1 /* 2 * Copyright (c) 2004, 2005 The DragonFly Project. All rights reserved. 3 * 4 * This code is derived from software contributed to The DragonFly Project 5 * by Jeffrey M. Hsu. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 3. Neither the name of The DragonFly Project nor the names of its 16 * contributors may be used to endorse or promote products derived 17 * from this software without specific, prior written permission. 18 * 19 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS 20 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT 21 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS 22 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE 23 * COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, 24 * INCIDENTAL, SPECIAL, EXEMPLARY OR CONSEQUENTIAL DAMAGES (INCLUDING, 25 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; 26 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED 27 * AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, 28 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT 29 * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 30 * SUCH DAMAGE. 31 */ 32 33 /* 34 * Copyright (c) 2004, 2005 Jeffrey M. Hsu. All rights reserved. 35 * 36 * License terms: all terms for the DragonFly license above plus the following: 37 * 38 * 4. All advertising materials mentioning features or use of this software 39 * must display the following acknowledgement: 40 * 41 * This product includes software developed by Jeffrey M. Hsu 42 * for the DragonFly Project. 43 * 44 * This requirement may be waived with permission from Jeffrey Hsu. 45 * Permission will be granted to any DragonFly user for free. 46 * This requirement will sunset and may be removed on Jan 31, 2006, 47 * after which the standard DragonFly license (as shown above) will 48 * apply. 49 */ 50 51 /* 52 * Copyright (c) 1988, 1991, 1993 53 * The Regents of the University of California. All rights reserved. 54 * 55 * Redistribution and use in source and binary forms, with or without 56 * modification, are permitted provided that the following conditions 57 * are met: 58 * 1. Redistributions of source code must retain the above copyright 59 * notice, this list of conditions and the following disclaimer. 60 * 2. Redistributions in binary form must reproduce the above copyright 61 * notice, this list of conditions and the following disclaimer in the 62 * documentation and/or other materials provided with the distribution. 63 * 3. All advertising materials mentioning features or use of this software 64 * must display the following acknowledgement: 65 * This product includes software developed by the University of 66 * California, Berkeley and its contributors. 67 * 4. Neither the name of the University nor the names of its contributors 68 * may be used to endorse or promote products derived from this software 69 * without specific prior written permission. 70 * 71 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 72 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 73 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 74 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 75 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 76 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 77 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 78 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 79 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 80 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 81 * SUCH DAMAGE. 82 * 83 * @(#)rtsock.c 8.7 (Berkeley) 10/12/95 84 * $FreeBSD: src/sys/net/rtsock.c,v 1.44.2.11 2002/12/04 14:05:41 ru Exp $ 85 * $DragonFly: src/sys/net/rtsock.c,v 1.24 2005/03/04 03:37:43 hsu Exp $ 86 */ 87 88 #include <sys/param.h> 89 #include <sys/systm.h> 90 #include <sys/kernel.h> 91 #include <sys/sysctl.h> 92 #include <sys/proc.h> 93 #include <sys/malloc.h> 94 #include <sys/mbuf.h> 95 #include <sys/protosw.h> 96 #include <sys/socket.h> 97 #include <sys/socketvar.h> 98 #include <sys/domain.h> 99 100 #include <machine/stdarg.h> 101 102 #include <net/if.h> 103 #include <net/route.h> 104 #include <net/raw_cb.h> 105 106 MALLOC_DEFINE(M_RTABLE, "routetbl", "routing tables"); 107 108 static struct route_cb { 109 int ip_count; 110 int ip6_count; 111 int ipx_count; 112 int