[deliverable/linux.git] / net / ipv4 / inet_fragment.c

/*
 * inet fragments management
 *
 *		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.
 *
 * 		Authors:	Pavel Emelyanov <xemul@openvz.org>
 *				Started as consolidation of ipv4/ip_fragment.c,
 *				ipv6/reassembly. and ipv6 nf conntrack reassembly
 */

#include <linux/list.h>
#include <linux/spinlock.h>
#include <linux/module.h>
#include <linux/timer.h>
#include <linux/mm.h>
#include <linux/random.h>
#include <linux/skbuff.h>
#include <linux/rtnetlink.h>
#include <linux/slab.h>

#include <net/sock.h>
#include <net/inet_frag.h>
#include <net/inet_ecn.h>

#define INETFRAGS_EVICT_BUCKETS   128
#define INETFRAGS_EVICT_MAX	  512

/* don't rebuild inetfrag table with new secret more often than this */
#define INETFRAGS_MIN_REBUILD_INTERVAL (5 * HZ)

/* Given the OR values of all fragments, apply RFC 3168 5.3 requirements
 * Value : 0xff if frame should be dropped.
 *         0 or INET_ECN_CE value, to be ORed in to final iph->tos field
 */
const u8 ip_frag_ecn_table[16] = {
	/* at least one fragment had CE, and others ECT_0 or ECT_1 */
	[IPFRAG_ECN_CE | IPFRAG_ECN_ECT_0]			= INET_ECN_CE,
	[IPFRAG_ECN_CE | IPFRAG_ECN_ECT_1]			= INET_ECN_CE,
	[IPFRAG_ECN_CE | IPFRAG_ECN_ECT_0 | IPFRAG_ECN_ECT_1]	= INET_ECN_CE,

	/* invalid combinations : drop frame */
	[IPFRAG_ECN_NOT_ECT | IPFRAG_ECN_CE] = 0xff,
	[IPFRAG_ECN_NOT_ECT | IPFRAG_ECN_ECT_0] = 0xff,
	[IPFRAG_ECN_NOT_ECT | IPFRAG_ECN_ECT_1] = 0xff,
	[IPFRAG_ECN_NOT_ECT | IPFRAG_ECN_ECT_0 | IPFRAG_ECN_ECT_1] = 0xff,
	[IPFRAG_ECN_NOT_ECT | IPFRAG_ECN_CE | IPFRAG_ECN_ECT_0] = 0xff,
	[IPFRAG_ECN_NOT_ECT | IPFRAG_ECN_CE | IPFRAG_ECN_ECT_1] = 0xff,
	[IPFRAG_ECN_NOT_ECT | IPFRAG_ECN_CE | IPFRAG_ECN_ECT_0 | IPFRAG_ECN_ECT_1] = 0xff,
};
EXPORT_SYMBOL(ip_frag_ecn_table);

static unsigned int
inet_frag_hashfn(const struct inet_frags *f, const struct inet_frag_queue *q)
{
	return f->hashfn(q) & (INETFRAGS_HASHSZ - 1);
}

static bool inet_frag_may_rebuild(struct inet_frags *f)
{
	return time_after(jiffies,
	       f->last_rebuild_jiffies + INETFRAGS_MIN_REBUILD_INTERVAL);
}

static void inet_frag_secret_rebuild(struct inet_frags *f)
{
	int i;

	write_seqlock_bh(&f->rnd_seqlock);

	if (!inet_frag_may_rebuild(f))
		goto out;

	get_random_bytes(&f->rnd, sizeof(u32));

	for (i = 0; i < INETFRAGS_HASHSZ; i++) {
		struct inet_frag_bucket *hb;
		struct inet_frag_queue *q;
		struct hlist_node *n;

		hb = &f->hash[i];
		spin_lock(&hb->chain_lock);

		hlist_for_each_entry_safe(q, n, &hb->chain, list) {
			unsigned int hval = inet_frag_hashfn(f, q);

			if (hval != i) {
				struct inet_frag_bucket *hb_dest;

				hlist_del(&q->list);

				/* Relink to new hash chain. */
				hb_dest = &f->hash[hval];

