tcp_subr.c revision 1.198
1 /* $NetBSD: tcp_subr.c,v 1.198 2006/04/15 02:30:39 christos 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 1.198 2006/04/15 02:30:39 christos 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 int tcp_do_loopback_cksum = 0; 204 int tcp_sack_tp_maxholes = 32; 205 int tcp_sack_globalmaxholes = 1024; 206 int tcp_sack_globalholes = 0; 207 208 209 /* tcb hash */ 210 #ifndef TCBHASHSIZE 211 #define TCBHASHSIZE 128 212 #endif 213 int tcbhashsize = TCBHASHSIZE; 214 215 /* syn hash parameters */ 216 #define TCP_SYN_HASH_SIZE 293 217 #define TCP_SYN_BUCKET_SIZE 35 218 int tcp_syn_cache_size = TCP_SYN_HASH_SIZE; 219 int tcp_syn_cache_limit = TCP_SYN_HASH_SIZE*TCP_SYN_BUCKET_SIZE; 220 int tcp_syn_bucket_limit = 3*TCP_SYN_BUCKET_SIZE; 221 struct syn_cache_head tcp_syn_cache[TCP_SYN_HASH_SIZE]; 222 223 int tcp_freeq(struct tcpcb *); 224 225 #ifdef INET 226 void tcp_mtudisc_callback(struct in_addr); 227 #endif 228 #ifdef INET6 229 void tcp6_mtudisc_callback(struct in6_addr *); 230 #endif 231 232 #ifdef INET6 233 void tcp6_mtudisc(struct in6pcb *, int); 234 #endif 235 236 POOL_INIT(tcpcb_pool, sizeof(struct tcpcb), 0, 0, 0, "tcpcbpl", NULL); 237 238 #ifdef TCP_CSUM_COUNTERS 239 #include <sys/device.h> 240 241 #if defined(INET) 242 struct evcnt tcp_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 243 NULL, "tcp", "hwcsum bad"); 244 struct evcnt tcp_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 245 NULL, "tcp", "hwcsum ok"); 246 struct evcnt tcp_hwcsum_data = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 247 NULL, "tcp", "hwcsum data"); 248 struct evcnt tcp_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 249 NULL, "tcp", "swcsum"); 250 251 EVCNT_ATTACH_STATIC(tcp_hwcsum_bad); 252 EVCNT_ATTACH_STATIC(tcp_hwcsum_ok); 253 EVCNT_ATTACH_STATIC(tcp_hwcsum_data); 254 EVCNT_ATTACH_STATIC(tcp_swcsum); 255 #endif /* defined(INET) */ 256 257 #if defined(INET6) 258 struct evcnt tcp6_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 259 NULL, "tcp6", "hwcsum bad"); 260 struct evcnt tcp6_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 261 NULL, "tcp6", "hwcsum ok"); 262 struct evcnt tcp6_hwcsum_data = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 263 NULL, "tcp6", "hwcsum data"); 264 struct evcnt tcp6_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 265 NULL, "tcp6", "swcsum"); 266 267 EVCNT_ATTACH_STATIC(tcp6_hwcsum_bad); 268 EVCNT_ATTACH_STATIC(tcp6_hwcsum_ok); 269 EVCNT_ATTACH_STATIC(tcp6_hwcsum_data); 270 EVCNT_ATTACH_STATIC(tcp6_swcsum); 271 #endif /* defined(INET6) */ 272 #endif /* TCP_CSUM_COUNTERS */ 273 274 275 #ifdef TCP_OUTPUT_COUNTERS 276 #include <sys/device.h> 277 278 struct evcnt tcp_output_bigheader = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 279 NULL, "tcp", "output big header"); 280 struct evcnt tcp_output_predict_hit = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 281 NULL, "tcp", "output predict hit"); 282 struct evcnt tcp_output_predict_miss = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 283 NULL, "tcp", "output predict miss"); 284 struct evcnt tcp_output_copysmall = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 285 NULL, "tcp", "output copy small"); 286 struct evcnt tcp_output_copybig = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 287 NULL, "tcp", "output copy big"); 288 struct evcnt tcp_output_refbig = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 289 NULL, "tcp", "output reference big"); 290 291 EVCNT_ATTACH_STATIC(tcp_output_bigheader); 292 EVCNT_ATTACH_STATIC(tcp_output_predict_hit); 293 EVCNT_ATTACH_STATIC(tcp_output_predict_miss); 294 EVCNT_ATTACH_STATIC(tcp_output_copysmall); 295 EVCNT_ATTACH_STATIC(tcp_output_copybig); 296 EVCNT_ATTACH_STATIC(tcp_output_refbig); 297 298 #endif /* TCP_OUTPUT_COUNTERS */ 299 300 #ifdef TCP_REASS_COUNTERS 301 #include <sys/device.h> 302 303 struct evcnt tcp_reass_ = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 304 NULL, "tcp_reass", "calls"); 305 struct evcnt tcp_reass_empty = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 306 &tcp_reass_, "tcp_reass", "insert into empty queue"); 307 struct evcnt tcp_reass_iteration[8] = { 308 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", ">7 iterations"), 309 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "1 iteration"), 310 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "2 iterations"), 311 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "3 iterations"), 312 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "4 iterations"), 313 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "5 iterations"), 314 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "6 iterations"), 315 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "7 iterations"), 316 }; 317 struct evcnt tcp_reass_prependfirst = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 318 &tcp_reass_, "tcp_reass", "prepend to first"); 319 struct evcnt tcp_reass_prepend = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 320 &tcp_reass_, "tcp_reass", "prepend"); 321 struct evcnt tcp_reass_insert = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 322 &tcp_reass_, "tcp_reass", "insert"); 323 struct evcnt tcp_reass_inserttail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 324 &tcp_reass_, "tcp_reass", "insert at tail"); 325 struct evcnt tcp_reass_append = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 326 &tcp_reass_, "tcp_reass", "append"); 327 struct evcnt tcp_reass_appendtail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 328 &tcp_reass_, "tcp_reass", "append to tail fragment"); 329 struct evcnt tcp_reass_overlaptail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 330 &tcp_reass_, "tcp_reass", "overlap at end"); 331 struct evcnt tcp_reass_overlapfront = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 332 &tcp_reass_, "tcp_reass", "overlap at start"); 333 struct evcnt tcp_reass_segdup = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 334 &tcp_reass_, "tcp_reass", "duplicate segment"); 335 struct evcnt tcp_reass_fragdup = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 336 &tcp_reass_, "tcp_reass", "duplicate fragment"); 337 338 EVCNT_ATTACH_STATIC(tcp_reass_); 339 EVCNT_ATTACH_STATIC(tcp_reass_empty); 340 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 0); 341 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 1); 342 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 2); 343 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 3); 344 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 4); 345 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 5); 346 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 6); 347 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 7); 348 EVCNT_ATTACH_STATIC(tcp_reass_prependfirst); 349 EVCNT_ATTACH_STATIC(tcp_reass_prepend); 350 EVCNT_ATTACH_STATIC(tcp_reass_insert); 351 EVCNT_ATTACH_STATIC(tcp_reass_inserttail); 352 EVCNT_ATTACH_STATIC(tcp_reass_append); 353 EVCNT_ATTACH_STATIC(tcp_reass_appendtail); 354 EVCNT_ATTACH_STATIC(tcp_reass_overlaptail); 355 EVCNT_ATTACH_STATIC(tcp_reass_overlapfront); 356 EVCNT_ATTACH_STATIC(tcp_reass_segdup); 357 EVCNT_ATTACH_STATIC(tcp_reass_fragdup); 358 359 #endif /* TCP_REASS_COUNTERS */ 360 361 #ifdef MBUFTRACE 362 struct mowner tcp_mowner = { "tcp" }; 363 struct mowner tcp_rx_mowner = { "tcp", "rx" }; 364 struct mowner tcp_tx_mowner = { "tcp", "tx" }; 365 #endif 366 367 /* 368 * Tcp initialization 369 */ 370 void 371 tcp_init(void) 372 { 373 int hlen; 374 375 /* Initialize the TCPCB template. */ 376 tcp_tcpcb_template(); 377 378 in_pcbinit(&tcbtable, tcbhashsize, tcbhashsize); 379 380 hlen = sizeof(struct ip) + sizeof(struct tcphdr); 381 #ifdef INET6 382 if (sizeof(struct ip) < sizeof(struct ip6_hdr)) 383 hlen = sizeof(struct ip6_hdr) + sizeof(struct tcphdr); 384 #endif 385 if (max_protohdr < hlen) 386 max_protohdr = hlen; 387 if (max_linkhdr + hlen > MHLEN) 388 panic("tcp_init"); 389 390 #ifdef INET 391 icmp_mtudisc_callback_register(tcp_mtudisc_callback); 392 #endif 393 #ifdef INET6 394 icmp6_mtudisc_callback_register(tcp6_mtudisc_callback); 395 #endif 396 397 /* Initialize timer state. */ 398 tcp_timer_init(); 399 400 /* Initialize the compressed state engine. */ 401 syn_cache_init(); 402 403 MOWNER_ATTACH(&tcp_tx_mowner); 404 MOWNER_ATTACH(&tcp_rx_mowner); 405 MOWNER_ATTACH(&tcp_mowner); 406 } 407 408 /* 409 * Create template to be used to send tcp packets on a connection. 