tcp.c 93.2 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

H
Herbert Xu 已提交
250
#include <crypto/hash.h>
251
#include <linux/kernel.h>
L
Linus Torvalds 已提交
252 253 254 255
#include <linux/module.h>
#include <linux/types.h>
#include <linux/fcntl.h>
#include <linux/poll.h>
256
#include <linux/inet_diag.h>
L
Linus Torvalds 已提交
257 258
#include <linux/init.h>
#include <linux/fs.h>
J
Jens Axboe 已提交
259
#include <linux/skbuff.h>
260
#include <linux/scatterlist.h>
J
Jens Axboe 已提交
261 262 263
#include <linux/splice.h>
#include <linux/net.h>
#include <linux/socket.h>
L
Linus Torvalds 已提交
264 265
#include <linux/random.h>
#include <linux/bootmem.h>
266 267
#include <linux/highmem.h>
#include <linux/swap.h>
268
#include <linux/cache.h>
H
Herbert Xu 已提交
269
#include <linux/err.h>
270
#include <linux/time.h>
271
#include <linux/slab.h>
272
#include <linux/errqueue.h>
273
#include <linux/static_key.h>
L
Linus Torvalds 已提交
274 275

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

282
#include <linux/uaccess.h>
L
Linus Torvalds 已提交
283
#include <asm/ioctls.h>
284
#include <net/busy_poll.h>
L
Linus Torvalds 已提交
285

286 287
#include <trace/events/tcp.h>

288
struct percpu_counter tcp_orphan_count;
289 290
EXPORT_SYMBOL_GPL(tcp_orphan_count);

291 292
long sysctl_tcp_mem[3] __read_mostly;
EXPORT_SYMBOL(sysctl_tcp_mem);
L
Linus Torvalds 已提交
293

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

297 298 299 300 301
#if IS_ENABLED(CONFIG_SMC)
DEFINE_STATIC_KEY_FALSE(tcp_have_smc);
EXPORT_SYMBOL(tcp_have_smc);
#endif

302 303 304 305
/*
 * Current number of TCP sockets.
 */
struct percpu_counter tcp_sockets_allocated;
L
Linus Torvalds 已提交
306 307
EXPORT_SYMBOL(tcp_sockets_allocated);

J
Jens Axboe 已提交
308 309 310 311 312 313 314 315 316
/*
 * TCP splice context
 */
struct tcp_splice_state {
	struct pipe_inode_info *pipe;
	size_t len;
	unsigned int flags;
};

L
Linus Torvalds 已提交
317 318 319
/*
 * Pressure flag: try to collapse.
 * Technical note: it is used by multiple contexts non atomically.
320
 * All the __sk_mem_schedule() is of this nature: accounting
L
Linus Torvalds 已提交
321 322
 * is strict, actions are advisory and have some latency.
 */
323 324
unsigned long tcp_memory_pressure __read_mostly;
EXPORT_SYMBOL_GPL(tcp_memory_pressure);
L
Linus Torvalds 已提交
325

326
void tcp_enter_memory_pressure(struct sock *sk)
L
Linus Torvalds 已提交
327
{
328 329 330 331 332 333 334 335 336
	unsigned long val;

	if (tcp_memory_pressure)
		return;
	val = jiffies;

	if (!val)
		val--;
	if (!cmpxchg(&tcp_memory_pressure, 0, val))
P
Pavel Emelyanov 已提交
337
		NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMEMORYPRESSURES);
L
Linus Torvalds 已提交
338
}
339 340 341 342 343 344 345 346 347 348 349 350 351 352
EXPORT_SYMBOL_GPL(tcp_enter_memory_pressure);

void tcp_leave_memory_pressure(struct sock *sk)
{
	unsigned long val;

	if (!tcp_memory_pressure)
		return;
	val = xchg(&tcp_memory_pressure, 0);
	if (val)
		NET_ADD_STATS(sock_net(sk), LINUX_MIB_TCPMEMORYPRESSURESCHRONO,
			      jiffies_to_msecs(jiffies - val));
}
EXPORT_SYMBOL_GPL(tcp_leave_memory_pressure);
L
Linus Torvalds 已提交
353

354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390
/* 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;
}

391 392 393 394 395 396 397 398 399 400 401 402 403
static u64 tcp_compute_delivery_rate(const struct tcp_sock *tp)
{
	u32 rate = READ_ONCE(tp->rate_delivered);
	u32 intv = READ_ONCE(tp->rate_interval_us);
	u64 rate64 = 0;

	if (rate && intv) {
		rate64 = (u64)rate * tp->mss_cache * USEC_PER_SEC;
		do_div(rate64, intv);
	}
	return rate64;
}

404 405 406 407 408 409 410 411 412 413
/* 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);

414
	tp->out_of_order_queue = RB_ROOT;
415
	sk->tcp_rtx_queue = RB_ROOT;
416
	tcp_init_xmit_timers(sk);
E
Eric Dumazet 已提交
417
	INIT_LIST_HEAD(&tp->tsq_node);
418
	INIT_LIST_HEAD(&tp->tsorted_sent_queue);
419 420

	icsk->icsk_rto = TCP_TIMEOUT_INIT;
421
	tp->mdev_us = jiffies_to_usecs(TCP_TIMEOUT_INIT);
422
	minmax_reset(&tp->rtt_min, tcp_jiffies32, ~0U);
423 424 425 426 427 428 429 430

	/* 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;

431 432 433
	/* There's a bubble in the pipe until at least the first ACK. */
	tp->app_limited = ~0U;

434 435 436 437 438 439 440
	/* 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;

441
	tp->reordering = sock_net(sk)->ipv4.sysctl_tcp_reordering;
442
	tcp_assign_congestion_control(sk);
443

444
	tp->tsoffset = 0;
445
	tp->rack.reo_wnd_steps = 1;
446

447 448 449 450 451 452 453
	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;

454 455
	sk->sk_sndbuf = sock_net(sk)->ipv4.sysctl_tcp_wmem[1];
	sk->sk_rcvbuf = sock_net(sk)->ipv4.sysctl_tcp_rmem[1];
456 457 458 459 460

	sk_sockets_allocated_inc(sk);
}
EXPORT_SYMBOL(tcp_init_sock);

461 462 463 464 465 466 467 468 469 470 471 472
void tcp_init_transfer(struct sock *sk, int bpf_op)
{
	struct inet_connection_sock *icsk = inet_csk(sk);

	tcp_mtup_init(sk);
	icsk->icsk_af_ops->rebuild_header(sk);
	tcp_init_metrics(sk);
	tcp_call_bpf(sk, bpf_op);
	tcp_init_congestion_control(sk);
	tcp_init_buffer_space(sk);
}

E
Eric Dumazet 已提交
473
static void tcp_tx_timestamp(struct sock *sk, u16 tsflags)
474
{
E
Eric Dumazet 已提交
475 476
	struct sk_buff *skb = tcp_write_queue_tail(sk);

477
	if (tsflags && skb) {
478
		struct skb_shared_info *shinfo = skb_shinfo(skb);
479
		struct tcp_skb_cb *tcb = TCP_SKB_CB(skb);
480

481
		sock_tx_timestamp(sk, tsflags, &shinfo->tx_flags);
482 483 484
		if (tsflags & SOF_TIMESTAMPING_TX_ACK)
			tcb->txstamp_ack = 1;
		if (tsflags & SOF_TIMESTAMPING_TX_RECORD_MASK)
485 486
			shinfo->tskey = TCP_SKB_CB(skb)->seq + skb->len - 1;
	}
487 488
}

L
Linus Torvalds 已提交
489 490 491 492 493 494 495 496 497 498 499
/*
 *	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;
500
	const struct tcp_sock *tp = tcp_sk(sk);
501
	int state;
L
Linus Torvalds 已提交
502

503 504
	sock_rps_record_flow(sk);

E
Eric Dumazet 已提交
505
	sock_poll_wait(file, sk_sleep(sk), wait);
506 507 508

	state = sk_state_load(sk);
	if (state == TCP_LISTEN)
509
		return inet_csk_listen_poll(sk);
L
Linus Torvalds 已提交
510 511

	/* Socket is not locked. We are protected from async events
512 513
	 * by poll logic and correct handling of state changes
	 * made by other threads is impossible in any case.
L
Linus Torvalds 已提交
514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536
	 */

	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
537
	 * solve this dilemma. I would prefer, if POLLHUP were maskable,
L
Linus Torvalds 已提交
538 539
	 * then we could set it on SND_SHUTDOWN. BTW examples given
	 * in Stevens' books assume exactly this behaviour, it explains
540
	 * why POLLHUP is incompatible with POLLOUT.	--ANK
L
Linus Torvalds 已提交
541 542 543 544
	 *
	 * NOTE. Check for TCP_CLOSE is added. The goal is to prevent
	 * blocking on fresh not-connected or disconnected socket. --ANK
	 */
545
	if (sk->sk_shutdown == SHUTDOWN_MASK || state == TCP_CLOSE)
L
Linus Torvalds 已提交
546 547
		mask |= POLLHUP;
	if (sk->sk_shutdown & RCV_SHUTDOWN)
548
		mask |= POLLIN | POLLRDNORM | POLLRDHUP;
L
Linus Torvalds 已提交
549

550
	/* Connected or passive Fast Open socket? */
551 552
	if (state != TCP_SYN_SENT &&
	    (state != TCP_SYN_RECV || tp->fastopen_rsk)) {
553 554 555 556 557
		int target = sock_rcvlowat(sk, 0, INT_MAX);

		if (tp->urg_seq == tp->copied_seq &&
		    !sock_flag(sk, SOCK_URGINLINE) &&
		    tp->urg_data)
558
			target++;
559 560

		if (tp->rcv_nxt - tp->copied_seq >= target)
L
Linus Torvalds 已提交
561 562 563
			mask |= POLLIN | POLLRDNORM;

		if (!(sk->sk_shutdown & SEND_SHUTDOWN)) {
564
			if (sk_stream_is_writeable(sk)) {
L
Linus Torvalds 已提交
565 566
				mask |= POLLOUT | POLLWRNORM;
			} else {  /* send SIGIO later */
567
				sk_set_bit(SOCKWQ_ASYNC_NOSPACE, sk);
L
Linus Torvalds 已提交
568 569 570 571
				set_bit(SOCK_NOSPACE, &sk->sk_socket->flags);

				/* Race breaker. If space is freed after
				 * wspace test but before the flags are set,
572 573
				 * IO signal will be lost. Memory barrier
				 * pairs with the input side.
L
Linus Torvalds 已提交
574
				 */
575
				smp_mb__after_atomic();
576
				if (sk_stream_is_writeable(sk))
L
Linus Torvalds 已提交
577 578
					mask |= POLLOUT | POLLWRNORM;
			}
579 580
		} else
			mask |= POLLOUT | POLLWRNORM;
L
Linus Torvalds 已提交
581 582 583

		if (tp->urg_data & TCP_URG_VALID)
			mask |= POLLPRI;
584
	} else if (state == TCP_SYN_SENT && inet_sk(sk)->defer_connect) {
W
Wei Wang 已提交
585 586 587 588 589
		/* Active TCP fastopen socket with defer_connect
		 * Return POLLOUT so application can call write()
		 * in order for kernel to generate SYN+data
		 */
		mask |= POLLOUT | POLLWRNORM;
L
Linus Torvalds 已提交
590
	}
T
Tom Marshall 已提交
591 592
	/* This barrier is coupled with smp_wmb() in tcp_reset() */
	smp_rmb();
593
	if (sk->sk_err || !skb_queue_empty(&sk->sk_error_queue))
T
Tom Marshall 已提交
594 595
		mask |= POLLERR;

L
Linus Torvalds 已提交
596 597
	return mask;
}
E
Eric Dumazet 已提交
598
EXPORT_SYMBOL(tcp_poll);
L
Linus Torvalds 已提交
599 600 601 602 603

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
604
	bool slow;
L
Linus Torvalds 已提交
605 606 607 608 609 610

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

E
Eric Dumazet 已提交
611
		slow = lock_sock_fast(sk);
612
		answ = tcp_inq(sk);
E
Eric Dumazet 已提交
613
		unlock_sock_fast(sk, slow);
L
Linus Torvalds 已提交
614 615 616 617 618 619 620 621 622 623 624 625 626
		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;
627 628 629 630 631 632 633 634 635
	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
636 637
	default:
		return -ENOIOCTLCMD;
638
	}
L
Linus Torvalds 已提交
639 640 641

	return put_user(answ, (int __user *)arg);
}
E
Eric Dumazet 已提交
642
EXPORT_SYMBOL(tcp_ioctl);
L
Linus Torvalds 已提交
643 644 645

static inline void tcp_mark_push(struct tcp_sock *tp, struct sk_buff *skb)
{
E
Eric Dumazet 已提交
646
	TCP_SKB_CB(skb)->tcp_flags |= TCPHDR_PSH;
L
Linus Torvalds 已提交
647 648 649
	tp->pushed_seq = tp->write_seq;
}

E
Eric Dumazet 已提交
650
static inline bool forced_push(const struct tcp_sock *tp)
L
Linus Torvalds 已提交
651 652 653 654
{
	return after(tp->write_seq, tp->pushed_seq + (tp->max_window >> 1));
}

655
static void skb_entail(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
656
{
657
	struct tcp_sock *tp = tcp_sk(sk);
658 659 660 661
	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
662
	tcb->tcp_flags = TCPHDR_ACK;
663
	tcb->sacked  = 0;
664
	__skb_header_release(skb);
665
	tcp_add_write_queue_tail(sk, skb);
666 667
	sk->sk_wmem_queued += skb->truesize;
	sk_mem_charge(sk, skb->truesize);
668
	if (tp->nonagle & TCP_NAGLE_PUSH)
669
		tp->nonagle &= ~TCP_NAGLE_PUSH;
670 671

	tcp_slow_start_after_idle_check(sk);
L
Linus Torvalds 已提交
672 673
}

674
static inline void tcp_mark_urg(struct tcp_sock *tp, int flags)
L
Linus Torvalds 已提交
675
{
I
Ilpo Järvinen 已提交
676
	if (flags & MSG_OOB)
L
Linus Torvalds 已提交
677 678 679
		tp->snd_up = tp->write_seq;
}

