Home | History | Annotate | Line # | Download | only in netinet
raw_ip.c revision 1.42.6.3
      1 /*	$NetBSD: raw_ip.c,v 1.42.6.3 1999/11/30 13:35:34 itojun Exp $	*/
      2 
      3 /*
      4  * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
      5  * All rights reserved.
      6  *
      7  * Redistribution and use in source and binary forms, with or without
      8  * modification, are permitted provided that the following conditions
      9  * are met:
     10  * 1. Redistributions of source code must retain the above copyright
     11  *    notice, this list of conditions and the following disclaimer.
     12  * 2. Redistributions in binary form must reproduce the above copyright
     13  *    notice, this list of conditions and the following disclaimer in the
     14  *    documentation and/or other materials provided with the distribution.
     15  * 3. Neither the name of the project nor the names of its contributors
     16  *    may be used to endorse or promote products derived from this software
     17  *    without specific prior written permission.
     18  *
     19  * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
     20  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     21  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     22  * ARE DISCLAIMED.  IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
     23  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     24  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     25  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     26  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     27  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     28  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     29  * SUCH DAMAGE.
     30  */
     31 
     32 /*
     33  * Copyright (c) 1982, 1986, 1988, 1993
     34  *	The Regents of the University of California.  All rights reserved.
     35  *
     36  * Redistribution and use in source and binary forms, with or without
     37  * modification, are permitted provided that the following conditions
     38  * are met:
     39  * 1. Redistributions of source code must retain the above copyright
     40  *    notice, this list of conditions and the following disclaimer.
     41  * 2. Redistributions in binary form must reproduce the above copyright
     42  *    notice, this list of conditions and the following disclaimer in the
     43  *    documentation and/or other materials provided with the distribution.
     44  * 3. All advertising materials mentioning features or use of this software
     45  *    must display the following acknowledgement:
     46  *	This product includes software developed by the University of
     47  *	California, Berkeley and its contributors.
     48  * 4. Neither the name of the University nor the names of its contributors
     49  *    may be used to endorse or promote products derived from this software
     50  *    without specific prior written permission.
     51  *
     52  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     53  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     54  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     55  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     56  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     57  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     58  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     59  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     60  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     61  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     62  * SUCH DAMAGE.
     63  *
     64  *	@(#)raw_ip.c	8.7 (Berkeley) 5/15/95
     65  */
     66 
     67 #include "opt_ipsec.h"
     68 #include "opt_mrouting.h"
     69 
     70 #include <sys/param.h>
     71 #include <sys/malloc.h>
     72 #include <sys/mbuf.h>
     73 #include <sys/socket.h>
     74 #include <sys/protosw.h>
     75 #include <sys/socketvar.h>
     76 #include <sys/errno.h>
     77 #include <sys/systm.h>
     78 #include <sys/proc.h>
     79 
     80 #include <net/if.h>
     81 #include <net/route.h>
     82 
     83 #include <netinet/in.h>
     84 #include <netinet/in_systm.h>
     85 #include <netinet/ip.h>
     86 #include <netinet/ip_var.h>
     87 #include <netinet/ip_mroute.h>
     88 #include <netinet/in_pcb.h>
     89 #include <netinet/in_var.h>
     90 
     91 #include <machine/stdarg.h>
     92 
     93 #ifdef IPSEC
     94 #include <netinet6/ipsec.h>
     95 #endif /*IPSEC*/
     96 
     97 struct inpcbtable rawcbtable;
     98 
     99 int	 rip_bind __P((struct inpcb *, struct mbuf *));
    100 int	 rip_connect __P((struct inpcb *, struct mbuf *));
    101 void	 rip_disconnect __P((struct inpcb *));
    102 
    103 /*
    104  * Nominal space allocated to a raw ip socket.
    105  */
    106 #define	RIPSNDQ		8192
    107 #define	RIPRCVQ		8192
    108 
    109 /*
    110  * Raw interface to IP protocol.
    111  */
    112 
    113 /*
    114  * Initialize raw connection block q.
