1 /* $NetBSD: tcp_output.c,v 1.52 1999/09/23 02:21:31 itojun Exp $ */ 2 3 /* 4 %%% portions-copyright-nrl-95 5 Portions of this software are Copyright 1995-1998 by Randall Atkinson, 6 Ronald Lee, Daniel McDonald, Bao Phan, and Chris Winters. All Rights 7 Reserved. All rights under this copyright have been assigned to the US 8 Naval Research Laboratory (NRL). The NRL Copyright Notice and License 9 Agreement Version 1.1 (January 17, 1995) applies to these portions of the 10 software. 11 You should have received a copy of the license with this software. If you 12 didn't get a copy, you may request one from <license@ipv6.nrl.navy.mil>. 13 14 */ 15 16 /* 17 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. 18 * All rights reserved. 19 * 20 * Redistribution and use in source and binary forms, with or without 21 * modification, are permitted provided that the following conditions 22 * are met: 23 * 1. Redistributions of source code must retain the above copyright 24 * notice, this list of conditions and the following disclaimer. 25 * 2. Redistributions in binary form must reproduce the above copyright 26 * notice, this list of conditions and the following disclaimer in the 27 * documentation and/or other materials provided with the distribution. 28 * 3. Neither the name of the project nor the names of its contributors 29 * may be used to endorse or promote products derived from this software 30 * without specific prior written permission. 31 * 32 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND 33 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 34 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 35 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE 36 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 37 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 38 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 39 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 40 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 41 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 42 * SUCH DAMAGE. 43 */ 44 45 /*- 46 * Copyright (c) 1997, 1998 The NetBSD Foundation, Inc. 47 * All rights reserved. 48 * 49 * This code is derived from software contributed to The NetBSD Foundation 50 * by Jason R. Thorpe and Kevin M. Lahey of the Numerical Aerospace Simulation 51 * Facility, NASA Ames Research Center. 52 * 53 * Redistribution and use in source and binary forms, with or without 54 * modification, are permitted provided that the following conditions 55 * are met: 56 * 1. Redistributions of source code must retain the above copyright 57 * notice, this list of conditions and the following disclaimer. 58 * 2. Redistributions in binary form must reproduce the above copyright 59 * notice, this list of conditions and the following disclaimer in the 60 * documentation and/or other materials provided with the distribution. 61 * 3. All advertising materials mentioning features or use of this software 62 * must display the following acknowledgement: 63 * This product includes software developed by the NetBSD 64 * Foundation, Inc. and its contributors. 65 * 4. Neither the name of The NetBSD Foundation nor the names of its 66 * contributors may be used to endorse or promote products derived 67 * from this software without specific prior written permission. 68 * 69 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 70 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 71 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 72 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 73 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 74 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 75 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 76 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 77 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 78 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 79 * POSSIBILITY OF SUCH DAMAGE. 80 */ 81 82 /* 83 * Copyright (c) 1982, 1986, 1988, 1990, 1993, 1995 84 * The Regents of the University of California. All rights reserved. 