in_pcb.c revision 1.129.2.1
1 /* $NetBSD: in_pcb.c,v 1.129.2.1 2009/04/28 07:37:22 skrll 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 The NetBSD Foundation, Inc. 34 * All rights reserved. 35 * 36 * This code is derived from software contributed to The NetBSD Foundation 37 * by Public Access Networks Corporation ("Panix"). It was developed under 38 * contract to Panix by Eric Haszlakiewicz and Thor Lancelot Simon. 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, 1991, 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 * @(#)in_pcb.c 8.4 (Berkeley) 5/24/95 91 */ 92 93 #include <sys/cdefs.h> 94 __KERNEL_RCSID(0, "$NetBSD: in_pcb.c,v 1.129.2.1 2009/04/28 07:37:22 skrll Exp $"); 95 96 #include "opt_inet.h" 97 #include "opt_ipsec.h" 98 99 #include <sys/param.h> 100 #include <sys/systm.h> 101 #include <sys/malloc.h> 102 #include <sys/mbuf.h> 103 #include <sys/protosw.h> 104 #include <sys/socket.h> 105 #include <sys/socketvar.h> 106 #include <sys/ioctl.h> 107 #include <sys/errno.h> 108 #include <sys/time.h> 109 #include <sys/once.h> 110 #include <sys/pool.h> 111 #include <sys/proc.h> 112 #include <sys/kauth.h> 113 #include <sys/uidinfo.h> 114 #include <sys/domain.h> 115 116 #include <net/if.h> 117 #include <net/route.h> 118 119 #include <netinet/in.h> 120 #include <netinet/in_systm.h> 121 #include <netinet/ip.h> 122 #include <netinet/in_pcb.h> 123 #include <netinet/in_var.h> 124 #include <netinet/ip_var.h> 125 126 #ifdef INET6 127 #include <netinet/ip6.h> 128 #include <netinet6/ip6_var.h> 129 #include <netinet6/in6_pcb.h> 130 #endif 131 132 #ifdef IPSEC 133 #include <netinet6/ipsec.h> 134 #include <netkey/key.h> 135 #elif FAST_IPSEC 136 #include <netipsec/ipsec.h> 137 #include <netipsec/key.h> 138 #endif /* IPSEC */ 139 140 struct in_addr zeroin_addr; 141 142 #define INPCBHASH_PORT(table, lport) \ 143 &(table)->inpt_porthashtbl[ntohs(lport) & (table)->inpt_porthash] 144 #define INPCBHASH_BIND(table, laddr, lport) \ 145 &(table)->inpt_bindhashtbl[ \ 146 ((ntohl((laddr).s_addr) + ntohs(lport))) & (table)->inpt_bindhash] 147 #define INPCBHASH_CONNECT(table, faddr, fport, laddr, lport) \ 148 &(table)->inpt_connecthashtbl[ \ 149 ((ntohl((faddr).s_addr) + ntohs(fport)) + \ 150 (ntohl((laddr).s_addr) + ntohs(lport))) & (table)->inpt_connecthash] 151 152 int anonportmin = IPPORT_ANONMIN; 153 int anonportmax = IPPORT_ANONMAX; 154 int lowportmin = IPPORT_RESERVEDMIN; 155 int lowportmax = IPPORT_RESERVEDMAX; 156 157 static struct pool inpcb_pool; 158 159 static int 160 inpcb_poolinit(void) 161 { 162 163 pool_init(&inpcb_pool, sizeof(struct inpcb), 0, 0, 0, "inpcbpl", NULL, 164 IPL_NET); 165 return 0; 166 } 167 168 void 169 in_pcbinit(struct inpcbtable *table, int bindhashsize, int connecthashsize) 170 { 171 static ONCE_DECL(control); 172 173 CIRCLEQ_INIT(&table->inpt_queue); 174 table->inpt_porthashtbl = hashinit(bindhashsize, HASH_LIST, true, 175 &table->inpt_porthash); 176 table->inpt_bindhashtbl = hashinit(bindhashsize, HASH_LIST, true, 177 &table->inpt_bindhash); 178 table->inpt_connecthashtbl = hashinit(connecthashsize, HASH_LIST, true, 179 &table->inpt_connecthash); 180 table->inpt_lastlow = IPPORT_RESERVEDMAX; 181 table->inpt_lastport = (u_int16_t)anonportmax; 182 183 