in_pcb.c revision 1.103
1 /* $NetBSD: in_pcb.c,v 1.103 2006/07/23 22:06:13 ad 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 * 3. All advertising materials mentioning features or use of this software 49 * must display the following acknowledgement: 50 * This product includes software developed by the NetBSD 51 * Foundation, Inc. and its contributors. 52 * 4. Neither the name of The NetBSD Foundation nor the names of its 53 * contributors may be used to endorse or promote products derived 54 * from this software without specific prior written permission. 55 * 56 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 57 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 58 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 59 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 60 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 61 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 62 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 63 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 64 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 65 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 66 * POSSIBILITY OF SUCH DAMAGE. 67 */ 68 69 /* 70 * Copyright (c) 1982, 1986, 1991, 1993, 1995 71 * The Regents of the University of California. All rights reserved. 72 * 73 * Redistribution and use in source and binary forms, with or without 74 * modification, are permitted provided that the following conditions 75 * are met: 76 * 1. Redistributions of source code must retain the above copyright 77 * notice, this list of conditions and the following disclaimer. 78 * 2. Redistributions in binary form must reproduce the above copyright 79 * notice, this list of conditions and the following disclaimer in the 80 * documentation and/or other materials provided with the distribution. 81 * 3. Neither the name of the University nor the names of its contributors 82 * may be used to endorse or promote products derived from this software 83 * without specific prior written permission. 84 * 85 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 86 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 87 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 88 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 89 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 90 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 91 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 92 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 93 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 94 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 95 * SUCH DAMAGE. 96 * 97 * @(#)in_pcb.c 8.4 (Berkeley) 5/24/95 98 */ 99 100 #include <sys/cdefs.h> 101 __KERNEL_RCSID(0, "$NetBSD: in_pcb.c,v 1.103 2006/07/23 22:06:13 ad Exp $"); 102 103 #include "opt_inet.h" 104 #include "opt_ipsec.h" 105 106 #include <sys/param.h> 107 #include <sys/systm.h> 108 #include <sys/malloc.h> 109 #include <sys/mbuf.h> 110 #include <sys/protosw.h> 111 #include <sys/socket.h> 112 #include <sys/socketvar.h> 113 #include <sys/ioctl.h> 114 #include <sys/errno.h> 115 #include <sys/time.h> 116 #include <sys/pool.h> 117 #include <sys/proc.h> 118 #include <sys/kauth.h> 119 120 #include <net/if.h> 121 #include <net/route.h> 122 123 #include <netinet/in.h> 124 #include <netinet/in_systm.h> 125 #include <netinet/ip.h> 126 #include <netinet/in_pcb.h> 127 #include <netinet/in_var.h> 