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 1.11 2004/03/06 05:20:31 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 <net/if.h> 52 #include <net/route.h> 53 #include <net/raw_cb.h> 54 55 MALLOC_DEFINE(M_RTABLE, "routetbl", "routing tables"); 56 57 static struct sockaddr route_dst = { 2, PF_ROUTE, }; 58 static struct sockaddr route_src = { 2, PF_ROUTE, }; 59 static struct sockaddr sa_zero = { sizeof(sa_zero), AF_INET, }; 60 static struct sockproto route_proto = { PF_ROUTE, }; 61 62 struct walkarg { 63 int w_tmemsize; 64 int w_op, w_arg; 65 caddr_t w_tmem; 66 struct sysctl_req *w_req; 67 }; 68 69 static struct mbuf * 70 rt_msg1 (int, struct rt_addrinfo *); 71 static int rt_msg2 (int, struct rt_addrinfo *, caddr_t, struct walkarg *); 72 static int rt_xaddrs (caddr_t, caddr_t, struct rt_addrinfo *); 73 static int sysctl_dumpentry (struct radix_node *rn, void *vw); 74 static int sysctl_iflist (int af, struct walkarg *w); 75 static int route_output (struct mbuf *, struct socket *, 76 struct pr_output_info *); 77 static void rt_setmetrics (u_long, struct rt_metrics *, 78 struct rt_metrics *); 79 80 /* Sleazy use of local variables throughout file, warning!!!! */ 81 #define dst info.rti_info[RTAX_DST] 82 #define gate info.rti_info[RTAX_GATEWAY] 83 #define netmask info.rti_info[RTAX_NETMASK] 84 #define genmask info.rti_info[RTAX_GENMASK] 85 #define ifpaddr info.rti_info[RTAX_IFP] 86 #define ifaaddr info.rti_info[RTAX_IFA] 87 #define brdaddr info.rti_info[RTAX_BRD] 88 89 /* 90 * It really doesn't make any sense at all for this code to share much 91 * with raw_usrreq.c, since its functionality is so restricted. XXX 92 */ 93 static int 94 rts_abort(struct socket *so) 95 { 96 int s, error; 97 s = splnet(); 98 error = raw_usrreqs.pru_abort(so); 99 splx(s); 100 return error; 101 } 102 103 /* pru_accept is EOPNOTSUPP */ 104 105 static int 106 rts_attach(struct socket *so, int proto, struct pru_attach_info *ai) 107 { 108 struct rawcb *rp; 109 int s, error; 110 111 if (sotorawcb(so) != 0) 112 return EISCONN; /* XXX panic? */ 113 MALLOC(rp, struct rawcb *, sizeof *rp, M_PCB, M_WAITOK|M_ZERO); 114 if (rp == 0) 115 return ENOBUFS; 116 117 /* 118 * The splnet() is necessary to block protocols from sending 119 * error notifications (like RTM_REDIRECT or RTM_LOSING) while 120 * this PCB is extant but incompletely initialized. 121 * Probably we should try to do more of this work beforehand and 122 * eliminate the spl. 123 */ 124 s = splnet(); 125 so->so_pcb = (caddr_t)rp; 126 error = raw_attach(so, proto, ai->sb_rlimit); 127 rp = sotorawcb(so); 128 if (error) { 129 splx(s); 130 free(rp, M_PCB); 131 return error; 132 } 133 switch(rp->rcb_proto.sp_protocol) { 134 case AF_INET: 135 route_cb.ip_count++; 136 break; 137 case AF_INET6: 138 route_cb.ip6_count++; 139 break; 140 case AF_IPX: 141 route_cb.ipx_count++; 142 break; 143 case AF_NS: 144 route_cb.ns_count++; 145 break; 146 } 147 rp->rcb_faddr = &route_src; 148 route_cb.any_count++; 149 soisconnected(so); 150 so->so_options |= SO_USELOOPBACK; 151 splx(s); 152 return 0; 153 } 154 155 static int 156 rts_bind(struct socket *so, struct sockaddr *nam, struct thread *td) 157 { 158 int s, error; 159 s = splnet(); 160 error = raw_usrreqs.pru_bind(so, nam, td); /* xxx just EINVAL */ 161 splx(s); 162 return error; 163 } 164 165 static int 166 rts_connect(struct