raw_ip.c revision 118622
1/* 2 * Copyright (c) 1982, 1986, 1988, 1993 3 * The Regents of the University of California. All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 3. All advertising materials mentioning features or use of this software 14 * must display the following acknowledgement: 15 * This product includes software developed by the University of 16 * California, Berkeley and its contributors. 17 * 4. Neither the name of the University nor the names of its contributors 18 * may be used to endorse or promote products derived from this software 19 * without specific prior written permission. 20 * 21 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 24 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 31 * SUCH DAMAGE. 32 * 33 * @(#)raw_ip.c 8.7 (Berkeley) 5/15/95 34 * $FreeBSD: head/sys/netinet/raw_ip.c 118622 2003-08-07 18:16:59Z hsu $ 35 */ 36 37#include "opt_inet6.h" 38#include "opt_ipsec.h" 39#include "opt_mac.h" 40#include "opt_random_ip_id.h" 41 42#include <sys/param.h> 43#include <sys/kernel.h> 44#include <sys/lock.h> 45#include <sys/mac.h> 46#include <sys/malloc.h> 47#include <sys/mbuf.h> 48#include <sys/proc.h> 49#include <sys/protosw.h> 50#include <sys/signalvar.h> 51#include <sys/socket.h> 52#include <sys/socketvar.h> 53#include <sys/sx.h> 54#include <sys/sysctl.h> 55#include <sys/systm.h> 56 57#include <vm/uma.h> 58 59#include <net/if.h> 60#include <net/route.h> 61 62#include <netinet/in.h> 63#include <netinet/in_systm.h> 64#include <netinet/in_pcb.h> 65#include <netinet/in_var.h> 66#include <netinet/ip.h> 67#include <netinet/ip_var.h> 68#include <netinet/ip_mroute.h> 69 70#include <netinet/ip_fw.h> 71#include <netinet/ip_dummynet.h> 72 73#ifdef FAST_IPSEC 74#include <netipsec/ipsec.h> 75#endif /*FAST_IPSEC*/ 76 77#ifdef IPSEC 78#include <netinet6/ipsec.h> 79#endif /*IPSEC*/ 80 81struct inpcbhead ripcb; 82struct inpcbinfo ripcbinfo; 83 84/* control hooks for ipfw and dummynet */ 85ip_fw_ctl_t *ip_fw_ctl_ptr; 86ip_dn_ctl_t *ip_dn_ctl_ptr; 87 88/* 89 * hooks for multicast routing. They all default to NULL, 90 * so leave them not initialized and rely on BSS being set to 0. 91 */ 92 93/* The socket used to communicate with the multicast routing daemon. */ 94struct socket *ip_mrouter; 95 96/* The various mrouter and rsvp functions */ 97int (*ip_mrouter_set)(struct socket *, struct sockopt *); 98int (*ip_mrouter_get)(struct socket *, struct sockopt *); 99int (*ip_mrouter_done)(void); 100int (*ip_mforward)(struct ip *, struct ifnet *, struct mbuf *, 101 struct ip_moptions *); 102int (*mrt_ioctl)(int, caddr_t); 103int (*legal_vif_num)(int); 104u_long (*ip_mcast_src)(int); 105 106void (*rsvp_input_p)(struct mbuf *m, int off); 107int (*ip_rsvp_vif)(struct socket *, struct sockopt *); 108void (*ip_rsvp_force_done)(struct socket *); 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 */ 123void 