1 /* $NetBSD: tcp_subr.c,v 1.173 2004/09/15 09:21:22 yamt 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. Neither the name of the University nor the names of its contributors 82 * may be used to endorse or promote products derived from this software 83 * without specific prior written permission. 84 * 85 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 86 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 87 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 88 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 89 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 90 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 91 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 92 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 93 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 94 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 95 * SUCH DAMAGE. 96 * 97 * @(#)tcp_subr.c 8.2 (Berkeley) 5/24/95 98 */ 99 100 #include <sys/cdefs.h> 101 __KERNEL_RCSID(0, "$NetBSD: tcp_subr.c,v 1.173 2004/09/15 09:21:22 yamt Exp $"); 102 103 #include "opt_inet.h" 104 #include "opt_ipsec.h" 105 #include "opt_tcp_compat_42.h" 106 #include "opt_inet_csum.h" 107 #include "opt_mbuftrace.h" 108 #include "rnd.h" 109 110 #include <sys/param.h> 111 #include <sys/proc.h> 112 #include <sys/systm.h> 113 #include <sys/malloc.h> 114 #include <sys/mbuf.h> 115 #include <sys/socket.h> 116 #include <sys/socketvar.h> 117 #include <sys/protosw.h> 118 #include <sys/errno.h> 119 #include <sys/kernel.h> 120 #include <sys/pool.h> 121 #if NRND > 0 122 #include <sys/md5.h> 123 #include <sys/rnd.h> 124 #endif 125 126 #include <net/route.h> 127 #include <net/if.h> 128 129 #include <netinet/in.h> 130 #include <netinet/in_systm.h> 131 #include <netinet/ip.h> 132 #include <netinet/in_pcb.h> 133 #include <netinet/ip_var.h> 134 #include <netinet/ip_icmp.h> 135 136 #ifdef INET6 137 #ifndef INET 138 #include <netinet/in.h> 139 #endif 140 #include <netinet/ip6.h> 141 #include <netinet6/in6_pcb.h> 142 #include <netinet6/ip6_var.h> 143 #include <netinet6/in6_var.h> 144 #include <netinet6/ip6protosw.h> 145 #include <netinet/icmp6.h> 146 #include <netinet6/nd6.h> 147 #endif 148 149 #include <netinet/tcp.h> 150 #include <netinet/tcp_fsm.h> 151 #include <netinet/tcp_seq.h> 152 #include <netinet/tcp_timer.h> 153 #include <netinet/tcp_var.h> 154 #include <netinet/tcpip.h> 155 156 #ifdef IPSEC 157 #include <netinet6/ipsec.h> 158 #include <netkey/key.h> 159 #endif /*IPSEC*/ 160 161 #ifdef FAST_IPSEC 162 #include <netipsec/ipsec.h> 163 #include <netipsec/xform.h> 164 #ifdef INET6 165 #include <netipsec/ipsec6.h> 166 #endif 167 #include <netipsec/key.h> 168 #endif /* FAST_IPSEC*/ 169 170 171 struct inpcbtable tcbtable; /* head of queue of active tcpcb's */ 172 struct tcpstat tcpstat; /* tcp statistics */ 173 u_int32_t tcp_now; /* for RFC 1323 timestamps */ 174 175 /* patchable/settable parameters for tcp */ 176 int tcp_mssdflt = TCP_MSS; 177 int tcp_rttdflt = TCPTV_SRTTDFLT / PR_SLOWHZ; 178 int tcp_do_rfc1323 = 1; /* window scaling / timestamps (obsolete) */ 179 #if NRND > 0 180 int tcp_do_rfc1948 = 0; /* ISS by cryptographic hash */ 181 #endif 182 int tcp_do_sack = 1; /* selective acknowledgement */ 183 int tcp_do_win_scale = 1; /* RFC1323 window scaling */ 184 int tcp_do_timestamps = 1; /* RFC1323 timestamps */ 185 int tcp_do_newreno = 1; /* Use the New Reno algorithms */ 186 int tcp_ack_on_push = 0; /* set to enable immediate ACK-on-PUSH */ 187 #ifndef TCP_INIT_WIN 188 #define TCP_INIT_WIN 0 /* initial slow start window */ 189 #endif 190 #ifndef TCP_INIT_WIN_LOCAL 191 #define TCP_INIT_WIN_LOCAL 4 /* initial slow start window for local nets */ 192 #endif 193 int tcp_init_win = TCP_INIT_WIN; 194 int tcp_init_win_local = TCP_INIT_WIN_LOCAL; 195 int tcp_mss_ifmtu = 0; 196 #ifdef TCP_COMPAT_42 197 int tcp_compat_42 = 1; 198 #else 199 int tcp_compat_42 = 0; 200 #endif 201 int tcp_rst_ppslim = 100; /* 100pps */ 202 int tcp_ackdrop_ppslim = 100; /* 100pps */ 203 204 /* tcb hash */ 205 #ifndef TCBHASHSIZE 206 #define TCBHASHSIZE 128 207 #endif 208 int tcbhashsize = TCBHASHSIZE; 209 210 /* syn hash parameters */ 211 #define TCP_SYN_HASH_SIZE 293 212 #define TCP_SYN_BUCKET_SIZE 35 213 int tcp_syn_cache_size = TCP_SYN_HASH_SIZE; 214 int tcp_syn_cache_limit = TCP_SYN_HASH_SIZE*TCP_SYN_BUCKET_SIZE; 215 int tcp_syn_bucket_limit = 3*TCP_SYN_BUCKET_SIZE; 216 struct syn_cache_head tcp_syn_cache[TCP_SYN_HASH_SIZE]; 217 218 int tcp_freeq __P((struct tcpcb *)); 219 220 #ifdef INET 221 void tcp_mtudisc_callback __P((struct in_addr)); 222 #endif 223 #ifdef INET6 224 void tcp6_mtudisc_callback __P((struct in6_addr *)); 225 #endif 226 227 void tcp_mtudisc __P((struct inpcb *, int)); 228 #ifdef INET6 229 void tcp6_mtudisc __P((struct in6pcb *, int)); 230 #endif 231 232 POOL_INIT(tcpcb_pool, sizeof(struct tcpcb), 0, 0, 0, "tcpcbpl", NULL); 233 234 #ifdef TCP_CSUM_COUNTERS 235 #include <sys/device.h> 236 237 struct evcnt tcp_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 238 NULL, "tcp", "hwcsum bad"); 239 struct evcnt tcp_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 240 NULL, "tcp", "hwcsum ok"); 241 struct evcnt tcp_hwcsum_data = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 242 NULL, "tcp", "hwcsum data"); 243 struct evcnt tcp_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 244 NULL, "tcp", "swcsum"); 245 246 EVCNT_ATTACH_STATIC(tcp_hwcsum_bad); 247 EVCNT_ATTACH_STATIC(tcp_hwcsum_ok); 248 EVCNT_ATTACH_STATIC(tcp_hwcsum_data); 249 EVCNT_ATTACH_STATIC(tcp_swcsum); 250 #endif /* TCP_CSUM_COUNTERS */ 251 252 253 #ifdef TCP_OUTPUT_COUNTERS 254 #include <sys/device.h> 255 256 struct evcnt tcp_output_bigheader = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 257 NULL, "tcp", "output big header"); 258 struct evcnt tcp_output_predict_hit = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 259 NULL, "tcp", "output predict hit"); 260 struct evcnt tcp_output_predict_miss = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 261 NULL, "tcp", "output predict miss"); 262 struct evcnt tcp_output_copysmall = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 263 NULL, "tcp", "output copy small"); 264 struct evcnt tcp_output_copybig = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 265 NULL, "tcp", "output copy big"); 266 struct evcnt tcp_output_refbig = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 267 NULL, "tcp", "output reference big"); 268 269 EVCNT_ATTACH_STATIC(tcp_output_bigheader); 270 EVCNT_ATTACH_STATIC(tcp_output_predict_hit); 271 EVCNT_ATTACH_STATIC(tcp_output_predict_miss); 272 EVCNT_ATTACH_STATIC(tcp_output_copysmall); 273 EVCNT_ATTACH_STATIC(tcp_output_copybig); 274 EVCNT_ATTACH_STATIC(tcp_output_refbig); 275 276 #endif /* TCP_OUTPUT_COUNTERS */ 277 278 #ifdef TCP_REASS_COUNTERS 279 #include <sys/device.h> 280 281 struct evcnt tcp_reass_ = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 282 NULL, "tcp_reass", "calls"); 283 struct evcnt tcp_reass_empty = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 284 &tcp_reass_, "tcp_reass", "insert into empty queue"); 285 struct evcnt tcp_reass_iteration[8] = { 286 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", ">7 iterations"), 287 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "1 iteration"), 288 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "2 iterations"), 289 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "3 iterations"), 290 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "4 iterations"), 291 