sock.h 71.3 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/page_counter.h>
58
#include <linux/memcontrol.h>
59
#include <linux/static_key.h>
A
Al Viro 已提交
60
#include <linux/sched.h>
H
Herbert Xu 已提交
61
#include <linux/wait.h>
62
#include <linux/cgroup-defs.h>
63
#include <linux/rbtree.h>
L
Linus Torvalds 已提交
64
#include <linux/filter.h>
65
#include <linux/rculist_nulls.h>
66
#include <linux/poll.h>
L
Linus Torvalds 已提交
67

E
Eric Dumazet 已提交
68
#include <linux/atomic.h>
69
#include <linux/refcount.h>
L
Linus Torvalds 已提交
70 71
#include <net/dst.h>
#include <net/checksum.h>
E
Eric Dumazet 已提交
72
#include <net/tcp_states.h>
73
#include <linux/net_tstamp.h>
74
#include <net/smc.h>
75
#include <net/l3mdev.h>
L
Linus Torvalds 已提交
76 77 78 79 80 81 82 83 84 85 86 87 88

/*
 * 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
89
/* Validate arguments and do nothing */
90
static inline __printf(2, 3)
E
Eric Dumazet 已提交
91
void SOCK_DEBUG(const struct sock *sk, const char *msg, ...)
92 93
{
}
L
Linus Torvalds 已提交
94 95 96 97 98 99 100 101
#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;
102
	int			owned;
L
Linus Torvalds 已提交
103
	wait_queue_head_t	wq;
I
Ingo Molnar 已提交
104 105 106 107 108 109 110 111 112
	/*
	 * 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
113 114 115
} socket_lock_t;

struct sock;
116
struct proto;
117
struct net;
L
Linus Torvalds 已提交
118

119 120 121
typedef __u32 __bitwise __portpair;
typedef __u64 __bitwise __addrpair;

L
Linus Torvalds 已提交
122
/**
123
 *	struct sock_common - minimal network layer representation of sockets
124 125
 *	@skc_daddr: Foreign IPv4 addr
 *	@skc_rcv_saddr: Bound local IPv4 addr
E
Eric Dumazet 已提交
126
 *	@skc_hash: hash value used with various protocol lookup tables
127
 *	@skc_u16hashes: two u16 hash values used by UDP lookup tables
128 129
 *	@skc_dport: placeholder for inet_dport/tw_dport
 *	@skc_num: placeholder for inet_num/tw_num
130 131 132
 *	@skc_family: network address family
 *	@skc_state: Connection state
 *	@skc_reuse: %SO_REUSEADDR setting
T
Tom Herbert 已提交
133
 *	@skc_reuseport: %SO_REUSEPORT setting
134 135
 *	@skc_bound_dev_if: bound device index if != 0
 *	@skc_bind_node: bind hash linkage for various protocol lookup tables
136
 *	@skc_portaddr_node: second hash linkage for UDP/UDP-Lite protocol
137
 *	@skc_prot: protocol handlers inside a network family
138
 *	@skc_net: reference to the network namespace of this socket
139 140 141
 *	@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
142
 *	@skc_rx_queue_mapping: rx queue number for this connection
143 144 145
 *	@skc_flags: place holder for sk_flags
 *		%SO_LINGER (l_onoff), %SO_BROADCAST, %SO_KEEPALIVE,
 *		%SO_OOBINLINE settings, %SO_TIMESTAMPING settings
146
 *	@skc_incoming_cpu: record/match cpu processing incoming packets
147
 *	@skc_refcnt: reference count
148 149
 *
 *	This is the minimal network layer representation of sockets, the header
150 151
 *	for struct sock and struct inet_timewait_sock.
 */
L
Linus Torvalds 已提交
152
struct sock_common {
153
	/* skc_daddr and skc_rcv_saddr must be grouped on a 8 bytes aligned
E
Eric Dumazet 已提交
154
	 * address on 64bit arches : cf INET_MATCH()
E
Eric Dumazet 已提交
155
	 */
156
	union {
157
		__addrpair	skc_addrpair;
158 159 160 161 162
		struct {
			__be32	skc_daddr;
			__be32	skc_rcv_saddr;
		};
	};
163 164 165 166
	union  {
		unsigned int	skc_hash;
		__u16		skc_u16hashes[2];
	};
167 168
	/* skc_dport && skc_num must be grouped as well */
	union {
169
		__portpair	skc_portpair;
170 171 172 173 174 175
		struct {
			__be16	skc_dport;
			__u16	skc_num;
		};
	};

E
Eric Dumazet 已提交
176 177
	unsigned short		skc_family;
	volatile unsigned char	skc_state;
T
Tom Herbert 已提交
178
	unsigned char		skc_reuse:4;
179 180
	unsigned char		skc_reuseport:1;
	unsigned char		skc_ipv6only:1;
181
	unsigned char		skc_net_refcnt:1;
E
Eric Dumazet 已提交
182
	int			skc_bound_dev_if;
183 184
	union {
		struct hlist_node	skc_bind_node;
185
		struct hlist_node	skc_portaddr_node;
186
	};
187
	struct proto		*skc_prot;
188
	possible_net_t		skc_net;
189 190 191 192 193 194

#if IS_ENABLED(CONFIG_IPV6)
	struct in6_addr		skc_v6_daddr;
	struct in6_addr		skc_v6_rcv_saddr;
#endif

E
Eric Dumazet 已提交
195 196
	atomic64_t		skc_cookie;

197 198 199 200 201 202 203 204 205 206
	/* following fields are padding to force
	 * offset(struct sock, sk_refcnt) == 128 on 64bit arches
	 * assuming IPV6 is enabled. We use this padding differently
	 * for different kind of 'sockets'
	 */
	union {
		unsigned long	skc_flags;
		struct sock	*skc_listener; /* request_sock */
		struct inet_timewait_death_row *skc_tw_dr; /* inet_timewait_sock */
	};
207 208 209 210
	/*
	 * fields between dontcopy_begin/dontcopy_end
	 * are not copied in sock_copy()
	 */
211
	/* private: */
212
	int			skc_dontcopy_begin[0];
213
	/* public: */
214 215 216 217
	union {
		struct hlist_node	skc_node;
		struct hlist_nulls_node skc_nulls_node;
	};
218
	unsigned short		skc_tx_queue_mapping;
219 220 221
#ifdef CONFIG_XPS
	unsigned short		skc_rx_queue_mapping;
#endif
222 223 224
	union {
		int		skc_incoming_cpu;
		u32		skc_rcv_wnd;
225
		u32		skc_tw_rcv_nxt; /* struct tcp_timewait_sock  */
226
	};
227

228
	refcount_t		skc_refcnt;
229
	/* private: */
230
	int                     skc_dontcopy_end[0];
231 232 233
	union {
		u32		skc_rxhash;
		u32		skc_window_clamp;
234
		u32		skc_tw_snd_nxt; /* struct tcp_timewait_sock */
235
	};
236
	/* public: */
L
Linus Torvalds 已提交
237 238 239 240
};

/**
  *	struct sock - network layer representation of sockets
241
  *	@__sk_common: shared layout with inet_timewait_sock
242 243 244
  *	@sk_shutdown: mask of %SEND_SHUTDOWN and/or %RCV_SHUTDOWN
  *	@sk_userlocks: %SO_SNDBUF and %SO_RCVBUF settings
  *	@sk_lock:	synchronizer
245
  *	@sk_kern_sock: True if sock is using kernel lock classes
246
  *	@sk_rcvbuf: size of receive buffer in bytes
247
  *	@sk_wq: sock wait queue and async head
S
Shawn Bohrer 已提交
248
  *	@sk_rx_dst: receive input route used by early demux
249
  *	@sk_dst_cache: destination cache
250
  *	@sk_dst_pending_confirm: need to confirm neighbour
251 252 253
  *	@sk_policy: flow policy
  *	@sk_receive_queue: incoming packets
  *	@sk_wmem_alloc: transmit queue bytes committed
254
  *	@sk_tsq_flags: TCP Small Queues flags
255 256 257 258
  *	@sk_write_queue: Packet sending queue
  *	@sk_omem_alloc: "o" is "option" or "other"
  *	@sk_wmem_queued: persistent queue size
  *	@sk_forward_alloc: space allocated forward
E
Eliezer Tamir 已提交
259
  *	@sk_napi_id: id of the last napi context to receive data for sk
260
  *	@sk_ll_usec: usecs to busypoll when there is no data
261
  *	@sk_allocation: allocation mode
E
Eric Dumazet 已提交
262
  *	@sk_pacing_rate: Pacing rate (if supported by transport/packet scheduler)
263
  *	@sk_pacing_status: Pacing status (requested, handled by sch_fq)
264
  *	@sk_max_pacing_rate: Maximum pacing rate (%SO_MAX_PACING_RATE)
265
  *	@sk_sndbuf: size of send buffer in bytes
266
  *	@__sk_flags_offset: empty field used to determine location of bitfield
267
  *	@sk_padding: unused element for alignment
268 269
  *	@sk_no_check_tx: %SO_NO_CHECK setting, set checksum in TX packets
  *	@sk_no_check_rx: allow zero checksum in RX packets
270
  *	@sk_route_caps: route capabilities (e.g. %NETIF_F_TSO)
E
Eric Dumazet 已提交
271
  *	@sk_route_nocaps: forbidden route capabilities (e.g NETIF_F_GSO_MASK)
272
  *	@sk_gso_type: GSO type (e.g. %SKB_GSO_TCPV4)
273
  *	@sk_gso_max_size: Maximum GSO segment size to build
274
  *	@sk_gso_max_segs: Maximum number of GSO segments
275
  *	@sk_pacing_shift: scaling factor for TCP Small Queues
276 277 278 279
  *	@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
280 281
  *	@sk_prot_creator: sk_prot of original sock creator (see ipv6_setsockopt,
  *			  IPV6_ADDRFORM for instance)
282
  *	@sk_err: last error
W
Wang Chen 已提交
283 284
  *	@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
285
  *	@sk_drops: raw/udp drops counter
286 287
  *	@sk_ack_backlog: current listen backlog
  *	@sk_max_ack_backlog: listen backlog set in listen()
288
  *	@sk_uid: user id of owner
289 290 291
  *	@sk_priority: %SO_PRIORITY setting
  *	@sk_type: socket type (%SOCK_STREAM, etc)
  *	@sk_protocol: which protocol this socket belongs in this network family
292 293
  *	@sk_peer_pid: &struct pid for this socket's peer
  *	@sk_peer_cred: %SO_PEERCRED setting
294 295 296
  *	@sk_rcvlowat: %SO_RCVLOWAT setting
  *	@sk_rcvtimeo: %SO_RCVTIMEO setting
  *	@sk_sndtimeo: %SO_SNDTIMEO setting
297
  *	@sk_txhash: computed flow hash for use on transmit
298 299 300
  *	@sk_filter: socket filtering instructions
  *	@sk_timer: sock cleanup timer
  *	@sk_stamp: time stamp of last packet received
301
  *	@sk_stamp_seq: lock for accessing sk_stamp on 32 bit architectures only
302
  *	@sk_tsflags: SO_TIMESTAMPING socket options
303
  *	@sk_tskey: counter to disambiguate concurrent tstamp requests
W
Willem de Bruijn 已提交
304
  *	@sk_zckey: counter to order MSG_ZEROCOPY notifications
305 306
  *	@sk_socket: Identd and reporting IO signals
  *	@sk_user_data: RPC layer private data
307
  *	@sk_frag: cached page frag
308
  *	@sk_peek_off: current peek_offset value
309
  *	@sk_send_head: front of stuff to transmit
310
  *	@sk_security: used by security modules
311
  *	@sk_mark: generic packet mark
312
  *	@sk_cgrp_data: cgroup data for this cgroup
313
  *	@sk_memcg: this socket's memory cgroup association
314 315 316 317 318 319 320
  *	@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
321
  *	@sk_reuseport_cb: reuseport group container
322
  *	@sk_rcu: used during RCU grace period
323 324 325
  *	@sk_clockid: clockid used by time-based scheduling (SO_TXTIME)
  *	@sk_txtime_deadline_mode: set deadline mode for SO_TXTIME
  *	@sk_txtime_unused: unused txtime flags
326
  *	@sk_toa_data: tcp option address (toa) data
327
  */
L
Linus Torvalds 已提交
328 329
struct sock {
	/*
330
	 * Now struct inet_timewait_sock also uses sock_common, so please just
L
Linus Torvalds 已提交
331 332 333
	 * don't add nothing before this first member (__sk_common) --acme
	 */
	struct sock_common	__sk_common;
E
Eric Dumazet 已提交
334 335 336
#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
337
#define sk_tx_queue_mapping	__sk_common.skc_tx_queue_mapping
338 339 340
#ifdef CONFIG_XPS
#define sk_rx_queue_mapping	__sk_common.skc_rx_queue_mapping
#endif
E
Eric Dumazet 已提交
341