124rip_init() 125{ 126 INP_INFO_LOCK_INIT(&ripcbinfo, "rip"); 127 LIST_INIT(&ripcb); 128 ripcbinfo.listhead = &ripcb; 129 /* 130 * XXX We don't use the hash list for raw IP, but it's easier 131 * to allocate a one entry hash list than it is to check all 132 * over the place for hashbase == NULL. 133 */ 134 ripcbinfo.hashbase = hashinit(1, M_PCB, &ripcbinfo.hashmask); 135 ripcbinfo.porthashbase = hashinit(1, M_PCB, &ripcbinfo.porthashmask); 136 ripcbinfo.ipi_zone = uma_zcreate("ripcb", sizeof(struct inpcb), 137 NULL, NULL, NULL, NULL, UMA_ALIGN_PTR, UMA_ZONE_NOFREE); 138 uma_zone_set_max(ripcbinfo.ipi_zone, maxsockets); 139} 140 141static struct sockaddr_in ripsrc = { sizeof(ripsrc), AF_INET }; 142/* 143 * Setup generic address and protocol structures 144 * for raw_input routine, then pass them along with 145 * mbuf chain. 146 */ 147void 148rip_input(m, off) 149 struct mbuf *m; 150 int off; 151{ 152 register struct ip *ip = mtod(m, struct ip *); 153 register struct inpcb *inp; 154 struct inpcb *last = 0; 155 struct mbuf *opts = 0; 156 int proto = ip->ip_p; 157 158 ripsrc.sin_addr = ip->ip_src; 159 LIST_FOREACH(inp, &ripcb, inp_list) { 160#ifdef INET6 161 if ((inp->inp_vflag & INP_IPV4) == 0) 162 continue; 163#endif 164 if (inp->inp_ip_p && inp->inp_ip_p != proto) 165 continue; 166 if (inp->inp_laddr.s_addr && 167 inp->inp_laddr.s_addr != ip->ip_dst.s_addr) 168 continue; 169 if (inp->inp_faddr.s_addr && 170 inp->inp_faddr.s_addr != ip->ip_src.s_addr) 171 continue; 172 if (last) { 173 struct mbuf *n = m_copy(m, 0, (int)M_COPYALL); 174 int policyfail = 0; 175 176 if (n != NULL) { 177#ifdef IPSEC 178 /* check AH/ESP integrity. */ 179 if (ipsec4_in_reject_so(n, last->inp_socket)) { 180 policyfail = 1; 181 ipsecstat.in_polvio++; 182 /* do not inject data to pcb */ 183 } 184#endif /*IPSEC*/ 185#ifdef FAST_IPSEC 186 /* check AH/ESP integrity. */ 187 if (ipsec4_in_reject(n, last)) { 188 policyfail = 1; 189 /* do not inject data to pcb */ 190 } 191#endif /*FAST_IPSEC*/ 192#ifdef MAC 193 if (policyfail == 0 && 194 mac_check_socket_deliver(last->inp_socket, 195 n) != 0) 196 policyfail = 1; 197#endif 198 } 199 if (policyfail) 200 m_freem(n); 201 else if (n) { 202 if (last->inp_flags & INP_CONTROLOPTS || 203 last->inp_socket->so_options & SO_TIMESTAMP) 204 ip_savecontrol(last, &opts, ip, n); 205 if (sbappendaddr(&last->inp_socket->so_rcv, 206 (struct sockaddr *)&ripsrc, n, 207 opts) == 0) { 208 /* should notify about lost packet */ 209 m_freem(n); 210 if (opts) 211 m_freem(opts); 212 } else 213 sorwakeup(last->inp_socket); 214 opts = 0; 215 } 216 } 217 last = inp; 218 } 219 if (last) { 220#ifdef IPSEC 221 /* check AH/ESP integrity. */ 222 if (ipsec4_in_reject_so(m, last->inp_socket)) { 223 m_freem(m); 224 ipsecstat.in_polvio++; 225 ipstat.ips_delivered--; 226 /* do not inject data to pcb */ 227 return; 228 } 229#endif /*IPSEC*/ 230#ifdef