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "5 iterations"), 292 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "6 iterations"), 293 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "7 iterations"), 294 }; 295 struct evcnt tcp_reass_prependfirst = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 296 &tcp_reass_, "tcp_reass", "prepend to first"); 297 struct evcnt tcp_reass_prepend = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 298 &tcp_reass_, "tcp_reass", "prepend"); 299 struct evcnt tcp_reass_insert = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 300 &tcp_reass_, "tcp_reass", "insert"); 301 struct evcnt tcp_reass_inserttail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 302 &tcp_reass_, "tcp_reass", "insert at tail"); 303 struct evcnt tcp_reass_append = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 304 &tcp_reass_, "tcp_reass", "append"); 305 struct evcnt tcp_reass_appendtail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 306 &tcp_reass_, "tcp_reass", "append to tail fragment"); 307 struct evcnt tcp_reass_overlaptail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 308 &tcp_reass_, "tcp_reass", "overlap at end"); 309 struct evcnt tcp_reass_overlapfront = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 310 &tcp_reass_, "tcp_reass", "overlap at start"); 311 struct evcnt tcp_reass_segdup = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 312 &tcp_reass_, "tcp_reass", "duplicate segment"); 313 struct evcnt tcp_reass_fragdup = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 314 &tcp_reass_, "tcp_reass", "duplicate fragment"); 315 316 EVCNT_ATTACH_STATIC(tcp_reass_); 317 EVCNT_ATTACH_STATIC(tcp_reass_empty); 318 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 0); 319 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 1); 320 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 2); 321 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 3); 322 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 4); 323 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 5); 324 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 6); 325 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 7); 326 EVCNT_ATTACH_STATIC(tcp_reass_prependfirst); 327 EVCNT_ATTACH_STATIC(tcp_reass_prepend); 328 EVCNT_ATTACH_STATIC(tcp_reass_insert); 329 EVCNT_ATTACH_STATIC(tcp_reass_inserttail); 330 EVCNT_ATTACH_STATIC(tcp_reass_append); 331 EVCNT_ATTACH_STATIC(tcp_reass_appendtail); 332 EVCNT_ATTACH_STATIC(tcp_reass_overlaptail); 333 EVCNT_ATTACH_STATIC(tcp_reass_overlapfront); 334 EVCNT_ATTACH_STATIC(tcp_reass_segdup); 335 EVCNT_ATTACH_STATIC(tcp_reass_fragdup); 336 337 #endif /* TCP_REASS_COUNTERS */ 338 339 #ifdef MBUFTRACE 340 struct mowner tcp_mowner = { "tcp" }; 341 struct mowner tcp_rx_mowner = { "tcp", "rx" }; 342 struct mowner tcp_tx_mowner = { "tcp", "tx" }; 343 #endif 344 345 /* 346 * Tcp initialization 347 */ 348 void 349 tcp_init() 350 { 351 int hlen; 352 353 /* Initialize the TCPCB template. */ 354 tcp_tcpcb_template(); 355 356 in_pcbinit(&tcbtable, tcbhashsize, tcbhashsize); 357 358 hlen = sizeof(struct ip) + sizeof(struct tcphdr); 359 #ifdef INET6 360 if (sizeof(struct ip) < sizeof(struct ip6_hdr)) 361 hlen = sizeof(struct ip6_hdr) + sizeof(struct tcphdr); 362 #endif 363 if (max_protohdr < hlen) 364 max_protohdr = hlen; 365 if (max_linkhdr + hlen > MHLEN) 366 panic("tcp_init"); 367 368 #ifdef INET 369 icmp_mtudisc_callback_register(tcp_mtudisc_callback); 370 #endif 371 #ifdef INET6 372 icmp6_mtudisc_callback_register(tcp6_mtudisc_callback); 373 #endif 374 375 /* Initialize timer state. */ 376 tcp_timer_init(); 377 378 /* Initialize the compressed state engine. */ 379 syn_cache_init(); 380 381 MOWNER_ATTACH(&tcp_tx_mowner); 382 MOWNER_ATTACH(&tcp_rx_mowner); 383 MOWNER_ATTACH(&tcp_mowner); 384 } 385 386 /* 387 * Create template to be used to send tcp packets on a connection. 388 * Call after host entry created, allocates an mbuf and fills 389 * in a skeletal tcp/ip header, minimizing the amount of work 390 * necessary when the connection is used. 391 */ 392 struct mbuf * 393 tcp_template(tp) 394 struct tcpcb *tp; 395 { 396 struct inpcb *inp = tp->t_inpcb; 397 #ifdef INET6 398 struct in6pcb *in6p = tp->t_in6pcb; 399 #endif 400 struct tcphdr *n; 401 struct mbuf *m; 402 int hlen; 403 404 switch (tp->t_family) { 405 case AF_INET: 406 hlen = sizeof(struct ip); 407 if (inp) 408 break; 409 #ifdef INET6 410 if (in6p) { 411 /* mapped addr case */ 412 if (IN6_IS_ADDR_V4MAPPED(&in6p->in6p_laddr) 413 && IN6_IS_ADDR_V4MAPPED(&in6p->in6p_faddr)) 414 break; 415 } 416 #endif 417 return NULL; /*EINVAL*/ 418 #ifdef INET6 419 case AF_INET6: 420 hlen = sizeof(struct ip6_hdr); 421 if (in6p) { 422 /* more sainty check? */ 423 break; 424 } 425 return NULL; /*EINVAL*/ 426 #endif 427 default: 428 hlen = 0; /*pacify gcc*/ 429 return NULL; /*EAFNOSUPPORT*/ 430 } 431 #ifdef DIAGNOSTIC 432 if (hlen + sizeof(struct tcphdr) > MCLBYTES) 433 panic("mclbytes too small for t_template"); 434 #endif 435 m = tp->t_template; 436 if (m && m->m_len == hlen + sizeof(struct tcphdr)) 437 ; 438 else { 439 if (m) 440 m_freem(m); 441 m = tp->t_template = NULL; 442 MGETHDR(m, M_DONTWAIT, MT_HEADER); 443 if (m && hlen + sizeof(struct tcphdr) > MHLEN) { 444 MCLGET(m, M_DONTWAIT); 445 if ((m->m_flags & M_EXT) == 0) { 446 m_free(m); 447 m = NULL; 448 } 449 } 450 if (m == NULL) 451 return NULL; 452 MCLAIM(m, &tcp_mowner); 453 m->m_pkthdr.len = m->m_len = hlen + sizeof(struct tcphdr); 454 } 455 456 bzero(mtod(m, caddr_t), m->m_len); 457 458 n = (struct tcphdr *)(mtod(m, caddr_t) + hlen); 459 460 switch (tp->t_family) { 461 case AF_INET: 462 { 463 struct ipovly *ipov; 464 mtod(m, struct ip *)->ip_v = 4; 465 mtod(m, struct ip *)->ip_hl = hlen >> 2; 466 ipov = mtod(m, struct ipovly *); 467 ipov->ih_pr = IPPROTO_TCP; 468 ipov->ih_len = htons(sizeof(struct tcphdr)); 469 if (inp) { 470 ipov->ih_src = inp->inp_laddr; 471 ipov->ih_dst = inp->inp_faddr; 472 } 473 #ifdef INET6 474 else if (in6p) { 475 /* mapped addr case */ 476 bcopy(&in6p->in6p_laddr.s6_addr32[3], &ipov->ih_src, 477 sizeof(ipov->ih_src)); 478 bcopy(&in6p->in6p_faddr.s6_addr32[3], &ipov->ih_dst, 479 sizeof(ipov->ih_dst)); 480 } 481 #endif 482 /* 483 * Compute the pseudo-header portion of the checksum 484 * now. We incrementally add in the TCP option and 485 * payload lengths later, and then compute the TCP 486 * checksum right before the packet is sent off onto 487 * the wire. 488 */ 489 n->th_sum = in_cksum_phdr(ipov->ih_src.s_addr, 490 ipov->ih_dst.s_addr, 491 htons(sizeof(struct tcphdr) + IPPROTO_TCP)); 492 break; 493 } 494 #ifdef INET6 495 case AF_INET6: 496 { 497 struct ip6_hdr *ip6; 498 mtod(m, struct ip *)->ip_v = 6; 499 ip6 = mtod(m, struct ip6_hdr *); 500 ip6->ip6_nxt = IPPROTO_TCP; 501 ip6->ip6_plen = htons(sizeof(struct tcphdr)); 502 ip6->ip6_src = in6p->in6p_laddr; 503 ip6->ip6_dst = in6p->in6p_faddr; 504 ip6->ip6_flow = in6p->in6p_flowinfo & IPV6_FLOWINFO_MASK; 505 if (ip6_auto_flowlabel) { 506 ip6->ip6_flow &= ~IPV6_FLOWLABEL_MASK; 507 ip6->ip6_flow |= 508 (htonl(ip6_randomflowlabel()) & IPV6_FLOWLABEL_MASK); 509 } 510 ip6->ip6_vfc &= ~IPV6_VERSION_MASK; 511 ip6->ip6_vfc |= IPV6_VERSION; 512 513 /* 514 * Compute the pseudo-header portion of the checksum 515 * now. We incrementally add in the TCP option and 516 * payload lengths later, and then compute the TCP 517 * checksum right before the packet is sent off onto 518 * the wire. 519 */ 520 n->th_sum = in6_cksum_phdr(&in6p->in6p_laddr, 521 &in6p->in6p_faddr, htonl(sizeof(struct tcphdr)), 522 htonl(IPPROTO_TCP)); 523 break; 524 } 525 #endif 526 } 527 if (inp) { 528 n->th_sport = inp->inp_lport; 529 n->th_dport = inp->inp_fport; 530 } 531 #ifdef INET6 532 else if (in6p) { 533 n->th_sport = in6p->in6p_lport; 534 n->th_dport = in6p->in6p_fport; 535 } 536 #endif 537 n->th_seq = 0; 538 n->th_ack = 0; 539 n->th_x2 = 0; 540 n->th_off = 5; 541 n->th_flags = 0; 542 n->th_win = 0; 543 n->th_urp = 0; 544 return (m); 545 } 546 547 /* 548 * Send a single message to the TCP at address specified by 549 * the given TCP/IP header. If m == 0, then we make a copy 550 * of the tcpiphdr at ti and send directly to the addressed host. 551 * This is used to force keep alive messages out using the TCP 552 * template for a connection tp->t_template. If flags are given 553 * then we send a message back to the TCP which originated the 554 * segment ti, and discard the mbuf containing it and any other 555 * attached mbufs. 