1 /* $OpenBSD: tcp_var.h,v 1.113 2016/06/18 10:36:13 vgross Exp $ */ 2 /* $NetBSD: tcp_var.h,v 1.17 1996/02/13 23:44:24 christos Exp $ */ 3 4 /* 5 * Copyright (c) 1982, 1986, 1993, 1994 6 * The Regents of the University of California. All rights reserved. 7 * 8 * Redistribution and use in source and binary forms, with or without 9 * modification, are permitted provided that the following conditions 10 * are met: 11 * 1. Redistributions of source code must retain the above copyright 12 * notice, this list of conditions and the following disclaimer. 13 * 2. Redistributions in binary form must reproduce the above copyright 14 * notice, this list of conditions and the following disclaimer in the 15 * documentation and/or other materials provided with the distribution. 16 * 3. Neither the name of the University nor the names of its contributors 17 * may be used to endorse or promote products derived from this software 18 * without specific prior written permission. 19 * 20 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 21 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 22 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 23 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 24 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 25 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 26 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 27 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 28 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 29 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 30 * SUCH DAMAGE. 31 * 32 * @(#)tcp_var.h 8.3 (Berkeley) 4/10/94 33 */ 34 35 #ifndef _NETINET_TCP_VAR_H_ 36 #define _NETINET_TCP_VAR_H_ 37 38 #include <sys/timeout.h> 39 40 /* 41 * Kernel variables for tcp. 42 */ 43 44 struct sackblk { 45 tcp_seq start; /* start seq no. of sack block */ 46 tcp_seq end; /* end seq no. */ 47 }; 48 49 struct sackhole { 50 tcp_seq start; /* start seq no. of hole */ 51 tcp_seq end; /* end seq no. */ 52 int dups; /* number of dup(s)acks for this hole */ 53 tcp_seq rxmit; /* next seq. no in hole to be retransmitted */ 54 struct sackhole *next; /* next in list */ 55 }; 56 57 /* 58 * TCP sequence queue structures. 59 */ 60 TAILQ_HEAD(tcpqehead, tcpqent); 61 struct tcpqent { 62 TAILQ_ENTRY(tcpqent) tcpqe_q; 63 struct tcphdr *tcpqe_tcp; 64 struct mbuf *tcpqe_m; /* mbuf contains packet */ 65 }; 66 67 /* 68 * Tcp control block, one per tcp; fields: 69 */ 70 struct tcpcb { 71 struct tcpqehead t_segq; /* sequencing queue */ 72 struct timeout t_timer[TCPT_NTIMERS]; /* tcp timers */ 73 short t_state; /* state of this connection */ 74 short t_rxtshift; /* log(2) of rexmt exp. backoff */ 75 short t_rxtcur; /* current retransmit value */ 76 short t_dupacks; /* consecutive dup acks recd */ 77 u_short t_maxseg; /* maximum segment size */ 78 char t_force; /* 1 if forcing out a byte */ 79 u_int t_flags; 80 #define TF_ACKNOW 0x0001 /* ack peer immediately */ 81 #define TF_DELACK 0x0002 /* ack, but try to delay it */ 82 #define TF_NODELAY 0x0004 /* don't delay packets to coalesce */ 83 #define TF_NOOPT 0x0008 /* don't use tcp options */ 84 #define TF_SENTFIN 0x0010 /* have sent FIN */ 85 #define TF_REQ_SCALE 0x0020 /* have/will request window scaling */ 86 #define TF_RCVD_SCALE 0x0040 /* other side has requested scaling */ 87 #define TF_REQ_TSTMP 0x0080 /* have/will request timestamps */ 88 #define TF_RCVD_TSTMP 0x0100 /* a timestamp was received in SYN */ 89 #define TF_SACK_PERMIT 0x0200 /* other