output.c 22.3 KB
Newer Older
1
/* SCTP kernel implementation
L
Linus Torvalds 已提交
2 3 4 5
 * (C) Copyright IBM Corp. 2001, 2004
 * Copyright (c) 1999-2000 Cisco, Inc.
 * Copyright (c) 1999-2001 Motorola, Inc.
 *
6
 * This file is part of the SCTP kernel implementation
L
Linus Torvalds 已提交
7 8 9
 *
 * These functions handle output processing.
 *
10
 * This SCTP implementation is free software;
L
Linus Torvalds 已提交
11 12 13 14 15
 * 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.
 *
16
 * This SCTP implementation is distributed in the hope that it
L
Linus Torvalds 已提交
17 18 19 20 21 22
 * 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
23 24
 * along with GNU CC; see the file COPYING.  If not, see
 * <http://www.gnu.org/licenses/>.
L
Linus Torvalds 已提交
25 26 27
 *
 * Please send any bug reports or fixes you make to the
 * email address(es):
28
 *    lksctp developers <linux-sctp@vger.kernel.org>
L
Linus Torvalds 已提交
29 30 31 32 33 34 35 36
 *
 * Written or modified by:
 *    La Monte H.P. Yarroll <piggy@acm.org>
 *    Karl Knutson          <karl@athena.chicago.il.us>
 *    Jon Grimm             <jgrimm@austin.ibm.com>
 *    Sridhar Samudrala     <sri@us.ibm.com>
 */

37 38
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

L
Linus Torvalds 已提交
39 40 41 42 43 44 45
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/wait.h>
#include <linux/time.h>
#include <linux/ip.h>
#include <linux/ipv6.h>
#include <linux/init.h>
46
#include <linux/slab.h>
L
Linus Torvalds 已提交
47
#include <net/inet_ecn.h>
V
Vlad Yasevich 已提交
48
#include <net/ip.h>
L
Linus Torvalds 已提交
49
#include <net/icmp.h>
50
#include <net/net_namespace.h>
L
Linus Torvalds 已提交
51 52 53 54 55 56

#include <linux/socket.h> /* for sa_family_t */
#include <net/sock.h>

#include <net/sctp/sctp.h>
#include <net/sctp/sm.h>
57
#include <net/sctp/checksum.h>
L
Linus Torvalds 已提交
58 59

/* Forward declarations for private helpers. */
60 61
static sctp_xmit_t __sctp_packet_append_chunk(struct sctp_packet *packet,
					      struct sctp_chunk *chunk);
62
static sctp_xmit_t sctp_packet_can_append_data(struct sctp_packet *packet,
L
Linus Torvalds 已提交
63
					   struct sctp_chunk *chunk);
64 65 66 67 68
static void sctp_packet_append_data(struct sctp_packet *packet,
					   struct sctp_chunk *chunk);
static sctp_xmit_t sctp_packet_will_fit(struct sctp_packet *packet,
					struct sctp_chunk *chunk,
					u16 chunk_len);
L
Linus Torvalds 已提交
69

70 71 72 73 74 75 76 77 78 79 80
static void sctp_packet_reset(struct sctp_packet *packet)
{
	packet->size = packet->overhead;
	packet->has_cookie_echo = 0;
	packet->has_sack = 0;
	packet->has_data = 0;
	packet->has_auth = 0;
	packet->ipfragok = 0;
	packet->auth = NULL;
}

L
Linus Torvalds 已提交
81 82 83 84 85 86 87 88
/* Config a packet.
 * This appears to be a followup set of initializations.
 */
struct sctp_packet *sctp_packet_config(struct sctp_packet *packet,
				       __u32 vtag, int ecn_capable)
{
	struct sctp_chunk *chunk = NULL;

89
	pr_debug("%s: packet:%p vtag:0x%x\n", __func__, packet, vtag);
L
Linus Torvalds 已提交
90 91 92 93 94 95 96

	packet->vtag = vtag;

	if (ecn_capable && sctp_packet_empty(packet)) {
		chunk = sctp_get_ecne_prepend(packet->transport->asoc);