RUN_ONCE(&control, inpcb_poolinit); 184 } 185 186 int 187 in_pcballoc(struct socket *so, void *v) 188 { 189 struct inpcbtable *table = v; 190 struct inpcb *inp; 191 int s; 192 #if defined(IPSEC) || defined(FAST_IPSEC) 193 int error; 194 #endif 195 196 s = splnet(); 197 inp = pool_get(&inpcb_pool, PR_NOWAIT); 198 splx(s); 199 if (inp == NULL) 200 return (ENOBUFS); 201 memset((void *)inp, 0, sizeof(*inp)); 202 inp->inp_af = AF_INET; 203 inp->inp_table = table; 204 inp->inp_socket = so; 205 inp->inp_errormtu = -1; 206 #if defined(IPSEC) || defined(FAST_IPSEC) 207 error = ipsec_init_pcbpolicy(so, &inp->inp_sp); 208 if (error != 0) { 209 s = splnet(); 210 pool_put(&inpcb_pool, inp); 211 splx(s); 212 return error; 213 } 214 #endif 215 so->so_pcb = inp; 216 s = splnet(); 217 CIRCLEQ_INSERT_HEAD(&table->inpt_queue, &inp->inp_head, 218 inph_queue); 219 LIST_INSERT_HEAD(INPCBHASH_PORT(table, inp->inp_lport), &inp->inp_head, 220 inph_lhash); 221 in_pcbstate(inp, INP_ATTACHED); 222 splx(s); 223 return (0); 224 } 225 226 static int 227 in_pcbsetport(struct in_addr *laddr, struct inpcb *inp, 228 struct sockaddr_in *sin, kauth_cred_t cred) 229 { 230 struct inpcbtable *table = inp->inp_table; 231 struct socket *so = inp->inp_socket; 232 int cnt; 233 u_int16_t mymin, mymax; 234 u_int16_t *lastport; 235 u_int16_t lport = 0; 236 237 if (inp->inp_flags & INP_LOWPORT) { 238 #ifndef IPNOPRIVPORTS 239 if (kauth_authorize_network(cred, 240 KAUTH_NETWORK_BIND, 241 KAUTH_REQ_NETWORK_BIND_PRIVPORT, so, 242 sin, NULL)) 243 return (EACCES); 244 #endif 245 mymin = lowportmin; 246 mymax = lowportmax; 247 lastport = &table->inpt_lastlow; 248 } else { 249 mymin = anonportmin; 250 mymax = anonportmax; 251 lastport = &table->inpt_lastport; 252 } 253 if (mymin > mymax) { /* sanity check */ 254 u_int16_t swp; 255 256 swp = mymin; 257 mymin = mymax; 258 mymax = swp; 259 } 260 261 lport = *lastport - 1; 262 for (cnt = mymax - mymin + 1; cnt; cnt--, lport--) { 263 if (lport < mymin || lport > mymax) 264 lport = mymax; 265 if (!in_pcblookup_port(table, inp->inp_laddr, 266 htons(lport), 1)) 267 goto found; 268 } 269 270 return (EAGAIN); 271 272 found: 273 inp->inp_flags |= INP_ANONPORT; 274 *lastport = lport; 275 lport = htons(lport); 276 inp->inp_lport = lport; 277 in_pcbstate(inp, INP_BOUND); 278 279 return (0); 280 } 281 282 static int 283 in_pcbbind_addr(struct inpcb *inp, struct sockaddr_in *sin, kauth_cred_t cred) 284 { 285 if (sin->sin_family != AF_INET) 286 return (EAFNOSUPPORT); 287 288 if (!in_nullhost(sin->sin_addr)) { 289 struct in_ifaddr *ia = NULL; 290 291 INADDR_TO_IA(sin->sin_addr, ia); 292 /* check for broadcast addresses */ 293 if (ia == NULL) 294 ia = ifatoia(ifa_ifwithaddr(sintosa(sin))); 295 if (ia == NULL) 296 return (EADDRNOTAVAIL); 297 } 298 299 inp->inp_laddr = sin->sin_addr; 300 301 return (0); 302 } 303 304 static int 305 in_pcbbind_port(struct inpcb *inp, struct sockaddr_in *sin, kauth_cred_t cred) 306 { 307 struct inpcbtable *table = inp->inp_table; 308 struct socket *so = inp->inp_socket; 309 int reuseport = (so->so_options & SO_REUSEPORT); 310 int wild = 0, error; 311 312 if (IN_MULTICAST(sin->sin_addr.s_addr)) { 313 /* 314 * Treat SO_REUSEADDR as SO_REUSEPORT for multicast; 315 * allow complete duplication of binding if 316 * SO_REUSEPORT is set, or if SO_REUSEADDR is set 317 * and a multicast address is bound on both 318 * new and duplicated sockets. 