associola.c 46.1 KB
Newer Older
1
/* SCTP kernel implementation
L
Linus Torvalds 已提交
2 3 4 5 6 7
 * (C) Copyright IBM Corp. 2001, 2004
 * Copyright (c) 1999-2000 Cisco, Inc.
 * Copyright (c) 1999-2001 Motorola, Inc.
 * Copyright (c) 2001 Intel Corp.
 * Copyright (c) 2001 La Monte H.P. Yarroll
 *
8
 * This file is part of the SCTP kernel implementation
L
Linus Torvalds 已提交
9 10 11
 *
 * This module provides the abstraction for an SCTP association.
 *
12
 * This SCTP implementation is free software;
L
Linus Torvalds 已提交
13 14 15 16 17
 * 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, or (at your option)
 * any later version.
 *
18
 * This SCTP implementation is distributed in the hope that it
L
Linus Torvalds 已提交
19 20 21 22 23 24 25 26 27 28 29 30
 * will be useful, but WITHOUT ANY WARRANTY; without even the implied
 *                 ************************
 * warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
 * See the GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with GNU CC; see the file COPYING.  If not, write to
 * the Free Software Foundation, 59 Temple Place - Suite 330,
 * Boston, MA 02111-1307, USA.
 *
 * Please send any bug reports or fixes you make to the
 * email address(es):
31
 *    lksctp developers <linux-sctp@vger.kernel.org>
L
Linus Torvalds 已提交
32 33 34 35 36 37 38 39 40 41 42 43 44
 *
 * Written or modified by:
 *    La Monte H.P. Yarroll <piggy@acm.org>
 *    Karl Knutson          <karl@athena.chicago.il.us>
 *    Jon Grimm             <jgrimm@us.ibm.com>
 *    Xingang Guo           <xingang.guo@intel.com>
 *    Hui Huang             <hui.huang@nokia.com>
 *    Sridhar Samudrala	    <sri@us.ibm.com>
 *    Daisy Chang	    <daisyc@us.ibm.com>
 *    Ryan Layer	    <rmlayer@us.ibm.com>
 *    Kevin Gao             <kevin.gao@intel.com>
 */

45 46
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

L
Linus Torvalds 已提交
47 48 49 50 51 52 53 54 55 56 57 58
#include <linux/types.h>
#include <linux/fcntl.h>
#include <linux/poll.h>
#include <linux/init.h>

#include <linux/slab.h>
#include <linux/in.h>
#include <net/ipv6.h>
#include <net/sctp/sctp.h>
#include <net/sctp/sm.h>

/* Forward declarations for internal functions. */
D
David Howells 已提交
59
static void sctp_assoc_bh_rcv(struct work_struct *work);
60
static void sctp_assoc_free_asconf_acks(struct sctp_association *asoc);
61
static void sctp_assoc_free_asconf_queue(struct sctp_association *asoc);
L
Linus Torvalds 已提交
62 63 64 65 66 67 68 69

/* 1st Level Abstractions. */

/* Initialize a new association from provided memory. */
static struct sctp_association *sctp_association_init(struct sctp_association *asoc,
					  const struct sctp_endpoint *ep,
					  const struct sock *sk,
					  sctp_scope_t scope,
A
Al Viro 已提交
70
					  gfp_t gfp)
L
Linus Torvalds 已提交
71
{
72
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
73 74
	struct sctp_sock *sp;
	int i;
75 76
	sctp_paramhdr_t *p;
	int err;
L
Linus Torvalds 已提交
77 78 79 80 81 82 83

	/* Retrieve the SCTP per socket area.  */
	sp = sctp_sk((struct sock *)sk);

	/* Discarding const is appropriate here.  */
	asoc->ep = (struct sctp_endpoint *)ep;
	asoc->base.sk = (struct sock *)sk;
84 85

	sctp_endpoint_hold(asoc->ep);
L
Linus Torvalds 已提交
86 87 88 89 90 91 92
	sock_hold(asoc->base.sk);

	/* Initialize the common base substructure.  */
	asoc->base.type = SCTP_EP_TYPE_ASSOCIATION;

	/* Initialize the object handling fields.  */
	atomic_set(&asoc->base.refcnt, 1);
93
	asoc->base.dead = false;
L
Linus Torvalds 已提交
94 95 96 97 98

	/* Initialize the bind addr area.  */
	sctp_bind_addr_init(&asoc->base.bind_addr, ep->base.bind_addr.port);

	asoc->state = SCTP_STATE_CLOSED;
99
	asoc->cookie_life = ms_to_ktime(sp->assocparams.sasoc_cookie_life);
L
Linus Torvalds 已提交
100
	asoc->frag_point = 0;
101
	asoc->user_frag = sp->user_frag;
L
Linus Torvalds 已提交
102 103 104 105 106

	/* Set the association max_retrans and RTO values from the
	 * socket values.
	 */
	asoc->max_retrans = sp->assocparams.sasoc_asocmaxrxt;
107
	asoc->pf_retrans  = net->sctp.pf_retrans;
108

L
Linus Torvalds 已提交
109 110 111 112 113 114
	asoc->rto_initial = msecs_to_jiffies(sp->rtoinfo.srto_initial);
	asoc->rto_max = msecs_to_jiffies(sp->rtoinfo.srto_max);
	asoc->rto_min = msecs_to_jiffies(sp->rtoinfo.srto_min);

	asoc->overall_error_count = 0;

115 116 117 118 119 120 121 122 123 124 125 126 127
	/* Initialize the association's heartbeat interval based on the
	 * sock configured value.
	 */
	asoc->hbinterval = msecs_to_jiffies(sp->hbinterval);

	/* Initialize path max retrans value. */
	asoc->pathmaxrxt = sp->pathmaxrxt;

	/* Initialize default path MTU. */
	asoc->pathmtu = sp->pathmtu;

	/* Set association default SACK delay */
	asoc->sackdelay = msecs_to_jiffies(sp->sackdelay);
128
	asoc->sackfreq = sp->sackfreq;
129 130 131 132 133 134

	/* Set the association default flags controlling
	 * Heartbeat, SACK delay, and Path MTU Discovery.
	 */
	asoc->param_flags = sp->param_flags;

L
Linus Torvalds 已提交
135 136 137
	/* Initialize the maximum mumber of new data packets that can be sent
	 * in a burst.
	 */
138
	asoc->max_burst = sp->max_burst;
L
Linus Torvalds 已提交
139

140 141 142 143 144 145 146 147 148 149 150 151
	/* initialize association timers */
	asoc->timeouts[SCTP_EVENT_TIMEOUT_NONE] = 0;
	asoc->timeouts[SCTP_EVENT_TIMEOUT_T1_COOKIE] = asoc->rto_initial;
	asoc->timeouts[SCTP_EVENT_TIMEOUT_T1_INIT] = asoc->rto_initial;
	asoc->timeouts[SCTP_EVENT_TIMEOUT_T2_SHUTDOWN] = asoc->rto_initial;
	asoc->timeouts[SCTP_EVENT_TIMEOUT_T3_RTX] = 0;
	asoc->timeouts[SCTP_EVENT_TIMEOUT_T4_RTO] = 0;

	/* sctpimpguide Section 2.12.2
	 * If the 'T5-shutdown-guard' timer is used, it SHOULD be set to the
	 * recommended value of 5 times 'RTO.Max'.
	 */
152
	asoc->timeouts[SCTP_EVENT_TIMEOUT_T5_SHUTDOWN_GUARD]
153 154 155
		= 5 * asoc->rto_max;

	asoc->timeouts[SCTP_EVENT_TIMEOUT_HEARTBEAT] = 0;
156
	asoc->timeouts[SCTP_EVENT_TIMEOUT_SACK] = asoc->sackdelay;
157
	asoc->timeouts[SCTP_EVENT_TIMEOUT_AUTOCLOSE] =
158
		min_t(unsigned long, sp->autoclose, net->sctp.max_autoclose) * HZ;
159

160
	/* Initializes the timers */
161 162 163
	for (i = SCTP_EVENT_TIMEOUT_NONE; i < SCTP_NUM_TIMEOUT_TYPES; ++i)
		setup_timer(&asoc->timers[i], sctp_timer_events[i],
				(unsigned long)asoc);
L
Linus Torvalds 已提交
164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185

	/* Pull default initialization values from the sock options.
	 * Note: This assumes that the values have already been
	 * validated in the sock.
	 */
	asoc->c.sinit_max_instreams = sp->initmsg.sinit_max_instreams;
	asoc->c.sinit_num_ostreams  = sp->initmsg.sinit_num_ostreams;
	asoc->max_init_attempts	= sp->initmsg.sinit_max_attempts;

	asoc->max_init_timeo =
		 msecs_to_jiffies(sp->initmsg.sinit_max_init_timeo);