342 343
#define sk_dontcopy_begin	__sk_common.skc_dontcopy_begin
#define sk_dontcopy_end		__sk_common.skc_dontcopy_end
E
Eric Dumazet 已提交
344
#define sk_hash			__sk_common.skc_hash
E
Eric Dumazet 已提交
345
#define sk_portpair		__sk_common.skc_portpair
E
Eric Dumazet 已提交
346 347
#define sk_num			__sk_common.skc_num
#define sk_dport		__sk_common.skc_dport
E
Eric Dumazet 已提交
348 349 350
#define sk_addrpair		__sk_common.skc_addrpair
#define sk_daddr		__sk_common.skc_daddr
#define sk_rcv_saddr		__sk_common.skc_rcv_saddr
L
Linus Torvalds 已提交
351 352 353
#define sk_family		__sk_common.skc_family
#define sk_state		__sk_common.skc_state
#define sk_reuse		__sk_common.skc_reuse
T
Tom Herbert 已提交
354
#define sk_reuseport		__sk_common.skc_reuseport
355
#define sk_ipv6only		__sk_common.skc_ipv6only
356
#define sk_net_refcnt		__sk_common.skc_net_refcnt
L
Linus Torvalds 已提交
357 358
#define sk_bound_dev_if		__sk_common.skc_bound_dev_if
#define sk_bind_node		__sk_common.skc_bind_node
359
#define sk_prot			__sk_common.skc_prot
360
#define sk_net			__sk_common.skc_net
361 362
#define sk_v6_daddr		__sk_common.skc_v6_daddr
#define sk_v6_rcv_saddr	__sk_common.skc_v6_rcv_saddr
E
Eric Dumazet 已提交
363
#define sk_cookie		__sk_common.skc_cookie
364
#define sk_incoming_cpu		__sk_common.skc_incoming_cpu
365
#define sk_flags		__sk_common.skc_flags
366
#define sk_rxhash		__sk_common.skc_rxhash
367

L
Linus Torvalds 已提交
368
	socket_lock_t		sk_lock;
369 370 371
	atomic_t		sk_drops;
	int			sk_rcvlowat;
	struct sk_buff_head	sk_error_queue;
E
Eric Dumazet 已提交
372
	struct sk_buff_head	sk_receive_queue;
373 374 375 376
	/*
	 * 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
377 378 379
	 * Note : rmem_alloc is in this structure to fill a hole
	 * on 64bit arches, not because its logically part of
	 * backlog.
380 381
	 */
	struct {
E
Eric Dumazet 已提交
382 383 384 385
		atomic_t	rmem_alloc;
		int		len;
		struct sk_buff	*head;
		struct sk_buff	*tail;
386
	} sk_backlog;
E
Eric Dumazet 已提交
387
#define sk_rmem_alloc sk_backlog.rmem_alloc
E
Eric Dumazet 已提交
388

389
	int			sk_forward_alloc;
390
#ifdef CONFIG_NET_RX_BUSY_POLL
391
	unsigned int		sk_ll_usec;
392 393
	/* ===== mostly read cache line ===== */
	unsigned int		sk_napi_id;
E
Eric Dumazet 已提交
394 395 396 397
#endif
	int			sk_rcvbuf;

	struct sk_filter __rcu	*sk_filter;
398 399 400 401
	union {
		struct socket_wq __rcu	*sk_wq;
		struct socket_wq	*sk_wq_raw;
	};
402
#ifdef CONFIG_XFRM
403
	struct xfrm_policy __rcu *sk_policy[2];
404
#endif
E
Eric Dumazet 已提交
405
	struct dst_entry	*sk_rx_dst;
406
	struct dst_entry __rcu	*sk_dst_cache;
L
Linus Torvalds 已提交
407
	atomic_t		sk_omem_alloc;
408
	int			sk_sndbuf;
409 410 411

	/* ===== cache line for TX ===== */
	int			sk_wmem_queued;
412
	refcount_t		sk_wmem_alloc;
413
	unsigned long		sk_tsq_flags;
414 415 416 417
	union {
		struct sk_buff	*sk_send_head;
		struct rb_root	tcp_rtx_queue;
	};
L
Linus Torvalds 已提交
418
	struct sk_buff_head	sk_write_queue;
419 420
	__s32			sk_peek_off;
	int			sk_write_pending;
421
	__u32			sk_dst_pending_confirm;
422
	u32			sk_pacing_status; /* see enum sk_pacing */
423 424 425 426 427 428 429 430 431
	long			sk_sndtimeo;
	struct timer_list	sk_timer;
	__u32			sk_priority;
	__u32			sk_mark;
	u32			sk_pacing_rate; /* bytes per second */
	u32			sk_max_pacing_rate;
	struct page_frag	sk_frag;
	netdev_features_t	sk_route_caps;
	netdev_features_t	sk_route_nocaps;
432
	netdev_features_t	sk_route_forced_caps;
433 434 435 436
	int			sk_gso_type;
	unsigned int		sk_gso_max_size;
	gfp_t			sk_allocation;
	__u32			sk_txhash;
437 438 439 440 441

	/*
	 * Because of non atomicity rules, all
	 * changes are protected by socket lock.
	 */
442 443 444 445 446 447 448 449 450 451 452 453 454 455 456
	unsigned int		__sk_flags_offset[0];
#ifdef __BIG_ENDIAN_BITFIELD
#define SK_FL_PROTO_SHIFT  16
#define SK_FL_PROTO_MASK   0x00ff0000

#define SK_FL_TYPE_SHIFT   0
#define SK_FL_TYPE_MASK    0x0000ffff
#else
#define SK_FL_PROTO_SHIFT  8
#define SK_FL_PROTO_MASK   0x0000ff00

#define SK_FL_TYPE_SHIFT   16
#define SK_FL_TYPE_MASK    0xffff0000
#endif

457 458
	unsigned int		sk_padding : 1,
				sk_kern_sock : 1,
459 460
				sk_no_check_tx : 1,
				sk_no_check_rx : 1,
E
Eric Dumazet 已提交
461 462 463
				sk_userlocks : 4,
				sk_protocol  : 8,
				sk_type      : 16;
464
#define SK_PROTOCOL_MAX U8_MAX
465
	u16			sk_gso_max_segs;
466
	u8			sk_pacing_shift;
L
Linus Torvalds 已提交
467
	unsigned long	        sk_lingertime;
468
	struct proto		*sk_prot_creator;
L
Linus Torvalds 已提交
469 470 471
	rwlock_t		sk_callback_lock;
	int			sk_err,
				sk_err_soft;
472 473
	u32			sk_ack_backlog;
	u32			sk_max_ack_backlog;
474
	kuid_t			sk_uid;
475 476
	struct pid		*sk_peer_pid;
	const struct cred	*sk_peer_cred;
L
Linus Torvalds 已提交
477
	long			sk_rcvtimeo;
478
	ktime_t			sk_stamp;
479 480 481
#if BITS_PER_LONG==32
	seqlock_t		sk_stamp_seq;
#endif
482
	u16			sk_tsflags;
483
	u8			sk_shutdown;
484
	u32			sk_tskey;
W
Willem de Bruijn 已提交
485
	atomic_t		sk_zckey;
486 487 488

	u8			sk_clockid;
	u8			sk_txtime_deadline_mode : 1,
489 490
				sk_txtime_report_errors : 1,
				sk_txtime_unused : 6;
491

L
Linus Torvalds 已提交
492 493
	struct socket		*sk_socket;
	void			*sk_user_data;
A
Alexey Dobriyan 已提交
494
#ifdef CONFIG_SECURITY
L
Linus Torvalds 已提交
495
	void			*sk_security;
A
Alexey Dobriyan 已提交
496
#endif
497
	struct sock_cgroup_data	sk_cgrp_data;
498
	struct mem_cgroup	*sk_memcg;
L
Linus Torvalds 已提交
499
	void			(*sk_state_change)(struct sock *sk);
500
	void			(*sk_data_ready)(struct sock *sk);
L
Linus Torvalds 已提交
501 502
	void			(*sk_write_space)(struct sock *sk);
	void			(*sk_error_report)(struct sock *sk);
E
Eric Dumazet 已提交
503 504
	int			(*sk_backlog_rcv)(struct sock *sk,
						  struct sk_buff *skb);
505 506 507 508 509
#ifdef CONFIG_SOCK_VALIDATE_XMIT
	struct sk_buff*		(*sk_validate_xmit_skb)(struct sock *sk,
							struct net_device *dev,
							struct sk_buff *skb);
#endif
L
Linus Torvalds 已提交
510
	void                    (*sk_destruct)(struct sock *sk);
511
	struct sock_reuseport __rcu	*sk_reuseport_cb;
512
	__be32			sk_toa_data[16];
513
	struct rcu_head		sk_rcu;
L
Linus Torvalds 已提交
514 515
};

516 517 518 519 520 521
enum sk_pacing {
	SK_PACING_NONE		= 0,
	SK_PACING_NEEDED	= 1,
	SK_PACING_FQ		= 2,
};

522 523 524 525 526
#define __sk_user_data(sk) ((*((void __rcu **)&(sk)->sk_user_data)))

#define rcu_dereference_sk_user_data(sk)	rcu_dereference(__sk_user_data((sk)))
#define rcu_assign_sk_user_data(sk, ptr)	rcu_assign_pointer(__sk_user_data((sk)), ptr)

527 528 529 530 531 532 533 534 535 536 537
/*
 * 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

538 539
int sk_set_peek_off(struct sock *sk, int val);

540 541
static inline int sk_peek_offset(struct sock *sk, int flags)
{
542
	if (unlikely(flags & MSG_PEEK)) {
543
		return READ_ONCE(sk->sk_peek_off);
544 545 546
	}

	return 0;
547 548 549 550
}

static inline void sk_peek_offset_bwd(struct sock *sk, int val)
{
551 552 553 554 555
	s32 off = READ_ONCE(sk->sk_peek_off);

	if (unlikely(off >= 0)) {
		off = max_t(s32, off - val, 0);
		WRITE_ONCE(sk->sk_peek_off, off);
556 557 558 559 560
	}
}

static inline void sk_peek_offset_fwd(struct sock *sk, int val)
{
561
	sk_peek_offset_bwd(sk, -val);
562 563
}

L
Linus Torvalds 已提交
564 565 566
/*
 * Hashed lists helper routines
 */
L
Li Zefan 已提交
567 568 569 570 571
static inline struct sock *sk_entry(const struct hlist_node *node)
{
	return hlist_entry(node, struct sock, sk_node);
}

572
static inline struct sock *__sk_head(const struct hlist_head *head)
L
Linus Torvalds 已提交
573 574 575 576
{
	return hlist_entry(head->first, struct sock, sk_node);
}

577
static inline struct sock *sk_head(const struct hlist_head *head)
L
Linus Torvalds 已提交
578 579 580 581
{
	return hlist_empty(head) ? NULL : __sk_head(head);
}

582 583 584 585 586 587 588 589 590 591
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);
}

592
static inline struct sock *sk_next(const struct sock *sk)
L
Linus Torvalds 已提交
593
{
G
Geliang Tang 已提交
594
	return hlist_entry_safe(sk->sk_node.next, struct sock, sk_node);
L
Linus Torvalds 已提交
595 596
}

597 598 599 600 601 602 603 604
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
605
static inline bool sk_unhashed(const struct sock *sk)
L
Linus Torvalds 已提交
606 607 608 609
{
	return hlist_unhashed(&sk->sk_node);
}

E
Eric Dumazet 已提交
610
static inline bool sk_hashed(const struct sock *sk)
L
Linus Torvalds 已提交
611
{
A
Akinobu Mita 已提交
612
	return !sk_unhashed(sk);
L
Linus Torvalds 已提交
613 614
}

E
Eric Dumazet 已提交
615
static inline void sk_node_init(struct hlist_node *node)
L
Linus Torvalds 已提交
616 617 618 619
{
	node->pprev = NULL;
}

E
Eric Dumazet 已提交
620
static inline void sk_nulls_node_init(struct hlist_nulls_node *node)
621 622 623 624
{
	node->pprev = NULL;
}

E
Eric Dumazet 已提交
625
static inline void __sk_del_node(struct sock *sk)
L
Linus Torvalds 已提交
626 627 628 629
{
	__hlist_del(&sk->sk_node);
}

