struct ifaddr *
ifa_ifwithroute(int flags, struct sockaddr *dst, struct sockaddr *gateway,
u_int rtableid)
{
struct ifaddr *ifa;
#ifdef IPSEC
/*
* If the destination is a PF_KEY address, we'll look
* for the existence of a encap interface number or address
* in the options list of the gateway. By default, we'll return
* enc0.
*/
if (dst && (dst->sa_family == PF_KEY))
return (encap_findgwifa(gateway));
#endif
if ((flags & RTF_GATEWAY) == 0) {
/*
* If we are adding a route to an interface,
* and the interface is a pt to pt link
* we should search for the destination
* as our clue to the interface. Otherwise
* we can use the local address.
*/
ifa = NULL;
if (flags & RTF_HOST)
ifa = ifa_ifwithdstaddr(dst, rtableid);
if (ifa == NULL)
ifa = ifa_ifwithaddr(gateway, rtableid);
} else {
/*
* If we are adding a route to a remote net
* or host, the gateway may still be on the
* other end of a pt to pt link.
*/
ifa = ifa_ifwithdstaddr(gateway, rtableid);
}
if (ifa == NULL)
ifa = ifa_ifwithnet(gateway, rtableid);
if (ifa == NULL) {
struct rtentry *rt = rtalloc1(gateway, 0, rtable_l2(rtableid));
if (rt == NULL)
return (NULL);
rt->rt_refcnt--;
/* The gateway must be local if the same address family. */
if ((rt->rt_flags & RTF_GATEWAY) &&
rt_key(rt)->sa_family == dst->sa_family)
return (0);
if ((ifa = rt->rt_ifa) == NULL)
return (NULL);
}
if (ifa->ifa_addr->sa_family != dst->sa_family) {
struct ifaddr *oifa = ifa;
ifa = ifaof_ifpforaddr(dst, ifa->ifa_ifp);
if (ifa == NULL)
ifa = oifa;
}
return (ifa);
}
/*
* Provide an opportunity for a TOE driver to offload.
*/
int
tcp_offload_connect(struct socket *so, struct sockaddr *nam)
{
struct ifnet *ifp;
struct toedev *tod;
struct rtentry *rt;
int error = EOPNOTSUPP;
INP_WLOCK_ASSERT(sotoinpcb(so));
KASSERT(nam->sa_family == AF_INET || nam->sa_family == AF_INET6,
("%s: called with sa_family %d", __func__, nam->sa_family));
if (registered_toedevs == 0)
return (error);
rt = rtalloc1(nam, 0, 0);
if (rt)
RT_UNLOCK(rt);
else
return (EHOSTUNREACH);
ifp = rt->rt_ifp;
if (nam->sa_family == AF_INET && !(ifp->if_capenable & IFCAP_TOE4))
goto done;
if (nam->sa_family == AF_INET6 && !(ifp->if_capenable & IFCAP_TOE6))
goto done;
tod = TOEDEV(ifp);
if (tod != NULL)
error = tod->tod_connect(tod, so, rt, nam);
done:
RTFREE(rt);
return (error);
}
/*
* Packet routing routines.
*/
void
rtalloc(struct route *ro)
{
if (ro->ro_rt && ro->ro_rt->rt_ifp && (ro->ro_rt->rt_flags & RTF_UP))
return; /* XXX */
ro->ro_rt = rtalloc1(&ro->ro_dst, 1, 0);
}
struct hcentry *
hc_get(struct sockaddr *sa)
{
u_long hash;
struct hcentry *hc;
struct hctable *hct;
int s;
hct = &hctable[sa->sa_family];
if (hct->hct_nentries == 0)
return 0;
hash = hct->hct_cb->hccb_hash(sa, hct->hct_nentries);
hc = hct->hct_heads[hash].lh_first;
for (; hc; hc = hc->hc_link.le_next) {
if (cmpsa(hc->hc_host, sa) == 0)
break;
}
if (hc == 0)
return 0;
s = splnet();
if (hc->hc_rt && (hc->hc_rt->rt_flags & RTF_UP) == 0) {
RTFREE(hc->hc_rt);
hc->hc_rt = 0;
}
if (hc->hc_rt == 0) {
hc->hc_rt = rtalloc1(hc->hc_host, 1, 0);
}
hc_ref(hc);
splx(s);
/* XXX move to front of list? */
return hc;
}
/*
* Start output on the mpe interface.
*/
void
mpestart(struct ifnet *ifp)
{
struct mbuf *m;
struct sockaddr *sa = (struct sockaddr *)&mpedst;
int s;
sa_family_t af;
struct rtentry *rt;
for (;;) {
s = splnet();
IFQ_DEQUEUE(&ifp->if_snd, m);
splx(s);
if (m == NULL)
return;
af = *mtod(m, sa_family_t *);
m_adj(m, sizeof(af));
switch (af) {
case AF_INET:
bzero(sa, sizeof(struct sockaddr_in));
satosin(sa)->sin_family = af;
satosin(sa)->sin_len = sizeof(struct sockaddr_in);
bcopy(mtod(m, caddr_t), &satosin(sa)->sin_addr,
sizeof(in_addr_t));
m_adj(m, sizeof(in_addr_t));
break;
default:
m_freem(m);
continue;
}
rt = rtalloc1(sa, RT_REPORT, 0);
if (rt == NULL) {
/* no route give up */
m_freem(m);
continue;
}
#if NBPFILTER > 0
if (ifp->if_bpf) {
/* remove MPLS label before passing packet to bpf */
m->m_data += sizeof(struct shim_hdr);
m->m_len -= sizeof(struct shim_hdr);
m->m_pkthdr.len -= sizeof(struct shim_hdr);
bpf_mtap_af(ifp->if_bpf, af, m, BPF_DIRECTION_OUT);
m->m_data -= sizeof(struct shim_hdr);
m->m_len += sizeof(struct shim_hdr);
m->m_pkthdr.len += sizeof(struct shim_hdr);
}
#endif
/* XXX lie, but mpls_output will only look at sa_family */
sa->sa_family = AF_MPLS;
mpls_output(rt->rt_ifp, m, sa, rt);
RTFREE(rt);
}
}
void
rtalloc_mpath(struct route *ro, int hash)
{
struct radix_node *rn0, *rn;
int n;
/*
* XXX we don't attempt to lookup cached route again; what should
* be done for sendto(3) case?
*/
if (ro->ro_rt && ro->ro_rt->rt_ifp && (ro->ro_rt->rt_flags & RTF_UP))
return; /* XXX */
#ifdef __FreeBSD__
ro->ro_rt = rtalloc1(&ro->ro_dst, 1, 0UL);
#else
ro->ro_rt = rtalloc1(&ro->ro_dst, 1);
#endif
/* if the route does not exist or it is not multipath, don't care */
if (!ro->ro_rt || !rn_mpath_next((struct radix_node *)ro->ro_rt))
return;
/* beyond here, we use rn as the master copy */
rn0 = rn = (struct radix_node *)ro->ro_rt;
n = rn_mpath_count(rn0);
/* gw selection by Modulo-N Hash (RFC2991) XXX need improvement? */
hash += hashjitter;
hash %= n;
while (hash-- > 0 && rn) {
/* stay within the multipath routes */
if (rn->rn_dupedkey && rn->rn_mask != rn->rn_dupedkey->rn_mask)
break;
rn = rn->rn_dupedkey;
}
/* XXX try filling rt_gwroute and avoid unreachable gw */
/* if gw selection fails, use the first match (default) */
if (!rn)
return;
rtfree(ro->ro_rt);
ro->ro_rt = (struct rtentry *)rn;
ro->ro_rt->rt_refcnt++;
}
/*
* allocate a route, potentially using multipath to select the peer.
*/
void
rtalloc_mpath(struct route *ro, u_int32_t *srcaddrp)
{
#if defined(INET) || defined(INET6)
struct radix_node *rn;
int hash, npaths, threshold;
#endif
/*
* return a cached entry if it is still valid, otherwise we increase
* the risk of disrupting local flows.
*/
if (ro->ro_rt && ro->ro_rt->rt_ifp && (ro->ro_rt->rt_flags & RTF_UP))
return;
ro->ro_rt = rtalloc1(&ro->ro_dst, RT_REPORT, ro->ro_tableid);
/* if the route does not exist or it is not multipath, don't care */
if (!ro->ro_rt || !(ro->ro_rt->rt_flags & RTF_MPATH))
return;
/* check if multipath routing is enabled for the specified protocol */
if (!(0
#ifdef INET
|| (ipmultipath && ro->ro_dst.sa_family == AF_INET)
#endif
#ifdef INET6
|| (ip6_multipath && ro->ro_dst.sa_family == AF_INET6)
#endif
))
return;
#if defined(INET) || defined(INET6)
/* gw selection by Hash-Threshold (RFC 2992) */
rn = (struct radix_node *)ro->ro_rt;
npaths = rn_mpath_count(rn);
hash = rn_mpath_hash(ro, srcaddrp) & 0xffff;
threshold = 1 + (0xffff / npaths);
while (hash > threshold && rn) {
/* stay within the multipath routes */
if (rn_mpath_next(rn, 0) == NULL)
break;
rn = rn->rn_dupedkey;
hash -= threshold;
}
/* XXX try filling rt_gwroute and avoid unreachable gw */
/* if gw selection fails, use the first match (default) */
if (!rn)
return;
rtfree(ro->ro_rt);
ro->ro_rt = (struct rtentry *)rn;
ro->ro_rt->rt_refcnt++;
#endif
}
int
tcp_offload_connect(struct socket *so, struct sockaddr *nam)
{
struct ifnet *ifp;
struct toedev *tdev;
struct rtentry *rt;
int error;
if (toedev_registration_count == 0)
return (EINVAL);
/*
* Look up the route used for the connection to
* determine if it uses an interface capable of
* offloading the connection.
*/
rt = rtalloc1(nam, 0 /*report*/, 0 /*ignflags*/);
if (rt)
RT_UNLOCK(rt);
else
return (EHOSTUNREACH);
ifp = rt->rt_ifp;
if ((ifp->if_capenable & IFCAP_TOE) == 0) {
error = EINVAL;
goto fail;
}
tdev = TOEDEV(ifp);
if (tdev == NULL) {
error = EPERM;
goto fail;
}
if (tdev->tod_can_offload(tdev, so) == 0) {
error = EPERM;
goto fail;
}
return (tdev->tod_connect(tdev, so, rt, nam));
fail:
RTFREE(rt);
return (error);
}
/*
* Force a routing table entry to the specified
* destination to go through the given gateway.
* Normally called as a result of a routing redirect
* message from the network layer.
*
* N.B.: must be called at splsoftnet
*/
void
rtredirect(struct sockaddr *dst, struct sockaddr *gateway,
struct sockaddr *netmask, int flags, struct sockaddr *src,
struct rtentry **rtp, u_int rdomain)
{
struct rtentry *rt;
int error = 0;
u_int32_t *stat = NULL;
struct rt_addrinfo info;
struct ifaddr *ifa;
struct ifnet *ifp = NULL;
splsoftassert(IPL_SOFTNET);
/* verify the gateway is directly reachable */
if ((ifa = ifa_ifwithnet(gateway, rdomain)) == NULL) {
error = ENETUNREACH;
goto out;
}
ifp = ifa->ifa_ifp;
rt = rtalloc1(dst, 0, rdomain);
/*
* If the redirect isn't from our current router for this dst,
* it's either old or wrong. If it redirects us to ourselves,
* we have a routing loop, perhaps as a result of an interface
* going down recently.
