1 /* 2 * Copyright (c) 1988, 1991, 1993 3 * The Regents of the University of California. All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 3. All advertising materials mentioning features or use of this software 14 * must display the following acknowledgement: 15 * This product includes software developed by the University of 16 * California, Berkeley and its contributors. 17 * 4. Neither the name of the University nor the names of its contributors 18 * may be used to endorse or promote products derived from this software 19 * without specific prior written permission. 20 * 21 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 24 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 31 * SUCH DAMAGE. 32 * 33 * @(#)rtsock.c 8.7 (Berkeley) 10/12/95 34 * $FreeBSD: src/sys/net/rtsock.c,v 1.44.2.11 2002/12/04 14:05:41 ru Exp $ 35 * $DragonFly: src/sys/net/rtsock.c,v 1.16 2004/12/15 00:11:04 hsu Exp $ 36 */ 37 38 39 #include <sys/param.h> 40 #include <sys/systm.h> 41 #include <sys/kernel.h> 42 #include <sys/sysctl.h> 43 #include <sys/proc.h> 44 #include <sys/malloc.h> 45 #include <sys/mbuf.h> 46 #include <sys/protosw.h> 47 #include <sys/socket.h> 48 #include <sys/socketvar.h> 49 #include <sys/domain.h> 50 51 #include <machine/stdarg.h> 52 53 #include <net/if.h> 54 #include <net/route.h> 55 #include <net/raw_cb.h> 56 57 MALLOC_DEFINE(M_RTABLE, "routetbl", "routing tables"); 58 59 static struct sockaddr route_dst = { 2, PF_ROUTE, }; 60 static struct sockaddr route_src = { 2, PF_ROUTE, }; 61 static struct sockaddr sa_zero = { sizeof(sa_zero), AF_INET, }; 62 static struct sockproto route_proto = { PF_ROUTE, }; 63 64 struct walkarg { 65 int w_tmemsize; 66 int w_op, w_arg; 67 caddr_t w_tmem; 68 struct sysctl_req *w_req; 69 }; 70 71 static struct mbuf * 72 rt_msg1 (int, struct rt_addrinfo *); 73 static int rt_msg2 (int, struct rt_addrinfo *, caddr_t, struct walkarg *); 74 static int rt_xaddrs (char *, char *, struct rt_addrinfo *); 75 static int sysctl_dumpentry (struct radix_node *rn, void *vw); 76 static int sysctl_iflist (int af, struct walkarg *w); 77 static int route_output(struct mbuf *, struct socket *, ...); 78 static void rt_setmetrics (u_long, struct rt_metrics *, 79 struct rt_metrics *); 80 81 /* 82 * It really doesn't make any sense at all for this code to share much 83 * with raw_usrreq.c, since its functionality is so restricted. XXX 84 */ 85 static int 86 rts_abort(struct socket *so) 87 { 88 int s, error; 89 s = splnet(); 90 error = raw_usrreqs.pru_abort(so); 91 splx(s); 92 return error; 93 } 94 95 /* pru_accept is EOPNOTSUPP */ 96 97 static int 98 rts_attach(struct socket *so, int proto, struct pru_attach_info *ai) 99 { 100 struct rawcb *rp; 101 int s, error; 102 103 if (sotorawcb(so) != NULL) 104 return EISCONN; /* XXX panic? */ 105 MALLOC(rp, struct rawcb *, sizeof *rp, M_PCB, M_WAITOK|M_ZERO); 106 if (rp == NULL) 107 return ENOBUFS; 108 109 /* 110 * The splnet() is necessary to block protocols from sending 111 * error notifications (like RTM_REDIRECT or RTM_LOSING) while 112 * this PCB is extant but incompletely initialized. 