side said I could SACK */ 90 #define TF_SIGNATURE 0x0400 /* require TCP MD5 signature */ 91 #ifdef TCP_ECN 92 #define TF_ECN_PERMIT 0x00008000 /* other side said I could ECN */ 93 #define TF_RCVD_CE 0x00010000 /* send ECE in subsequent segs */ 94 #define TF_SEND_CWR 0x00020000 /* send CWR in next seg */ 95 #define TF_DISABLE_ECN 0x00040000 /* disable ECN for this connection */ 96 #endif 97 #define TF_LASTIDLE 0x00100000 /* no outstanding ACK on last send */ 98 #define TF_DEAD 0x00200000 /* dead and to-be-released */ 99 #define TF_PMTUD_PEND 0x00400000 /* Path MTU Discovery pending */ 100 #define TF_NEEDOUTPUT 0x00800000 /* call tcp_output after tcp_input */ 101 #define TF_BLOCKOUTPUT 0x01000000 /* avert tcp_output during tcp_input */ 102 #define TF_NOPUSH 0x02000000 /* don't push */ 103 104 struct mbuf *t_template; /* skeletal packet for transmit */ 105 struct inpcb *t_inpcb; /* back pointer to internet pcb */ 106 struct timeout t_delack_to; /* delayed ACK callback */ 107 /* 108 * The following fields are used as in the protocol specification. 109 * See RFC793, Dec. 1981, page 21. 110 */ 111 /* send sequence variables */ 112 tcp_seq snd_una; /* send unacknowledged */ 113 tcp_seq snd_nxt; /* send next */ 114 tcp_seq snd_up; /* send urgent pointer */ 115 tcp_seq snd_wl1; /* window update seg seq number */ 116 tcp_seq snd_wl2; /* window update seg ack number */ 117 tcp_seq iss; /* initial send sequence number */ 118 u_long snd_wnd; /* send window */ 119 #if 1 /*def TCP_SACK*/ 120 int sack_enable; /* enable SACK for this connection */ 121 int snd_numholes; /* number of holes seen by sender */ 122 struct sackhole *snd_holes; /* linked list of holes (sorted) */ 123 #if 1 /*defined(TCP_SACK) && defined(TCP_FACK)*/ 124 tcp_seq snd_fack; /* for FACK congestion control */ 125 u_long snd_awnd; /* snd_nxt - snd_fack + */ 126 /* retransmitted data */ 127 int retran_data; /* amount of outstanding retx. data */ 128 #endif /* TCP_FACK */ 129 #endif /* TCP_SACK */ 130 #if 1 /*defined(TCP_SACK) || defined(TCP_ECN)*/ 131 tcp_seq snd_last; /* for use in fast recovery */ 132 #endif 133 /* receive sequence variables */ 134 u_long rcv_wnd; /* receive window */ 135 tcp_seq rcv_nxt; /* receive next */ 136 tcp_seq rcv_up; /* receive urgent pointer */ 137 tcp_seq irs; /* initial receive sequence number */ 138 #if 1 /*def TCP_SACK*/ 139 tcp_seq rcv_lastsack; /* last seq number(+1) sack'd by rcv'r*/ 140 int rcv_numsacks; /* # distinct sack blks present */ 141 struct sackblk sackblks[MAX_SACK_BLKS]; /* seq nos. of sack blocks */ 142 #endif 143 144 /* 145 * Additional variables for this implementation. 146 */ 147 /* receive variables */ 148 tcp_seq rcv_adv; /* advertised window */ 149 /* retransmit variables */ 150 tcp_seq snd_max; /* highest sequence number sent; 151 * used to recognize retransmits 152 */ 153 /* congestion control (for slow start, source quench, retransmit after loss) */ 154 u_long snd_cwnd; /* congestion-controlled window */ 155 u_long snd_ssthresh; /* snd_cwnd size threshold for 156 * for slow start exponential to 157 * linear switch 158 */ 159 160 /* auto-sizing variables */ 161 u_int rfbuf_cnt; /* recv buffer autoscaling byte count */ 162 u_int32_t rfbuf_ts; /* recv buffer autoscaling time stamp */ 163 164 u_short t_maxopd; /* mss plus options */ 165 u_short t_peermss; /* peer's maximum segment size */ 166 167 /* 168 * transmit timing stuff. See below for scale of srtt and rttvar. 169 * "Variance" is actually smoothed difference. 