reassembly.c 19.1 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 45 46 47 48 49
 *	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>

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

#include <net/ipv6.h>
50
#include <net/ip6_route.h>
L
Linus Torvalds 已提交
51 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 58
int sysctl_ip6frag_high_thresh __read_mostly = 256*1024;
int sysctl_ip6frag_low_thresh __read_mostly = 192*1024;
L
Linus Torvalds 已提交
59

60
int sysctl_ip6frag_time __read_mostly = IPV6_FRAG_TIMEOUT;
L
Linus Torvalds 已提交
61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76

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
{
77
	struct hlist_node	list;
L
Linus Torvalds 已提交
78 79
	struct list_head lru_list;		/* lru list member	*/

A
Al Viro 已提交
80
	__be32			id;		/* fragment id		*/
L
Linus Torvalds 已提交
81 82 83 84 85 86 87 88 89 90
	struct in6_addr		saddr;
	struct in6_addr		daddr;

	spinlock_t		lock;
	atomic_t		refcnt;
	struct timer_list	timer;		/* expire timer		*/
	struct sk_buff		*fragments;
	int			len;
	int			meat;
	int			iif;
91
	ktime_t			stamp;
L
Linus Torvalds 已提交
92 93 94 95 96 97 98 99 100 101 102 103
	unsigned int		csum;
	__u8			last_in;	/* has first/last segment arrived? */
#define COMPLETE		4
#define FIRST_IN		2
#define LAST_IN			1
	__u16			nhoffset;
};

/* Hash table. */

#define IP6Q_HASHSZ	64

104
static struct hlist_head ip6_frag_hash[IP6Q_HASHSZ];
L
Linus Torvalds 已提交
105 106 107 108 109 110 111
static DEFINE_RWLOCK(ip6_frag_lock);
static u32 ip6_frag_hash_rnd;
static LIST_HEAD(ip6_frag_lru_list);
int ip6_frag_nqueues = 0;

static __inline__ void __fq_unlink(struct frag_queue *fq)
{
112
	hlist_del(&fq->list);
L
Linus Torvalds 已提交
113 114 115 116 117 118 119 120 121 122 123
	list_del(&fq->lru_list);
	ip6_frag_nqueues--;
}

static __inline__ void fq_unlink(struct frag_queue *fq)
{
	write_lock(&ip6_frag_lock);
	__fq_unlink(fq);
	write_unlock(&ip6_frag_lock);
}

124 125 126 127
/*
 * 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
128
static unsigned int ip6qhashfn(__be32 id, struct in6_addr *saddr,
L
Linus Torvalds 已提交
129 130 131 132
			       struct in6_addr *daddr)
{
	u32 a, b, c;

A
Al Viro 已提交
133 134 135
	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
136 137 138 139 140 141

	a += JHASH_GOLDEN_RATIO;
	b += JHASH_GOLDEN_RATIO;
	c += ip6_frag_hash_rnd;
	__jhash_mix(a, b, c);

A
Al Viro 已提交
142 143 144
	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
145 146
	__jhash_mix(a, b, c);

A
Al Viro 已提交
147 148 149
	a += (__force u32)daddr->s6_addr32[2];
	b += (__force u32)daddr->s6_addr32[3];
	c += (__force u32)id;
L
Linus Torvalds 已提交
150 151 152 153 154 155
	__jhash_mix(a, b, c);

	return c & (IP6Q_HASHSZ - 1);
}

static struct timer_list ip6_frag_secret_timer;
156
int sysctl_ip6frag_secret_interval __read_mostly = 10 * 60 * HZ;
L
Linus Torvalds 已提交
157 158 159 160 161 162 163 164 165 166

static void ip6_frag_secret_rebuild(unsigned long dummy)
{
	unsigned long now = jiffies;
	int i;

	write_lock(&ip6_frag_lock);
	get_random_bytes(&ip6_frag_hash_rnd, sizeof(u32));
	for (i = 0; i < IP6Q_HASHSZ; i++) {
		struct frag_queue *q;
167
		struct hlist_node *p, *n;
L
Linus Torvalds 已提交
168

