tcp_subr.c revision 1.140
1 /* $NetBSD: tcp_subr.c,v 1.140 2003/06/23 11:02:15 martin Exp $ */ 2 3 /* 4 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. 5 * All rights reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 3. Neither the name of the project nor the names of its contributors 16 * may be used to endorse or promote products derived from this software 17 * without specific prior written permission. 18 * 19 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND 20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 22 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE 23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 29 * SUCH DAMAGE. 30 */ 31 32 /*- 33 * Copyright (c) 1997, 1998, 2000, 2001 The NetBSD Foundation, Inc. 34 * All rights reserved. 35 * 36 * This code is derived from software contributed to The NetBSD Foundation 37 * by Jason R. Thorpe and Kevin M. Lahey of the Numerical Aerospace Simulation 38 * Facility, NASA Ames Research Center. 39 * 40 * Redistribution and use in source and binary forms, with or without 41 * modification, are permitted provided that the following conditions 42 * are met: 43 * 1. Redistributions of source code must retain the above copyright 44 * notice, this list of conditions and the following disclaimer. 45 * 2. Redistributions in binary form must reproduce the above copyright 46 * notice, this list of conditions and the following disclaimer in the 47 * documentation and/or other materials provided with the distribution. 48 * 3. All advertising materials mentioning features or use of this software 49 * must display the following acknowledgement: 50 * This product includes software developed by the NetBSD 51 * Foundation, Inc. and its contributors. 52 * 4. Neither the name of The NetBSD Foundation nor the names of its 53 * contributors may be used to endorse or promote products derived 54 * from this software without specific prior written permission. 55 * 56 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 57 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 58 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 59 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 60 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 61 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 62 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 63 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 64 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 65 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 66 * POSSIBILITY OF SUCH DAMAGE. 67 */ 68 69 /* 70 * Copyright (c) 1982, 1986, 1988, 1990, 1993, 1995 71 * The Regents of the University of California. All rights reserved. 72 * 73 * Redistribution and use in source and binary forms, with or without 74 * modification, are permitted provided that the following conditions 75 * are met: 76 * 1. Redistributions of source code must retain the above copyright 77 * notice, this list of conditions and the following disclaimer. 78 * 2. Redistributions in binary form must reproduce the above copyright 79 * notice, this list of conditions and the following disclaimer in the 80 * documentation and/or other materials provided with the distribution. 81 * 3. All advertising materials mentioning features or use of this software 82 * must display the following acknowledgement: 83 * This product includes software developed by the University of 84 * California, Berkeley and its contributors. 85 * 4. Neither the name of the University nor the names of its contributors 86 * may be used to endorse or promote products derived from this software 87 * without specific prior written permission. 88 * 89 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 90 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 91 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 92 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 93 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 94 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 95 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 96 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 97 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 98 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 99 * SUCH DAMAGE. 100 * 101 * @(#)tcp_subr.c 8.2 (Berkeley) 5/24/95 102 */ 103 104 #include <sys/cdefs.h> 105 __KERNEL_RCSID(0, "$NetBSD: tcp_subr.c,v 1.140 2003/06/23 11:02:15 martin Exp $"); 106 107 #include "opt_inet.h" 108 #include "opt_ipsec.h" 109 #include "opt_tcp_compat_42.h" 110 #include "opt_inet_csum.h" 111 #include "opt_mbuftrace.h" 112 #include "rnd.h" 113 114 #include <sys/param.h> 115 #include <sys/proc.h> 116 #include <sys/systm.h> 117 #include <sys/malloc.h> 118 #include <sys/mbuf.h> 119 #include <sys/socket.h> 120 #include <sys/socketvar.h> 121 #include <sys/protosw.h> 122 #include <sys/errno.h> 123 #include <sys/kernel.h> 124 #include <sys/pool.h> 125 #if NRND > 0 126 #include <sys/md5.h> 127 #include <sys/rnd.h> 128 #endif 129 130 #include <net/route.h> 131 #include <net/if.h> 132 133 #include <netinet/in.h> 134 #include <netinet/in_systm.h> 135 #include <netinet/ip.h> 136 #include <netinet/in_pcb.h> 137 #include <netinet/ip_var.h> 138 #include <netinet/ip_icmp.h> 139 140 #ifdef INET6 141 #ifndef INET 142 #include <netinet/in.h> 143 #endif 144 #include <netinet/ip6.h> 145 #include <netinet6/in6_pcb.h> 146 #include <netinet6/ip6_var.h> 147 #include <netinet6/in6_var.h> 148 #include <netinet6/ip6protosw.h> 149 #include <netinet/icmp6.h> 150 #include <netinet6/nd6.h> 151 #endif 152 153 #include <netinet/tcp.h> 154 #include <netinet/tcp_fsm.h> 155 #include <netinet/tcp_seq.h> 156 #include <netinet/tcp_timer.h> 157 #include <netinet/tcp_var.h> 158 #include <netinet/tcpip.h> 159 160 #ifdef IPSEC 161 #include <netinet6/ipsec.h> 162 #endif /*IPSEC*/ 163 164 #ifdef INET6 165 struct in6pcb tcb6; 166 #endif 167 168 struct inpcbtable tcbtable; /* head of queue of active tcpcb's */ 169 struct tcpstat tcpstat; /* tcp statistics */ 170 u_int32_t tcp_now; /* for RFC 1323 timestamps */ 171 172 /* patchable/settable parameters for tcp */ 173 int tcp_mssdflt = TCP_MSS; 174 int tcp_rttdflt = TCPTV_SRTTDFLT / PR_SLOWHZ; 175 int tcp_do_rfc1323 = 1; /* window scaling / timestamps (obsolete) */ 176 #if NRND > 0 177 int tcp_do_rfc1948 = 0; /* ISS by cryptographic hash */ 178 #endif 179 int tcp_do_sack = 1; /* selective acknowledgement */ 180 int tcp_do_win_scale = 1; /* RFC1323 window scaling */ 181 int tcp_do_timestamps = 1; /* RFC1323 timestamps */ 182 int tcp_do_newreno = 0; /* Use the New Reno algorithms */ 183 int tcp_ack_on_push = 0; /* set to enable immediate ACK-on-PUSH */ 184 int tcp_init_win = 1; /* initial slow start window */ 185 int tcp_init_win_local = 4; /* initial slow start window for local nets */ 186 int tcp_mss_ifmtu = 0; 187 #ifdef TCP_COMPAT_42 188 int tcp_compat_42 = 1; 189 #else 190 int tcp_compat_42 = 0; 191 #endif 192 int tcp_rst_ppslim = 100; /* 100pps */ 193 194 /* tcb hash */ 195 #ifndef TCBHASHSIZE 196 #define TCBHASHSIZE 128 197 #endif 198 int tcbhashsize = TCBHASHSIZE; 199 200 /* syn hash parameters */ 201 #define TCP_SYN_HASH_SIZE 293 202 #define TCP_SYN_BUCKET_SIZE 35 203 int tcp_syn_cache_size = TCP_SYN_HASH_SIZE; 204 int tcp_syn_cache_limit = TCP_SYN_HASH_SIZE*TCP_SYN_BUCKET_SIZE; 205 int tcp_syn_bucket_limit = 3*TCP_SYN_BUCKET_SIZE; 206 struct syn_cache_head tcp_syn_cache[TCP_SYN_HASH_SIZE]; 207 208 int tcp_freeq __P((struct tcpcb *)); 209 210 #ifdef INET 211 void tcp_mtudisc_callback __P((struct in_addr)); 212 #endif 213 #ifdef INET6 214 void tcp6_mtudisc_callback __P((struct in6_addr *)); 215 #endif 216 217 void tcp_mtudisc __P((struct inpcb *, int)); 