ns_count; 113 int any_count; 114 } route_cb; 115 116 static const struct sockaddr route_src = { 2, PF_ROUTE, }; 117 118 struct walkarg { 119 int w_tmemsize; 120 int w_op, w_arg; 121 void *w_tmem; 122 struct sysctl_req *w_req; 123 }; 124 125 static struct mbuf * 126 rt_msg_mbuf (int, struct rt_addrinfo *); 127 static void rt_msg_buffer (int, struct rt_addrinfo *, void *buf, int len); 128 static int rt_msgsize (int type, struct rt_addrinfo *rtinfo); 129 static int rt_xaddrs (char *, char *, struct rt_addrinfo *); 130 static int sysctl_dumpentry (struct radix_node *rn, void *vw); 131 static int sysctl_iflist (int af, struct walkarg *w); 132 static int route_output(struct mbuf *, struct socket *, ...); 133 static void rt_setmetrics (u_long, struct rt_metrics *, 134 struct rt_metrics *); 135 136 /* 137 * It really doesn't make any sense at all for this code to share much 138 * with raw_usrreq.c, since its functionality is so restricted. XXX 139 */ 140 static int 141 rts_abort(struct socket *so) 142 { 143 int s, error; 144 145 s = splnet(); 146 error = raw_usrreqs.pru_abort(so); 147 splx(s); 148 return error; 149 } 150 151 /* pru_accept is EOPNOTSUPP */ 152 153 static int 154 rts_attach(struct socket *so, int proto, struct pru_attach_info *ai) 155 { 156 struct rawcb *rp; 157 int s, error; 158 159 if (sotorawcb(so) != NULL) 160 return EISCONN; /* XXX panic? */ 161 162 rp = malloc(sizeof *rp, M_PCB, M_WAITOK | M_ZERO); 163 if (rp == NULL) 164 return ENOBUFS; 165 166 /* 167 * The splnet() is necessary to block protocols from sending 168 * error notifications (like RTM_REDIRECT or RTM_LOSING) while 169 * this PCB is extant but incompletely initialized. 170 * Probably we should try to do more of this work beforehand and 171 * eliminate the spl. 172 */ 173 s = splnet(); 174 so->so_pcb = rp; 175 error = raw_attach(so, proto, ai->sb_rlimit); 176 rp = sotorawcb(so); 177 if (error) { 178 splx(s); 179 free(rp, M_PCB); 180 return error; 181 } 182 switch(rp->rcb_proto.sp_protocol) { 183 case AF_INET: 184 route_cb.ip_count++; 185 break; 186 case AF_INET6: 187 route_cb.ip6_count++; 188 break; 189 case AF_IPX: 190 route_cb.ipx_count++; 191 break; 192 case AF_NS: 193 route_cb.ns_count++; 194 break; 195 } 196 rp->rcb_faddr = &route_src; 197 route_cb.any_count++; 198 soisconnected(so); 199 so->so_options |= SO_USELOOPBACK; 200 splx(s); 201 return 0; 202 } 203 204 static int 205 rts_bind(struct socket *so, struct sockaddr *nam, struct thread *td) 206 { 207 int s, error; 208 209 s = splnet(); 210 error = raw_usrreqs.pru_bind(so, nam, td); /* xxx just EINVAL */ 211 splx(s); 212 return error; 213 } 214 215 static int 216 rts_connect(struct socket *so, struct sockaddr *nam, struct thread *td) 217 { 218 int s, error; 219 220 s = splnet(); 221 error = raw_usrreqs.pru_connect(so, nam, td); /* XXX just EINVAL */ 222 splx(s); 223 return error; 224 } 225 226 /* pru_connect2 is EOPNOTSUPP */ 227 /* pru_control is EOPNOTSUPP */ 228 229 static int 230 rts_detach(struct socket *so) 231 { 232 struct rawcb *rp = sotorawcb(so); 233 int s, error; 234 235 s = splnet(); 236 if (rp != NULL) { 237 switch(rp->rcb_proto.sp_protocol) { 238 case AF_INET: 239 route_cb.ip_count--; 240 break; 241 case AF_INET6: 242 route_cb.ip6_count--; 243 break; 244 case AF_IPX: 245 route_cb.ipx_count--; 246 break; 247 case AF_NS: 248 route_cb.ns_count--; 249 break; 250 } 251 route_cb.any_count--; 252 } 253 error = raw_usrreqs.pru_detach(so); 254 splx(s); 255 return error; 256 } 257 258 static int 259 rts_disconnect(struct socket *so) 260 { 261 int s, error; 262 263 s = splnet(); 264 error = raw_usrreqs.pru_disconnect(so); 265 splx(s); 266 return error; 267 } 268 269 /* pru_listen is EOPNOTSUPP */ 270 271 static int 272 rts_peeraddr(struct socket *so, struct sockaddr **nam) 273 { 274 int s, error; 275 276 s = splnet(); 277 error = raw_usrreqs.pru_peeraddr(so, nam); 278 splx(s); 279 return error; 280 } 281 282 /* pru_rcvd is EOPNOTSUPP */ 283 /* pru_rcvoob is EOPNOTSUPP */ 284 285 static int 286 rts_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam, 287 struct mbuf *control, struct thread *td) 288 { 289 int s, error; 290 291 s = splnet(); 292 error = raw_usrreqs.pru_send(so, flags, m, nam, control, td); 293 splx(s); 294 return error; 295 } 296 297 /* pru_sense is