in_pcb.c revision 1.27.4.1
1 /* $NetBSD: in_pcb.c,v 1.27.4.1 1996/12/10 11:39:02 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) 102 register void *v; 103 struct mbuf *nam; 104 { 105 register struct inpcb *inp = v; 106 register struct socket *so = inp->inp_socket; 107 register struct inpcbtable *table = inp->inp_table; 108 register struct sockaddr_in *sin; 109 struct proc *p = curproc; /* XXX */ 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) { 123 sin = mtod(nam, struct sockaddr_in *); 124 if (nam->m_len != sizeof (*sin)) 125 return (EINVAL); 126 #ifdef notdef 127 /* 128 * We should check the family, but old programs 129 * incorrectly fail to initialize it. 130 */ 131 if (sin->sin_family != AF_INET) 132 return (EAFNOSUPPORT); 133 #endif 134 lport = sin->sin_port; 135 if (IN_MULTICAST(sin->sin_addr.s_addr)) { 136 /* 137 * Treat SO_REUSEADDR as SO_REUSEPORT for multicast; 138 * allow complete duplication of binding if 139 * SO_REUSEPORT is set, or if SO_REUSEADDR is set 140 * and a multicast address is bound on both 141 * new and duplicated sockets. 142 */ 143 if (so->so_options & SO_REUSEADDR) 144 reuseport = SO_REUSEADDR|SO_REUSEPORT; 145 } else if (sin->sin_addr.s_addr != INADDR_ANY) { 146 sin->sin_port = 0; /* yech... */ 147 if (ifa_ifwithaddr(sintosa(sin)) == 0) 148 return (EADDRNOTAVAIL); 149 } 150 if (lport) { 151 struct inpcb *t; 152 153 /* GROSS */ 154 if (ntohs(lport) < IPPORT_RESERVED && 155 (error = suser(p->p_ucred, &p->p_acflag))) 156 return (EACCES); 157 t = in_pcblookup(table, zeroin_addr, 0, 158 sin->sin_addr, lport, wild); 159 if (t && (reuseport & t->inp_socket->so_options) == 0) 160 return (EADDRINUSE); 161 } 162 inp->inp_laddr = sin->sin_addr; 163 } 164 if (lport == 0) { 165 for (lport = table->inpt_lastport + 1; 166 lport < IPPORT_USERRESERVED; lport++) 167 if (!in_pcblookup(table, zeroin_addr, 0, inp->inp_laddr, 168 htons(lport), wild)) 169 goto found; 170 for (lport = IPPORT_RESERVED; 171 lport <= table->inpt_lastport; lport++) 172 if (!in_pcblookup(table, zeroin_addr, 0, inp->inp_laddr, 173 htons(lport), wild)) 174 goto found; 175 return (EAGAIN); 176 found: 177 table->inpt_lastport = lport; 178 lport = htons(lport); 179 } 180 inp->inp_lport = lport; 181 in_pcbrehash(inp); 182 return (0); 183 } 184 185 /* 186 * Connect from a socket to a specified address. 187 * Both address and port must be specified in argument sin. 188 * If don't have a local address for this socket yet, 189 * then pick one. 190 */ 191 int 192 in_pcbconnect(v, nam) 193 register void *v; 194 struct mbuf *nam; 195 { 196 register struct inpcb *inp = v; 197 struct in_ifaddr *ia; 198 struct sockaddr_in *ifaddr = NULL; 199 register struct sockaddr_in *sin = mtod(nam, struct sockaddr_in *); 200 201 if (nam->m_len != sizeof (*sin)) 202 return (EINVAL); 203 if (sin->sin_family != AF_INET) 204 return (EAFNOSUPPORT); 205 if (sin->sin_port == 0) 206 return (EADDRNOTAVAIL); 207 if (in_ifaddr.tqh_first != 0) { 208 /* 209 * If the destination address is INADDR_ANY, 210 * use the primary local address. 211 * If the supplied address is INADDR_BROADCAST, 212 * and the primary interface supports broadcast, 213 * choose the broadcast address for that interface. 