E
Eric Dumazet 已提交
680
/* If a not yet filled skb is pushed, do not send it if
681
 * we have data packets in Qdisc or NIC queues :
E
Eric Dumazet 已提交
682 683 684 685
 * Because TX completion will happen shortly, it gives a chance
 * to coalesce future sendmsg() payload into this skb, without
 * need for a timer, and with no latency trade off.
 * As packets containing data payload have a bigger truesize
686 687 688
 * than pure acks (dataless) packets, the last checks prevent
 * autocorking if we only have an ACK in Qdisc/NIC queues,
 * or if TX completion was delayed after we processed ACK packet.
E
Eric Dumazet 已提交
689 690 691
 */
static bool tcp_should_autocork(struct sock *sk, struct sk_buff *skb,
				int size_goal)
L
Linus Torvalds 已提交
692
{
E
Eric Dumazet 已提交
693
	return skb->len < size_goal &&
694
	       sock_net(sk)->ipv4.sysctl_tcp_autocorking &&
695
	       skb != tcp_write_queue_head(sk) &&
696
	       refcount_read(&sk->sk_wmem_alloc) > skb->truesize;
E
Eric Dumazet 已提交
697 698 699 700 701 702 703
}

static void tcp_push(struct sock *sk, int flags, int mss_now,
		     int nonagle, int size_goal)
{
	struct tcp_sock *tp = tcp_sk(sk);
	struct sk_buff *skb;
704

E
Eric Dumazet 已提交
705
	skb = tcp_write_queue_tail(sk);
706 707
	if (!skb)
		return;
E
Eric Dumazet 已提交
708 709 710 711 712 713 714 715
	if (!(flags & MSG_MORE) || forced_push(tp))
		tcp_mark_push(tp, skb);

	tcp_mark_urg(tp, flags);

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

		/* avoid atomic op if TSQ_THROTTLED bit is already set */
716
		if (!test_bit(TSQ_THROTTLED, &sk->sk_tsq_flags)) {
E
Eric Dumazet 已提交
717
			NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPAUTOCORKING);
718
			set_bit(TSQ_THROTTLED, &sk->sk_tsq_flags);
E
Eric Dumazet 已提交
719
		}
720 721 722
		/* It is possible TX completion already happened
		 * before we set TSQ_THROTTLED.
		 */
723
		if (refcount_read(&sk->sk_wmem_alloc) > skb->truesize)
724
			return;
L
Linus Torvalds 已提交
725
	}
E
Eric Dumazet 已提交
726 727 728 729 730

	if (flags & MSG_MORE)
		nonagle = TCP_NAGLE_CORK;

	__tcp_push_pending_frames(sk, mss_now, nonagle);
L
Linus Torvalds 已提交
731 732
}

733 734
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
735 736
{
	struct tcp_splice_state *tss = rd_desc->arg.data;
737
	int ret;
J
Jens Axboe 已提交
738

739
	ret = skb_splice_bits(skb, skb->sk, offset, tss->pipe,
A
Al Viro 已提交
740
			      min(rd_desc->count, len), tss->flags);
741 742 743
	if (ret > 0)
		rd_desc->count -= ret;
	return ret;
J
Jens Axboe 已提交
744 745 746 747 748 749 750
}

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,
751
		.count	  = tss->len,
J
Jens Axboe 已提交
752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782
	};

	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;

783
	sock_rps_record_flow(sk);
J
Jens Axboe 已提交
784 785 786 787 788 789 790 791 792 793
	/*
	 * We can't seek on a socket input
	 */
	if (unlikely(*ppos))
		return -ESPIPE;

	ret = spliced = 0;

	lock_sock(sk);

794
	timeo = sock_rcvtimeo(sk, sock->file->f_flags & O_NONBLOCK);
J
Jens Axboe 已提交
795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822
	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;
			}
823 824 825 826 827 828
			/* if __tcp_splice_read() got nothing while we have
			 * an skb in receive queue, we do not want to loop.
			 * This might happen with URG data.
			 */
			if (!skb_queue_empty(&sk->sk_receive_queue))
				break;
829
			sk_wait_data(sk, &timeo, NULL);
J
Jens Axboe 已提交
830 831 832 833 834 835 836 837 838
			if (signal_pending(current)) {
				ret = sock_intr_errno(timeo);
				break;
			}
			continue;
		}
		tss.len -= ret;
		spliced += ret;

839 840
		if (!timeo)
			break;
J
Jens Axboe 已提交
841 842 843 844
		release_sock(sk);
		lock_sock(sk);

		if (sk->sk_err || sk->sk_state == TCP_CLOSE ||
845
		    (sk->sk_shutdown & RCV_SHUTDOWN) ||
J
Jens Axboe 已提交
846 847 848 849 850 851 852 853 854 855 856
		    signal_pending(current))
			break;
	}

	release_sock(sk);

	if (spliced)
		return spliced;

	return ret;
}
E
Eric Dumazet 已提交
857
EXPORT_SYMBOL(tcp_splice_read);
J
Jens Axboe 已提交
858

859 860
struct sk_buff *sk_stream_alloc_skb(struct sock *sk, int size, gfp_t gfp,
				    bool force_schedule)
861 862 863 864 865 866
{
	struct sk_buff *skb;

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

867 868 869
	if (unlikely(tcp_under_memory_pressure(sk)))
		sk_mem_reclaim_partial(sk);

870
	skb = alloc_skb_fclone(size + sk->sk_prot->max_header, gfp);
871
	if (likely(skb)) {
872
		bool mem_scheduled;
873

874 875
		if (force_schedule) {
			mem_scheduled = true;
876 877
			sk_forced_mem_schedule(sk, skb->truesize);
		} else {
878
			mem_scheduled = sk_wmem_schedule(sk, skb->truesize);
879
		}
880
		if (likely(mem_scheduled)) {
881
			skb_reserve(skb, sk->sk_prot->max_header);
882 883 884 885
			/*
			 * Make sure that we have exactly size bytes
			 * available to the caller, no more, no less.
			 */
E
Eric Dumazet 已提交
886
			skb->reserved_tailroom = skb->end - skb->tail - size;
887
			INIT_LIST_HEAD(&skb->tcp_tsorted_anchor);
888 889 890 891
			return skb;
		}
		__kfree_skb(skb);
	} else {
892
		sk->sk_prot->enter_memory_pressure(sk);
893 894 895 896 897
		sk_stream_moderate_sndbuf(sk);
	}
	return NULL;
}

I
Ilpo Järvinen 已提交
898 899 900 901
static unsigned int tcp_xmit_size_goal(struct sock *sk, u32 mss_now,
				       int large_allowed)
{
	struct tcp_sock *tp = tcp_sk(sk);
902
	u32 new_size_goal, size_goal;
E
Eric Dumazet 已提交
903 904 905 906

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

907 908
	/* Note : tcp_tso_autosize() will eventually split this later */
	new_size_goal = sk->sk_gso_max_size - 1 - MAX_TCP_HEADER;
E
Eric Dumazet 已提交
909 910 911 912 913 914 915 916 917
	new_size_goal = tcp_bound_to_half_wnd(tp, new_size_goal);

	/* We try hard to avoid divides here */
	size_goal = tp->gso_segs * mss_now;
	if (unlikely(new_size_goal < size_goal ||
		     new_size_goal >= size_goal + mss_now)) {
		tp->gso_segs = min_t(u16, new_size_goal / mss_now,
				     sk->sk_gso_max_segs);
		size_goal = tp->gso_segs * mss_now;
I
Ilpo Järvinen 已提交
918 919
	}

E
Eric Dumazet 已提交
920
	return max(size_goal, mss_now);
I
Ilpo Järvinen 已提交
921 922 923 924 925 926 927 928 929 930 931 932
}

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;
}

933 934
ssize_t do_tcp_sendpages(struct sock *sk, struct page *page, int offset,
			 size_t size, int flags)
L
Linus Torvalds 已提交
935 936
{
	struct tcp_sock *tp = tcp_sk(sk);
937
	int mss_now, size_goal;
L
Linus Torvalds 已提交
938 939 940 941
	int err;
	ssize_t copied;
	long timeo = sock_sndtimeo(sk, flags & MSG_DONTWAIT);

942 943 944 945 946 947
	/* 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)) {
948 949
		err = sk_stream_wait_connect(sk, &timeo);
		if (err != 0)
L
Linus Torvalds 已提交
950
			goto out_err;
951
	}
L
Linus Torvalds 已提交
952

953
	sk_clear_bit(SOCKWQ_ASYNC_NOSPACE, sk);
L
Linus Torvalds 已提交
954

I
Ilpo Järvinen 已提交
955
	mss_now = tcp_send_mss(sk, &size_goal, flags);
L
Linus Torvalds 已提交
956 957 958 959
	copied = 0;

	err = -EPIPE;
	if (sk->sk_err || (sk->sk_shutdown & SEND_SHUTDOWN))
960
		goto out_err;
L
Linus Torvalds 已提交
961

962
	while (size > 0) {
963
		struct sk_buff *skb = tcp_write_queue_tail(sk);
964 965
		int copy, i;
		bool can_coalesce;
L
Linus Torvalds 已提交
966

967
		if (!skb || (copy = size_goal - skb->len) <= 0 ||
968
		    !tcp_skb_can_collapse_to(skb)) {
L
Linus Torvalds 已提交
969 970 971 972
new_segment:
			if (!sk_stream_memory_free(sk))
				goto wait_for_sndbuf;

973
			skb = sk_stream_alloc_skb(sk, 0, sk->sk_allocation,
974
					tcp_rtx_and_write_queues_empty(sk));
L
Linus Torvalds 已提交
975 976 977
			if (!skb)
				goto wait_for_memory;

978
			skb_entail(sk, skb);
979
			copy = size_goal;
L
Linus Torvalds 已提交
980 981 982 983 984 985 986
		}

		if (copy > size)
			copy = size;

		i = skb_shinfo(skb)->nr_frags;
		can_coalesce = skb_can_coalesce(skb, i, page, offset);
H
Hans Westgaard Ry 已提交
987
		if (!can_coalesce && i >= sysctl_max_skb_frags) {
L
Linus Torvalds 已提交
988 989 990
			tcp_mark_push(tp, skb);
			goto new_segment;
		}
991
		if (!sk_wmem_schedule(sk, copy))
L
Linus Torvalds 已提交
992
			goto wait_for_memory;
993

L
Linus Torvalds 已提交
994
		if (can_coalesce) {
E
Eric Dumazet 已提交
995
			skb_frag_size_add(&skb_shinfo(skb)->frags[i - 1], copy);
L
Linus Torvalds 已提交
996 997 998 999
		} else {
			get_page(page);
			skb_fill_page_desc(skb, i, page, offset, copy);
		}
1000
		skb_shinfo(skb)->tx_flags |= SKBTX_SHARED_FRAG;
1001

L
Linus Torvalds 已提交
1002 1003 1004 1005
		skb->len += copy;
		skb->data_len += copy;
		skb->truesize += copy;
		sk->sk_wmem_queued += copy;
1006
		sk_mem_charge(sk, copy);
1007
		skb->ip_summed = CHECKSUM_PARTIAL;
L
Linus Torvalds 已提交
1008 1009
		tp->write_seq += copy;
		TCP_SKB_CB(skb)->end_seq += copy;
1010
		tcp_skb_pcount_set(skb, 0);
L
Linus Torvalds 已提交
1011 1012

		if (!copied)
E
Eric Dumazet 已提交
1013
			TCP_SKB_CB(skb)->tcp_flags &= ~TCPHDR_PSH;
L
Linus Torvalds 已提交
1014 1015

		copied += copy;
1016
		offset += copy;
1017
		size -= copy;
1018
		if (!size)
L
Linus Torvalds 已提交
1019 1020
			goto out;

1021
		if (skb->len < size_goal || (flags & MSG_OOB))
L
Linus Torvalds 已提交
1022 1023 1024 1025
			continue;

		if (forced_push(tp)) {
			tcp_mark_push(tp, skb);
1026
			__tcp_push_pending_frames(sk, mss_now, TCP_NAGLE_PUSH);
1027
		} else if (skb == tcp_send_head(sk))
L
Linus Torvalds 已提交
1028 1029 1030 1031 1032 1033
			tcp_push_one(sk, mss_now);
		continue;

wait_for_sndbuf:
		set_bit(SOCK_NOSPACE, &sk->sk_socket->flags);
wait_for_memory:
E
Eric Dumazet 已提交
1034 1035
		tcp_push(sk, flags & ~MSG_MORE, mss_now,
			 TCP_NAGLE_PUSH, size_goal);
L
Linus Torvalds 已提交
1036

1037 1038
		err = sk_stream_wait_memory(sk, &timeo);
		if (err != 0)
L
Linus Torvalds 已提交
1039 1040
			goto do_error;

I
Ilpo Järvinen 已提交
1041
		mss_now = tcp_send_mss(sk, &size_goal, flags);
L
Linus Torvalds 已提交
1042 1043 1044
	}

out:
1045
	if (copied) {
E
Eric Dumazet 已提交
1046
		tcp_tx_timestamp(sk, sk->sk_tsflags);
1047 1048 1049
		if (!(flags & MSG_SENDPAGE_NOTLAST))
			tcp_push(sk, flags, mss_now, tp->nonagle, size_goal);
	}
L
Linus Torvalds 已提交
1050 1051 1052 1053 1054 1055
	return copied;

do_error:
	if (copied)
		goto out;
out_err:
1056
	/* make sure we wake any epoll edge trigger waiter */
1057 1058
	if (unlikely(skb_queue_len(&sk->sk_write_queue) == 0 &&
		     err == -EAGAIN)) {
1059
		sk->sk_write_space(sk);
1060 1061
		tcp_chrono_stop(sk, TCP_CHRONO_SNDBUF_LIMITED);
	}
L
Linus Torvalds 已提交
1062 1063
	return sk_stream_error(sk, flags, err);
}
1064
EXPORT_SYMBOL_GPL(do_tcp_sendpages);
L
Linus Torvalds 已提交
1065

1066 1067
int tcp_sendpage_locked(struct sock *sk, struct page *page, int offset,
			size_t size, int flags)
L
Linus Torvalds 已提交
1068 1069
{
	if (!(sk->sk_route_caps & NETIF_F_SG) ||
1070
	    !sk_check_csum_caps(sk))
1071
		return sock_no_sendpage_locked(sk, page, offset, size, flags);
L
Linus Torvalds 已提交
1072

1073 1074
	tcp_rate_check_app_limited(sk);  /* is sending application-limited? */

1075 1076
	return do_tcp_sendpages(sk, page, offset, size, flags);
}
1077
EXPORT_SYMBOL_GPL(tcp_sendpage_locked);
1078 1079 1080 1081 1082 1083 1084 1085

int tcp_sendpage(struct sock *sk, struct page *page, int offset,
		 size_t size, int flags)
{
	int ret;

	lock_sock(sk);
	ret = tcp_sendpage_locked(sk, page, offset, size, flags);
L
Linus Torvalds 已提交
1086
	release_sock(sk);
1087 1088

	return ret;
L
Linus Torvalds 已提交
1089
}
E
Eric Dumazet 已提交
1090
EXPORT_SYMBOL(tcp_sendpage);
L
Linus Torvalds 已提交
1091