socket *so, struct sockaddr *nam, struct thread *td) 167 { 168 int s, error; 169 s = splnet(); 170 error = raw_usrreqs.pru_connect(so, nam, td); /* XXX just EINVAL */ 171 splx(s); 172 return error; 173 } 174 175 /* pru_connect2 is EOPNOTSUPP */ 176 /* pru_control is EOPNOTSUPP */ 177 178 static int 179 rts_detach(struct socket *so) 180 { 181 struct rawcb *rp = sotorawcb(so); 182 int s, error; 183 184 s = splnet(); 185 if (rp != 0) { 186 switch(rp->rcb_proto.sp_protocol) { 187 case AF_INET: 188 route_cb.ip_count--; 189 break; 190 case AF_INET6: 191 route_cb.ip6_count--; 192 break; 193 case AF_IPX: 194 route_cb.ipx_count--; 195 break; 196 case AF_NS: 197 route_cb.ns_count--; 198 break; 199 } 200 route_cb.any_count--; 201 } 202 error = raw_usrreqs.pru_detach(so); 203 splx(s); 204 return error; 205 } 206 207 static int 208 rts_disconnect(struct socket *so) 209 { 210 int s, error; 211 s = splnet(); 212 error = raw_usrreqs.pru_disconnect(so); 213 splx(s); 214 return error; 215 } 216 217 /* pru_listen is EOPNOTSUPP */ 218 219 static int 220 rts_peeraddr(struct socket *so, struct sockaddr **nam) 221 { 222 int s, error; 223 s = splnet(); 224 error = raw_usrreqs.pru_peeraddr(so, nam); 225 splx(s); 226 return error; 227 } 228 229 /* pru_rcvd is EOPNOTSUPP */ 230 /* pru_rcvoob is EOPNOTSUPP */ 231 232 static int 233 rts_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam, 234 struct mbuf *control, struct thread *td) 235 { 236 int s, error; 237 s = splnet(); 238 error = raw_usrreqs.pru_send(so, flags, m, nam, control, td); 239 splx(s); 240 return error; 241 } 242 243 /* pru_sense is null */ 244 245 static int 246 rts_shutdown(struct socket *so) 247 { 248 int s, error; 249 s = splnet(); 250 error = raw_usrreqs.pru_shutdown(so); 251 splx(s); 252 return error; 253 } 254 255 static int 256 rts_sockaddr(struct socket *so, struct sockaddr **nam) 257 { 258 int s, error; 259 s = splnet(); 260 error = raw_usrreqs.pru_sockaddr(so, nam); 261 splx(s); 262 return error; 263 } 264 265 static struct pr_usrreqs route_usrreqs = { 266 rts_abort, pru_accept_notsupp, rts_attach, rts_bind, rts_connect, 267 pru_connect2_notsupp, pru_control_notsupp, rts_detach, rts_disconnect, 268 pru_listen_notsupp, rts_peeraddr, pru_rcvd_notsupp, pru_rcvoob_notsupp, 269 rts_send, pru_sense_null, rts_shutdown, rts_sockaddr, 270 sosend, soreceive, sopoll 271 }; 272 273 /*ARGSUSED*/ 274 static int 275 route_output(struct mbuf *m, struct socket *so, struct pr_output_info *oi) 276 { 277 struct rt_msghdr *rtm = 0; 278 struct rtentry *rt = 0; 279 struct rtentry *saved_nrt = 0; 280 struct radix_node_head *rnh; 281 struct rt_addrinfo info; 282 int len, error = 0; 283 struct ifnet *ifp = 0; 284 struct ifaddr *ifa = 0; 285 286 #define senderr(e) { error = e; goto flush;} 287 if (m == 0 || ((m->m_len < sizeof(long)) && 288 (m = m_pullup(m, sizeof(long))) == 0)) 289 return (ENOBUFS); 290 if ((m->m_flags & M_PKTHDR) == 0) 291 panic("route_output"); 292 len = m->m_pkthdr.len; 293 if (len < sizeof(*rtm) || 294 len != mtod(m, struct rt_msghdr *)->rtm_msglen) { 295 dst = 0; 296 senderr(EINVAL); 297 } 298 R_Malloc(rtm, struct rt_msghdr *, len); 299 if (rtm == 0) { 300 dst = 0; 301 senderr(ENOBUFS); 302 } 303 m_copydata(m, 0, len, (caddr_t)rtm); 304 if (rtm->rtm_version != RTM_VERSION) { 305 dst = 0; 306 senderr(EPROTONOSUPPORT); 307 } 308 