bsd/netinet/raw_ip.c

   1 /*
   2  * Copyright (c) 2000 Apple Computer, Inc. All rights reserved.
   3  *
   4  * @APPLE_LICENSE_HEADER_START@
   5  *
   6  * The contents of this file constitute Original Code as defined in and
   7  * are subject to the Apple Public Source License Version 1.1 (the
   8  * "License").  You may not use this file except in compliance with the
   9  * License.  Please obtain a copy of the License at
  10  * http://www.apple.com/publicsource and read it before using this file.
  11  *
  12  * This Original Code and all software distributed under the License are
  13  * distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY KIND, EITHER
  14  * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
  15  * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
  16  * FITNESS FOR A PARTICULAR PURPOSE OR NON-INFRINGEMENT.  Please see the
  17  * License for the specific language governing rights and limitations
  18  * under the License.
  19  *
  20  * @APPLE_LICENSE_HEADER_END@
  21  */
  22 /*
  23  * Copyright (c) 1982, 1986, 1988, 1993
  24  *      The Regents of the University of California.  All rights reserved.
  25  *
  26  * Redistribution and use in source and binary forms, with or without
  27  * modification, are permitted provided that the following conditions
  28  * are met:
  29  * 1. Redistributions of source code must retain the above copyright
  30  *    notice, this list of conditions and the following disclaimer.
  31  * 2. Redistributions in binary form must reproduce the above copyright
  32  *    notice, this list of conditions and the following disclaimer in the
  33  *    documentation and/or other materials provided with the distribution.
  34  * 3. All advertising materials mentioning features or use of this software
  35  *    must display the following acknowledgement:
  36  *      This product includes software developed by the University of
  37  *      California, Berkeley and its contributors.
  38  * 4. Neither the name of the University nor the names of its contributors
  39  *    may be used to endorse or promote products derived from this software
  40  *    without specific prior written permission.
  41  *
  42  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  43  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  44  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  45  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  46  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  47  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  48  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  49  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  50  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  51  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  52  * SUCH DAMAGE.
  53  *
  54  *      @(#)raw_ip.c    8.7 (Berkeley) 5/15/95
  55  */
  56
  57 #include <sys/param.h>
  58 #include <sys/systm.h>
  59 #include <sys/kernel.h>
  60 #include <sys/malloc.h>
  61 #include <sys/mbuf.h>
  62 #include <sys/proc.h>
  63 #include <sys/domain.h>
  64 #include <sys/protosw.h>
  65 #include <sys/socket.h>
  66 #include <sys/socketvar.h>
  67 #include <sys/sysctl.h>
  68
  69 #if __FreeBSD__
  70 #include <vm/vm_zone.h>
  71 #endif
  72
  73 #include <net/if.h>
  74 #include <net/route.h>
  75
  76 #define _IP_VHL
  77 #include <netinet/in.h>
  78 #include <netinet/in_systm.h>
  79 #include <netinet/ip.h>
  80 #include <netinet/in_pcb.h>
  81 #include <netinet/in_var.h>
  82 #include <netinet/ip_var.h>
  83 #include <netinet/ip_mroute.h>
  84
  85 #include <netinet/ip_fw.h>
  86
  87 #if IPSEC
  88 #include <netinet6/ipsec.h>
  89 #endif /*IPSEC*/
  90
  91 #if DUMMYNET
  92 #include <netinet/ip_dummynet.h>
  93 #endif
  94
  95 #if IPSEC
  96 extern int ipsec_bypass;
  97 extern lck_mtx_t *sadb_mutex;
  98 #endif
  99
 100 extern u_long  route_generation;
 101 struct  inpcbhead ripcb;
 102 struct  inpcbinfo ripcbinfo;
 103
 104 /* control hooks for ipfw and dummynet */
 105 ip_fw_ctl_t *ip_fw_ctl_ptr;
 106 #if DUMMYNET
 107 ip_dn_ctl_t *ip_dn_ctl_ptr;
 108 #endif /* DUMMYNET */
 109
 110 /*
 111  * Nominal space allocated to a raw ip socket.
 112  */
 113 #define RIPSNDQ         8192
 114 #define RIPRCVQ         8192
 115
 116 /*
 117  * Raw interface to IP protocol.