169
		hlist_for_each_entry_safe(q, p, n, &ip6_frag_hash[i], list) {
L
Linus Torvalds 已提交
170 171 172 173 174
			unsigned int hval = ip6qhashfn(q->id,
						       &q->saddr,
						       &q->daddr);

			if (hval != i) {
175
				hlist_del(&q->list);
L
Linus Torvalds 已提交
176 177

				/* Relink to new hash chain. */
178 179
				hlist_add_head(&q->list,
					       &ip6_frag_hash[hval]);
L
Linus Torvalds 已提交
180

181
			}
L
Linus Torvalds 已提交
182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209
		}
	}
	write_unlock(&ip6_frag_lock);

	mod_timer(&ip6_frag_secret_timer, now + sysctl_ip6frag_secret_interval);
}

atomic_t ip6_frag_mem = ATOMIC_INIT(0);

/* Memory Tracking Functions. */
static inline void frag_kfree_skb(struct sk_buff *skb, int *work)
{
	if (work)
		*work -= skb->truesize;
	atomic_sub(skb->truesize, &ip6_frag_mem);
	kfree_skb(skb);
}

static inline void frag_free_queue(struct frag_queue *fq, int *work)
{
	if (work)
		*work -= sizeof(struct frag_queue);
	atomic_sub(sizeof(struct frag_queue), &ip6_frag_mem);
	kfree(fq);
}

static inline struct frag_queue *frag_alloc_queue(void)
{
210
	struct frag_queue *fq = kzalloc(sizeof(struct frag_queue), GFP_ATOMIC);
L
Linus Torvalds 已提交
211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260

	if(!fq)
		return NULL;
	atomic_add(sizeof(struct frag_queue), &ip6_frag_mem);
	return fq;
}

/* Destruction primitives. */

/* Complete destruction of fq. */
static void ip6_frag_destroy(struct frag_queue *fq, int *work)
{
	struct sk_buff *fp;

	BUG_TRAP(fq->last_in&COMPLETE);
	BUG_TRAP(del_timer(&fq->timer) == 0);

	/* Release all fragment data. */
	fp = fq->fragments;
	while (fp) {
		struct sk_buff *xp = fp->next;

		frag_kfree_skb(fp, work);
		fp = xp;
	}

	frag_free_queue(fq, work);
}

static __inline__ void fq_put(struct frag_queue *fq, int *work)
{
	if (atomic_dec_and_test(&fq->refcnt))
		ip6_frag_destroy(fq, work);
}

/* 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)
{
	if (del_timer(&fq->timer))
		atomic_dec(&fq->refcnt);

	if (!(fq->last_in & COMPLETE)) {
		fq_unlink(fq);
		atomic_dec(&fq->refcnt);
		fq->last_in |= COMPLETE;
	}
}

261
static void ip6_evictor(struct inet6_dev *idev)
L
Linus Torvalds 已提交
262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287
{
	struct frag_queue *fq;
	struct list_head *tmp;
	int work;

	work = atomic_read(&ip6_frag_mem) - sysctl_ip6frag_low_thresh;
	if (work <= 0)
		return;

	while(work > 0) {
		read_lock(&ip6_frag_lock);
		if (list_empty(&ip6_frag_lru_list)) {
			read_unlock(&ip6_frag_lock);
			return;
		}
		tmp = ip6_frag_lru_list.next;
		fq = list_entry(tmp, struct frag_queue, lru_list);
		atomic_inc(&fq->refcnt);
		read_unlock(&ip6_frag_lock);

		spin_lock(&fq->lock);
		if (!(fq->last_in&COMPLETE))
			fq_kill(fq);
		spin_unlock(&fq->lock);

		fq_put(fq, &work);
288
		IP6_INC_STATS_BH(idev, IPSTATS_MIB_REASMFAILS);
L
Linus Torvalds 已提交
289 290 291 292 293 294
	}
}

static void ip6_frag_expire(unsigned long data)
{
	struct frag_queue *fq = (struct frag_queue *) data;
295
	struct net_device *dev = NULL;
L
Linus Torvalds 已提交
296 297 298 299 300 301 302 303

	spin_lock(&fq->lock);

	if (fq->last_in & COMPLETE)
		goto out;

	fq_kill(fq);