1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109
/* Do not bother using a page frag for very small frames.
 * But use this heuristic only for the first skb in write queue.
 *
 * Having no payload in skb->head allows better SACK shifting
 * in tcp_shift_skb_data(), reducing sack/rack overhead, because
 * write queue has less skbs.
 * Each skb can hold up to MAX_SKB_FRAGS * 32Kbytes, or ~0.5 MB.
 * This also speeds up tso_fragment(), since it wont fallback
 * to tcp_fragment().
 */
static int linear_payload_sz(bool first_skb)
{
	if (first_skb)
		return SKB_WITH_OVERHEAD(2048 - MAX_TCP_HEADER);
	return 0;
}

static int select_size(const struct sock *sk, bool sg, bool first_skb)
L
Linus Torvalds 已提交
1110
{
1111
	const struct tcp_sock *tp = tcp_sk(sk);
1112
	int tmp = tp->mss_cache;
L
Linus Torvalds 已提交
1113

K
Krishna Kumar 已提交
1114
	if (sg) {
1115
		if (sk_can_gso(sk)) {
1116
			tmp = linear_payload_sz(first_skb);
1117
		} else {
1118 1119 1120 1121 1122 1123 1124
			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
1125 1126 1127 1128

	return tmp;
}

1129 1130
void tcp_free_fastopen_req(struct tcp_sock *tp)
{
1131
	if (tp->fastopen_req) {
1132 1133 1134 1135 1136
		kfree(tp->fastopen_req);
		tp->fastopen_req = NULL;
	}
}

1137 1138
static int tcp_sendmsg_fastopen(struct sock *sk, struct msghdr *msg,
				int *copied, size_t size)
1139 1140
{
	struct tcp_sock *tp = tcp_sk(sk);
W
Wei Wang 已提交
1141
	struct inet_sock *inet = inet_sk(sk);
1142
	struct sockaddr *uaddr = msg->msg_name;
1143 1144
	int err, flags;

1145
	if (!(sock_net(sk)->ipv4.sysctl_tcp_fastopen & TFO_CLIENT_ENABLE) ||
1146 1147
	    (uaddr && msg->msg_namelen >= sizeof(uaddr->sa_family) &&
	     uaddr->sa_family == AF_UNSPEC))
1148
		return -EOPNOTSUPP;
1149
	if (tp->fastopen_req)
1150 1151 1152 1153
		return -EALREADY; /* Another Fast Open is in progress */

	tp->fastopen_req = kzalloc(sizeof(struct tcp_fastopen_request),
				   sk->sk_allocation);
1154
	if (unlikely(!tp->fastopen_req))
1155 1156
		return -ENOBUFS;
	tp->fastopen_req->data = msg;
1157
	tp->fastopen_req->size = size;
1158

W
Wei Wang 已提交
1159 1160 1161 1162 1163 1164 1165 1166 1167
	if (inet->defer_connect) {
		err = tcp_connect(sk);
		/* Same failure procedure as in tcp_v4/6_connect */
		if (err) {
			tcp_set_state(sk, TCP_CLOSE);
			inet->inet_dport = 0;
			sk->sk_route_caps = 0;
		}
	}
1168
	flags = (msg->msg_flags & MSG_DONTWAIT) ? O_NONBLOCK : 0;
1169
	err = __inet_stream_connect(sk->sk_socket, uaddr,
1170
				    msg->msg_namelen, flags, 1);
1171 1172 1173 1174 1175 1176 1177 1178
	/* fastopen_req could already be freed in __inet_stream_connect
	 * if the connection times out or gets rst
	 */
	if (tp->fastopen_req) {
		*copied = tp->fastopen_req->copied;
		tcp_free_fastopen_req(tp);
		inet->defer_connect = 0;
	}
1179 1180 1181
	return err;
}

1182
int tcp_sendmsg_locked(struct sock *sk, struct msghdr *msg, size_t size)
L
Linus Torvalds 已提交
1183 1184
{
	struct tcp_sock *tp = tcp_sk(sk);
W
Willem de Bruijn 已提交
1185
	struct ubuf_info *uarg = NULL;
L
Linus Torvalds 已提交
1186
	struct sk_buff *skb;
1187
	struct sockcm_cookie sockc;
1188 1189
	int flags, err, copied = 0;
	int mss_now = 0, size_goal, copied_syn = 0;
1190
	bool process_backlog = false;
1191
	bool sg;
L
Linus Torvalds 已提交
1192 1193 1194
	long timeo;

	flags = msg->msg_flags;
W
Willem de Bruijn 已提交
1195 1196 1197 1198 1199 1200 1201

	if (flags & MSG_ZEROCOPY && size) {
		if (sk->sk_state != TCP_ESTABLISHED) {
			err = -EINVAL;
			goto out_err;
		}

1202
		skb = tcp_write_queue_tail(sk);
W
Willem de Bruijn 已提交
1203 1204 1205 1206 1207 1208 1209 1210 1211 1212
		uarg = sock_zerocopy_realloc(sk, size, skb_zcopy(skb));
		if (!uarg) {
			err = -ENOBUFS;
			goto out_err;
		}

		if (!(sk_check_csum_caps(sk) && sk->sk_route_caps & NETIF_F_SG))
			uarg->zerocopy = 0;
	}

W
Wei Wang 已提交
1213
	if (unlikely(flags & MSG_FASTOPEN || inet_sk(sk)->defer_connect)) {
1214
		err = tcp_sendmsg_fastopen(sk, msg, &copied_syn, size);
1215 1216 1217 1218 1219 1220
		if (err == -EINPROGRESS && copied_syn > 0)
			goto out;
		else if (err)
			goto out_err;
	}

L
Linus Torvalds 已提交
1221 1222
	timeo = sock_sndtimeo(sk, flags & MSG_DONTWAIT);

1223 1224
	tcp_rate_check_app_limited(sk);  /* is sending application-limited? */

1225 1226 1227 1228 1229 1230
	/* 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)) {
1231 1232
		err = sk_stream_wait_connect(sk, &timeo);
		if (err != 0)
1233
			goto do_error;
1234
	}
L
Linus Torvalds 已提交
1235

P
Pavel Emelyanov 已提交
1236 1237 1238
	if (unlikely(tp->repair)) {
		if (tp->repair_queue == TCP_RECV_QUEUE) {
			copied = tcp_send_rcvq(sk, msg, size);
1239
			goto out_nopush;
P
Pavel Emelyanov 已提交
1240 1241 1242 1243 1244 1245 1246 1247 1248
		}

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

		/* 'common' sending to sendq */
	}

1249 1250 1251 1252 1253 1254 1255 1256 1257
	sockc.tsflags = sk->sk_tsflags;
	if (msg->msg_controllen) {
		err = sock_cmsg_send(sk, msg, &sockc);
		if (unlikely(err)) {
			err = -EINVAL;
			goto out_err;
		}
	}

L
Linus Torvalds 已提交
1258
	/* This should be in poll */
1259
	sk_clear_bit(SOCKWQ_ASYNC_NOSPACE, sk);
L
Linus Torvalds 已提交
1260 1261 1262 1263

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

1264 1265 1266
restart:
	mss_now = tcp_send_mss(sk, &size_goal, flags);

L
Linus Torvalds 已提交
1267 1268
	err = -EPIPE;
	if (sk->sk_err || (sk->sk_shutdown & SEND_SHUTDOWN))
1269
		goto do_error;
L
Linus Torvalds 已提交
1270

1271
	sg = !!(sk->sk_route_caps & NETIF_F_SG);
K
Krishna Kumar 已提交
1272

A
Al Viro 已提交
1273
	while (msg_data_left(msg)) {
1274 1275
		int copy = 0;
		int max = size_goal;
L
Linus Torvalds 已提交
1276

1277
		skb = tcp_write_queue_tail(sk);
1278
		if (skb) {
1279 1280 1281
			if (skb->ip_summed == CHECKSUM_NONE)
				max = mss_now;
			copy = max - skb->len;
1282
		}
L
Linus Torvalds 已提交
1283

1284
		if (copy <= 0 || !tcp_skb_can_collapse_to(skb)) {
1285 1286
			bool first_skb;

L
Linus Torvalds 已提交
1287
new_segment:
1288 1289 1290 1291 1292
			/* Allocate new segment. If the interface is SG,
			 * allocate skb fitting to single page.
			 */
			if (!sk_stream_memory_free(sk))
				goto wait_for_sndbuf;
L
Linus Torvalds 已提交
1293

1294 1295
			if (process_backlog && sk_flush_backlog(sk)) {
				process_backlog = false;
1296
				goto restart;
1297
			}
1298
			first_skb = tcp_rtx_and_write_queues_empty(sk);
1299
			skb = sk_stream_alloc_skb(sk,
1300
						  select_size(sk, sg, first_skb),
1301
						  sk->sk_allocation,
1302
						  first_skb);
1303 1304
			if (!skb)
				goto wait_for_memory;
L
Linus Torvalds 已提交
1305

1306
			process_backlog = true;
1307 1308 1309
			/*
			 * Check whether we can use HW checksum.
			 */
1310
			if (sk_check_csum_caps(sk))
1311
				skb->ip_summed = CHECKSUM_PARTIAL;
L
Linus Torvalds 已提交
1312

1313 1314 1315
			skb_entail(sk, skb);
			copy = size_goal;
			max = size_goal;
1316

1317 1318 1319 1320 1321 1322 1323
			/* All packets are restored as if they have
			 * already been sent. skb_mstamp isn't set to
			 * avoid wrong rtt estimation.
			 */
			if (tp->repair)
				TCP_SKB_CB(skb)->sacked |= TCPCB_REPAIRED;
		}
L
Linus Torvalds 已提交
1324

1325
		/* Try to append data to the end of skb. */
A
Al Viro 已提交
1326 1327
		if (copy > msg_data_left(msg))
			copy = msg_data_left(msg);
1328 1329 1330 1331 1332 1333 1334 1335

		/* Where to copy to? */
		if (skb_availroom(skb) > 0) {
			/* We have some space in skb head. Superb! */
			copy = min_t(int, copy, skb_availroom(skb));
			err = skb_add_data_nocache(sk, skb, &msg->msg_iter, copy);
			if (err)
				goto do_fault;
W
Willem de Bruijn 已提交
1336
		} else if (!uarg || !uarg->zerocopy) {
1337 1338 1339 1340 1341 1342
			bool merge = true;
			int i = skb_shinfo(skb)->nr_frags;
			struct page_frag *pfrag = sk_page_frag(sk);

			if (!sk_page_frag_refill(sk, pfrag))
				goto wait_for_memory;
1343

1344 1345
			if (!skb_can_coalesce(skb, i, pfrag->page,
					      pfrag->offset)) {
1346
				if (i >= sysctl_max_skb_frags || !sg) {
1347 1348
					tcp_mark_push(tp, skb);
					goto new_segment;
L
Linus Torvalds 已提交
1349
				}
1350
				merge = false;
L
Linus Torvalds 已提交
1351 1352
			}

1353 1354 1355 1356
			copy = min_t(int, copy, pfrag->size - pfrag->offset);

			if (!sk_wmem_schedule(sk, copy))
				goto wait_for_memory;
L
Linus Torvalds 已提交
1357

1358 1359 1360 1361 1362 1363
			err = skb_copy_to_page_nocache(sk, &msg->msg_iter, skb,
						       pfrag->page,
						       pfrag->offset,
						       copy);
			if (err)
				goto do_error;
L
Linus Torvalds 已提交
1364

1365 1366 1367 1368 1369 1370
			/* Update the skb. */
			if (merge) {
				skb_frag_size_add(&skb_shinfo(skb)->frags[i - 1], copy);
			} else {
				skb_fill_page_desc(skb, i, pfrag->page,
						   pfrag->offset, copy);
1371
				page_ref_inc(pfrag->page);
1372
			}
1373
			pfrag->offset += copy;
W
Willem de Bruijn 已提交
1374 1375 1376 1377 1378 1379 1380
		} else {
			err = skb_zerocopy_iter_stream(sk, skb, msg, copy, uarg);
			if (err == -EMSGSIZE || err == -EEXIST)
				goto new_segment;
			if (err < 0)
				goto do_error;
			copy = err;
1381 1382 1383 1384 1385 1386 1387 1388
		}

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

		tp->write_seq += copy;
		TCP_SKB_CB(skb)->end_seq += copy;
		tcp_skb_pcount_set(skb, 0);
L
Linus Torvalds 已提交
1389

1390
		copied += copy;
A
Al Viro 已提交
1391
		if (!msg_data_left(msg)) {
1392 1393
			if (unlikely(flags & MSG_EOR))
				TCP_SKB_CB(skb)->eor = 1;
1394 1395
			goto out;
		}
L
Linus Torvalds 已提交
1396

1397
		if (skb->len < max || (flags & MSG_OOB) || unlikely(tp->repair))
L
Linus Torvalds 已提交
1398 1399
			continue;

1400 1401 1402 1403 1404 1405 1406
		if (forced_push(tp)) {
			tcp_mark_push(tp, skb);
			__tcp_push_pending_frames(sk, mss_now, TCP_NAGLE_PUSH);
		} else if (skb == tcp_send_head(sk))
			tcp_push_one(sk, mss_now);
		continue;

L
Linus Torvalds 已提交
1407
wait_for_sndbuf:
1408
		set_bit(SOCK_NOSPACE, &sk->sk_socket->flags);
L
Linus Torvalds 已提交
1409
wait_for_memory:
1410 1411 1412
		if (copied)
			tcp_push(sk, flags & ~MSG_MORE, mss_now,
				 TCP_NAGLE_PUSH, size_goal);
L
Linus Torvalds 已提交
1413

1414 1415
		err = sk_stream_wait_memory(sk, &timeo);
		if (err != 0)
1416
			goto do_error;
L
Linus Torvalds 已提交
1417

