sock.h 61.0 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9
/*
 * INET		An implementation of the TCP/IP protocol suite for the LINUX
 *		operating system.  INET is implemented using the  BSD Socket
 *		interface as the means of communication with the user level.
 *
 *		Definitions for the AF_INET socket handler.
 *
 * Version:	@(#)sock.h	1.0.4	05/13/93
 *
10
 * Authors:	Ross Biro
L
Linus Torvalds 已提交
11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
 *		Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *		Corey Minyard <wf-rch!minyard@relay.EU.net>
 *		Florian La Roche <flla@stud.uni-sb.de>
 *
 * Fixes:
 *		Alan Cox	:	Volatiles in skbuff pointers. See
 *					skbuff comments. May be overdone,
 *					better to prove they can be removed
 *					than the reverse.
 *		Alan Cox	:	Added a zapped field for tcp to note
 *					a socket is reset and must stay shut up
 *		Alan Cox	:	New fields for options
 *	Pauline Middelink	:	identd support
 *		Alan Cox	:	Eliminate low level recv/recvfrom
 *		David S. Miller	:	New socket lookup architecture.
 *              Steve Whitehouse:       Default routines for sock_ops
 *              Arnaldo C. Melo :	removed net_pinfo, tp_pinfo and made
 *              			protinfo be just a void pointer, as the
 *              			protocol specific parts were moved to
 *              			respective headers and ipv4/v6, etc now
 *              			use private slabcaches for its socks
 *              Pedro Hortas	:	New flags field for socket options
 *
 *
 *		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 _SOCK_H
#define _SOCK_H

43
#include <linux/hardirq.h>
44
#include <linux/kernel.h>
L
Linus Torvalds 已提交
45
#include <linux/list.h>
46
#include <linux/list_nulls.h>
L
Linus Torvalds 已提交
47 48
#include <linux/timer.h>
#include <linux/cache.h>
49
#include <linux/bitops.h>
I
Ingo Molnar 已提交
50
#include <linux/lockdep.h>
L
Linus Torvalds 已提交
51 52
#include <linux/netdevice.h>
#include <linux/skbuff.h>	/* struct sk_buff */
A
Al Viro 已提交
53
#include <linux/mm.h>
L
Linus Torvalds 已提交
54
#include <linux/security.h>
55
#include <linux/slab.h>
56
#include <linux/uaccess.h>
57
#include <linux/memcontrol.h>
G
Glauber Costa 已提交
58
#include <linux/res_counter.h>
59
#include <linux/static_key.h>
A
Al Viro 已提交
60 61
#include <linux/aio.h>
#include <linux/sched.h>
L
Linus Torvalds 已提交
62 63

#include <linux/filter.h>
64
#include <linux/rculist_nulls.h>
65
#include <linux/poll.h>
L
Linus Torvalds 已提交
66

E
Eric Dumazet 已提交
67
#include <linux/atomic.h>
L
Linus Torvalds 已提交
68 69 70
#include <net/dst.h>
#include <net/checksum.h>

71 72
struct cgroup;
struct cgroup_subsys;
73
#ifdef CONFIG_NET
74 75
int mem_cgroup_sockets_init(struct mem_cgroup *memcg, struct cgroup_subsys *ss);
void mem_cgroup_sockets_destroy(struct mem_cgroup *memcg);
76 77
#else
static inline
78
int mem_cgroup_sockets_init(struct mem_cgroup *memcg, struct cgroup_subsys *ss)
79 80 81 82
{
	return 0;
}
static inline
83
void mem_cgroup_sockets_destroy(struct mem_cgroup *memcg)
84 85 86
{
}
#endif
L
Linus Torvalds 已提交
87 88 89 90 91 92 93 94 95 96 97 98
/*
 * This structure really needs to be cleaned up.
 * Most of it is for TCP, and not used by any of
 * the other protocols.
 */

/* Define this to get the SOCK_DBG debugging facility. */
#define SOCK_DEBUGGING
#ifdef SOCK_DEBUGGING
#define SOCK_DEBUG(sk, msg...) do { if ((sk) && sock_flag((sk), SOCK_DBG)) \
					printk(KERN_DEBUG msg); } while (0)
#else
99
/* Validate arguments and do nothing */
100
static inline __printf(2, 3)
E
Eric Dumazet 已提交
101
void SOCK_DEBUG(const struct sock *sk, const char *msg, ...)
102 103
{
}
L
Linus Torvalds 已提交
104 105 106 107 108 109 110 111
#endif

/* This is the per-socket lock.  The spinlock provides a synchronization
 * between user contexts and software interrupt processing, whereas the
 * mini-semaphore synchronizes multiple users amongst themselves.
 */
typedef struct {
	spinlock_t		slock;
112
	int			owned;
L
Linus Torvalds 已提交
113
	wait_queue_head_t	wq;
I
Ingo Molnar 已提交
114 115 116 117 118 119 120 121 122
	/*
	 * We express the mutex-alike socket_lock semantics
	 * to the lock validator by explicitly managing
	 * the slock as a lock variant (in addition to
	 * the slock itself):
	 */
#ifdef CONFIG_DEBUG_LOCK_ALLOC
	struct lockdep_map dep_map;
#endif
L
Linus Torvalds 已提交
123 124 125
} socket_lock_t;

struct sock;
126
struct proto;
127
struct net;
L
Linus Torvalds 已提交
128 129

/**
130
 *	struct sock_common - minimal network layer representation of sockets
131 132
 *	@skc_daddr: Foreign IPv4 addr
 *	@skc_rcv_saddr: Bound local IPv4 addr
E
Eric Dumazet 已提交
133
 *	@skc_hash: hash value used with various protocol lookup tables
134
 *	@skc_u16hashes: two u16 hash values used by UDP lookup tables
135 136 137 138 139
 *	@skc_family: network address family
 *	@skc_state: Connection state
 *	@skc_reuse: %SO_REUSEADDR setting
 *	@skc_bound_dev_if: bound device index if != 0
 *	@skc_bind_node: bind hash linkage for various protocol lookup tables
140
 *	@skc_portaddr_node: second hash linkage for UDP/UDP-Lite protocol
141
 *	@skc_prot: protocol handlers inside a network family
142
 *	@skc_net: reference to the network namespace of this socket
143 144 145 146
 *	@skc_node: main hash linkage for various protocol lookup tables
 *	@skc_nulls_node: main hash linkage for TCP/UDP/UDP-Lite protocol
 *	@skc_tx_queue_mapping: tx queue number for this connection
 *	@skc_refcnt: reference count
147 148
 *
 *	This is the minimal network layer representation of sockets, the header
149 150
 *	for struct sock and struct inet_timewait_sock.
 */
L
Linus Torvalds 已提交
151
struct sock_common {
152 153
	/* skc_daddr and skc_rcv_saddr must be grouped :
	 * cf INET_MATCH() and INET_TW_MATCH()
E
Eric Dumazet 已提交
154
	 */
155 156
	__be32			skc_daddr;
	__be32			skc_rcv_saddr;
E
Eric Dumazet 已提交
157

158 159 160 161
	union  {
		unsigned int	skc_hash;
		__u16		skc_u16hashes[2];
	};
E
Eric Dumazet 已提交
162 163 164 165
	unsigned short		skc_family;
	volatile unsigned char	skc_state;
	unsigned char		skc_reuse;
	int			skc_bound_dev_if;
166 167 168 169
	union {
		struct hlist_node	skc_bind_node;
		struct hlist_nulls_node skc_portaddr_node;
	};
170
	struct proto		*skc_prot;
171
#ifdef CONFIG_NET_NS
172
	struct net	 	*skc_net;
173
#endif
174 175 176 177
	/*
	 * fields between dontcopy_begin/dontcopy_end
	 * are not copied in sock_copy()
	 */
178
	/* private: */
179
	int			skc_dontcopy_begin[0];
180
	/* public: */
181 182 183 184 185 186
	union {
		struct hlist_node	skc_node;
		struct hlist_nulls_node skc_nulls_node;
	};
	int			skc_tx_queue_mapping;
	atomic_t		skc_refcnt;
187
	/* private: */
188
	int                     skc_dontcopy_end[0];
189
	/* public: */
L
Linus Torvalds 已提交
190 191
};

G
Glauber Costa 已提交
192
struct cg_proto;
L
Linus Torvalds 已提交
193 194
/**
  *	struct sock - network layer representation of sockets
195
  *	@__sk_common: shared layout with inet_timewait_sock
196 197 198 199
  *	@sk_shutdown: mask of %SEND_SHUTDOWN and/or %RCV_SHUTDOWN
  *	@sk_userlocks: %SO_SNDBUF and %SO_RCVBUF settings
  *	@sk_lock:	synchronizer
  *	@sk_rcvbuf: size of receive buffer in bytes
200
  *	@sk_wq: sock wait queue and async head
E
Eric Dumazet 已提交
201
  *	@sk_rx_dst: receive input route used by early tcp demux
202 203 204 205 206 207
  *	@sk_dst_cache: destination cache
  *	@sk_dst_lock: destination cache lock
  *	@sk_policy: flow policy
  *	@sk_receive_queue: incoming packets
  *	@sk_wmem_alloc: transmit queue bytes committed
  *	@sk_write_queue: Packet sending queue
208
  *	@sk_async_wait_queue: DMA copied packets
209 210 211 212 213
  *	@sk_omem_alloc: "o" is "option" or "other"
  *	@sk_wmem_queued: persistent queue size
  *	@sk_forward_alloc: space allocated forward
  *	@sk_allocation: allocation mode
  *	@sk_sndbuf: size of send buffer in bytes
W
Wang Chen 已提交
214
  *	@sk_flags: %SO_LINGER (l_onoff), %SO_BROADCAST, %SO_KEEPALIVE,
215
  *		   %SO_OOBINLINE settings, %SO_TIMESTAMPING settings
216 217
  *	@sk_no_check: %SO_NO_CHECK setting, wether or not checkup packets
  *	@sk_route_caps: route capabilities (e.g. %NETIF_F_TSO)
E
Eric Dumazet 已提交
218
  *	@sk_route_nocaps: forbidden route capabilities (e.g NETIF_F_GSO_MASK)
219
  *	@sk_gso_type: GSO type (e.g. %SKB_GSO_TCPV4)
220
  *	@sk_gso_max_size: Maximum GSO segment size to build
221 222 223 224
  *	@sk_lingertime: %SO_LINGER l_linger setting
  *	@sk_backlog: always used with the per-socket spinlock held
  *	@sk_callback_lock: used with the callbacks in the end of this struct
  *	@sk_error_queue: rarely used
W
Wang Chen 已提交
225 226
  *	@sk_prot_creator: sk_prot of original sock creator (see ipv6_setsockopt,
  *			  IPV6_ADDRFORM for instance)
227
  *	@sk_err: last error
W
Wang Chen 已提交
228 229
  *	@sk_err_soft: errors that don't cause failure but are the cause of a
  *		      persistent failure not just 'timed out'
E
Eric Dumazet 已提交
230
  *	@sk_drops: raw/udp drops counter
231 232 233
  *	@sk_ack_backlog: current listen backlog
  *	@sk_max_ack_backlog: listen backlog set in listen()
  *	@sk_priority: %SO_PRIORITY setting
234
  *	@sk_cgrp_prioidx: socket group's priority map index
235 236
  *	@sk_type: socket type (%SOCK_STREAM, etc)
  *	@sk_protocol: which protocol this socket belongs in this network family
237 238
  *	@sk_peer_pid: &struct pid for this socket's peer
  *	@sk_peer_cred: %SO_PEERCRED setting
239 240 241
  *	@sk_rcvlowat: %SO_RCVLOWAT setting
  *	@sk_rcvtimeo: %SO_RCVTIMEO setting
  *	@sk_sndtimeo: %SO_SNDTIMEO setting
242
  *	@sk_rxhash: flow hash received from netif layer
243 244 245 246 247 248 249 250
  *	@sk_filter: socket filtering instructions
  *	@sk_protinfo: private area, net family specific, when not using slab
  *	@sk_timer: sock cleanup timer
  *	@sk_stamp: time stamp of last packet received
  *	@sk_socket: Identd and reporting IO signals
  *	@sk_user_data: RPC layer private data
  *	@sk_sndmsg_page: cached page for sendmsg
  *	@sk_sndmsg_off: cached offset for sendmsg
251
  *	@sk_peek_off: current peek_offset value
252
  *	@sk_send_head: front of stuff to transmit
253
  *	@sk_security: used by security modules
254
  *	@sk_mark: generic packet mark
255
  *	@sk_classid: this socket's cgroup classid
G
Glauber Costa 已提交
256
  *	@sk_cgrp: this socket's cgroup-specific proto data
257 258 259 260 261 262 263
  *	@sk_write_pending: a write to stream socket waits to start
  *	@sk_state_change: callback to indicate change in the state of the sock
  *	@sk_data_ready: callback to indicate there is data to be processed
  *	@sk_write_space: callback to indicate there is bf sending space available
  *	@sk_error_report: callback to indicate errors (e.g. %MSG_ERRQUEUE)
  *	@sk_backlog_rcv: callback to process the backlog
  *	@sk_destruct: called at sock freeing time, i.e. when all refcnt == 0
L
Linus Torvalds 已提交
264 265 266
 */
struct sock {
	/*
267
	 * Now struct inet_timewait_sock also uses sock_common, so please just
L
Linus Torvalds 已提交
268 269 270
	 * don't add nothing before this first member (__sk_common) --acme
	 */
	struct sock_common	__sk_common;
E
Eric Dumazet 已提交
271 272 273
#define sk_node			__sk_common.skc_node
#define sk_nulls_node		__sk_common.skc_nulls_node
#define sk_refcnt		__sk_common.skc_refcnt
K
Krishna Kumar 已提交
274
#define sk_tx_queue_mapping	__sk_common.skc_tx_queue_mapping
E
Eric Dumazet 已提交
275