*/
#define equal(a1, a2) \
((a1)->sa_len == (a2)->sa_len && \
bcmp((caddr_t)(a1), (caddr_t)(a2), (a1)->sa_len) == 0)
if (!(flags & RTF_DONE) && rt &&
(!equal(src, rt->rt_gateway) || rt->rt_ifa != ifa))
error = EINVAL;
else if (ifa_ifwithaddr(gateway, rdomain) != NULL)
error = EHOSTUNREACH;
if (error)
goto done;
/*
* Create a new entry if we just got back a wildcard entry
* or the lookup failed. This is necessary for hosts
* which use routing redirects generated by smart gateways
* to dynamically build the routing tables.
*/
if ((rt == NULL) || (rt_mask(rt) && rt_mask(rt)->sa_len < 2))
goto create;
/*
* Don't listen to the redirect if it's
* for a route to an interface.
*/
if (rt->rt_flags & RTF_GATEWAY) {
if (((rt->rt_flags & RTF_HOST) == 0) && (flags & RTF_HOST)) {
/*
* Changing from route to net => route to host.
* Create new route, rather than smashing route to net.
*/
create:
if (rt)
rtfree(rt);
flags |= RTF_GATEWAY | RTF_DYNAMIC;
bzero(&info, sizeof(info));
info.rti_info[RTAX_DST] = dst;
info.rti_info[RTAX_GATEWAY] = gateway;
info.rti_info[RTAX_NETMASK] = netmask;
info.rti_ifa = ifa;
info.rti_flags = flags;
rt = NULL;
error = rtrequest1(RTM_ADD, &info, RTP_DEFAULT, &rt,
rdomain);
if (rt != NULL)
flags = rt->rt_flags;
stat = &rtstat.rts_dynamic;
} else {
/*
* Smash the current notion of the gateway to
* this destination. Should check about netmask!!!
*/
rt->rt_flags |= RTF_MODIFIED;
flags |= RTF_MODIFIED;
stat = &rtstat.rts_newgateway;
rt_setgate(rt, rt_key(rt), gateway, rdomain);
}
} else
error = EHOSTUNREACH;
done:
if (rt) {
if (rtp && !error)
*rtp = rt;
else
rtfree(rt);
}
out:
if (error)
rtstat.rts_badredirect++;
else if (stat != NULL)
(*stat)++;
bzero((caddr_t)&info, sizeof(info));
info.rti_info[RTAX_DST] = dst;
info.rti_info[RTAX_GATEWAY] = gateway;
info.rti_info[RTAX_NETMASK] = netmask;
info.rti_info[RTAX_AUTHOR] = src;
rt_missmsg(RTM_REDIRECT, &info, flags, ifp, error, rdomain);
}
int
rtrequest1(int req, struct rt_addrinfo *info, struct rtentry **ret_nrt,
u_int tableid)
{
int s = splsoftnet(); int error = 0;
struct rtentry *rt, *crt;
struct radix_node *rn;
struct radix_node_head *rnh;
struct ifaddr *ifa;
struct sockaddr *ndst;
struct sockaddr_rtlabel *sa_rl;
#define senderr(x) { error = x ; goto bad; }
if ((rnh = rt_gettable(info->rti_info[RTAX_DST]->sa_family, tableid)) ==
NULL)
senderr(EAFNOSUPPORT);
if (info->rti_flags & RTF_HOST)
info->rti_info[RTAX_NETMASK] = NULL;
switch (req) {
case RTM_DELETE:
if ((rn = rnh->rnh_lookup(info->rti_info[RTAX_DST],
info->rti_info[RTAX_NETMASK], rnh)) == NULL)
senderr(ESRCH);
rt = (struct rtentry *)rn;
#ifndef SMALL_KERNEL
/*
* if we got multipath routes, we require users to specify
* a matching RTAX_GATEWAY.
*/
if (rn_mpath_capable(rnh)) {
rt = rt_mpath_matchgate(rt,
info->rti_info[RTAX_GATEWAY]);
rn = (struct radix_node *)rt;
if (!rt)
senderr(ESRCH);
}
#endif
if ((rn = rnh->rnh_deladdr(info->rti_info[RTAX_DST],
info->rti_info[RTAX_NETMASK], rnh, rn)) == NULL)
senderr(ESRCH);
rt = (struct rtentry *)rn;
/* clean up any cloned children */
if ((rt->rt_flags & RTF_CLONING) != 0)
rtflushclone(rnh, rt);
if (rn->rn_flags & (RNF_ACTIVE | RNF_ROOT))
panic ("rtrequest delete");
if (rt->rt_gwroute) {
rt = rt->rt_gwroute; RTFREE(rt);
(rt = (struct rtentry *)rn)->rt_gwroute = NULL;
}
if (rt->rt_parent) {
rt->rt_parent->rt_refcnt--;
rt->rt_parent = NULL;
}
#ifndef SMALL_KERNEL
if (rn_mpath_capable(rnh)) {
if ((rn = rnh->rnh_lookup(info->rti_info[RTAX_DST],
info->rti_info[RTAX_NETMASK], rnh)) != NULL &&
rn_mpath_next(rn) == NULL)
((struct rtentry *)rn)->rt_flags &= ~RTF_MPATH;
}
#endif
rt->rt_flags &= ~RTF_UP;
if ((ifa = rt->rt_ifa) && ifa->ifa_rtrequest)
ifa->ifa_rtrequest(RTM_DELETE, rt, info);
rttrash++;
if (ret_nrt)
*ret_nrt = rt;
else if (rt->rt_refcnt <= 0) {
rt->rt_refcnt++;
rtfree(rt);
}
break;
case RTM_RESOLVE:
if (ret_nrt == NULL || (rt = *ret_nrt) == NULL)
senderr(EINVAL);
if ((rt->rt_flags & RTF_CLONING) == 0)
senderr(EINVAL);
ifa = rt->rt_ifa;
info->rti_flags = rt->rt_flags & ~(RTF_CLONING | RTF_STATIC);
info->rti_flags |= RTF_CLONED;
info->rti_info[RTAX_GATEWAY] = rt->rt_gateway;
if ((info->rti_info[RTAX_NETMASK] = rt->rt_genmask) == NULL)
info->rti_flags |= RTF_HOST;
goto makeroute;
case RTM_ADD:
if (info->rti_ifa == 0 && (error = rt_getifa(info)))
senderr(error);
ifa = info->rti_ifa;
makeroute:
rt = pool_get(&rtentry_pool, PR_NOWAIT);
if (rt == NULL)
senderr(ENOBUFS);
Bzero(rt, sizeof(*rt));
rt->rt_flags = RTF_UP | info->rti_flags;
LIST_INIT(&rt->rt_timer);
if (rt_setgate(rt, info->rti_info[RTAX_DST],
info->rti_info[RTAX_GATEWAY], tableid)) {
pool_put(&rtentry_pool, rt);
senderr(ENOBUFS);
}
ndst = rt_key(rt);
if (info->rti_info[RTAX_NETMASK] != NULL) {
rt_maskedcopy(info->rti_info[RTAX_DST], ndst,
info->rti_info[RTAX_NETMASK]);
} else
Bcopy(info->rti_info[RTAX_DST], ndst,
info->rti_info[RTAX_DST]->sa_len);
#ifndef SMALL_KERNEL
/* do not permit exactly the same dst/mask/gw pair */
if (rn_mpath_capable(rnh) &&
rt_mpath_conflict(rnh, rt, info->rti_info[RTAX_NETMASK],
info->rti_flags & RTF_MPATH)) {
if (rt->rt_gwroute)
rtfree(rt->rt_gwroute);
Free(rt_key(rt));
pool_put(&rtentry_pool, rt);
senderr(EEXIST);
}
#endif
if (info->rti_info[RTAX_LABEL] != NULL) {
sa_rl = (struct sockaddr_rtlabel *)
info->rti_info[RTAX_LABEL];
rt->rt_labelid = rtlabel_name2id(sa_rl->sr_label);
}
ifa->ifa_refcnt++;
rt->rt_ifa = ifa;
rt->rt_ifp = ifa->ifa_ifp;
if (req == RTM_RESOLVE) {
/*
* Copy both metrics and a back pointer to the cloned
* route's parent.
*/
rt->rt_rmx = (*ret_nrt)->rt_rmx; /* copy metrics */
rt->rt_parent = *ret_nrt; /* Back ptr. to parent. */
rt->rt_parent->rt_refcnt++;
}
rn = rnh->rnh_addaddr((caddr_t)ndst,
(caddr_t)info->rti_info[RTAX_NETMASK], rnh, rt->rt_nodes);
if (rn == NULL && (crt = rtalloc1(ndst, 0, tableid)) != NULL) {
/* overwrite cloned route */
if ((crt->rt_flags & RTF_CLONED) != 0) {
rtdeletemsg(crt, tableid);
rn = rnh->rnh_addaddr((caddr_t)ndst,
(caddr_t)info->rti_info[RTAX_NETMASK],
rnh, rt->rt_nodes);
}
RTFREE(crt);
}
if (rn == 0) {
IFAFREE(ifa);
if ((rt->rt_flags & RTF_CLONED) != 0 && rt->rt_parent)
rtfree(rt->rt_parent);
if (rt->rt_gwroute)
rtfree(rt->rt_gwroute);
Free(rt_key(rt));
pool_put(&rtentry_pool, rt);
senderr(EEXIST);
}
#ifndef SMALL_KERNEL
if (rn_mpath_capable(rnh) &&
(rn = rnh->rnh_lookup(info->rti_info[RTAX_DST],
info->rti_info[RTAX_NETMASK], rnh)) != NULL) {
if (rn_mpath_next(rn) == NULL)
((struct rtentry *)rn)->rt_flags &= ~RTF_MPATH;
else
((struct rtentry *)rn)->rt_flags |= RTF_MPATH;
}
#endif
if (ifa->ifa_rtrequest)
ifa->ifa_rtrequest(req, rt, info);
if (ret_nrt) {
*ret_nrt = rt;
rt->rt_refcnt++;
}
if ((rt->rt_flags & RTF_CLONING) != 0) {
/* clean up any cloned children */
rtflushclone(rnh, rt);
}
if_group_routechange(info->rti_info[RTAX_DST],
info->rti_info[RTAX_NETMASK]);
break;
}
bad:
splx(s);
return (error);
}
/*
* Input a Neighbor Solicitation Message.
