1 /* $NetBSD: tcp_subr.c,v 1.257 2014/11/10 18:52:51 maxv 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, 2008 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 * 49 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 50 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 51 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 52 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 53 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 54 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 55 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 56 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 57 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 58 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 59 * POSSIBILITY OF SUCH DAMAGE. 60 */ 61 62 /* 63 * Copyright (c) 1982, 1986, 1988, 1990, 1993, 1995 64 * The Regents of the University of California. All rights reserved. 65 * 66 * Redistribution and use in source and binary forms, with or without 67 * modification, are permitted provided that the following conditions 68 * are met: 69 * 1. Redistributions of source code must retain the above copyright 70 * notice, this list of conditions and the following disclaimer. 71 * 2. Redistributions in binary form must reproduce the above copyright 72 * notice, this list of conditions and the following disclaimer in the 73 * documentation and/or other materials provided with the distribution. 74 * 3. Neither the name of the University nor the names of its contributors 75 * may be used to endorse or promote products derived from this software 76 * without specific prior written permission. 77 * 78 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 79 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 80 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 81 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 82 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 83 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 84 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 85 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 86 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 87 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 88 * SUCH DAMAGE. 89 * 90 * @(#)tcp_subr.c 8.2 (Berkeley) 5/24/95 91 */ 92 93 #include <sys/cdefs.h> 94 __KERNEL_RCSID(0, "$NetBSD: tcp_subr.c,v 1.257 2014/11/10 18:52:51 maxv Exp $"); 95 96 #include "opt_inet.h" 97 #include "opt_ipsec.h" 98 #include "opt_tcp_compat_42.h" 99 #include "opt_inet_csum.h" 100 #include "opt_mbuftrace.h" 101 102 #include <sys/param.h> 103 #include <sys/atomic.h> 104 #include <sys/proc.h> 105 #include <sys/systm.h> 106 #include <sys/mbuf.h> 107 #include <sys/once.h> 108 #include <sys/socket.h> 109 #include <sys/socketvar.h> 110 #include <sys/protosw.h> 111 #include <sys/errno.h> 112 #include <sys/kernel.h> 113 #include <sys/pool.h> 114 #include <sys/md5.h> 115 #include <sys/cprng.h> 116 117 #include <net/route.h> 118 #include <net/if.h> 119 120 #include <netinet/in.h> 121 #include <netinet/in_systm.h> 122 #include <netinet/ip.h> 123 #include <netinet/in_pcb.h> 124 #include <netinet/ip_var.h> 125 #include <netinet/ip_icmp.h> 126 127 #ifdef INET6 128 #ifndef INET 129 #include <netinet/in.h> 130 #endif 131 #include <netinet/ip6.h> 132 #include <netinet6/in6_pcb.h> 133 #include <netinet6/ip6_var.h> 134 #include <netinet6/in6_var.h> 135 #include <netinet6/ip6protosw.h> 136 #include <netinet/icmp6.h> 137 #include <netinet6/nd6.h> 138 #endif 139 140 #include <netinet/tcp.h> 141 #include <netinet/tcp_fsm.h> 142 #include <netinet/tcp_seq.h> 143 #include <netinet/tcp_timer.h> 144 #include <netinet/tcp_var.h> 145 #include <netinet/tcp_vtw.h> 146 #include <netinet/tcp_private.h> 147 #include <netinet/tcp_congctl.h> 148 #include <netinet/tcpip.h> 149 150 #ifdef IPSEC 151 #include <netipsec/ipsec.h> 152 #include <netipsec/xform.h> 153 #ifdef INET6 154 #include <netipsec/ipsec6.h> 155 #endif 156 #include <netipsec/key.h> 157 #endif /* IPSEC*/ 158 159 160 struct inpcbtable tcbtable; /* head of queue of active tcpcb's */ 161 u_int32_t tcp_now; /* slow ticks, for RFC 1323 timestamps */ 162 163 percpu_t *tcpstat_percpu; 164 165 /* patchable/settable parameters for tcp */ 166 int tcp_mssdflt = TCP_MSS; 167 int tcp_minmss = TCP_MINMSS; 168 int tcp_rttdflt = TCPTV_SRTTDFLT / PR_SLOWHZ; 169 int tcp_do_rfc1323 = 1; /* window scaling / timestamps (obsolete) */ 170 int tcp_do_rfc1948 = 0; /* ISS by cryptographic hash */ 171 int tcp_do_sack = 1; /* selective acknowledgement */ 172 int tcp_do_win_scale = 1; /* RFC1323 window scaling */ 173 int tcp_do_timestamps = 1; /* RFC1323 timestamps */ 174 int tcp_ack_on_push = 0; /* set to enable immediate ACK-on-PUSH */ 175 int tcp_do_ecn = 0; /* Explicit Congestion Notification */ 176 #ifndef TCP_INIT_WIN 177 #define TCP_INIT_WIN 4 /* initial slow start window */ 178 #endif 179 #ifndef TCP_INIT_WIN_LOCAL 180 #define TCP_INIT_WIN_LOCAL 4 /* initial slow start window for local nets */ 181 #endif 182 /* 183 * Up to 5 we scale linearly, to reach 3 * 1460; then (iw) * 1460. 184 * This is to simulate current behavior for iw == 4 185 */ 186 int tcp_init_win_max[] = { 187 1 * 1460, 188 1 * 1460, 189 2 * 1460, 190 2 * 1460, 191 3 * 1460, 192 5 * 1460, 193 6 * 1460, 194 7 * 1460, 195 8 * 1460, 196 9 * 1460, 197 10 * 1460 198 }; 199 int tcp_init_win = TCP_INIT_WIN; 200 int tcp_init_win_local = TCP_INIT_WIN_LOCAL; 201 int tcp_mss_ifmtu = 0; 202 #ifdef TCP_COMPAT_42 203 int tcp_compat_42 = 1; 204 #else 205 int tcp_compat_42 = 0; 206 #endif 207 int tcp_rst_ppslim = 100; /* 100pps */ 208 int tcp_ackdrop_ppslim = 100; /* 100pps */ 209 int tcp_do_loopback_cksum = 0; 210 int tcp_do_abc = 1; /* RFC3465 Appropriate byte counting. */ 211 int tcp_abc_aggressive = 1; /* 1: L=2*SMSS 0: L=1*SMSS */ 212 int tcp_sack_tp_maxholes = 32; 213 int tcp_sack_globalmaxholes = 1024; 214 int tcp_sack_globalholes = 0; 215 int tcp_ecn_maxretries = 1; 216 int tcp_msl_enable = 1; /* enable TIME_WAIT truncation */ 217 int tcp_msl_loop = PR_SLOWHZ; /* MSL for loopback */ 218 int tcp_msl_local = 5 * PR_SLOWHZ; /* MSL for 'local' */ 219 int tcp_msl_remote = TCPTV_MSL; /* MSL otherwise */ 220 int tcp_msl_remote_threshold = TCPTV_SRTTDFLT; /* RTT threshold */ 221 int tcp_rttlocal = 0; /* Use RTT to decide who's 'local' */ 222 223 int tcp4_vtw_enable = 0; /* 1 to enable */ 224 int tcp6_vtw_enable = 0; /* 1 to enable */ 225 int tcp_vtw_was_enabled = 0; 226 int tcp_vtw_entries = 1 << 4; /* 16 vestigial TIME_WAIT entries */ 227 228 /* tcb hash */ 229 #ifndef TCBHASHSIZE 230 #define TCBHASHSIZE 128 231 #endif 232 int tcbhashsize = TCBHASHSIZE; 233 234 /* syn hash parameters */ 235 #define TCP_SYN_HASH_SIZE 293 236 #define TCP_SYN_BUCKET_SIZE 35 237 int tcp_syn_cache_size = TCP_SYN_HASH_SIZE; 238 int tcp_syn_cache_limit = TCP_SYN_HASH_SIZE*TCP_SYN_BUCKET_SIZE; 239 int tcp_syn_bucket_limit = 3*TCP_SYN_BUCKET_SIZE; 240 struct syn_cache_head tcp_syn_cache[TCP_SYN_HASH_SIZE]; 241 242 int tcp_freeq(struct tcpcb *); 243 244 #ifdef INET 245 static void tcp_mtudisc_callback(struct in_addr); 246 #endif 247 248 #ifdef INET6 249 void tcp6_mtudisc(struct in6pcb *, int); 250 #endif 251 252 static struct pool tcpcb_pool; 253 254 static int tcp_drainwanted; 255 256 #ifdef TCP_CSUM_COUNTERS 257 #include <sys/device.h> 258 259 #if defined(INET) 260 struct evcnt tcp_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 261 NULL, "tcp", "hwcsum bad"); 262 struct evcnt tcp_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 263 NULL, "tcp", "hwcsum ok"); 264 struct evcnt tcp_hwcsum_data = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 265 NULL, "tcp", "hwcsum data"); 266 struct evcnt tcp_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 267 NULL, "tcp", "swcsum"); 268 269 EVCNT_ATTACH_STATIC(tcp_hwcsum_bad); 270 EVCNT_ATTACH_STATIC(tcp_hwcsum_ok); 271 EVCNT_ATTACH_STATIC(tcp_hwcsum_data); 272 EVCNT_ATTACH_STATIC(tcp_swcsum); 273 #endif /* defined(INET) */ 274 275 #if defined(INET6) 276 struct evcnt tcp6_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 277 NULL, "tcp6", "hwcsum bad"); 278 struct evcnt tcp6_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 279 NULL, "tcp6", "hwcsum ok"); 280 struct evcnt tcp6_hwcsum_data = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 281 NULL, "tcp6", "hwcsum data"); 282 struct evcnt tcp6_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 283 NULL, "tcp6", "swcsum"); 284 285 EVCNT_ATTACH_STATIC(tcp6_hwcsum_bad); 286 EVCNT_ATTACH_STATIC(tcp6_hwcsum_ok); 287 EVCNT_ATTACH_STATIC(tcp6_hwcsum_data); 288 EVCNT_ATTACH_STATIC(tcp6_swcsum); 