tcp_output.c revision 1.115
1 /* $NetBSD: tcp_output.c,v 1.115 2004/12/15 04:25:19 thorpej Exp $ */ 2 3 /* 4 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. 5 * 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. Neither the name of the project nor the names of its contributors 16 * may be used to endorse or promote products derived from this software 17 * without specific prior written permission. 18 * 19 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND 20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 22 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE 23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 29 * SUCH DAMAGE. 30 */ 31 32 /* 33 * @(#)COPYRIGHT 1.1 (NRL) 17 January 1995 34 * 35 * NRL grants permission for redistribution and use in source and binary 36 * forms, with or without modification, of the software and documentation 37 * created at NRL provided that the following conditions are met: 38 * 39 * 1. Redistributions of source code must retain the above copyright 40 * notice, this list of conditions and the following disclaimer. 41 * 2. Redistributions in binary form must reproduce the above copyright 42 * notice, this list of conditions and the following disclaimer in the 43 * documentation and/or other materials provided with the distribution. 44 * 3. All advertising materials mentioning features or use of this software 45 * must display the following acknowledgements: 46 * This product includes software developed by the University of 47 * California, Berkeley and its contributors. 48 * This product includes software developed at the Information 49 * Technology Division, US Naval Research Laboratory. 50 * 4. Neither the name of the NRL nor the names of its contributors 51 * may be used to endorse or promote products derived from this software 52 * without specific prior written permission. 53 * 54 * THE SOFTWARE PROVIDED BY NRL IS PROVIDED BY NRL AND CONTRIBUTORS ``AS 55 * IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 56 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A 57 * PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL NRL OR 58 * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, 59 * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, 60 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR 61 * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF 62 * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING 63 * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS 64 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 65 * 66 * The views and conclusions contained in the software and documentation 67 * are those of the authors and should not be interpreted as representing 68 * official policies, either expressed or implied, of the US Naval 69 * Research Laboratory (NRL). 70 */ 71 72 /*- 73 * Copyright (c) 1997, 1998, 2001 The NetBSD Foundation, Inc. 74 * All rights reserved. 75 * 76 * This code is derived from software contributed to The NetBSD Foundation 77 * by Jason R. Thorpe and Kevin M. Lahey of the Numerical Aerospace Simulation 78 * Facility, NASA Ames Research Center. 79 * 80 * Redistribution and use in source and binary forms, with or without 81 * modification, are permitted provided that the following conditions 82 * are met: 83 * 1. Redistributions of source code must retain the above copyright 84 * notice, this list of conditions and the following disclaimer. 85 * 2. Redistributions in binary form must reproduce the above copyright 86 * notice, this list of conditions and the following disclaimer in the 87 * documentation and/or other materials provided with the distribution. 88 * 3. All advertising materials mentioning features or use of this software 89 * must display the following acknowledgement: 90 * This product includes software developed by the NetBSD 91 * Foundation, Inc. and its contributors. 92 * 4. Neither the name of The NetBSD Foundation nor the names of its 93 * contributors may be used to endorse or promote products derived 94 * from this software without specific prior written permission. 95 * 96 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 97 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 98 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 99 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 100 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 101 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 102 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 103 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 104 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 105 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 106 * POSSIBILITY OF SUCH DAMAGE. 107 */ 108 109 /* 110 * Copyright (c) 1982, 1986, 1988, 1990, 1993, 1995 111 * The Regents of the University of California. All rights reserved. 112 * 113 * Redistribution and use in source and binary forms, with or without 114 * modification, are permitted provided that the following conditions 115 * are met: 116 * 1. Redistributions of source code must retain the above copyright 117 * notice, this list of conditions and the following disclaimer. 