rtm->rtm_pid = oi->p_pid; 309 bzero(&info, sizeof(info)); 310 info.rti_addrs = rtm->rtm_addrs; 311 if (rt_xaddrs((caddr_t)(rtm + 1), len + (caddr_t)rtm, &info)) { 312 dst = 0; 313 senderr(EINVAL); 314 } 315 info.rti_flags = rtm->rtm_flags; 316 if (dst == 0 || (dst->sa_family >= AF_MAX) 317 || (gate != 0 && (gate->sa_family >= AF_MAX))) 318 senderr(EINVAL); 319 if (genmask) { 320 struct radix_node *t; 321 t = rn_addmask((caddr_t)genmask, 0, 1); 322 if (t && Bcmp((caddr_t *)genmask + 1, (caddr_t *)t->rn_key + 1, 323 *(u_char *)t->rn_key - 1) == 0) 324 genmask = (struct sockaddr *)(t->rn_key); 325 else 326 senderr(ENOBUFS); 327 } 328 329 /* 330 * Verify that the caller has the appropriate privilege; RTM_GET 331 * is the only operation the non-superuser is allowed. 332 */ 333 if (rtm->rtm_type != RTM_GET && suser_cred(so->so_cred, 0) != 0) 334 senderr(EPERM); 335 336 switch (rtm->rtm_type) { 337 338 case RTM_ADD: 339 if (gate == 0) 340 senderr(EINVAL); 341 error = rtrequest1(RTM_ADD, &info, &saved_nrt); 342 if (error == 0 && saved_nrt) { 343 rt_setmetrics(rtm->rtm_inits, 344 &rtm->rtm_rmx, &saved_nrt->rt_rmx); 345 saved_nrt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits); 346 saved_nrt->rt_rmx.rmx_locks |= 347 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks); 348 saved_nrt->rt_refcnt--; 349 saved_nrt->rt_genmask = genmask; 350 } 351 break; 352 353 case RTM_DELETE: 354 error = rtrequest1(RTM_DELETE, &info, &saved_nrt); 355 if (error == 0) { 356 if ((rt = saved_nrt)) 357 rt->rt_refcnt++; 358 goto report; 359 } 360 break; 361 362 case RTM_GET: 363 case RTM_CHANGE: 364 case RTM_LOCK: 365 if ((rnh = rt_tables[dst->sa_family]) == 0) { 366 senderr(EAFNOSUPPORT); 367 } else if ((rt = (struct rtentry *) 368 rnh->rnh_lookup(dst, netmask, rnh)) != NULL) 369 rt->rt_refcnt++; 370 else 371 senderr(ESRCH); 372 switch(rtm->rtm_type) { 373 374 case RTM_GET: 375 report: 376 dst = rt_key(rt); 377 gate = rt->rt_gateway; 378 netmask = rt_mask(rt); 379 genmask = rt->rt_genmask; 380 if (rtm->rtm_addrs & (RTA_IFP | RTA_IFA)) { 381 ifp = rt->rt_ifp; 382 if (ifp) { 383 ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 384 ifaaddr = rt->rt_ifa->ifa_addr; 385 if (ifp->if_flags & IFF_POINTOPOINT) 386 brdaddr = rt->rt_ifa->ifa_dstaddr; 387 rtm->rtm_index = ifp->if_index; 388 } else { 389 ifpaddr = 0; 390 ifaaddr = 0; 391 } 392 } 393 len = rt_msg2(rtm->rtm_type, &info, (caddr_t)0, 394 (struct walkarg *)0); 395 if (len > rtm->rtm_msglen) { 396 struct rt_msghdr *new_rtm; 397 R_Malloc(new_rtm, struct rt_msghdr *, len); 398 if (new_rtm == 0) 399 senderr(ENOBUFS); 400 Bcopy(rtm, new_rtm, rtm->rtm_msglen); 401 Free(rtm); rtm = new_rtm; 402 } 403 (void)rt_msg2(rtm->rtm_type, &info, (caddr_t)rtm, 404 (struct walkarg *)0); 405 rtm->rtm_flags = rt->rt_flags; 406 rtm->rtm_rmx = rt->rt_rmx; 407 rtm->rtm_addrs = info.rti_addrs; 408 break; 409 410 case RTM_CHANGE: 411 /* new gateway could require new ifaddr, ifp; 412 flags may also be different; ifp may be specified 413 by ll sockaddr when protocol address is ambiguous */ 414 #define equal(a1, a2) (bcmp((caddr_t)(a1), (caddr_t)(a2), (a1)->sa_len) == 0) 415 if ((rt->rt_flags & RTF_GATEWAY && gate != NULL) || 416 ifpaddr != NULL || 417 (ifaaddr != NULL && 418 !equal(ifaaddr, rt->rt_ifa->ifa_addr))) { 419 if ((error = rt_getifa(&info)) != 0) 420 senderr(error); 421 } 422 