2 * SPDX-License-Identifier: BSD-3-Clause
4 * Copyright (c) 1982, 1986, 1993
5 * The Regents of the University of California. All rights reserved.
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
10 * 1. Redistributions of source code must retain the above copyright
11 * notice, this list of conditions and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution.
15 * 3. Neither the name of the University nor the names of its contributors
16 * may be used to endorse or promote products derived from this software
17 * without specific prior written permission.
19 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
31 * @(#)tcp.h 8.1 (Berkeley) 6/10/93
35 #ifndef _NETINET_TCP_H_
36 #define _NETINET_TCP_H_
38 #include <sys/cdefs.h>
39 #include <sys/types.h>
43 typedef u_int32_t tcp_seq;
45 #define tcp6_seq tcp_seq /* for KAME src sync over BSD*'s */
46 #define tcp6hdr tcphdr /* for KAME src sync over BSD*'s */
50 * Per RFC 793, September, 1981.
53 u_short th_sport; /* source port */
54 u_short th_dport; /* destination port */
55 tcp_seq th_seq; /* sequence number */
56 tcp_seq th_ack; /* acknowledgement number */
57 #if BYTE_ORDER == LITTLE_ENDIAN
58 u_char th_x2:4, /* (unused) */
59 th_off:4; /* data offset */
61 #if BYTE_ORDER == BIG_ENDIAN
62 u_char th_off:4, /* data offset */
63 th_x2:4; /* (unused) */
74 #define TH_FLAGS (TH_FIN|TH_SYN|TH_RST|TH_PUSH|TH_ACK|TH_URG|TH_ECE|TH_CWR)
75 #define PRINT_TH_FLAGS "\20\1FIN\2SYN\3RST\4PUSH\5ACK\6URG\7ECE\10CWR"
77 u_short th_win; /* window */
78 u_short th_sum; /* checksum */
79 u_short th_urp; /* urgent pointer */
84 #define TCPOPT_PAD 0 /* padding after EOL */
88 #define TCPOPT_MAXSEG 2
89 #define TCPOLEN_MAXSEG 4
90 #define TCPOPT_WINDOW 3
91 #define TCPOLEN_WINDOW 3
92 #define TCPOPT_SACK_PERMITTED 4
93 #define TCPOLEN_SACK_PERMITTED 2
95 #define TCPOLEN_SACKHDR 2
96 #define TCPOLEN_SACK 8 /* 2*sizeof(tcp_seq) */
97 #define TCPOPT_TIMESTAMP 8
98 #define TCPOLEN_TIMESTAMP 10
99 #define TCPOLEN_TSTAMP_APPA (TCPOLEN_TIMESTAMP+2) /* appendix A */
100 #define TCPOPT_SIGNATURE 19 /* Keyed MD5: RFC 2385 */
101 #define TCPOLEN_SIGNATURE 18
102 #define TCPOPT_FAST_OPEN 34
103 #define TCPOLEN_FAST_OPEN_EMPTY 2
105 /* Miscellaneous constants */
106 #define MAX_SACK_BLKS 6 /* Max # SACK blocks stored at receiver side */
107 #define TCP_MAX_SACK 4 /* MAX # SACKs sent in any segment */
111 * The default maximum segment size (MSS) to be used for new TCP connections
112 * when path MTU discovery is not enabled.
114 * RFC879 derives the default MSS from the largest datagram size hosts are
115 * minimally required to handle directly or through IP reassembly minus the
116 * size of the IP and TCP header. With IPv6 the minimum MTU is specified
119 * For IPv4 the MSS is 576 - sizeof(struct tcpiphdr)
120 * For IPv6 the MSS is IPV6_MMTU - sizeof(struct ip6_hdr) - sizeof(struct tcphdr)
122 * We use explicit numerical definition here to avoid header pollution.
125 #define TCP6_MSS 1220
128 * Limit the lowest MSS we accept for path MTU discovery and the TCP SYN MSS
129 * option. Allowing low values of MSS can consume significant resources and
130 * be used to mount a resource exhaustion attack.
131 * Connections requesting lower MSS values will be rounded up to this value
132 * and the IP_DF flag will be cleared to allow fragmentation along the path.
134 * See tcp_subr.c tcp_minmss SYSCTL declaration for more comments. Setting
135 * it to "0" disables the minmss check.
137 * The default value is fine for TCP across the Internet's smallest official
138 * link MTU (256 bytes for AX.25 packet radio). However, a connection is very
139 * unlikely to come across such low MTU interfaces these days (anno domini 2003).
