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_low_rtobase; 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 struct sackblock *freecache; /* one slot free block cache */ 127 }; 128 129 struct netmsg_tcp_timer; 130 131 /* 132 * Tcp control block, one per tcp; fields: 133 * Organized for 16 byte cacheline efficiency. 134 */ 135 struct tcpcb { 136 struct tsegqe_head t_segq; 137 int t_dupacks; /* consecutive dup acks recd */ 138 int tt_cpu; /* sanity check the cpu */ 139 140 struct tcp_callout *tt_rexmt; /* retransmit timer */ 141 struct tcp_callout *tt_persist;/* retransmit persistence */ 142 struct tcp_callout *tt_keep; /* keepalive */ 143 struct tcp_callout *tt_2msl; /* 2*msl TIME_WAIT timer */ 144 struct tcp_callout *tt_delack; /* delayed ACK timer */ 145 struct netmsg_tcp_timer *tt_msg; /* timer message */ 146 147 struct inpcb *t_inpcb; /* back pointer to internet pcb */ 148 int t_state; /* state of this connection */ 149 u_int t_flags; 150 #define TF_ACKNOW 0x00000001 /* ack peer immediately */ 151 #define TF_DELACK 0x00000002 /* ack, but try to delay it */ 152 #define TF_NODELAY 0x00000004 /* don't delay packets to coalesce */ 153 #define TF_NOOPT 0x00000008 /* don't use tcp options */ 154 #define TF_SENTFIN 0x00000010 /* have sent FIN */ 155 #define TF_REQ_SCALE 0x00000020 /* have/will request window scaling */ 156 #define TF_RCVD_SCALE 0x00000040 /* other side has requested scaling */ 157 #define TF_REQ_TSTMP 0x00000080 /* have/will request timestamps */ 158 #define TF_RCVD_TSTMP 0x00000100 /* a timestamp was received in SYN */ 159 #define TF_SACK_PERMITTED 0x00000200 /* other side said I could SACK */ 160 #define TF_NEEDSYN 0x00000400 /* send SYN (implicit state) */ 161 #define TF_NEEDFIN 0x00000800 /* send FIN (implicit state) */ 162 #define TF_NOPUSH 0x00001000 /* don't push */ 163 #define TF_LISTEN 0x00002000 /* listen(2) has been called */ 164 #define TF_SIGNATURE 0x00004000 /* require MD5 digests (RFC2385) */ 165 #define TF_FASTKEEP 0x00008000 /* use a faster tcp_keepidle */ 166 #define TF_MORETOCOME 0x00010000 /* More data to be appended to sock */ 167 #define TF_UNUSED00 0x00020000 /* unused */ 168 #define TF_LASTIDLE 0x00040000 /* connection was previously idle */ 169 #define TF_RXWIN0SENT 0x00080000 /* sent a receiver win 0 in response */ 170 #define TF_FASTRECOVERY 0x00100000 /* in NewReno Fast Recovery */ 171 #define TF_WASFRECOVERY 0x00200000 /* was in NewReno Fast Recovery */ 172 #define TF_FIRSTACCACK 0x00400000 /* Look for 1st acceptable ACK. */ 173 #define TF_FASTREXMT 0x00800000 /* Did Fast Retransmit. */ 174 #define TF_EARLYREXMT 0x01000000 /* Did Early (Fast) Retransmit. */ 175 #define TF_FORCE 0x02000000 /* Set if forcing out a byte */ 176 #define TF_ONOUTPUTQ 0x04000000 /* on t_outputq list */ 177 #define TF_DUPSEG 0x08000000 /* last seg a duplicate */ 178 #define TF_ENCLOSESEG 0x10000000 /* enclosing SACK block */ 179 #define TF_SACKLEFT 0x20000000 /* send SACK blocks from left side */ 180 #define TF_KEEPALIVE 0x40000000 /* temporary keepalive */ 181 #define TF_RXRESIZED 0x80000000 /* rcvbuf was resized */ 182 tcp_seq snd_up; /* send urgent pointer */ 183 184 tcp_seq snd_una; /* send unacknowledged */ 185 tcp_seq snd_recover; /* for use with NewReno Fast Recovery */ 186 tcp_seq snd_max; /* highest sequence number sent; 187 * used to recognize retransmits */ 188 tcp_seq snd_nxt; /* send next */ 189 190 tcp_seq snd_wl1; /* window update seg seq number */ 191 tcp_seq snd_wl2; /* window update seg ack number */ 192 tcp_seq iss; /* initial send sequence number */ 193 tcp_seq irs; /* initial receive sequence number */ 194 195 tcp_seq rcv_nxt; /* receive next */ 196 tcp_seq rcv_adv; /* advertised window */ 197 u_long rcv_wnd; /* receive window */ 198 tcp_seq rcv_up; /* receive urgent pointer */ 199 200 u_long snd_wnd; /* send window */ 201 u_long snd_cwnd; /* congestion-controlled window */ 202 u_long snd_wacked; /* bytes acked in one send window */ 203 u_long snd_ssthresh; /* snd_cwnd size threshold for 204 * for slow start exponential to 205 * linear switch */ 206 207 int t_rxtcur; /* current retransmit value (ticks) */ 208 u_int t_maxseg; /* maximum segment size */ 209 int t_srtt; /* smoothed round-trip time */ 210 int t_rttvar; /* variance in round-trip time */ 211 212 u_int t_maxopd; /* mss plus options */ 213 214 u_long t_rcvtime; /* inactivity time */ 215 u_long t_starttime; /* time connection was established */ 216 int t_rtttime; /* round trip time */ 217 tcp_seq t_rtseq; /* sequence number being timed */ 218 219 int t_rxtshift; /* log(2) of rexmt exp. backoff */ 220 u_int t_rttmin; /* minimum rtt allowed */ 221 u_int t_rttbest; /* best rtt we've seen */ 222 u_long t_rttupdated; /* number of times rtt sampled */ 223 u_long max_sndwnd; /* largest window peer has offered */ 224 225 int t_softerror; /* possible error not yet reported */ 226 /* out-of-band data */ 227 char t_oobflags; /* have some */ 228 char t_iobc; /* input character */ 229 #define TCPOOB_HAVEDATA 0x01 230 #define TCPOOB_HADDATA 0x02 231 232 /* RFC 1323 variables */ 233 u_char snd_scale; /* window scaling for send window */ 234 u_char rcv_scale; /* window scaling for recv window */ 235 u_char request_r_scale; /* pending window scaling */ 236 u_char requested_s_scale; 237 u_long ts_recent; /* timestamp echo data */ 238 239 u_long ts_recent_age; /* when last updated */ 240 tcp_seq last_ack_sent; 241 242 /* experimental */ 243 u_long snd_cwnd_prev; /* cwnd prior to retransmit */ 244 u_long snd_wacked_prev; /* prior bytes acked in send window */ 245 u_long snd_ssthresh_prev; /* ssthresh prior to retransmit */ 246 tcp_seq snd_recover_prev; /* snd_recover prior to retransmit */ 247 u_long t_badrxtwin; /* window for retransmit recovery */ 248 u_long t_rexmtTS; /* timestamp of last retransmit */ 249 u_char snd_limited; /* segments limited transmitted */ 250 251 tcp_seq rexmt_high; /* highest seq # retransmitted + 1 */ 252 tcp_seq snd_max_rexmt; /* snd_max when rexmting snd_una */ 253 struct scoreboard scb; /* sack scoreboard */ 254 struct raw_sackblock reportblk; /* incoming segment or D-SACK block */ 255 struct raw_sackblock encloseblk; 256 int nsackhistory; 257 struct raw_sackblock