556 * 557 * In any case the ack and sequence number of the transmitted 558 * segment are as specified by the parameters. 559 */ 560 int 561 tcp_respond(tp, template, m, th0, ack, seq, flags) 562 struct tcpcb *tp; 563 struct mbuf *template; 564 struct mbuf *m; 565 struct tcphdr *th0; 566 tcp_seq ack, seq; 567 int flags; 568 { 569 struct route *ro; 570 int error, tlen, win = 0; 571 int hlen; 572 struct ip *ip; 573 #ifdef INET6 574 struct ip6_hdr *ip6; 575 #endif 576 int family; /* family on packet, not inpcb/in6pcb! */ 577 struct tcphdr *th; 578 struct socket *so; 579 580 if (tp != NULL && (flags & TH_RST) == 0) { 581 #ifdef DIAGNOSTIC 582 if (tp->t_inpcb && tp->t_in6pcb) 583 panic("tcp_respond: both t_inpcb and t_in6pcb are set"); 584 #endif 585 #ifdef INET 586 if (tp->t_inpcb) 587 win = sbspace(&tp->t_inpcb->inp_socket->so_rcv); 588 #endif 589 #ifdef INET6 590 if (tp->t_in6pcb) 591 win = sbspace(&tp->t_in6pcb->in6p_socket->so_rcv); 592 #endif 593 } 594 595 th = NULL; /* Quell uninitialized warning */ 596 ip = NULL; 597 #ifdef INET6 598 ip6 = NULL; 599 #endif 600 if (m == 0) { 601 if (!template) 602 return EINVAL; 603 604 /* get family information from template */ 605 switch (mtod(template, struct ip *)->ip_v) { 606 case 4: 607 family = AF_INET; 608 hlen = sizeof(struct ip); 609 break; 610 #ifdef INET6 611 case 6: 612 family = AF_INET6; 613 hlen = sizeof(struct ip6_hdr); 614 break; 615 #endif 616 default: 617 return EAFNOSUPPORT; 618 } 619 620 MGETHDR(m, M_DONTWAIT, MT_HEADER); 621 if (m) { 622 MCLAIM(m, &tcp_tx_mowner); 623 MCLGET(m, M_DONTWAIT); 624 if ((m->m_flags & M_EXT) == 0) { 625 m_free(m); 626 m = NULL; 627 } 628 } 629 if (m == NULL) 630 return (ENOBUFS); 631 632 if (tcp_compat_42) 633 tlen = 1; 634 else 635 tlen = 0; 636 637 m->m_data += max_linkhdr; 638 bcopy(mtod(template, caddr_t), mtod(m, caddr_t), 639 template->m_len); 640 switch (family) { 641 case AF_INET: 642 ip = mtod(m, struct ip *); 643 th = (struct tcphdr *)(ip + 1); 644 break; 645 #ifdef INET6 646 case AF_INET6: 647 ip6 = mtod(m, struct ip6_hdr *); 648 th = (struct tcphdr *)(ip6 + 1); 649 break; 650 #endif 651 #if 0 652 default: 653 /* noone will visit here */ 654 m_freem(m); 655 return EAFNOSUPPORT; 656 #endif 657 } 658 flags = TH_ACK; 659 } else { 660 661 if ((m->m_flags & M_PKTHDR) == 0) { 662 #if 0 663 printf("non PKTHDR to tcp_respond\n"); 664 #endif 665 m_freem(m); 666 return EINVAL; 667 } 668 #ifdef DIAGNOSTIC 669 if (!th0) 670 panic("th0 == NULL in tcp_respond"); 671 #endif 672 673 /* get family information from m */ 674 switch (mtod(m, struct ip *)->ip_v) { 675 case 4: 676 family = AF_INET; 677 hlen = sizeof(struct ip); 678 ip = mtod(m, struct ip *); 679 break; 680 #ifdef INET6 681 case 6: 682 family = AF_INET6; 683 hlen = sizeof(struct ip6_hdr); 684 ip6 = mtod(m, struct ip6_hdr *); 685 break; 686 #endif 687 default: 688 m_freem(m); 689 return EAFNOSUPPORT; 690 } 691 if ((flags & TH_SYN) == 0 || sizeof(*th0) > (th0->th_off << 2)) 692 tlen = sizeof(*th0); 693 else 694 tlen = th0->th_off << 2; 695 696 if (m->m_len > hlen + tlen && (m->m_flags & M_EXT) == 0 && 697 mtod(m, caddr_t) + hlen == (caddr_t)th0) { 698 m->m_len = hlen + tlen; 699 m_freem(m->m_next); 700 m->m_next = NULL; 701 } else { 702 struct mbuf *n; 703 704 #ifdef DIAGNOSTIC 705 if (max_linkhdr + hlen + tlen > MCLBYTES) { 706 m_freem(m); 707 return EMSGSIZE; 708 } 709 #endif 710 MGETHDR(n, M_DONTWAIT, MT_HEADER); 711 if (n && max_linkhdr + hlen + tlen > MHLEN) { 712 MCLGET(n, M_DONTWAIT); 713 if ((n->m_flags & M_EXT) == 0) { 714 m_freem(n); 715 n = NULL; 716 } 717 } 718 if (!n) { 719 m_freem(m); 720 return ENOBUFS; 721 } 722 723 MCLAIM(n, &tcp_tx_mowner); 724 n->m_data += max_linkhdr; 725 n->m_len = hlen + tlen; 726 m_copyback(n, 0, hlen, mtod(m, caddr_t)); 727 m_copyback(n, hlen, tlen, (caddr_t)th0); 728 729 m_freem(m); 730 m = n; 731 n = NULL; 732 } 733 734 #define xchg(a,b,type) { type t; t=a; a=b; b=t; } 735 switch (family) { 736 case AF_INET: 737 ip = mtod(m, struct ip *); 738 th = (struct tcphdr *)(ip + 1); 739 ip->ip_p = IPPROTO_TCP; 740 xchg(ip->ip_dst, ip->ip_src, struct in_addr); 741 ip->ip_p = IPPROTO_TCP; 742 break; 743 #ifdef INET6 744 case AF_INET6: 745 ip6 = mtod(m, struct ip6_hdr *); 746 th = (struct tcphdr *)(ip6 + 1); 747 ip6->ip6_nxt = IPPROTO_TCP; 748 xchg(ip6->ip6_dst, ip6->ip6_src, struct in6_addr); 749 ip6->ip6_nxt = IPPROTO_TCP; 750 break; 751 #endif 752 #if 0 753 default: 754 /* noone will visit here */ 755 m_freem(m); 756 return EAFNOSUPPORT; 757 #endif 758 } 759 xchg(th->th_dport, th->th_sport, u_int16_t); 760 #undef xchg 761 tlen = 0; /*be friendly with the following code*/ 762 } 763 th->th_seq = htonl(seq); 764 th->th_ack = htonl(ack); 765 th->th_x2 = 0; 766 if ((flags & TH_SYN) == 0) { 767 if (tp) 768 win >>= tp->rcv_scale; 769 if (win > TCP_MAXWIN) 770 win = TCP_MAXWIN; 771 th->th_win = htons((u_int16_t)win); 772 th->th_off = sizeof (struct tcphdr) >> 2; 773 tlen += sizeof(*th); 774 } else 775 tlen += th->th_off << 2; 776 m->m_len = hlen + tlen; 777 m->m_pkthdr.len = hlen + tlen; 778 m->m_pkthdr.rcvif = (struct ifnet *) 0; 779 th->th_flags = flags; 780 th->th_urp = 0; 781 782 switch (family) { 783 #ifdef INET 784 case AF_INET: 785 { 786 struct ipovly *ipov = (struct ipovly *)ip; 787 bzero(ipov->ih_x1, sizeof ipov->ih_x1); 788 ipov->ih_len = htons((u_int16_t)tlen); 789 790 th->th_sum = 0; 791 th->th_sum = in_cksum(m, hlen + tlen); 792 ip->ip_len = htons(hlen + tlen); 793 ip->ip_ttl = ip_defttl; 794 break; 795 } 796 #endif 797 #ifdef INET6 798 case AF_INET6: 799 { 800 th->th_sum = 0; 801 th->th_sum = in6_cksum(m, IPPROTO_TCP, sizeof(struct ip6_hdr), 802 tlen); 803 ip6->ip6_plen = ntohs(tlen); 804 if (tp && tp->t_in6pcb) { 805 struct ifnet *oifp; 806 ro = (struct route *)&tp->t_in6pcb->in6p_route; 807 oifp = ro->ro_rt ? ro->ro_rt->rt_ifp : NULL; 808 ip6->ip6_hlim = in6_selecthlim(tp->t_in6pcb, oifp); 809 } else 810 ip6->ip6_hlim = ip6_defhlim; 811 ip6->ip6_flow &= ~IPV6_FLOWINFO_MASK; 812 if (ip6_auto_flowlabel) { 813 ip6->ip6_flow |= 814 (htonl(ip6_randomflowlabel()) & IPV6_FLOWLABEL_MASK); 815 } 816 break; 817 } 818 #endif 819 } 820 821 if (tp && tp->t_inpcb) 822 so = tp->t_inpcb->inp_socket; 823 #ifdef INET6 824 else if (tp && tp->t_in6pcb) 825 so = tp->t_in6pcb->in6p_socket; 826 #endif 827 else 828 so = NULL; 829 830 if (tp != NULL && tp->t_inpcb != NULL) { 831 ro = &tp->t_inpcb->inp_route; 832 #ifdef DIAGNOSTIC 833 if (family != AF_INET) 834 panic("tcp_respond: address family mismatch"); 835 if (!in_hosteq(ip->ip_dst, tp->t_inpcb->inp_faddr)) { 836 panic("tcp_respond: ip_dst %x != inp_faddr %x", 837 ntohl(ip->ip_dst.s_addr), 838 ntohl(tp->t_inpcb->inp_faddr.s_addr)); 839 } 840 #endif 841 } 842 #ifdef INET6 843 else if (tp != NULL && tp->t_in6pcb != NULL) { 844 ro = (struct route *)&tp->t_in6pcb->in6p_route; 845 #ifdef DIAGNOSTIC 846 if (family == AF_INET) { 847 if (!IN6_IS_ADDR_V4MAPPED(&tp->t_in6pcb->in6p_faddr)) 848 panic("tcp_respond: not mapped addr"); 849 if (bcmp(&ip->ip_dst, 850 &tp->t_in6pcb->in6p_faddr.s6_addr32[3], 851 sizeof(ip->ip_dst)) != 0) { 852 panic("tcp_respond: ip_dst != in6p_faddr"); 853 } 854 } else if (family == AF_INET6) { 855 if (!IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst, 856 &tp->t_in6pcb->in6p_faddr)) 857 panic("tcp_respond: ip6_dst != in6p_faddr"); 858 } else 859 panic("tcp_respond: address family mismatch"); 860 #endif 861 } 862 #endif 863 else 864 ro = NULL; 865 866 switch (family) { 867 #ifdef INET 868 case AF_INET: 869 error = ip_output(m, NULL, ro, 870 (tp && tp->t_mtudisc ? IP_MTUDISC : 0), 871 (struct ip_moptions *)0, so); 872 break; 873 #endif 874 #ifdef INET6 875 case AF_INET6: 876 error = ip6_output(m, NULL, (struct route_in6 *)ro, 0, 877 (struct ip6_moptions *)0, so, NULL); 878 break; 879 #endif 880 default: 881 error = EAFNOSUPPORT; 882 break; 883 } 884 885 return (error); 886 } 887 888 /* 889 * Template TCPCB. Rather than zeroing a new TCPCB and initializing 890 * a bunch of members individually, we maintain this template for the 891 * static and mostly-static components of the TCPCB, and copy it into 892 * the new TCPCB instead. 893 */ 894 static struct tcpcb tcpcb_template = { 895 /* 896 * If TCP_NTIMERS ever changes, we'll need to update this 897 * initializer. 898 */ 899 .t_timer = { 900 CALLOUT_INITIALIZER, 901 CALLOUT_INITIALIZER, 902 CALLOUT_INITIALIZER, 903 CALLOUT_INITIALIZER, 904 }, 905 .t_delack_ch = CALLOUT_INITIALIZER, 906 907 .t_srtt = TCPTV_SRTTBASE, 908 .t_rttmin = TCPTV_MIN, 909 910 .snd_cwnd = TCP_MAXWIN << TCP_MAX_WINSHIFT, 911 .snd_ssthresh = TCP_MAXWIN << TCP_MAX_WINSHIFT, 912 }; 913 914 /* 915 * Updates the TCPCB template whenever a parameter that would affect 916 * the template is changed. 917 */ 918 void 919 tcp_tcpcb_template(void) 920 { 921 struct tcpcb *tp = &tcpcb_template; 922 int flags; 923 924 tp->t_peermss = tcp_mssdflt; 925 tp->t_ourmss = tcp_mssdflt; 926 tp->t_segsz = tcp_mssdflt; 927 928 flags = 0; 929 if (tcp_do_rfc1323 && tcp_do_win_scale) 930 flags |= TF_REQ_SCALE; 931 if (tcp_do_rfc1323 && tcp_do_timestamps) 932 flags |= TF_REQ_TSTMP; 933 if (tcp_do_sack == 2) 934 flags |= TF_WILL_SACK; 935 else if (tcp_do_sack == 1) 936 flags |= TF_WILL_SACK|TF_IGNR_RXSACK; 937 flags |= TF_CANT_TXSACK; 938 tp->t_flags = flags; 939 940 /* 941 * Init srtt to TCPTV_SRTTBASE (0), so we can tell that we have no 942 * rtt estimate. Set rttvar so that srtt + 2 * rttvar gives 943 * reasonable initial retransmit time. 944 */ 945 tp->t_rttvar = tcp_rttdflt * PR_SLOWHZ << (TCP_RTTVAR_SHIFT + 2 - 1); 946 TCPT_RANGESET(tp->t_rxtcur, TCP_REXMTVAL(tp), 947 TCPTV_MIN, TCPTV_REXMTMAX); 948 } 949 950 /* 951 * Create a new TCP control block, making an 952 * empty reassembly queue and hooking it to the argument 953 * protocol control block. 954 */ 955 struct tcpcb * 956 tcp_newtcpcb(family, aux) 957 int family; /* selects inpcb, or in6pcb */ 958 void *aux; 959 { 960 struct tcpcb *tp; 961 int i; 962 963 /* XXX Consider using a pool_cache for speed. */ 964 tp = pool_get(&tcpcb_pool, PR_NOWAIT); 965 if (tp == NULL) 966 return (NULL); 967 memcpy(tp, &tcpcb_template, sizeof(*tp)); 968 TAILQ_INIT(&tp->segq); 969 TAILQ_INIT(&tp->timeq); 970 tp->t_family = family; /* may be overridden later on */ 971 LIST_INIT(&tp->t_sc); /* XXX can template this */ 972 973 /* Don't sweat this loop; hopefully the compiler will unroll it. */ 974 for (i = 0; i < TCPT_NTIMERS; i++) 975 TCP_TIMER_INIT(tp, i); 976 977 switch (family) { 978 case AF_INET: 979 { 980 struct inpcb *inp = (struct inpcb *)aux; 981 982 inp->inp_ip.ip_ttl = ip_defttl; 983 inp->inp_ppcb = (caddr_t)tp; 984 985 tp->t_inpcb = inp; 986 tp->t_mtudisc = ip_mtudisc; 987 break; 988 } 989 #ifdef INET6 990 case AF_INET6: 991 { 992 struct in6pcb *in6p = (struct in6pcb *)aux; 993 994 in6p->in6p_ip6.ip6_hlim = in6_selecthlim(in6p, 995 in6p->in6p_route.ro_rt ? in6p->in6p_route.ro_rt->rt_ifp 996 : NULL); 997 in6p->in6p_ppcb = (caddr_t)tp; 998 999 tp->t_in6pcb = in6p; 1000 /* for IPv6, always try to run path MTU discovery */ 1001 tp->t_mtudisc = 1; 1002 break; 1003 } 1004 #endif /* INET6 */ 1005 default: 1006 pool_put(&tcpcb_pool, tp); 1007 return (NULL); 1008 } 1009 1010 /* 1011 * Initialize our timebase. When we send timestamps, we take 1012 * the delta from tcp_now -- this means each connection always 1013 * gets a timebase of 0, which makes it, among other things, 1014 * more difficult to determine how long a system has been up, 1015 * and thus how many TCP sequence increments have occurred. 1016 */ 1017 tp->ts_timebase = tcp_now; 1018 1019 return (tp); 1020 } 1021 1022 /* 1023 * Drop a TCP connection, reporting 1024 * the specified error. If connection is synchronized, 1025 * then send a RST to peer. 1026 */ 1027 struct tcpcb * 1028 tcp_drop(tp, errno) 1029 struct tcpcb *tp; 1030 int errno; 1031 { 1032 struct socket *so = NULL; 1033 1034 #ifdef DIAGNOSTIC 1035 if (tp->t_inpcb && tp->t_in6pcb) 1036 panic("tcp_drop: both t_inpcb and t_in6pcb are set"); 1037 #endif 1038 #ifdef INET 1039 if (tp->t_inpcb) 1040 so = tp->t_inpcb->inp_socket; 1041 #endif 1042 #ifdef INET6 1043 if (tp->t_in6pcb) 1044 so = tp->t_in6pcb->in6p_socket; 1045 #endif 1046 if (!so) 1047 return NULL; 1048 1049 if (TCPS_HAVERCVDSYN(tp->t_state)) { 1050 tp->t_state = TCPS_CLOSED; 1051 (void) tcp_output(tp); 1052 tcpstat.tcps_drops++; 1053 } else 1054 tcpstat.tcps_conndrops++; 1055 if (errno == ETIMEDOUT && tp->t_softerror) 1056 errno = tp->t_softerror; 1057 so->so_error = errno; 1058 return (tcp_close(tp)); 1059 } 1060 1061 /* 1062 * Return whether this tcpcb is marked as dead, indicating 1063 * to the calling timer function that no further action should 1064 * be taken, as we are about to release this tcpcb. The release 1065 * of the storage will be done if this is the last timer running. 1066 * 1067 * This should be called from the callout handler function after 1068 * callout_ack() is done, so that the number of invoking timer 1069 * functions is 0. 1070 */ 1071 int 1072 tcp_isdead(tp) 1073 struct tcpcb *tp; 1074 { 1075 int dead = (tp->t_flags & TF_DEAD); 1076 1077 if (__predict_false(dead)) { 1078 if (tcp_timers_invoking(tp) > 0) 1079 /* not quite there yet -- count separately? */ 1080 return dead; 1081 tcpstat.tcps_delayed_free++; 1082 pool_put(&tcpcb_pool, tp); 1083 } 1084 return dead; 1085 } 1086 1087 /* 1088 * Close a TCP control block: 1089 * discard all space held by the tcp 1090 * discard internet protocol block 1091 * wake up any sleepers 1092 */ 1093 struct tcpcb * 1094 tcp_close(tp) 1095 struct tcpcb *tp; 1096 { 1097 struct inpcb *inp; 1098 #ifdef INET6 1099 struct in6pcb *in6p; 1100 #endif 1101 struct socket *so; 1102 #ifdef RTV_RTT 1103 struct rtentry *rt; 1104 #endif 1105 struct route *ro; 1106 1107 inp = tp->t_inpcb; 1108 #ifdef INET6 1109 in6p = tp->t_in6pcb; 1110 #endif 1111 so = NULL; 1112 ro = NULL; 1113 if (inp) { 1114 so = inp->inp_socket; 1115 ro = &inp->inp_route; 1116 } 1117 #ifdef INET6 1118 else if (in6p) { 1119 so = in6p->in6p_socket; 1120 ro = (struct route *)&in6p->in6p_route; 1121 } 1122 #endif 1123 1124 #ifdef RTV_RTT 1125 /* 1126 * If we sent enough data to get some meaningful characteristics, 1127 * save them in the routing entry. 