289 #endif /* defined(INET6) */ 290 #endif /* TCP_CSUM_COUNTERS */ 291 292 293 #ifdef TCP_OUTPUT_COUNTERS 294 #include <sys/device.h> 295 296 struct evcnt tcp_output_bigheader = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 297 NULL, "tcp", "output big header"); 298 struct evcnt tcp_output_predict_hit = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 299 NULL, "tcp", "output predict hit"); 300 struct evcnt tcp_output_predict_miss = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 301 NULL, "tcp", "output predict miss"); 302 struct evcnt tcp_output_copysmall = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 303 NULL, "tcp", "output copy small"); 304 struct evcnt tcp_output_copybig = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 305 NULL, "tcp", "output copy big"); 306 struct evcnt tcp_output_refbig = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 307 NULL, "tcp", "output reference big"); 308 309 EVCNT_ATTACH_STATIC(tcp_output_bigheader); 310 EVCNT_ATTACH_STATIC(tcp_output_predict_hit); 311 EVCNT_ATTACH_STATIC(tcp_output_predict_miss); 312 EVCNT_ATTACH_STATIC(tcp_output_copysmall); 313 EVCNT_ATTACH_STATIC(tcp_output_copybig); 314 EVCNT_ATTACH_STATIC(tcp_output_refbig); 315 316 #endif /* TCP_OUTPUT_COUNTERS */ 317 318 #ifdef TCP_REASS_COUNTERS 319 #include <sys/device.h> 320 321 struct evcnt tcp_reass_ = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 322 NULL, "tcp_reass", "calls"); 323 struct evcnt tcp_reass_empty = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 324 &tcp_reass_, "tcp_reass", "insert into empty queue"); 325 struct evcnt tcp_reass_iteration[8] = { 326 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", ">7 iterations"), 327 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "1 iteration"), 328 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "2 iterations"), 329 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "3 iterations"), 330 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "4 iterations"), 331 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "5 iterations"), 332 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "6 iterations"), 333 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "7 iterations"), 334 }; 335 struct evcnt tcp_reass_prependfirst = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 336 &tcp_reass_, "tcp_reass", "prepend to first"); 337 struct evcnt tcp_reass_prepend = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 338 &tcp_reass_, "tcp_reass", "prepend"); 339 struct evcnt tcp_reass_insert = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 340 &tcp_reass_, "tcp_reass", "insert"); 341 struct evcnt tcp_reass_inserttail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 342 &tcp_reass_, "tcp_reass", "insert at tail"); 343 struct evcnt tcp_reass_append = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 344 &tcp_reass_, "tcp_reass", "append"); 345 struct evcnt tcp_reass_appendtail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 346 &tcp_reass_, "tcp_reass", "append to tail fragment"); 347 struct evcnt tcp_reass_overlaptail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 348 &tcp_reass_, "tcp_reass", "overlap at end"); 349 struct evcnt tcp_reass_overlapfront = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 350 &tcp_reass_, "tcp_reass", "overlap at start"); 351 struct evcnt tcp_reass_segdup = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 352 &tcp_reass_, "tcp_reass", "duplicate segment"); 353 struct evcnt tcp_reass_fragdup = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 354 &tcp_reass_, "tcp_reass", "duplicate fragment"); 355 356 EVCNT_ATTACH_STATIC(tcp_reass_); 357 EVCNT_ATTACH_STATIC(tcp_reass_empty); 358 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 0); 359 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 1); 360 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 2); 361 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 3); 362 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 4); 363 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 5); 364 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 6); 365 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 7); 366 EVCNT_ATTACH_STATIC(tcp_reass_prependfirst); 367 EVCNT_ATTACH_STATIC(tcp_reass_prepend); 368 EVCNT_ATTACH_STATIC(tcp_reass_insert); 369 EVCNT_ATTACH_STATIC(tcp_reass_inserttail); 370 EVCNT_ATTACH_STATIC(tcp_reass_append); 371 EVCNT_ATTACH_STATIC(tcp_reass_appendtail); 372 EVCNT_ATTACH_STATIC(tcp_reass_overlaptail); 373 EVCNT_ATTACH_STATIC(tcp_reass_overlapfront); 374 EVCNT_ATTACH_STATIC(tcp_reass_segdup); 375 EVCNT_ATTACH_STATIC(tcp_reass_fragdup); 376 377 #endif /* TCP_REASS_COUNTERS */ 378 379 #ifdef MBUFTRACE 380 struct mowner tcp_mowner = MOWNER_INIT("tcp", ""); 381 struct mowner tcp_rx_mowner = MOWNER_INIT("tcp", "rx"); 382 struct mowner tcp_tx_mowner = MOWNER_INIT("tcp", "tx"); 383 struct mowner tcp_sock_mowner = MOWNER_INIT("tcp", "sock"); 384 struct mowner tcp_sock_rx_mowner = MOWNER_INIT("tcp", "sock rx"); 385 struct mowner tcp_sock_tx_mowner = MOWNER_INIT("tcp", "sock tx"); 386 #endif 387 388 callout_t tcp_slowtimo_ch; 389 390 static int 391 do_tcpinit(void) 392 { 393 394 in_pcbinit(&tcbtable, tcbhashsize, tcbhashsize); 395 pool_init(&tcpcb_pool, sizeof(struct tcpcb), 0, 0, 0, "tcpcbpl", 396 NULL, IPL_SOFTNET); 397 398 tcp_usrreq_init(); 399 400 /* Initialize timer state. */ 401 tcp_timer_init(); 402 403 /* Initialize the compressed state engine. */ 404 syn_cache_init(); 405 406 /* Initialize the congestion control algorithms. */ 407 tcp_congctl_init(); 408 409 /* Initialize the TCPCB template. */ 410 tcp_tcpcb_template(); 411 412 /* Initialize reassembly queue */ 413 tcpipqent_init(); 414 415 /* SACK */ 416 tcp_sack_init(); 417 418 MOWNER_ATTACH(&tcp_tx_mowner); 419 MOWNER_ATTACH(&tcp_rx_mowner); 420 MOWNER_ATTACH(&tcp_reass_mowner); 421 MOWNER_ATTACH(&tcp_sock_mowner); 422 MOWNER_ATTACH(&tcp_sock_tx_mowner); 423 MOWNER_ATTACH(&tcp_sock_rx_mowner); 424 MOWNER_ATTACH(&tcp_mowner); 425 426 tcpstat_percpu = percpu_alloc(sizeof(uint64_t) * TCP_NSTATS); 427 428 vtw_earlyinit(); 429 430 callout_init(&tcp_slowtimo_ch, CALLOUT_MPSAFE); 431 callout_reset(&tcp_slowtimo_ch, 1, tcp_slowtimo, NULL); 432 433 return 0; 434 } 435 436 void 437 tcp_init_common(unsigned basehlen) 438 { 439 static ONCE_DECL(dotcpinit); 440 unsigned hlen = basehlen + sizeof(struct tcphdr); 441 unsigned oldhlen; 442 443 if (max_linkhdr + hlen > MHLEN) 444 panic("tcp_init"); 445 while ((oldhlen = max_protohdr) < hlen) 446 atomic_cas_uint(&max_protohdr, oldhlen, hlen); 447 448 RUN_ONCE(&dotcpinit, do_tcpinit); 449 } 450 451 /* 452 * Tcp initialization 453 */ 454 void 455 tcp_init(void) 456 { 457 458 icmp_mtudisc_callback_register(tcp_mtudisc_callback); 459 460 tcp_init_common(sizeof(struct ip)); 461 } 462 463 /* 464 * Create template to be used to send tcp packets on a connection. 465 * Call after host entry created, allocates an mbuf and fills 466 * in a skeletal tcp/ip header, minimizing the amount of work 467 * necessary when the connection is used. 468 */ 469 struct mbuf * 470 tcp_template(struct tcpcb *tp) 471 { 472 struct inpcb *inp = tp->t_inpcb; 473 #ifdef INET6 474 struct in6pcb *in6p = tp->t_in6pcb; 475 #endif 476 struct tcphdr *n; 477 struct mbuf *m; 478 int hlen; 479 480 switch (tp->t_family) { 481 case AF_INET: 482 hlen = sizeof(struct ip); 483 if (inp) 484 break; 485 #ifdef INET6 486 if (in6p) { 487 /* mapped addr case */ 488 if (IN6_IS_ADDR_V4MAPPED(&in6p->in6p_laddr) 489 && IN6_IS_ADDR_V4MAPPED(&in6p->in6p_faddr)) 490 break; 491 } 492 #endif 493 return NULL; /*EINVAL*/ 494 #ifdef INET6 495 case AF_INET6: 496 hlen = sizeof(struct ip6_hdr); 497 if (in6p) { 498 /* more sainty check? */ 499 break; 500 } 501 return NULL; /*EINVAL*/ 502 #endif 503 default: 504 hlen = 0; /*pacify gcc*/ 505 return NULL; /*EAFNOSUPPORT*/ 506 } 507 #ifdef DIAGNOSTIC 508 if (hlen + sizeof(struct tcphdr) > MCLBYTES) 509 panic("mclbytes too small for t_template"); 510 #endif 511 m = tp->t_template; 512 if (m && m->m_len == hlen + sizeof(struct tcphdr)) 513 ; 514 else { 515 if (m) 516 m_freem(m); 517 m = tp->t_template = NULL; 518 MGETHDR(m, M_DONTWAIT, MT_HEADER); 519 if (m && hlen + sizeof(struct tcphdr) > MHLEN) { 520 MCLGET(m, M_DONTWAIT); 521 if ((m->m_flags & M_EXT) == 0) { 522 m_free(m); 523 m = NULL; 524 } 525 } 526 if (m == NULL) 527 return NULL; 528 MCLAIM(m, &tcp_mowner); 529 m->m_pkthdr.len = m->m_len = hlen + sizeof(struct tcphdr); 530 } 531 532 memset(mtod(m, void *), 0, m->m_len); 533 534 n = (struct tcphdr *)(mtod(m, char *) + hlen); 