/*
* Receive data
*/
static int
ngipi_rcvdata(hook_p hook, item_p item)
{
struct mbuf *m;
NGI_GET_M(item, m);
NG_FREE_ITEM(item);
if (curthread->td_ng_outbound)
netisr_queue(NETISR_IP, m);
else
netisr_dispatch(NETISR_IP, m);
return 0;
}
int
gre_mobile_input(struct mbuf **mp, int *offp, int proto)
{
static const uint32_t af = AF_INET;
struct mbuf *m = *mp;
struct ip *ip = mtod(m, struct ip *);
struct mobip_h *mip = mtod(m, struct mobip_h *);
struct gre_softc *sc;
u_char osrc = 0;
int msiz, hlen;
hlen = *offp;
if ((sc = gre_lookup(m, IPPROTO_MOBILE)) == NULL) {
/* No matching tunnel or tunnel is down. */
m_freem(m);
return(IPPROTO_DONE);
}
sc->sc_if.if_ipackets++;
sc->sc_if.if_ibytes += m->m_pkthdr.len;
if(ntohs(mip->mh.proto) & MOB_H_SBIT) {
osrc = 1;
msiz = MOB_H_SIZ_L;
mip->mi.ip_src.s_addr = mip->mh.osrc;
} else {
msiz = MOB_H_SIZ_S;
}
mip->mi.ip_dst.s_addr = mip->mh.odst;
mip->mi.ip_p = (ntohs(mip->mh.proto) >> 8);
if (gre_in_cksum((u_short*)&mip->mh,msiz) != 0) {
m_freem(m);
return(IPPROTO_DONE);
}
bcopy((caddr_t)(ip) + (ip->ip_hl << 2) + msiz, (caddr_t)(ip) +
(ip->ip_hl << 2), m->m_len - msiz - (ip->ip_hl << 2));
m->m_len -= msiz;
m->m_pkthdr.len -= msiz;
/*
* On FreeBSD, rip_input() supplies us with ip->ip_len
* already converted into host byteorder and also decreases
* it by the lengh of IP header, however, ip_input() expects
* that this field is in the original format (network byteorder
* and full size of IP packet), so that adjust accordingly.
*/
ip->ip_len = htons(ip->ip_len + sizeof(struct ip) - msiz);
ip->ip_sum = 0;
ip->ip_sum = in_cksum(m, (ip->ip_hl << 2));
if (sc->sc_if.if_bpf)
bpf_ptap(sc->sc_if.if_bpf, m, &af, sizeof(af));
m->m_pkthdr.rcvif = &sc->sc_if;
netisr_queue(NETISR_IP, m);
return(IPPROTO_DONE);
}
static int
gre_input2(struct mbuf *m ,int hlen, u_char proto)
{
static const uint32_t af = AF_INET;
struct greip *gip = mtod(m, struct greip *);
int isr;
struct gre_softc *sc;
u_short flags;
if ((sc = gre_lookup(m, proto)) == NULL) {
/* No matching tunnel or tunnel is down. */
return (0);
}
sc->sc_if.if_ipackets++;
sc->sc_if.if_ibytes += m->m_pkthdr.len;
switch (proto) {
case IPPROTO_GRE:
hlen += sizeof (struct gre_h);
/* process GRE flags as packet can be of variable len */
flags = ntohs(gip->gi_flags);
/* Checksum & Offset are present */
if ((flags & GRE_CP) | (flags & GRE_RP))
hlen += 4;
/* We don't support routing fields (variable length) */
if (flags & GRE_RP)
return(0);
if (flags & GRE_KP)
hlen += 4;
if (flags & GRE_SP)
hlen +=4;
switch (ntohs(gip->gi_ptype)) { /* ethertypes */
case ETHERTYPE_IP:
case WCCP_PROTOCOL_TYPE:
isr = NETISR_IP;
break;
#ifdef NETATALK
case ETHERTYPE_ATALK:
isr = NETISR_ATALK1;
break;
#endif
case ETHERTYPE_IPV6:
/* FALLTHROUGH */
default: /* others not yet supported */
return(0);
}
break;
default:
/* others not yet supported */
return(0);
}
m->m_data += hlen;
m->m_len -= hlen;
m->m_pkthdr.len -= hlen;
if (sc->sc_if.if_bpf)
bpf_ptap(sc->sc_if.if_bpf, m, &af, sizeof(af));
m->m_pkthdr.rcvif = &sc->sc_if;
netisr_queue(isr, m);
return(1); /* packet is done, no further processing needed */
}
/*
* icintr()
*/
static void
icintr (device_t dev, int event, char *ptr)
{
struct ic_softc *sc = (struct ic_softc *)device_get_softc(dev);
int unit = device_get_unit(dev);
int len;
struct mbuf *top;
crit_enter();
switch (event) {
case INTR_GENERAL:
case INTR_START:
sc->ic_cp = sc->ic_ifbuf;
sc->ic_xfercnt = 0;
break;
case INTR_STOP:
/* if any error occured during transfert,
* drop the packet */
if (sc->ic_iferrs)
goto err;
if ((len = sc->ic_xfercnt) == 0)
break; /* ignore */
if (len <= ICHDRLEN)
goto err;
len -= ICHDRLEN;
sc->ic_if.if_ipackets ++;
sc->ic_if.if_ibytes += len;
BPF_TAP(&sc->ic_if, sc->ic_ifbuf, len + ICHDRLEN);
top = m_devget(sc->ic_ifbuf + ICHDRLEN, len, 0, &sc->ic_if, 0);
if (top)
netisr_queue(NETISR_IP, top);
break;
err:
kprintf("ic%d: errors (%d)!\n", unit, sc->ic_iferrs);
sc->ic_iferrs = 0; /* reset error count */
sc->ic_if.if_ierrors ++;
break;
case INTR_RECEIVE:
if (sc->ic_xfercnt >= sc->ic_if.if_mtu+ICHDRLEN) {
sc->ic_iferrs ++;
} else {
*sc->ic_cp++ = *ptr;
sc->ic_xfercnt ++;
}
break;
case INTR_NOACK: /* xfer terminated by master */
break;
case INTR_TRANSMIT:
*ptr = 0xff; /* XXX */
break;
case INTR_ERROR:
sc->ic_iferrs ++;
break;
default:
panic("%s: unknown event (%d)!", __func__, event);
}
crit_exit();
}
/*
* Forward a packet. If some error occurs return the sender
* an icmp packet. Note we can't always generate a meaningful
* icmp message because icmp doesn't have a large enough repertoire
* of codes and types.
*
* If not forwarding, just drop the packet. This could be confusing
* if ipforwarding was zero but some routing protocol was advancing
* us as a gateway to somewhere. However, we must let the routing
* protocol deal with that.
*
*/
void
ip6_forward(struct mbuf *m, int srcrt)
{
struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *);
struct sockaddr_in6 *dst = NULL;
struct rtentry *rt = NULL;
struct route_in6 rin6;
int error, type = 0, code = 0;
struct mbuf *mcopy = NULL;
struct ifnet *origifp; /* maybe unnecessary */
u_int32_t inzone, outzone;
struct in6_addr src_in6, dst_in6, odst;
struct m_tag *fwd_tag;
char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN];
/*
* Do not forward packets to multicast destination (should be handled
* by ip6_mforward().
* Do not forward packets with unspecified source. It was discussed
* in July 2000, on the ipngwg mailing list.