410 * Call after host entry created, allocates an mbuf and fills 411 * in a skeletal tcp/ip header, minimizing the amount of work 412 * necessary when the connection is used. 413 */ 414 struct mbuf * 415 tcp_template(struct tcpcb *tp) 416 { 417 struct inpcb *inp = tp->t_inpcb; 418 #ifdef INET6 419 struct in6pcb *in6p = tp->t_in6pcb; 420 #endif 421 struct tcphdr *n; 422 struct mbuf *m; 423 int hlen; 424 425 switch (tp->t_family) { 426 case AF_INET: 427 hlen = sizeof(struct ip); 428 if (inp) 429 break; 430 #ifdef INET6 431 if (in6p) { 432 /* mapped addr case */ 433 if (IN6_IS_ADDR_V4MAPPED(&in6p->in6p_laddr) 434 && IN6_IS_ADDR_V4MAPPED(&in6p->in6p_faddr)) 435 break; 436 } 437 #endif 438 return NULL; /*EINVAL*/ 439 #ifdef INET6 440 case AF_INET6: 441 hlen = sizeof(struct ip6_hdr); 442 if (in6p) { 443 /* more sainty check? */ 444 break; 445 } 446 return NULL; /*EINVAL*/ 447 #endif 448 default: 449 hlen = 0; /*pacify gcc*/ 450 return NULL; /*EAFNOSUPPORT*/ 451 } 452 #ifdef DIAGNOSTIC 453 if (hlen + sizeof(struct tcphdr) > MCLBYTES) 454 panic("mclbytes too small for t_template"); 455 #endif 456 m = tp->t_template; 457 if (m && m->m_len == hlen + sizeof(struct tcphdr)) 458 ; 459 else { 460 if (m) 461 m_freem(m); 462 m = tp->t_template = NULL; 463 MGETHDR(m, M_DONTWAIT, MT_HEADER); 464 if (m && hlen + sizeof(struct tcphdr) > MHLEN) { 465 MCLGET(m, M_DONTWAIT); 466 if ((m->m_flags & M_EXT) == 0) { 467 m_free(m); 468 m = NULL; 469 } 470 } 471 if (m == NULL) 472 return NULL; 473 MCLAIM(m, &tcp_mowner); 474 m->m_pkthdr.len = m->m_len = hlen + sizeof(struct tcphdr); 475 } 476 477 bzero(mtod(m, caddr_t), m->m_len); 478 479 n = (struct tcphdr *)(mtod(m, caddr_t) + hlen); 480 481 switch (tp->t_family) { 482 case AF_INET: 483 { 484 struct ipovly *ipov; 485 mtod(m, struct ip *)->ip_v = 4; 486 mtod(m, struct ip *)->ip_hl = hlen >> 2; 487 ipov = mtod(m, struct ipovly *); 488 ipov->ih_pr = IPPROTO_TCP; 489 ipov->ih_len = htons(sizeof(struct tcphdr)); 490 if (inp) { 491 ipov->ih_src = inp->inp_laddr; 492 ipov->ih_dst = inp->inp_faddr; 493 } 494 #ifdef INET6 495 else if (in6p) { 496 /* mapped addr case */ 497 bcopy(&in6p->in6p_laddr.s6_addr32[3], &ipov->ih_src, 498 sizeof(ipov->ih_src)); 499 bcopy(&in6p->in6p_faddr.s6_addr32[3], &ipov->ih_dst, 500 sizeof(ipov->ih_dst)); 501 } 502 #endif 503 /* 504 * Compute the pseudo-header portion of the checksum 505 * now. We incrementally add in the TCP option and 506 * payload lengths later, and then compute the TCP 507 * checksum right before the packet is sent off onto 508 * the wire. 509 */ 510 n->th_sum = in_cksum_phdr(ipov->ih_src.s_addr, 511 ipov->ih_dst.s_addr, 512 htons(sizeof(struct tcphdr) + IPPROTO_TCP)); 513 break; 514 } 515 #ifdef INET6 516 case AF_INET6: 517 { 518 struct ip6_hdr *ip6; 519 mtod(m, struct ip *)->ip_v = 6; 520 ip6 = mtod(m, struct ip6_hdr *); 521 ip6->ip6_nxt = IPPROTO_TCP; 522 ip6->ip6_plen = htons(sizeof(struct tcphdr)); 523 ip6->ip6_src = in6p->in6p_laddr; 524 ip6->ip6_dst = in6p->in6p_faddr; 525 ip6->ip6_flow = in6p->in6p_flowinfo & IPV6_FLOWINFO_MASK; 526 if (ip6_auto_flowlabel) { 527 ip6->ip6_flow &= ~IPV6_FLOWLABEL_MASK; 528 ip6->ip6_flow |= 529 (htonl(ip6_randomflowlabel()) & IPV6_FLOWLABEL_MASK); 530 } 531 ip6->ip6_vfc &= ~IPV6_VERSION_MASK; 532 ip6->ip6_vfc |= IPV6_VERSION; 533 534 /* 535 * Compute the pseudo-header portion of the checksum 536 * now. We incrementally add in the TCP option and 537 * payload lengths later, and then compute the TCP 538 * checksum right before the packet is sent off onto 539 * the wire. 540 */ 541 n->th_sum = in6_cksum_phdr(&in6p->in6p_laddr, 542 &in6p->in6p_faddr, htonl(sizeof(struct tcphdr)), 543 htonl(IPPROTO_TCP)); 544 break; 545 } 546 #endif 547 } 548 if (inp) { 549 n->th_sport = inp->inp_lport; 550 n->th_dport = inp->inp_fport; 551 } 552 #ifdef INET6 553 else if (in6p) { 554 n->th_sport = in6p->in6p_lport; 555 n->th_dport = in6p->in6p_fport; 556 } 557 #endif 558 n->th_seq = 0; 559 n->th_ack = 0; 560 n->th_x2 = 0; 561 n->th_off = 5; 562 n->th_flags = 0; 563 n->th_win = 0; 564 n->th_urp = 0; 565 return (m); 566 } 567 568 /* 569 * Send a single message to the TCP at address specified by 570 * the given TCP/IP header. If m == 0, then we make a copy 571 * of the tcpiphdr at ti and send directly to the addressed host. 572 * This is used to force keep alive messages out using the TCP 573 * template for a connection tp->t_template. If flags are given 574 * then we send a message back to the TCP which originated the 575 * segment ti, and discard the mbuf containing it and any other 576 * attached mbufs. 577 * 578 * In any case the ack and sequence number of the transmitted 579 * segment are as specified by the parameters. 580 */ 581 int 582 tcp_respond(struct tcpcb *tp, struct mbuf *template, struct mbuf *m, 583 struct tcphdr *th0, tcp_seq ack, tcp_seq seq, int flags) 584 { 585 struct route *ro; 586 int error, tlen, win = 0; 587 int hlen; 588 struct ip *ip; 589 #ifdef INET6 590 struct ip6_hdr *ip6; 591 #endif 592 int family; /* family on packet, not inpcb/in6pcb! */ 593 struct tcphdr *th; 594 struct socket *so; 595 596 if (tp != NULL && (flags & TH_RST) == 0) { 597 #ifdef DIAGNOSTIC 598 if (tp->t_inpcb && tp->t_in6pcb) 599 panic("tcp_respond: both t_inpcb and t_in6pcb are set"); 600 #endif 601 #ifdef INET 602 if (tp->t_inpcb) 603 win = sbspace(&tp->t_inpcb->inp_socket->so_rcv); 604 #endif 605 #ifdef INET6 606 if (tp->t_in6pcb) 607 win = sbspace(&tp->t_in6pcb->in6p_socket->so_rcv); 608 #endif 609 } 610 611 th = NULL; /* Quell uninitialized warning */ 612 ip = NULL; 613 #ifdef INET6 614 ip6 = NULL; 615 #endif 616 if (m == 0) { 617 if (!template) 618 return EINVAL; 619 620 /* get family information from template */ 621 switch (mtod(template, struct ip *)->ip_v) { 622 case 4: 623 family = AF_INET; 624 hlen = sizeof(struct ip); 625 break; 626 #ifdef INET6 627 case 6: 628 family = AF_INET6; 629 hlen = sizeof(struct ip6_hdr); 630 break; 631 #endif 632 default: 633 return EAFNOSUPPORT; 634 } 635 636 MGETHDR(m, M_DONTWAIT, MT_HEADER); 637 if (m) { 638 MCLAIM(m, &tcp_tx_mowner); 639 MCLGET(m, M_DONTWAIT); 640 if ((m->m_flags & M_EXT) == 0) { 641 m_free(m); 642 m = NULL; 643 } 644 } 645 if (m == NULL) 646 return (ENOBUFS); 647 648 if (tcp_compat_42) 649 tlen = 1; 650 else 651 tlen = 0; 652 653 m->m_data += max_linkhdr; 654 bcopy(mtod(template, caddr_t), mtod(m, caddr_t), 655 template->m_len); 656 switch (family) { 657 case AF_INET: 658 ip = mtod(m, struct ip *); 659 th = (struct tcphdr *)(ip + 1); 660 break; 661 #ifdef INET6 662 case AF_INET6: 663 ip6 = mtod(m, struct ip6_hdr *); 664 th = (struct tcphdr *)(ip6 + 1); 665 break; 666 #endif 667 #if 0 668 default: 669 /* noone will visit here */ 670 m_freem(m); 671 return EAFNOSUPPORT; 672 #endif 673 } 674 flags = TH_ACK; 675 } else { 676 677 if ((m->m_flags & M_PKTHDR) == 0) { 678 #if 0 679 printf("non PKTHDR to tcp_respond\n"); 680 #endif 681 m_freem(m); 682 return EINVAL; 683 } 684 #ifdef DIAGNOSTIC 685 if (!th0) 686 panic("th0 == NULL in tcp_respond"); 687 #endif 688 689 /* get family information from m */ 690 switch (mtod(m, struct ip *)->ip_v) { 691 case 4: 692 family = AF_INET; 693 hlen = sizeof(struct ip); 694 ip = mtod(m, struct ip *); 695 break; 696 #ifdef INET6 697 case 6: 698 family = AF_INET6; 699 hlen = sizeof(struct ip6_hdr); 700 ip6 = mtod(m, struct ip6_hdr *); 701 break; 702 #endif 703 default: 704 m_freem(m); 705 return EAFNOSUPPORT; 706 } 707 /* clear h/w csum flags inherited from rx packet */ 708 m->m_pkthdr.csum_flags = 0; 709 710 if ((flags & TH_SYN) == 0 || sizeof(*th0) > (th0->th_off << 2)) 711 tlen = sizeof(*th0); 712 else 713 tlen = th0->th_off << 2; 714 715 if (m->m_len > hlen + tlen && (m->m_flags & M_EXT) == 0 && 716 mtod(m, caddr_t) + hlen == (caddr_t)th0) { 717 m->m_len = hlen + tlen; 718 m_freem(m->m_next); 719 m->m_next = NULL; 720 } else { 721 struct mbuf *n; 722 723 #ifdef DIAGNOSTIC 724 if (max_linkhdr + hlen + tlen > MCLBYTES) { 725 m_freem(m); 726 return EMSGSIZE; 727 } 728 #endif 729 MGETHDR(n, M_DONTWAIT, MT_HEADER); 730 if (n && max_linkhdr + hlen + tlen > MHLEN) { 731 MCLGET(n, M_DONTWAIT); 732 if ((n->m_flags & M_EXT) == 0) { 733 m_freem(n); 734 n = NULL; 735 } 736 } 737 if (!n) { 738 m_freem(m); 739 return ENOBUFS; 740 } 741 742 MCLAIM(n, &tcp_tx_mowner); 743 n->m_data += max_linkhdr; 744 n->m_len = hlen + tlen; 745 m_copyback(n, 0, hlen, mtod(m, caddr_t)); 746 m_copyback(n, hlen, tlen, (caddr_t)th0); 747 748 m_freem(m); 749 m = n; 750 n = NULL; 751 } 752 753 #define xchg(a,b,type) { type t; t=a; a=b; b=t; } 754 switch (family) { 755 case AF_INET: 756 ip = mtod(m, struct ip *); 757 th = (struct tcphdr *)(ip + 1); 758 ip->ip_p = IPPROTO_TCP; 759 xchg(ip->ip_dst, ip->ip_src, struct in_addr); 760 ip->ip_p = IPPROTO_TCP; 761 break; 762 #ifdef INET6 763 case AF_INET6: 764 ip6 = mtod(m, struct ip6_hdr *); 765 th = (struct tcphdr *)(ip6 + 1); 766 ip6->ip6_nxt = IPPROTO_TCP; 767 xchg(ip6->ip6_dst, ip6->ip6_src, struct in6_addr); 768 ip6->ip6_nxt = IPPROTO_TCP; 769 break; 770 #endif 771 #if 0 772 default: 773 /* noone will visit here */ 774 m_freem(m); 775 return EAFNOSUPPORT; 776 #endif 777 } 778 xchg(th->th_dport, th->th_sport, u_int16_t); 779 #undef xchg 780 tlen = 0; /*be friendly with the following code*/ 781 } 782 th->th_seq = htonl(seq); 783 th->th_ack = htonl(ack); 784 th->th_x2 = 0; 785 if ((flags & TH_SYN) == 0) { 786 if (tp) 787 win >>= tp->rcv_scale; 788 if (win > TCP_MAXWIN) 789 win = TCP_MAXWIN; 790 th->th_win = htons((u_int16_t)win); 791 th->th_off = sizeof (struct tcphdr) >> 2; 792 tlen += sizeof(*th); 793 } else 794 tlen += th->th_off << 2; 795 m->m_len = hlen + tlen; 796 m->m_pkthdr.len = hlen + tlen; 797 m->m_pkthdr.rcvif = (struct ifnet *) 0; 798 th->th_flags = flags; 799 th->th_urp = 0; 800 801 switch (family) { 802 #ifdef INET 803 case AF_INET: 804 { 805 struct ipovly *ipov = (struct ipovly *)ip; 806 bzero(ipov->ih_x1, sizeof ipov->ih_x1); 807 ipov->ih_len = htons((u_int16_t)tlen); 808 809 th->th_sum = 0; 810 th->th_sum = in_cksum(m, hlen + tlen); 811 ip->ip_len = htons(hlen + tlen); 812 ip->ip_ttl = ip_defttl; 813 break; 814 } 815 #endif 816 #ifdef INET6 817 case AF_INET6: 818 { 819 th->th_sum = 0; 820 th->th_sum = in6_cksum(m, IPPROTO_TCP, sizeof(struct ip6_hdr), 821 tlen); 822 ip6->ip6_plen = htons(tlen); 823 if (tp && tp->t_in6pcb) { 824 struct ifnet *oifp; 825 ro = (struct route *)&tp->t_in6pcb->in6p_route; 826 oifp = ro->ro_rt ? ro->ro_rt->rt_ifp : NULL; 827 ip6->ip6_hlim = in6_selecthlim(tp->t_in6pcb, oifp); 828 } else 829 ip6->ip6_hlim = ip6_defhlim; 830 ip6->ip6_flow &= ~IPV6_FLOWINFO_MASK; 831 if (ip6_auto_flowlabel) { 832 ip6->ip6_flow |= 833 (htonl(ip6_randomflowlabel()) & IPV6_FLOWLABEL_MASK); 834 } 835 break; 836 } 837 #endif 838 } 839 840 if (tp && tp->t_inpcb) 841 so = tp->t_inpcb->inp_socket; 842 #ifdef INET6 843 else if (tp && tp->t_in6pcb) 844 so = tp->t_in6pcb->in6p_socket; 845 #endif 846 else 847 so = NULL; 848 849 if (tp != NULL && tp->t_inpcb != NULL) { 850 ro = &tp->t_inpcb->inp_route; 851 #ifdef DIAGNOSTIC 852 if (family != AF_INET) 853 panic("tcp_respond: address family mismatch"); 854 if (!in_hosteq(ip->ip_dst, tp->t_inpcb->inp_faddr)) { 855 panic("tcp_respond: ip_dst %x != inp_faddr %x", 856 ntohl(ip->ip_dst.s_addr), 857 ntohl(tp->t_inpcb->inp_faddr.s_addr)); 858 } 859 #endif 860 } 861 #ifdef INET6 862 else if (tp != NULL && tp->t_in6pcb != NULL) { 863 ro = (struct route *)&tp->t_in6pcb->in6p_route; 864 #ifdef DIAGNOSTIC 865 if (family == AF_INET) { 866 if (!IN6_IS_ADDR_V4MAPPED(&tp->t_in6pcb->in6p_faddr)) 867 panic("tcp_respond: not mapped addr"); 868 if (bcmp(&ip->ip_dst, 869 &tp->t_in6pcb->in6p_faddr.s6_addr32[3], 870 sizeof(ip->ip_dst)) != 0) { 871 panic("tcp_respond: ip_dst != in6p_faddr"); 872 } 873 } else if (family == AF_INET6) { 874 if (!IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst, 875 &tp->t_in6pcb->in6p_faddr)) 876 panic("tcp_respond: ip6_dst != in6p_faddr"); 877 } else 878 panic("tcp_respond: address family mismatch"); 879 #endif 880 } 881 #endif 882 else 883 ro = NULL; 884 885 switch (family) { 886 #ifdef INET 887 case AF_INET: 888 error = ip_output(m, NULL, ro, 889 (tp && tp->t_mtudisc ? IP_MTUDISC : 0), 890 (struct ip_moptions *)0, so); 891 break; 892 #endif 893 #ifdef INET6 894 case AF_INET6: 895 error = ip6_output(m, NULL, (struct route_in6 *)ro, 0, 896 (struct ip6_moptions *)0, so, NULL); 897 break; 898 #endif 899 default: 900 error = EAFNOSUPPORT; 901 break; 902 } 903 904 return (error); 905 } 906 907 /* 908 * Template TCPCB. Rather than zeroing a new TCPCB and initializing 909 * a bunch of members individually, we maintain this template for the 910 * static and mostly-static components of the TCPCB, and copy it into 911 * the new TCPCB instead. 912 */ 913 static struct tcpcb tcpcb_template = { 914 /* 915 * If TCP_NTIMERS ever changes, we'll need to update this 916 * initializer. 917 */ 918 .t_timer = { 919 CALLOUT_INITIALIZER, 920 CALLOUT_INITIALIZER, 921 CALLOUT_INITIALIZER, 922 CALLOUT_INITIALIZER, 923 }, 924 .t_delack_ch = CALLOUT_INITIALIZER, 925 926 .t_srtt = TCPTV_SRTTBASE, 927 .t_rttmin = TCPTV_MIN, 928 929 .snd_cwnd = TCP_MAXWIN << TCP_MAX_WINSHIFT, 930 .snd_ssthresh = TCP_MAXWIN << TCP_MAX_WINSHIFT, 931 .snd_numholes = 0, 932 933 .t_partialacks = -1, 934 }; 935 936 /* 937 * Updates the TCPCB template whenever a parameter that would affect 938 * the template is changed. 939 */ 940 void 941 tcp_tcpcb_template(void) 942 { 943 struct tcpcb *tp = &tcpcb_template; 944 int flags; 945 946 tp->t_peermss = tcp_mssdflt; 947 tp->t_ourmss = tcp_mssdflt; 948 tp->t_segsz = tcp_mssdflt; 949 950 flags = 0; 951 if (tcp_do_rfc1323 && tcp_do_win_scale) 952 flags |= TF_REQ_SCALE; 953 if (tcp_do_rfc1323 && tcp_do_timestamps) 954 flags |= TF_REQ_TSTMP; 955 tp->t_flags = flags; 956 957 /* 958 * Init srtt to TCPTV_SRTTBASE (0), so we can tell that we have no 959 * rtt estimate. Set rttvar so that srtt + 2 * rttvar gives 960 * reasonable initial retransmit time. 961 */ 962 tp->t_rttvar = tcp_rttdflt * PR_SLOWHZ << (TCP_RTTVAR_SHIFT + 2 - 1); 963 TCPT_RANGESET(tp->t_rxtcur, TCP_REXMTVAL(tp), 964 TCPTV_MIN, TCPTV_REXMTMAX); 965 } 966 967 /* 968 * Create a new TCP control block, making an 969 * empty reassembly queue and hooking it to the argument 970 * protocol control block. 