535 536 switch (tp->t_family) { 537 case AF_INET: 538 { 539 struct ipovly *ipov; 540 mtod(m, struct ip *)->ip_v = 4; 541 mtod(m, struct ip *)->ip_hl = hlen >> 2; 542 ipov = mtod(m, struct ipovly *); 543 ipov->ih_pr = IPPROTO_TCP; 544 ipov->ih_len = htons(sizeof(struct tcphdr)); 545 if (inp) { 546 ipov->ih_src = inp->inp_laddr; 547 ipov->ih_dst = inp->inp_faddr; 548 } 549 #ifdef INET6 550 else if (in6p) { 551 /* mapped addr case */ 552 bcopy(&in6p->in6p_laddr.s6_addr32[3], &ipov->ih_src, 553 sizeof(ipov->ih_src)); 554 bcopy(&in6p->in6p_faddr.s6_addr32[3], &ipov->ih_dst, 555 sizeof(ipov->ih_dst)); 556 } 557 #endif 558 /* 559 * Compute the pseudo-header portion of the checksum 560 * now. We incrementally add in the TCP option and 561 * payload lengths later, and then compute the TCP 562 * checksum right before the packet is sent off onto 563 * the wire. 564 */ 565 n->th_sum = in_cksum_phdr(ipov->ih_src.s_addr, 566 ipov->ih_dst.s_addr, 567 htons(sizeof(struct tcphdr) + IPPROTO_TCP)); 568 break; 569 } 570 #ifdef INET6 571 case AF_INET6: 572 { 573 struct ip6_hdr *ip6; 574 mtod(m, struct ip *)->ip_v = 6; 575 ip6 = mtod(m, struct ip6_hdr *); 576 ip6->ip6_nxt = IPPROTO_TCP; 577 ip6->ip6_plen = htons(sizeof(struct tcphdr)); 578 ip6->ip6_src = in6p->in6p_laddr; 579 ip6->ip6_dst = in6p->in6p_faddr; 580 ip6->ip6_flow = in6p->in6p_flowinfo & IPV6_FLOWINFO_MASK; 581 if (ip6_auto_flowlabel) { 582 ip6->ip6_flow &= ~IPV6_FLOWLABEL_MASK; 583 ip6->ip6_flow |= 584 (htonl(ip6_randomflowlabel()) & IPV6_FLOWLABEL_MASK); 585 } 586 ip6->ip6_vfc &= ~IPV6_VERSION_MASK; 587 ip6->ip6_vfc |= IPV6_VERSION; 588 589 /* 590 * Compute the pseudo-header portion of the checksum 591 * now. We incrementally add in the TCP option and 592 * payload lengths later, and then compute the TCP 593 * checksum right before the packet is sent off onto 594 * the wire. 595 */ 596 n->th_sum = in6_cksum_phdr(&in6p->in6p_laddr, 597 &in6p->in6p_faddr, htonl(sizeof(struct tcphdr)), 598 htonl(IPPROTO_TCP)); 599 break; 600 } 601 #endif 602 } 603 if (inp) { 604 n->th_sport = inp->inp_lport; 605 n->th_dport = inp->inp_fport; 606 } 607 #ifdef INET6 608 else if (in6p) { 609 n->th_sport = in6p->in6p_lport; 610 n->th_dport = in6p->in6p_fport; 611 } 612 #endif 613 n->th_seq = 0; 614 n->th_ack = 0; 615 n->th_x2 = 0; 616 n->th_off = 5; 617 n->th_flags = 0; 618 n->th_win = 0; 619 n->th_urp = 0; 620 return (m); 621 } 622 623 /* 624 * Send a single message to the TCP at address specified by 625 * the given TCP/IP header. If m == 0, then we make a copy 626 * of the tcpiphdr at ti and send directly to the addressed host. 627 * This is used to force keep alive messages out using the TCP 628 * template for a connection tp->t_template. If flags are given 629 * then we send a message back to the TCP which originated the 630 * segment ti, and discard the mbuf containing it and any other 631 * attached mbufs. 632 * 633 * In any case the ack and sequence number of the transmitted 634 * segment are as specified by the parameters. 635 */ 636 int 637 tcp_respond(struct tcpcb *tp, struct mbuf *mtemplate, struct mbuf *m, 638 struct tcphdr *th0, tcp_seq ack, tcp_seq seq, int flags) 639 { 640 #ifdef INET6 641 struct rtentry *rt; 642 #endif 643 struct route *ro; 644 int error, tlen, win = 0; 645 int hlen; 646 struct ip *ip; 647 #ifdef INET6 648 struct ip6_hdr *ip6; 649 #endif 650 int family; /* family on packet, not inpcb/in6pcb! */ 651 struct tcphdr *th; 652 struct socket *so; 653 654 if (tp != NULL && (flags & TH_RST) == 0) { 655 #ifdef DIAGNOSTIC 656 if (tp->t_inpcb && tp->t_in6pcb) 657 panic("tcp_respond: both t_inpcb and t_in6pcb are set"); 658 #endif 659 #ifdef INET 660 if (tp->t_inpcb) 661 win = sbspace(&tp->t_inpcb->inp_socket->so_rcv); 662 #endif 663 #ifdef INET6 664 if (tp->t_in6pcb) 665 win = sbspace(&tp->t_in6pcb->in6p_socket->so_rcv); 666 #endif 667 } 668 669 th = NULL; /* Quell uninitialized warning */ 670 ip = NULL; 671 #ifdef INET6 672 ip6 = NULL; 673 #endif 674 if (m == 0) { 675 if (!mtemplate) 676 return EINVAL; 677 678 /* get family information from template */ 679 switch (mtod(mtemplate, struct ip *)->ip_v) { 680 case 4: 681 family = AF_INET; 682 hlen = sizeof(struct ip); 683 break; 684 #ifdef INET6 685 case 6: 686 family = AF_INET6; 687 hlen = sizeof(struct ip6_hdr); 688 break; 689 #endif 690 default: 691 return EAFNOSUPPORT; 692 } 693 694 MGETHDR(m, M_DONTWAIT, MT_HEADER); 695 if (m) { 696 MCLAIM(m, &tcp_tx_mowner); 697 MCLGET(m, M_DONTWAIT); 698 if ((m->m_flags & M_EXT) == 0) { 699 m_free(m); 700 m = NULL; 701 } 702 } 703 if (m == NULL) 704 return (ENOBUFS); 705 706 if (tcp_compat_42) 707 tlen = 1; 708 else 709 tlen = 0; 710 711 m->m_data += max_linkhdr; 712 bcopy(mtod(mtemplate, void *), mtod(m, void *), 713 mtemplate->m_len); 714 switch (family) { 715 case AF_INET: 716 ip = mtod(m, struct ip *); 717 th = (struct tcphdr *)(ip + 1); 718 break; 719 #ifdef INET6 720 case AF_INET6: 721 ip6 = mtod(m, struct ip6_hdr *); 722 th = (struct tcphdr *)(ip6 + 1); 723 break; 724 #endif 725 #if 0 726 default: 727 /* noone will visit here */ 728 m_freem(m); 729 return EAFNOSUPPORT; 730 #endif 731 } 732 flags = TH_ACK; 733 } else { 734 735 if ((m->m_flags & M_PKTHDR) == 0) { 736 #if 0 737 printf("non PKTHDR to tcp_respond\n"); 738 #endif 739 m_freem(m); 740 return EINVAL; 741 } 742 #ifdef DIAGNOSTIC 743 if (!th0) 744 panic("th0 == NULL in tcp_respond"); 745 #endif 746 747 /* get family information from m */ 748 switch (mtod(m, struct ip *)->ip_v) { 749 case 4: 750 family = AF_INET; 751 hlen = sizeof(struct ip); 752 ip = mtod(m, struct ip *); 753 break; 754 #ifdef INET6 755 case 6: 756 family = AF_INET6; 757 hlen = sizeof(struct ip6_hdr); 758 ip6 = mtod(m, struct ip6_hdr *); 759 break; 760 #endif 761 default: 762 m_freem(m); 763 return EAFNOSUPPORT; 764 } 765 /* clear h/w csum flags inherited from rx packet */ 766 m->m_pkthdr.csum_flags = 0; 767 768 if ((flags & TH_SYN) == 0 || sizeof(*th0) > (th0->th_off << 2)) 769 tlen = sizeof(*th0); 770 else 771 tlen = th0->th_off << 2; 772 773 if (m->m_len > hlen + tlen && (m->m_flags & M_EXT) == 0 && 774 mtod(m, char *) + hlen == (char *)th0) { 775 m->m_len = hlen + tlen; 776 m_freem(m->m_next); 777 m->m_next = NULL; 778 } else { 779 struct mbuf *n; 780 781 #ifdef DIAGNOSTIC 782 if (max_linkhdr + hlen + tlen > MCLBYTES) { 783 m_freem(m); 784 return EMSGSIZE; 785 } 786 #endif 787 MGETHDR(n, M_DONTWAIT, MT_HEADER); 788 if (n && max_linkhdr + hlen + tlen > MHLEN) { 789 MCLGET(n, M_DONTWAIT); 790 if ((n->m_flags & M_EXT) == 0) { 791 m_freem(n); 792 n = NULL; 793 } 794 } 795 if (!n) { 796 m_freem(m); 797 return ENOBUFS; 798 } 799 800 MCLAIM(n, &tcp_tx_mowner); 801 n->m_data += max_linkhdr; 802 n->m_len = hlen + tlen; 803 m_copyback(n, 0, hlen, mtod(m, void *)); 804 m_copyback(n, hlen, tlen, (void *)th0); 805 806 m_freem(m); 807 m = n; 808 n = NULL; 809 } 810 811 #define xchg(a,b,type) { type t; t=a; a=b; b=t; } 812 switch (family) { 813 case AF_INET: 814 ip = mtod(m, struct ip *); 815 th = (struct tcphdr *)(ip + 1); 816 ip->ip_p = IPPROTO_TCP; 817 xchg(ip->ip_dst, ip->ip_src, struct in_addr); 818 ip->ip_p = IPPROTO_TCP; 819 break; 820 #ifdef INET6 821 case AF_INET6: 822 ip6 = mtod(m, struct ip6_hdr *); 823 th = (struct tcphdr *)(ip6 + 1); 824 ip6->ip6_nxt = IPPROTO_TCP; 825 xchg(ip6->ip6_dst, ip6->ip6_src, struct in6_addr); 826 ip6->ip6_nxt = IPPROTO_TCP; 827 break; 828 #endif 829 #if 0 830 default: 831 /* noone will visit here */ 832 m_freem(m); 833 return EAFNOSUPPORT; 834 #endif 835 } 836 xchg(th->th_dport, th->th_sport, u_int16_t); 837 #undef xchg 838 tlen = 0; /*be friendly with the following code*/ 839 } 840 th->th_seq = htonl(seq); 841 th->th_ack = htonl(ack); 842 th->th_x2 = 0; 843 if ((flags & TH_SYN) == 0) { 844 if (tp) 845 win >>= tp->rcv_scale; 846 if (win > TCP_MAXWIN) 847 win = TCP_MAXWIN; 848 th->th_win = htons((u_int16_t)win); 849 th->th_off = sizeof (struct tcphdr) >> 2; 850 tlen += sizeof(*th); 851 } else 852 tlen += th->th_off << 2; 853 m->m_len = hlen + tlen; 854 m->m_pkthdr.len = hlen + tlen; 855 m->m_pkthdr.rcvif = NULL; 856 th->th_flags = flags; 857 th->th_urp = 0; 858 859 switch (family) { 860 #ifdef INET 861 case AF_INET: 862 { 863 struct ipovly *ipov = (struct ipovly *)ip; 864 memset(ipov->ih_x1, 0, sizeof ipov->ih_x1); 865 ipov->ih_len = htons((u_int16_t)tlen); 866 867 th->th_sum = 0; 868 th->th_sum = in_cksum(m, hlen + tlen); 869 ip->ip_len = htons(hlen + tlen); 870 ip->ip_ttl = ip_defttl; 871 break; 872 } 873 #endif 874 #ifdef INET6 875 case AF_INET6: 876 { 877 th->th_sum = 0; 878 th->th_sum = in6_cksum(m, IPPROTO_TCP, sizeof(struct ip6_hdr), 879 