tcp_var.h revision 1.31
1 /* $NetBSD: tcp_var.h,v 1.31 1997/12/17 06:06:41 thorpej Exp $ */ 2 3 /* 4 * Copyright (c) 1982, 1986, 1993, 1994 5 * The Regents of the University of California. All rights reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 3. All advertising materials mentioning features or use of this software 16 * must display the following acknowledgement: 17 * This product includes software developed by the University of 18 * California, Berkeley and its contributors. 19 * 4. Neither the name of the University nor the names of its contributors 20 * may be used to endorse or promote products derived from this software 21 * without specific prior written permission. 22 * 23 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 24 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 25 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 26 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 27 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 28 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 29 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 30 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 31 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 32 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 33 * SUCH DAMAGE. 34 * 35 * @(#)tcp_var.h 8.3 (Berkeley) 4/10/94 36 */ 37 38 /* 39 * Kernel variables for tcp. 40 */ 41 42 /* 43 * Tcp control block, one per tcp; fields: 44 */ 45 struct tcpcb { 46 struct ipqehead segq; /* sequencing queue */ 47 short t_state; /* state of this connection */ 48 short t_timer[TCPT_NTIMERS]; /* tcp timers */ 49 short t_rxtshift; /* log(2) of rexmt exp. backoff */ 50 short t_rxtcur; /* current retransmit value */ 51 short t_dupacks; /* consecutive dup acks recd */ 52 u_short t_peermss; /* peer's maximum segment size */ 53 u_short t_ourmss; /* our's maximum segment size */ 54 u_short t_segsz; /* current segment size in use */ 55 char t_force; /* 1 if forcing out a byte */ 56 u_short t_flags; 57 #define TF_ACKNOW 0x0001 /* ack peer immediately */ 58 #define TF_DELACK 0x0002 /* ack, but try to delay it */ 59 #define TF_NODELAY 0x0004 /* don't delay packets to coalesce */ 60 #define TF_NOOPT 0x0008 /* don't use tcp options */ 61 #define TF_SENTFIN 0x0010 /* have sent FIN */ 62 #define TF_REQ_SCALE 0x0020 /* have/will request window scaling */ 63 #define TF_RCVD_SCALE 0x0040 /* other side has requested scaling */ 64 #define TF_REQ_TSTMP 0x0080 /* have/will request timestamps */ 65 #define TF_RCVD_TSTMP 0x0100 /* a timestamp was received in SYN */ 66 #define TF_SACK_PERMIT 0x0200 /* other side said I could SACK */ 67 68 struct tcpiphdr *t_template; /* skeletal packet for transmit */ 69 struct inpcb *t_inpcb; /* back pointer to internet pcb */ 70 /* 71 * The following fields are used as in the protocol specification. 72 * See RFC783, Dec. 1981, page 21. 73 */ 74 /* send sequence variables */ 75 tcp_seq snd_una; /* send unacknowledged */ 76 tcp_seq snd_nxt; /* send next */ 77 tcp_seq snd_up; /* send urgent pointer */ 78 tcp_seq snd_wl1; /* window update seg seq number */ 79 tcp_seq snd_wl2; /* window update seg ack number */ 80 tcp_seq iss; /* initial send sequence number */ 81 u_long snd_wnd; /* send window */ 82 /* receive sequence variables */ 83 u_long rcv_wnd; /* receive window */ 84 tcp_seq rcv_nxt; /* receive next */ 85 tcp_seq rcv_up; /* receive urgent pointer */ 86 tcp_seq irs; /* initial receive sequence number */ 87 /* 88 * Additional variables for this implementation. 