971 */ 972 /* family selects inpcb, or in6pcb */ 973 struct tcpcb * 974 tcp_newtcpcb(int family, void *aux) 975 { 976 struct tcpcb *tp; 977 int i; 978 979 /* XXX Consider using a pool_cache for speed. */ 980 tp = pool_get(&tcpcb_pool, PR_NOWAIT); 981 if (tp == NULL) 982 return (NULL); 983 memcpy(tp, &tcpcb_template, sizeof(*tp)); 984 TAILQ_INIT(&tp->segq); 985 TAILQ_INIT(&tp->timeq); 986 tp->t_family = family; /* may be overridden later on */ 987 TAILQ_INIT(&tp->snd_holes); 988 LIST_INIT(&tp->t_sc); /* XXX can template this */ 989 990 /* Don't sweat this loop; hopefully the compiler will unroll it. */ 991 for (i = 0; i < TCPT_NTIMERS; i++) 992 TCP_TIMER_INIT(tp, i); 993 994 switch (family) { 995 case AF_INET: 996 { 997 struct inpcb *inp = (struct inpcb *)aux; 998 999 inp->inp_ip.ip_ttl = ip_defttl; 1000 inp->inp_ppcb = (caddr_t)tp; 1001 1002 tp->t_inpcb = inp; 1003 tp->t_mtudisc = ip_mtudisc; 1004 break; 1005 } 1006 #ifdef INET6 1007 case AF_INET6: 1008 { 1009 struct in6pcb *in6p = (struct in6pcb *)aux; 1010 1011 in6p->in6p_ip6.ip6_hlim = in6_selecthlim(in6p, 1012 in6p->in6p_route.ro_rt ? in6p->in6p_route.ro_rt->rt_ifp 1013 : NULL); 1014 in6p->in6p_ppcb = (caddr_t)tp; 1015 1016 tp->t_in6pcb = in6p; 1017 /* for IPv6, always try to run path MTU discovery */ 1018 tp->t_mtudisc = 1; 1019 break; 1020 } 1021 #endif /* INET6 */ 1022 default: 1023 pool_put(&tcpcb_pool, tp); 1024 return (NULL); 1025 } 1026 1027 /* 1028 * Initialize our timebase. When we send timestamps, we take 1029 * the delta from tcp_now -- this means each connection always 1030 * gets a timebase of 0, which makes it, among other things, 1031 * more difficult to determine how long a system has been up, 1032 * and thus how many TCP sequence increments have occurred. 1033 */ 1034 tp->ts_timebase = tcp_now; 1035 1036 return (tp); 1037 } 1038 1039 /* 1040 * Drop a TCP connection, reporting 1041 * the specified error. If connection is synchronized, 1042 * then send a RST to peer. 1043 */ 1044 struct tcpcb * 1045 tcp_drop(struct tcpcb *tp, int errno) 1046 { 1047 struct socket *so = NULL; 1048 1049 #ifdef DIAGNOSTIC 1050 if (tp->t_inpcb && tp->t_in6pcb) 1051 panic("tcp_drop: both t_inpcb and t_in6pcb are set"); 1052 #endif 1053 #ifdef INET 1054 if (tp->t_inpcb) 1055 so = tp->t_inpcb->inp_socket; 1056 #endif 1057 #ifdef INET6 1058 if (tp->t_in6pcb) 1059 so = tp->t_in6pcb->in6p_socket; 1060 #endif 1061 if (!so) 1062 return NULL; 1063 1064 if (TCPS_HAVERCVDSYN(tp->t_state)) { 1065 tp->t_state = TCPS_CLOSED; 1066 (void) tcp_output(tp); 1067 tcpstat.tcps_drops++; 1068 } else 1069 tcpstat.tcps_conndrops++; 1070 if (errno == ETIMEDOUT && tp->t_softerror) 1071 errno = tp->t_softerror; 1072 so->so_error = errno; 1073 return (tcp_close(tp)); 1074 } 1075 1076 /* 1077 * Return whether this tcpcb is marked as dead, indicating 1078 * to the calling timer function that no further action should 1079 * be taken, as we are about to release this tcpcb. The release 1080 * of the storage will be done if this is the last timer running. 1081 * 1082 * This should be called from the callout handler function after 1083 * callout_ack() is done, so that the number of invoking timer 1084 * functions is 0. 1085 */ 1086 int 1087 tcp_isdead(struct tcpcb *tp) 1088 { 1089 int dead = (tp->t_flags & TF_DEAD); 1090 1091 if (__predict_false(dead)) { 1092 if (tcp_timers_invoking(tp) > 0) 1093 /* not quite there yet -- count separately? */ 1094 return dead; 1095 tcpstat.tcps_delayed_free++; 1096 pool_put(&tcpcb_pool, tp); 1097 } 1098 return dead; 1099 } 1100 1101 /* 1102 * Close a TCP control block: 1103 * discard all space held by the tcp 1104 * discard internet protocol block 1105 * wake up any sleepers 1106 */ 1107 struct tcpcb * 1108 tcp_close(struct tcpcb *tp) 1109 { 1110 struct inpcb *inp; 1111 #ifdef INET6 1112 struct in6pcb *in6p; 1113 #endif 1114 struct socket *so; 1115 #ifdef RTV_RTT 1116 struct rtentry *rt; 1117 #endif 1118 struct route *ro; 1119 1120 inp = tp->t_inpcb; 1121 #ifdef INET6 1122 in6p = tp->t_in6pcb; 1123 #endif 1124 so = NULL; 1125 ro = NULL; 1126 if (inp) { 1127 so = inp->inp_socket; 1128 ro = &inp->inp_route; 1129 } 1130 #ifdef INET6 1131 else if (in6p) { 1132 so = in6p->in6p_socket; 1133 ro = (struct route *)&in6p->in6p_route; 1134 } 1135 #endif 1136 1137 #ifdef RTV_RTT 1138 /* 1139 * If we sent enough data to get some meaningful characteristics, 1140 * save them in the routing entry. 'Enough' is arbitrarily 1141 * defined as the sendpipesize (default 4K) * 16. This would 1142 * give us 16 rtt samples assuming we only get one sample per 1143 * window (the usual case on a long haul net). 16 samples is 1144 * enough for the srtt filter to converge to within 5% of the correct 1145 * value; fewer samples and we could save a very bogus rtt. 1146 * 1147 * Don't update the default route's characteristics and don't 1148 * update anything that the user "locked". 1149 */ 1150 if (SEQ_LT(tp->iss + so->so_snd.sb_hiwat * 16, tp->snd_max) && 1151 ro && (rt = ro->ro_rt) && 1152 !in_nullhost(satosin(rt_key(rt))->sin_addr)) { 1153 u_long i = 0; 1154 1155 if ((rt->rt_rmx.rmx_locks & RTV_RTT) == 0) { 1156 i = tp->t_srtt * 1157 ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTT_SHIFT + 2)); 1158 if (rt->rt_rmx.rmx_rtt && i) 1159 /* 1160 * filter this update to half the old & half 1161 * the new values, converting scale. 1162 * See route.h and tcp_var.h for a 1163 * description of the scaling constants. 1164 */ 1165 rt->rt_rmx.rmx_rtt = 1166 (rt->rt_rmx.rmx_rtt + i) / 2; 1167 else 1168 rt->rt_rmx.rmx_rtt = i; 1169 } 1170 if ((rt->rt_rmx.rmx_locks & RTV_RTTVAR) == 0) { 1171 i = tp->t_rttvar * 1172 ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTTVAR_SHIFT + 2)); 1173 if (rt->rt_rmx.rmx_rttvar && i) 1174 rt->rt_rmx.rmx_rttvar = 1175 (rt->rt_rmx.rmx_rttvar + i) / 2; 1176 else 1177 rt->rt_rmx.rmx_rttvar = i; 1178 } 1179 /* 1180 * update the pipelimit (ssthresh) if it has been updated 1181 * already or if a pipesize was specified & the threshhold 1182 * got below half the pipesize. I.e., wait for bad news 1183 * before we start updating, then update on both good 1184 * and bad news. 1185 */ 1186 if (((rt->rt_rmx.rmx_locks & RTV_SSTHRESH) == 0 && 1187 (i = tp->snd_ssthresh) && rt->rt_rmx.rmx_ssthresh) || 1188 i < (rt->rt_rmx.rmx_sendpipe / 2)) { 1189 /* 1190 * convert the limit from user data bytes to 1191 * packets then to packet data bytes. 