sock.h 76.1 KB
Newer Older
1
/* SPDX-License-Identifier: GPL-2.0-or-later */
Linus Torvalds's avatar
Linus Torvalds committed
2 3 4 5 6 7 8 9 10
/*
 * 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
 *
11
 * Authors:	Ross Biro
Linus Torvalds's avatar
Linus Torvalds committed
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
 *		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
 */
#ifndef _SOCK_H
#define _SOCK_H

38
#include <linux/hardirq.h>
39
#include <linux/kernel.h>
Linus Torvalds's avatar
Linus Torvalds committed
40
#include <linux/list.h>
41
#include <linux/list_nulls.h>
Linus Torvalds's avatar
Linus Torvalds committed
42 43
#include <linux/timer.h>
#include <linux/cache.h>
44
#include <linux/bitops.h>
45
#include <linux/lockdep.h>
Linus Torvalds's avatar
Linus Torvalds committed
46 47
#include <linux/netdevice.h>
#include <linux/skbuff.h>	/* struct sk_buff */
Al Viro's avatar
Al Viro committed
48
#include <linux/mm.h>
Linus Torvalds's avatar
Linus Torvalds committed
49
#include <linux/security.h>
50
#include <linux/slab.h>
51
#include <linux/uaccess.h>
52
#include <linux/page_counter.h>
53
#include <linux/memcontrol.h>
54
#include <linux/static_key.h>
Al Viro's avatar
Al Viro committed
55
#include <linux/sched.h>
56
#include <linux/wait.h>
57
#include <linux/cgroup-defs.h>
58
#include <linux/rbtree.h>
Linus Torvalds's avatar
Linus Torvalds committed
59
#include <linux/filter.h>
60
#include <linux/rculist_nulls.h>
61
#include <linux/poll.h>
62
#include <linux/sockptr.h>
Linus Torvalds's avatar
Linus Torvalds committed
63

64
#include <linux/atomic.h>
65
#include <linux/refcount.h>
Linus Torvalds's avatar
Linus Torvalds committed
66 67
#include <net/dst.h>
#include <net/checksum.h>
68
#include <net/tcp_states.h>
69
#include <linux/net_tstamp.h>
70
#include <net/l3mdev.h>
Linus Torvalds's avatar
Linus Torvalds committed
71 72 73 74 75 76 77 78 79 80 81 82 83

/*
 * 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
84
/* Validate arguments and do nothing */
85
static inline __printf(2, 3)
86
void SOCK_DEBUG(const struct sock *sk, const char *msg, ...)
87 88
{
}
Linus Torvalds's avatar
Linus Torvalds committed
89 90 91 92 93 94 95 96
#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;
97
	int			owned;
Linus Torvalds's avatar
Linus Torvalds committed
98
	wait_queue_head_t	wq;
99 100 101 102 103 104 105 106 107
	/*
	 * 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
Linus Torvalds's avatar
Linus Torvalds committed
108 109 110
} socket_lock_t;

struct sock;
111
struct proto;
112
struct net;
Linus Torvalds's avatar
Linus Torvalds committed
113

114 115 116
typedef __u32 __bitwise __portpair;
typedef __u64 __bitwise __addrpair;

Linus Torvalds's avatar
Linus Torvalds committed
117
/**
118
 *	struct sock_common - minimal network layer representation of sockets
119 120
 *	@skc_daddr: Foreign IPv4 addr
 *	@skc_rcv_saddr: Bound local IPv4 addr
121
 *	@skc_addrpair: 8-byte-aligned __u64 union of @skc_daddr & @skc_rcv_saddr
Eric Dumazet's avatar
Eric Dumazet committed
122
 *	@skc_hash: hash value used with various protocol lookup tables
123
 *	@skc_u16hashes: two u16 hash values used by UDP lookup tables
124 125
 *	@skc_dport: placeholder for inet_dport/tw_dport
 *	@skc_num: placeholder for inet_num/tw_num
126
 *	@skc_portpair: __u32 union of @skc_dport & @skc_num
127 128 129
 *	@skc_family: network address family
 *	@skc_state: Connection state
 *	@skc_reuse: %SO_REUSEADDR setting
Tom Herbert's avatar
Tom Herbert committed
130
 *	@skc_reuseport: %SO_REUSEPORT setting
131 132
 *	@skc_ipv6only: socket is IPV6 only
 *	@skc_net_refcnt: socket is using net ref counting
133 134
 *	@skc_bound_dev_if: bound device index if != 0
 *	@skc_bind_node: bind hash linkage for various protocol lookup tables
135
 *	@skc_portaddr_node: second hash linkage for UDP/UDP-Lite protocol
136
 *	@skc_prot: protocol handlers inside a network family
137
 *	@skc_net: reference to the network namespace of this socket
138 139 140
 *	@skc_v6_daddr: IPV6 destination address
 *	@skc_v6_rcv_saddr: IPV6 source address
 *	@skc_cookie: socket's cookie value
141 142 143
 *	@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
144
 *	@skc_rx_queue_mapping: rx queue number for this connection
145 146 147
 *	@skc_flags: place holder for sk_flags
 *		%SO_LINGER (l_onoff), %SO_BROADCAST, %SO_KEEPALIVE,
 *		%SO_OOBINLINE settings, %SO_TIMESTAMPING settings
148 149 150 151
 *	@skc_listener: connection request listener socket (aka rsk_listener)
 *		[union with @skc_flags]
 *	@skc_tw_dr: (aka tw_dr) ptr to &struct inet_timewait_death_row
 *		[union with @skc_flags]
152
 *	@skc_incoming_cpu: record/match cpu processing incoming packets
153 154 155 156
 *	@skc_rcv_wnd: (aka rsk_rcv_wnd) TCP receive window size (possibly scaled)
 *		[union with @skc_incoming_cpu]
 *	@skc_tw_rcv_nxt: (aka tw_rcv_nxt) TCP window next expected seq number
 *		[union with @skc_incoming_cpu]
157
 *	@skc_refcnt: reference count
158 159
 *
 *	This is the minimal network layer representation of sockets, the header
160 161
 *	for struct sock and struct inet_timewait_sock.
 */
Linus Torvalds's avatar
Linus Torvalds committed
162
struct sock_common {
163
	/* skc_daddr and skc_rcv_saddr must be grouped on a 8 bytes aligned
Eric Dumazet's avatar
Eric Dumazet committed
164
	 * address on 64bit arches : cf INET_MATCH()
Eric Dumazet's avatar
Eric Dumazet committed
165
	 */
166
	union {
167
		__addrpair	skc_addrpair;
168 169 170 171 172
		struct {
			__be32	skc_daddr;
			__be32	skc_rcv_saddr;
		};
	};
173 174 175 176
	union  {
		unsigned int	skc_hash;
		__u16		skc_u16hashes[2];
	};
177 178
	/* skc_dport && skc_num must be grouped as well */
	union {
179
		__portpair	skc_portpair;
180 181 182 183 184 185
		struct {
			__be16	skc_dport;
			__u16	skc_num;
		};
	};

Eric Dumazet's avatar
Eric Dumazet committed
186 187
	unsigned short		skc_family;
	volatile unsigned char	skc_state;
Tom Herbert's avatar
Tom Herbert committed
188
	unsigned char		skc_reuse:4;
189 190
	unsigned char		skc_reuseport:1;
	unsigned char		skc_ipv6only:1;
191
	unsigned char		skc_net_refcnt:1;
Eric Dumazet's avatar
Eric Dumazet committed
192
	int			skc_bound_dev_if;
193 194
	union {
		struct hlist_node	skc_bind_node;
195
		struct hlist_node	skc_portaddr_node;
196
	};
197
	struct proto		*skc_prot;
198
	possible_net_t		skc_net;
199 200 201 202 203 204

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

205 206
	atomic64_t		skc_cookie;

207 208 209 210 211 212 213 214 215 216
	/* 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 */
	};
217 218 219 220
	/*
	 * fields between dontcopy_begin/dontcopy_end
	 * are not copied in sock_copy()
	 */
221
	/* private: */
222
	int			skc_dontcopy_begin[0];
223
	/* public: */
224 225 226 227
	union {
		struct hlist_node	skc_node;
		struct hlist_nulls_node skc_nulls_node;
	};
228
	unsigned short		skc_tx_queue_mapping;
229 230 231
#ifdef CONFIG_XPS
	unsigned short		skc_rx_queue_mapping;
#endif
232 233 234
	union {
		int		skc_incoming_cpu;
		u32		skc_rcv_wnd;
235
		u32		skc_tw_rcv_nxt; /* struct tcp_timewait_sock  */
236
	};
237

238
	refcount_t		skc_refcnt;
239
	/* private: */
240
	int                     skc_dontcopy_end[0];
241 242 243
	union {
		u32		skc_rxhash;
		u32		skc_window_clamp;
244
		u32		skc_tw_snd_nxt; /* struct tcp_timewait_sock */
245
	};
246
	/* public: */
Linus Torvalds's avatar
Linus Torvalds committed
247 248
};

249
struct bpf_local_storage;
250

Linus Torvalds's avatar
Linus Torvalds committed
251 252
/**
  *	struct sock - network layer representation of sockets
253
  *	@__sk_common: shared layout with inet_timewait_sock
254 255 256
  *	@sk_shutdown: mask of %SEND_SHUTDOWN and/or %RCV_SHUTDOWN
  *	@sk_userlocks: %SO_SNDBUF and %SO_RCVBUF settings
  *	@sk_lock:	synchronizer
257
  *	@sk_kern_sock: True if sock is using kernel lock classes
258
  *	@sk_rcvbuf: size of receive buffer in bytes
259
  *	@sk_wq: sock wait queue and async head
260
  *	@sk_rx_dst: receive input route used by early demux
261
  *	@sk_dst_cache: destination cache
262
  *	@sk_dst_pending_confirm: need to confirm neighbour
263
  *	@sk_policy: flow policy
264
  *	@sk_rx_skb_cache: cache copy of recently accessed RX skb
265 266
  *	@sk_receive_queue: incoming packets
  *	@sk_wmem_alloc: transmit queue bytes committed
267
  *	@sk_tsq_flags: TCP Small Queues flags
268 269 270 271
  *	@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
272
  *	@sk_napi_id: id of the last napi context to receive data for sk
273
  *	@sk_ll_usec: usecs to busypoll when there is no data
274
  *	@sk_allocation: allocation mode
275
  *	@sk_pacing_rate: Pacing rate (if supported by transport/packet scheduler)
276
  *	@sk_pacing_status: Pacing status (requested, handled by sch_fq)
277
  *	@sk_max_pacing_rate: Maximum pacing rate (%SO_MAX_PACING_RATE)
278
  *	@sk_sndbuf: size of send buffer in bytes
279
  *	@__sk_flags_offset: empty field used to determine location of bitfield
280
  *	@sk_padding: unused element for alignment
281 282
  *	@sk_no_check_tx: %SO_NO_CHECK setting, set checksum in TX packets
  *	@sk_no_check_rx: allow zero checksum in RX packets
283
  *	@sk_route_caps: route capabilities (e.g. %NETIF_F_TSO)
284
  *	@sk_route_nocaps: forbidden route capabilities (e.g NETIF_F_GSO_MASK)
285 286
  *	@sk_route_forced_caps: static, forced route capabilities
  *		(set in tcp_init_sock())
287
  *	@sk_gso_type: GSO type (e.g. %SKB_GSO_TCPV4)
288
  *	@sk_gso_max_size: Maximum GSO segment size to build
289
  *	@sk_gso_max_segs: Maximum number of GSO segments
290
  *	@sk_pacing_shift: scaling factor for TCP Small Queues
291 292 293 294
  *	@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
Wang Chen's avatar
Wang Chen committed
295 296
  *	@sk_prot_creator: sk_prot of original sock creator (see ipv6_setsockopt,
  *			  IPV6_ADDRFORM for instance)
297
  *	@sk_err: last error
Wang Chen's avatar
Wang Chen committed
298 299
  *	@sk_err_soft: errors that don't cause failure but are the cause of a
  *		      persistent failure not just 'timed out'
Eric Dumazet's avatar
Eric Dumazet committed
300
  *	@sk_drops: raw/udp drops counter
301 302
  *	@sk_ack_backlog: current listen backlog
  *	@sk_max_ack_backlog: listen backlog set in listen()
303
  *	@sk_uid: user id of owner
304 305 306
  *	@sk_priority: %SO_PRIORITY setting
  *	@sk_type: socket type (%SOCK_STREAM, etc)
  *	@sk_protocol: which protocol this socket belongs in this network family
307 308
  *	@sk_peer_pid: &struct pid for this socket's peer
  *	@sk_peer_cred: %SO_PEERCRED setting
309 310 311
  *	@sk_rcvlowat: %SO_RCVLOWAT setting
  *	@sk_rcvtimeo: %SO_RCVTIMEO setting
  *	@sk_sndtimeo: %SO_SNDTIMEO setting
312
  *	@sk_txhash: computed flow hash for use on transmit
313 314 315
  *	@sk_filter: socket filtering instructions
  *	@sk_timer: sock cleanup timer
  *	@sk_stamp: time stamp of last packet received
316
  *	@sk_stamp_seq: lock for accessing sk_stamp on 32 bit architectures only
317
  *	@sk_tsflags: SO_TIMESTAMPING socket options
318
  *	@sk_tskey: counter to disambiguate concurrent tstamp requests
Willem de Bruijn's avatar
Willem de Bruijn committed
319
  *	@sk_zckey: counter to order MSG_ZEROCOPY notifications
320 321
  *	@sk_socket: Identd and reporting IO signals
  *	@sk_user_data: RPC layer private data
322
  *	@sk_frag: cached page frag
323
  *	@sk_peek_off: current peek_offset value
324
  *	@sk_send_head: front of stuff to transmit
325 326
  *	@tcp_rtx_queue: TCP re-transmit queue [union with @sk_send_head]
  *	@sk_tx_skb_cache: cache copy of recently accessed TX skb
327
  *	@sk_security: used by security modules
328
  *	@sk_mark: generic packet mark
329
  *	@sk_cgrp_data: cgroup data for this cgroup
330
  *	@sk_memcg: this socket's memory cgroup association
331 332 333 334 335 336
  *	@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
337
  *	@sk_validate_xmit_skb: ptr to an optional validate function
338
  *	@sk_destruct: called at sock freeing time, i.e. when all refcnt == 0
339
  *	@sk_reuseport_cb: reuseport group container
340
  *	@sk_bpf_storage: ptr to cache and control for bpf_sk_storage
341
  *	@sk_rcu: used during RCU grace period
342 343
  *	@sk_clockid: clockid used by time-based scheduling (SO_TXTIME)
  *	@sk_txtime_deadline_mode: set deadline mode for SO_TXTIME
344
  *	@sk_txtime_report_errors: set report errors mode for SO_TXTIME
345
  *	@sk_txtime_unused: unused txtime flags
346
  */
Linus Torvalds's avatar
Linus Torvalds committed
347 348
struct sock {
	/*
349
	 * Now struct inet_timewait_sock also uses sock_common, so please just
Linus Torvalds's avatar
Linus Torvalds committed
350 351 352
	 * don't add nothing before this first member (__sk_common) --acme
	 */
	struct sock_common	__sk_common;
Eric Dumazet's avatar
Eric Dumazet committed
353 354 355
#define sk_node			__sk_common.skc_node
#define sk_nulls_node		__sk_common.skc_nulls_node
#define sk_refcnt		__sk_common.skc_refcnt
356
#define sk_tx_queue_mapping	__sk_common.skc_tx_queue_mapping
357 358 359
#ifdef CONFIG_XPS
#define sk_rx_queue_mapping	__sk_common.skc_rx_queue_mapping
#endif
Eric Dumazet's avatar
Eric Dumazet committed
360