'Enough' is arbitrarily 1128 * defined as the sendpipesize (default 4K) * 16. This would 1129 * give us 16 rtt samples assuming we only get one sample per 1130 * window (the usual case on a long haul net). 16 samples is 1131 * enough for the srtt filter to converge to within 5% of the correct 1132 * value; fewer samples and we could save a very bogus rtt. 1133 * 1134 * Don't update the default route's characteristics and don't 1135 * update anything that the user "locked". 1136 */ 1137 if (SEQ_LT(tp->iss + so->so_snd.sb_hiwat * 16, tp->snd_max) && 1138 ro && (rt = ro->ro_rt) && 1139 !in_nullhost(satosin(rt_key(rt))->sin_addr)) { 1140 u_long i = 0; 1141 1142 if ((rt->rt_rmx.rmx_locks & RTV_RTT) == 0) { 1143 i = tp->t_srtt * 1144 ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTT_SHIFT + 2)); 1145 if (rt->rt_rmx.rmx_rtt && i) 1146 /* 1147 * filter this update to half the old & half 1148 * the new values, converting scale. 1149 * See route.h and tcp_var.h for a 1150 * description of the scaling constants. 1151 */ 1152 rt->rt_rmx.rmx_rtt = 1153 (rt->rt_rmx.rmx_rtt + i) / 2; 1154 else 1155 rt->rt_rmx.rmx_rtt = i; 1156 } 1157 if ((rt->rt_rmx.rmx_locks & RTV_RTTVAR) == 0) { 1158 i = tp->t_rttvar * 1159 ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTTVAR_SHIFT + 2)); 1160 if (rt->rt_rmx.rmx_rttvar && i) 1161 rt->rt_rmx.rmx_rttvar = 1162 (rt->rt_rmx.rmx_rttvar + i) / 2; 1163 else 1164 rt->rt_rmx.rmx_rttvar = i; 1165 } 1166 /* 1167 * update the pipelimit (ssthresh) if it has been updated 1168 * already or if a pipesize was specified & the threshhold 1169 * got below half the pipesize. I.e., wait for bad news 1170 * before we start updating, then update on both good 1171 * and bad news. 1172 */ 1173 if (((rt->rt_rmx.rmx_locks & RTV_SSTHRESH) == 0 && 1174 (i = tp->snd_ssthresh) && rt->rt_rmx.rmx_ssthresh) || 1175 i < (rt->rt_rmx.rmx_sendpipe / 2)) { 1176 /* 1177 * convert the limit from user data bytes to 1178 * packets then to packet data bytes. 1179 */ 1180 i = (i + tp->t_segsz / 2) / tp->t_segsz; 1181 if (i < 2) 1182 i = 2; 1183 i *= (u_long)(tp->t_segsz + sizeof (struct tcpiphdr)); 1184 if (rt->rt_rmx.rmx_ssthresh) 1185 rt->rt_rmx.rmx_ssthresh = 1186 (rt->rt_rmx.rmx_ssthresh + i) / 2; 1187 else 1188 rt->rt_rmx.rmx_ssthresh = i; 1189 } 1190 } 1191 #endif /* RTV_RTT */ 1192 /* free the reassembly queue, if any */ 1193 TCP_REASS_LOCK(tp); 1194 (void) tcp_freeq(tp); 1195 TCP_REASS_UNLOCK(tp); 1196 1197 tcp_canceltimers(tp); 1198 TCP_CLEAR_DELACK(tp); 1199 syn_cache_cleanup(tp); 1200 1201 if (tp->t_template) { 1202 m_free(tp->t_template); 1203 tp->t_template = NULL; 1204 } 1205 if (tcp_timers_invoking(tp)) 1206 tp->t_flags |= TF_DEAD; 1207 else 1208 pool_put(&tcpcb_pool, tp); 1209 1210 if (inp) { 1211 inp->inp_ppcb = 0; 1212 soisdisconnected(so); 1213 in_pcbdetach(inp); 1214 } 1215 #ifdef INET6 1216 else if (in6p) { 1217 in6p->in6p_ppcb = 0; 1218 soisdisconnected(so); 1219 in6_pcbdetach(in6p); 1220 } 1221 #endif 1222 tcpstat.tcps_closed++; 1223 return ((struct tcpcb *)0); 1224 } 1225 1226 int 1227 tcp_freeq(tp) 1228 struct tcpcb *tp; 1229 { 1230 struct ipqent *qe; 1231 int rv = 0; 1232 #ifdef TCPREASS_DEBUG 1233 int i = 0; 1234 #endif 1235 1236 TCP_REASS_LOCK_CHECK(tp); 1237 1238 while ((qe = TAILQ_FIRST(&tp->segq)) != NULL) { 1239 #ifdef TCPREASS_DEBUG 1240 printf("tcp_freeq[%p,%d]: %u:%u(%u) 0x%02x\n", 1241 tp, i++, qe->ipqe_seq, qe->ipqe_seq + qe->ipqe_len, 1242 qe->ipqe_len, qe->ipqe_flags & (TH_SYN|TH_FIN|TH_RST)); 1243 #endif 1244 TAILQ_REMOVE(&tp->segq, qe, ipqe_q); 1245 TAILQ_REMOVE(&tp->timeq, qe, ipqe_timeq); 1246 m_freem(qe->ipqe_m); 1247 pool_put(&tcpipqent_pool, qe); 1248 rv = 1; 1249 } 1250 return (rv); 1251 } 1252 1253 /* 1254 * Protocol drain routine. Called when memory is in short supply. 1255 */ 1256 void 1257 tcp_drain() 1258 { 1259 struct inpcb_hdr *inph; 1260 struct tcpcb *tp; 1261 1262 /* 1263 * Free the sequence queue of all TCP connections. 1264 */ 1265 CIRCLEQ_FOREACH(inph, &tcbtable.inpt_queue, inph_queue) { 1266 switch (inph->inph_af) { 1267 case AF_INET: 1268 tp = intotcpcb((struct inpcb *)inph); 1269 break; 1270 #ifdef INET6 1271 case AF_INET6: 1272 tp = in6totcpcb((struct in6pcb *)inph); 1273 break; 1274 #endif 1275 default: 1276 tp = NULL; 1277 break; 1278 } 1279 if (tp != NULL) { 1280 /* 1281 * We may be called from a device's interrupt 1282 * context. If the tcpcb is already busy, 1283 * just bail out now. 1284 */ 1285 if (tcp_reass_lock_try(tp) == 0) 1286 continue; 1287 if (tcp_freeq(tp)) 1288 tcpstat.tcps_connsdrained++; 1289 TCP_REASS_UNLOCK(tp); 1290 } 1291 } 1292 } 1293 1294 /* 1295 * Notify a tcp user of an asynchronous error; 1296 * store error as soft error, but wake up user 1297 * (for now, won't do anything until can select for soft error). 1298 */ 1299 void 1300 tcp_notify(inp, error) 1301 struct inpcb *inp; 1302 int error; 1303 { 1304 struct tcpcb *tp = (struct tcpcb *)inp->inp_ppcb; 1305 struct socket *so = inp->inp_socket; 1306 1307 /* 1308 * Ignore some errors if we are hooked up. 1309 * If connection hasn't completed, has retransmitted several times, 1310 * and receives a second error, give up now. This is better 1311 * than waiting a long time to establish a connection that 1312 * can never complete. 1313 */ 1314 if (tp->t_state == TCPS_ESTABLISHED && 1315 (error == EHOSTUNREACH || error == ENETUNREACH || 1316 error == EHOSTDOWN)) { 1317 return; 1318 } else if (TCPS_HAVEESTABLISHED(tp->t_state) == 0 && 1319 tp->t_rxtshift > 3 && tp->t_softerror) 1320 so->so_error = error; 1321 else 1322 tp->t_softerror = error; 1323 wakeup((caddr_t) &so->so_timeo); 1324 sorwakeup(so); 1325 sowwakeup(so); 1326 } 1327 1328 #ifdef INET6 1329 void 1330 tcp6_notify(in6p, error) 1331 struct in6pcb *in6p; 1332 int error; 1333 { 1334 struct tcpcb *tp = (struct tcpcb *)in6p->in6p_ppcb; 1335 struct socket *so = in6p->in6p_socket; 1336 1337 /* 1338 * Ignore some errors if we are hooked up. 1339 * If connection hasn't completed, has retransmitted several times, 1340 * and receives a second error, give up now. This is better 1341 * than waiting a long time to establish a connection that 1342 * can never complete. 1343 */ 1344 if (tp->t_state == TCPS_ESTABLISHED && 1345 (error == EHOSTUNREACH || error == ENETUNREACH || 1346 error == EHOSTDOWN)) { 1347 return; 1348 } else if (TCPS_HAVEESTABLISHED(tp->t_state) == 0 && 1349 tp->t_rxtshift > 3 && tp->t_softerror) 1350 so->so_error = error; 1351 else 1352 tp->t_softerror = error; 1353 wakeup((caddr_t) &so->so_timeo); 1354 sorwakeup(so); 1355 sowwakeup(so); 1356 } 1357 #endif 1358 1359 #ifdef INET6 1360 void 1361 tcp6_ctlinput(cmd, sa, d) 1362 int cmd; 1363 struct sockaddr *sa; 1364 void *d; 1365 { 1366 struct tcphdr th; 1367 void (*notify) __P((struct in6pcb *, int)) = tcp6_notify; 1368 int nmatch; 1369 struct ip6_hdr *ip6; 1370 const struct sockaddr_in6 *sa6_src = NULL; 1371 struct sockaddr_in6 *sa6 = (struct sockaddr_in6 *)sa; 1372 struct mbuf *m; 1373 int off; 1374 1375 if (sa->sa_family != AF_INET6 || 1376 sa->sa_len != sizeof(struct sockaddr_in6)) 1377 return; 1378 if ((unsigned)cmd >= PRC_NCMDS) 1379 return; 1380 else if (cmd == PRC_QUENCH) { 1381 /* XXX there's no PRC_QUENCH in IPv6 */ 1382 notify = tcp6_quench; 1383 } else if (PRC_IS_REDIRECT(cmd)) 1384 notify = in6_rtchange, d = NULL; 1385 else if (cmd == PRC_MSGSIZE) 1386 ; /* special code is present, see below */ 1387 else if (cmd == PRC_HOSTDEAD) 1388 d = NULL; 1389 else if (inet6ctlerrmap[cmd] == 0) 1390 return; 1391 1392 /* if the parameter is from icmp6, decode it. */ 1393 if (d != NULL) { 1394 struct ip6ctlparam *ip6cp = (struct ip6ctlparam *)d; 1395 m = ip6cp->ip6c_m; 1396 ip6 = ip6cp->ip6c_ip6; 1397 off = ip6cp->ip6c_off; 1398 sa6_src = ip6cp->ip6c_src; 1399 } else { 1400 m = NULL; 1401 ip6 = NULL; 1402 sa6_src = &sa6_any; 1403 off = 0; 1404 } 1405 1406 if (ip6) { 1407 /* 1408 * XXX: We assume that when ip6 is non NULL, 1409 * M and OFF are valid. 