[unix-history] / usr / src / sys / netinet / ip_output.c

/*
 * Copyright (c) 1982, 1986, 1988, 1990 Regents of the University of California.
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms are permitted
 * provided that the above copyright notice and this paragraph are
 * duplicated in all such forms and that any documentation,
 * advertising materials, and other materials related to such
 * distribution and use acknowledge that the software was developed
 * by the University of California, Berkeley.  The name of the
 * University may not be used to endorse or promote products derived
 * from this software without specific prior written permission.
 * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR
 * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED
 * WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE.
 *
 *	@(#)ip_output.c	7.19 (Berkeley) %G%
 */

#include "param.h"
#include "malloc.h"
#include "mbuf.h"
#include "errno.h"
#include "protosw.h"
#include "socket.h"
#include "socketvar.h"

#include "../net/if.h"
#include "../net/route.h"

#include "in.h"
#include "in_systm.h"
#include "ip.h"
#include "in_pcb.h"
#include "in_var.h"
#include "ip_var.h"

#ifdef vax
#include "machine/mtpr.h"
#endif

struct mbuf *ip_insertoptions();

/*
 * IP output.  The packet in mbuf chain m contains a skeletal IP
 * header (with len, off, ttl, proto, tos, src, dst).
 * The mbuf chain containing the packet will be freed.
 * The mbuf opt, if present, will not be freed.
 */
ip_output(m0, opt, ro, flags)
	struct mbuf *m0;
	struct mbuf *opt;
	struct route *ro;
	int flags;
{
	register struct ip *ip, *mhip;
	register struct ifnet *ifp;
	register struct mbuf *m = m0;
	register int hlen = sizeof (struct ip);
	int len, off, error = 0;
	struct route iproute;
	struct sockaddr_in *dst;
	struct in_ifaddr *ia;

if ((m->m_flags & M_PKTHDR) == 0)
panic("ip_output no HDR");
	if (opt) {
		m = ip_insertoptions(m, opt, &len);
		hlen = len;
	}
	ip = mtod(m, struct ip *);
	/*
	 * Fill in IP header.
	 */
	if ((flags & IP_FORWARDING) == 0) {
		ip->ip_v = IPVERSION;
		ip->ip_off &= IP_DF;
		ip->ip_id = htons(ip_id++);
		ip->ip_hl = hlen >> 2;
	} else {
		hlen = ip->ip_hl << 2;
		ipstat.ips_localout++;
	}
	/*
	 * Route packet.
	 */
	if (ro == 0) {
		ro = &iproute;
		bzero((caddr_t)ro, sizeof (*ro));
	}
	dst = (struct sockaddr_in *)&ro->ro_dst;
	/*
	 * If there is a cached route,
	 * check that it is to the same destination
	 * and is still up.  If not, free it and try again.
	 */
	if (ro->ro_rt && ((ro->ro_rt->rt_flags & RTF_UP) == 0 ||
	   dst->sin_addr.s_addr != ip->ip_dst.s_addr)) {
		RTFREE(ro->ro_rt);
		ro->ro_rt = (struct rtentry *)0;
	}
	if (ro->ro_rt == 0) {
		dst->sin_family = AF_INET;
		dst->sin_len = sizeof(*dst);
		dst->sin_addr = ip->ip_dst;
	}
	/*
	 * If routing to interface only,
	 * short circuit routing lookup.
	 */
	if (flags & IP_ROUTETOIF) {

		ia = (struct in_ifaddr *)ifa_ifwithdstaddr((struct sockaddr *)dst);
		if (ia == 0)
			ia = in_iaonnetof(in_netof(ip->ip_dst));
		if (ia == 0) {
			error = ENETUNREACH;
			goto bad;
		}
		ifp = ia->ia_ifp;
	} else {
		if (ro->ro_rt == 0)
			rtalloc(ro);
		if (ro->ro_rt == 0) {
			error = EHOSTUNREACH;
			goto bad;
		}
		ia = (struct in_ifaddr *)ro->ro_rt->rt_ifa;
		ifp = ro->ro_rt->rt_ifp;
		ro->ro_rt->rt_use++;
		if (ro->ro_rt->rt_flags & RTF_GATEWAY)
			dst = (struct sockaddr_in *)ro->ro_rt->rt_gateway;
	}
#ifndef notdef
	/*
	 * If source address not specified yet, use address
	 * of outgoing interface.
	 */
	if (ip->ip_src.s_addr == INADDR_ANY)
		ip->ip_src = IA_SIN(ia)->sin_addr;
#endif
	/*
	 * Look for broadcast address and
	 * and verify user is allowed to send
	 * such a packet.
	 */
	if (in_broadcast(dst->sin_addr)) {
		if ((ifp->if_flags & IFF_BROADCAST) == 0) {
			error = EADDRNOTAVAIL;
			goto bad;
		}
		if ((flags & IP_ALLOWBROADCAST) == 0) {
			error = EACCES;
			goto bad;
		}
		/* don't allow broadcast messages to be fragmented */
		if (ip->ip_len > ifp->if_mtu) {
			error = EMSGSIZE;
			goto bad;
		}
		m->m_flags |= M_BCAST;
	}

