tcp_subr.c revision 1.134
1 /* $NetBSD: tcp_subr.c,v 1.134 2002/09/25 11:19:23 itojun 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 (c) 1997, 1998, 2000, 2001 The NetBSD Foundation, Inc. 34 * All rights reserved. 35 * 36 * This code is derived from software contributed to The NetBSD Foundation 37 * by Jason R. Thorpe and Kevin M. Lahey of the Numerical Aerospace Simulation 38 * Facility, NASA Ames Research Center. 39 * 40 * Redistribution and use in source and binary forms, with or without 41 * modification, are permitted provided that the following conditions 42 * are met: 43 * 1. Redistributions of source code must retain the above copyright 44 * notice, this list of conditions and the following disclaimer. 45 * 2. Redistributions in binary form must reproduce the above copyright 46 * notice, this list of conditions and the following disclaimer in the 47 * documentation and/or other materials provided with the distribution. 48 * 3. All advertising materials mentioning features or use of this software 49 * must display the following acknowledgement: 50 * This product includes software developed by the NetBSD 51 * Foundation, Inc. and its contributors. 52 * 4. Neither the name of The NetBSD Foundation nor the names of its 53 * contributors may be used to endorse or promote products derived 54 * from this software without specific prior written permission. 55 * 56 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 57 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 58 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 59 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 60 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 61 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 62 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 63 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 64 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 65 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 66 * POSSIBILITY OF SUCH DAMAGE. 67 */ 68 69 /* 70 * Copyright (c) 1982, 1986, 1988, 1990, 1993, 1995 71 * The Regents of the University of California. All rights reserved. 72 * 73 * Redistribution and use in source and binary forms, with or without 74 * modification, are permitted provided that the following conditions 75 * are met: 76 * 1. Redistributions of source code must retain the above copyright 77 * notice, this list of conditions and the following disclaimer. 78 * 2. Redistributions in binary form must reproduce the above copyright 79 * notice, this list of conditions and the following disclaimer in the 80 * documentation and/or other materials provided with the distribution. 81 * 3. All advertising materials mentioning features or use of this software 82 * must display the following acknowledgement: 83 * This product includes software developed by the University of 84 * California, Berkeley and its contributors. 85 * 4. Neither the name of the University nor the names of its contributors 86 * may be used to endorse or promote products derived from this software 87 * without specific prior written permission. 88 * 89 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 90 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 91 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 92 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 93 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 94 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 95 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 96 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 97 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 98 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 99 * SUCH DAMAGE. 100 * 101 * @(#)tcp_subr.c 8.2 (Berkeley) 5/24/95 102 */ 103 104 #include <sys/cdefs.h> 105 __KERNEL_RCSID(0, "$NetBSD: tcp_subr.c,v 1.134 2002/09/25 11:19:23 itojun Exp $"); 106 107 #include "opt_inet.h" 108 #include "opt_ipsec.h" 109 #include "opt_tcp_compat_42.h" 110 #include "opt_inet_csum.h" 111 #include "rnd.h" 112 113 #include <sys/param.h> 114 #include <sys/proc.h> 115 #include <sys/systm.h> 116 #include <sys/malloc.h> 117 #include <sys/mbuf.h> 118 #include <sys/socket.h> 119 #include <sys/socketvar.h> 120 #include <sys/protosw.h> 121 #include <sys/errno.h> 122 #include <sys/kernel.h> 123 #include <sys/pool.h> 124 #if NRND > 0 125 #include <sys/md5.h> 126 #include <sys/rnd.h> 127 #endif 128 129 #include <net/route.h> 130 #include <net/if.h> 131 132 #include <netinet/in.h> 133 #include <netinet/in_systm.h> 134 #include <netinet/ip.h> 135 #include <netinet/in_pcb.h> 136 #include <netinet/ip_var.h> 137 #include <netinet/ip_icmp.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 #include <netinet6/in6_var.h> 147 #include <netinet6/ip6protosw.h> 148 #include <netinet/icmp6.h> 149 #include <netinet6/nd6.h> 150 #endif 151 152 #include <netinet/tcp.h> 153 #include <netinet/tcp_fsm.h> 154 #include <netinet/tcp_seq.h> 155 #include <netinet/tcp_timer.h> 156 #include <netinet/tcp_var.h> 157 #include <netinet/tcpip.h> 158 159 #ifdef IPSEC 160 #include <netinet6/ipsec.h> 161 #endif /*IPSEC*/ 162 163 #ifdef INET6 164 struct in6pcb tcb6; 165 #endif 166 167 struct inpcbtable tcbtable; /* head of queue of active tcpcb's */ 168 struct tcpstat tcpstat; /* tcp statistics */ 169 u_int32_t tcp_now; /* for RFC 1323 timestamps */ 170 171 /* patchable/settable parameters for tcp */ 172 int tcp_mssdflt = TCP_MSS; 173 int tcp_rttdflt = TCPTV_SRTTDFLT / PR_SLOWHZ; 174 int tcp_do_rfc1323 = 1; /* window scaling / timestamps (obsolete) */ 175 #if NRND > 0 176 int tcp_do_rfc1948 = 0; /* ISS by cryptographic hash */ 177 #endif 178 int tcp_do_sack = 1; /* selective acknowledgement */ 179 int tcp_do_win_scale = 1; /* RFC1323 window scaling */ 180 int tcp_do_timestamps = 1; /* RFC1323 timestamps */ 181 int tcp_do_newreno = 0; /* Use the New Reno algorithms */ 182 int tcp_ack_on_push = 0; /* set to enable immediate ACK-on-PUSH */ 183 int tcp_init_win = 1; 184 int tcp_mss_ifmtu = 0; 185 #ifdef TCP_COMPAT_42 186 int tcp_compat_42 = 1; 187 #else 188 int tcp_compat_42 = 0; 189 #endif 190 int tcp_rst_ppslim = 100; /* 100pps */ 191 192 /* tcb hash */ 193 #ifndef TCBHASHSIZE 194 #define TCBHASHSIZE 128 195 #endif 196 int tcbhashsize = TCBHASHSIZE; 197 198 /* syn hash parameters */ 199 #define TCP_SYN_HASH_SIZE 293 200 #define TCP_SYN_BUCKET_SIZE 35 201 int tcp_syn_cache_size = TCP_SYN_HASH_SIZE; 202 int tcp_syn_cache_limit = TCP_SYN_HASH_SIZE*TCP_SYN_BUCKET_SIZE; 203 int tcp_syn_bucket_limit = 3*TCP_SYN_BUCKET_SIZE; 204 struct syn_cache_head tcp_syn_cache[TCP_SYN_HASH_SIZE]; 205 206 int tcp_freeq __P((struct tcpcb *)); 207 208 #ifdef INET 209 void tcp_mtudisc_callback __P((struct in_addr)); 210 #endif 211 #ifdef INET6 212 void tcp6_mtudisc_callback __P((struct in6_addr *)); 213 #endif 214 215 void tcp_mtudisc __P((struct inpcb *, int)); 216 #ifdef INET6 217 void tcp6_mtudisc __P((struct in6pcb *, int)); 218 #endif 219 220 struct pool tcpcb_pool; 221 222 #ifdef TCP_CSUM_COUNTERS 223 #include <sys/device.h> 224 225 struct evcnt tcp_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 226 NULL, "tcp", "hwcsum bad"); 227 struct evcnt tcp_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 228 NULL, "tcp", "hwcsum ok"); 229 struct evcnt tcp_hwcsum_data = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 230 NULL, "tcp", "hwcsum data"); 231 struct evcnt tcp_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 232 NULL, "tcp", "swcsum"); 233 #endif /* TCP_CSUM_COUNTERS */ 234 235 #ifdef