIT)) &&
	    !tcp_passive_fastopen(sk)) {
L
Linus Torvalds 已提交
890 891
		if ((err = sk_stream_wait_connect(sk, &timeo)) != 0)
			goto out_err;
892
	}
L
Linus Torvalds 已提交
893 894 895

	clear_bit(SOCK_ASYNC_NOSPACE, &sk->sk_socket->flags);

I
Ilpo Järvinen 已提交
896
	mss_now = tcp_send_mss(sk, &size_goal, flags);
L
Linus Torvalds 已提交
897 898 899 900
	copied = 0;

	err = -EPIPE;
	if (sk->sk_err || (sk->sk_shutdown & SEND_SHUTDOWN))
901
		goto out_err;
L
Linus Torvalds 已提交
902

903
	while (size > 0) {
904
		struct sk_buff *skb = tcp_write_queue_tail(sk);
905 906
		int copy, i;
		bool can_coalesce;
L
Linus Torvalds 已提交
907

908
		if (!tcp_send_head(sk) || (copy = size_goal - skb->len) <= 0) {
L
Linus Torvalds 已提交
909 910 911 912
new_segment:
			if (!sk_stream_memory_free(sk))
				goto wait_for_sndbuf;

913
			skb = sk_stream_alloc_skb(sk, 0, sk->sk_allocation);
L
Linus Torvalds 已提交
914 915 916
			if (!skb)
				goto wait_for_memory;

917
			skb_entail(sk, skb);
918
			copy = size_goal;
L
Linus Torvalds 已提交
919 920 921 922 923 924 925 926 927 928 929
		}

		if (copy > size)
			copy = size;

		i = skb_shinfo(skb)->nr_frags;
		can_coalesce = skb_can_coalesce(skb, i, page, offset);
		if (!can_coalesce && i >= MAX_SKB_FRAGS) {
			tcp_mark_push(tp, skb);
			goto new_segment;
		}
930
		if (!sk_wmem_schedule(sk, copy))
L
Linus Torvalds 已提交
931
			goto wait_for_memory;
932

L
Linus Torvalds 已提交
933
		if (can_coalesce) {
E
Eric Dumazet 已提交
934
			skb_frag_size_add(&skb_shinfo(skb)->frags[i - 1], copy);
L
Linus Torvalds 已提交
935 936 937 938
		} else {
			get_page(page);
			skb_fill_page_desc(skb, i, page, offset, copy);
		}
939
		skb_shinfo(skb)->tx_flags |= SKBTX_SHARED_FRAG;
940

L
Linus Torvalds 已提交
941 942 943 944
		skb->len += copy;
		skb->data_len += copy;
		skb->truesize += copy;
		sk->sk_wmem_queued += copy;
945
		sk_mem_charge(sk, copy);
946
		skb->ip_summed = CHECKSUM_PARTIAL;
L
Linus Torvalds 已提交
947 948
		tp->write_seq += copy;
		TCP_SKB_CB(skb)->end_seq += copy;
949
		skb_shinfo(skb)->gso_segs = 0;
L
Linus Torvalds 已提交
950 951

		if (!copied)
E
Eric Dumazet 已提交
952
			TCP_SKB_CB(skb)->tcp_flags &= ~TCPHDR_PSH;
L
Linus Torvalds 已提交
953 954

		copied += copy;
955 956
		offset += copy;
		if (!(size -= copy))
L
Linus Torvalds 已提交
957 958
			goto out;

959
		if (skb->len < size_goal || (flags & MSG_OOB))
L
Linus Torvalds 已提交
960 961 962 963
			continue;

		if (forced_push(tp)) {
			tcp_mark_push(tp, skb);
964
			__tcp_push_pending_frames(sk, mss_now, TCP_NAGLE_PUSH);
965
		} else if (skb == tcp_send_head(sk))
L
Linus Torvalds 已提交
966 967 968 969 970 971
			tcp_push_one(sk, mss_now);
		continue;

wait_for_sndbuf:
		set_bit(SOCK_NOSPACE, &sk->sk_socket->flags);
wait_for_memory:
E
Eric Dumazet 已提交
972 973
		tcp_push(sk, flags & ~MSG_MORE, mss_now,
			 TCP_NAGLE_PUSH, size_goal);
L
Linus Torvalds 已提交
974 975 976 977

		if ((err = sk_stream_wait_memory(sk, &timeo)) != 0)
			goto do_error;

I
Ilpo Järvinen 已提交
978
		mss_now = tcp_send_mss(sk, &size_goal, flags);
L
Linus Torvalds 已提交
979 980 981
	}

out:
982
	if (copied && !(flags & MSG_SENDPAGE_NOTLAST))
E
Eric Dumazet 已提交
983
		tcp_push(sk, flags, mss_now, tp->nonagle, size_goal);
L
Linus Torvalds 已提交
984 985 986 987 988 989 990 991 992
	return copied;

do_error:
	if (copied)
		goto out;
out_err:
	return sk_stream_error(sk, flags, err);
}

993 994
int tcp_sendpage(struct sock *sk, struct page *page, int offset,
		 size_t size, int flags)
L
Linus Torvalds 已提交
995 996 997 998
{
	ssize_t res;

	if (!(sk->sk_route_caps & NETIF_F_SG) ||
999
	    !(sk->sk_route_caps & NETIF_F_ALL_CSUM))
1000 1001
		return sock_no_sendpage(sk->sk_socket, page, offset, size,
					flags);
L
Linus Torvalds 已提交
1002 1003

	lock_sock(sk);
1004
	res = do_tcp_sendpages(sk, page, offset, size, flags);
L
Linus Torvalds 已提交
1005 1006 1007
	release_sock(sk);
	return res;
}
E
Eric Dumazet 已提交
1008
EXPORT_SYMBOL(tcp_sendpage);
L
Linus Torvalds 已提交
1009

1010
static inline int select_size(const struct sock *sk, bool sg)
L
Linus Torvalds 已提交
1011
{
1012
	const struct tcp_sock *tp = tcp_sk(sk);
1013
	int tmp = tp->mss_cache;
L
Linus Torvalds 已提交
1014

K
Krishna Kumar 已提交
1015
	if (sg) {
1016 1017 1018 1019 1020 1021
		if (sk_can_gso(sk)) {
			/* Small frames wont use a full page:
			 * Payload will immediately follow tcp header.
			 */
			tmp = SKB_WITH_OVERHEAD(2048 - MAX_TCP_HEADER);
		} else {
1022 1023 1024 1025 1026 1027 1028
			int pgbreak = SKB_MAX_HEAD(MAX_TCP_HEADER);

			if (tmp >= pgbreak &&
			    tmp <= pgbreak + (MAX_SKB_FRAGS - 1) * PAGE_SIZE)
				tmp = pgbreak;
		}
	}
L
Linus Torvalds 已提交
1029 1030 1031 1032

	return tmp;
}

1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064
void tcp_free_fastopen_req(struct tcp_sock *tp)
{
	if (tp->fastopen_req != NULL) {
		kfree(tp->fastopen_req);
		tp->fastopen_req = NULL;
	}
}

static int tcp_sendmsg_fastopen(struct sock *sk, struct msghdr *msg, int *size)
{
	struct tcp_sock *tp = tcp_sk(sk);
	int err, flags;

	if (!(sysctl_tcp_fastopen & TFO_CLIENT_ENABLE))
		return -EOPNOTSUPP;
	if (tp->fastopen_req != NULL)
		return -EALREADY; /* Another Fast Open is in progress */

	tp->fastopen_req = kzalloc(sizeof(struct tcp_fastopen_request),
				   sk->sk_allocation);
	if (unlikely(tp->fastopen_req == NULL))
		return -ENOBUFS;
	tp->fastopen_req->data = msg;

	flags = (msg->msg_flags & MSG_DONTWAIT) ? O_NONBLOCK : 0;
	err = __inet_stream_connect(sk->sk_socket, msg->msg_name,
				    msg->msg_namelen, flags);
	*size = tp->fastopen_req->copied;
	tcp_free_fastopen_req(tp);
	return err;
}

1065
int tcp_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
L
Linus Torvalds 已提交
1066 1067 1068 1069 1070
		size_t size)
{
	struct iovec *iov;
	struct tcp_sock *tp = tcp_sk(sk);
	struct sk_buff *skb;
1071 1072
	int iovlen, flags, err, copied = 0;
	int mss_now = 0, size_goal, copied_syn = 0, offset = 0;
1073
	bool sg;
L
Linus Torvalds 已提交
1074 1075 1076 1077 1078
	long timeo;

	lock_sock(sk);

	flags = msg->msg_flags;
1079 1080 1081 1082 1083 1084 1085 1086 1087
	if (flags & MSG_FASTOPEN) {
		err = tcp_sendmsg_fastopen(sk, msg, &copied_syn);
		if (err == -EINPROGRESS && copied_syn > 0)
			goto out;
		else if (err)
			goto out_err;
		offset = copied_syn;
	}

L
Linus Torvalds 已提交
1088 1089
	timeo = sock_sndtimeo(sk, flags & MSG_DONTWAIT);

1090 1091 1092 1093 1094 1095
	/* Wait for a connection to finish. One exception is TCP Fast Open
	 * (passive side) where data is allowed to be sent before a connection
	 * is fully established.
	 */
	if (((1 << sk->sk_state) & ~(TCPF_ESTABLISHED | TCPF_CLOSE_WAIT)) &&
	    !tcp_passive_fastopen(sk)) {
L
Linus Torvalds 已提交
1096
		if ((err = sk_stream_wait_connect(sk, &timeo)) != 0)
1097
			goto do_error;
1098
	}
L
Linus Torvalds 已提交
1099

P
Pavel Emelyanov 已提交
1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112
	if (unlikely(tp->repair)) {
		if (tp->repair_queue == TCP_RECV_QUEUE) {
			copied = tcp_send_rcvq(sk, msg, size);
			goto out;
		}

		err = -EINVAL;
		if (tp->repair_queue == TCP_NO_QUEUE)
			goto out_err;

		/* 'common' sending to sendq */
	}

L
Linus Torvalds 已提交
1113 1114 1115
	/* This should be in poll */
	clear_bit(SOCK_ASYNC_NOSPACE, &sk->sk_socket->flags);

I
Ilpo Järvinen 已提交
1116
	mss_now = tcp_send_mss(sk, &size_goal, flags);
L
Linus Torvalds 已提交
1117 1118 1119 1120 1121 1122 1123 1124

	/* Ok commence sending. */
	iovlen = msg->msg_iovlen;
	iov = msg->msg_iov;
	copied = 0;

	err = -EPIPE;
	if (sk->sk_err || (sk->sk_shutdown & SEND_SHUTDOWN))
1125
		goto out_err;
L
Linus Torvalds 已提交
1126

1127
	sg = !!(sk->sk_route_caps & NETIF_F_SG);
K
Krishna Kumar 已提交
1128

L
Linus Torvalds 已提交
1129
	while (--iovlen >= 0) {
1130
		size_t seglen = iov->iov_len;
L
Linus Torvalds 已提交
1131 1132 1133
		unsigned char __user *from = iov->iov_base;

		iov++;
1134 1135 1136 1137 1138 1139 1140 1141 1142
		if (unlikely(offset > 0)) {  /* Skip bytes copied in SYN */
			if (offset >= seglen) {
				offset -= seglen;
				continue;
			}
			seglen -= offset;
			from += offset;
			offset = 0;
		}
L
Linus Torvalds 已提交
1143 1144

		while (seglen > 0) {
1145 1146
			int copy = 0;
			int max = size_goal;
L
Linus Torvalds 已提交
1147

1148
			skb = tcp_write_queue_tail(sk);
1149 1150 1151 1152 1153
			if (tcp_send_head(sk)) {
				if (skb->ip_summed == CHECKSUM_NONE)
					max = mss_now;
				copy = max - skb->len;
			}
L
Linus Torvalds 已提交
1154

1155
			if (copy <= 0) {
L
Linus Torvalds 已提交
1156 1157 1158 1159 1160 1161 1162
new_segment:
				/* Allocate new segment. If the interface is SG,
				 * allocate skb fitting to single page.
				 */
				if (!sk_stream_memory_free(sk))
					goto wait_for_sndbuf;

K
Krishna Kumar 已提交
1163 1164 1165
				skb = sk_stream_alloc_skb(sk,
							  select_size(sk, sg),
							  sk->sk_allocation);
L
Linus Torvalds 已提交
1166 1167 1168
				if (!skb)
					goto wait_for_memory;

1169 1170 1171 1172 1173 1174 1175
				/*
				 * All packets are restored as if they have
				 * already been sent.
				 */
				if (tp->repair)
					TCP_SKB_CB(skb)->when = tcp_time_stamp;

