[mirror_ubuntu-bionic-kernel.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);

	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;
	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);
7eb95156 PE	95
b24b8a24 PE	96	setup_timer(&f->secret_timer, inet_frag_secret_rebuild,
b24b8a24 PE	97	(unsigned long)f);
3b4bc4a2	98	f->secret_timer.expires = jiffies + f->secret_interval;
321a3a99	99	add_timer(&f->secret_timer);
7eb95156 PE	100	}
	101	EXPORT_SYMBOL(inet_frags_init);
	102
e5a2bb84 PE	103	void inet_frags_init_net(struct netns_frags *nf)
	104	{
	105	nf->nqueues = 0;
d433673e	106	init_frag_mem_limit(nf);
3140c25c	107	INIT_LIST_HEAD(&nf->lru_list);
3ef0eb0d	108	spin_lock_init(&nf->lru_lock);
e5a2bb84 PE	109	}
	110	EXPORT_SYMBOL(inet_frags_init_net);
	111
7eb95156 PE	112	void inet_frags_fini(struct inet_frags *f)
7eb95156 PE	113	{
321a3a99	114	del_timer(&f->secret_timer);
7eb95156 PE	115	}
7eb95156 PE	116	EXPORT_SYMBOL(inet_frags_fini);
277e650d	117
81566e83 PE	118	void inet_frags_exit_net(struct netns_frags nf, struct inet_frags f)
	119	{
	120	nf->low_thresh = 0;
e8e16b70 DM	121
e8e16b70 DM	122	local_bh_disable();
6b102865	123	inet_frag_evictor(nf, f, true);
e8e16b70	124	local_bh_enable();
6d7b857d JDB	125
6d7b857d JDB	126	percpu_counter_destroy(&nf->mem);
81566e83 PE	127	}
	128	EXPORT_SYMBOL(inet_frags_exit_net);
	129
277e650d PE	130	static inline void fq_unlink(struct inet_frag_queue fq, struct inet_frags f)
277e650d PE	131	{
19952cc4 JDB	132	struct inet_frag_bucket *hb;
	133	unsigned int hash;
	134
	135	read_lock(&f->lock);
	136	hash = f->hashfn(fq);
	137	hb = &f->hash[hash];
	138
	139	spin_lock(&hb->chain_lock);
277e650d	140	hlist_del(&fq->list);
19952cc4 JDB	141	spin_unlock(&hb->chain_lock);
	142
	143	read_unlock(&f->lock);
3ef0eb0d	144	inet_frag_lru_del(fq);
277e650d PE	145	}
	146
	147	void inet_frag_kill(struct inet_frag_queue fq, struct inet_frags f)
	148	{
	149	if (del_timer(&fq->timer))
	150	atomic_dec(&fq->refcnt);
	151
bc578a54	152	if (!(fq->last_in & INET_FRAG_COMPLETE)) {
277e650d PE	153	fq_unlink(fq, f);
277e650d PE	154	atomic_dec(&fq->refcnt);
bc578a54	155	fq->last_in \|= INET_FRAG_COMPLETE;
277e650d PE	156	}
277e650d PE	157	}
277e650d	158	EXPORT_SYMBOL(inet_frag_kill);
1e4b8287	159
6ddc0822	160	static inline void frag_kfree_skb(struct netns_frags nf, struct inet_frags f,
d433673e	161	struct sk_buff *skb)
1e4b8287	162	{
1e4b8287 PE	163	if (f->skb_free)
	164	f->skb_free(skb);
	165	kfree_skb(skb);
	166	}
	167
	168	void inet_frag_destroy(struct inet_frag_queue q, struct inet_frags f,
	169	int *work)
	170	{
	171	struct sk_buff *fp;
6ddc0822	172	struct netns_frags *nf;
d433673e	173	unsigned int sum, sum_truesize = 0;
1e4b8287	174
547b792c IJ	175	WARN_ON(!(q->last_in & INET_FRAG_COMPLETE));