if (gate != NULL && 423 (error = rt_setgate(rt, rt_key(rt), gate)) != 0) 424 senderr(error); 425 if ((ifa = info.rti_ifa) != NULL) { 426 struct ifaddr *oifa = rt->rt_ifa; 427 if (oifa != ifa) { 428 if (oifa && oifa->ifa_rtrequest) 429 oifa->ifa_rtrequest(RTM_DELETE, rt, 430 &info); 431 IFAFREE(rt->rt_ifa); 432 rt->rt_ifa = ifa; 433 ifa->ifa_refcnt++; 434 rt->rt_ifp = info.rti_ifp; 435 } 436 } 437 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, 438 &rt->rt_rmx); 439 if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest) 440 rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt, &info); 441 if (genmask) 442 rt->rt_genmask = genmask; 443 /* 444 * Fall into 445 */ 446 case RTM_LOCK: 447 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits); 448 rt->rt_rmx.rmx_locks |= 449 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks); 450 break; 451 } 452 break; 453 454 default: 455 senderr(EOPNOTSUPP); 456 } 457 458 flush: 459 if (rtm) { 460 if (error) 461 rtm->rtm_errno = error; 462 else 463 rtm->rtm_flags |= RTF_DONE; 464 } 465 if (rt) 466 rtfree(rt); 467 { 468 struct rawcb *rp = 0; 469 /* 470 * Check to see if we don't want our own messages. 471 */ 472 if ((so->so_options & SO_USELOOPBACK) == 0) { 473 if (route_cb.any_count <= 1) { 474 if (rtm) 475 Free(rtm); 476 m_freem(m); 477 return (error); 478 } 479 /* There is another listener, so construct message */ 480 rp = sotorawcb(so); 481 } 482 if (rtm) { 483 m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm); 484 if (m->m_pkthdr.len < rtm->rtm_msglen) { 485 m_freem(m); 486 m = NULL; 487 } else if (m->m_pkthdr.len > rtm->rtm_msglen) 488 m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len); 489 Free(rtm); 490 } 491 if (rp) 492 rp->rcb_proto.sp_family = 0; /* Avoid us */ 493 if (dst) 494 route_proto.sp_protocol = dst->sa_family; 495 if (m) 496 raw_input(m, &route_proto, &route_src, &route_dst); 497 if (rp) 498 rp->rcb_proto.sp_family = PF_ROUTE; 499 } 500 return (error); 501 } 502 503 static void 504 rt_setmetrics(which, in, out) 505 u_long which; 506 struct rt_metrics *in, *out; 507 { 508 #define metric(f, e) if (which & (f)) out->e = in->e; 509 metric(RTV_RPIPE, rmx_recvpipe); 510 metric(RTV_SPIPE, rmx_sendpipe); 511 metric(RTV_SSTHRESH, rmx_ssthresh); 512 metric(RTV_RTT, rmx_rtt); 513 metric(RTV_RTTVAR, rmx_rttvar); 514 metric(RTV_HOPCOUNT, rmx_hopcount); 515 metric(RTV_MTU, rmx_mtu); 516 metric(RTV_EXPIRE, rmx_expire); 517 #undef metric 518 } 519 520 #define ROUNDUP(a) \ 521 ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long)) 522 #define ADVANCE(x, n) (x += ROUNDUP((n)->sa_len)) 523 524 525 /* 526 * Extract the addresses of the passed sockaddrs. 527 * Do a little sanity checking so as to avoid bad memory references. 528 * This data is derived straight from userland. 529 */ 530 static int 531 rt_xaddrs(cp, cplim, rtinfo) 532 caddr_t cp, cplim; 533 struct rt_addrinfo *rtinfo; 534 { 535 struct sockaddr *sa; 536 int i; 537 538 for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) { 539 if ((rtinfo->rti_addrs & (1 << i)) == 0) 540 continue; 541 sa = (struct sockaddr *)cp; 542 /* 543 * It won't fit. 544 */ 545 if ( (cp + sa->sa_len) > cplim ) { 546 return (EINVAL); 547 } 548 549 /* 550 * there are no more.. quit now 551 * If there are more bits, they are in error. 552 * I've seen this. route(1) can evidently generate these. 