in.c revision 1.178
1 /* $NetBSD: in.c,v 1.178 2016/08/01 03:15:30 ozaki-r 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 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.c 8.4 (Berkeley) 1/9/95 91 */ 92 93 #include <sys/cdefs.h> 94 __KERNEL_RCSID(0, "$NetBSD: in.c,v 1.178 2016/08/01 03:15:30 ozaki-r Exp $"); 95 96 #include "arp.h" 97 98 #ifdef _KERNEL_OPT 99 #include "opt_inet.h" 100 #include "opt_inet_conf.h" 101 #include "opt_mrouting.h" 102 #endif 103 104 #include <sys/param.h> 105 #include <sys/ioctl.h> 106 #include <sys/errno.h> 107 #include <sys/kernel.h> 108 #include <sys/malloc.h> 109 #include <sys/socket.h> 110 #include <sys/socketvar.h> 111 #include <sys/sysctl.h> 112 #include <sys/systm.h> 113 #include <sys/proc.h> 114 #include <sys/syslog.h> 115 #include <sys/kauth.h> 116 #include <sys/kmem.h> 117 118 #include <sys/cprng.h> 119 120 #include <net/if.h> 121 #include <net/route.h> 122 #include <net/pfil.h> 123 124 #include <net/if_arp.h> 125 #include <net/if_ether.h> 126 #include <net/if_types.h> 127 #include <net/if_llatbl.h> 128 #include <net/if_dl.h> 129 130 #include <netinet/in_systm.h> 131 #include <netinet/in.h> 132 #include <netinet/in_var.h> 133 #include <netinet/ip.h> 134 #include <netinet/ip_var.h> 135 #include <netinet/in_ifattach.h> 136 #include <netinet/in_pcb.h> 137 #include <netinet/in_selsrc.h> 138 #include <netinet/if_inarp.h> 139 #include <netinet/ip_mroute.h> 140 #include <netinet/igmp_var.h> 141 142 #ifdef IPSELSRC 143 #include <netinet/in_selsrc.h> 144 #endif 145 146 static u_int in_mask2len(struct in_addr *); 147 static void in_len2mask(struct in_addr *, u_int); 148 static int in_lifaddr_ioctl(struct socket *, u_long, void *, 149 struct ifnet *); 150 151 static int in_addprefix(struct in_ifaddr *, int); 152 static int in_scrubprefix(struct in_ifaddr *); 153 static void in_sysctl_init(struct sysctllog **); 154 155 #ifndef SUBNETSARELOCAL 156 #define SUBNETSARELOCAL 1 157 #endif 158 159 #ifndef HOSTZEROBROADCAST 160 #define HOSTZEROBROADCAST 0 161 #endif 162 163 /* Note: 61, 127, 251, 509, 1021, 2039 are good. */ 164 #ifndef IN_MULTI_HASH_SIZE 165 #define IN_MULTI_HASH_SIZE 509 166 #endif 167 168 static int subnetsarelocal = SUBNETSARELOCAL; 169 static int hostzeroisbroadcast = HOSTZEROBROADCAST; 170 171 /* 172 * This list is used to keep track of in_multi chains which belong to 173 * deleted interface addresses. We use in_ifaddr so that a chain head 174 * won't be deallocated until all multicast address record are deleted. 175 */ 176 177 LIST_HEAD(in_multihashhead, in_multi); /* Type of the hash head */ 178 179 static struct pool inmulti_pool; 180 static u_int in_multientries; 181 static struct in_multihashhead *in_multihashtbl; 182 static u_long in_multihash; 183 static krwlock_t in_multilock; 184 185 #define IN_MULTI_HASH(x, ifp) \ 186 (in_multihashtbl[(u_long)((x) ^ (ifp->if_index)) % IN_MULTI_HASH_SIZE]) 187 188 /* XXX DEPRECATED. Keep them to avoid breaking kvm(3) users. */ 189 struct in_ifaddrhashhead * in_ifaddrhashtbl; 190 u_long in_ifaddrhash; 191 struct in_ifaddrhead in_ifaddrhead; 192 static kmutex_t in_ifaddr_lock; 193 194 struct pslist_head * in_ifaddrhashtbl_pslist; 195 u_long in_ifaddrhash_pslist; 196 struct pslist_head in_ifaddrhead_pslist; 197 198 void 199 in_init(void) 200 { 201 pool_init(&inmulti_pool, sizeof(struct in_multi), 0, 0, 0, "inmltpl", 202 NULL, IPL_SOFTNET); 203 TAILQ_INIT(&in_ifaddrhead); 204 PSLIST_INIT(&in_ifaddrhead_pslist); 205 206 in_ifaddrhashtbl = hashinit(IN_IFADDR_HASH_SIZE, HASH_LIST, true, 207 &in_ifaddrhash); 208 209 in_ifaddrhashtbl_pslist = hashinit(IN_IFADDR_HASH_SIZE, HASH_PSLIST, 210 true, &in_ifaddrhash_pslist); 211 mutex_init(&in_ifaddr_lock, MUTEX_DEFAULT, IPL_NONE); 212 213 in_multihashtbl = hashinit(IN_IFADDR_HASH_SIZE, HASH_LIST, true, 214 &in_multihash); 215 rw_init(&in_multilock); 216 217 in_sysctl_init(NULL); 218 } 219 220 /* 221 * Return 1 if an internet address is for a ``local'' host 222 * (one to which we have a connection). If subnetsarelocal 223 * is true, this includes other subnets of the local net. 224 * Otherwise, it includes only the directly-connected (sub)nets. 225 */ 226 int 227 in_localaddr(struct in_addr in) 228 { 229 struct in_ifaddr *ia; 230 int localaddr = 0; 231 int s = pserialize_read_enter(); 232 233 if (subnetsarelocal) { 234 IN_ADDRLIST_READER_FOREACH(ia) { 235 if ((in.s_addr & ia->ia_netmask) == ia->ia_net) { 236 localaddr = 1; 237 break; 238 } 239 } 240 } else { 241 IN_ADDRLIST_READER_FOREACH(ia) { 242 if ((in.s_addr & ia->ia_subnetmask) == ia->ia_subnet) { 243 localaddr = 1; 244 break; 245 } 246 } 247 } 248 pserialize_read_exit(s); 249 250 return localaddr; 251 } 252 253 /* 254 * Determine whether an IP address is in a reserved set of addresses 255 * that may not be forwarded, or whether datagrams to that destination 256 * may be forwarded. 257 */ 258 int 259 in_canforward(struct in_addr in) 260 { 261 u_int32_t net; 262 263 if (IN_EXPERIMENTAL(in.s_addr) || IN_MULTICAST(in.s_addr)) 264 return (0); 265 if (IN_CLASSA(in.s_addr)) { 266 net = in.s_addr & IN_CLASSA_NET; 267 if (net == 0 || net == htonl(IN_LOOPBACKNET << IN_CLASSA_NSHIFT)) 268 return (0); 269 } 270 return (1); 271 } 272 273 /* 274 * Trim a mask in a sockaddr 275 */ 276 void 277 in_socktrim(struct sockaddr_in *ap) 278 { 279 char *cplim = (char *) &ap->sin_addr; 280 char *cp = (char *) (&ap->sin_addr + 1); 281 282 ap->sin_len = 0; 283 while (--cp >= cplim) 284 if (*cp) { 285 (ap)->sin_len = cp - (char *) (ap) + 1; 286 break; 287 } 288 } 289 290 /* 291 * Routine to take an Internet address and convert into a 292 * "dotted quad" representation for printing. 293 */ 294 const char * 295 in_fmtaddr(struct in_addr addr) 296 { 297 static char buf[sizeof("123.456.789.123")]; 298 299 addr.s_addr = ntohl(addr.s_addr); 300 301 snprintf(buf, sizeof(buf), "%d.%d.%d.%d", 302 (addr.s_addr >> 24) & 0xFF, 303 (addr.s_addr >> 16) & 0xFF, 304 (addr.s_addr >> 8) & 0xFF, 305 (addr.s_addr >> 0) & 0xFF); 306 return buf; 307 } 308 309 /* 310 * Maintain the "in_maxmtu" variable, which is the largest 311 * mtu for non-local interfaces with AF_INET addresses assigned 312 * to them that are up. 313 */ 314 unsigned long in_maxmtu; 315 316 void 317 in_setmaxmtu(void) 318 { 319 struct in_ifaddr *ia; 320 struct ifnet *ifp; 321 unsigned long maxmtu = 0; 322 int s = pserialize_read_enter(); 323 324 IN_ADDRLIST_READER_FOREACH(ia) { 325 if ((ifp = ia->ia_ifp) == 0) 326 continue; 327 if ((ifp->if_flags & (IFF_UP|IFF_LOOPBACK)) != IFF_UP) 328 continue; 329 if (ifp->if_mtu > maxmtu) 330 maxmtu = ifp->if_mtu; 331 } 332 if (maxmtu) 333 in_maxmtu = maxmtu; 334 pserialize_read_exit(s); 335 } 336 337 static u_int 338 in_mask2len(struct in_addr *mask) 339 { 340 u_int x, y; 341 u_char *p; 342 343 p = (u_char *)mask; 344 for (x = 0; x < sizeof(*mask); x++) { 345 if (p[x] != 0xff) 346 break; 347 } 348 y = 0; 349 if (x < sizeof(*mask)) { 350 for (y = 0; y < NBBY; y++) { 351 if ((p[x] & (0x80 >> y)) == 0) 352 break; 353 } 354 } 355 return x * NBBY + y; 356 } 357 358 static void 359 in_len2mask(struct in_addr *mask, u_int len) 360 { 361 u_int i; 362 u_char *p; 363 364 p = (u_char *)mask; 365 memset(mask, 0, sizeof(*mask)); 366 for (i = 0; i < len / NBBY; i++) 367 p[i] = 0xff; 368 if (len % NBBY) 369 p[i] = (0xff00 >> (len % NBBY)) & 0xff; 370 } 371 372 /* 373 * Generic internet control operations (ioctl's). 