bsd/netinet/raw_ip.c

   1 /*
   2  * Copyright (c) 2000-2007 Apple Inc. All rights reserved.
   3  *
   4  * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
   5  *
   6  * This file contains Original Code and/or Modifications of Original Code
   7  * as defined in and that are subject to the Apple Public Source License
   8  * Version 2.0 (the 'License'). You may not use this file except in
   9  * compliance with the License. The rights granted to you under the License
  10  * may not be used to create, or enable the creation or redistribution of,
  11  * unlawful or unlicensed copies of an Apple operating system, or to
  12  * circumvent, violate, or enable the circumvention or violation of, any
  13  * terms of an Apple operating system software license agreement.
  14  *
  15  * Please obtain a copy of the License at
  16  * http://www.opensource.apple.com/apsl/ and read it before using this file.
  17  *
  18  * The Original Code and all software distributed under the License are
  19  * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
  20  * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
  21  * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
  22  * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
  23  * Please see the License for the specific language governing rights and
  24  * limitations under the License.
  25  *
  26  * @APPLE_OSREFERENCE_LICENSE_HEADER_END@
  27  */
  28 /*
  29  * Copyright (c) 1982, 1986, 1988, 1993
  30  *      The Regents of the University of California.  All rights reserved.
  31  *
  32  * Redistribution and use in source and binary forms, with or without
  33  * modification, are permitted provided that the following conditions
  34  * are met:
  35  * 1. Redistributions of source code must retain the above copyright
  36  *    notice, this list of conditions and the following disclaimer.
  37  * 2. Redistributions in binary form must reproduce the above copyright
  38  *    notice, this list of conditions and the following disclaimer in the
  39  *    documentation and/or other materials provided with the distribution.
  40  * 3. All advertising materials mentioning features or use of this software
  41  *    must display the following acknowledgement:
  42  *      This product includes software developed by the University of
  43  *      California, Berkeley and its contributors.
  44  * 4. 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  *      @(#)raw_ip.c    8.7 (Berkeley) 5/15/95
  61  */
  62 /*
  63  * NOTICE: This file was modified by SPARTA, Inc. in 2005 to introduce
  64  * support for mandatory and extensible security protections.  This notice
  65  * is included in support of clause 2.2 (b) of the Apple Public License,
  66  * Version 2.0.
  67  */
  68
  69 #include <sys/param.h>
  70 #include <sys/systm.h>
  71 #include <sys/kernel.h>
  72 #include <sys/malloc.h>
  73 #include <sys/mbuf.h>
  74 #include <sys/proc.h>
  75 #include <sys/domain.h>
  76 #include <sys/protosw.h>
  77 #include <sys/socket.h>
  78 #include <sys/socketvar.h>
  79 #include <sys/sysctl.h>
  80 #include <libkern/OSAtomic.h>
  81 #include <kern/zalloc.h>
  82
  83 #include <pexpert/pexpert.h>
  84
  85 #include <net/if.h>
  86 #include <net/route.h>
  87
  88 #define _IP_VHL
  89 #include <netinet/in.h>
  90 #include <netinet/in_systm.h>
  91 #include <netinet/ip.h>
  92 #include <netinet/in_pcb.h>
  93 #include <netinet/in_var.h>
  94 #include <netinet/ip_var.h>
  95 #include <netinet/ip_mroute.h>
  96
  97 #include <netinet/ip_fw.h>
  98
  99 #if IPSEC
 100 #include <netinet6/ipsec.h>
 101 #endif /*IPSEC*/
 102
 103 #if DUMMYNET
 104 #include <netinet/ip_dummynet.h>
 105 #endif
 106
 107 #if CONFIG_MACF_NET
 108 #include <security/mac_framework.h>
 109 #endif /* MAC_NET */
 110
 111 int load_ipfw(void);
 112 int rip_detach(struct socket *);
 113 int rip_abort(struct socket *);
 114 int rip_disconnect(struct socket *);
 115 int rip_bind(struct socket *, struct sockaddr *, struct proc *);
 116 int rip_connect(struct socket *, struct sockaddr *, struct proc *);
 117 int rip_shutdown(struct socket *);
 118
 119 #if IPSEC
 120 extern int ipsec_bypass;
 121 #endif
 122
 123 extern u_long  route_generation;
 124 struct  inpcbhead ripcb;
 125 struct  inpcbinfo ripcbinfo;
 126
 127 /* control hooks for ipfw and dummynet */
 128 #if IPFIREWALL
 129 ip_fw_ctl_t *ip_fw_ctl_ptr;
 130 #if DUMMYNET
 131 ip_dn_ctl_t *ip_dn_ctl_ptr;
 132 #endif /* DUMMYNET */
 133 #endif /* IPFIREWALL */
 134
 135 /*
 136  * Nominal space allocated to a raw ip socket.