sackhistory[MAX_SACK_REPORT_BLOCKS]; /* reported */ 258 TAILQ_ENTRY(tcpcb) t_outputq; /* tcp_output needed list */ 259 260 /* bandwith limitation */ 261 u_long snd_bandwidth; /* calculated bandwidth or 0 */ 262 u_long snd_bwnd; /* bandwidth-controlled window */ 263 int t_bw_rtttime; /* used for bandwidth calculation */ 264 tcp_seq t_bw_rtseq; /* used for bandwidth calculation */ 265 266 /* anti DoS counters */ 267 u_long rcv_second; /* start of interval second */ 268 u_long rcv_pps; /* received packets per second */ 269 u_long rcv_byps; /* received bytes per second */ 270 271 u_int32_t rfbuf_ts; /* recv buffer autoscaling timestamp */ 272 int rfbuf_cnt; /* recv buffer autoscaling byte count */ 273 274 int t_keepinit; /* time to establish connection */ 275 276 int t_keepidle; /* time before keepalive probes begin */ 277 int t_keepintvl; /* time between keepalive probes */ 278 int t_keepcnt; /* maximum number of keepalive probes */ 279 int t_maxidle; /* time to drop after starting probes */ 280 281 int t_rxtsyn; /* time spent in SYN or SYN|ACK rexmt */ 282 }; 283 284 #define IN_FASTRECOVERY(tp) (tp->t_flags & TF_FASTRECOVERY) 285 #define ENTER_FASTRECOVERY(tp) tp->t_flags |= TF_FASTRECOVERY 286 #define EXIT_FASTRECOVERY(tp) tp->t_flags &= ~TF_FASTRECOVERY 287 288 #ifdef TCP_SIGNATURE 289 /* 290 * Defines which are needed by the xform_tcp module and tcp_[in|out]put 291 * for SADB verification and lookup. 292 */ 293 #define TCP_SIGLEN 16 /* length of computed digest in bytes */ 294 #define TCP_KEYLEN_MIN 1 /* minimum length of TCP-MD5 key */ 295 #define TCP_KEYLEN_MAX 80 /* maximum length of TCP-MD5 key */ 296 /* 297 * Only a single SA per host may be specified at this time. An SPI is 298 * needed in order for the KEY_ALLOCSA() lookup to work. 299 */ 300 #define TCP_SIG_SPI 0x1000 301 #endif /* TCP_SIGNATURE */ 302 303 /* 304 * TCP statistics. 305 */ 306 struct tcp_stats { 307 u_long tcps_connattempt; /* connections initiated */ 308 u_long tcps_accepts; /* connections accepted */ 309 u_long tcps_connects; /* connections established */ 310 u_long tcps_drops; /* connections dropped */ 311 u_long tcps_conndrops; /* embryonic connections dropped */ 312 u_long tcps_minmssdrops; /* average minmss too low drops */ 313 u_long tcps_closed; /* conn. closed (includes drops) */ 314 u_long tcps_segstimed; /* segs where we tried to get rtt */ 315 u_long tcps_rttupdated; /* times we succeeded */ 316 u_long tcps_delack; /* delayed acks sent */ 317 u_long tcps_timeoutdrop; /* conn. dropped in rxmt timeout */ 318 u_long tcps_rexmttimeo; /* retransmit timeouts */ 319 u_long tcps_persisttimeo; /* persist timeouts */ 320 u_long tcps_keeptimeo; /* keepalive timeouts */ 321 u_long tcps_keepprobe; /* keepalive probes sent */ 322 u_long tcps_keepdrops; /* connections dropped in keepalive */ 323 324 u_long tcps_sndtotal; /* total packets sent */ 325 u_long tcps_sndpack; /* data packets sent */ 326 u_long tcps_sndbyte; /* data bytes sent */ 327 u_long tcps_sndrexmitpack; /* data packets retransmitted */ 328 u_long tcps_sndrexmitbyte; /* data bytes retransmitted */ 329 u_long