in_pcb.c revision 1.27.4.2
1 /* $NetBSD: in_pcb.c,v 1.27.4.2 1996/12/11 04:01:01 mycroft Exp $ */ 2 3 /* 4 * Copyright (c) 1982, 1986, 1991, 1993 5 * The Regents of the University of California. 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. All advertising materials mentioning features or use of this software 16 * must display the following acknowledgement: 17 * This product includes software developed by the University of 18 * California, Berkeley and its contributors. 19 * 4. Neither the name of the University nor the names of its contributors 20 * may be used to endorse or promote products derived from this software 21 * without specific prior written permission. 22 * 23 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 24 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 25 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 26 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 27 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 28 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 29 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 30 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 31 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 32 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 33 * SUCH DAMAGE. 34 * 35 * @(#)in_pcb.c 8.2 (Berkeley) 1/4/94 36 */ 37 38 #include <sys/param.h> 39 #include <sys/systm.h> 40 #include <sys/malloc.h> 41 #include <sys/mbuf.h> 42 #include <sys/protosw.h> 43 #include <sys/socket.h> 44 #include <sys/socketvar.h> 45 #include <sys/ioctl.h> 46 #include <sys/errno.h> 47 #include <sys/time.h> 48 #include <sys/proc.h> 49 50 #include <net/if.h> 51 #include <net/route.h> 52 53 #include <netinet/in.h> 54 #include <netinet/in_systm.h> 55 #include <netinet/ip.h> 56 #include <netinet/in_pcb.h> 57 #include <netinet/in_var.h> 58 #include <netinet/ip_var.h> 59 60 struct in_addr zeroin_addr; 61 62 #define INPCBHASH(table, faddr, fport, laddr, lport) \ 63 &(table)->inpt_hashtbl[(ntohl((faddr)->s_addr) + ntohs((fport)) + ntohs((lport))) & (table->inpt_hash)] 64 65 void 66 in_pcbinit(table, hashsize) 67 struct inpcbtable *table; 68 int hashsize; 69 { 70 71 CIRCLEQ_INIT(&table->inpt_queue); 72 table->inpt_hashtbl = hashinit(hashsize, M_PCB, &table->inpt_hash); 73 table->inpt_lastport = IPPORT_RESERVED; 74 } 75 76 int 77 in_pcballoc(so, v) 78 struct socket *so; 79 void *v; 80 { 81 struct inpcbtable *table = v; 82 register struct inpcb *inp; 83 int s; 84 85 MALLOC(inp, struct inpcb *, sizeof(*inp), M_PCB, M_WAITOK); 86 if (inp == NULL) 87 return (ENOBUFS); 88 bzero((caddr_t)inp, sizeof(*inp)); 89 inp->inp_table = table; 90 inp->inp_socket = so; 91 s = splnet(); 92 CIRCLEQ_INSERT_HEAD(&table->inpt_queue, inp, inp_queue); 93 LIST_INSERT_HEAD(INPCBHASH(table, &inp->inp_faddr, inp->inp_fport, 94 &inp->inp_laddr, inp->inp_lport), inp, inp_hash); 95 splx(s); 96 so->so_pcb = inp; 97 return (0); 98 } 99 100 int 101 in_pcbbind(v, nam, p) 102 void *v; 103 struct mbuf *nam; 104 struct proc *p; 105 { 106 register struct inpcb *inp = v; 107 register struct socket *so = inp->inp_socket; 108 register struct inpcbtable *table = inp->inp_table; 109 register struct sockaddr_in *sin; 110 u_int16_t lport = 0; 111 int wild = 0, reuseport = (so->so_options & SO_REUSEPORT); 112 int error; 113 114 if (in_ifaddr.tqh_first == 0) 115 return (EADDRNOTAVAIL); 116 if (inp->inp_lport || inp->inp_laddr.s_addr != INADDR_ANY) 117 return (EINVAL); 118 