*
* Based on RFC 2461
* Based on RFC 2462 (duplicate address detection)
*/
void
nd6_ns_input(struct mbuf *m, int off, int icmp6len)
{
struct ifnet *ifp = m->m_pkthdr.rcvif;
struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *);
struct nd_neighbor_solicit *nd_ns;
struct in6_addr saddr6 = ip6->ip6_src;
struct in6_addr daddr6 = ip6->ip6_dst;
struct in6_addr taddr6;
struct in6_addr myaddr6;
char *lladdr = NULL;
struct ifaddr *ifa = NULL;
int lladdrlen = 0;
int anycast = 0, proxy = 0, tentative = 0;
int tlladdr;
int rflag;
union nd_opts ndopts;
struct sockaddr_dl proxydl;
char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN];
rflag = (V_ip6_forwarding) ? ND_NA_FLAG_ROUTER : 0;
if (ND_IFINFO(ifp)->flags & ND6_IFF_ACCEPT_RTADV && V_ip6_norbit_raif)
rflag = 0;
#ifndef PULLDOWN_TEST
IP6_EXTHDR_CHECK(m, off, icmp6len,);
nd_ns = (struct nd_neighbor_solicit *)((caddr_t)ip6 + off);
#else
IP6_EXTHDR_GET(nd_ns, struct nd_neighbor_solicit *, m, off, icmp6len);
if (nd_ns == NULL) {
ICMP6STAT_INC(icp6s_tooshort);
return;
}
#endif
ip6 = mtod(m, struct ip6_hdr *); /* adjust pointer for safety */
taddr6 = nd_ns->nd_ns_target;
if (in6_setscope(&taddr6, ifp, NULL) != 0)
goto bad;
if (ip6->ip6_hlim != 255) {
nd6log((LOG_ERR,
"nd6_ns_input: invalid hlim (%d) from %s to %s on %s\n",
ip6->ip6_hlim, ip6_sprintf(ip6bufs, &ip6->ip6_src),
ip6_sprintf(ip6bufd, &ip6->ip6_dst), if_name(ifp)));
goto bad;
}
if (IN6_IS_ADDR_UNSPECIFIED(&saddr6)) {
/* dst has to be a solicited node multicast address. */
if (daddr6.s6_addr16[0] == IPV6_ADDR_INT16_MLL &&
/* don't check ifindex portion */
daddr6.s6_addr32[1] == 0 &&
daddr6.s6_addr32[2] == IPV6_ADDR_INT32_ONE &&
daddr6.s6_addr8[12] == 0xff) {
; /* good */
} else {
nd6log((LOG_INFO, "nd6_ns_input: bad DAD packet "
"(wrong ip6 dst)\n"));
goto bad;
}
} else if (!V_nd6_onlink_ns_rfc4861) {
struct sockaddr_in6 src_sa6;
/*
* According to recent IETF discussions, it is not a good idea
* to accept a NS from an address which would not be deemed
* to be a neighbor otherwise. This point is expected to be
* clarified in future revisions of the specification.
*/
bzero(&src_sa6, sizeof(src_sa6));
src_sa6.sin6_family = AF_INET6;
src_sa6.sin6_len = sizeof(src_sa6);
src_sa6.sin6_addr = saddr6;
if (nd6_is_addr_neighbor(&src_sa6, ifp) == 0) {
nd6log((LOG_INFO, "nd6_ns_input: "
"NS packet from non-neighbor\n"));
goto bad;
}
}
if (IN6_IS_ADDR_MULTICAST(&taddr6)) {
nd6log((LOG_INFO, "nd6_ns_input: bad NS target (multicast)\n"));
goto bad;
}
icmp6len -= sizeof(*nd_ns);
nd6_option_init(nd_ns + 1, icmp6len, &ndopts);
if (nd6_options(&ndopts) < 0) {
nd6log((LOG_INFO,
"nd6_ns_input: invalid ND option, ignored\n"));
/* nd6_options have incremented stats */
goto freeit;
}
if (ndopts.nd_opts_src_lladdr) {
lladdr = (char *)(ndopts.nd_opts_src_lladdr + 1);
lladdrlen = ndopts.nd_opts_src_lladdr->nd_opt_len << 3;
}
if (IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_src) && lladdr) {
nd6log((LOG_INFO, "nd6_ns_input: bad DAD packet "
"(link-layer address option)\n"));
goto bad;
}
/*
* Attaching target link-layer address to the NA?
* (RFC 2461 7.2.4)
*
* NS IP dst is unicast/anycast MUST NOT add
* NS IP dst is solicited-node multicast MUST add
*
* In implementation, we add target link-layer address by default.
* We do not add one in MUST NOT cases.
*/
if (!IN6_IS_ADDR_MULTICAST(&daddr6))
tlladdr = 0;
else
tlladdr = 1;
/*
* Target address (taddr6) must be either:
* (1) Valid unicast/anycast address for my receiving interface,
* (2) Unicast address for which I'm offering proxy service, or
* (3) "tentative" address on which DAD is being performed.
*/
/* (1) and (3) check. */
if (ifp->if_carp)
ifa = (*carp_iamatch6_p)(ifp, &taddr6);
if (ifa == NULL)
ifa = (struct ifaddr *)in6ifa_ifpwithaddr(ifp, &taddr6);
/* (2) check. */
if (ifa == NULL) {
struct rtentry *rt;
struct sockaddr_in6 tsin6;
int need_proxy;
#ifdef RADIX_MPATH
struct route_in6 ro;
#endif
bzero(&tsin6, sizeof tsin6);
tsin6.sin6_len = sizeof(struct sockaddr_in6);
tsin6.sin6_family = AF_INET6;
tsin6.sin6_addr = taddr6;
#ifdef RADIX_MPATH
bzero(&ro, sizeof(ro));
ro.ro_dst = tsin6;
rtalloc_mpath((struct route *)&ro, RTF_ANNOUNCE);
rt = ro.ro_rt;
#else
rt = rtalloc1((struct sockaddr *)&tsin6, 0, 0);
#endif
need_proxy = (rt && (rt->rt_flags & RTF_ANNOUNCE) != 0 &&
rt->rt_gateway->sa_family == AF_LINK);
if (rt != NULL) {
/*
* Make a copy while we can be sure that rt_gateway
* is still stable before unlocking to avoid lock
* order problems. proxydl will only be used if
* proxy will be set in the next block.
*/
if (need_proxy)
proxydl = *SDL(rt->rt_gateway);
RTFREE_LOCKED(rt);
}
if (need_proxy) {
/*
* proxy NDP for single entry
*/
ifa = (struct ifaddr *)in6ifa_ifpforlinklocal(ifp,
IN6_IFF_NOTREADY|IN6_IFF_ANYCAST);
if (ifa)
proxy = 1;
}
}
if (ifa == NULL) {
/*
* We've got an NS packet, and we don't have that adddress
* assigned for us. We MUST silently ignore it.
* See RFC2461 7.2.3.
*/
goto freeit;
}
myaddr6 = *IFA_IN6(ifa);
anycast = ((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_ANYCAST;
tentative = ((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_TENTATIVE;
if (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_DUPLICATED)
goto freeit;
if (lladdr && ((ifp->if_addrlen + 2 + 7) & ~7) != lladdrlen) {
nd6log((LOG_INFO, "nd6_ns_input: lladdrlen mismatch for %s "
"(if %d, NS packet %d)\n",
ip6_sprintf(ip6bufs, &taddr6),
ifp->if_addrlen, lladdrlen - 2));
goto bad;
}
if (IN6_ARE_ADDR_EQUAL(&myaddr6, &saddr6)) {
nd6log((LOG_INFO, "nd6_ns_input: duplicate IP6 address %s\n",
ip6_sprintf(ip6bufs, &saddr6)));
goto freeit;
}
/*
* We have neighbor solicitation packet, with target address equals to
* one of my tentative address.
*
* src addr how to process?
* --- ---
* multicast of course, invalid (rejected in ip6_input)
* unicast somebody is doing address resolution -> ignore
* unspec dup address detection
*
* The processing is defined in RFC 2462.
*/
if (tentative) {
/*
* If source address is unspecified address, it is for
* duplicate address detection.
*
* If not, the packet is for addess resolution;
* silently ignore it.
*/
if (IN6_IS_ADDR_UNSPECIFIED(&saddr6))
nd6_dad_ns_input(ifa);
goto freeit;
}
/*
* If the source address is unspecified address, entries must not
* be created or updated.
* It looks that sender is performing DAD. Output NA toward
* all-node multicast address, to tell the sender that I'm using
* the address.
* S bit ("solicited") must be zero.
*/
if (IN6_IS_ADDR_UNSPECIFIED(&saddr6)) {
struct in6_addr in6_all;
in6_all = in6addr_linklocal_allnodes;
if (in6_setscope(&in6_all, ifp, NULL) != 0)
goto bad;
nd6_na_output(ifp, &in6_all, &taddr6,
((anycast || proxy || !tlladdr) ? 0 : ND_NA_FLAG_OVERRIDE) |
rflag, tlladdr, proxy ? (struct sockaddr *)&proxydl : NULL);
goto freeit;
}
nd6_cache_lladdr(ifp, &saddr6, lladdr, lladdrlen,
ND_NEIGHBOR_SOLICIT, 0);
nd6_na_output(ifp, &saddr6, &taddr6,
((anycast || proxy || !tlladdr) ? 0 : ND_NA_FLAG_OVERRIDE) |
rflag | ND_NA_FLAG_SOLICITED, tlladdr,
proxy ? (struct sockaddr *)&proxydl : NULL);
freeit:
if (ifa != NULL)
ifa_free(ifa);
m_freem(m);
return;
bad:
nd6log((LOG_ERR, "nd6_ns_input: src=%s\n",
ip6_sprintf(ip6bufs, &saddr6)));
nd6log((LOG_ERR, "nd6_ns_input: dst=%s\n",
ip6_sprintf(ip6bufs, &daddr6)));
nd6log((LOG_ERR, "nd6_ns_input: tgt=%s\n",
ip6_sprintf(ip6bufs, &taddr6)));
ICMP6STAT_INC(icp6s_badns);
if (ifa != NULL)
ifa_free(ifa);
m_freem(m);
}
/*
* llcintr() handles all LLC frames (except ISO CLNS ones for the time being)
* and tries to pass them on to the appropriate network layer entity.
*/
void
llcintr()
{
register struct mbuf *m;
register int i;
register int frame_kind;
register u_char cmdrsp;
struct llc_linkcb *linkp;
struct npaidbentry *sapinfo = NULL;
struct sdl_hdr *sdlhdr;
struct llc *frame;
long expected_len;
struct ifnet *ifp;
struct rtentry *llrt;
struct rtentry *nlrt;
for (;;) {
i = splimp();
IF_DEQUEUE(&llcintrq, m);
splx(i);
if (m == 0)
break;
#ifdef DIAGNOSTIC
if ((m->m_flags & M_PKTHDR) == 0)
panic("llcintr no HDR");
#endif
/*
* Get ifp this packet was received on
*/
ifp = m->m_pkthdr.rcvif;
sdlhdr = mtod(m, struct sdl_hdr *);
/*
* [Copied from net/ip_input.c]
*
* Check that the amount of data in the buffers is
* at least as much as the LLC header tells us.
* Trim mbufs if longer than expected.
* Drop packets if shorter than we think they are.