tlen); 880 ip6->ip6_plen = htons(tlen); 881 if (tp && tp->t_in6pcb) { 882 struct ifnet *oifp; 883 ro = &tp->t_in6pcb->in6p_route; 884 oifp = (rt = rtcache_validate(ro)) != NULL ? rt->rt_ifp 885 : NULL; 886 ip6->ip6_hlim = in6_selecthlim(tp->t_in6pcb, oifp); 887 } else 888 ip6->ip6_hlim = ip6_defhlim; 889 ip6->ip6_flow &= ~IPV6_FLOWINFO_MASK; 890 if (ip6_auto_flowlabel) { 891 ip6->ip6_flow |= 892 (htonl(ip6_randomflowlabel()) & IPV6_FLOWLABEL_MASK); 893 } 894 break; 895 } 896 #endif 897 } 898 899 if (tp && tp->t_inpcb) 900 so = tp->t_inpcb->inp_socket; 901 #ifdef INET6 902 else if (tp && tp->t_in6pcb) 903 so = tp->t_in6pcb->in6p_socket; 904 #endif 905 else 906 so = NULL; 907 908 if (tp != NULL && tp->t_inpcb != NULL) { 909 ro = &tp->t_inpcb->inp_route; 910 #ifdef DIAGNOSTIC 911 if (family != AF_INET) 912 panic("tcp_respond: address family mismatch"); 913 if (!in_hosteq(ip->ip_dst, tp->t_inpcb->inp_faddr)) { 914 panic("tcp_respond: ip_dst %x != inp_faddr %x", 915 ntohl(ip->ip_dst.s_addr), 916 ntohl(tp->t_inpcb->inp_faddr.s_addr)); 917 } 918 #endif 919 } 920 #ifdef INET6 921 else if (tp != NULL && tp->t_in6pcb != NULL) { 922 ro = (struct route *)&tp->t_in6pcb->in6p_route; 923 #ifdef DIAGNOSTIC 924 if (family == AF_INET) { 925 if (!IN6_IS_ADDR_V4MAPPED(&tp->t_in6pcb->in6p_faddr)) 926 panic("tcp_respond: not mapped addr"); 927 if (memcmp(&ip->ip_dst, 928 &tp->t_in6pcb->in6p_faddr.s6_addr32[3], 929 sizeof(ip->ip_dst)) != 0) { 930 panic("tcp_respond: ip_dst != in6p_faddr"); 931 } 932 } else if (family == AF_INET6) { 933 if (!IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst, 934 &tp->t_in6pcb->in6p_faddr)) 935 panic("tcp_respond: ip6_dst != in6p_faddr"); 936 } else 937 panic("tcp_respond: address family mismatch"); 938 #endif 939 } 940 #endif 941 else 942 ro = NULL; 943 944 switch (family) { 945 #ifdef INET 946 case AF_INET: 947 error = ip_output(m, NULL, ro, 948 (tp && tp->t_mtudisc ? IP_MTUDISC : 0), NULL, so); 949 break; 950 #endif 951 #ifdef INET6 952 case AF_INET6: 953 error = ip6_output(m, NULL, ro, 0, NULL, so, NULL); 954 break; 955 #endif 956 default: 957 error = EAFNOSUPPORT; 958 break; 959 } 960 961 return (error); 962 } 963 964 /* 965 * Template TCPCB. Rather than zeroing a new TCPCB and initializing 966 * a bunch of members individually, we maintain this template for the 967 * static and mostly-static components of the TCPCB, and copy it into 968 * the new TCPCB instead. 969 */ 970 static struct tcpcb tcpcb_template = { 971 .t_srtt = TCPTV_SRTTBASE, 972 .t_rttmin = TCPTV_MIN, 973 974 .snd_cwnd = TCP_MAXWIN << TCP_MAX_WINSHIFT, 975 .snd_ssthresh = TCP_MAXWIN << TCP_MAX_WINSHIFT, 976 .snd_numholes = 0, 977 .snd_cubic_wmax = 0, 978 .snd_cubic_wmax_last = 0, 979 .snd_cubic_ctime = 0, 980 981 .t_partialacks = -1, 982 .t_bytes_acked = 0, 983 }; 984 985 /* 986 * Updates the TCPCB template whenever a parameter that would affect 987 * the template is changed. 988 */ 989 void 990 tcp_tcpcb_template(void) 991 { 992 struct tcpcb *tp = &tcpcb_template; 993 int flags; 994 995 tp->t_peermss = tcp_mssdflt; 996 tp->t_ourmss = tcp_mssdflt; 997 tp->t_segsz = tcp_mssdflt; 998 999 flags = 0; 1000 if (tcp_do_rfc1323 && tcp_do_win_scale) 1001 flags |= TF_REQ_SCALE; 1002 if (tcp_do_rfc1323 && tcp_do_timestamps) 1003 flags |= TF_REQ_TSTMP; 1004 tp->t_flags = flags; 1005 1006 /* 1007 * Init srtt to TCPTV_SRTTBASE (0), so we can tell that we have no 1008 * rtt estimate. Set rttvar so that srtt + 2 * rttvar gives 1009 * reasonable initial retransmit time. 1010 */ 1011 tp->t_rttvar = tcp_rttdflt * PR_SLOWHZ << (TCP_RTTVAR_SHIFT + 2 - 1); 1012 TCPT_RANGESET(tp->t_rxtcur, TCP_REXMTVAL(tp), 1013 TCPTV_MIN, TCPTV_REXMTMAX); 1014 1015 /* Keep Alive */ 1016 tp->t_keepinit = tcp_keepinit; 1017 tp->t_keepidle = tcp_keepidle; 1018 tp->t_keepintvl = tcp_keepintvl; 1019 tp->t_keepcnt = tcp_keepcnt; 1020 tp->t_maxidle = tp->t_keepcnt * tp->t_keepintvl; 1021 1022 /* MSL */ 1023 tp->t_msl = TCPTV_MSL; 1024 } 1025 1026 /* 1027 * Create a new TCP control block, making an 1028 * empty reassembly queue and hooking it to the argument 1029 * protocol control block. 1030 */ 1031 /* family selects inpcb, or in6pcb */ 1032 struct tcpcb * 1033 tcp_newtcpcb(int family, void *aux) 1034 { 1035 #ifdef INET6 1036 struct rtentry *rt; 1037 #endif 1038 struct tcpcb *tp; 1039 int i; 1040 1041 /* XXX Consider using a pool_cache for speed. */ 1042 tp = pool_get(&tcpcb_pool, PR_NOWAIT); /* splsoftnet via tcp_usrreq */ 1043 if (tp == NULL) 1044 return (NULL); 1045 memcpy(tp, &tcpcb_template, sizeof(*tp)); 1046 TAILQ_INIT(&tp->segq); 1047 TAILQ_INIT(&tp->timeq); 1048 tp->t_family = family; /* may be overridden later on */ 1049 TAILQ_INIT(&tp->snd_holes); 1050 LIST_INIT(&tp->t_sc); /* XXX can template this */ 1051 1052 /* Don't sweat this loop; hopefully the compiler will unroll it. */ 1053 for (i = 0; i < TCPT_NTIMERS; i++) { 1054 callout_init(&tp->t_timer[i], CALLOUT_MPSAFE); 1055 TCP_TIMER_INIT(tp, i); 1056 } 1057 callout_init(&tp->t_delack_ch, CALLOUT_MPSAFE); 1058 1059 switch (family) { 1060 case AF_INET: 1061 { 1062 struct inpcb *inp = (struct inpcb *)aux; 1063 1064 inp->inp_ip.ip_ttl = ip_defttl; 1065 inp->inp_ppcb = (void *)tp; 1066 1067 tp->t_inpcb = inp; 1068 tp->t_mtudisc = ip_mtudisc; 1069 break; 1070 } 1071 #ifdef INET6 1072 case AF_INET6: 1073 { 1074 struct in6pcb *in6p = (struct in6pcb *)aux; 1075 1076 in6p->in6p_ip6.ip6_hlim = in6_selecthlim(in6p, 1077 (rt = rtcache_validate(&in6p->in6p_route)) != NULL 1078 ? rt->rt_ifp 1079 : NULL); 1080 in6p->in6p_ppcb = (void *)tp; 1081 1082 tp->t_in6pcb = in6p; 1083 /* for IPv6, always try to run path MTU discovery */ 1084 tp->t_mtudisc = 1; 1085 break; 1086 } 1087 #endif /* INET6 */ 1088 default: 1089 for (i = 0; i < TCPT_NTIMERS; i++) 1090 callout_destroy(&tp->t_timer[i]); 1091 callout_destroy(&tp->t_delack_ch); 1092 pool_put(&tcpcb_pool, tp); /* splsoftnet via tcp_usrreq */ 1093 return (NULL); 1094 } 1095 1096 /* 1097 * Initialize our timebase. When we send timestamps, we take 1098 * the delta from tcp_now -- this means each connection always 1099 * gets a timebase of 1, which makes it, among other things, 1100 * more difficult to determine how long a system has been up, 1101 * and thus how many TCP sequence increments have occurred. 1102 * 1103 * We start with 1, because 0 doesn't work with linux, which 1104 * considers timestamp 0 in a SYN packet as a bug and disables 1105 * timestamps. 1106 */ 1107 tp->ts_timebase = tcp_now - 1; 1108 1109 tcp_congctl_select(tp, tcp_congctl_global_name); 1110 1111 return (tp); 1112 } 1113 1114 /* 1115 * Drop a TCP connection, reporting 1116 * the specified error. If connection is synchronized, 1117 * then send a RST to peer. 1118 */ 1119 struct tcpcb * 1120 tcp_drop(struct tcpcb *tp, int errno) 1121 { 1122 struct socket *so = NULL; 1123 1124 #ifdef DIAGNOSTIC 1125 if (tp->t_inpcb && tp->t_in6pcb) 1126 panic("tcp_drop: both t_inpcb and t_in6pcb are set"); 1127 #endif 1128 #ifdef INET 1129 if (tp->t_inpcb) 1130 so = tp->t_inpcb->inp_socket; 1131 #endif 1132 #ifdef INET6 1133 if (tp->t_in6pcb) 1134 so = tp->t_in6pcb->in6p_socket; 1135 #endif 1136 if (!so) 1137 return NULL; 1138 1139 if (TCPS_HAVERCVDSYN(tp->t_state)) { 1140 tp->t_state = TCPS_CLOSED; 1141 (void) tcp_output(tp); 1142 TCP_STATINC(TCP_STAT_DROPS); 1143 } else 1144 TCP_STATINC(TCP_STAT_CONNDROPS); 1145 if (errno == ETIMEDOUT && tp->t_softerror) 1146 errno = tp->t_softerror; 1147 so->so_error = errno; 1148 return (tcp_close(tp)); 1149 } 1150 1151 /* 1152 * Close a TCP control block: 1153 * discard all space held by the tcp 1154 * discard internet protocol block 1155 * wake up any sleepers 1156 */ 1157 struct tcpcb * 1158 tcp_close(struct tcpcb *tp) 1159 { 1160 struct inpcb *inp; 1161 #ifdef INET6 1162 struct in6pcb *in6p; 1163 #endif 1164 struct socket *so; 1165 #ifdef RTV_RTT 1166 struct rtentry *rt; 1167 #endif 1168 struct route *ro; 1169 int j; 1170 1171 inp = tp->t_inpcb; 1172 #ifdef INET6 1173 in6p = tp->t_in6pcb; 1174 #endif 1175 so = NULL; 1176 ro = NULL; 1177 if (inp) { 1178 so = inp->inp_socket; 1179 ro = &inp->inp_route; 1180 } 1181 #ifdef INET6 1182 else if (in6p) { 1183 so = in6p->in6p_socket; 1184 ro = (struct route *)&in6p->in6p_route; 1185 } 1186 #endif 1187 1188 #ifdef RTV_RTT 1189 /* 1190 * If we sent enough data to get some meaningful characteristics, 1191 * save them in the routing entry. 'Enough' is arbitrarily 1192 * defined as the sendpipesize (default 4K) * 16. This would 1193 * give us 16 rtt samples assuming we only get one sample per 1194 * window (the usual case on a long haul net). 