*/
if ((m->m_flags & (M_BCAST|M_MCAST)) != 0 ||
IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) ||
IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_src)) {
IP6STAT_INC(ip6s_cantforward);
/* XXX in6_ifstat_inc(rt->rt_ifp, ifs6_in_discard) */
if (V_ip6_log_time + V_ip6_log_interval < time_uptime) {
V_ip6_log_time = time_uptime;
log(LOG_DEBUG,
"cannot forward "
"from %s to %s nxt %d received on %s\n",
ip6_sprintf(ip6bufs, &ip6->ip6_src),
ip6_sprintf(ip6bufd, &ip6->ip6_dst),
ip6->ip6_nxt,
if_name(m->m_pkthdr.rcvif));
}
m_freem(m);
return;
}
if (
#ifdef IPSTEALTH
V_ip6stealth == 0 &&
#endif
ip6->ip6_hlim <= IPV6_HLIMDEC) {
/* XXX in6_ifstat_inc(rt->rt_ifp, ifs6_in_discard) */
icmp6_error(m, ICMP6_TIME_EXCEEDED,
ICMP6_TIME_EXCEED_TRANSIT, 0);
return;
}
/*
* Save at most ICMPV6_PLD_MAXLEN (= the min IPv6 MTU -
* size of IPv6 + ICMPv6 headers) bytes of the packet in case
* we need to generate an ICMP6 message to the src.
* Thanks to M_EXT, in most cases copy will not occur.
*
* It is important to save it before IPsec processing as IPsec
* processing may modify the mbuf.
*/
mcopy = m_copym(m, 0, imin(m->m_pkthdr.len, ICMPV6_PLD_MAXLEN),
M_NOWAIT);
#ifdef IPSTEALTH
if (V_ip6stealth == 0)
#endif
ip6->ip6_hlim -= IPV6_HLIMDEC;
#if defined(IPSEC) || defined(IPSEC_SUPPORT)
if (IPSEC_ENABLED(ipv6)) {
if ((error = IPSEC_FORWARD(ipv6, m)) != 0) {
/* mbuf consumed by IPsec */
m_freem(mcopy);
if (error != EINPROGRESS)
IP6STAT_INC(ip6s_cantforward);
return;
}
/* No IPsec processing required */
}
#endif
again:
bzero(&rin6, sizeof(struct route_in6));
dst = (struct sockaddr_in6 *)&rin6.ro_dst;
dst->sin6_len = sizeof(struct sockaddr_in6);
dst->sin6_family = AF_INET6;
dst->sin6_addr = ip6->ip6_dst;
again2:
rin6.ro_rt = in6_rtalloc1((struct sockaddr *)dst, 0, 0, M_GETFIB(m));
rt = rin6.ro_rt;
if (rin6.ro_rt != NULL)
RT_UNLOCK(rin6.ro_rt);
else {
IP6STAT_INC(ip6s_noroute);
in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_noroute);
if (mcopy) {
icmp6_error(mcopy, ICMP6_DST_UNREACH,
ICMP6_DST_UNREACH_NOROUTE, 0);
}
goto bad;
}
/*
* Source scope check: if a packet can't be delivered to its
* destination for the reason that the destination is beyond the scope
* of the source address, discard the packet and return an icmp6
* destination unreachable error with Code 2 (beyond scope of source
* address). We use a local copy of ip6_src, since in6_setscope()
* will possibly modify its first argument.
* [draft-ietf-ipngwg-icmp-v3-04.txt, Section 3.1]
*/
src_in6 = ip6->ip6_src;
if (in6_setscope(&src_in6, rt->rt_ifp, &outzone)) {
/* XXX: this should not happen */
IP6STAT_INC(ip6s_cantforward);
IP6STAT_INC(ip6s_badscope);
goto bad;
}
if (in6_setscope(&src_in6, m->m_pkthdr.rcvif, &inzone)) {
IP6STAT_INC(ip6s_cantforward);
IP6STAT_INC(ip6s_badscope);
goto bad;
}
if (inzone != outzone) {
IP6STAT_INC(ip6s_cantforward);
IP6STAT_INC(ip6s_badscope);
in6_ifstat_inc(rt->rt_ifp, ifs6_in_discard);
if (V_ip6_log_time + V_ip6_log_interval < time_uptime) {
V_ip6_log_time = time_uptime;
log(LOG_DEBUG,
"cannot forward "
"src %s, dst %s, nxt %d, rcvif %s, outif %s\n",
ip6_sprintf(ip6bufs, &ip6->ip6_src),
ip6_sprintf(ip6bufd, &ip6->ip6_dst),
ip6->ip6_nxt,
if_name(m->m_pkthdr.rcvif), if_name(rt->rt_ifp));
}
if (mcopy)
icmp6_error(mcopy, ICMP6_DST_UNREACH,
ICMP6_DST_UNREACH_BEYONDSCOPE, 0);
goto bad;
}
/*
* Destination scope check: if a packet is going to break the scope
* zone of packet's destination address, discard it. This case should
* usually be prevented by appropriately-configured routing table, but
* we need an explicit check because we may mistakenly forward the
* packet to a different zone by (e.g.) a default route.
*/
dst_in6 = ip6->ip6_dst;
if (in6_setscope(&dst_in6, m->m_pkthdr.rcvif, &inzone) != 0 ||
in6_setscope(&dst_in6, rt->rt_ifp, &outzone) != 0 ||
inzone != outzone) {
IP6STAT_INC(ip6s_cantforward);
IP6STAT_INC(ip6s_badscope);
goto bad;
}
if (rt->rt_flags & RTF_GATEWAY)
dst = (struct sockaddr_in6 *)rt->rt_gateway;
/*
* If we are to forward the packet using the same interface
* as one we got the packet from, perhaps we should send a redirect
* to sender to shortcut a hop.
* Only send redirect if source is sending directly to us,
* and if packet was not source routed (or has any options).
* Also, don't send redirect if forwarding using a route
* modified by a redirect.
*/
if (V_ip6_sendredirects && rt->rt_ifp == m->m_pkthdr.rcvif && !srcrt &&
(rt->rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) == 0) {
if ((rt->rt_ifp->if_flags & IFF_POINTOPOINT) != 0) {
/*
* If the incoming interface is equal to the outgoing
* one, and the link attached to the interface is
* point-to-point, then it will be highly probable
* that a routing loop occurs. Thus, we immediately
* drop the packet and send an ICMPv6 error message.
*
* type/code is based on suggestion by Rich Draves.
* not sure if it is the best pick.
*/
icmp6_error(mcopy, ICMP6_DST_UNREACH,
ICMP6_DST_UNREACH_ADDR, 0);
goto bad;
}
type = ND_REDIRECT;
}
/*
* Fake scoped addresses. Note that even link-local source or
* destinaion can appear, if the originating node just sends the
* packet to us (without address resolution for the destination).
* Since both icmp6_error and icmp6_redirect_output fill the embedded
* link identifiers, we can do this stuff after making a copy for
* returning an error.
*/
if ((rt->rt_ifp->if_flags & IFF_LOOPBACK) != 0) {
/*
* See corresponding comments in ip6_output.
* XXX: but is it possible that ip6_forward() sends a packet
* to a loopback interface? I don't think so, and thus
* I bark here. ([email protected])
* XXX: it is common to route invalid packets to loopback.
* also, the codepath will be visited on use of ::1 in
* rthdr. (itojun)
*/
#if 1
if (0)
#else
if ((rt->rt_flags & (RTF_BLACKHOLE|RTF_REJECT)) == 0)
#endif
{
printf("ip6_forward: outgoing interface is loopback. "
"src %s, dst %s, nxt %d, rcvif %s, outif %s\n",
ip6_sprintf(ip6bufs, &ip6->ip6_src),
ip6_sprintf(ip6bufd, &ip6->ip6_dst),
ip6->ip6_nxt, if_name(m->m_pkthdr.rcvif),
if_name(rt->rt_ifp));
}
/* we can just use rcvif in forwarding. */
origifp = m->m_pkthdr.rcvif;
}
else
origifp = rt->rt_ifp;
/*
* clear embedded scope identifiers if necessary.