170 */ 171 uint32_t t_rcvtime; /* time last segment received */ 172 uint32_t t_rtttime; /* time we started measuring rtt */ 173 tcp_seq t_rtseq; /* sequence number being timed */ 174 short t_srtt; /* smoothed round-trip time */ 175 short t_rttvar; /* variance in round-trip time */ 176 u_short t_rttmin; /* minimum rtt allowed */ 177 u_long max_sndwnd; /* largest window peer has offered */ 178 179 /* out-of-band data */ 180 char t_oobflags; /* have some */ 181 char t_iobc; /* input character */ 182 #define TCPOOB_HAVEDATA 0x01 183 #define TCPOOB_HADDATA 0x02 184 short t_softerror; /* possible error not yet reported */ 185 186 /* RFC 1323 variables */ 187 u_char snd_scale; /* window scaling for send window */ 188 u_char rcv_scale; /* window scaling for recv window */ 189 u_char request_r_scale; /* pending window scaling */ 190 u_char requested_s_scale; 191 u_int32_t ts_recent; /* timestamp echo data */ 192 u_int32_t ts_modulate; /* modulation on timestamp */ 193 u_int32_t ts_recent_age; /* when last updated */ 194 tcp_seq last_ack_sent; 195 196 /* pointer for syn cache entries*/ 197 LIST_HEAD(, syn_cache) t_sc; /* list of entries by this tcb */ 198 199 /* Path-MTU Discovery Information */ 200 u_int t_pmtud_mss_acked; /* MSS acked, lower bound for MTU */ 201 u_int t_pmtud_mtu_sent; /* MTU used, upper bound for MTU */ 202 tcp_seq t_pmtud_th_seq; /* TCP SEQ from ICMP payload */ 203 u_int t_pmtud_nextmtu; /* Advertised Next-Hop MTU from ICMP */ 204 u_short t_pmtud_ip_len; /* IP length from ICMP payload */ 205 u_short t_pmtud_ip_hl; /* IP header length from ICMP payload */ 206 207 int pf; 208 209 struct timeout t_reap_to; /* delayed cleanup timeout */ 210 }; 211 212 #define intotcpcb(ip) ((struct tcpcb *)(ip)->inp_ppcb) 213 #define sototcpcb(so) (intotcpcb(sotoinpcb(so))) 214 215 #ifdef _KERNEL 216 extern int tcp_delack_ticks; 217 void tcp_delack(void *); 218 219 #define TCP_INIT_DELACK(tp) \ 220 timeout_set(&(tp)->t_delack_to, tcp_delack, tp) 221 222 #define TCP_RESTART_DELACK(tp) \ 223 timeout_add(&(tp)->t_delack_to, tcp_delack_ticks) 224 225 #define TCP_SET_DELACK(tp) \ 226 do { \ 227 if (((tp)->t_flags & TF_DELACK) == 0) { \ 228 (tp)->t_flags |= TF_DELACK; \ 229 TCP_RESTART_DELACK(tp); \ 230 } \ 231 } while (/* CONSTCOND */ 0) 232 233 #define TCP_CLEAR_DELACK(tp) \ 234 do { \ 235 if ((tp)->t_flags & TF_DELACK) { \ 236 (tp)->t_flags &= ~TF_DELACK; \ 237 timeout_del(&(tp)->t_delack_to); \ 238 } \ 239 } while (/* CONSTCOND */ 0) 240 241 /* 242 * Handy way of passing around TCP option info. 243 */ 244 struct tcp_opt_info { 245 int ts_present; 246 u_int32_t ts_val; 247 u_int32_t ts_ecr; 248 u_int16_t maxseg; 249 }; 250 251 /* 252 * Data for the TCP compressed state engine. 253 */ 254 255 #define TCP_SYN_HASH_SIZE 293 256 #define TCP_SYN_BUCKET_SIZE 35 257 258 union syn_cache_sa { 259 struct sockaddr sa; 260 struct sockaddr_in sin; 261 struct sockaddr_in6 sin6; 262 }; 263 264 struct syn_cache { 265 TAILQ_ENTRY(syn_cache) sc_bucketq; /* link on bucket list */ 266 struct timeout sc_timer; /* rexmt timer */ 267 union { /* cached route */ 268 struct route route4; 269 #ifdef INET6 270 struct route_in6 route6; 271 #endif 272 } sc_route_u; 273 #define sc_route4 sc_route_u.route4 274 #ifdef INET6 275 #define sc_route6 sc_route_u.route6 276 #endif 277 long sc_win; /* advertised window */ 278 struct syn_cache_head *sc_buckethead; /* our bucket index */ 279 struct syn_cache_set *sc_set; /* our syn cache set */ 280 u_int32_t sc_hash; 281 u_int32_t