bsd/netinet/tcp_var.h

   1 /*
   2  * Copyright (c) 2006 Apple Computer, Inc. All Rights Reserved.
   3  *
   4  * @APPLE_LICENSE_OSREFERENCE_HEADER_START@
   5  *
   6  * This file contains Original Code and/or Modifications of Original Code
   7  * as defined in and that are subject to the Apple Public Source License
   8  * Version 2.0 (the 'License'). You may not use this file except in
   9  * compliance with the License.  The rights granted to you under the
  10  * License may not be used to create, or enable the creation or
  11  * redistribution of, unlawful or unlicensed copies of an Apple operating
  12  * system, or to circumvent, violate, or enable the circumvention or
  13  * violation of, any terms of an Apple operating system software license
  14  * agreement.
  15  *
  16  * Please obtain a copy of the License at
  17  * http://www.opensource.apple.com/apsl/ and read it before using this
  18  * file.
  19  *
  20  * The Original Code and all software distributed under the License are
  21  * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
  22  * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
  23  * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
  24  * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
  25  * Please see the License for the specific language governing rights and
  26  * limitations under the License.
  27  *
  28  * @APPLE_LICENSE_OSREFERENCE_HEADER_END@
  29  */
  30 /*
  31  * Copyright (c) 1982, 1986, 1993, 1994, 1995
  32  *      The Regents of the University of California.  All rights reserved.
  33  *
  34  * Redistribution and use in source and binary forms, with or without
  35  * modification, are permitted provided that the following conditions
  36  * are met:
  37  * 1. Redistributions of source code must retain the above copyright
  38  *    notice, this list of conditions and the following disclaimer.
  39  * 2. Redistributions in binary form must reproduce the above copyright
  40  *    notice, this list of conditions and the following disclaimer in the
  41  *    documentation and/or other materials provided with the distribution.
  42  * 3. All advertising materials mentioning features or use of this software
  43  *    must display the following acknowledgement:
  44  *      This product includes software developed by the University of
  45  *      California, Berkeley and its contributors.
  46  * 4. Neither the name of the University nor the names of its contributors
  47  *    may be used to endorse or promote products derived from this software
  48  *    without specific prior written permission.
  49  *
  50  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  51  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  52  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  53  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  54  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  55  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  56  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  57  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  58  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  59  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  60  * SUCH DAMAGE.
  61  *
  62  *      @(#)tcp_var.h   8.4 (Berkeley) 5/24/95
  63  * $FreeBSD: src/sys/netinet/tcp_var.h,v 1.56.2.8 2001/08/22 00:59:13 silby Exp $
  64  */
  65
  66 #ifndef _NETINET_TCP_VAR_H_
  67 #define _NETINET_TCP_VAR_H_
  68 #include <sys/appleapiopts.h>
  69 #include <sys/queue.h>
  70 #include <netinet/in_pcb.h>
  71 #include <netinet/tcp_timer.h>
  72
  73 #ifdef KERNEL_PRIVATE
  74
  75 #define N_TIME_WAIT_SLOTS   128                /* must be power of 2 */
  76
  77 /*
  78  * Kernel variables for tcp.
  79  */
  80
  81 /* TCP segment queue entry */
  82 struct tseg_qent {
  83         LIST_ENTRY(tseg_qent) tqe_q;
  84         int     tqe_len;                /* TCP segment data length */
  85         struct  tcphdr *tqe_th;         /* a pointer to tcp header */
  86         struct  mbuf    *tqe_m;         /* mbuf contains packet */
  87 };
  88 LIST_HEAD(tsegqe_head, tseg_qent);
  89 extern int      tcp_reass_maxseg;
  90 extern int      tcp_reass_qsize;
  91 #ifdef MALLOC_DECLARE
  92 MALLOC_DECLARE(M_TSEGQ);
  93 #endif
  94
  95 struct sackblk {
  96         tcp_seq start;          /* start seq no. of sack block */
  97         tcp_seq end;            /* end seq no. */
  98 };
  99
 100 struct sackhole {
 101         tcp_seq start;          /* start seq no. of hole */
 102         tcp_seq end;            /* end seq no. */
 103         tcp_seq rxmit;          /* next seq. no in hole to be retransmitted */
 104         TAILQ_ENTRY(sackhole) scblink;  /* scoreboard linkage */
 105 };
 106
 107 struct sackhint {
 108         struct sackhole *nexthole;
 109         int     sack_bytes_rexmit;
 110 };
 111
 112 struct tcptemp {
 113         u_char  tt_ipgen[40]; /* the size must be of max ip header, now IPv6 */
 114         struct  tcphdr tt_t;
 115 };
 116
 117 #define tcp6cb          tcpcb  /* for KAME src sync over BSD*'s */
 118
 119 /*
 120  * Tcp control block, one per tcp; fields:
 121  * Organized for 16 byte cacheline efficiency.
