bsd/netinet/raw_ip.c

   1 /*
   2  * Copyright (c) 2006 Apple Computer, Inc. All Rights Reserved.
   3  *
   4  * @APPLE_LICENSE_OSREFERENCE_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
  10  * License may not be used to create, or enable the creation or
  11  * redistribution of, unlawful or unlicensed copies of an Apple operating
  12  * system, or to circumvent, violate, or enable the circumvention or
  13  * violation of, any terms of an Apple operating system software license
  14  * agreement.
  15  *
  16  * Please obtain a copy of the License at
  17  * http://www.opensource.apple.com/apsl/ and read it before using this
  18  * file.
  19  *
  20  * The Original Code and all software distributed under the License are
  21  * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
  22  * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
  23  * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
  24  * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
  25  * Please see the License for the specific language governing rights and
  26  * limitations under the License.
  27  *
  28  * @APPLE_LICENSE_OSREFERENCE_HEADER_END@
  29  */
  30 /*
  31  * Copyright (c) 1982, 1986, 1988, 1993
  32  *      The Regents of the University of California.  All rights reserved.
  33  *
  34  * Redistribution and use in source and binary forms, with or without
  35  * modification, are permitted provided that the following conditions
  36  * are met:
  37  * 1. Redistributions of source code must retain the above copyright
  38  *    notice, this list of conditions and the following disclaimer.
  39  * 2. Redistributions in binary form must reproduce the above copyright
  40  *    notice, this list of conditions and the following disclaimer in the
  41  *    documentation and/or other materials provided with the distribution.
  42  * 3. All advertising materials mentioning features or use of this software
  43  *    must display the following acknowledgement:
  44  *      This product includes software developed by the University of
  45  *      California, Berkeley and its contributors.
  46  * 4. Neither the name of the University nor the names of its contributors
  47  *    may be used to endorse or promote products derived from this software
  48  *    without specific prior written permission.
  49  *
  50  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  51  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  52  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  53  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  54  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  55  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  56  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  57  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  58  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  59  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  60  * SUCH DAMAGE.
  61  *
  62  *      @(#)raw_ip.c    8.7 (Berkeley) 5/15/95
  63  */
  64
  65 #include <sys/param.h>
  66 #include <sys/systm.h>
  67 #include <sys/kernel.h>
  68 #include <sys/malloc.h>
  69 #include <sys/mbuf.h>
  70 #include <sys/proc.h>
  71 #include <sys/domain.h>
  72 #include <sys/protosw.h>
  73 #include <sys/socket.h>
  74 #include <sys/socketvar.h>
  75 #include <sys/sysctl.h>
  76
  77 #if __FreeBSD__
  78 #include <vm/vm_zone.h>
  79 #endif
  80
  81 #include <net/if.h>
  82 #include <net/route.h>
  83
  84 #define _IP_VHL
  85 #include <netinet/in.h>
  86 #include <netinet/in_systm.h>
  87 #include <netinet/ip.h>
  88 #include <netinet/in_pcb.h>
  89 #include <netinet/in_var.h>
  90 #include <netinet/ip_var.h>
  91 #include <netinet/ip_mroute.h>
  92
  93 #include <netinet/ip_fw.h>
  94
  95 #if IPSEC
  96 #include <netinet6/ipsec.h>
  97 #endif /*IPSEC*/
  98
  99 #if DUMMYNET
 100 #include <netinet/ip_dummynet.h>
 101 #endif
 102
 103 #if IPSEC
 104 extern int ipsec_bypass;
 105 extern lck_mtx_t *sadb_mutex;
 106 #endif
 107
 108 extern u_long  route_generation;
 109 struct  inpcbhead ripcb;
 110 struct  inpcbinfo ripcbinfo;
 111
 112 /* control hooks for ipfw and dummynet */
 113 ip_fw_ctl_t *ip_fw_ctl_ptr;
 114 #if DUMMYNET
 115 ip_dn_ctl_t *ip_dn_ctl_ptr;
 116 #endif /* DUMMYNET */
 117
 118 /*
 119  * Nominal space allocated to a raw ip socket.
 120  */
 121 #define RIPSNDQ         8192
 122 #define RIPRCVQ         8192
 123
 124 /*
 125  * Raw interface to IP protocol.
