1 /* 2 * Copyright (c) 1988, 1991, 1993 3 * The Regents of the University of California. All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 3. All advertising materials mentioning features or use of this software 14 * must display the following acknowledgement: 15 * This product includes software developed by the University of 16 * California, Berkeley and its contributors. 17 * 4. Neither the name of the University nor the names of its contributors 18 * may be used to endorse or promote products derived from this software 19 * without specific prior written permission. 20 * 21 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 24 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 31 * SUCH DAMAGE. 32 * 33 * @(#)rtsock.c 8.5 (Berkeley) 11/2/94 34 * $FreeBSD$ 35 */ 36 37 38 #include <sys/param.h> 39 #include <sys/systm.h> 40 #include <sys/kernel.h> 41 #include <sys/sysctl.h> 42 #include <sys/proc.h> 43 #include <sys/malloc.h> 44 #include <sys/mbuf.h> 45 #include <sys/socket.h> 46 #include <sys/socketvar.h> 47 #include <sys/domain.h> 48 #include <sys/protosw.h> 49 #include <sys/jail.h> 50 51 #include <net/if.h> 52 #include <net/route.h> 53 #include <net/raw_cb.h> 54 55 MALLOC_DEFINE(M_RTABLE, "routetbl", "routing tables"); 56 57 static struct sockaddr route_dst = { 2, PF_ROUTE, }; 58 static struct sockaddr route_src = { 2, PF_ROUTE, }; 59 static struct sockaddr sa_zero = { sizeof(sa_zero), AF_INET, }; 60 static struct sockproto route_proto = { PF_ROUTE, }; 61 62 struct walkarg { 63 int w_tmemsize; 64 int w_op, w_arg; 65 caddr_t w_tmem; 66 struct sysctl_req *w_req; 67 }; 68 69 static struct mbuf * 70 rt_msg1 __P((int, struct rt_addrinfo *)); 71 static int rt_msg2 __P((int, 72 struct rt_addrinfo *, caddr_t, struct walkarg *)); 73 static int rt_xaddrs __P((caddr_t, caddr_t, struct rt_addrinfo *)); 74 static int sysctl_dumpentry __P((struct radix_node *rn, void *vw)); 75 static int sysctl_iflist __P((int af, struct walkarg *w)); 76 static int route_output __P((struct mbuf *, struct socket *)); 77 static void rt_setmetrics __P((u_long, struct rt_metrics *, struct rt_metrics *)); 78 79 /* Sleazy use of local variables throughout file, warning!!!! */ 80 #define dst info.rti_info[RTAX_DST] 81 #define gate info.rti_info[RTAX_GATEWAY] 82 #define netmask info.rti_info[RTAX_NETMASK] 83 #define genmask info.rti_info[RTAX_GENMASK] 84 #define ifpaddr info.rti_info[RTAX_IFP] 85 #define ifaaddr info.rti_info[RTAX_IFA] 86 #define brdaddr info.rti_info[RTAX_BRD] 87 88 /* 89 * It really doesn't make any sense at all for this code to share much 90 * with raw_usrreq.c, since its functionality is so restricted. XXX 91 */ 92 static int 93 rts_abort(struct socket *so) 94 { 95 int s, error; 96 s = splnet(); 97 error = raw_usrreqs.pru_abort(so); 98 splx(s); 99 return error; 100 } 101 102 /* pru_accept is EOPNOTSUPP */ 103 104 static int 105 rts_attach(struct socket *so, int proto, struct proc *p) 106 { 107 struct rawcb *rp; 108 int s, error; 109 110 if (sotorawcb(so) != 0) 111 return EISCONN; /* XXX panic? */ 112 /* XXX */ 113 MALLOC(rp, struct rawcb *, sizeof *rp, M_PCB, M_WAITOK | M_ZERO); 114 if (rp == 0) 115 return ENOBUFS; 116 117 /* 118 * The splnet() is necessary to block protocols from sending 119 * error notifications (like RTM_REDIRECT or RTM_LOSING) while 120 * this PCB is extant but incompletely initialized. 