 122  */
 123 struct tcpcb {
 124         struct  tsegqe_head t_segq;
 125         int     t_dupacks;              /* consecutive dup acks recd */
 126         struct  tcptemp *unused;        /* unused now: was t_template */
 127
 128         int     t_timer[TCPT_NTIMERS];  /* tcp timers */
 129
 130         struct  inpcb *t_inpcb;         /* back pointer to internet pcb */
 131         int     t_state;                /* state of this connection */
 132         u_int   t_flags;
 133 #define TF_ACKNOW       0x00001         /* ack peer immediately */
 134 #define TF_DELACK       0x00002         /* ack, but try to delay it */
 135 #define TF_NODELAY      0x00004         /* don't delay packets to coalesce */
 136 #define TF_NOOPT        0x00008         /* don't use tcp options */
 137 #define TF_SENTFIN      0x00010         /* have sent FIN */
 138 #define TF_REQ_SCALE    0x00020         /* have/will request window scaling */
 139 #define TF_RCVD_SCALE   0x00040         /* other side has requested scaling */
 140 #define TF_REQ_TSTMP    0x00080         /* have/will request timestamps */
 141 #define TF_RCVD_TSTMP   0x00100         /* a timestamp was received in SYN */
 142 #define TF_SACK_PERMIT  0x00200         /* other side said I could SACK */
 143 #define TF_NEEDSYN      0x00400         /* send SYN (implicit state) */
 144 #define TF_NEEDFIN      0x00800         /* send FIN (implicit state) */
 145 #define TF_NOPUSH       0x01000         /* don't push */
 146 #define TF_REQ_CC       0x02000         /* have/will request CC */
 147 #define TF_RCVD_CC      0x04000         /* a CC was received in SYN */
 148 #define TF_SENDCCNEW    0x08000         /* send CCnew instead of CC in SYN */
 149 #define TF_MORETOCOME   0x10000         /* More data to be appended to sock */
 150 #define TF_LQ_OVERFLOW  0x20000         /* listen queue overflow */
 151 #define TF_RXWIN0SENT   0x40000         /* sent a receiver win 0 in response */
 152 #define TF_SLOWLINK     0x80000         /* route is a on a modem speed link */
 153
 154
 155 #define TF_LASTIDLE     0x100000                /* connection was previously idle */
 156 #define TF_FASTRECOVERY 0x200000        /* in NewReno Fast Recovery */
 157 #define TF_WASFRECOVERY 0x400000        /* was in NewReno Fast Recovery */
 158 #define TF_SIGNATURE    0x800000        /* require MD5 digests (RFC2385) */
 159 #define TF_MAXSEGSNT    0x1000000       /* last segment sent was a full segment */
 160
 161         int     t_force;                /* 1 if forcing out a byte */
 162
 163         tcp_seq snd_una;                /* send unacknowledged */
 164         tcp_seq snd_max;                /* highest sequence number sent;
 165                                          * used to recognize retransmits
 166                                          */
 167         tcp_seq snd_nxt;                /* send next */
 168         tcp_seq snd_up;                 /* send urgent pointer */
 169
 170         tcp_seq snd_wl1;                /* window update seg seq number */
 171         tcp_seq snd_wl2;                /* window update seg ack number */
 172         tcp_seq iss;                    /* initial send sequence number */
 173         tcp_seq irs;                    /* initial receive sequence number */
 174
 175         tcp_seq rcv_nxt;                /* receive next */
 176         tcp_seq rcv_adv;                /* advertised window */
 177         u_long  rcv_wnd;                /* receive window */
 178         tcp_seq rcv_up;                 /* receive urgent pointer */
 179
 180         u_long  snd_wnd;                /* send window */
 181         u_long  snd_cwnd;               /* congestion-controlled window */
 182         u_long  snd_ssthresh;           /* snd_cwnd size threshold for
 183                                          * for slow start exponential to
 184                                          * linear switch
 185                                          */
 186         u_int   t_maxopd;               /* mss plus options */
 187
 188         u_long  t_rcvtime;              /* inactivity time */
 189         u_long  t_starttime;            /* time connection was established */
