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) 2003, 2004 Jeffrey M. Hsu. All rights reserved. 36 * 37 * License terms: all terms for the DragonFly license above plus the following: 38 * 39 * 4. All advertising materials mentioning features or use of this software 40 * must display the following acknowledgement: 41 * 42 * This product includes software developed by Jeffrey M. Hsu 43 * for the DragonFly Project. 44 * 45 * This requirement may be waived with permission from Jeffrey Hsu. 46 * This requirement will sunset and may be removed on July 8 2005, 47 * after which the standard DragonFly license (as shown above) will 48 * apply. 49 */ 50 51 /* 52 * Copyright (c) 1982, 1986, 1993, 1994, 1995 53 * The Regents of the University of California. All rights reserved. 54 * 55 * Redistribution and use in source and binary forms, with or without 56 * modification, are permitted provided that the following conditions 57 * are met: 58 * 1. Redistributions of source code must retain the above copyright 59 * notice, this list of conditions and the following disclaimer. 60 * 2. Redistributions in binary form must reproduce the above copyright 61 * notice, this list of conditions and the following disclaimer in the 62 * documentation and/or other materials provided with the distribution. 63 * 3. All advertising materials mentioning features or use of this software 64 * must display the following acknowledgement: 65 * This product includes software developed by the University of 66 * California, Berkeley and its contributors. 67 * 4. Neither the name of the University nor the names of its contributors 68 * may be used to endorse or promote products derived from this software 69 * without specific prior written permission. 70 * 71 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 72 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 73 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 74 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 75 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 76 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 77 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 78 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 79 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 80 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 81 * SUCH DAMAGE. 82 * 83 * @(#)tcp_var.h 8.4 (Berkeley) 5/24/95 84 * $FreeBSD: src/sys/netinet/tcp_var.h,v 1.56.2.13 2003/02/03 02:34:07 hsu Exp $ 85 * $DragonFly: src/sys/netinet/tcp_var.h,v 1.23 2004/07/08 22:07:35 hsu Exp $ 86 */ 87 88 #ifndef _NETINET_TCP_VAR_H_ 89 #define _NETINET_TCP_VAR_H_ 90 91 #include <netinet/in_pcb.h> /* needed for in_conninfo, inp_gen_t */ 92 93 /* 94 * Kernel variables for tcp. 95 */ 96 extern int tcp_do_rfc1323; 97 extern int tcp_do_rfc1644; 98 99 /* TCP segment queue entry */ 100 struct tseg_qent { 101 LIST_ENTRY(tseg_qent) tqe_q; 102 int tqe_len; /* TCP segment data length */ 103 struct tcphdr *tqe_th; /* a pointer to tcp header */ 104 struct mbuf *tqe_m; /* mbuf contains packet */ 105 }; 106 LIST_HEAD(tsegqe_head, tseg_qent); 107 extern int tcp_reass_maxseg; 108 extern int tcp_reass_qsize; 109 #ifdef MALLOC_DECLARE 110 MALLOC_DECLARE(M_TSEGQ); 111 #endif 112 113 struct tcptemp { 114 u_char tt_ipgen[40]; /* the size must be of max ip header, now IPv6 */ 115 struct tcphdr tt_t; 116 }; 117 118 #define tcp6cb tcpcb /* for KAME src sync over BSD*'s */ 119 120 /* 121 * Tcp control block, one per tcp; fields: 122 * Organized for 16 byte cacheline efficiency. 