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 1.14 2004/06/03 15:04:51 joerg 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 (caddr_t, caddr_t, 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 /* Sleazy use of local variables throughout file, warning!!!! */ 82 #define dst info.rti_info[RTAX_DST] 83 #define gate info.rti_info[RTAX_GATEWAY] 84 #define netmask info.rti_info[RTAX_NETMASK] 85 #define genmask info.rti_info[RTAX_GENMASK] 86 #define ifpaddr info.rti_info[RTAX_IFP] 87 #define ifaaddr info.rti_info[RTAX_IFA] 88 #define brdaddr info.rti_info[RTAX_BRD] 89 90 /* 91 * It really doesn't make any sense at all for this code to share much 92 * with raw_usrreq.c, since its functionality is so restricted. XXX 93 */ 94 static int 95 rts_abort(struct socket *so) 96 { 97 int s, error; 98 s = splnet(); 99 error = raw_usrreqs.pru_abort(so); 100 splx(s); 101 return error; 102 } 103 104 /* pru_accept is EOPNOTSUPP */ 105 106 static int 107 rts_attach(struct socket *so, int proto, struct pru_attach_info *ai) 108 { 109 struct rawcb *rp; 110 int s, error; 111 112 if (sotorawcb(so) != 0) 113 return EISCONN; /* XXX panic? */ 114 MALLOC(rp, struct rawcb *, sizeof *rp, M_PCB, M_WAITOK|M_ZERO); 115 if (rp == 0) 116 return ENOBUFS; 117 118 /* 119 * The splnet() is necessary to block protocols from sending 120 * error notifications (like RTM_REDIRECT or RTM_LOSING) while 121 * this PCB is extant but incompletely initialized. 122 * Probably we should try to do more of this work beforehand and 123 * eliminate the spl. 124 */ 125 s = splnet(); 126 so->so_pcb = (caddr_t)rp; 127 error = raw_attach(so, proto, ai->sb_rlimit); 128 rp = sotorawcb(so); 129 if (error) { 130 splx(s); 131 free(rp, M_PCB); 132 return error; 133 } 134 switch(rp->rcb_proto.sp_protocol) { 135 case AF_INET: 136 route_cb.ip_count++; 137 break; 138 case AF_INET6: 139 route_cb.ip6_count++; 140 break; 141 case AF_IPX: 142 route_cb.ipx_count++; 143 break; 144 case AF_NS: 145 route_cb.ns_count++; 146 break; 147 } 148 rp->rcb_faddr = &route_src; 149 route_cb.any_count++; 150 soisconnected(so); 151 so->so_options |= SO_USELOOPBACK; 152 splx(s); 153 return 0; 154 } 155 156 static int 157 rts_bind(struct socket *so, struct sockaddr *nam, struct thread *td) 158 { 159 int s, error; 160 s = splnet(); 161 error = raw_usrreqs.pru_bind(so, nam, td); /* xxx just EINVAL */ 162 splx(s); 163 return error; 164 } 165 166 static int 167 rts_connect(struct socket *so, struct sockaddr *nam, struct thread *td) 168 { 169 int s, error; 170 s = splnet(); 171 error = raw_usrreqs.pru_connect(so, nam, td); /* XXX just EINVAL */ 172 splx(s); 173 return error; 174 } 175 176 /* pru_connect2 is EOPNOTSUPP */ 177 /* pru_control is EOPNOTSUPP */ 178 179 static int 180 rts_detach(struct socket *so) 181 { 182 struct rawcb *rp = sotorawcb(so); 183 int s, error; 184 185 s = splnet(); 186 if (rp != 0) { 187 switch(rp->rcb_proto.sp_protocol) { 188 case AF_INET: 189 route_cb.ip_count--; 190 break; 191 case AF_INET6: 192 route_cb.ip6_count--; 193 break; 194 case AF_IPX: 195 route_cb.ipx_count--; 196 break; 197 case AF_NS: 198 route_cb.ns_count--; 199 break; 200 } 201 route_cb.any_count--; 202 } 203 error = raw_usrreqs.pru_detach(so); 204 splx(s); 205 return