128 #include <netinet/ip_var.h> 129 130 #ifdef INET6 131 #include <netinet/ip6.h> 132 #include <netinet6/ip6_var.h> 133 #include <netinet6/in6_pcb.h> 134 #endif 135 136 #ifdef IPSEC 137 #include <netinet6/ipsec.h> 138 #include <netkey/key.h> 139 #elif FAST_IPSEC 140 #include <netipsec/ipsec.h> 141 #include <netipsec/key.h> 142 #endif /* IPSEC */ 143 144 struct in_addr zeroin_addr; 145 146 #define INPCBHASH_PORT(table, lport) \ 147 &(table)->inpt_porthashtbl[ntohs(lport) & (table)->inpt_porthash] 148 #define INPCBHASH_BIND(table, laddr, lport) \ 149 &(table)->inpt_bindhashtbl[ \ 150 ((ntohl((laddr).s_addr) + ntohs(lport))) & (table)->inpt_bindhash] 151 #define INPCBHASH_CONNECT(table, faddr, fport, laddr, lport) \ 152 &(table)->inpt_connecthashtbl[ \ 153 ((ntohl((faddr).s_addr) + ntohs(fport)) + \ 154 (ntohl((laddr).s_addr) + ntohs(lport))) & (table)->inpt_connecthash] 155 156 int anonportmin = IPPORT_ANONMIN; 157 int anonportmax = IPPORT_ANONMAX; 158 int lowportmin = IPPORT_RESERVEDMIN; 159 int lowportmax = IPPORT_RESERVEDMAX; 160 161 POOL_INIT(inpcb_pool, sizeof(struct inpcb), 0, 0, 0, "inpcbpl", NULL); 162 163 void 164 in_pcbinit(struct inpcbtable *table, int bindhashsize, int connecthashsize) 165 { 166 167 CIRCLEQ_INIT(&table->inpt_queue); 168 table->inpt_porthashtbl = hashinit(bindhashsize, HASH_LIST, M_PCB, 169 M_WAITOK, &table->inpt_porthash); 170 table->inpt_bindhashtbl = hashinit(bindhashsize, HASH_LIST, M_PCB, 171 M_WAITOK, &table->inpt_bindhash); 172 table->inpt_connecthashtbl = hashinit(connecthashsize, HASH_LIST, 173 M_PCB, M_WAITOK, &table->inpt_connecthash); 174 table->inpt_lastlow = IPPORT_RESERVEDMAX; 175 table->inpt_lastport = (u_int16_t)anonportmax; 176 } 177 178 int 179 in_pcballoc(struct socket *so, void *v) 180 { 181 struct inpcbtable *table = v; 182 struct inpcb *inp; 183 int s; 184 #if defined(IPSEC) || defined(FAST_IPSEC) 185 int error; 186 #endif 187 188 inp = pool_get(&inpcb_pool, PR_NOWAIT); 189 if (inp == NULL) 190 return (ENOBUFS); 191 bzero((caddr_t)inp, sizeof(*inp)); 192 inp->inp_af = AF_INET; 193 inp->inp_table = table; 194 inp->inp_socket = so; 195 inp->inp_errormtu = -1; 196 #if defined(IPSEC) || defined(FAST_IPSEC) 197 error = ipsec_init_pcbpolicy(so, &inp->inp_sp); 198 if (error != 0) { 199 pool_put(&inpcb_pool, inp); 200 return error; 201 } 202 #endif 203 so->so_pcb = inp; 204 s = splnet(); 205 CIRCLEQ_INSERT_HEAD(&table->inpt_queue, &inp->inp_head, 206 inph_queue); 207 LIST_INSERT_HEAD(INPCBHASH_PORT(table, inp->inp_lport), &inp->inp_head, 208 inph_lhash); 209 in_pcbstate(inp, INP_ATTACHED); 210 splx(s); 211 return (0); 212 } 213 214 int 215 in_pcbbind(void *v, struct mbuf *nam, struct lwp *l) 216 { 217 struct in_ifaddr *ia = NULL; 218 struct inpcb *inp = v; 219 struct socket *so = inp->inp_socket; 220 struct inpcbtable *table = inp->inp_table; 221 struct sockaddr_in *sin; 222 u_int16_t lport = 0; 223 int wild = 0, reuseport = (so->so_options & SO_REUSEPORT); 224 225 if (inp->inp_af != AF_INET) 226 return (EINVAL); 227 228 if (TAILQ_FIRST(&in_ifaddrhead) == 0) 229 return (EADDRNOTAVAIL); 230 if (inp->inp_lport || !in_nullhost(inp->inp_laddr)) 231 return (EINVAL); 232 if ((so->so_options & (SO_REUSEADDR|SO_REUSEPORT)) == 0) 233 wild = 1; 234 if (nam == 0) 235 goto noname; 236 sin = mtod(nam, struct sockaddr_in *); 237 if (nam->m_len != sizeof (*sin)) 238 return (EINVAL); 239 if (sin->sin_family != AF_INET) 240 return (EAFNOSUPPORT); 241 lport = sin->sin_port; 242 if (IN_MULTICAST(sin->sin_addr.s_addr)) { 243 /* 244 * Treat SO_REUSEADDR as SO_REUSEPORT for multicast; 245 * allow complete duplication of binding if 246 * SO_REUSEPORT is set, or if SO_REUSEADDR is set 247 * and a multicast address is bound on both 248 * new and duplicated sockets. 