if ((so->so_options & (SO_REUSEADDR|SO_REUSEPORT)) == 0 && 119 ((so->so_proto->pr_flags & PR_CONNREQUIRED) == 0 || 120 (so->so_options & SO_ACCEPTCONN) == 0)) 121 wild = INPLOOKUP_WILDCARD; 122 if (nam == 0) 123 goto noname; 124 sin = mtod(nam, struct sockaddr_in *); 125 if (nam->m_len != sizeof (*sin)) 126 return (EINVAL); 127 #ifdef notdef 128 /* 129 * We should check the family, but old programs 130 * incorrectly fail to initialize it. 131 */ 132 if (sin->sin_family != AF_INET) 133 return (EAFNOSUPPORT); 134 #endif 135 lport = sin->sin_port; 136 if (IN_MULTICAST(sin->sin_addr.s_addr)) { 137 /* 138 * Treat SO_REUSEADDR as SO_REUSEPORT for multicast; 139 * allow complete duplication of binding if 140 * SO_REUSEPORT is set, or if SO_REUSEADDR is set 141 * and a multicast address is bound on both 142 * new and duplicated sockets. 143 */ 144 if (so->so_options & SO_REUSEADDR) 145 reuseport = SO_REUSEADDR|SO_REUSEPORT; 146 } else if (sin->sin_addr.s_addr != INADDR_ANY) { 147 sin->sin_port = 0; /* yech... */ 148 if (ifa_ifwithaddr(sintosa(sin)) == 0) 149 return (EADDRNOTAVAIL); 150 } 151 if (lport) { 152 struct inpcb *t; 153 154 /* GROSS */ 155 if (ntohs(lport) < IPPORT_RESERVED && 156 (p == 0 || (error = suser(p->p_ucred, &p->p_acflag)))) 157 return (EACCES); 158 t = in_pcblookup(table, zeroin_addr, 0, 159 sin->sin_addr, lport, wild); 160 if (t && (reuseport & t->inp_socket->so_options) == 0) 161 return (EADDRINUSE); 162 } 163 inp->inp_laddr = sin->sin_addr; 164 noname: 165 if (lport == 0) { 166 for (lport = table->inpt_lastport + 1; 167 lport < IPPORT_USERRESERVED; lport++) 168 if (!in_pcblookup(table, zeroin_addr, 0, inp->inp_laddr, 169 htons(lport), wild)) 170 goto found; 171 for (lport = IPPORT_RESERVED; 172 lport <= table->inpt_lastport; lport++) 173 if (!in_pcblookup(table, zeroin_addr, 0, inp->inp_laddr, 174 htons(lport), wild)) 175 goto found; 176 return (EAGAIN); 177 found: 178 table->inpt_lastport = lport; 179 lport = htons(lport); 180 } 181 inp->inp_lport = lport; 182 in_pcbrehash(inp); 183 return (0); 184 } 185 186 /* 187 * Connect from a socket to a specified address. 188 * Both address and port must be specified in argument sin. 189 * If don't have a local address for this socket yet, 190 * then pick one. 191 */ 192 int 193 in_pcbconnect(v, nam) 194 register void *v; 195 struct mbuf *nam; 196 { 197 register struct inpcb *inp = v; 198 struct in_ifaddr *ia; 199 struct sockaddr_in *ifaddr = NULL; 200 register struct sockaddr_in *sin = mtod(nam, struct sockaddr_in *); 201 202 if (nam->m_len != sizeof (*sin)) 203 return (EINVAL); 204 if (sin->sin_family != AF_INET) 205 return (EAFNOSUPPORT); 206 if (sin->sin_port == 0) 207 return (EADDRNOTAVAIL); 208 if (in_ifaddr.tqh_first != 0) { 209 /* 210 * If the destination address is INADDR_ANY, 211 * use the primary local address. 212 * If the supplied address is INADDR_BROADCAST, 213 * and the primary interface supports broadcast, 214 * choose the broadcast address for that interface. 