89 */ 90 /* receive variables */ 91 tcp_seq rcv_adv; /* advertised window */ 92 /* retransmit variables */ 93 tcp_seq snd_max; /* highest sequence number sent; 94 * used to recognize retransmits 95 */ 96 /* congestion control (for slow start, source quench, retransmit after loss) */ 97 u_long snd_cwnd; /* congestion-controlled window */ 98 u_long snd_ssthresh; /* snd_cwnd size threshhold for 99 * for slow start exponential to 100 * linear switch 101 */ 102 /* 103 * transmit timing stuff. See below for scale of srtt and rttvar. 104 * "Variance" is actually smoothed difference. 105 */ 106 short t_idle; /* inactivity time */ 107 short t_rtt; /* round trip time */ 108 tcp_seq t_rtseq; /* sequence number being timed */ 109 short t_srtt; /* smoothed round-trip time */ 110 short t_rttvar; /* variance in round-trip time */ 111 short t_rttmin; /* minimum rtt allowed */ 112 u_long max_sndwnd; /* largest window peer has offered */ 113 114 /* out-of-band data */ 115 char t_oobflags; /* have some */ 116 char t_iobc; /* input character */ 117 #define TCPOOB_HAVEDATA 0x01 118 #define TCPOOB_HADDATA 0x02 119 short t_softerror; /* possible error not yet reported */ 120 121 /* RFC 1323 variables */ 122 u_char snd_scale; /* window scaling for send window */ 123 u_char rcv_scale; /* window scaling for recv window */ 124 u_char request_r_scale; /* pending window scaling */ 125 u_char requested_s_scale; 126 u_int32_t ts_recent; /* timestamp echo data */ 127 u_int32_t ts_recent_age; /* when last updated */ 128 tcp_seq last_ack_sent; 129 130 /* TUBA stuff */ 131 caddr_t t_tuba_pcb; /* next level down pcb for TCP over z */ 132 }; 133 134 /* 135 * Handy way of passing around TCP option info. 136 */ 137 struct tcp_opt_info { 138 int ts_present; 139 u_int32_t ts_val; 140 u_int32_t ts_ecr; 141 u_int16_t maxseg; 142 }; 143 144 /* 145 * This structure should not exceed 32 bytes. 146 * XXX On the Alpha, it's already 36-bytes, which rounds to 40. 147 * XXX Need to eliminate the pointer. 148 * 149 * XXX We've blown 32 bytes on non-Alpha systems, too, since we're 150 * XXX storing the maxseg we advertised to the peer. Should we 151 * XXX create another malloc bucket? Should we care? 152 */ 153 struct syn_cache { 154 struct syn_cache *sc_next; 155 u_int32_t sc_tstmp : 1, 156 sc_hash : 31; 157 struct in_addr sc_src; 158 struct in_addr sc_dst; 159 tcp_seq sc_irs; 160 tcp_seq sc_iss; 161 u_int16_t sc_sport; 162 u_int16_t sc_dport; 163 u_int16_t sc_peermaxseg; 164 u_int16_t sc_ourmaxseg; 165 u_int8_t sc_timer; 166 u_int8_t sc_request_r_scale : 4, 167 sc_requested_s_scale : 4; 168 }; 169 170 struct syn_cache_head { 171 struct syn_cache *sch_first; /* First entry in the bucket */ 172 struct syn_cache *sch_last; /* Last entry in the bucket */ 173 struct syn_cache_head *sch_headq; /* The next non-empty bucket */ 174 int16_t sch_timer_sum; /* Total time in this bucket */ 175 u_int16_t sch_length; /* @ # elements in bucket */ 176 }; 177 178 #define intotcpcb(ip) ((struct tcpcb *)(ip)->inp_ppcb) 179 #define sototcpcb(so) (intotcpcb(sotoinpcb(so))) 180 181 /* 182 * The smoothed round-trip time and estimated variance 183 * are stored as fixed point numbers scaled by the values below. 