 126  */
 127
 128 /*
 129  * Initialize raw connection block q.
 130  */
 131 void
 132 rip_init()
 133 {
 134         struct inpcbinfo *pcbinfo;
 135
 136         LIST_INIT(&ripcb);
 137         ripcbinfo.listhead = &ripcb;
 138         /*
 139          * XXX We don't use the hash list for raw IP, but it's easier
 140          * to allocate a one entry hash list than it is to check all
 141          * over the place for hashbase == NULL.
 142          */
 143         ripcbinfo.hashbase = hashinit(1, M_PCB, &ripcbinfo.hashmask);
 144         ripcbinfo.porthashbase = hashinit(1, M_PCB, &ripcbinfo.porthashmask);
 145
 146         ripcbinfo.ipi_zone = (void *) zinit(sizeof(struct inpcb),
 147                                             (4096 * sizeof(struct inpcb)),
 148                                             4096, "ripzone");
 149
 150         pcbinfo = &ripcbinfo;
 151         /*
 152          * allocate lock group attribute and group for udp pcb mutexes
 153          */
 154         pcbinfo->mtx_grp_attr = lck_grp_attr_alloc_init();
 155         lck_grp_attr_setdefault(pcbinfo->mtx_grp_attr);
 156
 157         pcbinfo->mtx_grp = lck_grp_alloc_init("ripcb", pcbinfo->mtx_grp_attr);
 158
 159         /*
 160          * allocate the lock attribute for udp pcb mutexes
 161          */
 162         pcbinfo->mtx_attr = lck_attr_alloc_init();
 163         lck_attr_setdefault(pcbinfo->mtx_attr);
 164
 165         if ((pcbinfo->mtx = lck_rw_alloc_init(pcbinfo->mtx_grp, pcbinfo->mtx_attr)) == NULL)
 166                 return; /* pretty much dead if this fails... */
 167
 168 }
 169
 170 static struct   sockaddr_in ripsrc = { sizeof(ripsrc), AF_INET };
 171 /*
 172  * Setup generic address and protocol structures
 173  * for raw_input routine, then pass them along with
 174  * mbuf chain.
 175  */
 176 void
 177 rip_input(m, iphlen)
 178         struct mbuf *m;
 179         int iphlen;
 180 {
 181         register struct ip *ip = mtod(m, struct ip *);
 182         register struct inpcb *inp;
 183         struct inpcb *last = 0;
 184         struct mbuf *opts = 0;
 185         int skipit;
 186
 187         ripsrc.sin_addr = ip->ip_src;
 188         lck_rw_lock_shared(ripcbinfo.mtx);
 189         LIST_FOREACH(inp, &ripcb, inp_list) {
 190 #if INET6
 191                 if ((inp->inp_vflag & INP_IPV4) == 0)
 192                         continue;
 193 #endif
 194                 if (inp->inp_ip_p && (inp->inp_ip_p != ip->ip_p))
 195                         continue;
 196                 if (inp->inp_laddr.s_addr &&
 197                   inp->inp_laddr.s_addr != ip->ip_dst.s_addr)
 198                         continue;
 199                 if (inp->inp_faddr.s_addr &&
 200                   inp->inp_faddr.s_addr != ip->ip_src.s_addr)
 201                         continue;
 202                 if (last) {
 203                         struct mbuf *n = m_copy(m, 0, (int)M_COPYALL);
 204
 205 #if IPSEC
 206                         /* check AH/ESP integrity. */
 207                         skipit = 0;
 208                         if (ipsec_bypass == 0 && n) {
 209                                 lck_mtx_lock(sadb_mutex);
 210                                 if (ipsec4_in_reject_so(n, last->inp_socket)) {
 211                                         m_freem(n);
 212                                         ipsecstat.in_polvio++;
 213                                         /* do not inject data to pcb */
 214                                         skipit = 1;
 215                                 }
 216                                 lck_mtx_unlock(sadb_mutex);
 217                         }
 218 #endif /*IPSEC*/
 219                         if (n && skipit == 0) {
 220                                 int error = 0;
 221                                 if (last->inp_flags & INP_CONTROLOPTS ||
 222                                     last->inp_socket->so_options & SO_TIMESTAMP)
 223                                     ip_savecontrol(last, &opts, ip, n);
 224                                 if (last->inp_flags & INP_STRIPHDR) {
 225                                         n->m_len -= iphlen;
 226                                         n->m_pkthdr.len -= iphlen;
 227                                         n->m_data += iphlen;
 228                                 }
 229 // ###LOCK need to lock that socket?