 137  */
 138 #define RIPSNDQ         8192
 139 #define RIPRCVQ         8192
 140
 141 /*
 142  * Raw interface to IP protocol.
 143  */
 144
 145 /*
 146  * Initialize raw connection block q.
 147  */
 148 void
 149 rip_init()
 150 {
 151         struct inpcbinfo *pcbinfo;
 152
 153         LIST_INIT(&ripcb);
 154         ripcbinfo.listhead = &ripcb;
 155         /*
 156          * XXX We don't use the hash list for raw IP, but it's easier
 157          * to allocate a one entry hash list than it is to check all
 158          * over the place for hashbase == NULL.
 159          */
 160         ripcbinfo.hashbase = hashinit(1, M_PCB, &ripcbinfo.hashmask);
 161         ripcbinfo.porthashbase = hashinit(1, M_PCB, &ripcbinfo.porthashmask);
 162
 163         ripcbinfo.ipi_zone = (void *) zinit(sizeof(struct inpcb),
 164                                             (4096 * sizeof(struct inpcb)),
 165                                             4096, "ripzone");
 166
 167         pcbinfo = &ripcbinfo;
 168         /*
 169          * allocate lock group attribute and group for udp pcb mutexes
 170          */
 171         pcbinfo->mtx_grp_attr = lck_grp_attr_alloc_init();
 172
 173         pcbinfo->mtx_grp = lck_grp_alloc_init("ripcb", pcbinfo->mtx_grp_attr);
 174
 175         /*
 176          * allocate the lock attribute for udp pcb mutexes
 177          */
 178         pcbinfo->mtx_attr = lck_attr_alloc_init();
 179
 180         if ((pcbinfo->mtx = lck_rw_alloc_init(pcbinfo->mtx_grp, pcbinfo->mtx_attr)) == NULL)
 181                 return; /* pretty much dead if this fails... */
 182
 183 }
 184
 185 static struct   sockaddr_in ripsrc = { sizeof(ripsrc), AF_INET , 0, {0}, {0,0,0,0,0,0,0,0,} };
 186 /*
 187  * Setup generic address and protocol structures
 188  * for raw_input routine, then pass them along with
 189  * mbuf chain.
 190  */
 191 void
 192 rip_input(m, iphlen)
 193         struct mbuf *m;
 194         int iphlen;
 195 {
 196         register struct ip *ip = mtod(m, struct ip *);
 197         register struct inpcb *inp;
 198         struct inpcb *last = 0;
 199         struct mbuf *opts = 0;
 200         int skipit;
 201
 202         ripsrc.sin_addr = ip->ip_src;
 203         lck_rw_lock_shared(ripcbinfo.mtx);
 204         LIST_FOREACH(inp, &ripcb, inp_list) {
 205 #if INET6
 206                 if ((inp->inp_vflag & INP_IPV4) == 0)
 207                         continue;
 208 #endif
 209                 if (inp->inp_ip_p && (inp->inp_ip_p != ip->ip_p))
 210                         continue;
 211                 if (inp->inp_laddr.s_addr &&
 212                   inp->inp_laddr.s_addr != ip->ip_dst.s_addr)
 213                         continue;
 214                 if (inp->inp_faddr.s_addr &&
 215                   inp->inp_faddr.s_addr != ip->ip_src.s_addr)
 216                         continue;
 217                 if (last) {
 218                         struct mbuf *n = m_copy(m, 0, (int)M_COPYALL);
 219
 220 #if IPSEC
 221                         /* check AH/ESP integrity. */
 222                         skipit = 0;
 223                         if (ipsec_bypass == 0 && n) {
 224                                 if (ipsec4_in_reject_so(n, last->inp_socket)) {
 225                                         m_freem(n);
 226                                         IPSEC_STAT_INCREMENT(ipsecstat.in_polvio);
 227                                         /* do not inject data to pcb */
 228                                         skipit = 1;
 229                                 }
 230                         }
 231 #endif /*IPSEC*/
 232 #if CONFIG_MACF_NET
 233                         if (n && skipit == 0) {
 234                                 if (mac_inpcb_check_deliver(last, n, AF_INET,
 235                                     SOCK_RAW) != 0)
 236                                         skipit = 1;
 237                         }
 238 #endif
 239                         if (n && skipit == 0) {
 240                                 int error = 0;
 241                                 if (last->inp_flags & INP_CONTROLOPTS ||
 242                                     last->inp_socket->so_options & SO_TIMESTAMP)
 243                                     ip_savecontrol(last, &opts, ip, n);
 244                                 if (last->inp_flags & INP_STRIPHDR) {
 245                                         n->m_len -= iphlen;
 246                                         n->m_pkthdr.len -= iphlen;
 247                                         n->m_data += iphlen;
 248                                 }
 249 // ###LOCK need to lock that socket?
