ip_output.c revision 1.233.2.3
1 /* $NetBSD: ip_output.c,v 1.233.2.3 2015/09/22 12:06:11 skrll 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) 1998 The NetBSD Foundation, Inc. 34 * All rights reserved. 35 * 36 * This code is derived from software contributed to The NetBSD Foundation 37 * by Public Access Networks Corporation ("Panix"). It was developed under 38 * contract to Panix by Eric Haszlakiewicz and Thor Lancelot Simon. 39 * 40 * Redistribution and use in source and binary forms, with or without 41 * modification, are permitted provided that the following conditions 42 * are met: 43 * 1. Redistributions of source code must retain the above copyright 44 * notice, this list of conditions and the following disclaimer. 45 * 2. Redistributions in binary form must reproduce the above copyright 46 * notice, this list of conditions and the following disclaimer in the 47 * documentation and/or other materials provided with the distribution. 48 * 49 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 50 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 51 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 52 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 53 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 54 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 55 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 56 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 57 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 58 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 59 * POSSIBILITY OF SUCH DAMAGE. 60 */ 61 62 /* 63 * Copyright (c) 1982, 1986, 1988, 1990, 1993 64 * The Regents of the University of California. All rights reserved. 65 * 66 * Redistribution and use in source and binary forms, with or without 67 * modification, are permitted provided that the following conditions 68 * are met: 69 * 1. Redistributions of source code must retain the above copyright 70 * notice, this list of conditions and the following disclaimer. 71 * 2. Redistributions in binary form must reproduce the above copyright 72 * notice, this list of conditions and the following disclaimer in the 73 * documentation and/or other materials provided with the distribution. 74 * 3. Neither the name of the University nor the names of its contributors 75 * may be used to endorse or promote products derived from this software 76 * without specific prior written permission. 77 * 78 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 79 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 80 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 81 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 82 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 83 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 84 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 85 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 86 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 87 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 88 * SUCH DAMAGE. 89 * 90 * @(#)ip_output.c 8.3 (Berkeley) 1/21/94 91 */ 92 93 #include <sys/cdefs.h> 94 __KERNEL_RCSID(0, "$NetBSD: ip_output.c,v 1.233.2.3 2015/09/22 12:06:11 skrll Exp $"); 95 96 #ifdef _KERNEL_OPT 97 #include "opt_inet.h" 98 #include "opt_ipsec.h" 99 #include "opt_mrouting.h" 100 #include "opt_net_mpsafe.h" 101 #include "opt_mpls.h" 102 #endif 103 104 #include <sys/param.h> 105 #include <sys/kmem.h> 106 #include <sys/mbuf.h> 107 #include <sys/protosw.h> 108 #include <sys/socket.h> 109 #include <sys/socketvar.h> 110 #include <sys/kauth.h> 111 #ifdef IPSEC 112 #include <sys/domain.h> 113 #endif 114 #include <sys/systm.h> 115 116 #include <net/if.h> 117 #include <net/if_types.h> 118 #include <net/route.h> 119 #include <net/pfil.h> 120 121 #include <netinet/in.h> 122 #include <netinet/in_systm.h> 123 #include <netinet/ip.h> 124 #include <netinet/in_pcb.h> 125 #include <netinet/in_var.h> 126 #include <netinet/ip_var.h> 127 #include <netinet/ip_private.h> 128 #include <netinet/in_offload.h> 129 #include <netinet/portalgo.h> 130 #include <netinet/udp.h> 131 132 #ifdef INET6 133 #include <netinet6/ip6_var.h> 134 #endif 135 136 #ifdef MROUTING 137 #include <netinet/ip_mroute.h> 138 #endif 139 140 #ifdef IPSEC 141 #include <netipsec/ipsec.h> 142 #include <netipsec/key.h> 143 #endif 144 145 #ifdef MPLS 146 #include <netmpls/mpls.h> 147 #include <netmpls/mpls_var.h> 148 #endif 149 150 static int ip_pcbopts(struct inpcb *, const struct sockopt *); 151 static struct mbuf *ip_insertoptions(struct mbuf *, struct mbuf *, int *); 152 static struct ifnet *ip_multicast_if(struct in_addr *, int *); 153 static void ip_mloopback(struct ifnet *, struct mbuf *, 154 const struct sockaddr_in *); 155 156 extern pfil_head_t *inet_pfil_hook; /* XXX */ 157 158 int ip_do_loopback_cksum = 0; 159 160 static bool 161 ip_hresolv_needed(const struct ifnet * const ifp) 162 { 163 switch (ifp->if_type) { 164 case IFT_ARCNET: 165 case IFT_ATM: 166 case IFT_ECONET: 167 case IFT_ETHER: 168 case IFT_FDDI: 169 case IFT_HIPPI: 170 case IFT_IEEE1394: 171 case IFT_ISO88025: 172 case IFT_SLIP: 173 return true; 174 default: 175 return false; 176 } 177 } 178 179 static int 180 klock_if_output(struct ifnet * const ifp, struct mbuf * const m, 181 const struct sockaddr * const dst, struct rtentry *rt) 182 { 183 int error; 184 185 #ifndef NET_MPSAFE 186 KERNEL_LOCK(1, NULL); 187 #endif 188 189 error = (*ifp->if_output)(ifp, m, dst, rt); 190 191 #ifndef NET_MPSAFE 192 KERNEL_UNLOCK_ONE(NULL); 193 #endif 194 195 return error; 196 } 197 198 /* 199 * Send an IP packet to a host. 200 * 201 * If necessary, resolve the arbitrary IP route, rt0, to an IP host route before 202 * calling ifp's output routine. 203 */ 204 int 205 ip_hresolv_output(struct ifnet * const ifp0, struct mbuf * const m, 206 const struct sockaddr * const dst, struct rtentry *rt00) 207 { 208 int error = 0; 209 struct ifnet *ifp = ifp0; 210 struct rtentry *rt, *rt0, *gwrt; 211 212 #define RTFREE_IF_NEEDED(_rt) \ 213 if ((_rt) != NULL && (_rt) != rt00) \ 214 rtfree((_rt)); 215 216 rt0 = rt00; 217 retry: 218 if (!ip_hresolv_needed(ifp)) { 219 rt = rt0; 220 goto out; 221 } 222 223 if (rt0 == NULL) { 224 rt = NULL; 225 goto out; 226 } 227 228 rt = rt0; 229 230 /* 231 * The following block is highly questionable. How did we get here 232 * with a !RTF_UP route? Does rtalloc1() always return an RTF_UP 233 * route? 