    115  */
    116 void
    117 rip_init()
    118 {
    119 
    120 	in_pcbinit(&rawcbtable, 1, 1);
    121 }
    122 
    123 static struct	sockaddr_in ripsrc = { sizeof(ripsrc), AF_INET };
    124 
    125 /*
    126  * Setup generic address and protocol structures
    127  * for raw_input routine, then pass them along with
    128  * mbuf chain.
    129  */
    130 void
    131 #if __STDC__
    132 rip_input(struct mbuf *m, ...)
    133 #else
    134 rip_input(m, va_alist)
    135 	struct mbuf *m;
    136 	va_dcl
    137 #endif
    138 {
    139 	int off, proto;
    140 	register struct ip *ip = mtod(m, struct ip *);
    141 	register struct inpcb *inp;
    142 	struct inpcb *last = 0;
    143 	struct mbuf *opts = 0;
    144 	struct sockaddr_in ripsrc;
    145 	va_list ap;
    146 
    147 	va_start(ap, m);
    148 	off = va_arg(ap, int);
    149 	proto = va_arg(ap, int);
    150 	va_end(ap);
    151 
    152 	ripsrc.sin_family = AF_INET;
    153 	ripsrc.sin_len = sizeof(struct sockaddr_in);
    154 	ripsrc.sin_addr = ip->ip_src;
    155 	ripsrc.sin_port = 0;
    156 	bzero((caddr_t)ripsrc.sin_zero, sizeof(ripsrc.sin_zero));
    157 
    158 	/*
    159 	 * XXX Compatibility: programs using raw IP expect ip_len
    160 	 * XXX to have the header length subtracted.
    161 	 */
    162 	ip->ip_len -= ip->ip_hl << 2;
    163 
    164 	for (inp = rawcbtable.inpt_queue.cqh_first;
    165 	    inp != (struct inpcb *)&rawcbtable.inpt_queue;
    166 	    inp = inp->inp_queue.cqe_next) {
    167 		if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != proto)
    168 			continue;
    169 		if (!in_nullhost(inp->inp_laddr) &&
    170 		    !in_hosteq(inp->inp_laddr, ip->ip_dst))
    171 			continue;
    172 		if (!in_nullhost(inp->inp_faddr) &&
    173 		    !in_hosteq(inp->inp_faddr, ip->ip_src))
    174 			continue;
    175 		if (last) {
    176 			struct mbuf *n;
    177 			if ((n = m_copy(m, 0, (int)M_COPYALL)) != NULL) {
    178 				if (last->inp_flags & INP_CONTROLOPTS ||
    179 				    last->inp_socket->so_options & SO_TIMESTAMP)
    180 					ip_savecontrol(last, &opts, ip, n);
    181 				if (sbappendaddr(&last->inp_socket->so_rcv,
    182 				    sintosa(&ripsrc), n, opts) == 0) {
    183 					/* should notify about lost packet */
    184 					m_freem(n);
    185 					if (opts)
    186 						m_freem(opts);
    187 				} else
    188 					sorwakeup(last->inp_socket);
    189 				opts = NULL;
    190 			}
    191 		}
    192 		last = inp;
    193 	}
    194 	if (last) {
    195 		if (last->inp_flags & INP_CONTROLOPTS ||
    196 		    last->inp_socket->so_options & SO_TIMESTAMP)
    197 			ip_savecontrol(last, &opts, ip, m);
    198 		if (sbappendaddr(&last->inp_socket->so_rcv,
    199 		    sintosa(&ripsrc), m, opts) == 0) {
    200 			m_freem(m);
    201 			if (opts)
    202 				m_freem(opts);
    203 		} else
    204 			sorwakeup(last->inp_socket);
    205 	} else {
    206 		m_freem(m);
    207 		ipstat.ips_noproto++;
    208 		ipstat.ips_delivered--;
    209 	}
    210 	return;
    211 }
    212 
    213 /*
    214  * Generate IP header and pass packet to ip_output.
    215  * Tack on options user may have setup with control call.
    216  */
    217 int
    218 #if __STDC__
    219 rip_output(struct mbuf *m, ...)
