Home | History | Annotate | Line # | Download | only in netinet
raw_ip.c revision 1.53
      1 /*	$NetBSD: raw_ip.c,v 1.53 2000/03/30 13:25:04 augustss 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/ip_icmp.h>
     89 #include <netinet/in_pcb.h>
     90 #include <netinet/in_var.h>
     91 
     92 #include <machine/stdarg.h>
     93 
     94 #ifdef IPSEC
     95 #include <netinet6/ipsec.h>
     96 #endif /*IPSEC*/
     97 
     98 struct inpcbtable rawcbtable;
     99 
    100 int	 rip_bind __P((struct inpcb *, struct mbuf *));
    101 int	 rip_connect __P((struct inpcb *, struct mbuf *));
    102 void	 rip_disconnect __P((struct inpcb *));
    103 
    104 /*
    105  * Nominal space allocated to a raw ip socket.
    106  */
    107 #define	RIPSNDQ		8192
    108 #define	RIPRCVQ		8192
    109 
    110 /*
    111  * Raw interface to IP protocol.
    112  */
    113 
    114 /*
    115  * Initialize raw connection block q.
    116  */
    117 void
    118 rip_init()
    119 {
    120 
    121 	in_pcbinit(&rawcbtable, 1, 1);
    122 }
    123 
    124 static struct	sockaddr_in ripsrc = { sizeof(ripsrc), AF_INET };
    125 
    126 /*
    127  * Setup generic address and protocol structures
    128  * for raw_input routine, then pass them along with
    129  * mbuf chain.
    130  */
    131 void
    132 #if __STDC__
    133 rip_input(struct mbuf *m, ...)
    134 #else
    135 rip_input(m, va_alist)
    136 	struct mbuf *m;
    137 	va_dcl
    138 #endif
    139 {
    140 	int off, proto;
    141 	struct ip *ip = mtod(m, struct ip *);
    142 	struct inpcb *inp;
    143 	struct inpcb *last = 0;
    144 	struct mbuf *opts = 0;
    145 	struct sockaddr_in ripsrc;
    146 	va_list ap;
    147 
    148 	va_start(ap, m);
    149 	off = va_arg(ap, int);
    150 	proto = va_arg(ap, int);
    151 	va_end(ap);
    152 
    153 	ripsrc.sin_family = AF_INET;
    154 	ripsrc.sin_len = sizeof(struct sockaddr_in);
    155 	ripsrc.sin_addr = ip->ip_src;
    156 	ripsrc.sin_port = 0;
    157 	bzero((caddr_t)ripsrc.sin_zero, sizeof(ripsrc.sin_zero));
    158 
    159 	/*
    160 	 * XXX Compatibility: programs using raw IP expect ip_len
    161 	 * XXX to have the header length subtracted.
    162 	 */
    163 	ip->ip_len -= ip->ip_hl << 2;
    164 
    165 	for (inp = rawcbtable.inpt_queue.cqh_first;
    166 	    inp != (struct inpcb *)&rawcbtable.inpt_queue;
    167 	    inp = inp->inp_queue.cqe_next) {
    168 		if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != proto)
    169 			continue;
    170 		if (!in_nullhost(inp->inp_laddr) &&
    171 		    !in_hosteq(inp->inp_laddr, ip->ip_dst))
    172 			continue;
    173 		if (!in_nullhost(inp->inp_faddr) &&
    174 		    !in_hosteq(inp->inp_faddr, ip->ip_src))
    175 			continue;
    176 		if (last) {
    177 			struct mbuf *n;
    178 			if ((n = m_copy(m, 0, (int)M_COPYALL)) != NULL) {
    179 				if (last->inp_flags & INP_CONTROLOPTS ||
    180 				    last->inp_socket->so_options & SO_TIMESTAMP)
    181 					ip_savecontrol(last, &opts, ip, n);
    182 				if (sbappendaddr(&last->inp_socket->so_rcv,
    183 				    sintosa(&ripsrc), n, opts) == 0) {
    184 					/* should notify about lost packet */
    185 					m_freem(n);
    186 					if (opts)
    187 						m_freem(opts);
    188 				} else
    189 					sorwakeup(last->inp_socket);
    190 				opts = NULL;
    191 			}
    192 		}
    193 		last = inp;
    194 	}
    195 	if (last) {
    196 		if (last->inp_flags & INP_CONTROLOPTS ||
    197 		    last->inp_socket->so_options & SO_TIMESTAMP)
    198 			ip_savecontrol(last, &opts, ip, m);
    199 		if (sbappendaddr(&last->inp_socket->so_rcv,
    200 		    sintosa(&ripsrc), m, opts) == 0) {
    201 			m_freem(m);
    202 			if (opts)
    203 				m_freem(opts);
    204 		} else
    205 			sorwakeup(last->inp_socket);
    206 	} else {
    207 		if (inetsw[ip_protox[ip->ip_p]].pr_input == rip_input) {
    208 			icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_PROTOCOL,
    209 			    0, 0);
    210 			ipstat.ips_noproto++;
    211 			ipstat.ips_delivered--;
    212 		} else
    213 			m_freem(m);
    214 	}
    215 	return;
    216 }
    217 
    218 /*
    219  * Generate IP header and pass packet to ip_output.
