tcp_minisocks.c 24.8 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7
/*
 * INET		An implementation of the TCP/IP protocol suite for the LINUX
 *		operating system.  INET is implemented using the  BSD Socket
 *		interface as the means of communication with the user level.
 *
 *		Implementation of the Transmission Control Protocol(TCP).
 *
8
 * Authors:	Ross Biro
L
Linus Torvalds 已提交
9 10 11 12 13 14 15 16 17 18 19 20 21 22
 *		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>
 */

#include <linux/mm.h>
#include <linux/module.h>
23
#include <linux/slab.h>
L
Linus Torvalds 已提交
24 25 26 27 28 29
#include <linux/sysctl.h>
#include <linux/workqueue.h>
#include <net/tcp.h>
#include <net/inet_common.h>
#include <net/xfrm.h>

30
int sysctl_tcp_syncookies __read_mostly = 1;
31 32
EXPORT_SYMBOL(sysctl_tcp_syncookies);

33
int sysctl_tcp_abort_on_overflow __read_mostly;
L
Linus Torvalds 已提交
34

35 36 37
struct inet_timewait_death_row tcp_death_row = {
	.sysctl_max_tw_buckets = NR_FILE * 2,
	.period		= TCP_TIMEWAIT_LEN / INET_TWDR_TWKILL_SLOTS,
38
	.death_lock	= __SPIN_LOCK_UNLOCKED(tcp_death_row.death_lock),
39 40 41 42
	.hashinfo	= &tcp_hashinfo,
	.tw_timer	= TIMER_INITIALIZER(inet_twdr_hangman, 0,
					    (unsigned long)&tcp_death_row),
	.twkill_work	= __WORK_INITIALIZER(tcp_death_row.twkill_work,
43
					     inet_twdr_twkill_work),
44 45 46 47 48 49 50 51
/* Short-time timewait calendar */

	.twcal_hand	= -1,
	.twcal_timer	= TIMER_INITIALIZER(inet_twdr_twcal_tick, 0,
					    (unsigned long)&tcp_death_row),
};
EXPORT_SYMBOL_GPL(tcp_death_row);

E
Eric Dumazet 已提交
52
static bool tcp_in_window(u32 seq, u32 end_seq, u32 s_win, u32 e_win)
L
Linus Torvalds 已提交
53 54
{
	if (seq == s_win)
E
Eric Dumazet 已提交
55
		return true;
L
Linus Torvalds 已提交
56
	if (after(end_seq, s_win) && before(seq, e_win))
E
Eric Dumazet 已提交
57
		return true;
E
Eric Dumazet 已提交
58
	return seq == e_win && seq == end_seq;
L
Linus Torvalds 已提交
59 60
}

61
/*
L
Linus Torvalds 已提交
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
 * * Main purpose of TIME-WAIT state is to close connection gracefully,
 *   when one of ends sits in LAST-ACK or CLOSING retransmitting FIN
 *   (and, probably, tail of data) and one or more our ACKs are lost.
 * * What is TIME-WAIT timeout? It is associated with maximal packet
 *   lifetime in the internet, which results in wrong conclusion, that
 *   it is set to catch "old duplicate segments" wandering out of their path.
 *   It is not quite correct. This timeout is calculated so that it exceeds
 *   maximal retransmission timeout enough to allow to lose one (or more)
 *   segments sent by peer and our ACKs. This time may be calculated from RTO.
 * * When TIME-WAIT socket receives RST, it means that another end
 *   finally closed and we are allowed to kill TIME-WAIT too.
 * * Second purpose of TIME-WAIT is catching old duplicate segments.
 *   Well, certainly it is pure paranoia, but if we load TIME-WAIT
 *   with this semantics, we MUST NOT kill TIME-WAIT state with RSTs.
 * * If we invented some more clever way to catch duplicates
 *   (f.e. based on PAWS), we could truncate TIME-WAIT to several RTOs.
 *
 * The algorithm below is based on FORMAL INTERPRETATION of RFCs.
 * When you compare it to RFCs, please, read section SEGMENT ARRIVES
 * from the very beginning.
 *
 * NOTE. With recycling (and later with fin-wait-2) TW bucket
 * is _not_ stateless. It means, that strictly speaking we must
 * spinlock it. I do not want! Well, probability of misbehaviour
 * is ridiculously low and, seems, we could use some mb() tricks
 * to avoid misread sequence numbers, states etc.  --ANK
88 89
 *
 * We don't need to initialize tmp_out.sack_ok as we don't use the results
L
Linus Torvalds 已提交
90 91
 */
enum tcp_tw_status
92 93
tcp_timewait_state_process(struct inet_timewait_sock *tw, struct sk_buff *skb,
			   const struct tcphdr *th)
L
Linus Torvalds 已提交
94 95
{
	struct tcp_options_received tmp_opt;
96
	const u8 *hash_location;
97
	struct tcp_timewait_sock *tcptw = tcp_twsk((struct sock *)tw);
E
Eric Dumazet 已提交
98
	bool paws_reject = false;
L
Linus Torvalds 已提交
99