* in6_clearscope will touch the addresses only when necessary.
*/
in6_clearscope(&ip6->ip6_src);
in6_clearscope(&ip6->ip6_dst);
/* Jump over all PFIL processing if hooks are not active. */
if (!PFIL_HOOKED(&V_inet6_pfil_hook))
goto pass;
odst = ip6->ip6_dst;
/* Run through list of hooks for output packets. */
error = pfil_run_hooks(&V_inet6_pfil_hook, &m, rt->rt_ifp, PFIL_OUT, NULL);
if (error != 0 || m == NULL)
goto freecopy; /* consumed by filter */
ip6 = mtod(m, struct ip6_hdr *);
/* See if destination IP address was changed by packet filter. */
if (!IN6_ARE_ADDR_EQUAL(&odst, &ip6->ip6_dst)) {
m->m_flags |= M_SKIP_FIREWALL;
/* If destination is now ourself drop to ip6_input(). */
if (in6_localip(&ip6->ip6_dst))
m->m_flags |= M_FASTFWD_OURS;
else {
RTFREE(rt);
goto again; /* Redo the routing table lookup. */
}
}
/* See if local, if yes, send it to netisr. */
if (m->m_flags & M_FASTFWD_OURS) {
if (m->m_pkthdr.rcvif == NULL)
m->m_pkthdr.rcvif = V_loif;
if (m->m_pkthdr.csum_flags & CSUM_DELAY_DATA_IPV6) {
m->m_pkthdr.csum_flags |=
CSUM_DATA_VALID_IPV6 | CSUM_PSEUDO_HDR;
m->m_pkthdr.csum_data = 0xffff;
}
#ifdef SCTP
if (m->m_pkthdr.csum_flags & CSUM_SCTP_IPV6)
m->m_pkthdr.csum_flags |= CSUM_SCTP_VALID;
#endif
error = netisr_queue(NETISR_IPV6, m);
goto out;
}
/* Or forward to some other address? */
if ((m->m_flags & M_IP6_NEXTHOP) &&
(fwd_tag = m_tag_find(m, PACKET_TAG_IPFORWARD, NULL)) != NULL) {
dst = (struct sockaddr_in6 *)&rin6.ro_dst;
bcopy((fwd_tag+1), dst, sizeof(struct sockaddr_in6));
m->m_flags |= M_SKIP_FIREWALL;
m->m_flags &= ~M_IP6_NEXTHOP;
m_tag_delete(m, fwd_tag);
RTFREE(rt);
goto again2;
}
pass:
/* See if the size was changed by the packet filter. */
if (m->m_pkthdr.len > IN6_LINKMTU(rt->rt_ifp)) {
in6_ifstat_inc(rt->rt_ifp, ifs6_in_toobig);
if (mcopy)
icmp6_error(mcopy, ICMP6_PACKET_TOO_BIG, 0,
IN6_LINKMTU(rt->rt_ifp));
goto bad;
}
error = nd6_output_ifp(rt->rt_ifp, origifp, m, dst, NULL);
if (error) {
in6_ifstat_inc(rt->rt_ifp, ifs6_out_discard);
IP6STAT_INC(ip6s_cantforward);
} else {
IP6STAT_INC(ip6s_forward);
in6_ifstat_inc(rt->rt_ifp, ifs6_out_forward);
if (type)
IP6STAT_INC(ip6s_redirectsent);
else {
if (mcopy)
goto freecopy;
}
}
if (mcopy == NULL)
goto out;
switch (error) {
case 0:
if (type == ND_REDIRECT) {
icmp6_redirect_output(mcopy, rt);
goto out;
}
goto freecopy;
case EMSGSIZE:
/* xxx MTU is constant in PPP? */
goto freecopy;
case ENOBUFS:
/* Tell source to slow down like source quench in IP? */
goto freecopy;
case ENETUNREACH: /* shouldn't happen, checked above */
case EHOSTUNREACH:
case ENETDOWN:
case EHOSTDOWN:
default:
type = ICMP6_DST_UNREACH;
code = ICMP6_DST_UNREACH_ADDR;
break;
}
icmp6_error(mcopy, type, code, 0);
goto out;
freecopy:
m_freem(mcopy);
goto out;
bad:
m_freem(m);
out:
if (rt != NULL)
RTFREE(rt);
}
/*
* Forward a packet. If some error occurs return the sender
* an icmp packet. Note we can't always generate a meaningful
* icmp message because icmp doesn't have a large enough repertoire
* of codes and types.
*
* If not forwarding, just drop the packet. This could be confusing
* if ipforwarding was zero but some routing protocol was advancing
* us as a gateway to somewhere. However, we must let the routing
* protocol deal with that.
*
*/
void
ip6_forward(struct mbuf *m, int srcrt)
{
struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *);
struct sockaddr_in6 *dst = NULL;
struct rtentry *rt = NULL;
struct route_in6 rin6;
int error, type = 0, code = 0;
struct mbuf *mcopy = NULL;
struct ifnet *origifp; /* maybe unnecessary */
u_int32_t inzone, outzone;
struct in6_addr src_in6, dst_in6, odst;
#ifdef IPSEC
struct secpolicy *sp = NULL;
#endif
#ifdef SCTP
int sw_csum;
#endif
struct m_tag *fwd_tag;
char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN];
#ifdef IPSEC
/*
* Check AH/ESP integrity.
*/
/*
* Don't increment ip6s_cantforward because this is the check
* before forwarding packet actually.
*/
if (ipsec6_in_reject(m, NULL)) {
IPSEC6STAT_INC(ips_in_polvio);
m_freem(m);
return;
}
#endif /* IPSEC */
/*
* Do not forward packets to multicast destination (should be handled
* by ip6_mforward().
* Do not forward packets with unspecified source. It was discussed
* in July 2000, on the ipngwg mailing list.
*/
if ((m->m_flags & (M_BCAST|M_MCAST)) != 0 ||
IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) ||
IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_src)) {
IP6STAT_INC(ip6s_cantforward);
/* XXX in6_ifstat_inc(rt->rt_ifp, ifs6_in_discard) */
if (V_ip6_log_time + V_ip6_log_interval < time_uptime) {
V_ip6_log_time = time_uptime;
log(LOG_DEBUG,
"cannot forward "
"from %s to %s nxt %d received on %s\n",
ip6_sprintf(ip6bufs, &ip6->ip6_src),
ip6_sprintf(ip6bufd, &ip6->ip6_dst),
ip6->ip6_nxt,
if_name(m->m_pkthdr.rcvif));
}
m_freem(m);
return;
}
#ifdef IPSTEALTH
if (!V_ip6stealth) {
#endif
if (ip6->ip6_hlim <= IPV6_HLIMDEC) {
/* XXX in6_ifstat_inc(rt->rt_ifp, ifs6_in_discard) */
icmp6_error(m, ICMP6_TIME_EXCEEDED,
ICMP6_TIME_EXCEED_TRANSIT, 0);
return;
}
ip6->ip6_hlim -= IPV6_HLIMDEC;
#ifdef IPSTEALTH
}
#endif
/*
* Save at most ICMPV6_PLD_MAXLEN (= the min IPv6 MTU -
* size of IPv6 + ICMPv6 headers) bytes of the packet in case
* we need to generate an ICMP6 message to the src.
