1 /* $NetBSD: rtsock.c,v 1.72 2004/10/23 19:13:22 christos Exp $ */ 2 3 /* 4 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. 5 * All rights reserved. 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 project nor the names of its contributors 16 * may be used to endorse or promote products derived from this software 17 * without specific prior written permission. 18 * 19 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND 20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 22 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE 23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 29 * SUCH DAMAGE. 30 */ 31 32 /* 33 * Copyright (c) 1988, 1991, 1993 34 * The Regents of the University of California. All rights reserved. 35 * 36 * Redistribution and use in source and binary forms, with or without 37 * modification, are permitted provided that the following conditions 38 * are met: 39 * 1. Redistributions of source code must retain the above copyright 40 * notice, this list of conditions and the following disclaimer. 41 * 2. Redistributions in binary form must reproduce the above copyright 42 * notice, this list of conditions and the following disclaimer in the 43 * documentation and/or other materials provided with the distribution. 44 * 3. Neither the name of the University nor the names of its contributors 45 * may be used to endorse or promote products derived from this software 46 * without specific prior written permission. 47 * 48 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 49 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 50 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 51 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 52 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 53 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 54 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 55 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 56 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 57 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 58 * SUCH DAMAGE. 59 * 60 * @(#)rtsock.c 8.7 (Berkeley) 10/12/95 61 */ 62 63 #include <sys/cdefs.h> 64 __KERNEL_RCSID(0, "$NetBSD: rtsock.c,v 1.72 2004/10/23 19:13:22 christos Exp $"); 65 66 #include "opt_inet.h" 67 68 #include <sys/param.h> 69 #include <sys/systm.h> 70 #include <sys/proc.h> 71 #include <sys/mbuf.h> 72 #include <sys/socket.h> 73 #include <sys/socketvar.h> 74 #include <sys/domain.h> 75 #include <sys/protosw.h> 76 #include <sys/sysctl.h> 77 78 #include <net/if.h> 79 #include <net/route.h> 80 #include <net/raw_cb.h> 81 82 #include <machine/stdarg.h> 83 84 extern struct domain routedomain; /* or at least forward */ 85 86 struct sockaddr route_dst = { 2, PF_ROUTE, }; 87 struct sockaddr route_src = { 2, PF_ROUTE, }; 88 struct sockproto route_proto = { PF_ROUTE, }; 89 90 struct walkarg { 91 int w_op; 92 int w_arg; 93 int w_given; 94 int w_needed; 95 caddr_t w_where; 96 int w_tmemsize; 97 int w_tmemneeded; 98 caddr_t w_tmem; 99 }; 100 101 static struct mbuf *rt_msg1(int, struct rt_addrinfo *, caddr_t, int); 102 static int rt_msg2(int, struct rt_addrinfo *, caddr_t, struct walkarg *, int *); 103 static int rt_xaddrs(u_char, const char *, const char *, struct rt_addrinfo *); 104 static int sysctl_dumpentry(struct radix_node *, void *); 105 static int sysctl_iflist(int, struct walkarg *, int); 106 static int sysctl_rtable(SYSCTLFN_PROTO); 107 static __inline void rt_adjustcount(int, int); 108 109 /* Sleazy use of local variables throughout file, warning!!!! */ 110 #define dst info.rti_info[RTAX_DST] 111 #define gate info.rti_info[RTAX_GATEWAY] 112 #define netmask info.rti_info[RTAX_NETMASK] 113 #define genmask info.rti_info[RTAX_GENMASK] 114 #define ifpaddr info.rti_info[RTAX_IFP] 115 #define ifaaddr info.rti_info[RTAX_IFA] 116 #define brdaddr info.rti_info[RTAX_BRD] 117 118 static __inline void 119 rt_adjustcount(int af, int cnt) 120 { 121 route_cb.any_count += cnt; 122 switch (af) { 123 case AF_INET: 