 118  */
 119
 120 /*
 121  * Initialize raw connection block q.
 122  */
 123 void
 124 rip_init()
 125 {
 126         struct inpcbinfo *pcbinfo;
 127
 128         LIST_INIT(&ripcb);
 129         ripcbinfo.listhead = &ripcb;
 130         /*
 131          * XXX We don't use the hash list for raw IP, but it's easier
 132          * to allocate a one entry hash list than it is to check all
 133          * over the place for hashbase == NULL.
 134          */
 135         ripcbinfo.hashbase = hashinit(1, M_PCB, &ripcbinfo.hashmask);
 136         ripcbinfo.porthashbase = hashinit(1, M_PCB, &ripcbinfo.porthashmask);
 137
 138         ripcbinfo.ipi_zone = (void *) zinit(sizeof(struct inpcb),
 139                                             (4096 * sizeof(struct inpcb)),
 140                                             4096, "ripzone");
 141
 142         pcbinfo = &ripcbinfo;
 143         /*
 144          * allocate lock group attribute and group for udp pcb mutexes
 145          */
 146         pcbinfo->mtx_grp_attr = lck_grp_attr_alloc_init();
 147
 148         pcbinfo->mtx_grp = lck_grp_alloc_init("ripcb", pcbinfo->mtx_grp_attr);
 149
 150         /*
 151          * allocate the lock attribute for udp pcb mutexes
 152          */
 153         pcbinfo->mtx_attr = lck_attr_alloc_init();
 154
 155         if ((pcbinfo->mtx = lck_rw_alloc_init(pcbinfo->mtx_grp, pcbinfo->mtx_attr)) == NULL)
 156                 return; /* pretty much dead if this fails... */
 157
 158 }
 159
 160 static struct   sockaddr_in ripsrc = { sizeof(ripsrc), AF_INET };
 161 /*
 162  * Setup generic address and protocol structures
 163  * for raw_input routine, then pass them along with
 164  * mbuf chain.
 165  */
 166 void
 167 rip_input(m, iphlen)
 168         struct mbuf *m;
 169         int iphlen;
 170 {
 171         register struct ip *ip = mtod(m, struct ip *);
 172         register struct inpcb *inp;
 173         struct inpcb *last = 0;
 174         struct mbuf *opts = 0;
 175         int skipit;
 176
 177         ripsrc.sin_addr = ip->ip_src;
 178         lck_rw_lock_shared(ripcbinfo.mtx);
 179         LIST_FOREACH(inp, &ripcb, inp_list) {
 180 #if INET6
 181                 if ((inp->inp_vflag & INP_IPV4) == 0)
 182                         continue;
 183 #endif
 184                 if (inp->inp_ip_p && (inp->inp_ip_p != ip->ip_p))
 185                         continue;
 186                 if (inp->inp_laddr.s_addr &&
 187                   inp->inp_laddr.s_addr != ip->ip_dst.s_addr)
 188                         continue;
 189                 if (inp->inp_faddr.s_addr &&
 190                   inp->inp_faddr.s_addr != ip->ip_src.s_addr)
 191                         continue;
 192                 if (last) {
 193                         struct mbuf *n = m_copy(m, 0, (int)M_COPYALL);
 194
 195 #if IPSEC
 196                         /* check AH/ESP integrity. */
 197                         skipit = 0;
 198                         if (ipsec_bypass == 0 && n) {
 199                                 lck_mtx_lock(sadb_mutex);
 200                                 if (ipsec4_in_reject_so(n, last->inp_socket)) {
 201                                         m_freem(n);
 202                                         ipsecstat.in_polvio++;
 203                                         /* do not inject data to pcb */
 204                                         skipit = 1;
 205                                 }
 206                                 lck_mtx_unlock(sadb_mutex);
 207                         }
 208 #endif /*IPSEC*/
 209                         if (n && skipit == 0) {
 210                                 int error = 0;
 211                                 if (last->inp_flags & INP_CONTROLOPTS ||
 212                                     last->inp_socket->so_options & SO_TIMESTAMP)
 213                                     ip_savecontrol(last, &opts, ip, n);
 214                                 if (last->inp_flags & INP_STRIPHDR) {
 215                                         n->m_len -= iphlen;
 216                                         n->m_pkthdr.len -= iphlen;
 217                                         n->m_data += iphlen;
 218                                 }
 219 // ###LOCK need to lock that socket?