276 277
#define sk_dontcopy_begin	__sk_common.skc_dontcopy_begin
#define sk_dontcopy_end		__sk_common.skc_dontcopy_end
E
Eric Dumazet 已提交
278
#define sk_hash			__sk_common.skc_hash
L
Linus Torvalds 已提交
279 280 281 282 283
#define sk_family		__sk_common.skc_family
#define sk_state		__sk_common.skc_state
#define sk_reuse		__sk_common.skc_reuse
#define sk_bound_dev_if		__sk_common.skc_bound_dev_if
#define sk_bind_node		__sk_common.skc_bind_node
284
#define sk_prot			__sk_common.skc_prot
285
#define sk_net			__sk_common.skc_net
L
Linus Torvalds 已提交
286
	socket_lock_t		sk_lock;
E
Eric Dumazet 已提交
287
	struct sk_buff_head	sk_receive_queue;
288 289 290 291
	/*
	 * The backlog queue is special, it is always used with
	 * the per-socket spinlock held and requires low latency
	 * access. Therefore we special case it's implementation.
E
Eric Dumazet 已提交
292 293 294
	 * Note : rmem_alloc is in this structure to fill a hole
	 * on 64bit arches, not because its logically part of
	 * backlog.
295 296
	 */
	struct {
E
Eric Dumazet 已提交
297 298 299 300
		atomic_t	rmem_alloc;
		int		len;
		struct sk_buff	*head;
		struct sk_buff	*tail;
301
	} sk_backlog;
E
Eric Dumazet 已提交
302 303 304 305 306 307 308 309 310
#define sk_rmem_alloc sk_backlog.rmem_alloc
	int			sk_forward_alloc;
#ifdef CONFIG_RPS
	__u32			sk_rxhash;
#endif
	atomic_t		sk_drops;
	int			sk_rcvbuf;

	struct sk_filter __rcu	*sk_filter;
311
	struct socket_wq __rcu	*sk_wq;
E
Eric Dumazet 已提交
312 313 314 315 316

#ifdef CONFIG_NET_DMA
	struct sk_buff_head	sk_async_wait_queue;
#endif

317
#ifdef CONFIG_XFRM
L
Linus Torvalds 已提交
318
	struct xfrm_policy	*sk_policy[2];
319
#endif
E
Eric Dumazet 已提交
320
	unsigned long 		sk_flags;
E
Eric Dumazet 已提交
321
	struct dst_entry	*sk_rx_dst;
E
Eric Dumazet 已提交
322
	struct dst_entry	*sk_dst_cache;
E
Eric Dumazet 已提交
323
	spinlock_t		sk_dst_lock;
L
Linus Torvalds 已提交
324 325
	atomic_t		sk_wmem_alloc;
	atomic_t		sk_omem_alloc;
326
	int			sk_sndbuf;
L
Linus Torvalds 已提交
327
	struct sk_buff_head	sk_write_queue;
E
Eric Dumazet 已提交
328 329 330 331 332 333 334
	kmemcheck_bitfield_begin(flags);
	unsigned int		sk_shutdown  : 2,
				sk_no_check  : 2,
				sk_userlocks : 4,
				sk_protocol  : 8,
				sk_type      : 16;
	kmemcheck_bitfield_end(flags);
L
Linus Torvalds 已提交
335
	int			sk_wmem_queued;
A
Al Viro 已提交
336
	gfp_t			sk_allocation;
337 338
	netdev_features_t	sk_route_caps;
	netdev_features_t	sk_route_nocaps;
339
	int			sk_gso_type;
340
	unsigned int		sk_gso_max_size;
341
	int			sk_rcvlowat;
L
Linus Torvalds 已提交
342 343
	unsigned long	        sk_lingertime;
	struct sk_buff_head	sk_error_queue;
344
	struct proto		*sk_prot_creator;
L
Linus Torvalds 已提交
345 346 347 348 349 350
	rwlock_t		sk_callback_lock;
	int			sk_err,
				sk_err_soft;
	unsigned short		sk_ack_backlog;
	unsigned short		sk_max_ack_backlog;
	__u32			sk_priority;
351 352 353
#ifdef CONFIG_CGROUPS
	__u32			sk_cgrp_prioidx;
#endif
354 355
	struct pid		*sk_peer_pid;
	const struct cred	*sk_peer_cred;
L
Linus Torvalds 已提交
356 357 358 359
	long			sk_rcvtimeo;
	long			sk_sndtimeo;
	void			*sk_protinfo;
	struct timer_list	sk_timer;
360
	ktime_t			sk_stamp;
L
Linus Torvalds 已提交
361 362 363 364 365
	struct socket		*sk_socket;
	void			*sk_user_data;
	struct page		*sk_sndmsg_page;
	struct sk_buff		*sk_send_head;
	__u32			sk_sndmsg_off;
366
	__s32			sk_peek_off;
L
Linus Torvalds 已提交
367
	int			sk_write_pending;
A
Alexey Dobriyan 已提交
368
#ifdef CONFIG_SECURITY
L
Linus Torvalds 已提交
369
	void			*sk_security;
A
Alexey Dobriyan 已提交
370
#endif
371
	__u32			sk_mark;
372
	u32			sk_classid;
G
Glauber Costa 已提交
373
	struct cg_proto		*sk_cgrp;
L
Linus Torvalds 已提交
374 375 376 377
	void			(*sk_state_change)(struct sock *sk);
	void			(*sk_data_ready)(struct sock *sk, int bytes);
	void			(*sk_write_space)(struct sock *sk);
	void			(*sk_error_report)(struct sock *sk);
E
Eric Dumazet 已提交
378 379
	int			(*sk_backlog_rcv)(struct sock *sk,
						  struct sk_buff *skb);
L
Linus Torvalds 已提交
380 381 382
	void                    (*sk_destruct)(struct sock *sk);
};

383 384 385 386 387 388 389 390 391 392 393
/*
 * SK_CAN_REUSE and SK_NO_REUSE on a socket mean that the socket is OK
 * or not whether his port will be reused by someone else. SK_FORCE_REUSE
 * on a socket means that the socket will reuse everybody else's port
 * without looking at the other's sk_reuse value.
 */

#define SK_NO_REUSE	0
#define SK_CAN_REUSE	1
#define SK_FORCE_REUSE	2

394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417
static inline int sk_peek_offset(struct sock *sk, int flags)
{
	if ((flags & MSG_PEEK) && (sk->sk_peek_off >= 0))
		return sk->sk_peek_off;
	else
		return 0;
}

static inline void sk_peek_offset_bwd(struct sock *sk, int val)
{
	if (sk->sk_peek_off >= 0) {
		if (sk->sk_peek_off >= val)
			sk->sk_peek_off -= val;
		else
			sk->sk_peek_off = 0;
	}
}

static inline void sk_peek_offset_fwd(struct sock *sk, int val)
{
	if (sk->sk_peek_off >= 0)
		sk->sk_peek_off += val;
}

L
Linus Torvalds 已提交
418 419 420
/*
 * Hashed lists helper routines
 */
L
Li Zefan 已提交
421 422 423 424 425
static inline struct sock *sk_entry(const struct hlist_node *node)
{
	return hlist_entry(node, struct sock, sk_node);
}

426
static inline struct sock *__sk_head(const struct hlist_head *head)
L
Linus Torvalds 已提交
427 428 429 430
{
	return hlist_entry(head->first, struct sock, sk_node);
}

431
static inline struct sock *sk_head(const struct hlist_head *head)
L
Linus Torvalds 已提交
432 433 434 435
{
	return hlist_empty(head) ? NULL : __sk_head(head);
}

436 437 438 439 440 441 442 443 444 445
static inline struct sock *__sk_nulls_head(const struct hlist_nulls_head *head)
{
	return hlist_nulls_entry(head->first, struct sock, sk_nulls_node);
}

static inline struct sock *sk_nulls_head(const struct hlist_nulls_head *head)
{
	return hlist_nulls_empty(head) ? NULL : __sk_nulls_head(head);
}

446
static inline struct sock *sk_next(const struct sock *sk)
L
Linus Torvalds 已提交
447 448 449 450 451
{
	return sk->sk_node.next ?
		hlist_entry(sk->sk_node.next, struct sock, sk_node) : NULL;
}

452 453 454 455 456 457 458 459
static inline struct sock *sk_nulls_next(const struct sock *sk)
{
	return (!is_a_nulls(sk->sk_nulls_node.next)) ?
		hlist_nulls_entry(sk->sk_nulls_node.next,
				  struct sock, sk_nulls_node) :
		NULL;
}

E
Eric Dumazet 已提交
460
static inline bool sk_unhashed(const struct sock *sk)
L
Linus Torvalds 已提交
461 462 463 464
{
	return hlist_unhashed(&sk->sk_node);
}

E
Eric Dumazet 已提交
465
static inline bool sk_hashed(const struct sock *sk)
L
Linus Torvalds 已提交
466
{
A
Akinobu Mita 已提交
467
	return !sk_unhashed(sk);
L
Linus Torvalds 已提交
468 469
}

E
Eric Dumazet 已提交
470
static inline void sk_node_init(struct hlist_node *node)
L
Linus Torvalds 已提交
471 472 473 474
{
	node->pprev = NULL;
}

E
Eric Dumazet 已提交
475
static inline void sk_nulls_node_init(struct hlist_nulls_node *node)
476 477 478 479
{
	node->pprev = NULL;
}

E
Eric Dumazet 已提交
480
static inline void __sk_del_node(struct sock *sk)
L
Linus Torvalds 已提交
481 482 483 484
{
	__hlist_del(&sk->sk_node);
}

485
/* NB: equivalent to hlist_del_init_rcu */
E
Eric Dumazet 已提交
486
static inline bool __sk_del_node_init(struct sock *sk)
L
Linus Torvalds 已提交
487 488 489 490
{
	if (sk_hashed(sk)) {
		__sk_del_node(sk);
		sk_node_init(&sk->sk_node);
E
Eric Dumazet 已提交
491
		return true;
L
Linus Torvalds 已提交
492
	}
E
Eric Dumazet 已提交
493
	return false;
L
Linus Torvalds 已提交
494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514
}

/* Grab socket reference count. This operation is valid only
   when sk is ALREADY grabbed f.e. it is found in hash table
   or a list and the lookup is made under lock preventing hash table
   modifications.
 */

static inline void sock_hold(struct sock *sk)
{
	atomic_inc(&sk->sk_refcnt);
}

/* Ungrab socket in the context, which assumes that socket refcnt
   cannot hit zero, f.e. it is true in context of any socketcall.
 */
static inline void __sock_put(struct sock *sk)
{
	atomic_dec(&sk->sk_refcnt);
}

E
Eric Dumazet 已提交
515
static inline bool sk_del_node_init(struct sock *sk)
L
Linus Torvalds 已提交
516
{
E
Eric Dumazet 已提交
517
	bool rc = __sk_del_node_init(sk);
L
Linus Torvalds 已提交
518 519 520 521 522 523 524 525

	if (rc) {
		/* paranoid for a while -acme */
		WARN_ON(atomic_read(&sk->sk_refcnt) == 1);
		__sock_put(sk);
	}
	return rc;
}
526
#define sk_del_node_init_rcu(sk)	sk_del_node_init(sk)
L
Linus Torvalds 已提交
527