630
/* NB: equivalent to hlist_del_init_rcu */
E
Eric Dumazet 已提交
631
static inline bool __sk_del_node_init(struct sock *sk)
L
Linus Torvalds 已提交
632 633 634 635
{
	if (sk_hashed(sk)) {
		__sk_del_node(sk);
		sk_node_init(&sk->sk_node);
E
Eric Dumazet 已提交
636
		return true;
L
Linus Torvalds 已提交
637
	}
E
Eric Dumazet 已提交
638
	return false;
L
Linus Torvalds 已提交
639 640 641 642 643 644 645 646
}

/* 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.
 */

647
static __always_inline void sock_hold(struct sock *sk)
L
Linus Torvalds 已提交
648
{
649
	refcount_inc(&sk->sk_refcnt);
L
Linus Torvalds 已提交
650 651 652 653 654
}

/* Ungrab socket in the context, which assumes that socket refcnt
   cannot hit zero, f.e. it is true in context of any socketcall.
 */
655
static __always_inline void __sock_put(struct sock *sk)
L
Linus Torvalds 已提交
656
{
657
	refcount_dec(&sk->sk_refcnt);
L
Linus Torvalds 已提交
658 659
}

E
Eric Dumazet 已提交
660
static inline bool sk_del_node_init(struct sock *sk)
L
Linus Torvalds 已提交
661
{
E
Eric Dumazet 已提交
662
	bool rc = __sk_del_node_init(sk);
L
Linus Torvalds 已提交
663 664 665

	if (rc) {
		/* paranoid for a while -acme */
666
		WARN_ON(refcount_read(&sk->sk_refcnt) == 1);
L
Linus Torvalds 已提交
667 668 669 670
		__sock_put(sk);
	}
	return rc;
}
671
#define sk_del_node_init_rcu(sk)	sk_del_node_init(sk)
L
Linus Torvalds 已提交
672

E
Eric Dumazet 已提交
673
static inline bool __sk_nulls_del_node_init_rcu(struct sock *sk)
674 675
{
	if (sk_hashed(sk)) {
676
		hlist_nulls_del_init_rcu(&sk->sk_nulls_node);
E
Eric Dumazet 已提交
677
		return true;
678
	}
E
Eric Dumazet 已提交
679
	return false;
680 681
}

E
Eric Dumazet 已提交
682
static inline bool sk_nulls_del_node_init_rcu(struct sock *sk)
683
{
E
Eric Dumazet 已提交
684
	bool rc = __sk_nulls_del_node_init_rcu(sk);
685 686 687

	if (rc) {
		/* paranoid for a while -acme */
688
		WARN_ON(refcount_read(&sk->sk_refcnt) == 1);
689 690 691 692 693
		__sock_put(sk);
	}
	return rc;
}

E
Eric Dumazet 已提交
694
static inline void __sk_add_node(struct sock *sk, struct hlist_head *list)
L
Linus Torvalds 已提交
695 696 697 698
{
	hlist_add_head(&sk->sk_node, list);
}

E
Eric Dumazet 已提交
699
static inline void sk_add_node(struct sock *sk, struct hlist_head *list)
L
Linus Torvalds 已提交
700 701 702 703 704
{
	sock_hold(sk);
	__sk_add_node(sk, list);
}

E
Eric Dumazet 已提交
705
static inline void sk_add_node_rcu(struct sock *sk, struct hlist_head *list)
706 707
{
	sock_hold(sk);
708 709 710 711 712
	if (IS_ENABLED(CONFIG_IPV6) && sk->sk_reuseport &&
	    sk->sk_family == AF_INET6)
		hlist_add_tail_rcu(&sk->sk_node, list);
	else
		hlist_add_head_rcu(&sk->sk_node, list);
713 714
}

715 716 717 718 719 720
static inline void sk_add_node_tail_rcu(struct sock *sk, struct hlist_head *list)
{
	sock_hold(sk);
	hlist_add_tail_rcu(&sk->sk_node, list);
}

E
Eric Dumazet 已提交
721
static inline void __sk_nulls_add_node_rcu(struct sock *sk, struct hlist_nulls_head *list)
722
{
723
	hlist_nulls_add_head_rcu(&sk->sk_nulls_node, list);
724 725
}

E
Eric Dumazet 已提交
726
static inline void sk_nulls_add_node_rcu(struct sock *sk, struct hlist_nulls_head *list)
727 728
{
	sock_hold(sk);
729
	__sk_nulls_add_node_rcu(sk, list);
730 731
}

E
Eric Dumazet 已提交
732
static inline void __sk_del_bind_node(struct sock *sk)
L
Linus Torvalds 已提交
733 734 735 736
{
	__hlist_del(&sk->sk_bind_node);
}

E
Eric Dumazet 已提交
737
static inline void sk_add_bind_node(struct sock *sk,
L
Linus Torvalds 已提交
738 739 740 741 742
					struct hlist_head *list)
{
	hlist_add_head(&sk->sk_bind_node, list);
}

743 744 745 746
#define sk_for_each(__sk, list) \
	hlist_for_each_entry(__sk, list, sk_node)
#define sk_for_each_rcu(__sk, list) \
	hlist_for_each_entry_rcu(__sk, list, sk_node)
747 748 749 750
#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)
751 752
#define sk_for_each_from(__sk) \
	hlist_for_each_entry_from(__sk, sk_node)
753 754 755
#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)
756 757 758 759
#define sk_for_each_safe(__sk, tmp, list) \
	hlist_for_each_entry_safe(__sk, tmp, list, sk_node)
#define sk_for_each_bound(__sk, list) \
	hlist_for_each_entry(__sk, list, sk_bind_node)
L
Linus Torvalds 已提交
760

761
/**
762
 * sk_for_each_entry_offset_rcu - iterate over a list at a given struct offset
763 764 765 766 767 768
 * @tpos:	the type * to use as a loop cursor.
 * @pos:	the &struct hlist_node to use as a loop cursor.
 * @head:	the head for your list.
 * @offset:	offset of hlist_node within the struct.
 *
 */
769
#define sk_for_each_entry_offset_rcu(tpos, pos, head, offset)		       \
770
	for (pos = rcu_dereference(hlist_first_rcu(head));		       \
771
	     pos != NULL &&						       \
772
		({ tpos = (typeof(*tpos) *)((void *)pos - offset); 1;});       \
773
	     pos = rcu_dereference(hlist_next_rcu(pos)))
774

E
Eric W. Biederman 已提交
775 776 777 778 779 780 781 782 783
static inline struct user_namespace *sk_user_ns(struct sock *sk)
{
	/* Careful only use this in a context where these parameters
	 * can not change and must all be valid, such as recvmsg from
	 * userspace.
	 */
	return sk->sk_socket->file->f_cred->user_ns;
}

L
Linus Torvalds 已提交
784 785 786 787 788 789 790 791 792 793 794 795 796 797
/* 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 */
798
	SOCK_RCVTSTAMPNS, /* %SO_TIMESTAMPNS setting */
L
Linus Torvalds 已提交
799 800
	SOCK_LOCALROUTE, /* route locally only, %SO_DONTROUTE setting */
	SOCK_QUEUE_SHRUNK, /* write queue has been shrunk recently */
801
	SOCK_MEMALLOC, /* VM depends on this socket for swapping */
802
	SOCK_TIMESTAMPING_RX_SOFTWARE,  /* %SOF_TIMESTAMPING_RX_SOFTWARE */
E
Eric Dumazet 已提交
803
	SOCK_FASYNC, /* fasync() active */
804
	SOCK_RXQ_OVFL,
805
	SOCK_ZEROCOPY, /* buffers from userspace */
806
	SOCK_WIFI_STATUS, /* push wifi status to userspace */
807 808 809 810
	SOCK_NOFCS, /* Tell NIC not to do the Ethernet FCS.
		     * Will use last 4 bytes of packet sent from
		     * user-space instead.
		     */
811
	SOCK_FILTER_LOCKED, /* Filter cannot be changed anymore */
812
	SOCK_SELECT_ERR_QUEUE, /* Wake select on error queue */
813
	SOCK_RCU_FREE, /* wait rcu grace period in sk_destruct() */
814
	SOCK_TXTIME,
L
Linus Torvalds 已提交
815 816
};

817 818
#define SK_FLAGS_TIMESTAMP ((1UL << SOCK_TIMESTAMP) | (1UL << SOCK_TIMESTAMPING_RX_SOFTWARE))

R
Ralf Baechle 已提交
819 820 821 822 823
static inline void sock_copy_flags(struct sock *nsk, struct sock *osk)
{
	nsk->sk_flags = osk->sk_flags;
}

L
Linus Torvalds 已提交
824 825 826 827 828 829 830 831 832 833
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
834
static inline bool sock_flag(const struct sock *sk, enum sock_flags flag)
L
Linus Torvalds 已提交
835 836 837 838
{
	return test_bit(flag, &sk->sk_flags);
}

839
#ifdef CONFIG_NET
840
DECLARE_STATIC_KEY_FALSE(memalloc_socks_key);
841 842
static inline int sk_memalloc_socks(void)
{
843
	return static_branch_unlikely(&memalloc_socks_key);
844 845 846 847 848 849 850 851 852 853
}
#else

static inline int sk_memalloc_socks(void)
{
	return 0;
}

#endif

854
static inline gfp_t sk_gfp_mask(const struct sock *sk, gfp_t gfp_mask)
855
{
856
	return gfp_mask | (sk->sk_allocation & __GFP_MEMALLOC);
857 858
}

L
Linus Torvalds 已提交
859 860 861 862 863 864 865 866 867 868
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
869
static inline bool sk_acceptq_is_full(const struct sock *sk)
L
Linus Torvalds 已提交
870
{
871
	return sk->sk_ack_backlog > sk->sk_max_ack_backlog;
L
Linus Torvalds 已提交
872 873 874 875 876
}

/*
 * Compute minimal free write space needed to queue new packets.
 */
E
Eric Dumazet 已提交
877
static inline int sk_stream_min_wspace(const struct sock *sk)
L
Linus Torvalds 已提交
878
{
879
	return sk->sk_wmem_queued >> 1;
L
Linus Torvalds 已提交
880 881
}

E
Eric Dumazet 已提交
882
static inline int sk_stream_wspace(const struct sock *sk)
L
Linus Torvalds 已提交
883 884 885 886
{
	return sk->sk_sndbuf - sk->sk_wmem_queued;
}

887
void sk_stream_write_space(struct sock *sk);
L
Linus Torvalds 已提交
888

Z
Zhu Yi 已提交
889
/* OOB backlog add */
Z
Zhu Yi 已提交
890
static inline void __sk_add_backlog(struct sock *sk, struct sk_buff *skb)
891
{
E
Eric Dumazet 已提交
892
	/* dont let skb dst not refcounted, we are going to leave rcu lock */
E
Eric Dumazet 已提交
893
	skb_dst_force(skb);
E
Eric Dumazet 已提交
894 895 896 897

	if (!sk->sk_backlog.tail)
		sk->sk_backlog.head = skb;
	else
898
		sk->sk_backlog.tail->next = skb;
E
Eric Dumazet 已提交
899 900

	sk->sk_backlog.tail = skb;
901 902
	skb->next = NULL;
}
L
Linus Torvalds 已提交
903

904 905
/*
 * Take into account size of receive queue and backlog queue
E
Eric Dumazet 已提交
906 907
 * Do not take into account this skb truesize,
 * to allow even a single big packet to come.
908
 */
909
static inline bool sk_rcvqueues_full(const struct sock *sk, unsigned int limit)
910 911 912
{
	unsigned int qsize = sk->sk_backlog.len + atomic_read(&sk->sk_rmem_alloc);

913
	return qsize > limit;
914 915
}

Z
Zhu Yi 已提交
916
/* The per-socket spinlock must be held here. */
917 918
static inline __must_check int sk_add_backlog(struct sock *sk, struct sk_buff *skb,
					      unsigned int limit)
Z
Zhu Yi 已提交
919
{
920
	if (sk_rcvqueues_full(sk, limit))
Z
Zhu Yi 已提交
921 922
		return -ENOBUFS;

923 924 925 926 927 928 929 930
	/*
	 * If the skb was allocated from pfmemalloc reserves, only
	 * allow SOCK_MEMALLOC sockets to use it as this socket is
	 * helping free memory
	 */
	if (skb_pfmemalloc(skb) && !sock_flag(sk, SOCK_MEMALLOC))
		return -ENOMEM;

Z
Zhu Yi 已提交
931
	__sk_add_backlog(sk, skb);
Z
Zhu Yi 已提交
932 933 934 935
	sk->sk_backlog.len += skb->truesize;
	return 0;
}

936
int __sk_backlog_rcv(struct sock *sk, struct sk_buff *skb);
937

P
Peter Zijlstra 已提交
938 939
static inline int sk_backlog_rcv(struct sock *sk, struct sk_buff *skb)
{
940 941 942
	if (sk_memalloc_socks() && skb_pfmemalloc(skb))
		return __sk_backlog_rcv(sk, skb);

