tcp.c 86.8 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7
/*
 * INET		An implementation of the TCP/IP protocol suite for the LINUX
 *		operating system.  INET is implemented using the  BSD Socket
 *		interface as the means of communication with the user level.
 *
 *		Implementation of the Transmission Control Protocol(TCP).
 *
8
 * Authors:	Ross Biro
L
Linus Torvalds 已提交
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247
 *		Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *		Mark Evans, <evansmp@uhura.aston.ac.uk>
 *		Corey Minyard <wf-rch!minyard@relay.EU.net>
 *		Florian La Roche, <flla@stud.uni-sb.de>
 *		Charles Hedrick, <hedrick@klinzhai.rutgers.edu>
 *		Linus Torvalds, <torvalds@cs.helsinki.fi>
 *		Alan Cox, <gw4pts@gw4pts.ampr.org>
 *		Matthew Dillon, <dillon@apollo.west.oic.com>
 *		Arnt Gulbrandsen, <agulbra@nvg.unit.no>
 *		Jorge Cwik, <jorge@laser.satlink.net>
 *
 * Fixes:
 *		Alan Cox	:	Numerous verify_area() calls
 *		Alan Cox	:	Set the ACK bit on a reset
 *		Alan Cox	:	Stopped it crashing if it closed while
 *					sk->inuse=1 and was trying to connect
 *					(tcp_err()).
 *		Alan Cox	:	All icmp error handling was broken
 *					pointers passed where wrong and the
 *					socket was looked up backwards. Nobody
 *					tested any icmp error code obviously.
 *		Alan Cox	:	tcp_err() now handled properly. It
 *					wakes people on errors. poll
 *					behaves and the icmp error race
 *					has gone by moving it into sock.c
 *		Alan Cox	:	tcp_send_reset() fixed to work for
 *					everything not just packets for
 *					unknown sockets.
 *		Alan Cox	:	tcp option processing.
 *		Alan Cox	:	Reset tweaked (still not 100%) [Had
 *					syn rule wrong]
 *		Herp Rosmanith  :	More reset fixes
 *		Alan Cox	:	No longer acks invalid rst frames.
 *					Acking any kind of RST is right out.
 *		Alan Cox	:	Sets an ignore me flag on an rst
 *					receive otherwise odd bits of prattle
 *					escape still
 *		Alan Cox	:	Fixed another acking RST frame bug.
 *					Should stop LAN workplace lockups.
 *		Alan Cox	: 	Some tidyups using the new skb list
 *					facilities
 *		Alan Cox	:	sk->keepopen now seems to work
 *		Alan Cox	:	Pulls options out correctly on accepts
 *		Alan Cox	:	Fixed assorted sk->rqueue->next errors
 *		Alan Cox	:	PSH doesn't end a TCP read. Switched a
 *					bit to skb ops.
 *		Alan Cox	:	Tidied tcp_data to avoid a potential
 *					nasty.
 *		Alan Cox	:	Added some better commenting, as the
 *					tcp is hard to follow
 *		Alan Cox	:	Removed incorrect check for 20 * psh
 *	Michael O'Reilly	:	ack < copied bug fix.
 *	Johannes Stille		:	Misc tcp fixes (not all in yet).
 *		Alan Cox	:	FIN with no memory -> CRASH
 *		Alan Cox	:	Added socket option proto entries.
 *					Also added awareness of them to accept.
 *		Alan Cox	:	Added TCP options (SOL_TCP)
 *		Alan Cox	:	Switched wakeup calls to callbacks,
 *					so the kernel can layer network
 *					sockets.
 *		Alan Cox	:	Use ip_tos/ip_ttl settings.
 *		Alan Cox	:	Handle FIN (more) properly (we hope).
 *		Alan Cox	:	RST frames sent on unsynchronised
 *					state ack error.
 *		Alan Cox	:	Put in missing check for SYN bit.
 *		Alan Cox	:	Added tcp_select_window() aka NET2E
 *					window non shrink trick.
 *		Alan Cox	:	Added a couple of small NET2E timer
 *					fixes
 *		Charles Hedrick :	TCP fixes
 *		Toomas Tamm	:	TCP window fixes
 *		Alan Cox	:	Small URG fix to rlogin ^C ack fight
 *		Charles Hedrick	:	Rewrote most of it to actually work
 *		Linus		:	Rewrote tcp_read() and URG handling
 *					completely
 *		Gerhard Koerting:	Fixed some missing timer handling
 *		Matthew Dillon  :	Reworked TCP machine states as per RFC
 *		Gerhard Koerting:	PC/TCP workarounds
 *		Adam Caldwell	:	Assorted timer/timing errors
 *		Matthew Dillon	:	Fixed another RST bug
 *		Alan Cox	:	Move to kernel side addressing changes.
 *		Alan Cox	:	Beginning work on TCP fastpathing
 *					(not yet usable)
 *		Arnt Gulbrandsen:	Turbocharged tcp_check() routine.
 *		Alan Cox	:	TCP fast path debugging
 *		Alan Cox	:	Window clamping
 *		Michael Riepe	:	Bug in tcp_check()
 *		Matt Dillon	:	More TCP improvements and RST bug fixes
 *		Matt Dillon	:	Yet more small nasties remove from the
 *					TCP code (Be very nice to this man if
 *					tcp finally works 100%) 8)
 *		Alan Cox	:	BSD accept semantics.
 *		Alan Cox	:	Reset on closedown bug.
 *	Peter De Schrijver	:	ENOTCONN check missing in tcp_sendto().
 *		Michael Pall	:	Handle poll() after URG properly in
 *					all cases.
 *		Michael Pall	:	Undo the last fix in tcp_read_urg()
 *					(multi URG PUSH broke rlogin).
 *		Michael Pall	:	Fix the multi URG PUSH problem in
 *					tcp_readable(), poll() after URG
 *					works now.
 *		Michael Pall	:	recv(...,MSG_OOB) never blocks in the
 *					BSD api.
 *		Alan Cox	:	Changed the semantics of sk->socket to
 *					fix a race and a signal problem with
 *					accept() and async I/O.
 *		Alan Cox	:	Relaxed the rules on tcp_sendto().
 *		Yury Shevchuk	:	Really fixed accept() blocking problem.
 *		Craig I. Hagan  :	Allow for BSD compatible TIME_WAIT for
 *					clients/servers which listen in on
 *					fixed ports.
 *		Alan Cox	:	Cleaned the above up and shrank it to
 *					a sensible code size.
 *		Alan Cox	:	Self connect lockup fix.
 *		Alan Cox	:	No connect to multicast.
 *		Ross Biro	:	Close unaccepted children on master
 *					socket close.
 *		Alan Cox	:	Reset tracing code.
 *		Alan Cox	:	Spurious resets on shutdown.
 *		Alan Cox	:	Giant 15 minute/60 second timer error
 *		Alan Cox	:	Small whoops in polling before an
 *					accept.
 *		Alan Cox	:	Kept the state trace facility since
 *					it's handy for debugging.
 *		Alan Cox	:	More reset handler fixes.
 *		Alan Cox	:	Started rewriting the code based on
 *					the RFC's for other useful protocol
 *					references see: Comer, KA9Q NOS, and
 *					for a reference on the difference
 *					between specifications and how BSD
 *					works see the 4.4lite source.
 *		A.N.Kuznetsov	:	Don't time wait on completion of tidy
 *					close.
 *		Linus Torvalds	:	Fin/Shutdown & copied_seq changes.
 *		Linus Torvalds	:	Fixed BSD port reuse to work first syn
 *		Alan Cox	:	Reimplemented timers as per the RFC
 *					and using multiple timers for sanity.
 *		Alan Cox	:	Small bug fixes, and a lot of new
 *					comments.
 *		Alan Cox	:	Fixed dual reader crash by locking
 *					the buffers (much like datagram.c)
 *		Alan Cox	:	Fixed stuck sockets in probe. A probe
 *					now gets fed up of retrying without
 *					(even a no space) answer.
 *		Alan Cox	:	Extracted closing code better
 *		Alan Cox	:	Fixed the closing state machine to
 *					resemble the RFC.
 *		Alan Cox	:	More 'per spec' fixes.
 *		Jorge Cwik	:	Even faster checksumming.
 *		Alan Cox	:	tcp_data() doesn't ack illegal PSH
 *					only frames. At least one pc tcp stack
 *					generates them.
 *		Alan Cox	:	Cache last socket.
 *		Alan Cox	:	Per route irtt.
 *		Matt Day	:	poll()->select() match BSD precisely on error
 *		Alan Cox	:	New buffers
 *		Marc Tamsky	:	Various sk->prot->retransmits and
 *					sk->retransmits misupdating fixed.
 *					Fixed tcp_write_timeout: stuck close,
 *					and TCP syn retries gets used now.
 *		Mark Yarvis	:	In tcp_read_wakeup(), don't send an
 *					ack if state is TCP_CLOSED.
 *		Alan Cox	:	Look up device on a retransmit - routes may
 *					change. Doesn't yet cope with MSS shrink right
 *					but it's a start!
 *		Marc Tamsky	:	Closing in closing fixes.
 *		Mike Shaver	:	RFC1122 verifications.
 *		Alan Cox	:	rcv_saddr errors.
 *		Alan Cox	:	Block double connect().
 *		Alan Cox	:	Small hooks for enSKIP.
 *		Alexey Kuznetsov:	Path MTU discovery.
 *		Alan Cox	:	Support soft errors.
 *		Alan Cox	:	Fix MTU discovery pathological case
 *					when the remote claims no mtu!
 *		Marc Tamsky	:	TCP_CLOSE fix.
 *		Colin (G3TNE)	:	Send a reset on syn ack replies in
 *					window but wrong (fixes NT lpd problems)
 *		Pedro Roque	:	Better TCP window handling, delayed ack.
 *		Joerg Reuter	:	No modification of locked buffers in
 *					tcp_do_retransmit()
 *		Eric Schenk	:	Changed receiver side silly window
 *					avoidance algorithm to BSD style
 *					algorithm. This doubles throughput
 *					against machines running Solaris,
 *					and seems to result in general
 *					improvement.
 *	Stefan Magdalinski	:	adjusted tcp_readable() to fix FIONREAD
 *	Willy Konynenberg	:	Transparent proxying support.
 *	Mike McLagan		:	Routing by source
 *		Keith Owens	:	Do proper merging with partial SKB's in
 *					tcp_do_sendmsg to avoid burstiness.
 *		Eric Schenk	:	Fix fast close down bug with
 *					shutdown() followed by close().
 *		Andi Kleen 	:	Make poll agree with SIGIO
 *	Salvatore Sanfilippo	:	Support SO_LINGER with linger == 1 and
 *					lingertime == 0 (RFC 793 ABORT Call)
 *	Hirokazu Takahashi	:	Use copy_from_user() instead of
 *					csum_and_copy_from_user() if possible.
 *
 *		This program is free software; you can redistribute it and/or
 *		modify it under the terms of the GNU General Public License
 *		as published by the Free Software Foundation; either version
 *		2 of the License, or(at your option) any later version.
 *
 * Description of States:
 *
 *	TCP_SYN_SENT		sent a connection request, waiting for ack
 *
 *	TCP_SYN_RECV		received a connection request, sent ack,
 *				waiting for final ack in three-way handshake.
 *
 *	TCP_ESTABLISHED		connection established
 *
 *	TCP_FIN_WAIT1		our side has shutdown, waiting to complete
 *				transmission of remaining buffered data
 *
 *	TCP_FIN_WAIT2		all buffered data sent, waiting for remote
 *				to shutdown
 *
 *	TCP_CLOSING		both sides have shutdown but we still have
 *				data we have to finish sending
 *
 *	TCP_TIME_WAIT		timeout to catch resent junk before entering
 *				closed, can only be entered from FIN_WAIT2
 *				or CLOSING.  Required because the other end
 *				may not have gotten our last ACK causing it
 *				to retransmit the data packet (which we ignore)
 *
 *	TCP_CLOSE_WAIT		remote side has shutdown and is waiting for
 *				us to finish writing our data and to shutdown
 *				(we have to close() to move on to LAST_ACK)
 *
 *	TCP_LAST_ACK		out side has shutdown after remote has
 *				shutdown.  There may still be data in our
 *				buffer that we have to finish sending
 *
 *	TCP_CLOSE		socket is finished
 */

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

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

#include <net/icmp.h>
273
#include <net/inet_common.h>
L
Linus Torvalds 已提交
274 275 276
#include <net/tcp.h>
#include <net/xfrm.h>
#include <net/ip.h>
C
Chris Leech 已提交
277
#include <net/netdma.h>
J
Jens Axboe 已提交
278
#include <net/sock.h>
L
Linus Torvalds 已提交
279 280 281 282

#include <asm/uaccess.h>
#include <asm/ioctls.h>

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

285
struct percpu_counter tcp_orphan_count;
286 287
EXPORT_SYMBOL_GPL(tcp_orphan_count);

288 289
int sysctl_tcp_wmem[3] __read_mostly;
int sysctl_tcp_rmem[3] __read_mostly;
L
Linus Torvalds 已提交
290 291 292 293

EXPORT_SYMBOL(sysctl_tcp_rmem);
EXPORT_SYMBOL(sysctl_tcp_wmem);

E
Eric Dumazet 已提交
294
atomic_long_t tcp_memory_allocated;	/* Current allocated memory. */
L
Linus Torvalds 已提交
295
EXPORT_SYMBOL(tcp_memory_allocated);
296 297 298 299 300

/*
 * Current number of TCP sockets.
 */
struct percpu_counter tcp_sockets_allocated;
L
Linus Torvalds 已提交
301 302
EXPORT_SYMBOL(tcp_sockets_allocated);