218 #ifdef INET6 219 void tcp6_mtudisc __P((struct in6pcb *, int)); 220 #endif 221 222 struct pool tcpcb_pool; 223 224 #ifdef TCP_CSUM_COUNTERS 225 #include <sys/device.h> 226 227 struct evcnt tcp_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 228 NULL, "tcp", "hwcsum bad"); 229 struct evcnt tcp_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 230 NULL, "tcp", "hwcsum ok"); 231 struct evcnt tcp_hwcsum_data = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 232 NULL, "tcp", "hwcsum data"); 233 struct evcnt tcp_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 234 NULL, "tcp", "swcsum"); 235 #endif /* TCP_CSUM_COUNTERS */ 236 237 #ifdef TCP_OUTPUT_COUNTERS 238 #include <sys/device.h> 239 240 struct evcnt tcp_output_bigheader = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 241 NULL, "tcp", "output big header"); 242 struct evcnt tcp_output_copysmall = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 243 NULL, "tcp", "output copy small"); 244 struct evcnt tcp_output_copybig = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 245 NULL, "tcp", "output copy big"); 246 struct evcnt tcp_output_refbig = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 247 NULL, "tcp", "output reference big"); 248 #endif /* TCP_OUTPUT_COUNTERS */ 249 250 #ifdef TCP_REASS_COUNTERS 251 #include <sys/device.h> 252 253 struct evcnt tcp_reass_ = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 254 NULL, "tcp_reass", "calls"); 255 struct evcnt tcp_reass_empty = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 256 &tcp_reass_, "tcp_reass", "insert into empty queue"); 257 struct evcnt tcp_reass_iteration[8] = { 258 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", ">7 iterations"), 259 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "1 iteration"), 260 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "2 iterations"), 261 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "3 iterations"), 262 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "4 iterations"), 263 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "5 iterations"), 264 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "6 iterations"), 265 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "7 iterations"), 266 }; 267 struct evcnt tcp_reass_prependfirst = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 268 &tcp_reass_, "tcp_reass", "prepend to first"); 269 struct evcnt tcp_reass_prepend = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 270 &tcp_reass_, "tcp_reass", "prepend"); 271 struct evcnt tcp_reass_insert = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 272 &tcp_reass_, "tcp_reass", "insert"); 273 struct evcnt tcp_reass_inserttail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 274 &tcp_reass_, "tcp_reass", "insert at tail"); 275 struct evcnt tcp_reass_append = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 276 &tcp_reass_, "tcp_reass", "append"); 277 struct evcnt tcp_reass_appendtail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 278 &tcp_reass_, "tcp_reass", "append to tail fragment"); 279 struct evcnt tcp_reass_overlaptail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 280 &tcp_reass_, "tcp_reass", "overlap at end"); 281 struct evcnt tcp_reass_overlapfront = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 282 &tcp_reass_, "tcp_reass", "overlap at start"); 283 struct evcnt tcp_reass_segdup = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 284 &tcp_reass_, "tcp_reass", "duplicate segment"); 285 struct evcnt tcp_reass_fragdup = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 286 &tcp_reass_, "tcp_reass", "duplicate fragment"); 287 288 #endif /* TCP_REASS_COUNTERS */ 289 290 #ifdef MBUFTRACE 291 struct mowner tcp_mowner = { "tcp" }; 292 struct mowner tcp_rx_mowner = { "tcp", "rx" }; 293 struct mowner tcp_tx_mowner = { "tcp", "tx" }; 294 #endif 295 296 /* 297 * Tcp initialization 298 */ 299 void 300 tcp_init() 301 { 302 int hlen; 303 304 pool_init(&tcpcb_pool, sizeof(struct tcpcb), 0, 0, 0, "tcpcbpl", 305 NULL); 306 in_pcbinit(&tcbtable, tcbhashsize, tcbhashsize); 307 #ifdef INET6 308 tcb6.in6p_next = tcb6.in6p_prev = &tcb6; 309 #endif 310 311 hlen = sizeof(struct ip) + sizeof(struct tcphdr); 312 #ifdef INET6 313 if (sizeof(struct ip) < sizeof(struct ip6_hdr)) 314 hlen = sizeof(struct ip6_hdr) + sizeof(struct tcphdr); 315 #endif 316 if (max_protohdr < hlen) 317 max_protohdr = hlen; 318 if (max_linkhdr + hlen > MHLEN) 319 panic("tcp_init"); 320 321 #ifdef INET 322 icmp_mtudisc_callback_register(tcp_mtudisc_callback); 323 #endif 324 #ifdef INET6 325 icmp6_mtudisc_callback_register(tcp6_mtudisc_callback); 326 #endif 327 328 /* Initialize timer state. */ 329 tcp_timer_init(); 330 331 /* Initialize the compressed state engine. */ 332 syn_cache_init(); 333 334 #ifdef TCP_CSUM_COUNTERS 335 evcnt_attach_static(&tcp_hwcsum_bad); 336 evcnt_attach_static(&tcp_hwcsum_ok); 337 evcnt_attach_static(&tcp_hwcsum_data); 338 evcnt_attach_static(&tcp_swcsum); 339 #endif /* TCP_CSUM_COUNTERS */ 340 341 #ifdef TCP_OUTPUT_COUNTERS 342 evcnt_attach_static(&tcp_output_bigheader); 343 evcnt_attach_static(&tcp_output_copysmall); 344 evcnt_attach_static(&tcp_output_copybig); 345 evcnt_attach_static(&tcp_output_refbig); 346 #endif /* TCP_OUTPUT_COUNTERS */ 347 348 #ifdef TCP_REASS_COUNTERS 349 evcnt_attach_static(&tcp_reass_); 350 evcnt_attach_static(&tcp_reass_empty); 351 evcnt_attach_static(&tcp_reass_iteration[0]); 352 evcnt_attach_static(&tcp_reass_iteration[1]); 353 evcnt_attach_static(&tcp_reass_iteration[2]); 354 evcnt_attach_static(&tcp_reass_iteration[3]); 355 evcnt_attach_static(&tcp_reass_iteration[4]); 356 evcnt_attach_static(&tcp_reass_iteration[5]); 357 evcnt_attach_static(&tcp_reass_iteration[6]); 358 evcnt_attach_static(&tcp_reass_iteration[7]); 359 evcnt_attach_static(&tcp_reass_prependfirst); 360 evcnt_attach_static(&tcp_reass_prepend); 361 evcnt_attach_static(&tcp_reass_insert); 362 evcnt_attach_static(&tcp_reass_inserttail); 363 evcnt_attach_static(&tcp_reass_append); 364 evcnt_attach_static(&tcp_reass_appendtail); 365 evcnt_attach_static(&tcp_reass_overlaptail); 366 evcnt_attach_static(&tcp_reass_overlapfront); 367 evcnt_attach_static(&tcp_reass_segdup); 368 evcnt_attach_static(&tcp_reass_fragdup); 369 #endif /* TCP_REASS_COUNTERS */ 370 371 MOWNER_ATTACH(&tcp_tx_mowner); 372 MOWNER_ATTACH(&tcp_rx_mowner); 373 MOWNER_ATTACH(&tcp_mowner); 374 } 375 376 /* 377 * Create template to be used to send tcp packets on a connection. 378 * Call after host entry created, allocates an mbuf and fills 379 * in a skeletal tcp/ip header, minimizing the amount of work 380 * necessary when the connection is used. 381 */ 382 struct mbuf * 383 tcp_template(tp) 384 struct tcpcb *tp; 385 { 386 struct inpcb *inp = tp->t_inpcb; 387 #ifdef INET6 388 struct in6pcb *in6p = tp->t_in6pcb; 389 #endif 390 struct tcphdr *n; 391 struct mbuf *m; 392 int hlen; 393 394 switch (tp->t_family) { 395 case AF_INET: 396 hlen = sizeof(struct ip); 397 if (inp) 398 break; 399 #ifdef INET6 400 if (in6p) { 401 /* mapped addr case */ 402 if (IN6_IS_ADDR_V4MAPPED(&in6p->in6p_laddr) 403 && IN6_IS_ADDR_V4MAPPED(&in6p->in6p_faddr)) 404 break; 405 } 406 #endif 407 return NULL; /*EINVAL*/ 408 #ifdef INET6 409 case AF_INET6: 410 hlen = sizeof(struct ip6_hdr); 411 if (in6p) { 412 /* more sainty check? */ 413 break; 414 } 415 return NULL; /*EINVAL*/ 416 #endif 417 default: 418 hlen = 0; /*pacify gcc*/ 419 return NULL; /*EAFNOSUPPORT*/ 420 } 421 #ifdef DIAGNOSTIC 422 if (hlen + sizeof(struct tcphdr) > MCLBYTES) 423 panic("mclbytes too small for t_template"); 424 #endif 425 m = tp->t_template; 426 if (m && m->m_len == hlen + sizeof(struct tcphdr)) 427 ; 428 else { 429 if (m) 430 m_freem(m); 431 m = tp->t_template = NULL; 432 MGETHDR(m, M_DONTWAIT, MT_HEADER); 433 if (m && hlen + sizeof(struct tcphdr) > MHLEN) { 434 MCLGET(m, M_DONTWAIT); 435 if ((m->m_flags & M_EXT) == 0) { 436 m_free(m); 437 m = NULL; 438 } 439 } 440 if (m == NULL) 441 return NULL; 442 MCLAIM(m, &tcp_mowner); 443 m->m_pkthdr.len = m->m_len = hlen + sizeof(struct tcphdr); 444 } 445 446 bzero(mtod(m, caddr_t), m->m_len); 447 448 n = (struct tcphdr *)(mtod(m, caddr_t) + hlen); 449 450 switch (tp->t_family) { 451 case AF_INET: 452 { 453 struct ipovly *ipov; 454 mtod(m, struct ip *)->ip_v = 4; 455 ipov = mtod(m, struct ipovly *); 456 ipov->ih_pr = IPPROTO_TCP; 457 ipov->ih_len = htons(sizeof(struct tcphdr)); 458 if (inp) { 459 ipov->ih_src = inp->inp_laddr; 460 ipov->ih_dst = inp->inp_faddr; 461 } 462 #ifdef INET6 463 else if (in6p) { 464 /* mapped addr case */ 465 bcopy(&in6p->in6p_laddr.s6_addr32[3], &ipov->ih_src, 466 sizeof(ipov->ih_src)); 467 bcopy(&in6p->in6p_faddr.s6_addr32[3], &ipov->ih_dst, 468 sizeof(ipov->ih_dst)); 469 } 470 #endif 471 /* 472 * Compute the pseudo-header portion of the checksum 473 * now. We incrementally add in the TCP option and 474 * payload lengths later, and then compute the TCP 475 * checksum right before the packet is sent off onto 476 * the wire. 477 */ 478 n->th_sum = in_cksum_phdr(ipov->ih_src.s_addr, 479 ipov->ih_dst.s_addr, 480 htons(sizeof(struct tcphdr) + IPPROTO_TCP)); 481 break; 482 } 483 #ifdef INET6 484 case AF_INET6: 485 { 486 struct ip6_hdr *ip6; 487 mtod(m, struct ip *)->ip_v = 6; 488 ip6 = mtod(m, struct ip6_hdr *); 489 ip6->ip6_nxt = IPPROTO_TCP; 490 ip6->ip6_plen = htons(sizeof(struct tcphdr)); 491 ip6->ip6_src = in6p->in6p_laddr; 492 ip6->ip6_dst = in6p->in6p_faddr; 493 ip6->ip6_flow = in6p->in6p_flowinfo & IPV6_FLOWINFO_MASK; 494 if (ip6_auto_flowlabel) { 495 ip6->ip6_flow &= ~IPV6_FLOWLABEL_MASK; 496 ip6->ip6_flow |= 497 (htonl(ip6_flow_seq++) & IPV6_FLOWLABEL_MASK); 498 } 499 ip6->ip6_vfc &= ~IPV6_VERSION_MASK; 500 ip6->ip6_vfc |= IPV6_VERSION; 501 502 /* 503 * Compute the pseudo-header portion of the checksum 504 * now. We incrementally add in the TCP option and 505 * payload lengths later, and then compute the TCP 506 * checksum right before the packet is sent off onto 507 * the wire. 508 */ 509 n->th_sum = in6_cksum_phdr(&in6p->in6p_laddr, 510 &in6p->in6p_faddr, htonl(sizeof(struct tcphdr)), 511 htonl(IPPROTO_TCP)); 512 break; 513 } 514 #endif 515 } 516 if (inp) { 517 n->th_sport = inp->inp_lport; 518 n->th_dport = inp->inp_fport; 519 } 520 #ifdef INET6 521 else if (in6p) { 522 n->th_sport = in6p->in6p_lport; 523 n->th_dport = in6p->in6p_fport; 524 } 525 #endif 526 n->th_seq = 0; 527 n->th_ack = 0; 528 n->th_x2 = 0; 529 n->th_off = 5; 530 n->th_flags = 0; 531 n->th_win = 0; 532 n->th_urp = 0; 533 return (m); 534 } 535 536 /* 537 * Send a single message to the TCP at address specified by 538 * the given TCP/IP header. If m == 0, then we make a copy 539 * of the tcpiphdr at ti and send directly to the addressed host. 540 * This is used to force keep alive messages out using the TCP 541 * template for a connection tp->t_template. If flags are given 542 * then we send a message back to the TCP which originated the 543 * segment ti, and discard the mbuf containing it and any other 544 * attached mbufs. 545 * 546 * In any case the ack and sequence number of the transmitted 547 * segment are as specified by the parameters. 548 */ 549 int 550 tcp_respond(tp, template, m, th0, ack, seq, flags) 551 struct tcpcb *tp; 552 struct mbuf *template; 553 struct mbuf *m; 554 struct tcphdr *th0; 555 tcp_seq ack, seq; 556 int flags; 557 { 558 struct route *ro; 559 int error, tlen, win = 0; 560 int hlen; 561 struct ip *ip; 562 #ifdef INET6 563 struct ip6_hdr *ip6; 564 #endif 565 int family; /* family on packet, not inpcb/in6pcb! */ 566 struct tcphdr *th; 567 568 if (tp != NULL && (flags & TH_RST) == 0) { 569 #ifdef DIAGNOSTIC 570 if (tp->t_inpcb && tp->t_in6pcb) 571 panic("tcp_respond: both t_inpcb and t_in6pcb are set"); 572 #endif 573 #ifdef INET 574 if (tp->t_inpcb) 575 win = sbspace(&tp->t_inpcb->inp_socket->so_rcv); 576 #endif 577 #ifdef INET6 578 if (tp->t_in6pcb) 579 win = sbspace(&tp->t_in6pcb->in6p_socket->so_rcv); 580 #endif 581 } 582 583 th = NULL; /* Quell uninitialized warning */ 584 ip = NULL; 585 #ifdef INET6 586 ip6 = NULL; 587 #endif 588 if (m == 0) { 589 if (!template) 590 return EINVAL; 591 592 /* get family information from template */ 593 switch (mtod(template, struct ip *)->ip_v) { 594 case 4: 595 family = AF_INET; 596 hlen = sizeof(struct ip); 597 break; 598 #ifdef INET6 599 case 6: 600 family = AF_INET6; 601 hlen = sizeof(struct ip6_hdr); 602 break; 603 #endif 604 default: 605 return EAFNOSUPPORT; 606 } 607 608 MGETHDR(m, M_DONTWAIT, MT_HEADER); 609 if (m) { 610 MCLAIM(m, &tcp_tx_mowner); 611 MCLGET(m, M_DONTWAIT); 612 if ((m->m_flags & M_EXT) == 0) { 613 m_free(m); 614 m = NULL; 615 } 616 } 617 if (m == NULL) 618 return (ENOBUFS); 619 620 if (tcp_compat_42) 621 tlen = 1; 622 else 623 tlen = 0; 624 625 m->m_data += max_linkhdr; 626 bcopy(mtod(template, caddr_t), mtod(m, caddr_t), 627 template->m_len); 628 switch (family) { 629 case AF_INET: 630 ip = mtod(m, struct ip *); 631 th = (struct tcphdr *)(ip + 1); 632 break; 633 #ifdef INET6 634 case AF_INET6: 635 ip6 = mtod(m, struct ip6_hdr *); 636 th = (struct tcphdr *)(ip6 + 1); 637 break; 638 #endif 639 #if 0 640 default: 641 /* noone will visit here */ 642 m_freem(m); 643 return EAFNOSUPPORT; 644 #endif 645 } 646 flags = TH_ACK; 647 } else { 648 649 if ((m->m_flags & M_PKTHDR) == 0) { 650 #if 0 651 printf("non PKTHDR to tcp_respond\n"); 652 #endif 653 m_freem(m); 654 return EINVAL; 655 } 656 #ifdef DIAGNOSTIC 657 if (!th0) 658 panic("th0 == NULL in tcp_respond"); 659 #endif 660 661 /* get family information from m */ 662 switch (mtod(m, struct ip *)->ip_v) { 663 case 4: 664 family = AF_INET; 665 hlen = sizeof(struct ip); 666 ip = mtod(m, struct ip *); 667 break; 668 #ifdef INET6 669 case 6: 670 family = AF_INET6; 671 hlen = sizeof(struct ip6_hdr); 672 ip6 = mtod(m, struct ip6_hdr *); 673 break; 674 #endif 675 default: 676 m_freem(m); 677 return EAFNOSUPPORT; 678 } 679 if ((flags & TH_SYN) == 0 || sizeof(*th0) > (th0->th_off << 2)) 680 tlen = sizeof(*th0); 681 else 682 tlen = th0->th_off << 2; 683 684 if (m->m_len > hlen + tlen && (m->m_flags & M_EXT) == 0 && 685 mtod(m, caddr_t) + hlen == (caddr_t)th0) { 686 m->m_len = hlen + tlen; 687 m_freem(m->m_next); 688 m->m_next = NULL; 689 } else { 690 struct mbuf *n; 691 692 #ifdef DIAGNOSTIC 693 if (max_linkhdr + hlen + tlen > MCLBYTES) { 694 m_freem(m); 695 return EMSGSIZE; 696 } 697 #endif 698 MGETHDR(n, M_DONTWAIT, MT_HEADER); 699 if (n && max_linkhdr + hlen + tlen > MHLEN) { 700 MCLGET(n, M_DONTWAIT); 701 if ((n->m_flags & M_EXT) == 0) { 702 m_freem(n); 703 n = NULL; 704 } 705 } 706 if (!n) { 707 m_freem(m); 708 return ENOBUFS; 709 } 710 711 MCLAIM(n, &tcp_tx_mowner); 712 n->m_data += max_linkhdr; 713 n->m_len = hlen + tlen; 714 m_copyback(n, 0, hlen, mtod(m, caddr_t)); 715 m_copyback(n, hlen, tlen, (caddr_t)th0); 716 717 m_freem(m); 718 m = n; 719 n = NULL; 720 } 721 722 #define xchg(a,b,type) { type t; t=a; a=b; b=t; } 723 switch (family) { 724 case AF_INET: 725 ip = mtod(m, struct ip *); 726 th = (struct tcphdr *)(ip + 1); 727 ip->ip_p = IPPROTO_TCP; 728 xchg(ip->ip_dst, ip->ip_src, struct in_addr); 729 ip->ip_p = IPPROTO_TCP; 730 break; 731 #ifdef INET6 732 case AF_INET6: 733 ip6 = mtod(m, struct ip6_hdr *); 734 th = (struct tcphdr *)(ip6 + 1); 735 ip6->ip6_nxt = IPPROTO_TCP; 736 xchg(ip6->ip6_dst, ip6->ip6_src, struct in6_addr); 737 ip6->ip6_nxt = IPPROTO_TCP; 738 