	/* Allocate storage for the ssnmap after the inbound and outbound
	 * streams have been negotiated during Init.
	 */
	asoc->ssnmap = NULL;

	/* Set the local window size for receive.
	 * This is also the rcvbuf space per association.
	 * RFC 6 - A SCTP receiver MUST be able to receive a minimum of
	 * 1500 bytes in one SCTP packet.
	 */
186
	if ((sk->sk_rcvbuf/2) < SCTP_DEFAULT_MINWINDOW)
L
Linus Torvalds 已提交
187 188
		asoc->rwnd = SCTP_DEFAULT_MINWINDOW;
	else
189
		asoc->rwnd = sk->sk_rcvbuf/2;
L
Linus Torvalds 已提交
190 191 192 193

	asoc->a_rwnd = asoc->rwnd;

	asoc->rwnd_over = 0;
194
	asoc->rwnd_press = 0;
L
Linus Torvalds 已提交
195 196 197 198 199 200 201

	/* Use my own max window until I learn something better.  */
	asoc->peer.rwnd = SCTP_DEFAULT_MAXWINDOW;

	/* Set the sndbuf size for transmit.  */
	asoc->sndbuf_used = 0;

202 203 204
	/* Initialize the receive memory counter */
	atomic_set(&asoc->rmem_alloc, 0);

L
Linus Torvalds 已提交
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
	init_waitqueue_head(&asoc->wait);

	asoc->c.my_vtag = sctp_generate_tag(ep);
	asoc->peer.i.init_tag = 0;     /* INIT needs a vtag of 0. */
	asoc->c.peer_vtag = 0;
	asoc->c.my_ttag   = 0;
	asoc->c.peer_ttag = 0;
	asoc->c.my_port = ep->base.bind_addr.port;

	asoc->c.initial_tsn = sctp_generate_tsn(ep);

	asoc->next_tsn = asoc->c.initial_tsn;

	asoc->ctsn_ack_point = asoc->next_tsn - 1;
	asoc->adv_peer_ack_point = asoc->ctsn_ack_point;
	asoc->highest_sacked = asoc->ctsn_ack_point;
	asoc->last_cwr_tsn = asoc->ctsn_ack_point;
	asoc->unack_data = 0;

	/* ADDIP Section 4.1 Asconf Chunk Procedures
	 *
	 * When an endpoint has an ASCONF signaled change to be sent to the
	 * remote endpoint it should do the following:
	 * ...
	 * A2) a serial number should be assigned to the chunk. The serial
	 * number SHOULD be a monotonically increasing number. The serial
	 * numbers SHOULD be initialized at the start of the
	 * association to the same value as the initial TSN.
	 */
	asoc->addip_serial = asoc->c.initial_tsn;

236
	INIT_LIST_HEAD(&asoc->addip_chunk_list);
237
	INIT_LIST_HEAD(&asoc->asconf_ack_list);
L
Linus Torvalds 已提交
238 239 240

	/* Make an empty list of remote transport addresses.  */
	INIT_LIST_HEAD(&asoc->peer.transport_addr_list);
F
Frank Filz 已提交
241
	asoc->peer.transport_count = 0;
L
Linus Torvalds 已提交
242 243 244 245 246 247 248 249 250 251 252 253 254

	/* RFC 2960 5.1 Normal Establishment of an Association
	 *
	 * After the reception of the first data chunk in an
	 * association the endpoint must immediately respond with a
	 * sack to acknowledge the data chunk.  Subsequent
	 * acknowledgements should be done as described in Section
	 * 6.2.
	 *
	 * [We implement this by telling a new association that it
	 * already received one packet.]
	 */
	asoc->peer.sack_needed = 1;
255
	asoc->peer.sack_cnt = 0;
256
	asoc->peer.sack_generation = 1;
L
Linus Torvalds 已提交
257

258 259 260 261
	/* Assume that the peer will tell us if he recognizes ASCONF
	 * as part of INIT exchange.
	 * The sctp_addip_noauth option is there for backward compatibilty
	 * and will revert old behavior.
L
Linus Torvalds 已提交
262
	 */
263
	asoc->peer.asconf_capable = 0;
264
	if (net->sctp.addip_noauth)
265
		asoc->peer.asconf_capable = 1;
266 267
	asoc->asconf_addr_del_pending = NULL;
	asoc->src_out_of_asoc_ok = 0;
268
	asoc->new_transport = NULL;
L
Linus Torvalds 已提交
269 270 271

	/* Create an input queue.  */
	sctp_inq_init(&asoc->base.inqueue);
D
David Howells 已提交
272
	sctp_inq_set_th_handler(&asoc->base.inqueue, sctp_assoc_bh_rcv);
L
Linus Torvalds 已提交
273 274 275 276 277 278 279

	/* Create an output queue.  */
	sctp_outq_init(asoc, &asoc->outqueue);

	if (!sctp_ulpq_init(&asoc->ulpq, asoc))
		goto fail_init;

280
	memset(&asoc->peer.tsn_map, 0, sizeof(struct sctp_tsnmap));
L
Linus Torvalds 已提交
281 282 283 284 285 286 287 288 289

	asoc->need_ecne = 0;

	asoc->assoc_id = 0;

	/* Assume that peer would support both address types unless we are
	 * told otherwise.
	 */
	asoc->peer.ipv4_address = 1;
290 291
	if (asoc->base.sk->sk_family == PF_INET6)
		asoc->peer.ipv6_address = 1;
L
Linus Torvalds 已提交
292 293 294 295 296 297 298 299 300
	INIT_LIST_HEAD(&asoc->asocs);

	asoc->autoclose = sp->autoclose;

	asoc->default_stream = sp->default_stream;
	asoc->default_ppid = sp->default_ppid;
	asoc->default_flags = sp->default_flags;
	asoc->default_context = sp->default_context;
	asoc->default_timetolive = sp->default_timetolive;
301
	asoc->default_rcv_context = sp->default_rcv_context;
L
Linus Torvalds 已提交
302

303 304 305
	/* SCTP_GET_ASSOC_STATS COUNTERS */
	memset(&asoc->stats, 0, sizeof(struct sctp_priv_assoc_stats));

306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329
	/* AUTH related initializations */
	INIT_LIST_HEAD(&asoc->endpoint_shared_keys);
	err = sctp_auth_asoc_copy_shkeys(ep, asoc, gfp);
	if (err)
		goto fail_init;

	asoc->active_key_id = ep->active_key_id;
	asoc->asoc_shared_key = NULL;

	asoc->default_hmac_id = 0;
	/* Save the hmacs and chunks list into this association */
	if (ep->auth_hmacs_list)
		memcpy(asoc->c.auth_hmacs, ep->auth_hmacs_list,
			ntohs(ep->auth_hmacs_list->param_hdr.length));
	if (ep->auth_chunk_list)
		memcpy(asoc->c.auth_chunks, ep->auth_chunk_list,
			ntohs(ep->auth_chunk_list->param_hdr.length));

	/* Get the AUTH random number for this association */
	p = (sctp_paramhdr_t *)asoc->c.auth_random;
	p->type = SCTP_PARAM_RANDOM;
	p->length = htons(sizeof(sctp_paramhdr_t) + SCTP_AUTH_RANDOM_LENGTH);
	get_random_bytes(p+1, SCTP_AUTH_RANDOM_LENGTH);

L
Linus Torvalds 已提交
330 331 332 333
	return asoc;

fail_init:
	sock_put(asoc->base.sk);
334
	sctp_endpoint_put(asoc->ep);
L
Linus Torvalds 已提交
335 336 337 338 339 340
	return NULL;
}

/* Allocate and initialize a new association */
struct sctp_association *sctp_association_new(const struct sctp_endpoint *ep,
					 const struct sock *sk,
A
Alexey Dobriyan 已提交
341
					 sctp_scope_t scope,
A
Al Viro 已提交
342
					 gfp_t gfp)
L
Linus Torvalds 已提交
343 344 345
{
	struct sctp_association *asoc;

346
	asoc = kzalloc(sizeof(*asoc), gfp);
L
Linus Torvalds 已提交
347 348 349 350 351 352 353
	if (!asoc)
		goto fail;

	if (!sctp_association_init(asoc, ep, sk, scope, gfp))
		goto fail_init;

	SCTP_DBG_OBJCNT_INC(assoc);
354 355

	pr_debug("Created asoc %p\n", asoc);
L
Linus Torvalds 已提交
356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374

	return asoc;

fail_init:
	kfree(asoc);
fail:
	return NULL;
}

/* Free this association if possible.  There may still be users, so
 * the actual deallocation may be delayed.
 */
void sctp_association_free(struct sctp_association *asoc)
{
	struct sock *sk = asoc->base.sk;
	struct sctp_transport *transport;
	struct list_head *pos, *temp;
	int i;

