Annotation of XNU/bsd/netinet/ip_divert.c, revision 1.1.1.1

1.1       root        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:  */
                     55: 
                     56: #if ISFB31
                     57: #include "opt_inet.h"
                     58: #include "opt_ipfw.h"
                     59: #include "opt_ipdivert.h"
                     60: #endif
                     61: 
                     62: #ifndef INET
                     63: #error "IPDIVERT requires INET."
                     64: #endif
                     65: 
                     66: #include <sys/param.h>
                     67: #include <sys/malloc.h>
                     68: #include <sys/mbuf.h>
                     69: #include <sys/socket.h>
                     70: #include <sys/protosw.h>
                     71: #include <sys/socketvar.h>
                     72: #include <sys/systm.h>
                     73: #include <sys/proc.h>
                     74: 
                     75: #if ISFB31
                     76: #include <vm/vm_zone.h>
                     77: #endif
                     78: 
                     79: #include <net/if.h>
                     80: #include <net/route.h>
                     81: 
                     82: #include <netinet/in.h>
                     83: #include <netinet/in_systm.h>
                     84: #include <netinet/ip.h>
                     85: #include <netinet/in_pcb.h>
                     86: #include <netinet/in_var.h>
                     87: #include <netinet/ip_var.h>
                     88: 
                     89: /*
                     90:  * Divert sockets
                     91:  */
                     92: 
                     93: /*
                     94:  * Allocate enough space to hold a full IP packet
                     95:  */
                     96: #define        DIVSNDQ         (65536 + 100)
                     97: #define        DIVRCVQ         (65536 + 100)
                     98: 
                     99: /* Global variables */
                    100: 
                    101: /*
                    102:  * ip_input() and ip_output() set this secret value before calling us to
                    103:  * let us know which divert port to divert a packet to; this is done so
                    104:  * we can use the existing prototype for struct protosw's pr_input().
                    105:  * This is stored in host order.
                    106:  */
                    107: u_short ip_divert_port;
                    108: 
                    109: /*
                    110:  * A 16 bit cookie is passed to the user process.
                    111:  * The user process can send it back to help the caller know something
                    112:  * about where the packet came from.
                    113:  *
                    114:  * If IPFW is the caller then the cookie is the rule that sent
                    115:  * us here. On reinjection is is the rule after which processing
                    116:  * should continue. Leaving it the same will make processing start
                    117:  * at the rule number after that which sent it here. Setting it to
                    118:  * 0 will restart processing at the beginning. 
                    119:  */
                    120: u_int16_t ip_divert_cookie;
                    121: 
                    122: /* Internal variables */
                    123: 
                    124: static struct inpcbhead divcb;
                    125: static struct inpcbinfo divcbinfo;
                    126: 
                    127: static u_long  div_sendspace = DIVSNDQ;        /* XXX sysctl ? */
                    128: static u_long  div_recvspace = DIVRCVQ;        /* XXX sysctl ? */
                    129: 
                    130: /* Optimization: have this preinitialized */
                    131: static struct sockaddr_in divsrc = { sizeof(divsrc), AF_INET };
                    132: 
                    133: /* Internal functions */
                    134: 
                    135: static int div_output(struct socket *so,
                    136:                struct mbuf *m, struct sockaddr *addr, struct mbuf *control);
                    137: 
                    138: /*
                    139:  * Initialize divert connection block queue.
                    140:  */
                    141: void
                    142: div_init(void)
                    143: {
                    144:        LIST_INIT(&divcb);
                    145:        divcbinfo.listhead = &divcb;
                    146:        /*
                    147:         * XXX We don't use the hash list for divert IP, but it's easier
                    148:         * to allocate a one entry hash list than it is to check all
                    149:         * over the place for hashbase == NULL.
                    150:         */
                    151:        divcbinfo.hashbase = hashinit(1, M_PCB, &divcbinfo.hashmask);
                    152:        divcbinfo.porthashbase = hashinit(1, M_PCB, &divcbinfo.porthashmask);
                    153:        divcbinfo.ipi_zone = (void *) zinit(sizeof(struct inpcb),(maxsockets * sizeof(struct inpcb)),
                    154:                                   4096, "divzone");
                    155: 
                    156: /*
                    157:  * ### LD 08/03: init IP forwarding at this point [ipfw is not a module yet]
                    158:  */
                    159:        ip_fw_init();
                    160: }
                    161: 
                    162: /*
                    163:  * Setup generic address and protocol structures
                    164:  * for div_input routine, then pass them along with
                    165:  * mbuf chain. ip->ip_len is assumed to have had
                    166:  * the header length (hlen) subtracted out already.