123 */ 124 struct tcpcb { 125 struct tsegqe_head t_segq; 126 int t_dupacks; /* consecutive dup acks recd */ 127 struct tcptemp *unused; /* unused */ 128 struct callout *tt_rexmt; /* retransmit timer */ 129 130 struct callout *tt_persist; /* retransmit persistence */ 131 struct callout *tt_keep; /* keepalive */ 132 struct callout *tt_2msl; /* 2*msl TIME_WAIT timer */ 133 struct callout *tt_delack; /* delayed ACK timer */ 134 135 struct inpcb *t_inpcb; /* back pointer to internet pcb */ 136 int t_state; /* state of this connection */ 137 u_int t_flags; 138 #define TF_ACKNOW 0x00000001 /* ack peer immediately */ 139 #define TF_DELACK 0x00000002 /* ack, but try to delay it */ 140 #define TF_NODELAY 0x00000004 /* don't delay packets to coalesce */ 141 #define TF_NOOPT 0x00000008 /* don't use tcp options */ 142 #define TF_SENTFIN 0x00000010 /* have sent FIN */ 143 #define TF_REQ_SCALE 0x00000020 /* have/will request window scaling */ 144 #define TF_RCVD_SCALE 0x00000040 /* other side has requested scaling */ 145 #define TF_REQ_TSTMP 0x00000080 /* have/will request timestamps */ 146 #define TF_RCVD_TSTMP 0x00000100 /* a timestamp was received in SYN */ 147 #define TF_SACK_PERMIT 0x00000200 /* other side said I could SACK */ 148 #define TF_NEEDSYN 0x00000400 /* send SYN (implicit state) */ 149 #define TF_NEEDFIN 0x00000800 /* send FIN (implicit state) */ 150 #define TF_NOPUSH 0x00001000 /* don't push */ 151 #define TF_REQ_CC 0x00002000 /* have/will request CC */ 152 #define TF_RCVD_CC 0x00004000 /* a CC was received in SYN */ 153 #define TF_SENDCCNEW 0x00008000 /* send CCnew instead of CC in SYN */ 154 #define TF_MORETOCOME 0x00010000 /* More data to be appended to sock */ 155 #define TF_LQ_OVERFLOW 0x00020000 /* listen queue overflow */ 156 #define TF_LASTIDLE 0x00040000 /* connection was previously idle */ 157 #define TF_RXWIN0SENT 0x00080000 /* sent a receiver win 0 in response */ 158 #define TF_FASTRECOVERY 0x00100000 /* in NewReno Fast Recovery */ 159 #define TF_WASFRECOVERY 0x00200000 /* was in NewReno Fast Recovery */ 160 #define TF_FIRSTACCACK 0x00400000 /* Look for 1st acceptable ACK. */ 161 #define TF_FASTREXMT 0x00800000 /* Did Fast Retransmit. */ 162 #define TF_EARLYREXMT 0x01000000 /* Did Early (Fast) Retransmit. */ 163 #define TF_FORCE 0x02000000 /* Set if forcing out a byte */ 164 tcp_seq snd_up; /* send urgent pointer */ 165 166 tcp_seq snd_una; /* send unacknowledged */ 167 tcp_seq snd_recover; /* for use with NewReno Fast Recovery */ 168 tcp_seq snd_max; /* highest sequence number sent; 169 * used to recognize retransmits 170 */ 171 tcp_seq snd_nxt; /* send next */ 172 173 tcp_seq snd_wl1; /* window update seg seq number */ 174 tcp_seq snd_wl2; /* window update seg ack number */ 175 tcp_seq iss; /* initial send sequence number */ 176 tcp_seq irs; /* initial receive sequence number */ 177 178 tcp_seq rcv_nxt; /* receive next */ 179 tcp_seq rcv_adv; /* advertised window */ 180 u_long rcv_wnd; /* receive window */ 181 tcp_seq rcv_up; /* receive urgent pointer */ 182 183 u_long snd_wnd; /* send window */ 184 u_long snd_cwnd; /* congestion-controlled window */ 185 u_long snd_bwnd; /* bandwidth-controlled window */ 186 u_long snd_ssthresh; /* snd_cwnd size threshold for 187 * for slow start