1418
		mss_now = tcp_send_mss(sk, &size_goal, flags);
L
Linus Torvalds 已提交
1419 1420 1421
	}

out:
1422
	if (copied) {
E
Eric Dumazet 已提交
1423
		tcp_tx_timestamp(sk, sockc.tsflags);
E
Eric Dumazet 已提交
1424
		tcp_push(sk, flags, mss_now, tp->nonagle, size_goal);
1425
	}
1426
out_nopush:
W
Willem de Bruijn 已提交
1427
	sock_zerocopy_put(uarg);
1428
	return copied + copied_syn;
L
Linus Torvalds 已提交
1429 1430 1431

do_fault:
	if (!skb->len) {
1432 1433 1434 1435 1436
		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);
1437
		sk_wmem_free_skb(sk, skb);
L
Linus Torvalds 已提交
1438 1439 1440
	}

do_error:
1441
	if (copied + copied_syn)
L
Linus Torvalds 已提交
1442 1443
		goto out;
out_err:
W
Willem de Bruijn 已提交
1444
	sock_zerocopy_put_abort(uarg);
L
Linus Torvalds 已提交
1445
	err = sk_stream_error(sk, flags, err);
1446
	/* make sure we wake any epoll edge trigger waiter */
1447 1448
	if (unlikely(skb_queue_len(&sk->sk_write_queue) == 0 &&
		     err == -EAGAIN)) {
1449
		sk->sk_write_space(sk);
1450 1451
		tcp_chrono_stop(sk, TCP_CHRONO_SNDBUF_LIMITED);
	}
L
Linus Torvalds 已提交
1452 1453
	return err;
}
1454
EXPORT_SYMBOL_GPL(tcp_sendmsg_locked);
1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465

int tcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t size)
{
	int ret;

	lock_sock(sk);
	ret = tcp_sendmsg_locked(sk, msg, size);
	release_sock(sk);

	return ret;
}
E
Eric Dumazet 已提交
1466
EXPORT_SYMBOL(tcp_sendmsg);
L
Linus Torvalds 已提交
1467 1468 1469 1470 1471 1472

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

1473
static int tcp_recv_urg(struct sock *sk, struct msghdr *msg, int len, int flags)
L
Linus Torvalds 已提交
1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496
{
	struct tcp_sock *tp = tcp_sk(sk);

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

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

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

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

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

		if (len > 0) {
			if (!(flags & MSG_TRUNC))
A
Al Viro 已提交
1497
				err = memcpy_to_msg(msg, &c, 1);
L
Linus Torvalds 已提交
1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516
			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
1517 1518 1519 1520 1521 1522 1523
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 */

1524 1525 1526 1527 1528 1529 1530
	skb_rbtree_walk(skb, &sk->tcp_rtx_queue) {
		err = skb_copy_datagram_msg(skb, 0, msg, skb->len);
		if (err)
			return err;
		copied += skb->len;
	}

P
Pavel Emelyanov 已提交
1531
	skb_queue_walk(&sk->sk_write_queue, skb) {
1532
		err = skb_copy_datagram_msg(skb, 0, msg, skb->len);
P
Pavel Emelyanov 已提交
1533 1534 1535 1536 1537 1538 1539 1540 1541
		if (err)
			break;

		copied += skb->len;
	}

	return err ?: copied;
}

L
Linus Torvalds 已提交
1542 1543 1544 1545 1546 1547
/* 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.
 */
1548
static void tcp_cleanup_rbuf(struct sock *sk, int copied)
L
Linus Torvalds 已提交
1549 1550
{
	struct tcp_sock *tp = tcp_sk(sk);
E
Eric Dumazet 已提交
1551
	bool time_to_ack = false;
L
Linus Torvalds 已提交
1552 1553 1554

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

1555
	WARN(skb && !before(tp->copied_seq, TCP_SKB_CB(skb)->end_seq),
J
Joe Perches 已提交
1556
	     "cleanup rbuf bug: copied %X seq %X rcvnxt %X\n",
1557
	     tp->copied_seq, TCP_SKB_CB(skb)->end_seq, tp->rcv_nxt);
L
Linus Torvalds 已提交
1558

1559 1560
	if (inet_csk_ack_scheduled(sk)) {
		const struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
1561 1562
		   /* Delayed ACKs frequently hit locked sockets during bulk
		    * receive. */
1563
		if (icsk->icsk_ack.blocked ||
L
Linus Torvalds 已提交
1564
		    /* Once-per-two-segments ACK was not sent by tcp_input.c */
1565
		    tp->rcv_nxt - tp->rcv_wup > icsk->icsk_ack.rcv_mss ||
L
Linus Torvalds 已提交
1566 1567 1568 1569 1570 1571
		    /*
		     * 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.
		     */
1572 1573 1574 1575 1576
		    (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
1577
			time_to_ack = true;
L
Linus Torvalds 已提交
1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598
	}

	/* 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
1599
				time_to_ack = true;
L
Linus Torvalds 已提交
1600 1601 1602 1603 1604 1605
		}
	}
	if (time_to_ack)
		tcp_send_ack(sk);
}

1606
static struct sk_buff *tcp_recv_skb(struct sock *sk, u32 seq, u32 *off)
L
Linus Torvalds 已提交
1607 1608 1609 1610
{
	struct sk_buff *skb;
	u32 offset;

1611
	while ((skb = skb_peek(&sk->sk_receive_queue)) != NULL) {
L
Linus Torvalds 已提交
1612
		offset = seq - TCP_SKB_CB(skb)->seq;
1613 1614
		if (unlikely(TCP_SKB_CB(skb)->tcp_flags & TCPHDR_SYN)) {
			pr_err_once("%s: found a SYN, please report !\n", __func__);
L
Linus Torvalds 已提交
1615
			offset--;
1616
		}
1617
		if (offset < skb->len || (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN)) {
L
Linus Torvalds 已提交
1618 1619 1620
			*off = offset;
			return skb;
		}
1621 1622 1623 1624
		/* This looks weird, but this can happen if TCP collapsing
		 * splitted a fat GRO packet, while we released socket lock
		 * in skb_splice_bits()
		 */
D
Dan Williams 已提交
1625
		sk_eat_skb(sk, skb);
L
Linus Torvalds 已提交
1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653
	}
	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) {
1654 1655
			int used;
			size_t len;
L
Linus Torvalds 已提交
1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666

			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);
1667
			if (used <= 0) {
1668 1669 1670 1671
				if (!copied)
					copied = used;
				break;
			} else if (used <= len) {
L
Linus Torvalds 已提交
1672 1673 1674 1675
				seq += used;
				copied += used;
				offset += used;
			}
1676
			/* If recv_actor drops the lock (e.g. TCP splice
1677 1678 1679 1680
			 * receive) the skb pointer might be invalid when
			 * getting here: tcp_collapse might have deleted it
			 * while aggregating skbs from the socket queue.
			 */
1681 1682
			skb = tcp_recv_skb(sk, seq - 1, &offset);
			if (!skb)
L
Linus Torvalds 已提交
1683
				break;
1684 1685 1686 1687 1688
			/* 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
1689
		}
1690
		if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN) {
D
Dan Williams 已提交
1691
			sk_eat_skb(sk, skb);
L
Linus Torvalds 已提交
1692 1693 1694
			++seq;
			break;
		}
D
Dan Williams 已提交
1695
		sk_eat_skb(sk, skb);
L
Linus Torvalds 已提交
1696 1697
		if (!desc->count)
			break;
1698
		tp->copied_seq = seq;
L
Linus Torvalds 已提交
1699 1700 1701 1702 1703 1704
	}
	tp->copied_seq = seq;

	tcp_rcv_space_adjust(sk);

	/* Clean up data we have read: This will do ACK frames. */
1705 1706
	if (copied > 0) {
		tcp_recv_skb(sk, seq, &offset);
1707
		tcp_cleanup_rbuf(sk, copied);
1708
	}
L
Linus Torvalds 已提交
1709 1710
	return copied;
}
E
Eric Dumazet 已提交
1711
EXPORT_SYMBOL(tcp_read_sock);
L
Linus Torvalds 已提交
1712

1713 1714 1715 1716 1717 1718
int tcp_peek_len(struct socket *sock)
{
	return tcp_inq(sock->sk);
}
EXPORT_SYMBOL(tcp_peek_len);

1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773
static void tcp_update_recv_tstamps(struct sk_buff *skb,
				    struct scm_timestamping *tss)
{
	if (skb->tstamp)
		tss->ts[0] = ktime_to_timespec(skb->tstamp);
	else
		tss->ts[0] = (struct timespec) {0};

	if (skb_hwtstamps(skb)->hwtstamp)
		tss->ts[2] = ktime_to_timespec(skb_hwtstamps(skb)->hwtstamp);
	else
		tss->ts[2] = (struct timespec) {0};
}

/* Similar to __sock_recv_timestamp, but does not require an skb */
void tcp_recv_timestamp(struct msghdr *msg, const struct sock *sk,
			struct scm_timestamping *tss)
{
	struct timeval tv;
	bool has_timestamping = false;

	if (tss->ts[0].tv_sec || tss->ts[0].tv_nsec) {
		if (sock_flag(sk, SOCK_RCVTSTAMP)) {
			if (sock_flag(sk, SOCK_RCVTSTAMPNS)) {
				put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMPNS,
					 sizeof(tss->ts[0]), &tss->ts[0]);
			} else {
				tv.tv_sec = tss->ts[0].tv_sec;
				tv.tv_usec = tss->ts[0].tv_nsec / 1000;

				put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMP,
					 sizeof(tv), &tv);
			}
		}

		if (sk->sk_tsflags & SOF_TIMESTAMPING_SOFTWARE)
			has_timestamping = true;
		else
			tss->ts[0] = (struct timespec) {0};
	}

	if (tss->ts[2].tv_sec || tss->ts[2].tv_nsec) {
		if (sk->sk_tsflags & SOF_TIMESTAMPING_RAW_HARDWARE)
			has_timestamping = true;
		else
			tss->ts[2] = (struct timespec) {0};
	}

	if (has_timestamping) {
		tss->ts[1] = (struct timespec) {0};
		put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMPING,
			 sizeof(*tss), tss);
	}
}

L
Linus Torvalds 已提交
1774 1775 1776 1777 1778 1779 1780 1781
/*
 *	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.
 */

1782 1783
int tcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, int nonblock,
		int flags, int *addr_len)
L
Linus Torvalds 已提交
1784 1785 1786 1787 1788 1789 1790 1791 1792
{
	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;
1793
	struct sk_buff *skb, *last;
I
Ilpo Järvinen 已提交
1794
	u32 urg_hole = 0;
1795 1796
	struct scm_timestamping tss;
	bool has_tss = false;
L
Linus Torvalds 已提交
1797

1798
	if (unlikely(flags & MSG_ERRQUEUE))
1799
		return inet_recv_error(sk, msg, len, addr_len);
1800

1801 1802 1803
	if (sk_can_busy_loop(sk) && skb_queue_empty(&sk->sk_receive_queue) &&
	    (sk->sk_state == TCP_ESTABLISHED))
		sk_busy_loop(sk, nonblock);
1804

L
Linus Torvalds 已提交
1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816
	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
1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831
	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
1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854
	seq = &tp->copied_seq;
	if (flags & MSG_PEEK) {
		peek_seq = tp->copied_seq;
		seq = &peek_seq;
	}

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

	do {
		u32 offset;

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

		/* Next get a buffer. */

1855
		last = skb_peek_tail(&sk->sk_receive_queue);
1856
		skb_queue_walk(&sk->sk_receive_queue, skb) {
1857
			last = skb;
L
Linus Torvalds 已提交
1858 1859 1860
			/* Now that we have two receive queues this
			 * shouldn't happen.
			 */
1861
			if (WARN(before(*seq, TCP_SKB_CB(skb)->seq),
J
Joe Perches 已提交
1862 1863 1864
				 "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
1865
				break;
1866

L
Linus Torvalds 已提交
1867
			offset = *seq - TCP_SKB_CB(skb)->seq;
1868 1869
			if (unlikely(TCP_SKB_CB(skb)->tcp_flags & TCPHDR_SYN)) {
				pr_err_once("%s: found a SYN, please report !\n", __func__);
L
Linus Torvalds 已提交
1870
				offset--;
1871
			}
L
Linus Torvalds 已提交
1872 1873
			if (offset < skb->len)
				goto found_ok_skb;
1874
			if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN)
L
Linus Torvalds 已提交
1875
				goto found_fin_ok;
J
Joe Perches 已提交
1876 1877 1878
			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);
1879
		}
L
Linus Torvalds 已提交
1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890

		/* 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 ||
1891
			    signal_pending(current))
L
Linus Torvalds 已提交
1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926
				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;
			}
		}

1927
		tcp_cleanup_rbuf(sk, copied);
L
Linus Torvalds 已提交
1928 1929 1930 1931 1932

		if (copied >= target) {
			/* Do not sleep, just process backlog. */
			release_sock(sk);
			lock_sock(sk);
1933 1934 1935
		} else {
			sk_wait_data(sk, &timeo, last);
		}
L
Linus Torvalds 已提交
1936

I
Ilpo Järvinen 已提交
1937 1938
		if ((flags & MSG_PEEK) &&
		    (peek_seq - copied - urg_hole != tp->copied_seq)) {
1939 1940 1941
			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
1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958
			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
1959
						urg_hole++;
L
Linus Torvalds 已提交
1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970
						offset++;
						used--;
						if (!used)
							goto skip_copy;
					}
				} else
					used = urg_offset;
			}
		}

		if (!(flags & MSG_TRUNC)) {
1971
			err = skb_copy_datagram_msg(skb, offset, msg, used);
D
Dan Williams 已提交
1972 1973 1974 1975 1976
			if (err) {
				/* Exception. Bailout! */
				if (!copied)
					copied = -EFAULT;
				break;
L
Linus Torvalds 已提交
1977 1978 1979 1980 1981 1982 1983 1984 1985 1986
			}
		}

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

		tcp_rcv_space_adjust(sk);

skip_copy:
1987
		if (tp->urg_data && after(tp->copied_seq, tp->urg_seq)) {
L
Linus Torvalds 已提交
1988
			tp->urg_data = 0;
1989 1990
			tcp_fast_path_check(sk);
		}
L
Linus Torvalds 已提交
1991 1992 1993
		if (used + offset < skb->len)
			continue;

1994 1995 1996 1997
		if (TCP_SKB_CB(skb)->has_rxtstamp) {
			tcp_update_recv_tstamps(skb, &tss);
			has_tss = true;
		}
1998
		if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN)
L
Linus Torvalds 已提交
1999
			goto found_fin_ok;
D
Dan Williams 已提交
2000 2001
		if (!(flags & MSG_PEEK))
			sk_eat_skb(sk, skb);
L
Linus Torvalds 已提交
2002 2003 2004 2005 2006
		continue;

	found_fin_ok:
		/* Process the FIN. */
		++*seq;
D
Dan Williams 已提交
2007 2008
		if (!(flags & MSG_PEEK))
			sk_eat_skb(sk, skb);
L
Linus Torvalds 已提交
2009 2010 2011 2012 2013 2014 2015
		break;
	} while (len > 0);

	/* According to UNIX98, msg_name/msg_namelen are ignored
	 * on connected socket. I was just happy when found this 8) --ANK
	 */

2016 2017 2018
	if (has_tss)
		tcp_recv_timestamp(msg, sk, &tss);

