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 * 3. All advertising materials mentioning features or use of this software 14 * must display the following acknowledgement: 15 * This product includes software developed by the University of 16 * California, Berkeley and its contributors. 17 * 4. Neither the name of the University nor the names of its contributors 18 * may be used to endorse or promote products derived from this software 19 * without specific prior written permission. 20 * 21 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 24 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 31 * SUCH DAMAGE. 32 * 33 * @(#)tcp_var.h 8.4 (Berkeley) 5/24/95 34 * $FreeBSD$ 35 */ 36 37 #ifndef _NETINET_TCP_VAR_H_ 38 #define _NETINET_TCP_VAR_H_ 39 /* 40 * Kernel variables for tcp. 41 */ 42 43 /* 44 * Tcp control block, one per tcp; fields: 45 * Organized for 16 byte cacheline efficiency. 46 */ 47 struct tcpcb { 48 struct mbuf *t_segq; 49 int t_dupacks; /* consecutive dup acks recd */ 50 struct tcpiphdr *t_template; /* skeletal packet for transmit */ 51 52 struct callout *tt_rexmt; /* retransmit timer */ 53 struct callout *tt_persist; /* retransmit persistence */ 54 struct callout *tt_keep; /* keepalive */ 55 struct callout *tt_2msl; /* 2*msl TIME_WAIT timer */ 56 struct callout *tt_delack; /* delayed ACK timer */ 57 58 struct inpcb *t_inpcb; /* back pointer to internet pcb */ 59 int t_state; /* state of this connection */ 60 u_int t_flags; 61 #define TF_ACKNOW 0x00001 /* ack peer immediately */ 62 #define TF_DELACK 0x00002 /* ack, but try to delay it */ 63 #define TF_NODELAY 0x00004 /* don't delay packets to coalesce */ 64 #define TF_NOOPT 0x00008 /* don't use tcp options */ 65 #define TF_SENTFIN 0x00010 /* have sent FIN */ 66 #define TF_REQ_SCALE 0x00020 /* have/will request window scaling */ 67 #define TF_RCVD_SCALE 0x00040 /* other side has requested scaling */ 68 #define TF_REQ_TSTMP 0x00080 /* have/will request timestamps */ 69 #define TF_RCVD_TSTMP 0x00100 /* a timestamp was received in SYN */ 70 #define TF_SACK_PERMIT 0x00200 /* other side said I could SACK */ 71 #define TF_NEEDSYN 0x00400 /* send SYN (implicit state) */ 72 #define TF_NEEDFIN 0x00800 /* send FIN (implicit state) */ 73 #define TF_NOPUSH 0x01000 /* don't push */ 74 #define TF_REQ_CC 0x02000 /* have/will request CC */ 75 #define TF_RCVD_CC 0x04000 /* a CC was received in SYN */ 76 #define TF_SENDCCNEW 0x08000 /* send CCnew instead of CC in SYN */ 77 #define TF_MORETOCOME 0x10000 /* More data to be appended to sock */ 78 int t_force; /* 1 if forcing out a byte */ 79 80 tcp_seq snd_una; /* send unacknowledged */ 81 tcp_seq snd_max; /* highest sequence number sent; 82 * used to recognize retransmits 83 */ 84 tcp_seq snd_nxt; /* send next */ 85 tcp_seq snd_up; /* send urgent pointer */ 86 87 tcp_seq snd_wl1; /* window update seg seq number */ 88 tcp_seq snd_wl2; /* window update seg ack number */ 89 tcp_seq iss; /* initial send sequence number */ 90 tcp_seq irs; /* initial receive sequence number */ 91 92 tcp_seq rcv_nxt; /* receive next */ 93 tcp_seq