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 #ifdef _KERNEL 39 #include <net/vnet.h> 40 41 /* 42 * Kernel variables for tcp. 43 */ 44 VNET_DECLARE(int, tcp_do_rfc1323); 45 #define V_tcp_do_rfc1323 VNET(tcp_do_rfc1323) 46 47 #endif /* _KERNEL */ 48 49 /* TCP segment queue entry */ 50 struct tseg_qent { 51 LIST_ENTRY(tseg_qent) tqe_q; 52 int tqe_len; /* TCP segment data length */ 53 struct tcphdr *tqe_th; /* a pointer to tcp header */ 54 struct mbuf *tqe_m; /* mbuf contains packet */ 55 }; 56 LIST_HEAD(tsegqe_head, tseg_qent); 57 58 struct sackblk { 59 tcp_seq start; /* start seq no. of sack block */ 60 tcp_seq end; /* end seq no. */ 61 }; 62 63 struct sackhole { 64 tcp_seq start; /* start seq no. of hole */ 65 tcp_seq end; /* end seq no. */ 66 tcp_seq rxmit; /* next seq. no in hole to be retransmitted */ 67 TAILQ_ENTRY(sackhole) scblink; /* scoreboard linkage */ 68 }; 69 70 struct sackhint { 71 struct sackhole *nexthole; 72 int sack_bytes_rexmit; 73 tcp_seq last_sack_ack; /* Most recent/largest sacked ack */ 74 75 int ispare; /* explicit pad for 64bit alignment */ 76 uint64_t _pad[2]; /* 1 sacked_bytes, 1 TBD */ 77 }; 78 79 struct tcptemp { 80 u_char tt_ipgen[40]; /* the size must be of max ip header, now IPv6 */ 81 struct tcphdr tt_t; 82 }; 83 84 #define tcp6cb tcpcb /* for KAME src sync over BSD*'s */ 85 86 /* Neighbor Discovery, Neighbor Unreachability Detection Upper layer hint. */ 87 #ifdef INET6 88 #define ND6_HINT(tp) \ 89 do { \ 90 if ((tp) && (tp)->t_inpcb && \ 91 ((tp)->t_inpcb->inp_vflag & INP_IPV6) != 0) \ 92 nd6_nud_hint(NULL, NULL, 0); \ 93 } while (0) 94 #else 95 #define ND6_HINT(tp) 96 #endif 97 98 /* 99 * Tcp control block, one per tcp; fields: 100 * Organized for 16 byte cacheline efficiency. 101 */ 102 struct tcpcb { 103 struct tsegqe_head t_segq; /* segment reassembly queue */ 104 void *t_pspare[2]; /* new reassembly queue */ 105 int t_segqlen; /* segment reassembly queue length */ 106 int t_dupacks; /* consecutive dup acks recd */ 107 108 struct tcp_timer *t_timers; /* All the TCP timers in one struct */ 109 110 struct inpcb *t_inpcb; /* back pointer to internet pcb */ 111 int t_state; /* state of this connection */ 112 u_int t_flags; 113 114 struct vnet *t_vnet; /* back pointer to parent vnet */ 115 116 tcp_seq snd_una; /* send unacknowledged */ 117 tcp_seq snd_max; /* highest sequence number sent; 118 * used to recognize retransmits 119 */ 120 tcp_seq snd_nxt; /* send next */ 121 tcp_seq snd_up; /* send urgent pointer */ 122 123 tcp_seq snd_wl1; /* window update seg seq number */ 124 tcp_seq snd_wl2; /* window update seg ack number */ 125 tcp_seq iss; /* initial send sequence number */ 126 tcp_seq irs; /* initial receive sequence number */ 127 128 tcp_seq rcv_nxt; /* receive next */ 129 tcp_seq rcv_adv; /* advertised window */ 130 u_long rcv_wnd; /* receive window */ 131 tcp_seq rcv_up; /* receive urgent pointer */ 132 133 u_long snd_wnd; /* send window */ 134 u_long snd_cwnd; /* congestion-controlled window */ 135 u_long snd_spare1; /* unused */ 136 u_long snd_ssthresh; /* snd_cwnd size threshold for 137 * for slow start exponential to 138 * linear switch 139 */ 140 u_long snd_spare2; /* unused */ 141 tcp_seq snd_recover; /* for use in NewReno Fast Recovery */ 142 143 u_int t_maxopd; /* mss plus options */ 144 145 u_int t_rcvtime; /* inactivity time */ 146 u_int t_starttime; /* time connection was established */ 147 u_int t_rtttime; /* RTT measurement start time */ 148 tcp_seq t_rtseq; /* sequence number being timed */ 149 150 u_int t_bw_spare1; /* unused */ 151 tcp_seq t_bw_spare2; /* unused */ 152 153 int t_rxtcur; /* current retransmit value (ticks) */ 154 u_int t_maxseg; /* maximum segment size */ 155 int t_srtt; /* smoothed round-trip time */ 156 int t_rttvar; /* variance in round-trip time */ 157 158 int t_rxtshift; /* log(2) of rexmt exp. backoff */ 159 u_int t_rttmin; /* minimum rtt allowed */ 160 u_int t_rttbest; /* best rtt we've seen */ 161 u_long t_rttupdated; /* number of times rtt sampled */ 162 u_long max_sndwnd; /* largest window peer has offered */ 163 164 int t_softerror; /* possible error not yet reported */ 165 /* out-of-band data */ 166 char t_oobflags; /* have some */ 167 char t_iobc; /* input character */ 168 /* RFC 1323 variables */ 169 u_char snd_scale; /* window scaling for send window */ 170 u_char rcv_scale; /* window scaling for recv window */ 171 u_char request_r_scale; /* pending window scaling */ 172 u_int32_t ts_recent; /* timestamp echo data */ 173 u_int ts_recent_age; /* when last updated */ 174 u_int32_t ts_offset; /* our timestamp offset */ 175 176 tcp_seq last_ack_sent; 177 /* experimental */ 178 u_long snd_cwnd_prev; /* cwnd prior to retransmit */ 179 u_long snd_ssthresh_prev; /* ssthresh prior to retransmit */ 180 tcp_seq snd_recover_prev; /* snd_recover prior to retransmit */ 181 int t_sndzerowin; /* zero-window updates sent */ 182 u_int t_badrxtwin; /* window for retransmit recovery */ 183 u_char snd_limited; /* segments limited transmitted */ 184 /* SACK related state */ 185 int snd_numholes; /* number of holes seen by sender */ 186 TAILQ_HEAD(sackhole_head, sackhole) snd_holes; 187 /* SACK scoreboard (sorted) */ 188 tcp_seq snd_fack; /* last seq number(+1) sack'd by rcv'r*/ 189 int rcv_numsacks; /* # distinct sack blks present */ 190 struct sackblk sackblks[MAX_SACK_BLKS]; /* seq nos. of sack blocks */ 191 tcp_seq sack_newdata; /* New data xmitted in this recovery 192 episode starts at this seq number */ 193 struct sackhint sackhint; /* SACK scoreboard hint */ 194 int t_rttlow; /* smallest observerved RTT */ 195 u_int32_t rfbuf_ts; /* recv buffer autoscaling timestamp */ 196 int rfbuf_cnt; /* recv buffer autoscaling byte count */ 197 struct toedev *tod; /* toedev handling this connection */ 198 int t_sndrexmitpack; /* retransmit packets sent */ 199 int t_rcvoopack; /* out-of-order packets received */ 200 void *t_toe; /* TOE pcb pointer */ 201 int t_bytes_acked; /* # bytes acked during current RTT */ 202 struct cc_algo *cc_algo; /* congestion control algorithm */ 203 struct cc_var *ccv; /* congestion control specific vars */ 204 struct osd *osd; /* storage for Khelp module data */ 205 206 u_int t_keepinit; /* time to establish connection */ 207 u_int t_keepidle; /* time before keepalive probes begin */ 208 u_int t_keepintvl; /* interval between keepalives */ 209 u_int t_keepcnt; /* number of keepalives before close */ 210 211 uint32_t t_ispare[8]; /* 5 UTO, 3 TBD */ 212 void *t_pspare2[4]; /* 4 TBD */ 213 uint64_t _pad[6]; /* 6 TBD (1-2 CC/RTT?) */ 214 }; 215 216 /* 217 * Flags and utility macros for the t_flags field. 218 */ 219 #define TF_ACKNOW 0x000001 /* ack peer immediately */ 220 #define TF_DELACK 0x000002 /* ack, but try to delay it */ 221 #define TF_NODELAY 0x000004 /* don't delay packets to coalesce */ 222 #define TF_NOOPT 0x000008 /* don't use tcp options */ 223 #define TF_SENTFIN 0x000010 /* have sent FIN */ 224 #define TF_REQ_SCALE 0x000020 /* have/will request window scaling */ 225 #define TF_RCVD_SCALE 0x000040 /* other side has requested scaling */ 226 #define TF_REQ_TSTMP 0x000080 /* have/will request timestamps */ 227 #define TF_RCVD_TSTMP 0x000100 /* a timestamp was received in SYN */ 228 #define TF_SACK_PERMIT 0x000200 /* other side said I could SACK */ 229 #define TF_NEEDSYN 0x000400 /* send SYN (implicit state) */ 230 #define TF_NEEDFIN 0x000800 /* send FIN (implicit state) */ 231 #define TF_NOPUSH 0x001000 /* don't push */ 232 #define TF_PREVVALID 0x002000 /* saved values for bad rxmit valid */ 233 #define TF_MORETOCOME 0x010000 /* More data to be appended to sock */ 234 #define TF_LQ_OVERFLOW 0x020000 /* listen queue overflow */ 235 #define TF_LASTIDLE 0x040000 /* connection was previously idle */ 236 #define TF_RXWIN0SENT 0x080000 /* sent a receiver win 0 in response */ 237 #define TF_FASTRECOVERY 0x100000 /* in NewReno Fast Recovery */ 238 #define TF_WASFRECOVERY 0x200000 /* was in NewReno Fast Recovery */ 239 #define TF_SIGNATURE 0x400000 /* require MD5 digests (RFC2385) */ 240 #define TF_FORCEDATA 0x800000 /* force out a byte */ 241 #define TF_TSO 0x1000000 /* TSO enabled on this connection */ 242 #define TF_TOE 0x2000000 /* this connection is offloaded */ 243 #define TF_ECN_PERMIT 0x4000000 /* connection ECN-ready */ 244 #define TF_ECN_SND_CWR 0x8000000 /* ECN CWR in queue */ 245 #define TF_ECN_SND_ECE 0x10000000 /* ECN ECE in queue */ 246 #define TF_CONGRECOVERY 0x20000000 /* congestion recovery mode */ 247 #define TF_WASCRECOVERY 0x40000000 /* was in congestion recovery */ 248 249 #define IN_FASTRECOVERY(t_flags) (t_flags & TF_FASTRECOVERY) 250 #define ENTER_FASTRECOVERY(t_flags) t_flags |= TF_FASTRECOVERY 251 #define EXIT_FASTRECOVERY(t_flags) t_flags &= ~TF_FASTRECOVERY 252 253 #define IN_CONGRECOVERY(t_flags) (t_flags & TF_CONGRECOVERY) 254 #define ENTER_CONGRECOVERY(t_flags) t_flags |= TF_CONGRECOVERY 255 #define EXIT_CONGRECOVERY(t_flags) t_flags &= ~TF_CONGRECOVERY 256 257 #define IN_RECOVERY(t_flags) (t_flags & (TF_CONGRECOVERY | TF_FASTRECOVERY)) 258 #define ENTER_RECOVERY(t_flags) t_flags |= (TF_CONGRECOVERY | TF_FASTRECOVERY) 259 #define EXIT_RECOVERY(t_flags) t_flags &= ~(TF_CONGRECOVERY | TF_FASTRECOVERY) 260 261 #define BYTES_THIS_ACK(tp, th) (th->th_ack - tp->snd_una) 262 263 /* 264 * Flags for the t_oobflags field. 265 */ 266 #define TCPOOB_HAVEDATA 0x01 267 #define TCPOOB_HADDATA 0x02 268 269 #ifdef TCP_SIGNATURE 270 /* 271 * Defines which are needed by the xform_tcp module and tcp_[in|out]put 272 * for SADB verification and lookup. 273 */ 274 #define TCP_SIGLEN 16 /* length of computed digest in bytes */ 275 #define TCP_KEYLEN_MIN 1 /* minimum length of TCP-MD5 key */ 276 #define TCP_KEYLEN_MAX 80 /* maximum length of TCP-MD5 key */ 277 /* 278 * Only a single SA per host may be specified at this time. An SPI is 279 * needed in order for the KEY_ALLOCSA() lookup to work. 280 */ 281 #define TCP_SIG_SPI 0x1000 282 #endif /* TCP_SIGNATURE */ 283 284 /* 285 * Structure to hold TCP options that are only used during segment 286 * processing (in tcp_input), but not held in the tcpcb. 287 * It's basically used to reduce the number of parameters 288 * to tcp_dooptions and tcp_addoptions. 289 * The binary order of the to_flags is relevant for packing of the 290 * options in tcp_addoptions. 291 */ 292 struct tcpopt { 293 u_int64_t to_flags; /* which options are present */ 294 #define TOF_MSS 0x0001 /* maximum segment size */ 295 #define TOF_SCALE 0x0002 /* window scaling */ 296 #define TOF_SACKPERM 0x0004 /* SACK permitted */ 297 #define TOF_TS 0x0010 /* timestamp */ 298 #define TOF_SIGNATURE 0x0040 /* TCP-MD5 signature option (RFC2385) */ 299 #define TOF_SACK 0x0080 /* Peer sent SACK option */ 300 #define TOF_MAXOPT 0x0100 301 u_int32_t to_tsval; /* new timestamp */ 302 u_int32_t to_tsecr; /* reflected timestamp */ 303 u_char *to_sacks; /* pointer to the first SACK blocks */ 304 u_char *to_signature; /* pointer to the TCP-MD5 signature */ 305 u_int16_t to_mss; /* maximum segment size */ 306 u_int8_t to_wscale; /* window scaling */ 307 u_int8_t to_nsacks; /* number of SACK blocks */ 308 u_int32_t to_spare; /* UTO */ 309 }; 310 311 /* 312 * Flags for tcp_dooptions. 313 */ 314 #define TO_SYN 0x01 /* parse SYN-only options */ 315 316 struct hc_metrics_lite { /* must stay in sync with hc_metrics */ 317 u_long rmx_mtu; /* MTU for this path */ 318 u_long rmx_ssthresh; /* outbound gateway buffer limit */ 319 u_long rmx_rtt; /* estimated round trip time */ 320 u_long rmx_rttvar; /* estimated rtt variance */ 321 u_long rmx_bandwidth; /* estimated bandwidth */ 322 u_long rmx_cwnd; /* congestion window */ 323 u_long rmx_sendpipe; /* outbound delay-bandwidth product */ 324 u_long rmx_recvpipe; /* inbound delay-bandwidth product */ 325 }; 326 327 #ifndef _NETINET_IN_PCB_H_ 328 struct in_conninfo; 329 #endif /* _NETINET_IN_PCB_H_ */ 330 331 struct tcptw { 332 struct inpcb *tw_inpcb; /* XXX back pointer to internet pcb */ 333 tcp_seq snd_nxt; 334 tcp_seq rcv_nxt; 335 tcp_seq iss; 336 tcp_seq irs; 337 u_short last_win; /* cached window value */ 338 u_short tw_so_options; /* copy of so_options */ 339 struct ucred *tw_cred; /* user credentials */ 340 u_int32_t t_recent; 341 u_int32_t ts_offset; /* our timestamp offset */ 342 u_int t_starttime; 343 int tw_time; 344 TAILQ_ENTRY(tcptw) tw_2msl; 345 }; 346 347 #define intotcpcb(ip) ((struct tcpcb *)(ip)->inp_ppcb) 348 #define intotw(ip) ((struct tcptw *)(ip)->inp_ppcb) 349 #define sototcpcb(so) (intotcpcb(sotoinpcb(so))) 350 351 /* 352 * The smoothed round-trip time and estimated variance 353 * are stored as fixed point numbers scaled by the values below. 