L
Linus Torvalds 已提交
2019
	/* Clean up data we have read: This will do ACK frames. */
2020
	tcp_cleanup_rbuf(sk, copied);
L
Linus Torvalds 已提交
2021 2022 2023 2024 2025 2026 2027 2028 2029

	release_sock(sk);
	return copied;

out:
	release_sock(sk);
	return err;

recv_urg:
2030
	err = tcp_recv_urg(sk, msg, len, flags);
L
Linus Torvalds 已提交
2031
	goto out;
P
Pavel Emelyanov 已提交
2032 2033 2034 2035

recv_sndq:
	err = tcp_peek_sndq(sk, msg, len);
	goto out;
L
Linus Torvalds 已提交
2036
}
E
Eric Dumazet 已提交
2037
EXPORT_SYMBOL(tcp_recvmsg);
L
Linus Torvalds 已提交
2038

I
Ilpo Järvinen 已提交
2039 2040 2041 2042
void tcp_set_state(struct sock *sk, int state)
{
	int oldstate = sk->sk_state;

2043 2044
	trace_tcp_set_state(sk, oldstate, state);

I
Ilpo Järvinen 已提交
2045 2046 2047
	switch (state) {
	case TCP_ESTABLISHED:
		if (oldstate != TCP_ESTABLISHED)
P
Pavel Emelyanov 已提交
2048
			TCP_INC_STATS(sock_net(sk), TCP_MIB_CURRESTAB);
I
Ilpo Järvinen 已提交
2049 2050 2051 2052
		break;

	case TCP_CLOSE:
		if (oldstate == TCP_CLOSE_WAIT || oldstate == TCP_ESTABLISHED)
P
Pavel Emelyanov 已提交
2053
			TCP_INC_STATS(sock_net(sk), TCP_MIB_ESTABRESETS);
I
Ilpo Järvinen 已提交
2054 2055 2056 2057

		sk->sk_prot->unhash(sk);
		if (inet_csk(sk)->icsk_bind_hash &&
		    !(sk->sk_userlocks & SOCK_BINDPORT_LOCK))
2058
			inet_put_port(sk);
I
Ilpo Järvinen 已提交
2059 2060
		/* fall through */
	default:
2061
		if (oldstate == TCP_ESTABLISHED)
P
Pavel Emelyanov 已提交
2062
			TCP_DEC_STATS(sock_net(sk), TCP_MIB_CURRESTAB);
I
Ilpo Järvinen 已提交
2063 2064 2065 2066 2067
	}

	/* Change state AFTER socket is unhashed to avoid closed
	 * socket sitting in hash tables.
	 */
2068
	sk_state_store(sk, state);
I
Ilpo Järvinen 已提交
2069 2070

#ifdef STATE_TRACE
2071
	SOCK_DEBUG(sk, "TCP sk=%p, State %s -> %s\n", sk, statename[oldstate], statename[state]);
I
Ilpo Järvinen 已提交
2072 2073 2074 2075
#endif
}
EXPORT_SYMBOL_GPL(tcp_set_state);

L
Linus Torvalds 已提交
2076 2077 2078 2079 2080 2081 2082
/*
 *	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.
 */

2083
static const unsigned char new_state[16] = {
L
Linus Torvalds 已提交
2084
  /* current state:        new state:      action:	*/
2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097
  [0 /* (Invalid) */]	= TCP_CLOSE,
  [TCP_ESTABLISHED]	= TCP_FIN_WAIT1 | TCP_ACTION_FIN,
  [TCP_SYN_SENT]	= TCP_CLOSE,
  [TCP_SYN_RECV]	= TCP_FIN_WAIT1 | TCP_ACTION_FIN,
  [TCP_FIN_WAIT1]	= TCP_FIN_WAIT1,
  [TCP_FIN_WAIT2]	= TCP_FIN_WAIT2,
  [TCP_TIME_WAIT]	= TCP_CLOSE,
  [TCP_CLOSE]		= TCP_CLOSE,
  [TCP_CLOSE_WAIT]	= TCP_LAST_ACK  | TCP_ACTION_FIN,
  [TCP_LAST_ACK]	= TCP_LAST_ACK,
  [TCP_LISTEN]		= TCP_CLOSE,
  [TCP_CLOSING]		= TCP_CLOSING,
  [TCP_NEW_SYN_RECV]	= TCP_CLOSE,	/* should not happen ! */
L
Linus Torvalds 已提交
2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111
};

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
2112
 *	that we don't receive shut down or sock_set_flag(sk, SOCK_DEAD).
L
Linus Torvalds 已提交
2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132
 */

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
2133
EXPORT_SYMBOL(tcp_shutdown);
L
Linus Torvalds 已提交
2134

A
Arun Sharma 已提交
2135 2136 2137 2138 2139 2140 2141
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);

2142 2143 2144 2145
	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
2146 2147 2148
	return too_many_orphans || out_of_socket_memory;
}

L
Linus Torvalds 已提交
2149 2150 2151 2152
void tcp_close(struct sock *sk, long timeout)
{
	struct sk_buff *skb;
	int data_was_unread = 0;
H
Herbert Xu 已提交
2153
	int state;
L
Linus Torvalds 已提交
2154 2155 2156 2157 2158 2159 2160 2161

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

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

		/* Special case. */
2162
		inet_csk_listen_stop(sk);
L
Linus Torvalds 已提交
2163 2164 2165 2166 2167 2168 2169 2170 2171

		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) {
2172 2173 2174 2175
		u32 len = TCP_SKB_CB(skb)->end_seq - TCP_SKB_CB(skb)->seq;

		if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN)
			len--;
L
Linus Torvalds 已提交
2176 2177 2178 2179
		data_was_unread += len;
		__kfree_skb(skb);
	}

2180
	sk_mem_reclaim(sk);
L
Linus Torvalds 已提交
2181

2182 2183 2184 2185
	/* If socket has been already reset (e.g. in tcp_reset()) - kill it. */
	if (sk->sk_state == TCP_CLOSE)
		goto adjudge_to_death;

2186 2187 2188 2189 2190 2191
	/* 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
2192
	 */
P
Pavel Emelyanov 已提交
2193 2194 2195
	if (unlikely(tcp_sk(sk)->repair)) {
		sk->sk_prot->disconnect(sk, 0);
	} else if (data_was_unread) {
L
Linus Torvalds 已提交
2196
		/* Unread data was tossed, zap the connection. */
2197
		NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPABORTONCLOSE);
L
Linus Torvalds 已提交
2198
		tcp_set_state(sk, TCP_CLOSE);
2199
		tcp_send_active_reset(sk, sk->sk_allocation);
L
Linus Torvalds 已提交
2200 2201 2202
	} else if (sock_flag(sk, SOCK_LINGER) && !sk->sk_lingertime) {
		/* Check zero linger _after_ checking for unread data. */
		sk->sk_prot->disconnect(sk, 0);
2203
		NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPABORTONDATA);
L
Linus Torvalds 已提交
2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228
	} 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
2229 2230 2231 2232
		 * 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
2233 2234 2235 2236 2237 2238 2239
		 */
		tcp_send_fin(sk);
	}

	sk_stream_wait_close(sk, timeout);

adjudge_to_death:
H
Herbert Xu 已提交
2240 2241 2242 2243
	state = sk->sk_state;
	sock_hold(sk);
	sock_orphan(sk);

L
Linus Torvalds 已提交
2244 2245 2246 2247 2248
	/* 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
2249
	 *  to finish close. No need to check for user refs.
L
Linus Torvalds 已提交
2250 2251 2252
	 */
	local_bh_disable();
	bh_lock_sock(sk);
2253
	WARN_ON(sock_owned_by_user(sk));
L
Linus Torvalds 已提交
2254

H
Herbert Xu 已提交
2255 2256
	percpu_counter_inc(sk->sk_prot->orphan_count);

H
Herbert Xu 已提交
2257 2258 2259
	/* 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
2260 2261 2262 2263

	/*	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.
2264
	 *	We use a 1 minute timeout (about the same as BSD) then kill
L
Linus Torvalds 已提交
2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279
	 *	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);
2280
			__NET_INC_STATS(sock_net(sk),
2281
					LINUX_MIB_TCPABORTONLINGER);
L
Linus Torvalds 已提交
2282
		} else {
2283
			const int tmo = tcp_fin_time(sk);
L
Linus Torvalds 已提交
2284 2285

			if (tmo > TCP_TIMEWAIT_LEN) {
2286 2287
				inet_csk_reset_keepalive_timer(sk,
						tmo - TCP_TIMEWAIT_LEN);
L
Linus Torvalds 已提交
2288 2289 2290 2291 2292 2293 2294
			} else {
				tcp_time_wait(sk, TCP_FIN_WAIT2, tmo);
				goto out;
			}
		}
	}
	if (sk->sk_state != TCP_CLOSE) {
2295
		sk_mem_reclaim(sk);
A
Arun Sharma 已提交
2296
		if (tcp_check_oom(sk, 0)) {
L
Linus Torvalds 已提交
2297 2298
			tcp_set_state(sk, TCP_CLOSE);
			tcp_send_active_reset(sk, GFP_ATOMIC);
2299
			__NET_INC_STATS(sock_net(sk),
2300
					LINUX_MIB_TCPABORTONMEMORY);
L
Linus Torvalds 已提交
2301 2302 2303
		}
	}

2304 2305 2306 2307 2308 2309
	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.
		 */
2310
		if (req)
2311
			reqsk_fastopen_remove(sk, req, false);
2312
		inet_csk_destroy_sock(sk);
2313
	}
L
Linus Torvalds 已提交
2314 2315 2316 2317 2318 2319 2320
	/* Otherwise, socket is reprieved until protocol close. */

out:
	bh_unlock_sock(sk);
	local_bh_enable();
	sock_put(sk);
}
E
Eric Dumazet 已提交
2321
EXPORT_SYMBOL(tcp_close);
L
Linus Torvalds 已提交
2322 2323 2324

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

E
Eric Dumazet 已提交
2325
static inline bool tcp_need_reset(int state)
L
Linus Torvalds 已提交
2326 2327 2328 2329 2330 2331
{
	return (1 << state) &
	       (TCPF_ESTABLISHED | TCPF_CLOSE_WAIT | TCPF_FIN_WAIT1 |
		TCPF_FIN_WAIT2 | TCPF_SYN_RECV);
}

2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347
static void tcp_rtx_queue_purge(struct sock *sk)
{
	struct rb_node *p = rb_first(&sk->tcp_rtx_queue);

	while (p) {
		struct sk_buff *skb = rb_to_skb(p);

		p = rb_next(p);
		/* Since we are deleting whole queue, no need to
		 * list_del(&skb->tcp_tsorted_anchor)
		 */
		tcp_rtx_queue_unlink(skb, sk);
		sk_wmem_free_skb(sk, skb);
	}
}

2348 2349 2350 2351 2352 2353 2354 2355 2356
void tcp_write_queue_purge(struct sock *sk)
{
	struct sk_buff *skb;

	tcp_chrono_stop(sk, TCP_CHRONO_BUSY);
	while ((skb = __skb_dequeue(&sk->sk_write_queue)) != NULL) {
		tcp_skb_tsorted_anchor_cleanup(skb);
		sk_wmem_free_skb(sk, skb);
	}
2357
	tcp_rtx_queue_purge(sk);
2358 2359 2360 2361 2362
	INIT_LIST_HEAD(&tcp_sk(sk)->tsorted_sent_queue);
	sk_mem_reclaim(sk);
	tcp_clear_all_retrans_hints(tcp_sk(sk));
}

L
Linus Torvalds 已提交
2363 2364 2365
int tcp_disconnect(struct sock *sk, int flags)
{
	struct inet_sock *inet = inet_sk(sk);
2366
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
2367 2368 2369 2370 2371 2372 2373 2374 2375
	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) {
2376
		inet_csk_listen_stop(sk);
P
Pavel Emelyanov 已提交
2377 2378
	} else if (unlikely(tp->repair)) {
		sk->sk_err = ECONNABORTED;
L
Linus Torvalds 已提交
2379 2380 2381
	} 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
2382
		/* The last check adjusts for discrepancy of Linux wrt. RFC
L
Linus Torvalds 已提交
2383 2384 2385 2386 2387 2388 2389 2390 2391
		 * 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);
2392
	tcp_write_queue_purge(sk);
2393
	tcp_fastopen_active_disable_ofo_check(sk);
2394
	skb_rbtree_purge(&tp->out_of_order_queue);
L
Linus Torvalds 已提交
2395

E
Eric Dumazet 已提交
2396
	inet->inet_dport = 0;
L
Linus Torvalds 已提交
2397 2398 2399 2400 2401 2402

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

	sk->sk_shutdown = 0;
	sock_reset_flag(sk, SOCK_DONE);
2403
	tp->srtt_us = 0;
2404 2405
	tp->write_seq += tp->max_window + 2;
	if (tp->write_seq == 0)
L
Linus Torvalds 已提交
2406
		tp->write_seq = 1;
2407
	icsk->icsk_backoff = 0;
L
Linus Torvalds 已提交
2408
	tp->snd_cwnd = 2;
2409
	icsk->icsk_probes_out = 0;
L
Linus Torvalds 已提交
2410
	tp->packets_out = 0;
I
Ilpo Järvinen 已提交
2411
	tp->snd_ssthresh = TCP_INFINITE_SSTHRESH;
L
Linus Torvalds 已提交
2412
	tp->snd_cwnd_cnt = 0;
2413
	tp->window_clamp = 0;
2414
	tcp_set_ca_state(sk, TCP_CA_Open);
L
Linus Torvalds 已提交
2415
	tcp_clear_retrans(tp);
2416
	inet_csk_delack_init(sk);
2417 2418 2419 2420
	/* Initialize rcv_mss to TCP_MIN_MSS to avoid division by 0
	 * issue in __tcp_select_window()
	 */
	icsk->icsk_ack.rcv_mss = TCP_MIN_MSS;
2421
	memset(&tp->rx_opt, 0, sizeof(tp->rx_opt));
L
Linus Torvalds 已提交
2422
	__sk_dst_reset(sk);
2423 2424
	dst_release(sk->sk_rx_dst);
	sk->sk_rx_dst = NULL;
2425
	tcp_saved_syn_free(tp);
L
Linus Torvalds 已提交
2426

2427 2428 2429 2430
	/* Clean up fastopen related fields */
	tcp_free_fastopen_req(tp);
	inet->defer_connect = 0;

E
Eric Dumazet 已提交
2431
	WARN_ON(inet->inet_num && !icsk->icsk_bind_hash);
L
Linus Torvalds 已提交
2432 2433 2434 2435

	sk->sk_error_report(sk);
	return err;
}
E
Eric Dumazet 已提交
2436
EXPORT_SYMBOL(tcp_disconnect);
L
Linus Torvalds 已提交
2437