E
Eric Dumazet 已提交
528
static inline bool __sk_nulls_del_node_init_rcu(struct sock *sk)
529 530
{
	if (sk_hashed(sk)) {
531
		hlist_nulls_del_init_rcu(&sk->sk_nulls_node);
E
Eric Dumazet 已提交
532
		return true;
533
	}
E
Eric Dumazet 已提交
534
	return false;
535 536
}

E
Eric Dumazet 已提交
537
static inline bool sk_nulls_del_node_init_rcu(struct sock *sk)
538
{
E
Eric Dumazet 已提交
539
	bool rc = __sk_nulls_del_node_init_rcu(sk);
540 541 542 543 544 545 546 547 548

	if (rc) {
		/* paranoid for a while -acme */
		WARN_ON(atomic_read(&sk->sk_refcnt) == 1);
		__sock_put(sk);
	}
	return rc;
}

E
Eric Dumazet 已提交
549
static inline void __sk_add_node(struct sock *sk, struct hlist_head *list)
L
Linus Torvalds 已提交
550 551 552 553
{
	hlist_add_head(&sk->sk_node, list);
}

E
Eric Dumazet 已提交
554
static inline void sk_add_node(struct sock *sk, struct hlist_head *list)
L
Linus Torvalds 已提交
555 556 557 558 559
{
	sock_hold(sk);
	__sk_add_node(sk, list);
}

E
Eric Dumazet 已提交
560
static inline void sk_add_node_rcu(struct sock *sk, struct hlist_head *list)
561 562 563 564 565
{
	sock_hold(sk);
	hlist_add_head_rcu(&sk->sk_node, list);
}

E
Eric Dumazet 已提交
566
static inline void __sk_nulls_add_node_rcu(struct sock *sk, struct hlist_nulls_head *list)
567
{
568
	hlist_nulls_add_head_rcu(&sk->sk_nulls_node, list);
569 570
}

E
Eric Dumazet 已提交
571
static inline void sk_nulls_add_node_rcu(struct sock *sk, struct hlist_nulls_head *list)
572 573
{
	sock_hold(sk);
574
	__sk_nulls_add_node_rcu(sk, list);
575 576
}

E
Eric Dumazet 已提交
577
static inline void __sk_del_bind_node(struct sock *sk)
L
Linus Torvalds 已提交
578 579 580 581
{
	__hlist_del(&sk->sk_bind_node);
}

E
Eric Dumazet 已提交
582
static inline void sk_add_bind_node(struct sock *sk,
L
Linus Torvalds 已提交
583 584 585 586 587 588 589
					struct hlist_head *list)
{
	hlist_add_head(&sk->sk_bind_node, list);
}

#define sk_for_each(__sk, node, list) \
	hlist_for_each_entry(__sk, node, list, sk_node)
590 591
#define sk_for_each_rcu(__sk, node, list) \
	hlist_for_each_entry_rcu(__sk, node, list, sk_node)
592 593 594 595
#define sk_nulls_for_each(__sk, node, list) \
	hlist_nulls_for_each_entry(__sk, node, list, sk_nulls_node)
#define sk_nulls_for_each_rcu(__sk, node, list) \
	hlist_nulls_for_each_entry_rcu(__sk, node, list, sk_nulls_node)
L
Linus Torvalds 已提交
596 597 598
#define sk_for_each_from(__sk, node) \
	if (__sk && ({ node = &(__sk)->sk_node; 1; })) \
		hlist_for_each_entry_from(__sk, node, sk_node)
599 600 601
#define sk_nulls_for_each_from(__sk, node) \
	if (__sk && ({ node = &(__sk)->sk_nulls_node; 1; })) \
		hlist_nulls_for_each_entry_from(__sk, node, sk_nulls_node)
L
Linus Torvalds 已提交
602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620
#define sk_for_each_safe(__sk, node, tmp, list) \
	hlist_for_each_entry_safe(__sk, node, tmp, list, sk_node)
#define sk_for_each_bound(__sk, node, list) \
	hlist_for_each_entry(__sk, node, list, sk_bind_node)

/* Sock flags */
enum sock_flags {
	SOCK_DEAD,
	SOCK_DONE,
	SOCK_URGINLINE,
	SOCK_KEEPOPEN,
	SOCK_LINGER,
	SOCK_DESTROY,
	SOCK_BROADCAST,
	SOCK_TIMESTAMP,
	SOCK_ZAPPED,
	SOCK_USE_WRITE_QUEUE, /* whether to call sk->sk_write_space in sock_wfree */
	SOCK_DBG, /* %SO_DEBUG setting */
	SOCK_RCVTSTAMP, /* %SO_TIMESTAMP setting */
621
	SOCK_RCVTSTAMPNS, /* %SO_TIMESTAMPNS setting */
L
Linus Torvalds 已提交
622 623
	SOCK_LOCALROUTE, /* route locally only, %SO_DONTROUTE setting */
	SOCK_QUEUE_SHRUNK, /* write queue has been shrunk recently */
624 625 626 627 628 629 630
	SOCK_TIMESTAMPING_TX_HARDWARE,  /* %SOF_TIMESTAMPING_TX_HARDWARE */
	SOCK_TIMESTAMPING_TX_SOFTWARE,  /* %SOF_TIMESTAMPING_TX_SOFTWARE */
	SOCK_TIMESTAMPING_RX_HARDWARE,  /* %SOF_TIMESTAMPING_RX_HARDWARE */
	SOCK_TIMESTAMPING_RX_SOFTWARE,  /* %SOF_TIMESTAMPING_RX_SOFTWARE */
	SOCK_TIMESTAMPING_SOFTWARE,     /* %SOF_TIMESTAMPING_SOFTWARE */
	SOCK_TIMESTAMPING_RAW_HARDWARE, /* %SOF_TIMESTAMPING_RAW_HARDWARE */
	SOCK_TIMESTAMPING_SYS_HARDWARE, /* %SOF_TIMESTAMPING_SYS_HARDWARE */
E
Eric Dumazet 已提交
631
	SOCK_FASYNC, /* fasync() active */
632
	SOCK_RXQ_OVFL,
633
	SOCK_ZEROCOPY, /* buffers from userspace */
634
	SOCK_WIFI_STATUS, /* push wifi status to userspace */
635 636 637 638
	SOCK_NOFCS, /* Tell NIC not to do the Ethernet FCS.
		     * Will use last 4 bytes of packet sent from
		     * user-space instead.
		     */
L
Linus Torvalds 已提交
639 640
};

R
Ralf Baechle 已提交
641 642 643 644 645
static inline void sock_copy_flags(struct sock *nsk, struct sock *osk)
{
	nsk->sk_flags = osk->sk_flags;
}

L
Linus Torvalds 已提交
646 647 648 649 650 651 652 653 654 655
static inline void sock_set_flag(struct sock *sk, enum sock_flags flag)
{
	__set_bit(flag, &sk->sk_flags);
}

static inline void sock_reset_flag(struct sock *sk, enum sock_flags flag)
{
	__clear_bit(flag, &sk->sk_flags);
}

E
Eric Dumazet 已提交
656
static inline bool sock_flag(const struct sock *sk, enum sock_flags flag)
L
Linus Torvalds 已提交
657 658 659 660
{
	return test_bit(flag, &sk->sk_flags);
}

661 662 663 664 665
static inline gfp_t sk_gfp_atomic(struct sock *sk, gfp_t gfp_mask)
{
	return GFP_ATOMIC;
}

L
Linus Torvalds 已提交
666 667 668 669 670 671 672 673 674 675
static inline void sk_acceptq_removed(struct sock *sk)
{
	sk->sk_ack_backlog--;
}

static inline void sk_acceptq_added(struct sock *sk)
{
	sk->sk_ack_backlog++;
}

E
Eric Dumazet 已提交
676
static inline bool sk_acceptq_is_full(const struct sock *sk)
L
Linus Torvalds 已提交
677
{
678
	return sk->sk_ack_backlog > sk->sk_max_ack_backlog;
L
Linus Torvalds 已提交
679 680 681 682 683
}

/*
 * Compute minimal free write space needed to queue new packets.
 */
E
Eric Dumazet 已提交
684
static inline int sk_stream_min_wspace(const struct sock *sk)
L
Linus Torvalds 已提交
685
{
686
	return sk->sk_wmem_queued >> 1;
L
Linus Torvalds 已提交
687 688
}

E
Eric Dumazet 已提交
689
static inline int sk_stream_wspace(const struct sock *sk)
L
Linus Torvalds 已提交
690 691 692 693 694 695
{
	return sk->sk_sndbuf - sk->sk_wmem_queued;
}

extern void sk_stream_write_space(struct sock *sk);

E
Eric Dumazet 已提交
696
static inline bool sk_stream_memory_free(const struct sock *sk)
L
Linus Torvalds 已提交
697 698 699 700
{
	return sk->sk_wmem_queued < sk->sk_sndbuf;
}

Z
Zhu Yi 已提交
701
/* OOB backlog add */
Z
Zhu Yi 已提交
702
static inline void __sk_add_backlog(struct sock *sk, struct sk_buff *skb)
703
{
E
Eric Dumazet 已提交
704 705 706 707 708 709
	/* dont let skb dst not refcounted, we are going to leave rcu lock */
	skb_dst_force(skb);

	if (!sk->sk_backlog.tail)
		sk->sk_backlog.head = skb;
	else
710
		sk->sk_backlog.tail->next = skb;
E
Eric Dumazet 已提交
711 712

	sk->sk_backlog.tail = skb;
713 714
	skb->next = NULL;
}
L
Linus Torvalds 已提交
715

716 717
/*
 * Take into account size of receive queue and backlog queue
E
Eric Dumazet 已提交
718 719
 * Do not take into account this skb truesize,
 * to allow even a single big packet to come.
720
 */
721 722
static inline bool sk_rcvqueues_full(const struct sock *sk, const struct sk_buff *skb,
				     unsigned int limit)
723 724 725
{
	unsigned int qsize = sk->sk_backlog.len + atomic_read(&sk->sk_rmem_alloc);

726
	return qsize > limit;
727 728
}

Z
Zhu Yi 已提交
729
/* The per-socket spinlock must be held here. */
730 731
static inline __must_check int sk_add_backlog(struct sock *sk, struct sk_buff *skb,
					      unsigned int limit)
Z
Zhu Yi 已提交
732
{
733
	if (sk_rcvqueues_full(sk, skb, limit))
Z
Zhu Yi 已提交
734 735
		return -ENOBUFS;

Z
Zhu Yi 已提交
736
	__sk_add_backlog(sk, skb);
Z
Zhu Yi 已提交
737 738 739 740
	sk->sk_backlog.len += skb->truesize;
	return 0;
}

P
Peter Zijlstra 已提交
741 742 743 744 745
static inline int sk_backlog_rcv(struct sock *sk, struct sk_buff *skb)
{
	return sk->sk_backlog_rcv(sk, skb);
}

746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769
static inline void sock_rps_record_flow(const struct sock *sk)
{
#ifdef CONFIG_RPS
	struct rps_sock_flow_table *sock_flow_table;

	rcu_read_lock();
	sock_flow_table = rcu_dereference(rps_sock_flow_table);
	rps_record_sock_flow(sock_flow_table, sk->sk_rxhash);
	rcu_read_unlock();
#endif
}

static inline void sock_rps_reset_flow(const struct sock *sk)
{
#ifdef CONFIG_RPS
	struct rps_sock_flow_table *sock_flow_table;

	rcu_read_lock();
	sock_flow_table = rcu_dereference(rps_sock_flow_table);
	rps_reset_sock_flow(sock_flow_table, sk->sk_rxhash);
	rcu_read_unlock();
#endif
}

770 771
static inline void sock_rps_save_rxhash(struct sock *sk,
					const struct sk_buff *skb)
772 773
{
#ifdef CONFIG_RPS
774
	if (unlikely(sk->sk_rxhash != skb->rxhash)) {
775
		sock_rps_reset_flow(sk);
776
		sk->sk_rxhash = skb->rxhash;
777 778 779 780
	}
#endif
}

781 782 783 784 785 786 787 788
static inline void sock_rps_reset_rxhash(struct sock *sk)
{
#ifdef CONFIG_RPS
	sock_rps_reset_flow(sk);
	sk->sk_rxhash = 0;
#endif
}