*
* Layout of mbuf chain at this point:
*
* +-------------------------------+----+ -\
* | sockaddr_dl src - sdlhdr_src | 20 | \
* +-------------------------------+----+ |
* | sockaddr_dl dst - sdlhdr_dst | 20 | > sizeof(struct sdl_hdr) == 44
* +-------------------------------+----+ |
* | LLC frame len - sdlhdr_len | 04 | /
* +-------------------------------+----+ -/
* /
* | m_next
* \
* +----------------------------+----+ -\
* | llc DSAP | 01 | \
* +----------------------------+----+ |
* | llc SSAP | 01 | |
* +----------------------------+----+ > sdlhdr_len
* | llc control | 01 | |
* +----------------------------+----+ |
* | ... | | /
* -/
*
* Thus the we expect to have exactly
* (sdlhdr->sdlhdr_len+sizeof(struct sdl_hdr)) in the mbuf chain
*/
expected_len = sdlhdr->sdlhdr_len + sizeof(struct sdl_hdr);
if (m->m_pkthdr.len < expected_len) {
m_freem(m);
continue;
}
if (m->m_pkthdr.len > expected_len) {
if (m->m_len == m->m_pkthdr.len) {
m->m_len = expected_len;
m->m_pkthdr.len = expected_len;
} else
m_adj(m, expected_len - m->m_pkthdr.len);
}
/*
* Get llc header
*/
if (m->m_len > sizeof(struct sdl_hdr))
frame = mtod((struct mbuf *)((struct sdl_hdr*)(m+1)),
struct llc *);
else frame = mtod(m->m_next, struct llc *);
if (frame == (struct llc *) NULL)
panic("llcintr no llc header");
/*
* Now check for bogus I/S frame, i.e. those with a control
* field telling us they're an I/S frame yet their length
* is less than the established I/S frame length (DSAP + SSAP +
* control + N(R)&P/F = 4) --- we drop those suckers
*/
if (((frame->llc_control & 0x03) != 0x03)
&& ((expected_len - sizeof(struct sdl_hdr)) < LLC_ISFRAMELEN)) {
m_freem(m);
printf("llc: hurz error\n");
continue;
}
/*
* Get link control block for the addressed link connection.
* If there is none we take care of it later on.
*/
cmdrsp = (frame->llc_ssap & 0x01);
frame->llc_ssap &= ~0x01;
llrt = rtalloc1((struct sockaddr *)&sdlhdr->sdlhdr_src, 0);
if (llrt)
llrt->rt_refcnt--;
#ifdef notyet
else
llrt = npaidb_enter(&sdlhdr->sdlhdr_src, 0, 0, 0);
#endif /* notyet */
else {
/*
* MPLS Label Switch Engine
*/
static int
mpls_lse(struct mbuf *m)
{
struct sockaddr_mpls dst;
union mpls_shim tshim, *htag;
struct rtentry *rt = NULL;
int error = ENOBUFS;
uint psize = sizeof(struct sockaddr_mpls);
bool push_back_alert = false;
if (m->m_len < sizeof(union mpls_shim) &&
(m = m_pullup(m, sizeof(union mpls_shim))) == NULL)
goto done;
dst.smpls_len = sizeof(struct sockaddr_mpls);
dst.smpls_family = AF_MPLS;
dst.smpls_addr.s_addr = ntohl(mtod(m, union mpls_shim *)->s_addr);
/* Check if we're accepting MPLS Frames */
error = EINVAL;
if (!mpls_accept)
goto done;
/* TTL decrement */
if ((m = mpls_ttl_dec(m)) == NULL)
goto done;
/* RFC 4182 */
if (mpls_rfc4182 != 0)
while((dst.smpls_addr.shim.label == MPLS_LABEL_IPV4NULL ||
dst.smpls_addr.shim.label == MPLS_LABEL_IPV6NULL) &&
__predict_false(dst.smpls_addr.shim.bos == 0))
TRIM_LABEL;
/* RFC 3032 Section 2.1 Page 4 */
if (__predict_false(dst.smpls_addr.shim.label == MPLS_LABEL_RTALERT) &&
dst.smpls_addr.shim.bos == 0) {
TRIM_LABEL;
push_back_alert = true;
}
if (dst.smpls_addr.shim.label <= MPLS_LABEL_RESMAX) {
/* Don't swap reserved labels */
switch (dst.smpls_addr.shim.label) {
#ifdef INET
case MPLS_LABEL_IPV4NULL:
/* Pop shim and push mbuf to IP stack */
if (dst.smpls_addr.shim.bos)
error = mpls_unlabel_inet(m);
break;
#endif
#ifdef INET6
case MPLS_LABEL_IPV6NULL:
/* Pop shim and push mbuf to IPv6 stack */
if (dst.smpls_addr.shim.bos)
error = mpls_unlabel_inet6(m);
break;
#endif
case MPLS_LABEL_RTALERT: /* Yeah, I'm all alerted */
case MPLS_LABEL_IMPLNULL: /* This is logical only */
default: /* Rest are not allowed */
break;
}
goto done;
}
/* Check if we should do MPLS forwarding */
error = EHOSTUNREACH;
if (!mpls_forwarding)
goto done;
/* Get a route to dst */
dst.smpls_addr.shim.ttl =
dst.smpls_addr.shim.bos =
dst.smpls_addr.shim.exp = 0;
dst.smpls_addr.s_addr = htonl(dst.smpls_addr.s_addr);
if ((rt = rtalloc1((const struct sockaddr*)&dst, 1)) == NULL)
goto done;
/* MPLS packet with no MPLS tagged route ? */
if ((rt->rt_flags & RTF_GATEWAY) == 0 ||
rt_gettag(rt) == NULL ||
rt_gettag(rt)->sa_family != AF_MPLS)
goto done;
tshim.s_addr = MPLS_GETSADDR(rt);
/* Swap labels */
if ((m->m_len < sizeof(union mpls_shim)) &&
(m = m_pullup(m, sizeof(union mpls_shim))) == 0) {
error = ENOBUFS;
goto done;
}
/* Replace only the label */
htag = mtod(m, union mpls_shim *);
htag->s_addr = ntohl(htag->s_addr);
htag->shim.label = tshim.shim.label;
htag->s_addr = htonl(htag->s_addr);
/* check if there is anything more to prepend */
htag = &((struct sockaddr_mpls*)rt_gettag(rt))->smpls_addr;
while (psize <= rt_gettag(rt)->sa_len - sizeof(tshim)) {
htag++;
memset(&tshim, 0, sizeof(tshim));
tshim.s_addr = ntohl(htag->s_addr);
tshim.shim.bos = tshim.shim.exp = 0;
tshim.shim.ttl = mpls_defttl;
if (tshim.shim.label != MPLS_LABEL_IMPLNULL &&
((m = mpls_prepend_shim(m, &tshim)) == NULL))
return ENOBUFS;
psize += sizeof(tshim);
}
if (__predict_false(push_back_alert == true)) {
/* re-add the router alert label */
memset(&tshim, 0, sizeof(tshim));
tshim.s_addr = MPLS_LABEL_RTALERT;
tshim.shim.bos = tshim.shim.exp = 0;
tshim.shim.ttl = mpls_defttl;
if ((m = mpls_prepend_shim(m, &tshim)) == NULL)
return ENOBUFS;
}
error = mpls_send_frame(m, rt->rt_ifp, rt);
done:
if (error != 0 && m != NULL)
m_freem(m);
if (rt != NULL)
RTFREE(rt);
return error;
}
/*
* Ethernet output routine.
* Encapsulate a packet of type family for the local net.
* Assumes that ifp is actually pointer to arpcom structure.
*/
int
ether_output(struct ifnet *ifp0, struct mbuf *m0, struct sockaddr *dst,
struct rtentry *rt0)
{
u_int16_t etype;
int s, len, error = 0;
u_char edst[ETHER_ADDR_LEN];
u_char *esrc;
struct mbuf *m = m0;
struct rtentry *rt;
struct mbuf *mcopy = NULL;
struct ether_header *eh;
struct arpcom *ac = (struct arpcom *)ifp0;
short mflags;
struct ifnet *ifp = ifp0;
#ifdef DIAGNOSTIC
if (ifp->if_rdomain != rtable_l2(m->m_pkthdr.ph_rtableid)) {
printf("%s: trying to send packet on wrong domain. "
"if %d vs. mbuf %d, AF %d\n", ifp->if_xname,
ifp->if_rdomain, rtable_l2(m->m_pkthdr.ph_rtableid),
dst->sa_family);
}
#endif
#if NTRUNK > 0
/* restrict transmission on trunk members to bpf only */
if (ifp->if_type == IFT_IEEE8023ADLAG &&
(m_tag_find(m, PACKET_TAG_DLT, NULL) == NULL))
senderr(EBUSY);
#endif
#if NCARP > 0
if (ifp->if_type == IFT_CARP) {
ifp = ifp->if_carpdev;
ac = (struct arpcom *)ifp;
if ((ifp0->if_flags & (IFF_UP|IFF_RUNNING)) !=
(IFF_UP|IFF_RUNNING))
senderr(ENETDOWN);
}
#endif /* NCARP > 0 */
if ((ifp->if_flags & (IFF_UP|IFF_RUNNING)) != (IFF_UP|IFF_RUNNING))
senderr(ENETDOWN);
if ((rt = rt0) != NULL) {
if ((rt->rt_flags & RTF_UP) == 0) {
if ((rt0 = rt = rtalloc1(dst, RT_REPORT,
m->m_pkthdr.ph_rtableid)) != NULL)
rt->rt_refcnt--;
else
senderr(EHOSTUNREACH);
}
if (rt->rt_flags & RTF_GATEWAY) {
if (rt->rt_gwroute == NULL)
goto lookup;
if (((rt = rt->rt_gwroute)->rt_flags & RTF_UP) == 0) {
rtfree(rt);
rt = rt0;
lookup:
rt->rt_gwroute = rtalloc1(rt->rt_gateway,
RT_REPORT, ifp->if_rdomain);
if ((rt = rt->rt_gwroute) == NULL)
senderr(EHOSTUNREACH);
}
}
if (rt->rt_flags & RTF_REJECT)
if (rt->rt_rmx.rmx_expire == 0 ||
time_second < rt->rt_rmx.rmx_expire)
senderr(rt == rt0 ? EHOSTDOWN : EHOSTUNREACH);
}
esrc = ac->ac_enaddr;
switch (dst->sa_family) {
#ifdef INET
case AF_INET:
if (!arpresolve(ac, rt, m, dst, edst))
return (0); /* if not yet resolved */
/* If broadcasting on a simplex interface, loopback a copy */
if ((m->m_flags & M_BCAST) && (ifp->if_flags & IFF_SIMPLEX) &&
!m->m_pkthdr.pf.routed)
mcopy = m_copy(m, 0, (int)M_COPYALL);
etype = htons(ETHERTYPE_IP);
break;
#endif
#ifdef INET6
case AF_INET6:
if (!nd6_storelladdr(ifp, rt, m, dst, (u_char *)edst))
return (0); /* it must be impossible, but... */
etype = htons(ETHERTYPE_IPV6);
break;
#endif
#ifdef MPLS
case AF_MPLS:
if (rt)
dst = rt_key(rt);
else
senderr(EHOSTUNREACH);
if (!ISSET(ifp->if_xflags, IFXF_MPLS))
senderr(ENETUNREACH);
switch (dst->sa_family) {
case AF_LINK:
if (((struct sockaddr_dl *)dst)->sdl_alen <
sizeof(edst))
senderr(EHOSTUNREACH);
memcpy(edst, LLADDR((struct sockaddr_dl *)dst),
sizeof(edst));
break;
case AF_INET:
if (!arpresolve(ac, rt, m, dst, edst))
return (0); /* if not yet resolved */
break;
default:
senderr(EHOSTUNREACH);
}
/* XXX handling for simplex devices in case of M/BCAST ?? */
if (m->m_flags & (M_BCAST | M_MCAST))
etype = htons(ETHERTYPE_MPLS_MCAST);
else
etype = htons(ETHERTYPE_MPLS);
break;
#endif /* MPLS */
case pseudo_AF_HDRCMPLT:
eh = (struct ether_header *)dst->sa_data;
esrc = eh->ether_shost;
/* FALLTHROUGH */
case AF_UNSPEC:
eh = (struct ether_header *)dst->sa_data;
memcpy(edst, eh->ether_dhost, sizeof(edst));
/* AF_UNSPEC doesn't swap the byte order of the ether_type. */
etype = eh->ether_type;
break;
default:
printf("%s: can't handle af%d\n", ifp->if_xname,
dst->sa_family);
senderr(EAFNOSUPPORT);
}
/* XXX Should we feed-back an unencrypted IPsec packet ? */
if (mcopy)
(void) looutput(ifp, mcopy, dst, rt);
#if NCARP > 0
if (ifp0 != ifp && ifp0->if_type == IFT_CARP)
esrc = carp_get_srclladdr(ifp0, esrc);
#endif
if (ether_addheader(&m, ifp, etype, esrc, edst) == -1)
senderr(ENOBUFS);
#if NBRIDGE > 0
/*
* Interfaces that are bridgeports need special handling for output.