                    167:  * We tell whether the packet was incoming or outgoing
                    168:  * by seeing if hlen == 0, which is a hack.
                    169:  */
                    170: void
                    171: div_input(struct mbuf *m, int hlen)
                    172: {
                    173:        struct ip *ip;
                    174:        struct inpcb *inp;
                    175:        struct socket *sa;
                    176: 
                    177:        /* Sanity check */
                    178:        if (ip_divert_port == 0)
                    179:                panic("div_input: port is 0");
                    180: 
                    181:        /* Assure header */
                    182:        if (m->m_len < sizeof(struct ip) &&
                    183:            (m = m_pullup(m, sizeof(struct ip))) == 0) {
                    184:                return;
                    185:        }
                    186:        ip = mtod(m, struct ip *);
                    187: 
                    188:        /* Record divert cookie */
                    189:        divsrc.sin_port = ip_divert_cookie;
                    190:        ip_divert_cookie = 0;
                    191: 
                    192:        /* Restore packet header fields */
                    193:        ip->ip_len += hlen;
                    194:        HTONS(ip->ip_len);
                    195:        HTONS(ip->ip_off);
                    196: 
                    197:        /*
                    198:         * Record receive interface address, if any
                    199:         * But only for incoming packets.
                    200:         */
                    201:        divsrc.sin_addr.s_addr = 0;
                    202:        if (hlen) {
                    203:                struct ifaddr *ifa;
                    204: 
                    205: #if DIAGNOSTIC
                    206:                /* Sanity check */
                    207:                if (!(m->m_flags & M_PKTHDR))
                    208:                        panic("div_input: no pkt hdr");
                    209: #endif
                    210: 
                    211:                /* More fields affected by ip_input() */
                    212:                HTONS(ip->ip_id);
                    213: 
                    214:                /* Find IP address for receive interface */
                    215:                for (ifa = m->m_pkthdr.rcvif->if_addrhead.tqh_first;
                    216:                    ifa != NULL; ifa = ifa->ifa_link.tqe_next) {
                    217:                        if (ifa->ifa_addr == NULL)
                    218:                                continue;
                    219:                        if (ifa->ifa_addr->sa_family != AF_INET)
                    220:                                continue;
                    221:                        divsrc.sin_addr =
                    222:                            ((struct sockaddr_in *) ifa->ifa_addr)->sin_addr;
                    223:                        break;
                    224:                }
                    225:        }
                    226:        /*
                    227:         * Record the incoming interface name whenever we have one.
                    228:         */
                    229:        bzero(&divsrc.sin_zero, sizeof(divsrc.sin_zero));
                    230:        if (m->m_pkthdr.rcvif) {
                    231:                /*
                    232:                 * Hide the actual interface name in there in the 
                    233:                 * sin_zero array. XXX This needs to be moved to a
                    234:                 * different sockaddr type for divert, e.g.
                    235:                 * sockaddr_div with multiple fields like 
                    236:                 * sockaddr_dl. Presently we have only 7 bytes
                    237:                 * but that will do for now as most interfaces
                    238:                 * are 4 or less + 2 or less bytes for unit.
                    239:                 * There is probably a faster way of doing this,
                    240:                 * possibly taking it from the sockaddr_dl on the iface.
                    241:                 * This solves the problem of a P2P link and a LAN interface
                    242:                 * having the same address, which can result in the wrong
                    243:                 * interface being assigned to the packet when fed back
                    244:                 * into the divert socket. Theoretically if the daemon saves
                    245:                 * and re-uses the sockaddr_in as suggested in the man pages,
                    246:                 * this iface name will come along for the ride.
                    247:                 * (see div_output for the other half of this.)
                    248:                 */ 
                    249:                snprintf(divsrc.sin_zero, sizeof(divsrc.sin_zero),
                    250:                        "%s%d", m->m_pkthdr.rcvif->if_name,
                    251:                        m->m_pkthdr.rcvif->if_unit);
                    252:        }
                    253: 
                    254:        /* Put packet on socket queue, if any */
                    255:        sa = NULL;
                    256:        for (inp = divcb.lh_first; inp != NULL; inp = inp->inp_list.le_next) {
                    257:                if (inp->inp_lport == htons(ip_divert_port))
                    258:                        sa = inp->inp_socket;
                    259:        }
                    260:        ip_divert_port = 0;
                    261:        if (sa) {
                    262:                if (sbappendaddr(&sa->so_rcv, (struct sockaddr *)&divsrc,
                    263:                                m, (struct mbuf *)0) == 0)
                    264:                        m_freem(m);
                    265:                else
                    266:                        sorwakeup(sa);
                    267:        } else {
                    268:                m_freem(m);
                    269:                ipstat.ips_noproto++;
                    270:                ipstat.ips_delivered--;
                    271:         }
                    272: }
                    273: 
                    274: /*
                    275:  * Deliver packet back into the IP processing machinery.
