in_pcb.c revision 1.154
1 /* $NetBSD: in_pcb.c,v 1.154 2014/11/25 15:04:37 seanb 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) 1998, 2011 The NetBSD Foundation, Inc. 34 * All rights reserved. 35 * 36 * This code is derived from software contributed to The NetBSD Foundation 37 * by Coyote Point Systems, Inc. 38 * This code is derived from software contributed to The NetBSD Foundation 39 * by Public Access Networks Corporation ("Panix"). It was developed under 40 * contract to Panix by Eric Haszlakiewicz and Thor Lancelot Simon. 41 * 42 * Redistribution and use in source and binary forms, with or without 43 * modification, are permitted provided that the following conditions 44 * are met: 45 * 1. Redistributions of source code must retain the above copyright 46 * notice, this list of conditions and the following disclaimer. 47 * 2. Redistributions in binary form must reproduce the above copyright 48 * notice, this list of conditions and the following disclaimer in the 49 * documentation and/or other materials provided with the distribution. 50 * 51 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 52 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 53 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 54 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 55 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 56 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 57 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 58 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 59 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 60 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 61 * POSSIBILITY OF SUCH DAMAGE. 62 */ 63 64 /* 65 * Copyright (c) 1982, 1986, 1991, 1993, 1995 66 * The Regents of the University of California. All rights reserved. 67 * 68 * Redistribution and use in source and binary forms, with or without 69 * modification, are permitted provided that the following conditions 70 * are met: 71 * 1. Redistributions of source code must retain the above copyright 72 * notice, this list of conditions and the following disclaimer. 73 * 2. Redistributions in binary form must reproduce the above copyright 74 * notice, this list of conditions and the following disclaimer in the 75 * documentation and/or other materials provided with the distribution. 76 * 3. Neither the name of the University nor the names of its contributors 77 * may be used to endorse or promote products derived from this software 78 * without specific prior written permission. 79 * 80 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 81 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 82 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 83 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 84 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 85 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 86 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 87 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 88 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 89 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 90 * SUCH DAMAGE. 91 * 92 * @(#)in_pcb.c 8.4 (Berkeley) 5/24/95 93 */ 94 95 #include <sys/cdefs.h> 96 __KERNEL_RCSID(0, "$NetBSD: in_pcb.c,v 1.154 2014/11/25 15:04:37 seanb Exp $"); 97 98 #include "opt_inet.h" 99 #include "opt_ipsec.h" 100 101 #include <sys/param.h> 102 #include <sys/systm.h> 103 #include <sys/mbuf.h> 104 #include <sys/protosw.h> 105 #include <sys/socket.h> 106 #include <sys/socketvar.h> 107 #include <sys/ioctl.h> 108 #include <sys/errno.h> 109 #include <sys/time.h> 110 #include <sys/once.h> 111 #include <sys/pool.h> 112 #include <sys/proc.h> 113 #include <sys/kauth.h> 114 #include <sys/uidinfo.h> 115 #include <sys/domain.h> 116 117 #include <net/if.h> 118 #include <net/route.