234 */ 235 if ((rt->rt_flags & RTF_UP) == 0) { 236 rt = rtalloc1(dst, 1); 237 if (rt == NULL) { 238 error = EHOSTUNREACH; 239 goto bad; 240 } 241 rt0 = rt; 242 if (rt->rt_ifp != ifp) { 243 ifp = rt->rt_ifp; 244 goto retry; 245 } 246 } 247 248 if ((rt->rt_flags & RTF_GATEWAY) == 0) 249 goto out; 250 251 gwrt = rt_get_gwroute(rt); 252 RTFREE_IF_NEEDED(rt); 253 rt = gwrt; 254 if (rt == NULL || (rt->rt_flags & RTF_UP) == 0) { 255 if (rt != NULL) { 256 RTFREE_IF_NEEDED(rt); 257 rt = rt0; 258 } 259 if (rt == NULL) { 260 error = EHOSTUNREACH; 261 goto bad; 262 } 263 gwrt = rtalloc1(rt->rt_gateway, 1); 264 rt_set_gwroute(rt, gwrt); 265 RTFREE_IF_NEEDED(rt); 266 rt = gwrt; 267 if (rt == NULL) { 268 error = EHOSTUNREACH; 269 goto bad; 270 } 271 /* the "G" test below also prevents rt == rt0 */ 272 if ((rt->rt_flags & RTF_GATEWAY) != 0 || rt->rt_ifp != ifp) { 273 if (rt0->rt_gwroute != NULL) 274 rtfree(rt0->rt_gwroute); 275 rt0->rt_gwroute = NULL; 276 error = EHOSTUNREACH; 277 goto bad; 278 } 279 } 280 if ((rt->rt_flags & RTF_REJECT) != 0) { 281 if (rt->rt_rmx.rmx_expire == 0 || 282 time_uptime < rt->rt_rmx.rmx_expire) { 283 error = (rt == rt0) ? EHOSTDOWN : EHOSTUNREACH; 284 goto bad; 285 } 286 } 287 288 out: 289 #ifdef MPLS 290 if (rt0 != NULL && rt_gettag(rt0) != NULL && 291 rt_gettag(rt0)->sa_family == AF_MPLS && 292 (m->m_flags & (M_MCAST | M_BCAST)) == 0 && 293 ifp->if_type == IFT_ETHER) { 294 union mpls_shim msh; 295 msh.s_addr = MPLS_GETSADDR(rt0); 296 if (msh.shim.label != MPLS_LABEL_IMPLNULL) { 297 struct m_tag *mtag; 298 /* 299 * XXX tentative solution to tell ether_output 300 * it's MPLS. Need some more efficient solution. 301 */ 302 mtag = m_tag_get(PACKET_TAG_MPLS, 303 sizeof(int) /* dummy */, 304 M_NOWAIT); 305 if (mtag == NULL) { 306 error = ENOMEM; 307 goto bad; 308 } 309 m_tag_prepend(m, mtag); 310 } 311 } 312 #endif 313 314 error = klock_if_output(ifp, m, dst, rt); 315 goto exit; 316 317 bad: 318 if (m != NULL) 319 m_freem(m); 320 exit: 321 RTFREE_IF_NEEDED(rt); 322 323 return error; 324 325 #undef RTFREE_IF_NEEDED 326 } 327 328 /* 329 * IP output. The packet in mbuf chain m contains a skeletal IP 330 * header (with len, off, ttl, proto, tos, src, dst). 331 * The mbuf chain containing the packet will be freed. 332 * The mbuf opt, if present, will not be freed. 333 */ 334 int 335 ip_output(struct mbuf *m0, ...) 336 { 337 struct rtentry *rt; 338 struct ip *ip; 339 struct ifnet *ifp; 340 struct mbuf *m = m0; 341 int hlen = sizeof (struct ip); 342 int len, error = 0; 343 struct route iproute; 344 const struct sockaddr_in *dst; 345 struct in_ifaddr *ia; 346 int isbroadcast; 347 struct mbuf *opt; 348 struct route *ro; 349 int flags, sw_csum; 350 u_long mtu; 351 struct ip_moptions *imo; 352 struct socket *so; 353 va_list ap; 354 #ifdef IPSEC 355 struct secpolicy *sp = NULL; 356 #endif 357 bool natt_frag = false; 358 bool rtmtu_nolock; 359 union { 360 struct sockaddr dst; 361 struct sockaddr_in dst4; 362 } u; 363 struct sockaddr *rdst = &u.dst; /* real IP destination, as opposed 364 * to the nexthop 365 */ 366 367 len = 0; 368 va_start(ap, m0); 369 opt = va_arg(ap, struct mbuf *); 370 ro = va_arg(ap, struct route *); 371 flags = va_arg(ap, int); 372 imo = va_arg(ap, struct ip_moptions *); 373 so = va_arg(ap, struct socket *); 374 va_end(ap); 375 376 MCLAIM(m, &ip_tx_mowner); 377 378 KASSERT((m->m_flags & M_PKTHDR) != 0); 379 KASSERT((m->m_pkthdr.csum_flags & (M_CSUM_TCPv6|M_CSUM_UDPv6)) == 0); 380 KASSERT((m->m_pkthdr.csum_flags & (M_CSUM_TCPv4|M_CSUM_UDPv4)) != 381 (M_CSUM_TCPv4|M_CSUM_UDPv4)); 382 383 if (opt) { 384 m = ip_insertoptions(m, opt, &len); 385 if (len >= sizeof(struct ip)) 386 hlen = len; 387 } 388 ip = mtod(m, struct ip *); 389 390 /* 391 * Fill in IP header. 392 */ 393 if ((flags & (IP_FORWARDING|IP_RAWOUTPUT)) == 0) { 394 ip->ip_v = IPVERSION; 395 ip->ip_off = htons(0); 396 /* ip->ip_id filled in after we find out source ia */ 397 ip->ip_hl = hlen >> 2; 398 IP_STATINC(IP_STAT_LOCALOUT); 399 } else { 400 hlen = ip->ip_hl << 2; 401 } 402 403 /* 404 * Route packet. 405 */ 406 if (ro == NULL) { 407 memset(&iproute, 0, sizeof(iproute)); 408 ro = &iproute; 409 } 410 sockaddr_in_init(&u.dst4, &ip->ip_dst, 0); 411 dst = satocsin(rtcache_getdst(ro)); 412 413 /* 414 * If there is a cached route, check that it is to the same 415 * destination and is still up. If not, free it and try again. 416 * The address family should also be checked in case of sharing 417 * the cache with IPv6. 418 */ 419 if (dst && (dst->sin_family != AF_INET || 420 !in_hosteq(dst->sin_addr, ip->ip_dst))) 421 rtcache_free(ro); 422 423 if ((rt = rtcache_validate(ro)) == NULL && 424 (rt = rtcache_update(ro, 1)) == NULL) { 425 dst = &u.dst4; 426 error = rtcache_setdst(ro, &u.dst); 427 if (error != 0) 428 goto bad; 429 } 430 431 /* 432 * If routing to interface only, short circuit routing lookup. 433 */ 434 if (flags & IP_ROUTETOIF) { 435 if ((ia = ifatoia(ifa_ifwithladdr(sintocsa(dst)))) == NULL) { 436 IP_STATINC(IP_STAT_NOROUTE); 437 error = ENETUNREACH; 438 goto bad; 439 } 440 ifp = ia->ia_ifp; 441 mtu = ifp->if_mtu; 442 ip->ip_ttl = 1; 443 isbroadcast = in_broadcast(dst->sin_addr, ifp); 444 } else if ((IN_MULTICAST(ip->ip_dst.s_addr) || 445 ip->ip_dst.s_addr == INADDR_BROADCAST) && 446 imo != NULL && imo->imo_multicast_ifp != NULL) { 447 ifp = imo->imo_multicast_ifp; 448 mtu = ifp->if_mtu; 449 IFP_TO_IA(ifp, ia); 450 isbroadcast = 0; 451 } else { 452 if (rt == NULL) 453 rt = rtcache_init(ro); 454 if (rt == NULL) { 455 IP_STATINC(IP_STAT_NOROUTE); 456 error = EHOSTUNREACH; 457 goto bad; 458 } 459 ia = ifatoia(rt->rt_ifa); 460 ifp = rt->rt_ifp; 461 if ((mtu = rt->rt_rmx.rmx_mtu) == 0) 462 mtu = ifp->if_mtu; 463 rt->rt_use++; 464 if (rt->rt_flags & RTF_GATEWAY) 465 dst = satosin(rt->rt_gateway); 466 if (rt->rt_flags & RTF_HOST) 467 isbroadcast = rt->rt_flags & RTF_BROADCAST; 468 else 469 isbroadcast = in_broadcast(dst->sin_addr, ifp); 470 } 471 rtmtu_nolock = rt && (rt->rt_rmx.rmx_locks & RTV_MTU) == 0; 472 473 if (IN_MULTICAST(ip->ip_dst.s_addr) || 474 (ip->ip_dst.s_addr == INADDR_BROADCAST)) { 475 bool inmgroup; 476 477 m->m_flags |= (ip->ip_dst.s_addr == INADDR_BROADCAST) ? 