S
Stephen Hemminger 已提交
789 790 791 792 793 794 795 796 797 798 799
#define sk_wait_event(__sk, __timeo, __condition)			\
	({	int __rc;						\
		release_sock(__sk);					\
		__rc = __condition;					\
		if (!__rc) {						\
			*(__timeo) = schedule_timeout(*(__timeo));	\
		}							\
		lock_sock(__sk);					\
		__rc = __condition;					\
		__rc;							\
	})
L
Linus Torvalds 已提交
800 801 802 803 804 805 806 807 808

extern int sk_stream_wait_connect(struct sock *sk, long *timeo_p);
extern int sk_stream_wait_memory(struct sock *sk, long *timeo_p);
extern void sk_stream_wait_close(struct sock *sk, long timeo_p);
extern int sk_stream_error(struct sock *sk, int flags, int err);
extern void sk_stream_kill_queues(struct sock *sk);

extern int sk_wait_data(struct sock *sk, long *timeo);

809
struct request_sock_ops;
810
struct timewait_sock_ops;
811
struct inet_hashinfo;
812
struct raw_hashinfo;
813
struct module;
814

L
Linus Torvalds 已提交
815 816 817 818 819
/* Networking protocol blocks we attach to sockets.
 * socket layer -> transport layer interface
 * transport -> network interface is defined by struct inet_proto
 */
struct proto {
E
Eric Dumazet 已提交
820
	void			(*close)(struct sock *sk,
L
Linus Torvalds 已提交
821 822
					long timeout);
	int			(*connect)(struct sock *sk,
E
Eric Dumazet 已提交
823
					struct sockaddr *uaddr,
L
Linus Torvalds 已提交
824 825 826
					int addr_len);
	int			(*disconnect)(struct sock *sk, int flags);

E
Eric Dumazet 已提交
827
	struct sock *		(*accept)(struct sock *sk, int flags, int *err);
L
Linus Torvalds 已提交
828 829 830 831

	int			(*ioctl)(struct sock *sk, int cmd,
					 unsigned long arg);
	int			(*init)(struct sock *sk);
832
	void			(*destroy)(struct sock *sk);
L
Linus Torvalds 已提交
833
	void			(*shutdown)(struct sock *sk, int how);
E
Eric Dumazet 已提交
834
	int			(*setsockopt)(struct sock *sk, int level,
L
Linus Torvalds 已提交
835
					int optname, char __user *optval,
836
					unsigned int optlen);
E
Eric Dumazet 已提交
837 838 839
	int			(*getsockopt)(struct sock *sk, int level,
					int optname, char __user *optval,
					int __user *option);
A
Alexey Dobriyan 已提交
840
#ifdef CONFIG_COMPAT
841 842 843
	int			(*compat_setsockopt)(struct sock *sk,
					int level,
					int optname, char __user *optval,
844
					unsigned int optlen);
845 846 847 848
	int			(*compat_getsockopt)(struct sock *sk,
					int level,
					int optname, char __user *optval,
					int __user *option);
849 850
	int			(*compat_ioctl)(struct sock *sk,
					unsigned int cmd, unsigned long arg);
A
Alexey Dobriyan 已提交
851
#endif
L
Linus Torvalds 已提交
852 853 854 855
	int			(*sendmsg)(struct kiocb *iocb, struct sock *sk,
					   struct msghdr *msg, size_t len);
	int			(*recvmsg)(struct kiocb *iocb, struct sock *sk,
					   struct msghdr *msg,
E
Eric Dumazet 已提交
856 857
					   size_t len, int noblock, int flags,
					   int *addr_len);
L
Linus Torvalds 已提交
858 859
	int			(*sendpage)(struct sock *sk, struct page *page,
					int offset, size_t size, int flags);
E
Eric Dumazet 已提交
860
	int			(*bind)(struct sock *sk,
L
Linus Torvalds 已提交
861 862
					struct sockaddr *uaddr, int addr_len);

E
Eric Dumazet 已提交
863
	int			(*backlog_rcv) (struct sock *sk,
L
Linus Torvalds 已提交
864 865
						struct sk_buff *skb);

E
Eric Dumazet 已提交
866
	void		(*release_cb)(struct sock *sk);
867
	void		(*mtu_reduced)(struct sock *sk);
E
Eric Dumazet 已提交
868

L
Linus Torvalds 已提交
869 870 871
	/* Keeping track of sk's, looking them up, and port selection methods. */
	void			(*hash)(struct sock *sk);
	void			(*unhash)(struct sock *sk);
E
Eric Dumazet 已提交
872
	void			(*rehash)(struct sock *sk);
L
Linus Torvalds 已提交
873
	int			(*get_port)(struct sock *sk, unsigned short snum);
874
	void			(*clear_sk)(struct sock *sk, int size);
L
Linus Torvalds 已提交
875

876
	/* Keeping track of sockets in use */
877
#ifdef CONFIG_PROC_FS
878
	unsigned int		inuse_idx;
879
#endif
880

L
Linus Torvalds 已提交
881
	/* Memory pressure */
882
	void			(*enter_memory_pressure)(struct sock *sk);
E
Eric Dumazet 已提交
883
	atomic_long_t		*memory_allocated;	/* Current allocated memory. */
884
	struct percpu_counter	*sockets_allocated;	/* Current number of sockets. */
L
Linus Torvalds 已提交
885 886 887
	/*
	 * Pressure flag: try to collapse.
	 * Technical note: it is used by multiple contexts non atomically.
888
	 * All the __sk_mem_schedule() is of this nature: accounting
L
Linus Torvalds 已提交
889 890 891
	 * is strict, actions are advisory and have some latency.
	 */
	int			*memory_pressure;
E
Eric Dumazet 已提交
892
	long			*sysctl_mem;
L
Linus Torvalds 已提交
893 894 895
	int			*sysctl_wmem;
	int			*sysctl_rmem;
	int			max_header;
896
	bool			no_autobind;
L
Linus Torvalds 已提交
897

898
	struct kmem_cache	*slab;
L
Linus Torvalds 已提交
899
	unsigned int		obj_size;
900
	int			slab_flags;
L
Linus Torvalds 已提交
901

902
	struct percpu_counter	*orphan_count;
903

904
	struct request_sock_ops	*rsk_prot;
905
	struct timewait_sock_ops *twsk_prot;
906

907 908
	union {
		struct inet_hashinfo	*hashinfo;
909
		struct udp_table	*udp_table;
910
		struct raw_hashinfo	*raw_hash;
911
	} h;
912

L
Linus Torvalds 已提交
913 914 915 916 917
	struct module		*owner;

	char			name[32];

	struct list_head	node;
918 919 920
#ifdef SOCK_REFCNT_DEBUG
	atomic_t		socks;
#endif
A
Andrew Morton 已提交
921
#ifdef CONFIG_MEMCG_KMEM
G
Glauber Costa 已提交
922 923 924 925 926 927
	/*
	 * cgroup specific init/deinit functions. Called once for all
	 * protocols that implement it, from cgroups populate function.
	 * This function has to setup any files the protocol want to
	 * appear in the kmem cgroup filesystem.
	 */
928
	int			(*init_cgroup)(struct mem_cgroup *memcg,
G
Glauber Costa 已提交
929
					       struct cgroup_subsys *ss);
930
	void			(*destroy_cgroup)(struct mem_cgroup *memcg);
G
Glauber Costa 已提交
931 932 933 934
	struct cg_proto		*(*proto_cgroup)(struct mem_cgroup *memcg);
#endif
};

935 936 937 938 939 940 941 942 943 944
/*
 * Bits in struct cg_proto.flags
 */
enum cg_proto_flags {
	/* Currently active and new sockets should be assigned to cgroups */
	MEMCG_SOCK_ACTIVE,
	/* It was ever activated; we must disarm static keys on destruction */
	MEMCG_SOCK_ACTIVATED,
};

G
Glauber Costa 已提交
945 946 947 948 949 950
struct cg_proto {
	void			(*enter_memory_pressure)(struct sock *sk);
	struct res_counter	*memory_allocated;	/* Current allocated memory. */
	struct percpu_counter	*sockets_allocated;	/* Current number of sockets. */
	int			*memory_pressure;
	long			*sysctl_mem;
951
	unsigned long		flags;
G
Glauber Costa 已提交
952 953 954 955 956 957 958 959 960 961
	/*
	 * memcg field is used to find which memcg we belong directly
	 * Each memcg struct can hold more than one cg_proto, so container_of
	 * won't really cut.
	 *
	 * The elegant solution would be having an inverse function to
	 * proto_cgroup in struct proto, but that means polluting the structure
	 * for everybody, instead of just for memcg users.
	 */
	struct mem_cgroup	*memcg;
L
Linus Torvalds 已提交
962 963 964 965 966
};

extern int proto_register(struct proto *prot, int alloc_slab);
extern void proto_unregister(struct proto *prot);

967 968 969 970 971 972 973 974 975 976
static inline bool memcg_proto_active(struct cg_proto *cg_proto)
{
	return test_bit(MEMCG_SOCK_ACTIVE, &cg_proto->flags);
}

static inline bool memcg_proto_activated(struct cg_proto *cg_proto)
{
	return test_bit(MEMCG_SOCK_ACTIVATED, &cg_proto->flags);
}

977 978 979 980 981 982 983 984 985 986 987 988 989
#ifdef SOCK_REFCNT_DEBUG
static inline void sk_refcnt_debug_inc(struct sock *sk)
{
	atomic_inc(&sk->sk_prot->socks);
}

static inline void sk_refcnt_debug_dec(struct sock *sk)
{
	atomic_dec(&sk->sk_prot->socks);
	printk(KERN_DEBUG "%s socket %p released, %d are still alive\n",
	       sk->sk_prot->name, sk, atomic_read(&sk->sk_prot->socks));
}

G
Glauber Costa 已提交
990
inline void sk_refcnt_debug_release(const struct sock *sk)
991 992 993 994 995 996 997 998 999 1000 1001
{
	if (atomic_read(&sk->sk_refcnt) != 1)
		printk(KERN_DEBUG "Destruction of the %s socket %p delayed, refcnt=%d\n",
		       sk->sk_prot->name, sk, atomic_read(&sk->sk_refcnt));
}
#else /* SOCK_REFCNT_DEBUG */
#define sk_refcnt_debug_inc(sk) do { } while (0)
#define sk_refcnt_debug_dec(sk) do { } while (0)
#define sk_refcnt_debug_release(sk) do { } while (0)
#endif /* SOCK_REFCNT_DEBUG */

A
Andrew Morton 已提交
1002
#if defined(CONFIG_MEMCG_KMEM) && defined(CONFIG_NET)
1003
extern struct static_key memcg_socket_limit_enabled;
G
Glauber Costa 已提交
1004 1005 1006 1007 1008
static inline struct cg_proto *parent_cg_proto(struct proto *proto,
					       struct cg_proto *cg_proto)
{
	return proto->proto_cgroup(parent_mem_cgroup(cg_proto->memcg));
}
1009
#define mem_cgroup_sockets_enabled static_key_false(&memcg_socket_limit_enabled)
G
Glauber Costa 已提交
1010 1011 1012 1013 1014 1015 1016 1017 1018 1019
#else
#define mem_cgroup_sockets_enabled 0
static inline struct cg_proto *parent_cg_proto(struct proto *proto,
					       struct cg_proto *cg_proto)
{
	return NULL;
}
#endif


1020 1021 1022 1023 1024 1025 1026 1027 1028
static inline bool sk_has_memory_pressure(const struct sock *sk)
{
	return sk->sk_prot->memory_pressure != NULL;
}

static inline bool sk_under_memory_pressure(const struct sock *sk)
{
	if (!sk->sk_prot->memory_pressure)
		return false;
G
Glauber Costa 已提交
1029 1030 1031 1032

	if (mem_cgroup_sockets_enabled && sk->sk_cgrp)
		return !!*sk->sk_cgrp->memory_pressure;

1033 1034 1035 1036 1037 1038 1039
	return !!*sk->sk_prot->memory_pressure;
}

static inline void sk_leave_memory_pressure(struct sock *sk)
{
	int *memory_pressure = sk->sk_prot->memory_pressure;

G
Glauber Costa 已提交
1040 1041 1042 1043
	if (!memory_pressure)
		return;

	if (*memory_pressure)
1044
		*memory_pressure = 0;
G
Glauber Costa 已提交
1045 1046 1047 1048 1049 1050 1051 1052 1053 1054

	if (mem_cgroup_sockets_enabled && sk->sk_cgrp) {
		struct cg_proto *cg_proto = sk->sk_cgrp;
		struct proto *prot = sk->sk_prot;

		for (; cg_proto; cg_proto = parent_cg_proto(prot, cg_proto))
			if (*cg_proto->memory_pressure)
				*cg_proto->memory_pressure = 0;
	}