		/* If there a is a prepend chunk stick it on the list before
97 98
		 * any other chunks get appended.
		 */
L
Linus Torvalds 已提交
99 100 101 102 103 104 105 106 107 108 109 110 111 112 113
		if (chunk)
			sctp_packet_append_chunk(packet, chunk);
	}

	return packet;
}

/* Initialize the packet structure. */
struct sctp_packet *sctp_packet_init(struct sctp_packet *packet,
				     struct sctp_transport *transport,
				     __u16 sport, __u16 dport)
{
	struct sctp_association *asoc = transport->asoc;
	size_t overhead;

114
	pr_debug("%s: packet:%p transport:%p\n", __func__, packet, transport);
L
Linus Torvalds 已提交
115 116 117 118

	packet->transport = transport;
	packet->source_port = sport;
	packet->destination_port = dport;
119
	INIT_LIST_HEAD(&packet->chunk_list);
L
Linus Torvalds 已提交
120
	if (asoc) {
121 122
		struct sctp_sock *sp = sctp_sk(asoc->base.sk);
		overhead = sp->pf->af->net_header_len;
L
Linus Torvalds 已提交
123 124 125 126 127
	} else {
		overhead = sizeof(struct ipv6hdr);
	}
	overhead += sizeof(struct sctphdr);
	packet->overhead = overhead;
128
	sctp_packet_reset(packet);
L
Linus Torvalds 已提交
129
	packet->vtag = 0;
130

L
Linus Torvalds 已提交
131 132 133 134 135 136
	return packet;
}

/* Free a packet.  */
void sctp_packet_free(struct sctp_packet *packet)
{
137
	struct sctp_chunk *chunk, *tmp;
L
Linus Torvalds 已提交
138

139
	pr_debug("%s: packet:%p\n", __func__, packet);
L
Linus Torvalds 已提交
140

141 142
	list_for_each_entry_safe(chunk, tmp, &packet->chunk_list, list) {
		list_del_init(&chunk->list);
L
Linus Torvalds 已提交
143
		sctp_chunk_free(chunk);
144
	}
L
Linus Torvalds 已提交
145 146 147 148 149 150 151 152 153 154
}

/* This routine tries to append the chunk to the offered packet. If adding
 * the chunk causes the packet to exceed the path MTU and COOKIE_ECHO chunk
 * is not present in the packet, it transmits the input packet.
 * Data can be bundled with a packet containing a COOKIE_ECHO chunk as long
 * as it can fit in the packet, but any more data that does not fit in this
 * packet can be sent only after receiving the COOKIE_ACK.
 */
sctp_xmit_t sctp_packet_transmit_chunk(struct sctp_packet *packet,
155
				       struct sctp_chunk *chunk,
156
				       int one_packet, gfp_t gfp)
L
Linus Torvalds 已提交
157 158 159 160
{
	sctp_xmit_t retval;
	int error = 0;

161
	pr_debug("%s: packet:%p chunk:%p\n", __func__, packet, chunk);
L
Linus Torvalds 已提交
162 163 164 165

	switch ((retval = (sctp_packet_append_chunk(packet, chunk)))) {
	case SCTP_XMIT_PMTU_FULL:
		if (!packet->has_cookie_echo) {
166
			error = sctp_packet_transmit(packet, gfp);
L
Linus Torvalds 已提交
167 168 169 170 171 172
			if (error < 0)
				chunk->skb->sk->sk_err = -error;

			/* If we have an empty packet, then we can NOT ever
			 * return PMTU_FULL.
			 */
173 174 175
			if (!one_packet)
				retval = sctp_packet_append_chunk(packet,
								  chunk);
L
Linus Torvalds 已提交
176 177 178 179 180
		}
		break;

	case SCTP_XMIT_RWND_FULL:
	case SCTP_XMIT_OK:
181
	case SCTP_XMIT_DELAY:
L
Linus Torvalds 已提交
182
		break;
183
	}
L
Linus Torvalds 已提交
184 185 186 187

	return retval;
}

188 189 190 191 192 193 194 195 196 197 198 199 200 201 202
/* Try to bundle an auth chunk into the packet. */
static sctp_xmit_t sctp_packet_bundle_auth(struct sctp_packet *pkt,
					   struct sctp_chunk *chunk)
{
	struct sctp_association *asoc = pkt->transport->asoc;
	struct sctp_chunk *auth;
	sctp_xmit_t retval = SCTP_XMIT_OK;