break; 739 #endif 740 #if 0 741 default: 742 /* noone will visit here */ 743 m_freem(m); 744 return EAFNOSUPPORT; 745 #endif 746 } 747 xchg(th->th_dport, th->th_sport, u_int16_t); 748 #undef xchg 749 tlen = 0; /*be friendly with the following code*/ 750 } 751 th->th_seq = htonl(seq); 752 th->th_ack = htonl(ack); 753 th->th_x2 = 0; 754 if ((flags & TH_SYN) == 0) { 755 if (tp) 756 win >>= tp->rcv_scale; 757 if (win > TCP_MAXWIN) 758 win = TCP_MAXWIN; 759 th->th_win = htons((u_int16_t)win); 760 th->th_off = sizeof (struct tcphdr) >> 2; 761 tlen += sizeof(*th); 762 } else 763 tlen += th->th_off << 2; 764 m->m_len = hlen + tlen; 765 m->m_pkthdr.len = hlen + tlen; 766 m->m_pkthdr.rcvif = (struct ifnet *) 0; 767 th->th_flags = flags; 768 th->th_urp = 0; 769 770 switch (family) { 771 #ifdef INET 772 case AF_INET: 773 { 774 struct ipovly *ipov = (struct ipovly *)ip; 775 bzero(ipov->ih_x1, sizeof ipov->ih_x1); 776 ipov->ih_len = htons((u_int16_t)tlen); 777 778 th->th_sum = 0; 779 th->th_sum = in_cksum(m, hlen + tlen); 780 ip->ip_len = htons(hlen + tlen); 781 ip->ip_ttl = ip_defttl; 782 break; 783 } 784 #endif 785 #ifdef INET6 786 case AF_INET6: 787 { 788 th->th_sum = 0; 789 th->th_sum = in6_cksum(m, IPPROTO_TCP, sizeof(struct ip6_hdr), 790 tlen); 791 ip6->ip6_plen = ntohs(tlen); 792 if (tp && tp->t_in6pcb) { 793 struct ifnet *oifp; 794 ro = (struct route *)&tp->t_in6pcb->in6p_route; 795 oifp = ro->ro_rt ? ro->ro_rt->rt_ifp : NULL; 796 ip6->ip6_hlim = in6_selecthlim(tp->t_in6pcb, oifp); 797 } else 798 ip6->ip6_hlim = ip6_defhlim; 799 ip6->ip6_flow &= ~IPV6_FLOWINFO_MASK; 800 if (ip6_auto_flowlabel) { 801 ip6->ip6_flow |= 802 (htonl(ip6_flow_seq++) & IPV6_FLOWLABEL_MASK); 803 } 804 break; 805 } 806 #endif 807 } 808 809 #ifdef IPSEC 810 (void)ipsec_setsocket(m, NULL); 811 #endif /*IPSEC*/ 812 813 if (tp != NULL && tp->t_inpcb != NULL) { 814 ro = &tp->t_inpcb->inp_route; 815 #ifdef IPSEC 816 if (ipsec_setsocket(m, tp->t_inpcb->inp_socket) != 0) { 817 m_freem(m); 818 return ENOBUFS; 819 } 820 #endif 821 #ifdef DIAGNOSTIC 822 if (family != AF_INET) 823 panic("tcp_respond: address family mismatch"); 824 if (!in_hosteq(ip->ip_dst, tp->t_inpcb->inp_faddr)) { 825 panic("tcp_respond: ip_dst %x != inp_faddr %x", 826 ntohl(ip->ip_dst.s_addr), 827 ntohl(tp->t_inpcb->inp_faddr.s_addr)); 828 } 829 #endif 830 } 831 #ifdef INET6 832 else if (tp != NULL && tp->t_in6pcb != NULL) { 833 ro = (struct route *)&tp->t_in6pcb->in6p_route; 834 #ifdef IPSEC 835 if (ipsec_setsocket(m, tp->t_in6pcb->in6p_socket) != 0) { 836 m_freem(m); 837 return ENOBUFS; 838 } 839 #endif 840 #ifdef DIAGNOSTIC 841 if (family == AF_INET) { 842 if (!IN6_IS_ADDR_V4MAPPED(&tp->t_in6pcb->in6p_faddr)) 843 panic("tcp_respond: not mapped addr"); 844 if (bcmp(&ip->ip_dst, 845 &tp->t_in6pcb->in6p_faddr.s6_addr32[3], 846 sizeof(ip->ip_dst)) != 0) { 847 panic("tcp_respond: ip_dst != in6p_faddr"); 848 } 849 } else if (family == AF_INET6) { 850 if (!IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst, 851 &tp->t_in6pcb->in6p_faddr)) 852 panic("tcp_respond: ip6_dst != in6p_faddr"); 853 } else 854 panic("tcp_respond: address family mismatch"); 855 #endif 856 } 857 #endif 858 else 859 ro = NULL; 860 861 switch (family) { 862 #ifdef INET 863 case AF_INET: 864 error = ip_output(m, NULL, ro, 865 (tp && tp->t_mtudisc ? IP_MTUDISC : 0), 866 NULL); 867 break; 868 #endif 869 #ifdef INET6 870 case AF_INET6: 871 error = ip6_output(m, NULL, (struct route_in6 *)ro, 0, NULL, 872 NULL); 873 break; 874 #endif 875 default: 876 error = EAFNOSUPPORT; 877 break; 878 } 879 880 return (error); 881 } 882 883 /* 884 * Create a new TCP control block, making an 885 * empty reassembly queue and hooking it to the argument 886 * protocol control block. 887 */ 888 struct tcpcb * 889 tcp_newtcpcb(family, aux) 890 int family; /* selects inpcb, or in6pcb */ 891 void *aux; 892 { 893 struct tcpcb *tp; 894 int i; 895 896 switch (family) { 897 case PF_INET: 898 break; 899 #ifdef INET6 900 case PF_INET6: 901 break; 902 #endif 903 default: 904 return NULL; 905 } 906 907 tp = pool_get(&tcpcb_pool, PR_NOWAIT); 908 if (tp == NULL) 909 return (NULL); 910 bzero((caddr_t)tp, sizeof(struct tcpcb)); 911 TAILQ_INIT(&tp->segq); 912 TAILQ_INIT(&tp->timeq); 913 tp->t_family = family; /* may be overridden later on */ 914 tp->t_peermss = tcp_mssdflt; 915 tp->t_ourmss = tcp_mssdflt; 916 tp->t_segsz = tcp_mssdflt; 917 LIST_INIT(&tp->t_sc); 918 919 callout_init(&tp->t_delack_ch); 920 for (i = 0; i < TCPT_NTIMERS; i++) 921 TCP_TIMER_INIT(tp, i); 922 923 tp->t_flags = 0; 924 if (tcp_do_rfc1323 && tcp_do_win_scale) 925 tp->t_flags |= TF_REQ_SCALE; 926 if (tcp_do_rfc1323 && tcp_do_timestamps) 927 tp->t_flags |= TF_REQ_TSTMP; 928 if (tcp_do_sack == 2) 929 tp->t_flags |= TF_WILL_SACK; 930 else if (tcp_do_sack == 1) 931 tp->t_flags |= TF_WILL_SACK|TF_IGNR_RXSACK; 932 tp->t_flags |= TF_CANT_TXSACK; 933 switch (family) { 934 case PF_INET: 935 tp->t_inpcb = (struct inpcb *)aux; 936 tp->t_mtudisc = ip_mtudisc; 937 break; 938 #ifdef INET6 939 case PF_INET6: 940 tp->t_in6pcb = (struct in6pcb *)aux; 941 /* for IPv6, always try to run path MTU discovery */ 942 tp->t_mtudisc = 1; 943 break; 944 #endif 945 } 946 /* 947 * Init srtt to TCPTV_SRTTBASE (0), so we can tell that we have no 948 * rtt estimate. Set rttvar so that srtt + 2 * rttvar gives 949 * reasonable initial retransmit time. 950 */ 951 tp->t_srtt = TCPTV_SRTTBASE; 952 tp->t_rttvar = tcp_rttdflt * PR_SLOWHZ << (TCP_RTTVAR_SHIFT + 2 - 1); 953 tp->t_rttmin = TCPTV_MIN; 954 TCPT_RANGESET(tp->t_rxtcur, TCP_REXMTVAL(tp), 955 TCPTV_MIN, TCPTV_REXMTMAX); 956 tp->snd_cwnd = TCP_MAXWIN << TCP_MAX_WINSHIFT; 957 tp->snd_ssthresh = TCP_MAXWIN << TCP_MAX_WINSHIFT; 958 if (family == AF_INET) { 959 struct inpcb *inp = (struct inpcb *)aux; 960 inp->inp_ip.ip_ttl = ip_defttl; 961 inp->inp_ppcb = (caddr_t)tp; 962 } 963 #ifdef INET6 964 else if (family == AF_INET6) { 965 struct in6pcb *in6p = (struct in6pcb *)aux; 966 in6p->in6p_ip6.ip6_hlim = in6_selecthlim(in6p, 967 in6p->in6p_route.ro_rt ? in6p->in6p_route.ro_rt->rt_ifp 968 : NULL); 969 in6p->in6p_ppcb = (caddr_t)tp; 970 } 971 #endif 972 973 /* 974 * Initialize our timebase. When we send timestamps, we take 975 * the delta from tcp_now -- this means each connection always 976 * gets a timebase of 0, which makes it, among other things, 977 * more difficult to determine how long a system has been up, 978 * and thus how many TCP sequence increments have occurred. 979 */ 980 tp->ts_timebase = tcp_now; 981 982 return (tp); 983 } 984 985 /* 986 * Drop a TCP connection, reporting 987 * the specified error. If connection is synchronized, 988 * then send a RST to peer. 989 */ 990 struct tcpcb * 991 tcp_drop(tp, errno) 992 struct tcpcb *tp; 993 int errno; 994 { 995 struct socket *so = NULL; 996 997 #ifdef DIAGNOSTIC 998 if (tp->t_inpcb && tp->t_in6pcb) 999 panic("tcp_drop: both t_inpcb and t_in6pcb are set"); 1000 #endif 1001 #ifdef INET 1002 if (tp->t_inpcb) 1003 so = tp->t_inpcb->inp_socket; 1004 #endif 1005 #ifdef INET6 1006 if (tp->t_in6pcb) 1007 so = tp->t_in6pcb->in6p_socket; 1008 #endif 1009 if (!so) 1010 return NULL; 1011 1012 if (TCPS_HAVERCVDSYN(tp->t_state)) { 1013 tp->t_state = TCPS_CLOSED; 1014 (void) tcp_output(tp); 1015 tcpstat.tcps_drops++; 1016 } else 1017 tcpstat.tcps_conndrops++; 1018 if (errno == ETIMEDOUT && tp->t_softerror) 1019 errno = tp->t_softerror; 1020 so->so_error = errno; 1021 return (tcp_close(tp)); 1022 } 1023 1024 /* 1025 * Close a TCP control block: 1026 * discard all space held by the tcp 1027 * discard internet protocol block 1028 * wake up any sleepers 1029 */ 1030 struct tcpcb * 1031 tcp_close(tp) 1032 struct tcpcb *tp; 1033 { 1034 struct inpcb *inp; 1035 #ifdef INET6 1036 struct in6pcb *in6p; 1037 #endif 1038 struct socket *so; 1039 #ifdef RTV_RTT 1040 struct rtentry *rt; 1041 #endif 1042 struct route *ro; 1043 1044 inp = tp->t_inpcb; 1045 #ifdef INET6 1046 in6p = tp->t_in6pcb; 1047 #endif 1048 so = NULL; 1049 ro = NULL; 1050 if (inp) { 1051 so = inp->inp_socket; 1052 ro = &inp->inp_route; 1053 } 1054 #ifdef INET6 1055 else if (in6p) { 1056 so = in6p->in6p_socket; 1057 ro = (struct route *)&in6p->in6p_route; 1058 } 1059 #endif 1060 1061 #ifdef RTV_RTT 1062 /* 1063 * If we sent enough data to get some meaningful characteristics, 1064 * save them in the routing entry. 