121 * Probably we should try to do more of this work beforehand and 122 * eliminate the spl. 123 */ 124 s = splnet(); 125 so->so_pcb = (caddr_t)rp; 126 error = raw_usrreqs.pru_attach(so, proto, p); 127 rp = sotorawcb(so); 128 if (error) { 129 splx(s); 130 so->so_pcb = NULL; 131 free(rp, M_PCB); 132 return error; 133 } 134 switch(rp->rcb_proto.sp_protocol) { 135 case AF_INET: 136 route_cb.ip_count++; 137 break; 138 case AF_INET6: 139 route_cb.ip6_count++; 140 break; 141 case AF_IPX: 142 route_cb.ipx_count++; 143 break; 144 case AF_NS: 145 route_cb.ns_count++; 146 break; 147 } 148 rp->rcb_faddr = &route_src; 149 route_cb.any_count++; 150 soisconnected(so); 151 so->so_options |= SO_USELOOPBACK; 152 splx(s); 153 return 0; 154 } 155 156 static int 157 rts_bind(struct socket *so, struct sockaddr *nam, struct proc *p) 158 { 159 int s, error; 160 s = splnet(); 161 error = raw_usrreqs.pru_bind(so, nam, p); /* xxx just EINVAL */ 162 splx(s); 163 return error; 164 } 165 166 static int 167 rts_connect(struct socket *so, struct sockaddr *nam, struct proc *p) 168 { 169 int s, error; 170 s = splnet(); 171 error = raw_usrreqs.pru_connect(so, nam, p); /* XXX just EINVAL */ 172 splx(s); 173 return error; 174 } 175 176 /* pru_connect2 is EOPNOTSUPP */ 177 /* pru_control is EOPNOTSUPP */ 178 179 static int 180 rts_detach(struct socket *so) 181 { 182 struct rawcb *rp = sotorawcb(so); 183 int s, error; 184 185 s = splnet(); 186 if (rp != 0) { 187 switch(rp->rcb_proto.sp_protocol) { 188 case AF_INET: 189 route_cb.ip_count--; 190 break; 191 case AF_INET6: 192 route_cb.ip6_count--; 193 break; 194 case AF_IPX: 195 route_cb.ipx_count--; 196 break; 197 case AF_NS: 198 route_cb.ns_count--; 199 break; 200 } 201 route_cb.any_count--; 202 } 203 error = raw_usrreqs.pru_detach(so); 204 splx(s); 205 return error; 206 } 207 208 static int 209 rts_disconnect(struct socket *so) 210 { 211 int s, error; 212 s = splnet(); 213 error = raw_usrreqs.pru_disconnect(so); 214 splx(s); 215 return error; 216 } 217 218 /* pru_listen is EOPNOTSUPP */ 219 220 static int 221 rts_peeraddr(struct socket *so, struct sockaddr **nam) 222 { 223 int s, error; 224 s = splnet(); 225 error = raw_usrreqs.pru_peeraddr(so, nam); 226 splx(s); 227 return error; 228 } 229 230 /* pru_rcvd is EOPNOTSUPP */ 231 /* pru_rcvoob is EOPNOTSUPP */ 232 233 static int 234 rts_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam, 235 struct mbuf *control, struct proc *p) 236 { 237 int s, error; 238 s = splnet(); 239 error = raw_usrreqs.pru_send(so, flags, m, nam, control, p); 240 splx(s); 241 return error; 242 } 243 244 /* pru_sense is null */ 245 246 static int 247 rts_shutdown(struct socket *so) 248 { 249 int s, error; 250 s = splnet(); 251 error = raw_usrreqs.pru_shutdown(so); 252 splx(s); 253 return error; 254 } 255 256 static int 257 rts_sockaddr(struct socket *so, struct sockaddr **nam) 258 { 259 int s, error; 260 s = splnet(); 261 error = raw_usrreqs.pru_sockaddr(so, nam); 262 splx(s); 263 return error; 264 } 265 266 static struct pr_usrreqs route_usrreqs = { 267 rts_abort, pru_accept_notsupp, rts_attach, rts_bind, rts_connect, 268 pru_connect2_notsupp, pru_control_notsupp, rts_detach, rts_disconnect, 269 pru_listen_notsupp, rts_peeraddr, pru_rcvd_notsupp, pru_rcvoob_notsupp, 270 rts_send, pru_sense_null, rts_shutdown, rts_sockaddr, 