                    276:  *
                    277:  * If no address specified, or address is 0.0.0.0, send to ip_output();
                    278:  * otherwise, send to ip_input() and mark as having been received on
                    279:  * the interface with that address.
                    280:  */
                    281: static int
                    282: div_output(so, m, addr, control)
                    283:        struct socket *so;
                    284:        register struct mbuf *m;
                    285:        struct sockaddr *addr;
                    286:        struct mbuf *control;
                    287: {
                    288:        register struct inpcb *const inp = sotoinpcb(so);
                    289:        register struct ip *const ip = mtod(m, struct ip *);
                    290:        struct sockaddr_in *sin = (struct sockaddr_in *)addr;
                    291:        int error = 0;
                    292: 
                    293:        if (control)
                    294:                m_freem(control);               /* XXX */
                    295: 
                    296:        /* Loopback avoidance and state recovery */
                    297:        if (sin) {
                    298:                int     len = 0;
                    299:                char    *c = sin->sin_zero;
                    300: 
                    301:                ip_divert_cookie = sin->sin_port;
                    302: 
                    303:                /*
                    304:                 * Find receive interface with the given name or IP address.
                    305:                 * The name is user supplied data so don't trust it's size or 
                    306:                 * that it is zero terminated. The name has priority.
                    307:                 * We are presently assuming that the sockaddr_in 
                    308:                 * has not been replaced by a sockaddr_div, so we limit it
                    309:                 * to 16 bytes in total. the name is stuffed (if it exists)
                    310:                 * in the sin_zero[] field.
                    311:                 */
                    312:                while (*c++ && (len++ < sizeof(sin->sin_zero)));
                    313:                if ((len > 0) && (len < sizeof(sin->sin_zero)))
                    314:                        m->m_pkthdr.rcvif = ifunit(sin->sin_zero);
                    315:        } else {
                    316:                ip_divert_cookie = 0;
                    317:        }
                    318: 
                    319:        /* Reinject packet into the system as incoming or outgoing */
                    320:        if (!sin || sin->sin_addr.s_addr == 0) {
                    321:                /*
                    322:                 * Don't allow both user specified and setsockopt options,
                    323:                 * and don't allow packet length sizes that will crash
                    324:                 */
                    325:                if (((ip->ip_hl != (sizeof (*ip) >> 2)) && inp->inp_options) ||
                    326:                     ((u_short)ntohs(ip->ip_len) > m->m_pkthdr.len)) {
                    327:                        error = EINVAL;
                    328:                        goto cantsend;
                    329:                }
                    330: 
                    331:                /* Convert fields to host order for ip_output() */
                    332:                NTOHS(ip->ip_len);
                    333:                NTOHS(ip->ip_off);
                    334: 
                    335:                /* Send packet to output processing */
                    336:                ipstat.ips_rawout++;                    /* XXX */
                    337:                error = ip_output(m, inp->inp_options, &inp->inp_route,
                    338:                        (so->so_options & SO_DONTROUTE) |
                    339:                        IP_ALLOWBROADCAST | IP_RAWOUTPUT, inp->inp_moptions);
                    340:        } else {
                    341:                struct  ifaddr *ifa;
                    342: 
                    343:                /* If no luck with the name above. check by IP address.  */
                    344:                if (m->m_pkthdr.rcvif == NULL) {
                    345:                        /*
                    346:                         * Make sure there are no distractions
                    347:                         * for ifa_ifwithaddr. Clear the port and the ifname.
                    348:                         * Maybe zap all 8 bytes at once using a 64bit write?