tcps_sndsackrtopack; /* packets sent by SACK after RTO */ 330 u_long tcps_sndsackrtobyte; /* bytes sent by SACK after RTO */ 331 u_long tcps_sndfastrexmit; /* Fast Retransmissions */ 332 u_long tcps_sndearlyrexmit; /* early Fast Retransmissions */ 333 u_long tcps_sndlimited; /* Limited Transmit packets */ 334 u_long tcps_sndrtobad; /* spurious RTO retransmissions */ 335 u_long tcps_sndfastrexmitbad; /* spurious Fast Retransmissions */ 336 u_long tcps_sndearlyrexmitbad; /* spurious early Fast Retransmissions, 337 a subset of tcps_sndfastrexmitbad */ 338 u_long tcps_eifeldetected; /* Eifel-detected spurious rexmits */ 339 u_long tcps_rttcantdetect; /* Eifel but not 1/2 RTT-detectable */ 340 u_long tcps_rttdetected; /* RTT-detected spurious RTO rexmits */ 341 u_long tcps_sndacks; /* ack-only packets sent */ 342 u_long tcps_sndprobe; /* window probes sent */ 343 u_long tcps_sndurg; /* packets sent with URG only */ 344 u_long tcps_sndwinup; /* window update-only packets sent */ 345 u_long tcps_sndctrl; /* control (SYN|FIN|RST) packets sent */ 346 u_long tcps_sndsackpack; /* packets sent by SACK recovery alg */ 347 u_long tcps_sndsackbyte; /* bytes sent by SACK recovery */ 348 u_long tcps_snduna3; /* re-retransmit snd_una on 3 new seg */ 349 u_long tcps_snduna1; /* re-retransmit snd_una on 1 new seg */ 350 u_long tcps_sndsackopt; /* SACK options sent */ 351 u_long tcps_snddsackopt; /* D-SACK options sent */ 352 353 u_long tcps_rcvtotal; /* total packets received */ 354 u_long tcps_rcvpack; /* packets received in sequence */ 355 u_long tcps_rcvbyte; /* bytes received in sequence */ 356 u_long tcps_rcvbadsum; /* packets received with ccksum errs */ 357 u_long tcps_rcvbadoff; /* packets received with bad offset */ 358 u_long tcps_rcvmemdrop; /* packets dropped for lack of memory */ 359 u_long tcps_rcvshort; /* packets received too short */ 360 u_long tcps_rcvduppack; /* duplicate-only packets received */ 361 u_long tcps_rcvdupbyte; /* duplicate-only bytes received */ 362 u_long tcps_rcvpartduppack; /* packets with some duplicate data */ 363 u_long tcps_rcvpartdupbyte; /* dup. bytes in part-dup. packets */ 364 u_long tcps_rcvoopack; /* out-of-order packets received */ 365 u_long tcps_rcvoobyte; /* out-of-order bytes received */ 366 u_long tcps_rcvpackafterwin; /* packets with data after window */ 367 u_long tcps_rcvbyteafterwin; /* bytes rcvd after window */ 368 u_long tcps_rcvafterclose; /* packets rcvd after "close" */ 369 u_long tcps_rcvwinprobe; /* rcvd window probe packets */ 370 u_long tcps_rcvdupack; /* rcvd duplicate acks */ 371 u_long tcps_rcvacktoomuch; /* rcvd acks for unsent data */ 372 u_long tcps_rcvackpack; /* rcvd ack packets */ 373 u_long tcps_rcvackbyte; /* bytes acked by rcvd acks */ 374 u_long tcps_rcvwinupd; /* rcvd window update packets */ 375 u_long tcps_pawsdrop; /* segments dropped due to PAWS */ 376 u_long tcps_predack; /* times hdr predict ok for acks */ 377 u_long tcps_preddat; /* times hdr predict ok for data pkts */ 378 u_long tcps_pcbcachemiss; 379 u_long tcps_cachedrtt; /* times cached RTT in route updated */ 380 u_long