				/* This is the only place where we take
				 * another chain_lock while already holding
				 * one.  As this will not run concurrently,
				 * we cannot deadlock on hb_dest lock below, if its
				 * already locked it will be released soon since
				 * other caller cannot be waiting for hb lock
				 * that we've taken above.
				 */
				spin_lock_nested(&hb_dest->chain_lock,
						 SINGLE_DEPTH_NESTING);
				hlist_add_head(&q->list, &hb_dest->chain);
				spin_unlock(&hb_dest->chain_lock);
			}
		}
		spin_unlock(&hb->chain_lock);
	}

	f->rebuild = false;
	f->last_rebuild_jiffies = jiffies;
out:
	write_sequnlock_bh(&f->rnd_seqlock);
}

static bool inet_fragq_should_evict(const struct inet_frag_queue *q)
{
	return q->net->low_thresh == 0 ||
	       frag_mem_limit(q->net) >= q->net->low_thresh;
}

static unsigned int
inet_evict_bucket(struct inet_frags *f, struct inet_frag_bucket *hb)
{
	struct inet_frag_queue *fq;
	struct hlist_node *n;
	unsigned int evicted = 0;
	HLIST_HEAD(expired);

evict_again:
	spin_lock(&hb->chain_lock);

	hlist_for_each_entry_safe(fq, n, &hb->chain, list) {
		if (!inet_fragq_should_evict(fq))
			continue;

		if (!del_timer(&fq->timer)) {
			/* q expiring right now thus increment its refcount so
			 * it won't be freed under us and wait until the timer
			 * has finished executing then destroy it
			 */
			atomic_inc(&fq->refcnt);
			spin_unlock(&hb->chain_lock);
			del_timer_sync(&fq->timer);
			WARN_ON(atomic_read(&fq->refcnt) != 1);
			inet_frag_put(fq, f);
			goto evict_again;
		}

		fq->flags |= INET_FRAG_EVICTED;
		hlist_del(&fq->list);
		hlist_add_head(&fq->list, &expired);
		++evicted;
	}

	spin_unlock(&hb->chain_lock);

	hlist_for_each_entry_safe(fq, n, &expired, list)
		f->frag_expire((unsigned long) fq);

	return evicted;
}

static void inet_frag_worker(struct work_struct *work)
{
	unsigned int budget = INETFRAGS_EVICT_BUCKETS;
	unsigned int i, evicted = 0;
	struct inet_frags *f;

	f = container_of(work, struct inet_frags, frags_work);

	BUILD_BUG_ON(INETFRAGS_EVICT_BUCKETS >= INETFRAGS_HASHSZ);

	local_bh_disable();

	for (i = ACCESS_ONCE(f->next_bucket); budget; --budget) {
		evicted += inet_evict_bucket(f, &f->hash[i]);
		i = (i + 1) & (INETFRAGS_HASHSZ - 1);
		if (evicted > INETFRAGS_EVICT_MAX)
			break;
	}

	f->next_bucket = i;

	local_bh_enable();

	if (f->rebuild && inet_frag_may_rebuild(f))
		inet_frag_secret_rebuild(f);
}

static void inet_frag_schedule_worker(struct inet_frags *f)
{
	if (unlikely(!work_pending(&f->frags_work)))
		schedule_work(&f->frags_work);
}

int inet_frags_init(struct inet_frags *f)
{
	int i;

	INIT_WORK(&f->frags_work, inet_frag_worker);

	for (i = 0; i < INETFRAGS_HASHSZ; i++) {
		struct inet_frag_bucket *hb = &f->hash[i];

		spin_lock_init(&hb->chain_lock);
		INIT_HLIST_HEAD(&hb->chain);
	}

	seqlock_init(&f->rnd_seqlock);
	f->last_rebuild_jiffies = 0;
	f->frags_cachep = kmem_cache_create(f->frags_cache_name, f->qsize, 0, 0,
					    NULL);
	if (!f->frags_cachep)
		return -ENOMEM;

	return 0;
}
EXPORT_SYMBOL(inet_frags_init);

void inet_frags_init_net(struct netns_frags *nf)
{
	init_frag_mem_limit(nf);
}
EXPORT_SYMBOL(inet_frags_init_net);

void inet_frags_fini(struct inet_frags *f)
{
	cancel_work_sync(&f->frags_work);
	kmem_cache_destroy(f->frags_cachep);
}
EXPORT_SYMBOL(inet_frags_fini);

void inet_frags_exit_net(struct netns_frags *nf, struct inet_frags *f)
{
	unsigned int seq;
	int i;

	nf->low_thresh = 0;
	local_bh_disable();

evict_again:
	seq = read_seqbegin(&f->rnd_seqlock);

	for (i = 0; i < INETFRAGS_HASHSZ ; i++)
		inet_evict_bucket(f, &f->hash[i]);

	if (read_seqretry(&f->rnd_seqlock, seq))
		goto evict_again;

	local_bh_enable();

	percpu_counter_destroy(&nf->mem);
}
EXPORT_SYMBOL(inet_frags_exit_net);

static struct inet_frag_bucket *
get_frag_bucket_locked(struct inet_frag_queue *fq, struct inet_frags *f)
__acquires(hb->chain_lock)
{
	struct inet_frag_bucket *hb;
	unsigned int seq, hash;

 restart:
	seq = read_seqbegin(&f->rnd_seqlock);

