1 /* $NetBSD: tcp_var.h,v 1.94 2002/11/02 07:28:15 perry Exp $ */ 2 3 /* 4 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. 5 * All rights reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 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 project 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. 18 * 19 * THIS SOFTWARE IS PROVIDED BY THE PROJECT 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 PROJECT 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 29 * SUCH DAMAGE. 30 */ 31 32 /* 33 * @(#)COPYRIGHT 1.1 (NRL) 17 January 1995 34 * 35 * NRL grants permission for redistribution and use in source and binary 36 * forms, with or without modification, of the software and documentation 37 * created at NRL provided that the following conditions are met: 38 * 39 * 1. Redistributions of source code must retain the above copyright 40 * notice, this list of conditions and the following disclaimer. 41 * 2. Redistributions in binary form must reproduce the above copyright 42 * notice, this list of conditions and the following disclaimer in the 43 * documentation and/or other materials provided with the distribution. 44 * 3. All advertising materials mentioning features or use of this software 45 * must display the following acknowledgements: 46 * This product includes software developed by the University of 47 * California, Berkeley and its contributors. 48 * This product includes software developed at the Information 49 * Technology Division, US Naval Research Laboratory. 50 * 4. Neither the name of the NRL nor the names of its contributors 51 * may be used to endorse or promote products derived from this software 52 * without specific prior written permission. 53 * 54 * THE SOFTWARE PROVIDED BY NRL IS PROVIDED BY NRL AND CONTRIBUTORS ``AS 55 * IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 56 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A 57 * PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL NRL OR 58 * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, 59 * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, 60 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR 61 * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF 62 * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING 63 * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS 64 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 65 * 66 * The views and conclusions contained in the software and documentation 67 * are those of the authors and should not be interpreted as representing 68 * official policies, either expressed or implied, of the US Naval 69 * Research Laboratory (NRL). 70 */ 71 72 /*- 73 * Copyright (c) 1997, 1998, 1999, 2001 The NetBSD Foundation, Inc. 74 * All rights reserved. 75 * 76 * This code is derived from software contributed to The NetBSD Foundation 77 * by Jason R. Thorpe of the Numerical Aerospace Simulation Facility, 78 * NASA Ames Research Center. 79 * 80 * Redistribution and use in source and binary forms, with or without 81 * modification, are permitted provided that the following conditions 82 * are met: 83 * 1. Redistributions of source code must retain the above copyright 84 * notice, this list of conditions and the following disclaimer. 85 * 2. Redistributions in binary form must reproduce the above copyright 86 * notice, this list of conditions and the following disclaimer in the 87 * documentation and/or other materials provided with the distribution. 88 * 3. All advertising materials mentioning features or use of this software 89 * must display the following acknowledgement: 90 * This product includes software developed by the NetBSD 91 * Foundation, Inc. and its contributors. 