124 route_cb.ip_count += cnt; 125 return; 126 #ifdef INET6 127 case AF_INET6: 128 route_cb.ip6_count += cnt; 129 return; 130 #endif 131 case AF_IPX: 132 route_cb.ipx_count += cnt; 133 return; 134 case AF_NS: 135 route_cb.ns_count += cnt; 136 return; 137 case AF_ISO: 138 route_cb.iso_count += cnt; 139 return; 140 } 141 } 142 143 /*ARGSUSED*/ 144 int 145 route_usrreq(struct socket *so, int req, struct mbuf *m, struct mbuf *nam, 146 struct mbuf *control, struct proc *p) 147 { 148 int error = 0; 149 struct rawcb *rp = sotorawcb(so); 150 int s; 151 152 if (req == PRU_ATTACH) { 153 MALLOC(rp, struct rawcb *, sizeof(*rp), M_PCB, M_WAITOK); 154 if ((so->so_pcb = rp) != NULL) 155 memset(so->so_pcb, 0, sizeof(*rp)); 156 157 } 158 if (req == PRU_DETACH && rp) 159 rt_adjustcount(rp->rcb_proto.sp_protocol, -1); 160 s = splsoftnet(); 161 162 /* 163 * Don't call raw_usrreq() in the attach case, because 164 * we want to allow non-privileged processes to listen on 165 * and send "safe" commands to the routing socket. 166 */ 167 if (req == PRU_ATTACH) { 168 if (p == 0) 169 error = EACCES; 170 else 171 error = raw_attach(so, (int)(long)nam); 172 } else 173 error = raw_usrreq(so, req, m, nam, control, p); 174 175 rp = sotorawcb(so); 176 if (req == PRU_ATTACH && rp) { 177 if (error) { 178 free((caddr_t)rp, M_PCB); 179 splx(s); 180 return (error); 181 } 182 rt_adjustcount(rp->rcb_proto.sp_protocol, 1); 183 rp->rcb_laddr = &route_src; 184 rp->rcb_faddr = &route_dst; 185 soisconnected(so); 186 so->so_options |= SO_USELOOPBACK; 187 } 188 splx(s); 189 return (error); 190 } 191 192 /*ARGSUSED*/ 193 int 194 route_output(struct mbuf *m, ...) 195 { 196 struct rt_msghdr *rtm = 0; 197 struct radix_node *rn = 0; 198 struct rtentry *rt = 0; 199 struct rtentry *saved_nrt = 0; 200 struct radix_node_head *rnh; 201 struct rt_addrinfo info; 202 int len, error = 0; 203 struct ifnet *ifp = 0; 204 struct ifaddr *ifa = 0; 205 struct socket *so; 206 va_list ap; 207 sa_family_t family; 208 209 va_start(ap, m); 210 so = va_arg(ap, struct socket *); 211 va_end(ap); 212 213 #define senderr(e) do { error = e; goto flush;} while (/*CONSTCOND*/ 0) 214 if (m == 0 || ((m->m_len < sizeof(int32_t)) && 215 (m = m_pullup(m, sizeof(int32_t))) == 0)) 216 return (ENOBUFS); 217 if ((m->m_flags & M_PKTHDR) == 0) 218 panic("route_output"); 219 len = m->m_pkthdr.len; 220 if (len < sizeof(*rtm) || 221 len != mtod(m, struct rt_msghdr *)->rtm_msglen) { 222 dst = 0; 223 senderr(EINVAL); 224 } 225 R_Malloc(rtm, struct rt_msghdr *, len); 226 if (rtm == 0) { 227 dst = 0; 228 senderr(ENOBUFS); 229 } 230 m_copydata(m, 0, len, (caddr_t)rtm); 231 if (rtm->rtm_version != RTM_VERSION) { 232 dst = 0; 233 senderr(EPROTONOSUPPORT); 234 } 235 rtm->rtm_pid = curproc->p_pid; 236 memset(&info, 0, sizeof(info)); 237 info.rti_addrs = rtm->rtm_addrs; 238 if (rt_xaddrs(rtm->rtm_type, (caddr_t)(rtm + 1), len + (caddr_t)rtm, &info)) 239 senderr(EINVAL); 240 info.rti_flags = rtm->rtm_flags; 241 if (dst == 0 || (dst->sa_family >= AF_MAX)) 242 senderr(EINVAL); 243 if (gate != 0 && (gate->sa_family >= AF_MAX)) 244 senderr(EINVAL); 245 if (genmask) { 246 struct radix_node *t; 247 t = rn_addmask((caddr_t)genmask, 0, 1); 248 if (t && genmask->sa_len >= ((struct sockaddr *)t->rn_key)->sa_len && 249 Bcmp((caddr_t *)genmask + 1, (caddr_t *)t->rn_key + 1, 250 ((struct sockaddr *)t->rn_key)->sa_len) - 1) 251 genmask = (struct sockaddr *)(t->rn_key); 252 else 253 senderr(ENOBUFS); 254 } 255 256 /* 257 * Verify that the caller has the appropriate privilege; RTM_GET 258 * is the only operation the non-superuser is allowed. 