tcp.h 58.2 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1
2
3
4
5
6
7
8
9
/*
 * INET		An implementation of the TCP/IP protocol suite for the LINUX
 *		operating system.  INET is implemented using the  BSD Socket
 *		interface as the means of communication with the user level.
 *
 *		Definitions for the TCP module.
 *
 * Version:	@(#)tcp.h	1.0.5	05/23/93
 *
10
 * Authors:	Ross Biro
Linus Torvalds's avatar
Linus Torvalds committed
11
12
13
14
15
16
17
18
19
20
21
22
23
24
 *		Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *
 *		This program is free software; you can redistribute it and/or
 *		modify it under the terms of the GNU General Public License
 *		as published by the Free Software Foundation; either version
 *		2 of the License, or (at your option) any later version.
 */
#ifndef _TCP_H
#define _TCP_H

#define FASTRETRANS_DEBUG 1

#include <linux/list.h>
#include <linux/tcp.h>
25
#include <linux/bug.h>
Linus Torvalds's avatar
Linus Torvalds committed
26
27
28
#include <linux/slab.h>
#include <linux/cache.h>
#include <linux/percpu.h>
29
#include <linux/skbuff.h>
30
#include <linux/cryptohash.h>
31
#include <linux/kref.h>
32
#include <linux/ktime.h>
33
34

#include <net/inet_connection_sock.h>
35
#include <net/inet_timewait_sock.h>
36
#include <net/inet_hashtables.h>
Linus Torvalds's avatar
Linus Torvalds committed
37
#include <net/checksum.h>
38
#include <net/request_sock.h>
Linus Torvalds's avatar
Linus Torvalds committed
39
40
41
#include <net/sock.h>
#include <net/snmp.h>
#include <net/ip.h>
42
#include <net/tcp_states.h>
43
#include <net/inet_ecn.h>
44
#include <net/dst.h>
45

Linus Torvalds's avatar
Linus Torvalds committed
46
#include <linux/seq_file.h>
47
#include <linux/memcontrol.h>
Linus Torvalds's avatar
Linus Torvalds committed
48

49
extern struct inet_hashinfo tcp_hashinfo;
Linus Torvalds's avatar
Linus Torvalds committed
50

51
extern struct percpu_counter tcp_orphan_count;
52
void tcp_time_wait(struct sock *sk, int state, int timeo);
Linus Torvalds's avatar
Linus Torvalds committed
53
54

#define MAX_TCP_HEADER	(128 + MAX_HEADER)
Adam Langley's avatar
Adam Langley committed
55
#define MAX_TCP_OPTION_SPACE 40
Linus Torvalds's avatar
Linus Torvalds committed
56

57
/*
Linus Torvalds's avatar
Linus Torvalds committed
58
 * Never offer a window over 32767 without using window scaling. Some
59
 * poor stacks do signed 16bit maths!
Linus Torvalds's avatar
Linus Torvalds committed
60
61
62
63
64
65
 */
#define MAX_TCP_WINDOW		32767U

/* Minimal accepted MSS. It is (60+60+8) - (20+20). */
#define TCP_MIN_MSS		88U

John Heffner's avatar
John Heffner committed
66
/* The least MTU to use for probing */
67
#define TCP_BASE_MSS		1024
John Heffner's avatar
John Heffner committed
68

69
70
71
/* probing interval, default to 10 minutes as per RFC4821 */
#define TCP_PROBE_INTERVAL	600

72
73
74
/* Specify interval when tcp mtu probing will stop */
#define TCP_PROBE_THRESHOLD	8

Linus Torvalds's avatar
Linus Torvalds committed
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
/* After receiving this amount of duplicate ACKs fast retransmit starts. */
#define TCP_FASTRETRANS_THRESH 3

/* Maximal number of ACKs sent quickly to accelerate slow-start. */
#define TCP_MAX_QUICKACKS	16U

/* urg_data states */
#define TCP_URG_VALID	0x0100
#define TCP_URG_NOTYET	0x0200
#define TCP_URG_READ	0x0400

#define TCP_RETR1	3	/*
				 * This is how many retries it does before it
				 * tries to figure out if the gateway is
				 * down. Minimal RFC value is 3; it corresponds
				 * to ~3sec-8min depending on RTO.
				 */

#define TCP_RETR2	15	/*
				 * This should take at least
				 * 90 minutes to time out.
				 * RFC1122 says that the limit is 100 sec.
				 * 15 is ~13-30min depending on RTO.
				 */

100
101
102
103
104
105
106
107
#define TCP_SYN_RETRIES	 6	/* This is how many retries are done
				 * when active opening a connection.
				 * RFC1122 says the minimum retry MUST
				 * be at least 180secs.  Nevertheless
				 * this value is corresponding to
				 * 63secs of retransmission with the
				 * current initial RTO.
				 */
Linus Torvalds's avatar
Linus Torvalds committed
108

109
110
111
112
113
114
#define TCP_SYNACK_RETRIES 5	/* This is how may retries are done
				 * when passive opening a connection.
				 * This is corresponding to 31secs of
				 * retransmission with the current
				 * initial RTO.
				 */
Linus Torvalds's avatar
Linus Torvalds committed
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134

#define TCP_TIMEWAIT_LEN (60*HZ) /* how long to wait to destroy TIME-WAIT
				  * state, about 60 seconds	*/
#define TCP_FIN_TIMEOUT	TCP_TIMEWAIT_LEN
                                 /* BSD style FIN_WAIT2 deadlock breaker.
				  * It used to be 3min, new value is 60sec,
				  * to combine FIN-WAIT-2 timeout with
				  * TIME-WAIT timer.
				  */

#define TCP_DELACK_MAX	((unsigned)(HZ/5))	/* maximal time to delay before sending an ACK */
#if HZ >= 100
#define TCP_DELACK_MIN	((unsigned)(HZ/25))	/* minimal time to delay before sending an ACK */
#define TCP_ATO_MIN	((unsigned)(HZ/25))
#else
#define TCP_DELACK_MIN	4U
#define TCP_ATO_MIN	4U
#endif
#define TCP_RTO_MAX	((unsigned)(120*HZ))
#define TCP_RTO_MIN	((unsigned)(HZ/5))
135
#define TCP_TIMEOUT_INIT ((unsigned)(1*HZ))	/* RFC6298 2.1 initial RTO value	*/
136
137
138
139
140
141
#define TCP_TIMEOUT_FALLBACK ((unsigned)(3*HZ))	/* RFC 1122 initial RTO value, now
						 * used as a fallback RTO for the
						 * initial data transmission if no
						 * valid RTT sample has been acquired,
						 * most likely due to retrans in 3WHS.
						 */
Linus Torvalds's avatar
Linus Torvalds committed
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171

#define TCP_RESOURCE_PROBE_INTERVAL ((unsigned)(HZ/2U)) /* Maximal interval between probes
					                 * for local resources.
					                 */

#define TCP_KEEPALIVE_TIME	(120*60*HZ)	/* two hours */
#define TCP_KEEPALIVE_PROBES	9		/* Max of 9 keepalive probes	*/
#define TCP_KEEPALIVE_INTVL	(75*HZ)

#define MAX_TCP_KEEPIDLE	32767
#define MAX_TCP_KEEPINTVL	32767
#define MAX_TCP_KEEPCNT		127
#define MAX_TCP_SYNCNT		127