 250                                 if (sbappendaddr(&last->inp_socket->so_rcv,
 251                                     (struct sockaddr *)&ripsrc, n,
 252                                     opts, &error) != 0) {
 253                                         sorwakeup(last->inp_socket);
 254                                 }
 255                                 else {
 256                                         if (error) {
 257                                                 /* should notify about lost packet */
 258                                                 kprintf("rip_input can't append to socket\n");
 259                                         }
 260                                 }
 261                                 opts = 0;
 262                         }
 263                 }
 264                 last = inp;
 265         }
 266         lck_rw_done(ripcbinfo.mtx);
 267 #if IPSEC
 268         /* check AH/ESP integrity. */
 269         skipit = 0;
 270         if (ipsec_bypass == 0 && last) {
 271                 if (ipsec4_in_reject_so(m, last->inp_socket)) {
 272                         m_freem(m);
 273                         IPSEC_STAT_INCREMENT(ipsecstat.in_polvio);
 274                         OSAddAtomic(1, (SInt32*)&ipstat.ips_delivered);
 275                         /* do not inject data to pcb */
 276                         skipit = 1;
 277                 }
 278         }
 279 #endif /*IPSEC*/
 280 #if CONFIG_MACF_NET
 281         if (last && skipit == 0) {
 282                 if (mac_inpcb_check_deliver(last, m, AF_INET, SOCK_RAW) != 0)
 283                         skipit = 1;
 284         }
 285 #endif
 286         if (skipit == 0) {
 287                 if (last) {
 288                         if (last->inp_flags & INP_CONTROLOPTS ||
 289                                 last->inp_socket->so_options & SO_TIMESTAMP)
 290                                 ip_savecontrol(last, &opts, ip, m);
 291                         if (last->inp_flags & INP_STRIPHDR) {
 292                                 m->m_len -= iphlen;
 293                                 m->m_pkthdr.len -= iphlen;
 294                                 m->m_data += iphlen;
 295                         }
 296                         if (sbappendaddr(&last->inp_socket->so_rcv,
 297                                 (struct sockaddr *)&ripsrc, m, opts, NULL) != 0) {
 298                                 sorwakeup(last->inp_socket);
 299                         } else {
 300                                 kprintf("rip_input(2) can't append to socket\n");
 301                         }
 302                 } else {
 303                         m_freem(m);
 304                         OSAddAtomic(1, (SInt32*)&ipstat.ips_noproto);
 305                         OSAddAtomic(-1, (SInt32*)&ipstat.ips_delivered);
 306                 }
 307         }
 308 }
 309
 310 /*
 311  * Generate IP header and pass packet to ip_output.
 312  * Tack on options user may have setup with control call.
 313  */
 314 int
 315 rip_output(m, so, dst)
 316         register struct mbuf *m;
 317         struct socket *so;
 318         u_long dst;
 319 {
 320         register struct ip *ip;
 321         register struct inpcb *inp = sotoinpcb(so);
 322         int flags = (so->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST;
 323
 324         /*
 325          * If the user handed us a complete IP packet, use it.
 326          * Otherwise, allocate an mbuf for a header and fill it in.