478 M_BCAST : M_MCAST; 479 /* 480 * See if the caller provided any multicast options 481 */ 482 if (imo != NULL) 483 ip->ip_ttl = imo->imo_multicast_ttl; 484 else 485 ip->ip_ttl = IP_DEFAULT_MULTICAST_TTL; 486 487 /* 488 * if we don't know the outgoing ifp yet, we can't generate 489 * output 490 */ 491 if (!ifp) { 492 IP_STATINC(IP_STAT_NOROUTE); 493 error = ENETUNREACH; 494 goto bad; 495 } 496 497 /* 498 * If the packet is multicast or broadcast, confirm that 499 * the outgoing interface can transmit it. 500 */ 501 if (((m->m_flags & M_MCAST) && 502 (ifp->if_flags & IFF_MULTICAST) == 0) || 503 ((m->m_flags & M_BCAST) && 504 (ifp->if_flags & (IFF_BROADCAST|IFF_POINTOPOINT)) == 0)) { 505 IP_STATINC(IP_STAT_NOROUTE); 506 error = ENETUNREACH; 507 goto bad; 508 } 509 /* 510 * If source address not specified yet, use an address 511 * of outgoing interface. 512 */ 513 if (in_nullhost(ip->ip_src)) { 514 struct in_ifaddr *xia; 515 struct ifaddr *xifa; 516 517 IFP_TO_IA(ifp, xia); 518 if (!xia) { 519 error = EADDRNOTAVAIL; 520 goto bad; 521 } 522 xifa = &xia->ia_ifa; 523 if (xifa->ifa_getifa != NULL) { 524 xia = ifatoia((*xifa->ifa_getifa)(xifa, rdst)); 525 if (xia == NULL) { 526 error = EADDRNOTAVAIL; 527 goto bad; 528 } 529 } 530 ip->ip_src = xia->ia_addr.sin_addr; 531 } 532 533 inmgroup = in_multi_group(ip->ip_dst, ifp, flags); 534 if (inmgroup && (imo == NULL || imo->imo_multicast_loop)) { 535 /* 536 * If we belong to the destination multicast group 537 * on the outgoing interface, and the caller did not 538 * forbid loopback, loop back a copy. 539 */ 540 ip_mloopback(ifp, m, &u.dst4); 541 } 542 #ifdef MROUTING 543 else { 544 /* 545 * If we are acting as a multicast router, perform 546 * multicast forwarding as if the packet had just 547 * arrived on the interface to which we are about 548 * to send. The multicast forwarding function 549 * recursively calls this function, using the 550 * IP_FORWARDING flag to prevent infinite recursion. 551 * 552 * Multicasts that are looped back by ip_mloopback(), 553 * above, will be forwarded by the ip_input() routine, 554 * if necessary. 555 */ 556 extern struct socket *ip_mrouter; 557 558 if (ip_mrouter && (flags & IP_FORWARDING) == 0) { 559 if (ip_mforward(m, ifp) != 0) { 560 m_freem(m); 561 goto done; 562 } 563 } 564 } 565 #endif 566 /* 567 * Multicasts with a time-to-live of zero may be looped- 568 * back, above, but must not be transmitted on a network. 569 * Also, multicasts addressed to the loopback interface 570 * are not sent -- the above call to ip_mloopback() will 571 * loop back a copy if this host actually belongs to the 572 * destination group on the loopback interface. 573 */ 574 if (ip->ip_ttl == 0 || (ifp->if_flags & IFF_LOOPBACK) != 0) { 575 m_freem(m); 576 goto done; 577 } 578 goto sendit; 579 } 580 581 /* 582 * If source address not specified yet, use address 583 * of outgoing interface. 584 */ 585 if (in_nullhost(ip->ip_src)) { 586 struct ifaddr *xifa; 587 588 xifa = &ia->ia_ifa; 589 if (xifa->ifa_getifa != NULL) { 590 ia = ifatoia((*xifa->ifa_getifa)(xifa, rdst)); 591 if (ia == NULL) { 592 error = EADDRNOTAVAIL; 593 goto bad; 594 } 595 } 596 ip->ip_src = ia->ia_addr.sin_addr; 597 } 598 599 /* 600 * packets with Class-D address as source are not valid per 601 * RFC 1112 602 */ 603 if (IN_MULTICAST(ip->ip_src.s_addr)) { 604 IP_STATINC(IP_STAT_ODROPPED); 605 error = EADDRNOTAVAIL; 606 goto bad; 607 } 608 609 /* 610 * Look for broadcast address and and verify user is allowed to 611 * send such a packet. 612 */ 613 if (isbroadcast) { 614 if ((ifp->if_flags & IFF_BROADCAST) == 0) { 615 error = EADDRNOTAVAIL; 616 goto bad; 617 } 618 if ((flags & IP_ALLOWBROADCAST) == 0) { 619 error = EACCES; 620 goto bad; 621 } 622 /* don't allow broadcast messages to be fragmented */ 623 if (ntohs(ip->ip_len) > ifp->if_mtu) { 624 error = EMSGSIZE; 625 goto bad; 626 } 627 m->m_flags |= M_BCAST; 628 } else 629 m->m_flags &= ~M_BCAST; 630 631 sendit: 632 if ((flags & (IP_FORWARDING|IP_NOIPNEWID)) == 0) { 633 if (m->m_pkthdr.len < IP_MINFRAGSIZE) { 634 ip->ip_id = 0; 635 } else if ((m->m_pkthdr.csum_flags & M_CSUM_TSOv4) == 0) { 636 ip->ip_id = ip_newid(ia); 637 } else { 638 639 /* 640 * TSO capable interfaces (typically?) increment 641 * ip_id for each segment. 642 * "allocate" enough ids here to increase the chance 643 * for them to be unique. 644 * 645 * note that the following calculation is not 646 * needed to be precise. wasting some ip_id is fine. 647 */ 648 649 unsigned int segsz = m->m_pkthdr.segsz; 650 unsigned int datasz = ntohs(ip->ip_len) - hlen; 651 unsigned int num = howmany(datasz, segsz); 652 653 ip->ip_id = ip_newid_range(ia, num); 654 } 655 } 656 657 /* 658 * If we're doing Path MTU Discovery, we need to set DF unless 659 * the route's MTU is locked. 660 */ 661 if ((flags & IP_MTUDISC) != 0 && rtmtu_nolock) { 662 ip->ip_off |= htons(IP_DF); 663 } 664 665 #ifdef IPSEC 666 if (ipsec_used) { 667 bool ipsec_done = false; 668 669 /* Perform IPsec processing, if any. */ 670 error = ipsec4_output(m, so, flags, &sp, &mtu, &natt_frag, 671 &ipsec_done); 672 if (error || ipsec_done) 673 goto done; 674 } 675 #endif 676 677 /* 678 * Run through list of hooks for output packets. 679 */ 680 error = pfil_run_hooks(inet_pfil_hook, &m, ifp, PFIL_OUT); 681 if (error) 682 goto done; 683 if (m == NULL) 684 goto done; 685 686 ip = mtod(m, struct ip *); 687 hlen = ip->ip_hl << 2; 688 689 m->m_pkthdr.csum_data |= hlen << 16; 690 691 #if IFA_STATS 692 /* 693 * search for the source address structure to 694 * maintain output statistics. 