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/in_var.h> 125 #include <netinet/ip_var.h> 126 #include <netinet/portalgo.h> 127 128 #ifdef INET6 129 #include <netinet/ip6.h> 130 #include <netinet6/ip6_var.h> 131 #include <netinet6/in6_pcb.h> 132 #endif 133 134 #ifdef IPSEC 135 #include <netipsec/ipsec.h> 136 #include <netipsec/key.h> 137 #endif /* IPSEC */ 138 139 #include <netinet/tcp_vtw.h> 140 141 struct in_addr zeroin_addr; 142 143 #define INPCBHASH_PORT(table, lport) \ 144 &(table)->inpt_porthashtbl[ntohs(lport) & (table)->inpt_porthash] 145 #define INPCBHASH_BIND(table, laddr, lport) \ 146 &(table)->inpt_bindhashtbl[ \ 147 ((ntohl((laddr).s_addr) + ntohs(lport))) & (table)->inpt_bindhash] 148 #define INPCBHASH_CONNECT(table, faddr, fport, laddr, lport) \ 149 &(table)->inpt_connecthashtbl[ \ 150 ((ntohl((faddr).s_addr) + ntohs(fport)) + \ 151 (ntohl((laddr).s_addr) + ntohs(lport))) & (table)->inpt_connecthash] 152 153 int anonportmin = IPPORT_ANONMIN; 154 int anonportmax = IPPORT_ANONMAX; 155 int lowportmin = IPPORT_RESERVEDMIN; 156 int lowportmax = IPPORT_RESERVEDMAX; 157 158 static struct pool inpcb_pool; 159 160 static int 161 inpcb_poolinit(void) 162 { 163 164 pool_init(&inpcb_pool, sizeof(struct inpcb), 0, 0, 0, "inpcbpl", NULL, 165 IPL_NET); 166 return 0; 167 } 168 169 void 170 in_pcbinit(struct inpcbtable *table, int bindhashsize, int connecthashsize) 171 { 172 static ONCE_DECL(control); 173 174 TAILQ_INIT(&table->inpt_queue); 175 table->inpt_porthashtbl = hashinit(bindhashsize, HASH_LIST, true, 176 &table->inpt_porthash); 177 table->inpt_bindhashtbl = hashinit(bindhashsize, HASH_LIST, true, 178 &table->inpt_bindhash); 179 table->inpt_connecthashtbl = hashinit(connecthashsize, HASH_LIST, true, 180 &table->inpt_connecthash); 181 table->inpt_lastlow = IPPORT_RESERVEDMAX; 182 table->inpt_lastport = (u_int16_t)anonportmax; 183 184 RUN_ONCE(&control, inpcb_poolinit); 185 } 186 187 int 188 in_pcballoc(struct socket *so, void *v) 189 { 190 struct inpcbtable *table = v; 191 struct inpcb *inp; 192 int s; 193 194 s = splnet(); 195 inp = pool_get(&inpcb_pool, PR_NOWAIT); 196 splx(s); 197 if (inp == NULL) 198 return (ENOBUFS); 199 memset(inp, 0, sizeof(*inp)); 200 inp->inp_af = AF_INET; 201 inp->inp_table = table; 202 inp->inp_socket = so; 203 inp->inp_errormtu = -1; 204 inp->inp_portalgo = PORTALGO_DEFAULT; 205 inp->inp_bindportonsend = false; 206 #if defined(IPSEC) 207 if (ipsec_enabled) { 208 int error = ipsec_init_pcbpolicy(so, &inp->inp_sp); 209 if (error != 0) { 210 s = splnet(); 211 pool_put(&inpcb_pool, inp); 212 splx(s); 213 return error; 214 } 215 } 216 #endif 217 so->so_pcb = inp; 218 s = splnet(); 219 TAILQ_INSERT_HEAD(&table->inpt_queue, &inp->inp_head, inph_queue); 220 LIST_INSERT_HEAD(INPCBHASH_PORT(table, inp->inp_lport), &inp->inp_head, 221 inph_lhash); 222 in_pcbstate(inp, INP_ATTACHED); 223 splx(s); 224 return (0); 225 } 226 227 static int 228 in_pcbsetport(struct sockaddr_in *sin, struct inpcb *inp, kauth_cred_t cred) 229 { 230 struct inpcbtable *table = inp->inp_table; 231 struct socket *so = inp->inp_socket; 232 u_int16_t *lastport; 233 u_int16_t lport = 0; 234 enum kauth_network_req req; 235 int error; 236 237 if (inp->inp_flags & INP_LOWPORT) { 238 #ifndef IPNOPRIVPORTS 239 req = KAUTH_REQ_NETWORK_BIND_PRIVPORT; 240 #else 241 req = KAUTH_REQ_NETWORK_BIND_PORT; 242 #endif 