null */ 298 299 static int 300 rts_shutdown(struct socket *so) 301 { 302 int s, error; 303 304 s = splnet(); 305 error = raw_usrreqs.pru_shutdown(so); 306 splx(s); 307 return error; 308 } 309 310 static int 311 rts_sockaddr(struct socket *so, struct sockaddr **nam) 312 { 313 int s, error; 314 315 s = splnet(); 316 error = raw_usrreqs.pru_sockaddr(so, nam); 317 splx(s); 318 return error; 319 } 320 321 static struct pr_usrreqs route_usrreqs = { 322 rts_abort, pru_accept_notsupp, rts_attach, rts_bind, rts_connect, 323 pru_connect2_notsupp, pru_control_notsupp, rts_detach, rts_disconnect, 324 pru_listen_notsupp, rts_peeraddr, pru_rcvd_notsupp, pru_rcvoob_notsupp, 325 rts_send, pru_sense_null, rts_shutdown, rts_sockaddr, 326 sosend, soreceive, sopoll 327 }; 328 329 static __inline sa_family_t 330 familyof(struct sockaddr *sa) 331 { 332 return (sa != NULL ? sa->sa_family : 0); 333 } 334 335 static void 336 rts_input(struct mbuf *m, sa_family_t family) 337 { 338 static const struct sockaddr route_dst = { 2, PF_ROUTE, }; 339 struct sockproto route_proto = { PF_ROUTE, family }; 340 341 raw_input(m, &route_proto, &route_src, &route_dst); 342 } 343 344 static void * 345 reallocbuf(void *ptr, size_t len, size_t olen) 346 { 347 void *newptr; 348 349 newptr = malloc(len, M_RTABLE, M_INTWAIT | M_NULLOK); 350 if (newptr == NULL) 351 return NULL; 352 bcopy(ptr, newptr, olen); 353 free(ptr, M_RTABLE); 354 return (newptr); 355 } 356 357 /* 358 * Internal helper routine for route_output(). 359 */ 360 static int 361 fillrtmsg(struct rt_msghdr **prtm, struct rtentry *rt, 362 struct rt_addrinfo *rtinfo) 363 { 364 int msglen; 365 struct rt_msghdr *rtm = *prtm; 366 367 /* Fill in rt_addrinfo for call to rt_msg_buffer(). */ 368 rtinfo->rti_dst = rt_key(rt); 369 rtinfo->rti_gateway = rt->rt_gateway; 370 rtinfo->rti_netmask = rt_mask(rt); /* might be NULL */ 371 rtinfo->rti_genmask = rt->rt_genmask; /* might be NULL */ 372 if (rtm->rtm_addrs & (RTA_IFP | RTA_IFA)) { 373 if (rt->rt_ifp != NULL) { 374 rtinfo->rti_ifpaddr = 375 TAILQ_FIRST(&rt->rt_ifp->if_addrhead)->ifa_addr; 376 rtinfo->rti_ifaaddr = rt->rt_ifa->ifa_addr; 377 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT) 378 rtinfo->rti_bcastaddr = rt->rt_ifa->ifa_dstaddr; 379 rtm->rtm_index = rt->rt_ifp->if_index; 380 } else { 381 rtinfo->rti_ifpaddr = NULL; 382 rtinfo->rti_ifaaddr = NULL; 383 } 384 } 385 386 msglen = rt_msgsize(rtm->rtm_type, rtinfo); 387 if (rtm->rtm_msglen < msglen) { 388 rtm = reallocbuf(rtm, msglen, rtm->rtm_msglen); 389 if (rtm == NULL) 390 return (ENOBUFS); 391 *prtm = rtm; 392 } 393 rt_msg_buffer(rtm->rtm_type, rtinfo, rtm, msglen); 394 395 rtm->rtm_flags = rt->rt_flags; 396 rtm->rtm_rmx = rt->rt_rmx; 397 rtm->rtm_addrs = rtinfo->rti_addrs; 398 399 return (0); 400 } 401 402 /*ARGSUSED*/ 403 static int 404 route_output(struct mbuf *m, struct socket *so, ...) 405 { 406 struct rt_msghdr *rtm = NULL; 407 struct rtentry *rt = NULL; 408 struct rtentry *saved_nrt = NULL; 409 struct radix_node_head *rnh; 410 struct ifaddr *ifa = NULL; 411 struct rawcb *rp = NULL; 412 struct pr_output_info *oi; 413 struct rt_addrinfo rtinfo; 414 int len, error = 0; 415 __va_list ap; 416 417 __va_start(ap, so); 418 oi = __va_arg(ap, struct pr_output_info *); 419 __va_end(ap); 420 421 #define gotoerr(e) { error = e; goto flush;} 422 423 if (m == NULL || 424 (m->m_len < sizeof(long) && 425 (m = m_pullup(m, sizeof(long))) == NULL)) 426 return (ENOBUFS); 427 if (!(m->m_flags & M_PKTHDR)) 428 panic("route_output"); 429 len = m->m_pkthdr.len; 430 if (len < sizeof(struct rt_msghdr) || 431 len != mtod(m, struct rt_msghdr *)->rtm_msglen) { 432 rtinfo.rti_dst = NULL; 433 gotoerr(EINVAL); 434 } 435 rtm = malloc(len, M_RTABLE, M_INTWAIT | M_NULLOK); 436 if (rtm == NULL) { 437 rtinfo.rti_dst = NULL; 438 gotoerr(ENOBUFS); 439 } 440 m_copydata(m, 0, len, (caddr_t)rtm); 441 if (rtm->rtm_version != RTM_VERSION) { 442 rtinfo.rti_dst = NULL; 443 gotoerr(EPROTONOSUPPORT); 444 } 445 rtm->rtm_pid = oi->p_pid; 446 bzero(&rtinfo, sizeof(struct rt_addrinfo)); 447 rtinfo.rti_addrs = rtm->rtm_addrs; 448 if (rt_xaddrs((char *)(rtm + 1), (char *)rtm + len, &rtinfo) != 0) { 449 rtinfo.rti_dst = NULL; 450 gotoerr(EINVAL); 451 } 452 rtinfo.rti_flags = rtm->rtm_flags; 453 if (rtinfo.rti_dst == NULL || rtinfo.rti_dst->sa_family >= AF_MAX || 454 (rtinfo.rti_gateway && rtinfo.rti_gateway->sa_family >= AF_MAX)) 455 gotoerr(EINVAL); 456 457 if (rtinfo.rti_genmask != NULL) { 458 struct radix_node *n; 459 460 #define clen(s) (*(u_char *)(s)) 461 n = rn_addmask((char *)rtinfo.rti_genmask, TRUE, 1); 462 if (n != NULL && 463 rtinfo.rti_genmask->sa_len >= clen(n->rn_key) && 464 bcmp((char *)rtinfo.rti_genmask + 1, 465 (char *)n->rn_key + 1, clen(n->rn_key) - 1) == 0) 466 rtinfo.rti_genmask = (struct sockaddr *)n->rn_key; 467 else 468 gotoerr(ENOBUFS); 469 } 470 471 /* 472 * Verify that the caller has the appropriate privilege; RTM_GET 473 * is the only operation the non-superuser is allowed. 474 */ 475 if (rtm->rtm_type != RTM_GET && suser_cred(so->so_cred, 0) != 0) 476 gotoerr(EPERM); 477 478 switch (rtm->rtm_type) { 479 case RTM_ADD: 480 if (rtinfo.rti_gateway == NULL) 481 gotoerr(EINVAL); 482 error = rtrequest1(RTM_ADD, &rtinfo, &saved_nrt); 483 if (error == 0 && saved_nrt != NULL) { 484 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, 485 &saved_nrt->rt_rmx); 486 saved_nrt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits); 487 saved_nrt->rt_rmx.rmx_locks |= 488 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks); 489 --saved_nrt->rt_refcnt; 490 saved_nrt->rt_genmask = rtinfo.rti_genmask; 491 } 492 break; 493 case RTM_DELETE: 494 error = rtrequest1(RTM_DELETE, &rtinfo, &saved_nrt); 495 if (error == 0) { 496 if ((rt = saved_nrt)) 497 rt->rt_refcnt++; 498 if (fillrtmsg(&rtm, rt, &rtinfo) != 0) 499 gotoerr(ENOBUFS); 500 } 501 break; 502 case RTM_GET: 503 case RTM_CHANGE: 504 case RTM_LOCK: 505 if ((rnh = rt_tables[rtinfo.rti_dst->sa_family]) == NULL) 506 gotoerr(EAFNOSUPPORT); 507 rt = (struct rtentry *) 508 rnh->rnh_lookup((char *)rtinfo.rti_dst, 509 (char *)rtinfo.rti_netmask, rnh); 510 if (rt == NULL) 511 gotoerr(ESRCH); 512 rt->rt_refcnt++; 513 514 switch(rtm->rtm_type) { 515 case RTM_GET: 516 if (fillrtmsg(&rtm, rt, &rtinfo) != 0) 517 gotoerr(ENOBUFS); 518 break; 519 case RTM_CHANGE: 520 /* 521 * new gateway could require new ifaddr, ifp; 522 * flags may also be different; ifp may be specified 523 * by ll sockaddr when protocol address is ambiguous 524 */ 525 if (((rt->rt_flags & RTF_GATEWAY) && 526 rtinfo.rti_gateway != NULL) || 527 rtinfo.rti_ifpaddr != NULL || 528 (rtinfo.rti_ifaaddr != NULL && 529 sa_equal(rtinfo.rti_ifaaddr, 530 rt->rt_ifa->ifa_addr))) { 531 error = rt_getifa(&rtinfo); 532 if (error != 0) 533 gotoerr(error); 534 } 535 if (rtinfo.rti_gateway != NULL) { 536 error = rt_setgate(rt, rt_key(rt), 537 rtinfo.rti_gateway); 538 if (error != 0) 539 gotoerr(error); 540 } 541 if ((ifa = rtinfo.rti_ifa) != NULL) { 542 struct ifaddr *oifa = rt->rt_ifa; 543 544 if (oifa != ifa) { 545 if (oifa && oifa->ifa_rtrequest) 546 oifa->ifa_rtrequest(RTM_DELETE, 547 rt, &rtinfo); 548 IFAFREE(rt->rt_ifa); 549 IFAREF(ifa); 550 rt->rt_ifa = ifa; 551 rt->rt_ifp = rtinfo.rti_ifp; 552 } 553 } 554 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, 555 &rt->rt_rmx); 556 if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest) 557 rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt, &rtinfo); 558 if (rtinfo.rti_genmask != NULL) 559 rt->rt_genmask = rtinfo.rti_genmask; 560 /* 561 * Fall into 562 */ 563 case RTM_LOCK: 564 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits); 565 rt->rt_rmx.rmx_locks |= 566 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks); 567 break; 568 } 569 570 break; 571 default: 572 gotoerr(EOPNOTSUPP); 573 } 574 575 flush: 576 if (rtm != NULL) { 577 if (error != 0) 578 rtm->rtm_errno = error; 579 else 580 rtm->rtm_flags |= RTF_DONE; 581 } 582 if (rt != NULL) 583 rtfree(rt); 584 /* 585 * Check to see if we don't want our own messages. 586 */ 587 if (!