rcv_adv; /* advertised window */ 94 u_long rcv_wnd; /* receive window */ 95 tcp_seq rcv_up; /* receive urgent pointer */ 96 97 u_long snd_wnd; /* send window */ 98 u_long snd_cwnd; /* congestion-controlled window */ 99 u_long snd_ssthresh; /* snd_cwnd size threshold for 100 * for slow start exponential to 101 * linear switch 102 */ 103 u_int t_maxopd; /* mss plus options */ 104 105 u_long t_rcvtime; /* inactivity time */ 106 u_long t_starttime; /* time connection was established */ 107 int t_rtttime; /* round trip time */ 108 tcp_seq t_rtseq; /* sequence number being timed */ 109 110 int t_rxtcur; /* current retransmit value (ticks) */ 111 u_int t_maxseg; /* maximum segment size */ 112 int t_srtt; /* smoothed round-trip time */ 113 int t_rttvar; /* variance in round-trip time */ 114 115 int t_rxtshift; /* log(2) of rexmt exp. backoff */ 116 u_int t_rttmin; /* minimum rtt allowed */ 117 u_long t_rttupdated; /* number of times rtt sampled */ 118 u_long max_sndwnd; /* largest window peer has offered */ 119 120 int t_softerror; /* possible error not yet reported */ 121 /* out-of-band data */ 122 char t_oobflags; /* have some */ 123 char t_iobc; /* input character */ 124 #define TCPOOB_HAVEDATA 0x01 125 #define TCPOOB_HADDATA 0x02 126 /* RFC 1323 variables */ 127 u_char snd_scale; /* window scaling for send window */ 128 u_char rcv_scale; /* window scaling for recv window */ 129 u_char request_r_scale; /* pending window scaling */ 130 u_char requested_s_scale; 131 u_long ts_recent; /* timestamp echo data */ 132 133 u_long ts_recent_age; /* when last updated */ 134 tcp_seq last_ack_sent; 135 /* RFC 1644 variables */ 136 tcp_cc cc_send; /* send connection count */ 137 tcp_cc cc_recv; /* receive connection count */ 138 /* experimental */ 139 u_long snd_cwnd_prev; /* cwnd prior to retransmit */ 140 u_long snd_ssthresh_prev; /* ssthresh prior to retransmit */ 141 u_long t_badrxtwin; /* window for retransmit recovery */ 142 }; 143 144 /* 145 * Structure to hold TCP options that are only used during segment 146 * processing (in tcp_input), but not held in the tcpcb. 147 * It's basically used to reduce the number of parameters 148 * to tcp_dooptions. 149 */ 150 struct tcpopt { 151 u_long to_flag; /* which options are present */ 152 #define TOF_TS 0x0001 /* timestamp */ 153 #define TOF_CC 0x0002 /* CC and CCnew are exclusive */ 154 #define TOF_CCNEW 0x0004 155 #define TOF_CCECHO 0x0008 156 u_long to_tsval; 157 u_long to_tsecr; 158 tcp_cc to_cc; /* holds CC or CCnew */ 159 tcp_cc to_ccecho; 160 }; 161 162 /* 163 * The TAO cache entry which is stored in the protocol family specific 164 * portion of the route metrics. 165 */ 166 struct rmxp_tao { 167 tcp_cc tao_cc; /* latest CC in valid SYN */ 168 tcp_cc tao_ccsent; /* latest CC sent to peer */ 169 u_short tao_mssopt; /* peer's cached MSS */ 170 #ifdef notyet 171 u_short tao_flags; /* cache status flags */ 172 #define TAOF_DONT 0x0001 /* peer doesn't understand rfc1644 */ 173 #define TAOF_OK 0x0002 /* peer does understand rfc1644 */ 174 #define TAOF_UNDEF 0 /* we don't know yet */ 175 #endif /* notyet */ 176 }; 177 #define rmx_taop(r) ((struct rmxp_tao *)(r).rmx_filler) 178 179 #define intotcpcb(ip) ((struct tcpcb *)(ip)->inp_ppcb) 180 #define sototcpcb(so) (intotcpcb(sotoinpcb(so))) 181 182 /* 183 * The smoothed round-trip time and estimated variance 184 * are stored as fixed point numbers scaled by the values below. 