375 376 377 378 379 380 381 382 383 384 385 386
	/* Only real associations count against the endpoint, so
	 * don't bother for if this is a temporary association.
	 */
	if (!asoc->temp) {
		list_del(&asoc->asocs);

		/* Decrement the backlog value for a TCP-style listening
		 * socket.
		 */
		if (sctp_style(sk, TCP) && sctp_sstate(sk, LISTENING))
			sk->sk_ack_backlog--;
	}
L
Linus Torvalds 已提交
387 388 389 390

	/* Mark as dead, so other users can know this structure is
	 * going away.
	 */
391
	asoc->base.dead = true;
L
Linus Torvalds 已提交
392 393 394 395 396 397 398 399 400 401

	/* Dispose of any data lying around in the outqueue. */
	sctp_outq_free(&asoc->outqueue);

	/* Dispose of any pending messages for the upper layer. */
	sctp_ulpq_free(&asoc->ulpq);

	/* Dispose of any pending chunks on the inqueue. */
	sctp_inq_free(&asoc->base.inqueue);

402 403
	sctp_tsnmap_free(&asoc->peer.tsn_map);

L
Linus Torvalds 已提交
404 405 406 407 408 409 410 411 412 413 414 415
	/* Free ssnmap storage. */
	sctp_ssnmap_free(asoc->ssnmap);

	/* Clean up the bound address list. */
	sctp_bind_addr_free(&asoc->base.bind_addr);

	/* Do we need to go through all of our timers and
	 * delete them?   To be safe we will try to delete all, but we
	 * should be able to go through and make a guess based
	 * on our state.
	 */
	for (i = SCTP_EVENT_TIMEOUT_NONE; i < SCTP_NUM_TIMEOUT_TYPES; ++i) {
416
		if (del_timer(&asoc->timers[i]))
L
Linus Torvalds 已提交
417 418 419 420
			sctp_association_put(asoc);
	}

	/* Free peer's cached cookie. */
J
Jesper Juhl 已提交
421
	kfree(asoc->peer.cookie);
422 423 424
	kfree(asoc->peer.peer_random);
	kfree(asoc->peer.peer_chunks);
	kfree(asoc->peer.peer_hmacs);
L
Linus Torvalds 已提交
425 426 427 428

	/* Release the transport structures. */
	list_for_each_safe(pos, temp, &asoc->peer.transport_addr_list) {
		transport = list_entry(pos, struct sctp_transport, transports);
429
		list_del_rcu(pos);
L
Linus Torvalds 已提交
430 431 432
		sctp_transport_free(transport);
	}

F
Frank Filz 已提交
433 434
	asoc->peer.transport_count = 0;

435
	sctp_asconf_queue_teardown(asoc);
L
Linus Torvalds 已提交
436

437 438 439 440
	/* Free pending address space being deleted */
	if (asoc->asconf_addr_del_pending != NULL)
		kfree(asoc->asconf_addr_del_pending);

441 442 443 444 445 446
	/* AUTH - Free the endpoint shared keys */
	sctp_auth_destroy_keys(&asoc->endpoint_shared_keys);

	/* AUTH - Free the association shared key */
	sctp_auth_key_put(asoc->asoc_shared_key);

L
Linus Torvalds 已提交
447 448 449 450 451 452
	sctp_association_put(asoc);
}

/* Cleanup and free up an association. */
static void sctp_association_destroy(struct sctp_association *asoc)
{
453 454 455 456
	if (unlikely(!asoc->base.dead)) {
		WARN(1, "Attempt to destroy undead association %p!\n", asoc);
		return;
	}
L
Linus Torvalds 已提交
457 458 459 460 461 462 463 464 465 466

	sctp_endpoint_put(asoc->ep);
	sock_put(asoc->base.sk);

	if (asoc->assoc_id != 0) {
		spin_lock_bh(&sctp_assocs_id_lock);
		idr_remove(&sctp_assocs_id, asoc->assoc_id);
		spin_unlock_bh(&sctp_assocs_id_lock);
	}

467
	WARN_ON(atomic_read(&asoc->rmem_alloc));
468

469 470
	kfree(asoc);
	SCTP_DBG_OBJCNT_DEC(assoc);
L
Linus Torvalds 已提交
471 472 473 474 475 476
}

/* Change the primary destination address for the peer. */
void sctp_assoc_set_primary(struct sctp_association *asoc,
			    struct sctp_transport *transport)
{
477 478 479 480 481 482 483 484 485
	int changeover = 0;

	/* it's a changeover only if we already have a primary path
	 * that we are changing
	 */
	if (asoc->peer.primary_path != NULL &&
	    asoc->peer.primary_path != transport)
		changeover = 1 ;

L
Linus Torvalds 已提交
486 487 488 489 490 491 492 493 494
	asoc->peer.primary_path = transport;

	/* Set a default msg_name for events. */
	memcpy(&asoc->peer.primary_addr, &transport->ipaddr,
	       sizeof(union sctp_addr));

	/* If the primary path is changing, assume that the
	 * user wants to use this new path.
	 */
495 496
	if ((transport->state == SCTP_ACTIVE) ||
	    (transport->state == SCTP_UNKNOWN))
L
Linus Torvalds 已提交
497 498 499 500 501 502 503 504 505 506 507 508
		asoc->peer.active_path = transport;

	/*
	 * SFR-CACC algorithm:
	 * Upon the receipt of a request to change the primary
	 * destination address, on the data structure for the new
	 * primary destination, the sender MUST do the following:
	 *
	 * 1) If CHANGEOVER_ACTIVE is set, then there was a switch
	 * to this destination address earlier. The sender MUST set
	 * CYCLING_CHANGEOVER to indicate that this switch is a
	 * double switch to the same destination address.
509 510 511
	 *
	 * Really, only bother is we have data queued or outstanding on
	 * the association.
L
Linus Torvalds 已提交
512
	 */
513 514 515
	if (!asoc->outqueue.outstanding_bytes && !asoc->outqueue.out_qlen)
		return;

L
Linus Torvalds 已提交
516
	if (transport->cacc.changeover_active)
517
		transport->cacc.cycling_changeover = changeover;
L
Linus Torvalds 已提交
518 519 520 521

	/* 2) The sender MUST set CHANGEOVER_ACTIVE to indicate that
	 * a changeover has occurred.
	 */
522
	transport->cacc.changeover_active = changeover;
L
Linus Torvalds 已提交
523 524 525 526 527 528 529

	/* 3) The sender MUST store the next TSN to be sent in
	 * next_tsn_at_change.
	 */
	transport->cacc.next_tsn_at_change = asoc->next_tsn;
}

F
Frank Filz 已提交
530 531 532 533 534 535 536
/* Remove a transport from an association.  */
void sctp_assoc_rm_peer(struct sctp_association *asoc,
			struct sctp_transport *peer)
{
	struct list_head	*pos;
	struct sctp_transport	*transport;

537 538
	pr_debug("%s: association:%p addr:%pISpc\n",
		 __func__, asoc, &peer->ipaddr.sa);
F
Frank Filz 已提交
539 540 541 542 543 544 545 546

	/* If we are to remove the current retran_path, update it
	 * to the next peer before removing this peer from the list.
	 */
	if (asoc->peer.retran_path == peer)
		sctp_assoc_update_retran_path(asoc);

	/* Remove this peer from the list. */
547
	list_del_rcu(&peer->transports);
F
Frank Filz 已提交
548 549 550 551 552 553 554 555 556 557

	/* Get the first transport of asoc. */
	pos = asoc->peer.transport_addr_list.next;
	transport = list_entry(pos, struct sctp_transport, transports);

	/* Update any entries that match the peer to be deleted. */
	if (asoc->peer.primary_path == peer)
		sctp_assoc_set_primary(asoc, transport);
	if (asoc->peer.active_path == peer)
		asoc->peer.active_path = transport;
558 559
	if (asoc->peer.retran_path == peer)
		asoc->peer.retran_path = transport;
F
Frank Filz 已提交
560 561 562 563 564 565 566 567 568 569 570
	if (asoc->peer.last_data_from == peer)
		asoc->peer.last_data_from = transport;

	/* If we remove the transport an INIT was last sent to, set it to
	 * NULL. Combined with the update of the retran path above, this
	 * will cause the next INIT to be sent to the next available
	 * transport, maintaining the cycle.
	 */
	if (asoc->init_last_sent_to == peer)
		asoc->init_last_sent_to = NULL;

571 572 573 574 575 576 577 578
	/* If we remove the transport an SHUTDOWN was last sent to, set it
	 * to NULL. Combined with the update of the retran path above, this
	 * will cause the next SHUTDOWN to be sent to the next available
	 * transport, maintaining the cycle.
	 */
	if (asoc->shutdown_last_sent_to == peer)
		asoc->shutdown_last_sent_to = NULL;