118 * 2. Redistributions in binary form must reproduce the above copyright 119 * notice, this list of conditions and the following disclaimer in the 120 * documentation and/or other materials provided with the distribution. 121 * 3. Neither the name of the University nor the names of its contributors 122 * may be used to endorse or promote products derived from this software 123 * without specific prior written permission. 124 * 125 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 126 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 127 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 128 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 129 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 130 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 131 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 132 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 133 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 134 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 135 * SUCH DAMAGE. 136 * 137 * @(#)tcp_output.c 8.4 (Berkeley) 5/24/95 138 */ 139 140 #include <sys/cdefs.h> 141 __KERNEL_RCSID(0, "$NetBSD: tcp_output.c,v 1.115 2004/12/15 04:25:19 thorpej Exp $"); 142 143 #include "opt_inet.h" 144 #include "opt_ipsec.h" 145 #include "opt_tcp_debug.h" 146 147 #include <sys/param.h> 148 #include <sys/systm.h> 149 #include <sys/malloc.h> 150 #include <sys/mbuf.h> 151 #include <sys/protosw.h> 152 #include <sys/socket.h> 153 #include <sys/socketvar.h> 154 #include <sys/errno.h> 155 #include <sys/domain.h> 156 #include <sys/kernel.h> 157 #ifdef TCP_SIGNATURE 158 #include <sys/md5.h> 159 #endif 160 161 #include <net/if.h> 162 #include <net/route.h> 163 164 #include <netinet/in.h> 165 #include <netinet/in_systm.h> 166 #include <netinet/ip.h> 167 #include <netinet/in_pcb.h> 168 #include <netinet/ip_var.h> 169 170 #ifdef INET6 171 #ifndef INET 172 #include <netinet/in.h> 173 #endif 174 #include <netinet/ip6.h> 175 #include <netinet6/in6_var.h> 176 #include <netinet6/ip6_var.h> 177 #include <netinet6/in6_pcb.h> 178 #include <netinet6/nd6.h> 179 #endif 180 181 #ifdef FAST_IPSEC 182 #include <netipsec/ipsec.h> 183 #include <netipsec/key.h> 184 #endif /* FAST_IPSEC*/ 185 #ifdef IPSEC 186 #include <netinet6/ipsec.h> 187 #endif 188 189 #include <netinet/tcp.h> 190 #define TCPOUTFLAGS 191 #include <netinet/tcp_fsm.h> 192 #include <netinet/tcp_seq.h> 193 #include <netinet/tcp_timer.h> 194 #include <netinet/tcp_var.h> 195 #include <netinet/tcpip.h> 196 #include <netinet/tcp_debug.h> 197 198 #ifdef IPSEC 199 #include <netkey/key.h> 200 #endif 201 202 #ifdef notyet 203 extern struct mbuf *m_copypack(); 204 #endif 205 206 #define MAX_TCPOPTLEN 32 /* max # bytes that go in options */ 207 208 /* 209 * Knob to enable Congestion Window Monitoring, and control the 210 * the burst size it allows. Default burst is 4 packets, per 211 * the Internet draft. 212 */ 213 int tcp_cwm = 0; 214 int tcp_cwm_burstsize = 4; 215 216 #ifdef TCP_OUTPUT_COUNTERS 217 #include <sys/device.h> 218 219 extern struct evcnt tcp_output_bigheader; 220 extern struct evcnt tcp_output_predict_hit; 221 extern struct evcnt tcp_output_predict_miss; 222 extern struct evcnt tcp_output_copysmall; 223 extern struct evcnt tcp_output_copybig; 224 extern struct evcnt tcp_output_refbig; 225 226 #define TCP_OUTPUT_COUNTER_INCR(ev) (ev)->ev_count++ 227 #else 228 229 #define TCP_OUTPUT_COUNTER_INCR(ev) /* nothing */ 230 231 #endif /* TCP_OUTPUT_COUNTERS */ 232 233 static 234 #ifndef GPROF 235 __inline 236 #endif 237 int 238 tcp_segsize(struct tcpcb *tp, int *txsegsizep, int *rxsegsizep) 239 { 240 #ifdef INET 241 struct inpcb *inp = tp->t_inpcb; 242 #endif 243 #ifdef INET6 244 struct in6pcb *in6p = tp->t_in6pcb; 245 #endif 246 struct socket *so = NULL; 247 struct rtentry *rt; 248 struct ifnet *ifp; 249 int size; 250 int iphlen; 251 int optlen; 252 253 #ifdef DIAGNOSTIC 254 if (tp->t_inpcb && tp->t_in6pcb) 255 panic("tcp_segsize: both t_inpcb and t_in6pcb are set"); 256 #endif 257 switch (tp->t_family) { 258 #ifdef INET 259 case AF_INET: 260 iphlen = sizeof(struct ip); 261 break; 262 #endif 263 #ifdef INET6 264 case AF_INET6: 265 iphlen = sizeof(struct ip6_hdr); 266 break; 267 #endif 268 default: 269 size = tcp_mssdflt; 270 goto out; 271 } 272 273 rt = NULL; 274 #ifdef INET 275 if (inp) { 276 rt = in_pcbrtentry(inp); 277 so = inp->inp_socket; 278 } 279 #endif 280 #ifdef INET6 281 if (in6p) { 282 rt = in6_pcbrtentry(in6p); 283 so = in6p->in6p_socket; 284 } 285 #endif 286 if (rt == NULL) { 287 size = tcp_mssdflt; 288 goto out; 289 } 290 291 ifp = rt->rt_ifp; 292 293 size = tcp_mssdflt; 294 if (tp->t_mtudisc && rt->rt_rmx.rmx_mtu != 0) { 295 #ifdef INET6 296 if (in6p && rt->rt_rmx.rmx_mtu < IPV6_MMTU) { 297 /* 298 * RFC2460 section 5, last paragraph: if path MTU is 299 * smaller than 1280, use 1280 as packet size and 300 * attach fragment header. 