 190         int     t_rtttime;              /* round trip time */
 191         tcp_seq t_rtseq;                /* sequence number being timed */
 192
 193         int     t_rxtcur;               /* current retransmit value (ticks) */
 194         u_int   t_maxseg;               /* maximum segment size */
 195         int     t_srtt;                 /* smoothed round-trip time */
 196         int     t_rttvar;               /* variance in round-trip time */
 197
 198         int     t_rxtshift;             /* log(2) of rexmt exp. backoff */
 199         u_int   t_rttmin;               /* minimum rtt allowed */
 200         u_long  t_rttupdated;           /* number of times rtt sampled */
 201         u_long  max_sndwnd;             /* largest window peer has offered */
 202
 203         int     t_softerror;            /* possible error not yet reported */
 204 /* out-of-band data */
 205         char    t_oobflags;             /* have some */
 206         char    t_iobc;                 /* input character */
 207 #define TCPOOB_HAVEDATA 0x01
 208 #define TCPOOB_HADDATA  0x02
 209 /* RFC 1323 variables */
 210         u_char  snd_scale;              /* window scaling for send window */
 211         u_char  rcv_scale;              /* window scaling for recv window */
 212         u_char  request_r_scale;        /* pending window scaling */
 213         u_char  requested_s_scale;
 214         u_long  ts_recent;              /* timestamp echo data */
 215
 216         u_long  ts_recent_age;          /* when last updated */
 217         tcp_seq last_ack_sent;
 218 /* RFC 1644 variables */
 219         tcp_cc  cc_send;                /* send connection count */
 220         tcp_cc  cc_recv;                /* receive connection count */
 221         tcp_seq snd_recover;    /* for use in NewReno Fast Recovery */
 222 /* experimental */
 223         u_long  snd_cwnd_prev;          /* cwnd prior to retransmit */
 224         u_long  snd_ssthresh_prev;      /* ssthresh prior to retransmit */
 225         u_long  t_badrxtwin;            /* window for retransmit recovery */
 226
 227         int     t_keepidle;             /* keepalive idle timer (override global if > 0) */
 228         int     t_lastchain;            /* amount of packets chained last time around */
 229
 230 /* 3529618 MSS overload prevention */
 231         u_long  rcv_reset;
 232         u_long  rcv_pps;
 233         u_long  rcv_byps;
 234         tcp_seq snd_high;               /* for use in NewReno Fast Recovery */
 235         tcp_seq snd_high_prev;  /* snd_high prior to retransmit */
 236
 237         tcp_seq snd_recover_prev;       /* snd_recover prior to retransmit */
 238         u_char  snd_limited;            /* segments limited transmitted */
 239 /* anti DoS counters */
 240         u_long  rcv_second;             /* start of interval second */
 241 /* SACK related state */
 242         int     sack_enable;            /* enable SACK for this connection */
 243         int     snd_numholes;           /* number of holes seen by sender */
 244
 245         TAILQ_HEAD(sackhole_head, sackhole) snd_holes;
 246                                                 /* SACK scoreboard (sorted) */
 247         tcp_seq snd_fack;               /* last seq number(+1) sack'd by rcv'r*/
 248         int     rcv_numsacks;           /* # distinct sack blks present */
 249         struct sackblk sackblks[MAX_SACK_BLKS]; /* seq nos. of sack blocks */
 250         tcp_seq sack_newdata;           /* New data xmitted in this recovery
 251                                            episode starts at this seq number */
 252         struct sackhint sackhint;       /* SACK scoreboard hint */
 253         int     t_rttlow;               /* smallest observerved RTT */
 254 };
 255
 256 #define IN_FASTRECOVERY(tp)     (tp->t_flags & TF_FASTRECOVERY)
 257 #define ENTER_FASTRECOVERY(tp)  tp->t_flags |= TF_FASTRECOVERY
 258 #define EXIT_FASTRECOVERY(tp)   tp->t_flags &= ~TF_FASTRECOVERY
 259
 260
 261 /*
 262  * Structure to hold TCP options that are only used during segment
 263  * processing (in tcp_input), but not held in the tcpcb.
 264  * It's basically used to reduce the number of parameters
 265  * to tcp_dooptions.