695 */ 696 INADDR_TO_IA(ip->ip_src, ia); 697 #endif 698 699 /* Maybe skip checksums on loopback interfaces. */ 700 if (IN_NEED_CHECKSUM(ifp, M_CSUM_IPv4)) { 701 m->m_pkthdr.csum_flags |= M_CSUM_IPv4; 702 } 703 sw_csum = m->m_pkthdr.csum_flags & ~ifp->if_csum_flags_tx; 704 /* 705 * If small enough for mtu of path, or if using TCP segmentation 706 * offload, can just send directly. 707 */ 708 if (ntohs(ip->ip_len) <= mtu || 709 (m->m_pkthdr.csum_flags & M_CSUM_TSOv4) != 0) { 710 const struct sockaddr *sa; 711 712 #if IFA_STATS 713 if (ia) 714 ia->ia_ifa.ifa_data.ifad_outbytes += ntohs(ip->ip_len); 715 #endif 716 /* 717 * Always initialize the sum to 0! Some HW assisted 718 * checksumming requires this. 719 */ 720 ip->ip_sum = 0; 721 722 if ((m->m_pkthdr.csum_flags & M_CSUM_TSOv4) == 0) { 723 /* 724 * Perform any checksums that the hardware can't do 725 * for us. 726 * 727 * XXX Does any hardware require the {th,uh}_sum 728 * XXX fields to be 0? 729 */ 730 if (sw_csum & M_CSUM_IPv4) { 731 KASSERT(IN_NEED_CHECKSUM(ifp, M_CSUM_IPv4)); 732 ip->ip_sum = in_cksum(m, hlen); 733 m->m_pkthdr.csum_flags &= ~M_CSUM_IPv4; 734 } 735 if (sw_csum & (M_CSUM_TCPv4|M_CSUM_UDPv4)) { 736 if (IN_NEED_CHECKSUM(ifp, 737 sw_csum & (M_CSUM_TCPv4|M_CSUM_UDPv4))) { 738 in_delayed_cksum(m); 739 } 740 m->m_pkthdr.csum_flags &= 741 ~(M_CSUM_TCPv4|M_CSUM_UDPv4); 742 } 743 } 744 745 sa = (m->m_flags & M_MCAST) ? sintocsa(rdst) : sintocsa(dst); 746 if (__predict_true( 747 (m->m_pkthdr.csum_flags & M_CSUM_TSOv4) == 0 || 748 (ifp->if_capenable & IFCAP_TSOv4) != 0)) { 749 error = ip_hresolv_output(ifp, m, sa, rt); 750 } else { 751 error = ip_tso_output(ifp, m, sa, rt); 752 } 753 goto done; 754 } 755 756 /* 757 * We can't use HW checksumming if we're about to 758 * to fragment the packet. 759 * 760 * XXX Some hardware can do this. 761 */ 762 if (m->m_pkthdr.csum_flags & (M_CSUM_TCPv4|M_CSUM_UDPv4)) { 763 if (IN_NEED_CHECKSUM(ifp, 764 m->m_pkthdr.csum_flags & (M_CSUM_TCPv4|M_CSUM_UDPv4))) { 765 in_delayed_cksum(m); 766 } 767 m->m_pkthdr.csum_flags &= ~(M_CSUM_TCPv4|M_CSUM_UDPv4); 768 } 769 770 /* 771 * Too large for interface; fragment if possible. 772 * Must be able to put at least 8 bytes per fragment. 773 */ 774 if (ntohs(ip->ip_off) & IP_DF) { 775 if (flags & IP_RETURNMTU) { 776 struct inpcb *inp; 777 778 KASSERT(so && solocked(so)); 779 inp = sotoinpcb(so); 780 inp->inp_errormtu = mtu; 781 } 782 error = EMSGSIZE; 783 IP_STATINC(IP_STAT_CANTFRAG); 784 goto bad; 785 } 786 787 error = ip_fragment(m, ifp, mtu); 788 if (error) { 789 m = NULL; 790 goto bad; 791 } 792 793 for (; m; m = m0) { 794 m0 = m->m_nextpkt; 795 m->m_nextpkt = 0; 796 if (error) { 797 m_freem(m); 798 continue; 799 } 800 #if IFA_STATS 801 if (ia) 802 ia->ia_ifa.ifa_data.ifad_outbytes += ntohs(ip->ip_len); 803 #endif 804 /* 805 * If we get there, the packet has not been handled by 806 * IPsec whereas it should have. Now that it has been 807 * fragmented, re-inject it in ip_output so that IPsec 808 * processing can occur. 809 */ 810 if (natt_frag) { 811 error = ip_output(m, opt, ro, 812 flags | IP_RAWOUTPUT | IP_NOIPNEWID, 813 imo, so); 814 } else { 815 KASSERT((m->m_pkthdr.csum_flags & 816 (M_CSUM_UDPv4 | M_CSUM_TCPv4)) == 0); 817 error = ip_hresolv_output(ifp, m, 818 (m->m_flags & M_MCAST) ? 819 sintocsa(rdst) : sintocsa(dst), rt); 820 } 821 } 822 if (error == 0) { 823 IP_STATINC(IP_STAT_FRAGMENTED); 824 } 825 done: 826 if (ro == &iproute) { 827 rtcache_free(&iproute); 828 } 829 #ifdef IPSEC 830 if (sp) { 831 KEY_FREESP(&sp); 832 } 833 #endif 834 return error; 835 bad: 836 m_freem(m); 837 goto done; 838 } 839 840 int 841 ip_fragment(struct mbuf *m, struct ifnet *ifp, u_long mtu) 842 { 843 struct ip *ip, *mhip; 844 struct mbuf *m0; 845 int len, hlen, off; 846 int mhlen, firstlen; 847 struct mbuf **mnext; 848 int sw_csum = m->m_pkthdr.csum_flags; 849 int fragments = 0; 850 int s; 851 int error = 0; 852 853 ip = mtod(m, struct ip *); 854 hlen = ip->ip_hl << 2; 855 if (ifp != NULL) 856 sw_csum &= ~ifp->if_csum_flags_tx; 857 858 len = (mtu - hlen) &~ 7; 859 if (len < 8) { 860 m_freem(m); 861 return (EMSGSIZE); 862 } 863 864 firstlen = len; 865 mnext = &m->m_nextpkt; 866 867 /* 868 * Loop through length of segment after first fragment, 869 * make new header and copy data of each part and link onto chain. 870 */ 871 m0 = m; 872 mhlen = sizeof (struct ip); 873 for (off = hlen + len; off < ntohs(ip->ip_len); off += len) { 874 MGETHDR(m, M_DONTWAIT, MT_HEADER); 875 if (m == 0) { 876 error = ENOBUFS; 877 IP_STATINC(IP_STAT_ODROPPED); 878 goto sendorfree; 879 } 880 MCLAIM(m, m0->m_owner); 881 *mnext = m; 882 mnext = &m->m_nextpkt; 883 m->m_data += max_linkhdr; 884 mhip = mtod(m, struct ip *); 885 *mhip = *ip; 886 /* we must inherit MCAST and BCAST flags */ 887 m->m_flags |= m0->m_flags & (M_MCAST|M_BCAST); 888 if (hlen > sizeof (struct ip)) { 889 mhlen = ip_optcopy(ip, mhip) + sizeof (struct ip); 890 mhip->ip_hl = mhlen >> 2; 891 } 892 m->m_len = mhlen; 893 mhip->ip_off = ((off - hlen) >> 3) + 894 (ntohs(ip->ip_off) & ~IP_MF); 895 if (ip->ip_off & htons(IP_MF)) 896 mhip->ip_off |= IP_MF; 897 if (off + len >= ntohs(ip->ip_len)) 898 len = ntohs(ip->ip_len) - off; 899 else 900 mhip->ip_off |= IP_MF; 901 HTONS(mhip->ip_off); 902 mhip->ip_len = htons((u_int16_t)(len + mhlen)); 903 m->m_next = m_copym(m0, off, len, M_DONTWAIT); 904 if (m->m_next == 0) { 905 error = ENOBUFS; /* ??? */ 906 IP_STATINC(IP_STAT_ODROPPED); 907 goto sendorfree; 908 } 909 m->m_pkthdr.len = mhlen + len; 910 m->m_pkthdr.rcvif = NULL; 911 mhip->ip_sum = 0; 912 KASSERT((m->m_pkthdr.csum_flags & M_CSUM_IPv4) == 0); 913 if (sw_csum & M_CSUM_IPv4) { 914 mhip->ip_sum = in_cksum(m, mhlen); 915 } else { 916 /* 917 * checksum is hw-offloaded or not necessary. 918 */ 919 m->m_pkthdr.csum_flags |= 920 m0->m_pkthdr.csum_flags & M_CSUM_IPv4; 921 m->m_pkthdr.csum_data |= mhlen << 16; 922 KASSERT(!(ifp != NULL && 923 IN_NEED_CHECKSUM(ifp, M_CSUM_IPv4)) || 924 (m->m_pkthdr.csum_flags & M_CSUM_IPv4) != 0); 925 } 926 IP_STATINC(IP_STAT_OFRAGMENTS); 927 fragments++; 928 } 929 /* 930 * Update first fragment by trimming what's been copied out 931 * and updating header, then send each fragment (in order). 