85 * 86 * Redistribution and use in source and binary forms, with or without 87 * modification, are permitted provided that the following conditions 88 * are met: 89 * 1. Redistributions of source code must retain the above copyright 90 * notice, this list of conditions and the following disclaimer. 91 * 2. Redistributions in binary form must reproduce the above copyright 92 * notice, this list of conditions and the following disclaimer in the 93 * documentation and/or other materials provided with the distribution. 94 * 3. All advertising materials mentioning features or use of this software 95 * must display the following acknowledgement: 96 * This product includes software developed by the University of 97 * California, Berkeley and its contributors. 98 * 4. Neither the name of the University nor the names of its contributors 99 * may be used to endorse or promote products derived from this software 100 * without specific prior written permission. 101 * 102 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 103 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 104 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 105 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 106 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 107 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 108 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 109 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 110 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 111 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 112 * SUCH DAMAGE. 113 * 114 * @(#)tcp_output.c 8.4 (Berkeley) 5/24/95 115 */ 116 117 #include "opt_inet.h" 118 #include "opt_ipsec.h" 119 120 #include <sys/param.h> 121 #include <sys/systm.h> 122 #include <sys/malloc.h> 123 #include <sys/mbuf.h> 124 #include <sys/protosw.h> 125 #include <sys/socket.h> 126 #include <sys/socketvar.h> 127 #include <sys/errno.h> 128 #include <sys/domain.h> 129 130 #include <net/if.h> 131 #include <net/route.h> 132 133 #include <netinet/in.h> 134 #include <netinet/in_systm.h> 135 #include <netinet/ip.h> 136 #include <netinet/in_pcb.h> 137 #include <netinet/ip_var.h> 138 139 #ifdef INET6 140 #ifndef INET 141 #include <netinet/in.h> 142 #endif 143 #include <netinet/ip6.h> 144 #include <netinet6/in6_pcb.h> 145 #include <netinet6/ip6_var.h> 146 #endif 147 148 #include <netinet/tcp.h> 149 #define TCPOUTFLAGS 150 #include <netinet/tcp_fsm.h> 151 #include <netinet/tcp_seq.h> 152 #include <netinet/tcp_timer.h> 153 #include <netinet/tcp_var.h> 154 #include <netinet/tcpip.h> 155 #include <netinet/tcp_debug.h> 156 157 #ifdef notyet 158 extern struct mbuf *m_copypack(); 159 #endif 160 161 #define MAX_TCPOPTLEN 32 /* max # bytes that go in options */ 162 163 /* 164 * Knob to enable Congestion Window Monitoring, and control the 165 * the burst size it allows. Default burst is 4 packets, per 166 * the Internet draft. 167 */ 168 int tcp_cwm = 0; 169 int tcp_cwm_burstsize = 4; 170 171 static __inline void tcp_segsize __P((struct tcpcb *, int *, int *)); 172 static __inline void 173 tcp_segsize(tp, txsegsizep, rxsegsizep) 174 struct tcpcb *tp; 175 int *txsegsizep, *rxsegsizep; 176 { 177 struct inpcb *inp = tp->t_inpcb; 178 #ifdef INET6 179 struct in6pcb *in6p = tp->t_in6pcb; 180 #endif 181 struct rtentry *rt; 182 struct ifnet *ifp; 183 int size; 184 int iphlen; 185 186 switch (tp->t_family) { 187 case AF_INET: 188 iphlen = sizeof(struct ip); 189 break; 190 #ifdef INET6 191 case AF_INET6: 192 iphlen = sizeof(struct ip6_hdr); 193 break; 194 #endif 195 default: 196 size = tcp_mssdflt; 197 goto out; 198 } 199 200 if (inp) 201 rt = in_pcbrtentry(inp); 202 #if defined(INET6) && !defined(TCP6) 203 else if (in6p) 204 rt = in6_pcbrtentry(in6p); 205 #endif 206 else 207 rt = NULL; 208 if (rt == NULL) { 209 size = tcp_mssdflt; 210 goto out; 211 } 212 213 ifp = rt->rt_ifp; 214 215 size = tcp_mssdflt; 216 if (rt->rt_rmx.rmx_mtu != 0) 217 size = rt->rt_rmx.rmx_mtu - iphlen - sizeof(struct tcphdr); 218 else if (ip_mtudisc || ifp->if_flags & IFF_LOOPBACK) 219 size = ifp->if_mtu - iphlen - sizeof(struct tcphdr); 220 else if (inp && in_localaddr(inp->inp_faddr)) 221 size = ifp->if_mtu - iphlen - sizeof(struct tcphdr); 222 #ifdef INET6 223 else if (in6p) { 224 if (IN6_IS_ADDR_V4MAPPED(&in6p->in6p_faddr)) { 