1 /*- 2 * Copyright (c) 1982, 1986, 1993, 1994, 1995 3 * The Regents of the University of California. All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 4. Neither the name of the University nor the names of its contributors 14 * may be used to endorse or promote products derived from this software 15 * without specific prior written permission. 16 * 17 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 20 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 27 * SUCH DAMAGE. 28 * 29 * @(#)tcp_var.h 8.4 (Berkeley) 5/24/95 30 * $FreeBSD$ 31 */ 32 33 #ifndef _NETINET_TCP_VAR_H_ 34 #define _NETINET_TCP_VAR_H_ 35 36 #include <netinet/tcp.h> 37 38 struct vnet; 39 40 /* 41 * Kernel variables for tcp. 42 */ 43 #ifdef VIMAGE_GLOBALS 44 extern int tcp_do_rfc1323; 45 #endif 46 47 /* TCP segment queue entry */ 48 struct tseg_qent { 49 LIST_ENTRY(tseg_qent) tqe_q; 50 int tqe_len; /* TCP segment data length */ 51 struct tcphdr *tqe_th; /* a pointer to tcp header */ 52 struct mbuf *tqe_m; /* mbuf contains packet */ 53 }; 54 LIST_HEAD(tsegqe_head, tseg_qent); 55 #ifdef VIMAGE_GLOBALS 56 extern int tcp_reass_qsize; 57 #endif 58 extern struct uma_zone *tcp_reass_zone; 59 60 struct sackblk { 61 tcp_seq start; /* start seq no. of sack block */ 62 tcp_seq end; /* end seq no. */ 63 }; 64 65 struct sackhole { 66 tcp_seq start; /* start seq no. of hole */ 67 tcp_seq end; /* end seq no. */ 68 tcp_seq rxmit; /* next seq. no in hole to be retransmitted */ 69 TAILQ_ENTRY(sackhole) scblink; /* scoreboard linkage */ 70 }; 71 72 struct sackhint { 73 struct sackhole *nexthole; 74 int sack_bytes_rexmit; 75 }; 76 77 struct tcptemp { 78 u_char tt_ipgen[40]; /* the size must be of max ip header, now IPv6 */ 79 struct tcphdr tt_t; 80 }; 81 82 #define tcp6cb tcpcb /* for KAME src sync over BSD*'s */ 83 84 /* Neighbor Discovery, Neighbor Unreachability Detection Upper layer hint. */ 85 #ifdef INET6 86 #define ND6_HINT(tp) \ 87 do { \ 88 if ((tp) && (tp)->t_inpcb && \ 89 ((tp)->t_inpcb->inp_vflag & INP_IPV6) != 0) \ 90 nd6_nud_hint(NULL, NULL, 0); \ 91 } while (0) 92 #else 93 #define ND6_HINT(tp) 94 #endif 95 96 /* 97 * Tcp control block, one per tcp; fields: 98 * Organized for 16 byte cacheline efficiency. 99 */ 100 struct tcpcb { 101 struct tsegqe_head t_segq; /* segment reassembly queue */ 102 int t_segqlen; /* segment reassembly queue length */ 103 int t_dupacks; /* consecutive dup acks recd */ 104 105 struct tcp_timer *t_timers; /* All the TCP timers in one struct */ 106 107 struct inpcb *t_inpcb; /* back pointer to internet pcb */ 108 int t_state; /* state of this connection */ 109 u_int t_flags; 110 111 struct vnet *t_vnet; /* back pointer to parent vnet */ 112 113 tcp_seq snd_una; /* send unacknowledged */ 114 tcp_seq snd_max; /* highest sequence number sent; 115 * used to recognize retransmits 116 */ 117 tcp_seq snd_nxt; /* send next */ 118 tcp_seq snd_up; /* send urgent pointer */ 119 120 tcp_seq snd_wl1; /* window update seg seq number */ 121 tcp_seq snd_wl2; /* window update seg ack number */ 122 tcp_seq iss; /* initial send sequence number */ 123 tcp_seq irs; /* initial receive sequence number */ 124 125 tcp_seq rcv_nxt; /* receive next */ 126 tcp_seq rcv_adv; /* advertised window */ 127 u_long rcv_wnd; /* receive