(so->so_options & SO_USELOOPBACK)) { 588 if (route_cb.any_count <= 1) { 589 if (rtm != NULL) 590 free(rtm, M_RTABLE); 591 m_freem(m); 592 return (error); 593 } 594 /* There is another listener, so construct message */ 595 rp = sotorawcb(so); 596 } 597 if (rtm != NULL) { 598 m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm); 599 if (m->m_pkthdr.len < rtm->rtm_msglen) { 600 m_freem(m); 601 m = NULL; 602 } else if (m->m_pkthdr.len > rtm->rtm_msglen) 603 m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len); 604 free(rtm, M_RTABLE); 605 } 606 if (rp != NULL) 607 rp->rcb_proto.sp_family = 0; /* Avoid us */ 608 if (m != NULL) 609 rts_input(m, familyof(rtinfo.rti_dst)); 610 if (rp != NULL) 611 rp->rcb_proto.sp_family = PF_ROUTE; 612 return (error); 613 } 614 615 static void 616 rt_setmetrics(u_long which, struct rt_metrics *in, struct rt_metrics *out) 617 { 618 #define setmetric(flag, elt) if (which & (flag)) out->elt = in->elt; 619 setmetric(RTV_RPIPE, rmx_recvpipe); 620 setmetric(RTV_SPIPE, rmx_sendpipe); 621 setmetric(RTV_SSTHRESH, rmx_ssthresh); 622 setmetric(RTV_RTT, rmx_rtt); 623 setmetric(RTV_RTTVAR, rmx_rttvar); 624 setmetric(RTV_HOPCOUNT, rmx_hopcount); 625 setmetric(RTV_MTU, rmx_mtu); 626 setmetric(RTV_EXPIRE, rmx_expire); 627 #undef setmetric 628 } 629 630 #define ROUNDUP(a) \ 631 ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long)) 632 633 /* 634 * Extract the addresses of the passed sockaddrs. 635 * Do a little sanity checking so as to avoid bad memory references. 636 * This data is derived straight from userland. 637 */ 638 static int 639 rt_xaddrs(char *cp, char *cplim, struct rt_addrinfo *rtinfo) 640 { 641 struct sockaddr *sa; 642 int i; 643 644 for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) { 645 if ((rtinfo->rti_addrs & (1 << i)) == 0) 646 continue; 647 sa = (struct sockaddr *)cp; 648 /* 649 * It won't fit. 650 */ 651 if ((cp + sa->sa_len) > cplim) { 652 return (EINVAL); 653 } 654 655 /* 656 * There are no more... Quit now. 657 * If there are more bits, they are in error. 658 * I've seen this. route(1) can evidently generate these. 659 * This causes kernel to core dump. 660 * For compatibility, if we see this, point to a safe address. 661 */ 662 if (sa->sa_len == 0) { 663 static struct sockaddr sa_zero = { 664 sizeof sa_zero, AF_INET, 665 }; 666 667 rtinfo->rti_info[i] = &sa_zero; 668 return (0); /* should be EINVAL but for compat */ 669 } 670 671 /* Accept the sockaddr. */ 672 rtinfo->rti_info[i] = sa; 673 cp += ROUNDUP(sa->sa_len); 674 } 675 return (0); 676 } 677 678 static int 679 rt_msghdrsize(int type) 680 { 681 switch (type) { 682 case RTM_DELADDR: 683 case RTM_NEWADDR: 684 return sizeof(struct ifa_msghdr); 685 case RTM_DELMADDR: 686 case RTM_NEWMADDR: 687 return sizeof(struct ifma_msghdr); 688 case RTM_IFINFO: 689 return sizeof(struct if_msghdr); 690 case RTM_IFANNOUNCE: 691 return sizeof(struct if_announcemsghdr); 692 default: 693 return sizeof(struct rt_msghdr); 694 } 695 } 696 697 static int 698 rt_msgsize(int type, struct rt_addrinfo *rtinfo) 699 { 700 int len, i; 701 702 len = rt_msghdrsize(type); 703 for (i = 0; i < RTAX_MAX; i++) { 704 if (rtinfo->rti_info[i] != NULL) 705 len += ROUNDUP(rtinfo->rti_info[i]->sa_len); 706 } 707 len = ALIGN(len); 708 return len; 709 } 710 711 /* 712 * Build a routing message in a buffer. 713 * Copy the addresses in the rtinfo->rti_info[] sockaddr array 714 * to the end of the buffer after the message header. 715 * 716 * Set the rtinfo->rti_addrs bitmask of addresses present in rtinfo->rti_info[]. 717 * This side-effect can be avoided if we reorder the addrs bitmask field in all 718 * the route messages to line up so we can set it here instead of back in the 719 * calling routine. 720 */ 721 static void 722 rt_msg_buffer(int type, struct rt_addrinfo *rtinfo, void *buf, int msglen) 723 { 724 struct rt_msghdr *rtm; 725 char *cp; 726 int dlen, i; 727 728 rtm = (struct rt_msghdr *) buf; 729 rtm->rtm_version = RTM_VERSION; 730 rtm->rtm_type = type; 731 rtm->rtm_msglen = msglen; 732 733 cp = (char *)buf + rt_msghdrsize(type); 734 rtinfo->rti_addrs = 0; 735 for (i = 0; i < RTAX_MAX; i++) { 736 struct sockaddr *sa; 737 738 if ((sa = rtinfo->rti_info[i]) == NULL) 739 continue; 740 rtinfo->rti_addrs |= (1 << i); 741 dlen = ROUNDUP(sa->sa_len); 742 bcopy(sa, cp, dlen); 743 cp += dlen; 744 } 745 } 746 747 /* 748 * Build a routing message in a mbuf chain. 749 * Copy the addresses in the rtinfo->rti_info[] sockaddr array 750 * to the end of the mbuf after the message header. 751 * 752 * Set the rtinfo->rti_addrs bitmask of addresses present in rtinfo->rti_info[]. 753 * This side-effect can be avoided if we reorder the addrs bitmask field in all 754 * the route messages to line up so we can set it here instead of back in the 755 * calling routine. 