reassembly.c 18.5 KB
Newer Older
L
Linus Torvalds 已提交
1 2
/*
 *	IPv6 fragment reassembly
3
 *	Linux INET6 implementation
L
Linus Torvalds 已提交
4 5
 *
 *	Authors:
6
 *	Pedro Roque		<roque@di.fc.ul.pt>
L
Linus Torvalds 已提交
7 8 9 10 11 12 13 14 15 16 17
 *
 *	$Id: reassembly.c,v 1.26 2001/03/07 22:00:57 davem Exp $
 *
 *	Based on: net/ipv4/ip_fragment.c
 *
 *	This program is free software; you can redistribute it and/or
 *      modify it under the terms of the GNU General Public License
 *      as published by the Free Software Foundation; either version
 *      2 of the License, or (at your option) any later version.
 */

18 19
/*
 *	Fixes:
L
Linus Torvalds 已提交
20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44
 *	Andi Kleen	Make it work with multiple hosts.
 *			More RFC compliance.
 *
 *      Horst von Brand Add missing #include <linux/string.h>
 *	Alexey Kuznetsov	SMP races, threading, cleanup.
 *	Patrick McHardy		LRU queue of frag heads for evictor.
 *	Mitsuru KANDA @USAGI	Register inet6_protocol{}.
 *	David Stevens and
 *	YOSHIFUJI,H. @USAGI	Always remove fragment header to
 *				calculate ICV correctly.
 */
#include <linux/errno.h>
#include <linux/types.h>
#include <linux/string.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/jiffies.h>
#include <linux/net.h>
#include <linux/list.h>
#include <linux/netdevice.h>
#include <linux/in6.h>
#include <linux/ipv6.h>
#include <linux/icmpv6.h>
#include <linux/random.h>
#include <linux/jhash.h>
45
#include <linux/skbuff.h>
L
Linus Torvalds 已提交
46 47 48 49 50

#include <net/sock.h>
#include <net/snmp.h>

#include <net/ipv6.h>
51
#include <net/ip6_route.h>
L
Linus Torvalds 已提交
52 53 54 55 56
#include <net/protocol.h>
#include <net/transp_v6.h>
#include <net/rawv6.h>
#include <net/ndisc.h>
#include <net/addrconf.h>
57
#include <net/inet_frag.h>
L
Linus Torvalds 已提交
58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73

struct ip6frag_skb_cb
{
	struct inet6_skb_parm	h;
	int			offset;
};

#define FRAG6_CB(skb)	((struct ip6frag_skb_cb*)((skb)->cb))


/*
 *	Equivalent of ipv4 struct ipq
 */

struct frag_queue
{
74
	struct inet_frag_queue	q;
L
Linus Torvalds 已提交
75

A
Al Viro 已提交
76
	__be32			id;		/* fragment id		*/
L
Linus Torvalds 已提交
77 78 79 80 81 82 83 84
	struct in6_addr		saddr;
	struct in6_addr		daddr;

	int			iif;
	unsigned int		csum;
	__u16			nhoffset;
};

85
static struct inet_frags ip6_frags;
L
Linus Torvalds 已提交
86

87
int ip6_frag_nqueues(struct net *net)
88
{
89
	return net->ipv6.frags.nqueues;
90
}
L
Linus Torvalds 已提交
91

92
int ip6_frag_mem(struct net *net)
93
{
94
	return atomic_read(&net->ipv6.frags.mem);
95
}
L
Linus Torvalds 已提交
96

97 98 99
static int ip6_frag_reasm(struct frag_queue *fq, struct sk_buff *prev,
			  struct net_device *dev);

100 101 102 103
/*
 * callers should be careful not to use the hash value outside the ipfrag_lock
 * as doing so could race with ipfrag_hash_rnd being recalculated.
 */
A
Al Viro 已提交
104
static unsigned int ip6qhashfn(__be32 id, struct in6_addr *saddr,
L
Linus Torvalds 已提交
105 106 107 108
			       struct in6_addr *daddr)
{
	u32 a, b, c;