*/
if (ifp->if_bridgeport) {
struct m_tag *mtag;
/*
* Check if this packet has already been sent out through
* this bridgeport, in which case we simply send it out
* without further bridge processing.
*/
for (mtag = m_tag_find(m, PACKET_TAG_BRIDGE, NULL); mtag;
mtag = m_tag_find(m, PACKET_TAG_BRIDGE, mtag)) {
#ifdef DEBUG
/* Check that the information is there */
if (mtag->m_tag_len != sizeof(caddr_t)) {
error = EINVAL;
goto bad;
}
#endif
if (!memcmp(&ifp->if_bridgeport, mtag + 1,
sizeof(caddr_t)))
break;
}
if (mtag == NULL) {
/* Attach a tag so we can detect loops */
mtag = m_tag_get(PACKET_TAG_BRIDGE, sizeof(caddr_t),
M_NOWAIT);
if (mtag == NULL) {
error = ENOBUFS;
goto bad;
}
memcpy(mtag + 1, &ifp->if_bridgeport, sizeof(caddr_t));
m_tag_prepend(m, mtag);
error = bridge_output(ifp, m, NULL, NULL);
return (error);
}
}
#endif
mflags = m->m_flags;
len = m->m_pkthdr.len;
s = splnet();
/*
* Queue message on interface, and start output if interface
* not yet active.
*/
IFQ_ENQUEUE(&ifp->if_snd, m, NULL, error);
if (error) {
/* mbuf is already freed */
splx(s);
return (error);
}
ifp->if_obytes += len;
#if NCARP > 0
if (ifp != ifp0)
ifp0->if_obytes += len;
#endif /* NCARP > 0 */
if (mflags & M_MCAST)
ifp->if_omcasts++;
if_start(ifp);
splx(s);
return (error);
bad:
if (m)
m_freem(m);
return (error);
}
/*
* Input a Neighbor Solicitation Message.
*
* Based on RFC 2461
* Based on RFC 2462 (duplicate address detection)
*/
void
nd6_ns_input(struct mbuf *m, int off, int icmp6len)
{
struct ifnet *ifp = m->m_pkthdr.rcvif;
struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *);
struct nd_neighbor_solicit *nd_ns;
struct in6_addr saddr6 = ip6->ip6_src;
struct in6_addr daddr6 = ip6->ip6_dst;
struct in6_addr taddr6;
struct in6_addr myaddr6;
char *lladdr = NULL;
struct ifaddr *ifa;
int lladdrlen = 0;
int anycast = 0, proxy = 0, tentative = 0;
int router = ip6_forwarding;
int tlladdr;
union nd_opts ndopts;
const struct sockaddr_dl *proxydl = NULL;
IP6_EXTHDR_GET(nd_ns, struct nd_neighbor_solicit *, m, off, icmp6len);
if (nd_ns == NULL) {
ICMP6_STATINC(ICMP6_STAT_TOOSHORT);
return;
}
ip6 = mtod(m, struct ip6_hdr *); /* adjust pointer for safety */
taddr6 = nd_ns->nd_ns_target;
if (in6_setscope(&taddr6, ifp, NULL) != 0)
goto bad;
if (ip6->ip6_hlim != 255) {
nd6log((LOG_ERR,
"nd6_ns_input: invalid hlim (%d) from %s to %s on %s\n",
ip6->ip6_hlim, ip6_sprintf(&ip6->ip6_src),
ip6_sprintf(&ip6->ip6_dst), if_name(ifp)));
goto bad;
}
if (IN6_IS_ADDR_UNSPECIFIED(&saddr6)) {
/* dst has to be a solicited node multicast address. */
/* don't check ifindex portion */
if (daddr6.s6_addr16[0] == IPV6_ADDR_INT16_MLL &&
daddr6.s6_addr32[1] == 0 &&
daddr6.s6_addr32[2] == IPV6_ADDR_INT32_ONE &&
daddr6.s6_addr8[12] == 0xff) {
; /* good */
} else {
nd6log((LOG_INFO, "nd6_ns_input: bad DAD packet "
"(wrong ip6 dst)\n"));
goto bad;
}
} else {
struct sockaddr_in6 ssin6;
/*
* Make sure the source address is from a neighbor's address.
*/
sockaddr_in6_init(&ssin6, &saddr6, 0, 0, 0);
if (nd6_is_addr_neighbor(&ssin6, ifp) == 0) {
nd6log((LOG_INFO, "nd6_ns_input: "
"NS packet from non-neighbor\n"));
goto bad;
}
}
if (IN6_IS_ADDR_MULTICAST(&taddr6)) {
nd6log((LOG_INFO, "nd6_ns_input: bad NS target (multicast)\n"));
goto bad;
}
icmp6len -= sizeof(*nd_ns);
nd6_option_init(nd_ns + 1, icmp6len, &ndopts);
if (nd6_options(&ndopts) < 0) {
nd6log((LOG_INFO,
"nd6_ns_input: invalid ND option, ignored\n"));
/* nd6_options have incremented stats */
goto freeit;
}
if (ndopts.nd_opts_src_lladdr) {
lladdr = (char *)(ndopts.nd_opts_src_lladdr + 1);
lladdrlen = ndopts.nd_opts_src_lladdr->nd_opt_len << 3;
}
if (IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_src) && lladdr) {
nd6log((LOG_INFO, "nd6_ns_input: bad DAD packet "
"(link-layer address option)\n"));
goto bad;
}
/*
* Attaching target link-layer address to the NA?
* (RFC 2461 7.2.4)
*
* NS IP dst is multicast MUST add
* Otherwise MAY be omitted
*
* In this implementation, we omit the target link-layer address
* in the "MAY" case.
*/
#if 0 /* too much! */
ifa = (struct ifaddr *)in6ifa_ifpwithaddr(ifp, &daddr6);
if (ifa && (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_ANYCAST))
tlladdr = 0;
else
#endif
if (!IN6_IS_ADDR_MULTICAST(&daddr6))
tlladdr = 0;
else
tlladdr = 1;
/*
* Target address (taddr6) must be either:
* (1) Valid unicast/anycast address for my receiving interface,
* (2) Unicast address for which I'm offering proxy service, or
* (3) "tentative" address on which DAD is being performed.
*/
/* (1) and (3) check. */
#if NCARP > 0
if (ifp->if_carp && ifp->if_type != IFT_CARP)
ifa = carp_iamatch6(ifp->if_carp, &taddr6);
else
ifa = NULL;
if (!ifa)
ifa = (struct ifaddr *)in6ifa_ifpwithaddr(ifp, &taddr6);
#else
ifa = (struct ifaddr *)in6ifa_ifpwithaddr(ifp, &taddr6);
#endif
/* (2) check. */
if (ifa == NULL) {
struct rtentry *rt;
struct sockaddr_in6 tsin6;
sockaddr_in6_init(&tsin6, &taddr6, 0, 0, 0);
rt = rtalloc1((struct sockaddr *)&tsin6, 0);
if (rt && (rt->rt_flags & RTF_ANNOUNCE) != 0 &&
rt->rt_gateway->sa_family == AF_LINK) {
/*
* proxy NDP for single entry
*/
ifa = (struct ifaddr *)in6ifa_ifpforlinklocal(ifp,
IN6_IFF_NOTREADY|IN6_IFF_ANYCAST);
if (ifa) {
proxy = 1;
proxydl = satocsdl(rt->rt_gateway);
router = 0; /* XXX */
}
}
if (rt)
rtfree(rt);
}
if (ifa == NULL) {
/*
* We've got an NS packet, and we don't have that address
* assigned for us. We MUST silently ignore it.
* See RFC2461 7.2.3.
*/
goto freeit;
}
myaddr6 = *IFA_IN6(ifa);
anycast = ((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_ANYCAST;
tentative = ((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_TENTATIVE;
if (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_DUPLICATED)
goto freeit;
if (lladdr && ((ifp->if_addrlen + 2 + 7) & ~7) != lladdrlen) {
nd6log((LOG_INFO, "nd6_ns_input: lladdrlen mismatch for %s "
"(if %d, NS packet %d)\n",
ip6_sprintf(&taddr6), ifp->if_addrlen, lladdrlen - 2));
goto bad;
}
if (IN6_ARE_ADDR_EQUAL(&myaddr6, &saddr6)) {
nd6log((LOG_INFO, "nd6_ns_input: duplicate IP6 address %s\n",
ip6_sprintf(&saddr6)));
goto freeit;
}
/*
* We have neighbor solicitation packet, with target address equals to
* one of my tentative address.
*
* src addr how to process?
* --- ---
* multicast of course, invalid (rejected in ip6_input)
* unicast somebody is doing address resolution -> ignore
* unspec dup address detection
*
* The processing is defined in RFC 2462.
*/
if (tentative) {
/*
* If source address is unspecified address, it is for
* duplicate address detection.
*
* If not, the packet is for addess resolution;
* silently ignore it.
*/
if (IN6_IS_ADDR_UNSPECIFIED(&saddr6))
nd6_dad_ns_input(ifa);
goto freeit;
}
/*
* If the source address is unspecified address, entries must not
* be created or updated.
* It looks that sender is performing DAD. Output NA toward
* all-node multicast address, to tell the sender that I'm using
* the address.
* S bit ("solicited") must be zero.