100
	tmp_opt.saw_tstamp = 0;
101
	if (th->doff > (sizeof(*th) >> 2) && tcptw->tw_ts_recent_stamp) {
Y
Yuchung Cheng 已提交
102
		tcp_parse_options(skb, &tmp_opt, &hash_location, 0, NULL);
L
Linus Torvalds 已提交
103 104

		if (tmp_opt.saw_tstamp) {
105 106
			tmp_opt.ts_recent	= tcptw->tw_ts_recent;
			tmp_opt.ts_recent_stamp	= tcptw->tw_ts_recent_stamp;
I
Ilpo Järvinen 已提交
107
			paws_reject = tcp_paws_reject(&tmp_opt, th->rst);
L
Linus Torvalds 已提交
108 109 110 111 112 113 114 115 116
		}
	}

	if (tw->tw_substate == TCP_FIN_WAIT2) {
		/* Just repeat all the checks of tcp_rcv_state_process() */

		/* Out of window, send ACK */
		if (paws_reject ||
		    !tcp_in_window(TCP_SKB_CB(skb)->seq, TCP_SKB_CB(skb)->end_seq,
117 118
				   tcptw->tw_rcv_nxt,
				   tcptw->tw_rcv_nxt + tcptw->tw_rcv_wnd))
L
Linus Torvalds 已提交
119 120 121 122 123
			return TCP_TW_ACK;

		if (th->rst)
			goto kill;

124
		if (th->syn && !before(TCP_SKB_CB(skb)->seq, tcptw->tw_rcv_nxt))
L
Linus Torvalds 已提交
125 126 127
			goto kill_with_rst;

		/* Dup ACK? */
128 129
		if (!th->ack ||
		    !after(TCP_SKB_CB(skb)->end_seq, tcptw->tw_rcv_nxt) ||
L
Linus Torvalds 已提交
130
		    TCP_SKB_CB(skb)->end_seq == TCP_SKB_CB(skb)->seq) {
131
			inet_twsk_put(tw);
L
Linus Torvalds 已提交
132 133 134 135 136 137 138
			return TCP_TW_SUCCESS;
		}

		/* New data or FIN. If new data arrive after half-duplex close,
		 * reset.
		 */
		if (!th->fin ||
139
		    TCP_SKB_CB(skb)->end_seq != tcptw->tw_rcv_nxt + 1) {
L
Linus Torvalds 已提交
140
kill_with_rst:
141
			inet_twsk_deschedule(tw, &tcp_death_row);
142
			inet_twsk_put(tw);
L
Linus Torvalds 已提交
143 144 145 146
			return TCP_TW_RST;
		}

		/* FIN arrived, enter true time-wait state. */
147 148
		tw->tw_substate	  = TCP_TIME_WAIT;
		tcptw->tw_rcv_nxt = TCP_SKB_CB(skb)->end_seq;
L
Linus Torvalds 已提交
149
		if (tmp_opt.saw_tstamp) {
150
			tcptw->tw_ts_recent_stamp = get_seconds();
151
			tcptw->tw_ts_recent	  = tmp_opt.rcv_tsval;
L
Linus Torvalds 已提交
152 153
		}

154 155 156
		if (tcp_death_row.sysctl_tw_recycle &&
		    tcptw->tw_ts_recent_stamp &&
		    tcp_tw_remember_stamp(tw))
157 158
			inet_twsk_schedule(tw, &tcp_death_row, tw->tw_timeout,
					   TCP_TIMEWAIT_LEN);
L
Linus Torvalds 已提交
159
		else
160 161
			inet_twsk_schedule(tw, &tcp_death_row, TCP_TIMEWAIT_LEN,
					   TCP_TIMEWAIT_LEN);
L
Linus Torvalds 已提交
162 163 164 165 166 167 168 169 170 171
		return TCP_TW_ACK;
	}

	/*
	 *	Now real TIME-WAIT state.
	 *
	 *	RFC 1122:
	 *	"When a connection is [...] on TIME-WAIT state [...]
	 *	[a TCP] MAY accept a new SYN from the remote TCP to
	 *	reopen the connection directly, if it:
172
	 *
L
Linus Torvalds 已提交
173 174 175 176 177
	 *	(1)  assigns its initial sequence number for the new
	 *	connection to be larger than the largest sequence
	 *	number it used on the previous connection incarnation,
	 *	and
	 *
178
	 *	(2)  returns to TIME-WAIT state if the SYN turns out
L
Linus Torvalds 已提交
179 180 181 182
	 *	to be an old duplicate".
	 */

	if (!paws_reject &&
183
	    (TCP_SKB_CB(skb)->seq == tcptw->tw_rcv_nxt &&
L
Linus Torvalds 已提交
184 185 186 187
	     (TCP_SKB_CB(skb)->seq == TCP_SKB_CB(skb)->end_seq || th->rst))) {
		/* In window segment, it may be only reset or bare ack. */