225 /* mapped addr case */ 226 struct in_addr d; 227 bcopy(&in6p->in6p_faddr.s6_addr32[3], &d, sizeof(d)); 228 if (in_localaddr(d)) 229 size = ifp->if_mtu - iphlen - sizeof(struct tcphdr); 230 } else { 231 if (in6_localaddr(&in6p->in6p_faddr)) 232 size = ifp->if_mtu - iphlen - sizeof(struct tcphdr); 233 } 234 } 235 #endif 236 size -= tcp_optlen(tp); 237 /* 238 * XXX tp->t_ourmss should have the right size, but without this code 239 * fragmentation will occur... need more investigation 240 */ 241 if (inp) { 242 #ifdef IPSEC 243 size -= ipsec4_hdrsiz_tcp(tp); 244 #endif 245 size -= ip_optlen(inp); 246 } 247 #ifdef INET6 248 else if (in6p && tp->t_family == AF_INET) { 249 #ifdef IPSEC 250 size -= ipsec4_hdrsiz_tcp(tp); 251 #endif 252 /* XXX size -= ip_optlen(in6p); */ 253 } 254 else if (in6p && tp->t_family == AF_INET6) { 255 #if defined(IPSEC) && !defined(TCP6) 256 size -= ipsec6_hdrsiz_tcp(tp); 257 #endif 258 size -= ip6_optlen(in6p); 259 } 260 #endif 261 262 out: 263 /* 264 * *rxsegsizep holds *estimated* inbound segment size (estimation 265 * assumes that path MTU is the same for both ways). this is only 266 * for silly window avoidance, do not use the value for other purposes. 267 * 268 * ipseclen is subtracted from both sides, this may not be right. 269 * I'm not quite sure about this (could someone comment). 270 */ 271 *txsegsizep = min(tp->t_peermss, size); 272 *rxsegsizep = min(tp->t_ourmss, size); 273 274 if (*txsegsizep != tp->t_segsz) { 275 /* 276 * If the new segment size is larger, we don't want to 277 * mess up the congestion window, but if it is smaller 278 * we'll have to reduce the congestion window to ensure 279 * that we don't get into trouble with initial windows 280 * and the rest. In any case, if the segment size 281 * has changed, chances are the path has, too, and 282 * our congestion window will be different. 283 */ 284 if (*txsegsizep < tp->t_segsz) { 285 tp->snd_cwnd = max((tp->snd_cwnd / tp->t_segsz) 286 * *txsegsizep, *txsegsizep); 287 tp->snd_ssthresh = max((tp->snd_ssthresh / tp->t_segsz) 288 * *txsegsizep, *txsegsizep); 289 } 290 tp->t_segsz = *txsegsizep; 291 } 292 } 293 294 /* 295 * Tcp output routine: figure out what should be sent and send it. 296 */ 297 int 298 tcp_output(tp) 299 register struct tcpcb *tp; 300 { 301 struct socket *so; 302 struct route *ro; 303 struct rtentry *rt; 304 long len, win; 305 int off, flags, error; 306 register struct mbuf *m; 307 struct ip *ip; 308 #ifdef INET6 309 struct ip6_hdr *ip6; 310 #endif 311 register struct tcphdr *th; 312 u_char opt[MAX_TCPOPTLEN]; 313 unsigned optlen, hdrlen; 314 int idle, sendalot, txsegsize, rxsegsize; 315 int maxburst = TCP_MAXBURST; 316 int af; /* address family on the wire */ 317 int iphdrlen; 318 319 so = NULL; 320 ro = NULL; 321 if (tp->t_inpcb) { 322 so = tp->t_inpcb->inp_socket; 323 ro = &tp->t_inpcb->inp_route; 324 } 325 #ifdef INET6 326 else if (tp->t_in6pcb) { 327 so = tp->t_in6pcb->in6p_socket; 328 ro = (struct route *)&tp->t_in6pcb->in6p_route; 329 } 330 #endif 331 332 switch (af = tp->t_family) { 333 case AF_INET: 334 if (tp->t_inpcb) 335 break; 336 #ifdef INET6 337 /* mapped addr case */ 338 if (tp->t_in6pcb) 339 break; 340 #endif 341 return EINVAL; 342 #ifdef INET6 343 case AF_INET6: 344 if (tp->t_in6pcb) 345 break; 346 return EINVAL; 347 #endif 348 default: 349 return EAFNOSUPPORT; 350 } 351 352 tcp_segsize(tp, &txsegsize, &rxsegsize); 353 354 idle = (tp->snd_max == tp->snd_una); 355 356 /* 357 * Restart Window computation. From draft-floyd-incr-init-win-03: 358 * 359 * Optionally, a TCP MAY set the restart window to the 360 * minimum of the value used for the initial window and 361 * the current value of cwnd (in other words, using a 362 * larger value for the restart window should never increase 363 * the size of cwnd). 364 */ 365 if (tcp_cwm) { 366 /* 367 * Hughes/Touch/Heidemann Congestion Window Monitoring. 368 * Count the number of packets currently pending 369 * acknowledgement, and limit our congestion window 370 * to a pre-determined allowed burst size plus that count. 371 * This prevents bursting once all pending packets have 372 * been acknowledged (i.e. transmission is idle). 