exponential to 188 * linear switch 189 */ 190 u_long snd_bandwidth; /* calculated bandwidth or 0 */ 191 192 u_int t_maxopd; /* mss plus options */ 193 194 u_long t_rcvtime; /* inactivity time */ 195 u_long t_starttime; /* time connection was established */ 196 int t_rtttime; /* round trip time */ 197 tcp_seq t_rtseq; /* sequence number being timed */ 198 199 int t_bw_rtttime; /* used for bandwidth calculation */ 200 tcp_seq t_bw_rtseq; /* used for bandwidth calculation */ 201 202 int t_rxtcur; /* current retransmit value (ticks) */ 203 u_int t_maxseg; /* maximum segment size */ 204 int t_srtt; /* smoothed round-trip time */ 205 int t_rttvar; /* variance in round-trip time */ 206 207 int t_rxtshift; /* log(2) of rexmt exp. backoff */ 208 u_int t_rttmin; /* minimum rtt allowed */ 209 u_int t_rttbest; /* best rtt we've seen */ 210 u_long t_rttupdated; /* number of times rtt sampled */ 211 u_long max_sndwnd; /* largest window peer has offered */ 212 213 int t_softerror; /* possible error not yet reported */ 214 /* out-of-band data */ 215 char t_oobflags; /* have some */ 216 char t_iobc; /* input character */ 217 #define TCPOOB_HAVEDATA 0x01 218 #define TCPOOB_HADDATA 0x02 219 /* RFC 1323 variables */ 220 u_char snd_scale; /* window scaling for send window */ 221 u_char rcv_scale; /* window scaling for recv window */ 222 u_char request_r_scale; /* pending window scaling */ 223 u_char requested_s_scale; 224 u_long ts_recent; /* timestamp echo data */ 225 226 u_long ts_recent_age; /* when last updated */ 227 tcp_seq last_ack_sent; 228 /* RFC 1644 variables */ 229 tcp_cc cc_send; /* send connection count */ 230 tcp_cc cc_recv; /* receive connection count */ 231 /* experimental */ 232 u_long snd_cwnd_prev; /* cwnd prior to retransmit */ 233 u_long snd_ssthresh_prev; /* ssthresh prior to retransmit */ 234 tcp_seq snd_recover_prev; /* snd_recover prior to retransmit */ 235 u_long t_badrxtwin; /* window for retransmit recovery */ 236 u_long t_rexmtTS; /* timestamp of last retransmit */ 237 u_char snd_limited; /* segments limited transmitted */ 238 }; 239 240 #define IN_FASTRECOVERY(tp) (tp->t_flags & TF_FASTRECOVERY) 241 #define ENTER_FASTRECOVERY(tp) tp->t_flags |= TF_FASTRECOVERY 242 #define EXIT_FASTRECOVERY(tp) tp->t_flags &= ~TF_FASTRECOVERY 243 244 #ifdef _KERNEL 245 246 #if defined(SMP) 247 #define _GD mycpu 248 #define tcpstat tcpstats_ary[_GD->gd_cpuid] 249 #else /* !SMP */ 250 #define tcpstat tcpstats_ary[0] 251 #endif 252 253 struct tcp_stats; 254 extern struct tcp_stats tcpstats_ary[MAXCPU]; 255 #endif 256 257 /* 258 * TCP statistics. 259 */ 260 struct tcp_stats { 261 u_long tcps_connattempt; /* connections initiated */ 262 u_long tcps_accepts; /* connections accepted */ 263 u_long tcps_connects; /* connections established */ 264 u_long tcps_drops; /* connections dropped */ 265 u_long tcps_conndrops; /* embryonic connections dropped */ 266 u_long tcps_closed; /* conn. closed (includes drops) */ 267 u_long tcps_segstimed; /* segs where we tried to get rtt */ 268 u_long tcps_rttupdated; /* times we succeeded */ 269 u_long tcps_delack; /* delayed acks sent */ 270 u_long tcps_timeoutdrop; /* conn. dropped in rxmt timeout */ 271 u_long tcps_rexmttimeo; /* retransmit timeouts */ 272 u_long tcps_persisttimeo; /* persist timeouts */ 273 u_long tcps_keeptimeo; /* keepalive timeouts */ 274 u_long tcps_keepprobe; /* keepalive probes sent */ 