579 580 581 582 583 584 585
	/* If we remove the transport an ASCONF was last sent to, set it to
	 * NULL.
	 */
	if (asoc->addip_last_asconf &&
	    asoc->addip_last_asconf->transport == peer)
		asoc->addip_last_asconf->transport = NULL;

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
	/* If we have something on the transmitted list, we have to
	 * save it off.  The best place is the active path.
	 */
	if (!list_empty(&peer->transmitted)) {
		struct sctp_transport *active = asoc->peer.active_path;
		struct sctp_chunk *ch;

		/* Reset the transport of each chunk on this list */
		list_for_each_entry(ch, &peer->transmitted,
					transmitted_list) {
			ch->transport = NULL;
			ch->rtt_in_progress = 0;
		}

		list_splice_tail_init(&peer->transmitted,
					&active->transmitted);

		/* Start a T3 timer here in case it wasn't running so
		 * that these migrated packets have a chance to get
		 * retrnasmitted.
		 */
		if (!timer_pending(&active->T3_rtx_timer))
			if (!mod_timer(&active->T3_rtx_timer,
					jiffies + active->rto))
				sctp_transport_hold(active);
	}

F
Frank Filz 已提交
613 614 615 616 617
	asoc->peer.transport_count--;

	sctp_transport_free(peer);
}

L
Linus Torvalds 已提交
618 619 620
/* Add a transport address to an association.  */
struct sctp_transport *sctp_assoc_add_peer(struct sctp_association *asoc,
					   const union sctp_addr *addr,
A
Al Viro 已提交
621
					   const gfp_t gfp,
F
Frank Filz 已提交
622
					   const int peer_state)
L
Linus Torvalds 已提交
623
{
624
	struct net *net = sock_net(asoc->base.sk);
L
Linus Torvalds 已提交
625 626 627 628 629 630 631
	struct sctp_transport *peer;
	struct sctp_sock *sp;
	unsigned short port;

	sp = sctp_sk(asoc->base.sk);

	/* AF_INET and AF_INET6 share common port field. */
632
	port = ntohs(addr->v4.sin_port);
L
Linus Torvalds 已提交
633

634 635
	pr_debug("%s: association:%p addr:%pISpc state:%d\n", __func__,
		 asoc, &addr->sa, peer_state);
F
Frank Filz 已提交
636

L
Linus Torvalds 已提交
637 638 639 640 641 642
	/* Set the port if it has not been set yet.  */
	if (0 == asoc->peer.port)
		asoc->peer.port = port;

	/* Check to see if this is a duplicate. */
	peer = sctp_assoc_lookup_paddr(asoc, addr);
F
Frank Filz 已提交
643
	if (peer) {
644 645 646 647
		/* An UNKNOWN state is only set on transports added by
		 * user in sctp_connectx() call.  Such transports should be
		 * considered CONFIRMED per RFC 4960, Section 5.4.
		 */
648
		if (peer->state == SCTP_UNKNOWN) {
649
			peer->state = SCTP_ACTIVE;
650
		}
L
Linus Torvalds 已提交
651
		return peer;
F
Frank Filz 已提交
652
	}
L
Linus Torvalds 已提交
653

654
	peer = sctp_transport_new(net, addr, gfp);
L
Linus Torvalds 已提交
655 656 657 658 659
	if (!peer)
		return NULL;

	sctp_transport_set_owner(peer, asoc);

660 661 662 663 664 665 666 667
	/* Initialize the peer's heartbeat interval based on the
	 * association configured value.
	 */
	peer->hbinterval = asoc->hbinterval;

	/* Set the path max_retrans.  */
	peer->pathmaxrxt = asoc->pathmaxrxt;

668 669 670
	/* And the partial failure retrnas threshold */
	peer->pf_retrans = asoc->pf_retrans;

671 672 673 674
	/* Initialize the peer's SACK delay timeout based on the
	 * association configured value.
	 */
	peer->sackdelay = asoc->sackdelay;
675
	peer->sackfreq = asoc->sackfreq;
676 677 678 679 680 681

	/* Enable/disable heartbeat, SACK delay, and path MTU discovery
	 * based on association setting.
	 */
	peer->param_flags = asoc->param_flags;

682 683
	sctp_transport_route(peer, NULL, sp);

L
Linus Torvalds 已提交
684
	/* Initialize the pmtu of the transport. */
685 686 687 688 689 690
	if (peer->param_flags & SPP_PMTUD_DISABLE) {
		if (asoc->pathmtu)
			peer->pathmtu = asoc->pathmtu;
		else
			peer->pathmtu = SCTP_DEFAULT_MAXSEGMENT;
	}
L
Linus Torvalds 已提交
691 692 693 694 695 696

	/* If this is the first transport addr on this association,
	 * initialize the association PMTU to the peer's PMTU.
	 * If not and the current association PMTU is higher than the new
	 * peer's PMTU, reset the association PMTU to the new peer's PMTU.
	 */
697 698
	if (asoc->pathmtu)
		asoc->pathmtu = min_t(int, peer->pathmtu, asoc->pathmtu);
L
Linus Torvalds 已提交
699
	else
700
		asoc->pathmtu = peer->pathmtu;
L
Linus Torvalds 已提交
701

702 703 704
	pr_debug("%s: association:%p PMTU set to %d\n", __func__, asoc,
		 asoc->pathmtu);

705
	peer->pmtu_pending = 0;
L
Linus Torvalds 已提交
706

707
	asoc->frag_point = sctp_frag_point(asoc, asoc->pathmtu);
L
Linus Torvalds 已提交
708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724

	/* The asoc->peer.port might not be meaningful yet, but
	 * initialize the packet structure anyway.
	 */
	sctp_packet_init(&peer->packet, peer, asoc->base.bind_addr.port,
			 asoc->peer.port);

	/* 7.2.1 Slow-Start
	 *
	 * o The initial cwnd before DATA transmission or after a sufficiently
	 *   long idle period MUST be set to
	 *      min(4*MTU, max(2*MTU, 4380 bytes))
	 *
	 * o The initial value of ssthresh MAY be arbitrarily high
	 *   (for example, implementations MAY use the size of the
	 *   receiver advertised window).
	 */
725
	peer->cwnd = min(4*asoc->pathmtu, max_t(__u32, 2*asoc->pathmtu, 4380));
L
Linus Torvalds 已提交
726 727 728 729 730 731 732 733 734

	/* At this point, we may not have the receiver's advertised window,
	 * so initialize ssthresh to the default value and it will be set
	 * later when we process the INIT.
	 */
	peer->ssthresh = SCTP_DEFAULT_MAXWINDOW;

	peer->partial_bytes_acked = 0;
	peer->flight_size = 0;
735
	peer->burst_limited = 0;
L
Linus Torvalds 已提交
736 737 738

	/* Set the transport's RTO.initial value */
	peer->rto = asoc->rto_initial;
739
	sctp_max_rto(asoc, peer);
L
Linus Torvalds 已提交
740

F
Frank Filz 已提交
741 742 743
	/* Set the peer's active state. */
	peer->state = peer_state;

L
Linus Torvalds 已提交
744
	/* Attach the remote transport to our asoc.  */
745
	list_add_tail_rcu(&peer->transports, &asoc->peer.transport_addr_list);
F
Frank Filz 已提交
746
	asoc->peer.transport_count++;
L
Linus Torvalds 已提交
747 748 749 750 751 752 753

	/* If we do not yet have a primary path, set one.  */
	if (!asoc->peer.primary_path) {
		sctp_assoc_set_primary(asoc, peer);
		asoc->peer.retran_path = peer;
	}

754 755
	if (asoc->peer.active_path == asoc->peer.retran_path &&
	    peer->state != SCTP_UNCONFIRMED) {
L
Linus Torvalds 已提交
756
		asoc->peer.retran_path = peer;
F
Frank Filz 已提交
757
	}
L
Linus Torvalds 已提交
758 759 760 761 762 763 764 765 766 767 768 769 770 771 772

	return peer;
}

/* Delete a transport address from an association.  */
void sctp_assoc_del_peer(struct sctp_association *asoc,
			 const union sctp_addr *addr)
{
	struct list_head	*pos;
	struct list_head	*temp;
	struct sctp_transport	*transport;

	list_for_each_safe(pos, temp, &asoc->peer.transport_addr_list) {
		transport = list_entry(pos, struct sctp_transport, transports);
		if (sctp_cmp_addr_exact(addr, &transport->ipaddr)) {
F
Frank Filz 已提交
773 774
			/* Do book keeping for removing the peer and free it. */
			sctp_assoc_rm_peer(asoc, transport);
L
Linus Torvalds 已提交
775 776 777 778 779 780 781 782 783 784 785 786 787 788
			break;
		}
	}
}

/* Lookup a transport by address. */
struct sctp_transport *sctp_assoc_lookup_paddr(
					const struct sctp_association *asoc,
					const union sctp_addr *address)
{
	struct sctp_transport *t;