1055 1056 1057 1058
}

static inline void sk_enter_memory_pressure(struct sock *sk)
{
G
Glauber Costa 已提交
1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070
	if (!sk->sk_prot->enter_memory_pressure)
		return;

	if (mem_cgroup_sockets_enabled && sk->sk_cgrp) {
		struct cg_proto *cg_proto = sk->sk_cgrp;
		struct proto *prot = sk->sk_prot;

		for (; cg_proto; cg_proto = parent_cg_proto(prot, cg_proto))
			cg_proto->enter_memory_pressure(sk);
	}

	sk->sk_prot->enter_memory_pressure(sk);
1071 1072 1073 1074 1075
}

static inline long sk_prot_mem_limits(const struct sock *sk, int index)
{
	long *prot = sk->sk_prot->sysctl_mem;
G
Glauber Costa 已提交
1076 1077
	if (mem_cgroup_sockets_enabled && sk->sk_cgrp)
		prot = sk->sk_cgrp->sysctl_mem;
1078 1079 1080
	return prot[index];
}

G
Glauber Costa 已提交
1081 1082 1083 1084 1085 1086 1087
static inline void memcg_memory_allocated_add(struct cg_proto *prot,
					      unsigned long amt,
					      int *parent_status)
{
	struct res_counter *fail;
	int ret;

1088 1089
	ret = res_counter_charge_nofail(prot->memory_allocated,
					amt << PAGE_SHIFT, &fail);
G
Glauber Costa 已提交
1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106
	if (ret < 0)
		*parent_status = OVER_LIMIT;
}

static inline void memcg_memory_allocated_sub(struct cg_proto *prot,
					      unsigned long amt)
{
	res_counter_uncharge(prot->memory_allocated, amt << PAGE_SHIFT);
}

static inline u64 memcg_memory_allocated_read(struct cg_proto *prot)
{
	u64 ret;
	ret = res_counter_read_u64(prot->memory_allocated, RES_USAGE);
	return ret >> PAGE_SHIFT;
}

1107 1108 1109 1110
static inline long
sk_memory_allocated(const struct sock *sk)
{
	struct proto *prot = sk->sk_prot;
G
Glauber Costa 已提交
1111 1112 1113
	if (mem_cgroup_sockets_enabled && sk->sk_cgrp)
		return memcg_memory_allocated_read(sk->sk_cgrp);

1114 1115 1116 1117
	return atomic_long_read(prot->memory_allocated);
}

static inline long
G
Glauber Costa 已提交
1118
sk_memory_allocated_add(struct sock *sk, int amt, int *parent_status)
1119 1120
{
	struct proto *prot = sk->sk_prot;
G
Glauber Costa 已提交
1121 1122 1123 1124 1125 1126 1127 1128

	if (mem_cgroup_sockets_enabled && sk->sk_cgrp) {
		memcg_memory_allocated_add(sk->sk_cgrp, amt, parent_status);
		/* update the root cgroup regardless */
		atomic_long_add_return(amt, prot->memory_allocated);
		return memcg_memory_allocated_read(sk->sk_cgrp);
	}

1129 1130 1131 1132
	return atomic_long_add_return(amt, prot->memory_allocated);
}

static inline void
1133
sk_memory_allocated_sub(struct sock *sk, int amt)
1134 1135
{
	struct proto *prot = sk->sk_prot;
G
Glauber Costa 已提交
1136

1137
	if (mem_cgroup_sockets_enabled && sk->sk_cgrp)
G
Glauber Costa 已提交
1138 1139
		memcg_memory_allocated_sub(sk->sk_cgrp, amt);

1140 1141 1142 1143 1144 1145
	atomic_long_sub(amt, prot->memory_allocated);
}

static inline void sk_sockets_allocated_dec(struct sock *sk)
{
	struct proto *prot = sk->sk_prot;
G
Glauber Costa 已提交
1146 1147 1148 1149 1150 1151 1152 1153

	if (mem_cgroup_sockets_enabled && sk->sk_cgrp) {
		struct cg_proto *cg_proto = sk->sk_cgrp;

		for (; cg_proto; cg_proto = parent_cg_proto(prot, cg_proto))
			percpu_counter_dec(cg_proto->sockets_allocated);
	}

1154 1155 1156 1157 1158 1159
	percpu_counter_dec(prot->sockets_allocated);
}

static inline void sk_sockets_allocated_inc(struct sock *sk)
{
	struct proto *prot = sk->sk_prot;
G
Glauber Costa 已提交
1160 1161 1162 1163 1164 1165 1166 1167

	if (mem_cgroup_sockets_enabled && sk->sk_cgrp) {
		struct cg_proto *cg_proto = sk->sk_cgrp;

		for (; cg_proto; cg_proto = parent_cg_proto(prot, cg_proto))
			percpu_counter_inc(cg_proto->sockets_allocated);
	}

1168 1169 1170 1171 1172 1173 1174 1175
	percpu_counter_inc(prot->sockets_allocated);
}

static inline int
sk_sockets_allocated_read_positive(struct sock *sk)
{
	struct proto *prot = sk->sk_prot;

G
Glauber Costa 已提交
1176
	if (mem_cgroup_sockets_enabled && sk->sk_cgrp)
1177
		return percpu_counter_read_positive(sk->sk_cgrp->sockets_allocated);
G
Glauber Costa 已提交
1178

1179
	return percpu_counter_read_positive(prot->sockets_allocated);
1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201
}

static inline int
proto_sockets_allocated_sum_positive(struct proto *prot)
{
	return percpu_counter_sum_positive(prot->sockets_allocated);
}

static inline long
proto_memory_allocated(struct proto *prot)
{
	return atomic_long_read(prot->memory_allocated);
}

static inline bool
proto_memory_pressure(struct proto *prot)
{
	if (!prot->memory_pressure)
		return false;
	return !!*prot->memory_pressure;
}

1202 1203

#ifdef CONFIG_PROC_FS
L
Linus Torvalds 已提交
1204
/* Called with local bh disabled */
1205 1206
extern void sock_prot_inuse_add(struct net *net, struct proto *prot, int inc);
extern int sock_prot_inuse_get(struct net *net, struct proto *proto);
1207
#else
E
Eric Dumazet 已提交
1208
static inline void sock_prot_inuse_add(struct net *net, struct proto *prot,
1209
		int inc)
1210 1211 1212 1213
{
}
#endif

L
Linus Torvalds 已提交
1214

1215 1216 1217 1218 1219 1220 1221 1222 1223
/* With per-bucket locks this operation is not-atomic, so that
 * this version is not worse.
 */
static inline void __sk_prot_rehash(struct sock *sk)
{
	sk->sk_prot->unhash(sk);
	sk->sk_prot->hash(sk);
}

1224 1225
void sk_prot_clear_portaddr_nulls(struct sock *sk, int size);

L
Linus Torvalds 已提交
1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278
/* About 10 seconds */
#define SOCK_DESTROY_TIME (10*HZ)

/* Sockets 0-1023 can't be bound to unless you are superuser */
#define PROT_SOCK	1024

#define SHUTDOWN_MASK	3
#define RCV_SHUTDOWN	1
#define SEND_SHUTDOWN	2

#define SOCK_SNDBUF_LOCK	1
#define SOCK_RCVBUF_LOCK	2
#define SOCK_BINDADDR_LOCK	4
#define SOCK_BINDPORT_LOCK	8

/* sock_iocb: used to kick off async processing of socket ios */
struct sock_iocb {
	struct list_head	list;

	int			flags;
	int			size;
	struct socket		*sock;
	struct sock		*sk;
	struct scm_cookie	*scm;
	struct msghdr		*msg, async_msg;
	struct kiocb		*kiocb;
};

static inline struct sock_iocb *kiocb_to_siocb(struct kiocb *iocb)
{
	return (struct sock_iocb *)iocb->private;
}

static inline struct kiocb *siocb_to_kiocb(struct sock_iocb *si)
{
	return si->kiocb;
}

struct socket_alloc {
	struct socket socket;
	struct inode vfs_inode;
};

static inline struct socket *SOCKET_I(struct inode *inode)
{
	return &container_of(inode, struct socket_alloc, vfs_inode)->socket;
}

static inline struct inode *SOCK_INODE(struct socket *socket)
{
	return &container_of(socket, struct socket_alloc, socket)->vfs_inode;
}

1279 1280 1281 1282 1283
/*
 * Functions for memory accounting
 */
extern int __sk_mem_schedule(struct sock *sk, int size, int kind);
extern void __sk_mem_reclaim(struct sock *sk);
L
Linus Torvalds 已提交
1284

1285 1286 1287 1288
#define SK_MEM_QUANTUM ((int)PAGE_SIZE)
#define SK_MEM_QUANTUM_SHIFT ilog2(SK_MEM_QUANTUM)
#define SK_MEM_SEND	0
#define SK_MEM_RECV	1
L
Linus Torvalds 已提交
1289

1290
static inline int sk_mem_pages(int amt)
L
Linus Torvalds 已提交
1291
{
1292
	return (amt + SK_MEM_QUANTUM - 1) >> SK_MEM_QUANTUM_SHIFT;
L
Linus Torvalds 已提交
1293 1294
}

E
Eric Dumazet 已提交
1295
static inline bool sk_has_account(struct sock *sk)
L
Linus Torvalds 已提交
1296
{
1297 1298
	/* return true if protocol supports memory accounting */
	return !!sk->sk_prot->memory_allocated;
L
Linus Torvalds 已提交
1299 1300
}

E
Eric Dumazet 已提交
1301
static inline bool sk_wmem_schedule(struct sock *sk, int size)
L
Linus Torvalds 已提交
1302
{
1303
	if (!sk_has_account(sk))
E
Eric Dumazet 已提交
1304
		return true;
1305 1306
	return size <= sk->sk_forward_alloc ||
		__sk_mem_schedule(sk, size, SK_MEM_SEND);
L
Linus Torvalds 已提交
1307 1308
}

E
Eric Dumazet 已提交
1309
static inline bool sk_rmem_schedule(struct sock *sk, int size)
H
Herbert Xu 已提交
1310
{
1311
	if (!sk_has_account(sk))
E
Eric Dumazet 已提交
1312
		return true;
H
Herbert Xu 已提交
1313
	return size <= sk->sk_forward_alloc ||
1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324
		__sk_mem_schedule(sk, size, SK_MEM_RECV);
}

static inline void sk_mem_reclaim(struct sock *sk)
{
	if (!sk_has_account(sk))
		return;
	if (sk->sk_forward_alloc >= SK_MEM_QUANTUM)
		__sk_mem_reclaim(sk);
}

1325 1326 1327 1328 1329 1330 1331 1332
static inline void sk_mem_reclaim_partial(struct sock *sk)
{
	if (!sk_has_account(sk))
		return;
	if (sk->sk_forward_alloc > SK_MEM_QUANTUM)
		__sk_mem_reclaim(sk);
}

1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352
static inline void sk_mem_charge(struct sock *sk, int size)
{
	if (!sk_has_account(sk))
		return;
	sk->sk_forward_alloc -= size;
}

static inline void sk_mem_uncharge(struct sock *sk, int size)
{
	if (!sk_has_account(sk))
		return;
	sk->sk_forward_alloc += size;
}

static inline void sk_wmem_free_skb(struct sock *sk, struct sk_buff *skb)
{
	sock_set_flag(sk, SOCK_QUEUE_SHRUNK);
	sk->sk_wmem_queued -= skb->truesize;
	sk_mem_uncharge(sk, skb->truesize);
	__kfree_skb(skb);
H
Herbert Xu 已提交
1353 1354
}

L
Linus Torvalds 已提交
1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367
/* Used by processes to "lock" a socket state, so that
 * interrupts and bottom half handlers won't change it
 * from under us. It essentially blocks any incoming
 * packets, so that we won't get any new data or any
 * packets that change the state of the socket.
 *
 * While locked, BH processing will add new packets to
 * the backlog queue.  This queue is processed by the
 * owner of the socket lock right before it is released.
 *
 * Since ~2.3.5 it is also exclusive sleep lock serializing
 * accesses from user process context.
 */
1368
#define sock_owned_by_user(sk)	((sk)->sk_lock.owned)
L
Linus Torvalds 已提交
1369