window */ 128 tcp_seq rcv_up; /* receive urgent pointer */ 129 130 u_long snd_wnd; /* send window */ 131 u_long snd_cwnd; /* congestion-controlled window */ 132 u_long snd_bwnd; /* bandwidth-controlled window */ 133 u_long snd_ssthresh; /* snd_cwnd size threshold for 134 * for slow start exponential to 135 * linear switch 136 */ 137 u_long snd_bandwidth; /* calculated bandwidth or 0 */ 138 tcp_seq snd_recover; /* for use in NewReno Fast Recovery */ 139 140 u_int t_maxopd; /* mss plus options */ 141 142 u_long t_rcvtime; /* inactivity time */ 143 u_long t_starttime; /* time connection was established */ 144 int t_rtttime; /* round trip time */ 145 tcp_seq t_rtseq; /* sequence number being timed */ 146 147 int t_bw_rtttime; /* used for bandwidth calculation */ 148 tcp_seq t_bw_rtseq; /* used for bandwidth calculation */ 149 150 int t_rxtcur; /* current retransmit value (ticks) */ 151 u_int t_maxseg; /* maximum segment size */ 152 int t_srtt; /* smoothed round-trip time */ 153 int t_rttvar; /* variance in round-trip time */ 154 155 int t_rxtshift; /* log(2) of rexmt exp. backoff */ 156 u_int t_rttmin; /* minimum rtt allowed */ 157 u_int t_rttbest; /* best rtt we've seen */ 158 u_long t_rttupdated; /* number of times rtt sampled */ 159 u_long max_sndwnd; /* largest window peer has offered */ 160 161 int t_softerror; /* possible error not yet reported */ 162 /* out-of-band data */ 163 char t_oobflags; /* have some */ 164 char t_iobc; /* input character */ 165 /* RFC 1323 variables */ 166 u_char snd_scale; /* window scaling for send window */ 167 u_char rcv_scale; /* window scaling for recv window */ 168 u_char request_r_scale; /* pending window scaling */ 169 u_int32_t ts_recent; /* timestamp echo data */ 170 u_long ts_recent_age; /* when last updated */ 171 u_int32_t ts_offset; /* our timestamp offset */ 172 173 tcp_seq last_ack_sent; 174 /* experimental */ 175 u_long snd_cwnd_prev; /* cwnd prior to retransmit */ 176 u_long snd_ssthresh_prev; /* ssthresh prior to retransmit */ 177 tcp_seq snd_recover_prev; /* snd_recover prior to retransmit */ 178 u_long t_badrxtwin; /* window for retransmit recovery */ 179 u_char snd_limited; /* segments limited transmitted */ 180 /* SACK related state */ 181 int snd_numholes; /* number of holes seen by sender */ 182 TAILQ_HEAD(sackhole_head, sackhole) snd_holes; 183 /* SACK scoreboard (sorted) */ 184 tcp_seq snd_fack; /* last seq number(+1) sack'd by rcv'r*/ 185 int rcv_numsacks; /* # distinct sack blks present */ 186 struct sackblk sackblks[MAX_SACK_BLKS]; /* seq nos. of sack blocks */ 187 tcp_seq sack_newdata; /* New data xmitted in this recovery 188 episode starts at this seq number */ 189 struct sackhint sackhint; /* SACK scoreboard hint */ 190 int t_rttlow; /* smallest observerved RTT */ 191 u_int32_t rfbuf_ts; /* recv buffer autoscaling timestamp */ 192 int rfbuf_cnt; /* recv buffer autoscaling byte count */ 193 void *t_pspare[3]; /* toe usrreqs / toepcb * / congestion algo / 1 general use */ 194 struct toe_usrreqs *t_tu; /* offload operations vector */ 195 void *t_toe; /* TOE pcb pointer */ 196 int t_bytes_acked; /* # bytes acked during current RTT */ 197 }; 198 199 /* 200 * Flags and utility macros for the t_flags field. 