547b792c IJ	176	WARN_ON(del_timer(&q->timer) != 0);
1e4b8287 PE	177
	178	/* Release all fragment data. */
	179	fp = q->fragments;
6ddc0822	180	nf = q->net;
1e4b8287 PE	181	while (fp) {
	182	struct sk_buff *xp = fp->next;
	183
d433673e JDB	184	sum_truesize += fp->truesize;
d433673e JDB	185	frag_kfree_skb(nf, f, fp);
1e4b8287 PE	186	fp = xp;
1e4b8287 PE	187	}
d433673e	188	sum = sum_truesize + f->qsize;
1e4b8287	189	if (work)
d433673e JDB	190	*work -= sum;
d433673e JDB	191	sub_frag_mem_limit(q, sum);
1e4b8287	192
c9547709 PE	193	if (f->destructor)
	194	f->destructor(q);
	195	kfree(q);
1e4b8287 PE	196
	197	}
	198	EXPORT_SYMBOL(inet_frag_destroy);
8e7999c4	199
6b102865	200	int inet_frag_evictor(struct netns_frags nf, struct inet_frags f, bool force)
8e7999c4 PE	201	{
	202	struct inet_frag_queue *q;
	203	int work, evicted = 0;
	204
6b102865	205	if (!force) {
d433673e	206	if (frag_mem_limit(nf) <= nf->high_thresh)
6b102865 AW	207	return 0;
	208	}
	209
d433673e	210	work = frag_mem_limit(nf) - nf->low_thresh;
e588e2f2	211	while (work > 0 \|\| force) {
3ef0eb0d JDB	212	spin_lock(&nf->lru_lock);
3ef0eb0d JDB	213
3140c25c	214	if (list_empty(&nf->lru_list)) {
3ef0eb0d	215	spin_unlock(&nf->lru_lock);
8e7999c4 PE	216	break;
	217	}
	218
3140c25c	219	q = list_first_entry(&nf->lru_list,
8e7999c4 PE	220	struct inet_frag_queue, lru_list);
8e7999c4 PE	221	atomic_inc(&q->refcnt);
68399ac3 JDB	222	/* Remove q from list to avoid several CPUs grabbing it */
	223	list_del_init(&q->lru_list);
	224
3ef0eb0d	225	spin_unlock(&nf->lru_lock);
8e7999c4 PE	226
8e7999c4 PE	227	spin_lock(&q->lock);
bc578a54	228	if (!(q->last_in & INET_FRAG_COMPLETE))
8e7999c4 PE	229	inet_frag_kill(q, f);
	230	spin_unlock(&q->lock);
	231
	232	if (atomic_dec_and_test(&q->refcnt))
	233	inet_frag_destroy(q, f, &work);
	234	evicted++;
	235	}
	236
	237	return evicted;
	238	}
	239	EXPORT_SYMBOL(inet_frag_evictor);
2588fe1d	240
ac18e750 PE	241	static struct inet_frag_queue inet_frag_intern(struct netns_frags nf,
ac18e750 PE	242	struct inet_frag_queue qp_in, struct inet_frags f,
9a375803	243	void *arg)
2588fe1d	244	{
19952cc4	245	struct inet_frag_bucket *hb;
2588fe1d	246	struct inet_frag_queue *qp;
9a375803	247	unsigned int hash;
2588fe1d	248
19952cc4	249	read_lock(&f->lock); /* Protects against hash rebuild */
9a375803 PE	250	/*
	251	* While we stayed w/o the lock other CPU could update
	252	* the rnd seed, so we need to re-calculate the hash
	253	* chain. Fortunatelly the qp_in can be used to get one.
	254	*/
	255	hash = f->hashfn(qp_in);
19952cc4 JDB	256	hb = &f->hash[hash];
	257	spin_lock(&hb->chain_lock);
	258
2588fe1d PE	259	#ifdef CONFIG_SMP
	260	/* With SMP race we have to recheck hash table, because
	261	* such entry could be created on other cpu, while we
19952cc4	262	* released the hash bucket lock.