184 * For convenience, these scales are also used in smoothing the average 185 * (smoothed = (1/scale)sample + ((scale-1)/scale)smoothed). 186 * With these scales, srtt has 3 bits to the right of the binary point, 187 * and thus an "ALPHA" of 0.875. rttvar has 2 bits to the right of the 188 * binary point, and is smoothed with an ALPHA of 0.75. 189 */ 190 #define TCP_RTT_SHIFT 3 /* shift for srtt; 3 bits frac. */ 191 #define TCP_RTTVAR_SHIFT 2 /* multiplier for rttvar; 2 bits */ 192 193 /* 194 * The initial retransmission should happen at rtt + 4 * rttvar. 195 * Because of the way we do the smoothing, srtt and rttvar 196 * will each average +1/2 tick of bias. When we compute 197 * the retransmit timer, we want 1/2 tick of rounding and 198 * 1 extra tick because of +-1/2 tick uncertainty in the 199 * firing of the timer. The bias will give us exactly the 200 * 1.5 tick we need. But, because the bias is 201 * statistical, we have to test that we don't drop below 202 * the minimum feasible timer (which is 2 ticks). 203 * This macro assumes that the value of 1<<TCP_RTTVAR_SHIFT 204 * is the same as the multiplier for rttvar. 205 */ 206 #define TCP_REXMTVAL(tp) \ 207 ((((tp)->t_srtt >> TCP_RTT_SHIFT) + (tp)->t_rttvar) >> 2) 208 209 #ifdef _KERNEL 210 /* 211 * Compute the initial window for slow start. 212 */ 213 extern int tcp_init_win; 214 #define TCP_INITIAL_WINDOW(segsz) \ 215 ((tcp_init_win == 0) ? (min(4 * (segsz), max(2 * (segsz), 4380))) : \ 216 ((segsz) * tcp_init_win)) 217 #endif /* _KERNEL */ 218 219 /* 220 * TCP statistics. 221 * Many of these should be kept per connection, 222 * but that's inconvenient at the moment. 223 */ 224 struct tcpstat { 225 u_long tcps_connattempt; /* connections initiated */ 226 u_long tcps_accepts; /* connections accepted */ 227 u_long tcps_connects; /* connections established */ 228 u_long tcps_drops; /* connections dropped */ 229 u_long tcps_conndrops; /* embryonic connections dropped */ 230 u_long tcps_closed; /* conn. closed (includes drops) */ 231 u_long tcps_segstimed; /* segs where we tried to get rtt */ 232 u_long tcps_rttupdated; /* times we succeeded */ 233 u_long tcps_delack; /* delayed acks sent */ 234 u_long tcps_timeoutdrop; /* conn. dropped in rxmt timeout */ 235 u_long tcps_rexmttimeo; /* retransmit timeouts */ 236 u_long tcps_persisttimeo; /* persist timeouts */ 237 u_long tcps_keeptimeo; /* keepalive timeouts */ 238 u_long tcps_keepprobe; /* keepalive probes sent */ 239 u_long tcps_keepdrops; /* connections dropped in keepalive */ 240 u_long tcps_persistdrops; /* connections dropped in persist */ 241 u_long tcps_connsdrained; /* connections drained due to memory 242 shortage */ 243 244 u_long tcps_sndtotal; /* total packets sent */ 245 u_long tcps_sndpack; /* data packets sent */ 246 u_long tcps_sndbyte; /* data bytes sent */ 247 u_long tcps_sndrexmitpack; /* data packets retransmitted */ 248 u_long tcps_sndrexmitbyte; /* data bytes retransmitted */ 249 u_long tcps_sndacks; /* ack-only packets sent */ 250 u_long tcps_sndprobe; /* window probes sent */ 251 u_long tcps_sndurg; /* packets sent with URG only */ 252 u_long tcps_sndwinup; /* window update-only packets sent */ 253 u_long tcps_sndctrl; /* control (SYN|FIN|RST) packets sent */ 254 255 u_long tcps_rcvtotal; /* total packets received */ 256 u_long tcps_rcvpack; /* packets received in sequence */ 257 u_long tcps_rcvbyte; /* bytes received in sequence */ 258 u_long