259 */ 260 if (rtm->rtm_type != RTM_GET && 261 suser(curproc->p_ucred, &curproc->p_acflag) != 0) 262 senderr(EACCES); 263 264 switch (rtm->rtm_type) { 265 266 case RTM_ADD: 267 if (gate == 0) 268 senderr(EINVAL); 269 error = rtrequest1(rtm->rtm_type, &info, &saved_nrt); 270 if (error == 0 && saved_nrt) { 271 rt_setmetrics(rtm->rtm_inits, 272 &rtm->rtm_rmx, &saved_nrt->rt_rmx); 273 saved_nrt->rt_refcnt--; 274 saved_nrt->rt_genmask = genmask; 275 } 276 break; 277 278 case RTM_DELETE: 279 error = rtrequest1(rtm->rtm_type, &info, &saved_nrt); 280 if (error == 0) { 281 (rt = saved_nrt)->rt_refcnt++; 282 goto report; 283 } 284 break; 285 286 case RTM_GET: 287 case RTM_CHANGE: 288 case RTM_LOCK: 289 if ((rnh = rt_tables[dst->sa_family]) == 0) { 290 senderr(EAFNOSUPPORT); 291 } 292 rn = rnh->rnh_lookup(dst, netmask, rnh); 293 if (rn == NULL || (rn->rn_flags & RNF_ROOT) != 0) { 294 senderr(ESRCH); 295 } 296 rt = (struct rtentry *)rn; 297 rt->rt_refcnt++; 298 if (rtm->rtm_type != RTM_GET) {/* XXX: too grotty */ 299 struct radix_node *rn; 300 extern struct radix_node_head *mask_rnhead; 301 302 if (Bcmp(dst, rt_key(rt), dst->sa_len) != 0) 303 senderr(ESRCH); 304 if (netmask && (rn = rn_search(netmask, 305 mask_rnhead->rnh_treetop))) 306 netmask = (struct sockaddr *)rn->rn_key; 307 for (rn = rt->rt_nodes; rn; rn = rn->rn_dupedkey) 308 if (netmask == (struct sockaddr *)rn->rn_mask) 309 break; 310 if (rn == 0) 311 senderr(ETOOMANYREFS); 312 rt = (struct rtentry *)rn; 313 } 314 315 switch (rtm->rtm_type) { 316 case RTM_GET: 317 report: 318 dst = rt_key(rt); 319 gate = rt->rt_gateway; 320 netmask = rt_mask(rt); 321 genmask = rt->rt_genmask; 322 if (rtm->rtm_addrs & (RTA_IFP | RTA_IFA)) { 323 if ((ifp = rt->rt_ifp) != NULL) { 324 ifpaddr = TAILQ_FIRST(&ifp->if_addrlist)->ifa_addr; 325 ifaaddr = rt->rt_ifa->ifa_addr; 326 if (ifp->if_flags & IFF_POINTOPOINT) 327 brdaddr = rt->rt_ifa->ifa_dstaddr; 328 else 329 brdaddr = 0; 330 rtm->rtm_index = ifp->if_index; 331 } else { 332 ifpaddr = 0; 333 ifaaddr = 0; 334 } 335 } 336 (void)rt_msg2(rtm->rtm_type, &info, (caddr_t)0, 337 (struct walkarg *)0, &len); 338 if (len > rtm->rtm_msglen) { 339 struct rt_msghdr *new_rtm; 340 R_Malloc(new_rtm, struct rt_msghdr *, len); 341 if (new_rtm == 0) 342 senderr(ENOBUFS); 343 Bcopy(rtm, new_rtm, rtm->rtm_msglen); 344 Free(rtm); rtm = new_rtm; 345 } 346 (void)rt_msg2(rtm->rtm_type, &info, (caddr_t)rtm, 347 (struct walkarg *)0, 0); 348 rtm->rtm_flags = rt->rt_flags; 349 rtm->rtm_rmx = rt->rt_rmx; 350 rtm->rtm_addrs = info.rti_addrs; 351 break; 352 353 case RTM_CHANGE: 354 /* 355 * new gateway could require new ifaddr, ifp; 356 * flags may also be different; ifp may be specified 357 * by ll sockaddr when protocol address is ambiguous 358 */ 359 if ((error = rt_getifa(&info)) != 0) 360 senderr(error); 361 if (gate && rt_setgate(rt, rt_key(rt), gate)) 362 senderr(EDQUOT); 363 /* new gateway could require new ifaddr, ifp; 364 flags may also be different; ifp may be specified 365 by ll sockaddr when protocol address is ambiguous */ 366 if (ifpaddr && (ifa = ifa_ifwithnet(ifpaddr)) && 367 (ifp = ifa->ifa_ifp) && (ifaaddr || gate)) 368 ifa = ifaof_ifpforaddr(ifaaddr ? ifaaddr : gate, 369 ifp); 370 else if ((ifaaddr && (ifa = ifa_ifwithaddr(ifaaddr))) || 371 (gate && (ifa = ifa_ifwithroute(rt->rt_flags, 372 rt_key(rt), gate)))) 373 ifp = ifa->ifa_ifp; 374 if (ifa) { 375 struct ifaddr *oifa = rt->rt_ifa; 376 if (oifa != ifa) { 377 if (oifa && oifa->ifa_rtrequest) 378 oifa->ifa_rtrequest(RTM_DELETE, rt, 379 &info); 380 IFAFREE(rt->rt_ifa); 381 rt->rt_ifa = ifa; 382 IFAREF(rt->rt_ifa); 383 rt->rt_ifp = ifp; 384 } 385 } 386 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, 387 &rt->rt_rmx); 388 if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest) 389 rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt, &info); 390 if (genmask) 391 rt->rt_genmask = genmask; 392 /* 393 * Fall into 394 */ 395 case RTM_LOCK: 396 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits); 397 rt->rt_rmx.rmx_locks |= 398 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks); 399 break; 400 } 401 break; 402 403 default: 404 senderr(EOPNOTSUPP); 405 } 406 407 flush: 408 if (rtm) { 409 if (error) 410 rtm->rtm_errno = error; 411 else 412 rtm->rtm_flags |= RTF_DONE; 413 } 414 family = dst ? dst->sa_family : 0; 415 if (rt) 416 rtfree(rt); 417 { 418 struct rawcb *rp = 0; 419 /* 420 * Check to see if we don't want our own messages. 