L
Linus Torvalds 已提交
1176 1177 1178
				/*
				 * Check whether we can use HW checksum.
				 */
1179
				if (sk->sk_route_caps & NETIF_F_ALL_CSUM)
1180
					skb->ip_summed = CHECKSUM_PARTIAL;
L
Linus Torvalds 已提交
1181

1182
				skb_entail(sk, skb);
1183
				copy = size_goal;
1184
				max = size_goal;
L
Linus Torvalds 已提交
1185 1186 1187 1188 1189 1190 1191
			}

			/* Try to append data to the end of skb. */
			if (copy > seglen)
				copy = seglen;

			/* Where to copy to? */
1192
			if (skb_availroom(skb) > 0) {
L
Linus Torvalds 已提交
1193
				/* We have some space in skb head. Superb! */
1194
				copy = min_t(int, copy, skb_availroom(skb));
1195 1196
				err = skb_add_data_nocache(sk, skb, from, copy);
				if (err)
L
Linus Torvalds 已提交
1197 1198
					goto do_fault;
			} else {
1199
				bool merge = true;
L
Linus Torvalds 已提交
1200
				int i = skb_shinfo(skb)->nr_frags;
1201 1202 1203 1204 1205 1206 1207 1208 1209 1210
				struct page_frag *pfrag = sk_page_frag(sk);

				if (!sk_page_frag_refill(sk, pfrag))
					goto wait_for_memory;

				if (!skb_can_coalesce(skb, i, pfrag->page,
						      pfrag->offset)) {
					if (i == MAX_SKB_FRAGS || !sg) {
						tcp_mark_push(tp, skb);
						goto new_segment;
L
Linus Torvalds 已提交
1211
					}
1212 1213
					merge = false;
				}
1214

1215
				copy = min_t(int, copy, pfrag->size - pfrag->offset);
1216

1217
				if (!sk_wmem_schedule(sk, copy))
1218
					goto wait_for_memory;
L
Linus Torvalds 已提交
1219

1220
				err = skb_copy_to_page_nocache(sk, from, skb,
1221 1222 1223 1224
							       pfrag->page,
							       pfrag->offset,
							       copy);
				if (err)
L
Linus Torvalds 已提交
1225 1226 1227 1228
					goto do_error;

				/* Update the skb. */
				if (merge) {
E
Eric Dumazet 已提交
1229
					skb_frag_size_add(&skb_shinfo(skb)->frags[i - 1], copy);
L
Linus Torvalds 已提交
1230
				} else {
1231 1232 1233
					skb_fill_page_desc(skb, i, pfrag->page,
							   pfrag->offset, copy);
					get_page(pfrag->page);
L
Linus Torvalds 已提交
1234
				}
1235
				pfrag->offset += copy;
L
Linus Torvalds 已提交
1236 1237 1238
			}

			if (!copied)
E
Eric Dumazet 已提交
1239
				TCP_SKB_CB(skb)->tcp_flags &= ~TCPHDR_PSH;
L
Linus Torvalds 已提交
1240 1241 1242

			tp->write_seq += copy;
			TCP_SKB_CB(skb)->end_seq += copy;
1243
			skb_shinfo(skb)->gso_segs = 0;
L
Linus Torvalds 已提交
1244 1245 1246 1247 1248 1249

			from += copy;
			copied += copy;
			if ((seglen -= copy) == 0 && iovlen == 0)
				goto out;

P
Pavel Emelyanov 已提交
1250
			if (skb->len < max || (flags & MSG_OOB) || unlikely(tp->repair))
L
Linus Torvalds 已提交
1251 1252 1253 1254
				continue;

			if (forced_push(tp)) {
				tcp_mark_push(tp, skb);
1255
				__tcp_push_pending_frames(sk, mss_now, TCP_NAGLE_PUSH);
1256
			} else if (skb == tcp_send_head(sk))
L
Linus Torvalds 已提交
1257 1258 1259 1260 1261 1262
				tcp_push_one(sk, mss_now);
			continue;

wait_for_sndbuf:
			set_bit(SOCK_NOSPACE, &sk->sk_socket->flags);
wait_for_memory:
1263
			if (copied)
E
Eric Dumazet 已提交
1264 1265
				tcp_push(sk, flags & ~MSG_MORE, mss_now,
					 TCP_NAGLE_PUSH, size_goal);
L
Linus Torvalds 已提交
1266 1267 1268 1269

			if ((err = sk_stream_wait_memory(sk, &timeo)) != 0)
				goto do_error;

I
Ilpo Järvinen 已提交
1270
			mss_now = tcp_send_mss(sk, &size_goal, flags);
L
Linus Torvalds 已提交
1271 1272 1273 1274
		}
	}

out:
1275
	if (copied)
E
Eric Dumazet 已提交
1276
		tcp_push(sk, flags, mss_now, tp->nonagle, size_goal);
L
Linus Torvalds 已提交
1277
	release_sock(sk);
1278
	return copied + copied_syn;
L
Linus Torvalds 已提交
1279 1280 1281

do_fault:
	if (!skb->len) {
1282 1283 1284 1285 1286
		tcp_unlink_write_queue(skb, sk);
		/* It is the one place in all of TCP, except connection
		 * reset, where we can be unlinking the send_head.
		 */
		tcp_check_send_head(sk, skb);
1287
		sk_wmem_free_skb(sk, skb);
L
Linus Torvalds 已提交
1288 1289 1290
	}

do_error:
1291
	if (copied + copied_syn)
L
Linus Torvalds 已提交
1292 1293 1294 1295 1296 1297
		goto out;
out_err:
	err = sk_stream_error(sk, flags, err);
	release_sock(sk);
	return err;
}
E
Eric Dumazet 已提交
1298
EXPORT_SYMBOL(tcp_sendmsg);
L
Linus Torvalds 已提交
1299 1300 1301 1302 1303 1304

/*
 *	Handle reading urgent data. BSD has very simple semantics for
 *	this, no blocking and very strange errors 8)
 */

1305
static int tcp_recv_urg(struct sock *sk, struct msghdr *msg, int len, int flags)
L
Linus Torvalds 已提交
1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348
{
	struct tcp_sock *tp = tcp_sk(sk);

	/* No URG data to read. */
	if (sock_flag(sk, SOCK_URGINLINE) || !tp->urg_data ||
	    tp->urg_data == TCP_URG_READ)
		return -EINVAL;	/* Yes this is right ! */

	if (sk->sk_state == TCP_CLOSE && !sock_flag(sk, SOCK_DONE))
		return -ENOTCONN;

	if (tp->urg_data & TCP_URG_VALID) {
		int err = 0;
		char c = tp->urg_data;

		if (!(flags & MSG_PEEK))
			tp->urg_data = TCP_URG_READ;

		/* Read urgent data. */
		msg->msg_flags |= MSG_OOB;

		if (len > 0) {
			if (!(flags & MSG_TRUNC))
				err = memcpy_toiovec(msg->msg_iov, &c, 1);
			len = 1;
		} else
			msg->msg_flags |= MSG_TRUNC;

		return err ? -EFAULT : len;
	}

	if (sk->sk_state == TCP_CLOSE || (sk->sk_shutdown & RCV_SHUTDOWN))
		return 0;

	/* Fixed the recv(..., MSG_OOB) behaviour.  BSD docs and
	 * the available implementations agree in this case:
	 * this call should never block, independent of the
	 * blocking state of the socket.
	 * Mike <pall@rz.uni-karlsruhe.de>
	 */
	return -EAGAIN;
}

P
Pavel Emelyanov 已提交
1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366
static int tcp_peek_sndq(struct sock *sk, struct msghdr *msg, int len)
{
	struct sk_buff *skb;
	int copied = 0, err = 0;

	/* XXX -- need to support SO_PEEK_OFF */

	skb_queue_walk(&sk->sk_write_queue, skb) {
		err = skb_copy_datagram_iovec(skb, 0, msg->msg_iov, skb->len);
		if (err)
			break;

		copied += skb->len;
	}

	return err ?: copied;
}

L
Linus Torvalds 已提交
1367 1368 1369 1370 1371 1372
/* Clean up the receive buffer for full frames taken by the user,
 * then send an ACK if necessary.  COPIED is the number of bytes
 * tcp_recvmsg has given to the user so far, it speeds up the
 * calculation of whether or not we must ACK for the sake of
 * a window update.
 */
1373
void tcp_cleanup_rbuf(struct sock *sk, int copied)
L
Linus Torvalds 已提交
1374 1375
{
	struct tcp_sock *tp = tcp_sk(sk);
E
Eric Dumazet 已提交
1376
	bool time_to_ack = false;
L
Linus Torvalds 已提交
1377 1378 1379

	struct sk_buff *skb = skb_peek(&sk->sk_receive_queue);

1380
	WARN(skb && !before(tp->copied_seq, TCP_SKB_CB(skb)->end_seq),
J
Joe Perches 已提交
1381
	     "cleanup rbuf bug: copied %X seq %X rcvnxt %X\n",
1382
	     tp->copied_seq, TCP_SKB_CB(skb)->end_seq, tp->rcv_nxt);
L
Linus Torvalds 已提交
1383

1384 1385
	if (inet_csk_ack_scheduled(sk)) {
		const struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
1386 1387
		   /* Delayed ACKs frequently hit locked sockets during bulk
		    * receive. */
1388
		if (icsk->icsk_ack.blocked ||
L
Linus Torvalds 已提交
1389
		    /* Once-per-two-segments ACK was not sent by tcp_input.c */
1390
		    tp->rcv_nxt - tp->rcv_wup > icsk->icsk_ack.rcv_mss ||
L
Linus Torvalds 已提交
1391 1392 1393 1394 1395 1396
		    /*
		     * If this read emptied read buffer, we send ACK, if
		     * connection is not bidirectional, user drained
		     * receive buffer and there was a small segment
		     * in queue.
		     */
1397 1398 1399 1400 1401
		    (copied > 0 &&
		     ((icsk->icsk_ack.pending & ICSK_ACK_PUSHED2) ||
		      ((icsk->icsk_ack.pending & ICSK_ACK_PUSHED) &&
		       !icsk->icsk_ack.pingpong)) &&
		      !atomic_read(&sk->sk_rmem_alloc)))
E
Eric Dumazet 已提交
1402
			time_to_ack = true;
L
Linus Torvalds 已提交
1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423
	}

	/* We send an ACK if we can now advertise a non-zero window
	 * which has been raised "significantly".
	 *
	 * Even if window raised up to infinity, do not send window open ACK
	 * in states, where we will not receive more. It is useless.
	 */
	if (copied > 0 && !time_to_ack && !(sk->sk_shutdown & RCV_SHUTDOWN)) {
		__u32 rcv_window_now = tcp_receive_window(tp);

		/* Optimize, __tcp_select_window() is not cheap. */
		if (2*rcv_window_now <= tp->window_clamp) {
			__u32 new_window = __tcp_select_window(sk);

			/* Send ACK now, if this read freed lots of space
			 * in our buffer. Certainly, new_window is new window.
			 * We can advertise it now, if it is not less than current one.
			 * "Lots" means "at least twice" here.
			 */
			if (new_window && new_window >= 2 * rcv_window_now)
E
Eric Dumazet 已提交
1424
				time_to_ack = true;
L
Linus Torvalds 已提交
1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435
		}
	}
	if (time_to_ack)
		tcp_send_ack(sk);
}

static void tcp_prequeue_process(struct sock *sk)
{
	struct sk_buff *skb;
	struct tcp_sock *tp = tcp_sk(sk);

1436
	NET_INC_STATS_USER(sock_net(sk), LINUX_MIB_TCPPREQUEUED);
L
Linus Torvalds 已提交
1437 1438 1439 1440 1441

	/* RX process wants to run with disabled BHs, though it is not
	 * necessary */
	local_bh_disable();
	while ((skb = __skb_dequeue(&tp->ucopy.prequeue)) != NULL)
P
Peter Zijlstra 已提交
1442
		sk_backlog_rcv(sk, skb);
L
Linus Torvalds 已提交
1443 1444 1445 1446 1447 1448
	local_bh_enable();

	/* Clear memory counter. */
	tp->ucopy.memory = 0;
}

1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459
#ifdef CONFIG_NET_DMA
static void tcp_service_net_dma(struct sock *sk, bool wait)
{
	dma_cookie_t done, used;
	dma_cookie_t last_issued;
	struct tcp_sock *tp = tcp_sk(sk);

	if (!tp->ucopy.dma_chan)
		return;

	last_issued = tp->ucopy.dma_cookie;
1460
	dma_async_issue_pending(tp->ucopy.dma_chan);
1461 1462

	do {
1463
		if (dma_async_is_tx_complete(tp->ucopy.dma_chan,
1464
					      last_issued, &done,
1465
					      &used) == DMA_COMPLETE) {
1466 1467 1468 1469 1470 1471 1472
			/* Safe to free early-copied skbs now */
			__skb_queue_purge(&sk->sk_async_wait_queue);
			break;
		} else {
			struct sk_buff *skb;
			while ((skb = skb_peek(&sk->sk_async_wait_queue)) &&
			       (dma_async_is_complete(skb->dma_cookie, done,
1473
						      used) == DMA_COMPLETE)) {
1474 1475 1476 1477 1478 1479 1480 1481
				__skb_dequeue(&sk->sk_async_wait_queue);
				kfree_skb(skb);
			}
		}
	} while (wait);
}
#endif