	hash = inet_frag_hashfn(f, fq);
	hb = &f->hash[hash];

	spin_lock(&hb->chain_lock);
	if (read_seqretry(&f->rnd_seqlock, seq)) {
		spin_unlock(&hb->chain_lock);
		goto restart;
	}

	return hb;
}

static inline void fq_unlink(struct inet_frag_queue *fq, struct inet_frags *f)
{
	struct inet_frag_bucket *hb;

	hb = get_frag_bucket_locked(fq, f);
	hlist_del(&fq->list);
	spin_unlock(&hb->chain_lock);
}

void inet_frag_kill(struct inet_frag_queue *fq, struct inet_frags *f)
{
	if (del_timer(&fq->timer))
		atomic_dec(&fq->refcnt);

	if (!(fq->flags & INET_FRAG_COMPLETE)) {
		fq_unlink(fq, f);
		atomic_dec(&fq->refcnt);
		fq->flags |= INET_FRAG_COMPLETE;
	}
}
EXPORT_SYMBOL(inet_frag_kill);

static inline void frag_kfree_skb(struct netns_frags *nf, struct inet_frags *f,
				  struct sk_buff *skb)
{
	if (f->skb_free)
		f->skb_free(skb);
	kfree_skb(skb);
}

void inet_frag_destroy(struct inet_frag_queue *q, struct inet_frags *f)
{
	struct sk_buff *fp;
	struct netns_frags *nf;
	unsigned int sum, sum_truesize = 0;

	WARN_ON(!(q->flags & INET_FRAG_COMPLETE));
	WARN_ON(del_timer(&q->timer) != 0);

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

		sum_truesize += fp->truesize;
		frag_kfree_skb(nf, f, fp);
		fp = xp;
	}
	sum = sum_truesize + f->qsize;
	sub_frag_mem_limit(q, sum);

	if (f->destructor)
		f->destructor(q);
	kmem_cache_free(f->frags_cachep, q);
}
EXPORT_SYMBOL(inet_frag_destroy);

static struct inet_frag_queue *inet_frag_intern(struct netns_frags *nf,
						struct inet_frag_queue *qp_in,
						struct inet_frags *f,
						void *arg)
{
	struct inet_frag_bucket *hb = get_frag_bucket_locked(qp_in, f);
	struct inet_frag_queue *qp;

#ifdef CONFIG_SMP
	/* With SMP race we have to recheck hash table, because
	 * such entry could have been created on other cpu before
	 * we acquired hash bucket lock.
	 */
	hlist_for_each_entry(qp, &hb->chain, list) {
		if (qp->net == nf && f->match(qp, arg)) {
			atomic_inc(&qp->refcnt);
			spin_unlock(&hb->chain_lock);
			qp_in->flags |= INET_FRAG_COMPLETE;
			inet_frag_put(qp_in, f);
			return qp;
		}
	}
#endif
	qp = qp_in;
	if (!mod_timer(&qp->timer, jiffies + nf->timeout))
		atomic_inc(&qp->refcnt);

	atomic_inc(&qp->refcnt);
	hlist_add_head(&qp->list, &hb->chain);

	spin_unlock(&hb->chain_lock);

	return qp;
}

static struct inet_frag_queue *inet_frag_alloc(struct netns_frags *nf,
					       struct inet_frags *f,
					       void *arg)
{
	struct inet_frag_queue *q;

	if (frag_mem_limit(nf) > nf->high_thresh) {
		inet_frag_schedule_worker(f);
		return NULL;
	}

	q = kmem_cache_zalloc(f->frags_cachep, GFP_ATOMIC);
	if (q == NULL)
		return NULL;

	q->net = nf;
	f->constructor(q, arg);
	add_frag_mem_limit(q, f->qsize);

	setup_timer(&q->timer, f->frag_expire, (unsigned long)q);
	spin_lock_init(&q->lock);
	atomic_set(&q->refcnt, 1);

	return q;
}

static struct inet_frag_queue *inet_frag_create(struct netns_frags *nf,
						struct inet_frags *f,
						void *arg)
{
	struct inet_frag_queue *q;

	q = inet_frag_alloc(nf, f, arg);
	if (q == NULL)
		return NULL;

	return inet_frag_intern(nf, q, f, arg);
}

struct inet_frag_queue *inet_frag_find(struct netns_frags *nf,
				       struct inet_frags *f, void *key,
				       unsigned int hash)
{
	struct inet_frag_bucket *hb;
	struct inet_frag_queue *q;
	int depth = 0;

	if (frag_mem_limit(nf) > nf->low_thresh)
		inet_frag_schedule_worker(f);

	hash &= (INETFRAGS_HASHSZ - 1);
	hb = &f->hash[hash];

