1 /* 2 * Copyright (c) 2003, 2004 Jeffrey M. Hsu. All rights reserved. 3 * Copyright (c) 2003, 2004 The DragonFly Project. All rights reserved. 4 * 5 * This code is derived from software contributed to The DragonFly Project 6 * by Jeffrey M. Hsu. 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 DragonFly Project nor the names of its 17 * contributors may be used to endorse or promote products derived 18 * from this software without specific, prior written permission. 19 * 20 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS 21 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT 22 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS 23 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE 24 * COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, 25 * INCIDENTAL, SPECIAL, EXEMPLARY OR CONSEQUENTIAL DAMAGES (INCLUDING, 26 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; 27 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED 28 * AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, 29 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT 30 * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 31 * SUCH DAMAGE. 32 */ 33 34 /* 35 * Copyright (c) 1982, 1986, 1993, 1994, 1995 36 * The Regents of the University of California. All rights reserved. 37 * 38 * Redistribution and use in source and binary forms, with or without 39 * modification, are permitted provided that the following conditions 40 * are met: 41 * 1. Redistributions of source code must retain the above copyright 42 * notice, this list of conditions and the following disclaimer. 43 * 2. Redistributions in binary form must reproduce the above copyright 44 * notice, this list of conditions and the following disclaimer in the 45 * documentation and/or other materials provided with the distribution. 46 * 3. All advertising materials mentioning features or use of this software 47 * must display the following acknowledgement: 48 * This product includes software developed by the University of 49 * California, Berkeley and its contributors. 50 * 4. Neither the name of the University 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 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 55 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 56 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 57 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 58 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 59 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 60 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 61 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 62 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 63 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 64 * SUCH DAMAGE. 65 * 66 * @(#)tcp_var.h 8.4 (Berkeley) 5/24/95 67 * $FreeBSD: src/sys/netinet/tcp_var.h,v 1.56.2.13 2003/02/03 02:34:07 hsu Exp $ 68 * $DragonFly: src/sys/netinet/tcp_var.h,v 1.42 2008/10/27 02:56:30 sephe Exp $ 69 */ 70 71 #ifndef _NETINET_TCP_VAR_H_ 72 #define _NETINET_TCP_VAR_H_ 73 74 #ifndef _NETINET_IN_PCB_H_ 75 #include <netinet/in_pcb.h> /* needed for in_conninfo, inp_gen_t */ 76 #endif 77 #ifndef _NETINET_TCP_H_ 78 #include <netinet/tcp.h> 79 #endif 80 81 /* 82 * Kernel variables for tcp. 83 */ 84 extern int tcp_do_rfc1323; 85 extern int tcp_do_rfc3390; 86 extern int tcp_do_sack; 