1 /* $NetBSD: tcp_subr.c,v 1.200 2006/10/05 17:35:19 tls 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 1.200 2006/10/05 17:35:19 tls 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 int tcp_do_ecn = 0; /* Explicit Congestion Notification */ 188 #ifndef TCP_INIT_WIN 189 #define TCP_INIT_WIN 0 /* initial slow start window */ 190 #endif 191 #ifndef TCP_INIT_WIN_LOCAL 192 #define TCP_INIT_WIN_LOCAL 4 /* initial slow start window for local nets */ 193 #endif 194 int tcp_init_win = TCP_INIT_WIN; 195 int tcp_init_win_local = TCP_INIT_WIN_LOCAL; 196 int tcp_mss_ifmtu = 0; 197 #ifdef TCP_COMPAT_42 198 int tcp_compat_42 = 1; 199 #else 200 int tcp_compat_42 = 0; 201 #endif 202 int tcp_rst_ppslim = 100; /* 100pps */ 203 int tcp_ackdrop_ppslim = 100; /* 100pps */ 204 int tcp_do_loopback_cksum = 0; 205 int tcp_sack_tp_maxholes = 32; 206 int tcp_sack_globalmaxholes = 1024; 207 int tcp_sack_globalholes = 0; 208 int tcp_ecn_maxretries = 1; 209 210 211 /* tcb hash */ 212 #ifndef TCBHASHSIZE 213 #define TCBHASHSIZE 128 214 #endif 215 int tcbhashsize = TCBHASHSIZE; 216 217 /* syn hash parameters */ 218 #define TCP_SYN_HASH_SIZE 293 219 #define TCP_SYN_BUCKET_SIZE 35 220 int tcp_syn_cache_size = TCP_SYN_HASH_SIZE; 221 int tcp_syn_cache_limit = TCP_SYN_HASH_SIZE*TCP_SYN_BUCKET_SIZE; 222 int tcp_syn_bucket_limit = 3*TCP_SYN_BUCKET_SIZE; 223 struct syn_cache_head tcp_syn_cache[TCP_SYN_HASH_SIZE]; 224 225 int tcp_freeq(struct tcpcb *); 226 227 #ifdef INET 228 void tcp_mtudisc_callback(struct in_addr); 229 #endif 230 #ifdef INET6 231 void tcp6_mtudisc_callback(struct in6_addr *); 232 #endif 233 234 #ifdef INET6 235 void tcp6_mtudisc(struct in6pcb *, int); 236 #endif 237 238 POOL_INIT(tcpcb_pool, sizeof(struct tcpcb), 0, 0, 0, "tcpcbpl", NULL); 239 240 #ifdef TCP_CSUM_COUNTERS 241 #include <sys/device.h> 242 243 #if defined(INET) 244 struct evcnt tcp_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 245 NULL, "tcp", "hwcsum bad"); 246 struct evcnt tcp_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 247 NULL, "tcp", "hwcsum ok"); 248 struct evcnt tcp_hwcsum_data = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 249 NULL, "tcp", "hwcsum data"); 250 struct evcnt tcp_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 251 NULL, "tcp", "swcsum"); 252 253 EVCNT_ATTACH_STATIC(tcp_hwcsum_bad); 254 EVCNT_ATTACH_STATIC(tcp_hwcsum_ok); 255 EVCNT_ATTACH_STATIC(tcp_hwcsum_data); 256 EVCNT_ATTACH_STATIC(tcp_swcsum); 257 #endif /* defined(INET) */ 258 259 #if defined(INET6) 260 struct evcnt tcp6_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 261 NULL, "tcp6", "hwcsum bad"); 262 struct evcnt tcp6_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 263 NULL, "tcp6", "hwcsum ok"); 264 struct evcnt tcp6_hwcsum_data = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 265 NULL, "tcp6", "hwcsum data"); 266 struct evcnt tcp6_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 267 NULL, "tcp6", "swcsum"); 268 269 EVCNT_ATTACH_STATIC(tcp6_hwcsum_bad); 270 EVCNT_ATTACH_STATIC(tcp6_hwcsum_ok); 271 EVCNT_ATTACH_STATIC(tcp6_hwcsum_data); 272 EVCNT_ATTACH_STATIC(tcp6_swcsum); 273 #endif /* defined(INET6) */ 274 #endif /* TCP_CSUM_COUNTERS */ 275 276 277 #ifdef TCP_OUTPUT_COUNTERS 278 #include <sys/device.h> 279 280 struct evcnt tcp_output_bigheader = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 281 NULL, "tcp", "output big header"); 282 struct evcnt tcp_output_predict_hit = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 283 NULL, "tcp", "output predict hit"); 284 struct evcnt tcp_output_predict_miss = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 285 NULL, "tcp", "output predict miss"); 286 struct evcnt tcp_output_copysmall = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 287 NULL, "tcp", "output copy small"); 288 struct evcnt tcp_output_copybig = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 289 NULL, "tcp", "output copy big"); 290 struct evcnt tcp_output_refbig = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 291 NULL, "tcp", "output reference big"); 292 293 EVCNT_ATTACH_STATIC(tcp_output_bigheader); 294 EVCNT_ATTACH_STATIC(tcp_output_predict_hit); 295 EVCNT_ATTACH_STATIC(tcp_output_predict_miss); 296 EVCNT_ATTACH_STATIC(tcp_output_copysmall); 297 EVCNT_ATTACH_STATIC(tcp_output_copybig); 298 EVCNT_ATTACH_STATIC(tcp_output_refbig); 299 300 #endif /* TCP_OUTPUT_COUNTERS */ 301 302 #ifdef TCP_REASS_COUNTERS 303 #include <sys/device.h> 304 305 struct evcnt tcp_reass_ = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 306 NULL, "tcp_reass", "calls"); 307 struct evcnt tcp_reass_empty = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 308 &tcp_reass_, "tcp_reass", "insert into empty queue"); 309 struct evcnt tcp_reass_iteration[8] = { 310 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", ">7 iterations"), 311 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "1 iteration"), 312 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "2 iterations"), 313 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "3 iterations"), 314 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "4 iterations"), 315 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "5 iterations"), 316 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "6 iterations"), 317 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "7 iterations"), 318 }; 319 struct evcnt tcp_reass_prependfirst = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 320 &tcp_reass_, "tcp_reass", "prepend to first"); 321 struct evcnt tcp_reass_prepend = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 322 &tcp_reass_, "tcp_reass", "prepend"); 323 struct evcnt tcp_reass_insert = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 324 &tcp_reass_, "tcp_reass", "insert"); 325 struct evcnt tcp_reass_inserttail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 326 &tcp_reass_, "tcp_reass", "insert at tail"); 327 struct evcnt tcp_reass_append = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 328 &tcp_reass_, "tcp_reass", "append"); 329 struct evcnt tcp_reass_appendtail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 330 &tcp_reass_, "tcp_reass", "append to tail fragment"); 331 struct evcnt tcp_reass_overlaptail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 332 &tcp_reass_, "tcp_reass", "overlap at end"); 333 struct evcnt tcp_reass_overlapfront = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 334 &tcp_reass_, "tcp_reass", "overlap at start"); 335 struct evcnt tcp_reass_segdup = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 336 &tcp_reass_, "tcp_reass", "duplicate segment"); 337 struct evcnt tcp_reass_fragdup = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 338 &tcp_reass_, "tcp_reass", "duplicate fragment"); 339 340 EVCNT_ATTACH_STATIC(tcp_reass_); 341 EVCNT_ATTACH_STATIC(tcp_reass_empty); 342 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 0); 343 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 1); 344 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 2); 345 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 3); 346 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 4); 347 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 5); 348 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 6); 349 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 7); 350 EVCNT_ATTACH_STATIC(tcp_reass_prependfirst); 351 EVCNT_ATTACH_STATIC(tcp_reass_prepend); 352 EVCNT_ATTACH_STATIC(tcp_reass_insert); 353 EVCNT_ATTACH_STATIC(tcp_reass_inserttail); 354 