	/*
	 * If small enough for interface, can just send directly.
	 */
	if (ip->ip_len <= ifp->if_mtu) {
		ip->ip_len = htons((u_short)ip->ip_len);
		ip->ip_off = htons((u_short)ip->ip_off);
		ip->ip_sum = 0;
		ip->ip_sum = in_cksum(m, hlen);
		error = (*ifp->if_output)(ifp, m,
				(struct sockaddr *)dst, ro->ro_rt);
		goto done;
	}
	ipstat.ips_fragmented++;
	/*
	 * Too large for interface; fragment if possible.
	 * Must be able to put at least 8 bytes per fragment.
	 */
	if (ip->ip_off & IP_DF) {
		error = EMSGSIZE;
		goto bad;
	}
	len = (ifp->if_mtu - hlen) &~ 7;
	if (len < 8) {
		error = EMSGSIZE;
		goto bad;
	}

    {
	int mhlen, firstlen = len;
	struct mbuf **mnext = &m->m_nextpkt;

	/*
	 * Loop through length of segment after first fragment,
	 * make new header and copy data of each part and link onto chain.
	 */
	m0 = m;
	mhlen = sizeof (struct ip);
	for (off = hlen + len; off < ip->ip_len; off += len) {
		MGETHDR(m, M_DONTWAIT, MT_HEADER);
		if (m == 0) {
			error = ENOBUFS;
			goto sendorfree;
		}
		m->m_data += max_linkhdr;
		mhip = mtod(m, struct ip *);
		*mhip = *ip;
		if (hlen > sizeof (struct ip)) {
			mhlen = ip_optcopy(ip, mhip) + sizeof (struct ip);
			mhip->ip_hl = mhlen >> 2;
		}
		m->m_len = mhlen;
		mhip->ip_off = ((off - hlen) >> 3) + (ip->ip_off & ~IP_MF);
		if (ip->ip_off & IP_MF)
			mhip->ip_off |= IP_MF;
		if (off + len >= ip->ip_len)
			len = ip->ip_len - off;
		else
			mhip->ip_off |= IP_MF;
		mhip->ip_len = htons((u_short)(len + mhlen));
		m->m_next = m_copy(m0, off, len);
		if (m->m_next == 0) {
			error = ENOBUFS;	/* ??? */
			goto sendorfree;
		}
		m->m_pkthdr.len = mhlen + len;
		m->m_pkthdr.rcvif = (struct ifnet *)0;
		mhip->ip_off = htons((u_short)mhip->ip_off);
		mhip->ip_sum = 0;
		mhip->ip_sum = in_cksum(m, mhlen);
		*mnext = m;
		mnext = &m->m_nextpkt;
		ipstat.ips_ofragments++;
	}
	/*
	 * Update first fragment by trimming what's been copied out
	 * and updating header, then send each fragment (in order).
	 */
	m = m0;
	m_adj(m, hlen + firstlen - ip->ip_len);
	m->m_pkthdr.len = hlen + firstlen;
	ip->ip_len = htons((u_short)m->m_pkthdr.len);
	ip->ip_off = htons((u_short)(ip->ip_off | IP_MF));
	ip->ip_sum = 0;
	ip->ip_sum = in_cksum(m, hlen);
sendorfree:
	for (m = m0; m; m = m0) {
		m0 = m->m_nextpkt;
		m->m_nextpkt = 0;
		if (error == 0)
			error = (*ifp->if_output)(ifp, m,
			    (struct sockaddr *)dst, ro->ro_rt);
		else
			m_freem(m);
	}
    }
done:
	if (ro == &iproute && (flags & IP_ROUTETOIF) == 0 && ro->ro_rt)
		RTFREE(ro->ro_rt);
	return (error);
bad:
	m_freem(m0);
	goto done;
}

/*
 * Insert IP options into preformed packet.
 * Adjust IP destination as required for IP source routing,
 * as indicated by a non-zero in_addr at the start of the options.
 */
struct mbuf *
ip_insertoptions(m, opt, phlen)
	register struct mbuf *m;
	struct mbuf *opt;
	int *phlen;
{
	register struct ipoption *p = mtod(opt, struct ipoption *);
	struct mbuf *n;
	register struct ip *ip = mtod(m, struct ip *);
	unsigned optlen;

	optlen = opt->m_len - sizeof(p->ipopt_dst);
	if (p->ipopt_dst.s_addr)
		ip->ip_dst = p->ipopt_dst;
	if (m->m_flags & M_EXT || m->m_data - optlen < m->m_pktdat) {
		MGETHDR(n, M_DONTWAIT, MT_HEADER);
		if (n == 0)
			return (m);
		n->m_pkthdr.len = m->m_pkthdr.len + optlen;
		m->m_len -= sizeof(struct ip);
		m->m_data += sizeof(struct ip);
		n->m_next = m;
		m = n;
		m->m_len = optlen + sizeof(struct ip);
		m->m_data += max_linkhdr;
		bcopy((caddr_t)ip, mtod(m, caddr_t), sizeof(struct ip));
	} else {
		m->m_data -= optlen;
		m->m_len += optlen;
		m->m_pkthdr.len += optlen;
		ovbcopy((caddr_t)ip, mtod(m, caddr_t), sizeof(struct ip));
	}
	ip = mtod(m, struct ip *);
	bcopy((caddr_t)p->ipopt_list, (caddr_t)(ip + 1), (unsigned)optlen);
	*phlen = sizeof(struct ip) + optlen;
	ip->ip_len += optlen;
	return (m);
}