87 extern int tcp_do_smartsack; 88 extern int tcp_aggregate_acks; 89 90 /* TCP segment queue entry */ 91 struct tseg_qent { 92 LIST_ENTRY(tseg_qent) tqe_q; 93 int tqe_len; /* TCP segment data length */ 94 struct tcphdr *tqe_th; /* a pointer to tcp header */ 95 struct mbuf *tqe_m; /* mbuf contains packet */ 96 }; 97 LIST_HEAD(tsegqe_head, tseg_qent); 98 extern int tcp_reass_maxseg; 99 extern int tcp_reass_qsize; 100 #ifdef MALLOC_DECLARE 101 MALLOC_DECLARE(M_TSEGQ); 102 #endif 103 104 struct tcptemp { 105 u_char tt_ipgen[40]; /* the size must be of max ip header, now IPv6 */ 106 struct tcphdr tt_t; 107 }; 108 109 #define tcp6cb tcpcb /* for KAME src sync over BSD*'s */ 110 111 struct raw_sackblock { /* covers [start, end) */ 112 tcp_seq rblk_start; 113 tcp_seq rblk_end; 114 }; 115 116 /* maximum number of SACK blocks that will fit in the TCP option space */ 117 #define MAX_SACK_REPORT_BLOCKS 4 118 119 TAILQ_HEAD(sackblock_list, sackblock); 120 121 struct scoreboard { 122 int nblocks; 123 struct sackblock_list sackblocks; 124 tcp_seq lostseq; /* passed SACK lost test */ 125 struct sackblock *lastfound; /* search hint */ 126 }; 127 128 struct netmsg_tcp_timer; 129 130 /* 131 * Tcp control block, one per tcp; fields: 132 * Organized for 16 byte cacheline efficiency. 133 */ 134 struct tcpcb { 135 struct tsegqe_head t_segq; 136 int t_dupacks; /* consecutive dup acks recd */ 137 int tt_cpu; /* sanity check the cpu */ 138 139 struct tcp_callout *tt_rexmt; /* retransmit timer */ 140 struct tcp_callout *tt_persist;/* retransmit persistence */ 141 struct tcp_callout *tt_keep; /* keepalive */ 142 struct tcp_callout *tt_2msl; /* 2*msl TIME_WAIT timer */ 143 struct tcp_callout *tt_delack; /* delayed ACK timer */ 144 struct netmsg_tcp_timer *tt_msg; /* timer message */ 145 146 struct inpcb *t_inpcb; /* back pointer to internet pcb */ 147 int t_state; /* state of this connection */ 148 u_int t_flags; 149 #define TF_ACKNOW 0x00000001 /* ack peer immediately */ 150 #define TF_DELACK 0x00000002 /* ack, but try to delay it */ 151 #define TF_NODELAY 0x00000004 /* don't delay packets to coalesce */ 152 #define TF_NOOPT 0x00000008 /* don't use tcp options */ 153 #define TF_SENTFIN 0x00000010 /* have sent FIN */ 154 #define TF_REQ_SCALE 0x00000020 /* have/will request window scaling */ 155 #define TF_RCVD_SCALE 0x00000040 /* other side has requested scaling */ 156 #define TF_REQ_TSTMP 0x00000080 /* have/will request timestamps */ 157 #define TF_RCVD_TSTMP 0x00000100 /* a timestamp was received in SYN */ 158 #define TF_SACK_PERMITTED 0x00000200 /* other side said I could SACK */ 159 #define TF_NEEDSYN 0x00000400 /* send SYN (implicit state) */ 160 #define TF_NEEDFIN 0x00000800 /* send FIN (implicit state) */ 161 #define TF_NOPUSH 0x00001000 /* don't push */ 162 /* 0x00001000 - 0x00008000 were used for T/TCP */ 163 #define TF_MORETOCOME 0x00010000 /* More data to be appended to sock */ 164 #define TF_LQ_OVERFLOW 0x00020000 /* listen queue overflow */ 165 #define TF_LASTIDLE 0x00040000 /* connection was previously idle */ 166 #define TF_RXWIN0SENT 0x00080000 /* sent a receiver win 0 in response */ 167 #define TF_FASTRECOVERY 0x00100000 /* in NewReno Fast Recovery */ 168 #define TF_WASFRECOVERY 0x00200000 /* was in NewReno Fast Recovery */ 169 #define TF_FIRSTACCACK 0x00400000 /* Look for 1st