EVCNT_ATTACH_STATIC(tcp_reass_append); 355 EVCNT_ATTACH_STATIC(tcp_reass_appendtail); 356 EVCNT_ATTACH_STATIC(tcp_reass_overlaptail); 357 EVCNT_ATTACH_STATIC(tcp_reass_overlapfront); 358 EVCNT_ATTACH_STATIC(tcp_reass_segdup); 359 EVCNT_ATTACH_STATIC(tcp_reass_fragdup); 360 361 #endif /* TCP_REASS_COUNTERS */ 362 363 #ifdef MBUFTRACE 364 struct mowner tcp_mowner = { "tcp" }; 365 struct mowner tcp_rx_mowner = { "tcp", "rx" }; 366 struct mowner tcp_tx_mowner = { "tcp", "tx" }; 367 #endif 368 369 /* 370 * Tcp initialization 371 */ 372 void 373 tcp_init(void) 374 { 375 int hlen; 376 377 /* Initialize the TCPCB template. */ 378 tcp_tcpcb_template(); 379 380 in_pcbinit(&tcbtable, tcbhashsize, tcbhashsize); 381 382 hlen = sizeof(struct ip) + sizeof(struct tcphdr); 383 #ifdef INET6 384 if (sizeof(struct ip) < sizeof(struct ip6_hdr)) 385 hlen = sizeof(struct ip6_hdr) + sizeof(struct tcphdr); 386 #endif 387 if (max_protohdr < hlen) 388 max_protohdr = hlen; 389 if (max_linkhdr + hlen > MHLEN) 390 panic("tcp_init"); 391 392 #ifdef INET 393 icmp_mtudisc_callback_register(tcp_mtudisc_callback); 394 #endif 395 #ifdef INET6 396 icmp6_mtudisc_callback_register(tcp6_mtudisc_callback); 397 #endif 398 399 /* Initialize timer state. */ 400 tcp_timer_init(); 401 402 /* Initialize the compressed state engine. */ 403 syn_cache_init(); 404 405 MOWNER_ATTACH(&tcp_tx_mowner); 406 MOWNER_ATTACH(&tcp_rx_mowner); 407 MOWNER_ATTACH(&tcp_mowner); 408 } 409 410 /* 411 * Create template to be used to send tcp packets on a connection. 412 * Call after host entry created, allocates an mbuf and fills 413 * in a skeletal tcp/ip header, minimizing the amount of work 414 * necessary when the connection is used. 415 */ 416 struct mbuf * 417 tcp_template(struct tcpcb *tp) 418 { 419 struct inpcb *inp = tp->t_inpcb; 420 #ifdef INET6 421 struct in6pcb *in6p = tp->t_in6pcb; 422 #endif 423 struct tcphdr *n; 424 struct mbuf *m; 425 int hlen; 426 427 switch (tp->t_family) { 428 case AF_INET: 429 hlen = sizeof(struct ip); 430 if (inp) 431 break; 432 #ifdef INET6 433 if (in6p) { 434 /* mapped addr case */ 435 if (IN6_IS_ADDR_V4MAPPED(&in6p->in6p_laddr) 436 && IN6_IS_ADDR_V4MAPPED(&in6p->in6p_faddr)) 437 break; 438 } 439 #endif 440 return NULL; /*EINVAL*/ 441 #ifdef INET6 442 case AF_INET6: 443 hlen = sizeof(struct ip6_hdr); 444 if (in6p) { 445 /* more sainty check? */ 446 break; 447 } 448 return NULL; /*EINVAL*/ 449 #endif 450 default: 451 hlen = 0; /*pacify gcc*/ 452 return NULL; /*EAFNOSUPPORT*/ 453 } 454 #ifdef DIAGNOSTIC 455 if (hlen + sizeof(struct tcphdr) > MCLBYTES) 456 panic("mclbytes too small for t_template"); 457 #endif 458 m = tp->t_template; 459 if (m && m->m_len == hlen + sizeof(struct tcphdr)) 460 ; 461 else { 462 if (m) 463 m_freem(m); 464 m = tp->t_template = NULL; 465 MGETHDR(m, M_DONTWAIT, MT_HEADER); 466 if (m && hlen + sizeof(struct tcphdr) > MHLEN) { 467 MCLGET(m, M_DONTWAIT); 468 if ((m->m_flags & M_EXT) == 0) { 469 m_free(m); 470 m = NULL; 471 } 472 } 473 if (m == NULL) 474 return NULL; 475 MCLAIM(m, &tcp_mowner); 476 m->m_pkthdr.len = m->m_len = hlen + sizeof(struct tcphdr); 477 } 478 479 bzero(mtod(m, caddr_t), m->m_len); 480 481 n = (struct tcphdr *)(mtod(m, caddr_t) + hlen); 482 483 switch (tp->t_family) { 484 case AF_INET: 485 { 486 struct ipovly *ipov; 487 mtod(m, struct ip *)->ip_v = 4; 488 mtod(m, struct ip *)->ip_hl = hlen >> 2; 489 ipov = mtod(m, struct ipovly *); 490 ipov->ih_pr = IPPROTO_TCP; 491 ipov->ih_len = htons(sizeof(struct tcphdr)); 492 if (inp) { 493 ipov->ih_src = inp->inp_laddr; 494 ipov->ih_dst = inp->inp_faddr; 495 } 496 #ifdef INET6 497 else if (in6p) { 498 /* mapped addr case */ 499 bcopy(&in6p->in6p_laddr.s6_addr32[3], &ipov->ih_src, 500 sizeof(ipov->ih_src)); 501 bcopy(&in6p->in6p_faddr.s6_addr32[3], &ipov->ih_dst, 502 sizeof(ipov->ih_dst)); 503 } 504 #endif 505 /* 506 * Compute the pseudo-header portion of the checksum 507 * now. We incrementally add in the TCP option and 508 * payload lengths later, and then compute the TCP 509 * checksum right before the packet is sent off onto 510 * the wire. 511 */ 512 n->th_sum = in_cksum_phdr(ipov->ih_src.s_addr, 513 ipov->ih_dst.s_addr, 514 htons(sizeof(struct tcphdr) + IPPROTO_TCP)); 515 break; 516 } 517 #ifdef INET6 518 case AF_INET6: 519 { 520 struct ip6_hdr *ip6; 521 mtod(m, struct ip *)->ip_v = 6; 522 ip6 = mtod(m, struct ip6_hdr *); 523 ip6->ip6_nxt = IPPROTO_TCP; 524 ip6->ip6_plen = htons(sizeof(struct tcphdr)); 525 ip6->ip6_src = in6p->in6p_laddr; 526 ip6->ip6_dst = in6p->in6p_faddr; 527 ip6->ip6_flow = in6p->in6p_flowinfo & IPV6_FLOWINFO_MASK; 528 if (ip6_auto_flowlabel) { 529 ip6->ip6_flow &= ~IPV6_FLOWLABEL_MASK; 530 ip6->ip6_flow |= 531 (htonl(ip6_randomflowlabel()) & IPV6_FLOWLABEL_MASK); 532 } 533 ip6->ip6_vfc &= ~IPV6_VERSION_MASK; 534 ip6->ip6_vfc |= IPV6_VERSION; 535 536 /* 537 * Compute the pseudo-header portion of the checksum 538 * now. We incrementally add in the TCP option and 539 * payload lengths later, and then compute the TCP 540 * checksum right before the packet is sent off onto 541 * the wire. 542 */ 543 n->th_sum = in6_cksum_phdr(&in6p->in6p_laddr, 544 &in6p->in6p_faddr, htonl(sizeof(struct tcphdr)), 545 htonl(IPPROTO_TCP)); 546 break; 547 } 548 #endif 549 } 550 if (inp) { 551 n->th_sport = inp->inp_lport; 552 n->th_dport = inp->inp_fport; 553 } 554 #ifdef INET6 555 else if (in6p) { 556 n->th_sport = in6p->in6p_lport; 557 n->th_dport = in6p->in6p_fport; 558 } 559 #endif 560 n->th_seq = 0; 561 n->th_ack = 0; 562 n->th_x2 = 0; 563 n->th_off = 5; 564 n->th_flags = 0; 565 n->th_win = 0; 566 n->th_urp = 0; 567 return (m); 568 } 569 570 /* 571 * Send a single message to the TCP at address specified by 572 * the given TCP/IP header. If m == 0, then we make a copy 573 * of the tcpiphdr at ti and send directly to the addressed host. 574 * This is used to force keep alive messages out using the TCP 575 * template for a connection tp->t_template. If flags are given 576 * then we send a message back to the TCP which originated the 577 * segment ti, and discard the mbuf containing it and any other 578 * attached mbufs. 579 * 580 * In any case the ack and sequence number of the transmitted 581 * segment are as specified by the parameters. 582 */ 583 int 584 tcp_respond(struct tcpcb *tp, struct mbuf *template, struct mbuf *m, 585 struct tcphdr *th0, tcp_seq ack, tcp_seq seq, int flags) 586 { 587 struct route *ro; 588 int error, tlen, win = 0; 589 int hlen; 590 struct ip *ip; 591 #ifdef INET6 592 struct ip6_hdr *ip6; 593 #endif 594 int family; /* family on packet, not inpcb/in6pcb! */ 595 struct tcphdr *th; 596 struct socket *so; 597 598 if (tp != NULL && (flags & TH_RST) == 0) { 599 #ifdef DIAGNOSTIC 600 if (tp->t_inpcb && tp->t_in6pcb) 601 panic("tcp_respond: both t_inpcb and t_in6pcb are set"); 602 #endif 603 #ifdef INET 604 if (tp->t_inpcb) 605 win = sbspace(&tp->t_inpcb->inp_socket->so_rcv); 606 #endif 607 #ifdef INET6 608 if (tp->t_in6pcb) 609 win = sbspace(&tp->t_in6pcb->in6p_socket->so_rcv); 610 #endif 611 } 612 613 th = NULL; /* Quell uninitialized warning */ 614 ip = NULL; 615 #ifdef INET6 616 ip6 = NULL; 617 #endif 618 if (m == 0) { 619 if (!template) 620 return EINVAL; 621 622 /* get family information from template */ 623 switch (mtod(template, struct ip *)->ip_v) { 624 case 4: 625 family = AF_INET; 626 hlen = sizeof(struct ip); 627 break; 628 #ifdef INET6 629 case 6: 630 family = AF_INET6; 631 hlen = sizeof(struct ip6_hdr); 632 break; 633 #endif 634 default: 635 return EAFNOSUPPORT; 636 } 637 638 MGETHDR(m, M_DONTWAIT, MT_HEADER); 