[mirror_ubuntu-bionic-kernel.git] / include / net / inet_frag.h

#ifndef __NET_FRAG_H__
#define __NET_FRAG_H__

#include <linux/percpu_counter.h>

struct netns_frags {
	int			nqueues;
	struct list_head	lru_list;
	spinlock_t		lru_lock;

	/* The percpu_counter "mem" need to be cacheline aligned.
	 *  mem.count must not share cacheline with other writers
	 */
	struct percpu_counter   mem ____cacheline_aligned_in_smp;

	/* sysctls */
	int			timeout;
	int			high_thresh;
	int			low_thresh;
};

struct inet_frag_queue {
	spinlock_t		lock;
	struct timer_list	timer;      /* when will this queue expire? */
	struct list_head	lru_list;   /* lru list member */
	struct hlist_node	list;
	atomic_t		refcnt;
	struct sk_buff		*fragments; /* list of received fragments */
	struct sk_buff		*fragments_tail;
	ktime_t			stamp;
	int			len;        /* total length of orig datagram */
	int			meat;
	__u8			last_in;    /* first/last segment arrived? */

#define INET_FRAG_COMPLETE	4
#define INET_FRAG_FIRST_IN	2
#define INET_FRAG_LAST_IN	1

	u16			max_size;

	struct netns_frags	*net;
};

#define INETFRAGS_HASHSZ	1024

/* averaged:
 * max_depth = default ipfrag_high_thresh / INETFRAGS_HASHSZ /
 *	       rounded up (SKB_TRUELEN(0) + sizeof(struct ipq or
 *	       struct frag_queue))
 */
#define INETFRAGS_MAXDEPTH		128

struct inet_frag_bucket {
	struct hlist_head	chain;
	spinlock_t		chain_lock;
};

struct inet_frags {
	struct inet_frag_bucket	hash[INETFRAGS_HASHSZ];
	/* This rwlock is a global lock (seperate per IPv4, IPv6 and
	 * netfilter). Important to keep this on a seperate cacheline.
	 * Its primarily a rebuild protection rwlock.
	 */
	rwlock_t		lock ____cacheline_aligned_in_smp;
	int			secret_interval;
	struct timer_list	secret_timer;

	/* The first call to hashfn is responsible to initialize
	 * rnd. This is best done with net_get_random_once.
	 */
	u32			rnd;
	int			qsize;

	unsigned int		(*hashfn)(struct inet_frag_queue *);
	bool			(*match)(struct inet_frag_queue *q, void *arg);
	void			(*constructor)(struct inet_frag_queue *q,
						void *arg);
	void			(*destructor)(struct inet_frag_queue *);
	void			(*skb_free)(struct sk_buff *);
	void			(*frag_expire)(unsigned long data);
};

void inet_frags_init(struct inet_frags *);
void inet_frags_fini(struct inet_frags *);

void inet_frags_init_net(struct netns_frags *nf);
void inet_frags_exit_net(struct netns_frags *nf, struct inet_frags *f);

void inet_frag_kill(struct inet_frag_queue *q, struct inet_frags *f);
void inet_frag_destroy(struct inet_frag_queue *q,
				struct inet_frags *f, int *work);
int inet_frag_evictor(struct netns_frags *nf, struct inet_frags *f, bool force);
struct inet_frag_queue *inet_frag_find(struct netns_frags *nf,
		struct inet_frags *f, void *key, unsigned int hash)
	__releases(&f->lock);
void inet_frag_maybe_warn_overflow(struct inet_frag_queue *q,
				   const char *prefix);

static inline void inet_frag_put(struct inet_frag_queue *q, struct inet_frags *f)
{
	if (atomic_dec_and_test(&q->refcnt))
		inet_frag_destroy(q, f, NULL);
}