A
Al Viro 已提交
109 110 111
	a = (__force u32)saddr->s6_addr32[0];
	b = (__force u32)saddr->s6_addr32[1];
	c = (__force u32)saddr->s6_addr32[2];
L
Linus Torvalds 已提交
112 113 114

	a += JHASH_GOLDEN_RATIO;
	b += JHASH_GOLDEN_RATIO;
115
	c += ip6_frags.rnd;
L
Linus Torvalds 已提交
116 117
	__jhash_mix(a, b, c);

A
Al Viro 已提交
118 119 120
	a += (__force u32)saddr->s6_addr32[3];
	b += (__force u32)daddr->s6_addr32[0];
	c += (__force u32)daddr->s6_addr32[1];
L
Linus Torvalds 已提交
121 122
	__jhash_mix(a, b, c);

A
Al Viro 已提交
123 124 125
	a += (__force u32)daddr->s6_addr32[2];
	b += (__force u32)daddr->s6_addr32[3];
	c += (__force u32)id;
L
Linus Torvalds 已提交
126 127
	__jhash_mix(a, b, c);

128
	return c & (INETFRAGS_HASHSZ - 1);
L
Linus Torvalds 已提交
129 130
}

131
static unsigned int ip6_hashfn(struct inet_frag_queue *q)
L
Linus Torvalds 已提交
132
{
133
	struct frag_queue *fq;
L
Linus Torvalds 已提交
134

135 136
	fq = container_of(q, struct frag_queue, q);
	return ip6qhashfn(fq->id, &fq->saddr, &fq->daddr);
L
Linus Torvalds 已提交
137 138
}

139 140 141 142 143 144 145 146 147 148 149 150
int ip6_frag_match(struct inet_frag_queue *q, void *a)
{
	struct frag_queue *fq;
	struct ip6_create_arg *arg = a;

	fq = container_of(q, struct frag_queue, q);
	return (fq->id == arg->id &&
			ipv6_addr_equal(&fq->saddr, arg->src) &&
			ipv6_addr_equal(&fq->daddr, arg->dst));
}
EXPORT_SYMBOL(ip6_frag_match);

L
Linus Torvalds 已提交
151
/* Memory Tracking Functions. */
152 153
static inline void frag_kfree_skb(struct netns_frags *nf,
		struct sk_buff *skb, int *work)
L
Linus Torvalds 已提交
154 155 156
{
	if (work)
		*work -= skb->truesize;
157
	atomic_sub(skb->truesize, &nf->mem);
L
Linus Torvalds 已提交
158 159 160
	kfree_skb(skb);
}

161
void ip6_frag_init(struct inet_frag_queue *q, void *a)
L
Linus Torvalds 已提交
162
{
163 164 165 166 167 168
	struct frag_queue *fq = container_of(q, struct frag_queue, q);
	struct ip6_create_arg *arg = a;

	fq->id = arg->id;
	ipv6_addr_copy(&fq->saddr, arg->src);
	ipv6_addr_copy(&fq->daddr, arg->dst);
L
Linus Torvalds 已提交
169
}
170
EXPORT_SYMBOL(ip6_frag_init);
L
Linus Torvalds 已提交
171 172 173

/* Destruction primitives. */

174
static __inline__ void fq_put(struct frag_queue *fq)
L
Linus Torvalds 已提交
175
{
P
Pavel Emelyanov 已提交
176
	inet_frag_put(&fq->q, &ip6_frags);
L
Linus Torvalds 已提交
177 178 179 180 181 182 183
}

/* Kill fq entry. It is not destroyed immediately,
 * because caller (and someone more) holds reference count.
 */
static __inline__ void fq_kill(struct frag_queue *fq)
{
184
	inet_frag_kill(&fq->q, &ip6_frags);
L
Linus Torvalds 已提交
185 186
}

187
static void ip6_evictor(struct net *net, struct inet6_dev *idev)
L
Linus Torvalds 已提交
188
{
189 190
	int evicted;