P
Peter Zijlstra 已提交
943 944 945
	return sk->sk_backlog_rcv(sk, skb);
}

E
Eric Dumazet 已提交
946 947
static inline void sk_incoming_cpu_update(struct sock *sk)
{
948 949 950 951
	int cpu = raw_smp_processor_id();

	if (unlikely(sk->sk_incoming_cpu != cpu))
		sk->sk_incoming_cpu = cpu;
E
Eric Dumazet 已提交
952 953
}

954
static inline void sock_rps_record_flow_hash(__u32 hash)
955 956 957 958 959 960
{
#ifdef CONFIG_RPS
	struct rps_sock_flow_table *sock_flow_table;

	rcu_read_lock();
	sock_flow_table = rcu_dereference(rps_sock_flow_table);
961
	rps_record_sock_flow(sock_flow_table, hash);
962 963 964 965
	rcu_read_unlock();
#endif
}

966 967
static inline void sock_rps_record_flow(const struct sock *sk)
{
968
#ifdef CONFIG_RPS
E
Eric Dumazet 已提交
969 970 971 972 973 974 975 976 977 978 979 980 981 982
	if (static_key_false(&rfs_needed)) {
		/* Reading sk->sk_rxhash might incur an expensive cache line
		 * miss.
		 *
		 * TCP_ESTABLISHED does cover almost all states where RFS
		 * might be useful, and is cheaper [1] than testing :
		 *	IPv4: inet_sk(sk)->inet_daddr
		 * 	IPv6: ipv6_addr_any(&sk->sk_v6_daddr)
		 * OR	an additional socket flag
		 * [1] : sk_state and sk_prot are in the same cache line.
		 */
		if (sk->sk_state == TCP_ESTABLISHED)
			sock_rps_record_flow_hash(sk->sk_rxhash);
	}
983
#endif
984 985
}

986 987
static inline void sock_rps_save_rxhash(struct sock *sk,
					const struct sk_buff *skb)
988 989
{
#ifdef CONFIG_RPS
990
	if (unlikely(sk->sk_rxhash != skb->hash))
991
		sk->sk_rxhash = skb->hash;
992 993 994
#endif
}

995 996 997 998 999 1000 1001
static inline void sock_rps_reset_rxhash(struct sock *sk)
{
#ifdef CONFIG_RPS
	sk->sk_rxhash = 0;
#endif
}

W
WANG Cong 已提交
1002
#define sk_wait_event(__sk, __timeo, __condition, __wait)		\
S
Stephen Hemminger 已提交
1003 1004 1005 1006
	({	int __rc;						\
		release_sock(__sk);					\
		__rc = __condition;					\
		if (!__rc) {						\
W
WANG Cong 已提交
1007 1008 1009
			*(__timeo) = wait_woken(__wait,			\
						TASK_INTERRUPTIBLE,	\
						*(__timeo));		\
S
Stephen Hemminger 已提交
1010
		}							\
W
WANG Cong 已提交
1011
		sched_annotate_sleep();					\
S
Stephen Hemminger 已提交
1012 1013 1014 1015
		lock_sock(__sk);					\
		__rc = __condition;					\
		__rc;							\
	})
L
Linus Torvalds 已提交
1016

1017 1018 1019 1020 1021 1022 1023
int sk_stream_wait_connect(struct sock *sk, long *timeo_p);
int sk_stream_wait_memory(struct sock *sk, long *timeo_p);
void sk_stream_wait_close(struct sock *sk, long timeo_p);
int sk_stream_error(struct sock *sk, int flags, int err);
void sk_stream_kill_queues(struct sock *sk);
void sk_set_memalloc(struct sock *sk);
void sk_clear_memalloc(struct sock *sk);
L
Linus Torvalds 已提交
1024

1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035
void __sk_flush_backlog(struct sock *sk);

static inline bool sk_flush_backlog(struct sock *sk)
{
	if (unlikely(READ_ONCE(sk->sk_backlog.tail))) {
		__sk_flush_backlog(sk);
		return true;
	}
	return false;
}

1036
int sk_wait_data(struct sock *sk, long *timeo, const struct sk_buff *skb);
L
Linus Torvalds 已提交
1037

1038
struct request_sock_ops;
1039
struct timewait_sock_ops;
1040
struct inet_hashinfo;
1041
struct raw_hashinfo;
1042
struct smc_hashinfo;
1043
struct module;
1044

E
Eric Dumazet 已提交
1045
/*
1046
 * caches using SLAB_TYPESAFE_BY_RCU should let .next pointer from nulls nodes
E
Eric Dumazet 已提交
1047 1048 1049 1050 1051 1052 1053 1054 1055 1056
 * un-modified. Special care is taken when initializing object to zero.
 */
static inline void sk_prot_clear_nulls(struct sock *sk, int size)
{
	if (offsetof(struct sock, sk_node.next) != 0)
		memset(sk, 0, offsetof(struct sock, sk_node.next));
	memset(&sk->sk_node.pprev, 0,
	       size - offsetof(struct sock, sk_node.pprev));
}

L
Linus Torvalds 已提交
1057 1058 1059 1060
/* Networking protocol blocks we attach to sockets.
 * socket layer -> transport layer interface
 */
struct proto {
E
Eric Dumazet 已提交
1061
	void			(*close)(struct sock *sk,
L
Linus Torvalds 已提交
1062
					long timeout);
A
Andrey Ignatov 已提交
1063 1064 1065
	int			(*pre_connect)(struct sock *sk,
					struct sockaddr *uaddr,
					int addr_len);
L
Linus Torvalds 已提交
1066
	int			(*connect)(struct sock *sk,
E
Eric Dumazet 已提交
1067
					struct sockaddr *uaddr,
L
Linus Torvalds 已提交
1068 1069 1070
					int addr_len);
	int			(*disconnect)(struct sock *sk, int flags);

1071 1072
	struct sock *		(*accept)(struct sock *sk, int flags, int *err,
					  bool kern);
L
Linus Torvalds 已提交
1073 1074 1075 1076

	int			(*ioctl)(struct sock *sk, int cmd,
					 unsigned long arg);
	int			(*init)(struct sock *sk);
1077
	void			(*destroy)(struct sock *sk);
L
Linus Torvalds 已提交
1078
	void			(*shutdown)(struct sock *sk, int how);
E
Eric Dumazet 已提交
1079
	int			(*setsockopt)(struct sock *sk, int level,
L
Linus Torvalds 已提交
1080
					int optname, char __user *optval,
1081
					unsigned int optlen);
E
Eric Dumazet 已提交
1082 1083 1084
	int			(*getsockopt)(struct sock *sk, int level,
					int optname, char __user *optval,
					int __user *option);
1085
	void			(*keepalive)(struct sock *sk, int valbool);
A
Alexey Dobriyan 已提交
1086
#ifdef CONFIG_COMPAT
1087 1088 1089
	int			(*compat_setsockopt)(struct sock *sk,
					int level,
					int optname, char __user *optval,
1090
					unsigned int optlen);
1091 1092 1093 1094
	int			(*compat_getsockopt)(struct sock *sk,
					int level,
					int optname, char __user *optval,
					int __user *option);
1095 1096
	int			(*compat_ioctl)(struct sock *sk,
					unsigned int cmd, unsigned long arg);
A
Alexey Dobriyan 已提交
1097
#endif
1098 1099 1100
	int			(*sendmsg)(struct sock *sk, struct msghdr *msg,
					   size_t len);
	int			(*recvmsg)(struct sock *sk, struct msghdr *msg,
E
Eric Dumazet 已提交
1101 1102
					   size_t len, int noblock, int flags,
					   int *addr_len);
L
Linus Torvalds 已提交
1103 1104
	int			(*sendpage)(struct sock *sk, struct page *page,
					int offset, size_t size, int flags);
E
Eric Dumazet 已提交
1105
	int			(*bind)(struct sock *sk,
L
Linus Torvalds 已提交
1106 1107
					struct sockaddr *uaddr, int addr_len);

E
Eric Dumazet 已提交
1108
	int			(*backlog_rcv) (struct sock *sk,
L
Linus Torvalds 已提交
1109 1110
						struct sk_buff *skb);

E
Eric Dumazet 已提交
1111 1112
	void		(*release_cb)(struct sock *sk);

L
Linus Torvalds 已提交
1113
	/* Keeping track of sk's, looking them up, and port selection methods. */
1114
	int			(*hash)(struct sock *sk);
L
Linus Torvalds 已提交
1115
	void			(*unhash)(struct sock *sk);
E
Eric Dumazet 已提交
1116
	void			(*rehash)(struct sock *sk);
L
Linus Torvalds 已提交
1117 1118
	int			(*get_port)(struct sock *sk, unsigned short snum);

1119
	/* Keeping track of sockets in use */
1120
#ifdef CONFIG_PROC_FS
1121
	unsigned int		inuse_idx;
1122
#endif
1123

1124
	bool			(*stream_memory_free)(const struct sock *sk);
1125
	bool			(*stream_memory_read)(const struct sock *sk);
L
Linus Torvalds 已提交
1126
	/* Memory pressure */
1127
	void			(*enter_memory_pressure)(struct sock *sk);
1128
	void			(*leave_memory_pressure)(struct sock *sk);
E
Eric Dumazet 已提交
1129
	atomic_long_t		*memory_allocated;	/* Current allocated memory. */
1130
	struct percpu_counter	*sockets_allocated;	/* Current number of sockets. */
L
Linus Torvalds 已提交
1131 1132 1133
	/*
	 * Pressure flag: try to collapse.
	 * Technical note: it is used by multiple contexts non atomically.
1134
	 * All the __sk_mem_schedule() is of this nature: accounting
L
Linus Torvalds 已提交
1135 1136
	 * is strict, actions are advisory and have some latency.
	 */
1137
	unsigned long		*memory_pressure;
E
Eric Dumazet 已提交
1138
	long			*sysctl_mem;
1139

L
Linus Torvalds 已提交
1140 1141
	int			*sysctl_wmem;
	int			*sysctl_rmem;
1142 1143 1144
	u32			sysctl_wmem_offset;
	u32			sysctl_rmem_offset;

L
Linus Torvalds 已提交
1145
	int			max_header;
1146
	bool			no_autobind;
L
Linus Torvalds 已提交
1147

1148
	struct kmem_cache	*slab;
L
Linus Torvalds 已提交
1149
	unsigned int		obj_size;
1150
	slab_flags_t		slab_flags;
1151 1152
	unsigned int		useroffset;	/* Usercopy region offset */
	unsigned int		usersize;	/* Usercopy region size */
L
Linus Torvalds 已提交
1153

1154
	struct percpu_counter	*orphan_count;
1155

1156
	struct request_sock_ops	*rsk_prot;
1157
	struct timewait_sock_ops *twsk_prot;
1158

1159 1160
	union {
		struct inet_hashinfo	*hashinfo;
1161
		struct udp_table	*udp_table;
1162
		struct raw_hashinfo	*raw_hash;
1163
		struct smc_hashinfo	*smc_hash;
1164
	} h;
1165

L
Linus Torvalds 已提交
1166 1167 1168 1169 1170
	struct module		*owner;

	char			name[32];

	struct list_head	node;
1171 1172
#ifdef SOCK_REFCNT_DEBUG
	atomic_t		socks;
G
Glauber Costa 已提交
1173
#endif
1174
	int			(*diag_destroy)(struct sock *sk, int err);
1175
} __randomize_layout;
G
Glauber Costa 已提交
1176

1177 1178
int proto_register(struct proto *prot, int alloc_slab);
void proto_unregister(struct proto *prot);
1179
int sock_load_diag_module(int family, int protocol);
L
Linus Torvalds 已提交
1180

1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193
#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));
}

1194
static inline void sk_refcnt_debug_release(const struct sock *sk)
1195
{
1196
	if (refcount_read(&sk->sk_refcnt) != 1)
1197
		printk(KERN_DEBUG "Destruction of the %s socket %p delayed, refcnt=%d\n",
1198
		       sk->sk_prot->name, sk, refcount_read(&sk->sk_refcnt));
1199 1200 1201 1202 1203 1204 1205
}
#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 */

1206 1207 1208 1209 1210 1211 1212 1213 1214
static inline bool sk_stream_memory_free(const struct sock *sk)
{
	if (sk->sk_wmem_queued >= sk->sk_sndbuf)
		return false;

	return sk->sk_prot->stream_memory_free ?
		sk->sk_prot->stream_memory_free(sk) : true;
}

1215 1216
static inline bool sk_stream_is_writeable(const struct sock *sk)
{
1217 1218
	return sk_stream_wspace(sk) >= sk_stream_min_wspace(sk) &&
	       sk_stream_memory_free(sk);
1219
}
G
Glauber Costa 已提交
1220