932 */ 933 m = m0; 934 m_adj(m, hlen + firstlen - ntohs(ip->ip_len)); 935 m->m_pkthdr.len = hlen + firstlen; 936 ip->ip_len = htons((u_int16_t)m->m_pkthdr.len); 937 ip->ip_off |= htons(IP_MF); 938 ip->ip_sum = 0; 939 if (sw_csum & M_CSUM_IPv4) { 940 ip->ip_sum = in_cksum(m, hlen); 941 m->m_pkthdr.csum_flags &= ~M_CSUM_IPv4; 942 } else { 943 /* 944 * checksum is hw-offloaded or not necessary. 945 */ 946 KASSERT(!(ifp != NULL && IN_NEED_CHECKSUM(ifp, M_CSUM_IPv4)) || 947 (m->m_pkthdr.csum_flags & M_CSUM_IPv4) != 0); 948 KASSERT(M_CSUM_DATA_IPv4_IPHL(m->m_pkthdr.csum_data) >= 949 sizeof(struct ip)); 950 } 951 sendorfree: 952 /* 953 * If there is no room for all the fragments, don't queue 954 * any of them. 955 */ 956 if (ifp != NULL) { 957 s = splnet(); 958 if (ifp->if_snd.ifq_maxlen - ifp->if_snd.ifq_len < fragments && 959 error == 0) { 960 error = ENOBUFS; 961 IP_STATINC(IP_STAT_ODROPPED); 962 IFQ_INC_DROPS(&ifp->if_snd); 963 } 964 splx(s); 965 } 966 if (error) { 967 for (m = m0; m; m = m0) { 968 m0 = m->m_nextpkt; 969 m->m_nextpkt = NULL; 970 m_freem(m); 971 } 972 } 973 return (error); 974 } 975 976 /* 977 * Process a delayed payload checksum calculation. 978 */ 979 void 980 in_delayed_cksum(struct mbuf *m) 981 { 982 struct ip *ip; 983 u_int16_t csum, offset; 984 985 ip = mtod(m, struct ip *); 986 offset = ip->ip_hl << 2; 987 csum = in4_cksum(m, 0, offset, ntohs(ip->ip_len) - offset); 988 if (csum == 0 && (m->m_pkthdr.csum_flags & M_CSUM_UDPv4) != 0) 989 csum = 0xffff; 990 991 offset += M_CSUM_DATA_IPv4_OFFSET(m->m_pkthdr.csum_data); 992 993 if ((offset + sizeof(u_int16_t)) > m->m_len) { 994 /* This happen when ip options were inserted 995 printf("in_delayed_cksum: pullup len %d off %d proto %d\n", 996 m->m_len, offset, ip->ip_p); 997 */ 998 m_copyback(m, offset, sizeof(csum), (void *) &csum); 999 } else 1000 *(u_int16_t *)(mtod(m, char *) + offset) = csum; 1001 } 1002 1003 /* 1004 * Determine the maximum length of the options to be inserted; 1005 * we would far rather allocate too much space rather than too little. 1006 */ 1007 1008 u_int 1009 ip_optlen(struct inpcb *inp) 1010 { 1011 struct mbuf *m = inp->inp_options; 1012 1013 if (m && m->m_len > offsetof(struct ipoption, ipopt_dst)) { 1014 return (m->m_len - offsetof(struct ipoption, ipopt_dst)); 1015 } 1016 return 0; 1017 } 1018 1019 /* 1020 * Insert IP options into preformed packet. 1021 * Adjust IP destination as required for IP source routing, 1022 * as indicated by a non-zero in_addr at the start of the options. 1023 */ 1024 static struct mbuf * 1025 ip_insertoptions(struct mbuf *m, struct mbuf *opt, int *phlen) 1026 { 1027 struct ipoption *p = mtod(opt, struct ipoption *); 1028 struct mbuf *n; 1029 struct ip *ip = mtod(m, struct ip *); 1030 unsigned optlen; 1031 1032 optlen = opt->m_len - sizeof(p->ipopt_dst); 1033 if (optlen + ntohs(ip->ip_len) > IP_MAXPACKET) 1034 return (m); /* XXX should fail */ 1035 if (!in_nullhost(p->ipopt_dst)) 1036 ip->ip_dst = p->ipopt_dst; 1037 if (M_READONLY(m) || M_LEADINGSPACE(m) < optlen) { 1038 MGETHDR(n, M_DONTWAIT, MT_HEADER); 1039 if (n == 0) 1040 return (m); 1041 MCLAIM(n, m->m_owner); 1042 M_MOVE_PKTHDR(n, m); 1043 m->m_len -= sizeof(struct ip); 1044 m->m_data += sizeof(struct ip); 1045 n->m_next = m; 1046 m = n; 1047 m->m_len = optlen + sizeof(struct ip); 1048 m->m_data += max_linkhdr; 1049 bcopy((void *)ip, mtod(m, void *), sizeof(struct ip)); 1050 } else { 1051 m->m_data -= optlen; 1052 m->m_len += optlen; 1053 memmove(mtod(m, void *), ip, sizeof(struct ip)); 1054 } 1055 m->m_pkthdr.len += optlen; 1056 ip = mtod(m, struct ip *); 1057 bcopy((void *)p->ipopt_list, (void *)(ip + 1), (unsigned)optlen); 1058 *phlen = sizeof(struct ip) + optlen; 1059 ip->ip_len = htons(ntohs(ip->ip_len) + optlen); 1060 return (m); 1061 } 1062 1063 /* 1064 * Copy options from ip to jp, 1065 * omitting those not copied during fragmentation. 1066 */ 1067 int 1068 ip_optcopy(struct ip *ip, struct ip *jp) 1069 { 1070 u_char *cp, *dp; 1071 int opt, optlen, cnt; 1072 1073 cp = (u_char *)(ip + 1); 1074 dp = (u_char *)(jp + 1); 1075 cnt = (ip->ip_hl << 2) - sizeof (struct ip); 1076 for (; cnt > 0; cnt -= optlen, cp += optlen) { 1077 opt = cp[0]; 1078 if (opt == IPOPT_EOL) 1079 break; 1080 if (opt == IPOPT_NOP) { 1081 /* Preserve for IP mcast tunnel's LSRR alignment. */ 1082 *dp++ = IPOPT_NOP; 1083 optlen = 1; 1084 continue; 1085 } 1086 1087 KASSERT(cnt >= IPOPT_OLEN + sizeof(*cp)); 1088 optlen = cp[IPOPT_OLEN]; 1089 KASSERT(optlen >= IPOPT_OLEN + sizeof(*cp) && optlen < cnt); 1090 1091 /* Invalid lengths should have been caught by ip_dooptions. */ 1092 if (optlen > cnt) 1093 optlen = cnt; 1094 if (IPOPT_COPIED(opt)) { 1095 bcopy((void *)cp, (void *)dp, (unsigned)optlen); 1096 dp += optlen; 1097 } 1098 } 1099 for (optlen = dp - (u_char *)(jp+1); optlen & 0x3; optlen++) 1100 *dp++ = IPOPT_EOL; 1101 return (optlen); 1102 } 1103 1104 /* 1105 * IP socket option processing. 1106 */ 1107 int 1108 ip_ctloutput(int op, struct socket *so, struct sockopt *sopt) 1109 { 1110 struct inpcb *inp = sotoinpcb(so); 1111 struct ip *ip = &inp->inp_ip; 1112 int inpflags = inp->inp_flags; 1113 int optval = 0, error = 0; 1114 1115 if (sopt->sopt_level != IPPROTO_IP) { 1116 if (sopt->sopt_level == SOL_SOCKET && sopt->sopt_name == SO_NOHEADER) 1117 return 0; 1118 return ENOPROTOOPT; 1119 } 1120 1121 switch (op) { 1122 case PRCO_SETOPT: 1123 switch (sopt->sopt_name) { 1124 case IP_OPTIONS: 1125 #ifdef notyet 1126 case IP_RETOPTS: 1127 #endif 1128 error = ip_pcbopts(inp, sopt); 1129 break; 1130 1131 case IP_TOS: 1132 case IP_TTL: 1133 case IP_MINTTL: 1134 case IP_PKTINFO: 1135 case IP_RECVOPTS: 1136 case IP_RECVRETOPTS: 1137 case IP_RECVDSTADDR: 1138 case IP_RECVIF: 1139 case IP_RECVPKTINFO: 1140 case IP_RECVTTL: 1141 error = sockopt_getint(sopt, &optval); 1142 if (error) 1143 break; 1144 1145 switch (sopt->sopt_name) { 1146 case IP_TOS: 1147 ip->ip_tos = optval; 1148 break; 1149 1150 case IP_TTL: 1151 ip->ip_ttl = optval; 1152 break; 1153 1154 case IP_MINTTL: 1155 if (optval > 0 && optval <= MAXTTL) 1156 inp->inp_ip_minttl = optval; 1157 else 1158 error = EINVAL; 1159 break; 1160 #define OPTSET(bit) \ 1161 if (optval) \ 