215 */ 216 if (sin->sin_addr.s_addr == INADDR_ANY) 217 sin->sin_addr = in_ifaddr.tqh_first->ia_addr.sin_addr; 218 else if (sin->sin_addr.s_addr == INADDR_BROADCAST && 219 (in_ifaddr.tqh_first->ia_ifp->if_flags & IFF_BROADCAST)) 220 sin->sin_addr = in_ifaddr.tqh_first->ia_broadaddr.sin_addr; 221 } 222 if (inp->inp_laddr.s_addr == INADDR_ANY) { 223 register struct route *ro; 224 225 ia = (struct in_ifaddr *)0; 226 /* 227 * If route is known or can be allocated now, 228 * our src addr is taken from the i/f, else punt. 229 */ 230 ro = &inp->inp_route; 231 if (ro->ro_rt && 232 (satosin(&ro->ro_dst)->sin_addr.s_addr != 233 sin->sin_addr.s_addr || 234 inp->inp_socket->so_options & SO_DONTROUTE)) { 235 RTFREE(ro->ro_rt); 236 ro->ro_rt = (struct rtentry *)0; 237 } 238 if ((inp->inp_socket->so_options & SO_DONTROUTE) == 0 && /*XXX*/ 239 (ro->ro_rt == (struct rtentry *)0 || 240 ro->ro_rt->rt_ifp == (struct ifnet *)0)) { 241 /* No route yet, so try to acquire one */ 242 ro->ro_dst.sa_family = AF_INET; 243 ro->ro_dst.sa_len = sizeof(struct sockaddr_in); 244 satosin(&ro->ro_dst)->sin_addr = sin->sin_addr; 245 rtalloc(ro); 246 } 247 /* 248 * If we found a route, use the address 249 * corresponding to the outgoing interface 250 * unless it is the loopback (in case a route 251 * to our address on another net goes to loopback). 252 */ 253 if (ro->ro_rt && !(ro->ro_rt->rt_ifp->if_flags & IFF_LOOPBACK)) 254 ia = ifatoia(ro->ro_rt->rt_ifa); 255 if (ia == 0) { 256 u_int16_t fport = sin->sin_port; 257 258 sin->sin_port = 0; 259 ia = ifatoia(ifa_ifwithladdr(sintosa(sin))); 260 sin->sin_port = fport; 261 if (ia == 0) 262 ia = in_ifaddr.tqh_first; 263 if (ia == 0) 264 return (EADDRNOTAVAIL); 265 } 266 /* 267 * If the destination address is multicast and an outgoing 268 * interface has been set as a multicast option, use the 269 * address of that interface as our source address. 270 */ 271 if (IN_MULTICAST(sin->sin_addr.s_addr) && 272 inp->inp_moptions != NULL) { 273 struct ip_moptions *imo; 274 struct ifnet *ifp; 275 276 imo = inp->inp_moptions; 277 if (imo->imo_multicast_ifp != NULL) { 278 ifp = imo->imo_multicast_ifp; 279 for (ia = in_ifaddr.tqh_first; ia != 0; 280 ia = ia->ia_list.tqe_next) 281 if (ia->ia_ifp == ifp) 282 break; 283 if (ia == 0) 284 return (EADDRNOTAVAIL); 285 } 286 } 287 ifaddr = satosin(&ia->ia_addr); 288 } 289 if (in_pcbhashlookup(inp->inp_table, sin->sin_addr, sin->sin_port, 290 inp->inp_laddr.s_addr ? inp->inp_laddr : ifaddr->sin_addr, 291 inp->inp_lport) != 0) 292 return (EADDRINUSE); 293 if (inp->inp_laddr.s_addr == INADDR_ANY) { 294 if (inp->inp_lport == 0) 295 (void)in_pcbbind(inp, (struct mbuf *)0, 296 (struct proc *)0); 297 inp->inp_laddr = ifaddr->sin_addr; 298 } 299 inp->inp_faddr = sin->sin_addr; 300 inp->inp_fport = sin->sin_port; 301 in_pcbrehash(inp); 302 return (0); 303 } 304 305 void 306 in_pcbdisconnect(v) 307 void *v; 308 { 309 struct inpcb *inp = v; 310 311 inp->inp_faddr.s_addr = INADDR_ANY; 312 inp->inp_fport = 0; 313 in_pcbrehash(inp); 314 if (inp->inp_socket->so_state & SS_NOFDREF) 315 in_pcbdetach(inp); 316 } 317 318 void 319 in_pcbdetach(v) 320 void *v; 321 { 322 struct inpcb *inp = v; 323 struct socket *so = inp->inp_socket; 324 int s; 325 326 so->so_pcb = 0; 327 sofree(so); 328 if (inp->inp_options) 329 (void)m_free(inp->inp_options); 330 if (inp->inp_route.ro_rt) 331 rtfree(inp->inp_route.ro_rt); 332 ip_freemoptions(inp->inp_moptions); 333 s = splnet(); 334 LIST_REMOVE(inp, inp_hash); 335 CIRCLEQ_REMOVE(&inp->inp_table->inpt_queue, inp, inp_queue); 336 splx(s); 337 FREE(inp, M_PCB); 338 } 339 340 void 341 in_setsockaddr(inp, nam) 342 register struct inpcb *inp; 343 struct mbuf *nam; 344 { 345 register struct sockaddr_in *sin; 346 347 nam->m_len = sizeof (*sin); 348 sin = mtod(nam, struct sockaddr_in *); 349 bzero((caddr_t)sin, sizeof (*sin)); 350 sin->sin_family = AF_INET; 351 sin->sin_len = sizeof(*sin); 