    220 #else
    221 rip_output(m, va_alist)
    222 	struct mbuf *m;
    223 	va_dcl
    224 #endif
    225 {
    226 	register struct inpcb *inp;
    227 	register struct ip *ip;
    228 	struct mbuf *opts;
    229 	int flags;
    230 	va_list ap;
    231 
    232 	va_start(ap, m);
    233 	inp = va_arg(ap, struct inpcb *);
    234 	va_end(ap);
    235 
    236 	flags =
    237 	    (inp->inp_socket->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST
    238 	    | IP_RETURNMTU;
    239 
    240 	/*
    241 	 * If the user handed us a complete IP packet, use it.
    242 	 * Otherwise, allocate an mbuf for a header and fill it in.
    243 	 */
    244 	if ((inp->inp_flags & INP_HDRINCL) == 0) {
    245 		if ((m->m_pkthdr.len + sizeof(struct ip)) > IP_MAXPACKET) {
    246 			m_freem(m);
    247 			return (EMSGSIZE);
    248 		}
    249 		M_PREPEND(m, sizeof(struct ip), M_WAIT);
    250 		ip = mtod(m, struct ip *);
    251 		ip->ip_tos = 0;
    252 		ip->ip_off = 0;
    253 		ip->ip_p = inp->inp_ip.ip_p;
    254 		ip->ip_len = m->m_pkthdr.len;
    255 		ip->ip_src = inp->inp_laddr;
    256 		ip->ip_dst = inp->inp_faddr;
    257 		ip->ip_ttl = MAXTTL;
    258 		opts = inp->inp_options;
    259 	} else {
    260 		if (m->m_pkthdr.len > IP_MAXPACKET) {
    261 			m_freem(m);
    262 			return (EMSGSIZE);
    263 		}
    264 		ip = mtod(m, struct ip *);
    265 		if (m->m_pkthdr.len != ip->ip_len) {
    266 			m_freem(m);
    267 			return (EINVAL);
    268 		}
    269 		if (ip->ip_id == 0)
    270 			ip->ip_id = htons(ip_id++);
    271 		opts = NULL;
    272 		/* XXX prevent ip_output from overwriting header fields */
    273 		flags |= IP_RAWOUTPUT;
    274 		ipstat.ips_rawout++;
    275 	}
    276 #ifdef IPSEC
    277 	m->m_pkthdr.rcvif = (struct ifnet *)inp->inp_socket;	/*XXX*/
    278 #endif /*IPSEC*/
    279 	return (ip_output(m, opts, &inp->inp_route, flags, inp->inp_moptions, &inp->inp_errormtu));
    280 }
    281 
    282 /*
    283  * Raw IP socket option processing.
    284  */
    285 int
    286 rip_ctloutput(op, so, level, optname, m)
    287 	int op;
    288 	struct socket *so;
    289 	int level, optname;
    290 	struct mbuf **m;
    291 {
    292 	register struct inpcb *inp = sotoinpcb(so);
    293 	int error = 0;
    294 
    295 	if (level != IPPROTO_IP) {
    296 		error = ENOPROTOOPT;
    297 		if (op == PRCO_SETOPT && *m != 0)
    298 			(void) m_free(*m);
    299 	} else switch (op) {
    300 
    301 	case PRCO_SETOPT:
    302 		switch (optname) {
    303 		case IP_HDRINCL:
    304 			if (*m == 0 || (*m)->m_len < sizeof (int))
    305 				error = EINVAL;
    306 			else {
    307 				if (*mtod(*m, int *))
    308 					inp->inp_flags |= INP_HDRINCL;
    309 				else
    310 					inp->inp_flags &= ~INP_HDRINCL;
    311 			}
    312 			if (*m != 0)
    313 				(void) m_free(*m);
    314 			break;
    315 
    316 #ifdef MROUTING
    317 		case MRT_INIT:
    318 		case MRT_DONE:
    319 		case MRT_ADD_VIF:
    320 		case MRT_DEL_VIF:
    321 		case MRT_ADD_MFC:
    322 		case MRT_DEL_MFC:
    323 		case MRT_ASSERT:
    324 			error = ip_mrouter_set(so, optname, m);
    325 			break;
    326 #endif
    327 