		if (th->rst) {
S
Stephen Hemminger 已提交
188
			/* This is TIME_WAIT assassination, in two flavors.
L
Linus Torvalds 已提交
189 190 191 192 193
			 * Oh well... nobody has a sufficient solution to this
			 * protocol bug yet.
			 */
			if (sysctl_tcp_rfc1337 == 0) {
kill:
194
				inet_twsk_deschedule(tw, &tcp_death_row);
195
				inet_twsk_put(tw);
L
Linus Torvalds 已提交
196 197 198
				return TCP_TW_SUCCESS;
			}
		}
199 200
		inet_twsk_schedule(tw, &tcp_death_row, TCP_TIMEWAIT_LEN,
				   TCP_TIMEWAIT_LEN);
L
Linus Torvalds 已提交
201 202

		if (tmp_opt.saw_tstamp) {
203
			tcptw->tw_ts_recent	  = tmp_opt.rcv_tsval;
204
			tcptw->tw_ts_recent_stamp = get_seconds();
L
Linus Torvalds 已提交
205 206
		}

207
		inet_twsk_put(tw);
L
Linus Torvalds 已提交
208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228
		return TCP_TW_SUCCESS;
	}

	/* Out of window segment.

	   All the segments are ACKed immediately.

	   The only exception is new SYN. We accept it, if it is
	   not old duplicate and we are not in danger to be killed
	   by delayed old duplicates. RFC check is that it has
	   newer sequence number works at rates <40Mbit/sec.
	   However, if paws works, it is reliable AND even more,
	   we even may relax silly seq space cutoff.

	   RED-PEN: we violate main RFC requirement, if this SYN will appear
	   old duplicate (i.e. we receive RST in reply to SYN-ACK),
	   we must return socket to time-wait state. It is not good,
	   but not fatal yet.
	 */

	if (th->syn && !th->rst && !th->ack && !paws_reject &&
229 230 231 232
	    (after(TCP_SKB_CB(skb)->seq, tcptw->tw_rcv_nxt) ||
	     (tmp_opt.saw_tstamp &&
	      (s32)(tcptw->tw_ts_recent - tmp_opt.rcv_tsval) < 0))) {
		u32 isn = tcptw->tw_snd_nxt + 65535 + 2;
L
Linus Torvalds 已提交
233 234 235 236 237 238 239
		if (isn == 0)
			isn++;
		TCP_SKB_CB(skb)->when = isn;
		return TCP_TW_SYN;
	}

	if (paws_reject)
240
		NET_INC_STATS_BH(twsk_net(tw), LINUX_MIB_PAWSESTABREJECTED);
L
Linus Torvalds 已提交
241

S
Stephen Hemminger 已提交
242
	if (!th->rst) {
L
Linus Torvalds 已提交
243 244 245 246 247 248 249
		/* In this case we must reset the TIMEWAIT timer.
		 *
		 * If it is ACKless SYN it may be both old duplicate
		 * and new good SYN with random sequence number <rcv_nxt.
		 * Do not reschedule in the last case.
		 */
		if (paws_reject || th->ack)
250 251
			inet_twsk_schedule(tw, &tcp_death_row, TCP_TIMEWAIT_LEN,
					   TCP_TIMEWAIT_LEN);
L
Linus Torvalds 已提交
252 253 254 255 256 257

		/* Send ACK. Note, we do not put the bucket,
		 * it will be released by caller.
		 */
		return TCP_TW_ACK;
	}
258
	inet_twsk_put(tw);
L
Linus Torvalds 已提交
259 260
	return TCP_TW_SUCCESS;
}
E
Eric Dumazet 已提交
261
EXPORT_SYMBOL(tcp_timewait_state_process);
L
Linus Torvalds 已提交
262

263
/*
L
Linus Torvalds 已提交
264
 * Move a socket to time-wait or dead fin-wait-2 state.
265
 */
L
Linus Torvalds 已提交
266 267
void tcp_time_wait(struct sock *sk, int state, int timeo)
{
268
	struct inet_timewait_sock *tw = NULL;
269
	const struct inet_connection_sock *icsk = inet_csk(sk);
270
	const struct tcp_sock *tp = tcp_sk(sk);
E
Eric Dumazet 已提交
271
	bool recycle_ok = false;
L
Linus Torvalds 已提交
272

D
David S. Miller 已提交
273
	if (tcp_death_row.sysctl_tw_recycle && tp->rx_opt.ts_recent_stamp)
274
		recycle_ok = tcp_remember_stamp(sk);
L
Linus Torvalds 已提交
275

276
	if (tcp_death_row.tw_count < tcp_death_row.sysctl_max_tw_buckets)
277
		tw = inet_twsk_alloc(sk, state);
L
Linus Torvalds 已提交
278

279 280
	if (tw != NULL) {
		struct tcp_timewait_sock *tcptw = tcp_twsk((struct sock *)tw);
281
		const int rto = (icsk->icsk_rto << 2) - (icsk->icsk_rto >> 1);
282
		struct inet_sock *inet = inet_sk(sk);
283