acceptable ACK. */ 170 #define TF_FASTREXMT 0x00800000 /* Did Fast Retransmit. */ 171 #define TF_EARLYREXMT 0x01000000 /* Did Early (Fast) Retransmit. */ 172 #define TF_FORCE 0x02000000 /* Set if forcing out a byte */ 173 #define TF_ONOUTPUTQ 0x04000000 /* on t_outputq list */ 174 #define TF_DUPSEG 0x08000000 /* last seg a duplicate */ 175 #define TF_ENCLOSESEG 0x10000000 /* enclosing SACK block */ 176 #define TF_SACKLEFT 0x20000000 /* send SACK blocks from left side */ 177 #define TF_KEEPALIVE 0x40000000 /* temporary keepalive */ 178 #define TF_RXRESIZED 0x80000000 /* rcvbuf was resized */ 179 tcp_seq snd_up; /* send urgent pointer */ 180 181 tcp_seq snd_una; /* send unacknowledged */ 182 tcp_seq snd_recover; /* for use with NewReno Fast Recovery */ 183 tcp_seq snd_max; /* highest sequence number sent; 184 * used to recognize retransmits */ 185 tcp_seq snd_nxt; /* send next */ 186 187 tcp_seq snd_wl1; /* window update seg seq number */ 188 tcp_seq snd_wl2; /* window update seg ack number */ 189 tcp_seq iss; /* initial send sequence number */ 190 tcp_seq irs; /* initial receive sequence number */ 191 192 tcp_seq rcv_nxt; /* receive next */ 193 tcp_seq rcv_adv; /* advertised window */ 194 u_long rcv_wnd; /* receive window */ 195 tcp_seq rcv_up; /* receive urgent pointer */ 196 197 u_long snd_wnd; /* send window */ 198 u_long snd_cwnd; /* congestion-controlled window */ 199 u_long snd_wacked; /* bytes acked in one send window */ 200 u_long snd_ssthresh; /* snd_cwnd size threshold for 201 * for slow start exponential to 202 * linear switch */ 203 204 int t_rxtcur; /* current retransmit value (ticks) */ 205 u_int t_maxseg; /* maximum segment size */ 206 int t_srtt; /* smoothed round-trip time */ 207 int t_rttvar; /* variance in round-trip time */ 208 209 u_int t_maxopd; /* mss plus options */ 210 211 u_long t_rcvtime; /* inactivity time */ 212 u_long t_starttime; /* time connection was established */ 213 int t_rtttime; /* round trip time */ 214 tcp_seq t_rtseq; /* sequence number being timed */ 215 216 int t_rxtshift; /* log(2) of rexmt exp. backoff */ 217 u_int t_rttmin; /* minimum rtt allowed */ 218 u_int t_rttbest; /* best rtt we've seen */ 219 u_long t_rttupdated; /* number of times rtt sampled */ 220 u_long max_sndwnd; /* largest window peer has offered */ 221 222 int t_softerror; /* possible error not yet reported */ 223 /* out-of-band data */ 224 char t_oobflags; /* have some */ 225 char t_iobc; /* input character */ 226 #define TCPOOB_HAVEDATA 0x01 227 #define TCPOOB_HADDATA 0x02 228 229 /* RFC 1323 variables */ 230 u_char snd_scale; /* window scaling for send window */ 231 u_char rcv_scale; /* window scaling for recv window */ 232 u_char request_r_scale; /* pending window scaling */ 233 u_char requested_s_scale; 234 u_long ts_recent; /* timestamp echo data */ 235 236 u_long ts_recent_age; /* when last updated */ 237 tcp_seq last_ack_sent; 238 239 /* experimental */ 240 u_long snd_cwnd_prev; /* cwnd prior to retransmit */ 241 u_long snd_wacked_prev; /* prior bytes acked in send window */ 242 u_long snd_ssthresh_prev; /* ssthresh prior to retransmit */ 243 tcp_seq snd_recover_prev; /* snd_recover prior to retransmit */ 244 u_long t_badrxtwin; /* window for retransmit recovery */ 245 u_long t_rexmtTS; /* timestamp of last retransmit */ 246 u_char snd_limited; /* segments