191
	evicted = inet_frag_evictor(&net->ipv6.frags, &ip6_frags);
192 193
	if (evicted)
		IP6_ADD_STATS_BH(idev, IPSTATS_MIB_REASMFAILS, evicted);
L
Linus Torvalds 已提交
194 195 196 197
}

static void ip6_frag_expire(unsigned long data)
{
198
	struct frag_queue *fq;
199
	struct net_device *dev = NULL;
L
Linus Torvalds 已提交
200

201 202
	fq = container_of((struct inet_frag_queue *)data, struct frag_queue, q);

203
	spin_lock(&fq->q.lock);
L
Linus Torvalds 已提交
204

205
	if (fq->q.last_in & COMPLETE)
L
Linus Torvalds 已提交
206 207 208 209
		goto out;

	fq_kill(fq);

210
	dev = dev_get_by_index(&init_net, fq->iif);
211 212 213 214 215 216 217
	if (!dev)
		goto out;

	rcu_read_lock();
	IP6_INC_STATS_BH(__in6_dev_get(dev), IPSTATS_MIB_REASMTIMEOUT);
	IP6_INC_STATS_BH(__in6_dev_get(dev), IPSTATS_MIB_REASMFAILS);
	rcu_read_unlock();
L
Linus Torvalds 已提交
218

219
	/* Don't send error if the first segment did not arrive. */
220
	if (!(fq->q.last_in&FIRST_IN) || !fq->q.fragments)
221 222 223 224 225 226 227
		goto out;

	/*
	   But use as source device on which LAST ARRIVED
	   segment was received. And do not use fq->dev
	   pointer directly, device might already disappeared.
	 */
228 229
	fq->q.fragments->dev = dev;
	icmpv6_send(fq->q.fragments, ICMPV6_TIME_EXCEED, ICMPV6_EXC_FRAGTIME, 0, dev);
L
Linus Torvalds 已提交
230
out:
231 232
	if (dev)
		dev_put(dev);
233
	spin_unlock(&fq->q.lock);
234
	fq_put(fq);
L
Linus Torvalds 已提交
235 236
}

237
static __inline__ struct frag_queue *
238
fq_find(struct net *net, __be32 id, struct in6_addr *src, struct in6_addr *dst,
239
	struct inet6_dev *idev)
L
Linus Torvalds 已提交
240
{
241 242
	struct inet_frag_queue *q;
	struct ip6_create_arg arg;
243
	unsigned int hash;
L
Linus Torvalds 已提交
244

245 246 247
	arg.id = id;
	arg.src = src;
	arg.dst = dst;
248
	hash = ip6qhashfn(id, src, dst);
L
Linus Torvalds 已提交
249

250
	q = inet_frag_find(&net->ipv6.frags, &ip6_frags, &arg, hash);
251 252
	if (q == NULL)
		goto oom;
L
Linus Torvalds 已提交
253

254
	return container_of(q, struct frag_queue, q);
L
Linus Torvalds 已提交
255 256

oom:
257
	IP6_INC_STATS_BH(idev, IPSTATS_MIB_REASMFAILS);
L
Linus Torvalds 已提交
258 259 260
	return NULL;
}

261
static int ip6_frag_queue(struct frag_queue *fq, struct sk_buff *skb,
L
Linus Torvalds 已提交
262 263 264
			   struct frag_hdr *fhdr, int nhoff)
{
	struct sk_buff *prev, *next;
265
	struct net_device *dev;
L
Linus Torvalds 已提交
266 267
	int offset, end;

268
	if (fq->q.last_in & COMPLETE)
L
Linus Torvalds 已提交
269 270 271
		goto err;

	offset = ntohs(fhdr->frag_off) & ~0x7;
272 273
	end = offset + (ntohs(ipv6_hdr(skb)->payload_len) -
			((u8 *)(fhdr + 1) - (u8 *)(ipv6_hdr(skb) + 1)));
L
Linus Torvalds 已提交
274 275

	if ((unsigned int)end > IPV6_MAXPLEN) {
276 277
		IP6_INC_STATS_BH(ip6_dst_idev(skb->dst),
				 IPSTATS_MIB_INHDRERRORS);
278 279 280
		icmpv6_param_prob(skb, ICMPV6_HDR_FIELD,
				  ((u8 *)&fhdr->frag_off -
				   skb_network_header(skb)));
281
		return -1;
L
Linus Torvalds 已提交
282 283
	}