#define TCP_SYNQ_INTERVAL	(HZ/5)	/* Period of SYNACK timer */

#define TCP_PAWS_24DAYS	(60 * 60 * 24 * 24)
#define TCP_PAWS_MSL	60		/* Per-host timestamps are invalidated
					 * after this time. It should be equal
					 * (or greater than) TCP_TIMEWAIT_LEN
					 * to provide reliability equal to one
					 * provided by timewait state.
					 */
#define TCP_PAWS_WINDOW	1		/* Replay window for per-host
					 * timestamps. It must be less than
					 * minimal timewait lifetime.
					 */
/*
 *	TCP option
 */
172

Linus Torvalds's avatar
Linus Torvalds committed
173
174
175
176
177
178
179
#define TCPOPT_NOP		1	/* Padding */
#define TCPOPT_EOL		0	/* End of options */
#define TCPOPT_MSS		2	/* Segment size negotiating */
#define TCPOPT_WINDOW		3	/* Window scaling */
#define TCPOPT_SACK_PERM        4       /* SACK Permitted */
#define TCPOPT_SACK             5       /* SACK Block */
#define TCPOPT_TIMESTAMP	8	/* Better RTT estimations/PAWS */
180
#define TCPOPT_MD5SIG		19	/* MD5 Signature (RFC2385) */
181
#define TCPOPT_FASTOPEN		34	/* Fast open (RFC7413) */
Yuchung Cheng's avatar
Yuchung Cheng committed
182
183
184
185
186
#define TCPOPT_EXP		254	/* Experimental */
/* Magic number to be after the option value for sharing TCP
 * experimental options. See draft-ietf-tcpm-experimental-options-00.txt
 */
#define TCPOPT_FASTOPEN_MAGIC	0xF989
Linus Torvalds's avatar
Linus Torvalds committed
187
188
189
190
191
192
193
194
195

/*
 *     TCP option lengths
 */

#define TCPOLEN_MSS            4
#define TCPOLEN_WINDOW         3
#define TCPOLEN_SACK_PERM      2
#define TCPOLEN_TIMESTAMP      10
196
#define TCPOLEN_MD5SIG         18
197
#define TCPOLEN_FASTOPEN_BASE  2
Yuchung Cheng's avatar
Yuchung Cheng committed
198
#define TCPOLEN_EXP_FASTOPEN_BASE  4
Linus Torvalds's avatar
Linus Torvalds committed
199
200
201
202
203
204
205
206

/* But this is what stacks really send out. */
#define TCPOLEN_TSTAMP_ALIGNED		12
#define TCPOLEN_WSCALE_ALIGNED		4
#define TCPOLEN_SACKPERM_ALIGNED	4
#define TCPOLEN_SACK_BASE		2
#define TCPOLEN_SACK_BASE_ALIGNED	4
#define TCPOLEN_SACK_PERBLOCK		8
207
#define TCPOLEN_MD5SIG_ALIGNED		20
Adam Langley's avatar
Adam Langley committed
208
#define TCPOLEN_MSS_ALIGNED		4
Linus Torvalds's avatar
Linus Torvalds committed
209
210
211
212

/* Flags in tp->nonagle */
#define TCP_NAGLE_OFF		1	/* Nagle's algo is disabled */
#define TCP_NAGLE_CORK		2	/* Socket is corked	    */
Stephen Hemminger's avatar
Stephen Hemminger committed
213
#define TCP_NAGLE_PUSH		4	/* Cork is overridden for already queued data */
Linus Torvalds's avatar
Linus Torvalds committed
214

215
216
217
/* TCP thin-stream limits */
#define TCP_THIN_LINEAR_RETRIES 6       /* After 6 linear retries, do exp. backoff */

218
/* TCP initial congestion window as per rfc6928 */
219
220
#define TCP_INIT_CWND		10

221
222
/* Bit Flags for sysctl_tcp_fastopen */
#define	TFO_CLIENT_ENABLE	1
223
#define	TFO_SERVER_ENABLE	2
224
#define	TFO_CLIENT_NO_COOKIE	4	/* Data in SYN w/o cookie option */
225

226
227
228
229
/* Accept SYN data w/o any cookie option */
#define	TFO_SERVER_COOKIE_NOT_REQD	0x200

/* Force enable TFO on all listeners, i.e., not requiring the
230
 * TCP_FASTOPEN socket option.
231
232
233
 */
#define	TFO_SERVER_WO_SOCKOPT1	0x400

234
235
extern struct inet_timewait_death_row tcp_death_row;

Linus Torvalds's avatar
Linus Torvalds committed
236
237
238
239
/* sysctl variables for tcp */
extern int sysctl_tcp_timestamps;
extern int sysctl_tcp_window_scaling;
extern int sysctl_tcp_sack;
Yuchung Cheng's avatar
Yuchung Cheng committed
240
extern int sysctl_tcp_fastopen;
Linus Torvalds's avatar
Linus Torvalds committed
241
242
243
244
245
246
247
extern int sysctl_tcp_retrans_collapse;
extern int sysctl_tcp_stdurg;
extern int sysctl_tcp_rfc1337;
extern int sysctl_tcp_abort_on_overflow;
extern int sysctl_tcp_max_orphans;
extern int sysctl_tcp_fack;
extern int sysctl_tcp_reordering;
248
extern int sysctl_tcp_max_reordering;
Linus Torvalds's avatar
Linus Torvalds committed
249
extern int sysctl_tcp_dsack;
250
extern long sysctl_tcp_mem[3];
Linus Torvalds's avatar
Linus Torvalds committed
251
252
253
254
255
256
257
258
259
260
extern int sysctl_tcp_wmem[3];
extern int sysctl_tcp_rmem[3];
extern int sysctl_tcp_app_win;
extern int sysctl_tcp_adv_win_scale;
extern int sysctl_tcp_tw_reuse;
extern int sysctl_tcp_frto;
extern int sysctl_tcp_low_latency;
extern int sysctl_tcp_nometrics_save;
extern int sysctl_tcp_moderate_rcvbuf;
extern int sysctl_tcp_tso_win_divisor;
261
extern int sysctl_tcp_workaround_signed_windows;
262
extern int sysctl_tcp_slow_start_after_idle;
263
extern int sysctl_tcp_thin_linear_timeouts;
Andreas Petlund's avatar
Andreas Petlund committed
264
extern int sysctl_tcp_thin_dupack;
Yuchung Cheng's avatar
Yuchung Cheng committed
265
extern int sysctl_tcp_early_retrans;
Eric Dumazet's avatar
Eric Dumazet committed
266
extern int sysctl_tcp_limit_output_bytes;
Eric Dumazet's avatar
Eric Dumazet committed
267
extern int sysctl_tcp_challenge_ack_limit;
268
extern int sysctl_tcp_min_tso_segs;
269
extern int sysctl_tcp_min_rtt_wlen;
Eric Dumazet's avatar
Eric Dumazet committed
270
extern int sysctl_tcp_autocorking;
271
extern int sysctl_tcp_invalid_ratelimit;
272
273
extern int sysctl_tcp_pacing_ss_ratio;
extern int sysctl_tcp_pacing_ca_ratio;
Linus Torvalds's avatar
Linus Torvalds committed
274