249 */ 250 if (so->so_options & SO_REUSEADDR) 251 reuseport = SO_REUSEADDR|SO_REUSEPORT; 252 } else if (!in_nullhost(sin->sin_addr)) { 253 sin->sin_port = 0; /* yech... */ 254 INADDR_TO_IA(sin->sin_addr, ia); 255 /* check for broadcast addresses */ 256 if (ia == NULL) 257 ia = ifatoia(ifa_ifwithaddr(sintosa(sin))); 258 if (ia == NULL) 259 return (EADDRNOTAVAIL); 260 } 261 if (lport) { 262 struct inpcb *t; 263 #ifdef INET6 264 struct in6pcb *t6; 265 struct in6_addr mapped; 266 #endif 267 #ifndef IPNOPRIVPORTS 268 /* GROSS */ 269 if (ntohs(lport) < IPPORT_RESERVED && 270 (l == 0 || kauth_authorize_generic(l->l_cred, 271 KAUTH_GENERIC_ISSUSER, &l->l_acflag))) 272 return (EACCES); 273 #endif 274 #ifdef INET6 275 memset(&mapped, 0, sizeof(mapped)); 276 mapped.s6_addr16[5] = 0xffff; 277 memcpy(&mapped.s6_addr32[3], &sin->sin_addr, 278 sizeof(mapped.s6_addr32[3])); 279 t6 = in6_pcblookup_port(table, &mapped, lport, wild); 280 if (t6 && (reuseport & t6->in6p_socket->so_options) == 0) 281 return (EADDRINUSE); 282 #endif 283 if (so->so_uidinfo->ui_uid && !IN_MULTICAST(sin->sin_addr.s_addr)) { 284 t = in_pcblookup_port(table, sin->sin_addr, lport, 1); 285 /* 286 * XXX: investigate ramifications of loosening this 287 * restriction so that as long as both ports have 288 * SO_REUSEPORT allow the bind 289 */ 290 if (t && 291 (!in_nullhost(sin->sin_addr) || 292 !in_nullhost(t->inp_laddr) || 293 (t->inp_socket->so_options & SO_REUSEPORT) == 0) 294 && (so->so_uidinfo->ui_uid != t->inp_socket->so_uidinfo->ui_uid)) { 295 return (EADDRINUSE); 296 } 297 } 298 t = in_pcblookup_port(table, sin->sin_addr, lport, wild); 299 if (t && (reuseport & t->inp_socket->so_options) == 0) 300 return (EADDRINUSE); 301 } 302 inp->inp_laddr = sin->sin_addr; 303 304 noname: 305 if (lport == 0) { 306 int cnt; 307 u_int16_t mymin, mymax; 308 u_int16_t *lastport; 309 310 if (inp->inp_flags & INP_LOWPORT) { 311 #ifndef IPNOPRIVPORTS 312 if (l == 0 || kauth_authorize_generic(l->l_cred, 313 KAUTH_GENERIC_ISSUSER, &l->l_acflag)) 314 return (EACCES); 315 #endif 316 mymin = lowportmin; 317 mymax = lowportmax; 318 lastport = &table->inpt_lastlow; 319 } else { 320 mymin = anonportmin; 321 mymax = anonportmax; 322 lastport = &table->inpt_lastport; 323 } 324 if (mymin > mymax) { /* sanity check */ 325 u_int16_t swp; 326 327 swp = mymin; 328 mymin = mymax; 329 mymax = swp; 330 } 331 332 lport = *lastport - 1; 333 for (cnt = mymax - mymin + 1; cnt; cnt--, lport--) { 334 if (lport < mymin || lport > mymax) 335 lport = mymax; 336 if (!in_pcblookup_port(table, inp->inp_laddr, 337 htons(lport), 1)) 338 goto found; 339 } 340 if (!in_nullhost(inp->inp_laddr)) 341 inp->inp_laddr.s_addr = INADDR_ANY; 342 return (EAGAIN); 343 found: 344 inp->inp_flags |= INP_ANONPORT; 345 *lastport = lport; 346 lport = htons(lport); 347 } 348 inp->inp_lport = lport; 349 LIST_REMOVE(&inp->inp_head, inph_lhash); 350 LIST_INSERT_HEAD(INPCBHASH_PORT(table, inp->inp_lport), &inp->inp_head, 351 inph_lhash); 352 in_pcbstate(inp, INP_BOUND); 353 return (0); 354 } 355 356 /* 357 * Connect from a socket to a specified address. 