1370 1371 1372 1373 1374 1375 1376
/*
 * Macro so as to not evaluate some arguments when
 * lockdep is not enabled.
 *
 * Mark both the sk_lock and the sk_lock.slock as a
 * per-address-family lock class.
 */
E
Eric Dumazet 已提交
1377
#define sock_lock_init_class_and_name(sk, sname, skey, name, key)	\
1378
do {									\
1379
	sk->sk_lock.owned = 0;						\
1380 1381 1382 1383 1384
	init_waitqueue_head(&sk->sk_lock.wq);				\
	spin_lock_init(&(sk)->sk_lock.slock);				\
	debug_check_no_locks_freed((void *)&(sk)->sk_lock,		\
			sizeof((sk)->sk_lock));				\
	lockdep_set_class_and_name(&(sk)->sk_lock.slock,		\
E
Eric Dumazet 已提交
1385
				(skey), (sname));				\
1386 1387 1388
	lockdep_init_map(&(sk)->sk_lock.dep_map, (name), (key), 0);	\
} while (0)

H
Harvey Harrison 已提交
1389
extern void lock_sock_nested(struct sock *sk, int subclass);
1390 1391 1392 1393 1394 1395

static inline void lock_sock(struct sock *sk)
{
	lock_sock_nested(sk, 0);
}

H
Harvey Harrison 已提交
1396
extern void release_sock(struct sock *sk);
L
Linus Torvalds 已提交
1397 1398 1399

/* BH context may only use the following locking interface. */
#define bh_lock_sock(__sk)	spin_lock(&((__sk)->sk_lock.slock))
1400 1401 1402
#define bh_lock_sock_nested(__sk) \
				spin_lock_nested(&((__sk)->sk_lock.slock), \
				SINGLE_DEPTH_NESTING)
L
Linus Torvalds 已提交
1403 1404
#define bh_unlock_sock(__sk)	spin_unlock(&((__sk)->sk_lock.slock))

1405 1406 1407 1408 1409 1410 1411 1412 1413 1414
extern bool lock_sock_fast(struct sock *sk);
/**
 * unlock_sock_fast - complement of lock_sock_fast
 * @sk: socket
 * @slow: slow mode
 *
 * fast unlock socket for user context.
 * If slow mode is on, we call regular release_sock()
 */
static inline void unlock_sock_fast(struct sock *sk, bool slow)
E
Eric Dumazet 已提交
1415
{
1416 1417 1418 1419
	if (slow)
		release_sock(sk);
	else
		spin_unlock_bh(&sk->sk_lock.slock);
E
Eric Dumazet 已提交
1420 1421 1422
}


1423
extern struct sock		*sk_alloc(struct net *net, int family,
A
Al Viro 已提交
1424
					  gfp_t priority,
1425
					  struct proto *prot);
L
Linus Torvalds 已提交
1426
extern void			sk_free(struct sock *sk);
1427
extern void			sk_release_kernel(struct sock *sk);
1428 1429
extern struct sock		*sk_clone_lock(const struct sock *sk,
					       const gfp_t priority);
L
Linus Torvalds 已提交
1430 1431 1432

extern struct sk_buff		*sock_wmalloc(struct sock *sk,
					      unsigned long size, int force,
A
Al Viro 已提交
1433
					      gfp_t priority);
L
Linus Torvalds 已提交
1434 1435
extern struct sk_buff		*sock_rmalloc(struct sock *sk,
					      unsigned long size, int force,
A
Al Viro 已提交
1436
					      gfp_t priority);
L
Linus Torvalds 已提交
1437 1438
extern void			sock_wfree(struct sk_buff *skb);
extern void			sock_rfree(struct sk_buff *skb);
D
David S. Miller 已提交
1439
extern void			sock_edemux(struct sk_buff *skb);
L
Linus Torvalds 已提交
1440 1441 1442

extern int			sock_setsockopt(struct socket *sock, int level,
						int op, char __user *optval,
1443
						unsigned int optlen);
L
Linus Torvalds 已提交
1444 1445

extern int			sock_getsockopt(struct socket *sock, int level,
E
Eric Dumazet 已提交
1446
						int op, char __user *optval,
L
Linus Torvalds 已提交
1447
						int __user *optlen);
E
Eric Dumazet 已提交
1448
extern struct sk_buff		*sock_alloc_send_skb(struct sock *sk,
L
Linus Torvalds 已提交
1449 1450 1451
						     unsigned long size,
						     int noblock,
						     int *errcode);
E
Eric Dumazet 已提交
1452
extern struct sk_buff		*sock_alloc_send_pskb(struct sock *sk,
H
Herbert Xu 已提交
1453 1454 1455 1456
						      unsigned long header_len,
						      unsigned long data_len,
						      int noblock,
						      int *errcode);
V
Victor Fusco 已提交
1457
extern void *sock_kmalloc(struct sock *sk, int size,
A
Al Viro 已提交
1458
			  gfp_t priority);
L
Linus Torvalds 已提交
1459 1460 1461
extern void sock_kfree_s(struct sock *sk, void *mem, int size);
extern void sk_send_sigurg(struct sock *sk);

1462 1463 1464 1465 1466 1467 1468 1469
#ifdef CONFIG_CGROUPS
extern void sock_update_classid(struct sock *sk);
#else
static inline void sock_update_classid(struct sock *sk)
{
}
#endif

L
Linus Torvalds 已提交
1470 1471 1472 1473
/*
 * Functions to fill in entries in struct proto_ops when a protocol
 * does not implement a particular function.
 */
E
Eric Dumazet 已提交
1474
extern int                      sock_no_bind(struct socket *,
L
Linus Torvalds 已提交
1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492
					     struct sockaddr *, int);
extern int                      sock_no_connect(struct socket *,
						struct sockaddr *, int, int);
extern int                      sock_no_socketpair(struct socket *,
						   struct socket *);
extern int                      sock_no_accept(struct socket *,
					       struct socket *, int);
extern int                      sock_no_getname(struct socket *,
						struct sockaddr *, int *, int);
extern unsigned int             sock_no_poll(struct file *, struct socket *,
					     struct poll_table_struct *);
extern int                      sock_no_ioctl(struct socket *, unsigned int,
					      unsigned long);
extern int			sock_no_listen(struct socket *, int);
extern int                      sock_no_shutdown(struct socket *, int);
extern int			sock_no_getsockopt(struct socket *, int , int,
						   char __user *, int __user *);
extern int			sock_no_setsockopt(struct socket *, int, int,
1493
						   char __user *, unsigned int);
L
Linus Torvalds 已提交
1494 1495 1496 1497 1498 1499 1500 1501 1502
extern int                      sock_no_sendmsg(struct kiocb *, struct socket *,
						struct msghdr *, size_t);
extern int                      sock_no_recvmsg(struct kiocb *, struct socket *,
						struct msghdr *, size_t, int);
extern int			sock_no_mmap(struct file *file,
					     struct socket *sock,
					     struct vm_area_struct *vma);
extern ssize_t			sock_no_sendpage(struct socket *sock,
						struct page *page,
E
Eric Dumazet 已提交
1503
						int offset, size_t size,
L
Linus Torvalds 已提交
1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514
						int flags);

/*
 * Functions to fill in entries in struct proto_ops when a protocol
 * uses the inet style.
 */
extern int sock_common_getsockopt(struct socket *sock, int level, int optname,
				  char __user *optval, int __user *optlen);
extern int sock_common_recvmsg(struct kiocb *iocb, struct socket *sock,
			       struct msghdr *msg, size_t size, int flags);
extern int sock_common_setsockopt(struct socket *sock, int level, int optname,
1515
				  char __user *optval, unsigned int optlen);
1516 1517 1518
extern int compat_sock_common_getsockopt(struct socket *sock, int level,
		int optname, char __user *optval, int __user *optlen);
extern int compat_sock_common_setsockopt(struct socket *sock, int level,
1519
		int optname, char __user *optval, unsigned int optlen);
L
Linus Torvalds 已提交
1520 1521 1522 1523 1524 1525

extern void sk_common_release(struct sock *sk);

/*
 *	Default socket callbacks and setup code
 */
E
Eric Dumazet 已提交
1526

L
Linus Torvalds 已提交
1527 1528 1529
/* Initialise core socket variables */
extern void sock_init_data(struct socket *sock, struct sock *sk);

E
Eric Dumazet 已提交
1530 1531
extern void sk_filter_release_rcu(struct rcu_head *rcu);

1532
/**
1533
 *	sk_filter_release - release a socket filter
1534 1535 1536 1537 1538
 *	@fp: filter to remove
 *
 *	Remove a filter from a socket and release its resources.
 */

1539 1540 1541
static inline void sk_filter_release(struct sk_filter *fp)
{
	if (atomic_dec_and_test(&fp->refcnt))
1542
		call_rcu(&fp->rcu, sk_filter_release_rcu);
1543 1544 1545
}

static inline void sk_filter_uncharge(struct sock *sk, struct sk_filter *fp)
L
Linus Torvalds 已提交
1546 1547 1548 1549
{
	unsigned int size = sk_filter_len(fp);

	atomic_sub(size, &sk->sk_omem_alloc);
1550
	sk_filter_release(fp);
L
Linus Torvalds 已提交
1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590
}

static inline void sk_filter_charge(struct sock *sk, struct sk_filter *fp)
{
	atomic_inc(&fp->refcnt);
	atomic_add(sk_filter_len(fp), &sk->sk_omem_alloc);
}

/*
 * Socket reference counting postulates.
 *
 * * Each user of socket SHOULD hold a reference count.
 * * Each access point to socket (an hash table bucket, reference from a list,
 *   running timer, skb in flight MUST hold a reference count.
 * * When reference count hits 0, it means it will never increase back.
 * * When reference count hits 0, it means that no references from
 *   outside exist to this socket and current process on current CPU
 *   is last user and may/should destroy this socket.
 * * sk_free is called from any context: process, BH, IRQ. When
 *   it is called, socket has no references from outside -> sk_free
 *   may release descendant resources allocated by the socket, but
 *   to the time when it is called, socket is NOT referenced by any
 *   hash tables, lists etc.
 * * Packets, delivered from outside (from network or from another process)
 *   and enqueued on receive/error queues SHOULD NOT grab reference count,
 *   when they sit in queue. Otherwise, packets will leak to hole, when
 *   socket is looked up by one cpu and unhasing is made by another CPU.
 *   It is true for udp/raw, netlink (leak to receive and error queues), tcp
 *   (leak to backlog). Packet socket does all the processing inside
 *   BR_NETPROTO_LOCK, so that it has not this race condition. UNIX sockets
 *   use separate SMP lock, so that they are prone too.
 */

/* Ungrab socket and destroy it, if it was the last reference. */
static inline void sock_put(struct sock *sk)
{
	if (atomic_dec_and_test(&sk->sk_refcnt))
		sk_free(sk);
}

1591 1592
extern int sk_receive_skb(struct sock *sk, struct sk_buff *skb,
			  const int nested);
1593

K
Krishna Kumar 已提交
1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605
static inline void sk_tx_queue_set(struct sock *sk, int tx_queue)
{
	sk->sk_tx_queue_mapping = tx_queue;
}

static inline void sk_tx_queue_clear(struct sock *sk)
{
	sk->sk_tx_queue_mapping = -1;
}

static inline int sk_tx_queue_get(const struct sock *sk)
{
1606
	return sk ? sk->sk_tx_queue_mapping : -1;
K
Krishna Kumar 已提交
1607 1608
}

1609 1610
static inline void sk_set_socket(struct sock *sk, struct socket *sock)
{
K
Krishna Kumar 已提交
1611
	sk_tx_queue_clear(sk);
1612 1613 1614
	sk->sk_socket = sock;
}

E
Eric Dumazet 已提交
1615 1616
static inline wait_queue_head_t *sk_sleep(struct sock *sk)
{
1617 1618
	BUILD_BUG_ON(offsetof(struct socket_wq, wait) != 0);
	return &rcu_dereference_raw(sk->sk_wq)->wait;
E
Eric Dumazet 已提交
1619
}
L
Linus Torvalds 已提交
1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630
/* Detach socket from process context.
 * Announce socket dead, detach it from wait queue and inode.
 * Note that parent inode held reference count on this struct sock,
 * we do not release it in this function, because protocol
 * probably wants some additional cleanups or even continuing
 * to work with this socket (TCP).
 */
static inline void sock_orphan(struct sock *sk)
{
	write_lock_bh(&sk->sk_callback_lock);
	sock_set_flag(sk, SOCK_DEAD);
1631
	sk_set_socket(sk, NULL);
1632
	sk->sk_wq  = NULL;
L
Linus Torvalds 已提交
1633 1634 1635 1636 1637 1638
	write_unlock_bh(&sk->sk_callback_lock);
}

static inline void sock_graft(struct sock *sk, struct socket *parent)
{
	write_lock_bh(&sk->sk_callback_lock);
1639
	sk->sk_wq = parent->wq;
L
Linus Torvalds 已提交
1640
	parent->sk = sk;
1641
	sk_set_socket(sk, parent);
1642
	security_sock_graft(sk, parent);
L
Linus Torvalds 已提交
1643 1644 1645 1646 1647 1648 1649 1650 1651
	write_unlock_bh(&sk->sk_callback_lock);
}

extern int sock_i_uid(struct sock *sk);
extern unsigned long sock_i_ino(struct sock *sk);

static inline struct dst_entry *
__sk_dst_get(struct sock *sk)
{
1652
	return rcu_dereference_check(sk->sk_dst_cache, sock_owned_by_user(sk) ||
E
Eric Dumazet 已提交
1653
						       lockdep_is_held(&sk->sk_lock.slock));
L
Linus Torvalds 已提交
1654 1655 1656 1657 1658 1659 1660
}

static inline struct dst_entry *
sk_dst_get(struct sock *sk)
{
	struct dst_entry *dst;