'Enough' is arbitrarily 1065 * defined as the sendpipesize (default 4K) * 16. This would 1066 * give us 16 rtt samples assuming we only get one sample per 1067 * window (the usual case on a long haul net). 16 samples is 1068 * enough for the srtt filter to converge to within 5% of the correct 1069 * value; fewer samples and we could save a very bogus rtt. 1070 * 1071 * Don't update the default route's characteristics and don't 1072 * update anything that the user "locked". 1073 */ 1074 if (SEQ_LT(tp->iss + so->so_snd.sb_hiwat * 16, tp->snd_max) && 1075 ro && (rt = ro->ro_rt) && 1076 !in_nullhost(satosin(rt_key(rt))->sin_addr)) { 1077 u_long i = 0; 1078 1079 if ((rt->rt_rmx.rmx_locks & RTV_RTT) == 0) { 1080 i = tp->t_srtt * 1081 ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTT_SHIFT + 2)); 1082 if (rt->rt_rmx.rmx_rtt && i) 1083 /* 1084 * filter this update to half the old & half 1085 * the new values, converting scale. 1086 * See route.h and tcp_var.h for a 1087 * description of the scaling constants. 1088 */ 1089 rt->rt_rmx.rmx_rtt = 1090 (rt->rt_rmx.rmx_rtt + i) / 2; 1091 else 1092 rt->rt_rmx.rmx_rtt = i; 1093 } 1094 if ((rt->rt_rmx.rmx_locks & RTV_RTTVAR) == 0) { 1095 i = tp->t_rttvar * 1096 ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTTVAR_SHIFT + 2)); 1097 if (rt->rt_rmx.rmx_rttvar && i) 1098 rt->rt_rmx.rmx_rttvar = 1099 (rt->rt_rmx.rmx_rttvar + i) / 2; 1100 else 1101 rt->rt_rmx.rmx_rttvar = i; 1102 } 1103 /* 1104 * update the pipelimit (ssthresh) if it has been updated 1105 * already or if a pipesize was specified & the threshhold 1106 * got below half the pipesize. I.e., wait for bad news 1107 * before we start updating, then update on both good 1108 * and bad news. 1109 */ 1110 if (((rt->rt_rmx.rmx_locks & RTV_SSTHRESH) == 0 && 1111 (i = tp->snd_ssthresh) && rt->rt_rmx.rmx_ssthresh) || 1112 i < (rt->rt_rmx.rmx_sendpipe / 2)) { 1113 /* 1114 * convert the limit from user data bytes to 1115 * packets then to packet data bytes. 1116 */ 1117 i = (i + tp->t_segsz / 2) / tp->t_segsz; 1118 if (i < 2) 1119 i = 2; 1120 i *= (u_long)(tp->t_segsz + sizeof (struct tcpiphdr)); 1121 if (rt->rt_rmx.rmx_ssthresh) 1122 rt->rt_rmx.rmx_ssthresh = 1123 (rt->rt_rmx.rmx_ssthresh + i) / 2; 1124 else 1125 rt->rt_rmx.rmx_ssthresh = i; 1126 } 1127 } 1128 #endif /* RTV_RTT */ 1129 /* free the reassembly queue, if any */ 1130 TCP_REASS_LOCK(tp); 1131 (void) tcp_freeq(tp); 1132 TCP_REASS_UNLOCK(tp); 1133 1134 tcp_canceltimers(tp); 1135 TCP_CLEAR_DELACK(tp); 1136 syn_cache_cleanup(tp); 1137 1138 if (tp->t_template) { 1139 m_free(tp->t_template); 1140 tp->t_template = NULL; 1141 } 1142 pool_put(&tcpcb_pool, tp); 1143 if (inp) { 1144 inp->inp_ppcb = 0; 1145 soisdisconnected(so); 1146 in_pcbdetach(inp); 1147 } 1148 #ifdef INET6 1149 else if (in6p) { 1150 in6p->in6p_ppcb = 0; 1151 soisdisconnected(so); 1152 in6_pcbdetach(in6p); 1153 } 1154 #endif 1155 tcpstat.tcps_closed++; 1156 return ((struct tcpcb *)0); 1157 } 1158 1159 int 1160 tcp_freeq(tp) 1161 struct tcpcb *tp; 1162 { 1163 struct ipqent *qe; 1164 int rv = 0; 1165 #ifdef TCPREASS_DEBUG 1166 int i = 0; 1167 #endif 1168 1169 TCP_REASS_LOCK_CHECK(tp); 1170 1171 while ((qe = TAILQ_FIRST(&tp->segq)) != NULL) { 1172 #ifdef TCPREASS_DEBUG 1173 printf("tcp_freeq[%p,%d]: %u:%u(%u) 0x%02x\n", 1174 tp, i++, qe->ipqe_seq, qe->ipqe_seq + qe->ipqe_len, 1175 qe->ipqe_len, qe->ipqe_flags & (TH_SYN|TH_FIN|TH_RST)); 1176 #endif 1177 TAILQ_REMOVE(&tp->segq, qe, ipqe_q); 1178 TAILQ_REMOVE(&tp->timeq, qe, ipqe_timeq); 1179 m_freem(qe->ipqe_m); 1180 pool_put(&ipqent_pool, qe); 1181 rv = 1; 1182 } 1183 return (rv); 1184 } 1185 1186 /* 1187 * Protocol drain routine. Called when memory is in short supply. 1188 */ 1189 void 1190 tcp_drain() 1191 { 1192 struct inpcb *inp; 1193 struct tcpcb *tp; 1194 1195 /* 1196 * Free the sequence queue of all TCP connections. 1197 */ 1198 inp = CIRCLEQ_FIRST(&tcbtable.inpt_queue); 1199 if (inp) /* XXX */ 1200 CIRCLEQ_FOREACH(inp, &tcbtable.inpt_queue, inp_queue) { 1201 if ((tp = intotcpcb(inp)) != NULL) { 1202 /* 1203 * We may be called from a device's interrupt 1204 * context. If the tcpcb is already busy, 1205 * just bail out now. 1206 */ 1207 if (tcp_reass_lock_try(tp) == 0) 1208 continue; 1209 if (tcp_freeq(tp)) 1210 tcpstat.tcps_connsdrained++; 1211 TCP_REASS_UNLOCK(tp); 1212 } 1213 } 1214 } 1215 1216 #ifdef INET6 1217 void 1218 tcp6_drain() 1219 { 1220 struct in6pcb *in6p; 1221 struct tcpcb *tp; 1222 struct in6pcb *head = &tcb6; 1223 1224 /* 1225 * Free the sequence queue of all TCP connections. 1226 */ 1227 for (in6p = head->in6p_next; in6p != head; in6p = in6p->in6p_next) { 1228 if ((tp = in6totcpcb(in6p)) != NULL) { 1229 /* 1230 * We may be called from a device's interrupt 1231 * context. If the tcpcb is already busy, 1232 * just bail out now. 1233 */ 1234 if (tcp_reass_lock_try(tp) == 0) 1235 continue; 1236 if (tcp_freeq(tp)) 1237 tcpstat.tcps_connsdrained++; 1238 TCP_REASS_UNLOCK(tp); 1239 } 1240 } 1241 } 1242 #endif 1243 1244 /* 1245 * Notify a tcp user of an asynchronous error; 1246 * store error as soft error, but wake up user 1247 * (for now, won't do anything until can select for soft error). 1248 */ 1249 void 1250 tcp_notify(inp, error) 1251 struct inpcb *inp; 1252 int error; 1253 { 1254 struct tcpcb *tp = (struct tcpcb *)inp->inp_ppcb; 1255 struct socket *so = inp->inp_socket; 1256 1257 /* 1258 * Ignore some errors if we are hooked up. 1259 * If connection hasn't completed, has retransmitted several times, 1260 * and receives a second error, give up now. This is better 1261 * than waiting a long time to establish a connection that 1262 * can never complete. 1263 */ 1264 if (tp->t_state == TCPS_ESTABLISHED && 1265 (error == EHOSTUNREACH || error == ENETUNREACH || 1266 error == EHOSTDOWN)) { 1267 return; 1268 } else if (TCPS_HAVEESTABLISHED(tp->t_state) == 0 && 1269 tp->t_rxtshift > 3 && tp->t_softerror) 1270 so->so_error = error; 1271 else 1272 tp->t_softerror = error; 1273 wakeup((caddr_t) &so->so_timeo); 1274 sorwakeup(so); 1275 sowwakeup(so); 1276 } 1277 1278 #ifdef INET6 1279 void 1280 tcp6_notify(in6p, error) 1281 struct in6pcb *in6p; 1282 int error; 1283 { 1284 struct tcpcb *tp = (struct tcpcb *)in6p->in6p_ppcb; 1285 struct socket *so = in6p->in6p_socket; 1286 1287 /* 1288 * Ignore some errors if we are hooked up. 1289 * If connection hasn't completed, has retransmitted several times, 1290 * and receives a second error, give up now. This is better 1291 * than waiting a long time to establish a connection that 1292 * can never complete. 