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 * $FreeBSD$ 35 */ 36 37 38 #include <sys/param.h> 39 #include <sys/systm.h> 40 #include <sys/kernel.h> 41 #include <sys/sysctl.h> 42 #include <sys/proc.h> 43 #include <sys/malloc.h> 44 #include <sys/mbuf.h> 45 #include <sys/socket.h> 46 #include <sys/socketvar.h> 47 #include <sys/domain.h> 48 #include <sys/protosw.h> 49 #include <sys/jail.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 static void rt_setmetrics(u_long, struct rt_metrics *, struct rt_metrics *); 77 78 /* Sleazy use of local variables throughout file, warning!!!! */ 79 #define dst info.rti_info[RTAX_DST] 80 #define gate info.rti_info[RTAX_GATEWAY] 81 #define netmask info.rti_info[RTAX_NETMASK] 82 #define genmask info.rti_info[RTAX_GENMASK] 83 #define ifpaddr info.rti_info[RTAX_IFP] 84 #define ifaaddr info.rti_info[RTAX_IFA] 85 #define brdaddr info.rti_info[RTAX_BRD] 86 87 /* 88 * It really doesn't make any sense at all for this code to share much 89 * with raw_usrreq.c, since its functionality is so restricted. XXX 90 */ 91 static int 92 rts_abort(struct socket *so) 93 { 94 int s, error; 95 s = splnet(); 96 error = raw_usrreqs.pru_abort(so); 97 splx(s); 98 return error; 99 } 100 101 /* pru_accept is EOPNOTSUPP */ 102 103 static int 104 rts_attach(struct socket *so, int proto, struct thread *td) 105 { 106 struct rawcb *rp; 107 int s, error; 108 109 if (sotorawcb(so) != 0) 110 return EISCONN; /* XXX panic? */ 111 /* XXX */ 112 MALLOC(rp, struct rawcb *, sizeof *rp, M_PCB, M_WAITOK | M_ZERO); 113 if (rp == 0) 114 return ENOBUFS; 115 116 /* 117 * The splnet() is necessary to block protocols from sending 118 * error notifications (like RTM_REDIRECT or RTM_LOSING) while 119 * this PCB is extant but incompletely initialized. 120 * Probably we should try to do more of this work beforehand and 121 * eliminate the spl. 122 */ 123 s = splnet(); 124 so->so_pcb = (caddr_t)rp; 125 error = raw_attach(so, proto); 126 rp = sotorawcb(so); 127 if (error) { 128 splx(s); 129 so->so_pcb = NULL; 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(m, so) 276 register struct mbuf *m; 277 struct socket *so; 278 { 279 register struct rt_msghdr *rtm = 0; 280 register struct rtentry *rt = 0; 281 struct rtentry *saved_nrt = 0; 282 struct radix_node_head *rnh; 283 struct rt_addrinfo info; 284 int len, error = 0; 285 struct ifnet *ifp = 0; 286 struct ifaddr *ifa = 0; 287 288 #define senderr(e) { error = e; goto flush;} 289 if (m == 0 || ((m->m_len < sizeof(long)) && 290 (m = m_pullup(m, sizeof(long))) == 0)) 291 return (ENOBUFS); 292 if ((m->m_flags & M_PKTHDR) == 0) 293 panic("route_output"); 294 len = m->m_pkthdr.len; 295 if (len < sizeof(*rtm) || 296 len != mtod(m, struct rt_msghdr *)->rtm_msglen) { 297 dst = 0; 298 senderr(EINVAL); 299 } 300 R_Malloc(rtm, struct rt_msghdr *, len); 301 if (rtm == 0) { 302 dst = 0; 303 senderr(ENOBUFS); 304 } 305 m_copydata(m, 0, len, (caddr_t)rtm); 306 if (rtm->rtm_version != RTM_VERSION) { 307 dst = 0; 308 senderr(EPROTONOSUPPORT); 309 } 310 rtm->rtm_pid = curproc->p_pid; 311 bzero(&info, sizeof(info)); 312 info.rti_addrs = rtm->rtm_addrs; 313 if (rt_xaddrs((caddr_t)(rtm + 1), len + (caddr_t)rtm, &info)) { 314 dst = 0; 315 senderr(EINVAL); 316 } 317 info.rti_flags = rtm->rtm_flags; 318 if (dst == 0 || (dst->sa_family >= AF_MAX) 319 || (gate != 0 && (gate->sa_family >= AF_MAX))) 320 senderr(EINVAL); 321 if (genmask) { 322 struct radix_node *t; 323 t = rn_addmask((caddr_t)genmask, 0, 1); 324 if (t && Bcmp((caddr_t *)genmask + 1, (caddr_t *)t->rn_key + 1, 325 *(u_char *)t->rn_key - 1) == 0) 326 genmask = (struct sockaddr *)(t->rn_key); 327 else 328 senderr(ENOBUFS); 329 } 330 331 /* 332 * Verify that the caller has the appropriate privilege; RTM_GET 333 * is the only operation the non-superuser is allowed. 334 */ 335 if (rtm->rtm_type != RTM_GET && (error = suser(curproc)) != 0) 336 senderr(error); 337 338 switch (rtm->rtm_type) { 339 340 case RTM_ADD: 341 if (gate == 0) 342 senderr(EINVAL); 343 error = rtrequest1(RTM_ADD, &info, &saved_nrt); 344 if (error == 0 && saved_nrt) { 345 rt_setmetrics(rtm->rtm_inits, 346 &rtm->rtm_rmx, &saved_nrt->rt_rmx); 347 saved_nrt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits); 348 saved_nrt->rt_rmx.rmx_locks |= 349 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks); 350 saved_nrt->rt_refcnt--; 351 saved_nrt->rt_genmask = genmask; 352 } 353 break; 354 355 case RTM_DELETE: 356 error = rtrequest1(RTM_DELETE, &info, &saved_nrt); 357 if (error == 0) { 358 if ((rt = saved_nrt)) 359 rt->rt_refcnt++; 360 goto report; 361 } 362 break; 363 364 case RTM_GET: 365 case RTM_CHANGE: 366 case RTM_LOCK: 367 if ((rnh = rt_tables[dst->sa_family]) == 0) { 368 senderr(EAFNOSUPPORT); 369 } else if ((rt = (struct rtentry *) 370 rnh->rnh_lookup(dst, netmask, rnh)) != NULL) 371 rt->rt_refcnt++; 372 else 373 senderr(ESRCH); 374 switch(rtm->rtm_type) { 375 376 case RTM_GET: 377 report: 378 dst = rt_key(rt); 379 gate = rt->rt_gateway; 380 netmask = rt_mask(rt); 381 genmask = rt->rt_genmask; 382 if (rtm->rtm_addrs & (RTA_IFP | RTA_IFA)) { 383 ifp = rt->rt_ifp; 384 if (ifp) { 385 ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 386 ifaaddr = rt->rt_ifa->ifa_addr; 387 if (ifp->if_flags & IFF_POINTOPOINT) 388 brdaddr = rt->rt_ifa->ifa_dstaddr; 389 rtm->rtm_index = ifp->if_index; 390 } else { 391 ifpaddr = 0; 392 ifaaddr = 0; 393 } 394 } 395 len = rt_msg2(rtm->rtm_type, &info, (caddr_t)0, 396 (struct walkarg *)0); 397 if (len > rtm->rtm_msglen) { 398 struct rt_msghdr *new_rtm; 399 R_Malloc(new_rtm, struct rt_msghdr *, len); 400 if (new_rtm == 0) 401 senderr(ENOBUFS); 402 Bcopy(rtm, new_rtm, rtm->rtm_msglen); 403 Free(rtm); rtm = new_rtm; 404 } 405 (void)rt_msg2(rtm->rtm_type, &info, (caddr_t)rtm, 406 (struct walkarg *)0); 407 rtm->rtm_flags = rt->rt_flags; 408 rtm->rtm_rmx = rt->rt_rmx; 409 rtm->rtm_addrs = info.rti_addrs; 410 break; 411 412 case RTM_CHANGE: 413 /* new gateway could require new ifaddr, ifp; 414 flags may also be different; ifp may be specified 415 by ll sockaddr when protocol address is ambiguous */ 416 #define equal(a1, a2) (bcmp((caddr_t)(a1), (caddr_t)(a2), (a1)->sa_len) == 0) 417 if ((rt->rt_flags & RTF_GATEWAY && gate != NULL) || 418 ifpaddr != NULL || 419 (ifaaddr != NULL && 420 !equal(ifaaddr, rt->rt_ifa->ifa_addr))) { 