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