214 */ 215 if (sin->sin_addr.s_addr == INADDR_ANY) 216 sin->sin_addr = in_ifaddr.tqh_first->ia_addr.sin_addr; 217 else if (sin->sin_addr.s_addr == INADDR_BROADCAST && 218 (in_ifaddr.tqh_first->ia_ifp->if_flags & IFF_BROADCAST)) 219 sin->sin_addr = in_ifaddr.tqh_first->ia_broadaddr.sin_addr; 220 } 221 if (inp->inp_laddr.s_addr == INADDR_ANY) { 222 register struct route *ro; 223 224 ia = (struct in_ifaddr *)0; 225 /* 226 * If route is known or can be allocated now, 227 * our src addr is taken from the i/f, else punt. 228 */ 229 ro = &inp->inp_route; 230 if (ro->ro_rt && 231 (satosin(&ro->ro_dst)->sin_addr.s_addr != 232 sin->sin_addr.s_addr || 233 inp->inp_socket->so_options & SO_DONTROUTE)) { 234 RTFREE(ro->ro_rt); 235 ro->ro_rt = (struct rtentry *)0; 236 } 237 if ((inp->inp_socket->so_options & SO_DONTROUTE) == 0 && /*XXX*/ 238 (ro->ro_rt == (struct rtentry *)0 || 239 ro->ro_rt->rt_ifp == (struct ifnet *)0)) { 240 /* No route yet, so try to acquire one */ 241 ro->ro_dst.sa_family = AF_INET; 242 ro->ro_dst.sa_len = sizeof(struct sockaddr_in); 243 satosin(&ro->ro_dst)->sin_addr = sin->sin_addr; 244 rtalloc(ro); 245 } 246 /* 247 * If we found a route, use the address 248 * corresponding to the outgoing interface 249 * unless it is the loopback (in case a route 250 * to our address on another net goes to loopback). 251 */ 252 if (ro->ro_rt && !(ro->ro_rt->rt_ifp->if_flags & IFF_LOOPBACK)) 253 ia = ifatoia(ro->ro_rt->rt_ifa); 254 if (ia == 0) { 255 u_int16_t fport = sin->sin_port; 256 257 sin->sin_port = 0; 258 ia = ifatoia(ifa_ifwithladdr(sintosa(sin))); 259 sin->sin_port = fport; 260 if (ia == 0) 261 ia = in_ifaddr.tqh_first; 262 if (ia == 0) 263 return (EADDRNOTAVAIL); 264 } 265 /* 266 * If the destination address is multicast and an outgoing 267 * interface has been set as a multicast option, use the 268 * address of that interface as our source address. 269 */ 270 if (IN_MULTICAST(sin->sin_addr.s_addr) && 271 inp->inp_moptions != NULL) { 272 struct ip_moptions *imo; 273 struct ifnet *ifp; 274 275 imo = inp->inp_moptions; 276 if (imo->imo_multicast_ifp != NULL) { 277 ifp = imo->imo_multicast_ifp; 278 for (ia = in_ifaddr.tqh_first; ia != 0; 279 ia = ia->ia_list.tqe_next) 280 if (ia->ia_ifp == ifp) 281 break; 282 if (ia == 0) 283 return (EADDRNOTAVAIL); 284 } 285 } 286 ifaddr = satosin(&ia->ia_addr); 287 } 288 if (in_pcbhashlookup(inp->inp_table, sin->sin_addr, sin->sin_port, 289 inp->inp_laddr.s_addr ? inp->inp_laddr : ifaddr->sin_addr, 290 inp->inp_lport) != 0) 291 return (EADDRINUSE); 292 if (inp->inp_laddr.s_addr == INADDR_ANY) { 293 if (inp->inp_lport == 0) 294 (void)in_pcbbind(inp, (struct mbuf *)0); 295 inp->inp_laddr = ifaddr->sin_addr; 296 } 297 inp->inp_faddr = sin->sin_addr; 298 inp->inp_fport = sin->sin_port; 299 in_pcbrehash(inp); 300 return (0); 301 } 302 303 void 304 in_pcbdisconnect(v) 305 void *v; 306 { 307 struct inpcb *inp = v; 308 309 inp->inp_faddr.s_addr = INADDR_ANY; 310 inp->inp_fport = 0; 311 in_pcbrehash(inp); 312 if (inp->inp_socket->so_state & SS_NOFDREF) 313 in_pcbdetach(inp); 314 } 315 316 void 317 in_pcbdetach(v) 318 void *v; 319 { 320 struct inpcb *inp = v; 321 struct socket *so = inp->inp_socket; 322 int s; 323 324 so->so_pcb = 0; 325 sofree(so); 326 if (inp->inp_options) 327 (void)m_free(inp->inp_options); 328 if (inp->inp_route.ro_rt) 329 rtfree(inp->inp_route.ro_rt); 330 ip_freemoptions(inp->inp_moptions); 331 s = splnet(); 332 LIST_REMOVE(inp, inp_hash); 333 CIRCLEQ_REMOVE(&inp->inp_table->inpt_queue, inp, inp_queue); 334 splx(s); 335 FREE(inp, M_PCB); 336 } 337 338 void 339 in_setsockaddr(inp, nam) 340 register struct inpcb *inp; 341 struct mbuf *nam; 342 { 343 register struct sockaddr_in *sin; 344 345 nam->m_len = sizeof (*sin); 346 sin = mtod(nam, struct sockaddr_in *); 347 bzero((caddr_t)sin, sizeof (*sin)); 348 sin->sin_family = AF_INET; 349 sin->sin_len = sizeof(*sin); 350 sin->sin_port = inp->inp_lport; 351 sin->sin_addr = inp->inp_laddr; 352 } 353 354 void 355 in_setpeeraddr(inp, nam) 356 struct inpcb *inp; 357 struct mbuf *nam; 358 { 359 register struct sockaddr_in *sin; 360 361 nam->m_len = sizeof (*sin); 362 sin = mtod(nam, struct sockaddr_in *); 363 bzero((caddr_t)sin, sizeof (*sin)); 364 sin->sin_family = AF_INET; 365 sin->sin_len = sizeof(*sin); 366 sin->sin_port = inp->inp_fport; 367 sin->sin_addr = inp->inp_faddr; 368 } 369 370 /* 371 * Pass some notification to all connections of a protocol 372 * associated with address dst. The local address and/or port numbers 373 * may be specified to limit the search. The "usual action" will be 374 * taken, depending on the ctlinput cmd. The caller must filter any 375 * cmds that are uninteresting (e.g., no error in the map). 376 * Call the protocol specific routine (if any) to report 377 * any errors for each matching socket. 378 * 379 * Must be called at splsoftnet. 380 */ 381 void 382 in_pcbnotify(table, dst, fport_arg, laddr, lport_arg, errno, notify) 383 struct inpcbtable *table; 384 struct sockaddr *dst; 385 u_int fport_arg, lport_arg; 386 struct in_addr laddr; 387 int errno; 388 void (*notify) __P((struct inpcb *, int)); 389 { 390 register struct inpcb *inp, *oinp; 391 struct in_addr faddr; 392 u_int16_t fport = fport_arg, lport = lport_arg; 393 394 if (dst->sa_family != AF_INET) 395 return; 396 faddr = satosin(dst)->sin_addr; 397 if (faddr.s_addr == INADDR_ANY) 398 return; 399 400 for (inp = table->inpt_queue.cqh_first; 401 inp != (struct inpcb *)&table->inpt_queue;) { 402 if (inp->inp_faddr.s_addr != faddr.s_addr || 403 inp->inp_socket == 0 || 404 inp->inp_fport != fport || 405 inp->inp_lport != lport || 406 inp->inp_laddr.s_addr != laddr.s_addr) { 407 inp = inp->inp_queue.cqe_next; 408 continue; 409 } 410 oinp = inp; 411 inp = inp->inp_queue.cqe_next; 412 if (notify) 413 (*notify)(oinp, errno); 414 } 415 } 416 417 void 418 in_pcbnotifyall(table, dst, errno, notify) 419 struct inpcbtable *table; 420 struct sockaddr *dst; 421 int errno; 422 void (*notify) __P((struct inpcb *, int)); 423 { 424 register struct inpcb *inp, *oinp; 425 struct in_addr faddr; 426 427 if (dst->sa_family != AF_INET) 428 return; 429 faddr = satosin(dst)->sin_addr; 430 if (faddr.s_addr == INADDR_ANY) 431 return; 432 433 for (inp = table->inpt_queue.cqh_first; 434 inp != (struct inpcb *)&table->inpt_queue;) { 435 if (inp->inp_faddr.s_addr != faddr.s_addr || 436 inp->inp_socket == 0) { 437 inp = inp->inp_queue.cqe_next; 438 continue; 439 } 440 oinp = inp; 441 inp = inp->inp_queue.cqe_next; 442 if (notify) 443 (*notify)(oinp, errno); 444 } 445 } 446 447 /* 448 * Check for alternatives when higher level complains 449 * about service problems. For now, invalidate cached 450 * routing information. If the route was created dynamically 451 * (by a redirect), time to try a default gateway again. 452 */ 453 void 454 in_losing(inp) 455 struct inpcb *inp; 456 { 457 register struct rtentry *rt; 458 struct rt_addrinfo info; 459 460 if ((rt = inp->inp_route.ro_rt)) { 461 inp->inp_route.ro_rt = 0; 462 bzero((caddr_t)&info, sizeof(info)); 463 info.rti_info[RTAX_DST] = &inp->inp_route.ro_dst; 464 info.rti_info[RTAX_GATEWAY] = rt->rt_gateway; 465 info.rti_info[RTAX_NETMASK] = rt_mask(rt); 466 rt_missmsg(RTM_LOSING, &info, rt->rt_flags, 0); 467 if (rt->rt_flags & RTF_DYNAMIC) 468 (void) rtrequest(RTM_DELETE, rt_key(rt), 469 rt->rt_gateway, rt_mask(rt), rt->rt_flags, 470 (struct rtentry **)0); 471 else 472 /* 473 * A new route can be allocated 474 * the next time output is attempted. 