Eric Dumazet's avatar
Eric Dumazet committed
275
extern atomic_long_t tcp_memory_allocated;
276
extern struct percpu_counter tcp_sockets_allocated;
Linus Torvalds's avatar
Linus Torvalds committed
277
278
extern int tcp_memory_pressure;

279
280
281
/* optimized version of sk_under_memory_pressure() for TCP sockets */
static inline bool tcp_under_memory_pressure(const struct sock *sk)
{
282
283
	if (mem_cgroup_sockets_enabled && sk->sk_memcg &&
	    mem_cgroup_under_socket_pressure(sk->sk_memcg))
284
		return true;
285
286
287

	return tcp_memory_pressure;
}
Linus Torvalds's avatar
Linus Torvalds committed
288
289
290
291
292
/*
 * The next routines deal with comparing 32 bit unsigned ints
 * and worry about wraparound (automatic with unsigned arithmetic).
 */

Eric Dumazet's avatar
Eric Dumazet committed
293
static inline bool before(__u32 seq1, __u32 seq2)
Linus Torvalds's avatar
Linus Torvalds committed
294
{
295
        return (__s32)(seq1-seq2) < 0;
Linus Torvalds's avatar
Linus Torvalds committed
296
}
297
#define after(seq2, seq1) 	before(seq1, seq2)
Linus Torvalds's avatar
Linus Torvalds committed
298
299

/* is s2<=s1<=s3 ? */
Eric Dumazet's avatar
Eric Dumazet committed
300
static inline bool between(__u32 seq1, __u32 seq2, __u32 seq3)
Linus Torvalds's avatar
Linus Torvalds committed
301
302
303
304
{
	return seq3 - seq2 >= seq1 - seq2;
}

Arun Sharma's avatar
Arun Sharma committed
305
306
307
308
309
310
311
312
static inline bool tcp_out_of_memory(struct sock *sk)
{
	if (sk->sk_wmem_queued > SOCK_MIN_SNDBUF &&
	    sk_memory_allocated(sk) > sk_prot_mem_limits(sk, 2))
		return true;
	return false;
}

313
314
void sk_forced_mem_schedule(struct sock *sk, int size);

315
static inline bool tcp_too_many_orphans(struct sock *sk, int shift)
316
{
317
318
319
320
321
322
323
324
325
	struct percpu_counter *ocp = sk->sk_prot->orphan_count;
	int orphans = percpu_counter_read_positive(ocp);

	if (orphans << shift > sysctl_tcp_max_orphans) {
		orphans = percpu_counter_sum_positive(ocp);
		if (orphans << shift > sysctl_tcp_max_orphans)
			return true;
	}
	return false;
326
}
Linus Torvalds's avatar
Linus Torvalds committed
327

328
bool tcp_check_oom(struct sock *sk, int shift);
Arun Sharma's avatar
Arun Sharma committed
329

330

Linus Torvalds's avatar
Linus Torvalds committed
331
332
extern struct proto tcp_prot;

333
#define TCP_INC_STATS(net, field)	SNMP_INC_STATS((net)->mib.tcp_statistics, field)
Eric Dumazet's avatar
Eric Dumazet committed
334
#define __TCP_INC_STATS(net, field)	__SNMP_INC_STATS((net)->mib.tcp_statistics, field)
335
#define TCP_DEC_STATS(net, field)	SNMP_DEC_STATS((net)->mib.tcp_statistics, field)
336
#define TCP_ADD_STATS(net, field, val)	SNMP_ADD_STATS((net)->mib.tcp_statistics, field, val)
Linus Torvalds's avatar
Linus Torvalds committed
337

338
339
340
341
342
343
344
345
346
347
void tcp_tasklet_init(void);

void tcp_v4_err(struct sk_buff *skb, u32);

void tcp_shutdown(struct sock *sk, int how);

void tcp_v4_early_demux(struct sk_buff *skb);
int tcp_v4_rcv(struct sk_buff *skb);

int tcp_v4_tw_remember_stamp(struct inet_timewait_sock *tw);
348
int tcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t size);
349
350
351
352
353
354
355
int tcp_sendpage(struct sock *sk, struct page *page, int offset, size_t size,
		 int flags);
void tcp_release_cb(struct sock *sk);
void tcp_wfree(struct sk_buff *skb);
void tcp_write_timer_handler(struct sock *sk);
void tcp_delack_timer_handler(struct sock *sk);
int tcp_ioctl(struct sock *sk, int cmd, unsigned long arg);
356
int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb);
357
358
359
360
361
362
363
364
void tcp_rcv_established(struct sock *sk, struct sk_buff *skb,
			 const struct tcphdr *th, unsigned int len);
void tcp_rcv_space_adjust(struct sock *sk);
int tcp_twsk_unique(struct sock *sk, struct sock *sktw, void *twp);
void tcp_twsk_destructor(struct sock *sk);
ssize_t tcp_splice_read(struct socket *sk, loff_t *ppos,
			struct pipe_inode_info *pipe, size_t len,
			unsigned int flags);
Jens Axboe's avatar
Jens Axboe committed
365

366
367
static inline void tcp_dec_quickack_mode(struct sock *sk,
					 const unsigned int pkts)
Linus Torvalds's avatar
Linus Torvalds committed
368
{
369
	struct inet_connection_sock *icsk = inet_csk(sk);
370

371
372
373
	if (icsk->icsk_ack.quick) {
		if (pkts >= icsk->icsk_ack.quick) {
			icsk->icsk_ack.quick = 0;
374
			/* Leaving quickack mode we deflate ATO. */
375
			icsk->icsk_ack.ato   = TCP_ATO_MIN;
376
		} else
377
			icsk->icsk_ack.quick -= pkts;
Linus Torvalds's avatar
Linus Torvalds committed
378
379
380
	}
}

381
382
383
#define	TCP_ECN_OK		1
#define	TCP_ECN_QUEUE_CWR	2
#define	TCP_ECN_DEMAND_CWR	4
384
#define	TCP_ECN_SEEN		8
385

Eric Dumazet's avatar
Eric Dumazet committed
386
enum tcp_tw_status {
Linus Torvalds's avatar
Linus Torvalds committed
387
388
389
390
391
392
393
	TCP_TW_SUCCESS = 0,
	TCP_TW_RST = 1,
	TCP_TW_ACK = 2,
	TCP_TW_SYN = 3
};