/*
 * Copy options from ip to jp,
 * omitting those not copied during fragmentation.
 */
ip_optcopy(ip, jp)
	struct ip *ip, *jp;
{
	register u_char *cp, *dp;
	int opt, optlen, cnt;

	cp = (u_char *)(ip + 1);
	dp = (u_char *)(jp + 1);
	cnt = (ip->ip_hl << 2) - sizeof (struct ip);
	for (; cnt > 0; cnt -= optlen, cp += optlen) {
		opt = cp[0];
		if (opt == IPOPT_EOL)
			break;
		if (opt == IPOPT_NOP)
			optlen = 1;
		else
			optlen = cp[IPOPT_OLEN];
		/* bogus lengths should have been caught by ip_dooptions */
		if (optlen > cnt)
			optlen = cnt;
		if (IPOPT_COPIED(opt)) {
			bcopy((caddr_t)cp, (caddr_t)dp, (unsigned)optlen);
			dp += optlen;
		}
	}
	for (optlen = dp - (u_char *)(jp+1); optlen & 0x3; optlen++)
		*dp++ = IPOPT_EOL;
	return (optlen);
}

/*
 * IP socket option processing.
 */
ip_ctloutput(op, so, level, optname, mp)
	int op;
	struct socket *so;
	int level, optname;
	struct mbuf **mp;
{
	register struct inpcb *inp = sotoinpcb(so);
	register struct mbuf *m = *mp;
	register int optval;
	int error = 0;

	if (level != IPPROTO_IP)
		error = EINVAL;
	else switch (op) {

	case PRCO_SETOPT:
		switch (optname) {
		case IP_OPTIONS:
		case IP_RETOPTS:
			return (ip_pcbopts(optname, &inp->inp_options, m));

		case IP_TOS:
		case IP_TTL:
		case IP_RECVOPTS:
		case IP_RECVRETOPTS:
		case IP_RECVDSTADDR:
			if (m->m_len != sizeof(int))
				error = EINVAL;
			else {
				optval = *mtod(m, int *);
				switch (op) {

				case IP_TOS:
					inp->inp_ip.ip_tos = optval;
					break;

				case IP_TTL:
					inp->inp_ip.ip_tos = optval;
					break;
#define	OPTSET(bit) \
	if (optval) \
		inp->inp_flags |= bit; \
	else \
		inp->inp_flags &= ~bit;

				case IP_RECVOPTS:
					OPTSET(INP_RECVOPTS);
					break;

				case IP_RECVRETOPTS:
					OPTSET(INP_RECVRETOPTS);
					break;

				case IP_RECVDSTADDR:
					OPTSET(INP_RECVDSTADDR);
					break;
				}
			}
			break;
#undef OPTSET

		default:
			error = EINVAL;
			break;
		}
		if (m)
			(void)m_free(m);
		break;

	case PRCO_GETOPT:
		switch (optname) {
		case IP_OPTIONS:
			*mp = m = m_get(M_WAIT, MT_SOOPTS);
			if (inp->inp_options) {
				m->m_len = inp->inp_options->m_len;
				bcopy(mtod(inp->inp_options, caddr_t),
				    mtod(m, caddr_t), (unsigned)m->m_len);
			} else
				m->m_len = 0;
			break;

		case IP_TOS:
		case IP_TTL:
		case IP_RECVOPTS:
		case IP_RECVRETOPTS:
		case IP_RECVDSTADDR:
			*mp = m = m_get(M_WAIT, MT_SOOPTS);
			m->m_len = sizeof(int);
			switch (op) {

			case IP_TOS:
				optval = inp->inp_ip.ip_tos;
				break;

			case IP_TTL:
				optval = inp->inp_ip.ip_tos;
				break;

#define	OPTBIT(bit)	(inp->inp_flags & bit ? 1 : 0)

			case IP_RECVOPTS:
				optval = OPTBIT(INP_RECVOPTS);
				break;

			case IP_RECVRETOPTS:
				optval = OPTBIT(INP_RECVRETOPTS);
				break;

			case IP_RECVDSTADDR:
				optval = OPTBIT(INP_RECVDSTADDR);
				break;
			}
			*mtod(m, int *) = optval;
			break;

		default:
			error = EINVAL;
			break;
		}
		break;
	}
	return (error);
}

/*
 * Set up IP options in pcb for insertion in output packets.
 * Store in mbuf with pointer in pcbopt, adding pseudo-option
 * with destination address if source routed.
 */
ip_pcbopts(pcbopt, m)
	struct mbuf **pcbopt;
	register struct mbuf *m;
{
	register cnt, optlen;
	register u_char *cp;
	u_char opt;

	/* turn off any old options */
	if (*pcbopt)
		(void)m_free(*pcbopt);
	*pcbopt = 0;
	if (m == (struct mbuf *)0 || m->m_len == 0) {
		/*
		 * Only turning off any previous options.
		 */
		if (m)
			(void)m_free(m);
		return (0);
	}

#ifndef	vax
	if (m->m_len % sizeof(long))
		goto bad;
#endif
	/*
	 * IP first-hop destination address will be stored before
	 * actual options; move other options back
	 * and clear it when none present.
	 */
	if (m->m_data + m->m_len + sizeof(struct in_addr) >= &m->m_dat[MLEN])
		goto bad;
	cnt = m->m_len;
	m->m_len += sizeof(struct in_addr);
	cp = mtod(m, u_char *) + sizeof(struct in_addr);
	ovbcopy(mtod(m, caddr_t), (caddr_t)cp, (unsigned)cnt);
	bzero(mtod(m, caddr_t), sizeof(struct in_addr));