1410 */ 1411 1412 /* check if we can safely examine src and dst ports */ 1413 if (m->m_pkthdr.len < off + sizeof(th)) { 1414 if (cmd == PRC_MSGSIZE) 1415 icmp6_mtudisc_update((struct ip6ctlparam *)d, 0); 1416 return; 1417 } 1418 1419 bzero(&th, sizeof(th)); 1420 m_copydata(m, off, sizeof(th), (caddr_t)&th); 1421 1422 if (cmd == PRC_MSGSIZE) { 1423 int valid = 0; 1424 1425 /* 1426 * Check to see if we have a valid TCP connection 1427 * corresponding to the address in the ICMPv6 message 1428 * payload. 1429 */ 1430 if (in6_pcblookup_connect(&tcbtable, &sa6->sin6_addr, 1431 th.th_dport, (struct in6_addr *)&sa6_src->sin6_addr, 1432 th.th_sport, 0)) 1433 valid++; 1434 1435 /* 1436 * Depending on the value of "valid" and routing table 1437 * size (mtudisc_{hi,lo}wat), we will: 1438 * - recalcurate the new MTU and create the 1439 * corresponding routing entry, or 1440 * - ignore the MTU change notification. 1441 */ 1442 icmp6_mtudisc_update((struct ip6ctlparam *)d, valid); 1443 1444 /* 1445 * no need to call in6_pcbnotify, it should have been 1446 * called via callback if necessary 1447 */ 1448 return; 1449 } 1450 1451 nmatch = in6_pcbnotify(&tcbtable, sa, th.th_dport, 1452 (struct sockaddr *)sa6_src, th.th_sport, cmd, NULL, notify); 1453 if (nmatch == 0 && syn_cache_count && 1454 (inet6ctlerrmap[cmd] == EHOSTUNREACH || 1455 inet6ctlerrmap[cmd] == ENETUNREACH || 1456 inet6ctlerrmap[cmd] == EHOSTDOWN)) 1457 syn_cache_unreach((struct sockaddr *)sa6_src, 1458 sa, &th); 1459 } else { 1460 (void) in6_pcbnotify(&tcbtable, sa, 0, 1461 (struct sockaddr *)sa6_src, 0, cmd, NULL, notify); 1462 } 1463 } 1464 #endif 1465 1466 #ifdef INET 1467 /* assumes that ip header and tcp header are contiguous on mbuf */ 1468 void * 1469 tcp_ctlinput(cmd, sa, v) 1470 int cmd; 1471 struct sockaddr *sa; 1472 void *v; 1473 { 1474 struct ip *ip = v; 1475 struct tcphdr *th; 1476 struct icmp *icp; 1477 extern const int inetctlerrmap[]; 1478 void (*notify) __P((struct inpcb *, int)) = tcp_notify; 1479 int errno; 1480 int nmatch; 1481 #ifdef INET6 1482 struct in6_addr src6, dst6; 1483 #endif 1484 1485 if (sa->sa_family != AF_INET || 1486 sa->sa_len != sizeof(struct sockaddr_in)) 1487 return NULL; 1488 if ((unsigned)cmd >= PRC_NCMDS) 1489 return NULL; 1490 errno = inetctlerrmap[cmd]; 1491 if (cmd == PRC_QUENCH) 1492 notify = tcp_quench; 1493 else if (PRC_IS_REDIRECT(cmd)) 1494 notify = in_rtchange, ip = 0; 1495 else if (cmd == PRC_MSGSIZE && ip && ip->ip_v == 4) { 1496 /* 1497 * Check to see if we have a valid TCP connection 1498 * corresponding to the address in the ICMP message 1499 * payload. 1500 * 1501 * Boundary check is made in icmp_input(), with ICMP_ADVLENMIN. 1502 */ 1503 th = (struct tcphdr *)((caddr_t)ip + (ip->ip_hl << 2)); 1504 #ifdef INET6 1505 memset(&src6, 0, sizeof(src6)); 1506 memset(&dst6, 0, sizeof(dst6)); 1507 src6.s6_addr16[5] = dst6.s6_addr16[5] = 0xffff; 1508 memcpy(&src6.s6_addr32[3], &ip->ip_src, sizeof(struct in_addr)); 1509 memcpy(&dst6.s6_addr32[3], &ip->ip_dst, sizeof(struct in_addr)); 1510 #endif 1511 if (in_pcblookup_connect(&tcbtable, ip->ip_dst, th->th_dport, 1512 ip->ip_src, th->th_sport) != NULL) 1513 ; 1514 #ifdef INET6 1515 else if (in6_pcblookup_connect(&tcbtable, &dst6, 1516 th->th_dport, &src6, th->th_sport, 0) != NULL) 1517 ; 1518 #endif 1519 else 1520 return NULL; 1521 1522 /* 1523 * Now that we've validated that we are actually communicating 1524 * with the host indicated in the ICMP message, locate the 1525 * ICMP header, recalculate the new MTU, and create the 1526 * corresponding routing entry. 1527 */ 1528 icp = (struct icmp *)((caddr_t)ip - 1529 offsetof(struct icmp, icmp_ip)); 1530 icmp_mtudisc(icp, ip->ip_dst); 1531 1532 return NULL; 1533 } else if (cmd == PRC_HOSTDEAD) 1534 ip = 0; 1535 else if (errno == 0) 1536 return NULL; 1537 if (ip && ip->ip_v == 4 && sa->sa_family == AF_INET) { 1538 th = (struct tcphdr *)((caddr_t)ip + (ip->ip_hl << 2)); 1539 nmatch = in_pcbnotify(&tcbtable, satosin(sa)->sin_addr, 1540 th->th_dport, ip->ip_src, th->th_sport, errno, notify); 1541 if (nmatch == 0 && syn_cache_count && 1542 (inetctlerrmap[cmd] == EHOSTUNREACH || 1543 inetctlerrmap[cmd] == ENETUNREACH || 1544 inetctlerrmap[cmd] == EHOSTDOWN)) { 1545 struct sockaddr_in sin; 1546 bzero(&sin, sizeof(sin)); 1547 sin.sin_len = sizeof(sin); 1548 sin.sin_family = AF_INET; 1549 sin.sin_port = th->th_sport; 1550 sin.sin_addr = ip->ip_src; 1551 syn_cache_unreach((struct sockaddr *)&sin, sa, th); 1552 } 1553 1554 /* XXX mapped address case */ 1555 } else 1556 in_pcbnotifyall(&tcbtable, satosin(sa)->sin_addr, errno, 1557 notify); 1558 return NULL; 1559 } 1560 1561 /* 1562 * When a source quence is received, we are being notifed of congestion. 1563 * Close the congestion window down to the Loss Window (one segment). 1564 * We will gradually open it again as we proceed. 1565 */ 1566 void 1567 tcp_quench(inp, errno) 1568 struct inpcb *inp; 1569 int errno; 1570 { 1571 struct tcpcb *tp = intotcpcb(inp); 1572 1573 if (tp) 1574 tp->snd_cwnd = tp->t_segsz; 1575 } 1576 #endif 1577 1578 #ifdef INET6 1579 void 1580 tcp6_quench(in6p, errno) 1581 struct in6pcb *in6p; 1582 int errno; 1583 { 1584 struct tcpcb *tp = in6totcpcb(in6p); 1585 1586 if (tp) 1587 tp->snd_cwnd = tp->t_segsz; 1588 } 1589 #endif 1590 1591 #ifdef INET 1592 /* 1593 * Path MTU Discovery handlers. 1594 */ 1595 void 1596 tcp_mtudisc_callback(faddr) 1597 struct in_addr faddr; 1598 { 1599 #ifdef INET6 1600 struct in6_addr in6; 1601 #endif 1602 1603 in_pcbnotifyall(&tcbtable, faddr, EMSGSIZE, tcp_mtudisc); 1604 #ifdef INET6 1605 memset(&in6, 0, sizeof(in6)); 1606 in6.s6_addr16[5] = 0xffff; 1607 memcpy(&in6.s6_addr32[3], &faddr, sizeof(struct in_addr)); 1608 tcp6_mtudisc_callback(&in6); 1609 #endif 1610 } 1611 1612 /* 1613 * On receipt of path MTU corrections, flush old route and replace it 1614 * with the new one. Retransmit all unacknowledged packets, to ensure 1615 * that all packets will be received. 1616 */ 1617 void 1618 tcp_mtudisc(inp, errno) 1619 struct inpcb *inp; 1620 int errno; 1621 { 1622 struct tcpcb *tp = intotcpcb(inp); 1623 struct rtentry *rt = in_pcbrtentry(inp); 1624 1625 if (tp != 0) { 1626 if (rt != 0) { 1627 /* 1628 * If this was not a host route, remove and realloc. 1629 */ 1630 if ((rt->rt_flags & RTF_HOST) == 0) { 1631 in_rtchange(inp, errno); 1632 if ((rt = in_pcbrtentry(inp)) == 0) 1633 return; 1634 } 1635 1636 /* 1637 * Slow start out of the error condition. We 1638 * use the MTU because we know it's smaller 1639 * than the previously transmitted segment. 1640 * 1641 * Note: This is more conservative than the 1642 * suggestion in draft-floyd-incr-init-win-03. 1643 */ 1644 if (rt->rt_rmx.rmx_mtu != 0) 1645 tp->snd_cwnd = 1646 TCP_INITIAL_WINDOW(tcp_init_win, 1647 rt->rt_rmx.rmx_mtu); 1648 } 1649 1650 /* 1651 * Resend unacknowledged packets. 1652 */ 1653 tp->snd_nxt = tp->snd_una; 1654 tcp_output(tp); 1655 } 1656 } 1657 #endif 1658 1659 #ifdef INET6 1660 /* 1661 * Path MTU Discovery handlers. 