    328 		default:
    329 			error = ip_ctloutput(op, so, level, optname, m);
    330 			break;
    331 		}
    332 		break;
    333 
    334 	case PRCO_GETOPT:
    335 		switch (optname) {
    336 		case IP_HDRINCL:
    337 			*m = m_get(M_WAIT, M_SOOPTS);
    338 			(*m)->m_len = sizeof (int);
    339 			*mtod(*m, int *) = inp->inp_flags & INP_HDRINCL ? 1 : 0;
    340 			break;
    341 
    342 #ifdef MROUTING
    343 		case MRT_VERSION:
    344 		case MRT_ASSERT:
    345 			error = ip_mrouter_get(so, optname, m);
    346 			break;
    347 #endif
    348 
    349 		default:
    350 			error = ip_ctloutput(op, so, level, optname, m);
    351 			break;
    352 		}
    353 		break;
    354 	}
    355 	return (error);
    356 }
    357 
    358 int
    359 rip_bind(inp, nam)
    360 	struct inpcb *inp;
    361 	struct mbuf *nam;
    362 {
    363 	struct sockaddr_in *addr = mtod(nam, struct sockaddr_in *);
    364 
    365 	if (nam->m_len != sizeof(*addr))
    366 		return (EINVAL);
    367 	if (ifnet.tqh_first == 0)
    368 		return (EADDRNOTAVAIL);
    369 	if (addr->sin_family != AF_INET &&
    370 	    addr->sin_family != AF_IMPLINK)
    371 		return (EAFNOSUPPORT);
    372 	if (!in_nullhost(addr->sin_addr) &&
    373 	    ifa_ifwithaddr(sintosa(addr)) == 0)
    374 		return (EADDRNOTAVAIL);
    375 	inp->inp_laddr = addr->sin_addr;
    376 	return (0);
    377 }
    378 
    379 int
    380 rip_connect(inp, nam)
    381 	struct inpcb *inp;
    382 	struct mbuf *nam;
    383 {
    384 	struct sockaddr_in *addr = mtod(nam, struct sockaddr_in *);
    385 
    386 	if (nam->m_len != sizeof(*addr))
    387 		return (EINVAL);
    388 	if (ifnet.tqh_first == 0)
    389 		return (EADDRNOTAVAIL);
    390 	if (addr->sin_family != AF_INET &&
    391 	    addr->sin_family != AF_IMPLINK)
    392 		return (EAFNOSUPPORT);
    393 	inp->inp_faddr = addr->sin_addr;
    394 	return (0);
    395 }
    396 
    397 void
    398 rip_disconnect(inp)
    399 	struct inpcb *inp;
    400 {
    401 
    402 	inp->inp_faddr = zeroin_addr;
    403 }
    404 
    405 u_long	rip_sendspace = RIPSNDQ;
    406 u_long	rip_recvspace = RIPRCVQ;
    407 
    408 /*ARGSUSED*/
    409 int
    410 rip_usrreq(so, req, m, nam, control, p)
    411 	register struct socket *so;
    412 	int req;
    413 	struct mbuf *m, *nam, *control;
    414 	struct proc *p;
    415 {
    416 	register struct inpcb *inp;
    417 	int s;
    418 	register int error = 0;
    419 #ifdef MROUTING
    420 	extern struct socket *ip_mrouter;
    421 #endif
    422 
    423 	if (req == PRU_CONTROL)
    424 		return (in_control(so, (long)m, (caddr_t)nam,
    425 		    (struct ifnet *)control, p));
    426 
    427 	s = splsoftnet();
    428 	inp = sotoinpcb(so);
    429 #ifdef DIAGNOSTIC
    430 	if (req != PRU_SEND && req != PRU_SENDOOB && control)
    431 		panic("rip_usrreq: unexpected control mbuf");
    432 #endif
    433 	if (inp == 0 && req != PRU_ATTACH) {
    434 		error = EINVAL;
    435 		goto release;
    436 	}
    437 
    438 	switch (req) {
    439 
    440 	case PRU_ATTACH:
    441 		if (inp != 0) {
    442 			error = EISCONN;
    443 			break;
    444 		}
    445 		if (p == 0 || (error = suser(p->p_ucred, &p->p_acflag))) {