92 * 4. Neither the name of The NetBSD Foundation nor the names of its 93 * contributors may be used to endorse or promote products derived 94 * from this software without specific prior written permission. 95 * 96 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 97 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 98 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 99 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 100 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 101 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 102 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 103 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 104 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 105 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 106 * POSSIBILITY OF SUCH DAMAGE. 107 */ 108 109 /* 110 * Copyright (c) 1982, 1986, 1993, 1994, 1995 111 * The Regents of the University of California. All rights reserved. 112 * 113 * Redistribution and use in source and binary forms, with or without 114 * modification, are permitted provided that the following conditions 115 * are met: 116 * 1. Redistributions of source code must retain the above copyright 117 * notice, this list of conditions and the following disclaimer. 118 * 2. Redistributions in binary form must reproduce the above copyright 119 * notice, this list of conditions and the following disclaimer in the 120 * documentation and/or other materials provided with the distribution. 121 * 3. All advertising materials mentioning features or use of this software 122 * must display the following acknowledgement: 123 * This product includes software developed by the University of 124 * California, Berkeley and its contributors. 125 * 4. Neither the name of the University nor the names of its contributors 126 * may be used to endorse or promote products derived from this software 127 * without specific prior written permission. 128 * 129 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 130 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 131 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 132 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 133 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 134 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 135 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 136 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 137 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 138 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 139 * SUCH DAMAGE. 140 * 141 * @(#)tcp_var.h 8.4 (Berkeley) 5/24/95 142 */ 143 144 #ifndef _NETINET_TCP_VAR_H_ 145 #define _NETINET_TCP_VAR_H_ 146 147 #if defined(_KERNEL_OPT) 148 #include "opt_inet.h" 149 #endif 150 151 /* 152 * Kernel variables for tcp. 153 */ 154 155 #include <sys/callout.h> 156 157 /* 158 * Tcp control block, one per tcp; fields: 159 */ 160 struct tcpcb { 161 int t_family; /* address family on the wire */ 162 struct ipqehead segq; /* sequencing queue */ 163 struct callout t_timer[TCPT_NTIMERS];/* tcp timers */ 164 short t_state; /* state of this connection */ 165 short t_rxtshift; /* log(2) of rexmt exp. backoff */ 166 uint32_t t_rxtcur; /* current retransmit value */ 167 short t_dupacks; /* consecutive dup acks recd */ 168 u_short t_peermss; /* peer's maximum segment size */ 169 u_short t_ourmss; /* our's maximum segment size */ 170 u_short t_segsz; /* current segment size in use */ 171 char t_force; /* 1 if forcing out a byte */ 172 u_int t_flags; 173 #define TF_ACKNOW 0x0001 /* ack peer immediately */ 174 #define TF_DELACK 0x0002 /* ack, but try to delay it */ 175 #define TF_NODELAY 0x0004 /* don't delay packets to coalesce */ 176 #define TF_NOOPT 0x0008 /* don't use tcp options */ 177 #define TF_REQ_SCALE 0x0020 /* have/will request window scaling */ 178 #define TF_RCVD_SCALE 0x0040 /* other side has requested scaling */ 