 220                                 if (sbappendaddr(&last->inp_socket->so_rcv,
 221                                     (struct sockaddr *)&ripsrc, n,
 222                                     opts, &error) != 0) {
 223                                         sorwakeup(last->inp_socket);
 224                                 }
 225                                 else {
 226                                         if (error) {
 227                                                 /* should notify about lost packet */
 228                                                 kprintf("rip_input can't append to socket\n");
 229                                         }
 230                                 }
 231                                 opts = 0;
 232                         }
 233                 }
 234                 last = inp;
 235         }
 236         lck_rw_done(ripcbinfo.mtx);
 237 #if IPSEC
 238         /* check AH/ESP integrity. */
 239         skipit = 0;
 240         if (ipsec_bypass == 0 && last) {
 241                 lck_mtx_lock(sadb_mutex);
 242                 if (ipsec4_in_reject_so(m, last->inp_socket)) {
 243                         m_freem(m);
 244                         ipsecstat.in_polvio++;
 245                         ipstat.ips_delivered--;
 246                         /* do not inject data to pcb */
 247                         skipit = 1;
 248                 }
 249                 lck_mtx_unlock(sadb_mutex);
 250         }
 251 #endif /*IPSEC*/
 252         if (skipit == 0) {
 253                 if (last) {
 254                         if (last->inp_flags & INP_CONTROLOPTS ||
 255                                 last->inp_socket->so_options & SO_TIMESTAMP)
 256                                 ip_savecontrol(last, &opts, ip, m);
 257                         if (last->inp_flags & INP_STRIPHDR) {
 258                                 m->m_len -= iphlen;
 259                                 m->m_pkthdr.len -= iphlen;
 260                                 m->m_data += iphlen;
 261                         }
 262                         if (sbappendaddr(&last->inp_socket->so_rcv,
 263                                 (struct sockaddr *)&ripsrc, m, opts, NULL) != 0) {
 264                                 sorwakeup(last->inp_socket);
 265                         } else {
 266                                 kprintf("rip_input(2) can't append to socket\n");
 267                         }
 268                 } else {
 269                         m_freem(m);
 270                         ipstat.ips_noproto++;
 271                         ipstat.ips_delivered--;
 272                 }
 273         }
 274 }
 275
 276 /*
 277  * Generate IP header and pass packet to ip_output.
 278  * Tack on options user may have setup with control call.
 279  */
 280 int
 281 rip_output(m, so, dst)
 282         register struct mbuf *m;
 283         struct socket *so;
 284         u_long dst;
 285 {
 286         register struct ip *ip;
 287         register struct inpcb *inp = sotoinpcb(so);
 288         int flags = (so->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST;
 289
 290         /*
 291          * If the user handed us a complete IP packet, use it.
 292          * Otherwise, allocate an mbuf for a header and fill it in.
 293          */
 294         if ((inp->inp_flags & INP_HDRINCL) == 0) {
 295                 if (m->m_pkthdr.len + sizeof(struct ip) > IP_MAXPACKET) {
 296                         m_freem(m);
 297                         return(EMSGSIZE);
 298                 }
 299                 M_PREPEND(m, sizeof(struct ip), M_WAIT);
 300                 ip = mtod(m, struct ip *);
 301                 ip->ip_tos = inp->inp_ip_tos;
 302                 ip->ip_off = 0;
 303                 ip->ip_p = inp->inp_ip_p;
 304                 ip->ip_len = m->m_pkthdr.len;
 305                 ip->ip_src = inp->inp_laddr;
 306                 ip->ip_dst.s_addr = dst;
 307                 ip->ip_ttl = inp->inp_ip_ttl;
 308         } else {
 309                 if (m->m_pkthdr.len > IP_MAXPACKET) {
 310                         m_freem(m);
 311                         return(EMSGSIZE);
 312                 }
 313                 ip = mtod(m, struct ip *);
 314                 /* don't allow both user specified and setsockopt options,