1482
static struct sk_buff *tcp_recv_skb(struct sock *sk, u32 seq, u32 *off)
L
Linus Torvalds 已提交
1483 1484 1485 1486
{
	struct sk_buff *skb;
	u32 offset;

1487
	while ((skb = skb_peek(&sk->sk_receive_queue)) != NULL) {
L
Linus Torvalds 已提交
1488
		offset = seq - TCP_SKB_CB(skb)->seq;
1489
		if (tcp_hdr(skb)->syn)
L
Linus Torvalds 已提交
1490
			offset--;
1491
		if (offset < skb->len || tcp_hdr(skb)->fin) {
L
Linus Torvalds 已提交
1492 1493 1494
			*off = offset;
			return skb;
		}
1495 1496 1497 1498 1499
		/* This looks weird, but this can happen if TCP collapsing
		 * splitted a fat GRO packet, while we released socket lock
		 * in skb_splice_bits()
		 */
		sk_eat_skb(sk, skb, false);
L
Linus Torvalds 已提交
1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527
	}
	return NULL;
}

/*
 * This routine provides an alternative to tcp_recvmsg() for routines
 * that would like to handle copying from skbuffs directly in 'sendfile'
 * fashion.
 * Note:
 *	- It is assumed that the socket was locked by the caller.
 *	- The routine does not block.
 *	- At present, there is no support for reading OOB data
 *	  or for 'peeking' the socket using this routine
 *	  (although both would be easy to implement).
 */
int tcp_read_sock(struct sock *sk, read_descriptor_t *desc,
		  sk_read_actor_t recv_actor)
{
	struct sk_buff *skb;
	struct tcp_sock *tp = tcp_sk(sk);
	u32 seq = tp->copied_seq;
	u32 offset;
	int copied = 0;

	if (sk->sk_state == TCP_LISTEN)
		return -ENOTCONN;
	while ((skb = tcp_recv_skb(sk, seq, &offset)) != NULL) {
		if (offset < skb->len) {
1528 1529
			int used;
			size_t len;
L
Linus Torvalds 已提交
1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540

			len = skb->len - offset;
			/* Stop reading if we hit a patch of urgent data */
			if (tp->urg_data) {
				u32 urg_offset = tp->urg_seq - seq;
				if (urg_offset < len)
					len = urg_offset;
				if (!len)
					break;
			}
			used = recv_actor(desc, skb, offset, len);
1541
			if (used <= 0) {
1542 1543 1544 1545
				if (!copied)
					copied = used;
				break;
			} else if (used <= len) {
L
Linus Torvalds 已提交
1546 1547 1548 1549
				seq += used;
				copied += used;
				offset += used;
			}
1550
			/* If recv_actor drops the lock (e.g. TCP splice
1551 1552 1553 1554
			 * receive) the skb pointer might be invalid when
			 * getting here: tcp_collapse might have deleted it
			 * while aggregating skbs from the socket queue.
			 */
1555 1556
			skb = tcp_recv_skb(sk, seq - 1, &offset);
			if (!skb)
L
Linus Torvalds 已提交
1557
				break;
1558 1559 1560 1561 1562
			/* TCP coalescing might have appended data to the skb.
			 * Try to splice more frags
			 */
			if (offset + 1 != skb->len)
				continue;
L
Linus Torvalds 已提交
1563
		}
1564
		if (tcp_hdr(skb)->fin) {
E
Eric Dumazet 已提交
1565
			sk_eat_skb(sk, skb, false);
L
Linus Torvalds 已提交
1566 1567 1568
			++seq;
			break;
		}
E
Eric Dumazet 已提交
1569
		sk_eat_skb(sk, skb, false);
L
Linus Torvalds 已提交
1570 1571
		if (!desc->count)
			break;
1572
		tp->copied_seq = seq;
L
Linus Torvalds 已提交
1573 1574 1575 1576 1577 1578
	}
	tp->copied_seq = seq;

	tcp_rcv_space_adjust(sk);

	/* Clean up data we have read: This will do ACK frames. */
1579 1580
	if (copied > 0) {
		tcp_recv_skb(sk, seq, &offset);
1581
		tcp_cleanup_rbuf(sk, copied);
1582
	}
L
Linus Torvalds 已提交
1583 1584
	return copied;
}
E
Eric Dumazet 已提交
1585
EXPORT_SYMBOL(tcp_read_sock);
L
Linus Torvalds 已提交
1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606

/*
 *	This routine copies from a sock struct into the user buffer.
 *
 *	Technical note: in 2.3 we work on _locked_ socket, so that
 *	tricks with *seq access order and skb->users are not required.
 *	Probably, code can be easily improved even more.
 */

int tcp_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
		size_t len, int nonblock, int flags, int *addr_len)
{
	struct tcp_sock *tp = tcp_sk(sk);
	int copied = 0;
	u32 peek_seq;
	u32 *seq;
	unsigned long used;
	int err;
	int target;		/* Read at least this many bytes */
	long timeo;
	struct task_struct *user_recv = NULL;
E
Eric Dumazet 已提交
1607
	bool copied_early = false;
1608
	struct sk_buff *skb;
I
Ilpo Järvinen 已提交
1609
	u32 urg_hole = 0;
L
Linus Torvalds 已提交
1610

1611 1612 1613
	if (sk_can_busy_loop(sk) && skb_queue_empty(&sk->sk_receive_queue) &&
	    (sk->sk_state == TCP_ESTABLISHED))
		sk_busy_loop(sk, nonblock);
1614

L
Linus Torvalds 已提交
1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626
	lock_sock(sk);

	err = -ENOTCONN;
	if (sk->sk_state == TCP_LISTEN)
		goto out;

	timeo = sock_rcvtimeo(sk, nonblock);

	/* Urgent data needs to be handled specially. */
	if (flags & MSG_OOB)
		goto recv_urg;

P
Pavel Emelyanov 已提交
1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641
	if (unlikely(tp->repair)) {
		err = -EPERM;
		if (!(flags & MSG_PEEK))
			goto out;

		if (tp->repair_queue == TCP_SEND_QUEUE)
			goto recv_sndq;

		err = -EINVAL;
		if (tp->repair_queue == TCP_NO_QUEUE)
			goto out;

		/* 'common' recv queue MSG_PEEK-ing */
	}

L
Linus Torvalds 已提交
1642 1643 1644 1645 1646 1647 1648 1649
	seq = &tp->copied_seq;
	if (flags & MSG_PEEK) {
		peek_seq = tp->copied_seq;
		seq = &peek_seq;
	}

	target = sock_rcvlowat(sk, flags & MSG_WAITALL, len);

C
Chris Leech 已提交
1650 1651 1652
#ifdef CONFIG_NET_DMA
	tp->ucopy.dma_chan = NULL;
	preempt_disable();
1653
	skb = skb_peek_tail(&sk->sk_receive_queue);
A
Andrew Morton 已提交
1654 1655 1656 1657 1658 1659 1660 1661
	{
		int available = 0;

		if (skb)
			available = TCP_SKB_CB(skb)->seq + skb->len - (*seq);
		if ((available < target) &&
		    (len > sysctl_tcp_dma_copybreak) && !(flags & MSG_PEEK) &&
		    !sysctl_tcp_low_latency &&
1662
		    net_dma_find_channel()) {
A
Andrew Morton 已提交
1663 1664 1665 1666 1667 1668 1669
			preempt_enable_no_resched();
			tp->ucopy.pinned_list =
					dma_pin_iovec_pages(msg->msg_iov, len);
		} else {
			preempt_enable_no_resched();
		}
	}
C
Chris Leech 已提交
1670 1671
#endif

L
Linus Torvalds 已提交
1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686
	do {
		u32 offset;

		/* Are we at urgent data? Stop if we have read anything or have SIGURG pending. */
		if (tp->urg_data && tp->urg_seq == *seq) {
			if (copied)
				break;
			if (signal_pending(current)) {
				copied = timeo ? sock_intr_errno(timeo) : -EAGAIN;
				break;
			}
		}

		/* Next get a buffer. */

1687
		skb_queue_walk(&sk->sk_receive_queue, skb) {
L
Linus Torvalds 已提交
1688 1689 1690
			/* Now that we have two receive queues this
			 * shouldn't happen.
			 */
1691
			if (WARN(before(*seq, TCP_SKB_CB(skb)->seq),
J
Joe Perches 已提交
1692 1693 1694
				 "recvmsg bug: copied %X seq %X rcvnxt %X fl %X\n",
				 *seq, TCP_SKB_CB(skb)->seq, tp->rcv_nxt,
				 flags))
L
Linus Torvalds 已提交
1695
				break;
1696

L
Linus Torvalds 已提交
1697
			offset = *seq - TCP_SKB_CB(skb)->seq;
1698
			if (tcp_hdr(skb)->syn)
L
Linus Torvalds 已提交
1699 1700 1701
				offset--;
			if (offset < skb->len)
				goto found_ok_skb;
1702
			if (tcp_hdr(skb)->fin)
L
Linus Torvalds 已提交
1703
				goto found_fin_ok;
J
Joe Perches 已提交
1704 1705 1706
			WARN(!(flags & MSG_PEEK),
			     "recvmsg bug 2: copied %X seq %X rcvnxt %X fl %X\n",
			     *seq, TCP_SKB_CB(skb)->seq, tp->rcv_nxt, flags);
1707
		}
L
Linus Torvalds 已提交
1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718

		/* Well, if we have backlog, try to process it now yet. */

		if (copied >= target && !sk->sk_backlog.tail)
			break;

		if (copied) {
			if (sk->sk_err ||
			    sk->sk_state == TCP_CLOSE ||
			    (sk->sk_shutdown & RCV_SHUTDOWN) ||
			    !timeo ||
1719
			    signal_pending(current))
L
Linus Torvalds 已提交
1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754
				break;
		} else {
			if (sock_flag(sk, SOCK_DONE))
				break;

			if (sk->sk_err) {
				copied = sock_error(sk);
				break;
			}

			if (sk->sk_shutdown & RCV_SHUTDOWN)
				break;

			if (sk->sk_state == TCP_CLOSE) {
				if (!sock_flag(sk, SOCK_DONE)) {
					/* This occurs when user tries to read
					 * from never connected socket.
					 */
					copied = -ENOTCONN;
					break;
				}
				break;
			}

			if (!timeo) {
				copied = -EAGAIN;
				break;
			}

			if (signal_pending(current)) {
				copied = sock_intr_errno(timeo);
				break;
			}
		}

1755
		tcp_cleanup_rbuf(sk, copied);
L
Linus Torvalds 已提交
1756

1757
		if (!sysctl_tcp_low_latency && tp->ucopy.task == user_recv) {
L
Linus Torvalds 已提交
1758 1759 1760 1761 1762 1763 1764 1765 1766
			/* Install new reader */
			if (!user_recv && !(flags & (MSG_TRUNC | MSG_PEEK))) {
				user_recv = current;
				tp->ucopy.task = user_recv;
				tp->ucopy.iov = msg->msg_iov;
			}

			tp->ucopy.len = len;

1767 1768
			WARN_ON(tp->copied_seq != tp->rcv_nxt &&
				!(flags & (MSG_PEEK | MSG_TRUNC)));
L
Linus Torvalds 已提交
1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795

			/* Ugly... If prequeue is not empty, we have to
			 * process it before releasing socket, otherwise
			 * order will be broken at second iteration.
			 * More elegant solution is required!!!
			 *
			 * Look: we have the following (pseudo)queues:
			 *
			 * 1. packets in flight
			 * 2. backlog
			 * 3. prequeue
			 * 4. receive_queue
			 *
			 * Each queue can be processed only if the next ones
			 * are empty. At this point we have empty receive_queue.
			 * But prequeue _can_ be not empty after 2nd iteration,
			 * when we jumped to start of loop because backlog
			 * processing added something to receive_queue.
			 * We cannot release_sock(), because backlog contains
			 * packets arrived _after_ prequeued ones.
			 *
			 * Shortly, algorithm is clear --- to process all
			 * the queues in order. We could make it more directly,
			 * requeueing packets from backlog to prequeue, if
			 * is not empty. It is more elegant, but eats cycles,
			 * unfortunately.
			 */
1796
			if (!skb_queue_empty(&tp->ucopy.prequeue))
L
Linus Torvalds 已提交
1797 1798 1799 1800 1801
				goto do_prequeue;