394
395
396
397
enum tcp_tw_status tcp_timewait_state_process(struct inet_timewait_sock *tw,
					      struct sk_buff *skb,
					      const struct tcphdr *th);
struct sock *tcp_check_req(struct sock *sk, struct sk_buff *skb,
398
			   struct request_sock *req, bool fastopen);
399
400
int tcp_child_process(struct sock *parent, struct sock *child,
		      struct sk_buff *skb);
401
void tcp_enter_loss(struct sock *sk);
402
403
404
405
406
void tcp_clear_retrans(struct tcp_sock *tp);
void tcp_update_metrics(struct sock *sk);
void tcp_init_metrics(struct sock *sk);
void tcp_metrics_init(void);
bool tcp_peer_is_proven(struct request_sock *req, struct dst_entry *dst,
407
			bool paws_check, bool timestamps);
408
409
410
411
412
413
414
415
416
417
418
419
420
bool tcp_remember_stamp(struct sock *sk);
bool tcp_tw_remember_stamp(struct inet_timewait_sock *tw);
void tcp_fetch_timewait_stamp(struct sock *sk, struct dst_entry *dst);
void tcp_disable_fack(struct tcp_sock *tp);
void tcp_close(struct sock *sk, long timeout);
void tcp_init_sock(struct sock *sk);
unsigned int tcp_poll(struct file *file, struct socket *sock,
		      struct poll_table_struct *wait);
int tcp_getsockopt(struct sock *sk, int level, int optname,
		   char __user *optval, int __user *optlen);
int tcp_setsockopt(struct sock *sk, int level, int optname,
		   char __user *optval, unsigned int optlen);
int compat_tcp_getsockopt(struct sock *sk, int level, int optname,
Changli Gao's avatar
Changli Gao committed
421
			  char __user *optval, int __user *optlen);
422
int compat_tcp_setsockopt(struct sock *sk, int level, int optname,
Changli Gao's avatar
Changli Gao committed
423
			  char __user *optval, unsigned int optlen);
424
void tcp_set_keepalive(struct sock *sk, int val);
425
void tcp_syn_ack_timeout(const struct request_sock *req);
426
427
int tcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, int nonblock,
		int flags, int *addr_len);
428
429
430
431
void tcp_parse_options(const struct sk_buff *skb,
		       struct tcp_options_received *opt_rx,
		       int estab, struct tcp_fastopen_cookie *foc);
const u8 *tcp_parse_md5sig_option(const struct tcphdr *th);
432

Linus Torvalds's avatar
Linus Torvalds committed
433
434
435
436
/*
 *	TCP v4 functions exported for the inet6 API
 */

437
void tcp_v4_send_check(struct sock *sk, struct sk_buff *skb);
438
void tcp_v4_mtu_reduced(struct sock *sk);
439
void tcp_req_err(struct sock *sk, u32 seq, bool abort);
440
int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb);
441
struct sock *tcp_create_openreq_child(const struct sock *sk,
442
443
				      struct request_sock *req,
				      struct sk_buff *skb);
444
void tcp_ca_openreq_child(struct sock *sk, const struct dst_entry *dst);
445
struct sock *tcp_v4_syn_recv_sock(const struct sock *sk, struct sk_buff *skb,
446
				  struct request_sock *req,
447
448
449
				  struct dst_entry *dst,
				  struct request_sock *req_unhash,
				  bool *own_req);
450
451
452
int tcp_v4_do_rcv(struct sock *sk, struct sk_buff *skb);
int tcp_v4_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len);
int tcp_connect(struct sock *sk);
453
454
455
456
457
enum tcp_synack_type {
	TCP_SYNACK_NORMAL,
	TCP_SYNACK_FASTOPEN,
	TCP_SYNACK_COOKIE,
};
458
struct sk_buff *tcp_make_synack(const struct sock *sk, struct dst_entry *dst,
459
				struct request_sock *req,
460
				struct tcp_fastopen_cookie *foc,
461
				enum tcp_synack_type synack_type);
462
int tcp_disconnect(struct sock *sk, int flags);
Linus Torvalds's avatar
Linus Torvalds committed
463

Pavel Emelyanov's avatar
Pavel Emelyanov committed
464
void tcp_finish_connect(struct sock *sk, struct sk_buff *skb);
465
int tcp_send_rcvq(struct sock *sk, struct msghdr *msg, size_t size);
466
void inet_sk_rx_dst_set(struct sock *sk, const struct sk_buff *skb);
Linus Torvalds's avatar
Linus Torvalds committed
467
468

/* From syncookies.c */
469
470
471
struct sock *tcp_get_cookie_sock(struct sock *sk, struct sk_buff *skb,
				 struct request_sock *req,
				 struct dst_entry *dst);
472
473
int __cookie_v4_check(const struct iphdr *iph, const struct tcphdr *th,
		      u32 cookie);
Cong Wang's avatar
Cong Wang committed
474
struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb);
475
#ifdef CONFIG_SYN_COOKIES
476

477
/* Syncookies use a monotonic timer which increments every 60 seconds.
478
479
480
 * This counter is used both as a hash input and partially encoded into
 * the cookie value.  A cookie is only validated further if the delta
 * between the current counter value and the encoded one is less than this,
481
 * i.e. a sent cookie is valid only at most for 2*60 seconds (or less if
482
483
 * the counter advances immediately after a cookie is generated).
 */
484
485
486
487
488
489
#define MAX_SYNCOOKIE_AGE	2
#define TCP_SYNCOOKIE_PERIOD	(60 * HZ)
#define TCP_SYNCOOKIE_VALID	(MAX_SYNCOOKIE_AGE * TCP_SYNCOOKIE_PERIOD)

/* syncookies: remember time of last synqueue overflow
 * But do not dirty this field too often (once per second is enough)
490
 * It is racy as we do not hold a lock, but race is very minor.
491
 */
492
static inline void tcp_synq_overflow(const struct sock *sk)
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
{
	unsigned long last_overflow = tcp_sk(sk)->rx_opt.ts_recent_stamp;
	unsigned long now = jiffies;

	if (time_after(now, last_overflow + HZ))
		tcp_sk(sk)->rx_opt.ts_recent_stamp = now;
}

/* syncookies: no recent synqueue overflow on this listening socket? */
static inline bool tcp_synq_no_recent_overflow(const struct sock *sk)
{
	unsigned long last_overflow = tcp_sk(sk)->rx_opt.ts_recent_stamp;

	return time_after(jiffies, last_overflow + TCP_SYNCOOKIE_VALID);
}
508
509
510

static inline u32 tcp_cookie_time(void)
{
511
512
	u64 val = get_jiffies_64();

513
	do_div(val, TCP_SYNCOOKIE_PERIOD);
514
	return val;
515
516
}

517
518
u32 __cookie_v4_init_sequence(const struct iphdr *iph, const struct tcphdr *th,
			      u16 *mssp);
519
__u32 cookie_v4_init_sequence(const struct sk_buff *skb, __u16 *mss);
520
__u32 cookie_init_timestamp(struct request_sock *req);
521
522
bool cookie_timestamp_decode(struct tcp_options_received *opt);
bool cookie_ecn_ok(const struct tcp_options_received *opt,
523
		   const struct net *net, const struct dst_entry *dst);
524

525
/* From net/ipv6/syncookies.c */
526
527
528
int __cookie_v6_check(const struct ipv6hdr *iph, const struct tcphdr *th,
		      u32 cookie);
struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb);
529

530
531
u32 __cookie_v6_init_sequence(const struct ipv6hdr *iph,
			      const struct tcphdr *th, u16 *mssp);
532
__u32 cookie_v6_init_sequence(const struct sk_buff *skb, __u16 *mss);
533
#endif
Linus Torvalds's avatar
Linus Torvalds committed
534
535
/* tcp_output.c */