1221 1222 1223 1224 1225 1226 1227 1228 1229 1230
static inline int sk_under_cgroup_hierarchy(struct sock *sk,
					    struct cgroup *ancestor)
{
#ifdef CONFIG_SOCK_CGROUP_DATA
	return cgroup_is_descendant(sock_cgroup_ptr(&sk->sk_cgrp_data),
				    ancestor);
#else
	return -ENOTSUPP;
#endif
}
1231

1232 1233 1234 1235 1236 1237 1238 1239 1240
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
1241

1242 1243
	if (mem_cgroup_sockets_enabled && sk->sk_memcg &&
	    mem_cgroup_under_socket_pressure(sk->sk_memcg))
1244
		return true;
G
Glauber Costa 已提交
1245

1246
	return !!*sk->sk_prot->memory_pressure;
1247 1248 1249 1250 1251
}

static inline long
sk_memory_allocated(const struct sock *sk)
{
1252
	return atomic_long_read(sk->sk_prot->memory_allocated);
1253 1254 1255
}

static inline long
1256
sk_memory_allocated_add(struct sock *sk, int amt)
1257
{
1258
	return atomic_long_add_return(amt, sk->sk_prot->memory_allocated);
1259 1260 1261
}

static inline void
1262
sk_memory_allocated_sub(struct sock *sk, int amt)
1263
{
1264
	atomic_long_sub(amt, sk->sk_prot->memory_allocated);
1265 1266 1267 1268
}

static inline void sk_sockets_allocated_dec(struct sock *sk)
{
1269
	percpu_counter_dec(sk->sk_prot->sockets_allocated);
1270 1271 1272 1273
}

static inline void sk_sockets_allocated_inc(struct sock *sk)
{
1274
	percpu_counter_inc(sk->sk_prot->sockets_allocated);
1275 1276 1277 1278 1279
}

static inline int
sk_sockets_allocated_read_positive(struct sock *sk)
{
1280
	return percpu_counter_read_positive(sk->sk_prot->sockets_allocated);
1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302
}

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;
}

1303 1304

#ifdef CONFIG_PROC_FS
L
Linus Torvalds 已提交
1305
/* Called with local bh disabled */
1306 1307
void sock_prot_inuse_add(struct net *net, struct proto *prot, int inc);
int sock_prot_inuse_get(struct net *net, struct proto *proto);
1308
int sock_inuse_get(struct net *net);
1309
#else
E
Eric Dumazet 已提交
1310
static inline void sock_prot_inuse_add(struct net *net, struct proto *prot,
1311
		int inc)
1312 1313 1314 1315
{
}
#endif

L
Linus Torvalds 已提交
1316

1317 1318 1319
/* With per-bucket locks this operation is not-atomic, so that
 * this version is not worse.
 */
1320
static inline int __sk_prot_rehash(struct sock *sk)
1321 1322
{
	sk->sk_prot->unhash(sk);
1323
	return sk->sk_prot->hash(sk);
1324 1325
}

L
Linus Torvalds 已提交
1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355
/* 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

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;
}

1356 1357 1358
/*
 * Functions for memory accounting
 */
1359
int __sk_mem_raise_allocated(struct sock *sk, int size, int amt, int kind);
1360
int __sk_mem_schedule(struct sock *sk, int size, int kind);
1361
void __sk_mem_reduce_allocated(struct sock *sk, int amount);
E
Eric Dumazet 已提交
1362
void __sk_mem_reclaim(struct sock *sk, int amount);
L
Linus Torvalds 已提交
1363

E
Eric Dumazet 已提交
1364 1365 1366 1367
/* We used to have PAGE_SIZE here, but systems with 64KB pages
 * do not necessarily have 16x time more memory than 4KB ones.
 */
#define SK_MEM_QUANTUM 4096
1368 1369 1370
#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
1371

E
Eric Dumazet 已提交
1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384
/* sysctl_mem values are in pages, we convert them in SK_MEM_QUANTUM units */
static inline long sk_prot_mem_limits(const struct sock *sk, int index)
{
	long val = sk->sk_prot->sysctl_mem[index];

#if PAGE_SIZE > SK_MEM_QUANTUM
	val <<= PAGE_SHIFT - SK_MEM_QUANTUM_SHIFT;
#elif PAGE_SIZE < SK_MEM_QUANTUM
	val >>= SK_MEM_QUANTUM_SHIFT - PAGE_SHIFT;
#endif
	return val;
}

1385
static inline int sk_mem_pages(int amt)
L
Linus Torvalds 已提交
1386
{
1387
	return (amt + SK_MEM_QUANTUM - 1) >> SK_MEM_QUANTUM_SHIFT;
L
Linus Torvalds 已提交
1388 1389
}

E
Eric Dumazet 已提交
1390
static inline bool sk_has_account(struct sock *sk)
L
Linus Torvalds 已提交
1391
{
1392 1393
	/* return true if protocol supports memory accounting */
	return !!sk->sk_prot->memory_allocated;
L
Linus Torvalds 已提交
1394 1395
}

E
Eric Dumazet 已提交
1396
static inline bool sk_wmem_schedule(struct sock *sk, int size)
L
Linus Torvalds 已提交
1397
{
1398
	if (!sk_has_account(sk))
E
Eric Dumazet 已提交
1399
		return true;
1400 1401
	return size <= sk->sk_forward_alloc ||
		__sk_mem_schedule(sk, size, SK_MEM_SEND);
L
Linus Torvalds 已提交
1402 1403
}

1404
static inline bool
1405
sk_rmem_schedule(struct sock *sk, struct sk_buff *skb, int size)
H
Herbert Xu 已提交
1406
{
1407
	if (!sk_has_account(sk))
E
Eric Dumazet 已提交
1408
		return true;
1409 1410 1411
	return size<= sk->sk_forward_alloc ||
		__sk_mem_schedule(sk, size, SK_MEM_RECV) ||
		skb_pfmemalloc(skb);
1412 1413 1414 1415 1416 1417 1418
}

static inline void sk_mem_reclaim(struct sock *sk)
{
	if (!sk_has_account(sk))
		return;
	if (sk->sk_forward_alloc >= SK_MEM_QUANTUM)
E
Eric Dumazet 已提交
1419
		__sk_mem_reclaim(sk, sk->sk_forward_alloc);
1420 1421
}

1422 1423 1424 1425 1426
static inline void sk_mem_reclaim_partial(struct sock *sk)
{
	if (!sk_has_account(sk))
		return;
	if (sk->sk_forward_alloc > SK_MEM_QUANTUM)
E
Eric Dumazet 已提交
1427
		__sk_mem_reclaim(sk, sk->sk_forward_alloc - 1);
1428 1429
}

1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441
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;
1442 1443 1444 1445 1446 1447 1448 1449 1450 1451

	/* Avoid a possible overflow.
	 * TCP send queues can make this happen, if sk_mem_reclaim()
	 * is not called and more than 2 GBytes are released at once.
	 *
	 * If we reach 2 MBytes, reclaim 1 MBytes right now, there is
	 * no need to hold that much forward allocation anyway.
	 */
	if (unlikely(sk->sk_forward_alloc >= 1 << 21))
		__sk_mem_reclaim(sk, 1 << 20);
1452 1453 1454 1455 1456 1457 1458 1459
}

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
1460 1461
}

1462 1463
static inline void sock_release_ownership(struct sock *sk)
{
1464 1465 1466 1467 1468 1469
	if (sk->sk_lock.owned) {
		sk->sk_lock.owned = 0;

		/* The sk_lock has mutex_unlock() semantics: */
		mutex_release(&sk->sk_lock.dep_map, 1, _RET_IP_);
	}
1470 1471
}

1472 1473 1474 1475 1476 1477 1478
/*
 * 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
1479
#define sock_lock_init_class_and_name(sk, sname, skey, name, key)	\
1480
do {									\
1481
	sk->sk_lock.owned = 0;						\
1482 1483 1484 1485 1486
	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
1487
				(skey), (sname));				\
1488 1489 1490
	lockdep_init_map(&(sk)->sk_lock.dep_map, (name), (key), 0);	\
} while (0)

1491
#ifdef CONFIG_LOCKDEP
1492
static inline bool lockdep_sock_is_held(const struct sock *sk)
1493 1494 1495 1496
{
	return lockdep_is_held(&sk->sk_lock) ||
	       lockdep_is_held(&sk->sk_lock.slock);
}
1497
#endif
1498

1499
void lock_sock_nested(struct sock *sk, int subclass);
1500 1501 1502 1503 1504 1505

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

1506
void __release_sock(struct sock *sk);
1507
void release_sock(struct sock *sk);
L
Linus Torvalds 已提交
1508 1509 1510

/* BH context may only use the following locking interface. */
#define bh_lock_sock(__sk)	spin_lock(&((__sk)->sk_lock.slock))
1511 1512 1513
#define bh_lock_sock_nested(__sk) \
				spin_lock_nested(&((__sk)->sk_lock.slock), \
				SINGLE_DEPTH_NESTING)
L
Linus Torvalds 已提交
1514 1515
#define bh_unlock_sock(__sk)	spin_unlock(&((__sk)->sk_lock.slock))

1516
bool lock_sock_fast(struct sock *sk);
1517 1518 1519 1520 1521 1522 1523 1524 1525
/**
 * 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
1526
{
1527 1528 1529 1530
	if (slow)
		release_sock(sk);
	else
		spin_unlock_bh(&sk->sk_lock.slock);
E
Eric Dumazet 已提交
1531 1532
}

1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546
/* 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.
 */

1547
static inline void sock_owned_by_me(const struct sock *sk)
1548 1549
{
#ifdef CONFIG_LOCKDEP
1550
	WARN_ON_ONCE(!lockdep_sock_is_held(sk) && debug_locks);
1551
#endif
1552 1553 1554 1555 1556
}

static inline bool sock_owned_by_user(const struct sock *sk)
{
	sock_owned_by_me(sk);
1557 1558 1559
	return sk->sk_lock.owned;
}

T
Tom Herbert 已提交
1560 1561 1562 1563 1564
static inline bool sock_owned_by_user_nocheck(const struct sock *sk)
{
	return sk->sk_lock.owned;
}

1565 1566 1567 1568 1569 1570 1571
/* no reclassification while locks are held */
static inline bool sock_allow_reclassification(const struct sock *csk)
{
	struct sock *sk = (struct sock *)csk;

	return !sk->sk_lock.owned && !spin_is_locked(&sk->sk_lock.slock);
}
E
Eric Dumazet 已提交
1572

1573
struct sock *sk_alloc(struct net *net, int family, gfp_t priority,
1574
		      struct proto *prot, int kern);
1575
void sk_free(struct sock *sk);
1576
void sk_destruct(struct sock *sk);
1577
struct sock *sk_clone_lock(const struct sock *sk, const gfp_t priority);
1578
void sk_free_unlock_clone(struct sock *sk);
1579 1580 1581

struct sk_buff *sock_wmalloc(struct sock *sk, unsigned long size, int force,
			     gfp_t priority);
E
Eric Dumazet 已提交
1582
void __sock_wfree(struct sk_buff *skb);
1583
void sock_wfree(struct sk_buff *skb);
1584 1585
struct sk_buff *sock_omalloc(struct sock *sk, unsigned long size,
			     gfp_t priority);
1586 1587
void skb_orphan_partial(struct sk_buff *skb);
void sock_rfree(struct sk_buff *skb);
1588
void sock_efree(struct sk_buff *skb);
1589
#ifdef CONFIG_INET
1590
void sock_edemux(struct sk_buff *skb);
1591
#else
1592
#define sock_edemux sock_efree
1593
#endif
1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606

int sock_setsockopt(struct socket *sock, int level, int op,
		    char __user *optval, unsigned int optlen);

int sock_getsockopt(struct socket *sock, int level, int op,
		    char __user *optval, int __user *optlen);
struct sk_buff *sock_alloc_send_skb(struct sock *sk, unsigned long size,
				    int noblock, int *errcode);
struct sk_buff *sock_alloc_send_pskb(struct sock *sk, unsigned long header_len,
				     unsigned long data_len, int noblock,
				     int *errcode, int max_page_order);
void *sock_kmalloc(struct sock *sk, int size, gfp_t priority);
void sock_kfree_s(struct sock *sk, void *mem, int size);
1607
void sock_kzfree_s(struct sock *sk, void *mem, int size);
1608
void sk_send_sigurg(struct sock *sk);
L
Linus Torvalds 已提交
1609

E
Edward Jee 已提交
1610
struct sockcm_cookie {
1611
	u64 transmit_time;
E
Edward Jee 已提交
1612
	u32 mark;
1613
	u16 tsflags;
E
Edward Jee 已提交
1614 1615
};