421 if ((error = rt_getifa(&info)) != 0) 422 senderr(error); 423 } 424 if (gate != NULL && 425 (error = rt_setgate(rt, rt_key(rt), gate)) != 0) 426 senderr(error); 427 if ((ifa = info.rti_ifa) != NULL) { 428 register struct ifaddr *oifa = rt->rt_ifa; 429 if (oifa != ifa) { 430 if (oifa && oifa->ifa_rtrequest) 431 oifa->ifa_rtrequest(RTM_DELETE, rt, 432 &info); 433 IFAFREE(rt->rt_ifa); 434 rt->rt_ifa = ifa; 435 ifa->ifa_refcnt++; 436 rt->rt_ifp = info.rti_ifp; 437 } 438 } 439 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, 440 &rt->rt_rmx); 441 if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest) 442 rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt, &info); 443 if (genmask) 444 rt->rt_genmask = genmask; 445 /* 446 * Fall into 447 */ 448 case RTM_LOCK: 449 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits); 450 rt->rt_rmx.rmx_locks |= 451 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks); 452 break; 453 } 454 break; 455 456 default: 457 senderr(EOPNOTSUPP); 458 } 459 460 flush: 461 if (rtm) { 462 if (error) 463 rtm->rtm_errno = error; 464 else 465 rtm->rtm_flags |= RTF_DONE; 466 } 467 if (rt) 468 rtfree(rt); 469 { 470 register struct rawcb *rp = 0; 471 /* 472 * Check to see if we don't want our own messages. 473 */ 474 if ((so->so_options & SO_USELOOPBACK) == 0) { 475 if (route_cb.any_count <= 1) { 476 if (rtm) 477 Free(rtm); 478 m_freem(m); 479 return (error); 480 } 481 /* There is another listener, so construct message */ 482 rp = sotorawcb(so); 483 } 484 if (rtm) { 485 m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm); 486 if (m->m_pkthdr.len < rtm->rtm_msglen) { 487 m_freem(m); 488 m = NULL; 489 } else if (m->m_pkthdr.len > rtm->rtm_msglen) 490 m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len); 491 Free(rtm); 492 } 493 if (rp) 494 rp->rcb_proto.sp_family = 0; /* Avoid us */ 495 if (dst) 496 route_proto.sp_protocol = dst->sa_family; 497 if (m) 498 raw_input(m, &route_proto, &route_src, &route_dst); 499 if (rp) 500 rp->rcb_proto.sp_family = PF_ROUTE; 501 } 502 return (error); 503 } 504 505 static void 506 rt_setmetrics(which, in, out) 507 u_long which; 508 register struct rt_metrics *in, *out; 509 { 510 #define metric(f, e) if (which & (f)) out->e = in->e; 511 metric(RTV_RPIPE, rmx_recvpipe); 512 metric(RTV_SPIPE, rmx_sendpipe); 513 metric(RTV_SSTHRESH, rmx_ssthresh); 514 metric(RTV_RTT, rmx_rtt); 515 metric(RTV_RTTVAR, rmx_rttvar); 516 metric(RTV_HOPCOUNT, rmx_hopcount); 517 metric(RTV_MTU, rmx_mtu); 518 metric(RTV_EXPIRE, rmx_expire); 519 #undef metric 520 } 521 522 #define ROUNDUP(a) \ 523 ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long)) 524 #define ADVANCE(x, n) (x += ROUNDUP((n)->sa_len)) 525 526 527 /* 528 * Extract the addresses of the passed sockaddrs. 529 * Do a little sanity checking so as to avoid bad memory references. 530 * This data is derived straight from userland. 