361 362
#define sk_dontcopy_begin	__sk_common.skc_dontcopy_begin
#define sk_dontcopy_end		__sk_common.skc_dontcopy_end
Eric Dumazet's avatar
Eric Dumazet committed
363
#define sk_hash			__sk_common.skc_hash
364
#define sk_portpair		__sk_common.skc_portpair
Eric Dumazet's avatar
Eric Dumazet committed
365 366
#define sk_num			__sk_common.skc_num
#define sk_dport		__sk_common.skc_dport
367 368 369
#define sk_addrpair		__sk_common.skc_addrpair
#define sk_daddr		__sk_common.skc_daddr
#define sk_rcv_saddr		__sk_common.skc_rcv_saddr
Linus Torvalds's avatar
Linus Torvalds committed
370 371 372
#define sk_family		__sk_common.skc_family
#define sk_state		__sk_common.skc_state
#define sk_reuse		__sk_common.skc_reuse
Tom Herbert's avatar
Tom Herbert committed
373
#define sk_reuseport		__sk_common.skc_reuseport
374
#define sk_ipv6only		__sk_common.skc_ipv6only
375
#define sk_net_refcnt		__sk_common.skc_net_refcnt
Linus Torvalds's avatar
Linus Torvalds committed
376 377
#define sk_bound_dev_if		__sk_common.skc_bound_dev_if
#define sk_bind_node		__sk_common.skc_bind_node
378
#define sk_prot			__sk_common.skc_prot
379
#define sk_net			__sk_common.skc_net
380 381
#define sk_v6_daddr		__sk_common.skc_v6_daddr
#define sk_v6_rcv_saddr	__sk_common.skc_v6_rcv_saddr
382
#define sk_cookie		__sk_common.skc_cookie
383
#define sk_incoming_cpu		__sk_common.skc_incoming_cpu
384
#define sk_flags		__sk_common.skc_flags
385
#define sk_rxhash		__sk_common.skc_rxhash
386

Linus Torvalds's avatar
Linus Torvalds committed
387
	socket_lock_t		sk_lock;
388 389 390
	atomic_t		sk_drops;
	int			sk_rcvlowat;
	struct sk_buff_head	sk_error_queue;
391
	struct sk_buff		*sk_rx_skb_cache;
392
	struct sk_buff_head	sk_receive_queue;
393 394 395 396
	/*
	 * 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.
397 398 399
	 * Note : rmem_alloc is in this structure to fill a hole
	 * on 64bit arches, not because its logically part of
	 * backlog.
400 401
	 */
	struct {
402 403 404 405
		atomic_t	rmem_alloc;
		int		len;
		struct sk_buff	*head;
		struct sk_buff	*tail;
406
	} sk_backlog;
407
#define sk_rmem_alloc sk_backlog.rmem_alloc
408

409
	int			sk_forward_alloc;
410
#ifdef CONFIG_NET_RX_BUSY_POLL
411
	unsigned int		sk_ll_usec;
412 413
	/* ===== mostly read cache line ===== */
	unsigned int		sk_napi_id;
414 415 416 417
#endif
	int			sk_rcvbuf;

	struct sk_filter __rcu	*sk_filter;
418 419
	union {
		struct socket_wq __rcu	*sk_wq;
420
		/* private: */
421
		struct socket_wq	*sk_wq_raw;
422
		/* public: */
423
	};
424
#ifdef CONFIG_XFRM
425
	struct xfrm_policy __rcu *sk_policy[2];
426
#endif
Eric Dumazet's avatar
Eric Dumazet committed
427
	struct dst_entry	*sk_rx_dst;
428
	struct dst_entry __rcu	*sk_dst_cache;
Linus Torvalds's avatar
Linus Torvalds committed
429
	atomic_t		sk_omem_alloc;
430
	int			sk_sndbuf;
431 432 433

	/* ===== cache line for TX ===== */
	int			sk_wmem_queued;
434
	refcount_t		sk_wmem_alloc;
435
	unsigned long		sk_tsq_flags;
436 437 438 439
	union {
		struct sk_buff	*sk_send_head;
		struct rb_root	tcp_rtx_queue;
	};
440
	struct sk_buff		*sk_tx_skb_cache;
Linus Torvalds's avatar
Linus Torvalds committed
441
	struct sk_buff_head	sk_write_queue;
442 443
	__s32			sk_peek_off;
	int			sk_write_pending;
444
	__u32			sk_dst_pending_confirm;
445
	u32			sk_pacing_status; /* see enum sk_pacing */
446 447 448 449
	long			sk_sndtimeo;
	struct timer_list	sk_timer;
	__u32			sk_priority;
	__u32			sk_mark;
450 451
	unsigned long		sk_pacing_rate; /* bytes per second */
	unsigned long		sk_max_pacing_rate;
452 453 454
	struct page_frag	sk_frag;
	netdev_features_t	sk_route_caps;
	netdev_features_t	sk_route_nocaps;
455
	netdev_features_t	sk_route_forced_caps;
456 457 458 459
	int			sk_gso_type;
	unsigned int		sk_gso_max_size;
	gfp_t			sk_allocation;
	__u32			sk_txhash;
460 461 462 463 464

	/*
	 * Because of non atomicity rules, all
	 * changes are protected by socket lock.
	 */
465
	u8			sk_padding : 1,
466
				sk_kern_sock : 1,
467 468
				sk_no_check_tx : 1,
				sk_no_check_rx : 1,
469
				sk_userlocks : 4;
470
	u8			sk_pacing_shift;
471 472 473
	u16			sk_type;
	u16			sk_protocol;
	u16			sk_gso_max_segs;
Linus Torvalds's avatar
Linus Torvalds committed
474
	unsigned long	        sk_lingertime;
475
	struct proto		*sk_prot_creator;
Linus Torvalds's avatar
Linus Torvalds committed
476 477 478
	rwlock_t		sk_callback_lock;
	int			sk_err,
				sk_err_soft;
479 480
	u32			sk_ack_backlog;
	u32			sk_max_ack_backlog;
481
	kuid_t			sk_uid;
482 483
	struct pid		*sk_peer_pid;
	const struct cred	*sk_peer_cred;
Linus Torvalds's avatar
Linus Torvalds committed
484
	long			sk_rcvtimeo;
485
	ktime_t			sk_stamp;
486 487 488
#if BITS_PER_LONG==32
	seqlock_t		sk_stamp_seq;
#endif
489
	u16			sk_tsflags;
490
	u8			sk_shutdown;
491
	u32			sk_tskey;
Willem de Bruijn's avatar
Willem de Bruijn committed
492
	atomic_t		sk_zckey;
493 494 495

	u8			sk_clockid;
	u8			sk_txtime_deadline_mode : 1,
496 497
				sk_txtime_report_errors : 1,
				sk_txtime_unused : 6;
498

Linus Torvalds's avatar
Linus Torvalds committed
499 500
	struct socket		*sk_socket;
	void			*sk_user_data;
Alexey Dobriyan's avatar
Alexey Dobriyan committed
501
#ifdef CONFIG_SECURITY
Linus Torvalds's avatar
Linus Torvalds committed
502
	void			*sk_security;
Alexey Dobriyan's avatar
Alexey Dobriyan committed
503
#endif
504
	struct sock_cgroup_data	sk_cgrp_data;
505
	struct mem_cgroup	*sk_memcg;
Linus Torvalds's avatar
Linus Torvalds committed
506
	void			(*sk_state_change)(struct sock *sk);
507
	void			(*sk_data_ready)(struct sock *sk);
Linus Torvalds's avatar
Linus Torvalds committed
508 509
	void			(*sk_write_space)(struct sock *sk);
	void			(*sk_error_report)(struct sock *sk);
510 511
	int			(*sk_backlog_rcv)(struct sock *sk,
						  struct sk_buff *skb);
512 513 514 515 516
#ifdef CONFIG_SOCK_VALIDATE_XMIT
	struct sk_buff*		(*sk_validate_xmit_skb)(struct sock *sk,
							struct net_device *dev,
							struct sk_buff *skb);
#endif
Linus Torvalds's avatar
Linus Torvalds committed
517
	void                    (*sk_destruct)(struct sock *sk);
518
	struct sock_reuseport __rcu	*sk_reuseport_cb;
519
#ifdef CONFIG_BPF_SYSCALL
520
	struct bpf_local_storage __rcu	*sk_bpf_storage;
521
#endif
522
	struct rcu_head		sk_rcu;
Linus Torvalds's avatar
Linus Torvalds committed
523 524
};

525 526 527 528 529 530
enum sk_pacing {
	SK_PACING_NONE		= 0,
	SK_PACING_NEEDED	= 1,
	SK_PACING_FQ		= 2,
};

531 532 533 534 535 536
/* Pointer stored in sk_user_data might not be suitable for copying
 * when cloning the socket. For instance, it can point to a reference
 * counted object. sk_user_data bottom bit is set if pointer must not
 * be copied.
 */
#define SK_USER_DATA_NOCOPY	1UL
537 538
#define SK_USER_DATA_BPF	2UL	/* Managed by BPF */
#define SK_USER_DATA_PTRMASK	~(SK_USER_DATA_NOCOPY | SK_USER_DATA_BPF)
539 540 541 542 543 544 545 546 547 548

/**
 * sk_user_data_is_nocopy - Test if sk_user_data pointer must not be copied
 * @sk: socket
 */
static inline bool sk_user_data_is_nocopy(const struct sock *sk)
{
	return ((uintptr_t)sk->sk_user_data & SK_USER_DATA_NOCOPY);
}

549 550
#define __sk_user_data(sk) ((*((void __rcu **)&(sk)->sk_user_data)))

