tcp.c 79.9 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>
J
Jens Axboe 已提交
277
#include <net/sock.h>
L
Linus Torvalds 已提交
278 279 280

#include <asm/uaccess.h>
#include <asm/ioctls.h>
281
#include <net/busy_poll.h>
L
Linus Torvalds 已提交
282

283
int sysctl_tcp_fin_timeout __read_mostly = TCP_FIN_TIMEOUT;
L
Linus Torvalds 已提交
284

E
Eric Dumazet 已提交
285 286
int sysctl_tcp_min_tso_segs __read_mostly = 2;

E
Eric Dumazet 已提交
287 288
int sysctl_tcp_autocorking __read_mostly = 1;

289
struct percpu_counter tcp_orphan_count;
290 291
EXPORT_SYMBOL_GPL(tcp_orphan_count);

292
long sysctl_tcp_mem[3] __read_mostly;
293 294
int sysctl_tcp_wmem[3] __read_mostly;
int sysctl_tcp_rmem[3] __read_mostly;
L
Linus Torvalds 已提交
295

296
EXPORT_SYMBOL(sysctl_tcp_mem);
L
Linus Torvalds 已提交
297 298 299
EXPORT_SYMBOL(sysctl_tcp_rmem);
EXPORT_SYMBOL(sysctl_tcp_wmem);

E
Eric Dumazet 已提交
300
atomic_long_t tcp_memory_allocated;	/* Current allocated memory. */
L
Linus Torvalds 已提交
301
EXPORT_SYMBOL(tcp_memory_allocated);
302 303 304 305 306

/*
 * Current number of TCP sockets.
 */
struct percpu_counter tcp_sockets_allocated;
L
Linus Torvalds 已提交
307 308
EXPORT_SYMBOL(tcp_sockets_allocated);

J
Jens Axboe 已提交
309 310 311 312 313 314 315 316 317
/*
 * TCP splice context
 */
struct tcp_splice_state {
	struct pipe_inode_info *pipe;
	size_t len;
	unsigned int flags;
};

L
Linus Torvalds 已提交
318 319 320
/*
 * Pressure flag: try to collapse.
 * Technical note: it is used by multiple contexts non atomically.
321
 * All the __sk_mem_schedule() is of this nature: accounting
L
Linus Torvalds 已提交
322 323
 * is strict, actions are advisory and have some latency.
 */
324
int tcp_memory_pressure __read_mostly;
L
Linus Torvalds 已提交
325 326
EXPORT_SYMBOL(tcp_memory_pressure);

327
void tcp_enter_memory_pressure(struct sock *sk)
L
Linus Torvalds 已提交
328 329
{
	if (!tcp_memory_pressure) {
P
Pavel Emelyanov 已提交
330
		NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMEMORYPRESSURES);
L
Linus Torvalds 已提交
331 332 333 334 335
		tcp_memory_pressure = 1;
	}
}
EXPORT_SYMBOL(tcp_enter_memory_pressure);

336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372
/* 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;
}

373 374 375 376 377 378 379 380 381 382
/* 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);

383
	__skb_queue_head_init(&tp->out_of_order_queue);
384 385
	tcp_init_xmit_timers(sk);
	tcp_prequeue_init(tp);
E
Eric Dumazet 已提交
386
	INIT_LIST_HEAD(&tp->tsq_node);
387 388

	icsk->icsk_rto = TCP_TIMEOUT_INIT;
389
	tp->mdev_us = jiffies_to_usecs(TCP_TIMEOUT_INIT);
390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405

	/* 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
406
	tcp_enable_early_retrans(tp);
407
	tcp_assign_congestion_control(sk);
408

409 410
	tp->tsoffset = 0;

411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427
	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);

428
static void tcp_tx_timestamp(struct sock *sk, struct sk_buff *skb)
429
{
430 431
	if (sk->sk_tsflags) {
		struct skb_shared_info *shinfo = skb_shinfo(skb);
432

433 434 435 436
		sock_tx_timestamp(sk, &shinfo->tx_flags);
		if (shinfo->tx_flags & SKBTX_ANY_TSTAMP)
			shinfo->tskey = TCP_SKB_CB(skb)->seq + skb->len - 1;
	}
437 438
}

L
Linus Torvalds 已提交
439 440 441 442 443 444 445 446 447 448 449
/*
 *	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;
450
	const struct tcp_sock *tp = tcp_sk(sk);
L
Linus Torvalds 已提交
451

452 453
	sock_rps_record_flow(sk);

E
Eric Dumazet 已提交
454
	sock_poll_wait(file, sk_sleep(sk), wait);
L
Linus Torvalds 已提交
455
	if (sk->sk_state == TCP_LISTEN)
456
		return inet_csk_listen_poll(sk);
L
Linus Torvalds 已提交
457 458

	/* Socket is not locked. We are protected from async events
459 460
	 * by poll logic and correct handling of state changes
	 * made by other threads is impossible in any case.
L
Linus Torvalds 已提交
461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483
	 */

	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
484
	 * solve this dilemma. I would prefer, if POLLHUP were maskable,
L
Linus Torvalds 已提交
485 486
	 * then we could set it on SND_SHUTDOWN. BTW examples given
	 * in Stevens' books assume exactly this behaviour, it explains
487
	 * why POLLHUP is incompatible with POLLOUT.	--ANK
L
Linus Torvalds 已提交
488 489 490 491 492 493 494
	 *
	 * 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)
495
		mask |= POLLIN | POLLRDNORM | POLLRDHUP;
L
Linus Torvalds 已提交
496

497 498 499
	/* Connected or passive Fast Open socket? */
	if (sk->sk_state != TCP_SYN_SENT &&
	    (sk->sk_state != TCP_SYN_RECV || tp->fastopen_rsk != NULL)) {
500 501 502 503 504
		int target = sock_rcvlowat(sk, 0, INT_MAX);

		if (tp->urg_seq == tp->copied_seq &&
		    !sock_flag(sk, SOCK_URGINLINE) &&
		    tp->urg_data)
505
			target++;
506

L
Linus Torvalds 已提交
507 508 509
		/* Potential race condition. If read of tp below will
		 * escape above sk->sk_state, we can be illegally awaken
		 * in SYN_* states. */
510
		if (tp->rcv_nxt - tp->copied_seq >= target)
L
Linus Torvalds 已提交
511 512 513
			mask |= POLLIN | POLLRDNORM;

		if (!(sk->sk_shutdown & SEND_SHUTDOWN)) {
514
			if (sk_stream_is_writeable(sk)) {
L
Linus Torvalds 已提交
515 516 517 518 519 520 521 522 523 524
				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.
				 */
525
				if (sk_stream_is_writeable(sk))
L
Linus Torvalds 已提交
526 527
					mask |= POLLOUT | POLLWRNORM;
			}
528 529
		} else
			mask |= POLLOUT | POLLWRNORM;
L
Linus Torvalds 已提交
530 531 532 533

		if (tp->urg_data & TCP_URG_VALID)
			mask |= POLLPRI;
	}
T
Tom Marshall 已提交
534 535
	/* This barrier is coupled with smp_wmb() in tcp_reset() */
	smp_rmb();
536
	if (sk->sk_err || !skb_queue_empty(&sk->sk_error_queue))
T
Tom Marshall 已提交
537 538
		mask |= POLLERR;

L
Linus Torvalds 已提交
539 540
	return mask;
}
E
Eric Dumazet 已提交
541
EXPORT_SYMBOL(tcp_poll);
L
Linus Torvalds 已提交
542 543 544 545 546

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
547
	bool slow;
L
Linus Torvalds 已提交
548 549 550 551 552 553

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

E
Eric Dumazet 已提交
554
		slow = lock_sock_fast(sk);
L
Linus Torvalds 已提交
555 556 557 558 559 560
		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)) {
561

L
Linus Torvalds 已提交
562 563
			answ = tp->rcv_nxt - tp->copied_seq;

E
Eric Dumazet 已提交
564 565 566
			/* Subtract 1, if FIN was received */
			if (answ && sock_flag(sk, SOCK_DONE))
				answ--;
L
Linus Torvalds 已提交
567 568
		} else
			answ = tp->urg_seq - tp->copied_seq;
E
Eric Dumazet 已提交
569
		unlock_sock_fast(sk, slow);
L
Linus Torvalds 已提交
570 571 572 573 574 575 576 577 578 579 580 581 582
		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;
583 584 585 586 587 588 589 590 591
	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
592 593
	default:
		return -ENOIOCTLCMD;
594
	}
L
Linus Torvalds 已提交
595 596 597

	return put_user(answ, (int __user *)arg);
}
E
Eric Dumazet 已提交
598
EXPORT_SYMBOL(tcp_ioctl);
L
Linus Torvalds 已提交
599 600 601

static inline void tcp_mark_push(struct tcp_sock *tp, struct sk_buff *skb)
{
E
Eric Dumazet 已提交
602
	TCP_SKB_CB(skb)->tcp_flags |= TCPHDR_PSH;
L
Linus Torvalds 已提交
603 604 605
	tp->pushed_seq = tp->write_seq;
}

E
Eric Dumazet 已提交
606
static inline bool forced_push(const struct tcp_sock *tp)
L
Linus Torvalds 已提交
607 608 609 610
{
	return after(tp->write_seq, tp->pushed_seq + (tp->max_window >> 1));
}

611
static void skb_entail(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
612
{
613
	struct tcp_sock *tp = tcp_sk(sk);
614 615 616 617
	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
618
	tcb->tcp_flags = TCPHDR_ACK;
619
	tcb->sacked  = 0;
620
	__skb_header_release(skb);
621
	tcp_add_write_queue_tail(sk, skb);
622 623
	sk->sk_wmem_queued += skb->truesize;
	sk_mem_charge(sk, skb->truesize);
624
	if (tp->nonagle & TCP_NAGLE_PUSH)
625
		tp->nonagle &= ~TCP_NAGLE_PUSH;
L
Linus Torvalds 已提交
626 627
}

628
static inline void tcp_mark_urg(struct tcp_sock *tp, int flags)
L
Linus Torvalds 已提交
629
{
I
Ilpo Järvinen 已提交
630
	if (flags & MSG_OOB)
L
Linus Torvalds 已提交
631 632 633
		tp->snd_up = tp->write_seq;
}