531 */ 532 static int 533 rt_xaddrs(cp, cplim, rtinfo) 534 register caddr_t cp, cplim; 535 register struct rt_addrinfo *rtinfo; 536 { 537 register struct sockaddr *sa; 538 register int i; 539 540 for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) { 541 if ((rtinfo->rti_addrs & (1 << i)) == 0) 542 continue; 543 sa = (struct sockaddr *)cp; 544 /* 545 * It won't fit. 546 */ 547 if ( (cp + sa->sa_len) > cplim ) { 548 return (EINVAL); 549 } 550 551 /* 552 * there are no more.. quit now 553 * If there are more bits, they are in error. 554 * I've seen this. route(1) can evidently generate these. 555 * This causes kernel to core dump. 556 * for compatibility, If we see this, point to a safe address. 557 */ 558 if (sa->sa_len == 0) { 559 rtinfo->rti_info[i] = &sa_zero; 560 return (0); /* should be EINVAL but for compat */ 561 } 562 563 /* accept it */ 564 rtinfo->rti_info[i] = sa; 565 ADVANCE(cp, sa); 566 } 567 return (0); 568 } 569 570 static struct mbuf * 571 rt_msg1(type, rtinfo) 572 int type; 573 register struct rt_addrinfo *rtinfo; 574 { 575 register struct rt_msghdr *rtm; 576 register struct mbuf *m; 577 register int i; 578 register struct sockaddr *sa; 579 int len, dlen; 580 581 switch (type) { 582 583 case RTM_DELADDR: 584 case RTM_NEWADDR: 585 len = sizeof(struct ifa_msghdr); 586 break; 587 588 case RTM_DELMADDR: 589 case RTM_NEWMADDR: 590 len = sizeof(struct ifma_msghdr); 591 break; 592 593 case RTM_IFINFO: 594 len = sizeof(struct if_msghdr); 595 break; 596 597 case RTM_IFANNOUNCE: 598 len = sizeof(struct if_announcemsghdr); 599 break; 600 601 default: 602 len = sizeof(struct rt_msghdr); 603 } 604 if (len > MCLBYTES) 605 panic("rt_msg1"); 606 m = m_gethdr(M_DONTWAIT, MT_DATA); 607 if (m && len > MHLEN) { 608 MCLGET(m, M_DONTWAIT); 609 if ((m->m_flags & M_EXT) == 0) { 610 m_free(m); 611 m = NULL; 612 } 613 } 614 if (m == 0) 615 return (m); 616 m->m_pkthdr.len = m->m_len = len; 617 m->m_pkthdr.rcvif = 0; 618 rtm = mtod(m, struct rt_msghdr *); 619 bzero((caddr_t)rtm, len); 620 for (i = 0; i < RTAX_MAX; i++) { 621 if ((sa = rtinfo->rti_info[i]) == NULL) 622 continue; 623 rtinfo->rti_addrs |= (1 << i); 624 dlen = ROUNDUP(sa->sa_len); 625 m_copyback(m, len, dlen, (caddr_t)sa); 626 len += dlen; 627 } 628 if (m->m_pkthdr.len != len) { 629 m_freem(m); 630 return (NULL); 631 } 632 rtm->rtm_msglen = len; 633 rtm->rtm_version = RTM_VERSION; 634 rtm->rtm_type = type; 635 return (m); 636 } 637 638 static int 639 rt_msg2(type, rtinfo, cp, w) 640 int type; 641 register struct rt_addrinfo *rtinfo; 642 caddr_t cp; 643 struct walkarg *w; 644 { 645 register int i; 646 int len, dlen, second_time = 0; 647 caddr_t cp0; 648 649 rtinfo->rti_addrs = 0; 650 again: 651 switch (type) { 652 653 case RTM_DELADDR: 654 case RTM_NEWADDR: 655 len = sizeof(struct ifa_msghdr); 656 break; 657 658 case RTM_IFINFO: 659 len = sizeof(struct if_msghdr); 660 break; 661 662 default: 663 len = sizeof(struct rt_msghdr); 664 } 665 cp0 = cp; 666 if (cp0) 667 cp += len; 668 for (i = 0; i < RTAX_MAX; i++) { 669 register struct sockaddr *sa; 670 671 if ((sa = rtinfo->rti_info[i]) == 0) 672 continue; 673 rtinfo->rti_addrs |= (1 << i); 674 dlen = ROUNDUP(sa->sa_len); 675 if (cp) { 676 bcopy((caddr_t)sa, cp, (unsigned)dlen); 677 cp += dlen; 678 } 679 len += dlen; 680 } 681 len = ALIGN(len); 682 if (cp == 0 && w != NULL && !second_time) { 683 register struct walkarg *rw = w; 684 685 if (rw->w_req) { 686 if (rw->w_tmemsize < len) { 687 if (rw->w_tmem) 688 free(rw->w_tmem, M_RTABLE); 689 rw->w_tmem = (caddr_t) 690 malloc(len, M_RTABLE, M_NOWAIT); 691 if (rw->w_tmem) 692 rw->w_tmemsize = len; 693 } 694 if (rw->w_tmem) { 695 cp = rw->w_tmem; 696 second_time = 1; 697 goto again; 698 } 699 } 700 } 701 if (cp) { 702 register 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 register struct rt_addrinfo *rtinfo; 721 { 722 register struct rt_msghdr *rtm; 723 register struct mbuf *m; 724 struct sockaddr *sa = rtinfo->rti_info[RTAX_DST]; 725 726 if (route_cb.any_count == 0) 727 return; 728 m = rt_msg1(type, rtinfo); 729 if (m == 0) 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 register struct ifnet *ifp; 746 { 747 register 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((caddr_t)&info, sizeof(info)); 754 m = rt_msg1(RTM_IFINFO, &info); 755 if (m == 0) 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 = 0; 762 route_proto.sp_protocol = 0; 763 raw_input(m, &route_proto, &route_src, &route_dst); 764 } 765 766 /* 767 * This is called to generate messages from the routing socket 768 * indicating a network interface has had addresses associated with it. 769 * if we ever reverse the logic and replace messages TO the routing 770 * socket indicate a request to configure interfaces, then it will 771 * be unnecessary as the routing socket will automatically generate 772 * copies of it. 773 */ 774 void 775 rt_newaddrmsg(cmd, ifa, error, rt) 776 int cmd, error; 777 register struct ifaddr *ifa; 778 register struct rtentry *rt; 779 { 780 struct rt_addrinfo info; 781 struct sockaddr *sa = 0; 782 int pass; 783 struct mbuf *m = 0; 784 struct ifnet *ifp = ifa->ifa_ifp; 785 786 if (route_cb.any_count == 0) 787 return; 788 for (pass = 1; pass < 3; pass++) { 789 bzero((caddr_t)&info, sizeof(info)); 790 if ((cmd == RTM_ADD && pass == 1) || 791 (cmd == RTM_DELETE && pass == 2)) { 792 register struct ifa_msghdr *ifam; 793 int ncmd = cmd == RTM_ADD ? RTM_NEWADDR : RTM_DELADDR; 794 795 ifaaddr = sa = ifa->ifa_addr; 796 ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 797 netmask = ifa->ifa_netmask; 798 brdaddr = ifa->ifa_dstaddr; 799 if ((m = rt_msg1(ncmd, &info)) == NULL) 800 continue; 801 ifam = mtod(m, struct ifa_msghdr *); 802 ifam->ifam_index = ifp->if_index; 803 ifam->ifam_metric = ifa->ifa_metric; 804 ifam->ifam_flags = ifa->ifa_flags; 805 ifam->ifam_addrs = info.rti_addrs; 806 } 807 if ((cmd == RTM_ADD && pass == 2) || 808 (cmd == RTM_DELETE && pass == 1)) { 809 register struct rt_msghdr *rtm; 810 811 if (rt == 0) 812 continue; 813 netmask = rt_mask(rt); 814 dst = sa = rt_key(rt); 815 gate = rt->rt_gateway; 816 if ((m = rt_msg1(cmd, &info)) == NULL) 817 continue; 818 rtm = mtod(m, struct rt_msghdr *); 819 rtm->rtm_index = ifp->if_index; 820 rtm->rtm_flags |= rt->rt_flags; 821 rtm->rtm_errno = error; 822 rtm->rtm_addrs = info.rti_addrs; 823 } 824 route_proto.sp_protocol = sa ? sa->sa_family : 0; 825 raw_input(m, &route_proto, &route_src, &route_dst); 826 } 827 } 828 829 /* 830 * This is the analogue to the rt_newaddrmsg which performs the same 831 * function but for multicast group memberhips. This is easier since 832 * there is no route state to worry about. 