 230                                 if (sbappendaddr(&last->inp_socket->so_rcv,
 231                                     (struct sockaddr *)&ripsrc, n,
 232                                     opts, &error) != 0) {
 233                                         sorwakeup(last->inp_socket);
 234                                 }
 235                                 else {
 236                                         if (error) {
 237                                                 /* should notify about lost packet */
 238                                                 kprintf("rip_input can't append to socket\n");
 239                                         }
 240                                 }
 241                                 opts = 0;
 242                         }
 243                 }
 244                 last = inp;
 245         }
 246         lck_rw_done(ripcbinfo.mtx);
 247 #if IPSEC
 248         /* check AH/ESP integrity. */
 249         skipit = 0;
 250         if (ipsec_bypass == 0 && last) {
 251                 lck_mtx_lock(sadb_mutex);
 252                 if (ipsec4_in_reject_so(m, last->inp_socket)) {
 253                         m_freem(m);
 254                         ipsecstat.in_polvio++;
 255                         ipstat.ips_delivered--;
 256                         /* do not inject data to pcb */
 257                         skipit = 1;
 258                 }
 259                 lck_mtx_unlock(sadb_mutex);
 260         }
 261 #endif /*IPSEC*/
 262         if (skipit == 0) {
 263                 if (last) {
 264                         if (last->inp_flags & INP_CONTROLOPTS ||
 265                                 last->inp_socket->so_options & SO_TIMESTAMP)
 266                                 ip_savecontrol(last, &opts, ip, m);
 267                         if (last->inp_flags & INP_STRIPHDR) {
 268                                 m->m_len -= iphlen;
 269                                 m->m_pkthdr.len -= iphlen;
 270                                 m->m_data += iphlen;
 271                         }
 272                         if (sbappendaddr(&last->inp_socket->so_rcv,
 273                                 (struct sockaddr *)&ripsrc, m, opts, NULL) != 0) {
 274                                 sorwakeup(last->inp_socket);
 275                         } else {
 276                                 kprintf("rip_input(2) can't append to socket\n");
 277                         }
 278                 } else {
 279                         m_freem(m);
 280                         ipstat.ips_noproto++;
 281                         ipstat.ips_delivered--;
 282                 }
 283         }
 284 }
 285
 286 /*
 287  * Generate IP header and pass packet to ip_output.
 288  * Tack on options user may have setup with control call.
 289  */
 290 int
 291 rip_output(m, so, dst)
 292         register struct mbuf *m;
 293         struct socket *so;
 294         u_long dst;
 295 {
 296         register struct ip *ip;
 297         register struct inpcb *inp = sotoinpcb(so);
 298         int flags = (so->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST;
 299
 300         /*
 301          * If the user handed us a complete IP packet, use it.
 302          * Otherwise, allocate an mbuf for a header and fill it in.