	/* if we don't have an association, we can't do authentication */
	if (!asoc)
		return retval;

	/* See if this is an auth chunk we are bundling or if
	 * auth is already bundled.
	 */
203
	if (chunk->chunk_hdr->type == SCTP_CID_AUTH || pkt->has_auth)
204 205 206 207 208 209 210 211 212 213 214 215
		return retval;

	/* if the peer did not request this chunk to be authenticated,
	 * don't do it
	 */
	if (!chunk->auth)
		return retval;

	auth = sctp_make_auth(asoc);
	if (!auth)
		return retval;

216 217 218 219
	retval = __sctp_packet_append_chunk(pkt, auth);

	if (retval != SCTP_XMIT_OK)
		sctp_chunk_free(auth);
220 221 222 223

	return retval;
}

L
Linus Torvalds 已提交
224 225 226 227 228 229 230 231 232 233 234 235
/* Try to bundle a SACK with the packet. */
static sctp_xmit_t sctp_packet_bundle_sack(struct sctp_packet *pkt,
					   struct sctp_chunk *chunk)
{
	sctp_xmit_t retval = SCTP_XMIT_OK;

	/* If sending DATA and haven't aleady bundled a SACK, try to
	 * bundle one in to the packet.
	 */
	if (sctp_chunk_is_data(chunk) && !pkt->has_sack &&
	    !pkt->has_cookie_echo) {
		struct sctp_association *asoc;
D
Doug Graham 已提交
236
		struct timer_list *timer;
L
Linus Torvalds 已提交
237
		asoc = pkt->transport->asoc;
D
Doug Graham 已提交
238
		timer = &asoc->timers[SCTP_EVENT_TIMEOUT_SACK];
L
Linus Torvalds 已提交
239

D
Doug Graham 已提交
240 241
		/* If the SACK timer is running, we have a pending SACK */
		if (timer_pending(timer)) {
L
Linus Torvalds 已提交
242
			struct sctp_chunk *sack;
243 244 245 246 247

			if (pkt->transport->sack_generation !=
			    pkt->transport->asoc->peer.sack_generation)
				return retval;

L
Linus Torvalds 已提交
248 249 250
			asoc->a_rwnd = asoc->rwnd;
			sack = sctp_make_sack(asoc);
			if (sack) {
251 252 253 254 255
				retval = __sctp_packet_append_chunk(pkt, sack);
				if (retval != SCTP_XMIT_OK) {
					sctp_chunk_free(sack);
					goto out;
				}
L
Linus Torvalds 已提交
256
				asoc->peer.sack_needed = 0;
D
Doug Graham 已提交
257
				if (del_timer(timer))
L
Linus Torvalds 已提交
258 259 260 261
					sctp_association_put(asoc);
			}
		}
	}
262
out:
L
Linus Torvalds 已提交
263 264 265
	return retval;
}

266

L
Linus Torvalds 已提交
267 268 269
/* Append a chunk to the offered packet reporting back any inability to do
 * so.
 */
270 271
static sctp_xmit_t __sctp_packet_append_chunk(struct sctp_packet *packet,
					      struct sctp_chunk *chunk)
L
Linus Torvalds 已提交
272 273 274 275
{
	sctp_xmit_t retval = SCTP_XMIT_OK;
	__u16 chunk_len = WORD_ROUND(ntohs(chunk->chunk_hdr->length));

276 277 278 279
	/* Check to see if this chunk will fit into the packet */
	retval = sctp_packet_will_fit(packet, chunk, chunk_len);
	if (retval != SCTP_XMIT_OK)
		goto finish;
L
Linus Torvalds 已提交
280

281
	/* We believe that this chunk is OK to add to the packet */
282
	switch (chunk->chunk_hdr->type) {
283
	case SCTP_CID_DATA:
284 285
		/* Account for the data being in the packet */
		sctp_packet_append_data(packet, chunk);
L
Linus Torvalds 已提交
286 287
		/* Disallow SACK bundling after DATA. */
		packet->has_sack = 1;
288 289 290 291
		/* Disallow AUTH bundling after DATA */
		packet->has_auth = 1;
		/* Let it be knows that packet has DATA in it */
		packet->has_data = 1;
292 293
		/* timestamp the chunk for rtx purposes */
		chunk->sent_at = jiffies;
294
		break;
295
	case SCTP_CID_COOKIE_ECHO:
L
Linus Torvalds 已提交
296
		packet->has_cookie_echo = 1;
297 298
		break;