421 */ 422 if ((so->so_options & SO_USELOOPBACK) == 0) { 423 if (route_cb.any_count <= 1) { 424 if (rtm) 425 Free(rtm); 426 m_freem(m); 427 return (error); 428 } 429 /* There is another listener, so construct message */ 430 rp = sotorawcb(so); 431 } 432 if (rtm) { 433 m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm); 434 if (m->m_pkthdr.len < rtm->rtm_msglen) { 435 m_freem(m); 436 m = NULL; 437 } else if (m->m_pkthdr.len > rtm->rtm_msglen) 438 m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len); 439 Free(rtm); 440 } 441 if (rp) 442 rp->rcb_proto.sp_family = 0; /* Avoid us */ 443 if (family) 444 route_proto.sp_protocol = family; 445 if (m) 446 raw_input(m, &route_proto, &route_src, &route_dst); 447 if (rp) 448 rp->rcb_proto.sp_family = PF_ROUTE; 449 } 450 return (error); 451 } 452 453 void 454 rt_setmetrics(u_long which, const struct rt_metrics *in, struct rt_metrics *out) 455 { 456 #define metric(f, e) if (which & (f)) out->e = in->e; 457 metric(RTV_RPIPE, rmx_recvpipe); 458 metric(RTV_SPIPE, rmx_sendpipe); 459 metric(RTV_SSTHRESH, rmx_ssthresh); 460 metric(RTV_RTT, rmx_rtt); 461 metric(RTV_RTTVAR, rmx_rttvar); 462 metric(RTV_HOPCOUNT, rmx_hopcount); 463 metric(RTV_MTU, rmx_mtu); 464 metric(RTV_EXPIRE, rmx_expire); 465 #undef metric 466 } 467 468 #define ROUNDUP(a) \ 469 ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long)) 470 #define ADVANCE(x, n) (x += ROUNDUP((n)->sa_len)) 471 472 static int 473 rt_xaddrs(u_char rtmtype, const char *cp, const char *cplim, struct rt_addrinfo *rtinfo) 474 { 475 const struct sockaddr *sa = NULL; /* Quell compiler warning */ 476 int i; 477 478 for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) { 479 if ((rtinfo->rti_addrs & (1 << i)) == 0) 480 continue; 481 rtinfo->rti_info[i] = sa = (struct sockaddr *)cp; 482 ADVANCE(cp, sa); 483 } 484 485 /* Check for extra addresses specified, except RTM_GET asking for interface info. */ 486 if (rtmtype == RTM_GET) { 487 if (((rtinfo->rti_addrs & (~((1 << RTAX_IFP) | (1 << RTAX_IFA)))) & (~0 << i)) != 0) 488 return (1); 489 } else { 490 if ((rtinfo->rti_addrs & (~0 << i)) != 0) 491 return (1); 492 } 493 /* Check for bad data length. */ 494 if (cp != cplim) { 495 if (i == RTAX_NETMASK + 1 && 496 cp - ROUNDUP(sa->sa_len) + sa->sa_len == cplim) 497 /* 498 * The last sockaddr was netmask. 499 * We accept this for now for the sake of old 500 * binaries or third party softwares. 501 */ 502 ; 503 else 504 return (1); 505 } 506 return (0); 507 } 508 509 static struct mbuf * 510 rt_msg1(int type, struct rt_addrinfo *rtinfo, caddr_t data, int datalen) 511 { 512 struct rt_msghdr *rtm; 513 struct mbuf *m; 514 int i; 515 const struct sockaddr *sa; 516 int len, dlen; 517 518 m = m_gethdr(M_DONTWAIT, MT_DATA); 519 if (m == 0) 520 return (m); 521 MCLAIM(m, &routedomain.dom_mowner); 522 switch (type) { 523 524 case RTM_DELADDR: 525 case RTM_NEWADDR: 526 len = sizeof(struct ifa_msghdr); 527 break; 528 529 #ifdef COMPAT_14 530 case RTM_OIFINFO: 531 len = sizeof(struct if_msghdr14); 532 break; 533 #endif 534 535 case RTM_IFINFO: 536 len = sizeof(struct if_msghdr); 537 break; 538 539 case RTM_IFANNOUNCE: 540 len = sizeof(struct if_announcemsghdr); 541 break; 542 543 default: 544 len = sizeof(struct rt_msghdr); 545 } 546 if (len > MHLEN + MLEN) 547 panic("rt_msg1: message too long"); 548 else if (len > MHLEN) { 549 m->m_next = m_get(M_DONTWAIT, MT_DATA); 550 if (m->m_next == NULL) { 551 m_freem(m); 552 return (NULL); 553 } 554 MCLAIM(m->m_next, m->m_owner); 555 m->m_pkthdr.len = len; 556 m->m_len = MHLEN; 557 m->m_next->m_len = len - MHLEN; 558 } else { 559 m->m_pkthdr.len = m->m_len = len; 560 } 561 m->m_pkthdr.rcvif = 0; 562 m_copyback(m, 0, datalen, data); 563 rtm = mtod(m, struct rt_msghdr *); 564 for (i = 0; i < RTAX_MAX; i++) { 565 if ((sa = rtinfo->rti_info[i]) == NULL) 566 continue; 567 rtinfo->rti_addrs |= (1 << i); 568 dlen = ROUNDUP(sa->sa_len); 569 m_copyback(m, len, dlen, (caddr_t)sa); 570 len += dlen; 571 } 572 if (m->m_pkthdr.len != len) { 573 m_freem(m); 574 return (NULL); 575 } 576 rtm->rtm_msglen = len; 577 rtm->rtm_version = RTM_VERSION; 578 rtm->rtm_type = type; 579 return (m); 580 } 581 582 /* 583 * rt_msg2 584 * 585 * fills 'cp' or 'w'.w_tmem with the routing socket message and 586 * returns the length of the message in 'lenp'. 