354 * For convenience, these scales are also used in smoothing the average 355 * (smoothed = (1/scale)sample + ((scale-1)/scale)smoothed). 356 * With these scales, srtt has 3 bits to the right of the binary point, 357 * and thus an "ALPHA" of 0.875. rttvar has 2 bits to the right of the 358 * binary point, and is smoothed with an ALPHA of 0.75. 359 */ 360 #define TCP_RTT_SCALE 32 /* multiplier for srtt; 3 bits frac. */ 361 #define TCP_RTT_SHIFT 5 /* shift for srtt; 3 bits frac. */ 362 #define TCP_RTTVAR_SCALE 16 /* multiplier for rttvar; 2 bits */ 363 #define TCP_RTTVAR_SHIFT 4 /* shift for rttvar; 2 bits */ 364 #define TCP_DELTA_SHIFT 2 /* see tcp_input.c */ 365 366 /* 367 * The initial retransmission should happen at rtt + 4 * rttvar. 368 * Because of the way we do the smoothing, srtt and rttvar 369 * will each average +1/2 tick of bias. When we compute 370 * the retransmit timer, we want 1/2 tick of rounding and 371 * 1 extra tick because of +-1/2 tick uncertainty in the 372 * firing of the timer. The bias will give us exactly the 373 * 1.5 tick we need. But, because the bias is 374 * statistical, we have to test that we don't drop below 375 * the minimum feasible timer (which is 2 ticks). 376 * This version of the macro adapted from a paper by Lawrence 377 * Brakmo and Larry Peterson which outlines a problem caused 378 * by insufficient precision in the original implementation, 379 * which results in inappropriately large RTO values for very 380 * fast networks. 381 */ 382 #define TCP_REXMTVAL(tp) \ 383 max((tp)->t_rttmin, (((tp)->t_srtt >> (TCP_RTT_SHIFT - TCP_DELTA_SHIFT)) \ 384 + (tp)->t_rttvar) >> TCP_DELTA_SHIFT) 385 386 /* 387 * TCP statistics. 388 * Many of these should be kept per connection, 389 * but that's inconvenient at the moment. 390 */ 391 struct tcpstat { 392 uint64_t tcps_connattempt; /* connections initiated */ 393 uint64_t tcps_accepts; /* connections accepted */ 394 uint64_t tcps_connects; /* connections established */ 395 uint64_t tcps_drops; /* connections dropped */ 396 uint64_t tcps_conndrops; /* embryonic connections dropped */ 397 uint64_t tcps_minmssdrops; /* average minmss too low drops */ 398 uint64_t tcps_closed; /* conn. closed (includes drops) */ 399 uint64_t tcps_segstimed; /* segs where we tried to get rtt */ 400 uint64_t tcps_rttupdated; /* times we succeeded */ 401 uint64_t tcps_delack; /* delayed acks sent */ 402 uint64_t tcps_timeoutdrop; /* conn. dropped in rxmt timeout */ 403 uint64_t tcps_rexmttimeo; /* retransmit timeouts */ 404 uint64_t tcps_persisttimeo; /* persist timeouts */ 405 uint64_t tcps_keeptimeo; /* keepalive timeouts */ 406 uint64_t tcps_keepprobe; /* keepalive probes sent */ 407 uint64_t tcps_keepdrops; /* connections dropped in keepalive */ 408 409 uint64_t tcps_sndtotal; /* total packets sent */ 410 uint64_t tcps_sndpack; /* data packets sent */ 411 uint64_t tcps_sndbyte; /* data bytes sent */ 412 uint64_t tcps_sndrexmitpack; /* data packets retransmitted */ 413 uint64_t tcps_sndrexmitbyte; /* data bytes retransmitted */ 414 uint64_t tcps_sndrexmitbad; /* unnecessary packet retransmissions */ 415 uint64_t tcps_sndacks; /* ack-only packets sent */ 416 uint64_t tcps_sndprobe; /* window probes sent */ 417 uint64_t tcps_sndurg; /* packets sent with URG only */ 418 uint64_t tcps_sndwinup; /* window update-only packets sent */ 419 uint64_t tcps_sndctrl; /* control (SYN|FIN|RST) packets sent */ 420 421 uint64_t tcps_rcvtotal; /* total packets received */ 422 uint64_t tcps_rcvpack; /* packets received in sequence */ 423 uint64_t tcps_rcvbyte; /* bytes received in sequence */ 424 uint64_t tcps_rcvbadsum; /* packets received with ccksum errs */ 425 uint64_t tcps_rcvbadoff; /* packets received with bad offset */ 426 uint64_t tcps_rcvmemdrop; /* packets dropped for lack of memory */ 427 uint64_t tcps_rcvshort; /* packets received too short */ 428 uint64_t tcps_rcvduppack; /* duplicate-only packets received */ 