374 * Ifp is 0 if not an interface-specific ioctl. 375 */ 376 /* ARGSUSED */ 377 static int 378 in_control0(struct socket *so, u_long cmd, void *data, struct ifnet *ifp) 379 { 380 struct ifreq *ifr = (struct ifreq *)data; 381 struct in_ifaddr *ia = NULL; 382 struct in_aliasreq *ifra = (struct in_aliasreq *)data; 383 struct sockaddr_in oldaddr; 384 int error, hostIsNew, maskIsNew; 385 int newifaddr = 0; 386 bool run_hook = false; 387 bool need_reinsert = false; 388 struct psref psref; 389 int bound; 390 391 switch (cmd) { 392 case SIOCALIFADDR: 393 case SIOCDLIFADDR: 394 case SIOCGLIFADDR: 395 if (ifp == NULL) 396 return EINVAL; 397 return in_lifaddr_ioctl(so, cmd, data, ifp); 398 case SIOCGIFADDRPREF: 399 case SIOCSIFADDRPREF: 400 if (ifp == NULL) 401 return EINVAL; 402 return ifaddrpref_ioctl(so, cmd, data, ifp); 403 } 404 405 bound = curlwp_bind(); 406 /* 407 * Find address for this interface, if it exists. 408 */ 409 if (ifp != NULL) 410 ia = in_get_ia_from_ifp_psref(ifp, &psref); 411 412 hostIsNew = 1; /* moved here to appease gcc */ 413 switch (cmd) { 414 case SIOCAIFADDR: 415 case SIOCDIFADDR: 416 case SIOCGIFALIAS: 417 case SIOCGIFAFLAG_IN: 418 if (ifra->ifra_addr.sin_family == AF_INET) { 419 int s; 420 421 if (ia != NULL) 422 ia4_release(ia, &psref); 423 s = pserialize_read_enter(); 424 IN_ADDRHASH_READER_FOREACH(ia, 425 ifra->ifra_addr.sin_addr.s_addr) { 426 if (ia->ia_ifp == ifp && 427 in_hosteq(ia->ia_addr.sin_addr, 428 ifra->ifra_addr.sin_addr)) 429 break; 430 } 431 if (ia != NULL) 432 ia4_acquire(ia, &psref); 433 pserialize_read_exit(s); 434 } 435 if ((cmd == SIOCDIFADDR || 436 cmd == SIOCGIFALIAS || 437 cmd == SIOCGIFAFLAG_IN) && 438 ia == NULL) { 439 error = EADDRNOTAVAIL; 440 goto out; 441 } 442 443 if (cmd == SIOCDIFADDR && 444 ifra->ifra_addr.sin_family == AF_UNSPEC) { 445 ifra->ifra_addr.sin_family = AF_INET; 446 } 447 /* FALLTHROUGH */ 448 case SIOCSIFADDR: 449 if (ia == NULL || ia->ia_addr.sin_family != AF_INET) 450 ; 451 else if (ifra->ifra_addr.sin_len == 0) { 452 ifra->ifra_addr = ia->ia_addr; 453 hostIsNew = 0; 454 } else if (in_hosteq(ia->ia_addr.sin_addr, 455 ifra->ifra_addr.sin_addr)) 456 hostIsNew = 0; 457 /* FALLTHROUGH */ 458 case SIOCSIFDSTADDR: 459 if (ifra->ifra_addr.sin_family != AF_INET) { 460 error = EAFNOSUPPORT; 461 goto out; 462 } 463 /* FALLTHROUGH */ 464 case SIOCSIFNETMASK: 465 if (ifp == NULL) 466 panic("in_control"); 467 468 if (cmd == SIOCGIFALIAS || cmd == SIOCGIFAFLAG_IN) 469 break; 470 471 if (ia == NULL && 472 (cmd == SIOCSIFNETMASK || cmd == SIOCSIFDSTADDR)) { 473 error = EADDRNOTAVAIL; 474 goto out; 475 } 476 477 if (kauth_authorize_network(curlwp->l_cred, KAUTH_NETWORK_INTERFACE, 478 KAUTH_REQ_NETWORK_INTERFACE_SETPRIV, ifp, (void *)cmd, 479 NULL) != 0) { 480 error = EPERM; 481 goto out; 482 } 483 484 if (ia == NULL) { 485 ia = malloc(sizeof(*ia), M_IFADDR, M_WAITOK|M_ZERO); 486 if (ia == NULL) { 487 error = ENOBUFS; 488 goto out; 489 } 490 ia->ia_ifa.ifa_addr = sintosa(&ia->ia_addr); 491 ia->ia_ifa.ifa_dstaddr = sintosa(&ia->ia_dstaddr); 492 ia->ia_ifa.ifa_netmask = sintosa(&ia->ia_sockmask); 493 #ifdef IPSELSRC 494 ia->ia_ifa.ifa_getifa = in_getifa; 495 #else /* IPSELSRC */ 496 ia->ia_ifa.ifa_getifa = NULL; 497 #endif /* IPSELSRC */ 498 ia->ia_sockmask.sin_len = 8; 499 ia->ia_sockmask.sin_family = AF_INET; 500 if (ifp->if_flags & IFF_BROADCAST) { 501 ia->ia_broadaddr.sin_len = sizeof(ia->ia_addr); 502 ia->ia_broadaddr.sin_family = AF_INET; 503 } 504 ia->ia_ifp = ifp; 505 ia->ia_idsalt = cprng_fast32() % 65535; 506 LIST_INIT(&ia->ia_multiaddrs); 507 IN_ADDRHASH_ENTRY_INIT(ia); 508 IN_ADDRLIST_ENTRY_INIT(ia); 509 ifa_psref_init(&ia->ia_ifa); 510 511 newifaddr = 1; 512 } 513 break; 514 515 case SIOCSIFBRDADDR: 516 if (kauth_authorize_network(curlwp->l_cred, KAUTH_NETWORK_INTERFACE, 517 KAUTH_REQ_NETWORK_INTERFACE_SETPRIV, ifp, (void *)cmd, 518 NULL) != 0) { 519 error = EPERM; 520 goto out; 521 } 522 /* FALLTHROUGH */ 523 524 case SIOCGIFADDR: 525 case SIOCGIFNETMASK: 526 case SIOCGIFDSTADDR: 527 case SIOCGIFBRDADDR: 528 if (ia == NULL) { 529 error = EADDRNOTAVAIL; 530 goto out; 531 } 532 break; 533 } 534 error = 0; 535 switch (cmd) { 536 537 case SIOCGIFADDR: 538 ifreq_setaddr(cmd, ifr, sintocsa(&ia->ia_addr)); 539 break; 540 541 case SIOCGIFBRDADDR: 542 if ((ifp->if_flags & IFF_BROADCAST) == 0) { 543 error = EINVAL; 544 goto out; 545 } 546 ifreq_setdstaddr(cmd, ifr, sintocsa(&ia->ia_broadaddr)); 547 break; 548 549 case SIOCGIFDSTADDR: 550 if ((ifp->if_flags & IFF_POINTOPOINT) == 0) { 551 error = EINVAL; 552 goto out; 553 } 554 ifreq_setdstaddr(cmd, ifr, sintocsa(&ia->ia_dstaddr)); 555 break; 556 557 case SIOCGIFNETMASK: 558 /* 559 * We keep the number of trailing zero bytes the sin_len field 560 * of ia_sockmask, so we fix this before we pass it back to 561 * userland. 562 */ 563 oldaddr = ia->ia_sockmask; 564 oldaddr.sin_len = sizeof(struct sockaddr_in); 565 ifreq_setaddr(cmd, ifr, (const void *)&oldaddr); 566 break; 567 568 case SIOCSIFDSTADDR: 569 if ((ifp->if_flags & IFF_POINTOPOINT) == 0) { 570 error = EINVAL; 571 goto out; 572 } 573 oldaddr = ia->ia_dstaddr; 574 ia->ia_dstaddr = *satocsin(ifreq_getdstaddr(cmd, ifr)); 575 if ((error = if_addr_init(ifp, &ia->ia_ifa, false)) != 0) { 576 ia->ia_dstaddr = oldaddr; 577 goto out; 578 } 579 if (ia->ia_flags & IFA_ROUTE) { 580 ia->ia_ifa.ifa_dstaddr = sintosa(&oldaddr); 581 rtinit(&ia->ia_ifa, RTM_DELETE, RTF_HOST); 582 ia->ia_ifa.ifa_dstaddr = sintosa(&ia->ia_dstaddr); 583 rtinit(&ia->ia_ifa, RTM_ADD, RTF_HOST|RTF_UP); 584 } 585 break; 586 587 case SIOCSIFBRDADDR: 588 if ((ifp->if_flags & IFF_BROADCAST) == 0) { 589 error = EINVAL; 590 goto out; 591 } 592 ia->ia_broadaddr = *satocsin(ifreq_getbroadaddr(cmd, ifr)); 593 break; 594 595 case SIOCSIFADDR: 596 if (!newifaddr) { 597 mutex_enter(&in_ifaddr_lock); 598 LIST_REMOVE(ia, ia_hash); 599 IN_ADDRHASH_WRITER_REMOVE(ia); 600 mutex_exit(&in_ifaddr_lock); 601 need_reinsert = true; 602 } 603 error = in_ifinit(ifp, ia, satocsin(ifreq_getaddr(cmd, ifr)), 604 1, hostIsNew); 605 606 run_hook = true; 607 break; 608 609 case SIOCSIFNETMASK: 610 in_ifscrub(ifp, ia); 611 ia->ia_sockmask = *satocsin(ifreq_getaddr(cmd, ifr)); 612 ia->ia_subnetmask = ia->ia_sockmask.sin_addr.s_addr; 613 if (!newifaddr) { 614 mutex_enter(&in_ifaddr_lock); 615 LIST_REMOVE(ia, ia_hash); 616 IN_ADDRHASH_WRITER_REMOVE(ia); 617 mutex_exit(&in_ifaddr_lock); 618 need_reinsert = true; 619 } 620 error = in_ifinit(ifp, ia, NULL, 0, 0); 621 break; 622 623 case SIOCAIFADDR: 624 maskIsNew = 0; 625 if (ifra->ifra_mask.sin_len) { 626 /* Only scrub if we control the prefix route, 627 * otherwise userland gets a bogus message */ 628 if ((ia->ia_flags & IFA_ROUTE)) 629 in_ifscrub(ifp, ia); 630 ia->ia_sockmask = ifra->ifra_mask; 631 ia->ia_subnetmask = ia->ia_sockmask.sin_addr.s_addr; 632 maskIsNew = 1; 633 } 634 if ((ifp->if_flags & IFF_POINTOPOINT) && 635 (ifra->ifra_dstaddr.sin_family == AF_INET)) { 636 /* Only scrub if we control the prefix route, 637 * otherwise userland gets a bogus message */ 638 if ((ia->ia_flags & IFA_ROUTE)) 639 