275 u_long tcps_keepdrops; /* connections dropped in keepalive */ 276 277 u_long tcps_sndtotal; /* total packets sent */ 278 u_long tcps_sndpack; /* data packets sent */ 279 u_long tcps_sndbyte; /* data bytes sent */ 280 u_long tcps_sndrexmitpack; /* data packets retransmitted */ 281 u_long tcps_sndrexmitbyte; /* data bytes retransmitted */ 282 u_long tcps_sndfastrexmit; /* Fast Retransmissions */ 283 u_long tcps_sndearlyrexmit; /* early Fast Retransmissions */ 284 u_long tcps_sndlimited; /* Limited Transmit packets */ 285 u_long tcps_sndrtobad; /* spurious RTO retransmissions */ 286 u_long tcps_sndfastrexmitbad; /* spurious Fast Retransmissions */ 287 u_long tcps_sndearlyrexmitbad; /* spurious early Fast Retransmissions, 288 a subset of tcps_sndfastrexmitbad */ 289 u_long tcps_eifeldetected; /* Eifel-detected spurious rexmits */ 290 u_long tcps_rttcantdetect; /* Eifel but not 1/2 RTT-detectable */ 291 u_long tcps_rttdetected; /* RTT-detected spurious RTO rexmits */ 292 u_long tcps_sndacks; /* ack-only packets sent */ 293 u_long tcps_sndprobe; /* window probes sent */ 294 u_long tcps_sndurg; /* packets sent with URG only */ 295 u_long tcps_sndwinup; /* window update-only packets sent */ 296 u_long tcps_sndctrl; /* control (SYN|FIN|RST) packets sent */ 297 298 u_long tcps_rcvtotal; /* total packets received */ 299 u_long tcps_rcvpack; /* packets received in sequence */ 300 u_long tcps_rcvbyte; /* bytes received in sequence */ 301 u_long tcps_rcvbadsum; /* packets received with ccksum errs */ 302 u_long tcps_rcvbadoff; /* packets received with bad offset */ 303 u_long tcps_rcvmemdrop; /* packets dropped for lack of memory */ 304 u_long tcps_rcvshort; /* packets received too short */ 305 u_long tcps_rcvduppack; /* duplicate-only packets received */ 306 u_long tcps_rcvdupbyte; /* duplicate-only bytes received */ 307 u_long tcps_rcvpartduppack; /* packets with some duplicate data */ 308 u_long tcps_rcvpartdupbyte; /* dup. bytes in part-dup. packets */ 309 u_long tcps_rcvoopack; /* out-of-order packets received */ 310 u_long tcps_rcvoobyte; /* out-of-order bytes received */ 311 u_long tcps_rcvpackafterwin; /* packets with data after window */ 312 u_long tcps_rcvbyteafterwin; /* bytes rcvd after window */ 313 u_long tcps_rcvafterclose; /* packets rcvd after "close" */ 314 u_long tcps_rcvwinprobe; /* rcvd window probe packets */ 315 u_long tcps_rcvdupack; /* rcvd duplicate acks */ 316 u_long tcps_rcvacktoomuch; /* rcvd acks for unsent data */ 317 u_long tcps_rcvackpack; /* rcvd ack packets */ 318 u_long tcps_rcvackbyte; /* bytes acked by rcvd acks */ 319 u_long tcps_rcvwinupd; /* rcvd window update packets */ 320 u_long tcps_pawsdrop; /* segments dropped due to PAWS */ 321 u_long tcps_predack; /* times hdr predict ok for acks */ 322 u_long tcps_preddat; /* times hdr predict ok for data pkts */ 323 u_long tcps_pcbcachemiss; 324 u_long tcps_cachedrtt; /* times cached RTT in route updated */ 325 u_long tcps_cachedrttvar; /* times cached rttvar updated */ 326 u_long tcps_cachedssthresh; /* times cached ssthresh updated */ 327 u_long tcps_usedrtt; /* times RTT initialized from route */ 328 u_long tcps_usedrttvar; /* times RTTVAR initialized from rt */ 329 u_long tcps_usedssthresh; /* times ssthresh initialized from rt*/ 330 u_long tcps_persistdrop; /* timeout in persist state */ 331 u_long tcps_badsyn; /* bogus SYN, e.g. premature ACK */ 332 