271 sosend, soreceive, sopoll 272 }; 273 274 /*ARGSUSED*/ 275 static int 276 route_output(m, so) 277 register struct mbuf *m; 278 struct socket *so; 279 { 280 register struct rt_msghdr *rtm = 0; 281 register struct rtentry *rt = 0; 282 struct rtentry *saved_nrt = 0; 283 struct radix_node_head *rnh; 284 struct rt_addrinfo info; 285 int len, error = 0; 286 struct ifnet *ifp = 0; 287 struct ifaddr *ifa = 0; 288 289 #define senderr(e) { error = e; goto flush;} 290 if (m == 0 || ((m->m_len < sizeof(long)) && 291 (m = m_pullup(m, sizeof(long))) == 0)) 292 return (ENOBUFS); 293 if ((m->m_flags & M_PKTHDR) == 0) 294 panic("route_output"); 295 len = m->m_pkthdr.len; 296 if (len < sizeof(*rtm) || 297 len != mtod(m, struct rt_msghdr *)->rtm_msglen) { 298 dst = 0; 299 senderr(EINVAL); 300 } 301 R_Malloc(rtm, struct rt_msghdr *, len); 302 if (rtm == 0) { 303 dst = 0; 304 senderr(ENOBUFS); 305 } 306 m_copydata(m, 0, len, (caddr_t)rtm); 307 if (rtm->rtm_version != RTM_VERSION) { 308 dst = 0; 309 senderr(EPROTONOSUPPORT); 310 } 311 rtm->rtm_pid = curproc->p_pid; 312 info.rti_addrs = rtm->rtm_addrs; 313 if (rt_xaddrs((caddr_t)(rtm + 1), len + (caddr_t)rtm, &info)) { 314 dst = 0; 315 senderr(EINVAL); 316 } 317 if (dst == 0 || (dst->sa_family >= AF_MAX) 318 || (gate != 0 && (gate->sa_family >= AF_MAX))) 319 senderr(EINVAL); 320 if (genmask) { 321 struct radix_node *t; 322 t = rn_addmask((caddr_t)genmask, 0, 1); 323 if (t && Bcmp((caddr_t *)genmask + 1, (caddr_t *)t->rn_key + 1, 324 *(u_char *)t->rn_key - 1) == 0) 325 genmask = (struct sockaddr *)(t->rn_key); 326 else 327 senderr(ENOBUFS); 328 } 329 switch (rtm->rtm_type) { 330 331 case RTM_ADD: 332 if (gate == 0) 333 senderr(EINVAL); 334 error = rtrequest(RTM_ADD, dst, gate, netmask, 335 rtm->rtm_flags, &saved_nrt); 336 if (error == 0 && saved_nrt) { 337 rt_setmetrics(rtm->rtm_inits, 338 &rtm->rtm_rmx, &saved_nrt->rt_rmx); 339 saved_nrt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits); 340 saved_nrt->rt_rmx.rmx_locks |= 341 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks); 342 saved_nrt->rt_refcnt--; 343 saved_nrt->rt_genmask = genmask; 344 } 345 break; 346 347 case RTM_DELETE: 348 error = rtrequest(RTM_DELETE, dst, gate, netmask, 349 rtm->rtm_flags, &saved_nrt); 350 if (error == 0) { 351 if ((rt = saved_nrt)) 352 rt->rt_refcnt++; 353 goto report; 354 } 355 break; 356 357 case RTM_GET: 358 case RTM_CHANGE: 359 case RTM_LOCK: 360 if ((rnh = rt_tables[dst->sa_family]) == 0) { 361 senderr(EAFNOSUPPORT); 362 } else if ((rt = (struct rtentry *) 363 rnh->rnh_lookup(dst, netmask, rnh)) != NULL) 364 rt->rt_refcnt++; 365 else 366 senderr(ESRCH); 367 switch(rtm->rtm_type) { 368 369 case RTM_GET: 370 report: 371 dst = rt_key(rt); 372 gate = rt->rt_gateway; 373 netmask = rt_mask(rt); 374 genmask = rt->rt_genmask; 375 if (rtm->rtm_addrs & (RTA_IFP | RTA_IFA)) { 376 ifp = rt->rt_ifp; 377 if (ifp) { 378 ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 379 ifaaddr = rt->rt_ifa->ifa_addr; 380 rtm->rtm_index = ifp->if_index; 381 } else { 382 ifpaddr = 0; 383 ifaaddr = 0; 384 } 385 } 386 len = rt_msg2(rtm->rtm_type, &info, (caddr_t)0, 387 (struct walkarg *)0); 388 if (len > rtm->rtm_msglen) { 389 struct rt_msghdr *new_rtm; 390 R_Malloc(new_rtm, struct rt_msghdr *, len); 391 if (new_rtm == 0) 392 senderr(ENOBUFS); 393 