304
	dev = dev_get_by_index(&init_net, fq->iif);
305 306 307 308 309 310 311
	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
312

313 314 315 316 317 318 319 320 321 322 323
	/* Don't send error if the first segment did not arrive. */
	if (!(fq->last_in&FIRST_IN) || !fq->fragments)
		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.
	 */
	fq->fragments->dev = dev;
	icmpv6_send(fq->fragments, ICMPV6_TIME_EXCEED, ICMPV6_EXC_FRAGTIME, 0, dev);
L
Linus Torvalds 已提交
324
out:
325 326
	if (dev)
		dev_put(dev);
L
Linus Torvalds 已提交
327 328 329 330 331 332 333
	spin_unlock(&fq->lock);
	fq_put(fq, NULL);
}

/* Creation primitives. */


334
static struct frag_queue *ip6_frag_intern(struct frag_queue *fq_in)
L
Linus Torvalds 已提交
335 336
{
	struct frag_queue *fq;
337
	unsigned int hash;
338 339 340
#ifdef CONFIG_SMP
	struct hlist_node *n;
#endif
L
Linus Torvalds 已提交
341 342

	write_lock(&ip6_frag_lock);
343
	hash = ip6qhashfn(fq_in->id, &fq_in->saddr, &fq_in->daddr);
L
Linus Torvalds 已提交
344
#ifdef CONFIG_SMP
345
	hlist_for_each_entry(fq, n, &ip6_frag_hash[hash], list) {
346
		if (fq->id == fq_in->id &&
L
Linus Torvalds 已提交
347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362
		    ipv6_addr_equal(&fq_in->saddr, &fq->saddr) &&
		    ipv6_addr_equal(&fq_in->daddr, &fq->daddr)) {
			atomic_inc(&fq->refcnt);
			write_unlock(&ip6_frag_lock);
			fq_in->last_in |= COMPLETE;
			fq_put(fq_in, NULL);
			return fq;
		}
	}
#endif
	fq = fq_in;

	if (!mod_timer(&fq->timer, jiffies + sysctl_ip6frag_time))
		atomic_inc(&fq->refcnt);

	atomic_inc(&fq->refcnt);
363
	hlist_add_head(&fq->list, &ip6_frag_hash[hash]);
L
Linus Torvalds 已提交
364 365 366 367 368 369 370 371 372
	INIT_LIST_HEAD(&fq->lru_list);
	list_add_tail(&fq->lru_list, &ip6_frag_lru_list);
	ip6_frag_nqueues++;
	write_unlock(&ip6_frag_lock);
	return fq;
}


static struct frag_queue *
A
Al Viro 已提交
373
ip6_frag_create(__be32 id, struct in6_addr *src, struct in6_addr *dst,
374
		struct inet6_dev *idev)
L
Linus Torvalds 已提交
375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390
{
	struct frag_queue *fq;

	if ((fq = frag_alloc_queue()) == NULL)
		goto oom;

	fq->id = id;
	ipv6_addr_copy(&fq->saddr, src);
	ipv6_addr_copy(&fq->daddr, dst);

	init_timer(&fq->timer);
	fq->timer.function = ip6_frag_expire;
	fq->timer.data = (long) fq;
	spin_lock_init(&fq->lock);
	atomic_set(&fq->refcnt, 1);