error; 206 } 207 208 static int 209 rts_disconnect(struct socket *so) 210 { 211 int s, error; 212 s = splnet(); 213 error = raw_usrreqs.pru_disconnect(so); 214 splx(s); 215 return error; 216 } 217 218 /* pru_listen is EOPNOTSUPP */ 219 220 static int 221 rts_peeraddr(struct socket *so, struct sockaddr **nam) 222 { 223 int s, error; 224 s = splnet(); 225 error = raw_usrreqs.pru_peeraddr(so, nam); 226 splx(s); 227 return error; 228 } 229 230 /* pru_rcvd is EOPNOTSUPP */ 231 /* pru_rcvoob is EOPNOTSUPP */ 232 233 static int 234 rts_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam, 235 struct mbuf *control, struct thread *td) 236 { 237 int s, error; 238 s = splnet(); 239 error = raw_usrreqs.pru_send(so, flags, m, nam, control, td); 240 splx(s); 241 return error; 242 } 243 244 /* pru_sense is null */ 245 246 static int 247 rts_shutdown(struct socket *so) 248 { 249 int s, error; 250 s = splnet(); 251 error = raw_usrreqs.pru_shutdown(so); 252 splx(s); 253 return error; 254 } 255 256 static int 257 rts_sockaddr(struct socket *so, struct sockaddr **nam) 258 { 259 int s, error; 260 s = splnet(); 261 error = raw_usrreqs.pru_sockaddr(so, nam); 262 splx(s); 263 return error; 264 } 265 266 static struct pr_usrreqs route_usrreqs = { 267 rts_abort, pru_accept_notsupp, rts_attach, rts_bind, rts_connect, 268 pru_connect2_notsupp, pru_control_notsupp, rts_detach, rts_disconnect, 269 pru_listen_notsupp, rts_peeraddr, pru_rcvd_notsupp, pru_rcvoob_notsupp, 270 rts_send, pru_sense_null, rts_shutdown, rts_sockaddr, 271 sosend, soreceive, sopoll 272 }; 273 274 /*ARGSUSED*/ 275 static int 276 route_output(struct mbuf *m, struct socket *so, ...) 277 { 278 struct rt_msghdr *rtm = 0; 279 struct rtentry *rt = 0; 280 struct rtentry *saved_nrt = 0; 281 struct radix_node_head *rnh; 282 struct rt_addrinfo info; 283 int len, error = 0; 284 struct ifnet *ifp = 0; 285 struct ifaddr *ifa = 0; 286 struct pr_output_info *oi; 287 __va_list ap; 288 289 __va_start(ap, so); 290 oi = __va_arg(ap, struct pr_output_info *); 291 __va_end(ap); 292 293 #define senderr(e) { error = e; goto flush;} 294 if (m == 0 || ((m->m_len < sizeof(long)) && 295 (m = m_pullup(m, sizeof(long))) == 0)) 296 return (ENOBUFS); 297 if ((m->m_flags & M_PKTHDR) == 0) 298 panic("route_output"); 299 len = m->m_pkthdr.len; 300 if (len < sizeof(*rtm) || 301 len != mtod(m, struct rt_msghdr *)->rtm_msglen) { 302 dst = 0; 303 senderr(EINVAL); 304 } 305 R_Malloc(rtm, struct rt_msghdr *, len); 306 if (rtm == 0) { 307 dst = 0; 308 senderr(ENOBUFS); 309 } 310 m_copydata(m, 0, len, (caddr_t)rtm); 311 if (rtm->rtm_version != RTM_VERSION) { 312 dst = 0; 313 senderr(EPROTONOSUPPORT); 314 } 315 rtm->rtm_pid = oi->p_pid; 316 bzero(&info, sizeof(info)); 317 info.rti_addrs = rtm->rtm_addrs; 318 if (rt_xaddrs((caddr_t)(rtm + 1), len + (caddr_t)rtm, &info)) { 319 dst = 0; 320 senderr(EINVAL); 321 } 322 info.rti_flags = rtm->rtm_flags; 323 if (dst == 0 || (dst->sa_family >= AF_MAX) 324 || (gate != 0 && (gate->sa_family >= AF_MAX))) 325 senderr(EINVAL); 326 if (genmask) { 327 struct radix_node *t; 328 t = rn_addmask((caddr_t)genmask, 0, 1); 329 if (t && Bcmp((caddr_t *)genmask + 1, (caddr_t *)t->rn_key + 1, 330 *(u_char *)t->rn_key - 1) == 0) 331 genmask = (struct sockaddr *)(t->rn_key); 332 else 333 senderr(ENOBUFS); 334 } 335 336 /* 337 * Verify that the caller has the appropriate privilege; RTM_GET 338 * is the only operation the non-superuser is allowed. 