limited transmitted */ 247 248 tcp_seq rexmt_high; /* highest seq # retransmitted + 1 */ 249 tcp_seq snd_max_rexmt; /* snd_max when rexmting snd_una */ 250 struct scoreboard scb; /* sack scoreboard */ 251 struct raw_sackblock reportblk; /* incoming segment or D-SACK block */ 252 struct raw_sackblock encloseblk; 253 int nsackhistory; 254 struct raw_sackblock sackhistory[MAX_SACK_REPORT_BLOCKS]; /* reported */ 255 TAILQ_ENTRY(tcpcb) t_outputq; /* tcp_output needed list */ 256 257 /* bandwith limitation */ 258 u_long snd_bandwidth; /* calculated bandwidth or 0 */ 259 u_long snd_bwnd; /* bandwidth-controlled window */ 260 int t_bw_rtttime; /* used for bandwidth calculation */ 261 tcp_seq t_bw_rtseq; /* used for bandwidth calculation */ 262 263 /* anti DoS counters */ 264 u_long rcv_second; /* start of interval second */ 265 u_long rcv_pps; /* received packets per second */ 266 u_long rcv_byps; /* received bytes per second */ 267 268 u_int32_t rfbuf_ts; /* recv buffer autoscaling timestamp */ 269 int rfbuf_cnt; /* recv buffer autoscaling byte count */ 270 }; 271 272 #define IN_FASTRECOVERY(tp) (tp->t_flags & TF_FASTRECOVERY) 273 #define ENTER_FASTRECOVERY(tp) tp->t_flags |= TF_FASTRECOVERY 274 #define EXIT_FASTRECOVERY(tp) tp->t_flags &= ~TF_FASTRECOVERY 275 276 /* 277 * TCP statistics. 278 */ 279 struct tcp_stats { 280 u_long tcps_connattempt; /* connections initiated */ 281 u_long tcps_accepts; /* connections accepted */ 282 u_long tcps_connects; /* connections established */ 283 u_long tcps_drops; /* connections dropped */ 284 u_long tcps_conndrops; /* embryonic connections dropped */ 285 u_long tcps_minmssdrops; /* average minmss too low drops */ 286 u_long tcps_closed; /* conn. closed (includes drops) */ 287 u_long tcps_segstimed; /* segs where we tried to get rtt */ 288 u_long tcps_rttupdated; /* times we succeeded */ 289 u_long tcps_delack; /* delayed acks sent */ 290 u_long tcps_timeoutdrop; /* conn. dropped in rxmt timeout */ 291 u_long tcps_rexmttimeo; /* retransmit timeouts */ 292 u_long tcps_persisttimeo; /* persist timeouts */ 293 u_long tcps_keeptimeo; /* keepalive timeouts */ 294 u_long tcps_keepprobe; /* keepalive probes sent */ 295 u_long tcps_keepdrops; /* connections dropped in keepalive */ 296 297 u_long tcps_sndtotal; /* total packets sent */ 298 u_long tcps_sndpack; /* data packets sent */ 299 u_long tcps_sndbyte; /* data bytes sent */ 300 u_long tcps_sndrexmitpack; /* data packets retransmitted */ 301 u_long tcps_sndrexmitbyte; /* data bytes retransmitted */ 302 u_long tcps_sndfastrexmit; /* Fast Retransmissions */ 303 u_long tcps_sndearlyrexmit; /* early Fast Retransmissions */ 304 u_long tcps_sndlimited; /* Limited Transmit packets */ 305 u_long tcps_sndrtobad; /* spurious RTO retransmissions */ 306 u_long tcps_sndfastrexmitbad; /* spurious Fast Retransmissions */ 307 u_long tcps_sndearlyrexmitbad; /* spurious early Fast Retransmissions, 308 a subset of tcps_sndfastrexmitbad */ 309 u_long tcps_eifeldetected; /* Eifel-detected spurious rexmits */ 310 u_long tcps_rttcantdetect; /* Eifel but not 1/2 RTT-detectable */ 311 u_long tcps_rttdetected; /* RTT-detected spurious RTO rexmits */ 312 u_long tcps_sndacks; /* ack-only packets sent */ 313 u_long tcps_sndprobe; /* window probes sent */ 314 u_long tcps_sndurg; /* packets sent with URG only */ 315 u_long