FAST_IPSEC 231 /* check AH/ESP integrity. */ 232 if (ipsec4_in_reject(m, last)) { 233 m_freem(m); 234 ipstat.ips_delivered--; 235 /* do not inject data to pcb */ 236 return; 237 } 238#endif /*FAST_IPSEC*/ 239#ifdef MAC 240 if (mac_check_socket_deliver(last->inp_socket, m) != 0) { 241 m_freem(m); 242 ipstat.ips_delivered--; 243 return; 244 } 245#endif 246 if (last->inp_flags & INP_CONTROLOPTS || 247 last->inp_socket->so_options & SO_TIMESTAMP) 248 ip_savecontrol(last, &opts, ip, m); 249 if (sbappendaddr(&last->inp_socket->so_rcv, 250 (struct sockaddr *)&ripsrc, m, opts) == 0) { 251 m_freem(m); 252 if (opts) 253 m_freem(opts); 254 } else 255 sorwakeup(last->inp_socket); 256 } else { 257 m_freem(m); 258 ipstat.ips_noproto++; 259 ipstat.ips_delivered--; 260 } 261} 262 263/* 264 * Generate IP header and pass packet to ip_output. 265 * Tack on options user may have setup with control call. 266 */ 267int 268rip_output(m, so, dst) 269 struct mbuf *m; 270 struct socket *so; 271 u_long dst; 272{ 273 register struct ip *ip; 274 register struct inpcb *inp = sotoinpcb(so); 275 int flags = (so->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST; 276 277#ifdef MAC 278 mac_create_mbuf_from_socket(so, m); 279#endif 280 281 /* 282 * If the user handed us a complete IP packet, use it. 283 * Otherwise, allocate an mbuf for a header and fill it in. 284 */ 285 if ((inp->inp_flags & INP_HDRINCL) == 0) { 286 if (m->m_pkthdr.len + sizeof(struct ip) > IP_MAXPACKET) { 287 m_freem(m); 288 return(EMSGSIZE); 289 } 290 M_PREPEND(m, sizeof(struct ip), M_TRYWAIT); 291 ip = mtod(m, struct ip *); 292 ip->ip_tos = inp->inp_ip_tos; 293 ip->ip_off = 0; 294 ip->ip_p = inp->inp_ip_p; 295 ip->ip_len = m->m_pkthdr.len; 296 ip->ip_src = inp->inp_laddr; 297 ip->ip_dst.s_addr = dst; 298 ip->ip_ttl = inp->inp_ip_ttl; 299 } else { 300 if (m->m_pkthdr.len > IP_MAXPACKET) { 301 m_freem(m); 302 return(EMSGSIZE); 303 } 304 ip = mtod(m, struct ip *); 305 /* don't allow both user specified and setsockopt options, 306 and don't allow packet length sizes that will crash */ 307 if (((ip->ip_hl != (sizeof (*ip) >> 2)) 308 && inp->inp_options) 309 || (ip->ip_len > m->m_pkthdr.len) 310 || (ip->ip_len < (ip->ip_hl << 2))) { 311 m_freem(m); 312 return EINVAL; 313 } 314 if (ip->ip_id == 0) 315#ifdef RANDOM_IP_ID 316 ip->ip_id = ip_randomid(); 317#else 318 ip->ip_id = htons(ip_id++); 319#endif 320 /* XXX prevent ip_output from overwriting header fields */ 321 flags |= IP_RAWOUTPUT; 322 ipstat.ips_rawout++; 323 } 324 325 return (ip_output(m, inp->inp_options, &inp->inp_route, flags, 326 inp->inp_moptions, inp)); 327} 328 329/* 330 * Raw IP socket option processing. 331 * 332 * Note that access to all of the IP administrative functions here is 333 * implicitly protected by suser() as gaining access to a raw socket 334 * requires either that the thread pass a suser() check, or that it be 335 * passed a raw socket by another thread that has passed a suser() check. 