Bcopy(rtm, new_rtm, rtm->rtm_msglen); 394 Free(rtm); rtm = new_rtm; 395 } 396 (void)rt_msg2(rtm->rtm_type, &info, (caddr_t)rtm, 397 (struct walkarg *)0); 398 rtm->rtm_flags = rt->rt_flags; 399 rtm->rtm_rmx = rt->rt_rmx; 400 rtm->rtm_addrs = info.rti_addrs; 401 break; 402 403 case RTM_CHANGE: 404 if (gate && (error = rt_setgate(rt, rt_key(rt), gate))) 405 senderr(error); 406 407 /* 408 * If they tried to change things but didn't specify 409 * the required gateway, then just use the old one. 410 * This can happen if the user tries to change the 411 * flags on the default route without changing the 412 * default gateway. Changing flags still doesn't work. 413 */ 414 if ((rt->rt_flags & RTF_GATEWAY) && !gate) 415 gate = rt->rt_gateway; 416 417 /* new gateway could require new ifaddr, ifp; 418 flags may also be different; ifp may be specified 419 by ll sockaddr when protocol address is ambiguous */ 420 if (ifpaddr && (ifa = ifa_ifwithnet(ifpaddr)) && 421 (ifp = ifa->ifa_ifp) && (ifaaddr || gate)) 422 ifa = ifaof_ifpforaddr(ifaaddr ? ifaaddr : gate, 423 ifp); 424 else if ((ifaaddr && (ifa = ifa_ifwithaddr(ifaaddr))) || 425 (gate && (ifa = ifa_ifwithroute(rt->rt_flags, 426 rt_key(rt), gate)))) 427 ifp = ifa->ifa_ifp; 428 if (ifa) { 429 register struct ifaddr *oifa = rt->rt_ifa; 430 if (oifa != ifa) { 431 if (oifa && oifa->ifa_rtrequest) 432 oifa->ifa_rtrequest(RTM_DELETE, 433 rt, gate); 434 IFAFREE(rt->rt_ifa); 435 rt->rt_ifa = ifa; 436 ifa->ifa_refcnt++; 437 rt->rt_ifp = ifp; 438 } 439 } 440 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, 441 &rt->rt_rmx); 442 if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest) 443 rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt, gate); 444 if (genmask) 445 rt->rt_genmask = genmask; 446 /* 447 * Fall into 448 */ 449 case RTM_LOCK: 450 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits); 451 rt->rt_rmx.rmx_locks |= 452 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks); 453 break; 454 } 455 break; 456 457 default: 458 senderr(EOPNOTSUPP); 459 } 460 461 flush: 462 if (rtm) { 463 if (error) 464 rtm->rtm_errno = error; 465 else 466 rtm->rtm_flags |= RTF_DONE; 467 } 468 if (rt) 469 rtfree(rt); 470 { 471 register struct rawcb *rp = 0; 472 /* 473 * Check to see if we don't want our own messages. 474 */ 475 if ((so->so_options & SO_USELOOPBACK) == 0) { 476 if (route_cb.any_count <= 1) { 477 if (rtm) 478 Free(rtm); 479 m_freem(m); 480 return (error); 481 } 482 /* There is another listener, so construct message */ 483 rp = sotorawcb(so); 484 } 485 if (rtm) { 486 m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm); 487 if (m->m_pkthdr.len < rtm->rtm_msglen) { 488 m_freem(m); 489 m = NULL; 490 } else if (m->m_pkthdr.len > rtm->rtm_msglen) 491 m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len); 492 Free(rtm); 493 } 494 if (rp) 495 rp->rcb_proto.sp_family = 0; /* Avoid us */ 496 if (dst) 497 route_proto.sp_protocol = dst->sa_family; 498 if (m) 499 raw_input(m, &route_proto, &route_src, &route_dst); 500 if (rp) 501 rp->rcb_proto.sp_family = PF_ROUTE; 502 } 503 return (error); 504 } 505 506 static void 507 rt_setmetrics(which, in, out) 508 u_long which; 509 register struct rt_metrics *in, *out; 510 { 511 #define metric(f, e) if (which & (f)) out->e = in->e; 512 metric(RTV_RPIPE, rmx_recvpipe); 513 metric(RTV_SPIPE, rmx_sendpipe); 514 metric(RTV_SSTHRESH, rmx_ssthresh); 515 metric(RTV_RTT, rmx_rtt); 516 metric(RTV_RTTVAR, rmx_rttvar); 517 metric(RTV_HOPCOUNT, rmx_hopcount); 518 metric(RTV_MTU, rmx_mtu); 519 metric(RTV_EXPIRE, rmx_expire); 520 #undef metric 521 } 522 523 #define ROUNDUP(a) \ 524 ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long)) 525 #define ADVANCE(x, n) (x += ROUNDUP((n)->sa_len)) 526 527 528 /* 529 * Extract the addresses of the passed sockaddrs. 