299
	case SCTP_CID_SACK:
L
Linus Torvalds 已提交
300
		packet->has_sack = 1;
301 302
		if (chunk->asoc)
			chunk->asoc->stats.osacks++;
303 304
		break;

305
	case SCTP_CID_AUTH:
306 307 308 309
		packet->has_auth = 1;
		packet->auth = chunk;
		break;
	}
L
Linus Torvalds 已提交
310 311

	/* It is OK to send this chunk.  */
312
	list_add_tail(&chunk->list, &packet->chunk_list);
L
Linus Torvalds 已提交
313 314 315 316 317 318
	packet->size += chunk_len;
	chunk->transport = packet->transport;
finish:
	return retval;
}

319 320 321 322 323 324 325 326
/* Append a chunk to the offered packet reporting back any inability to do
 * so.
 */
sctp_xmit_t sctp_packet_append_chunk(struct sctp_packet *packet,
				     struct sctp_chunk *chunk)
{
	sctp_xmit_t retval = SCTP_XMIT_OK;

327
	pr_debug("%s: packet:%p chunk:%p\n", __func__, packet, chunk);
328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354

	/* Data chunks are special.  Before seeing what else we can
	 * bundle into this packet, check to see if we are allowed to
	 * send this DATA.
	 */
	if (sctp_chunk_is_data(chunk)) {
		retval = sctp_packet_can_append_data(packet, chunk);
		if (retval != SCTP_XMIT_OK)
			goto finish;
	}

	/* Try to bundle AUTH chunk */
	retval = sctp_packet_bundle_auth(packet, chunk);
	if (retval != SCTP_XMIT_OK)
		goto finish;

	/* Try to bundle SACK chunk */
	retval = sctp_packet_bundle_sack(packet, chunk);
	if (retval != SCTP_XMIT_OK)
		goto finish;

	retval = __sctp_packet_append_chunk(packet, chunk);

finish:
	return retval;
}

355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373
static void sctp_packet_release_owner(struct sk_buff *skb)
{
	sk_free(skb->sk);
}

static void sctp_packet_set_owner_w(struct sk_buff *skb, struct sock *sk)
{
	skb_orphan(skb);
	skb->sk = sk;
	skb->destructor = sctp_packet_release_owner;

	/*
	 * The data chunks have already been accounted for in sctp_sendmsg(),
	 * therefore only reserve a single byte to keep socket around until
	 * the packet has been transmitted.
	 */
	atomic_inc(&sk->sk_wmem_alloc);
}

L
Linus Torvalds 已提交
374 375 376 377 378
/* All packets are sent to the network through this function from
 * sctp_outq_tail().
 *
 * The return value is a normal kernel error return value.
 */
379
int sctp_packet_transmit(struct sctp_packet *packet, gfp_t gfp)
L
Linus Torvalds 已提交
380 381 382 383 384
{
	struct sctp_transport *tp = packet->transport;
	struct sctp_association *asoc = tp->asoc;
	struct sctphdr *sh;
	struct sk_buff *nskb;
385
	struct sctp_chunk *chunk, *tmp;
L
Linus Torvalds 已提交
386 387 388 389
	struct sock *sk;
	int err = 0;
	int padding;		/* How much padding do we need?  */
	__u8 has_data = 0;
390
	struct dst_entry *dst;
391
	unsigned char *auth = NULL;	/* pointer to auth in skb data */
L
Linus Torvalds 已提交
392

393
	pr_debug("%s: packet:%p\n", __func__, packet);
L
Linus Torvalds 已提交
394 395

	/* Do NOT generate a chunkless packet. */
396
	if (list_empty(&packet->chunk_list))
L
Linus Torvalds 已提交
397 398 399
		return err;

	/* Set up convenience variables... */
400
	chunk = list_entry(packet->chunk_list.next, struct sctp_chunk, list);
L
Linus Torvalds 已提交
401 402 403
	sk = chunk->skb->sk;