536
537
u32 tcp_tso_autosize(const struct sock *sk, unsigned int mss_now,
		     int min_tso_segs);
538
539
540
void __tcp_push_pending_frames(struct sock *sk, unsigned int cur_mss,
			       int nonagle);
bool tcp_may_send_now(struct sock *sk);
541
542
int __tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb, int segs);
int tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb, int segs);
543
544
545
546
void tcp_retransmit_timer(struct sock *sk);
void tcp_xmit_retransmit_queue(struct sock *);
void tcp_simple_retransmit(struct sock *);
int tcp_trim_head(struct sock *, struct sk_buff *, u32);
547
int tcp_fragment(struct sock *, struct sk_buff *, u32, unsigned int, gfp_t);
548
549
550

void tcp_send_probe0(struct sock *);
void tcp_send_partial(struct sock *);
551
int tcp_write_wakeup(struct sock *, int mib);
552
553
554
555
556
557
558
559
void tcp_send_fin(struct sock *sk);
void tcp_send_active_reset(struct sock *sk, gfp_t priority);
int tcp_send_synack(struct sock *);
void tcp_push_one(struct sock *, unsigned int mss_now);
void tcp_send_ack(struct sock *sk);
void tcp_send_delayed_ack(struct sock *sk);
void tcp_send_loss_probe(struct sock *sk);
bool tcp_schedule_loss_probe(struct sock *sk);
560
561
void tcp_skb_collapse_tstamp(struct sk_buff *skb,
			     const struct sk_buff *next_skb);
Linus Torvalds's avatar
Linus Torvalds committed
562

563
/* tcp_input.c */
564
565
void tcp_resume_early_retransmit(struct sock *sk);
void tcp_rearm_rto(struct sock *sk);
566
void tcp_synack_rtt_meas(struct sock *sk, struct request_sock *req);
567
void tcp_reset(struct sock *sk);
Yuchung Cheng's avatar
Yuchung Cheng committed
568
void tcp_skb_mark_lost_uncond_verify(struct tcp_sock *tp, struct sk_buff *skb);
569
void tcp_fin(struct sock *sk);
570

Linus Torvalds's avatar
Linus Torvalds committed
571
/* tcp_timer.c */
572
void tcp_init_xmit_timers(struct sock *);
573
574
575
576
static inline void tcp_clear_xmit_timers(struct sock *sk)
{
	inet_csk_clear_xmit_timers(sk);
}
Linus Torvalds's avatar
Linus Torvalds committed
577

578
579
unsigned int tcp_sync_mss(struct sock *sk, u32 pmtu);
unsigned int tcp_current_mss(struct sock *sk);
Ilpo Järvinen's avatar
Ilpo Järvinen committed
580
581
582
583

/* Bound MSS / TSO packet size with the half of the window */
static inline int tcp_bound_to_half_wnd(struct tcp_sock *tp, int pktsize)
{
584
585
586
587
588
589
590
591
592
	int cutoff;

	/* When peer uses tiny windows, there is no use in packetizing
	 * to sub-MSS pieces for the sake of SWS or making sure there
	 * are enough packets in the pipe for fast recovery.
	 *
	 * On the other hand, for extremely large MSS devices, handling
	 * smaller than MSS windows in this way does make sense.
	 */
593
	if (tp->max_window > TCP_MSS_DEFAULT)
594
595
596
597
598
599
		cutoff = (tp->max_window >> 1);
	else
		cutoff = tp->max_window;

	if (cutoff && pktsize > cutoff)
		return max_t(int, cutoff, 68U - tp->tcp_header_len);
Ilpo Järvinen's avatar
Ilpo Järvinen committed
600
601
602
	else
		return pktsize;
}
Linus Torvalds's avatar
Linus Torvalds committed
603

604
/* tcp.c */
605
void tcp_get_info(struct sock *, struct tcp_info *);
Linus Torvalds's avatar
Linus Torvalds committed
606
607

/* Read 'sendfile()'-style from a TCP socket */
608
609
int tcp_read_sock(struct sock *sk, read_descriptor_t *desc,
		  sk_read_actor_t recv_actor);
Linus Torvalds's avatar
Linus Torvalds committed
610

611
void tcp_initialize_rcv_mss(struct sock *sk);
Linus Torvalds's avatar
Linus Torvalds committed
612

613
614
615
616
int tcp_mtu_to_mss(struct sock *sk, int pmtu);
int tcp_mss_to_mtu(struct sock *sk, int mss);
void tcp_mtup_init(struct sock *sk);
void tcp_init_buffer_space(struct sock *sk);
John Heffner's avatar
John Heffner committed
617

618
619
620
621
622
623
624
625
static inline void tcp_bound_rto(const struct sock *sk)
{
	if (inet_csk(sk)->icsk_rto > TCP_RTO_MAX)
		inet_csk(sk)->icsk_rto = TCP_RTO_MAX;
}

static inline u32 __tcp_set_rto(const struct tcp_sock *tp)
{
626
	return usecs_to_jiffies((tp->srtt_us >> 3) + tp->rttvar_us);
627
628
}

Stephen Hemminger's avatar
Stephen Hemminger committed
629
static inline void __tcp_fast_path_on(struct tcp_sock *tp, u32 snd_wnd)
Linus Torvalds's avatar
Linus Torvalds committed
630
631
632
633
634
635
{
	tp->pred_flags = htonl((tp->tcp_header_len << 26) |
			       ntohl(TCP_FLAG_ACK) |
			       snd_wnd);
}

Stephen Hemminger's avatar
Stephen Hemminger committed
636
static inline void tcp_fast_path_on(struct tcp_sock *tp)
Linus Torvalds's avatar
Linus Torvalds committed
637
638
639
640
{
	__tcp_fast_path_on(tp, tp->snd_wnd >> tp->rx_opt.snd_wscale);
}

641
static inline void tcp_fast_path_check(struct sock *sk)
Linus Torvalds's avatar
Linus Torvalds committed
642
{
643
644
	struct tcp_sock *tp = tcp_sk(sk);

645
	if (RB_EMPTY_ROOT(&tp->out_of_order_queue) &&
Linus Torvalds's avatar
Linus Torvalds committed
646
647
648
649
650
651
	    tp->rcv_wnd &&
	    atomic_read(&sk->sk_rmem_alloc) < sk->sk_rcvbuf &&
	    !tp->urg_data)
		tcp_fast_path_on(tp);
}

652
653
654
/* Compute the actual rto_min value */
static inline u32 tcp_rto_min(struct sock *sk)
{
655
	const struct dst_entry *dst = __sk_dst_get(sk);
656
657
658
659
660
661
662
	u32 rto_min = TCP_RTO_MIN;

	if (dst && dst_metric_locked(dst, RTAX_RTO_MIN))
		rto_min = dst_metric_rtt(dst, RTAX_RTO_MIN);
	return rto_min;
}

663
664
665
666
667
static inline u32 tcp_rto_min_us(struct sock *sk)
{
	return jiffies_to_usecs(tcp_rto_min(sk));
}

668
669
670
671
672
static inline bool tcp_ca_dst_locked(const struct dst_entry *dst)
{
	return dst_metric_locked(dst, RTAX_CC_ALGO);
}

673
674
675
/* Minimum RTT in usec. ~0 means not available. */
static inline u32 tcp_min_rtt(const struct tcp_sock *tp)
{
676
	return minmax_get(&tp->rtt_min);
677
678
}