TCP_OUTPUT_COUNTERS 236 #include <sys/device.h> 237 238 struct evcnt tcp_output_bigheader = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 239 NULL, "tcp", "output big header"); 240 struct evcnt tcp_output_copysmall = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 241 NULL, "tcp", "output copy small"); 242 struct evcnt tcp_output_copybig = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 243 NULL, "tcp", "output copy big"); 244 struct evcnt tcp_output_refbig = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 245 NULL, "tcp", "output reference big"); 246 #endif /* TCP_OUTPUT_COUNTERS */ 247 248 #ifdef TCP_REASS_COUNTERS 249 #include <sys/device.h> 250 251 struct evcnt tcp_reass_ = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 252 NULL, "tcp_reass", "calls"); 253 struct evcnt tcp_reass_empty = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 254 &tcp_reass_, "tcp_reass", "insert into empty queue"); 255 struct evcnt tcp_reass_iteration[8] = { 256 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", ">7 iterations"), 257 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "1 iteration"), 258 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "2 iterations"), 259 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "3 iterations"), 260 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "4 iterations"), 261 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "5 iterations"), 262 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "6 iterations"), 263 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "7 iterations"), 264 }; 265 struct evcnt tcp_reass_prependfirst = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 266 &tcp_reass_, "tcp_reass", "prepend to first"); 267 struct evcnt tcp_reass_prepend = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 268 &tcp_reass_, "tcp_reass", "prepend"); 269 struct evcnt tcp_reass_insert = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 270 &tcp_reass_, "tcp_reass", "insert"); 271 struct evcnt tcp_reass_inserttail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 272 &tcp_reass_, "tcp_reass", "insert at tail"); 273 struct evcnt tcp_reass_append = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 274 &tcp_reass_, "tcp_reass", "append"); 275 struct evcnt tcp_reass_appendtail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 276 &tcp_reass_, "tcp_reass", "append to tail fragment"); 277 struct evcnt tcp_reass_overlaptail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 278 &tcp_reass_, "tcp_reass", "overlap at end"); 279 struct evcnt tcp_reass_overlapfront = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 280 &tcp_reass_, "tcp_reass", "overlap at start"); 281 struct evcnt tcp_reass_segdup = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 282 &tcp_reass_, "tcp_reass", "duplicate segment"); 283 struct evcnt tcp_reass_fragdup = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 284 &tcp_reass_, "tcp_reass", "duplicate fragment"); 285 286 #endif /* TCP_REASS_COUNTERS */ 287 288 /* 289 * Tcp initialization 290 */ 291 void 292 tcp_init() 293 { 294 int hlen; 295 296 pool_init(&tcpcb_pool, sizeof(struct tcpcb), 0, 0, 0, "tcpcbpl", 297 NULL); 298 in_pcbinit(&tcbtable, tcbhashsize, tcbhashsize); 299 #ifdef INET6 300 tcb6.in6p_next = tcb6.in6p_prev = &tcb6; 301 #endif 302 303 hlen = sizeof(struct ip) + sizeof(struct tcphdr); 304 #ifdef INET6 305 if (sizeof(struct ip) < sizeof(struct ip6_hdr)) 306 hlen = sizeof(struct ip6_hdr) + sizeof(struct tcphdr); 307 #endif 308 if (max_protohdr < hlen) 309 max_protohdr = hlen; 310 if (max_linkhdr + hlen > MHLEN) 311 panic("tcp_init"); 312 313 #ifdef INET 314 icmp_mtudisc_callback_register(tcp_mtudisc_callback); 315 #endif 316 #ifdef INET6 317 icmp6_mtudisc_callback_register(tcp6_mtudisc_callback); 318 #endif 319 320 /* Initialize timer state. */ 321 tcp_timer_init(); 322 323 /* Initialize the compressed state engine. */ 324 syn_cache_init(); 325 326 #ifdef TCP_CSUM_COUNTERS 327 evcnt_attach_static(&tcp_hwcsum_bad); 328 evcnt_attach_static(&tcp_hwcsum_ok); 329 evcnt_attach_static(&tcp_hwcsum_data); 330 evcnt_attach_static(&tcp_swcsum); 331 #endif /* TCP_CSUM_COUNTERS */ 332 333 #ifdef TCP_OUTPUT_COUNTERS 334 evcnt_attach_static(&tcp_output_bigheader); 335 evcnt_attach_static(&tcp_output_copysmall); 336 evcnt_attach_static(&tcp_output_copybig); 337 evcnt_attach_static(&tcp_output_refbig); 338 #endif /* TCP_OUTPUT_COUNTERS */ 339 340 #ifdef TCP_REASS_COUNTERS 341 evcnt_attach_static(&tcp_reass_); 342 evcnt_attach_static(&tcp_reass_empty); 343 evcnt_attach_static(&tcp_reass_iteration[0]); 344 evcnt_attach_static(&tcp_reass_iteration[1]); 345 evcnt_attach_static(&tcp_reass_iteration[2]); 346 evcnt_attach_static(&tcp_reass_iteration[3]); 347 evcnt_attach_static(&tcp_reass_iteration[4]); 348 evcnt_attach_static(&tcp_reass_iteration[5]); 349 evcnt_attach_static(&tcp_reass_iteration[6]); 350 evcnt_attach_static(&tcp_reass_iteration[7]); 351 evcnt_attach_static(&tcp_reass_prependfirst); 352 evcnt_attach_static(&tcp_reass_prepend); 353 evcnt_attach_static(&tcp_reass_insert); 354 evcnt_attach_static(&tcp_reass_inserttail); 355 evcnt_attach_static(&tcp_reass_append); 356 evcnt_attach_static(&tcp_reass_appendtail); 357 evcnt_attach_static(&tcp_reass_overlaptail); 358 evcnt_attach_static(&tcp_reass_overlapfront); 359 evcnt_attach_static(&tcp_reass_segdup); 360 evcnt_attach_static(&tcp_reass_fragdup); 361 #endif /* TCP_REASS_COUNTERS */ 362 } 363 364 /* 365 * Create template to be used to send tcp packets on a connection. 366 * Call after host entry created, allocates an mbuf and fills 367 * in a skeletal tcp/ip header, minimizing the amount of work 368 * necessary when the connection is used. 369 */ 370 struct mbuf * 371 tcp_template(tp) 372 struct tcpcb *tp; 373 { 374 struct inpcb *inp = tp->t_inpcb; 375 #ifdef INET6 376 struct in6pcb *in6p = tp->t_in6pcb; 377 #endif 378 struct tcphdr *n; 379 struct mbuf *m; 380 int hlen; 381 382 switch (tp->t_family) { 383 case AF_INET: 384 hlen = sizeof(struct ip); 385 if (inp) 386 break; 387 #ifdef INET6 388 if (in6p) { 389 /* mapped addr case */ 390 if (IN6_IS_ADDR_V4MAPPED(&in6p->in6p_laddr) 391 && IN6_IS_ADDR_V4MAPPED(&in6p->in6p_faddr)) 392 break; 393 } 394 #endif 395 return NULL; /*EINVAL*/ 396 #ifdef INET6 397 case AF_INET6: 398 hlen = sizeof(struct ip6_hdr); 399 if (in6p) { 400 /* more sainty check? */ 401 break; 402 } 403 return NULL; /*EINVAL*/ 404 #endif 405 default: 406 hlen = 0; /*pacify gcc*/ 407 return NULL; /*EAFNOSUPPORT*/ 408 } 409 #ifdef DIAGNOSTIC 410 if (hlen + sizeof(struct tcphdr) > MCLBYTES) 411 panic("mclbytes too small for t_template"); 412 #endif 413 m = tp->t_template; 414 if (m && m->m_len == hlen + sizeof(struct tcphdr)) 415 ; 416 else { 417 if (m) 418 m_freem(m); 419 m = tp->t_template = NULL; 420 MGETHDR(m, M_DONTWAIT, MT_HEADER); 421 if (m && hlen + sizeof(struct tcphdr) > MHLEN) { 422 MCLGET(m, M_DONTWAIT); 423 if ((m->m_flags & M_EXT) == 0) { 424 m_free(m); 425 m = NULL; 426 } 427 } 428 if (m == NULL) 429 return NULL; 430 m->m_pkthdr.len = m->m_len = hlen + sizeof(struct tcphdr); 431 } 432 433 bzero(mtod(m, caddr_t), m->m_len); 434 435 n = (struct tcphdr *)(mtod(m, caddr_t) + hlen); 436 437 switch (tp->t_family) { 438 case AF_INET: 439 { 440 struct ipovly *ipov; 441 mtod(m, struct ip *)->ip_v = 4; 442 ipov = mtod(m, struct ipovly *); 443 ipov->ih_pr = IPPROTO_TCP; 444 ipov->ih_len = htons(sizeof(struct tcphdr)); 445 if (inp) { 446 ipov->ih_src = inp->inp_laddr; 447 ipov->ih_dst = inp->inp_faddr; 