sc_timestamp; /* timestamp from SYN */ 282 u_int32_t sc_modulate; /* our timestamp modulator */ 283 #if 0 284 u_int32_t sc_timebase; /* our local timebase */ 285 #endif 286 union syn_cache_sa sc_src; 287 union syn_cache_sa sc_dst; 288 tcp_seq sc_irs; 289 tcp_seq sc_iss; 290 u_int sc_rtableid; 291 u_int sc_rxtcur; /* current rxt timeout */ 292 u_int sc_rxttot; /* total time spend on queues */ 293 u_short sc_rxtshift; /* for computing backoff */ 294 u_short sc_flags; 295 296 #define SCF_UNREACH 0x0001 /* we've had an unreach error */ 297 #define SCF_TIMESTAMP 0x0002 /* peer will do timestamps */ 298 #define SCF_DEAD 0x0004 /* this entry to be released */ 299 #define SCF_SACK_PERMIT 0x0008 /* permit sack */ 300 #define SCF_ECN_PERMIT 0x0010 /* permit ecn */ 301 #define SCF_SIGNATURE 0x0020 /* enforce tcp signatures */ 302 303 struct mbuf *sc_ipopts; /* IP options */ 304 u_int16_t sc_peermaxseg; 305 u_int16_t sc_ourmaxseg; 306 u_int sc_request_r_scale : 4, 307 sc_requested_s_scale : 4; 308 309 struct tcpcb *sc_tp; /* tcb for listening socket */ 310 LIST_ENTRY(syn_cache) sc_tpq; /* list of entries by same tp */ 311 }; 312 313 struct syn_cache_head { 314 TAILQ_HEAD(, syn_cache) sch_bucket; /* bucket entries */ 315 u_short sch_length; /* # entries in bucket */ 316 }; 317 318 struct syn_cache_set { 319 struct syn_cache_head scs_buckethead[TCP_SYN_HASH_SIZE]; 320 int scs_count; 321 int scs_use; 322 u_int32_t scs_random[5]; 323 }; 324 325 #endif /* _KERNEL */ 326 327 /* 328 * The smoothed round-trip time and estimated variance 329 * are stored as fixed point numbers scaled by the values below. 330 * For convenience, these scales are also used in smoothing the average 331 * (smoothed = (1/scale)sample + ((scale-1)/scale)smoothed). 332 * With these scales, srtt has 5 bits to the right of the binary point, 333 * and thus an "ALPHA" of 0.875. rttvar has 4 bits to the right of the 334 * binary point, and is smoothed with an ALPHA of 0.75. 335 */ 336 #define TCP_RTT_SHIFT 3 /* shift for srtt; 5 bits frac. */ 337 #define TCP_RTTVAR_SHIFT 2 /* shift for rttvar; 4 bits */ 338 #define TCP_RTT_BASE_SHIFT 2 /* remaining 2 bit shift */ 339 #define TCP_RTT_MAX (1<<9) /* maximum rtt */ 340 341 /* 342 * The initial retransmission should happen at rtt + 4 * rttvar. 343 * Because of the way we do the smoothing, srtt and rttvar 344 * will each average +1/2 tick of bias. When we compute 345 * the retransmit timer, we want 1/2 tick of rounding and 346 * 1 extra tick because of +-1/2 tick uncertainty in the 347 * firing of the timer. The bias will give us exactly the 348 * 1.5 tick we need. But, because the bias is 349 * statistical, we have to test that we don't drop below 350 * the minimum feasible timer (which is 2 ticks). 351 * This macro assumes that the value of (1 << TCP_RTTVAR_SHIFT) 352 * is the same as the multiplier for rttvar. 353 */ 354 #define TCP_REXMTVAL(tp) \ 355 ((((tp)->t_srtt >> TCP_RTT_SHIFT) + (tp)->t_rttvar) >> TCP_RTT_BASE_SHIFT) 356 357 /* 358 * TCP statistics. 359 * Many of these should be kept per connection, 360 * but that's inconvenient at the moment. 