185 * For convenience, these scales are also used in smoothing the average 186 * (smoothed = (1/scale)sample + ((scale-1)/scale)smoothed). 187 * With these scales, srtt has 3 bits to the right of the binary point, 188 * and thus an "ALPHA" of 0.875. rttvar has 2 bits to the right of the 189 * binary point, and is smoothed with an ALPHA of 0.75. 190 */ 191 #define TCP_RTT_SCALE 32 /* multiplier for srtt; 3 bits frac. */ 192 #define TCP_RTT_SHIFT 5 /* shift for srtt; 3 bits frac. */ 193 #define TCP_RTTVAR_SCALE 16 /* multiplier for rttvar; 2 bits */ 194 #define TCP_RTTVAR_SHIFT 4 /* shift for rttvar; 2 bits */ 195 #define TCP_DELTA_SHIFT 2 /* see tcp_input.c */ 196 197 /* 198 * The initial retransmission should happen at rtt + 4 * rttvar. 199 * Because of the way we do the smoothing, srtt and rttvar 200 * will each average +1/2 tick of bias. When we compute 201 * the retransmit timer, we want 1/2 tick of rounding and 202 * 1 extra tick because of +-1/2 tick uncertainty in the 203 * firing of the timer. The bias will give us exactly the 204 * 1.5 tick we need. But, because the bias is 205 * statistical, we have to test that we don't drop below 206 * the minimum feasible timer (which is 2 ticks). 207 * This version of the macro adapted from a paper by Lawrence 208 * Brakmo and Larry Peterson which outlines a problem caused 209 * by insufficient precision in the original implementation, 210 * which results in inappropriately large RTO values for very 211 * fast networks. 212 */ 213 #define TCP_REXMTVAL(tp) \ 214 max((tp)->t_rttmin, (((tp)->t_srtt >> (TCP_RTT_SHIFT - TCP_DELTA_SHIFT)) \ 215 + (tp)->t_rttvar) >> TCP_DELTA_SHIFT) 216 217 /* 218 * TCP statistics. 219 * Many of these should be kept per connection, 220 * but that's inconvenient at the moment. 221 */ 222 struct tcpstat { 223 u_long tcps_connattempt; /* connections initiated */ 224 u_long tcps_accepts; /* connections accepted */ 225 u_long tcps_connects; /* connections established */ 226 u_long tcps_drops; /* connections dropped */ 227 u_long tcps_conndrops; /* embryonic connections dropped */ 228 u_long tcps_closed; /* conn. closed (includes drops) */ 229 u_long tcps_segstimed; /* segs where we tried to get rtt */ 230 u_long tcps_rttupdated; /* times we succeeded */ 231 u_long tcps_delack; /* delayed acks sent */ 232 u_long tcps_timeoutdrop; /* conn. dropped in rxmt timeout */ 233 u_long tcps_rexmttimeo; /* retransmit timeouts */ 234 u_long tcps_persisttimeo; /* persist timeouts */ 235 u_long tcps_keeptimeo; /* keepalive timeouts */ 236 u_long tcps_keepprobe; /* keepalive probes sent */ 237 u_long tcps_keepdrops; /* connections dropped in keepalive */ 238 239 u_long tcps_sndtotal; /* total packets sent */ 240 u_long tcps_sndpack; /* data