*/
if (IN6_IS_ADDR_UNSPECIFIED(&saddr6)) {
struct in6_addr in6_all;
in6_all = in6addr_linklocal_allnodes;
if (in6_setscope(&in6_all, ifp, NULL) != 0)
goto bad;
nd6_na_output(ifp, &in6_all, &taddr6,
((anycast || proxy || !tlladdr) ? 0 : ND_NA_FLAG_OVERRIDE) |
(ip6_forwarding ? ND_NA_FLAG_ROUTER : 0),
tlladdr, (const struct sockaddr *)proxydl);
goto freeit;
}
nd6_cache_lladdr(ifp, &saddr6, lladdr, lladdrlen, ND_NEIGHBOR_SOLICIT, 0);
nd6_na_output(ifp, &saddr6, &taddr6,
((anycast || proxy || !tlladdr) ? 0 : ND_NA_FLAG_OVERRIDE) |
(router ? ND_NA_FLAG_ROUTER : 0) | ND_NA_FLAG_SOLICITED,
tlladdr, (const struct sockaddr *)proxydl);
freeit:
m_freem(m);
return;
bad:
nd6log((LOG_ERR, "nd6_ns_input: src=%s\n", ip6_sprintf(&saddr6)));
nd6log((LOG_ERR, "nd6_ns_input: dst=%s\n", ip6_sprintf(&daddr6)));
nd6log((LOG_ERR, "nd6_ns_input: tgt=%s\n", ip6_sprintf(&taddr6)));
ICMP6_STATINC(ICMP6_STAT_BADNS);
m_freem(m);
}
/*
* prepend shim and deliver
*/
static int
mpls_output(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *dst, struct rtentry *rt)
{
union mpls_shim mh, *pms;
struct rtentry *rt1;
int err;
uint psize = sizeof(struct sockaddr_mpls);
if ((ifp->if_flags & (IFF_UP|IFF_RUNNING)) != (IFF_UP|IFF_RUNNING)) {
m_freem(m);
return ENETDOWN;
}
if (rt_gettag(rt) == NULL || rt_gettag(rt)->sa_family != AF_MPLS) {
m_freem(m);
return EINVAL;
}
bpf_mtap_af(ifp, dst->sa_family, m);
memset(&mh, 0, sizeof(mh));
mh.s_addr = MPLS_GETSADDR(rt);
mh.shim.bos = 1;
mh.shim.exp = 0;
mh.shim.ttl = mpls_defttl;
pms = &((struct sockaddr_mpls*)rt_gettag(rt))->smpls_addr;
while (psize <= rt_gettag(rt)->sa_len - sizeof(mh)) {
pms++;
if (mh.shim.label != MPLS_LABEL_IMPLNULL &&
((m = mpls_prepend_shim(m, &mh)) == NULL))
return ENOBUFS;
memset(&mh, 0, sizeof(mh));
mh.s_addr = ntohl(pms->s_addr);
mh.shim.bos = mh.shim.exp = 0;
mh.shim.ttl = mpls_defttl;
psize += sizeof(mh);
}
switch(dst->sa_family) {
#ifdef INET
case AF_INET:
m = mpls_label_inet(m, &mh, psize - sizeof(struct sockaddr_mpls));
break;
#endif
#ifdef INET6
case AF_INET6:
m = mpls_label_inet6(m, &mh, psize - sizeof(struct sockaddr_mpls));
break;
#endif
default:
m = mpls_prepend_shim(m, &mh);
break;
}
if (m == NULL) {
IF_DROP(&ifp->if_snd);
ifp->if_oerrors++;
return ENOBUFS;
}
ifp->if_opackets++;
ifp->if_obytes += m->m_pkthdr.len;
if ((rt1=rtalloc1(rt->rt_gateway, 1)) == NULL) {
m_freem(m);
return EHOSTUNREACH;
}
err = mpls_send_frame(m, rt1->rt_ifp, rt);
RTFREE(rt1);
return err;
}
/*
* Do what we need to do when inserting a route.
*/
static struct radix_node *
in6_addroute(void *v_arg, void *n_arg, struct radix_node_head *head,
struct radix_node *treenodes)
{
struct rtentry *rt = (struct rtentry *)treenodes;
struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)rt_key(rt);
struct radix_node *ret;
if (IN6_IS_ADDR_MULTICAST(&sin6->sin6_addr))
rt->rt_flags |= RTF_MULTICAST;
/*
* A little bit of help for both IPv6 output and input:
* For local addresses, we make sure that RTF_LOCAL is set,
* with the thought that this might one day be used to speed up
* ip_input().
*
* We also mark routes to multicast addresses as such, because
* it's easy to do and might be useful (but this is much more
* dubious since it's so easy to inspect the address). (This
* is done above.)
*
* XXX
* should elaborate the code.
*/
if (rt->rt_flags & RTF_HOST) {
if (IN6_ARE_ADDR_EQUAL(&satosin6(rt->rt_ifa->ifa_addr)
->sin6_addr,
&sin6->sin6_addr)) {
rt->rt_flags |= RTF_LOCAL;
}
}
if (!rt->rt_rmx.rmx_mtu && rt->rt_ifp)
rt->rt_rmx.rmx_mtu = IN6_LINKMTU(rt->rt_ifp);
ret = rn_addroute(v_arg, n_arg, head, treenodes);
if (ret == NULL && rt->rt_flags & RTF_HOST) {
struct rtentry *rt2;
/*
* We are trying to add a host route, but can't.
* Find out if it is because of an
* ARP entry and delete it if so.
*/
rt2 = rtalloc1((struct sockaddr *)sin6, 0, RTF_CLONING);
if (rt2) {
if (rt2->rt_flags & RTF_LLINFO &&
rt2->rt_flags & RTF_HOST &&
rt2->rt_gateway &&
rt2->rt_gateway->sa_family == AF_LINK) {
rtexpunge(rt2);
RTFREE_LOCKED(rt2);
ret = rn_addroute(v_arg, n_arg, head,
treenodes);
} else
RTFREE_LOCKED(rt2);
}
} else if (ret == NULL && rt->rt_flags & RTF_CLONING) {
struct rtentry *rt2;
/*
* We are trying to add a net route, but can't.
* The following case should be allowed, so we'll make a
* special check for this:
* Two IPv6 addresses with the same prefix is assigned
* to a single interrface.
* # ifconfig if0 inet6 3ffe:0501::1 prefix 64 alias (*1)
* # ifconfig if0 inet6 3ffe:0501::2 prefix 64 alias (*2)
* In this case, (*1) and (*2) want to add the same
* net route entry, 3ffe:0501:: -> if0.
* This case should not raise an error.
*/
rt2 = rtalloc1((struct sockaddr *)sin6, 0, RTF_CLONING);
if (rt2) {
if ((rt2->rt_flags & (RTF_CLONING|RTF_HOST|RTF_GATEWAY))
== RTF_CLONING
&& rt2->rt_gateway
&& rt2->rt_gateway->sa_family == AF_LINK
&& rt2->rt_ifp == rt->rt_ifp) {
ret = rt2->rt_nodes;
}
RTFREE_LOCKED(rt2);
}
}
return ret;
}
STATUS arpResolve
(
char *targetAddr, /* name or Internet address of target */
char *pHwAddr, /* where to return the H/W address */
int numTries, /* number of times to try ARPing (-1 means try
forever) */
int numTicks /* number of ticks between ARPs */
)
{
struct ifnet * pIf = NULL;
struct sockaddr_in sockInetAddr;
struct rtentry * pRt;
unsigned long addr;
int retVal = 0;
if (pHwAddr == NULL || numTries < -1 || numTries == 0) /* user messed up */
{
errno = S_arpLib_INVALID_ARGUMENT;
return (ERROR);
}
/* the 'targetAddr' can either be the hostname or the actual Internet
* address.
*/
if ((addr = (unsigned long) hostGetByName (targetAddr)) == ERROR &&
(addr = inet_addr (targetAddr)) == ERROR)
{
errno = S_arpLib_INVALID_HOST;
return (ERROR);
}
bzero ((caddr_t)&sockInetAddr, sizeof (sockInetAddr));
sockInetAddr.sin_len = sizeof(struct sockaddr_in);
sockInetAddr.sin_family = AF_INET;
sockInetAddr.sin_addr.s_addr = addr;
/*
* Get associated local interface's ifnet. This search also
* clones an empty ARP entry from the interface route if one
* does not already exist.
*/
pRt = rtalloc1 ( (struct sockaddr *)&sockInetAddr, 1);
if (pRt == NULL)
{
errno = S_arpLib_INVALID_HOST;
return (ERROR);
}
pIf = pRt->rt_ifp;
if (pIf == NULL)
{
rtfree (pRt);
errno = S_arpLib_INVALID_HOST;
return (ERROR);
}
/* return 0xffffffffffff for broadcast Internet address */
if (in_broadcast (sockInetAddr.sin_addr, pIf))
{
bcopy ((char *) etherbroadcastaddr, pHwAddr,
sizeof (etherbroadcastaddr));
rtfree (pRt);
return (OK);
}
/* Try to resolve the Ethernet address by calling arpresolve() which
* may send out ARP request messages out onto the Ethernet wire.
*/
while ((numTries == -1 || numTries-- > 0) &&
(retVal = arpresolve ((struct arpcom *) pIf,
(struct rtentry *)pRt,
(struct mbuf *) NULL,
(struct sockaddr *)&sockInetAddr,
(UCHAR *)pHwAddr))
== 0)
if (numTries) /* don't delay after last arp */
taskDelay (numTicks);
rtfree (pRt);
if (retVal == 0) /* unsuccessful resolution */
{
errno = S_arpLib_INVALID_HOST;
return (ERROR);
}
return (OK);
}
int
rtrequest1(int req, struct rt_addrinfo *info, u_int8_t prio,
struct rtentry **ret_nrt, u_int tableid)
{
int s = splsoftnet(); int error = 0;
struct rtentry *rt, *crt;
struct radix_node *rn;
struct radix_node_head *rnh;
struct ifaddr *ifa;
struct sockaddr *ndst;
struct sockaddr_rtlabel *sa_rl, sa_rl2;
#ifdef MPLS
struct sockaddr_mpls *sa_mpls;
#endif
#define senderr(x) { error = x ; goto bad; }
if ((rnh = rt_gettable(info->rti_info[RTAX_DST]->sa_family, tableid)) ==
NULL)
senderr(EAFNOSUPPORT);
if (info->rti_flags & RTF_HOST)
info->rti_info[RTAX_NETMASK] = NULL;
switch (req) {
case RTM_DELETE:
if ((rn = rnh->rnh_lookup(info->rti_info[RTAX_DST],
info->rti_info[RTAX_NETMASK], rnh)) == NULL)
senderr(ESRCH);
rt = (struct rtentry *)rn;
#ifndef SMALL_KERNEL
/*
* if we got multipath routes, we require users to specify
* a matching RTAX_GATEWAY.