301 */ 302 size = IPV6_MMTU - iphlen - sizeof(struct ip6_frag) - 303 sizeof(struct tcphdr); 304 } else 305 size = rt->rt_rmx.rmx_mtu - iphlen - 306 sizeof(struct tcphdr); 307 #else 308 size = rt->rt_rmx.rmx_mtu - iphlen - sizeof(struct tcphdr); 309 #endif 310 } else if (ifp->if_flags & IFF_LOOPBACK) 311 size = ifp->if_mtu - iphlen - sizeof(struct tcphdr); 312 #ifdef INET 313 else if (inp && tp->t_mtudisc) 314 size = ifp->if_mtu - iphlen - sizeof(struct tcphdr); 315 else if (inp && in_localaddr(inp->inp_faddr)) 316 size = ifp->if_mtu - iphlen - sizeof(struct tcphdr); 317 #endif 318 #ifdef INET6 319 else if (in6p) { 320 #ifdef INET 321 if (IN6_IS_ADDR_V4MAPPED(&in6p->in6p_faddr)) { 322 /* mapped addr case */ 323 struct in_addr d; 324 bcopy(&in6p->in6p_faddr.s6_addr32[3], &d, sizeof(d)); 325 if (tp->t_mtudisc || in_localaddr(d)) 326 size = ifp->if_mtu - iphlen - sizeof(struct tcphdr); 327 } else 328 #endif 329 { 330 /* 331 * for IPv6, path MTU discovery is always turned on, 332 * or the node must use packet size <= 1280. 333 */ 334 size = tp->t_mtudisc ? IN6_LINKMTU(ifp) : IPV6_MMTU; 335 size -= (iphlen + sizeof(struct tcphdr)); 336 } 337 } 338 #endif 339 out: 340 /* 341 * Now we must make room for whatever extra TCP/IP options are in 342 * the packet. 343 */ 344 optlen = tcp_optlen(tp); 345 346 /* 347 * XXX tp->t_ourmss should have the right size, but without this code 348 * fragmentation will occur... need more investigation 349 */ 350 #ifdef INET 351 if (inp) { 352 #if defined(IPSEC) || defined(FAST_IPSEC) 353 if (! IPSEC_PCB_SKIP_IPSEC(inp->inp_sp, IPSEC_DIR_OUTBOUND)) 354 optlen += ipsec4_hdrsiz_tcp(tp); 355 #endif 356 optlen += ip_optlen(inp); 357 } 358 #endif 359 #ifdef INET6 360 #ifdef INET 361 if (in6p && tp->t_family == AF_INET) { 362 #if defined(IPSEC) || defined(FAST_IPSEC) 363 if (! IPSEC_PCB_SKIP_IPSEC(in6p->in6p_sp, IPSEC_DIR_OUTBOUND)) 364 optlen += ipsec4_hdrsiz_tcp(tp); 365 #endif 366 /* XXX size -= ip_optlen(in6p); */ 367 } else 368 #endif 369 if (in6p && tp->t_family == AF_INET6) { 370 #ifdef IPSEC 371 if (! IPSEC_PCB_SKIP_IPSEC(in6p->in6p_sp, IPSEC_DIR_OUTBOUND)) 372 optlen += ipsec6_hdrsiz_tcp(tp); 373 #endif 374 optlen += ip6_optlen(in6p); 375 } 376 #endif 377 size -= optlen; 378 379 /* there may not be any room for data if mtu is too small */ 380 if (size < 0) 381 return (EMSGSIZE); 382 383 /* 384 * *rxsegsizep holds *estimated* inbound segment size (estimation 385 * assumes that path MTU is the same for both ways). this is only 386 * for silly window avoidance, do not use the value for other purposes. 387 * 388 * ipseclen is subtracted from both sides, this may not be right. 389 * I'm not quite sure about this (could someone comment). 390 */ 391 *txsegsizep = min(tp->t_peermss - optlen, size); 392 /* 393 * Never send more than half a buffer full. This insures that we can 394 * always keep 2 packets on the wire, no matter what SO_SNDBUF is, and 395 * therefore acks will never be delayed unless we run out of data to 396 * transmit. 397 */ 398 if (so) 399 *txsegsizep = min(so->so_snd.sb_hiwat >> 1, *txsegsizep); 400 *rxsegsizep = min(tp->t_ourmss - optlen, size); 401 402 if (*txsegsizep != tp->t_segsz) { 403 /* 404 * If the new segment size is larger, we don't want to 405 * mess up the congestion window, but if it is smaller 406 * we'll have to reduce the congestion window to ensure 407 * that we don't get into trouble with initial windows 408 * and the rest. In any case, if the segment size 409 * has changed, chances are the path has, too, and 410 * our congestion window will be different. 411 */ 412 if (*txsegsizep < tp->t_segsz) { 413 tp->snd_cwnd = max((tp->snd_cwnd / tp->t_segsz) 414 * *txsegsizep, *txsegsizep); 415 tp->snd_ssthresh = max((tp->snd_ssthresh / tp->t_segsz) 416 * *txsegsizep, *txsegsizep); 417 } 418 tp->t_segsz = *txsegsizep; 419 } 420 421 return (0); 422 } 423 424 static 425 #ifndef GPROF 426 __inline 427 #endif 428 int 429 tcp_build_datapkt(struct tcpcb *tp, struct socket *so, int off, 430 long len, int hdrlen, struct mbuf **mp) 431 { 432 struct mbuf *m, *m0; 433 434 if (tp->t_force && len == 1) 435 tcpstat.tcps_sndprobe++; 436 else if (SEQ_LT(tp->snd_nxt, tp->snd_max)) { 437 tcpstat.tcps_sndrexmitpack++; 438 tcpstat.tcps_sndrexmitbyte += len; 439 } else { 440 tcpstat.tcps_sndpack++; 441 tcpstat.tcps_sndbyte += len; 442 } 443 #ifdef notyet 444 if ((m = m_copypack(so->so_snd.sb_mb, off, 445 (int)len, max_linkhdr + hdrlen)) == 0) 446 return (ENOBUFS); 447 /* 448 * m_copypack left space for our hdr; use it. 449 */ 450 m->m_len += hdrlen; 451 m->m_data -= hdrlen; 452 #else 453 MGETHDR(m, M_DONTWAIT, MT_HEADER); 454 if (__predict_false(m == NULL)) 455 return (ENOBUFS); 456 MCLAIM(m, &tcp_tx_mowner); 457 458 /* 459 * XXX Because other code assumes headers will fit in 460 * XXX one header mbuf. 461 * 462 * (This code should almost *never* be run.) 463 */ 464 if (__predict_false((max_linkhdr + hdrlen) > MHLEN)) { 465 TCP_OUTPUT_COUNTER_INCR(&tcp_output_bigheader); 466 MCLGET(m, M_DONTWAIT); 467 if ((m->m_flags & M_EXT) == 0) { 468 m_freem(m); 469 return (ENOBUFS); 470 } 471 } 472 473 m->m_data += max_linkhdr; 474 m->m_len = hdrlen; 475 476 /* 477 * To avoid traversing the whole sb_mb chain for correct 478 * data to send, remember last sent mbuf, its offset and 479 * the sent size. When called the next time, see if the 480 * data to send is directly following the previous transfer. 