* Thanks to M_EXT, in most cases copy will not occur.
*
* It is important to save it before IPsec processing as IPsec
* processing may modify the mbuf.
*/
mcopy = m_copy(m, 0, imin(m->m_pkthdr.len, ICMPV6_PLD_MAXLEN));
#ifdef IPSEC
/* get a security policy for this packet */
sp = ipsec_getpolicybyaddr(m, IPSEC_DIR_OUTBOUND,
IP_FORWARDING, &error);
if (sp == NULL) {
IPSEC6STAT_INC(ips_out_inval);
IP6STAT_INC(ip6s_cantforward);
if (mcopy) {
#if 0
/* XXX: what icmp ? */
#else
m_freem(mcopy);
#endif
}
m_freem(m);
return;
}
error = 0;
/* check policy */
switch (sp->policy) {
case IPSEC_POLICY_DISCARD:
/*
* This packet is just discarded.
*/
IPSEC6STAT_INC(ips_out_polvio);
IP6STAT_INC(ip6s_cantforward);
KEY_FREESP(&sp);
if (mcopy) {
#if 0
/* XXX: what icmp ? */
#else
m_freem(mcopy);
#endif
}
m_freem(m);
return;
case IPSEC_POLICY_BYPASS:
case IPSEC_POLICY_NONE:
/* no need to do IPsec. */
KEY_FREESP(&sp);
goto skip_ipsec;
case IPSEC_POLICY_IPSEC:
if (sp->req == NULL) {
/* XXX should be panic ? */
printf("ip6_forward: No IPsec request specified.\n");
IP6STAT_INC(ip6s_cantforward);
KEY_FREESP(&sp);
if (mcopy) {
#if 0
/* XXX: what icmp ? */
#else
m_freem(mcopy);
#endif
}
m_freem(m);
return;
}
/* do IPsec */
break;
case IPSEC_POLICY_ENTRUST:
default:
/* should be panic ?? */
printf("ip6_forward: Invalid policy found. %d\n", sp->policy);
KEY_FREESP(&sp);
goto skip_ipsec;
}
{
struct ipsecrequest *isr = NULL;
/*
* when the kernel forwards a packet, it is not proper to apply
* IPsec transport mode to the packet is not proper. this check
* avoid from this.
* at present, if there is even a transport mode SA request in the
* security policy, the kernel does not apply IPsec to the packet.
* this check is not enough because the following case is valid.
* ipsec esp/tunnel/xxx-xxx/require esp/transport//require;
*/
for (isr = sp->req; isr; isr = isr->next) {
if (isr->saidx.mode == IPSEC_MODE_ANY)
goto doipsectunnel;
if (isr->saidx.mode == IPSEC_MODE_TUNNEL)
goto doipsectunnel;
}
/*
* if there's no need for tunnel mode IPsec, skip.
*/
if (!isr)
goto skip_ipsec;
doipsectunnel:
/*
* All the extension headers will become inaccessible
* (since they can be encrypted).
* Don't panic, we need no more updates to extension headers
* on inner IPv6 packet (since they are now encapsulated).
*
* IPv6 [ESP|AH] IPv6 [extension headers] payload
*/
/*
* If we need to encapsulate the packet, do it here
* ipsec6_proces_packet will send the packet using ip6_output
*/
error = ipsec6_process_packet(m, sp->req);
KEY_FREESP(&sp);
if (error == EJUSTRETURN) {
/*
* We had a SP with a level of 'use' and no SA. We
* will just continue to process the packet without
* IPsec processing.
*/
error = 0;
goto skip_ipsec;
}
if (error) {
/* mbuf is already reclaimed in ipsec6_process_packet. */
switch (error) {
case EHOSTUNREACH:
case ENETUNREACH:
case EMSGSIZE:
case ENOBUFS:
case ENOMEM:
break;
default:
printf("ip6_output (ipsec): error code %d\n", error);
/* FALLTHROUGH */
case ENOENT:
/* don't show these error codes to the user */
break;
}
IP6STAT_INC(ip6s_cantforward);
if (mcopy) {
#if 0
/* XXX: what icmp ? */
#else
m_freem(mcopy);
#endif
}
return;
} else {
/*
* In the FAST IPSec case we have already
* re-injected the packet and it has been freed
* by the ipsec_done() function. So, just clean
* up after ourselves.
*/
m = NULL;
goto freecopy;
}
}
skip_ipsec:
#endif
again:
bzero(&rin6, sizeof(struct route_in6));
dst = (struct sockaddr_in6 *)&rin6.ro_dst;
dst->sin6_len = sizeof(struct sockaddr_in6);
dst->sin6_family = AF_INET6;
dst->sin6_addr = ip6->ip6_dst;
again2:
rin6.ro_rt = in6_rtalloc1((struct sockaddr *)dst, 0, 0, M_GETFIB(m));
if (rin6.ro_rt != NULL)
RT_UNLOCK(rin6.ro_rt);
else {
IP6STAT_INC(ip6s_noroute);
in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_noroute);
if (mcopy) {
icmp6_error(mcopy, ICMP6_DST_UNREACH,
ICMP6_DST_UNREACH_NOROUTE, 0);
}
goto bad;
}
rt = rin6.ro_rt;
/*
* Source scope check: if a packet can't be delivered to its
* destination for the reason that the destination is beyond the scope
* of the source address, discard the packet and return an icmp6
* destination unreachable error with Code 2 (beyond scope of source
* address). We use a local copy of ip6_src, since in6_setscope()
* will possibly modify its first argument.
* [draft-ietf-ipngwg-icmp-v3-04.txt, Section 3.1]
*/
src_in6 = ip6->ip6_src;
if (in6_setscope(&src_in6, rt->rt_ifp, &outzone)) {
/* XXX: this should not happen */
IP6STAT_INC(ip6s_cantforward);
IP6STAT_INC(ip6s_badscope);
goto bad;
}
if (in6_setscope(&src_in6, m->m_pkthdr.rcvif, &inzone)) {
IP6STAT_INC(ip6s_cantforward);
IP6STAT_INC(ip6s_badscope);
goto bad;
}
if (inzone != outzone) {
IP6STAT_INC(ip6s_cantforward);
IP6STAT_INC(ip6s_badscope);
in6_ifstat_inc(rt->rt_ifp, ifs6_in_discard);
if (V_ip6_log_time + V_ip6_log_interval < time_uptime) {
V_ip6_log_time = time_uptime;
log(LOG_DEBUG,
"cannot forward "
"src %s, dst %s, nxt %d, rcvif %s, outif %s\n",
ip6_sprintf(ip6bufs, &ip6->ip6_src),
ip6_sprintf(ip6bufd, &ip6->ip6_dst),
ip6->ip6_nxt,
if_name(m->m_pkthdr.rcvif), if_name(rt->rt_ifp));
}
if (mcopy)
icmp6_error(mcopy, ICMP6_DST_UNREACH,
ICMP6_DST_UNREACH_BEYONDSCOPE, 0);
goto bad;
}
/*
* Destination scope check: if a packet is going to break the scope
* zone of packet's destination address, discard it. This case should
* usually be prevented by appropriately-configured routing table, but
* we need an explicit check because we may mistakenly forward the
* packet to a different zone by (e.g.) a default route.