u_long tcps_mturesent; /* resends due to MTU discovery */ 333 u_long tcps_listendrop; /* listen queue overflows */ 334 335 u_long tcps_sc_added; /* entry added to syncache */ 336 u_long tcps_sc_retransmitted; /* syncache entry was retransmitted */ 337 u_long tcps_sc_dupsyn; /* duplicate SYN packet */ 338 u_long tcps_sc_dropped; /* could not reply to packet */ 339 u_long tcps_sc_completed; /* successful extraction of entry */ 340 u_long tcps_sc_bucketoverflow; /* syncache per-bucket limit hit */ 341 u_long tcps_sc_cacheoverflow; /* syncache cache limit hit */ 342 u_long tcps_sc_reset; /* RST removed entry from syncache */ 343 u_long tcps_sc_stale; /* timed out or listen socket gone */ 344 u_long tcps_sc_aborted; /* syncache entry aborted */ 345 u_long tcps_sc_badack; /* removed due to bad ACK */ 346 u_long tcps_sc_unreach; /* ICMP unreachable received */ 347 u_long tcps_sc_zonefail; /* zalloc() failed */ 348 u_long tcps_sc_sendcookie; /* SYN cookie sent */ 349 u_long tcps_sc_recvcookie; /* SYN cookie received */ 350 }; 351 352 /* 353 * Structure to hold TCP options that are only used during segment 354 * processing (in tcp_input), but not held in the tcpcb. 355 * It's basically used to reduce the number of parameters 356 * to tcp_dooptions. 357 */ 358 struct tcpopt { 359 u_long to_flags; /* which options are present */ 360 #define TOF_TS 0x0001 /* timestamp */ 361 #define TOF_CC 0x0002 /* CC and CCnew are exclusive */ 362 #define TOF_CCNEW 0x0004 363 #define TOF_CCECHO 0x0008 364 #define TOF_MSS 0x0010 365 #define TOF_SCALE 0x0020 366 u_int32_t to_tsval; 367 u_int32_t to_tsecr; 368 tcp_cc to_cc; /* holds CC or CCnew */ 369 tcp_cc to_ccecho; 370 u_int16_t to_mss; 371 u_int8_t to_requested_s_scale; 372 u_int8_t to_pad; 373 }; 374 375 struct syncache { 376 inp_gen_t sc_inp_gencnt; /* pointer check */ 377 struct tcpcb *sc_tp; /* tcb for listening socket */ 378 struct mbuf *sc_ipopts; /* source route */ 379 struct in_conninfo sc_inc; /* addresses */ 380 #define sc_route sc_inc.inc_route 381 #define sc_route6 sc_inc.inc6_route 382 u_int32_t sc_tsrecent; 383 tcp_cc sc_cc_send; /* holds CC or CCnew */ 384 tcp_cc sc_cc_recv; 385 tcp_seq sc_irs; /* seq from peer */ 386 tcp_seq sc_iss; /* our ISS */ 387 u_long sc_rxttime; /* retransmit time */ 388 u_int16_t sc_rxtslot; /* retransmit counter */ 389 u_int16_t sc_peer_mss; /* peer's MSS */ 390 u_int16_t sc_wnd; /* advertised window */ 391 u_int8_t sc_requested_s_scale:4, 392 sc_request_r_scale:4; 393 u_int8_t sc_flags; 394 #define SCF_NOOPT 0x01 /* no TCP options */ 395 #define SCF_WINSCALE 0x02 /* negotiated window scaling */ 396 #define SCF_TIMESTAMP 0x04 /* negotiated timestamps */ 397 #define SCF_CC 0x08 /* negotiated CC */ 398 #define SCF_UNREACH 0x10 /* icmp unreachable received */ 399 #define SCF_KEEPROUTE 0x20 /* keep cloned route */ 400 TAILQ_ENTRY(syncache) sc_hash; 401 TAILQ_ENTRY(syncache) sc_timerq; 402 }; 403 404 struct syncache_head { 405 TAILQ_HEAD(, syncache) sch_bucket; 406 u_int sch_length; 407 }; 408 409 /* 410 * The TAO cache entry which is stored in the protocol family specific 411 * portion of the route metrics. 