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 1.21 2005/01/26 23:09:57 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 char *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 static int 358 fillrtmsg(struct rt_msghdr **prtm, struct rtentry *rt, 359 struct rt_addrinfo *rtinfo) 360 { 361 int msglen; 362 struct rt_msghdr *rtm = *prtm; 363 364 /* Fill in rt_addrinfo for call to rt_msg_buffer(). */ 365 rtinfo->rti_dst = rt_key(rt); 366 rtinfo->rti_gateway = rt->rt_gateway; 367 rtinfo->rti_netmask = rt_mask(rt); /* might be NULL */ 368 rtinfo->rti_genmask = rt->rt_genmask; /* might be NULL */ 369 if (rtm->rtm_addrs & (RTA_IFP | RTA_IFA)) { 370 if (rt->rt_ifp != NULL) { 371 rtinfo->rti_ifpaddr = 372 TAILQ_FIRST(&rt->rt_ifp->if_addrhead)->ifa_addr; 373 rtinfo->rti_ifaaddr = rt->rt_ifa->ifa_addr; 374 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT) 375 rtinfo->rti_bcastaddr = rt->rt_ifa->ifa_dstaddr; 376 rtm->rtm_index = rt->rt_ifp->if_index; 377 } else { 378 rtinfo->rti_ifpaddr = NULL; 379 rtinfo->rti_ifaaddr = NULL; 380 } 381 } 382 383 msglen = rt_msgsize(rtm->rtm_type, rtinfo); 384 if (rtm->rtm_msglen < msglen) { 385 rtm = reallocbuf(rtm, msglen, rtm->rtm_msglen); 386 if (rtm == NULL) 387 return (ENOBUFS); 388 *prtm = rtm; 389 } 390 rt_msg_buffer(rtm->rtm_type, rtinfo, rtm, msglen); 391 392 rtm->rtm_flags = rt->rt_flags; 393 rtm->rtm_rmx = rt->rt_rmx; 394 rtm->rtm_addrs = rtinfo->rti_addrs; 395 396 return (0); 397 } 398 399 /*ARGSUSED*/ 400 static int 401 route_output(struct mbuf *m, struct socket *so, ...) 402 { 403 struct rt_msghdr *rtm = NULL; 404 struct rtentry *rt = NULL; 405 struct rtentry *saved_nrt = NULL; 406 struct radix_node_head *rnh; 407 struct ifaddr *ifa = NULL; 408 struct rawcb *rp = NULL; 409 struct pr_output_info *oi; 410 struct rt_addrinfo rtinfo; 411 int len, error = 0; 412 __va_list ap; 413 414 __va_start(ap, so); 415 oi = __va_arg(ap, struct pr_output_info *); 416 __va_end(ap); 417 418 #define gotoerr(e) { error = e; goto flush;} 419 420 if (m == NULL || 421 (m->m_len < sizeof(long) && 422 (m = m_pullup(m, sizeof(long))) == NULL)) 423 return (ENOBUFS); 424 if (!(m->m_flags & M_PKTHDR)) 425 panic("route_output"); 426 len = m->m_pkthdr.len; 427 if (len < sizeof(struct rt_msghdr) || 428 len != mtod(m, struct rt_msghdr *)->rtm_msglen) { 429 rtinfo.rti_dst = NULL; 430 gotoerr(EINVAL); 431 } 432 rtm = malloc(len, M_RTABLE, M_INTWAIT | M_NULLOK); 433 if (rtm == NULL) { 434 rtinfo.rti_dst = NULL; 435 gotoerr(ENOBUFS); 436 } 437 m_copydata(m, 0, len, (caddr_t)rtm); 438 if (rtm->rtm_version != RTM_VERSION) { 439 rtinfo.rti_dst = NULL; 440 gotoerr(EPROTONOSUPPORT); 441 } 442 rtm->rtm_pid = oi->p_pid; 443 bzero(&rtinfo, sizeof(struct rt_addrinfo)); 444 rtinfo.rti_addrs = rtm->rtm_addrs; 445 if (rt_xaddrs((char *)(rtm + 1), (char *)rtm + len, &rtinfo) != 0) { 446 rtinfo.rti_dst = NULL; 447 gotoerr(EINVAL); 448 } 449 rtinfo.rti_flags = rtm->rtm_flags; 450 if (rtinfo.rti_dst == NULL || rtinfo.rti_dst->sa_family >= AF_MAX || 451 (rtinfo.rti_gateway && rtinfo.rti_gateway->sa_family >= AF_MAX)) 452 gotoerr(EINVAL); 453 454 if (rtinfo.rti_genmask != NULL) { 455 struct radix_node *n; 456 457 #define clen(s) (*(u_char *)(s)) 458 n = rn_addmask((char *)rtinfo.rti_genmask, TRUE, 1); 459 if (n != NULL && 460 rtinfo.rti_genmask->sa_len >= clen(n->rn_key) && 461 bcmp((char *)rtinfo.rti_genmask + 1, 462 (char *)n->rn_key + 1, clen(n->rn_key) - 1) == 0) 463 rtinfo.rti_genmask = (struct sockaddr *)n->rn_key; 464 else 465 gotoerr(ENOBUFS); 466 } 467 468 /* 469 * Verify that the caller has the appropriate privilege; RTM_GET 470 * is the only operation the non-superuser is allowed. 471 */ 472 if (rtm->rtm_type != RTM_GET && suser_cred(so->so_cred, 0) != 0) 473 gotoerr(EPERM); 474 475 switch (rtm->rtm_type) { 476 case RTM_ADD: 477 if (rtinfo.rti_gateway == NULL) 478 gotoerr(EINVAL); 479 error = rtrequest1(RTM_ADD, &rtinfo, &saved_nrt); 480 if (error == 0 && saved_nrt != NULL) { 481 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, 482 &saved_nrt->rt_rmx); 483 saved_nrt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits); 484 saved_nrt->rt_rmx.rmx_locks |= 485 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks); 486 --saved_nrt->rt_refcnt; 487 saved_nrt->rt_genmask = rtinfo.rti_genmask; 488 } 489 break; 490 case RTM_DELETE: 491 error = rtrequest1(RTM_DELETE, &rtinfo, &saved_nrt); 492 if (error == 0) { 493 if ((rt = saved_nrt)) 494 rt->rt_refcnt++; 495 if (fillrtmsg(&rtm, rt, &rtinfo) != 0) 496 gotoerr(ENOBUFS); 497 } 498 break; 499 case RTM_GET: 500 case RTM_CHANGE: 501 case RTM_LOCK: 502 if ((rnh = rt_tables[rtinfo.rti_dst->sa_family]) == NULL) 503 gotoerr(EAFNOSUPPORT); 504 rt = (struct rtentry *) 505 rnh->rnh_lookup((char *)rtinfo.rti_dst, 506 (char *)rtinfo.rti_netmask, rnh); 507 if (rt == NULL) 508 gotoerr(ESRCH); 509 rt->rt_refcnt++; 510 511 switch(rtm->rtm_type) { 512 case RTM_GET: 513 if (fillrtmsg(&rtm, rt, &rtinfo) != 0) 514 gotoerr(ENOBUFS); 515 break; 516 case RTM_CHANGE: 517 /* 518 * new gateway could require new ifaddr, ifp; 519 * flags may also be different; ifp may be specified 520 * by ll sockaddr when protocol address is ambiguous 521 */ 522 if (((rt->rt_flags & RTF_GATEWAY) && 523 rtinfo.rti_gateway != NULL) || 524 rtinfo.rti_ifpaddr != NULL || 525 (rtinfo.rti_ifaaddr != NULL && 526 sa_equal(rtinfo.rti_ifaaddr, 527 rt->rt_ifa->ifa_addr))) { 528 error = rt_getifa(&rtinfo); 529 if (error != 0) 530 gotoerr(error); 531 } 532 if (rtinfo.rti_gateway != NULL) { 533 error = rt_setgate(rt, rt_key(rt), 534 rtinfo.rti_gateway); 535 if (error != 0) 536 gotoerr(error); 537 } 538 if ((ifa = rtinfo.rti_ifa) != NULL) { 539 struct ifaddr *oifa = rt->rt_ifa; 540 541 if (oifa != ifa) { 542 if (oifa && oifa->ifa_rtrequest) 543 oifa->ifa_rtrequest(RTM_DELETE, 544 rt, &rtinfo); 545 IFAFREE(rt->rt_ifa); 546 IFAREF(ifa); 547 rt->rt_ifa = ifa; 548 rt->rt_ifp = rtinfo.rti_ifp; 549 } 550 } 551 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, 552 &rt->rt_rmx); 553 if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest) 554 rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt, &rtinfo); 555 if (rtinfo.rti_genmask != NULL) 556 rt->rt_genmask = rtinfo.rti_genmask; 557 /* 558 * Fall into 559 */ 560 case RTM_LOCK: 561 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits); 562 rt->rt_rmx.rmx_locks |= 563 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks); 564 break; 565 } 566 567 break; 568 default: 569 gotoerr(EOPNOTSUPP); 570 } 571 572 flush: 573 if (rtm != NULL) { 574 if (error != 0) 575 rtm->rtm_errno = error; 576 else 577 rtm->rtm_flags |= RTF_DONE; 578 } 579 if (rt != NULL) 580 rtfree(rt); 581 /* 582 * Check to see if we don't want our own messages. 583 */ 584 if (!(so->so_options & SO_USELOOPBACK)) { 585 if (route_cb.any_count <= 1) { 586 if (rtm != NULL) 587 free(rtm, M_RTABLE); 588 m_freem(m); 589 return (error); 590 } 591 /* There is another listener, so construct message */ 592 rp = sotorawcb(so); 593 } 594 if (rtm != NULL) { 595 m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm); 596 if (m->m_pkthdr.len < rtm->rtm_msglen) { 597 m_freem(m); 598 m = NULL; 599 } else if (m->m_pkthdr.len > rtm->rtm_msglen) 600 m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len); 601 free(rtm, M_RTABLE); 602 } 603 if (rp != NULL) 604 rp->rcb_proto.sp_family = 0; /* Avoid us */ 605 if (m != NULL) 606 rts_input(m, familyof(rtinfo.rti_dst)); 607 if (rp != NULL) 608 rp->rcb_proto.sp_family = PF_ROUTE; 609 return (error); 610 } 611 612 static void 613 rt_setmetrics(u_long which, struct rt_metrics *in, struct rt_metrics *out) 614 { 615 #define setmetric(flag, elt) if (which & (flag)) out->elt = in->elt; 616 setmetric(RTV_RPIPE, rmx_recvpipe); 617 setmetric(RTV_SPIPE, rmx_sendpipe); 618 setmetric(RTV_SSTHRESH, rmx_ssthresh); 619 setmetric(RTV_RTT, rmx_rtt); 620 setmetric(RTV_RTTVAR, rmx_rttvar); 621 setmetric(RTV_HOPCOUNT, rmx_hopcount); 622 setmetric(RTV_MTU, rmx_mtu); 623 setmetric(RTV_EXPIRE, rmx_expire); 624 #undef setmetric 625 } 626 627 #define ROUNDUP(a) \ 628 ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long)) 629 630 /* 631 * Extract the addresses of the passed sockaddrs. 632 * Do a little sanity checking so as to avoid bad memory references. 633 * This data is derived straight from userland. 634 */ 635 static int 636 rt_xaddrs(char *cp, char *cplim, struct rt_addrinfo *rtinfo) 637 { 638 struct sockaddr *sa; 639 int i; 640 641 for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) { 642 if ((rtinfo->rti_addrs & (1 << i)) == 0) 643 continue; 644 sa = (struct sockaddr *)cp; 645 /* 646 * It won't fit. 647 */ 648 if ((cp + sa->sa_len) > cplim) { 649 return (EINVAL); 650 } 651 652 /* 653 * There are no more... Quit now. 654 * If there are more bits, they are in error. 655 * I've seen this. route(1) can evidently generate these. 656 * This causes kernel to core dump. 657 * For compatibility, if we see this, point to a safe address. 