1162 inpflags |= bit; \ 1163 else \ 1164 inpflags &= ~bit; 1165 1166 case IP_PKTINFO: 1167 OPTSET(INP_PKTINFO); 1168 break; 1169 1170 case IP_RECVOPTS: 1171 OPTSET(INP_RECVOPTS); 1172 break; 1173 1174 case IP_RECVPKTINFO: 1175 OPTSET(INP_RECVPKTINFO); 1176 break; 1177 1178 case IP_RECVRETOPTS: 1179 OPTSET(INP_RECVRETOPTS); 1180 break; 1181 1182 case IP_RECVDSTADDR: 1183 OPTSET(INP_RECVDSTADDR); 1184 break; 1185 1186 case IP_RECVIF: 1187 OPTSET(INP_RECVIF); 1188 break; 1189 1190 case IP_RECVTTL: 1191 OPTSET(INP_RECVTTL); 1192 break; 1193 } 1194 break; 1195 #undef OPTSET 1196 1197 case IP_MULTICAST_IF: 1198 case IP_MULTICAST_TTL: 1199 case IP_MULTICAST_LOOP: 1200 case IP_ADD_MEMBERSHIP: 1201 case IP_DROP_MEMBERSHIP: 1202 error = ip_setmoptions(&inp->inp_moptions, sopt); 1203 break; 1204 1205 case IP_PORTRANGE: 1206 error = sockopt_getint(sopt, &optval); 1207 if (error) 1208 break; 1209 1210 switch (optval) { 1211 case IP_PORTRANGE_DEFAULT: 1212 case IP_PORTRANGE_HIGH: 1213 inpflags &= ~(INP_LOWPORT); 1214 break; 1215 1216 case IP_PORTRANGE_LOW: 1217 inpflags |= INP_LOWPORT; 1218 break; 1219 1220 default: 1221 error = EINVAL; 1222 break; 1223 } 1224 break; 1225 1226 case IP_PORTALGO: 1227 error = sockopt_getint(sopt, &optval); 1228 if (error) 1229 break; 1230 1231 error = portalgo_algo_index_select( 1232 (struct inpcb_hdr *)inp, optval); 1233 break; 1234 1235 #if defined(IPSEC) 1236 case IP_IPSEC_POLICY: 1237 if (ipsec_enabled) { 1238 error = ipsec4_set_policy(inp, sopt->sopt_name, 1239 sopt->sopt_data, sopt->sopt_size, 1240 curlwp->l_cred); 1241 break; 1242 } 1243 /*FALLTHROUGH*/ 1244 #endif /* IPSEC */ 1245 1246 default: 1247 error = ENOPROTOOPT; 1248 break; 1249 } 1250 break; 1251 1252 case PRCO_GETOPT: 1253 switch (sopt->sopt_name) { 1254 case IP_OPTIONS: 1255 case IP_RETOPTS: { 1256 struct mbuf *mopts = inp->inp_options; 1257 1258 if (mopts) { 1259 struct mbuf *m; 1260 1261 m = m_copym(mopts, 0, M_COPYALL, M_DONTWAIT); 1262 if (m == NULL) { 1263 error = ENOBUFS; 1264 break; 1265 } 1266 error = sockopt_setmbuf(sopt, m); 1267 } 1268 break; 1269 } 1270 case IP_PKTINFO: 1271 case IP_TOS: 1272 case IP_TTL: 1273 case IP_MINTTL: 1274 case IP_RECVOPTS: 1275 case IP_RECVRETOPTS: 1276 case IP_RECVDSTADDR: 1277 case IP_RECVIF: 1278 case IP_RECVPKTINFO: 1279 case IP_RECVTTL: 1280 case IP_ERRORMTU: 1281 switch (sopt->sopt_name) { 1282 case IP_TOS: 1283 optval = ip->ip_tos; 1284 break; 1285 1286 case IP_TTL: 1287 optval = ip->ip_ttl; 1288 break; 1289 1290 case IP_MINTTL: 1291 optval = inp->inp_ip_minttl; 1292 break; 1293 1294 case IP_ERRORMTU: 1295 optval = inp->inp_errormtu; 1296 break; 1297 1298 #define OPTBIT(bit) (inpflags & bit ? 1 : 0) 1299 1300 case IP_PKTINFO: 1301 optval = OPTBIT(INP_PKTINFO); 1302 break; 1303 1304 case IP_RECVOPTS: 1305 optval = OPTBIT(INP_RECVOPTS); 1306 break; 1307 1308 case IP_RECVPKTINFO: 1309 optval = OPTBIT(INP_RECVPKTINFO); 1310 break; 1311 1312 case IP_RECVRETOPTS: 1313 optval = OPTBIT(INP_RECVRETOPTS); 1314 break; 1315 1316 case IP_RECVDSTADDR: 1317 optval = OPTBIT(INP_RECVDSTADDR); 1318 break; 1319 1320 case IP_RECVIF: 1321 optval = OPTBIT(INP_RECVIF); 1322 break; 1323 1324 case IP_RECVTTL: 1325 optval = OPTBIT(INP_RECVTTL); 1326 break; 1327 } 1328 error = sockopt_setint(sopt, optval); 1329 break; 1330 1331 #if 0 /* defined(IPSEC) */ 1332 case IP_IPSEC_POLICY: 1333 { 1334 struct mbuf *m = NULL; 1335 1336 /* XXX this will return EINVAL as sopt is empty */ 1337 error = ipsec4_get_policy(inp, sopt->sopt_data, 1338 sopt->sopt_size, &m); 1339 if (error == 0) 1340 error = sockopt_setmbuf(sopt, m); 1341 break; 1342 } 1343 #endif /*IPSEC*/ 1344 1345 case IP_MULTICAST_IF: 1346 case IP_MULTICAST_TTL: 1347 case IP_MULTICAST_LOOP: 1348 case IP_ADD_MEMBERSHIP: 1349 case IP_DROP_MEMBERSHIP: 1350 error = ip_getmoptions(inp->inp_moptions, sopt); 1351 break; 1352 1353 case IP_PORTRANGE: 1354 if (inpflags & INP_LOWPORT) 1355 optval = IP_PORTRANGE_LOW; 1356 else 1357 optval = IP_PORTRANGE_DEFAULT; 1358 error = sockopt_setint(sopt, optval); 1359 break; 1360 1361 case IP_PORTALGO: 1362 optval = inp->inp_portalgo; 1363 error = sockopt_setint(sopt, optval); 1364 break; 1365 1366 default: 1367 error = ENOPROTOOPT; 1368 break; 1369 } 1370 break; 1371 } 1372 1373 if (!error) { 1374 inp->inp_flags = inpflags; 1375 } 1376 return error; 1377 } 1378 1379 /* 1380 * Set up IP options in pcb for insertion in output packets. 1381 * Store in mbuf with pointer in pcbopt, adding pseudo-option 1382 * with destination address if source routed. 1383 */ 1384 static int 1385 ip_pcbopts(struct inpcb *inp, const struct sockopt *sopt) 1386 { 1387 struct mbuf *m; 1388 const u_char *cp; 1389 u_char *dp; 1390 int cnt; 1391 1392 /* Turn off any old options. */ 1393 if (inp->inp_options) { 1394 m_free(inp->inp_options); 1395 } 1396 inp->inp_options = NULL; 1397 if ((cnt = sopt->sopt_size) == 0) { 1398 /* Only turning off any previous options. */ 1399 return 0; 1400 } 1401 cp = sopt->sopt_data; 1402 1403 #ifndef __vax__ 1404 if (cnt % sizeof(int32_t)) 1405 return (EINVAL); 1406 #endif 1407 1408 m = m_get(M_DONTWAIT, MT_SOOPTS); 1409 if (m == NULL) 1410 return (ENOBUFS); 1411 1412 dp = mtod(m, u_char *); 1413 memset(dp, 0, sizeof(struct in_addr)); 1414 dp += sizeof(struct in_addr); 1415 m->m_len = sizeof(struct in_addr); 1416 1417 /* 1418 * IP option list according to RFC791. Each option is of the form 1419 * 1420 * [optval] [olen] [(olen - 2) data bytes] 1421 * 1422 * We validate the list and copy options to an mbuf for prepending 1423 * to data packets. The IP first-hop destination address will be 1424 * stored before actual options and is zero if unset. 1425 */ 1426 while (cnt > 0) { 1427 uint8_t optval, olen, offset; 1428 1429 optval = cp[IPOPT_OPTVAL]; 1430 1431 if (optval == IPOPT_EOL || optval == IPOPT_NOP) { 1432 olen = 1; 1433 } else { 1434 if (cnt < IPOPT_OLEN + 1) 1435 goto bad; 1436 1437 olen = cp[IPOPT_OLEN]; 1438 if (olen < IPOPT_OLEN + 1 || olen > cnt) 1439 goto bad; 1440 } 1441 1442 if (optval == IPOPT_LSRR || optval == IPOPT_SSRR) { 1443 /* 1444 * user process specifies route as: 1445 * ->A->B->C->D 1446 * D must be our final destination (but we can't 1447 * check that since we may not have connected yet). 