361 */ 362 struct tcpstat { 363 u_int32_t tcps_connattempt; /* connections initiated */ 364 u_int32_t tcps_accepts; /* connections accepted */ 365 u_int32_t tcps_connects; /* connections established */ 366 u_int32_t tcps_drops; /* connections dropped */ 367 u_int32_t tcps_conndrops; /* embryonic connections dropped */ 368 u_int32_t tcps_closed; /* conn. closed (includes drops) */ 369 u_int32_t tcps_segstimed; /* segs where we tried to get rtt */ 370 u_int32_t tcps_rttupdated; /* times we succeeded */ 371 u_int32_t tcps_delack; /* delayed acks sent */ 372 u_int32_t tcps_timeoutdrop; /* conn. dropped in rxmt timeout */ 373 u_int32_t tcps_rexmttimeo; /* retransmit timeouts */ 374 u_int32_t tcps_persisttimeo; /* persist timeouts */ 375 u_int32_t tcps_persistdrop; /* connections dropped in persist */ 376 u_int32_t tcps_keeptimeo; /* keepalive timeouts */ 377 u_int32_t tcps_keepprobe; /* keepalive probes sent */ 378 u_int32_t tcps_keepdrops; /* connections dropped in keepalive */ 379 380 u_int32_t tcps_sndtotal; /* total packets sent */ 381 u_int32_t tcps_sndpack; /* data packets sent */ 382 u_int64_t tcps_sndbyte; /* data bytes sent */ 383 u_int32_t tcps_sndrexmitpack; /* data packets retransmitted */ 384 u_int64_t tcps_sndrexmitbyte; /* data bytes retransmitted */ 385 u_int64_t tcps_sndrexmitfast; /* Fast retransmits */ 386 u_int32_t tcps_sndacks; /* ack-only packets sent */ 387 u_int32_t tcps_sndprobe; /* window probes sent */ 388 u_int32_t tcps_sndurg; /* packets sent with URG only */ 389 u_int32_t tcps_sndwinup; /* window update-only packets sent */ 390 u_int32_t tcps_sndctrl; /* control (SYN|FIN|RST) packets sent */ 391 392 u_int32_t tcps_rcvtotal; /* total packets received */ 393 u_int32_t tcps_rcvpack; /* packets received in sequence */ 394 u_int64_t tcps_rcvbyte; /* bytes received in sequence */ 395 u_int32_t tcps_rcvbadsum; /* packets received with ccksum errs */ 396 u_int32_t tcps_rcvbadoff; /* packets received with bad offset */ 397 u_int32_t tcps_rcvmemdrop; /* packets dropped for lack of memory */ 398 u_int32_t tcps_rcvnosec; /* packets dropped for lack of ipsec */ 399 u_int32_t tcps_rcvshort; /* packets received too short */ 400 u_int32_t tcps_rcvduppack; /* duplicate-only packets received */ 401 u_int64_t tcps_rcvdupbyte; /* duplicate-only bytes received */ 402 u_int32_t tcps_rcvpartduppack; /* packets with some duplicate data */ 403 u_int64_t tcps_rcvpartdupbyte; /* dup. bytes in part-dup. packets */ 404 u_int32_t tcps_rcvoopack; /* out-of-order packets received */ 405 u_int64_t tcps_rcvoobyte; /* out-of-order bytes received */ 406 u_int32_t tcps_rcvpackafterwin; /* packets with data after window */ 407 u_int64_t tcps_rcvbyteafterwin; /* bytes rcvd after window */ 408 u_int32_t tcps_rcvafterclose; /* packets rcvd after "close" */ 409 u_int32_t tcps_rcvwinprobe; /* rcvd window probe packets */ 410 u_int32_t tcps_rcvdupack; /* rcvd duplicate acks */ 411 u_int32_t tcps_rcvacktoomuch; /* rcvd acks for unsent data */ 412 u_int32_t tcps_rcvacktooold; /* rcvd acks for old data */ 413 u_int32_t tcps_rcvackpack; /* rcvd ack packets */ 414 u_int64_t tcps_rcvackbyte; /* bytes acked by rcvd acks */ 415 u_int32_t tcps_rcvwinupd; /* rcvd window update packets */ 416 u_int32_t tcps_pawsdrop; /* segments dropped due to PAWS */ 417 u_int32_t tcps_predack; /* times hdr predict ok for acks */ 418 u_int32_t tcps_preddat; /* times hdr predict ok for data pkts */ 419 420 u_int32_t tcps_pcbhashmiss; /* input packets missing pcb hash */ 421 u_int32_t tcps_noport; /* no socket on port */ 422 u_int32_t tcps_badsyn; /* SYN packet with src==dst rcv'ed */ 423 u_int32_t tcps_dropsyn; /* SYN packet dropped */ 424 425 u_int32_t tcps_rcvbadsig; /* rcvd bad/missing TCP signatures */ 426 u_int64_t tcps_rcvgoodsig; /* rcvd good TCP signatures */ 427 u_int32_t tcps_inswcsum; /* input software-checksummed packets */ 428 