J
Jens Axboe 已提交
303 304 305 306 307 308 309 310 311
/*
 * TCP splice context
 */
struct tcp_splice_state {
	struct pipe_inode_info *pipe;
	size_t len;
	unsigned int flags;
};

L
Linus Torvalds 已提交
312 313 314
/*
 * Pressure flag: try to collapse.
 * Technical note: it is used by multiple contexts non atomically.
315
 * All the __sk_mem_schedule() is of this nature: accounting
L
Linus Torvalds 已提交
316 317
 * is strict, actions are advisory and have some latency.
 */
318
int tcp_memory_pressure __read_mostly;
L
Linus Torvalds 已提交
319 320
EXPORT_SYMBOL(tcp_memory_pressure);

321
void tcp_enter_memory_pressure(struct sock *sk)
L
Linus Torvalds 已提交
322 323
{
	if (!tcp_memory_pressure) {
P
Pavel Emelyanov 已提交
324
		NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMEMORYPRESSURES);
L
Linus Torvalds 已提交
325 326 327 328 329
		tcp_memory_pressure = 1;
	}
}
EXPORT_SYMBOL(tcp_enter_memory_pressure);

330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366
/* 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;
}

367 368 369 370 371 372 373 374 375 376 377 378 379
/* Address-family independent initialization for a tcp_sock.
 *
 * NOTE: A lot of things set to zero explicitly by call to
 *       sk_alloc() so need not be done here.
 */
void tcp_init_sock(struct sock *sk)
{
	struct inet_connection_sock *icsk = inet_csk(sk);
	struct tcp_sock *tp = tcp_sk(sk);

	skb_queue_head_init(&tp->out_of_order_queue);
	tcp_init_xmit_timers(sk);
	tcp_prequeue_init(tp);
E
Eric Dumazet 已提交
380
	INIT_LIST_HEAD(&tp->tsq_node);
381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399

	icsk->icsk_rto = TCP_TIMEOUT_INIT;
	tp->mdev = TCP_TIMEOUT_INIT;

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

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

	tp->reordering = sysctl_tcp_reordering;
Y
Yuchung Cheng 已提交
400
	tcp_enable_early_retrans(tp);
401 402
	icsk->icsk_ca_ops = &tcp_init_congestion_ops;

403 404
	tp->tsoffset = 0;

405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425
	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;

	/* Presumed zeroed, in order of appearance:
	 *	cookie_in_always, cookie_out_never,
	 *	s_data_constant, s_data_in, s_data_out
	 */
	sk->sk_sndbuf = sysctl_tcp_wmem[1];
	sk->sk_rcvbuf = sysctl_tcp_rmem[1];

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

L
Linus Torvalds 已提交
426 427 428 429 430 431 432 433 434 435 436
/*
 *	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;
437
	const struct tcp_sock *tp = tcp_sk(sk);
L
Linus Torvalds 已提交
438

E
Eric Dumazet 已提交
439
	sock_poll_wait(file, sk_sleep(sk), wait);
L
Linus Torvalds 已提交
440
	if (sk->sk_state == TCP_LISTEN)
441
		return inet_csk_listen_poll(sk);
L
Linus Torvalds 已提交
442 443

	/* Socket is not locked. We are protected from async events
444 445
	 * by poll logic and correct handling of state changes
	 * made by other threads is impossible in any case.
L
Linus Torvalds 已提交
446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468
	 */

	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
469
	 * solve this dilemma. I would prefer, if POLLHUP were maskable,
L
Linus Torvalds 已提交
470 471
	 * then we could set it on SND_SHUTDOWN. BTW examples given
	 * in Stevens' books assume exactly this behaviour, it explains
472
	 * why POLLHUP is incompatible with POLLOUT.	--ANK
L
Linus Torvalds 已提交
473 474 475 476 477 478 479
	 *
	 * 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)
480
		mask |= POLLIN | POLLRDNORM | POLLRDHUP;
L
Linus Torvalds 已提交
481

482 483 484
	/* Connected or passive Fast Open socket? */
	if (sk->sk_state != TCP_SYN_SENT &&
	    (sk->sk_state != TCP_SYN_RECV || tp->fastopen_rsk != NULL)) {
485 486 487 488 489
		int target = sock_rcvlowat(sk, 0, INT_MAX);

		if (tp->urg_seq == tp->copied_seq &&
		    !sock_flag(sk, SOCK_URGINLINE) &&
		    tp->urg_data)
490
			target++;
491

L
Linus Torvalds 已提交
492 493 494
		/* Potential race condition. If read of tp below will
		 * escape above sk->sk_state, we can be illegally awaken
		 * in SYN_* states. */
495
		if (tp->rcv_nxt - tp->copied_seq >= target)
L
Linus Torvalds 已提交
496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512
			mask |= POLLIN | POLLRDNORM;

		if (!(sk->sk_shutdown & SEND_SHUTDOWN)) {
			if (sk_stream_wspace(sk) >= sk_stream_min_wspace(sk)) {
				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.
				 */
				if (sk_stream_wspace(sk) >= sk_stream_min_wspace(sk))
					mask |= POLLOUT | POLLWRNORM;
			}
513 514
		} else
			mask |= POLLOUT | POLLWRNORM;
L
Linus Torvalds 已提交
515 516 517 518

		if (tp->urg_data & TCP_URG_VALID)
			mask |= POLLPRI;
	}
T
Tom Marshall 已提交
519 520 521 522 523
	/* This barrier is coupled with smp_wmb() in tcp_reset() */
	smp_rmb();
	if (sk->sk_err)
		mask |= POLLERR;

L
Linus Torvalds 已提交
524 525
	return mask;
}
E
Eric Dumazet 已提交
526
EXPORT_SYMBOL(tcp_poll);
L
Linus Torvalds 已提交
527 528 529 530 531

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
532
	bool slow;
L
Linus Torvalds 已提交
533 534 535 536 537 538

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

E
Eric Dumazet 已提交
539
		slow = lock_sock_fast(sk);
L
Linus Torvalds 已提交
540 541 542 543 544 545
		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)) {
546

L
Linus Torvalds 已提交
547 548
			answ = tp->rcv_nxt - tp->copied_seq;

E
Eric Dumazet 已提交
549 550 551
			/* Subtract 1, if FIN was received */
			if (answ && sock_flag(sk, SOCK_DONE))
				answ--;
L
Linus Torvalds 已提交
552 553
		} else
			answ = tp->urg_seq - tp->copied_seq;
E
Eric Dumazet 已提交
554
		unlock_sock_fast(sk, slow);
L
Linus Torvalds 已提交
555 556 557 558 559 560 561 562 563 564 565 566 567
		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;
568 569 570 571 572 573 574 575 576
	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
577 578
	default:
		return -ENOIOCTLCMD;
579
	}
L
Linus Torvalds 已提交
580 581 582

	return put_user(answ, (int __user *)arg);
}
E
Eric Dumazet 已提交
583
EXPORT_SYMBOL(tcp_ioctl);
L
Linus Torvalds 已提交
584 585 586

static inline void tcp_mark_push(struct tcp_sock *tp, struct sk_buff *skb)
{
E
Eric Dumazet 已提交
587
	TCP_SKB_CB(skb)->tcp_flags |= TCPHDR_PSH;
L
Linus Torvalds 已提交
588 589 590
	tp->pushed_seq = tp->write_seq;
}

E
Eric Dumazet 已提交
591
static inline bool forced_push(const struct tcp_sock *tp)
L
Linus Torvalds 已提交
592 593 594 595
{
	return after(tp->write_seq, tp->pushed_seq + (tp->max_window >> 1));
}

596
static inline void skb_entail(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
597
{
598
	struct tcp_sock *tp = tcp_sk(sk);
599 600 601 602
	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
603
	tcb->tcp_flags = TCPHDR_ACK;
604
	tcb->sacked  = 0;
L
Linus Torvalds 已提交
605
	skb_header_release(skb);
606
	tcp_add_write_queue_tail(sk, skb);
607 608
	sk->sk_wmem_queued += skb->truesize;
	sk_mem_charge(sk, skb->truesize);
609
	if (tp->nonagle & TCP_NAGLE_PUSH)
610
		tp->nonagle &= ~TCP_NAGLE_PUSH;
L
Linus Torvalds 已提交
611 612
}

613
static inline void tcp_mark_urg(struct tcp_sock *tp, int flags)
L
Linus Torvalds 已提交
614
{
I
Ilpo Järvinen 已提交
615
	if (flags & MSG_OOB)
L
Linus Torvalds 已提交
616 617 618
		tp->snd_up = tp->write_seq;
}

619 620
static inline void tcp_push(struct sock *sk, int flags, int mss_now,
			    int nonagle)
L
Linus Torvalds 已提交
621
{
622
	if (tcp_send_head(sk)) {
623 624
		struct tcp_sock *tp = tcp_sk(sk);

L
Linus Torvalds 已提交
625
		if (!(flags & MSG_MORE) || forced_push(tp))
626 627 628
			tcp_mark_push(tp, tcp_write_queue_tail(sk));

		tcp_mark_urg(tp, flags);
629
		__tcp_push_pending_frames(sk, mss_now,
L
Linus Torvalds 已提交
630 631 632 633
					  (flags & MSG_MORE) ? TCP_NAGLE_CORK : nonagle);
	}
}

634 635
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
636 637
{
	struct tcp_splice_state *tss = rd_desc->arg.data;
638
	int ret;
J
Jens Axboe 已提交
639

640 641
	ret = skb_splice_bits(skb, offset, tss->pipe, min(rd_desc->count, len),
			      tss->flags);
642 643 644
	if (ret > 0)
		rd_desc->count -= ret;
	return ret;
J
Jens Axboe 已提交
645 646 647 648 649 650 651
}

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,
652
		.count	  = tss->len,
J
Jens Axboe 已提交
653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683
	};

	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;

684
	sock_rps_record_flow(sk);
J
Jens Axboe 已提交
685 686 687 688 689 690 691 692 693 694
	/*
	 * We can't seek on a socket input
	 */
	if (unlikely(*ppos))
		return -ESPIPE;

	ret = spliced = 0;

	lock_sock(sk);

695
	timeo = sock_rcvtimeo(sk, sock->file->f_flags & O_NONBLOCK);
J
Jens Axboe 已提交
696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733
	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;

734 735
		if (!timeo)
			break;
J
Jens Axboe 已提交
736 737 738 739
		release_sock(sk);
		lock_sock(sk);

		if (sk->sk_err || sk->sk_state == TCP_CLOSE ||
740
		    (sk->sk_shutdown & RCV_SHUTDOWN) ||
J
Jens Axboe 已提交
741 742 743 744 745 746 747 748 749 750 751
		    signal_pending(current))
			break;
	}

	release_sock(sk);

	if (spliced)
		return spliced;

	return ret;
}
E
Eric Dumazet 已提交
752
EXPORT_SYMBOL(tcp_splice_read);
J
Jens Axboe 已提交
753

754
struct sk_buff *sk_stream_alloc_skb(struct sock *sk, int size, gfp_t gfp)
755 756 757 758 759 760 761 762
{
	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) {
763
		if (sk_wmem_schedule(sk, skb->truesize)) {
764
			skb_reserve(skb, sk->sk_prot->max_header);
765 766 767 768
			/*
			 * Make sure that we have exactly size bytes
			 * available to the caller, no more, no less.
			 */
E
Eric Dumazet 已提交
769
			skb->reserved_tailroom = skb->end - skb->tail - size;
770 771 772 773
			return skb;
		}
		__kfree_skb(skb);
	} else {
774
		sk->sk_prot->enter_memory_pressure(sk);
775 776 777 778 779
		sk_stream_moderate_sndbuf(sk);
	}
	return NULL;
}

I
Ilpo Järvinen 已提交
780 781 782 783
static unsigned int tcp_xmit_size_goal(struct sock *sk, u32 mss_now,
				       int large_allowed)
{
	struct tcp_sock *tp = tcp_sk(sk);
784
	u32 xmit_size_goal, old_size_goal;
I
Ilpo Järvinen 已提交
785 786 787 788 789 790 791 792 793

	xmit_size_goal = mss_now;

	if (large_allowed && sk_can_gso(sk)) {
		xmit_size_goal = ((sk->sk_gso_max_size - 1) -
				  inet_csk(sk)->icsk_af_ops->net_header_len -
				  inet_csk(sk)->icsk_ext_hdr_len -
				  tp->tcp_header_len);

E
Eric Dumazet 已提交
794 795 796 797
		/* TSQ : try to have two TSO segments in flight */
		xmit_size_goal = min_t(u32, xmit_size_goal,
				       sysctl_tcp_limit_output_bytes >> 1);

I
Ilpo Järvinen 已提交
798
		xmit_size_goal = tcp_bound_to_half_wnd(tp, xmit_size_goal);
799 800 801 802 803 804 805 806

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

		if (likely(old_size_goal <= xmit_size_goal &&
			   old_size_goal + mss_now > xmit_size_goal)) {
			xmit_size_goal = old_size_goal;
		} else {
807 808 809
			tp->xmit_size_goal_segs =
				min_t(u16, xmit_size_goal / mss_now,
				      sk->sk_gso_max_segs);
810 811
			xmit_size_goal = tp->xmit_size_goal_segs * mss_now;
		}
I
Ilpo Järvinen 已提交
812 813
	}

814
	return max(xmit_size_goal, mss_now);
I
Ilpo Järvinen 已提交
815 816 817 818 819 820 821 822 823 824 825 826
}

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;
}

827 828
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
829 830
{
	struct tcp_sock *tp = tcp_sk(sk);
831
	int mss_now, size_goal;
L
Linus Torvalds 已提交
832 833 834 835
	int err;
	ssize_t copied;
	long timeo = sock_sndtimeo(sk, flags & MSG_DONTWAIT);