373 * 374 * XXX Link this to Initial Window? 375 */ 376 tp->snd_cwnd = min(tp->snd_cwnd, 377 (tcp_cwm_burstsize * txsegsize) + 378 (tp->snd_nxt - tp->snd_una)); 379 } else { 380 if (idle && tp->t_idle >= tp->t_rxtcur) { 381 /* 382 * We have been idle for "a while" and no acks are 383 * expected to clock out any data we send -- 384 * slow start to get ack "clock" running again. 385 */ 386 tp->snd_cwnd = min(tp->snd_cwnd, 387 TCP_INITIAL_WINDOW(tcp_init_win, txsegsize)); 388 } 389 } 390 391 again: 392 /* 393 * Determine length of data that should be transmitted, and 394 * flags that should be used. If there is some data or critical 395 * controls (SYN, RST) to send, then transmit; otherwise, 396 * investigate further. 397 */ 398 sendalot = 0; 399 off = tp->snd_nxt - tp->snd_una; 400 win = min(tp->snd_wnd, tp->snd_cwnd); 401 402 flags = tcp_outflags[tp->t_state]; 403 /* 404 * If in persist timeout with window of 0, send 1 byte. 405 * Otherwise, if window is small but nonzero 406 * and timer expired, we will send what we can 407 * and go to transmit state. 408 */ 409 if (tp->t_force) { 410 if (win == 0) { 411 /* 412 * If we still have some data to send, then 413 * clear the FIN bit. Usually this would 414 * happen below when it realizes that we 415 * aren't sending all the data. However, 416 * if we have exactly 1 byte of unset data, 417 * then it won't clear the FIN bit below, 418 * and if we are in persist state, we wind 419 * up sending the packet without recording 420 * that we sent the FIN bit. 421 * 422 * We can't just blindly clear the FIN bit, 423 * because if we don't have any more data 424 * to send then the probe will be the FIN 425 * itself. 426 */ 427 if (off < so->so_snd.sb_cc) 428 flags &= ~TH_FIN; 429 win = 1; 430 } else { 431 TCP_TIMER_DISARM(tp, TCPT_PERSIST); 432 tp->t_rxtshift = 0; 433 } 434 } 435 436 if (win < so->so_snd.sb_cc) { 437 len = win - off; 438 flags &= ~TH_FIN; 439 } else 440 len = so->so_snd.sb_cc - off; 441 442 if (len < 0) { 443 /* 444 * If FIN has been sent but not acked, 445 * but we haven't been called to retransmit, 446 * len will be -1. Otherwise, window shrank 447 * after we sent into it. If window shrank to 0, 448 * cancel pending retransmit, pull snd_nxt back 449 * to (closed) window, and set the persist timer 450 * if it isn't already going. If the window didn't 451 * close completely, just wait for an ACK. 452 * 453 * If we have a pending FIN, either it has already been 454 * transmitted or it is outside the window, so drop it. 455 * If the FIN has been transmitted, but this is not a 456 * retransmission, then len must be -1. Therefore we also 457 * prevent here the sending of `gratuitous FINs'. This 458 * eliminates the need to check for that case below (e.g. 459 * to back up snd_nxt before the FIN so that the sequence 460 * number is correct). 461 */ 462 len = 0; 463 flags &= ~TH_FIN; 464 if (win == 0) { 465 TCP_TIMER_DISARM(tp, TCPT_REXMT); 466 tp->t_rxtshift = 0; 467 tp->snd_nxt = tp->snd_una; 468 if (TCP_TIMER_ISARMED(tp, TCPT_PERSIST) == 0) 469 tcp_setpersist(tp); 470 } 471 } 472 if (len > txsegsize) { 473 len = txsegsize; 474 flags &= ~TH_FIN; 475 sendalot = 1; 476 } 477 478 win = sbspace(&so->so_rcv); 479 480 /* 481 * Sender silly window avoidance. If connection is idle 482 * and can send all data, a maximum segment, 483 * at least a maximum default-size segment do it, 484 * or are forced, do it; otherwise don't bother. 485 * If peer's buffer is tiny, then send 486 * when window is at least half open. 487 * If retransmitting (possibly after persist timer forced us 488 * to send into a small window), then must resend. 489 */ 490 if (len) { 491 if (len == txsegsize) 492 goto send; 493 if ((so->so_state & SS_MORETOCOME) == 0 && 494 ((idle || tp->t_flags & TF_NODELAY) && 495 len + off >= so->so_snd.sb_cc)) 496 goto send; 497 if (tp->t_force) 498 goto send; 499 if (len >= tp->max_sndwnd / 2) 500 goto send; 501 if (SEQ_LT(tp->snd_nxt, tp->snd_max)) 502 goto send; 503 } 504 505 /* 506 * Compare available window to amount of window known to peer 507 * (as advertised window less next expected input). If the 508 * difference is at least twice the size of the largest segment 509 * we expect to receive (i.e. two segments) or at least 50% of 510 * the maximum possible window, then want to send a window update 511 * to peer. 512 */ 513 if (win > 0) { 514 /* 515 * "adv" is the amount we can increase the window, 516 * taking into account that we are limited by 517 * TCP_MAXWIN << tp->rcv_scale. 