658 */ 659 if (sa->sa_len == 0) { 660 static struct sockaddr sa_zero = { 661 sizeof sa_zero, AF_INET, 662 }; 663 664 rtinfo->rti_info[i] = &sa_zero; 665 return (0); /* should be EINVAL but for compat */ 666 } 667 668 /* Accept the sockaddr. */ 669 rtinfo->rti_info[i] = sa; 670 cp += ROUNDUP(sa->sa_len); 671 } 672 return (0); 673 } 674 675 static int 676 rt_msghdrsize(int type) 677 { 678 switch (type) { 679 case RTM_DELADDR: 680 case RTM_NEWADDR: 681 return sizeof(struct ifa_msghdr); 682 case RTM_DELMADDR: 683 case RTM_NEWMADDR: 684 return sizeof(struct ifma_msghdr); 685 case RTM_IFINFO: 686 return sizeof(struct if_msghdr); 687 case RTM_IFANNOUNCE: 688 return sizeof(struct if_announcemsghdr); 689 default: 690 return sizeof(struct rt_msghdr); 691 } 692 } 693 694 static int 695 rt_msgsize(int type, struct rt_addrinfo *rtinfo) 696 { 697 int len, i; 698 699 len = rt_msghdrsize(type); 700 for (i = 0; i < RTAX_MAX; i++) { 701 if (rtinfo->rti_info[i] != NULL) 702 len += ROUNDUP(rtinfo->rti_info[i]->sa_len); 703 } 704 len = ALIGN(len); 705 return len; 706 } 707 708 /* 709 * Build a routing message in a buffer. 710 * Copy the addresses in the rtinfo->rti_info[] sockaddr array 711 * to the end of the buffer after the message header. 712 * 713 * Set the rtinfo->rti_addrs bitmask of addresses present in rtinfo->rti_info[]. 714 * This side-effect can be avoided if we reorder the addrs bitmask field in all 715 * the route messages to line up so we can set it here instead of back in the 716 * calling routine. 717 */ 718 static void 719 rt_msg_buffer(int type, struct rt_addrinfo *rtinfo, void *buf, int msglen) 720 { 721 struct rt_msghdr *rtm; 722 char *cp; 723 int dlen, i; 724 725 rtm = (struct rt_msghdr *) buf; 726 rtm->rtm_version = RTM_VERSION; 727 rtm->rtm_type = type; 728 rtm->rtm_msglen = msglen; 729 730 cp = (char *)buf + rt_msghdrsize(type); 731 rtinfo->rti_addrs = 0; 732 for (i = 0; i < RTAX_MAX; i++) { 733 struct sockaddr *sa; 734 735 if ((sa = rtinfo->rti_info[i]) == NULL) 736 continue; 737 rtinfo->rti_addrs |= (1 << i); 738 dlen = ROUNDUP(sa->sa_len); 739 bcopy(sa, cp, dlen); 740 cp += dlen; 741 } 742 } 743 744 /* 745 * Build a routing message in a mbuf chain. 746 * Copy the addresses in the rtinfo->rti_info[] sockaddr array 747 * to the end of the mbuf after the message header. 748 * 749 * Set the rtinfo->rti_addrs bitmask of addresses present in rtinfo->rti_info[]. 750 * This side-effect can be avoided if we reorder the addrs bitmask field in all 751 * the route messages to line up so we can set it here instead of back in the 752 * calling routine. 753 */ 754 static struct mbuf * 755 rt_msg_mbuf(int type, struct rt_addrinfo *rtinfo) 756 { 757 struct mbuf *m; 758 struct rt_msghdr *rtm; 759 int hlen, len; 760 int i; 761 762 hlen = rt_msghdrsize(type); 763 KASSERT(hlen <= MCLBYTES, ("rt_msg_mbuf: hlen %d doesn't fit", hlen)); 764 765 m = m_gethdr(MB_DONTWAIT, MT_DATA); 766 if (m == NULL) 767 return (NULL); 768 if (hlen > MHLEN) { 769 MCLGET(m, MB_DONTWAIT); 770 if (!