836 837 838 839 840 841
	/* 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
842 843
		if ((err = sk_stream_wait_connect(sk, &timeo)) != 0)
			goto out_err;
844
	}
L
Linus Torvalds 已提交
845 846 847

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

I
Ilpo Järvinen 已提交
848
	mss_now = tcp_send_mss(sk, &size_goal, flags);
L
Linus Torvalds 已提交
849 850 851 852
	copied = 0;

	err = -EPIPE;
	if (sk->sk_err || (sk->sk_shutdown & SEND_SHUTDOWN))
853
		goto out_err;
L
Linus Torvalds 已提交
854

855
	while (size > 0) {
856
		struct sk_buff *skb = tcp_write_queue_tail(sk);
857 858
		int copy, i;
		bool can_coalesce;
L
Linus Torvalds 已提交
859

860
		if (!tcp_send_head(sk) || (copy = size_goal - skb->len) <= 0) {
L
Linus Torvalds 已提交
861 862 863 864
new_segment:
			if (!sk_stream_memory_free(sk))
				goto wait_for_sndbuf;

865
			skb = sk_stream_alloc_skb(sk, 0, sk->sk_allocation);
L
Linus Torvalds 已提交
866 867 868
			if (!skb)
				goto wait_for_memory;

869
			skb_entail(sk, skb);
870
			copy = size_goal;
L
Linus Torvalds 已提交
871 872 873 874 875 876 877 878 879 880 881
		}

		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;
		}
882
		if (!sk_wmem_schedule(sk, copy))
L
Linus Torvalds 已提交
883
			goto wait_for_memory;
884

L
Linus Torvalds 已提交
885
		if (can_coalesce) {
E
Eric Dumazet 已提交
886
			skb_frag_size_add(&skb_shinfo(skb)->frags[i - 1], copy);
L
Linus Torvalds 已提交
887 888 889 890
		} else {
			get_page(page);
			skb_fill_page_desc(skb, i, page, offset, copy);
		}
891
		skb_shinfo(skb)->tx_flags |= SKBTX_SHARED_FRAG;
892

L
Linus Torvalds 已提交
893 894 895 896
		skb->len += copy;
		skb->data_len += copy;
		skb->truesize += copy;
		sk->sk_wmem_queued += copy;
897
		sk_mem_charge(sk, copy);
898
		skb->ip_summed = CHECKSUM_PARTIAL;
L
Linus Torvalds 已提交
899 900
		tp->write_seq += copy;
		TCP_SKB_CB(skb)->end_seq += copy;
901
		skb_shinfo(skb)->gso_segs = 0;
L
Linus Torvalds 已提交
902 903

		if (!copied)
E
Eric Dumazet 已提交
904
			TCP_SKB_CB(skb)->tcp_flags &= ~TCPHDR_PSH;
L
Linus Torvalds 已提交
905 906

		copied += copy;
907 908
		offset += copy;
		if (!(size -= copy))
L
Linus Torvalds 已提交
909 910
			goto out;

911
		if (skb->len < size_goal || (flags & MSG_OOB))
L
Linus Torvalds 已提交
912 913 914 915
			continue;

		if (forced_push(tp)) {
			tcp_mark_push(tp, skb);
916
			__tcp_push_pending_frames(sk, mss_now, TCP_NAGLE_PUSH);
917
		} else if (skb == tcp_send_head(sk))
L
Linus Torvalds 已提交
918 919 920 921 922 923
			tcp_push_one(sk, mss_now);
		continue;

wait_for_sndbuf:
		set_bit(SOCK_NOSPACE, &sk->sk_socket->flags);
wait_for_memory:
924
		tcp_push(sk, flags & ~MSG_MORE, mss_now, TCP_NAGLE_PUSH);
L
Linus Torvalds 已提交
925 926 927 928

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

I
Ilpo Järvinen 已提交
929
		mss_now = tcp_send_mss(sk, &size_goal, flags);
L
Linus Torvalds 已提交
930 931 932
	}

out:
933
	if (copied && !(flags & MSG_SENDPAGE_NOTLAST))
934
		tcp_push(sk, flags, mss_now, tp->nonagle);
L
Linus Torvalds 已提交
935 936 937 938 939 940 941 942 943
	return copied;

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

944 945
int tcp_sendpage(struct sock *sk, struct page *page, int offset,
		 size_t size, int flags)
L
Linus Torvalds 已提交
946 947 948 949
{
	ssize_t res;

	if (!(sk->sk_route_caps & NETIF_F_SG) ||
950
	    !(sk->sk_route_caps & NETIF_F_ALL_CSUM))
951 952
		return sock_no_sendpage(sk->sk_socket, page, offset, size,
					flags);
L
Linus Torvalds 已提交
953 954

	lock_sock(sk);
955
	res = do_tcp_sendpages(sk, page, offset, size, flags);
L
Linus Torvalds 已提交
956 957 958
	release_sock(sk);
	return res;
}
E
Eric Dumazet 已提交
959
EXPORT_SYMBOL(tcp_sendpage);
L
Linus Torvalds 已提交
960

961
static inline int select_size(const struct sock *sk, bool sg)
L
Linus Torvalds 已提交
962
{
963
	const struct tcp_sock *tp = tcp_sk(sk);
964
	int tmp = tp->mss_cache;
L
Linus Torvalds 已提交
965

K
Krishna Kumar 已提交
966
	if (sg) {
967 968 969 970 971 972
		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 {
973 974 975 976 977 978 979
			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
980 981 982 983

	return tmp;
}

984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015
void tcp_free_fastopen_req(struct tcp_sock *tp)
{
	if (tp->fastopen_req != NULL) {
		kfree(tp->fastopen_req);
		tp->fastopen_req = NULL;
	}
}

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

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

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

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

1016
int tcp_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
L
Linus Torvalds 已提交
1017 1018 1019 1020 1021
		size_t size)
{
	struct iovec *iov;
	struct tcp_sock *tp = tcp_sk(sk);
	struct sk_buff *skb;
1022 1023
	int iovlen, flags, err, copied = 0;
	int mss_now = 0, size_goal, copied_syn = 0, offset = 0;
1024
	bool sg;
L
Linus Torvalds 已提交
1025 1026 1027 1028 1029
	long timeo;

	lock_sock(sk);

	flags = msg->msg_flags;
1030 1031 1032 1033 1034 1035 1036 1037 1038
	if (flags & MSG_FASTOPEN) {
		err = tcp_sendmsg_fastopen(sk, msg, &copied_syn);
		if (err == -EINPROGRESS && copied_syn > 0)
			goto out;
		else if (err)
			goto out_err;
		offset = copied_syn;
	}

L
Linus Torvalds 已提交
1039 1040
	timeo = sock_sndtimeo(sk, flags & MSG_DONTWAIT);

1041 1042 1043 1044 1045 1046
	/* 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
1047
		if ((err = sk_stream_wait_connect(sk, &timeo)) != 0)
1048
			goto do_error;
1049
	}
L
Linus Torvalds 已提交
1050

P
Pavel Emelyanov 已提交
1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063
	if (unlikely(tp->repair)) {
		if (tp->repair_queue == TCP_RECV_QUEUE) {
			copied = tcp_send_rcvq(sk, msg, size);
			goto out;
		}

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

		/* 'common' sending to sendq */
	}

L
Linus Torvalds 已提交
1064 1065 1066
	/* This should be in poll */
	clear_bit(SOCK_ASYNC_NOSPACE, &sk->sk_socket->flags);

I
Ilpo Järvinen 已提交
1067
	mss_now = tcp_send_mss(sk, &size_goal, flags);
L
Linus Torvalds 已提交
1068 1069 1070 1071 1072 1073 1074 1075

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

	err = -EPIPE;
	if (sk->sk_err || (sk->sk_shutdown & SEND_SHUTDOWN))
1076
		goto out_err;
L
Linus Torvalds 已提交
1077

1078
	sg = !!(sk->sk_route_caps & NETIF_F_SG);
K
Krishna Kumar 已提交
1079

L
Linus Torvalds 已提交
1080
	while (--iovlen >= 0) {
1081
		size_t seglen = iov->iov_len;
L
Linus Torvalds 已提交
1082 1083 1084
		unsigned char __user *from = iov->iov_base;

		iov++;
1085 1086 1087 1088 1089 1090 1091 1092 1093
		if (unlikely(offset > 0)) {  /* Skip bytes copied in SYN */
			if (offset >= seglen) {
				offset -= seglen;
				continue;
			}
			seglen -= offset;
			from += offset;
			offset = 0;
		}
L
Linus Torvalds 已提交
1094 1095

		while (seglen > 0) {
1096 1097
			int copy = 0;
			int max = size_goal;
L
Linus Torvalds 已提交
1098

1099
			skb = tcp_write_queue_tail(sk);
1100 1101 1102 1103 1104
			if (tcp_send_head(sk)) {
				if (skb->ip_summed == CHECKSUM_NONE)
					max = mss_now;
				copy = max - skb->len;
			}
L
Linus Torvalds 已提交
1105

1106
			if (copy <= 0) {
L
Linus Torvalds 已提交
1107 1108 1109 1110 1111 1112 1113
new_segment:
				/* Allocate new segment. If the interface is SG,
				 * allocate skb fitting to single page.
				 */
				if (!sk_stream_memory_free(sk))
					goto wait_for_sndbuf;

K
Krishna Kumar 已提交
1114 1115 1116
				skb = sk_stream_alloc_skb(sk,
							  select_size(sk, sg),
							  sk->sk_allocation);
L
Linus Torvalds 已提交
1117 1118 1119 1120 1121 1122
				if (!skb)
					goto wait_for_memory;

				/*
				 * Check whether we can use HW checksum.
				 */
1123
				if (sk->sk_route_caps & NETIF_F_ALL_CSUM)
1124
					skb->ip_summed = CHECKSUM_PARTIAL;
L
Linus Torvalds 已提交
1125

1126
				skb_entail(sk, skb);
1127
				copy = size_goal;
1128
				max = size_goal;
L
Linus Torvalds 已提交
1129 1130 1131 1132 1133 1134 1135
			}

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

			/* Where to copy to? */
1136
			if (skb_availroom(skb) > 0) {
L
Linus Torvalds 已提交
1137
				/* We have some space in skb head. Superb! */
1138
				copy = min_t(int, copy, skb_availroom(skb));
1139 1140
				err = skb_add_data_nocache(sk, skb, from, copy);
				if (err)
L
Linus Torvalds 已提交
1141 1142
					goto do_fault;
			} else {
1143
				bool merge = true;
L
Linus Torvalds 已提交
1144
				int i = skb_shinfo(skb)->nr_frags;
1145 1146 1147 1148 1149 1150 1151 1152 1153 1154
				struct page_frag *pfrag = sk_page_frag(sk);

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

				if (!skb_can_coalesce(skb, i, pfrag->page,
						      pfrag->offset)) {
					if (i == MAX_SKB_FRAGS || !sg) {
						tcp_mark_push(tp, skb);
						goto new_segment;
L
Linus Torvalds 已提交
1155
					}
1156 1157
					merge = false;
				}
1158

1159
				copy = min_t(int, copy, pfrag->size - pfrag->offset);
1160

1161
				if (!sk_wmem_schedule(sk, copy))
1162
					goto wait_for_memory;
L
Linus Torvalds 已提交
1163

1164
				err = skb_copy_to_page_nocache(sk, from, skb,
1165 1166 1167 1168
							       pfrag->page,
							       pfrag->offset,
							       copy);
				if (err)
L
Linus Torvalds 已提交
1169 1170 1171 1172
					goto do_error;

				/* Update the skb. */
				if (merge) {
E
Eric Dumazet 已提交
1173
					skb_frag_size_add(&skb_shinfo(skb)->frags[i - 1], copy);
L
Linus Torvalds 已提交
1174
				} else {
1175 1176 1177
					skb_fill_page_desc(skb, i, pfrag->page,
							   pfrag->offset, copy);
					get_page(pfrag->page);
L
Linus Torvalds 已提交
1178
				}
1179
				pfrag->offset += copy;
L
Linus Torvalds 已提交
1180 1181 1182
			}

			if (!copied)
E
Eric Dumazet 已提交
1183
				TCP_SKB_CB(skb)->tcp_flags &= ~TCPHDR_PSH;
L
Linus Torvalds 已提交
1184 1185 1186

			tp->write_seq += copy;
			TCP_SKB_CB(skb)->end_seq += copy;
1187
			skb_shinfo(skb)->gso_segs = 0;
L
Linus Torvalds 已提交
1188 1189 1190 1191 1192 1193

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

P
Pavel Emelyanov 已提交
1194
			if (skb->len < max || (flags & MSG_OOB) || unlikely(tp->repair))
L
Linus Torvalds 已提交
1195 1196 1197 1198
				continue;

			if (forced_push(tp)) {
				tcp_mark_push(tp, skb);
1199
				__tcp_push_pending_frames(sk, mss_now, TCP_NAGLE_PUSH);
1200
			} else if (skb == tcp_send_head(sk))
L
Linus Torvalds 已提交
1201 1202 1203 1204 1205 1206
				tcp_push_one(sk, mss_now);
			continue;

wait_for_sndbuf:
			set_bit(SOCK_NOSPACE, &sk->sk_socket->flags);
wait_for_memory:
1207
			if (copied)
1208
				tcp_push(sk, flags & ~MSG_MORE, mss_now, TCP_NAGLE_PUSH);
L
Linus Torvalds 已提交
1209 1210 1211 1212