336 * If FreeBSD moves to a more fine-grained access control mechanism, 337 * additional checks will need to be placed here if the raw IP attachment 338 * check is not equivilent the the check required for these 339 * administrative operations; in some cases, these checks are already 340 * present. 341 */ 342int 343rip_ctloutput(so, sopt) 344 struct socket *so; 345 struct sockopt *sopt; 346{ 347 struct inpcb *inp = sotoinpcb(so); 348 int error, optval; 349 350 if (sopt->sopt_level != IPPROTO_IP) 351 return (EINVAL); 352 353 error = 0; 354 355 switch (sopt->sopt_dir) { 356 case SOPT_GET: 357 switch (sopt->sopt_name) { 358 case IP_HDRINCL: 359 optval = inp->inp_flags & INP_HDRINCL; 360 error = sooptcopyout(sopt, &optval, sizeof optval); 361 break; 362 363 case IP_FW_ADD: /* ADD actually returns the body... */ 364 case IP_FW_GET: 365 if (IPFW_LOADED) 366 error = ip_fw_ctl_ptr(sopt); 367 else 368 error = ENOPROTOOPT; 369 break; 370 371 case IP_DUMMYNET_GET: 372 if (DUMMYNET_LOADED) 373 error = ip_dn_ctl_ptr(sopt); 374 else 375 error = ENOPROTOOPT; 376 break ; 377 378 case MRT_INIT: 379 case MRT_DONE: 380 case MRT_ADD_VIF: 381 case MRT_DEL_VIF: 382 case MRT_ADD_MFC: 383 case MRT_DEL_MFC: 384 case MRT_VERSION: 385 case MRT_ASSERT: 386 case MRT_API_SUPPORT: 387 case MRT_API_CONFIG: 388 case MRT_ADD_BW_UPCALL: 389 case MRT_DEL_BW_UPCALL: 390 error = ip_mrouter_get ? ip_mrouter_get(so, sopt) : 391 EOPNOTSUPP; 392 break; 393 394 default: 395 error = ip_ctloutput(so, sopt); 396 break; 397 } 398 break; 399 400 case SOPT_SET: 401 switch (sopt->sopt_name) { 402 case IP_HDRINCL: 403 error = sooptcopyin(sopt, &optval, sizeof optval, 404 sizeof optval); 405 if (error) 406 break; 407 if (optval) 408 inp->inp_flags |= INP_HDRINCL; 409 else 410 inp->inp_flags &= ~INP_HDRINCL; 411 break; 412 413 case IP_FW_ADD: 414 case IP_FW_DEL: 415 case IP_FW_FLUSH: 416 case IP_FW_ZERO: 417 case IP_FW_RESETLOG: 418 if (IPFW_LOADED) 419 error = ip_fw_ctl_ptr(sopt); 420 else 421 error = ENOPROTOOPT; 422 break; 423 424 case IP_DUMMYNET_CONFIGURE: 425 case IP_DUMMYNET_DEL: 426 case IP_DUMMYNET_FLUSH: 427 if (DUMMYNET_LOADED) 428 error = ip_dn_ctl_ptr(sopt); 429 else 430 error = ENOPROTOOPT ; 431 break ; 432 433 case IP_RSVP_ON: 434 error = ip_rsvp_init(so); 435 break; 436 437 case IP_RSVP_OFF: 438 error = ip_rsvp_done(); 439 break; 440 441 case IP_RSVP_VIF_ON: 442 case IP_RSVP_VIF_OFF: 443 error = ip_rsvp_vif ? 