352 sin->sin_port = inp->inp_lport; 353 sin->sin_addr = inp->inp_laddr; 354 } 355 356 void 357 in_setpeeraddr(inp, nam) 358 struct inpcb *inp; 359 struct mbuf *nam; 360 { 361 register struct sockaddr_in *sin; 362 363 nam->m_len = sizeof (*sin); 364 sin = mtod(nam, struct sockaddr_in *); 365 bzero((caddr_t)sin, sizeof (*sin)); 366 sin->sin_family = AF_INET; 367 sin->sin_len = sizeof(*sin); 368 sin->sin_port = inp->inp_fport; 369 sin->sin_addr = inp->inp_faddr; 370 } 371 372 /* 373 * Pass some notification to all connections of a protocol 374 * associated with address dst. The local address and/or port numbers 375 * may be specified to limit the search. The "usual action" will be 376 * taken, depending on the ctlinput cmd. The caller must filter any 377 * cmds that are uninteresting (e.g., no error in the map). 378 * Call the protocol specific routine (if any) to report 379 * any errors for each matching socket. 380 * 381 * Must be called at splsoftnet. 382 */ 383 void 384 in_pcbnotify(table, dst, fport_arg, laddr, lport_arg, errno, notify) 385 struct inpcbtable *table; 386 struct sockaddr *dst; 387 u_int fport_arg, lport_arg; 388 struct in_addr laddr; 389 int errno; 390 void (*notify) __P((struct inpcb *, int)); 391 { 392 register struct inpcb *inp, *oinp; 393 struct in_addr faddr; 394 u_int16_t fport = fport_arg, lport = lport_arg; 395 396 if (dst->sa_family != AF_INET) 397 return; 398 faddr = satosin(dst)->sin_addr; 399 if (faddr.s_addr == INADDR_ANY) 400 return; 401 402 for (inp = table->inpt_queue.cqh_first; 403 inp != (struct inpcb *)&table->inpt_queue;) { 404 if (inp->inp_faddr.s_addr != faddr.s_addr || 405 inp->inp_socket == 0 || 406 inp->inp_fport != fport || 407 inp->inp_lport != lport || 408 inp->inp_laddr.s_addr != laddr.s_addr) { 409 inp = inp->inp_queue.cqe_next; 410 continue; 411 } 412 oinp = inp; 413 inp = inp->inp_queue.cqe_next; 414 if (notify) 415 (*notify)(oinp, errno); 416 } 417 } 418 419 void 420 in_pcbnotifyall(table, dst, errno, notify) 421 struct inpcbtable *table; 422 struct sockaddr *dst; 423 int errno; 424 void (*notify) __P((struct inpcb *, int)); 425 { 426 register struct inpcb *inp, *oinp; 427 struct in_addr faddr; 428 429 if (dst->sa_family != AF_INET) 430 return; 431 faddr = satosin(dst)->sin_addr; 432 if (faddr.s_addr == INADDR_ANY) 433 return; 434 435 for (inp = table->inpt_queue.cqh_first; 436 inp != (struct inpcb *)&table->inpt_queue;) { 437 if (inp->inp_faddr.s_addr != faddr.s_addr || 438 inp->inp_socket == 0) { 439 inp = inp->inp_queue.cqe_next; 440 continue; 441 } 442 oinp = inp; 443 inp = inp->inp_queue.cqe_next; 444 if (notify) 445 (*notify)(oinp, errno); 446 } 447 } 448 449 /* 450 * Check for alternatives when higher level complains 451 * about service problems. For now, invalidate cached 452 * routing information. If the route was created dynamically 453 * (by a redirect), time to try a default gateway again. 454 */ 455 void 456 in_losing(inp) 457 struct inpcb *inp; 458 { 459 register struct rtentry *rt; 460 struct rt_addrinfo info; 461 462 if ((rt = inp->inp_route.ro_rt)) { 463 inp->inp_route.ro_rt = 0; 464 bzero((caddr_t)&info, sizeof(info)); 465 info.rti_info[RTAX_DST] = &inp->inp_route.ro_dst; 466 info.rti_info[RTAX_GATEWAY] = rt->rt_gateway; 467 info.rti_info[RTAX_NETMASK] = rt_mask(rt); 468 rt_missmsg(RTM_LOSING, &info, rt->rt_flags, 0); 469 if (rt->rt_flags & RTF_DYNAMIC) 470 (void) rtrequest(RTM_DELETE, rt_key(rt), 471 rt->rt_gateway, rt_mask(rt), rt->rt_flags, 472 (struct rtentry **)0); 473 else 474 /* 475 * A new route can be allocated 476 * the next time output is attempted. 