481 * This is important for large TCP windows. 482 */ 483 if (off == 0 || tp->t_lastm == NULL || 484 (tp->t_lastoff + tp->t_lastlen) != off) { 485 TCP_OUTPUT_COUNTER_INCR(&tcp_output_predict_miss); 486 /* 487 * Either a new packet or a retransmit. 488 * Start from the beginning. 489 */ 490 tp->t_lastm = so->so_snd.sb_mb; 491 tp->t_inoff = off; 492 } else { 493 TCP_OUTPUT_COUNTER_INCR(&tcp_output_predict_hit); 494 tp->t_inoff += tp->t_lastlen; 495 } 496 497 /* Traverse forward to next packet */ 498 while (tp->t_inoff > 0) { 499 if (tp->t_lastm == NULL) 500 panic("tp->t_lastm == NULL"); 501 if (tp->t_inoff < tp->t_lastm->m_len) 502 break; 503 tp->t_inoff -= tp->t_lastm->m_len; 504 tp->t_lastm = tp->t_lastm->m_next; 505 } 506 507 tp->t_lastoff = off; 508 tp->t_lastlen = len; 509 m0 = tp->t_lastm; 510 off = tp->t_inoff; 511 512 if (len <= M_TRAILINGSPACE(m)) { 513 m_copydata(m0, off, (int) len, mtod(m, caddr_t) + hdrlen); 514 m->m_len += len; 515 TCP_OUTPUT_COUNTER_INCR(&tcp_output_copysmall); 516 } else { 517 m->m_next = m_copy(m0, off, (int) len); 518 if (m->m_next == NULL) { 519 m_freem(m); 520 return (ENOBUFS); 521 } 522 #ifdef TCP_OUTPUT_COUNTERS 523 if (m->m_next->m_flags & M_EXT) 524 TCP_OUTPUT_COUNTER_INCR(&tcp_output_refbig); 525 else 526 TCP_OUTPUT_COUNTER_INCR(&tcp_output_copybig); 527 #endif /* TCP_OUTPUT_COUNTERS */ 528 } 529 #endif 530 531 *mp = m; 532 return (0); 533 } 534 535 /* 536 * Tcp output routine: figure out what should be sent and send it. 537 */ 538 int 539 tcp_output(tp) 540 struct tcpcb *tp; 541 { 542 struct socket *so; 543 struct route *ro; 544 long len, win; 545 int off, flags, error; 546 struct mbuf *m; 547 struct ip *ip; 548 #ifdef INET6 549 struct ip6_hdr *ip6; 550 #endif 551 struct tcphdr *th; 552 u_char opt[MAX_TCPOPTLEN]; 553 unsigned optlen, hdrlen; 554 int idle, sendalot, txsegsize, rxsegsize; 555 int maxburst = TCP_MAXBURST; 556 int af; /* address family on the wire */ 557 int iphdrlen; 558 #ifdef TCP_SIGNATURE 559 int sigoff = 0; 560 #endif 561 562 #ifdef DIAGNOSTIC 563 if (tp->t_inpcb && tp->t_in6pcb) 564 panic("tcp_output: both t_inpcb and t_in6pcb are set"); 565 #endif 566 so = NULL; 567 ro = NULL; 568 if (tp->t_inpcb) { 569 so = tp->t_inpcb->inp_socket; 570 ro = &tp->t_inpcb->inp_route; 571 } 572 #ifdef INET6 573 else if (tp->t_in6pcb) { 574 so = tp->t_in6pcb->in6p_socket; 575 ro = (struct route *)&tp->t_in6pcb->in6p_route; 576 } 577 #endif 578 579 switch (af = tp->t_family) { 580 #ifdef INET 581 case AF_INET: 582 if (tp->t_inpcb) 583 break; 584 #ifdef INET6 585 /* mapped addr case */ 586 if (tp->t_in6pcb) 587 break; 588 #endif 589 return (EINVAL); 590 #endif 591 #ifdef INET6 592 case AF_INET6: 593 if (tp->t_in6pcb) 594 break; 595 return (EINVAL); 596 #endif 597 default: 598 return (EAFNOSUPPORT); 599 } 600 601 if (tcp_segsize(tp, &txsegsize, &rxsegsize)) 602 return (EMSGSIZE); 603 604 idle = (tp->snd_max == tp->snd_una); 605 606 /* 607 * Restart Window computation. From draft-floyd-incr-init-win-03: 608 * 609 * Optionally, a TCP MAY set the restart window to the 610 * minimum of the value used for the initial window and 611 * the current value of cwnd (in other words, using a 612 * larger value for the restart window should never increase 613 * the size of cwnd). 614 */ 615 if (tcp_cwm) { 616 /* 617 * Hughes/Touch/Heidemann Congestion Window Monitoring. 618 * Count the number of packets currently pending 619 * acknowledgement, and limit our congestion window 620 * to a pre-determined allowed burst size plus that count. 621 * This prevents bursting once all pending packets have 622 * been acknowledged (i.e. transmission is idle). 623 * 624 * XXX Link this to Initial Window? 625 */ 626 tp->snd_cwnd = min(tp->snd_cwnd, 627 (tcp_cwm_burstsize * txsegsize) + 628 (tp->snd_nxt - tp->snd_una)); 629 } else { 630 if (idle && (tcp_now - tp->t_rcvtime) >= tp->t_rxtcur) { 631 /* 632 * We have been idle for "a while" and no acks are 633 * expected to clock out any data we send -- 634 * slow start to get ack "clock" running again. 635 */ 636 int ss = tcp_init_win; 637 #ifdef INET 638 if (tp->t_inpcb && 639 in_localaddr(tp->t_inpcb->inp_faddr)) 640 ss = tcp_init_win_local; 641 #endif 642 #ifdef INET6 643 if (tp->t_in6pcb && 644 in6_localaddr(&tp->t_in6pcb->in6p_faddr)) 645 ss = tcp_init_win_local; 646 #endif 647 tp->snd_cwnd = min(tp->snd_cwnd, 648 TCP_INITIAL_WINDOW(ss, txsegsize)); 649 } 650 } 651 652 again: 653 /* 654 * Determine length of data that should be transmitted, and 655 * flags that should be used. If there is some data or critical 656 * controls (SYN, RST) to send, then transmit; otherwise, 657 * investigate further. 658 */ 659 sendalot = 0; 660 off = tp->snd_nxt - tp->snd_una; 661 win = min(tp->snd_wnd, tp->snd_cwnd); 662 663 flags = tcp_outflags[tp->t_state]; 664 /* 665 * If in persist timeout with window of 0, send 1 byte. 666 * Otherwise, if window is small but nonzero 667 * and timer expired, we will send what we can 668 * and go to transmit state. 