tcps_cachedrttvar; /* times cached rttvar updated */ 381 u_long tcps_cachedssthresh; /* times cached ssthresh updated */ 382 u_long tcps_usedrtt; /* times RTT initialized from route */ 383 u_long tcps_usedrttvar; /* times RTTVAR initialized from rt */ 384 u_long tcps_usedssthresh; /* times ssthresh initialized from rt*/ 385 u_long tcps_persistdrop; /* timeout in persist state */ 386 u_long tcps_badsyn; /* bogus SYN, e.g. premature ACK */ 387 u_long tcps_mturesent; /* resends due to MTU discovery */ 388 u_long tcps_listendrop; /* listen queue overflows */ 389 u_long tcps_rcvbadsackopt; /* rcvd illegal SACK options */ 390 391 u_long tcps_sacksbupdate; /* times SACK scoreboard updated */ 392 u_long tcps_sacksboverflow; /* times SACK scoreboard overflowed */ 393 u_long tcps_sacksbreused; /* times SACK sb-block reused */ 394 u_long tcps_sacksbfailed; /* times SACK sb update failed */ 395 u_long tcps_sacksbfast; /* timee SACK sb-block uses cache */ 396 397 u_long tcps_sc_added; /* entry added to syncache */ 398 u_long tcps_sc_retransmitted; /* syncache entry was retransmitted */ 399 u_long tcps_sc_dupsyn; /* duplicate SYN packet */ 400 u_long tcps_sc_dropped; /* could not reply to packet */ 401 u_long tcps_sc_completed; /* successful extraction of entry */ 402 u_long tcps_sc_bucketoverflow; /* syncache per-bucket limit hit */ 403 u_long tcps_sc_cacheoverflow; /* syncache cache limit hit */ 404 u_long tcps_sc_reset; /* RST removed entry from syncache */ 405 u_long tcps_sc_stale; /* timed out or listen socket gone */ 406 u_long tcps_sc_aborted; /* syncache entry aborted */ 407 u_long tcps_sc_badack; /* removed due to bad ACK */ 408 u_long tcps_sc_unreach; /* ICMP unreachable received */ 409 u_long tcps_sc_zonefail; /* zalloc() failed */ 410 u_long tcps_sc_sendcookie; /* SYN cookie sent */ 411 u_long tcps_sc_recvcookie; /* SYN cookie received */ 412 }; 413 414 #ifdef _KERNEL 415 416 #if defined(SMP) 417 #define tcpstat tcpstats_percpu[mycpuid] 418 #else 419 #define tcpstat tcpstats_percpu[0] 420 #endif 421 422 struct sockopt; 423 424 extern struct tcp_stats tcpstats_percpu[MAXCPU]; 425 426 static const int tcprexmtthresh = 3; 427 #endif 428 429 /* 430 * Structure to hold TCP options that are only used during segment 431 * processing (in tcp_input), but not held in the tcpcb. 432 * It's basically used to reduce the number of parameters 433 * to tcp_dooptions. 434 */ 435 struct tcpopt { 436 u_long to_flags; /* which options are present */ 437 #define TOF_TS 0x0001 /* timestamp */ 438 #define TOF_MSS 0x0010 439 #define TOF_SCALE 0x0020 440 #define TOF_SACK_PERMITTED 0x0040 441 #define TOF_SACK 0x0080 442 #define TOF_SIGNATURE 0x0100 /* signature option present */ 443 #define TOF_SIGLEN 0x0200 /* sigature length valid (RFC2385) */ 444 u_int32_t to_tsval; 445 u_int32_t to_tsecr; 446 u_int16_t to_mss; 447 u_int8_t to_requested_s_scale; 448 u_int8_t to_nsackblocks; 449 struct raw_sackblock *to_sackblocks; 450 }; 451 452 struct syncache { 453 inp_gen_t sc_inp_gencnt; /* pointer check */ 454 struct tcpcb *sc_tp; /* tcb for listening socket */ 455 struct mbuf *sc_ipopts; /* source route */ 456 struct in_conninfo sc_inc; /* addresses */ 457 #define sc_route sc_inc.inc_route 458 #define sc_route6 sc_inc.inc6_route 459 u_int32_t sc_tsrecent; 460 tcp_seq sc_irs; /* seq from peer */ 461 tcp_seq sc_iss; /* our ISS */ 462 u_long sc_rxttime; /* retransmit time */ 463 u_int16_t sc_rxtslot; /* retransmit counter */ 464 u_int16_t sc_peer_mss; /* peer's MSS */ 465 u_int16_t sc_wnd; /* advertised window */ 466 u_int8_t sc_requested_s_scale:4, 467 sc_request_r_scale:4; 468 u_int8_t sc_flags; 469 #define SCF_NOOPT 0x01 /* no TCP options */ 470 #define SCF_WINSCALE 0x02 /* negotiated window scaling */ 471 #define SCF_TIMESTAMP 0x04 /* negotiated timestamps */ 472 #define SCF_UNUSED 0x08 /* unused */ 473 #define SCF_UNREACH 0x10 /* icmp unreachable received */ 474 #define SCF_SACK_PERMITTED 0x20 /* saw SACK permitted option */ 475 #define SCF_SIGNATURE 0x40 /* send MD5 digests */ 476 #define SCF_MARKER 0x80 /* not a real entry */ 477 int sc_rxtused; /* time spent in SYN|ACK rxt */ 478 TAILQ_ENTRY(syncache) sc_hash; 479 TAILQ_ENTRY(syncache) sc_timerq; 480 }; 481 482 struct syncache_head { 483 TAILQ_HEAD(, syncache) sch_bucket; 484 u_int sch_length; 485 }; 486 487 #define intotcpcb(ip) ((struct tcpcb *)(ip)->inp_ppcb) 488 #define sototcpcb(so) (intotcpcb(sotoinpcb(so))) 489 490 /* 491 * The smoothed round-trip time and estimated variance 492 * are stored as fixed point numbers scaled by the values below. 493 * For convenience, these scales are also used in smoothing the average 494 * (smoothed = (1/scale)sample + ((scale-1)/scale)smoothed). 495 * With these scales, srtt has 3 bits to the right of the binary point, 496 * and thus an "ALPHA" of 0.875. rttvar has 2 bits to the right of the 497 * binary point, and is smoothed with an ALPHA of 0.75. 498 */ 499 #define TCP_RTT_SCALE 32 /* multiplier for srtt; 3 bits frac. */ 500 #define TCP_RTT_SHIFT 5 /* shift for srtt; 3 bits frac. */ 501 #define TCP_RTTVAR_SCALE 16 /* multiplier for rttvar; 2 bits */ 502 #define TCP_RTTVAR_SHIFT 4 /* shift for rttvar; 2 bits */ 503 #define TCP_DELTA_SHIFT 2 /* see tcp_input.c */ 504 505 /* 506 * The initial retransmission should happen at rtt + 4 * rttvar. 507 * Because of the way we do the smoothing, srtt and rttvar 508 * will each average +1/2 tick of bias. When we compute 509 * the retransmit timer, we want 1/2 tick of rounding and 510 * 1 extra tick because of +-1/2 tick uncertainty in the 511 * firing of the timer. The bias will give us exactly the 512 * 1.5 tick we need. But, because the bias is 513 * statistical, we have to test that we don't drop below 514 * the minimum feasible timer (which is 2 ticks). 515 * This version of the macro adapted from a paper by Lawrence 516 * Brakmo and Larry Peterson which outlines a problem caused 517 * by insufficient precision in the original implementation, 518 * which results in inappropriately large RTO values for very 519 * fast networks. 520 */ 521 #define TCP_REXMTVAL(tp) \ 522 max((tp)->t_rttmin, (((tp)->t_srtt >> (TCP_RTT_SHIFT - TCP_DELTA_SHIFT)) \ 523 + (tp)->t_rttvar) >> TCP_DELTA_SHIFT) 524 525 /* 526 * TCB structure exported to user-land via sysctl(3). 