284
		tw->tw_transparent	= inet->transparent;
L
Linus Torvalds 已提交
285
		tw->tw_rcv_wscale	= tp->rx_opt.rcv_wscale;
286 287 288 289 290
		tcptw->tw_rcv_nxt	= tp->rcv_nxt;
		tcptw->tw_snd_nxt	= tp->snd_nxt;
		tcptw->tw_rcv_wnd	= tcp_receive_window(tp);
		tcptw->tw_ts_recent	= tp->rx_opt.ts_recent;
		tcptw->tw_ts_recent_stamp = tp->rx_opt.ts_recent_stamp;
L
Linus Torvalds 已提交
291

E
Eric Dumazet 已提交
292
#if IS_ENABLED(CONFIG_IPV6)
L
Linus Torvalds 已提交
293 294
		if (tw->tw_family == PF_INET6) {
			struct ipv6_pinfo *np = inet6_sk(sk);
295
			struct inet6_timewait_sock *tw6;
L
Linus Torvalds 已提交
296

297 298
			tw->tw_ipv6_offset = inet6_tw_offset(sk->sk_prot);
			tw6 = inet6_twsk((struct sock *)tw);
A
Alexey Dobriyan 已提交
299 300
			tw6->tw_v6_daddr = np->daddr;
			tw6->tw_v6_rcv_saddr = np->rcv_saddr;
301
			tw->tw_tclass = np->tclass;
302
			tw->tw_ipv6only = np->ipv6only;
303
		}
L
Linus Torvalds 已提交
304
#endif
305 306 307 308 309 310 311 312 313 314

#ifdef CONFIG_TCP_MD5SIG
		/*
		 * The timewait bucket does not have the key DB from the
		 * sock structure. We just make a quick copy of the
		 * md5 key being used (if indeed we are using one)
		 * so the timewait ack generating code has the key.
		 */
		do {
			struct tcp_md5sig_key *key;
E
Eric Dumazet 已提交
315
			tcptw->tw_md5_key = NULL;
316 317
			key = tp->af_specific->md5_lookup(sk, sk);
			if (key != NULL) {
E
Eric Dumazet 已提交
318 319
				tcptw->tw_md5_key = kmemdup(key, sizeof(*key), GFP_ATOMIC);
				if (tcptw->tw_md5_key && tcp_alloc_md5sig_pool(sk) == NULL)
320 321
					BUG();
			}
S
Stephen Hemminger 已提交
322
		} while (0);
323 324
#endif

L
Linus Torvalds 已提交
325
		/* Linkage updates. */
326
		__inet_twsk_hashdance(tw, sk, &tcp_hashinfo);
L
Linus Torvalds 已提交
327 328 329 330 331 332 333 334 335 336 337 338 339

		/* Get the TIME_WAIT timeout firing. */
		if (timeo < rto)
			timeo = rto;

		if (recycle_ok) {
			tw->tw_timeout = rto;
		} else {
			tw->tw_timeout = TCP_TIMEWAIT_LEN;
			if (state == TCP_TIME_WAIT)
				timeo = TCP_TIMEWAIT_LEN;
		}

340 341
		inet_twsk_schedule(tw, &tcp_death_row, timeo,
				   TCP_TIMEWAIT_LEN);
342
		inet_twsk_put(tw);
L
Linus Torvalds 已提交
343 344 345 346 347
	} else {
		/* Sorry, if we're out of memory, just CLOSE this
		 * socket up.  We've got bigger problems than
		 * non-graceful socket closings.
		 */
348
		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPTIMEWAITOVERFLOW);
L
Linus Torvalds 已提交
349 350 351 352 353 354
	}

	tcp_update_metrics(sk);
	tcp_done(sk);
}

355 356
void tcp_twsk_destructor(struct sock *sk)
{
D
David S. Miller 已提交
357
#ifdef CONFIG_TCP_MD5SIG
358
	struct tcp_timewait_sock *twsk = tcp_twsk(sk);
359

E
Eric Dumazet 已提交
360
	if (twsk->tw_md5_key) {
361
		tcp_free_md5sig_pool();
E
Eric Dumazet 已提交
362 363
		kfree_rcu(twsk->tw_md5_key, rcu);
	}
364 365 366 367
#endif
}
EXPORT_SYMBOL_GPL(tcp_twsk_destructor);

368 369 370 371 372 373
static inline void TCP_ECN_openreq_child(struct tcp_sock *tp,
					 struct request_sock *req)
{
	tp->ecn_flags = inet_rsk(req)->ecn_ok ? TCP_ECN_OK : 0;
}

L
Linus Torvalds 已提交
374 375 376 377 378 379
/* This is not only more efficient than what we used to do, it eliminates
 * a lot of code duplication between IPv4/IPv6 SYN recv processing. -DaveM
 *
 * Actually, we could lots of memory writes here. tp of listening
 * socket contains all necessary default parameters.
 */
380
struct sock *tcp_create_openreq_child(struct sock *sk, struct request_sock *req, struct sk_buff *skb)
L
Linus Torvalds 已提交
381
{
382
	struct sock *newsk = inet_csk_clone_lock(sk, req, GFP_ATOMIC);
L
Linus Torvalds 已提交
383