639 if (m) { 640 MCLAIM(m, &tcp_tx_mowner); 641 MCLGET(m, M_DONTWAIT); 642 if ((m->m_flags & M_EXT) == 0) { 643 m_free(m); 644 m = NULL; 645 } 646 } 647 if (m == NULL) 648 return (ENOBUFS); 649 650 if (tcp_compat_42) 651 tlen = 1; 652 else 653 tlen = 0; 654 655 m->m_data += max_linkhdr; 656 bcopy(mtod(template, caddr_t), mtod(m, caddr_t), 657 template->m_len); 658 switch (family) { 659 case AF_INET: 660 ip = mtod(m, struct ip *); 661 th = (struct tcphdr *)(ip + 1); 662 break; 663 #ifdef INET6 664 case AF_INET6: 665 ip6 = mtod(m, struct ip6_hdr *); 666 th = (struct tcphdr *)(ip6 + 1); 667 break; 668 #endif 669 #if 0 670 default: 671 /* noone will visit here */ 672 m_freem(m); 673 return EAFNOSUPPORT; 674 #endif 675 } 676 flags = TH_ACK; 677 } else { 678 679 if ((m->m_flags & M_PKTHDR) == 0) { 680 #if 0 681 printf("non PKTHDR to tcp_respond\n"); 682 #endif 683 m_freem(m); 684 return EINVAL; 685 } 686 #ifdef DIAGNOSTIC 687 if (!th0) 688 panic("th0 == NULL in tcp_respond"); 689 #endif 690 691 /* get family information from m */ 692 switch (mtod(m, struct ip *)->ip_v) { 693 case 4: 694 family = AF_INET; 695 hlen = sizeof(struct ip); 696 ip = mtod(m, struct ip *); 697 break; 698 #ifdef INET6 699 case 6: 700 family = AF_INET6; 701 hlen = sizeof(struct ip6_hdr); 702 ip6 = mtod(m, struct ip6_hdr *); 703 break; 704 #endif 705 default: 706 m_freem(m); 707 return EAFNOSUPPORT; 708 } 709 /* clear h/w csum flags inherited from rx packet */ 710 m->m_pkthdr.csum_flags = 0; 711 712 if ((flags & TH_SYN) == 0 || sizeof(*th0) > (th0->th_off << 2)) 713 tlen = sizeof(*th0); 714 else 715 tlen = th0->th_off << 2; 716 717 if (m->m_len > hlen + tlen && (m->m_flags & M_EXT) == 0 && 718 mtod(m, caddr_t) + hlen == (caddr_t)th0) { 719 m->m_len = hlen + tlen; 720 m_freem(m->m_next); 721 m->m_next = NULL; 722 } else { 723 struct mbuf *n; 724 725 #ifdef DIAGNOSTIC 726 if (max_linkhdr + hlen + tlen > MCLBYTES) { 727 m_freem(m); 728 return EMSGSIZE; 729 } 730 #endif 731 MGETHDR(n, M_DONTWAIT, MT_HEADER); 732 if (n && max_linkhdr + hlen + tlen > MHLEN) { 733 MCLGET(n, M_DONTWAIT); 734 if ((n->m_flags & M_EXT) == 0) { 735 m_freem(n); 736 n = NULL; 737 } 738 } 739 if (!n) { 740 m_freem(m); 741 return ENOBUFS; 742 } 743 744 MCLAIM(n, &tcp_tx_mowner); 745 n->m_data += max_linkhdr; 746 n->m_len = hlen + tlen; 747 m_copyback(n, 0, hlen, mtod(m, caddr_t)); 748 m_copyback(n, hlen, tlen, (caddr_t)th0); 749 750 m_freem(m); 751 m = n; 752 n = NULL; 753 } 754 755 #define xchg(a,b,type) { type t; t=a; a=b; b=t; } 756 switch (family) { 757 case AF_INET: 758 ip = mtod(m, struct ip *); 759 th = (struct tcphdr *)(ip + 1); 760 ip->ip_p = IPPROTO_TCP; 761 xchg(ip->ip_dst, ip->ip_src, struct in_addr); 762 ip->ip_p = IPPROTO_TCP; 763 break; 764 #ifdef INET6 765 case AF_INET6: 766 ip6 = mtod(m, struct ip6_hdr *); 767 th = (struct tcphdr *)(ip6 + 1); 768 ip6->ip6_nxt = IPPROTO_TCP; 769 xchg(ip6->ip6_dst, ip6->ip6_src, struct in6_addr); 770 ip6->ip6_nxt = IPPROTO_TCP; 771 break; 772 #endif 773 #if 0 774 default: 775 /* noone will visit here */ 776 m_freem(m); 777 return EAFNOSUPPORT; 778 #endif 779 } 780 xchg(th->th_dport, th->th_sport, u_int16_t); 781 #undef xchg 782 tlen = 0; /*be friendly with the following code*/ 783 } 784 th->th_seq = htonl(seq); 785 th->th_ack = htonl(ack); 786 th->th_x2 = 0; 787 if ((flags & TH_SYN) == 0) { 788 if (tp) 789 win >>= tp->rcv_scale; 790 if (win > TCP_MAXWIN) 791 win = TCP_MAXWIN; 792 th->th_win = htons((u_int16_t)win); 793 th->th_off = sizeof (struct tcphdr) >> 2; 794 tlen += sizeof(*th); 795 } else 796 tlen += th->th_off << 2; 797 m->m_len = hlen + tlen; 798 m->m_pkthdr.len = hlen + tlen; 799 m->m_pkthdr.rcvif = (struct ifnet *) 0; 800 th->th_flags = flags; 801 th->th_urp = 0; 802 803 switch (family) { 804 #ifdef INET 805 case AF_INET: 806 { 807 struct ipovly *ipov = (struct ipovly *)ip; 808 bzero(ipov->ih_x1, sizeof ipov->ih_x1); 809 ipov->ih_len = htons((u_int16_t)tlen); 810 811 th->th_sum = 0; 812 th->th_sum = in_cksum(m, hlen + tlen); 813 ip->ip_len = htons(hlen + tlen); 814 ip->ip_ttl = ip_defttl; 815 break; 816 } 817 #endif 818 #ifdef INET6 819 case AF_INET6: 820 { 821 th->th_sum = 0; 822 th->th_sum = in6_cksum(m, IPPROTO_TCP, sizeof(struct ip6_hdr), 823 tlen); 824 ip6->ip6_plen = htons(tlen); 825 if (tp && tp->t_in6pcb) { 826 struct ifnet *oifp; 827 ro = (struct route *)&tp->t_in6pcb->in6p_route; 828 oifp = ro->ro_rt ? ro->ro_rt->rt_ifp : NULL; 829 ip6->ip6_hlim = in6_selecthlim(tp->t_in6pcb, oifp); 830 } else 831 ip6->ip6_hlim = ip6_defhlim; 832 ip6->ip6_flow &= ~IPV6_FLOWINFO_MASK; 833 if (ip6_auto_flowlabel) { 834 ip6->ip6_flow |= 835 (htonl(ip6_randomflowlabel()) & IPV6_FLOWLABEL_MASK); 836 } 837 break; 838 } 839 #endif 840 } 841 842 if (tp && tp->t_inpcb) 843 so = tp->t_inpcb->inp_socket; 844 #ifdef INET6 845 else if (tp && tp->t_in6pcb) 846 so = tp->t_in6pcb->in6p_socket; 847 #endif 848 else 849 so = NULL; 850 851 if (tp != NULL && tp->t_inpcb != NULL) { 852 ro = &tp->t_inpcb->inp_route; 853 #ifdef DIAGNOSTIC 854 if (family != AF_INET) 855 panic("tcp_respond: address family mismatch"); 856 if (!in_hosteq(ip->ip_dst, tp->t_inpcb->inp_faddr)) { 857 panic("tcp_respond: ip_dst %x != inp_faddr %x", 858 ntohl(ip->ip_dst.s_addr), 859 ntohl(tp->t_inpcb->inp_faddr.s_addr)); 860 } 861 #endif 862 } 863 #ifdef INET6 864 else if (tp != NULL && tp->t_in6pcb != NULL) { 865 ro = (struct route *)&tp->t_in6pcb->in6p_route; 866 #ifdef DIAGNOSTIC 867 if (family == AF_INET) { 868 if (!IN6_IS_ADDR_V4MAPPED(&tp->t_in6pcb->in6p_faddr)) 869 panic("tcp_respond: not mapped addr"); 870 if (bcmp(&ip->ip_dst, 871 &tp->t_in6pcb->in6p_faddr.s6_addr32[3], 872 sizeof(ip->ip_dst)) != 0) { 873 panic("tcp_respond: ip_dst != in6p_faddr"); 874 } 875 } else if (family == AF_INET6) { 876 if (!IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst, 877 &tp->t_in6pcb->in6p_faddr)) 878 panic("tcp_respond: ip6_dst != in6p_faddr"); 879 } else 880 panic("tcp_respond: address family mismatch"); 881 #endif 882 } 883 #endif 884 else 885 ro = NULL; 886 887 switch (family) { 888 #ifdef INET 889 case AF_INET: 890 error = ip_output(m, NULL, ro, 891 (tp && tp->t_mtudisc ? IP_MTUDISC : 0), 892 (struct ip_moptions *)0, so); 893 break; 894 #endif 895 #ifdef INET6 896 case AF_INET6: 897 error = ip6_output(m, NULL, (struct route_in6 *)ro, 0, 898 (struct ip6_moptions *)0, so, NULL); 899 break; 900 #endif 901 default: 902 error = EAFNOSUPPORT; 903 break; 904 } 905 906 return (error); 907 } 908 909 /* 910 * Template TCPCB. Rather than zeroing a new TCPCB and initializing 911 * a bunch of members individually, we maintain this template for the 912 * static and mostly-static components of the TCPCB, and copy it into 913 * the new TCPCB instead. 914 */ 915 static struct tcpcb tcpcb_template = { 916 /* 917 * If TCP_NTIMERS ever changes, we'll need to update this 918 * initializer. 919 */ 920 .t_timer = { 921 CALLOUT_INITIALIZER, 922 CALLOUT_INITIALIZER, 923 CALLOUT_INITIALIZER, 924 CALLOUT_INITIALIZER, 925 }, 926 .t_delack_ch = CALLOUT_INITIALIZER, 927 928 .t_srtt = TCPTV_SRTTBASE, 929 .t_rttmin = TCPTV_MIN, 930 931 .snd_cwnd = TCP_MAXWIN << TCP_MAX_WINSHIFT, 932 .snd_ssthresh = TCP_MAXWIN << TCP_MAX_WINSHIFT, 933 .snd_numholes = 0, 934 935 .t_partialacks = -1, 936 }; 937 938 /* 939 * Updates the TCPCB template whenever a parameter that would affect 940 * the template is changed. 