475 */ 476 rtfree(rt); 477 } 478 } 479 480 /* 481 * After a routing change, flush old routing 482 * and allocate a (hopefully) better one. 483 */ 484 void 485 in_rtchange(inp, errno) 486 register struct inpcb *inp; 487 int errno; 488 { 489 if (inp->inp_route.ro_rt) { 490 rtfree(inp->inp_route.ro_rt); 491 inp->inp_route.ro_rt = 0; 492 /* 493 * A new route can be allocated the next time 494 * output is attempted. 495 */ 496 } 497 } 498 499 struct inpcb * 500 in_pcblookup(table, faddr, fport_arg, laddr, lport_arg, flags) 501 struct inpcbtable *table; 502 struct in_addr faddr, laddr; 503 u_int fport_arg, lport_arg; 504 int flags; 505 { 506 register struct inpcb *inp, *match = 0; 507 int matchwild = 3, wildcard; 508 u_int16_t fport = fport_arg, lport = lport_arg; 509 510 for (inp = table->inpt_queue.cqh_first; 511 inp != (struct inpcb *)&table->inpt_queue; 512 inp = inp->inp_queue.cqe_next) { 513 if (inp->inp_lport != lport) 514 continue; 515 wildcard = 0; 516 if (inp->inp_faddr.s_addr != INADDR_ANY) { 517 if (faddr.s_addr == INADDR_ANY) 518 wildcard++; 519 else if (inp->inp_faddr.s_addr != faddr.s_addr || 520 inp->inp_fport != fport) 521 continue; 522 } else { 523 if (faddr.s_addr != INADDR_ANY) 524 wildcard++; 525 } 526 if (inp->inp_laddr.s_addr != INADDR_ANY) { 527 if (laddr.s_addr == INADDR_ANY) 528 wildcard++; 529 else if (inp->inp_laddr.s_addr != laddr.s_addr) 530 continue; 531 } else { 532 if (laddr.s_addr != INADDR_ANY) 533 wildcard++; 534 } 535 if (wildcard && (flags & INPLOOKUP_WILDCARD) == 0) 536 continue; 537 if (wildcard < matchwild) { 538 match = inp; 539 matchwild = wildcard; 540 if (matchwild == 0) 541 break; 542 } 543 } 544 return (match); 545 } 546 547 void 548 in_pcbrehash(inp) 549 struct inpcb *inp; 550 { 551 struct inpcbtable *table = inp->inp_table; 552 int s; 553 554 s = splnet(); 555 LIST_REMOVE(inp, inp_hash); 556 LIST_INSERT_HEAD(INPCBHASH(table, &inp->inp_faddr, inp->inp_fport, 557 &inp->inp_laddr, inp->inp_lport), inp, inp_hash); 558 splx(s); 559 } 560 561 #ifdef DIAGNOSTIC 562 int in_pcbnotifymiss = 0; 563 #endif 564 565 struct inpcb * 566 in_pcbhashlookup(table, faddr, fport_arg, laddr, lport_arg) 567 struct inpcbtable *table; 568 struct in_addr faddr, laddr; 569 u_int fport_arg, lport_arg; 570 { 571 struct inpcbhead *head; 572 register struct inpcb *inp; 573 u_int16_t fport = fport_arg, lport = lport_arg; 574 575 head = INPCBHASH(table, &faddr, fport, &laddr, lport); 576 for (inp = head->lh_first; inp != NULL; inp = inp->inp_hash.le_next) { 577 if (inp->inp_faddr.s_addr == faddr.s_addr && 578 inp->inp_fport == fport && 579 inp->inp_lport == lport && 580 inp->inp_laddr.s_addr == laddr.s_addr) { 581 /* 582 * Move this PCB to the head of hash chain so that 583 * repeated accesses are quicker. This is analogous to 584 * the historic single-entry PCB cache. 585 */ 586 if (inp != head->lh_first) { 587 LIST_REMOVE(inp, inp_hash); 588 LIST_INSERT_HEAD(head, inp, inp_hash); 589 } 590 break; 591 } 592 } 593 #ifdef DIAGNOSTIC 594 if (inp == NULL && in_pcbnotifymiss) { 595 printf("in_pcbhashlookup: faddr=%08x fport=%d laddr=%08x lport=%d\n", 596 ntohl(faddr.s_addr), ntohs(fport), 597 ntohl(laddr.s_addr), ntohs(lport)); 598 } 599 #endif 600 return (inp); 601 } 602
Indexes created Mon Oct 20 05:10:11 GMT 2025