358 * Both address and port must be specified in argument sin. 359 * If don't have a local address for this socket yet, 360 * then pick one. 361 */ 362 int 363 in_pcbconnect(void *v, struct mbuf *nam, struct lwp *l) 364 { 365 struct inpcb *inp = v; 366 struct in_ifaddr *ia = NULL; 367 struct sockaddr_in *ifaddr = NULL; 368 struct sockaddr_in *sin = mtod(nam, struct sockaddr_in *); 369 int error; 370 371 if (inp->inp_af != AF_INET) 372 return (EINVAL); 373 374 if (nam->m_len != sizeof (*sin)) 375 return (EINVAL); 376 if (sin->sin_family != AF_INET) 377 return (EAFNOSUPPORT); 378 if (sin->sin_port == 0) 379 return (EADDRNOTAVAIL); 380 if (TAILQ_FIRST(&in_ifaddrhead) != 0) { 381 /* 382 * If the destination address is INADDR_ANY, 383 * use any local address (likely loopback). 384 * If the supplied address is INADDR_BROADCAST, 385 * use the broadcast address of an interface 386 * which supports broadcast. (loopback does not) 387 */ 388 389 if (in_nullhost(sin->sin_addr)) { 390 sin->sin_addr = 391 TAILQ_FIRST(&in_ifaddrhead)->ia_addr.sin_addr; 392 } else if (sin->sin_addr.s_addr == INADDR_BROADCAST) { 393 TAILQ_FOREACH(ia, &in_ifaddrhead, ia_list) { 394 if (ia->ia_ifp->if_flags & IFF_BROADCAST) { 395 sin->sin_addr = 396 ia->ia_broadaddr.sin_addr; 397 break; 398 } 399 } 400 } 401 } 402 /* 403 * If we haven't bound which network number to use as ours, 404 * we will use the number of the outgoing interface. 405 * This depends on having done a routing lookup, which 406 * we will probably have to do anyway, so we might 407 * as well do it now. On the other hand if we are 408 * sending to multiple destinations we may have already 409 * done the lookup, so see if we can use the route 410 * from before. In any case, we only 411 * chose a port number once, even if sending to multiple 412 * destinations. 413 */ 414 if (in_nullhost(inp->inp_laddr)) { 415 int xerror; 416 ifaddr = in_selectsrc(sin, &inp->inp_route, 417 inp->inp_socket->so_options, inp->inp_moptions, &xerror); 418 if (ifaddr == NULL) { 419 if (xerror == 0) 420 xerror = EADDRNOTAVAIL; 421 return xerror; 422 } 423 INADDR_TO_IA(ifaddr->sin_addr, ia); 424 if (ia == NULL) 425 return (EADDRNOTAVAIL); 426 } 427 if (in_pcblookup_connect(inp->inp_table, sin->sin_addr, sin->sin_port, 428 !in_nullhost(inp->inp_laddr) ? inp->inp_laddr : ifaddr->sin_addr, 429 inp->inp_lport) != 0) 430 return (EADDRINUSE); 431 if (in_nullhost(inp->inp_laddr)) { 432 if (inp->inp_lport == 0) { 433 error = in_pcbbind(inp, NULL, l); 434 /* 435 * This used to ignore the return value 436 * completely, but we need to check for 437 * ephemeral port shortage. 438 * And attempts to request low ports if not root. 439 */ 440 if (error != 0) 441 return (error); 442 } 443 inp->inp_laddr = ifaddr->sin_addr; 444 } 445 inp->inp_faddr = sin->sin_addr; 446 inp->inp_fport = sin->sin_port; 447 in_pcbstate(inp, INP_CONNECTED); 448 #if defined(IPSEC) || defined(FAST_IPSEC) 449 if (inp->inp_socket->so_type == SOCK_STREAM) 450 ipsec_pcbconn(inp->inp_sp); 451 #endif 452 return (0); 453 } 454 455 void 456 in_pcbdisconnect(void *v) 457 { 458 struct inpcb *inp = v; 459 460 if (inp->inp_af != AF_INET) 461 return; 462 463 inp->inp_faddr = zeroin_addr; 464 inp->inp_fport = 0; 465 in_pcbstate(inp, INP_BOUND); 466 #if defined(IPSEC) || defined(FAST_IPSEC) 467 ipsec_pcbdisconn(inp->inp_sp); 468 #endif 469 if (inp->inp_socket->so_state & SS_NOFDREF) 470 in_pcbdetach(inp); 471 } 472 473 void 474 in_pcbdetach(void *v) 475 { 476 struct inpcb *inp = v; 477 struct socket *so = inp->inp_socket; 478 int s; 479 480 if (inp->inp_af != AF_INET) 481 return; 482 483 #if defined(IPSEC) || defined(FAST_IPSEC) 484 ipsec4_delete_pcbpolicy(inp); 485 #endif /*IPSEC*/ 486 so->so_pcb = 0; 487 sofree(so); 488 if (inp->inp_options) 489 (void)m_free(inp->inp_options); 490 if (inp->inp_route.ro_rt) 491 rtfree(inp->inp_route.ro_rt); 492 ip_freemoptions(inp->inp_moptions); 493 s = splnet(); 494 in_pcbstate(inp, INP_ATTACHED); 495 LIST_REMOVE(&inp->inp_head, inph_lhash); 496 CIRCLEQ_REMOVE(&inp->inp_table->inpt_queue, &inp->inp_head, 497 inph_queue); 498 splx(s); 499 pool_put(&inpcb_pool, inp); 500 } 501 502 void 503 in_setsockaddr(struct inpcb *inp, struct mbuf *nam) 504 { 505 struct sockaddr_in *sin; 506 507 if (inp->inp_af != AF_INET) 508 return; 509 510 nam->m_len = sizeof (*sin); 511 sin = mtod(nam, struct sockaddr_in *); 512 bzero((caddr_t)sin, sizeof (*sin)); 513 sin->sin_family = AF_INET; 514 sin->sin_len = sizeof(*sin); 515 sin->sin_port = inp->inp_lport; 516 sin->sin_addr = inp->inp_laddr; 517 } 518 519 void 520 in_setpeeraddr(struct inpcb *inp, struct mbuf *nam) 521 { 522 struct sockaddr_in *sin; 523 524 if (inp->inp_af != AF_INET) 525 return; 526 527 nam->m_len = sizeof (*sin); 528 sin = mtod(nam, struct sockaddr_in *); 529 bzero((caddr_t)sin, sizeof (*sin)); 530 sin->sin_family = AF_INET; 531 sin->sin_len = sizeof(*sin); 532 sin->sin_port = inp->inp_fport; 533 sin->sin_addr = inp->inp_faddr; 534 } 535 536 /* 537 * Pass some notification to all connections of a protocol 538 * associated with address dst. The local address and/or port numbers 539 * may be specified to limit the search. The "usual action" will be 540 * taken, depending on the ctlinput cmd. The caller must filter any 541 * cmds that are uninteresting (e.g., no error in the map). 542 * Call the protocol specific routine (if any) to report 543 * any errors for each matching socket. 544 * 545 * Must be called at splsoftnet. 546 */ 547 int 548 in_pcbnotify(struct inpcbtable *table, struct in_addr faddr, u_int fport_arg, 549 struct in_addr laddr, u_int lport_arg, int errno, 550 void (*notify)(struct inpcb *, int)) 551 { 552 struct inpcbhead *head; 553 struct inpcb *inp, *ninp; 554 u_int16_t fport = fport_arg, lport = lport_arg; 555 int nmatch; 556 557 if (in_nullhost(faddr) || notify == 0) 558 return (0); 559 560 nmatch = 0; 561 head = INPCBHASH_CONNECT(table, faddr, fport, laddr, lport); 562 for (inp = (struct inpcb *)LIST_FIRST(head); inp != NULL; inp = ninp) { 563 ninp = (struct inpcb *)LIST_NEXT(inp, inp_hash); 564 if (inp->inp_af != AF_INET) 565 continue; 566 if (in_hosteq(inp->inp_faddr, faddr) && 567 inp->inp_fport == fport && 568 inp->inp_lport == lport && 569 in_hosteq(inp->inp_laddr, laddr)) { 570 (*notify)(inp, errno); 571 nmatch++; 572 } 573 } 574 return (nmatch); 575 } 576 577 void 578 in_pcbnotifyall(struct inpcbtable *table, struct in_addr faddr, int errno, 579 void (*notify)(struct inpcb *, int)) 580 { 581 struct inpcb *inp, *ninp; 582 583 if (in_nullhost(faddr) || notify == 0) 584 return; 585 586 for (inp = (struct inpcb *)CIRCLEQ_FIRST(&table->inpt_queue); 587 inp != (void *)&table->inpt_queue; 588 inp = ninp) { 589 ninp = (struct inpcb *)CIRCLEQ_NEXT(inp, inp_queue); 590 if (inp->inp_af != AF_INET) 591 continue; 