msg.c 15.4 KB
Newer Older
P
Per Liden 已提交
1 2
/*
 * net/tipc/msg.c: TIPC message header routines
3
 *
4
 * Copyright (c) 2000-2006, 2014-2015, Ericsson AB
5
 * Copyright (c) 2005, 2010-2011, Wind River Systems
P
Per Liden 已提交
6 7
 * All rights reserved.
 *
P
Per Liden 已提交
8
 * Redistribution and use in source and binary forms, with or without
P
Per Liden 已提交
9 10
 * modification, are permitted provided that the following conditions are met:
 *
P
Per Liden 已提交
11 12 13 14 15 16 17 18
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. Neither the names of the copyright holders nor the names of its
 *    contributors may be used to endorse or promote products derived from
 *    this software without specific prior written permission.
P
Per Liden 已提交
19
 *
P
Per Liden 已提交
20 21 22 23 24 25 26 27 28 29 30 31 32 33
 * Alternatively, this software may be distributed under the terms of the
 * GNU General Public License ("GPL") version 2 as published by the Free
 * Software Foundation.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
 * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
 * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
P
Per Liden 已提交
34 35 36
 * POSSIBILITY OF SUCH DAMAGE.
 */

37
#include <net/sock.h>
P
Per Liden 已提交
38 39
#include "core.h"
#include "msg.h"
40 41
#include "addr.h"
#include "name_table.h"
P
Per Liden 已提交
42

43 44
#define MAX_FORWARD_SIZE 1024

45
static unsigned int align(unsigned int i)
46
{
47
	return (i + 3) & ~3u;
48 49
}

Y
Ying Xue 已提交
50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72
/**
 * tipc_buf_acquire - creates a TIPC message buffer
 * @size: message size (including TIPC header)
 *
 * Returns a new buffer with data pointers set to the specified size.
 *
 * NOTE: Headroom is reserved to allow prepending of a data link header.
 *       There may also be unrequested tailroom present at the buffer's end.
 */
struct sk_buff *tipc_buf_acquire(u32 size)
{
	struct sk_buff *skb;
	unsigned int buf_size = (BUF_HEADROOM + size + 3) & ~3u;

	skb = alloc_skb_fclone(buf_size, GFP_ATOMIC);
	if (skb) {
		skb_reserve(skb, BUF_HEADROOM);
		skb_put(skb, size);
		skb->next = NULL;
	}
	return skb;
}

73 74
void tipc_msg_init(u32 own_node, struct tipc_msg *m, u32 user, u32 type,
		   u32 hsize, u32 dnode)
75 76 77 78 79 80
{
	memset(m, 0, hsize);
	msg_set_version(m);
	msg_set_user(m, user);
	msg_set_hdr_sz(m, hsize);
	msg_set_size(m, hsize);
81
	msg_set_prevnode(m, own_node);
82
	msg_set_type(m, type);
83
	if (hsize > SHORT_H_SIZE) {
84 85
		msg_set_orignode(m, own_node);
		msg_set_destnode(m, dnode);
86 87 88
	}
}

89
struct sk_buff *tipc_msg_create(uint user, uint type,
90 91
				uint hdr_sz, uint data_sz, u32 dnode,
				u32 onode, u32 dport, u32 oport, int errcode)
92 93 94 95 96 97 98 99 100
{
	struct tipc_msg *msg;
	struct sk_buff *buf;

	buf = tipc_buf_acquire(hdr_sz + data_sz);
	if (unlikely(!buf))
		return NULL;

	msg = buf_msg(buf);
101
	tipc_msg_init(onode, msg, user, type, hdr_sz, dnode);
102 103 104 105 106 107 108 109 110
	msg_set_size(msg, hdr_sz + data_sz);
	msg_set_origport(msg, oport);
	msg_set_destport(msg, dport);
	msg_set_errcode(msg, errcode);
	if (hdr_sz > SHORT_H_SIZE) {
		msg_set_orignode(msg, onode);
		msg_set_destnode(msg, dnode);
	}
	return buf;
111 112
}

113
/* tipc_buf_append(): Append a buffer to the fragment list of another buffer
114 115 116
 * @*headbuf: in:  NULL for first frag, otherwise value returned from prev call
 *            out: set when successful non-complete reassembly, otherwise NULL
 * @*buf:     in:  the buffer to append. Always defined
S
stephen hemminger 已提交
117
 *            out: head buf after successful complete reassembly, otherwise NULL
118
 * Returns 1 when reassembly complete, otherwise 0
119 120 121 122 123 124
 */
int tipc_buf_append(struct sk_buff **headbuf, struct sk_buff **buf)
{
	struct sk_buff *head = *headbuf;
	struct sk_buff *frag = *buf;
	struct sk_buff *tail;
125 126
	struct tipc_msg *msg;
	u32 fragid;
127
	int delta;
128
	bool headstolen;
129