tcps_sndwinup; /* window update-only packets sent */ 316 u_long tcps_sndctrl; /* control (SYN|FIN|RST) packets sent */ 317 u_long tcps_sndsackpack; /* packets sent by SACK recovery alg */ 318 u_long tcps_sndsackbyte; /* bytes sent by SACK recovery */ 319 u_long tcps_snduna3; /* re-retransmit snd_una on 3 new seg */ 320 u_long tcps_snduna1; /* re-retransmit snd_una on 1 new seg */ 321 322 u_long tcps_rcvtotal; /* total packets received */ 323 u_long tcps_rcvpack; /* packets received in sequence */ 324 u_long tcps_rcvbyte; /* bytes received in sequence */ 325 u_long tcps_rcvbadsum; /* packets received with ccksum errs */ 326 u_long tcps_rcvbadoff; /* packets received with bad offset */ 327 u_long tcps_rcvmemdrop; /* packets dropped for lack of memory */ 328 u_long tcps_rcvshort; /* packets received too short */ 329 u_long tcps_rcvduppack; /* duplicate-only packets received */ 330 u_long tcps_rcvdupbyte; /* duplicate-only bytes received */ 331 u_long tcps_rcvpartduppack; /* packets with some duplicate data */ 332 u_long tcps_rcvpartdupbyte; /* dup. bytes in part-dup. packets */ 333 u_long tcps_rcvoopack; /* out-of-order packets received */ 334 u_long tcps_rcvoobyte; /* out-of-order bytes received */ 335 u_long tcps_rcvpackafterwin; /* packets with data after window */ 336 u_long tcps_rcvbyteafterwin; /* bytes rcvd after window */ 337 u_long tcps_rcvafterclose; /* packets rcvd after "close" */ 338 u_long tcps_rcvwinprobe; /* rcvd window probe packets */ 339 u_long tcps_rcvdupack; /* rcvd duplicate acks */ 340 u_long tcps_rcvacktoomuch; /* rcvd acks for unsent data */ 341 u_long tcps_rcvackpack; /* rcvd ack packets */ 342 u_long tcps_rcvackbyte; /* bytes acked by rcvd acks */ 343 u_long tcps_rcvwinupd; /* rcvd window update packets */ 344 u_long tcps_pawsdrop; /* segments dropped due to PAWS */ 345 u_long tcps_predack; /* times hdr predict ok for acks */ 346 u_long tcps_preddat; /* times hdr predict ok for data pkts */ 347 u_long tcps_pcbcachemiss; 348 u_long tcps_cachedrtt; /* times cached RTT in route updated */ 349 u_long tcps_cachedrttvar; /* times cached rttvar updated */ 350 u_long tcps_cachedssthresh; /* times cached ssthresh updated */ 351 u_long tcps_usedrtt; /* times RTT initialized from route */ 352 u_long tcps_usedrttvar; /* times RTTVAR initialized from rt */ 353 u_long tcps_usedssthresh; /* times ssthresh initialized from rt*/ 354 u_long tcps_persistdrop; /* timeout in persist state */ 355 u_long tcps_badsyn; /* bogus SYN, e.g. premature ACK */ 356 u_long tcps_mturesent; /* resends due to MTU discovery */ 357 u_long tcps_listendrop; /* listen queue overflows */ 358 359 u_long tcps_sc_added; /* entry added to syncache */ 360 u_long tcps_sc_retransmitted; /* syncache entry was retransmitted */ 361 u_long tcps_sc_dupsyn; /* duplicate SYN packet */ 362 u_long tcps_sc_dropped; /* could not reply to packet */ 363 u_long tcps_sc_completed; /* successful extraction of entry */ 364 u_long tcps_sc_bucketoverflow; /* syncache per-bucket limit hit */ 365 u_long tcps_sc_cacheoverflow; /* syncache cache limit hit */ 366 u_long tcps_sc_reset; /* RST removed entry from syncache */ 367 u_long tcps_sc_stale; /* timed out or listen socket gone */ 368 u_long tcps_sc_aborted; /* syncache entry aborted */ 369 u_long tcps_sc_badack; /* removed due to bad ACK */ 370 u_long tcps_sc_unreach; /* ICMP