 266  */
 267 struct tcpopt {
 268         u_long  to_flags;               /* which options are present */
 269 #define TOF_TS          0x0001          /* timestamp */
 270 #define TOF_MSS         0x0010
 271 #define TOF_SCALE       0x0020
 272 #define TOF_SIGNATURE   0x0040  /* signature option present */
 273 #define TOF_SIGLEN      0x0080  /* signature length valid (RFC2385) */
 274 #define TOF_SACK        0x0100          /* Peer sent SACK option */
 275         u_long          to_tsval;
 276         u_long          to_tsecr;
 277         u_int16_t       to_mss;
 278         u_int8_t        to_requested_s_scale;
 279         u_int8_t        to_nsacks;      /* number of SACK blocks */
 280         u_char          *to_sacks;      /* pointer to the first SACK blocks */
 281 };
 282
 283 /*
 284  * The TAO cache entry which is stored in the protocol family specific
 285  * portion of the route metrics.
 286  */
 287 struct rmxp_tao {
 288         tcp_cc  tao_cc;                 /* latest CC in valid SYN */
 289         tcp_cc  tao_ccsent;             /* latest CC sent to peer */
 290         u_short tao_mssopt;             /* peer's cached MSS */
 291 #ifdef notyet
 292         u_short tao_flags;              /* cache status flags */
 293 #define TAOF_DONT       0x0001          /* peer doesn't understand rfc1644 */
 294 #define TAOF_OK         0x0002          /* peer does understand rfc1644 */
 295 #define TAOF_UNDEF      0               /* we don't know yet */
 296 #endif /* notyet */
 297 };
 298 #define rmx_taop(r)     ((struct rmxp_tao *)(r).rmx_filler)
 299
 300 #define intotcpcb(ip)   ((struct tcpcb *)(ip)->inp_ppcb)
 301 #define sototcpcb(so)   (intotcpcb(sotoinpcb(so)))
 302
 303 /*
 304  * The smoothed round-trip time and estimated variance
 305  * are stored as fixed point numbers scaled by the values below.
 306  * For convenience, these scales are also used in smoothing the average
 307  * (smoothed = (1/scale)sample + ((scale-1)/scale)smoothed).
 308  * With these scales, srtt has 3 bits to the right of the binary point,
 309  * and thus an "ALPHA" of 0.875.  rttvar has 2 bits to the right of the
 310  * binary point, and is smoothed with an ALPHA of 0.75.
 311  */
 312 #define TCP_RTT_SCALE           32      /* multiplier for srtt; 3 bits frac. */
 313 #define TCP_RTT_SHIFT           5       /* shift for srtt; 3 bits frac. */
 314 #define TCP_RTTVAR_SCALE        16      /* multiplier for rttvar; 2 bits */
 315 #define TCP_RTTVAR_SHIFT        4       /* shift for rttvar; 2 bits */
 316 #define TCP_DELTA_SHIFT         2       /* see tcp_input.c */
 317
 318 /*
 319  * The initial retransmission should happen at rtt + 4 * rttvar.
 320  * Because of the way we do the smoothing, srtt and rttvar
 321  * will each average +1/2 tick of bias.  When we compute
 322  * the retransmit timer, we want 1/2 tick of rounding and
 323  * 1 extra tick because of +-1/2 tick uncertainty in the
 324  * firing of the timer.  The bias will give us exactly the
 325  * 1.5 tick we need.  But, because the bias is
 326  * statistical, we have to test that we don't drop below
 327  * the minimum feasible timer (which is 2 ticks).
 328  * This version of the macro adapted from a paper by Lawrence
 329  * Brakmo and Larry Peterson which outlines a problem caused
 330  * by insufficient precision in the original implementation,
 331  * which results in inappropriately large RTO values for very
 332  * fast networks.
