[deliverable/linux.git] / include / net / request_sock.h

/*
 * NET		Generic infrastructure for Network protocols.
 *
 *		Definitions for request_sock 
 *
 * Authors:	Arnaldo Carvalho de Melo <acme@conectiva.com.br>
 *
 * 		From code originally in include/net/tcp.h
 *
 *		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.
 */
#ifndef _REQUEST_SOCK_H
#define _REQUEST_SOCK_H

#include <linux/slab.h>
#include <linux/spinlock.h>
#include <linux/types.h>

#include <net/sock.h>

struct request_sock;
struct sk_buff;
struct dst_entry;
struct proto;

struct request_sock_ops {
	int		family;
	kmem_cache_t	*slab;
	int		obj_size;
	int		(*rtx_syn_ack)(struct sock *sk,
				       struct request_sock *req,
				       struct dst_entry *dst);
	void		(*send_ack)(struct sk_buff *skb,
				    struct request_sock *req);
	void		(*send_reset)(struct sk_buff *skb);
	void		(*destructor)(struct request_sock *req);
};

/* struct request_sock - mini sock to represent a connection request
 */
struct request_sock {
	struct request_sock		*dl_next; /* Must be first member! */
	u16				mss;
	u8				retrans;
	u8				__pad;
	/* The following two fields can be easily recomputed I think -AK */
	u32				window_clamp; /* window clamp at creation time */
	u32				rcv_wnd;	  /* rcv_wnd offered first time */
	u32				ts_recent;
	unsigned long			expires;
	struct request_sock_ops		*rsk_ops;
	struct sock			*sk;
};

static inline struct request_sock *reqsk_alloc(struct request_sock_ops *ops)
{
	struct request_sock *req = kmem_cache_alloc(ops->slab, SLAB_ATOMIC);

	if (req != NULL)
		req->rsk_ops = ops;

	return req;
}

static inline void __reqsk_free(struct request_sock *req)
{
	kmem_cache_free(req->rsk_ops->slab, req);
}

static inline void reqsk_free(struct request_sock *req)
{
	req->rsk_ops->destructor(req);
	__reqsk_free(req);
}

extern int sysctl_max_syn_backlog;

/** struct listen_sock - listen state
 *
 * @max_qlen_log - log_2 of maximal queued SYNs/REQUESTs
 */
struct listen_sock {
	u8			max_qlen_log;
	/* 3 bytes hole, try to use */
	int			qlen;
	int			qlen_young;
	int			clock_hand;
	u32			hash_rnd;
	struct request_sock	*syn_table[0];
};

/** struct request_sock_queue - queue of request_socks
 *
 * @rskq_accept_head - FIFO head of established children
 * @rskq_accept_tail - FIFO tail of established children
 * @syn_wait_lock - serializer
 *
 * %syn_wait_lock is necessary only to avoid proc interface having to grab the main
 * lock sock while browsing the listening hash (otherwise it's deadlock prone).
 *
 * This lock is acquired in read mode only from listening_get_next() seq_file
 * op and it's acquired in write mode _only_ from code that is actively
 * changing rskq_accept_head. All readers that are holding the master sock lock
 * don't need to grab this lock in read mode too as rskq_accept_head. writes
 * are always protected from the main sock lock.
 */
struct request_sock_queue {
	struct request_sock	*rskq_accept_head;
	struct request_sock	*rskq_accept_tail;
	rwlock_t		syn_wait_lock;
	struct listen_sock	*listen_opt;
};

extern int reqsk_queue_alloc(struct request_sock_queue *queue,
			     const int nr_table_entries);

static inline struct listen_sock *reqsk_queue_yank_listen_sk(struct request_sock_queue *queue)
{
	struct listen_sock *lopt;

	write_lock_bh(&queue->syn_wait_lock);
	lopt = queue->listen_opt;
	queue->listen_opt = NULL;
	write_unlock_bh(&queue->syn_wait_lock);

	return lopt;
}

static inline void reqsk_queue_destroy(struct request_sock_queue *queue)
{
	kfree(reqsk_queue_yank_listen_sk(queue));
}

static inline struct request_sock *
	reqsk_queue_yank_acceptq(struct request_sock_queue *queue)
{
	struct request_sock *req = queue->rskq_accept_head;

	queue->rskq_accept_head = queue->rskq_accept_head = NULL;
	return req;
}

static inline int reqsk_queue_empty(struct request_sock_queue *queue)
{
	return queue->rskq_accept_head == NULL;
}