tcps_rcvbadsum; /* packets received with ccksum errs */ 259 u_long tcps_rcvbadoff; /* packets received with bad offset */ 260 u_long tcps_rcvmemdrop; /* packets dropped for lack of memory */ 261 u_long tcps_rcvshort; /* packets received too short */ 262 u_long tcps_rcvduppack; /* duplicate-only packets received */ 263 u_long tcps_rcvdupbyte; /* duplicate-only bytes received */ 264 u_long tcps_rcvpartduppack; /* packets with some duplicate data */ 265 u_long tcps_rcvpartdupbyte; /* dup. bytes in part-dup. packets */ 266 u_long tcps_rcvoopack; /* out-of-order packets received */ 267 u_long tcps_rcvoobyte; /* out-of-order bytes received */ 268 u_long tcps_rcvpackafterwin; /* packets with data after window */ 269 u_long tcps_rcvbyteafterwin; /* bytes rcvd after window */ 270 u_long tcps_rcvafterclose; /* packets rcvd after "close" */ 271 u_long tcps_rcvwinprobe; /* rcvd window probe packets */ 272 u_long tcps_rcvdupack; /* rcvd duplicate acks */ 273 u_long tcps_rcvacktoomuch; /* rcvd acks for unsent data */ 274 u_long tcps_rcvackpack; /* rcvd ack packets */ 275 u_long tcps_rcvackbyte; /* bytes acked by rcvd acks */ 276 u_long tcps_rcvwinupd; /* rcvd window update packets */ 277 u_long tcps_pawsdrop; /* segments dropped due to PAWS */ 278 u_long tcps_predack; /* times hdr predict ok for acks */ 279 u_long tcps_preddat; /* times hdr predict ok for data pkts */ 280 281 u_long tcps_pcbhashmiss; /* input packets missing pcb hash */ 282 u_long tcps_noport; /* no socket on port */ 283 u_long tcps_badsyn; /* received ack for which we have 284 no SYN in compressed state */ 285 286 /* These statistics deal with the SYN cache. */ 287 u_long tcps_sc_added; /* # of entries added */ 288 u_long tcps_sc_completed; /* # of connections completed */ 289 u_long tcps_sc_timed_out; /* # of entries timed out */ 290 u_long tcps_sc_overflowed; /* # dropped due to overflow */ 291 u_long tcps_sc_reset; /* # dropped due to RST */ 292 u_long tcps_sc_unreach; /* # dropped due to ICMP unreach */ 293 u_long tcps_sc_bucketoverflow; /* # dropped due to bucket overflow */ 294 u_long tcps_sc_aborted; /* # of entries aborted (no mem) */ 295 u_long tcps_sc_dupesyn; /* # of duplicate SYNs received */ 296 u_long tcps_sc_dropped; /* # of SYNs dropped (no route/mem) */ 297 u_long tcps_sc_collisions; /* # of hash collisions */ 298 }; 299 300 /* 301 * Names for TCP sysctl objects. 302 */ 303 #define TCPCTL_RFC1323 1 /* RFC1323 timestamps/scaling */ 304 #define TCPCTL_SENDSPACE 2 /* default send buffer */ 305 #define TCPCTL_RECVSPACE 3 /* default recv buffer */ 306 #define TCPCTL_MSSDFLT 4 /* default seg size */ 307 #define TCPCTL_SYN_CACHE_LIMIT 5 /* max size of comp. state engine */ 308 #define TCPCTL_SYN_BUCKET_LIMIT 6 /* max size of hash bucket */ 309 #define TCPCTL_SYN_CACHE_INTER 7 /* interval of comp. state timer */ 310 #define TCPCTL_INIT_WIN 8 /* initial window */ 311 #define TCPCTL_MAXID 9 312 313 #define TCPCTL_NAMES { \ 314 { 0, 0 }, \ 315 { "rfc1323", CTLTYPE_INT }, \ 316 { "sendspace", CTLTYPE_INT }, \ 317 { "recvspace", CTLTYPE_INT }, \ 318 { "mssdflt", CTLTYPE_INT }, \ 319 { "syn_cache_limit", CTLTYPE_INT }, \ 320 { "syn_bucket_limit", CTLTYPE_INT }, \ 321 { "syn_cache_interval", CTLTYPE_INT },\ 322 { "init_win", CTLTYPE_INT }, \ 323 } 324 325 #ifdef _KERNEL 326 struct inpcbtable tcbtable; /* head of queue of active tcpcb's */ 327 struct tcpstat tcpstat; /* tcp