[linux-2.6-block.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 <linux/bug.h>

#include <net/sock.h>

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

struct request_sock_ops {
	int		family;
	int		obj_size;
	struct kmem_cache	*slab;
	char		*slab_name;
	int		(*rtx_syn_ack)(struct sock *sk,
				       struct request_sock *req);
	void		(*send_ack)(struct sock *sk, struct sk_buff *skb,
				    struct request_sock *req);
	void		(*send_reset)(struct sock *sk,
				      struct sk_buff *skb);
	void		(*destructor)(struct request_sock *req);
	void		(*syn_ack_timeout)(struct sock *sk,
					   struct request_sock *req);
};

int inet_rtx_syn_ack(struct sock *parent, struct request_sock *req);

/* struct request_sock - mini sock to represent a connection request
 */
struct request_sock {
	struct sock_common		__req_common;
#define rsk_refcnt			__req_common.skc_refcnt

	struct request_sock		*dl_next;
	u16				mss;
	u8				num_retrans; /* number of retransmits */
	u8				cookie_ts:1; /* syncookie: encode tcpopts in timestamp */
	u8				num_timeout:7; /* number of timeouts */
	/* 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;
	const struct request_sock_ops	*rsk_ops;
	struct sock			*sk;
	u32				secid;
	u32				peer_secid;
};

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

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

	return req;
}

static inline struct request_sock *inet_reqsk(struct sock *sk)
{
	return (struct request_sock *)sk;
}

static inline void reqsk_free(struct request_sock *req)
{
	/* temporary debugging */
	WARN_ON_ONCE(atomic_read(&req->rsk_refcnt) != 0);

	req->rsk_ops->destructor(req);
	kmem_cache_free(req->rsk_ops->slab, req);
}

static inline void reqsk_put(struct request_sock *req)
{
	/* temporary debugging, until req sock are put into ehash table */
	WARN_ON_ONCE(atomic_read(&req->rsk_refcnt) != 1);

	if (atomic_dec_and_test(&req->rsk_refcnt))
		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;
	u8			synflood_warned;
	/* 2 bytes hole, try to use */
	int			qlen;
	int			qlen_young;
	int			clock_hand;
	u32			hash_rnd;
	u32			nr_table_entries;
	struct request_sock	*syn_table[0];
};

/*
 * For a TCP Fast Open listener -
 *	lock - protects the access to all the reqsk, which is co-owned by
 *		the listener and the child socket.
 *	qlen - pending TFO requests (still in TCP_SYN_RECV).
 *	max_qlen - max TFO reqs allowed before TFO is disabled.
 *
 *	XXX (TFO) - ideally these fields can be made as part of "listen_sock"
 *	structure above. But there is some implementation difficulty due to
 *	listen_sock being part of request_sock_queue hence will be freed when
 *	a listener is stopped. But TFO related fields may continue to be
 *	accessed even after a listener is closed, until its sk_refcnt drops
 *	to 0 implying no more outstanding TFO reqs. One solution is to keep
 *	listen_opt around until	sk_refcnt drops to 0. But there is some other
 *	complexity that needs to be resolved. E.g., a listener can be disabled
 *	temporarily through shutdown()->tcp_disconnect(), and re-enabled later.
 */
struct fastopen_queue {
	struct request_sock	*rskq_rst_head; /* Keep track of past TFO */
	struct request_sock	*rskq_rst_tail; /* requests that caused RST.
						 * This is part of the defense
						 * against spoofing attack.
						 */
	spinlock_t	lock;
	int		qlen;		/* # of pending (TCP_SYN_RECV) reqs */
	int		max_qlen;	/* != 0 iff TFO is currently enabled */
};

/** struct request_sock_queue - queue of request_socks
 *
 * @rskq_accept_head - FIFO head of established children
 * @rskq_accept_tail - FIFO tail of established children
 * @rskq_defer_accept - User waits for some data after accept()
 * @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;
	u8			rskq_defer_accept;
	/* 3 bytes hole, try to pack */
	struct listen_sock	*listen_opt;
	struct fastopen_queue	*fastopenq; /* This is non-NULL iff TFO has been
					     * enabled on this listener. Check
					     * max_qlen != 0 in fastopen_queue
					     * to determine if TFO is enabled
					     * right at this moment.
					     */
};

int reqsk_queue_alloc(struct request_sock_queue *queue,
		      unsigned int nr_table_entries);

void __reqsk_queue_destroy(struct request_sock_queue *queue);
void reqsk_queue_destroy(struct request_sock_queue *queue);
void reqsk_fastopen_remove(struct sock *sk, struct request_sock *req,
			   bool reset);