201 */ 202 #define TF_ACKNOW 0x000001 /* ack peer immediately */ 203 #define TF_DELACK 0x000002 /* ack, but try to delay it */ 204 #define TF_NODELAY 0x000004 /* don't delay packets to coalesce */ 205 #define TF_NOOPT 0x000008 /* don't use tcp options */ 206 #define TF_SENTFIN 0x000010 /* have sent FIN */ 207 #define TF_REQ_SCALE 0x000020 /* have/will request window scaling */ 208 #define TF_RCVD_SCALE 0x000040 /* other side has requested scaling */ 209 #define TF_REQ_TSTMP 0x000080 /* have/will request timestamps */ 210 #define TF_RCVD_TSTMP 0x000100 /* a timestamp was received in SYN */ 211 #define TF_SACK_PERMIT 0x000200 /* other side said I could SACK */ 212 #define TF_NEEDSYN 0x000400 /* send SYN (implicit state) */ 213 #define TF_NEEDFIN 0x000800 /* send FIN (implicit state) */ 214 #define TF_NOPUSH 0x001000 /* don't push */ 215 #define TF_MORETOCOME 0x010000 /* More data to be appended to sock */ 216 #define TF_LQ_OVERFLOW 0x020000 /* listen queue overflow */ 217 #define TF_LASTIDLE 0x040000 /* connection was previously idle */ 218 #define TF_RXWIN0SENT 0x080000 /* sent a receiver win 0 in response */ 219 #define TF_FASTRECOVERY 0x100000 /* in NewReno Fast Recovery */ 220 #define TF_WASFRECOVERY 0x200000 /* was in NewReno Fast Recovery */ 221 #define TF_SIGNATURE 0x400000 /* require MD5 digests (RFC2385) */ 222 #define TF_FORCEDATA 0x800000 /* force out a byte */ 223 #define TF_TSO 0x1000000 /* TSO enabled on this connection */ 224 #define TF_TOE 0x2000000 /* this connection is offloaded */ 225 #define TF_ECN_PERMIT 0x4000000 /* connection ECN-ready */ 226 #define TF_ECN_SND_CWR 0x8000000 /* ECN CWR in queue */ 227 #define TF_ECN_SND_ECE 0x10000000 /* ECN ECE in queue */ 228 229 #define IN_FASTRECOVERY(tp) (tp->t_flags & TF_FASTRECOVERY) 230 #define ENTER_FASTRECOVERY(tp) tp->t_flags |= TF_FASTRECOVERY 231 #define EXIT_FASTRECOVERY(tp) tp->t_flags &= ~TF_FASTRECOVERY 232 233 /* 234 * Flags for the t_oobflags field. 235 */ 236 #define TCPOOB_HAVEDATA 0x01 237 #define TCPOOB_HADDATA 0x02 238 239 #ifdef TCP_SIGNATURE 240 /* 241 * Defines which are needed by the xform_tcp module and tcp_[in|out]put 242 * for SADB verification and lookup. 243 */ 244 #define TCP_SIGLEN 16 /* length of computed digest in bytes */ 245 #define TCP_KEYLEN_MIN 1 /* minimum length of TCP-MD5 key */ 246 #define TCP_KEYLEN_MAX 80 /* maximum length of TCP-MD5 key */ 247 /* 248 * Only a single SA per host may be specified at this time. An SPI is 249 * needed in order for the KEY_ALLOCSA() lookup to work. 250 */ 251 #define TCP_SIG_SPI 0x1000 252 #endif /* TCP_SIGNATURE */ 253 254 /* 255 * Structure to hold TCP options that are only used during segment 256 * processing (in tcp_input), but not held in the tcpcb. 257 * It's basically used to reduce the number of parameters 258 * to tcp_dooptions and tcp_addoptions. 259 * The binary order of the to_flags is relevant for packing of the 260 * options in tcp_addoptions. 