669 */ 670 if (tp->t_force) { 671 if (win == 0) { 672 /* 673 * If we still have some data to send, then 674 * clear the FIN bit. Usually this would 675 * happen below when it realizes that we 676 * aren't sending all the data. However, 677 * if we have exactly 1 byte of unset data, 678 * then it won't clear the FIN bit below, 679 * and if we are in persist state, we wind 680 * up sending the packet without recording 681 * that we sent the FIN bit. 682 * 683 * We can't just blindly clear the FIN bit, 684 * because if we don't have any more data 685 * to send then the probe will be the FIN 686 * itself. 687 */ 688 if (off < so->so_snd.sb_cc) 689 flags &= ~TH_FIN; 690 win = 1; 691 } else { 692 TCP_TIMER_DISARM(tp, TCPT_PERSIST); 693 tp->t_rxtshift = 0; 694 } 695 } 696 697 if (win < so->so_snd.sb_cc) { 698 len = win - off; 699 flags &= ~TH_FIN; 700 } else 701 len = so->so_snd.sb_cc - off; 702 703 if (len < 0) { 704 /* 705 * If FIN has been sent but not acked, 706 * but we haven't been called to retransmit, 707 * len will be -1. Otherwise, window shrank 708 * after we sent into it. If window shrank to 0, 709 * cancel pending retransmit, pull snd_nxt back 710 * to (closed) window, and set the persist timer 711 * if it isn't already going. If the window didn't 712 * close completely, just wait for an ACK. 713 * 714 * If we have a pending FIN, either it has already been 715 * transmitted or it is outside the window, so drop it. 716 * If the FIN has been transmitted, but this is not a 717 * retransmission, then len must be -1. Therefore we also 718 * prevent here the sending of `gratuitous FINs'. This 719 * eliminates the need to check for that case below (e.g. 720 * to back up snd_nxt before the FIN so that the sequence 721 * number is correct). 722 */ 723 len = 0; 724 flags &= ~TH_FIN; 725 if (win == 0) { 726 TCP_TIMER_DISARM(tp, TCPT_REXMT); 727 tp->t_rxtshift = 0; 728 tp->snd_nxt = tp->snd_una; 729 if (TCP_TIMER_ISARMED(tp, TCPT_PERSIST) == 0) 730 tcp_setpersist(tp); 731 } 732 } 733 if (len > txsegsize) { 734 len = txsegsize; 735 flags &= ~TH_FIN; 736 sendalot = 1; 737 } 738 739 win = sbspace(&so->so_rcv); 740 741 /* 742 * Sender silly window avoidance. If connection is idle 743 * and can send all data, a maximum segment, 744 * at least a maximum default-size segment do it, 745 * or are forced, do it; otherwise don't bother. 746 * If peer's buffer is tiny, then send 747 * when window is at least half open. 748 * If retransmitting (possibly after persist timer forced us 749 * to send into a small window), then must resend. 750 */ 751 if (len) { 752 if (len == txsegsize) 753 goto send; 754 if ((so->so_state & SS_MORETOCOME) == 0 && 755 ((idle || tp->t_flags & TF_NODELAY) && 756 len + off >= so->so_snd.sb_cc)) 757 goto send; 758 if (tp->t_force) 759 goto send; 760 if (len >= tp->max_sndwnd / 2) 761 goto send; 762 if (SEQ_LT(tp->snd_nxt, tp->snd_max)) 763 goto send; 764 } 765 766 /* 767 * Compare available window to amount of window known to peer 768 * (as advertised window less next expected input). If the 769 * difference is at least twice the size of the largest segment 770 * we expect to receive (i.e. two segments) or at least 50% of 771 * the maximum possible window, then want to send a window update 772 * to peer. 773 */ 774 if (win > 0) { 775 /* 776 * "adv" is the amount we can increase the window, 777 * taking into account that we are limited by 778 * TCP_MAXWIN << tp->rcv_scale. 779 */ 780 long adv = min(win, (long)TCP_MAXWIN << tp->rcv_scale) - 781 (tp->rcv_adv - tp->rcv_nxt); 782 783 if (adv >= (long) (2 * rxsegsize)) 784 goto send; 785 if (2 * adv >= (long) so->so_rcv.sb_hiwat) 786 goto send; 787 } 788 789 /* 790 * Send if we owe peer an ACK. 791 */ 792 if (tp->t_flags & TF_ACKNOW) 793 goto send; 794 if (flags & (TH_SYN|TH_FIN|TH_RST)) 795 goto send; 796 if (SEQ_GT(tp->snd_up, tp->snd_una)) 797 goto send; 798 799 /* 800 * TCP window updates are not reliable, rather a polling protocol 801 * using ``persist'' packets is used to insure receipt of window 802 * updates. The three ``states'' for the output side are: 803 * idle not doing retransmits or persists 804 * persisting to move a small or zero window 805 * (re)transmitting and thereby not persisting 806 * 807 * tp->t_timer[TCPT_PERSIST] 808 * is set when we are in persist state. 809 * tp->t_force 810 * is set when we are called to send a persist packet. 811 * tp->t_timer[TCPT_REXMT] 812 * is set when we are retransmitting 813 * The output side is idle when both timers are zero. 814 * 815 * If send window is too small, there is data to transmit, and no 816 * retransmit or persist is pending, then go to persist state. 817 * If nothing happens soon, send when timer expires: 818 * if window is nonzero, transmit what we can, 819 * otherwise force out a byte. 820 */ 821 if (so->so_snd.sb_cc && TCP_TIMER_ISARMED(tp, TCPT_REXMT) == 0 && 822 TCP_TIMER_ISARMED(tp, TCPT_PERSIST) == 0) { 823 tp->t_rxtshift = 0; 824 tcp_setpersist(tp); 825 } 826 827 /* 828 * No reason to send a segment, just return. 