1662 */ 1663 void 1664 tcp6_mtudisc_callback(faddr) 1665 struct in6_addr *faddr; 1666 { 1667 struct sockaddr_in6 sin6; 1668 1669 bzero(&sin6, sizeof(sin6)); 1670 sin6.sin6_family = AF_INET6; 1671 sin6.sin6_len = sizeof(struct sockaddr_in6); 1672 sin6.sin6_addr = *faddr; 1673 (void) in6_pcbnotify(&tcbtable, (struct sockaddr *)&sin6, 0, 1674 (struct sockaddr *)&sa6_any, 0, PRC_MSGSIZE, NULL, tcp6_mtudisc); 1675 } 1676 1677 void 1678 tcp6_mtudisc(in6p, errno) 1679 struct in6pcb *in6p; 1680 int errno; 1681 { 1682 struct tcpcb *tp = in6totcpcb(in6p); 1683 struct rtentry *rt = in6_pcbrtentry(in6p); 1684 1685 if (tp != 0) { 1686 if (rt != 0) { 1687 /* 1688 * If this was not a host route, remove and realloc. 1689 */ 1690 if ((rt->rt_flags & RTF_HOST) == 0) { 1691 in6_rtchange(in6p, errno); 1692 if ((rt = in6_pcbrtentry(in6p)) == 0) 1693 return; 1694 } 1695 1696 /* 1697 * Slow start out of the error condition. We 1698 * use the MTU because we know it's smaller 1699 * than the previously transmitted segment. 1700 * 1701 * Note: This is more conservative than the 1702 * suggestion in draft-floyd-incr-init-win-03. 1703 */ 1704 if (rt->rt_rmx.rmx_mtu != 0) 1705 tp->snd_cwnd = 1706 TCP_INITIAL_WINDOW(tcp_init_win, 1707 rt->rt_rmx.rmx_mtu); 1708 } 1709 1710 /* 1711 * Resend unacknowledged packets. 1712 */ 1713 tp->snd_nxt = tp->snd_una; 1714 tcp_output(tp); 1715 } 1716 } 1717 #endif /* INET6 */ 1718 1719 /* 1720 * Compute the MSS to advertise to the peer. Called only during 1721 * the 3-way handshake. If we are the server (peer initiated 1722 * connection), we are called with a pointer to the interface 1723 * on which the SYN packet arrived. If we are the client (we 1724 * initiated connection), we are called with a pointer to the 1725 * interface out which this connection should go. 1726 * 1727 * NOTE: Do not subtract IP option/extension header size nor IPsec 1728 * header size from MSS advertisement. MSS option must hold the maximum 1729 * segment size we can accept, so it must always be: 1730 * max(if mtu) - ip header - tcp header 1731 */ 1732 u_long 1733 tcp_mss_to_advertise(ifp, af) 1734 const struct ifnet *ifp; 1735 int af; 1736 { 1737 extern u_long in_maxmtu; 1738 u_long mss = 0; 1739 u_long hdrsiz; 1740 1741 /* 1742 * In order to avoid defeating path MTU discovery on the peer, 1743 * we advertise the max MTU of all attached networks as our MSS, 1744 * per RFC 1191, section 3.1. 1745 * 1746 * We provide the option to advertise just the MTU of 1747 * the interface on which we hope this connection will 1748 * be receiving. If we are responding to a SYN, we 1749 * will have a pretty good idea about this, but when 1750 * initiating a connection there is a bit more doubt. 1751 * 1752 * We also need to ensure that loopback has a large enough 1753 * MSS, as the loopback MTU is never included in in_maxmtu. 1754 */ 1755 1756 if (ifp != NULL) 1757 switch (af) { 1758 case AF_INET: 1759 mss = ifp->if_mtu; 1760 break; 1761 #ifdef INET6 1762 case AF_INET6: 1763 mss = IN6_LINKMTU(ifp); 1764 break; 1765 #endif 1766 } 1767 1768 if (tcp_mss_ifmtu == 0) 1769 switch (af) { 1770 case AF_INET: 1771 mss = max(in_maxmtu, mss); 1772 break; 1773 #ifdef INET6 1774 case AF_INET6: 1775 mss = max(in6_maxmtu, mss); 1776 break; 1777 #endif 1778 } 1779 1780 switch (af) { 1781 case AF_INET: 1782 hdrsiz = sizeof(struct ip); 1783 break; 1784 #ifdef INET6 1785 case AF_INET6: 1786 hdrsiz = sizeof(struct ip6_hdr); 1787 break; 1788 #endif 1789 default: 1790 hdrsiz = 0; 1791 break; 1792 } 1793 hdrsiz += sizeof(struct tcphdr); 1794 if (mss > hdrsiz) 1795 mss -= hdrsiz; 1796 1797 mss = max(tcp_mssdflt, mss); 1798 return (mss); 1799 } 1800 1801 /* 1802 * Set connection variables based on the peer's advertised MSS. 1803 * We are passed the TCPCB for the actual connection. If we 1804 * are the server, we are called by the compressed state engine 1805 * when the 3-way handshake is complete. If we are the client, 1806 * we are called when we receive the SYN,ACK from the server. 1807 * 1808 * NOTE: Our advertised MSS value must be initialized in the TCPCB 1809 * before this routine is called! 1810 */ 1811 void 1812 tcp_mss_from_peer(tp, offer) 1813 struct tcpcb *tp; 1814 int offer; 1815 { 1816 struct socket *so; 1817 #if defined(RTV_SPIPE) || defined(RTV_SSTHRESH) 1818 struct rtentry *rt; 1819 #endif 1820 u_long bufsize; 1821 int mss; 1822 1823 #ifdef DIAGNOSTIC 1824 if (tp->t_inpcb && tp->t_in6pcb) 1825 panic("tcp_mss_from_peer: both t_inpcb and t_in6pcb are set"); 1826 #endif 1827 so = NULL; 1828 rt = NULL; 1829 #ifdef INET 1830 if (tp->t_inpcb) { 1831 so = tp->t_inpcb->inp_socket; 1832 #if defined(RTV_SPIPE) || defined(RTV_SSTHRESH) 1833 rt = in_pcbrtentry(tp->t_inpcb); 1834 #endif 1835 } 1836 #endif 1837 #ifdef INET6 1838 if (tp->t_in6pcb) { 1839 so = tp->t_in6pcb->in6p_socket; 1840 #if defined(RTV_SPIPE) || defined(RTV_SSTHRESH) 1841 rt = in6_pcbrtentry(tp->t_in6pcb); 1842 #endif 1843 } 1844 #endif 1845 1846 /* 1847 * As per RFC1122, use the default MSS value, unless they 1848 * sent us an offer. Do not accept offers less than 256 bytes. 1849 */ 1850 mss = tcp_mssdflt; 1851 if (offer) 1852 mss = offer; 1853 mss = max(mss, 256); /* sanity */ 1854 tp->t_peermss = mss; 1855 mss -= tcp_optlen(tp); 1856 #ifdef INET 1857 if (tp->t_inpcb) 1858 mss -= ip_optlen(tp->t_inpcb); 1859 #endif 1860 #ifdef INET6 1861 if (tp->t_in6pcb) 1862 mss -= ip6_optlen(tp->t_in6pcb); 1863 #endif 1864 1865 /* 1866 * If there's a pipesize, change the socket buffer to that size. 1867 * Make the socket buffer an integral number of MSS units. If 1868 * the MSS is larger than the socket buffer, artificially decrease 1869 * the MSS. 1870 */ 1871 #ifdef RTV_SPIPE 1872 if (rt != NULL && rt->rt_rmx.rmx_sendpipe != 0) 1873 bufsize = rt->rt_rmx.rmx_sendpipe; 1874 else 1875 #endif 1876 bufsize = so->so_snd.sb_hiwat; 1877 if (bufsize < mss) 1878 mss = bufsize; 1879 else { 1880 bufsize = roundup(bufsize, mss); 1881 if (bufsize > sb_max) 1882 bufsize = sb_max; 1883 (void) sbreserve(&so->so_snd, bufsize, so); 1884 } 1885 tp->t_segsz = mss; 1886 1887 #ifdef RTV_SSTHRESH 1888 if (rt != NULL && rt->rt_rmx.rmx_ssthresh) { 1889 /* 1890 * There's some sort of gateway or interface buffer 1891 * limit on the path. Use this to set the slow 1892 * start threshold, but set the threshold to no less 1893 * than 2 * MSS. 1894 */ 1895 tp->snd_ssthresh = max(2 * mss, rt->rt_rmx.rmx_ssthresh); 1896 } 1897 #endif 1898 } 1899 1900 /* 1901 * Processing necessary when a TCP connection is established. 1902 */ 1903 void 1904 tcp_established(tp) 1905 struct tcpcb *tp; 1906 { 1907 struct socket *so; 1908 #ifdef RTV_RPIPE 1909 struct rtentry *rt; 1910 #endif 1911 u_long bufsize; 1912 1913 #ifdef DIAGNOSTIC 1914 if (tp->t_inpcb && tp->t_in6pcb) 1915 panic("tcp_established: both t_inpcb and t_in6pcb are set"); 1916 #endif 1917 so = NULL; 1918 rt = NULL; 1919 #ifdef INET 1920 if (tp->t_inpcb) { 1921 so = tp->t_inpcb->inp_socket; 1922 #if defined(RTV_RPIPE) 1923 rt = in_pcbrtentry(tp->t_inpcb); 1924 #endif 1925 } 1926 #endif 1927 #ifdef INET6 1928 if (tp->t_in6pcb) { 1929 so = tp->t_in6pcb->in6p_socket; 1930 #if defined(RTV_RPIPE) 1931 rt = in6_pcbrtentry(tp->t_in6pcb); 1932 #endif 1933 } 1934 #endif 1935 1936 tp->t_state = TCPS_ESTABLISHED; 1937 TCP_TIMER_ARM(tp, TCPT_KEEP, tcp_keepidle); 1938 1939 #ifdef RTV_RPIPE 1940 if (rt != NULL && rt->rt_rmx.rmx_recvpipe != 0) 1941 bufsize = rt->rt_rmx.rmx_recvpipe; 1942 else 1943 #endif 1944 bufsize = so->so_rcv.sb_hiwat; 1945 if (bufsize > tp->t_ourmss) { 1946 bufsize = roundup(bufsize, tp->t_ourmss); 1947 if (bufsize > sb_max) 1948 bufsize = sb_max; 1949 (void) sbreserve(&so->so_rcv, bufsize, so); 1950 } 1951 } 1952 1953 /* 1954 * Check if there's an initial rtt or rttvar. Convert from the 1955 * route-table units to scaled multiples of the slow timeout timer. 