			/* __ Set realtime policy in scheduler __ */
		}

1802
#ifdef CONFIG_NET_DMA
1803 1804 1805 1806 1807 1808
		if (tp->ucopy.dma_chan) {
			if (tp->rcv_wnd == 0 &&
			    !skb_queue_empty(&sk->sk_async_wait_queue)) {
				tcp_service_net_dma(sk, true);
				tcp_cleanup_rbuf(sk, copied);
			} else
1809
				dma_async_issue_pending(tp->ucopy.dma_chan);
1810
		}
1811
#endif
L
Linus Torvalds 已提交
1812 1813 1814 1815 1816 1817 1818
		if (copied >= target) {
			/* Do not sleep, just process backlog. */
			release_sock(sk);
			lock_sock(sk);
		} else
			sk_wait_data(sk, &timeo);

C
Chris Leech 已提交
1819
#ifdef CONFIG_NET_DMA
1820
		tcp_service_net_dma(sk, false);  /* Don't block */
C
Chris Leech 已提交
1821 1822 1823
		tp->ucopy.wakeup = 0;
#endif

L
Linus Torvalds 已提交
1824 1825 1826 1827 1828 1829
		if (user_recv) {
			int chunk;

			/* __ Restore normal policy in scheduler __ */

			if ((chunk = len - tp->ucopy.len) != 0) {
1830
				NET_ADD_STATS_USER(sock_net(sk), LINUX_MIB_TCPDIRECTCOPYFROMBACKLOG, chunk);
L
Linus Torvalds 已提交
1831 1832 1833 1834 1835
				len -= chunk;
				copied += chunk;
			}

			if (tp->rcv_nxt == tp->copied_seq &&
1836
			    !skb_queue_empty(&tp->ucopy.prequeue)) {
L
Linus Torvalds 已提交
1837 1838 1839 1840
do_prequeue:
				tcp_prequeue_process(sk);

				if ((chunk = len - tp->ucopy.len) != 0) {
1841
					NET_ADD_STATS_USER(sock_net(sk), LINUX_MIB_TCPDIRECTCOPYFROMPREQUEUE, chunk);
L
Linus Torvalds 已提交
1842 1843 1844 1845 1846
					len -= chunk;
					copied += chunk;
				}
			}
		}
I
Ilpo Järvinen 已提交
1847 1848
		if ((flags & MSG_PEEK) &&
		    (peek_seq - copied - urg_hole != tp->copied_seq)) {
1849 1850 1851
			net_dbg_ratelimited("TCP(%s:%d): Application bug, race in MSG_PEEK\n",
					    current->comm,
					    task_pid_nr(current));
L
Linus Torvalds 已提交
1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868
			peek_seq = tp->copied_seq;
		}
		continue;

	found_ok_skb:
		/* Ok so how much can we use? */
		used = skb->len - offset;
		if (len < used)
			used = len;

		/* Do we have urgent data here? */
		if (tp->urg_data) {
			u32 urg_offset = tp->urg_seq - *seq;
			if (urg_offset < used) {
				if (!urg_offset) {
					if (!sock_flag(sk, SOCK_URGINLINE)) {
						++*seq;
I
Ilpo Järvinen 已提交
1869
						urg_hole++;
L
Linus Torvalds 已提交
1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880
						offset++;
						used--;
						if (!used)
							goto skip_copy;
					}
				} else
					used = urg_offset;
			}
		}

		if (!(flags & MSG_TRUNC)) {
C
Chris Leech 已提交
1881 1882
#ifdef CONFIG_NET_DMA
			if (!tp->ucopy.dma_chan && tp->ucopy.pinned_list)
1883
				tp->ucopy.dma_chan = net_dma_find_channel();
C
Chris Leech 已提交
1884 1885 1886 1887 1888 1889 1890 1891 1892

			if (tp->ucopy.dma_chan) {
				tp->ucopy.dma_cookie = dma_skb_copy_datagram_iovec(
					tp->ucopy.dma_chan, skb, offset,
					msg->msg_iov, used,
					tp->ucopy.pinned_list);

				if (tp->ucopy.dma_cookie < 0) {

1893 1894
					pr_alert("%s: dma_cookie < 0\n",
						 __func__);
C
Chris Leech 已提交
1895 1896 1897 1898 1899 1900

					/* Exception. Bailout! */
					if (!copied)
						copied = -EFAULT;
					break;
				}
1901

1902
				dma_async_issue_pending(tp->ucopy.dma_chan);
1903

C
Chris Leech 已提交
1904
				if ((offset + used) == skb->len)
E
Eric Dumazet 已提交
1905
					copied_early = true;
C
Chris Leech 已提交
1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917

			} else
#endif
			{
				err = skb_copy_datagram_iovec(skb, offset,
						msg->msg_iov, used);
				if (err) {
					/* Exception. Bailout! */
					if (!copied)
						copied = -EFAULT;
					break;
				}
L
Linus Torvalds 已提交
1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929
			}
		}

		*seq += used;
		copied += used;
		len -= used;

		tcp_rcv_space_adjust(sk);

skip_copy:
		if (tp->urg_data && after(tp->copied_seq, tp->urg_seq)) {
			tp->urg_data = 0;
1930
			tcp_fast_path_check(sk);
L
Linus Torvalds 已提交
1931 1932 1933 1934
		}
		if (used + offset < skb->len)
			continue;

1935
		if (tcp_hdr(skb)->fin)
L
Linus Torvalds 已提交
1936
			goto found_fin_ok;
C
Chris Leech 已提交
1937 1938
		if (!(flags & MSG_PEEK)) {
			sk_eat_skb(sk, skb, copied_early);
E
Eric Dumazet 已提交
1939
			copied_early = false;
C
Chris Leech 已提交
1940
		}
L
Linus Torvalds 已提交
1941 1942 1943 1944 1945
		continue;

	found_fin_ok:
		/* Process the FIN. */
		++*seq;
C
Chris Leech 已提交
1946 1947
		if (!(flags & MSG_PEEK)) {
			sk_eat_skb(sk, skb, copied_early);
E
Eric Dumazet 已提交
1948
			copied_early = false;
C
Chris Leech 已提交
1949
		}
L
Linus Torvalds 已提交
1950 1951 1952 1953
		break;
	} while (len > 0);

	if (user_recv) {
1954
		if (!skb_queue_empty(&tp->ucopy.prequeue)) {
L
Linus Torvalds 已提交
1955 1956 1957 1958 1959 1960 1961
			int chunk;

			tp->ucopy.len = copied > 0 ? len : 0;

			tcp_prequeue_process(sk);

			if (copied > 0 && (chunk = len - tp->ucopy.len) != 0) {
1962
				NET_ADD_STATS_USER(sock_net(sk), LINUX_MIB_TCPDIRECTCOPYFROMPREQUEUE, chunk);
L
Linus Torvalds 已提交
1963 1964 1965 1966 1967 1968 1969 1970 1971
				len -= chunk;
				copied += chunk;
			}
		}

		tp->ucopy.task = NULL;
		tp->ucopy.len = 0;
	}

C
Chris Leech 已提交
1972
#ifdef CONFIG_NET_DMA
1973 1974
	tcp_service_net_dma(sk, true);  /* Wait for queue to drain */
	tp->ucopy.dma_chan = NULL;
C
Chris Leech 已提交
1975 1976 1977 1978 1979 1980 1981

	if (tp->ucopy.pinned_list) {
		dma_unpin_iovec_pages(tp->ucopy.pinned_list);
		tp->ucopy.pinned_list = NULL;
	}
#endif

L
Linus Torvalds 已提交
1982 1983 1984 1985 1986
	/* According to UNIX98, msg_name/msg_namelen are ignored
	 * on connected socket. I was just happy when found this 8) --ANK
	 */

	/* Clean up data we have read: This will do ACK frames. */
1987
	tcp_cleanup_rbuf(sk, copied);
L
Linus Torvalds 已提交
1988 1989 1990 1991 1992 1993 1994 1995 1996

	release_sock(sk);
	return copied;

out:
	release_sock(sk);
	return err;

recv_urg:
1997
	err = tcp_recv_urg(sk, msg, len, flags);
L
Linus Torvalds 已提交
1998
	goto out;
P
Pavel Emelyanov 已提交
1999 2000 2001 2002

recv_sndq:
	err = tcp_peek_sndq(sk, msg, len);
	goto out;
L
Linus Torvalds 已提交
2003
}
E
Eric Dumazet 已提交
2004
EXPORT_SYMBOL(tcp_recvmsg);
L
Linus Torvalds 已提交
2005

I
Ilpo Järvinen 已提交
2006 2007 2008 2009 2010 2011 2012
void tcp_set_state(struct sock *sk, int state)
{
	int oldstate = sk->sk_state;

	switch (state) {
	case TCP_ESTABLISHED:
		if (oldstate != TCP_ESTABLISHED)
P
Pavel Emelyanov 已提交
2013
			TCP_INC_STATS(sock_net(sk), TCP_MIB_CURRESTAB);
I
Ilpo Järvinen 已提交
2014 2015 2016 2017
		break;

	case TCP_CLOSE:
		if (oldstate == TCP_CLOSE_WAIT || oldstate == TCP_ESTABLISHED)
P
Pavel Emelyanov 已提交
2018
			TCP_INC_STATS(sock_net(sk), TCP_MIB_ESTABRESETS);
I
Ilpo Järvinen 已提交
2019 2020 2021 2022

		sk->sk_prot->unhash(sk);
		if (inet_csk(sk)->icsk_bind_hash &&
		    !(sk->sk_userlocks & SOCK_BINDPORT_LOCK))
2023
			inet_put_port(sk);
I
Ilpo Järvinen 已提交
2024 2025
		/* fall through */
	default:
2026
		if (oldstate == TCP_ESTABLISHED)
P
Pavel Emelyanov 已提交
2027
			TCP_DEC_STATS(sock_net(sk), TCP_MIB_CURRESTAB);
I
Ilpo Järvinen 已提交
2028 2029 2030 2031 2032 2033 2034 2035
	}

	/* Change state AFTER socket is unhashed to avoid closed
	 * socket sitting in hash tables.
	 */
	sk->sk_state = state;

#ifdef STATE_TRACE
2036
	SOCK_DEBUG(sk, "TCP sk=%p, State %s -> %s\n", sk, statename[oldstate], statename[state]);
I
Ilpo Järvinen 已提交
2037 2038 2039 2040
#endif
}
EXPORT_SYMBOL_GPL(tcp_set_state);

L
Linus Torvalds 已提交
2041 2042 2043 2044 2045 2046 2047
/*
 *	State processing on a close. This implements the state shift for
 *	sending our FIN frame. Note that we only send a FIN for some
 *	states. A shutdown() may have already sent the FIN, or we may be
 *	closed.
 */

2048
static const unsigned char new_state[16] = {
L
Linus Torvalds 已提交
2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075
  /* current state:        new state:      action:	*/
  /* (Invalid)		*/ TCP_CLOSE,
  /* TCP_ESTABLISHED	*/ TCP_FIN_WAIT1 | TCP_ACTION_FIN,
  /* TCP_SYN_SENT	*/ TCP_CLOSE,
  /* TCP_SYN_RECV	*/ TCP_FIN_WAIT1 | TCP_ACTION_FIN,
  /* TCP_FIN_WAIT1	*/ TCP_FIN_WAIT1,
  /* TCP_FIN_WAIT2	*/ TCP_FIN_WAIT2,
  /* TCP_TIME_WAIT	*/ TCP_CLOSE,
  /* TCP_CLOSE		*/ TCP_CLOSE,
  /* TCP_CLOSE_WAIT	*/ TCP_LAST_ACK  | TCP_ACTION_FIN,
  /* TCP_LAST_ACK	*/ TCP_LAST_ACK,
  /* TCP_LISTEN		*/ TCP_CLOSE,
  /* TCP_CLOSING	*/ TCP_CLOSING,
};

static int tcp_close_state(struct sock *sk)
{
	int next = (int)new_state[sk->sk_state];
	int ns = next & TCP_STATE_MASK;

	tcp_set_state(sk, ns);

	return next & TCP_ACTION_FIN;
}

/*
 *	Shutdown the sending side of a connection. Much like close except
2076
 *	that we don't receive shut down or sock_set_flag(sk, SOCK_DEAD).
L
Linus Torvalds 已提交
2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096
 */

void tcp_shutdown(struct sock *sk, int how)
{
	/*	We need to grab some memory, and put together a FIN,
	 *	and then put it into the queue to be sent.
	 *		Tim MacKenzie(tym@dibbler.cs.monash.edu.au) 4 Dec '92.
	 */
	if (!(how & SEND_SHUTDOWN))
		return;