412 */ 413 struct rmxp_tao { 414 tcp_cc tao_cc; /* latest CC in valid SYN */ 415 tcp_cc tao_ccsent; /* latest CC sent to peer */ 416 u_short tao_mssopt; /* peer's cached MSS */ 417 #ifdef notyet 418 u_short tao_flags; /* cache status flags */ 419 #define TAOF_DONT 0x0001 /* peer doesn't understand rfc1644 */ 420 #define TAOF_OK 0x0002 /* peer does understand rfc1644 */ 421 #define TAOF_UNDEF 0 /* we don't know yet */ 422 #endif /* notyet */ 423 }; 424 #define rmx_taop(r) ((struct rmxp_tao *)(r).rmx_filler) 425 426 #define intotcpcb(ip) ((struct tcpcb *)(ip)->inp_ppcb) 427 #define sototcpcb(so) (intotcpcb(sotoinpcb(so))) 428 429 /* 430 * The smoothed round-trip time and estimated variance 431 * are stored as fixed point numbers scaled by the values below. 432 * For convenience, these scales are also used in smoothing the average 433 * (smoothed = (1/scale)sample + ((scale-1)/scale)smoothed). 434 * With these scales, srtt has 3 bits to the right of the binary point, 435 * and thus an "ALPHA" of 0.875. rttvar has 2 bits to the right of the 436 * binary point, and is smoothed with an ALPHA of 0.75. 437 */ 438 #define TCP_RTT_SCALE 32 /* multiplier for srtt; 3 bits frac. */ 439 #define TCP_RTT_SHIFT 5 /* shift for srtt; 3 bits frac. */ 440 #define TCP_RTTVAR_SCALE 16 /* multiplier for rttvar; 2 bits */ 441 #define TCP_RTTVAR_SHIFT 4 /* shift for rttvar; 2 bits */ 442 #define TCP_DELTA_SHIFT 2 /* see tcp_input.c */ 443 444 /* 445 * The initial retransmission should happen at rtt + 4 * rttvar. 446 * Because of the way we do the smoothing, srtt and rttvar 447 * will each average +1/2 tick of bias. When we compute 448 * the retransmit timer, we want 1/2 tick of rounding and 449 * 1 extra tick because of +-1/2 tick uncertainty in the 450 * firing of the timer. The bias will give us exactly the 451 * 1.5 tick we need. But, because the bias is 452 * statistical, we have to test that we don't drop below 453 * the minimum feasible timer (which is 2 ticks). 454 * This version of the macro adapted from a paper by Lawrence 455 * Brakmo and Larry Peterson which outlines a problem caused 456 * by insufficient precision in the original implementation, 457 * which results in inappropriately large RTO values for very 458 * fast networks. 459 */ 460 #define TCP_REXMTVAL(tp) \ 461 max((tp)->t_rttmin, (((tp)->t_srtt >> (TCP_RTT_SHIFT - TCP_DELTA_SHIFT)) \ 462 + (tp)->t_rttvar) >> TCP_DELTA_SHIFT) 463 464 /* 465 * TCB structure exported to user-land via sysctl(3). 466 * Evil hack: declare only if in_pcb.h and sys/socketvar.h have been 467 * included. Not all of our clients do. 468 */ 469 #if defined(_NETINET_IN_PCB_H_) && defined(_SYS_SOCKETVAR_H_) 470 struct xtcpcb { 471 size_t xt_len; 472 struct inpcb xt_inp; 473 struct tcpcb xt_tp; 474 struct xsocket xt_socket; 475 u_quad_t xt_alignment_hack; 476 }; 477 #endif 478 479 /* 480 * Names for TCP sysctl objects 481 */ 482 #define TCPCTL_DO_RFC1323 1 /* use RFC-1323 extensions */ 483 #define TCPCTL_DO_RFC1644 2 /* use RFC-1644 extensions */ 484 #define TCPCTL_MSSDFLT 3 /* MSS default */ 485 #define TCPCTL_STATS 4 /* statistics (read-only) */ 486 #define TCPCTL_RTTDFLT 5 /* default RTT estimate */ 487 #define TCPCTL_KEEPIDLE 6 /* keepalive idle timer */ 488 #define TCPCTL_KEEPINTVL 7 /* interval to send keepalives */ 489 #define TCPCTL_SENDSPACE 8 /* send buffer space */ 490 #define TCPCTL_RECVSPACE 9 /* receive buffer space */ 491 #define TCPCTL_KEEPINIT 10 /* timeout for establishing syn */ 492 #define TCPCTL_PCBLIST 11 /* list of all outstanding PCBs */ 493 #define TCPCTL_DELACKTIME 12 /* time before sending delayed ACK */ 494 #define TCPCTL_V6MSSDFLT 13 /* MSS default