	/* Cycle through all transports searching for a peer address. */

789 790
	list_for_each_entry(t, &asoc->peer.transport_addr_list,
			transports) {
L
Linus Torvalds 已提交
791 792 793 794 795 796 797
		if (sctp_cmp_addr_exact(address, &t->ipaddr))
			return t;
	}

	return NULL;
}

798 799 800 801 802 803 804 805 806 807 808 809 810 811 812
/* Remove all transports except a give one */
void sctp_assoc_del_nonprimary_peers(struct sctp_association *asoc,
				     struct sctp_transport *primary)
{
	struct sctp_transport	*temp;
	struct sctp_transport	*t;

	list_for_each_entry_safe(t, temp, &asoc->peer.transport_addr_list,
				 transports) {
		/* if the current transport is not the primary one, delete it */
		if (t != primary)
			sctp_assoc_rm_peer(asoc, t);
	}
}

L
Linus Torvalds 已提交
813 814 815 816 817 818 819 820 821 822 823 824 825
/* Engage in transport control operations.
 * Mark the transport up or down and send a notification to the user.
 * Select and update the new active and retran paths.
 */
void sctp_assoc_control_transport(struct sctp_association *asoc,
				  struct sctp_transport *transport,
				  sctp_transport_cmd_t command,
				  sctp_sn_error_t error)
{
	struct sctp_transport *t = NULL;
	struct sctp_transport *first;
	struct sctp_transport *second;
	struct sctp_ulpevent *event;
826
	struct sockaddr_storage addr;
L
Linus Torvalds 已提交
827
	int spc_state = 0;
828
	bool ulp_notify = true;
L
Linus Torvalds 已提交
829 830 831 832

	/* Record the transition on the transport.  */
	switch (command) {
	case SCTP_TRANSPORT_UP:
833 834 835 836 837 838 839 840 841
		/* If we are moving from UNCONFIRMED state due
		 * to heartbeat success, report the SCTP_ADDR_CONFIRMED
		 * state to the user, otherwise report SCTP_ADDR_AVAILABLE.
		 */
		if (SCTP_UNCONFIRMED == transport->state &&
		    SCTP_HEARTBEAT_SUCCESS == error)
			spc_state = SCTP_ADDR_CONFIRMED;
		else
			spc_state = SCTP_ADDR_AVAILABLE;
842 843 844 845 846 847 848 849
		/* Don't inform ULP about transition from PF to
		 * active state and set cwnd to 1, see SCTP
		 * Quick failover draft section 5.1, point 5
		 */
		if (transport->state == SCTP_PF) {
			ulp_notify = false;
			transport->cwnd = 1;
		}
F
Frank Filz 已提交
850
		transport->state = SCTP_ACTIVE;
L
Linus Torvalds 已提交
851 852 853
		break;

	case SCTP_TRANSPORT_DOWN:
854 855 856
		/* If the transport was never confirmed, do not transition it
		 * to inactive state.  Also, release the cached route since
		 * there may be a better route next time.
857 858 859
		 */
		if (transport->state != SCTP_UNCONFIRMED)
			transport->state = SCTP_INACTIVE;
860 861 862 863
		else {
			dst_release(transport->dst);
			transport->dst = NULL;
		}
864

L
Linus Torvalds 已提交
865 866 867
		spc_state = SCTP_ADDR_UNREACHABLE;
		break;

868 869 870 871 872
	case SCTP_TRANSPORT_PF:
		transport->state = SCTP_PF;
		ulp_notify = false;
		break;

L
Linus Torvalds 已提交
873 874
	default:
		return;
875
	}
L
Linus Torvalds 已提交
876 877 878 879

	/* Generate and send a SCTP_PEER_ADDR_CHANGE notification to the
	 * user.
	 */
880 881 882 883 884 885 886 887 888
	if (ulp_notify) {
		memset(&addr, 0, sizeof(struct sockaddr_storage));
		memcpy(&addr, &transport->ipaddr,
		       transport->af_specific->sockaddr_len);
		event = sctp_ulpevent_make_peer_addr_change(asoc, &addr,
					0, spc_state, error, GFP_ATOMIC);
		if (event)
			sctp_ulpq_tail_event(&asoc->ulpq, event);
	}
L
Linus Torvalds 已提交
889 890 891 892 893 894 895 896 897 898 899

	/* Select new active and retran paths. */

	/* Look for the two most recently used active transports.
	 *
	 * This code produces the wrong ordering whenever jiffies
	 * rolls over, but we still get usable transports, so we don't
	 * worry about it.
	 */
	first = NULL; second = NULL;

900 901
	list_for_each_entry(t, &asoc->peer.transport_addr_list,
			transports) {
L
Linus Torvalds 已提交
902

903
		if ((t->state == SCTP_INACTIVE) ||
904 905
		    (t->state == SCTP_UNCONFIRMED) ||
		    (t->state == SCTP_PF))
L
Linus Torvalds 已提交
906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924
			continue;
		if (!first || t->last_time_heard > first->last_time_heard) {
			second = first;
			first = t;
		}
		if (!second || t->last_time_heard > second->last_time_heard)
			second = t;
	}

	/* RFC 2960 6.4 Multi-Homed SCTP Endpoints
	 *
	 * By default, an endpoint should always transmit to the
	 * primary path, unless the SCTP user explicitly specifies the
	 * destination transport address (and possibly source
	 * transport address) to use.
	 *
	 * [If the primary is active but not most recent, bump the most
	 * recently used transport.]
	 */
925 926
	if (((asoc->peer.primary_path->state == SCTP_ACTIVE) ||
	     (asoc->peer.primary_path->state == SCTP_UNKNOWN)) &&
L
Linus Torvalds 已提交
927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020
	    first != asoc->peer.primary_path) {
		second = first;
		first = asoc->peer.primary_path;
	}

	/* If we failed to find a usable transport, just camp on the
	 * primary, even if it is inactive.
	 */
	if (!first) {
		first = asoc->peer.primary_path;
		second = asoc->peer.primary_path;
	}

	/* Set the active and retran transports.  */
	asoc->peer.active_path = first;
	asoc->peer.retran_path = second;
}

/* Hold a reference to an association. */
void sctp_association_hold(struct sctp_association *asoc)
{
	atomic_inc(&asoc->base.refcnt);
}

/* Release a reference to an association and cleanup
 * if there are no more references.
 */
void sctp_association_put(struct sctp_association *asoc)
{
	if (atomic_dec_and_test(&asoc->base.refcnt))
		sctp_association_destroy(asoc);
}

/* Allocate the next TSN, Transmission Sequence Number, for the given
 * association.
 */
__u32 sctp_association_get_next_tsn(struct sctp_association *asoc)
{
	/* From Section 1.6 Serial Number Arithmetic:
	 * Transmission Sequence Numbers wrap around when they reach
	 * 2**32 - 1.  That is, the next TSN a DATA chunk MUST use
	 * after transmitting TSN = 2*32 - 1 is TSN = 0.
	 */
	__u32 retval = asoc->next_tsn;
	asoc->next_tsn++;
	asoc->unack_data++;

	return retval;
}

/* Compare two addresses to see if they match.  Wildcard addresses
 * only match themselves.
 */
int sctp_cmp_addr_exact(const union sctp_addr *ss1,
			const union sctp_addr *ss2)
{
	struct sctp_af *af;

	af = sctp_get_af_specific(ss1->sa.sa_family);
	if (unlikely(!af))
		return 0;

	return af->cmp_addr(ss1, ss2);
}

/* Return an ecne chunk to get prepended to a packet.
 * Note:  We are sly and return a shared, prealloced chunk.  FIXME:
 * No we don't, but we could/should.
 */
struct sctp_chunk *sctp_get_ecne_prepend(struct sctp_association *asoc)
{
	struct sctp_chunk *chunk;

	/* Send ECNE if needed.
	 * Not being able to allocate a chunk here is not deadly.
	 */
	if (asoc->need_ecne)
		chunk = sctp_make_ecne(asoc, asoc->last_ecne_tsn);
	else
		chunk = NULL;

	return chunk;
}

/*
 * Find which transport this TSN was sent on.
 */
struct sctp_transport *sctp_assoc_lookup_tsn(struct sctp_association *asoc,
					     __u32 tsn)
{
	struct sctp_transport *active;
	struct sctp_transport *match;
	struct sctp_transport *transport;
	struct sctp_chunk *chunk;
1021
	__be32 key = htonl(tsn);
L
Linus Torvalds 已提交
1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041

	match = NULL;

	/*
	 * FIXME: In general, find a more efficient data structure for
	 * searching.
	 */