	for (; cnt > 0; cnt -= optlen, cp += optlen) {
		opt = cp[IPOPT_OPTVAL];
		if (opt == IPOPT_EOL)
			break;
		if (opt == IPOPT_NOP)
			optlen = 1;
		else {
			optlen = cp[IPOPT_OLEN];
			if (optlen <= IPOPT_OLEN || optlen > cnt)
				goto bad;
		}
		switch (opt) {

		default:
			break;

		case IPOPT_LSRR:
		case IPOPT_SSRR:
			/*
			 * user process specifies route as:
			 *	->A->B->C->D
			 * D must be our final destination (but we can't
			 * check that since we may not have connected yet).
			 * A is first hop destination, which doesn't appear in
			 * actual IP option, but is stored before the options.
			 */
			if (optlen < IPOPT_MINOFF - 1 + sizeof(struct in_addr))
				goto bad;
			m->m_len -= sizeof(struct in_addr);
			cnt -= sizeof(struct in_addr);
			optlen -= sizeof(struct in_addr);
			cp[IPOPT_OLEN] = optlen;
			/*
			 * Move first hop before start of options.
			 */
			bcopy((caddr_t)&cp[IPOPT_OFFSET+1], mtod(m, caddr_t),
			    sizeof(struct in_addr));
			/*
			 * Then copy rest of options back
			 * to close up the deleted entry.
			 */
			ovbcopy((caddr_t)(&cp[IPOPT_OFFSET+1] +
			    sizeof(struct in_addr)),
			    (caddr_t)&cp[IPOPT_OFFSET+1],
			    (unsigned)cnt + sizeof(struct in_addr));
			break;
		}
	}
	if (m->m_len > MAX_IPOPTLEN + sizeof(struct in_addr))
		goto bad;
	*pcbopt = m;
	return (0);