	/* Allocate the new skb.  */
404
	nskb = alloc_skb(packet->size + MAX_HEADER, gfp);
L
Linus Torvalds 已提交
405 406 407 408
	if (!nskb)
		goto nomem;

	/* Make sure the outbound skb has enough header room reserved. */
409
	skb_reserve(nskb, packet->overhead + MAX_HEADER);
L
Linus Torvalds 已提交
410 411 412 413

	/* Set the owning socket so that we know where to get the
	 * destination IP address.
	 */
414
	sctp_packet_set_owner_w(nskb, sk);
L
Linus Torvalds 已提交
415

416
	if (!sctp_transport_dst_check(tp)) {
417 418
		sctp_transport_route(tp, NULL, sctp_sk(sk));
		if (asoc && (asoc->param_flags & SPP_PMTUD_ENABLE)) {
419
			sctp_assoc_sync_pmtu(sk, asoc);
420 421
		}
	}
E
Eric Dumazet 已提交
422
	dst = dst_clone(tp->dst);
423
	if (!dst)
424
		goto no_route;
425
	skb_dst_set(nskb, dst);
426

L
Linus Torvalds 已提交
427 428
	/* Build the SCTP header.  */
	sh = (struct sctphdr *)skb_push(nskb, sizeof(struct sctphdr));
429
	skb_reset_transport_header(nskb);
L
Linus Torvalds 已提交
430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462
	sh->source = htons(packet->source_port);
	sh->dest   = htons(packet->destination_port);

	/* From 6.8 Adler-32 Checksum Calculation:
	 * After the packet is constructed (containing the SCTP common
	 * header and one or more control or DATA chunks), the
	 * transmitter shall:
	 *
	 * 1) Fill in the proper Verification Tag in the SCTP common
	 *    header and initialize the checksum field to 0's.
	 */
	sh->vtag     = htonl(packet->vtag);
	sh->checksum = 0;

	/**
	 * 6.10 Bundling
	 *
	 *    An endpoint bundles chunks by simply including multiple
	 *    chunks in one outbound SCTP packet.  ...
	 */

	/**
	 * 3.2  Chunk Field Descriptions
	 *
	 * The total length of a chunk (including Type, Length and
	 * Value fields) MUST be a multiple of 4 bytes.  If the length
	 * of the chunk is not a multiple of 4 bytes, the sender MUST
	 * pad the chunk with all zero bytes and this padding is not
	 * included in the chunk length field.  The sender should
	 * never pad with more than 3 bytes.
	 *
	 * [This whole comment explains WORD_ROUND() below.]
	 */
463 464 465

	pr_debug("***sctp_transmit_packet***\n");

466 467
	list_for_each_entry_safe(chunk, tmp, &packet->chunk_list, list) {
		list_del_init(&chunk->list);
L
Linus Torvalds 已提交
468
		if (sctp_chunk_is_data(chunk)) {
469 470 471 472 473 474
			/* 6.3.1 C4) When data is in flight and when allowed
			 * by rule C5, a new RTT measurement MUST be made each
			 * round trip.  Furthermore, new RTT measurements
			 * SHOULD be made no more than once per round-trip
			 * for a given destination transport address.
			 */
L
Linus Torvalds 已提交
475

476
			if (!chunk->resent && !tp->rto_pending) {
477 478
				chunk->rtt_in_progress = 1;
				tp->rto_pending = 1;
479
			}
480

L
Linus Torvalds 已提交
481 482 483 484 485 486 487
			has_data = 1;
		}

		padding = WORD_ROUND(chunk->skb->len) - chunk->skb->len;
		if (padding)
			memset(skb_put(chunk->skb, padding), 0, padding);

488 489 490 491 492 493 494 495
		/* if this is the auth chunk that we are adding,
		 * store pointer where it will be added and put
		 * the auth into the packet.
		 */
		if (chunk == packet->auth)
			auth = skb_tail_pointer(nskb);

		memcpy(skb_put(nskb, chunk->skb->len),
496
			       chunk->skb->data, chunk->skb->len);
L
Linus Torvalds 已提交
497