284 285
	if (skb->ip_summed == CHECKSUM_COMPLETE) {
		const unsigned char *nh = skb_network_header(skb);
286
		skb->csum = csum_sub(skb->csum,
287 288 289
				     csum_partial(nh, (u8 *)(fhdr + 1) - nh,
						  0));
	}
L
Linus Torvalds 已提交
290 291 292 293 294 295

	/* Is this the final fragment? */
	if (!(fhdr->frag_off & htons(IP6_MF))) {
		/* If we already have some bits beyond end
		 * or have different end, the segment is corrupted.
		 */
296 297
		if (end < fq->q.len ||
		    ((fq->q.last_in & LAST_IN) && end != fq->q.len))
L
Linus Torvalds 已提交
298
			goto err;
299 300
		fq->q.last_in |= LAST_IN;
		fq->q.len = end;
L
Linus Torvalds 已提交
301 302 303 304 305 306 307 308
	} else {
		/* Check if the fragment is rounded to 8 bytes.
		 * Required by the RFC.
		 */
		if (end & 0x7) {
			/* RFC2460 says always send parameter problem in
			 * this case. -DaveM
			 */
309 310
			IP6_INC_STATS_BH(ip6_dst_idev(skb->dst),
					 IPSTATS_MIB_INHDRERRORS);
311
			icmpv6_param_prob(skb, ICMPV6_HDR_FIELD,
L
Linus Torvalds 已提交
312
					  offsetof(struct ipv6hdr, payload_len));
313
			return -1;
L
Linus Torvalds 已提交
314
		}
315
		if (end > fq->q.len) {
L
Linus Torvalds 已提交
316
			/* Some bits beyond end -> corruption. */
317
			if (fq->q.last_in & LAST_IN)
L
Linus Torvalds 已提交
318
				goto err;
319
			fq->q.len = end;
L
Linus Torvalds 已提交
320 321 322 323 324 325 326 327 328
		}
	}

	if (end == offset)
		goto err;

	/* Point into the IP datagram 'data' part. */
	if (!pskb_pull(skb, (u8 *) (fhdr + 1) - skb->data))
		goto err;
329

330 331
	if (pskb_trim_rcsum(skb, end - offset))
		goto err;
L
Linus Torvalds 已提交
332 333 334 335 336 337

	/* Find out which fragments are in front and at the back of us
	 * in the chain of fragments so far.  We must know where to put
	 * this fragment, right?
	 */
	prev = NULL;
338
	for(next = fq->q.fragments; next != NULL; next = next->next) {
L
Linus Torvalds 已提交
339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374
		if (FRAG6_CB(next)->offset >= offset)
			break;	/* bingo! */
		prev = next;
	}

	/* We found where to put this one.  Check for overlap with
	 * preceding fragment, and, if needed, align things so that
	 * any overlaps are eliminated.
	 */
	if (prev) {
		int i = (FRAG6_CB(prev)->offset + prev->len) - offset;

		if (i > 0) {
			offset += i;
			if (end <= offset)
				goto err;
			if (!pskb_pull(skb, i))
				goto err;
			if (skb->ip_summed != CHECKSUM_UNNECESSARY)
				skb->ip_summed = CHECKSUM_NONE;
		}
	}

	/* Look for overlap with succeeding segments.
	 * If we can merge fragments, do it.
	 */
	while (next && FRAG6_CB(next)->offset < end) {
		int i = end - FRAG6_CB(next)->offset; /* overlap is 'i' bytes */

		if (i < next->len) {
			/* Eat head of the next overlapped fragment
			 * and leave the loop. The next ones cannot overlap.
			 */
			if (!pskb_pull(next, i))
				goto err;
			FRAG6_CB(next)->offset += i;	/* next fragment */
375
			fq->q.meat -= i;
L
Linus Torvalds 已提交
376 377 378 379 380 381 382 383 384 385 386 387 388 389
			if (next->ip_summed != CHECKSUM_UNNECESSARY)
				next->ip_summed = CHECKSUM_NONE;
			break;
		} else {
			struct sk_buff *free_it = next;