1192 */ 1193 i = (i + tp->t_segsz / 2) / tp->t_segsz; 1194 if (i < 2) 1195 i = 2; 1196 i *= (u_long)(tp->t_segsz + sizeof (struct tcpiphdr)); 1197 if (rt->rt_rmx.rmx_ssthresh) 1198 rt->rt_rmx.rmx_ssthresh = 1199 (rt->rt_rmx.rmx_ssthresh + i) / 2; 1200 else 1201 rt->rt_rmx.rmx_ssthresh = i; 1202 } 1203 } 1204 #endif /* RTV_RTT */ 1205 /* free the reassembly queue, if any */ 1206 TCP_REASS_LOCK(tp); 1207 (void) tcp_freeq(tp); 1208 TCP_REASS_UNLOCK(tp); 1209 1210 /* free the SACK holes list. */ 1211 tcp_free_sackholes(tp); 1212 1213 tcp_canceltimers(tp); 1214 TCP_CLEAR_DELACK(tp); 1215 syn_cache_cleanup(tp); 1216 1217 if (tp->t_template) { 1218 m_free(tp->t_template); 1219 tp->t_template = NULL; 1220 } 1221 if (tcp_timers_invoking(tp)) 1222 tp->t_flags |= TF_DEAD; 1223 else 1224 pool_put(&tcpcb_pool, tp); 1225 1226 if (inp) { 1227 inp->inp_ppcb = 0; 1228 soisdisconnected(so); 1229 in_pcbdetach(inp); 1230 } 1231 #ifdef INET6 1232 else if (in6p) { 1233 in6p->in6p_ppcb = 0; 1234 soisdisconnected(so); 1235 in6_pcbdetach(in6p); 1236 } 1237 #endif 1238 tcpstat.tcps_closed++; 1239 return ((struct tcpcb *)0); 1240 } 1241 1242 int 1243 tcp_freeq(tp) 1244 struct tcpcb *tp; 1245 { 1246 struct ipqent *qe; 1247 int rv = 0; 1248 #ifdef TCPREASS_DEBUG 1249 int i = 0; 1250 #endif 1251 1252 TCP_REASS_LOCK_CHECK(tp); 1253 1254 while ((qe = TAILQ_FIRST(&tp->segq)) != NULL) { 1255 #ifdef TCPREASS_DEBUG 1256 printf("tcp_freeq[%p,%d]: %u:%u(%u) 0x%02x\n", 1257 tp, i++, qe->ipqe_seq, qe->ipqe_seq + qe->ipqe_len, 1258 qe->ipqe_len, qe->ipqe_flags & (TH_SYN|TH_FIN|TH_RST)); 1259 #endif 1260 TAILQ_REMOVE(&tp->segq, qe, ipqe_q); 1261 TAILQ_REMOVE(&tp->timeq, qe, ipqe_timeq); 1262 m_freem(qe->ipqe_m); 1263 tcpipqent_free(qe); 1264 rv = 1; 1265 } 1266 tp->t_segqlen = 0; 1267 KASSERT(TAILQ_EMPTY(&tp->timeq)); 1268 return (rv); 1269 } 1270 1271 /* 1272 * Protocol drain routine. Called when memory is in short supply. 1273 */ 1274 void 1275 tcp_drain(void) 1276 { 1277 struct inpcb_hdr *inph; 1278 struct tcpcb *tp; 1279 1280 /* 1281 * Free the sequence queue of all TCP connections. 1282 */ 1283 CIRCLEQ_FOREACH(inph, &tcbtable.inpt_queue, inph_queue) { 1284 switch (inph->inph_af) { 1285 case AF_INET: 1286 tp = intotcpcb((struct inpcb *)inph); 1287 break; 1288 #ifdef INET6 1289 case AF_INET6: 1290 tp = in6totcpcb((struct in6pcb *)inph); 1291 break; 1292 #endif 1293 default: 1294 tp = NULL; 1295 break; 1296 } 1297 if (tp != NULL) { 1298 /* 1299 * We may be called from a device's interrupt 1300 * context. If the tcpcb is already busy, 1301 * just bail out now. 1302 */ 1303 if (tcp_reass_lock_try(tp) == 0) 1304 continue; 1305 if (tcp_freeq(tp)) 1306 tcpstat.tcps_connsdrained++; 1307 TCP_REASS_UNLOCK(tp); 1308 } 1309 } 1310 } 1311 1312 /* 1313 * Notify a tcp user of an asynchronous error; 1314 * store error as soft error, but wake up user 1315 * (for now, won't do anything until can select for soft error). 1316 */ 1317 void 1318 tcp_notify(struct inpcb *inp, int error) 1319 { 1320 struct tcpcb *tp = (struct tcpcb *)inp->inp_ppcb; 1321 struct socket *so = inp->inp_socket; 1322 1323 /* 1324 * Ignore some errors if we are hooked up. 1325 * If connection hasn't completed, has retransmitted several times, 1326 * and receives a second error, give up now. This is better 1327 * than waiting a long time to establish a connection that 1328 * can never complete. 1329 */ 1330 if (tp->t_state == TCPS_ESTABLISHED && 1331 (error == EHOSTUNREACH || error == ENETUNREACH || 1332 error == EHOSTDOWN)) { 1333 return; 1334 } else if (TCPS_HAVEESTABLISHED(tp->t_state) == 0 && 1335 tp->t_rxtshift > 3 && tp->t_softerror) 1336 so->so_error = error; 1337 else 1338 tp->t_softerror = error; 1339 wakeup((caddr_t) &so->so_timeo); 1340 sorwakeup(so); 1341 sowwakeup(so); 1342 } 1343 1344 #ifdef INET6 1345 void 1346 tcp6_notify(struct in6pcb *in6p, int error) 1347 { 1348 struct tcpcb *tp = (struct tcpcb *)in6p->in6p_ppcb; 1349 struct socket *so = in6p->in6p_socket; 1350 1351 /* 1352 * Ignore some errors if we are hooked up. 1353 * If connection hasn't completed, has retransmitted several times, 1354 * and receives a second error, give up now. This is better 1355 * than waiting a long time to establish a connection that 1356 * can never complete. 1357 */ 1358 if (tp->t_state == TCPS_ESTABLISHED && 1359 (error == EHOSTUNREACH || error == ENETUNREACH || 1360 error == EHOSTDOWN)) { 1361 return; 1362 } else if (TCPS_HAVEESTABLISHED(tp->t_state) == 0 && 1363 tp->t_rxtshift > 3 && tp->t_softerror) 1364 so->so_error = error; 1365 else 1366 tp->t_softerror = error; 1367 wakeup((caddr_t) &so->so_timeo); 1368 sorwakeup(so); 1369 sowwakeup(so); 1370 } 1371 #endif 1372 1373 #ifdef INET6 1374 void 1375 tcp6_ctlinput(int cmd, struct sockaddr *sa, void *d) 1376 { 1377 struct tcphdr th; 1378 void (*notify)(struct in6pcb *, int) = tcp6_notify; 1379 int nmatch; 1380 struct ip6_hdr *ip6; 1381 const struct sockaddr_in6 *sa6_src = NULL; 1382 struct sockaddr_in6 *sa6 = (struct sockaddr_in6 *)sa; 1383 struct mbuf *m; 1384 int off; 1385 1386 if (sa->sa_family != AF_INET6 || 1387 sa->sa_len != sizeof(struct sockaddr_in6)) 1388 return; 1389 if ((unsigned)cmd >= PRC_NCMDS) 1390 return; 1391 else if (cmd == PRC_QUENCH) { 1392 /* 1393 * Don't honor ICMP Source Quench messages meant for 1394 * TCP connections. 1395 */ 1396 return; 1397 } else if (PRC_IS_REDIRECT(cmd)) 1398 notify = in6_rtchange, d = NULL; 1399 else if (cmd == PRC_MSGSIZE) 1400 ; /* special code is present, see below */ 1401 else if (cmd == PRC_HOSTDEAD) 1402 d = NULL; 1403 else if (inet6ctlerrmap[cmd] == 0) 1404 return; 1405 1406 /* if the parameter is from icmp6, decode it. */ 1407 if (d != NULL) { 1408 struct ip6ctlparam *ip6cp = (struct ip6ctlparam *)d; 1409 m = ip6cp->ip6c_m; 1410 ip6 = ip6cp->ip6c_ip6; 1411 off = ip6cp->ip6c_off; 1412 sa6_src = ip6cp->ip6c_src; 1413 } else { 1414 m = NULL; 1415 ip6 = NULL; 1416 sa6_src = &sa6_any; 1417 off = 0; 1418 } 1419 1420 if (ip6) { 1421 /* 1422 * XXX: We assume that when ip6 is non NULL, 1423 * M and OFF are valid. 1424 */ 1425 1426 /* check if we can safely examine src and dst ports */ 1427 if (m->m_pkthdr.len < off + sizeof(th)) { 1428 if (cmd == PRC_MSGSIZE) 1429 icmp6_mtudisc_update((struct ip6ctlparam *)d, 0); 1430 return; 1431 } 1432 1433 bzero(&th, sizeof(th)); 1434 m_copydata(m, off, sizeof(th), (caddr_t)&th); 1435 1436 if (cmd == PRC_MSGSIZE) { 1437 int valid = 0; 1438 1439 /* 1440 * Check to see if we have a valid TCP connection 1441 * corresponding to the address in the ICMPv6 message 1442 * payload. 1443 */ 1444 if (in6_pcblookup_connect(&tcbtable, &sa6->sin6_addr, 1445 th.th_dport, (const struct in6_addr *)&sa6_src->sin6_addr, 1446 th.th_sport, 0)) 1447 valid++; 1448 1449 /* 1450 * Depending on the value of "valid" and routing table 1451 * size (mtudisc_{hi,lo}wat), we will: 1452 * - recalcurate the new MTU and create the 1453 * corresponding routing entry, or 1454 * - ignore the MTU change notification. 1455 */ 1456 icmp6_mtudisc_update((struct ip6ctlparam *)d, valid); 1457 1458 /* 1459 * no need to call in6_pcbnotify, it should have been 1460 * called via callback if necessary 1461 */ 1462 return; 1463 } 1464 1465 nmatch = in6_pcbnotify(&tcbtable, sa, th.th_dport, 1466 (const struct sockaddr *)sa6_src, th.th_sport, cmd, NULL, notify); 1467 if (nmatch == 0 && syn_cache_count && 1468 (inet6ctlerrmap[cmd] == EHOSTUNREACH || 1469 inet6ctlerrmap[cmd] == ENETUNREACH || 1470 inet6ctlerrmap[cmd] == EHOSTDOWN)) 1471 syn_cache_unreach((const struct sockaddr *)sa6_src, 1472 sa, &th); 1473 } else { 1474 (void) in6_pcbnotify(&tcbtable, sa, 0, 1475 (const struct sockaddr *)sa6_src, 0, cmd, NULL, notify); 1476 } 1477 } 1478 #endif 1479 1480 #ifdef INET 1481 /* assumes that ip header and tcp header are contiguous on mbuf */ 1482 void * 1483 tcp_ctlinput(int cmd, struct sockaddr *sa, void *v) 1484 { 1485 struct ip *ip = v; 1486 struct tcphdr *th; 1487 struct icmp *icp; 1488 extern const int inetctlerrmap[]; 1489 void (*notify)(struct inpcb *, int) = tcp_notify; 1490 int errno; 1491 int nmatch; 1492 struct tcpcb *tp; 1493 u_int mtu; 1494 tcp_seq seq; 1495 struct inpcb *inp; 1496 #ifdef INET6 1497 struct in6pcb *in6p; 1498 struct in6_addr src6, dst6; 1499 #endif 1500 1501 if (sa->sa_family != AF_INET || 1502 sa->sa_len != sizeof(struct sockaddr_in)) 1503 return NULL; 1504 if ((unsigned)cmd >= PRC_NCMDS) 1505 return NULL; 1506 errno = inetctlerrmap[cmd]; 1507 if (cmd == PRC_QUENCH) 1508 /* 1509 * Don't honor ICMP Source Quench messages meant for 1510 * TCP connections. 