551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568
#define rcu_dereference_sk_user_data(sk)				\
({									\
	void *__tmp = rcu_dereference(__sk_user_data((sk)));		\
	(void *)((uintptr_t)__tmp & SK_USER_DATA_PTRMASK);		\
})
#define rcu_assign_sk_user_data(sk, ptr)				\
({									\
	uintptr_t __tmp = (uintptr_t)(ptr);				\
	WARN_ON_ONCE(__tmp & ~SK_USER_DATA_PTRMASK);			\
	rcu_assign_pointer(__sk_user_data((sk)), __tmp);		\
})
#define rcu_assign_sk_user_data_nocopy(sk, ptr)				\
({									\
	uintptr_t __tmp = (uintptr_t)(ptr);				\
	WARN_ON_ONCE(__tmp & ~SK_USER_DATA_PTRMASK);			\
	rcu_assign_pointer(__sk_user_data((sk)),			\
			   __tmp | SK_USER_DATA_NOCOPY);		\
})
569

570 571 572 573 574 575 576 577 578 579 580
/*
 * 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

581 582
int sk_set_peek_off(struct sock *sk, int val);

583 584
static inline int sk_peek_offset(struct sock *sk, int flags)
{
585
	if (unlikely(flags & MSG_PEEK)) {
586
		return READ_ONCE(sk->sk_peek_off);
587 588 589
	}

	return 0;
590 591 592 593
}

static inline void sk_peek_offset_bwd(struct sock *sk, int val)
{
594 595 596 597 598
	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);
599 600 601 602 603
	}
}

static inline void sk_peek_offset_fwd(struct sock *sk, int val)
{
604
	sk_peek_offset_bwd(sk, -val);
605 606
}

Linus Torvalds's avatar
Linus Torvalds committed
607 608 609
/*
 * Hashed lists helper routines
 */
Li Zefan's avatar
Li Zefan committed
610 611 612 613 614
static inline struct sock *sk_entry(const struct hlist_node *node)
{
	return hlist_entry(node, struct sock, sk_node);
}

615
static inline struct sock *__sk_head(const struct hlist_head *head)
Linus Torvalds's avatar
Linus Torvalds committed
616 617 618 619
{
	return hlist_entry(head->first, struct sock, sk_node);
}

620
static inline struct sock *sk_head(const struct hlist_head *head)
Linus Torvalds's avatar
Linus Torvalds committed
621 622 623 624
{
	return hlist_empty(head) ? NULL : __sk_head(head);
}

625 626 627 628 629 630 631 632 633 634
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);
}

635
static inline struct sock *sk_next(const struct sock *sk)
Linus Torvalds's avatar
Linus Torvalds committed
636
{
Geliang Tang's avatar
Geliang Tang committed
637
	return hlist_entry_safe(sk->sk_node.next, struct sock, sk_node);
Linus Torvalds's avatar
Linus Torvalds committed
638 639
}

640 641 642 643 644 645 646 647
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;
}

648
static inline bool sk_unhashed(const struct sock *sk)
Linus Torvalds's avatar
Linus Torvalds committed
649 650 651 652
{
	return hlist_unhashed(&sk->sk_node);
}

653
static inline bool sk_hashed(const struct sock *sk)
Linus Torvalds's avatar
Linus Torvalds committed
654
{
Akinobu Mita's avatar
Akinobu Mita committed
655
	return !sk_unhashed(sk);
Linus Torvalds's avatar
Linus Torvalds committed
656 657
}

658
static inline void sk_node_init(struct hlist_node *node)
Linus Torvalds's avatar
Linus Torvalds committed
659 660 661 662
{
	node->pprev = NULL;
}

663
static inline void sk_nulls_node_init(struct hlist_nulls_node *node)
664 665 666 667
{
	node->pprev = NULL;
}

668
static inline void __sk_del_node(struct sock *sk)
Linus Torvalds's avatar
Linus Torvalds committed
669 670 671 672
{
	__hlist_del(&sk->sk_node);
}

673
/* NB: equivalent to hlist_del_init_rcu */
674
static inline bool __sk_del_node_init(struct sock *sk)
Linus Torvalds's avatar
Linus Torvalds committed
675 676 677 678
{
	if (sk_hashed(sk)) {
		__sk_del_node(sk);
		sk_node_init(&sk->sk_node);
679
		return true;
Linus Torvalds's avatar
Linus Torvalds committed
680
	}
681
	return false;
Linus Torvalds's avatar
Linus Torvalds committed
682 683 684 685 686 687 688 689
}

/* 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.
 */

690
static __always_inline void sock_hold(struct sock *sk)
Linus Torvalds's avatar
Linus Torvalds committed
691
{
692
	refcount_inc(&sk->sk_refcnt);
Linus Torvalds's avatar
Linus Torvalds committed
693 694 695 696 697
}

/* Ungrab socket in the context, which assumes that socket refcnt
   cannot hit zero, f.e. it is true in context of any socketcall.
 */
698
static __always_inline void __sock_put(struct sock *sk)
Linus Torvalds's avatar
Linus Torvalds committed
699
{
700
	refcount_dec(&sk->sk_refcnt);
Linus Torvalds's avatar
Linus Torvalds committed
701 702
}

703
static inline bool sk_del_node_init(struct sock *sk)
Linus Torvalds's avatar
Linus Torvalds committed
704
{
705
	bool rc = __sk_del_node_init(sk);
Linus Torvalds's avatar
Linus Torvalds committed
706 707 708

	if (rc) {
		/* paranoid for a while -acme */
709
		WARN_ON(refcount_read(&sk->sk_refcnt) == 1);
Linus Torvalds's avatar
Linus Torvalds committed
710 711 712 713
		__sock_put(sk);
	}
	return rc;
}
714
#define sk_del_node_init_rcu(sk)	sk_del_node_init(sk)
Linus Torvalds's avatar
Linus Torvalds committed
715

716
static inline bool __sk_nulls_del_node_init_rcu(struct sock *sk)
717 718
{
	if (sk_hashed(sk)) {
719
		hlist_nulls_del_init_rcu(&sk->sk_nulls_node);
720
		return true;
721
	}
722
	return false;
723 724
}

725
static inline bool sk_nulls_del_node_init_rcu(struct sock *sk)
726
{
727
	bool rc = __sk_nulls_del_node_init_rcu(sk);
728 729 730

	if (rc) {
		/* paranoid for a while -acme */
731
		WARN_ON(refcount_read(&sk->sk_refcnt) == 1);
732 733 734 735 736
		__sock_put(sk);
	}
	return rc;
}

737
static inline void __sk_add_node(struct sock *sk, struct hlist_head *list)
Linus Torvalds's avatar
Linus Torvalds committed
738 739 740 741
{
	hlist_add_head(&sk->sk_node, list);
}

742
static inline void sk_add_node(struct sock *sk, struct hlist_head *list)
Linus Torvalds's avatar
Linus Torvalds committed
743 744 745 746 747
{
	sock_hold(sk);
	__sk_add_node(sk, list);
}

748
static inline void sk_add_node_rcu(struct sock *sk, struct hlist_head *list)
749 750
{
	sock_hold(sk);
751 752 753 754 755
	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);
756 757
}

758 759 760 761 762 763
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);
}

764
static inline void __sk_nulls_add_node_rcu(struct sock *sk, struct hlist_nulls_head *list)
765
{
766
	hlist_nulls_add_head_rcu(&sk->sk_nulls_node, list);
767 768
}

769 770 771 772 773
static inline void __sk_nulls_add_node_tail_rcu(struct sock *sk, struct hlist_nulls_head *list)
{
	hlist_nulls_add_tail_rcu(&sk->sk_nulls_node, list);
}

774
static inline void sk_nulls_add_node_rcu(struct sock *sk, struct hlist_nulls_head *list)
775 776
{
	sock_hold(sk);
777
	__sk_nulls_add_node_rcu(sk, list);
778 779
}

780
static inline void __sk_del_bind_node(struct sock *sk)
Linus Torvalds's avatar
Linus Torvalds committed
781 782 783 784
{
	__hlist_del(&sk->sk_bind_node);
}

785
static inline void sk_add_bind_node(struct sock *sk,
Linus Torvalds's avatar
Linus Torvalds committed
786 787 788 789 790
					struct hlist_head *list)
{
	hlist_add_head(&sk->sk_bind_node, list);
}

791 792 793 794
#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)
795 796 797 798
#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)
799 800
#define sk_for_each_from(__sk) \
	hlist_for_each_entry_from(__sk, sk_node)
801 802 803
#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)
804 805 806 807
#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)
Linus Torvalds's avatar
Linus Torvalds committed
808

809
/**
810
 * sk_for_each_entry_offset_rcu - iterate over a list at a given struct offset
811 812 813 814 815 816
 * @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.
 *
 */
817
#define sk_for_each_entry_offset_rcu(tpos, pos, head, offset)		       \
818
	for (pos = rcu_dereference(hlist_first_rcu(head));		       \
819
	     pos != NULL &&						       \
820
		({ tpos = (typeof(*tpos) *)((void *)pos - offset); 1;});       \
821
	     pos = rcu_dereference(hlist_next_rcu(pos)))
822

823 824 825 826 827 828 829 830 831
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;
}

Linus Torvalds's avatar
Linus Torvalds committed
832 833 834 835 836 837 838 839 840 841 842 843 844 845
/* 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 */
846
	SOCK_RCVTSTAMPNS, /* %SO_TIMESTAMPNS setting */
Linus Torvalds's avatar
Linus Torvalds committed
847 848
	SOCK_LOCALROUTE, /* route locally only, %SO_DONTROUTE setting */
	SOCK_QUEUE_SHRUNK, /* write queue has been shrunk recently */
849
	SOCK_MEMALLOC, /* VM depends on this socket for swapping */
850
	SOCK_TIMESTAMPING_RX_SOFTWARE,  /* %SOF_TIMESTAMPING_RX_SOFTWARE */
Eric Dumazet's avatar
Eric Dumazet committed
851
	SOCK_FASYNC, /* fasync() active */
852
	SOCK_RXQ_OVFL,
853
	SOCK_ZEROCOPY, /* buffers from userspace */
854
	SOCK_WIFI_STATUS, /* push wifi status to userspace */
855 856 857 858
	SOCK_NOFCS, /* Tell NIC not to do the Ethernet FCS.
		     * Will use last 4 bytes of packet sent from
		     * user-space instead.
		     */
859
	SOCK_FILTER_LOCKED, /* Filter cannot be changed anymore */
860
	SOCK_SELECT_ERR_QUEUE, /* Wake select on error queue */
861
	SOCK_RCU_FREE, /* wait rcu grace period in sk_destruct() */
862
	SOCK_TXTIME,
Jason Wang's avatar
Jason Wang committed
863
	SOCK_XDP, /* XDP is attached */
864
	SOCK_TSTAMP_NEW, /* Indicates 64 bit timestamps always */
Linus Torvalds's avatar
Linus Torvalds committed
865 866
};

867 868
#define SK_FLAGS_TIMESTAMP ((1UL << SOCK_TIMESTAMP) | (1UL << SOCK_TIMESTAMPING_RX_SOFTWARE))

869 870 871 872 873
static inline void sock_copy_flags(struct sock *nsk, struct sock *osk)
{
	nsk->sk_flags = osk->sk_flags;
}

Linus Torvalds's avatar
Linus Torvalds committed
874 875 876 877 878 879 880 881 882 883
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);
}

884 885 886 887 888 889 890 891 892
static inline void sock_valbool_flag(struct sock *sk, enum sock_flags bit,
				     int valbool)
{
	if (valbool)
		sock_set_flag(sk, bit);
	else
		sock_reset_flag(sk, bit);
}

Eric Dumazet's avatar
Eric Dumazet committed
893
static inline bool sock_flag(const struct sock *sk, enum sock_flags flag)
Linus Torvalds's avatar
Linus Torvalds committed
894 895 896 897
{
	return test_bit(flag, &sk->sk_flags);
}

898
#ifdef CONFIG_NET
899
DECLARE_STATIC_KEY_FALSE(memalloc_socks_key);
900 901
static inline int sk_memalloc_socks(void)
{
902
	return static_branch_unlikely(&memalloc_socks_key);
903
}
904 905

void __receive_sock(struct file *file);
906 907 908 909 910 911 912
#else

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

913 914
static inline void __receive_sock(struct file *file)
{ }
915 916
#endif

917
static inline gfp_t sk_gfp_mask(const struct sock *sk, gfp_t gfp_mask)
918
{
919
	return gfp_mask | (sk->sk_allocation & __GFP_MEMALLOC);
920 921
}