261 */ 262 struct tcpopt { 263 u_long to_flags; /* which options are present */ 264 #define TOF_MSS 0x0001 /* maximum segment size */ 265 #define TOF_SCALE 0x0002 /* window scaling */ 266 #define TOF_SACKPERM 0x0004 /* SACK permitted */ 267 #define TOF_TS 0x0010 /* timestamp */ 268 #define TOF_SIGNATURE 0x0040 /* TCP-MD5 signature option (RFC2385) */ 269 #define TOF_SACK 0x0080 /* Peer sent SACK option */ 270 #define TOF_MAXOPT 0x0100 271 u_int32_t to_tsval; /* new timestamp */ 272 u_int32_t to_tsecr; /* reflected timestamp */ 273 u_int16_t to_mss; /* maximum segment size */ 274 u_int8_t to_wscale; /* window scaling */ 275 u_int8_t to_nsacks; /* number of SACK blocks */ 276 u_char *to_sacks; /* pointer to the first SACK blocks */ 277 u_char *to_signature; /* pointer to the TCP-MD5 signature */ 278 }; 279 280 /* 281 * Flags for tcp_dooptions. 282 */ 283 #define TO_SYN 0x01 /* parse SYN-only options */ 284 285 struct hc_metrics_lite { /* must stay in sync with hc_metrics */ 286 u_long rmx_mtu; /* MTU for this path */ 287 u_long rmx_ssthresh; /* outbound gateway buffer limit */ 288 u_long rmx_rtt; /* estimated round trip time */ 289 u_long rmx_rttvar; /* estimated rtt variance */ 290 u_long rmx_bandwidth; /* estimated bandwidth */ 291 u_long rmx_cwnd; /* congestion window */ 292 u_long rmx_sendpipe; /* outbound delay-bandwidth product */ 293 u_long rmx_recvpipe; /* inbound delay-bandwidth product */ 294 }; 295 296 #ifndef _NETINET_IN_PCB_H_ 297 struct in_conninfo; 298 #endif /* _NETINET_IN_PCB_H_ */ 299 300 struct tcptw { 301 struct inpcb *tw_inpcb; /* XXX back pointer to internet pcb */ 302 tcp_seq snd_nxt; 303 tcp_seq rcv_nxt; 304 tcp_seq iss; 305 tcp_seq irs; 306 u_short last_win; /* cached window value */ 307 u_short tw_so_options; /* copy of so_options */ 308 struct ucred *tw_cred; /* user credentials */ 309 u_long t_recent; 310 u_int32_t ts_offset; /* our timestamp offset */ 311 u_long t_starttime; 312 int tw_time; 313 TAILQ_ENTRY(tcptw) tw_2msl; 314 }; 315 316 #define intotcpcb(ip) ((struct tcpcb *)(ip)->inp_ppcb) 317 #define intotw(ip) ((struct tcptw *)(ip)->inp_ppcb) 318 #define sototcpcb(so) (intotcpcb(sotoinpcb(so))) 319 320 /* 321 * The smoothed round-trip time and estimated variance 322 * are stored as fixed point numbers scaled by the values below. 323 * For convenience, these scales are also used in smoothing the average 324 * (smoothed = (1/scale)sample + ((scale-1)/scale)smoothed). 325 * With these scales, srtt has 3 bits to the right of the binary point, 326 * and thus an "ALPHA" of 0.875. rttvar has 2 bits to the right of the 327 * binary point, and is smoothed with an ALPHA of 0.75. 328 */ 329 #define TCP_RTT_SCALE 32 /* multiplier for srtt; 3 bits frac. */ 330 #define TCP_RTT_SHIFT 5 /* shift for srtt; 3 bits frac. */ 331 #define TCP_RTTVAR_SCALE 16 /* multiplier for rttvar; 2 bits */ 332 #define TCP_RTTVAR_SHIFT 4 /* shift for rttvar; 2 bits */ 333 #define TCP_DELTA_SHIFT 2 /* see tcp_input.c */ 334 335 /* 336 * The initial retransmission should happen at rtt + 4 * rttvar. 337 * Because of the way we do the smoothing, srtt and rttvar 338 * will each average +1/2 tick of bias. When we compute 339 * the retransmit timer, we want 1/2 tick of rounding and 340 * 1 extra tick because of +-1/2 tick uncertainty in the 341 * firing of the timer. The bias will give us exactly the 342 * 1.5 tick we need. But, because the bias is 343 * statistical, we have to test that we don't drop below 344 * the minimum feasible timer (which is 2 ticks). 