1511 */ 1512 return NULL; 1513 else if (PRC_IS_REDIRECT(cmd)) 1514 notify = in_rtchange, ip = 0; 1515 else if (cmd == PRC_MSGSIZE && ip && ip->ip_v == 4) { 1516 /* 1517 * Check to see if we have a valid TCP connection 1518 * corresponding to the address in the ICMP message 1519 * payload. 1520 * 1521 * Boundary check is made in icmp_input(), with ICMP_ADVLENMIN. 1522 */ 1523 th = (struct tcphdr *)((caddr_t)ip + (ip->ip_hl << 2)); 1524 #ifdef INET6 1525 memset(&src6, 0, sizeof(src6)); 1526 memset(&dst6, 0, sizeof(dst6)); 1527 src6.s6_addr16[5] = dst6.s6_addr16[5] = 0xffff; 1528 memcpy(&src6.s6_addr32[3], &ip->ip_src, sizeof(struct in_addr)); 1529 memcpy(&dst6.s6_addr32[3], &ip->ip_dst, sizeof(struct in_addr)); 1530 #endif 1531 if ((inp = in_pcblookup_connect(&tcbtable, ip->ip_dst, 1532 th->th_dport, ip->ip_src, th->th_sport)) != NULL) 1533 #ifdef INET6 1534 in6p = NULL; 1535 #else 1536 ; 1537 #endif 1538 #ifdef INET6 1539 else if ((in6p = in6_pcblookup_connect(&tcbtable, &dst6, 1540 th->th_dport, &src6, th->th_sport, 0)) != NULL) 1541 ; 1542 #endif 1543 else 1544 return NULL; 1545 1546 /* 1547 * Now that we've validated that we are actually communicating 1548 * with the host indicated in the ICMP message, locate the 1549 * ICMP header, recalculate the new MTU, and create the 1550 * corresponding routing entry. 1551 */ 1552 icp = (struct icmp *)((caddr_t)ip - 1553 offsetof(struct icmp, icmp_ip)); 1554 if (inp) { 1555 if ((tp = intotcpcb(inp)) == NULL) 1556 return NULL; 1557 } 1558 #ifdef INET6 1559 else if (in6p) { 1560 if ((tp = in6totcpcb(in6p)) == NULL) 1561 return NULL; 1562 } 1563 #endif 1564 else 1565 return NULL; 1566 seq = ntohl(th->th_seq); 1567 if (SEQ_LT(seq, tp->snd_una) || SEQ_GT(seq, tp->snd_max)) 1568 return NULL; 1569 /* 1570 * If the ICMP message advertises a Next-Hop MTU 1571 * equal or larger than the maximum packet size we have 1572 * ever sent, drop the message. 1573 */ 1574 mtu = (u_int)ntohs(icp->icmp_nextmtu); 1575 if (mtu >= tp->t_pmtud_mtu_sent) 1576 return NULL; 1577 if (mtu >= tcp_hdrsz(tp) + tp->t_pmtud_mss_acked) { 1578 /* 1579 * Calculate new MTU, and create corresponding 1580 * route (traditional PMTUD). 1581 */ 1582 tp->t_flags &= ~TF_PMTUD_PEND; 1583 icmp_mtudisc(icp, ip->ip_dst); 1584 } else { 1585 /* 1586 * Record the information got in the ICMP 1587 * message; act on it later. 1588 * If we had already recorded an ICMP message, 1589 * replace the old one only if the new message 1590 * refers to an older TCP segment 1591 */ 1592 if (tp->t_flags & TF_PMTUD_PEND) { 1593 if (SEQ_LT(tp->t_pmtud_th_seq, seq)) 1594 return NULL; 1595 } else 1596 tp->t_flags |= TF_PMTUD_PEND; 1597 tp->t_pmtud_th_seq = seq; 1598 tp->t_pmtud_nextmtu = icp->icmp_nextmtu; 1599 tp->t_pmtud_ip_len = icp->icmp_ip.ip_len; 1600 tp->t_pmtud_ip_hl = icp->icmp_ip.ip_hl; 1601 } 1602 return NULL; 1603 } else if (cmd == PRC_HOSTDEAD) 1604 ip = 0; 1605 else if (errno == 0) 1606 return NULL; 1607 if (ip && ip->ip_v == 4 && sa->sa_family == AF_INET) { 1608 th = (struct tcphdr *)((caddr_t)ip + (ip->ip_hl << 2)); 1609 nmatch = in_pcbnotify(&tcbtable, satosin(sa)->sin_addr, 1610 th->th_dport, ip->ip_src, th->th_sport, errno, notify); 1611 if (nmatch == 0 && syn_cache_count && 1612 (inetctlerrmap[cmd] == EHOSTUNREACH || 1613 inetctlerrmap[cmd] == ENETUNREACH || 1614 inetctlerrmap[cmd] == EHOSTDOWN)) { 1615 struct sockaddr_in sin; 1616 bzero(&sin, sizeof(sin)); 1617 sin.sin_len = sizeof(sin); 1618 sin.sin_family = AF_INET; 1619 sin.sin_port = th->th_sport; 1620 sin.sin_addr = ip->ip_src; 1621 syn_cache_unreach((struct sockaddr *)&sin, sa, th); 1622 } 1623 1624 /* XXX mapped address case */ 1625 } else 1626 in_pcbnotifyall(&tcbtable, satosin(sa)->sin_addr, errno, 1627 notify); 1628 return NULL; 1629 } 1630 1631 /* 1632 * When a source quench is received, we are being notified of congestion. 1633 * Close the congestion window down to the Loss Window (one segment). 1634 * We will gradually open it again as we proceed. 1635 */ 1636 void 1637 tcp_quench(struct inpcb *inp, int errno) 1638 { 1639 struct tcpcb *tp = intotcpcb(inp); 1640 1641 if (tp) 1642 tp->snd_cwnd = tp->t_segsz; 1643 } 1644 #endif 1645 1646 #ifdef INET6 1647 void 1648 tcp6_quench(struct in6pcb *in6p, int errno) 1649 { 1650 struct tcpcb *tp = in6totcpcb(in6p); 1651 1652 if (tp) 1653 tp->snd_cwnd = tp->t_segsz; 1654 } 1655 #endif 1656 1657 #ifdef INET 1658 /* 1659 * Path MTU Discovery handlers. 1660 */ 1661 void 1662 tcp_mtudisc_callback(struct in_addr faddr) 1663 { 1664 #ifdef INET6 1665 struct in6_addr in6; 1666 #endif 1667 1668 in_pcbnotifyall(&tcbtable, faddr, EMSGSIZE, tcp_mtudisc); 1669 #ifdef INET6 1670 memset(&in6, 0, sizeof(in6)); 1671 in6.s6_addr16[5] = 0xffff; 1672 memcpy(&in6.s6_addr32[3], &faddr, sizeof(struct in_addr)); 1673 tcp6_mtudisc_callback(&in6); 1674 #endif 1675 } 1676 1677 /* 1678 * On receipt of path MTU corrections, flush old route and replace it 1679 * with the new one. Retransmit all unacknowledged packets, to ensure 1680 * that all packets will be received. 1681 */ 1682 void 1683 tcp_mtudisc(struct inpcb *inp, int errno) 1684 { 1685 struct tcpcb *tp = intotcpcb(inp); 1686 struct rtentry *rt = in_pcbrtentry(inp); 1687 1688 if (tp != 0) { 1689 if (rt != 0) { 1690 /* 1691 * If this was not a host route, remove and realloc. 1692 */ 1693 if ((rt->rt_flags & RTF_HOST) == 0) { 1694 in_rtchange(inp, errno); 1695 if ((rt = in_pcbrtentry(inp)) == 0) 1696 return; 1697 } 1698 1699 /* 1700 * Slow start out of the error condition. We 1701 * use the MTU because we know it's smaller 1702 * than the previously transmitted segment. 1703 * 1704 * Note: This is more conservative than the 1705 * suggestion in draft-floyd-incr-init-win-03. 1706 */ 1707 if (rt->rt_rmx.rmx_mtu != 0) 1708 tp->snd_cwnd = 1709 TCP_INITIAL_WINDOW(tcp_init_win, 1710 rt->rt_rmx.rmx_mtu); 1711 } 1712 1713 /* 1714 * Resend unacknowledged packets. 1715 */ 1716 tp->snd_nxt = tp->snd_una; 1717 tcp_output(tp); 1718 } 1719 } 1720 #endif 1721 1722 #ifdef INET6 1723 /* 1724 * Path MTU Discovery handlers. 1725 */ 1726 void 1727 tcp6_mtudisc_callback(struct in6_addr *faddr) 1728 { 1729 struct sockaddr_in6 sin6; 1730 1731 bzero(&sin6, sizeof(sin6)); 1732 sin6.sin6_family = AF_INET6; 1733 sin6.sin6_len = sizeof(struct sockaddr_in6); 1734 sin6.sin6_addr = *faddr; 1735 (void) in6_pcbnotify(&tcbtable, (struct sockaddr *)&sin6, 0, 1736 (const struct sockaddr *)&sa6_any, 0, PRC_MSGSIZE, NULL, tcp6_mtudisc); 1737 } 1738 1739 void 1740 tcp6_mtudisc(struct in6pcb *in6p, int errno) 1741 { 1742 struct tcpcb *tp = in6totcpcb(in6p); 1743 struct rtentry *rt = in6_pcbrtentry(in6p); 1744 1745 if (tp != 0) { 1746 if (rt != 0) { 1747 /* 1748 * If this was not a host route, remove and realloc. 1749 */ 1750 if ((rt->rt_flags & RTF_HOST) == 0) { 1751 in6_rtchange(in6p, errno); 1752 if ((rt = in6_pcbrtentry(in6p)) == 0) 1753 return; 1754 } 1755 1756 /* 1757 * Slow start out of the error condition. We 1758 * use the MTU because we know it's smaller 1759 * than the previously transmitted segment. 