/* Memory Tracking Functions. */

/* The default percpu_counter batch size is not big enough to scale to
 * fragmentation mem acct sizes.
 * The mem size of a 64K fragment is approx:
 *  (44 fragments * 2944 truesize) + frag_queue struct(200) = 129736 bytes
 */
static unsigned int frag_percpu_counter_batch = 130000;

static inline int frag_mem_limit(struct netns_frags *nf)
{
	return percpu_counter_read(&nf->mem);
}

static inline void sub_frag_mem_limit(struct inet_frag_queue *q, int i)
{
	__percpu_counter_add(&q->net->mem, -i, frag_percpu_counter_batch);
}

static inline void add_frag_mem_limit(struct inet_frag_queue *q, int i)
{
	__percpu_counter_add(&q->net->mem, i, frag_percpu_counter_batch);
}

static inline void init_frag_mem_limit(struct netns_frags *nf)
{
	percpu_counter_init(&nf->mem, 0);
}

static inline int sum_frag_mem_limit(struct netns_frags *nf)
{
	int res;

	local_bh_disable();
	res = percpu_counter_sum_positive(&nf->mem);
	local_bh_enable();

	return res;
}

static inline void inet_frag_lru_move(struct inet_frag_queue *q)
{
	spin_lock(&q->net->lru_lock);
	if (!list_empty(&q->lru_list))
		list_move_tail(&q->lru_list, &q->net->lru_list);
	spin_unlock(&q->net->lru_lock);
}

static inline void inet_frag_lru_del(struct inet_frag_queue *q)
{
	spin_lock(&q->net->lru_lock);
	list_del_init(&q->lru_list);
	q->net->nqueues--;
	spin_unlock(&q->net->lru_lock);
}

static inline void inet_frag_lru_add(struct netns_frags *nf,
				     struct inet_frag_queue *q)
{
	spin_lock(&nf->lru_lock);
	list_add_tail(&q->lru_list, &nf->lru_list);
	q->net->nqueues++;
	spin_unlock(&nf->lru_lock);
}

/* RFC 3168 support :
 * We want to check ECN values of all fragments, do detect invalid combinations.
 * In ipq->ecn, we store the OR value of each ip4_frag_ecn() fragment value.
 */
#define	IPFRAG_ECN_NOT_ECT	0x01 /* one frag had ECN_NOT_ECT */
#define	IPFRAG_ECN_ECT_1	0x02 /* one frag had ECN_ECT_1 */
#define	IPFRAG_ECN_ECT_0	0x04 /* one frag had ECN_ECT_0 */
#define	IPFRAG_ECN_CE		0x08 /* one frag had ECN_CE */

extern const u8 ip_frag_ecn_table[16];