179 #define TF_REQ_TSTMP 0x0080 /* have/will request timestamps */ 180 #define TF_RCVD_TSTMP 0x0100 /* a timestamp was received in SYN */ 181 #define TF_SACK_PERMIT 0x0200 /* other side said I could SACK */ 182 #define TF_SYN_REXMT 0x0400 /* rexmit timer fired on SYN */ 183 #define TF_WILL_SACK 0x0800 /* try to use SACK */ 184 #define TF_CANT_TXSACK 0x1000 /* other side said I could not SACK */ 185 #define TF_IGNR_RXSACK 0x2000 /* ignore received SACK blocks */ 186 #define TF_REASSEMBLING 0x4000 /* we're busy reassembling */ 187 188 189 struct mbuf *t_template; /* skeletal packet for transmit */ 190 struct inpcb *t_inpcb; /* back pointer to internet pcb */ 191 struct in6pcb *t_in6pcb; /* back pointer to internet pcb */ 192 struct callout t_delack_ch; /* delayed ACK callout */ 193 /* 194 * The following fields are used as in the protocol specification. 195 * See RFC783, Dec. 1981, page 21. 196 */ 197 /* send sequence variables */ 198 tcp_seq snd_una; /* send unacknowledged */ 199 tcp_seq snd_nxt; /* send next */ 200 tcp_seq snd_up; /* send urgent pointer */ 201 tcp_seq snd_wl1; /* window update seg seq number */ 202 tcp_seq snd_wl2; /* window update seg ack number */ 203 tcp_seq iss; /* initial send sequence number */ 204 u_long snd_wnd; /* send window */ 205 tcp_seq snd_recover; /* for use in fast recovery */ 206 /* receive sequence variables */ 207 u_long rcv_wnd; /* receive window */ 208 tcp_seq rcv_nxt; /* receive next */ 209 tcp_seq rcv_up; /* receive urgent pointer */ 210 tcp_seq irs; /* initial receive sequence number */ 211 /* 212 * Additional variables for this implementation. 213 */ 214 /* receive variables */ 215 tcp_seq rcv_adv; /* advertised window */ 216 /* retransmit variables */ 217 tcp_seq snd_max; /* highest sequence number sent; 218 * used to recognize retransmits 219 */ 220 /* congestion control (for slow start, source quench, retransmit after loss) */ 221 u_long snd_cwnd; /* congestion-controlled window */ 222 u_long snd_ssthresh; /* snd_cwnd size threshhold for 223 * for slow start exponential to 224 * linear switch 225 */ 226 /* 227 * transmit timing stuff. See below for scale of srtt and rttvar. 228 * "Variance" is actually smoothed difference. 229 */ 230 uint32_t t_rcvtime; /* time last segment received */ 231 uint32_t t_rtttime; /* time we started measuring rtt */ 232 tcp_seq t_rtseq; /* sequence number being timed */ 233 int32_t t_srtt; /* smoothed round-trip time */ 234 int32_t t_rttvar; /* variance in round-trip time */ 235 uint32_t t_rttmin; /* minimum rtt allowed */ 236 u_long max_sndwnd; /* largest window peer has offered */ 237 238 /* out-of-band data */ 239 char t_oobflags; /* have some */ 240 char t_iobc; /* input character */ 241 #define TCPOOB_HAVEDATA 0x01 242 #define TCPOOB_HADDATA 0x02 243 short t_softerror; /* possible error not yet reported */ 244 245 /* RFC 1323 variables */ 246 u_char snd_scale; /* window scaling for send window */ 247 u_char rcv_scale; /* window scaling for recv window */ 248 u_char request_r_scale; /* pending window scaling */ 249 u_char requested_s_scale; 250 u_int32_t ts_recent; /* timestamp echo data */ 251 u_int32_t ts_recent_age; /* when last updated */ 252 u_int32_t ts_timebase; /* our timebase */ 253 tcp_seq last_ack_sent; 254 255 /* SACK stuff */ 256 struct ipqehead timeq; /* time sequenced queue (for SACK) */ 257 258 /* path MTU discovery blackhole detection */ 259 int t_mtudisc; /* perform mtudisc for this tcb */ 260 261 /* pointer for syn cache entries*/ 262 LIST_HEAD(, syn_cache) t_sc; /* list of entries by this tcb */ 263 }; 264 265 #ifdef _KERNEL 266 /* 267 * TCP reassembly queue locks. 