u_int32_t tcps_outswcsum; /* output software-checksummed packets */ 429 430 /* ECN stats */ 431 u_int32_t tcps_ecn_accepts; /* ecn connections accepted */ 432 u_int32_t tcps_ecn_rcvece; /* # of rcvd ece */ 433 u_int32_t tcps_ecn_rcvcwr; /* # of rcvd cwr */ 434 u_int32_t tcps_ecn_rcvce; /* # of rcvd ce in ip header */ 435 u_int32_t tcps_ecn_sndect; /* # of cwr sent */ 436 u_int32_t tcps_ecn_sndece; /* # of ece sent */ 437 u_int32_t tcps_ecn_sndcwr; /* # of cwr sent */ 438 u_int32_t tcps_cwr_ecn; /* # of cwnd reduced by ecn */ 439 u_int32_t tcps_cwr_frecovery; /* # of cwnd reduced by fastrecovery */ 440 u_int32_t tcps_cwr_timeout; /* # of cwnd reduced by timeout */ 441 442 /* These statistics deal with the SYN cache. */ 443 u_int64_t tcps_sc_added; /* # of entries added */ 444 u_int64_t tcps_sc_completed; /* # of connections completed */ 445 u_int64_t tcps_sc_timed_out; /* # of entries timed out */ 446 u_int64_t tcps_sc_overflowed; /* # dropped due to overflow */ 447 u_int64_t tcps_sc_reset; /* # dropped due to RST */ 448 u_int64_t tcps_sc_unreach; /* # dropped due to ICMP unreach */ 449 u_int64_t tcps_sc_bucketoverflow;/* # dropped due to bucket overflow */ 450 u_int64_t tcps_sc_aborted; /* # of entries aborted (no mem) */ 451 u_int64_t tcps_sc_dupesyn; /* # of duplicate SYNs received */ 452 u_int64_t tcps_sc_dropped; /* # of SYNs dropped (no route/mem) */ 453 u_int64_t tcps_sc_collisions; /* # of hash collisions */ 454 u_int64_t tcps_sc_retransmitted;/* # of retransmissions */ 455 u_int64_t tcps_sc_seedrandom; /* # of syn cache seeds with random */ 456 457 u_int64_t tcps_conndrained; /* # of connections drained */ 458 459 u_int64_t tcps_sack_recovery_episode; /* SACK recovery episodes */ 460 u_int64_t tcps_sack_rexmits; /* SACK rexmit segments */ 461 u_int64_t tcps_sack_rexmit_bytes; /* SACK rexmit bytes */ 462 u_int64_t tcps_sack_rcv_opts; /* SACK options received */ 463 u_int64_t tcps_sack_snd_opts; /* SACK options sent */ 464 }; 465 466 /* 467 * Names for TCP sysctl objects. 468 */ 469 470 #define TCPCTL_RFC1323 1 /* enable/disable RFC1323 timestamps/scaling */ 471 #define TCPCTL_KEEPINITTIME 2 /* TCPT_KEEP value */ 472 #define TCPCTL_KEEPIDLE 3 /* allow tcp_keepidle to be changed */ 473 #define TCPCTL_KEEPINTVL 4 /* allow tcp_keepintvl to be changed */ 474 #define TCPCTL_SLOWHZ 5 /* return kernel idea of PR_SLOWHZ */ 475 #define TCPCTL_BADDYNAMIC 6 /* return bad dynamic port bitmap */ 476 #define TCPCTL_RECVSPACE 7 /* receive buffer space */ 477 #define TCPCTL_SENDSPACE 8 /* send buffer space */ 478 #define TCPCTL_IDENT 9 /* get connection owner */ 479 #define TCPCTL_SACK 10 /* selective acknowledgement, rfc 2018 */ 480 #define TCPCTL_MSSDFLT 11 /* Default maximum segment size */ 481 #define TCPCTL_RSTPPSLIMIT 12 /* RST pps limit */ 482 #define TCPCTL_ACK_ON_PUSH 13 /* ACK immediately on PUSH */ 483 #define TCPCTL_ECN 14 /* RFC3168 ECN */ 484 #define TCPCTL_SYN_CACHE_LIMIT 15 /* max size of comp. state engine */ 485 #define TCPCTL_SYN_BUCKET_LIMIT 16 /* max size of hash bucket */ 486 #define TCPCTL_RFC3390 17 /* enable/disable RFC3390 increased cwnd */ 487 #define TCPCTL_REASS_LIMIT 18 /* max entries for tcp reass queues */ 488 #define TCPCTL_DROP 19 /* drop tcp connection */ 489 #define TCPCTL_SACKHOLE_LIMIT 20 /* max entries for tcp sack queues */ 490 #define TCPCTL_STATS 21 /* TCP statistics */ 491 #define TCPCTL_ALWAYS_KEEPALIVE 22 /* assume SO_KEEPALIVE is always set */ 492 #define TCPCTL_SYN_USE_LIMIT 23 /* number of