*/
if (rn_mpath_capable(rnh)) {
rt = rt_mpath_matchgate(rt,
info->rti_info[RTAX_GATEWAY], prio);
rn = (struct radix_node *)rt;
if (!rt)
senderr(ESRCH);
}
#endif
if ((rn = rnh->rnh_deladdr(info->rti_info[RTAX_DST],
info->rti_info[RTAX_NETMASK], rnh, rn)) == NULL)
senderr(ESRCH);
rt = (struct rtentry *)rn;
/* clean up any cloned children */
if ((rt->rt_flags & RTF_CLONING) != 0)
rtflushclone(rnh, rt);
if (rn->rn_flags & (RNF_ACTIVE | RNF_ROOT))
panic ("rtrequest delete");
if (rt->rt_gwroute) {
rt = rt->rt_gwroute; RTFREE(rt);
(rt = (struct rtentry *)rn)->rt_gwroute = NULL;
}
if (rt->rt_parent) {
rt->rt_parent->rt_refcnt--;
rt->rt_parent = NULL;
}
#ifndef SMALL_KERNEL
if (rn_mpath_capable(rnh)) {
if ((rn = rnh->rnh_lookup(info->rti_info[RTAX_DST],
info->rti_info[RTAX_NETMASK], rnh)) != NULL &&
rn_mpath_next(rn, 0) == NULL)
((struct rtentry *)rn)->rt_flags &= ~RTF_MPATH;
}
#endif
rt->rt_flags &= ~RTF_UP;
if ((ifa = rt->rt_ifa) && ifa->ifa_rtrequest)
ifa->ifa_rtrequest(RTM_DELETE, rt, info);
rttrash++;
if (ret_nrt)
*ret_nrt = rt;
else if (rt->rt_refcnt <= 0) {
rt->rt_refcnt++;
rtfree(rt);
}
break;
case RTM_RESOLVE:
if (ret_nrt == NULL || (rt = *ret_nrt) == NULL)
senderr(EINVAL);
if ((rt->rt_flags & RTF_CLONING) == 0)
senderr(EINVAL);
ifa = rt->rt_ifa;
info->rti_flags = rt->rt_flags & ~(RTF_CLONING | RTF_STATIC);
info->rti_flags |= RTF_CLONED;
info->rti_info[RTAX_GATEWAY] = rt->rt_gateway;
if ((info->rti_info[RTAX_NETMASK] = rt->rt_genmask) == NULL)
info->rti_flags |= RTF_HOST;
info->rti_info[RTAX_LABEL] =
rtlabel_id2sa(rt->rt_labelid, &sa_rl2);
goto makeroute;
case RTM_ADD:
if (info->rti_ifa == 0 && (error = rt_getifa(info, tableid)))
senderr(error);
ifa = info->rti_ifa;
makeroute:
rt = pool_get(&rtentry_pool, PR_NOWAIT | PR_ZERO);
if (rt == NULL)
senderr(ENOBUFS);
rt->rt_flags = info->rti_flags;
if (prio == 0)
prio = ifa->ifa_ifp->if_priority + RTP_STATIC;
rt->rt_priority = prio; /* init routing priority */
if ((LINK_STATE_IS_UP(ifa->ifa_ifp->if_link_state) ||
ifa->ifa_ifp->if_link_state == LINK_STATE_UNKNOWN) &&
ifa->ifa_ifp->if_flags & IFF_UP)
rt->rt_flags |= RTF_UP;
else {
rt->rt_flags &= ~RTF_UP;
rt->rt_priority |= RTP_DOWN;
}
LIST_INIT(&rt->rt_timer);
if (rt_setgate(rt, info->rti_info[RTAX_DST],
info->rti_info[RTAX_GATEWAY], tableid)) {
pool_put(&rtentry_pool, rt);
senderr(ENOBUFS);
}
ndst = rt_key(rt);
if (info->rti_info[RTAX_NETMASK] != NULL) {
rt_maskedcopy(info->rti_info[RTAX_DST], ndst,
info->rti_info[RTAX_NETMASK]);
} else
Bcopy(info->rti_info[RTAX_DST], ndst,
info->rti_info[RTAX_DST]->sa_len);
#ifndef SMALL_KERNEL
/* do not permit exactly the same dst/mask/gw pair */
if (rn_mpath_capable(rnh) &&
rt_mpath_conflict(rnh, rt, info->rti_info[RTAX_NETMASK],
info->rti_flags & RTF_MPATH)) {
if (rt->rt_gwroute)
rtfree(rt->rt_gwroute);
Free(rt_key(rt));
pool_put(&rtentry_pool, rt);
senderr(EEXIST);
}
#endif
if (info->rti_info[RTAX_LABEL] != NULL) {
sa_rl = (struct sockaddr_rtlabel *)
info->rti_info[RTAX_LABEL];
rt->rt_labelid = rtlabel_name2id(sa_rl->sr_label);
}
#ifdef MPLS
/* We have to allocate additional space for MPLS infos */
if (info->rti_info[RTAX_SRC] != NULL ||
info->rti_info[RTAX_DST]->sa_family == AF_MPLS) {
struct rt_mpls *rt_mpls;
sa_mpls = (struct sockaddr_mpls *)
info->rti_info[RTAX_SRC];
rt->rt_llinfo = (caddr_t)malloc(sizeof(struct rt_mpls),
M_TEMP, M_NOWAIT|M_ZERO);
if (rt->rt_llinfo == NULL) {
if (rt->rt_gwroute)
rtfree(rt->rt_gwroute);
Free(rt_key(rt));
pool_put(&rtentry_pool, rt);
senderr(ENOMEM);
}
rt_mpls = (struct rt_mpls *)rt->rt_llinfo;
if (sa_mpls != NULL)
rt_mpls->mpls_label = sa_mpls->smpls_label;
rt_mpls->mpls_operation = info->rti_mpls;
/* XXX: set experimental bits */
rt->rt_flags |= RTF_MPLS;
}
#endif
ifa->ifa_refcnt++;
rt->rt_ifa = ifa;
rt->rt_ifp = ifa->ifa_ifp;
if (req == RTM_RESOLVE) {
/*
* Copy both metrics and a back pointer to the cloned
* route's parent.
*/
rt->rt_rmx = (*ret_nrt)->rt_rmx; /* copy metrics */
rt->rt_priority = (*ret_nrt)->rt_priority;
rt->rt_parent = *ret_nrt; /* Back ptr. to parent. */
rt->rt_parent->rt_refcnt++;
}
rn = rnh->rnh_addaddr((caddr_t)ndst,
(caddr_t)info->rti_info[RTAX_NETMASK], rnh, rt->rt_nodes,
rt->rt_priority);
if (rn == NULL && (crt = rtalloc1(ndst, 0, tableid)) != NULL) {
/* overwrite cloned route */
if ((crt->rt_flags & RTF_CLONED) != 0) {
rtdeletemsg(crt, tableid);
rn = rnh->rnh_addaddr((caddr_t)ndst,
(caddr_t)info->rti_info[RTAX_NETMASK],
rnh, rt->rt_nodes, rt->rt_priority);
}
RTFREE(crt);
}
if (rn == 0) {
IFAFREE(ifa);
if ((rt->rt_flags & RTF_CLONED) != 0 && rt->rt_parent)
rtfree(rt->rt_parent);
if (rt->rt_gwroute)
rtfree(rt->rt_gwroute);
Free(rt_key(rt));
pool_put(&rtentry_pool, rt);
senderr(EEXIST);
}
#ifndef SMALL_KERNEL
if (rn_mpath_capable(rnh) &&
(rn = rnh->rnh_lookup(info->rti_info[RTAX_DST],
info->rti_info[RTAX_NETMASK], rnh)) != NULL &&
(rn = rn_mpath_prio(rn, prio)) != NULL) {
if (rn_mpath_next(rn, 0) == NULL)
((struct rtentry *)rn)->rt_flags &= ~RTF_MPATH;
else
((struct rtentry *)rn)->rt_flags |= RTF_MPATH;
}
#endif
if (ifa->ifa_rtrequest)
ifa->ifa_rtrequest(req, rt, info);
if (ret_nrt) {
*ret_nrt = rt;
rt->rt_refcnt++;
}
if ((rt->rt_flags & RTF_CLONING) != 0) {
/* clean up any cloned children */
rtflushclone(rnh, rt);
}
if_group_routechange(info->rti_info[RTAX_DST],
info->rti_info[RTAX_NETMASK]);
break;
}
bad:
splx(s);
return (error);
}
/*
* Ethernet output routine.
* Encapsulate a packet of type family for the local net.
* Assumes that ifp is actually pointer to ethercom structure.
*/
int
ssh_interceptor_ether_output(struct ifnet *ifp, struct mbuf *m0,
struct sockaddr *dst, struct rtentry *rt0)
{
u_int16_t etype = 0;
int s, error = 0, hdrcmplt = 0;
u_char esrc[6], edst[6];
struct mbuf *m = m0;
struct rtentry *rt;
struct mbuf *mcopy = (struct mbuf *)0;
struct ether_header *eh, ehd;
#ifdef INET
struct arphdr *ah;
#endif /* INET */
#ifdef NETATALK
struct at_ifaddr *aa;
#endif /* NETATALK */
if ((ifp->if_flags & (IFF_UP|IFF_RUNNING)) != (IFF_UP|IFF_RUNNING))
senderr(ENETDOWN);
ifp->if_lastchange = time;
if ((rt = rt0) != NULL) {
if ((rt->rt_flags & RTF_UP) == 0) {
if ((rt0 = rt = rtalloc1(dst, 1)) != NULL) {
rt->rt_refcnt--;
if (rt->rt_ifp != ifp)
return (*rt->rt_ifp->if_output)
(ifp, m0, dst, rt);
} else
senderr(EHOSTUNREACH);
}
if ((rt->rt_flags & RTF_GATEWAY) && dst->sa_family != AF_NS) {
if (rt->rt_gwroute == 0)
goto lookup;
if (((rt = rt->rt_gwroute)->rt_flags & RTF_UP) == 0) {
rtfree(rt); rt = rt0;
lookup: rt->rt_gwroute = rtalloc1(rt->rt_gateway, 1);
if ((rt = rt->rt_gwroute) == 0)
senderr(EHOSTUNREACH);
/* the "G" test below also prevents rt == rt0 */
if ((rt->rt_flags & RTF_GATEWAY) ||
(rt->rt_ifp != ifp)) {
rt->rt_refcnt--;
rt0->rt_gwroute = 0;
senderr(EHOSTUNREACH);
}
}
}
if (rt->rt_flags & RTF_REJECT)
if (rt->rt_rmx.rmx_expire == 0 ||
time.tv_sec < rt->rt_rmx.rmx_expire)
senderr(rt == rt0 ? EHOSTDOWN : EHOSTUNREACH);
}
switch (dst->sa_family) {
#ifdef INET
case AF_INET:
if (m->m_flags & M_BCAST)
bcopy((caddr_t)etherbroadcastaddr, (caddr_t)edst,
sizeof(edst));
else if (m->m_flags & M_MCAST) {
ETHER_MAP_IP_MULTICAST(&SIN(dst)->sin_addr,
(caddr_t)edst)
} else if (!arpresolve(ifp, rt, m, dst, edst))
return (0); /* if not yet resolved */
/* If broadcasting on a simplex interface, loopback a copy */
if ((m->m_flags & M_BCAST) && (ifp->if_flags & IFF_SIMPLEX))
mcopy = m_copy(m, 0, (int)M_COPYALL);
etype = htons(ETHERTYPE_IP);
break;
case AF_ARP:
ah = mtod(m, struct arphdr *);
if (m->m_flags & M_BCAST)
bcopy((caddr_t)etherbroadcastaddr, (caddr_t)edst,
sizeof(edst));
else
bcopy((caddr_t)ar_tha(ah),
(caddr_t)edst, sizeof(edst));
ah->ar_hrd = htons(ARPHRD_ETHER);
switch(ntohs(ah->ar_op)) {
case ARPOP_REVREQUEST:
case ARPOP_REVREPLY:
etype = htons(ETHERTYPE_REVARP);
break;
case ARPOP_REQUEST:
case ARPOP_REPLY:
default:
etype = htons(ETHERTYPE_ARP);
}
break;
#endif
#ifdef INET6
case AF_INET6:
#ifdef OLDIP6OUTPUT
if (!nd6_resolve(ifp, rt, m, dst, (u_char *)edst))
return(0); /* if not yet resolves */
#else
if (!nd6_storelladdr(ifp, rt, m, dst, (u_char *)edst)){
/* this must be impossible, so we bark */
printf("nd6_storelladdr failed\n");
return(0);
}
#endif /* OLDIP6OUTPUT */
etype = htons(ETHERTYPE_IPV6);
break;
#endif
#ifdef NETATALK
case AF_APPLETALK:
if (!aarpresolve(ifp, m, (struct sockaddr_at *)dst, edst)) {
#ifdef NETATALKDEBUG
printf("aarpresolv failed\n");
#endif /* NETATALKDEBUG */
return (0);
}
/*
* ifaddr is the first thing in at_ifaddr
*/
aa = (struct at_ifaddr *) at_ifawithnet(
(struct sockaddr_at *)dst, ifp);
if (aa == NULL)
goto bad;
/*
* In the phase 2 case, we need to prepend an mbuf for the
* llc header. Since we must preserve the value of m,
* which is passed to us by value, we m_copy() the first
* mbuf, and use it for our llc header.