E
Eric Dumazet 已提交
2438
static inline bool tcp_can_repair_sock(const struct sock *sk)
P
Pavel Emelyanov 已提交
2439
{
2440
	return ns_capable(sock_net(sk)->user_ns, CAP_NET_ADMIN) &&
2441
		(sk->sk_state != TCP_LISTEN);
P
Pavel Emelyanov 已提交
2442 2443
}

2444 2445 2446 2447 2448 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
static int tcp_repair_set_window(struct tcp_sock *tp, char __user *optbuf, int len)
{
	struct tcp_repair_window opt;

	if (!tp->repair)
		return -EPERM;

	if (len != sizeof(opt))
		return -EINVAL;

	if (copy_from_user(&opt, optbuf, sizeof(opt)))
		return -EFAULT;

	if (opt.max_window < opt.snd_wnd)
		return -EINVAL;

	if (after(opt.snd_wl1, tp->rcv_nxt + opt.rcv_wnd))
		return -EINVAL;

	if (after(opt.rcv_wup, tp->rcv_nxt))
		return -EINVAL;

	tp->snd_wl1	= opt.snd_wl1;
	tp->snd_wnd	= opt.snd_wnd;
	tp->max_window	= opt.max_window;

	tp->rcv_wnd	= opt.rcv_wnd;
	tp->rcv_wup	= opt.rcv_wup;

	return 0;
}

2476
static int tcp_repair_options_est(struct sock *sk,
2477
		struct tcp_repair_opt __user *optbuf, unsigned int len)
2478
{
2479
	struct tcp_sock *tp = tcp_sk(sk);
2480
	struct tcp_repair_opt opt;
2481

2482 2483
	while (len >= sizeof(opt)) {
		if (copy_from_user(&opt, optbuf, sizeof(opt)))
2484 2485 2486
			return -EFAULT;

		optbuf++;
2487
		len -= sizeof(opt);
2488

2489 2490 2491
		switch (opt.opt_code) {
		case TCPOPT_MSS:
			tp->rx_opt.mss_clamp = opt.opt_val;
2492
			tcp_mtup_init(sk);
2493
			break;
2494
		case TCPOPT_WINDOW:
2495 2496 2497 2498
			{
				u16 snd_wscale = opt.opt_val & 0xFFFF;
				u16 rcv_wscale = opt.opt_val >> 16;

2499
				if (snd_wscale > TCP_MAX_WSCALE || rcv_wscale > TCP_MAX_WSCALE)
2500
					return -EFBIG;
2501

2502 2503 2504 2505
				tp->rx_opt.snd_wscale = snd_wscale;
				tp->rx_opt.rcv_wscale = rcv_wscale;
				tp->rx_opt.wscale_ok = 1;
			}
2506 2507
			break;
		case TCPOPT_SACK_PERM:
2508 2509 2510
			if (opt.opt_val != 0)
				return -EINVAL;

2511 2512 2513
			tp->rx_opt.sack_ok |= TCP_SACK_SEEN;
			break;
		case TCPOPT_TIMESTAMP:
2514 2515 2516
			if (opt.opt_val != 0)
				return -EINVAL;

2517 2518 2519 2520 2521 2522 2523 2524
			tp->rx_opt.tstamp_ok = 1;
			break;
		}
	}

	return 0;
}

L
Linus Torvalds 已提交
2525 2526 2527
/*
 *	Socket option code for TCP.
 */
2528
static int do_tcp_setsockopt(struct sock *sk, int level,
2529
		int optname, char __user *optval, unsigned int optlen)
L
Linus Torvalds 已提交
2530 2531
{
	struct tcp_sock *tp = tcp_sk(sk);
2532
	struct inet_connection_sock *icsk = inet_csk(sk);
2533
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
2534 2535 2536
	int val;
	int err = 0;

2537 2538 2539
	/* These are data/string values, all the others are ints */
	switch (optname) {
	case TCP_CONGESTION: {
2540 2541 2542 2543 2544 2545
		char name[TCP_CA_NAME_MAX];

		if (optlen < 1)
			return -EINVAL;

		val = strncpy_from_user(name, optval,
2546
					min_t(long, TCP_CA_NAME_MAX-1, optlen));
2547 2548 2549 2550 2551
		if (val < 0)
			return -EFAULT;
		name[val] = 0;

		lock_sock(sk);
2552
		err = tcp_set_congestion_control(sk, name, true, true);
2553 2554 2555
		release_sock(sk);
		return err;
	}
D
Dave Watson 已提交
2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573
	case TCP_ULP: {
		char name[TCP_ULP_NAME_MAX];

		if (optlen < 1)
			return -EINVAL;

		val = strncpy_from_user(name, optval,
					min_t(long, TCP_ULP_NAME_MAX - 1,
					      optlen));
		if (val < 0)
			return -EFAULT;
		name[val] = 0;

		lock_sock(sk);
		err = tcp_set_ulp(sk, name);
		release_sock(sk);
		return err;
	}
2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584
	case TCP_FASTOPEN_KEY: {
		__u8 key[TCP_FASTOPEN_KEY_LENGTH];

		if (optlen != sizeof(key))
			return -EINVAL;

		if (copy_from_user(key, optval, optlen))
			return -EFAULT;

		return tcp_fastopen_reset_cipher(net, sk, key, sizeof(key));
	}
2585 2586 2587
	default:
		/* fallthru */
		break;
2588
	}
2589

L
Linus Torvalds 已提交
2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601
	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
2602 2603
		 * know which interface is going to be used
		 */
2604
		if (val && (val < TCP_MIN_MSS || val > MAX_TCP_WINDOW)) {
L
Linus Torvalds 已提交
2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621
			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;
2622
			tcp_push_pending_frames(sk);
L
Linus Torvalds 已提交
2623 2624 2625 2626 2627
		} else {
			tp->nonagle &= ~TCP_NAGLE_OFF;
		}
		break;

A
Andreas Petlund 已提交
2628 2629 2630 2631 2632 2633 2634
	case TCP_THIN_LINEAR_TIMEOUTS:
		if (val < 0 || val > 1)
			err = -EINVAL;
		else
			tp->thin_lto = val;
		break;

A
Andreas Petlund 已提交
2635 2636 2637 2638 2639
	case TCP_THIN_DUPACK:
		if (val < 0 || val > 1)
			err = -EINVAL;
		break;

P
Pavel Emelyanov 已提交
2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675
	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;

2676 2677 2678 2679
	case TCP_REPAIR_OPTIONS:
		if (!tp->repair)
			err = -EINVAL;
		else if (sk->sk_state == TCP_ESTABLISHED)
2680
			err = tcp_repair_options_est(sk,
2681 2682
					(struct tcp_repair_opt __user *)optval,
					optlen);
2683 2684 2685 2686
		else
			err = -EPERM;
		break;

L
Linus Torvalds 已提交
2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704
	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;
2705
			tcp_push_pending_frames(sk);
L
Linus Torvalds 已提交
2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716
		}
		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))) {
2717
				u32 elapsed = keepalive_time_elapsed(tp);
L
Linus Torvalds 已提交
2718 2719 2720 2721
				if (tp->keepalive_time > elapsed)
					elapsed = tp->keepalive_time - elapsed;
				else
					elapsed = 0;
2722
				inet_csk_reset_keepalive_timer(sk, elapsed);
L
Linus Torvalds 已提交
2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741
			}
		}
		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
2742
			icsk->icsk_syn_retries = val;
L
Linus Torvalds 已提交
2743 2744
		break;

2745 2746 2747 2748 2749 2750 2751
	case TCP_SAVE_SYN:
		if (val < 0 || val > 1)
			err = -EINVAL;
		else
			tp->save_syn = val;
		break;

L
Linus Torvalds 已提交
2752 2753 2754
	case TCP_LINGER2:
		if (val < 0)
			tp->linger2 = -1;
2755
		else if (val > net->ipv4.sysctl_tcp_fin_timeout / HZ)
L
Linus Torvalds 已提交
2756 2757 2758 2759 2760 2761
			tp->linger2 = 0;
		else
			tp->linger2 = val * HZ;
		break;

	case TCP_DEFER_ACCEPT:
2762 2763 2764 2765
		/* 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
2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781
		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) {
2782
			icsk->icsk_ack.pingpong = 1;
L
Linus Torvalds 已提交
2783
		} else {
2784
			icsk->icsk_ack.pingpong = 0;
L
Linus Torvalds 已提交
2785 2786
			if ((1 << sk->sk_state) &
			    (TCPF_ESTABLISHED | TCPF_CLOSE_WAIT) &&
2787 2788
			    inet_csk_ack_scheduled(sk)) {
				icsk->icsk_ack.pending |= ICSK_ACK_PUSHED;
2789
				tcp_cleanup_rbuf(sk, 1);
L
Linus Torvalds 已提交
2790
				if (!(val & 1))
2791
					icsk->icsk_ack.pingpong = 1;
L
Linus Torvalds 已提交
2792 2793 2794 2795
			}
		}
		break;

2796 2797
#ifdef CONFIG_TCP_MD5SIG
	case TCP_MD5SIG:
2798
	case TCP_MD5SIG_EXT:
2799
		/* Read the IP->Key mappings from userspace */
2800
		err = tp->af_specific->md5_parse(sk, optname, optval, optlen);
2801 2802
		break;
#endif
2803
	case TCP_USER_TIMEOUT:
2804
		/* Cap the max time in ms TCP will retry or probe the window
2805 2806
		 * before giving up and aborting (ETIMEDOUT) a connection.
		 */
2807 2808 2809 2810
		if (val < 0)
			err = -EINVAL;
		else
			icsk->icsk_user_timeout = msecs_to_jiffies(val);
2811
		break;
2812 2813 2814

	case TCP_FASTOPEN:
		if (val >= 0 && ((1 << sk->sk_state) & (TCPF_CLOSE |
2815
		    TCPF_LISTEN))) {
2816
			tcp_fastopen_init_key_once(net);
2817

2818
			fastopen_queue_tune(sk, val);
2819
		} else {
2820
			err = -EINVAL;
2821
		}
2822
		break;
W
Wei Wang 已提交
2823 2824 2825
	case TCP_FASTOPEN_CONNECT:
		if (val > 1 || val < 0) {
			err = -EINVAL;
2826
		} else if (net->ipv4.sysctl_tcp_fastopen & TFO_CLIENT_ENABLE) {
W
Wei Wang 已提交
2827 2828 2829 2830 2831 2832 2833 2834
			if (sk->sk_state == TCP_CLOSE)
				tp->fastopen_connect = val;
			else
				err = -EINVAL;
		} else {
			err = -EOPNOTSUPP;
		}
		break;
2835 2836 2837 2838 2839 2840 2841 2842
	case TCP_FASTOPEN_NO_COOKIE:
		if (val > 1 || val < 0)
			err = -EINVAL;
		else if (!((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN)))
			err = -EINVAL;
		else
			tp->fastopen_no_cookie = val;
		break;
2843 2844 2845 2846
	case TCP_TIMESTAMP:
		if (!tp->repair)
			err = -EPERM;
		else
2847
			tp->tsoffset = val - tcp_time_stamp_raw();
2848
		break;
2849 2850 2851
	case TCP_REPAIR_WINDOW:
		err = tcp_repair_set_window(tp, optval, optlen);
		break;
2852 2853 2854 2855
	case TCP_NOTSENT_LOWAT:
		tp->notsent_lowat = val;
		sk->sk_write_space(sk);
		break;
L
Linus Torvalds 已提交
2856 2857 2858
	default:
		err = -ENOPROTOOPT;
		break;
2859 2860
	}

L
Linus Torvalds 已提交
2861 2862 2863 2864
	release_sock(sk);
	return err;
}

2865
int tcp_setsockopt(struct sock *sk, int level, int optname, char __user *optval,
2866
		   unsigned int optlen)
2867
{
2868
	const struct inet_connection_sock *icsk = inet_csk(sk);
2869 2870 2871 2872 2873 2874

	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
2875
EXPORT_SYMBOL(tcp_setsockopt);
2876 2877

#ifdef CONFIG_COMPAT
2878
int compat_tcp_setsockopt(struct sock *sk, int level, int optname,
2879
			  char __user *optval, unsigned int optlen)
2880
{
2881 2882 2883
	if (level != SOL_TCP)
		return inet_csk_compat_setsockopt(sk, level, optname,
						  optval, optlen);
2884 2885
	return do_tcp_setsockopt(sk, level, optname, optval, optlen);
}
2886
EXPORT_SYMBOL(compat_tcp_setsockopt);
2887 2888
#endif

2889 2890 2891 2892 2893 2894 2895 2896 2897
static void tcp_get_info_chrono_stats(const struct tcp_sock *tp,
				      struct tcp_info *info)
{
	u64 stats[__TCP_CHRONO_MAX], total = 0;
	enum tcp_chrono i;

	for (i = TCP_CHRONO_BUSY; i < __TCP_CHRONO_MAX; ++i) {
		stats[i] = tp->chrono_stat[i - 1];
		if (i == tp->chrono_type)
2898
			stats[i] += tcp_jiffies32 - tp->chrono_start;
2899 2900 2901 2902 2903 2904 2905 2906 2907
		stats[i] *= USEC_PER_SEC / HZ;
		total += stats[i];
	}

	info->tcpi_busy_time = total;
	info->tcpi_rwnd_limited = stats[TCP_CHRONO_RWND_LIMITED];
	info->tcpi_sndbuf_limited = stats[TCP_CHRONO_SNDBUF_LIMITED];
}

L
Linus Torvalds 已提交
2908
/* Return information about state of tcp endpoint in API format. */
2909
void tcp_get_info(struct sock *sk, struct tcp_info *info)
L
Linus Torvalds 已提交
2910
{
2911
	const struct tcp_sock *tp = tcp_sk(sk); /* iff sk_type == SOCK_STREAM */
2912
	const struct inet_connection_sock *icsk = inet_csk(sk);
2913
	u32 now;
2914
	u64 rate64;
2915
	bool slow;
2916
	u32 rate;
L
Linus Torvalds 已提交
2917 2918

	memset(info, 0, sizeof(*info));
2919 2920
	if (sk->sk_type != SOCK_STREAM)
		return;
L
Linus Torvalds 已提交
2921

2922 2923
	info->tcpi_state = sk_state_load(sk);

2924 2925 2926
	/* Report meaningful fields for all TCP states, including listeners */
	rate = READ_ONCE(sk->sk_pacing_rate);
	rate64 = rate != ~0U ? rate : ~0ULL;
2927
	info->tcpi_pacing_rate = rate64;
2928 2929 2930

	rate = READ_ONCE(sk->sk_max_pacing_rate);
	rate64 = rate != ~0U ? rate : ~0ULL;
2931
	info->tcpi_max_pacing_rate = rate64;
2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944

	info->tcpi_reordering = tp->reordering;
	info->tcpi_snd_cwnd = tp->snd_cwnd;

	if (info->tcpi_state == TCP_LISTEN) {
		/* listeners aliased fields :
		 * tcpi_unacked -> Number of children ready for accept()
		 * tcpi_sacked  -> max backlog
		 */
		info->tcpi_unacked = sk->sk_ack_backlog;
		info->tcpi_sacked = sk->sk_max_ack_backlog;
		return;
	}
2945 2946 2947

	slow = lock_sock_fast(sk);