static inline void reqsk_queue_unlink(struct request_sock_queue *queue,
				      struct request_sock *req,
				      struct request_sock **prev_req)
{
	write_lock(&queue->syn_wait_lock);
	*prev_req = req->dl_next;
	write_unlock(&queue->syn_wait_lock);
}

static inline void reqsk_queue_add(struct request_sock_queue *queue,
				   struct request_sock *req,
				   struct sock *parent,
				   struct sock *child)
{
	req->sk = child;
	sk_acceptq_added(parent);

	if (queue->rskq_accept_head == NULL)
		queue->rskq_accept_head = req;
	else
		queue->rskq_accept_tail->dl_next = req;

	queue->rskq_accept_tail = req;
	req->dl_next = NULL;
}

static inline struct request_sock *reqsk_queue_remove(struct request_sock_queue *queue)
{
	struct request_sock *req = queue->rskq_accept_head;

	BUG_TRAP(req != NULL);

	queue->rskq_accept_head = req->dl_next;
	if (queue->rskq_accept_head == NULL)
		queue->rskq_accept_tail = NULL;

	return req;
}

static inline struct sock *reqsk_queue_get_child(struct request_sock_queue *queue,
						 struct sock *parent)
{
	struct request_sock *req = reqsk_queue_remove(queue);
	struct sock *child = req->sk;

	BUG_TRAP(child != NULL);

	sk_acceptq_removed(parent);
	__reqsk_free(req);
	return child;
}

static inline int reqsk_queue_removed(struct request_sock_queue *queue,
				      struct request_sock *req)
{
	struct listen_sock *lopt = queue->listen_opt;

	if (req->retrans == 0)
		--lopt->qlen_young;

	return --lopt->qlen;
}

static inline int reqsk_queue_added(struct request_sock_queue *queue)
{
	struct listen_sock *lopt = queue->listen_opt;
	const int prev_qlen = lopt->qlen;

	lopt->qlen_young++;
	lopt->qlen++;
	return prev_qlen;
}

static inline int reqsk_queue_len(struct request_sock_queue *queue)
{
	return queue->listen_opt != NULL ? queue->listen_opt->qlen : 0;
}

static inline int reqsk_queue_len_young(struct request_sock_queue *queue)
{
	return queue->listen_opt->qlen_young;
}

static inline int reqsk_queue_is_full(struct request_sock_queue *queue)
{
	return queue->listen_opt->qlen >> queue->listen_opt->max_qlen_log;
}

static inline void reqsk_queue_hash_req(struct request_sock_queue *queue,
					u32 hash, struct request_sock *req,
					unsigned timeout)
{
	struct listen_sock *lopt = queue->listen_opt;

	req->expires = jiffies + timeout;
	req->retrans = 0;
	req->sk = NULL;
	req->dl_next = lopt->syn_table[hash];

	write_lock(&queue->syn_wait_lock);
	lopt->syn_table[hash] = req;
	write_unlock(&queue->syn_wait_lock);
}