444 ip_rsvp_vif(so, sopt) : EINVAL; 445 break; 446 447 case MRT_INIT: 448 case MRT_DONE: 449 case MRT_ADD_VIF: 450 case MRT_DEL_VIF: 451 case MRT_ADD_MFC: 452 case MRT_DEL_MFC: 453 case MRT_VERSION: 454 case MRT_ASSERT: 455 case MRT_API_SUPPORT: 456 case MRT_API_CONFIG: 457 case MRT_ADD_BW_UPCALL: 458 case MRT_DEL_BW_UPCALL: 459 error = ip_mrouter_set ? ip_mrouter_set(so, sopt) : 460 EOPNOTSUPP; 461 break; 462 463 default: 464 error = ip_ctloutput(so, sopt); 465 break; 466 } 467 break; 468 } 469 470 return (error); 471} 472 473/* 474 * This function exists solely to receive the PRC_IFDOWN messages which 475 * are sent by if_down(). It looks for an ifaddr whose ifa_addr is sa, 476 * and calls in_ifadown() to remove all routes corresponding to that address. 477 * It also receives the PRC_IFUP messages from if_up() and reinstalls the 478 * interface routes. 479 */ 480void 481rip_ctlinput(cmd, sa, vip) 482 int cmd; 483 struct sockaddr *sa; 484 void *vip; 485{ 486 struct in_ifaddr *ia; 487 struct ifnet *ifp; 488 int err; 489 int flags; 490 491 switch (cmd) { 492 case PRC_IFDOWN: 493 TAILQ_FOREACH(ia, &in_ifaddrhead, ia_link) { 494 if (ia->ia_ifa.ifa_addr == sa 495 && (ia->ia_flags & IFA_ROUTE)) { 496 /* 497 * in_ifscrub kills the interface route. 498 */ 499 in_ifscrub(ia->ia_ifp, ia); 500 /* 501 * in_ifadown gets rid of all the rest of 502 * the routes. This is not quite the right 503 * thing to do, but at least if we are running 504 * a routing process they will come back. 505 */ 506 in_ifadown(&ia->ia_ifa, 0); 507 break; 508 } 509 } 510 break; 511 512 case PRC_IFUP: 513 TAILQ_FOREACH(ia, &in_ifaddrhead, ia_link) { 514 if (ia->ia_ifa.ifa_addr == sa) 515 break; 516 } 517 if (ia == 0 || (ia->ia_flags & IFA_ROUTE)) 518 return; 519 flags = RTF_UP; 520 ifp = ia->ia_ifa.ifa_ifp; 521 522 if ((ifp->if_flags & IFF_LOOPBACK) 523 || (ifp->if_flags & IFF_POINTOPOINT)) 524 flags |= RTF_HOST; 525 526 err = rtinit(&ia->ia_ifa, RTM_ADD, flags); 527 if (err == 0) 528 ia->ia_flags |= IFA_ROUTE; 529 break; 530 } 531} 532 533u_long rip_sendspace = RIPSNDQ; 534u_long rip_recvspace = RIPRCVQ; 535int rip_olddiverterror = 1; 536 537SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW, 538 &rip_sendspace, 0, "Maximum outgoing raw IP datagram size"); 539SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW, 540 &rip_recvspace, 0, "Maximum incoming raw IP datagram size"); 541SYSCTL_INT(_net_inet_raw, OID_AUTO, olddiverterror, CTLFLAG_RW, 542 &rip_olddiverterror, 0, "Return an error when creating an 'old' DIVERT socket"); 543 544static int 545rip_attach(struct socket *so, int proto, struct thread *td) 546{ 547 struct inpcb *inp; 548 int error, s; 549 550 inp = sotoinpcb(so); 551 if (inp) 552 panic("rip_attach"); 553 if (td && (error = suser(td)) != 0) 554 return error; 555 556 if (proto >= IPPROTO_MAX || proto < 0) 557 return EPROTONOSUPPORT; 558 559 /* To be removed before 5.2 */ 560 if (rip_olddiverterror && proto == IPPROTO_OLD_DIVERT) { 561 printf("Old IPDIVERT program needs to be recompiled, or new IP proto 254 user needs sysctl net.inet.raw.olddiverterror=0\n"); 562 return EPROTONOSUPPORT; 563 } 564 565 error = soreserve(so, rip_sendspace, rip_recvspace); 566 if (error) 567 return error; 568 s = splnet(); 569 error = in_pcballoc(so, &ripcbinfo, td); 570 splx(s); 571 if (error) 572 return error; 573 inp = (struct inpcb *)so->so_pcb; 574 inp->inp_vflag |= INP_IPV4; 575 inp->inp_ip_p = proto; 576 inp->inp_ip_ttl = ip_defttl; 577 return 0; 578} 579 580static int 581rip_detach(struct socket *so) 582{ 583 struct inpcb *inp; 584 585 inp = sotoinpcb(so); 586 if (inp == 0) 587 panic("rip_detach"); 588 if (so == ip_mrouter && ip_mrouter_done) 589 ip_mrouter_done(); 590 if (ip_rsvp_force_done) 591 ip_rsvp_force_done(so); 592 if (so == ip_rsvpd) 593 ip_rsvp_done(); 594 in_pcbdetach(inp); 595 return 0; 596} 597 598static int 599rip_abort(struct socket *so) 600{ 601 soisdisconnected(so); 602 if (so->so_state & SS_NOFDREF) 603 return rip_detach(so); 604 return 0; 605} 606 607static int 608rip_disconnect(struct socket *so) 609{ 610 if ((so->so_state & SS_ISCONNECTED) == 0) 611 return ENOTCONN; 612 return rip_abort(so); 613} 614 615static int 616rip_bind(struct socket *so, struct sockaddr *nam, struct thread *td) 617{ 618 struct inpcb *inp = sotoinpcb(so); 619 struct sockaddr_in *addr = (struct sockaddr_in *)nam; 620 621 if (nam->sa_len != sizeof(*addr)) 622 return EINVAL; 623 624 if (TAILQ_EMPTY(&ifnet) || ((addr->sin_family != AF_INET) && 625 (addr->sin_family != AF_IMPLINK)) || 626 (addr->sin_addr.s_addr && 627 ifa_ifwithaddr((struct sockaddr *)addr) == 0)) 628 return EADDRNOTAVAIL; 629 inp->inp_laddr = addr->sin_addr; 630 return 0; 631} 632 633static int 634rip_connect(struct socket *so, struct sockaddr *nam, struct thread *td) 635{ 636 struct inpcb *inp = sotoinpcb(so); 637 struct sockaddr_in *addr = (struct sockaddr_in *)nam; 638 639 if (nam->sa_len != sizeof(*addr)) 640 return EINVAL; 641 if (TAILQ_EMPTY(&ifnet)) 642 return EADDRNOTAVAIL; 643 if ((addr->sin_family != AF_INET) && 644 (addr->sin_family != AF_IMPLINK)) 645 return EAFNOSUPPORT; 646 inp->inp_faddr = addr->sin_addr; 647 soisconnected(so); 648 return 0; 649} 650 651static int 652rip_shutdown(struct socket *so) 653{ 654 socantsendmore(so); 655 return 0; 656} 657 658static int 659rip_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam, 660 struct mbuf *control, struct thread *td) 661{ 662 struct inpcb *inp = sotoinpcb(so); 663 register u_long dst; 664 665 if (so->so_state & SS_ISCONNECTED) { 666 if (nam) { 667 m_freem(m); 668 return EISCONN; 669 } 670 dst = inp->inp_faddr.s_addr; 671 } else { 672 if (nam == NULL) { 673 m_freem(m); 674 return ENOTCONN; 675 } 676 dst = ((struct sockaddr_in *)nam)->sin_addr.s_addr; 677 } 678 return rip_output(m, so, dst); 679} 680 681static int 682rip_pcblist(SYSCTL_HANDLER_ARGS) 683{ 684 int error, i, n, s; 685 struct inpcb *inp, **inp_list; 686 inp_gen_t gencnt; 687 struct xinpgen xig; 688 689 /* 690 * The process of preparing the TCB list is too time-consuming and 691 * resource-intensive to repeat twice on every request. 