	spin_lock(&hb->chain_lock);
	hlist_for_each_entry(q, &hb->chain, list) {
		if (q->net == nf && f->match(q, key)) {
			atomic_inc(&q->refcnt);
			spin_unlock(&hb->chain_lock);
			return q;
		}
		depth++;
	}
	spin_unlock(&hb->chain_lock);

	if (depth <= INETFRAGS_MAXDEPTH)
		return inet_frag_create(nf, f, key);

	if (inet_frag_may_rebuild(f)) {
		if (!f->rebuild)
			f->rebuild = true;
		inet_frag_schedule_worker(f);
	}

	return ERR_PTR(-ENOBUFS);
}
EXPORT_SYMBOL(inet_frag_find);

void inet_frag_maybe_warn_overflow(struct inet_frag_queue *q,
				   const char *prefix)
{
	static const char msg[] = "inet_frag_find: Fragment hash bucket"
		" list length grew over limit " __stringify(INETFRAGS_MAXDEPTH)
		". Dropping fragment.\n";

	if (PTR_ERR(q) == -ENOBUFS)
		LIMIT_NETDEBUG(KERN_WARNING "%s%s", prefix, msg);
}
EXPORT_SYMBOL(inet_frag_maybe_warn_overflow);
Commit	Line	Data
7eb95156 PE	1	/*
	2	* inet fragments management
	3	*
	4	* This program is free software; you can redistribute it and/or
	5	* modify it under the terms of the GNU General Public License
	6	* as published by the Free Software Foundation; either version
	7	* 2 of the License, or (at your option) any later version.
	8	*
	9	* Authors: Pavel Emelyanov <xemul@openvz.org>
	10	* Started as consolidation of ipv4/ip_fragment.c,
	11	* ipv6/reassembly. and ipv6 nf conntrack reassembly
	12	*/
	13
	14	#include <linux/list.h>
	15	#include <linux/spinlock.h>
	16	#include <linux/module.h>
	17	#include <linux/timer.h>
	18	#include <linux/mm.h>
321a3a99	19	#include <linux/random.h>
1e4b8287 PE	20	#include <linux/skbuff.h>
1e4b8287 PE	21	#include <linux/rtnetlink.h>
5a0e3ad6	22	#include <linux/slab.h>
7eb95156	23
5a3da1fe	24	#include <net/sock.h>
7eb95156	25	#include <net/inet_frag.h>
be991971 HFS	26	#include <net/inet_ecn.h>
be991971 HFS	27
b13d3cbf FW	28	#define INETFRAGS_EVICT_BUCKETS 128
	29	#define INETFRAGS_EVICT_MAX 512
	30
e3a57d18 FW	31	/* don't rebuild inetfrag table with new secret more often than this */
	32	#define INETFRAGS_MIN_REBUILD_INTERVAL (5 * HZ)
	33
be991971 HFS	34	/* Given the OR values of all fragments, apply RFC 3168 5.3 requirements
	35	* Value : 0xff if frame should be dropped.
	36	* 0 or INET_ECN_CE value, to be ORed in to final iph->tos field
	37	*/
	38	const u8 ip_frag_ecn_table[16] = {
	39	/* at least one fragment had CE, and others ECT_0 or ECT_1 */
	40	[IPFRAG_ECN_CE \| IPFRAG_ECN_ECT_0] = INET_ECN_CE,
	41	[IPFRAG_ECN_CE \| IPFRAG_ECN_ECT_1] = INET_ECN_CE,
	42	[IPFRAG_ECN_CE \| IPFRAG_ECN_ECT_0 \| IPFRAG_ECN_ECT_1] = INET_ECN_CE,
	43
	44	/* invalid combinations : drop frame */
	45	[IPFRAG_ECN_NOT_ECT \| IPFRAG_ECN_CE] = 0xff,
	46	[IPFRAG_ECN_NOT_ECT \| IPFRAG_ECN_ECT_0] = 0xff,
	47	[IPFRAG_ECN_NOT_ECT \| IPFRAG_ECN_ECT_1] = 0xff,
	48	[IPFRAG_ECN_NOT_ECT \| IPFRAG_ECN_ECT_0 \| IPFRAG_ECN_ECT_1] = 0xff,
	49	[IPFRAG_ECN_NOT_ECT \| IPFRAG_ECN_CE \| IPFRAG_ECN_ECT_0] = 0xff,
	50	[IPFRAG_ECN_NOT_ECT \| IPFRAG_ECN_CE \| IPFRAG_ECN_ECT_1] = 0xff,
	51	[IPFRAG_ECN_NOT_ECT \| IPFRAG_ECN_CE \| IPFRAG_ECN_ECT_0 \| IPFRAG_ECN_ECT_1] = 0xff,
	52	};
	53	EXPORT_SYMBOL(ip_frag_ecn_table);
7eb95156	54
fb3cfe6e FW	55	static unsigned int