243 244 lastport = &table->inpt_lastlow; 245 } else { 246 req = KAUTH_REQ_NETWORK_BIND_PORT; 247 248 lastport = &table->inpt_lastport; 249 } 250 251 /* XXX-kauth: KAUTH_REQ_NETWORK_BIND_AUTOASSIGN_{,PRIV}PORT */ 252 error = kauth_authorize_network(cred, KAUTH_NETWORK_BIND, req, so, sin, 253 NULL); 254 if (error) 255 return (EACCES); 256 257 /* 258 * Use RFC6056 randomized port selection 259 */ 260 error = portalgo_randport(&lport, &inp->inp_head, cred); 261 if (error) 262 return error; 263 264 inp->inp_flags |= INP_ANONPORT; 265 *lastport = lport; 266 lport = htons(lport); 267 inp->inp_lport = lport; 268 in_pcbstate(inp, INP_BOUND); 269 270 return (0); 271 } 272 273 static int 274 in_pcbbind_addr(struct inpcb *inp, struct sockaddr_in *sin, kauth_cred_t cred) 275 { 276 if (sin->sin_family != AF_INET) 277 return (EAFNOSUPPORT); 278 279 if (IN_MULTICAST(sin->sin_addr.s_addr)) { 280 /* Always succeed; port reuse handled in in_pcbbind_port(). */ 281 } else if (!in_nullhost(sin->sin_addr)) { 282 struct in_ifaddr *ia = NULL; 283 284 INADDR_TO_IA(sin->sin_addr, ia); 285 /* check for broadcast addresses */ 286 if (ia == NULL) 287 ia = ifatoia(ifa_ifwithaddr(sintosa(sin))); 288 if (ia == NULL) 289 return (EADDRNOTAVAIL); 290 } 291 292 inp->inp_laddr = sin->sin_addr; 293 294 return (0); 295 } 296 297 static int 298 in_pcbbind_port(struct inpcb *inp, struct sockaddr_in *sin, kauth_cred_t cred) 299 { 300 struct inpcbtable *table = inp->inp_table; 301 struct socket *so = inp->inp_socket; 302 int reuseport = (so->so_options & SO_REUSEPORT); 303 int wild = 0, error; 304 305 if (IN_MULTICAST(sin->sin_addr.s_addr)) { 306 /* 307 * Treat SO_REUSEADDR as SO_REUSEPORT for multicast; 308 * allow complete duplication of binding if 309 * SO_REUSEPORT is set, or if SO_REUSEADDR is set 310 * and a multicast address is bound on both 311 * new and duplicated sockets. 312 */ 313 if (so->so_options & SO_REUSEADDR) 314 reuseport = SO_REUSEADDR|SO_REUSEPORT; 315 } 316 317 if (sin->sin_port == 0) { 318 error = in_pcbsetport(sin, inp, cred); 319 if (error) 320 return (error); 321 } else { 322 struct inpcb *t; 323 vestigial_inpcb_t vestige; 324 #ifdef INET6 325 struct in6pcb *t6; 326 struct in6_addr mapped; 327 #endif 328 enum kauth_network_req req; 329 330 if ((so->so_options & (SO_REUSEADDR|SO_REUSEPORT)) == 0) 331 wild = 1; 332 333 #ifndef IPNOPRIVPORTS 334 if (ntohs(sin->sin_port) < IPPORT_RESERVED) 335 req = KAUTH_REQ_NETWORK_BIND_PRIVPORT; 336 else 337 #endif /* !IPNOPRIVPORTS */ 338 req = KAUTH_REQ_NETWORK_BIND_PORT; 339 340 error = kauth_authorize_network(cred, KAUTH_NETWORK_BIND, req, 341 so, sin, NULL); 342 if (error) 343 return (EACCES); 344 345 #ifdef INET6 346 memset(&mapped, 0, sizeof(mapped)); 347 mapped.s6_addr16[5] = 0xffff; 348 memcpy(&mapped.s6_addr32[3], &sin->sin_addr, 349 sizeof(mapped.s6_addr32[3])); 350 t6 = in6_pcblookup_port(table, &mapped, sin->sin_port, wild, &vestige); 351 if (t6 && (reuseport & t6->in6p_socket->so_options) == 0) 352 return (EADDRINUSE); 353 if (!t6 && vestige.valid) { 354 if (!!reuseport != !!vestige.reuse_port) { 355 return EADDRINUSE; 356 } 357 } 358 #endif 359 360 /* XXX-kauth */ 361 if (so->so_uidinfo->ui_uid && !IN_MULTICAST(sin->sin_addr.s_addr)) { 362 t = in_pcblookup_port(table, sin->sin_addr, sin->sin_port, 1, &vestige); 363 /* 364 * XXX: investigate ramifications of loosening this 365 * restriction so that as long as both ports have 366 * SO_REUSEPORT allow the bind 367 */ 368 if (t && 369 (!in_nullhost(sin->sin_addr) || 370 !in_nullhost(t->inp_laddr) || 371 (t->inp_socket->so_options & SO_REUSEPORT) == 