1616 1617 1618 1619 1620 1621
static inline void sockcm_init(struct sockcm_cookie *sockc,
			       const struct sock *sk)
{
	*sockc = (struct sockcm_cookie) { .tsflags = sk->sk_tsflags };
}

1622 1623
int __sock_cmsg_send(struct sock *sk, struct msghdr *msg, struct cmsghdr *cmsg,
		     struct sockcm_cookie *sockc);
E
Edward Jee 已提交
1624 1625 1626
int sock_cmsg_send(struct sock *sk, struct msghdr *msg,
		   struct sockcm_cookie *sockc);

L
Linus Torvalds 已提交
1627 1628 1629 1630
/*
 * Functions to fill in entries in struct proto_ops when a protocol
 * does not implement a particular function.
 */
1631 1632 1633
int sock_no_bind(struct socket *, struct sockaddr *, int);
int sock_no_connect(struct socket *, struct sockaddr *, int, int);
int sock_no_socketpair(struct socket *, struct socket *);
1634
int sock_no_accept(struct socket *, struct socket *, int, bool);
1635
int sock_no_getname(struct socket *, struct sockaddr *, int);
1636 1637 1638 1639 1640
int sock_no_ioctl(struct socket *, unsigned int, unsigned long);
int sock_no_listen(struct socket *, int);
int sock_no_shutdown(struct socket *, int);
int sock_no_getsockopt(struct socket *, int , int, char __user *, int __user *);
int sock_no_setsockopt(struct socket *, int, int, char __user *, unsigned int);
1641
int sock_no_sendmsg(struct socket *, struct msghdr *, size_t);
1642
int sock_no_sendmsg_locked(struct sock *sk, struct msghdr *msg, size_t len);
1643
int sock_no_recvmsg(struct socket *, struct msghdr *, size_t, int);
1644 1645 1646 1647
int sock_no_mmap(struct file *file, struct socket *sock,
		 struct vm_area_struct *vma);
ssize_t sock_no_sendpage(struct socket *sock, struct page *page, int offset,
			 size_t size, int flags);
1648 1649
ssize_t sock_no_sendpage_locked(struct sock *sk, struct page *page,
				int offset, size_t size, int flags);
L
Linus Torvalds 已提交
1650 1651 1652 1653 1654

/*
 * Functions to fill in entries in struct proto_ops when a protocol
 * uses the inet style.
 */
1655
int sock_common_getsockopt(struct socket *sock, int level, int optname,
L
Linus Torvalds 已提交
1656
				  char __user *optval, int __user *optlen);
1657 1658
int sock_common_recvmsg(struct socket *sock, struct msghdr *msg, size_t size,
			int flags);
1659
int sock_common_setsockopt(struct socket *sock, int level, int optname,
1660
				  char __user *optval, unsigned int optlen);
1661
int compat_sock_common_getsockopt(struct socket *sock, int level,
1662
		int optname, char __user *optval, int __user *optlen);
1663
int compat_sock_common_setsockopt(struct socket *sock, int level,
1664
		int optname, char __user *optval, unsigned int optlen);
L
Linus Torvalds 已提交
1665

1666
void sk_common_release(struct sock *sk);
L
Linus Torvalds 已提交
1667 1668 1669 1670

/*
 *	Default socket callbacks and setup code
 */
E
Eric Dumazet 已提交
1671

L
Linus Torvalds 已提交
1672
/* Initialise core socket variables */
1673
void sock_init_data(struct socket *sock, struct sock *sk);
L
Linus Torvalds 已提交
1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702

/*
 * 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)
{
1703
	if (refcount_dec_and_test(&sk->sk_refcnt))
L
Linus Torvalds 已提交
1704 1705
		sk_free(sk);
}
E
Eric Dumazet 已提交
1706
/* Generic version of sock_put(), dealing with all sockets
1707
 * (TCP_TIMEWAIT, TCP_NEW_SYN_RECV, ESTABLISHED...)
E
Eric Dumazet 已提交
1708 1709
 */
void sock_gen_put(struct sock *sk);
L
Linus Torvalds 已提交
1710

1711
int __sk_receive_skb(struct sock *sk, struct sk_buff *skb, const int nested,
1712
		     unsigned int trim_cap, bool refcounted);
1713 1714 1715
static inline int sk_receive_skb(struct sock *sk, struct sk_buff *skb,
				 const int nested)
{
1716
	return __sk_receive_skb(sk, skb, nested, 1, true);
1717
}
1718

K
Krishna Kumar 已提交
1719 1720
static inline void sk_tx_queue_set(struct sock *sk, int tx_queue)
{
1721 1722 1723
	/* sk_tx_queue_mapping accept only upto a 16-bit value */
	if (WARN_ON_ONCE((unsigned short)tx_queue >= USHRT_MAX))
		return;
K
Krishna Kumar 已提交
1724 1725 1726
	sk->sk_tx_queue_mapping = tx_queue;
}

1727 1728
#define NO_QUEUE_MAPPING	USHRT_MAX

K
Krishna Kumar 已提交
1729 1730
static inline void sk_tx_queue_clear(struct sock *sk)
{
1731
	sk->sk_tx_queue_mapping = NO_QUEUE_MAPPING;
K
Krishna Kumar 已提交
1732 1733 1734 1735
}

static inline int sk_tx_queue_get(const struct sock *sk)
{
1736 1737 1738 1739
	if (sk && sk->sk_tx_queue_mapping != NO_QUEUE_MAPPING)
		return sk->sk_tx_queue_mapping;

	return -1;
K
Krishna Kumar 已提交
1740 1741
}

1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762
static inline void sk_rx_queue_set(struct sock *sk, const struct sk_buff *skb)
{
#ifdef CONFIG_XPS
	if (skb_rx_queue_recorded(skb)) {
		u16 rx_queue = skb_get_rx_queue(skb);

		if (WARN_ON_ONCE(rx_queue == NO_QUEUE_MAPPING))
			return;

		sk->sk_rx_queue_mapping = rx_queue;
	}
#endif
}

static inline void sk_rx_queue_clear(struct sock *sk)
{
#ifdef CONFIG_XPS
	sk->sk_rx_queue_mapping = NO_QUEUE_MAPPING;
#endif
}

1763 1764 1765 1766 1767 1768 1769 1770 1771 1772
#ifdef CONFIG_XPS
static inline int sk_rx_queue_get(const struct sock *sk)
{
	if (sk && sk->sk_rx_queue_mapping != NO_QUEUE_MAPPING)
		return sk->sk_rx_queue_mapping;

	return -1;
}
#endif

1773 1774
static inline void sk_set_socket(struct sock *sk, struct socket *sock)
{
K
Krishna Kumar 已提交
1775
	sk_tx_queue_clear(sk);
1776 1777 1778
	sk->sk_socket = sock;
}

E
Eric Dumazet 已提交
1779 1780
static inline wait_queue_head_t *sk_sleep(struct sock *sk)
{
1781 1782
	BUILD_BUG_ON(offsetof(struct socket_wq, wait) != 0);
	return &rcu_dereference_raw(sk->sk_wq)->wait;
E
Eric Dumazet 已提交
1783
}
L
Linus Torvalds 已提交
1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794
/* 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);
1795
	sk_set_socket(sk, NULL);
1796
	sk->sk_wq  = NULL;
L
Linus Torvalds 已提交
1797 1798 1799 1800 1801
	write_unlock_bh(&sk->sk_callback_lock);
}

static inline void sock_graft(struct sock *sk, struct socket *parent)
{
1802
	WARN_ON(parent->sk);
L
Linus Torvalds 已提交
1803
	write_lock_bh(&sk->sk_callback_lock);
1804
	rcu_assign_pointer(sk->sk_wq, parent->wq);
L
Linus Torvalds 已提交
1805
	parent->sk = sk;
1806
	sk_set_socket(sk, parent);
1807
	sk->sk_uid = SOCK_INODE(parent)->i_uid;
1808
	security_sock_graft(sk, parent);
L
Linus Torvalds 已提交
1809 1810 1811
	write_unlock_bh(&sk->sk_callback_lock);
}

1812 1813
kuid_t sock_i_uid(struct sock *sk);
unsigned long sock_i_ino(struct sock *sk);
L
Linus Torvalds 已提交
1814

1815 1816 1817 1818 1819
static inline kuid_t sock_net_uid(const struct net *net, const struct sock *sk)
{
	return sk ? sk->sk_uid : make_kuid(net->user_ns, 0);
}

1820
static inline u32 net_tx_rndhash(void)
1821
{
1822 1823 1824 1825
	u32 v = prandom_u32();

	return v ?: 1;
}
1826

1827 1828 1829
static inline void sk_set_txhash(struct sock *sk)
{
	sk->sk_txhash = net_tx_rndhash();
1830 1831
}

1832 1833 1834 1835 1836 1837
static inline void sk_rethink_txhash(struct sock *sk)
{
	if (sk->sk_txhash)
		sk_set_txhash(sk);
}

L
Linus Torvalds 已提交
1838 1839 1840
static inline struct dst_entry *
__sk_dst_get(struct sock *sk)
{
1841 1842
	return rcu_dereference_check(sk->sk_dst_cache,
				     lockdep_sock_is_held(sk));
L
Linus Torvalds 已提交
1843 1844 1845 1846 1847 1848 1849
}

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

E
Eric Dumazet 已提交
1850 1851
	rcu_read_lock();
	dst = rcu_dereference(sk->sk_dst_cache);
1852 1853
	if (dst && !atomic_inc_not_zero(&dst->__refcnt))
		dst = NULL;
E
Eric Dumazet 已提交
1854
	rcu_read_unlock();
L
Linus Torvalds 已提交
1855 1856 1857
	return dst;
}

E
Eric Dumazet 已提交
1858 1859 1860 1861
static inline void dst_negative_advice(struct sock *sk)
{
	struct dst_entry *ndst, *dst = __sk_dst_get(sk);

1862 1863
	sk_rethink_txhash(sk);

E
Eric Dumazet 已提交
1864 1865 1866 1867 1868
	if (dst && dst->ops->negative_advice) {
		ndst = dst->ops->negative_advice(dst);

		if (ndst != dst) {
			rcu_assign_pointer(sk->sk_dst_cache, ndst);
Z
ZHAO Gang 已提交
1869
			sk_tx_queue_clear(sk);
1870
			sk->sk_dst_pending_confirm = 0;
E
Eric Dumazet 已提交
1871 1872 1873 1874
		}
	}
}

L
Linus Torvalds 已提交
1875 1876 1877 1878 1879
static inline void
__sk_dst_set(struct sock *sk, struct dst_entry *dst)
{
	struct dst_entry *old_dst;

K
Krishna Kumar 已提交
1880
	sk_tx_queue_clear(sk);
1881
	sk->sk_dst_pending_confirm = 0;
1882 1883
	old_dst = rcu_dereference_protected(sk->sk_dst_cache,
					    lockdep_sock_is_held(sk));
E
Eric Dumazet 已提交
1884
	rcu_assign_pointer(sk->sk_dst_cache, dst);
L
Linus Torvalds 已提交
1885 1886 1887 1888 1889 1890
	dst_release(old_dst);
}

static inline void
sk_dst_set(struct sock *sk, struct dst_entry *dst)
{
1891 1892 1893
	struct dst_entry *old_dst;

	sk_tx_queue_clear(sk);
1894
	sk->sk_dst_pending_confirm = 0;
1895
	old_dst = xchg((__force struct dst_entry **)&sk->sk_dst_cache, dst);
1896
	dst_release(old_dst);
L
Linus Torvalds 已提交
1897 1898 1899 1900 1901
}

static inline void
__sk_dst_reset(struct sock *sk)
{
E
Eric Dumazet 已提交
1902
	__sk_dst_set(sk, NULL);
L
Linus Torvalds 已提交
1903 1904 1905 1906 1907
}

static inline void
sk_dst_reset(struct sock *sk)
{
1908
	sk_dst_set(sk, NULL);
L
Linus Torvalds 已提交
1909 1910
}

1911
struct dst_entry *__sk_dst_check(struct sock *sk, u32 cookie);
L
Linus Torvalds 已提交
1912

1913
struct dst_entry *sk_dst_check(struct sock *sk, u32 cookie);
L
Linus Torvalds 已提交
1914

1915 1916 1917 1918 1919 1920
static inline void sk_dst_confirm(struct sock *sk)
{
	if (!sk->sk_dst_pending_confirm)
		sk->sk_dst_pending_confirm = 1;
}

1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934
static inline void sock_confirm_neigh(struct sk_buff *skb, struct neighbour *n)
{
	if (skb_get_dst_pending_confirm(skb)) {
		struct sock *sk = skb->sk;
		unsigned long now = jiffies;