553 * This causes kernel to core dump. 554 * for compatibility, If we see this, point to a safe address. 555 */ 556 if (sa->sa_len == 0) { 557 rtinfo->rti_info[i] = &sa_zero; 558 return (0); /* should be EINVAL but for compat */ 559 } 560 561 /* accept it */ 562 rtinfo->rti_info[i] = sa; 563 ADVANCE(cp, sa); 564 } 565 return (0); 566 } 567 568 static struct mbuf * 569 rt_msg1(type, rtinfo) 570 int type; 571 struct rt_addrinfo *rtinfo; 572 { 573 struct rt_msghdr *rtm; 574 struct mbuf *m; 575 int i; 576 struct sockaddr *sa; 577 int len, dlen; 578 579 switch (type) { 580 581 case RTM_DELADDR: 582 case RTM_NEWADDR: 583 len = sizeof(struct ifa_msghdr); 584 break; 585 586 case RTM_DELMADDR: 587 case RTM_NEWMADDR: 588 len = sizeof(struct ifma_msghdr); 589 break; 590 591 case RTM_IFINFO: 592 len = sizeof(struct if_msghdr); 593 break; 594 595 case RTM_IFANNOUNCE: 596 len = sizeof(struct if_announcemsghdr); 597 break; 598 599 default: 600 len = sizeof(struct rt_msghdr); 601 } 602 if (len > MCLBYTES) 603 panic("rt_msg1"); 604 m = m_gethdr(M_DONTWAIT, MT_DATA); 605 if (m && len > MHLEN) { 606 MCLGET(m, M_DONTWAIT); 607 if ((m->m_flags & M_EXT) == 0) { 608 m_free(m); 609 m = NULL; 610 } 611 } 612 if (m == 0) 613 return (m); 614 m->m_pkthdr.len = m->m_len = len; 615 m->m_pkthdr.rcvif = 0; 616 rtm = mtod(m, struct rt_msghdr *); 617 bzero((caddr_t)rtm, len); 618 for (i = 0; i < RTAX_MAX; i++) { 619 if ((sa = rtinfo->rti_info[i]) == NULL) 620 continue; 621 rtinfo->rti_addrs |= (1 << i); 622 dlen = ROUNDUP(sa->sa_len); 623 m_copyback(m, len, dlen, (caddr_t)sa); 624 len += dlen; 625 } 626 if (m->m_pkthdr.len != len) { 627 m_freem(m); 628 return (NULL); 629 } 630 rtm->rtm_msglen = len; 631 rtm->rtm_version = RTM_VERSION; 632 rtm->rtm_type = type; 633 return (m); 634 } 635 636 static int 637 rt_msg2(type, rtinfo, cp, w) 638 int type; 639 struct rt_addrinfo *rtinfo; 640 caddr_t cp; 641 struct walkarg *w; 642 { 643 int i; 644 int len, dlen, second_time = 0; 645 caddr_t cp0; 646 647 rtinfo->rti_addrs = 0; 648 again: 649 switch (type) { 650 651 case RTM_DELADDR: 652 case RTM_NEWADDR: 653 len = sizeof(struct ifa_msghdr); 654 break; 655 656 case RTM_IFINFO: 657 len = sizeof(struct if_msghdr); 658 break; 659 660 default: 661 len = sizeof(struct rt_msghdr); 662 } 663 cp0 = cp; 664 if (cp0) 665 cp += len; 666 for (i = 0; i < RTAX_MAX; i++) { 667 struct sockaddr *sa; 668 669 if ((sa = rtinfo->rti_info[i]) == 0) 670 continue; 671 rtinfo->rti_addrs |= (1 << i); 672 dlen = ROUNDUP(sa->sa_len); 673 if (cp) { 674 bcopy((caddr_t)sa, cp, (unsigned)dlen); 675 cp += dlen; 676 } 677 len += dlen; 678 } 679 len = ALIGN(len); 680 if (cp == 0 && w != NULL && !second_time) { 681 struct walkarg *rw = w; 682 683 if (rw->w_req) { 684 if (rw->w_tmemsize < len) { 685 if (rw->w_tmem) 686 free(rw->w_tmem, M_RTABLE); 687 rw->w_tmem = (caddr_t) 688 malloc(len, M_RTABLE, M_NOWAIT); 689 if (rw->w_tmem) 690 rw->w_tmemsize = len; 691 } 692 if (rw->w_tmem) { 693 cp = rw->w_tmem; 694 second_time = 1; 695 goto again; 696 } 697 } 698 } 699 if (cp) { 700 struct rt_msghdr *rtm = (struct rt_msghdr *)cp0; 701 702 rtm->rtm_version = RTM_VERSION; 703 rtm->rtm_type = type; 704 rtm->rtm_msglen = len; 705 } 706 return (len); 707 } 708 709 /* 710 * This routine is called to generate a message from the routing 711 * socket indicating that a redirect has occured, a routing lookup 712 * has failed, or that a protocol has detected timeouts to a particular 713 * destination. 