statistics */ 328 u_int32_t tcp_now; /* for RFC 1323 timestamps */ 329 extern int tcp_do_rfc1323; /* enabled/disabled? */ 330 extern int tcp_mssdflt; /* default seg size */ 331 extern int tcp_init_win; /* initial window */ 332 extern int tcp_syn_cache_limit; /* max entries for compressed state engine */ 333 extern int tcp_syn_bucket_limit;/* max entries per hash bucket */ 334 extern int tcp_syn_cache_interval; /* compressed state timer */ 335 336 extern int tcp_syn_cache_size; 337 extern int tcp_syn_cache_timeo; 338 extern struct syn_cache_head tcp_syn_cache[], *tcp_syn_cache_first; 339 extern u_long syn_cache_count; 340 341 int tcp_attach __P((struct socket *)); 342 void tcp_canceltimers __P((struct tcpcb *)); 343 struct tcpcb * 344 tcp_close __P((struct tcpcb *)); 345 void *tcp_ctlinput __P((int, struct sockaddr *, void *)); 346 int tcp_ctloutput __P((int, struct socket *, int, int, struct mbuf **)); 347 struct tcpcb * 348 tcp_disconnect __P((struct tcpcb *)); 349 struct tcpcb * 350 tcp_drop __P((struct tcpcb *, int)); 351 void tcp_dooptions __P((struct tcpcb *, 352 u_char *, int, struct tcpiphdr *, struct tcp_opt_info *)); 353 void tcp_drain __P((void)); 354 void tcp_established __P((struct tcpcb *)); 355 void tcp_fasttimo __P((void)); 356 void tcp_init __P((void)); 357 void tcp_input __P((struct mbuf *, ...)); 358 int tcp_mss_to_advertise __P((const struct tcpcb *)); 359 void tcp_mss_from_peer __P((struct tcpcb *, int)); 360 void tcp_mtudisc __P((struct inpcb *, int)); 361 struct tcpcb * 362 tcp_newtcpcb __P((struct inpcb *)); 363 void tcp_notify __P((struct inpcb *, int)); 364 int tcp_output __P((struct tcpcb *)); 365 void tcp_pulloutofband __P((struct socket *, 366 struct tcpiphdr *, struct mbuf *)); 367 void tcp_quench __P((struct inpcb *, int)); 368 int tcp_reass __P((struct tcpcb *, struct tcpiphdr *, struct mbuf *)); 369 int tcp_respond __P((struct tcpcb *, 370 struct tcpiphdr *, struct mbuf *, tcp_seq, tcp_seq, int)); 371 void tcp_rmx_rtt __P((struct tcpcb *)); 372 void tcp_setpersist __P((struct tcpcb *)); 373 void tcp_slowtimo __P((void)); 374 struct tcpiphdr * 375 tcp_template __P((struct tcpcb *)); 376 struct tcpcb * 377 tcp_timers __P((struct tcpcb *, int)); 378 void tcp_trace __P((int, int, struct tcpcb *, struct tcpiphdr *, int)); 379 struct tcpcb * 380 tcp_usrclosed __P((struct tcpcb *)); 381 int tcp_sysctl __P((int *, u_int, void *, size_t *, void *, size_t)); 382 int tcp_usrreq __P((struct socket *, 383 int, struct mbuf *, struct mbuf *, struct mbuf *, struct proc *)); 384 void tcp_xmit_timer __P((struct tcpcb *, int)); 385 tcp_seq tcp_new_iss __P((void *, u_long, tcp_seq)); 386 387 int syn_cache_add __P((struct socket *, struct mbuf *, u_char *, 388 int, struct tcp_opt_info *)); 389 void syn_cache_unreach __P((struct ip *, struct tcphdr *)); 390 struct socket * 391 syn_cache_get __P((struct socket *so, struct mbuf *)); 392 void syn_cache_insert __P((struct syn_cache *, struct syn_cache ***, 393 struct syn_cache_head **)); 394 struct syn_cache * 395 syn_cache_lookup __P((struct tcpiphdr *, struct syn_cache ***, 396 struct syn_cache_head **)); 397 void syn_cache_reset __P((struct tcpiphdr *)); 398 int syn_cache_respond __P((struct syn_cache *, struct mbuf *, 399 register struct tcpiphdr *, long, u_long)); 400 void syn_cache_timer __P((int)); 401 402 #endif 403
Indexes created Tue Sep 23 22:09:46 GMT 2025