uses before reseeding hash */ 493 #define TCPCTL_ROOTONLY 24 /* return root only port bitmap */ 494 #define TCPCTL_MAXID 25 495 496 #define TCPCTL_NAMES { \ 497 { 0, 0 }, \ 498 { "rfc1323", CTLTYPE_INT }, \ 499 { "keepinittime", CTLTYPE_INT }, \ 500 { "keepidle", CTLTYPE_INT }, \ 501 { "keepintvl", CTLTYPE_INT }, \ 502 { "slowhz", CTLTYPE_INT }, \ 503 { "baddynamic", CTLTYPE_STRUCT }, \ 504 { NULL, 0 }, \ 505 { NULL, 0 }, \ 506 { "ident", CTLTYPE_STRUCT }, \ 507 { "sack", CTLTYPE_INT }, \ 508 { "mssdflt", CTLTYPE_INT }, \ 509 { "rstppslimit", CTLTYPE_INT }, \ 510 { "ackonpush", CTLTYPE_INT }, \ 511 { "ecn", CTLTYPE_INT }, \ 512 { "syncachelimit", CTLTYPE_INT }, \ 513 { "synbucketlimit", CTLTYPE_INT }, \ 514 { "rfc3390", CTLTYPE_INT }, \ 515 { "reasslimit", CTLTYPE_INT }, \ 516 { "drop", CTLTYPE_STRUCT }, \ 517 { "sackholelimit", CTLTYPE_INT }, \ 518 { "stats", CTLTYPE_STRUCT }, \ 519 { "always_keepalive", CTLTYPE_INT }, \ 520 { "synuselimit", CTLTYPE_INT }, \ 521 { "rootonly", CTLTYPE_STRUCT }, \ 522 } 523 524 #define TCPCTL_VARS { \ 525 NULL, \ 526 &tcp_do_rfc1323, \ 527 &tcptv_keep_init, \ 528 &tcp_keepidle, \ 529 &tcp_keepintvl, \ 530 NULL, \ 531 NULL, \ 532 NULL, \ 533 NULL, \ 534 NULL, \ 535 NULL, \ 536 &tcp_mssdflt, \ 537 &tcp_rst_ppslim, \ 538 &tcp_ack_on_push, \ 539 NULL, \ 540 &tcp_syn_cache_limit, \ 541 &tcp_syn_bucket_limit, \ 542 &tcp_do_rfc3390, \ 543 NULL, \ 544 NULL, \ 545 NULL, \ 546 NULL, \ 547 NULL, \ 548 NULL, \ 549 NULL \ 550 } 551 552 struct tcp_ident_mapping { 553 struct sockaddr_storage faddr, laddr; 554 int euid, ruid; 555 u_int rdomain; 556 }; 557 558 #ifdef _KERNEL 559 extern struct inpcbtable tcbtable; /* head of queue of active tcpcb's */ 560 extern struct tcpstat tcpstat; /* tcp statistics */ 561 extern u_int32_t tcp_now; /* for RFC 1323 timestamps */ 562 extern int tcp_do_rfc1323; /* enabled/disabled? */ 563 extern int tcptv_keep_init; /* time to keep alive the initial SYN packet */ 564 extern int tcp_mssdflt; /* default maximum segment size */ 565 extern int tcp_rst_ppslim; /* maximum outgoing RST packet per second */ 566 extern int tcp_ack_on_push; /* ACK immediately on PUSH */ 567 #ifdef TCP_SACK 568 extern int tcp_do_sack; /* SACK enabled/disabled */ 569 extern struct pool sackhl_pool; 570 extern int tcp_sackhole_limit; /* max entries for tcp sack queues */ 571 #endif 572 extern int tcp_do_ecn; /* RFC3168 ECN enabled/disabled? */ 573 extern int tcp_do_rfc3390; /* RFC3390 Increasing TCP's Initial Window */ 574 575 extern struct pool tcpqe_pool; 576 extern int tcp_reass_limit; /* max entries for tcp reass queues */ 577 578 extern int tcp_syn_cache_limit; /* max entries for compressed state engine */ 579 extern int tcp_syn_bucket_limit;/* max entries per hash bucket */ 580 extern int tcp_syn_use_limit; /* number of uses before reseeding hash */ 581 extern struct syn_cache_set tcp_syn_cache[]; 582 583 int tcp_attach(struct socket *); 584 void tcp_canceltimers(struct tcpcb *); 585 struct tcpcb * 586 tcp_close(struct tcpcb *); 587 void tcp_reaper(void *); 588 int tcp_freeq(struct tcpcb *); 589 #ifdef INET6 590 void tcp6_ctlinput(int, struct sockaddr *, u_int, void *); 591 #endif 592 void *tcp_ctlinput(int, struct sockaddr *, u_int, void *); 593 int tcp_ctloutput(int, struct socket *, int, int, struct mbuf **); 594 struct tcpcb * 595 tcp_disconnect(struct tcpcb *); 596 struct tcpcb * 597 tcp_drop(struct tcpcb *, int); 598 int tcp_dooptions(struct tcpcb *, u_char *, int, struct tcphdr *, 599 struct