714 */ 715 void 716 rt_missmsg(type, rtinfo, flags, error) 717 int type, flags, error; 718 struct rt_addrinfo *rtinfo; 719 { 720 struct rt_msghdr *rtm; 721 struct mbuf *m; 722 struct sockaddr *sa = rtinfo->rti_info[RTAX_DST]; 723 724 if (route_cb.any_count == 0) 725 return; 726 m = rt_msg1(type, rtinfo); 727 if (m == 0) 728 return; 729 rtm = mtod(m, struct rt_msghdr *); 730 rtm->rtm_flags = RTF_DONE | flags; 731 rtm->rtm_errno = error; 732 rtm->rtm_addrs = rtinfo->rti_addrs; 733 route_proto.sp_protocol = sa ? sa->sa_family : 0; 734 raw_input(m, &route_proto, &route_src, &route_dst); 735 } 736 737 /* 738 * This routine is called to generate a message from the routing 739 * socket indicating that the status of a network interface has changed. 740 */ 741 void 742 rt_ifmsg(ifp) 743 struct ifnet *ifp; 744 { 745 struct if_msghdr *ifm; 746 struct mbuf *m; 747 struct rt_addrinfo info; 748 749 if (route_cb.any_count == 0) 750 return; 751 bzero((caddr_t)&info, sizeof(info)); 752 m = rt_msg1(RTM_IFINFO, &info); 753 if (m == 0) 754 return; 755 ifm = mtod(m, struct if_msghdr *); 756 ifm->ifm_index = ifp->if_index; 757 ifm->ifm_flags = (u_short)ifp->if_flags; 758 ifm->ifm_data = ifp->if_data; 759 ifm->ifm_addrs = 0; 760 route_proto.sp_protocol = 0; 761 raw_input(m, &route_proto, &route_src, &route_dst); 762 } 763 764 static void 765 rt_ifamsg(int cmd, struct ifaddr *ifa) 766 { 767 struct ifa_msghdr *ifam; 768 struct rt_addrinfo info; 769 struct mbuf *m; 770 struct sockaddr *sa; 771 struct ifnet *ifp = ifa->ifa_ifp; 772 773 bzero(&info, sizeof(info)); 774 ifaaddr = sa = ifa->ifa_addr; 775 ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 776 netmask = ifa->ifa_netmask; 777 brdaddr = ifa->ifa_dstaddr; 778 if ((m = rt_msg1(cmd, &info)) == NULL) 779 return; 780 ifam = mtod(m, struct ifa_msghdr *); 781 ifam->ifam_index = ifp->if_index; 782 ifam->ifam_metric = ifa->ifa_metric; 783 ifam->ifam_flags = ifa->ifa_flags; 784 ifam->ifam_addrs = info.rti_addrs; 785 786 route_proto.sp_protocol = sa ? sa->sa_family : 0; 787 raw_input(m, &route_proto, &route_src, &route_dst); 788 } 789 790 static void 791 rt_rtmsg(int cmd, struct ifaddr *ifa, int error, struct rtentry *rt) 792 { 793 struct rt_msghdr *rtm; 794 struct rt_addrinfo info; 795 struct mbuf *m; 796 struct sockaddr *sa; 797 struct ifnet *ifp = ifa->ifa_ifp; 798 799 if (rt == NULL) 800 return; 801 bzero(&info, sizeof(info)); 802 netmask = rt_mask(rt); 803 dst = sa = rt_key(rt); 804 gate = rt->rt_gateway; 805 if ((m = rt_msg1(cmd, &info)) == NULL) 806 return; 807 rtm = mtod(m, struct rt_msghdr *); 808 rtm->rtm_index = ifp->if_index; 809 rtm->rtm_flags |= rt->rt_flags; 810 rtm->rtm_errno = error; 811 rtm->rtm_addrs = info.rti_addrs; 812 813 route_proto.sp_protocol = sa ? sa->sa_family : 0; 814 raw_input(m, &route_proto, &route_src, &route_dst); 815 } 816 817 /* 818 * This is called to generate messages from the routing socket 819 * indicating a network interface has had addresses associated with it. 820 * if we ever reverse the logic and replace messages TO the routing 821 * socket indicate a request to configure interfaces, then it will 822 * be unnecessary as the routing socket will automatically generate 823 * copies of it. 