2948
	info->tcpi_ca_state = icsk->icsk_ca_state;
2949
	info->tcpi_retransmits = icsk->icsk_retransmits;
2950
	info->tcpi_probes = icsk->icsk_probes_out;
2951
	info->tcpi_backoff = icsk->icsk_backoff;
L
Linus Torvalds 已提交
2952 2953 2954

	if (tp->rx_opt.tstamp_ok)
		info->tcpi_options |= TCPI_OPT_TIMESTAMPS;
2955
	if (tcp_is_sack(tp))
L
Linus Torvalds 已提交
2956 2957 2958 2959 2960
		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;
2961
	}
L
Linus Torvalds 已提交
2962

E
Eric Dumazet 已提交
2963
	if (tp->ecn_flags & TCP_ECN_OK)
L
Linus Torvalds 已提交
2964
		info->tcpi_options |= TCPI_OPT_ECN;
E
Eric Dumazet 已提交
2965 2966
	if (tp->ecn_flags & TCP_ECN_SEEN)
		info->tcpi_options |= TCPI_OPT_ECN_SEEN;
Y
Yuchung Cheng 已提交
2967 2968
	if (tp->syn_data_acked)
		info->tcpi_options |= TCPI_OPT_SYN_DATA;
L
Linus Torvalds 已提交
2969

2970 2971
	info->tcpi_rto = jiffies_to_usecs(icsk->icsk_rto);
	info->tcpi_ato = jiffies_to_usecs(icsk->icsk_ack.ato);
2972
	info->tcpi_snd_mss = tp->mss_cache;
2973
	info->tcpi_rcv_mss = icsk->icsk_ack.rcv_mss;
L
Linus Torvalds 已提交
2974

2975 2976 2977
	info->tcpi_unacked = tp->packets_out;
	info->tcpi_sacked = tp->sacked_out;

L
Linus Torvalds 已提交
2978 2979 2980
	info->tcpi_lost = tp->lost_out;
	info->tcpi_retrans = tp->retrans_out;

2981
	now = tcp_jiffies32;
L
Linus Torvalds 已提交
2982
	info->tcpi_last_data_sent = jiffies_to_msecs(now - tp->lsndtime);
2983
	info->tcpi_last_data_recv = jiffies_to_msecs(now - icsk->icsk_ack.lrcvtime);
L
Linus Torvalds 已提交
2984 2985
	info->tcpi_last_ack_recv = jiffies_to_msecs(now - tp->rcv_tstamp);

2986
	info->tcpi_pmtu = icsk->icsk_pmtu_cookie;
L
Linus Torvalds 已提交
2987
	info->tcpi_rcv_ssthresh = tp->rcv_ssthresh;
2988 2989
	info->tcpi_rtt = tp->srtt_us >> 3;
	info->tcpi_rttvar = tp->mdev_us >> 2;
L
Linus Torvalds 已提交
2990 2991 2992
	info->tcpi_snd_ssthresh = tp->snd_ssthresh;
	info->tcpi_advmss = tp->advmss;

2993
	info->tcpi_rcv_rtt = tp->rcv_rtt_est.rtt_us >> 3;
L
Linus Torvalds 已提交
2994 2995 2996
	info->tcpi_rcv_space = tp->rcvq_space.space;

	info->tcpi_total_retrans = tp->total_retrans;
2997

2998 2999
	info->tcpi_bytes_acked = tp->bytes_acked;
	info->tcpi_bytes_received = tp->bytes_received;
3000
	info->tcpi_notsent_bytes = max_t(int, 0, tp->write_seq - tp->snd_nxt);
3001
	tcp_get_info_chrono_stats(tp, info);
3002

3003 3004
	info->tcpi_segs_out = tp->segs_out;
	info->tcpi_segs_in = tp->segs_in;
3005 3006

	info->tcpi_min_rtt = tcp_min_rtt(tp);
3007 3008
	info->tcpi_data_segs_in = tp->data_segs_in;
	info->tcpi_data_segs_out = tp->data_segs_out;
Y
Yuchung Cheng 已提交
3009 3010

	info->tcpi_delivery_rate_app_limited = tp->rate_app_limited ? 1 : 0;
3011 3012
	rate64 = tcp_compute_delivery_rate(tp);
	if (rate64)
3013
		info->tcpi_delivery_rate = rate64;
3014
	unlock_sock_fast(sk, slow);
L
Linus Torvalds 已提交
3015 3016 3017
}
EXPORT_SYMBOL_GPL(tcp_get_info);

3018 3019 3020 3021 3022
struct sk_buff *tcp_get_timestamping_opt_stats(const struct sock *sk)
{
	const struct tcp_sock *tp = tcp_sk(sk);
	struct sk_buff *stats;
	struct tcp_info info;
3023 3024
	u64 rate64;
	u32 rate;
3025

3026 3027 3028
	stats = alloc_skb(7 * nla_total_size_64bit(sizeof(u64)) +
			  3 * nla_total_size(sizeof(u32)) +
			  2 * nla_total_size(sizeof(u8)), GFP_ATOMIC);
3029 3030 3031 3032 3033 3034 3035 3036 3037 3038
	if (!stats)
		return NULL;

	tcp_get_info_chrono_stats(tp, &info);
	nla_put_u64_64bit(stats, TCP_NLA_BUSY,
			  info.tcpi_busy_time, TCP_NLA_PAD);
	nla_put_u64_64bit(stats, TCP_NLA_RWND_LIMITED,
			  info.tcpi_rwnd_limited, TCP_NLA_PAD);
	nla_put_u64_64bit(stats, TCP_NLA_SNDBUF_LIMITED,
			  info.tcpi_sndbuf_limited, TCP_NLA_PAD);
3039 3040 3041 3042
	nla_put_u64_64bit(stats, TCP_NLA_DATA_SEGS_OUT,
			  tp->data_segs_out, TCP_NLA_PAD);
	nla_put_u64_64bit(stats, TCP_NLA_TOTAL_RETRANS,
			  tp->total_retrans, TCP_NLA_PAD);
3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056

	rate = READ_ONCE(sk->sk_pacing_rate);
	rate64 = rate != ~0U ? rate : ~0ULL;
	nla_put_u64_64bit(stats, TCP_NLA_PACING_RATE, rate64, TCP_NLA_PAD);

	rate64 = tcp_compute_delivery_rate(tp);
	nla_put_u64_64bit(stats, TCP_NLA_DELIVERY_RATE, rate64, TCP_NLA_PAD);

	nla_put_u32(stats, TCP_NLA_SND_CWND, tp->snd_cwnd);
	nla_put_u32(stats, TCP_NLA_REORDERING, tp->reordering);
	nla_put_u32(stats, TCP_NLA_MIN_RTT, tcp_min_rtt(tp));

	nla_put_u8(stats, TCP_NLA_RECUR_RETRANS, inet_csk(sk)->icsk_retransmits);
	nla_put_u8(stats, TCP_NLA_DELIVERY_RATE_APP_LMT, !!tp->rate_app_limited);
3057 3058 3059
	return stats;
}

3060 3061
static int do_tcp_getsockopt(struct sock *sk, int level,
		int optname, char __user *optval, int __user *optlen)
L
Linus Torvalds 已提交
3062
{
3063
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
3064
	struct tcp_sock *tp = tcp_sk(sk);
3065
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077
	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:
3078
		val = tp->mss_cache;
L
Linus Torvalds 已提交
3079 3080
		if (!val && ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN)))
			val = tp->rx_opt.user_mss;
3081 3082
		if (tp->repair)
			val = tp->rx_opt.mss_clamp;
L
Linus Torvalds 已提交
3083 3084 3085 3086 3087 3088 3089 3090
		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
3091
		val = keepalive_time_when(tp) / HZ;
L
Linus Torvalds 已提交
3092 3093
		break;
	case TCP_KEEPINTVL:
E
Eric Dumazet 已提交
3094
		val = keepalive_intvl_when(tp) / HZ;
L
Linus Torvalds 已提交
3095 3096
		break;
	case TCP_KEEPCNT:
E
Eric Dumazet 已提交
3097
		val = keepalive_probes(tp);
L
Linus Torvalds 已提交
3098 3099
		break;
	case TCP_SYNCNT:
3100
		val = icsk->icsk_syn_retries ? : net->ipv4.sysctl_tcp_syn_retries;
L
Linus Torvalds 已提交
3101 3102 3103 3104
		break;
	case TCP_LINGER2:
		val = tp->linger2;
		if (val >= 0)
3105
			val = (val ? : net->ipv4.sysctl_tcp_fin_timeout) / HZ;
L
Linus Torvalds 已提交
3106 3107
		break;
	case TCP_DEFER_ACCEPT:
3108 3109
		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
3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128
		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;
	}
3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148
	case TCP_CC_INFO: {
		const struct tcp_congestion_ops *ca_ops;
		union tcp_cc_info info;
		size_t sz = 0;
		int attr;

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

		ca_ops = icsk->icsk_ca_ops;
		if (ca_ops && ca_ops->get_info)
			sz = ca_ops->get_info(sk, ~0U, &attr, &info);

		len = min_t(unsigned int, len, sz);
		if (put_user(len, optlen))
			return -EFAULT;
		if (copy_to_user(optval, &info, len))
			return -EFAULT;
		return 0;
	}
L
Linus Torvalds 已提交
3149
	case TCP_QUICKACK:
3150
		val = !icsk->icsk_ack.pingpong;
L
Linus Torvalds 已提交
3151
		break;
3152 3153 3154 3155 3156 3157 3158

	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;
3159
		if (copy_to_user(optval, icsk->icsk_ca_ops->name, len))
3160 3161
			return -EFAULT;
		return 0;
3162

D
Dave Watson 已提交
3163 3164 3165 3166
	case TCP_ULP:
		if (get_user(len, optlen))
			return -EFAULT;
		len = min_t(unsigned int, len, TCP_ULP_NAME_MAX);
3167 3168 3169 3170 3171
		if (!icsk->icsk_ulp_ops) {
			if (put_user(0, optlen))
				return -EFAULT;
			return 0;
		}
D
Dave Watson 已提交
3172 3173 3174 3175 3176 3177
		if (put_user(len, optlen))
			return -EFAULT;
		if (copy_to_user(optval, icsk->icsk_ulp_ops->name, len))
			return -EFAULT;
		return 0;

3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199
	case TCP_FASTOPEN_KEY: {
		__u8 key[TCP_FASTOPEN_KEY_LENGTH];
		struct tcp_fastopen_context *ctx;

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

		rcu_read_lock();
		ctx = rcu_dereference(icsk->icsk_accept_queue.fastopenq.ctx);
		if (ctx)
			memcpy(key, ctx->key, sizeof(key));
		else
			len = 0;
		rcu_read_unlock();

		len = min_t(unsigned int, len, sizeof(key));
		if (put_user(len, optlen))
			return -EFAULT;
		if (copy_to_user(optval, key, len))
			return -EFAULT;
		return 0;
	}
3200 3201 3202
	case TCP_THIN_LINEAR_TIMEOUTS:
		val = tp->thin_lto;
		break;
Y
Yuchung Cheng 已提交
3203

3204
	case TCP_THIN_DUPACK:
Y
Yuchung Cheng 已提交
3205
		val = 0;
3206
		break;
3207

P
Pavel Emelyanov 已提交
3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 3218
	case TCP_REPAIR:
		val = tp->repair;
		break;

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

3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240
	case TCP_REPAIR_WINDOW: {
		struct tcp_repair_window opt;

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

		if (len != sizeof(opt))
			return -EINVAL;

		if (!tp->repair)
			return -EPERM;

		opt.snd_wl1	= tp->snd_wl1;
		opt.snd_wnd	= tp->snd_wnd;
		opt.max_window	= tp->max_window;
		opt.rcv_wnd	= tp->rcv_wnd;
		opt.rcv_wup	= tp->rcv_wup;

		if (copy_to_user(optval, &opt, len))
			return -EFAULT;
		return 0;
	}
P
Pavel Emelyanov 已提交
3241 3242 3243 3244 3245 3246 3247 3248 3249
	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;

3250 3251 3252
	case TCP_USER_TIMEOUT:
		val = jiffies_to_msecs(icsk->icsk_user_timeout);
		break;
3253 3254

	case TCP_FASTOPEN:
3255
		val = icsk->icsk_accept_queue.fastopenq.max_qlen;
3256 3257
		break;

W
Wei Wang 已提交
3258 3259 3260 3261
	case TCP_FASTOPEN_CONNECT:
		val = tp->fastopen_connect;
		break;

3262 3263 3264 3265
	case TCP_FASTOPEN_NO_COOKIE:
		val = tp->fastopen_no_cookie;
		break;

3266
	case TCP_TIMESTAMP:
3267
		val = tcp_time_stamp_raw() + tp->tsoffset;
3268
		break;
3269 3270 3271
	case TCP_NOTSENT_LOWAT:
		val = tp->notsent_lowat;
		break;
3272 3273 3274 3275 3276 3277 3278 3279 3280
	case TCP_SAVE_SYN:
		val = tp->save_syn;
		break;
	case TCP_SAVED_SYN: {
		if (get_user(len, optlen))
			return -EFAULT;

		lock_sock(sk);
		if (tp->saved_syn) {
3281 3282 3283 3284 3285 3286 3287 3288 3289
			if (len < tp->saved_syn[0]) {
				if (put_user(tp->saved_syn[0], optlen)) {
					release_sock(sk);
					return -EFAULT;
				}
				release_sock(sk);
				return -EINVAL;
			}
			len = tp->saved_syn[0];
3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307
			if (put_user(len, optlen)) {
				release_sock(sk);
				return -EFAULT;
			}
			if (copy_to_user(optval, tp->saved_syn + 1, len)) {
				release_sock(sk);
				return -EFAULT;
			}
			tcp_saved_syn_free(tp);
			release_sock(sk);
		} else {
			release_sock(sk);
			len = 0;
			if (put_user(len, optlen))
				return -EFAULT;
		}
		return 0;
	}
L
Linus Torvalds 已提交
3308 3309
	default:
		return -ENOPROTOOPT;
3310
	}
L
Linus Torvalds 已提交
3311 3312 3313 3314 3315 3316 3317 3318

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

3319 3320 3321 3322 3323 3324 3325 3326 3327 3328
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
3329
EXPORT_SYMBOL(tcp_getsockopt);
3330 3331