941 */ 942 void 943 tcp_tcpcb_template(void) 944 { 945 struct tcpcb *tp = &tcpcb_template; 946 int flags; 947 948 tp->t_peermss = tcp_mssdflt; 949 tp->t_ourmss = tcp_mssdflt; 950 tp->t_segsz = tcp_mssdflt; 951 952 flags = 0; 953 if (tcp_do_rfc1323 && tcp_do_win_scale) 954 flags |= TF_REQ_SCALE; 955 if (tcp_do_rfc1323 && tcp_do_timestamps) 956 flags |= TF_REQ_TSTMP; 957 tp->t_flags = flags; 958 959 /* 960 * Init srtt to TCPTV_SRTTBASE (0), so we can tell that we have no 961 * rtt estimate. Set rttvar so that srtt + 2 * rttvar gives 962 * reasonable initial retransmit time. 963 */ 964 tp->t_rttvar = tcp_rttdflt * PR_SLOWHZ << (TCP_RTTVAR_SHIFT + 2 - 1); 965 TCPT_RANGESET(tp->t_rxtcur, TCP_REXMTVAL(tp), 966 TCPTV_MIN, TCPTV_REXMTMAX); 967 } 968 969 /* 970 * Create a new TCP control block, making an 971 * empty reassembly queue and hooking it to the argument 972 * protocol control block. 973 */ 974 /* family selects inpcb, or in6pcb */ 975 struct tcpcb * 976 tcp_newtcpcb(int family, void *aux) 977 { 978 struct tcpcb *tp; 979 int i; 980 981 /* XXX Consider using a pool_cache for speed. */ 982 tp = pool_get(&tcpcb_pool, PR_NOWAIT); /* splsoftnet via tcp_usrreq */ 983 if (tp == NULL) 984 return (NULL); 985 memcpy(tp, &tcpcb_template, sizeof(*tp)); 986 TAILQ_INIT(&tp->segq); 987 TAILQ_INIT(&tp->timeq); 988 tp->t_family = family; /* may be overridden later on */ 989 TAILQ_INIT(&tp->snd_holes); 990 LIST_INIT(&tp->t_sc); /* XXX can template this */ 991 992 /* Don't sweat this loop; hopefully the compiler will unroll it. */ 993 for (i = 0; i < TCPT_NTIMERS; i++) 994 TCP_TIMER_INIT(tp, i); 995 996 switch (family) { 997 case AF_INET: 998 { 999 struct inpcb *inp = (struct inpcb *)aux; 1000 1001 inp->inp_ip.ip_ttl = ip_defttl; 1002 inp->inp_ppcb = (caddr_t)tp; 1003 1004 tp->t_inpcb = inp; 1005 tp->t_mtudisc = ip_mtudisc; 1006 break; 1007 } 1008 #ifdef INET6 1009 case AF_INET6: 1010 { 1011 struct in6pcb *in6p = (struct in6pcb *)aux; 1012 1013 in6p->in6p_ip6.ip6_hlim = in6_selecthlim(in6p, 1014 in6p->in6p_route.ro_rt ? in6p->in6p_route.ro_rt->rt_ifp 1015 : NULL); 1016 in6p->in6p_ppcb = (caddr_t)tp; 1017 1018 tp->t_in6pcb = in6p; 1019 /* for IPv6, always try to run path MTU discovery */ 1020 tp->t_mtudisc = 1; 1021 break; 1022 } 1023 #endif /* INET6 */ 1024 default: 1025 pool_put(&tcpcb_pool, tp); /* splsoftnet via tcp_usrreq */ 1026 return (NULL); 1027 } 1028 1029 /* 1030 * Initialize our timebase. When we send timestamps, we take 1031 * the delta from tcp_now -- this means each connection always 1032 * gets a timebase of 0, which makes it, among other things, 1033 * more difficult to determine how long a system has been up, 1034 * and thus how many TCP sequence increments have occurred. 1035 */ 1036 tp->ts_timebase = tcp_now; 1037 1038 return (tp); 1039 } 1040 1041 /* 1042 * Drop a TCP connection, reporting 1043 * the specified error. If connection is synchronized, 1044 * then send a RST to peer. 1045 */ 1046 struct tcpcb * 1047 tcp_drop(struct tcpcb *tp, int errno) 1048 { 1049 struct socket *so = NULL; 1050 1051 #ifdef DIAGNOSTIC 1052 if (tp->t_inpcb && tp->t_in6pcb) 1053 panic("tcp_drop: both t_inpcb and t_in6pcb are set"); 1054 #endif 1055 #ifdef INET 1056 if (tp->t_inpcb) 1057 so = tp->t_inpcb->inp_socket; 1058 #endif 1059 #ifdef INET6 1060 if (tp->t_in6pcb) 1061 so = tp->t_in6pcb->in6p_socket; 1062 #endif 1063 if (!so) 1064 return NULL; 1065 1066 if (TCPS_HAVERCVDSYN(tp->t_state)) { 1067 tp->t_state = TCPS_CLOSED; 1068 (void) tcp_output(tp); 1069 tcpstat.tcps_drops++; 1070 } else 1071 tcpstat.tcps_conndrops++; 1072 if (errno == ETIMEDOUT && tp->t_softerror) 1073 errno = tp->t_softerror; 1074 so->so_error = errno; 1075 return (tcp_close(tp)); 1076 } 1077 1078 /* 1079 * Return whether this tcpcb is marked as dead, indicating 1080 * to the calling timer function that no further action should 1081 * be taken, as we are about to release this tcpcb. The release 1082 * of the storage will be done if this is the last timer running. 1083 * 1084 * This should be called from the callout handler function after 1085 * callout_ack() is done, so that the number of invoking timer 1086 * functions is 0. 1087 */ 1088 int 1089 tcp_isdead(struct tcpcb *tp) 1090 { 1091 int dead = (tp->t_flags & TF_DEAD); 1092 1093 if (__predict_false(dead)) { 1094 if (tcp_timers_invoking(tp) > 0) 1095 /* not quite there yet -- count separately? */ 1096 return dead; 1097 tcpstat.tcps_delayed_free++; 1098 pool_put(&tcpcb_pool, tp); /* splsoftnet via tcp_timer.c */ 1099 } 1100 return dead; 1101 } 1102 1103 /* 1104 * Close a TCP control block: 1105 * discard all space held by the tcp 1106 * discard internet protocol block 1107 * wake up any sleepers 1108 */ 1109 struct tcpcb * 1110 tcp_close(struct tcpcb *tp) 1111 { 1112 struct inpcb *inp; 1113 #ifdef INET6 1114 struct in6pcb *in6p; 1115 #endif 1116 struct socket *so; 1117 #ifdef RTV_RTT 1118 struct rtentry *rt; 1119 #endif 1120 struct route *ro; 1121 int s; 1122 1123 inp = tp->t_inpcb; 1124 #ifdef INET6 1125 in6p = tp->t_in6pcb; 1126 #endif 1127 so = NULL; 1128 ro = NULL; 1129 if (inp) { 1130 so = inp->inp_socket; 1131 ro = &inp->inp_route; 1132 } 1133 #ifdef INET6 1134 else if (in6p) { 1135 so = in6p->in6p_socket; 1136 ro = (struct route *)&in6p->in6p_route; 1137 } 1138 #endif 1139 1140 #ifdef RTV_RTT 1141 /* 1142 * If we sent enough data to get some meaningful characteristics, 1143 * save them in the routing entry. 'Enough' is arbitrarily 1144 * defined as the sendpipesize (default 4K) * 16. This would 1145 * give us 16 rtt samples assuming we only get one sample per 1146 * window (the usual case on a long haul net). 16 samples is 1147 * enough for the srtt filter to converge to within 5% of the correct 1148 * value; fewer samples and we could save a very bogus rtt. 1149 * 1150 * Don't update the default route's characteristics and don't 1151 * update anything that the user "locked". 1152 */ 1153 if (SEQ_LT(tp->iss + so->so_snd.sb_hiwat * 16, tp->snd_max) && 1154 ro && (rt = ro->ro_rt) && 1155 !in_nullhost(satosin(rt_key(rt))->sin_addr)) { 1156 u_long i = 0; 1157 1158 if ((rt->rt_rmx.rmx_locks & RTV_RTT) == 0) { 1159 i = tp->t_srtt * 1160 ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTT_SHIFT + 2)); 1161 if (rt->rt_rmx.rmx_rtt && i) 1162 /* 1163 * filter this update to half the old & half 1164 * the new values, converting scale. 1165 * See route.h and tcp_var.h for a 1166 * description of the scaling constants. 1167 */ 1168 rt->rt_rmx.rmx_rtt = 1169 (rt->rt_rmx.rmx_rtt + i) / 2; 1170 else 1171 rt->rt_rmx.rmx_rtt = i; 1172 } 1173 if ((rt->rt_rmx.rmx_locks & RTV_RTTVAR) == 0) { 1174 i = tp->t_rttvar * 1175 ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTTVAR_SHIFT + 2)); 1176 if (rt->rt_rmx.rmx_rttvar && i) 1177 rt->rt_rmx.rmx_rttvar = 1178 (rt->rt_rmx.rmx_rttvar + i) / 2; 1179 else 1180 rt->rt_rmx.rmx_rttvar = i; 1181 } 1182 /* 1183 * update the pipelimit (ssthresh) if it has been updated 1184 * already or if a pipesize was specified & the threshhold 1185 * got below half the pipesize. I.e., wait for bad news 1186 * before we start updating, then update on both good 1187 * and bad news. 1188 */ 1189 if (((rt->rt_rmx.rmx_locks & RTV_SSTHRESH) == 0 && 1190 (i = tp->snd_ssthresh) && rt->rt_rmx.rmx_ssthresh) || 1191 i < (rt->rt_rmx.rmx_sendpipe / 2)) { 1192 /* 1193 * convert the limit from user data bytes to 1194 * packets then to packet data bytes. 