1293 */ 1294 if (tp->t_state == TCPS_ESTABLISHED && 1295 (error == EHOSTUNREACH || error == ENETUNREACH || 1296 error == EHOSTDOWN)) { 1297 return; 1298 } else if (TCPS_HAVEESTABLISHED(tp->t_state) == 0 && 1299 tp->t_rxtshift > 3 && tp->t_softerror) 1300 so->so_error = error; 1301 else 1302 tp->t_softerror = error; 1303 wakeup((caddr_t) &so->so_timeo); 1304 sorwakeup(so); 1305 sowwakeup(so); 1306 } 1307 #endif 1308 1309 #ifdef INET6 1310 void 1311 tcp6_ctlinput(cmd, sa, d) 1312 int cmd; 1313 struct sockaddr *sa; 1314 void *d; 1315 { 1316 struct tcphdr th; 1317 void (*notify) __P((struct in6pcb *, int)) = tcp6_notify; 1318 int nmatch; 1319 struct ip6_hdr *ip6; 1320 const struct sockaddr_in6 *sa6_src = NULL; 1321 struct sockaddr_in6 *sa6 = (struct sockaddr_in6 *)sa; 1322 struct mbuf *m; 1323 int off; 1324 1325 if (sa->sa_family != AF_INET6 || 1326 sa->sa_len != sizeof(struct sockaddr_in6)) 1327 return; 1328 if ((unsigned)cmd >= PRC_NCMDS) 1329 return; 1330 else if (cmd == PRC_QUENCH) { 1331 /* XXX there's no PRC_QUENCH in IPv6 */ 1332 notify = tcp6_quench; 1333 } else if (PRC_IS_REDIRECT(cmd)) 1334 notify = in6_rtchange, d = NULL; 1335 else if (cmd == PRC_MSGSIZE) 1336 ; /* special code is present, see below */ 1337 else if (cmd == PRC_HOSTDEAD) 1338 d = NULL; 1339 else if (inet6ctlerrmap[cmd] == 0) 1340 return; 1341 1342 /* if the parameter is from icmp6, decode it. */ 1343 if (d != NULL) { 1344 struct ip6ctlparam *ip6cp = (struct ip6ctlparam *)d; 1345 m = ip6cp->ip6c_m; 1346 ip6 = ip6cp->ip6c_ip6; 1347 off = ip6cp->ip6c_off; 1348 sa6_src = ip6cp->ip6c_src; 1349 } else { 1350 m = NULL; 1351 ip6 = NULL; 1352 sa6_src = &sa6_any; 1353 } 1354 1355 if (ip6) { 1356 /* 1357 * XXX: We assume that when ip6 is non NULL, 1358 * M and OFF are valid. 1359 */ 1360 1361 /* check if we can safely examine src and dst ports */ 1362 if (m->m_pkthdr.len < off + sizeof(th)) { 1363 if (cmd == PRC_MSGSIZE) 1364 icmp6_mtudisc_update((struct ip6ctlparam *)d, 0); 1365 return; 1366 } 1367 1368 bzero(&th, sizeof(th)); 1369 m_copydata(m, off, sizeof(th), (caddr_t)&th); 1370 1371 if (cmd == PRC_MSGSIZE) { 1372 int valid = 0; 1373 1374 /* 1375 * Check to see if we have a valid TCP connection 1376 * corresponding to the address in the ICMPv6 message 1377 * payload. 1378 */ 1379 if (in6_pcblookup_connect(&tcb6, &sa6->sin6_addr, 1380 th.th_dport, (struct in6_addr *)&sa6_src->sin6_addr, 1381 th.th_sport, 0)) 1382 valid++; 1383 1384 /* 1385 * Depending on the value of "valid" and routing table 1386 * size (mtudisc_{hi,lo}wat), we will: 1387 * - recalcurate the new MTU and create the 1388 * corresponding routing entry, or 1389 * - ignore the MTU change notification. 1390 */ 1391 icmp6_mtudisc_update((struct ip6ctlparam *)d, valid); 1392 1393 /* 1394 * no need to call in6_pcbnotify, it should have been 1395 * called via callback if necessary 1396 */ 1397 return; 1398 } 1399 1400 nmatch = in6_pcbnotify(&tcb6, sa, th.th_dport, 1401 (struct sockaddr *)sa6_src, th.th_sport, cmd, NULL, notify); 1402 if (nmatch == 0 && syn_cache_count && 1403 (inet6ctlerrmap[cmd] == EHOSTUNREACH || 1404 inet6ctlerrmap[cmd] == ENETUNREACH || 1405 inet6ctlerrmap[cmd] == EHOSTDOWN)) 1406 syn_cache_unreach((struct sockaddr *)sa6_src, 1407 sa, &th); 1408 } else { 1409 (void) in6_pcbnotify(&tcb6, sa, 0, (struct sockaddr *)sa6_src, 1410 0, cmd, NULL, notify); 1411 } 1412 } 1413 #endif 1414 1415 #ifdef INET 1416 /* assumes that ip header and tcp header are contiguous on mbuf */ 1417 void * 1418 tcp_ctlinput(cmd, sa, v) 1419 int cmd; 1420 struct sockaddr *sa; 1421 void *v; 1422 { 1423 struct ip *ip = v; 1424 struct tcphdr *th; 1425 struct icmp *icp; 1426 extern const int inetctlerrmap[]; 1427 void (*notify) __P((struct inpcb *, int)) = tcp_notify; 1428 int errno; 1429 int nmatch; 1430 #ifdef INET6 1431 struct in6_addr src6, dst6; 1432 #endif 1433 1434 if (sa->sa_family != AF_INET || 1435 sa->sa_len != sizeof(struct sockaddr_in)) 1436 return NULL; 1437 if ((unsigned)cmd >= PRC_NCMDS) 1438 return NULL; 1439 errno = inetctlerrmap[cmd]; 1440 if (cmd == PRC_QUENCH) 1441 notify = tcp_quench; 1442 else if (PRC_IS_REDIRECT(cmd)) 1443 notify = in_rtchange, ip = 0; 1444 else if (cmd == PRC_MSGSIZE && ip && ip->ip_v == 4) { 1445 /* 1446 * Check to see if we have a valid TCP connection 1447 * corresponding to the address in the ICMP message 1448 * payload. 1449 * 1450 * Boundary check is made in icmp_input(), with ICMP_ADVLENMIN. 1451 */ 1452 th = (struct tcphdr *)((caddr_t)ip + (ip->ip_hl << 2)); 1453 #ifdef INET6 1454 memset(&src6, 0, sizeof(src6)); 1455 memset(&dst6, 0, sizeof(dst6)); 1456 src6.s6_addr16[5] = dst6.s6_addr16[5] = 0xffff; 1457 memcpy(&src6.s6_addr32[3], &ip->ip_src, sizeof(struct in_addr)); 1458 memcpy(&dst6.s6_addr32[3], &ip->ip_dst, sizeof(struct in_addr)); 1459 #endif 1460 if (in_pcblookup_connect(&tcbtable, ip->ip_dst, th->th_dport, 1461 ip->ip_src, th->th_sport) != NULL) 1462 ; 1463 #ifdef INET6 1464 else if (in6_pcblookup_connect(&tcb6, &dst6, 1465 th->th_dport, &src6, th->th_sport, 0) != NULL) 1466 ; 1467 #endif 1468 else 1469 return NULL; 1470 1471 /* 1472 * Now that we've validated that we are actually communicating 1473 * with the host indicated in the ICMP message, locate the 1474 * ICMP header, recalculate the new MTU, and create the 1475 * corresponding routing entry. 1476 */ 1477 icp = (struct icmp *)((caddr_t)ip - 1478 offsetof(struct icmp, icmp_ip)); 1479 icmp_mtudisc(icp, ip->ip_dst); 1480 1481 return NULL; 1482 } else if (cmd == PRC_HOSTDEAD) 1483 ip = 0; 1484 else if (errno == 0) 1485 return NULL; 1486 if (ip && ip->ip_v == 4 && sa->sa_family == AF_INET) { 1487 th = (struct tcphdr *)((caddr_t)ip + (ip->ip_hl << 2)); 1488 nmatch = in_pcbnotify(&tcbtable, satosin(sa)->sin_addr, 1489 th->th_dport, ip->ip_src, th->th_sport, errno, notify); 1490 if (nmatch == 0 && syn_cache_count && 1491 (inetctlerrmap[cmd] == EHOSTUNREACH || 1492 inetctlerrmap[cmd] == ENETUNREACH || 1493 inetctlerrmap[cmd] == EHOSTDOWN)) { 1494 struct sockaddr_in sin; 1495 bzero(&sin, sizeof(sin)); 1496 sin.sin_len = sizeof(sin); 1497 sin.sin_family = AF_INET; 1498 sin.sin_port = th->th_sport; 1499 sin.sin_addr = ip->ip_src; 1500 syn_cache_unreach((struct sockaddr *)&sin, sa, th); 1501 } 1502 1503 /* XXX mapped address case */ 1504 } else 1505 in_pcbnotifyall(&tcbtable, satosin(sa)->sin_addr, errno, 1506 notify); 1507 return NULL; 1508 } 1509 1510 /* 1511 * When a source quence is received, we are being notifed of congestion. 1512 * Close the congestion window down to the Loss Window (one segment). 1513 * We will gradually open it again as we proceed. 1514 */ 1515 void 1516 tcp_quench(inp, errno) 1517 struct inpcb *inp; 1518 int errno; 1519 { 1520 struct tcpcb *tp = intotcpcb(inp); 1521 1522 if (tp) 1523 tp->snd_cwnd = tp->t_segsz; 1524 } 1525 #endif 1526 1527 #ifdef INET6 1528 void 1529 tcp6_quench(in6p, errno) 1530 struct in6pcb *in6p; 1531 int errno; 1532 { 1533 struct tcpcb *tp = in6totcpcb(in6p); 1534 1535 if (tp) 1536 tp->snd_cwnd = tp->t_segsz; 1537 } 1538 #endif 1539 1540 #ifdef INET 1541 /* 1542 * Path MTU Discovery handlers. 1543 */ 1544 void 1545 tcp_mtudisc_callback(faddr) 1546 struct in_addr faddr; 1547 { 1548 #ifdef INET6 1549 struct in6_addr in6; 1550 #endif 1551 1552 in_pcbnotifyall(&tcbtable, faddr, EMSGSIZE, tcp_mtudisc); 1553 #ifdef INET6 1554 memset(&in6, 0, sizeof(in6)); 1555 in6.s6_addr16[5] = 0xffff; 1556 memcpy(&in6.s6_addr32[3], &faddr, sizeof(struct in_addr)); 1557 tcp6_mtudisc_callback(&in6); 1558 #endif 1559 } 1560 1561 /* 1562 * On receipt of path MTU corrections, flush old route and replace it 1563 * with the new one. Retransmit all unacknowledged packets, to ensure 1564 * that all packets will be received. 1565 */ 1566 void 1567 tcp_mtudisc(inp, errno) 1568 struct inpcb *inp; 1569 int errno; 1570 { 1571 struct tcpcb *tp = intotcpcb(inp); 1572 struct rtentry *rt = in_pcbrtentry(inp); 1573 1574 if (tp != 0) { 1575 if (rt != 0) { 1576 /* 1577 * If this was not a host route, remove and realloc. 1578 */ 1579 if ((rt->rt_flags & RTF_HOST) == 0) { 1580 in_rtchange(inp, errno); 1581 if ((rt = in_pcbrtentry(inp)) == 0) 1582 return; 1583 } 1584 1585 /* 1586 * Slow start out of the error condition. We 1587 * use the MTU because we know it's smaller 1588 * than the previously transmitted segment. 