384
	if (newsk != NULL) {
385
		const struct inet_request_sock *ireq = inet_rsk(req);
386
		struct tcp_request_sock *treq = tcp_rsk(req);
387
		struct inet_connection_sock *newicsk = inet_csk(newsk);
388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414
		struct tcp_sock *newtp = tcp_sk(newsk);
		struct tcp_sock *oldtp = tcp_sk(sk);
		struct tcp_cookie_values *oldcvp = oldtp->cookie_values;

		/* TCP Cookie Transactions require space for the cookie pair,
		 * as it differs for each connection.  There is no need to
		 * copy any s_data_payload stored at the original socket.
		 * Failure will prevent resuming the connection.
		 *
		 * Presumed copied, in order of appearance:
		 *	cookie_in_always, cookie_out_never
		 */
		if (oldcvp != NULL) {
			struct tcp_cookie_values *newcvp =
				kzalloc(sizeof(*newtp->cookie_values),
					GFP_ATOMIC);

			if (newcvp != NULL) {
				kref_init(&newcvp->kref);
				newcvp->cookie_desired =
						oldcvp->cookie_desired;
				newtp->cookie_values = newcvp;
			} else {
				/* Not Yet Implemented */
				newtp->cookie_values = NULL;
			}
		}
L
Linus Torvalds 已提交
415 416 417

		/* Now setup tcp_sock */
		newtp->pred_flags = 0;
418 419 420 421 422 423 424

		newtp->rcv_wup = newtp->copied_seq =
		newtp->rcv_nxt = treq->rcv_isn + 1;

		newtp->snd_sml = newtp->snd_una =
		newtp->snd_nxt = newtp->snd_up =
			treq->snt_isn + 1 + tcp_s_data_size(oldtp);
L
Linus Torvalds 已提交
425 426

		tcp_prequeue_init(newtp);
E
Eric Dumazet 已提交
427
		INIT_LIST_HEAD(&newtp->tsq_node);
L
Linus Torvalds 已提交
428

429
		tcp_init_wl(newtp, treq->rcv_isn);
L
Linus Torvalds 已提交
430 431 432

		newtp->srtt = 0;
		newtp->mdev = TCP_TIMEOUT_INIT;
433
		newicsk->icsk_rto = TCP_TIMEOUT_INIT;
L
Linus Torvalds 已提交
434 435 436 437 438

		newtp->packets_out = 0;
		newtp->retrans_out = 0;
		newtp->sacked_out = 0;
		newtp->fackets_out = 0;
I
Ilpo Järvinen 已提交
439
		newtp->snd_ssthresh = TCP_INFINITE_SSTHRESH;
Y
Yuchung Cheng 已提交
440
		tcp_enable_early_retrans(newtp);
L
Linus Torvalds 已提交
441 442 443 444 445 446

		/* 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
		 */
447
		newtp->snd_cwnd = TCP_INIT_CWND;
L
Linus Torvalds 已提交
448 449 450 451 452
		newtp->snd_cwnd_cnt = 0;

		newtp->frto_counter = 0;
		newtp->frto_highmark = 0;

453 454 455
		if (newicsk->icsk_ca_ops != &tcp_init_congestion_ops &&
		    !try_module_get(newicsk->icsk_ca_ops->owner))
			newicsk->icsk_ca_ops = &tcp_init_congestion_ops;
456

457
		tcp_set_ca_state(newsk, TCP_CA_Open);
L
Linus Torvalds 已提交
458 459
		tcp_init_xmit_timers(newsk);
		skb_queue_head_init(&newtp->out_of_order_queue);
460 461
		newtp->write_seq = newtp->pushed_seq =
			treq->snt_isn + 1 + tcp_s_data_size(oldtp);
L
Linus Torvalds 已提交
462 463 464 465 466

		newtp->rx_opt.saw_tstamp = 0;

		newtp->rx_opt.dsack = 0;
		newtp->rx_opt.num_sacks = 0;
467

L
Linus Torvalds 已提交
468 469 470
		newtp->urg_data = 0;

		if (sock_flag(newsk, SOCK_KEEPOPEN))
471 472
			inet_csk_reset_keepalive_timer(newsk,
						       keepalive_time_when(newtp));
L
Linus Torvalds 已提交
473

474
		newtp->rx_opt.tstamp_ok = ireq->tstamp_ok;
S
Stephen Hemminger 已提交
475
		if ((newtp->rx_opt.sack_ok = ireq->sack_ok) != 0) {
L
Linus Torvalds 已提交
476
			if (sysctl_tcp_fack)
477
				tcp_enable_fack(newtp);
L
Linus Torvalds 已提交
478 479 480 481
		}
		newtp->window_clamp = req->window_clamp;
		newtp->rcv_ssthresh = req->rcv_wnd;
		newtp->rcv_wnd = req->rcv_wnd;
482
		newtp->rx_opt.wscale_ok = ireq->wscale_ok;
L
Linus Torvalds 已提交
483
		if (newtp->rx_opt.wscale_ok) {
484 485
			newtp->rx_opt.snd_wscale = ireq->snd_wscale;
			newtp->rx_opt.rcv_wscale = ireq->rcv_wscale;
L
Linus Torvalds 已提交
486 487 488 489
		} else {
			newtp->rx_opt.snd_wscale = newtp->rx_opt.rcv_wscale = 0;
			newtp->window_clamp = min(newtp->window_clamp, 65535U);
		}
490 491
		newtp->snd_wnd = (ntohs(tcp_hdr(skb)->window) <<
				  newtp->rx_opt.snd_wscale);
L
Linus Torvalds 已提交
492 493 494 495
		newtp->max_window = newtp->snd_wnd;