1195 */ 1196 i = (i + tp->t_segsz / 2) / tp->t_segsz; 1197 if (i < 2) 1198 i = 2; 1199 i *= (u_long)(tp->t_segsz + sizeof (struct tcpiphdr)); 1200 if (rt->rt_rmx.rmx_ssthresh) 1201 rt->rt_rmx.rmx_ssthresh = 1202 (rt->rt_rmx.rmx_ssthresh + i) / 2; 1203 else 1204 rt->rt_rmx.rmx_ssthresh = i; 1205 } 1206 } 1207 #endif /* RTV_RTT */ 1208 /* free the reassembly queue, if any */ 1209 TCP_REASS_LOCK(tp); 1210 (void) tcp_freeq(tp); 1211 TCP_REASS_UNLOCK(tp); 1212 1213 /* free the SACK holes list. */ 1214 tcp_free_sackholes(tp); 1215 1216 tcp_canceltimers(tp); 1217 TCP_CLEAR_DELACK(tp); 1218 syn_cache_cleanup(tp); 1219 1220 if (tp->t_template) { 1221 m_free(tp->t_template); 1222 tp->t_template = NULL; 1223 } 1224 if (tcp_timers_invoking(tp)) 1225 tp->t_flags |= TF_DEAD; 1226 else 1227 { 1228 s = splsoftnet(); /* do we trust all callers of tcp_close()? */ 1229 pool_put(&tcpcb_pool, tp); 1230 splx(s); 1231 } 1232 1233 if (inp) { 1234 inp->inp_ppcb = 0; 1235 soisdisconnected(so); 1236 in_pcbdetach(inp); 1237 } 1238 #ifdef INET6 1239 else if (in6p) { 1240 in6p->in6p_ppcb = 0; 1241 soisdisconnected(so); 1242 in6_pcbdetach(in6p); 1243 } 1244 #endif 1245 tcpstat.tcps_closed++; 1246 return ((struct tcpcb *)0); 1247 } 1248 1249 int 1250 tcp_freeq(tp) 1251 struct tcpcb *tp; 1252 { 1253 struct ipqent *qe; 1254 int rv = 0; 1255 #ifdef TCPREASS_DEBUG 1256 int i = 0; 1257 #endif 1258 1259 TCP_REASS_LOCK_CHECK(tp); 1260 1261 while ((qe = TAILQ_FIRST(&tp->segq)) != NULL) { 1262 #ifdef TCPREASS_DEBUG 1263 printf("tcp_freeq[%p,%d]: %u:%u(%u) 0x%02x\n", 1264 tp, i++, qe->ipqe_seq, qe->ipqe_seq + qe->ipqe_len, 1265 qe->ipqe_len, qe->ipqe_flags & (TH_SYN|TH_FIN|TH_RST)); 1266 #endif 1267 TAILQ_REMOVE(&tp->segq, qe, ipqe_q); 1268 TAILQ_REMOVE(&tp->timeq, qe, ipqe_timeq); 1269 m_freem(qe->ipqe_m); 1270 tcpipqent_free(qe); 1271 rv = 1; 1272 } 1273 tp->t_segqlen = 0; 1274 KASSERT(TAILQ_EMPTY(&tp->timeq)); 1275 return (rv); 1276 } 1277 1278 /* 1279 * Protocol drain routine. Called when memory is in short supply. 1280 */ 1281 void 1282 tcp_drain(void) 1283 { 1284 struct inpcb_hdr *inph; 1285 struct tcpcb *tp; 1286 1287 /* 1288 * Free the sequence queue of all TCP connections. 1289 */ 1290 CIRCLEQ_FOREACH(inph, &tcbtable.inpt_queue, inph_queue) { 1291 switch (inph->inph_af) { 1292 case AF_INET: 1293 tp = intotcpcb((struct inpcb *)inph); 1294 break; 1295 #ifdef INET6 1296 case AF_INET6: 1297 tp = in6totcpcb((struct in6pcb *)inph); 1298 break; 1299 #endif 1300 default: 1301 tp = NULL; 1302 break; 1303 } 1304 if (tp != NULL) { 1305 /* 1306 * We may be called from a device's interrupt 1307 * context. If the tcpcb is already busy, 1308 * just bail out now. 1309 */ 1310 if (tcp_reass_lock_try(tp) == 0) 1311 continue; 1312 if (tcp_freeq(tp)) 1313 tcpstat.tcps_connsdrained++; 1314 TCP_REASS_UNLOCK(tp); 1315 } 1316 } 1317 } 1318 1319 /* 1320 * Notify a tcp user of an asynchronous error; 1321 * store error as soft error, but wake up user 1322 * (for now, won't do anything until can select for soft error). 1323 */ 1324 void 1325 tcp_notify(struct inpcb *inp, int error) 1326 { 1327 struct tcpcb *tp = (struct tcpcb *)inp->inp_ppcb; 1328 struct socket *so = inp->inp_socket; 1329 1330 /* 1331 * Ignore some errors if we are hooked up. 1332 * If connection hasn't completed, has retransmitted several times, 1333 * and receives a second error, give up now. This is better 1334 * than waiting a long time to establish a connection that 1335 * can never complete. 1336 */ 1337 if (tp->t_state == TCPS_ESTABLISHED && 1338 (error == EHOSTUNREACH || error == ENETUNREACH || 1339 error == EHOSTDOWN)) { 1340 return; 1341 } else if (TCPS_HAVEESTABLISHED(tp->t_state) == 0 && 1342 tp->t_rxtshift > 3 && tp->t_softerror) 1343 so->so_error = error; 1344 else 1345 tp->t_softerror = error; 1346 wakeup((caddr_t) &so->so_timeo); 1347 sorwakeup(so); 1348 sowwakeup(so); 1349 } 1350 1351 #ifdef INET6 1352 void 1353 tcp6_notify(struct in6pcb *in6p, int error) 1354 { 1355 struct tcpcb *tp = (struct tcpcb *)in6p->in6p_ppcb; 1356 struct socket *so = in6p->in6p_socket; 1357 1358 /* 1359 * Ignore some errors if we are hooked up. 1360 * If connection hasn't completed, has retransmitted several times, 1361 * and receives a second error, give up now. This is better 1362 * than waiting a long time to establish a connection that 1363 * can never complete. 1364 */ 1365 if (tp->t_state == TCPS_ESTABLISHED && 1366 (error == EHOSTUNREACH || error == ENETUNREACH || 1367 error == EHOSTDOWN)) { 1368 return; 1369 } else if (TCPS_HAVEESTABLISHED(tp->t_state) == 0 && 1370 tp->t_rxtshift > 3 && tp->t_softerror) 1371 so->so_error = error; 1372 else 1373 tp->t_softerror = error; 1374 wakeup((caddr_t) &so->so_timeo); 1375 sorwakeup(so); 1376 sowwakeup(so); 1377 } 1378 #endif 1379 1380 #ifdef INET6 1381 void 1382 tcp6_ctlinput(int cmd, struct sockaddr *sa, void *d) 1383 { 1384 struct tcphdr th; 1385 void (*notify)(struct in6pcb *, int) = tcp6_notify; 1386 int nmatch; 1387 struct ip6_hdr *ip6; 1388 const struct sockaddr_in6 *sa6_src = NULL; 1389 struct sockaddr_in6 *sa6 = (struct sockaddr_in6 *)sa; 1390 struct mbuf *m; 1391 int off; 1392 1393 if (sa->sa_family != AF_INET6 || 1394 sa->sa_len != sizeof(struct sockaddr_in6)) 1395 return; 1396 if ((unsigned)cmd >= PRC_NCMDS) 1397 return; 1398 else if (cmd == PRC_QUENCH) { 1399 /* 1400 * Don't honor ICMP Source Quench messages meant for 1401 * TCP connections. 1402 */ 1403 return; 1404 } else if (PRC_IS_REDIRECT(cmd)) 1405 notify = in6_rtchange, d = NULL; 1406 else if (cmd == PRC_MSGSIZE) 1407 ; /* special code is present, see below */ 1408 else if (cmd == PRC_HOSTDEAD) 1409 d = NULL; 1410 else if (inet6ctlerrmap[cmd] == 0) 1411 return; 1412 1413 /* if the parameter is from icmp6, decode it. */ 1414 if (d != NULL) { 1415 struct ip6ctlparam *ip6cp = (struct ip6ctlparam *)d; 1416 m = ip6cp->ip6c_m; 1417 ip6 = ip6cp->ip6c_ip6; 1418 off = ip6cp->ip6c_off; 1419 sa6_src = ip6cp->ip6c_src; 1420 } else { 1421 m = NULL; 1422 ip6 = NULL; 1423 sa6_src = &sa6_any; 1424 off = 0; 1425 } 1426 1427 if (ip6) { 1428 /* 1429 * XXX: We assume that when ip6 is non NULL, 1430 * M and OFF are valid. 1431 */ 1432 1433 /* check if we can safely examine src and dst ports */ 1434 if (m->m_pkthdr.len < off + sizeof(th)) { 1435 if (cmd == PRC_MSGSIZE) 1436 icmp6_mtudisc_update((struct ip6ctlparam *)d, 0); 1437 return; 1438 } 1439 1440 bzero(&th, sizeof(th)); 1441 m_copydata(m, off, sizeof(th), (caddr_t)&th); 1442 1443 if (cmd == PRC_MSGSIZE) { 1444 int valid = 0; 1445 1446 /* 1447 * Check to see if we have a valid TCP connection 1448 * corresponding to the address in the ICMPv6 message 1449 * payload. 1450 */ 1451 if (in6_pcblookup_connect(&tcbtable, &sa6->sin6_addr, 1452 th.th_dport, (const struct in6_addr *)&sa6_src->sin6_addr, 1453 th.th_sport, 0)) 1454 valid++; 1455 1456 /* 1457 * Depending on the value of "valid" and routing table 1458 * size (mtudisc_{hi,lo}wat), we will: 1459 * - recalcurate the new MTU and create the 1460 * corresponding routing entry, or 1461 * - ignore the MTU change notification. 