268 */ 269 static __inline int tcp_reass_lock_try __P((struct tcpcb *)) 270 __attribute__((__unused__)); 271 static __inline void tcp_reass_unlock __P((struct tcpcb *)) 272 __attribute__((__unused__)); 273 274 static __inline int 275 tcp_reass_lock_try(tp) 276 struct tcpcb *tp; 277 { 278 int s; 279 280 /* 281 * Use splvm() -- we're blocking things that would cause 282 * mbuf allocation. 283 */ 284 s = splvm(); 285 if (tp->t_flags & TF_REASSEMBLING) { 286 splx(s); 287 return (0); 288 } 289 tp->t_flags |= TF_REASSEMBLING; 290 splx(s); 291 return (1); 292 } 293 294 static __inline void 295 tcp_reass_unlock(tp) 296 struct tcpcb *tp; 297 { 298 int s; 299 300 s = splvm(); 301 tp->t_flags &= ~TF_REASSEMBLING; 302 splx(s); 303 } 304 305 #ifdef DIAGNOSTIC 306 #define TCP_REASS_LOCK(tp) \ 307 do { \ 308 if (tcp_reass_lock_try(tp) == 0) { \ 309 printf("%s:%d: tcpcb %p reass already locked\n", \ 310 __FILE__, __LINE__, tp); \ 311 panic("tcp_reass_lock"); \ 312 } \ 313 } while (/*CONSTCOND*/ 0) 314 #define TCP_REASS_LOCK_CHECK(tp) \ 315 do { \ 316 if (((tp)->t_flags & TF_REASSEMBLING) == 0) { \ 317 printf("%s:%d: tcpcb %p reass lock not held\n", \ 318 __FILE__, __LINE__, tp); \ 319 panic("tcp reass lock check"); \ 320 } \ 321 } while (/*CONSTCOND*/ 0) 322 #else 323 #define TCP_REASS_LOCK(tp) (void) tcp_reass_lock_try((tp)) 324 #define TCP_REASS_LOCK_CHECK(tp) /* nothing */ 325 #endif 326 327 #define TCP_REASS_UNLOCK(tp) tcp_reass_unlock((tp)) 328 #endif /* _KERNEL */ 329 330 /* 331 * Queue for delayed ACK processing. 332 */ 333 #ifdef _KERNEL 334 extern int tcp_delack_ticks; 335 void tcp_delack(void *); 336 337 #define TCP_RESTART_DELACK(tp) \ 338 callout_reset(&(tp)->t_delack_ch, tcp_delack_ticks, \ 339 tcp_delack, tp) 340 341 #define TCP_SET_DELACK(tp) \ 342 do { \ 343 if (((tp)->t_flags & TF_DELACK) == 0) { \ 344 (tp)->t_flags |= TF_DELACK; \ 345 TCP_RESTART_DELACK(tp); \ 346 } \ 347 } while (/*CONSTCOND*/0) 348 349 #define TCP_CLEAR_DELACK(tp) \ 350 do { \ 351 if ((tp)->t_flags & TF_DELACK) { \ 352 (tp)->t_flags &= ~TF_DELACK; \ 353 callout_stop(&(tp)->t_delack_ch); \ 354 } \ 355 } while (/*CONSTCOND*/0) 356 #endif /* _KERNEL */ 357 358 /* 359 * Compute the current timestamp for a connection. 360 */ 361 #define TCP_TIMESTAMP(tp) (tcp_now - (tp)->ts_timebase) 362 363 /* 364 * Handy way of passing around TCP option info. 365 */ 366 struct tcp_opt_info { 367 int ts_present; 368 u_int32_t ts_val; 369 u_int32_t ts_ecr; 370 u_int16_t maxseg; 371 }; 372 373 /* 374 * Data for the TCP compressed state engine. 375 */ 376 union syn_cache_sa { 377 struct sockaddr sa; 378 struct sockaddr_in sin; 379 #if 1 /*def INET6*/ 380 struct sockaddr_in6 sin6; 381 #endif 382 }; 383 384 struct syn_cache { 385 TAILQ_ENTRY(syn_cache) sc_bucketq; /* link on bucket list */ 386 struct callout sc_timer; /* rexmt timer */ 387 union { /* cached route */ 388 struct route route4; 389 #ifdef INET6 390 struct route_in6 route6; 391 #endif 392 } sc_route_u; 393 #define sc_route4 sc_route_u.route4 394 #ifdef INET6 395 #define sc_route6 sc_route_u.route6 396 #endif 397 long sc_win; /* advertised window */ 398 int sc_bucketidx; /* our bucket index */ 399 u_int32_t sc_hash; 400 u_int32_t sc_timestamp; /* timestamp from SYN */ 401 u_int32_t sc_timebase; /* our local timebase */ 402 union syn_cache_sa sc_src; 403 union syn_cache_sa sc_dst; 404 tcp_seq sc_irs; 405 tcp_seq sc_iss; 406 u_int sc_rxtcur; /* current rxt timeout */ 407 u_int sc_rxttot; /* total time spend on queues */ 408 u_short sc_rxtshift; /* for computing backoff */ 409 u_short sc_flags; 410 411 #define SCF_UNREACH 0x0001 /* we've had an unreach error */ 412 #define SCF_TIMESTAMP 0x0002 /* peer will do timestamps */ 413 414 struct mbuf *sc_ipopts; /* IP options */ 415 u_int16_t sc_peermaxseg; 416 u_int16_t sc_ourmaxseg; 417 u_int8_t sc_request_r_scale : 4, 418 sc_requested_s_scale : 4; 419 420 struct tcpcb *sc_tp; /* tcb for listening socket */ 421 LIST_ENTRY(syn_cache) sc_tpq; /* list