339 */ 340 if (rtm->rtm_type != RTM_GET && suser_cred(so->so_cred, 0) != 0) 341 senderr(EPERM); 342 343 switch (rtm->rtm_type) { 344 345 case RTM_ADD: 346 if (gate == 0) 347 senderr(EINVAL); 348 error = rtrequest1(RTM_ADD, &info, &saved_nrt); 349 if (error == 0 && saved_nrt) { 350 rt_setmetrics(rtm->rtm_inits, 351 &rtm->rtm_rmx, &saved_nrt->rt_rmx); 352 saved_nrt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits); 353 saved_nrt->rt_rmx.rmx_locks |= 354 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks); 355 saved_nrt->rt_refcnt--; 356 saved_nrt->rt_genmask = genmask; 357 } 358 break; 359 360 case RTM_DELETE: 361 error = rtrequest1(RTM_DELETE, &info, &saved_nrt); 362 if (error == 0) { 363 if ((rt = saved_nrt)) 364 rt->rt_refcnt++; 365 goto report; 366 } 367 break; 368 369 case RTM_GET: 370 case RTM_CHANGE: 371 case RTM_LOCK: 372 if ((rnh = rt_tables[dst->sa_family]) == 0) { 373 senderr(EAFNOSUPPORT); 374 } else if ((rt = (struct rtentry *) 375 rnh->rnh_lookup(dst, netmask, rnh)) != NULL) 376 rt->rt_refcnt++; 377 else 378 senderr(ESRCH); 379 switch(rtm->rtm_type) { 380 381 case RTM_GET: 382 report: 383 dst = rt_key(rt); 384 gate = rt->rt_gateway; 385 netmask = rt_mask(rt); 386 genmask = rt->rt_genmask; 387 if (rtm->rtm_addrs & (RTA_IFP | RTA_IFA)) { 388 ifp = rt->rt_ifp; 389 if (ifp) { 390 ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 391 ifaaddr = rt->rt_ifa->ifa_addr; 392 if (ifp->if_flags & IFF_POINTOPOINT) 393 brdaddr = rt->rt_ifa->ifa_dstaddr; 394 rtm->rtm_index = ifp->if_index; 395 } else { 396 ifpaddr = 0; 397 ifaaddr = 0; 398 } 399 } 400 len = rt_msg2(rtm->rtm_type, &info, (caddr_t)0, 401 (struct walkarg *)0); 402 if (len > rtm->rtm_msglen) { 403 struct rt_msghdr *new_rtm; 404 R_Malloc(new_rtm, struct rt_msghdr *, len); 405 if (new_rtm == 0) 406 senderr(ENOBUFS); 407 Bcopy(rtm, new_rtm, rtm->rtm_msglen); 408 Free(rtm); rtm = new_rtm; 409 } 410 (void)rt_msg2(rtm->rtm_type, &info, (caddr_t)rtm, 411 (struct walkarg *)0); 412 rtm->rtm_flags = rt->rt_flags; 413 rtm->rtm_rmx = rt->rt_rmx; 414 rtm->rtm_addrs = info.rti_addrs; 415 break; 416 417 case RTM_CHANGE: 418 /* new gateway could require new ifaddr, ifp; 419 flags may also be different; ifp may be specified 420 by ll sockaddr when protocol address is ambiguous */ 421 #define equal(a1, a2) (bcmp((caddr_t)(a1), (caddr_t)(a2), (a1)->sa_len) == 0) 422 if ((rt->rt_flags & RTF_GATEWAY && gate != NULL) || 423 ifpaddr != NULL || 424 (ifaaddr != NULL && 425 !equal(ifaaddr, rt->rt_ifa->ifa_addr))) { 426 if ((error = rt_getifa(&info)) != 0) 427 senderr(error); 428 } 429 if (gate != NULL && 430 (error = rt_setgate(rt, rt_key(rt), gate)) != 0) 431 senderr(error); 432 if ((ifa = info.rti_ifa) != NULL) { 433 struct ifaddr *oifa = rt->rt_ifa; 434 if (oifa != ifa) { 435 if (oifa && oifa->ifa_rtrequest) 436 oifa->ifa_rtrequest(RTM_DELETE, rt, 437 &info); 438 IFAFREE(rt->rt_ifa); 439 rt->rt_ifa = ifa; 440 ifa->ifa_refcnt++; 441 rt->rt_ifp = info.rti_ifp; 442 } 443 } 444 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, 445 &rt->rt_rmx); 446 if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest) 447 rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt, &info); 448 if (genmask) 449 rt->rt_genmask = genmask; 450 /* 451 * Fall into 452 */ 453 case RTM_LOCK: 454 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits); 455 rt->rt_rmx.rmx_locks |= 456 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks); 457 break; 458 } 459 break; 460 461 default: 462 senderr(EOPNOTSUPP); 463 } 464 465 flush: 466 if (rtm) { 467 if (error) 468 rtm->rtm_errno = error; 469 else 470 rtm->rtm_flags |= RTF_DONE; 471 } 472 if (rt) 473 rtfree(rt); 474 { 475 struct rawcb *rp = 0; 476 /* 477 * Check to see if we don't want our own messages. 478 */ 479 if ((so->so_options & SO_USELOOPBACK) == 0) { 480 if (route_cb.any_count <= 1) { 481 if (rtm) 482 Free(rtm); 483 m_freem(m); 484 return (error); 485 } 486 /* There is another listener, so construct message */ 487 rp = sotorawcb(so); 488 } 489 if (rtm) { 490 m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm); 491 if (m->m_pkthdr.len < rtm->rtm_msglen) { 492 m_freem(m); 493 m = NULL; 494 } else if (m->m_pkthdr.len > rtm->rtm_msglen) 495 m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len); 496 Free(rtm); 497 } 498 if (rp) 499 rp->rcb_proto.sp_family = 0; /* Avoid us */ 500 if (dst) 501 route_proto.sp_protocol = dst->sa_family; 502 if (m) 503 raw_input(m, &route_proto, &route_src, &route_dst); 504 if (rp) 505 rp->rcb_proto.sp_family = PF_ROUTE; 506 } 507 return (error); 508 } 509 510 static void 511 rt_setmetrics(which, in, out) 512 u_long which; 513 struct rt_metrics *in, *out; 514 { 515 #define metric(f, e) if (which & (f)) out->e = in->e; 516 metric(RTV_RPIPE, rmx_recvpipe); 517 metric(RTV_SPIPE, rmx_sendpipe); 518 metric(RTV_SSTHRESH, rmx_ssthresh); 519 metric(RTV_RTT, rmx_rtt); 520 metric(RTV_RTTVAR, rmx_rttvar); 521 metric(RTV_HOPCOUNT, rmx_hopcount); 522 metric(RTV_MTU, rmx_mtu); 523 metric(RTV_EXPIRE, rmx_expire); 524 #undef metric 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 /* 533 * Extract the addresses of the passed sockaddrs. 534 * Do a little sanity checking so as to avoid bad memory references. 535 * This data is derived straight from userland. 536 */ 537 static int 538 rt_xaddrs(cp, cplim, rtinfo) 539 caddr_t cp, cplim; 540 struct rt_addrinfo *rtinfo; 541 { 542 struct sockaddr *sa; 543 int i; 544 545 for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) { 546 if ((rtinfo->rti_addrs & (1 << i)) == 0) 547 continue; 548 sa = (struct sockaddr *)cp; 549 /* 550 * It won't fit. 551 */ 552 if ( (cp + sa->sa_len) > cplim ) { 553 return (EINVAL); 554 } 555 556 /* 557 * there are no more.. quit now 558 * If there are more bits, they are in error. 559 * I've seen this. route(1) can evidently generate these. 560 * This causes kernel to core dump. 561 * for compatibility, If we see this, point to a safe address. 