824 */ 825 void 826 rt_newaddrmsg(cmd, ifa, error, rt) 827 int cmd, error; 828 struct ifaddr *ifa; 829 struct rtentry *rt; 830 { 831 if (route_cb.any_count == 0) 832 return; 833 834 if (cmd == RTM_ADD) { 835 rt_ifamsg(RTM_NEWADDR, ifa); 836 rt_rtmsg(RTM_ADD, ifa, error, rt); 837 } else { 838 KASSERT((cmd == RTM_DELETE), ("unknown cmd %d", cmd)); 839 rt_rtmsg(RTM_DELETE, ifa, error, rt); 840 rt_ifamsg(RTM_DELADDR, ifa); 841 } 842 } 843 844 /* 845 * This is the analogue to the rt_newaddrmsg which performs the same 846 * function but for multicast group memberhips. This is easier since 847 * there is no route state to worry about. 848 */ 849 void 850 rt_newmaddrmsg(cmd, ifma) 851 int cmd; 852 struct ifmultiaddr *ifma; 853 { 854 struct rt_addrinfo info; 855 struct mbuf *m = 0; 856 struct ifnet *ifp = ifma->ifma_ifp; 857 struct ifma_msghdr *ifmam; 858 859 if (route_cb.any_count == 0) 860 return; 861 862 bzero((caddr_t)&info, sizeof(info)); 863 ifaaddr = ifma->ifma_addr; 864 if (ifp && TAILQ_FIRST(&ifp->if_addrhead)) 865 ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 866 else 867 ifpaddr = NULL; 868 /* 869 * If a link-layer address is present, present it as a ``gateway'' 870 * (similarly to how ARP entries, e.g., are presented). 871 */ 872 gate = ifma->ifma_lladdr; 873 if ((m = rt_msg1(cmd, &info)) == NULL) 874 return; 875 ifmam = mtod(m, struct ifma_msghdr *); 876 ifmam->ifmam_index = ifp->if_index; 877 ifmam->ifmam_addrs = info.rti_addrs; 878 route_proto.sp_protocol = ifma->ifma_addr->sa_family; 879 raw_input(m, &route_proto, &route_src, &route_dst); 880 } 881 882 /* 883 * This is called to generate routing socket messages indicating 884 * network interface arrival and departure. 885 */ 886 void 887 rt_ifannouncemsg(ifp, what) 888 struct ifnet *ifp; 889 int what; 890 { 891 struct if_announcemsghdr *ifan; 892 struct mbuf *m; 893 struct rt_addrinfo info; 894 895 if (route_cb.any_count == 0) 896 return; 897 bzero((caddr_t)&info, sizeof(info)); 898 m = rt_msg1(RTM_IFANNOUNCE, &info); 899 if (m == NULL) 900 return; 901 ifan = mtod(m, struct if_announcemsghdr *); 902 ifan->ifan_index = ifp->if_index; 903 strlcpy(ifan->ifan_name, ifp->if_xname, sizeof(ifan->ifan_name)); 904 ifan->ifan_what = what; 905 route_proto.sp_protocol = 0; 906 raw_input(m, &route_proto, &route_src, &route_dst); 907 } 908 909 /* 910 * This is used in dumping the kernel table via sysctl(). 911 */ 912 int 913 sysctl_dumpentry(rn, vw) 914 struct radix_node *rn; 915 void *vw; 916 { 917 struct walkarg *w = vw; 918 struct rtentry *rt = (struct rtentry *)rn; 919 int error = 0, size; 920 struct rt_addrinfo info; 921 922 if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg)) 923 return 0; 924 bzero((caddr_t)&info, sizeof(info)); 925 dst = rt_key(rt); 926 gate = rt->rt_gateway; 927 netmask = rt_mask(rt); 928 genmask = rt->rt_genmask; 929 if (rt->rt_ifp) { 930 ifpaddr = TAILQ_FIRST(&rt->rt_ifp->if_addrhead)->ifa_addr; 931 ifaaddr = rt->rt_ifa->ifa_addr; 932 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT) 933 brdaddr = rt->rt_ifa->ifa_dstaddr; 934 } 935 size = rt_msg2(RTM_GET, &info, 0, w); 936 if (w->w_req && w->w_tmem) { 937 struct rt_msghdr *rtm = (struct rt_msghdr *)w->w_tmem; 938 939 rtm->rtm_flags = rt->rt_flags; 940 rtm->rtm_use = rt->rt_use; 941 rtm->rtm_rmx = rt->rt_rmx; 942 rtm->rtm_index = rt->rt_ifp->if_index; 