of entries by same tp */ 422 }; 423 424 struct syn_cache_head { 425 TAILQ_HEAD(, syn_cache) sch_bucket; /* bucket entries */ 426 u_short sch_length; /* # entries in bucket */ 427 }; 428 429 #define intotcpcb(ip) ((struct tcpcb *)(ip)->inp_ppcb) 430 #ifdef INET6 431 #define in6totcpcb(ip) ((struct tcpcb *)(ip)->in6p_ppcb) 432 #endif 433 #ifndef INET6 434 #define sototcpcb(so) (intotcpcb(sotoinpcb(so))) 435 #else 436 #define sototcpcb(so) (((so)->so_proto->pr_domain->dom_family == AF_INET) \ 437 ? intotcpcb(sotoinpcb(so)) \ 438 : in6totcpcb(sotoin6pcb(so))) 439 #endif 440 441 /* 442 * The smoothed round-trip time and estimated variance 443 * are stored as fixed point numbers scaled by the values below. 444 * For convenience, these scales are also used in smoothing the average 445 * (smoothed = (1/scale)sample + ((scale-1)/scale)smoothed). 446 * With these scales, srtt has 3 bits to the right of the binary point, 447 * and thus an "ALPHA" of 0.875. rttvar has 2 bits to the right of the 448 * binary point, and is smoothed with an ALPHA of 0.75. 449 */ 450 #define TCP_RTT_SHIFT 3 /* shift for srtt; 3 bits frac. */ 451 #define TCP_RTTVAR_SHIFT 2 /* multiplier for rttvar; 2 bits */ 452 453 /* 454 * The initial retransmission should happen at rtt + 4 * rttvar. 455 * Because of the way we do the smoothing, srtt and rttvar 456 * will each average +1/2 tick of bias. When we compute 457 * the retransmit timer, we want 1/2 tick of rounding and 458 * 1 extra tick because of +-1/2 tick uncertainty in the 459 * firing of the timer. The bias will give us exactly the 460 * 1.5 tick we need. But, because the bias is 461 * statistical, we have to test that we don't drop below 462 * the minimum feasible timer (which is 2 ticks). 463 * This macro assumes that the value of 1<<TCP_RTTVAR_SHIFT 464 * is the same as the multiplier for rttvar. 465 */ 466 #define TCP_REXMTVAL(tp) \ 467 ((((tp)->t_srtt >> TCP_RTT_SHIFT) + (tp)->t_rttvar) >> 2) 468 469 /* 470 * Compute the initial window for slow start. 471 */ 472 #define TCP_INITIAL_WINDOW(iw, segsz) \ 473 (((iw) == 0) ? (min(4 * (segsz), max(2 * (segsz), 4380))) : \ 474 ((segsz) * (iw))) 475 476 /* 477 * TCP statistics. 478 * Many of these should be kept per connection, 479 * but that's inconvenient at the moment. 480 */ 481 struct tcpstat { 482 u_quad_t tcps_connattempt; /* connections initiated */ 483 u_quad_t tcps_accepts; /* connections accepted */ 484 u_quad_t tcps_connects; /* connections established */ 485 u_quad_t tcps_drops; /* connections dropped */ 486 u_quad_t tcps_conndrops; /* embryonic connections dropped */ 487 u_quad_t tcps_closed; /* conn. closed (includes drops) */ 488 u_quad_t tcps_segstimed; /* segs where we tried to get rtt */ 489 u_quad_t tcps_rttupdated; /* times we succeeded */ 490 u_quad_t tcps_delack; /* delayed acks sent */ 491 u_quad_t tcps_timeoutdrop; /* conn. dropped in rxmt timeout */ 492 u_quad_t tcps_rexmttimeo; /* retransmit timeouts */ 493 u_quad_t tcps_persisttimeo; /* persist timeouts */ 494 u_quad_t tcps_keeptimeo; /* keepalive timeouts */ 495 u_quad_t tcps_keepprobe; /* keepalive probes sent */ 496 u_quad_t tcps_keepdrops; /* connections dropped in keepalive */ 497 u_quad_t tcps_persistdrops; /* connections dropped in persist */ 498 u_quad_t tcps_connsdrained; /* connections drained due to memory 499 shortage */ 500 u_quad_t tcps_pmtublackhole; /* PMTUD blackhole detected */ 501 502 u_quad_t tcps_sndtotal; /* total packets sent */ 503 u_quad_t tcps_sndpack; /* data packets sent */ 504 u_quad_t tcps_sndbyte; /* data bytes sent */ 505 u_quad_t tcps_sndrexmitpack; /* data packets retransmitted */ 506 u_quad_t tcps_sndrexmitbyte; /* data bytes retransmitted */ 507 u_quad_t tcps_sndacks; /* ack-only packets sent */ 508 u_quad_t tcps_sndprobe; /* window probes sent */ 509 u_quad_t tcps_sndurg; /* packets sent with URG only */ 510 u_quad_t tcps_sndwinup; /* window update-only packets sent */ 511 u_quad_t tcps_sndctrl; /* control (SYN|FIN|RST) packets sent */ 512 513 u_quad_t tcps_rcvtotal; /* total packets received */ 514 u_quad_t tcps_rcvpack; /* packets received in sequence */ 515 u_quad_t tcps_rcvbyte; /* bytes received in sequence */ 516 u_quad_t tcps_rcvbadsum; /* packets received with ccksum errs */ 517 u_quad_t tcps_rcvbadoff; /* packets received with bad offset */ 518 u_quad_t tcps_rcvmemdrop; /* packets dropped for lack of memory */ 519 u_quad_t tcps_rcvshort; /* packets received too short */ 520 u_quad_t tcps_rcvduppack; /* duplicate-only packets received */ 521 u_quad_t tcps_rcvdupbyte; /* duplicate-only bytes received */ 522 u_quad_t tcps_rcvpartduppack; /* packets with some duplicate data */ 523 u_quad_t tcps_rcvpartdupbyte; /* dup. bytes in part-dup. packets */ 524 u_quad_t tcps_rcvoopack; /* out-of-order packets received */ 525 u_quad_t tcps_rcvoobyte; /* out-of-order bytes received */ 526 u_quad_t tcps_rcvpackafterwin; /* packets with data after window */ 527 u_quad_t tcps_rcvbyteafterwin; /* bytes rcvd after window */ 528 u_quad_t tcps_rcvafterclose; /* packets rcvd after "close" */ 529 u_quad_t tcps_rcvwinprobe; /* rcvd window probe packets */ 530 u_quad_t tcps_rcvdupack; /* rcvd duplicate acks */ 531 u_quad_t tcps_rcvacktoomuch; /* rcvd acks for unsent data */ 532 u_quad_t tcps_rcvackpack; /* rcvd ack packets */ 533 u_quad_t tcps_rcvackbyte; /* bytes acked by rcvd acks */ 534 u_quad_t tcps_rcvwinupd; /* rcvd window update packets */ 535 u_quad_t tcps_pawsdrop; /* segments dropped due to PAWS */ 536 u_quad_t tcps_predack; /* times hdr predict ok for acks */ 537 u_quad_t tcps_preddat; /* times hdr predict ok for data pkts */ 538 539 u_quad_t tcps_pcbhashmiss; /* input packets missing pcb hash */ 540 u_quad_t tcps_noport; /* no socket on port */ 541 u_quad_t tcps_badsyn; /* received ack for which we have 542 no SYN in compressed state */ 543 544 /* These statistics deal with the SYN cache. */ 545 u_quad_t tcps_sc_added; /* # of entries added */ 546 u_quad_t tcps_sc_completed; /* # of connections completed */ 547 u_quad_t tcps_sc_timed_out; /* # of entries timed out */ 548 u_quad_t tcps_sc_overflowed; /* # dropped due to overflow */ 549 u_quad_t tcps_sc_reset; /* # dropped due to RST */ 550 u_quad_t tcps_sc_unreach; /* # dropped due to ICMP unreach */ 551 u_quad_t tcps_sc_bucketoverflow;/* # dropped due to bucket overflow */ 552 u_quad_t tcps_sc_aborted; /* # of entries aborted (no mem) */ 553 u_quad_t tcps_sc_dupesyn; /* # of duplicate SYNs received */ 554 u_quad_t tcps_sc_dropped; /* # of SYNs dropped (no route/mem) */ 555 u_quad_t tcps_sc_collisions; /* # of hash collisions */ 556 u_quad_t tcps_sc_retransmitted; /* # of retransmissions */ 557 558 u_quad_t tcps_selfquench; /* # of ENOBUFS we get on output */ 559 }; 560 561 /* 562 * Names for TCP sysctl objects. 563 */ 564 #define TCPCTL_RFC1323 1 /* RFC1323 timestamps/scaling */ 565 #define TCPCTL_SENDSPACE 2 /* default send buffer */ 566 #define TCPCTL_RECVSPACE 3 /* default recv buffer */ 567 #define TCPCTL_MSSDFLT 4 /* default seg size */ 568 #define TCPCTL_SYN_CACHE_LIMIT 5 /* max size of comp. state engine */ 569 #define TCPCTL_SYN_BUCKET_LIMIT 6 /* max size of hash bucket */ 570 #if 0 /*obsoleted*/ 571 #define TCPCTL_SYN_CACHE_INTER 7 /* interval of comp. state timer */ 572 #endif 573 #define TCPCTL_INIT_WIN 8 /* initial window */ 574 #define TCPCTL_MSS_IFMTU 9 /* mss from interface, not in_maxmtu */ 575 #define TCPCTL_SACK 10 /* RFC2018 selective acknowledgement */ 576 #define TCPCTL_WSCALE 11 /* RFC1323 window scaling */ 577 #define TCPCTL_TSTAMP 12 /* RFC1323 timestamps */ 578 #define TCPCTL_COMPAT_42 13 /* 