0) 372 && (so->so_uidinfo->ui_uid != t->inp_socket->so_uidinfo->ui_uid)) { 373 return (EADDRINUSE); 374 } 375 if (!t && vestige.valid) { 376 if ((!in_nullhost(sin->sin_addr) 377 || !in_nullhost(vestige.laddr.v4) 378 || !vestige.reuse_port) 379 && so->so_uidinfo->ui_uid != vestige.uid) { 380 return EADDRINUSE; 381 } 382 } 383 } 384 t = in_pcblookup_port(table, sin->sin_addr, sin->sin_port, wild, &vestige); 385 if (t && (reuseport & t->inp_socket->so_options) == 0) 386 return (EADDRINUSE); 387 if (!t 388 && vestige.valid 389 && !(reuseport && vestige.reuse_port)) 390 return EADDRINUSE; 391 392 inp->inp_lport = sin->sin_port; 393 in_pcbstate(inp, INP_BOUND); 394 } 395 396 LIST_REMOVE(&inp->inp_head, inph_lhash); 397 LIST_INSERT_HEAD(INPCBHASH_PORT(table, inp->inp_lport), &inp->inp_head, 398 inph_lhash); 399 400 return (0); 401 } 402 403 int 404 in_pcbbind(void *v, struct mbuf *nam, struct lwp *l) 405 { 406 struct inpcb *inp = v; 407 struct sockaddr_in *sin = NULL; /* XXXGCC */ 408 struct sockaddr_in lsin; 409 int error; 410 411 if (inp->inp_af != AF_INET) 412 return (EINVAL); 413 414 if (TAILQ_FIRST(&in_ifaddrhead) == 0) 415 return (EADDRNOTAVAIL); 416 if (inp->inp_lport || !in_nullhost(inp->inp_laddr)) 417 return (EINVAL); 418 419 if (nam != NULL) { 420 sin = mtod(nam, struct sockaddr_in *); 421 if (nam->m_len != sizeof (*sin)) 422 return (EINVAL); 423 } else { 424 lsin = *((const struct sockaddr_in *) 425 inp->inp_socket->so_proto->pr_domain->dom_sa_any); 426 sin = &lsin; 427 } 428 429 /* Bind address. */ 430 error = in_pcbbind_addr(inp, sin, l->l_cred); 431 if (error) 432 return (error); 433 434 /* Bind port. */ 435 error = in_pcbbind_port(inp, sin, l->l_cred); 436 if (error) { 437 inp->inp_laddr.s_addr = INADDR_ANY; 438 439 return (error); 440 } 441 442 return (0); 443 } 444 445 /* 446 * Connect from a socket to a specified address. 447 * Both address and port must be specified in argument sin. 448 * If don't have a local address for this socket yet, 449 * then pick one. 450 */ 451 int 452 in_pcbconnect(void *v, struct mbuf *nam, struct lwp *l) 453 { 454 struct inpcb *inp = v; 455 struct in_ifaddr *ia = NULL; 456 struct sockaddr_in *ifaddr = NULL; 457 struct sockaddr_in *sin = mtod(nam, struct sockaddr_in *); 458 vestigial_inpcb_t vestige; 459 int error; 460 461 if (inp->inp_af != AF_INET) 462 return (EINVAL); 463 464 if (nam->m_len != sizeof (*sin)) 465 return (EINVAL); 466 if (sin->sin_family != AF_INET) 467 return (EAFNOSUPPORT); 468 if (sin->sin_port == 0) 469 return (EADDRNOTAVAIL); 470 471 if (IN_MULTICAST(sin->sin_addr.s_addr) && 472 inp->inp_socket->so_type == SOCK_STREAM) 473 return EADDRNOTAVAIL; 474 475 if (TAILQ_FIRST(&in_ifaddrhead) != 0) { 476 /* 477 * If the destination address is INADDR_ANY, 478 * use any local address (likely loopback). 479 * If the supplied address is INADDR_BROADCAST, 480 * use the broadcast address of an interface 481 * which supports broadcast. (loopback does not) 482 */ 483 484 if (in_nullhost(sin->sin_addr)) { 485 sin->sin_addr = 486 TAILQ_FIRST(&in_ifaddrhead)->ia_addr.sin_addr; 487 } else if (sin->sin_addr.s_addr == INADDR_BROADCAST) { 488 TAILQ_FOREACH(ia, &in_ifaddrhead, ia_list) { 489 if (ia->ia_ifp->if_flags & IFF_BROADCAST) { 490 sin->sin_addr = 491 ia->ia_broadaddr.sin_addr; 492 break; 493 } 494 } 495 } 496 } 497 /* 498 * If we haven't bound which network number to use as ours, 499 * we will use the number of the outgoing interface. 