518 */ 519 long adv = min(win, (long)TCP_MAXWIN << tp->rcv_scale) - 520 (tp->rcv_adv - tp->rcv_nxt); 521 522 if (adv >= (long) (2 * rxsegsize)) 523 goto send; 524 if (2 * adv >= (long) so->so_rcv.sb_hiwat) 525 goto send; 526 } 527 528 /* 529 * Send if we owe peer an ACK. 530 */ 531 if (tp->t_flags & TF_ACKNOW) 532 goto send; 533 if (flags & (TH_SYN|TH_FIN|TH_RST)) 534 goto send; 535 if (SEQ_GT(tp->snd_up, tp->snd_una)) 536 goto send; 537 538 /* 539 * TCP window updates are not reliable, rather a polling protocol 540 * using ``persist'' packets is used to insure receipt of window 541 * updates. The three ``states'' for the output side are: 542 * idle not doing retransmits or persists 543 * persisting to move a small or zero window 544 * (re)transmitting and thereby not persisting 545 * 546 * tp->t_timer[TCPT_PERSIST] 547 * is set when we are in persist state. 548 * tp->t_force 549 * is set when we are called to send a persist packet. 550 * tp->t_timer[TCPT_REXMT] 551 * is set when we are retransmitting 552 * The output side is idle when both timers are zero. 553 * 554 * If send window is too small, there is data to transmit, and no 555 * retransmit or persist is pending, then go to persist state. 556 * If nothing happens soon, send when timer expires: 557 * if window is nonzero, transmit what we can, 558 * otherwise force out a byte. 559 */ 560 if (so->so_snd.sb_cc && TCP_TIMER_ISARMED(tp, TCPT_REXMT) == 0 && 561 TCP_TIMER_ISARMED(tp, TCPT_PERSIST) == 0) { 562 tp->t_rxtshift = 0; 563 tcp_setpersist(tp); 564 } 565 566 /* 567 * No reason to send a segment, just return. 568 */ 569 return (0); 570 571 send: 572 /* 573 * Before ESTABLISHED, force sending of initial options 574 * unless TCP set not to do any options. 575 * NOTE: we assume that the IP/TCP header plus TCP options 576 * always fit in a single mbuf, leaving room for a maximum 577 * link header, i.e. 578 * max_linkhdr + sizeof (struct tcpiphdr) + optlen <= MCLBYTES 579 */ 580 optlen = 0; 581 switch (af) { 582 case AF_INET: 583 iphdrlen = sizeof(struct ip) + sizeof(struct tcphdr); 584 break; 585 #ifdef INET6 586 case AF_INET6: 587 iphdrlen = sizeof(struct ip6_hdr) + sizeof(struct tcphdr); 588 break; 589 #endif 590 default: /*pacify gcc*/ 591 iphdrlen = 0; 592 break; 593 } 594 hdrlen = iphdrlen; 595 if (flags & TH_SYN) { 596 struct rtentry *rt; 597 598 if (tp->t_inpcb) 599 rt = in_pcbrtentry(tp->t_inpcb); 600 #if defined(INET6) && !defined(TCP6) 601 else if (tp->t_in6pcb) 602 rt = in6_pcbrtentry(tp->t_in6pcb); 603 #endif 604 else 605 rt = NULL; 606 607 tp->snd_nxt = tp->iss; 608 tp->t_ourmss = tcp_mss_to_advertise(rt != NULL ? 609 rt->rt_ifp : NULL, af); 610 if ((tp->t_flags & TF_NOOPT) == 0) { 611 opt[0] = TCPOPT_MAXSEG; 612 opt[1] = 4; 613 opt[2] = (tp->t_ourmss >> 8) & 0xff; 614 opt[3] = tp->t_ourmss & 0xff; 615 optlen = 4; 616 617 if ((tp->t_flags & TF_REQ_SCALE) && 618 ((flags & TH_ACK) == 0 || 619 (tp->t_flags & TF_RCVD_SCALE))) { 620 *((u_int32_t *) (opt + optlen)) = htonl( 621 TCPOPT_NOP << 24 | 622 TCPOPT_WINDOW << 16 | 623 TCPOLEN_WINDOW << 8 | 624 tp->request_r_scale); 625 optlen += 4; 626 } 627 } 628 } 629 630 /* 631 * Send a timestamp and echo-reply if this is a SYN and our side 632 * wants to use timestamps (TF_REQ_TSTMP is set) or both our side 633 * and our peer have sent timestamps in our SYN's. 634 */ 635 if ((tp->t_flags & (TF_REQ_TSTMP|TF_NOOPT)) == TF_REQ_TSTMP && 636 (flags & TH_RST) == 0 && 637 ((flags & (TH_SYN|TH_ACK)) == TH_SYN || 638 (tp->t_flags & TF_RCVD_TSTMP))) { 639 u_int32_t *lp = (u_int32_t *)(opt + optlen); 640 641 /* Form timestamp option as shown in appendix A of RFC 1323. */ 642 *lp++ = htonl(TCPOPT_TSTAMP_HDR); 643 *lp++ = htonl(tcp_now); 644 *lp = htonl(tp->ts_recent); 645 optlen += TCPOLEN_TSTAMP_APPA; 646 } 647 648 hdrlen += optlen; 649 650 #ifdef DIAGNOSTIC 651 if (len > txsegsize) 652 panic("tcp data to be sent is larger than segment"); 653 if (max_linkhdr + hdrlen > MCLBYTES) 654 panic("tcphdr too big"); 655 #endif 656 657 /* 658 * Grab a header mbuf, attaching a copy of data to 659 * be transmitted, and initialize the header from 660 * the template for sends on this connection. 