[GitHub/mt8127/android_kernel_alcatel_ttab.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>

/* 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 void inet_frag_secret_rebuild(unsigned long dummy)
{
	struct inet_frags *f = (struct inet_frags *)dummy;
	unsigned long now = jiffies;
	int i;

	/* Per bucket lock NOT needed here, due to write lock protection */
	write_lock(&f->lock);

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

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

				hlist_del(&q->list);

				/* Relink to new hash chain. */
				hb_dest = &f->hash[hval];
				hlist_add_head(&q->list, &hb_dest->chain);
			}
		}
	}
	write_unlock(&f->lock);

	mod_timer(&f->secret_timer, now + f->secret_interval);
}

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

	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);
	}
	rwlock_init(&f->lock);

	f->rnd = (u32) ((num_physpages ^ (num_physpages>>7)) ^
				   (jiffies ^ (jiffies >> 6)));

	setup_timer(&f->secret_timer, inet_frag_secret_rebuild,
			(unsigned long)f);
	f->secret_timer.expires = jiffies + f->secret_interval;
	add_timer(&f->secret_timer);
}
EXPORT_SYMBOL(inet_frags_init);

void inet_frags_init_net(struct netns_frags *nf)
{
	nf->nqueues = 0;
	init_frag_mem_limit(nf);
	INIT_LIST_HEAD(&nf->lru_list);
	spin_lock_init(&nf->lru_lock);
}
EXPORT_SYMBOL(inet_frags_init_net);

void inet_frags_fini(struct inet_frags *f)
{
	del_timer(&f->secret_timer);
}
EXPORT_SYMBOL(inet_frags_fini);

void inet_frags_exit_net(struct netns_frags *nf, struct inet_frags *f)
{
	nf->low_thresh = 0;

	local_bh_disable();
	inet_frag_evictor(nf, f, true);
	local_bh_enable();

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

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

	read_lock(&f->lock);
	hash = f->hashfn(fq);
	hb = &f->hash[hash];

	spin_lock(&hb->chain_lock);
	hlist_del(&fq->list);
	spin_unlock(&hb->chain_lock);

	read_unlock(&f->lock);
	inet_frag_lru_del(fq);
}

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

	if (!(fq->last_in & INET_FRAG_COMPLETE)) {
		fq_unlink(fq, f);
		atomic_dec(&fq->refcnt);
		fq->last_in |= 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,
					int *work)
{
	struct sk_buff *fp;
	struct netns_frags *nf;
	unsigned int sum, sum_truesize = 0;

	WARN_ON(!(q->last_in & 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;
	if (work)
		*work -= sum;
	sub_frag_mem_limit(q, sum);

	if (f->destructor)
		f->destructor(q);
	kfree(q);

}
EXPORT_SYMBOL(inet_frag_destroy);

int inet_frag_evictor(struct netns_frags *nf, struct inet_frags *f, bool force)
{
	struct inet_frag_queue *q;
	int work, evicted = 0;

	if (!force) {
		if (frag_mem_limit(nf) <= nf->high_thresh)
			return 0;
	}

	work = frag_mem_limit(nf) - nf->low_thresh;
	while (work > 0 || force) {
		spin_lock(&nf->lru_lock);

		if (list_empty(&nf->lru_list)) {
			spin_unlock(&nf->lru_lock);
			break;
		}

		q = list_first_entry(&nf->lru_list,
				struct inet_frag_queue, lru_list);
		atomic_inc(&q->refcnt);
		/* Remove q from list to avoid several CPUs grabbing it */
		list_del_init(&q->lru_list);

		spin_unlock(&nf->lru_lock);

		spin_lock(&q->lock);
		if (!(q->last_in & INET_FRAG_COMPLETE))
			inet_frag_kill(q, f);
		spin_unlock(&q->lock);

		if (atomic_dec_and_test(&q->refcnt))
			inet_frag_destroy(q, f, &work);
		evicted++;
	}

	return evicted;
}
EXPORT_SYMBOL(inet_frag_evictor);

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;
	struct inet_frag_queue *qp;
#ifdef CONFIG_SMP
#endif
	unsigned int hash;

	read_lock(&f->lock); /* Protects against hash rebuild */
	/*
	 * While we stayed w/o the lock other CPU could update
	 * the rnd seed, so we need to re-calculate the hash
	 * chain. Fortunatelly the qp_in can be used to get one.
	 */
	hash = f->hashfn(qp_in);
	hb = &f->hash[hash];
	spin_lock(&hb->chain_lock);