	56	inet_frag_hashfn(const struct inet_frags f, const struct inet_frag_queue q)
	57	{
	58	return f->hashfn(q) & (INETFRAGS_HASHSZ - 1);
	59	}
	60
e3a57d18 FW	61	static bool inet_frag_may_rebuild(struct inet_frags *f)
	62	{
	63	return time_after(jiffies,
	64	f->last_rebuild_jiffies + INETFRAGS_MIN_REBUILD_INTERVAL);
	65	}
	66
	67	static void inet_frag_secret_rebuild(struct inet_frags *f)
321a3a99	68	{
321a3a99 PE	69	int i;
321a3a99 PE	70
ab1c724f	71	write_seqlock_bh(&f->rnd_seqlock);
e3a57d18 FW	72
	73	if (!inet_frag_may_rebuild(f))
	74	goto out;
19952cc4	75
321a3a99	76	get_random_bytes(&f->rnd, sizeof(u32));
e3a57d18	77
321a3a99	78	for (i = 0; i < INETFRAGS_HASHSZ; i++) {
19952cc4	79	struct inet_frag_bucket *hb;
321a3a99	80	struct inet_frag_queue *q;
b67bfe0d	81	struct hlist_node *n;
321a3a99	82
19952cc4	83	hb = &f->hash[i];
ab1c724f FW	84	spin_lock(&hb->chain_lock);
ab1c724f FW	85
19952cc4	86	hlist_for_each_entry_safe(q, n, &hb->chain, list) {
fb3cfe6e	87	unsigned int hval = inet_frag_hashfn(f, q);
321a3a99 PE	88
321a3a99 PE	89	if (hval != i) {
19952cc4 JDB	90	struct inet_frag_bucket *hb_dest;
19952cc4 JDB	91
321a3a99 PE	92	hlist_del(&q->list);
	93
	94	/* Relink to new hash chain. */
19952cc4	95	hb_dest = &f->hash[hval];
ab1c724f FW	96
	97	/* This is the only place where we take
	98	* another chain_lock while already holding
	99	* one. As this will not run concurrently,
	100	* we cannot deadlock on hb_dest lock below, if its
	101	* already locked it will be released soon since
	102	* other caller cannot be waiting for hb lock
	103	* that we've taken above.
	104	*/
	105	spin_lock_nested(&hb_dest->chain_lock,
	106	SINGLE_DEPTH_NESTING);
19952cc4	107	hlist_add_head(&q->list, &hb_dest->chain);
ab1c724f	108	spin_unlock(&hb_dest->chain_lock);
321a3a99 PE	109	}
321a3a99 PE	110	}
ab1c724f	111	spin_unlock(&hb->chain_lock);
321a3a99	112	}
321a3a99	113
e3a57d18 FW	114	f->rebuild = false;
	115	f->last_rebuild_jiffies = jiffies;
	116	out:
ab1c724f	117	write_sequnlock_bh(&f->rnd_seqlock);
321a3a99 PE	118	}
321a3a99 PE	119
b13d3cbf FW	120	static bool inet_fragq_should_evict(const struct inet_frag_queue *q)
	121	{
	122	return q->net->low_thresh == 0 \|\|
	123	frag_mem_limit(q->net) >= q->net->low_thresh;
	124	}
	125
	126	static unsigned int
	127	inet_evict_bucket(struct inet_frags f, struct inet_frag_bucket hb)
	128	{
	129	struct inet_frag_queue *fq;
	130	struct hlist_node *n;
	131	unsigned int evicted = 0;
	132	HLIST_HEAD(expired);
	133
	134	evict_again:
	135	spin_lock(&hb->chain_lock);
	136
	137	hlist_for_each_entry_safe(fq, n, &hb->chain, list) {
	138	if (!inet_fragq_should_evict(fq))
	139	continue;
	140
	141	if (!del_timer(&fq->timer)) {
	142	/* q expiring right now thus increment its refcount so
	143	* it won't be freed under us and wait until the timer
	144	* has finished executing then destroy it
	145	*/
	146	atomic_inc(&fq->refcnt);
	147	spin_unlock(&hb->chain_lock);
	148	del_timer_sync(&fq->timer);
	149	WARN_ON(atomic_read(&fq->refcnt) != 1);
	150	inet_frag_put(fq, f);
	151	goto evict_again;
	152	}
	153
06aa8b8a	154	fq->flags \|= INET_FRAG_EVICTED;
b13d3cbf FW	155	hlist_del(&fq->list);
	156	hlist_add_head(&fq->list, &expired);
	157	++evicted;
	158	}