1956 * Called only during the 3-way handshake. 1957 */ 1958 void 1959 tcp_rmx_rtt(tp) 1960 struct tcpcb *tp; 1961 { 1962 #ifdef RTV_RTT 1963 struct rtentry *rt = NULL; 1964 int rtt; 1965 1966 #ifdef DIAGNOSTIC 1967 if (tp->t_inpcb && tp->t_in6pcb) 1968 panic("tcp_rmx_rtt: both t_inpcb and t_in6pcb are set"); 1969 #endif 1970 #ifdef INET 1971 if (tp->t_inpcb) 1972 rt = in_pcbrtentry(tp->t_inpcb); 1973 #endif 1974 #ifdef INET6 1975 if (tp->t_in6pcb) 1976 rt = in6_pcbrtentry(tp->t_in6pcb); 1977 #endif 1978 if (rt == NULL) 1979 return; 1980 1981 if (tp->t_srtt == 0 && (rtt = rt->rt_rmx.rmx_rtt)) { 1982 /* 1983 * XXX The lock bit for MTU indicates that the value 1984 * is also a minimum value; this is subject to time. 1985 */ 1986 if (rt->rt_rmx.rmx_locks & RTV_RTT) 1987 TCPT_RANGESET(tp->t_rttmin, 1988 rtt / (RTM_RTTUNIT / PR_SLOWHZ), 1989 TCPTV_MIN, TCPTV_REXMTMAX); 1990 tp->t_srtt = rtt / 1991 ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTT_SHIFT + 2)); 1992 if (rt->rt_rmx.rmx_rttvar) { 1993 tp->t_rttvar = rt->rt_rmx.rmx_rttvar / 1994 ((RTM_RTTUNIT / PR_SLOWHZ) >> 1995 (TCP_RTTVAR_SHIFT + 2)); 1996 } else { 1997 /* Default variation is +- 1 rtt */ 1998 tp->t_rttvar = 1999 tp->t_srtt >> (TCP_RTT_SHIFT - TCP_RTTVAR_SHIFT); 2000 } 2001 TCPT_RANGESET(tp->t_rxtcur, 2002 ((tp->t_srtt >> 2) + tp->t_rttvar) >> (1 + 2), 2003 tp->t_rttmin, TCPTV_REXMTMAX); 2004 } 2005 #endif 2006 } 2007 2008 tcp_seq tcp_iss_seq = 0; /* tcp initial seq # */ 2009 #if NRND > 0 2010 u_int8_t tcp_iss_secret[16]; /* 128 bits; should be plenty */ 2011 #endif 2012 2013 /* 2014 * Get a new sequence value given a tcp control block 2015 */ 2016 tcp_seq 2017 tcp_new_iss(struct tcpcb *tp, tcp_seq addin) 2018 { 2019 2020 #ifdef INET 2021 if (tp->t_inpcb != NULL) { 2022 return (tcp_new_iss1(&tp->t_inpcb->inp_laddr, 2023 &tp->t_inpcb->inp_faddr, tp->t_inpcb->inp_lport, 2024 tp->t_inpcb->inp_fport, sizeof(tp->t_inpcb->inp_laddr), 2025 addin)); 2026 } 2027 #endif 2028 #ifdef INET6 2029 if (tp->t_in6pcb != NULL) { 2030 return (tcp_new_iss1(&tp->t_in6pcb->in6p_laddr, 2031 &tp->t_in6pcb->in6p_faddr, tp->t_in6pcb->in6p_lport, 2032 tp->t_in6pcb->in6p_fport, sizeof(tp->t_in6pcb->in6p_laddr), 2033 addin)); 2034 } 2035 #endif 2036 /* Not possible. */ 2037 panic("tcp_new_iss"); 2038 } 2039 2040 /* 2041 * This routine actually generates a new TCP initial sequence number. 2042 */ 2043 tcp_seq 2044 tcp_new_iss1(void *laddr, void *faddr, u_int16_t lport, u_int16_t fport, 2045 size_t addrsz, tcp_seq addin) 2046 { 2047 tcp_seq tcp_iss; 2048 2049 #if NRND > 0 2050 static int beenhere; 2051 2052 /* 2053 * If we haven't been here before, initialize our cryptographic 2054 * hash secret. 2055 */ 2056 if (beenhere == 0) { 2057 rnd_extract_data(tcp_iss_secret, sizeof(tcp_iss_secret), 2058 RND_EXTRACT_ANY); 2059 beenhere = 1; 2060 } 2061 2062 if (tcp_do_rfc1948) { 2063 MD5_CTX ctx; 2064 u_int8_t hash[16]; /* XXX MD5 knowledge */ 2065 2066 /* 2067 * Compute the base value of the ISS. It is a hash 2068 * of (saddr, sport, daddr, dport, secret). 2069 */ 2070 MD5Init(&ctx); 2071 2072 MD5Update(&ctx, (u_char *) laddr, addrsz); 2073 MD5Update(&ctx, (u_char *) &lport, sizeof(lport)); 2074 2075 MD5Update(&ctx, (u_char *) faddr, addrsz); 2076 MD5Update(&ctx, (u_char *) &fport, sizeof(fport)); 2077 2078 MD5Update(&ctx, tcp_iss_secret, sizeof(tcp_iss_secret)); 2079 2080 MD5Final(hash, &ctx); 2081 2082 memcpy(&tcp_iss, hash, sizeof(tcp_iss)); 2083 2084 /* 2085 * Now increment our "timer", and add it in to 2086 * the computed value. 2087 * 2088 * XXX Use `addin'? 2089 * XXX TCP_ISSINCR too large to use? 2090 */ 2091 tcp_iss_seq += TCP_ISSINCR; 2092 #ifdef TCPISS_DEBUG 2093 printf("ISS hash 0x%08x, ", tcp_iss); 2094 #endif 2095 tcp_iss += tcp_iss_seq + addin; 2096 #ifdef TCPISS_DEBUG 2097 printf("new ISS 0x%08x\n", tcp_iss); 2098 #endif 2099 } else 2100 #endif /* NRND > 0 */ 2101 { 2102 /* 2103 * Randomize. 2104 */ 2105 #if NRND > 0 2106 rnd_extract_data(&tcp_iss, sizeof(tcp_iss), RND_EXTRACT_ANY); 2107 #else 2108 tcp_iss = arc4random(); 2109 #endif 2110 2111 /* 2112 * If we were asked to add some amount to a known value, 2113 * we will take a random value obtained above, mask off 2114 * the upper bits, and add in the known value. We also 2115 * add in a constant to ensure that we are at least a 2116 * certain distance from the original value. 2117 * 2118 * This is used when an old connection is in timed wait 2119 * and we have a new one coming in, for instance. 2120 */ 2121 if (addin != 0) { 2122 #ifdef TCPISS_DEBUG 2123 printf("Random %08x, ", tcp_iss); 2124 #endif 2125 tcp_iss &= TCP_ISS_RANDOM_MASK; 2126 tcp_iss += addin + TCP_ISSINCR; 2127 #ifdef TCPISS_DEBUG 2128 printf("Old ISS %08x, ISS %08x\n", addin, tcp_iss); 2129 #endif 2130 } else { 2131 tcp_iss &= TCP_ISS_RANDOM_MASK; 2132 tcp_iss += tcp_iss_seq; 2133 tcp_iss_seq += TCP_ISSINCR; 2134 #ifdef TCPISS_DEBUG 2135 printf("ISS %08x\n", tcp_iss); 2136 #endif 2137 } 2138 } 2139 2140 if (tcp_compat_42) { 2141 /* 2142 * Limit it to the positive range for really old TCP 2143 * implementations. 2144 * Just AND off the top bit instead of checking if 2145 * is set first - saves a branch 50% of the time. 2146 */ 2147 tcp_iss &= 0x7fffffff; /* XXX */ 2148 } 2149 2150 return (tcp_iss); 2151 } 2152 2153 #if defined(IPSEC) || defined(FAST_IPSEC) 2154 /* compute ESP/AH header size for TCP, including outer IP header. */ 2155 size_t 2156 ipsec4_hdrsiz_tcp(tp) 2157 struct tcpcb *tp; 2158 { 2159 struct inpcb *inp; 2160 size_t hdrsiz; 2161 2162 /* XXX mapped addr case (tp->t_in6pcb) */ 2163 if (!tp || !tp->t_template || !(inp = tp->t_inpcb)) 2164 return 0; 2165 switch (tp->t_family) { 2166 case AF_INET: 2167 /* XXX: should use currect direction. */ 2168 hdrsiz = ipsec4_hdrsiz(tp->t_template, IPSEC_DIR_OUTBOUND, inp); 2169 break; 2170 default: 2171 hdrsiz = 0; 2172 break; 2173 } 2174 2175 return hdrsiz; 2176 } 2177 2178 #ifdef INET6 2179 size_t 2180 ipsec6_hdrsiz_tcp(tp) 2181 struct tcpcb *tp; 2182 { 2183 struct in6pcb *in6p; 2184 size_t hdrsiz; 2185 2186 if (!tp || !tp->t_template || !(in6p = tp->t_in6pcb)) 2187 return 0; 2188 switch (tp->t_family) { 2189 case AF_INET6: 2190 /* XXX: should use currect direction. */ 2191 hdrsiz = ipsec6_hdrsiz(tp->t_template, IPSEC_DIR_OUTBOUND, in6p); 2192 break; 2193 case AF_INET: 2194 /* mapped address case - tricky */ 2195 default: 2196 hdrsiz = 0; 2197 break; 2198 } 2199 2200 return hdrsiz; 2201 } 2202 #endif 2203 #endif /*IPSEC*/ 2204 2205 /* 2206 * Determine the length of the TCP options for this connection. 2207 * 2208 * XXX: What do we do for SACK, when we add that? Just reserve 2209 * all of the space? Otherwise we can't exactly be incrementing 2210 * cwnd by an amount that varies depending on the amount we last 2211 * had to SACK! 2212 */ 2213 2214 u_int 2215 tcp_optlen(tp) 2216 struct tcpcb *tp; 2217 { 2218 u_int optlen; 2219 2220 optlen = 0; 2221 if ((tp->t_flags & (TF_REQ_TSTMP|TF_RCVD_TSTMP|TF_NOOPT)) == 2222 (TF_REQ_TSTMP | TF_RCVD_TSTMP)) 2223 optlen += TCPOLEN_TSTAMP_APPA; 2224 2225 #ifdef TCP_SIGNATURE 2226 #if defined(INET6) && defined(FAST_IPSEC) 2227 if (tp->t_family == AF_INET) 2228 #endif 2229 if (tp->t_flags & TF_SIGNATURE) 2230 optlen += TCPOLEN_SIGNATURE + 2; 2231 #endif /* TCP_SIGNATURE */ 2232 2233 return optlen; 2234 } 2235