113 * Probably we should try to do more of this work beforehand and 114 * eliminate the spl. 115 */ 116 s = splnet(); 117 so->so_pcb = rp; 118 error = raw_attach(so, proto, ai->sb_rlimit); 119 rp = sotorawcb(so); 120 if (error) { 121 splx(s); 122 free(rp, M_PCB); 123 return error; 124 } 125 switch(rp->rcb_proto.sp_protocol) { 126 case AF_INET: 127 route_cb.ip_count++; 128 break; 129 case AF_INET6: 130 route_cb.ip6_count++; 131 break; 132 case AF_IPX: 133 route_cb.ipx_count++; 134 break; 135 case AF_NS: 136 route_cb.ns_count++; 137 break; 138 } 139 rp->rcb_faddr = &route_src; 140 route_cb.any_count++; 141 soisconnected(so); 142 so->so_options |= SO_USELOOPBACK; 143 splx(s); 144 return 0; 145 } 146 147 static int 148 rts_bind(struct socket *so, struct sockaddr *nam, struct thread *td) 149 { 150 int s, error; 151 s = splnet(); 152 error = raw_usrreqs.pru_bind(so, nam, td); /* xxx just EINVAL */ 153 splx(s); 154 return error; 155 } 156 157 static int 158 rts_connect(struct socket *so, struct sockaddr *nam, struct thread *td) 159 { 160 int s, error; 161 s = splnet(); 162 error = raw_usrreqs.pru_connect(so, nam, td); /* XXX just EINVAL */ 163 splx(s); 164 return error; 165 } 166 167 /* pru_connect2 is EOPNOTSUPP */ 168 /* pru_control is EOPNOTSUPP */ 169 170 static int 171 rts_detach(struct socket *so) 172 { 173 struct rawcb *rp = sotorawcb(so); 174 int s, error; 175 176 s = splnet(); 177 if (rp != NULL) { 178 switch(rp->rcb_proto.sp_protocol) { 179 case AF_INET: 180 route_cb.ip_count--; 181 break; 182 case AF_INET6: 183 route_cb.ip6_count--; 184 break; 185 case AF_IPX: 186 route_cb.ipx_count--; 187 break; 188 case AF_NS: 189 route_cb.ns_count--; 190 break; 191 } 192 route_cb.any_count--; 193 } 194 error = raw_usrreqs.pru_detach(so); 195 splx(s); 196 return error; 197 } 198 199 static int 200 rts_disconnect(struct socket *so) 201 { 202 int s, error; 203 s = splnet(); 204 error = raw_usrreqs.pru_disconnect(so); 205 splx(s); 206 return error; 207 } 208 209 /* pru_listen is EOPNOTSUPP */ 210 211 static int 212 rts_peeraddr(struct socket *so, struct sockaddr **nam) 213 { 214 int s, error; 215 s = splnet(); 216 error = raw_usrreqs.pru_peeraddr(so, nam); 217 splx(s); 218 return error; 219 } 220 221 /* pru_rcvd is EOPNOTSUPP */ 222 /* pru_rcvoob is EOPNOTSUPP */ 223 224 static int 225 rts_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam, 226 struct mbuf *control, struct thread *td) 227 { 228 int s, error; 229 s = splnet(); 230 error = raw_usrreqs.pru_send(so, flags, m, nam, control, td); 231 splx(s); 232 return error; 233 } 234 235 /* pru_sense is null */ 236 237 static int 238 rts_shutdown(struct socket *so) 239 { 240 int s, error; 241 s = splnet(); 242 error = raw_usrreqs.pru_shutdown(so); 243 splx(s); 244 return error; 245 } 246 247 static int 248 rts_sockaddr(struct socket *so, struct sockaddr **nam) 249 { 250 int s, error; 251 s = splnet(); 252 error = raw_usrreqs.pru_sockaddr(so, nam); 253 splx(s); 254 return error; 255 } 256 257 static struct pr_usrreqs route_usrreqs = { 258 rts_abort, pru_accept_notsupp, rts_attach, rts_bind, rts_connect, 259 pru_connect2_notsupp, pru_control_notsupp, rts_detach, rts_disconnect, 260 pru_listen_notsupp, rts_peeraddr, pru_rcvd_notsupp, pru_rcvoob_notsupp, 261 rts_send, pru_sense_null, rts_shutdown, rts_sockaddr, 262 sosend, soreceive, sopoll 263 }; 264 265 /*ARGSUSED*/ 266 static int 267 route_output(struct mbuf *m, struct socket *so, ...) 268 { 269 struct rt_msghdr *rtm = NULL; 270 struct rtentry *rt = NULL; 271 struct rtentry *saved_nrt = NULL; 272 struct radix_node_head *rnh; 273 struct ifnet *ifp = NULL; 274 struct ifaddr *ifa = NULL; 275 struct pr_output_info *oi; 276 struct rt_addrinfo info; 277 int len, error = 0; 278 __va_list ap; 279 280 __va_start(ap, so); 281 oi = __va_arg(ap, struct pr_output_info *); 282 __va_end(ap); 283 284 #define gotoerr(e) { error = e; goto flush;} 285 if (m == NULL || ((m->m_len < sizeof(long)) && 286 (m = m_pullup(m, sizeof(long))) == NULL)) 287 return (ENOBUFS); 288 if (!