			/* Old fragment is completely overridden with
			 * new one drop it.
			 */
			next = next->next;

			if (prev)
				prev->next = next;
			else
390
				fq->q.fragments = next;
L
Linus Torvalds 已提交
391

392
			fq->q.meat -= free_it->len;
393
			frag_kfree_skb(fq->q.net, free_it, NULL);
L
Linus Torvalds 已提交
394 395 396 397 398 399 400 401 402 403
		}
	}

	FRAG6_CB(skb)->offset = offset;

	/* Insert this fragment in the chain of fragments. */
	skb->next = next;
	if (prev)
		prev->next = skb;
	else
404
		fq->q.fragments = skb;
L
Linus Torvalds 已提交
405

406 407 408 409 410
	dev = skb->dev;
	if (dev) {
		fq->iif = dev->ifindex;
		skb->dev = NULL;
	}
411 412
	fq->q.stamp = skb->tstamp;
	fq->q.meat += skb->len;
413
	atomic_add(skb->truesize, &fq->q.net->mem);
L
Linus Torvalds 已提交
414 415 416 417 418 419

	/* The first fragment.
	 * nhoffset is obtained from the first fragment, of course.
	 */
	if (offset == 0) {
		fq->nhoffset = nhoff;
420
		fq->q.last_in |= FIRST_IN;
L
Linus Torvalds 已提交
421
	}
422

423
	if (fq->q.last_in == (FIRST_IN | LAST_IN) && fq->q.meat == fq->q.len)
424 425
		return ip6_frag_reasm(fq, prev, dev);

426 427 428
	write_lock(&ip6_frags.lock);
	list_move_tail(&fq->q.lru_list, &ip6_frags.lru_list);
	write_unlock(&ip6_frags.lock);
429
	return -1;
L
Linus Torvalds 已提交
430 431

err:
432
	IP6_INC_STATS(ip6_dst_idev(skb->dst), IPSTATS_MIB_REASMFAILS);
L
Linus Torvalds 已提交
433
	kfree_skb(skb);
434
	return -1;
L
Linus Torvalds 已提交
435 436 437 438 439 440 441 442 443 444 445
}

/*
 *	Check if this packet is complete.
 *	Returns NULL on failure by any reason, and pointer
 *	to current nexthdr field in reassembled frame.
 *
 *	It is called with locked fq, and caller must check that
 *	queue is eligible for reassembly i.e. it is not COMPLETE,
 *	the last and the first frames arrived and all the bits are here.
 */
446
static int ip6_frag_reasm(struct frag_queue *fq, struct sk_buff *prev,
L
Linus Torvalds 已提交
447 448
			  struct net_device *dev)
{
449
	struct sk_buff *fp, *head = fq->q.fragments;
L
Linus Torvalds 已提交
450 451 452 453 454
	int    payload_len;
	unsigned int nhoff;

	fq_kill(fq);

455 456 457 458 459 460 461 462 463 464 465
	/* Make the one we just received the head. */
	if (prev) {
		head = prev->next;
		fp = skb_clone(head, GFP_ATOMIC);

		if (!fp)
			goto out_oom;

		fp->next = head->next;
		prev->next = fp;

466 467
		skb_morph(head, fq->q.fragments);
		head->next = fq->q.fragments->next;
468

469 470
		kfree_skb(fq->q.fragments);
		fq->q.fragments = head;
471 472
	}

L
Linus Torvalds 已提交
473 474 475 476
	BUG_TRAP(head != NULL);
	BUG_TRAP(FRAG6_CB(head)->offset == 0);

	/* Unfragmented part is taken from the first segment. */
477
	payload_len = ((head->data - skb_network_header(head)) -
478
		       sizeof(struct ipv6hdr) + fq->q.len -
479
		       sizeof(struct frag_hdr));
L
Linus Torvalds 已提交
480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506
	if (payload_len > IPV6_MAXPLEN)
		goto out_oversize;