#endif /* _REQUEST_SOCK_H */
Commit	Line	Data
2e6599cb ACM	1	/*
	2	* NET Generic infrastructure for Network protocols.
	3	*
	4	* Definitions for request_sock
	5	*
	6	* Authors: Arnaldo Carvalho de Melo <acme@conectiva.com.br>
	7	*
	8	* From code originally in include/net/tcp.h
	9	*
	10	* This program is free software; you can redistribute it and/or
	11	* modify it under the terms of the GNU General Public License
	12	* as published by the Free Software Foundation; either version
	13	* 2 of the License, or (at your option) any later version.
	14	*/
	15	#ifndef _REQUEST_SOCK_H
	16	#define _REQUEST_SOCK_H
	17
	18	#include <linux/slab.h>
0e87506f	19	#include <linux/spinlock.h>
2e6599cb	20	#include <linux/types.h>
0e87506f	21
2e6599cb ACM	22	#include <net/sock.h>
2e6599cb ACM	23
60236fdd	24	struct request_sock;
2e6599cb ACM	25	struct sk_buff;
	26	struct dst_entry;
	27	struct proto;
	28
60236fdd	29	struct request_sock_ops {
2e6599cb ACM	30	int family;
	31	kmem_cache_t *slab;
	32	int obj_size;
	33	int (rtx_syn_ack)(struct sock sk,
60236fdd	34	struct request_sock *req,
2e6599cb ACM	35	struct dst_entry *dst);
2e6599cb ACM	36	void (send_ack)(struct sk_buff skb,
60236fdd	37	struct request_sock *req);
2e6599cb	38	void (send_reset)(struct sk_buff skb);
60236fdd	39	void (destructor)(struct request_sock req);
2e6599cb ACM	40	};
2e6599cb ACM	41
60236fdd	42	/* struct request_sock - mini sock to represent a connection request
2e6599cb	43	*/
60236fdd ACM	44	struct request_sock {
60236fdd ACM	45	struct request_sock dl_next; / Must be first member! */
2e6599cb ACM	46	u16 mss;
	47	u8 retrans;
	48	u8 __pad;
	49	/* The following two fields can be easily recomputed I think -AK */
	50	u32 window_clamp; /* window clamp at creation time */
	51	u32 rcv_wnd; /* rcv_wnd offered first time */
	52	u32 ts_recent;
	53	unsigned long expires;
60236fdd	54	struct request_sock_ops *rsk_ops;
2e6599cb ACM	55	struct sock *sk;
	56	};
	57
60236fdd	58	static inline struct request_sock reqsk_alloc(struct request_sock_ops ops)
2e6599cb	59	{
60236fdd	60	struct request_sock *req = kmem_cache_alloc(ops->slab, SLAB_ATOMIC);
2e6599cb ACM	61
2e6599cb ACM	62	if (req != NULL)
60236fdd	63	req->rsk_ops = ops;
2e6599cb ACM	64
	65	return req;
	66	}
	67
60236fdd	68	static inline void __reqsk_free(struct request_sock *req)
2e6599cb	69	{
60236fdd	70	kmem_cache_free(req->rsk_ops->slab, req);
2e6599cb ACM	71	}
2e6599cb ACM	72
60236fdd	73	static inline void reqsk_free(struct request_sock *req)
2e6599cb	74	{
60236fdd ACM	75	req->rsk_ops->destructor(req);
60236fdd ACM	76	__reqsk_free(req);
2e6599cb ACM	77	}
2e6599cb ACM	78
0e87506f ACM	79	extern int sysctl_max_syn_backlog;
0e87506f ACM	80
2ad69c55	81	/** struct listen_sock - listen state
0e87506f ACM	82	*
	83	* @max_qlen_log - log_2 of maximal queued SYNs/REQUESTs
	84	*/
2ad69c55	85	struct listen_sock {
0e87506f ACM	86	u8 max_qlen_log;
	87	/* 3 bytes hole, try to use */
	88	int qlen;
	89	int qlen_young;
	90	int clock_hand;
	91	u32 hash_rnd;
	92	struct request_sock *syn_table[0];
	93	};
	94
	95	/** struct request_sock_queue - queue of request_socks
	96	*
	97	* @rskq_accept_head - FIFO head of established children
	98	* @rskq_accept_tail - FIFO tail of established children
	99	* @syn_wait_lock - serializer
	100	*
	101	* %syn_wait_lock is necessary only to avoid proc interface having to grab the main
	102	* lock sock while browsing the listening hash (otherwise it's deadlock prone).
	103	*
	104	* This lock is acquired in read mode only from listening_get_next() seq_file
	105	* op and it's acquired in write mode _only_ from code that is actively
	106	* changing rskq_accept_head. All readers that are holding the master sock lock
	107	* don't need to grab this lock in read mode too as rskq_accept_head. writes
	108	* are always protected from the main sock lock.
	109	*/
	110	struct request_sock_queue {
	111	struct request_sock *rskq_accept_head;
	112	struct request_sock *rskq_accept_tail;
	113	rwlock_t syn_wait_lock;
2ad69c55	114	struct listen_sock *listen_opt;