 303          */
 304         if ((inp->inp_flags & INP_HDRINCL) == 0) {
 305                 if (m->m_pkthdr.len + sizeof(struct ip) > IP_MAXPACKET) {
 306                         m_freem(m);
 307                         return(EMSGSIZE);
 308                 }
 309                 M_PREPEND(m, sizeof(struct ip), M_WAIT);
 310                 ip = mtod(m, struct ip *);
 311                 ip->ip_tos = inp->inp_ip_tos;
 312                 ip->ip_off = 0;
 313                 ip->ip_p = inp->inp_ip_p;
 314                 ip->ip_len = m->m_pkthdr.len;
 315                 ip->ip_src = inp->inp_laddr;
 316                 ip->ip_dst.s_addr = dst;
 317                 ip->ip_ttl = inp->inp_ip_ttl;
 318         } else {
 319                 if (m->m_pkthdr.len > IP_MAXPACKET) {
 320                         m_freem(m);
 321                         return(EMSGSIZE);
 322                 }
 323                 ip = mtod(m, struct ip *);
 324                 /* don't allow both user specified and setsockopt options,
 325                    and don't allow packet length sizes that will crash */
 326                 if (((IP_VHL_HL(ip->ip_vhl) != (sizeof (*ip) >> 2))
 327                      && inp->inp_options)
 328                     || (ip->ip_len > m->m_pkthdr.len)
 329                     || (ip->ip_len < (IP_VHL_HL(ip->ip_vhl) << 2))) {
 330                         m_freem(m);
 331                         return EINVAL;
 332                 }
 333                 if (ip->ip_id == 0)
 334 #if RANDOM_IP_ID
 335                         ip->ip_id = ip_randomid();
 336 #else
 337                         ip->ip_id = htons(ip_id++);
 338 #endif
 339                 /* XXX prevent ip_output from overwriting header fields */
 340                 flags |= IP_RAWOUTPUT;
 341                 ipstat.ips_rawout++;
 342         }
 343
 344 #if IPSEC
 345         if (ipsec_bypass == 0 && ipsec_setsocket(m, so) != 0) {
 346                 m_freem(m);
 347                 return ENOBUFS;
 348         }
 349 #endif /*IPSEC*/
 350
 351         if (inp->inp_route.ro_rt && inp->inp_route.ro_rt->generation_id != route_generation) {
 352                 rtfree(inp->inp_route.ro_rt);
 353                 inp->inp_route.ro_rt = (struct rtentry *)0;
 354         }
 355
 356         return (ip_output_list(m, 0, inp->inp_options, &inp->inp_route, flags,
 357                           inp->inp_moptions));
 358 }
 359
 360 extern int
 361 load_ipfw()
 362 {
 363         kern_return_t   err;
 364
 365         ipfw_init();
 366
 367 #if DUMMYNET
 368         if (!DUMMYNET_LOADED)
 369                 ip_dn_init();
 370 #endif /* DUMMYNET */
 371         err = 0;
 372
 373         return err == 0 && ip_fw_ctl_ptr == NULL ? -1 : err;
 374 }
 375
 376 /*
 377  * Raw IP socket option processing.
 378  */
 379 int
 380 rip_ctloutput(so, sopt)
 381         struct socket *so;
 382         struct sockopt *sopt;
 383 {
 384         struct  inpcb *inp = sotoinpcb(so);
 385         int     error, optval;
 386
 387         if (sopt->sopt_level != IPPROTO_IP)
 388                 return (EINVAL);
 389
 390         error = 0;
 391
 392         switch (sopt->sopt_dir) {
 393         case SOPT_GET:
 394                 switch (sopt->sopt_name) {
 395                 case IP_HDRINCL:
 396                         optval = inp->inp_flags & INP_HDRINCL;
 397                         error = sooptcopyout(sopt, &optval, sizeof optval);
 398                         break;
 399
 400         case IP_STRIPHDR:
 401             optval = inp->inp_flags & INP_STRIPHDR;
 402             error = sooptcopyout(sopt, &optval, sizeof optval);
 403             break;
 404
 405                 case IP_FW_ADD:
 406                 case IP_FW_GET:
 407                 case IP_OLD_FW_ADD:
 408                 case IP_OLD_FW_GET:
 409                         if (ip_fw_ctl_ptr == 0)
 410                                 error = load_ipfw();
 411                         if (ip_fw_ctl_ptr && error == 0)
 412                                 error = ip_fw_ctl_ptr(sopt);
 413                         else
 414                                 error = ENOPROTOOPT;
 415                         break;
 416
 417 #if DUMMYNET
 418                 case IP_DUMMYNET_GET:
 419                         if (DUMMYNET_LOADED)
 420                                 error = ip_dn_ctl_ptr(sopt);
 421                         else
 422                                 error = ENOPROTOOPT;
 423                         break ;
 424 #endif /* DUMMYNET */
 425
 426                 case MRT_INIT:
 427                 case MRT_DONE:
 428                 case MRT_ADD_VIF:
 429                 case MRT_DEL_VIF:
 430                 case MRT_ADD_MFC:
 431                 case MRT_DEL_MFC:
 432                 case MRT_VERSION:
 433                 case MRT_ASSERT:
 434                         error = ip_mrouter_get(so, sopt);
 435                         break;
 436
 437                 default:
 438                         error = ip_ctloutput(so, sopt);
 439                         break;
 440                 }
 441                 break;
 442
 443         case SOPT_SET:
 444                 switch (sopt->sopt_name) {
 445                 case IP_HDRINCL:
 446                         error = sooptcopyin(sopt, &optval, sizeof optval,
 447                                             sizeof optval);
 448                         if (error)
 449                                 break;
 450                         if (optval)
 451                                 inp->inp_flags |= INP_HDRINCL;
 452                         else
 453                                 inp->inp_flags &= ~INP_HDRINCL;
 454                         break;
 455
 456         case IP_STRIPHDR:
 457             error = sooptcopyin(sopt, &optval, sizeof optval,
 458                         sizeof optval);
 459             if (error)
 460                 break;
 461             if (optval)
 462                 inp->inp_flags |= INP_STRIPHDR;
 463             else
 464                 inp->inp_flags &= ~INP_STRIPHDR;
 465             break;
 466
 467
 468                 case IP_FW_ADD:
 469                 case IP_FW_DEL:
 470                 case IP_FW_FLUSH:
 471                 case IP_FW_ZERO:
 472                 case IP_FW_RESETLOG:
 473                 case IP_OLD_FW_ADD:
 474                 case IP_OLD_FW_DEL:
 475                 case IP_OLD_FW_FLUSH:
 476                 case IP_OLD_FW_ZERO:
 477                 case IP_OLD_FW_RESETLOG:
 478                         if (ip_fw_ctl_ptr == 0)
 479                                 error = load_ipfw();
 480                         if (ip_fw_ctl_ptr && error == 0)
 481                                 error = ip_fw_ctl_ptr(sopt);
 482                         else
 483                                 error = ENOPROTOOPT;
 484                         break;
 485
 486 #if DUMMYNET
 487                 case IP_DUMMYNET_CONFIGURE:
 488                 case IP_DUMMYNET_DEL:
 489                 case IP_DUMMYNET_FLUSH:
 490                         if (DUMMYNET_LOADED)
 491                                 error = ip_dn_ctl_ptr(sopt);
 492                         else
 493                                 error = ENOPROTOOPT ;
 494                         break ;
 495 #endif
 496
 497                 case IP_RSVP_ON:
 498                         error = ip_rsvp_init(so);
 499                         break;
 500
 501                 case IP_RSVP_OFF:
 502                         error = ip_rsvp_done();
 503                         break;
 504
 505                         /* XXX - should be combined */
 506                 case IP_RSVP_VIF_ON:
 507                         error = ip_rsvp_vif_init(so, sopt);
 508                         break;
 509
 510                 case IP_RSVP_VIF_OFF:
 511                         error = ip_rsvp_vif_done(so, sopt);
 512                         break;
 513
 514                 case MRT_INIT:
 515                 case MRT_DONE:
 516                 case MRT_ADD_VIF:
 517                 case MRT_DEL_VIF:
 518                 case MRT_ADD_MFC:
 519                 case MRT_DEL_MFC:
 520                 case MRT_VERSION:
 521                 case MRT_ASSERT:
 522                         error = ip_mrouter_set(so, sopt);
 523                         break;
 524
 525                 default:
 526                         error = ip_ctloutput(so, sopt);
 527                         break;
 528                 }
 529                 break;
 530         }
 531
 532         return (error);
 533 }
 534
 535 /*
 536  * This function exists solely to receive the PRC_IFDOWN messages which
 537  * are sent by if_down().  It looks for an ifaddr whose ifa_addr is sa,
 538  * and calls in_ifadown() to remove all routes corresponding to that address.
 539  * It also receives the PRC_IFUP messages from if_up() and reinstalls the
 540  * interface routes.
 541  */
 542 void
 543 rip_ctlinput(cmd, sa, vip)
 544         int cmd;
 545         struct sockaddr *sa;
 546         void *vip;
 547 {
 548         struct in_ifaddr *ia;
 549         struct ifnet *ifp;
 550         int err;
 551         int flags;
 552
 553         switch (cmd) {
 554         case PRC_IFDOWN:
 555                 lck_mtx_lock(rt_mtx);
 556                 for (ia = in_ifaddrhead.tqh_first; ia;
 557                      ia = ia->ia_link.tqe_next) {
 558                         if (ia->ia_ifa.ifa_addr == sa
 559                             && (ia->ia_flags & IFA_ROUTE)) {
 560                                 /*
 561                                  * in_ifscrub kills the interface route.