unreachable received */ 371 u_long tcps_sc_zonefail; /* zalloc() failed */ 372 u_long tcps_sc_sendcookie; /* SYN cookie sent */ 373 u_long tcps_sc_recvcookie; /* SYN cookie received */ 374 }; 375 376 #ifdef _KERNEL 377 378 #if defined(SMP) 379 #define tcpstat tcpstats_percpu[mycpuid] 380 #else 381 #define tcpstat tcpstats_percpu[0] 382 #endif 383 384 struct sockopt; 385 386 extern struct tcp_stats tcpstats_percpu[MAXCPU]; 387 388 static const int tcprexmtthresh = 3; 389 #endif 390 391 /* 392 * Structure to hold TCP options that are only used during segment 393 * processing (in tcp_input), but not held in the tcpcb. 394 * It's basically used to reduce the number of parameters 395 * to tcp_dooptions. 396 */ 397 struct tcpopt { 398 u_long to_flags; /* which options are present */ 399 #define TOF_TS 0x0001 /* timestamp */ 400 #define TOF_MSS 0x0010 401 #define TOF_SCALE 0x0020 402 #define TOF_SACK_PERMITTED 0x0040 403 #define TOF_SACK 0x0080 404 u_int32_t to_tsval; 405 u_int32_t to_tsecr; 406 u_int16_t to_mss; 407 u_int8_t to_requested_s_scale; 408 u_int8_t to_nsackblocks; 409 struct raw_sackblock *to_sackblocks; 410 }; 411 412 struct syncache { 413 inp_gen_t sc_inp_gencnt; /* pointer check */ 414 struct tcpcb *sc_tp; /* tcb for listening socket */ 415 struct mbuf *sc_ipopts; /* source route */ 416 struct in_conninfo sc_inc; /* addresses */ 417 #define sc_route sc_inc.inc_route 418 #define sc_route6 sc_inc.inc6_route 419 u_int32_t sc_tsrecent; 420 tcp_seq sc_irs; /* seq from peer */ 421 tcp_seq sc_iss; /* our ISS */ 422 u_long sc_rxttime; /* retransmit time */ 423 u_int16_t sc_rxtslot; /* retransmit counter */ 424 u_int16_t sc_peer_mss; /* peer's MSS */ 425 u_int16_t sc_wnd; /* advertised window */ 426 u_int8_t sc_requested_s_scale:4, 427 sc_request_r_scale:4; 428 u_int8_t sc_flags; 429 #define SCF_NOOPT 0x01 /* no TCP options */ 430 #define SCF_WINSCALE 0x02 /* negotiated window scaling */ 431 #define SCF_TIMESTAMP 0x04 /* negotiated timestamps */ 432 #define SCF_UNREACH 0x10 /* icmp unreachable received */ 433 #define SCF_SACK_PERMITTED 0x20 /* saw SACK permitted option */ 434 TAILQ_ENTRY(syncache) sc_hash; 435 TAILQ_ENTRY(syncache) sc_timerq; 436 }; 437 438 struct syncache_head { 439 TAILQ_HEAD(, syncache) sch_bucket; 440 u_int sch_length; 441 }; 442 443 #define intotcpcb(ip) ((struct tcpcb *)(ip)->inp_ppcb) 444 #define sototcpcb(so) (intotcpcb(sotoinpcb(so))) 445 446 /* 447 * The smoothed round-trip time and estimated variance 448 * are stored as fixed point numbers scaled by the values below. 449 * For convenience, these scales are also used in smoothing the average 450 * (smoothed = (1/scale)sample + ((scale-1)/scale)smoothed). 451 * With these scales, srtt has 3 bits to the right of the binary point, 452 * and thus an "ALPHA" of 0.875. rttvar has 2 bits to the right of the 453 * binary point, and is smoothed with an ALPHA of 0.75. 454 */ 455 #define TCP_RTT_SCALE 32 /* multiplier for srtt; 3 bits frac. */ 456 #define TCP_RTT_SHIFT 5 /* shift for srtt; 3 bits frac. */ 457 #define TCP_RTTVAR_SCALE 16 /* multiplier for rttvar; 2 bits */ 458 #define TCP_RTTVAR_SHIFT 4 /* shift for rttvar; 2 bits */ 459 #define TCP_DELTA_SHIFT 2 /* see tcp_input.c */ 460 461 /* 462 * The initial retransmission should happen at rtt + 4 * rttvar. 