527 * Evil hack: declare only if in_pcb.h and sys/socketvar.h have been 528 * included. Not all of our clients do. 529 */ 530 #if defined(_NETINET_IN_PCB_H_) && defined(_SYS_SOCKETVAR_H_) 531 struct xtcpcb { 532 size_t xt_len; 533 struct inpcb xt_inp; 534 struct tcpcb xt_tp; 535 struct xsocket xt_socket; 536 u_quad_t xt_alignment_hack; 537 }; 538 #endif 539 540 /* 541 * Names for TCP sysctl objects 542 */ 543 #define TCPCTL_DO_RFC1323 1 /* use RFC-1323 extensions */ 544 /* 2 was TCPCTL_DO_RFC1644 */ 545 #define TCPCTL_MSSDFLT 3 /* MSS default */ 546 #define TCPCTL_STATS 4 /* statistics (read-only) */ 547 #define TCPCTL_RTTDFLT 5 /* default RTT estimate */ 548 #define TCPCTL_KEEPIDLE 6 /* keepalive idle timer */ 549 #define TCPCTL_KEEPINTVL 7 /* interval to send keepalives */ 550 #define TCPCTL_SENDSPACE 8 /* send buffer space */ 551 #define TCPCTL_RECVSPACE 9 /* receive buffer space */ 552 #define TCPCTL_KEEPINIT 10 /* timeout for establishing syn */ 553 #define TCPCTL_PCBLIST 11 /* list of all outstanding PCBs */ 554 #define TCPCTL_DELACKTIME 12 /* time before sending delayed ACK */ 555 #define TCPCTL_V6MSSDFLT 13 /* MSS default for IPv6 */ 556 #define TCPCTL_MAXID 14 557 558 #define TCPCTL_NAMES { \ 559 { 0, 0 }, \ 560 { "rfc1323", CTLTYPE_INT }, \ 561 { "reserved", CTLTYPE_INT}, /* was rfc1644 */ \ 562 { "mssdflt", CTLTYPE_INT }, \ 563 { "stats", CTLTYPE_STRUCT }, \ 564 { "rttdflt", CTLTYPE_INT }, \ 565 { "keepidle", CTLTYPE_INT }, \ 566 { "keepintvl", CTLTYPE_INT }, \ 567 { "sendspace", CTLTYPE_INT }, \ 568 { "recvspace", CTLTYPE_INT }, \ 569 { "keepinit", CTLTYPE_INT }, \ 570 { "pcblist", CTLTYPE_STRUCT }, \ 571 { "delacktime", CTLTYPE_INT }, \ 572 { "v6mssdflt", CTLTYPE_INT }, \ 573 } 574 575 #ifdef _KERNEL 576 #ifdef SYSCTL_DECL 577 SYSCTL_DECL(_net_inet_tcp); 578 #endif 579 580 #define TCP_DO_SACK(tp) ((tp)->t_flags & TF_SACK_PERMITTED) 581 #define TCP_SACK_BLKEND(len, thflags) \ 582 ((len) + (((thflags) & TH_FIN) != 0)) 583 584 TAILQ_HEAD(tcpcbackqhead,tcpcb); 585 586 extern struct inpcbinfo tcbinfo[]; 587 extern struct tcpcbackqhead tcpcbackq[]; 588 589 extern int tcp_mssdflt; /* XXX */ 590 extern int tcp_minmss; 591 extern int tcp_delack_enabled; 592 extern int path_mtu_discovery; 593 594 union netmsg; 595 596 int tcp_addrcpu(in_addr_t faddr, in_port_t fport, 597 in_addr_t laddr, in_port_t lport); 598 struct lwkt_port * 599 tcp_addrport(in_addr_t faddr, in_port_t fport, 600 in_addr_t laddr, in_port_t lport); 601 struct lwkt_port *tcp_addrport0(void); 602 void tcp_canceltimers (struct tcpcb *); 603 struct tcpcb * 604 tcp_close (struct tcpcb *); 605 void tcp_ctlinput(union netmsg *); 606 void tcp_ctloutput(union netmsg *); 607 struct tcpcb * 608 tcp_drop (struct tcpcb *, int); 609 void tcp_drain (void); 610 void tcp_fasttimo (void); 611 void tcp_init (void); 612 void tcp_thread_init (void); 613 int tcp_input (struct mbuf **, int *, int); 614 void tcp_mss (struct tcpcb *, int); 615 int tcp_mssopt (struct tcpcb *); 616 void tcp_drop_syn_sent (struct inpcb *, int); 