	159
	160	spin_unlock(&hb->chain_lock);
	161
	162	hlist_for_each_entry_safe(fq, n, &expired, list)
	163	f->frag_expire((unsigned long) fq);
	164
	165	return evicted;
	166	}
	167
	168	static void inet_frag_worker(struct work_struct *work)
	169	{
	170	unsigned int budget = INETFRAGS_EVICT_BUCKETS;
	171	unsigned int i, evicted = 0;
	172	struct inet_frags *f;
	173
	174	f = container_of(work, struct inet_frags, frags_work);
	175
	176	BUILD_BUG_ON(INETFRAGS_EVICT_BUCKETS >= INETFRAGS_HASHSZ);
	177
ab1c724f	178	local_bh_disable();
b13d3cbf FW	179
	180	for (i = ACCESS_ONCE(f->next_bucket); budget; --budget) {
	181	evicted += inet_evict_bucket(f, &f->hash[i]);
	182	i = (i + 1) & (INETFRAGS_HASHSZ - 1);
	183	if (evicted > INETFRAGS_EVICT_MAX)
	184	break;
	185	}
	186
	187	f->next_bucket = i;
	188
ab1c724f FW	189	local_bh_enable();
ab1c724f FW	190
e3a57d18 FW	191	if (f->rebuild && inet_frag_may_rebuild(f))
e3a57d18 FW	192	inet_frag_secret_rebuild(f);
b13d3cbf FW	193	}
	194
	195	static void inet_frag_schedule_worker(struct inet_frags *f)
	196	{
	197	if (unlikely(!work_pending(&f->frags_work)))
	198	schedule_work(&f->frags_work);
	199	}
	200
d4ad4d22	201	int inet_frags_init(struct inet_frags *f)
7eb95156 PE	202	{
	203	int i;
	204
b13d3cbf FW	205	INIT_WORK(&f->frags_work, inet_frag_worker);
b13d3cbf FW	206
19952cc4 JDB	207	for (i = 0; i < INETFRAGS_HASHSZ; i++) {
19952cc4 JDB	208	struct inet_frag_bucket *hb = &f->hash[i];
7eb95156	209
19952cc4 JDB	210	spin_lock_init(&hb->chain_lock);
	211	INIT_HLIST_HEAD(&hb->chain);
	212	}
ab1c724f FW	213
ab1c724f FW	214	seqlock_init(&f->rnd_seqlock);
e3a57d18	215	f->last_rebuild_jiffies = 0;
d4ad4d22 NA	216	f->frags_cachep = kmem_cache_create(f->frags_cache_name, f->qsize, 0, 0,
	217	NULL);
	218	if (!f->frags_cachep)
	219	return -ENOMEM;
	220
	221	return 0;
7eb95156 PE	222	}
	223	EXPORT_SYMBOL(inet_frags_init);
	224
e5a2bb84 PE	225	void inet_frags_init_net(struct netns_frags *nf)
e5a2bb84 PE	226	{
d433673e	227	init_frag_mem_limit(nf);
e5a2bb84 PE	228	}
	229	EXPORT_SYMBOL(inet_frags_init_net);
	230
7eb95156 PE	231	void inet_frags_fini(struct inet_frags *f)
7eb95156 PE	232	{
b13d3cbf	233	cancel_work_sync(&f->frags_work);
d4ad4d22	234	kmem_cache_destroy(f->frags_cachep);
7eb95156 PE	235	}
7eb95156 PE	236	EXPORT_SYMBOL(inet_frags_fini);
277e650d	237
81566e83 PE	238	void inet_frags_exit_net(struct netns_frags nf, struct inet_frags f)
81566e83 PE	239	{
ab1c724f	240	unsigned int seq;
b13d3cbf FW	241	int i;
b13d3cbf FW	242
81566e83	243	nf->low_thresh = 0;
ab1c724f	244	local_bh_disable();
e8e16b70	245
ab1c724f FW	246	evict_again:
ab1c724f FW	247	seq = read_seqbegin(&f->rnd_seqlock);
b13d3cbf FW	248
	249	for (i = 0; i < INETFRAGS_HASHSZ ; i++)
	250	inet_evict_bucket(f, &f->hash[i]);
	251
ab1c724f FW	252	if (read_seqretry(&f->rnd_seqlock, seq))
	253	goto evict_again;
	254
	255	local_bh_enable();
6d7b857d JDB	256
6d7b857d JDB	257	percpu_counter_destroy(&nf->mem);
81566e83 PE	258	}
	259	EXPORT_SYMBOL(inet_frags_exit_net);
	260
ab1c724f FW	261	static struct inet_frag_bucket *
	262	get_frag_bucket_locked(struct inet_frag_queue fq, struct inet_frags f)
	263	__acquires(hb->chain_lock)
277e650d	264	{