in_ifscrub(ifp, ia); 640 ia->ia_dstaddr = ifra->ifra_dstaddr; 641 maskIsNew = 1; /* We lie; but the effect's the same */ 642 } 643 if (ifra->ifra_addr.sin_family == AF_INET && 644 (hostIsNew || maskIsNew)) { 645 if (!newifaddr) { 646 mutex_enter(&in_ifaddr_lock); 647 LIST_REMOVE(ia, ia_hash); 648 IN_ADDRHASH_WRITER_REMOVE(ia); 649 mutex_exit(&in_ifaddr_lock); 650 need_reinsert = true; 651 } 652 error = in_ifinit(ifp, ia, &ifra->ifra_addr, 0, 653 hostIsNew); 654 } 655 if ((ifp->if_flags & IFF_BROADCAST) && 656 (ifra->ifra_broadaddr.sin_family == AF_INET)) 657 ia->ia_broadaddr = ifra->ifra_broadaddr; 658 run_hook = true; 659 break; 660 661 case SIOCGIFALIAS: 662 ifra->ifra_mask = ia->ia_sockmask; 663 if ((ifp->if_flags & IFF_POINTOPOINT) && 664 (ia->ia_dstaddr.sin_family == AF_INET)) 665 ifra->ifra_dstaddr = ia->ia_dstaddr; 666 else if ((ifp->if_flags & IFF_BROADCAST) && 667 (ia->ia_broadaddr.sin_family == AF_INET)) 668 ifra->ifra_broadaddr = ia->ia_broadaddr; 669 else 670 memset(&ifra->ifra_broadaddr, 0, 671 sizeof(ifra->ifra_broadaddr)); 672 break; 673 674 case SIOCGIFAFLAG_IN: 675 ifr->ifr_addrflags = ia->ia4_flags; 676 break; 677 678 case SIOCDIFADDR: 679 ia4_release(ia, &psref); 680 in_purgeaddr(&ia->ia_ifa); 681 ia = NULL; 682 run_hook = true; 683 break; 684 685 #ifdef MROUTING 686 case SIOCGETVIFCNT: 687 case SIOCGETSGCNT: 688 error = mrt_ioctl(so, cmd, data); 689 break; 690 #endif /* MROUTING */ 691 692 default: 693 error = ENOTTY; 694 goto out; 695 } 696 697 /* 698 * XXX insert regardless of error to make in_purgeaddr below work. 699 * Need to improve. 700 */ 701 if (newifaddr) { 702 ifaref(&ia->ia_ifa); 703 ifa_insert(ifp, &ia->ia_ifa); 704 705 mutex_enter(&in_ifaddr_lock); 706 TAILQ_INSERT_TAIL(&in_ifaddrhead, ia, ia_list); 707 IN_ADDRLIST_WRITER_INSERT_TAIL(ia); 708 LIST_INSERT_HEAD(&IN_IFADDR_HASH(ia->ia_addr.sin_addr.s_addr), 709 ia, ia_hash); 710 IN_ADDRHASH_WRITER_INSERT_HEAD(ia); 711 mutex_exit(&in_ifaddr_lock); 712 } else if (need_reinsert) { 713 mutex_enter(&in_ifaddr_lock); 714 LIST_INSERT_HEAD(&IN_IFADDR_HASH(ia->ia_addr.sin_addr.s_addr), 715 ia, ia_hash); 716 IN_ADDRHASH_WRITER_INSERT_HEAD(ia); 717 mutex_exit(&in_ifaddr_lock); 718 } 719 720 if (error == 0) { 721 if (run_hook) 722 (void)pfil_run_hooks(if_pfil, 723 (struct mbuf **)cmd, ifp, PFIL_IFADDR); 724 } else if (newifaddr) { 725 KASSERT(ia != NULL); 726 in_purgeaddr(&ia->ia_ifa); 727 ia = NULL; 728 } 729 730 out: 731 if (!newifaddr && ia != NULL) 732 ia4_release(ia, &psref); 733 curlwp_bindx(bound); 734 return error; 735 } 736 737 int 738 in_control(struct socket *so, u_long cmd, void *data, struct ifnet *ifp) 739 { 740 int error; 741 742 mutex_enter(softnet_lock); 743 error = in_control0(so, cmd, data, ifp); 744 mutex_exit(softnet_lock); 745 746 return error; 747 } 748 749 /* Add ownaddr as loopback rtentry. */ 750 static void 751 in_ifaddlocal(struct ifaddr *ifa) 752 { 753 struct in_ifaddr *ia; 754 755 ia = (struct in_ifaddr *)ifa; 756 if (ia->ia_addr.sin_addr.s_addr == INADDR_ANY || 757 (ia->ia_ifp->if_flags & IFF_POINTOPOINT && 758 in_hosteq(ia->ia_dstaddr.sin_addr, ia->ia_addr.sin_addr))) 759 { 760 rt_newaddrmsg(RTM_NEWADDR, ifa, 0, NULL); 761 return; 762 } 763 764 rt_ifa_addlocal(ifa); 765 } 766 767 /* Remove loopback entry of ownaddr */ 768 static void 769 in_ifremlocal(struct ifaddr *ifa) 770 { 771 struct in_ifaddr *ia, *p; 772 struct ifaddr *alt_ifa = NULL; 773 int ia_count = 0; 774 int s; 775 struct psref psref; 776 int bound = curlwp_bind(); 777 778 ia = (struct in_ifaddr *)ifa; 779 /* Delete the entry if exactly one ifaddr matches the 780 * address, ifa->ifa_addr. */ 781 s = pserialize_read_enter(); 782 IN_ADDRLIST_READER_FOREACH(p) { 783 if (!in_hosteq(p->ia_addr.sin_addr, ia->ia_addr.sin_addr)) 784 continue; 785 if (p->ia_ifp != ia->ia_ifp) 786 alt_ifa = &p->ia_ifa; 787 if (++ia_count > 1 && alt_ifa != NULL) 788 break; 789 } 790 if (alt_ifa != NULL && ia_count > 1) 791 ifa_acquire(alt_ifa, &psref); 792 pserialize_read_exit(s); 793 794 if (ia_count == 0) 795 goto out; 796 797 rt_ifa_remlocal(ifa, ia_count == 1 ? NULL : alt_ifa); 798 if (alt_ifa != NULL && ia_count > 1) 799 ifa_release(alt_ifa, &psref); 800 out: 801 curlwp_bindx(bound); 802 } 803 804 /* 805 * Depends on it isn't called in concurrent. It should be guaranteed 806 * by ifa->ifa_ifp's ioctl lock. The possible callers are in_control 807 * and if_purgeaddrs; the former is called iva ifa->ifa_ifp's ioctl 808 * and the latter is called via ifa->ifa_ifp's if_detach. The functions 809 * never be executed in concurrent. 810 */ 811 void 812 in_purgeaddr(struct ifaddr *ifa) 813 { 814 struct ifnet *ifp = ifa->ifa_ifp; 815 struct in_ifaddr *ia = (void *) ifa; 816 817 KASSERT(!ifa_held(ifa)); 818 819 /* stop DAD processing */ 820 if (ia->ia_dad_stop != NULL) 821 ia->ia_dad_stop(ifa); 822 823 in_ifscrub(ifp, ia); 824 in_ifremlocal(ifa); 825 if (ia->ia_allhosts != NULL) 826 in_delmulti(ia->ia_allhosts); 827 828 mutex_enter(&in_ifaddr_lock); 829 LIST_REMOVE(ia, ia_hash); 830 IN_ADDRHASH_WRITER_REMOVE(ia); 831 TAILQ_REMOVE(&in_ifaddrhead, ia, ia_list); 832 IN_ADDRLIST_WRITER_REMOVE(ia); 833 ifa_remove(ifp, &ia->ia_ifa); 834 mutex_exit(&in_ifaddr_lock); 835 836 IN_ADDRHASH_ENTRY_DESTROY(ia); 837 IN_ADDRLIST_ENTRY_DESTROY(ia); 838 ifafree(&ia->ia_ifa); 839 in_setmaxmtu(); 840 } 841 842 void 843 in_purgeif(struct ifnet *ifp) /* MUST be called at splsoftnet() */ 844 { 845 if_purgeaddrs(ifp, AF_INET, in_purgeaddr); 846 igmp_purgeif(ifp); /* manipulates pools */ 847 #ifdef MROUTING 848 ip_mrouter_detach(ifp); 849 #endif 850 } 851 852 /* 853 * SIOC[GAD]LIFADDR. 854 * SIOCGLIFADDR: get first address. (???) 855 * SIOCGLIFADDR with IFLR_PREFIX: 856 * get first address that matches the specified prefix. 857 * SIOCALIFADDR: add the specified address. 858 * SIOCALIFADDR with IFLR_PREFIX: 859 * EINVAL since we can't deduce hostid part of the address. 860 * SIOCDLIFADDR: delete the specified address. 861 * SIOCDLIFADDR with IFLR_PREFIX: 862 * delete the first address that matches the specified prefix. 863 * return values: 864 * EINVAL on invalid parameters 865 * EADDRNOTAVAIL on prefix match failed/specified address not found 866 * other values may be returned from in_ioctl() 867 */ 868 static int 869 in_lifaddr_ioctl(struct socket *so, u_long cmd, void *data, 870 struct ifnet *ifp) 871 { 872 struct if_laddrreq *iflr = (struct if_laddrreq *)data; 873 struct ifaddr *ifa; 874 struct sockaddr *sa; 875 876 /* sanity checks */ 877 if (data == NULL || ifp == NULL) { 878 panic("invalid argument to in_lifaddr_ioctl"); 879 /*NOTRECHED*/ 880 } 881 882 switch (cmd) { 883 case SIOCGLIFADDR: 884 /* address must be specified on GET with IFLR_PREFIX */ 885 if ((iflr->flags & IFLR_PREFIX) == 0) 886 break; 887 /*FALLTHROUGH*/ 888 case SIOCALIFADDR: 889 case SIOCDLIFADDR: 890 /* address must be specified on ADD and DELETE */ 891 sa = (struct sockaddr *)&iflr->addr; 892 if (sa->sa_family != AF_INET) 893 return EINVAL; 894 if (sa->sa_len != sizeof(struct sockaddr_in)) 895 return EINVAL; 896 /* XXX need improvement */ 897 sa = (struct sockaddr *)&iflr->dstaddr; 898 if (sa->sa_family != AF_UNSPEC && sa->sa_family != AF_INET) 899 return EINVAL; 900 if (sa->sa_len != 0 && sa->sa_len != sizeof(struct