		if (newtp->rx_opt.tstamp_ok) {
			newtp->rx_opt.ts_recent = req->ts_recent;
496
			newtp->rx_opt.ts_recent_stamp = get_seconds();
L
Linus Torvalds 已提交
497 498 499 500 501
			newtp->tcp_header_len = sizeof(struct tcphdr) + TCPOLEN_TSTAMP_ALIGNED;
		} else {
			newtp->rx_opt.ts_recent_stamp = 0;
			newtp->tcp_header_len = sizeof(struct tcphdr);
		}
502 503 504 505 506
#ifdef CONFIG_TCP_MD5SIG
		newtp->md5sig_info = NULL;	/*XXX*/
		if (newtp->af_specific->md5_lookup(sk, newsk))
			newtp->tcp_header_len += TCPOLEN_MD5SIG_ALIGNED;
#endif
507
		if (skb->len >= TCP_MSS_DEFAULT + newtp->tcp_header_len)
508
			newicsk->icsk_ack.last_seg_size = skb->len - newtp->tcp_header_len;
L
Linus Torvalds 已提交
509 510
		newtp->rx_opt.mss_clamp = req->mss;
		TCP_ECN_openreq_child(newtp, req);
511
		newtp->fastopen_rsk = NULL;
Y
Yuchung Cheng 已提交
512
		newtp->syn_data_acked = 0;
L
Linus Torvalds 已提交
513

514
		TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_PASSIVEOPENS);
L
Linus Torvalds 已提交
515 516 517
	}
	return newsk;
}
E
Eric Dumazet 已提交
518
EXPORT_SYMBOL(tcp_create_openreq_child);
L
Linus Torvalds 已提交
519

520
/*
521 522 523 524 525 526
 * Process an incoming packet for SYN_RECV sockets represented as a
 * request_sock. Normally sk is the listener socket but for TFO it
 * points to the child socket.
 *
 * XXX (TFO) - The current impl contains a special check for ack
 * validation and inside tcp_v4_reqsk_send_ack(). Can we do better?
527 528
 *
 * We don't need to initialize tmp_opt.sack_ok as we don't use the results
L
Linus Torvalds 已提交
529 530
 */

531
struct sock *tcp_check_req(struct sock *sk, struct sk_buff *skb,
532
			   struct request_sock *req,
533 534
			   struct request_sock **prev,
			   bool fastopen)
L
Linus Torvalds 已提交
535
{
536
	struct tcp_options_received tmp_opt;
537
	const u8 *hash_location;
538
	struct sock *child;
539
	const struct tcphdr *th = tcp_hdr(skb);
540
	__be32 flg = tcp_flag_word(th) & (TCP_FLAG_RST|TCP_FLAG_SYN|TCP_FLAG_ACK);
E
Eric Dumazet 已提交
541
	bool paws_reject = false;
L
Linus Torvalds 已提交
542

543 544
	BUG_ON(fastopen == (sk->sk_state == TCP_LISTEN));

545 546
	tmp_opt.saw_tstamp = 0;
	if (th->doff > (sizeof(struct tcphdr)>>2)) {
Y
Yuchung Cheng 已提交
547
		tcp_parse_options(skb, &tmp_opt, &hash_location, 0, NULL);
L
Linus Torvalds 已提交
548 549 550 551 552 553 554

		if (tmp_opt.saw_tstamp) {
			tmp_opt.ts_recent = req->ts_recent;
			/* We do not store true stamp, but it is not required,
			 * it can be estimated (approximately)
			 * from another data.
			 */
555
			tmp_opt.ts_recent_stamp = get_seconds() - ((TCP_TIMEOUT_INIT/HZ)<<req->num_timeout);
I
Ilpo Järvinen 已提交
556
			paws_reject = tcp_paws_reject(&tmp_opt, th->rst);
L
Linus Torvalds 已提交
557 558 559 560
		}
	}

	/* Check for pure retransmitted SYN. */
561
	if (TCP_SKB_CB(skb)->seq == tcp_rsk(req)->rcv_isn &&
L
Linus Torvalds 已提交
562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579
	    flg == TCP_FLAG_SYN &&
	    !paws_reject) {
		/*
		 * RFC793 draws (Incorrectly! It was fixed in RFC1122)
		 * this case on figure 6 and figure 8, but formal
		 * protocol description says NOTHING.
		 * To be more exact, it says that we should send ACK,
		 * because this segment (at least, if it has no data)
		 * is out of window.
		 *
		 *  CONCLUSION: RFC793 (even with RFC1122) DOES NOT
		 *  describe SYN-RECV state. All the description
		 *  is wrong, we cannot believe to it and should
		 *  rely only on common sense and implementation
		 *  experience.
		 *
		 * Enforce "SYN-ACK" according to figure 8, figure 6
		 * of RFC793, fixed by RFC1122.
580 581 582
		 *
		 * Note that even if there is new data in the SYN packet
		 * they will be thrown away too.
L
Linus Torvalds 已提交
583
		 */
584
		inet_rtx_syn_ack(sk, req);
L
Linus Torvalds 已提交
585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616
		return NULL;
	}

	/* Further reproduces section "SEGMENT ARRIVES"
	   for state SYN-RECEIVED of RFC793.
	   It is broken, however, it does not work only
	   when SYNs are crossed.