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 = 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;

	WARN_ON(req == NULL);

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

	return req;
}

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

	if (req->num_timeout == 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(const struct request_sock_queue *queue)
{
	return queue->listen_opt != NULL ? queue->listen_opt->qlen : 0;
}

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

static inline int reqsk_queue_is_full(const 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 long timeout)
{
	struct listen_sock *lopt = queue->listen_opt;

	req->expires = jiffies + timeout;
	req->num_retrans = 0;
	req->num_timeout = 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>
547b792c	21	#include <linux/bug.h>
0e87506f	22
2e6599cb ACM	23	#include <net/sock.h>
2e6599cb ACM	24
60236fdd	25	struct request_sock;
2e6599cb ACM	26	struct sk_buff;
	27	struct dst_entry;
	28	struct proto;
	29
60236fdd	30	struct request_sock_ops {
2e6599cb	31	int family;
2e6599cb	32	int obj_size;
e18b890b	33	struct kmem_cache *slab;
7e56b5d6	34	char *slab_name;
2e6599cb	35	int (rtx_syn_ack)(struct sock sk,
1a2c6181	36	struct request_sock *req);
6edafaaf	37	void (send_ack)(struct sock sk, struct sk_buff *skb,
60236fdd	38	struct request_sock *req);
cfb6eeb4 YH	39	void (send_reset)(struct sock sk,
cfb6eeb4 YH	40	struct sk_buff *skb);
60236fdd	41	void (destructor)(struct request_sock req);
72659ecc OP	42	void (syn_ack_timeout)(struct sock sk,
72659ecc OP	43	struct request_sock *req);
2e6599cb ACM	44	};
2e6599cb ACM	45
c0f4502a	46	int inet_rtx_syn_ack(struct sock parent, struct request_sock req);
e6c022a4	47
60236fdd	48	/* struct request_sock - mini sock to represent a connection request
2e6599cb	49	*/
60236fdd	50	struct request_sock {
634fb979	51	struct sock_common __req_common;
1e2e0117 ED	52	#define rsk_refcnt __req_common.skc_refcnt
1e2e0117 ED	53
3fb62c5d	54	struct request_sock *dl_next;
2e6599cb	55	u16 mss;
e6c022a4 ED	56	u8 num_retrans; /* number of retransmits */
	57	u8 cookie_ts:1; /* syncookie: encode tcpopts in timestamp */
	58	u8 num_timeout:7; /* number of timeouts */
2e6599cb ACM	59	/* The following two fields can be easily recomputed I think -AK */
	60	u32 window_clamp; /* window clamp at creation time */
	61	u32 rcv_wnd; /* rcv_wnd offered first time */
	62	u32 ts_recent;
	63	unsigned long expires;
72a3effa	64	const struct request_sock_ops *rsk_ops;
2e6599cb	65	struct sock *sk;
4237c75c	66	u32 secid;
6b877699	67	u32 peer_secid;
2e6599cb ACM	68	};
2e6599cb ACM	69
72a3effa	70	static inline struct request_sock reqsk_alloc(const struct request_sock_ops ops)
2e6599cb	71	{
54e6ecb2	72	struct request_sock *req = kmem_cache_alloc(ops->slab, GFP_ATOMIC);
2e6599cb ACM	73
2e6599cb ACM	74	if (req != NULL)
60236fdd	75	req->rsk_ops = ops;
2e6599cb ACM	76
	77	return req;
	78	}
	79
41b822c5 ED	80	static inline struct request_sock inet_reqsk(struct sock sk)
	81	{
	82	return (struct request_sock *)sk;
	83	}
	84
60236fdd	85	static inline void reqsk_free(struct request_sock *req)
2e6599cb	86	{
13854e5a ED	87	/* temporary debugging */
	88	WARN_ON_ONCE(atomic_read(&req->rsk_refcnt) != 0);
	89
60236fdd	90	req->rsk_ops->destructor(req);
13854e5a	91	kmem_cache_free(req->rsk_ops->slab, req);
2e6599cb ACM	92	}
2e6599cb ACM	93
1e2e0117 ED	94	static inline void reqsk_put(struct request_sock *req)
1e2e0117 ED	95	{
13854e5a ED	96	/* temporary debugging, until req sock are put into ehash table */
	97	WARN_ON_ONCE(atomic_read(&req->rsk_refcnt) != 1);