530 * Do a little sanity checking so as to avoid bad memory references. 531 * This data is derived straight from userland. 532 */ 533 static int 534 rt_xaddrs(cp, cplim, rtinfo) 535 register caddr_t cp, cplim; 536 register struct rt_addrinfo *rtinfo; 537 { 538 register struct sockaddr *sa; 539 register int i; 540 541 bzero(rtinfo->rti_info, sizeof(rtinfo->rti_info)); 542 for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) { 543 if ((rtinfo->rti_addrs & (1 << i)) == 0) 544 continue; 545 sa = (struct sockaddr *)cp; 546 /* 547 * It won't fit. 548 */ 549 if ( (cp + sa->sa_len) > cplim ) { 550 return (EINVAL); 551 } 552 553 /* 554 * there are no more.. quit now 555 * If there are more bits, they are in error. 556 * I've seen this. route(1) can evidently generate these. 557 * This causes kernel to core dump. 558 * for compatibility, If we see this, point to a safe address. 559 */ 560 if (sa->sa_len == 0) { 561 rtinfo->rti_info[i] = &sa_zero; 562 return (0); /* should be EINVAL but for compat */ 563 } 564 565 /* accept it */ 566 rtinfo->rti_info[i] = sa; 567 ADVANCE(cp, sa); 568 } 569 return (0); 570 } 571 572 static struct mbuf * 573 rt_msg1(type, rtinfo) 574 int type; 575 register struct rt_addrinfo *rtinfo; 576 { 577 register struct rt_msghdr *rtm; 578 register struct mbuf *m; 579 register int i; 580 register struct sockaddr *sa; 581 int len, dlen; 582 583 switch (type) { 584 585 case RTM_DELADDR: 586 case RTM_NEWADDR: 587 len = sizeof(struct ifa_msghdr); 588 break; 589 590 case RTM_DELMADDR: 591 case RTM_NEWMADDR: 592 len = sizeof(struct ifma_msghdr); 593 break; 594 595 case RTM_IFINFO: 596 len = sizeof(struct if_msghdr); 597 break; 598 599 default: 600 len = sizeof(struct rt_msghdr); 601 } 602 if (len > MCLBYTES) 603 panic("rt_msg1"); 604 m = m_gethdr(M_DONTWAIT, MT_DATA); 605 if (m && len > MHLEN) { 606 MCLGET(m, M_DONTWAIT); 607 if ((m->m_flags & M_EXT) == 0) { 608 m_free(m); 609 m = NULL; 610 } 611 } 612 if (m == 0) 613 return (m); 614 m->m_pkthdr.len = m->m_len = len; 615 m->m_pkthdr.rcvif = 0; 616 rtm = mtod(m, struct rt_msghdr *); 617 bzero((caddr_t)rtm, len); 618 for (i = 0; i < RTAX_MAX; i++) { 619 if ((sa = rtinfo->rti_info[i]) == NULL) 620 continue; 621 rtinfo->rti_addrs |= (1 << i); 622 dlen = ROUNDUP(sa->sa_len); 623 m_copyback(m, len, dlen, (caddr_t)sa); 624 len += dlen; 625 } 626 if (m->m_pkthdr.len != len) { 627 m_freem(m); 628 return (NULL); 629 } 630 rtm->rtm_msglen = len; 631 rtm->rtm_version = RTM_VERSION; 632 rtm->rtm_type = type; 633 return (m); 634 } 635 636 static int 637 rt_msg2(type, rtinfo, cp, w) 638 int type; 639 register struct rt_addrinfo *rtinfo; 640 caddr_t cp; 641 struct walkarg *w; 642 { 643 register int i; 644 int len, dlen, second_time = 0; 645 caddr_t cp0; 646 647 rtinfo->rti_addrs = 0; 648 again: 649 switch (type) { 650 651 case RTM_DELADDR: 652 case RTM_NEWADDR: 653 len = sizeof(struct