(m->m_flags & M_PKTHDR)) 289 panic("route_output"); 290 len = m->m_pkthdr.len; 291 if (len < sizeof(*rtm) || 292 len != mtod(m, struct rt_msghdr *)->rtm_msglen) { 293 info.sa_dst = NULL; 294 gotoerr(EINVAL); 295 } 296 R_Malloc(rtm, struct rt_msghdr *, len); 297 if (rtm == NULL) { 298 info.sa_dst = NULL; 299 gotoerr(ENOBUFS); 300 } 301 m_copydata(m, 0, len, (caddr_t)rtm); 302 if (rtm->rtm_version != RTM_VERSION) { 303 info.sa_dst = NULL; 304 gotoerr(EPROTONOSUPPORT); 305 } 306 rtm->rtm_pid = oi->p_pid; 307 bzero(&info, sizeof(info)); 308 info.rti_addrs = rtm->rtm_addrs; 309 if (rt_xaddrs((char *)(rtm + 1), len + (char *)rtm, &info)) { 310 info.sa_dst = NULL; 311 gotoerr(EINVAL); 312 } 313 info.rti_flags = rtm->rtm_flags; 314 if (info.sa_dst == NULL || info.sa_dst->sa_family >= AF_MAX || 315 (info.sa_gateway != NULL && (info.sa_gateway->sa_family >= AF_MAX))) 316 gotoerr(EINVAL); 317 318 if (info.sa_genmask != NULL) { 319 struct radix_node *t; 320 int klen; 321 322 t = rn_addmask((char *)info.sa_genmask, TRUE, 1); 323 if (t != NULL && 324 info.sa_genmask->sa_len >= (klen = *(u_char *)t->rn_key) && 325 bcmp((char *)info.sa_genmask + 1, (char *)t->rn_key + 1, 326 klen - 1) == 0) 327 info.sa_genmask = (struct sockaddr *)(t->rn_key); 328 else 329 gotoerr(ENOBUFS); 330 } 331 332 /* 333 * Verify that the caller has the appropriate privilege; RTM_GET 334 * is the only operation the non-superuser is allowed. 335 */ 336 if (rtm->rtm_type != RTM_GET && suser_cred(so->so_cred, 0) != 0) 337 gotoerr(EPERM); 338 339 switch (rtm->rtm_type) { 340 341 case RTM_ADD: 342 if (info.sa_gateway == NULL) 343 gotoerr(EINVAL); 344 error = rtrequest1(RTM_ADD, &info, &saved_nrt); 345 if (error == 0 && saved_nrt != NULL) { 346 rt_setmetrics(rtm->rtm_inits, 347 &rtm->rtm_rmx, &saved_nrt->rt_rmx); 348 saved_nrt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits); 349 saved_nrt->rt_rmx.rmx_locks |= 350 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks); 351 saved_nrt->rt_refcnt--; 352 saved_nrt->rt_genmask = info.sa_genmask; 353 } 354 break; 355 356 case RTM_DELETE: 357 error = rtrequest1(RTM_DELETE, &info, &saved_nrt); 358 if (error == 0) { 359 if ((rt = saved_nrt)) 360 rt->rt_refcnt++; 361 goto report; 362 } 363 break; 364 365 case RTM_GET: 366 case RTM_CHANGE: 367 case RTM_LOCK: 368 if ((rnh = rt_tables[info.sa_dst->sa_family]) == NULL) { 369 gotoerr(EAFNOSUPPORT); 370 } else if ((rt = (struct rtentry *) rnh->rnh_lookup( 371 (char *)info.sa_dst, (char *)info.sa_netmask, rnh)) != NULL) 372 rt->rt_refcnt++; 373 else 374 gotoerr(ESRCH); 375 switch(rtm->rtm_type) { 376 377 case RTM_GET: 378 report: 379 info.sa_dst = rt_key(rt); 380 info.sa_gateway = rt->rt_gateway; 381 info.sa_netmask = rt_mask(rt); 382 info.sa_genmask = rt->rt_genmask; 383 if (rtm->rtm_addrs & (RTA_IFP | RTA_IFA)) { 384 ifp = rt->rt_ifp; 385 if (ifp) { 386 info.sa_ifpaddr = 387 TAILQ_FIRST(&ifp->if_addrhead)-> 388 ifa_addr; 389 info.sa_ifaaddr = rt->rt_ifa->ifa_addr; 390 if (ifp->if_flags & IFF_POINTOPOINT) 391 info.sa_bcastaddr = 392 rt->rt_ifa->ifa_dstaddr; 393 rtm->rtm_index = ifp->if_index; 394 } else { 395 info.sa_ifpaddr = NULL; 396 info.sa_ifaaddr = NULL; 397 } 398 } 399 len = rt_msg2(rtm->rtm_type, &info, NULL, NULL); 400 if (len > rtm->rtm_msglen) { 401 struct rt_msghdr *new_rtm; 402 R_Malloc(new_rtm, struct rt_msghdr *, len); 403 if (new_rtm == NULL) 404 gotoerr(ENOBUFS); 405 bcopy(rtm, new_rtm, rtm->rtm_msglen); 406 Free(rtm); rtm = new_rtm; 407 } 408 (void)rt_msg2(rtm->rtm_type, &info, (caddr_t)rtm, NULL); 