Linus Torvalds's avatar
Linus Torvalds committed
679
680
681
682
/* Compute the actual receive window we are currently advertising.
 * Rcv_nxt can be after the window if our peer push more data
 * than the offered window.
 */
Stephen Hemminger's avatar
Stephen Hemminger committed
683
static inline u32 tcp_receive_window(const struct tcp_sock *tp)
Linus Torvalds's avatar
Linus Torvalds committed
684
685
686
687
688
689
690
691
692
693
694
695
{
	s32 win = tp->rcv_wup + tp->rcv_wnd - tp->rcv_nxt;

	if (win < 0)
		win = 0;
	return (u32) win;
}

/* Choose a new window, without checks for shrinking, and without
 * scaling applied to the result.  The caller does these things
 * if necessary.  This is a "raw" window selection.
 */
696
u32 __tcp_select_window(struct sock *sk);
Linus Torvalds's avatar
Linus Torvalds committed
697

Pavel Emelyanov's avatar
Pavel Emelyanov committed
698
699
void tcp_send_window_probe(struct sock *sk);

Linus Torvalds's avatar
Linus Torvalds committed
700
701
/* TCP timestamps are only 32-bits, this causes a slight
 * complication on 64-bit systems since we store a snapshot
702
703
 * of jiffies in the buffer control blocks below.  We decided
 * to use only the low 32-bits of jiffies and hide the ugly
Linus Torvalds's avatar
Linus Torvalds committed
704
705
706
707
 * casts with the following macro.
 */
#define tcp_time_stamp		((__u32)(jiffies))

708
709
710
711
712
713
static inline u32 tcp_skb_timestamp(const struct sk_buff *skb)
{
	return skb->skb_mstamp.stamp_jiffies;
}


Changli Gao's avatar
Changli Gao committed
714
715
716
717
718
719
720
721
722
723
724
#define tcp_flag_byte(th) (((u_int8_t *)th)[13])

#define TCPHDR_FIN 0x01
#define TCPHDR_SYN 0x02
#define TCPHDR_RST 0x04
#define TCPHDR_PSH 0x08
#define TCPHDR_ACK 0x10
#define TCPHDR_URG 0x20
#define TCPHDR_ECE 0x40
#define TCPHDR_CWR 0x80

725
726
#define TCPHDR_SYN_ECN	(TCPHDR_SYN | TCPHDR_ECE | TCPHDR_CWR)

Stephen Hemminger's avatar
Stephen Hemminger committed
727
/* This is what the send packet queuing engine uses to pass
728
729
730
731
 * TCP per-packet control information to the transmission code.
 * We also store the host-order sequence numbers in here too.
 * This is 44 bytes if IPV6 is enabled.
 * If this grows please adjust skbuff.h:skbuff->cb[xxx] size appropriately.
Linus Torvalds's avatar
Linus Torvalds committed
732
733
734
735
 */
struct tcp_skb_cb {
	__u32		seq;		/* Starting sequence number	*/
	__u32		end_seq;	/* SEQ + FIN + SYN + datalen	*/
736
737
738
739
	union {
		/* Note : tcp_tw_isn is used in input path only
		 *	  (isn chosen by tcp_timewait_state_process())
		 *
740
741
		 * 	  tcp_gso_segs/size are used in write queue only,
		 *	  cf tcp_skb_pcount()/tcp_skb_mss()
742
743
		 */
		__u32		tcp_tw_isn;
744
745
746
747
		struct {
			u16	tcp_gso_segs;
			u16	tcp_gso_size;
		};
748
	};
Eric Dumazet's avatar
Eric Dumazet committed
749
	__u8		tcp_flags;	/* TCP header flags. (tcp[13])	*/
750

Linus Torvalds's avatar
Linus Torvalds committed
751
752
753
754
755
	__u8		sacked;		/* State flags for SACK/FACK.	*/
#define TCPCB_SACKED_ACKED	0x01	/* SKB ACK'd by a SACK block	*/
#define TCPCB_SACKED_RETRANS	0x02	/* SKB retransmitted		*/
#define TCPCB_LOST		0x04	/* SKB is lost			*/
#define TCPCB_TAGBITS		0x07	/* All tag bits			*/
756
#define TCPCB_REPAIRED		0x10	/* SKB repaired (no skb_mstamp)	*/
Linus Torvalds's avatar
Linus Torvalds committed
757
#define TCPCB_EVER_RETRANS	0x80	/* Ever retransmitted frame	*/
758
759
#define TCPCB_RETRANS		(TCPCB_SACKED_RETRANS|TCPCB_EVER_RETRANS| \
				TCPCB_REPAIRED)
Linus Torvalds's avatar
Linus Torvalds committed
760

761
	__u8		ip_dsfield;	/* IPv4 tos or IPv6 dsfield	*/
762
	__u8		txstamp_ack:1,	/* Record TX timestamp for ack? */
763
764
			eor:1,		/* Is skb MSG_EOR marked? */
			unused:6;
Linus Torvalds's avatar
Linus Torvalds committed
765
	__u32		ack_seq;	/* Sequence number ACK'd	*/
766
	union {
767
		struct {
768
			/* There is space for up to 24 bytes */
769
770
771
			__u32 in_flight:30,/* Bytes in flight at transmit */
			      is_app_limited:1, /* cwnd not fully used? */
			      unused:1;
772
773
774
775
776
777
			/* pkts S/ACKed so far upon tx of skb, incl retrans: */
			__u32 delivered;
			/* start of send pipeline phase */
			struct skb_mstamp first_tx_mstamp;
			/* when we reached the "delivered" count */
			struct skb_mstamp delivered_mstamp;
778
779
780
		} tx;   /* only used for outgoing skbs */
		union {
			struct inet_skb_parm	h4;
781
#if IS_ENABLED(CONFIG_IPV6)
782
			struct inet6_skb_parm	h6;
783
#endif
784
785
		} header;	/* For incoming skbs */
	};
Linus Torvalds's avatar
Linus Torvalds committed
786
787
788
789
};

#define TCP_SKB_CB(__skb)	((struct tcp_skb_cb *)&((__skb)->cb[0]))

Eric Dumazet's avatar
Eric Dumazet committed
790

791
#if IS_ENABLED(CONFIG_IPV6)
Eric Dumazet's avatar
Eric Dumazet committed
792
793
794
795
796
/* This is the variant of inet6_iif() that must be used by TCP,
 * as TCP moves IP6CB into a different location in skb->cb[]
 */
static inline int tcp_v6_iif(const struct sk_buff *skb)
{
797
	bool l3_slave = ipv6_l3mdev_skb(TCP_SKB_CB(skb)->header.h6.flags);
798
799

	return l3_slave ? skb->skb_iif : TCP_SKB_CB(skb)->header.h6.iif;
Eric Dumazet's avatar
Eric Dumazet committed
800
}
801
#endif
Eric Dumazet's avatar
Eric Dumazet committed
802

803
804
805
806
807
/* TCP_SKB_CB reference means this can not be used from early demux */
static inline bool inet_exact_dif_match(struct net *net, struct sk_buff *skb)
{
#if IS_ENABLED(CONFIG_NET_L3_MASTER_DEV)
	if (!net->ipv4.sysctl_tcp_l3mdev_accept &&
808
	    skb && ipv4_l3mdev_skb(TCP_SKB_CB(skb)->header.h4.flags))
809
810
811
812
813
		return true;
#endif
	return false;
}

Linus Torvalds's avatar
Linus Torvalds committed
814
815
/* Due to TSO, an SKB can be composed of multiple actual
 * packets.  To keep these tracked properly, we use this.
816
 */
Linus Torvalds's avatar
Linus Torvalds committed
817
static inline int tcp_skb_pcount(const struct sk_buff *skb)
818
{
819
820
	return TCP_SKB_CB(skb)->tcp_gso_segs;
}
821