E
Eric Dumazet 已提交
1661 1662
	rcu_read_lock();
	dst = rcu_dereference(sk->sk_dst_cache);
L
Linus Torvalds 已提交
1663 1664
	if (dst)
		dst_hold(dst);
E
Eric Dumazet 已提交
1665
	rcu_read_unlock();
L
Linus Torvalds 已提交
1666 1667 1668
	return dst;
}

E
Eric Dumazet 已提交
1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684
extern void sk_reset_txq(struct sock *sk);

static inline void dst_negative_advice(struct sock *sk)
{
	struct dst_entry *ndst, *dst = __sk_dst_get(sk);

	if (dst && dst->ops->negative_advice) {
		ndst = dst->ops->negative_advice(dst);

		if (ndst != dst) {
			rcu_assign_pointer(sk->sk_dst_cache, ndst);
			sk_reset_txq(sk);
		}
	}
}

L
Linus Torvalds 已提交
1685 1686 1687 1688 1689
static inline void
__sk_dst_set(struct sock *sk, struct dst_entry *dst)
{
	struct dst_entry *old_dst;

K
Krishna Kumar 已提交
1690
	sk_tx_queue_clear(sk);
1691 1692 1693 1694 1695
	/*
	 * This can be called while sk is owned by the caller only,
	 * with no state that can be checked in a rcu_dereference_check() cond
	 */
	old_dst = rcu_dereference_raw(sk->sk_dst_cache);
E
Eric Dumazet 已提交
1696
	rcu_assign_pointer(sk->sk_dst_cache, dst);
L
Linus Torvalds 已提交
1697 1698 1699 1700 1701 1702
	dst_release(old_dst);
}

static inline void
sk_dst_set(struct sock *sk, struct dst_entry *dst)
{
E
Eric Dumazet 已提交
1703
	spin_lock(&sk->sk_dst_lock);
L
Linus Torvalds 已提交
1704
	__sk_dst_set(sk, dst);
E
Eric Dumazet 已提交
1705
	spin_unlock(&sk->sk_dst_lock);
L
Linus Torvalds 已提交
1706 1707 1708 1709 1710
}

static inline void
__sk_dst_reset(struct sock *sk)
{
E
Eric Dumazet 已提交
1711
	__sk_dst_set(sk, NULL);
L
Linus Torvalds 已提交
1712 1713 1714 1715 1716
}

static inline void
sk_dst_reset(struct sock *sk)
{
E
Eric Dumazet 已提交
1717
	spin_lock(&sk->sk_dst_lock);
L
Linus Torvalds 已提交
1718
	__sk_dst_reset(sk);
E
Eric Dumazet 已提交
1719
	spin_unlock(&sk->sk_dst_lock);
L
Linus Torvalds 已提交
1720 1721
}

1722
extern struct dst_entry *__sk_dst_check(struct sock *sk, u32 cookie);
L
Linus Torvalds 已提交
1723

1724
extern struct dst_entry *sk_dst_check(struct sock *sk, u32 cookie);
L
Linus Torvalds 已提交
1725

E
Eric Dumazet 已提交
1726
static inline bool sk_can_gso(const struct sock *sk)
1727 1728 1729 1730
{
	return net_gso_ok(sk->sk_route_caps, sk->sk_gso_type);
}

1731
extern void sk_setup_caps(struct sock *sk, struct dst_entry *dst);
1732

1733
static inline void sk_nocaps_add(struct sock *sk, netdev_features_t flags)
E
Eric Dumazet 已提交
1734 1735 1736 1737 1738
{
	sk->sk_route_nocaps |= flags;
	sk->sk_route_caps &= ~flags;
}

1739 1740
static inline int skb_do_copy_data_nocache(struct sock *sk, struct sk_buff *skb,
					   char __user *from, char *to,
1741
					   int copy, int offset)
1742 1743 1744 1745 1746 1747
{
	if (skb->ip_summed == CHECKSUM_NONE) {
		int err = 0;
		__wsum csum = csum_and_copy_from_user(from, to, copy, 0, &err);
		if (err)
			return err;
1748
		skb->csum = csum_block_add(skb->csum, csum, offset);
1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761
	} else if (sk->sk_route_caps & NETIF_F_NOCACHE_COPY) {
		if (!access_ok(VERIFY_READ, from, copy) ||
		    __copy_from_user_nocache(to, from, copy))
			return -EFAULT;
	} else if (copy_from_user(to, from, copy))
		return -EFAULT;

	return 0;
}

static inline int skb_add_data_nocache(struct sock *sk, struct sk_buff *skb,
				       char __user *from, int copy)
{
1762
	int err, offset = skb->len;
1763

1764 1765
	err = skb_do_copy_data_nocache(sk, skb, from, skb_put(skb, copy),
				       copy, offset);
1766
	if (err)
1767
		__skb_trim(skb, offset);
1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778

	return err;
}

static inline int skb_copy_to_page_nocache(struct sock *sk, char __user *from,
					   struct sk_buff *skb,
					   struct page *page,
					   int off, int copy)
{
	int err;

1779 1780
	err = skb_do_copy_data_nocache(sk, skb, from, page_address(page) + off,
				       copy, skb->len);
1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791
	if (err)
		return err;

	skb->len	     += copy;
	skb->data_len	     += copy;
	skb->truesize	     += copy;
	sk->sk_wmem_queued   += copy;
	sk_mem_charge(sk, copy);
	return 0;
}

L
Linus Torvalds 已提交
1792 1793 1794 1795 1796 1797
static inline int skb_copy_to_page(struct sock *sk, char __user *from,
				   struct sk_buff *skb, struct page *page,
				   int off, int copy)
{
	if (skb->ip_summed == CHECKSUM_NONE) {
		int err = 0;
1798
		__wsum csum = csum_and_copy_from_user(from,
L
Linus Torvalds 已提交
1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810
						     page_address(page) + off,
							    copy, 0, &err);
		if (err)
			return err;
		skb->csum = csum_block_add(skb->csum, csum, skb->len);
	} else if (copy_from_user(page_address(page) + off, from, copy))
		return -EFAULT;

	skb->len	     += copy;
	skb->data_len	     += copy;
	skb->truesize	     += copy;
	sk->sk_wmem_queued   += copy;
1811
	sk_mem_charge(sk, copy);
L
Linus Torvalds 已提交
1812 1813 1814
	return 0;
}

1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842
/**
 * sk_wmem_alloc_get - returns write allocations
 * @sk: socket
 *
 * Returns sk_wmem_alloc minus initial offset of one
 */
static inline int sk_wmem_alloc_get(const struct sock *sk)
{
	return atomic_read(&sk->sk_wmem_alloc) - 1;
}

/**
 * sk_rmem_alloc_get - returns read allocations
 * @sk: socket
 *
 * Returns sk_rmem_alloc
 */
static inline int sk_rmem_alloc_get(const struct sock *sk)
{
	return atomic_read(&sk->sk_rmem_alloc);
}

/**
 * sk_has_allocations - check if allocations are outstanding
 * @sk: socket
 *
 * Returns true if socket has write or read allocations
 */
E
Eric Dumazet 已提交
1843
static inline bool sk_has_allocations(const struct sock *sk)
1844 1845 1846 1847
{
	return sk_wmem_alloc_get(sk) || sk_rmem_alloc_get(sk);
}

1848
/**
1849
 * wq_has_sleeper - check if there are any waiting processes
R
Randy Dunlap 已提交
1850
 * @wq: struct socket_wq
1851
 *
1852
 * Returns true if socket_wq has waiting processes
1853
 *
1854
 * The purpose of the wq_has_sleeper and sock_poll_wait is to wrap the memory
1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866
 * barrier call. They were added due to the race found within the tcp code.
 *
 * Consider following tcp code paths:
 *
 * CPU1                  CPU2
 *
 * sys_select            receive packet
 *   ...                 ...
 *   __add_wait_queue    update tp->rcv_nxt
 *   ...                 ...
 *   tp->rcv_nxt check   sock_def_readable
 *   ...                 {
1867 1868 1869 1870
 *   schedule               rcu_read_lock();
 *                          wq = rcu_dereference(sk->sk_wq);
 *                          if (wq && waitqueue_active(&wq->wait))
 *                              wake_up_interruptible(&wq->wait)
1871 1872 1873 1874 1875 1876 1877
 *                          ...
 *                       }
 *
 * The race for tcp fires when the __add_wait_queue changes done by CPU1 stay
 * in its cache, and so does the tp->rcv_nxt update on CPU2 side.  The CPU1
 * could then endup calling schedule and sleep forever if there are no more
 * data on the socket.
1878
 *
1879
 */
1880
static inline bool wq_has_sleeper(struct socket_wq *wq)
1881
{
E
Eric Dumazet 已提交
1882
	/* We need to be sure we are in sync with the
1883 1884 1885 1886
	 * add_wait_queue modifications to the wait queue.
	 *
	 * This memory barrier is paired in the sock_poll_wait.
	 */
1887 1888
	smp_mb();
	return wq && waitqueue_active(&wq->wait);
1889 1890 1891 1892 1893 1894 1895 1896
}

/**
 * sock_poll_wait - place memory barrier behind the poll_wait call.
 * @filp:           file
 * @wait_address:   socket wait queue
 * @p:              poll_table
 *
1897
 * See the comments in the wq_has_sleeper function.
1898 1899 1900 1901
 */
static inline void sock_poll_wait(struct file *filp,
		wait_queue_head_t *wait_address, poll_table *p)
{
1902
	if (!poll_does_not_wait(p) && wait_address) {
1903
		poll_wait(filp, wait_address, p);
E
Eric Dumazet 已提交
1904
		/* We need to be sure we are in sync with the
1905 1906
		 * socket flags modification.
		 *
1907
		 * This memory barrier is paired in the wq_has_sleeper.
E
Eric Dumazet 已提交
1908
		 */
1909 1910 1911 1912
		smp_mb();
	}
}

L
Linus Torvalds 已提交
1913
/*
E
Eric Dumazet 已提交
1914
 *	Queue a received datagram if it will fit. Stream and sequenced
L
Linus Torvalds 已提交
1915 1916 1917
 *	protocols can't normally use this as they need to fit buffers in
 *	and play with them.
 *
E
Eric Dumazet 已提交
1918
 *	Inlined as it's very short and called for pretty much every
L
Linus Torvalds 已提交
1919 1920 1921 1922 1923
 *	packet ever received.
 */

static inline void skb_set_owner_w(struct sk_buff *skb, struct sock *sk)
{
1924
	skb_orphan(skb);
L
Linus Torvalds 已提交
1925 1926
	skb->sk = sk;
	skb->destructor = sock_wfree;
1927 1928 1929 1930 1931
	/*
	 * We used to take a refcount on sk, but following operation
	 * is enough to guarantee sk_free() wont free this sock until
	 * all in-flight packets are completed
	 */
L
Linus Torvalds 已提交
1932 1933 1934 1935 1936
	atomic_add(skb->truesize, &sk->sk_wmem_alloc);
}

static inline void skb_set_owner_r(struct sk_buff *skb, struct sock *sk)
{
1937
	skb_orphan(skb);
L
Linus Torvalds 已提交
1938 1939 1940
	skb->sk = sk;
	skb->destructor = sock_rfree;
	atomic_add(skb->truesize, &sk->sk_rmem_alloc);
1941
	sk_mem_charge(sk, skb->truesize);
L
Linus Torvalds 已提交
1942 1943
}