 327          */
 328         if ((inp->inp_flags & INP_HDRINCL) == 0) {
 329                 if (m->m_pkthdr.len + sizeof(struct ip) > IP_MAXPACKET) {
 330                         m_freem(m);
 331                         return(EMSGSIZE);
 332                 }
 333                 M_PREPEND(m, sizeof(struct ip), M_WAIT);
 334                 ip = mtod(m, struct ip *);
 335                 ip->ip_tos = inp->inp_ip_tos;
 336                 ip->ip_off = 0;
 337                 ip->ip_p = inp->inp_ip_p;
 338                 ip->ip_len = m->m_pkthdr.len;
 339                 ip->ip_src = inp->inp_laddr;
 340                 ip->ip_dst.s_addr = dst;
 341                 ip->ip_ttl = inp->inp_ip_ttl;
 342         } else {
 343                 if (m->m_pkthdr.len > IP_MAXPACKET) {
 344                         m_freem(m);
 345                         return(EMSGSIZE);
 346                 }
 347                 ip = mtod(m, struct ip *);
 348                 /* don't allow both user specified and setsockopt options,
 349                    and don't allow packet length sizes that will crash */
 350                 if (((IP_VHL_HL(ip->ip_vhl) != (sizeof (*ip) >> 2))
 351                      && inp->inp_options)
 352                     || (ip->ip_len > m->m_pkthdr.len)
 353                     || (ip->ip_len < (IP_VHL_HL(ip->ip_vhl) << 2))) {
 354                         m_freem(m);
 355                         return EINVAL;
 356                 }
 357                 if (ip->ip_id == 0)
 358 #if RANDOM_IP_ID
 359                         ip->ip_id = ip_randomid();
 360 #else
 361                         ip->ip_id = htons(ip_id++);
 362 #endif
 363                 /* XXX prevent ip_output from overwriting header fields */
 364                 flags |= IP_RAWOUTPUT;
 365                 OSAddAtomic(1, (SInt32*)&ipstat.ips_rawout);
 366         }
 367
 368 #if IPSEC
 369         if (ipsec_bypass == 0 && ipsec_setsocket(m, so) != 0) {
 370                 m_freem(m);
 371                 return ENOBUFS;
 372         }
 373 #endif /*IPSEC*/
 374
 375         if (inp->inp_route.ro_rt && inp->inp_route.ro_rt->generation_id != route_generation) {
 376                 rtfree(inp->inp_route.ro_rt);
 377                 inp->inp_route.ro_rt = (struct rtentry *)0;
 378         }
 379
 380 #if CONFIG_MACF_NET
 381         mac_mbuf_label_associate_inpcb(inp, m);
 382 #endif
 383
 384 #if CONFIG_FORCE_OUT_IFP
 385         return (ip_output_list(m, 0, inp->inp_options, &inp->inp_route, flags,
 386                           inp->inp_moptions, inp->pdp_ifp));
 387 #else
 388         return (ip_output_list(m, 0, inp->inp_options, &inp->inp_route, flags,
 389                           inp->inp_moptions, NULL));
 390 #endif
 391 }
 392
 393 #if IPFIREWALL
 394 int
 395 load_ipfw(void)
 396 {
 397         kern_return_t   err;
 398
 399         ipfw_init();
 400
 401 #if DUMMYNET
 402         if (!DUMMYNET_LOADED)
 403                 ip_dn_init();
 404 #endif /* DUMMYNET */
 405         err = 0;
 406
 407         return err == 0 && ip_fw_ctl_ptr == NULL ? -1 : err;
 408 }
 409 #endif /* IPFIREWALL */
 410
 411 /*
 412  * Raw IP socket option processing.
 413  */
 414 int
 415 rip_ctloutput(so, sopt)
 416         struct socket *so;
 417         struct sockopt *sopt;
 418 {
 419         struct  inpcb *inp = sotoinpcb(so);
 420         int     error, optval;
 421
 422         if (sopt->sopt_level != IPPROTO_IP)
 423                 return (EINVAL);
 424
 425         error = 0;
 426
 427         switch (sopt->sopt_dir) {
 428         case SOPT_GET:
 429                 switch (sopt->sopt_name) {
 430                 case IP_HDRINCL:
 431                         optval = inp->inp_flags & INP_HDRINCL;
 432                         error = sooptcopyout(sopt, &optval, sizeof optval);
 433                         break;
 434
 435         case IP_STRIPHDR:
 436             optval = inp->inp_flags & INP_STRIPHDR;
 437             error = sooptcopyout(sopt, &optval, sizeof optval);
 438             break;
 439
 440 #if IPFIREWALL
 441                 case IP_FW_ADD:
 442                 case IP_FW_GET:
 443                 case IP_OLD_FW_ADD:
 444                 case IP_OLD_FW_GET:
 445                         if (ip_fw_ctl_ptr == 0)
 446                                 error = load_ipfw();
 447                         if (ip_fw_ctl_ptr && error == 0)
 448                                 error = ip_fw_ctl_ptr(sopt);
 449                         else
 450                                 error = ENOPROTOOPT;
 451                         break;
 452 #endif IPFIREWALL
 453
 454 #if DUMMYNET
 455                 case IP_DUMMYNET_GET:
 456                         if (DUMMYNET_LOADED)