391
	return ip6_frag_intern(fq);
L
Linus Torvalds 已提交
392 393

oom:
394
	IP6_INC_STATS_BH(idev, IPSTATS_MIB_REASMFAILS);
L
Linus Torvalds 已提交
395 396 397 398
	return NULL;
}

static __inline__ struct frag_queue *
A
Al Viro 已提交
399
fq_find(__be32 id, struct in6_addr *src, struct in6_addr *dst,
400
	struct inet6_dev *idev)
L
Linus Torvalds 已提交
401 402
{
	struct frag_queue *fq;
403
	struct hlist_node *n;
404
	unsigned int hash;
L
Linus Torvalds 已提交
405 406

	read_lock(&ip6_frag_lock);
407
	hash = ip6qhashfn(id, src, dst);
408
	hlist_for_each_entry(fq, n, &ip6_frag_hash[hash], list) {
409
		if (fq->id == id &&
L
Linus Torvalds 已提交
410 411 412 413 414 415 416 417 418
		    ipv6_addr_equal(src, &fq->saddr) &&
		    ipv6_addr_equal(dst, &fq->daddr)) {
			atomic_inc(&fq->refcnt);
			read_unlock(&ip6_frag_lock);
			return fq;
		}
	}
	read_unlock(&ip6_frag_lock);

419
	return ip6_frag_create(id, src, dst, idev);
L
Linus Torvalds 已提交
420 421 422
}


423
static void ip6_frag_queue(struct frag_queue *fq, struct sk_buff *skb,
L
Linus Torvalds 已提交
424 425 426 427 428 429 430 431 432
			   struct frag_hdr *fhdr, int nhoff)
{
	struct sk_buff *prev, *next;
	int offset, end;

	if (fq->last_in & COMPLETE)
		goto err;

	offset = ntohs(fhdr->frag_off) & ~0x7;
433 434
	end = offset + (ntohs(ipv6_hdr(skb)->payload_len) -
			((u8 *)(fhdr + 1) - (u8 *)(ipv6_hdr(skb) + 1)));
L
Linus Torvalds 已提交
435 436

	if ((unsigned int)end > IPV6_MAXPLEN) {
437 438
		IP6_INC_STATS_BH(ip6_dst_idev(skb->dst),
				 IPSTATS_MIB_INHDRERRORS);
439 440 441
		icmpv6_param_prob(skb, ICMPV6_HDR_FIELD,
				  ((u8 *)&fhdr->frag_off -
				   skb_network_header(skb)));
442
		return;
L
Linus Torvalds 已提交
443 444
	}

445 446
	if (skb->ip_summed == CHECKSUM_COMPLETE) {
		const unsigned char *nh = skb_network_header(skb);
447
		skb->csum = csum_sub(skb->csum,
448 449 450
				     csum_partial(nh, (u8 *)(fhdr + 1) - nh,
						  0));
	}
L
Linus Torvalds 已提交
451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469

	/* 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.
		 */
		if (end < fq->len ||
		    ((fq->last_in & LAST_IN) && end != fq->len))
			goto err;
		fq->last_in |= LAST_IN;
		fq->len = end;
	} 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
			 */
470 471
			IP6_INC_STATS_BH(ip6_dst_idev(skb->dst),
					 IPSTATS_MIB_INHDRERRORS);
472
			icmpv6_param_prob(skb, ICMPV6_HDR_FIELD,
L
Linus Torvalds 已提交
473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489
					  offsetof(struct ipv6hdr, payload_len));
			return;
		}
		if (end > fq->len) {
			/* Some bits beyond end -> corruption. */
			if (fq->last_in & LAST_IN)
				goto err;
			fq->len = end;
		}
	}

	if (end == offset)
		goto err;

	/* Point into the IP datagram 'data' part. */
	if (!pskb_pull(skb, (u8 *) (fhdr + 1) - skb->data))
		goto err;
490

491 492
	if (pskb_trim_rcsum(skb, end - offset))
		goto err;
L
Linus Torvalds 已提交
493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569

	/* 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;
	for(next = fq->fragments; next != NULL; next = next->next) {
		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 */
			fq->meat -= i;
			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
				fq->fragments = next;

			fq->meat -= free_it->len;
			frag_kfree_skb(free_it, NULL);
		}
	}

	FRAG6_CB(skb)->offset = offset;

	/* Insert this fragment in the chain of fragments. */
	skb->next = next;
	if (prev)
		prev->next = skb;
	else
		fq->fragments = skb;

	if (skb->dev)
		fq->iif = skb->dev->ifindex;
	skb->dev = NULL;
570
	fq->stamp = skb->tstamp;
L
Linus Torvalds 已提交
571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586
	fq->meat += skb->len;
	atomic_add(skb->truesize, &ip6_frag_mem);