E
Eric Dumazet 已提交
1944
extern void sk_reset_timer(struct sock *sk, struct timer_list *timer,
L
Linus Torvalds 已提交
1945 1946
			   unsigned long expires);

E
Eric Dumazet 已提交
1947
extern void sk_stop_timer(struct sock *sk, struct timer_list *timer);
L
Linus Torvalds 已提交
1948

1949
extern int sock_queue_rcv_skb(struct sock *sk, struct sk_buff *skb);
L
Linus Torvalds 已提交
1950

1951
extern int sock_queue_err_skb(struct sock *sk, struct sk_buff *skb);
L
Linus Torvalds 已提交
1952 1953 1954 1955

/*
 *	Recover an error report and clear atomically
 */
E
Eric Dumazet 已提交
1956

L
Linus Torvalds 已提交
1957 1958
static inline int sock_error(struct sock *sk)
{
1959 1960 1961 1962
	int err;
	if (likely(!sk->sk_err))
		return 0;
	err = xchg(&sk->sk_err, 0);
L
Linus Torvalds 已提交
1963 1964 1965 1966 1967 1968 1969 1970 1971
	return -err;
}

static inline unsigned long sock_wspace(struct sock *sk)
{
	int amt = 0;

	if (!(sk->sk_shutdown & SEND_SHUTDOWN)) {
		amt = sk->sk_sndbuf - atomic_read(&sk->sk_wmem_alloc);
E
Eric Dumazet 已提交
1972
		if (amt < 0)
L
Linus Torvalds 已提交
1973 1974 1975 1976 1977 1978 1979
			amt = 0;
	}
	return amt;
}

static inline void sk_wake_async(struct sock *sk, int how, int band)
{
E
Eric Dumazet 已提交
1980
	if (sock_flag(sk, SOCK_FASYNC))
L
Linus Torvalds 已提交
1981 1982 1983 1984
		sock_wake_async(sk->sk_socket, how, band);
}

#define SOCK_MIN_SNDBUF 2048
E
Eric Dumazet 已提交
1985 1986 1987 1988 1989
/*
 * Since sk_rmem_alloc sums skb->truesize, even a small frame might need
 * sizeof(sk_buff) + MTU + padding, unless net driver perform copybreak
 */
#define SOCK_MIN_RCVBUF (2048 + sizeof(struct sk_buff))
L
Linus Torvalds 已提交
1990 1991 1992 1993

static inline void sk_stream_moderate_sndbuf(struct sock *sk)
{
	if (!(sk->sk_userlocks & SOCK_SNDBUF_LOCK)) {
1994
		sk->sk_sndbuf = min(sk->sk_sndbuf, sk->sk_wmem_queued >> 1);
L
Linus Torvalds 已提交
1995 1996 1997 1998
		sk->sk_sndbuf = max(sk->sk_sndbuf, SOCK_MIN_SNDBUF);
	}
}

1999
struct sk_buff *sk_stream_alloc_skb(struct sock *sk, int size, gfp_t gfp);
L
Linus Torvalds 已提交
2000 2001 2002 2003 2004

static inline struct page *sk_stream_alloc_page(struct sock *sk)
{
	struct page *page = NULL;

2005 2006
	page = alloc_pages(sk->sk_allocation, 0);
	if (!page) {
2007
		sk_enter_memory_pressure(sk);
L
Linus Torvalds 已提交
2008 2009 2010 2011 2012 2013 2014 2015
		sk_stream_moderate_sndbuf(sk);
	}
	return page;
}

/*
 *	Default write policy as shown to user space via poll/select/SIGIO
 */
E
Eric Dumazet 已提交
2016
static inline bool sock_writeable(const struct sock *sk)
L
Linus Torvalds 已提交
2017
{
2018
	return atomic_read(&sk->sk_wmem_alloc) < (sk->sk_sndbuf >> 1);
L
Linus Torvalds 已提交
2019 2020
}

A
Al Viro 已提交
2021
static inline gfp_t gfp_any(void)
L
Linus Torvalds 已提交
2022
{
2023
	return in_softirq() ? GFP_ATOMIC : GFP_KERNEL;
L
Linus Torvalds 已提交
2024 2025
}

E
Eric Dumazet 已提交
2026
static inline long sock_rcvtimeo(const struct sock *sk, bool noblock)
L
Linus Torvalds 已提交
2027 2028 2029 2030
{
	return noblock ? 0 : sk->sk_rcvtimeo;
}

E
Eric Dumazet 已提交
2031
static inline long sock_sndtimeo(const struct sock *sk, bool noblock)
L
Linus Torvalds 已提交
2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048
{
	return noblock ? 0 : sk->sk_sndtimeo;
}

static inline int sock_rcvlowat(const struct sock *sk, int waitall, int len)
{
	return (waitall ? len : min_t(int, sk->sk_rcvlowat, len)) ? : 1;
}

/* Alas, with timeout socket operations are not restartable.
 * Compare this to poll().
 */
static inline int sock_intr_errno(long timeo)
{
	return timeo == MAX_SCHEDULE_TIMEOUT ? -ERESTARTSYS : -EINTR;
}

2049 2050
extern void __sock_recv_timestamp(struct msghdr *msg, struct sock *sk,
	struct sk_buff *skb);
2051 2052
extern void __sock_recv_wifi_status(struct msghdr *msg, struct sock *sk,
	struct sk_buff *skb);
2053

E
Eric Dumazet 已提交
2054
static inline void
L
Linus Torvalds 已提交
2055 2056
sock_recv_timestamp(struct msghdr *msg, struct sock *sk, struct sk_buff *skb)
{
2057
	ktime_t kt = skb->tstamp;
2058
	struct skb_shared_hwtstamps *hwtstamps = skb_hwtstamps(skb);
2059

2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076
	/*
	 * generate control messages if
	 * - receive time stamping in software requested (SOCK_RCVTSTAMP
	 *   or SOCK_TIMESTAMPING_RX_SOFTWARE)
	 * - software time stamp available and wanted
	 *   (SOCK_TIMESTAMPING_SOFTWARE)
	 * - hardware time stamps available and wanted
	 *   (SOCK_TIMESTAMPING_SYS_HARDWARE or
	 *   SOCK_TIMESTAMPING_RAW_HARDWARE)
	 */
	if (sock_flag(sk, SOCK_RCVTSTAMP) ||
	    sock_flag(sk, SOCK_TIMESTAMPING_RX_SOFTWARE) ||
	    (kt.tv64 && sock_flag(sk, SOCK_TIMESTAMPING_SOFTWARE)) ||
	    (hwtstamps->hwtstamp.tv64 &&
	     sock_flag(sk, SOCK_TIMESTAMPING_RAW_HARDWARE)) ||
	    (hwtstamps->syststamp.tv64 &&
	     sock_flag(sk, SOCK_TIMESTAMPING_SYS_HARDWARE)))
2077 2078
		__sock_recv_timestamp(msg, sk, skb);
	else
2079
		sk->sk_stamp = kt;
2080 2081 2082

	if (sock_flag(sk, SOCK_WIFI_STATUS) && skb->wifi_acked_valid)
		__sock_recv_wifi_status(msg, sk, skb);
L
Linus Torvalds 已提交
2083 2084
}

2085 2086 2087 2088 2089 2090 2091 2092 2093 2094
extern void __sock_recv_ts_and_drops(struct msghdr *msg, struct sock *sk,
				     struct sk_buff *skb);

static inline void sock_recv_ts_and_drops(struct msghdr *msg, struct sock *sk,
					  struct sk_buff *skb)
{
#define FLAGS_TS_OR_DROPS ((1UL << SOCK_RXQ_OVFL)			| \
			   (1UL << SOCK_RCVTSTAMP)			| \
			   (1UL << SOCK_TIMESTAMPING_RX_SOFTWARE)	| \
			   (1UL << SOCK_TIMESTAMPING_SOFTWARE)		| \
E
Eric Dumazet 已提交
2095
			   (1UL << SOCK_TIMESTAMPING_RAW_HARDWARE)	| \
2096 2097 2098 2099 2100 2101 2102
			   (1UL << SOCK_TIMESTAMPING_SYS_HARDWARE))

	if (sk->sk_flags & FLAGS_TS_OR_DROPS)
		__sock_recv_ts_and_drops(msg, sk, skb);
	else
		sk->sk_stamp = skb->tstamp;
}
2103

2104 2105 2106
/**
 * sock_tx_timestamp - checks whether the outgoing packet is to be time stamped
 * @sk:		socket sending this packet
2107
 * @tx_flags:	filled with instructions for time stamping
2108 2109 2110 2111
 *
 * Currently only depends on SOCK_TIMESTAMPING* flags. Returns error code if
 * parameters are invalid.
 */
2112
extern int sock_tx_timestamp(struct sock *sk, __u8 *tx_flags);
2113

L
Linus Torvalds 已提交
2114 2115
/**
 * sk_eat_skb - Release a skb if it is no longer needed
2116 2117
 * @sk: socket to eat this skb from
 * @skb: socket buffer to eat
R
Randy Dunlap 已提交
2118
 * @copied_early: flag indicating whether DMA operations copied this data early
L
Linus Torvalds 已提交
2119 2120 2121 2122
 *
 * This routine must be called with interrupts disabled or with the socket
 * locked so that the sk_buff queue operation is ok.
*/
2123
#ifdef CONFIG_NET_DMA
E
Eric Dumazet 已提交
2124
static inline void sk_eat_skb(struct sock *sk, struct sk_buff *skb, bool copied_early)
2125 2126 2127 2128 2129 2130 2131 2132
{
	__skb_unlink(skb, &sk->sk_receive_queue);
	if (!copied_early)
		__kfree_skb(skb);
	else
		__skb_queue_tail(&sk->sk_async_wait_queue, skb);
}
#else
E
Eric Dumazet 已提交
2133
static inline void sk_eat_skb(struct sock *sk, struct sk_buff *skb, bool copied_early)
L
Linus Torvalds 已提交
2134 2135 2136 2137
{
	__skb_unlink(skb, &sk->sk_receive_queue);
	__kfree_skb(skb);
}
2138
#endif
L
Linus Torvalds 已提交
2139

2140 2141 2142
static inline
struct net *sock_net(const struct sock *sk)
{
E
Eric Dumazet 已提交
2143
	return read_pnet(&sk->sk_net);
2144 2145 2146
}

static inline
2147
void sock_net_set(struct sock *sk, struct net *net)
2148
{
E
Eric Dumazet 已提交
2149
	write_pnet(&sk->sk_net, net);
2150 2151
}

2152 2153
/*
 * Kernel sockets, f.e. rtnl or icmp_socket, are a part of a namespace.
L
Lucas De Marchi 已提交
2154
 * They should not hold a reference to a namespace in order to allow
2155 2156 2157 2158 2159
 * to stop it.
 * Sockets after sk_change_net should be released using sk_release_kernel
 */
static inline void sk_change_net(struct sock *sk, struct net *net)
{
2160
	put_net(sock_net(sk));
2161
	sock_net_set(sk, hold_net(net));
2162 2163
}

2164 2165
static inline struct sock *skb_steal_sock(struct sk_buff *skb)
{
2166
	if (skb->sk) {
2167 2168 2169 2170 2171 2172 2173 2174 2175
		struct sock *sk = skb->sk;

		skb->destructor = NULL;
		skb->sk = NULL;
		return sk;
	}
	return NULL;
}

2176
extern void sock_enable_timestamp(struct sock *sk, int flag);
L
Linus Torvalds 已提交
2177
extern int sock_get_timestamp(struct sock *, struct timeval __user *);
2178
extern int sock_get_timestampns(struct sock *, struct timespec __user *);
L
Linus Torvalds 已提交
2179

E
Eric Dumazet 已提交
2180 2181
/*
 *	Enable debug/info messages
L
Linus Torvalds 已提交
2182
 */
2183 2184 2185
extern int net_msg_warn;
#define NETDEBUG(fmt, args...) \
	do { if (net_msg_warn) printk(fmt,##args); } while (0)
L
Linus Torvalds 已提交
2186

2187 2188
#define LIMIT_NETDEBUG(fmt, args...) \
	do { if (net_msg_warn && net_ratelimit()) printk(fmt,##args); } while(0)
L
Linus Torvalds 已提交
2189 2190 2191 2192

extern __u32 sysctl_wmem_max;
extern __u32 sysctl_rmem_max;

2193 2194
extern void sk_init(void);

2195 2196
extern int sysctl_optmem_max;

2197 2198 2199
extern __u32 sysctl_wmem_default;
extern __u32 sysctl_rmem_default;

L
Linus Torvalds 已提交
2200
#endif	/* _SOCK_H */