409 rtm->rtm_flags = rt->rt_flags; 410 rtm->rtm_rmx = rt->rt_rmx; 411 rtm->rtm_addrs = info.rti_addrs; 412 break; 413 414 case RTM_CHANGE: 415 /* 416 * new gateway could require new ifaddr, ifp; 417 * flags may also be different; ifp may be specified 418 * by ll sockaddr when protocol address is ambiguous 419 */ 420 if (((rt->rt_flags & RTF_GATEWAY) && 421 info.sa_gateway != NULL) || 422 info.sa_ifpaddr != NULL || 423 (info.sa_ifaaddr != NULL && 424 bcmp(info.sa_ifaaddr, rt->rt_ifa->ifa_addr, 425 info.sa_ifaaddr->sa_len) == 0)) { 426 if ((error = rt_getifa(&info)) != 0) 427 gotoerr(error); 428 } 429 if (info.sa_gateway != NULL && 430 (error = rt_setgate(rt, rt_key(rt), 431 info.sa_gateway)) != 0) 432 gotoerr(error); 433 if ((ifa = info.rti_ifa) != NULL) { 434 struct ifaddr *oifa = rt->rt_ifa; 435 436 if (oifa != ifa) { 437 if (oifa && oifa->ifa_rtrequest) 438 oifa->ifa_rtrequest(RTM_DELETE, rt, 439 &info); 440 IFAFREE(rt->rt_ifa); 441 rt->rt_ifa = ifa; 442 ifa->ifa_refcnt++; 443 rt->rt_ifp = info.rti_ifp; 444 } 445 } 446 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, 447 &rt->rt_rmx); 448 if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest) 449 rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt, &info); 450 if (info.sa_genmask != NULL) 451 rt->rt_genmask = info.sa_genmask; 452 /* 453 * Fall into 454 */ 455 case RTM_LOCK: 456 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits); 457 rt->rt_rmx.rmx_locks |= 458 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks); 459 break; 460 } 461 break; 462 463 default: 464 gotoerr(EOPNOTSUPP); 465 } 466 467 flush: 468 if (rtm) { 469 if (error) 470 rtm->rtm_errno = error; 471 else 472 rtm->rtm_flags |= RTF_DONE; 473 } 474 if (rt) 475 rtfree(rt); 476 { 477 struct rawcb *rp = NULL; 478 /* 479 * Check to see if we don't want our own messages. 480 */ 481 if (!(so->so_options & SO_USELOOPBACK)) { 482 if (route_cb.any_count <= 1) { 483 if (rtm) 484 Free(rtm); 485 m_freem(m); 486 return (error); 487 } 488 /* There is another listener, so construct message */ 489 rp = sotorawcb(so); 490 } 491 if (rtm) { 492 m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm); 493 if (m->m_pkthdr.len < rtm->rtm_msglen) { 494 m_freem(m); 495 m = NULL; 496 } else if (m->m_pkthdr.len > rtm->rtm_msglen) 497 m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len); 498 Free(rtm); 499 } 500 if (rp != NULL) 501 rp->rcb_proto.sp_family = 0; /* Avoid us */ 502 if (info.sa_dst != NULL) 503 route_proto.sp_protocol = info.sa_dst->sa_family; 504 if (m != NULL) 505 raw_input(m, &route_proto, &route_src, &route_dst); 506 if (rp != NULL) 507 rp->rcb_proto.sp_family = PF_ROUTE; 508 } 509 return (error); 510 } 511 512 static void 513 rt_setmetrics(u_long which, struct rt_metrics *in, struct rt_metrics *out) 514 { 515 #define setmetric(flag, elt) if (which & (flag)) out->elt = in->elt; 516 setmetric(RTV_RPIPE, rmx_recvpipe); 517 setmetric(RTV_SPIPE, rmx_sendpipe); 518 setmetric(RTV_SSTHRESH, rmx_ssthresh); 519 setmetric(RTV_RTT, rmx_rtt); 520 setmetric(RTV_RTTVAR, rmx_rttvar); 521 setmetric(RTV_HOPCOUNT, rmx_hopcount); 522 setmetric(RTV_MTU, rmx_mtu); 523 setmetric(RTV_EXPIRE, rmx_expire); 524 #undef setmetric 525 } 526 527 #define ROUNDUP(a) \ 528 ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long)) 529 #define ADVANCE(x, n) (x += ROUNDUP((n)->sa_len)) 530 531 /* 532 * Extract the addresses of the passed sockaddrs. 533 * Do a little sanity checking so as to avoid bad memory references. 534 * This data is derived straight from userland. 