448 } 449 #ifdef INET6 450 else if (in6p) { 451 /* mapped addr case */ 452 bcopy(&in6p->in6p_laddr.s6_addr32[3], &ipov->ih_src, 453 sizeof(ipov->ih_src)); 454 bcopy(&in6p->in6p_faddr.s6_addr32[3], &ipov->ih_dst, 455 sizeof(ipov->ih_dst)); 456 } 457 #endif 458 /* 459 * Compute the pseudo-header portion of the checksum 460 * now. We incrementally add in the TCP option and 461 * payload lengths later, and then compute the TCP 462 * checksum right before the packet is sent off onto 463 * the wire. 464 */ 465 n->th_sum = in_cksum_phdr(ipov->ih_src.s_addr, 466 ipov->ih_dst.s_addr, 467 htons(sizeof(struct tcphdr) + IPPROTO_TCP)); 468 break; 469 } 470 #ifdef INET6 471 case AF_INET6: 472 { 473 struct ip6_hdr *ip6; 474 mtod(m, struct ip *)->ip_v = 6; 475 ip6 = mtod(m, struct ip6_hdr *); 476 ip6->ip6_nxt = IPPROTO_TCP; 477 ip6->ip6_plen = htons(sizeof(struct tcphdr)); 478 ip6->ip6_src = in6p->in6p_laddr; 479 ip6->ip6_dst = in6p->in6p_faddr; 480 ip6->ip6_flow = in6p->in6p_flowinfo & IPV6_FLOWINFO_MASK; 481 if (ip6_auto_flowlabel) { 482 ip6->ip6_flow &= ~IPV6_FLOWLABEL_MASK; 483 ip6->ip6_flow |= 484 (htonl(ip6_flow_seq++) & IPV6_FLOWLABEL_MASK); 485 } 486 ip6->ip6_vfc &= ~IPV6_VERSION_MASK; 487 ip6->ip6_vfc |= IPV6_VERSION; 488 489 /* 490 * Compute the pseudo-header portion of the checksum 491 * now. We incrementally add in the TCP option and 492 * payload lengths later, and then compute the TCP 493 * checksum right before the packet is sent off onto 494 * the wire. 495 */ 496 n->th_sum = in6_cksum_phdr(&in6p->in6p_laddr, 497 &in6p->in6p_faddr, htonl(sizeof(struct tcphdr)), 498 htonl(IPPROTO_TCP)); 499 break; 500 } 501 #endif 502 } 503 if (inp) { 504 n->th_sport = inp->inp_lport; 505 n->th_dport = inp->inp_fport; 506 } 507 #ifdef INET6 508 else if (in6p) { 509 n->th_sport = in6p->in6p_lport; 510 n->th_dport = in6p->in6p_fport; 511 } 512 #endif 513 n->th_seq = 0; 514 n->th_ack = 0; 515 n->th_x2 = 0; 516 n->th_off = 5; 517 n->th_flags = 0; 518 n->th_win = 0; 519 n->th_urp = 0; 520 return (m); 521 } 522 523 /* 524 * Send a single message to the TCP at address specified by 525 * the given TCP/IP header. If m == 0, then we make a copy 526 * of the tcpiphdr at ti and send directly to the addressed host. 527 * This is used to force keep alive messages out using the TCP 528 * template for a connection tp->t_template. If flags are given 529 * then we send a message back to the TCP which originated the 530 * segment ti, and discard the mbuf containing it and any other 531 * attached mbufs. 532 * 533 * In any case the ack and sequence number of the transmitted 534 * segment are as specified by the parameters. 535 */ 536 int 537 tcp_respond(tp, template, m, th0, ack, seq, flags) 538 struct tcpcb *tp; 539 struct mbuf *template; 540 struct mbuf *m; 541 struct tcphdr *th0; 542 tcp_seq ack, seq; 543 int flags; 544 { 545 struct route *ro; 546 int error, tlen, win = 0; 547 int hlen; 548 struct ip *ip; 549 #ifdef INET6 550 struct ip6_hdr *ip6; 551 #endif 552 int family; /* family on packet, not inpcb/in6pcb! */ 553 struct tcphdr *th; 554 555 if (tp != NULL && (flags & TH_RST) == 0) { 556 #ifdef DIAGNOSTIC 557 if (tp->t_inpcb && tp->t_in6pcb) 558 panic("tcp_respond: both t_inpcb and t_in6pcb are set"); 559 #endif 560 #ifdef INET 561 if (tp->t_inpcb) 562 win = sbspace(&tp->t_inpcb->inp_socket->so_rcv); 563 #endif 564 #ifdef INET6 565 if (tp->t_in6pcb) 566 win = sbspace(&tp->t_in6pcb->in6p_socket->so_rcv); 567 #endif 568 } 569 570 ip = NULL; 571 #ifdef INET6 572 ip6 = NULL; 573 #endif 574 if (m == 0) { 575 if (!template) 576 return EINVAL; 577 578 /* get family information from template */ 579 switch (mtod(template, struct ip *)->ip_v) { 580 case 4: 581 family = AF_INET; 582 hlen = sizeof(struct ip); 583 break; 584 #ifdef INET6 585 case 6: 586 family = AF_INET6; 587 hlen = sizeof(struct ip6_hdr); 588 break; 589 #endif 590 default: 591 return EAFNOSUPPORT; 592 } 593 594 MGETHDR(m, M_DONTWAIT, MT_HEADER); 595 if (m) { 596 MCLGET(m, M_DONTWAIT); 597 if ((m->m_flags & M_EXT) == 0) { 598 m_free(m); 599 m = NULL; 600 } 601 } 602 if (m == NULL) 603 return (ENOBUFS); 604 605 if (tcp_compat_42) 606 tlen = 1; 607 else 608 tlen = 0; 609 610 m->m_data += max_linkhdr; 611 bcopy(mtod(template, caddr_t), mtod(m, caddr_t), 612 template->m_len); 613 switch (family) { 614 case AF_INET: 615 ip = mtod(m, struct ip *); 616 th = (struct tcphdr *)(ip + 1); 617 break; 618 #ifdef INET6 619 case AF_INET6: 620 ip6 = mtod(m, struct ip6_hdr *); 621 th = (struct tcphdr *)(ip6 + 1); 622 break; 623 #endif 624 #if 0 625 default: 626 /* noone will visit here */ 627 m_freem(m); 628 return EAFNOSUPPORT; 629 #endif 630 } 631 flags = TH_ACK; 632 } else { 633 634 if ((m->m_flags & M_PKTHDR) == 0) { 635 #if 0 636 printf("non PKTHDR to tcp_respond\n"); 637 #endif 638 m_freem(m); 639 return EINVAL; 640 } 641 #ifdef DIAGNOSTIC 642 if (!th0) 643 panic("th0 == NULL in tcp_respond"); 644 #endif 645 646 /* get family information from m */ 647 switch (mtod(m, struct ip *)->ip_v) { 648 case 4: 649 family = AF_INET; 650 hlen = sizeof(struct ip); 651 ip = mtod(m, struct ip *); 652 break; 653 #ifdef INET6 654 case 6: 655 family = AF_INET6; 656 hlen = sizeof(struct ip6_hdr); 657 ip6 = mtod(m, struct ip6_hdr *); 658 break; 659 #endif 660 default: 661 m_freem(m); 662 return EAFNOSUPPORT; 663 } 664 if ((flags & TH_SYN) == 0 || sizeof(*th0) > (th0->th_off << 2)) 665 tlen = sizeof(*th0); 666 else 667 tlen = th0->th_off << 2; 668 669 if (m->m_len > hlen + tlen && (m->m_flags & M_EXT) == 0 && 670 mtod(m, caddr_t) + hlen == (caddr_t)th0) { 671 m->m_len = hlen + tlen; 672 m_freem(m->m_next); 673 m->m_next = NULL; 674 } else { 675 struct mbuf *n; 676 677 #ifdef DIAGNOSTIC 678 if (max_linkhdr + hlen + tlen > MCLBYTES) { 679 m_freem(m); 680 return EMSGSIZE; 681 } 682 #endif 683 MGETHDR(n, M_DONTWAIT, MT_HEADER); 684 if (n && max_linkhdr + hlen + tlen > MHLEN) { 685 MCLGET(n, M_DONTWAIT); 686 if ((n->m_flags & M_EXT) == 0) { 687 m_freem(n); 688 n = NULL; 689 } 690 } 691 if (!n) { 692 m_freem(m); 693 return ENOBUFS; 694 } 695 696 n->m_data += max_linkhdr; 697 n->m_len = hlen + tlen; 698 m_copyback(n, 0, hlen, mtod(m, caddr_t)); 699 m_copyback(n, hlen, tlen, (caddr_t)th0); 700 701 m_freem(m); 702 m = n; 703 n = NULL; 704 } 705 706 #define xchg(a,b,type) { type t; t=a; a=b; b=t; } 707 switch (family) { 708 case AF_INET: 709 ip = mtod(m, struct ip *); 710 th = (struct tcphdr *)(ip + 1); 711 ip->ip_p = IPPROTO_TCP; 712 xchg(ip->ip_dst, ip->ip_src, struct in_addr); 713 ip->ip_p = IPPROTO_TCP; 714 break; 715 #ifdef INET6 716 case AF_INET6: 717 ip6 = mtod(m, struct ip6_hdr *); 718 th = (struct tcphdr *)(ip6 + 1); 719 ip6->ip6_nxt = IPPROTO_TCP; 720 xchg(ip6->ip6_dst, ip6->ip6_src, struct in6_addr); 721 ip6->ip6_nxt = IPPROTO_TCP; 722 break; 723 #endif 724 #if 0 725 default: 726 /* noone will visit here */ 727 m_freem(m); 728 return EAFNOSUPPORT; 729 #endif 730 } 731 xchg(th->th_dport, th->th_sport, u_int16_t); 732 #undef xchg 733 tlen = 0; /*be friendly with the following code*/ 734 } 735 th->th_seq = htonl(seq); 736 th->th_ack = htonl(ack); 737 th->th_x2 = 0; 738 if ((flags & TH_SYN) == 0) { 739 if (tp) 740 win >>= tp->rcv_scale; 741 if (win > TCP_MAXWIN) 742 win = TCP_MAXWIN; 743 th->th_win = htons((u_int16_t)win); 744 th->th_off = sizeof (struct tcphdr) >> 2; 745 tlen += sizeof(*th); 746 } else 747 tlen += th->th_off << 2; 748 m->m_len = hlen + tlen; 749 m->m_pkthdr.len = hlen + tlen; 750 m->m_pkthdr.rcvif = (struct ifnet *) 0; 751 th->th_flags = flags; 752 th->th_urp = 0; 753 754 switch (family) { 755 #ifdef INET 756 case AF_INET: 757 { 758 struct ipovly *ipov = (struct ipovly *)ip; 759 bzero(ipov->ih_x1, sizeof ipov->ih_x1); 760 ipov->ih_len = htons((u_int16_t)tlen); 761 762 th->th_sum = 0; 763 th->th_sum = in_cksum(m, hlen + tlen); 764 ip->ip_len = htons(hlen + tlen); 765 ip->ip_ttl = ip_defttl; 766 break; 767 } 768 #endif 769 #ifdef INET6 770 case AF_INET6: 771 { 772 th->th_sum = 0; 773 th->th_sum = in6_cksum(m, IPPROTO_TCP, sizeof(struct ip6_hdr), 774 tlen); 775 ip6->ip6_plen = ntohs(tlen); 776 if (tp && tp->t_in6pcb) { 777 struct ifnet *oifp; 778 ro = (struct route *)&tp->t_in6pcb->in6p_route; 779 oifp = ro->ro_rt ? ro->ro_rt->rt_ifp : NULL; 780 ip6->ip6_hlim = in6_selecthlim(tp->t_in6pcb, oifp); 781 } else 782 ip6->ip6_hlim = ip6_defhlim; 783 ip6->ip6_flow &= ~IPV6_FLOWINFO_MASK; 784 if (ip6_auto_flowlabel) { 785 ip6->ip6_flow |= 786 (htonl(ip6_flow_seq++) & IPV6_FLOWLABEL_MASK); 787 } 788 break; 789 } 790 #endif 791 } 792 793 #ifdef IPSEC 794 (void)ipsec_setsocket(m, NULL); 795 #endif /*IPSEC*/ 796 797 if (tp != NULL && tp->t_inpcb != NULL) { 798 ro = &tp->t_inpcb->inp_route; 799 #ifdef IPSEC 800 if (ipsec_setsocket(m, tp->t_inpcb->inp_socket) != 0) { 801 m_freem(m); 802 return ENOBUFS; 803 } 804 #endif 805 #ifdef DIAGNOSTIC 806 if (family != AF_INET) 807 panic("tcp_respond: address family mismatch"); 808 if (!in_hosteq(ip->ip_dst, tp->t_inpcb->inp_faddr)) { 809 panic("tcp_respond: ip_dst %x != inp_faddr %x", 810 ntohl(ip->ip_dst.s_addr), 811 ntohl(tp->t_inpcb->inp_faddr.s_addr)); 812 } 813 #endif 814 } 815 #ifdef INET6 816 else if (tp != NULL && tp->t_in6pcb != NULL) { 817 ro = (struct route *)&tp->t_in6pcb->in6p_route; 818 #ifdef IPSEC 819 if (ipsec_setsocket(m, tp->t_in6pcb->in6p_socket) != 0) { 820 m_freem(m); 821 return ENOBUFS; 822 } 823 #endif 824 #ifdef DIAGNOSTIC 825 if (family == AF_INET) { 826 if (!IN6_IS_ADDR_V4MAPPED(&tp->t_in6pcb->in6p_faddr)) 827 panic("tcp_respond: not mapped addr"); 828 if (bcmp(&ip->ip_dst, 829 &tp->t_in6pcb->in6p_faddr.s6_addr32[3], 830 sizeof(ip->ip_dst)) != 0) { 831 panic("tcp_respond: ip_dst != in6p_faddr"); 832 } 833 } else if (family == AF_INET6) { 834 if (!IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst, 835 &tp->t_in6pcb->in6p_faddr)) 836 panic("tcp_respond: ip6_dst != in6p_faddr"); 837 } else 838 panic("tcp_respond: address family mismatch"); 839 #endif 840 } 841 #endif 842 else 843 ro = NULL; 844 845 switch (family) { 846 #ifdef INET 847 case AF_INET: 848 error = ip_output(m, NULL, ro, 849 (tp && tp->t_mtudisc ? IP_MTUDISC : 0), 850 NULL); 851 break; 852 #endif 853 #ifdef INET6 854 case AF_INET6: 855 error = ip6_output(m, NULL, (struct route_in6 *)ro, 0, NULL, 856 NULL); 857 break; 858 #endif 859 default: 860 error = EAFNOSUPPORT; 861 break; 862 } 863 864 return (error); 865 } 866 867 /* 868 * Create a new TCP control block, making an 869 * empty reassembly queue and hooking it to the argument 870 * protocol control block. 871 */ 872 struct tcpcb * 873 tcp_newtcpcb(family, aux) 874 int family; /* selects inpcb, or in6pcb */ 875 void *aux; 876 { 877 struct tcpcb *tp; 878 int i; 879 880 switch (family) { 881 case PF_INET: 882 break; 883 #ifdef INET6 884 case PF_INET6: 885 break; 886 #endif 887 default: 888 return NULL; 889 } 890 891 tp = pool_get(&tcpcb_pool, PR_NOWAIT); 892 if (tp == NULL) 893 return (NULL); 894 bzero((caddr_t)tp, sizeof(struct tcpcb)); 895 TAILQ_INIT(&tp->segq); 896 TAILQ_INIT(&tp->timeq); 897 tp->t_family = family; /* may be overridden later on */ 898 tp->t_peermss = tcp_mssdflt; 899 tp->t_ourmss = tcp_mssdflt; 900 tp->t_segsz = tcp_mssdflt; 901 LIST_INIT(&tp->t_sc); 902 903 callout_init(&tp->t_delack_ch); 904 for (i = 0; i < TCPT_NTIMERS; i++) 905 TCP_TIMER_INIT(tp, i); 906 907 tp->t_flags = 0; 908 if (tcp_do_rfc1323 && tcp_do_win_scale) 909 tp->t_flags |= TF_REQ_SCALE; 910 if (tcp_do_rfc1323 && tcp_do_timestamps) 911 tp->t_flags |= TF_REQ_TSTMP; 912 if (tcp_do_sack == 2) 913 tp->t_flags |= TF_WILL_SACK; 914 else if (tcp_do_sack == 1) 915 tp->t_flags |= TF_WILL_SACK|TF_IGNR_RXSACK; 916 tp->t_flags |= TF_CANT_TXSACK; 917 switch (family) { 918 case PF_INET: 919 tp->t_inpcb = (struct inpcb *)aux; 920 tp->t_mtudisc = ip_mtudisc; 921 break; 922 #ifdef INET6 923 case PF_INET6: 924 tp->t_in6pcb = (struct in6pcb *)aux; 925 /* for IPv6, always try to run path MTU discovery */ 926 tp->t_mtudisc = 1; 927 break; 928 #endif 929 } 930 /* 931 * Init srtt to TCPTV_SRTTBASE (0), so we can tell that we have no 932 * rtt estimate. Set rttvar so that srtt + 2 * rttvar gives 933 * reasonable initial retransmit time. 934 */ 935 tp->t_srtt = TCPTV_SRTTBASE; 936 tp->t_rttvar = tcp_rttdflt * PR_SLOWHZ << (TCP_RTTVAR_SHIFT + 2 - 1); 937 tp->t_rttmin = TCPTV_MIN; 938 TCPT_RANGESET(tp->t_rxtcur, TCP_REXMTVAL(tp), 939 TCPTV_MIN, TCPTV_REXMTMAX); 940 tp->snd_cwnd = TCP_MAXWIN << TCP_MAX_WINSHIFT; 941 tp->snd_ssthresh = TCP_MAXWIN << TCP_MAX_WINSHIFT; 942 if (family == AF_INET) { 943 struct inpcb *inp = (struct inpcb *)aux; 944 inp->inp_ip.ip_ttl = ip_defttl; 945 inp->inp_ppcb = (caddr_t)tp; 946 } 947 #ifdef INET6 948 else if (family == AF_INET6) { 949 struct in6pcb *in6p = (struct in6pcb *)aux; 950 in6p->in6p_ip6.ip6_hlim = in6_selecthlim(in6p, 951 in6p->in6p_route.ro_rt ? in6p->in6p_route.ro_rt->rt_ifp 952 : NULL); 953 in6p->in6p_ppcb = (caddr_t)tp; 954 } 955 #endif 956 957 /* 958 * Initialize our timebase. When we send timestamps, we take 959 * the delta from tcp_now -- this means each connection always 960 * gets a timebase of 0, which makes it, among other things, 961 * more difficult to determine how long a system has been up, 962 * and thus how many TCP sequence increments have occurred. 963 */ 964 tp->ts_timebase = tcp_now; 965 966 return (tp); 967 } 968 969 /* 970 * Drop a TCP connection, reporting 971 * the specified error. If connection is synchronized, 972 * then send a RST to peer. 973 */ 974 struct tcpcb * 975 tcp_drop(tp, errno) 976 struct tcpcb *tp; 977 int errno; 978 { 979 struct socket *so = NULL; 980 981 #ifdef DIAGNOSTIC 982 if (tp->t_inpcb && tp->t_in6pcb) 983 panic("tcp_drop: both t_inpcb and t_in6pcb are set"); 984 #endif 985 #ifdef INET 986 if (tp->t_inpcb) 987 so = tp->t_inpcb->inp_socket; 988 #endif 989 #ifdef INET6 990 if (tp->t_in6pcb) 991 so = tp->t_in6pcb->in6p_socket; 992 #endif 993 if (!so) 994 return NULL; 995 996 if (TCPS_HAVERCVDSYN(tp->t_state)) { 997 tp->t_state = TCPS_CLOSED; 998 (void) tcp_output(tp); 999 tcpstat.tcps_drops++; 1000 } else 1001 tcpstat.tcps_conndrops++; 1002 if (errno == ETIMEDOUT && tp->t_softerror) 1003 errno = tp->t_softerror; 1004 so->so_error = errno; 1005 return (tcp_close(tp)); 1006 } 1007 1008 /* 1009 * Close a TCP control block: 1010 * discard all space held by the tcp 1011 * discard internet protocol block 1012 * wake up any sleepers 1013 */ 1014 struct tcpcb * 1015 tcp_close(tp) 1016 struct tcpcb *tp; 1017 { 1018 struct inpcb *inp; 1019 #ifdef INET6 1020 struct in6pcb *in6p; 1021 #endif 1022 struct socket *so; 1023 #ifdef RTV_RTT 1024 struct rtentry *rt; 1025 #endif 1026 struct route *ro; 1027 1028 inp = tp->t_inpcb; 1029 #ifdef INET6 1030 in6p = tp->t_in6pcb; 1031 #endif 1032 so = NULL; 1033 ro = NULL; 1034 if (inp) { 1035 so = inp->inp_socket; 1036 ro = &inp->inp_route; 1037 } 1038 #ifdef INET6 1039 else if (in6p) { 1040 so = in6p->in6p_socket; 1041 ro = (struct route *)&in6p->in6p_route; 1042 } 1043 #endif 1044 1045 #ifdef RTV_RTT 1046 /* 1047 * If we sent enough data to get some meaningful characteristics, 1048 * save them in the routing entry. 