    220  * Tack on options user may have setup with control call.
    221  */
    222 int
    223 #if __STDC__
    224 rip_output(struct mbuf *m, ...)
    225 #else
    226 rip_output(m, va_alist)
    227 	struct mbuf *m;
    228 	va_dcl
    229 #endif
    230 {
    231 	struct inpcb *inp;
    232 	struct ip *ip;
    233 	struct mbuf *opts;
    234 	int flags;
    235 	va_list ap;
    236 
    237 	va_start(ap, m);
    238 	inp = va_arg(ap, struct inpcb *);
    239 	va_end(ap);
    240 
    241 	flags =
    242 	    (inp->inp_socket->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST
    243 	    | IP_RETURNMTU;
    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 = 0;
    257 		ip->ip_off = 0;
    258 		ip->ip_p = inp->inp_ip.ip_p;
    259 		ip->ip_len = m->m_pkthdr.len;
    260 		ip->ip_src = inp->inp_laddr;
    261 		ip->ip_dst = inp->inp_faddr;
    262 		ip->ip_ttl = MAXTTL;
    263 		opts = inp->inp_options;
    264 	} else {
    265 		if (m->m_pkthdr.len > IP_MAXPACKET) {
    266 			m_freem(m);
    267 			return (EMSGSIZE);
    268 		}
    269 		ip = mtod(m, struct ip *);
    270 		if (m->m_pkthdr.len != ip->ip_len) {
    271 			m_freem(m);
    272 			return (EINVAL);
    273 		}
    274 		if (ip->ip_id == 0)
    275 			ip->ip_id = htons(ip_id++);
    276 		opts = NULL;
    277 		/* XXX prevent ip_output from overwriting header fields */
    278 		flags |= IP_RAWOUTPUT;
    279 		ipstat.ips_rawout++;
    280 	}
    281 #ifdef IPSEC
    282 	ipsec_setsocket(m, inp->inp_socket);
    283 #endif /*IPSEC*/
    284 	return (ip_output(m, opts, &inp->inp_route, flags, inp->inp_moptions, &inp->inp_errormtu));
    285 }
    286 
    287 /*
    288  * Raw IP socket option processing.