16 samples is 1195 * enough for the srtt filter to converge to within 5% of the correct 1196 * value; fewer samples and we could save a very bogus rtt. 1197 * 1198 * Don't update the default route's characteristics and don't 1199 * update anything that the user "locked". 1200 */ 1201 if (SEQ_LT(tp->iss + so->so_snd.sb_hiwat * 16, tp->snd_max) && 1202 ro && (rt = rtcache_validate(ro)) != NULL && 1203 !in_nullhost(satocsin(rt_getkey(rt))->sin_addr)) { 1204 u_long i = 0; 1205 1206 if ((rt->rt_rmx.rmx_locks & RTV_RTT) == 0) { 1207 i = tp->t_srtt * 1208 ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTT_SHIFT + 2)); 1209 if (rt->rt_rmx.rmx_rtt && i) 1210 /* 1211 * filter this update to half the old & half 1212 * the new values, converting scale. 1213 * See route.h and tcp_var.h for a 1214 * description of the scaling constants. 1215 */ 1216 rt->rt_rmx.rmx_rtt = 1217 (rt->rt_rmx.rmx_rtt + i) / 2; 1218 else 1219 rt->rt_rmx.rmx_rtt = i; 1220 } 1221 if ((rt->rt_rmx.rmx_locks & RTV_RTTVAR) == 0) { 1222 i = tp->t_rttvar * 1223 ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTTVAR_SHIFT + 2)); 1224 if (rt->rt_rmx.rmx_rttvar && i) 1225 rt->rt_rmx.rmx_rttvar = 1226 (rt->rt_rmx.rmx_rttvar + i) / 2; 1227 else 1228 rt->rt_rmx.rmx_rttvar = i; 1229 } 1230 /* 1231 * update the pipelimit (ssthresh) if it has been updated 1232 * already or if a pipesize was specified & the threshhold 1233 * got below half the pipesize. I.e., wait for bad news 1234 * before we start updating, then update on both good 1235 * and bad news. 1236 */ 1237 if (((rt->rt_rmx.rmx_locks & RTV_SSTHRESH) == 0 && 1238 (i = tp->snd_ssthresh) && rt->rt_rmx.rmx_ssthresh) || 1239 i < (rt->rt_rmx.rmx_sendpipe / 2)) { 1240 /* 1241 * convert the limit from user data bytes to 1242 * packets then to packet data bytes. 1243 */ 1244 i = (i + tp->t_segsz / 2) / tp->t_segsz; 1245 if (i < 2) 1246 i = 2; 1247 i *= (u_long)(tp->t_segsz + sizeof (struct tcpiphdr)); 1248 if (rt->rt_rmx.rmx_ssthresh) 1249 rt->rt_rmx.rmx_ssthresh = 1250 (rt->rt_rmx.rmx_ssthresh + i) / 2; 1251 else 1252 rt->rt_rmx.rmx_ssthresh = i; 1253 } 1254 } 1255 #endif /* RTV_RTT */ 1256 /* free the reassembly queue, if any */ 1257 TCP_REASS_LOCK(tp); 1258 (void) tcp_freeq(tp); 1259 TCP_REASS_UNLOCK(tp); 1260 1261 /* free the SACK holes list. */ 1262 tcp_free_sackholes(tp); 1263 tcp_congctl_release(tp); 1264 syn_cache_cleanup(tp); 1265 1266 if (tp->t_template) { 1267 m_free(tp->t_template); 1268 tp->t_template = NULL; 1269 } 1270 1271 /* 1272 * Detaching the pcb will unlock the socket/tcpcb, and stopping 1273 * the timers can also drop the lock. We need to prevent access 1274 * to the tcpcb as it's half torn down. Flag the pcb as dead 1275 * (prevents access by timers) and only then detach it. 1276 */ 1277 tp->t_flags |= TF_DEAD; 1278 if (inp) { 1279 inp->inp_ppcb = 0; 1280 soisdisconnected(so); 1281 in_pcbdetach(inp); 1282 } 1283 #ifdef INET6 1284 else if (in6p) { 1285 in6p->in6p_ppcb = 0; 1286 soisdisconnected(so); 1287 in6_pcbdetach(in6p); 1288 } 1289 #endif 1290 /* 1291 * pcb is no longer visble elsewhere, so we can safely release 1292 * the lock in callout_halt() if needed. 1293 */ 1294 TCP_STATINC(TCP_STAT_CLOSED); 1295 for (j = 0; j < TCPT_NTIMERS; j++) { 1296 callout_halt(&tp->t_timer[j], softnet_lock); 1297 callout_destroy(&tp->t_timer[j]); 1298 } 1299 callout_halt(&tp->t_delack_ch, softnet_lock); 1300 callout_destroy(&tp->t_delack_ch); 1301 pool_put(&tcpcb_pool, tp); 1302 1303 return NULL; 1304 } 1305 1306 int 1307 tcp_freeq(struct tcpcb *tp) 1308 { 1309 struct ipqent *qe; 1310 int rv = 0; 1311 #ifdef TCPREASS_DEBUG 1312 int i = 0; 1313 #endif 1314 1315 TCP_REASS_LOCK_CHECK(tp); 1316 1317 while ((qe = TAILQ_FIRST(&tp->segq)) != NULL) { 1318 #ifdef TCPREASS_DEBUG 1319 printf("tcp_freeq[%p,%d]: %u:%u(%u) 0x%02x\n", 1320 tp, i++, qe->ipqe_seq, qe->ipqe_seq + qe->ipqe_len, 1321 qe->ipqe_len, qe->ipqe_flags & (TH_SYN|TH_FIN|TH_RST)); 1322 #endif 1323 TAILQ_REMOVE(&tp->segq, qe, ipqe_q); 1324 TAILQ_REMOVE(&tp->timeq, qe, ipqe_timeq); 1325 m_freem(qe->ipqe_m); 1326 tcpipqent_free(qe); 1327 rv = 1; 1328 } 1329 tp->t_segqlen = 0; 1330 KASSERT(TAILQ_EMPTY(&tp->timeq)); 1331 return (rv); 1332 } 1333 1334 void 1335 tcp_fasttimo(void) 1336 { 1337 if (tcp_drainwanted) { 1338 tcp_drain(); 1339 tcp_drainwanted = 0; 1340 } 1341 } 1342 1343 void 1344 tcp_drainstub(void) 1345 { 1346 tcp_drainwanted = 1; 1347 } 1348 1349 /* 1350 * Protocol drain routine. Called when memory is in short supply. 1351 * Called from pr_fasttimo thus a callout context. 1352 */ 1353 void 1354 tcp_drain(void) 1355 { 1356 struct inpcb_hdr *inph; 1357 struct tcpcb *tp; 1358 1359 mutex_enter(softnet_lock); 1360 KERNEL_LOCK(1, NULL); 1361 1362 /* 1363 * Free the sequence queue of all TCP connections. 1364 */ 1365 TAILQ_FOREACH(inph, &tcbtable.inpt_queue, inph_queue) { 1366 switch (inph->inph_af) { 1367 case AF_INET: 1368 tp = intotcpcb((struct inpcb *)inph); 1369 break; 1370 #ifdef INET6 1371 case AF_INET6: 1372 tp = in6totcpcb((struct in6pcb *)inph); 1373 break; 1374 #endif 1375 default: 1376 tp = NULL; 1377 break; 1378 } 1379 if (tp != NULL) { 1380 /* 1381 * We may be called from a device's interrupt 1382 * context. If the tcpcb is already busy, 1383 * just bail out now. 1384 */ 1385 if (tcp_reass_lock_try(tp) == 0) 1386 continue; 1387 if (tcp_freeq(tp)) 1388 TCP_STATINC(TCP_STAT_CONNSDRAINED); 1389 TCP_REASS_UNLOCK(tp); 1390 } 1391 } 1392 1393 KERNEL_UNLOCK_ONE(NULL); 1394 mutex_exit(softnet_lock); 1395 } 1396 1397 /* 1398 * Notify a tcp user of an asynchronous error; 1399 * store error as soft error, but wake up user 1400 * (for now, won't do anything until can select for soft error). 1401 */ 1402 void 1403 tcp_notify(struct inpcb *inp, int error) 1404 { 1405 struct tcpcb *tp = (struct tcpcb *)inp->inp_ppcb; 1406 struct socket *so = inp->inp_socket; 1407 1408 /* 1409 * Ignore some errors if we are hooked up. 1410 * If connection hasn't completed, has retransmitted several times, 1411 * and receives a second error, give up now. This is better 1412 * than waiting a long time to establish a connection that 1413 * can never complete. 1414 */ 1415 if (tp->t_state == TCPS_ESTABLISHED && 1416 (error == EHOSTUNREACH || error == ENETUNREACH || 1417 error == EHOSTDOWN)) { 1418 return; 1419 } else if (TCPS_HAVEESTABLISHED(tp->t_state) == 0 && 1420 tp->t_rxtshift > 3 && tp->t_softerror) 1421 so->so_error = error; 1422 else 1423 tp->t_softerror = error; 1424 cv_broadcast(&so->so_cv); 1425 sorwakeup(so); 1426 sowwakeup(so); 1427 } 1428 1429 #ifdef INET6 1430 void 1431 tcp6_notify(struct in6pcb *in6p, int error) 1432 { 1433 struct tcpcb *tp = (struct tcpcb *)in6p->in6p_ppcb; 1434 struct socket *so = in6p->in6p_socket; 1435 1436 /* 1437 * Ignore some errors if we are hooked up. 1438 * If connection hasn't completed, has retransmitted several times, 1439 * and receives a second error, give up now. This is better 1440 * than waiting a long time to establish a connection that 1441 * can never complete. 1442 */ 1443 if (tp->t_state == TCPS_ESTABLISHED && 1444 (error == EHOSTUNREACH || error == ENETUNREACH || 1445 error == EHOSTDOWN)) { 1446 return; 1447 } else if (TCPS_HAVEESTABLISHED(tp->t_state) == 0 && 1448 tp->t_rxtshift > 3 && tp->t_softerror) 1449 so->so_error = error; 1450 else 1451 tp->t_softerror = error; 1452 cv_broadcast(&so->so_cv); 1453 sorwakeup(so); 1454 sowwakeup(so); 1455 } 1456 #endif 1457 1458 #ifdef INET6 1459 void * 1460 tcp6_ctlinput(int cmd, const struct sockaddr *sa, void *d) 1461 { 1462 struct tcphdr th; 1463 void (*notify)(struct in6pcb *, int) = tcp6_notify; 1464 int nmatch; 1465 struct ip6_hdr *ip6; 1466 const struct sockaddr_in6 *sa6_src = NULL; 1467 const struct sockaddr_in6 *sa6 = (const struct sockaddr_in6 *)sa; 1468 struct mbuf *m; 1469 int off; 1470 1471 if (sa->sa_family != AF_INET6 || 1472 sa->sa_len != sizeof(struct sockaddr_in6)) 1473 return NULL; 1474 if ((unsigned)cmd >= PRC_NCMDS) 1475 return NULL; 1476 else if (cmd == PRC_QUENCH) { 1477 /* 1478 * Don't honor ICMP Source Quench messages meant for 1479 * TCP connections. 1480 */ 1481 return NULL; 1482 } else if (PRC_IS_REDIRECT(cmd)) 1483 notify = in6_rtchange, d = NULL; 1484 else if (cmd == PRC_MSGSIZE) 1485 ; /* special code is present, see below */ 1486 else if (cmd == PRC_HOSTDEAD) 1487 d = NULL; 1488 else if (inet6ctlerrmap[cmd] == 0) 1489 return NULL; 1490 1491 /* if the parameter is from icmp6, decode it. */ 1492 if (d != NULL) { 1493 struct ip6ctlparam *ip6cp = (struct ip6ctlparam *)d; 1494 m = ip6cp->ip6c_m; 1495 ip6 = ip6cp->ip6c_ip6; 1496 off = ip6cp->ip6c_off; 1497 sa6_src = ip6cp->ip6c_src; 1498 } else { 1499 m = NULL; 1500 ip6 = NULL; 1501 sa6_src = &sa6_any; 1502 off = 0; 1503 } 1504 1505 if (ip6) { 1506 /* 1507 * XXX: We assume that when ip6 is non NULL, 1508 * M and OFF are valid. 