 457                                 error = ip_dn_ctl_ptr(sopt);
 458                         else
 459                                 error = ENOPROTOOPT;
 460                         break ;
 461 #endif /* DUMMYNET */
 462
 463 #if MROUTING
 464                 case MRT_INIT:
 465                 case MRT_DONE:
 466                 case MRT_ADD_VIF:
 467                 case MRT_DEL_VIF:
 468                 case MRT_ADD_MFC:
 469                 case MRT_DEL_MFC:
 470                 case MRT_VERSION:
 471                 case MRT_ASSERT:
 472                         error = ip_mrouter_get(so, sopt);
 473                         break;
 474 #endif /* MROUTING */
 475
 476                 default:
 477                         error = ip_ctloutput(so, sopt);
 478                         break;
 479                 }
 480                 break;
 481
 482         case SOPT_SET:
 483                 switch (sopt->sopt_name) {
 484                 case IP_HDRINCL:
 485                         error = sooptcopyin(sopt, &optval, sizeof optval,
 486                                             sizeof optval);
 487                         if (error)
 488                                 break;
 489                         if (optval)
 490                                 inp->inp_flags |= INP_HDRINCL;
 491                         else
 492                                 inp->inp_flags &= ~INP_HDRINCL;
 493                         break;
 494
 495         case IP_STRIPHDR:
 496             error = sooptcopyin(sopt, &optval, sizeof optval,
 497                         sizeof optval);
 498             if (error)
 499                 break;
 500             if (optval)
 501                 inp->inp_flags |= INP_STRIPHDR;
 502             else
 503                 inp->inp_flags &= ~INP_STRIPHDR;
 504             break;
 505
 506
 507 #if IPFIREWALL
 508                 case IP_FW_ADD:
 509                 case IP_FW_DEL:
 510                 case IP_FW_FLUSH:
 511                 case IP_FW_ZERO:
 512                 case IP_FW_RESETLOG:
 513                 case IP_OLD_FW_ADD:
 514                 case IP_OLD_FW_DEL:
 515                 case IP_OLD_FW_FLUSH:
 516                 case IP_OLD_FW_ZERO:
 517                 case IP_OLD_FW_RESETLOG:
 518                         if (ip_fw_ctl_ptr == 0)
 519                                 error = load_ipfw();
 520                         if (ip_fw_ctl_ptr && error == 0)
 521                                 error = ip_fw_ctl_ptr(sopt);
 522                         else
 523                                 error = ENOPROTOOPT;
 524                         break;
 525 #endif /* IPFIREWALL */
 526
 527 #if DUMMYNET
 528                 case IP_DUMMYNET_CONFIGURE:
 529                 case IP_DUMMYNET_DEL:
 530                 case IP_DUMMYNET_FLUSH:
 531                         if (DUMMYNET_LOADED)
 532                                 error = ip_dn_ctl_ptr(sopt);
 533                         else
 534                                 error = ENOPROTOOPT ;
 535                         break ;
 536 #endif
 537
 538 #if MROUTING
 539                 case IP_RSVP_ON:
 540                         error = ip_rsvp_init(so);
 541                         break;
 542
 543                 case IP_RSVP_OFF:
 544                         error = ip_rsvp_done();
 545                         break;
 546
 547                         /* XXX - should be combined */
 548                 case IP_RSVP_VIF_ON:
 549                         error = ip_rsvp_vif_init(so, sopt);
 550                         break;
 551
 552                 case IP_RSVP_VIF_OFF:
 553                         error = ip_rsvp_vif_done(so, sopt);
 554                         break;
 555
 556                 case MRT_INIT:
 557                 case MRT_DONE:
 558                 case MRT_ADD_VIF:
 559                 case MRT_DEL_VIF:
 560                 case MRT_ADD_MFC:
 561                 case MRT_DEL_MFC:
 562                 case MRT_VERSION:
 563                 case MRT_ASSERT:
 564                         error = ip_mrouter_set(so, sopt);
 565                         break;
 566 #endif /* MROUTING */
 567
 568                 default:
 569                         error = ip_ctloutput(so, sopt);
 570                         break;
 571                 }
 572                 break;
 573         }
 574
 575         return (error);
 576 }
 577
 578 /*
 579  * This function exists solely to receive the PRC_IFDOWN messages which
 580  * are sent by if_down().  It looks for an ifaddr whose ifa_addr is sa,
 581  * and calls in_ifadown() to remove all routes corresponding to that address.
 582  * It also receives the PRC_IFUP messages from if_up() and reinstalls the
 583  * interface routes.