19952cc4	265	struct inet_frag_bucket *hb;
ab1c724f FW	266	unsigned int seq, hash;
	267
	268	restart:
	269	seq = read_seqbegin(&f->rnd_seqlock);
19952cc4	270
fb3cfe6e	271	hash = inet_frag_hashfn(f, fq);
19952cc4 JDB	272	hb = &f->hash[hash];
	273
	274	spin_lock(&hb->chain_lock);
ab1c724f FW	275	if (read_seqretry(&f->rnd_seqlock, seq)) {
	276	spin_unlock(&hb->chain_lock);
	277	goto restart;
	278	}
	279
	280	return hb;
	281	}
	282
	283	static inline void fq_unlink(struct inet_frag_queue fq, struct inet_frags f)
	284	{
	285	struct inet_frag_bucket *hb;
	286
	287	hb = get_frag_bucket_locked(fq, f);
277e650d	288	hlist_del(&fq->list);
19952cc4	289	spin_unlock(&hb->chain_lock);
277e650d PE	290	}
	291
	292	void inet_frag_kill(struct inet_frag_queue fq, struct inet_frags f)
	293	{
	294	if (del_timer(&fq->timer))
	295	atomic_dec(&fq->refcnt);
	296
06aa8b8a	297	if (!(fq->flags & INET_FRAG_COMPLETE)) {
277e650d PE	298	fq_unlink(fq, f);
277e650d PE	299	atomic_dec(&fq->refcnt);
06aa8b8a	300	fq->flags \|= INET_FRAG_COMPLETE;
277e650d PE	301	}
277e650d PE	302	}
277e650d	303	EXPORT_SYMBOL(inet_frag_kill);
1e4b8287	304
6ddc0822	305	static inline void frag_kfree_skb(struct netns_frags nf, struct inet_frags f,
06aa8b8a	306	struct sk_buff *skb)
1e4b8287	307	{
1e4b8287 PE	308	if (f->skb_free)
	309	f->skb_free(skb);
	310	kfree_skb(skb);
	311	}
	312
3fd588eb	313	void inet_frag_destroy(struct inet_frag_queue q, struct inet_frags f)
1e4b8287 PE	314	{
1e4b8287 PE	315	struct sk_buff *fp;
6ddc0822	316	struct netns_frags *nf;
d433673e	317	unsigned int sum, sum_truesize = 0;
1e4b8287	318
06aa8b8a	319	WARN_ON(!(q->flags & INET_FRAG_COMPLETE));
547b792c	320	WARN_ON(del_timer(&q->timer) != 0);
1e4b8287 PE	321
	322	/* Release all fragment data. */
	323	fp = q->fragments;
6ddc0822	324	nf = q->net;
1e4b8287 PE	325	while (fp) {
	326	struct sk_buff *xp = fp->next;
	327
d433673e JDB	328	sum_truesize += fp->truesize;
d433673e JDB	329	frag_kfree_skb(nf, f, fp);
1e4b8287 PE	330	fp = xp;
1e4b8287 PE	331	}
d433673e	332	sum = sum_truesize + f->qsize;
d433673e	333	sub_frag_mem_limit(q, sum);
1e4b8287	334
c9547709 PE	335	if (f->destructor)
c9547709 PE	336	f->destructor(q);
d4ad4d22	337	kmem_cache_free(f->frags_cachep, q);
1e4b8287 PE	338	}
1e4b8287 PE	339	EXPORT_SYMBOL(inet_frag_destroy);
8e7999c4	340
ac18e750	341	static struct inet_frag_queue inet_frag_intern(struct netns_frags nf,
f926e236 NA	342	struct inet_frag_queue *qp_in,
	343	struct inet_frags *f,
	344	void *arg)
2588fe1d	345	{
ab1c724f	346	struct inet_frag_bucket *hb = get_frag_bucket_locked(qp_in, f);
2588fe1d	347	struct inet_frag_queue *qp;
19952cc4	348
2588fe1d PE	349	#ifdef CONFIG_SMP
2588fe1d PE	350	/* With SMP race we have to recheck hash table, because
ab1c724f FW	351	* such entry could have been created on other cpu before
ab1c724f FW	352	* we acquired hash bucket lock.
2588fe1d	353	*/
19952cc4	354	hlist_for_each_entry(qp, &hb->chain, list) {
ac18e750	355	if (qp->net == nf && f->match(qp, arg)) {
2588fe1d	356	atomic_inc(&qp->refcnt);
19952cc4	357	spin_unlock(&hb->chain_lock);
06aa8b8a	358	qp_in->flags \|= INET_FRAG_COMPLETE;
2588fe1d PE	359	inet_frag_put(qp_in, f);
	360	return qp;
	361	}
	362	}
	363	#endif