	/* If we've already sent a FIN, or it's a closed state, skip this. */
	if ((1 << sk->sk_state) &
	    (TCPF_ESTABLISHED | TCPF_SYN_SENT |
	     TCPF_SYN_RECV | TCPF_CLOSE_WAIT)) {
		/* Clear out any half completed packets.  FIN if needed. */
		if (tcp_close_state(sk))
			tcp_send_fin(sk);
	}
}
E
Eric Dumazet 已提交
2097
EXPORT_SYMBOL(tcp_shutdown);
L
Linus Torvalds 已提交
2098

A
Arun Sharma 已提交
2099 2100 2101 2102 2103 2104 2105
bool tcp_check_oom(struct sock *sk, int shift)
{
	bool too_many_orphans, out_of_socket_memory;

	too_many_orphans = tcp_too_many_orphans(sk, shift);
	out_of_socket_memory = tcp_out_of_memory(sk);

2106 2107 2108 2109
	if (too_many_orphans)
		net_info_ratelimited("too many orphaned sockets\n");
	if (out_of_socket_memory)
		net_info_ratelimited("out of memory -- consider tuning tcp_mem\n");
A
Arun Sharma 已提交
2110 2111 2112
	return too_many_orphans || out_of_socket_memory;
}

L
Linus Torvalds 已提交
2113 2114 2115 2116
void tcp_close(struct sock *sk, long timeout)
{
	struct sk_buff *skb;
	int data_was_unread = 0;
H
Herbert Xu 已提交
2117
	int state;
L
Linus Torvalds 已提交
2118 2119 2120 2121 2122 2123 2124 2125

	lock_sock(sk);
	sk->sk_shutdown = SHUTDOWN_MASK;

	if (sk->sk_state == TCP_LISTEN) {
		tcp_set_state(sk, TCP_CLOSE);

		/* Special case. */
2126
		inet_csk_listen_stop(sk);
L
Linus Torvalds 已提交
2127 2128 2129 2130 2131 2132 2133 2134 2135 2136

		goto adjudge_to_death;
	}

	/*  We need to flush the recv. buffs.  We do this only on the
	 *  descriptor close, not protocol-sourced closes, because the
	 *  reader process may not have drained the data yet!
	 */
	while ((skb = __skb_dequeue(&sk->sk_receive_queue)) != NULL) {
		u32 len = TCP_SKB_CB(skb)->end_seq - TCP_SKB_CB(skb)->seq -
2137
			  tcp_hdr(skb)->fin;
L
Linus Torvalds 已提交
2138 2139 2140 2141
		data_was_unread += len;
		__kfree_skb(skb);
	}

2142
	sk_mem_reclaim(sk);
L
Linus Torvalds 已提交
2143

2144 2145 2146 2147
	/* If socket has been already reset (e.g. in tcp_reset()) - kill it. */
	if (sk->sk_state == TCP_CLOSE)
		goto adjudge_to_death;

2148 2149 2150 2151 2152 2153
	/* As outlined in RFC 2525, section 2.17, we send a RST here because
	 * data was lost. To witness the awful effects of the old behavior of
	 * always doing a FIN, run an older 2.1.x kernel or 2.0.x, start a bulk
	 * GET in an FTP client, suspend the process, wait for the client to
	 * advertise a zero window, then kill -9 the FTP client, wheee...
	 * Note: timeout is always zero in such a case.
L
Linus Torvalds 已提交
2154
	 */
P
Pavel Emelyanov 已提交
2155 2156 2157
	if (unlikely(tcp_sk(sk)->repair)) {
		sk->sk_prot->disconnect(sk, 0);
	} else if (data_was_unread) {
L
Linus Torvalds 已提交
2158
		/* Unread data was tossed, zap the connection. */
2159
		NET_INC_STATS_USER(sock_net(sk), LINUX_MIB_TCPABORTONCLOSE);
L
Linus Torvalds 已提交
2160
		tcp_set_state(sk, TCP_CLOSE);
2161
		tcp_send_active_reset(sk, sk->sk_allocation);
L
Linus Torvalds 已提交
2162 2163 2164
	} else if (sock_flag(sk, SOCK_LINGER) && !sk->sk_lingertime) {
		/* Check zero linger _after_ checking for unread data. */
		sk->sk_prot->disconnect(sk, 0);
2165
		NET_INC_STATS_USER(sock_net(sk), LINUX_MIB_TCPABORTONDATA);
L
Linus Torvalds 已提交
2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190
	} else if (tcp_close_state(sk)) {
		/* We FIN if the application ate all the data before
		 * zapping the connection.
		 */

		/* RED-PEN. Formally speaking, we have broken TCP state
		 * machine. State transitions:
		 *
		 * TCP_ESTABLISHED -> TCP_FIN_WAIT1
		 * TCP_SYN_RECV	-> TCP_FIN_WAIT1 (forget it, it's impossible)
		 * TCP_CLOSE_WAIT -> TCP_LAST_ACK
		 *
		 * are legal only when FIN has been sent (i.e. in window),
		 * rather than queued out of window. Purists blame.
		 *
		 * F.e. "RFC state" is ESTABLISHED,
		 * if Linux state is FIN-WAIT-1, but FIN is still not sent.
		 *
		 * The visible declinations are that sometimes
		 * we enter time-wait state, when it is not required really
		 * (harmless), do not send active resets, when they are
		 * required by specs (TCP_ESTABLISHED, TCP_CLOSE_WAIT, when
		 * they look as CLOSING or LAST_ACK for Linux)
		 * Probably, I missed some more holelets.
		 * 						--ANK
2191 2192 2193 2194
		 * XXX (TFO) - To start off we don't support SYN+ACK+FIN
		 * in a single packet! (May consider it later but will
		 * probably need API support or TCP_CORK SYN-ACK until
		 * data is written and socket is closed.)
L
Linus Torvalds 已提交
2195 2196 2197 2198 2199 2200 2201
		 */
		tcp_send_fin(sk);
	}

	sk_stream_wait_close(sk, timeout);

adjudge_to_death:
H
Herbert Xu 已提交
2202 2203 2204 2205
	state = sk->sk_state;
	sock_hold(sk);
	sock_orphan(sk);

L
Linus Torvalds 已提交
2206 2207 2208 2209 2210 2211 2212 2213 2214
	/* It is the last release_sock in its life. It will remove backlog. */
	release_sock(sk);


	/* Now socket is owned by kernel and we acquire BH lock
	   to finish close. No need to check for user refs.
	 */
	local_bh_disable();
	bh_lock_sock(sk);
2215
	WARN_ON(sock_owned_by_user(sk));
L
Linus Torvalds 已提交
2216

H
Herbert Xu 已提交
2217 2218
	percpu_counter_inc(sk->sk_prot->orphan_count);

H
Herbert Xu 已提交
2219 2220 2221
	/* Have we already been destroyed by a softirq or backlog? */
	if (state != TCP_CLOSE && sk->sk_state == TCP_CLOSE)
		goto out;
L
Linus Torvalds 已提交
2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241

	/*	This is a (useful) BSD violating of the RFC. There is a
	 *	problem with TCP as specified in that the other end could
	 *	keep a socket open forever with no application left this end.
	 *	We use a 3 minute timeout (about the same as BSD) then kill
	 *	our end. If they send after that then tough - BUT: long enough
	 *	that we won't make the old 4*rto = almost no time - whoops
	 *	reset mistake.
	 *
	 *	Nope, it was not mistake. It is really desired behaviour
	 *	f.e. on http servers, when such sockets are useless, but
	 *	consume significant resources. Let's do it with special
	 *	linger2	option.					--ANK
	 */

	if (sk->sk_state == TCP_FIN_WAIT2) {
		struct tcp_sock *tp = tcp_sk(sk);
		if (tp->linger2 < 0) {
			tcp_set_state(sk, TCP_CLOSE);
			tcp_send_active_reset(sk, GFP_ATOMIC);
2242 2243
			NET_INC_STATS_BH(sock_net(sk),
					LINUX_MIB_TCPABORTONLINGER);
L
Linus Torvalds 已提交
2244
		} else {
2245
			const int tmo = tcp_fin_time(sk);
L
Linus Torvalds 已提交
2246 2247

			if (tmo > TCP_TIMEWAIT_LEN) {
2248 2249
				inet_csk_reset_keepalive_timer(sk,
						tmo - TCP_TIMEWAIT_LEN);
L
Linus Torvalds 已提交
2250 2251 2252 2253 2254 2255 2256
			} else {
				tcp_time_wait(sk, TCP_FIN_WAIT2, tmo);
				goto out;
			}
		}
	}
	if (sk->sk_state != TCP_CLOSE) {
2257
		sk_mem_reclaim(sk);
A
Arun Sharma 已提交
2258
		if (tcp_check_oom(sk, 0)) {
L
Linus Torvalds 已提交
2259 2260
			tcp_set_state(sk, TCP_CLOSE);
			tcp_send_active_reset(sk, GFP_ATOMIC);
2261 2262
			NET_INC_STATS_BH(sock_net(sk),
					LINUX_MIB_TCPABORTONMEMORY);
L
Linus Torvalds 已提交
2263 2264 2265
		}
	}

2266 2267 2268 2269 2270 2271 2272 2273
	if (sk->sk_state == TCP_CLOSE) {
		struct request_sock *req = tcp_sk(sk)->fastopen_rsk;
		/* We could get here with a non-NULL req if the socket is
		 * aborted (e.g., closed with unread data) before 3WHS
		 * finishes.
		 */
		if (req != NULL)
			reqsk_fastopen_remove(sk, req, false);
2274
		inet_csk_destroy_sock(sk);
2275
	}
L
Linus Torvalds 已提交
2276 2277 2278 2279 2280 2281 2282
	/* Otherwise, socket is reprieved until protocol close. */

out:
	bh_unlock_sock(sk);
	local_bh_enable();
	sock_put(sk);
}
E
Eric Dumazet 已提交
2283
EXPORT_SYMBOL(tcp_close);
L
Linus Torvalds 已提交
2284 2285 2286

/* These states need RST on ABORT according to RFC793 */

E
Eric Dumazet 已提交
2287
static inline bool tcp_need_reset(int state)
L
Linus Torvalds 已提交
2288 2289 2290 2291 2292 2293 2294 2295 2296
{
	return (1 << state) &
	       (TCPF_ESTABLISHED | TCPF_CLOSE_WAIT | TCPF_FIN_WAIT1 |
		TCPF_FIN_WAIT2 | TCPF_SYN_RECV);
}

int tcp_disconnect(struct sock *sk, int flags)
{
	struct inet_sock *inet = inet_sk(sk);
2297
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
2298 2299 2300 2301 2302 2303 2304 2305 2306
	struct tcp_sock *tp = tcp_sk(sk);
	int err = 0;
	int old_state = sk->sk_state;

	if (old_state != TCP_CLOSE)
		tcp_set_state(sk, TCP_CLOSE);

	/* ABORT function of RFC793 */
	if (old_state == TCP_LISTEN) {
2307
		inet_csk_listen_stop(sk);
P
Pavel Emelyanov 已提交
2308 2309
	} else if (unlikely(tp->repair)) {
		sk->sk_err = ECONNABORTED;
L
Linus Torvalds 已提交
2310 2311 2312
	} else if (tcp_need_reset(old_state) ||
		   (tp->snd_nxt != tp->write_seq &&
		    (1 << old_state) & (TCPF_CLOSING | TCPF_LAST_ACK))) {
S
Stephen Hemminger 已提交
2313
		/* The last check adjusts for discrepancy of Linux wrt. RFC
L
Linus Torvalds 已提交
2314 2315 2316 2317 2318 2319 2320 2321 2322
		 * states
		 */
		tcp_send_active_reset(sk, gfp_any());
		sk->sk_err = ECONNRESET;
	} else if (old_state == TCP_SYN_SENT)
		sk->sk_err = ECONNRESET;

	tcp_clear_xmit_timers(sk);
	__skb_queue_purge(&sk->sk_receive_queue);
2323
	tcp_write_queue_purge(sk);
L
Linus Torvalds 已提交
2324
	__skb_queue_purge(&tp->out_of_order_queue);
C
Chris Leech 已提交
2325 2326 2327
#ifdef CONFIG_NET_DMA
	__skb_queue_purge(&sk->sk_async_wait_queue);
#endif
L
Linus Torvalds 已提交
2328