756 */ 757 static struct mbuf * 758 rt_msg_mbuf(int type, struct rt_addrinfo *rtinfo) 759 { 760 struct mbuf *m; 761 struct rt_msghdr *rtm; 762 int hlen, len; 763 int i; 764 765 hlen = rt_msghdrsize(type); 766 KASSERT(hlen <= MCLBYTES, ("rt_msg_mbuf: hlen %d doesn't fit", hlen)); 767 768 m = m_gethdr(MB_DONTWAIT, MT_DATA); 769 if (m == NULL) 770 return (NULL); 771 if (hlen > MHLEN) { 772 MCLGET(m, MB_DONTWAIT); 773 if (!(m->m_flags & M_EXT)) { 774 m_free(m); 775 return (NULL); 776 } 777 } 778 m->m_pkthdr.len = m->m_len = hlen; 779 m->m_pkthdr.rcvif = NULL; 780 rtinfo->rti_addrs = 0; 781 len = hlen; 782 for (i = 0; i < RTAX_MAX; i++) { 783 struct sockaddr *sa; 784 int dlen; 785 786 if ((sa = rtinfo->rti_info[i]) == NULL) 787 continue; 788 rtinfo->rti_addrs |= (1 << i); 789 dlen = ROUNDUP(sa->sa_len); 790 m_copyback(m, len, dlen, (caddr_t)sa); /* can grow mbuf chain */ 791 len += dlen; 792 } 793 if (m->m_pkthdr.len != len) { /* one of the m_copyback() calls failed */ 794 m_freem(m); 795 return (NULL); 796 } 797 rtm = mtod(m, struct rt_msghdr *); 798 bzero(rtm, hlen); 799 rtm->rtm_msglen = len; 800 rtm->rtm_version = RTM_VERSION; 801 rtm->rtm_type = type; 802 return (m); 803 } 804 805 /* 806 * This routine is called to generate a message from the routing 807 * socket indicating that a redirect has occurred, a routing lookup 808 * has failed, or that a protocol has detected timeouts to a particular 809 * destination. 810 */ 811 void 812 rt_missmsg(int type, struct rt_addrinfo *rtinfo, int flags, int error) 813 { 814 struct sockaddr *dst = rtinfo->rti_info[RTAX_DST]; 815 struct rt_msghdr *rtm; 816 struct mbuf *m; 817 818 if (route_cb.any_count == 0) 819 return; 820 m = rt_msg_mbuf(type, rtinfo); 821 if (m == NULL) 822 return; 823 rtm = mtod(m, struct rt_msghdr *); 824 rtm->rtm_flags = RTF_DONE | flags; 825 rtm->rtm_errno = error; 826 rtm->rtm_addrs = rtinfo->rti_addrs; 827 rts_input(m, familyof(dst)); 828 } 829 830 void 831 rt_dstmsg(int type, struct sockaddr *dst, int error) 832 { 833 struct rt_msghdr *rtm; 834 struct rt_addrinfo addrs; 835 struct mbuf *m; 836 837 if (route_cb.any_count == 0) 838 return; 839 bzero(&addrs, sizeof(struct rt_addrinfo)); 840 addrs.rti_info[RTAX_DST] = dst; 841 m = rt_msg_mbuf(type, &addrs); 842 if (m == NULL) 843 return; 844 rtm = mtod(m, struct rt_msghdr *); 845 rtm->rtm_flags = RTF_DONE; 846 rtm->rtm_errno = error; 847 rtm->rtm_addrs = addrs.rti_addrs; 848 rts_input(m, familyof(dst)); 849 } 850 851 /* 852 * This routine is called to generate a message from the routing 853 * socket indicating that the status of a network interface has changed. 854 */ 855 void 856 rt_ifmsg(struct ifnet *ifp) 857 { 858 struct if_msghdr *ifm; 859 struct mbuf *m; 860 struct rt_addrinfo rtinfo; 861 862 if (route_cb.any_count == 0) 863 return; 864 bzero(&rtinfo, sizeof(struct rt_addrinfo)); 865 m = rt_msg_mbuf(RTM_IFINFO, &rtinfo); 866 if (m == NULL) 867 return; 868 ifm = mtod(m, struct if_msghdr *); 869 ifm->ifm_index = ifp->if_index; 870 ifm->ifm_flags = (u_short)ifp->if_flags; 871 ifm->ifm_data = ifp->if_data; 872 ifm->ifm_addrs = 0; 873 rts_input(m, 0); 874 } 875 876 static void 877 rt_ifamsg(int cmd, struct ifaddr *ifa) 878 { 879 struct ifa_msghdr *ifam; 880 struct rt_addrinfo rtinfo; 881 struct mbuf *m; 882 struct ifnet *ifp = ifa->ifa_ifp; 883 884 bzero(&rtinfo, sizeof(struct rt_addrinfo)); 885 rtinfo.rti_ifaaddr = ifa->ifa_addr; 886 rtinfo.rti_ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 887 rtinfo.rti_netmask = ifa->ifa_netmask; 888 rtinfo.rti_bcastaddr = ifa->ifa_dstaddr; 889 890 m = rt_msg_mbuf(cmd, &rtinfo); 891 if (m == NULL) 892 return; 893 894 ifam = mtod(m, struct ifa_msghdr *); 895 ifam->ifam_index = ifp->if_index; 896 ifam->ifam_metric = ifa->ifa_metric; 897 ifam->ifam_flags = ifa->ifa_flags; 898 ifam->ifam_addrs = rtinfo.rti_addrs; 899 900 rts_input(m, familyof(ifa->ifa_addr)); 901 } 902 903 void 904 rt_rtmsg(int cmd, struct rtentry *rt, struct ifnet *ifp, int error) 905 { 906 struct rt_msghdr *rtm; 907 struct rt_addrinfo rtinfo; 908 struct mbuf *m; 909 struct sockaddr *dst; 910 911 if (rt == NULL) 912 return; 913 914 bzero(&rtinfo, sizeof(struct rt_addrinfo)); 915 rtinfo.rti_dst = dst = rt_key(rt); 916 rtinfo.rti_gateway = rt->rt_gateway; 917 