		/* avoid dirtying neighbour */
		if (n->confirmed != now)
			n->confirmed = now;
		if (sk && sk->sk_dst_pending_confirm)
			sk->sk_dst_pending_confirm = 0;
	}
}

1935 1936
bool sk_mc_loop(struct sock *sk);

E
Eric Dumazet 已提交
1937
static inline bool sk_can_gso(const struct sock *sk)
1938 1939 1940 1941
{
	return net_gso_ok(sk->sk_route_caps, sk->sk_gso_type);
}

1942
void sk_setup_caps(struct sock *sk, struct dst_entry *dst);
1943

1944
static inline void sk_nocaps_add(struct sock *sk, netdev_features_t flags)
E
Eric Dumazet 已提交
1945 1946 1947 1948 1949
{
	sk->sk_route_nocaps |= flags;
	sk->sk_route_caps &= ~flags;
}

1950
static inline int skb_do_copy_data_nocache(struct sock *sk, struct sk_buff *skb,
1951
					   struct iov_iter *from, char *to,
1952
					   int copy, int offset)
1953 1954
{
	if (skb->ip_summed == CHECKSUM_NONE) {
1955
		__wsum csum = 0;
1956
		if (!csum_and_copy_from_iter_full(to, copy, &csum, from))
1957
			return -EFAULT;
1958
		skb->csum = csum_block_add(skb->csum, csum, offset);
1959
	} else if (sk->sk_route_caps & NETIF_F_NOCACHE_COPY) {
1960
		if (!copy_from_iter_full_nocache(to, copy, from))
1961
			return -EFAULT;
1962
	} else if (!copy_from_iter_full(to, copy, from))
1963 1964 1965 1966 1967 1968
		return -EFAULT;

	return 0;
}

static inline int skb_add_data_nocache(struct sock *sk, struct sk_buff *skb,
1969
				       struct iov_iter *from, int copy)
1970
{
1971
	int err, offset = skb->len;
1972

1973 1974
	err = skb_do_copy_data_nocache(sk, skb, from, skb_put(skb, copy),
				       copy, offset);
1975
	if (err)
1976
		__skb_trim(skb, offset);
1977 1978 1979 1980

	return err;
}

1981
static inline int skb_copy_to_page_nocache(struct sock *sk, struct iov_iter *from,
1982 1983 1984 1985 1986 1987
					   struct sk_buff *skb,
					   struct page *page,
					   int off, int copy)
{
	int err;

1988 1989
	err = skb_do_copy_data_nocache(sk, skb, from, page_address(page) + off,
				       copy, skb->len);
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000
	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;
}

2001 2002 2003 2004 2005 2006 2007 2008
/**
 * 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)
{
2009
	return refcount_read(&sk->sk_wmem_alloc) - 1;
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028
}

/**
 * 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
2029
static inline bool sk_has_allocations(const struct sock *sk)
2030 2031 2032 2033
{
	return sk_wmem_alloc_get(sk) || sk_rmem_alloc_get(sk);
}

2034
/**
H
Herbert Xu 已提交
2035
 * skwq_has_sleeper - check if there are any waiting processes
R
Randy Dunlap 已提交
2036
 * @wq: struct socket_wq
2037
 *
2038
 * Returns true if socket_wq has waiting processes
2039
 *
H
Herbert Xu 已提交
2040
 * The purpose of the skwq_has_sleeper and sock_poll_wait is to wrap the memory
2041 2042
 * barrier call. They were added due to the race found within the tcp code.
 *
2043
 * Consider following tcp code paths::
2044
 *
2045 2046
 *   CPU1                CPU2
 *   sys_select          receive packet
2047 2048 2049 2050 2051
 *   ...                 ...
 *   __add_wait_queue    update tp->rcv_nxt
 *   ...                 ...
 *   tp->rcv_nxt check   sock_def_readable
 *   ...                 {
2052 2053 2054 2055
 *   schedule               rcu_read_lock();
 *                          wq = rcu_dereference(sk->sk_wq);
 *                          if (wq && waitqueue_active(&wq->wait))
 *                              wake_up_interruptible(&wq->wait)
2056 2057 2058 2059 2060 2061 2062
 *                          ...
 *                       }
 *
 * 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.
2063
 *
2064
 */
H
Herbert Xu 已提交
2065
static inline bool skwq_has_sleeper(struct socket_wq *wq)
2066
{
H
Herbert Xu 已提交
2067
	return wq && wq_has_sleeper(&wq->wait);
2068 2069 2070 2071 2072
}

/**
 * sock_poll_wait - place memory barrier behind the poll_wait call.
 * @filp:           file
2073
 * @sock:           socket to wait on
2074 2075
 * @p:              poll_table
 *
2076
 * See the comments in the wq_has_sleeper function.
2077 2078 2079 2080 2081 2082
 *
 * Do not derive sock from filp->private_data here. An SMC socket establishes
 * an internal TCP socket that is used in the fallback case. All socket
 * operations on the SMC socket are then forwarded to the TCP socket. In case of
 * poll, the filp->private_data pointer references the SMC socket because the
 * TCP socket has no file assigned.
2083
 */
2084 2085
static inline void sock_poll_wait(struct file *filp, struct socket *sock,
				  poll_table *p)
2086
{
2087 2088
	if (!poll_does_not_wait(p)) {
		poll_wait(filp, &sock->wq->wait, p);
E
Eric Dumazet 已提交
2089
		/* We need to be sure we are in sync with the
2090 2091
		 * socket flags modification.
		 *
2092
		 * This memory barrier is paired in the wq_has_sleeper.
E
Eric Dumazet 已提交
2093
		 */
2094 2095 2096 2097
		smp_mb();
	}
}

2098 2099 2100 2101 2102 2103 2104 2105
static inline void skb_set_hash_from_sk(struct sk_buff *skb, struct sock *sk)
{
	if (sk->sk_txhash) {
		skb->l4_hash = 1;
		skb->hash = sk->sk_txhash;
	}
}

2106 2107
void skb_set_owner_w(struct sk_buff *skb, struct sock *sk);

L
Linus Torvalds 已提交
2108
/*
E
Eric Dumazet 已提交
2109
 *	Queue a received datagram if it will fit. Stream and sequenced
L
Linus Torvalds 已提交
2110 2111 2112
 *	protocols can't normally use this as they need to fit buffers in
 *	and play with them.
 *
E
Eric Dumazet 已提交
2113
 *	Inlined as it's very short and called for pretty much every
L
Linus Torvalds 已提交
2114 2115 2116 2117
 *	packet ever received.
 */
static inline void skb_set_owner_r(struct sk_buff *skb, struct sock *sk)
{
2118
	skb_orphan(skb);
L
Linus Torvalds 已提交
2119 2120 2121
	skb->sk = sk;
	skb->destructor = sock_rfree;
	atomic_add(skb->truesize, &sk->sk_rmem_alloc);
2122
	sk_mem_charge(sk, skb->truesize);
L
Linus Torvalds 已提交
2123 2124
}

2125 2126
void sk_reset_timer(struct sock *sk, struct timer_list *timer,
		    unsigned long expires);
L
Linus Torvalds 已提交
2127

2128
void sk_stop_timer(struct sock *sk, struct timer_list *timer);
L
Linus Torvalds 已提交
2129

2130 2131
int __sk_queue_drop_skb(struct sock *sk, struct sk_buff_head *sk_queue,
			struct sk_buff *skb, unsigned int flags,
2132 2133
			void (*destructor)(struct sock *sk,
					   struct sk_buff *skb));
2134
int __sock_queue_rcv_skb(struct sock *sk, struct sk_buff *skb);
2135
int sock_queue_rcv_skb(struct sock *sk, struct sk_buff *skb);
L
Linus Torvalds 已提交
2136

2137
int sock_queue_err_skb(struct sock *sk, struct sk_buff *skb);
2138
struct sk_buff *sock_dequeue_err_skb(struct sock *sk);
L
Linus Torvalds 已提交
2139 2140 2141 2142

/*
 *	Recover an error report and clear atomically
 */
E
Eric Dumazet 已提交
2143

L
Linus Torvalds 已提交
2144 2145
static inline int sock_error(struct sock *sk)
{
2146 2147 2148 2149
	int err;
	if (likely(!sk->sk_err))
		return 0;
	err = xchg(&sk->sk_err, 0);
L
Linus Torvalds 已提交
2150 2151 2152 2153 2154 2155 2156 2157
	return -err;
}

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

	if (!(sk->sk_shutdown & SEND_SHUTDOWN)) {
2158
		amt = sk->sk_sndbuf - refcount_read(&sk->sk_wmem_alloc);
E
Eric Dumazet 已提交
2159
		if (amt < 0)
L
Linus Torvalds 已提交
2160 2161 2162 2163 2164
			amt = 0;
	}
	return amt;
}

2165 2166 2167 2168
/* Note:
 *  We use sk->sk_wq_raw, from contexts knowing this
 *  pointer is not NULL and cannot disappear/change.
 */
2169
static inline void sk_set_bit(int nr, struct sock *sk)
L
Linus Torvalds 已提交
2170
{
2171 2172
	if ((nr == SOCKWQ_ASYNC_NOSPACE || nr == SOCKWQ_ASYNC_WAITDATA) &&
	    !sock_flag(sk, SOCK_FASYNC))
2173 2174
		return;

2175
	set_bit(nr, &sk->sk_wq_raw->flags);
2176 2177 2178 2179
}

static inline void sk_clear_bit(int nr, struct sock *sk)
{
2180 2181
	if ((nr == SOCKWQ_ASYNC_NOSPACE || nr == SOCKWQ_ASYNC_WAITDATA) &&
	    !sock_flag(sk, SOCK_FASYNC))
2182 2183
		return;

2184
	clear_bit(nr, &sk->sk_wq_raw->flags);
2185 2186
}

2187
static inline void sk_wake_async(const struct sock *sk, int how, int band)
L
Linus Torvalds 已提交
2188
{
2189 2190 2191 2192 2193
	if (sock_flag(sk, SOCK_FASYNC)) {
		rcu_read_lock();
		sock_wake_async(rcu_dereference(sk->sk_wq), how, band);
		rcu_read_unlock();
	}
L
Linus Torvalds 已提交
2194 2195
}

2196 2197 2198 2199
/* Since sk_{r,w}mem_alloc sums skb->truesize, even a small frame might
 * need sizeof(sk_buff) + MTU + padding, unless net driver perform copybreak.
 * Note: for send buffers, TCP works better if we can build two skbs at
 * minimum.
E
Eric Dumazet 已提交
2200
 */
2201
#define TCP_SKB_MIN_TRUESIZE	(2048 + SKB_DATA_ALIGN(sizeof(struct sk_buff)))
2202 2203 2204

#define SOCK_MIN_SNDBUF		(TCP_SKB_MIN_TRUESIZE * 2)
#define SOCK_MIN_RCVBUF		 TCP_SKB_MIN_TRUESIZE
L
Linus Torvalds 已提交
2205 2206 2207 2208

static inline void sk_stream_moderate_sndbuf(struct sock *sk)
{
	if (!(sk->sk_userlocks & SOCK_SNDBUF_LOCK)) {
2209
		sk->sk_sndbuf = min(sk->sk_sndbuf, sk->sk_wmem_queued >> 1);
2210
		sk->sk_sndbuf = max_t(u32, sk->sk_sndbuf, SOCK_MIN_SNDBUF);
L
Linus Torvalds 已提交
2211 2212 2213
	}
}

2214 2215
struct sk_buff *sk_stream_alloc_skb(struct sock *sk, int size, gfp_t gfp,
				    bool force_schedule);
L
Linus Torvalds 已提交
2216

2217 2218 2219 2220 2221 2222 2223 2224
/**
 * sk_page_frag - return an appropriate page_frag
 * @sk: socket
 *
 * If socket allocation mode allows current thread to sleep, it means its
 * safe to use the per task page_frag instead of the per socket one.
 */
static inline struct page_frag *sk_page_frag(struct sock *sk)
L
Linus Torvalds 已提交
2225
{
2226
	if (gfpflags_allow_blocking(sk->sk_allocation))
2227
		return &current->task_frag;
L
Linus Torvalds 已提交
2228

2229
	return &sk->sk_frag;
L
Linus Torvalds 已提交
2230 2231
}

2232
bool sk_page_frag_refill(struct sock *sk, struct page_frag *pfrag);
2233

2234
int sk_alloc_sg(struct sock *sk, int len, struct scatterlist *sg,
2235
		int sg_start, int *sg_curr, unsigned int *sg_size,
2236 2237
		int first_coalesce);