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

I
Ilpo Järvinen 已提交
1213
			mss_now = tcp_send_mss(sk, &size_goal, flags);
L
Linus Torvalds 已提交
1214 1215 1216 1217
		}
	}

out:
1218
	if (copied)
1219
		tcp_push(sk, flags, mss_now, tp->nonagle);
L
Linus Torvalds 已提交
1220
	release_sock(sk);
1221
	return copied + copied_syn;
L
Linus Torvalds 已提交
1222 1223 1224

do_fault:
	if (!skb->len) {
1225 1226 1227 1228 1229
		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);
1230
		sk_wmem_free_skb(sk, skb);
L
Linus Torvalds 已提交
1231 1232 1233
	}

do_error:
1234
	if (copied + copied_syn)
L
Linus Torvalds 已提交
1235 1236 1237 1238 1239 1240
		goto out;
out_err:
	err = sk_stream_error(sk, flags, err);
	release_sock(sk);
	return err;
}
E
Eric Dumazet 已提交
1241
EXPORT_SYMBOL(tcp_sendmsg);
L
Linus Torvalds 已提交
1242 1243 1244 1245 1246 1247

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

1248
static int tcp_recv_urg(struct sock *sk, struct msghdr *msg, int len, int flags)
L
Linus Torvalds 已提交
1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291
{
	struct tcp_sock *tp = tcp_sk(sk);

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

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

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

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

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

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

		return err ? -EFAULT : len;
	}

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

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

P
Pavel Emelyanov 已提交
1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309
static int tcp_peek_sndq(struct sock *sk, struct msghdr *msg, int len)
{
	struct sk_buff *skb;
	int copied = 0, err = 0;

	/* XXX -- need to support SO_PEEK_OFF */

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

		copied += skb->len;
	}

	return err ?: copied;
}

L
Linus Torvalds 已提交
1310 1311 1312 1313 1314 1315
/* 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.
 */
1316
void tcp_cleanup_rbuf(struct sock *sk, int copied)
L
Linus Torvalds 已提交
1317 1318
{
	struct tcp_sock *tp = tcp_sk(sk);
E
Eric Dumazet 已提交
1319
	bool time_to_ack = false;
L
Linus Torvalds 已提交
1320 1321 1322

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

1323
	WARN(skb && !before(tp->copied_seq, TCP_SKB_CB(skb)->end_seq),
J
Joe Perches 已提交
1324
	     "cleanup rbuf bug: copied %X seq %X rcvnxt %X\n",
1325
	     tp->copied_seq, TCP_SKB_CB(skb)->end_seq, tp->rcv_nxt);
L
Linus Torvalds 已提交
1326

1327 1328
	if (inet_csk_ack_scheduled(sk)) {
		const struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
1329 1330
		   /* Delayed ACKs frequently hit locked sockets during bulk
		    * receive. */
1331
		if (icsk->icsk_ack.blocked ||
L
Linus Torvalds 已提交
1332
		    /* Once-per-two-segments ACK was not sent by tcp_input.c */
1333
		    tp->rcv_nxt - tp->rcv_wup > icsk->icsk_ack.rcv_mss ||
L
Linus Torvalds 已提交
1334 1335 1336 1337 1338 1339
		    /*
		     * 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.
		     */
1340 1341 1342 1343 1344
		    (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
1345
			time_to_ack = true;
L
Linus Torvalds 已提交
1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366
	}

	/* 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
1367
				time_to_ack = true;
L
Linus Torvalds 已提交
1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378
		}
	}
	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);

1379
	NET_INC_STATS_USER(sock_net(sk), LINUX_MIB_TCPPREQUEUED);
L
Linus Torvalds 已提交
1380 1381 1382 1383 1384

	/* 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
1385
		sk_backlog_rcv(sk, skb);
L
Linus Torvalds 已提交
1386 1387 1388 1389 1390 1391
	local_bh_enable();

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

1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402
#ifdef CONFIG_NET_DMA
static void tcp_service_net_dma(struct sock *sk, bool wait)
{
	dma_cookie_t done, used;
	dma_cookie_t last_issued;
	struct tcp_sock *tp = tcp_sk(sk);

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

	last_issued = tp->ucopy.dma_cookie;
1403
	dma_async_issue_pending(tp->ucopy.dma_chan);
1404 1405

	do {
1406
		if (dma_async_is_tx_complete(tp->ucopy.dma_chan,
1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424
					      last_issued, &done,
					      &used) == DMA_SUCCESS) {
			/* Safe to free early-copied skbs now */
			__skb_queue_purge(&sk->sk_async_wait_queue);
			break;
		} else {
			struct sk_buff *skb;
			while ((skb = skb_peek(&sk->sk_async_wait_queue)) &&
			       (dma_async_is_complete(skb->dma_cookie, done,
						      used) == DMA_SUCCESS)) {
				__skb_dequeue(&sk->sk_async_wait_queue);
				kfree_skb(skb);
			}
		}
	} while (wait);
}
#endif

1425
static struct sk_buff *tcp_recv_skb(struct sock *sk, u32 seq, u32 *off)
L
Linus Torvalds 已提交
1426 1427 1428 1429
{
	struct sk_buff *skb;
	u32 offset;

1430
	while ((skb = skb_peek(&sk->sk_receive_queue)) != NULL) {
L
Linus Torvalds 已提交
1431
		offset = seq - TCP_SKB_CB(skb)->seq;
1432
		if (tcp_hdr(skb)->syn)
L
Linus Torvalds 已提交
1433
			offset--;
1434
		if (offset < skb->len || tcp_hdr(skb)->fin) {
L
Linus Torvalds 已提交
1435 1436 1437
			*off = offset;
			return skb;
		}
1438 1439 1440 1441 1442
		/* This looks weird, but this can happen if TCP collapsing
		 * splitted a fat GRO packet, while we released socket lock
		 * in skb_splice_bits()
		 */
		sk_eat_skb(sk, skb, false);
L
Linus Torvalds 已提交
1443 1444 1445 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
	}
	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) {
1471 1472
			int used;
			size_t len;
L
Linus Torvalds 已提交
1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483

			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);
1484
			if (used <= 0) {
1485 1486 1487 1488
				if (!copied)
					copied = used;
				break;
			} else if (used <= len) {
L
Linus Torvalds 已提交
1489 1490 1491 1492
				seq += used;
				copied += used;
				offset += used;
			}
1493
			/* If recv_actor drops the lock (e.g. TCP splice
1494 1495 1496 1497
			 * receive) the skb pointer might be invalid when
			 * getting here: tcp_collapse might have deleted it
			 * while aggregating skbs from the socket queue.
			 */
1498 1499
			skb = tcp_recv_skb(sk, seq - 1, &offset);
			if (!skb)
L
Linus Torvalds 已提交
1500
				break;
1501 1502 1503 1504 1505
			/* 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
1506
		}
1507
		if (tcp_hdr(skb)->fin) {
E
Eric Dumazet 已提交
1508
			sk_eat_skb(sk, skb, false);
L
Linus Torvalds 已提交
1509 1510 1511
			++seq;
			break;
		}
E
Eric Dumazet 已提交
1512
		sk_eat_skb(sk, skb, false);
L
Linus Torvalds 已提交
1513 1514
		if (!desc->count)
			break;
1515
		tp->copied_seq = seq;
L
Linus Torvalds 已提交
1516 1517 1518 1519 1520 1521
	}
	tp->copied_seq = seq;

	tcp_rcv_space_adjust(sk);

	/* Clean up data we have read: This will do ACK frames. */
1522 1523
	if (copied > 0) {
		tcp_recv_skb(sk, seq, &offset);
1524
		tcp_cleanup_rbuf(sk, copied);
1525
	}
L
Linus Torvalds 已提交
1526 1527
	return copied;
}
E
Eric Dumazet 已提交
1528
EXPORT_SYMBOL(tcp_read_sock);
L
Linus Torvalds 已提交
1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549

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

int tcp_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
		size_t len, int nonblock, int flags, int *addr_len)
{
	struct tcp_sock *tp = tcp_sk(sk);
	int copied = 0;
	u32 peek_seq;
	u32 *seq;
	unsigned long used;
	int err;
	int target;		/* Read at least this many bytes */
	long timeo;
	struct task_struct *user_recv = NULL;
E
Eric Dumazet 已提交
1550
	bool copied_early = false;
1551
	struct sk_buff *skb;
I
Ilpo Järvinen 已提交
1552
	u32 urg_hole = 0;
L
Linus Torvalds 已提交
1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565

	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
1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580
	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
1581 1582 1583 1584 1585 1586 1587 1588
	seq = &tp->copied_seq;
	if (flags & MSG_PEEK) {
		peek_seq = tp->copied_seq;
		seq = &peek_seq;
	}

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

C
Chris Leech 已提交
1589 1590 1591
#ifdef CONFIG_NET_DMA
	tp->ucopy.dma_chan = NULL;
	preempt_disable();
1592
	skb = skb_peek_tail(&sk->sk_receive_queue);
A
Andrew Morton 已提交
1593 1594 1595 1596 1597 1598 1599 1600
	{
		int available = 0;

		if (skb)
			available = TCP_SKB_CB(skb)->seq + skb->len - (*seq);
		if ((available < target) &&
		    (len > sysctl_tcp_dma_copybreak) && !(flags & MSG_PEEK) &&
		    !sysctl_tcp_low_latency &&
1601
		    net_dma_find_channel()) {
A
Andrew Morton 已提交
1602 1603 1604 1605 1606 1607 1608
			preempt_enable_no_resched();
			tp->ucopy.pinned_list =
					dma_pin_iovec_pages(msg->msg_iov, len);
		} else {
			preempt_enable_no_resched();
		}
	}
C
Chris Leech 已提交
1609 1610
#endif

L
Linus Torvalds 已提交
1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625
	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. */

1626
		skb_queue_walk(&sk->sk_receive_queue, skb) {
L
Linus Torvalds 已提交
1627 1628 1629
			/* Now that we have two receive queues this
			 * shouldn't happen.
			 */
1630
			if (WARN(before(*seq, TCP_SKB_CB(skb)->seq),
J
Joe Perches 已提交
1631 1632 1633
				 "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
1634
				break;
1635

L
Linus Torvalds 已提交
1636
			offset = *seq - TCP_SKB_CB(skb)->seq;
1637
			if (tcp_hdr(skb)->syn)
L
Linus Torvalds 已提交
1638 1639 1640
				offset--;
			if (offset < skb->len)
				goto found_ok_skb;
1641
			if (tcp_hdr(skb)->fin)
L
Linus Torvalds 已提交
1642
				goto found_fin_ok;
J
Joe Perches 已提交
1643 1644 1645
			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);
1646
		}
L
Linus Torvalds 已提交
1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657

		/* 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 ||
1658
			    signal_pending(current))
L
Linus Torvalds 已提交
1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693
				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;
			}
		}

1694
		tcp_cleanup_rbuf(sk, copied);
L
Linus Torvalds 已提交
1695

1696
		if (!sysctl_tcp_low_latency && tp->ucopy.task == user_recv) {
L
Linus Torvalds 已提交
1697 1698 1699 1700 1701 1702 1703 1704 1705
			/* Install new reader */
			if (!user_recv && !(flags & (MSG_TRUNC | MSG_PEEK))) {
				user_recv = current;
				tp->ucopy.task = user_recv;
				tp->ucopy.iov = msg->msg_iov;
			}

			tp->ucopy.len = len;

1706 1707
			WARN_ON(tp->copied_seq != tp->rcv_nxt &&
				!(flags & (MSG_PEEK | MSG_TRUNC)));
L
Linus Torvalds 已提交
1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734

			/* 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.
			 */
1735
			if (!skb_queue_empty(&tp->ucopy.prequeue))
L
Linus Torvalds 已提交
1736 1737 1738 1739 1740
				goto do_prequeue;

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

1741
#ifdef CONFIG_NET_DMA
1742 1743 1744 1745 1746 1747
		if (tp->ucopy.dma_chan) {
			if (tp->rcv_wnd == 0 &&
			    !skb_queue_empty(&sk->sk_async_wait_queue)) {
				tcp_service_net_dma(sk, true);
				tcp_cleanup_rbuf(sk, copied);
			} else
1748
				dma_async_issue_pending(tp->ucopy.dma_chan);
1749
		}
1750
#endif
L
Linus Torvalds 已提交
1751 1752 1753 1754 1755 1756 1757
		if (copied >= target) {
			/* Do not sleep, just process backlog. */
			release_sock(sk);
			lock_sock(sk);
		} else
			sk_wait_data(sk, &timeo);

C
Chris Leech 已提交
1758
#ifdef CONFIG_NET_DMA
1759
		tcp_service_net_dma(sk, false);  /* Don't block */
C
Chris Leech 已提交
1760 1761 1762
		tp->ucopy.wakeup = 0;
#endif

L
Linus Torvalds 已提交
1763 1764 1765 1766 1767 1768
		if (user_recv) {
			int chunk;

			/* __ Restore normal policy in scheduler __ */

			if ((chunk = len - tp->ucopy.len) != 0) {
1769
				NET_ADD_STATS_USER(sock_net(sk), LINUX_MIB_TCPDIRECTCOPYFROMBACKLOG, chunk);
L
Linus Torvalds 已提交
1770 1771 1772 1773 1774
				len -= chunk;
				copied += chunk;
			}

			if (tp->rcv_nxt == tp->copied_seq &&
1775
			    !skb_queue_empty(&tp->ucopy.prequeue)) {
L
Linus Torvalds 已提交
1776 1777 1778 1779
do_prequeue:
				tcp_prequeue_process(sk);

				if ((chunk = len - tp->ucopy.len) != 0) {
1780
					NET_ADD_STATS_USER(sock_net(sk), LINUX_MIB_TCPDIRECTCOPYFROMPREQUEUE, chunk);
L
Linus Torvalds 已提交
1781 1782 1783 1784 1785
					len -= chunk;
					copied += chunk;
				}
			}
		}
I
Ilpo Järvinen 已提交
1786 1787
		if ((flags & MSG_PEEK) &&
		    (peek_seq - copied - urg_hole != tp->copied_seq)) {
1788 1789 1790
			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
1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807
			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
1808
						urg_hole++;
L
Linus Torvalds 已提交
1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819
						offset++;
						used--;
						if (!used)
							goto skip_copy;
					}
				} else
					used = urg_offset;
			}
		}

		if (!(flags & MSG_TRUNC)) {
C
Chris Leech 已提交
1820 1821
#ifdef CONFIG_NET_DMA
			if (!tp->ucopy.dma_chan && tp->ucopy.pinned_list)
1822
				tp->ucopy.dma_chan = net_dma_find_channel();
C
Chris Leech 已提交
1823 1824 1825 1826 1827 1828 1829 1830 1831