mbuf *, int, struct tcp_opt_info *, u_int); 600 void tcp_init(void); 601 #ifdef INET6 602 int tcp6_input(struct mbuf **, int *, int); 603 #endif 604 void tcp_input(struct mbuf *, ...); 605 int tcp_mss(struct tcpcb *, int); 606 void tcp_mss_update(struct tcpcb *); 607 u_int tcp_hdrsz(struct tcpcb *); 608 void tcp_mtudisc(struct inpcb *, int); 609 void tcp_mtudisc_increase(struct inpcb *, int); 610 #ifdef INET6 611 void tcp6_mtudisc(struct inpcb *, int); 612 void tcp6_mtudisc_callback(struct sockaddr_in6 *, u_int); 613 #endif 614 struct tcpcb * 615 tcp_newtcpcb(struct inpcb *); 616 void tcp_notify(struct inpcb *, int); 617 int tcp_output(struct tcpcb *); 618 void tcp_pulloutofband(struct socket *, u_int, struct mbuf *, int); 619 int tcp_reass(struct tcpcb *, struct tcphdr *, struct mbuf *, int *); 620 void tcp_rscale(struct tcpcb *, u_long); 621 void tcp_respond(struct tcpcb *, caddr_t, struct tcphdr *, tcp_seq, 622 tcp_seq, int, u_int); 623 void tcp_setpersist(struct tcpcb *); 624 void tcp_update_sndspace(struct tcpcb *); 625 void tcp_update_rcvspace(struct tcpcb *); 626 void tcp_slowtimo(void); 627 struct mbuf * 628 tcp_template(struct tcpcb *); 629 void tcp_trace(short, short, struct tcpcb *, caddr_t, int, int); 630 struct tcpcb * 631 tcp_usrclosed(struct tcpcb *); 632 int tcp_sysctl(int *, u_int, void *, size_t *, void *, size_t); 633 int tcp_usrreq(struct socket *, 634 int, struct mbuf *, struct mbuf *, struct mbuf *, struct proc *); 635 void tcp_xmit_timer(struct tcpcb *, int); 636 void tcpdropoldhalfopen(struct tcpcb *, u_int16_t); 637 #ifdef TCP_SACK 638 void tcp_sack_option(struct tcpcb *,struct tcphdr *,u_char *,int); 639 void tcp_update_sack_list(struct tcpcb *tp, tcp_seq, tcp_seq); 640 void tcp_del_sackholes(struct tcpcb *, struct tcphdr *); 641 void tcp_clean_sackreport(struct tcpcb *tp); 642 void tcp_sack_adjust(struct tcpcb *tp); 643 struct sackhole * 644 tcp_sack_output(struct tcpcb *tp); 645 int tcp_sack_partialack(struct tcpcb *, struct tcphdr *); 646 #ifdef DEBUG 647 void tcp_print_holes(struct tcpcb *tp); 648 #endif 649 #endif /* TCP_SACK */ 650 #if defined(TCP_SACK) 651 int tcp_newreno(struct tcpcb *, struct tcphdr *); 652 u_long tcp_seq_subtract(u_long, u_long ); 653 #endif /* TCP_SACK */ 654 #ifdef TCP_SIGNATURE 655 int tcp_signature_apply(caddr_t, caddr_t, unsigned int); 656 int tcp_signature(struct tdb *, int, struct mbuf *, struct tcphdr *, 657 int, int, char *); 658 #endif /* TCP_SIGNATURE */ 659 void tcp_set_iss_tsm(struct tcpcb *); 660 661 int syn_cache_add(struct sockaddr *, struct sockaddr *, 662 struct tcphdr *, unsigned int, struct socket *, 663 struct mbuf *, u_char *, int, struct tcp_opt_info *, tcp_seq *); 664 void syn_cache_unreach(struct sockaddr *, struct sockaddr *, 665 struct tcphdr *, u_int); 666 struct socket *syn_cache_get(struct sockaddr *, struct sockaddr *, 667 struct tcphdr *, unsigned int, unsigned int, 668 struct socket *so, struct mbuf *); 669 void syn_cache_init(void); 670 void syn_cache_insert(struct syn_cache *, struct tcpcb *); 671 struct syn_cache *syn_cache_lookup(struct sockaddr *, struct sockaddr *, 672 struct syn_cache_head **, u_int); 673 void syn_cache_reset(struct sockaddr *, struct sockaddr *, 674 struct tcphdr *, u_int); 675 int syn_cache_respond(struct syn_cache *, struct mbuf *); 676 void syn_cache_timer(void *); 677 void syn_cache_cleanup(struct tcpcb *); 678 void syn_cache_reaper(void *); 679 680 #endif /* _KERNEL */ 681 #endif /* _NETINET_TCP_VAR_H_ */ 682