L
Linus Torvalds 已提交
2238 2239 2240
/*
 *	Default write policy as shown to user space via poll/select/SIGIO
 */
E
Eric Dumazet 已提交
2241
static inline bool sock_writeable(const struct sock *sk)
L
Linus Torvalds 已提交
2242
{
2243
	return refcount_read(&sk->sk_wmem_alloc) < (sk->sk_sndbuf >> 1);
L
Linus Torvalds 已提交
2244 2245
}

A
Al Viro 已提交
2246
static inline gfp_t gfp_any(void)
L
Linus Torvalds 已提交
2247
{
2248
	return in_softirq() ? GFP_ATOMIC : GFP_KERNEL;
L
Linus Torvalds 已提交
2249 2250
}

E
Eric Dumazet 已提交
2251
static inline long sock_rcvtimeo(const struct sock *sk, bool noblock)
L
Linus Torvalds 已提交
2252 2253 2254 2255
{
	return noblock ? 0 : sk->sk_rcvtimeo;
}

E
Eric Dumazet 已提交
2256
static inline long sock_sndtimeo(const struct sock *sk, bool noblock)
L
Linus Torvalds 已提交
2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273
{
	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;
}

2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287
struct sock_skb_cb {
	u32 dropcount;
};

/* Store sock_skb_cb at the end of skb->cb[] so protocol families
 * using skb->cb[] would keep using it directly and utilize its
 * alignement guarantee.
 */
#define SOCK_SKB_CB_OFFSET ((FIELD_SIZEOF(struct sk_buff, cb) - \
			    sizeof(struct sock_skb_cb)))

#define SOCK_SKB_CB(__skb) ((struct sock_skb_cb *)((__skb)->cb + \
			    SOCK_SKB_CB_OFFSET))

2288
#define sock_skb_cb_check_size(size) \
2289
	BUILD_BUG_ON((size) > SOCK_SKB_CB_OFFSET)
2290

2291 2292 2293
static inline void
sock_skb_set_dropcount(const struct sock *sk, struct sk_buff *skb)
{
2294 2295
	SOCK_SKB_CB(skb)->dropcount = sock_flag(sk, SOCK_RXQ_OVFL) ?
						atomic_read(&sk->sk_drops) : 0;
2296 2297
}

2298 2299 2300 2301 2302 2303 2304
static inline void sk_drops_add(struct sock *sk, const struct sk_buff *skb)
{
	int segs = max_t(u16, 1, skb_shinfo(skb)->gso_segs);

	atomic_add(segs, &sk->sk_drops);
}

2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332
static inline ktime_t sock_read_timestamp(struct sock *sk)
{
#if BITS_PER_LONG==32
	unsigned int seq;
	ktime_t kt;

	do {
		seq = read_seqbegin(&sk->sk_stamp_seq);
		kt = sk->sk_stamp;
	} while (read_seqretry(&sk->sk_stamp_seq, seq));

	return kt;
#else
	return sk->sk_stamp;
#endif
}

static inline void sock_write_timestamp(struct sock *sk, ktime_t kt)
{
#if BITS_PER_LONG==32
	write_seqlock(&sk->sk_stamp_seq);
	sk->sk_stamp = kt;
	write_sequnlock(&sk->sk_stamp_seq);
#else
	sk->sk_stamp = kt;
#endif
}

2333 2334 2335 2336
void __sock_recv_timestamp(struct msghdr *msg, struct sock *sk,
			   struct sk_buff *skb);
void __sock_recv_wifi_status(struct msghdr *msg, struct sock *sk,
			     struct sk_buff *skb);
2337

E
Eric Dumazet 已提交
2338
static inline void
L
Linus Torvalds 已提交
2339 2340
sock_recv_timestamp(struct msghdr *msg, struct sock *sk, struct sk_buff *skb)
{
2341
	ktime_t kt = skb->tstamp;
2342
	struct skb_shared_hwtstamps *hwtstamps = skb_hwtstamps(skb);
2343

2344 2345
	/*
	 * generate control messages if
2346
	 * - receive time stamping in software requested
2347 2348 2349 2350
	 * - software time stamp available and wanted
	 * - hardware time stamps available and wanted
	 */
	if (sock_flag(sk, SOCK_RCVTSTAMP) ||
2351
	    (sk->sk_tsflags & SOF_TIMESTAMPING_RX_SOFTWARE) ||
T
Thomas Gleixner 已提交
2352 2353
	    (kt && sk->sk_tsflags & SOF_TIMESTAMPING_SOFTWARE) ||
	    (hwtstamps->hwtstamp &&
2354
	     (sk->sk_tsflags & SOF_TIMESTAMPING_RAW_HARDWARE)))
2355 2356
		__sock_recv_timestamp(msg, sk, skb);
	else
2357
		sock_write_timestamp(sk, kt);
2358 2359 2360

	if (sock_flag(sk, SOCK_WIFI_STATUS) && skb->wifi_acked_valid)
		__sock_recv_wifi_status(msg, sk, skb);
L
Linus Torvalds 已提交
2361 2362
}

2363 2364
void __sock_recv_ts_and_drops(struct msghdr *msg, struct sock *sk,
			      struct sk_buff *skb);
2365

2366
#define SK_DEFAULT_STAMP (-1L * NSEC_PER_SEC)
2367 2368 2369 2370
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)			| \
2371 2372 2373
			   (1UL << SOCK_RCVTSTAMP))
#define TSFLAGS_ANY	  (SOF_TIMESTAMPING_SOFTWARE			| \
			   SOF_TIMESTAMPING_RAW_HARDWARE)
2374

2375
	if (sk->sk_flags & FLAGS_TS_OR_DROPS || sk->sk_tsflags & TSFLAGS_ANY)
2376
		__sock_recv_ts_and_drops(msg, sk, skb);
2377
	else if (unlikely(sock_flag(sk, SOCK_TIMESTAMP)))
2378
		sock_write_timestamp(sk, skb->tstamp);
2379
	else if (unlikely(sk->sk_stamp == SK_DEFAULT_STAMP))
2380
		sock_write_timestamp(sk, 0);
2381
}
2382

2383
void __sock_tx_timestamp(__u16 tsflags, __u8 *tx_flags);
2384

2385 2386 2387
/**
 * sock_tx_timestamp - checks whether the outgoing packet is to be time stamped
 * @sk:		socket sending this packet
2388
 * @tsflags:	timestamping flags to use
2389 2390
 * @tx_flags:	completed with instructions for time stamping
 *
2391
 * Note: callers should take care of initial ``*tx_flags`` value (usually 0)
2392
 */
2393 2394
static inline void sock_tx_timestamp(const struct sock *sk, __u16 tsflags,
				     __u8 *tx_flags)
2395
{
2396 2397
	if (unlikely(tsflags))
		__sock_tx_timestamp(tsflags, tx_flags);
2398 2399 2400
	if (unlikely(sock_flag(sk, SOCK_WIFI_STATUS)))
		*tx_flags |= SKBTX_WIFI_STATUS;
}
2401

L
Linus Torvalds 已提交
2402 2403
/**
 * sk_eat_skb - Release a skb if it is no longer needed
2404 2405
 * @sk: socket to eat this skb from
 * @skb: socket buffer to eat
L
Linus Torvalds 已提交
2406 2407 2408 2409
 *
 * This routine must be called with interrupts disabled or with the socket
 * locked so that the sk_buff queue operation is ok.
*/
D
Dan Williams 已提交
2410
static inline void sk_eat_skb(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
2411 2412 2413 2414 2415
{
	__skb_unlink(skb, &sk->sk_receive_queue);
	__kfree_skb(skb);
}

2416 2417 2418
static inline
struct net *sock_net(const struct sock *sk)
{
E
Eric Dumazet 已提交
2419
	return read_pnet(&sk->sk_net);
2420 2421 2422
}

static inline
2423
void sock_net_set(struct sock *sk, struct net *net)
2424
{
E
Eric Dumazet 已提交
2425
	write_pnet(&sk->sk_net, net);
2426 2427
}

2428 2429
static inline struct sock *skb_steal_sock(struct sk_buff *skb)
{
2430
	if (skb->sk) {
2431 2432 2433 2434 2435 2436 2437 2438 2439
		struct sock *sk = skb->sk;

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

E
Eric Dumazet 已提交
2440 2441 2442 2443 2444 2445 2446 2447
/* This helper checks if a socket is a full socket,
 * ie _not_ a timewait or request socket.
 */
static inline bool sk_fullsock(const struct sock *sk)
{
	return (1 << sk->sk_state) & ~(TCPF_TIME_WAIT | TCPF_NEW_SYN_RECV);
}

2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463
/* Checks if this SKB belongs to an HW offloaded socket
 * and whether any SW fallbacks are required based on dev.
 */
static inline struct sk_buff *sk_validate_xmit_skb(struct sk_buff *skb,
						   struct net_device *dev)
{
#ifdef CONFIG_SOCK_VALIDATE_XMIT
	struct sock *sk = skb->sk;

	if (sk && sk_fullsock(sk) && sk->sk_validate_xmit_skb)
		skb = sk->sk_validate_xmit_skb(sk, dev, skb);
#endif

	return skb;
}

2464 2465 2466 2467 2468 2469 2470 2471
/* This helper checks if a socket is a LISTEN or NEW_SYN_RECV
 * SYNACK messages can be attached to either ones (depending on SYNCOOKIE)
 */
static inline bool sk_listener(const struct sock *sk)
{
	return (1 << sk->sk_state) & (TCPF_LISTEN | TCPF_NEW_SYN_RECV);
}

2472 2473 2474 2475 2476
void sock_enable_timestamp(struct sock *sk, int flag);
int sock_get_timestamp(struct sock *, struct timeval __user *);
int sock_get_timestampns(struct sock *, struct timespec __user *);
int sock_recv_errqueue(struct sock *sk, struct msghdr *msg, int len, int level,
		       int type);
L
Linus Torvalds 已提交
2477

2478 2479 2480 2481 2482
bool sk_ns_capable(const struct sock *sk,
		   struct user_namespace *user_ns, int cap);
bool sk_capable(const struct sock *sk, int cap);
bool sk_net_capable(const struct sock *sk, int cap);

2483 2484
void sk_get_meminfo(const struct sock *sk, u32 *meminfo);

2485 2486 2487 2488 2489 2490 2491 2492 2493 2494
/* Take into consideration the size of the struct sk_buff overhead in the
 * determination of these values, since that is non-constant across
 * platforms.  This makes socket queueing behavior and performance
 * not depend upon such differences.
 */
#define _SK_MEM_PACKETS		256
#define _SK_MEM_OVERHEAD	SKB_TRUESIZE(256)
#define SK_WMEM_MAX		(_SK_MEM_OVERHEAD * _SK_MEM_PACKETS)
#define SK_RMEM_MAX		(_SK_MEM_OVERHEAD * _SK_MEM_PACKETS)

L
Linus Torvalds 已提交
2495 2496 2497
extern __u32 sysctl_wmem_max;
extern __u32 sysctl_rmem_max;

2498
extern int sysctl_tstamp_allow_data;
2499 2500
extern int sysctl_optmem_max;

2501 2502 2503
extern __u32 sysctl_wmem_default;
extern __u32 sysctl_rmem_default;

2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521
static inline int sk_get_wmem0(const struct sock *sk, const struct proto *proto)
{
	/* Does this proto have per netns sysctl_wmem ? */
	if (proto->sysctl_wmem_offset)
		return *(int *)((void *)sock_net(sk) + proto->sysctl_wmem_offset);

	return *proto->sysctl_wmem;
}

static inline int sk_get_rmem0(const struct sock *sk, const struct proto *proto)
{
	/* Does this proto have per netns sysctl_rmem ? */
	if (proto->sysctl_rmem_offset)
		return *(int *)((void *)sock_net(sk) + proto->sysctl_rmem_offset);

	return *proto->sysctl_rmem;
}

2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532
/* Default TCP Small queue budget is ~1 ms of data (1sec >> 10)
 * Some wifi drivers need to tweak it to get more chunks.
 * They can use this helper from their ndo_start_xmit()
 */
static inline void sk_pacing_shift_update(struct sock *sk, int val)
{
	if (!sk || !sk_fullsock(sk) || sk->sk_pacing_shift == val)
		return;
	sk->sk_pacing_shift = val;
}

2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551
/* if a socket is bound to a device, check that the given device
 * index is either the same or that the socket is bound to an L3
 * master device and the given device index is also enslaved to
 * that L3 master
 */
static inline bool sk_dev_equal_l3scope(struct sock *sk, int dif)
{
	int mdif;

	if (!sk->sk_bound_dev_if || sk->sk_bound_dev_if == dif)
		return true;

	mdif = l3mdev_master_ifindex_by_index(sock_net(sk), dif);
	if (mdif && mdif == sk->sk_bound_dev_if)
		return true;

	return false;
}

L
Linus Torvalds 已提交
2552
#endif	/* _SOCK_H */