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

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

1832 1833
					pr_alert("%s: dma_cookie < 0\n",
						 __func__);
C
Chris Leech 已提交
1834 1835 1836 1837 1838 1839

					/* Exception. Bailout! */
					if (!copied)
						copied = -EFAULT;
					break;
				}
1840

1841
				dma_async_issue_pending(tp->ucopy.dma_chan);
1842

C
Chris Leech 已提交
1843
				if ((offset + used) == skb->len)
E
Eric Dumazet 已提交
1844
					copied_early = true;
C
Chris Leech 已提交
1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856

			} else
#endif
			{
				err = skb_copy_datagram_iovec(skb, offset,
						msg->msg_iov, used);
				if (err) {
					/* Exception. Bailout! */
					if (!copied)
						copied = -EFAULT;
					break;
				}
L
Linus Torvalds 已提交
1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868
			}
		}

		*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;
1869
			tcp_fast_path_check(sk);
L
Linus Torvalds 已提交
1870 1871 1872 1873
		}
		if (used + offset < skb->len)
			continue;

1874
		if (tcp_hdr(skb)->fin)
L
Linus Torvalds 已提交
1875
			goto found_fin_ok;
C
Chris Leech 已提交
1876 1877
		if (!(flags & MSG_PEEK)) {
			sk_eat_skb(sk, skb, copied_early);
E
Eric Dumazet 已提交
1878
			copied_early = false;
C
Chris Leech 已提交
1879
		}
L
Linus Torvalds 已提交
1880 1881 1882 1883 1884
		continue;

	found_fin_ok:
		/* Process the FIN. */
		++*seq;
C
Chris Leech 已提交
1885 1886
		if (!(flags & MSG_PEEK)) {
			sk_eat_skb(sk, skb, copied_early);
E
Eric Dumazet 已提交
1887
			copied_early = false;
C
Chris Leech 已提交
1888
		}
L
Linus Torvalds 已提交
1889 1890 1891 1892
		break;
	} while (len > 0);

	if (user_recv) {
1893
		if (!skb_queue_empty(&tp->ucopy.prequeue)) {
L
Linus Torvalds 已提交
1894 1895 1896 1897 1898 1899 1900
			int chunk;

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

			tcp_prequeue_process(sk);

			if (copied > 0 && (chunk = len - tp->ucopy.len) != 0) {
1901
				NET_ADD_STATS_USER(sock_net(sk), LINUX_MIB_TCPDIRECTCOPYFROMPREQUEUE, chunk);
L
Linus Torvalds 已提交
1902 1903 1904 1905 1906 1907 1908 1909 1910
				len -= chunk;
				copied += chunk;
			}
		}

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

C
Chris Leech 已提交
1911
#ifdef CONFIG_NET_DMA
1912 1913
	tcp_service_net_dma(sk, true);  /* Wait for queue to drain */
	tp->ucopy.dma_chan = NULL;
C
Chris Leech 已提交
1914 1915 1916 1917 1918 1919 1920

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

L
Linus Torvalds 已提交
1921 1922 1923 1924 1925
	/* 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. */
1926
	tcp_cleanup_rbuf(sk, copied);
L
Linus Torvalds 已提交
1927 1928 1929 1930 1931 1932 1933 1934 1935

	release_sock(sk);
	return copied;

out:
	release_sock(sk);
	return err;

recv_urg:
1936
	err = tcp_recv_urg(sk, msg, len, flags);
L
Linus Torvalds 已提交
1937
	goto out;
P
Pavel Emelyanov 已提交
1938 1939 1940 1941

recv_sndq:
	err = tcp_peek_sndq(sk, msg, len);
	goto out;
L
Linus Torvalds 已提交
1942
}
E
Eric Dumazet 已提交
1943
EXPORT_SYMBOL(tcp_recvmsg);
L
Linus Torvalds 已提交
1944

I
Ilpo Järvinen 已提交
1945 1946 1947 1948 1949 1950 1951
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
1952
			TCP_INC_STATS(sock_net(sk), TCP_MIB_CURRESTAB);
I
Ilpo Järvinen 已提交
1953 1954 1955 1956
		break;

	case TCP_CLOSE:
		if (oldstate == TCP_CLOSE_WAIT || oldstate == TCP_ESTABLISHED)
P
Pavel Emelyanov 已提交
1957
			TCP_INC_STATS(sock_net(sk), TCP_MIB_ESTABRESETS);
I
Ilpo Järvinen 已提交
1958 1959 1960 1961

		sk->sk_prot->unhash(sk);
		if (inet_csk(sk)->icsk_bind_hash &&
		    !(sk->sk_userlocks & SOCK_BINDPORT_LOCK))
1962
			inet_put_port(sk);
I
Ilpo Järvinen 已提交
1963 1964
		/* fall through */
	default:
1965
		if (oldstate == TCP_ESTABLISHED)
P
Pavel Emelyanov 已提交
1966
			TCP_DEC_STATS(sock_net(sk), TCP_MIB_CURRESTAB);
I
Ilpo Järvinen 已提交
1967 1968 1969 1970 1971 1972 1973 1974
	}

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

#ifdef STATE_TRACE
1975
	SOCK_DEBUG(sk, "TCP sk=%p, State %s -> %s\n", sk, statename[oldstate], statename[state]);
I
Ilpo Järvinen 已提交
1976 1977 1978 1979
#endif
}
EXPORT_SYMBOL_GPL(tcp_set_state);

L
Linus Torvalds 已提交
1980 1981 1982 1983 1984 1985 1986
/*
 *	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.
 */

1987
static const unsigned char new_state[16] = {
L
Linus Torvalds 已提交
1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
  /* current state:        new state:      action:	*/
  /* (Invalid)		*/ TCP_CLOSE,
  /* TCP_ESTABLISHED	*/ TCP_FIN_WAIT1 | TCP_ACTION_FIN,
  /* TCP_SYN_SENT	*/ TCP_CLOSE,
  /* TCP_SYN_RECV	*/ TCP_FIN_WAIT1 | TCP_ACTION_FIN,
  /* TCP_FIN_WAIT1	*/ TCP_FIN_WAIT1,
  /* TCP_FIN_WAIT2	*/ TCP_FIN_WAIT2,
  /* TCP_TIME_WAIT	*/ TCP_CLOSE,
  /* TCP_CLOSE		*/ TCP_CLOSE,
  /* TCP_CLOSE_WAIT	*/ TCP_LAST_ACK  | TCP_ACTION_FIN,
  /* TCP_LAST_ACK	*/ TCP_LAST_ACK,
  /* TCP_LISTEN		*/ TCP_CLOSE,
  /* TCP_CLOSING	*/ TCP_CLOSING,
};

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

	tcp_set_state(sk, ns);

	return next & TCP_ACTION_FIN;
}

/*
 *	Shutdown the sending side of a connection. Much like close except
2015
 *	that we don't receive shut down or sock_set_flag(sk, SOCK_DEAD).
L
Linus Torvalds 已提交
2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035
 */

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
2036
EXPORT_SYMBOL(tcp_shutdown);
L
Linus Torvalds 已提交
2037

A
Arun Sharma 已提交
2038 2039 2040 2041 2042 2043 2044
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);

2045 2046 2047 2048
	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
2049 2050 2051
	return too_many_orphans || out_of_socket_memory;
}

L
Linus Torvalds 已提交
2052 2053 2054 2055
void tcp_close(struct sock *sk, long timeout)
{
	struct sk_buff *skb;
	int data_was_unread = 0;
H
Herbert Xu 已提交
2056
	int state;
L
Linus Torvalds 已提交
2057 2058 2059 2060 2061 2062 2063 2064

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

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

		/* Special case. */
2065
		inet_csk_listen_stop(sk);
L
Linus Torvalds 已提交
2066 2067 2068 2069 2070 2071 2072 2073 2074 2075

		goto adjudge_to_death;
	}

	/*  We need to flush the recv. buffs.  We do this only on the
	 *  descriptor close, not protocol-sourced closes, because the
	 *  reader process may not have drained the data yet!
	 */
	while ((skb = __skb_dequeue(&sk->sk_receive_queue)) != NULL) {
		u32 len = TCP_SKB_CB(skb)->end_seq - TCP_SKB_CB(skb)->seq -
2076
			  tcp_hdr(skb)->fin;
L
Linus Torvalds 已提交
2077 2078 2079 2080
		data_was_unread += len;
		__kfree_skb(skb);
	}

2081
	sk_mem_reclaim(sk);
L
Linus Torvalds 已提交
2082

2083 2084 2085 2086
	/* If socket has been already reset (e.g. in tcp_reset()) - kill it. */
	if (sk->sk_state == TCP_CLOSE)
		goto adjudge_to_death;

2087 2088 2089 2090 2091 2092
	/* 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
2093
	 */
P
Pavel Emelyanov 已提交
2094 2095 2096
	if (unlikely(tcp_sk(sk)->repair)) {
		sk->sk_prot->disconnect(sk, 0);
	} else if (data_was_unread) {
L
Linus Torvalds 已提交
2097
		/* Unread data was tossed, zap the connection. */
2098
		NET_INC_STATS_USER(sock_net(sk), LINUX_MIB_TCPABORTONCLOSE);
L
Linus Torvalds 已提交
2099
		tcp_set_state(sk, TCP_CLOSE);
2100
		tcp_send_active_reset(sk, sk->sk_allocation);
L
Linus Torvalds 已提交
2101 2102 2103
	} else if (sock_flag(sk, SOCK_LINGER) && !sk->sk_lingertime) {
		/* Check zero linger _after_ checking for unread data. */
		sk->sk_prot->disconnect(sk, 0);
2104
		NET_INC_STATS_USER(sock_net(sk), LINUX_MIB_TCPABORTONDATA);
L
Linus Torvalds 已提交
2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129
	} 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
2130 2131 2132 2133
		 * 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
2134 2135 2136 2137 2138 2139 2140
		 */
		tcp_send_fin(sk);
	}

	sk_stream_wait_close(sk, timeout);

adjudge_to_death:
H
Herbert Xu 已提交
2141 2142 2143 2144
	state = sk->sk_state;
	sock_hold(sk);
	sock_orphan(sk);

L
Linus Torvalds 已提交
2145 2146 2147 2148 2149 2150 2151 2152 2153
	/* 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);
2154
	WARN_ON(sock_owned_by_user(sk));
L
Linus Torvalds 已提交
2155

H
Herbert Xu 已提交
2156 2157
	percpu_counter_inc(sk->sk_prot->orphan_count);

H
Herbert Xu 已提交
2158 2159 2160
	/* 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
2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180

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

	if (sk->sk_state == TCP_FIN_WAIT2) {
		struct tcp_sock *tp = tcp_sk(sk);
		if (tp->linger2 < 0) {
			tcp_set_state(sk, TCP_CLOSE);
			tcp_send_active_reset(sk, GFP_ATOMIC);
2181 2182
			NET_INC_STATS_BH(sock_net(sk),
					LINUX_MIB_TCPABORTONLINGER);
L
Linus Torvalds 已提交
2183
		} else {
2184
			const int tmo = tcp_fin_time(sk);
L
Linus Torvalds 已提交
2185 2186

			if (tmo > TCP_TIMEWAIT_LEN) {
2187 2188
				inet_csk_reset_keepalive_timer(sk,
						tmo - TCP_TIMEWAIT_LEN);
L
Linus Torvalds 已提交
2189 2190 2191 2192 2193 2194 2195
			} else {
				tcp_time_wait(sk, TCP_FIN_WAIT2, tmo);
				goto out;
			}
		}
	}
	if (sk->sk_state != TCP_CLOSE) {
2196
		sk_mem_reclaim(sk);
A
Arun Sharma 已提交
2197
		if (tcp_check_oom(sk, 0)) {
L
Linus Torvalds 已提交
2198 2199
			tcp_set_state(sk, TCP_CLOSE);
			tcp_send_active_reset(sk, GFP_ATOMIC);
2200 2201
			NET_INC_STATS_BH(sock_net(sk),
					LINUX_MIB_TCPABORTONMEMORY);
L
Linus Torvalds 已提交
2202 2203 2204
		}
	}

2205 2206 2207 2208 2209 2210 2211 2212
	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);
2213
		inet_csk_destroy_sock(sk);
2214
	}
L
Linus Torvalds 已提交
2215 2216 2217 2218 2219 2220 2221
	/* Otherwise, socket is reprieved until protocol close. */

out:
	bh_unlock_sock(sk);
	local_bh_enable();
	sock_put(sk);
}
E
Eric Dumazet 已提交
2222
EXPORT_SYMBOL(tcp_close);
L
Linus Torvalds 已提交
2223 2224 2225