1760 * 1761 * Note: This is more conservative than the 1762 * suggestion in draft-floyd-incr-init-win-03. 1763 */ 1764 if (rt->rt_rmx.rmx_mtu != 0) 1765 tp->snd_cwnd = 1766 TCP_INITIAL_WINDOW(tcp_init_win, 1767 rt->rt_rmx.rmx_mtu); 1768 } 1769 1770 /* 1771 * Resend unacknowledged packets. 1772 */ 1773 tp->snd_nxt = tp->snd_una; 1774 tcp_output(tp); 1775 } 1776 } 1777 #endif /* INET6 */ 1778 1779 /* 1780 * Compute the MSS to advertise to the peer. Called only during 1781 * the 3-way handshake. If we are the server (peer initiated 1782 * connection), we are called with a pointer to the interface 1783 * on which the SYN packet arrived. If we are the client (we 1784 * initiated connection), we are called with a pointer to the 1785 * interface out which this connection should go. 1786 * 1787 * NOTE: Do not subtract IP option/extension header size nor IPsec 1788 * header size from MSS advertisement. MSS option must hold the maximum 1789 * segment size we can accept, so it must always be: 1790 * max(if mtu) - ip header - tcp header 1791 */ 1792 u_long 1793 tcp_mss_to_advertise(const struct ifnet *ifp, int af) 1794 { 1795 extern u_long in_maxmtu; 1796 u_long mss = 0; 1797 u_long hdrsiz; 1798 1799 /* 1800 * In order to avoid defeating path MTU discovery on the peer, 1801 * we advertise the max MTU of all attached networks as our MSS, 1802 * per RFC 1191, section 3.1. 1803 * 1804 * We provide the option to advertise just the MTU of 1805 * the interface on which we hope this connection will 1806 * be receiving. If we are responding to a SYN, we 1807 * will have a pretty good idea about this, but when 1808 * initiating a connection there is a bit more doubt. 1809 * 1810 * We also need to ensure that loopback has a large enough 1811 * MSS, as the loopback MTU is never included in in_maxmtu. 1812 */ 1813 1814 if (ifp != NULL) 1815 switch (af) { 1816 case AF_INET: 1817 mss = ifp->if_mtu; 1818 break; 1819 #ifdef INET6 1820 case AF_INET6: 1821 mss = IN6_LINKMTU(ifp); 1822 break; 1823 #endif 1824 } 1825 1826 if (tcp_mss_ifmtu == 0) 1827 switch (af) { 1828 case AF_INET: 1829 mss = max(in_maxmtu, mss); 1830 break; 1831 #ifdef INET6 1832 case AF_INET6: 1833 mss = max(in6_maxmtu, mss); 1834 break; 1835 #endif 1836 } 1837 1838 switch (af) { 1839 case AF_INET: 1840 hdrsiz = sizeof(struct ip); 1841 break; 1842 #ifdef INET6 1843 case AF_INET6: 1844 hdrsiz = sizeof(struct ip6_hdr); 1845 break; 1846 #endif 1847 default: 1848 hdrsiz = 0; 1849 break; 1850 } 1851 hdrsiz += sizeof(struct tcphdr); 1852 if (mss > hdrsiz) 1853 mss -= hdrsiz; 1854 1855 mss = max(tcp_mssdflt, mss); 1856 return (mss); 1857 } 1858 1859 /* 1860 * Set connection variables based on the peer's advertised MSS. 1861 * We are passed the TCPCB for the actual connection. If we 1862 * are the server, we are called by the compressed state engine 1863 * when the 3-way handshake is complete. If we are the client, 1864 * we are called when we receive the SYN,ACK from the server. 1865 * 1866 * NOTE: Our advertised MSS value must be initialized in the TCPCB 1867 * before this routine is called! 1868 */ 1869 void 1870 tcp_mss_from_peer(struct tcpcb *tp, int offer) 1871 { 1872 struct socket *so; 1873 #if defined(RTV_SPIPE) || defined(RTV_SSTHRESH) 1874 struct rtentry *rt; 1875 #endif 1876 u_long bufsize; 1877 int mss; 1878 1879 #ifdef DIAGNOSTIC 1880 if (tp->t_inpcb && tp->t_in6pcb) 1881 panic("tcp_mss_from_peer: both t_inpcb and t_in6pcb are set"); 1882 #endif 1883 so = NULL; 1884 rt = NULL; 1885 #ifdef INET 1886 if (tp->t_inpcb) { 1887 so = tp->t_inpcb->inp_socket; 1888 #if defined(RTV_SPIPE) || defined(RTV_SSTHRESH) 1889 rt = in_pcbrtentry(tp->t_inpcb); 1890 #endif 1891 } 1892 #endif 1893 #ifdef INET6 1894 if (tp->t_in6pcb) { 1895 so = tp->t_in6pcb->in6p_socket; 1896 #if defined(RTV_SPIPE) || defined(RTV_SSTHRESH) 1897 rt = in6_pcbrtentry(tp->t_in6pcb); 1898 #endif 1899 } 1900 #endif 1901 1902 /* 1903 * As per RFC1122, use the default MSS value, unless they 1904 * sent us an offer. Do not accept offers less than 256 bytes. 1905 */ 1906 mss = tcp_mssdflt; 1907 if (offer) 1908 mss = offer; 1909 mss = max(mss, 256); /* sanity */ 1910 tp->t_peermss = mss; 1911 mss -= tcp_optlen(tp); 1912 #ifdef INET 1913 if (tp->t_inpcb) 1914 mss -= ip_optlen(tp->t_inpcb); 1915 #endif 1916 #ifdef INET6 1917 if (tp->t_in6pcb) 1918 mss -= ip6_optlen(tp->t_in6pcb); 1919 #endif 1920 1921 /* 1922 * If there's a pipesize, change the socket buffer to that size. 1923 * Make the socket buffer an integral number of MSS units. If 1924 * the MSS is larger than the socket buffer, artificially decrease 1925 * the MSS. 1926 */ 1927 #ifdef RTV_SPIPE 1928 if (rt != NULL && rt->rt_rmx.rmx_sendpipe != 0) 1929 bufsize = rt->rt_rmx.rmx_sendpipe; 1930 else 1931 #endif 1932 { 1933 KASSERT(so != NULL); 1934 bufsize = so->so_snd.sb_hiwat; 1935 } 1936 if (bufsize < mss) 1937 mss = bufsize; 1938 else { 1939 bufsize = roundup(bufsize, mss); 1940 if (bufsize > sb_max) 1941 bufsize = sb_max; 1942 (void) sbreserve(&so->so_snd, bufsize, so); 1943 } 1944 tp->t_segsz = mss; 1945 1946 #ifdef RTV_SSTHRESH 1947 if (rt != NULL && rt->rt_rmx.rmx_ssthresh) { 1948 /* 1949 * There's some sort of gateway or interface buffer 1950 * limit on the path. Use this to set the slow 1951 * start threshold, but set the threshold to no less 1952 * than 2 * MSS. 1953 */ 1954 tp->snd_ssthresh = max(2 * mss, rt->rt_rmx.rmx_ssthresh); 1955 } 1956 #endif 1957 } 1958 1959 /* 1960 * Processing necessary when a TCP connection is established. 1961 */ 1962 void 1963 tcp_established(struct tcpcb *tp) 1964 { 1965 struct socket *so; 1966 #ifdef RTV_RPIPE 1967 struct rtentry *rt; 1968 #endif 1969 u_long bufsize; 1970 1971 #ifdef DIAGNOSTIC 1972 if (tp->t_inpcb && tp->t_in6pcb) 1973 panic("tcp_established: both t_inpcb and t_in6pcb are set"); 1974 #endif 1975 so = NULL; 1976 rt = NULL; 1977 #ifdef INET 1978 if (tp->t_inpcb) { 1979 so = tp->t_inpcb->inp_socket; 1980 #if defined(RTV_RPIPE) 1981 rt = in_pcbrtentry(tp->t_inpcb); 1982 #endif 1983 } 1984 #endif 1985 #ifdef INET6 1986 if (tp->t_in6pcb) { 1987 so = tp->t_in6pcb->in6p_socket; 1988 #if defined(RTV_RPIPE) 1989 rt = in6_pcbrtentry(tp->t_in6pcb); 1990 #endif 1991 } 1992 #endif 1993 1994 tp->t_state = TCPS_ESTABLISHED; 1995 TCP_TIMER_ARM(tp, TCPT_KEEP, tcp_keepidle); 1996 1997 #ifdef RTV_RPIPE 1998 if (rt != NULL && rt->rt_rmx.rmx_recvpipe != 0) 1999 bufsize = rt->rt_rmx.rmx_recvpipe; 2000 else 2001 #endif 2002 { 2003 KASSERT(so != NULL); 2004 bufsize = so->so_rcv.sb_hiwat; 2005 } 2006 if (bufsize > tp->t_ourmss) { 2007 bufsize = roundup(bufsize, tp->t_ourmss); 2008 if (bufsize > sb_max) 2009 bufsize = sb_max; 2010 (void) sbreserve(&so->so_rcv, bufsize, so); 2011 } 2012 } 2013 2014 /* 2015 * Check if there's an initial rtt or rttvar. Convert from the 2016 * route-table units to scaled multiples of the slow timeout timer. 2017 * Called only during the 3-way handshake. 