 333  */
 334 #define TCP_REXMTVAL(tp) \
 335         max((tp)->t_rttmin, (((tp)->t_srtt >> (TCP_RTT_SHIFT - TCP_DELTA_SHIFT))  \
 336           + (tp)->t_rttvar) >> TCP_DELTA_SHIFT)
 337
 338 /*
 339  * Jaguar compatible TCP control block, for xtcpcb
 340  * Does not have the old fields
 341  */
 342 struct otcpcb {
 343 #else
 344 struct tseg_qent;
 345 LIST_HEAD(tsegqe_head, tseg_qent);
 346
 347 struct tcpcb {
 348 #endif /* KERNEL_PRIVATE */
 349         struct  tsegqe_head t_segq;
 350         int     t_dupacks;              /* consecutive dup acks recd */
 351         struct  tcptemp *unused;        /* unused now: was t_template */
 352
 353         int     t_timer[TCPT_NTIMERS];  /* tcp timers */
 354
 355         struct  inpcb *t_inpcb;         /* back pointer to internet pcb */
 356         int     t_state;                /* state of this connection */
 357         u_int   t_flags;
 358 #define TF_ACKNOW       0x00001         /* ack peer immediately */
 359 #define TF_DELACK       0x00002         /* ack, but try to delay it */
 360 #define TF_NODELAY      0x00004         /* don't delay packets to coalesce */
 361 #define TF_NOOPT        0x00008         /* don't use tcp options */
 362 #define TF_SENTFIN      0x00010         /* have sent FIN */
 363 #define TF_REQ_SCALE    0x00020         /* have/will request window scaling */
 364 #define TF_RCVD_SCALE   0x00040         /* other side has requested scaling */
 365 #define TF_REQ_TSTMP    0x00080         /* have/will request timestamps */
 366 #define TF_RCVD_TSTMP   0x00100         /* a timestamp was received in SYN */
 367 #define TF_SACK_PERMIT  0x00200         /* other side said I could SACK */
 368 #define TF_NEEDSYN      0x00400         /* send SYN (implicit state) */
 369 #define TF_NEEDFIN      0x00800         /* send FIN (implicit state) */
 370 #define TF_NOPUSH       0x01000         /* don't push */
 371 #define TF_REQ_CC       0x02000         /* have/will request CC */
 372 #define TF_RCVD_CC      0x04000         /* a CC was received in SYN */
 373 #define TF_SENDCCNEW    0x08000         /* send CCnew instead of CC in SYN */
 374 #define TF_MORETOCOME   0x10000         /* More data to be appended to sock */
 375 #define TF_LQ_OVERFLOW  0x20000         /* listen queue overflow */
 376 #define TF_RXWIN0SENT   0x40000         /* sent a receiver win 0 in response */
 377 #define TF_SLOWLINK     0x80000         /* route is a on a modem speed link */
 378
 379         int     t_force;                /* 1 if forcing out a byte */
 380
 381         tcp_seq snd_una;                /* send unacknowledged */
 382         tcp_seq snd_max;                /* highest sequence number sent;
 383                                          * used to recognize retransmits
 384                                          */
 385         tcp_seq snd_nxt;                /* send next */
 386         tcp_seq snd_up;                 /* send urgent pointer */
 387
 388         tcp_seq snd_wl1;                /* window update seg seq number */
 389         tcp_seq snd_wl2;                /* window update seg ack number */
 390         tcp_seq iss;                    /* initial send sequence number */
 391         tcp_seq irs;                    /* initial receive sequence number */
 392
 393         tcp_seq rcv_nxt;                /* receive next */
 394         tcp_seq rcv_adv;                /* advertised window */
 395         u_long  rcv_wnd;                /* receive window */
 396         tcp_seq rcv_up;                 /* receive urgent pointer */
 397
 398         u_long  snd_wnd;                /* send window */
 399         u_long  snd_cwnd;               /* congestion-controlled window */
 400         u_long  snd_ssthresh;           /* snd_cwnd size threshold for
 401                                          * for slow start exponential to
 402                                          * linear switch
 403                                          */
 404         u_int   t_maxopd;               /* mss plus options */
 405
 406         u_long  t_rcvtime;              /* inactivity time */
 407         u_long  t_starttime;            /* time connection was established */
 408         int     t_rtttime;              /* round trip time */
 409         tcp_seq t_rtseq;                /* sequence number being timed */
 410
 411         int     t_rxtcur;               /* current retransmit value (ticks) */
 412         u_int   t_maxseg;               /* maximum segment size */
 413         int     t_srtt;                 /* smoothed round-trip time */
 414         int     t_rttvar;               /* variance in round-trip time */
 415
 416         int     t_rxtshift;             /* log(2) of rexmt exp. backoff */
 417         u_int   t_rttmin;               /* minimum rtt allowed */
 418         u_long  t_rttupdated;           /* number of times rtt sampled */
 419         u_long  max_sndwnd;             /* largest window peer has offered */
 420
 421         int     t_softerror;            /* possible error not yet reported */
 422 /* out-of-band data */
 423         char    t_oobflags;             /* have some */
 424         char    t_iobc;                 /* input character */
 425 #define TCPOOB_HAVEDATA 0x01
 426 #define TCPOOB_HADDATA  0x02
 427 /* RFC 1323 variables */
 428         u_char  snd_scale;              /* window scaling for send window */
 429         u_char  rcv_scale;              /* window scaling for recv window */
 430         u_char  request_r_scale;        /* pending window scaling */
 431         u_char  requested_s_scale;
 432         u_long  ts_recent;              /* timestamp echo data */
 433
 434         u_long  ts_recent_age;          /* when last updated */
 435         tcp_seq last_ack_sent;
 436 /* RFC 1644 variables */
 437         tcp_cc  cc_send;                /* send connection count */
 438         tcp_cc  cc_recv;                /* receive connection count */
 439         tcp_seq snd_recover;    /* for use in fast recovery */
 440 /* experimental */
 441         u_long  snd_cwnd_prev;          /* cwnd prior to retransmit */
 442         u_long  snd_ssthresh_prev;      /* ssthresh prior to retransmit */
 443         u_long  t_badrxtwin;            /* window for retransmit recovery */
 444 };
 445
 446 /*
 447  * TCP statistics.