587 * 588 * if walkarg is 0, cp is expected to be 0 or a buffer large enough to hold 589 * the message 590 * otherwise walkarg's w_needed is updated and if the user buffer is 591 * specified and w_needed indicates space exists the information is copied 592 * into the temp space (w_tmem). w_tmem is [re]allocated if necessary, 593 * if the allocation fails ENOBUFS is returned. 594 */ 595 static int 596 rt_msg2(int type, struct rt_addrinfo *rtinfo, caddr_t cp, struct walkarg *w, 597 int *lenp) 598 { 599 int i; 600 int len, dlen, second_time = 0; 601 caddr_t cp0; 602 603 rtinfo->rti_addrs = 0; 604 again: 605 switch (type) { 606 607 case RTM_DELADDR: 608 case RTM_NEWADDR: 609 len = sizeof(struct ifa_msghdr); 610 break; 611 #ifdef COMPAT_14 612 case RTM_OIFINFO: 613 len = sizeof(struct if_msghdr14); 614 break; 615 #endif 616 617 case RTM_IFINFO: 618 len = sizeof(struct if_msghdr); 619 break; 620 621 default: 622 len = sizeof(struct rt_msghdr); 623 } 624 if ((cp0 = cp) != NULL) 625 cp += len; 626 for (i = 0; i < RTAX_MAX; i++) { 627 const struct sockaddr *sa; 628 629 if ((sa = rtinfo->rti_info[i]) == 0) 630 continue; 631 rtinfo->rti_addrs |= (1 << i); 632 dlen = ROUNDUP(sa->sa_len); 633 if (cp) { 634 bcopy(sa, cp, (unsigned)dlen); 635 cp += dlen; 636 } 637 len += dlen; 638 } 639 if (cp == 0 && w != NULL && !second_time) { 640 struct walkarg *rw = w; 641 642 rw->w_needed += len; 643 if (rw->w_needed <= 0 && rw->w_where) { 644 if (rw->w_tmemsize < len) { 645 if (rw->w_tmem) 646 free(rw->w_tmem, M_RTABLE); 647 rw->w_tmem = (caddr_t) malloc(len, M_RTABLE, 648 M_NOWAIT); 649 if (rw->w_tmem) 650 rw->w_tmemsize = len; 651 } 652 if (rw->w_tmem) { 653 cp = rw->w_tmem; 654 second_time = 1; 655 goto again; 656 } else { 657 rw->w_tmemneeded = len; 658 return (ENOBUFS); 659 } 660 } 661 } 662 if (cp) { 663 struct rt_msghdr *rtm = (struct rt_msghdr *)cp0; 664 665 rtm->rtm_version = RTM_VERSION; 666 rtm->rtm_type = type; 667 rtm->rtm_msglen = len; 668 } 669 if (lenp) 670 *lenp = len; 671 return (0); 672 } 673 674 /* 675 * This routine is called to generate a message from the routing 676 * socket indicating that a redirect has occurred, a routing lookup 677 * has failed, or that a protocol has detected timeouts to a particular 678 * destination. 679 */ 680 void 681 rt_missmsg(int type, struct rt_addrinfo *rtinfo, int flags, int error) 682 { 683 struct rt_msghdr rtm; 684 struct mbuf *m; 685 const struct sockaddr *sa = rtinfo->rti_info[RTAX_DST]; 686 687 if (route_cb.any_count == 0) 688 return; 689 memset(&rtm, 0, sizeof(rtm)); 690 rtm.rtm_flags = RTF_DONE | flags; 691 rtm.rtm_errno = error; 692 m = rt_msg1(type, rtinfo, (caddr_t)&rtm, sizeof(rtm)); 693 if (m == 0) 694 return; 695 mtod(m, struct rt_msghdr *)->rtm_addrs = rtinfo->rti_addrs; 696 route_proto.sp_protocol = sa ? sa->sa_family : 0; 697 raw_input(m, &route_proto, &route_src, &route_dst); 698 } 699 700 /* 701 * This routine is called to generate a message from the routing 702 * socket indicating that the status of a network interface has changed. 