	/*
	 * The general strategy is to search each transport's transmitted
	 * list.   Return which transport this TSN lives on.
	 *
	 * Let's be hopeful and check the active_path first.
	 * Another optimization would be to know if there is only one
	 * outbound path and not have to look for the TSN at all.
	 *
	 */

	active = asoc->peer.active_path;

1042 1043
	list_for_each_entry(chunk, &active->transmitted,
			transmitted_list) {
L
Linus Torvalds 已提交
1044 1045 1046 1047 1048 1049 1050 1051

		if (key == chunk->subh.data_hdr->tsn) {
			match = active;
			goto out;
		}
	}

	/* If not found, go search all the other transports. */
1052 1053
	list_for_each_entry(transport, &asoc->peer.transport_addr_list,
			transports) {
L
Linus Torvalds 已提交
1054 1055

		if (transport == active)
1056
			continue;
1057 1058
		list_for_each_entry(chunk, &transport->transmitted,
				transmitted_list) {
L
Linus Torvalds 已提交
1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070
			if (key == chunk->subh.data_hdr->tsn) {
				match = transport;
				goto out;
			}
		}
	}
out:
	return match;
}

/* Is this the association we are looking for? */
struct sctp_transport *sctp_assoc_is_match(struct sctp_association *asoc,
1071
					   struct net *net,
L
Linus Torvalds 已提交
1072 1073 1074 1075 1076
					   const union sctp_addr *laddr,
					   const union sctp_addr *paddr)
{
	struct sctp_transport *transport;

1077
	if ((htons(asoc->base.bind_addr.port) == laddr->v4.sin_port) &&
1078 1079
	    (htons(asoc->peer.port) == paddr->v4.sin_port) &&
	    net_eq(sock_net(asoc->base.sk), net)) {
L
Linus Torvalds 已提交
1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094
		transport = sctp_assoc_lookup_paddr(asoc, paddr);
		if (!transport)
			goto out;

		if (sctp_bind_addr_match(&asoc->base.bind_addr, laddr,
					 sctp_sk(asoc->base.sk)))
			goto out;
	}
	transport = NULL;

out:
	return transport;
}

/* Do delayed input processing.  This is scheduled by sctp_rcv(). */
D
David Howells 已提交
1095
static void sctp_assoc_bh_rcv(struct work_struct *work)
L
Linus Torvalds 已提交
1096
{
D
David Howells 已提交
1097 1098 1099
	struct sctp_association *asoc =
		container_of(work, struct sctp_association,
			     base.inqueue.immediate);
1100
	struct net *net = sock_net(asoc->base.sk);
L
Linus Torvalds 已提交
1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116
	struct sctp_endpoint *ep;
	struct sctp_chunk *chunk;
	struct sctp_inq *inqueue;
	int state;
	sctp_subtype_t subtype;
	int error = 0;

	/* The association should be held so we should be safe. */
	ep = asoc->ep;

	inqueue = &asoc->base.inqueue;
	sctp_association_hold(asoc);
	while (NULL != (chunk = sctp_inq_pop(inqueue))) {
		state = asoc->state;
		subtype = SCTP_ST_CHUNK(chunk->chunk_hdr->type);

1117 1118 1119 1120 1121 1122 1123 1124 1125 1126
		/* SCTP-AUTH, Section 6.3:
		 *    The receiver has a list of chunk types which it expects
		 *    to be received only after an AUTH-chunk.  This list has
		 *    been sent to the peer during the association setup.  It
		 *    MUST silently discard these chunks if they are not placed
		 *    after an AUTH chunk in the packet.
		 */
		if (sctp_auth_recv_cid(subtype.chunk, asoc) && !chunk->auth)
			continue;

L
Linus Torvalds 已提交
1127 1128 1129 1130 1131
		/* Remember where the last DATA chunk came from so we
		 * know where to send the SACK.
		 */
		if (sctp_chunk_is_data(chunk))
			asoc->peer.last_data_from = chunk->transport;
1132
		else {
1133
			SCTP_INC_STATS(net, SCTP_MIB_INCTRLCHUNKS);
1134 1135 1136 1137
			asoc->stats.ictrlchunks++;
			if (chunk->chunk_hdr->type == SCTP_CID_SACK)
				asoc->stats.isacks++;
		}
L
Linus Torvalds 已提交
1138 1139 1140 1141 1142

		if (chunk->transport)
			chunk->transport->last_time_heard = jiffies;

		/* Run through the state machine. */
1143
		error = sctp_do_sm(net, SCTP_EVENT_T_CHUNK, subtype,
L
Linus Torvalds 已提交
1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201
				   state, ep, asoc, chunk, GFP_ATOMIC);

		/* Check to see if the association is freed in response to
		 * the incoming chunk.  If so, get out of the while loop.
		 */
		if (asoc->base.dead)
			break;

		/* If there is an error on chunk, discard this packet. */
		if (error && chunk)
			chunk->pdiscard = 1;
	}
	sctp_association_put(asoc);
}

/* This routine moves an association from its old sk to a new sk.  */
void sctp_assoc_migrate(struct sctp_association *assoc, struct sock *newsk)
{
	struct sctp_sock *newsp = sctp_sk(newsk);
	struct sock *oldsk = assoc->base.sk;

	/* Delete the association from the old endpoint's list of
	 * associations.
	 */
	list_del_init(&assoc->asocs);

	/* Decrement the backlog value for a TCP-style socket. */
	if (sctp_style(oldsk, TCP))
		oldsk->sk_ack_backlog--;

	/* Release references to the old endpoint and the sock.  */
	sctp_endpoint_put(assoc->ep);
	sock_put(assoc->base.sk);

	/* Get a reference to the new endpoint.  */
	assoc->ep = newsp->ep;
	sctp_endpoint_hold(assoc->ep);

	/* Get a reference to the new sock.  */
	assoc->base.sk = newsk;
	sock_hold(assoc->base.sk);

	/* Add the association to the new endpoint's list of associations.  */
	sctp_endpoint_add_asoc(newsp->ep, assoc);
}

/* Update an association (possibly from unexpected COOKIE-ECHO processing).  */
void sctp_assoc_update(struct sctp_association *asoc,
		       struct sctp_association *new)
{
	struct sctp_transport *trans;
	struct list_head *pos, *temp;

	/* Copy in new parameters of peer. */
	asoc->c = new->c;
	asoc->peer.rwnd = new->peer.rwnd;
	asoc->peer.sack_needed = new->peer.sack_needed;
	asoc->peer.i = new->peer.i;
1202 1203
	sctp_tsnmap_init(&asoc->peer.tsn_map, SCTP_TSN_MAP_INITIAL,
			 asoc->peer.i.initial_tsn, GFP_ATOMIC);
L
Linus Torvalds 已提交
1204 1205 1206 1207

	/* Remove any peer addresses not present in the new association. */
	list_for_each_safe(pos, temp, &asoc->peer.transport_addr_list) {
		trans = list_entry(pos, struct sctp_transport, transports);
1208 1209 1210 1211
		if (!sctp_assoc_lookup_paddr(new, &trans->ipaddr)) {
			sctp_assoc_rm_peer(asoc, trans);
			continue;
		}
1212 1213 1214

		if (asoc->state >= SCTP_STATE_ESTABLISHED)
			sctp_transport_reset(trans);
L
Linus Torvalds 已提交
1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231
	}

	/* If the case is A (association restart), use
	 * initial_tsn as next_tsn. If the case is B, use
	 * current next_tsn in case data sent to peer
	 * has been discarded and needs retransmission.
	 */
	if (asoc->state >= SCTP_STATE_ESTABLISHED) {
		asoc->next_tsn = new->next_tsn;
		asoc->ctsn_ack_point = new->ctsn_ack_point;
		asoc->adv_peer_ack_point = new->adv_peer_ack_point;

		/* Reinitialize SSN for both local streams
		 * and peer's streams.
		 */
		sctp_ssnmap_clear(asoc->ssnmap);

1232 1233 1234 1235 1236 1237
		/* Flush the ULP reassembly and ordered queue.
		 * Any data there will now be stale and will
		 * cause problems.
		 */
		sctp_ulpq_flush(&asoc->ulpq);

1238 1239 1240 1241 1242 1243
		/* reset the overall association error count so
		 * that the restarted association doesn't get torn
		 * down on the next retransmission timer.
		 */
		asoc->overall_error_count = 0;