 315                    and don't allow packet length sizes that will crash */
 316                 if (((IP_VHL_HL(ip->ip_vhl) != (sizeof (*ip) >> 2))
 317                      && inp->inp_options)
 318                     || (ip->ip_len > m->m_pkthdr.len)
 319                     || (ip->ip_len < (IP_VHL_HL(ip->ip_vhl) << 2))) {
 320                         m_freem(m);
 321                         return EINVAL;
 322                 }
 323                 if (ip->ip_id == 0)
 324 #if RANDOM_IP_ID
 325                         ip->ip_id = ip_randomid();
 326 #else
 327                         ip->ip_id = htons(ip_id++);
 328 #endif
 329                 /* XXX prevent ip_output from overwriting header fields */
 330                 flags |= IP_RAWOUTPUT;
 331                 ipstat.ips_rawout++;
 332         }
 333
 334 #if IPSEC
 335         if (ipsec_bypass == 0 && ipsec_setsocket(m, so) != 0) {
 336                 m_freem(m);
 337                 return ENOBUFS;
 338         }
 339 #endif /*IPSEC*/
 340
 341         if (inp->inp_route.ro_rt && inp->inp_route.ro_rt->generation_id != route_generation) {
 342                 rtfree(inp->inp_route.ro_rt);
 343                 inp->inp_route.ro_rt = (struct rtentry *)0;
 344         }
 345
 346         return (ip_output_list(m, 0, inp->inp_options, &inp->inp_route, flags,
 347                           inp->inp_moptions));
 348 }
 349
 350 extern int
 351 load_ipfw()
 352 {
 353         kern_return_t   err;
 354
 355         ipfw_init();
 356
 357 #if DUMMYNET
 358         if (!DUMMYNET_LOADED)
 359                 ip_dn_init();
 360 #endif /* DUMMYNET */
 361         err = 0;
 362
 363         return err == 0 && ip_fw_ctl_ptr == NULL ? -1 : err;
 364 }
 365
 366 /*
 367  * Raw IP socket option processing.
 368  */
 369 int
 370 rip_ctloutput(so, sopt)
 371         struct socket *so;
 372         struct sockopt *sopt;
 373 {
 374         struct  inpcb *inp = sotoinpcb(so);
 375         int     error, optval;
 376
 377         if (sopt->sopt_level != IPPROTO_IP)
 378                 return (EINVAL);
 379
 380         error = 0;
 381
 382         switch (sopt->sopt_dir) {
 383         case SOPT_GET:
 384                 switch (sopt->sopt_name) {
 385                 case IP_HDRINCL:
 386                         optval = inp->inp_flags & INP_HDRINCL;
 387                         error = sooptcopyout(sopt, &optval, sizeof optval);
 388                         break;
 389
 390         case IP_STRIPHDR:
 391             optval = inp->inp_flags & INP_STRIPHDR;
 392             error = sooptcopyout(sopt, &optval, sizeof optval);
 393             break;
 394
 395                 case IP_FW_ADD:
 396                 case IP_FW_GET:
 397                 case IP_OLD_FW_ADD:
 398                 case IP_OLD_FW_GET:
 399                         if (ip_fw_ctl_ptr == 0)
 400                                 error = load_ipfw();
 401                         if (ip_fw_ctl_ptr && error == 0)
 402                                 error = ip_fw_ctl_ptr(sopt);
 403                         else
 404                                 error = ENOPROTOOPT;
 405                         break;
 406
 407 #if DUMMYNET
 408                 case IP_DUMMYNET_GET:
 409                         if (DUMMYNET_LOADED)
 410                                 error = ip_dn_ctl_ptr(sopt);
 411                         else
 412                                 error = ENOPROTOOPT;
 413                         break ;
 414 #endif /* DUMMYNET */
 415
 416                 case MRT_INIT:
 417                 case MRT_DONE:
 418                 case MRT_ADD_VIF:
 419                 case MRT_DEL_VIF:
 420                 case MRT_ADD_MFC:
 421                 case MRT_DEL_MFC:
 422                 case MRT_VERSION:
 423                 case MRT_ASSERT:
 424                         error = ip_mrouter_get(so, sopt);
 425                         break;
 426
 427                 default:
 428                         error = ip_ctloutput(so, sopt);
 429                         break;
 430                 }
 431                 break;
 432
 433         case SOPT_SET:
 434                 switch (sopt->sopt_name) {
 435                 case IP_HDRINCL:
 436                         error = sooptcopyin(sopt, &optval, sizeof optval,