429 uint64_t tcps_rcvdupbyte; /* duplicate-only bytes received */ 430 uint64_t tcps_rcvpartduppack; /* packets with some duplicate data */ 431 uint64_t tcps_rcvpartdupbyte; /* dup. bytes in part-dup. packets */ 432 uint64_t tcps_rcvoopack; /* out-of-order packets received */ 433 uint64_t tcps_rcvoobyte; /* out-of-order bytes received */ 434 uint64_t tcps_rcvpackafterwin; /* packets with data after window */ 435 uint64_t tcps_rcvbyteafterwin; /* bytes rcvd after window */ 436 uint64_t tcps_rcvafterclose; /* packets rcvd after "close" */ 437 uint64_t tcps_rcvwinprobe; /* rcvd window probe packets */ 438 uint64_t tcps_rcvdupack; /* rcvd duplicate acks */ 439 uint64_t tcps_rcvacktoomuch; /* rcvd acks for unsent data */ 440 uint64_t tcps_rcvackpack; /* rcvd ack packets */ 441 uint64_t tcps_rcvackbyte; /* bytes acked by rcvd acks */ 442 uint64_t tcps_rcvwinupd; /* rcvd window update packets */ 443 uint64_t tcps_pawsdrop; /* segments dropped due to PAWS */ 444 uint64_t tcps_predack; /* times hdr predict ok for acks */ 445 uint64_t tcps_preddat; /* times hdr predict ok for data pkts */ 446 uint64_t tcps_pcbcachemiss; 447 uint64_t tcps_cachedrtt; /* times cached RTT in route updated */ 448 uint64_t tcps_cachedrttvar; /* times cached rttvar updated */ 449 uint64_t tcps_cachedssthresh; /* times cached ssthresh updated */ 450 uint64_t tcps_usedrtt; /* times RTT initialized from route */ 451 uint64_t tcps_usedrttvar; /* times RTTVAR initialized from rt */ 452 uint64_t tcps_usedssthresh; /* times ssthresh initialized from rt*/ 453 uint64_t tcps_persistdrop; /* timeout in persist state */ 454 uint64_t tcps_badsyn; /* bogus SYN, e.g. premature ACK */ 455 uint64_t tcps_mturesent; /* resends due to MTU discovery */ 456 uint64_t tcps_listendrop; /* listen queue overflows */ 457 uint64_t tcps_badrst; /* ignored RSTs in the window */ 458 459 uint64_t tcps_sc_added; /* entry added to syncache */ 460 uint64_t tcps_sc_retransmitted; /* syncache entry was retransmitted */ 461 uint64_t tcps_sc_dupsyn; /* duplicate SYN packet */ 462 uint64_t tcps_sc_dropped; /* could not reply to packet */ 463 uint64_t tcps_sc_completed; /* successful extraction of entry */ 464 uint64_t tcps_sc_bucketoverflow;/* syncache per-bucket limit hit */ 465 uint64_t tcps_sc_cacheoverflow; /* syncache cache limit hit */ 466 uint64_t tcps_sc_reset; /* RST removed entry from syncache */ 467 uint64_t tcps_sc_stale; /* timed out or listen socket gone */ 468 uint64_t tcps_sc_aborted; /* syncache entry aborted */ 469 uint64_t tcps_sc_badack; /* removed due to bad ACK */ 470 uint64_t tcps_sc_unreach; /* ICMP unreachable received */ 471 uint64_t tcps_sc_zonefail; /* zalloc() failed */ 472 uint64_t tcps_sc_sendcookie; /* SYN cookie sent */ 473 uint64_t tcps_sc_recvcookie; /* SYN cookie received */ 474 475 uint64_t tcps_hc_added; /* entry added to hostcache */ 476 uint64_t tcps_hc_bucketoverflow;/* hostcache per bucket limit hit */ 477 478 uint64_t tcps_finwait2_drops; /* Drop FIN_WAIT_2 connection after time limit */ 479 480 /* SACK related stats */ 481 uint64_t tcps_sack_recovery_episode; /* SACK recovery episodes */ 482 uint64_t tcps_sack_rexmits; /* SACK rexmit segments */ 483 uint64_t tcps_sack_rexmit_bytes; /* SACK rexmit bytes */ 484 uint64_t tcps_sack_rcv_blocks; /* SACK blocks (options) received */ 485 uint64_t tcps_sack_send_blocks; /* SACK blocks (options) sent */ 486 uint64_t tcps_sack_sboverflow; /* times scoreboard overflowed */ 487 488 /* ECN related stats */ 489 uint64_t tcps_ecn_ce; /* ECN Congestion Experienced */ 490 uint64_t tcps_ecn_ect0; /* ECN Capable Transport */ 491 uint64_t tcps_ecn_ect1; /* ECN Capable Transport */ 492 uint64_t tcps_ecn_shs; /* ECN successful handshakes */ 493 uint64_t tcps_ecn_rcwnd; /* # times ECN reduced the cwnd */ 494 495 /* TCP_SIGNATURE related stats */ 496 uint64_t tcps_sig_rcvgoodsig; /* Total matching signature received */ 497 uint64_t tcps_sig_rcvbadsig; /* Total bad signature received */ 498 uint64_t tcps_sig_err_buildsig; /* Mismatching signature received */ 499 uint64_t tcps_sig_err_sigopt; /* No signature expected by socket */ 500 uint64_t tcps_sig_err_nosigopt; /* No