592 if (in_hosteq(inp->inp_faddr, faddr)) 593 (*notify)(inp, errno); 594 } 595 } 596 597 void 598 in_pcbpurgeif0(struct inpcbtable *table, struct ifnet *ifp) 599 { 600 struct inpcb *inp, *ninp; 601 struct ip_moptions *imo; 602 int i, gap; 603 604 for (inp = (struct inpcb *)CIRCLEQ_FIRST(&table->inpt_queue); 605 inp != (void *)&table->inpt_queue; 606 inp = ninp) { 607 ninp = (struct inpcb *)CIRCLEQ_NEXT(inp, inp_queue); 608 if (inp->inp_af != AF_INET) 609 continue; 610 imo = inp->inp_moptions; 611 if (imo != NULL) { 612 /* 613 * Unselect the outgoing interface if it is being 614 * detached. 615 */ 616 if (imo->imo_multicast_ifp == ifp) 617 imo->imo_multicast_ifp = NULL; 618 619 /* 620 * Drop multicast group membership if we joined 621 * through the interface being detached. 622 */ 623 for (i = 0, gap = 0; i < imo->imo_num_memberships; 624 i++) { 625 if (imo->imo_membership[i]->inm_ifp == ifp) { 626 in_delmulti(imo->imo_membership[i]); 627 gap++; 628 } else if (gap != 0) 629 imo->imo_membership[i - gap] = 630 imo->imo_membership[i]; 631 } 632 imo->imo_num_memberships -= gap; 633 } 634 } 635 } 636 637 void 638 in_pcbpurgeif(struct inpcbtable *table, struct ifnet *ifp) 639 { 640 struct inpcb *inp, *ninp; 641 642 for (inp = (struct inpcb *)CIRCLEQ_FIRST(&table->inpt_queue); 643 inp != (void *)&table->inpt_queue; 644 inp = ninp) { 645 ninp = (struct inpcb *)CIRCLEQ_NEXT(inp, inp_queue); 646 if (inp->inp_af != AF_INET) 647 continue; 648 if (inp->inp_route.ro_rt != NULL && 649 inp->inp_route.ro_rt->rt_ifp == ifp) 650 in_rtchange(inp, 0); 651 } 652 } 653 654 /* 655 * Check for alternatives when higher level complains 656 * about service problems. For now, invalidate cached 657 * routing information. If the route was created dynamically 658 * (by a redirect), time to try a default gateway again. 659 */ 660 void 661 in_losing(struct inpcb *inp) 662 { 663 struct rtentry *rt; 664 struct rt_addrinfo info; 665 666 if (inp->inp_af != AF_INET) 667 return; 668 669 if ((rt = inp->inp_route.ro_rt)) { 670 inp->inp_route.ro_rt = 0; 671 bzero((caddr_t)&info, sizeof(info)); 672 info.rti_info[RTAX_DST] = &inp->inp_route.ro_dst; 673 info.rti_info[RTAX_GATEWAY] = rt->rt_gateway; 674 info.rti_info[RTAX_NETMASK] = rt_mask(rt); 675 rt_missmsg(RTM_LOSING, &info, rt->rt_flags, 0); 676 if (rt->rt_flags & RTF_DYNAMIC) 677 (void) rtrequest(RTM_DELETE, rt_key(rt), 678 rt->rt_gateway, rt_mask(rt), rt->rt_flags, 679 (struct rtentry **)0); 680 else 681 /* 682 * A new route can be allocated 683 * the next time output is attempted. 684 */ 685 rtfree(rt); 686 } 687 } 688 689 /* 690 * After a routing change, flush old routing 691 * and allocate a (hopefully) better one. 692 */ 693 void 694 in_rtchange(struct inpcb *inp, int errno) 695 { 696 697 if (inp->inp_af != AF_INET) 698 return; 699 700 if (inp->inp_route.ro_rt) { 701 rtfree(inp->inp_route.ro_rt); 702 inp->inp_route.ro_rt = 0; 703 /* 704 * A new route can be allocated the next time 705 * output is attempted. 