 437                                             sizeof optval);
 438                         if (error)
 439                                 break;
 440                         if (optval)
 441                                 inp->inp_flags |= INP_HDRINCL;
 442                         else
 443                                 inp->inp_flags &= ~INP_HDRINCL;
 444                         break;
 445
 446         case IP_STRIPHDR:
 447             error = sooptcopyin(sopt, &optval, sizeof optval,
 448                         sizeof optval);
 449             if (error)
 450                 break;
 451             if (optval)
 452                 inp->inp_flags |= INP_STRIPHDR;
 453             else
 454                 inp->inp_flags &= ~INP_STRIPHDR;
 455             break;
 456
 457
 458                 case IP_FW_ADD:
 459                 case IP_FW_DEL:
 460                 case IP_FW_FLUSH:
 461                 case IP_FW_ZERO:
 462                 case IP_FW_RESETLOG:
 463                 case IP_OLD_FW_ADD:
 464                 case IP_OLD_FW_DEL:
 465                 case IP_OLD_FW_FLUSH:
 466                 case IP_OLD_FW_ZERO:
 467                 case IP_OLD_FW_RESETLOG:
 468                         if (ip_fw_ctl_ptr == 0)
 469                                 error = load_ipfw();
 470                         if (ip_fw_ctl_ptr && error == 0)
 471                                 error = ip_fw_ctl_ptr(sopt);
 472                         else
 473                                 error = ENOPROTOOPT;
 474                         break;
 475
 476 #if DUMMYNET
 477                 case IP_DUMMYNET_CONFIGURE:
 478                 case IP_DUMMYNET_DEL:
 479                 case IP_DUMMYNET_FLUSH:
 480                         if (DUMMYNET_LOADED)
 481                                 error = ip_dn_ctl_ptr(sopt);
 482                         else
 483                                 error = ENOPROTOOPT ;
 484                         break ;
 485 #endif
 486
 487                 case IP_RSVP_ON:
 488                         error = ip_rsvp_init(so);
 489                         break;
 490
 491                 case IP_RSVP_OFF:
 492                         error = ip_rsvp_done();
 493                         break;
 494
 495                         /* XXX - should be combined */
 496                 case IP_RSVP_VIF_ON:
 497                         error = ip_rsvp_vif_init(so, sopt);
 498                         break;
 499
 500                 case IP_RSVP_VIF_OFF:
 501                         error = ip_rsvp_vif_done(so, sopt);
 502                         break;
 503
 504                 case MRT_INIT:
 505                 case MRT_DONE:
 506                 case MRT_ADD_VIF:
 507                 case MRT_DEL_VIF:
 508                 case MRT_ADD_MFC:
 509                 case MRT_DEL_MFC:
 510                 case MRT_VERSION:
 511                 case MRT_ASSERT:
 512                         error = ip_mrouter_set(so, sopt);
 513                         break;
 514
 515                 default:
 516                         error = ip_ctloutput(so, sopt);
 517                         break;
 518                 }
 519                 break;
 520         }
 521
 522         return (error);
 523 }
 524
 525 /*
 526  * This function exists solely to receive the PRC_IFDOWN messages which
 527  * are sent by if_down().  It looks for an ifaddr whose ifa_addr is sa,
 528  * and calls in_ifadown() to remove all routes corresponding to that address.
 529  * It also receives the PRC_IFUP messages from if_up() and reinstalls the
 530  * interface routes.
 531  */
 532 void
 533 rip_ctlinput(cmd, sa, vip)
 534         int cmd;
 535         struct sockaddr *sa;
 536         void *vip;
 537 {
 538         struct in_ifaddr *ia;
 539         struct ifnet *ifp;
 540         int err;
 541         int flags;
 542
 543         switch (cmd) {
 544         case PRC_IFDOWN:
 545                 lck_mtx_lock(rt_mtx);
 546                 for (ia = in_ifaddrhead.tqh_first; ia;
 547                      ia = ia->ia_link.tqe_next) {
 548                         if (ia->ia_ifa.ifa_addr == sa
 549                             && (ia->ia_flags & IFA_ROUTE)) {
 550                                 /*
 551                                  * in_ifscrub kills the interface route.