500 * This depends on having done a routing lookup, which 501 * we will probably have to do anyway, so we might 502 * as well do it now. On the other hand if we are 503 * sending to multiple destinations we may have already 504 * done the lookup, so see if we can use the route 505 * from before. In any case, we only 506 * chose a port number once, even if sending to multiple 507 * destinations. 508 */ 509 if (in_nullhost(inp->inp_laddr)) { 510 int xerror; 511 ifaddr = in_selectsrc(sin, &inp->inp_route, 512 inp->inp_socket->so_options, inp->inp_moptions, &xerror); 513 if (ifaddr == NULL) { 514 if (xerror == 0) 515 xerror = EADDRNOTAVAIL; 516 return xerror; 517 } 518 INADDR_TO_IA(ifaddr->sin_addr, ia); 519 if (ia == NULL) 520 return (EADDRNOTAVAIL); 521 } 522 if (in_pcblookup_connect(inp->inp_table, sin->sin_addr, sin->sin_port, 523 !in_nullhost(inp->inp_laddr) ? inp->inp_laddr : ifaddr->sin_addr, 524 inp->inp_lport, &vestige) != 0 525 || vestige.valid) 526 return (EADDRINUSE); 527 if (in_nullhost(inp->inp_laddr)) { 528 if (inp->inp_lport == 0) { 529 error = in_pcbbind(inp, NULL, l); 530 /* 531 * This used to ignore the return value 532 * completely, but we need to check for 533 * ephemeral port shortage. 534 * And attempts to request low ports if not root. 535 */ 536 if (error != 0) 537 return (error); 538 } 539 inp->inp_laddr = ifaddr->sin_addr; 540 } 541 inp->inp_faddr = sin->sin_addr; 542 inp->inp_fport = sin->sin_port; 543 544 /* Late bind, if needed */ 545 if (inp->inp_bindportonsend) { 546 struct sockaddr_in lsin = *((const struct sockaddr_in *) 547 inp->inp_socket->so_proto->pr_domain->dom_sa_any); 548 lsin.sin_addr = inp->inp_laddr; 549 lsin.sin_port = 0; 550 551 if ((error = in_pcbbind_port(inp, &lsin, l->l_cred)) != 0) 552 return error; 553 } 554 555 in_pcbstate(inp, INP_CONNECTED); 556 #if defined(IPSEC) 557 if (ipsec_enabled && inp->inp_socket->so_type == SOCK_STREAM) 558 ipsec_pcbconn(inp->inp_sp); 559 #endif 560 return (0); 561 } 562 563 void 564 in_pcbdisconnect(void *v) 565 { 566 struct inpcb *inp = v; 567 568 if (inp->inp_af != AF_INET) 569 return; 570 571 inp->inp_faddr = zeroin_addr; 572 inp->inp_fport = 0; 573 in_pcbstate(inp, INP_BOUND); 574 #if defined(IPSEC) 575 if (ipsec_enabled) 576 ipsec_pcbdisconn(inp->inp_sp); 577 #endif 578 if (inp->inp_socket->so_state & SS_NOFDREF) 579 in_pcbdetach(inp); 580 } 581 582 void 583 in_pcbdetach(void *v) 584 { 585 struct inpcb *inp = v; 586 struct socket *so = inp->inp_socket; 587 int s; 588 589 if (inp->inp_af != AF_INET) 590 return; 591 592 #if defined(IPSEC) 593 if (ipsec_enabled) 594 ipsec4_delete_pcbpolicy(inp); 595 #endif 596 so->so_pcb = NULL; 597 598 s = splnet(); 599 in_pcbstate(inp, INP_ATTACHED); 600 LIST_REMOVE(&inp->inp_head, inph_lhash); 601 TAILQ_REMOVE(&inp->inp_table->inpt_queue, &inp->inp_head, inph_queue); 602 splx(s); 603 604 if (inp->inp_options) { 605 m_free(inp->inp_options); 606 } 607 rtcache_free(&inp->inp_route); 608 ip_freemoptions(inp->inp_moptions); 609 sofree(so); /* drops the socket's lock */ 610 611 pool_put(&inpcb_pool, inp); 612 mutex_enter(softnet_lock); /* reacquire the softnet_lock */ 613 } 614 615 void 616 in_setsockaddr(struct inpcb *inp, struct mbuf *nam) 617 { 618 struct sockaddr_in *sin; 619 620 if (inp->inp_af != AF_INET) 621 return; 622 623 sin = mtod(nam, struct sockaddr_in *); 624 sockaddr_in_init(sin, &inp->inp_laddr, inp->inp_lport); 625 nam->m_len = sin->sin_len; 626 } 627 628 void 629 in_setpeeraddr(struct inpcb *inp, struct mbuf *nam) 630 { 631 struct sockaddr_in *sin; 632 633 if (inp->inp_af != AF_INET) 634 return; 635 636 sin = mtod(nam, struct sockaddr_in *); 637 sockaddr_in_init(sin, &inp->inp_faddr, inp->inp_fport); 638 nam->m_len = sin->sin_len; 639 } 640 641 /* 642 * Pass some notification to all connections of a protocol 643 * associated with address dst. The local address and/or port numbers 644 * may be specified to limit the search. The "usual action" will be 645 * taken, depending on the ctlinput cmd. The caller must filter any 646 * cmds that are uninteresting (e.g., no error in the map). 