535 */ 536 static int 537 rt_xaddrs(char *cp, char *cplim, struct rt_addrinfo *rtinfo) 538 { 539 struct sockaddr *sa; 540 int i; 541 542 for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) { 543 if ((rtinfo->rti_addrs & (1 << i)) == 0) 544 continue; 545 sa = (struct sockaddr *)cp; 546 /* 547 * It won't fit. 548 */ 549 if ( (cp + sa->sa_len) > cplim ) { 550 return (EINVAL); 551 } 552 553 /* 554 * There are no more... Quit now. 555 * If there are more bits, they are in error. 556 * I've seen this. route(1) can evidently generate these. 557 * This causes kernel to core dump. 558 * For compatibility, if we see this, point to a safe address. 559 */ 560 if (sa->sa_len == 0) { 561 rtinfo->rti_info[i] = &sa_zero; 562 return (0); /* should be EINVAL but for compat */ 563 } 564 565 /* Accept the sockaddr. */ 566 rtinfo->rti_info[i] = sa; 567 ADVANCE(cp, sa); 568 } 569 return (0); 570 } 571 572 static struct mbuf * 573 rt_msg1(type, rtinfo) 574 int type; 575 struct rt_addrinfo *rtinfo; 576 { 577 struct rt_msghdr *rtm; 578 struct mbuf *m; 579 int i; 580 struct sockaddr *sa; 581 int len, dlen; 582 583 switch (type) { 584 585 case RTM_DELADDR: 586 case RTM_NEWADDR: 587 len = sizeof(struct ifa_msghdr); 588 break; 589 590 case RTM_DELMADDR: 591 case RTM_NEWMADDR: 592 len = sizeof(struct ifma_msghdr); 593 break; 594 595 case RTM_IFINFO: 596 len = sizeof(struct if_msghdr); 597 break; 598 599 case RTM_IFANNOUNCE: 600 len = sizeof(struct if_announcemsghdr); 601 break; 602 603 default: 604 len = sizeof(struct rt_msghdr); 605 } 606 if (len > MCLBYTES) 607 panic("rt_msg1"); 608 m = m_gethdr(MB_DONTWAIT, MT_DATA); 609 if (m && len > MHLEN) { 610 MCLGET(m, MB_DONTWAIT); 611 if (!(m->m_flags & M_EXT)) { 612 m_free(m); 613 m = NULL; 614 } 615 } 616 if (m == NULL) 617 return (m); 618 m->m_pkthdr.len = m->m_len = len; 619 m->m_pkthdr.rcvif = NULL; 620 rtm = mtod(m, struct rt_msghdr *); 621 bzero(rtm, len); 622 for (i = 0; i < RTAX_MAX; i++) { 623 if ((sa = rtinfo->rti_info[i]) == NULL) 624 continue; 625 rtinfo->rti_addrs |= (1 << i); 626 dlen = ROUNDUP(sa->sa_len); 627 m_copyback(m, len, dlen, (caddr_t)sa); 628 len += dlen; 629 } 630 if (m->m_pkthdr.len != len) { 631 m_freem(m); 632 return (NULL); 633 } 634 rtm->rtm_msglen = len; 635 rtm->rtm_version = RTM_VERSION; 636 rtm->rtm_type = type; 637 return (m); 638 } 639 640 static int 641 rt_msg2(int type, struct rt_addrinfo *rtinfo, caddr_t cp, struct walkarg *w) 642 { 643 int i; 644 int len, dlen; 645 boolean_t second_time = FALSE; 646 caddr_t cp0; 647 648 rtinfo->rti_addrs = NULL; 649 again: 650 switch (type) { 651 652 case RTM_DELADDR: 653 case RTM_NEWADDR: 654 len = sizeof(struct ifa_msghdr); 655 break; 656 657 case RTM_IFINFO: 658 len = sizeof(struct if_msghdr); 659 break; 660 661 default: 662 len = sizeof(struct rt_msghdr); 663 } 664 cp0 = cp; 665 if (cp != NULL) 666 cp += len; 667 668 for (i = 0; i < RTAX_MAX; i++) { 669 struct sockaddr *sa; 670 671 if ((sa = rtinfo->rti_info[i]) == NULL) 672 continue; 673 rtinfo->rti_addrs |= (1 << i); 674 dlen = ROUNDUP(sa->sa_len); 675 if (cp != NULL) { 676 bcopy(sa, cp, dlen); 677 cp += dlen; 678 } 679 len += dlen; 680 } 681 len = ALIGN(len); 682 if (cp == NULL && w != NULL && !second_time) { 683 struct walkarg *rw = w; 684 685 if (rw->w_req != NULL) { 686 if (rw->w_tmemsize < len) { 687 if (rw->w_tmem) 688 free(rw->w_tmem, M_RTABLE); 689 rw->w_tmem = malloc(len, M_RTABLE, 690 M_INTWAIT | M_NULLOK); 691 if (rw->w_tmem) 692 rw->w_tmemsize = len; 693 } 694 if (rw->w_tmem != NULL) { 695 cp = rw->w_tmem; 696 second_time = TRUE; 697 goto again; 698 } 699 } 700 } 701 if (cp != NULL) { 702 struct rt_msghdr *rtm = (struct rt_msghdr *)cp0; 703 704 rtm->rtm_version = RTM_VERSION; 705 rtm->rtm_type = type; 706 rtm->rtm_msglen = len; 707 } 708 return (len); 709 } 710 711 /* 712 * This routine is called to generate a message from the routing 713 * socket indicating that a redirect has occured, a routing lookup 714 * has failed, or that a protocol has detected timeouts to a particular 715 * destination. 