130 131 132 133 134 135
	if (!frag)
		goto err;

	msg = buf_msg(frag);
	fragid = msg_type(msg);
	frag->next = NULL;
136 137 138
	skb_pull(frag, msg_hdr_sz(msg));

	if (fragid == FIRST_FRAGMENT) {
139 140 141 142
		if (unlikely(head))
			goto err;
		if (unlikely(skb_unclone(frag, GFP_ATOMIC)))
			goto err;
143 144
		head = *headbuf = frag;
		skb_frag_list_init(head);
145
		TIPC_SKB_CB(head)->tail = NULL;
146
		*buf = NULL;
147 148
		return 0;
	}
149

150
	if (!head)
151 152
		goto err;

153 154 155
	if (skb_try_coalesce(head, frag, &headstolen, &delta)) {
		kfree_skb_partial(frag, headstolen);
	} else {
156
		tail = TIPC_SKB_CB(head)->tail;
157 158 159 160 161 162 163 164 165
		if (!skb_has_frag_list(head))
			skb_shinfo(head)->frag_list = frag;
		else
			tail->next = frag;
		head->truesize += frag->truesize;
		head->data_len += frag->len;
		head->len += frag->len;
		TIPC_SKB_CB(head)->tail = frag;
	}
166

167
	if (fragid == LAST_FRAGMENT) {
168 169 170
		TIPC_SKB_CB(head)->validated = false;
		if (unlikely(!tipc_msg_validate(head)))
			goto err;
171 172 173 174 175 176 177
		*buf = head;
		TIPC_SKB_CB(head)->tail = NULL;
		*headbuf = NULL;
		return 1;
	}
	*buf = NULL;
	return 0;
178
err:
179 180
	pr_warn_ratelimited("Unable to build fragment list\n");
	kfree_skb(*buf);
181 182
	kfree_skb(*headbuf);
	*buf = *headbuf = NULL;
183 184
	return 0;
}
185

186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227
/* tipc_msg_validate - validate basic format of received message
 *
 * This routine ensures a TIPC message has an acceptable header, and at least
 * as much data as the header indicates it should.  The routine also ensures
 * that the entire message header is stored in the main fragment of the message
 * buffer, to simplify future access to message header fields.
 *
 * Note: Having extra info present in the message header or data areas is OK.
 * TIPC will ignore the excess, under the assumption that it is optional info
 * introduced by a later release of the protocol.
 */
bool tipc_msg_validate(struct sk_buff *skb)
{
	struct tipc_msg *msg;
	int msz, hsz;

	if (unlikely(TIPC_SKB_CB(skb)->validated))
		return true;
	if (unlikely(!pskb_may_pull(skb, MIN_H_SIZE)))
		return false;

	hsz = msg_hdr_sz(buf_msg(skb));
	if (unlikely(hsz < MIN_H_SIZE) || (hsz > MAX_H_SIZE))
		return false;
	if (unlikely(!pskb_may_pull(skb, hsz)))
		return false;

	msg = buf_msg(skb);
	if (unlikely(msg_version(msg) != TIPC_VERSION))
		return false;

	msz = msg_size(msg);
	if (unlikely(msz < hsz))
		return false;
	if (unlikely((msz - hsz) > TIPC_MAX_USER_MSG_SIZE))
		return false;
	if (unlikely(skb->len < msz))
		return false;

	TIPC_SKB_CB(skb)->validated = true;
	return true;
}
228 229

/**
230
 * tipc_msg_build - create buffer chain containing specified header and data
231
 * @mhdr: Message header, to be prepended to data
232
 * @m: User message
233 234
 * @dsz: Total length of user data
 * @pktmax: Max packet size that can be used
235 236
 * @list: Buffer or chain of buffers to be returned to caller
 *
237 238
 * Returns message data size or errno: -ENOMEM, -EFAULT
 */
239
int tipc_msg_build(struct tipc_msg *mhdr, struct msghdr *m,
240
		   int offset, int dsz, int pktmax, struct sk_buff_head *list)
241 242 243 244 245 246 247 248
{
	int mhsz = msg_hdr_sz(mhdr);
	int msz = mhsz + dsz;
	int pktno = 1;
	int pktsz;
	int pktrem = pktmax;
	int drem = dsz;
	struct tipc_msg pkthdr;
249
	struct sk_buff *skb;
250 251
	char *pktpos;
	int rc;
252