319 */ 320 if (so->so_options & SO_REUSEADDR) 321 reuseport = SO_REUSEADDR|SO_REUSEPORT; 322 } 323 324 if (sin->sin_port == 0) { 325 error = in_pcbsetport(&inp->inp_laddr, inp, sin, cred); 326 if (error) 327 return (error); 328 } else { 329 struct inpcb *t; 330 #ifdef INET6 331 struct in6pcb *t6; 332 struct in6_addr mapped; 333 #endif 334 enum kauth_network_req req; 335 336 if ((so->so_options & (SO_REUSEADDR|SO_REUSEPORT)) == 0) 337 wild = 1; 338 339 #ifndef IPNOPRIVPORTS 340 if (ntohs(sin->sin_port) < IPPORT_RESERVED) 341 req = KAUTH_REQ_NETWORK_BIND_PRIVPORT; 342 else 343 #endif /* !IPNOPRIVPORTS */ 344 req = KAUTH_REQ_NETWORK_BIND_PORT; 345 346 error = kauth_authorize_network(cred, KAUTH_NETWORK_BIND, req, 347 so, sin, NULL); 348 if (error) 349 return (error); 350 351 #ifdef INET6 352 memset(&mapped, 0, sizeof(mapped)); 353 mapped.s6_addr16[5] = 0xffff; 354 memcpy(&mapped.s6_addr32[3], &sin->sin_addr, 355 sizeof(mapped.s6_addr32[3])); 356 t6 = in6_pcblookup_port(table, &mapped, sin->sin_port, wild); 357 if (t6 && (reuseport & t6->in6p_socket->so_options) == 0) 358 return (EADDRINUSE); 359 #endif 360 361 /* XXX-kauth */ 362 if (so->so_uidinfo->ui_uid && !IN_MULTICAST(sin->sin_addr.s_addr)) { 363 t = in_pcblookup_port(table, sin->sin_addr, sin->sin_port, 1); 364 /* 365 * XXX: investigate ramifications of loosening this 366 * restriction so that as long as both ports have 367 * SO_REUSEPORT allow the bind 368 */ 369 if (t && 370 (!in_nullhost(sin->sin_addr) || 371 !in_nullhost(t->inp_laddr) || 372 (t->inp_socket->so_options & SO_REUSEPORT) == 0) 373 && (so->so_uidinfo->ui_uid != t->inp_socket->so_uidinfo->ui_uid)) { 374 return (EADDRINUSE); 375 } 376 } 377 t = in_pcblookup_port(table, sin->sin_addr, sin->sin_port, wild); 378 if (t && (reuseport & t->inp_socket->so_options) == 0) 379 return (EADDRINUSE); 380 381 inp->inp_lport = sin->sin_port; 382 in_pcbstate(inp, INP_BOUND); 383 } 384 385 LIST_REMOVE(&inp->inp_head, inph_lhash); 386 LIST_INSERT_HEAD(INPCBHASH_PORT(table, inp->inp_lport), &inp->inp_head, 387 inph_lhash); 388 389 return (0); 390 } 391 392 int 393 in_pcbbind(void *v, struct mbuf *nam, struct lwp *l) 394 { 395 struct inpcb *inp = v; 396 struct sockaddr_in *sin = NULL; /* XXXGCC */ 397 int error; 398 399 if (inp->inp_af != AF_INET) 400 return (EINVAL); 401 402 if (TAILQ_FIRST(&in_ifaddrhead) == 0) 403 return (EADDRNOTAVAIL); 404 if (inp->inp_lport || !in_nullhost(inp->inp_laddr)) 405 return (EINVAL); 406 407 if (nam != NULL) { 408 sin = mtod(nam, struct sockaddr_in *); 409 if (nam->m_len != sizeof (*sin)) 410 return (EINVAL); 411 } else { 412 sin = (struct sockaddr_in *) 413 __UNCONST(inp->inp_socket->so_proto->pr_domain->dom_sa_any); 414 } 415 416 /* Bind address. */ 417 error = in_pcbbind_addr(inp, sin, l->l_cred); 418 if (error) 419 return (error); 420 421 /* Bind port. */ 422 error = in_pcbbind_port(inp, sin, l->l_cred); 423 if (error) { 424 inp->inp_laddr.s_addr = INADDR_ANY; 425 426 return (error); 427 } 428 429 return (0); 430 } 431 432 /* 433 * Connect from a socket to a specified address. 