E
Eric Dumazet 已提交
634
/* If a not yet filled skb is pushed, do not send it if
635
 * we have data packets in Qdisc or NIC queues :
E
Eric Dumazet 已提交
636 637 638 639
 * 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
640 641 642
 * than pure acks (dataless) packets, the last checks prevent
 * autocorking if we only have an ACK in Qdisc/NIC queues,
 * or if TX completion was delayed after we processed ACK packet.
E
Eric Dumazet 已提交
643 644 645
 */
static bool tcp_should_autocork(struct sock *sk, struct sk_buff *skb,
				int size_goal)
L
Linus Torvalds 已提交
646
{
E
Eric Dumazet 已提交
647 648
	return skb->len < size_goal &&
	       sysctl_tcp_autocorking &&
649
	       skb != tcp_write_queue_head(sk) &&
E
Eric Dumazet 已提交
650 651 652 653 654 655 656 657
	       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;
658

E
Eric Dumazet 已提交
659 660
	if (!tcp_send_head(sk))
		return;
661

E
Eric Dumazet 已提交
662 663 664 665 666 667 668 669 670 671 672 673 674
	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);
		}
675 676 677 678 679
		/* It is possible TX completion already happened
		 * before we set TSQ_THROTTLED.
		 */
		if (atomic_read(&sk->sk_wmem_alloc) > skb->truesize)
			return;
L
Linus Torvalds 已提交
680
	}
E
Eric Dumazet 已提交
681 682 683 684 685

	if (flags & MSG_MORE)
		nonagle = TCP_NAGLE_CORK;

	__tcp_push_pending_frames(sk, mss_now, nonagle);
L
Linus Torvalds 已提交
686 687
}

688 689
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
690 691
{
	struct tcp_splice_state *tss = rd_desc->arg.data;
692
	int ret;
J
Jens Axboe 已提交
693

694 695
	ret = skb_splice_bits(skb, offset, tss->pipe, min(rd_desc->count, len),
			      tss->flags);
696 697 698
	if (ret > 0)
		rd_desc->count -= ret;
	return ret;
J
Jens Axboe 已提交
699 700 701 702 703 704 705
}

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,
706
		.count	  = tss->len,
J
Jens Axboe 已提交
707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737
	};

	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;

738
	sock_rps_record_flow(sk);
J
Jens Axboe 已提交
739 740 741 742 743 744 745 746 747 748
	/*
	 * We can't seek on a socket input
	 */
	if (unlikely(*ppos))
		return -ESPIPE;

	ret = spliced = 0;

	lock_sock(sk);

749
	timeo = sock_rcvtimeo(sk, sock->file->f_flags & O_NONBLOCK);
J
Jens Axboe 已提交
750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787
	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;

788 789
		if (!timeo)
			break;
J
Jens Axboe 已提交
790 791 792 793
		release_sock(sk);
		lock_sock(sk);

		if (sk->sk_err || sk->sk_state == TCP_CLOSE ||
794
		    (sk->sk_shutdown & RCV_SHUTDOWN) ||
J
Jens Axboe 已提交
795 796 797 798 799 800 801 802 803 804 805
		    signal_pending(current))
			break;
	}

	release_sock(sk);

	if (spliced)
		return spliced;

	return ret;
}
E
Eric Dumazet 已提交
806
EXPORT_SYMBOL(tcp_splice_read);
J
Jens Axboe 已提交
807

808
struct sk_buff *sk_stream_alloc_skb(struct sock *sk, int size, gfp_t gfp)
809 810 811 812 813 814 815 816
{
	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) {
817
		if (sk_wmem_schedule(sk, skb->truesize)) {
818
			skb_reserve(skb, sk->sk_prot->max_header);
819 820 821 822
			/*
			 * Make sure that we have exactly size bytes
			 * available to the caller, no more, no less.
			 */
E
Eric Dumazet 已提交
823
			skb->reserved_tailroom = skb->end - skb->tail - size;
824 825 826 827
			return skb;
		}
		__kfree_skb(skb);
	} else {
828
		sk->sk_prot->enter_memory_pressure(sk);
829 830 831 832 833
		sk_stream_moderate_sndbuf(sk);
	}
	return NULL;
}

I
Ilpo Järvinen 已提交
834 835 836 837
static unsigned int tcp_xmit_size_goal(struct sock *sk, u32 mss_now,
				       int large_allowed)
{
	struct tcp_sock *tp = tcp_sk(sk);
838
	u32 new_size_goal, size_goal;
E
Eric Dumazet 已提交
839 840 841 842

	if (!large_allowed || !sk_can_gso(sk))
		return mss_now;

843 844
	/* Note : tcp_tso_autosize() will eventually split this later */
	new_size_goal = sk->sk_gso_max_size - 1 - MAX_TCP_HEADER;
E
Eric Dumazet 已提交
845 846 847 848 849 850 851 852 853
	new_size_goal = tcp_bound_to_half_wnd(tp, new_size_goal);

	/* We try hard to avoid divides here */
	size_goal = tp->gso_segs * mss_now;
	if (unlikely(new_size_goal < size_goal ||
		     new_size_goal >= size_goal + mss_now)) {
		tp->gso_segs = min_t(u16, new_size_goal / mss_now,
				     sk->sk_gso_max_segs);
		size_goal = tp->gso_segs * mss_now;
I
Ilpo Järvinen 已提交
854 855
	}

E
Eric Dumazet 已提交
856
	return max(size_goal, mss_now);
I
Ilpo Järvinen 已提交
857 858 859 860 861 862 863 864 865 866 867 868
}

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;
}

869 870
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
871 872
{
	struct tcp_sock *tp = tcp_sk(sk);
873
	int mss_now, size_goal;
L
Linus Torvalds 已提交
874 875 876 877
	int err;
	ssize_t copied;
	long timeo = sock_sndtimeo(sk, flags & MSG_DONTWAIT);

878 879 880 881 882 883
	/* 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
884 885
		if ((err = sk_stream_wait_connect(sk, &timeo)) != 0)
			goto out_err;
886
	}
L
Linus Torvalds 已提交
887 888 889

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

I
Ilpo Järvinen 已提交
890
	mss_now = tcp_send_mss(sk, &size_goal, flags);
L
Linus Torvalds 已提交
891 892 893 894
	copied = 0;

	err = -EPIPE;
	if (sk->sk_err || (sk->sk_shutdown & SEND_SHUTDOWN))
895
		goto out_err;
L
Linus Torvalds 已提交
896

897
	while (size > 0) {
898
		struct sk_buff *skb = tcp_write_queue_tail(sk);
899 900
		int copy, i;
		bool can_coalesce;
L
Linus Torvalds 已提交
901

902
		if (!tcp_send_head(sk) || (copy = size_goal - skb->len) <= 0) {
L
Linus Torvalds 已提交
903 904 905 906
new_segment:
			if (!sk_stream_memory_free(sk))
				goto wait_for_sndbuf;

907
			skb = sk_stream_alloc_skb(sk, 0, sk->sk_allocation);
L
Linus Torvalds 已提交
908 909 910
			if (!skb)
				goto wait_for_memory;

911
			skb_entail(sk, skb);
912
			copy = size_goal;
L
Linus Torvalds 已提交
913 914 915 916 917 918 919 920 921 922 923
		}

		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;
		}
924
		if (!sk_wmem_schedule(sk, copy))
L
Linus Torvalds 已提交
925
			goto wait_for_memory;
926

L
Linus Torvalds 已提交
927
		if (can_coalesce) {
E
Eric Dumazet 已提交
928
			skb_frag_size_add(&skb_shinfo(skb)->frags[i - 1], copy);
L
Linus Torvalds 已提交
929 930 931 932
		} else {
			get_page(page);
			skb_fill_page_desc(skb, i, page, offset, copy);
		}
933
		skb_shinfo(skb)->tx_flags |= SKBTX_SHARED_FRAG;
934

L
Linus Torvalds 已提交
935 936 937 938
		skb->len += copy;
		skb->data_len += copy;
		skb->truesize += copy;
		sk->sk_wmem_queued += copy;
939
		sk_mem_charge(sk, copy);
940
		skb->ip_summed = CHECKSUM_PARTIAL;
L
Linus Torvalds 已提交
941 942
		tp->write_seq += copy;
		TCP_SKB_CB(skb)->end_seq += copy;
943
		tcp_skb_pcount_set(skb, 0);
L
Linus Torvalds 已提交
944 945

		if (!copied)
E
Eric Dumazet 已提交
946
			TCP_SKB_CB(skb)->tcp_flags &= ~TCPHDR_PSH;
L
Linus Torvalds 已提交
947 948

		copied += copy;
949
		offset += copy;
950 951
		if (!(size -= copy)) {
			tcp_tx_timestamp(sk, skb);
L
Linus Torvalds 已提交
952
			goto out;
953
		}
L
Linus Torvalds 已提交
954

955
		if (skb->len < size_goal || (flags & MSG_OOB))
L
Linus Torvalds 已提交
956 957 958 959
			continue;

		if (forced_push(tp)) {
			tcp_mark_push(tp, skb);
960
			__tcp_push_pending_frames(sk, mss_now, TCP_NAGLE_PUSH);
961
		} else if (skb == tcp_send_head(sk))
L
Linus Torvalds 已提交
962 963 964 965 966 967
			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
968 969
		tcp_push(sk, flags & ~MSG_MORE, mss_now,
			 TCP_NAGLE_PUSH, size_goal);
L
Linus Torvalds 已提交
970 971 972 973

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

I
Ilpo Järvinen 已提交
974
		mss_now = tcp_send_mss(sk, &size_goal, flags);
L
Linus Torvalds 已提交
975 976 977
	}

out:
978
	if (copied && !(flags & MSG_SENDPAGE_NOTLAST))
E
Eric Dumazet 已提交
979
		tcp_push(sk, flags, mss_now, tp->nonagle, size_goal);
L
Linus Torvalds 已提交
980 981 982 983 984 985 986 987 988
	return copied;