1448 * A is first hop destination, which doesn't appear in 1449 * actual IP option, but is stored before the options. 1450 */ 1451 if (olen < IPOPT_OFFSET + 1 + sizeof(struct in_addr)) 1452 goto bad; 1453 1454 offset = cp[IPOPT_OFFSET]; 1455 memcpy(mtod(m, u_char *), cp + IPOPT_OFFSET + 1, 1456 sizeof(struct in_addr)); 1457 1458 cp += sizeof(struct in_addr); 1459 cnt -= sizeof(struct in_addr); 1460 olen -= sizeof(struct in_addr); 1461 1462 if (m->m_len + olen > MAX_IPOPTLEN + sizeof(struct in_addr)) 1463 goto bad; 1464 1465 memcpy(dp, cp, olen); 1466 dp[IPOPT_OPTVAL] = optval; 1467 dp[IPOPT_OLEN] = olen; 1468 dp[IPOPT_OFFSET] = offset; 1469 break; 1470 } else { 1471 if (m->m_len + olen > MAX_IPOPTLEN + sizeof(struct in_addr)) 1472 goto bad; 1473 1474 memcpy(dp, cp, olen); 1475 break; 1476 } 1477 1478 dp += olen; 1479 m->m_len += olen; 1480 1481 if (optval == IPOPT_EOL) 1482 break; 1483 1484 cp += olen; 1485 cnt -= olen; 1486 } 1487 1488 inp->inp_options = m; 1489 return 0; 1490 bad: 1491 (void)m_free(m); 1492 return EINVAL; 1493 } 1494 1495 /* 1496 * following RFC1724 section 3.3, 0.0.0.0/8 is interpreted as interface index. 1497 */ 1498 static struct ifnet * 1499 ip_multicast_if(struct in_addr *a, int *ifindexp) 1500 { 1501 int ifindex; 1502 struct ifnet *ifp = NULL; 1503 struct in_ifaddr *ia; 1504 1505 if (ifindexp) 1506 *ifindexp = 0; 1507 if (ntohl(a->s_addr) >> 24 == 0) { 1508 ifindex = ntohl(a->s_addr) & 0xffffff; 1509 ifp = if_byindex(ifindex); 1510 if (!ifp) 1511 return NULL; 1512 if (ifindexp) 1513 *ifindexp = ifindex; 1514 } else { 1515 LIST_FOREACH(ia, &IN_IFADDR_HASH(a->s_addr), ia_hash) { 1516 if (in_hosteq(ia->ia_addr.sin_addr, *a) && 1517 (ia->ia_ifp->if_flags & IFF_MULTICAST) != 0) { 1518 ifp = ia->ia_ifp; 1519 break; 1520 } 1521 } 1522 } 1523 return ifp; 1524 } 1525 1526 static int 1527 ip_getoptval(const struct sockopt *sopt, u_int8_t *val, u_int maxval) 1528 { 1529 u_int tval; 1530 u_char cval; 1531 int error; 1532 1533 if (sopt == NULL) 1534 return EINVAL; 1535 1536 switch (sopt->sopt_size) { 1537 case sizeof(u_char): 1538 error = sockopt_get(sopt, &cval, sizeof(u_char)); 1539 tval = cval; 1540 break; 1541 1542 case sizeof(u_int): 1543 error = sockopt_get(sopt, &tval, sizeof(u_int)); 1544 break; 1545 1546 default: 1547 error = EINVAL; 1548 } 1549 1550 if (error) 1551 return error; 1552 1553 if (tval > maxval) 1554 return EINVAL; 1555 1556 *val = tval; 1557 return 0; 1558 } 1559 1560 static int 1561 ip_get_membership(const struct sockopt *sopt, struct ifnet **ifp, 1562 struct in_addr *ia, bool add) 1563 { 1564 int error; 1565 struct ip_mreq mreq; 1566 1567 error = sockopt_get(sopt, &mreq, sizeof(mreq)); 1568 if (error) 1569 return error; 1570 1571 if (!IN_MULTICAST(mreq.imr_multiaddr.s_addr)) 1572 return EINVAL; 1573 1574 memcpy(ia, &mreq.imr_multiaddr, sizeof(*ia)); 1575 1576 if (in_nullhost(mreq.imr_interface)) { 1577 union { 1578 struct sockaddr dst; 1579 struct sockaddr_in dst4; 1580 } u; 1581 struct route ro; 1582 1583 if (!add) { 1584 *ifp = NULL; 1585 return 0; 1586 } 1587 /* 1588 * If no interface address was provided, use the interface of 1589 * the route to the given multicast address. 1590 */ 1591 struct rtentry *rt; 1592 memset(&ro, 0, sizeof(ro)); 1593 1594 sockaddr_in_init(&u.dst4, ia, 0); 1595 error = rtcache_setdst(&ro, &u.dst); 1596 if (error != 0) 1597 return error; 1598 *ifp = (rt = rtcache_init(&ro)) != NULL ? rt->rt_ifp : NULL; 1599 rtcache_free(&ro); 1600 } else { 1601 *ifp = ip_multicast_if(&mreq.imr_interface, NULL); 1602 if (!add && *ifp == NULL) 1603 return EADDRNOTAVAIL; 1604 } 1605 return 0; 1606 } 1607 1608 /* 1609 * Add a multicast group membership. 1610 * Group must be a valid IP multicast address. 1611 */ 1612 static int 1613 ip_add_membership(struct ip_moptions *imo, const struct sockopt *sopt) 1614 { 1615 struct ifnet *ifp; 1616 struct in_addr ia; 1617 int i, error; 1618 1619 if (sopt->sopt_size == sizeof(struct ip_mreq)) 1620 error = ip_get_membership(sopt, &ifp, &ia, true); 1621 else 1622 #ifdef INET6 1623 error = ip6_get_membership(sopt, &ifp, &ia, sizeof(ia)); 1624 #else 1625 return EINVAL; 1626 #endif 1627 1628 if (error) 1629 return error; 1630 1631 /* 1632 * See if we found an interface, and confirm that it 1633 * supports multicast. 1634 */ 1635 if (ifp == NULL || (ifp->if_flags & IFF_MULTICAST) == 0) 1636 return EADDRNOTAVAIL; 1637 1638 /* 1639 * See if the membership already exists or if all the 1640 * membership slots are full. 1641 */ 1642 for (i = 0; i < imo->imo_num_memberships; ++i) { 1643 if (imo->imo_membership[i]->inm_ifp == ifp && 1644 in_hosteq(imo->imo_membership[i]->inm_addr, ia)) 1645 break; 1646 } 1647 if (i < imo->imo_num_memberships) 1648 return EADDRINUSE; 1649 1650 if (i == IP_MAX_MEMBERSHIPS) 1651 return ETOOMANYREFS; 1652 1653 /* 1654 * Everything looks good; add a new record to the multicast 1655 * address list for the given interface. 1656 */ 1657 if ((imo->imo_membership[i] = in_addmulti(&ia, ifp)) == NULL) 1658 return ENOBUFS; 1659 1660 ++imo->imo_num_memberships; 1661 return 0; 1662 } 1663 1664 /* 1665 * Drop a multicast group membership. 1666 * Group must be a valid IP multicast address. 1667 */ 1668 static int 1669 ip_drop_membership(struct ip_moptions *imo, const struct sockopt *sopt) 1670 { 1671 struct in_addr ia; 1672 struct ifnet *ifp; 1673 int i, error; 1674 1675 if (sopt->sopt_size == sizeof(struct ip_mreq)) 1676 error = ip_get_membership(sopt, &ifp, &ia, false); 1677 else 1678 #ifdef INET6 1679 error = ip6_get_membership(sopt, &ifp, &ia, sizeof(ia)); 1680 #else 1681 return EINVAL; 1682 #endif 1683 1684 if (error) 1685 return error; 1686 1687 /* 1688 * Find the membership in the membership array. 1689 */ 1690 for (i = 0; i < imo->imo_num_memberships; ++i) { 1691 if ((ifp == NULL || 1692 imo->imo_membership[i]->inm_ifp == ifp) && 1693 in_hosteq(imo->imo_membership[i]->inm_addr, ia)) 1694 break; 1695 } 1696 if (i == imo->imo_num_memberships) 1697 return EADDRNOTAVAIL; 1698 1699 /* 1700 * Give up the multicast address record to which the 1701 * membership points. 