498 499 500 501 502 503 504
		pr_debug("*** Chunk:%p[%s] %s 0x%x, length:%d, chunk->skb->len:%d, "
			 "rtt_in_progress:%d\n", chunk,
			 sctp_cname(SCTP_ST_CHUNK(chunk->chunk_hdr->type)),
			 chunk->has_tsn ? "TSN" : "No TSN",
			 chunk->has_tsn ? ntohl(chunk->subh.data_hdr->tsn) : 0,
			 ntohs(chunk->chunk_hdr->length), chunk->skb->len,
			 chunk->rtt_in_progress);
L
Linus Torvalds 已提交
505 506 507 508 509 510 511

		/*
		 * If this is a control chunk, this is our last
		 * reference. Free data chunks after they've been
		 * acknowledged or have failed.
		 */
		if (!sctp_chunk_is_data(chunk))
512
			sctp_chunk_free(chunk);
L
Linus Torvalds 已提交
513 514
	}

515 516 517 518 519 520 521 522 523 524 525
	/* SCTP-AUTH, Section 6.2
	 *    The sender MUST calculate the MAC as described in RFC2104 [2]
	 *    using the hash function H as described by the MAC Identifier and
	 *    the shared association key K based on the endpoint pair shared key
	 *    described by the shared key identifier.  The 'data' used for the
	 *    computation of the AUTH-chunk is given by the AUTH chunk with its
	 *    HMAC field set to zero (as shown in Figure 6) followed by all
	 *    chunks that are placed after the AUTH chunk in the SCTP packet.
	 */
	if (auth)
		sctp_auth_calculate_hmac(asoc, nskb,
526 527
					 (struct sctp_auth_chunk *)auth,
					 gfp);
528 529 530 531 532 533 534 535

	/* 2) Calculate the Adler-32 checksum of the whole packet,
	 *    including the SCTP common header and all the
	 *    chunks.
	 *
	 * Note: Adler-32 is no longer applicable, as has been replaced
	 * by CRC32-C as described in <draft-ietf-tsvwg-sctpcsum-02.txt>.
	 */
536
	if (!sctp_checksum_disable) {
537
		if (!(dst->dev->features & NETIF_F_SCTP_CRC) ||
538
		    (dst_xfrm(dst) != NULL) || packet->ipfragok) {
539
			sh->checksum = sctp_compute_cksum(nskb, 0);
540
		} else {
L
Lucas De Marchi 已提交
541
			/* no need to seed pseudo checksum for SCTP */
542
			nskb->ip_summed = CHECKSUM_PARTIAL;
543
			nskb->csum_start = skb_transport_header(nskb) - nskb->head;
544 545 546
			nskb->csum_offset = offsetof(struct sctphdr, checksum);
		}
	}
L
Linus Torvalds 已提交
547 548 549 550 551 552 553 554 555 556 557 558 559

	/* IP layer ECN support
	 * From RFC 2481
	 *  "The ECN-Capable Transport (ECT) bit would be set by the
	 *   data sender to indicate that the end-points of the
	 *   transport protocol are ECN-capable."
	 *
	 * Now setting the ECT bit all the time, as it should not cause
	 * any problems protocol-wise even if our peer ignores it.
	 *
	 * Note: The works for IPv6 layer checks this bit too later
	 * in transmission.  See IP6_ECN_flow_xmit().
	 */
560
	tp->af_specific->ecn_capable(nskb->sk);
L
Linus Torvalds 已提交
561 562 563 564 565 566 567

	/* Set up the IP options.  */
	/* BUG: not implemented
	 * For v4 this all lives somewhere in sk->sk_opt...
	 */

	/* Dump that on IP!  */
568 569 570 571 572 573 574
	if (asoc) {
		asoc->stats.opackets++;
		if (asoc->peer.last_sent_to != tp)
			/* Considering the multiple CPU scenario, this is a
			 * "correcter" place for last_sent_to.  --xguo
			 */
			asoc->peer.last_sent_to = tp;
L
Linus Torvalds 已提交
575 576 577 578 579 580 581
	}

	if (has_data) {
		struct timer_list *timer;
		unsigned long timeout;