822
823
824
static inline void tcp_skb_pcount_set(struct sk_buff *skb, int segs)
{
	TCP_SKB_CB(skb)->tcp_gso_segs = segs;
825
826
}

827
static inline void tcp_skb_pcount_add(struct sk_buff *skb, int segs)
Linus Torvalds's avatar
Linus Torvalds committed
828
{
829
	TCP_SKB_CB(skb)->tcp_gso_segs += segs;
Linus Torvalds's avatar
Linus Torvalds committed
830
831
}

832
/* This is valid iff skb is in write queue and tcp_skb_pcount() > 1. */
Linus Torvalds's avatar
Linus Torvalds committed
833
834
static inline int tcp_skb_mss(const struct sk_buff *skb)
{
835
	return TCP_SKB_CB(skb)->tcp_gso_size;
Linus Torvalds's avatar
Linus Torvalds committed
836
837
}

838
839
840
841
842
static inline bool tcp_skb_can_collapse_to(const struct sk_buff *skb)
{
	return likely(!TCP_SKB_CB(skb)->eor);
}

843
844
845
846
847
848
/* Events passed to congestion control interface */
enum tcp_ca_event {
	CA_EVENT_TX_START,	/* first transmit when no packets in flight */
	CA_EVENT_CWND_RESTART,	/* congestion window restart */
	CA_EVENT_COMPLETE_CWR,	/* end of congestion recovery */
	CA_EVENT_LOSS,		/* loss timeout */
849
850
851
852
	CA_EVENT_ECN_NO_CE,	/* ECT set, but not CE marked */
	CA_EVENT_ECN_IS_CE,	/* received CE marked IP packet */
	CA_EVENT_DELAYED_ACK,	/* Delayed ack is sent */
	CA_EVENT_NON_DELAYED_ACK,
853
854
};

855
/* Information about inbound ACK, passed to cong_ops->in_ack_event() */
856
enum tcp_ca_ack_event_flags {
857
858
859
	CA_ACK_SLOWPATH		= (1 << 0),	/* In slow path processing */
	CA_ACK_WIN_UPDATE	= (1 << 1),	/* ACK updated window */
	CA_ACK_ECE		= (1 << 2),	/* ECE bit is set on ack */
860
861
862
863
864
865
};

/*
 * Interface for adding new TCP congestion control handlers
 */
#define TCP_CA_NAME_MAX	16
866
867
868
#define TCP_CA_MAX	128
#define TCP_CA_BUF_MAX	(TCP_CA_NAME_MAX*TCP_CA_MAX)

869
870
#define TCP_CA_UNSPEC	0

871
/* Algorithm can be set on socket without CAP_NET_ADMIN privileges */
872
#define TCP_CONG_NON_RESTRICTED 0x1
873
874
/* Requires ECN/ECT set on all packets */
#define TCP_CONG_NEEDS_ECN	0x2
875

876
877
union tcp_cc_info;

878
879
880
struct ack_sample {
	u32 pkts_acked;
	s32 rtt_us;
881
	u32 in_flight;
882
883
};

884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
/* A rate sample measures the number of (original/retransmitted) data
 * packets delivered "delivered" over an interval of time "interval_us".
 * The tcp_rate.c code fills in the rate sample, and congestion
 * control modules that define a cong_control function to run at the end
 * of ACK processing can optionally chose to consult this sample when
 * setting cwnd and pacing rate.
 * A sample is invalid if "delivered" or "interval_us" is negative.
 */
struct rate_sample {
	struct	skb_mstamp prior_mstamp; /* starting timestamp for interval */
	u32  prior_delivered;	/* tp->delivered at "prior_mstamp" */
	s32  delivered;		/* number of packets delivered over interval */
	long interval_us;	/* time for tp->delivered to incr "delivered" */
	long rtt_us;		/* RTT of last (S)ACKed packet (or -1) */
	int  losses;		/* number of packets marked lost upon ACK */
	u32  acked_sacked;	/* number of packets newly (S)ACKed upon ACK */
	u32  prior_in_flight;	/* in flight before this ACK */
901
	bool is_app_limited;	/* is sample from packet with bubble in pipe? */
902
903
904
	bool is_retrans;	/* is sample from retransmission? */
};

905
906
struct tcp_congestion_ops {
	struct list_head	list;
907
908
	u32 key;
	u32 flags;
909
910

	/* initialize private data (optional) */
911
	void (*init)(struct sock *sk);
912
	/* cleanup private data  (optional) */
913
	void (*release)(struct sock *sk);
914
915

	/* return slow start threshold (required) */
916
	u32 (*ssthresh)(struct sock *sk);
917
	/* do new cwnd calculation (required) */
918
	void (*cong_avoid)(struct sock *sk, u32 ack, u32 acked);
919
	/* call before changing ca_state (optional) */
920
	void (*set_state)(struct sock *sk, u8 new_state);
921
	/* call when cwnd event occurs (optional) */
922
	void (*cwnd_event)(struct sock *sk, enum tcp_ca_event ev);
923
924
	/* call when ack arrives (optional) */
	void (*in_ack_event)(struct sock *sk, u32 flags);
925
	/* new value of cwnd after loss (optional) */
926
	u32  (*undo_cwnd)(struct sock *sk);
927
	/* hook for packet ack accounting (optional) */
928
	void (*pkts_acked)(struct sock *sk, const struct ack_sample *sample);
929
930
	/* suggest number of segments for each skb to transmit (optional) */
	u32 (*tso_segs_goal)(struct sock *sk);
931
932
	/* returns the multiplier used in tcp_sndbuf_expand (optional) */
	u32 (*sndbuf_expand)(struct sock *sk);
933
934
935
936
	/* call when packets are delivered to update cwnd and pacing rate,
	 * after all the ca_state processing. (optional)
	 */
	void (*cong_control)(struct sock *sk, const struct rate_sample *rs);
937
	/* get info for inet_diag (optional) */
938
939
	size_t (*get_info)(struct sock *sk, u32 ext, int *attr,
			   union tcp_cc_info *info);
940
941
942
943
944

	char 		name[TCP_CA_NAME_MAX];
	struct module 	*owner;
};

945
946
int tcp_register_congestion_control(struct tcp_congestion_ops *type);
void tcp_unregister_congestion_control(struct tcp_congestion_ops *type);
947

948
void tcp_assign_congestion_control(struct sock *sk);
949
950
951
952
953
954
955
956
void tcp_init_congestion_control(struct sock *sk);
void tcp_cleanup_congestion_control(struct sock *sk);
int tcp_set_default_congestion_control(const char *name);
void tcp_get_default_congestion_control(char *name);
void tcp_get_available_congestion_control(char *buf, size_t len);
void tcp_get_allowed_congestion_control(char *buf, size_t len);
int tcp_set_allowed_congestion_control(char *allowed);
int tcp_set_congestion_control(struct sock *sk, const char *name);
Neal Cardwell's avatar
Neal Cardwell committed
957
958
u32 tcp_slow_start(struct tcp_sock *tp, u32 acked);
void tcp_cong_avoid_ai(struct tcp_sock *tp, u32 w, u32 acked);
959