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

989 990
int tcp_sendpage(struct sock *sk, struct page *page, int offset,
		 size_t size, int flags)
L
Linus Torvalds 已提交
991 992 993 994
{
	ssize_t res;

	if (!(sk->sk_route_caps & NETIF_F_SG) ||
995
	    !(sk->sk_route_caps & NETIF_F_ALL_CSUM))
996 997
		return sock_no_sendpage(sk->sk_socket, page, offset, size,
					flags);
L
Linus Torvalds 已提交
998 999

	lock_sock(sk);
1000
	res = do_tcp_sendpages(sk, page, offset, size, flags);
L
Linus Torvalds 已提交
1001 1002 1003
	release_sock(sk);
	return res;
}
E
Eric Dumazet 已提交
1004
EXPORT_SYMBOL(tcp_sendpage);
L
Linus Torvalds 已提交
1005

1006
static inline int select_size(const struct sock *sk, bool sg)
L
Linus Torvalds 已提交
1007
{
1008
	const struct tcp_sock *tp = tcp_sk(sk);
1009
	int tmp = tp->mss_cache;
L
Linus Torvalds 已提交
1010

K
Krishna Kumar 已提交
1011
	if (sg) {
1012 1013 1014 1015 1016 1017
		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 {
1018 1019 1020 1021 1022 1023 1024
			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
1025 1026 1027 1028

	return tmp;
}

1029 1030 1031 1032 1033 1034 1035 1036
void tcp_free_fastopen_req(struct tcp_sock *tp)
{
	if (tp->fastopen_req != NULL) {
		kfree(tp->fastopen_req);
		tp->fastopen_req = NULL;
	}
}

1037 1038
static int tcp_sendmsg_fastopen(struct sock *sk, struct msghdr *msg,
				int *copied, size_t size)
1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052
{
	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;
1053
	tp->fastopen_req->size = size;
1054 1055 1056 1057

	flags = (msg->msg_flags & MSG_DONTWAIT) ? O_NONBLOCK : 0;
	err = __inet_stream_connect(sk->sk_socket, msg->msg_name,
				    msg->msg_namelen, flags);
1058
	*copied = tp->fastopen_req->copied;
1059 1060 1061 1062
	tcp_free_fastopen_req(tp);
	return err;
}

1063
int tcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t size)
L
Linus Torvalds 已提交
1064 1065 1066
{
	struct tcp_sock *tp = tcp_sk(sk);
	struct sk_buff *skb;
1067 1068
	int flags, err, copied = 0;
	int mss_now = 0, size_goal, copied_syn = 0;
1069
	bool sg;
L
Linus Torvalds 已提交
1070 1071 1072 1073 1074
	long timeo;

	lock_sock(sk);

	flags = msg->msg_flags;
1075
	if (flags & MSG_FASTOPEN) {
1076
		err = tcp_sendmsg_fastopen(sk, msg, &copied_syn, size);
1077 1078 1079 1080 1081 1082
		if (err == -EINPROGRESS && copied_syn > 0)
			goto out;
		else if (err)
			goto out_err;
	}

L
Linus Torvalds 已提交
1083 1084
	timeo = sock_sndtimeo(sk, flags & MSG_DONTWAIT);

1085 1086 1087 1088 1089 1090
	/* 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
1091
		if ((err = sk_stream_wait_connect(sk, &timeo)) != 0)
1092
			goto do_error;
1093
	}
L
Linus Torvalds 已提交
1094

P
Pavel Emelyanov 已提交
1095 1096 1097
	if (unlikely(tp->repair)) {
		if (tp->repair_queue == TCP_RECV_QUEUE) {
			copied = tcp_send_rcvq(sk, msg, size);
1098
			goto out_nopush;
P
Pavel Emelyanov 已提交
1099 1100 1101 1102 1103 1104 1105 1106 1107
		}

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

		/* 'common' sending to sendq */
	}

L
Linus Torvalds 已提交
1108 1109 1110
	/* This should be in poll */
	clear_bit(SOCK_ASYNC_NOSPACE, &sk->sk_socket->flags);

I
Ilpo Järvinen 已提交
1111
	mss_now = tcp_send_mss(sk, &size_goal, flags);
L
Linus Torvalds 已提交
1112 1113 1114 1115 1116 1117

	/* Ok commence sending. */
	copied = 0;

	err = -EPIPE;
	if (sk->sk_err || (sk->sk_shutdown & SEND_SHUTDOWN))
1118
		goto out_err;
L
Linus Torvalds 已提交
1119

1120
	sg = !!(sk->sk_route_caps & NETIF_F_SG);
K
Krishna Kumar 已提交
1121

1122 1123 1124
	while (iov_iter_count(&msg->msg_iter)) {
		int copy = 0;
		int max = size_goal;
L
Linus Torvalds 已提交
1125

1126 1127 1128 1129 1130
		skb = tcp_write_queue_tail(sk);
		if (tcp_send_head(sk)) {
			if (skb->ip_summed == CHECKSUM_NONE)
				max = mss_now;
			copy = max - skb->len;
1131
		}
L
Linus Torvalds 已提交
1132

1133
		if (copy <= 0) {
L
Linus Torvalds 已提交
1134
new_segment:
1135 1136 1137 1138 1139
			/* Allocate new segment. If the interface is SG,
			 * allocate skb fitting to single page.
			 */
			if (!sk_stream_memory_free(sk))
				goto wait_for_sndbuf;
L
Linus Torvalds 已提交
1140

1141 1142 1143 1144 1145
			skb = sk_stream_alloc_skb(sk,
						  select_size(sk, sg),
						  sk->sk_allocation);
			if (!skb)
				goto wait_for_memory;
L
Linus Torvalds 已提交
1146

1147 1148 1149 1150 1151
			/*
			 * Check whether we can use HW checksum.
			 */
			if (sk->sk_route_caps & NETIF_F_ALL_CSUM)
				skb->ip_summed = CHECKSUM_PARTIAL;
L
Linus Torvalds 已提交
1152

1153 1154 1155
			skb_entail(sk, skb);
			copy = size_goal;
			max = size_goal;
1156

1157 1158 1159 1160 1161 1162 1163
			/* All packets are restored as if they have
			 * already been sent. skb_mstamp isn't set to
			 * avoid wrong rtt estimation.
			 */
			if (tp->repair)
				TCP_SKB_CB(skb)->sacked |= TCPCB_REPAIRED;
		}
L
Linus Torvalds 已提交
1164

1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182
		/* Try to append data to the end of skb. */
		if (copy > iov_iter_count(&msg->msg_iter))
			copy = iov_iter_count(&msg->msg_iter);

		/* Where to copy to? */
		if (skb_availroom(skb) > 0) {
			/* We have some space in skb head. Superb! */
			copy = min_t(int, copy, skb_availroom(skb));
			err = skb_add_data_nocache(sk, skb, &msg->msg_iter, copy);
			if (err)
				goto do_fault;
		} else {
			bool merge = true;
			int i = skb_shinfo(skb)->nr_frags;
			struct page_frag *pfrag = sk_page_frag(sk);

			if (!sk_page_frag_refill(sk, pfrag))
				goto wait_for_memory;
1183

1184 1185 1186 1187 1188
			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
1189
				}
1190
				merge = false;
L
Linus Torvalds 已提交
1191 1192
			}

1193 1194 1195 1196
			copy = min_t(int, copy, pfrag->size - pfrag->offset);

			if (!sk_wmem_schedule(sk, copy))
				goto wait_for_memory;
L
Linus Torvalds 已提交
1197

1198 1199 1200 1201 1202 1203
			err = skb_copy_to_page_nocache(sk, &msg->msg_iter, skb,
						       pfrag->page,
						       pfrag->offset,
						       copy);
			if (err)
				goto do_error;
L
Linus Torvalds 已提交
1204

1205 1206 1207 1208 1209 1210 1211
			/* Update the skb. */
			if (merge) {
				skb_frag_size_add(&skb_shinfo(skb)->frags[i - 1], copy);
			} else {
				skb_fill_page_desc(skb, i, pfrag->page,
						   pfrag->offset, copy);
				get_page(pfrag->page);
1212
			}
1213 1214 1215 1216 1217 1218 1219 1220 1221
			pfrag->offset += copy;
		}

		if (!copied)
			TCP_SKB_CB(skb)->tcp_flags &= ~TCPHDR_PSH;

		tp->write_seq += copy;
		TCP_SKB_CB(skb)->end_seq += copy;
		tcp_skb_pcount_set(skb, 0);
L
Linus Torvalds 已提交
1222

1223 1224 1225 1226 1227
		copied += copy;
		if (!iov_iter_count(&msg->msg_iter)) {
			tcp_tx_timestamp(sk, skb);
			goto out;
		}
L
Linus Torvalds 已提交
1228

1229
		if (skb->len < max || (flags & MSG_OOB) || unlikely(tp->repair))
L
Linus Torvalds 已提交
1230 1231
			continue;

1232 1233 1234 1235 1236 1237 1238
		if (forced_push(tp)) {
			tcp_mark_push(tp, skb);
			__tcp_push_pending_frames(sk, mss_now, TCP_NAGLE_PUSH);
		} else if (skb == tcp_send_head(sk))
			tcp_push_one(sk, mss_now);
		continue;

L
Linus Torvalds 已提交
1239
wait_for_sndbuf:
1240
		set_bit(SOCK_NOSPACE, &sk->sk_socket->flags);
L
Linus Torvalds 已提交
1241
wait_for_memory:
1242 1243 1244
		if (copied)
			tcp_push(sk, flags & ~MSG_MORE, mss_now,
				 TCP_NAGLE_PUSH, size_goal);
L
Linus Torvalds 已提交
1245

1246 1247
		if ((err = sk_stream_wait_memory(sk, &timeo)) != 0)
			goto do_error;
L
Linus Torvalds 已提交
1248