661 */ 662 if (len) { 663 if (tp->t_force && len == 1) 664 tcpstat.tcps_sndprobe++; 665 else if (SEQ_LT(tp->snd_nxt, tp->snd_max)) { 666 tcpstat.tcps_sndrexmitpack++; 667 tcpstat.tcps_sndrexmitbyte += len; 668 } else { 669 tcpstat.tcps_sndpack++; 670 tcpstat.tcps_sndbyte += len; 671 } 672 #ifdef notyet 673 if ((m = m_copypack(so->so_snd.sb_mb, off, 674 (int)len, max_linkhdr + hdrlen)) == 0) { 675 error = ENOBUFS; 676 goto out; 677 } 678 /* 679 * m_copypack left space for our hdr; use it. 680 */ 681 m->m_len += hdrlen; 682 m->m_data -= hdrlen; 683 #else 684 MGETHDR(m, M_DONTWAIT, MT_HEADER); 685 if (m != NULL) { 686 MCLGET(m, M_DONTWAIT); 687 if ((m->m_flags & M_EXT) == 0) { 688 m_freem(m); 689 m = NULL; 690 } 691 } 692 if (m == NULL) { 693 error = ENOBUFS; 694 goto out; 695 } 696 m->m_data += max_linkhdr; 697 m->m_len = hdrlen; 698 if (len <= MCLBYTES - hdrlen - max_linkhdr) { 699 m_copydata(so->so_snd.sb_mb, off, (int) len, 700 mtod(m, caddr_t) + hdrlen); 701 m->m_len += len; 702 } else { 703 m->m_next = m_copy(so->so_snd.sb_mb, off, (int) len); 704 if (m->m_next == 0) { 705 m_freem(m); 706 error = ENOBUFS; 707 goto out; 708 } 709 } 710 #endif 711 /* 712 * If we're sending everything we've got, set PUSH. 713 * (This will keep happy those implementations which only 714 * give data to the user when a buffer fills or 715 * a PUSH comes in.) 716 */ 717 if (off + len == so->so_snd.sb_cc) 718 flags |= TH_PUSH; 719 } else { 720 if (tp->t_flags & TF_ACKNOW) 721 tcpstat.tcps_sndacks++; 722 else if (flags & (TH_SYN|TH_FIN|TH_RST)) 723 tcpstat.tcps_sndctrl++; 724 else if (SEQ_GT(tp->snd_up, tp->snd_una)) 725 tcpstat.tcps_sndurg++; 726 else 727 tcpstat.tcps_sndwinup++; 728 729 MGETHDR(m, M_DONTWAIT, MT_HEADER); 730 if (m != NULL) { 731 MCLGET(m, M_DONTWAIT); 732 if ((m->m_flags & M_EXT) == 0) { 733 m_freem(m); 734 m = NULL; 735 } 736 } 737 if (m == NULL) { 738 error = ENOBUFS; 739 goto out; 740 } 741 m->m_data += max_linkhdr; 742 m->m_len = hdrlen; 743 } 744 m->m_pkthdr.rcvif = (struct ifnet *)0; 745 switch (af) { 746 case AF_INET: 747 ip = mtod(m, struct ip *); 748 #ifdef INET6 749 ip6 = NULL; 750 #endif 751 th = (struct tcphdr *)(ip + 1); 752 break; 753 #ifdef INET6 754 case AF_INET6: 755 ip = NULL; 756 ip6 = mtod(m, struct ip6_hdr *); 757 th = (struct tcphdr *)(ip6 + 1); 758 break; 759 #endif 760 default: /*pacify gcc*/ 761 ip = NULL; 762 #ifdef INET6 763 ip6 = NULL; 764 #endif 765 th = NULL; 766 break; 767 } 768 if (tp->t_template == 0) 769 panic("tcp_output"); 770 if (tp->t_template->m_len < iphdrlen) 771 panic("tcp_output"); 772 bcopy(mtod(tp->t_template, caddr_t), mtod(m, caddr_t), iphdrlen); 773 774 /* 775 * If we are doing retransmissions, then snd_nxt will 776 * not reflect the first unsent octet. For ACK only 777 * packets, we do not want the sequence number of the 778 * retransmitted packet, we want the sequence number 779 * of the next unsent octet. So, if there is no data 780 * (and no SYN or FIN), use snd_max instead of snd_nxt 781 * when filling in ti_seq. But if we are in persist 782 * state, snd_max might reflect one byte beyond the 783 * right edge of the window, so use snd_nxt in that 784 * case, since we know we aren't doing a retransmission. 785 * (retransmit and persist are mutually exclusive...) 786 */ 787 if (len || (flags & (TH_SYN|TH_FIN)) || 788 TCP_TIMER_ISARMED(tp, TCPT_PERSIST)) 789 th->th_seq = htonl(tp->snd_nxt); 790 else 791 th->th_seq = htonl(tp->snd_max); 792 th->th_ack = htonl(tp->rcv_nxt); 793 if (optlen) { 794 bcopy((caddr_t)opt, (caddr_t)(th + 1), optlen); 795 th->th_off = (sizeof (struct tcphdr) + optlen) >> 2; 796 } 797 th->th_flags = flags; 798 /* 799 * Calculate receive window. Don't shrink window, 800 * but avoid silly window syndrome. 