 552                                  */
 553                                 in_ifscrub(ia->ia_ifp, ia, 1);
 554                                 /*
 555                                  * in_ifadown gets rid of all the rest of
 556                                  * the routes.  This is not quite the right
 557                                  * thing to do, but at least if we are running
 558                                  * a routing process they will come back.
 559                                  */
 560                                 in_ifadown(&ia->ia_ifa, 1);
 561                                 break;
 562                         }
 563                 }
 564                 lck_mtx_unlock(rt_mtx);
 565                 break;
 566
 567         case PRC_IFUP:
 568                 lck_mtx_lock(rt_mtx);
 569                 for (ia = in_ifaddrhead.tqh_first; ia;
 570                      ia = ia->ia_link.tqe_next) {
 571                         if (ia->ia_ifa.ifa_addr == sa)
 572                                 break;
 573                 }
 574                 if (ia == 0 || (ia->ia_flags & IFA_ROUTE)) {
 575                         lck_mtx_unlock(rt_mtx);
 576                         return;
 577                 }
 578                 flags = RTF_UP;
 579                 ifp = ia->ia_ifa.ifa_ifp;
 580
 581                 if ((ifp->if_flags & IFF_LOOPBACK)
 582                     || (ifp->if_flags & IFF_POINTOPOINT))
 583                         flags |= RTF_HOST;
 584
 585                 err = rtinit_locked(&ia->ia_ifa, RTM_ADD, flags);
 586                 lck_mtx_unlock(rt_mtx);
 587                 if (err == 0)
 588                         ia->ia_flags |= IFA_ROUTE;
 589                 break;
 590         }
 591 }
 592
 593 u_long  rip_sendspace = RIPSNDQ;
 594 u_long  rip_recvspace = RIPRCVQ;
 595
 596 SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW,
 597     &rip_sendspace, 0, "Maximum outgoing raw IP datagram size");
 598 SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW,
 599     &rip_recvspace, 0, "Maximum incoming raw IP datagram size");
 600
 601 static int
 602 rip_attach(struct socket *so, int proto, struct proc *p)
 603 {
 604         struct inpcb *inp;
 605         int error, s;
 606
 607         inp = sotoinpcb(so);
 608         if (inp)
 609                 panic("rip_attach");
 610 #if __APPLE__
 611         if ((so->so_state & SS_PRIV) == 0)
 612                 return (EPERM);
 613 #else
 614         if (p && (error = suser(p)) != 0)
 615                 return error;
 616 #endif
 617
 618         error = soreserve(so, rip_sendspace, rip_recvspace);
 619         if (error)
 620                 return error;
 621         s = splnet();
 622         error = in_pcballoc(so, &ripcbinfo, p);
 623         splx(s);
 624         if (error)
 625                 return error;
 626         inp = (struct inpcb *)so->so_pcb;
 627         inp->inp_vflag |= INP_IPV4;
 628         inp->inp_ip_p = proto;
 629         inp->inp_ip_ttl = ip_defttl;
 630         return 0;
 631 }
 632
 633 __private_extern__ int
 634 rip_detach(struct socket *so)
 635 {
 636         struct inpcb *inp;
 637
 638         inp = sotoinpcb(so);
 639         if (inp == 0)
 640                 panic("rip_detach");
 641         if (so == ip_mrouter)
 642                 ip_mrouter_done();
 643         ip_rsvp_force_done(so);
 644         if (so == ip_rsvpd)
 645                 ip_rsvp_done();
 646         in_pcbdetach(inp);
 647         return 0;
 648 }
 649
 650 __private_extern__ int
 651 rip_abort(struct socket *so)
 652 {
 653         soisdisconnected(so);
 654         return rip_detach(so);
 655 }
 656
 657 __private_extern__ int
 658 rip_disconnect(struct socket *so)
 659 {
 660         if ((so->so_state & SS_ISCONNECTED) == 0)
 661                 return ENOTCONN;
 662         return rip_abort(so);
 663 }
 664
 665 __private_extern__ int
 666 rip_bind(struct socket *so, struct sockaddr *nam, struct proc *p)
 667 {
 668         struct inpcb *inp = sotoinpcb(so);