943 rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0; 944 rtm->rtm_addrs = info.rti_addrs; 945 error = SYSCTL_OUT(w->w_req, (caddr_t)rtm, size); 946 return (error); 947 } 948 return (error); 949 } 950 951 int 952 sysctl_iflist(af, w) 953 int af; 954 struct walkarg *w; 955 { 956 struct ifnet *ifp; 957 struct ifaddr *ifa; 958 struct rt_addrinfo info; 959 int len, error = 0; 960 961 bzero((caddr_t)&info, sizeof(info)); 962 TAILQ_FOREACH(ifp, &ifnet, if_link) { 963 if (w->w_arg && w->w_arg != ifp->if_index) 964 continue; 965 ifa = TAILQ_FIRST(&ifp->if_addrhead); 966 ifpaddr = ifa->ifa_addr; 967 len = rt_msg2(RTM_IFINFO, &info, (caddr_t)0, w); 968 ifpaddr = 0; 969 if (w->w_req && w->w_tmem) { 970 struct if_msghdr *ifm; 971 972 ifm = (struct if_msghdr *)w->w_tmem; 973 ifm->ifm_index = ifp->if_index; 974 ifm->ifm_flags = (u_short)ifp->if_flags; 975 ifm->ifm_data = ifp->if_data; 976 ifm->ifm_addrs = info.rti_addrs; 977 error = SYSCTL_OUT(w->w_req,(caddr_t)ifm, len); 978 if (error) 979 return (error); 980 } 981 while ((ifa = TAILQ_NEXT(ifa, ifa_link)) != 0) { 982 if (af && af != ifa->ifa_addr->sa_family) 983 continue; 984 if (curproc->p_ucred->cr_prison && prison_if(curthread, ifa->ifa_addr)) 985 continue; 986 ifaaddr = ifa->ifa_addr; 987 netmask = ifa->ifa_netmask; 988 brdaddr = ifa->ifa_dstaddr; 989 len = rt_msg2(RTM_NEWADDR, &info, 0, w); 990 if (w->w_req && w->w_tmem) { 991 struct ifa_msghdr *ifam; 992 993 ifam = (struct ifa_msghdr *)w->w_tmem; 994 ifam->ifam_index = ifa->ifa_ifp->if_index; 995 ifam->ifam_flags = ifa->ifa_flags; 996 ifam->ifam_metric = ifa->ifa_metric; 997 ifam->ifam_addrs = info.rti_addrs; 998 error = SYSCTL_OUT(w->w_req, w->w_tmem, len); 999 if (error) 1000 return (error); 1001 } 1002 } 1003 ifaaddr = netmask = brdaddr = 0; 1004 } 1005 return (0); 1006 } 1007 1008 static int 1009 sysctl_rtsock(SYSCTL_HANDLER_ARGS) 1010 { 1011 int *name = (int *)arg1; 1012 u_int namelen = arg2; 1013 struct radix_node_head *rnh; 1014 int i, s, error = EINVAL; 1015 u_char af; 1016 struct walkarg w; 1017 1018 name ++; 1019 namelen--; 1020 if (req->newptr) 1021 return (EPERM); 1022 if (namelen != 3) 1023 return (EINVAL); 1024 af = name[0]; 1025 Bzero(&w, sizeof(w)); 1026 w.w_op = name[1]; 1027 w.w_arg = name[2]; 1028 w.w_req = req; 1029 1030 s = splnet(); 1031 switch (w.w_op) { 1032 1033 case NET_RT_DUMP: 1034 case NET_RT_FLAGS: 1035 for (i = 1; i <= AF_MAX; i++) 1036 if ((rnh = rt_tables[i]) && (af == 0 || af == i) && 1037 (error = rnh->rnh_walktree(rnh, 1038 sysctl_dumpentry, &w))) 1039 break; 1040 break; 1041 1042 case NET_RT_IFLIST: 1043 error = sysctl_iflist(af, &w); 1044 } 1045 splx(s); 1046 if (w.w_tmem) 1047 free(w.w_tmem, M_RTABLE); 1048 return (error); 1049 } 1050 1051 SYSCTL_NODE(_net, PF_ROUTE, routetable, CTLFLAG_RD, sysctl_rtsock, ""); 1052 1053 /* 1054 * Definitions of protocols supported in the ROUTE domain. 1055 */ 1056 1057 extern struct domain routedomain; /* or at least forward */ 1058 1059 static struct protosw routesw[] = { 1060 { SOCK_RAW, &routedomain, 0, PR_ATOMIC|PR_ADDR, 1061 0, route_output, raw_ctlinput, 0, 1062 cpu0_soport, 1063 raw_init, 0, 0, 0, 1064 &route_usrreqs 1065 } 1066 }; 1067 1068 static struct domain routedomain = 1069 { PF_ROUTE, "route", 0, 0, 0, 1070 routesw, &routesw[sizeof(routesw)/sizeof(routesw[0])] }; 1071 1072 DOMAIN_SET(route); 1073