*/
dst_in6 = ip6->ip6_dst;
if (in6_setscope(&dst_in6, m->m_pkthdr.rcvif, &inzone) != 0 ||
in6_setscope(&dst_in6, rt->rt_ifp, &outzone) != 0 ||
inzone != outzone) {
IP6STAT_INC(ip6s_cantforward);
IP6STAT_INC(ip6s_badscope);
goto bad;
}
if (m->m_pkthdr.len > IN6_LINKMTU(rt->rt_ifp)) {
in6_ifstat_inc(rt->rt_ifp, ifs6_in_toobig);
if (mcopy) {
u_long mtu;
#ifdef IPSEC
struct secpolicy *sp;
int ipsecerror;
size_t ipsechdrsiz;
#endif /* IPSEC */
mtu = IN6_LINKMTU(rt->rt_ifp);
#ifdef IPSEC
/*
* When we do IPsec tunnel ingress, we need to play
* with the link value (decrement IPsec header size
* from mtu value). The code is much simpler than v4
* case, as we have the outgoing interface for
* encapsulated packet as "rt->rt_ifp".
*/
sp = ipsec_getpolicybyaddr(mcopy, IPSEC_DIR_OUTBOUND,
IP_FORWARDING, &ipsecerror);
if (sp) {
ipsechdrsiz = ipsec_hdrsiz(mcopy,
IPSEC_DIR_OUTBOUND, NULL);
if (ipsechdrsiz < mtu)
mtu -= ipsechdrsiz;
}
/*
* if mtu becomes less than minimum MTU,
* tell minimum MTU (and I'll need to fragment it).
*/
if (mtu < IPV6_MMTU)
mtu = IPV6_MMTU;
#endif /* IPSEC */
icmp6_error(mcopy, ICMP6_PACKET_TOO_BIG, 0, mtu);
}
goto bad;
}
if (rt->rt_flags & RTF_GATEWAY)
dst = (struct sockaddr_in6 *)rt->rt_gateway;
/*
* If we are to forward the packet using the same interface
* as one we got the packet from, perhaps we should send a redirect
* to sender to shortcut a hop.
* Only send redirect if source is sending directly to us,
* and if packet was not source routed (or has any options).
* Also, don't send redirect if forwarding using a route
* modified by a redirect.
*/
if (V_ip6_sendredirects && rt->rt_ifp == m->m_pkthdr.rcvif && !srcrt &&
(rt->rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) == 0) {
if ((rt->rt_ifp->if_flags & IFF_POINTOPOINT) != 0) {
/*
* If the incoming interface is equal to the outgoing
* one, and the link attached to the interface is
* point-to-point, then it will be highly probable
* that a routing loop occurs. Thus, we immediately
* drop the packet and send an ICMPv6 error message.
*
* type/code is based on suggestion by Rich Draves.
* not sure if it is the best pick.
*/
icmp6_error(mcopy, ICMP6_DST_UNREACH,
ICMP6_DST_UNREACH_ADDR, 0);
goto bad;
}
type = ND_REDIRECT;
}
/*
* Fake scoped addresses. Note that even link-local source or
* destinaion can appear, if the originating node just sends the
* packet to us (without address resolution for the destination).
* Since both icmp6_error and icmp6_redirect_output fill the embedded
* link identifiers, we can do this stuff after making a copy for
* returning an error.
*/
if ((rt->rt_ifp->if_flags & IFF_LOOPBACK) != 0) {
/*
* See corresponding comments in ip6_output.
* XXX: but is it possible that ip6_forward() sends a packet
* to a loopback interface? I don't think so, and thus
* I bark here. ([email protected])
* XXX: it is common to route invalid packets to loopback.
* also, the codepath will be visited on use of ::1 in
* rthdr. (itojun)
*/
#if 1
if (0)
#else
if ((rt->rt_flags & (RTF_BLACKHOLE|RTF_REJECT)) == 0)
#endif
{
printf("ip6_forward: outgoing interface is loopback. "
"src %s, dst %s, nxt %d, rcvif %s, outif %s\n",
ip6_sprintf(ip6bufs, &ip6->ip6_src),
ip6_sprintf(ip6bufd, &ip6->ip6_dst),
ip6->ip6_nxt, if_name(m->m_pkthdr.rcvif),
if_name(rt->rt_ifp));
}
/* we can just use rcvif in forwarding. */
origifp = m->m_pkthdr.rcvif;
}
else
origifp = rt->rt_ifp;
/*
* clear embedded scope identifiers if necessary.
* in6_clearscope will touch the addresses only when necessary.
*/
in6_clearscope(&ip6->ip6_src);
in6_clearscope(&ip6->ip6_dst);
/* Jump over all PFIL processing if hooks are not active. */
if (!PFIL_HOOKED(&V_inet6_pfil_hook))
goto pass;
odst = ip6->ip6_dst;
/* Run through list of hooks for output packets. */
error = pfil_run_hooks(&V_inet6_pfil_hook, &m, rt->rt_ifp, PFIL_OUT, NULL);
if (error != 0 || m == NULL)
goto freecopy; /* consumed by filter */
ip6 = mtod(m, struct ip6_hdr *);
/* See if destination IP address was changed by packet filter. */
if (!IN6_ARE_ADDR_EQUAL(&odst, &ip6->ip6_dst)) {
m->m_flags |= M_SKIP_FIREWALL;
/* If destination is now ourself drop to ip6_input(). */
if (in6_localip(&ip6->ip6_dst)) {
m->m_flags |= M_FASTFWD_OURS;
if (m->m_pkthdr.rcvif == NULL)
m->m_pkthdr.rcvif = V_loif;
if (m->m_pkthdr.csum_flags & CSUM_DELAY_DATA_IPV6) {
m->m_pkthdr.csum_flags |=
CSUM_DATA_VALID_IPV6 | CSUM_PSEUDO_HDR;
m->m_pkthdr.csum_data = 0xffff;
}
#ifdef SCTP
if (m->m_pkthdr.csum_flags & CSUM_SCTP_IPV6)
m->m_pkthdr.csum_flags |= CSUM_SCTP_VALID;
#endif
error = netisr_queue(NETISR_IPV6, m);
goto out;
} else
goto again; /* Redo the routing table lookup. */
}
/* See if local, if yes, send it to netisr. */
if (m->m_flags & M_FASTFWD_OURS) {
if (m->m_pkthdr.rcvif == NULL)
m->m_pkthdr.rcvif = V_loif;
if (m->m_pkthdr.csum_flags & CSUM_DELAY_DATA_IPV6) {
m->m_pkthdr.csum_flags |=
CSUM_DATA_VALID_IPV6 | CSUM_PSEUDO_HDR;
m->m_pkthdr.csum_data = 0xffff;
}
#ifdef SCTP
if (m->m_pkthdr.csum_flags & CSUM_SCTP_IPV6)
m->m_pkthdr.csum_flags |= CSUM_SCTP_VALID;
#endif
error = netisr_queue(NETISR_IPV6, m);
goto out;
}
/* Or forward to some other address? */
if ((m->m_flags & M_IP6_NEXTHOP) &&
(fwd_tag = m_tag_find(m, PACKET_TAG_IPFORWARD, NULL)) != NULL) {
dst = (struct sockaddr_in6 *)&rin6.ro_dst;
bcopy((fwd_tag+1), dst, sizeof(struct sockaddr_in6));
m->m_flags |= M_SKIP_FIREWALL;
m->m_flags &= ~M_IP6_NEXTHOP;
m_tag_delete(m, fwd_tag);
goto again2;
}
pass:
error = nd6_output(rt->rt_ifp, origifp, m, dst, rt);
if (error) {
in6_ifstat_inc(rt->rt_ifp, ifs6_out_discard);
IP6STAT_INC(ip6s_cantforward);
} else {
IP6STAT_INC(ip6s_forward);
in6_ifstat_inc(rt->rt_ifp, ifs6_out_forward);
if (type)
IP6STAT_INC(ip6s_redirectsent);
else {
if (mcopy)
goto freecopy;
}
}
if (mcopy == NULL)
goto out;
switch (error) {
case 0:
if (type == ND_REDIRECT) {
icmp6_redirect_output(mcopy, rt);
goto out;
}
goto freecopy;
case EMSGSIZE:
/* xxx MTU is constant in PPP? */
goto freecopy;
case ENOBUFS:
/* Tell source to slow down like source quench in IP? */
goto freecopy;
case ENETUNREACH: /* shouldn't happen, checked above */
case EHOSTUNREACH:
case ENETDOWN:
case EHOSTDOWN:
default:
type = ICMP6_DST_UNREACH;
code = ICMP6_DST_UNREACH_ADDR;
break;
}
icmp6_error(mcopy, type, code, 0);
goto out;
freecopy:
m_freem(mcopy);
goto out;
bad:
m_freem(m);
out:
if (rt != NULL)
RTFREE(rt);
}
/*
* Decapsulate. Does the real work and is called from gre_input()
* (above). Returns an mbuf back if packet is not yet processed,
* and NULL if it needs no further processing. proto is the protocol
* number of the "calling" foo_input() routine.