Linus Torvalds's avatar
Linus Torvalds committed
922 923
static inline void sk_acceptq_removed(struct sock *sk)
{
924
	WRITE_ONCE(sk->sk_ack_backlog, sk->sk_ack_backlog - 1);
Linus Torvalds's avatar
Linus Torvalds committed
925 926 927 928
}

static inline void sk_acceptq_added(struct sock *sk)
{
929
	WRITE_ONCE(sk->sk_ack_backlog, sk->sk_ack_backlog + 1);
Linus Torvalds's avatar
Linus Torvalds committed
930 931
}

932
static inline bool sk_acceptq_is_full(const struct sock *sk)
Linus Torvalds's avatar
Linus Torvalds committed
933
{
934
	return READ_ONCE(sk->sk_ack_backlog) > READ_ONCE(sk->sk_max_ack_backlog);
Linus Torvalds's avatar
Linus Torvalds committed
935 936 937 938 939
}

/*
 * Compute minimal free write space needed to queue new packets.
 */
940
static inline int sk_stream_min_wspace(const struct sock *sk)
Linus Torvalds's avatar
Linus Torvalds committed
941
{
942
	return READ_ONCE(sk->sk_wmem_queued) >> 1;
Linus Torvalds's avatar
Linus Torvalds committed
943 944
}

945
static inline int sk_stream_wspace(const struct sock *sk)
Linus Torvalds's avatar
Linus Torvalds committed
946
{
947 948 949 950 951 952
	return READ_ONCE(sk->sk_sndbuf) - READ_ONCE(sk->sk_wmem_queued);
}

static inline void sk_wmem_queued_add(struct sock *sk, int val)
{
	WRITE_ONCE(sk->sk_wmem_queued, sk->sk_wmem_queued + val);
Linus Torvalds's avatar
Linus Torvalds committed
953 954
}

955
void sk_stream_write_space(struct sock *sk);
Linus Torvalds's avatar
Linus Torvalds committed
956

Zhu Yi's avatar
Zhu Yi committed
957
/* OOB backlog add */
Zhu Yi's avatar
Zhu Yi committed
958
static inline void __sk_add_backlog(struct sock *sk, struct sk_buff *skb)
959
{
960
	/* dont let skb dst not refcounted, we are going to leave rcu lock */
961
	skb_dst_force(skb);
962 963

	if (!sk->sk_backlog.tail)
964
		WRITE_ONCE(sk->sk_backlog.head, skb);
965
	else
966
		sk->sk_backlog.tail->next = skb;
967

968
	WRITE_ONCE(sk->sk_backlog.tail, skb);
969 970
	skb->next = NULL;
}
Linus Torvalds's avatar
Linus Torvalds committed
971

972 973
/*
 * Take into account size of receive queue and backlog queue
Eric Dumazet's avatar
Eric Dumazet committed
974 975
 * Do not take into account this skb truesize,
 * to allow even a single big packet to come.
976
 */
977
static inline bool sk_rcvqueues_full(const struct sock *sk, unsigned int limit)
978 979 980
{
	unsigned int qsize = sk->sk_backlog.len + atomic_read(&sk->sk_rmem_alloc);

981
	return qsize > limit;
982 983
}

Zhu Yi's avatar
Zhu Yi committed
984
/* The per-socket spinlock must be held here. */
985 986
static inline __must_check int sk_add_backlog(struct sock *sk, struct sk_buff *skb,
					      unsigned int limit)
Zhu Yi's avatar
Zhu Yi committed
987
{
988
	if (sk_rcvqueues_full(sk, limit))
Zhu Yi's avatar
Zhu Yi committed
989 990
		return -ENOBUFS;

991 992 993 994 995 996 997 998
	/*
	 * 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;

Zhu Yi's avatar
Zhu Yi committed
999
	__sk_add_backlog(sk, skb);
Zhu Yi's avatar
Zhu Yi committed
1000 1001 1002 1003
	sk->sk_backlog.len += skb->truesize;
	return 0;
}

1004
int __sk_backlog_rcv(struct sock *sk, struct sk_buff *skb);
1005

1006 1007
static inline int sk_backlog_rcv(struct sock *sk, struct sk_buff *skb)
{
1008 1009 1010
	if (sk_memalloc_socks() && skb_pfmemalloc(skb))
		return __sk_backlog_rcv(sk, skb);

1011 1012 1013
	return sk->sk_backlog_rcv(sk, skb);
}

1014 1015
static inline void sk_incoming_cpu_update(struct sock *sk)
{
1016 1017
	int cpu = raw_smp_processor_id();

1018 1019
	if (unlikely(READ_ONCE(sk->sk_incoming_cpu) != cpu))
		WRITE_ONCE(sk->sk_incoming_cpu, cpu);
1020 1021
}

1022
static inline void sock_rps_record_flow_hash(__u32 hash)
1023 1024 1025 1026 1027 1028
{
#ifdef CONFIG_RPS
	struct rps_sock_flow_table *sock_flow_table;

	rcu_read_lock();
	sock_flow_table = rcu_dereference(rps_sock_flow_table);
1029
	rps_record_sock_flow(sock_flow_table, hash);
1030 1031 1032 1033
	rcu_read_unlock();
#endif
}

1034 1035
static inline void sock_rps_record_flow(const struct sock *sk)
{
1036
#ifdef CONFIG_RPS
1037
	if (static_branch_unlikely(&rfs_needed)) {
Eric Dumazet's avatar
Eric Dumazet committed
1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050
		/* 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);
	}
1051
#endif
1052 1053
}

1054 1055
static inline void sock_rps_save_rxhash(struct sock *sk,
					const struct sk_buff *skb)
1056 1057
{
#ifdef CONFIG_RPS
1058
	if (unlikely(sk->sk_rxhash != skb->hash))
1059
		sk->sk_rxhash = skb->hash;
1060 1061 1062
#endif
}

1063 1064 1065 1066 1067 1068 1069
static inline void sock_rps_reset_rxhash(struct sock *sk)
{
#ifdef CONFIG_RPS
	sk->sk_rxhash = 0;
#endif
}

1070
#define sk_wait_event(__sk, __timeo, __condition, __wait)		\
1071 1072 1073 1074
	({	int __rc;						\
		release_sock(__sk);					\
		__rc = __condition;					\
		if (!__rc) {						\
1075 1076 1077
			*(__timeo) = wait_woken(__wait,			\
						TASK_INTERRUPTIBLE,	\
						*(__timeo));		\
1078
		}							\
1079
		sched_annotate_sleep();					\
1080 1081 1082 1083
		lock_sock(__sk);					\
		__rc = __condition;					\
		__rc;							\
	})
Linus Torvalds's avatar
Linus Torvalds committed
1084

1085 1086 1087 1088 1089 1090 1091
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);
Linus Torvalds's avatar
Linus Torvalds committed
1092

1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103
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;
}

1104
int sk_wait_data(struct sock *sk, long *timeo, const struct sk_buff *skb);
Linus Torvalds's avatar
Linus Torvalds committed
1105

1106
struct request_sock_ops;
1107
struct timewait_sock_ops;
1108
struct inet_hashinfo;
1109
struct raw_hashinfo;
1110
struct smc_hashinfo;
1111
struct module;
1112

1113
/*
1114
 * caches using SLAB_TYPESAFE_BY_RCU should let .next pointer from nulls nodes
1115 1116 1117 1118 1119 1120 1121 1122 1123 1124
 * 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));
}

Linus Torvalds's avatar
Linus Torvalds committed
1125 1126 1127 1128
/* Networking protocol blocks we attach to sockets.
 * socket layer -> transport layer interface
 */
struct proto {
1129
	void			(*close)(struct sock *sk,
Linus Torvalds's avatar
Linus Torvalds committed
1130
					long timeout);
1131 1132 1133
	int			(*pre_connect)(struct sock *sk,
					struct sockaddr *uaddr,
					int addr_len);
Linus Torvalds's avatar
Linus Torvalds committed
1134
	int			(*connect)(struct sock *sk,
1135
					struct sockaddr *uaddr,
Linus Torvalds's avatar
Linus Torvalds committed
1136 1137 1138
					int addr_len);
	int			(*disconnect)(struct sock *sk, int flags);

1139 1140
	struct sock *		(*accept)(struct sock *sk, int flags, int *err,
					  bool kern);
Linus Torvalds's avatar
Linus Torvalds committed
1141 1142 1143 1144

	int			(*ioctl)(struct sock *sk, int cmd,
					 unsigned long arg);
	int			(*init)(struct sock *sk);
1145
	void			(*destroy)(struct sock *sk);
Linus Torvalds's avatar
Linus Torvalds committed
1146
	void			(*shutdown)(struct sock *sk, int how);
1147
	int			(*setsockopt)(struct sock *sk, int level,
1148
					int optname, sockptr_t optval,
1149
					unsigned int optlen);
1150 1151 1152
	int			(*getsockopt)(struct sock *sk, int level,
					int optname, char __user *optval,
					int __user *option);
1153
	void			(*keepalive)(struct sock *sk, int valbool);
1154
#ifdef CONFIG_COMPAT
1155 1156
	int			(*compat_ioctl)(struct sock *sk,
					unsigned int cmd, unsigned long arg);
1157
#endif
1158 1159 1160
	int			(*sendmsg)(struct sock *sk, struct msghdr *msg,
					   size_t len);
	int			(*recvmsg)(struct sock *sk, struct msghdr *msg,
1161 1162
					   size_t len, int noblock, int flags,
					   int *addr_len);
Linus Torvalds's avatar
Linus Torvalds committed
1163 1164
	int			(*sendpage)(struct sock *sk, struct page *page,
					int offset, size_t size, int flags);
1165
	int			(*bind)(struct sock *sk,
1166 1167 1168
					struct sockaddr *addr, int addr_len);
	int			(*bind_add)(struct sock *sk,
					struct sockaddr *addr, int addr_len);
Linus Torvalds's avatar
Linus Torvalds committed
1169

1170
	int			(*backlog_rcv) (struct sock *sk,
Linus Torvalds's avatar
Linus Torvalds committed
1171 1172
						struct sk_buff *skb);

Eric Dumazet's avatar
Eric Dumazet committed
1173 1174
	void		(*release_cb)(struct sock *sk);

Linus Torvalds's avatar
Linus Torvalds committed
1175
	/* Keeping track of sk's, looking them up, and port selection methods. */
1176
	int			(*hash)(struct sock *sk);
Linus Torvalds's avatar
Linus Torvalds committed
1177
	void			(*unhash)(struct sock *sk);
Eric Dumazet's avatar
Eric Dumazet committed
1178
	void			(*rehash)(struct sock *sk);
Linus Torvalds's avatar
Linus Torvalds committed
1179 1180
	int			(*get_port)(struct sock *sk, unsigned short snum);

1181
	/* Keeping track of sockets in use */
1182
#ifdef CONFIG_PROC_FS
1183
	unsigned int		inuse_idx;
1184
#endif
1185

1186
	bool			(*stream_memory_free)(const struct sock *sk, int wake);
1187
	bool			(*stream_memory_read)(const struct sock *sk);
Linus Torvalds's avatar
Linus Torvalds committed
1188
	/* Memory pressure */
1189
	void			(*enter_memory_pressure)(struct sock *sk);
1190
	void			(*leave_memory_pressure)(struct sock *sk);
Eric Dumazet's avatar
Eric Dumazet committed
1191
	atomic_long_t		*memory_allocated;	/* Current allocated memory. */
1192
	struct percpu_counter	*sockets_allocated;	/* Current number of sockets. */
Linus Torvalds's avatar
Linus Torvalds committed
1193 1194 1195
	/*
	 * Pressure flag: try to collapse.
	 * Technical note: it is used by multiple contexts non atomically.
1196
	 * All the __sk_mem_schedule() is of this nature: accounting
Linus Torvalds's avatar
Linus Torvalds committed
1197 1198
	 * is strict, actions are advisory and have some latency.
	 */
1199
	unsigned long		*memory_pressure;
Eric Dumazet's avatar
Eric Dumazet committed
1200
	long			*sysctl_mem;
1201

Linus Torvalds's avatar
Linus Torvalds committed
1202 1203
	int			*sysctl_wmem;
	int			*sysctl_rmem;
1204 1205 1206
	u32			sysctl_wmem_offset;
	u32			sysctl_rmem_offset;

Linus Torvalds's avatar
Linus Torvalds committed
1207
	int			max_header;
1208
	bool			no_autobind;
Linus Torvalds's avatar
Linus Torvalds committed
1209

1210
	struct kmem_cache	*slab;
Linus Torvalds's avatar
Linus Torvalds committed
1211
	unsigned int		obj_size;
1212
	slab_flags_t		slab_flags;
1213 1214
	unsigned int		useroffset;	/* Usercopy region offset */
	unsigned int		usersize;	/* Usercopy region size */
Linus Torvalds's avatar
Linus Torvalds committed
1215