345 * This version of the macro adapted from a paper by Lawrence 346 * Brakmo and Larry Peterson which outlines a problem caused 347 * by insufficient precision in the original implementation, 348 * which results in inappropriately large RTO values for very 349 * fast networks. 350 */ 351 #define TCP_REXMTVAL(tp) \ 352 max((tp)->t_rttmin, (((tp)->t_srtt >> (TCP_RTT_SHIFT - TCP_DELTA_SHIFT)) \ 353 + (tp)->t_rttvar) >> TCP_DELTA_SHIFT) 354 355 /* 356 * TCP statistics. 357 * Many of these should be kept per connection, 358 * but that's inconvenient at the moment. 359 */ 360 struct tcpstat { 361 u_long tcps_connattempt; /* connections initiated */ 362 u_long tcps_accepts; /* connections accepted */ 363 u_long tcps_connects; /* connections established */ 364 u_long tcps_drops; /* connections dropped */ 365 u_long tcps_conndrops; /* embryonic connections dropped */ 366 u_long tcps_minmssdrops; /* average minmss too low drops */ 367 u_long tcps_closed; /* conn. closed (includes drops) */ 368 u_long tcps_segstimed; /* segs where we tried to get rtt */ 369 u_long tcps_rttupdated; /* times we succeeded */ 370 u_long tcps_delack; /* delayed acks sent */ 371 u_long tcps_timeoutdrop; /* conn. dropped in rxmt timeout */ 372 u_long tcps_rexmttimeo; /* retransmit timeouts */ 373 u_long tcps_persisttimeo; /* persist timeouts */ 374 u_long tcps_keeptimeo; /* keepalive timeouts */ 375 u_long tcps_keepprobe; /* keepalive probes sent */ 376 u_long tcps_keepdrops; /* connections dropped in keepalive */ 377 378 u_long tcps_sndtotal; /* total packets sent */ 379 u_long tcps_sndpack; /* data packets sent */ 380 u_long tcps_sndbyte; /* data bytes sent */ 381 u_long tcps_sndrexmitpack; /* data packets retransmitted */ 382 u_long tcps_sndrexmitbyte; /* data bytes retransmitted */ 383 u_long tcps_sndrexmitbad; /* unnecessary packet retransmissions */ 384 u_long tcps_sndacks; /* ack-only packets sent */ 385 u_long tcps_sndprobe; /* window probes sent */ 386 u_long tcps_sndurg; /* packets sent with URG only */ 387 u_long tcps_sndwinup; /* window update-only packets sent */ 388 u_long tcps_sndctrl; /* control (SYN|FIN|RST) packets sent */ 389 390 u_long tcps_rcvtotal; /* total packets received */ 391 u_long tcps_rcvpack; /* packets received in sequence */ 392 u_long tcps_rcvbyte; /* bytes received in sequence */ 393 u_long tcps_rcvbadsum; /* packets received with ccksum errs */ 394 u_long tcps_rcvbadoff; /* packets received with bad offset */ 395 u_long tcps_rcvmemdrop; /* packets dropped for lack of memory */ 396 u_long tcps_rcvshort; /* packets received too short */ 397 u_long tcps_rcvduppack; /* duplicate-only packets received */ 398 u_long tcps_rcvdupbyte; /* duplicate-only bytes received */ 399 u_long tcps_rcvpartduppack; /* packets with some duplicate data */ 400 u_long tcps_rcvpartdupbyte; /* dup. bytes in part-dup. packets */ 401 u_long tcps_rcvoopack; /* out-of-order packets received */ 402 u_long tcps_rcvoobyte; /* out-of-order bytes received */ 403 u_long tcps_rcvpackafterwin; /* packets with data after window */ 404 u_long tcps_rcvbyteafterwin; /* bytes rcvd after window */ 405 u_long tcps_rcvafterclose; /* packets rcvd after "close" */ 406 u_long tcps_rcvwinprobe; /* rcvd window probe packets */ 407 u_long tcps_rcvdupack; /* rcvd duplicate acks */ 408 u_long tcps_rcvacktoomuch; /* rcvd acks