*/
static struct mbuf *
gre_input2(struct mbuf *m ,int hlen, u_char proto)
{
struct greip *gip;
int isr;
struct gre_softc *sc;
u_int16_t flags;
u_int32_t af;
if ((sc = gre_lookup(m, proto)) == NULL) {
/* No matching tunnel or tunnel is down. */
return (m);
}
if (m->m_len < sizeof(*gip)) {
m = m_pullup(m, sizeof(*gip));
if (m == NULL)
return (NULL);
}
gip = mtod(m, struct greip *);
GRE2IFP(sc)->if_ipackets++;
GRE2IFP(sc)->if_ibytes += m->m_pkthdr.len;
switch (proto) {
case IPPROTO_GRE:
hlen += sizeof(struct gre_h);
/* process GRE flags as packet can be of variable len */
flags = ntohs(gip->gi_flags);
/* Checksum & Offset are present */
if ((flags & GRE_CP) | (flags & GRE_RP))
hlen += 4;
/* We don't support routing fields (variable length) */
if (flags & GRE_RP)
return (m);
if (flags & GRE_KP)
hlen += 4;
if (flags & GRE_SP)
hlen += 4;
switch (ntohs(gip->gi_ptype)) { /* ethertypes */
case WCCP_PROTOCOL_TYPE:
if (sc->wccp_ver == WCCP_V2)
hlen += 4;
/* FALLTHROUGH */
case ETHERTYPE_IP: /* shouldn't need a schednetisr(), */
isr = NETISR_IP;/* as we are in ip_input */
af = AF_INET;
break;
#ifdef INET6
case ETHERTYPE_IPV6:
isr = NETISR_IPV6;
af = AF_INET6;
break;
#endif
#ifdef NETATALK
case ETHERTYPE_ATALK:
isr = NETISR_ATALK1;
af = AF_APPLETALK;
break;
#endif
default:
/* Others not yet supported. */
return (m);
}
break;
default:
/* Others not yet supported. */
return (m);
}
if (hlen > m->m_pkthdr.len) {
m_freem(m);
return (NULL);
}
/* Unlike NetBSD, in FreeBSD m_adj() adjusts m->m_pkthdr.len as well */
m_adj(m, hlen);
if (bpf_peers_present(GRE2IFP(sc)->if_bpf)) {
bpf_mtap2(GRE2IFP(sc)->if_bpf, &af, sizeof(af), m);
}
if ((GRE2IFP(sc)->if_flags & IFF_MONITOR) != 0) {
m_freem(m);
return(NULL);
}
m->m_pkthdr.rcvif = GRE2IFP(sc);
netisr_queue(isr, m);
/* Packet is done, no further processing needed. */
return (NULL);
}
/*
* if_simloop()
*
* This function is to support software emulation of hardware loopback,
* i.e., for interfaces with the IFF_SIMPLEX attribute. Since they can't
* hear their own broadcasts, we create a copy of the packet that we
* would normally receive via a hardware loopback.
*
* This function expects the packet to include the media header of length hlen.
*/
int
if_simloop(struct ifnet *ifp, struct mbuf *m, int af, int hlen)
{
int isr;
M_ASSERTPKTHDR(m);
m_tag_delete_nonpersistent(m);
m->m_pkthdr.rcvif = ifp;
#if 0
/*
* Let BPF see incoming packet in the following manner:
* - Emulated packet loopback for a simplex interface
* (net/if_ethersubr.c)
* -> passes it to ifp's BPF
* - IPv4/v6 multicast packet loopback (netinet(6)/ip(6)_output.c)
* -> not passes it to any BPF
* - Normal packet loopback from myself to myself (net/if_loop.c)
* -> passes to lo0's BPF (even in case of IPv6, where ifp!=lo0)
*/
if (hlen > 0) {
if (bpf_peers_present(ifp->if_bpf)) {
bpf_mtap(ifp->if_bpf, m);
}
} else {
if (bpf_peers_present(V_loif->if_bpf)) {
if ((m->m_flags & M_MCAST) == 0 || V_loif == ifp) {
/* XXX beware sizeof(af) != 4 */
u_int32_t af1 = af;
/*
* We need to prepend the address family.
*/
bpf_mtap2(V_loif->if_bpf, &af1, sizeof(af1), m);
}
}
}
#endif
/* Strip away media header */
if (hlen > 0) {
m_adj(m, hlen);
#ifndef __NO_STRICT_ALIGNMENT
/*
* Some archs do not like unaligned data, so
* we move data down in the first mbuf.