706 */ 707 } 708 /* XXX SHOULD NOTIFY HIGHER-LEVEL PROTOCOLS */ 709 } 710 711 struct inpcb * 712 in_pcblookup_port(struct inpcbtable *table, struct in_addr laddr, 713 u_int lport_arg, int lookup_wildcard) 714 { 715 struct inpcbhead *head; 716 struct inpcb_hdr *inph; 717 struct inpcb *inp, *match = 0; 718 int matchwild = 3, wildcard; 719 u_int16_t lport = lport_arg; 720 721 head = INPCBHASH_PORT(table, lport); 722 LIST_FOREACH(inph, head, inph_lhash) { 723 inp = (struct inpcb *)inph; 724 if (inp->inp_af != AF_INET) 725 continue; 726 727 if (inp->inp_lport != lport) 728 continue; 729 wildcard = 0; 730 if (!in_nullhost(inp->inp_faddr)) 731 wildcard++; 732 if (in_nullhost(inp->inp_laddr)) { 733 if (!in_nullhost(laddr)) 734 wildcard++; 735 } else { 736 if (in_nullhost(laddr)) 737 wildcard++; 738 else { 739 if (!in_hosteq(inp->inp_laddr, laddr)) 740 continue; 741 } 742 } 743 if (wildcard && !lookup_wildcard) 744 continue; 745 if (wildcard < matchwild) { 746 match = inp; 747 matchwild = wildcard; 748 if (matchwild == 0) 749 break; 750 } 751 } 752 return (match); 753 } 754 755 #ifdef DIAGNOSTIC 756 int in_pcbnotifymiss = 0; 757 #endif 758 759 struct inpcb * 760 in_pcblookup_connect(struct inpcbtable *table, 761 struct in_addr faddr, u_int fport_arg, 762 struct in_addr laddr, u_int lport_arg) 763 { 764 struct inpcbhead *head; 765 struct inpcb_hdr *inph; 766 struct inpcb *inp; 767 u_int16_t fport = fport_arg, lport = lport_arg; 768 769 head = INPCBHASH_CONNECT(table, faddr, fport, laddr, lport); 770 LIST_FOREACH(inph, head, inph_hash) { 771 inp = (struct inpcb *)inph; 772 if (inp->inp_af != AF_INET) 773 continue; 774 775 if (in_hosteq(inp->inp_faddr, faddr) && 776 inp->inp_fport == fport && 777 inp->inp_lport == lport && 778 in_hosteq(inp->inp_laddr, laddr)) 779 goto out; 780 } 781 #ifdef DIAGNOSTIC 782 if (in_pcbnotifymiss) { 783 printf("in_pcblookup_connect: faddr=%08x fport=%d laddr=%08x lport=%d\n", 784 ntohl(faddr.s_addr), ntohs(fport), 785 ntohl(laddr.s_addr), ntohs(lport)); 786 } 787 #endif 788 return (0); 789 790 out: 791 /* Move this PCB to the head of hash chain. */ 792 inph = &inp->inp_head; 793 if (inph != LIST_FIRST(head)) { 794 LIST_REMOVE(inph, inph_hash); 795 LIST_INSERT_HEAD(head, inph, inph_hash); 796 } 797 return (inp); 798 } 799 800 struct inpcb * 801 in_pcblookup_bind(struct inpcbtable *table, 802 struct in_addr laddr, u_int lport_arg) 803 { 804 struct inpcbhead *head; 805 struct inpcb_hdr *inph; 806 struct inpcb *inp; 807 u_int16_t lport = lport_arg; 808 809 head = INPCBHASH_BIND(table, laddr, lport); 810 LIST_FOREACH(inph, head, inph_hash) { 811 inp = (struct inpcb *)inph; 812 if (inp->inp_af != AF_INET) 813 continue; 814 815 if (inp->inp_lport == lport && 816 in_hosteq(inp->inp_laddr, laddr)) 817 goto out; 818 } 819 head = INPCBHASH_BIND(table, zeroin_addr, lport); 820 LIST_FOREACH(inph, head, inph_hash) { 821 inp = (struct inpcb *)inph; 822 if (inp->inp_af != AF_INET) 823 continue; 824 825 if (inp->inp_lport == lport && 826 in_hosteq(inp->inp_laddr, zeroin_addr)) 827 goto out; 828 } 829 #ifdef DIAGNOSTIC 830 if (in_pcbnotifymiss) { 831 printf("in_pcblookup_bind: laddr=%08x lport=%d\n", 832 ntohl(laddr.s_addr), ntohs(lport)); 833 } 834 #endif 835 return (0); 836 837 out: 838 /* Move this PCB to the head of hash chain. */ 839 inph = &inp->inp_head; 840 if (inph != LIST_FIRST(head)) { 841 LIST_REMOVE(inph, inph_hash); 842 LIST_INSERT_HEAD(head, inph, inph_hash); 843 } 844 return (inp); 845 } 846 847 void 848 in_pcbstate(struct inpcb *inp, int state) 849 { 850 851 if (inp->inp_af != AF_INET) 852 return; 853 854 if (inp->inp_state > INP_ATTACHED) 855 LIST_REMOVE(&inp->inp_head, inph_hash); 856 857 switch (state) { 858 case INP_BOUND: 859 LIST_INSERT_HEAD(INPCBHASH_BIND(inp->inp_table, 860 inp->inp_laddr, inp->inp_lport), &inp->inp_head, 861 inph_hash); 862 