for IPv6 */ 495 #define TCPCTL_MAXID 14 496 497 #define TCPCTL_NAMES { \ 498 { 0, 0 }, \ 499 { "rfc1323", CTLTYPE_INT }, \ 500 { "rfc1644", CTLTYPE_INT }, \ 501 { "mssdflt", CTLTYPE_INT }, \ 502 { "stats", CTLTYPE_STRUCT }, \ 503 { "rttdflt", CTLTYPE_INT }, \ 504 { "keepidle", CTLTYPE_INT }, \ 505 { "keepintvl", CTLTYPE_INT }, \ 506 { "sendspace", CTLTYPE_INT }, \ 507 { "recvspace", CTLTYPE_INT }, \ 508 { "keepinit", CTLTYPE_INT }, \ 509 { "pcblist", CTLTYPE_STRUCT }, \ 510 { "delacktime", CTLTYPE_INT }, \ 511 { "v6mssdflt", CTLTYPE_INT }, \ 512 } 513 514 #ifdef _KERNEL 515 #ifdef SYSCTL_DECL 516 SYSCTL_DECL(_net_inet_tcp); 517 #endif 518 519 extern struct inpcbinfo tcbinfo[]; 520 extern int tcp_mssdflt; /* XXX */ 521 extern int tcp_delack_enabled; 522 extern int tcp_do_newreno; 523 extern int path_mtu_discovery; 524 extern int ss_fltsz; 525 extern int ss_fltsz_local; 526 527 int tcp_addrcpu(in_addr_t faddr, in_port_t fport, 528 in_addr_t laddr, in_port_t lport); 529 struct lwkt_port * 530 tcp_addrport(in_addr_t faddr, in_port_t fport, 531 in_addr_t laddr, in_port_t lport); 532 void tcp_canceltimers (struct tcpcb *); 533 struct tcpcb * 534 tcp_close (struct tcpcb *); 535 void tcp_ctlinput (int, struct sockaddr *, void *); 536 int tcp_ctloutput (struct socket *, struct sockopt *); 537 struct lwkt_port * 538 tcp_cport(int cpu); 539 struct tcpcb * 540 tcp_drop (struct tcpcb *, int); 541 void tcp_drain (void); 542 void tcp_fasttimo (void); 543 struct rmxp_tao * 544 tcp_gettaocache (struct in_conninfo *); 545 void tcp_init (void); 546 void tcp_thread_init (void); 547 void tcp_input (struct mbuf *, ...); 548 void tcp_mss (struct tcpcb *, int); 549 int tcp_mssopt (struct tcpcb *); 550 void tcp_drop_syn_sent (struct inpcb *, int); 551 void tcp_mtudisc (struct inpcb *, int); 552 struct tcpcb * 553 tcp_newtcpcb (struct inpcb *); 554 int tcp_output (struct tcpcb *); 555 void tcp_quench (struct inpcb *, int); 556 void tcp_respond (struct tcpcb *, void *, 557 struct tcphdr *, struct mbuf *, tcp_seq, tcp_seq, int); 558 struct rtentry * 559 tcp_rtlookup (struct in_conninfo *); 560 void tcp_save_congestion_state(struct tcpcb *tp); 561 void tcp_revert_congestion_state(struct tcpcb *tp); 562 void tcp_setpersist (struct tcpcb *); 563 void tcp_slowtimo (void); 564 struct tcptemp *tcp_maketemplate (struct tcpcb *); 565 void tcp_freetemplate (struct tcptemp *); 566 void tcp_fillheaders (struct tcpcb *, void *, void *); 567 struct lwkt_port * 568 tcp_soport(struct socket *, struct sockaddr *nam, int req); 569 struct tcpcb * 570 tcp_timers (struct tcpcb *, int); 571 void tcp_trace (int, int, struct tcpcb *, void *, struct tcphdr *, 572 int); 573 void tcp_xmit_bandwidth_limit(struct tcpcb *tp, tcp_seq ack_seq); 574 void syncache_init(void); 575 void syncache_unreach(struct in_conninfo *, struct tcphdr *); 576 int syncache_expand(struct in_conninfo *, struct tcphdr *, 577 struct socket **, struct mbuf *); 578 int syncache_add(struct in_conninfo *, struct tcpopt *, 579 struct tcphdr *, struct socket **, struct mbuf *); 580 void syncache_chkrst(struct in_conninfo *, struct tcphdr *); 581 void syncache_badack(struct in_conninfo *); 582 583 extern struct pr_usrreqs tcp_usrreqs; 584 extern u_long tcp_sendspace; 585 extern u_long tcp_recvspace; 586 tcp_seq tcp_new_isn (struct tcpcb *); 587 588 #endif /* _KERNEL */ 589 590 #endif /* _NETINET_TCP_VAR_H_ */ 591