*/
if (mtod(m, vm_offset_t) & 3) {
KASSERT(hlen >= 3, ("if_simloop: hlen too small"));
bcopy(m->m_data,
(char *)(mtod(m, vm_offset_t)
- (mtod(m, vm_offset_t) & 3)),
m->m_len);
m->m_data -= (mtod(m,vm_offset_t) & 3);
}
#endif
}
/* Deliver to upper layer protocol */
switch (af) {
#ifdef INET
case AF_INET:
isr = NETISR_IP;
break;
#endif
#ifdef INET6
case AF_INET6:
m->m_flags |= M_LOOP;
isr = NETISR_IPV6;
break;
#endif
#ifdef IPX
case AF_IPX:
isr = NETISR_IPX;
break;
#endif
#ifdef NETATALK
case AF_APPLETALK:
isr = NETISR_ATALK2;
break;
#endif
default:
printf("if_simloop: can't handle af=%d\n", af);
m_freem(m);
return (EAFNOSUPPORT);
}
ifp->if_ipackets++;
ifp->if_ibytes += m->m_pkthdr.len;
netisr_queue(isr, m); /* mbuf is free'd on failure. */
return (0);
}
static inline int
ip_output_pfil(struct mbuf **mp, struct ifnet *ifp, struct inpcb *inp,
struct sockaddr_in *dst, int *fibnum, int *error)
{
struct m_tag *fwd_tag = NULL;
struct mbuf *m;
struct in_addr odst;
struct ip *ip;
m = *mp;
ip = mtod(m, struct ip *);
/* Run through list of hooks for output packets. */
odst.s_addr = ip->ip_dst.s_addr;
*error = pfil_run_hooks(&V_inet_pfil_hook, mp, ifp, PFIL_OUT, inp);
m = *mp;
if ((*error) != 0 || m == NULL)
return 1; /* Finished */
ip = mtod(m, struct ip *);
/* See if destination IP address was changed by packet filter. */
if (odst.s_addr != ip->ip_dst.s_addr) {
m->m_flags |= M_SKIP_FIREWALL;
/* If destination is now ourself drop to ip_input(). */
if (in_localip(ip->ip_dst)) {
m->m_flags |= M_FASTFWD_OURS;
if (m->m_pkthdr.rcvif == NULL)
m->m_pkthdr.rcvif = V_loif;
if (m->m_pkthdr.csum_flags & CSUM_DELAY_DATA) {
m->m_pkthdr.csum_flags |=
CSUM_DATA_VALID | CSUM_PSEUDO_HDR;
m->m_pkthdr.csum_data = 0xffff;
}
m->m_pkthdr.csum_flags |=
CSUM_IP_CHECKED | CSUM_IP_VALID;
#ifdef SCTP
if (m->m_pkthdr.csum_flags & CSUM_SCTP)
m->m_pkthdr.csum_flags |= CSUM_SCTP_VALID;
#endif
*error = netisr_queue(NETISR_IP, m);
return 1; /* Finished */
}
bzero(dst, sizeof(*dst));
dst->sin_family = AF_INET;
dst->sin_len = sizeof(*dst);
dst->sin_addr = ip->ip_dst;
return -1; /* Reloop */
}
/* See if fib was changed by packet filter. */
if ((*fibnum) != M_GETFIB(m)) {
m->m_flags |= M_SKIP_FIREWALL;
*fibnum = M_GETFIB(m);
return -1; /* Reloop for FIB change */
}
/* See if local, if yes, send it to netisr with IP_FASTFWD_OURS. */
if (m->m_flags & M_FASTFWD_OURS) {
if (m->m_pkthdr.rcvif == NULL)
m->m_pkthdr.rcvif = V_loif;
if (m->m_pkthdr.csum_flags & CSUM_DELAY_DATA) {
m->m_pkthdr.csum_flags |=
CSUM_DATA_VALID | CSUM_PSEUDO_HDR;
m->m_pkthdr.csum_data = 0xffff;
}
#ifdef SCTP
if (m->m_pkthdr.csum_flags & CSUM_SCTP)
m->m_pkthdr.csum_flags |= CSUM_SCTP_VALID;
#endif
m->m_pkthdr.csum_flags |=
CSUM_IP_CHECKED | CSUM_IP_VALID;
*error = netisr_queue(NETISR_IP, m);
return 1; /* Finished */
}
/* Or forward to some other address? */
if ((m->m_flags & M_IP_NEXTHOP) &&
((fwd_tag = m_tag_find(m, PACKET_TAG_IPFORWARD, NULL)) != NULL)) {
bcopy((fwd_tag+1), dst, sizeof(struct sockaddr_in));
m->m_flags |= M_SKIP_FIREWALL;
m->m_flags &= ~M_IP_NEXTHOP;
m_tag_delete(m, fwd_tag);
return -1; /* Reloop for CHANGE of dst */
}
return 0;
}
void
mpls_input(struct mbuf *m)
{
struct mpls *mpls = NULL;
mpls_label_t label;
M_ASSERTPKTHDR(m);
mplsstat.mplss_total++;
/* length checks already performed at mpls_demux() */
KASSERT(m->m_pkthdr.len >= sizeof(struct mpls),
("mpls_input: mpls header too small"));
again:
if (m->m_len < sizeof(struct mpls)) {
m = m_pullup(m, sizeof(struct mpls));
if (m == NULL) {
mplsstat.mplss_toosmall++;
return;
}
}
mpls = mtod(m, struct mpls*);
label = MPLS_LABEL(ntohl(mpls->mpls_shim));
switch (label) {
case 0:
/*
* Label 0: represents "IPv4 Explicit NULL Label".
*/
if (MPLS_STACK(ntohl(mpls->mpls_shim))) {
/* Decapsulate the ip datagram from the mpls frame. */
m_adj(m, sizeof(struct mpls));
netisr_queue(NETISR_IP, m);
return;
}
goto again; /* If not the bottom label, per RFC4182. */
case 1:
/*
* Label 1: represents "Router Alert Label" and is valid
* anywhere except at the bottom of the stack.
*/
break;
case 2:
/*
* Label 2: represents "IPv6 Explicit NULL Label".
*/
if (MPLS_STACK(ntohl(mpls->mpls_shim))) {
/* Decapsulate the ip datagram from the mpls frame. */
m_adj(m, sizeof(struct mpls));
netisr_queue(NETISR_IPV6, m);
return;
}
goto again; /* If not the bottom label, per RFC4182. */
case 3:
/*
* Label 3: represents the "Implicit NULL Label" and must not
* appear on the wire.
*/
break;
default:
/*
* Labels 4 - 15: reserved, drop them.
*/
if (label <= 15) {
mplsstat.mplss_reserved++;
m_freem(m);
return;
}
if (mplsforwarding) {
mpls_forward(m);
return;
} else {
mplsstat.mplss_cantforward++;
m_freem(m);
return;
}
}
mplsstat.mplss_invalid++;
m_freem(m);
}
int
esp4_input(struct mbuf **mp, int *offp, int proto)
{
int off;
struct ip *ip;
struct esp *esp;
struct esptail esptail;
struct mbuf *m;
u_int32_t spi;
struct secasvar *sav = NULL;
size_t taillen;
u_int16_t nxt;
const struct esp_algorithm *algo;
int ivlen;
size_t hlen;
size_t esplen;
off = *offp;
m = *mp;
*mp = NULL;
/* sanity check for alignment. */
if (off % 4 != 0 || m->m_pkthdr.len % 4 != 0) {
ipseclog((LOG_ERR, "IPv4 ESP input: packet alignment problem "
"(off=%d, pktlen=%d)\n", off, m->m_pkthdr.len));
ipsecstat.in_inval++;
goto bad;
}
if (m->m_len < off + ESPMAXLEN) {
m = m_pullup(m, off + ESPMAXLEN);
if (!m) {
ipseclog((LOG_DEBUG,
"IPv4 ESP input: can't pullup in esp4_input\n"));
ipsecstat.in_inval++;
goto bad;
}
}
ip = mtod(m, struct ip *);
esp = (struct esp *)(((u_int8_t *)ip) + off);
#ifdef _IP_VHL
hlen = IP_VHL_HL(ip->ip_vhl) << 2;
#else
hlen = ip->ip_hl << 2;
#endif
/* find the sassoc. */
spi = esp->esp_spi;
if ((sav = key_allocsa(AF_INET,
(caddr_t)&ip->ip_src, (caddr_t)&ip->ip_dst,
IPPROTO_ESP, spi)) == 0) {
ipseclog((LOG_WARNING,
"IPv4 ESP input: no key association found for spi %u\n",
(u_int32_t)ntohl(spi)));
ipsecstat.in_nosa++;
goto bad;
}
KEYDEBUG(KEYDEBUG_IPSEC_STAMP,
kprintf("DP esp4_input called to allocate SA:%p\n", sav));
if (sav->state != SADB_SASTATE_MATURE
&& sav->state != SADB_SASTATE_DYING) {
ipseclog((LOG_DEBUG,
"IPv4 ESP input: non-mature/dying SA found for spi %u\n",
(u_int32_t)ntohl(spi)));
ipsecstat.in_badspi++;
goto bad;
}
algo = esp_algorithm_lookup(sav->alg_enc);
if (!algo) {
ipseclog((LOG_DEBUG, "IPv4 ESP input: "
"unsupported encryption algorithm for spi %u\n",
(u_int32_t)ntohl(spi)));
ipsecstat.in_badspi++;
goto bad;
}
/* check if we have proper ivlen information */
ivlen = sav->ivlen;
if (ivlen < 0) {
ipseclog((LOG_ERR, "improper ivlen in IPv4 ESP input: %s %s\n",
ipsec4_logpacketstr(ip, spi), ipsec_logsastr(sav)));
ipsecstat.in_inval++;
goto bad;
}
if (!((sav->flags & SADB_X_EXT_OLD) == 0 && sav->replay
&& (sav->alg_auth && sav->key_auth)))
goto noreplaycheck;
if (sav->alg_auth == SADB_X_AALG_NULL ||
sav->alg_auth == SADB_AALG_NONE)
goto noreplaycheck;
/*
* check for sequence number.