 584  */
 585 void
 586 rip_ctlinput(
 587         int cmd,
 588         struct sockaddr *sa,
 589         __unused void *vip)
 590 {
 591         struct in_ifaddr *ia;
 592         struct ifnet *ifp;
 593         int err;
 594         int flags;
 595
 596         switch (cmd) {
 597         case PRC_IFDOWN:
 598                 lck_mtx_lock(rt_mtx);
 599                 for (ia = in_ifaddrhead.tqh_first; ia;
 600                      ia = ia->ia_link.tqe_next) {
 601                         if (ia->ia_ifa.ifa_addr == sa
 602                             && (ia->ia_flags & IFA_ROUTE)) {
 603                                 /*
 604                                  * in_ifscrub kills the interface route.
 605                                  */
 606                                 in_ifscrub(ia->ia_ifp, ia, 1);
 607                                 /*
 608                                  * in_ifadown gets rid of all the rest of
 609                                  * the routes.  This is not quite the right
 610                                  * thing to do, but at least if we are running
 611                                  * a routing process they will come back.
 612                                  */
 613                                 in_ifadown(&ia->ia_ifa, 1);
 614                                 break;
 615                         }
 616                 }
 617                 lck_mtx_unlock(rt_mtx);
 618                 break;
 619
 620         case PRC_IFUP:
 621                 lck_mtx_lock(rt_mtx);
 622                 for (ia = in_ifaddrhead.tqh_first; ia;
 623                      ia = ia->ia_link.tqe_next) {
 624                         if (ia->ia_ifa.ifa_addr == sa)
 625                                 break;
 626                 }
 627                 if (ia == 0 || (ia->ia_flags & IFA_ROUTE)) {
 628                         lck_mtx_unlock(rt_mtx);
 629                         return;
 630                 }
 631                 flags = RTF_UP;
 632                 ifp = ia->ia_ifa.ifa_ifp;
 633
 634                 if ((ifp->if_flags & IFF_LOOPBACK)
 635                     || (ifp->if_flags & IFF_POINTOPOINT))
 636                         flags |= RTF_HOST;
 637
 638                 err = rtinit_locked(&ia->ia_ifa, RTM_ADD, flags);
 639                 lck_mtx_unlock(rt_mtx);
 640                 if (err == 0)
 641                         ia->ia_flags |= IFA_ROUTE;
 642                 break;
 643         }
 644 }
 645
 646 u_long  rip_sendspace = RIPSNDQ;
 647 u_long  rip_recvspace = RIPRCVQ;
 648
 649 SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW,
 650     &rip_sendspace, 0, "Maximum outgoing raw IP datagram size");
 651 SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW,
 652     &rip_recvspace, 0, "Maximum incoming raw IP datagram size");
 653
 654 static int
 655 rip_attach(struct socket *so, int proto, struct proc *p)
 656 {
 657         struct inpcb *inp;
 658         int error;
 659
 660         inp = sotoinpcb(so);
 661         if (inp)
 662                 panic("rip_attach");
 663         if ((so->so_state & SS_PRIV) == 0)
 664                 return (EPERM);
 665
 666         error = soreserve(so, rip_sendspace, rip_recvspace);
 667         if (error)
 668                 return error;
 669         error = in_pcballoc(so, &ripcbinfo, p);
 670         if (error)
 671                 return error;
 672         inp = (struct inpcb *)so->so_pcb;
 673         inp->inp_vflag |= INP_IPV4;
 674         inp->inp_ip_p = proto;
 675         inp->inp_ip_ttl = ip_defttl;
 676         return 0;
 677 }
 678
 679 __private_extern__ int
 680 rip_detach(struct socket *so)
 681 {
 682         struct inpcb *inp;
 683
 684         inp = sotoinpcb(so);
 685         if (inp == 0)
 686                 panic("rip_detach");
 687 #if MROUTING
 688         if (so == ip_mrouter)
 689                 ip_mrouter_done();
 690         ip_rsvp_force_done(so);
 691         if (so == ip_rsvpd)
 692                 ip_rsvp_done();
 693 #endif /* MROUTING */
 694         in_pcbdetach(inp);
 695         return 0;
 696 }
 697
 698 __private_extern__ int
 699 rip_abort(struct socket *so)
 700 {
 701         soisdisconnected(so);
 702         return rip_detach(so);
 703 }
 704
 705 __private_extern__ int
 706 rip_disconnect(struct socket *so)
 707 {
 708         if ((so->so_state & SS_ISCONNECTED) == 0)
 709                 return ENOTCONN;