253 254 255 256
	msg_set_size(mhdr, msz);

	/* No fragmentation needed? */
	if (likely(msz <= pktmax)) {
257 258
		skb = tipc_buf_acquire(msz);
		if (unlikely(!skb))
259
			return -ENOMEM;
260
		skb_orphan(skb);
261 262 263
		__skb_queue_tail(list, skb);
		skb_copy_to_linear_data(skb, mhdr, mhsz);
		pktpos = skb->data + mhsz;
A
Al Viro 已提交
264
		if (copy_from_iter(pktpos, dsz, &m->msg_iter) == dsz)
265 266 267 268 269 270
			return dsz;
		rc = -EFAULT;
		goto error;
	}

	/* Prepare reusable fragment header */
271 272
	tipc_msg_init(msg_prevnode(mhdr), &pkthdr, MSG_FRAGMENTER,
		      FIRST_FRAGMENT, INT_H_SIZE, msg_destnode(mhdr));
273 274 275 276
	msg_set_size(&pkthdr, pktmax);
	msg_set_fragm_no(&pkthdr, pktno);

	/* Prepare first fragment */
277 278
	skb = tipc_buf_acquire(pktmax);
	if (!skb)
279
		return -ENOMEM;
280
	skb_orphan(skb);
281 282 283
	__skb_queue_tail(list, skb);
	pktpos = skb->data;
	skb_copy_to_linear_data(skb, &pkthdr, INT_H_SIZE);
284 285
	pktpos += INT_H_SIZE;
	pktrem -= INT_H_SIZE;
286
	skb_copy_to_linear_data_offset(skb, INT_H_SIZE, mhdr, mhsz);
287 288 289 290 291 292 293
	pktpos += mhsz;
	pktrem -= mhsz;

	do {
		if (drem < pktrem)
			pktrem = drem;

A
Al Viro 已提交
294
		if (copy_from_iter(pktpos, pktrem, &m->msg_iter) != pktrem) {
295 296 297 298 299 300 301 302 303 304 305 306 307
			rc = -EFAULT;
			goto error;
		}
		drem -= pktrem;

		if (!drem)
			break;

		/* Prepare new fragment: */
		if (drem < (pktmax - INT_H_SIZE))
			pktsz = drem + INT_H_SIZE;
		else
			pktsz = pktmax;
308 309
		skb = tipc_buf_acquire(pktsz);
		if (!skb) {
310 311 312
			rc = -ENOMEM;
			goto error;
		}
313
		skb_orphan(skb);
314
		__skb_queue_tail(list, skb);
315 316 317
		msg_set_type(&pkthdr, FRAGMENT);
		msg_set_size(&pkthdr, pktsz);
		msg_set_fragm_no(&pkthdr, ++pktno);
318 319
		skb_copy_to_linear_data(skb, &pkthdr, INT_H_SIZE);
		pktpos = skb->data + INT_H_SIZE;
320 321 322
		pktrem = pktsz - INT_H_SIZE;

	} while (1);
323
	msg_set_type(buf_msg(skb), LAST_FRAGMENT);
324 325
	return dsz;
error:
326 327
	__skb_queue_purge(list);
	__skb_queue_head_init(list);
328 329 330
	return rc;
}

331 332
/**
 * tipc_msg_bundle(): Append contents of a buffer to tail of an existing one
J
Jon Paul Maloy 已提交
333
 * @bskb: the buffer to append to ("bundle")
334
 * @skb:  buffer to be appended
335 336 337 338
 * @mtu:  max allowable size for the bundle buffer
 * Consumes buffer if successful
 * Returns true if bundling could be performed, otherwise false
 */
J
Jon Paul Maloy 已提交
339
bool tipc_msg_bundle(struct sk_buff *bskb, struct sk_buff *skb, u32 mtu)
340
{
J
Jon Paul Maloy 已提交
341
	struct tipc_msg *bmsg;
342
	struct tipc_msg *msg = buf_msg(skb);
J
Jon Paul Maloy 已提交
343
	unsigned int bsz;
344
	unsigned int msz = msg_size(msg);
J
Jon Paul Maloy 已提交
345
	u32 start, pad;
346 347 348 349
	u32 max = mtu - INT_H_SIZE;

	if (likely(msg_user(msg) == MSG_FRAGMENTER))
		return false;
J
Jon Paul Maloy 已提交
350 351 352 353 354 355 356
	if (!bskb)
		return false;
	bmsg = buf_msg(bskb);
	bsz = msg_size(bmsg);
	start = align(bsz);
	pad = start - bsz;