463 * Because of the way we do the smoothing, srtt and rttvar 464 * will each average +1/2 tick of bias. When we compute 465 * the retransmit timer, we want 1/2 tick of rounding and 466 * 1 extra tick because of +-1/2 tick uncertainty in the 467 * firing of the timer. The bias will give us exactly the 468 * 1.5 tick we need. But, because the bias is 469 * statistical, we have to test that we don't drop below 470 * the minimum feasible timer (which is 2 ticks). 471 * This version of the macro adapted from a paper by Lawrence 472 * Brakmo and Larry Peterson which outlines a problem caused 473 * by insufficient precision in the original implementation, 474 * which results in inappropriately large RTO values for very 475 * fast networks. 476 */ 477 #define TCP_REXMTVAL(tp) \ 478 max((tp)->t_rttmin, (((tp)->t_srtt >> (TCP_RTT_SHIFT - TCP_DELTA_SHIFT)) \ 479 + (tp)->t_rttvar) >> TCP_DELTA_SHIFT) 480 481 /* 482 * TCB structure exported to user-land via sysctl(3). 483 * Evil hack: declare only if in_pcb.h and sys/socketvar.h have been 484 * included. Not all of our clients do. 485 */ 486 #if defined(_NETINET_IN_PCB_H_) && defined(_SYS_SOCKETVAR_H_) 487 struct xtcpcb { 488 size_t xt_len; 489 struct inpcb xt_inp; 490 struct tcpcb xt_tp; 491 struct xsocket xt_socket; 492 u_quad_t xt_alignment_hack; 493 }; 494 #endif 495 496 /* 497 * Names for TCP sysctl objects 498 */ 499 #define TCPCTL_DO_RFC1323 1 /* use RFC-1323 extensions */ 500 /* 2 was TCPCTL_DO_RFC1644 */ 501 #define TCPCTL_MSSDFLT 3 /* MSS default */ 502 #define TCPCTL_STATS 4 /* statistics (read-only) */ 503 #define TCPCTL_RTTDFLT 5 /* default RTT estimate */ 504 #define TCPCTL_KEEPIDLE 6 /* keepalive idle timer */ 505 #define TCPCTL_KEEPINTVL 7 /* interval to send keepalives */ 506 #define TCPCTL_SENDSPACE 8 /* send buffer space */ 507 #define TCPCTL_RECVSPACE 9 /* receive buffer space */ 508 #define TCPCTL_KEEPINIT 10 /* timeout for establishing syn */ 509 #define TCPCTL_PCBLIST 11 /* list of all outstanding PCBs */ 510 #define TCPCTL_DELACKTIME 12 /* time before sending delayed ACK */ 511 #define TCPCTL_V6MSSDFLT 13 /* MSS default for IPv6 */ 512 #define TCPCTL_MAXID 14 513 514 #define TCPCTL_NAMES { \ 515 { 0, 0 }, \ 516 { "rfc1323", CTLTYPE_INT }, \ 517 { "reserved", CTLTYPE_INT}, /* was rfc1644 */ \ 518 { "mssdflt", CTLTYPE_INT }, \ 519 { "stats", CTLTYPE_STRUCT }, \ 520 { "rttdflt", CTLTYPE_INT }, \ 521 { "keepidle", CTLTYPE_INT }, \ 522 { "keepintvl", CTLTYPE_INT }, \ 523 { "sendspace", CTLTYPE_INT }, \ 524 { "recvspace", CTLTYPE_INT }, \ 525 { "keepinit", CTLTYPE_INT }, \ 526 { "pcblist", CTLTYPE_STRUCT }, \ 527 { "delacktime", CTLTYPE_INT }, \ 528 { "v6mssdflt", CTLTYPE_INT }, \ 529 } 530 531 #ifdef _KERNEL 532 #ifdef SYSCTL_DECL 533 SYSCTL_DECL(_net_inet_tcp); 534 #endif 535 536 #define TCP_DO_SACK(tp) ((tp)->t_flags & TF_SACK_PERMITTED) 537 538 TAILQ_HEAD(tcpcbackqhead,tcpcb); 539 540 extern struct inpcbinfo tcbinfo[]; 541 extern struct tcpcbackqhead tcpcbackq[]; 542 543 extern int tcp_mssdflt; /* XXX */ 544 extern int tcp_minmss; 545 extern int tcp_delack_enabled; 546 extern int path_mtu_discovery; 547 548 int tcp_addrcpu(in_addr_t faddr, in_port_t fport, 549 in_addr_t laddr, in_port_t lport); 550 struct lwkt_port * 551 tcp_addrport(in_addr_t faddr, in_port_t fport, 552 in_addr_t laddr, in_port_t lport); 553 struct lwkt_port *tcp_addrport0(void); 