#ifdef CONFIG_SMP
	/* With SMP race we have to recheck hash table, because
	 * such entry could be created on other cpu, while we
	 * released the 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);
			read_unlock(&f->lock);
			qp_in->last_in |= 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);
	inet_frag_lru_add(nf, qp);
	spin_unlock(&hb->chain_lock);
	read_unlock(&f->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;

	q = kzalloc(f->qsize, 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);
	INIT_LIST_HEAD(&q->lru_list);

	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)
	__releases(&f->lock)
{
	struct inet_frag_bucket *hb;
	struct inet_frag_queue *q;
	int depth = 0;

	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);
			read_unlock(&f->lock);
			return q;
		}
		depth++;
	}
	spin_unlock(&hb->chain_lock);
	read_unlock(&f->lock);

	if (depth <= INETFRAGS_MAXDEPTH)
		return inet_frag_create(nf, f, key);
	else
		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>
	27
	28	/* Given the OR values of all fragments, apply RFC 3168 5.3 requirements
	29	* Value : 0xff if frame should be dropped.
	30	* 0 or INET_ECN_CE value, to be ORed in to final iph->tos field
	31	*/
	32	const u8 ip_frag_ecn_table[16] = {
	33	/* at least one fragment had CE, and others ECT_0 or ECT_1 */
	34	[IPFRAG_ECN_CE \| IPFRAG_ECN_ECT_0] = INET_ECN_CE,
	35	[IPFRAG_ECN_CE \| IPFRAG_ECN_ECT_1] = INET_ECN_CE,
	36	[IPFRAG_ECN_CE \| IPFRAG_ECN_ECT_0 \| IPFRAG_ECN_ECT_1] = INET_ECN_CE,
	37
	38	/* invalid combinations : drop frame */
	39	[IPFRAG_ECN_NOT_ECT \| IPFRAG_ECN_CE] = 0xff,
	40	[IPFRAG_ECN_NOT_ECT \| IPFRAG_ECN_ECT_0] = 0xff,
	41	[IPFRAG_ECN_NOT_ECT \| IPFRAG_ECN_ECT_1] = 0xff,
	42	[IPFRAG_ECN_NOT_ECT \| IPFRAG_ECN_ECT_0 \| IPFRAG_ECN_ECT_1] = 0xff,
	43	[IPFRAG_ECN_NOT_ECT \| IPFRAG_ECN_CE \| IPFRAG_ECN_ECT_0] = 0xff,
	44	[IPFRAG_ECN_NOT_ECT \| IPFRAG_ECN_CE \| IPFRAG_ECN_ECT_1] = 0xff,
	45	[IPFRAG_ECN_NOT_ECT \| IPFRAG_ECN_CE \| IPFRAG_ECN_ECT_0 \| IPFRAG_ECN_ECT_1] = 0xff,
	46	};
	47	EXPORT_SYMBOL(ip_frag_ecn_table);
7eb95156	48
321a3a99 PE	49	static void inet_frag_secret_rebuild(unsigned long dummy)
	50	{
	51	struct inet_frags f = (struct inet_frags )dummy;
	52	unsigned long now = jiffies;
	53	int i;
	54
19952cc4	55	/* Per bucket lock NOT needed here, due to write lock protection */
321a3a99	56	write_lock(&f->lock);
19952cc4	57
321a3a99 PE	58	get_random_bytes(&f->rnd, sizeof(u32));
321a3a99 PE	59	for (i = 0; i < INETFRAGS_HASHSZ; i++) {
19952cc4	60	struct inet_frag_bucket *hb;
321a3a99	61	struct inet_frag_queue *q;
b67bfe0d	62	struct hlist_node *n;
321a3a99	63
19952cc4 JDB	64	hb = &f->hash[i];
19952cc4 JDB	65	hlist_for_each_entry_safe(q, n, &hb->chain, list) {
321a3a99 PE	66	unsigned int hval = f->hashfn(q);
	67
	68	if (hval != i) {
19952cc4 JDB	69	struct inet_frag_bucket *hb_dest;
19952cc4 JDB	70
321a3a99 PE	71	hlist_del(&q->list);
	72
	73	/* Relink to new hash chain. */
19952cc4 JDB	74	hb_dest = &f->hash[hval];
19952cc4 JDB	75	hlist_add_head(&q->list, &hb_dest->chain);
321a3a99 PE	76	}
	77	}