1249
		mss_now = tcp_send_mss(sk, &size_goal, flags);
L
Linus Torvalds 已提交
1250 1251 1252
	}

out:
1253
	if (copied)
E
Eric Dumazet 已提交
1254
		tcp_push(sk, flags, mss_now, tp->nonagle, size_goal);
1255
out_nopush:
L
Linus Torvalds 已提交
1256
	release_sock(sk);
1257
	return copied + copied_syn;
L
Linus Torvalds 已提交
1258 1259 1260

do_fault:
	if (!skb->len) {
1261 1262 1263 1264 1265
		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);
1266
		sk_wmem_free_skb(sk, skb);
L
Linus Torvalds 已提交
1267 1268 1269
	}

do_error:
1270
	if (copied + copied_syn)
L
Linus Torvalds 已提交
1271 1272 1273 1274 1275 1276
		goto out;
out_err:
	err = sk_stream_error(sk, flags, err);
	release_sock(sk);
	return err;
}
E
Eric Dumazet 已提交
1277
EXPORT_SYMBOL(tcp_sendmsg);
L
Linus Torvalds 已提交
1278 1279 1280 1281 1282 1283

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

1284
static int tcp_recv_urg(struct sock *sk, struct msghdr *msg, int len, int flags)
L
Linus Torvalds 已提交
1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307
{
	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))
A
Al Viro 已提交
1308
				err = memcpy_to_msg(msg, &c, 1);
L
Linus Torvalds 已提交
1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327
			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
1328 1329 1330 1331 1332 1333 1334 1335
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) {
1336
		err = skb_copy_datagram_msg(skb, 0, msg, skb->len);
P
Pavel Emelyanov 已提交
1337 1338 1339 1340 1341 1342 1343 1344 1345
		if (err)
			break;

		copied += skb->len;
	}

	return err ?: copied;
}

L
Linus Torvalds 已提交
1346 1347 1348 1349 1350 1351
/* 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.
 */
1352
static void tcp_cleanup_rbuf(struct sock *sk, int copied)
L
Linus Torvalds 已提交
1353 1354
{
	struct tcp_sock *tp = tcp_sk(sk);
E
Eric Dumazet 已提交
1355
	bool time_to_ack = false;
L
Linus Torvalds 已提交
1356 1357 1358

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

1359
	WARN(skb && !before(tp->copied_seq, TCP_SKB_CB(skb)->end_seq),
J
Joe Perches 已提交
1360
	     "cleanup rbuf bug: copied %X seq %X rcvnxt %X\n",
1361
	     tp->copied_seq, TCP_SKB_CB(skb)->end_seq, tp->rcv_nxt);
L
Linus Torvalds 已提交
1362

1363 1364
	if (inet_csk_ack_scheduled(sk)) {
		const struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
1365 1366
		   /* Delayed ACKs frequently hit locked sockets during bulk
		    * receive. */
1367
		if (icsk->icsk_ack.blocked ||
L
Linus Torvalds 已提交
1368
		    /* Once-per-two-segments ACK was not sent by tcp_input.c */
1369
		    tp->rcv_nxt - tp->rcv_wup > icsk->icsk_ack.rcv_mss ||
L
Linus Torvalds 已提交
1370 1371 1372 1373 1374 1375
		    /*
		     * 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.
		     */
1376 1377 1378 1379 1380
		    (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
1381
			time_to_ack = true;
L
Linus Torvalds 已提交
1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402
	}

	/* 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
1403
				time_to_ack = true;
L
Linus Torvalds 已提交
1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414
		}
	}
	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);

1415
	NET_INC_STATS_USER(sock_net(sk), LINUX_MIB_TCPPREQUEUED);
L
Linus Torvalds 已提交
1416 1417 1418 1419 1420

	/* 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
1421
		sk_backlog_rcv(sk, skb);
L
Linus Torvalds 已提交
1422 1423 1424 1425 1426 1427
	local_bh_enable();

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

1428
static struct sk_buff *tcp_recv_skb(struct sock *sk, u32 seq, u32 *off)
L
Linus Torvalds 已提交
1429 1430 1431 1432
{
	struct sk_buff *skb;
	u32 offset;

1433
	while ((skb = skb_peek(&sk->sk_receive_queue)) != NULL) {
L
Linus Torvalds 已提交
1434
		offset = seq - TCP_SKB_CB(skb)->seq;
1435
		if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_SYN)
L
Linus Torvalds 已提交
1436
			offset--;
1437
		if (offset < skb->len || (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN)) {
L
Linus Torvalds 已提交
1438 1439 1440
			*off = offset;
			return skb;
		}
1441 1442 1443 1444
		/* This looks weird, but this can happen if TCP collapsing
		 * splitted a fat GRO packet, while we released socket lock
		 * in skb_splice_bits()
		 */
D
Dan Williams 已提交
1445
		sk_eat_skb(sk, skb);
L
Linus Torvalds 已提交
1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473
	}
	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) {
1474 1475
			int used;
			size_t len;
L
Linus Torvalds 已提交
1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486

			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);
1487
			if (used <= 0) {
1488 1489 1490 1491
				if (!copied)
					copied = used;
				break;
			} else if (used <= len) {
L
Linus Torvalds 已提交
1492 1493 1494 1495
				seq += used;
				copied += used;
				offset += used;
			}
1496
			/* If recv_actor drops the lock (e.g. TCP splice
1497 1498 1499 1500
			 * receive) the skb pointer might be invalid when
			 * getting here: tcp_collapse might have deleted it
			 * while aggregating skbs from the socket queue.
			 */
1501 1502
			skb = tcp_recv_skb(sk, seq - 1, &offset);
			if (!skb)
L
Linus Torvalds 已提交
1503
				break;
1504 1505 1506 1507 1508
			/* 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
1509
		}
1510
		if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN) {
D
Dan Williams 已提交
1511
			sk_eat_skb(sk, skb);
L
Linus Torvalds 已提交
1512 1513 1514
			++seq;
			break;
		}
D
Dan Williams 已提交
1515
		sk_eat_skb(sk, skb);
L
Linus Torvalds 已提交
1516 1517
		if (!desc->count)
			break;
1518
		tp->copied_seq = seq;
L
Linus Torvalds 已提交
1519 1520 1521 1522 1523 1524
	}
	tp->copied_seq = seq;

	tcp_rcv_space_adjust(sk);

	/* Clean up data we have read: This will do ACK frames. */
1525 1526
	if (copied > 0) {
		tcp_recv_skb(sk, seq, &offset);
1527
		tcp_cleanup_rbuf(sk, copied);
1528
	}
L
Linus Torvalds 已提交
1529 1530
	return copied;
}
E
Eric Dumazet 已提交
1531
EXPORT_SYMBOL(tcp_read_sock);
L
Linus Torvalds 已提交
1532 1533 1534 1535 1536 1537 1538 1539 1540

/*
 *	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.
 */

1541 1542
int tcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, int nonblock,
		int flags, int *addr_len)
L
Linus Torvalds 已提交
1543 1544 1545 1546 1547 1548 1549 1550 1551 1552
{
	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;
1553
	struct sk_buff *skb;
I
Ilpo Järvinen 已提交
1554
	u32 urg_hole = 0;
L
Linus Torvalds 已提交
1555

1556
	if (unlikely(flags & MSG_ERRQUEUE))
1557
		return inet_recv_error(sk, msg, len, addr_len);
1558

1559 1560 1561
	if (sk_can_busy_loop(sk) && skb_queue_empty(&sk->sk_receive_queue) &&
	    (sk->sk_state == TCP_ESTABLISHED))
		sk_busy_loop(sk, nonblock);
1562

L
Linus Torvalds 已提交
1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574
	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
1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589
	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
1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612
	seq = &tp->copied_seq;
	if (flags & MSG_PEEK) {
		peek_seq = tp->copied_seq;
		seq = &peek_seq;
	}

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

	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. */

1613
		skb_queue_walk(&sk->sk_receive_queue, skb) {
L
Linus Torvalds 已提交
1614 1615 1616
			/* Now that we have two receive queues this
			 * shouldn't happen.
			 */
1617
			if (WARN(before(*seq, TCP_SKB_CB(skb)->seq),
J
Joe Perches 已提交
1618 1619 1620
				 "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
1621
				break;
1622

L
Linus Torvalds 已提交
1623
			offset = *seq - TCP_SKB_CB(skb)->seq;
1624
			if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_SYN)
L
Linus Torvalds 已提交
1625 1626 1627
				offset--;
			if (offset < skb->len)
				goto found_ok_skb;
1628
			if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN)
L
Linus Torvalds 已提交
1629
				goto found_fin_ok;
J
Joe Perches 已提交
1630 1631 1632
			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);
1633
		}
L
Linus Torvalds 已提交
1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644

		/* 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 ||
1645
			    signal_pending(current))
L
Linus Torvalds 已提交
1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680
				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;
			}
		}

1681
		tcp_cleanup_rbuf(sk, copied);
L
Linus Torvalds 已提交
1682

1683
		if (!sysctl_tcp_low_latency && tp->ucopy.task == user_recv) {
L
Linus Torvalds 已提交
1684 1685 1686 1687
			/* Install new reader */
			if (!user_recv && !(flags & (MSG_TRUNC | MSG_PEEK))) {
				user_recv = current;
				tp->ucopy.task = user_recv;
1688
				tp->ucopy.msg = msg;
L
Linus Torvalds 已提交
1689 1690 1691 1692
			}

			tp->ucopy.len = len;

1693 1694
			WARN_ON(tp->copied_seq != tp->rcv_nxt &&
				!(flags & (MSG_PEEK | MSG_TRUNC)));
L
Linus Torvalds 已提交
1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721

			/* 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.
			 */
1722
			if (!skb_queue_empty(&tp->ucopy.prequeue))
L
Linus Torvalds 已提交
1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740
				goto do_prequeue;

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

		if (copied >= target) {
			/* Do not sleep, just process backlog. */
			release_sock(sk);
			lock_sock(sk);
		} else
			sk_wait_data(sk, &timeo);

		if (user_recv) {
			int chunk;