 562                                  */
 563                                 in_ifscrub(ia->ia_ifp, ia, 1);
 564                                 /*
 565                                  * in_ifadown gets rid of all the rest of
 566                                  * the routes.  This is not quite the right
 567                                  * thing to do, but at least if we are running
 568                                  * a routing process they will come back.
 569                                  */
 570                                 in_ifadown(&ia->ia_ifa, 1);
 571                                 break;
 572                         }
 573                 }
 574                 lck_mtx_unlock(rt_mtx);
 575                 break;
 576
 577         case PRC_IFUP:
 578                 lck_mtx_lock(rt_mtx);
 579                 for (ia = in_ifaddrhead.tqh_first; ia;
 580                      ia = ia->ia_link.tqe_next) {
 581                         if (ia->ia_ifa.ifa_addr == sa)
 582                                 break;
 583                 }
 584                 if (ia == 0 || (ia->ia_flags & IFA_ROUTE)) {
 585                         lck_mtx_unlock(rt_mtx);
 586                         return;
 587                 }
 588                 flags = RTF_UP;
 589                 ifp = ia->ia_ifa.ifa_ifp;
 590
 591                 if ((ifp->if_flags & IFF_LOOPBACK)
 592                     || (ifp->if_flags & IFF_POINTOPOINT))
 593                         flags |= RTF_HOST;
 594
 595                 err = rtinit_locked(&ia->ia_ifa, RTM_ADD, flags);
 596                 lck_mtx_unlock(rt_mtx);
 597                 if (err == 0)
 598                         ia->ia_flags |= IFA_ROUTE;
 599                 break;
 600         }
 601 }
 602
 603 u_long  rip_sendspace = RIPSNDQ;
 604 u_long  rip_recvspace = RIPRCVQ;
 605
 606 SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW,
 607     &rip_sendspace, 0, "Maximum outgoing raw IP datagram size");
 608 SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW,
 609     &rip_recvspace, 0, "Maximum incoming raw IP datagram size");
 610
 611 static int
 612 rip_attach(struct socket *so, int proto, struct proc *p)
 613 {
 614         struct inpcb *inp;
 615         int error, s;
 616
 617         inp = sotoinpcb(so);
 618         if (inp)
 619                 panic("rip_attach");
 620 #if __APPLE__
 621         if ((so->so_state & SS_PRIV) == 0)
 622                 return (EPERM);
 623 #else
 624         if (p && (error = suser(p)) != 0)
 625                 return error;
 626 #endif
 627
 628         error = soreserve(so, rip_sendspace, rip_recvspace);
 629         if (error)
 630                 return error;
 631         s = splnet();
 632         error = in_pcballoc(so, &ripcbinfo, p);
 633         splx(s);
 634         if (error)
 635                 return error;
 636         inp = (struct inpcb *)so->so_pcb;
 637         inp->inp_vflag |= INP_IPV4;
 638         inp->inp_ip_p = proto;
 639         inp->inp_ip_ttl = ip_defttl;
 640         return 0;
 641 }
 642
 643 __private_extern__ int
 644 rip_detach(struct socket *so)
 645 {
 646         struct inpcb *inp;
 647
 648         inp = sotoinpcb(so);
 649         if (inp == 0)
 650                 panic("rip_detach");
 651         if (so == ip_mrouter)
 652                 ip_mrouter_done();
 653         ip_rsvp_force_done(so);
 654         if (so == ip_rsvpd)
 655                 ip_rsvp_done();
 656         in_pcbdetach(inp);
 657         return 0;
 658 }
 659
 660 __private_extern__ int
 661 rip_abort(struct socket *so)
 662 {
 663         soisdisconnected(so);
 664         return rip_detach(so);
 665 }
 666
 667 __private_extern__ int
 668 rip_disconnect(struct socket *so)
 669 {
 670         if ((so->so_state & SS_ISCONNECTED) == 0)
 671                 return ENOTCONN;
 672         return rip_abort(so);
 673 }
 674