434 * Both address and port must be specified in argument sin. 435 * If don't have a local address for this socket yet, 436 * then pick one. 437 */ 438 int 439 in_pcbconnect(void *v, struct mbuf *nam, struct lwp *l) 440 { 441 struct inpcb *inp = v; 442 struct in_ifaddr *ia = NULL; 443 struct sockaddr_in *ifaddr = NULL; 444 struct sockaddr_in *sin = mtod(nam, struct sockaddr_in *); 445 int error; 446 447 if (inp->inp_af != AF_INET) 448 return (EINVAL); 449 450 if (nam->m_len != sizeof (*sin)) 451 return (EINVAL); 452 if (sin->sin_family != AF_INET) 453 return (EAFNOSUPPORT); 454 if (sin->sin_port == 0) 455 return (EADDRNOTAVAIL); 456 if (TAILQ_FIRST(&in_ifaddrhead) != 0) { 457 /* 458 * If the destination address is INADDR_ANY, 459 * use any local address (likely loopback). 460 * If the supplied address is INADDR_BROADCAST, 461 * use the broadcast address of an interface 462 * which supports broadcast. (loopback does not) 463 */ 464 465 if (in_nullhost(sin->sin_addr)) { 466 sin->sin_addr = 467 TAILQ_FIRST(&in_ifaddrhead)->ia_addr.sin_addr; 468 } else if (sin->sin_addr.s_addr == INADDR_BROADCAST) { 469 TAILQ_FOREACH(ia, &in_ifaddrhead, ia_list) { 470 if (ia->ia_ifp->if_flags & IFF_BROADCAST) { 471 sin->sin_addr = 472 ia->ia_broadaddr.sin_addr; 473 break; 474 } 475 } 476 } 477 } 478 /* 479 * If we haven't bound which network number to use as ours, 480 * we will use the number of the outgoing interface. 481 * This depends on having done a routing lookup, which 482 * we will probably have to do anyway, so we might 483 * as well do it now. On the other hand if we are 484 * sending to multiple destinations we may have already 485 * done the lookup, so see if we can use the route 486 * from before. In any case, we only 487 * chose a port number once, even if sending to multiple 488 * destinations. 489 */ 490 if (in_nullhost(inp->inp_laddr)) { 491 int xerror; 492 ifaddr = in_selectsrc(sin, &inp->inp_route, 493 inp->inp_socket->so_options, inp->inp_moptions, &xerror); 494 if (ifaddr == NULL) { 495 if (xerror == 0) 496 xerror = EADDRNOTAVAIL; 497 return xerror; 498 } 499 INADDR_TO_IA(ifaddr->sin_addr, ia); 500 if (ia == NULL) 501 return (EADDRNOTAVAIL); 502 } 503 if (in_pcblookup_connect(inp->inp_table, sin->sin_addr, sin->sin_port, 504 !in_nullhost(inp->inp_laddr) ? inp->inp_laddr : ifaddr->sin_addr, 505 inp->inp_lport) != 0) 506 return (EADDRINUSE); 507 if (in_nullhost(inp->inp_laddr)) { 508 if (inp->inp_lport == 0) { 509 error = in_pcbbind(inp, NULL, l); 510 /* 511 * This used to ignore the return value 512 * completely, but we need to check for 513 * ephemeral port shortage. 514 * And attempts to request low ports if not root. 515 */ 516 if (error != 0) 517 return (error); 518 } 519 inp->inp_laddr = ifaddr->sin_addr; 520 } 521 inp->inp_faddr = sin->sin_addr; 522 inp->inp_fport = sin->sin_port; 523 in_pcbstate(inp, INP_CONNECTED); 524 #if defined(IPSEC) || defined(FAST_IPSEC) 525 if (inp->inp_socket->so_type == SOCK_STREAM) 526 ipsec_pcbconn(inp->inp_sp); 527 #endif 528 return (0); 529 } 530 531 void 532 in_pcbdisconnect(void *v) 533 { 534 struct inpcb *inp = v; 535 536 if (inp->inp_af != AF_INET) 537 return; 538 539 inp->inp_faddr = zeroin_addr; 540 inp->inp_fport = 0; 541 