1702 */ 1703 in_delmulti(imo->imo_membership[i]); 1704 1705 /* 1706 * Remove the gap in the membership array. 1707 */ 1708 for (++i; i < imo->imo_num_memberships; ++i) 1709 imo->imo_membership[i-1] = imo->imo_membership[i]; 1710 --imo->imo_num_memberships; 1711 return 0; 1712 } 1713 1714 /* 1715 * Set the IP multicast options in response to user setsockopt(). 1716 */ 1717 int 1718 ip_setmoptions(struct ip_moptions **pimo, const struct sockopt *sopt) 1719 { 1720 struct ip_moptions *imo = *pimo; 1721 struct in_addr addr; 1722 struct ifnet *ifp; 1723 int ifindex, error = 0; 1724 1725 if (!imo) { 1726 /* 1727 * No multicast option buffer attached to the pcb; 1728 * allocate one and initialize to default values. 1729 */ 1730 imo = kmem_intr_alloc(sizeof(*imo), KM_NOSLEEP); 1731 if (imo == NULL) 1732 return ENOBUFS; 1733 1734 imo->imo_multicast_ifp = NULL; 1735 imo->imo_multicast_addr.s_addr = INADDR_ANY; 1736 imo->imo_multicast_ttl = IP_DEFAULT_MULTICAST_TTL; 1737 imo->imo_multicast_loop = IP_DEFAULT_MULTICAST_LOOP; 1738 imo->imo_num_memberships = 0; 1739 *pimo = imo; 1740 } 1741 1742 switch (sopt->sopt_name) { 1743 case IP_MULTICAST_IF: 1744 /* 1745 * Select the interface for outgoing multicast packets. 1746 */ 1747 error = sockopt_get(sopt, &addr, sizeof(addr)); 1748 if (error) 1749 break; 1750 1751 /* 1752 * INADDR_ANY is used to remove a previous selection. 1753 * When no interface is selected, a default one is 1754 * chosen every time a multicast packet is sent. 1755 */ 1756 if (in_nullhost(addr)) { 1757 imo->imo_multicast_ifp = NULL; 1758 break; 1759 } 1760 /* 1761 * The selected interface is identified by its local 1762 * IP address. Find the interface and confirm that 1763 * it supports multicasting. 1764 */ 1765 ifp = ip_multicast_if(&addr, &ifindex); 1766 if (ifp == NULL || (ifp->if_flags & IFF_MULTICAST) == 0) { 1767 error = EADDRNOTAVAIL; 1768 break; 1769 } 1770 imo->imo_multicast_ifp = ifp; 1771 if (ifindex) 1772 imo->imo_multicast_addr = addr; 1773 else 1774 imo->imo_multicast_addr.s_addr = INADDR_ANY; 1775 break; 1776 1777 case IP_MULTICAST_TTL: 1778 /* 1779 * Set the IP time-to-live for outgoing multicast packets. 1780 */ 1781 error = ip_getoptval(sopt, &imo->imo_multicast_ttl, MAXTTL); 1782 break; 1783 1784 case IP_MULTICAST_LOOP: 1785 /* 1786 * Set the loopback flag for outgoing multicast packets. 1787 * Must be zero or one. 1788 */ 1789 error = ip_getoptval(sopt, &imo->imo_multicast_loop, 1); 1790 break; 1791 1792 case IP_ADD_MEMBERSHIP: /* IPV6_JOIN_GROUP */ 1793 error = ip_add_membership(imo, sopt); 1794 break; 1795 1796 case IP_DROP_MEMBERSHIP: /* IPV6_LEAVE_GROUP */ 1797 error = ip_drop_membership(imo, sopt); 1798 break; 1799 1800 default: 1801 error = EOPNOTSUPP; 1802 break; 1803 } 1804 1805 /* 1806 * If all options have default values, no need to keep the mbuf. 1807 */ 1808 if (imo->imo_multicast_ifp == NULL && 1809 imo->imo_multicast_ttl == IP_DEFAULT_MULTICAST_TTL && 1810 imo->imo_multicast_loop == IP_DEFAULT_MULTICAST_LOOP && 1811 imo->imo_num_memberships == 0) { 1812 kmem_free(imo, sizeof(*imo)); 1813 *pimo = NULL; 1814 } 1815 1816 return error; 1817 } 1818 1819 /* 1820 * Return the IP multicast options in response to user getsockopt(). 1821 */ 1822 int 1823 ip_getmoptions(struct ip_moptions *imo, struct sockopt *sopt) 1824 { 1825 struct in_addr addr; 1826 struct in_ifaddr *ia; 1827 uint8_t optval; 1828 int error = 0; 1829 1830 switch (sopt->sopt_name) { 1831 case IP_MULTICAST_IF: 1832 if (imo == NULL || imo->imo_multicast_ifp == NULL) 1833 addr = zeroin_addr; 1834 else if (imo->imo_multicast_addr.s_addr) { 1835 /* return the value user has set */ 1836 addr = imo->imo_multicast_addr; 1837 } else { 1838 IFP_TO_IA(imo->imo_multicast_ifp, ia); 1839 addr = ia ? ia->ia_addr.sin_addr : zeroin_addr; 1840 } 1841 error = sockopt_set(sopt, &addr, sizeof(addr)); 1842 break; 1843 1844 case IP_MULTICAST_TTL: 1845 optval = imo ? imo->imo_multicast_ttl 1846 : IP_DEFAULT_MULTICAST_TTL; 1847 1848 error = sockopt_set(sopt, &optval, sizeof(optval)); 1849 break; 1850 1851 case IP_MULTICAST_LOOP: 1852 optval = imo ? imo->imo_multicast_loop 1853 : IP_DEFAULT_MULTICAST_LOOP; 1854 1855 error = sockopt_set(sopt, &optval, sizeof(optval)); 1856 break; 1857 1858 default: 1859 error = EOPNOTSUPP; 1860 } 1861 1862 return error; 1863 } 1864 1865 /* 1866 * Discard the IP multicast options. 1867 */ 1868 void 1869 ip_freemoptions(struct ip_moptions *imo) 1870 { 1871 int i; 1872 1873 if (imo != NULL) { 1874 for (i = 0; i < imo->imo_num_memberships; ++i) 1875 in_delmulti(imo->imo_membership[i]); 1876 kmem_free(imo, sizeof(*imo)); 1877 } 1878 } 1879 1880 /* 1881 * Routine called from ip_output() to loop back a copy of an IP multicast 1882 * packet to the input queue of a specified interface. Note that this 1883 * calls the output routine of the loopback "driver", but with an interface 1884 * pointer that might NOT be lo0ifp -- easier than replicating that code here. 1885 */ 1886 static void 1887 ip_mloopback(struct ifnet *ifp, struct mbuf *m, const struct sockaddr_in *dst) 1888 { 1889 struct ip *ip; 1890 struct mbuf *copym; 1891 1892 copym = m_copypacket(m, M_DONTWAIT); 1893 if (copym != NULL && 1894 (copym->m_flags & M_EXT || copym->m_len < sizeof(struct ip))) 1895 copym = m_pullup(copym, sizeof(struct ip)); 1896 if (copym == NULL) 1897 return; 1898 /* 1899 * We don't bother to fragment if the IP length is greater 1900 * than the interface's MTU. Can this possibly matter? 1901 */ 1902 ip = mtod(copym, struct ip *); 1903 1904 if (copym->m_pkthdr.csum_flags & (M_CSUM_TCPv4|M_CSUM_UDPv4)) { 1905 in_delayed_cksum(copym); 1906 copym->m_pkthdr.csum_flags &= 1907 ~(M_CSUM_TCPv4|M_CSUM_UDPv4); 1908 } 1909 1910 ip->ip_sum = 0; 1911 ip->ip_sum = in_cksum(copym, ip->ip_hl << 2); 1912 #ifndef NET_MPSAFE 1913 KERNEL_LOCK(1, NULL); 1914 #endif 1915 (void)looutput(ifp, copym, sintocsa(dst), NULL); 1916 #ifndef NET_MPSAFE 1917 KERNEL_UNLOCK_ONE(NULL); 1918 #endif 1919 } 1920
Indexes created Wed Oct 01 12:09:54 GMT 2025