1509 */ 1510 1511 /* check if we can safely examine src and dst ports */ 1512 if (m->m_pkthdr.len < off + sizeof(th)) { 1513 if (cmd == PRC_MSGSIZE) 1514 icmp6_mtudisc_update((struct ip6ctlparam *)d, 0); 1515 return NULL; 1516 } 1517 1518 memset(&th, 0, sizeof(th)); 1519 m_copydata(m, off, sizeof(th), (void *)&th); 1520 1521 if (cmd == PRC_MSGSIZE) { 1522 int valid = 0; 1523 1524 /* 1525 * Check to see if we have a valid TCP connection 1526 * corresponding to the address in the ICMPv6 message 1527 * payload. 1528 */ 1529 if (in6_pcblookup_connect(&tcbtable, &sa6->sin6_addr, 1530 th.th_dport, 1531 (const struct in6_addr *)&sa6_src->sin6_addr, 1532 th.th_sport, 0, 0)) 1533 valid++; 1534 1535 /* 1536 * Depending on the value of "valid" and routing table 1537 * size (mtudisc_{hi,lo}wat), we will: 1538 * - recalcurate the new MTU and create the 1539 * corresponding routing entry, or 1540 * - ignore the MTU change notification. 1541 */ 1542 icmp6_mtudisc_update((struct ip6ctlparam *)d, valid); 1543 1544 /* 1545 * no need to call in6_pcbnotify, it should have been 1546 * called via callback if necessary 1547 */ 1548 return NULL; 1549 } 1550 1551 nmatch = in6_pcbnotify(&tcbtable, sa, th.th_dport, 1552 (const struct sockaddr *)sa6_src, th.th_sport, cmd, NULL, notify); 1553 if (nmatch == 0 && syn_cache_count && 1554 (inet6ctlerrmap[cmd] == EHOSTUNREACH || 1555 inet6ctlerrmap[cmd] == ENETUNREACH || 1556 inet6ctlerrmap[cmd] == EHOSTDOWN)) 1557 syn_cache_unreach((const struct sockaddr *)sa6_src, 1558 sa, &th); 1559 } else { 1560 (void) in6_pcbnotify(&tcbtable, sa, 0, 1561 (const struct sockaddr *)sa6_src, 0, cmd, NULL, notify); 1562 } 1563 1564 return NULL; 1565 } 1566 #endif 1567 1568 #ifdef INET 1569 /* assumes that ip header and tcp header are contiguous on mbuf */ 1570 void * 1571 tcp_ctlinput(int cmd, const struct sockaddr *sa, void *v) 1572 { 1573 struct ip *ip = v; 1574 struct tcphdr *th; 1575 struct icmp *icp; 1576 extern const int inetctlerrmap[]; 1577 void (*notify)(struct inpcb *, int) = tcp_notify; 1578 int errno; 1579 int nmatch; 1580 struct tcpcb *tp; 1581 u_int mtu; 1582 tcp_seq seq; 1583 struct inpcb *inp; 1584 #ifdef INET6 1585 struct in6pcb *in6p; 1586 struct in6_addr src6, dst6; 1587 #endif 1588 1589 if (sa->sa_family != AF_INET || 1590 sa->sa_len != sizeof(struct sockaddr_in)) 1591 return NULL; 1592 if ((unsigned)cmd >= PRC_NCMDS) 1593 return NULL; 1594 errno = inetctlerrmap[cmd]; 1595 if (cmd == PRC_QUENCH) 1596 /* 1597 * Don't honor ICMP Source Quench messages meant for 1598 * TCP connections. 1599 */ 1600 return NULL; 1601 else if (PRC_IS_REDIRECT(cmd)) 1602 notify = in_rtchange, ip = 0; 1603 else if (cmd == PRC_MSGSIZE && ip && ip->ip_v == 4) { 1604 /* 1605 * Check to see if we have a valid TCP connection 1606 * corresponding to the address in the ICMP message 1607 * payload. 1608 * 1609 * Boundary check is made in icmp_input(), with ICMP_ADVLENMIN. 1610 */ 1611 th = (struct tcphdr *)((char *)ip + (ip->ip_hl << 2)); 1612 #ifdef INET6 1613 memset(&src6, 0, sizeof(src6)); 1614 memset(&dst6, 0, sizeof(dst6)); 1615 src6.s6_addr16[5] = dst6.s6_addr16[5] = 0xffff; 1616 memcpy(&src6.s6_addr32[3], &ip->ip_src, sizeof(struct in_addr)); 1617 memcpy(&dst6.s6_addr32[3], &ip->ip_dst, sizeof(struct in_addr)); 1618 #endif 1619 if ((inp = in_pcblookup_connect(&tcbtable, ip->ip_dst, 1620 th->th_dport, ip->ip_src, th->th_sport, 0)) != NULL) 1621 #ifdef INET6 1622 in6p = NULL; 1623 #else 1624 ; 1625 #endif 1626 #ifdef INET6 1627 else if ((in6p = in6_pcblookup_connect(&tcbtable, &dst6, 1628 th->th_dport, &src6, th->th_sport, 0, 0)) != NULL) 1629 ; 1630 #endif 1631 else 1632 return NULL; 1633 1634 /* 1635 * Now that we've validated that we are actually communicating 1636 * with the host indicated in the ICMP message, locate the 1637 * ICMP header, recalculate the new MTU, and create the 1638 * corresponding routing entry. 1639 */ 1640 icp = (struct icmp *)((char *)ip - 1641 offsetof(struct icmp, icmp_ip)); 1642 if (inp) { 1643 if ((tp = intotcpcb(inp)) == NULL) 1644 return NULL; 1645 } 1646 #ifdef INET6 1647 else if (in6p) { 1648 if ((tp = in6totcpcb(in6p)) == NULL) 1649 return NULL; 1650 } 1651 #endif 1652 else 1653 return NULL; 1654 seq = ntohl(th->th_seq); 1655 if (SEQ_LT(seq, tp->snd_una) || SEQ_GT(seq, tp->snd_max)) 1656 return NULL; 1657 /* 1658 * If the ICMP message advertises a Next-Hop MTU 1659 * equal or larger than the maximum packet size we have 1660 * ever sent, drop the message. 1661 */ 1662 mtu = (u_int)ntohs(icp->icmp_nextmtu); 1663 if (mtu >= tp->t_pmtud_mtu_sent) 1664 return NULL; 1665 if (mtu >= tcp_hdrsz(tp) + tp->t_pmtud_mss_acked) { 1666 /* 1667 * Calculate new MTU, and create corresponding 1668 * route (traditional PMTUD). 1669 */ 1670 tp->t_flags &= ~TF_PMTUD_PEND; 1671 icmp_mtudisc(icp, ip->ip_dst); 1672 } else { 1673 /* 1674 * Record the information got in the ICMP 1675 * message; act on it later. 1676 * If we had already recorded an ICMP message, 1677 * replace the old one only if the new message 1678 * refers to an older TCP segment 1679 */ 1680 if (tp->t_flags & TF_PMTUD_PEND) { 1681 if (SEQ_LT(tp->t_pmtud_th_seq, seq)) 1682 return NULL; 1683 } else 1684 tp->t_flags |= TF_PMTUD_PEND; 1685 tp->t_pmtud_th_seq = seq; 1686 tp->t_pmtud_nextmtu = icp->icmp_nextmtu; 1687 tp->t_pmtud_ip_len = icp->icmp_ip.ip_len; 1688 tp->t_pmtud_ip_hl = icp->icmp_ip.ip_hl; 1689 } 1690 return NULL; 1691 } else if (cmd == PRC_HOSTDEAD) 1692 ip = 0; 1693 else if (errno == 0) 1694 return NULL; 1695 if (ip && ip->ip_v == 4 && sa->sa_family == AF_INET) { 1696 th = (struct tcphdr *)((char *)ip + (ip->ip_hl << 2)); 1697 nmatch = in_pcbnotify(&tcbtable, satocsin(sa)->sin_addr, 1698 th->th_dport, ip->ip_src, th->th_sport, errno, notify); 1699 if (nmatch == 0 && syn_cache_count && 1700 (inetctlerrmap[cmd] == EHOSTUNREACH || 1701 inetctlerrmap[cmd] == ENETUNREACH || 1702 inetctlerrmap[cmd] == EHOSTDOWN)) { 1703 struct sockaddr_in sin; 1704 memset(&sin, 0, sizeof(sin)); 1705 sin.sin_len = sizeof(sin); 1706 sin.sin_family = AF_INET; 1707 sin.sin_port = th->th_sport; 1708 sin.sin_addr = ip->ip_src; 1709 syn_cache_unreach((struct sockaddr *)&sin, sa, th); 1710 } 1711 1712 /* XXX mapped address case */ 1713 } else 1714 in_pcbnotifyall(&tcbtable, satocsin(sa)->sin_addr, errno, 1715 notify); 1716 return NULL; 1717 } 1718 1719 /* 1720 * When a source quench is received, we are being notified of congestion. 1721 * Close the congestion window down to the Loss Window (one segment). 1722 * We will gradually open it again as we proceed. 1723 */ 1724 void 1725 tcp_quench(struct inpcb *inp, int errno) 1726 { 1727 struct tcpcb *tp = intotcpcb(inp); 1728 1729 if (tp) { 1730 tp->snd_cwnd = tp->t_segsz; 1731 tp->t_bytes_acked = 0; 1732 } 1733 } 1734 #endif 1735 1736 #ifdef INET6 1737 void 1738 tcp6_quench(struct in6pcb *in6p, int errno) 1739 { 1740 struct tcpcb *tp = in6totcpcb(in6p); 1741 1742 if (tp) { 1743 tp->snd_cwnd = tp->t_segsz; 1744 tp->t_bytes_acked = 0; 1745 } 1746 } 1747 #endif 1748 1749 #ifdef INET 1750 /* 1751 * Path MTU Discovery handlers. 1752 */ 1753 void 1754 tcp_mtudisc_callback(struct in_addr faddr) 1755 { 1756 #ifdef INET6 1757 struct in6_addr in6; 1758 #endif 1759 1760 in_pcbnotifyall(&tcbtable, faddr, EMSGSIZE, tcp_mtudisc); 1761 #ifdef INET6 1762 memset(&in6, 0, sizeof(in6)); 1763 in6.s6_addr16[5] = 0xffff; 1764 memcpy(&in6.s6_addr32[3], &faddr, sizeof(struct in_addr)); 1765 tcp6_mtudisc_callback(&in6); 1766 #endif 1767 } 1768 1769 /* 1770 * On receipt of path MTU corrections, flush old route and replace it 1771 * with the new one. Retransmit all unacknowledged packets, to ensure 1772 * that all packets will be received. 1773 */ 1774 void 1775 tcp_mtudisc(struct inpcb *inp, int errno) 1776 { 1777 struct tcpcb *tp = intotcpcb(inp); 1778 struct rtentry *rt = in_pcbrtentry(inp); 1779 1780 if (tp != 0) { 1781 if (rt != 0) { 1782 /* 1783 * If this was not a host route, remove and realloc. 1784 */ 1785 if ((rt->rt_flags & RTF_HOST) == 0) { 1786 in_rtchange(inp, errno); 1787 if ((rt = in_pcbrtentry(inp)) == 0) 1788 return; 1789 } 1790 1791 /* 1792 * Slow start out of the error condition. We 1793 * use the MTU because we know it's smaller 1794 * than the previously transmitted segment. 1795 * 1796 * Note: This is more conservative than the 1797 * suggestion in draft-floyd-incr-init-win-03. 