signature provided by segment */ 501 502 uint64_t _pad[12]; /* 6 UTO, 6 TBD */ 503 }; 504 505 #ifdef _KERNEL 506 507 #include <sys/counter.h> 508 509 /* Should match 'struct tcpstat' above. */ 510 struct tcpstat_p { 511 counter_u64_t tcps_connattempt; 512 counter_u64_t tcps_accepts; 513 counter_u64_t tcps_connects; 514 counter_u64_t tcps_drops; 515 counter_u64_t tcps_conndrops; 516 counter_u64_t tcps_minmssdrops; 517 counter_u64_t tcps_closed; 518 counter_u64_t tcps_segstimed; 519 counter_u64_t tcps_rttupdated; 520 counter_u64_t tcps_delack; 521 counter_u64_t tcps_timeoutdrop; 522 counter_u64_t tcps_rexmttimeo; 523 counter_u64_t tcps_persisttimeo; 524 counter_u64_t tcps_keeptimeo; 525 counter_u64_t tcps_keepprobe; 526 counter_u64_t tcps_keepdrops; 527 counter_u64_t tcps_sndtotal; 528 counter_u64_t tcps_sndpack; 529 counter_u64_t tcps_sndbyte; 530 counter_u64_t tcps_sndrexmitpack; 531 counter_u64_t tcps_sndrexmitbyte; 532 counter_u64_t tcps_sndrexmitbad; 533 counter_u64_t tcps_sndacks; 534 counter_u64_t tcps_sndprobe; 535 counter_u64_t tcps_sndurg; 536 counter_u64_t tcps_sndwinup; 537 counter_u64_t tcps_sndctrl; 538 counter_u64_t tcps_rcvtotal; 539 counter_u64_t tcps_rcvpack; 540 counter_u64_t tcps_rcvbyte; 541 counter_u64_t tcps_rcvbadsum; 542 counter_u64_t tcps_rcvbadoff; 543 counter_u64_t tcps_rcvmemdrop; 544 counter_u64_t tcps_rcvshort; 545 counter_u64_t tcps_rcvduppack; 546 counter_u64_t tcps_rcvdupbyte; 547 counter_u64_t tcps_rcvpartduppack; 548 counter_u64_t tcps_rcvpartdupbyte; 549 counter_u64_t tcps_rcvoopack; 550 counter_u64_t tcps_rcvoobyte; 551 counter_u64_t tcps_rcvpackafterwin; 552 counter_u64_t tcps_rcvbyteafterwin; 553 counter_u64_t tcps_rcvafterclose; 554 counter_u64_t tcps_rcvwinprobe; 555 counter_u64_t tcps_rcvdupack; 556 counter_u64_t tcps_rcvacktoomuch; 557 counter_u64_t tcps_rcvackpack; 558 counter_u64_t tcps_rcvackbyte; 559 counter_u64_t tcps_rcvwinupd; 560 counter_u64_t tcps_pawsdrop; 561 counter_u64_t tcps_predack; 562 counter_u64_t tcps_preddat; 563 counter_u64_t tcps_pcbcachemiss; 564 counter_u64_t tcps_cachedrtt; 565 counter_u64_t tcps_cachedrttvar; 566 counter_u64_t tcps_cachedssthresh; 567 counter_u64_t tcps_usedrtt; 568 counter_u64_t tcps_usedrttvar; 569 counter_u64_t tcps_usedssthresh; 570 counter_u64_t tcps_persistdrop; 571 counter_u64_t tcps_badsyn; 572 counter_u64_t tcps_mturesent; 573 counter_u64_t tcps_listendrop; 574 counter_u64_t tcps_badrst; 575 counter_u64_t tcps_sc_added; 576 counter_u64_t tcps_sc_retransmitted; 577 counter_u64_t tcps_sc_dupsyn; 578 counter_u64_t tcps_sc_dropped; 579 counter_u64_t tcps_sc_completed; 580 counter_u64_t tcps_sc_bucketoverflow; 581 counter_u64_t tcps_sc_cacheoverflow; 582 counter_u64_t tcps_sc_reset; 583 counter_u64_t tcps_sc_stale; 584 counter_u64_t tcps_sc_aborted; 585 counter_u64_t tcps_sc_badack; 586 counter_u64_t tcps_sc_unreach; 587 counter_u64_t tcps_sc_zonefail; 588 counter_u64_t tcps_sc_sendcookie; 589 counter_u64_t tcps_sc_recvcookie; 590 counter_u64_t tcps_hc_added; 591 counter_u64_t tcps_hc_bucketoverflow; 592 counter_u64_t tcps_finwait2_drops; 593 counter_u64_t tcps_sack_recovery_episode; 594 counter_u64_t tcps_sack_rexmits; 595 counter_u64_t tcps_sack_rexmit_bytes; 596 counter_u64_t tcps_sack_rcv_blocks; 597 counter_u64_t tcps_sack_send_blocks; 598 counter_u64_t tcps_sack_sboverflow; 599 counter_u64_t tcps_ecn_ce; 600 counter_u64_t tcps_ecn_ect0; 601 counter_u64_t tcps_ecn_ect1; 602 counter_u64_t tcps_ecn_shs; 603 counter_u64_t tcps_ecn_rcwnd; 604 counter_u64_t tcps_sig_rcvgoodsig; 605 counter_u64_t tcps_sig_rcvbadsig; 606 counter_u64_t tcps_sig_err_buildsig; 607 counter_u64_t tcps_sig_err_sigopt; 608 counter_u64_t tcps_sig_err_nosigopt; 609 }; 610 611 VNET_DECLARE(struct tcpstat_p, tcpstatp); /* tcp statistics */ 612 #define V_tcpstatp VNET(tcpstatp) 613 614 /* 615 * In-kernel consumers can use these accessor macros directly to update 616 * stats. 617 */ 618 #define TCPSTAT_ADD(name, val) counter_u64_add(V_tcpstatp.name, (val)) 619 #define TCPSTAT_INC(name) TCPSTAT_ADD(name, 1) 620 621 /* 622 * Kernel module consumers must use this accessor macro. 