    446 			error = EACCES;
    447 			break;
    448 		}
    449 		if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) {
    450 			error = soreserve(so, rip_sendspace, rip_recvspace);
    451 			if (error)
    452 				break;
    453 		}
    454 		error = in_pcballoc(so, &rawcbtable);
    455 		if (error)
    456 			break;
    457 		inp = sotoinpcb(so);
    458 		inp->inp_ip.ip_p = (long)nam;
    459 #ifdef IPSEC
    460 		error = ipsec_init_policy(so, &inp->inp_sp);
    461 		if (error != 0) {
    462 			in_pcbdetach(inp);
    463 			break;
    464 		}
    465 #endif /*IPSEC*/
    466 		break;
    467 
    468 	case PRU_DETACH:
    469 #ifdef MROUTING
    470 		if (so == ip_mrouter)
    471 			ip_mrouter_done();
    472 #endif
    473 		in_pcbdetach(inp);
    474 		break;
    475 
    476 	case PRU_BIND:
    477 		error = rip_bind(inp, nam);
    478 		break;
    479 
    480 	case PRU_LISTEN:
    481 		error = EOPNOTSUPP;
    482 		break;
    483 
    484 	case PRU_CONNECT:
    485 		error = rip_connect(inp, nam);
    486 		if (error)
    487 			break;
    488 		soisconnected(so);
    489 		break;
    490 
    491 	case PRU_CONNECT2:
    492 		error = EOPNOTSUPP;
    493 		break;
    494 
    495 	case PRU_DISCONNECT:
    496 		soisdisconnected(so);
    497 		rip_disconnect(inp);
    498 		break;
    499 
    500 	/*
    501 	 * Mark the connection as being incapable of further input.
    502 	 */
    503 	case PRU_SHUTDOWN:
    504 		socantsendmore(so);
    505 		break;
    506 
    507 	case PRU_RCVD:
    508 		error = EOPNOTSUPP;
    509 		break;
    510 
    511 	/*
    512 	 * Ship a packet out.  The appropriate raw output
    513 	 * routine handles any massaging necessary.
    514 	 */
    515 	case PRU_SEND:
    516 		if (control && control->m_len) {
    517 			m_freem(control);
    518 			m_freem(m);
    519 			error = EINVAL;
    520 			break;
    521 		}
    522 	{
    523 		if (nam) {
    524 			if ((so->so_state & SS_ISCONNECTED) != 0) {
    525 				error = EISCONN;
    526 				goto die;
    527 			}
    528 			error = rip_connect(inp, nam);
    529 			if (error) {
    530 			die:
    531 				m_freem(m);
    532 				break;
    533 			}
    534 		} else {
    535 			if ((so->so_state & SS_ISCONNECTED) == 0) {
    536 				error = ENOTCONN;
    537 				goto die;
    538 			}
    539 		}
    540 		error = rip_output(m, inp);
    541 		if (nam)
    542 			rip_disconnect(inp);
    543 	}
    544 		break;
    545 
    546 	case PRU_SENSE:
    547 		/*
    548 		 * stat: don't bother with a blocksize.
    549 		 */
    550 		splx(s);
    551 		return (0);
    552 
    553 	case PRU_RCVOOB:
    554 		error = EOPNOTSUPP;
    555 		break;
    556 
    557 	case PRU_SENDOOB:
    558 		m_freem(control);
    559 		m_freem(m);
    560 		error = EOPNOTSUPP;
    561 		break;
    562 
    563 	case PRU_SOCKADDR:
    564 		in_setsockaddr(inp, nam);
    565 		break;
    566 
    567 	case PRU_PEERADDR:
    568 		in_setpeeraddr(inp, nam);
    569 		break;
    570 
    571 	default:
    572 		panic("rip_usrreq");
    573 	}
    574 
    575 release:
    576 	splx(s);
    577 	return (error);
    578 }
    579