                    349:                         */
                    350:                        bzero(sin->sin_zero, sizeof(sin->sin_zero));
                    351:                        /* *((u_int64_t *)sin->sin_zero) = 0; */ /* XXX ?? */
                    352:                        sin->sin_port = 0;
                    353:                        if (!(ifa = ifa_ifwithaddr((struct sockaddr *) sin))) {
                    354:                                error = EADDRNOTAVAIL;
                    355:                                goto cantsend;
                    356:                        }
                    357:                        m->m_pkthdr.rcvif = ifa->ifa_ifp;
                    358:                }
                    359: 
                    360:                /* Send packet to input processing */
                    361:                ip_input(m);
                    362:        }
                    363: 
                    364:        /* paranoid: Reset for next time (and other packets) */
                    365:        /* almost definitly already done in the ipfw filter but.. */
                    366:        ip_divert_cookie = 0;
                    367:        return error;
                    368: 
                    369: cantsend:
                    370:        ip_divert_cookie = 0;
                    371:        m_freem(m);
                    372:        return error;
                    373: }
                    374: 
                    375: static int
                    376: div_attach(struct socket *so, int proto, struct proc *p)
                    377: {
                    378:        struct inpcb *inp;
                    379:        int error, s;
                    380: 
                    381:        inp  = sotoinpcb(so);
                    382:        if (inp)
                    383:                panic("div_attach");
                    384:        if (p && (error = suser(p->p_ucred, &p->p_acflag)) != 0)
                    385:                return error;
                    386: 
                    387:        s = splnet();
                    388:        error = in_pcballoc(so, &divcbinfo, p);
                    389:        splx(s);
                    390:        if (error)
                    391:                return error;
                    392:        error = soreserve(so, div_sendspace, div_recvspace);
                    393:        if (error)
                    394:                return error;
                    395:        inp = (struct inpcb *)so->so_pcb;
                    396:        inp->inp_ip_p = proto;
                    397:        inp->inp_flags |= INP_HDRINCL;
                    398:        /* The socket is always "connected" because
                    399:           we always know "where" to send the packet */
                    400:        so->so_state |= SS_ISCONNECTED;
                    401:        return 0;
                    402: }
                    403: 
                    404: static int
                    405: div_detach(struct socket *so)
                    406: {
                    407:        struct inpcb *inp;
                    408: 
                    409:        inp = sotoinpcb(so);
                    410:        if (inp == 0)
                    411:                panic("div_detach");
                    412:        in_pcbdetach(inp);
                    413:        return 0;
                    414: }
                    415: 
                    416: static int
                    417: div_abort(struct socket *so)
                    418: {
                    419:        soisdisconnected(so);
                    420:        return div_detach(so);
                    421: }
                    422: 
                    423: static int
                    424: div_disconnect(struct socket *so)
                    425: {
                    426:        if ((so->so_state & SS_ISCONNECTED) == 0)
                    427:                return ENOTCONN;
                    428:        return div_abort(so);
                    429: }
                    430: 
                    431: static int
                    432: div_bind(struct socket *so, struct sockaddr *nam, struct proc *p)
                    433: {
                    434:        struct inpcb *inp;
                    435:        int s;
                    436:        int error;
                    437: 
                    438:        s = splnet();
                    439:        inp = sotoinpcb(so);
                    440:        error = in_pcbbind(inp, nam, p);
                    441:        splx(s);
                    442:        return 0;
                    443: }
                    444: 
                    445: static int
                    446: div_shutdown(struct socket *so)
                    447: {
                    448:        socantsendmore(so);
                    449:        return 0;
                    450: }
                    451: 
                    452: static int
                    453: div_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam,
                    454:         struct mbuf *control, struct proc *p)
                    455: {
                    456:        /* Packet must have a header (but that's about it) */
                    457:        if (m->m_len < sizeof (struct ip) ||
                    458:            (m = m_pullup(m, sizeof (struct ip))) == 0) {
                    459:                ipstat.ips_toosmall++;
                    460:                m_freem(m);
                    461:                return EINVAL;
                    462:        }
                    463: 
                    464:        /* Send packet */
                    465:        return div_output(so, m, nam, control);
                    466: }
                    467: 
                    468: struct pr_usrreqs div_usrreqs = {
                    469:        div_abort, pru_accept_notsupp, div_attach, div_bind,
                    470:        pru_connect_notsupp, pru_connect2_notsupp, in_control, div_detach,
                    471:        div_disconnect, pru_listen_notsupp, in_setpeeraddr, pru_rcvd_notsupp,
                    472:        pru_rcvoob_notsupp, div_send, pru_sense_null, div_shutdown,
                    473:        in_setsockaddr, sosend, soreceive, sopoll
                    474: };

unix.superglobalmegacorp.com

This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.