#endif
Commit	Line	Data
5ab11c98 PE	1	#ifndef __NET_FRAG_H__
	2	#define __NET_FRAG_H__
	3
6d7b857d JDB	4	#include <linux/percpu_counter.h>
6d7b857d JDB	5
ac18e750	6	struct netns_frags {
e5a2bb84	7	int nqueues;
3140c25c	8	struct list_head lru_list;
3ef0eb0d	9	spinlock_t lru_lock;
b2fd5321	10
6d7b857d JDB	11	/* The percpu_counter "mem" need to be cacheline aligned.
6d7b857d JDB	12	* mem.count must not share cacheline with other writers
cd39a789	13	*/
6d7b857d JDB	14	struct percpu_counter mem ____cacheline_aligned_in_smp;
6d7b857d JDB	15
b2fd5321 PE	16	/* sysctls */
b2fd5321 PE	17	int timeout;
e31e0bdc PE	18	int high_thresh;
e31e0bdc PE	19	int low_thresh;
ac18e750 PE	20	};
ac18e750 PE	21
5ab11c98	22	struct inet_frag_queue {
5ab11c98	23	spinlock_t lock;
5ab11c98	24	struct timer_list timer; /* when will this queue expire? */
6e34a8b3 JDB	25	struct list_head lru_list; /* lru list member */
	26	struct hlist_node list;
	27	atomic_t refcnt;
5ab11c98	28	struct sk_buff fragments; / list of received fragments */
d6bebca9	29	struct sk_buff *fragments_tail;
5ab11c98 PE	30	ktime_t stamp;
	31	int len; /* total length of orig datagram */
	32	int meat;
	33	__u8 last_in; /* first/last segment arrived? */
	34
bc578a54 JP	35	#define INET_FRAG_COMPLETE 4
	36	#define INET_FRAG_FIRST_IN 2
	37	#define INET_FRAG_LAST_IN 1
5f2d04f1 PM	38
5f2d04f1 PM	39	u16 max_size;
6e34a8b3 JDB	40
6e34a8b3 JDB	41	struct netns_frags *net;
5ab11c98 PE	42	};
5ab11c98 PE	43
a4c4009f	44	#define INETFRAGS_HASHSZ 1024
7eb95156	45
5a3da1fe HFS	46	/* averaged:
	47	* max_depth = default ipfrag_high_thresh / INETFRAGS_HASHSZ /
	48	* rounded up (SKB_TRUELEN(0) + sizeof(struct ipq or
	49	* struct frag_queue))
	50	*/
	51	#define INETFRAGS_MAXDEPTH 128
	52
19952cc4 JDB	53	struct inet_frag_bucket {
	54	struct hlist_head chain;
	55	spinlock_t chain_lock;
	56	};
	57
7eb95156	58	struct inet_frags {
19952cc4	59	struct inet_frag_bucket hash[INETFRAGS_HASHSZ];
5f8e1e8b JDB	60	/* This rwlock is a global lock (seperate per IPv4, IPv6 and
5f8e1e8b JDB	61	* netfilter). Important to keep this on a seperate cacheline.
19952cc4	62	* Its primarily a rebuild protection rwlock.
5f8e1e8b JDB	63	*/
5f8e1e8b JDB	64	rwlock_t lock ____cacheline_aligned_in_smp;
3b4bc4a2	65	int secret_interval;
7eb95156	66	struct timer_list secret_timer;
7088ad74 HFS	67
	68	/* The first call to hashfn is responsible to initialize
	69	* rnd. This is best done with net_get_random_once.
	70	*/
5f8e1e8b JDB	71	u32 rnd;
5f8e1e8b JDB	72	int qsize;
321a3a99 PE	73
321a3a99 PE	74	unsigned int (hashfn)(struct inet_frag_queue );
5f8e1e8b	75	bool (match)(struct inet_frag_queue q, void *arg);
c6fda282 PE	76	void (constructor)(struct inet_frag_queue q,
c6fda282 PE	77	void *arg);
1e4b8287 PE	78	void (destructor)(struct inet_frag_queue );
1e4b8287 PE	79	void (skb_free)(struct sk_buff );
e521db9d	80	void (*frag_expire)(unsigned long data);
7eb95156 PE	81	};
	82
	83	void inet_frags_init(struct inet_frags *);
	84	void inet_frags_fini(struct inet_frags *);
	85
e5a2bb84	86	void inet_frags_init_net(struct netns_frags *nf);
81566e83	87	void inet_frags_exit_net(struct netns_frags nf, struct inet_frags f);
e5a2bb84	88
277e650d	89	void inet_frag_kill(struct inet_frag_queue q, struct inet_frags f);
1e4b8287 PE	90	void inet_frag_destroy(struct inet_frag_queue *q,