	98
1e2e0117 ED	99	if (atomic_dec_and_test(&req->rsk_refcnt))
	100	reqsk_free(req);
	101	}
	102
0e87506f ACM	103	extern int sysctl_max_syn_backlog;
0e87506f ACM	104
2ad69c55	105	/** struct listen_sock - listen state
0e87506f ACM	106	*
	107	* @max_qlen_log - log_2 of maximal queued SYNs/REQUESTs
	108	*/
2ad69c55	109	struct listen_sock {
0e87506f	110	u8 max_qlen_log;
946cedcc ED	111	u8 synflood_warned;
946cedcc ED	112	/* 2 bytes hole, try to use */
0e87506f ACM	113	int qlen;
	114	int qlen_young;
	115	int clock_hand;
	116	u32 hash_rnd;
83e3609e	117	u32 nr_table_entries;
0e87506f ACM	118	struct request_sock *syn_table[0];
	119	};
	120
10467163 JC	121	/*
	122	* For a TCP Fast Open listener -
	123	* lock - protects the access to all the reqsk, which is co-owned by
	124	* the listener and the child socket.
	125	* qlen - pending TFO requests (still in TCP_SYN_RECV).
	126	* max_qlen - max TFO reqs allowed before TFO is disabled.
	127	*
	128	* XXX (TFO) - ideally these fields can be made as part of "listen_sock"
	129	* structure above. But there is some implementation difficulty due to
	130	* listen_sock being part of request_sock_queue hence will be freed when
	131	* a listener is stopped. But TFO related fields may continue to be
	132	* accessed even after a listener is closed, until its sk_refcnt drops
	133	* to 0 implying no more outstanding TFO reqs. One solution is to keep
	134	* listen_opt around until sk_refcnt drops to 0. But there is some other
	135	* complexity that needs to be resolved. E.g., a listener can be disabled
	136	* temporarily through shutdown()->tcp_disconnect(), and re-enabled later.
	137	*/
	138	struct fastopen_queue {
	139	struct request_sock rskq_rst_head; / Keep track of past TFO */
	140	struct request_sock rskq_rst_tail; / requests that caused RST.
	141	* This is part of the defense
	142	* against spoofing attack.
	143	*/
	144	spinlock_t lock;
	145	int qlen; /* # of pending (TCP_SYN_RECV) reqs */
	146	int max_qlen; /* != 0 iff TFO is currently enabled */
	147	};
	148
0e87506f ACM	149	/** struct request_sock_queue - queue of request_socks
	150	*
	151	* @rskq_accept_head - FIFO head of established children
	152	* @rskq_accept_tail - FIFO tail of established children
295f7324	153	* @rskq_defer_accept - User waits for some data after accept()
0e87506f ACM	154	* @syn_wait_lock - serializer
	155	*
	156	* %syn_wait_lock is necessary only to avoid proc interface having to grab the main
	157	* lock sock while browsing the listening hash (otherwise it's deadlock prone).
	158	*
	159	* This lock is acquired in read mode only from listening_get_next() seq_file
	160	* op and it's acquired in write mode _only_ from code that is actively
	161	* changing rskq_accept_head. All readers that are holding the master sock lock
	162	* don't need to grab this lock in read mode too as rskq_accept_head. writes
	163	* are always protected from the main sock lock.
	164	*/
	165	struct request_sock_queue {
	166	struct request_sock *rskq_accept_head;
	167	struct request_sock *rskq_accept_tail;
	168	rwlock_t syn_wait_lock;
ec0a1966 DM	169	u8 rskq_defer_accept;
ec0a1966 DM	170	/* 3 bytes hole, try to pack */
2ad69c55	171	struct listen_sock *listen_opt;
10467163 JC	172	struct fastopen_queue fastopenq; / This is non-NULL iff TFO has been
	173	* enabled on this listener. Check
	174	* max_qlen != 0 in fastopen_queue
	175	* to determine if TFO is enabled
	176	* right at this moment.
	177	*/
0e87506f ACM	178	};
0e87506f ACM	179
c0f4502a JP	180	int reqsk_queue_alloc(struct request_sock_queue *queue,