sockaddr_in)) 901 return EINVAL; 902 break; 903 default: /*shouldn't happen*/ 904 #if 0 905 panic("invalid cmd to in_lifaddr_ioctl"); 906 /*NOTREACHED*/ 907 #else 908 return EOPNOTSUPP; 909 #endif 910 } 911 if (sizeof(struct in_addr) * NBBY < iflr->prefixlen) 912 return EINVAL; 913 914 switch (cmd) { 915 case SIOCALIFADDR: 916 { 917 struct in_aliasreq ifra; 918 919 if (iflr->flags & IFLR_PREFIX) 920 return EINVAL; 921 922 /* copy args to in_aliasreq, perform ioctl(SIOCAIFADDR). */ 923 memset(&ifra, 0, sizeof(ifra)); 924 memcpy(ifra.ifra_name, iflr->iflr_name, 925 sizeof(ifra.ifra_name)); 926 927 memcpy(&ifra.ifra_addr, &iflr->addr, 928 ((struct sockaddr *)&iflr->addr)->sa_len); 929 930 if (((struct sockaddr *)&iflr->dstaddr)->sa_family) { /*XXX*/ 931 memcpy(&ifra.ifra_dstaddr, &iflr->dstaddr, 932 ((struct sockaddr *)&iflr->dstaddr)->sa_len); 933 } 934 935 ifra.ifra_mask.sin_family = AF_INET; 936 ifra.ifra_mask.sin_len = sizeof(struct sockaddr_in); 937 in_len2mask(&ifra.ifra_mask.sin_addr, iflr->prefixlen); 938 939 return in_control(so, SIOCAIFADDR, &ifra, ifp); 940 } 941 case SIOCGLIFADDR: 942 case SIOCDLIFADDR: 943 { 944 struct in_ifaddr *ia; 945 struct in_addr mask, candidate, match; 946 struct sockaddr_in *sin; 947 int cmp; 948 949 memset(&mask, 0, sizeof(mask)); 950 memset(&match, 0, sizeof(match)); /* XXX gcc */ 951 if (iflr->flags & IFLR_PREFIX) { 952 /* lookup a prefix rather than address. */ 953 in_len2mask(&mask, iflr->prefixlen); 954 955 sin = (struct sockaddr_in *)&iflr->addr; 956 match.s_addr = sin->sin_addr.s_addr; 957 match.s_addr &= mask.s_addr; 958 959 /* if you set extra bits, that's wrong */ 960 if (match.s_addr != sin->sin_addr.s_addr) 961 return EINVAL; 962 963 cmp = 1; 964 } else { 965 if (cmd == SIOCGLIFADDR) { 966 /* on getting an address, take the 1st match */ 967 cmp = 0; /*XXX*/ 968 } else { 969 /* on deleting an address, do exact match */ 970 in_len2mask(&mask, 32); 971 sin = (struct sockaddr_in *)&iflr->addr; 972 match.s_addr = sin->sin_addr.s_addr; 973 974 cmp = 1; 975 } 976 } 977 978 IFADDR_READER_FOREACH(ifa, ifp) { 979 if (ifa->ifa_addr->sa_family != AF_INET) 980 continue; 981 if (cmp == 0) 982 break; 983 candidate.s_addr = ((struct sockaddr_in *)ifa->ifa_addr)->sin_addr.s_addr; 984 candidate.s_addr &= mask.s_addr; 985 if (candidate.s_addr == match.s_addr) 986 break; 987 } 988 if (ifa == NULL) 989 return EADDRNOTAVAIL; 990 ia = (struct in_ifaddr *)ifa; 991 992 if (cmd == SIOCGLIFADDR) { 993 /* fill in the if_laddrreq structure */ 994 memcpy(&iflr->addr, &ia->ia_addr, ia->ia_addr.sin_len); 995 996 if ((ifp->if_flags & IFF_POINTOPOINT) != 0) { 997 memcpy(&iflr->dstaddr, &ia->ia_dstaddr, 998 ia->ia_dstaddr.sin_len); 999 } else 1000 memset(&iflr->dstaddr, 0, sizeof(iflr->dstaddr)); 1001 1002 iflr->prefixlen = 1003 in_mask2len(&ia->ia_sockmask.sin_addr); 1004 1005 iflr->flags = 0; /*XXX*/ 1006 1007 return 0; 1008 } else { 1009 struct in_aliasreq ifra; 1010 1011 /* fill in_aliasreq and do ioctl(SIOCDIFADDR) */ 1012 memset(&ifra, 0, sizeof(ifra)); 1013 memcpy(ifra.ifra_name, iflr->iflr_name, 1014 sizeof(ifra.ifra_name)); 1015 1016 memcpy(&ifra.ifra_addr, &ia->ia_addr, 1017 ia->ia_addr.sin_len); 1018 if ((ifp->if_flags & IFF_POINTOPOINT) != 0) { 1019 memcpy(&ifra.ifra_dstaddr, &ia->ia_dstaddr, 1020 ia->ia_dstaddr.sin_len); 1021 } 1022 memcpy(&ifra.ifra_dstaddr, &ia->ia_sockmask, 1023 ia->ia_sockmask.sin_len); 1024 1025 return in_control(so, SIOCDIFADDR, &ifra, ifp); 1026 } 1027 } 1028 } 1029 1030 return EOPNOTSUPP; /*just for safety*/ 1031 } 1032 1033 /* 1034 * Delete any existing route for an interface. 1035 */ 1036 void 1037 in_ifscrub(struct ifnet *ifp, struct in_ifaddr *ia) 1038 { 1039 1040 in_scrubprefix(ia); 1041 } 1042 1043 /* 1044 * Initialize an interface's internet address 1045 * and routing table entry. 1046 */ 1047 int 1048 in_ifinit(struct ifnet *ifp, struct in_ifaddr *ia, 1049 const struct sockaddr_in *sin, int scrub, int hostIsNew) 1050 { 1051 u_int32_t i; 1052 struct sockaddr_in oldaddr; 1053 int s = splnet(), flags = RTF_UP, error; 1054 1055 if (sin == NULL) 1056 sin = &ia->ia_addr; 1057 1058 /* 1059 * Set up new addresses. 1060 */ 1061 oldaddr = ia->ia_addr; 1062 ia->ia_addr = *sin; 1063 1064 /* Set IN_IFF flags early for if_addr_init() */ 1065 if (hostIsNew && if_do_dad(ifp) && !in_nullhost(ia->ia_addr.sin_addr)) { 1066 if (ifp->if_link_state == LINK_STATE_DOWN) 1067 ia->ia4_flags |= IN_IFF_DETACHED; 1068 else 1069 /* State the intent to try DAD if possible */ 1070 ia->ia4_flags |= IN_IFF_TRYTENTATIVE; 1071 } 1072 1073 /* 1074 * Give the interface a chance to initialize 1075 * if this is its first address, 1076 * and to validate the address if necessary. 1077 */ 1078 if ((error = if_addr_init(ifp, &ia->ia_ifa, true)) != 0) 1079 goto bad; 1080 /* Now clear the try tentative flag, it's job is done. */ 1081 ia->ia4_flags &= ~IN_IFF_TRYTENTATIVE; 1082 splx(s); 1083 1084 if (scrub) { 1085 ia->ia_ifa.ifa_addr = sintosa(&oldaddr); 1086 in_ifscrub(ifp, ia); 1087 ia->ia_ifa.ifa_addr = sintosa(&ia->ia_addr); 1088 } 1089 1090 /* Add the local route to the address */ 1091 in_ifaddlocal(&ia->ia_ifa); 1092 1093 i = ia->ia_addr.sin_addr.s_addr; 1094 if (IN_CLASSA(i)) 1095 ia->ia_netmask = IN_CLASSA_NET; 1096 else if (IN_CLASSB(i)) 1097 ia->ia_netmask = IN_CLASSB_NET; 1098 else 1099 ia->ia_netmask = IN_CLASSC_NET; 1100 /* 1101 * The subnet mask usually includes at least the standard network part, 1102 * but may may be smaller in the case of supernetting. 1103 * If it is set, we believe it. 1104 */ 1105 if (ia->ia_subnetmask == 0) { 1106 ia->ia_subnetmask = ia->ia_netmask; 1107 ia->ia_sockmask.sin_addr.s_addr = ia->ia_subnetmask; 1108 } else 1109 ia->ia_netmask &= ia->ia_subnetmask; 1110 1111 ia->ia_net = i & ia->ia_netmask; 1112 ia->ia_subnet = i & ia->ia_subnetmask; 1113 in_socktrim(&ia->ia_sockmask); 1114 /* re-calculate the "in_maxmtu" value */ 1115 in_setmaxmtu(); 1116 /* 1117 * Add route for the network. 1118 */ 1119 ia->ia_ifa.ifa_metric = ifp->if_metric; 1120 if (ifp->if_flags & IFF_BROADCAST) { 1121 ia->ia_broadaddr.sin_addr.s_addr = 1122 ia->ia_subnet | ~ia->ia_subnetmask; 1123 ia->ia_netbroadcast.s_addr = 1124 ia->ia_net | ~ia->ia_netmask; 1125 } else if (ifp->if_flags & IFF_LOOPBACK) { 1126 ia->ia_dstaddr = ia->ia_addr; 1127 flags |= RTF_HOST; 1128 } else if (ifp->if_flags & IFF_POINTOPOINT) { 1129 if (ia->ia_dstaddr.sin_family != AF_INET) 1130 return (0); 1131 flags |= RTF_HOST; 1132 } 1133 error = in_addprefix(ia, flags); 1134 /* 1135 * If the interface supports multicast, join the "all hosts" 1136 * multicast group on that interface. 1137 */ 1138 if ((ifp->if_flags & IFF_MULTICAST) != 0 && ia->ia_allhosts == NULL) { 1139 struct in_addr addr; 1140 1141 addr.s_addr = INADDR_ALLHOSTS_GROUP; 1142 ia->ia_allhosts = in_addmulti(&addr, ifp); 1143 } 1144 1145 if (hostIsNew && if_do_dad(ifp) && 1146 !in_nullhost(ia->ia_addr.sin_addr) && 1147 ia->ia4_flags & IN_IFF_TENTATIVE) 1148 ia->ia_dad_start((struct ifaddr *)ia); 1149 1150 return (error); 1151 bad: 1152 splx(s); 1153 ia->ia_addr = oldaddr; 1154 return (error); 1155 } 1156 1157 #define rtinitflags(x) \ 1158 ((((x)->ia_ifp->if_flags & (IFF_LOOPBACK | IFF_POINTOPOINT)) != 0) \ 1159 ? RTF_HOST : 0) 1160 1161 /* 1162 * add a route to prefix ("connected route" in cisco terminology). 