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

E
Eric Dumazet 已提交
2226
static inline bool tcp_need_reset(int state)
L
Linus Torvalds 已提交
2227 2228 2229 2230 2231 2232 2233 2234 2235
{
	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);
2236
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
2237 2238 2239 2240 2241 2242 2243 2244 2245
	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) {
2246
		inet_csk_listen_stop(sk);
P
Pavel Emelyanov 已提交
2247 2248
	} else if (unlikely(tp->repair)) {
		sk->sk_err = ECONNABORTED;
L
Linus Torvalds 已提交
2249 2250 2251
	} 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
2252
		/* The last check adjusts for discrepancy of Linux wrt. RFC
L
Linus Torvalds 已提交
2253 2254 2255 2256 2257 2258 2259 2260 2261
		 * 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);
2262
	tcp_write_queue_purge(sk);
L
Linus Torvalds 已提交
2263
	__skb_queue_purge(&tp->out_of_order_queue);
C
Chris Leech 已提交
2264 2265 2266
#ifdef CONFIG_NET_DMA
	__skb_queue_purge(&sk->sk_async_wait_queue);
#endif
L
Linus Torvalds 已提交
2267

E
Eric Dumazet 已提交
2268
	inet->inet_dport = 0;
L
Linus Torvalds 已提交
2269 2270 2271 2272 2273 2274 2275 2276 2277

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

	sk->sk_shutdown = 0;
	sock_reset_flag(sk, SOCK_DONE);
	tp->srtt = 0;
	if ((tp->write_seq += tp->max_window + 2) == 0)
		tp->write_seq = 1;
2278
	icsk->icsk_backoff = 0;
L
Linus Torvalds 已提交
2279
	tp->snd_cwnd = 2;
2280
	icsk->icsk_probes_out = 0;
L
Linus Torvalds 已提交
2281
	tp->packets_out = 0;
I
Ilpo Järvinen 已提交
2282
	tp->snd_ssthresh = TCP_INFINITE_SSTHRESH;
L
Linus Torvalds 已提交
2283
	tp->snd_cwnd_cnt = 0;
2284
	tp->window_clamp = 0;
2285
	tcp_set_ca_state(sk, TCP_CA_Open);
L
Linus Torvalds 已提交
2286
	tcp_clear_retrans(tp);
2287
	inet_csk_delack_init(sk);
2288
	tcp_init_send_head(sk);
2289
	memset(&tp->rx_opt, 0, sizeof(tp->rx_opt));
L
Linus Torvalds 已提交
2290 2291
	__sk_dst_reset(sk);

E
Eric Dumazet 已提交
2292
	WARN_ON(inet->inet_num && !icsk->icsk_bind_hash);
L
Linus Torvalds 已提交
2293 2294 2295 2296

	sk->sk_error_report(sk);
	return err;
}
E
Eric Dumazet 已提交
2297
EXPORT_SYMBOL(tcp_disconnect);
L
Linus Torvalds 已提交
2298

2299 2300 2301 2302 2303 2304 2305
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
2306
static inline bool tcp_can_repair_sock(const struct sock *sk)
P
Pavel Emelyanov 已提交
2307
{
2308
	return ns_capable(sock_net(sk)->user_ns, CAP_NET_ADMIN) &&
P
Pavel Emelyanov 已提交
2309 2310 2311
		((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_ESTABLISHED));
}

2312 2313
static int tcp_repair_options_est(struct tcp_sock *tp,
		struct tcp_repair_opt __user *optbuf, unsigned int len)
2314
{
2315
	struct tcp_repair_opt opt;
2316

2317 2318
	while (len >= sizeof(opt)) {
		if (copy_from_user(&opt, optbuf, sizeof(opt)))
2319 2320 2321
			return -EFAULT;

		optbuf++;
2322
		len -= sizeof(opt);
2323

2324 2325 2326
		switch (opt.opt_code) {
		case TCPOPT_MSS:
			tp->rx_opt.mss_clamp = opt.opt_val;
2327
			break;
2328
		case TCPOPT_WINDOW:
2329 2330 2331 2332 2333 2334
			{
				u16 snd_wscale = opt.opt_val & 0xFFFF;
				u16 rcv_wscale = opt.opt_val >> 16;

				if (snd_wscale > 14 || rcv_wscale > 14)
					return -EFBIG;
2335

2336 2337 2338 2339
				tp->rx_opt.snd_wscale = snd_wscale;
				tp->rx_opt.rcv_wscale = rcv_wscale;
				tp->rx_opt.wscale_ok = 1;
			}
2340 2341
			break;
		case TCPOPT_SACK_PERM:
2342 2343 2344
			if (opt.opt_val != 0)
				return -EINVAL;

2345 2346 2347 2348 2349
			tp->rx_opt.sack_ok |= TCP_SACK_SEEN;
			if (sysctl_tcp_fack)
				tcp_enable_fack(tp);
			break;
		case TCPOPT_TIMESTAMP:
2350 2351 2352
			if (opt.opt_val != 0)
				return -EINVAL;

2353 2354 2355 2356 2357 2358 2359 2360
			tp->rx_opt.tstamp_ok = 1;
			break;
		}
	}

	return 0;
}

L
Linus Torvalds 已提交
2361 2362 2363
/*
 *	Socket option code for TCP.
 */
2364
static int do_tcp_setsockopt(struct sock *sk, int level,
2365
		int optname, char __user *optval, unsigned int optlen)
L
Linus Torvalds 已提交
2366 2367
{
	struct tcp_sock *tp = tcp_sk(sk);
2368
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
2369 2370 2371
	int val;
	int err = 0;

2372 2373 2374
	/* These are data/string values, all the others are ints */
	switch (optname) {
	case TCP_CONGESTION: {
2375 2376 2377 2378 2379 2380
		char name[TCP_CA_NAME_MAX];

		if (optlen < 1)
			return -EINVAL;

		val = strncpy_from_user(name, optval,
2381
					min_t(long, TCP_CA_NAME_MAX-1, optlen));
2382 2383 2384 2385 2386
		if (val < 0)
			return -EFAULT;
		name[val] = 0;

		lock_sock(sk);
2387
		err = tcp_set_congestion_control(sk, name);
2388 2389 2390
		release_sock(sk);
		return err;
	}
2391 2392 2393
	default:
		/* fallthru */
		break;
2394
	}
2395

L
Linus Torvalds 已提交
2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408
	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 */
2409
		if (val < TCP_MIN_MSS || val > MAX_TCP_WINDOW) {
L
Linus Torvalds 已提交
2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426
			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;
2427
			tcp_push_pending_frames(sk);
L
Linus Torvalds 已提交
2428 2429 2430 2431 2432
		} else {
			tp->nonagle &= ~TCP_NAGLE_OFF;
		}
		break;

A
Andreas Petlund 已提交
2433 2434 2435 2436 2437 2438 2439
	case TCP_THIN_LINEAR_TIMEOUTS:
		if (val < 0 || val > 1)
			err = -EINVAL;
		else
			tp->thin_lto = val;
		break;

A
Andreas Petlund 已提交
2440 2441 2442 2443 2444
	case TCP_THIN_DUPACK:
		if (val < 0 || val > 1)
			err = -EINVAL;
		else
			tp->thin_dupack = val;
Y
Yuchung Cheng 已提交
2445 2446
			if (tp->thin_dupack)
				tcp_disable_early_retrans(tp);
A
Andreas Petlund 已提交
2447 2448
		break;

P
Pavel Emelyanov 已提交
2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484
	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;

2485 2486 2487 2488
	case TCP_REPAIR_OPTIONS:
		if (!tp->repair)
			err = -EINVAL;
		else if (sk->sk_state == TCP_ESTABLISHED)
2489 2490 2491
			err = tcp_repair_options_est(tp,
					(struct tcp_repair_opt __user *)optval,
					optlen);
2492 2493 2494 2495
		else
			err = -EPERM;
		break;

L
Linus Torvalds 已提交
2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513
	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;
2514
			tcp_push_pending_frames(sk);
L
Linus Torvalds 已提交
2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525
		}
		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))) {
2526
				u32 elapsed = keepalive_time_elapsed(tp);
L
Linus Torvalds 已提交
2527 2528 2529 2530
				if (tp->keepalive_time > elapsed)
					elapsed = tp->keepalive_time - elapsed;
				else
					elapsed = 0;
2531
				inet_csk_reset_keepalive_timer(sk, elapsed);
L
Linus Torvalds 已提交
2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550
			}
		}
		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
2551
			icsk->icsk_syn_retries = val;
L
Linus Torvalds 已提交
2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563
		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:
2564 2565 2566 2567
		/* 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
2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583
		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) {
2584
			icsk->icsk_ack.pingpong = 1;
L
Linus Torvalds 已提交
2585
		} else {
2586
			icsk->icsk_ack.pingpong = 0;
L
Linus Torvalds 已提交
2587 2588
			if ((1 << sk->sk_state) &
			    (TCPF_ESTABLISHED | TCPF_CLOSE_WAIT) &&
2589 2590
			    inet_csk_ack_scheduled(sk)) {
				icsk->icsk_ack.pending |= ICSK_ACK_PUSHED;
2591
				tcp_cleanup_rbuf(sk, 1);
L
Linus Torvalds 已提交
2592
				if (!(val & 1))
2593
					icsk->icsk_ack.pingpong = 1;
L
Linus Torvalds 已提交
2594 2595 2596 2597
			}
		}
		break;

2598 2599 2600 2601 2602 2603
#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
2604 2605 2606 2607
	case TCP_USER_TIMEOUT:
		/* Cap the max timeout in ms TCP will retry/retrans
		 * before giving up and aborting (ETIMEDOUT) a connection.
		 */
2608 2609 2610 2611
		if (val < 0)
			err = -EINVAL;
		else
			icsk->icsk_user_timeout = msecs_to_jiffies(val);
2612
		break;
2613 2614 2615 2616 2617 2618 2619 2620

	case TCP_FASTOPEN:
		if (val >= 0 && ((1 << sk->sk_state) & (TCPF_CLOSE |
		    TCPF_LISTEN)))
			err = fastopen_init_queue(sk, val);
		else
			err = -EINVAL;
		break;
2621 2622 2623 2624 2625 2626
	case TCP_TIMESTAMP:
		if (!tp->repair)
			err = -EPERM;
		else
			tp->tsoffset = val - tcp_time_stamp;
		break;
L
Linus Torvalds 已提交
2627 2628 2629
	default:
		err = -ENOPROTOOPT;
		break;
2630 2631
	}

L
Linus Torvalds 已提交
2632 2633 2634 2635
	release_sock(sk);
	return err;
}

2636
int tcp_setsockopt(struct sock *sk, int level, int optname, char __user *optval,
2637
		   unsigned int optlen)
2638
{
2639
	const struct inet_connection_sock *icsk = inet_csk(sk);
2640 2641 2642 2643 2644 2645

	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
2646
EXPORT_SYMBOL(tcp_setsockopt);
2647 2648

#ifdef CONFIG_COMPAT
2649
int compat_tcp_setsockopt(struct sock *sk, int level, int optname,
2650
			  char __user *optval, unsigned int optlen)
2651
{
2652 2653 2654
	if (level != SOL_TCP)
		return inet_csk_compat_setsockopt(sk, level, optname,
						  optval, optlen);
2655 2656
	return do_tcp_setsockopt(sk, level, optname, optval, optlen);
}
2657
EXPORT_SYMBOL(compat_tcp_setsockopt);
2658 2659
#endif

L
Linus Torvalds 已提交
2660
/* Return information about state of tcp endpoint in API format. */
2661
void tcp_get_info(const struct sock *sk, struct tcp_info *info)
L
Linus Torvalds 已提交
2662
{
2663
	const struct tcp_sock *tp = tcp_sk(sk);
2664
	const struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
2665 2666 2667 2668 2669
	u32 now = tcp_time_stamp;

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

	info->tcpi_state = sk->sk_state;
2670
	info->tcpi_ca_state = icsk->icsk_ca_state;
2671
	info->tcpi_retransmits = icsk->icsk_retransmits;
2672
	info->tcpi_probes = icsk->icsk_probes_out;
2673
	info->tcpi_backoff = icsk->icsk_backoff;
L
Linus Torvalds 已提交
2674 2675 2676

	if (tp->rx_opt.tstamp_ok)
		info->tcpi_options |= TCPI_OPT_TIMESTAMPS;
2677
	if (tcp_is_sack(tp))
L
Linus Torvalds 已提交
2678 2679 2680 2681 2682
		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;
2683
	}
L
Linus Torvalds 已提交
2684

E
Eric Dumazet 已提交
2685
	if (tp->ecn_flags & TCP_ECN_OK)
L
Linus Torvalds 已提交
2686
		info->tcpi_options |= TCPI_OPT_ECN;
E
Eric Dumazet 已提交
2687 2688
	if (tp->ecn_flags & TCP_ECN_SEEN)
		info->tcpi_options |= TCPI_OPT_ECN_SEEN;
Y
Yuchung Cheng 已提交
2689 2690
	if (tp->syn_data_acked)
		info->tcpi_options |= TCPI_OPT_SYN_DATA;
L
Linus Torvalds 已提交
2691

2692 2693
	info->tcpi_rto = jiffies_to_usecs(icsk->icsk_rto);
	info->tcpi_ato = jiffies_to_usecs(icsk->icsk_ack.ato);
2694
	info->tcpi_snd_mss = tp->mss_cache;
2695
	info->tcpi_rcv_mss = icsk->icsk_ack.rcv_mss;
L
Linus Torvalds 已提交
2696

2697 2698 2699 2700 2701 2702 2703
	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
2704 2705 2706 2707 2708
	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);
2709
	info->tcpi_last_data_recv = jiffies_to_msecs(now - icsk->icsk_ack.lrcvtime);
L
Linus Torvalds 已提交
2710 2711
	info->tcpi_last_ack_recv = jiffies_to_msecs(now - tp->rcv_tstamp);