(m->m_flags & M_EXT)) { 771 m_free(m); 772 return (NULL); 773 } 774 } 775 m->m_pkthdr.len = m->m_len = hlen; 776 m->m_pkthdr.rcvif = NULL; 777 rtinfo->rti_addrs = 0; 778 len = hlen; 779 for (i = 0; i < RTAX_MAX; i++) { 780 struct sockaddr *sa; 781 int dlen; 782 783 if ((sa = rtinfo->rti_info[i]) == NULL) 784 continue; 785 rtinfo->rti_addrs |= (1 << i); 786 dlen = ROUNDUP(sa->sa_len); 787 m_copyback(m, len, dlen, (caddr_t)sa); /* can grow mbuf chain */ 788 len += dlen; 789 } 790 if (m->m_pkthdr.len != len) { /* one of the m_copyback() calls failed */ 791 m_freem(m); 792 return (NULL); 793 } 794 rtm = mtod(m, struct rt_msghdr *); 795 bzero(rtm, hlen); 796 rtm->rtm_msglen = len; 797 rtm->rtm_version = RTM_VERSION; 798 rtm->rtm_type = type; 799 return (m); 800 } 801 802 /* 803 * This routine is called to generate a message from the routing 804 * socket indicating that a redirect has occurred, a routing lookup 805 * has failed, or that a protocol has detected timeouts to a particular 806 * destination. 807 */ 808 void 809 rt_missmsg(int type, struct rt_addrinfo *rtinfo, int flags, int error) 810 { 811 struct sockaddr *dst = rtinfo->rti_info[RTAX_DST]; 812 struct rt_msghdr *rtm; 813 struct mbuf *m; 814 815 if (route_cb.any_count == 0) 816 return; 817 m = rt_msg_mbuf(type, rtinfo); 818 if (m == NULL) 819 return; 820 rtm = mtod(m, struct rt_msghdr *); 821 rtm->rtm_flags = RTF_DONE | flags; 822 rtm->rtm_errno = error; 823 rtm->rtm_addrs = rtinfo->rti_addrs; 824 rts_input(m, familyof(dst)); 825 } 826 827 void 828 rt_dstmsg(int type, struct sockaddr *dst, int error) 829 { 830 struct rt_msghdr *rtm; 831 struct rt_addrinfo addrs; 832 struct mbuf *m; 833 834 if (route_cb.any_count == 0) 835 return; 836 bzero(&addrs, sizeof(struct rt_addrinfo)); 837 addrs.rti_info[RTAX_DST] = dst; 838 m = rt_msg_mbuf(type, &addrs); 839 if (m == NULL) 840 return; 841 rtm = mtod(m, struct rt_msghdr *); 842 rtm->rtm_flags = RTF_DONE; 843 rtm->rtm_errno = error; 844 rtm->rtm_addrs = addrs.rti_addrs; 845 rts_input(m, familyof(dst)); 846 } 847 848 /* 849 * This routine is called to generate a message from the routing 850 * socket indicating that the status of a network interface has changed. 851 */ 852 void 853 rt_ifmsg(struct ifnet *ifp) 854 { 855 struct if_msghdr *ifm; 856 struct mbuf *m; 857 struct rt_addrinfo rtinfo; 858 859 if (route_cb.any_count == 0) 860 return; 861 bzero(&rtinfo, sizeof(struct rt_addrinfo)); 862 m = rt_msg_mbuf(RTM_IFINFO, &rtinfo); 863 if (m == NULL) 864 return; 865 ifm = mtod(m, struct if_msghdr *); 866 ifm->ifm_index = ifp->if_index; 867 ifm->ifm_flags = (u_short)ifp->if_flags; 868 ifm->ifm_data = ifp->if_data; 869 ifm->ifm_addrs = 0; 870 rts_input(m, 0); 871 } 872 873 static void 874 rt_ifamsg(int cmd, struct ifaddr *ifa) 875 { 876 struct ifa_msghdr *ifam; 877 struct rt_addrinfo rtinfo; 878 struct mbuf *m; 879 struct ifnet *ifp = ifa->ifa_ifp; 880 881 bzero(&rtinfo, sizeof(struct rt_addrinfo)); 882 rtinfo.rti_ifaaddr = ifa->ifa_addr; 883 rtinfo.rti_ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 884 rtinfo.rti_netmask = ifa->ifa_netmask; 885 rtinfo.rti_bcastaddr = ifa->ifa_dstaddr; 886 887 m = rt_msg_mbuf(cmd, &rtinfo); 888 if (m == NULL) 889 return; 890 891 ifam = mtod(m, struct ifa_msghdr *); 892 ifam->ifam_index = ifp->if_index; 893 ifam->ifam_metric = ifa->ifa_metric; 894 ifam->ifam_flags = ifa->ifa_flags; 895 ifam->ifam_addrs = rtinfo.rti_addrs; 896 897 rts_input(m, familyof(ifa->ifa_addr)); 898 } 899 900 void 901 rt_rtmsg(int cmd, struct rtentry *rt, struct ifnet *ifp, int error) 902 { 903 struct rt_msghdr *rtm; 904 struct rt_addrinfo rtinfo; 905 struct mbuf *m; 906 struct sockaddr *dst; 907 908 if (rt == NULL) 909 return; 910 911 bzero(&rtinfo, sizeof(struct rt_addrinfo)); 912 rtinfo.rti_dst = dst = rt_key(rt); 913 rtinfo.rti_gateway = rt->rt_gateway; 914 rtinfo.rti_netmask = rt_mask(rt); 915 if (ifp != NULL) 916 rtinfo.rti_ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 917 rtinfo.rti_ifaaddr = rt->rt_ifa->ifa_addr; 918 919 m = rt_msg_mbuf(cmd, &rtinfo); 920 if (m == NULL) 921 return; 922 923 rtm = mtod(m, struct rt_msghdr *); 924 if (ifp != NULL) 925 rtm->rtm_index = ifp->if_index; 926 rtm->rtm_flags |= rt->rt_flags; 927 rtm->rtm_errno = error; 928 rtm->rtm_addrs = rtinfo.rti_addrs; 929 930 rts_input(m, familyof(dst)); 931 } 932 933 /* 934 * This is called to generate messages from the routing socket 935 * indicating a network interface has had addresses associated with it. 936 * if we ever reverse the logic and replace messages TO the routing 937 * socket indicate a request to configure interfaces, then it will 938 * be unnecessary as the routing socket will automatically generate 939 * copies of it. 940 */ 941 void 942 rt_newaddrmsg(int cmd, struct ifaddr *ifa, int error, struct rtentry *rt) 943 { 944 if (route_cb.any_count == 0) 945 return; 946 947 if (cmd == RTM_ADD) { 948 rt_ifamsg(RTM_NEWADDR, ifa); 949 rt_rtmsg(RTM_ADD, rt, ifa->ifa_ifp, error); 950 } else { 951 KASSERT((cmd == RTM_DELETE), ("unknown cmd %d", cmd)); 952 rt_rtmsg(RTM_DELETE, rt, ifa->ifa_ifp, error); 953 rt_ifamsg(RTM_DELADDR, ifa); 954 } 955 } 956 957 /* 958 * This is the analogue to the rt_newaddrmsg which performs the same 959 * function but for multicast group memberhips. This is easier since 960 * there is no route state to worry about. 961 */ 962 void 963 rt_newmaddrmsg(int cmd, struct ifmultiaddr *ifma) 964 { 965 struct rt_addrinfo rtinfo; 966 struct mbuf *m = NULL; 967 struct ifnet *ifp = ifma->ifma_ifp; 968 struct ifma_msghdr *ifmam; 969 970 if (route_cb.any_count == 0) 971 return; 972 973 bzero(&rtinfo, sizeof(struct rt_addrinfo)); 974 rtinfo.rti_ifaaddr = ifma->ifma_addr; 975 if (ifp != NULL) 976 rtinfo.rti_ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 977 else 978 rtinfo.rti_ifpaddr = NULL; 979 /* 980 * If a link-layer address is present, present it as a ``gateway'' 981 * (similarly to how ARP entries, e.g., are presented). 982 */ 983 rtinfo.rti_gateway = ifma->ifma_lladdr; 984 985 m = rt_msg_mbuf(cmd, &rtinfo); 986 if (m == NULL) 987 return; 988 989 ifmam = mtod(m, struct ifma_msghdr *); 990 ifmam->ifmam_index = ifp->if_index; 991 ifmam->ifmam_addrs = rtinfo.rti_addrs; 992 993 rts_input(m, familyof(ifma->ifma_addr)); 994 } 995 996 /* 997 * This is called to generate routing socket messages indicating 998 * network interface arrival and departure. 