'Enough' is arbitrarily 1049 * defined as the sendpipesize (default 4K) * 16. This would 1050 * give us 16 rtt samples assuming we only get one sample per 1051 * window (the usual case on a long haul net). 16 samples is 1052 * enough for the srtt filter to converge to within 5% of the correct 1053 * value; fewer samples and we could save a very bogus rtt. 1054 * 1055 * Don't update the default route's characteristics and don't 1056 * update anything that the user "locked". 1057 */ 1058 if (SEQ_LT(tp->iss + so->so_snd.sb_hiwat * 16, tp->snd_max) && 1059 ro && (rt = ro->ro_rt) && 1060 !in_nullhost(satosin(rt_key(rt))->sin_addr)) { 1061 u_long i = 0; 1062 1063 if ((rt->rt_rmx.rmx_locks & RTV_RTT) == 0) { 1064 i = tp->t_srtt * 1065 ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTT_SHIFT + 2)); 1066 if (rt->rt_rmx.rmx_rtt && i) 1067 /* 1068 * filter this update to half the old & half 1069 * the new values, converting scale. 1070 * See route.h and tcp_var.h for a 1071 * description of the scaling constants. 1072 */ 1073 rt->rt_rmx.rmx_rtt = 1074 (rt->rt_rmx.rmx_rtt + i) / 2; 1075 else 1076 rt->rt_rmx.rmx_rtt = i; 1077 } 1078 if ((rt->rt_rmx.rmx_locks & RTV_RTTVAR) == 0) { 1079 i = tp->t_rttvar * 1080 ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTTVAR_SHIFT + 2)); 1081 if (rt->rt_rmx.rmx_rttvar && i) 1082 rt->rt_rmx.rmx_rttvar = 1083 (rt->rt_rmx.rmx_rttvar + i) / 2; 1084 else 1085 rt->rt_rmx.rmx_rttvar = i; 1086 } 1087 /* 1088 * update the pipelimit (ssthresh) if it has been updated 1089 * already or if a pipesize was specified & the threshhold 1090 * got below half the pipesize. I.e., wait for bad news 1091 * before we start updating, then update on both good 1092 * and bad news. 1093 */ 1094 if (((rt->rt_rmx.rmx_locks & RTV_SSTHRESH) == 0 && 1095 (i = tp->snd_ssthresh) && rt->rt_rmx.rmx_ssthresh) || 1096 i < (rt->rt_rmx.rmx_sendpipe / 2)) { 1097 /* 1098 * convert the limit from user data bytes to 1099 * packets then to packet data bytes. 1100 */ 1101 i = (i + tp->t_segsz / 2) / tp->t_segsz; 1102 if (i < 2) 1103 i = 2; 1104 i *= (u_long)(tp->t_segsz + sizeof (struct tcpiphdr)); 1105 if (rt->rt_rmx.rmx_ssthresh) 1106 rt->rt_rmx.rmx_ssthresh = 1107 (rt->rt_rmx.rmx_ssthresh + i) / 2; 1108 else 1109 rt->rt_rmx.rmx_ssthresh = i; 1110 } 1111 } 1112 #endif /* RTV_RTT */ 1113 /* free the reassembly queue, if any */ 1114 TCP_REASS_LOCK(tp); 1115 (void) tcp_freeq(tp); 1116 TCP_REASS_UNLOCK(tp); 1117 1118 tcp_canceltimers(tp); 1119 TCP_CLEAR_DELACK(tp); 1120 syn_cache_cleanup(tp); 1121 1122 if (tp->t_template) { 1123 m_free(tp->t_template); 1124 tp->t_template = NULL; 1125 } 1126 pool_put(&tcpcb_pool, tp); 1127 if (inp) { 1128 inp->inp_ppcb = 0; 1129 soisdisconnected(so); 1130 in_pcbdetach(inp); 1131 } 1132 #ifdef INET6 1133 else if (in6p) { 1134 in6p->in6p_ppcb = 0; 1135 soisdisconnected(so); 1136 in6_pcbdetach(in6p); 1137 } 1138 #endif 1139 tcpstat.tcps_closed++; 1140 return ((struct tcpcb *)0); 1141 } 1142 1143 int 1144 tcp_freeq(tp) 1145 struct tcpcb *tp; 1146 { 1147 struct ipqent *qe; 1148 int rv = 0; 1149 #ifdef TCPREASS_DEBUG 1150 int i = 0; 1151 #endif 1152 1153 TCP_REASS_LOCK_CHECK(tp); 1154 1155 while ((qe = TAILQ_FIRST(&tp->segq)) != NULL) { 1156 #ifdef TCPREASS_DEBUG 1157 printf("tcp_freeq[%p,%d]: %u:%u(%u) 0x%02x\n", 1158 tp, i++, qe->ipqe_seq, qe->ipqe_seq + qe->ipqe_len, 1159 qe->ipqe_len, qe->ipqe_flags & (TH_SYN|TH_FIN|TH_RST)); 1160 #endif 1161 TAILQ_REMOVE(&tp->segq, qe, ipqe_q); 1162 TAILQ_REMOVE(&tp->timeq, qe, ipqe_timeq); 1163 m_freem(qe->ipqe_m); 1164 pool_put(&ipqent_pool, qe); 1165 rv = 1; 1166 } 1167 return (rv); 1168 } 1169 1170 /* 1171 * Protocol drain routine. Called when memory is in short supply. 1172 */ 1173 void 1174 tcp_drain() 1175 { 1176 struct inpcb *inp; 1177 struct tcpcb *tp; 1178 1179 /* 1180 * Free the sequence queue of all TCP connections. 1181 */ 1182 inp = CIRCLEQ_FIRST(&tcbtable.inpt_queue); 1183 if (inp) /* XXX */ 1184 CIRCLEQ_FOREACH(inp, &tcbtable.inpt_queue, inp_queue) { 1185 if ((tp = intotcpcb(inp)) != NULL) { 1186 /* 1187 * We may be called from a device's interrupt 1188 * context. If the tcpcb is already busy, 1189 * just bail out now. 1190 */ 1191 if (tcp_reass_lock_try(tp) == 0) 1192 continue; 1193 if (tcp_freeq(tp)) 1194 tcpstat.tcps_connsdrained++; 1195 TCP_REASS_UNLOCK(tp); 1196 } 1197 } 1198 } 1199 1200 #ifdef INET6 1201 void 1202 tcp6_drain() 1203 { 1204 struct in6pcb *in6p; 1205 struct tcpcb *tp; 1206 struct in6pcb *head = &tcb6; 1207 1208 /* 1209 * Free the sequence queue of all TCP connections. 1210 */ 1211 for (in6p = head->in6p_next; in6p != head; in6p = in6p->in6p_next) { 1212 if ((tp = in6totcpcb(in6p)) != NULL) { 1213 /* 1214 * We may be called from a device's interrupt 1215 * context. If the tcpcb is already busy, 1216 * just bail out now. 1217 */ 1218 if (tcp_reass_lock_try(tp) == 0) 1219 continue; 1220 if (tcp_freeq(tp)) 1221 tcpstat.tcps_connsdrained++; 1222 TCP_REASS_UNLOCK(tp); 1223 } 1224 } 1225 } 1226 #endif 1227 1228 /* 1229 * Notify a tcp user of an asynchronous error; 1230 * store error as soft error, but wake up user 1231 * (for now, won't do anything until can select for soft error). 1232 */ 1233 void 1234 tcp_notify(inp, error) 1235 struct inpcb *inp; 1236 int error; 1237 { 1238 struct tcpcb *tp = (struct tcpcb *)inp->inp_ppcb; 1239 struct socket *so = inp->inp_socket; 1240 1241 /* 1242 * Ignore some errors if we are hooked up. 1243 * If connection hasn't completed, has retransmitted several times, 1244 * and receives a second error, give up now. This is better 1245 * than waiting a long time to establish a connection that 1246 * can never complete. 1247 */ 1248 if (tp->t_state == TCPS_ESTABLISHED && 1249 (error == EHOSTUNREACH || error == ENETUNREACH || 1250 error == EHOSTDOWN)) { 1251 return; 1252 } else if (TCPS_HAVEESTABLISHED(tp->t_state) == 0 && 1253 tp->t_rxtshift > 3 && tp->t_softerror) 1254 so->so_error = error; 1255 else 1256 tp->t_softerror = error; 1257 wakeup((caddr_t) &so->so_timeo); 1258 sorwakeup(so); 1259 sowwakeup(so); 1260 } 1261 1262 #ifdef INET6 1263 void 1264 tcp6_notify(in6p, error) 1265 struct in6pcb *in6p; 1266 int error; 1267 { 1268 struct tcpcb *tp = (struct tcpcb *)in6p->in6p_ppcb; 1269 struct socket *so = in6p->in6p_socket; 1270 1271 /* 1272 * Ignore some errors if we are hooked up. 1273 * If connection hasn't completed, has retransmitted several times, 1274 * and receives a second error, give up now. This is better 1275 * than waiting a long time to establish a connection that 1276 * can never complete. 1277 */ 1278 if (tp->t_state == TCPS_ESTABLISHED && 1279 (error == EHOSTUNREACH || error == ENETUNREACH || 1280 error == EHOSTDOWN)) { 1281 return; 1282 } else if (TCPS_HAVEESTABLISHED(tp->t_state) == 0 && 1283 tp->t_rxtshift > 3 && tp->t_softerror) 1284 so->so_error = error; 1285 else 1286 tp->t_softerror = error; 1287 wakeup((caddr_t) &so->so_timeo); 1288 sorwakeup(so); 1289 sowwakeup(so); 1290 } 1291 #endif 1292 1293 #ifdef INET6 1294 void 1295 tcp6_ctlinput(cmd, sa, d) 1296 int cmd; 1297 struct sockaddr *sa; 1298 void *d; 1299 { 1300 struct tcphdr th; 1301 void (*notify) __P((struct in6pcb *, int)) = tcp6_notify; 1302 int nmatch; 1303 struct ip6_hdr *ip6; 1304 const struct sockaddr_in6 *sa6_src = NULL; 1305 struct sockaddr_in6 *sa6 = (struct sockaddr_in6 *)sa; 1306 struct mbuf *m; 1307 int off; 1308 1309 if (sa->sa_family != AF_INET6 || 1310 sa->sa_len != sizeof(struct sockaddr_in6)) 1311 return; 1312 if ((unsigned)cmd >= PRC_NCMDS) 1313 return; 1314 else if (cmd == PRC_QUENCH) { 1315 /* XXX there's no PRC_QUENCH in IPv6 */ 1316 notify = tcp6_quench; 1317 } else if (PRC_IS_REDIRECT(cmd)) 1318 notify = in6_rtchange, d = NULL; 1319 else if (cmd == PRC_MSGSIZE) 1320 ; /* special code is present, see below */ 1321 else if (cmd == PRC_HOSTDEAD) 1322 d = NULL; 1323 else if (inet6ctlerrmap[cmd] == 0) 1324 return; 1325 1326 /* if the parameter is from icmp6, decode it. */ 1327 if (d != NULL) { 1328 struct ip6ctlparam *ip6cp = (struct ip6ctlparam *)d; 1329 m = ip6cp->ip6c_m; 1330 ip6 = ip6cp->ip6c_ip6; 1331 off = ip6cp->ip6c_off; 1332 sa6_src = ip6cp->ip6c_src; 1333 } else { 1334 m = NULL; 1335 ip6 = NULL; 1336 sa6_src = &sa6_any; 1337 } 1338 1339 if (ip6) { 1340 /* 1341 * XXX: We assume that when ip6 is non NULL, 1342 * M and OFF are valid. 