    289  */
    290 int
    291 rip_ctloutput(op, so, level, optname, m)
    292 	int op;
    293 	struct socket *so;
    294 	int level, optname;
    295 	struct mbuf **m;
    296 {
    297 	struct inpcb *inp = sotoinpcb(so);
    298 	int error = 0;
    299 
    300 	if (level != IPPROTO_IP) {
    301 		error = ENOPROTOOPT;
    302 		if (op == PRCO_SETOPT && *m != 0)
    303 			(void) m_free(*m);
    304 	} else switch (op) {
    305 
    306 	case PRCO_SETOPT:
    307 		switch (optname) {
    308 		case IP_HDRINCL:
    309 			if (*m == 0 || (*m)->m_len < sizeof (int))
    310 				error = EINVAL;
    311 			else {
    312 				if (*mtod(*m, int *))
    313 					inp->inp_flags |= INP_HDRINCL;
    314 				else
    315 					inp->inp_flags &= ~INP_HDRINCL;
    316 			}
    317 			if (*m != 0)
    318 				(void) m_free(*m);
    319 			break;
    320 
    321 #ifdef MROUTING
    322 		case MRT_INIT:
    323 		case MRT_DONE:
    324 		case MRT_ADD_VIF:
    325 		case MRT_DEL_VIF:
    326 		case MRT_ADD_MFC:
    327 		case MRT_DEL_MFC:
    328 		case MRT_ASSERT:
    329 			error = ip_mrouter_set(so, optname, m);
    330 			break;
    331 #endif
    332 
    333 		default:
    334 			error = ip_ctloutput(op, so, level, optname, m);
    335 			break;
    336 		}
    337 		break;
    338 
    339 	case PRCO_GETOPT:
    340 		switch (optname) {
    341 		case IP_HDRINCL:
    342 			*m = m_get(M_WAIT, M_SOOPTS);
    343 			(*m)->m_len = sizeof (int);
    344 			*mtod(*m, int *) = inp->inp_flags & INP_HDRINCL ? 1 : 0;
    345 			break;
    346 
    347 #ifdef MROUTING
    348 		case MRT_VERSION:
    349 		case MRT_ASSERT:
    350 			error = ip_mrouter_get(so, optname, m);
    351 			break;
    352 #endif
    353 
    354 		default:
    355 			error = ip_ctloutput(op, so, level, optname, m);
    356 			break;
    357 		}
    358 		break;
    359 	}
    360 	return (error);
    361 }
    362 
    363 int
    364 rip_bind(inp, nam)
    365 	struct inpcb *inp;
    366 	struct mbuf *nam;
    367 {
    368 	struct sockaddr_in *addr = mtod(nam, struct sockaddr_in *);
    369 
    370 	if (nam->m_len != sizeof(*addr))
    371 		return (EINVAL);
    372 	if (ifnet.tqh_first == 0)
    373 		return (EADDRNOTAVAIL);
    374 	if (addr->sin_family != AF_INET &&
    375 	    addr->sin_family != AF_IMPLINK)
    376 		return (EAFNOSUPPORT);
    377 	if (!in_nullhost(addr->sin_addr) &&
    378 	    ifa_ifwithaddr(sintosa(addr)) == 0)
    379 		return (EADDRNOTAVAIL);
    380 	inp->inp_laddr = addr->sin_addr;
    381 	return (0);
    382 }
    383 
    384 int
    385 rip_connect(inp, nam)
    386 	struct inpcb *inp;
    387 	struct mbuf *nam;
    388 {
    389 	struct sockaddr_in *addr = mtod(nam, struct sockaddr_in *);
    390 
    391 	if (nam->m_len != sizeof(*addr))
    392 		return (EINVAL);
    393 	if (ifnet.tqh_first == 0)
    394 		return (EADDRNOTAVAIL);
    395 	if (addr->sin_family != AF_INET &&
    396 	    addr->sin_family != AF_IMPLINK)
    397 		return (EAFNOSUPPORT);
    398 	inp->inp_faddr = addr->sin_addr;
    399 	return (0);
    400 }
    401 
    402 void
    403 rip_disconnect(inp)
    404 	struct inpcb *inp;
    405 {
    406 
    407 	inp->inp_faddr = zeroin_addr;
    408 }
    409 
    410 u_long	rip_sendspace = RIPSNDQ;
    411 u_long	rip_recvspace = RIPRCVQ;
    412 
    413 /*ARGSUSED*/
    414 int
    415 rip_usrreq(so, req, m, nam, control, p)
    416 	struct socket *so;
    417 	int req;
    418 	struct mbuf *m, *nam, *control;
    419 	struct proc *p;
    420 {
    421 	struct inpcb *inp;
    422 	int s;
    423 	int error = 0;
    424 #ifdef MROUTING
    425 	extern struct socket *ip_mrouter;
    426 #endif
    427 
    428 	if (req == PRU_CONTROL)
    429 		return (in_control(so, (long)m, (caddr_t)nam,
    430 		    (struct ifnet *)control, p));
    431 
    432 	if (req == PRU_PURGEIF) {