999 */ 1000 void 1001 rt_ifannouncemsg(struct ifnet *ifp, int what) 1002 { 1003 struct rt_addrinfo addrinfo; 1004 struct mbuf *m; 1005 struct if_announcemsghdr *ifan; 1006 1007 if (route_cb.any_count == 0) 1008 return; 1009 1010 bzero(&addrinfo, sizeof addrinfo); 1011 m = rt_msg_mbuf(RTM_IFANNOUNCE, &addrinfo); 1012 if (m == NULL) 1013 return; 1014 1015 ifan = mtod(m, struct if_announcemsghdr *); 1016 ifan->ifan_index = ifp->if_index; 1017 strlcpy(ifan->ifan_name, ifp->if_xname, sizeof ifan->ifan_name); 1018 ifan->ifan_what = what; 1019 1020 rts_input(m, 0); 1021 } 1022 1023 static int 1024 resizewalkarg(struct walkarg *w, int len) 1025 { 1026 void *newptr; 1027 1028 newptr = malloc(len, M_RTABLE, M_INTWAIT | M_NULLOK); 1029 if (newptr == NULL) 1030 return (ENOMEM); 1031 if (w->w_tmem != NULL) 1032 free(w->w_tmem, M_RTABLE); 1033 w->w_tmem = newptr; 1034 w->w_tmemsize = len; 1035 return (0); 1036 } 1037 1038 /* 1039 * This is used in dumping the kernel table via sysctl(). 1040 */ 1041 int 1042 sysctl_dumpentry(struct radix_node *rn, void *vw) 1043 { 1044 struct walkarg *w = vw; 1045 struct rtentry *rt = (struct rtentry *)rn; 1046 struct rt_addrinfo rtinfo; 1047 int error, msglen; 1048 1049 if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg)) 1050 return 0; 1051 1052 bzero(&rtinfo, sizeof(struct rt_addrinfo)); 1053 rtinfo.rti_dst = rt_key(rt); 1054 rtinfo.rti_gateway = rt->rt_gateway; 1055 rtinfo.rti_netmask = rt_mask(rt); 1056 rtinfo.rti_genmask = rt->rt_genmask; 1057 if (rt->rt_ifp != NULL) { 1058 rtinfo.rti_ifpaddr = 1059 TAILQ_FIRST(&rt->rt_ifp->if_addrhead)->ifa_addr; 1060 rtinfo.rti_ifaaddr = rt->rt_ifa->ifa_addr; 1061 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT) 1062 rtinfo.rti_bcastaddr = rt->rt_ifa->ifa_dstaddr; 1063 } 1064 msglen = rt_msgsize(RTM_GET, &rtinfo); 1065 if (w->w_tmemsize < msglen && resizewalkarg(w, msglen) != 0) 1066 return (ENOMEM); 1067 rt_msg_buffer(RTM_GET, &rtinfo, w->w_tmem, msglen); 1068 if (w->w_req != NULL) { 1069 struct rt_msghdr *rtm = (struct rt_msghdr *)w->w_tmem; 1070 1071 rtm->rtm_flags = rt->rt_flags; 1072 rtm->rtm_use = rt->rt_use; 1073 rtm->rtm_rmx = rt->rt_rmx; 1074 rtm->rtm_index = rt->rt_ifp->if_index; 1075 rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0; 1076 rtm->rtm_addrs = rtinfo.rti_addrs; 1077 error = SYSCTL_OUT(w->w_req, rtm, msglen); 1078 return (error); 1079 } 1080 return (0); 1081 } 1082 1083 static int 1084 sysctl_iflist(int af, struct walkarg *w) 1085 { 1086 struct ifnet *ifp; 1087 struct ifaddr *ifa; 1088 struct rt_addrinfo rtinfo; 1089 int msglen, error; 1090 1091 bzero(&rtinfo, sizeof(struct rt_addrinfo)); 1092 TAILQ_FOREACH(ifp, &ifnet, if_link) { 1093 if (w->w_arg && w->w_arg != ifp->if_index) 1094 continue; 1095 ifa = TAILQ_FIRST(&ifp->if_addrhead); 1096 rtinfo.rti_ifpaddr = ifa->ifa_addr; 1097 msglen = rt_msgsize(RTM_IFINFO, &rtinfo); 1098 if (w->w_tmemsize < msglen && resizewalkarg(w, msglen) != 0) 1099 return (ENOMEM); 1100 