 710         return rip_abort(so);
 711 }
 712
 713 __private_extern__ int
 714 rip_bind(struct socket *so, struct sockaddr *nam, __unused struct proc *p)
 715 {
 716         struct inpcb *inp = sotoinpcb(so);
 717         struct sockaddr_in *addr = (struct sockaddr_in *)nam;
 718         struct ifaddr *ifa = NULL;
 719
 720         if (nam->sa_len != sizeof(*addr))
 721                 return EINVAL;
 722
 723         if (TAILQ_EMPTY(&ifnet_head) || ((addr->sin_family != AF_INET) &&
 724                                     (addr->sin_family != AF_IMPLINK)) ||
 725             (addr->sin_addr.s_addr &&
 726              (ifa = ifa_ifwithaddr((struct sockaddr *)addr)) == 0)) {
 727                 return EADDRNOTAVAIL;
 728         }
 729         else if (ifa) {
 730                 ifafree(ifa);
 731                 ifa = NULL;
 732         }
 733         inp->inp_laddr = addr->sin_addr;
 734         return 0;
 735 }
 736
 737 __private_extern__ int
 738 rip_connect(struct socket *so, struct sockaddr *nam, __unused  struct proc *p)
 739 {
 740         struct inpcb *inp = sotoinpcb(so);
 741         struct sockaddr_in *addr = (struct sockaddr_in *)nam;
 742
 743         if (nam->sa_len != sizeof(*addr))
 744                 return EINVAL;
 745         if (TAILQ_EMPTY(&ifnet_head))
 746                 return EADDRNOTAVAIL;
 747         if ((addr->sin_family != AF_INET) &&
 748             (addr->sin_family != AF_IMPLINK))
 749                 return EAFNOSUPPORT;
 750         inp->inp_faddr = addr->sin_addr;
 751         soisconnected(so);
 752         return 0;
 753 }
 754
 755 __private_extern__ int
 756 rip_shutdown(struct socket *so)
 757 {
 758         socantsendmore(so);
 759         return 0;
 760 }
 761
 762 __private_extern__ int
 763 rip_send(struct socket *so, __unused int flags, struct mbuf *m, struct sockaddr *nam,
 764          __unused struct mbuf *control, __unused struct proc *p)
 765 {
 766         struct inpcb *inp = sotoinpcb(so);
 767         register u_long dst;
 768
 769         if (so->so_state & SS_ISCONNECTED) {
 770                 if (nam) {
 771                         m_freem(m);
 772                         return EISCONN;
 773                 }
 774                 dst = inp->inp_faddr.s_addr;
 775         } else {
 776                 if (nam == NULL) {
 777                         m_freem(m);
 778                         return ENOTCONN;
 779                 }
 780                 dst = ((struct sockaddr_in *)nam)->sin_addr.s_addr;
 781         }
 782         return rip_output(m, so, dst);
 783 }
 784
 785 /* note: rip_unlock is called from different protos  instead of the generic socket_unlock,
 786  * it will handle the socket dealloc on last reference
 787  * */
 788 int
 789 rip_unlock(struct socket *so, int refcount, int debug)
 790 {
 791         int lr_saved;
 792         struct inpcb *inp = sotoinpcb(so);
 793
 794         if (debug == 0)
 795                 lr_saved = (unsigned int) __builtin_return_address(0);
 796         else lr_saved = debug;
 797
 798         if (refcount) {
 799                 if (so->so_usecount <= 0)
 800                         panic("rip_unlock: bad refoucnt so=%p val=%x\n", so, so->so_usecount);
 801                 so->so_usecount--;
 802                 if (so->so_usecount == 0 && (inp->inp_wantcnt == WNT_STOPUSING)) {
 803                         /* cleanup after last reference */
 804                         lck_mtx_unlock(so->so_proto->pr_domain->dom_mtx);
 805                         lck_rw_lock_exclusive(ripcbinfo.mtx);
 806                         in_pcbdispose(inp);
 807                         lck_rw_done(ripcbinfo.mtx);
 808                         return(0);
 809                 }
 810         }
 811         so->unlock_lr[so->next_unlock_lr] = (u_int32_t)lr_saved;
 812         so->next_unlock_lr = (so->next_unlock_lr+1) % SO_LCKDBG_MAX;
 813         lck_mtx_unlock(so->so_proto->pr_domain->dom_mtx);
 814         return(0);
 815 }
 816
 817 static int
 818 rip_pcblist SYSCTL_HANDLER_ARGS
 819 {
 820 #pragma unused(oidp, arg1, arg2)
 821         int error, i, n;
 822         struct inpcb *inp, **inp_list;
 823         inp_gen_t gencnt;
 824         struct xinpgen xig;
 825
 826         /*
 827          * The process of preparing the TCB list is too time-consuming and
 828          * resource-intensive to repeat twice on every request.