357 358 359 360 361 362
	if (unlikely(msg_user(msg) == CHANGEOVER_PROTOCOL))
		return false;
	if (unlikely(msg_user(msg) == BCAST_PROTOCOL))
		return false;
	if (likely(msg_user(bmsg) != MSG_BUNDLER))
		return false;
363
	if (unlikely(skb_tailroom(bskb) < (pad + msz)))
364 365 366 367
		return false;
	if (unlikely(max < (start + msz)))
		return false;

368 369
	skb_put(bskb, pad + msz);
	skb_copy_to_linear_data_offset(bskb, start, skb->data, msz);
370 371
	msg_set_size(bmsg, start + msz);
	msg_set_msgcnt(bmsg, msg_msgcnt(bmsg) + 1);
372
	kfree_skb(skb);
373 374 375
	return true;
}

376 377
/**
 *  tipc_msg_extract(): extract bundled inner packet from buffer
378
 *  @skb: buffer to be extracted from.
379
 *  @iskb: extracted inner buffer, to be returned
380 381
 *  @pos: position in outer message of msg to be extracted.
 *        Returns position of next msg
382 383 384 385 386
 *  Consumes outer buffer when last packet extracted
 *  Returns true when when there is an extracted buffer, otherwise false
 */
bool tipc_msg_extract(struct sk_buff *skb, struct sk_buff **iskb, int *pos)
{
387
	struct tipc_msg *msg;
388
	int imsz, offset;
389

390
	*iskb = NULL;
391
	if (unlikely(skb_linearize(skb)))
392 393
		goto none;

394
	msg = buf_msg(skb);
395 396
	offset = msg_hdr_sz(msg) + *pos;
	if (unlikely(offset > (msg_size(msg) - MIN_H_SIZE)))
397 398
		goto none;

399 400
	*iskb = skb_clone(skb, GFP_ATOMIC);
	if (unlikely(!*iskb))
401
		goto none;
402 403 404 405
	skb_pull(*iskb, offset);
	imsz = msg_size(buf_msg(*iskb));
	skb_trim(*iskb, imsz);
	if (unlikely(!tipc_msg_validate(*iskb)))
406 407 408 409 410
		goto none;
	*pos += align(imsz);
	return true;
none:
	kfree_skb(skb);
411
	kfree_skb(*iskb);
412 413 414 415
	*iskb = NULL;
	return false;
}

416 417
/**
 * tipc_msg_make_bundle(): Create bundle buf and append message to its tail
418 419 420
 * @list: the buffer chain
 * @skb: buffer to be appended and replaced
 * @mtu: max allowable size for the bundle buffer, inclusive header
421 422
 * @dnode: destination node for message. (Not always present in header)
 * Replaces buffer if successful
S
stephen hemminger 已提交
423
 * Returns true if success, otherwise false
424
 */
J
Jon Paul Maloy 已提交
425
bool tipc_msg_make_bundle(struct sk_buff **skb, u32 mtu, u32 dnode)
426
{
427
	struct sk_buff *bskb;
428
	struct tipc_msg *bmsg;
J
Jon Paul Maloy 已提交
429
	struct tipc_msg *msg = buf_msg(*skb);
430 431 432 433 434 435 436 437 438 439 440 441
	u32 msz = msg_size(msg);
	u32 max = mtu - INT_H_SIZE;

	if (msg_user(msg) == MSG_FRAGMENTER)
		return false;
	if (msg_user(msg) == CHANGEOVER_PROTOCOL)
		return false;
	if (msg_user(msg) == BCAST_PROTOCOL)
		return false;
	if (msz > (max / 2))
		return false;

442 443
	bskb = tipc_buf_acquire(max);
	if (!bskb)
444 445
		return false;

446 447
	skb_trim(bskb, INT_H_SIZE);
	bmsg = buf_msg(bskb);
448 449
	tipc_msg_init(msg_prevnode(msg), bmsg, MSG_BUNDLER, 0,
		      INT_H_SIZE, dnode);
450 451 452
	msg_set_seqno(bmsg, msg_seqno(msg));
	msg_set_ack(bmsg, msg_ack(msg));
	msg_set_bcast_ack(bmsg, msg_bcast_ack(msg));
J
Jon Paul Maloy 已提交
453 454 455
	tipc_msg_bundle(bskb, *skb, mtu);
	*skb = bskb;
	return true;
456
}
457 458 459 460 461 462 463 464 465