			/* __ Restore normal policy in scheduler __ */

			if ((chunk = len - tp->ucopy.len) != 0) {
1741
				NET_ADD_STATS_USER(sock_net(sk), LINUX_MIB_TCPDIRECTCOPYFROMBACKLOG, chunk);
L
Linus Torvalds 已提交
1742 1743 1744 1745 1746
				len -= chunk;
				copied += chunk;
			}

			if (tp->rcv_nxt == tp->copied_seq &&
1747
			    !skb_queue_empty(&tp->ucopy.prequeue)) {
L
Linus Torvalds 已提交
1748 1749 1750 1751
do_prequeue:
				tcp_prequeue_process(sk);

				if ((chunk = len - tp->ucopy.len) != 0) {
1752
					NET_ADD_STATS_USER(sock_net(sk), LINUX_MIB_TCPDIRECTCOPYFROMPREQUEUE, chunk);
L
Linus Torvalds 已提交
1753 1754 1755 1756 1757
					len -= chunk;
					copied += chunk;
				}
			}
		}
I
Ilpo Järvinen 已提交
1758 1759
		if ((flags & MSG_PEEK) &&
		    (peek_seq - copied - urg_hole != tp->copied_seq)) {
1760 1761 1762
			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
1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779
			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
1780
						urg_hole++;
L
Linus Torvalds 已提交
1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791
						offset++;
						used--;
						if (!used)
							goto skip_copy;
					}
				} else
					used = urg_offset;
			}
		}

		if (!(flags & MSG_TRUNC)) {
1792
			err = skb_copy_datagram_msg(skb, offset, msg, used);
D
Dan Williams 已提交
1793 1794 1795 1796 1797
			if (err) {
				/* Exception. Bailout! */
				if (!copied)
					copied = -EFAULT;
				break;
L
Linus Torvalds 已提交
1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809
			}
		}

		*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;
1810
			tcp_fast_path_check(sk);
L
Linus Torvalds 已提交
1811 1812 1813 1814
		}
		if (used + offset < skb->len)
			continue;

1815
		if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN)
L
Linus Torvalds 已提交
1816
			goto found_fin_ok;
D
Dan Williams 已提交
1817 1818
		if (!(flags & MSG_PEEK))
			sk_eat_skb(sk, skb);
L
Linus Torvalds 已提交
1819 1820 1821 1822 1823
		continue;

	found_fin_ok:
		/* Process the FIN. */
		++*seq;
D
Dan Williams 已提交
1824 1825
		if (!(flags & MSG_PEEK))
			sk_eat_skb(sk, skb);
L
Linus Torvalds 已提交
1826 1827 1828 1829
		break;
	} while (len > 0);

	if (user_recv) {
1830
		if (!skb_queue_empty(&tp->ucopy.prequeue)) {
L
Linus Torvalds 已提交
1831 1832 1833 1834 1835 1836 1837
			int chunk;

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

			tcp_prequeue_process(sk);

			if (copied > 0 && (chunk = len - tp->ucopy.len) != 0) {
1838
				NET_ADD_STATS_USER(sock_net(sk), LINUX_MIB_TCPDIRECTCOPYFROMPREQUEUE, chunk);
L
Linus Torvalds 已提交
1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852
				len -= chunk;
				copied += chunk;
			}
		}

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

	/* 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. */
1853
	tcp_cleanup_rbuf(sk, copied);
L
Linus Torvalds 已提交
1854 1855 1856 1857 1858 1859 1860 1861 1862

	release_sock(sk);
	return copied;

out:
	release_sock(sk);
	return err;

recv_urg:
1863
	err = tcp_recv_urg(sk, msg, len, flags);
L
Linus Torvalds 已提交
1864
	goto out;
P
Pavel Emelyanov 已提交
1865 1866 1867 1868

recv_sndq:
	err = tcp_peek_sndq(sk, msg, len);
	goto out;
L
Linus Torvalds 已提交
1869
}
E
Eric Dumazet 已提交
1870
EXPORT_SYMBOL(tcp_recvmsg);
L
Linus Torvalds 已提交
1871

I
Ilpo Järvinen 已提交
1872 1873 1874 1875 1876 1877 1878
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
1879
			TCP_INC_STATS(sock_net(sk), TCP_MIB_CURRESTAB);
I
Ilpo Järvinen 已提交
1880 1881 1882 1883
		break;

	case TCP_CLOSE:
		if (oldstate == TCP_CLOSE_WAIT || oldstate == TCP_ESTABLISHED)
P
Pavel Emelyanov 已提交
1884
			TCP_INC_STATS(sock_net(sk), TCP_MIB_ESTABRESETS);
I
Ilpo Järvinen 已提交
1885 1886 1887 1888

		sk->sk_prot->unhash(sk);
		if (inet_csk(sk)->icsk_bind_hash &&
		    !(sk->sk_userlocks & SOCK_BINDPORT_LOCK))
1889
			inet_put_port(sk);
I
Ilpo Järvinen 已提交
1890 1891
		/* fall through */
	default:
1892
		if (oldstate == TCP_ESTABLISHED)
P
Pavel Emelyanov 已提交
1893
			TCP_DEC_STATS(sock_net(sk), TCP_MIB_CURRESTAB);
I
Ilpo Järvinen 已提交
1894 1895 1896 1897 1898 1899 1900 1901
	}

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

#ifdef STATE_TRACE
1902
	SOCK_DEBUG(sk, "TCP sk=%p, State %s -> %s\n", sk, statename[oldstate], statename[state]);
I
Ilpo Järvinen 已提交
1903 1904 1905 1906
#endif
}
EXPORT_SYMBOL_GPL(tcp_set_state);

L
Linus Torvalds 已提交
1907 1908 1909 1910 1911 1912 1913
/*
 *	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.
 */

1914
static const unsigned char new_state[16] = {
L
Linus Torvalds 已提交
1915
  /* current state:        new state:      action:	*/
1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928
  [0 /* (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,
  [TCP_NEW_SYN_RECV]	= TCP_CLOSE,	/* should not happen ! */
L
Linus Torvalds 已提交
1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942
};

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
1943
 *	that we don't receive shut down or sock_set_flag(sk, SOCK_DEAD).
L
Linus Torvalds 已提交
1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963
 */

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
1964
EXPORT_SYMBOL(tcp_shutdown);
L
Linus Torvalds 已提交
1965

A
Arun Sharma 已提交
1966 1967 1968 1969 1970 1971 1972
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);

1973 1974 1975 1976
	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
1977 1978 1979
	return too_many_orphans || out_of_socket_memory;
}

L
Linus Torvalds 已提交
1980 1981 1982 1983
void tcp_close(struct sock *sk, long timeout)
{
	struct sk_buff *skb;
	int data_was_unread = 0;
H
Herbert Xu 已提交
1984
	int state;
L
Linus Torvalds 已提交
1985 1986 1987 1988 1989 1990 1991 1992

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

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

		/* Special case. */
1993
		inet_csk_listen_stop(sk);
L
Linus Torvalds 已提交
1994 1995 1996 1997 1998 1999 2000 2001 2002

		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) {
2003 2004 2005 2006
		u32 len = TCP_SKB_CB(skb)->end_seq - TCP_SKB_CB(skb)->seq;

		if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN)
			len--;
L
Linus Torvalds 已提交
2007 2008 2009 2010
		data_was_unread += len;
		__kfree_skb(skb);
	}

2011
	sk_mem_reclaim(sk);
L
Linus Torvalds 已提交
2012

2013 2014 2015 2016
	/* If socket has been already reset (e.g. in tcp_reset()) - kill it. */
	if (sk->sk_state == TCP_CLOSE)
		goto adjudge_to_death;

2017 2018 2019 2020 2021 2022
	/* 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
2023
	 */
P
Pavel Emelyanov 已提交
2024 2025 2026
	if (unlikely(tcp_sk(sk)->repair)) {
		sk->sk_prot->disconnect(sk, 0);
	} else if (data_was_unread) {
L
Linus Torvalds 已提交
2027
		/* Unread data was tossed, zap the connection. */
2028
		NET_INC_STATS_USER(sock_net(sk), LINUX_MIB_TCPABORTONCLOSE);
L
Linus Torvalds 已提交
2029
		tcp_set_state(sk, TCP_CLOSE);
2030
		tcp_send_active_reset(sk, sk->sk_allocation);
L
Linus Torvalds 已提交
2031 2032 2033
	} else if (sock_flag(sk, SOCK_LINGER) && !sk->sk_lingertime) {
		/* Check zero linger _after_ checking for unread data. */
		sk->sk_prot->disconnect(sk, 0);
2034
		NET_INC_STATS_USER(sock_net(sk), LINUX_MIB_TCPABORTONDATA);
L
Linus Torvalds 已提交
2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059
	} 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
2060 2061 2062 2063
		 * 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
2064 2065 2066 2067 2068 2069 2070
		 */
		tcp_send_fin(sk);
	}

	sk_stream_wait_close(sk, timeout);

adjudge_to_death:
H
Herbert Xu 已提交
2071 2072 2073 2074
	state = sk->sk_state;
	sock_hold(sk);
	sock_orphan(sk);

L
Linus Torvalds 已提交
2075 2076 2077 2078 2079 2080 2081 2082 2083
	/* 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);
2084
	WARN_ON(sock_owned_by_user(sk));
L
Linus Torvalds 已提交
2085

H
Herbert Xu 已提交
2086 2087
	percpu_counter_inc(sk->sk_prot->orphan_count);

H
Herbert Xu 已提交
2088 2089 2090
	/* 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
2091 2092 2093 2094

	/*	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.
2095
	 *	We use a 1 minute timeout (about the same as BSD) then kill
L
Linus Torvalds 已提交
2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110
	 *	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);
2111 2112
			NET_INC_STATS_BH(sock_net(sk),
					LINUX_MIB_TCPABORTONLINGER);
L
Linus Torvalds 已提交
2113
		} else {
2114
			const int tmo = tcp_fin_time(sk);
L
Linus Torvalds 已提交
2115 2116

			if (tmo > TCP_TIMEWAIT_LEN) {
2117 2118
				inet_csk_reset_keepalive_timer(sk,
						tmo - TCP_TIMEWAIT_LEN);
L
Linus Torvalds 已提交
2119 2120 2121 2122 2123 2124 2125
			} else {
				tcp_time_wait(sk, TCP_FIN_WAIT2, tmo);
				goto out;
			}
		}
	}
	if (sk->sk_state != TCP_CLOSE) {
2126
		sk_mem_reclaim(sk);
A
Arun Sharma 已提交
2127
		if (tcp_check_oom(sk, 0)) {
L
Linus Torvalds 已提交
2128 2129
			tcp_set_state(sk, TCP_CLOSE);
			tcp_send_active_reset(sk, GFP_ATOMIC);
2130 2131
			NET_INC_STATS_BH(sock_net(sk),
					LINUX_MIB_TCPABORTONMEMORY);
L
Linus Torvalds 已提交
2132 2133 2134
		}
	}