	78	}
	79	write_unlock(&f->lock);
	80
3b4bc4a2	81	mod_timer(&f->secret_timer, now + f->secret_interval);
321a3a99 PE	82	}
321a3a99 PE	83
7eb95156 PE	84	void inet_frags_init(struct inet_frags *f)
	85	{
	86	int i;
	87
19952cc4 JDB	88	for (i = 0; i < INETFRAGS_HASHSZ; i++) {
19952cc4 JDB	89	struct inet_frag_bucket *hb = &f->hash[i];
7eb95156	90
19952cc4 JDB	91	spin_lock_init(&hb->chain_lock);
	92	INIT_HLIST_HEAD(&hb->chain);
	93	}
7eb95156 PE	94	rwlock_init(&f->lock);
	95
	96	f->rnd = (u32) ((num_physpages ^ (num_physpages>>7)) ^
	97	(jiffies ^ (jiffies >> 6)));
	98
b24b8a24 PE	99	setup_timer(&f->secret_timer, inet_frag_secret_rebuild,
b24b8a24 PE	100	(unsigned long)f);
3b4bc4a2	101	f->secret_timer.expires = jiffies + f->secret_interval;
321a3a99	102	add_timer(&f->secret_timer);
7eb95156 PE	103	}
	104	EXPORT_SYMBOL(inet_frags_init);
	105
e5a2bb84 PE	106	void inet_frags_init_net(struct netns_frags *nf)
	107	{
	108	nf->nqueues = 0;
d433673e	109	init_frag_mem_limit(nf);
3140c25c	110	INIT_LIST_HEAD(&nf->lru_list);
3ef0eb0d	111	spin_lock_init(&nf->lru_lock);
e5a2bb84 PE	112	}
	113	EXPORT_SYMBOL(inet_frags_init_net);
	114
7eb95156 PE	115	void inet_frags_fini(struct inet_frags *f)
7eb95156 PE	116	{
321a3a99	117	del_timer(&f->secret_timer);
7eb95156 PE	118	}
7eb95156 PE	119	EXPORT_SYMBOL(inet_frags_fini);
277e650d	120
81566e83 PE	121	void inet_frags_exit_net(struct netns_frags nf, struct inet_frags f)
	122	{
	123	nf->low_thresh = 0;
e8e16b70 DM	124
e8e16b70 DM	125	local_bh_disable();
6b102865	126	inet_frag_evictor(nf, f, true);
e8e16b70	127	local_bh_enable();
6d7b857d JDB	128
6d7b857d JDB	129	percpu_counter_destroy(&nf->mem);
81566e83 PE	130	}
	131	EXPORT_SYMBOL(inet_frags_exit_net);
	132
277e650d PE	133	static inline void fq_unlink(struct inet_frag_queue fq, struct inet_frags f)
277e650d PE	134	{
19952cc4 JDB	135	struct inet_frag_bucket *hb;
	136	unsigned int hash;
	137
	138	read_lock(&f->lock);
	139	hash = f->hashfn(fq);
	140	hb = &f->hash[hash];
	141
	142	spin_lock(&hb->chain_lock);
277e650d	143	hlist_del(&fq->list);
19952cc4 JDB	144	spin_unlock(&hb->chain_lock);
	145
	146	read_unlock(&f->lock);
3ef0eb0d	147	inet_frag_lru_del(fq);
277e650d PE	148	}
	149
	150	void inet_frag_kill(struct inet_frag_queue fq, struct inet_frags f)
	151	{
	152	if (del_timer(&fq->timer))
	153	atomic_dec(&fq->refcnt);
	154
bc578a54	155	if (!(fq->last_in & INET_FRAG_COMPLETE)) {
277e650d PE	156	fq_unlink(fq, f);
277e650d PE	157	atomic_dec(&fq->refcnt);
bc578a54	158	fq->last_in \|= INET_FRAG_COMPLETE;
277e650d PE	159	}
277e650d PE	160	}
277e650d	161	EXPORT_SYMBOL(inet_frag_kill);
1e4b8287	162
6ddc0822	163	static inline void frag_kfree_skb(struct netns_frags nf, struct inet_frags f,
d433673e	164	struct sk_buff *skb)
1e4b8287	165	{
1e4b8287 PE	166	if (f->skb_free)
	167	f->skb_free(skb);
	168	kfree_skb(skb);
	169	}
	170
	171	void inet_frag_destroy(struct inet_frag_queue q, struct inet_frags f,
	172	int *work)
	173	{
	174	struct sk_buff *fp;
6ddc0822	175	struct netns_frags *nf;
d433673e	176	unsigned int sum, sum_truesize = 0;
1e4b8287	177