1798 */ 1799 if (rt->rt_rmx.rmx_mtu != 0) 1800 tp->snd_cwnd = 1801 TCP_INITIAL_WINDOW(tcp_init_win, 1802 rt->rt_rmx.rmx_mtu); 1803 } 1804 1805 /* 1806 * Resend unacknowledged packets. 1807 */ 1808 tp->snd_nxt = tp->sack_newdata = tp->snd_una; 1809 tcp_output(tp); 1810 } 1811 } 1812 #endif 1813 1814 #ifdef INET6 1815 /* 1816 * Path MTU Discovery handlers. 1817 */ 1818 void 1819 tcp6_mtudisc_callback(struct in6_addr *faddr) 1820 { 1821 struct sockaddr_in6 sin6; 1822 1823 memset(&sin6, 0, sizeof(sin6)); 1824 sin6.sin6_family = AF_INET6; 1825 sin6.sin6_len = sizeof(struct sockaddr_in6); 1826 sin6.sin6_addr = *faddr; 1827 (void) in6_pcbnotify(&tcbtable, (struct sockaddr *)&sin6, 0, 1828 (const struct sockaddr *)&sa6_any, 0, PRC_MSGSIZE, NULL, tcp6_mtudisc); 1829 } 1830 1831 void 1832 tcp6_mtudisc(struct in6pcb *in6p, int errno) 1833 { 1834 struct tcpcb *tp = in6totcpcb(in6p); 1835 struct rtentry *rt = in6_pcbrtentry(in6p); 1836 1837 if (tp != 0) { 1838 if (rt != 0) { 1839 /* 1840 * If this was not a host route, remove and realloc. 1841 */ 1842 if ((rt->rt_flags & RTF_HOST) == 0) { 1843 in6_rtchange(in6p, errno); 1844 if ((rt = in6_pcbrtentry(in6p)) == 0) 1845 return; 1846 } 1847 1848 /* 1849 * Slow start out of the error condition. We 1850 * use the MTU because we know it's smaller 1851 * than the previously transmitted segment. 1852 * 1853 * Note: This is more conservative than the 1854 * suggestion in draft-floyd-incr-init-win-03. 1855 */ 1856 if (rt->rt_rmx.rmx_mtu != 0) 1857 tp->snd_cwnd = 1858 TCP_INITIAL_WINDOW(tcp_init_win, 1859 rt->rt_rmx.rmx_mtu); 1860 } 1861 1862 /* 1863 * Resend unacknowledged packets. 1864 */ 1865 tp->snd_nxt = tp->sack_newdata = tp->snd_una; 1866 tcp_output(tp); 1867 } 1868 } 1869 #endif /* INET6 */ 1870 1871 /* 1872 * Compute the MSS to advertise to the peer. Called only during 1873 * the 3-way handshake. If we are the server (peer initiated 1874 * connection), we are called with a pointer to the interface 1875 * on which the SYN packet arrived. If we are the client (we 1876 * initiated connection), we are called with a pointer to the 1877 * interface out which this connection should go. 1878 * 1879 * NOTE: Do not subtract IP option/extension header size nor IPsec 1880 * header size from MSS advertisement. MSS option must hold the maximum 1881 * segment size we can accept, so it must always be: 1882 * max(if mtu) - ip header - tcp header 1883 */ 1884 u_long 1885 tcp_mss_to_advertise(const struct ifnet *ifp, int af) 1886 { 1887 extern u_long in_maxmtu; 1888 u_long mss = 0; 1889 u_long hdrsiz; 1890 1891 /* 1892 * In order to avoid defeating path MTU discovery on the peer, 1893 * we advertise the max MTU of all attached networks as our MSS, 1894 * per RFC 1191, section 3.1. 1895 * 1896 * We provide the option to advertise just the MTU of 1897 * the interface on which we hope this connection will 1898 * be receiving. If we are responding to a SYN, we 1899 * will have a pretty good idea about this, but when 1900 * initiating a connection there is a bit more doubt. 1901 * 1902 * We also need to ensure that loopback has a large enough 1903 * MSS, as the loopback MTU is never included in in_maxmtu. 1904 */ 1905 1906 if (ifp != NULL) 1907 switch (af) { 1908 case AF_INET: 1909 mss = ifp->if_mtu; 1910 break; 1911 #ifdef INET6 1912 case AF_INET6: 1913 mss = IN6_LINKMTU(ifp); 1914 break; 1915 #endif 1916 } 1917 1918 if (tcp_mss_ifmtu == 0) 1919 switch (af) { 1920 case AF_INET: 1921 mss = max(in_maxmtu, mss); 1922 break; 1923 #ifdef INET6 1924 case AF_INET6: 1925 mss = max(in6_maxmtu, mss); 1926 break; 1927 #endif 1928 } 1929 1930 switch (af) { 1931 case AF_INET: 1932 hdrsiz = sizeof(struct ip); 1933 break; 1934 #ifdef INET6 1935 case AF_INET6: 1936 hdrsiz = sizeof(struct ip6_hdr); 1937 break; 1938 #endif 1939 default: 1940 hdrsiz = 0; 1941 break; 1942 } 1943 hdrsiz += sizeof(struct tcphdr); 1944 if (mss > hdrsiz) 1945 mss -= hdrsiz; 1946 1947 mss = max(tcp_mssdflt, mss); 1948 return (mss); 1949 } 1950 1951 /* 1952 * Set connection variables based on the peer's advertised MSS. 1953 * We are passed the TCPCB for the actual connection. If we 1954 * are the server, we are called by the compressed state engine 1955 * when the 3-way handshake is complete. If we are the client, 1956 * we are called when we receive the SYN,ACK from the server. 1957 * 1958 * NOTE: Our advertised MSS value must be initialized in the TCPCB 1959 * before this routine is called! 1960 */ 1961 void 1962 tcp_mss_from_peer(struct tcpcb *tp, int offer) 1963 { 1964 struct socket *so; 1965 #if defined(RTV_SPIPE) || defined(RTV_SSTHRESH) 1966 struct rtentry *rt; 1967 #endif 1968 u_long bufsize; 1969 int mss; 1970 1971 #ifdef DIAGNOSTIC 1972 if (tp->t_inpcb && tp->t_in6pcb) 1973 panic("tcp_mss_from_peer: 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_SPIPE) || defined(RTV_SSTHRESH) 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_SPIPE) || defined(RTV_SSTHRESH) 1989 rt = in6_pcbrtentry(tp->t_in6pcb); 1990 #endif 1991 } 1992 #endif 1993 1994 /* 1995 * As per RFC1122, use the default MSS value, unless they 1996 * sent us an offer. Do not accept offers less than 256 bytes. 1997 */ 1998 mss = tcp_mssdflt; 1999 if (offer) 2000 mss = offer; 2001 mss = max(mss, 256); /* sanity */ 2002 tp->t_peermss = mss; 2003 mss -= tcp_optlen(tp); 2004 #ifdef INET 2005 if (tp->t_inpcb) 2006 mss -= ip_optlen(tp->t_inpcb); 2007 #endif 2008 #ifdef INET6 2009 if (tp->t_in6pcb) 2010 mss -= ip6_optlen(tp->t_in6pcb); 2011 #endif 2012 2013 /* 2014 * If there's a pipesize, change the socket buffer to that size. 2015 * Make the socket buffer an integral number of MSS units. If 2016 * the MSS is larger than the socket buffer, artificially decrease 2017 * the MSS. 2018 */ 2019 #ifdef RTV_SPIPE 2020 if (rt != NULL && rt->rt_rmx.rmx_sendpipe != 0) 2021 bufsize = rt->rt_rmx.rmx_sendpipe; 2022 else 2023 #endif 2024 { 2025 KASSERT(so != NULL); 2026 bufsize = so->so_snd.sb_hiwat; 2027 } 2028 if (bufsize < mss) 2029 mss = bufsize; 2030 else { 2031 bufsize = roundup(bufsize, mss); 2032 if (bufsize > sb_max) 2033 bufsize = sb_max; 2034 (void) sbreserve(&so->so_snd, bufsize, so); 2035 } 2036 tp->t_segsz = mss; 2037 2038 #ifdef RTV_SSTHRESH 2039 if (rt != NULL && rt->rt_rmx.rmx_ssthresh) { 2040 /* 2041 * There's some sort of gateway or interface buffer 2042 * limit on the path. Use this to set the slow 2043 * start threshold, but set the threshold to no less 2044 * than 2 * MSS. 2045 */ 2046 tp->snd_ssthresh = max(2 * mss, rt->rt_rmx.rmx_ssthresh); 2047 } 2048 #endif 2049 } 2050 2051 /* 2052 * Processing necessary when a TCP connection is established. 2053 */ 2054 void 2055 tcp_established(struct tcpcb *tp) 2056 { 2057 struct socket *so; 2058 #ifdef RTV_RPIPE 2059 struct rtentry *rt; 2060 #endif 2061 u_long bufsize; 2062 2063 #ifdef DIAGNOSTIC 2064 if (tp->t_inpcb && tp->t_in6pcb) 2065 panic("tcp_established: both t_inpcb and t_in6pcb are set"); 2066 #endif 2067 so = NULL; 2068 rt = NULL; 2069 #ifdef INET 2070 /* This is a while() to reduce the dreadful stairstepping below */ 2071 while (tp->t_inpcb) { 2072 so = tp->t_inpcb->inp_socket; 2073 #if defined(RTV_RPIPE) 2074 rt = in_pcbrtentry(tp->t_inpcb); 2075 #endif 2076 if (__predict_true(tcp_msl_enable)) { 2077 if (tp->t_inpcb->inp_laddr.s_addr == INADDR_LOOPBACK) { 2078 tp->t_msl = tcp_msl_loop ? tcp_msl_loop : (TCPTV_MSL >> 2); 2079 break; 2080 } 2081 2082 if (__predict_false(tcp_rttlocal)) { 2083 /* This may be adjusted by tcp_input */ 2084 tp->t_msl = tcp_msl_local ? tcp_msl_local : (TCPTV_MSL >> 1); 2085 break; 2086 } 2087 if (in_localaddr(tp->t_inpcb->inp_faddr)) { 2088 tp->t_msl = tcp_msl_local ? tcp_msl_local : (TCPTV_MSL >> 1); 2089 break; 2090 } 2091 } 2092 tp->t_msl = tcp_msl_remote ? tcp_msl_remote : TCPTV_MSL; 2093 break; 2094 } 2095 #endif 2096 #ifdef INET6 2097 /* The !tp->t_inpcb lets the compiler know it can't be v4 *and* v6 */ 2098 while (!tp->t_inpcb && tp->t_in6pcb) { 2099 so = tp->t_in6pcb->in6p_socket; 2100 #if defined(RTV_RPIPE) 2101 rt = in6_pcbrtentry(tp->t_in6pcb); 2102 #endif 2103 if (__predict_true(tcp_msl_enable)) { 2104 extern const struct in6_addr in6addr_loopback; 2105 2106 if (IN6_ARE_ADDR_EQUAL(&tp->t_in6pcb->in6p_laddr, 2107 &in6addr_loopback)) { 2108 tp->t_msl = tcp_msl_loop ? tcp_msl_loop : (TCPTV_MSL >> 2); 2109 break; 2110 } 2111 2112 if (__predict_false(tcp_rttlocal)) { 2113 /* This may be adjusted by tcp_input */ 2114 tp->t_msl = tcp_msl_local ? tcp_msl_local : (TCPTV_MSL >> 1); 2115 break; 2116 } 2117 if (in6_localaddr(&tp->t_in6pcb->in6p_faddr)) { 2118 tp->t_msl = tcp_msl_local ? tcp_msl_local : (TCPTV_MSL >> 1); 2119 break; 2120 } 2121 } 2122 tp->t_msl = tcp_msl_remote ? tcp_msl_remote : TCPTV_MSL; 2123 break; 2124 } 2125 #endif 2126 2127 tp->t_state = TCPS_ESTABLISHED; 2128 TCP_TIMER_ARM(tp, TCPT_KEEP, tp->t_keepidle); 2129 2130 #ifdef RTV_RPIPE 2131 if (rt != NULL && rt->rt_rmx.rmx_recvpipe != 0) 2132 bufsize = rt->rt_rmx.rmx_recvpipe; 2133 else 2134 #endif 2135 { 2136 KASSERT(so != NULL); 2137 bufsize = so->so_rcv.sb_hiwat; 2138 } 2139 if (bufsize > tp->t_ourmss) { 2140 bufsize = roundup(bufsize, tp->t_ourmss); 2141 if (bufsize > sb_max) 2142 bufsize = sb_max; 2143 (void) sbreserve(&so->so_rcv, bufsize, so); 2144 } 2145 } 2146 2147 /* 2148 * Check if there's an initial rtt or rttvar. Convert from the 2149 * route-table units to scaled multiples of the slow timeout timer. 