617 void tcp_mtudisc (struct inpcb *, int); 618 struct tcpcb * 619 tcp_newtcpcb (struct inpcb *); 620 int tcp_output (struct tcpcb *); 621 void tcp_quench (struct inpcb *, int); 622 void tcp_respond (struct tcpcb *, void *, 623 struct tcphdr *, struct mbuf *, tcp_seq, tcp_seq, int); 624 struct rtentry * 625 tcp_rtlookup (struct in_conninfo *); 626 int tcp_sack_bytes_below(struct scoreboard *scb, tcp_seq seq); 627 void tcp_sack_destroy(struct scoreboard *scb); 628 void tcp_sack_cleanup(struct scoreboard *scb); 629 void tcp_sack_report_cleanup(struct tcpcb *tp); 630 int tcp_sack_ndsack_blocks(struct raw_sackblock *blocks, 631 const int numblocks, tcp_seq snd_una); 632 void tcp_sack_fill_report(struct tcpcb *tp, u_char *opt, u_int *plen); 633 boolean_t 634 tcp_sack_has_sacked(struct scoreboard *scb, u_int amount); 635 void tcp_sack_tcpcb_init(struct tcpcb *tp); 636 uint32_t tcp_sack_compute_pipe(struct tcpcb *tp); 637 boolean_t 638 tcp_sack_nextseg(struct tcpcb *tp, tcp_seq *nextrexmt, uint32_t *len, 639 boolean_t *losdup); 640 #ifdef later 641 void tcp_sack_revert_scoreboard(struct scoreboard *scb, tcp_seq snd_una, 642 u_int maxseg); 643 void tcp_sack_save_scoreboard(struct scoreboard *scb); 644 #endif 645 void tcp_sack_skip_sacked(struct scoreboard *scb, tcp_seq *prexmt); 646 void tcp_sack_update_scoreboard(struct tcpcb *tp, struct tcpopt *to); 647 void tcp_save_congestion_state(struct tcpcb *tp); 648 void tcp_revert_congestion_state(struct tcpcb *tp); 649 void tcp_setpersist (struct tcpcb *); 650 struct tcptemp *tcp_maketemplate (struct tcpcb *); 651 void tcp_freetemplate (struct tcptemp *); 652 void tcp_fillheaders (struct tcpcb *, void *, void *); 653 struct lwkt_port * 654 tcp_soport(struct socket *, struct sockaddr *, struct mbuf **); 655 struct lwkt_port * 656 tcp_ctlport(int, struct sockaddr *, void *); 657 struct tcpcb * 658 tcp_timers (struct tcpcb *, int); 659 void tcp_trace (short, short, struct tcpcb *, void *, struct tcphdr *, 660 int); 661 void tcp_xmit_bandwidth_limit(struct tcpcb *tp, tcp_seq ack_seq); 662 u_long tcp_initial_window(struct tcpcb *tp); 663 void tcp_timer_keep_activity(struct tcpcb *tp, int thflags); 664 void syncache_init(void); 665 void syncache_unreach(struct in_conninfo *, struct tcphdr *); 666 int syncache_expand(struct in_conninfo *, struct tcphdr *, 667 struct socket **, struct mbuf *); 668 int syncache_add(struct in_conninfo *, struct tcpopt *, 669 struct tcphdr *, struct socket *, struct mbuf *); 670 void syncache_chkrst(struct in_conninfo *, struct tcphdr *); 671 void syncache_badack(struct in_conninfo *); 672 void syncache_destroy(struct tcpcb *tp); 673 674 #ifdef TCP_SIGNATURE 675 int tcpsignature_apply(void *fstate, void *data, unsigned int len); 676 int tcpsignature_compute(struct mbuf *m, int len, int tcpoptlen, 677 u_char *buf, u_int direction); 678 #endif /* TCP_SIGNATURE */ 679 680 extern struct pr_usrreqs tcp_usrreqs; 681 extern u_long tcp_sendspace; 682 extern u_long tcp_recvspace; 683 tcp_seq tcp_new_isn (struct tcpcb *); 684 685 #endif /* _KERNEL */ 686 687 #endif /* _NETINET_TCP_VAR_H_ */ 688