bad:
	(void)m_free(m);
	return (EINVAL);
}
Commit	Line	Data
5afc289d	1	/*
76ea132e	2	* Copyright (c) 1982, 1986, 1988, 1990 Regents of the University of California.
2b6b6284	3	* All rights reserved.
5afc289d	4	*
2b6b6284	5	* Redistribution and use in source and binary forms are permitted
616d42db KB	6	* provided that the above copyright notice and this paragraph are
	7	* duplicated in all such forms and that any documentation,
	8	* advertising materials, and other materials related to such
	9	* distribution and use acknowledge that the software was developed
	10	* by the University of California, Berkeley. The name of the
	11	* University may not be used to endorse or promote products derived
	12	* from this software without specific prior written permission.
	13	* THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR
	14	* IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED
	15	* WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE.
2b6b6284	16	*
76ea132e	17	* @(#)ip_output.c 7.19 (Berkeley) %G%
5afc289d	18	*/
6e8b2eca	19
20666ad3	20	#include "param.h"
d1f75e79	21	#include "malloc.h"
20666ad3 JB	22	#include "mbuf.h"
20666ad3 JB	23	#include "errno.h"
53ba4464	24	#include "protosw.h"
20666ad3 JB	25	#include "socket.h"
20666ad3 JB	26	#include "socketvar.h"
f4d55810 SL	27
	28	#include "../net/if.h"
	29	#include "../net/route.h"
	30
20666ad3 JB	31	#include "in.h"
	32	#include "in_systm.h"
	33	#include "ip.h"
60c0bb09 KS	34	#include "in_pcb.h"
60c0bb09 KS	35	#include "in_var.h"
20666ad3	36	#include "ip_var.h"
f4d55810	37
b152d3c4	38	#ifdef vax
c18e27ed	39	#include "machine/mtpr.h"
b152d3c4	40	#endif
ac904f92	41
53ba4464 MK	42	struct mbuf *ip_insertoptions();
	43
	44	/*
	45	* IP output. The packet in mbuf chain m contains a skeletal IP
99fe25ae MK	46	* header (with len, off, ttl, proto, tos, src, dst).
	47	* The mbuf chain containing the packet will be freed.
	48	* The mbuf opt, if present, will not be freed.
53ba4464	49	*/
07136d15 MK	50	ip_output(m0, opt, ro, flags)
07136d15 MK	51	struct mbuf *m0;
8a13b737	52	struct mbuf *opt;
ee787340	53	struct route *ro;
0e3f761f	54	int flags;
ac904f92	55	{
07136d15	56	register struct ip ip, mhip;
8a13b737	57	register struct ifnet *ifp;
07136d15 MK	58	register struct mbuf *m = m0;
	59	register int hlen = sizeof (struct ip);
	60	int len, off, error = 0;
ee787340	61	struct route iproute;
a8671e7e	62	struct sockaddr_in *dst;
60c0bb09	63	struct in_ifaddr *ia;
ac904f92	64
d1f75e79 MK	65	if ((m->m_flags & M_PKTHDR) == 0)
d1f75e79 MK	66	panic("ip_output no HDR");
07136d15 MK	67	if (opt) {
	68	m = ip_insertoptions(m, opt, &len);
	69	hlen = len;
	70	}
53ba4464	71	ip = mtod(m, struct ip *);
e8d11875	72	/*
2b4b57cd	73	* Fill in IP header.
e8d11875	74	*/
0e3f761f SL	75	if ((flags & IP_FORWARDING) == 0) {
	76	ip->ip_v = IPVERSION;
	77	ip->ip_off &= IP_DF;
	78	ip->ip_id = htons(ip_id++);
4adcd589	79	ip->ip_hl = hlen >> 2;
ea9a9897	80	} else {
53ba4464	81	hlen = ip->ip_hl << 2;
ea9a9897 KS	82	ipstat.ips_localout++;
ea9a9897 KS	83	}
8a13b737	84	/*
a13c006d	85	* Route packet.
8a13b737	86	*/
ee787340 SL	87	if (ro == 0) {
	88	ro = &iproute;
	89	bzero((caddr_t)ro, sizeof (*ro));
	90	}
a8671e7e	91	dst = (struct sockaddr_in *)&ro->ro_dst;
ccb87262 MK	92	/*
	93	* If there is a cached route,
	94	* check that it is to the same destination
	95	* and is still up. If not, free it and try again.
	96	*/
	97	if (ro->ro_rt && ((ro->ro_rt->rt_flags & RTF_UP) == 0 \|\|
	98	dst->sin_addr.s_addr != ip->ip_dst.s_addr)) {
	99	RTFREE(ro->ro_rt);
	100	ro->ro_rt = (struct rtentry *)0;
	101	}
ee787340	102	if (ro->ro_rt == 0) {
a8671e7e	103	dst->sin_family = AF_INET;
d1f75e79	104	dst->sin_len = sizeof(*dst);
a8671e7e	105	dst->sin_addr = ip->ip_dst;
d55475b1 MK	106	}
	107	/*
	108	* If routing to interface only,
	109	* short circuit routing lookup.
	110	*/
	111	if (flags & IP_ROUTETOIF) {
52c6a991	112
9340d736	113	ia = (struct in_ifaddr )ifa_ifwithdstaddr((struct sockaddr )dst);
52c6a991 MK	114	if (ia == 0)
52c6a991 MK	115	ia = in_iaonnetof(in_netof(ip->ip_dst));
d55475b1 MK	116	if (ia == 0) {
	117	error = ENETUNREACH;
	118	goto bad;
a13c006d	119	}
d55475b1 MK	120	ifp = ia->ia_ifp;
d55475b1 MK	121	} else {
d55475b1 MK	122	if (ro->ro_rt == 0)
d55475b1 MK	123	rtalloc(ro);
76ea132e MK	124	if (ro->ro_rt == 0) {
76ea132e MK	125	error = EHOSTUNREACH;
d55475b1 MK	126	goto bad;
d55475b1 MK	127	}