829 */ 830 return (0); 831 832 send: 833 /* 834 * Before ESTABLISHED, force sending of initial options 835 * unless TCP set not to do any options. 836 * NOTE: we assume that the IP/TCP header plus TCP options 837 * always fit in a single mbuf, leaving room for a maximum 838 * link header, i.e. 839 * max_linkhdr + sizeof (struct tcpiphdr) + optlen <= MCLBYTES 840 */ 841 optlen = 0; 842 switch (af) { 843 #ifdef INET 844 case AF_INET: 845 iphdrlen = sizeof(struct ip) + sizeof(struct tcphdr); 846 break; 847 #endif 848 #ifdef INET6 849 case AF_INET6: 850 iphdrlen = sizeof(struct ip6_hdr) + sizeof(struct tcphdr); 851 break; 852 #endif 853 default: /*pacify gcc*/ 854 iphdrlen = 0; 855 break; 856 } 857 hdrlen = iphdrlen; 858 if (flags & TH_SYN) { 859 struct rtentry *rt; 860 861 rt = NULL; 862 #ifdef INET 863 if (tp->t_inpcb) 864 rt = in_pcbrtentry(tp->t_inpcb); 865 #endif 866 #ifdef INET6 867 if (tp->t_in6pcb) 868 rt = in6_pcbrtentry(tp->t_in6pcb); 869 #endif 870 871 tp->snd_nxt = tp->iss; 872 tp->t_ourmss = tcp_mss_to_advertise(rt != NULL ? 873 rt->rt_ifp : NULL, af); 874 if ((tp->t_flags & TF_NOOPT) == 0) { 875 opt[0] = TCPOPT_MAXSEG; 876 opt[1] = 4; 877 opt[2] = (tp->t_ourmss >> 8) & 0xff; 878 opt[3] = tp->t_ourmss & 0xff; 879 optlen = 4; 880 881 if ((tp->t_flags & TF_REQ_SCALE) && 882 ((flags & TH_ACK) == 0 || 883 (tp->t_flags & TF_RCVD_SCALE))) { 884 *((u_int32_t *) (opt + optlen)) = htonl( 885 TCPOPT_NOP << 24 | 886 TCPOPT_WINDOW << 16 | 887 TCPOLEN_WINDOW << 8 | 888 tp->request_r_scale); 889 optlen += 4; 890 } 891 } 892 } 893 894 /* 895 * Send a timestamp and echo-reply if this is a SYN and our side 896 * wants to use timestamps (TF_REQ_TSTMP is set) or both our side 897 * and our peer have sent timestamps in our SYN's. 898 */ 899 if ((tp->t_flags & (TF_REQ_TSTMP|TF_NOOPT)) == TF_REQ_TSTMP && 900 (flags & TH_RST) == 0 && 901 ((flags & (TH_SYN|TH_ACK)) == TH_SYN || 902 (tp->t_flags & TF_RCVD_TSTMP))) { 903 u_int32_t *lp = (u_int32_t *)(opt + optlen); 904 905 /* Form timestamp option as shown in appendix A of RFC 1323. */ 906 *lp++ = htonl(TCPOPT_TSTAMP_HDR); 907 *lp++ = htonl(TCP_TIMESTAMP(tp)); 908 *lp = htonl(tp->ts_recent); 909 optlen += TCPOLEN_TSTAMP_APPA; 910 } 911 912 #ifdef TCP_SIGNATURE 913 #if defined(INET6) && defined(FAST_IPSEC) 914 if (tp->t_family == AF_INET) 915 #endif 916 if (tp->t_flags & TF_SIGNATURE) { 917 u_char *bp; 918 /* 919 * Initialize TCP-MD5 option (RFC2385) 920 */ 921 bp = (u_char *)opt + optlen; 922 *bp++ = TCPOPT_SIGNATURE; 923 *bp++ = TCPOLEN_SIGNATURE; 924 sigoff = optlen + 2; 925 bzero(bp, TCP_SIGLEN); 926 bp += TCP_SIGLEN; 927 optlen += TCPOLEN_SIGNATURE; 928 /* 929 * Terminate options list and maintain 32-bit alignment. 930 */ 931 *bp++ = TCPOPT_NOP; 932 *bp++ = TCPOPT_EOL; 933 optlen += 2; 934 } 935 #endif /* TCP_SIGNATURE */ 936 937 hdrlen += optlen; 938 939 #ifdef DIAGNOSTIC 940 if (len > txsegsize) 941 panic("tcp data to be sent is larger than segment"); 942 if (max_linkhdr + hdrlen > MCLBYTES) 943 panic("tcphdr too big"); 944 #endif 945 946 /* 947 * Grab a header mbuf, attaching a copy of data to 948 * be transmitted, and initialize the header from 949 * the template for sends on this connection. 950 */ 951 if (len) { 952 error = tcp_build_datapkt(tp, so, off, len, hdrlen, &m); 953 if (error) 954 goto out; 955 /* 956 * If we're sending everything we've got, set PUSH. 957 * (This will keep happy those implementations which only 958 * give data to the user when a buffer fills or 959 * a PUSH comes in.) 960 */ 961 if (off + len == so->so_snd.sb_cc) 962 flags |= TH_PUSH; 963 } else { 964 if (tp->t_flags & TF_ACKNOW) 965 tcpstat.tcps_sndacks++; 966 else if (flags & (TH_SYN|TH_FIN|TH_RST)) 967 tcpstat.tcps_sndctrl++; 968 else if (SEQ_GT(tp->snd_up, tp->snd_una)) 969 tcpstat.tcps_sndurg++; 970 else 971 tcpstat.tcps_sndwinup++; 972 973 MGETHDR(m, M_DONTWAIT, MT_HEADER); 974 if (m != NULL && max_linkhdr + hdrlen > MHLEN) { 975 MCLGET(m, M_DONTWAIT); 976 if ((m->m_flags & M_EXT) == 0) { 977 m_freem(m); 978 m = NULL; 979 } 980 } 981 if (m == NULL) { 982 error = ENOBUFS; 983 goto out; 984 } 985 MCLAIM(m, &tcp_tx_mowner); 986 m->m_data += max_linkhdr; 987 m->m_len = hdrlen; 988 } 989 m->m_pkthdr.rcvif = (struct ifnet *)0; 990 switch (af) { 991 #ifdef INET 992 case AF_INET: 993 ip = mtod(m, struct ip *); 994 #ifdef INET6 995 ip6 = NULL; 996 #endif 997 th = (struct tcphdr *)(ip + 1); 998 break; 999 #endif 1000 #ifdef INET6 1001 case AF_INET6: 1002 ip = NULL; 1003 ip6 = mtod(m, struct ip6_hdr *); 1004 th = (struct tcphdr *)(ip6 + 1); 1005 break; 1006 #endif 1007 default: /*pacify gcc*/ 1008 ip = NULL; 1009 #ifdef INET6 1010 ip6 = NULL; 1011 #endif 1012 th = NULL; 1013 break; 1014 } 1015 if (tp->t_template == 0) 1016 panic("tcp_output"); 1017 if (tp->t_template->m_len < iphdrlen) 1018 panic("tcp_output"); 1019 bcopy(mtod(tp->t_template, caddr_t), mtod(m, caddr_t), iphdrlen); 1020 1021 /* 1022 * If we are doing retransmissions, then snd_nxt will 1023 * not reflect the first unsent octet. For ACK only 1024 * packets, we do not want the sequence number of the 1025 * retransmitted packet, we want the sequence number 1026 * of the next unsent octet. So, if there is no data 1027 * (and no SYN or FIN), use snd_max instead of snd_nxt 1028 * when filling in ti_seq. But if we are in persist 1029 * state, snd_max might reflect one byte beyond the 1030 * right edge of the window, so use snd_nxt in that 1031 * case, since we know we aren't doing a retransmission. 