in_pcbstate(inp, INP_BOUND); 542 #if defined(IPSEC) || defined(FAST_IPSEC) 543 ipsec_pcbdisconn(inp->inp_sp); 544 #endif 545 if (inp->inp_socket->so_state & SS_NOFDREF) 546 in_pcbdetach(inp); 547 } 548 549 void 550 in_pcbdetach(void *v) 551 { 552 struct inpcb *inp = v; 553 struct socket *so = inp->inp_socket; 554 int s; 555 556 if (inp->inp_af != AF_INET) 557 return; 558 559 #if defined(IPSEC) || defined(FAST_IPSEC) 560 ipsec4_delete_pcbpolicy(inp); 561 #endif /*IPSEC*/ 562 so->so_pcb = 0; 563 if (inp->inp_options) 564 (void)m_free(inp->inp_options); 565 rtcache_free(&inp->inp_route); 566 ip_freemoptions(inp->inp_moptions); 567 s = splnet(); 568 in_pcbstate(inp, INP_ATTACHED); 569 LIST_REMOVE(&inp->inp_head, inph_lhash); 570 CIRCLEQ_REMOVE(&inp->inp_table->inpt_queue, &inp->inp_head, 571 inph_queue); 572 pool_put(&inpcb_pool, inp); 573 splx(s); 574 sofree(so); /* drops the socket's lock */ 575 mutex_enter(softnet_lock); /* reacquire the softnet_lock */ 576 } 577 578 void 579 in_setsockaddr(struct inpcb *inp, struct mbuf *nam) 580 { 581 struct sockaddr_in *sin; 582 583 if (inp->inp_af != AF_INET) 584 return; 585 586 sin = mtod(nam, struct sockaddr_in *); 587 sockaddr_in_init(sin, &inp->inp_laddr, inp->inp_lport); 588 nam->m_len = sin->sin_len; 589 } 590 591 void 592 in_setpeeraddr(struct inpcb *inp, struct mbuf *nam) 593 { 594 struct sockaddr_in *sin; 595 596 if (inp->inp_af != AF_INET) 597 return; 598 599 sin = mtod(nam, struct sockaddr_in *); 600 sockaddr_in_init(sin, &inp->inp_faddr, inp->inp_fport); 601 nam->m_len = sin->sin_len; 602 } 603 604 /* 605 * Pass some notification to all connections of a protocol 606 * associated with address dst. The local address and/or port numbers 607 * may be specified to limit the search. The "usual action" will be 608 * taken, depending on the ctlinput cmd. The caller must filter any 609 * cmds that are uninteresting (e.g., no error in the map). 610 * Call the protocol specific routine (if any) to report 611 * any errors for each matching socket. 612 * 613 * Must be called at splsoftnet. 614 */ 615 int 616 in_pcbnotify(struct inpcbtable *table, struct in_addr faddr, u_int fport_arg, 617 struct in_addr laddr, u_int lport_arg, int errno, 618 void (*notify)(struct inpcb *, int)) 619 { 620 struct inpcbhead *head; 621 struct inpcb *inp, *ninp; 622 u_int16_t fport = fport_arg, lport = lport_arg; 623 int nmatch; 624 625 if (in_nullhost(faddr) || notify == 0) 626 return (0); 627 628 nmatch = 0; 629 head = INPCBHASH_CONNECT(table, faddr, fport, laddr, lport); 630 for (inp = (struct inpcb *)LIST_FIRST(head); inp != NULL; inp = ninp) { 631 ninp = (struct inpcb *)LIST_NEXT(inp, inp_hash); 632 if (inp->inp_af != AF_INET) 633 continue; 634 if (in_hosteq(inp->inp_faddr, faddr) && 635 inp->inp_fport == fport && 636 inp->inp_lport == lport && 637 in_hosteq(inp->inp_laddr, laddr)) { 638 (*notify)(inp, errno); 639 nmatch++; 640 } 641 } 642 return (nmatch); 643 } 644 645 void 646 in_pcbnotifyall(struct inpcbtable *table, struct in_addr faddr, int errno, 647 void (*notify)(struct inpcb *, int)) 648 { 649 struct inpcb *inp, *ninp; 650 651 if (in_nullhost(faddr) || notify == 0) 652 return; 653 654 for (inp = (struct inpcb *)CIRCLEQ_FIRST(&table->inpt_queue); 655 