ifa_msghdr); 654 break; 655 656 case RTM_IFINFO: 657 len = sizeof(struct if_msghdr); 658 break; 659 660 default: 661 len = sizeof(struct rt_msghdr); 662 } 663 cp0 = cp; 664 if (cp0) 665 cp += len; 666 for (i = 0; i < RTAX_MAX; i++) { 667 register struct sockaddr *sa; 668 669 if ((sa = rtinfo->rti_info[i]) == 0) 670 continue; 671 rtinfo->rti_addrs |= (1 << i); 672 dlen = ROUNDUP(sa->sa_len); 673 if (cp) { 674 bcopy((caddr_t)sa, cp, (unsigned)dlen); 675 cp += dlen; 676 } 677 len += dlen; 678 } 679 if (cp == 0 && w != NULL && !second_time) { 680 register struct walkarg *rw = w; 681 682 if (rw->w_req) { 683 if (rw->w_tmemsize < len) { 684 if (rw->w_tmem) 685 free(rw->w_tmem, M_RTABLE); 686 rw->w_tmem = (caddr_t) 687 malloc(len, M_RTABLE, M_NOWAIT); 688 if (rw->w_tmem) 689 rw->w_tmemsize = len; 690 } 691 if (rw->w_tmem) { 692 cp = rw->w_tmem; 693 second_time = 1; 694 goto again; 695 } 696 } 697 } 698 if (cp) { 699 register struct rt_msghdr *rtm = (struct rt_msghdr *)cp0; 700 701 rtm->rtm_version = RTM_VERSION; 702 rtm->rtm_type = type; 703 rtm->rtm_msglen = len; 704 } 705 return (len); 706 } 707 708 /* 709 * This routine is called to generate a message from the routing 710 * socket indicating that a redirect has occured, a routing lookup 711 * has failed, or that a protocol has detected timeouts to a particular 712 * destination. 713 */ 714 void 715 rt_missmsg(type, rtinfo, flags, error) 716 int type, flags, error; 717 register struct rt_addrinfo *rtinfo; 718 { 719 register struct rt_msghdr *rtm; 720 register struct mbuf *m; 721 struct sockaddr *sa = rtinfo->rti_info[RTAX_DST]; 722 723 if (route_cb.any_count == 0) 724 return; 725 m = rt_msg1(type, rtinfo); 726 if (m == 0) 727 return; 728 rtm = mtod(m, struct rt_msghdr *); 729 rtm->rtm_flags = RTF_DONE | flags; 730 rtm->rtm_errno = error; 731 rtm->rtm_addrs = rtinfo->rti_addrs; 732 route_proto.sp_protocol = sa ? sa->sa_family : 0; 733 raw_input(m, &route_proto, &route_src, &route_dst); 734 } 735 736 /* 737 * This routine is called to generate a message from the routing 738 * socket indicating that the status of a network interface has changed. 739 */ 740 void 741 rt_ifmsg(ifp) 742 register struct ifnet *ifp; 743 { 744 register struct if_msghdr *ifm; 745 struct mbuf *m; 746 struct rt_addrinfo info; 747 748 if (route_cb.any_count == 0) 749 return; 750 bzero((caddr_t)&info, sizeof(info)); 751 m = rt_msg1(RTM_IFINFO, &info); 752 if (m == 0) 753 return; 754 ifm = mtod(m, struct if_msghdr *); 755 ifm->ifm_index = ifp->if_index; 756 ifm->ifm_flags = (u_short)ifp->if_flags; 757 ifm->ifm_data = ifp->if_data; 758 ifm->ifm_addrs = 0; 759 route_proto.sp_protocol = 0; 760 raw_input(m, &route_proto, &route_src, &route_dst); 761 } 762 763 /* 764 * This is called to generate messages from the routing socket 765 * indicating a network interface has had addresses associated with it. 766 * if we ever reverse the logic and replace messages TO the routing 767 * socket indicate a request to configure interfaces, then it will 768 * be unnecessary as the routing socket will automatically generate 769 * copies of it. 770 */ 771 void 772 rt_newaddrmsg(cmd, ifa, error, rt) 773 int cmd, error; 774 register struct ifaddr *ifa; 775 register struct rtentry *rt; 776 { 777 struct rt_addrinfo info; 778 struct sockaddr *sa = 0; 779 int pass; 780 struct mbuf *m = 0; 781 