2135 2136 2137 2138 2139 2140 2141 2142
	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);
2143
		inet_csk_destroy_sock(sk);
2144
	}
L
Linus Torvalds 已提交
2145 2146 2147 2148 2149 2150 2151
	/* Otherwise, socket is reprieved until protocol close. */

out:
	bh_unlock_sock(sk);
	local_bh_enable();
	sock_put(sk);
}
E
Eric Dumazet 已提交
2152
EXPORT_SYMBOL(tcp_close);
L
Linus Torvalds 已提交
2153 2154 2155

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

E
Eric Dumazet 已提交
2156
static inline bool tcp_need_reset(int state)
L
Linus Torvalds 已提交
2157 2158 2159 2160 2161 2162 2163 2164 2165
{
	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);
2166
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
2167 2168 2169 2170 2171 2172 2173 2174 2175
	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) {
2176
		inet_csk_listen_stop(sk);
P
Pavel Emelyanov 已提交
2177 2178
	} else if (unlikely(tp->repair)) {
		sk->sk_err = ECONNABORTED;
L
Linus Torvalds 已提交
2179 2180 2181
	} 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
2182
		/* The last check adjusts for discrepancy of Linux wrt. RFC
L
Linus Torvalds 已提交
2183 2184 2185 2186 2187 2188 2189 2190 2191
		 * 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);
2192
	tcp_write_queue_purge(sk);
L
Linus Torvalds 已提交
2193 2194
	__skb_queue_purge(&tp->out_of_order_queue);

E
Eric Dumazet 已提交
2195
	inet->inet_dport = 0;
L
Linus Torvalds 已提交
2196 2197 2198 2199 2200 2201

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

	sk->sk_shutdown = 0;
	sock_reset_flag(sk, SOCK_DONE);
2202
	tp->srtt_us = 0;
L
Linus Torvalds 已提交
2203 2204
	if ((tp->write_seq += tp->max_window + 2) == 0)
		tp->write_seq = 1;
2205
	icsk->icsk_backoff = 0;
L
Linus Torvalds 已提交
2206
	tp->snd_cwnd = 2;
2207
	icsk->icsk_probes_out = 0;
L
Linus Torvalds 已提交
2208
	tp->packets_out = 0;
I
Ilpo Järvinen 已提交
2209
	tp->snd_ssthresh = TCP_INFINITE_SSTHRESH;
L
Linus Torvalds 已提交
2210
	tp->snd_cwnd_cnt = 0;
2211
	tp->window_clamp = 0;
2212
	tcp_set_ca_state(sk, TCP_CA_Open);
L
Linus Torvalds 已提交
2213
	tcp_clear_retrans(tp);
2214
	inet_csk_delack_init(sk);
2215
	tcp_init_send_head(sk);
2216
	memset(&tp->rx_opt, 0, sizeof(tp->rx_opt));
L
Linus Torvalds 已提交
2217 2218
	__sk_dst_reset(sk);

E
Eric Dumazet 已提交
2219
	WARN_ON(inet->inet_num && !icsk->icsk_bind_hash);
L
Linus Torvalds 已提交
2220 2221 2222 2223

	sk->sk_error_report(sk);
	return err;
}
E
Eric Dumazet 已提交
2224
EXPORT_SYMBOL(tcp_disconnect);
L
Linus Torvalds 已提交
2225

2226 2227 2228 2229 2230 2231 2232
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
2233
static inline bool tcp_can_repair_sock(const struct sock *sk)
P
Pavel Emelyanov 已提交
2234
{
2235
	return ns_capable(sock_net(sk)->user_ns, CAP_NET_ADMIN) &&
P
Pavel Emelyanov 已提交
2236 2237 2238
		((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_ESTABLISHED));
}

2239 2240
static int tcp_repair_options_est(struct tcp_sock *tp,
		struct tcp_repair_opt __user *optbuf, unsigned int len)
2241
{
2242
	struct tcp_repair_opt opt;
2243

2244 2245
	while (len >= sizeof(opt)) {
		if (copy_from_user(&opt, optbuf, sizeof(opt)))
2246 2247 2248
			return -EFAULT;

		optbuf++;
2249
		len -= sizeof(opt);
2250

2251 2252 2253
		switch (opt.opt_code) {
		case TCPOPT_MSS:
			tp->rx_opt.mss_clamp = opt.opt_val;
2254
			break;
2255
		case TCPOPT_WINDOW:
2256 2257 2258 2259 2260 2261
			{
				u16 snd_wscale = opt.opt_val & 0xFFFF;
				u16 rcv_wscale = opt.opt_val >> 16;

				if (snd_wscale > 14 || rcv_wscale > 14)
					return -EFBIG;
2262

2263 2264 2265 2266
				tp->rx_opt.snd_wscale = snd_wscale;
				tp->rx_opt.rcv_wscale = rcv_wscale;
				tp->rx_opt.wscale_ok = 1;
			}
2267 2268
			break;
		case TCPOPT_SACK_PERM:
2269 2270 2271
			if (opt.opt_val != 0)
				return -EINVAL;

2272 2273 2274 2275 2276
			tp->rx_opt.sack_ok |= TCP_SACK_SEEN;
			if (sysctl_tcp_fack)
				tcp_enable_fack(tp);
			break;
		case TCPOPT_TIMESTAMP:
2277 2278 2279
			if (opt.opt_val != 0)
				return -EINVAL;

2280 2281 2282 2283 2284 2285 2286 2287
			tp->rx_opt.tstamp_ok = 1;
			break;
		}
	}

	return 0;
}

L
Linus Torvalds 已提交
2288 2289 2290
/*
 *	Socket option code for TCP.
 */
2291
static int do_tcp_setsockopt(struct sock *sk, int level,
2292
		int optname, char __user *optval, unsigned int optlen)
L
Linus Torvalds 已提交
2293 2294
{
	struct tcp_sock *tp = tcp_sk(sk);
2295
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
2296 2297 2298
	int val;
	int err = 0;

2299 2300 2301
	/* These are data/string values, all the others are ints */
	switch (optname) {
	case TCP_CONGESTION: {
2302 2303 2304 2305 2306 2307
		char name[TCP_CA_NAME_MAX];

		if (optlen < 1)
			return -EINVAL;

		val = strncpy_from_user(name, optval,
2308
					min_t(long, TCP_CA_NAME_MAX-1, optlen));
2309 2310 2311 2312 2313
		if (val < 0)
			return -EFAULT;
		name[val] = 0;

		lock_sock(sk);
2314
		err = tcp_set_congestion_control(sk, name);
2315 2316 2317
		release_sock(sk);
		return err;
	}
2318 2319 2320
	default:
		/* fallthru */
		break;
2321
	}
2322

L
Linus Torvalds 已提交
2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335
	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 */
2336
		if (val < TCP_MIN_MSS || val > MAX_TCP_WINDOW) {
L
Linus Torvalds 已提交
2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353
			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;
2354
			tcp_push_pending_frames(sk);
L
Linus Torvalds 已提交
2355 2356 2357 2358 2359
		} else {
			tp->nonagle &= ~TCP_NAGLE_OFF;
		}
		break;

A
Andreas Petlund 已提交
2360 2361 2362 2363 2364 2365 2366
	case TCP_THIN_LINEAR_TIMEOUTS:
		if (val < 0 || val > 1)
			err = -EINVAL;
		else
			tp->thin_lto = val;
		break;

A
Andreas Petlund 已提交
2367 2368 2369
	case TCP_THIN_DUPACK:
		if (val < 0 || val > 1)
			err = -EINVAL;
2370
		else {
A
Andreas Petlund 已提交
2371
			tp->thin_dupack = val;
Y
Yuchung Cheng 已提交
2372 2373
			if (tp->thin_dupack)
				tcp_disable_early_retrans(tp);
2374
		}
A
Andreas Petlund 已提交
2375 2376
		break;

P
Pavel Emelyanov 已提交
2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412
	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;

2413 2414 2415 2416
	case TCP_REPAIR_OPTIONS:
		if (!tp->repair)
			err = -EINVAL;
		else if (sk->sk_state == TCP_ESTABLISHED)
2417 2418 2419
			err = tcp_repair_options_est(tp,
					(struct tcp_repair_opt __user *)optval,
					optlen);
2420 2421 2422 2423
		else
			err = -EPERM;
		break;

L
Linus Torvalds 已提交
2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441
	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;
2442
			tcp_push_pending_frames(sk);
L
Linus Torvalds 已提交
2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453
		}
		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))) {
2454
				u32 elapsed = keepalive_time_elapsed(tp);
L
Linus Torvalds 已提交
2455 2456 2457 2458
				if (tp->keepalive_time > elapsed)
					elapsed = tp->keepalive_time - elapsed;
				else
					elapsed = 0;
2459
				inet_csk_reset_keepalive_timer(sk, elapsed);
L
Linus Torvalds 已提交
2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478
			}
		}
		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
2479
			icsk->icsk_syn_retries = val;
L
Linus Torvalds 已提交
2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491
		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:
2492 2493 2494 2495
		/* 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
2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511
		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) {
2512
			icsk->icsk_ack.pingpong = 1;
L
Linus Torvalds 已提交
2513
		} else {
2514
			icsk->icsk_ack.pingpong = 0;
L
Linus Torvalds 已提交
2515 2516
			if ((1 << sk->sk_state) &
			    (TCPF_ESTABLISHED | TCPF_CLOSE_WAIT) &&
2517 2518
			    inet_csk_ack_scheduled(sk)) {
				icsk->icsk_ack.pending |= ICSK_ACK_PUSHED;
2519
				tcp_cleanup_rbuf(sk, 1);
L
Linus Torvalds 已提交
2520
				if (!(val & 1))
2521
					icsk->icsk_ack.pingpong = 1;
L
Linus Torvalds 已提交
2522 2523 2524 2525
			}
		}
		break;