2018 */ 2019 void 2020 tcp_rmx_rtt(struct tcpcb *tp) 2021 { 2022 #ifdef RTV_RTT 2023 struct rtentry *rt = NULL; 2024 int rtt; 2025 2026 #ifdef DIAGNOSTIC 2027 if (tp->t_inpcb && tp->t_in6pcb) 2028 panic("tcp_rmx_rtt: both t_inpcb and t_in6pcb are set"); 2029 #endif 2030 #ifdef INET 2031 if (tp->t_inpcb) 2032 rt = in_pcbrtentry(tp->t_inpcb); 2033 #endif 2034 #ifdef INET6 2035 if (tp->t_in6pcb) 2036 rt = in6_pcbrtentry(tp->t_in6pcb); 2037 #endif 2038 if (rt == NULL) 2039 return; 2040 2041 if (tp->t_srtt == 0 && (rtt = rt->rt_rmx.rmx_rtt)) { 2042 /* 2043 * XXX The lock bit for MTU indicates that the value 2044 * is also a minimum value; this is subject to time. 2045 */ 2046 if (rt->rt_rmx.rmx_locks & RTV_RTT) 2047 TCPT_RANGESET(tp->t_rttmin, 2048 rtt / (RTM_RTTUNIT / PR_SLOWHZ), 2049 TCPTV_MIN, TCPTV_REXMTMAX); 2050 tp->t_srtt = rtt / 2051 ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTT_SHIFT + 2)); 2052 if (rt->rt_rmx.rmx_rttvar) { 2053 tp->t_rttvar = rt->rt_rmx.rmx_rttvar / 2054 ((RTM_RTTUNIT / PR_SLOWHZ) >> 2055 (TCP_RTTVAR_SHIFT + 2)); 2056 } else { 2057 /* Default variation is +- 1 rtt */ 2058 tp->t_rttvar = 2059 tp->t_srtt >> (TCP_RTT_SHIFT - TCP_RTTVAR_SHIFT); 2060 } 2061 TCPT_RANGESET(tp->t_rxtcur, 2062 ((tp->t_srtt >> 2) + tp->t_rttvar) >> (1 + 2), 2063 tp->t_rttmin, TCPTV_REXMTMAX); 2064 } 2065 #endif 2066 } 2067 2068 tcp_seq tcp_iss_seq = 0; /* tcp initial seq # */ 2069 #if NRND > 0 2070 u_int8_t tcp_iss_secret[16]; /* 128 bits; should be plenty */ 2071 #endif 2072 2073 /* 2074 * Get a new sequence value given a tcp control block 2075 */ 2076 tcp_seq 2077 tcp_new_iss(struct tcpcb *tp, tcp_seq addin) 2078 { 2079 2080 #ifdef INET 2081 if (tp->t_inpcb != NULL) { 2082 return (tcp_new_iss1(&tp->t_inpcb->inp_laddr, 2083 &tp->t_inpcb->inp_faddr, tp->t_inpcb->inp_lport, 2084 tp->t_inpcb->inp_fport, sizeof(tp->t_inpcb->inp_laddr), 2085 addin)); 2086 } 2087 #endif 2088 #ifdef INET6 2089 if (tp->t_in6pcb != NULL) { 2090 return (tcp_new_iss1(&tp->t_in6pcb->in6p_laddr, 2091 &tp->t_in6pcb->in6p_faddr, tp->t_in6pcb->in6p_lport, 2092 tp->t_in6pcb->in6p_fport, sizeof(tp->t_in6pcb->in6p_laddr), 2093 addin)); 2094 } 2095 #endif 2096 /* Not possible. */ 2097 panic("tcp_new_iss"); 2098 } 2099 2100 /* 2101 * This routine actually generates a new TCP initial sequence number. 2102 */ 2103 tcp_seq 2104 tcp_new_iss1(void *laddr, void *faddr, u_int16_t lport, u_int16_t fport, 2105 size_t addrsz, tcp_seq addin) 2106 { 2107 tcp_seq tcp_iss; 2108 2109 #if NRND > 0 2110 static int beenhere; 2111 2112 /* 2113 * If we haven't been here before, initialize our cryptographic 2114 * hash secret. 2115 */ 2116 if (beenhere == 0) { 2117 rnd_extract_data(tcp_iss_secret, sizeof(tcp_iss_secret), 2118 RND_EXTRACT_ANY); 2119 beenhere = 1; 2120 } 2121 2122 if (tcp_do_rfc1948) { 2123 MD5_CTX ctx; 2124 u_int8_t hash[16]; /* XXX MD5 knowledge */ 2125 2126 /* 2127 * Compute the base value of the ISS. It is a hash 2128 * of (saddr, sport, daddr, dport, secret). 2129 */ 2130 MD5Init(&ctx); 2131 2132 MD5Update(&ctx, (u_char *) laddr, addrsz); 2133 MD5Update(&ctx, (u_char *) &lport, sizeof(lport)); 2134 2135 MD5Update(&ctx, (u_char *) faddr, addrsz); 2136 MD5Update(&ctx, (u_char *) &fport, sizeof(fport)); 2137 2138 MD5Update(&ctx, tcp_iss_secret, sizeof(tcp_iss_secret)); 2139 2140 MD5Final(hash, &ctx); 2141 2142 memcpy(&tcp_iss, hash, sizeof(tcp_iss)); 2143 2144 /* 2145 * Now increment our "timer", and add it in to 2146 * the computed value. 2147 * 2148 * XXX Use `addin'? 2149 * XXX TCP_ISSINCR too large to use? 2150 */ 2151 tcp_iss_seq += TCP_ISSINCR; 2152 #ifdef TCPISS_DEBUG 2153 printf("ISS hash 0x%08x, ", tcp_iss); 2154 #endif 2155 tcp_iss += tcp_iss_seq + addin; 2156 #ifdef TCPISS_DEBUG 2157 printf("new ISS 0x%08x\n", tcp_iss); 2158 #endif 2159 } else 2160 #endif /* NRND > 0 */ 2161 { 2162 /* 2163 * Randomize. 2164 */ 2165 #if NRND > 0 2166 rnd_extract_data(&tcp_iss, sizeof(tcp_iss), RND_EXTRACT_ANY); 2167 #else 2168 tcp_iss = arc4random(); 2169 #endif 2170 2171 /* 2172 * If we were asked to add some amount to a known value, 2173 * we will take a random value obtained above, mask off 2174 * the upper bits, and add in the known value. We also 2175 * add in a constant to ensure that we are at least a 2176 * certain distance from the original value. 2177 * 2178 * This is used when an old connection is in timed wait 2179 * and we have a new one coming in, for instance. 2180 */ 2181 if (addin != 0) { 2182 #ifdef TCPISS_DEBUG 2183 printf("Random %08x, ", tcp_iss); 2184 #endif 2185 tcp_iss &= TCP_ISS_RANDOM_MASK; 2186 tcp_iss += addin + TCP_ISSINCR; 2187 #ifdef TCPISS_DEBUG 2188 printf("Old ISS %08x, ISS %08x\n", addin, tcp_iss); 2189 #endif 2190 } else { 2191 tcp_iss &= TCP_ISS_RANDOM_MASK; 2192 tcp_iss += tcp_iss_seq; 2193 tcp_iss_seq += TCP_ISSINCR; 2194 #ifdef TCPISS_DEBUG 2195 printf("ISS %08x\n", tcp_iss); 2196 #endif 2197 } 2198 } 2199 2200 if (tcp_compat_42) { 2201 /* 2202 * Limit it to the positive range for really old TCP 2203 * implementations. 2204 * Just AND off the top bit instead of checking if 2205 * is set first - saves a branch 50% of the time. 2206 */ 2207 tcp_iss &= 0x7fffffff; /* XXX */ 2208 } 2209 2210 return (tcp_iss); 2211 } 2212 2213 #if defined(IPSEC) || defined(FAST_IPSEC) 2214 /* compute ESP/AH header size for TCP, including outer IP header. */ 2215 size_t 2216 ipsec4_hdrsiz_tcp(struct tcpcb *tp) 2217 { 2218 struct inpcb *inp; 2219 size_t hdrsiz; 2220 2221 /* XXX mapped addr case (tp->t_in6pcb) */ 2222 if (!tp || !tp->t_template || !(inp = tp->t_inpcb)) 2223 return 0; 2224 switch (tp->t_family) { 2225 case AF_INET: 2226 /* XXX: should use currect direction. */ 2227 hdrsiz = ipsec4_hdrsiz(tp->t_template, IPSEC_DIR_OUTBOUND, inp); 2228 break; 2229 default: 2230 hdrsiz = 0; 2231 break; 2232 } 2233 2234 return hdrsiz; 2235 } 2236 2237 #ifdef INET6 2238 size_t 2239 ipsec6_hdrsiz_tcp(struct tcpcb *tp) 2240 { 2241 struct in6pcb *in6p; 2242 size_t hdrsiz; 2243 2244 if (!tp || !tp->t_template || !(in6p = tp->t_in6pcb)) 2245 return 0; 2246 switch (tp->t_family) { 2247 case AF_INET6: 2248 /* XXX: should use currect direction. */ 2249 hdrsiz = ipsec6_hdrsiz(tp->t_template, IPSEC_DIR_OUTBOUND, in6p); 2250 break; 2251 case AF_INET: 2252 /* mapped address case - tricky */ 2253 default: 2254 hdrsiz = 0; 2255 break; 2256 } 2257 2258 return hdrsiz; 2259 } 2260 #endif 2261 #endif /*IPSEC*/ 2262 2263 /* 2264 * Determine the length of the TCP options for this connection. 2265 * 2266 * XXX: What do we do for SACK, when we add that? Just reserve 2267 * all of the space? Otherwise we can't exactly be incrementing 2268 * cwnd by an amount that varies depending on the amount we last 2269 * had to SACK! 2270 */ 2271 2272 u_int 2273 tcp_optlen(struct tcpcb *tp) 2274 { 2275 u_int optlen; 2276 2277 optlen = 0; 2278 if ((tp->t_flags & (TF_REQ_TSTMP|TF_RCVD_TSTMP|TF_NOOPT)) == 2279 (TF_REQ_TSTMP | TF_RCVD_TSTMP)) 2280 optlen += TCPOLEN_TSTAMP_APPA; 2281 2282 #ifdef TCP_SIGNATURE 2283 #if defined(INET6) && defined(FAST_IPSEC) 2284 if (tp->t_family == AF_INET) 2285 #endif 2286 if (tp->t_flags & TF_SIGNATURE) 2287 optlen += TCPOLEN_SIGNATURE + 2; 2288 #endif /* TCP_SIGNATURE */ 2289 2290 return optlen; 2291 } 2292 2293 u_int 2294 tcp_hdrsz(struct tcpcb *tp) 2295 { 2296 u_int hlen; 2297 2298 switch (tp->t_family) { 2299 #ifdef INET6 2300 case AF_INET6: 2301 hlen = sizeof(struct ip6_hdr); 2302 break; 2303 #endif 2304 case AF_INET: 2305 hlen = sizeof(struct ip); 2306 break; 2307 default: 2308 hlen = 0; 2309 break; 2310 } 2311 hlen += sizeof(struct tcphdr); 2312 2313 if ((tp->t_flags & (TF_REQ_TSTMP|TF_NOOPT)) == TF_REQ_TSTMP && 2314 (tp->t_flags & TF_RCVD_TSTMP) == TF_RCVD_TSTMP) 2315 hlen += TCPOLEN_TSTAMP_APPA; 2316 #ifdef TCP_SIGNATURE 2317 if (tp->t_flags & TF_SIGNATURE) 2318 hlen += TCPOLEN_SIGLEN; 2319 #endif 2320 return hlen; 2321 } 2322
Indexes created Mon Sep 22 05:09:51 GMT 2025