	/* The first fragment.
	 * nhoffset is obtained from the first fragment, of course.
	 */
	if (offset == 0) {
		fq->nhoffset = nhoff;
		fq->last_in |= FIRST_IN;
	}
	write_lock(&ip6_frag_lock);
	list_move_tail(&fq->lru_list, &ip6_frag_lru_list);
	write_unlock(&ip6_frag_lock);
	return;

err:
587
	IP6_INC_STATS(ip6_dst_idev(skb->dst), IPSTATS_MIB_REASMFAILS);
L
Linus Torvalds 已提交
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
	kfree_skb(skb);
}

/*
 *	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.
 */
static int ip6_frag_reasm(struct frag_queue *fq, struct sk_buff **skb_in,
			  struct net_device *dev)
{
	struct sk_buff *fp, *head = fq->fragments;
	int    payload_len;
	unsigned int nhoff;

	fq_kill(fq);

	BUG_TRAP(head != NULL);
	BUG_TRAP(FRAG6_CB(head)->offset == 0);

	/* Unfragmented part is taken from the first segment. */
613 614 615
	payload_len = ((head->data - skb_network_header(head)) -
		       sizeof(struct ipv6hdr) + fq->len -
		       sizeof(struct frag_hdr));
L
Linus Torvalds 已提交
616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648
	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;
		atomic_add(clone->truesize, &ip6_frag_mem);
	}

	/* We have to remove fragment header from datagram and to relocate
	 * header in order to calculate ICV correctly. */
	nhoff = fq->nhoffset;
649
	skb_network_header(head)[nhoff] = skb_transport_header(head)[0];
650
	memmove(head->head + sizeof(struct frag_hdr), head->head,
L
Linus Torvalds 已提交
651
		(head->data - head->head) - sizeof(struct frag_hdr));
652 653
	head->mac_header += sizeof(struct frag_hdr);
	head->network_header += sizeof(struct frag_hdr);
L
Linus Torvalds 已提交
654 655

	skb_shinfo(head)->frag_list = head->next;
656
	skb_reset_transport_header(head);
657
	skb_push(head, head->data - skb_network_header(head));
L
Linus Torvalds 已提交
658 659 660 661 662 663 664
	atomic_sub(head->truesize, &ip6_frag_mem);

	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;
665
		else if (head->ip_summed == CHECKSUM_COMPLETE)
L
Linus Torvalds 已提交
666 667 668 669 670 671 672
			head->csum = csum_add(head->csum, fp->csum);
		head->truesize += fp->truesize;
		atomic_sub(fp->truesize, &ip6_frag_mem);
	}

	head->next = NULL;
	head->dev = dev;
673
	head->tstamp = fq->stamp;
674
	ipv6_hdr(head)->payload_len = htons(payload_len);
675
	IP6CB(head)->nhoff = nhoff;
L
Linus Torvalds 已提交
676 677 678 679

	*skb_in = head;

	/* Yes, and fold redundant checksum back. 8) */
680
	if (head->ip_summed == CHECKSUM_COMPLETE)
681
		head->csum = csum_partial(skb_network_header(head),
682
					  skb_network_header_len(head),
683
					  head->csum);
L
Linus Torvalds 已提交
684

685 686 687
	rcu_read_lock();
	IP6_INC_STATS_BH(__in6_dev_get(dev), IPSTATS_MIB_REASMOKS);
	rcu_read_unlock();
L
Linus Torvalds 已提交
688 689 690 691 692 693 694 695 696 697 698
	fq->fragments = NULL;
	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:
699 700 701
	rcu_read_lock();
	IP6_INC_STATS_BH(__in6_dev_get(dev), IPSTATS_MIB_REASMFAILS);
	rcu_read_unlock();
L
Linus Torvalds 已提交
702 703 704
	return -1;
}

