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