1589 * 1590 * Note: This is more conservative than the 1591 * suggestion in draft-floyd-incr-init-win-03. 1592 */ 1593 if (rt->rt_rmx.rmx_mtu != 0) 1594 tp->snd_cwnd = 1595 TCP_INITIAL_WINDOW(tcp_init_win, 1596 rt->rt_rmx.rmx_mtu); 1597 } 1598 1599 /* 1600 * Resend unacknowledged packets. 1601 */ 1602 tp->snd_nxt = tp->snd_una; 1603 tcp_output(tp); 1604 } 1605 } 1606 #endif 1607 1608 #ifdef INET6 1609 /* 1610 * Path MTU Discovery handlers. 1611 */ 1612 void 1613 tcp6_mtudisc_callback(faddr) 1614 struct in6_addr *faddr; 1615 { 1616 struct sockaddr_in6 sin6; 1617 1618 bzero(&sin6, sizeof(sin6)); 1619 sin6.sin6_family = AF_INET6; 1620 sin6.sin6_len = sizeof(struct sockaddr_in6); 1621 sin6.sin6_addr = *faddr; 1622 (void) in6_pcbnotify(&tcb6, (struct sockaddr *)&sin6, 0, 1623 (struct sockaddr *)&sa6_any, 0, PRC_MSGSIZE, NULL, tcp6_mtudisc); 1624 } 1625 1626 void 1627 tcp6_mtudisc(in6p, errno) 1628 struct in6pcb *in6p; 1629 int errno; 1630 { 1631 struct tcpcb *tp = in6totcpcb(in6p); 1632 struct rtentry *rt = in6_pcbrtentry(in6p); 1633 1634 if (tp != 0) { 1635 if (rt != 0) { 1636 /* 1637 * If this was not a host route, remove and realloc. 1638 */ 1639 if ((rt->rt_flags & RTF_HOST) == 0) { 1640 in6_rtchange(in6p, errno); 1641 if ((rt = in6_pcbrtentry(in6p)) == 0) 1642 return; 1643 } 1644 1645 /* 1646 * Slow start out of the error condition. We 1647 * use the MTU because we know it's smaller 1648 * than the previously transmitted segment. 1649 * 1650 * Note: This is more conservative than the 1651 * suggestion in draft-floyd-incr-init-win-03. 1652 */ 1653 if (rt->rt_rmx.rmx_mtu != 0) 1654 tp->snd_cwnd = 1655 TCP_INITIAL_WINDOW(tcp_init_win, 1656 rt->rt_rmx.rmx_mtu); 1657 } 1658 1659 /* 1660 * Resend unacknowledged packets. 1661 */ 1662 tp->snd_nxt = tp->snd_una; 1663 tcp_output(tp); 1664 } 1665 } 1666 #endif /* INET6 */ 1667 1668 /* 1669 * Compute the MSS to advertise to the peer. Called only during 1670 * the 3-way handshake. If we are the server (peer initiated 1671 * connection), we are called with a pointer to the interface 1672 * on which the SYN packet arrived. If we are the client (we 1673 * initiated connection), we are called with a pointer to the 1674 * interface out which this connection should go. 1675 * 1676 * NOTE: Do not subtract IP option/extension header size nor IPsec 1677 * header size from MSS advertisement. MSS option must hold the maximum 1678 * segment size we can accept, so it must always be: 1679 * max(if mtu) - ip header - tcp header 1680 */ 1681 u_long 1682 tcp_mss_to_advertise(ifp, af) 1683 const struct ifnet *ifp; 1684 int af; 1685 { 1686 extern u_long in_maxmtu; 1687 u_long mss = 0; 1688 u_long hdrsiz; 1689 1690 /* 1691 * In order to avoid defeating path MTU discovery on the peer, 1692 * we advertise the max MTU of all attached networks as our MSS, 1693 * per RFC 1191, section 3.1. 1694 * 1695 * We provide the option to advertise just the MTU of 1696 * the interface on which we hope this connection will 1697 * be receiving. If we are responding to a SYN, we 1698 * will have a pretty good idea about this, but when 1699 * initiating a connection there is a bit more doubt. 1700 * 1701 * We also need to ensure that loopback has a large enough 1702 * MSS, as the loopback MTU is never included in in_maxmtu. 1703 */ 1704 1705 if (ifp != NULL) 1706 switch (af) { 1707 case AF_INET: 1708 mss = ifp->if_mtu; 1709 break; 1710 #ifdef INET6 1711 case AF_INET6: 1712 mss = IN6_LINKMTU(ifp); 1713 break; 1714 #endif 1715 } 1716 1717 if (tcp_mss_ifmtu == 0) 1718 switch (af) { 1719 case AF_INET: 1720 mss = max(in_maxmtu, mss); 1721 break; 1722 #ifdef INET6 1723 case AF_INET6: 1724 mss = max(in6_maxmtu, mss); 1725 break; 1726 #endif 1727 } 1728 1729 switch (af) { 1730 case AF_INET: 1731 hdrsiz = sizeof(struct ip); 1732 break; 1733 #ifdef INET6 1734 case AF_INET6: 1735 hdrsiz = sizeof(struct ip6_hdr); 1736 break; 1737 #endif 1738 default: 1739 hdrsiz = 0; 1740 break; 1741 } 1742 hdrsiz += sizeof(struct tcphdr); 1743 if (mss > hdrsiz) 1744 mss -= hdrsiz; 1745 1746 mss = max(tcp_mssdflt, mss); 1747 return (mss); 1748 } 1749 1750 /* 1751 * Set connection variables based on the peer's advertised MSS. 1752 * We are passed the TCPCB for the actual connection. If we 1753 * are the server, we are called by the compressed state engine 1754 * when the 3-way handshake is complete. If we are the client, 1755 * we are called when we receive the SYN,ACK from the server. 1756 * 1757 * NOTE: Our advertised MSS value must be initialized in the TCPCB 1758 * before this routine is called! 1759 */ 1760 void 1761 tcp_mss_from_peer(tp, offer) 1762 struct tcpcb *tp; 1763 int offer; 1764 { 1765 struct socket *so; 1766 #if defined(RTV_SPIPE) || defined(RTV_SSTHRESH) 1767 struct rtentry *rt; 1768 #endif 1769 u_long bufsize; 1770 int mss; 1771 1772 #ifdef DIAGNOSTIC 1773 if (tp->t_inpcb && tp->t_in6pcb) 1774 panic("tcp_mss_from_peer: both t_inpcb and t_in6pcb are set"); 1775 #endif 1776 so = NULL; 1777 rt = NULL; 1778 #ifdef INET 1779 if (tp->t_inpcb) { 1780 so = tp->t_inpcb->inp_socket; 1781 #if defined(RTV_SPIPE) || defined(RTV_SSTHRESH) 1782 rt = in_pcbrtentry(tp->t_inpcb); 1783 #endif 1784 } 1785 #endif 1786 #ifdef INET6 1787 if (tp->t_in6pcb) { 1788 so = tp->t_in6pcb->in6p_socket; 1789 #if defined(RTV_SPIPE) || defined(RTV_SSTHRESH) 1790 rt = in6_pcbrtentry(tp->t_in6pcb); 1791 #endif 1792 } 1793 #endif 1794 1795 /* 1796 * As per RFC1122, use the default MSS value, unless they 1797 * sent us an offer. Do not accept offers less than 32 bytes. 1798 */ 1799 mss = tcp_mssdflt; 1800 if (offer) 1801 mss = offer; 1802 mss = max(mss, 32); /* sanity */ 1803 tp->t_peermss = mss; 1804 mss -= tcp_optlen(tp); 1805 #ifdef INET 1806 if (tp->t_inpcb) 1807 mss -= ip_optlen(tp->t_inpcb); 1808 #endif 1809 #ifdef INET6 1810 if (tp->t_in6pcb) 1811 mss -= ip6_optlen(tp->t_in6pcb); 1812 #endif 1813 1814 /* 1815 * If there's a pipesize, change the socket buffer to that size. 1816 * Make the socket buffer an integral number of MSS units. If 1817 * the MSS is larger than the socket buffer, artificially decrease 1818 * the MSS. 1819 */ 1820 #ifdef RTV_SPIPE 1821 if (rt != NULL && rt->rt_rmx.rmx_sendpipe != 0) 1822 bufsize = rt->rt_rmx.rmx_sendpipe; 1823 else 1824 #endif 1825 bufsize = so->so_snd.sb_hiwat; 1826 if (bufsize < mss) 1827 mss = bufsize; 1828 else { 1829 bufsize = roundup(bufsize, mss); 1830 if (bufsize > sb_max) 1831 bufsize = sb_max; 1832 (void) sbreserve(&so->so_snd, bufsize); 1833 } 1834 tp->t_segsz = mss; 1835 1836 #ifdef RTV_SSTHRESH 1837 if (rt != NULL && rt->rt_rmx.rmx_ssthresh) { 1838 /* 1839 * There's some sort of gateway or interface buffer 1840 * limit on the path. Use this to set the slow 1841 * start threshold, but set the threshold to no less 1842 * than 2 * MSS. 1843 */ 1844 tp->snd_ssthresh = max(2 * mss, rt->rt_rmx.rmx_ssthresh); 1845 } 1846 #endif 1847 } 1848 1849 /* 1850 * Processing necessary when a TCP connection is established. 1851 */ 1852 void 1853 tcp_established(tp) 1854 struct tcpcb *tp; 1855 { 1856 struct socket *so; 1857 #ifdef RTV_RPIPE 1858 struct rtentry *rt; 1859 #endif 1860 u_long bufsize; 1861 1862 #ifdef DIAGNOSTIC 1863 if (tp->t_inpcb && tp->t_in6pcb) 1864 panic("tcp_established: both t_inpcb and t_in6pcb are set"); 1865 #endif 1866 so = NULL; 1867 rt = NULL; 1868 #ifdef INET 1869 if (tp->t_inpcb) { 1870 so = tp->t_inpcb->inp_socket; 1871 #if defined(RTV_RPIPE) 1872 rt = in_pcbrtentry(tp->t_inpcb); 1873 #endif 1874 } 1875 #endif 1876 #ifdef INET6 1877 if (tp->t_in6pcb) { 1878 so = tp->t_in6pcb->in6p_socket; 1879 #if defined(RTV_RPIPE) 1880 rt = in6_pcbrtentry(tp->t_in6pcb); 1881 #endif 1882 } 1883 #endif 1884 1885 tp->t_state = TCPS_ESTABLISHED; 1886 TCP_TIMER_ARM(tp, TCPT_KEEP, tcp_keepidle); 1887 1888 #ifdef RTV_RPIPE 1889 if (rt != NULL && rt->rt_rmx.rmx_recvpipe != 0) 1890 bufsize = rt->rt_rmx.rmx_recvpipe; 1891 else 1892 #endif 1893 bufsize = so->so_rcv.sb_hiwat; 1894 if (bufsize > tp->t_ourmss) { 1895 bufsize = roundup(bufsize, tp->t_ourmss); 1896 if (bufsize > sb_max) 1897 bufsize = sb_max; 1898 (void) sbreserve(&so->so_rcv, bufsize); 1899 } 1900 } 1901 1902 /* 1903 * Check if there's an initial rtt or rttvar. Convert from the 1904 * route-table units to scaled multiples of the slow timeout timer. 