 448  * Many of these should be kept per connection,
 449  * but that's inconvenient at the moment.
 450  */
 451 struct  tcpstat {
 452         u_long  tcps_connattempt;       /* connections initiated */
 453         u_long  tcps_accepts;           /* connections accepted */
 454         u_long  tcps_connects;          /* connections established */
 455         u_long  tcps_drops;             /* connections dropped */
 456         u_long  tcps_conndrops;         /* embryonic connections dropped */
 457         u_long  tcps_closed;            /* conn. closed (includes drops) */
 458         u_long  tcps_segstimed;         /* segs where we tried to get rtt */
 459         u_long  tcps_rttupdated;        /* times we succeeded */
 460         u_long  tcps_delack;            /* delayed acks sent */
 461         u_long  tcps_timeoutdrop;       /* conn. dropped in rxmt timeout */
 462         u_long  tcps_rexmttimeo;        /* retransmit timeouts */
 463         u_long  tcps_persisttimeo;      /* persist timeouts */
 464         u_long  tcps_keeptimeo;         /* keepalive timeouts */
 465         u_long  tcps_keepprobe;         /* keepalive probes sent */
 466         u_long  tcps_keepdrops;         /* connections dropped in keepalive */
 467
 468         u_long  tcps_sndtotal;          /* total packets sent */
 469         u_long  tcps_sndpack;           /* data packets sent */
 470         u_long  tcps_sndbyte;           /* data bytes sent */
 471         u_long  tcps_sndrexmitpack;     /* data packets retransmitted */
 472         u_long  tcps_sndrexmitbyte;     /* data bytes retransmitted */
 473         u_long  tcps_sndacks;           /* ack-only packets sent */
 474         u_long  tcps_sndprobe;          /* window probes sent */
 475         u_long  tcps_sndurg;            /* packets sent with URG only */
 476         u_long  tcps_sndwinup;          /* window update-only packets sent */
 477         u_long  tcps_sndctrl;           /* control (SYN|FIN|RST) packets sent */
 478
 479         u_long  tcps_rcvtotal;          /* total packets received */
 480         u_long  tcps_rcvpack;           /* packets received in sequence */
 481         u_long  tcps_rcvbyte;           /* bytes received in sequence */
 482         u_long  tcps_rcvbadsum;         /* packets received with ccksum errs */
 483         u_long  tcps_rcvbadoff;         /* packets received with bad offset */
 484         u_long  tcps_rcvmemdrop;        /* packets dropped for lack of memory */
 485         u_long  tcps_rcvshort;          /* packets received too short */
 486         u_long  tcps_rcvduppack;        /* duplicate-only packets received */
 487         u_long  tcps_rcvdupbyte;        /* duplicate-only bytes received */
 488         u_long  tcps_rcvpartduppack;    /* packets with some duplicate data */
 489         u_long  tcps_rcvpartdupbyte;    /* dup. bytes in part-dup. packets */
 490         u_long  tcps_rcvoopack;         /* out-of-order packets received */
 491         u_long  tcps_rcvoobyte;         /* out-of-order bytes received */
 492         u_long  tcps_rcvpackafterwin;   /* packets with data after window */
 493         u_long  tcps_rcvbyteafterwin;   /* bytes rcvd after window */
 494         u_long  tcps_rcvafterclose;     /* packets rcvd after "close" */
 495         u_long  tcps_rcvwinprobe;       /* rcvd window probe packets */
 496         u_long  tcps_rcvdupack;         /* rcvd duplicate acks */
 497         u_long  tcps_rcvacktoomuch;     /* rcvd acks for unsent data */
 498         u_long  tcps_rcvackpack;        /* rcvd ack packets */