692 */ 693 if (req->oldptr == 0) { 694 n = ripcbinfo.ipi_count; 695 req->oldidx = 2 * (sizeof xig) 696 + (n + n/8) * sizeof(struct xinpcb); 697 return 0; 698 } 699 700 if (req->newptr != 0) 701 return EPERM; 702 703 /* 704 * OK, now we're committed to doing something. 705 */ 706 s = splnet(); 707 gencnt = ripcbinfo.ipi_gencnt; 708 n = ripcbinfo.ipi_count; 709 splx(s); 710 711 xig.xig_len = sizeof xig; 712 xig.xig_count = n; 713 xig.xig_gen = gencnt; 714 xig.xig_sogen = so_gencnt; 715 error = SYSCTL_OUT(req, &xig, sizeof xig); 716 if (error) 717 return error; 718 719 inp_list = malloc(n * sizeof *inp_list, M_TEMP, M_WAITOK); 720 if (inp_list == 0) 721 return ENOMEM; 722 723 s = splnet(); 724 for (inp = LIST_FIRST(ripcbinfo.listhead), i = 0; inp && i < n; 725 inp = LIST_NEXT(inp, inp_list)) { 726 if (inp->inp_gencnt <= gencnt) { 727 if (cr_canseesocket(req->td->td_ucred, 728 inp->inp_socket)) 729 continue; 730 inp_list[i++] = inp; 731 } 732 } 733 splx(s); 734 n = i; 735 736 error = 0; 737 for (i = 0; i < n; i++) { 738 inp = inp_list[i]; 739 if (inp->inp_gencnt <= gencnt) { 740 struct xinpcb xi; 741 xi.xi_len = sizeof xi; 742 /* XXX should avoid extra copy */ 743 bcopy(inp, &xi.xi_inp, sizeof *inp); 744 if (inp->inp_socket) 745 sotoxsocket(inp->inp_socket, &xi.xi_socket); 746 error = SYSCTL_OUT(req, &xi, sizeof xi); 747 } 748 } 749 if (!error) { 750 /* 751 * Give the user an updated idea of our state. 752 * If the generation differs from what we told 753 * her before, she knows that something happened 754 * while we were processing this request, and it 755 * might be necessary to retry. 756 */ 757 s = splnet(); 758 xig.xig_gen = ripcbinfo.ipi_gencnt; 759 xig.xig_sogen = so_gencnt; 760 xig.xig_count = ripcbinfo.ipi_count; 761 splx(s); 762 error = SYSCTL_OUT(req, &xig, sizeof xig); 763 } 764 free(inp_list, M_TEMP); 765 return error; 766} 767 768/* 769 * This is the wrapper function for in_setsockaddr. We just pass down 770 * the pcbinfo for in_setpeeraddr to lock. 771 */ 772static int 773rip_sockaddr(struct socket *so, struct sockaddr **nam) 774{ 775 return (in_setsockaddr(so, nam, &ripcbinfo)); 776} 777 778/* 779 * This is the wrapper function for in_setpeeraddr. We just pass down 780 * the pcbinfo for in_setpeeraddr to lock. 781 */ 782static int 783rip_peeraddr(struct socket *so, struct sockaddr **nam) 784{ 785 return (in_setpeeraddr(so, nam, &ripcbinfo)); 786} 787 788 789SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist, CTLFLAG_RD, 0, 0, 790 rip_pcblist, "S,xinpcb", "List of active raw IP sockets"); 791 792struct pr_usrreqs rip_usrreqs = { 793 rip_abort, pru_accept_notsupp, rip_attach, rip_bind, rip_connect, 794 pru_connect2_notsupp, in_control, rip_detach, rip_disconnect, 795 pru_listen_notsupp, rip_peeraddr, pru_rcvd_notsupp, 796 pru_rcvoob_notsupp, rip_send, pru_sense_null, rip_shutdown, 797 rip_sockaddr, sosend, soreceive, sopoll 798}; 799