60c0bb09	128	ia = (struct in_ifaddr *)ro->ro_rt->rt_ifa;
76ea132e	129	ifp = ro->ro_rt->rt_ifp;
d55475b1	130	ro->ro_rt->rt_use++;
39e4cc50	131	if (ro->ro_rt->rt_flags & RTF_GATEWAY)
d1f75e79	132	dst = (struct sockaddr_in *)ro->ro_rt->rt_gateway;
0e3f761f	133	}
5afc289d MK	134	#ifndef notdef
	135	/*
	136	* If source address not specified yet, use address
	137	* of outgoing interface.
	138	*/
60c0bb09 KS	139	if (ip->ip_src.s_addr == INADDR_ANY)
60c0bb09 KS	140	ip->ip_src = IA_SIN(ia)->sin_addr;
5afc289d	141	#endif
a13c006d	142	/*
af099287 SL	143	* Look for broadcast address and
af099287 SL	144	* and verify user is allowed to send
19f96414	145	* such a packet.
a13c006d	146	*/
dfce8240	147	if (in_broadcast(dst->sin_addr)) {
19f96414 SL	148	if ((ifp->if_flags & IFF_BROADCAST) == 0) {
	149	error = EADDRNOTAVAIL;
	150	goto bad;
	151	}
0e3f761f	152	if ((flags & IP_ALLOWBROADCAST) == 0) {
a13c006d	153	error = EACCES;
ee787340	154	goto bad;
8a2f82db	155	}
19f96414 SL	156	/* don't allow broadcast messages to be fragmented */
	157	if (ip->ip_len > ifp->if_mtu) {
	158	error = EMSGSIZE;
	159	goto bad;
	160	}
d1f75e79	161	m->m_flags \|= M_BCAST;
ee787340	162	}
ac904f92	163
2b4b57cd BJ	164	/*
	165	* If small enough for interface, can just send directly.
	166	*/
8a13b737 BJ	167	if (ip->ip_len <= ifp->if_mtu) {
	168	ip->ip_len = htons((u_short)ip->ip_len);
	169	ip->ip_off = htons((u_short)ip->ip_off);
	170	ip->ip_sum = 0;
	171	ip->ip_sum = in_cksum(m, hlen);
60c0bb09 KS	172	error = (*ifp->if_output)(ifp, m,
60c0bb09 KS	173	(struct sockaddr *)dst, ro->ro_rt);
a13c006d	174	goto done;
cdad2eb1	175	}
ea9a9897	176	ipstat.ips_fragmented++;
2b4b57cd BJ	177	/*
	178	* Too large for interface; fragment if possible.
	179	* Must be able to put at least 8 bytes per fragment.
	180	*/
8a2f82db SL	181	if (ip->ip_off & IP_DF) {
8a2f82db SL	182	error = EMSGSIZE;
2b4b57cd	183	goto bad;
8a2f82db	184	}
8a13b737	185	len = (ifp->if_mtu - hlen) &~ 7;
8a2f82db SL	186	if (len < 8) {
8a2f82db SL	187	error = EMSGSIZE;
2b4b57cd	188	goto bad;
8a2f82db	189	}
2b4b57cd	190
e537fed1 MK	191	{
e537fed1 MK	192	int mhlen, firstlen = len;
d1f75e79	193	struct mbuf **mnext = &m->m_nextpkt;
e537fed1	194
2b4b57cd	195	/*
e537fed1 MK	196	* Loop through length of segment after first fragment,
e537fed1 MK	197	* make new header and copy data of each part and link onto chain.
2b4b57cd	198	*/
07136d15	199	m0 = m;
e537fed1 MK	200	mhlen = sizeof (struct ip);
e537fed1 MK	201	for (off = hlen + len; off < ip->ip_len; off += len) {
d1f75e79	202	MGETHDR(m, M_DONTWAIT, MT_HEADER);
07136d15	203	if (m == 0) {
8a2f82db	204	error = ENOBUFS;
f688491d	205	goto sendorfree;
8a2f82db	206	}
d1f75e79	207	m->m_data += max_linkhdr;
07136d15	208	mhip = mtod(m, struct ip *);
2b4b57cd	209	mhip = ip;
4ad99bae	210	if (hlen > sizeof (struct ip)) {
e537fed1	211	mhlen = ip_optcopy(ip, mhip) + sizeof (struct ip);
07136d15	212	mhip->ip_hl = mhlen >> 2;
e537fed1	213	}
07136d15	214	m->m_len = mhlen;
e537fed1	215	mhip->ip_off = ((off - hlen) >> 3) + (ip->ip_off & ~IP_MF);
4adcd589 MK	216	if (ip->ip_off & IP_MF)
4adcd589 MK	217	mhip->ip_off \|= IP_MF;
e537fed1 MK	218	if (off + len >= ip->ip_len)
e537fed1 MK	219	len = ip->ip_len - off;
07136d15	220	else
2b4b57cd	221	mhip->ip_off \|= IP_MF;
07136d15 MK	222	mhip->ip_len = htons((u_short)(len + mhlen));
	223	m->m_next = m_copy(m0, off, len);
	224	if (m->m_next == 0) {
8a2f82db	225	error = ENOBUFS; /* ??? */
e537fed1	226	goto sendorfree;
98444525	227	}
d1f75e79 MK	228	m->m_pkthdr.len = mhlen + len;
d1f75e79 MK	229	m->m_pkthdr.rcvif = (struct ifnet *)0;
0b49870f	230	mhip->ip_off = htons((u_short)mhip->ip_off);
0b49870f	231	mhip->ip_sum = 0;
07136d15	232	mhip->ip_sum = in_cksum(m, mhlen);
e537fed1	233	*mnext = m;
d1f75e79	234	mnext = &m->m_nextpkt;
ea9a9897	235	ipstat.ips_ofragments++;
2b4b57cd	236	}
e537fed1 MK	237	/*
	238	* Update first fragment by trimming what's been copied out
	239	* and updating header, then send each fragment (in order).
	240	*/
9d14dd83 KM	241	m = m0;
9d14dd83 KM	242	m_adj(m, hlen + firstlen - ip->ip_len);
d1f75e79 MK	243	m->m_pkthdr.len = hlen + firstlen;
d1f75e79 MK	244	ip->ip_len = htons((u_short)m->m_pkthdr.len);
4ee97dd8	245	ip->ip_off = htons((u_short)(ip->ip_off \| IP_MF));
e537fed1	246	ip->ip_sum = 0;
9d14dd83	247	ip->ip_sum = in_cksum(m, hlen);
e537fed1 MK	248	sendorfree:
e537fed1 MK	249	for (m = m0; m; m = m0) {
d1f75e79 MK	250	m0 = m->m_nextpkt;