L
Linus Torvalds 已提交
1244 1245
	} else {
		/* Add any peer addresses from the new association. */
1246 1247
		list_for_each_entry(trans, &new->peer.transport_addr_list,
				transports) {
L
Linus Torvalds 已提交
1248 1249
			if (!sctp_assoc_lookup_paddr(asoc, &trans->ipaddr))
				sctp_assoc_add_peer(asoc, &trans->ipaddr,
1250
						    GFP_ATOMIC, trans->state);
L
Linus Torvalds 已提交
1251 1252 1253 1254 1255 1256 1257 1258 1259
		}

		asoc->ctsn_ack_point = asoc->next_tsn - 1;
		asoc->adv_peer_ack_point = asoc->ctsn_ack_point;
		if (!asoc->ssnmap) {
			/* Move the ssnmap. */
			asoc->ssnmap = new->ssnmap;
			new->ssnmap = NULL;
		}
1260 1261 1262 1263 1264 1265 1266

		if (!asoc->assoc_id) {
			/* get a new association id since we don't have one
			 * yet.
			 */
			sctp_assoc_set_id(asoc, GFP_ATOMIC);
		}
L
Linus Torvalds 已提交
1267
	}
1268

1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285
	/* SCTP-AUTH: Save the peer parameters from the new assocaitions
	 * and also move the association shared keys over
	 */
	kfree(asoc->peer.peer_random);
	asoc->peer.peer_random = new->peer.peer_random;
	new->peer.peer_random = NULL;

	kfree(asoc->peer.peer_chunks);
	asoc->peer.peer_chunks = new->peer.peer_chunks;
	new->peer.peer_chunks = NULL;

	kfree(asoc->peer.peer_hmacs);
	asoc->peer.peer_hmacs = new->peer.peer_hmacs;
	new->peer.peer_hmacs = NULL;

	sctp_auth_key_put(asoc->asoc_shared_key);
	sctp_auth_asoc_init_active_key(asoc, GFP_ATOMIC);
L
Linus Torvalds 已提交
1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298
}

/* Update the retran path for sending a retransmitted packet.
 * Round-robin through the active transports, else round-robin
 * through the inactive transports as this is the next best thing
 * we can try.
 */
void sctp_assoc_update_retran_path(struct sctp_association *asoc)
{
	struct sctp_transport *t, *next;
	struct list_head *head = &asoc->peer.transport_addr_list;
	struct list_head *pos;

G
Gui Jianfeng 已提交
1299 1300 1301
	if (asoc->peer.transport_count == 1)
		return;

L
Linus Torvalds 已提交
1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315
	/* Find the next transport in a round-robin fashion. */
	t = asoc->peer.retran_path;
	pos = &t->transports;
	next = NULL;

	while (1) {
		/* Skip the head. */
		if (pos->next == head)
			pos = head->next;
		else
			pos = pos->next;

		t = list_entry(pos, struct sctp_transport, transports);

G
Gui Jianfeng 已提交
1316 1317 1318 1319 1320 1321 1322 1323 1324
		/* We have exhausted the list, but didn't find any
		 * other active transports.  If so, use the next
		 * transport.
		 */
		if (t == asoc->peer.retran_path) {
			t = next;
			break;
		}

L
Linus Torvalds 已提交
1325 1326
		/* Try to find an active transport. */

1327 1328
		if ((t->state == SCTP_ACTIVE) ||
		    (t->state == SCTP_UNKNOWN)) {
L
Linus Torvalds 已提交
1329 1330 1331 1332 1333
			break;
		} else {
			/* Keep track of the next transport in case
			 * we don't find any active transport.
			 */
1334
			if (t->state != SCTP_UNCONFIRMED && !next)
L
Linus Torvalds 已提交
1335 1336 1337 1338
				next = t;
		}
	}

1339 1340
	if (t)
		asoc->peer.retran_path = t;
1341 1342
	else
		t = asoc->peer.retran_path;
F
Frank Filz 已提交
1343

1344 1345
	pr_debug("%s: association:%p addr:%pISpc\n", __func__, asoc,
		 &t->ipaddr.sa);
F
Frank Filz 已提交
1346 1347
}

1348 1349 1350
/* Choose the transport for sending retransmit packet.  */
struct sctp_transport *sctp_assoc_choose_alter_transport(
	struct sctp_association *asoc, struct sctp_transport *last_sent_to)
F
Frank Filz 已提交
1351
{
1352 1353
	/* If this is the first time packet is sent, use the active path,
	 * else use the retran path. If the last packet was sent over the
F
Frank Filz 已提交
1354 1355
	 * retran path, update the retran path and use it.
	 */
1356
	if (!last_sent_to)
L
Linus Torvalds 已提交
1357 1358
		return asoc->peer.active_path;
	else {
1359
		if (last_sent_to == asoc->peer.retran_path)
L
Linus Torvalds 已提交
1360 1361 1362 1363 1364 1365 1366 1367
			sctp_assoc_update_retran_path(asoc);
		return asoc->peer.retran_path;
	}
}

/* Update the association's pmtu and frag_point by going through all the
 * transports. This routine is called when a transport's PMTU has changed.
 */
1368
void sctp_assoc_sync_pmtu(struct sock *sk, struct sctp_association *asoc)
L
Linus Torvalds 已提交
1369 1370 1371 1372 1373 1374 1375 1376
{
	struct sctp_transport *t;
	__u32 pmtu = 0;

	if (!asoc)
		return;

	/* Get the lowest pmtu of all the transports. */
1377 1378
	list_for_each_entry(t, &asoc->peer.transport_addr_list,
				transports) {
1379
		if (t->pmtu_pending && t->dst) {
1380
			sctp_transport_update_pmtu(sk, t, dst_mtu(t->dst));
1381 1382
			t->pmtu_pending = 0;
		}
1383 1384
		if (!pmtu || (t->pathmtu < pmtu))
			pmtu = t->pathmtu;
L
Linus Torvalds 已提交
1385 1386 1387
	}

	if (pmtu) {
1388
		asoc->pathmtu = pmtu;
1389
		asoc->frag_point = sctp_frag_point(asoc, pmtu);
L
Linus Torvalds 已提交
1390 1391
	}

1392 1393
	pr_debug("%s: asoc:%p, pmtu:%d, frag_point:%d\n", __func__, asoc,
		 asoc->pathmtu, asoc->frag_point);
L
Linus Torvalds 已提交
1394 1395 1396 1397 1398
}

/* Should we send a SACK to update our peer? */
static inline int sctp_peer_needs_update(struct sctp_association *asoc)
{
1399
	struct net *net = sock_net(asoc->base.sk);
L
Linus Torvalds 已提交
1400 1401 1402 1403 1404 1405
	switch (asoc->state) {
	case SCTP_STATE_ESTABLISHED:
	case SCTP_STATE_SHUTDOWN_PENDING:
	case SCTP_STATE_SHUTDOWN_RECEIVED:
	case SCTP_STATE_SHUTDOWN_SENT:
		if ((asoc->rwnd > asoc->a_rwnd) &&
1406
		    ((asoc->rwnd - asoc->a_rwnd) >= max_t(__u32,
1407
			   (asoc->base.sk->sk_rcvbuf >> net->sctp.rwnd_upd_shift),
1408
			   asoc->pathmtu)))
L
Linus Torvalds 已提交
1409 1410 1411 1412 1413 1414 1415 1416 1417
			return 1;
		break;
	default:
		break;
	}
	return 0;
}

/* Increase asoc's rwnd by len and send any window update SACK if needed. */
1418
void sctp_assoc_rwnd_increase(struct sctp_association *asoc, unsigned int len)
L
Linus Torvalds 已提交
1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433
{
	struct sctp_chunk *sack;
	struct timer_list *timer;

	if (asoc->rwnd_over) {
		if (asoc->rwnd_over >= len) {
			asoc->rwnd_over -= len;
		} else {
			asoc->rwnd += (len - asoc->rwnd_over);
			asoc->rwnd_over = 0;
		}
	} else {
		asoc->rwnd += len;
	}

1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444
	/* If we had window pressure, start recovering it
	 * once our rwnd had reached the accumulated pressure
	 * threshold.  The idea is to recover slowly, but up
	 * to the initial advertised window.
	 */
	if (asoc->rwnd_press && asoc->rwnd >= asoc->rwnd_press) {
		int change = min(asoc->pathmtu, asoc->rwnd_press);
		asoc->rwnd += change;
		asoc->rwnd_press -= change;
	}

1445 1446 1447
	pr_debug("%s: asoc:%p rwnd increased by %d to (%u, %u) - %u\n",
		 __func__, asoc, len, asoc->rwnd, asoc->rwnd_over,
		 asoc->a_rwnd);
L
Linus Torvalds 已提交
1448 1449 1450 1451 1452 1453 1454 1455

	/* Send a window update SACK if the rwnd has increased by at least the
	 * minimum of the association's PMTU and half of the receive buffer.
	 * The algorithm used is similar to the one described in
	 * Section 4.2.3.3 of RFC 1122.
	 */
	if (sctp_peer_needs_update(asoc)) {
		asoc->a_rwnd = asoc->rwnd;
1456 1457 1458 1459 1460

		pr_debug("%s: sending window update SACK- asoc:%p rwnd:%u "
			 "a_rwnd:%u\n", __func__, asoc, asoc->rwnd,
			 asoc->a_rwnd);