E
Eric Dumazet 已提交
2329
	inet->inet_dport = 0;
L
Linus Torvalds 已提交
2330 2331 2332 2333 2334 2335 2336 2337 2338

	if (!(sk->sk_userlocks & SOCK_BINDADDR_LOCK))
		inet_reset_saddr(sk);

	sk->sk_shutdown = 0;
	sock_reset_flag(sk, SOCK_DONE);
	tp->srtt = 0;
	if ((tp->write_seq += tp->max_window + 2) == 0)
		tp->write_seq = 1;
2339
	icsk->icsk_backoff = 0;
L
Linus Torvalds 已提交
2340
	tp->snd_cwnd = 2;
2341
	icsk->icsk_probes_out = 0;
L
Linus Torvalds 已提交
2342
	tp->packets_out = 0;
I
Ilpo Järvinen 已提交
2343
	tp->snd_ssthresh = TCP_INFINITE_SSTHRESH;
L
Linus Torvalds 已提交
2344
	tp->snd_cwnd_cnt = 0;
2345
	tp->window_clamp = 0;
2346
	tcp_set_ca_state(sk, TCP_CA_Open);
L
Linus Torvalds 已提交
2347
	tcp_clear_retrans(tp);
2348
	inet_csk_delack_init(sk);
2349
	tcp_init_send_head(sk);
2350
	memset(&tp->rx_opt, 0, sizeof(tp->rx_opt));
L
Linus Torvalds 已提交
2351 2352
	__sk_dst_reset(sk);

E
Eric Dumazet 已提交
2353
	WARN_ON(inet->inet_num && !icsk->icsk_bind_hash);
L
Linus Torvalds 已提交
2354 2355 2356 2357

	sk->sk_error_report(sk);
	return err;
}
E
Eric Dumazet 已提交
2358
EXPORT_SYMBOL(tcp_disconnect);
L
Linus Torvalds 已提交
2359

2360 2361 2362 2363 2364 2365 2366
void tcp_sock_destruct(struct sock *sk)
{
	inet_sock_destruct(sk);

	kfree(inet_csk(sk)->icsk_accept_queue.fastopenq);
}

E
Eric Dumazet 已提交
2367
static inline bool tcp_can_repair_sock(const struct sock *sk)
P
Pavel Emelyanov 已提交
2368
{
2369
	return ns_capable(sock_net(sk)->user_ns, CAP_NET_ADMIN) &&
P
Pavel Emelyanov 已提交
2370 2371 2372
		((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_ESTABLISHED));
}

2373 2374
static int tcp_repair_options_est(struct tcp_sock *tp,
		struct tcp_repair_opt __user *optbuf, unsigned int len)
2375
{
2376
	struct tcp_repair_opt opt;
2377

2378 2379
	while (len >= sizeof(opt)) {
		if (copy_from_user(&opt, optbuf, sizeof(opt)))
2380 2381 2382
			return -EFAULT;

		optbuf++;
2383
		len -= sizeof(opt);
2384

2385 2386 2387
		switch (opt.opt_code) {
		case TCPOPT_MSS:
			tp->rx_opt.mss_clamp = opt.opt_val;
2388
			break;
2389
		case TCPOPT_WINDOW:
2390 2391 2392 2393 2394 2395
			{
				u16 snd_wscale = opt.opt_val & 0xFFFF;
				u16 rcv_wscale = opt.opt_val >> 16;

				if (snd_wscale > 14 || rcv_wscale > 14)
					return -EFBIG;
2396

2397 2398 2399 2400
				tp->rx_opt.snd_wscale = snd_wscale;
				tp->rx_opt.rcv_wscale = rcv_wscale;
				tp->rx_opt.wscale_ok = 1;
			}
2401 2402
			break;
		case TCPOPT_SACK_PERM:
2403 2404 2405
			if (opt.opt_val != 0)
				return -EINVAL;

2406 2407 2408 2409 2410
			tp->rx_opt.sack_ok |= TCP_SACK_SEEN;
			if (sysctl_tcp_fack)
				tcp_enable_fack(tp);
			break;
		case TCPOPT_TIMESTAMP:
2411 2412 2413
			if (opt.opt_val != 0)
				return -EINVAL;

2414 2415 2416 2417 2418 2419 2420 2421
			tp->rx_opt.tstamp_ok = 1;
			break;
		}
	}

	return 0;
}

L
Linus Torvalds 已提交
2422 2423 2424
/*
 *	Socket option code for TCP.
 */
2425
static int do_tcp_setsockopt(struct sock *sk, int level,
2426
		int optname, char __user *optval, unsigned int optlen)
L
Linus Torvalds 已提交
2427 2428
{
	struct tcp_sock *tp = tcp_sk(sk);
2429
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
2430 2431 2432
	int val;
	int err = 0;

2433 2434 2435
	/* These are data/string values, all the others are ints */
	switch (optname) {
	case TCP_CONGESTION: {
2436 2437 2438 2439 2440 2441
		char name[TCP_CA_NAME_MAX];

		if (optlen < 1)
			return -EINVAL;

		val = strncpy_from_user(name, optval,
2442
					min_t(long, TCP_CA_NAME_MAX-1, optlen));
2443 2444 2445 2446 2447
		if (val < 0)
			return -EFAULT;
		name[val] = 0;

		lock_sock(sk);
2448
		err = tcp_set_congestion_control(sk, name);
2449 2450 2451
		release_sock(sk);
		return err;
	}
2452 2453 2454
	default:
		/* fallthru */
		break;
2455
	}
2456

L
Linus Torvalds 已提交
2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469
	if (optlen < sizeof(int))
		return -EINVAL;

	if (get_user(val, (int __user *)optval))
		return -EFAULT;

	lock_sock(sk);

	switch (optname) {
	case TCP_MAXSEG:
		/* Values greater than interface MTU won't take effect. However
		 * at the point when this call is done we typically don't yet
		 * know which interface is going to be used */
2470
		if (val < TCP_MIN_MSS || val > MAX_TCP_WINDOW) {
L
Linus Torvalds 已提交
2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487
			err = -EINVAL;
			break;
		}
		tp->rx_opt.user_mss = val;
		break;

	case TCP_NODELAY:
		if (val) {
			/* TCP_NODELAY is weaker than TCP_CORK, so that
			 * this option on corked socket is remembered, but
			 * it is not activated until cork is cleared.
			 *
			 * However, when TCP_NODELAY is set we make
			 * an explicit push, which overrides even TCP_CORK
			 * for currently queued segments.
			 */
			tp->nonagle |= TCP_NAGLE_OFF|TCP_NAGLE_PUSH;
2488
			tcp_push_pending_frames(sk);
L
Linus Torvalds 已提交
2489 2490 2491 2492 2493
		} else {
			tp->nonagle &= ~TCP_NAGLE_OFF;
		}
		break;

A
Andreas Petlund 已提交
2494 2495 2496 2497 2498 2499 2500
	case TCP_THIN_LINEAR_TIMEOUTS:
		if (val < 0 || val > 1)
			err = -EINVAL;
		else
			tp->thin_lto = val;
		break;

A
Andreas Petlund 已提交
2501 2502 2503
	case TCP_THIN_DUPACK:
		if (val < 0 || val > 1)
			err = -EINVAL;
2504
		else {
A
Andreas Petlund 已提交
2505
			tp->thin_dupack = val;
Y
Yuchung Cheng 已提交
2506 2507
			if (tp->thin_dupack)
				tcp_disable_early_retrans(tp);
2508
		}
A
Andreas Petlund 已提交
2509 2510
		break;

P
Pavel Emelyanov 已提交
2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546
	case TCP_REPAIR:
		if (!tcp_can_repair_sock(sk))
			err = -EPERM;
		else if (val == 1) {
			tp->repair = 1;
			sk->sk_reuse = SK_FORCE_REUSE;
			tp->repair_queue = TCP_NO_QUEUE;
		} else if (val == 0) {
			tp->repair = 0;
			sk->sk_reuse = SK_NO_REUSE;
			tcp_send_window_probe(sk);
		} else
			err = -EINVAL;

		break;

	case TCP_REPAIR_QUEUE:
		if (!tp->repair)
			err = -EPERM;
		else if (val < TCP_QUEUES_NR)
			tp->repair_queue = val;
		else
			err = -EINVAL;
		break;

	case TCP_QUEUE_SEQ:
		if (sk->sk_state != TCP_CLOSE)
			err = -EPERM;
		else if (tp->repair_queue == TCP_SEND_QUEUE)
			tp->write_seq = val;
		else if (tp->repair_queue == TCP_RECV_QUEUE)
			tp->rcv_nxt = val;
		else
			err = -EINVAL;
		break;

2547 2548 2549 2550
	case TCP_REPAIR_OPTIONS:
		if (!tp->repair)
			err = -EINVAL;
		else if (sk->sk_state == TCP_ESTABLISHED)
2551 2552 2553
			err = tcp_repair_options_est(tp,
					(struct tcp_repair_opt __user *)optval,
					optlen);
2554 2555 2556 2557
		else
			err = -EPERM;
		break;

L
Linus Torvalds 已提交
2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575
	case TCP_CORK:
		/* When set indicates to always queue non-full frames.
		 * Later the user clears this option and we transmit
		 * any pending partial frames in the queue.  This is
		 * meant to be used alongside sendfile() to get properly
		 * filled frames when the user (for example) must write
		 * out headers with a write() call first and then use
		 * sendfile to send out the data parts.
		 *
		 * TCP_CORK can be set together with TCP_NODELAY and it is
		 * stronger than TCP_NODELAY.
		 */
		if (val) {
			tp->nonagle |= TCP_NAGLE_CORK;
		} else {
			tp->nonagle &= ~TCP_NAGLE_CORK;
			if (tp->nonagle&TCP_NAGLE_OFF)
				tp->nonagle |= TCP_NAGLE_PUSH;
2576
			tcp_push_pending_frames(sk);
L
Linus Torvalds 已提交
2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587
		}
		break;

	case TCP_KEEPIDLE:
		if (val < 1 || val > MAX_TCP_KEEPIDLE)
			err = -EINVAL;
		else {
			tp->keepalive_time = val * HZ;
			if (sock_flag(sk, SOCK_KEEPOPEN) &&
			    !((1 << sk->sk_state) &
			      (TCPF_CLOSE | TCPF_LISTEN))) {
2588
				u32 elapsed = keepalive_time_elapsed(tp);
L
Linus Torvalds 已提交
2589 2590 2591 2592
				if (tp->keepalive_time > elapsed)
					elapsed = tp->keepalive_time - elapsed;
				else
					elapsed = 0;
2593
				inet_csk_reset_keepalive_timer(sk, elapsed);
L
Linus Torvalds 已提交
2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612
			}
		}
		break;
	case TCP_KEEPINTVL:
		if (val < 1 || val > MAX_TCP_KEEPINTVL)
			err = -EINVAL;
		else
			tp->keepalive_intvl = val * HZ;
		break;
	case TCP_KEEPCNT:
		if (val < 1 || val > MAX_TCP_KEEPCNT)
			err = -EINVAL;
		else
			tp->keepalive_probes = val;
		break;
	case TCP_SYNCNT:
		if (val < 1 || val > MAX_TCP_SYNCNT)
			err = -EINVAL;
		else
2613
			icsk->icsk_syn_retries = val;
L
Linus Torvalds 已提交
2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625
		break;

	case TCP_LINGER2:
		if (val < 0)
			tp->linger2 = -1;
		else if (val > sysctl_tcp_fin_timeout / HZ)
			tp->linger2 = 0;
		else
			tp->linger2 = val * HZ;
		break;

	case TCP_DEFER_ACCEPT:
2626 2627 2628 2629
		/* Translate value in seconds to number of retransmits */
		icsk->icsk_accept_queue.rskq_defer_accept =
			secs_to_retrans(val, TCP_TIMEOUT_INIT / HZ,
					TCP_RTO_MAX / HZ);
L
Linus Torvalds 已提交
2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645
		break;

	case TCP_WINDOW_CLAMP:
		if (!val) {
			if (sk->sk_state != TCP_CLOSE) {
				err = -EINVAL;
				break;
			}
			tp->window_clamp = 0;
		} else
			tp->window_clamp = val < SOCK_MIN_RCVBUF / 2 ?
						SOCK_MIN_RCVBUF / 2 : val;
		break;

	case TCP_QUICKACK:
		if (!val) {
2646
			icsk->icsk_ack.pingpong = 1;
L
Linus Torvalds 已提交
2647
		} else {
2648
			icsk->icsk_ack.pingpong = 0;
L
Linus Torvalds 已提交
2649 2650
			if ((1 << sk->sk_state) &
			    (TCPF_ESTABLISHED | TCPF_CLOSE_WAIT) &&
2651 2652
			    inet_csk_ack_scheduled(sk)) {
				icsk->icsk_ack.pending |= ICSK_ACK_PUSHED;
2653
				tcp_cleanup_rbuf(sk, 1);
L
Linus Torvalds 已提交
2654
				if (!(val & 1))
2655
					icsk->icsk_ack.pingpong = 1;
L
Linus Torvalds 已提交
2656 2657 2658 2659
			}
		}
		break;