d1f75e79 MK	251	m->m_nextpkt = 0;
e537fed1 MK	252	if (error == 0)
e537fed1 MK	253	error = (*ifp->if_output)(ifp, m,
60c0bb09	254	(struct sockaddr *)dst, ro->ro_rt);
e537fed1 MK	255	else
	256	m_freem(m);
	257	}
	258	}
a13c006d	259	done:
0e3f761f	260	if (ro == &iproute && (flags & IP_ROUTETOIF) == 0 && ro->ro_rt)
a13c006d	261	RTFREE(ro->ro_rt);
8a2f82db	262	return (error);
e537fed1 MK	263	bad:
	264	m_freem(m0);
	265	goto done;
2b4b57cd BJ	266	}
2b4b57cd BJ	267
53ba4464 MK	268	/*
	269	* Insert IP options into preformed packet.
	270	* Adjust IP destination as required for IP source routing,
	271	* as indicated by a non-zero in_addr at the start of the options.
	272	*/
	273	struct mbuf *
	274	ip_insertoptions(m, opt, phlen)
	275	register struct mbuf *m;
	276	struct mbuf *opt;
	277	int *phlen;
	278	{
	279	register struct ipoption p = mtod(opt, struct ipoption );
	280	struct mbuf *n;
	281	register struct ip ip = mtod(m, struct ip );
8011f5df	282	unsigned optlen;
53ba4464 MK	283
	284	optlen = opt->m_len - sizeof(p->ipopt_dst);
	285	if (p->ipopt_dst.s_addr)
	286	ip->ip_dst = p->ipopt_dst;
d1f75e79 MK	287	if (m->m_flags & M_EXT \|\| m->m_data - optlen < m->m_pktdat) {
d1f75e79 MK	288	MGETHDR(n, M_DONTWAIT, MT_HEADER);
53ba4464 MK	289	if (n == 0)
53ba4464 MK	290	return (m);
d1f75e79	291	n->m_pkthdr.len = m->m_pkthdr.len + optlen;
53ba4464	292	m->m_len -= sizeof(struct ip);
d1f75e79	293	m->m_data += sizeof(struct ip);
53ba4464 MK	294	n->m_next = m;
53ba4464 MK	295	m = n;
53ba4464	296	m->m_len = optlen + sizeof(struct ip);
d1f75e79	297	m->m_data += max_linkhdr;
53ba4464 MK	298	bcopy((caddr_t)ip, mtod(m, caddr_t), sizeof(struct ip));
53ba4464 MK	299	} else {
d1f75e79	300	m->m_data -= optlen;
53ba4464	301	m->m_len += optlen;
d1f75e79	302	m->m_pkthdr.len += optlen;
53ba4464 MK	303	ovbcopy((caddr_t)ip, mtod(m, caddr_t), sizeof(struct ip));
	304	}
	305	ip = mtod(m, struct ip *);
8011f5df	306	bcopy((caddr_t)p->ipopt_list, (caddr_t)(ip + 1), (unsigned)optlen);
53ba4464 MK	307	*phlen = sizeof(struct ip) + optlen;
	308	ip->ip_len += optlen;
	309	return (m);
	310	}
	311
2b4b57cd	312	/*
e537fed1 MK	313	* Copy options from ip to jp,
e537fed1 MK	314	* omitting those not copied during fragmentation.
2b4b57cd	315	*/
e537fed1	316	ip_optcopy(ip, jp)
2b4b57cd	317	struct ip ip, jp;
2b4b57cd BJ	318	{
	319	register u_char cp, dp;
	320	int opt, optlen, cnt;
	321
	322	cp = (u_char *)(ip + 1);
	323	dp = (u_char *)(jp + 1);
	324	cnt = (ip->ip_hl << 2) - sizeof (struct ip);
	325	for (; cnt > 0; cnt -= optlen, cp += optlen) {
	326	opt = cp[0];
	327	if (opt == IPOPT_EOL)
	328	break;
	329	if (opt == IPOPT_NOP)
	330	optlen = 1;
	331	else
53ba4464	332	optlen = cp[IPOPT_OLEN];
e537fed1 MK	333	/* bogus lengths should have been caught by ip_dooptions */
	334	if (optlen > cnt)
	335	optlen = cnt;
	336	if (IPOPT_COPIED(opt)) {
4ad99bae	337	bcopy((caddr_t)cp, (caddr_t)dp, (unsigned)optlen);
2b4b57cd	338	dp += optlen;
ac904f92	339	}
e8d11875	340	}
2b4b57cd BJ	341	for (optlen = dp - (u_char *)(jp+1); optlen & 0x3; optlen++)
	342	*dp++ = IPOPT_EOL;
	343	return (optlen);
ac904f92	344	}
53ba4464 MK	345
	346	/*
	347	* IP socket option processing.
	348	*/
76ea132e	349	ip_ctloutput(op, so, level, optname, mp)
53ba4464 MK	350	int op;
	351	struct socket *so;
	352	int level, optname;
76ea132e	353	struct mbuf **mp;
53ba4464	354	{
76ea132e MK	355	register struct inpcb *inp = sotoinpcb(so);
	356	register struct mbuf m = mp;
	357	register int optval;
53ba4464	358	int error = 0;
53ba4464 MK	359
	360	if (level != IPPROTO_IP)
	361	error = EINVAL;
	362	else switch (op) {
	363
	364	case PRCO_SETOPT:
	365	switch (optname) {
	366	case IP_OPTIONS:
76ea132e MK	367	case IP_RETOPTS:
	368	return (ip_pcbopts(optname, &inp->inp_options, m));
	369
	370	case IP_TOS:
	371	case IP_TTL:
	372	case IP_RECVOPTS:
	373	case IP_RECVRETOPTS:
	374	case IP_RECVDSTADDR:
	375	if (m->m_len != sizeof(int))
	376	error = EINVAL;
	377	else {
	378	optval = mtod(m, int );
	379	switch (op) {
	380
	381	case IP_TOS:
	382	inp->inp_ip.ip_tos = optval;
	383	break;
	384
	385	case IP_TTL:
	386	inp->inp_ip.ip_tos = optval;
	387	break;
	388	#define OPTSET(bit) \
	389	if (optval) \
	390	inp->inp_flags \|= bit; \
	391	else \
	392	inp->inp_flags &= ~bit;
	393
	394	case IP_RECVOPTS:
	395	OPTSET(INP_RECVOPTS);
	396	break;
	397
	398	case IP_RECVRETOPTS:
	399	OPTSET(INP_RECVRETOPTS);
	400	break;
	401
	402	case IP_RECVDSTADDR:
	403	OPTSET(INP_RECVDSTADDR);