141 #define TCP_MINMSS 216
143 #define TCP_MAXWIN 65535 /* largest value for (unscaled) window */
144 #define TTCP_CLIENT_SND_WND 4096 /* dflt send window for T/TCP client */
146 #define TCP_MAX_WINSHIFT 14 /* maximum window shift */
148 #define TCP_MAXBURST 4 /* maximum segments in a burst */
150 #define TCP_MAXHLEN (0xf<<2) /* max length of header in bytes */
151 #define TCP_MAXOLEN (TCP_MAXHLEN - sizeof(struct tcphdr))
152 /* max space left for options */
154 #define TCP_FASTOPEN_MIN_COOKIE_LEN 4 /* Per RFC7413 */
155 #define TCP_FASTOPEN_MAX_COOKIE_LEN 16 /* Per RFC7413 */
156 #define TCP_FASTOPEN_PSK_LEN 16 /* Same as TCP_FASTOPEN_KEY_LEN */
157 #endif /* __BSD_VISIBLE */
160 * User-settable options (used with setsockopt). These are discrete
161 * values and are not masked together. Some values appear to be
162 * bitmasks for historical reasons.
164 #define TCP_NODELAY 1 /* don't delay send to coalesce packets */
166 #define TCP_MAXSEG 2 /* set maximum segment size */
167 #define TCP_NOPUSH 4 /* don't push last block of write */
168 #define TCP_NOOPT 8 /* don't use TCP options */
169 #define TCP_MD5SIG 16 /* use MD5 digests (RFC2385) */
170 #define TCP_INFO 32 /* retrieve tcp_info structure */
171 #define TCP_LOG 34 /* configure event logging for connection */
172 #define TCP_LOGBUF 35 /* retrieve event log for connection */
173 #define TCP_LOGID 36 /* configure log ID to correlate connections */
174 #define TCP_LOGDUMP 37 /* dump connection log events to device */
175 #define TCP_LOGDUMPID 38 /* dump events from connections with same ID to
177 #define TCP_CONGESTION 64 /* get/set congestion control algorithm */
178 #define TCP_CCALGOOPT 65 /* get/set cc algorithm specific options */
179 #define TCP_DELACK 72 /* socket option for delayed ack */
180 #define TCP_KEEPINIT 128 /* N, time to establish connection */
181 #define TCP_KEEPIDLE 256 /* L,N,X start keeplives after this period */
182 #define TCP_KEEPINTVL 512 /* L,N interval between keepalives */
183 #define TCP_KEEPCNT 1024 /* L,N number of keepalives before close */
184 #define TCP_FASTOPEN 1025 /* enable TFO / was created via TFO */
185 #define TCP_PCAP_OUT 2048 /* number of output packets to keep */
186 #define TCP_PCAP_IN 4096 /* number of input packets to keep */
187 #define TCP_FUNCTION_BLK 8192 /* Set the tcp function pointers to the specified stack */
188 /* Options for Rack and BBR */
189 #define TCP_RACK_PROP 1051 /* RACK proportional rate reduction (bool) */
190 #define TCP_RACK_TLP_REDUCE 1052 /* RACK TLP cwnd reduction (bool) */
191 #define TCP_RACK_PACE_REDUCE 1053 /* RACK Pacing reduction factor (divisor) */
192 #define TCP_RACK_PACE_MAX_SEG 1054 /* Max segments in a pace */
193 #define TCP_RACK_PACE_ALWAYS 1055 /* Use the always pace method */
194 #define TCP_RACK_PROP_RATE 1056 /* The proportional reduction rate */
195 #define TCP_RACK_PRR_SENDALOT 1057 /* Allow PRR to send more than one seg */
196 #define TCP_RACK_MIN_TO 1058 /* Minimum time between rack t-o's in ms */
197 #define TCP_RACK_EARLY_RECOV 1059 /* Should recovery happen early (bool) */
198 #define TCP_RACK_EARLY_SEG 1060 /* If early recovery max segments */
199 #define TCP_RACK_REORD_THRESH 1061 /* RACK reorder threshold (shift amount) */
200 #define TCP_RACK_REORD_FADE 1062 /* Does reordering fade after ms time */
201 #define TCP_RACK_TLP_THRESH 1063 /* RACK TLP theshold i.e. srtt+(srtt/N) */
202 #define TCP_RACK_PKT_DELAY 1064 /* RACK added ms i.e. rack-rtt + reord + N */
203 #define TCP_RACK_TLP_INC_VAR 1065 /* Does TLP include rtt variance in t-o */
204 #define TCP_BBR_IWINTSO 1067 /* Initial TSO window for BBRs first sends */
205 #define TCP_BBR_RECFORCE 1068 /* Enter recovery force out a segment disregard pacer no longer valid */
206 #define TCP_BBR_STARTUP_PG 1069 /* Startup pacing gain */
207 #define TCP_BBR_DRAIN_PG 1070 /* Drain pacing gain */
208 #define TCP_BBR_RWND_IS_APP 1071 /* Rwnd limited is considered app limited */
209 #define TCP_BBR_PROBE_RTT_INT 1072 /* How long in useconds between probe-rtt */