2660 2661 2662 2663 2664 2665
#ifdef CONFIG_TCP_MD5SIG
	case TCP_MD5SIG:
		/* Read the IP->Key mappings from userspace */
		err = tp->af_specific->md5_parse(sk, optval, optlen);
		break;
#endif
2666 2667 2668 2669
	case TCP_USER_TIMEOUT:
		/* Cap the max timeout in ms TCP will retry/retrans
		 * before giving up and aborting (ETIMEDOUT) a connection.
		 */
2670 2671 2672 2673
		if (val < 0)
			err = -EINVAL;
		else
			icsk->icsk_user_timeout = msecs_to_jiffies(val);
2674
		break;
2675 2676 2677 2678 2679 2680 2681 2682

	case TCP_FASTOPEN:
		if (val >= 0 && ((1 << sk->sk_state) & (TCPF_CLOSE |
		    TCPF_LISTEN)))
			err = fastopen_init_queue(sk, val);
		else
			err = -EINVAL;
		break;
2683 2684 2685 2686 2687 2688
	case TCP_TIMESTAMP:
		if (!tp->repair)
			err = -EPERM;
		else
			tp->tsoffset = val - tcp_time_stamp;
		break;
2689 2690 2691 2692
	case TCP_NOTSENT_LOWAT:
		tp->notsent_lowat = val;
		sk->sk_write_space(sk);
		break;
L
Linus Torvalds 已提交
2693 2694 2695
	default:
		err = -ENOPROTOOPT;
		break;
2696 2697
	}

L
Linus Torvalds 已提交
2698 2699 2700 2701
	release_sock(sk);
	return err;
}

2702
int tcp_setsockopt(struct sock *sk, int level, int optname, char __user *optval,
2703
		   unsigned int optlen)
2704
{
2705
	const struct inet_connection_sock *icsk = inet_csk(sk);
2706 2707 2708 2709 2710 2711

	if (level != SOL_TCP)
		return icsk->icsk_af_ops->setsockopt(sk, level, optname,
						     optval, optlen);
	return do_tcp_setsockopt(sk, level, optname, optval, optlen);
}
E
Eric Dumazet 已提交
2712
EXPORT_SYMBOL(tcp_setsockopt);
2713 2714

#ifdef CONFIG_COMPAT
2715
int compat_tcp_setsockopt(struct sock *sk, int level, int optname,
2716
			  char __user *optval, unsigned int optlen)
2717
{
2718 2719 2720
	if (level != SOL_TCP)
		return inet_csk_compat_setsockopt(sk, level, optname,
						  optval, optlen);
2721 2722
	return do_tcp_setsockopt(sk, level, optname, optval, optlen);
}
2723
EXPORT_SYMBOL(compat_tcp_setsockopt);
2724 2725
#endif

L
Linus Torvalds 已提交
2726
/* Return information about state of tcp endpoint in API format. */
2727
void tcp_get_info(const struct sock *sk, struct tcp_info *info)
L
Linus Torvalds 已提交
2728
{
2729
	const struct tcp_sock *tp = tcp_sk(sk);
2730
	const struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
2731 2732 2733 2734 2735
	u32 now = tcp_time_stamp;

	memset(info, 0, sizeof(*info));

	info->tcpi_state = sk->sk_state;
2736
	info->tcpi_ca_state = icsk->icsk_ca_state;
2737
	info->tcpi_retransmits = icsk->icsk_retransmits;
2738
	info->tcpi_probes = icsk->icsk_probes_out;
2739
	info->tcpi_backoff = icsk->icsk_backoff;
L
Linus Torvalds 已提交
2740 2741 2742

	if (tp->rx_opt.tstamp_ok)
		info->tcpi_options |= TCPI_OPT_TIMESTAMPS;
2743
	if (tcp_is_sack(tp))
L
Linus Torvalds 已提交
2744 2745 2746 2747 2748
		info->tcpi_options |= TCPI_OPT_SACK;
	if (tp->rx_opt.wscale_ok) {
		info->tcpi_options |= TCPI_OPT_WSCALE;
		info->tcpi_snd_wscale = tp->rx_opt.snd_wscale;
		info->tcpi_rcv_wscale = tp->rx_opt.rcv_wscale;
2749
	}
L
Linus Torvalds 已提交
2750

E
Eric Dumazet 已提交
2751
	if (tp->ecn_flags & TCP_ECN_OK)
L
Linus Torvalds 已提交
2752
		info->tcpi_options |= TCPI_OPT_ECN;
E
Eric Dumazet 已提交
2753 2754
	if (tp->ecn_flags & TCP_ECN_SEEN)
		info->tcpi_options |= TCPI_OPT_ECN_SEEN;
Y
Yuchung Cheng 已提交
2755 2756
	if (tp->syn_data_acked)
		info->tcpi_options |= TCPI_OPT_SYN_DATA;
L
Linus Torvalds 已提交
2757

2758 2759
	info->tcpi_rto = jiffies_to_usecs(icsk->icsk_rto);
	info->tcpi_ato = jiffies_to_usecs(icsk->icsk_ack.ato);
2760
	info->tcpi_snd_mss = tp->mss_cache;
2761
	info->tcpi_rcv_mss = icsk->icsk_ack.rcv_mss;
L
Linus Torvalds 已提交
2762

2763 2764 2765 2766 2767 2768 2769
	if (sk->sk_state == TCP_LISTEN) {
		info->tcpi_unacked = sk->sk_ack_backlog;
		info->tcpi_sacked = sk->sk_max_ack_backlog;
	} else {
		info->tcpi_unacked = tp->packets_out;
		info->tcpi_sacked = tp->sacked_out;
	}
L
Linus Torvalds 已提交
2770 2771 2772 2773 2774
	info->tcpi_lost = tp->lost_out;
	info->tcpi_retrans = tp->retrans_out;
	info->tcpi_fackets = tp->fackets_out;

	info->tcpi_last_data_sent = jiffies_to_msecs(now - tp->lsndtime);
2775
	info->tcpi_last_data_recv = jiffies_to_msecs(now - icsk->icsk_ack.lrcvtime);
L
Linus Torvalds 已提交
2776 2777
	info->tcpi_last_ack_recv = jiffies_to_msecs(now - tp->rcv_tstamp);

2778
	info->tcpi_pmtu = icsk->icsk_pmtu_cookie;
L
Linus Torvalds 已提交
2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793
	info->tcpi_rcv_ssthresh = tp->rcv_ssthresh;
	info->tcpi_rtt = jiffies_to_usecs(tp->srtt)>>3;
	info->tcpi_rttvar = jiffies_to_usecs(tp->mdev)>>2;
	info->tcpi_snd_ssthresh = tp->snd_ssthresh;
	info->tcpi_snd_cwnd = tp->snd_cwnd;
	info->tcpi_advmss = tp->advmss;
	info->tcpi_reordering = tp->reordering;

	info->tcpi_rcv_rtt = jiffies_to_usecs(tp->rcv_rtt_est.rtt)>>3;
	info->tcpi_rcv_space = tp->rcvq_space.space;

	info->tcpi_total_retrans = tp->total_retrans;
}
EXPORT_SYMBOL_GPL(tcp_get_info);

2794 2795
static int do_tcp_getsockopt(struct sock *sk, int level,
		int optname, char __user *optval, int __user *optlen)
L
Linus Torvalds 已提交
2796
{
2797
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810
	struct tcp_sock *tp = tcp_sk(sk);
	int val, len;

	if (get_user(len, optlen))
		return -EFAULT;

	len = min_t(unsigned int, len, sizeof(int));

	if (len < 0)
		return -EINVAL;

	switch (optname) {
	case TCP_MAXSEG:
2811
		val = tp->mss_cache;
L
Linus Torvalds 已提交
2812 2813
		if (!val && ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN)))
			val = tp->rx_opt.user_mss;
2814 2815
		if (tp->repair)
			val = tp->rx_opt.mss_clamp;
L
Linus Torvalds 已提交
2816 2817 2818 2819 2820 2821 2822 2823
		break;
	case TCP_NODELAY:
		val = !!(tp->nonagle&TCP_NAGLE_OFF);
		break;
	case TCP_CORK:
		val = !!(tp->nonagle&TCP_NAGLE_CORK);
		break;
	case TCP_KEEPIDLE:
E
Eric Dumazet 已提交
2824
		val = keepalive_time_when(tp) / HZ;
L
Linus Torvalds 已提交
2825 2826
		break;
	case TCP_KEEPINTVL:
E
Eric Dumazet 已提交
2827
		val = keepalive_intvl_when(tp) / HZ;
L
Linus Torvalds 已提交
2828 2829
		break;
	case TCP_KEEPCNT:
E
Eric Dumazet 已提交
2830
		val = keepalive_probes(tp);
L
Linus Torvalds 已提交
2831 2832
		break;
	case TCP_SYNCNT:
2833
		val = icsk->icsk_syn_retries ? : sysctl_tcp_syn_retries;
L
Linus Torvalds 已提交
2834 2835 2836 2837 2838 2839 2840
		break;
	case TCP_LINGER2:
		val = tp->linger2;
		if (val >= 0)
			val = (val ? : sysctl_tcp_fin_timeout) / HZ;
		break;
	case TCP_DEFER_ACCEPT:
2841 2842
		val = retrans_to_secs(icsk->icsk_accept_queue.rskq_defer_accept,
				      TCP_TIMEOUT_INIT / HZ, TCP_RTO_MAX / HZ);
L
Linus Torvalds 已提交
2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862
		break;
	case TCP_WINDOW_CLAMP:
		val = tp->window_clamp;
		break;
	case TCP_INFO: {
		struct tcp_info info;

		if (get_user(len, optlen))
			return -EFAULT;

		tcp_get_info(sk, &info);

		len = min_t(unsigned int, len, sizeof(info));
		if (put_user(len, optlen))
			return -EFAULT;
		if (copy_to_user(optval, &info, len))
			return -EFAULT;
		return 0;
	}
	case TCP_QUICKACK:
2863
		val = !icsk->icsk_ack.pingpong;
L
Linus Torvalds 已提交
2864
		break;
2865 2866 2867 2868 2869 2870 2871

	case TCP_CONGESTION:
		if (get_user(len, optlen))
			return -EFAULT;
		len = min_t(unsigned int, len, TCP_CA_NAME_MAX);
		if (put_user(len, optlen))
			return -EFAULT;
2872
		if (copy_to_user(optval, icsk->icsk_ca_ops->name, len))
2873 2874
			return -EFAULT;
		return 0;
2875

2876 2877 2878 2879 2880 2881
	case TCP_THIN_LINEAR_TIMEOUTS:
		val = tp->thin_lto;
		break;
	case TCP_THIN_DUPACK:
		val = tp->thin_dupack;
		break;
2882

P
Pavel Emelyanov 已提交
2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902
	case TCP_REPAIR:
		val = tp->repair;
		break;

	case TCP_REPAIR_QUEUE:
		if (tp->repair)
			val = tp->repair_queue;
		else
			return -EINVAL;
		break;

	case TCP_QUEUE_SEQ:
		if (tp->repair_queue == TCP_SEND_QUEUE)
			val = tp->write_seq;
		else if (tp->repair_queue == TCP_RECV_QUEUE)
			val = tp->rcv_nxt;
		else
			return -EINVAL;
		break;

2903 2904 2905
	case TCP_USER_TIMEOUT:
		val = jiffies_to_msecs(icsk->icsk_user_timeout);
		break;
2906 2907 2908
	case TCP_TIMESTAMP:
		val = tcp_time_stamp + tp->tsoffset;
		break;
2909 2910 2911
	case TCP_NOTSENT_LOWAT:
		val = tp->notsent_lowat;
		break;
L
Linus Torvalds 已提交
2912 2913
	default:
		return -ENOPROTOOPT;
2914
	}
L
Linus Torvalds 已提交
2915 2916 2917 2918 2919 2920 2921 2922

	if (put_user(len, optlen))
		return -EFAULT;
	if (copy_to_user(optval, &val, len))
		return -EFAULT;
	return 0;
}

2923 2924 2925 2926 2927 2928 2929 2930 2931 2932
int tcp_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
		   int __user *optlen)
{
	struct inet_connection_sock *icsk = inet_csk(sk);

	if (level != SOL_TCP)
		return icsk->icsk_af_ops->getsockopt(sk, level, optname,
						     optval, optlen);
	return do_tcp_getsockopt(sk, level, optname, optval, optlen);
}
E
Eric Dumazet 已提交
2933
EXPORT_SYMBOL(tcp_getsockopt);
2934 2935

#ifdef CONFIG_COMPAT
2936 2937
int compat_tcp_getsockopt(struct sock *sk, int level, int optname,
			  char __user *optval, int __user *optlen)
2938
{
2939 2940 2941
	if (level != SOL_TCP)
		return inet_csk_compat_getsockopt(sk, level, optname,
						  optval, optlen);
2942 2943
	return do_tcp_getsockopt(sk, level, optname, optval, optlen);
}
2944
EXPORT_SYMBOL(compat_tcp_getsockopt);
2945
#endif
L
Linus Torvalds 已提交
2946

2947
#ifdef CONFIG_TCP_MD5SIG
2948 2949
static struct tcp_md5sig_pool __percpu *tcp_md5sig_pool __read_mostly;
static DEFINE_MUTEX(tcp_md5sig_mutex);
2950