1163 * does nothing if there's some interface address with the same prefix already. 1164 */ 1165 static int 1166 in_addprefix(struct in_ifaddr *target, int flags) 1167 { 1168 struct in_ifaddr *ia; 1169 struct in_addr prefix, mask, p; 1170 int error; 1171 int s; 1172 1173 if ((flags & RTF_HOST) != 0) 1174 prefix = target->ia_dstaddr.sin_addr; 1175 else { 1176 prefix = target->ia_addr.sin_addr; 1177 mask = target->ia_sockmask.sin_addr; 1178 prefix.s_addr &= mask.s_addr; 1179 } 1180 1181 s = pserialize_read_enter(); 1182 IN_ADDRLIST_READER_FOREACH(ia) { 1183 if (rtinitflags(ia)) 1184 p = ia->ia_dstaddr.sin_addr; 1185 else { 1186 p = ia->ia_addr.sin_addr; 1187 p.s_addr &= ia->ia_sockmask.sin_addr.s_addr; 1188 } 1189 1190 if (prefix.s_addr != p.s_addr) 1191 continue; 1192 1193 /* 1194 * if we got a matching prefix route inserted by other 1195 * interface address, we don't need to bother 1196 * 1197 * XXX RADIX_MPATH implications here? -dyoung 1198 */ 1199 if (ia->ia_flags & IFA_ROUTE) { 1200 pserialize_read_exit(s); 1201 return 0; 1202 } 1203 } 1204 pserialize_read_exit(s); 1205 1206 /* 1207 * noone seem to have prefix route. insert it. 1208 */ 1209 error = rtinit(&target->ia_ifa, RTM_ADD, flags); 1210 if (error == 0) 1211 target->ia_flags |= IFA_ROUTE; 1212 else if (error == EEXIST) { 1213 /* 1214 * the fact the route already exists is not an error. 1215 */ 1216 error = 0; 1217 } 1218 return error; 1219 } 1220 1221 /* 1222 * remove a route to prefix ("connected route" in cisco terminology). 1223 * re-installs the route by using another interface address, if there's one 1224 * with the same prefix (otherwise we lose the route mistakenly). 1225 */ 1226 static int 1227 in_scrubprefix(struct in_ifaddr *target) 1228 { 1229 struct in_ifaddr *ia; 1230 struct in_addr prefix, mask, p; 1231 int error; 1232 int s; 1233 1234 /* If we don't have IFA_ROUTE we should still inform userland */ 1235 if ((target->ia_flags & IFA_ROUTE) == 0) 1236 return 0; 1237 1238 if (rtinitflags(target)) 1239 prefix = target->ia_dstaddr.sin_addr; 1240 else { 1241 prefix = target->ia_addr.sin_addr; 1242 mask = target->ia_sockmask.sin_addr; 1243 prefix.s_addr &= mask.s_addr; 1244 } 1245 1246 s = pserialize_read_enter(); 1247 IN_ADDRLIST_READER_FOREACH(ia) { 1248 if (rtinitflags(ia)) 1249 p = ia->ia_dstaddr.sin_addr; 1250 else { 1251 p = ia->ia_addr.sin_addr; 1252 p.s_addr &= ia->ia_sockmask.sin_addr.s_addr; 1253 } 1254 1255 if (prefix.s_addr != p.s_addr) 1256 continue; 1257 1258 /* 1259 * if we got a matching prefix route, move IFA_ROUTE to him 1260 */ 1261 if ((ia->ia_flags & IFA_ROUTE) == 0) { 1262 struct psref psref; 1263 int bound = curlwp_bind(); 1264 1265 ia4_acquire(ia, &psref); 1266 pserialize_read_exit(s); 1267 1268 rtinit(&target->ia_ifa, RTM_DELETE, 1269 rtinitflags(target)); 1270 target->ia_flags &= ~IFA_ROUTE; 1271 1272 error = rtinit(&ia->ia_ifa, RTM_ADD, 1273 rtinitflags(ia) | RTF_UP); 1274 if (error == 0) 1275 ia->ia_flags |= IFA_ROUTE; 1276 1277 ia4_release(ia, &psref); 1278 curlwp_bindx(bound); 1279 1280 return error; 1281 } 1282 } 1283 pserialize_read_exit(s); 1284 1285 /* 1286 * noone seem to have prefix route. remove it. 1287 */ 1288 rtinit(&target->ia_ifa, RTM_DELETE, rtinitflags(target)); 1289 target->ia_flags &= ~IFA_ROUTE; 1290 return 0; 1291 } 1292 1293 #undef rtinitflags 1294 1295 /* 1296 * Return 1 if the address might be a local broadcast address. 1297 */ 1298 int 1299 in_broadcast(struct in_addr in, struct ifnet *ifp) 1300 { 1301 struct ifaddr *ifa; 1302 int s; 1303 1304 KASSERT(ifp != NULL); 1305 1306 if (in.s_addr == INADDR_BROADCAST || 1307 in_nullhost(in)) 1308 return 1; 1309 if ((ifp->if_flags & IFF_BROADCAST) == 0) 1310 return 0; 1311 /* 1312 * Look through the list of addresses for a match 1313 * with a broadcast address. 1314 */ 1315 #define ia (ifatoia(ifa)) 1316 s = pserialize_read_enter(); 1317 IFADDR_READER_FOREACH(ifa, ifp) { 1318 if (ifa->ifa_addr->sa_family == AF_INET && 1319 !in_hosteq(in, ia->ia_addr.sin_addr) && 1320 (in_hosteq(in, ia->ia_broadaddr.sin_addr) || 1321 in_hosteq(in, ia->ia_netbroadcast) || 1322 (hostzeroisbroadcast && 1323 /* 1324 * Check for old-style (host 0) broadcast. 1325 */ 1326 (in.s_addr == ia->ia_subnet || 1327 in.s_addr == ia->ia_net)))) { 1328 pserialize_read_exit(s); 1329 return 1; 1330 } 1331 } 1332 pserialize_read_exit(s); 1333 return (0); 1334 #undef ia 1335 } 1336 1337 /* 1338 * perform DAD when interface becomes IFF_UP. 1339 */ 1340 void 1341 in_if_link_up(struct ifnet *ifp) 1342 { 1343 struct ifaddr *ifa; 1344 struct in_ifaddr *ia; 1345 1346 /* Ensure it's sane to run DAD */ 1347 if (ifp->if_link_state == LINK_STATE_DOWN) 1348 return; 1349 if ((ifp->if_flags & (IFF_UP|IFF_RUNNING)) != (IFF_UP|IFF_RUNNING)) 1350 return; 1351 1352 IFADDR_READER_FOREACH(ifa, ifp) { 1353 if (ifa->ifa_addr->sa_family != AF_INET) 1354 continue; 1355 ia = (struct in_ifaddr *)ifa; 1356 1357 /* If detached then mark as tentative */ 1358 if (ia->ia4_flags & IN_IFF_DETACHED) { 1359 ia->ia4_flags &= ~IN_IFF_DETACHED; 1360 if (if_do_dad(ifp) && ia->ia_dad_start != NULL) 1361 ia->ia4_flags |= IN_IFF_TENTATIVE; 1362 else if ((ia->ia4_flags & IN_IFF_TENTATIVE) == 0) 1363 rt_newaddrmsg(RTM_NEWADDR, ifa, 0, NULL); 1364 } 1365 1366 if (ia->ia4_flags & IN_IFF_TENTATIVE) { 1367 /* Clear the duplicated flag as we're starting DAD. */ 1368 ia->ia4_flags &= ~IN_IFF_DUPLICATED; 1369 ia->ia_dad_start(ifa); 1370 } 1371 } 1372 } 1373 1374 void 1375 in_if_up(struct ifnet *ifp) 1376 { 1377 1378 /* interface may not support link state, so bring it up also */ 1379 in_if_link_up(ifp); 1380 } 1381 1382 /* 1383 * Mark all addresses as detached. 1384 */ 1385 void 1386 in_if_link_down(struct ifnet *ifp) 1387 { 1388 struct ifaddr *ifa; 1389 struct in_ifaddr *ia; 1390 1391 IFADDR_READER_FOREACH(ifa, ifp) { 1392 if (ifa->ifa_addr->sa_family != AF_INET) 1393 continue; 1394 ia = (struct in_ifaddr *)ifa; 1395 1396 /* Stop DAD processing */ 1397 if (ia->ia_dad_stop != NULL) 1398 ia->ia_dad_stop(ifa); 1399 1400 /* 1401 * Mark the address as detached. 1402 */ 1403 if (!(ia->ia4_flags & IN_IFF_DETACHED)) { 1404 ia->ia4_flags |= IN_IFF_DETACHED; 1405 ia->ia4_flags &= 1406 ~(IN_IFF_TENTATIVE | IN_IFF_DUPLICATED); 1407 rt_newaddrmsg(RTM_NEWADDR, ifa, 0, NULL); 1408 } 1409 } 1410 } 1411 1412 void 1413 in_if_down(struct ifnet *ifp) 1414 { 1415 1416 in_if_link_down(ifp); 1417 } 1418 1419 void 1420 in_if_link_state_change(struct ifnet *ifp, int link_state) 1421 { 1422 1423 switch (link_state) { 1424 case LINK_STATE_DOWN: 1425 in_if_link_down(ifp); 1426 break; 1427 case LINK_STATE_UP: 1428 in_if_link_up(ifp); 1429 break; 1430 } 1431 } 1432 1433 /* 1434 * in_lookup_multi: look up the in_multi record for a given IP 1435 * multicast address on a given interface. If no matching record is 1436 * found, return NULL. 1437 */ 1438 struct in_multi * 1439 in_lookup_multi(struct in_addr addr, ifnet_t *ifp) 1440 { 1441 struct in_multi *inm; 1442 1443 KASSERT(rw_lock_held(&in_multilock)); 1444 1445 LIST_FOREACH(inm, &IN_MULTI_HASH(addr.s_addr, ifp), inm_list) { 1446 if (in_hosteq(inm->inm_addr, addr) && inm->inm_ifp == ifp) 1447 break; 1448 } 1449 return inm; 1450 } 1451 1452 /* 1453 * in_multi_group: check whether the address belongs to an IP multicast 1454 * group we are joined on this interface. Returns true or false. 