1216
	struct percpu_counter	*orphan_count;
1217

1218
	struct request_sock_ops	*rsk_prot;
1219
	struct timewait_sock_ops *twsk_prot;
1220

1221 1222
	union {
		struct inet_hashinfo	*hashinfo;
1223
		struct udp_table	*udp_table;
1224
		struct raw_hashinfo	*raw_hash;
1225
		struct smc_hashinfo	*smc_hash;
1226
	} h;
1227

Linus Torvalds's avatar
Linus Torvalds committed
1228 1229 1230 1231 1232
	struct module		*owner;

	char			name[32];

	struct list_head	node;
1233 1234
#ifdef SOCK_REFCNT_DEBUG
	atomic_t		socks;
Glauber Costa's avatar
Glauber Costa committed
1235
#endif
1236
	int			(*diag_destroy)(struct sock *sk, int err);
1237
} __randomize_layout;
Glauber Costa's avatar
Glauber Costa committed
1238

1239 1240
int proto_register(struct proto *prot, int alloc_slab);
void proto_unregister(struct proto *prot);
1241
int sock_load_diag_module(int family, int protocol);
Linus Torvalds's avatar
Linus Torvalds committed
1242

1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255
#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));
}

1256
static inline void sk_refcnt_debug_release(const struct sock *sk)
1257
{
1258
	if (refcount_read(&sk->sk_refcnt) != 1)
1259
		printk(KERN_DEBUG "Destruction of the %s socket %p delayed, refcnt=%d\n",
1260
		       sk->sk_prot->name, sk, refcount_read(&sk->sk_refcnt));
1261 1262 1263 1264 1265 1266 1267
}
#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 */

1268
static inline bool __sk_stream_memory_free(const struct sock *sk, int wake)
1269
{
1270
	if (READ_ONCE(sk->sk_wmem_queued) >= READ_ONCE(sk->sk_sndbuf))
1271 1272 1273
		return false;

	return sk->sk_prot->stream_memory_free ?
1274
		sk->sk_prot->stream_memory_free(sk, wake) : true;
1275 1276
}

1277 1278 1279 1280 1281 1282
static inline bool sk_stream_memory_free(const struct sock *sk)
{
	return __sk_stream_memory_free(sk, 0);
}

static inline bool __sk_stream_is_writeable(const struct sock *sk, int wake)
1283
{
1284
	return sk_stream_wspace(sk) >= sk_stream_min_wspace(sk) &&
1285 1286 1287 1288 1289 1290
	       __sk_stream_memory_free(sk, wake);
}

static inline bool sk_stream_is_writeable(const struct sock *sk)
{
	return __sk_stream_is_writeable(sk, 0);
1291
}
Glauber Costa's avatar
Glauber Costa committed
1292

1293 1294 1295 1296 1297 1298 1299 1300 1301 1302
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
}
1303

1304 1305 1306 1307 1308 1309 1310 1311 1312
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;
Glauber Costa's avatar
Glauber Costa committed
1313

1314 1315
	if (mem_cgroup_sockets_enabled && sk->sk_memcg &&
	    mem_cgroup_under_socket_pressure(sk->sk_memcg))
1316
		return true;
Glauber Costa's avatar
Glauber Costa committed
1317

1318
	return !!*sk->sk_prot->memory_pressure;
1319 1320 1321 1322 1323
}

static inline long
sk_memory_allocated(const struct sock *sk)
{
1324
	return atomic_long_read(sk->sk_prot->memory_allocated);
1325 1326 1327
}

static inline long
1328
sk_memory_allocated_add(struct sock *sk, int amt)
1329
{
1330
	return atomic_long_add_return(amt, sk->sk_prot->memory_allocated);
1331 1332 1333
}

static inline void
1334
sk_memory_allocated_sub(struct sock *sk, int amt)
1335
{
1336
	atomic_long_sub(amt, sk->sk_prot->memory_allocated);
1337 1338 1339 1340
}

static inline void sk_sockets_allocated_dec(struct sock *sk)
{
1341
	percpu_counter_dec(sk->sk_prot->sockets_allocated);
1342 1343 1344 1345
}

static inline void sk_sockets_allocated_inc(struct sock *sk)
{
1346
	percpu_counter_inc(sk->sk_prot->sockets_allocated);
1347 1348
}

1349
static inline u64
1350 1351
sk_sockets_allocated_read_positive(struct sock *sk)
{
1352
	return percpu_counter_read_positive(sk->sk_prot->sockets_allocated);
1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374
}

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;
}

1375 1376

#ifdef CONFIG_PROC_FS
Linus Torvalds's avatar
Linus Torvalds committed
1377
/* Called with local bh disabled */
1378 1379
void sock_prot_inuse_add(struct net *net, struct proto *prot, int inc);
int sock_prot_inuse_get(struct net *net, struct proto *proto);
1380
int sock_inuse_get(struct net *net);
1381
#else
1382
static inline void sock_prot_inuse_add(struct net *net, struct proto *prot,
1383
		int inc)
1384 1385 1386 1387
{
}
#endif

Linus Torvalds's avatar
Linus Torvalds committed
1388

1389 1390 1391
/* With per-bucket locks this operation is not-atomic, so that
 * this version is not worse.
 */
1392
static inline int __sk_prot_rehash(struct sock *sk)
1393 1394
{
	sk->sk_prot->unhash(sk);
1395
	return sk->sk_prot->hash(sk);
1396 1397
}

Linus Torvalds's avatar
Linus Torvalds committed
1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427
/* 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;
}

1428 1429 1430
/*
 * Functions for memory accounting
 */
1431
int __sk_mem_raise_allocated(struct sock *sk, int size, int amt, int kind);
1432
int __sk_mem_schedule(struct sock *sk, int size, int kind);
1433
void __sk_mem_reduce_allocated(struct sock *sk, int amount);
1434
void __sk_mem_reclaim(struct sock *sk, int amount);
Linus Torvalds's avatar
Linus Torvalds committed
1435

1436 1437 1438 1439
/* 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
1440 1441 1442
#define SK_MEM_QUANTUM_SHIFT ilog2(SK_MEM_QUANTUM)
#define SK_MEM_SEND	0
#define SK_MEM_RECV	1
Linus Torvalds's avatar
Linus Torvalds committed
1443

1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456
/* 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;
}

1457
static inline int sk_mem_pages(int amt)
Linus Torvalds's avatar
Linus Torvalds committed
1458
{
1459
	return (amt + SK_MEM_QUANTUM - 1) >> SK_MEM_QUANTUM_SHIFT;
Linus Torvalds's avatar
Linus Torvalds committed
1460 1461
}

1462
static inline bool sk_has_account(struct sock *sk)
Linus Torvalds's avatar
Linus Torvalds committed
1463
{
1464 1465
	/* return true if protocol supports memory accounting */
	return !!sk->sk_prot->memory_allocated;
Linus Torvalds's avatar
Linus Torvalds committed
1466 1467
}

1468
static inline bool sk_wmem_schedule(struct sock *sk, int size)
Linus Torvalds's avatar
Linus Torvalds committed
1469
{
1470
	if (!sk_has_account(sk))
1471
		return true;
1472 1473
	return size <= sk->sk_forward_alloc ||
		__sk_mem_schedule(sk, size, SK_MEM_SEND);
Linus Torvalds's avatar
Linus Torvalds committed
1474 1475
}

1476
static inline bool
1477
sk_rmem_schedule(struct sock *sk, struct sk_buff *skb, int size)
1478
{
1479
	if (!sk_has_account(sk))
1480
		return true;
1481 1482 1483
	return size<= sk->sk_forward_alloc ||
		__sk_mem_schedule(sk, size, SK_MEM_RECV) ||
		skb_pfmemalloc(skb);
1484 1485 1486 1487 1488 1489 1490
}

static inline void sk_mem_reclaim(struct sock *sk)
{
	if (!sk_has_account(sk))
		return;
	if (sk->sk_forward_alloc >= SK_MEM_QUANTUM)
1491
		__sk_mem_reclaim(sk, sk->sk_forward_alloc);
1492 1493
}

1494 1495 1496 1497 1498
static inline void sk_mem_reclaim_partial(struct sock *sk)
{
	if (!sk_has_account(sk))
		return;
	if (sk->sk_forward_alloc > SK_MEM_QUANTUM)
1499
		__sk_mem_reclaim(sk, sk->sk_forward_alloc - 1);
1500 1501
}

1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513
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;
1514 1515 1516 1517 1518 1519 1520 1521 1522 1523

	/* 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);
1524 1525
}

1526
DECLARE_STATIC_KEY_FALSE(tcp_tx_skb_cache_key);
1527 1528
static inline void sk_wmem_free_skb(struct sock *sk, struct sk_buff *skb)
{
1529
	sock_set_flag(sk, SOCK_QUEUE_SHRUNK);
1530
	sk_wmem_queued_add(sk, -skb->truesize);
1531
	sk_mem_uncharge(sk, skb->truesize);
1532 1533
	if (static_branch_unlikely(&tcp_tx_skb_cache_key) &&
	    !sk->sk_tx_skb_cache && !skb_cloned(skb)) {
Paolo Abeni's avatar
Paolo Abeni committed
1534
		skb_ext_reset(skb);
1535
		skb_zcopy_clear(skb, true);
1536 1537 1538
		sk->sk_tx_skb_cache = skb;
		return;
	}
1539
	__kfree_skb(skb);
1540 1541
}

1542 1543
static inline void sock_release_ownership(struct sock *sk)
{
1544 1545 1546 1547
	if (sk->sk_lock.owned) {
		sk->sk_lock.owned = 0;

		/* The sk_lock has mutex_unlock() semantics: */
1548
		mutex_release(&sk->sk_lock.dep_map, _RET_IP_);
1549
	}
1550 1551
}

1552 1553 1554 1555 1556 1557 1558
/*
 * 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.
 */
1559
#define sock_lock_init_class_and_name(sk, sname, skey, name, key)	\
1560
do {									\
1561
	sk->sk_lock.owned = 0;						\
1562 1563 1564 1565 1566
	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,		\
1567
				(skey), (sname));				\
1568 1569 1570
	lockdep_init_map(&(sk)->sk_lock.dep_map, (name), (key), 0);	\
} while (0)

1571
#ifdef CONFIG_LOCKDEP
1572
static inline bool lockdep_sock_is_held(const struct sock *sk)
1573 1574 1575 1576
{
	return lockdep_is_held(&sk->sk_lock) ||
	       lockdep_is_held(&sk->sk_lock.slock);
}
1577
#endif
1578

1579
void lock_sock_nested(struct sock *sk, int subclass);
1580 1581 1582 1583 1584 1585

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

1586
void __release_sock(struct sock *sk);
1587
void release_sock(struct sock *sk);
Linus Torvalds's avatar
Linus Torvalds committed
1588 1589 1590

/* BH context may only use the following locking interface. */
#define bh_lock_sock(__sk)	spin_lock(&((__sk)->sk_lock.slock))
1591 1592 1593
#define bh_lock_sock_nested(__sk) \
				spin_lock_nested(&((__sk)->sk_lock.slock), \
				SINGLE_DEPTH_NESTING)
Linus Torvalds's avatar
Linus Torvalds committed
1594 1595
#define bh_unlock_sock(__sk)	spin_unlock(&((__sk)->sk_lock.slock))

1596
bool lock_sock_fast(struct sock *sk);
1597 1598 1599 1600 1601 1602 1603 1604 1605
/**
 * 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)
Eric Dumazet's avatar
Eric Dumazet committed
1606
{
1607 1608 1609 1610
	if (slow)
		release_sock(sk);
	else
		spin_unlock_bh(&sk->sk_lock.slock);
Eric Dumazet's avatar
Eric Dumazet committed
1611 1612
}

1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626
/* 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.
 */

1627
static inline void sock_owned_by_me(const struct sock *sk)
1628 1629
{
#ifdef CONFIG_LOCKDEP
1630
	WARN_ON_ONCE(!lockdep_sock_is_held(sk) && debug_locks);
1631
#endif
1632 1633 1634 1635 1636
}

static inline bool sock_owned_by_user(const struct sock *sk)
{
	sock_owned_by_me(sk);
1637 1638 1639
	return sk->sk_lock.owned;
}

1640 1641 1642 1643 1644
static inline bool sock_owned_by_user_nocheck(const struct sock *sk)
{
	return sk->sk_lock.owned;
}

1645 1646 1647 1648 1649 1650 1651
/* 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);
}
Eric Dumazet's avatar
Eric Dumazet committed
1652