703 */ 704 void 705 rt_ifmsg(struct ifnet *ifp) 706 { 707 struct if_msghdr ifm; 708 #ifdef COMPAT_14 709 struct if_msghdr14 oifm; 710 #endif 711 struct mbuf *m; 712 struct rt_addrinfo info; 713 714 if (route_cb.any_count == 0) 715 return; 716 memset(&info, 0, sizeof(info)); 717 memset(&ifm, 0, sizeof(ifm)); 718 ifm.ifm_index = ifp->if_index; 719 ifm.ifm_flags = ifp->if_flags; 720 ifm.ifm_data = ifp->if_data; 721 ifm.ifm_addrs = 0; 722 m = rt_msg1(RTM_IFINFO, &info, (caddr_t)&ifm, sizeof(ifm)); 723 if (m == 0) 724 return; 725 route_proto.sp_protocol = 0; 726 raw_input(m, &route_proto, &route_src, &route_dst); 727 #ifdef COMPAT_14 728 memset(&info, 0, sizeof(info)); 729 memset(&oifm, 0, sizeof(oifm)); 730 oifm.ifm_index = ifp->if_index; 731 oifm.ifm_flags = ifp->if_flags; 732 oifm.ifm_data.ifi_type = ifp->if_data.ifi_type; 733 oifm.ifm_data.ifi_addrlen = ifp->if_data.ifi_addrlen; 734 oifm.ifm_data.ifi_hdrlen = ifp->if_data.ifi_hdrlen; 735 oifm.ifm_data.ifi_mtu = ifp->if_data.ifi_mtu; 736 oifm.ifm_data.ifi_metric = ifp->if_data.ifi_metric; 737 oifm.ifm_data.ifi_baudrate = ifp->if_data.ifi_baudrate; 738 oifm.ifm_data.ifi_ipackets = ifp->if_data.ifi_ipackets; 739 oifm.ifm_data.ifi_ierrors = ifp->if_data.ifi_ierrors; 740 oifm.ifm_data.ifi_opackets = ifp->if_data.ifi_opackets; 741 oifm.ifm_data.ifi_oerrors = ifp->if_data.ifi_oerrors; 742 oifm.ifm_data.ifi_collisions = ifp->if_data.ifi_collisions; 743 oifm.ifm_data.ifi_ibytes = ifp->if_data.ifi_ibytes; 744 oifm.ifm_data.ifi_obytes = ifp->if_data.ifi_obytes; 745 oifm.ifm_data.ifi_imcasts = ifp->if_data.ifi_imcasts; 746 oifm.ifm_data.ifi_omcasts = ifp->if_data.ifi_omcasts; 747 oifm.ifm_data.ifi_iqdrops = ifp->if_data.ifi_iqdrops; 748 oifm.ifm_data.ifi_noproto = ifp->if_data.ifi_noproto; 749 oifm.ifm_data.ifi_lastchange = ifp->if_data.ifi_lastchange; 750 oifm.ifm_addrs = 0; 751 m = rt_msg1(RTM_OIFINFO, &info, (caddr_t)&oifm, sizeof(oifm)); 752 if (m == 0) 753 return; 754 route_proto.sp_protocol = 0; 755 raw_input(m, &route_proto, &route_src, &route_dst); 756 #endif 757 } 758 759 /* 760 * This is called to generate messages from the routing socket 761 * indicating a network interface has had addresses associated with it. 762 * if we ever reverse the logic and replace messages TO the routing 763 * socket indicate a request to configure interfaces, then it will 764 * be unnecessary as the routing socket will automatically generate 765 * copies of it. 766 */ 767 void 768 rt_newaddrmsg(int cmd, struct ifaddr *ifa, int error, struct rtentry *rt) 769 { 770 struct rt_addrinfo info; 771 struct sockaddr *sa = NULL; 772 int pass; 773 struct mbuf *m = NULL; 774 struct ifnet *ifp = ifa->ifa_ifp; 775 776 if (route_cb.any_count == 0) 777 return; 778 for (pass = 1; pass < 3; pass++) { 779 memset(&info, 0, sizeof(info)); 780 if ((cmd == RTM_ADD && pass == 1) || 781 (cmd == RTM_DELETE && pass == 2)) { 782 struct ifa_msghdr ifam; 783 int ncmd = cmd == RTM_ADD ? RTM_NEWADDR : RTM_DELADDR; 784 785 ifaaddr = sa = ifa->ifa_addr; 786 ifpaddr = TAILQ_FIRST(&ifp->if_addrlist)->ifa_addr; 787 netmask = ifa->ifa_netmask; 788 brdaddr = ifa->ifa_dstaddr; 789 memset(&ifam, 0, sizeof(ifam)); 790 ifam.ifam_index = ifp->if_index; 791 ifam.ifam_metric = ifa->ifa_metric; 792 ifam.ifam_flags = ifa->ifa_flags; 793 m = rt_msg1(ncmd, &info, (caddr_t)&ifam, sizeof(ifam)); 794 if (m == NULL) 795 continue; 796 mtod(m, struct ifa_msghdr *)->ifam_addrs = 797 info.rti_addrs; 798 } 799 if ((cmd == RTM_ADD && pass == 2) || 800 (cmd == RTM_DELETE && pass == 1)) { 801 struct rt_msghdr rtm; 802 803 if (rt == 0) 804 continue; 805 netmask = rt_mask(rt); 806 dst = sa = rt_key(rt); 807 gate = rt->rt_gateway; 808 memset(&rtm, 0, sizeof(rtm)); 809 rtm.rtm_index = ifp->if_index; 810 rtm.rtm_flags |= rt->rt_flags; 811 rtm.rtm_errno = error; 812 m = rt_msg1(cmd, &info, (caddr_t)&rtm, sizeof(rtm)); 813 if (m == NULL) 814 continue; 815 mtod(m, struct rt_msghdr *)->rtm_addrs = info.rti_addrs; 816 } 817 route_proto.sp_protocol = sa ? sa->sa_family : 0; 818 raw_input(m, &route_proto, &route_src, &route_dst); 819 } 820 } 821 822 /* 823 * This is called to generate routing socket messages indicating 824 * network interface arrival and departure. 825 */ 826 void 827 rt_ifannouncemsg(struct ifnet *ifp, int what) 828 { 829 struct if_announcemsghdr ifan; 830 struct mbuf *m; 831 struct rt_addrinfo info; 832 833 if (route_cb.any_count == 0) 834 return; 835 memset(&info, 0, sizeof(info)); 836 memset(&ifan, 0, sizeof(ifan)); 837 ifan.ifan_index = ifp->if_index; 838 strlcpy(ifan.ifan_name, ifp->if_xname, sizeof(ifan.ifan_name)); 839 ifan.ifan_what = what; 840 m = rt_msg1(RTM_IFANNOUNCE, &info, (caddr_t)&ifan, sizeof(ifan)); 841 if (m == 0) 842 return; 843 route_proto.sp_protocol = 0; 844 raw_input(m, &route_proto, &route_src, &route_dst); 845 } 846 847 /* 848 * This is used in dumping the kernel table via sysctl(). 