547b792c IJ	178	WARN_ON(!(q->last_in & INET_FRAG_COMPLETE));
547b792c IJ	179	WARN_ON(del_timer(&q->timer) != 0);
1e4b8287 PE	180
	181	/* Release all fragment data. */
	182	fp = q->fragments;
6ddc0822	183	nf = q->net;
1e4b8287 PE	184	while (fp) {
	185	struct sk_buff *xp = fp->next;
	186
d433673e JDB	187	sum_truesize += fp->truesize;
d433673e JDB	188	frag_kfree_skb(nf, f, fp);
1e4b8287 PE	189	fp = xp;
1e4b8287 PE	190	}
d433673e	191	sum = sum_truesize + f->qsize;
1e4b8287	192	if (work)
d433673e JDB	193	*work -= sum;
d433673e JDB	194	sub_frag_mem_limit(q, sum);
1e4b8287	195
c9547709 PE	196	if (f->destructor)
	197	f->destructor(q);
	198	kfree(q);
1e4b8287 PE	199
	200	}
	201	EXPORT_SYMBOL(inet_frag_destroy);
8e7999c4	202
6b102865	203	int inet_frag_evictor(struct netns_frags nf, struct inet_frags f, bool force)
8e7999c4 PE	204	{
	205	struct inet_frag_queue *q;
	206	int work, evicted = 0;
	207
6b102865	208	if (!force) {
d433673e	209	if (frag_mem_limit(nf) <= nf->high_thresh)
6b102865 AW	210	return 0;
	211	}
	212
d433673e	213	work = frag_mem_limit(nf) - nf->low_thresh;
458b05e1	214	while (work > 0 \|\| force) {
3ef0eb0d JDB	215	spin_lock(&nf->lru_lock);
3ef0eb0d JDB	216
3140c25c	217	if (list_empty(&nf->lru_list)) {
3ef0eb0d	218	spin_unlock(&nf->lru_lock);
8e7999c4 PE	219	break;
	220	}
	221
3140c25c	222	q = list_first_entry(&nf->lru_list,
8e7999c4 PE	223	struct inet_frag_queue, lru_list);
8e7999c4 PE	224	atomic_inc(&q->refcnt);
68399ac3 JDB	225	/* Remove q from list to avoid several CPUs grabbing it */
	226	list_del_init(&q->lru_list);
	227
3ef0eb0d	228	spin_unlock(&nf->lru_lock);
8e7999c4 PE	229
8e7999c4 PE	230	spin_lock(&q->lock);
bc578a54	231	if (!(q->last_in & INET_FRAG_COMPLETE))
8e7999c4 PE	232	inet_frag_kill(q, f);
	233	spin_unlock(&q->lock);
	234
	235	if (atomic_dec_and_test(&q->refcnt))
	236	inet_frag_destroy(q, f, &work);
	237	evicted++;
	238	}
	239
	240	return evicted;
	241	}
	242	EXPORT_SYMBOL(inet_frag_evictor);
2588fe1d	243
ac18e750 PE	244	static struct inet_frag_queue inet_frag_intern(struct netns_frags nf,
ac18e750 PE	245	struct inet_frag_queue qp_in, struct inet_frags f,
9a375803	246	void *arg)
2588fe1d	247	{
19952cc4	248	struct inet_frag_bucket *hb;
2588fe1d PE	249	struct inet_frag_queue *qp;
2588fe1d PE	250	#ifdef CONFIG_SMP
2588fe1d	251	#endif
9a375803	252	unsigned int hash;
2588fe1d	253
19952cc4	254	read_lock(&f->lock); /* Protects against hash rebuild */
9a375803 PE	255	/*
	256	* While we stayed w/o the lock other CPU could update
	257	* the rnd seed, so we need to re-calculate the hash
	258	* chain. Fortunatelly the qp_in can be used to get one.
	259	*/
	260	hash = f->hashfn(qp_in);
19952cc4 JDB	261	hb = &f->hash[hash];
	262	spin_lock(&hb->chain_lock);
	263
2588fe1d PE	264	#ifdef CONFIG_SMP
	265	/* With SMP race we have to recheck hash table, because
	266	* such entry could be created on other cpu, while we
19952cc4	267	* released the hash bucket lock.
2588fe1d	268	*/
19952cc4	269	hlist_for_each_entry(qp, &hb->chain, list) {
ac18e750	270	if (qp->net == nf && f->match(qp, arg)) {