562 */ 563 if (sa->sa_len == 0) { 564 rtinfo->rti_info[i] = &sa_zero; 565 return (0); /* should be EINVAL but for compat */ 566 } 567 568 /* accept it */ 569 rtinfo->rti_info[i] = sa; 570 ADVANCE(cp, sa); 571 } 572 return (0); 573 } 574 575 static struct mbuf * 576 rt_msg1(type, rtinfo) 577 int type; 578 struct rt_addrinfo *rtinfo; 579 { 580 struct rt_msghdr *rtm; 581 struct mbuf *m; 582 int i; 583 struct sockaddr *sa; 584 int len, dlen; 585 586 switch (type) { 587 588 case RTM_DELADDR: 589 case RTM_NEWADDR: 590 len = sizeof(struct ifa_msghdr); 591 break; 592 593 case RTM_DELMADDR: 594 case RTM_NEWMADDR: 595 len = sizeof(struct ifma_msghdr); 596 break; 597 598 case RTM_IFINFO: 599 len = sizeof(struct if_msghdr); 600 break; 601 602 case RTM_IFANNOUNCE: 603 len = sizeof(struct if_announcemsghdr); 604 break; 605 606 default: 607 len = sizeof(struct rt_msghdr); 608 } 609 if (len > MCLBYTES) 610 panic("rt_msg1"); 611 m = m_gethdr(MB_DONTWAIT, MT_DATA); 612 if (m && len > MHLEN) { 613 MCLGET(m, MB_DONTWAIT); 614 if ((m->m_flags & M_EXT) == 0) { 615 m_free(m); 616 m = NULL; 617 } 618 } 619 if (m == 0) 620 return (m); 621 m->m_pkthdr.len = m->m_len = len; 622 m->m_pkthdr.rcvif = 0; 623 rtm = mtod(m, struct rt_msghdr *); 624 bzero((caddr_t)rtm, len); 625 for (i = 0; i < RTAX_MAX; i++) { 626 if ((sa = rtinfo->rti_info[i]) == NULL) 627 continue; 628 rtinfo->rti_addrs |= (1 << i); 629 dlen = ROUNDUP(sa->sa_len); 630 m_copyback(m, len, dlen, (caddr_t)sa); 631 len += dlen; 632 } 633 if (m->m_pkthdr.len != len) { 634 m_freem(m); 635 return (NULL); 636 } 637 rtm->rtm_msglen = len; 638 rtm->rtm_version = RTM_VERSION; 639 rtm->rtm_type = type; 640 return (m); 641 } 642 643 static int 644 rt_msg2(type, rtinfo, cp, w) 645 int type; 646 struct rt_addrinfo *rtinfo; 647 caddr_t cp; 648 struct walkarg *w; 649 { 650 int i; 651 int len, dlen, second_time = 0; 652 caddr_t cp0; 653 654 rtinfo->rti_addrs = 0; 655 again: 656 switch (type) { 657 658 case RTM_DELADDR: 659 case RTM_NEWADDR: 660 len = sizeof(struct ifa_msghdr); 661 break; 662 663 case RTM_IFINFO: 664 len = sizeof(struct if_msghdr); 665 break; 666 667 default: 668 len = sizeof(struct rt_msghdr); 669 } 670 cp0 = cp; 671 if (cp0) 672 cp += len; 673 for (i = 0; i < RTAX_MAX; i++) { 674 struct sockaddr *sa; 675 676 if ((sa = rtinfo->rti_info[i]) == 0) 677 continue; 678 rtinfo->rti_addrs |= (1 << i); 679 dlen = ROUNDUP(sa->sa_len); 680 if (cp) { 681 bcopy((caddr_t)sa, cp, (unsigned)dlen); 682 cp += dlen; 683 } 684 len += dlen; 685 } 686 len = ALIGN(len); 687 if (cp == 0 && w != NULL && !second_time) { 688 struct walkarg *rw = w; 689 690 if (rw->w_req) { 691 if (rw->w_tmemsize < len) { 692 if (rw->w_tmem) 693 free(rw->w_tmem, M_RTABLE); 694 rw->w_tmem = (caddr_t)malloc(len, M_RTABLE, 695 M_INTWAIT | M_NULLOK); 696 if (rw->w_tmem) 697 rw->w_tmemsize = len; 698 } 699 if (rw->w_tmem) { 700 cp = rw->w_tmem; 701 second_time = 1; 702 goto again; 703 } 704 } 705 } 706 if (cp) { 707 struct rt_msghdr *rtm = (struct rt_msghdr *)cp0; 708 709 rtm->rtm_version = RTM_VERSION; 710 rtm->rtm_type = type; 711 rtm->rtm_msglen = len; 712 } 713 return (len); 714 } 715 716 /* 717 * This routine is called to generate a message from the routing 718 * socket indicating that a redirect has occured, a routing lookup 719 * has failed, or that a protocol has detected timeouts to a particular 720 * destination. 