    433 		in_purgeif((struct ifnet *)control);
    434 		in_pcbpurgeif(&rawcbtable, (struct ifnet *)control);
    435 		return (0);
    436 	}
    437 
    438 	s = splsoftnet();
    439 	inp = sotoinpcb(so);
    440 #ifdef DIAGNOSTIC
    441 	if (req != PRU_SEND && req != PRU_SENDOOB && control)
    442 		panic("rip_usrreq: unexpected control mbuf");
    443 #endif
    444 	if (inp == 0 && req != PRU_ATTACH) {
    445 		error = EINVAL;
    446 		goto release;
    447 	}
    448 
    449 	switch (req) {
    450 
    451 	case PRU_ATTACH:
    452 		if (inp != 0) {
    453 			error = EISCONN;
    454 			break;
    455 		}
    456 		if (p == 0 || (error = suser(p->p_ucred, &p->p_acflag))) {
    457 			error = EACCES;
    458 			break;
    459 		}
    460 		if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) {
    461 			error = soreserve(so, rip_sendspace, rip_recvspace);
    462 			if (error)
    463 				break;
    464 		}
    465 		error = in_pcballoc(so, &rawcbtable);
    466 		if (error)
    467 			break;
    468 		inp = sotoinpcb(so);
    469 		inp->inp_ip.ip_p = (long)nam;
    470 #ifdef IPSEC
    471 		error = ipsec_init_policy(so, &inp->inp_sp);
    472 		if (error != 0) {
    473 			in_pcbdetach(inp);
    474 			break;
    475 		}
    476 #endif /*IPSEC*/
    477 		break;
    478 
    479 	case PRU_DETACH:
    480 #ifdef MROUTING
    481 		if (so == ip_mrouter)
    482 			ip_mrouter_done();
    483 #endif
    484 		in_pcbdetach(inp);
    485 		break;
    486 
    487 	case PRU_BIND:
    488 		error = rip_bind(inp, nam);
    489 		break;
    490 
    491 	case PRU_LISTEN:
    492 		error = EOPNOTSUPP;
    493 		break;
    494 
    495 	case PRU_CONNECT:
    496 		error = rip_connect(inp, nam);
    497 		if (error)
    498 			break;
    499 		soisconnected(so);
    500 		break;
    501 
    502 	case PRU_CONNECT2:
    503 		error = EOPNOTSUPP;
    504 		break;
    505 
    506 	case PRU_DISCONNECT:
    507 		soisdisconnected(so);
    508 		rip_disconnect(inp);
    509 		break;
    510 
    511 	/*
    512 	 * Mark the connection as being incapable of further input.
    513 	 */
    514 	case PRU_SHUTDOWN:
    515 		socantsendmore(so);
    516 		break;
    517 
    518 	case PRU_RCVD:
    519 		error = EOPNOTSUPP;
    520 		break;
    521 
    522 	/*
    523 	 * Ship a packet out.  The appropriate raw output
    524 	 * routine handles any massaging necessary.
    525 	 */
    526 	case PRU_SEND:
    527 		if (control && control->m_len) {
    528 			m_freem(control);
    529 			m_freem(m);
    530 			error = EINVAL;
    531 			break;
    532 		}
    533 	{
    534 		if (nam) {
    535 			if ((so->so_state & SS_ISCONNECTED) != 0) {
    536 				error = EISCONN;
    537 				goto die;
    538 			}
    539 			error = rip_connect(inp, nam);
    540 			if (error) {
    541 			die:
    542 				m_freem(m);
    543 				break;
    544 			}
    545 		} else {
    546 			if ((so->so_state & SS_ISCONNECTED) == 0) {
    547 				error = ENOTCONN;
    548 				goto die;
    549 			}
    550 		}
    551 		error = rip_output(m, inp);
    552 		if (nam)
    553 			rip_disconnect(inp);
    554 	}
    555 		break;
    556 
    557 	case PRU_SENSE:
    558 		/*
    559 		 * stat: don't bother with a blocksize.
    560 		 */
    561 		splx(s);
    562 		return (0);
    563 
    564 	case PRU_RCVOOB:
    565 		error = EOPNOTSUPP;
    566 		break;
    567 
    568 	case PRU_SENDOOB:
    569 		m_freem(control);
    570 		m_freem(m);
    571 		error = EOPNOTSUPP;
    572 		break;
    573 
    574 	case PRU_SOCKADDR:
    575 		in_setsockaddr(inp, nam);
    576 		break;
    577 
    578 	case PRU_PEERADDR:
    579 		in_setpeeraddr(inp, nam);
    580 		break;
    581 
    582 	default:
    583 		panic("rip_usrreq");
    584 	}
    585 
    586 release:
    587 	splx(s);
    588 	return (error);
    589 }
    590