 675 __private_extern__ int
 676 rip_bind(struct socket *so, struct sockaddr *nam, struct proc *p)
 677 {
 678         struct inpcb *inp = sotoinpcb(so);
 679         struct sockaddr_in *addr = (struct sockaddr_in *)nam;
 680         struct ifaddr *ifa = NULL;
 681
 682         if (nam->sa_len != sizeof(*addr))
 683                 return EINVAL;
 684
 685         if (TAILQ_EMPTY(&ifnet_head) || ((addr->sin_family != AF_INET) &&
 686                                     (addr->sin_family != AF_IMPLINK)) ||
 687             (addr->sin_addr.s_addr &&
 688              (ifa = ifa_ifwithaddr((struct sockaddr *)addr)) == 0)) {
 689                 return EADDRNOTAVAIL;
 690         }
 691         else if (ifa) {
 692                 ifafree(ifa);
 693                 ifa = NULL;
 694         }
 695         inp->inp_laddr = addr->sin_addr;
 696         return 0;
 697 }
 698
 699 __private_extern__ int
 700 rip_connect(struct socket *so, struct sockaddr *nam, struct proc *p)
 701 {
 702         struct inpcb *inp = sotoinpcb(so);
 703         struct sockaddr_in *addr = (struct sockaddr_in *)nam;
 704
 705         if (nam->sa_len != sizeof(*addr))
 706                 return EINVAL;
 707         if (TAILQ_EMPTY(&ifnet_head))
 708                 return EADDRNOTAVAIL;
 709         if ((addr->sin_family != AF_INET) &&
 710             (addr->sin_family != AF_IMPLINK))
 711                 return EAFNOSUPPORT;
 712         inp->inp_faddr = addr->sin_addr;
 713         soisconnected(so);
 714         return 0;
 715 }
 716
 717 __private_extern__ int
 718 rip_shutdown(struct socket *so)
 719 {
 720         socantsendmore(so);
 721         return 0;
 722 }
 723
 724 __private_extern__ int
 725 rip_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam,
 726          struct mbuf *control, struct proc *p)
 727 {
 728         struct inpcb *inp = sotoinpcb(so);
 729         register u_long dst;
 730
 731         if (so->so_state & SS_ISCONNECTED) {
 732                 if (nam) {
 733                         m_freem(m);
 734                         return EISCONN;
 735                 }
 736                 dst = inp->inp_faddr.s_addr;
 737         } else {
 738                 if (nam == NULL) {
 739                         m_freem(m);
 740                         return ENOTCONN;
 741                 }
 742                 dst = ((struct sockaddr_in *)nam)->sin_addr.s_addr;
 743         }
 744         return rip_output(m, so, dst);
 745 }
 746
 747 int
 748 rip_unlock(struct socket *so, int refcount, int debug)
 749 {
 750         int lr_saved;
 751         struct inpcb *inp = sotoinpcb(so);
 752 #ifdef __ppc__
 753         if (debug == 0) {
 754                 __asm__ volatile("mflr %0" : "=r" (lr_saved));
 755         }
 756         else lr_saved = debug;
 757 #endif
 758         if (refcount) {
 759                 if (so->so_usecount <= 0)
 760                         panic("rip_unlock: bad refoucnt so=%x val=%x\n", so, so->so_usecount);
 761                 so->so_usecount--;
 762                 if (so->so_usecount == 0 && (inp->inp_wantcnt == WNT_STOPUSING)) {
 763                         lck_mtx_unlock(so->so_proto->pr_domain->dom_mtx);
 764                         lck_rw_lock_exclusive(ripcbinfo.mtx);
 765                         in_pcbdispose(inp);
 766                         lck_rw_done(ripcbinfo.mtx);
 767                         return(0);
 768                 }
 769         }
 770         lck_mtx_unlock(so->so_proto->pr_domain->dom_mtx);
 771         return(0);
 772 }
 773
 774 static int
 775 rip_pcblist SYSCTL_HANDLER_ARGS
 776 {
 777         int error, i, n, s;
 778         struct inpcb *inp, **inp_list;
 779         inp_gen_t gencnt;
 780         struct xinpgen xig;
 781
 782         /*
 783          * The process of preparing the TCB list is too time-consuming and
 784          * resource-intensive to repeat twice on every request.