	/* Head of list must not be cloned. */
	if (skb_cloned(head) && pskb_expand_head(head, 0, 0, GFP_ATOMIC))
		goto out_oom;

	/* If the first fragment is fragmented itself, we split
	 * it to two chunks: the first with data and paged part
	 * and the second, holding only fragments. */
	if (skb_shinfo(head)->frag_list) {
		struct sk_buff *clone;
		int i, plen = 0;

		if ((clone = alloc_skb(0, GFP_ATOMIC)) == NULL)
			goto out_oom;
		clone->next = head->next;
		head->next = clone;
		skb_shinfo(clone)->frag_list = skb_shinfo(head)->frag_list;
		skb_shinfo(head)->frag_list = NULL;
		for (i=0; i<skb_shinfo(head)->nr_frags; i++)
			plen += skb_shinfo(head)->frags[i].size;
		clone->len = clone->data_len = head->data_len - plen;
		head->data_len -= clone->len;
		head->len -= clone->len;
		clone->csum = 0;
		clone->ip_summed = head->ip_summed;
507
		atomic_add(clone->truesize, &fq->q.net->mem);
L
Linus Torvalds 已提交
508 509 510 511 512
	}

	/* We have to remove fragment header from datagram and to relocate
	 * header in order to calculate ICV correctly. */
	nhoff = fq->nhoffset;
513
	skb_network_header(head)[nhoff] = skb_transport_header(head)[0];
514
	memmove(head->head + sizeof(struct frag_hdr), head->head,
L
Linus Torvalds 已提交
515
		(head->data - head->head) - sizeof(struct frag_hdr));
516 517
	head->mac_header += sizeof(struct frag_hdr);
	head->network_header += sizeof(struct frag_hdr);
L
Linus Torvalds 已提交
518 519

	skb_shinfo(head)->frag_list = head->next;
520
	skb_reset_transport_header(head);
521
	skb_push(head, head->data - skb_network_header(head));
522
	atomic_sub(head->truesize, &fq->q.net->mem);
L
Linus Torvalds 已提交
523 524 525 526 527 528

	for (fp=head->next; fp; fp = fp->next) {
		head->data_len += fp->len;
		head->len += fp->len;
		if (head->ip_summed != fp->ip_summed)
			head->ip_summed = CHECKSUM_NONE;
529
		else if (head->ip_summed == CHECKSUM_COMPLETE)
L
Linus Torvalds 已提交
530 531
			head->csum = csum_add(head->csum, fp->csum);
		head->truesize += fp->truesize;
532
		atomic_sub(fp->truesize, &fq->q.net->mem);
L
Linus Torvalds 已提交
533 534 535 536
	}

	head->next = NULL;
	head->dev = dev;
537
	head->tstamp = fq->q.stamp;
538
	ipv6_hdr(head)->payload_len = htons(payload_len);
539
	IP6CB(head)->nhoff = nhoff;
L
Linus Torvalds 已提交
540 541

	/* Yes, and fold redundant checksum back. 8) */
542
	if (head->ip_summed == CHECKSUM_COMPLETE)
543
		head->csum = csum_partial(skb_network_header(head),
544
					  skb_network_header_len(head),
545
					  head->csum);
L
Linus Torvalds 已提交
546

547 548 549
	rcu_read_lock();
	IP6_INC_STATS_BH(__in6_dev_get(dev), IPSTATS_MIB_REASMOKS);
	rcu_read_unlock();
550
	fq->q.fragments = NULL;
L
Linus Torvalds 已提交
551 552 553 554 555 556 557 558 559 560
	return 1;

out_oversize:
	if (net_ratelimit())
		printk(KERN_DEBUG "ip6_frag_reasm: payload len = %d\n", payload_len);
	goto out_fail;
out_oom:
	if (net_ratelimit())
		printk(KERN_DEBUG "ip6_frag_reasm: no memory for reassembly\n");
out_fail:
561 562 563
	rcu_read_lock();
	IP6_INC_STATS_BH(__in6_dev_get(dev), IPSTATS_MIB_REASMFAILS);
	rcu_read_unlock();
L
Linus Torvalds 已提交
564 565 566
	return -1;
}