inp != (void *)&table->inpt_queue; 656 inp = ninp) { 657 ninp = (struct inpcb *)CIRCLEQ_NEXT(inp, inp_queue); 658 if (inp->inp_af != AF_INET) 659 continue; 660 if (in_hosteq(inp->inp_faddr, faddr)) 661 (*notify)(inp, errno); 662 } 663 } 664 665 void 666 in_pcbpurgeif0(struct inpcbtable *table, struct ifnet *ifp) 667 { 668 struct inpcb *inp, *ninp; 669 struct ip_moptions *imo; 670 int i, gap; 671 672 for (inp = (struct inpcb *)CIRCLEQ_FIRST(&table->inpt_queue); 673 inp != (void *)&table->inpt_queue; 674 inp = ninp) { 675 ninp = (struct inpcb *)CIRCLEQ_NEXT(inp, inp_queue); 676 if (inp->inp_af != AF_INET) 677 continue; 678 imo = inp->inp_moptions; 679 if (imo != NULL) { 680 /* 681 * Unselect the outgoing interface if it is being 682 * detached. 683 */ 684 if (imo->imo_multicast_ifp == ifp) 685 imo->imo_multicast_ifp = NULL; 686 687 /* 688 * Drop multicast group membership if we joined 689 * through the interface being detached. 690 */ 691 for (i = 0, gap = 0; i < imo->imo_num_memberships; 692 i++) { 693 if (imo->imo_membership[i]->inm_ifp == ifp) { 694 in_delmulti(imo->imo_membership[i]); 695 gap++; 696 } else if (gap != 0) 697 imo->imo_membership[i - gap] = 698 imo->imo_membership[i]; 699 } 700 imo->imo_num_memberships -= gap; 701 } 702 } 703 } 704 705 void 706 in_pcbpurgeif(struct inpcbtable *table, struct ifnet *ifp) 707 { 708 struct rtentry *rt; 709 struct inpcb *inp, *ninp; 710 711 for (inp = (struct inpcb *)CIRCLEQ_FIRST(&table->inpt_queue); 712 inp != (void *)&table->inpt_queue; 713 inp = ninp) { 714 ninp = (struct inpcb *)CIRCLEQ_NEXT(inp, inp_queue); 715 if (inp->inp_af != AF_INET) 716 continue; 717 if ((rt = rtcache_validate(&inp->inp_route)) != NULL && 718 rt->rt_ifp == ifp) 719 in_rtchange(inp, 0); 720 } 721 } 722 723 /* 724 * Check for alternatives when higher level complains 725 * about service problems. For now, invalidate cached 726 * routing information. If the route was created dynamically 727 * (by a redirect), time to try a default gateway again. 728 */ 729 void 730 in_losing(struct inpcb *inp) 731 { 732 struct rtentry *rt; 733 struct rt_addrinfo info; 734 735 if (inp->inp_af != AF_INET) 736 return; 737 738 if ((rt = rtcache_validate(&inp->inp_route)) == NULL) 739 return; 740 741 memset(&info, 0, sizeof(info)); 742 info.rti_info[RTAX_DST] = rtcache_getdst(&inp->inp_route); 743 info.rti_info[RTAX_GATEWAY] = rt->rt_gateway; 744 info.rti_info[RTAX_NETMASK] = rt_mask(rt); 745 rt_missmsg(RTM_LOSING, &info, rt->rt_flags, 0); 746 if (rt->rt_flags & RTF_DYNAMIC) 747 (void) rtrequest(RTM_DELETE, rt_getkey(rt), 748 rt->rt_gateway, rt_mask(rt), rt->rt_flags, 749 NULL); 750 /* 751 * A new route can be allocated 752 * the next time output is attempted. 753 */ 754 rtcache_free(&inp->inp_route); 755 } 756 757 /* 758 * After a routing change, flush old routing. A new route can be 759 * allocated the next time output is attempted. 