623 */ 624 void kmod_tcpstat_inc(int statnum); 625 #define KMOD_TCPSTAT_INC(name) \ 626 kmod_tcpstat_inc(offsetof(struct tcpstat_p, name) / \ 627 sizeof(counter_u64_t)) 628 629 /* 630 * TCP specific helper hook point identifiers. 631 */ 632 #define HHOOK_TCP_EST_IN 0 633 #define HHOOK_TCP_EST_OUT 1 634 #define HHOOK_TCP_LAST HHOOK_TCP_EST_OUT 635 636 struct tcp_hhook_data { 637 struct tcpcb *tp; 638 struct tcphdr *th; 639 struct tcpopt *to; 640 long len; 641 int tso; 642 tcp_seq curack; 643 }; 644 #endif 645 646 /* 647 * TCB structure exported to user-land via sysctl(3). 648 * Evil hack: declare only if in_pcb.h and sys/socketvar.h have been 649 * included. Not all of our clients do. 650 */ 651 #if defined(_NETINET_IN_PCB_H_) && defined(_SYS_SOCKETVAR_H_) 652 struct xtcp_timer { 653 int tt_rexmt; /* retransmit timer */ 654 int tt_persist; /* retransmit persistence */ 655 int tt_keep; /* keepalive */ 656 int tt_2msl; /* 2*msl TIME_WAIT timer */ 657 int tt_delack; /* delayed ACK timer */ 658 int t_rcvtime; /* Time since last packet received */ 659 }; 660 struct xtcpcb { 661 size_t xt_len; 662 struct inpcb xt_inp; 663 struct tcpcb xt_tp; 664 struct xsocket xt_socket; 665 struct xtcp_timer xt_timer; 666 u_quad_t xt_alignment_hack; 667 }; 668 #endif 669 670 /* 671 * Names for TCP sysctl objects 672 */ 673 #define TCPCTL_DO_RFC1323 1 /* use RFC-1323 extensions */ 674 #define TCPCTL_MSSDFLT 3 /* MSS default */ 675 #define TCPCTL_STATS 4 /* statistics (read-only) */ 676 #define TCPCTL_RTTDFLT 5 /* default RTT estimate */ 677 #define TCPCTL_KEEPIDLE 6 /* keepalive idle timer */ 678 #define TCPCTL_KEEPINTVL 7 /* interval to send keepalives */ 679 #define TCPCTL_SENDSPACE 8 /* send buffer space */ 680 #define TCPCTL_RECVSPACE 9 /* receive buffer space */ 681 #define TCPCTL_KEEPINIT 10 /* timeout for establishing syn */ 682 #define TCPCTL_PCBLIST 11 /* list of all outstanding PCBs */ 683 #define TCPCTL_DELACKTIME 12 /* time before sending delayed ACK */ 684 #define TCPCTL_V6MSSDFLT 13 /* MSS default for IPv6 */ 685 #define TCPCTL_SACK 14 /* Selective Acknowledgement,rfc 2018 */ 686 #define TCPCTL_DROP 15 /* drop tcp connection */ 687 #define TCPCTL_MAXID 16 688 #define TCPCTL_FINWAIT2_TIMEOUT 17 689 690 #define TCPCTL_NAMES { \ 691 { 0, 0 }, \ 692 { "rfc1323", CTLTYPE_INT }, \ 693 { "mssdflt", CTLTYPE_INT }, \ 694 { "stats", CTLTYPE_STRUCT }, \ 695 { "rttdflt", CTLTYPE_INT }, \ 696 { "keepidle", CTLTYPE_INT }, \ 697 { "keepintvl", CTLTYPE_INT }, \ 698 { "sendspace", CTLTYPE_INT }, \ 699 { "recvspace", CTLTYPE_INT }, \ 700 { "keepinit", CTLTYPE_INT }, \ 701 { "pcblist", CTLTYPE_STRUCT }, \ 702 { "delacktime", CTLTYPE_INT }, \ 703 { "v6mssdflt", CTLTYPE_INT }, \ 704 { "maxid", CTLTYPE_INT }, \ 705 } 706 707 708 #ifdef _KERNEL 709 #ifdef SYSCTL_DECL 710 SYSCTL_DECL(_net_inet_tcp); 711 SYSCTL_DECL(_net_inet_tcp_sack); 712 MALLOC_DECLARE(M_TCPLOG); 713 #endif 714 715 VNET_DECLARE(struct inpcbhead, tcb); /* queue of active tcpcb's */ 716 VNET_DECLARE(struct inpcbinfo, tcbinfo); 717 extern int tcp_log_in_vain; 718 VNET_DECLARE(int, tcp_mssdflt); /* XXX */ 719 VNET_DECLARE(int, tcp_minmss); 720 VNET_DECLARE(int, tcp_delack_enabled); 721 VNET_DECLARE(int, tcp_do_rfc3390); 722 VNET_DECLARE(int, tcp_do_initcwnd10); 723 VNET_DECLARE(int, tcp_sendspace); 724 VNET_DECLARE(int, tcp_recvspace); 725 VNET_DECLARE(int, path_mtu_discovery); 726 VNET_DECLARE(int, tcp_do_rfc3465); 727 VNET_DECLARE(int, tcp_abc_l_var); 728 #define V_tcb VNET(tcb) 729 #define V_tcbinfo VNET(tcbinfo) 730 #define V_tcp_mssdflt VNET(tcp_mssdflt) 731 #define V_tcp_minmss VNET(tcp_minmss) 732 #define V_tcp_delack_enabled VNET(tcp_delack_enabled) 733 #define V_tcp_do_rfc3390 VNET(tcp_do_rfc3390) 734 #define V_tcp_do_initcwnd10 VNET(tcp_do_initcwnd10) 735 #define V_tcp_sendspace VNET(tcp_sendspace) 736 #define V_tcp_recvspace VNET(tcp_recvspace) 737 #define V_path_mtu_discovery VNET(path_mtu_discovery) 738 #define V_tcp_do_rfc3465 VNET(tcp_do_rfc3465) 739 #define V_tcp_abc_l_var VNET(tcp_abc_l_var) 740 741 VNET_DECLARE(int, tcp_do_sack); /* SACK enabled/disabled */ 742 VNET_DECLARE(int, tcp_sc_rst_sock_fail); /* RST on sock alloc failure */ 743 #define