break; 863 case INP_CONNECTED: 864 LIST_INSERT_HEAD(INPCBHASH_CONNECT(inp->inp_table, 865 inp->inp_faddr, inp->inp_fport, 866 inp->inp_laddr, inp->inp_lport), &inp->inp_head, 867 inph_hash); 868 break; 869 } 870 871 inp->inp_state = state; 872 } 873 874 struct rtentry * 875 in_pcbrtentry(struct inpcb *inp) 876 { 877 struct route *ro; 878 879 if (inp->inp_af != AF_INET) 880 return (NULL); 881 882 ro = &inp->inp_route; 883 884 if (ro->ro_rt && ((ro->ro_rt->rt_flags & RTF_UP) == 0 || 885 !in_hosteq(satosin(&ro->ro_dst)->sin_addr, inp->inp_faddr))) { 886 RTFREE(ro->ro_rt); 887 ro->ro_rt = (struct rtentry *)NULL; 888 } 889 if (ro->ro_rt == (struct rtentry *)NULL && 890 !in_nullhost(inp->inp_faddr)) { 891 bzero(&ro->ro_dst, sizeof(struct sockaddr_in)); 892 ro->ro_dst.sa_family = AF_INET; 893 ro->ro_dst.sa_len = sizeof(ro->ro_dst); 894 satosin(&ro->ro_dst)->sin_addr = inp->inp_faddr; 895 rtalloc(ro); 896 } 897 return (ro->ro_rt); 898 } 899 900 struct sockaddr_in * 901 in_selectsrc(struct sockaddr_in *sin, struct route *ro, 902 int soopts, struct ip_moptions *mopts, int *errorp) 903 { 904 struct in_ifaddr *ia; 905 906 ia = (struct in_ifaddr *)0; 907 /* 908 * If route is known or can be allocated now, 909 * our src addr is taken from the i/f, else punt. 910 * Note that we should check the address family of the cached 911 * destination, in case of sharing the cache with IPv6. 912 */ 913 if (ro->ro_rt && 914 (ro->ro_dst.sa_family != AF_INET || 915 !in_hosteq(satosin(&ro->ro_dst)->sin_addr, sin->sin_addr) || 916 soopts & SO_DONTROUTE)) { 917 RTFREE(ro->ro_rt); 918 ro->ro_rt = (struct rtentry *)0; 919 } 920 if ((soopts & SO_DONTROUTE) == 0 && /*XXX*/ 921 (ro->ro_rt == (struct rtentry *)0 || 922 ro->ro_rt->rt_ifp == (struct ifnet *)0)) { 923 /* No route yet, so try to acquire one */ 924 bzero(&ro->ro_dst, sizeof(struct sockaddr_in)); 925 ro->ro_dst.sa_family = AF_INET; 926 ro->ro_dst.sa_len = sizeof(struct sockaddr_in); 927 satosin(&ro->ro_dst)->sin_addr = sin->sin_addr; 928 rtalloc(ro); 929 } 930 /* 931 * If we found a route, use the address 932 * corresponding to the outgoing interface 933 * unless it is the loopback (in case a route 934 * to our address on another net goes to loopback). 935 * 936 * XXX Is this still true? Do we care? 937 */ 938 if (ro->ro_rt && !(ro->ro_rt->rt_ifp->if_flags & IFF_LOOPBACK)) 939 ia = ifatoia(ro->ro_rt->rt_ifa); 940 if (ia == NULL) { 941 u_int16_t fport = sin->sin_port; 942 943 sin->sin_port = 0; 944 ia = ifatoia(ifa_ifwithladdr(sintosa(sin))); 945 sin->sin_port = fport; 946 if (ia == 0) { 947 /* Find 1st non-loopback AF_INET address */ 948 TAILQ_FOREACH(ia, &in_ifaddrhead, ia_list) { 949 if (!(ia->ia_ifp->if_flags & IFF_LOOPBACK)) 950 break; 951 } 952 } 953 if (ia == NULL) { 954 *errorp = EADDRNOTAVAIL; 955 return NULL; 956 } 957 } 958 /* 959 * If the destination address is multicast and an outgoing 960 * interface has been set as a multicast option, use the 961 * address of that interface as our source address. 962 */ 963 if (IN_MULTICAST(sin->sin_addr.s_addr) && mopts != NULL) { 964 struct ip_moptions *imo; 965 struct ifnet *ifp; 966 967 imo = mopts; 968 if (imo->imo_multicast_ifp != NULL) { 969 ifp = imo->imo_multicast_ifp; 970 IFP_TO_IA(ifp, ia); /* XXX */ 971 if (ia == 0) { 972 *errorp = EADDRNOTAVAIL; 973 return NULL; 974 } 975 } 976 } 977 return satosin(&ia->ia_addr); 978 } 979
Indexes created Mon Sep 22 13:09:51 GMT 2025