2712
	info->tcpi_pmtu = icsk->icsk_pmtu_cookie;
L
Linus Torvalds 已提交
2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727
	info->tcpi_rcv_ssthresh = tp->rcv_ssthresh;
	info->tcpi_rtt = jiffies_to_usecs(tp->srtt)>>3;
	info->tcpi_rttvar = jiffies_to_usecs(tp->mdev)>>2;
	info->tcpi_snd_ssthresh = tp->snd_ssthresh;
	info->tcpi_snd_cwnd = tp->snd_cwnd;
	info->tcpi_advmss = tp->advmss;
	info->tcpi_reordering = tp->reordering;

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

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

2728 2729
static int do_tcp_getsockopt(struct sock *sk, int level,
		int optname, char __user *optval, int __user *optlen)
L
Linus Torvalds 已提交
2730
{
2731
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744
	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:
2745
		val = tp->mss_cache;
L
Linus Torvalds 已提交
2746 2747
		if (!val && ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN)))
			val = tp->rx_opt.user_mss;
2748 2749
		if (tp->repair)
			val = tp->rx_opt.mss_clamp;
L
Linus Torvalds 已提交
2750 2751 2752 2753 2754 2755 2756 2757
		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
2758
		val = keepalive_time_when(tp) / HZ;
L
Linus Torvalds 已提交
2759 2760
		break;
	case TCP_KEEPINTVL:
E
Eric Dumazet 已提交
2761
		val = keepalive_intvl_when(tp) / HZ;
L
Linus Torvalds 已提交
2762 2763
		break;
	case TCP_KEEPCNT:
E
Eric Dumazet 已提交
2764
		val = keepalive_probes(tp);
L
Linus Torvalds 已提交
2765 2766
		break;
	case TCP_SYNCNT:
2767
		val = icsk->icsk_syn_retries ? : sysctl_tcp_syn_retries;
L
Linus Torvalds 已提交
2768 2769 2770 2771 2772 2773 2774
		break;
	case TCP_LINGER2:
		val = tp->linger2;
		if (val >= 0)
			val = (val ? : sysctl_tcp_fin_timeout) / HZ;
		break;
	case TCP_DEFER_ACCEPT:
2775 2776
		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
2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796
		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:
2797
		val = !icsk->icsk_ack.pingpong;
L
Linus Torvalds 已提交
2798
		break;
2799 2800 2801 2802 2803 2804 2805

	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;
2806
		if (copy_to_user(optval, icsk->icsk_ca_ops->name, len))
2807 2808
			return -EFAULT;
		return 0;
2809

2810 2811 2812 2813 2814 2815
	case TCP_THIN_LINEAR_TIMEOUTS:
		val = tp->thin_lto;
		break;
	case TCP_THIN_DUPACK:
		val = tp->thin_dupack;
		break;
2816

P
Pavel Emelyanov 已提交
2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836
	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;

2837 2838 2839
	case TCP_USER_TIMEOUT:
		val = jiffies_to_msecs(icsk->icsk_user_timeout);
		break;
2840 2841 2842
	case TCP_TIMESTAMP:
		val = tcp_time_stamp + tp->tsoffset;
		break;
L
Linus Torvalds 已提交
2843 2844
	default:
		return -ENOPROTOOPT;
2845
	}
L
Linus Torvalds 已提交
2846 2847 2848 2849 2850 2851 2852 2853

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

2854 2855 2856 2857 2858 2859 2860 2861 2862 2863
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
2864
EXPORT_SYMBOL(tcp_getsockopt);
2865 2866

#ifdef CONFIG_COMPAT
2867 2868
int compat_tcp_getsockopt(struct sock *sk, int level, int optname,
			  char __user *optval, int __user *optlen)
2869
{
2870 2871 2872
	if (level != SOL_TCP)
		return inet_csk_compat_getsockopt(sk, level, optname,
						  optval, optlen);
2873 2874
	return do_tcp_getsockopt(sk, level, optname, optval, optlen);
}
2875
EXPORT_SYMBOL(compat_tcp_getsockopt);
2876
#endif
L
Linus Torvalds 已提交
2877

2878 2879
struct sk_buff *tcp_tso_segment(struct sk_buff *skb,
	netdev_features_t features)
H
Herbert Xu 已提交
2880 2881 2882
{
	struct sk_buff *segs = ERR_PTR(-EINVAL);
	struct tcphdr *th;
2883
	unsigned int thlen;
H
Herbert Xu 已提交
2884
	unsigned int seq;
2885
	__be32 delta;
H
Herbert Xu 已提交
2886
	unsigned int oldlen;
2887
	unsigned int mss;
E
Eric Dumazet 已提交
2888
	struct sk_buff *gso_skb = skb;
2889
	__sum16 newcheck;
H
Herbert Xu 已提交
2890 2891 2892 2893

	if (!pskb_may_pull(skb, sizeof(*th)))
		goto out;

2894
	th = tcp_hdr(skb);
H
Herbert Xu 已提交
2895 2896 2897 2898 2899 2900 2901
	thlen = th->doff * 4;
	if (thlen < sizeof(*th))
		goto out;

	if (!pskb_may_pull(skb, thlen))
		goto out;

2902
	oldlen = (u16)~skb->len;
H
Herbert Xu 已提交
2903 2904
	__skb_pull(skb, thlen);

2905 2906 2907 2908
	mss = skb_shinfo(skb)->gso_size;
	if (unlikely(skb->len <= mss))
		goto out;

2909 2910
	if (skb_gso_ok(skb, features | NETIF_F_GSO_ROBUST)) {
		/* Packet is from an untrusted source, reset gso_segs. */
2911 2912 2913 2914 2915 2916 2917
		int type = skb_shinfo(skb)->gso_type;

		if (unlikely(type &
			     ~(SKB_GSO_TCPV4 |
			       SKB_GSO_DODGY |
			       SKB_GSO_TCP_ECN |
			       SKB_GSO_TCPV6 |
2918
			       SKB_GSO_GRE |
2919
			       SKB_GSO_UDP_TUNNEL |
2920 2921 2922
			       0) ||
			     !(type & (SKB_GSO_TCPV4 | SKB_GSO_TCPV6))))
			goto out;
2923

2924
		skb_shinfo(skb)->gso_segs = DIV_ROUND_UP(skb->len, mss);
2925 2926 2927 2928 2929

		segs = NULL;
		goto out;
	}

2930
	segs = skb_segment(skb, features);
H
Herbert Xu 已提交
2931 2932 2933
	if (IS_ERR(segs))
		goto out;

2934
	delta = htonl(oldlen + (thlen + mss));
H
Herbert Xu 已提交
2935 2936

	skb = segs;
2937
	th = tcp_hdr(skb);
H
Herbert Xu 已提交
2938 2939
	seq = ntohl(th->seq);

2940 2941 2942
	newcheck = ~csum_fold((__force __wsum)((__force u32)th->check +
					       (__force u32)delta));

H
Herbert Xu 已提交
2943 2944
	do {
		th->fin = th->psh = 0;
2945
		th->check = newcheck;
H
Herbert Xu 已提交
2946

2947
		if (skb->ip_summed != CHECKSUM_PARTIAL)
2948 2949 2950
			th->check =
			     csum_fold(csum_partial(skb_transport_header(skb),
						    thlen, skb->csum));
H
Herbert Xu 已提交
2951

2952
		seq += mss;
H
Herbert Xu 已提交
2953
		skb = skb->next;
2954
		th = tcp_hdr(skb);
H
Herbert Xu 已提交
2955 2956 2957 2958 2959

		th->seq = htonl(seq);
		th->cwr = 0;
	} while (skb->next);

E
Eric Dumazet 已提交
2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970
	/* Following permits TCP Small Queues to work well with GSO :
	 * The callback to TCP stack will be called at the time last frag
	 * is freed at TX completion, and not right now when gso_skb
	 * is freed by GSO engine
	 */
	if (gso_skb->destructor == tcp_wfree) {
		swap(gso_skb->sk, skb->sk);
		swap(gso_skb->destructor, skb->destructor);
		swap(gso_skb->truesize, skb->truesize);
	}

2971
	delta = htonl(oldlen + (skb->tail - skb->transport_header) +
2972
		      skb->data_len);
2973 2974
	th->check = ~csum_fold((__force __wsum)((__force u32)th->check +
				(__force u32)delta));
2975
	if (skb->ip_summed != CHECKSUM_PARTIAL)
2976 2977
		th->check = csum_fold(csum_partial(skb_transport_header(skb),
						   thlen, skb->csum));
H
Herbert Xu 已提交
2978 2979 2980 2981

out:
	return segs;
}
H
Herbert Xu 已提交
2982
EXPORT_SYMBOL(tcp_tso_segment);
H
Herbert Xu 已提交
2983

H
Herbert Xu 已提交
2984 2985 2986 2987 2988 2989
struct sk_buff **tcp_gro_receive(struct sk_buff **head, struct sk_buff *skb)
{
	struct sk_buff **pp = NULL;
	struct sk_buff *p;
	struct tcphdr *th;
	struct tcphdr *th2;
2990
	unsigned int len;
H
Herbert Xu 已提交
2991
	unsigned int thlen;
2992
	__be32 flags;
H
Herbert Xu 已提交
2993
	unsigned int mss = 1;
2994 2995
	unsigned int hlen;
	unsigned int off;
H
Herbert Xu 已提交
2996
	int flush = 1;
H
Herbert Xu 已提交
2997
	int i;
H
Herbert Xu 已提交
2998

2999 3000 3001 3002 3003 3004 3005 3006
	off = skb_gro_offset(skb);
	hlen = off + sizeof(*th);
	th = skb_gro_header_fast(skb, off);
	if (skb_gro_header_hard(skb, hlen)) {
		th = skb_gro_header_slow(skb, hlen, off);
		if (unlikely(!th))
			goto out;
	}
H
Herbert Xu 已提交
3007 3008 3009 3010 3011

	thlen = th->doff * 4;
	if (thlen < sizeof(*th))
		goto out;

3012 3013 3014 3015 3016 3017
	hlen = off + thlen;
	if (skb_gro_header_hard(skb, hlen)) {
		th = skb_gro_header_slow(skb, hlen, off);
		if (unlikely(!th))
			goto out;
	}
H
Herbert Xu 已提交
3018

3019
	skb_gro_pull(skb, thlen);
H
Herbert Xu 已提交
3020

3021
	len = skb_gro_len(skb);
H
Herbert Xu 已提交
3022 3023 3024 3025 3026 3027 3028 3029
	flags = tcp_flag_word(th);

	for (; (p = *head); head = &p->next) {
		if (!NAPI_GRO_CB(p)->same_flow)
			continue;

		th2 = tcp_hdr(p);

H
Herbert Xu 已提交
3030
		if (*(u32 *)&th->source ^ *(u32 *)&th2->source) {
H
Herbert Xu 已提交
3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041
			NAPI_GRO_CB(p)->same_flow = 0;
			continue;
		}

		goto found;
	}

	goto out_check_final;

found:
	flush = NAPI_GRO_CB(p)->flush;
3042 3043 3044 3045
	flush |= (__force int)(flags & TCP_FLAG_CWR);
	flush |= (__force int)((flags ^ tcp_flag_word(th2)) &
		  ~(TCP_FLAG_CWR | TCP_FLAG_FIN | TCP_FLAG_PSH));
	flush |= (__force int)(th->ack_seq ^ th2->ack_seq);
3046
	for (i = sizeof(*th); i < thlen; i += 4)
H
Herbert Xu 已提交
3047 3048
		flush |= *(u32 *)((u8 *)th + i) ^
			 *(u32 *)((u8 *)th2 + i);
H
Herbert Xu 已提交
3049

H
Herbert Xu 已提交
3050
	mss = skb_shinfo(p)->gso_size;
H
Herbert Xu 已提交
3051

H
Herbert Xu 已提交
3052
	flush |= (len - 1) >= mss;
H
Herbert Xu 已提交
3053
	flush |= (ntohl(th2->seq) + skb_gro_len(p)) ^ ntohl(th->seq);
H
Herbert Xu 已提交
3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064

	if (flush || skb_gro_receive(head, skb)) {
		mss = 1;
		goto out_check_final;
	}

	p = *head;
	th2 = tcp_hdr(p);
	tcp_flag_word(th2) |= flags & (TCP_FLAG_FIN | TCP_FLAG_PSH);

out_check_final:
3065
	flush = len < mss;
3066 3067 3068
	flush |= (__force int)(flags & (TCP_FLAG_URG | TCP_FLAG_PSH |
					TCP_FLAG_RST | TCP_FLAG_SYN |
					TCP_FLAG_FIN));
H
Herbert Xu 已提交
3069 3070 3071 3072 3073 3074 3075 3076 3077

	if (p && (!NAPI_GRO_CB(skb)->same_flow || flush))
		pp = head;

out:
	NAPI_GRO_CB(skb)->flush |= flush;

	return pp;
}
H
Herbert Xu 已提交
3078
EXPORT_SYMBOL(tcp_gro_receive);
H
Herbert Xu 已提交
3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094

int tcp_gro_complete(struct sk_buff *skb)
{
	struct tcphdr *th = tcp_hdr(skb);

	skb->csum_start = skb_transport_header(skb) - skb->head;
	skb->csum_offset = offsetof(struct tcphdr, check);
	skb->ip_summed = CHECKSUM_PARTIAL;

	skb_shinfo(skb)->gso_segs = NAPI_GRO_CB(skb)->count;

	if (th->cwr)
		skb_shinfo(skb)->gso_type |= SKB_GSO_TCP_ECN;

	return 0;
}
H
Herbert Xu 已提交
3095
EXPORT_SYMBOL(tcp_gro_complete);
H
Herbert Xu 已提交
3096