1032 * (retransmit and persist are mutually exclusive...) 1033 */ 1034 if (len || (flags & (TH_SYN|TH_FIN)) || 1035 TCP_TIMER_ISARMED(tp, TCPT_PERSIST)) 1036 th->th_seq = htonl(tp->snd_nxt); 1037 else 1038 th->th_seq = htonl(tp->snd_max); 1039 th->th_ack = htonl(tp->rcv_nxt); 1040 if (optlen) { 1041 bcopy((caddr_t)opt, (caddr_t)(th + 1), optlen); 1042 th->th_off = (sizeof (struct tcphdr) + optlen) >> 2; 1043 } 1044 th->th_flags = flags; 1045 /* 1046 * Calculate receive window. Don't shrink window, 1047 * but avoid silly window syndrome. 1048 */ 1049 if (win < (long)(so->so_rcv.sb_hiwat / 4) && win < (long)rxsegsize) 1050 win = 0; 1051 if (win > (long)TCP_MAXWIN << tp->rcv_scale) 1052 win = (long)TCP_MAXWIN << tp->rcv_scale; 1053 if (win < (long)(int32_t)(tp->rcv_adv - tp->rcv_nxt)) 1054 win = (long)(int32_t)(tp->rcv_adv - tp->rcv_nxt); 1055 th->th_win = htons((u_int16_t) (win>>tp->rcv_scale)); 1056 if (SEQ_GT(tp->snd_up, tp->snd_nxt)) { 1057 u_int32_t urp = tp->snd_up - tp->snd_nxt; 1058 if (urp > IP_MAXPACKET) 1059 urp = IP_MAXPACKET; 1060 th->th_urp = htons((u_int16_t)urp); 1061 th->th_flags |= TH_URG; 1062 } else 1063 /* 1064 * If no urgent pointer to send, then we pull 1065 * the urgent pointer to the left edge of the send window 1066 * so that it doesn't drift into the send window on sequence 1067 * number wraparound. 1068 */ 1069 tp->snd_up = tp->snd_una; /* drag it along */ 1070 1071 #ifdef TCP_SIGNATURE 1072 #if defined(INET6) && defined(FAST_IPSEC) 1073 if (tp->t_family == AF_INET) /* XXX */ 1074 #endif 1075 if (sigoff && (tp->t_flags & TF_SIGNATURE)) { 1076 struct secasvar *sav; 1077 u_int8_t *sigp; 1078 1079 sav = tcp_signature_getsav(m, th); 1080 1081 if (sav == NULL) { 1082 if (m) 1083 m_freem(m); 1084 return (EPERM); 1085 } 1086 1087 m->m_pkthdr.len = hdrlen + len; 1088 sigp = (caddr_t)th + sizeof(*th) + sigoff; 1089 tcp_signature(m, th, (caddr_t)th - mtod(m, caddr_t), sav, sigp); 1090 1091 key_sa_recordxfer(sav, m); 1092 #ifdef FAST_IPSEC 1093 KEY_FREESAV(&sav); 1094 #else 1095 key_freesav(sav); 1096 #endif 1097 } 1098 #endif 1099 1100 /* 1101 * Set ourselves up to be checksummed just before the packet 1102 * hits the wire. Maybe skip checksums on loopback interfaces. 1103 */ 1104 switch (af) { 1105 #ifdef INET 1106 case AF_INET: 1107 if (__predict_true(ro->ro_rt == NULL || 1108 !(ro->ro_rt->rt_ifp->if_flags & 1109 IFF_LOOPBACK) || 1110 tcp_do_loopback_cksum)) 1111 m->m_pkthdr.csum_flags = M_CSUM_TCPv4; 1112 else 1113 m->m_pkthdr.csum_flags = 0; 1114 m->m_pkthdr.csum_data = offsetof(struct tcphdr, th_sum); 1115 if (len + optlen) { 1116 /* Fixup the pseudo-header checksum. */ 1117 /* XXXJRT Not IP Jumbogram safe. */ 1118 th->th_sum = in_cksum_addword(th->th_sum, 1119 htons((u_int16_t) (len + optlen))); 1120 } 1121 break; 1122 #endif 1123 #ifdef INET6 1124 case AF_INET6: 1125 /* 1126 * XXX Actually delaying the checksum is Hard 1127 * XXX (well, maybe not for Itojun, but it is 1128 * XXX for me), but we can still take advantage 1129 * XXX of the cached pseudo-header checksum. 1130 */ 1131 /* equals to hdrlen + len */ 1132 m->m_pkthdr.len = sizeof(struct ip6_hdr) 1133 + sizeof(struct tcphdr) + optlen + len; 1134 #ifdef notyet 1135 if (__predict_true(ro->ro_rt == NULL || 1136 !(ro->ro_rt->rt_ifp->if_flags & 1137 IFF_LOOPBACK) || 1138 tcp_do_loopback_cksum)) 1139 m->m_pkthdr.csum_flags = M_CSUM_TCPv6; 1140 else 1141 m->m_pkthdr.csum_flags = 0; 1142 m->m_pkthdr.csum_data = offsetof(struct tcphdr, th_sum); 1143 #endif 1144 if (len + optlen) { 1145 /* Fixup the pseudo-header checksum. */ 1146 /* XXXJRT: Not IPv6 Jumbogram safe. */ 1147 th->th_sum = in_cksum_addword(th->th_sum, 1148 htons((u_int16_t) (len + optlen))); 1149 } 1150 #ifndef notyet 1151 if (__predict_true(ro->ro_rt == NULL || 1152 !(ro->ro_rt->rt_ifp->if_flags & 1153 IFF_LOOPBACK) || 1154 tcp_do_loopback_cksum)) 1155 th->th_sum = in6_cksum(m, 0, sizeof(struct ip6_hdr), 1156 sizeof(struct tcphdr) + optlen + len); 1157 #endif 1158 break; 1159 #endif 1160 } 1161 1162 /* 1163 * In transmit state, time the transmission and arrange for 1164 * the retransmit. In persist state, just set snd_max. 1165 */ 1166 if (tp->t_force == 0 || TCP_TIMER_ISARMED(tp, TCPT_PERSIST) == 0) { 1167 tcp_seq startseq = tp->snd_nxt; 1168 1169 /* 1170 * Advance snd_nxt over sequence space of this segment. 1171 * There are no states in which we send both a SYN and a FIN, 1172 * so we collapse the tests for these flags. 