 669         struct sockaddr_in *addr = (struct sockaddr_in *)nam;
 670         struct ifaddr *ifa = NULL;
 671
 672         if (nam->sa_len != sizeof(*addr))
 673                 return EINVAL;
 674
 675         if (TAILQ_EMPTY(&ifnet_head) || ((addr->sin_family != AF_INET) &&
 676                                     (addr->sin_family != AF_IMPLINK)) ||
 677             (addr->sin_addr.s_addr &&
 678              (ifa = ifa_ifwithaddr((struct sockaddr *)addr)) == 0)) {
 679                 return EADDRNOTAVAIL;
 680         }
 681         else if (ifa) {
 682                 ifafree(ifa);
 683                 ifa = NULL;
 684         }
 685         inp->inp_laddr = addr->sin_addr;
 686         return 0;
 687 }
 688
 689 __private_extern__ int
 690 rip_connect(struct socket *so, struct sockaddr *nam, struct proc *p)
 691 {
 692         struct inpcb *inp = sotoinpcb(so);
 693         struct sockaddr_in *addr = (struct sockaddr_in *)nam;
 694
 695         if (nam->sa_len != sizeof(*addr))
 696                 return EINVAL;
 697         if (TAILQ_EMPTY(&ifnet_head))
 698                 return EADDRNOTAVAIL;
 699         if ((addr->sin_family != AF_INET) &&
 700             (addr->sin_family != AF_IMPLINK))
 701                 return EAFNOSUPPORT;
 702         inp->inp_faddr = addr->sin_addr;
 703         soisconnected(so);
 704         return 0;
 705 }
 706
 707 __private_extern__ int
 708 rip_shutdown(struct socket *so)
 709 {
 710         socantsendmore(so);
 711         return 0;
 712 }
 713
 714 __private_extern__ int
 715 rip_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam,
 716          struct mbuf *control, struct proc *p)
 717 {
 718         struct inpcb *inp = sotoinpcb(so);
 719         register u_long dst;
 720
 721         if (so->so_state & SS_ISCONNECTED) {
 722                 if (nam) {
 723                         m_freem(m);
 724                         return EISCONN;
 725                 }
 726                 dst = inp->inp_faddr.s_addr;
 727         } else {
 728                 if (nam == NULL) {
 729                         m_freem(m);
 730                         return ENOTCONN;
 731                 }
 732                 dst = ((struct sockaddr_in *)nam)->sin_addr.s_addr;
 733         }
 734         return rip_output(m, so, dst);
 735 }
 736
 737 /* note: rip_unlock is called from different protos  instead of the generic socket_unlock,
 738  * it will handle the socket dealloc on last reference
 739  * */
 740 int
 741 rip_unlock(struct socket *so, int refcount, int debug)
 742 {
 743         int lr_saved;
 744         struct inpcb *inp = sotoinpcb(so);
 745
 746         if (debug == 0)
 747                 lr_saved = (unsigned int) __builtin_return_address(0);
 748         else lr_saved = debug;
 749
 750         if (refcount) {
 751                 if (so->so_usecount <= 0)
 752                         panic("rip_unlock: bad refoucnt so=%x val=%x\n", so, so->so_usecount);
 753                 so->so_usecount--;
 754                 if (so->so_usecount == 0 && (inp->inp_wantcnt == WNT_STOPUSING)) {
 755                         /* cleanup after last reference */
 756                         lck_mtx_unlock(so->so_proto->pr_domain->dom_mtx);
 757                         lck_rw_lock_exclusive(ripcbinfo.mtx);
 758                         in_pcbdispose(inp);
 759                         lck_rw_done(ripcbinfo.mtx);
 760                         return(0);
 761                 }
 762         }
 763         so->unlock_lr[so->next_unlock_lr] = (u_int *)lr_saved;
 764         so->next_unlock_lr = (so->next_unlock_lr+1) % SO_LCKDBG_MAX;
 765         lck_mtx_unlock(so->so_proto->pr_domain->dom_mtx);
 766         return(0);
 767 }
 768
 769 static int
 770 rip_pcblist SYSCTL_HANDLER_ARGS
 771 {
 772         int error, i, n, s;
 773         struct inpcb *inp, **inp_list;
 774         inp_gen_t gencnt;
 775         struct xinpgen xig;
 776
 777         /*
 778          * The process of preparing the TCB list is too time-consuming and
 779          * resource-intensive to repeat twice on every request.