760 */ 761 void 762 in_rtchange(struct inpcb *inp, int errno) 763 { 764 765 if (inp->inp_af != AF_INET) 766 return; 767 768 rtcache_free(&inp->inp_route); 769 770 /* XXX SHOULD NOTIFY HIGHER-LEVEL PROTOCOLS */ 771 } 772 773 struct inpcb * 774 in_pcblookup_port(struct inpcbtable *table, struct in_addr laddr, 775 u_int lport_arg, int lookup_wildcard) 776 { 777 struct inpcbhead *head; 778 struct inpcb_hdr *inph; 779 struct inpcb *inp, *match = 0; 780 int matchwild = 3, wildcard; 781 u_int16_t lport = lport_arg; 782 783 head = INPCBHASH_PORT(table, lport); 784 LIST_FOREACH(inph, head, inph_lhash) { 785 inp = (struct inpcb *)inph; 786 if (inp->inp_af != AF_INET) 787 continue; 788 789 if (inp->inp_lport != lport) 790 continue; 791 wildcard = 0; 792 if (!in_nullhost(inp->inp_faddr)) 793 wildcard++; 794 if (in_nullhost(inp->inp_laddr)) { 795 if (!in_nullhost(laddr)) 796 wildcard++; 797 } else { 798 if (in_nullhost(laddr)) 799 wildcard++; 800 else { 801 if (!in_hosteq(inp->inp_laddr, laddr)) 802 continue; 803 } 804 } 805 if (wildcard && !lookup_wildcard) 806 continue; 807 if (wildcard < matchwild) { 808 match = inp; 809 matchwild = wildcard; 810 if (matchwild == 0) 811 break; 812 } 813 } 814 return (match); 815 } 816 817 #ifdef DIAGNOSTIC 818 int in_pcbnotifymiss = 0; 819 #endif 820 821 struct inpcb * 822 in_pcblookup_connect(struct inpcbtable *table, 823 struct in_addr faddr, u_int fport_arg, 824 struct in_addr laddr, u_int lport_arg) 825 { 826 struct inpcbhead *head; 827 struct inpcb_hdr *inph; 828 struct inpcb *inp; 829 u_int16_t fport = fport_arg, lport = lport_arg; 830 831 head = INPCBHASH_CONNECT(table, faddr, fport, laddr, lport); 832 LIST_FOREACH(inph, head, inph_hash) { 833 inp = (struct inpcb *)inph; 834 if (inp->inp_af != AF_INET) 835 continue; 836 837 if (in_hosteq(inp->inp_faddr, faddr) && 838 inp->inp_fport == fport && 839 inp->inp_lport == lport && 840 in_hosteq(inp->inp_laddr, laddr)) 841 goto out; 842 } 843 #ifdef DIAGNOSTIC 844 if (in_pcbnotifymiss) { 845 printf("in_pcblookup_connect: faddr=%08x fport=%d laddr=%08x lport=%d\n", 846 ntohl(faddr.s_addr), ntohs(fport), 847 ntohl(laddr.s_addr), ntohs(lport)); 848 } 849 #endif 850 return (0); 851 852 out: 853 /* Move this PCB to the head of hash chain. */ 854 inph = &inp->inp_head; 855 if (inph != LIST_FIRST(head)) { 856 LIST_REMOVE(inph, inph_hash); 857 LIST_INSERT_HEAD(head, inph, inph_hash); 858 } 859 return (inp); 860 } 861 862 struct inpcb * 863 in_pcblookup_bind(struct inpcbtable *table, 864 struct in_addr laddr, u_int lport_arg) 865 { 866 struct inpcbhead *head; 867 struct inpcb_hdr *inph; 868 struct inpcb *inp; 869 u_int16_t lport = lport_arg; 870 871 head = INPCBHASH_BIND(table, laddr, lport); 872 LIST_FOREACH(inph, head, inph_hash) { 873 inp = (struct inpcb *)inph; 874 if (inp->inp_af != AF_INET) 875 continue; 876 877 if (inp->inp_lport == lport && 878 in_hosteq(inp->inp_laddr, laddr)) 879 goto out; 880 } 881 head = INPCBHASH_BIND(table, zeroin_addr, lport); 882 LIST_FOREACH(inph, head, inph_hash) { 883 inp = (struct inpcb *)inph; 884 if (inp->inp_af != AF_INET) 885 continue; 886 887 if (inp->inp_lport == lport && 888 in_hosteq(inp->inp_laddr, zeroin_addr)) 889 goto out; 890 } 891 #ifdef DIAGNOSTIC 892 if (in_pcbnotifymiss) { 893 printf("in_pcblookup_bind: laddr=%08x lport=%d\n", 894 ntohl(laddr.s_addr), ntohs(lport)); 895 } 896 #endif 897 return (0); 898 899 out: 900 /* Move this PCB to the head of hash chain. */ 901 inph = &inp->inp_head; 902 if (inph != LIST_FIRST(head)) { 903 LIST_REMOVE(inph, inph_hash); 904 LIST_INSERT_HEAD(head, inph, inph_hash); 