	404	break;
	405	}
	406	}
	407	break;
	408	#undef OPTSET
53ba4464 MK	409
	410	default:
	411	error = EINVAL;
	412	break;
	413	}
76ea132e MK	414	if (m)
76ea132e MK	415	(void)m_free(m);
53ba4464 MK	416	break;
	417
	418	case PRCO_GETOPT:
	419	switch (optname) {
	420	case IP_OPTIONS:
76ea132e	421	*mp = m = m_get(M_WAIT, MT_SOOPTS);
53ba4464	422	if (inp->inp_options) {
76ea132e	423	m->m_len = inp->inp_options->m_len;
53ba4464	424	bcopy(mtod(inp->inp_options, caddr_t),
76ea132e	425	mtod(m, caddr_t), (unsigned)m->m_len);
53ba4464	426	} else
76ea132e MK	427	m->m_len = 0;
	428	break;
	429
	430	case IP_TOS:
	431	case IP_TTL:
	432	case IP_RECVOPTS:
	433	case IP_RECVRETOPTS:
	434	case IP_RECVDSTADDR:
	435	*mp = m = m_get(M_WAIT, MT_SOOPTS);
	436	m->m_len = sizeof(int);
	437	switch (op) {
	438
	439	case IP_TOS:
	440	optval = inp->inp_ip.ip_tos;
	441	break;
	442
	443	case IP_TTL:
	444	optval = inp->inp_ip.ip_tos;
	445	break;
	446
	447	#define OPTBIT(bit) (inp->inp_flags & bit ? 1 : 0)
	448
	449	case IP_RECVOPTS:
	450	optval = OPTBIT(INP_RECVOPTS);
	451	break;
	452
	453	case IP_RECVRETOPTS:
	454	optval = OPTBIT(INP_RECVRETOPTS);
	455	break;
	456
	457	case IP_RECVDSTADDR:
	458	optval = OPTBIT(INP_RECVDSTADDR);
	459	break;
	460	}
	461	mtod(m, int ) = optval;
53ba4464	462	break;
76ea132e	463
53ba4464 MK	464	default:
	465	error = EINVAL;
	466	break;
	467	}
	468	break;
	469	}
53ba4464 MK	470	return (error);
	471	}
	472
	473	/*
0c3fb1b4 MK	474	* Set up IP options in pcb for insertion in output packets.
	475	* Store in mbuf with pointer in pcbopt, adding pseudo-option
	476	* with destination address if source routed.
53ba4464	477	*/
0c3fb1b4 MK	478	ip_pcbopts(pcbopt, m)
	479	struct mbuf **pcbopt;
	480	register struct mbuf *m;
53ba4464 MK	481	{
	482	register cnt, optlen;
	483	register u_char *cp;
	484	u_char opt;
	485
	486	/* turn off any old options */
0c3fb1b4	487	if (*pcbopt)
8011f5df	488	(void)m_free(*pcbopt);
0c3fb1b4	489	*pcbopt = 0;
53ba4464 MK	490	if (m == (struct mbuf *)0 \|\| m->m_len == 0) {
	491	/*
	492	* Only turning off any previous options.
	493	*/
	494	if (m)
8011f5df	495	(void)m_free(m);
53ba4464 MK	496	return (0);
	497	}
	498
	499	#ifndef vax
	500	if (m->m_len % sizeof(long))
	501	goto bad;
	502	#endif
	503	/*
	504	* IP first-hop destination address will be stored before
	505	* actual options; move other options back
	506	* and clear it when none present.
	507	*/
d1f75e79	508	if (m->m_data + m->m_len + sizeof(struct in_addr) >= &m->m_dat[MLEN])
53ba4464	509	goto bad;
53ba4464 MK	510	cnt = m->m_len;
	511	m->m_len += sizeof(struct in_addr);
	512	cp = mtod(m, u_char *) + sizeof(struct in_addr);
8011f5df	513	ovbcopy(mtod(m, caddr_t), (caddr_t)cp, (unsigned)cnt);
53ba4464 MK	514	bzero(mtod(m, caddr_t), sizeof(struct in_addr));
	515
	516	for (; cnt > 0; cnt -= optlen, cp += optlen) {
	517	opt = cp[IPOPT_OPTVAL];
	518	if (opt == IPOPT_EOL)
	519	break;
	520	if (opt == IPOPT_NOP)
	521	optlen = 1;
	522	else {
	523	optlen = cp[IPOPT_OLEN];
	524	if (optlen <= IPOPT_OLEN \|\| optlen > cnt)
	525	goto bad;
	526	}
	527	switch (opt) {
	528
	529	default:
	530	break;
	531
	532	case IPOPT_LSRR:
	533	case IPOPT_SSRR:
	534	/*
	535	* user process specifies route as:
	536	* ->A->B->C->D
	537	* D must be our final destination (but we can't
	538	* check that since we may not have connected yet).
	539	* A is first hop destination, which doesn't appear in
	540	* actual IP option, but is stored before the options.
	541	*/
	542	if (optlen < IPOPT_MINOFF - 1 + sizeof(struct in_addr))
	543	goto bad;
	544	m->m_len -= sizeof(struct in_addr);
	545	cnt -= sizeof(struct in_addr);
	546	optlen -= sizeof(struct in_addr);
	547	cp[IPOPT_OLEN] = optlen;
	548	/*
	549	* Move first hop before start of options.
	550	*/
8011f5df	551	bcopy((caddr_t)&cp[IPOPT_OFFSET+1], mtod(m, caddr_t),
53ba4464 MK	552	sizeof(struct in_addr));
	553	/*
	554	* Then copy rest of options back
	555	* to close up the deleted entry.
	556	*/
8011f5df MK	557	ovbcopy((caddr_t)(&cp[IPOPT_OFFSET+1] +
	558	sizeof(struct in_addr)),
	559	(caddr_t)&cp[IPOPT_OFFSET+1],
	560	(unsigned)cnt + sizeof(struct in_addr));
53ba4464 MK	561	break;
	562	}
	563	}
d1f75e79 MK	564	if (m->m_len > MAX_IPOPTLEN + sizeof(struct in_addr))
d1f75e79 MK	565	goto bad;
0c3fb1b4	566	*pcbopt = m;
53ba4464 MK	567	return (0);
	568
	569	bad:
8011f5df	570	(void)m_free(m);
53ba4464 MK	571	return (EINVAL);
53ba4464 MK	572	}