struct ifnet *ifp = ifa->ifa_ifp; 782 783 if (route_cb.any_count == 0) 784 return; 785 for (pass = 1; pass < 3; pass++) { 786 bzero((caddr_t)&info, sizeof(info)); 787 if ((cmd == RTM_ADD && pass == 1) || 788 (cmd == RTM_DELETE && pass == 2)) { 789 register struct ifa_msghdr *ifam; 790 int ncmd = cmd == RTM_ADD ? RTM_NEWADDR : RTM_DELADDR; 791 792 ifaaddr = sa = ifa->ifa_addr; 793 ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 794 netmask = ifa->ifa_netmask; 795 brdaddr = ifa->ifa_dstaddr; 796 if ((m = rt_msg1(ncmd, &info)) == NULL) 797 continue; 798 ifam = mtod(m, struct ifa_msghdr *); 799 ifam->ifam_index = ifp->if_index; 800 ifam->ifam_metric = ifa->ifa_metric; 801 ifam->ifam_flags = ifa->ifa_flags; 802 ifam->ifam_addrs = info.rti_addrs; 803 } 804 if ((cmd == RTM_ADD && pass == 2) || 805 (cmd == RTM_DELETE && pass == 1)) { 806 register struct rt_msghdr *rtm; 807 808 if (rt == 0) 809 continue; 810 netmask = rt_mask(rt); 811 dst = sa = rt_key(rt); 812 gate = rt->rt_gateway; 813 if ((m = rt_msg1(cmd, &info)) == NULL) 814 continue; 815 rtm = mtod(m, struct rt_msghdr *); 816 rtm->rtm_index = ifp->if_index; 817 rtm->rtm_flags |= rt->rt_flags; 818 rtm->rtm_errno = error; 819 rtm->rtm_addrs = info.rti_addrs; 820 } 821 route_proto.sp_protocol = sa ? sa->sa_family : 0; 822 raw_input(m, &route_proto, &route_src, &route_dst); 823 } 824 } 825 826 /* 827 * This is the analogue to the rt_newaddrmsg which performs the same 828 * function but for multicast group memberhips. This is easier since 829 * there is no route state to worry about. 830 */ 831 void 832 rt_newmaddrmsg(cmd, ifma) 833 int cmd; 834 struct ifmultiaddr *ifma; 835 { 836 struct rt_addrinfo info; 837 struct mbuf *m = 0; 838 struct ifnet *ifp = ifma->ifma_ifp; 839 struct ifma_msghdr *ifmam; 840 841 if (route_cb.any_count == 0) 842 return; 843 844 bzero((caddr_t)&info, sizeof(info)); 845 ifaaddr = ifma->ifma_addr; 846 if (ifp && TAILQ_FIRST(&ifp->if_addrhead)) 847 ifpaddr = TAILQ_FIRST(&ifp->if_addrhead)->ifa_addr; 848 else 849 ifpaddr = NULL; 850 /* 851 * If a link-layer address is present, present it as a ``gateway'' 852 * (similarly to how ARP entries, e.g., are presented). 853 */ 854 gate = ifma->ifma_lladdr; 855 if ((m = rt_msg1(cmd, &info)) == NULL) 856 return; 857 ifmam = mtod(m, struct ifma_msghdr *); 858 ifmam->ifmam_index = ifp->if_index; 859 ifmam->ifmam_addrs = info.rti_addrs; 860 route_proto.sp_protocol = ifma->ifma_addr->sa_family; 861 raw_input(m, &route_proto, &route_src, &route_dst); 862 } 863 864 /* 865 * This is used in dumping the kernel table via sysctl(). 866 */ 867 int 868 sysctl_dumpentry(rn, vw) 869 struct radix_node *rn; 870 void *vw; 871 { 872 register struct walkarg *w = vw; 873 register struct rtentry *rt = (struct rtentry *)rn; 874 int error = 0, size; 875 struct rt_addrinfo info; 876 877 if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg)) 878 return 0; 879 bzero((caddr_t)&info, sizeof(info)); 880 dst = rt_key(rt); 881 gate = rt->rt_gateway; 882 netmask = rt_mask(rt); 883 genmask = rt->rt_genmask; 884 size = rt_msg2(RTM_GET, &info, 0, w); 885 if (w->w_req && w->w_tmem) { 886 register struct rt_msghdr *rtm = (struct rt_msghdr *)w->w_tmem; 887 888 rtm->rtm_flags = rt->rt_flags; 889 rtm->rtm_use = rt->rt_use; 890 rtm->rtm_rmx = rt->rt_rmx; 891 rtm->rtm_index = rt->rt_ifp->if_index; 892 rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0; 893 rtm->rtm_addrs = info.rti_addrs; 894 error = SYSCTL_OUT(w->w_req, (caddr_t)rtm, size); 895 return (error); 896 } 897 return (error); 898 } 899 900 int 901 sysctl_iflist(af, w) 902 int af; 903 register struct walkarg *w; 904 { 905 register struct ifnet *ifp; 906 register struct ifaddr *ifa; 907 struct rt_addrinfo info; 908 int len, error = 0; 909 910 bzero((caddr_t)&info, sizeof(info)); 911 TAILQ_FOREACH(ifp, &ifnet, if_link) { 912 if (w->w_arg && w->w_arg != ifp->if_index) 913 continue; 914 ifa = TAILQ_FIRST(&ifp->if_addrhead); 915 ifpaddr = ifa->ifa_addr; 916 len = rt_msg2(RTM_IFINFO, &info, (caddr_t)0, w); 917 ifpaddr = 0; 918 if (w->w_req && w->w_tmem) { 919 register struct if_msghdr *ifm; 920 921 ifm = (struct if_msghdr *)w->w_tmem; 922 ifm->ifm_index = ifp->if_index; 923 ifm->ifm_flags = (u_short)ifp->if_flags; 924 ifm->ifm_data = ifp->if_data; 925 ifm->ifm_addrs = info.rti_addrs; 926 error = SYSCTL_OUT(w->w_req,(caddr_t)ifm, len); 927 if (error) 928 return (error); 929 } 930 while ((ifa = TAILQ_NEXT(ifa, ifa_link)) != 0) { 931 if (af && af != ifa->ifa_addr->sa_family) 932 continue; 933 if (jailed(curproc->p_ucred) && 934 prison_if(curproc->p_ucred, ifa->ifa_addr)) 935 continue; 936 ifaaddr = ifa->ifa_addr; 937 netmask = ifa->ifa_netmask; 938 brdaddr = ifa->ifa_dstaddr; 939 len = rt_msg2(RTM_NEWADDR, &info, 0, w); 940 if (w->w_req && w->w_tmem) { 941 register struct ifa_msghdr *ifam; 942 943 ifam = (struct ifa_msghdr *)w->w_tmem; 944 ifam->ifam_index = ifa->ifa_ifp->if_index; 945 ifam->ifam_flags = ifa->ifa_flags; 946 ifam->ifam_metric = ifa->ifa_metric; 947 ifam->ifam_addrs = info.rti_addrs; 948 error = SYSCTL_OUT(w->w_req, w->w_tmem, len); 949 if (error) 950 return (error); 951 } 952 } 953 ifaaddr = netmask = brdaddr = 0; 954 } 955 return (0); 956 } 957 958 static int 959 sysctl_rtsock(SYSCTL_HANDLER_ARGS) 960 { 961 int *name = (int *)arg1; 962 u_int namelen = arg2; 963 register struct radix_node_head *rnh; 964 int i, s, error = EINVAL; 965 u_char af; 966 struct walkarg w; 967 968 name ++; 969 namelen--; 970 if (req->newptr) 971 return (EPERM); 972 if (namelen != 3) 973 return (EINVAL); 974 af = name[0]; 975 Bzero(&w, sizeof(w)); 976 w.w_op = name[1]; 977 w.w_arg = name[2]; 978 w.w_req = req; 979 980 s = splnet(); 981 switch (w.w_op) { 982 983 case NET_RT_DUMP: 984 case NET_RT_FLAGS: 985 for (i = 1; i <= AF_MAX; i++) 986 if ((rnh = rt_tables[i]) && (af == 0 || af == i) && 987 (error = rnh->rnh_walktree(rnh, 988 sysctl_dumpentry, &w))) 989 break; 990 break; 991 992 case NET_RT_IFLIST: 993 error = sysctl_iflist(af, &w); 994 } 995 splx(s); 996 if (w.w_tmem) 997 free(w.w_tmem, M_RTABLE); 998 return (error); 999 } 1000 1001 SYSCTL_NODE(_net, PF_ROUTE, routetable, CTLFLAG_RD, sysctl_rtsock, ""); 1002 1003 /* 1004 * Definitions of protocols supported in the ROUTE domain. 1005 */ 1006 1007 extern struct domain routedomain; /* or at least forward */ 1008 1009 static struct protosw routesw[] = { 1010 { SOCK_RAW, &routedomain, 0, PR_ATOMIC|PR_ADDR, 1011 0, route_output, raw_ctlinput, 0, 1012 0, 1013 raw_init, 0, 0, 0, 1014 &route_usrreqs 1015 } 1016 }; 1017 1018 static struct domain routedomain = 1019 { PF_ROUTE, "route", 0, 0, 0, 1020 routesw, &routesw[sizeof(routesw)/sizeof(routesw[0])] }; 1021 1022 DOMAIN_SET(route); 1023