/**
 * tipc_msg_reverse(): swap source and destination addresses and add error code
 * @buf:  buffer containing message to be reversed
 * @dnode: return value: node where to send message after reversal
 * @err:  error code to be set in message
 * Consumes buffer if failure
 * Returns true if success, otherwise false
 */
466
bool tipc_msg_reverse(u32 own_addr,  struct sk_buff *buf, u32 *dnode,
467
		      int err)
468 469 470 471 472 473
{
	struct tipc_msg *msg = buf_msg(buf);
	uint imp = msg_importance(msg);
	struct tipc_msg ohdr;
	uint rdsz = min_t(uint, msg_data_sz(msg), MAX_FORWARD_SIZE);

474
	if (skb_linearize(buf))
475
		goto exit;
476
	msg = buf_msg(buf);
477 478 479
	if (msg_dest_droppable(msg))
		goto exit;
	if (msg_errcode(msg))
480 481
		goto exit;
	memcpy(&ohdr, msg, msg_hdr_sz(msg));
482 483 484
	imp = min_t(uint, imp + 1, TIPC_CRITICAL_IMPORTANCE);
	if (msg_isdata(msg))
		msg_set_importance(msg, imp);
485 486 487
	msg_set_errcode(msg, err);
	msg_set_origport(msg, msg_destport(&ohdr));
	msg_set_destport(msg, msg_origport(&ohdr));
488
	msg_set_prevnode(msg, own_addr);
489 490 491 492 493 494 495 496 497 498 499
	if (!msg_short(msg)) {
		msg_set_orignode(msg, msg_destnode(&ohdr));
		msg_set_destnode(msg, msg_orignode(&ohdr));
	}
	msg_set_size(msg, msg_hdr_sz(msg) + rdsz);
	skb_trim(buf, msg_size(msg));
	skb_orphan(buf);
	*dnode = msg_orignode(&ohdr);
	return true;
exit:
	kfree_skb(buf);
500
	*dnode = 0;
501 502
	return false;
}
503 504

/**
505 506 507 508
 * tipc_msg_lookup_dest(): try to find new destination for named message
 * @skb: the buffer containing the message.
 * @dnode: return value: next-hop node, if destination found
 * @err: return value: error code to use, if message to be rejected
509
 * Does not consume buffer
510
 * Returns true if a destination is found, false otherwise
511
 */
512 513
bool tipc_msg_lookup_dest(struct net *net, struct sk_buff *skb,
			  u32 *dnode, int *err)
514
{
515
	struct tipc_msg *msg = buf_msg(skb);
516 517
	u32 dport;

518 519 520 521 522 523 524
	if (!msg_isdata(msg))
		return false;
	if (!msg_named(msg))
		return false;
	*err = -TIPC_ERR_NO_NAME;
	if (skb_linearize(skb))
		return false;
525
	if (msg_reroute_cnt(msg) > 0)
526
		return false;
527
	*dnode = addr_domain(net, msg_lookup_scope(msg));
528
	dport = tipc_nametbl_translate(net, msg_nametype(msg),
529
				       msg_nameinst(msg), dnode);
530
	if (!dport)
531
		return false;
532 533 534
	msg_incr_reroute_cnt(msg);
	msg_set_destnode(msg, *dnode);
	msg_set_destport(msg, dport);
535 536
	*err = TIPC_OK;
	return true;
537
}
538 539 540 541

/* tipc_msg_reassemble() - clone a buffer chain of fragments and
 *                         reassemble the clones into one message
 */
542
struct sk_buff *tipc_msg_reassemble(struct sk_buff_head *list)
543
{
544 545
	struct sk_buff *skb;
	struct sk_buff *frag = NULL;
546 547 548 549
	struct sk_buff *head = NULL;
	int hdr_sz;

	/* Copy header if single buffer */
550 551 552 553
	if (skb_queue_len(list) == 1) {
		skb = skb_peek(list);
		hdr_sz = skb_headroom(skb) + msg_hdr_sz(buf_msg(skb));
		return __pskb_copy(skb, hdr_sz, GFP_ATOMIC);
554 555 556
	}

	/* Clone all fragments and reassemble */
557 558
	skb_queue_walk(list, skb) {
		frag = skb_clone(skb, GFP_ATOMIC);
559 560 561 562 563 564 565 566 567 568 569 570 571 572
		if (!frag)
			goto error;
		frag->next = NULL;
		if (tipc_buf_append(&head, &frag))
			break;
		if (!head)
			goto error;
	}
	return frag;
error:
	pr_warn("Failed do clone local mcast rcv buffer\n");
	kfree_skb(head);
	return NULL;
}