1462 */ 1463 icmp6_mtudisc_update((struct ip6ctlparam *)d, valid); 1464 1465 /* 1466 * no need to call in6_pcbnotify, it should have been 1467 * called via callback if necessary 1468 */ 1469 return; 1470 } 1471 1472 nmatch = in6_pcbnotify(&tcbtable, sa, th.th_dport, 1473 (const struct sockaddr *)sa6_src, th.th_sport, cmd, NULL, notify); 1474 if (nmatch == 0 && syn_cache_count && 1475 (inet6ctlerrmap[cmd] == EHOSTUNREACH || 1476 inet6ctlerrmap[cmd] == ENETUNREACH || 1477 inet6ctlerrmap[cmd] == EHOSTDOWN)) 1478 syn_cache_unreach((const struct sockaddr *)sa6_src, 1479 sa, &th); 1480 } else { 1481 (void) in6_pcbnotify(&tcbtable, sa, 0, 1482 (const struct sockaddr *)sa6_src, 0, cmd, NULL, notify); 1483 } 1484 } 1485 #endif 1486 1487 #ifdef INET 1488 /* assumes that ip header and tcp header are contiguous on mbuf */ 1489 void * 1490 tcp_ctlinput(int cmd, struct sockaddr *sa, void *v) 1491 { 1492 struct ip *ip = v; 1493 struct tcphdr *th; 1494 struct icmp *icp; 1495 extern const int inetctlerrmap[]; 1496 void (*notify)(struct inpcb *, int) = tcp_notify; 1497 int errno; 1498 int nmatch; 1499 struct tcpcb *tp; 1500 u_int mtu; 1501 tcp_seq seq; 1502 struct inpcb *inp; 1503 #ifdef INET6 1504 struct in6pcb *in6p; 1505 struct in6_addr src6, dst6; 1506 #endif 1507 1508 if (sa->sa_family != AF_INET || 1509 sa->sa_len != sizeof(struct sockaddr_in)) 1510 return NULL; 1511 if ((unsigned)cmd >= PRC_NCMDS) 1512 return NULL; 1513 errno = inetctlerrmap[cmd]; 1514 if (cmd == PRC_QUENCH) 1515 /* 1516 * Don't honor ICMP Source Quench messages meant for 1517 * TCP connections. 1518 */ 1519 return NULL; 1520 else if (PRC_IS_REDIRECT(cmd)) 1521 notify = in_rtchange, ip = 0; 1522 else if (cmd == PRC_MSGSIZE && ip && ip->ip_v == 4) { 1523 /* 1524 * Check to see if we have a valid TCP connection 1525 * corresponding to the address in the ICMP message 1526 * payload. 1527 * 1528 * Boundary check is made in icmp_input(), with ICMP_ADVLENMIN. 1529 */ 1530 th = (struct tcphdr *)((caddr_t)ip + (ip->ip_hl << 2)); 1531 #ifdef INET6 1532 memset(&src6, 0, sizeof(src6)); 1533 memset(&dst6, 0, sizeof(dst6)); 1534 src6.s6_addr16[5] = dst6.s6_addr16[5] = 0xffff; 1535 memcpy(&src6.s6_addr32[3], &ip->ip_src, sizeof(struct in_addr)); 1536 memcpy(&dst6.s6_addr32[3], &ip->ip_dst, sizeof(struct in_addr)); 1537 #endif 1538 if ((inp = in_pcblookup_connect(&tcbtable, ip->ip_dst, 1539 th->th_dport, ip->ip_src, th->th_sport)) != NULL) 1540 #ifdef INET6 1541 in6p = NULL; 1542 #else 1543 ; 1544 #endif 1545 #ifdef INET6 1546 else if ((in6p = in6_pcblookup_connect(&tcbtable, &dst6, 1547 th->th_dport, &src6, th->th_sport, 0)) != NULL) 1548 ; 1549 #endif 1550 else 1551 return NULL; 1552 1553 /* 1554 * Now that we've validated that we are actually communicating 1555 * with the host indicated in the ICMP message, locate the 1556 * ICMP header, recalculate the new MTU, and create the 1557 * corresponding routing entry. 1558 */ 1559 icp = (struct icmp *)((caddr_t)ip - 1560 offsetof(struct icmp, icmp_ip)); 1561 if (inp) { 1562 if ((tp = intotcpcb(inp)) == NULL) 1563 return NULL; 1564 } 1565 #ifdef INET6 1566 else if (in6p) { 1567 if ((tp = in6totcpcb(in6p)) == NULL) 1568 return NULL; 1569 } 1570 #endif 1571 else 1572 return NULL; 1573 seq = ntohl(th->th_seq); 1574 if (SEQ_LT(seq, tp->snd_una) || SEQ_GT(seq, tp->snd_max)) 1575 return NULL; 1576 /* 1577 * If the ICMP message advertises a Next-Hop MTU 1578 * equal or larger than the maximum packet size we have 1579 * ever sent, drop the message. 1580 */ 1581 mtu = (u_int)ntohs(icp->icmp_nextmtu); 1582 if (mtu >= tp->t_pmtud_mtu_sent) 1583 return NULL; 1584 if (mtu >= tcp_hdrsz(tp) + tp->t_pmtud_mss_acked) { 1585 /* 1586 * Calculate new MTU, and create corresponding 1587 * route (traditional PMTUD). 1588 */ 1589 tp->t_flags &= ~TF_PMTUD_PEND; 1590 icmp_mtudisc(icp, ip->ip_dst); 1591 } else { 1592 /* 1593 * Record the information got in the ICMP 1594 * message; act on it later. 1595 * If we had already recorded an ICMP message, 1596 * replace the old one only if the new message 1597 * refers to an older TCP segment 1598 */ 1599 if (tp->t_flags & TF_PMTUD_PEND) { 1600 if (SEQ_LT(tp->t_pmtud_th_seq, seq)) 1601 return NULL; 1602 } else 1603 tp->t_flags |= TF_PMTUD_PEND; 1604 tp->t_pmtud_th_seq = seq; 1605 tp->t_pmtud_nextmtu = icp->icmp_nextmtu; 1606 tp->t_pmtud_ip_len = icp->icmp_ip.ip_len; 1607 tp->t_pmtud_ip_hl = icp->icmp_ip.ip_hl; 1608 } 1609 return NULL; 1610 } else if (cmd == PRC_HOSTDEAD) 1611 ip = 0; 1612 else if (errno == 0) 1613 return NULL; 1614 if (ip && ip->ip_v == 4 && sa->sa_family == AF_INET) { 1615 th = (struct tcphdr *)((caddr_t)ip + (ip->ip_hl << 2)); 1616 nmatch = in_pcbnotify(&tcbtable, satosin(sa)->sin_addr, 1617 th->th_dport, ip->ip_src, th->th_sport, errno, notify); 1618 if (nmatch == 0 && syn_cache_count && 1619 (inetctlerrmap[cmd] == EHOSTUNREACH || 1620 inetctlerrmap[cmd] == ENETUNREACH || 1621 inetctlerrmap[cmd] == EHOSTDOWN)) { 1622 struct sockaddr_in sin; 1623 bzero(&sin, sizeof(sin)); 1624 sin.sin_len = sizeof(sin); 1625 sin.sin_family = AF_INET; 1626 sin.sin_port = th->th_sport; 1627 sin.sin_addr = ip->ip_src; 1628 syn_cache_unreach((struct sockaddr *)&sin, sa, th); 1629 } 1630 1631 /* XXX mapped address case */ 1632 } else 1633 in_pcbnotifyall(&tcbtable, satosin(sa)->sin_addr, errno, 1634 notify); 1635 return NULL; 1636 } 1637 1638 /* 1639 * When a source quench is received, we are being notified of congestion. 1640 * Close the congestion window down to the Loss Window (one segment). 1641 * We will gradually open it again as we proceed. 1642 */ 1643 void 1644 tcp_quench(struct inpcb *inp, int errno) 1645 { 1646 struct tcpcb *tp = intotcpcb(inp); 1647 1648 if (tp) 1649 tp->snd_cwnd = tp->t_segsz; 1650 } 1651 #endif 1652 1653 #ifdef INET6 1654 void 1655 tcp6_quench(struct in6pcb *in6p, int errno) 1656 { 1657 struct tcpcb *tp = in6totcpcb(in6p); 1658 1659 if (tp) 1660 tp->snd_cwnd = tp->t_segsz; 1661 } 1662 #endif 1663 1664 #ifdef INET 1665 /* 1666 * Path MTU Discovery handlers. 1667 */ 1668 void 1669 tcp_mtudisc_callback(struct in_addr faddr) 1670 { 1671 #ifdef INET6 1672 struct in6_addr in6; 1673 #endif 1674 1675 in_pcbnotifyall(&tcbtable, faddr, EMSGSIZE, tcp_mtudisc); 1676 #ifdef INET6 1677 memset(&in6, 0, sizeof(in6)); 1678 in6.s6_addr16[5] = 0xffff; 1679 memcpy(&in6.s6_addr32[3], &faddr, sizeof(struct in_addr)); 1680 tcp6_mtudisc_callback(&in6); 1681 #endif 1682 } 1683 1684 /* 1685 * On receipt of path MTU corrections, flush old route and replace it 1686 * with the new one. Retransmit all unacknowledged packets, to ensure 1687 * that all packets will be received. 1688 */ 1689 void 1690 tcp_mtudisc(struct inpcb *inp, int errno) 1691 { 1692 struct tcpcb *tp = intotcpcb(inp); 1693 struct rtentry *rt = in_pcbrtentry(inp); 1694 1695 if (tp != 0) { 1696 if (rt != 0) { 1697 /* 1698 * If this was not a host route, remove and realloc. 1699 */ 1700 if ((rt->rt_flags & RTF_HOST) == 0) { 1701 in_rtchange(inp, errno); 1702 if ((rt = in_pcbrtentry(inp)) == 0) 1703 return; 1704 } 1705 1706 /* 1707 * Slow start out of the error condition. We 1708 * use the MTU because we know it's smaller 1709 * than the previously transmitted segment. 1710 * 1711 * Note: This is more conservative than the 1712 * suggestion in draft-floyd-incr-init-win-03. 1713 */ 1714 if (rt->rt_rmx.rmx_mtu != 0) 1715 tp->snd_cwnd = 1716 TCP_INITIAL_WINDOW(tcp_init_win, 1717 rt->rt_rmx.rmx_mtu); 1718 } 1719 1720 /* 1721 * Resend unacknowledged packets. 1722 */ 1723 tp->snd_nxt = tp->snd_una; 1724 tcp_output(tp); 1725 } 1726 } 1727 #endif 1728 1729 #ifdef INET6 1730 /* 1731 * Path MTU Discovery handlers. 1732 */ 1733 void 1734 tcp6_mtudisc_callback(struct in6_addr *faddr) 1735 { 1736 struct sockaddr_in6 sin6; 1737 1738 bzero(&sin6, sizeof(sin6)); 1739 sin6.sin6_family = AF_INET6; 1740 sin6.sin6_len = sizeof(struct sockaddr_in6); 1741 sin6.sin6_addr = *faddr; 1742 (void) in6_pcbnotify(&tcbtable, (struct sockaddr *)&sin6, 0, 1743 (const struct sockaddr *)&sa6_any, 0, PRC_MSGSIZE, NULL, tcp6_mtudisc); 1744 } 1745 1746 void 1747 tcp6_mtudisc(struct in6pcb *in6p, int errno) 1748 { 1749 struct tcpcb *tp = in6totcpcb(in6p); 1750 struct rtentry *rt = in6_pcbrtentry(in6p); 1751 1752 if (tp != 0) { 1753 if (rt != 0) { 1754 /* 1755 * If this was not a host route, remove and realloc. 