2588fe1d	263	*/
19952cc4	264	hlist_for_each_entry(qp, &hb->chain, list) {
ac18e750	265	if (qp->net == nf && f->match(qp, arg)) {
2588fe1d	266	atomic_inc(&qp->refcnt);
19952cc4 JDB	267	spin_unlock(&hb->chain_lock);
19952cc4 JDB	268	read_unlock(&f->lock);
bc578a54	269	qp_in->last_in \|= INET_FRAG_COMPLETE;
2588fe1d PE	270	inet_frag_put(qp_in, f);
	271	return qp;
	272	}
	273	}
	274	#endif
	275	qp = qp_in;
b2fd5321	276	if (!mod_timer(&qp->timer, jiffies + nf->timeout))
2588fe1d PE	277	atomic_inc(&qp->refcnt);
	278
	279	atomic_inc(&qp->refcnt);
19952cc4	280	hlist_add_head(&qp->list, &hb->chain);
24b9bf43	281	inet_frag_lru_add(nf, qp);
19952cc4 JDB	282	spin_unlock(&hb->chain_lock);
19952cc4 JDB	283	read_unlock(&f->lock);
24b9bf43	284
2588fe1d PE	285	return qp;
2588fe1d PE	286	}
e521db9d	287
ac18e750 PE	288	static struct inet_frag_queue inet_frag_alloc(struct netns_frags nf,
ac18e750 PE	289	struct inet_frags f, void arg)
e521db9d PE	290	{
	291	struct inet_frag_queue *q;
	292
	293	q = kzalloc(f->qsize, GFP_ATOMIC);
	294	if (q == NULL)
	295	return NULL;
	296
54db0cc2	297	q->net = nf;
c6fda282	298	f->constructor(q, arg);
d433673e JDB	299	add_frag_mem_limit(q, f->qsize);
d433673e JDB	300
e521db9d PE	301	setup_timer(&q->timer, f->frag_expire, (unsigned long)q);
	302	spin_lock_init(&q->lock);
	303	atomic_set(&q->refcnt, 1);
b56141ab	304	INIT_LIST_HEAD(&q->lru_list);
e521db9d PE	305
	306	return q;
	307	}
c6fda282	308
ac18e750	309	static struct inet_frag_queue inet_frag_create(struct netns_frags nf,
9a375803	310	struct inet_frags f, void arg)
c6fda282 PE	311	{
	312	struct inet_frag_queue *q;
	313
ac18e750	314	q = inet_frag_alloc(nf, f, arg);
c6fda282 PE	315	if (q == NULL)
	316	return NULL;
	317
9a375803	318	return inet_frag_intern(nf, q, f, arg);
c6fda282	319	}
abd6523d	320
ac18e750 PE	321	struct inet_frag_queue inet_frag_find(struct netns_frags nf,
ac18e750 PE	322	struct inet_frags f, void key, unsigned int hash)
56bca31f	323	__releases(&f->lock)
abd6523d	324	{
19952cc4	325	struct inet_frag_bucket *hb;
abd6523d	326	struct inet_frag_queue *q;
5a3da1fe	327	int depth = 0;
abd6523d	328
19952cc4 JDB	329	hb = &f->hash[hash];
	330
	331	spin_lock(&hb->chain_lock);
	332	hlist_for_each_entry(q, &hb->chain, list) {
ac18e750	333	if (q->net == nf && f->match(q, key)) {
abd6523d	334	atomic_inc(&q->refcnt);
19952cc4	335	spin_unlock(&hb->chain_lock);
abd6523d PE	336	read_unlock(&f->lock);
	337	return q;
	338	}
5a3da1fe	339	depth++;
abd6523d	340	}
19952cc4	341	spin_unlock(&hb->chain_lock);
abd6523d PE	342	read_unlock(&f->lock);
abd6523d PE	343
5a3da1fe HFS	344	if (depth <= INETFRAGS_MAXDEPTH)
	345	return inet_frag_create(nf, f, key);
	346	else
	347	return ERR_PTR(-ENOBUFS);
abd6523d PE	348	}
abd6523d PE	349	EXPORT_SYMBOL(inet_frag_find);
5a3da1fe HFS	350
	351	void inet_frag_maybe_warn_overflow(struct inet_frag_queue *q,
	352	const char *prefix)
	353	{
	354	static const char msg[] = "inet_frag_find: Fragment hash bucket"
	355	" list length grew over limit " __stringify(INETFRAGS_MAXDEPTH)
	356	". Dropping fragment.\n";
	357
	358	if (PTR_ERR(q) == -ENOBUFS)
	359	LIMIT_NETDEBUG(KERN_WARNING "%s%s", prefix, msg);
	360	}
	361	EXPORT_SYMBOL(inet_frag_maybe_warn_overflow);