		/* Restart the AUTOCLOSE timer when sending data. */
582 583
		if (sctp_state(asoc, ESTABLISHED) &&
		    asoc->timeouts[SCTP_EVENT_TIMEOUT_AUTOCLOSE]) {
L
Linus Torvalds 已提交
584 585 586 587 588 589 590 591
			timer = &asoc->timers[SCTP_EVENT_TIMEOUT_AUTOCLOSE];
			timeout = asoc->timeouts[SCTP_EVENT_TIMEOUT_AUTOCLOSE];

			if (!mod_timer(timer, jiffies + timeout))
				sctp_association_hold(asoc);
		}
	}

592
	pr_debug("***sctp_transmit_packet*** skb->len:%d\n", nskb->len);
L
Linus Torvalds 已提交
593

W
WANG Cong 已提交
594
	nskb->ignore_df = packet->ipfragok;
595
	tp->af_specific->sctp_xmit(nskb, tp);
L
Linus Torvalds 已提交
596 597

out:
598
	sctp_packet_reset(packet);
L
Linus Torvalds 已提交
599 600 601
	return err;
no_route:
	kfree_skb(nskb);
602 603 604

	if (asoc)
		IP_INC_STATS(sock_net(asoc->base.sk), IPSTATS_MIB_OUTNOROUTES);
L
Linus Torvalds 已提交
605 606 607 608 609 610 611 612 613 614 615 616 617 618

	/* FIXME: Returning the 'err' will effect all the associations
	 * associated with a socket, although only one of the paths of the
	 * association is unreachable.
	 * The real failure of a transport or association can be passed on
	 * to the user via notifications. So setting this error may not be
	 * required.
	 */
	 /* err = -EHOSTUNREACH; */
err:
	/* Control chunks are unreliable so just drop them.  DATA chunks
	 * will get resent or dropped later.
	 */

619 620
	list_for_each_entry_safe(chunk, tmp, &packet->chunk_list, list) {
		list_del_init(&chunk->list);
L
Linus Torvalds 已提交
621
		if (!sctp_chunk_is_data(chunk))
622
			sctp_chunk_free(chunk);
L
Linus Torvalds 已提交
623 624 625 626 627 628 629 630 631 632 633
	}
	goto out;
nomem:
	err = -ENOMEM;
	goto err;
}

/********************************************************************
 * 2nd Level Abstractions
 ********************************************************************/

634 635
/* This private function check to see if a chunk can be added */
static sctp_xmit_t sctp_packet_can_append_data(struct sctp_packet *packet,
L
Linus Torvalds 已提交
636 637
					   struct sctp_chunk *chunk)
{
638
	size_t datasize, rwnd, inflight, flight_size;
L
Linus Torvalds 已提交
639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656
	struct sctp_transport *transport = packet->transport;
	struct sctp_association *asoc = transport->asoc;
	struct sctp_outq *q = &asoc->outqueue;

	/* RFC 2960 6.1  Transmission of DATA Chunks
	 *
	 * A) At any given time, the data sender MUST NOT transmit new data to
	 * any destination transport address if its peer's rwnd indicates
	 * that the peer has no buffer space (i.e. rwnd is 0, see Section
	 * 6.2.1).  However, regardless of the value of rwnd (including if it
	 * is 0), the data sender can always have one DATA chunk in flight to
	 * the receiver if allowed by cwnd (see rule B below).  This rule
	 * allows the sender to probe for a change in rwnd that the sender
	 * missed due to the SACK having been lost in transit from the data
	 * receiver to the data sender.
	 */

	rwnd = asoc->peer.rwnd;
657 658
	inflight = q->outstanding_bytes;
	flight_size = transport->flight_size;
L
Linus Torvalds 已提交
659 660 661

	datasize = sctp_data_size(chunk);

662 663 664 665 666
	if (datasize > rwnd && inflight > 0)
		/* We have (at least) one data chunk in flight,
		 * so we can't fall back to rule 6.1 B).
		 */
		return SCTP_XMIT_RWND_FULL;
L
Linus Torvalds 已提交
667 668 669 670 671 672 673 674 675 676 677 678 679

	/* RFC 2960 6.1  Transmission of DATA Chunks
	 *
	 * B) At any given time, the sender MUST NOT transmit new data
	 * to a given transport address if it has cwnd or more bytes
	 * of data outstanding to that transport address.
	 */
	/* RFC 7.2.4 & the Implementers Guide 2.8.
	 *
	 * 3) ...
	 *    When a Fast Retransmit is being performed the sender SHOULD
	 *    ignore the value of cwnd and SHOULD NOT delay retransmission.
	 */
680 681 682
	if (chunk->fast_retransmit != SCTP_NEED_FRTX &&
	    flight_size >= transport->cwnd)
		return SCTP_XMIT_RWND_FULL;
L
Linus Torvalds 已提交
683 684 685 686 687 688 689