2150 * Called only during the 3-way handshake. 2151 */ 2152 void 2153 tcp_rmx_rtt(struct tcpcb *tp) 2154 { 2155 #ifdef RTV_RTT 2156 struct rtentry *rt = NULL; 2157 int rtt; 2158 2159 #ifdef DIAGNOSTIC 2160 if (tp->t_inpcb && tp->t_in6pcb) 2161 panic("tcp_rmx_rtt: both t_inpcb and t_in6pcb are set"); 2162 #endif 2163 #ifdef INET 2164 if (tp->t_inpcb) 2165 rt = in_pcbrtentry(tp->t_inpcb); 2166 #endif 2167 #ifdef INET6 2168 if (tp->t_in6pcb) 2169 rt = in6_pcbrtentry(tp->t_in6pcb); 2170 #endif 2171 if (rt == NULL) 2172 return; 2173 2174 if (tp->t_srtt == 0 && (rtt = rt->rt_rmx.rmx_rtt)) { 2175 /* 2176 * XXX The lock bit for MTU indicates that the value 2177 * is also a minimum value; this is subject to time. 2178 */ 2179 if (rt->rt_rmx.rmx_locks & RTV_RTT) 2180 TCPT_RANGESET(tp->t_rttmin, 2181 rtt / (RTM_RTTUNIT / PR_SLOWHZ), 2182 TCPTV_MIN, TCPTV_REXMTMAX); 2183 tp->t_srtt = rtt / 2184 ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTT_SHIFT + 2)); 2185 if (rt->rt_rmx.rmx_rttvar) { 2186 tp->t_rttvar = rt->rt_rmx.rmx_rttvar / 2187 ((RTM_RTTUNIT / PR_SLOWHZ) >> 2188 (TCP_RTTVAR_SHIFT + 2)); 2189 } else { 2190 /* Default variation is +- 1 rtt */ 2191 tp->t_rttvar = 2192 tp->t_srtt >> (TCP_RTT_SHIFT - TCP_RTTVAR_SHIFT); 2193 } 2194 TCPT_RANGESET(tp->t_rxtcur, 2195 ((tp->t_srtt >> 2) + tp->t_rttvar) >> (1 + 2), 2196 tp->t_rttmin, TCPTV_REXMTMAX); 2197 } 2198 #endif 2199 } 2200 2201 tcp_seq tcp_iss_seq = 0; /* tcp initial seq # */ 2202 u_int8_t tcp_iss_secret[16]; /* 128 bits; should be plenty */ 2203 2204 /* 2205 * Get a new sequence value given a tcp control block 2206 */ 2207 tcp_seq 2208 tcp_new_iss(struct tcpcb *tp, tcp_seq addin) 2209 { 2210 2211 #ifdef INET 2212 if (tp->t_inpcb != NULL) { 2213 return (tcp_new_iss1(&tp->t_inpcb->inp_laddr, 2214 &tp->t_inpcb->inp_faddr, tp->t_inpcb->inp_lport, 2215 tp->t_inpcb->inp_fport, sizeof(tp->t_inpcb->inp_laddr), 2216 addin)); 2217 } 2218 #endif 2219 #ifdef INET6 2220 if (tp->t_in6pcb != NULL) { 2221 return (tcp_new_iss1(&tp->t_in6pcb->in6p_laddr, 2222 &tp->t_in6pcb->in6p_faddr, tp->t_in6pcb->in6p_lport, 2223 tp->t_in6pcb->in6p_fport, sizeof(tp->t_in6pcb->in6p_laddr), 2224 addin)); 2225 } 2226 #endif 2227 /* Not possible. */ 2228 panic("tcp_new_iss"); 2229 } 2230 2231 /* 2232 * This routine actually generates a new TCP initial sequence number. 2233 */ 2234 tcp_seq 2235 tcp_new_iss1(void *laddr, void *faddr, u_int16_t lport, u_int16_t fport, 2236 size_t addrsz, tcp_seq addin) 2237 { 2238 tcp_seq tcp_iss; 2239 2240 static bool tcp_iss_gotten_secret; 2241 2242 /* 2243 * If we haven't been here before, initialize our cryptographic 2244 * hash secret. 2245 */ 2246 if (tcp_iss_gotten_secret == false) { 2247 cprng_strong(kern_cprng, 2248 tcp_iss_secret, sizeof(tcp_iss_secret), FASYNC); 2249 tcp_iss_gotten_secret = true; 2250 } 2251 2252 if (tcp_do_rfc1948) { 2253 MD5_CTX ctx; 2254 u_int8_t hash[16]; /* XXX MD5 knowledge */ 2255 2256 /* 2257 * Compute the base value of the ISS. It is a hash 2258 * of (saddr, sport, daddr, dport, secret). 2259 */ 2260 MD5Init(&ctx); 2261 2262 MD5Update(&ctx, (u_char *) laddr, addrsz); 2263 MD5Update(&ctx, (u_char *) &lport, sizeof(lport)); 2264 2265 MD5Update(&ctx, (u_char *) faddr, addrsz); 2266 MD5Update(&ctx, (u_char *) &fport, sizeof(fport)); 2267 2268 MD5Update(&ctx, tcp_iss_secret, sizeof(tcp_iss_secret)); 2269 2270 MD5Final(hash, &ctx); 2271 2272 memcpy(&tcp_iss, hash, sizeof(tcp_iss)); 2273 2274 /* 2275 * Now increment our "timer", and add it in to 2276 * the computed value. 2277 * 2278 * XXX Use `addin'? 2279 * XXX TCP_ISSINCR too large to use? 2280 */ 2281 tcp_iss_seq += TCP_ISSINCR; 2282 #ifdef TCPISS_DEBUG 2283 printf("ISS hash 0x%08x, ", tcp_iss); 2284 #endif 2285 tcp_iss += tcp_iss_seq + addin; 2286 #ifdef TCPISS_DEBUG 2287 printf("new ISS 0x%08x\n", tcp_iss); 2288 #endif 2289 } else { 2290 /* 2291 * Randomize. 2292 */ 2293 tcp_iss = cprng_fast32(); 2294 2295 /* 2296 * If we were asked to add some amount to a known value, 2297 * we will take a random value obtained above, mask off 2298 * the upper bits, and add in the known value. We also 2299 * add in a constant to ensure that we are at least a 2300 * certain distance from the original value. 2301 * 2302 * This is used when an old connection is in timed wait 2303 * and we have a new one coming in, for instance. 2304 */ 2305 if (addin != 0) { 2306 #ifdef TCPISS_DEBUG 2307 printf("Random %08x, ", tcp_iss); 2308 #endif 2309 tcp_iss &= TCP_ISS_RANDOM_MASK; 2310 tcp_iss += addin + TCP_ISSINCR; 2311 #ifdef TCPISS_DEBUG 2312 printf("Old ISS %08x, ISS %08x\n", addin, tcp_iss); 2313 #endif 2314 } else { 2315 tcp_iss &= TCP_ISS_RANDOM_MASK; 2316 tcp_iss += tcp_iss_seq; 2317 tcp_iss_seq += TCP_ISSINCR; 2318 #ifdef TCPISS_DEBUG 2319 printf("ISS %08x\n", tcp_iss); 2320 #endif 2321 } 2322 } 2323 2324 if (tcp_compat_42) { 2325 /* 2326 * Limit it to the positive range for really old TCP 2327 * implementations. 2328 * Just AND off the top bit instead of checking if 2329 * is set first - saves a branch 50% of the time. 2330 */ 2331 tcp_iss &= 0x7fffffff; /* XXX */ 2332 } 2333 2334 return (tcp_iss); 2335 } 2336 2337 #if defined(IPSEC) 2338 /* compute ESP/AH header size for TCP, including outer IP header. */ 2339 size_t 2340 ipsec4_hdrsiz_tcp(struct tcpcb *tp) 2341 { 2342 struct inpcb *inp; 2343 size_t hdrsiz; 2344 2345 /* XXX mapped addr case (tp->t_in6pcb) */ 2346 if (!tp || !tp->t_template || !(inp = tp->t_inpcb)) 2347 return 0; 2348 switch (tp->t_family) { 2349 case AF_INET: 2350 /* XXX: should use currect direction. */ 2351 hdrsiz = ipsec4_hdrsiz(tp->t_template, IPSEC_DIR_OUTBOUND, inp); 2352 break; 2353 default: 2354 hdrsiz = 0; 2355 break; 2356 } 2357 2358 return hdrsiz; 2359 } 2360 2361 #ifdef INET6 2362 size_t 2363 ipsec6_hdrsiz_tcp(struct tcpcb *tp) 2364 { 2365 struct in6pcb *in6p; 2366 size_t hdrsiz; 2367 2368 if (!tp || !tp->t_template || !(in6p = tp->t_in6pcb)) 2369 return 0; 2370 switch (tp->t_family) { 2371 case AF_INET6: 2372 /* XXX: should use currect direction. */ 2373 hdrsiz = ipsec6_hdrsiz(tp->t_template, IPSEC_DIR_OUTBOUND, in6p); 2374 break; 2375 case AF_INET: 2376 /* mapped address case - tricky */ 2377 default: 2378 hdrsiz = 0; 2379 break; 2380 } 2381 2382 return hdrsiz; 2383 } 2384 #endif 2385 #endif /*IPSEC*/ 2386 2387 /* 2388 * Determine the length of the TCP options for this connection. 2389 * 2390 * XXX: What do we do for SACK, when we add that? Just reserve 2391 * all of the space? Otherwise we can't exactly be incrementing 2392 * cwnd by an amount that varies depending on the amount we last 2393 * had to SACK! 2394 */ 2395 2396 u_int 2397 tcp_optlen(struct tcpcb *tp) 2398 { 2399 u_int optlen; 2400 2401 optlen = 0; 2402 if ((tp->t_flags & (TF_REQ_TSTMP|TF_RCVD_TSTMP|TF_NOOPT)) == 2403 (TF_REQ_TSTMP | TF_RCVD_TSTMP)) 2404 optlen += TCPOLEN_TSTAMP_APPA; 2405 2406 #ifdef TCP_SIGNATURE 2407 if (tp->t_flags & TF_SIGNATURE) 2408 optlen += TCPOLEN_SIGNATURE + 2; 2409 #endif /* TCP_SIGNATURE */ 2410 2411 return optlen; 2412 } 2413 2414 u_int 2415 tcp_hdrsz(struct tcpcb *tp) 2416 { 2417 u_int hlen; 2418 2419 switch (tp->t_family) { 2420 #ifdef INET6 2421 case AF_INET6: 2422 hlen = sizeof(struct ip6_hdr); 2423 break; 2424 #endif 2425 case AF_INET: 2426 hlen = sizeof(struct ip); 2427 break; 2428 default: 2429 hlen = 0; 2430 break; 2431 } 2432 hlen += sizeof(struct tcphdr); 2433 2434 if ((tp->t_flags & (TF_REQ_TSTMP|TF_NOOPT)) == TF_REQ_TSTMP && 2435 (tp->t_flags & TF_RCVD_TSTMP) == TF_RCVD_TSTMP) 2436 hlen += TCPOLEN_TSTAMP_APPA; 2437 #ifdef TCP_SIGNATURE 2438 if (tp->t_flags & TF_SIGNATURE) 2439 hlen += TCPOLEN_SIGLEN; 2440 #endif 2441 return hlen; 2442 } 2443 2444 void 2445 tcp_statinc(u_int stat) 2446 { 2447 2448 KASSERT(stat < TCP_NSTATS); 2449 TCP_STATINC(stat); 2450 } 2451 2452 void 2453 tcp_statadd(u_int stat, uint64_t val) 2454 { 2455 2456 KASSERT(stat < TCP_NSTATS); 2457 TCP_STATADD(stat, val); 2458 } 2459