c0f4502a JP	181	unsigned int nr_table_entries);
0e87506f	182
c0f4502a JP	183	void __reqsk_queue_destroy(struct request_sock_queue *queue);
	184	void reqsk_queue_destroy(struct request_sock_queue *queue);
	185	void reqsk_fastopen_remove(struct sock sk, struct request_sock req,
	186	bool reset);
83e3609e	187
0e87506f ACM	188	static inline struct request_sock *
	189	reqsk_queue_yank_acceptq(struct request_sock_queue *queue)
	190	{
	191	struct request_sock *req = queue->rskq_accept_head;
	192
3eb4801d	193	queue->rskq_accept_head = NULL;
0e87506f ACM	194	return req;
	195	}
	196
	197	static inline int reqsk_queue_empty(struct request_sock_queue *queue)
	198	{
	199	return queue->rskq_accept_head == NULL;
	200	}
	201
	202	static inline void reqsk_queue_unlink(struct request_sock_queue *queue,
	203	struct request_sock *req,
	204	struct request_sock **prev_req)
	205	{
	206	write_lock(&queue->syn_wait_lock);
	207	*prev_req = req->dl_next;
	208	write_unlock(&queue->syn_wait_lock);
	209	}
	210
	211	static inline void reqsk_queue_add(struct request_sock_queue *queue,
	212	struct request_sock *req,
	213	struct sock *parent,
	214	struct sock *child)
	215	{
	216	req->sk = child;
	217	sk_acceptq_added(parent);
	218
	219	if (queue->rskq_accept_head == NULL)
	220	queue->rskq_accept_head = req;
	221	else
	222	queue->rskq_accept_tail->dl_next = req;
	223
	224	queue->rskq_accept_tail = req;
	225	req->dl_next = NULL;
	226	}
	227
	228	static inline struct request_sock reqsk_queue_remove(struct request_sock_queue queue)
	229	{
	230	struct request_sock *req = queue->rskq_accept_head;
	231
547b792c	232	WARN_ON(req == NULL);
0e87506f ACM	233
	234	queue->rskq_accept_head = req->dl_next;
	235	if (queue->rskq_accept_head == NULL)
	236	queue->rskq_accept_tail = NULL;
	237
	238	return req;
	239	}
	240
0e87506f ACM	241	static inline int reqsk_queue_removed(struct request_sock_queue *queue,
	242	struct request_sock *req)
	243	{
2ad69c55	244	struct listen_sock *lopt = queue->listen_opt;
0e87506f	245
e6c022a4	246	if (req->num_timeout == 0)
0e87506f ACM	247	--lopt->qlen_young;
	248
	249	return --lopt->qlen;
	250	}
	251
	252	static inline int reqsk_queue_added(struct request_sock_queue *queue)
	253	{
2ad69c55	254	struct listen_sock *lopt = queue->listen_opt;
0e87506f ACM	255	const int prev_qlen = lopt->qlen;
	256
	257	lopt->qlen_young++;
	258	lopt->qlen++;
	259	return prev_qlen;
	260	}
	261
463c84b9	262	static inline int reqsk_queue_len(const struct request_sock_queue *queue)
0e87506f ACM	263	{
	264	return queue->listen_opt != NULL ? queue->listen_opt->qlen : 0;
	265	}
	266
463c84b9	267	static inline int reqsk_queue_len_young(const struct request_sock_queue *queue)
0e87506f ACM	268	{
	269	return queue->listen_opt->qlen_young;
	270	}
	271
463c84b9	272	static inline int reqsk_queue_is_full(const struct request_sock_queue *queue)
0e87506f ACM	273	{
	274	return queue->listen_opt->qlen >> queue->listen_opt->max_qlen_log;
	275	}
	276
	277	static inline void reqsk_queue_hash_req(struct request_sock_queue *queue,
	278	u32 hash, struct request_sock *req,
8129765a	279	unsigned long timeout)
0e87506f	280	{
2ad69c55	281	struct listen_sock *lopt = queue->listen_opt;
0e87506f ACM	282
0e87506f ACM	283	req->expires = jiffies + timeout;
e6c022a4 ED	284	req->num_retrans = 0;
e6c022a4 ED	285	req->num_timeout = 0;
0e87506f ACM	286	req->sk = NULL;
	287	req->dl_next = lopt->syn_table[hash];
	288
	289	write_lock(&queue->syn_wait_lock);
	290	lopt->syn_table[hash] = req;
	291	write_unlock(&queue->syn_wait_lock);
	292	}
	293
2e6599cb	294	#endif /* _REQUEST_SOCK_H */