L
Linus Torvalds 已提交
1461 1462 1463 1464 1465 1466 1467 1468 1469 1470
		sack = sctp_make_sack(asoc);
		if (!sack)
			return;

		asoc->peer.sack_needed = 0;

		sctp_outq_tail(&asoc->outqueue, sack);

		/* Stop the SACK timer.  */
		timer = &asoc->timers[SCTP_EVENT_TIMEOUT_SACK];
1471
		if (del_timer(timer))
L
Linus Torvalds 已提交
1472 1473 1474 1475 1476
			sctp_association_put(asoc);
	}
}

/* Decrease asoc's rwnd by len. */
1477
void sctp_assoc_rwnd_decrease(struct sctp_association *asoc, unsigned int len)
L
Linus Torvalds 已提交
1478
{
1479 1480 1481
	int rx_count;
	int over = 0;

1482 1483 1484 1485
	if (unlikely(!asoc->rwnd || asoc->rwnd_over))
		pr_debug("%s: association:%p has asoc->rwnd:%u, "
			 "asoc->rwnd_over:%u!\n", __func__, asoc,
			 asoc->rwnd, asoc->rwnd_over);
1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499

	if (asoc->ep->rcvbuf_policy)
		rx_count = atomic_read(&asoc->rmem_alloc);
	else
		rx_count = atomic_read(&asoc->base.sk->sk_rmem_alloc);

	/* If we've reached or overflowed our receive buffer, announce
	 * a 0 rwnd if rwnd would still be positive.  Store the
	 * the pottential pressure overflow so that the window can be restored
	 * back to original value.
	 */
	if (rx_count >= asoc->base.sk->sk_rcvbuf)
		over = 1;

L
Linus Torvalds 已提交
1500 1501
	if (asoc->rwnd >= len) {
		asoc->rwnd -= len;
1502
		if (over) {
1503
			asoc->rwnd_press += asoc->rwnd;
1504 1505
			asoc->rwnd = 0;
		}
L
Linus Torvalds 已提交
1506 1507 1508 1509
	} else {
		asoc->rwnd_over = len - asoc->rwnd;
		asoc->rwnd = 0;
	}
1510 1511 1512 1513

	pr_debug("%s: asoc:%p rwnd decreased by %d to (%u, %u, %u)\n",
		 __func__, asoc, len, asoc->rwnd, asoc->rwnd_over,
		 asoc->rwnd_press);
L
Linus Torvalds 已提交
1514 1515 1516 1517 1518
}

/* Build the bind address list for the association based on info from the
 * local endpoint and the remote peer.
 */
A
Alexey Dobriyan 已提交
1519
int sctp_assoc_set_bind_addr_from_ep(struct sctp_association *asoc,
1520
				     sctp_scope_t scope, gfp_t gfp)
L
Linus Torvalds 已提交
1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532
{
	int flags;

	/* Use scoping rules to determine the subset of addresses from
	 * the endpoint.
	 */
	flags = (PF_INET6 == asoc->base.sk->sk_family) ? SCTP_ADDR6_ALLOWED : 0;
	if (asoc->peer.ipv4_address)
		flags |= SCTP_ADDR4_PEERSUPP;
	if (asoc->peer.ipv6_address)
		flags |= SCTP_ADDR6_PEERSUPP;

1533 1534
	return sctp_bind_addr_copy(sock_net(asoc->base.sk),
				   &asoc->base.bind_addr,
L
Linus Torvalds 已提交
1535 1536 1537 1538 1539 1540
				   &asoc->ep->base.bind_addr,
				   scope, gfp, flags);
}

/* Build the association's bind address list from the cookie.  */
int sctp_assoc_set_bind_addr_from_cookie(struct sctp_association *asoc,
A
Alexey Dobriyan 已提交
1541
					 struct sctp_cookie *cookie,
A
Al Viro 已提交
1542
					 gfp_t gfp)
L
Linus Torvalds 已提交
1543 1544 1545 1546 1547 1548 1549 1550 1551
{
	int var_size2 = ntohs(cookie->peer_init->chunk_hdr.length);
	int var_size3 = cookie->raw_addr_list_len;
	__u8 *raw = (__u8 *)cookie->peer_init + var_size2;

	return sctp_raw_to_bind_addrs(&asoc->base.bind_addr, raw, var_size3,
				      asoc->ep->base.bind_addr.port, gfp);
}

1552 1553
/* Lookup laddr in the bind address list of an association. */
int sctp_assoc_lookup_laddr(struct sctp_association *asoc,
L
Linus Torvalds 已提交
1554 1555
			    const union sctp_addr *laddr)
{
1556
	int found = 0;
L
Linus Torvalds 已提交
1557 1558 1559

	if ((asoc->base.bind_addr.port == ntohs(laddr->v4.sin_port)) &&
	    sctp_bind_addr_match(&asoc->base.bind_addr, laddr,
1560
				 sctp_sk(asoc->base.sk)))
L
Linus Torvalds 已提交
1561 1562 1563 1564
		found = 1;

	return found;
}
1565 1566 1567 1568

/* Set an association id for a given association */
int sctp_assoc_set_id(struct sctp_association *asoc, gfp_t gfp)
{
T
Tejun Heo 已提交
1569 1570
	bool preload = gfp & __GFP_WAIT;
	int ret;
1571 1572 1573

	/* If the id is already assigned, keep it. */
	if (asoc->assoc_id)
T
Tejun Heo 已提交
1574
		return 0;
1575

T
Tejun Heo 已提交
1576 1577
	if (preload)
		idr_preload(gfp);
1578
	spin_lock_bh(&sctp_assocs_id_lock);
1579 1580
	/* 0 is not a valid assoc_id, must be >= 1 */
	ret = idr_alloc_cyclic(&sctp_assocs_id, asoc, 1, 0, GFP_NOWAIT);
1581
	spin_unlock_bh(&sctp_assocs_id_lock);
T
Tejun Heo 已提交
1582 1583 1584 1585
	if (preload)
		idr_preload_end();
	if (ret < 0)
		return ret;
1586

T
Tejun Heo 已提交
1587 1588
	asoc->assoc_id = (sctp_assoc_t)ret;
	return 0;
1589
}
1590

1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602
/* Free the ASCONF queue */
static void sctp_assoc_free_asconf_queue(struct sctp_association *asoc)
{
	struct sctp_chunk *asconf;
	struct sctp_chunk *tmp;

	list_for_each_entry_safe(asconf, tmp, &asoc->addip_chunk_list, list) {
		list_del_init(&asconf->list);
		sctp_chunk_free(asconf);
	}
}

1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621
/* Free asconf_ack cache */
static void sctp_assoc_free_asconf_acks(struct sctp_association *asoc)
{
	struct sctp_chunk *ack;
	struct sctp_chunk *tmp;

	list_for_each_entry_safe(ack, tmp, &asoc->asconf_ack_list,
				transmitted_list) {
		list_del_init(&ack->transmitted_list);
		sctp_chunk_free(ack);
	}
}

/* Clean up the ASCONF_ACK queue */
void sctp_assoc_clean_asconf_ack_cache(const struct sctp_association *asoc)
{
	struct sctp_chunk *ack;
	struct sctp_chunk *tmp;

L
Lucas De Marchi 已提交
1622
	/* We can remove all the entries from the queue up to
1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640
	 * the "Peer-Sequence-Number".
	 */
	list_for_each_entry_safe(ack, tmp, &asoc->asconf_ack_list,
				transmitted_list) {
		if (ack->subh.addip_hdr->serial ==
				htonl(asoc->peer.addip_serial))
			break;

		list_del_init(&ack->transmitted_list);
		sctp_chunk_free(ack);
	}
}

/* Find the ASCONF_ACK whose serial number matches ASCONF */
struct sctp_chunk *sctp_assoc_lookup_asconf_ack(
					const struct sctp_association *asoc,
					__be32 serial)
{
1641
	struct sctp_chunk *ack;
1642 1643 1644 1645 1646 1647 1648

	/* Walk through the list of cached ASCONF-ACKs and find the
	 * ack chunk whose serial number matches that of the request.
	 */
	list_for_each_entry(ack, &asoc->asconf_ack_list, transmitted_list) {
		if (ack->subh.addip_hdr->serial == serial) {
			sctp_chunk_hold(ack);
1649
			return ack;
1650 1651 1652
		}
	}

1653
	return NULL;
1654
}
1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667

void sctp_asconf_queue_teardown(struct sctp_association *asoc)
{
	/* Free any cached ASCONF_ACK chunk. */
	sctp_assoc_free_asconf_acks(asoc);

	/* Free the ASCONF queue. */
	sctp_assoc_free_asconf_queue(asoc);

	/* Free any cached ASCONF chunk. */
	if (asoc->addip_last_asconf)
		sctp_chunk_free(asoc->addip_last_asconf);
}