801 */ 802 if (win < (long)(so->so_rcv.sb_hiwat / 4) && win < (long)rxsegsize) 803 win = 0; 804 if (win > (long)TCP_MAXWIN << tp->rcv_scale) 805 win = (long)TCP_MAXWIN << tp->rcv_scale; 806 if (win < (long)(tp->rcv_adv - tp->rcv_nxt)) 807 win = (long)(tp->rcv_adv - tp->rcv_nxt); 808 th->th_win = htons((u_int16_t) (win>>tp->rcv_scale)); 809 if (SEQ_GT(tp->snd_up, tp->snd_nxt)) { 810 u_int32_t urp = tp->snd_up - tp->snd_nxt; 811 if (urp > IP_MAXPACKET) 812 urp = IP_MAXPACKET; 813 th->th_urp = htons((u_int16_t)urp); 814 th->th_flags |= TH_URG; 815 } else 816 /* 817 * If no urgent pointer to send, then we pull 818 * the urgent pointer to the left edge of the send window 819 * so that it doesn't drift into the send window on sequence 820 * number wraparound. 821 */ 822 tp->snd_up = tp->snd_una; /* drag it along */ 823 824 /* 825 * Put TCP length in extended header, and then 826 * checksum extended header and data. 827 */ 828 switch (af) { 829 case AF_INET: 830 { 831 struct ipovly *ipov = (struct ipovly *)ip; 832 if (len + optlen) 833 ipov->ih_len = htons((u_int16_t)(sizeof(struct tcphdr) + 834 optlen + len)); 835 bzero(ipov->ih_x1, sizeof(ipov->ih_x1)); 836 th->th_sum = 0; 837 th->th_sum = in_cksum(m, (int)(hdrlen + len)); 838 break; 839 } 840 #ifdef INET6 841 case AF_INET6: 842 /* equals to hdrlen + len */ 843 m->m_pkthdr.len = sizeof(struct ip6_hdr) 844 + sizeof(struct tcphdr) + optlen + len; 845 th->th_sum = 0; 846 th->th_sum = in6_cksum(m, IPPROTO_TCP, 847 sizeof(struct ip6_hdr), 848 sizeof(struct tcphdr) + optlen + len); 849 break; 850 #endif 851 } 852 853 /* 854 * In transmit state, time the transmission and arrange for 855 * the retransmit. In persist state, just set snd_max. 856 */ 857 if (tp->t_force == 0 || TCP_TIMER_ISARMED(tp, TCPT_PERSIST) == 0) { 858 tcp_seq startseq = tp->snd_nxt; 859 860 /* 861 * Advance snd_nxt over sequence space of this segment. 862 * There are no states in which we send both a SYN and a FIN, 863 * so we collapse the tests for these flags. 864 */ 865 if (flags & (TH_SYN|TH_FIN)) 866 tp->snd_nxt++; 867 tp->snd_nxt += len; 868 if (SEQ_GT(tp->snd_nxt, tp->snd_max)) { 869 tp->snd_max = tp->snd_nxt; 870 /* 871 * Time this transmission if not a retransmission and 872 * not currently timing anything. 873 */ 874 if (tp->t_rtt == 0) { 875 tp->t_rtt = 1; 876 tp->t_rtseq = startseq; 877 tcpstat.tcps_segstimed++; 878 } 879 } 880 881 /* 882 * Set retransmit timer if not currently set, 883 * and not doing an ack or a keep-alive probe. 884 * Initial value for retransmit timer is smoothed 885 * round-trip time + 2 * round-trip time variance. 886 * Initialize shift counter which is used for backoff 887 * of retransmit time. 888 */ 889 if (TCP_TIMER_ISARMED(tp, TCPT_REXMT) == 0 && 890 tp->snd_nxt != tp->snd_una) { 891 TCP_TIMER_ARM(tp, TCPT_REXMT, tp->t_rxtcur); 892 if (TCP_TIMER_ISARMED(tp, TCPT_PERSIST)) { 893 TCP_TIMER_DISARM(tp, TCPT_PERSIST); 894 tp->t_rxtshift = 0; 895 } 896 } 897 } else 898 if (SEQ_GT(tp->snd_nxt + len, tp->snd_max)) 899 tp->snd_max = tp->snd_nxt + len; 900 901 /* 902 * Trace. 903 */ 904 if (so->so_options & SO_DEBUG) { 905 /* 906 * need to recover version # field, which was overwritten 907 * on ip_cksum computation. 908 */ 909 struct ip *sip; 910 sip = mtod(m, struct ip *); 911 switch (af) { 912 case AF_INET: 913 sip->ip_v = 4; 914 break; 915 #ifdef INET6 916 case AF_INET6: 917 sip->ip_v = 6; 918 break; 919 #endif 920 } 921 tcp_trace(TA_OUTPUT, tp->t_state, tp, m, 0); 922 } 923 924 /* 925 * Fill in IP length and desired time to live and 926 * send to IP level. There should be a better way 927 * to handle ttl and tos; we could keep them in 928 * the template, but need a way to checksum without them. 929 */ 930 m->m_pkthdr.len = hdrlen + len; 931 932 switch (af) { 933 case AF_INET: 934 ip->ip_len = m->m_pkthdr.len; 935 if (tp->t_inpcb) { 936 ip->ip_ttl = tp->t_inpcb->inp_ip.ip_ttl; 937 ip->ip_tos = tp->t_inpcb->inp_ip.ip_tos; 938 } 939 #ifdef INET6 940 else if (tp->t_in6pcb) { 941 ip->ip_ttl = tp->t_in6pcb->in6p_ip6.ip6_hlim; 942 ip->ip_tos = 0; /*XXX*/ 943 } 944 #endif 945 break; 946 #ifdef INET6 947 case AF_INET6: 948 ip6->ip6_nxt = IPPROTO_TCP; 949 if (tp->t_in6pcb) 950 ip6->ip6_hlim = tp->t_in6pcb->in6p_ip6.ip6_hlim; 951 /* ip6->ip6_flow = ??? */ 952 /* ip6_plen will be filled in ip6_output(). */ 953 break; 954 #endif 955 } 956 957 /* 958 * If we're doing Path MTU discovery, we need to set DF unless 959 * the route's MTU is locked. If we lack a route, we need to 960 * look it up now. 961 * 962 * ip_output() could do this for us, but it's convenient to just 963 * do it here unconditionally. 