567
static int ipv6_frag_rcv(struct sk_buff *skb)
L
Linus Torvalds 已提交
568 569 570
{
	struct frag_hdr *fhdr;
	struct frag_queue *fq;
571
	struct ipv6hdr *hdr = ipv6_hdr(skb);
572
	struct net *net;
L
Linus Torvalds 已提交
573

574
	IP6_INC_STATS_BH(ip6_dst_idev(skb->dst), IPSTATS_MIB_REASMREQDS);
L
Linus Torvalds 已提交
575 576 577

	/* Jumbo payload inhibits frag. header */
	if (hdr->payload_len==0) {
578
		IP6_INC_STATS(ip6_dst_idev(skb->dst), IPSTATS_MIB_INHDRERRORS);
579 580
		icmpv6_param_prob(skb, ICMPV6_HDR_FIELD,
				  skb_network_header_len(skb));
L
Linus Torvalds 已提交
581 582
		return -1;
	}
583 584
	if (!pskb_may_pull(skb, (skb_transport_offset(skb) +
				 sizeof(struct frag_hdr)))) {
585
		IP6_INC_STATS(ip6_dst_idev(skb->dst), IPSTATS_MIB_INHDRERRORS);
586 587
		icmpv6_param_prob(skb, ICMPV6_HDR_FIELD,
				  skb_network_header_len(skb));
L
Linus Torvalds 已提交
588 589 590
		return -1;
	}

591
	hdr = ipv6_hdr(skb);
592
	fhdr = (struct frag_hdr *)skb_transport_header(skb);
L
Linus Torvalds 已提交
593 594 595

	if (!(fhdr->frag_off & htons(0xFFF9))) {
		/* It is not a fragmented frame */
596
		skb->transport_header += sizeof(struct frag_hdr);
597
		IP6_INC_STATS_BH(ip6_dst_idev(skb->dst), IPSTATS_MIB_REASMOKS);
L
Linus Torvalds 已提交
598

599
		IP6CB(skb)->nhoff = (u8 *)fhdr - skb_network_header(skb);
L
Linus Torvalds 已提交
600 601 602
		return 1;
	}

603
	net = skb->dev->nd_net;
604
	if (atomic_read(&net->ipv6.frags.mem) > net->ipv6.frags.high_thresh)
605
		ip6_evictor(net, ip6_dst_idev(skb->dst));
L
Linus Torvalds 已提交
606

607
	if ((fq = fq_find(net, fhdr->identification, &hdr->saddr, &hdr->daddr,
608
			  ip6_dst_idev(skb->dst))) != NULL) {
609
		int ret;
L
Linus Torvalds 已提交
610

611
		spin_lock(&fq->q.lock);
L
Linus Torvalds 已提交
612

613
		ret = ip6_frag_queue(fq, skb, fhdr, IP6CB(skb)->nhoff);
L
Linus Torvalds 已提交
614

615
		spin_unlock(&fq->q.lock);
616
		fq_put(fq);
L
Linus Torvalds 已提交
617 618 619
		return ret;
	}

620
	IP6_INC_STATS_BH(ip6_dst_idev(skb->dst), IPSTATS_MIB_REASMFAILS);
L
Linus Torvalds 已提交
621 622 623 624 625 626 627 628 629 630
	kfree_skb(skb);
	return -1;
}

static struct inet6_protocol frag_protocol =
{
	.handler	=	ipv6_frag_rcv,
	.flags		=	INET6_PROTO_NOPOLICY,
};