960
u32 tcp_reno_ssthresh(struct sock *sk);
961
void tcp_reno_cong_avoid(struct sock *sk, u32 ack, u32 acked);
962
extern struct tcp_congestion_ops tcp_reno;
963

964
struct tcp_congestion_ops *tcp_ca_find_key(u32 key);
965
u32 tcp_ca_get_key_by_name(const char *name, bool *ecn_ca);
966
#ifdef CONFIG_INET
967
char *tcp_ca_get_name_by_key(u32 key, char *buffer);
968
969
970
971
972
973
#else
static inline char *tcp_ca_get_name_by_key(u32 key, char *buffer)
{
	return NULL;
}
#endif
974

975
976
977
978
979
980
981
static inline bool tcp_ca_needs_ecn(const struct sock *sk)
{
	const struct inet_connection_sock *icsk = inet_csk(sk);

	return icsk->icsk_ca_ops->flags & TCP_CONG_NEEDS_ECN;
}

982
static inline void tcp_set_ca_state(struct sock *sk, const u8 ca_state)
983
{
984
985
986
987
988
	struct inet_connection_sock *icsk = inet_csk(sk);

	if (icsk->icsk_ca_ops->set_state)
		icsk->icsk_ca_ops->set_state(sk, ca_state);
	icsk->icsk_ca_state = ca_state;
989
990
}

991
static inline void tcp_ca_event(struct sock *sk, const enum tcp_ca_event event)
992
{
993
994
995
996
	const struct inet_connection_sock *icsk = inet_csk(sk);

	if (icsk->icsk_ca_ops->cwnd_event)
		icsk->icsk_ca_ops->cwnd_event(sk, event);
997
998
}

999
1000
1001
1002
1003
1004
/* From tcp_rate.c */
void tcp_rate_skb_sent(struct sock *sk, struct sk_buff *skb);
void tcp_rate_skb_delivered(struct sock *sk, struct sk_buff *skb,
			    struct rate_sample *rs);
void tcp_rate_gen(struct sock *sk, u32 delivered, u32 lost,
		  struct skb_mstamp *now, struct rate_sample *rs);
1005
void tcp_rate_check_app_limited(struct sock *sk);
1006

1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
/* These functions determine how the current flow behaves in respect of SACK
 * handling. SACK is negotiated with the peer, and therefore it can vary
 * between different flows.
 *
 * tcp_is_sack - SACK enabled
 * tcp_is_reno - No SACK
 * tcp_is_fack - FACK enabled, implies SACK enabled
 */
static inline int tcp_is_sack(const struct tcp_sock *tp)
{
	return tp->rx_opt.sack_ok;
}

Eric Dumazet's avatar
Eric Dumazet committed
1020
static inline bool tcp_is_reno(const struct tcp_sock *tp)
1021
1022
1023
1024
{
	return !tcp_is_sack(tp);
}

Eric Dumazet's avatar
Eric Dumazet committed
1025
static inline bool tcp_is_fack(const struct tcp_sock *tp)
1026
{
1027
	return tp->rx_opt.sack_ok & TCP_FACK_ENABLED;
1028
1029
1030
1031
}

static inline void tcp_enable_fack(struct tcp_sock *tp)
{
1032
	tp->rx_opt.sack_ok |= TCP_FACK_ENABLED;
1033
1034
}

Yuchung Cheng's avatar
Yuchung Cheng committed
1035
1036
1037
1038
1039
/* TCP early-retransmit (ER) is similar to but more conservative than
 * the thin-dupack feature.  Enable ER only if thin-dupack is disabled.
 */
static inline void tcp_enable_early_retrans(struct tcp_sock *tp)
{
1040
1041
	struct net *net = sock_net((struct sock *)tp);

Yuchung Cheng's avatar
Yuchung Cheng committed
1042
	tp->do_early_retrans = sysctl_tcp_early_retrans &&
Nandita Dukkipati's avatar
Nandita Dukkipati committed
1043
		sysctl_tcp_early_retrans < 4 && !sysctl_tcp_thin_dupack &&
1044
		net->ipv4.sysctl_tcp_reordering == 3;
Yuchung Cheng's avatar
Yuchung Cheng committed
1045
1046
1047
1048
1049
1050
1051
}

static inline void tcp_disable_early_retrans(struct tcp_sock *tp)
{
	tp->do_early_retrans = 0;
}

1052
1053
1054
1055
1056
static inline unsigned int tcp_left_out(const struct tcp_sock *tp)
{
	return tp->sacked_out + tp->lost_out;
}

Linus Torvalds's avatar
Linus Torvalds committed
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
/* This determines how many packets are "in the network" to the best
 * of our knowledge.  In many cases it is conservative, but where
 * detailed information is available from the receiver (via SACK
 * blocks etc.) we can make more aggressive calculations.
 *
 * Use this for decisions involving congestion control, use just
 * tp->packets_out to determine if the send queue is empty or not.
 *
 * Read this equation as:
 *
 *	"Packets sent once on transmission queue" MINUS
 *	"Packets left network, but not honestly ACKed yet" PLUS
 *	"Packets fast retransmitted"
 */
Stephen Hemminger's avatar
Stephen Hemminger committed
1071
static inline unsigned int tcp_packets_in_flight(const struct tcp_sock *tp)
Linus Torvalds's avatar
Linus Torvalds committed
1072
{
1073
	return tp->packets_out - tcp_left_out(tp) + tp->retrans_out;
Linus Torvalds's avatar
Linus Torvalds committed
1074
1075
}

1076
1077
#define TCP_INFINITE_SSTHRESH	0x7fffffff

1078
1079
static inline bool tcp_in_slow_start(const struct tcp_sock *tp)
{
1080
	return tp->snd_cwnd < tp->snd_ssthresh;
1081
1082
}

1083
1084
1085
1086
1087
static inline bool tcp_in_initial_slowstart(const struct tcp_sock *tp)
{
	return tp->snd_ssthresh >= TCP_INFINITE_SSTHRESH;
}

1088
1089
1090
1091
1092
1093
static inline bool tcp_in_cwnd_reduction(const struct sock *sk)
{
	return (TCPF_CA_CWR | TCPF_CA_Recovery) &
	       (1 << inet_csk(sk)->icsk_ca_state);
}

Linus Torvalds's avatar
Linus Torvalds committed
1094
/* If cwnd > ssthresh, we may raise ssthresh to be half-way to cwnd.
1095
 * The exception is cwnd reduction phase, when cwnd is decreasing towards
Linus Torvalds's avatar
Linus Torvalds committed
1096
1097
 * ssthresh.
 */
1098
static inline __u32 tcp_current_ssthresh(const struct sock *sk)
Linus Torvalds's avatar
Linus Torvalds committed
1099
{
1100
	const struct tcp_sock *tp = tcp_sk(sk);
1101

1102
	if (tcp_in_cwnd_reduction(sk))
Linus Torvalds's avatar
Linus Torvalds committed
1103
1104
1105
1106
1107
1108
1109
		return tp->snd_ssthresh;
	else
		return max(tp->snd_ssthresh,
			   ((tp->snd_cwnd >> 1) +
			    (tp->snd_cwnd >> 2)));
}

1110
1111
/* Use define here intentionally to get WARN_ON location shown at the caller */
#define tcp_verify_left_out(tp)	WARN_ON(tcp_left_out(tp) > tp->packets_out)