 780          */
 781         lck_rw_lock_exclusive(ripcbinfo.mtx);
 782         if (req->oldptr == USER_ADDR_NULL) {
 783                 n = ripcbinfo.ipi_count;
 784                 req->oldidx = 2 * (sizeof xig)
 785                         + (n + n/8) * sizeof(struct xinpcb);
 786                 lck_rw_done(ripcbinfo.mtx);
 787                 return 0;
 788         }
 789
 790         if (req->newptr != USER_ADDR_NULL) {
 791                 lck_rw_done(ripcbinfo.mtx);
 792                 return EPERM;
 793         }
 794
 795         /*
 796          * OK, now we're committed to doing something.
 797          */
 798         gencnt = ripcbinfo.ipi_gencnt;
 799         n = ripcbinfo.ipi_count;
 800
 801         bzero(&xig, sizeof(xig));
 802         xig.xig_len = sizeof xig;
 803         xig.xig_count = n;
 804         xig.xig_gen = gencnt;
 805         xig.xig_sogen = so_gencnt;
 806         error = SYSCTL_OUT(req, &xig, sizeof xig);
 807         if (error) {
 808                 lck_rw_done(ripcbinfo.mtx);
 809                 return error;
 810         }
 811     /*
 812      * We are done if there is no pcb
 813      */
 814     if (n == 0) {
 815         lck_rw_done(ripcbinfo.mtx);
 816         return 0;
 817     }
 818
 819         inp_list = _MALLOC(n * sizeof *inp_list, M_TEMP, M_WAITOK);
 820         if (inp_list == 0) {
 821                 lck_rw_done(ripcbinfo.mtx);
 822                 return ENOMEM;
 823         }
 824
 825         for (inp = ripcbinfo.listhead->lh_first, i = 0; inp && i < n;
 826              inp = inp->inp_list.le_next) {
 827                 if (inp->inp_gencnt <= gencnt && inp->inp_state != INPCB_STATE_DEAD)
 828                         inp_list[i++] = inp;
 829         }
 830         n = i;
 831
 832         error = 0;
 833         for (i = 0; i < n; i++) {
 834                 inp = inp_list[i];
 835                 if (inp->inp_gencnt <= gencnt && inp->inp_state != INPCB_STATE_DEAD) {
 836                         struct xinpcb xi;
 837
 838                         bzero(&xi, sizeof(xi));
 839                         xi.xi_len = sizeof xi;
 840                         /* XXX should avoid extra copy */
 841                         inpcb_to_compat(inp, &xi.xi_inp);
 842                         if (inp->inp_socket)
 843                                 sotoxsocket(inp->inp_socket, &xi.xi_socket);
 844                         error = SYSCTL_OUT(req, &xi, sizeof xi);
 845                 }
 846         }
 847         if (!error) {
 848                 /*
 849                  * Give the user an updated idea of our state.
 850                  * If the generation differs from what we told
 851                  * her before, she knows that something happened
 852                  * while we were processing this request, and it
 853                  * might be necessary to retry.
 854                  */
 855                 bzero(&xig, sizeof(xig));
 856                 xig.xig_len = sizeof xig;
 857                 xig.xig_gen = ripcbinfo.ipi_gencnt;
 858                 xig.xig_sogen = so_gencnt;
 859                 xig.xig_count = ripcbinfo.ipi_count;
 860                 error = SYSCTL_OUT(req, &xig, sizeof xig);
 861         }
 862         FREE(inp_list, M_TEMP);
 863         lck_rw_done(ripcbinfo.mtx);
 864         return error;
 865 }
 866
 867 SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist, CTLFLAG_RD, 0, 0,
 868             rip_pcblist, "S,xinpcb", "List of active raw IP sockets");
 869
 870 struct pr_usrreqs rip_usrreqs = {
 871         rip_abort, pru_accept_notsupp, rip_attach, rip_bind, rip_connect,
 872         pru_connect2_notsupp, in_control, rip_detach, rip_disconnect,
 873         pru_listen_notsupp, in_setpeeraddr, pru_rcvd_notsupp,
 874         pru_rcvoob_notsupp, rip_send, pru_sense_null, rip_shutdown,
 875         in_setsockaddr, sosend, soreceive, pru_sopoll_notsupp
 876 };