for unsent data */ 409 u_long tcps_rcvackpack; /* rcvd ack packets */ 410 u_long tcps_rcvackbyte; /* bytes acked by rcvd acks */ 411 u_long tcps_rcvwinupd; /* rcvd window update packets */ 412 u_long tcps_pawsdrop; /* segments dropped due to PAWS */ 413 u_long tcps_predack; /* times hdr predict ok for acks */ 414 u_long tcps_preddat; /* times hdr predict ok for data pkts */ 415 u_long tcps_pcbcachemiss; 416 u_long tcps_cachedrtt; /* times cached RTT in route updated */ 417 u_long tcps_cachedrttvar; /* times cached rttvar updated */ 418 u_long tcps_cachedssthresh; /* times cached ssthresh updated */ 419 u_long tcps_usedrtt; /* times RTT initialized from route */ 420 u_long tcps_usedrttvar; /* times RTTVAR initialized from rt */ 421 u_long tcps_usedssthresh; /* times ssthresh initialized from rt*/ 422 u_long tcps_persistdrop; /* timeout in persist state */ 423 u_long tcps_badsyn; /* bogus SYN, e.g. premature ACK */ 424 u_long tcps_mturesent; /* resends due to MTU discovery */ 425 u_long tcps_listendrop; /* listen queue overflows */ 426 u_long tcps_badrst; /* ignored RSTs in the window */ 427 428 u_long tcps_sc_added; /* entry added to syncache */ 429 u_long tcps_sc_retransmitted; /* syncache entry was retransmitted */ 430 u_long tcps_sc_dupsyn; /* duplicate SYN packet */ 431 u_long tcps_sc_dropped; /* could not reply to packet */ 432 u_long tcps_sc_completed; /* successful extraction of entry */ 433 u_long tcps_sc_bucketoverflow; /* syncache per-bucket limit hit */ 434 u_long tcps_sc_cacheoverflow; /* syncache cache limit hit */ 435 u_long tcps_sc_reset; /* RST removed entry from syncache */ 436 u_long tcps_sc_stale; /* timed out or listen socket gone */ 437 u_long tcps_sc_aborted; /* syncache entry aborted */ 438 u_long tcps_sc_badack; /* removed due to bad ACK */ 439 u_long tcps_sc_unreach; /* ICMP unreachable received */ 440 u_long tcps_sc_zonefail; /* zalloc() failed */ 441 u_long tcps_sc_sendcookie; /* SYN cookie sent */ 442 u_long tcps_sc_recvcookie; /* SYN cookie received */ 443 444 u_long tcps_hc_added; /* entry added to hostcache */ 445 u_long tcps_hc_bucketoverflow; /* hostcache per bucket limit hit */ 446 447 u_long tcps_finwait2_drops; /* Drop FIN_WAIT_2 connection after time limit */ 448 449 /* SACK related stats */ 450 u_long tcps_sack_recovery_episode; /* SACK recovery episodes */ 451 u_long tcps_sack_rexmits; /* SACK rexmit segments */ 452 u_long tcps_sack_rexmit_bytes; /* SACK rexmit bytes */ 453 u_long tcps_sack_rcv_blocks; /* SACK blocks (options) received */ 454 u_long tcps_sack_send_blocks; /* SACK blocks (options) sent */ 455 u_long tcps_sack_sboverflow; /* times scoreboard overflowed */ 456 457 /* ECN related stats */ 458 u_long tcps_ecn_ce; /* ECN Congestion Experienced */ 459 u_long tcps_ecn_ect0; /* ECN Capable Transport */ 460 u_long tcps_ecn_ect1; /* ECN Capable Transport */ 461 u_long tcps_ecn_shs; /* ECN successful handshakes */ 462 u_long tcps_ecn_rcwnd; /* # times ECN reduced the cwnd */ 463 }; 464 465 #ifdef _KERNEL 466 #define TCPSTAT_ADD(name, val) V_tcpstat.name += (val) 467 #define TCPSTAT_INC(name) TCPSTAT_ADD(name, 1) 468 #endif 469 470 /* 471 * TCB structure exported to user-land via sysctl(3). 