rtinfo.rti_netmask = rt_mask(rt); 918 if (ifp != NULL) 919 rtinfo.rti_ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 920 rtinfo.rti_ifaaddr = rt->rt_ifa->ifa_addr; 921 922 m = rt_msg_mbuf(cmd, &rtinfo); 923 if (m == NULL) 924 return; 925 926 rtm = mtod(m, struct rt_msghdr *); 927 if (ifp != NULL) 928 rtm->rtm_index = ifp->if_index; 929 rtm->rtm_flags |= rt->rt_flags; 930 rtm->rtm_errno = error; 931 rtm->rtm_addrs = rtinfo.rti_addrs; 932 933 rts_input(m, familyof(dst)); 934 } 935 936 /* 937 * This is called to generate messages from the routing socket 938 * indicating a network interface has had addresses associated with it. 939 * if we ever reverse the logic and replace messages TO the routing 940 * socket indicate a request to configure interfaces, then it will 941 * be unnecessary as the routing socket will automatically generate 942 * copies of it. 943 */ 944 void 945 rt_newaddrmsg(int cmd, struct ifaddr *ifa, int error, struct rtentry *rt) 946 { 947 if (route_cb.any_count == 0) 948 return; 949 950 if (cmd == RTM_ADD) { 951 rt_ifamsg(RTM_NEWADDR, ifa); 952 rt_rtmsg(RTM_ADD, rt, ifa->ifa_ifp, error); 953 } else { 954 KASSERT((cmd == RTM_DELETE), ("unknown cmd %d", cmd)); 955 rt_rtmsg(RTM_DELETE, rt, ifa->ifa_ifp, error); 956 rt_ifamsg(RTM_DELADDR, ifa); 957 } 958 } 959 960 /* 961 * This is the analogue to the rt_newaddrmsg which performs the same 962 * function but for multicast group memberhips. This is easier since 963 * there is no route state to worry about. 964 */ 965 void 966 rt_newmaddrmsg(int cmd, struct ifmultiaddr *ifma) 967 { 968 struct rt_addrinfo rtinfo; 969 struct mbuf *m = NULL; 970 struct ifnet *ifp = ifma->ifma_ifp; 971 struct ifma_msghdr *ifmam; 972 973 if (route_cb.any_count == 0) 974 return; 975 976 bzero(&rtinfo, sizeof(struct rt_addrinfo)); 977 rtinfo.rti_ifaaddr = ifma->ifma_addr; 978 if (ifp != NULL && TAILQ_FIRST(&ifp->if_addrhead) != NULL) 979 rtinfo.rti_ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 980 else 981 rtinfo.rti_ifpaddr = NULL; 982 /* 983 * If a link-layer address is present, present it as a ``gateway'' 984 * (similarly to how ARP entries, e.g., are presented). 985 */ 986 rtinfo.rti_gateway = ifma->ifma_lladdr; 987 988 m = rt_msg_mbuf(cmd, &rtinfo); 989 if (m == NULL) 990 return; 991 992 ifmam = mtod(m, struct ifma_msghdr *); 993 ifmam->ifmam_index = ifp->if_index; 994 ifmam->ifmam_addrs = rtinfo.rti_addrs; 995 996 rts_input(m, familyof(ifma->ifma_addr)); 997 } 998 999 /* 1000 * This is called to generate routing socket messages indicating 1001 * network interface arrival and departure. 1002 */ 1003 void 1004 rt_ifannouncemsg(struct ifnet *ifp, int what) 1005 { 1006 struct rt_addrinfo addrinfo; 1007 struct mbuf *m; 1008 struct if_announcemsghdr *ifan; 1009 1010 if (route_cb.any_count == 0) 1011 return; 1012 1013 bzero(&addrinfo, sizeof addrinfo); 1014 m = rt_msg_mbuf(RTM_IFANNOUNCE, &addrinfo); 1015 if (m == NULL) 1016 return; 1017 1018 ifan = mtod(m, struct if_announcemsghdr *); 1019 ifan->ifan_index = ifp->if_index; 1020 strlcpy(ifan->ifan_name, ifp->if_xname, sizeof ifan->ifan_name); 1021 ifan->ifan_what = what; 1022 1023 rts_input(m, 0); 1024 } 1025 1026 static int 1027 resizewalkarg(struct walkarg *w, int len) 1028 { 1029 void *newptr; 1030 1031 newptr = malloc(len, M_RTABLE, M_INTWAIT | M_NULLOK); 1032 if (newptr == NULL) 1033 return (ENOMEM); 1034 if (w->w_tmem != NULL) 1035 free(w->w_tmem, M_RTABLE); 1036 w->w_tmem = newptr; 1037 w->w_tmemsize = len; 1038 return (0); 1039 } 1040 1041 /* 1042 * This is used in dumping the kernel table via sysctl(). 1043 */ 1044 int 1045 sysctl_dumpentry(struct radix_node *rn, void *vw) 1046 { 1047 struct walkarg *w = vw; 1048 struct rtentry *rt = (struct rtentry *)rn; 1049 struct rt_addrinfo rtinfo; 1050 int error, msglen; 1051 1052 if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg)) 1053 return 0; 1054 1055 bzero(&rtinfo, sizeof(struct rt_addrinfo)); 1056 rtinfo.rti_dst = rt_key(rt); 1057 rtinfo.rti_gateway = rt->rt_gateway; 1058 rtinfo.rti_netmask = rt_mask(rt); 1059 rtinfo.rti_genmask = rt->rt_genmask; 1060 if (rt->rt_ifp != NULL) { 1061 rtinfo.rti_ifpaddr = 1062 TAILQ_FIRST(&rt->rt_ifp->if_addrhead)->ifa_addr; 1063 rtinfo.rti_ifaaddr = rt->rt_ifa->ifa_addr; 1064 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT) 