1756 */ 1757 if ((rt->rt_flags & RTF_HOST) == 0) { 1758 in6_rtchange(in6p, errno); 1759 if ((rt = in6_pcbrtentry(in6p)) == 0) 1760 return; 1761 } 1762 1763 /* 1764 * Slow start out of the error condition. We 1765 * use the MTU because we know it's smaller 1766 * than the previously transmitted segment. 1767 * 1768 * Note: This is more conservative than the 1769 * suggestion in draft-floyd-incr-init-win-03. 1770 */ 1771 if (rt->rt_rmx.rmx_mtu != 0) 1772 tp->snd_cwnd = 1773 TCP_INITIAL_WINDOW(tcp_init_win, 1774 rt->rt_rmx.rmx_mtu); 1775 } 1776 1777 /* 1778 * Resend unacknowledged packets. 1779 */ 1780 tp->snd_nxt = tp->snd_una; 1781 tcp_output(tp); 1782 } 1783 } 1784 #endif /* INET6 */ 1785 1786 /* 1787 * Compute the MSS to advertise to the peer. Called only during 1788 * the 3-way handshake. If we are the server (peer initiated 1789 * connection), we are called with a pointer to the interface 1790 * on which the SYN packet arrived. If we are the client (we 1791 * initiated connection), we are called with a pointer to the 1792 * interface out which this connection should go. 1793 * 1794 * NOTE: Do not subtract IP option/extension header size nor IPsec 1795 * header size from MSS advertisement. MSS option must hold the maximum 1796 * segment size we can accept, so it must always be: 1797 * max(if mtu) - ip header - tcp header 1798 */ 1799 u_long 1800 tcp_mss_to_advertise(const struct ifnet *ifp, int af) 1801 { 1802 extern u_long in_maxmtu; 1803 u_long mss = 0; 1804 u_long hdrsiz; 1805 1806 /* 1807 * In order to avoid defeating path MTU discovery on the peer, 1808 * we advertise the max MTU of all attached networks as our MSS, 1809 * per RFC 1191, section 3.1. 1810 * 1811 * We provide the option to advertise just the MTU of 1812 * the interface on which we hope this connection will 1813 * be receiving. If we are responding to a SYN, we 1814 * will have a pretty good idea about this, but when 1815 * initiating a connection there is a bit more doubt. 1816 * 1817 * We also need to ensure that loopback has a large enough 1818 * MSS, as the loopback MTU is never included in in_maxmtu. 1819 */ 1820 1821 if (ifp != NULL) 1822 switch (af) { 1823 case AF_INET: 1824 mss = ifp->if_mtu; 1825 break; 1826 #ifdef INET6 1827 case AF_INET6: 1828 mss = IN6_LINKMTU(ifp); 1829 break; 1830 #endif 1831 } 1832 1833 if (tcp_mss_ifmtu == 0) 1834 switch (af) { 1835 case AF_INET: 1836 mss = max(in_maxmtu, mss); 1837 break; 1838 #ifdef INET6 1839 case AF_INET6: 1840 mss = max(in6_maxmtu, mss); 1841 break; 1842 #endif 1843 } 1844 1845 switch (af) { 1846 case AF_INET: 1847 hdrsiz = sizeof(struct ip); 1848 break; 1849 #ifdef INET6 1850 case AF_INET6: 1851 hdrsiz = sizeof(struct ip6_hdr); 1852 break; 1853 #endif 1854 default: 1855 hdrsiz = 0; 1856 break; 1857 } 1858 hdrsiz += sizeof(struct tcphdr); 1859 if (mss > hdrsiz) 1860 mss -= hdrsiz; 1861 1862 mss = max(tcp_mssdflt, mss); 1863 return (mss); 1864 } 1865 1866 /* 1867 * Set connection variables based on the peer's advertised MSS. 1868 * We are passed the TCPCB for the actual connection. If we 1869 * are the server, we are called by the compressed state engine 1870 * when the 3-way handshake is complete. If we are the client, 1871 * we are called when we receive the SYN,ACK from the server. 1872 * 1873 * NOTE: Our advertised MSS value must be initialized in the TCPCB 1874 * before this routine is called! 1875 */ 1876 void 1877 tcp_mss_from_peer(struct tcpcb *tp, int offer) 1878 { 1879 struct socket *so; 1880 #if defined(RTV_SPIPE) || defined(RTV_SSTHRESH) 1881 struct rtentry *rt; 1882 #endif 1883 u_long bufsize; 1884 int mss; 1885 1886 #ifdef DIAGNOSTIC 1887 if (tp->t_inpcb && tp->t_in6pcb) 1888 panic("tcp_mss_from_peer: both t_inpcb and t_in6pcb are set"); 1889 #endif 1890 so = NULL; 1891 rt = NULL; 1892 #ifdef INET 1893 if (tp->t_inpcb) { 1894 so = tp->t_inpcb->inp_socket; 1895 #if defined(RTV_SPIPE) || defined(RTV_SSTHRESH) 1896 rt = in_pcbrtentry(tp->t_inpcb); 1897 #endif 1898 } 1899 #endif 1900 #ifdef INET6 1901 if (tp->t_in6pcb) { 1902 so = tp->t_in6pcb->in6p_socket; 1903 #if defined(RTV_SPIPE) || defined(RTV_SSTHRESH) 1904 rt = in6_pcbrtentry(tp->t_in6pcb); 1905 #endif 1906 } 1907 #endif 1908 1909 /* 1910 * As per RFC1122, use the default MSS value, unless they 1911 * sent us an offer. Do not accept offers less than 256 bytes. 1912 */ 1913 mss = tcp_mssdflt; 1914 if (offer) 1915 mss = offer; 1916 mss = max(mss, 256); /* sanity */ 1917 tp->t_peermss = mss; 1918 mss -= tcp_optlen(tp); 1919 #ifdef INET 1920 if (tp->t_inpcb) 1921 mss -= ip_optlen(tp->t_inpcb); 1922 #endif 1923 #ifdef INET6 1924 if (tp->t_in6pcb) 1925 mss -= ip6_optlen(tp->t_in6pcb); 1926 #endif 1927 1928 /* 1929 * If there's a pipesize, change the socket buffer to that size. 1930 * Make the socket buffer an integral number of MSS units. If 1931 * the MSS is larger than the socket buffer, artificially decrease 1932 * the MSS. 1933 */ 1934 #ifdef RTV_SPIPE 1935 if (rt != NULL && rt->rt_rmx.rmx_sendpipe != 0) 1936 bufsize = rt->rt_rmx.rmx_sendpipe; 1937 else 1938 #endif 1939 { 1940 KASSERT(so != NULL); 1941 bufsize = so->so_snd.sb_hiwat; 1942 } 1943 if (bufsize < mss) 1944 mss = bufsize; 1945 else { 1946 bufsize = roundup(bufsize, mss); 1947 if (bufsize > sb_max) 1948 bufsize = sb_max; 1949 (void) sbreserve(&so->so_snd, bufsize, so); 1950 } 1951 tp->t_segsz = mss; 1952 1953 #ifdef RTV_SSTHRESH 1954 if (rt != NULL && rt->rt_rmx.rmx_ssthresh) { 1955 /* 1956 * There's some sort of gateway or interface buffer 1957 * limit on the path. Use this to set the slow 1958 * start threshold, but set the threshold to no less 1959 * than 2 * MSS. 1960 */ 1961 tp->snd_ssthresh = max(2 * mss, rt->rt_rmx.rmx_ssthresh); 1962 } 1963 #endif 1964 } 1965 1966 /* 1967 * Processing necessary when a TCP connection is established. 1968 */ 1969 void 1970 tcp_established(struct tcpcb *tp) 1971 { 1972 struct socket *so; 1973 #ifdef RTV_RPIPE 1974 struct rtentry *rt; 1975 #endif 1976 u_long bufsize; 1977 1978 #ifdef DIAGNOSTIC 1979 if (tp->t_inpcb && tp->t_in6pcb) 1980 panic("tcp_established: both t_inpcb and t_in6pcb are set"); 1981 #endif 1982 so = NULL; 1983 rt = NULL; 1984 #ifdef INET 1985 if (tp->t_inpcb) { 1986 so = tp->t_inpcb->inp_socket; 1987 #if defined(RTV_RPIPE) 1988 rt = in_pcbrtentry(tp->t_inpcb); 1989 #endif 1990 } 1991 #endif 1992 #ifdef INET6 1993 if (tp->t_in6pcb) { 1994 so = tp->t_in6pcb->in6p_socket; 1995 #if defined(RTV_RPIPE) 1996 rt = in6_pcbrtentry(tp->t_in6pcb); 1997 #endif 1998 } 1999 #endif 2000 2001 tp->t_state = TCPS_ESTABLISHED; 2002 TCP_TIMER_ARM(tp, TCPT_KEEP, tcp_keepidle); 2003 2004 #ifdef RTV_RPIPE 2005 if (rt != NULL && rt->rt_rmx.rmx_recvpipe != 0) 2006 bufsize = rt->rt_rmx.rmx_recvpipe; 2007 else 2008 #endif 2009 { 2010 KASSERT(so != NULL); 2011 bufsize = so->so_rcv.sb_hiwat; 2012 } 2013 if (bufsize > tp->t_ourmss) { 2014 bufsize = roundup(bufsize, tp->t_ourmss); 2015 if (bufsize > sb_max) 2016 bufsize = sb_max; 2017 (void) sbreserve(&so->so_rcv, bufsize, so); 2018 } 2019 } 2020 2021 /* 2022 * Check if there's an initial rtt or rttvar. Convert from the 2023 * route-table units to scaled multiples of the slow timeout timer. 2024 * Called only during the 3-way handshake. 