packets sent */ 241 u_long tcps_sndbyte; /* data bytes sent */ 242 u_long tcps_sndrexmitpack; /* data packets retransmitted */ 243 u_long tcps_sndrexmitbyte; /* data bytes retransmitted */ 244 u_long tcps_sndacks; /* ack-only packets sent */ 245 u_long tcps_sndprobe; /* window probes sent */ 246 u_long tcps_sndurg; /* packets sent with URG only */ 247 u_long tcps_sndwinup; /* window update-only packets sent */ 248 u_long tcps_sndctrl; /* control (SYN|FIN|RST) packets sent */ 249 250 u_long tcps_rcvtotal; /* total packets received */ 251 u_long tcps_rcvpack; /* packets received in sequence */ 252 u_long tcps_rcvbyte; /* bytes received in sequence */ 253 u_long tcps_rcvbadsum; /* packets received with ccksum errs */ 254 u_long tcps_rcvbadoff; /* packets received with bad offset */ 255 u_long tcps_rcvshort; /* packets received too short */ 256 u_long tcps_rcvduppack; /* duplicate-only packets received */ 257 u_long tcps_rcvdupbyte; /* duplicate-only bytes received */ 258 u_long tcps_rcvpartduppack; /* packets with some duplicate data */ 259 u_long tcps_rcvpartdupbyte; /* dup. bytes in part-dup. packets */ 260 u_long tcps_rcvoopack; /* out-of-order packets received */ 261 u_long tcps_rcvoobyte; /* out-of-order bytes received */ 262 u_long tcps_rcvpackafterwin; /* packets with data after window */ 263 u_long tcps_rcvbyteafterwin; /* bytes rcvd after window */ 264 u_long tcps_rcvafterclose; /* packets rcvd after "close" */ 265 u_long tcps_rcvwinprobe; /* rcvd window probe packets */ 266 u_long tcps_rcvdupack; /* rcvd duplicate acks */ 267 u_long tcps_rcvacktoomuch; /* rcvd acks for unsent data */ 268 u_long tcps_rcvackpack; /* rcvd ack packets */ 269 u_long tcps_rcvackbyte; /* bytes acked by rcvd acks */ 270 u_long tcps_rcvwinupd; /* rcvd window update packets */ 271 u_long tcps_pawsdrop; /* segments dropped due to PAWS */ 272 u_long tcps_predack; /* times hdr predict ok for acks */ 273 u_long tcps_preddat; /* times hdr predict ok for data pkts */ 274 u_long tcps_pcbcachemiss; 275 u_long tcps_cachedrtt; /* times cached RTT in route updated */ 276 u_long tcps_cachedrttvar; /* times cached rttvar updated */ 277 u_long tcps_cachedssthresh; /* times cached ssthresh updated */ 278 u_long tcps_usedrtt; /* times RTT initialized from route */ 279 u_long tcps_usedrttvar; /* times RTTVAR initialized from rt */ 280 u_long tcps_usedssthresh; /* times ssthresh initialized from rt*/ 281 u_long tcps_persistdrop; /* timeout in persist state */ 282 u_long tcps_badsyn; /* bogus SYN, e.g. premature ACK */ 283 u_long tcps_mturesent; /* resends due to MTU discovery */ 284 u_long tcps_listendrop; /* listen queue overflows */ 285 }; 286 287 /* 288 * TCB structure exported to user-land via sysctl(3). 