1e4b8287 PE	91	struct inet_frags f, int work);
6b102865	92	int inet_frag_evictor(struct netns_frags nf, struct inet_frags f, bool force);
ac18e750	93	struct inet_frag_queue inet_frag_find(struct netns_frags nf,
56bca31f HE	94	struct inet_frags f, void key, unsigned int hash)
56bca31f HE	95	__releases(&f->lock);
5a3da1fe HFS	96	void inet_frag_maybe_warn_overflow(struct inet_frag_queue *q,
5a3da1fe HFS	97	const char *prefix);
277e650d	98
762cc408 PE	99	static inline void inet_frag_put(struct inet_frag_queue q, struct inet_frags f)
	100	{
	101	if (atomic_dec_and_test(&q->refcnt))
	102	inet_frag_destroy(q, f, NULL);
	103	}
	104
d433673e JDB	105	/* Memory Tracking Functions. */
d433673e JDB	106
6d7b857d JDB	107	/* The default percpu_counter batch size is not big enough to scale to
	108	* fragmentation mem acct sizes.
	109	* The mem size of a 64K fragment is approx:
	110	* (44 fragments * 2944 truesize) + frag_queue struct(200) = 129736 bytes
	111	*/
	112	static unsigned int frag_percpu_counter_batch = 130000;
	113
d433673e JDB	114	static inline int frag_mem_limit(struct netns_frags *nf)
d433673e JDB	115	{
6d7b857d	116	return percpu_counter_read(&nf->mem);
d433673e JDB	117	}
	118
	119	static inline void sub_frag_mem_limit(struct inet_frag_queue *q, int i)
	120	{
6d7b857d	121	__percpu_counter_add(&q->net->mem, -i, frag_percpu_counter_batch);
d433673e JDB	122	}
	123
	124	static inline void add_frag_mem_limit(struct inet_frag_queue *q, int i)
	125	{
6d7b857d	126	__percpu_counter_add(&q->net->mem, i, frag_percpu_counter_batch);
d433673e JDB	127	}
	128
	129	static inline void init_frag_mem_limit(struct netns_frags *nf)
	130	{
6d7b857d	131	percpu_counter_init(&nf->mem, 0);
d433673e JDB	132	}
	133
	134	static inline int sum_frag_mem_limit(struct netns_frags *nf)
	135	{
4cfb0485 ED	136	int res;
	137
	138	local_bh_disable();
	139	res = percpu_counter_sum_positive(&nf->mem);
	140	local_bh_enable();
	141
	142	return res;
d433673e JDB	143	}
d433673e JDB	144
3ef0eb0d JDB	145	static inline void inet_frag_lru_move(struct inet_frag_queue *q)
	146	{
	147	spin_lock(&q->net->lru_lock);
b56141ab KK	148	if (!list_empty(&q->lru_list))
b56141ab KK	149	list_move_tail(&q->lru_list, &q->net->lru_list);
3ef0eb0d JDB	150	spin_unlock(&q->net->lru_lock);
	151	}
	152
	153	static inline void inet_frag_lru_del(struct inet_frag_queue *q)
	154	{
	155	spin_lock(&q->net->lru_lock);
b56141ab	156	list_del_init(&q->lru_list);
1b5ab0de	157	q->net->nqueues--;
3ef0eb0d JDB	158	spin_unlock(&q->net->lru_lock);
	159	}
	160
	161	static inline void inet_frag_lru_add(struct netns_frags *nf,
	162	struct inet_frag_queue *q)
	163	{
	164	spin_lock(&nf->lru_lock);
	165	list_add_tail(&q->lru_list, &nf->lru_list);
1b5ab0de	166	q->net->nqueues++;
3ef0eb0d JDB	167	spin_unlock(&nf->lru_lock);
3ef0eb0d JDB	168	}
be991971 HFS	169
	170	/* RFC 3168 support :
	171	* We want to check ECN values of all fragments, do detect invalid combinations.
	172	* In ipq->ecn, we store the OR value of each ip4_frag_ecn() fragment value.
	173	*/
	174	#define IPFRAG_ECN_NOT_ECT 0x01 /* one frag had ECN_NOT_ECT */
	175	#define IPFRAG_ECN_ECT_1 0x02 /* one frag had ECN_ECT_1 */
	176	#define IPFRAG_ECN_ECT_0 0x04 /* one frag had ECN_ECT_0 */
	177	#define IPFRAG_ECN_CE 0x08 /* one frag had ECN_CE */
	178
	179	extern const u8 ip_frag_ecn_table[16];
	180
5ab11c98	181	#endif