472 * Evil hack: declare only if in_pcb.h and sys/socketvar.h have been 473 * included. Not all of our clients do. 474 */ 475 #if defined(_NETINET_IN_PCB_H_) && defined(_SYS_SOCKETVAR_H_) 476 struct xtcpcb { 477 size_t xt_len; 478 struct inpcb xt_inp; 479 struct tcpcb xt_tp; 480 struct xsocket xt_socket; 481 u_quad_t xt_alignment_hack; 482 }; 483 #endif 484 485 /* 486 * Names for TCP sysctl objects 487 */ 488 #define TCPCTL_DO_RFC1323 1 /* use RFC-1323 extensions */ 489 #define TCPCTL_MSSDFLT 3 /* MSS default */ 490 #define TCPCTL_STATS 4 /* statistics (read-only) */ 491 #define TCPCTL_RTTDFLT 5 /* default RTT estimate */ 492 #define TCPCTL_KEEPIDLE 6 /* keepalive idle timer */ 493 #define TCPCTL_KEEPINTVL 7 /* interval to send keepalives */ 494 #define TCPCTL_SENDSPACE 8 /* send buffer space */ 495 #define TCPCTL_RECVSPACE 9 /* receive buffer space */ 496 #define TCPCTL_KEEPINIT 10 /* timeout for establishing syn */ 497 #define TCPCTL_PCBLIST 11 /* list of all outstanding PCBs */ 498 #define TCPCTL_DELACKTIME 12 /* time before sending delayed ACK */ 499 #define TCPCTL_V6MSSDFLT 13 /* MSS default for IPv6 */ 500 #define TCPCTL_SACK 14 /* Selective Acknowledgement,rfc 2018 */ 501 #define TCPCTL_DROP 15 /* drop tcp connection */ 502 #define TCPCTL_MAXID 16 503 #define TCPCTL_FINWAIT2_TIMEOUT 17 504 505 #define TCPCTL_NAMES { \ 506 { 0, 0 }, \ 507 { "rfc1323", CTLTYPE_INT }, \ 508 { "mssdflt", CTLTYPE_INT }, \ 509 { "stats", CTLTYPE_STRUCT }, \ 510 { "rttdflt", CTLTYPE_INT }, \ 511 { "keepidle", CTLTYPE_INT }, \ 512 { "keepintvl", CTLTYPE_INT }, \ 513 { "sendspace", CTLTYPE_INT }, \ 514 { "recvspace", CTLTYPE_INT }, \ 515 { "keepinit", CTLTYPE_INT }, \ 516 { "pcblist", CTLTYPE_STRUCT }, \ 517 { "delacktime", CTLTYPE_INT }, \ 518 { "v6mssdflt", CTLTYPE_INT }, \ 519 { "maxid", CTLTYPE_INT }, \ 520 } 521 522 523 #ifdef _KERNEL 524 #ifdef SYSCTL_DECL 525 SYSCTL_DECL(_net_inet_tcp); 526 SYSCTL_DECL(_net_inet_tcp_sack); 527 MALLOC_DECLARE(M_TCPLOG); 528 #endif 529 530 extern int tcp_log_in_vain; 531 532 #ifdef VIMAGE_GLOBALS 533 extern struct inpcbhead tcb; /* head of queue of active tcpcb's */ 534 extern struct inpcbinfo tcbinfo; 535 extern struct tcpstat tcpstat; /* tcp statistics */ 536 extern int tcp_mssdflt; /* XXX */ 537 extern int tcp_minmss; 538 extern int tcp_delack_enabled; 539 extern int tcp_do_newreno; 540 extern int path_mtu_discovery; 541 extern int ss_fltsz; 542 extern int ss_fltsz_local; 543 544 extern int blackhole; 545 extern int drop_synfin; 546 extern int tcp_do_rfc3042; 547 extern int tcp_do_rfc3390; 548 extern int tcp_insecure_rst; 549 extern int tcp_do_autorcvbuf; 550 extern int tcp_autorcvbuf_inc; 551 extern int tcp_autorcvbuf_max; 552 extern int tcp_do_rfc3465; 553 extern int tcp_abc_l_var; 554 555 extern int tcp_do_tso; 556 extern int tcp_do_autosndbuf; 557 extern int tcp_autosndbuf_inc; 558 extern int tcp_autosndbuf_max; 559 560 extern int nolocaltimewait; 561 562 extern int tcp_do_sack; /* SACK enabled/disabled */ 563 extern int tcp_sack_maxholes; 564 extern int tcp_sack_globalmaxholes; 565 extern int tcp_sack_globalholes; 566 extern int tcp_sc_rst_sock_fail; /* RST on sock alloc failure */ 567 extern int tcp_do_ecn; /* TCP ECN enabled/disabled */ 568 extern int tcp_ecn_maxretries; 569 #endif /* VIMAGE_GLOBALS */ 570 571 int tcp_addoptions(struct