721 */ 722 void 723 rt_missmsg(type, rtinfo, flags, error) 724 int type, flags, error; 725 struct rt_addrinfo *rtinfo; 726 { 727 struct rt_msghdr *rtm; 728 struct mbuf *m; 729 struct sockaddr *sa = rtinfo->rti_info[RTAX_DST]; 730 731 if (route_cb.any_count == 0) 732 return; 733 m = rt_msg1(type, rtinfo); 734 if (m == 0) 735 return; 736 rtm = mtod(m, struct rt_msghdr *); 737 rtm->rtm_flags = RTF_DONE | flags; 738 rtm->rtm_errno = error; 739 rtm->rtm_addrs = rtinfo->rti_addrs; 740 route_proto.sp_protocol = sa ? sa->sa_family : 0; 741 raw_input(m, &route_proto, &route_src, &route_dst); 742 } 743 744 /* 745 * This routine is called to generate a message from the routing 746 * socket indicating that the status of a network interface has changed. 747 */ 748 void 749 rt_ifmsg(ifp) 750 struct ifnet *ifp; 751 { 752 struct if_msghdr *ifm; 753 struct mbuf *m; 754 struct rt_addrinfo info; 755 756 if (route_cb.any_count == 0) 757 return; 758 bzero((caddr_t)&info, sizeof(info)); 759 m = rt_msg1(RTM_IFINFO, &info); 760 if (m == 0) 761 return; 762 ifm = mtod(m, struct if_msghdr *); 763 ifm->ifm_index = ifp->if_index; 764 ifm->ifm_flags = (u_short)ifp->if_flags; 765 ifm->ifm_data = ifp->if_data; 766 ifm->ifm_addrs = 0; 767 route_proto.sp_protocol = 0; 768 raw_input(m, &route_proto, &route_src, &route_dst); 769 } 770 771 static void 772 rt_ifamsg(int cmd, struct ifaddr *ifa) 773 { 774 struct ifa_msghdr *ifam; 775 struct rt_addrinfo info; 776 struct mbuf *m; 777 struct sockaddr *sa; 778 struct ifnet *ifp = ifa->ifa_ifp; 779 780 bzero(&info, sizeof(info)); 781 ifaaddr = sa = ifa->ifa_addr; 782 ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 783 netmask = ifa->ifa_netmask; 784 brdaddr = ifa->ifa_dstaddr; 785 if ((m = rt_msg1(cmd, &info)) == NULL) 786 return; 787 ifam = mtod(m, struct ifa_msghdr *); 788 ifam->ifam_index = ifp->if_index; 789 ifam->ifam_metric = ifa->ifa_metric; 790 ifam->ifam_flags = ifa->ifa_flags; 791 ifam->ifam_addrs = info.rti_addrs; 792 793 route_proto.sp_protocol = sa ? sa->sa_family : 0; 794 raw_input(m, &route_proto, &route_src, &route_dst); 795 } 796 797 static void 798 rt_rtmsg(int cmd, struct ifaddr *ifa, int error, struct rtentry *rt) 799 { 800 struct rt_msghdr *rtm; 801 struct rt_addrinfo info; 802 struct mbuf *m; 803 struct sockaddr *sa; 804 struct ifnet *ifp = ifa->ifa_ifp; 805 806 if (rt == NULL) 807 return; 808 bzero(&info, sizeof(info)); 809 netmask = rt_mask(rt); 810 dst = sa = rt_key(rt); 811 gate = rt->rt_gateway; 812 if ((m = rt_msg1(cmd, &info)) == NULL) 813 return; 814 rtm = mtod(m, struct rt_msghdr *); 815 rtm->rtm_index = ifp->if_index; 816 rtm->rtm_flags |= rt->rt_flags; 817 rtm->rtm_errno = error; 818 rtm->rtm_addrs = info.rti_addrs; 819 820 route_proto.sp_protocol = sa ? sa->sa_family : 0; 821 raw_input(m, &route_proto, &route_src, &route_dst); 822 } 823 824 /* 825 * This is called to generate messages from the routing socket 826 * indicating a network interface has had addresses associated with it. 827 * if we ever reverse the logic and replace messages TO the routing 828 * socket indicate a request to configure interfaces, then it will 829 * be unnecessary as the routing socket will automatically generate 830 * copies of it. 831 */ 832 void 833 rt_newaddrmsg(cmd, ifa, error, rt) 834 int cmd, error; 835 struct ifaddr *ifa; 836 struct rtentry *rt; 837 { 838 if (route_cb.any_count == 0) 839 return; 840 841 if (cmd == RTM_ADD) { 842 rt_ifamsg(RTM_NEWADDR, ifa); 843 rt_rtmsg(RTM_ADD, ifa, error, rt); 844 } else { 845 KASSERT((cmd == RTM_DELETE), ("unknown cmd %d", cmd)); 846 rt_rtmsg(RTM_DELETE, ifa, error, rt); 847 rt_ifamsg(RTM_DELADDR, ifa); 848 } 849 } 850 851 /* 852 * This is the analogue to the rt_newaddrmsg which performs the same 853 * function but for multicast group memberhips. This is easier since 854 * there is no route state to worry about. 