833 */ 834 void 835 rt_newmaddrmsg(cmd, ifma) 836 int cmd; 837 struct ifmultiaddr *ifma; 838 { 839 struct rt_addrinfo info; 840 struct mbuf *m = 0; 841 struct ifnet *ifp = ifma->ifma_ifp; 842 struct ifma_msghdr *ifmam; 843 844 if (route_cb.any_count == 0) 845 return; 846 847 bzero((caddr_t)&info, sizeof(info)); 848 ifaaddr = ifma->ifma_addr; 849 if (ifp && TAILQ_FIRST(&ifp->if_addrhead)) 850 ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 851 else 852 ifpaddr = NULL; 853 /* 854 * If a link-layer address is present, present it as a ``gateway'' 855 * (similarly to how ARP entries, e.g., are presented). 856 */ 857 gate = ifma->ifma_lladdr; 858 if ((m = rt_msg1(cmd, &info)) == NULL) 859 return; 860 ifmam = mtod(m, struct ifma_msghdr *); 861 ifmam->ifmam_index = ifp->if_index; 862 ifmam->ifmam_addrs = info.rti_addrs; 863 route_proto.sp_protocol = ifma->ifma_addr->sa_family; 864 raw_input(m, &route_proto, &route_src, &route_dst); 865 } 866 867 /* 868 * This is called to generate routing socket messages indicating 869 * network interface arrival and departure. 870 */ 871 void 872 rt_ifannouncemsg(ifp, what) 873 struct ifnet *ifp; 874 int what; 875 { 876 struct if_announcemsghdr *ifan; 877 struct mbuf *m; 878 struct rt_addrinfo info; 879 880 if (route_cb.any_count == 0) 881 return; 882 bzero((caddr_t)&info, sizeof(info)); 883 m = rt_msg1(RTM_IFANNOUNCE, &info); 884 if (m == NULL) 885 return; 886 ifan = mtod(m, struct if_announcemsghdr *); 887 ifan->ifan_index = ifp->if_index; 888 snprintf(ifan->ifan_name, sizeof(ifan->ifan_name), 889 "%s%d", ifp->if_name, ifp->if_unit); 890 ifan->ifan_what = what; 891 route_proto.sp_protocol = 0; 892 raw_input(m, &route_proto, &route_src, &route_dst); 893 } 894 895 /* 896 * This is used in dumping the kernel table via sysctl(). 897 */ 898 int 899 sysctl_dumpentry(rn, vw) 900 struct radix_node *rn; 901 void *vw; 902 { 903 register struct walkarg *w = vw; 904 register struct rtentry *rt = (struct rtentry *)rn; 905 int error = 0, size; 906 struct rt_addrinfo info; 907 908 if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg)) 909 return 0; 910 bzero((caddr_t)&info, sizeof(info)); 911 dst = rt_key(rt); 912 gate = rt->rt_gateway; 913 netmask = rt_mask(rt); 914 genmask = rt->rt_genmask; 915 if (rt->rt_ifp) { 916 ifpaddr = TAILQ_FIRST(&rt->rt_ifp->if_addrhead)->ifa_addr; 917 ifaaddr = rt->rt_ifa->ifa_addr; 918 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT) 919 brdaddr = rt->rt_ifa->ifa_dstaddr; 920 } 921 size = rt_msg2(RTM_GET, &info, 0, w); 922 if (w->w_req && w->w_tmem) { 923 register struct rt_msghdr *rtm = (struct rt_msghdr *)w->w_tmem; 924 925 rtm->rtm_flags = rt->rt_flags; 926 rtm->rtm_use = rt->rt_use; 927 rtm->rtm_rmx = rt->rt_rmx; 928 rtm->rtm_index = rt->rt_ifp->if_index; 929 rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0; 930 rtm->rtm_addrs = info.rti_addrs; 931 error = SYSCTL_OUT(w->w_req, (caddr_t)rtm, size); 932 return (error); 933 } 934 return (error); 935 } 936 937 int 938 sysctl_iflist(af, w) 939 int af; 940 register struct