	364	qp = qp_in;
b2fd5321	365	if (!mod_timer(&qp->timer, jiffies + nf->timeout))
2588fe1d PE	366	atomic_inc(&qp->refcnt);
	367
	368	atomic_inc(&qp->refcnt);
19952cc4	369	hlist_add_head(&qp->list, &hb->chain);
3fd588eb	370
19952cc4	371	spin_unlock(&hb->chain_lock);
24b9bf43	372
2588fe1d PE	373	return qp;
2588fe1d PE	374	}
e521db9d	375
ac18e750	376	static struct inet_frag_queue inet_frag_alloc(struct netns_frags nf,
f926e236 NA	377	struct inet_frags *f,
f926e236 NA	378	void *arg)
e521db9d PE	379	{
	380	struct inet_frag_queue *q;
	381
b13d3cbf FW	382	if (frag_mem_limit(nf) > nf->high_thresh) {
b13d3cbf FW	383	inet_frag_schedule_worker(f);
86e93e47	384	return NULL;
b13d3cbf	385	}
86e93e47	386
d4ad4d22	387	q = kmem_cache_zalloc(f->frags_cachep, GFP_ATOMIC);
e521db9d PE	388	if (q == NULL)
	389	return NULL;
	390
54db0cc2	391	q->net = nf;
c6fda282	392	f->constructor(q, arg);
d433673e JDB	393	add_frag_mem_limit(q, f->qsize);
d433673e JDB	394
e521db9d PE	395	setup_timer(&q->timer, f->frag_expire, (unsigned long)q);
	396	spin_lock_init(&q->lock);
	397	atomic_set(&q->refcnt, 1);
	398
	399	return q;
	400	}
c6fda282	401
ac18e750	402	static struct inet_frag_queue inet_frag_create(struct netns_frags nf,
f926e236 NA	403	struct inet_frags *f,
f926e236 NA	404	void *arg)
c6fda282 PE	405	{
	406	struct inet_frag_queue *q;
	407
ac18e750	408	q = inet_frag_alloc(nf, f, arg);
c6fda282 PE	409	if (q == NULL)
	410	return NULL;
	411
9a375803	412	return inet_frag_intern(nf, q, f, arg);
c6fda282	413	}
abd6523d	414
ac18e750	415	struct inet_frag_queue inet_frag_find(struct netns_frags nf,
f926e236 NA	416	struct inet_frags f, void key,
f926e236 NA	417	unsigned int hash)
abd6523d	418	{
19952cc4	419	struct inet_frag_bucket *hb;
abd6523d	420	struct inet_frag_queue *q;
5a3da1fe	421	int depth = 0;
abd6523d	422
b13d3cbf FW	423	if (frag_mem_limit(nf) > nf->low_thresh)
b13d3cbf FW	424	inet_frag_schedule_worker(f);
86e93e47	425
fb3cfe6e	426	hash &= (INETFRAGS_HASHSZ - 1);
19952cc4 JDB	427	hb = &f->hash[hash];
	428
	429	spin_lock(&hb->chain_lock);
	430	hlist_for_each_entry(q, &hb->chain, list) {
ac18e750	431	if (q->net == nf && f->match(q, key)) {
abd6523d	432	atomic_inc(&q->refcnt);
19952cc4	433	spin_unlock(&hb->chain_lock);
abd6523d PE	434	return q;
abd6523d PE	435	}
5a3da1fe	436	depth++;
abd6523d	437	}
19952cc4	438	spin_unlock(&hb->chain_lock);
abd6523d	439
5a3da1fe HFS	440	if (depth <= INETFRAGS_MAXDEPTH)
5a3da1fe HFS	441	return inet_frag_create(nf, f, key);
e3a57d18 FW	442
e3a57d18 FW	443	if (inet_frag_may_rebuild(f)) {
ab1c724f FW	444	if (!f->rebuild)
ab1c724f FW	445	f->rebuild = true;
e3a57d18 FW	446	inet_frag_schedule_worker(f);
	447	}
	448
	449	return ERR_PTR(-ENOBUFS);
abd6523d PE	450	}
abd6523d PE	451	EXPORT_SYMBOL(inet_frag_find);
5a3da1fe HFS	452
	453	void inet_frag_maybe_warn_overflow(struct inet_frag_queue *q,
	454	const char *prefix)
	455	{
	456	static const char msg[] = "inet_frag_find: Fragment hash bucket"
	457	" list length grew over limit " __stringify(INETFRAGS_MAXDEPTH)
	458	". Dropping fragment.\n";
	459
	460	if (PTR_ERR(q) == -ENOBUFS)
	461	LIMIT_NETDEBUG(KERN_WARNING "%s%s", prefix, msg);
	462	}
	463	EXPORT_SYMBOL(inet_frag_maybe_warn_overflow);