210 #define TCP_BBR_ONE_RETRAN 1073 /* Is only one segment allowed out during retran */
211 #define TCP_BBR_STARTUP_LOSS_EXIT 1074 /* Do we exit a loss during startup if not 20% incr */
212 #define TCP_BBR_USE_LOWGAIN 1075 /* lower the gain in PROBE_BW enable */
213 #define TCP_BBR_LOWGAIN_THRESH 1076 /* Unused after 2.3 morphs to TSLIMITS >= 2.3 */
214 #define TCP_BBR_TSLIMITS 1076 /* Do we use experimental Timestamp limiting for our algo */
215 #define TCP_BBR_LOWGAIN_HALF 1077 /* Unused after 2.3 */
216 #define TCP_BBR_PACE_OH 1077 /* Reused in 4.2 for pacing overhead setting */
217 #define TCP_BBR_LOWGAIN_FD 1078 /* Unused after 2.3 */
218 #define TCP_BBR_HOLD_TARGET 1078 /* For 4.3 on */
219 #define TCP_BBR_USEDEL_RATE 1079 /* Enable use of delivery rate for loss recovery */
220 #define TCP_BBR_MIN_RTO 1080 /* Min RTO in milliseconds */
221 #define TCP_BBR_MAX_RTO 1081 /* Max RTO in milliseconds */
222 #define TCP_BBR_REC_OVER_HPTS 1082 /* Recovery override htps settings 0/1/3 */
223 #define TCP_BBR_UNLIMITED 1083 /* Not used before 2.3 and morphs to algorithm >= 2.3 */
224 #define TCP_BBR_ALGORITHM 1083 /* What measurement algo does BBR use netflix=0, google=1 */
225 #define TCP_BBR_DRAIN_INC_EXTRA 1084 /* Does the 3/4 drain target include the extra gain */
226 #define TCP_BBR_STARTUP_EXIT_EPOCH 1085 /* what epoch gets us out of startup */
227 #define TCP_BBR_PACE_PER_SEC 1086
228 #define TCP_BBR_PACE_DEL_TAR 1087
229 #define TCP_BBR_PACE_SEG_MAX 1088
230 #define TCP_BBR_PACE_SEG_MIN 1089
231 #define TCP_BBR_PACE_CROSS 1090
232 #define TCP_RACK_IDLE_REDUCE_HIGH 1092 /* Reduce the highest cwnd seen to IW on idle */
233 #define TCP_RACK_MIN_PACE 1093 /* Do we enforce rack min pace time */
234 #define TCP_RACK_MIN_PACE_SEG 1094 /* If so what is the seg threshould */
235 #define TCP_RACK_GP_INCREASE 1094 /* After 4.1 its the GP increase */
236 #define TCP_RACK_TLP_USE 1095
237 #define TCP_BBR_ACK_COMP_ALG 1096 /* Not used */
238 #define TCP_BBR_TMR_PACE_OH 1096 /* Recycled in 4.2 */
239 #define TCP_BBR_EXTRA_GAIN 1097
240 #define TCP_BBR_RACK_RTT_USE 1098 /* what RTT should we use 0, 1, or 2? */
241 #define TCP_BBR_RETRAN_WTSO 1099
242 #define TCP_DATA_AFTER_CLOSE 1100
243 #define TCP_BBR_PROBE_RTT_GAIN 1101
244 #define TCP_BBR_PROBE_RTT_LEN 1102
245 #define TCP_BBR_SEND_IWND_IN_TSO 1103 /* Do we burst out whole iwin size chunks at start? */
246 #define TCP_BBR_USE_RACK_CHEAT 1104 /* Do we use the rack cheat for pacing rxt's */
247 #define TCP_BBR_HDWR_PACE 1105 /* Enable/disable hardware pacing */
248 #define TCP_BBR_UTTER_MAX_TSO 1106 /* Do we enforce an utter max TSO size */
249 #define TCP_BBR_EXTRA_STATE 1107 /* Special exit-persist catch up */
250 #define TCP_BBR_FLOOR_MIN_TSO 1108 /* The min tso size */
251 #define TCP_BBR_MIN_TOPACEOUT 1109 /* Do we suspend pacing until */
252 #define TCP_BBR_TSTMP_RAISES 1110 /* Can a timestamp measurement raise the b/w */
253 #define TCP_BBR_POLICER_DETECT 1111 /* Turn on/off google mode policer detection */
256 /* Start of reserved space for third-party user-settable options. */
257 #define TCP_VENDOR SO_VENDOR
259 #define TCP_CA_NAME_MAX 16 /* max congestion control name length */
261 #define TCPI_OPT_TIMESTAMPS 0x01
262 #define TCPI_OPT_SACK 0x02
263 #define TCPI_OPT_WSCALE 0x04
264 #define TCPI_OPT_ECN 0x08
265 #define TCPI_OPT_TOE 0x10
267 /* Maximum length of log ID. */
268 #define TCP_LOG_ID_LEN 64
271 * The TCP_INFO socket option comes from the Linux 2.6 TCP API, and permits
272 * the caller to query certain information about the state of a TCP
273 * connection. We provide an overlapping set of fields with the Linux
274 * implementation, but since this is a fixed size structure, room has been
275 * left for growth. In order to maximize potential future compatibility with
276 * the Linux API, the same variable names and order have been adopted, and
277 * padding left to make room for omitted fields in case they are added later.