1173 */ 1174 if (flags & (TH_SYN|TH_FIN)) 1175 tp->snd_nxt++; 1176 tp->snd_nxt += len; 1177 if (SEQ_GT(tp->snd_nxt, tp->snd_max)) { 1178 tp->snd_max = tp->snd_nxt; 1179 /* 1180 * Time this transmission if not a retransmission and 1181 * not currently timing anything. 1182 */ 1183 if (tp->t_rtttime == 0) { 1184 tp->t_rtttime = tcp_now; 1185 tp->t_rtseq = startseq; 1186 tcpstat.tcps_segstimed++; 1187 } 1188 } 1189 1190 /* 1191 * Set retransmit timer if not currently set, 1192 * and not doing an ack or a keep-alive probe. 1193 * Initial value for retransmit timer is smoothed 1194 * round-trip time + 2 * round-trip time variance. 1195 * Initialize shift counter which is used for backoff 1196 * of retransmit time. 1197 */ 1198 if (TCP_TIMER_ISARMED(tp, TCPT_REXMT) == 0 && 1199 tp->snd_nxt != tp->snd_una) { 1200 if (TCP_TIMER_ISARMED(tp, TCPT_PERSIST)) { 1201 TCP_TIMER_DISARM(tp, TCPT_PERSIST); 1202 tp->t_rxtshift = 0; 1203 } 1204 TCP_TIMER_ARM(tp, TCPT_REXMT, tp->t_rxtcur); 1205 } 1206 } else 1207 if (SEQ_GT(tp->snd_nxt + len, tp->snd_max)) 1208 tp->snd_max = tp->snd_nxt + len; 1209 1210 #ifdef TCP_DEBUG 1211 /* 1212 * Trace. 1213 */ 1214 if (so->so_options & SO_DEBUG) 1215 tcp_trace(TA_OUTPUT, tp->t_state, tp, m, 0); 1216 #endif 1217 1218 /* 1219 * Fill in IP length and desired time to live and 1220 * send to IP level. There should be a better way 1221 * to handle ttl and tos; we could keep them in 1222 * the template, but need a way to checksum without them. 1223 */ 1224 m->m_pkthdr.len = hdrlen + len; 1225 1226 switch (af) { 1227 #ifdef INET 1228 case AF_INET: 1229 ip->ip_len = htons(m->m_pkthdr.len); 1230 if (tp->t_inpcb) { 1231 ip->ip_ttl = tp->t_inpcb->inp_ip.ip_ttl; 1232 ip->ip_tos = tp->t_inpcb->inp_ip.ip_tos; 1233 } 1234 #ifdef INET6 1235 else if (tp->t_in6pcb) { 1236 ip->ip_ttl = in6_selecthlim(tp->t_in6pcb, NULL); /*XXX*/ 1237 ip->ip_tos = 0; /*XXX*/ 1238 } 1239 #endif 1240 break; 1241 #endif 1242 #ifdef INET6 1243 case AF_INET6: 1244 ip6->ip6_nxt = IPPROTO_TCP; 1245 if (tp->t_in6pcb) { 1246 /* 1247 * we separately set hoplimit for every segment, since 1248 * the user might want to change the value via 1249 * setsockopt. Also, desired default hop limit might 1250 * be changed via Neighbor Discovery. 1251 */ 1252 ip6->ip6_hlim = in6_selecthlim(tp->t_in6pcb, 1253 ro->ro_rt ? ro->ro_rt->rt_ifp : NULL); 1254 } 1255 /* ip6->ip6_flow = ??? */ 1256 /* ip6_plen will be filled in ip6_output(). */ 1257 break; 1258 #endif 1259 } 1260 1261 switch (af) { 1262 #ifdef INET 1263 case AF_INET: 1264 { 1265 struct mbuf *opts; 1266 1267 if (tp->t_inpcb) 1268 opts = tp->t_inpcb->inp_options; 1269 else 1270 opts = NULL; 1271 error = ip_output(m, opts, ro, 1272 (tp->t_mtudisc ? IP_MTUDISC : 0) | 1273 (so->so_options & SO_DONTROUTE), 1274 (struct ip_moptions *)0, so); 1275 break; 1276 } 1277 #endif 1278 #ifdef INET6 1279 case AF_INET6: 1280 { 1281 struct ip6_pktopts *opts; 1282 1283 if (tp->t_in6pcb) 1284 opts = tp->t_in6pcb->in6p_outputopts; 1285 else 1286 opts = NULL; 1287 error = ip6_output(m, opts, (struct route_in6 *)ro, 1288 so->so_options & SO_DONTROUTE, 1289 (struct ip6_moptions *)0, so, NULL); 1290 break; 1291 } 1292 #endif 1293 default: 1294 error = EAFNOSUPPORT; 1295 break; 1296 } 1297 if (error) { 1298 out: 1299 if (error == ENOBUFS) { 1300 tcpstat.tcps_selfquench++; 1301 #ifdef INET 1302 if (tp->t_inpcb) 1303 tcp_quench(tp->t_inpcb, 0); 1304 #endif 1305 #ifdef INET6 1306 if (tp->t_in6pcb) 1307 tcp6_quench(tp->t_in6pcb, 0); 1308 #endif 1309 error = 0; 1310 } else if ((error == EHOSTUNREACH || error == ENETDOWN) && 1311 TCPS_HAVERCVDSYN(tp->t_state)) { 1312 tp->t_softerror = error; 1313 error = 0; 1314 } 1315 1316 /* Restart the delayed ACK timer, if necessary. */ 1317 if (tp->t_flags & TF_DELACK) 1318 TCP_RESTART_DELACK(tp); 1319 1320 return (error); 1321 } 1322 tcpstat.tcps_sndtotal++; 1323 if (tp->t_flags & TF_DELACK) 1324 tcpstat.tcps_delack++; 1325 1326 /* 1327 * Data sent (as far as we can tell). 1328 * If this advertises a larger window than any other segment, 1329 * then remember the size of the advertised window. 1330 * Any pending ACK has now been sent. 1331 */ 1332 if (win > 0 && SEQ_GT(tp->rcv_nxt+win, tp->rcv_adv)) 1333 tp->rcv_adv = tp->rcv_nxt + win; 1334 tp->last_ack_sent = tp->rcv_nxt; 1335 tp->t_flags &= ~TF_ACKNOW; 1336 TCP_CLEAR_DELACK(tp); 1337 #ifdef DIAGNOSTIC 1338 if (maxburst < 0) 1339 printf("tcp_output: maxburst exceeded by %d\n", -maxburst); 1340 #endif 1341 if (sendalot && (!tcp_do_newreno || --maxburst)) 1342 goto again; 1343 return (0); 1344 } 1345 1346 void 1347 tcp_setpersist(tp) 1348 struct tcpcb *tp; 1349 { 1350 int t = ((tp->t_srtt >> 2) + tp->t_rttvar) >> (1 + 2); 1351 int nticks; 1352 1353 if (TCP_TIMER_ISARMED(tp, TCPT_REXMT)) 1354 panic("tcp_output REXMT"); 1355 /* 1356 * Start/restart persistance timer. 1357 */ 1358 if (t < tp->t_rttmin) 1359 t = tp->t_rttmin; 1360 TCPT_RANGESET(nticks, t * tcp_backoff[tp->t_rxtshift], 1361 TCPTV_PERSMIN, TCPTV_PERSMAX); 1362 TCP_TIMER_ARM(tp, TCPT_PERSIST, nticks); 1363 if (tp->t_rxtshift < TCP_MAXRXTSHIFT) 1364 tp->t_rxtshift++; 1365 } 1366
Indexes created Tue Sep 23 14:10:03 GMT 2025