477 */ 478 rtfree(rt); 479 } 480 } 481 482 /* 483 * After a routing change, flush old routing 484 * and allocate a (hopefully) better one. 485 */ 486 void 487 in_rtchange(inp, errno) 488 register struct inpcb *inp; 489 int errno; 490 { 491 if (inp->inp_route.ro_rt) { 492 rtfree(inp->inp_route.ro_rt); 493 inp->inp_route.ro_rt = 0; 494 /* 495 * A new route can be allocated the next time 496 * output is attempted. 497 */ 498 } 499 } 500 501 struct inpcb * 502 in_pcblookup(table, faddr, fport_arg, laddr, lport_arg, flags) 503 struct inpcbtable *table; 504 struct in_addr faddr, laddr; 505 u_int fport_arg, lport_arg; 506 int flags; 507 { 508 register struct inpcb *inp, *match = 0; 509 int matchwild = 3, wildcard; 510 u_int16_t fport = fport_arg, lport = lport_arg; 511 512 for (inp = table->inpt_queue.cqh_first; 513 inp != (struct inpcb *)&table->inpt_queue; 514 inp = inp->inp_queue.cqe_next) { 515 if (inp->inp_lport != lport) 516 continue; 517 wildcard = 0; 518 if (inp->inp_faddr.s_addr != INADDR_ANY) { 519 if (faddr.s_addr == INADDR_ANY) 520 wildcard++; 521 else if (inp->inp_faddr.s_addr != faddr.s_addr || 522 inp->inp_fport != fport) 523 continue; 524 } else { 525 if (faddr.s_addr != INADDR_ANY) 526 wildcard++; 527 } 528 if (inp->inp_laddr.s_addr != INADDR_ANY) { 529 if (laddr.s_addr == INADDR_ANY) 530 wildcard++; 531 else if (inp->inp_laddr.s_addr != laddr.s_addr) 532 continue; 533 } else { 534 if (laddr.s_addr != INADDR_ANY) 535 wildcard++; 536 } 537 if (wildcard && (flags & INPLOOKUP_WILDCARD) == 0) 538 continue; 539 if (wildcard < matchwild) { 540 match = inp; 541 matchwild = wildcard; 542 if (matchwild == 0) 543 break; 544 } 545 } 546 return (match); 547 } 548 549 void 550 in_pcbrehash(inp) 551 struct inpcb *inp; 552 { 553 struct inpcbtable *table = inp->inp_table; 554 int s; 555 556 s = splnet(); 557 LIST_REMOVE(inp, inp_hash); 558 LIST_INSERT_HEAD(INPCBHASH(table, &inp->inp_faddr, inp->inp_fport, 559 &inp->inp_laddr, inp->inp_lport), inp, inp_hash); 560 splx(s); 561 } 562 563 #ifdef DIAGNOSTIC 564 int in_pcbnotifymiss = 0; 565 #endif 566 567 struct inpcb * 568 in_pcbhashlookup(table, faddr, fport_arg, laddr, lport_arg) 569 struct inpcbtable *table; 570 struct in_addr faddr, laddr; 571 u_int fport_arg, lport_arg; 572 { 573 struct inpcbhead *head; 574 register struct inpcb *inp; 575 u_int16_t fport = fport_arg, lport = lport_arg; 576 577 head = INPCBHASH(table, &faddr, fport, &laddr, lport); 578 for (inp = head->lh_first; inp != NULL; inp = inp->inp_hash.le_next) { 579 if (inp->inp_faddr.s_addr == faddr.s_addr && 580 inp->inp_fport == fport && 581 inp->inp_lport == lport && 582 inp->inp_laddr.s_addr == laddr.s_addr) { 583 /* 584 * Move this PCB to the head of hash chain so that 585 * repeated accesses are quicker. This is analogous to 586 * the historic single-entry PCB cache. 587 */ 588 if (inp != head->lh_first) { 589 LIST_REMOVE(inp, inp_hash); 590 LIST_INSERT_HEAD(head, inp, inp_hash); 591 } 592 break; 593 } 594 } 595 #ifdef DIAGNOSTIC 596 if (inp == NULL && in_pcbnotifymiss) { 597 printf("in_pcbhashlookup: faddr=%08x fport=%d laddr=%08x lport=%d\n", 598 ntohl(faddr.s_addr), ntohs(fport), 599 ntohl(laddr.s_addr), ntohs(lport)); 600 } 601 #endif 602 return (inp); 603 } 604
Indexes created Mon Oct 20 16:09:52 GMT 2025