	/* Nagle's algorithm to solve small-packet problem:
	 * Inhibit the sending of new chunks when new outgoing data arrives
	 * if any previously transmitted data on the connection remains
	 * unacknowledged.
	 */

690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716
	if (sctp_sk(asoc->base.sk)->nodelay)
		/* Nagle disabled */
		return SCTP_XMIT_OK;

	if (!sctp_packet_empty(packet))
		/* Append to packet */
		return SCTP_XMIT_OK;

	if (inflight == 0)
		/* Nothing unacked */
		return SCTP_XMIT_OK;

	if (!sctp_state(asoc, ESTABLISHED))
		return SCTP_XMIT_OK;

	/* Check whether this chunk and all the rest of pending data will fit
	 * or delay in hopes of bundling a full sized packet.
	 */
	if (chunk->skb->len + q->out_qlen >= transport->pathmtu - packet->overhead)
		/* Enough data queued to fill a packet */
		return SCTP_XMIT_OK;

	/* Don't delay large message writes that may have been fragmented */
	if (!chunk->msg->can_delay)
		return SCTP_XMIT_OK;

	/* Defer until all data acked or packet full */
717
	return SCTP_XMIT_DELAY;
718 719 720 721 722 723 724 725 726 727 728
}

/* This private function does management things when adding DATA chunk */
static void sctp_packet_append_data(struct sctp_packet *packet,
				struct sctp_chunk *chunk)
{
	struct sctp_transport *transport = packet->transport;
	size_t datasize = sctp_data_size(chunk);
	struct sctp_association *asoc = transport->asoc;
	u32 rwnd = asoc->peer.rwnd;

L
Linus Torvalds 已提交
729 730 731 732 733 734
	/* Keep track of how many bytes are in flight over this transport. */
	transport->flight_size += datasize;

	/* Keep track of how many bytes are in flight to the receiver. */
	asoc->outqueue.outstanding_bytes += datasize;

735
	/* Update our view of the receiver's rwnd. */
L
Linus Torvalds 已提交
736 737 738 739 740 741 742 743 744
	if (datasize < rwnd)
		rwnd -= datasize;
	else
		rwnd = 0;

	asoc->peer.rwnd = rwnd;
	/* Has been accepted for transmission. */
	if (!asoc->peer.prsctp_capable)
		chunk->msg->can_abandon = 0;
745 746
	sctp_chunk_assign_tsn(chunk);
	sctp_chunk_assign_ssn(chunk);
747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785
}

static sctp_xmit_t sctp_packet_will_fit(struct sctp_packet *packet,
					struct sctp_chunk *chunk,
					u16 chunk_len)
{
	size_t psize;
	size_t pmtu;
	int too_big;
	sctp_xmit_t retval = SCTP_XMIT_OK;

	psize = packet->size;
	pmtu  = ((packet->transport->asoc) ?
		(packet->transport->asoc->pathmtu) :
		(packet->transport->pathmtu));

	too_big = (psize + chunk_len > pmtu);

	/* Decide if we need to fragment or resubmit later. */
	if (too_big) {
		/* It's OK to fragmet at IP level if any one of the following
		 * is true:
		 * 	1. The packet is empty (meaning this chunk is greater
		 * 	   the MTU)
		 * 	2. The chunk we are adding is a control chunk
		 * 	3. The packet doesn't have any data in it yet and data
		 * 	requires authentication.
		 */
		if (sctp_packet_empty(packet) || !sctp_chunk_is_data(chunk) ||
		    (!packet->has_data && chunk->auth)) {
			/* We no longer do re-fragmentation.
			 * Just fragment at the IP layer, if we
			 * actually hit this condition
			 */
			packet->ipfragok = 1;
		} else {
			retval = SCTP_XMIT_PMTU_FULL;
		}
	}
L
Linus Torvalds 已提交
786 787 788

	return retval;
}