716 */ 717 void 718 rt_missmsg(type, rtinfo, flags, error) 719 int type, flags, error; 720 struct rt_addrinfo *rtinfo; 721 { 722 struct rt_msghdr *rtm; 723 struct mbuf *m; 724 struct sockaddr *sa = rtinfo->sa_dst; 725 726 if (route_cb.any_count == 0) 727 return; 728 m = rt_msg1(type, rtinfo); 729 if (m == NULL) 730 return; 731 rtm = mtod(m, struct rt_msghdr *); 732 rtm->rtm_flags = RTF_DONE | flags; 733 rtm->rtm_errno = error; 734 rtm->rtm_addrs = rtinfo->rti_addrs; 735 route_proto.sp_protocol = sa ? sa->sa_family : 0; 736 raw_input(m, &route_proto, &route_src, &route_dst); 737 } 738 739 /* 740 * This routine is called to generate a message from the routing 741 * socket indicating that the status of a network interface has changed. 742 */ 743 void 744 rt_ifmsg(ifp) 745 struct ifnet *ifp; 746 { 747 struct if_msghdr *ifm; 748 struct mbuf *m; 749 struct rt_addrinfo info; 750 751 if (route_cb.any_count == 0) 752 return; 753 bzero(&info, sizeof(info)); 754 m = rt_msg1(RTM_IFINFO, &info); 755 if (m == NULL) 756 return; 757 ifm = mtod(m, struct if_msghdr *); 758 ifm->ifm_index = ifp->if_index; 759 ifm->ifm_flags = (u_short)ifp->if_flags; 760 ifm->ifm_data = ifp->if_data; 761 ifm->ifm_addrs = NULL; 762 route_proto.sp_protocol = 0; 763 raw_input(m, &route_proto, &route_src, &route_dst); 764 } 765 766 static void 767 rt_ifamsg(int cmd, struct ifaddr *ifa) 768 { 769 struct ifa_msghdr *ifam; 770 struct rt_addrinfo info; 771 struct mbuf *m; 772 struct sockaddr *sa; 773 struct ifnet *ifp = ifa->ifa_ifp; 774 775 bzero(&info, sizeof(info)); 776 info.sa_ifaaddr = sa = ifa->ifa_addr; 777 info.sa_ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 778 info.sa_netmask = ifa->ifa_netmask; 779 info.sa_bcastaddr = ifa->ifa_dstaddr; 780 if ((m = rt_msg1(cmd, &info)) == NULL) 781 return; 782 ifam = mtod(m, struct ifa_msghdr *); 783 ifam->ifam_index = ifp->if_index; 784 ifam->ifam_metric = ifa->ifa_metric; 785 ifam->ifam_flags = ifa->ifa_flags; 786 ifam->ifam_addrs = info.rti_addrs; 787 788 route_proto.sp_protocol = sa ? sa->sa_family : 0; 789 raw_input(m, &route_proto, &route_src, &route_dst); 790 } 791 792 static void 793 rt_rtmsg(int cmd, struct ifaddr *ifa, int error, struct rtentry *rt) 794 { 795 struct rt_msghdr *rtm; 796 struct rt_addrinfo info; 797 struct mbuf *m; 798 struct sockaddr *sa; 799 struct ifnet *ifp = ifa->ifa_ifp; 800 801 if (rt == NULL) 802 return; 803 bzero(&info, sizeof(info)); 804 info.sa_netmask = rt_mask(rt); 805 info.sa_dst = sa = rt_key(rt); 806 info.sa_gateway = rt->rt_gateway; 807 if ((m = rt_msg1(cmd, &info)) == NULL) 808 return; 809 rtm = mtod(m, struct rt_msghdr *); 810 rtm->rtm_index = ifp->if_index; 811 rtm->rtm_flags |= rt->rt_flags; 812 rtm->rtm_errno = error; 813 rtm->rtm_addrs = info.rti_addrs; 814 815 route_proto.sp_protocol = sa ? sa->sa_family : 0; 816 raw_input(m, &route_proto, &route_src, &route_dst); 817 } 818 819 /* 820 * This is called to generate messages from the routing socket 821 * indicating a network interface has had addresses associated with it. 822 * if we ever reverse the logic and