*/
if (ipsec_chkreplay(ntohl(((struct newesp *)esp)->esp_seq), sav))
; /* okey */
else {
ipsecstat.in_espreplay++;
ipseclog((LOG_WARNING,
"replay packet in IPv4 ESP input: %s %s\n",
ipsec4_logpacketstr(ip, spi), ipsec_logsastr(sav)));
goto bad;
}
/* check ICV */
{
u_char sum0[AH_MAXSUMSIZE];
u_char sum[AH_MAXSUMSIZE];
const struct ah_algorithm *sumalgo;
size_t siz;
sumalgo = ah_algorithm_lookup(sav->alg_auth);
if (!sumalgo)
goto noreplaycheck;
siz = (((*sumalgo->sumsiz)(sav) + 3) & ~(4 - 1));
if (m->m_pkthdr.len < off + ESPMAXLEN + siz) {
ipsecstat.in_inval++;
goto bad;
}
if (AH_MAXSUMSIZE < siz) {
ipseclog((LOG_DEBUG,
"internal error: AH_MAXSUMSIZE must be larger than %lu\n",
(u_long)siz));
ipsecstat.in_inval++;
goto bad;
}
m_copydata(m, m->m_pkthdr.len - siz, siz, &sum0[0]);
if (esp_auth(m, off, m->m_pkthdr.len - off - siz, sav, sum)) {
ipseclog((LOG_WARNING, "auth fail in IPv4 ESP input: %s %s\n",
ipsec4_logpacketstr(ip, spi), ipsec_logsastr(sav)));
ipsecstat.in_espauthfail++;
goto bad;
}
if (bcmp(sum0, sum, siz) != 0) {
ipseclog((LOG_WARNING, "auth fail in IPv4 ESP input: %s %s\n",
ipsec4_logpacketstr(ip, spi), ipsec_logsastr(sav)));
ipsecstat.in_espauthfail++;
goto bad;
}
/* strip off the authentication data */
m_adj(m, -siz);
ip = mtod(m, struct ip *);
#ifdef IPLEN_FLIPPED
ip->ip_len = ip->ip_len - siz;
#else
ip->ip_len = htons(ntohs(ip->ip_len) - siz);
#endif
m->m_flags |= M_AUTHIPDGM;
ipsecstat.in_espauthsucc++;
}
/*
* update sequence number.
*/
if ((sav->flags & SADB_X_EXT_OLD) == 0 && sav->replay) {
if (ipsec_updatereplay(ntohl(((struct newesp *)esp)->esp_seq), sav)) {
ipsecstat.in_espreplay++;
goto bad;
}
}
noreplaycheck:
/* process main esp header. */
if (sav->flags & SADB_X_EXT_OLD) {
/* RFC 1827 */
esplen = sizeof(struct esp);
} else {
/* RFC 2406 */
if (sav->flags & SADB_X_EXT_DERIV)
esplen = sizeof(struct esp);
else
esplen = sizeof(struct newesp);
}
if (m->m_pkthdr.len < off + esplen + ivlen + sizeof(esptail)) {
ipseclog((LOG_WARNING,
"IPv4 ESP input: packet too short\n"));
ipsecstat.in_inval++;
goto bad;
}
if (m->m_len < off + esplen + ivlen) {
m = m_pullup(m, off + esplen + ivlen);
if (!m) {
ipseclog((LOG_DEBUG,
"IPv4 ESP input: can't pullup in esp4_input\n"));
ipsecstat.in_inval++;
goto bad;
}
}
/*
* pre-compute and cache intermediate key
*/
if (esp_schedule(algo, sav) != 0) {
ipsecstat.in_inval++;
goto bad;
}
/*
* decrypt the packet.
*/
if (!algo->decrypt)
panic("internal error: no decrypt function");
if ((*algo->decrypt)(m, off, sav, algo, ivlen)) {
/* m is already freed */
m = NULL;
ipseclog((LOG_ERR, "decrypt fail in IPv4 ESP input: %s\n",
ipsec_logsastr(sav)));
ipsecstat.in_inval++;
goto bad;
}
ipsecstat.in_esphist[sav->alg_enc]++;
m->m_flags |= M_DECRYPTED;
/*
* find the trailer of the ESP.
*/
m_copydata(m, m->m_pkthdr.len - sizeof(esptail), sizeof(esptail),
(caddr_t)&esptail);
nxt = esptail.esp_nxt;
taillen = esptail.esp_padlen + sizeof(esptail);
if (m->m_pkthdr.len < taillen ||
m->m_pkthdr.len - taillen < off + esplen + ivlen + sizeof(esptail)) {
ipseclog((LOG_WARNING,
"bad pad length in IPv4 ESP input: %s %s\n",
ipsec4_logpacketstr(ip, spi), ipsec_logsastr(sav)));
ipsecstat.in_inval++;
goto bad;
}
/* strip off the trailing pad area. */
m_adj(m, -taillen);
#ifdef IPLEN_FLIPPED
ip->ip_len = ip->ip_len - taillen;
#else
ip->ip_len = htons(ntohs(ip->ip_len) - taillen);
#endif
/* was it transmitted over the IPsec tunnel SA? */
if (ipsec4_tunnel_validate(m, off + esplen + ivlen, nxt, sav)) {
/*
* strip off all the headers that precedes ESP header.
* IP4 xx ESP IP4' payload -> IP4' payload
*
* XXX more sanity checks
* XXX relationship with gif?
*/
u_int8_t tos;
tos = ip->ip_tos;
m_adj(m, off + esplen + ivlen);
if (m->m_len < sizeof(*ip)) {
m = m_pullup(m, sizeof(*ip));
if (!m) {
ipsecstat.in_inval++;
goto bad;
}
}
ip = mtod(m, struct ip *);
/* ECN consideration. */
ip_ecn_egress(ip4_ipsec_ecn, &tos, &ip->ip_tos);
if (!key_checktunnelsanity(sav, AF_INET,
(caddr_t)&ip->ip_src, (caddr_t)&ip->ip_dst)) {
ipseclog((LOG_ERR, "ipsec tunnel address mismatch "
"in IPv4 ESP input: %s %s\n",
ipsec4_logpacketstr(ip, spi), ipsec_logsastr(sav)));
ipsecstat.in_inval++;
goto bad;
}
key_sa_recordxfer(sav, m);
if (ipsec_addhist(m, IPPROTO_ESP, spi) != 0 ||
ipsec_addhist(m, IPPROTO_IPV4, 0) != 0) {
ipsecstat.in_nomem++;
goto bad;
}
if (netisr_queue(NETISR_IP, m)) {
ipsecstat.in_inval++;
m = NULL;
goto bad;
}
nxt = IPPROTO_DONE;
} else {