631 632 633 634 635
#ifdef CONFIG_SYSCTL
static struct ctl_table ip6_frags_ctl_table[] = {
	{
		.ctl_name	= NET_IPV6_IP6FRAG_HIGH_THRESH,
		.procname	= "ip6frag_high_thresh",
636
		.data		= &init_net.ipv6.frags.high_thresh,
637 638 639 640 641 642 643
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= &proc_dointvec
	},
	{
		.ctl_name	= NET_IPV6_IP6FRAG_LOW_THRESH,
		.procname	= "ip6frag_low_thresh",
644
		.data		= &init_net.ipv6.frags.low_thresh,
645 646 647 648 649 650 651
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= &proc_dointvec
	},
	{
		.ctl_name	= NET_IPV6_IP6FRAG_TIME,
		.procname	= "ip6frag_time",
652
		.data		= &init_net.ipv6.frags.timeout,
653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= &proc_dointvec_jiffies,
		.strategy	= &sysctl_jiffies,
	},
	{
		.ctl_name	= NET_IPV6_IP6FRAG_SECRET_INTERVAL,
		.procname	= "ip6frag_secret_interval",
		.data		= &init_net.ipv6.sysctl.frags.secret_interval,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= &proc_dointvec_jiffies,
		.strategy	= &sysctl_jiffies
	},
	{ }
};

static int ip6_frags_sysctl_register(struct net *net)
{
672
	struct ctl_table *table;
673 674
	struct ctl_table_header *hdr;

675 676 677 678 679 680
	table = ip6_frags_ctl_table;
	if (net != &init_net) {
		table = kmemdup(table, sizeof(ip6_frags_ctl_table), GFP_KERNEL);
		if (table == NULL)
			goto err_alloc;

681 682
		table[0].data = &net->ipv6.frags.high_thresh;
		table[1].data = &net->ipv6.frags.low_thresh;
683
		table[2].data = &net->ipv6.frags.timeout;
684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707
		table[3].mode &= ~0222;
	}

	hdr = register_net_sysctl_table(net, net_ipv6_ctl_path, table);
	if (hdr == NULL)
		goto err_reg;

	net->ipv6.sysctl.frags_hdr = hdr;
	return 0;

err_reg:
	if (net != &init_net)
		kfree(table);
err_alloc:
	return -ENOMEM;
}

static void ip6_frags_sysctl_unregister(struct net *net)
{
	struct ctl_table *table;

	table = net->ipv6.sysctl.frags_hdr->ctl_table_arg;
	unregister_net_sysctl_table(net->ipv6.sysctl.frags_hdr);
	kfree(table);
708 709 710
}
#else
static inline int ip6_frags_sysctl_register(struct net *net)
711
{
712 713
	return 0;
}
714 715 716 717

static inline void ip6_frags_sysctl_unregister(struct net *net)
{
}
718
#endif
D
Daniel Lezcano 已提交
719

720 721
static int ipv6_frags_init_net(struct net *net)
{
722
	ip6_frags.ctl = &net->ipv6.sysctl.frags;
723

724 725
	net->ipv6.frags.high_thresh = 256 * 1024;
	net->ipv6.frags.low_thresh = 192 * 1024;
726
	net->ipv6.frags.timeout = IPV6_FRAG_TIMEOUT;
727 728
	net->ipv6.sysctl.frags.secret_interval = 10 * 60 * HZ;

729 730
	inet_frags_init_net(&net->ipv6.frags);

731
	return ip6_frags_sysctl_register(net);
732 733
}

734
int __init ipv6_frag_init(void)
L
Linus Torvalds 已提交
735
{
736
	int ret;
L
Linus Torvalds 已提交
737

738 739 740
	ret = inet6_add_protocol(&frag_protocol, IPPROTO_FRAGMENT);
	if (ret)
		goto out;
741

742 743
	ipv6_frags_init_net(&init_net);

744
	ip6_frags.hashfn = ip6_hashfn;
745
	ip6_frags.constructor = ip6_frag_init;
746
	ip6_frags.destructor = NULL;
747 748
	ip6_frags.skb_free = NULL;
	ip6_frags.qsize = sizeof(struct frag_queue);
749
	ip6_frags.match = ip6_frag_match;
750
	ip6_frags.frag_expire = ip6_frag_expire;
751
	inet_frags_init(&ip6_frags);
752 753 754 755 756 757 758 759
out:
	return ret;
}

void ipv6_frag_exit(void)
{
	inet_frags_fini(&ip6_frags);
	inet6_del_protocol(&frag_protocol, IPPROTO_FRAGMENT);
L
Linus Torvalds 已提交
760
}