1343 */ 1344 1345 /* check if we can safely examine src and dst ports */ 1346 if (m->m_pkthdr.len < off + sizeof(th)) { 1347 if (cmd == PRC_MSGSIZE) 1348 icmp6_mtudisc_update((struct ip6ctlparam *)d, 0); 1349 return; 1350 } 1351 1352 bzero(&th, sizeof(th)); 1353 m_copydata(m, off, sizeof(th), (caddr_t)&th); 1354 1355 if (cmd == PRC_MSGSIZE) { 1356 int valid = 0; 1357 1358 /* 1359 * Check to see if we have a valid TCP connection 1360 * corresponding to the address in the ICMPv6 message 1361 * payload. 1362 */ 1363 if (in6_pcblookup_connect(&tcb6, &sa6->sin6_addr, 1364 th.th_dport, (struct in6_addr *)&sa6_src->sin6_addr, 1365 th.th_sport, 0)) 1366 valid++; 1367 1368 /* 1369 * Depending on the value of "valid" and routing table 1370 * size (mtudisc_{hi,lo}wat), we will: 1371 * - recalcurate the new MTU and create the 1372 * corresponding routing entry, or 1373 * - ignore the MTU change notification. 1374 */ 1375 icmp6_mtudisc_update((struct ip6ctlparam *)d, valid); 1376 1377 /* 1378 * no need to call in6_pcbnotify, it should have been 1379 * called via callback if necessary 1380 */ 1381 return; 1382 } 1383 1384 nmatch = in6_pcbnotify(&tcb6, sa, th.th_dport, 1385 (struct sockaddr *)sa6_src, th.th_sport, cmd, NULL, notify); 1386 if (nmatch == 0 && syn_cache_count && 1387 (inet6ctlerrmap[cmd] == EHOSTUNREACH || 1388 inet6ctlerrmap[cmd] == ENETUNREACH || 1389 inet6ctlerrmap[cmd] == EHOSTDOWN)) 1390 syn_cache_unreach((struct sockaddr *)sa6_src, 1391 sa, &th); 1392 } else { 1393 (void) in6_pcbnotify(&tcb6, sa, 0, (struct sockaddr *)sa6_src, 1394 0, cmd, NULL, notify); 1395 } 1396 } 1397 #endif 1398 1399 #ifdef INET 1400 /* assumes that ip header and tcp header are contiguous on mbuf */ 1401 void * 1402 tcp_ctlinput(cmd, sa, v) 1403 int cmd; 1404 struct sockaddr *sa; 1405 void *v; 1406 { 1407 struct ip *ip = v; 1408 struct tcphdr *th; 1409 struct icmp *icp; 1410 extern const int inetctlerrmap[]; 1411 void (*notify) __P((struct inpcb *, int)) = tcp_notify; 1412 int errno; 1413 int nmatch; 1414 #ifdef INET6 1415 struct in6_addr src6, dst6; 1416 #endif 1417 1418 if (sa->sa_family != AF_INET || 1419 sa->sa_len != sizeof(struct sockaddr_in)) 1420 return NULL; 1421 if ((unsigned)cmd >= PRC_NCMDS) 1422 return NULL; 1423 errno = inetctlerrmap[cmd]; 1424 if (cmd == PRC_QUENCH) 1425 notify = tcp_quench; 1426 else if (PRC_IS_REDIRECT(cmd)) 1427 notify = in_rtchange, ip = 0; 1428 else if (cmd == PRC_MSGSIZE && ip && ip->ip_v == 4) { 1429 /* 1430 * Check to see if we have a valid TCP connection 1431 * corresponding to the address in the ICMP message 1432 * payload. 1433 * 1434 * Boundary check is made in icmp_input(), with ICMP_ADVLENMIN. 1435 */ 1436 th = (struct tcphdr *)((caddr_t)ip + (ip->ip_hl << 2)); 1437 #ifdef INET6 1438 memset(&src6, 0, sizeof(src6)); 1439 memset(&dst6, 0, sizeof(dst6)); 1440 src6.s6_addr16[5] = dst6.s6_addr16[5] = 0xffff; 1441 memcpy(&src6.s6_addr32[3], &ip->ip_src, sizeof(struct in_addr)); 1442 memcpy(&dst6.s6_addr32[3], &ip->ip_dst, sizeof(struct in_addr)); 1443 #endif 1444 if (in_pcblookup_connect(&tcbtable, ip->ip_dst, th->th_dport, 1445 ip->ip_src, th->th_sport) != NULL) 1446 ; 1447 #ifdef INET6 1448 else if (in6_pcblookup_connect(&tcb6, &dst6, 1449 th->th_dport, &src6, th->th_sport, 0) != NULL) 1450 ; 1451 #endif 1452 else 1453 return NULL; 1454 1455 /* 1456 * Now that we've validated that we are actually communicating 1457 * with the host indicated in the ICMP message, locate the 1458 * ICMP header, recalculate the new MTU, and create the 1459 * corresponding routing entry. 1460 */ 1461 icp = (struct icmp *)((caddr_t)ip - 1462 offsetof(struct icmp, icmp_ip)); 1463 icmp_mtudisc(icp, ip->ip_dst); 1464 1465 return NULL; 1466 } else if (cmd == PRC_HOSTDEAD) 1467 ip = 0; 1468 else if (errno == 0) 1469 return NULL; 1470 if (ip && ip->ip_v == 4 && sa->sa_family == AF_INET) { 1471 th = (struct tcphdr *)((caddr_t)ip + (ip->ip_hl << 2)); 1472 nmatch = in_pcbnotify(&tcbtable, satosin(sa)->sin_addr, 1473 th->th_dport, ip->ip_src, th->th_sport, errno, notify); 1474 if (nmatch == 0 && syn_cache_count && 1475 (inetctlerrmap[cmd] == EHOSTUNREACH || 1476 inetctlerrmap[cmd] == ENETUNREACH || 1477 inetctlerrmap[cmd] == EHOSTDOWN)) { 1478 struct sockaddr_in sin; 1479 bzero(&sin, sizeof(sin)); 1480 sin.sin_len = sizeof(sin); 1481 sin.sin_family = AF_INET; 1482 sin.sin_port = th->th_sport; 1483 sin.sin_addr = ip->ip_src; 1484 syn_cache_unreach((struct sockaddr *)&sin, sa, th); 1485 } 1486 1487 /* XXX mapped address case */ 1488 } else 1489 in_pcbnotifyall(&tcbtable, satosin(sa)->sin_addr, errno, 1490 notify); 1491 return NULL; 1492 } 1493 1494 /* 1495 * When a source quence is received, we are being notifed of congestion. 1496 * Close the congestion window down to the Loss Window (one segment). 1497 * We will gradually open it again as we proceed. 1498 */ 1499 void 1500 tcp_quench(inp, errno) 1501 struct inpcb *inp; 1502 int errno; 1503 { 1504 struct tcpcb *tp = intotcpcb(inp); 1505 1506 if (tp) 1507 tp->snd_cwnd = tp->t_segsz; 1508 } 1509 #endif 1510 1511 #ifdef INET6 1512 void 1513 tcp6_quench(in6p, errno) 1514 struct in6pcb *in6p; 1515 int errno; 1516 { 1517 struct tcpcb *tp = in6totcpcb(in6p); 1518 1519 if (tp) 1520 tp->snd_cwnd = tp->t_segsz; 1521 } 1522 #endif 1523 1524 #ifdef INET 1525 /* 1526 * Path MTU Discovery handlers. 1527 */ 1528 void 1529 tcp_mtudisc_callback(faddr) 1530 struct in_addr faddr; 1531 { 1532 #ifdef INET6 1533 struct in6_addr in6; 1534 #endif 1535 1536 in_pcbnotifyall(&tcbtable, faddr, EMSGSIZE, tcp_mtudisc); 1537 #ifdef INET6 1538 memset(&in6, 0, sizeof(in6)); 1539 in6.s6_addr16[5] = 0xffff; 1540 memcpy(&in6.s6_addr32[3], &faddr, sizeof(struct in_addr)); 1541 tcp6_mtudisc_callback(&in6); 1542 #endif 1543 } 1544 1545 /* 1546 * On receipt of path MTU corrections, flush old route and replace it 1547 * with the new one. Retransmit all unacknowledged packets, to ensure 1548 * that all packets will be received. 1549 */ 1550 void 1551 tcp_mtudisc(inp, errno) 1552 struct inpcb *inp; 1553 int errno; 1554 { 1555 struct tcpcb *tp = intotcpcb(inp); 1556 struct rtentry *rt = in_pcbrtentry(inp); 1557 1558 if (tp != 0) { 1559 if (rt != 0) { 1560 /* 1561 * If this was not a host route, remove and realloc. 1562 */ 1563 if ((rt->rt_flags & RTF_HOST) == 0) { 1564 in_rtchange(inp, errno); 1565 if ((rt = in_pcbrtentry(inp)) == 0) 1566 return; 1567 } 1568 1569 /* 1570 * Slow start out of the error condition. We 1571 * use the MTU because we know it's smaller 1572 * than the previously transmitted segment. 1573 * 1574 * Note: This is more conservative than the 1575 * suggestion in draft-floyd-incr-init-win-03. 