1065 rtinfo.rti_bcastaddr = rt->rt_ifa->ifa_dstaddr; 1066 } 1067 msglen = rt_msgsize(RTM_GET, &rtinfo); 1068 if (w->w_tmemsize < msglen && resizewalkarg(w, msglen) != 0) 1069 return (ENOMEM); 1070 rt_msg_buffer(RTM_GET, &rtinfo, w->w_tmem, msglen); 1071 if (w->w_req != NULL) { 1072 struct rt_msghdr *rtm = w->w_tmem; 1073 1074 rtm->rtm_flags = rt->rt_flags; 1075 rtm->rtm_use = rt->rt_use; 1076 rtm->rtm_rmx = rt->rt_rmx; 1077 rtm->rtm_index = rt->rt_ifp->if_index; 1078 rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0; 1079 rtm->rtm_addrs = rtinfo.rti_addrs; 1080 error = SYSCTL_OUT(w->w_req, rtm, msglen); 1081 return (error); 1082 } 1083 return (0); 1084 } 1085 1086 static int 1087 sysctl_iflist(int af, struct walkarg *w) 1088 { 1089 struct ifnet *ifp; 1090 struct ifaddr *ifa; 1091 struct rt_addrinfo rtinfo; 1092 int msglen, error; 1093 1094 bzero(&rtinfo, sizeof(struct rt_addrinfo)); 1095 TAILQ_FOREACH(ifp, &ifnet, if_link) { 1096 if (w->w_arg && w->w_arg != ifp->if_index) 1097 continue; 1098 ifa = TAILQ_FIRST(&ifp->if_addrhead); 1099 rtinfo.rti_ifpaddr = ifa->ifa_addr; 1100 msglen = rt_msgsize(RTM_IFINFO, &rtinfo); 1101 if (w->w_tmemsize < msglen && resizewalkarg(w, msglen) != 0) 1102 return (ENOMEM); 1103 rt_msg_buffer(RTM_IFINFO, &rtinfo, w->w_tmem, msglen); 1104 rtinfo.rti_ifpaddr = NULL; 1105 if (w->w_req != NULL && w->w_tmem != NULL) { 1106 struct if_msghdr *ifm = w->w_tmem; 1107 1108 ifm->ifm_index = ifp->if_index; 1109 ifm->ifm_flags = (u_short)ifp->if_flags; 1110 ifm->ifm_data = ifp->if_data; 1111 ifm->ifm_addrs = rtinfo.rti_addrs; 1112 error = SYSCTL_OUT(w->w_req, ifm, msglen); 1113 if (error) 1114 return (error); 1115 } 1116 while ((ifa = TAILQ_NEXT(ifa, ifa_link)) != NULL) { 1117 if (af && af != ifa->ifa_addr->sa_family) 1118 continue; 1119 if (curproc->p_ucred->cr_prison && 1120 prison_if(curthread, ifa->ifa_addr)) 1121 continue; 1122 rtinfo.rti_ifaaddr = ifa->ifa_addr; 1123 rtinfo.rti_netmask = ifa->ifa_netmask; 1124 rtinfo.rti_bcastaddr = ifa->ifa_dstaddr; 1125 msglen = rt_msgsize(RTM_NEWADDR, &rtinfo); 1126 if (w->w_tmemsize < msglen && 1127 resizewalkarg(w, msglen) != 0) 1128 return (ENOMEM); 1129 rt_msg_buffer(RTM_NEWADDR, &rtinfo, w->w_tmem, msglen); 1130 if (w->w_req != NULL) { 1131 struct ifa_msghdr *ifam = w->w_tmem; 1132 1133 ifam->ifam_index = ifa->ifa_ifp->if_index; 1134 ifam->ifam_flags = ifa->ifa_flags; 1135 ifam->ifam_metric = ifa->ifa_metric; 1136 ifam->ifam_addrs = rtinfo.rti_addrs; 1137 error = SYSCTL_OUT(w->w_req, w->w_tmem, msglen); 1138 if (error) 1139 return (error); 1140 } 1141 } 1142 rtinfo.rti_netmask = NULL; 1143 rtinfo.rti_ifaaddr = NULL; 1144 rtinfo.rti_bcastaddr = NULL; 1145 } 1146 return (0); 1147 } 1148 1149 static int 1150 sysctl_rtsock(SYSCTL_HANDLER_ARGS) 1151 { 1152 int *name = (int *)arg1; 1153 u_int namelen = arg2; 1154 struct radix_node_head *rnh; 1155 int i, s, error = EINVAL; 1156 u_char af; 1157 struct walkarg w; 1158 1159 name ++; 1160 namelen--; 1161 if (req->newptr) 1162 return (EPERM); 1163 if (namelen != 3) 1164 return (EINVAL); 1165 af = name[0]; 1166 bzero(&w, sizeof w); 1167 w.w_op = name[1]; 1168 w.w_arg = name[2]; 1169 w.w_req = req; 1170 1171 s = splnet(); 1172 switch (w.w_op) { 1173 1174 case NET_RT_DUMP: 1175 case NET_RT_FLAGS: 1176 for (i = 1; i <= AF_MAX; i++) 1177 if ((rnh = rt_tables[i]) && (af == 0 || af == i) && 1178 (error = rnh->rnh_walktree(rnh, 1179 sysctl_dumpentry, &w))) 1180 break; 1181 break; 1182 1183 case NET_RT_IFLIST: 1184 error = sysctl_iflist(af, &w); 1185 } 1186 splx(s); 1187 if (w.w_tmem != NULL) 1188 free(w.w_tmem, M_RTABLE); 1189 return (error); 1190 } 1191 1192 SYSCTL_NODE(_net, PF_ROUTE, routetable, CTLFLAG_RD, sysctl_rtsock, ""); 1193 1194 /* 1195 * Definitions of protocols supported in the ROUTE domain. 1196 */ 1197 1198 extern struct domain routedomain; /* or at least forward */ 1199 1200 static struct protosw routesw[] = { 1201 { SOCK_RAW, &routedomain, 0, PR_ATOMIC|PR_ADDR, 1202 0, route_output, raw_ctlinput, 0, 1203 cpu0_soport, 1204 raw_init, 0, 0, 0, 1205 &route_usrreqs 1206 } 1207 }; 1208 1209 static struct domain routedomain = { 1210 PF_ROUTE, "route", NULL, NULL, NULL, 1211 routesw, &routesw[(sizeof routesw)/(sizeof routesw[0])], 1212 }; 1213 1214 DOMAIN_SET(route); 1215