4.2BSD TCP bug work-arounds */ 579 #define TCPCTL_CWM 14 /* Congestion Window Monitoring */ 580 #define TCPCTL_CWM_BURSTSIZE 15 /* burst size allowed by CWM */ 581 #define TCPCTL_ACK_ON_PUSH 16 /* ACK immediately on PUSH */ 582 #define TCPCTL_KEEPIDLE 17 /* keepalive idle time */ 583 #define TCPCTL_KEEPINTVL 18 /* keepalive probe interval */ 584 #define TCPCTL_KEEPCNT 19 /* keepalive count */ 585 #define TCPCTL_SLOWHZ 20 /* PR_SLOWHZ (read-only) */ 586 #define TCPCTL_NEWRENO 21 /* NewReno Congestion Control */ 587 #define TCPCTL_LOG_REFUSED 22 /* Log refused connections */ 588 #if 0 /*obsoleted*/ 589 #define TCPCTL_RSTRATELIMIT 23 /* RST rate limit */ 590 #endif 591 #define TCPCTL_RSTPPSLIMIT 24 /* RST pps limit */ 592 #define TCPCTL_DELACK_TICKS 25 /* # ticks to delay ACK */ 593 #define TCPCTL_MAXID 26 594 595 #define TCPCTL_NAMES { \ 596 { 0, 0 }, \ 597 { "rfc1323", CTLTYPE_INT }, \ 598 { "sendspace", CTLTYPE_INT }, \ 599 { "recvspace", CTLTYPE_INT }, \ 600 { "mssdflt", CTLTYPE_INT }, \ 601 { "syn_cache_limit", CTLTYPE_INT }, \ 602 { "syn_bucket_limit", CTLTYPE_INT }, \ 603 { 0, 0 },\ 604 { "init_win", CTLTYPE_INT }, \ 605 { "mss_ifmtu", CTLTYPE_INT }, \ 606 { "sack", CTLTYPE_INT }, \ 607 { "win_scale", CTLTYPE_INT }, \ 608 { "timestamps", CTLTYPE_INT }, \ 609 { "compat_42", CTLTYPE_INT }, \ 610 { "cwm", CTLTYPE_INT }, \ 611 { "cwm_burstsize", CTLTYPE_INT }, \ 612 { "ack_on_push", CTLTYPE_INT }, \ 613 { "keepidle", CTLTYPE_INT }, \ 614 { "keepintvl", CTLTYPE_INT }, \ 615 { "keepcnt", CTLTYPE_INT }, \ 616 { "slowhz", CTLTYPE_INT }, \ 617 { "newreno", CTLTYPE_INT }, \ 618 { "log_refused",CTLTYPE_INT }, \ 619 { 0, 0 }, \ 620 { "rstppslimit", CTLTYPE_INT }, \ 621 { "delack_ticks", CTLTYPE_INT }, \ 622 } 623 624 #ifdef _KERNEL 625 extern struct inpcbtable tcbtable; /* head of queue of active tcpcb's */ 626 #ifdef INET6 627 extern struct in6pcb tcb6; 628 #endif 629 extern struct tcpstat tcpstat; /* tcp statistics */ 630 extern u_int32_t tcp_now; /* for RFC 1323 timestamps */ 631 extern int tcp_do_rfc1323; /* enabled/disabled? */ 632 extern int tcp_do_sack; /* SACK enabled/disabled? */ 633 extern int tcp_do_win_scale; /* RFC1323 window scaling enabled/disabled? */ 634 extern int tcp_do_timestamps; /* RFC1323 timestamps enabled/disabled? */ 635 extern int tcp_do_newreno; /* Use the New Reno algorithms */ 636 extern int tcp_mssdflt; /* default seg size */ 637 extern int tcp_init_win; /* initial window */ 638 extern int tcp_mss_ifmtu; /* take MSS from interface, not in_maxmtu */ 639 extern int tcp_compat_42; /* work around ancient broken TCP peers */ 640 extern int tcp_cwm; /* enable Congestion Window Monitoring */ 641 extern int tcp_cwm_burstsize; /* burst size allowed by CWM */ 642 extern int tcp_ack_on_push; /* ACK immediately on PUSH */ 643 extern int tcp_syn_cache_limit; /* max entries for compressed state engine */ 644 extern int tcp_syn_bucket_limit;/* max entries per hash bucket */ 645 extern int tcp_log_refused; /* log refused connections */ 646 647 extern int tcp_rst_ppslim; 648 649 extern int tcp_syn_cache_size; 650 extern struct syn_cache_head tcp_syn_cache[]; 651 extern u_long syn_cache_count; 652 653 #define TCPCTL_VARIABLES { \ 654 { 0 }, \ 655 { 1, 0, &tcp_do_rfc1323 }, \ 656 { 1, 0, &tcp_sendspace }, \ 657 { 1, 0, &tcp_recvspace }, \ 658 { 1, 0, &tcp_mssdflt }, \ 659 { 1, 0, &tcp_syn_cache_limit }, \ 660 { 1, 0, &tcp_syn_bucket_limit }, \ 661 { 0 }, \ 662 { 1, 0, &tcp_init_win }, \ 663 { 1, 0, &tcp_mss_ifmtu }, \ 664 { 1, 0, &tcp_do_sack }, \ 665 { 1, 0, &tcp_do_win_scale }, \ 666 { 1, 0, &tcp_do_timestamps }, \ 667 { 1, 0, &tcp_compat_42 }, \ 668 { 1, 0, &tcp_cwm }, \ 669 { 1, 0, &tcp_cwm_burstsize }, \ 670 { 1, 0, &tcp_ack_on_push }, \ 671 { 1, 0, &tcp_keepidle }, \ 672 { 1, 0, &tcp_keepintvl }, \ 673 { 1, 0, &tcp_keepcnt }, \ 674 { 1, 1, 0, PR_SLOWHZ }, \ 675 { 1, 0, &tcp_do_newreno }, \ 676 { 1, 0, &tcp_log_refused }, \ 677 { 0 }, \ 678 { 1, 0, &tcp_rst_ppslim }, \ 679 { 1, 0, &tcp_delack_ticks }, \ 680 } 681 682 #ifdef __NO_STRICT_ALIGNMENT 683 #define TCP_HDR_ALIGNED_P(th) 1 684 #else 685 #define TCP_HDR_ALIGNED_P(th) ((((vaddr_t) (th)) & 3) == 0) 686 #endif 687 688 int tcp_attach __P((struct socket *)); 689 void tcp_canceltimers __P((struct tcpcb *)); 690 struct tcpcb * 691 tcp_close __P((struct tcpcb *)); 692 #ifdef INET6 693 void tcp6_ctlinput __P((int, struct sockaddr *, void *)); 694 #endif 695 void *tcp_ctlinput __P((int, struct sockaddr *, void *)); 696 int tcp_ctloutput __P((int, struct socket *, int, int, struct mbuf **)); 697 struct tcpcb * 698 tcp_disconnect __P((struct tcpcb *)); 699 struct tcpcb * 700 tcp_drop __P((struct tcpcb *, int)); 701 void tcp_dooptions __P((struct tcpcb *, 702 u_char *, int, struct tcphdr *, struct tcp_opt_info *)); 703 void tcp_drain __P((void)); 704 #ifdef INET6 705 void tcp6_drain __P((void)); 706 #endif 707 void tcp_established __P((struct tcpcb *)); 708 void tcp_init __P((void)); 709 #ifdef INET6 710 int tcp6_input __P((struct mbuf **, int *, int)); 711 #endif 712 void tcp_input __P((struct mbuf *, ...)); 713 u_long tcp_mss_to_advertise __P((const struct ifnet *, int)); 714 void tcp_mss_from_peer __P((struct tcpcb *, int)); 715 struct tcpcb * 716 tcp_newtcpcb __P((int, void *)); 717 void tcp_notify __P((struct inpcb *, int)); 718 #ifdef INET6 719 void tcp6_notify __P((struct in6pcb *, int)); 720 #endif 721 u_int tcp_optlen __P((struct tcpcb *)); 722 int tcp_output __P((struct tcpcb *)); 723 void tcp_pulloutofband __P((struct socket *, 724 struct tcphdr *, struct mbuf *, int)); 725 void tcp_quench __P((struct inpcb *, int)); 726 #ifdef INET6 727 void tcp6_quench __P((struct in6pcb *, int)); 728 #endif 729 int tcp_reass __P((struct tcpcb *, struct tcphdr *, struct mbuf *, int *)); 730 int tcp_respond __P((struct tcpcb *, struct mbuf *, struct mbuf *, 731 struct tcphdr *, tcp_seq, tcp_seq, int)); 732 void tcp_rmx_rtt __P((struct tcpcb *)); 733 void tcp_setpersist __P((struct tcpcb *)); 734 void tcp_slowtimo __P((void)); 735 struct mbuf * 736 tcp_template __P((struct tcpcb *)); 737 void tcp_trace __P((int, int, struct tcpcb *, struct mbuf *, int)); 738 struct tcpcb * 739 tcp_usrclosed __P((struct tcpcb *)); 740 int tcp_sysctl __P((int *, u_int, void *, size_t *, void *, size_t)); 741 int tcp_usrreq __P((struct socket *, 742 int, struct mbuf *, struct mbuf *, struct mbuf *, struct proc *)); 743 void tcp_xmit_timer __P((struct tcpcb *, uint32_t)); 744 tcp_seq tcp_new_iss __P((struct tcpcb *, tcp_seq)); 745 tcp_seq tcp_new_iss1 __P((void *, void *, u_int16_t, u_int16_t, size_t, 746 tcp_seq)); 747 748 int syn_cache_add __P((struct sockaddr *, struct sockaddr *, 749 struct tcphdr *, unsigned int, struct socket *, 750 struct mbuf *, u_char *, int, struct tcp_opt_info *)); 751 void syn_cache_unreach __P((struct sockaddr *, struct sockaddr *, 752 struct tcphdr *)); 753 struct socket *syn_cache_get __P((struct sockaddr *, struct sockaddr *, 754 struct tcphdr *, unsigned int, unsigned int, 755 struct socket *so, struct mbuf *)); 756 void syn_cache_init __P((void)); 757 void syn_cache_insert __P((struct syn_cache *, struct tcpcb *)); 758 struct syn_cache *syn_cache_lookup __P((struct sockaddr *, struct sockaddr *, 759 struct syn_cache_head **)); 760 void syn_cache_reset __P((struct sockaddr *, struct sockaddr *, 761 struct tcphdr *)); 762 int syn_cache_respond __P((struct syn_cache *, struct mbuf *)); 763 void syn_cache_timer __P((void *)); 764 void syn_cache_cleanup __P((struct tcpcb *)); 765 766 int tcp_newreno __P((struct tcpcb *, struct tcphdr *)); 767 #endif 768 769 #endif /* _NETINET_TCP_VAR_H_ */ 770