 499         u_long  tcps_rcvackbyte;        /* bytes acked by rcvd acks */
 500         u_long  tcps_rcvwinupd;         /* rcvd window update packets */
 501         u_long  tcps_pawsdrop;          /* segments dropped due to PAWS */
 502         u_long  tcps_predack;           /* times hdr predict ok for acks */
 503         u_long  tcps_preddat;           /* times hdr predict ok for data pkts */
 504         u_long  tcps_pcbcachemiss;
 505         u_long  tcps_cachedrtt;         /* times cached RTT in route updated */
 506         u_long  tcps_cachedrttvar;      /* times cached rttvar updated */
 507         u_long  tcps_cachedssthresh;    /* times cached ssthresh updated */
 508         u_long  tcps_usedrtt;           /* times RTT initialized from route */
 509         u_long  tcps_usedrttvar;        /* times RTTVAR initialized from rt */
 510         u_long  tcps_usedssthresh;      /* times ssthresh initialized from rt*/
 511         u_long  tcps_persistdrop;       /* timeout in persist state */
 512         u_long  tcps_badsyn;            /* bogus SYN, e.g. premature ACK */
 513         u_long  tcps_mturesent;         /* resends due to MTU discovery */
 514         u_long  tcps_listendrop;        /* listen queue overflows */
 515
 516         /* SACK related stats */
 517         u_long  tcps_sack_recovery_episode; /* SACK recovery episodes */
 518         u_long  tcps_sack_rexmits;          /* SACK rexmit segments   */
 519         u_long  tcps_sack_rexmit_bytes;     /* SACK rexmit bytes      */
 520         u_long  tcps_sack_rcv_blocks;       /* SACK blocks (options) received */
 521         u_long  tcps_sack_send_blocks;      /* SACK blocks (options) sent     */
 522         u_long  tcps_sack_sboverflow;       /* SACK sendblock overflow   */
 523
 524 };
 525
 526 /*
 527  * TCB structure exported to user-land via sysctl(3).
 528  * Evil hack: declare only if in_pcb.h and sys/socketvar.h have been
 529  * included.  Not all of our clients do.
 530  */
 531 struct  xtcpcb {
 532         size_t  xt_len;
 533 #ifdef KERNEL_PRIVATE
 534         struct  inpcb_compat    xt_inp;
 535 #else
 536         struct  inpcb   xt_inp;
 537 #endif
 538 #ifdef KERNEL_PRIVATE
 539         struct  otcpcb  xt_tp;
 540 #else
 541         struct  tcpcb   xt_tp;
 542 #endif
 543         struct  xsocket xt_socket;
 544         u_quad_t        xt_alignment_hack;
 545 };
 546
 547 /*
 548  * Names for TCP sysctl objects
 549  */
 550 #define TCPCTL_DO_RFC1323       1       /* use RFC-1323 extensions */
 551 #define TCPCTL_DO_RFC1644       2       /* use RFC-1644 extensions */
 552 #define TCPCTL_MSSDFLT          3       /* MSS default */
 553 #define TCPCTL_STATS            4       /* statistics (read-only) */
 554 #define TCPCTL_RTTDFLT          5       /* default RTT estimate */
 555 #define TCPCTL_KEEPIDLE         6       /* keepalive idle timer */
 556 #define TCPCTL_KEEPINTVL        7       /* interval to send keepalives */
 557 #define TCPCTL_SENDSPACE        8       /* send buffer space */
 558 #define TCPCTL_RECVSPACE        9       /* receive buffer space */
 559 #define TCPCTL_KEEPINIT         10      /* timeout for establishing syn */
 560 #define TCPCTL_PCBLIST          11      /* list of all outstanding PCBs */
 561 #define TCPCTL_DELACKTIME       12      /* time before sending delayed ACK */
 562 #define TCPCTL_V6MSSDFLT        13      /* MSS default for IPv6 */
 563 #define TCPCTL_MAXID            14
 564
 565 #ifdef KERNEL_PRIVATE
 566 #define TCPCTL_NAMES { \
 567         { 0, 0 }, \
 568         { "rfc1323", CTLTYPE_INT }, \