705
static int ipv6_frag_rcv(struct sk_buff **skbp)
L
Linus Torvalds 已提交
706
{
707
	struct sk_buff *skb = *skbp;
L
Linus Torvalds 已提交
708 709 710
	struct net_device *dev = skb->dev;
	struct frag_hdr *fhdr;
	struct frag_queue *fq;
711
	struct ipv6hdr *hdr = ipv6_hdr(skb);
L
Linus Torvalds 已提交
712

713
	IP6_INC_STATS_BH(ip6_dst_idev(skb->dst), IPSTATS_MIB_REASMREQDS);
L
Linus Torvalds 已提交
714 715 716

	/* Jumbo payload inhibits frag. header */
	if (hdr->payload_len==0) {
717
		IP6_INC_STATS(ip6_dst_idev(skb->dst), IPSTATS_MIB_INHDRERRORS);
718 719
		icmpv6_param_prob(skb, ICMPV6_HDR_FIELD,
				  skb_network_header_len(skb));
L
Linus Torvalds 已提交
720 721
		return -1;
	}
722 723
	if (!pskb_may_pull(skb, (skb_transport_offset(skb) +
				 sizeof(struct frag_hdr)))) {
724
		IP6_INC_STATS(ip6_dst_idev(skb->dst), IPSTATS_MIB_INHDRERRORS);
725 726
		icmpv6_param_prob(skb, ICMPV6_HDR_FIELD,
				  skb_network_header_len(skb));
L
Linus Torvalds 已提交
727 728 729
		return -1;
	}

730
	hdr = ipv6_hdr(skb);
731
	fhdr = (struct frag_hdr *)skb_transport_header(skb);
L
Linus Torvalds 已提交
732 733 734

	if (!(fhdr->frag_off & htons(0xFFF9))) {
		/* It is not a fragmented frame */
735
		skb->transport_header += sizeof(struct frag_hdr);
736
		IP6_INC_STATS_BH(ip6_dst_idev(skb->dst), IPSTATS_MIB_REASMOKS);
L
Linus Torvalds 已提交
737

738
		IP6CB(skb)->nhoff = (u8 *)fhdr - skb_network_header(skb);
L
Linus Torvalds 已提交
739 740 741 742
		return 1;
	}

	if (atomic_read(&ip6_frag_mem) > sysctl_ip6frag_high_thresh)
743
		ip6_evictor(ip6_dst_idev(skb->dst));
L
Linus Torvalds 已提交
744

745 746
	if ((fq = fq_find(fhdr->identification, &hdr->saddr, &hdr->daddr,
			  ip6_dst_idev(skb->dst))) != NULL) {
L
Linus Torvalds 已提交
747 748 749 750
		int ret = -1;

		spin_lock(&fq->lock);

751
		ip6_frag_queue(fq, skb, fhdr, IP6CB(skb)->nhoff);
L
Linus Torvalds 已提交
752 753 754

		if (fq->last_in == (FIRST_IN|LAST_IN) &&
		    fq->meat == fq->len)
755
			ret = ip6_frag_reasm(fq, skbp, dev);
L
Linus Torvalds 已提交
756 757 758 759 760 761

		spin_unlock(&fq->lock);
		fq_put(fq, NULL);
		return ret;
	}

762
	IP6_INC_STATS_BH(ip6_dst_idev(skb->dst), IPSTATS_MIB_REASMFAILS);
L
Linus Torvalds 已提交
763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785
	kfree_skb(skb);
	return -1;
}

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

void __init ipv6_frag_init(void)
{
	if (inet6_add_protocol(&frag_protocol, IPPROTO_FRAGMENT) < 0)
		printk(KERN_ERR "ipv6_frag_init: Could not register protocol\n");

	ip6_frag_hash_rnd = (u32) ((num_physpages ^ (num_physpages>>7)) ^
				   (jiffies ^ (jiffies >> 6)));

	init_timer(&ip6_frag_secret_timer);
	ip6_frag_secret_timer.function = ip6_frag_secret_rebuild;
	ip6_frag_secret_timer.expires = jiffies + sysctl_ip6frag_secret_interval;
	add_timer(&ip6_frag_secret_timer);
}