289 * Evil hack: declare only if in_pcb.h and sys/socketvar.h have been 290 * included. Not all of our clients do. 291 */ 292 #if defined(_NETINET_IN_PCB_H_) && defined(_SYS_SOCKETVAR_H_) 293 struct xtcpcb { 294 size_t xt_len; 295 struct inpcb xt_inp; 296 struct tcpcb xt_tp; 297 struct xsocket xt_socket; 298 u_quad_t xt_alignment_hack; 299 }; 300 #endif 301 302 /* 303 * Names for TCP sysctl objects 304 */ 305 #define TCPCTL_DO_RFC1323 1 /* use RFC-1323 extensions */ 306 #define TCPCTL_DO_RFC1644 2 /* use RFC-1644 extensions */ 307 #define TCPCTL_MSSDFLT 3 /* MSS default */ 308 #define TCPCTL_STATS 4 /* statistics (read-only) */ 309 #define TCPCTL_RTTDFLT 5 /* default RTT estimate */ 310 #define TCPCTL_KEEPIDLE 6 /* keepalive idle timer */ 311 #define TCPCTL_KEEPINTVL 7 /* interval to send keepalives */ 312 #define TCPCTL_SENDSPACE 8 /* send buffer space */ 313 #define TCPCTL_RECVSPACE 9 /* receive buffer space */ 314 #define TCPCTL_KEEPINIT 10 /* receive buffer space */ 315 #define TCPCTL_PCBLIST 11 /* list of all outstanding PCBs */ 316 #define TCPCTL_DELACKTIME 12 /* time before sending delayed ACK */ 317 #define TCPCTL_MAXID 13 318 319 #define TCPCTL_NAMES { \ 320 { 0, 0 }, \ 321 { "rfc1323", CTLTYPE_INT }, \ 322 { "rfc1644", CTLTYPE_INT }, \ 323 { "mssdflt", CTLTYPE_INT }, \ 324 { "stats", CTLTYPE_STRUCT }, \ 325 { "rttdflt", CTLTYPE_INT }, \ 326 { "keepidle", CTLTYPE_INT }, \ 327 { "keepintvl", CTLTYPE_INT }, \ 328 { "sendspace", CTLTYPE_INT }, \ 329 { "recvspace", CTLTYPE_INT }, \ 330 { "keepinit", CTLTYPE_INT }, \ 331 { "pcblist", CTLTYPE_STRUCT }, \ 332 { "delacktime", CTLTYPE_INT }, \ 333 } 334 335 336 #ifdef KERNEL 337 #ifdef SYSCTL_DECL 338 SYSCTL_DECL(_net_inet_tcp); 339 #endif 340 341 extern struct inpcbhead tcb; /* head of queue of active tcpcb's */ 342 extern struct inpcbinfo tcbinfo; 343 extern struct tcpstat tcpstat; /* tcp statistics */ 344 extern int tcp_mssdflt; /* XXX */ 345 extern int tcp_delack_enabled; 346 extern int ss_fltsz; 347 extern int ss_fltsz_local; 348 349 void tcp_canceltimers __P((struct tcpcb *)); 350 struct tcpcb * 351 tcp_close __P((struct tcpcb *)); 352 void tcp_ctlinput __P((int, struct sockaddr *, void *)); 353 int tcp_ctloutput __P((struct socket *, struct sockopt *)); 354 struct tcpcb * 355 tcp_drop __P((struct tcpcb *, int)); 356 void tcp_drain __P((void)); 357 void tcp_fasttimo __P((void)); 358 struct rmxp_tao * 359 tcp_gettaocache __P((struct inpcb *)); 360 void tcp_init __P((void)); 361 void tcp_input __P((struct mbuf *, int)); 362 void tcp_mss __P((struct tcpcb *, int)); 363 int tcp_mssopt __P((struct tcpcb *)); 364 void tcp_mtudisc __P((struct inpcb *, int)); 365 struct tcpcb * 366 tcp_newtcpcb __P((struct inpcb *)); 367 int tcp_output __P((struct tcpcb *)); 368 void tcp_quench __P((struct inpcb *, int)); 369 void tcp_respond __P((struct tcpcb *, 370 struct tcpiphdr *, struct mbuf *, tcp_seq, tcp_seq, int)); 371 struct rtentry * 372 tcp_rtlookup __P((struct inpcb *)); 373 void tcp_setpersist __P((struct tcpcb *)); 374 void tcp_slowtimo __P((void)); 375 struct tcpiphdr * 376 tcp_template __P((struct tcpcb *)); 377 struct tcpcb * 378 tcp_timers __P((struct tcpcb *, int)); 379 void tcp_trace __P((int, int, struct tcpcb *, struct tcpiphdr *, int)); 380 381 extern struct pr_usrreqs tcp_usrreqs; 382 extern u_long tcp_sendspace; 383 extern u_long tcp_recvspace; 384 385 #endif /* KERNEL */ 386 387 #endif /* _NETINET_TCP_VAR_H_ */ 388