2025 */ 2026 void 2027 tcp_rmx_rtt(struct tcpcb *tp) 2028 { 2029 #ifdef RTV_RTT 2030 struct rtentry *rt = NULL; 2031 int rtt; 2032 2033 #ifdef DIAGNOSTIC 2034 if (tp->t_inpcb && tp->t_in6pcb) 2035 panic("tcp_rmx_rtt: both t_inpcb and t_in6pcb are set"); 2036 #endif 2037 #ifdef INET 2038 if (tp->t_inpcb) 2039 rt = in_pcbrtentry(tp->t_inpcb); 2040 #endif 2041 #ifdef INET6 2042 if (tp->t_in6pcb) 2043 rt = in6_pcbrtentry(tp->t_in6pcb); 2044 #endif 2045 if (rt == NULL) 2046 return; 2047 2048 if (tp->t_srtt == 0 && (rtt = rt->rt_rmx.rmx_rtt)) { 2049 /* 2050 * XXX The lock bit for MTU indicates that the value 2051 * is also a minimum value; this is subject to time. 2052 */ 2053 if (rt->rt_rmx.rmx_locks & RTV_RTT) 2054 TCPT_RANGESET(tp->t_rttmin, 2055 rtt / (RTM_RTTUNIT / PR_SLOWHZ), 2056 TCPTV_MIN, TCPTV_REXMTMAX); 2057 tp->t_srtt = rtt / 2058 ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTT_SHIFT + 2)); 2059 if (rt->rt_rmx.rmx_rttvar) { 2060 tp->t_rttvar = rt->rt_rmx.rmx_rttvar / 2061 ((RTM_RTTUNIT / PR_SLOWHZ) >> 2062 (TCP_RTTVAR_SHIFT + 2)); 2063 } else { 2064 /* Default variation is +- 1 rtt */ 2065 tp->t_rttvar = 2066 tp->t_srtt >> (TCP_RTT_SHIFT - TCP_RTTVAR_SHIFT); 2067 } 2068 TCPT_RANGESET(tp->t_rxtcur, 2069 ((tp->t_srtt >> 2) + tp->t_rttvar) >> (1 + 2), 2070 tp->t_rttmin, TCPTV_REXMTMAX); 2071 } 2072 #endif 2073 } 2074 2075 tcp_seq tcp_iss_seq = 0; /* tcp initial seq # */ 2076 #if NRND > 0 2077 u_int8_t tcp_iss_secret[16]; /* 128 bits; should be plenty */ 2078 #endif 2079 2080 /* 2081 * Get a new sequence value given a tcp control block 2082 */ 2083 tcp_seq 2084 tcp_new_iss(struct tcpcb *tp, tcp_seq addin) 2085 { 2086 2087 #ifdef INET 2088 if (tp->t_inpcb != NULL) { 2089 return (tcp_new_iss1(&tp->t_inpcb->inp_laddr, 2090 &tp->t_inpcb->inp_faddr, tp->t_inpcb->inp_lport, 2091 tp->t_inpcb->inp_fport, sizeof(tp->t_inpcb->inp_laddr), 2092 addin)); 2093 } 2094 #endif 2095 #ifdef INET6 2096 if (tp->t_in6pcb != NULL) { 2097 return (tcp_new_iss1(&tp->t_in6pcb->in6p_laddr, 2098 &tp->t_in6pcb->in6p_faddr, tp->t_in6pcb->in6p_lport, 2099 tp->t_in6pcb->in6p_fport, sizeof(tp->t_in6pcb->in6p_laddr), 2100 addin)); 2101 } 2102 #endif 2103 /* Not possible. */ 2104 panic("tcp_new_iss"); 2105 } 2106 2107 /* 2108 * This routine actually generates a new TCP initial sequence number. 2109 */ 2110 tcp_seq 2111 tcp_new_iss1(void *laddr, void *faddr, u_int16_t lport, u_int16_t fport, 2112 size_t addrsz, tcp_seq addin) 2113 { 2114 tcp_seq tcp_iss; 2115 2116 #if NRND > 0 2117 static int beenhere; 2118 2119 /* 2120 * If we haven't been here before, initialize our cryptographic 2121 * hash secret. 2122 */ 2123 if (beenhere == 0) { 2124 rnd_extract_data(tcp_iss_secret, sizeof(tcp_iss_secret), 2125 RND_EXTRACT_ANY); 2126 beenhere = 1; 2127 } 2128 2129 if (tcp_do_rfc1948) { 2130 MD5_CTX ctx; 2131 u_int8_t hash[16]; /* XXX MD5 knowledge */ 2132 2133 /* 2134 * Compute the base value of the ISS. It is a hash 2135 * of (saddr, sport, daddr, dport, secret). 2136 */ 2137 MD5Init(&ctx); 2138 2139 MD5Update(&ctx, (u_char *) laddr, addrsz); 2140 MD5Update(&ctx, (u_char *) &lport, sizeof(lport)); 2141 2142 MD5Update(&ctx, (u_char *) faddr, addrsz); 2143 MD5Update(&ctx, (u_char *) &fport, sizeof(fport)); 2144 2145 MD5Update(&ctx, tcp_iss_secret, sizeof(tcp_iss_secret)); 2146 2147 MD5Final(hash, &ctx); 2148 2149 memcpy(&tcp_iss, hash, sizeof(tcp_iss)); 2150 2151 /* 2152 * Now increment our "timer", and add it in to 2153 * the computed value. 2154 * 2155 * XXX Use `addin'? 2156 * XXX TCP_ISSINCR too large to use? 2157 */ 2158 tcp_iss_seq += TCP_ISSINCR; 2159 #ifdef TCPISS_DEBUG 2160 printf("ISS hash 0x%08x, ", tcp_iss); 2161 #endif 2162 tcp_iss += tcp_iss_seq + addin; 2163 #ifdef TCPISS_DEBUG 2164 printf("new ISS 0x%08x\n", tcp_iss); 2165 #endif 2166 } else 2167 #endif /* NRND > 0 */ 2168 { 2169 /* 2170 * Randomize. 2171 */ 2172 #if NRND > 0 2173 rnd_extract_data(&tcp_iss, sizeof(tcp_iss), RND_EXTRACT_ANY); 2174 #else 2175 tcp_iss = arc4random(); 2176 #endif 2177 2178 /* 2179 * If we were asked to add some amount to a known value, 2180 * we will take a random value obtained above, mask off 2181 * the upper bits, and add in the known value. We also 2182 * add in a constant to ensure that we are at least a 2183 * certain distance from the original value. 2184 * 2185 * This is used when an old connection is in timed wait 2186 * and we have a new one coming in, for instance. 2187 */ 2188 if (addin != 0) { 2189 #ifdef TCPISS_DEBUG 2190 printf("Random %08x, ", tcp_iss); 2191 #endif 2192 tcp_iss &= TCP_ISS_RANDOM_MASK; 2193 tcp_iss += addin + TCP_ISSINCR; 2194 #ifdef TCPISS_DEBUG 2195 printf("Old ISS %08x, ISS %08x\n", addin, tcp_iss); 2196 #endif 2197 } else { 2198 tcp_iss &= TCP_ISS_RANDOM_MASK; 2199 tcp_iss += tcp_iss_seq; 2200 tcp_iss_seq += TCP_ISSINCR; 2201 #ifdef TCPISS_DEBUG 2202 printf("ISS %08x\n", tcp_iss); 2203 #endif 2204 } 2205 } 2206 2207 if (tcp_compat_42) { 2208 /* 2209 * Limit it to the positive range for really old TCP 2210 * implementations. 2211 * Just AND off the top bit instead of checking if 2212 * is set first - saves a branch 50% of the time. 2213 */ 2214 tcp_iss &= 0x7fffffff; /* XXX */ 2215 } 2216 2217 return (tcp_iss); 2218 } 2219 2220 #if defined(IPSEC) || defined(FAST_IPSEC) 2221 /* compute ESP/AH header size for TCP, including outer IP header. */ 2222 size_t 2223 ipsec4_hdrsiz_tcp(struct tcpcb *tp) 2224 { 2225 struct inpcb *inp; 2226 size_t hdrsiz; 2227 2228 /* XXX mapped addr case (tp->t_in6pcb) */ 2229 if (!tp || !tp->t_template || !(inp = tp->t_inpcb)) 2230 return 0; 2231 switch (tp->t_family) { 2232 case AF_INET: 2233 /* XXX: should use currect direction. */ 2234 hdrsiz = ipsec4_hdrsiz(tp->t_template, IPSEC_DIR_OUTBOUND, inp); 2235 break; 2236 default: 2237 hdrsiz = 0; 2238 break; 2239 } 2240 2241 return hdrsiz; 2242 } 2243 2244 #ifdef INET6 2245 size_t 2246 ipsec6_hdrsiz_tcp(struct tcpcb *tp) 2247 { 2248 struct in6pcb *in6p; 2249 size_t hdrsiz; 2250 2251 if (!tp || !tp->t_template || !(in6p = tp->t_in6pcb)) 2252 return 0; 2253 switch (tp->t_family) { 2254 case AF_INET6: 2255 /* XXX: should use currect direction. */ 2256 hdrsiz = ipsec6_hdrsiz(tp->t_template, IPSEC_DIR_OUTBOUND, in6p); 2257 break; 2258 case AF_INET: 2259 /* mapped address case - tricky */ 2260 default: 2261 hdrsiz = 0; 2262 break; 2263 } 2264 2265 return hdrsiz; 2266 } 2267 #endif 2268 #endif /*IPSEC*/ 2269 2270 /* 2271 * Determine the length of the TCP options for this connection. 2272 * 2273 * XXX: What do we do for SACK, when we add that? Just reserve 2274 * all of the space? Otherwise we can't exactly be incrementing 2275 * cwnd by an amount that varies depending on the amount we last 2276 * had to SACK! 2277 */ 2278 2279 u_int 2280 tcp_optlen(struct tcpcb *tp) 2281 { 2282 u_int optlen; 2283 2284 optlen = 0; 2285 if ((tp->t_flags & (TF_REQ_TSTMP|TF_RCVD_TSTMP|TF_NOOPT)) == 2286 (TF_REQ_TSTMP | TF_RCVD_TSTMP)) 2287 optlen += TCPOLEN_TSTAMP_APPA; 2288 2289 #ifdef TCP_SIGNATURE 2290 #if defined(INET6) && defined(FAST_IPSEC) 2291 if (tp->t_family == AF_INET) 2292 #endif 2293 if (tp->t_flags & TF_SIGNATURE) 2294 optlen += TCPOLEN_SIGNATURE + 2; 2295 #endif /* TCP_SIGNATURE */ 2296 2297 return optlen; 2298 } 2299 2300 u_int 2301 tcp_hdrsz(struct tcpcb *tp) 2302 { 2303 u_int hlen; 2304 2305 switch (tp->t_family) { 2306 #ifdef INET6 2307 case AF_INET6: 2308 hlen = sizeof(struct ip6_hdr); 2309 break; 2310 #endif 2311 case AF_INET: 2312 hlen = sizeof(struct ip); 2313 break; 2314 default: 2315 hlen = 0; 2316 break; 2317 } 2318 hlen += sizeof(struct tcphdr); 2319 2320 if ((tp->t_flags & (TF_REQ_TSTMP|TF_NOOPT)) == TF_REQ_TSTMP && 2321 (tp->t_flags & TF_RCVD_TSTMP) == TF_RCVD_TSTMP) 2322 hlen += TCPOLEN_TSTAMP_APPA; 2323 #ifdef TCP_SIGNATURE 2324 if (tp->t_flags & TF_SIGNATURE) 2325 hlen += TCPOLEN_SIGLEN; 2326 #endif 2327 return hlen; 2328 } 2329