849 */ 850 static int 851 sysctl_dumpentry(struct radix_node *rn, void *v) 852 { 853 struct walkarg *w = v; 854 struct rtentry *rt = (struct rtentry *)rn; 855 int error = 0, size; 856 struct rt_addrinfo info; 857 858 if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg)) 859 return 0; 860 memset(&info, 0, sizeof(info)); 861 dst = rt_key(rt); 862 gate = rt->rt_gateway; 863 netmask = rt_mask(rt); 864 genmask = rt->rt_genmask; 865 if (rt->rt_ifp) { 866 ifpaddr = TAILQ_FIRST(&rt->rt_ifp->if_addrlist)->ifa_addr; 867 ifaaddr = rt->rt_ifa->ifa_addr; 868 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT) 869 brdaddr = rt->rt_ifa->ifa_dstaddr; 870 } 871 if ((error = rt_msg2(RTM_GET, &info, 0, w, &size))) 872 return (error); 873 if (w->w_where && w->w_tmem && w->w_needed <= 0) { 874 struct rt_msghdr *rtm = (struct rt_msghdr *)w->w_tmem; 875 876 rtm->rtm_flags = rt->rt_flags; 877 rtm->rtm_use = rt->rt_use; 878 rtm->rtm_rmx = rt->rt_rmx; 879 rtm->rtm_index = rt->rt_ifp->if_index; 880 rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0; 881 rtm->rtm_addrs = info.rti_addrs; 882 if ((error = copyout(rtm, w->w_where, size)) != 0) 883 w->w_where = NULL; 884 else 885 w->w_where += size; 886 } 887 return (error); 888 } 889 890 static int 891 sysctl_iflist(int af, struct walkarg *w, int type) 892 { 893 struct ifnet *ifp; 894 struct ifaddr *ifa; 895 struct rt_addrinfo info; 896 int len, error = 0; 897 898 memset(&info, 0, sizeof(info)); 899 TAILQ_FOREACH(ifp, &ifnet, if_list) { 900 if (w->w_arg && w->w_arg != ifp->if_index) 901 continue; 902 ifa = TAILQ_FIRST(&ifp->if_addrlist); 903 ifpaddr = ifa->ifa_addr; 904 switch (type) { 905 case NET_RT_IFLIST: 906 error = 907 rt_msg2(RTM_IFINFO, &info, (caddr_t)0, w, &len); 908 break; 909 #ifdef COMPAT_14 910 case NET_RT_OIFLIST: 911 error = 912 rt_msg2(RTM_OIFINFO, &info, (caddr_t)0, w, &len); 913 break; 914 #endif 915 default: 916 panic("sysctl_iflist(1)"); 917 } 918 if (error) 919 return (error); 920 ifpaddr = 0; 921 if (w->w_where && w->w_tmem && w->w_needed <= 0) { 922 switch (type) { 923 case NET_RT_IFLIST: { 924 struct if_msghdr *ifm; 925 926 ifm = (struct if_msghdr *)w->w_tmem; 927 ifm->ifm_index = ifp->if_index; 928 ifm->ifm_flags = ifp->if_flags; 929 ifm->ifm_data = ifp->if_data; 930 ifm->ifm_addrs = info.rti_addrs; 931 error = copyout(ifm, w->w_where, len); 932 if (error) 933 return (error); 934 w->w_where += len; 935 break; 936 } 937 938 #ifdef COMPAT_14 939 case NET_RT_OIFLIST: { 940 struct if_msghdr14 *ifm; 941 942 ifm = (struct if_msghdr14 *)w->w_tmem; 943 ifm->ifm_index = ifp->if_index; 944 ifm->ifm_flags = ifp->if_flags; 945 ifm->ifm_data.ifi_type = ifp->if_data.ifi_type; 946 ifm->ifm_data.ifi_addrlen = 947 ifp->if_data.ifi_addrlen; 948 ifm->ifm_data.ifi_hdrlen = 949 ifp->if_data.ifi_hdrlen; 950 ifm->ifm_data.ifi_mtu = ifp->if_data.ifi_mtu; 951 ifm->ifm_data.ifi_metric = 952 ifp->if_data.ifi_metric; 953 ifm->ifm_data.ifi_baudrate = 954 ifp->if_data.ifi_baudrate; 955 ifm->ifm_data.ifi_ipackets = 956 ifp->if_data.ifi_ipackets; 957 ifm->ifm_data.ifi_ierrors = 958 ifp->if_data.ifi_ierrors; 959 ifm->ifm_data.ifi_opackets = 960 ifp->if_data.ifi_opackets; 961 ifm->ifm_data.ifi_oerrors = 962 ifp->if_data.ifi_oerrors; 963 ifm->ifm_data.ifi_collisions = 964 ifp->if_data.ifi_collisions; 965 ifm->ifm_data.ifi_ibytes = 966 ifp->if_data.ifi_ibytes; 967 ifm->ifm_data.ifi_obytes = 968 ifp->if_data.ifi_obytes; 969 ifm->ifm_data.ifi_imcasts = 970 ifp->if_data.ifi_imcasts; 971 ifm->ifm_data.ifi_omcasts = 972 ifp->if_data.ifi_omcasts; 973 ifm->ifm_data.ifi_iqdrops = 974 ifp->if_data.ifi_iqdrops; 975 ifm->ifm_data.ifi_noproto = 976 ifp->if_data.ifi_noproto; 977 ifm->ifm_data.ifi_lastchange = 978 