2526 2527 2528 2529 2530 2531
#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
2532
	case TCP_USER_TIMEOUT:
2533
		/* Cap the max time in ms TCP will retry or probe the window
2534 2535
		 * before giving up and aborting (ETIMEDOUT) a connection.
		 */
2536 2537 2538 2539
		if (val < 0)
			err = -EINVAL;
		else
			icsk->icsk_user_timeout = msecs_to_jiffies(val);
2540
		break;
2541 2542 2543 2544 2545 2546 2547 2548

	case TCP_FASTOPEN:
		if (val >= 0 && ((1 << sk->sk_state) & (TCPF_CLOSE |
		    TCPF_LISTEN)))
			err = fastopen_init_queue(sk, val);
		else
			err = -EINVAL;
		break;
2549 2550 2551 2552 2553 2554
	case TCP_TIMESTAMP:
		if (!tp->repair)
			err = -EPERM;
		else
			tp->tsoffset = val - tcp_time_stamp;
		break;
2555 2556 2557 2558
	case TCP_NOTSENT_LOWAT:
		tp->notsent_lowat = val;
		sk->sk_write_space(sk);
		break;
L
Linus Torvalds 已提交
2559 2560 2561
	default:
		err = -ENOPROTOOPT;
		break;
2562 2563
	}

L
Linus Torvalds 已提交
2564 2565 2566 2567
	release_sock(sk);
	return err;
}

2568
int tcp_setsockopt(struct sock *sk, int level, int optname, char __user *optval,
2569
		   unsigned int optlen)
2570
{
2571
	const struct inet_connection_sock *icsk = inet_csk(sk);
2572 2573 2574 2575 2576 2577

	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
2578
EXPORT_SYMBOL(tcp_setsockopt);
2579 2580

#ifdef CONFIG_COMPAT
2581
int compat_tcp_setsockopt(struct sock *sk, int level, int optname,
2582
			  char __user *optval, unsigned int optlen)
2583
{
2584 2585 2586
	if (level != SOL_TCP)
		return inet_csk_compat_setsockopt(sk, level, optname,
						  optval, optlen);
2587 2588
	return do_tcp_setsockopt(sk, level, optname, optval, optlen);
}
2589
EXPORT_SYMBOL(compat_tcp_setsockopt);
2590 2591
#endif

L
Linus Torvalds 已提交
2592
/* Return information about state of tcp endpoint in API format. */
2593
void tcp_get_info(const struct sock *sk, struct tcp_info *info)
L
Linus Torvalds 已提交
2594
{
2595
	const struct tcp_sock *tp = tcp_sk(sk);
2596
	const struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
2597 2598 2599 2600 2601
	u32 now = tcp_time_stamp;

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

	info->tcpi_state = sk->sk_state;
2602
	info->tcpi_ca_state = icsk->icsk_ca_state;
2603
	info->tcpi_retransmits = icsk->icsk_retransmits;
2604
	info->tcpi_probes = icsk->icsk_probes_out;
2605
	info->tcpi_backoff = icsk->icsk_backoff;
L
Linus Torvalds 已提交
2606 2607 2608

	if (tp->rx_opt.tstamp_ok)
		info->tcpi_options |= TCPI_OPT_TIMESTAMPS;
2609
	if (tcp_is_sack(tp))
L
Linus Torvalds 已提交
2610 2611 2612 2613 2614
		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;
2615
	}
L
Linus Torvalds 已提交
2616

E
Eric Dumazet 已提交
2617
	if (tp->ecn_flags & TCP_ECN_OK)
L
Linus Torvalds 已提交
2618
		info->tcpi_options |= TCPI_OPT_ECN;
E
Eric Dumazet 已提交
2619 2620
	if (tp->ecn_flags & TCP_ECN_SEEN)
		info->tcpi_options |= TCPI_OPT_ECN_SEEN;
Y
Yuchung Cheng 已提交
2621 2622
	if (tp->syn_data_acked)
		info->tcpi_options |= TCPI_OPT_SYN_DATA;
L
Linus Torvalds 已提交
2623

2624 2625
	info->tcpi_rto = jiffies_to_usecs(icsk->icsk_rto);
	info->tcpi_ato = jiffies_to_usecs(icsk->icsk_ack.ato);
2626
	info->tcpi_snd_mss = tp->mss_cache;
2627
	info->tcpi_rcv_mss = icsk->icsk_ack.rcv_mss;
L
Linus Torvalds 已提交
2628

2629 2630 2631 2632 2633 2634 2635
	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
2636 2637 2638 2639 2640
	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);
2641
	info->tcpi_last_data_recv = jiffies_to_msecs(now - icsk->icsk_ack.lrcvtime);
L
Linus Torvalds 已提交
2642 2643
	info->tcpi_last_ack_recv = jiffies_to_msecs(now - tp->rcv_tstamp);

2644
	info->tcpi_pmtu = icsk->icsk_pmtu_cookie;
L
Linus Torvalds 已提交
2645
	info->tcpi_rcv_ssthresh = tp->rcv_ssthresh;
2646 2647
	info->tcpi_rtt = tp->srtt_us >> 3;
	info->tcpi_rttvar = tp->mdev_us >> 2;
L
Linus Torvalds 已提交
2648 2649 2650 2651 2652 2653 2654 2655 2656
	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;
2657 2658 2659 2660 2661

	info->tcpi_pacing_rate = sk->sk_pacing_rate != ~0U ?
					sk->sk_pacing_rate : ~0ULL;
	info->tcpi_max_pacing_rate = sk->sk_max_pacing_rate != ~0U ?
					sk->sk_max_pacing_rate : ~0ULL;
L
Linus Torvalds 已提交
2662 2663 2664
}
EXPORT_SYMBOL_GPL(tcp_get_info);

2665 2666
static int do_tcp_getsockopt(struct sock *sk, int level,
		int optname, char __user *optval, int __user *optlen)
L
Linus Torvalds 已提交
2667
{
2668
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681
	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:
2682
		val = tp->mss_cache;
L
Linus Torvalds 已提交
2683 2684
		if (!val && ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN)))
			val = tp->rx_opt.user_mss;
2685 2686
		if (tp->repair)
			val = tp->rx_opt.mss_clamp;
L
Linus Torvalds 已提交
2687 2688 2689 2690 2691 2692 2693 2694
		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
2695
		val = keepalive_time_when(tp) / HZ;
L
Linus Torvalds 已提交
2696 2697
		break;
	case TCP_KEEPINTVL:
E
Eric Dumazet 已提交
2698
		val = keepalive_intvl_when(tp) / HZ;
L
Linus Torvalds 已提交
2699 2700
		break;
	case TCP_KEEPCNT:
E
Eric Dumazet 已提交
2701
		val = keepalive_probes(tp);
L
Linus Torvalds 已提交
2702 2703
		break;
	case TCP_SYNCNT:
2704
		val = icsk->icsk_syn_retries ? : sysctl_tcp_syn_retries;
L
Linus Torvalds 已提交
2705 2706 2707 2708 2709 2710 2711
		break;
	case TCP_LINGER2:
		val = tp->linger2;
		if (val >= 0)
			val = (val ? : sysctl_tcp_fin_timeout) / HZ;
		break;
	case TCP_DEFER_ACCEPT:
2712 2713
		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
2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733
		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:
2734
		val = !icsk->icsk_ack.pingpong;
L
Linus Torvalds 已提交
2735
		break;
2736 2737 2738 2739 2740 2741 2742

	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;
2743
		if (copy_to_user(optval, icsk->icsk_ca_ops->name, len))
2744 2745
			return -EFAULT;
		return 0;
2746

2747 2748 2749 2750 2751 2752
	case TCP_THIN_LINEAR_TIMEOUTS:
		val = tp->thin_lto;
		break;
	case TCP_THIN_DUPACK:
		val = tp->thin_dupack;
		break;
2753

P
Pavel Emelyanov 已提交
2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773
	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;

2774 2775 2776
	case TCP_USER_TIMEOUT:
		val = jiffies_to_msecs(icsk->icsk_user_timeout);
		break;
2777 2778 2779 2780 2781 2782 2783 2784

	case TCP_FASTOPEN:
		if (icsk->icsk_accept_queue.fastopenq != NULL)
			val = icsk->icsk_accept_queue.fastopenq->max_qlen;
		else
			val = 0;
		break;

2785 2786 2787
	case TCP_TIMESTAMP:
		val = tcp_time_stamp + tp->tsoffset;
		break;
2788 2789 2790
	case TCP_NOTSENT_LOWAT:
		val = tp->notsent_lowat;
		break;
L
Linus Torvalds 已提交
2791 2792
	default:
		return -ENOPROTOOPT;
2793
	}
L
Linus Torvalds 已提交
2794 2795 2796 2797 2798 2799 2800 2801

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

2802 2803 2804 2805 2806 2807 2808 2809 2810 2811
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
2812
EXPORT_SYMBOL(tcp_getsockopt);
2813 2814

#ifdef CONFIG_COMPAT
2815 2816
int compat_tcp_getsockopt(struct sock *sk, int level, int optname,
			  char __user *optval, int __user *optlen)
2817
{
2818 2819 2820
	if (level != SOL_TCP)
		return inet_csk_compat_getsockopt(sk, level, optname,
						  optval, optlen);
2821 2822
	return do_tcp_getsockopt(sk, level, optname, optval, optlen);
}
2823
EXPORT_SYMBOL(compat_tcp_getsockopt);
2824
#endif
L
Linus Torvalds 已提交
2825

2826
#ifdef CONFIG_TCP_MD5SIG
E
Eric Dumazet 已提交
2827
static DEFINE_PER_CPU(struct tcp_md5sig_pool, tcp_md5sig_pool);
2828
static DEFINE_MUTEX(tcp_md5sig_mutex);
E
Eric Dumazet 已提交
2829
static bool tcp_md5sig_pool_populated = false;
2830

2831
static void __tcp_alloc_md5sig_pool(void)
2832 2833 2834 2835
{
	int cpu;

	for_each_possible_cpu(cpu) {
E
Eric Dumazet 已提交
2836 2837
		if (!per_cpu(tcp_md5sig_pool, cpu).md5_desc.tfm) {
			struct crypto_hash *hash;
2838