3097 3098
#ifdef CONFIG_TCP_MD5SIG
static unsigned long tcp_md5sig_users;
3099
static struct tcp_md5sig_pool __percpu *tcp_md5sig_pool;
3100 3101
static DEFINE_SPINLOCK(tcp_md5sig_pool_lock);

3102
static void __tcp_free_md5sig_pool(struct tcp_md5sig_pool __percpu *pool)
3103 3104
{
	int cpu;
3105

3106
	for_each_possible_cpu(cpu) {
3107 3108 3109 3110
		struct tcp_md5sig_pool *p = per_cpu_ptr(pool, cpu);

		if (p->md5_desc.tfm)
			crypto_free_hash(p->md5_desc.tfm);
3111 3112 3113 3114 3115 3116
	}
	free_percpu(pool);
}

void tcp_free_md5sig_pool(void)
{
3117
	struct tcp_md5sig_pool __percpu *pool = NULL;
3118

3119
	spin_lock_bh(&tcp_md5sig_pool_lock);
3120 3121 3122 3123
	if (--tcp_md5sig_users == 0) {
		pool = tcp_md5sig_pool;
		tcp_md5sig_pool = NULL;
	}
3124
	spin_unlock_bh(&tcp_md5sig_pool_lock);
3125 3126 3127 3128 3129
	if (pool)
		__tcp_free_md5sig_pool(pool);
}
EXPORT_SYMBOL(tcp_free_md5sig_pool);

3130
static struct tcp_md5sig_pool __percpu *
3131
__tcp_alloc_md5sig_pool(struct sock *sk)
3132 3133
{
	int cpu;
3134
	struct tcp_md5sig_pool __percpu *pool;
3135

3136
	pool = alloc_percpu(struct tcp_md5sig_pool);
3137 3138 3139 3140 3141 3142 3143
	if (!pool)
		return NULL;

	for_each_possible_cpu(cpu) {
		struct crypto_hash *hash;

		hash = crypto_alloc_hash("md5", 0, CRYPTO_ALG_ASYNC);
3144
		if (IS_ERR_OR_NULL(hash))
3145 3146
			goto out_free;

3147
		per_cpu_ptr(pool, cpu)->md5_desc.tfm = hash;
3148 3149 3150 3151 3152 3153 3154
	}
	return pool;
out_free:
	__tcp_free_md5sig_pool(pool);
	return NULL;
}

3155
struct tcp_md5sig_pool __percpu *tcp_alloc_md5sig_pool(struct sock *sk)
3156
{
3157
	struct tcp_md5sig_pool __percpu *pool;
E
Eric Dumazet 已提交
3158
	bool alloc = false;
3159 3160

retry:
3161
	spin_lock_bh(&tcp_md5sig_pool_lock);
3162 3163
	pool = tcp_md5sig_pool;
	if (tcp_md5sig_users++ == 0) {
E
Eric Dumazet 已提交
3164
		alloc = true;
3165
		spin_unlock_bh(&tcp_md5sig_pool_lock);
3166 3167
	} else if (!pool) {
		tcp_md5sig_users--;
3168
		spin_unlock_bh(&tcp_md5sig_pool_lock);
3169 3170 3171
		cpu_relax();
		goto retry;
	} else
3172
		spin_unlock_bh(&tcp_md5sig_pool_lock);
3173 3174 3175

	if (alloc) {
		/* we cannot hold spinlock here because this may sleep. */
3176
		struct tcp_md5sig_pool __percpu *p;
3177 3178

		p = __tcp_alloc_md5sig_pool(sk);
3179
		spin_lock_bh(&tcp_md5sig_pool_lock);
3180 3181
		if (!p) {
			tcp_md5sig_users--;
3182
			spin_unlock_bh(&tcp_md5sig_pool_lock);
3183 3184 3185 3186 3187
			return NULL;
		}
		pool = tcp_md5sig_pool;
		if (pool) {
			/* oops, it has already been assigned. */
3188
			spin_unlock_bh(&tcp_md5sig_pool_lock);
3189 3190 3191
			__tcp_free_md5sig_pool(p);
		} else {
			tcp_md5sig_pool = pool = p;
3192
			spin_unlock_bh(&tcp_md5sig_pool_lock);
3193 3194 3195 3196 3197 3198
		}
	}
	return pool;
}
EXPORT_SYMBOL(tcp_alloc_md5sig_pool);

E
Eric Dumazet 已提交
3199 3200 3201 3202 3203 3204 3205 3206 3207

/**
 *	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)
3208
{
3209
	struct tcp_md5sig_pool __percpu *p;
E
Eric Dumazet 已提交
3210 3211 3212 3213

	local_bh_disable();

	spin_lock(&tcp_md5sig_pool_lock);
3214 3215 3216
	p = tcp_md5sig_pool;
	if (p)
		tcp_md5sig_users++;
E
Eric Dumazet 已提交
3217 3218 3219
	spin_unlock(&tcp_md5sig_pool_lock);

	if (p)
3220
		return this_cpu_ptr(p);
3221

E
Eric Dumazet 已提交
3222 3223 3224 3225
	local_bh_enable();
	return NULL;
}
EXPORT_SYMBOL(tcp_get_md5sig_pool);
3226

E
Eric Dumazet 已提交
3227
void tcp_put_md5sig_pool(void)
3228
{
E
Eric Dumazet 已提交
3229
	local_bh_enable();
3230
	tcp_free_md5sig_pool();
3231
}
E
Eric Dumazet 已提交
3232
EXPORT_SYMBOL(tcp_put_md5sig_pool);
3233 3234

int tcp_md5_hash_header(struct tcp_md5sig_pool *hp,
3235
			const struct tcphdr *th)
3236 3237
{
	struct scatterlist sg;
3238
	struct tcphdr hdr;
3239 3240
	int err;

3241 3242 3243 3244
	/* We are not allowed to change tcphdr, make a local copy */
	memcpy(&hdr, th, sizeof(hdr));
	hdr.check = 0;

3245
	/* options aren't included in the hash */
3246 3247
	sg_init_one(&sg, &hdr, sizeof(hdr));
	err = crypto_hash_update(&hp->md5_desc, &sg, sizeof(hdr));
3248 3249 3250 3251 3252
	return err;
}
EXPORT_SYMBOL(tcp_md5_hash_header);

int tcp_md5_hash_skb_data(struct tcp_md5sig_pool *hp,
3253
			  const struct sk_buff *skb, unsigned int header_len)
3254 3255 3256 3257
{
	struct scatterlist sg;
	const struct tcphdr *tp = tcp_hdr(skb);
	struct hash_desc *desc = &hp->md5_desc;
3258 3259 3260
	unsigned int i;
	const unsigned int head_data_len = skb_headlen(skb) > header_len ?
					   skb_headlen(skb) - header_len : 0;
3261
	const struct skb_shared_info *shi = skb_shinfo(skb);
3262
	struct sk_buff *frag_iter;
3263 3264 3265 3266 3267 3268 3269 3270 3271

	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];
3272
		struct page *page = skb_frag_page(f);
E
Eric Dumazet 已提交
3273 3274
		sg_set_page(&sg, page, skb_frag_size(f), f->page_offset);
		if (crypto_hash_update(desc, &sg, skb_frag_size(f)))
3275 3276 3277
			return 1;
	}

3278 3279 3280 3281
	skb_walk_frags(skb, frag_iter)
		if (tcp_md5_hash_skb_data(hp, frag_iter, 0))
			return 1;

3282 3283 3284 3285
	return 0;
}
EXPORT_SYMBOL(tcp_md5_hash_skb_data);

3286
int tcp_md5_hash_key(struct tcp_md5sig_pool *hp, const struct tcp_md5sig_key *key)
3287 3288 3289 3290 3291 3292 3293 3294
{
	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);

3295 3296
#endif

A
Andi Kleen 已提交
3297 3298
void tcp_done(struct sock *sk)
{
3299 3300
	struct request_sock *req = tcp_sk(sk)->fastopen_rsk;

3301
	if (sk->sk_state == TCP_SYN_SENT || sk->sk_state == TCP_SYN_RECV)
3302
		TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_ATTEMPTFAILS);
A
Andi Kleen 已提交
3303 3304 3305

	tcp_set_state(sk, TCP_CLOSE);
	tcp_clear_xmit_timers(sk);
3306 3307
	if (req != NULL)
		reqsk_fastopen_remove(sk, req, false);
A
Andi Kleen 已提交
3308 3309 3310 3311 3312 3313 3314 3315 3316 3317

	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);

3318
extern struct tcp_congestion_ops tcp_reno;
L
Linus Torvalds 已提交
3319 3320 3321 3322

static __initdata unsigned long thash_entries;
static int __init set_thash_entries(char *str)
{
3323 3324
	ssize_t ret;

L
Linus Torvalds 已提交
3325 3326
	if (!str)
		return 0;
3327 3328 3329 3330 3331

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

L
Linus Torvalds 已提交
3332 3333 3334 3335
	return 1;
}
__setup("thash_entries=", set_thash_entries);

3336 3337 3338 3339 3340 3341 3342 3343 3344
void tcp_init_mem(struct net *net)
{
	unsigned long limit = nr_free_buffer_pages() / 8;
	limit = max(limit, 128UL);
	net->ipv4.sysctl_tcp_mem[0] = limit / 4 * 3;
	net->ipv4.sysctl_tcp_mem[1] = limit;
	net->ipv4.sysctl_tcp_mem[2] = net->ipv4.sysctl_tcp_mem[0] * 2;
}

L
Linus Torvalds 已提交
3345 3346 3347
void __init tcp_init(void)
{
	struct sk_buff *skb = NULL;
3348
	unsigned long limit;
3349
	int max_rshare, max_wshare, cnt;
3350
	unsigned int i;
L
Linus Torvalds 已提交
3351

3352
	BUILD_BUG_ON(sizeof(struct tcp_skb_cb) > sizeof(skb->cb));
L
Linus Torvalds 已提交
3353

3354
	percpu_counter_init(&tcp_sockets_allocated, 0);
3355
	percpu_counter_init(&tcp_orphan_count, 0);
3356 3357 3358
	tcp_hashinfo.bind_bucket_cachep =
		kmem_cache_create("tcp_bind_bucket",
				  sizeof(struct inet_bind_bucket), 0,
3359
				  SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
L
Linus Torvalds 已提交
3360 3361 3362 3363 3364 3365

	/* Size and allocate the main established and bind bucket
	 * hash tables.
	 *
	 * The methodology is similar to that of the buffer cache.
	 */
3366
	tcp_hashinfo.ehash =
L
Linus Torvalds 已提交
3367
		alloc_large_system_hash("TCP established",
3368
					sizeof(struct inet_ehash_bucket),
L
Linus Torvalds 已提交
3369
					thash_entries,
E
Eric Dumazet 已提交
3370
					17, /* one slot per 128 KB of memory */
3371
					0,
L
Linus Torvalds 已提交
3372
					NULL,
3373
					&tcp_hashinfo.ehash_mask,
3374
					0,
3375
					thash_entries ? 0 : 512 * 1024);
3376
	for (i = 0; i <= tcp_hashinfo.ehash_mask; i++) {
3377 3378
		INIT_HLIST_NULLS_HEAD(&tcp_hashinfo.ehash[i].chain, i);
		INIT_HLIST_NULLS_HEAD(&tcp_hashinfo.ehash[i].twchain, i);
L
Linus Torvalds 已提交
3379
	}
3380 3381
	if (inet_ehash_locks_alloc(&tcp_hashinfo))
		panic("TCP: failed to alloc ehash_locks");
3382
	tcp_hashinfo.bhash =
L
Linus Torvalds 已提交
3383
		alloc_large_system_hash("TCP bind",
3384
					sizeof(struct inet_bind_hashbucket),
3385
					tcp_hashinfo.ehash_mask + 1,
E
Eric Dumazet 已提交
3386
					17, /* one slot per 128 KB of memory */
3387
					0,
3388
					&tcp_hashinfo.bhash_size,
L
Linus Torvalds 已提交
3389
					NULL,
3390
					0,
L
Linus Torvalds 已提交
3391
					64 * 1024);
3392
	tcp_hashinfo.bhash_size = 1U << tcp_hashinfo.bhash_size;
3393 3394 3395
	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
3396 3397
	}

E
Eric Dumazet 已提交
3398 3399 3400 3401 3402 3403

	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
3404

3405
	tcp_init_mem(&init_net);
3406
	/* Set per-socket limits to no more than 1/128 the pressure threshold */
E
Eric Dumazet 已提交
3407
	limit = nr_free_buffer_pages() << (PAGE_SHIFT - 7);
3408 3409
	max_wshare = min(4UL*1024*1024, limit);
	max_rshare = min(6UL*1024*1024, limit);
3410

3411
	sysctl_tcp_wmem[0] = SK_MEM_QUANTUM;
3412
	sysctl_tcp_wmem[1] = 16*1024;
3413
	sysctl_tcp_wmem[2] = max(64*1024, max_wshare);
3414

3415
	sysctl_tcp_rmem[0] = SK_MEM_QUANTUM;
3416
	sysctl_tcp_rmem[1] = 87380;
3417
	sysctl_tcp_rmem[2] = max(87380, max_rshare);
L
Linus Torvalds 已提交
3418

3419
	pr_info("Hash tables configured (established %u bind %u)\n",
J
Joe Perches 已提交
3420
		tcp_hashinfo.ehash_mask + 1, tcp_hashinfo.bhash_size);
3421

3422 3423
	tcp_metrics_init();

3424
	tcp_register_congestion_control(&tcp_reno);
3425

E
Eric Dumazet 已提交
3426
	tcp_tasklet_init();
L
Linus Torvalds 已提交
3427
}