1653
struct sock *sk_alloc(struct net *net, int family, gfp_t priority,
1654
		      struct proto *prot, int kern);
1655
void sk_free(struct sock *sk);
1656
void sk_destruct(struct sock *sk);
1657
struct sock *sk_clone_lock(const struct sock *sk, const gfp_t priority);
1658
void sk_free_unlock_clone(struct sock *sk);
1659 1660 1661

struct sk_buff *sock_wmalloc(struct sock *sk, unsigned long size, int force,
			     gfp_t priority);
1662
void __sock_wfree(struct sk_buff *skb);
1663
void sock_wfree(struct sk_buff *skb);
1664 1665
struct sk_buff *sock_omalloc(struct sock *sk, unsigned long size,
			     gfp_t priority);
1666 1667
void skb_orphan_partial(struct sk_buff *skb);
void sock_rfree(struct sk_buff *skb);
1668
void sock_efree(struct sk_buff *skb);
1669
#ifdef CONFIG_INET
1670
void sock_edemux(struct sk_buff *skb);
1671
void sock_pfree(struct sk_buff *skb);
1672
#else
1673
#define sock_edemux sock_efree
1674
#endif
1675 1676

int sock_setsockopt(struct socket *sock, int level, int op,
1677
		    sockptr_t optval, unsigned int optlen);
1678 1679 1680

int sock_getsockopt(struct socket *sock, int level, int op,
		    char __user *optval, int __user *optlen);
1681 1682
int sock_gettstamp(struct socket *sock, void __user *userstamp,
		   bool timeval, bool time32);
1683 1684 1685 1686 1687 1688 1689
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);
1690
void sock_kzfree_s(struct sock *sk, void *mem, int size);
1691
void sk_send_sigurg(struct sock *sk);
Linus Torvalds's avatar
Linus Torvalds committed
1692

1693
struct sockcm_cookie {
1694
	u64 transmit_time;
1695
	u32 mark;
1696
	u16 tsflags;
1697 1698
};

1699 1700 1701 1702 1703 1704
static inline void sockcm_init(struct sockcm_cookie *sockc,
			       const struct sock *sk)
{
	*sockc = (struct sockcm_cookie) { .tsflags = sk->sk_tsflags };
}

1705 1706
int __sock_cmsg_send(struct sock *sk, struct msghdr *msg, struct cmsghdr *cmsg,
		     struct sockcm_cookie *sockc);
1707 1708 1709
int sock_cmsg_send(struct sock *sk, struct msghdr *msg,
		   struct sockcm_cookie *sockc);

Linus Torvalds's avatar
Linus Torvalds committed
1710 1711 1712 1713
/*
 * Functions to fill in entries in struct proto_ops when a protocol
 * does not implement a particular function.
 */
1714 1715 1716
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 *);
1717
int sock_no_accept(struct socket *, struct socket *, int, bool);
1718
int sock_no_getname(struct socket *, struct sockaddr *, int);
1719 1720 1721
int sock_no_ioctl(struct socket *, unsigned int, unsigned long);
int sock_no_listen(struct socket *, int);
int sock_no_shutdown(struct socket *, int);
1722
int sock_no_sendmsg(struct socket *, struct msghdr *, size_t);
1723
int sock_no_sendmsg_locked(struct sock *sk, struct msghdr *msg, size_t len);
1724
int sock_no_recvmsg(struct socket *, struct msghdr *, size_t, int);
1725 1726 1727 1728
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);
1729 1730
ssize_t sock_no_sendpage_locked(struct sock *sk, struct page *page,
				int offset, size_t size, int flags);
Linus Torvalds's avatar
Linus Torvalds committed
1731 1732 1733 1734 1735

/*
 * Functions to fill in entries in struct proto_ops when a protocol
 * uses the inet style.
 */
1736
int sock_common_getsockopt(struct socket *sock, int level, int optname,
Linus Torvalds's avatar
Linus Torvalds committed
1737
				  char __user *optval, int __user *optlen);
1738 1739
int sock_common_recvmsg(struct socket *sock, struct msghdr *msg, size_t size,
			int flags);
1740
int sock_common_setsockopt(struct socket *sock, int level, int optname,
1741
			   sockptr_t optval, unsigned int optlen);
Linus Torvalds's avatar
Linus Torvalds committed
1742

1743
void sk_common_release(struct sock *sk);
Linus Torvalds's avatar
Linus Torvalds committed
1744 1745 1746 1747

/*
 *	Default socket callbacks and setup code
 */
1748

Linus Torvalds's avatar
Linus Torvalds committed
1749
/* Initialise core socket variables */
1750
void sock_init_data(struct socket *sock, struct sock *sk);
Linus Torvalds's avatar
Linus Torvalds committed
1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779

/*
 * 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)
{
1780
	if (refcount_dec_and_test(&sk->sk_refcnt))
Linus Torvalds's avatar
Linus Torvalds committed
1781 1782
		sk_free(sk);
}
Eric Dumazet's avatar
Eric Dumazet committed
1783
/* Generic version of sock_put(), dealing with all sockets
1784
 * (TCP_TIMEWAIT, TCP_NEW_SYN_RECV, ESTABLISHED...)
Eric Dumazet's avatar
Eric Dumazet committed
1785 1786
 */
void sock_gen_put(struct sock *sk);
Linus Torvalds's avatar
Linus Torvalds committed
1787

1788
int __sk_receive_skb(struct sock *sk, struct sk_buff *skb, const int nested,
1789
		     unsigned int trim_cap, bool refcounted);
1790 1791 1792
static inline int sk_receive_skb(struct sock *sk, struct sk_buff *skb,
				 const int nested)
{
1793
	return __sk_receive_skb(sk, skb, nested, 1, true);
1794
}
1795

1796 1797
static inline void sk_tx_queue_set(struct sock *sk, int tx_queue)
{
1798 1799 1800
	/* sk_tx_queue_mapping accept only upto a 16-bit value */
	if (WARN_ON_ONCE((unsigned short)tx_queue >= USHRT_MAX))
		return;
1801 1802 1803
	sk->sk_tx_queue_mapping = tx_queue;
}

1804 1805
#define NO_QUEUE_MAPPING	USHRT_MAX

1806 1807
static inline void sk_tx_queue_clear(struct sock *sk)
{
1808
	sk->sk_tx_queue_mapping = NO_QUEUE_MAPPING;
1809 1810 1811 1812
}

static inline int sk_tx_queue_get(const struct sock *sk)
{
1813 1814 1815 1816
	if (sk && sk->sk_tx_queue_mapping != NO_QUEUE_MAPPING)
		return sk->sk_tx_queue_mapping;

	return -1;
1817 1818
}

1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839
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
}

1840 1841 1842 1843 1844 1845 1846 1847 1848 1849
#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

1850 1851 1852 1853 1854
static inline void sk_set_socket(struct sock *sk, struct socket *sock)
{
	sk->sk_socket = sock;
}

Eric Dumazet's avatar
Eric Dumazet committed
1855 1856
static inline wait_queue_head_t *sk_sleep(struct sock *sk)
{
1857 1858
	BUILD_BUG_ON(offsetof(struct socket_wq, wait) != 0);
	return &rcu_dereference_raw(sk->sk_wq)->wait;
Eric Dumazet's avatar
Eric Dumazet committed
1859
}
Linus Torvalds's avatar
Linus Torvalds committed
1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870
/* 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);
1871
	sk_set_socket(sk, NULL);
1872
	sk->sk_wq  = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
1873 1874 1875 1876 1877
	write_unlock_bh(&sk->sk_callback_lock);
}

static inline void sock_graft(struct sock *sk, struct socket *parent)
{
1878
	WARN_ON(parent->sk);
Linus Torvalds's avatar
Linus Torvalds committed
1879
	write_lock_bh(&sk->sk_callback_lock);
1880
	rcu_assign_pointer(sk->sk_wq, &parent->wq);
Linus Torvalds's avatar
Linus Torvalds committed
1881
	parent->sk = sk;
1882
	sk_set_socket(sk, parent);
1883
	sk->sk_uid = SOCK_INODE(parent)->i_uid;
1884
	security_sock_graft(sk, parent);
Linus Torvalds's avatar
Linus Torvalds committed
1885 1886 1887
	write_unlock_bh(&sk->sk_callback_lock);
}

1888 1889
kuid_t sock_i_uid(struct sock *sk);
unsigned long sock_i_ino(struct sock *sk);
Linus Torvalds's avatar
Linus Torvalds committed
1890

1891 1892 1893 1894 1895
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);
}

1896
static inline u32 net_tx_rndhash(void)
1897
{
1898 1899 1900 1901
	u32 v = prandom_u32();

	return v ?: 1;
}
1902

1903 1904 1905
static inline void sk_set_txhash(struct sock *sk)
{
	sk->sk_txhash = net_tx_rndhash();
1906 1907
}

1908 1909 1910 1911 1912 1913
static inline void sk_rethink_txhash(struct sock *sk)
{
	if (sk->sk_txhash)
		sk_set_txhash(sk);
}

Linus Torvalds's avatar
Linus Torvalds committed
1914 1915 1916
static inline struct dst_entry *
__sk_dst_get(struct sock *sk)
{
1917 1918
	return rcu_dereference_check(sk->sk_dst_cache,
				     lockdep_sock_is_held(sk));
Linus Torvalds's avatar
Linus Torvalds committed
1919 1920 1921 1922 1923 1924 1925
}

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

1926 1927
	rcu_read_lock();
	dst = rcu_dereference(sk->sk_dst_cache);
1928 1929
	if (dst && !atomic_inc_not_zero(&dst->__refcnt))
		dst = NULL;
1930
	rcu_read_unlock();
Linus Torvalds's avatar
Linus Torvalds committed
1931 1932 1933
	return dst;
}

1934 1935 1936 1937
static inline void dst_negative_advice(struct sock *sk)
{
	struct dst_entry *ndst, *dst = __sk_dst_get(sk);

1938 1939
	sk_rethink_txhash(sk);

1940 1941 1942 1943 1944
	if (dst && dst->ops->negative_advice) {
		ndst = dst->ops->negative_advice(dst);

		if (ndst != dst) {
			rcu_assign_pointer(sk->sk_dst_cache, ndst);
1945
			sk_tx_queue_clear(sk);
1946
			sk->sk_dst_pending_confirm = 0;
1947 1948 1949 1950
		}
	}
}

Linus Torvalds's avatar
Linus Torvalds committed
1951 1952 1953 1954 1955
static inline void
__sk_dst_set(struct sock *sk, struct dst_entry *dst)
{
	struct dst_entry *old_dst;

1956
	sk_tx_queue_clear(sk);
1957
	sk->sk_dst_pending_confirm = 0;
1958 1959
	old_dst = rcu_dereference_protected(sk->sk_dst_cache,
					    lockdep_sock_is_held(sk));
1960
	rcu_assign_pointer(sk->sk_dst_cache, dst);
Linus Torvalds's avatar
Linus Torvalds committed
1961 1962 1963 1964 1965 1966
	dst_release(old_dst);
}

static inline void
sk_dst_set(struct sock *sk, struct dst_entry *dst)
{
1967 1968 1969
	struct dst_entry *old_dst;

	sk_tx_queue_clear(sk);
1970
	sk->sk_dst_pending_confirm = 0;
1971
	old_dst = xchg((__force struct dst_entry **)&sk->sk_dst_cache, dst);
1972
	dst_release(old_dst);
Linus Torvalds's avatar
Linus Torvalds committed
1973 1974 1975 1976 1977
}

static inline void
__sk_dst_reset(struct sock *sk)
{
1978
	__sk_dst_set(sk, NULL);
Linus Torvalds's avatar
Linus Torvalds committed
1979 1980 1981 1982 1983
}

static inline void
sk_dst_reset(struct sock *sk)
{
1984
	sk_dst_set(sk, NULL);
Linus Torvalds's avatar
Linus Torvalds committed
1985 1986
}

1987
struct dst_entry *__sk_dst_check(struct sock *sk, u32 cookie);
Linus Torvalds's avatar
Linus Torvalds committed
1988

1989
struct dst_entry *sk_dst_check(struct sock *sk, u32 cookie);
Linus Torvalds's avatar
Linus Torvalds committed
1990

1991 1992
static inline void sk_dst_confirm(struct sock *sk)
{
1993 1994
	if (!READ_ONCE(sk->sk_dst_pending_confirm))
		WRITE_ONCE(sk->sk_dst_pending_confirm, 1);
1995 1996
}

1997 1998 1999 2000 2001 2002 2003
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 */
2004 2005 2006 2007
		if (READ_ONCE(n->confirmed) != now)
			WRITE_ONCE(n->confirmed, now);
		if (sk && READ_ONCE(sk->sk_dst_pending_confirm))
			WRITE_ONCE(sk->sk_dst_pending_confirm, 0);
2008 2009 2010
	}
}

2011 2012
bool sk_mc_loop(struct sock *sk);

2013
static inline bool sk_can_gso(const struct sock *sk)
2014 2015 2016 2017
{
	return net_gso_ok(sk->sk_route_caps, sk->sk_gso_type);
}