647 * Call the protocol specific routine (if any) to report 648 * any errors for each matching socket. 649 * 650 * Must be called at splsoftnet. 651 */ 652 int 653 in_pcbnotify(struct inpcbtable *table, struct in_addr faddr, u_int fport_arg, 654 struct in_addr laddr, u_int lport_arg, int errno, 655 void (*notify)(struct inpcb *, int)) 656 { 657 struct inpcbhead *head; 658 struct inpcb *inp, *ninp; 659 u_int16_t fport = fport_arg, lport = lport_arg; 660 int nmatch; 661 662 if (in_nullhost(faddr) || notify == 0) 663 return (0); 664 665 nmatch = 0; 666 head = INPCBHASH_CONNECT(table, faddr, fport, laddr, lport); 667 for (inp = (struct inpcb *)LIST_FIRST(head); inp != NULL; inp = ninp) { 668 ninp = (struct inpcb *)LIST_NEXT(inp, inp_hash); 669 if (inp->inp_af != AF_INET) 670 continue; 671 if (in_hosteq(inp->inp_faddr, faddr) && 672 inp->inp_fport == fport && 673 inp->inp_lport == lport && 674 in_hosteq(inp->inp_laddr, laddr)) { 675 (*notify)(inp, errno); 676 nmatch++; 677 } 678 } 679 return (nmatch); 680 } 681 682 void 683 in_pcbnotifyall(struct inpcbtable *table, struct in_addr faddr, int errno, 684 void (*notify)(struct inpcb *, int)) 685 { 686 struct inpcb_hdr *inph, *ninph; 687 688 if (in_nullhost(faddr) || notify == 0) 689 return; 690 691 TAILQ_FOREACH_SAFE(inph, &table->inpt_queue, inph_queue, ninph) { 692 struct inpcb *inp = (struct inpcb *)inph; 693 if (inp->inp_af != AF_INET) 694 continue; 695 if (in_hosteq(inp->inp_faddr, faddr)) 696 (*notify)(inp, errno); 697 } 698 } 699 700 void 701 in_purgeifmcast(struct ip_moptions *imo, struct ifnet *ifp) 702 { 703 int i, gap; 704 705 if (imo == NULL) 706 return; 707 708 /* 709 * Unselect the outgoing interface if it is being 710 * detached. 711 */ 712 if (imo->imo_multicast_ifp == ifp) 713 imo->imo_multicast_ifp = NULL; 714 715 /* 716 * Drop multicast group membership if we joined 717 * through the interface being detached. 718 */ 719 for (i = 0, gap = 0; i < imo->imo_num_memberships; i++) { 720 if (imo->imo_membership[i]->inm_ifp == ifp) { 721 in_delmulti(imo->imo_membership[i]); 722 gap++; 723 } else if (gap != 0) 724 imo->imo_membership[i - gap] = imo->imo_membership[i]; 725 } 726 imo->imo_num_memberships -= gap; 727 } 728 729 void 730 in_pcbpurgeif0(struct inpcbtable *table, struct ifnet *ifp) 731 { 732 struct inpcb_hdr *inph, *ninph; 733 734 TAILQ_FOREACH_SAFE(inph, &table->inpt_queue, inph_queue, ninph) { 735 struct inpcb *inp = (struct inpcb *)inph; 736 if (inp->inp_af != AF_INET) 737 continue; 738 in_purgeifmcast(inp->inp_moptions, ifp); 739 } 740 } 741 742 void 743 in_pcbpurgeif(struct inpcbtable *table, struct ifnet *ifp) 744 { 745 struct rtentry *rt; 746 struct inpcb_hdr *inph, *ninph; 747 748 TAILQ_FOREACH_SAFE(inph, &table->inpt_queue, inph_queue, ninph) { 749 struct inpcb *inp = (struct inpcb *)inph; 750 if (inp->inp_af != AF_INET) 751 continue; 752 if ((rt = rtcache_validate(&inp->inp_route)) != NULL && 753 rt->rt_ifp == ifp) 754 in_rtchange(inp, 0); 755 } 756 } 757 758 /* 759 * Check for alternatives when higher level complains 760 * about service problems. For now, invalidate cached 761 * routing information. If the route was created dynamically 762 * (by a redirect), time to try a default gateway again. 