bsd/netinet/raw_ip.c

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