replace messages TO the routing 823 * socket indicate a request to configure interfaces, then it will 824 * be unnecessary as the routing socket will automatically generate 825 * copies of it. 826 */ 827 void 828 rt_newaddrmsg(cmd, ifa, error, rt) 829 int cmd, error; 830 struct ifaddr *ifa; 831 struct rtentry *rt; 832 { 833 if (route_cb.any_count == 0) 834 return; 835 836 if (cmd == RTM_ADD) { 837 rt_ifamsg(RTM_NEWADDR, ifa); 838 rt_rtmsg(RTM_ADD, ifa, error, rt); 839 } else { 840 KASSERT((cmd == RTM_DELETE), ("unknown cmd %d", cmd)); 841 rt_rtmsg(RTM_DELETE, ifa, error, rt); 842 rt_ifamsg(RTM_DELADDR, ifa); 843 } 844 } 845 846 /* 847 * This is the analogue to the rt_newaddrmsg which performs the same 848 * function but for multicast group memberhips. This is easier since 849 * there is no route state to worry about. 850 */ 851 void 852 rt_newmaddrmsg(cmd, ifma) 853 int cmd; 854 struct ifmultiaddr *ifma; 855 { 856 struct rt_addrinfo info; 857 struct mbuf *m = NULL; 858 struct ifnet *ifp = ifma->ifma_ifp; 859 struct ifma_msghdr *ifmam; 860 861 if (route_cb.any_count == 0) 862 return; 863 864 bzero(&info, sizeof(info)); 865 info.sa_ifaaddr = ifma->ifma_addr; 866 if (ifp != NULL && TAILQ_FIRST(&ifp->if_addrhead) != NULL) 867 info.sa_ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 868 else 869 info.sa_ifpaddr = NULL; 870 /* 871 * If a link-layer address is present, present it as a ``gateway'' 872 * (similarly to how ARP entries, e.g., are presented). 873 */ 874 info.sa_gateway = ifma->ifma_lladdr; 875 if ((m = rt_msg1(cmd, &info)) == NULL) 876 return; 877 ifmam = mtod(m, struct ifma_msghdr *); 878 ifmam->ifmam_index = ifp->if_index; 879 ifmam->ifmam_addrs = info.rti_addrs; 880 route_proto.sp_protocol = ifma->ifma_addr->sa_family; 881 raw_input(m, &route_proto, &route_src, &route_dst); 882 } 883 884 /* 885 * This is called to generate routing socket messages indicating 886 * network interface arrival and departure. 887 */ 888 void 889 rt_ifannouncemsg(ifp, what) 890 struct ifnet *ifp; 891 int what; 892 { 893 struct if_announcemsghdr *ifan; 894 struct mbuf *m; 895 struct rt_addrinfo info; 896 897 if (route_cb.any_count == 0) 898 return; 899 bzero(&info, sizeof(info)); 900 m = rt_msg1(RTM_IFANNOUNCE, &info); 901 if (m == NULL) 902 return; 903 ifan = mtod(m, struct if_announcemsghdr *); 904 ifan->ifan_index = ifp->if_index; 905 strlcpy(ifan->ifan_name, ifp->if_xname, sizeof(ifan->ifan_name)); 906 ifan->ifan_what = what; 907 route_proto.sp_protocol = 0; 908 raw_input(m, &route_proto, &route_src, &route_dst); 909 } 910 911 /* 912 * This is used in dumping the kernel table via sysctl(). 913 */ 914 int 915 sysctl_dumpentry(rn, vw) 916 struct radix_node *rn; 917 void *vw; 918 { 919 struct walkarg *w = vw; 920 struct rtentry *rt = (struct rtentry *)rn; 921 int error = 0, size; 922 struct rt_addrinfo info; 923 924 if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg)) 925 return 0; 926 bzero(&info, sizeof(info)); 927 info.sa_dst = rt_key(rt); 928 info.sa_gateway = rt->rt_gateway; 929 info.sa_netmask = rt_mask(rt); 930 info.sa_genmask = rt->rt_genmask; 931 if (rt->rt_ifp != NULL) { 932 info.sa_ifpaddr = 933 TAILQ_FIRST(&rt->rt_ifp->if_addrhead)->ifa_addr; 934 info.sa_ifaaddr = rt->rt_ifa->ifa_addr; 935 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT) 936 info.sa_bcastaddr = rt->rt_ifa->ifa_dstaddr; 937 } 938 size = rt_msg2(RTM_GET, &info, NULL, w); 939 if (w->w_req != NULL && w->w_tmem != NULL) { 940 struct rt_msghdr *rtm = (struct rt_msghdr *)w->w_tmem; 941 942 rtm->rtm_flags = rt->rt_flags; 943 rtm->rtm_use = rt->rt_use; 944 rtm->rtm_rmx = rt->rt_rmx; 945 rtm->rtm_index = rt->rt_ifp->if_index; 946 rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0; 947 rtm->rtm_addrs = info.rti_addrs; 948 error = SYSCTL_OUT(w->w_req, (caddr_t)rtm, size); 949 return (error); 950 } 951 return (error); 952 } 953 954 int 955 sysctl_iflist(af, w) 956 int af; 957 struct walkarg *w; 958 { 959 struct ifnet *ifp; 960 struct ifaddr *ifa; 961 struct rt_addrinfo info; 962 int len, error = 0; 963 964 bzero(&info, sizeof(info)); 965 TAILQ_FOREACH(ifp, &ifnet, if_link) { 966 if (w->w_arg && w->w_arg != ifp->if_index) 967 continue; 968 ifa = TAILQ_FIRST(&ifp->if_addrhead); 969 info.sa_ifpaddr = ifa->ifa_addr; 970 len = rt_msg2(RTM_IFINFO, &info, NULL, w); 971 info.sa_ifpaddr = NULL; 972 if (w->w_req != NULL && w->w_tmem != NULL) { 973 struct if_msghdr *ifm; 974 975 ifm = (struct if_msghdr *)w->w_tmem; 976 ifm->ifm_index = ifp->if_index; 977 ifm->ifm_flags = (u_short)ifp->if_flags; 978 ifm->ifm_data = ifp->if_data; 979 ifm->ifm_addrs = info.rti_addrs; 980 error = SYSCTL_OUT(w->w_req,(caddr_t)ifm, len); 981 if (error) 982 return (error); 983 } 984 while ((ifa = TAILQ_NEXT(ifa, ifa_link)) != NULL) { 985 if (af && af != ifa->ifa_addr->sa_family) 986 continue; 987 if (curproc->p_ucred->cr_prison && prison_if(curthread, ifa->ifa_addr)) 988 continue; 989 info.sa_ifaaddr = ifa->ifa_addr; 990 info.sa_netmask = ifa->ifa_netmask; 991 info.sa_bcastaddr = ifa->ifa_dstaddr; 992 len = rt_msg2(RTM_NEWADDR, &info, NULL, w); 993 if (w->w_req && w->w_tmem) { 994 struct ifa_msghdr *ifam; 995 996 ifam = (struct ifa_msghdr *)w->w_tmem; 997 ifam->ifam_index = ifa->ifa_ifp->if_index; 998 ifam->ifam_flags = ifa->ifa_flags; 999 ifam->ifam_metric = ifa->ifa_metric; 1000 ifam->ifam_addrs = info.rti_addrs; 1001 error = SYSCTL_OUT(w->w_req, w->w_tmem, len); 1002 if (error) 1003 return (error); 1004 } 1005 } 1006 info.sa_netmask = info.sa_ifaaddr = info.sa_bcastaddr = NULL; 1007 } 1008 return (0); 1009 } 1010 1011 static int 1012 sysctl_rtsock(SYSCTL_HANDLER_ARGS) 1013 { 1014 int *name = (int *)arg1; 1015 u_int namelen = arg2; 1016 struct radix_node_head *rnh; 1017 int i, s, error = EINVAL; 1018 u_char af; 1019 struct walkarg w; 1020 1021 name ++; 1022 namelen--; 1023 if (req->newptr) 1024 return (EPERM); 1025 if (namelen != 3) 1026 return (EINVAL); 1027 af = name[0]; 1028 bzero(&w, sizeof(w)); 1029 w.w_op = name[1]; 1030 w.w_arg = name[2]; 1031 w.w_req = req; 1032 1033 s = splnet(); 1034 switch (w.w_op) { 1035 1036 case NET_RT_DUMP: 1037 case NET_RT_FLAGS: 1038 for (i = 1; i <= AF_MAX; i++) 1039 if ((rnh = rt_tables[i]) && (af == 0 || af == i) && 1040 (error = rnh->rnh_walktree(rnh, 1041 sysctl_dumpentry, &w))) 1042 break; 1043 break; 1044 1045 case NET_RT_IFLIST: 1046 error = sysctl_iflist(af, &w); 1047 } 1048 splx(s); 1049 if (w.w_tmem) 1050 free(w.w_tmem, M_RTABLE); 1051 return (error); 1052 } 1053 1054 SYSCTL_NODE(_net, PF_ROUTE, routetable, CTLFLAG_RD, sysctl_rtsock, ""); 1055 1056 /* 1057 * Definitions of protocols supported in the ROUTE domain. 1058 */ 1059 1060 extern struct domain routedomain; /* or at least forward */ 1061 1062 static struct protosw routesw[] = { 1063 { SOCK_RAW, &routedomain, 0, PR_ATOMIC|PR_ADDR, 1064 0, route_output, raw_ctlinput, 0, 1065 cpu0_soport, 1066 raw_init, 0, 0, 0, 1067 &route_usrreqs 1068 } 1069 }; 1070 1071 static struct domain routedomain = 1072 { PF_ROUTE, "route", 0, 0, 0, 1073 routesw, &routesw[sizeof(routesw)/sizeof(routesw[0])] }; 1074 1075 DOMAIN_SET(route); 1076