2018
void sk_setup_caps(struct sock *sk, struct dst_entry *dst);
2019

2020
static inline void sk_nocaps_add(struct sock *sk, netdev_features_t flags)
2021 2022 2023 2024 2025
{
	sk->sk_route_nocaps |= flags;
	sk->sk_route_caps &= ~flags;
}

2026
static inline int skb_do_copy_data_nocache(struct sock *sk, struct sk_buff *skb,
2027
					   struct iov_iter *from, char *to,
2028
					   int copy, int offset)
2029 2030
{
	if (skb->ip_summed == CHECKSUM_NONE) {
2031
		__wsum csum = 0;
2032
		if (!csum_and_copy_from_iter_full(to, copy, &csum, from))
2033
			return -EFAULT;
2034
		skb->csum = csum_block_add(skb->csum, csum, offset);
2035
	} else if (sk->sk_route_caps & NETIF_F_NOCACHE_COPY) {
2036
		if (!copy_from_iter_full_nocache(to, copy, from))
2037
			return -EFAULT;
2038
	} else if (!copy_from_iter_full(to, copy, from))
2039 2040 2041 2042 2043 2044
		return -EFAULT;

	return 0;
}

static inline int skb_add_data_nocache(struct sock *sk, struct sk_buff *skb,
2045
				       struct iov_iter *from, int copy)
2046
{
2047
	int err, offset = skb->len;
2048

2049 2050
	err = skb_do_copy_data_nocache(sk, skb, from, skb_put(skb, copy),
				       copy, offset);
2051
	if (err)
2052
		__skb_trim(skb, offset);
2053 2054 2055 2056

	return err;
}

2057
static inline int skb_copy_to_page_nocache(struct sock *sk, struct iov_iter *from,
2058 2059 2060 2061 2062 2063
					   struct sk_buff *skb,
					   struct page *page,
					   int off, int copy)
{
	int err;

2064 2065
	err = skb_do_copy_data_nocache(sk, skb, from, page_address(page) + off,
				       copy, skb->len);
2066 2067 2068 2069 2070 2071
	if (err)
		return err;

	skb->len	     += copy;
	skb->data_len	     += copy;
	skb->truesize	     += copy;
2072
	sk_wmem_queued_add(sk, copy);
2073 2074 2075 2076
	sk_mem_charge(sk, copy);
	return 0;
}

2077 2078 2079 2080
/**
 * sk_wmem_alloc_get - returns write allocations
 * @sk: socket
 *
2081
 * Return: sk_wmem_alloc minus initial offset of one
2082 2083 2084
 */
static inline int sk_wmem_alloc_get(const struct sock *sk)
{
2085
	return refcount_read(&sk->sk_wmem_alloc) - 1;
2086 2087 2088 2089 2090 2091
}

/**
 * sk_rmem_alloc_get - returns read allocations
 * @sk: socket
 *
2092
 * Return: sk_rmem_alloc
2093 2094 2095 2096 2097 2098 2099 2100 2101 2102
 */
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
 *
2103
 * Return: true if socket has write or read allocations
2104
 */
2105
static inline bool sk_has_allocations(const struct sock *sk)
2106 2107 2108 2109
{
	return sk_wmem_alloc_get(sk) || sk_rmem_alloc_get(sk);
}

2110
/**
2111
 * skwq_has_sleeper - check if there are any waiting processes
2112
 * @wq: struct socket_wq
2113
 *
2114
 * Return: true if socket_wq has waiting processes
2115
 *
2116
 * The purpose of the skwq_has_sleeper and sock_poll_wait is to wrap the memory
2117 2118
 * barrier call. They were added due to the race found within the tcp code.
 *
2119
 * Consider following tcp code paths::
2120
 *
2121 2122
 *   CPU1                CPU2
 *   sys_select          receive packet
2123 2124 2125 2126 2127
 *   ...                 ...
 *   __add_wait_queue    update tp->rcv_nxt
 *   ...                 ...
 *   tp->rcv_nxt check   sock_def_readable
 *   ...                 {
2128 2129 2130 2131
 *   schedule               rcu_read_lock();
 *                          wq = rcu_dereference(sk->sk_wq);
 *                          if (wq && waitqueue_active(&wq->wait))
 *                              wake_up_interruptible(&wq->wait)
2132 2133 2134 2135 2136 2137 2138
 *                          ...
 *                       }
 *
 * 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.
2139
 *
2140
 */
2141
static inline bool skwq_has_sleeper(struct socket_wq *wq)
2142
{
2143
	return wq && wq_has_sleeper(&wq->wait);
2144 2145 2146 2147 2148
}

/**
 * sock_poll_wait - place memory barrier behind the poll_wait call.
 * @filp:           file
2149
 * @sock:           socket to wait on
2150 2151
 * @p:              poll_table
 *
2152
 * See the comments in the wq_has_sleeper function.
2153
 */
2154 2155
static inline void sock_poll_wait(struct file *filp, struct socket *sock,
				  poll_table *p)
2156
{
2157
	if (!poll_does_not_wait(p)) {
2158
		poll_wait(filp, &sock->wq.wait, p);
2159
		/* We need to be sure we are in sync with the
2160 2161
		 * socket flags modification.
		 *
2162
		 * This memory barrier is paired in the wq_has_sleeper.
2163
		 */
2164 2165 2166 2167
		smp_mb();
	}
}

2168 2169 2170 2171 2172 2173 2174 2175
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;
	}
}

2176 2177
void skb_set_owner_w(struct sk_buff *skb, struct sock *sk);

Linus Torvalds's avatar
Linus Torvalds committed
2178
/*
2179
 *	Queue a received datagram if it will fit. Stream and sequenced
Linus Torvalds's avatar
Linus Torvalds committed
2180 2181 2182
 *	protocols can't normally use this as they need to fit buffers in
 *	and play with them.
 *
2183
 *	Inlined as it's very short and called for pretty much every
Linus Torvalds's avatar
Linus Torvalds committed
2184 2185 2186 2187
 *	packet ever received.
 */
static inline void skb_set_owner_r(struct sk_buff *skb, struct sock *sk)
{
2188
	skb_orphan(skb);
Linus Torvalds's avatar
Linus Torvalds committed
2189 2190 2191
	skb->sk = sk;
	skb->destructor = sock_rfree;
	atomic_add(skb->truesize, &sk->sk_rmem_alloc);
2192
	sk_mem_charge(sk, skb->truesize);
Linus Torvalds's avatar
Linus Torvalds committed
2193 2194
}

2195 2196
void sk_reset_timer(struct sock *sk, struct timer_list *timer,
		    unsigned long expires);
Linus Torvalds's avatar
Linus Torvalds committed
2197

2198
void sk_stop_timer(struct sock *sk, struct timer_list *timer);
Linus Torvalds's avatar
Linus Torvalds committed
2199

2200 2201
int __sk_queue_drop_skb(struct sock *sk, struct sk_buff_head *sk_queue,
			struct sk_buff *skb, unsigned int flags,
2202 2203
			void (*destructor)(struct sock *sk,
					   struct sk_buff *skb));
2204
int __sock_queue_rcv_skb(struct sock *sk, struct sk_buff *skb);
2205
int sock_queue_rcv_skb(struct sock *sk, struct sk_buff *skb);
Linus Torvalds's avatar
Linus Torvalds committed
2206

2207
int sock_queue_err_skb(struct sock *sk, struct sk_buff *skb);
2208
struct sk_buff *sock_dequeue_err_skb(struct sock *sk);
Linus Torvalds's avatar
Linus Torvalds committed
2209 2210 2211 2212

/*
 *	Recover an error report and clear atomically
 */
2213

Linus Torvalds's avatar
Linus Torvalds committed
2214 2215
static inline int sock_error(struct sock *sk)
{
2216 2217 2218 2219
	int err;
	if (likely(!sk->sk_err))
		return 0;
	err = xchg(&sk->sk_err, 0);
Linus Torvalds's avatar
Linus Torvalds committed
2220 2221 2222 2223 2224 2225 2226 2227
	return -err;
}

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

	if (!(sk->sk_shutdown & SEND_SHUTDOWN)) {
2228
		amt = sk->sk_sndbuf - refcount_read(&sk->sk_wmem_alloc);
2229
		if (amt < 0)
Linus Torvalds's avatar
Linus Torvalds committed
2230 2231 2232 2233 2234
			amt = 0;
	}
	return amt;
}

2235 2236 2237 2238
/* Note:
 *  We use sk->sk_wq_raw, from contexts knowing this
 *  pointer is not NULL and cannot disappear/change.
 */
2239
static inline void sk_set_bit(int nr, struct sock *sk)
Linus Torvalds's avatar
Linus Torvalds committed
2240
{
2241 2242
	if ((nr == SOCKWQ_ASYNC_NOSPACE || nr == SOCKWQ_ASYNC_WAITDATA) &&
	    !sock_flag(sk, SOCK_FASYNC))
2243 2244
		return;

2245
	set_bit(nr, &sk->sk_wq_raw->flags);
2246 2247 2248 2249
}

static inline void sk_clear_bit(int nr, struct sock *sk)
{
2250 2251
	if ((nr == SOCKWQ_ASYNC_NOSPACE || nr == SOCKWQ_ASYNC_WAITDATA) &&
	    !sock_flag(sk, SOCK_FASYNC))
2252 2253
		return;

2254
	clear_bit(nr, &sk->sk_wq_raw->flags);
2255 2256
}

2257
static inline void sk_wake_async(const struct sock *sk, int how, int band)
Linus Torvalds's avatar
Linus Torvalds committed
2258
{
2259 2260 2261 2262 2263
	if (sock_flag(sk, SOCK_FASYNC)) {
		rcu_read_lock();
		sock_wake_async(rcu_dereference(sk->sk_wq), how, band);
		rcu_read_unlock();
	}
Linus Torvalds's avatar
Linus Torvalds committed
2264 2265
}

2266 2267 2268 2269
/* 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.
Eric Dumazet's avatar
Eric Dumazet committed
2270
 */
2271
#define TCP_SKB_MIN_TRUESIZE	(2048 + SKB_DATA_ALIGN(sizeof(struct sk_buff)))
2272 2273 2274

#define SOCK_MIN_SNDBUF		(TCP_SKB_MIN_TRUESIZE * 2)
#define SOCK_MIN_RCVBUF		 TCP_SKB_MIN_TRUESIZE
Linus Torvalds's avatar
Linus Torvalds committed
2275 2276 2277

static inline void sk_stream_moderate_sndbuf(struct sock *sk)
{
2278 2279 2280 2281 2282 2283 2284 2285
	u32 val;

	if (sk->sk_userlocks & SOCK_SNDBUF_LOCK)
		return;

	val = min(sk->sk_sndbuf, sk->sk_wmem_queued >> 1);

	WRITE_ONCE(sk->sk_sndbuf, max_t(u32, val, SOCK_MIN_SNDBUF));
Linus Torvalds's avatar
Linus Torvalds committed
2286 2287
}

2288 2289
struct sk_buff *sk_stream_alloc_skb(struct sock *sk, int size, gfp_t gfp,
				    bool force_schedule);
Linus Torvalds's avatar
Linus Torvalds committed
2290

2291 2292 2293 2294
/**
 * sk_page_frag - return an appropriate page_frag
 * @sk: socket
 *
2295 2296 2297 2298 2299 2300 2301
 * Use the per task page_frag instead of the per socket one for
 * optimization when we know that we're in the normal context and owns
 * everything that's associated with %current.
 *
 * gfpflags_allow_blocking() isn't enough here as direct reclaim may nest
 * inside other socket operations and end up recursing into sk_page_frag()
 * while it's already in use.
2302 2303 2304
 *
 * Return: a per task page_frag if context allows that,
 * otherwise a per socket one.
2305 2306
 */
static inline struct page_frag *sk_page_frag(struct sock *sk)
Linus Torvalds's avatar
Linus Torvalds committed
2307
{
2308
	if (gfpflags_normal_context(sk->sk_allocation))
2309
		return &current->task_frag;
Linus Torvalds's avatar
Linus Torvalds committed
2310

2311
	return &sk->sk_frag;
Linus Torvalds's avatar
Linus Torvalds committed
2312 2313
}

2314
bool sk_page_frag_refill(struct sock *sk, struct page_frag *pfrag);
2315

Linus Torvalds's avatar
Linus Torvalds committed
2316 2317 2318
/*
 *	Default write policy as shown to user space via poll/select/SIGIO
 */
2319
static inline bool sock_writeable(const struct sock *sk)
Linus Torvalds's avatar
Linus Torvalds committed
2320
{
2321
	return refcount_read(&sk->sk_wmem_alloc) < (READ_ONCE(sk->sk_sndbuf) >> 1);
Linus Torvalds's avatar
Linus Torvalds committed
2322 2323
}