763 */ 764 void 765 in_losing(struct inpcb *inp) 766 { 767 struct rtentry *rt; 768 struct rt_addrinfo info; 769 770 if (inp->inp_af != AF_INET) 771 return; 772 773 if ((rt = rtcache_validate(&inp->inp_route)) == NULL) 774 return; 775 776 memset(&info, 0, sizeof(info)); 777 info.rti_info[RTAX_DST] = rtcache_getdst(&inp->inp_route); 778 info.rti_info[RTAX_GATEWAY] = rt->rt_gateway; 779 info.rti_info[RTAX_NETMASK] = rt_mask(rt); 780 rt_missmsg(RTM_LOSING, &info, rt->rt_flags, 0); 781 if (rt->rt_flags & RTF_DYNAMIC) 782 (void) rtrequest(RTM_DELETE, rt_getkey(rt), 783 rt->rt_gateway, rt_mask(rt), rt->rt_flags, 784 NULL); 785 /* 786 * A new route can be allocated 787 * the next time output is attempted. 788 */ 789 rtcache_free(&inp->inp_route); 790 } 791 792 /* 793 * After a routing change, flush old routing. A new route can be 794 * allocated the next time output is attempted. 795 */ 796 void 797 in_rtchange(struct inpcb *inp, int errno) 798 { 799 800 if (inp->inp_af != AF_INET) 801 return; 802 803 rtcache_free(&inp->inp_route); 804 805 /* XXX SHOULD NOTIFY HIGHER-LEVEL PROTOCOLS */ 806 } 807 808 struct inpcb * 809 in_pcblookup_port(struct inpcbtable *table, struct in_addr laddr, 810 u_int lport_arg, int lookup_wildcard, vestigial_inpcb_t *vp) 811 { 812 struct inpcbhead *head; 813 struct inpcb_hdr *inph; 814 struct inpcb *match = NULL; 815 int matchwild = 3; 816 int wildcard; 817 u_int16_t lport = lport_arg; 818 819 if (vp) 820 vp->valid = 0; 821 822 head = INPCBHASH_PORT(table, lport); 823 LIST_FOREACH(inph, head, inph_lhash) { 824 struct inpcb * const inp = (struct inpcb *)inph; 825 826 if (inp->inp_af != AF_INET) 827 continue; 828 if (inp->inp_lport != lport) 829 continue; 830 /* 831 * check if inp's faddr and laddr match with ours. 832 * our faddr is considered null. 833 * count the number of wildcard matches. (0 - 2) 834 * 835 * null null match 836 * A null wildcard match 837 * null B wildcard match 838 * A B non match 839 * A A match 840 */ 841 wildcard = 0; 842 if (!in_nullhost(inp->inp_faddr)) 843 wildcard++; 844 if (in_nullhost(inp->inp_laddr)) { 845 if (!in_nullhost(laddr)) 846 wildcard++; 847 } else { 848 if (in_nullhost(laddr)) 849 wildcard++; 850 else { 851 if (!in_hosteq(inp->inp_laddr, laddr)) 852 continue; 853 } 854 } 855 if (wildcard && !lookup_wildcard) 856 continue; 857 /* 858 * prefer an address with less wildcards. 859 */ 860 if (wildcard < matchwild) { 861 match = inp; 862 matchwild = wildcard; 863 if (matchwild == 0) 864 break; 865 } 866 } 867 if (match && matchwild == 0) 868 return match; 869 870 if (vp && table->vestige) { 871 void *state = (*table->vestige->init_ports4)(laddr, lport_arg, lookup_wildcard); 872 vestigial_inpcb_t better; 873 874 while (table->vestige 875 && (*table->vestige->next_port4)(state, vp)) { 876 877 if (vp->lport != lport) 878 continue; 879 wildcard = 0; 880 if (!in_nullhost(vp->faddr.v4)) 881 wildcard++; 882 if (in_nullhost(vp->laddr.v4)) { 883 if (!in_nullhost(laddr)) 884 wildcard++; 885 } else { 886 if (in_nullhost(laddr)) 887 wildcard++; 888 else { 889 if (!in_hosteq(vp->laddr.v4, laddr)) 890 continue; 891 } 892 } 893 if (wildcard && !lookup_wildcard) 894 continue; 895 if (wildcard < matchwild) { 896 better = *vp; 897 match = (void*)&better; 898 899 matchwild = wildcard; 900 if (matchwild == 0) 901 break; 902 } 903 } 904 905 if (match) { 906 if (match != (void*)&better) 907 return match; 908 else { 909 *vp = better; 910 return 0; 911 } 912 } 913 } 914 915 return (match); 916 } 917 918 #ifdef DIAGNOSTIC 919 int in_pcbnotifymiss = 0; 920 #endif 921 922 struct