tcpopt *, u_char *); 572 struct tcpcb * 573 tcp_close(struct tcpcb *); 574 void tcp_discardcb(struct tcpcb *); 575 void tcp_twstart(struct tcpcb *); 576 #if 0 577 int tcp_twrecycleable(struct tcptw *tw); 578 #endif 579 void tcp_twclose(struct tcptw *_tw, int _reuse); 580 void tcp_ctlinput(int, struct sockaddr *, void *); 581 int tcp_ctloutput(struct socket *, struct sockopt *); 582 struct tcpcb * 583 tcp_drop(struct tcpcb *, int); 584 void tcp_drain(void); 585 void tcp_fasttimo(void); 586 void tcp_init(void); 587 void tcp_fini(void *); 588 char *tcp_log_addrs(struct in_conninfo *, struct tcphdr *, void *, 589 const void *); 590 int tcp_reass(struct tcpcb *, struct tcphdr *, int *, struct mbuf *); 591 void tcp_reass_init(void); 592 void tcp_input(struct mbuf *, int); 593 u_long tcp_maxmtu(struct in_conninfo *, int *); 594 u_long tcp_maxmtu6(struct in_conninfo *, int *); 595 void tcp_mss_update(struct tcpcb *, int, struct hc_metrics_lite *, int *); 596 void tcp_mss(struct tcpcb *, int); 597 int tcp_mssopt(struct in_conninfo *); 598 struct inpcb * 599 tcp_drop_syn_sent(struct inpcb *, int); 600 struct inpcb * 601 tcp_mtudisc(struct inpcb *, int); 602 struct tcpcb * 603 tcp_newtcpcb(struct inpcb *); 604 int tcp_output(struct tcpcb *); 605 void tcp_respond(struct tcpcb *, void *, 606 struct tcphdr *, struct mbuf *, tcp_seq, tcp_seq, int); 607 void tcp_tw_init(void); 608 void tcp_tw_zone_change(void); 609 int tcp_twcheck(struct inpcb *, struct tcpopt *, struct tcphdr *, 610 struct mbuf *, int); 611 int tcp_twrespond(struct tcptw *, int); 612 void tcp_setpersist(struct tcpcb *); 613 #ifdef TCP_SIGNATURE 614 int tcp_signature_compute(struct mbuf *, int, int, int, u_char *, u_int); 615 #endif 616 void tcp_slowtimo(void); 617 struct tcptemp * 618 tcpip_maketemplate(struct inpcb *); 619 void tcpip_fillheaders(struct inpcb *, void *, void *); 620 void tcp_timer_activate(struct tcpcb *, int, u_int); 621 int tcp_timer_active(struct tcpcb *, int); 622 void tcp_trace(short, short, struct tcpcb *, void *, struct tcphdr *, int); 623 void tcp_xmit_bandwidth_limit(struct tcpcb *tp, tcp_seq ack_seq); 624 /* 625 * All tcp_hc_* functions are IPv4 and IPv6 (via in_conninfo) 626 */ 627 void tcp_hc_init(void); 628 void tcp_hc_get(struct in_conninfo *, struct hc_metrics_lite *); 629 u_long tcp_hc_getmtu(struct in_conninfo *); 630 void tcp_hc_updatemtu(struct in_conninfo *, u_long); 631 void tcp_hc_update(struct in_conninfo *, struct hc_metrics_lite *); 632 633 extern struct pr_usrreqs tcp_usrreqs; 634 extern u_long tcp_sendspace; 635 extern u_long tcp_recvspace; 636 tcp_seq tcp_new_isn(struct tcpcb *); 637 638 void tcp_sack_doack(struct tcpcb *, struct tcpopt *, tcp_seq); 639 void tcp_update_sack_list(struct tcpcb *tp, tcp_seq rcv_laststart, tcp_seq rcv_lastend); 640 void tcp_clean_sackreport(struct tcpcb *tp); 641 void tcp_sack_adjust(struct tcpcb *tp); 642 struct sackhole *tcp_sack_output(struct tcpcb *tp, int *sack_bytes_rexmt); 643 void tcp_sack_partialack(struct tcpcb *, struct tcphdr *); 644 void tcp_free_sackholes(struct tcpcb *tp); 645 int tcp_newreno(struct tcpcb *, struct tcphdr *); 646 u_long tcp_seq_subtract(u_long, u_long ); 647 648 #endif /* _KERNEL */ 649 650 #endif /* _NETINET_TCP_VAR_H_ */ 651