walkarg *w; 941 { 942 register struct ifnet *ifp; 943 register struct ifaddr *ifa; 944 struct rt_addrinfo info; 945 int len, error = 0; 946 947 bzero((caddr_t)&info, sizeof(info)); 948 TAILQ_FOREACH(ifp, &ifnet, if_link) { 949 if (w->w_arg && w->w_arg != ifp->if_index) 950 continue; 951 ifa = TAILQ_FIRST(&ifp->if_addrhead); 952 ifpaddr = ifa->ifa_addr; 953 len = rt_msg2(RTM_IFINFO, &info, (caddr_t)0, w); 954 ifpaddr = 0; 955 if (w->w_req && w->w_tmem) { 956 register struct if_msghdr *ifm; 957 958 ifm = (struct if_msghdr *)w->w_tmem; 959 ifm->ifm_index = ifp->if_index; 960 ifm->ifm_flags = (u_short)ifp->if_flags; 961 ifm->ifm_data = ifp->if_data; 962 ifm->ifm_addrs = info.rti_addrs; 963 error = SYSCTL_OUT(w->w_req,(caddr_t)ifm, len); 964 if (error) 965 goto done; 966 } 967 while ((ifa = TAILQ_NEXT(ifa, ifa_link)) != 0) { 968 if (af && af != ifa->ifa_addr->sa_family) 969 continue; 970 if (jailed(curthread->td_ucred) && 971 prison_if(curthread->td_ucred, ifa->ifa_addr)) 972 continue; 973 ifaaddr = ifa->ifa_addr; 974 netmask = ifa->ifa_netmask; 975 brdaddr = ifa->ifa_dstaddr; 976 len = rt_msg2(RTM_NEWADDR, &info, 0, w); 977 if (w->w_req && w->w_tmem) { 978 register struct ifa_msghdr *ifam; 979 980 ifam = (struct ifa_msghdr *)w->w_tmem; 981 ifam->ifam_index = ifa->ifa_ifp->if_index; 982 ifam->ifam_flags = ifa->ifa_flags; 983 ifam->ifam_metric = ifa->ifa_metric; 984 ifam->ifam_addrs = info.rti_addrs; 985 error = SYSCTL_OUT(w->w_req, w->w_tmem, len); 986 if (error) 987 goto done; 988 } 989 } 990 ifaaddr = netmask = brdaddr = 0; 991 } 992 done: 993 return (error); 994 } 995 996 static int 997 sysctl_rtsock(SYSCTL_HANDLER_ARGS) 998 { 999 int *name = (int *)arg1; 1000 u_int namelen = arg2; 1001 register struct radix_node_head *rnh; 1002 int i, s, error = EINVAL; 1003 u_char af; 1004 struct walkarg w; 1005 1006 name ++; 1007 namelen--; 1008 if (req->newptr) 1009 return (EPERM); 1010 if (namelen != 3) 1011 return ((namelen < 3) ? EISDIR : ENOTDIR); 1012 af = name[0]; 1013 Bzero(&w, sizeof(w)); 1014 w.w_op = name[1]; 1015 w.w_arg = name[2]; 1016 w.w_req = req; 1017 1018 s = splnet(); 1019 switch (w.w_op) { 1020 1021 case NET_RT_DUMP: 1022 case NET_RT_FLAGS: 1023 for (i = 1; i <= AF_MAX; i++) 1024 if ((rnh = rt_tables[i]) && (af == 0 || af == i) && 1025 (error = rnh->rnh_walktree(rnh, 1026 sysctl_dumpentry, &w))) 1027 break; 1028 break; 1029 1030 case NET_RT_IFLIST: 1031 error = sysctl_iflist(af, &w); 1032 } 1033 splx(s); 1034 if (w.w_tmem) 1035 free(w.w_tmem, M_RTABLE); 1036 return (error); 1037 } 1038 1039 SYSCTL_NODE(_net, PF_ROUTE, routetable, CTLFLAG_RD, sysctl_rtsock, ""); 1040 1041 /* 1042 * Definitions of protocols supported in the ROUTE domain. 1043 */ 1044 1045 extern struct domain routedomain; /* or at least forward */ 1046 1047 static struct protosw routesw[] = { 1048 { SOCK_RAW, &routedomain, 0, PR_ATOMIC|PR_ADDR, 1049 0, route_output, raw_ctlinput, 0, 1050 0, 1051 raw_init, 0, 0, 0, 1052 &route_usrreqs 1053 } 1054 }; 1055 1056 static struct domain routedomain = 1057 { PF_ROUTE, "route", 0, 0, 0, 1058 routesw, &routesw[sizeof(routesw)/sizeof(routesw[0])] }; 1059 1060 DOMAIN_SET(route); 1061