279 * XXX: This is currently an unstable ABI/API, in that it is expected to
283 u_int8_t tcpi_state; /* TCP FSM state. */
284 u_int8_t __tcpi_ca_state;
285 u_int8_t __tcpi_retransmits;
286 u_int8_t __tcpi_probes;
287 u_int8_t __tcpi_backoff;
288 u_int8_t tcpi_options; /* Options enabled on conn. */
289 u_int8_t tcpi_snd_wscale:4, /* RFC1323 send shift value. */
290 tcpi_rcv_wscale:4; /* RFC1323 recv shift value. */
292 u_int32_t tcpi_rto; /* Retransmission timeout (usec). */
293 u_int32_t __tcpi_ato;
294 u_int32_t tcpi_snd_mss; /* Max segment size for send. */
295 u_int32_t tcpi_rcv_mss; /* Max segment size for receive. */
297 u_int32_t __tcpi_unacked;
298 u_int32_t __tcpi_sacked;
299 u_int32_t __tcpi_lost;
300 u_int32_t __tcpi_retrans;
301 u_int32_t __tcpi_fackets;
303 /* Times; measurements in usecs. */
304 u_int32_t __tcpi_last_data_sent;
305 u_int32_t __tcpi_last_ack_sent; /* Also unimpl. on Linux? */
306 u_int32_t tcpi_last_data_recv; /* Time since last recv data. */
307 u_int32_t __tcpi_last_ack_recv;
309 /* Metrics; variable units. */
310 u_int32_t __tcpi_pmtu;
311 u_int32_t __tcpi_rcv_ssthresh;
312 u_int32_t tcpi_rtt; /* Smoothed RTT in usecs. */
313 u_int32_t tcpi_rttvar; /* RTT variance in usecs. */
314 u_int32_t tcpi_snd_ssthresh; /* Slow start threshold. */
315 u_int32_t tcpi_snd_cwnd; /* Send congestion window. */
316 u_int32_t __tcpi_advmss;
317 u_int32_t __tcpi_reordering;
319 u_int32_t __tcpi_rcv_rtt;
320 u_int32_t tcpi_rcv_space; /* Advertised recv window. */
322 /* FreeBSD extensions to tcp_info. */
323 u_int32_t tcpi_snd_wnd; /* Advertised send window. */
324 u_int32_t tcpi_snd_bwnd; /* No longer used. */
325 u_int32_t tcpi_snd_nxt; /* Next egress seqno */
326 u_int32_t tcpi_rcv_nxt; /* Next ingress seqno */
327 u_int32_t tcpi_toe_tid; /* HWTID for TOE endpoints */
328 u_int32_t tcpi_snd_rexmitpack; /* Retransmitted packets */
329 u_int32_t tcpi_rcv_ooopack; /* Out-of-order packets */
330 u_int32_t tcpi_snd_zerowin; /* Zero-sized windows sent */
332 /* Padding to grow without breaking ABI. */
333 u_int32_t __tcpi_pad[26]; /* Padding. */
337 * If this structure is provided when setting the TCP_FASTOPEN socket
338 * option, and the enable member is non-zero, a subsequent connect will use
339 * pre-shared key (PSK) mode using the provided key.
341 struct tcp_fastopen {
343 uint8_t psk[TCP_FASTOPEN_PSK_LEN];
346 #define TCP_FUNCTION_NAME_LEN_MAX 32
348 struct tcp_function_set {
349 char function_set_name[TCP_FUNCTION_NAME_LEN_MAX];
353 #endif /* !_NETINET_TCP_H_ */