1455 */ 1456 bool 1457 in_multi_group(struct in_addr addr, ifnet_t *ifp, int flags) 1458 { 1459 bool ingroup; 1460 1461 if (__predict_true(flags & IP_IGMP_MCAST) == 0) { 1462 rw_enter(&in_multilock, RW_READER); 1463 ingroup = in_lookup_multi(addr, ifp) != NULL; 1464 rw_exit(&in_multilock); 1465 } else { 1466 /* XXX Recursive call from ip_output(). */ 1467 KASSERT(rw_lock_held(&in_multilock)); 1468 ingroup = in_lookup_multi(addr, ifp) != NULL; 1469 } 1470 return ingroup; 1471 } 1472 1473 /* 1474 * Add an address to the list of IP multicast addresses for a given interface. 1475 */ 1476 struct in_multi * 1477 in_addmulti(struct in_addr *ap, ifnet_t *ifp) 1478 { 1479 struct sockaddr_in sin; 1480 struct in_multi *inm; 1481 1482 /* 1483 * See if address already in list. 1484 */ 1485 rw_enter(&in_multilock, RW_WRITER); 1486 inm = in_lookup_multi(*ap, ifp); 1487 if (inm != NULL) { 1488 /* 1489 * Found it; just increment the reference count. 1490 */ 1491 inm->inm_refcount++; 1492 rw_exit(&in_multilock); 1493 return inm; 1494 } 1495 1496 /* 1497 * New address; allocate a new multicast record. 1498 */ 1499 inm = pool_get(&inmulti_pool, PR_NOWAIT); 1500 if (inm == NULL) { 1501 rw_exit(&in_multilock); 1502 return NULL; 1503 } 1504 inm->inm_addr = *ap; 1505 inm->inm_ifp = ifp; 1506 inm->inm_refcount = 1; 1507 1508 /* 1509 * Ask the network driver to update its multicast reception 1510 * filter appropriately for the new address. 1511 */ 1512 sockaddr_in_init(&sin, ap, 0); 1513 if (if_mcast_op(ifp, SIOCADDMULTI, sintosa(&sin)) != 0) { 1514 rw_exit(&in_multilock); 1515 pool_put(&inmulti_pool, inm); 1516 return NULL; 1517 } 1518 1519 /* 1520 * Let IGMP know that we have joined a new IP multicast group. 1521 */ 1522 if (igmp_joingroup(inm) != 0) { 1523 rw_exit(&in_multilock); 1524 pool_put(&inmulti_pool, inm); 1525 return NULL; 1526 } 1527 LIST_INSERT_HEAD( 1528 &IN_MULTI_HASH(inm->inm_addr.s_addr, ifp), 1529 inm, inm_list); 1530 in_multientries++; 1531 rw_exit(&in_multilock); 1532 1533 return inm; 1534 } 1535 1536 /* 1537 * Delete a multicast address record. 1538 */ 1539 void 1540 in_delmulti(struct in_multi *inm) 1541 { 1542 struct sockaddr_in sin; 1543 1544 rw_enter(&in_multilock, RW_WRITER); 1545 if (--inm->inm_refcount > 0) { 1546 rw_exit(&in_multilock); 1547 return; 1548 } 1549 1550 /* 1551 * No remaining claims to this record; let IGMP know that 1552 * we are leaving the multicast group. 1553 */ 1554 igmp_leavegroup(inm); 1555 1556 /* 1557 * Notify the network driver to update its multicast reception 1558 * filter. 1559 */ 1560 sockaddr_in_init(&sin, &inm->inm_addr, 0); 1561 if_mcast_op(inm->inm_ifp, SIOCDELMULTI, sintosa(&sin)); 1562 1563 /* 1564 * Unlink from list. 1565 */ 1566 LIST_REMOVE(inm, inm_list); 1567 in_multientries--; 1568 rw_exit(&in_multilock); 1569 1570 pool_put(&inmulti_pool, inm); 1571 } 1572 1573 /* 1574 * in_next_multi: step through all of the in_multi records, one at a time. 1575 * The current position is remembered in "step", which the caller must 1576 * provide. in_first_multi(), below, must be called to initialize "step" 1577 * and get the first record. Both macros return a NULL "inm" when there 1578 * are no remaining records. 1579 */ 1580 struct in_multi * 1581 in_next_multi(struct in_multistep *step) 1582 { 1583 struct in_multi *inm; 1584 1585 KASSERT(rw_lock_held(&in_multilock)); 1586 1587 while (step->i_inm == NULL && step->i_n < IN_MULTI_HASH_SIZE) { 1588 step->i_inm = LIST_FIRST(&in_multihashtbl[++step->i_n]); 1589 } 1590 if ((inm = step->i_inm) != NULL) { 1591 step->i_inm = LIST_NEXT(inm, inm_list); 1592 } 1593 return inm; 1594 } 1595 1596 struct in_multi * 1597 in_first_multi(struct in_multistep *step) 1598 { 1599 KASSERT(rw_lock_held(&in_multilock)); 1600 1601 step->i_n = 0; 1602 step->i_inm = LIST_FIRST(&in_multihashtbl[0]); 1603 return in_next_multi(step); 1604 } 1605 1606 void 1607 in_multi_lock(int op) 1608 { 1609 rw_enter(&in_multilock, op); 1610 } 1611 1612 void 1613 in_multi_unlock(void) 1614 { 1615 rw_exit(&in_multilock); 1616 } 1617 1618 int 1619 in_multi_lock_held(void) 1620 { 1621 return rw_lock_held(&in_multilock); 1622 } 1623 1624 struct in_ifaddr * 1625 in_selectsrc(struct sockaddr_in *sin, struct route *ro, 1626 int soopts, struct ip_moptions *mopts, int *errorp, struct psref *psref) 1627 { 1628 struct rtentry *rt = NULL; 1629 struct in_ifaddr *ia = NULL; 1630 1631 KASSERT(ISSET(curlwp->l_pflag, LP_BOUND)); 1632 /* 1633 * If route is known or can be allocated now, take the 1634 * source address from the interface. Otherwise, punt. 1635 */ 1636 if ((soopts & SO_DONTROUTE) != 0) 1637 rtcache_free(ro); 1638 else { 1639 union { 1640 struct sockaddr dst; 1641 struct sockaddr_in dst4; 1642 } u; 1643 1644 sockaddr_in_init(&u.dst4, &sin->sin_addr, 0); 1645 rt = rtcache_lookup(ro, &u.dst); 1646 } 1647 /* 1648 * If we found a route, use the address 1649 * corresponding to the outgoing interface 1650 * unless it is the loopback (in case a route 1651 * to our address on another net goes to loopback). 1652 * 1653 * XXX Is this still true? Do we care? 1654 */ 1655 if (rt != NULL && (rt->rt_ifp->if_flags & IFF_LOOPBACK) == 0) { 1656 int s; 1657 struct ifaddr *ifa; 1658 /* 1659 * Just in case. May not need to do this workaround. 1660 * Revisit when working on rtentry MP-ification. 1661 */ 1662 s = pserialize_read_enter(); 1663 IFADDR_READER_FOREACH(ifa, rt->rt_ifp) { 1664 if (ifa == rt->rt_ifa) 1665 break; 1666 } 1667 if (ifa != NULL) 1668 ifa_acquire(ifa, psref); 1669 pserialize_read_exit(s); 1670 1671 ia = ifatoia(ifa); 1672 } 1673 if (ia == NULL) { 1674 u_int16_t fport = sin->sin_port; 1675 struct ifaddr *ifa; 1676 int s; 1677 1678 sin->sin_port = 0; 1679 ifa = ifa_ifwithladdr_psref(sintosa(sin), psref); 1680 sin->sin_port = fport; 1681 if (ifa == NULL) { 1682 /* Find 1st non-loopback AF_INET address */ 1683 s = pserialize_read_enter(); 1684 IN_ADDRLIST_READER_FOREACH(ia) { 1685 if (!(ia->ia_ifp->if_flags & IFF_LOOPBACK)) 1686 break; 1687 } 1688 if (ia != NULL) 1689 ia4_acquire(ia, psref); 1690 pserialize_read_exit(s); 1691 } else { 1692 /* ia is already referenced by psref */ 1693 ia = ifatoia(ifa); 1694 } 1695 if (ia == NULL) { 1696 *errorp = EADDRNOTAVAIL; 1697 return NULL; 1698 } 1699 } 1700 /* 1701 * If the destination address is multicast and an outgoing 1702 * interface has been set as a multicast option, use the 1703 * address of that interface as our source address. 