inpcb * 923 in_pcblookup_connect(struct inpcbtable *table, 924 struct in_addr faddr, u_int fport_arg, 925 struct in_addr laddr, u_int lport_arg, 926 vestigial_inpcb_t *vp) 927 { 928 struct inpcbhead *head; 929 struct inpcb_hdr *inph; 930 struct inpcb *inp; 931 u_int16_t fport = fport_arg, lport = lport_arg; 932 933 if (vp) 934 vp->valid = 0; 935 936 head = INPCBHASH_CONNECT(table, faddr, fport, laddr, lport); 937 LIST_FOREACH(inph, head, inph_hash) { 938 inp = (struct inpcb *)inph; 939 if (inp->inp_af != AF_INET) 940 continue; 941 942 if (in_hosteq(inp->inp_faddr, faddr) && 943 inp->inp_fport == fport && 944 inp->inp_lport == lport && 945 in_hosteq(inp->inp_laddr, laddr)) 946 goto out; 947 } 948 if (vp && table->vestige) { 949 if ((*table->vestige->lookup4)(faddr, fport_arg, 950 laddr, lport_arg, vp)) 951 return 0; 952 } 953 954 #ifdef DIAGNOSTIC 955 if (in_pcbnotifymiss) { 956 printf("in_pcblookup_connect: faddr=%08x fport=%d laddr=%08x lport=%d\n", 957 ntohl(faddr.s_addr), ntohs(fport), 958 ntohl(laddr.s_addr), ntohs(lport)); 959 } 960 #endif 961 return (0); 962 963 out: 964 /* Move this PCB to the head of hash chain. */ 965 inph = &inp->inp_head; 966 if (inph != LIST_FIRST(head)) { 967 LIST_REMOVE(inph, inph_hash); 968 LIST_INSERT_HEAD(head, inph, inph_hash); 969 } 970 return (inp); 971 } 972 973 struct inpcb * 974 in_pcblookup_bind(struct inpcbtable *table, 975 struct in_addr laddr, u_int lport_arg) 976 { 977 struct inpcbhead *head; 978 struct inpcb_hdr *inph; 979 struct inpcb *inp; 980 u_int16_t lport = lport_arg; 981 982 head = INPCBHASH_BIND(table, laddr, lport); 983 LIST_FOREACH(inph, head, inph_hash) { 984 inp = (struct inpcb *)inph; 985 if (inp->inp_af != AF_INET) 986 continue; 987 988 if (inp->inp_lport == lport && 989 in_hosteq(inp->inp_laddr, laddr)) 990 goto out; 991 } 992 head = INPCBHASH_BIND(table, zeroin_addr, lport); 993 LIST_FOREACH(inph, head, inph_hash) { 994 inp = (struct inpcb *)inph; 995 if (inp->inp_af != AF_INET) 996 continue; 997 998 if (inp->inp_lport == lport && 999 in_hosteq(inp->inp_laddr, zeroin_addr)) 1000 goto out; 1001 } 1002 #ifdef DIAGNOSTIC 1003 if (in_pcbnotifymiss) { 1004 printf("in_pcblookup_bind: laddr=%08x lport=%d\n", 1005 ntohl(laddr.s_addr), ntohs(lport)); 1006 } 1007 #endif 1008 return (0); 1009 1010 out: 1011 /* Move this PCB to the head of hash chain. */ 1012 inph = &inp->inp_head; 1013 if (inph != LIST_FIRST(head)) { 1014 LIST_REMOVE(inph, inph_hash); 1015 LIST_INSERT_HEAD(head, inph, inph_hash); 1016 } 1017 return (inp); 1018 } 1019 1020 void 1021 in_pcbstate(struct inpcb *inp, int state) 1022 { 1023 1024 if (inp->inp_af != AF_INET) 1025 return; 1026 1027 if (inp->inp_state > INP_ATTACHED) 1028 LIST_REMOVE(&inp->inp_head, inph_hash); 1029 1030 switch (state) { 1031 case INP_BOUND: 1032 LIST_INSERT_HEAD(INPCBHASH_BIND(inp->inp_table, 1033 inp->inp_laddr, inp->inp_lport), &inp->inp_head, 1034 inph_hash); 1035 break; 1036 case INP_CONNECTED: 1037 LIST_INSERT_HEAD(INPCBHASH_CONNECT(inp->inp_table, 1038 inp->inp_faddr, inp->inp_fport, 1039 inp->inp_laddr, inp->inp_lport), &inp->inp_head, 1040 inph_hash); 1041 break; 1042 } 1043 1044 inp->inp_state = state; 1045 } 1046 1047 struct rtentry * 1048 in_pcbrtentry(struct inpcb *inp) 1049 { 1050 struct route *ro; 1051 union { 1052 struct sockaddr dst; 1053 struct sockaddr_in dst4; 1054 } u; 1055 1056 if (inp->inp_af != AF_INET) 1057 return (NULL); 1058 1059 ro = &inp->inp_route; 1060 1061 sockaddr_in_init(&u.dst4, &inp->inp_faddr, 0); 1062 return rtcache_lookup(ro, &u.dst); 1063 } 1064
Indexes created Mon Sep 22 13:09:51 GMT 2025