0e87506f ACM	115	};
	116
	117	extern int reqsk_queue_alloc(struct request_sock_queue *queue,
	118	const int nr_table_entries);
	119
2ad69c55	120	static inline struct listen_sock reqsk_queue_yank_listen_sk(struct request_sock_queue queue)
0e87506f	121	{
2ad69c55	122	struct listen_sock *lopt;
0e87506f ACM	123
	124	write_lock_bh(&queue->syn_wait_lock);
	125	lopt = queue->listen_opt;
	126	queue->listen_opt = NULL;
	127	write_unlock_bh(&queue->syn_wait_lock);
	128
	129	return lopt;
	130	}
	131
	132	static inline void reqsk_queue_destroy(struct request_sock_queue *queue)
	133	{
	134	kfree(reqsk_queue_yank_listen_sk(queue));
	135	}
	136
	137	static inline struct request_sock *
	138	reqsk_queue_yank_acceptq(struct request_sock_queue *queue)
	139	{
	140	struct request_sock *req = queue->rskq_accept_head;
	141
	142	queue->rskq_accept_head = queue->rskq_accept_head = NULL;
	143	return req;
	144	}
	145
	146	static inline int reqsk_queue_empty(struct request_sock_queue *queue)
	147	{
	148	return queue->rskq_accept_head == NULL;
	149	}
	150
	151	static inline void reqsk_queue_unlink(struct request_sock_queue *queue,
	152	struct request_sock *req,
	153	struct request_sock **prev_req)
	154	{
	155	write_lock(&queue->syn_wait_lock);
	156	*prev_req = req->dl_next;
	157	write_unlock(&queue->syn_wait_lock);
	158	}
	159
	160	static inline void reqsk_queue_add(struct request_sock_queue *queue,
	161	struct request_sock *req,
	162	struct sock *parent,
	163	struct sock *child)
	164	{
	165	req->sk = child;
	166	sk_acceptq_added(parent);
	167
	168	if (queue->rskq_accept_head == NULL)
	169	queue->rskq_accept_head = req;
	170	else
	171	queue->rskq_accept_tail->dl_next = req;
	172
	173	queue->rskq_accept_tail = req;
	174	req->dl_next = NULL;
	175	}
	176
	177	static inline struct request_sock reqsk_queue_remove(struct request_sock_queue queue)
	178	{
	179	struct request_sock *req = queue->rskq_accept_head;
	180
	181	BUG_TRAP(req != NULL);
	182
	183	queue->rskq_accept_head = req->dl_next;
	184	if (queue->rskq_accept_head == NULL)
	185	queue->rskq_accept_tail = NULL;
	186
187	return req;
188	}
189
190	static inline struct sock reqsk_queue_get_child(struct request_sock_queue queue,
191	struct sock *parent)
192	{
193	struct request_sock *req = reqsk_queue_remove(queue);
194	struct sock *child = req->sk;
195
196	BUG_TRAP(child != NULL);
197
198	sk_acceptq_removed(parent);
199	__reqsk_free(req);
200	return child;
201	}
202
203	static inline int reqsk_queue_removed(struct request_sock_queue *queue,
204	struct request_sock *req)
205	{
2ad69c55	206	struct listen_sock *lopt = queue->listen_opt;
0e87506f ACM	207
	208	if (req->retrans == 0)
	209	--lopt->qlen_young;
	210
	211	return --lopt->qlen;
	212	}
	213
	214	static inline int reqsk_queue_added(struct request_sock_queue *queue)
	215	{
2ad69c55	216	struct listen_sock *lopt = queue->listen_opt;
0e87506f ACM	217	const int prev_qlen = lopt->qlen;
	218
	219	lopt->qlen_young++;
	220	lopt->qlen++;
	221	return prev_qlen;
	222	}
	223
	224	static inline int reqsk_queue_len(struct request_sock_queue *queue)
	225	{
	226	return queue->listen_opt != NULL ? queue->listen_opt->qlen : 0;
	227	}
	228
	229	static inline int reqsk_queue_len_young(struct request_sock_queue *queue)
	230	{
	231	return queue->listen_opt->qlen_young;
	232	}
	233
	234	static inline int reqsk_queue_is_full(struct request_sock_queue *queue)
	235	{
	236	return queue->listen_opt->qlen >> queue->listen_opt->max_qlen_log;
	237	}
	238
	239	static inline void reqsk_queue_hash_req(struct request_sock_queue *queue,
	240	u32 hash, struct request_sock *req,
	241	unsigned timeout)
	242	{
2ad69c55	243	struct listen_sock *lopt = queue->listen_opt;
0e87506f ACM	244
	245	req->expires = jiffies + timeout;
	246	req->retrans = 0;
	247	req->sk = NULL;
	248	req->dl_next = lopt->syn_table[hash];
	249
	250	write_lock(&queue->syn_wait_lock);
	251	lopt->syn_table[hash] = req;
	252	write_unlock(&queue->syn_wait_lock);
	253	}
	254
2e6599cb	255	#endif /* _REQUEST_SOCK_H */