1704 */ 1705 if (IN_MULTICAST(sin->sin_addr.s_addr) && mopts != NULL) { 1706 struct ip_moptions *imo; 1707 1708 imo = mopts; 1709 if (imo->imo_multicast_if_index != 0) { 1710 struct ifnet *ifp; 1711 int s; 1712 1713 if (ia != NULL) 1714 ia4_release(ia, psref); 1715 s = pserialize_read_enter(); 1716 ifp = if_byindex(imo->imo_multicast_if_index); 1717 if (ifp != NULL) { 1718 /* XXX */ 1719 ia = in_get_ia_from_ifp_psref(ifp, psref); 1720 } else 1721 ia = NULL; 1722 if (ia == NULL || ia->ia4_flags & IN_IFF_NOTREADY) { 1723 pserialize_read_exit(s); 1724 if (ia != NULL) 1725 ia4_release(ia, psref); 1726 *errorp = EADDRNOTAVAIL; 1727 return NULL; 1728 } 1729 pserialize_read_exit(s); 1730 } 1731 } 1732 if (ia->ia_ifa.ifa_getifa != NULL) { 1733 ia = ifatoia((*ia->ia_ifa.ifa_getifa)(&ia->ia_ifa, 1734 sintosa(sin))); 1735 if (ia == NULL) { 1736 *errorp = EADDRNOTAVAIL; 1737 return NULL; 1738 } 1739 /* FIXME NOMPSAFE */ 1740 ia4_acquire(ia, psref); 1741 } 1742 #ifdef GETIFA_DEBUG 1743 else 1744 printf("%s: missing ifa_getifa\n", __func__); 1745 #endif 1746 return ia; 1747 } 1748 1749 #if NARP > 0 1750 1751 struct in_llentry { 1752 struct llentry base; 1753 }; 1754 1755 #define IN_LLTBL_DEFAULT_HSIZE 32 1756 #define IN_LLTBL_HASH(k, h) \ 1757 (((((((k >> 8) ^ k) >> 8) ^ k) >> 8) ^ k) & ((h) - 1)) 1758 1759 /* 1760 * Do actual deallocation of @lle. 1761 * Called by LLE_FREE_LOCKED when number of references 1762 * drops to zero. 1763 */ 1764 static void 1765 in_lltable_destroy_lle(struct llentry *lle) 1766 { 1767 1768 LLE_WUNLOCK(lle); 1769 LLE_LOCK_DESTROY(lle); 1770 kmem_intr_free(lle, sizeof(*lle)); 1771 } 1772 1773 static struct llentry * 1774 in_lltable_new(struct in_addr addr4, u_int flags) 1775 { 1776 struct in_llentry *lle; 1777 1778 lle = kmem_intr_zalloc(sizeof(*lle), KM_NOSLEEP); 1779 if (lle == NULL) /* NB: caller generates msg */ 1780 return NULL; 1781 1782 /* 1783 * For IPv4 this will trigger "arpresolve" to generate 1784 * an ARP request. 1785 */ 1786 lle->base.la_expire = time_uptime; /* mark expired */ 1787 lle->base.r_l3addr.addr4 = addr4; 1788 lle->base.lle_refcnt = 1; 1789 lle->base.lle_free = in_lltable_destroy_lle; 1790 LLE_LOCK_INIT(&lle->base); 1791 callout_init(&lle->base.la_timer, CALLOUT_MPSAFE); 1792 1793 return (&lle->base); 1794 } 1795 1796 #define IN_ARE_MASKED_ADDR_EQUAL(d, a, m) ( \ 1797 (((ntohl((d).s_addr) ^ (a)->sin_addr.s_addr) & (m)->sin_addr.s_addr)) == 0 ) 1798 1799 static int 1800 in_lltable_match_prefix(const struct sockaddr *prefix, 1801 const struct sockaddr *mask, u_int flags, struct llentry *lle) 1802 { 1803 const struct sockaddr_in *pfx = (const struct sockaddr_in *)prefix; 1804 const struct sockaddr_in *msk = (const struct sockaddr_in *)mask; 1805 1806 /* 1807 * (flags & LLE_STATIC) means deleting all entries 1808 * including static ARP entries. 1809 */ 1810 if (IN_ARE_MASKED_ADDR_EQUAL(lle->r_l3addr.addr4, pfx, msk) && 1811 ((flags & LLE_STATIC) || !(lle->la_flags & LLE_STATIC))) 1812 return (1); 1813 1814 return (0); 1815 } 1816 1817 static void 1818 in_lltable_free_entry(struct lltable *llt, struct llentry *lle) 1819 { 1820 struct ifnet *ifp __diagused; 1821 size_t pkts_dropped; 1822 1823 LLE_WLOCK_ASSERT(lle); 1824 KASSERT(llt != NULL); 1825 1826 /* Unlink entry from table if not already */ 1827 if ((lle->la_flags & LLE_LINKED) != 0) { 1828 ifp = llt->llt_ifp; 1829 IF_AFDATA_WLOCK_ASSERT(ifp); 1830 lltable_unlink_entry(llt, lle); 1831 } 1832 1833 /* cancel timer */ 1834 if (callout_halt(&lle->lle_timer, &lle->lle_lock)) 1835 LLE_REMREF(lle); 1836 1837 /* Drop hold queue */ 1838 pkts_dropped = llentry_free(lle); 1839 arp_stat_add(ARP_STAT_DFRDROPPED, (uint64_t)pkts_dropped); 1840 } 1841 1842 static int 1843 in_lltable_rtcheck(struct ifnet *ifp, u_int flags, const struct sockaddr *l3addr) 1844 { 1845 struct rtentry *rt; 1846 int error = EINVAL; 1847 1848 KASSERTMSG(l3addr->sa_family == AF_INET, 1849 "sin_family %d", l3addr->sa_family); 1850 1851 rt = rtalloc1(l3addr, 0); 1852 if (rt == NULL) 1853 return error; 1854 1855 /* 1856 * If the gateway for an existing host route matches the target L3 1857 * address, which is a special route inserted by some implementation 1858 * such as MANET, and the interface is of the correct type, then 1859 * allow for ARP to proceed. 1860 */ 1861 if (rt->rt_flags & RTF_GATEWAY) { 1862 if (!(rt->rt_flags & RTF_HOST) || !rt->rt_ifp || 1863 rt->rt_ifp->if_type != IFT_ETHER || 1864 #ifdef __FreeBSD__ 1865 (rt->rt_ifp->if_flags & (IFF_NOARP | IFF_STATICARP)) != 0 || 1866 #else 1867 (rt->rt_ifp->if_flags & IFF_NOARP) != 0 || 1868 #endif 1869 memcmp(rt->rt_gateway->sa_data, l3addr->sa_data, 1870 sizeof(in_addr_t)) != 0) { 1871 goto error; 1872 } 1873 } 1874 1875 /* 1876 * Make sure that at least the destination address is covered 1877 * by the route. This is for handling the case where 2 or more 1878 * interfaces have the same prefix. An incoming packet arrives 1879 * on one interface and the corresponding outgoing packet leaves 1880 * another interface. 1881 */ 1882 if (!(rt->rt_flags & RTF_HOST) && rt->rt_ifp != ifp) { 1883 const char *sa, *mask, *addr, *lim; 1884 int len; 1885 1886 mask = (const char *)rt_mask(rt); 1887 /* 1888 * Just being extra cautious to avoid some custom 1889 * code getting into trouble. 1890 */ 1891 if (mask == NULL) 1892 goto error; 1893 1894 sa = (const char *)rt_getkey(rt); 1895 addr = (const char *)l3addr; 1896 len = ((const struct sockaddr_in *)l3addr)->sin_len; 1897 lim = addr + len; 1898 1899 for ( ; addr < lim; sa++, mask++, addr++) { 1900 if ((*sa ^ *addr) & *mask) { 1901 #ifdef DIAGNOSTIC 1902 log(LOG_INFO, "IPv4 address: \"%s\" is not on the network\n", 1903 inet_ntoa(((const struct sockaddr_in *)l3addr)->sin_addr)); 1904 #endif 1905 goto error; 1906 } 1907 } 1908 } 1909 1910 error = 0; 1911 error: 1912 rtfree(rt); 1913 return error; 1914 } 1915 1916 static inline uint32_t 1917 in_lltable_hash_dst(const struct in_addr dst, uint32_t hsize) 1918 { 1919 1920 return (IN_LLTBL_HASH(dst.s_addr, hsize)); 1921 } 1922 1923 static uint32_t 1924 in_lltable_hash(const struct llentry *lle, uint32_t hsize) 1925 { 1926 1927 return (in_lltable_hash_dst(lle->r_l3addr.addr4, hsize)); 1928 } 1929 1930 static void 1931 in_lltable_fill_sa_entry(const struct llentry *lle, struct sockaddr *sa) 1932 { 1933 struct sockaddr_in *sin; 1934 1935 sin = (struct sockaddr_in *)sa; 1936 memset(sin, 0, sizeof(*sin)); 1937 sin->sin_family = AF_INET; 1938 sin->sin_len = sizeof(*sin); 1939 sin->sin_addr = lle->r_l3addr.addr4; 1940 } 1941 1942 static inline struct llentry * 1943 in_lltable_find_dst(struct lltable *llt, struct in_addr dst) 1944 { 1945 struct llentry *lle; 1946 struct llentries *lleh; 1947 u_int hashidx; 1948 1949 hashidx = in_lltable_hash_dst(dst, llt->llt_hsize); 1950 lleh = &llt->lle_head[hashidx]; 1951 LIST_FOREACH(lle, lleh, lle_next) { 1952 if (lle->la_flags & LLE_DELETED) 1953 continue; 1954 if (lle->r_l3addr.addr4.s_addr == dst.s_addr) 1955 break; 1956 } 1957 1958 return (lle); 1959 } 1960 1961 static int 1962 in_lltable_delete(struct lltable *llt, u_int flags, 1963 const struct sockaddr *l3addr) 1964 { 1965 const struct sockaddr_in *sin = (const struct sockaddr_in *)l3addr; 1966 struct ifnet *ifp __diagused = llt->llt_ifp; 1967 struct llentry *lle; 1968 1969 IF_AFDATA_WLOCK_ASSERT(ifp); 1970 KASSERTMSG(l3addr->sa_family == AF_INET, 1971 "sin_family %d", l3addr->sa_family); 1972 1973 lle = in_lltable_find_dst(llt, sin->sin_addr); 1974 if (lle == NULL) { 1975 #ifdef DIAGNOSTIC 1976 log(LOG_INFO, "interface address is missing from cache = %p in delete\n", lle); 1977 #endif 1978 return (ENOENT); 1979 } 1980 1981 LLE_WLOCK(lle); 1982 lle->la_flags |= LLE_DELETED; 1983 #ifdef DIAGNOSTIC 1984 log(LOG_INFO, "ifaddr cache = %p is deleted\n", lle); 1985 #endif 1986 if ((lle->la_flags & (LLE_STATIC | LLE_IFADDR)) == LLE_STATIC) 1987 llentry_free(lle); 1988 else 1989 LLE_WUNLOCK(lle); 1990 1991 return (0); 1992 } 1993 1994 static struct llentry * 1995 in_lltable_create(struct lltable *llt, u_int flags, const struct sockaddr *l3addr) 1996 { 1997 const struct sockaddr_in *sin = (const struct sockaddr_in *)l3addr; 1998 struct ifnet *ifp = llt->llt_ifp; 1999 struct llentry *lle; 2000 2001 IF_AFDATA_WLOCK_ASSERT(ifp); 2002 KASSERTMSG(l3addr->sa_family == AF_INET, 2003 "sin_family %d", l3addr->sa_family); 2004 2005 lle = in_lltable_find_dst(llt, sin->sin_addr); 2006 2007 if (lle != NULL) { 2008 LLE_WLOCK(lle); 2009 return (lle); 2010 } 2011 2012 /* no existing record, we need to create new one */ 2013 2014 /* 2015 * A route that covers the given address must have 2016 * been installed 1st because we are doing a resolution, 2017 * verify this. 2018 */ 2019 if (!