rt_msg_buffer(RTM_IFINFO, &rtinfo, w->w_tmem, msglen); 1101 rtinfo.rti_ifpaddr = NULL; 1102 if (w->w_req != NULL && w->w_tmem != NULL) { 1103 struct if_msghdr *ifm; 1104 1105 ifm = (struct if_msghdr *)w->w_tmem; 1106 ifm->ifm_index = ifp->if_index; 1107 ifm->ifm_flags = (u_short)ifp->if_flags; 1108 ifm->ifm_data = ifp->if_data; 1109 ifm->ifm_addrs = rtinfo.rti_addrs; 1110 error = SYSCTL_OUT(w->w_req, ifm, msglen); 1111 if (error) 1112 return (error); 1113 } 1114 while ((ifa = TAILQ_NEXT(ifa, ifa_link)) != NULL) { 1115 if (af && af != ifa->ifa_addr->sa_family) 1116 continue; 1117 if (curproc->p_ucred->cr_prison && 1118 prison_if(curthread, ifa->ifa_addr)) 1119 continue; 1120 rtinfo.rti_ifaaddr = ifa->ifa_addr; 1121 rtinfo.rti_netmask = ifa->ifa_netmask; 1122 rtinfo.rti_bcastaddr = ifa->ifa_dstaddr; 1123 msglen = rt_msgsize(RTM_NEWADDR, &rtinfo); 1124 if (w->w_tmemsize < msglen && 1125 resizewalkarg(w, msglen) != 0) 1126 return (ENOMEM); 1127 rt_msg_buffer(RTM_NEWADDR, &rtinfo, w->w_tmem, msglen); 1128 if (w->w_req != NULL) { 1129 struct ifa_msghdr *ifam; 1130 1131 ifam = (struct ifa_msghdr *)w->w_tmem; 1132 ifam->ifam_index = ifa->ifa_ifp->if_index; 1133 ifam->ifam_flags = ifa->ifa_flags; 1134 ifam->ifam_metric = ifa->ifa_metric; 1135 ifam->ifam_addrs = rtinfo.rti_addrs; 1136 error = SYSCTL_OUT(w->w_req, w->w_tmem, msglen); 1137 if (error) 1138 return (error); 1139 } 1140 } 1141 rtinfo.rti_netmask = NULL; 1142 rtinfo.rti_ifaaddr = NULL; 1143 rtinfo.rti_bcastaddr = NULL; 1144 } 1145 return (0); 1146 } 1147 1148 static int 1149 sysctl_rtsock(SYSCTL_HANDLER_ARGS) 1150 { 1151 int *name = (int *)arg1; 1152 u_int namelen = arg2; 1153 struct radix_node_head *rnh; 1154 int i, s, error = EINVAL; 1155 u_char af; 1156 struct walkarg w; 1157 1158 name ++; 1159 namelen--; 1160 if (req->newptr) 1161 return (EPERM); 1162 if (namelen != 3) 1163 return (EINVAL); 1164 af = name[0]; 1165 bzero(&w, sizeof w); 1166 w.w_op = name[1]; 1167 w.w_arg = name[2]; 1168 w.w_req = req; 1169 1170 s = splnet(); 1171 switch (w.w_op) { 1172 1173 case NET_RT_DUMP: 1174 case NET_RT_FLAGS: 1175 for (i = 1; i <= AF_MAX; i++) 1176 if ((rnh = rt_tables[i]) && (af == 0 || af == i) && 1177 (error = rnh->rnh_walktree(rnh, 1178 sysctl_dumpentry, &w))) 1179 break; 1180 break; 1181 1182 case NET_RT_IFLIST: 1183 error = sysctl_iflist(af, &w); 1184 } 1185 splx(s); 1186 if (w.w_tmem != NULL) 1187 free(w.w_tmem, M_RTABLE); 1188 return (error); 1189 } 1190 1191 SYSCTL_NODE(_net, PF_ROUTE, routetable, CTLFLAG_RD, sysctl_rtsock, ""); 1192 1193 /* 1194 * Definitions of protocols supported in the ROUTE domain. 1195 */ 1196 1197 extern struct domain routedomain; /* or at least forward */ 1198 1199 static struct protosw routesw[] = { 1200 { SOCK_RAW, &routedomain, 0, PR_ATOMIC|PR_ADDR, 1201 0, route_output, raw_ctlinput, 0, 1202 cpu0_soport, 1203 raw_init, 0, 0, 0, 1204 &route_usrreqs 1205 } 1206 }; 1207 1208 static struct domain routedomain = 1209 { PF_ROUTE, "route", 0, 0, 0, 1210 routesw, &routesw[(sizeof routesw)/(sizeof routesw[0])] }; 1211 1212 DOMAIN_SET(route); 1213