554 void tcp_canceltimers (struct tcpcb *); 555 struct tcpcb * 556 tcp_close (struct tcpcb *); 557 void tcpmsg_service_loop (void *); 558 void tcp_ctlinput (int, struct sockaddr *, void *); 559 int tcp_ctloutput (struct socket *, struct sockopt *); 560 struct lwkt_port * 561 tcp_cport(int cpu); 562 struct tcpcb * 563 tcp_drop (struct tcpcb *, int); 564 void tcp_drain (void); 565 void tcp_fasttimo (void); 566 void tcp_init (void); 567 void tcp_thread_init (void); 568 void tcp_input (struct mbuf *, ...); 569 void tcp_mss (struct tcpcb *, int); 570 int tcp_mssopt (struct tcpcb *); 571 void tcp_drop_syn_sent (struct inpcb *, int); 572 void tcp_mtudisc (struct inpcb *, int); 573 struct tcpcb * 574 tcp_newtcpcb (struct inpcb *); 575 int tcp_output (struct tcpcb *); 576 void tcp_quench (struct inpcb *, int); 577 void tcp_respond (struct tcpcb *, void *, 578 struct tcphdr *, struct mbuf *, tcp_seq, tcp_seq, int); 579 struct rtentry * 580 tcp_rtlookup (struct in_conninfo *); 581 int tcp_sack_bytes_below(struct scoreboard *scb, tcp_seq seq); 582 void tcp_sack_cleanup(struct scoreboard *scb); 583 int tcp_sack_ndsack_blocks(struct raw_sackblock *blocks, 584 const int numblocks, tcp_seq snd_una); 585 void tcp_sack_fill_report(struct tcpcb *tp, u_char *opt, u_int *plen); 586 boolean_t 587 tcp_sack_has_sacked(struct scoreboard *scb, u_int amount); 588 void tcp_sack_tcpcb_init(struct tcpcb *tp); 589 uint32_t tcp_sack_compute_pipe(struct tcpcb *tp); 590 boolean_t 591 tcp_sack_nextseg(struct tcpcb *tp, tcp_seq *nextrexmt, uint32_t *len, 592 boolean_t *losdup); 593 #ifdef later 594 void tcp_sack_revert_scoreboard(struct scoreboard *scb, tcp_seq snd_una, 595 u_int maxseg); 596 void tcp_sack_save_scoreboard(struct scoreboard *scb); 597 #endif 598 void tcp_sack_skip_sacked(struct scoreboard *scb, tcp_seq *prexmt); 599 void tcp_sack_update_scoreboard(struct tcpcb *tp, struct tcpopt *to); 600 void tcp_save_congestion_state(struct tcpcb *tp); 601 void tcp_revert_congestion_state(struct tcpcb *tp); 602 void tcp_setpersist (struct tcpcb *); 603 void tcp_slowtimo (void); 604 struct tcptemp *tcp_maketemplate (struct tcpcb *); 605 void tcp_freetemplate (struct tcptemp *); 606 void tcp_fillheaders (struct tcpcb *, void *, void *); 607 struct lwkt_port * 608 tcp_soport(struct socket *, struct sockaddr *, struct mbuf **); 609 struct lwkt_port * 610 tcp_soport_attach(struct socket *); 611 struct lwkt_port * 612 tcp_ctlport(int, struct sockaddr *, void *); 613 struct tcpcb * 614 tcp_timers (struct tcpcb *, int); 615 void tcp_trace (short, short, struct tcpcb *, void *, struct tcphdr *, 616 int); 617 void tcp_xmit_bandwidth_limit(struct tcpcb *tp, tcp_seq ack_seq); 618 void syncache_init(void); 619 void syncache_unreach(struct in_conninfo *, struct tcphdr *); 620 int syncache_expand(struct in_conninfo *, struct tcphdr *, 621 struct socket **, struct mbuf *); 622 int syncache_add(struct in_conninfo *, struct tcpopt *, 623 struct tcphdr *, struct socket **, struct mbuf *); 624 void syncache_chkrst(struct in_conninfo *, struct tcphdr *); 625 void syncache_badack(struct in_conninfo *); 626 627 extern struct pr_usrreqs tcp_usrreqs; 628 extern u_long tcp_sendspace; 629 extern u_long tcp_recvspace; 630 tcp_seq tcp_new_isn (struct tcpcb *); 631 632 #endif /* _KERNEL */ 633 634 #endif /* _NETINET_TCP_VAR_H_ */ 635