1905 * Called only during the 3-way handshake. 1906 */ 1907 void 1908 tcp_rmx_rtt(tp) 1909 struct tcpcb *tp; 1910 { 1911 #ifdef RTV_RTT 1912 struct rtentry *rt = NULL; 1913 int rtt; 1914 1915 #ifdef DIAGNOSTIC 1916 if (tp->t_inpcb && tp->t_in6pcb) 1917 panic("tcp_rmx_rtt: both t_inpcb and t_in6pcb are set"); 1918 #endif 1919 #ifdef INET 1920 if (tp->t_inpcb) 1921 rt = in_pcbrtentry(tp->t_inpcb); 1922 #endif 1923 #ifdef INET6 1924 if (tp->t_in6pcb) 1925 rt = in6_pcbrtentry(tp->t_in6pcb); 1926 #endif 1927 if (rt == NULL) 1928 return; 1929 1930 if (tp->t_srtt == 0 && (rtt = rt->rt_rmx.rmx_rtt)) { 1931 /* 1932 * XXX The lock bit for MTU indicates that the value 1933 * is also a minimum value; this is subject to time. 1934 */ 1935 if (rt->rt_rmx.rmx_locks & RTV_RTT) 1936 TCPT_RANGESET(tp->t_rttmin, 1937 rtt / (RTM_RTTUNIT / PR_SLOWHZ), 1938 TCPTV_MIN, TCPTV_REXMTMAX); 1939 tp->t_srtt = rtt / 1940 ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTT_SHIFT + 2)); 1941 if (rt->rt_rmx.rmx_rttvar) { 1942 tp->t_rttvar = rt->rt_rmx.rmx_rttvar / 1943 ((RTM_RTTUNIT / PR_SLOWHZ) >> 1944 (TCP_RTTVAR_SHIFT + 2)); 1945 } else { 1946 /* Default variation is +- 1 rtt */ 1947 tp->t_rttvar = 1948 tp->t_srtt >> (TCP_RTT_SHIFT - TCP_RTTVAR_SHIFT); 1949 } 1950 TCPT_RANGESET(tp->t_rxtcur, 1951 ((tp->t_srtt >> 2) + tp->t_rttvar) >> (1 + 2), 1952 tp->t_rttmin, TCPTV_REXMTMAX); 1953 } 1954 #endif 1955 } 1956 1957 tcp_seq tcp_iss_seq = 0; /* tcp initial seq # */ 1958 #if NRND > 0 1959 u_int8_t tcp_iss_secret[16]; /* 128 bits; should be plenty */ 1960 #endif 1961 1962 /* 1963 * Get a new sequence value given a tcp control block 1964 */ 1965 tcp_seq 1966 tcp_new_iss(struct tcpcb *tp, tcp_seq addin) 1967 { 1968 1969 #ifdef INET 1970 if (tp->t_inpcb != NULL) { 1971 return (tcp_new_iss1(&tp->t_inpcb->inp_laddr, 1972 &tp->t_inpcb->inp_faddr, tp->t_inpcb->inp_lport, 1973 tp->t_inpcb->inp_fport, sizeof(tp->t_inpcb->inp_laddr), 1974 addin)); 1975 } 1976 #endif 1977 #ifdef INET6 1978 if (tp->t_in6pcb != NULL) { 1979 return (tcp_new_iss1(&tp->t_in6pcb->in6p_laddr, 1980 &tp->t_in6pcb->in6p_faddr, tp->t_in6pcb->in6p_lport, 1981 tp->t_in6pcb->in6p_fport, sizeof(tp->t_in6pcb->in6p_laddr), 1982 addin)); 1983 } 1984 #endif 1985 /* Not possible. */ 1986 panic("tcp_new_iss"); 1987 } 1988 1989 /* 1990 * This routine actually generates a new TCP initial sequence number. 1991 */ 1992 tcp_seq 1993 tcp_new_iss1(void *laddr, void *faddr, u_int16_t lport, u_int16_t fport, 1994 size_t addrsz, tcp_seq addin) 1995 { 1996 tcp_seq tcp_iss; 1997 1998 #if NRND > 0 1999 static int beenhere; 2000 2001 /* 2002 * If we haven't been here before, initialize our cryptographic 2003 * hash secret. 2004 */ 2005 if (beenhere == 0) { 2006 rnd_extract_data(tcp_iss_secret, sizeof(tcp_iss_secret), 2007 RND_EXTRACT_ANY); 2008 beenhere = 1; 2009 } 2010 2011 if (tcp_do_rfc1948) { 2012 MD5_CTX ctx; 2013 u_int8_t hash[16]; /* XXX MD5 knowledge */ 2014 2015 /* 2016 * Compute the base value of the ISS. It is a hash 2017 * of (saddr, sport, daddr, dport, secret). 2018 */ 2019 MD5Init(&ctx); 2020 2021 MD5Update(&ctx, (u_char *) laddr, addrsz); 2022 MD5Update(&ctx, (u_char *) &lport, sizeof(lport)); 2023 2024 MD5Update(&ctx, (u_char *) faddr, addrsz); 2025 MD5Update(&ctx, (u_char *) &fport, sizeof(fport)); 2026 2027 MD5Update(&ctx, tcp_iss_secret, sizeof(tcp_iss_secret)); 2028 2029 MD5Final(hash, &ctx); 2030 2031 memcpy(&tcp_iss, hash, sizeof(tcp_iss)); 2032 2033 /* 2034 * Now increment our "timer", and add it in to 2035 * the computed value. 2036 * 2037 * XXX Use `addin'? 2038 * XXX TCP_ISSINCR too large to use? 2039 */ 2040 tcp_iss_seq += TCP_ISSINCR; 2041 #ifdef TCPISS_DEBUG 2042 printf("ISS hash 0x%08x, ", tcp_iss); 2043 #endif 2044 tcp_iss += tcp_iss_seq + addin; 2045 #ifdef TCPISS_DEBUG 2046 printf("new ISS 0x%08x\n", tcp_iss); 2047 #endif 2048 } else 2049 #endif /* NRND > 0 */ 2050 { 2051 /* 2052 * Randomize. 2053 */ 2054 #if NRND > 0 2055 rnd_extract_data(&tcp_iss, sizeof(tcp_iss), RND_EXTRACT_ANY); 2056 #else 2057 tcp_iss = arc4random(); 2058 #endif 2059 2060 /* 2061 * If we were asked to add some amount to a known value, 2062 * we will take a random value obtained above, mask off 2063 * the upper bits, and add in the known value. We also 2064 * add in a constant to ensure that we are at least a 2065 * certain distance from the original value. 2066 * 2067 * This is used when an old connection is in timed wait 2068 * and we have a new one coming in, for instance. 2069 */ 2070 if (addin != 0) { 2071 #ifdef TCPISS_DEBUG 2072 printf("Random %08x, ", tcp_iss); 2073 #endif 2074 tcp_iss &= TCP_ISS_RANDOM_MASK; 2075 tcp_iss += addin + TCP_ISSINCR; 2076 #ifdef TCPISS_DEBUG 2077 printf("Old ISS %08x, ISS %08x\n", addin, tcp_iss); 2078 #endif 2079 } else { 2080 tcp_iss &= TCP_ISS_RANDOM_MASK; 2081 tcp_iss += tcp_iss_seq; 2082 tcp_iss_seq += TCP_ISSINCR; 2083 #ifdef TCPISS_DEBUG 2084 printf("ISS %08x\n", tcp_iss); 2085 #endif 2086 } 2087 } 2088 2089 if (tcp_compat_42) { 2090 /* 2091 * Limit it to the positive range for really old TCP 2092 * implementations. 2093 * Just AND off the top bit instead of checking if 2094 * is set first - saves a branch 50% of the time. 2095 */ 2096 tcp_iss &= 0x7fffffff; /* XXX */ 2097 } 2098 2099 return (tcp_iss); 2100 } 2101 2102 #ifdef IPSEC 2103 /* compute ESP/AH header size for TCP, including outer IP header. */ 2104 size_t 2105 ipsec4_hdrsiz_tcp(tp) 2106 struct tcpcb *tp; 2107 { 2108 struct inpcb *inp; 2109 size_t hdrsiz; 2110 2111 /* XXX mapped addr case (tp->t_in6pcb) */ 2112 if (!tp || !tp->t_template || !(inp = tp->t_inpcb)) 2113 return 0; 2114 switch (tp->t_family) { 2115 case AF_INET: 2116 /* XXX: should use currect direction. */ 2117 hdrsiz = ipsec4_hdrsiz(tp->t_template, IPSEC_DIR_OUTBOUND, inp); 2118 break; 2119 default: 2120 hdrsiz = 0; 2121 break; 2122 } 2123 2124 return hdrsiz; 2125 } 2126 2127 #ifdef INET6 2128 size_t 2129 ipsec6_hdrsiz_tcp(tp) 2130 struct tcpcb *tp; 2131 { 2132 struct in6pcb *in6p; 2133 size_t hdrsiz; 2134 2135 if (!tp || !tp->t_template || !(in6p = tp->t_in6pcb)) 2136 return 0; 2137 switch (tp->t_family) { 2138 case AF_INET6: 2139 /* XXX: should use currect direction. */ 2140 hdrsiz = ipsec6_hdrsiz(tp->t_template, IPSEC_DIR_OUTBOUND, in6p); 2141 break; 2142 case AF_INET: 2143 /* mapped address case - tricky */ 2144 default: 2145 hdrsiz = 0; 2146 break; 2147 } 2148 2149 return hdrsiz; 2150 } 2151 #endif 2152 #endif /*IPSEC*/ 2153 2154 /* 2155 * Determine the length of the TCP options for this connection. 2156 * 2157 * XXX: What do we do for SACK, when we add that? Just reserve 2158 * all of the space? Otherwise we can't exactly be incrementing 2159 * cwnd by an amount that varies depending on the amount we last 2160 * had to SACK! 2161 */ 2162 2163 u_int 2164 tcp_optlen(tp) 2165 struct tcpcb *tp; 2166 { 2167 if ((tp->t_flags & (TF_REQ_TSTMP|TF_RCVD_TSTMP|TF_NOOPT)) == 2168 (TF_REQ_TSTMP | TF_RCVD_TSTMP)) 2169 return TCPOLEN_TSTAMP_APPA; 2170 else 2171 return 0; 2172 } 2173
Indexes created Sun Sep 21 20:09:37 GMT 2025