855 */ 856 void 857 rt_newmaddrmsg(cmd, ifma) 858 int cmd; 859 struct ifmultiaddr *ifma; 860 { 861 struct rt_addrinfo info; 862 struct mbuf *m = 0; 863 struct ifnet *ifp = ifma->ifma_ifp; 864 struct ifma_msghdr *ifmam; 865 866 if (route_cb.any_count == 0) 867 return; 868 869 bzero((caddr_t)&info, sizeof(info)); 870 ifaaddr = ifma->ifma_addr; 871 if (ifp && TAILQ_FIRST(&ifp->if_addrhead)) 872 ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 873 else 874 ifpaddr = NULL; 875 /* 876 * If a link-layer address is present, present it as a ``gateway'' 877 * (similarly to how ARP entries, e.g., are presented). 878 */ 879 gate = ifma->ifma_lladdr; 880 if ((m = rt_msg1(cmd, &info)) == NULL) 881 return; 882 ifmam = mtod(m, struct ifma_msghdr *); 883 ifmam->ifmam_index = ifp->if_index; 884 ifmam->ifmam_addrs = info.rti_addrs; 885 route_proto.sp_protocol = ifma->ifma_addr->sa_family; 886 raw_input(m, &route_proto, &route_src, &route_dst); 887 } 888 889 /* 890 * This is called to generate routing socket messages indicating 891 * network interface arrival and departure. 892 */ 893 void 894 rt_ifannouncemsg(ifp, what) 895 struct ifnet *ifp; 896 int what; 897 { 898 struct if_announcemsghdr *ifan; 899 struct mbuf *m; 900 struct rt_addrinfo info; 901 902 if (route_cb.any_count == 0) 903 return; 904 bzero((caddr_t)&info, sizeof(info)); 905 m = rt_msg1(RTM_IFANNOUNCE, &info); 906 if (m == NULL) 907 return; 908 ifan = mtod(m, struct if_announcemsghdr *); 909 ifan->ifan_index = ifp->if_index; 910 strlcpy(ifan->ifan_name, ifp->if_xname, sizeof(ifan->ifan_name)); 911 ifan->ifan_what = what; 912 route_proto.sp_protocol = 0; 913 raw_input(m, &route_proto, &route_src, &route_dst); 914 } 915 916 /* 917 * This is used in dumping the kernel table via sysctl(). 918 */ 919 int 920 sysctl_dumpentry(rn, vw) 921 struct radix_node *rn; 922 void *vw; 923 { 924 struct walkarg *w = vw; 925 struct rtentry *rt = (struct rtentry *)rn; 926 int error = 0, size; 927 struct rt_addrinfo info; 928 929 if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg)) 930 return 0; 931 bzero((caddr_t)&info, sizeof(info)); 932 dst = rt_key(rt); 933 gate = rt->rt_gateway; 934 netmask = rt_mask(rt); 935 genmask = rt->rt_genmask; 936 if (rt->rt_ifp) { 937 ifpaddr = TAILQ_FIRST(&rt->rt_ifp->if_addrhead)->ifa_addr; 938 ifaaddr = rt->rt_ifa->ifa_addr; 939 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT) 940 brdaddr = rt->rt_ifa->ifa_dstaddr; 941 } 942 size = rt_msg2(RTM_GET, &info, 0, w); 943 if (w->w_req && w->w_tmem) { 944 struct rt_msghdr *rtm = (struct rt_msghdr *)w->w_tmem; 945 946 rtm->rtm_flags = rt->rt_flags; 947 rtm->rtm_use = rt->rt_use; 948 rtm->rtm_rmx = rt->rt_rmx; 949 rtm->rtm_index = rt->rt_ifp->if_index; 950 rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0; 951 rtm->rtm_addrs = info.rti_addrs; 952 error = SYSCTL_OUT(w->w_req, (caddr_t)rtm, size); 953 return (error); 954 } 955 return (error); 956 } 957 958 int 959 sysctl_iflist(af, w) 960 int af; 961 struct walkarg *w; 962 { 963 struct ifnet *ifp; 964 struct ifaddr *ifa; 965 struct rt_addrinfo info; 966 int len, error = 0; 967 968 bzero((caddr_t)&info, sizeof(info)); 969 TAILQ_FOREACH(ifp, &ifnet, if_link) { 970 if (w->w_arg && w->w_arg != ifp->if_index) 971 continue; 972 ifa = TAILQ_FIRST(&ifp->if_addrhead); 973 ifpaddr = ifa->ifa_addr; 974 len = rt_msg2(RTM_IFINFO, &info, (caddr_t)0, w); 975 ifpaddr = 0; 976 if (w->w_req && w->w_tmem) { 977 struct if_msghdr *ifm; 978 979 ifm = (struct if_msghdr *)w->w_tmem; 980 ifm->ifm_index = ifp->if_index; 981 ifm->ifm_flags = (u_short)ifp->if_flags; 982 ifm->ifm_data = ifp->if_data; 983 ifm->ifm_addrs = info.rti_addrs; 984 error = SYSCTL_OUT(w->w_req,(caddr_t)ifm, len); 985 if (error) 986 return (error); 987 } 988 while ((ifa = TAILQ_NEXT(ifa, ifa_link)) != 0) { 989 if (af && af != ifa->ifa_addr->sa_family) 990 continue; 991 if (curproc->p_ucred->cr_prison && prison_if(curthread, ifa->ifa_addr)) 992 continue; 993 ifaaddr = ifa->ifa_addr; 994 netmask = ifa->ifa_netmask; 995 brdaddr = ifa->ifa_dstaddr; 996 len = rt_msg2(RTM_NEWADDR, &info, 0, w); 997 if (w->w_req && w->w_tmem) { 998 struct ifa_msghdr *ifam; 999 1000 ifam = (struct ifa_msghdr *)w->w_tmem; 1001 ifam->ifam_index = ifa->ifa_ifp->if_index; 1002 ifam->ifam_flags = ifa->ifa_flags; 1003 ifam->ifam_metric = ifa->ifa_metric; 1004 ifam->ifam_addrs = info.rti_addrs; 1005 error = SYSCTL_OUT(w->w_req, w->w_tmem, len); 1006 if (error) 1007 return (error); 1008 } 1009 } 1010 ifaaddr = netmask = brdaddr = 0; 1011 } 1012 return (0); 1013 } 1014 1015 static int 1016 sysctl_rtsock(SYSCTL_HANDLER_ARGS) 1017 { 1018 int *name = (int *)arg1; 1019 u_int namelen = arg2; 1020 struct radix_node_head *rnh; 1021 int i, s, error = EINVAL; 1022 u_char af; 1023 struct walkarg w; 1024 1025 name ++; 1026 namelen--; 1027 if (req->newptr) 1028 return (EPERM); 1029 if (namelen != 3) 1030 return (EINVAL); 1031 af = name[0]; 1032 Bzero(&w, sizeof(w)); 1033 w.w_op = name[1]; 1034 w.w_arg = name[2]; 1035 w.w_req = req; 1036 1037 s = splnet(); 1038 switch (w.w_op) { 1039 1040 case NET_RT_DUMP: 1041 case NET_RT_FLAGS: 1042 for (i = 1; i <= AF_MAX; i++) 1043 if ((rnh = rt_tables[i]) && (af == 0 || af == i) && 1044 (error = rnh->rnh_walktree(rnh, 1045 sysctl_dumpentry, &w))) 1046 break; 1047 break; 1048 1049 case NET_RT_IFLIST: 1050 error = sysctl_iflist(af, &w); 1051 } 1052 splx(s); 1053 if (w.w_tmem) 1054 free(w.w_tmem, M_RTABLE); 1055 return (error); 1056 } 1057 1058 SYSCTL_NODE(_net, PF_ROUTE, routetable, CTLFLAG_RD, sysctl_rtsock, ""); 1059 1060 /* 1061 * Definitions of protocols supported in the ROUTE domain. 1062 */ 1063 1064 extern struct domain routedomain; /* or at least forward */ 1065 1066 static struct protosw routesw[] = { 1067 { SOCK_RAW, &routedomain, 0, PR_ATOMIC|PR_ADDR, 1068 0, route_output, raw_ctlinput, 0, 1069 cpu0_soport, 1070 raw_init, 0, 0, 0, 1071 &route_usrreqs 1072 } 1073 }; 1074 1075 static struct domain routedomain = 1076 { PF_ROUTE, "route", 0, 0, 0, 1077 routesw, &routesw[sizeof(routesw)/sizeof(routesw[0])] }; 1078 1079 DOMAIN_SET(route); 1080