E
Eric Dumazet 已提交
2839 2840 2841 2842 2843
			hash = crypto_alloc_hash("md5", 0, CRYPTO_ALG_ASYNC);
			if (IS_ERR_OR_NULL(hash))
				return;
			per_cpu(tcp_md5sig_pool, cpu).md5_desc.tfm = hash;
		}
2844
	}
E
Eric Dumazet 已提交
2845 2846
	/* before setting tcp_md5sig_pool_populated, we must commit all writes
	 * to memory. See smp_rmb() in tcp_get_md5sig_pool()
2847 2848
	 */
	smp_wmb();
E
Eric Dumazet 已提交
2849
	tcp_md5sig_pool_populated = true;
2850 2851
}

2852
bool tcp_alloc_md5sig_pool(void)
2853
{
E
Eric Dumazet 已提交
2854
	if (unlikely(!tcp_md5sig_pool_populated)) {
2855 2856
		mutex_lock(&tcp_md5sig_mutex);

E
Eric Dumazet 已提交
2857
		if (!tcp_md5sig_pool_populated)
2858 2859 2860
			__tcp_alloc_md5sig_pool();

		mutex_unlock(&tcp_md5sig_mutex);
2861
	}
E
Eric Dumazet 已提交
2862
	return tcp_md5sig_pool_populated;
2863 2864 2865
}
EXPORT_SYMBOL(tcp_alloc_md5sig_pool);

E
Eric Dumazet 已提交
2866 2867 2868 2869 2870 2871 2872 2873 2874

/**
 *	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)
2875
{
E
Eric Dumazet 已提交
2876
	local_bh_disable();
2877

E
Eric Dumazet 已提交
2878 2879 2880 2881 2882
	if (tcp_md5sig_pool_populated) {
		/* coupled with smp_wmb() in __tcp_alloc_md5sig_pool() */
		smp_rmb();
		return this_cpu_ptr(&tcp_md5sig_pool);
	}
E
Eric Dumazet 已提交
2883 2884 2885 2886
	local_bh_enable();
	return NULL;
}
EXPORT_SYMBOL(tcp_get_md5sig_pool);
2887

2888
int tcp_md5_hash_header(struct tcp_md5sig_pool *hp,
2889
			const struct tcphdr *th)
2890 2891
{
	struct scatterlist sg;
2892
	struct tcphdr hdr;
2893 2894
	int err;

2895 2896 2897 2898
	/* We are not allowed to change tcphdr, make a local copy */
	memcpy(&hdr, th, sizeof(hdr));
	hdr.check = 0;

2899
	/* options aren't included in the hash */
2900 2901
	sg_init_one(&sg, &hdr, sizeof(hdr));
	err = crypto_hash_update(&hp->md5_desc, &sg, sizeof(hdr));
2902 2903 2904 2905 2906
	return err;
}
EXPORT_SYMBOL(tcp_md5_hash_header);

int tcp_md5_hash_skb_data(struct tcp_md5sig_pool *hp,
2907
			  const struct sk_buff *skb, unsigned int header_len)
2908 2909 2910 2911
{
	struct scatterlist sg;
	const struct tcphdr *tp = tcp_hdr(skb);
	struct hash_desc *desc = &hp->md5_desc;
2912 2913 2914
	unsigned int i;
	const unsigned int head_data_len = skb_headlen(skb) > header_len ?
					   skb_headlen(skb) - header_len : 0;
2915
	const struct skb_shared_info *shi = skb_shinfo(skb);
2916
	struct sk_buff *frag_iter;
2917 2918 2919 2920 2921 2922 2923 2924 2925

	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
2926 2927 2928 2929 2930
		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
2931
		if (crypto_hash_update(desc, &sg, skb_frag_size(f)))
2932 2933 2934
			return 1;
	}

2935 2936 2937 2938
	skb_walk_frags(skb, frag_iter)
		if (tcp_md5_hash_skb_data(hp, frag_iter, 0))
			return 1;

2939 2940 2941 2942
	return 0;
}
EXPORT_SYMBOL(tcp_md5_hash_skb_data);

2943
int tcp_md5_hash_key(struct tcp_md5sig_pool *hp, const struct tcp_md5sig_key *key)
2944 2945 2946 2947 2948 2949 2950 2951
{
	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);

2952 2953
#endif

A
Andi Kleen 已提交
2954 2955
void tcp_done(struct sock *sk)
{
2956 2957
	struct request_sock *req = tcp_sk(sk)->fastopen_rsk;

2958
	if (sk->sk_state == TCP_SYN_SENT || sk->sk_state == TCP_SYN_RECV)
2959
		TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_ATTEMPTFAILS);
A
Andi Kleen 已提交
2960 2961 2962

	tcp_set_state(sk, TCP_CLOSE);
	tcp_clear_xmit_timers(sk);
2963 2964
	if (req != NULL)
		reqsk_fastopen_remove(sk, req, false);
A
Andi Kleen 已提交
2965 2966 2967 2968 2969 2970 2971 2972 2973 2974

	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);

2975
extern struct tcp_congestion_ops tcp_reno;
L
Linus Torvalds 已提交
2976 2977 2978 2979

static __initdata unsigned long thash_entries;
static int __init set_thash_entries(char *str)
{
2980 2981
	ssize_t ret;

L
Linus Torvalds 已提交
2982 2983
	if (!str)
		return 0;
2984 2985 2986 2987 2988

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

L
Linus Torvalds 已提交
2989 2990 2991 2992
	return 1;
}
__setup("thash_entries=", set_thash_entries);

F
Fabian Frederick 已提交
2993
static void __init tcp_init_mem(void)
2994 2995 2996
{
	unsigned long limit = nr_free_buffer_pages() / 8;
	limit = max(limit, 128UL);
2997 2998 2999
	sysctl_tcp_mem[0] = limit / 4 * 3;
	sysctl_tcp_mem[1] = limit;
	sysctl_tcp_mem[2] = sysctl_tcp_mem[0] * 2;
3000 3001
}

L
Linus Torvalds 已提交
3002 3003
void __init tcp_init(void)
{
3004
	unsigned long limit;
3005
	int max_rshare, max_wshare, cnt;
3006
	unsigned int i;
L
Linus Torvalds 已提交
3007

3008
	sock_skb_cb_check_size(sizeof(struct tcp_skb_cb));
L
Linus Torvalds 已提交
3009

3010 3011
	percpu_counter_init(&tcp_sockets_allocated, 0, GFP_KERNEL);
	percpu_counter_init(&tcp_orphan_count, 0, GFP_KERNEL);
3012 3013 3014
	tcp_hashinfo.bind_bucket_cachep =
		kmem_cache_create("tcp_bind_bucket",
				  sizeof(struct inet_bind_bucket), 0,
3015
				  SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
L
Linus Torvalds 已提交
3016 3017 3018 3019 3020 3021

	/* Size and allocate the main established and bind bucket
	 * hash tables.
	 *
	 * The methodology is similar to that of the buffer cache.
	 */
3022
	tcp_hashinfo.ehash =
L
Linus Torvalds 已提交
3023
		alloc_large_system_hash("TCP established",
3024
					sizeof(struct inet_ehash_bucket),
L
Linus Torvalds 已提交
3025
					thash_entries,
E
Eric Dumazet 已提交
3026
					17, /* one slot per 128 KB of memory */
3027
					0,
L
Linus Torvalds 已提交
3028
					NULL,
3029
					&tcp_hashinfo.ehash_mask,
3030
					0,
3031
					thash_entries ? 0 : 512 * 1024);
E
Eric Dumazet 已提交
3032
	for (i = 0; i <= tcp_hashinfo.ehash_mask; i++)
3033
		INIT_HLIST_NULLS_HEAD(&tcp_hashinfo.ehash[i].chain, i);
E
Eric Dumazet 已提交
3034

3035 3036
	if (inet_ehash_locks_alloc(&tcp_hashinfo))
		panic("TCP: failed to alloc ehash_locks");
3037
	tcp_hashinfo.bhash =
L
Linus Torvalds 已提交
3038
		alloc_large_system_hash("TCP bind",
3039
					sizeof(struct inet_bind_hashbucket),
3040
					tcp_hashinfo.ehash_mask + 1,
E
Eric Dumazet 已提交
3041
					17, /* one slot per 128 KB of memory */
3042
					0,
3043
					&tcp_hashinfo.bhash_size,
L
Linus Torvalds 已提交
3044
					NULL,
3045
					0,
L
Linus Torvalds 已提交
3046
					64 * 1024);
3047
	tcp_hashinfo.bhash_size = 1U << tcp_hashinfo.bhash_size;
3048 3049 3050
	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
3051 3052
	}

E
Eric Dumazet 已提交
3053 3054 3055 3056 3057 3058

	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
3059

3060
	tcp_init_mem();
3061
	/* Set per-socket limits to no more than 1/128 the pressure threshold */
E
Eric Dumazet 已提交
3062
	limit = nr_free_buffer_pages() << (PAGE_SHIFT - 7);
3063 3064
	max_wshare = min(4UL*1024*1024, limit);
	max_rshare = min(6UL*1024*1024, limit);
3065

3066
	sysctl_tcp_wmem[0] = SK_MEM_QUANTUM;
3067
	sysctl_tcp_wmem[1] = 16*1024;
3068
	sysctl_tcp_wmem[2] = max(64*1024, max_wshare);
3069

3070
	sysctl_tcp_rmem[0] = SK_MEM_QUANTUM;
3071
	sysctl_tcp_rmem[1] = 87380;
3072
	sysctl_tcp_rmem[2] = max(87380, max_rshare);
L
Linus Torvalds 已提交
3073

3074
	pr_info("Hash tables configured (established %u bind %u)\n",
J
Joe Perches 已提交
3075
		tcp_hashinfo.ehash_mask + 1, tcp_hashinfo.bhash_size);
3076

3077
	tcp_metrics_init();
3078
	BUG_ON(tcp_register_congestion_control(&tcp_reno) != 0);
E
Eric Dumazet 已提交
3079
	tcp_tasklet_init();
L
Linus Torvalds 已提交
3080
}