1576 */ 1577 if (rt->rt_rmx.rmx_mtu != 0) 1578 tp->snd_cwnd = 1579 TCP_INITIAL_WINDOW(tcp_init_win, 1580 rt->rt_rmx.rmx_mtu); 1581 } 1582 1583 /* 1584 * Resend unacknowledged packets. 1585 */ 1586 tp->snd_nxt = tp->snd_una; 1587 tcp_output(tp); 1588 } 1589 } 1590 #endif 1591 1592 #ifdef INET6 1593 /* 1594 * Path MTU Discovery handlers. 1595 */ 1596 void 1597 tcp6_mtudisc_callback(faddr) 1598 struct in6_addr *faddr; 1599 { 1600 struct sockaddr_in6 sin6; 1601 1602 bzero(&sin6, sizeof(sin6)); 1603 sin6.sin6_family = AF_INET6; 1604 sin6.sin6_len = sizeof(struct sockaddr_in6); 1605 sin6.sin6_addr = *faddr; 1606 (void) in6_pcbnotify(&tcb6, (struct sockaddr *)&sin6, 0, 1607 (struct sockaddr *)&sa6_any, 0, PRC_MSGSIZE, NULL, tcp6_mtudisc); 1608 } 1609 1610 void 1611 tcp6_mtudisc(in6p, errno) 1612 struct in6pcb *in6p; 1613 int errno; 1614 { 1615 struct tcpcb *tp = in6totcpcb(in6p); 1616 struct rtentry *rt = in6_pcbrtentry(in6p); 1617 1618 if (tp != 0) { 1619 if (rt != 0) { 1620 /* 1621 * If this was not a host route, remove and realloc. 1622 */ 1623 if ((rt->rt_flags & RTF_HOST) == 0) { 1624 in6_rtchange(in6p, errno); 1625 if ((rt = in6_pcbrtentry(in6p)) == 0) 1626 return; 1627 } 1628 1629 /* 1630 * Slow start out of the error condition. We 1631 * use the MTU because we know it's smaller 1632 * than the previously transmitted segment. 1633 * 1634 * Note: This is more conservative than the 1635 * suggestion in draft-floyd-incr-init-win-03. 1636 */ 1637 if (rt->rt_rmx.rmx_mtu != 0) 1638 tp->snd_cwnd = 1639 TCP_INITIAL_WINDOW(tcp_init_win, 1640 rt->rt_rmx.rmx_mtu); 1641 } 1642 1643 /* 1644 * Resend unacknowledged packets. 1645 */ 1646 tp->snd_nxt = tp->snd_una; 1647 tcp_output(tp); 1648 } 1649 } 1650 #endif /* INET6 */ 1651 1652 /* 1653 * Compute the MSS to advertise to the peer. Called only during 1654 * the 3-way handshake. If we are the server (peer initiated 1655 * connection), we are called with a pointer to the interface 1656 * on which the SYN packet arrived. If we are the client (we 1657 * initiated connection), we are called with a pointer to the 1658 * interface out which this connection should go. 1659 * 1660 * NOTE: Do not subtract IP option/extension header size nor IPsec 1661 * header size from MSS advertisement. MSS option must hold the maximum 1662 * segment size we can accept, so it must always be: 1663 * max(if mtu) - ip header - tcp header 1664 */ 1665 u_long 1666 tcp_mss_to_advertise(ifp, af) 1667 const struct ifnet *ifp; 1668 int af; 1669 { 1670 extern u_long in_maxmtu; 1671 u_long mss = 0; 1672 u_long hdrsiz; 1673 1674 /* 1675 * In order to avoid defeating path MTU discovery on the peer, 1676 * we advertise the max MTU of all attached networks as our MSS, 1677 * per RFC 1191, section 3.1. 1678 * 1679 * We provide the option to advertise just the MTU of 1680 * the interface on which we hope this connection will 1681 * be receiving. If we are responding to a SYN, we 1682 * will have a pretty good idea about this, but when 1683 * initiating a connection there is a bit more doubt. 1684 * 1685 * We also need to ensure that loopback has a large enough 1686 * MSS, as the loopback MTU is never included in in_maxmtu. 1687 */ 1688 1689 if (ifp != NULL) 1690 switch (af) { 1691 case AF_INET: 1692 mss = ifp->if_mtu; 1693 break; 1694 #ifdef INET6 1695 case AF_INET6: 1696 mss = IN6_LINKMTU(ifp); 1697 break; 1698 #endif 1699 } 1700 1701 if (tcp_mss_ifmtu == 0) 1702 switch (af) { 1703 case AF_INET: 1704 mss = max(in_maxmtu, mss); 1705 break; 1706 #ifdef INET6 1707 case AF_INET6: 1708 mss = max(in6_maxmtu, mss); 1709 break; 1710 #endif 1711 } 1712 1713 switch (af) { 1714 case AF_INET: 1715 hdrsiz = sizeof(struct ip); 1716 break; 1717 #ifdef INET6 1718 case AF_INET6: 1719 hdrsiz = sizeof(struct ip6_hdr); 1720 break; 1721 #endif 1722 default: 1723 hdrsiz = 0; 1724 break; 1725 } 1726 hdrsiz += sizeof(struct tcphdr); 1727 if (mss > hdrsiz) 1728 mss -= hdrsiz; 1729 1730 mss = max(tcp_mssdflt, mss); 1731 return (mss); 1732 } 1733 1734 /* 1735 * Set connection variables based on the peer's advertised MSS. 1736 * We are passed the TCPCB for the actual connection. If we 1737 * are the server, we are called by the compressed state engine 1738 * when the 3-way handshake is complete. If we are the client, 1739 * we are called when we receive the SYN,ACK from the server. 1740 * 1741 * NOTE: Our advertised MSS value must be initialized in the TCPCB 1742 * before this routine is called! 1743 */ 1744 void 1745 tcp_mss_from_peer(tp, offer) 1746 struct tcpcb *tp; 1747 int offer; 1748 { 1749 struct socket *so; 1750 #if defined(RTV_SPIPE) || defined(RTV_SSTHRESH) 1751 struct rtentry *rt; 1752 #endif 1753 u_long bufsize; 1754 int mss; 1755 1756 #ifdef DIAGNOSTIC 1757 if (tp->t_inpcb && tp->t_in6pcb) 1758 panic("tcp_mss_from_peer: both t_inpcb and t_in6pcb are set"); 1759 #endif 1760 so = NULL; 1761 rt = NULL; 1762 #ifdef INET 1763 if (tp->t_inpcb) { 1764 so = tp->t_inpcb->inp_socket; 1765 #if defined(RTV_SPIPE) || defined(RTV_SSTHRESH) 1766 rt = in_pcbrtentry(tp->t_inpcb); 1767 #endif 1768 } 1769 #endif 1770 #ifdef INET6 1771 if (tp->t_in6pcb) { 1772 so = tp->t_in6pcb->in6p_socket; 1773 #if defined(RTV_SPIPE) || defined(RTV_SSTHRESH) 1774 rt = in6_pcbrtentry(tp->t_in6pcb); 1775 #endif 1776 } 1777 #endif 1778 1779 /* 1780 * As per RFC1122, use the default MSS value, unless they 1781 * sent us an offer. Do not accept offers less than 32 bytes. 1782 */ 1783 mss = tcp_mssdflt; 1784 if (offer) 1785 mss = offer; 1786 mss = max(mss, 32); /* sanity */ 1787 tp->t_peermss = mss; 1788 mss -= tcp_optlen(tp); 1789 #ifdef INET 1790 if (tp->t_inpcb) 1791 mss -= ip_optlen(tp->t_inpcb); 1792 #endif 1793 #ifdef INET6 1794 if (tp->t_in6pcb) 1795 mss -= ip6_optlen(tp->t_in6pcb); 1796 #endif 1797 1798 /* 1799 * If there's a pipesize, change the socket buffer to that size. 1800 * Make the socket buffer an integral number of MSS units. If 1801 * the MSS is larger than the socket buffer, artificially decrease 1802 * the MSS. 1803 */ 1804 #ifdef RTV_SPIPE 1805 if (rt != NULL && rt->rt_rmx.rmx_sendpipe != 0) 1806 bufsize = rt->rt_rmx.rmx_sendpipe; 1807 else 1808 #endif 1809 bufsize = so->so_snd.sb_hiwat; 1810 if (bufsize < mss) 1811 mss = bufsize; 1812 else { 1813 bufsize = roundup(bufsize, mss); 1814 if (bufsize > sb_max) 1815 bufsize = sb_max; 1816 (void) sbreserve(&so->so_snd, bufsize); 1817 } 1818 tp->t_segsz = mss; 1819 1820 #ifdef RTV_SSTHRESH 1821 if (rt != NULL && rt->rt_rmx.rmx_ssthresh) { 1822 /* 1823 * There's some sort of gateway or interface buffer 1824 * limit on the path. Use this to set the slow 1825 * start threshold, but set the threshold to no less 1826 * than 2 * MSS. 1827 */ 1828 tp->snd_ssthresh = max(2 * mss, rt->rt_rmx.rmx_ssthresh); 1829 } 1830 #endif 1831 } 1832 1833 /* 1834 * Processing necessary when a TCP connection is established. 1835 */ 1836 void 1837 tcp_established(tp) 1838 struct tcpcb *tp; 1839 { 1840 struct socket *so; 1841 #ifdef RTV_RPIPE 1842 struct rtentry *rt; 1843 #endif 1844 u_long bufsize; 1845 1846 #ifdef DIAGNOSTIC 1847 if (tp->t_inpcb && tp->t_in6pcb) 1848 panic("tcp_established: both t_inpcb and t_in6pcb are set"); 1849 #endif 1850 so = NULL; 1851 rt = NULL; 1852 #ifdef INET 1853 if (tp->t_inpcb) { 1854 so = tp->t_inpcb->inp_socket; 1855 #if defined(RTV_RPIPE) 1856 rt = in_pcbrtentry(tp->t_inpcb); 1857 #endif 1858 } 1859 #endif 1860 #ifdef INET6 1861 if (tp->t_in6pcb) { 1862 so = tp->t_in6pcb->in6p_socket; 1863 #if defined(RTV_RPIPE) 1864 rt = in6_pcbrtentry(tp->t_in6pcb); 1865 #endif 1866 } 1867 #endif 1868 1869 tp->t_state = TCPS_ESTABLISHED; 1870 TCP_TIMER_ARM(tp, TCPT_KEEP, tcp_keepidle); 1871 1872 #ifdef RTV_RPIPE 1873 if (rt != NULL && rt->rt_rmx.rmx_recvpipe != 0) 1874 bufsize = rt->rt_rmx.rmx_recvpipe; 1875 else 1876 #endif 1877 bufsize = so->so_rcv.sb_hiwat; 1878 if (bufsize > tp->t_ourmss) { 1879 bufsize = roundup(bufsize, tp->t_ourmss); 1880 if (bufsize > sb_max) 1881 bufsize = sb_max; 1882 (void) sbreserve(&so->so_rcv, bufsize); 1883 } 1884 } 1885 1886 /* 1887 * Check if there's an initial rtt or rttvar. Convert from the 1888 * route-table units to scaled multiples of the slow timeout timer. 