 829          */
 830         lck_rw_lock_exclusive(ripcbinfo.mtx);
 831         if (req->oldptr == USER_ADDR_NULL) {
 832                 n = ripcbinfo.ipi_count;
 833                 req->oldidx = 2 * (sizeof xig)
 834                         + (n + n/8) * sizeof(struct xinpcb);
 835                 lck_rw_done(ripcbinfo.mtx);
 836                 return 0;
 837         }
 838
 839         if (req->newptr != USER_ADDR_NULL) {
 840                 lck_rw_done(ripcbinfo.mtx);
 841                 return EPERM;
 842         }
 843
 844         /*
 845          * OK, now we're committed to doing something.
 846          */
 847         gencnt = ripcbinfo.ipi_gencnt;
 848         n = ripcbinfo.ipi_count;
 849
 850         bzero(&xig, sizeof(xig));
 851         xig.xig_len = sizeof xig;
 852         xig.xig_count = n;
 853         xig.xig_gen = gencnt;
 854         xig.xig_sogen = so_gencnt;
 855         error = SYSCTL_OUT(req, &xig, sizeof xig);
 856         if (error) {
 857                 lck_rw_done(ripcbinfo.mtx);
 858                 return error;
 859         }
 860     /*
 861      * We are done if there is no pcb
 862      */
 863     if (n == 0) {
 864         lck_rw_done(ripcbinfo.mtx);
 865         return 0;
 866     }
 867
 868         inp_list = _MALLOC(n * sizeof *inp_list, M_TEMP, M_WAITOK);
 869         if (inp_list == 0) {
 870                 lck_rw_done(ripcbinfo.mtx);
 871                 return ENOMEM;
 872         }
 873
 874         for (inp = ripcbinfo.listhead->lh_first, i = 0; inp && i < n;
 875              inp = inp->inp_list.le_next) {
 876                 if (inp->inp_gencnt <= gencnt && inp->inp_state != INPCB_STATE_DEAD)
 877                         inp_list[i++] = inp;
 878         }
 879         n = i;
 880
 881         error = 0;
 882         for (i = 0; i < n; i++) {
 883                 inp = inp_list[i];
 884                 if (inp->inp_gencnt <= gencnt && inp->inp_state != INPCB_STATE_DEAD) {
 885                         struct xinpcb xi;
 886
 887                         bzero(&xi, sizeof(xi));
 888                         xi.xi_len = sizeof xi;
 889                         /* XXX should avoid extra copy */
 890                         inpcb_to_compat(inp, &xi.xi_inp);
 891                         if (inp->inp_socket)
 892                                 sotoxsocket(inp->inp_socket, &xi.xi_socket);
 893                         error = SYSCTL_OUT(req, &xi, sizeof xi);
 894                 }
 895         }
 896         if (!error) {
 897                 /*
 898                  * Give the user an updated idea of our state.
 899                  * If the generation differs from what we told
 900                  * her before, she knows that something happened
 901                  * while we were processing this request, and it
 902                  * might be necessary to retry.
 903                  */
 904                 bzero(&xig, sizeof(xig));
 905                 xig.xig_len = sizeof xig;
 906                 xig.xig_gen = ripcbinfo.ipi_gencnt;
 907                 xig.xig_sogen = so_gencnt;
 908                 xig.xig_count = ripcbinfo.ipi_count;
 909                 error = SYSCTL_OUT(req, &xig, sizeof xig);
 910         }
 911         FREE(inp_list, M_TEMP);
 912         lck_rw_done(ripcbinfo.mtx);
 913         return error;
 914 }
 915
 916 SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist, CTLFLAG_RD, 0, 0,
 917             rip_pcblist, "S,xinpcb", "List of active raw IP sockets");
 918
 919 struct pr_usrreqs rip_usrreqs = {
 920         rip_abort, pru_accept_notsupp, rip_attach, rip_bind, rip_connect,
 921         pru_connect2_notsupp, in_control, rip_detach, rip_disconnect,
 922         pru_listen_notsupp, in_setpeeraddr, pru_rcvd_notsupp,
 923         pru_rcvoob_notsupp, rip_send, pru_sense_null, rip_shutdown,
 924         in_setsockaddr, sosend, soreceive, pru_sopoll_notsupp
 925 };
 926 /* DSEP Review Done pl-20051213-v02 @3253 */