2588fe1d	271	atomic_inc(&qp->refcnt);
19952cc4 JDB	272	spin_unlock(&hb->chain_lock);
19952cc4 JDB	273	read_unlock(&f->lock);
bc578a54	274	qp_in->last_in \|= INET_FRAG_COMPLETE;
2588fe1d PE	275	inet_frag_put(qp_in, f);
	276	return qp;
	277	}
	278	}
	279	#endif
	280	qp = qp_in;
b2fd5321	281	if (!mod_timer(&qp->timer, jiffies + nf->timeout))
2588fe1d PE	282	atomic_inc(&qp->refcnt);
	283
	284	atomic_inc(&qp->refcnt);
19952cc4	285	hlist_add_head(&qp->list, &hb->chain);
f7a2e253	286	inet_frag_lru_add(nf, qp);
19952cc4 JDB	287	spin_unlock(&hb->chain_lock);
19952cc4 JDB	288	read_unlock(&f->lock);
f7a2e253	289
2588fe1d PE	290	return qp;
2588fe1d PE	291	}
e521db9d	292
ac18e750 PE	293	static struct inet_frag_queue inet_frag_alloc(struct netns_frags nf,
ac18e750 PE	294	struct inet_frags f, void arg)
e521db9d PE	295	{
	296	struct inet_frag_queue *q;
	297
	298	q = kzalloc(f->qsize, GFP_ATOMIC);
	299	if (q == NULL)
	300	return NULL;
	301
54db0cc2	302	q->net = nf;
c6fda282	303	f->constructor(q, arg);
d433673e JDB	304	add_frag_mem_limit(q, f->qsize);
d433673e JDB	305
e521db9d PE	306	setup_timer(&q->timer, f->frag_expire, (unsigned long)q);
	307	spin_lock_init(&q->lock);
	308	atomic_set(&q->refcnt, 1);
b56141ab	309	INIT_LIST_HEAD(&q->lru_list);
e521db9d PE	310
	311	return q;
	312	}
c6fda282	313
ac18e750	314	static struct inet_frag_queue inet_frag_create(struct netns_frags nf,
9a375803	315	struct inet_frags f, void arg)
c6fda282 PE	316	{
	317	struct inet_frag_queue *q;
	318
ac18e750	319	q = inet_frag_alloc(nf, f, arg);
c6fda282 PE	320	if (q == NULL)
	321	return NULL;
	322
9a375803	323	return inet_frag_intern(nf, q, f, arg);
c6fda282	324	}
abd6523d	325
ac18e750 PE	326	struct inet_frag_queue inet_frag_find(struct netns_frags nf,
ac18e750 PE	327	struct inet_frags f, void key, unsigned int hash)
56bca31f	328	__releases(&f->lock)
abd6523d	329	{
19952cc4	330	struct inet_frag_bucket *hb;
abd6523d	331	struct inet_frag_queue *q;
5a3da1fe	332	int depth = 0;
abd6523d	333
19952cc4 JDB	334	hb = &f->hash[hash];
	335
	336	spin_lock(&hb->chain_lock);
	337	hlist_for_each_entry(q, &hb->chain, list) {
ac18e750	338	if (q->net == nf && f->match(q, key)) {
abd6523d	339	atomic_inc(&q->refcnt);
19952cc4	340	spin_unlock(&hb->chain_lock);
abd6523d PE	341	read_unlock(&f->lock);
	342	return q;
	343	}
5a3da1fe	344	depth++;
abd6523d	345	}
19952cc4	346	spin_unlock(&hb->chain_lock);
abd6523d PE	347	read_unlock(&f->lock);
abd6523d PE	348
5a3da1fe HFS	349	if (depth <= INETFRAGS_MAXDEPTH)
	350	return inet_frag_create(nf, f, key);
	351	else
	352	return ERR_PTR(-ENOBUFS);
abd6523d PE	353	}
abd6523d PE	354	EXPORT_SYMBOL(inet_frag_find);
5a3da1fe HFS	355
	356	void inet_frag_maybe_warn_overflow(struct inet_frag_queue *q,
	357	const char *prefix)
	358	{
	359	static const char msg[] = "inet_frag_find: Fragment hash bucket"
	360	" list length grew over limit " __stringify(INETFRAGS_MAXDEPTH)
	361	". Dropping fragment.\n";
	362
	363	if (PTR_ERR(q) == -ENOBUFS)
	364	LIMIT_NETDEBUG(KERN_WARNING "%s%s", prefix, msg);
	365	}
	366	EXPORT_SYMBOL(inet_frag_maybe_warn_overflow);