 569         { "rfc1644", CTLTYPE_INT }, \
 570         { "mssdflt", CTLTYPE_INT }, \
 571         { "stats", CTLTYPE_STRUCT }, \
 572         { "rttdflt", CTLTYPE_INT }, \
 573         { "keepidle", CTLTYPE_INT }, \
 574         { "keepintvl", CTLTYPE_INT }, \
 575         { "sendspace", CTLTYPE_INT }, \
 576         { "recvspace", CTLTYPE_INT }, \
 577         { "keepinit", CTLTYPE_INT }, \
 578         { "pcblist", CTLTYPE_STRUCT }, \
 579         { "delacktime", CTLTYPE_INT }, \
 580         { "v6mssdflt", CTLTYPE_INT }, \
 581 }
 582
 583 #ifdef SYSCTL_DECL
 584 SYSCTL_DECL(_net_inet_tcp);
 585 #endif /* SYSCTL_DECL */
 586
 587 extern  struct inpcbhead tcb;           /* head of queue of active tcpcb's */
 588 extern  struct inpcbinfo tcbinfo;
 589 extern  struct tcpstat tcpstat; /* tcp statistics */
 590 extern  int tcp_mssdflt;        /* XXX */
 591 extern  int tcp_minmss;
 592 extern  int tcp_minmssoverload;
 593 extern  int tcp_do_newreno;
 594 extern  int ss_fltsz;
 595 extern  int ss_fltsz_local;
 596 #ifdef __APPLE__
 597 extern  u_long tcp_now;         /* for RFC 1323 timestamps */
 598 extern  int tcp_delack_enabled;
 599 #endif /* __APPLE__ */
 600
 601 extern  int tcp_do_sack;        /* SACK enabled/disabled */
 602
 603 void     tcp_canceltimers(struct tcpcb *);
 604 struct tcpcb *
 605          tcp_close(struct tcpcb *);
 606 void     tcp_ctlinput(int, struct sockaddr *, void *);
 607 int      tcp_ctloutput(struct socket *, struct sockopt *);
 608 struct tcpcb *
 609          tcp_drop(struct tcpcb *, int);
 610 void     tcp_drain(void);
 611 void     tcp_fasttimo(void);
 612 struct rmxp_tao *
 613          tcp_gettaocache(struct inpcb *);
 614 void     tcp_init(void);
 615 void     tcp_input(struct mbuf *, int);
 616 void     tcp_mss(struct tcpcb *, int);
 617 int      tcp_mssopt(struct tcpcb *);
 618 void     tcp_drop_syn_sent(struct inpcb *, int);
 619 void     tcp_mtudisc(struct inpcb *, int);
 620 struct tcpcb *
 621          tcp_newtcpcb(struct inpcb *);
 622 int      tcp_output(struct tcpcb *);
 623 void     tcp_quench(struct inpcb *, int);
 624 void     tcp_respond(struct tcpcb *, void *,
 625             struct tcphdr *, struct mbuf *, tcp_seq, tcp_seq, int);
 626 struct rtentry *
 627          tcp_rtlookup(struct inpcb *);
 628 void     tcp_setpersist(struct tcpcb *);
 629 void     tcp_slowtimo(void);
 630 struct tcptemp *
 631          tcp_maketemplate(struct tcpcb *);
 632 void     tcp_fillheaders(struct tcpcb *, void *, void *);
 633 struct tcpcb *
 634          tcp_timers(struct tcpcb *, int);
 635 void     tcp_trace(int, int, struct tcpcb *, void *, struct tcphdr *, int);
 636
 637 void     tcp_update_sack_list(struct tcpcb *tp, tcp_seq rcv_laststart, tcp_seq rcv_lastend);
 638 void     tcp_clean_sackreport(struct tcpcb *tp);
 639 void     tcp_sack_adjust(struct tcpcb *tp);
 640 struct sackhole *tcp_sack_output(struct tcpcb *tp, int *sack_bytes_rexmt);
 641 void     tcp_sack_partialack(struct tcpcb *, struct tcphdr *);
 642 void     tcp_free_sackholes(struct tcpcb *tp);
 643
 644
 645 int      tcp_lock (struct socket *, int, int);
 646 int      tcp_unlock (struct socket *, int, int);
 647 #ifdef _KERN_LOCKS_H_
 648 lck_mtx_t *      tcp_getlock (struct socket *, int);
 649 #else
 650 void *   tcp_getlock (struct socket *, int);
 651 #endif
 652
 653
 654 extern  struct pr_usrreqs tcp_usrreqs;
 655 extern  u_long tcp_sendspace;
 656 extern  u_long tcp_recvspace;
 657 tcp_seq tcp_new_isn(struct tcpcb *);
 658
 659 #endif /* KERNEL_RPIVATE */
 660
 661 #endif /* _NETINET_TCP_VAR_H_ */