 785          */
 786         lck_rw_lock_exclusive(ripcbinfo.mtx);
 787         if (req->oldptr == USER_ADDR_NULL) {
 788                 n = ripcbinfo.ipi_count;
 789                 req->oldidx = 2 * (sizeof xig)
 790                         + (n + n/8) * sizeof(struct xinpcb);
 791                 lck_rw_done(ripcbinfo.mtx);
 792                 return 0;
 793         }
 794
 795         if (req->newptr != USER_ADDR_NULL) {
 796                 lck_rw_done(ripcbinfo.mtx);
 797                 return EPERM;
 798         }
 799
 800         /*
 801          * OK, now we're committed to doing something.
 802          */
 803         gencnt = ripcbinfo.ipi_gencnt;
 804         n = ripcbinfo.ipi_count;
 805
 806         bzero(&xig, sizeof(xig));
 807         xig.xig_len = sizeof xig;
 808         xig.xig_count = n;
 809         xig.xig_gen = gencnt;
 810         xig.xig_sogen = so_gencnt;
 811         error = SYSCTL_OUT(req, &xig, sizeof xig);
 812         if (error) {
 813                 lck_rw_done(ripcbinfo.mtx);
 814                 return error;
 815         }
 816     /*
 817      * We are done if there is no pcb
 818      */
 819     if (n == 0) {
 820         lck_rw_done(ripcbinfo.mtx);
 821         return 0;
 822     }
 823
 824         inp_list = _MALLOC(n * sizeof *inp_list, M_TEMP, M_WAITOK);
 825         if (inp_list == 0) {
 826                 lck_rw_done(ripcbinfo.mtx);
 827                 return ENOMEM;
 828         }
 829
 830         for (inp = ripcbinfo.listhead->lh_first, i = 0; inp && i < n;
 831              inp = inp->inp_list.le_next) {
 832                 if (inp->inp_gencnt <= gencnt && inp->inp_state != INPCB_STATE_DEAD)
 833                         inp_list[i++] = inp;
 834         }
 835         n = i;
 836
 837         error = 0;
 838         for (i = 0; i < n; i++) {
 839                 inp = inp_list[i];
 840                 if (inp->inp_gencnt <= gencnt && inp->inp_state != INPCB_STATE_DEAD) {
 841                         struct xinpcb xi;
 842
 843                         bzero(&xi, sizeof(xi));
 844                         xi.xi_len = sizeof xi;
 845                         /* XXX should avoid extra copy */
 846                         inpcb_to_compat(inp, &xi.xi_inp);
 847                         if (inp->inp_socket)
 848                                 sotoxsocket(inp->inp_socket, &xi.xi_socket);
 849                         error = SYSCTL_OUT(req, &xi, sizeof xi);
 850                 }
 851         }
 852         if (!error) {
 853                 /*
 854                  * Give the user an updated idea of our state.
 855                  * If the generation differs from what we told
 856                  * her before, she knows that something happened
 857                  * while we were processing this request, and it
 858                  * might be necessary to retry.
 859                  */
 860                 bzero(&xig, sizeof(xig));
 861                 xig.xig_len = sizeof xig;
 862                 xig.xig_gen = ripcbinfo.ipi_gencnt;
 863                 xig.xig_sogen = so_gencnt;
 864                 xig.xig_count = ripcbinfo.ipi_count;
 865                 error = SYSCTL_OUT(req, &xig, sizeof xig);
 866         }
 867         FREE(inp_list, M_TEMP);
 868         lck_rw_done(ripcbinfo.mtx);
 869         return error;
 870 }
 871
 872 SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist, CTLFLAG_RD, 0, 0,
 873             rip_pcblist, "S,xinpcb", "List of active raw IP sockets");
 874
 875 struct pr_usrreqs rip_usrreqs = {
 876         rip_abort, pru_accept_notsupp, rip_attach, rip_bind, rip_connect,
 877         pru_connect2_notsupp, in_control, rip_detach, rip_disconnect,
 878         pru_listen_notsupp, in_setpeeraddr, pru_rcvd_notsupp,
 879         pru_rcvoob_notsupp, rip_send, pru_sense_null, rip_shutdown,
 880         in_setsockaddr, sosend, soreceive, pru_sopoll_notsupp
 881 };