964 */ 965 if ((rt = ro->ro_rt) == NULL || (rt->rt_flags & RTF_UP) == 0) { 966 if (ro->ro_rt != NULL) { 967 RTFREE(ro->ro_rt); 968 ro->ro_rt = NULL; 969 } 970 switch (af) { 971 case AF_INET: 972 { 973 struct sockaddr_in *dst; 974 dst = satosin(&ro->ro_dst); 975 dst->sin_family = AF_INET; 976 dst->sin_len = sizeof(*dst); 977 if (tp->t_inpcb) 978 dst->sin_addr = tp->t_inpcb->inp_faddr; 979 #ifdef INET6 980 else if (tp->t_in6pcb) { 981 bcopy(&tp->t_in6pcb->in6p_faddr.s6_addr32[3], 982 &dst->sin_addr, sizeof(dst->sin_addr)); 983 } 984 #endif 985 break; 986 } 987 #ifdef INET6 988 case AF_INET6: 989 { 990 struct sockaddr_in6 *dst; 991 dst = satosin6(&ro->ro_dst); 992 dst->sin6_family = AF_INET6; 993 dst->sin6_len = sizeof(*dst); 994 dst->sin6_addr = tp->t_in6pcb->in6p_faddr; 995 break; 996 } 997 #endif 998 } 999 rtalloc(ro); 1000 if ((rt = ro->ro_rt) == NULL) { 1001 m_freem(m); 1002 switch (af) { 1003 case AF_INET: 1004 ipstat.ips_noroute++; 1005 break; 1006 #ifdef INET6 1007 case AF_INET6: 1008 ip6stat.ip6s_noroute++; 1009 break; 1010 #endif 1011 } 1012 error = EHOSTUNREACH; 1013 goto out; 1014 } 1015 } 1016 #ifdef IPSEC 1017 m->m_pkthdr.rcvif = (struct ifnet *)so; 1018 #endif /*IPSEC*/ 1019 1020 switch (af) { 1021 case AF_INET: 1022 { 1023 struct mbuf *opts; 1024 1025 if (ip_mtudisc != 0 && (rt->rt_rmx.rmx_locks & RTV_MTU) == 0) 1026 ip->ip_off |= IP_DF; 1027 1028 #if BSD >= 43 1029 if (tp->t_inpcb) 1030 opts = tp->t_inpcb->inp_options; 1031 else 1032 opts = NULL; 1033 error = ip_output(m, opts, ro, 1034 so->so_options & SO_DONTROUTE, 0); 1035 #else 1036 opts = NULL; 1037 error = ip_output(m, opts, ro, so->so_options & SO_DONTROUTE); 1038 #endif 1039 break; 1040 } 1041 #ifdef INET6 1042 case AF_INET6: 1043 { 1044 struct ip6_pktopts *opts; 1045 1046 #if BSD >= 43 1047 if (tp->t_in6pcb) 1048 opts = tp->t_in6pcb->in6p_outputopts; 1049 else 1050 opts = NULL; 1051 error = ip6_output(m, opts, (struct route_in6 *)ro, 1052 so->so_options & SO_DONTROUTE, 0); 1053 #else 1054 opts = NULL; 1055 error = ip6_output(m, opts, (struct route_in6 *)ro, 1056 so->so_options & SO_DONTROUTE); 1057 #endif 1058 break; 1059 } 1060 #endif 1061 default: 1062 error = EAFNOSUPPORT; 1063 break; 1064 } 1065 if (error) { 1066 out: 1067 if (error == ENOBUFS) { 1068 if (tp->t_inpcb) 1069 tcp_quench(tp->t_inpcb, 0); 1070 #if 0 /*XXX def INET6*/ 1071 else if (tp->t_in6pcb) 1072 tcp6_quench(tp->t_in6pcb, 0); 1073 #endif 1074 return (0); 1075 } 1076 if ((error == EHOSTUNREACH || error == ENETDOWN) 1077 && TCPS_HAVERCVDSYN(tp->t_state)) { 1078 tp->t_softerror = error; 1079 return (0); 1080 } 1081 return (error); 1082 } 1083 tcpstat.tcps_sndtotal++; 1084 if (tp->t_flags & TF_DELACK) 1085 tcpstat.tcps_delack++; 1086 1087 /* 1088 * Data sent (as far as we can tell). 1089 * If this advertises a larger window than any other segment, 1090 * then remember the size of the advertised window. 1091 * Any pending ACK has now been sent. 1092 */ 1093 if (win > 0 && SEQ_GT(tp->rcv_nxt+win, tp->rcv_adv)) 1094 tp->rcv_adv = tp->rcv_nxt + win; 1095 tp->last_ack_sent = tp->rcv_nxt; 1096 tp->t_flags &= ~TF_ACKNOW; 1097 TCP_CLEAR_DELACK(tp); 1098 #ifdef DIAGNOSTIC 1099 if (maxburst < 0) 1100 printf("tcp_output: maxburst exceeded by %d\n", -maxburst); 1101 #endif 1102 if (sendalot && (!tcp_do_newreno || --maxburst)) 1103 goto again; 1104 return (0); 1105 } 1106 1107 void 1108 tcp_setpersist(tp) 1109 register struct tcpcb *tp; 1110 { 1111 register int t = ((tp->t_srtt >> 2) + tp->t_rttvar) >> (1 + 2); 1112 int nticks; 1113 1114 if (TCP_TIMER_ISARMED(tp, TCPT_REXMT)) 1115 panic("tcp_output REXMT"); 1116 /* 1117 * Start/restart persistance timer. 1118 */ 1119 if (t < tp->t_rttmin) 1120 t = tp->t_rttmin; 1121 TCPT_RANGESET(nticks, t * tcp_backoff[tp->t_rxtshift], 1122 TCPTV_PERSMIN, TCPTV_PERSMAX); 1123 TCP_TIMER_ARM(tp, TCPT_PERSIST, nticks); 1124 if (tp->t_rxtshift < TCP_MAXRXTSHIFT) 1125 tp->t_rxtshift++; 1126 } 1127