(flags & LLE_IFADDR) && 2020 in_lltable_rtcheck(ifp, flags, l3addr) != 0) 2021 return (NULL); 2022 2023 lle = in_lltable_new(sin->sin_addr, flags); 2024 if (lle == NULL) { 2025 log(LOG_INFO, "lla_lookup: new lle malloc failed\n"); 2026 return (NULL); 2027 } 2028 lle->la_flags = flags; 2029 if ((flags & LLE_IFADDR) == LLE_IFADDR) { 2030 memcpy(&lle->ll_addr, CLLADDR(ifp->if_sadl), ifp->if_addrlen); 2031 lle->la_flags |= (LLE_VALID | LLE_STATIC); 2032 } 2033 2034 lltable_link_entry(llt, lle); 2035 LLE_WLOCK(lle); 2036 2037 return (lle); 2038 } 2039 2040 /* 2041 * Return NULL if not found or marked for deletion. 2042 * If found return lle read locked. 2043 */ 2044 static struct llentry * 2045 in_lltable_lookup(struct lltable *llt, u_int flags, const struct sockaddr *l3addr) 2046 { 2047 const struct sockaddr_in *sin = (const struct sockaddr_in *)l3addr; 2048 struct llentry *lle; 2049 2050 IF_AFDATA_LOCK_ASSERT(llt->llt_ifp); 2051 KASSERTMSG(l3addr->sa_family == AF_INET, 2052 "sin_family %d", l3addr->sa_family); 2053 2054 lle = in_lltable_find_dst(llt, sin->sin_addr); 2055 2056 if (lle == NULL) 2057 return NULL; 2058 2059 if (flags & LLE_EXCLUSIVE) 2060 LLE_WLOCK(lle); 2061 else 2062 LLE_RLOCK(lle); 2063 2064 return lle; 2065 } 2066 2067 static int 2068 in_lltable_dump_entry(struct lltable *llt, struct llentry *lle, 2069 struct rt_walkarg *w) 2070 { 2071 struct sockaddr_in sin; 2072 2073 LLTABLE_LOCK_ASSERT(); 2074 2075 /* skip deleted entries */ 2076 if (lle->la_flags & LLE_DELETED) 2077 return 0; 2078 2079 sockaddr_in_init(&sin, &lle->r_l3addr.addr4, 0); 2080 2081 return lltable_dump_entry(llt, lle, w, sintosa(&sin)); 2082 } 2083 2084 #endif /* NARP > 0 */ 2085 2086 static int 2087 in_multicast_sysctl(SYSCTLFN_ARGS) 2088 { 2089 struct ifnet *ifp; 2090 struct ifaddr *ifa; 2091 struct in_ifaddr *ifa4; 2092 struct in_multi *inm; 2093 uint32_t tmp; 2094 int error; 2095 size_t written; 2096 struct psref psref; 2097 int bound; 2098 2099 if (namelen != 1) 2100 return EINVAL; 2101 2102 bound = curlwp_bind(); 2103 ifp = if_get_byindex(name[0], &psref); 2104 if (ifp == NULL) { 2105 curlwp_bindx(bound); 2106 return ENODEV; 2107 } 2108 2109 if (oldp == NULL) { 2110 *oldlenp = 0; 2111 IFADDR_FOREACH(ifa, ifp) { 2112 if (ifa->ifa_addr->sa_family != AF_INET) 2113 continue; 2114 ifa4 = (void *)ifa; 2115 LIST_FOREACH(inm, &ifa4->ia_multiaddrs, inm_list) { 2116 *oldlenp += 2 * sizeof(struct in_addr) + 2117 sizeof(uint32_t); 2118 } 2119 } 2120 if_put(ifp, &psref); 2121 curlwp_bindx(bound); 2122 return 0; 2123 } 2124 2125 error = 0; 2126 written = 0; 2127 IFADDR_FOREACH(ifa, ifp) { 2128 if (ifa->ifa_addr->sa_family != AF_INET) 2129 continue; 2130 ifa4 = (void *)ifa; 2131 LIST_FOREACH(inm, &ifa4->ia_multiaddrs, inm_list) { 2132 if (written + 2 * sizeof(struct in_addr) + 2133 sizeof(uint32_t) > *oldlenp) 2134 goto done; 2135 error = sysctl_copyout(l, &ifa4->ia_addr.sin_addr, 2136 oldp, sizeof(struct in_addr)); 2137 if (error) 2138 goto done; 2139 oldp = (char *)oldp + sizeof(struct in_addr); 2140 written += sizeof(struct in_addr); 2141 error = sysctl_copyout(l, &inm->inm_addr, 2142 oldp, sizeof(struct in_addr)); 2143 if (error) 2144 goto done; 2145 oldp = (char *)oldp + sizeof(struct in_addr); 2146 written += sizeof(struct in_addr); 2147 tmp = inm->inm_refcount; 2148 error = sysctl_copyout(l, &tmp, oldp, sizeof(tmp)); 2149 if (error) 2150 goto done; 2151 oldp = (char *)oldp + sizeof(tmp); 2152 written += sizeof(tmp); 2153 } 2154 } 2155 done: 2156 if_put(ifp, &psref); 2157 curlwp_bindx(bound); 2158 *oldlenp = written; 2159 return error; 2160 } 2161 2162 static void 2163 in_sysctl_init(struct sysctllog **clog) 2164 { 2165 sysctl_createv(clog, 0, NULL, NULL, 2166 CTLFLAG_PERMANENT, 2167 CTLTYPE_NODE, "inet", 2168 SYSCTL_DESCR("PF_INET related settings"), 2169 NULL, 0, NULL, 0, 2170 CTL_NET, PF_INET, CTL_EOL); 2171 sysctl_createv(clog, 0, NULL, NULL, 2172 CTLFLAG_PERMANENT, 2173 CTLTYPE_NODE, "multicast", 2174 SYSCTL_DESCR("Multicast information"), 2175 in_multicast_sysctl, 0, NULL, 0, 2176 CTL_NET, PF_INET, CTL_CREATE, CTL_EOL); 2177 sysctl_createv(clog, 0, NULL, NULL, 2178 CTLFLAG_PERMANENT, 2179 CTLTYPE_NODE, "ip", 2180 SYSCTL_DESCR("IPv4 related settings"), 2181 NULL, 0, NULL, 0, 2182 CTL_NET, PF_INET, IPPROTO_IP, CTL_EOL); 2183 2184 sysctl_createv(clog, 0, NULL, NULL, 2185 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 2186 CTLTYPE_INT, "subnetsarelocal", 2187 SYSCTL_DESCR("Whether logical subnets are considered " 2188 "local"), 2189 NULL, 0, &subnetsarelocal, 0, 2190 CTL_NET, PF_INET, IPPROTO_IP, 2191 IPCTL_SUBNETSARELOCAL, CTL_EOL); 2192 sysctl_createv(clog, 0, NULL, NULL, 2193 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 2194 CTLTYPE_INT, "hostzerobroadcast", 2195 SYSCTL_DESCR("All zeroes address is broadcast address"), 2196 NULL, 0, &hostzeroisbroadcast, 0, 2197 CTL_NET, PF_INET, IPPROTO_IP, 2198 IPCTL_HOSTZEROBROADCAST, CTL_EOL); 2199 } 2200 2201 #if NARP > 0 2202 2203 static struct lltable * 2204 in_lltattach(struct ifnet *ifp) 2205 { 2206 struct lltable *llt; 2207 2208 llt = lltable_allocate_htbl(IN_LLTBL_DEFAULT_HSIZE); 2209 llt->llt_af = AF_INET; 2210 llt->llt_ifp = ifp; 2211 2212 llt->llt_lookup = in_lltable_lookup; 2213 llt->llt_create = in_lltable_create; 2214 llt->llt_delete = in_lltable_delete; 2215 llt->llt_dump_entry = in_lltable_dump_entry; 2216 llt->llt_hash = in_lltable_hash; 2217 llt->llt_fill_sa_entry = in_lltable_fill_sa_entry; 2218 llt->llt_free_entry = in_lltable_free_entry; 2219 llt->llt_match_prefix = in_lltable_match_prefix; 2220 lltable_link(llt); 2221 2222 return (llt); 2223 } 2224 2225 #endif /* NARP > 0 */ 2226 2227 void * 2228 in_domifattach(struct ifnet *ifp) 2229 { 2230 struct in_ifinfo *ii; 2231 2232 ii = kmem_zalloc(sizeof(struct in_ifinfo), KM_SLEEP); 2233 KASSERT(ii != NULL); 2234 2235 #if NARP > 0 2236 ii->ii_llt = in_lltattach(ifp); 2237 #endif 2238 2239 #ifdef IPSELSRC 2240 ii->ii_selsrc = in_selsrc_domifattach(ifp); 2241 KASSERT(ii->ii_selsrc != NULL); 2242 #endif 2243 2244 return ii; 2245 } 2246 2247 void 2248 in_domifdetach(struct ifnet *ifp, void *aux) 2249 { 2250 struct in_ifinfo *ii = aux; 2251 2252 #ifdef IPSELSRC 2253 in_selsrc_domifdetach(ifp, ii->ii_selsrc); 2254 #endif 2255 #if NARP > 0 2256 lltable_free(ii->ii_llt); 2257 #endif 2258 kmem_free(ii, sizeof(struct in_ifinfo)); 2259 } 2260
Indexes created Sun Sep 21 20:09:37 GMT 2025