V_tcp_do_sack VNET(tcp_do_sack) 744 #define V_tcp_sc_rst_sock_fail VNET(tcp_sc_rst_sock_fail) 745 746 VNET_DECLARE(int, tcp_do_ecn); /* TCP ECN enabled/disabled */ 747 VNET_DECLARE(int, tcp_ecn_maxretries); 748 #define V_tcp_do_ecn VNET(tcp_do_ecn) 749 #define V_tcp_ecn_maxretries VNET(tcp_ecn_maxretries) 750 751 VNET_DECLARE(struct hhook_head *, tcp_hhh[HHOOK_TCP_LAST + 1]); 752 #define V_tcp_hhh VNET(tcp_hhh) 753 754 int tcp_addoptions(struct tcpopt *, u_char *); 755 int tcp_ccalgounload(struct cc_algo *unload_algo); 756 struct tcpcb * 757 tcp_close(struct tcpcb *); 758 void tcp_discardcb(struct tcpcb *); 759 void tcp_twstart(struct tcpcb *); 760 #if 0 761 int tcp_twrecycleable(struct tcptw *tw); 762 #endif 763 void tcp_twclose(struct tcptw *_tw, int _reuse); 764 void tcp_ctlinput(int, struct sockaddr *, void *); 765 int tcp_ctloutput(struct socket *, struct sockopt *); 766 struct tcpcb * 767 tcp_drop(struct tcpcb *, int); 768 void tcp_drain(void); 769 void tcp_init(void); 770 #ifdef VIMAGE 771 void tcp_destroy(void); 772 #endif 773 void tcp_fini(void *); 774 char *tcp_log_addrs(struct in_conninfo *, struct tcphdr *, void *, 775 const void *); 776 char *tcp_log_vain(struct in_conninfo *, struct tcphdr *, void *, 777 const void *); 778 int tcp_reass(struct tcpcb *, struct tcphdr *, int *, struct mbuf *); 779 void tcp_reass_init(void); 780 void tcp_reass_flush(struct tcpcb *); 781 #ifdef VIMAGE 782 void tcp_reass_destroy(void); 783 #endif 784 void tcp_input(struct mbuf *, int); 785 u_long tcp_maxmtu(struct in_conninfo *, int *); 786 u_long tcp_maxmtu6(struct in_conninfo *, int *); 787 void tcp_mss_update(struct tcpcb *, int, int, struct hc_metrics_lite *, 788 int *); 789 void tcp_mss(struct tcpcb *, int); 790 int tcp_mssopt(struct in_conninfo *); 791 struct inpcb * 792 tcp_drop_syn_sent(struct inpcb *, int); 793 struct inpcb * 794 tcp_mtudisc(struct inpcb *, int); 795 struct tcpcb * 796 tcp_newtcpcb(struct inpcb *); 797 int tcp_output(struct tcpcb *); 798 void tcp_respond(struct tcpcb *, void *, 799 struct tcphdr *, struct mbuf *, tcp_seq, tcp_seq, int); 800 void tcp_tw_init(void); 801 #ifdef VIMAGE 802 void tcp_tw_destroy(void); 803 #endif 804 void tcp_tw_zone_change(void); 805 int tcp_twcheck(struct inpcb *, struct tcpopt *, struct tcphdr *, 806 struct mbuf *, int); 807 int tcp_twrespond(struct tcptw *, int); 808 void tcp_setpersist(struct tcpcb *); 809 #ifdef TCP_SIGNATURE 810 int tcp_signature_compute(struct mbuf *, int, int, int, u_char *, u_int); 811 int tcp_signature_verify(struct mbuf *, int, int, int, struct tcpopt *, 812 struct tcphdr *, u_int); 813 #endif 814 void tcp_slowtimo(void); 815 struct tcptemp * 816 tcpip_maketemplate(struct inpcb *); 817 void tcpip_fillheaders(struct inpcb *, void *, void *); 818 void tcp_timer_activate(struct tcpcb *, int, u_int); 819 int tcp_timer_active(struct tcpcb *, int); 820 void tcp_trace(short, short, struct tcpcb *, void *, struct tcphdr *, int); 821 /* 822 * All tcp_hc_* functions are IPv4 and IPv6 (via in_conninfo) 823 */ 824 void tcp_hc_init(void); 825 #ifdef VIMAGE 826 void tcp_hc_destroy(void); 827 #endif 828 void tcp_hc_get(struct in_conninfo *, struct hc_metrics_lite *); 829 u_long tcp_hc_getmtu(struct in_conninfo *); 830 void tcp_hc_updatemtu(struct in_conninfo *, u_long); 831 void tcp_hc_update(struct in_conninfo *, struct hc_metrics_lite *); 832 833 extern struct pr_usrreqs tcp_usrreqs; 834 tcp_seq tcp_new_isn(struct tcpcb *); 835 836 void tcp_sack_doack(struct tcpcb *, struct tcpopt *, tcp_seq); 837 void tcp_update_sack_list(struct tcpcb *tp, tcp_seq rcv_laststart, tcp_seq rcv_lastend); 838 void tcp_clean_sackreport(struct tcpcb *tp); 839 void tcp_sack_adjust(struct tcpcb *tp); 840 struct sackhole *tcp_sack_output(struct tcpcb *tp, int *sack_bytes_rexmt); 841 void tcp_sack_partialack(struct tcpcb *, struct tcphdr *); 842 void tcp_free_sackholes(struct tcpcb *tp); 843 int tcp_newreno(struct tcpcb *, struct tcphdr *); 844 u_long tcp_seq_subtract(u_long, u_long ); 845 846 void cc_cong_signal(struct tcpcb *tp, struct tcphdr *th, uint32_t type); 847 848 #endif /* _KERNEL */ 849 850 #endif /* _NETINET_TCP_VAR_H_ */ 851