ifp->if_data.ifi_lastchange; 979 ifm->ifm_addrs = info.rti_addrs; 980 error = copyout(ifm, w->w_where, len); 981 if (error) 982 return (error); 983 w->w_where += len; 984 break; 985 } 986 #endif 987 default: 988 panic("sysctl_iflist(2)"); 989 } 990 } 991 while ((ifa = TAILQ_NEXT(ifa, ifa_list)) != NULL) { 992 if (af && af != ifa->ifa_addr->sa_family) 993 continue; 994 ifaaddr = ifa->ifa_addr; 995 netmask = ifa->ifa_netmask; 996 brdaddr = ifa->ifa_dstaddr; 997 if ((error = rt_msg2(RTM_NEWADDR, &info, 0, w, &len))) 998 return (error); 999 if (w->w_where && w->w_tmem && w->w_needed <= 0) { 1000 struct ifa_msghdr *ifam; 1001 1002 ifam = (struct ifa_msghdr *)w->w_tmem; 1003 ifam->ifam_index = ifa->ifa_ifp->if_index; 1004 ifam->ifam_flags = ifa->ifa_flags; 1005 ifam->ifam_metric = ifa->ifa_metric; 1006 ifam->ifam_addrs = info.rti_addrs; 1007 error = copyout(w->w_tmem, w->w_where, len); 1008 if (error) 1009 return (error); 1010 w->w_where += len; 1011 } 1012 } 1013 ifaaddr = netmask = brdaddr = 0; 1014 } 1015 return (0); 1016 } 1017 1018 static int 1019 sysctl_rtable(SYSCTLFN_ARGS) 1020 { 1021 void *where = oldp; 1022 size_t *given = oldlenp; 1023 const void *new = newp; 1024 struct radix_node_head *rnh; 1025 int i, s, error = EINVAL; 1026 u_char af; 1027 struct walkarg w; 1028 1029 if (namelen == 1 && name[0] == CTL_QUERY) 1030 return (sysctl_query(SYSCTLFN_CALL(rnode))); 1031 1032 if (new) 1033 return (EPERM); 1034 if (namelen != 3) 1035 return (EINVAL); 1036 af = name[0]; 1037 w.w_tmemneeded = 0; 1038 w.w_tmemsize = 0; 1039 w.w_tmem = NULL; 1040 again: 1041 /* we may return here if a later [re]alloc of the t_mem buffer fails */ 1042 if (w.w_tmemneeded) { 1043 w.w_tmem = (caddr_t) malloc(w.w_tmemneeded, M_RTABLE, M_WAITOK); 1044 w.w_tmemsize = w.w_tmemneeded; 1045 w.w_tmemneeded = 0; 1046 } 1047 w.w_op = name[1]; 1048 w.w_arg = name[2]; 1049 w.w_given = *given; 1050 w.w_needed = 0 - w.w_given; 1051 w.w_where = where; 1052 1053 s = splsoftnet(); 1054 switch (w.w_op) { 1055 1056 case NET_RT_DUMP: 1057 case NET_RT_FLAGS: 1058 for (i = 1; i <= AF_MAX; i++) 1059 if ((rnh = rt_tables[i]) && (af == 0 || af == i) && 1060 (error = (*rnh->rnh_walktree)(rnh, 1061 sysctl_dumpentry, &w))) 1062 break; 1063 break; 1064 1065 #ifdef COMPAT_14 1066 case NET_RT_OIFLIST: 1067 error = sysctl_iflist(af, &w, w.w_op); 1068 break; 1069 #endif 1070 1071 case NET_RT_IFLIST: 1072 error = sysctl_iflist(af, &w, w.w_op); 1073 } 1074 splx(s); 1075 1076 /* check to see if we couldn't allocate memory with NOWAIT */ 1077 if (error == ENOBUFS && w.w_tmem == 0 && w.w_tmemneeded) 1078 goto again; 1079 1080 if (w.w_tmem) 1081 free(w.w_tmem, M_RTABLE); 1082 w.w_needed += w.w_given; 1083 if (where) { 1084 *given = w.w_where - (caddr_t) where; 1085 if (*given < w.w_needed) 1086 return (ENOMEM); 1087 } else { 1088 *given = (11 * w.w_needed) / 10; 1089 } 1090 return (error); 1091 } 1092 1093 /* 1094 * Definitions of protocols supported in the ROUTE domain. 1095 */ 1096 1097 const struct protosw routesw[] = { 1098 { 1099 SOCK_RAW, &routedomain, 0, PR_ATOMIC|PR_ADDR, 1100 raw_input, route_output, raw_ctlinput, 0, 1101 route_usrreq, 1102 raw_init, 0, 0, 0, 1103 } }; 1104 1105 struct domain routedomain = { 1106 PF_ROUTE, "route", route_init, 0, 0, 1107 routesw, &routesw[sizeof(routesw)/sizeof(routesw[0])] 1108 }; 1109 1110 SYSCTL_SETUP(sysctl_net_route_setup, "sysctl net.route subtree setup") 1111 { 1112 sysctl_createv(clog, 0, NULL, NULL, 1113 CTLFLAG_PERMANENT, 1114 CTLTYPE_NODE, "net", NULL, 1115 NULL, 0, NULL, 0, 1116 CTL_NET, CTL_EOL); 1117 1118 sysctl_createv(clog, 0, NULL, NULL, 1119 CTLFLAG_PERMANENT, 1120 CTLTYPE_NODE, "route", 1121 SYSCTL_DESCR("PF_ROUTE information"), 1122 NULL, 0, NULL, 0, 1123 CTL_NET, PF_ROUTE, CTL_EOL); 1124 sysctl_createv(clog, 0, NULL, NULL, 1125 CTLFLAG_PERMANENT, 1126 CTLTYPE_NODE, "rtable", 1127 SYSCTL_DESCR("Routing table information"), 1128 sysctl_rtable, 0, NULL, 0, 1129 CTL_NET, PF_ROUTE, 0 /* any protocol */, CTL_EOL); 1130 } 1131