#ifdef CONFIG_COMPAT
3332 3333
int compat_tcp_getsockopt(struct sock *sk, int level, int optname,
			  char __user *optval, int __user *optlen)
3334
{
3335 3336 3337
	if (level != SOL_TCP)
		return inet_csk_compat_getsockopt(sk, level, optname,
						  optval, optlen);
3338 3339
	return do_tcp_getsockopt(sk, level, optname, optval, optlen);
}
3340
EXPORT_SYMBOL(compat_tcp_getsockopt);
3341
#endif
L
Linus Torvalds 已提交
3342

3343
#ifdef CONFIG_TCP_MD5SIG
E
Eric Dumazet 已提交
3344
static DEFINE_PER_CPU(struct tcp_md5sig_pool, tcp_md5sig_pool);
3345
static DEFINE_MUTEX(tcp_md5sig_mutex);
E
Eric Dumazet 已提交
3346
static bool tcp_md5sig_pool_populated = false;
3347

3348
static void __tcp_alloc_md5sig_pool(void)
3349
{
H
Herbert Xu 已提交
3350
	struct crypto_ahash *hash;
3351 3352
	int cpu;

H
Herbert Xu 已提交
3353
	hash = crypto_alloc_ahash("md5", 0, CRYPTO_ALG_ASYNC);
3354
	if (IS_ERR(hash))
H
Herbert Xu 已提交
3355 3356
		return;

3357
	for_each_possible_cpu(cpu) {
3358
		void *scratch = per_cpu(tcp_md5sig_pool, cpu).scratch;
H
Herbert Xu 已提交
3359
		struct ahash_request *req;
3360

3361 3362 3363 3364 3365 3366 3367 3368 3369
		if (!scratch) {
			scratch = kmalloc_node(sizeof(union tcp_md5sum_block) +
					       sizeof(struct tcphdr),
					       GFP_KERNEL,
					       cpu_to_node(cpu));
			if (!scratch)
				return;
			per_cpu(tcp_md5sig_pool, cpu).scratch = scratch;
		}
H
Herbert Xu 已提交
3370 3371 3372 3373 3374 3375 3376 3377 3378 3379
		if (per_cpu(tcp_md5sig_pool, cpu).md5_req)
			continue;

		req = ahash_request_alloc(hash, GFP_KERNEL);
		if (!req)
			return;

		ahash_request_set_callback(req, 0, NULL, NULL);

		per_cpu(tcp_md5sig_pool, cpu).md5_req = req;
3380
	}
E
Eric Dumazet 已提交
3381 3382
	/* before setting tcp_md5sig_pool_populated, we must commit all writes
	 * to memory. See smp_rmb() in tcp_get_md5sig_pool()
3383 3384
	 */
	smp_wmb();
E
Eric Dumazet 已提交
3385
	tcp_md5sig_pool_populated = true;
3386 3387
}

3388
bool tcp_alloc_md5sig_pool(void)
3389
{
E
Eric Dumazet 已提交
3390
	if (unlikely(!tcp_md5sig_pool_populated)) {
3391 3392
		mutex_lock(&tcp_md5sig_mutex);

E
Eric Dumazet 已提交
3393
		if (!tcp_md5sig_pool_populated)
3394 3395 3396
			__tcp_alloc_md5sig_pool();

		mutex_unlock(&tcp_md5sig_mutex);
3397
	}
E
Eric Dumazet 已提交
3398
	return tcp_md5sig_pool_populated;
3399 3400 3401
}
EXPORT_SYMBOL(tcp_alloc_md5sig_pool);

E
Eric Dumazet 已提交
3402 3403 3404 3405 3406 3407 3408 3409 3410

/**
 *	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)
3411
{
E
Eric Dumazet 已提交
3412
	local_bh_disable();
3413

E
Eric Dumazet 已提交
3414 3415 3416 3417 3418
	if (tcp_md5sig_pool_populated) {
		/* coupled with smp_wmb() in __tcp_alloc_md5sig_pool() */
		smp_rmb();
		return this_cpu_ptr(&tcp_md5sig_pool);
	}
E
Eric Dumazet 已提交
3419 3420 3421 3422
	local_bh_enable();
	return NULL;
}
EXPORT_SYMBOL(tcp_get_md5sig_pool);
3423

3424
int tcp_md5_hash_skb_data(struct tcp_md5sig_pool *hp,
3425
			  const struct sk_buff *skb, unsigned int header_len)
3426 3427 3428
{
	struct scatterlist sg;
	const struct tcphdr *tp = tcp_hdr(skb);
H
Herbert Xu 已提交
3429
	struct ahash_request *req = hp->md5_req;
3430 3431 3432
	unsigned int i;
	const unsigned int head_data_len = skb_headlen(skb) > header_len ?
					   skb_headlen(skb) - header_len : 0;
3433
	const struct skb_shared_info *shi = skb_shinfo(skb);
3434
	struct sk_buff *frag_iter;
3435 3436 3437 3438

	sg_init_table(&sg, 1);

	sg_set_buf(&sg, ((u8 *) tp) + header_len, head_data_len);
H
Herbert Xu 已提交
3439 3440
	ahash_request_set_crypt(req, &sg, NULL, head_data_len);
	if (crypto_ahash_update(req))
3441 3442 3443 3444
		return 1;

	for (i = 0; i < shi->nr_frags; ++i) {
		const struct skb_frag_struct *f = &shi->frags[i];
E
Eric Dumazet 已提交
3445 3446 3447 3448 3449
		unsigned int offset = f->page_offset;
		struct page *page = skb_frag_page(f) + (offset >> PAGE_SHIFT);

		sg_set_page(&sg, page, skb_frag_size(f),
			    offset_in_page(offset));
H
Herbert Xu 已提交
3450 3451
		ahash_request_set_crypt(req, &sg, NULL, skb_frag_size(f));
		if (crypto_ahash_update(req))
3452 3453 3454
			return 1;
	}

3455 3456 3457 3458
	skb_walk_frags(skb, frag_iter)
		if (tcp_md5_hash_skb_data(hp, frag_iter, 0))
			return 1;

3459 3460 3461 3462
	return 0;
}
EXPORT_SYMBOL(tcp_md5_hash_skb_data);

3463
int tcp_md5_hash_key(struct tcp_md5sig_pool *hp, const struct tcp_md5sig_key *key)
3464 3465 3466 3467
{
	struct scatterlist sg;

	sg_init_one(&sg, key->key, key->keylen);
H
Herbert Xu 已提交
3468 3469
	ahash_request_set_crypt(hp->md5_req, &sg, NULL, key->keylen);
	return crypto_ahash_update(hp->md5_req);
3470 3471 3472
}
EXPORT_SYMBOL(tcp_md5_hash_key);

3473 3474
#endif

A
Andi Kleen 已提交
3475 3476
void tcp_done(struct sock *sk)
{
3477 3478
	struct request_sock *req = tcp_sk(sk)->fastopen_rsk;

3479
	if (sk->sk_state == TCP_SYN_SENT || sk->sk_state == TCP_SYN_RECV)
3480
		TCP_INC_STATS(sock_net(sk), TCP_MIB_ATTEMPTFAILS);
A
Andi Kleen 已提交
3481 3482 3483

	tcp_set_state(sk, TCP_CLOSE);
	tcp_clear_xmit_timers(sk);
3484
	if (req)
3485
		reqsk_fastopen_remove(sk, req, false);
A
Andi Kleen 已提交
3486 3487 3488 3489 3490 3491 3492 3493 3494 3495

	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);

3496 3497 3498
int tcp_abort(struct sock *sk, int err)
{
	if (!sk_fullsock(sk)) {
3499 3500 3501 3502 3503 3504 3505 3506 3507
		if (sk->sk_state == TCP_NEW_SYN_RECV) {
			struct request_sock *req = inet_reqsk(sk);

			local_bh_disable();
			inet_csk_reqsk_queue_drop_and_put(req->rsk_listener,
							  req);
			local_bh_enable();
			return 0;
		}
3508 3509 3510 3511 3512 3513
		return -EOPNOTSUPP;
	}

	/* Don't race with userspace socket closes such as tcp_close. */
	lock_sock(sk);

3514 3515 3516 3517 3518
	if (sk->sk_state == TCP_LISTEN) {
		tcp_set_state(sk, TCP_CLOSE);
		inet_csk_listen_stop(sk);
	}

3519 3520 3521 3522 3523 3524 3525 3526 3527 3528 3529 3530 3531 3532 3533 3534 3535 3536 3537 3538 3539
	/* Don't race with BH socket closes such as inet_csk_listen_stop. */
	local_bh_disable();
	bh_lock_sock(sk);

	if (!sock_flag(sk, SOCK_DEAD)) {
		sk->sk_err = err;
		/* This barrier is coupled with smp_rmb() in tcp_poll() */
		smp_wmb();
		sk->sk_error_report(sk);
		if (tcp_need_reset(sk->sk_state))
			tcp_send_active_reset(sk, GFP_ATOMIC);
		tcp_done(sk);
	}

	bh_unlock_sock(sk);
	local_bh_enable();
	release_sock(sk);
	return 0;
}
EXPORT_SYMBOL_GPL(tcp_abort);

3540
extern struct tcp_congestion_ops tcp_reno;
L
Linus Torvalds 已提交
3541 3542 3543 3544

static __initdata unsigned long thash_entries;
static int __init set_thash_entries(char *str)
{
3545 3546
	ssize_t ret;

L
Linus Torvalds 已提交
3547 3548
	if (!str)
		return 0;
3549 3550 3551 3552 3553

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

L
Linus Torvalds 已提交
3554 3555 3556 3557
	return 1;
}
__setup("thash_entries=", set_thash_entries);

F
Fabian Frederick 已提交
3558
static void __init tcp_init_mem(void)
3559
{
E
Eric Dumazet 已提交
3560 3561
	unsigned long limit = nr_free_buffer_pages() / 16;

3562
	limit = max(limit, 128UL);
E
Eric Dumazet 已提交
3563 3564 3565
	sysctl_tcp_mem[0] = limit / 4 * 3;		/* 4.68 % */
	sysctl_tcp_mem[1] = limit;			/* 6.25 % */
	sysctl_tcp_mem[2] = sysctl_tcp_mem[0] * 2;	/* 9.37 % */
3566 3567
}

L
Linus Torvalds 已提交
3568 3569
void __init tcp_init(void)
{
3570
	int max_rshare, max_wshare, cnt;
3571
	unsigned long limit;
3572
	unsigned int i;
L
Linus Torvalds 已提交
3573

3574 3575
	BUILD_BUG_ON(sizeof(struct tcp_skb_cb) >
		     FIELD_SIZEOF(struct sk_buff, cb));
L
Linus Torvalds 已提交
3576

3577 3578
	percpu_counter_init(&tcp_sockets_allocated, 0, GFP_KERNEL);
	percpu_counter_init(&tcp_orphan_count, 0, GFP_KERNEL);
3579
	inet_hashinfo_init(&tcp_hashinfo);
3580 3581 3582
	tcp_hashinfo.bind_bucket_cachep =
		kmem_cache_create("tcp_bind_bucket",
				  sizeof(struct inet_bind_bucket), 0,
3583
				  SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
L
Linus Torvalds 已提交
3584 3585 3586 3587 3588 3589

	/* Size and allocate the main established and bind bucket
	 * hash tables.
	 *
	 * The methodology is similar to that of the buffer cache.
	 */
3590
	tcp_hashinfo.ehash =
L
Linus Torvalds 已提交
3591
		alloc_large_system_hash("TCP established",
3592
					sizeof(struct inet_ehash_bucket),
L
Linus Torvalds 已提交
3593
					thash_entries,
E
Eric Dumazet 已提交
3594
					17, /* one slot per 128 KB of memory */
3595
					0,
L
Linus Torvalds 已提交
3596
					NULL,
3597
					&tcp_hashinfo.ehash_mask,
3598
					0,
3599
					thash_entries ? 0 : 512 * 1024);
E
Eric Dumazet 已提交
3600
	for (i = 0; i <= tcp_hashinfo.ehash_mask; i++)
3601
		INIT_HLIST_NULLS_HEAD(&tcp_hashinfo.ehash[i].chain, i);
E
Eric Dumazet 已提交
3602

3603 3604
	if (inet_ehash_locks_alloc(&tcp_hashinfo))
		panic("TCP: failed to alloc ehash_locks");
3605
	tcp_hashinfo.bhash =
L
Linus Torvalds 已提交
3606
		alloc_large_system_hash("TCP bind",
3607
					sizeof(struct inet_bind_hashbucket),
3608
					tcp_hashinfo.ehash_mask + 1,
E
Eric Dumazet 已提交
3609
					17, /* one slot per 128 KB of memory */
3610
					0,
3611
					&tcp_hashinfo.bhash_size,
L
Linus Torvalds 已提交
3612
					NULL,
3613
					0,
L
Linus Torvalds 已提交
3614
					64 * 1024);
3615
	tcp_hashinfo.bhash_size = 1U << tcp_hashinfo.bhash_size;
3616 3617 3618
	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
3619 3620
	}

E
Eric Dumazet 已提交
3621 3622 3623

	cnt = tcp_hashinfo.ehash_mask + 1;
	sysctl_tcp_max_orphans = cnt / 2;
L
Linus Torvalds 已提交
3624

3625
	tcp_init_mem();
3626
	/* Set per-socket limits to no more than 1/128 the pressure threshold */
E
Eric Dumazet 已提交
3627
	limit = nr_free_buffer_pages() << (PAGE_SHIFT - 7);
3628 3629
	max_wshare = min(4UL*1024*1024, limit);
	max_rshare = min(6UL*1024*1024, limit);
3630

3631 3632 3633
	init_net.ipv4.sysctl_tcp_wmem[0] = SK_MEM_QUANTUM;
	init_net.ipv4.sysctl_tcp_wmem[1] = 16*1024;
	init_net.ipv4.sysctl_tcp_wmem[2] = max(64*1024, max_wshare);
3634

3635 3636 3637
	init_net.ipv4.sysctl_tcp_rmem[0] = SK_MEM_QUANTUM;
	init_net.ipv4.sysctl_tcp_rmem[1] = 87380;
	init_net.ipv4.sysctl_tcp_rmem[2] = max(87380, max_rshare);
L
Linus Torvalds 已提交
3638

3639
	pr_info("Hash tables configured (established %u bind %u)\n",
J
Joe Perches 已提交
3640
		tcp_hashinfo.ehash_mask + 1, tcp_hashinfo.bhash_size);
3641

3642
	tcp_v4_init();
3643
	tcp_metrics_init();
3644
	BUG_ON(tcp_register_congestion_control(&tcp_reno) != 0);
E
Eric Dumazet 已提交
3645
	tcp_tasklet_init();
L
Linus Torvalds 已提交
3646
}