2951
static void __tcp_free_md5sig_pool(struct tcp_md5sig_pool __percpu *pool)
2952 2953
{
	int cpu;
2954

2955
	for_each_possible_cpu(cpu) {
2956 2957 2958 2959
		struct tcp_md5sig_pool *p = per_cpu_ptr(pool, cpu);

		if (p->md5_desc.tfm)
			crypto_free_hash(p->md5_desc.tfm);
2960 2961 2962 2963
	}
	free_percpu(pool);
}

2964
static void __tcp_alloc_md5sig_pool(void)
2965 2966
{
	int cpu;
2967
	struct tcp_md5sig_pool __percpu *pool;
2968

2969
	pool = alloc_percpu(struct tcp_md5sig_pool);
2970
	if (!pool)
2971
		return;
2972 2973 2974 2975 2976

	for_each_possible_cpu(cpu) {
		struct crypto_hash *hash;

		hash = crypto_alloc_hash("md5", 0, CRYPTO_ALG_ASYNC);
2977
		if (IS_ERR_OR_NULL(hash))
2978 2979
			goto out_free;

2980
		per_cpu_ptr(pool, cpu)->md5_desc.tfm = hash;
2981
	}
2982 2983 2984 2985 2986 2987
	/* before setting tcp_md5sig_pool, we must commit all writes
	 * to memory. See ACCESS_ONCE() in tcp_get_md5sig_pool()
	 */
	smp_wmb();
	tcp_md5sig_pool = pool;
	return;
2988 2989 2990 2991
out_free:
	__tcp_free_md5sig_pool(pool);
}

2992
bool tcp_alloc_md5sig_pool(void)
2993
{
2994 2995 2996 2997 2998 2999 3000
	if (unlikely(!tcp_md5sig_pool)) {
		mutex_lock(&tcp_md5sig_mutex);

		if (!tcp_md5sig_pool)
			__tcp_alloc_md5sig_pool();

		mutex_unlock(&tcp_md5sig_mutex);
3001
	}
3002
	return tcp_md5sig_pool != NULL;
3003 3004 3005
}
EXPORT_SYMBOL(tcp_alloc_md5sig_pool);

E
Eric Dumazet 已提交
3006 3007 3008 3009 3010 3011 3012 3013 3014

/**
 *	tcp_get_md5sig_pool - get md5sig_pool for this user
 *
 *	We use percpu structure, so if we succeed, we exit with preemption
 *	and BH disabled, to make sure another thread or softirq handling
 *	wont try to get same context.
 */
struct tcp_md5sig_pool *tcp_get_md5sig_pool(void)
3015
{
3016
	struct tcp_md5sig_pool __percpu *p;
E
Eric Dumazet 已提交
3017 3018

	local_bh_disable();
3019
	p = ACCESS_ONCE(tcp_md5sig_pool);
E
Eric Dumazet 已提交
3020
	if (p)
3021
		return __this_cpu_ptr(p);
3022

E
Eric Dumazet 已提交
3023 3024 3025 3026
	local_bh_enable();
	return NULL;
}
EXPORT_SYMBOL(tcp_get_md5sig_pool);
3027

3028
int tcp_md5_hash_header(struct tcp_md5sig_pool *hp,
3029
			const struct tcphdr *th)
3030 3031
{
	struct scatterlist sg;
3032
	struct tcphdr hdr;
3033 3034
	int err;

3035 3036 3037 3038
	/* We are not allowed to change tcphdr, make a local copy */
	memcpy(&hdr, th, sizeof(hdr));
	hdr.check = 0;

3039
	/* options aren't included in the hash */
3040 3041
	sg_init_one(&sg, &hdr, sizeof(hdr));
	err = crypto_hash_update(&hp->md5_desc, &sg, sizeof(hdr));
3042 3043 3044 3045 3046
	return err;
}
EXPORT_SYMBOL(tcp_md5_hash_header);

int tcp_md5_hash_skb_data(struct tcp_md5sig_pool *hp,
3047
			  const struct sk_buff *skb, unsigned int header_len)
3048 3049 3050 3051
{
	struct scatterlist sg;
	const struct tcphdr *tp = tcp_hdr(skb);
	struct hash_desc *desc = &hp->md5_desc;
3052 3053 3054
	unsigned int i;
	const unsigned int head_data_len = skb_headlen(skb) > header_len ?
					   skb_headlen(skb) - header_len : 0;
3055
	const struct skb_shared_info *shi = skb_shinfo(skb);
3056
	struct sk_buff *frag_iter;
3057 3058 3059 3060 3061 3062 3063 3064 3065

	sg_init_table(&sg, 1);

	sg_set_buf(&sg, ((u8 *) tp) + header_len, head_data_len);
	if (crypto_hash_update(desc, &sg, head_data_len))
		return 1;

	for (i = 0; i < shi->nr_frags; ++i) {
		const struct skb_frag_struct *f = &shi->frags[i];
E
Eric Dumazet 已提交
3066 3067 3068 3069 3070
		unsigned int offset = f->page_offset;
		struct page *page = skb_frag_page(f) + (offset >> PAGE_SHIFT);

		sg_set_page(&sg, page, skb_frag_size(f),
			    offset_in_page(offset));
E
Eric Dumazet 已提交
3071
		if (crypto_hash_update(desc, &sg, skb_frag_size(f)))
3072 3073 3074
			return 1;
	}

3075 3076 3077 3078
	skb_walk_frags(skb, frag_iter)
		if (tcp_md5_hash_skb_data(hp, frag_iter, 0))
			return 1;

3079 3080 3081 3082
	return 0;
}
EXPORT_SYMBOL(tcp_md5_hash_skb_data);

3083
int tcp_md5_hash_key(struct tcp_md5sig_pool *hp, const struct tcp_md5sig_key *key)
3084 3085 3086 3087 3088 3089 3090 3091
{
	struct scatterlist sg;

	sg_init_one(&sg, key->key, key->keylen);
	return crypto_hash_update(&hp->md5_desc, &sg, key->keylen);
}
EXPORT_SYMBOL(tcp_md5_hash_key);

3092 3093
#endif

A
Andi Kleen 已提交
3094 3095
void tcp_done(struct sock *sk)
{
3096 3097
	struct request_sock *req = tcp_sk(sk)->fastopen_rsk;

3098
	if (sk->sk_state == TCP_SYN_SENT || sk->sk_state == TCP_SYN_RECV)
3099
		TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_ATTEMPTFAILS);
A
Andi Kleen 已提交
3100 3101 3102

	tcp_set_state(sk, TCP_CLOSE);
	tcp_clear_xmit_timers(sk);
3103 3104
	if (req != NULL)
		reqsk_fastopen_remove(sk, req, false);
A
Andi Kleen 已提交
3105 3106 3107 3108 3109 3110 3111 3112 3113 3114

	sk->sk_shutdown = SHUTDOWN_MASK;

	if (!sock_flag(sk, SOCK_DEAD))
		sk->sk_state_change(sk);
	else
		inet_csk_destroy_sock(sk);
}
EXPORT_SYMBOL_GPL(tcp_done);

3115
extern struct tcp_congestion_ops tcp_reno;
L
Linus Torvalds 已提交
3116 3117 3118 3119

static __initdata unsigned long thash_entries;
static int __init set_thash_entries(char *str)
{
3120 3121
	ssize_t ret;

L
Linus Torvalds 已提交
3122 3123
	if (!str)
		return 0;
3124 3125 3126 3127 3128

	ret = kstrtoul(str, 0, &thash_entries);
	if (ret)
		return 0;

L
Linus Torvalds 已提交
3129 3130 3131 3132
	return 1;
}
__setup("thash_entries=", set_thash_entries);

3133
static void tcp_init_mem(void)
3134 3135 3136
{
	unsigned long limit = nr_free_buffer_pages() / 8;
	limit = max(limit, 128UL);
3137 3138 3139
	sysctl_tcp_mem[0] = limit / 4 * 3;
	sysctl_tcp_mem[1] = limit;
	sysctl_tcp_mem[2] = sysctl_tcp_mem[0] * 2;
3140 3141
}

L
Linus Torvalds 已提交
3142 3143 3144
void __init tcp_init(void)
{
	struct sk_buff *skb = NULL;
3145
	unsigned long limit;
3146
	int max_rshare, max_wshare, cnt;
3147
	unsigned int i;
L
Linus Torvalds 已提交
3148

3149
	BUILD_BUG_ON(sizeof(struct tcp_skb_cb) > sizeof(skb->cb));
L
Linus Torvalds 已提交
3150

3151
	percpu_counter_init(&tcp_sockets_allocated, 0);
3152
	percpu_counter_init(&tcp_orphan_count, 0);
3153 3154 3155
	tcp_hashinfo.bind_bucket_cachep =
		kmem_cache_create("tcp_bind_bucket",
				  sizeof(struct inet_bind_bucket), 0,
3156
				  SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
L
Linus Torvalds 已提交
3157 3158 3159 3160 3161 3162

	/* Size and allocate the main established and bind bucket
	 * hash tables.
	 *
	 * The methodology is similar to that of the buffer cache.
	 */
3163
	tcp_hashinfo.ehash =
L
Linus Torvalds 已提交
3164
		alloc_large_system_hash("TCP established",
3165
					sizeof(struct inet_ehash_bucket),
L
Linus Torvalds 已提交
3166
					thash_entries,
E
Eric Dumazet 已提交
3167
					17, /* one slot per 128 KB of memory */
3168
					0,
L
Linus Torvalds 已提交
3169
					NULL,
3170
					&tcp_hashinfo.ehash_mask,
3171
					0,
3172
					thash_entries ? 0 : 512 * 1024);
E
Eric Dumazet 已提交
3173
	for (i = 0; i <= tcp_hashinfo.ehash_mask; i++)
3174
		INIT_HLIST_NULLS_HEAD(&tcp_hashinfo.ehash[i].chain, i);
E
Eric Dumazet 已提交
3175

3176 3177
	if (inet_ehash_locks_alloc(&tcp_hashinfo))
		panic("TCP: failed to alloc ehash_locks");
3178
	tcp_hashinfo.bhash =
L
Linus Torvalds 已提交
3179
		alloc_large_system_hash("TCP bind",
3180
					sizeof(struct inet_bind_hashbucket),
3181
					tcp_hashinfo.ehash_mask + 1,
E
Eric Dumazet 已提交
3182
					17, /* one slot per 128 KB of memory */
3183
					0,
3184
					&tcp_hashinfo.bhash_size,
L
Linus Torvalds 已提交
3185
					NULL,
3186
					0,
L
Linus Torvalds 已提交
3187
					64 * 1024);
3188
	tcp_hashinfo.bhash_size = 1U << tcp_hashinfo.bhash_size;
3189 3190 3191
	for (i = 0; i < tcp_hashinfo.bhash_size; i++) {
		spin_lock_init(&tcp_hashinfo.bhash[i].lock);
		INIT_HLIST_HEAD(&tcp_hashinfo.bhash[i].chain);
L
Linus Torvalds 已提交
3192 3193
	}

E
Eric Dumazet 已提交
3194 3195 3196 3197 3198 3199

	cnt = tcp_hashinfo.ehash_mask + 1;

	tcp_death_row.sysctl_max_tw_buckets = cnt / 2;
	sysctl_tcp_max_orphans = cnt / 2;
	sysctl_max_syn_backlog = max(128, cnt / 256);
L
Linus Torvalds 已提交
3200

3201
	tcp_init_mem();
3202
	/* Set per-socket limits to no more than 1/128 the pressure threshold */
E
Eric Dumazet 已提交
3203
	limit = nr_free_buffer_pages() << (PAGE_SHIFT - 7);
3204 3205
	max_wshare = min(4UL*1024*1024, limit);
	max_rshare = min(6UL*1024*1024, limit);
3206

3207
	sysctl_tcp_wmem[0] = SK_MEM_QUANTUM;
3208
	sysctl_tcp_wmem[1] = 16*1024;
3209
	sysctl_tcp_wmem[2] = max(64*1024, max_wshare);
3210

3211
	sysctl_tcp_rmem[0] = SK_MEM_QUANTUM;
3212
	sysctl_tcp_rmem[1] = 87380;
3213
	sysctl_tcp_rmem[2] = max(87380, max_rshare);
L
Linus Torvalds 已提交
3214

3215
	pr_info("Hash tables configured (established %u bind %u)\n",
J
Joe Perches 已提交
3216
		tcp_hashinfo.ehash_mask + 1, tcp_hashinfo.bhash_size);
3217

3218 3219
	tcp_metrics_init();

3220
	tcp_register_congestion_control(&tcp_reno);
3221

E
Eric Dumazet 已提交
3222
	tcp_tasklet_init();
L
Linus Torvalds 已提交
3223
}