1889 * Called only during the 3-way handshake. 1890 */ 1891 void 1892 tcp_rmx_rtt(tp) 1893 struct tcpcb *tp; 1894 { 1895 #ifdef RTV_RTT 1896 struct rtentry *rt = NULL; 1897 int rtt; 1898 1899 #ifdef DIAGNOSTIC 1900 if (tp->t_inpcb && tp->t_in6pcb) 1901 panic("tcp_rmx_rtt: both t_inpcb and t_in6pcb are set"); 1902 #endif 1903 #ifdef INET 1904 if (tp->t_inpcb) 1905 rt = in_pcbrtentry(tp->t_inpcb); 1906 #endif 1907 #ifdef INET6 1908 if (tp->t_in6pcb) 1909 rt = in6_pcbrtentry(tp->t_in6pcb); 1910 #endif 1911 if (rt == NULL) 1912 return; 1913 1914 if (tp->t_srtt == 0 && (rtt = rt->rt_rmx.rmx_rtt)) { 1915 /* 1916 * XXX The lock bit for MTU indicates that the value 1917 * is also a minimum value; this is subject to time. 1918 */ 1919 if (rt->rt_rmx.rmx_locks & RTV_RTT) 1920 TCPT_RANGESET(tp->t_rttmin, 1921 rtt / (RTM_RTTUNIT / PR_SLOWHZ), 1922 TCPTV_MIN, TCPTV_REXMTMAX); 1923 tp->t_srtt = rtt / 1924 ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTT_SHIFT + 2)); 1925 if (rt->rt_rmx.rmx_rttvar) { 1926 tp->t_rttvar = rt->rt_rmx.rmx_rttvar / 1927 ((RTM_RTTUNIT / PR_SLOWHZ) >> 1928 (TCP_RTTVAR_SHIFT + 2)); 1929 } else { 1930 /* Default variation is +- 1 rtt */ 1931 tp->t_rttvar = 1932 tp->t_srtt >> (TCP_RTT_SHIFT - TCP_RTTVAR_SHIFT); 1933 } 1934 TCPT_RANGESET(tp->t_rxtcur, 1935 ((tp->t_srtt >> 2) + tp->t_rttvar) >> (1 + 2), 1936 tp->t_rttmin, TCPTV_REXMTMAX); 1937 } 1938 #endif 1939 } 1940 1941 tcp_seq tcp_iss_seq = 0; /* tcp initial seq # */ 1942 #if NRND > 0 1943 u_int8_t tcp_iss_secret[16]; /* 128 bits; should be plenty */ 1944 #endif 1945 1946 /* 1947 * Get a new sequence value given a tcp control block 1948 */ 1949 tcp_seq 1950 tcp_new_iss(struct tcpcb *tp, tcp_seq addin) 1951 { 1952 1953 #ifdef INET 1954 if (tp->t_inpcb != NULL) { 1955 return (tcp_new_iss1(&tp->t_inpcb->inp_laddr, 1956 &tp->t_inpcb->inp_faddr, tp->t_inpcb->inp_lport, 1957 tp->t_inpcb->inp_fport, sizeof(tp->t_inpcb->inp_laddr), 1958 addin)); 1959 } 1960 #endif 1961 #ifdef INET6 1962 if (tp->t_in6pcb != NULL) { 1963 return (tcp_new_iss1(&tp->t_in6pcb->in6p_laddr, 1964 &tp->t_in6pcb->in6p_faddr, tp->t_in6pcb->in6p_lport, 1965 tp->t_in6pcb->in6p_fport, sizeof(tp->t_in6pcb->in6p_laddr), 1966 addin)); 1967 } 1968 #endif 1969 /* Not possible. */ 1970 panic("tcp_new_iss"); 1971 } 1972 1973 /* 1974 * This routine actually generates a new TCP initial sequence number. 1975 */ 1976 tcp_seq 1977 tcp_new_iss1(void *laddr, void *faddr, u_int16_t lport, u_int16_t fport, 1978 size_t addrsz, tcp_seq addin) 1979 { 1980 tcp_seq tcp_iss; 1981 1982 #if NRND > 0 1983 static int beenhere; 1984 1985 /* 1986 * If we haven't been here before, initialize our cryptographic 1987 * hash secret. 1988 */ 1989 if (beenhere == 0) { 1990 rnd_extract_data(tcp_iss_secret, sizeof(tcp_iss_secret), 1991 RND_EXTRACT_ANY); 1992 beenhere = 1; 1993 } 1994 1995 if (tcp_do_rfc1948) { 1996 MD5_CTX ctx; 1997 u_int8_t hash[16]; /* XXX MD5 knowledge */ 1998 1999 /* 2000 * Compute the base value of the ISS. It is a hash 2001 * of (saddr, sport, daddr, dport, secret). 2002 */ 2003 MD5Init(&ctx); 2004 2005 MD5Update(&ctx, (u_char *) laddr, addrsz); 2006 MD5Update(&ctx, (u_char *) &lport, sizeof(lport)); 2007 2008 MD5Update(&ctx, (u_char *) faddr, addrsz); 2009 MD5Update(&ctx, (u_char *) &fport, sizeof(fport)); 2010 2011 MD5Update(&ctx, tcp_iss_secret, sizeof(tcp_iss_secret)); 2012 2013 MD5Final(hash, &ctx); 2014 2015 memcpy(&tcp_iss, hash, sizeof(tcp_iss)); 2016 2017 /* 2018 * Now increment our "timer", and add it in to 2019 * the computed value. 2020 * 2021 * XXX Use `addin'? 2022 * XXX TCP_ISSINCR too large to use? 2023 */ 2024 tcp_iss_seq += TCP_ISSINCR; 2025 #ifdef TCPISS_DEBUG 2026 printf("ISS hash 0x%08x, ", tcp_iss); 2027 #endif 2028 tcp_iss += tcp_iss_seq + addin; 2029 #ifdef TCPISS_DEBUG 2030 printf("new ISS 0x%08x\n", tcp_iss); 2031 #endif 2032 } else 2033 #endif /* NRND > 0 */ 2034 { 2035 /* 2036 * Randomize. 2037 */ 2038 #if NRND > 0 2039 rnd_extract_data(&tcp_iss, sizeof(tcp_iss), RND_EXTRACT_ANY); 2040 #else 2041 tcp_iss = arc4random(); 2042 #endif 2043 2044 /* 2045 * If we were asked to add some amount to a known value, 2046 * we will take a random value obtained above, mask off 2047 * the upper bits, and add in the known value. We also 2048 * add in a constant to ensure that we are at least a 2049 * certain distance from the original value. 2050 * 2051 * This is used when an old connection is in timed wait 2052 * and we have a new one coming in, for instance. 2053 */ 2054 if (addin != 0) { 2055 #ifdef TCPISS_DEBUG 2056 printf("Random %08x, ", tcp_iss); 2057 #endif 2058 tcp_iss &= TCP_ISS_RANDOM_MASK; 2059 tcp_iss += addin + TCP_ISSINCR; 2060 #ifdef TCPISS_DEBUG 2061 printf("Old ISS %08x, ISS %08x\n", addin, tcp_iss); 2062 #endif 2063 } else { 2064 tcp_iss &= TCP_ISS_RANDOM_MASK; 2065 tcp_iss += tcp_iss_seq; 2066 tcp_iss_seq += TCP_ISSINCR; 2067 #ifdef TCPISS_DEBUG 2068 printf("ISS %08x\n", tcp_iss); 2069 #endif 2070 } 2071 } 2072 2073 if (tcp_compat_42) { 2074 /* 2075 * Limit it to the positive range for really old TCP 2076 * implementations. 2077 */ 2078 if (tcp_iss >= 0x80000000) 2079 tcp_iss &= 0x7fffffff; /* XXX */ 2080 } 2081 2082 return (tcp_iss); 2083 } 2084 2085 #ifdef IPSEC 2086 /* compute ESP/AH header size for TCP, including outer IP header. */ 2087 size_t 2088 ipsec4_hdrsiz_tcp(tp) 2089 struct tcpcb *tp; 2090 { 2091 struct inpcb *inp; 2092 size_t hdrsiz; 2093 2094 /* XXX mapped addr case (tp->t_in6pcb) */ 2095 if (!tp || !tp->t_template || !(inp = tp->t_inpcb)) 2096 return 0; 2097 switch (tp->t_family) { 2098 case AF_INET: 2099 /* XXX: should use currect direction. */ 2100 hdrsiz = ipsec4_hdrsiz(tp->t_template, IPSEC_DIR_OUTBOUND, inp); 2101 break; 2102 default: 2103 hdrsiz = 0; 2104 break; 2105 } 2106 2107 return hdrsiz; 2108 } 2109 2110 #ifdef INET6 2111 size_t 2112 ipsec6_hdrsiz_tcp(tp) 2113 struct tcpcb *tp; 2114 { 2115 struct in6pcb *in6p; 2116 size_t hdrsiz; 2117 2118 if (!tp || !tp->t_template || !(in6p = tp->t_in6pcb)) 2119 return 0; 2120 switch (tp->t_family) { 2121 case AF_INET6: 2122 /* XXX: should use currect direction. */ 2123 hdrsiz = ipsec6_hdrsiz(tp->t_template, IPSEC_DIR_OUTBOUND, in6p); 2124 break; 2125 case AF_INET: 2126 /* mapped address case - tricky */ 2127 default: 2128 hdrsiz = 0; 2129 break; 2130 } 2131 2132 return hdrsiz; 2133 } 2134 #endif 2135 #endif /*IPSEC*/ 2136 2137 /* 2138 * Determine the length of the TCP options for this connection. 2139 * 2140 * XXX: What do we do for SACK, when we add that? Just reserve 2141 * all of the space? Otherwise we can't exactly be incrementing 2142 * cwnd by an amount that varies depending on the amount we last 2143 * had to SACK! 2144 */ 2145 2146 u_int 2147 tcp_optlen(tp) 2148 struct tcpcb *tp; 2149 { 2150 if ((tp->t_flags & (TF_REQ_TSTMP|TF_RCVD_TSTMP|TF_NOOPT)) == 2151 (TF_REQ_TSTMP | TF_RCVD_TSTMP)) 2152 return TCPOLEN_TSTAMP_APPA; 2153 else 2154 return 0; 2155 } 2156
Indexes created Sun Sep 21 20:09:37 GMT 2025