905 } 906 return (inp); 907 } 908 909 void 910 in_pcbstate(struct inpcb *inp, int state) 911 { 912 913 if (inp->inp_af != AF_INET) 914 return; 915 916 if (inp->inp_state > INP_ATTACHED) 917 LIST_REMOVE(&inp->inp_head, inph_hash); 918 919 switch (state) { 920 case INP_BOUND: 921 LIST_INSERT_HEAD(INPCBHASH_BIND(inp->inp_table, 922 inp->inp_laddr, inp->inp_lport), &inp->inp_head, 923 inph_hash); 924 break; 925 case INP_CONNECTED: 926 LIST_INSERT_HEAD(INPCBHASH_CONNECT(inp->inp_table, 927 inp->inp_faddr, inp->inp_fport, 928 inp->inp_laddr, inp->inp_lport), &inp->inp_head, 929 inph_hash); 930 break; 931 } 932 933 inp->inp_state = state; 934 } 935 936 struct rtentry * 937 in_pcbrtentry(struct inpcb *inp) 938 { 939 struct route *ro; 940 union { 941 struct sockaddr dst; 942 struct sockaddr_in dst4; 943 } u; 944 945 if (inp->inp_af != AF_INET) 946 return (NULL); 947 948 ro = &inp->inp_route; 949 950 sockaddr_in_init(&u.dst4, &inp->inp_faddr, 0); 951 return rtcache_lookup(ro, &u.dst); 952 } 953 954 struct sockaddr_in * 955 in_selectsrc(struct sockaddr_in *sin, struct route *ro, 956 int soopts, struct ip_moptions *mopts, int *errorp) 957 { 958 struct rtentry *rt = NULL; 959 struct in_ifaddr *ia = NULL; 960 961 /* 962 * If route is known or can be allocated now, take the 963 * source address from the interface. Otherwise, punt. 964 */ 965 if ((soopts & SO_DONTROUTE) != 0) 966 rtcache_free(ro); 967 else { 968 union { 969 struct sockaddr dst; 970 struct sockaddr_in dst4; 971 } u; 972 973 sockaddr_in_init(&u.dst4, &sin->sin_addr, 0); 974 rt = rtcache_lookup(ro, &u.dst); 975 } 976 /* 977 * If we found a route, use the address 978 * corresponding to the outgoing interface 979 * unless it is the loopback (in case a route 980 * to our address on another net goes to loopback). 981 * 982 * XXX Is this still true? Do we care? 983 */ 984 if (rt != NULL && (rt->rt_ifp->if_flags & IFF_LOOPBACK) == 0) 985 ia = ifatoia(rt->rt_ifa); 986 if (ia == NULL) { 987 u_int16_t fport = sin->sin_port; 988 989 sin->sin_port = 0; 990 ia = ifatoia(ifa_ifwithladdr(sintosa(sin))); 991 sin->sin_port = fport; 992 if (ia == NULL) { 993 /* Find 1st non-loopback AF_INET address */ 994 TAILQ_FOREACH(ia, &in_ifaddrhead, ia_list) { 995 if (!(ia->ia_ifp->if_flags & IFF_LOOPBACK)) 996 break; 997 } 998 } 999 if (ia == NULL) { 1000 *errorp = EADDRNOTAVAIL; 1001 return NULL; 1002 } 1003 } 1004 /* 1005 * If the destination address is multicast and an outgoing 1006 * interface has been set as a multicast option, use the 1007 * address of that interface as our source address. 1008 */ 1009 if (IN_MULTICAST(sin->sin_addr.s_addr) && mopts != NULL) { 1010 struct ip_moptions *imo; 1011 struct ifnet *ifp; 1012 1013 imo = mopts; 1014 if (imo->imo_multicast_ifp != NULL) { 1015 ifp = imo->imo_multicast_ifp; 1016 IFP_TO_IA(ifp, ia); /* XXX */ 1017 if (ia == 0) { 1018 *errorp = EADDRNOTAVAIL; 1019 return NULL; 1020 } 1021 } 1022 } 1023 if (ia->ia_ifa.ifa_getifa != NULL) { 1024 ia = ifatoia((*ia->ia_ifa.ifa_getifa)(&ia->ia_ifa, 1025 sintosa(sin))); 1026 } 1027 #ifdef GETIFA_DEBUG 1028 else 1029 printf("%s: missing ifa_getifa\n", __func__); 1030 #endif 1031 return satosin(&ia->ia_addr); 1032 } 1033
Indexes created Wed Oct 15 03:09:54 GMT 2025