2324
static inline gfp_t gfp_any(void)
Linus Torvalds's avatar
Linus Torvalds committed
2325
{
2326
	return in_softirq() ? GFP_ATOMIC : GFP_KERNEL;
Linus Torvalds's avatar
Linus Torvalds committed
2327 2328
}

2329
static inline long sock_rcvtimeo(const struct sock *sk, bool noblock)
Linus Torvalds's avatar
Linus Torvalds committed
2330 2331 2332 2333
{
	return noblock ? 0 : sk->sk_rcvtimeo;
}

2334
static inline long sock_sndtimeo(const struct sock *sk, bool noblock)
Linus Torvalds's avatar
Linus Torvalds committed
2335 2336 2337 2338 2339 2340
{
	return noblock ? 0 : sk->sk_sndtimeo;
}

static inline int sock_rcvlowat(const struct sock *sk, int waitall, int len)
{
2341 2342 2343
	int v = waitall ? len : min_t(int, READ_ONCE(sk->sk_rcvlowat), len);

	return v ?: 1;
Linus Torvalds's avatar
Linus Torvalds committed
2344 2345 2346 2347 2348 2349 2350 2351 2352 2353
}

/* 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;
}

2354 2355 2356 2357 2358 2359 2360 2361
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.
 */
2362
#define SOCK_SKB_CB_OFFSET ((sizeof_field(struct sk_buff, cb) - \
2363 2364 2365 2366 2367
			    sizeof(struct sock_skb_cb)))

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

2368
#define sock_skb_cb_check_size(size) \
2369
	BUILD_BUG_ON((size) > SOCK_SKB_CB_OFFSET)
2370

2371 2372 2373
static inline void
sock_skb_set_dropcount(const struct sock *sk, struct sk_buff *skb)
{
2374 2375
	SOCK_SKB_CB(skb)->dropcount = sock_flag(sk, SOCK_RXQ_OVFL) ?
						atomic_read(&sk->sk_drops) : 0;
2376 2377
}

2378 2379 2380 2381 2382 2383 2384
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);
}

2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397
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
2398
	return READ_ONCE(sk->sk_stamp);
2399 2400 2401 2402 2403 2404 2405 2406 2407 2408
#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
2409
	WRITE_ONCE(sk->sk_stamp, kt);
2410 2411 2412
#endif
}

2413 2414 2415 2416
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);
2417

2418
static inline void
Linus Torvalds's avatar
Linus Torvalds committed
2419 2420
sock_recv_timestamp(struct msghdr *msg, struct sock *sk, struct sk_buff *skb)
{
2421
	ktime_t kt = skb->tstamp;
2422
	struct skb_shared_hwtstamps *hwtstamps = skb_hwtstamps(skb);
2423

2424 2425
	/*
	 * generate control messages if
2426
	 * - receive time stamping in software requested
2427 2428 2429 2430
	 * - software time stamp available and wanted
	 * - hardware time stamps available and wanted
	 */
	if (sock_flag(sk, SOCK_RCVTSTAMP) ||
2431
	    (sk->sk_tsflags & SOF_TIMESTAMPING_RX_SOFTWARE) ||
2432 2433
	    (kt && sk->sk_tsflags & SOF_TIMESTAMPING_SOFTWARE) ||
	    (hwtstamps->hwtstamp &&
2434
	     (sk->sk_tsflags & SOF_TIMESTAMPING_RAW_HARDWARE)))
2435 2436
		__sock_recv_timestamp(msg, sk, skb);
	else
2437
		sock_write_timestamp(sk, kt);
2438 2439 2440

	if (sock_flag(sk, SOCK_WIFI_STATUS) && skb->wifi_acked_valid)
		__sock_recv_wifi_status(msg, sk, skb);
Linus Torvalds's avatar
Linus Torvalds committed
2441 2442
}

2443 2444
void __sock_recv_ts_and_drops(struct msghdr *msg, struct sock *sk,
			      struct sk_buff *skb);
2445

2446
#define SK_DEFAULT_STAMP (-1L * NSEC_PER_SEC)
2447 2448 2449 2450
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)			| \
2451 2452 2453
			   (1UL << SOCK_RCVTSTAMP))
#define TSFLAGS_ANY	  (SOF_TIMESTAMPING_SOFTWARE			| \
			   SOF_TIMESTAMPING_RAW_HARDWARE)
2454

2455
	if (sk->sk_flags & FLAGS_TS_OR_DROPS || sk->sk_tsflags & TSFLAGS_ANY)
2456
		__sock_recv_ts_and_drops(msg, sk, skb);
2457
	else if (unlikely(sock_flag(sk, SOCK_TIMESTAMP)))
2458
		sock_write_timestamp(sk, skb->tstamp);
2459
	else if (unlikely(sk->sk_stamp == SK_DEFAULT_STAMP))
2460
		sock_write_timestamp(sk, 0);
2461
}
2462

2463
void __sock_tx_timestamp(__u16 tsflags, __u8 *tx_flags);
2464

2465
/**
2466
 * _sock_tx_timestamp - checks whether the outgoing packet is to be time stamped
2467
 * @sk:		socket sending this packet
2468
 * @tsflags:	timestamping flags to use
2469
 * @tx_flags:	completed with instructions for time stamping
2470
 * @tskey:      filled in with next sk_tskey (not for TCP, which uses seqno)
2471
 *
2472
 * Note: callers should take care of initial ``*tx_flags`` value (usually 0)
2473
 */
2474 2475
static inline void _sock_tx_timestamp(struct sock *sk, __u16 tsflags,
				      __u8 *tx_flags, __u32 *tskey)
2476
{
2477
	if (unlikely(tsflags)) {
2478
		__sock_tx_timestamp(tsflags, tx_flags);
2479 2480 2481 2482
		if (tsflags & SOF_TIMESTAMPING_OPT_ID && tskey &&
		    tsflags & SOF_TIMESTAMPING_TX_RECORD_MASK)
			*tskey = sk->sk_tskey++;
	}
2483 2484 2485
	if (unlikely(sock_flag(sk, SOCK_WIFI_STATUS)))
		*tx_flags |= SKBTX_WIFI_STATUS;
}
2486

2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498
static inline void sock_tx_timestamp(struct sock *sk, __u16 tsflags,
				     __u8 *tx_flags)
{
	_sock_tx_timestamp(sk, tsflags, tx_flags, NULL);
}

static inline void skb_setup_tx_timestamp(struct sk_buff *skb, __u16 tsflags)
{
	_sock_tx_timestamp(skb->sk, tsflags, &skb_shinfo(skb)->tx_flags,
			   &skb_shinfo(skb)->tskey);
}

2499
DECLARE_STATIC_KEY_FALSE(tcp_rx_skb_cache_key);
Linus Torvalds's avatar
Linus Torvalds committed
2500 2501
/**
 * sk_eat_skb - Release a skb if it is no longer needed
2502 2503
 * @sk: socket to eat this skb from
 * @skb: socket buffer to eat
Linus Torvalds's avatar
Linus Torvalds committed
2504 2505 2506 2507
 *
 * This routine must be called with interrupts disabled or with the socket
 * locked so that the sk_buff queue operation is ok.
*/
Dan Williams's avatar
Dan Williams committed
2508
static inline void sk_eat_skb(struct sock *sk, struct sk_buff *skb)
Linus Torvalds's avatar
Linus Torvalds committed
2509 2510
{
	__skb_unlink(skb, &sk->sk_receive_queue);
2511
	if (static_branch_unlikely(&tcp_rx_skb_cache_key) &&
2512 2513 2514 2515 2516
	    !sk->sk_rx_skb_cache) {
		sk->sk_rx_skb_cache = skb;
		skb_orphan(skb);
		return;
	}
Linus Torvalds's avatar
Linus Torvalds committed
2517 2518 2519
	__kfree_skb(skb);
}

2520 2521 2522
static inline
struct net *sock_net(const struct sock *sk)
{
Eric Dumazet's avatar
Eric Dumazet committed
2523
	return read_pnet(&sk->sk_net);
2524 2525 2526
}

static inline
2527
void sock_net_set(struct sock *sk, struct net *net)
2528
{
Eric Dumazet's avatar
Eric Dumazet committed
2529
	write_pnet(&sk->sk_net, net);
2530 2531
}

2532 2533 2534 2535 2536 2537 2538 2539 2540 2541
static inline bool
skb_sk_is_prefetched(struct sk_buff *skb)
{
#ifdef CONFIG_INET
	return skb->destructor == sock_pfree;
#else
	return false;
#endif /* CONFIG_INET */
}

2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556
/* 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);
}

static inline bool
sk_is_refcounted(struct sock *sk)
{
	/* Only full sockets have sk->sk_flags. */
	return !sk_fullsock(sk) || !sock_flag(sk, SOCK_RCU_FREE);
}

2557
/**
2558 2559 2560
 * skb_steal_sock - steal a socket from an sk_buff
 * @skb: sk_buff to steal the socket from
 * @refcounted: is set to true if the socket is reference-counted
2561 2562 2563
 */
static inline struct sock *
skb_steal_sock(struct sk_buff *skb, bool *refcounted)
2564
{
2565
	if (skb->sk) {
2566 2567
		struct sock *sk = skb->sk;

2568
		*refcounted = true;
2569 2570
		if (skb_sk_is_prefetched(skb))
			*refcounted = sk_is_refcounted(sk);
2571 2572 2573 2574
		skb->destructor = NULL;
		skb->sk = NULL;
		return sk;
	}
2575
	*refcounted = false;
2576 2577 2578
	return NULL;
}

2579 2580
/* Checks if this SKB belongs to an HW offloaded socket
 * and whether any SW fallbacks are required based on dev.
2581
 * Check decrypted mark in case skb_orphan() cleared socket.
2582 2583 2584 2585 2586 2587 2588
 */
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;

2589
	if (sk && sk_fullsock(sk) && sk->sk_validate_xmit_skb) {
2590
		skb = sk->sk_validate_xmit_skb(sk, dev, skb);
2591 2592 2593 2594 2595 2596 2597
#ifdef CONFIG_TLS_DEVICE
	} else if (unlikely(skb->decrypted)) {
		pr_warn_ratelimited("unencrypted skb with no associated socket - dropping\n");
		kfree_skb(skb);
		skb = NULL;
#endif
	}
2598 2599 2600 2601 2602
#endif

	return skb;
}

2603 2604 2605 2606 2607 2608 2609 2610
/* 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);
}

2611
void sock_enable_timestamp(struct sock *sk, enum sock_flags flag);
2612 2613
int sock_recv_errqueue(struct sock *sk, struct msghdr *msg, int len, int level,
		       int type);
Linus Torvalds's avatar
Linus Torvalds committed
2614

2615 2616 2617 2618 2619
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);

2620 2621
void sk_get_meminfo(const struct sock *sk, u32 *meminfo);

2622 2623 2624 2625 2626 2627 2628 2629 2630 2631
/* 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)

Linus Torvalds's avatar
Linus Torvalds committed
2632 2633 2634
extern __u32 sysctl_wmem_max;
extern __u32 sysctl_rmem_max;

2635
extern int sysctl_tstamp_allow_data;
2636 2637
extern int sysctl_optmem_max;

2638 2639 2640
extern __u32 sysctl_wmem_default;
extern __u32 sysctl_rmem_default;

2641 2642
DECLARE_STATIC_KEY_FALSE(net_high_order_alloc_disable_key);

2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660
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;
}

2661 2662 2663 2664 2665 2666
/* 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)
{
2667
	if (!sk || !sk_fullsock(sk) || READ_ONCE(sk->sk_pacing_shift) == val)
2668
		return;
2669
	WRITE_ONCE(sk->sk_pacing_shift, val);
2670 2671
}

2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690
/* 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;
}

2691 2692
void sock_def_readable(struct sock *sk);

2693
int sock_bindtoindex(struct sock *sk, int ifindex, bool lock_sk);
2694
void sock_enable_timestamps(struct sock *sk);
2695
void sock_no_linger(struct sock *sk);
2696
void sock_set_keepalive(struct sock *sk);
2697
void sock_set_priority(struct sock *sk, u32 priority);
2698
void sock_set_rcvbuf(struct sock *sk, int val);
2699
void sock_set_mark(struct sock *sk, u32 val);
2700
void sock_set_reuseaddr(struct sock *sk);
2701
void sock_set_reuseport(struct sock *sk);
2702
void sock_set_sndtimeo(struct sock *sk, s64 secs);
2703

2704 2705
int sock_bind_add(struct sock *sk, struct sockaddr *addr, int addr_len);

Linus Torvalds's avatar
Linus Torvalds committed
2706
#endif	/* _SOCK_H */