	   You would think that SYN crossing is impossible here, since
	   we should have a SYN_SENT socket (from connect()) on our end,
	   but this is not true if the crossed SYNs were sent to both
	   ends by a malicious third party.  We must defend against this,
	   and to do that we first verify the ACK (as per RFC793, page
	   36) and reset if it is invalid.  Is this a true full defense?
	   To convince ourselves, let us consider a way in which the ACK
	   test can still pass in this 'malicious crossed SYNs' case.
	   Malicious sender sends identical SYNs (and thus identical sequence
	   numbers) to both A and B:

		A: gets SYN, seq=7
		B: gets SYN, seq=7

	   By our good fortune, both A and B select the same initial
	   send sequence number of seven :-)

		A: sends SYN|ACK, seq=7, ack_seq=8
		B: sends SYN|ACK, seq=7, ack_seq=8

	   So we are now A eating this SYN|ACK, ACK test passes.  So
	   does sequence test, SYN is truncated, and thus we consider
	   it a bare ACK.

617 618 619 620
	   If icsk->icsk_accept_queue.rskq_defer_accept, we silently drop this
	   bare ACK.  Otherwise, we create an established connection.  Both
	   ends (listening sockets) accept the new incoming connection and try
	   to talk to each other. 8-)
L
Linus Torvalds 已提交
621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636

	   Note: This case is both harmless, and rare.  Possibility is about the
	   same as us discovering intelligent life on another plant tomorrow.

	   But generally, we should (RFC lies!) to accept ACK
	   from SYNACK both here and in tcp_rcv_state_process().
	   tcp_rcv_state_process() does not, hence, we do not too.

	   Note that the case is absolutely generic:
	   we cannot optimize anything here without
	   violating protocol. All the checks must be made
	   before attempt to create socket.
	 */

	/* RFC793 page 36: "If the connection is in any non-synchronized state ...
	 *                  and the incoming segment acknowledges something not yet
S
Stephen Hemminger 已提交
637
	 *                  sent (the segment carries an unacceptable ACK) ...
L
Linus Torvalds 已提交
638 639
	 *                  a reset is sent."
	 *
640 641 642 643
	 * Invalid ACK: reset will be sent by listening socket.
	 * Note that the ACK validity check for a Fast Open socket is done
	 * elsewhere and is checked directly against the child socket rather
	 * than req because user data may have been sent out.
L
Linus Torvalds 已提交
644
	 */
645
	if ((flg & TCP_FLAG_ACK) && !fastopen &&
646 647
	    (TCP_SKB_CB(skb)->ack_seq !=
	     tcp_rsk(req)->snt_isn + 1 + tcp_s_data_size(tcp_sk(sk))))
L
Linus Torvalds 已提交
648 649 650 651 652 653 654 655 656 657
		return sk;

	/* Also, it would be not so bad idea to check rcv_tsecr, which
	 * is essentially ACK extension and too early or too late values
	 * should cause reset in unsynchronized states.
	 */

	/* RFC793: "first check sequence number". */

	if (paws_reject || !tcp_in_window(TCP_SKB_CB(skb)->seq, TCP_SKB_CB(skb)->end_seq,
658
					  tcp_rsk(req)->rcv_nxt, tcp_rsk(req)->rcv_nxt + req->rcv_wnd)) {
L
Linus Torvalds 已提交
659 660
		/* Out of window: send ACK and drop. */
		if (!(flg & TCP_FLAG_RST))
661
			req->rsk_ops->send_ack(sk, skb, req);
L
Linus Torvalds 已提交
662
		if (paws_reject)
663
			NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_PAWSESTABREJECTED);
L
Linus Torvalds 已提交
664 665 666 667 668
		return NULL;
	}

	/* In sequence, PAWS is OK. */

669
	if (tmp_opt.saw_tstamp && !after(TCP_SKB_CB(skb)->seq, tcp_rsk(req)->rcv_nxt))
670
		req->ts_recent = tmp_opt.rcv_tsval;
L
Linus Torvalds 已提交
671

672 673 674 675 676
	if (TCP_SKB_CB(skb)->seq == tcp_rsk(req)->rcv_isn) {
		/* Truncate SYN, it is out of window starting
		   at tcp_rsk(req)->rcv_isn + 1. */
		flg &= ~TCP_FLAG_SYN;
	}
L
Linus Torvalds 已提交
677

678 679 680 681 682 683 684
	/* RFC793: "second check the RST bit" and
	 *	   "fourth, check the SYN bit"
	 */
	if (flg & (TCP_FLAG_RST|TCP_FLAG_SYN)) {
		TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_ATTEMPTFAILS);
		goto embryonic_reset;
	}
L
Linus Torvalds 已提交
685

686 687
	/* ACK sequence verified above, just make sure ACK is
	 * set.  If ACK not set, just silently drop the packet.
688 689 690
	 *
	 * XXX (TFO) - if we ever allow "data after SYN", the
	 * following check needs to be removed.
691 692 693
	 */
	if (!(flg & TCP_FLAG_ACK))
		return NULL;
694

695 696 697
	/* Got ACK for our SYNACK, so update baseline for SYNACK RTT sample. */
	if (tmp_opt.saw_tstamp && tmp_opt.rcv_tsecr)
		tcp_rsk(req)->snt_synack = tmp_opt.rcv_tsecr;
698
	else if (req->num_retrans) /* don't take RTT sample if retrans && ~TS */
699 700
		tcp_rsk(req)->snt_synack = 0;

701 702 703 704 705 706
	/* For Fast Open no more processing is needed (sk is the
	 * child socket).
	 */
	if (fastopen)
		return sk;

707
	/* While TCP_DEFER_ACCEPT is active, drop bare ACK. */
708
	if (req->num_timeout < inet_csk(sk)->icsk_accept_queue.rskq_defer_accept &&
709 710
	    TCP_SKB_CB(skb)->end_seq == tcp_rsk(req)->rcv_isn + 1) {
		inet_rsk(req)->acked = 1;
711
		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPDEFERACCEPTDROP);
712 713 714 715 716 717 718 719 720 721 722 723
		return NULL;
	}

	/* OK, ACK is valid, create big socket and
	 * feed this segment to it. It will repeat all
	 * the tests. THIS SEGMENT MUST MOVE SOCKET TO
	 * ESTABLISHED STATE. If it will be dropped after
	 * socket is created, wait for troubles.
	 */
	child = inet_csk(sk)->icsk_af_ops->syn_recv_sock(sk, skb, req, NULL);
	if (child == NULL)
		goto listen_overflow;
L
Linus Torvalds 已提交
724

725 726
	inet_csk_reqsk_queue_unlink(sk, req, prev);
	inet_csk_reqsk_queue_removed(sk, req);
L
Linus Torvalds 已提交
727

728 729
	inet_csk_reqsk_queue_add(sk, req, child);
	return child;
L
Linus Torvalds 已提交
730

731 732 733 734 735
listen_overflow:
	if (!sysctl_tcp_abort_on_overflow) {
		inet_rsk(req)->acked = 1;
		return NULL;
	}
L
Linus Torvalds 已提交
736

737
embryonic_reset:
738 739 740 741 742 743
	if (!(flg & TCP_FLAG_RST)) {
		/* Received a bad SYN pkt - for TFO We try not to reset
		 * the local connection unless it's really necessary to
		 * avoid becoming vulnerable to outside attack aiming at
		 * resetting legit local connections.
		 */
744
		req->rsk_ops->send_reset(sk, skb);
745 746 747 748 749 750 751 752
	} else if (fastopen) { /* received a valid RST pkt */
		reqsk_fastopen_remove(sk, req, true);
		tcp_reset(sk);
	}
	if (!fastopen) {
		inet_csk_reqsk_queue_drop(sk, req, prev);
		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_EMBRYONICRSTS);
	}
753
	return NULL;
L
Linus Torvalds 已提交
754
}
E
Eric Dumazet 已提交
755
EXPORT_SYMBOL(tcp_check_req);
L
Linus Torvalds 已提交
756 757 758 759 760

/*
 * Queue segment on the new socket if the new socket is active,
 * otherwise we just shortcircuit this and continue with
 * the new socket.
761 762 763 764 765 766
 *
 * For the vast majority of cases child->sk_state will be TCP_SYN_RECV
 * when entering. But other states are possible due to a race condition
 * where after __inet_lookup_established() fails but before the listener
 * locked is obtained, other packets cause the same connection to
 * be created.
L
Linus Torvalds 已提交
767 768 769 770 771 772 773 774 775
 */

int tcp_child_process(struct sock *parent, struct sock *child,
		      struct sk_buff *skb)
{
	int ret = 0;
	int state = child->sk_state;

	if (!sock_owned_by_user(child)) {
776 777
		ret = tcp_rcv_state_process(child, skb, tcp_hdr(skb),
					    skb->len);
L
Linus Torvalds 已提交
778 779 780 781 782 783 784 785
		/* Wakeup parent, send SIGIO */
		if (state == TCP_SYN_RECV && child->sk_state != state)
			parent->sk_data_ready(parent, 0);
	} else {
		/* Alas, it is possible again, because we do lookup
		 * in main socket hash table and lock on listening
		 * socket does not protect us more.
		 */
Z
Zhu Yi 已提交
786
		__sk_add_backlog(child, skb);
L
Linus Torvalds 已提交
787 788 789 790 791 792 793
	}

	bh_unlock_sock(child);
	sock_put(child);
	return ret;
}
EXPORT_SYMBOL(tcp_child_process);