*/
if (aa->aa_flags & AFA_PHASE2) {
struct llc llc;
M_PREPEND(m, sizeof(struct llc), M_DONTWAIT);
llc.llc_dsap = llc.llc_ssap = LLC_SNAP_LSAP;
llc.llc_control = LLC_UI;
bcopy(at_org_code, llc.llc_snap_org_code,
sizeof(llc.llc_snap_org_code));
llc.llc_snap_ether_type = htons(ETHERTYPE_ATALK);
bcopy(&llc, mtod(m, caddr_t), sizeof(struct llc));
} else {
etype = htons(ETHERTYPE_ATALK);
}
break;
#endif /* NETATALK */
#ifdef NS
case AF_NS:
etype = htons(ETHERTYPE_NS);
bcopy((caddr_t)&(((struct sockaddr_ns *)dst)->sns_addr.x_host),
(caddr_t)edst, sizeof (edst));
if (!bcmp((caddr_t)edst, (caddr_t)&ns_thishost, sizeof(edst)))
return (looutput(ifp, m, dst, rt));
/* If broadcasting on a simplex interface, loopback a copy */
if ((m->m_flags & M_BCAST) && (ifp->if_flags & IFF_SIMPLEX))
mcopy = m_copy(m, 0, (int)M_COPYALL);
break;
#endif
#ifdef IPX
case AF_IPX:
etype = htons(ETHERTYPE_IPX);
bcopy((caddr_t)&(((struct sockaddr_ipx *)dst)->sipx_addr.x_host),
(caddr_t)edst, sizeof (edst));
/* If broadcasting on a simplex interface, loopback a copy */
if ((m->m_flags & M_BCAST) && (ifp->if_flags & IFF_SIMPLEX))
mcopy = m_copy(m, 0, (int)M_COPYALL);
break;
#endif
#ifdef ISO
case AF_ISO: {
int snpalen;
struct llc *l;
struct sockaddr_dl *sdl;
if (rt && (sdl = (struct sockaddr_dl *)rt->rt_gateway) &&
sdl->sdl_family == AF_LINK && sdl->sdl_alen > 0) {
bcopy(LLADDR(sdl), (caddr_t)edst, sizeof(edst));
} else {
error = iso_snparesolve(ifp, (struct sockaddr_iso *)dst,
(char *)edst, &snpalen);
if (error)
goto bad; /* Not Resolved */
}
/* If broadcasting on a simplex interface, loopback a copy */
if (*edst & 1)
m->m_flags |= (M_BCAST|M_MCAST);
if ((m->m_flags & M_BCAST) && (ifp->if_flags & IFF_SIMPLEX) &&
(mcopy = m_copy(m, 0, (int)M_COPYALL))) {
M_PREPEND(mcopy, sizeof (*eh), M_DONTWAIT);
if (mcopy) {
eh = mtod(mcopy, struct ether_header *);
bcopy((caddr_t)edst,
(caddr_t)eh->ether_dhost, sizeof (edst));
bcopy(LLADDR(ifp->if_sadl),
(caddr_t)eh->ether_shost, sizeof (edst));
}
}
M_PREPEND(m, 3, M_DONTWAIT);
if (m == NULL)
return (0);
l = mtod(m, struct llc *);
l->llc_dsap = l->llc_ssap = LLC_ISO_LSAP;
l->llc_control = LLC_UI;
#ifdef ARGO_DEBUG
if (argo_debug[D_ETHER]) {
int i;
printf("unoutput: sending pkt to: ");
for (i=0; i<6; i++)
printf("%x ", edst[i] & 0xff);
printf("\n");
}
#endif
} break;
#endif /* ISO */
#ifdef LLC
/* case AF_NSAP: */
case AF_CCITT: {
struct sockaddr_dl *sdl =
(struct sockaddr_dl *) rt -> rt_gateway;
if (sdl && sdl->sdl_family == AF_LINK
&& sdl->sdl_alen > 0) {
bcopy(LLADDR(sdl), (char *)edst,
sizeof(edst));
} else goto bad; /* Not a link interface ? Funny ... */
if ((ifp->if_flags & IFF_SIMPLEX) && (*edst & 1) &&
(mcopy = m_copy(m, 0, (int)M_COPYALL))) {
M_PREPEND(mcopy, sizeof (*eh), M_DONTWAIT);
if (mcopy) {
eh = mtod(mcopy, struct ether_header *);
bcopy((caddr_t)edst,
(caddr_t)eh->ether_dhost, sizeof (edst));
bcopy(LLADDR(ifp->if_sadl),
(caddr_t)eh->ether_shost, sizeof (edst));
}
}
#ifdef LLC_DEBUG
{
int i;
struct llc *l = mtod(m, struct llc *);
printf("ether_output: sending LLC2 pkt to: ");
for (i=0; i<6; i++)
printf("%x ", edst[i] & 0xff);
printf(" len 0x%x dsap 0x%x ssap 0x%x control 0x%x\n",
m->m_pkthdr.len, l->llc_dsap & 0xff, l->llc_ssap &0xff,
l->llc_control & 0xff);
}
#endif /* LLC_DEBUG */
} break;
/*
* Do what we need to do when inserting a route.
*/
static struct radix_node *
in_addroute(void *v_arg, void *n_arg, struct radix_node_head *head,
struct radix_node *treenodes)
{
struct rtentry *rt = (struct rtentry *)treenodes;
struct sockaddr_in *sin = (struct sockaddr_in *)rt_key(rt);
struct radix_node *ret;
/*
* For IP, all unicast non-host routes are automatically cloning.
*/
if(IN_MULTICAST(ntohl(sin->sin_addr.s_addr)))
rt->rt_flags |= RTF_MULTICAST;
if(!(rt->rt_flags & (RTF_HOST | RTF_CLONING | RTF_MULTICAST))) {
rt->rt_flags |= RTF_PRCLONING;
}
/*
* A little bit of help for both IP output and input:
* For host routes, we make sure that RTF_BROADCAST
* is set for anything that looks like a broadcast address.
* This way, we can avoid an expensive call to in_broadcast()
* in ip_output() most of the time (because the route passed
* to ip_output() is almost always a host route).
*
* We also do the same for local addresses, with the thought
* that this might one day be used to speed up ip_input().
*
* We also mark routes to multicast addresses as such, because
* it's easy to do and might be useful (but this is much more
* dubious since it's so easy to inspect the address). (This
* is done above.)
*/
if (rt->rt_flags & RTF_HOST) {
if (in_broadcast(sin->sin_addr, rt->rt_ifp)) {
rt->rt_flags |= RTF_BROADCAST;
} else {
#define satosin(sa) ((struct sockaddr_in *)sa)
if (satosin(rt->rt_ifa->ifa_addr)->sin_addr.s_addr
== sin->sin_addr.s_addr)
rt->rt_flags |= RTF_LOCAL;
#undef satosin
}
}
/*
* We also specify a send and receive pipe size for every
* route added, to help TCP a bit. TCP doesn't actually
* want a true pipe size, which would be prohibitive in memory
* costs and is hard to compute anyway; it simply uses these
* values to size its buffers. So, we fill them in with the
* same values that TCP would have used anyway, and allow the
* installing program or the link layer to override these values
* as it sees fit. This will hopefully allow TCP more
* opportunities to save its ssthresh value.
*/
if (!rt->rt_rmx.rmx_sendpipe && !(rt->rt_rmx.rmx_locks & RTV_SPIPE))
rt->rt_rmx.rmx_sendpipe = tcp_sendspace;
if (!rt->rt_rmx.rmx_recvpipe && !(rt->rt_rmx.rmx_locks & RTV_RPIPE))
rt->rt_rmx.rmx_recvpipe = tcp_recvspace;
if (!rt->rt_rmx.rmx_mtu && !(rt->rt_rmx.rmx_locks & RTV_MTU)
&& rt->rt_ifp)
rt->rt_rmx.rmx_mtu = rt->rt_ifp->if_mtu;
ret = rn_addroute(v_arg, n_arg, head, treenodes);
if (ret == NULL && rt->rt_flags & RTF_HOST) {
struct rtentry *rt2;
/*
* We are trying to add a host route, but can't.
* Find out if it is because of an
* ARP entry and delete it if so.
*/
rt2 = rtalloc1((struct sockaddr *)sin, 0,
RTF_CLONING | RTF_PRCLONING);
if (rt2) {
if (rt2->rt_flags & RTF_LLINFO &&
rt2->rt_flags & RTF_HOST &&
rt2->rt_gateway &&
rt2->rt_gateway->sa_family == AF_LINK) {
rtrequest(RTM_DELETE,
(struct sockaddr *)rt_key(rt2),
rt2->rt_gateway,
rt_mask(rt2), rt2->rt_flags, 0);
ret = rn_addroute(v_arg, n_arg, head,
treenodes);
}
RTFREE(rt2);
}
}
return ret;
}
/*
* Do what we need to do when inserting a route.
*/
static struct radix_node *
in_addroute(void *v_arg, void *n_arg, struct radix_node_head *head,
struct radix_node *treenodes)
{
struct rtentry *rt = (struct rtentry *)treenodes;
struct sockaddr_in *sin = (struct sockaddr_in *)rt_key(rt);
struct radix_node *ret;
/*
* For IP, all unicast non-host routes are automatically cloning.
*/
if(IN_MULTICAST(ntohl(sin->sin_addr.s_addr)))
rt->rt_flags |= RTF_MULTICAST;
if(!(rt->rt_flags & (RTF_HOST | RTF_CLONING | RTF_MULTICAST))) {
rt->rt_flags |= RTF_PRCLONING;
}
/*
* A little bit of help for both IP output and input:
* For host routes, we make sure that RTF_BROADCAST
* is set for anything that looks like a broadcast address.
* This way, we can avoid an expensive call to in_broadcast()
* in ip_output() most of the time (because the route passed
* to ip_output() is almost always a host route).
*
* We also do the same for local addresses, with the thought
* that this might one day be used to speed up ip_input().
*
* We also mark routes to multicast addresses as such, because
* it's easy to do and might be useful (but this is much more
* dubious since it's so easy to inspect the address). (This
* is done above.)
*/
if (rt->rt_flags & RTF_HOST) {
if (in_broadcast(sin->sin_addr, rt->rt_ifp)) {
rt->rt_flags |= RTF_BROADCAST;
} else {
#define satosin(sa) ((struct sockaddr_in *)sa)
if (satosin(rt->rt_ifa->ifa_addr)->sin_addr.s_addr
== sin->sin_addr.s_addr)
rt->rt_flags |= RTF_LOCAL;
#undef satosin
}
}
if (!rt->rt_rmx.rmx_mtu && !(rt->rt_rmx.rmx_locks & RTV_MTU)
&& rt->rt_ifp)
rt->rt_rmx.rmx_mtu = rt->rt_ifp->if_mtu;
ret = rn_addroute(v_arg, n_arg, head, treenodes);
if (ret == NULL && rt->rt_flags & RTF_HOST) {
struct rtentry *rt2;
/*
* We are trying to add a host route, but can't.
* Find out if it is because of an
* ARP entry and delete it if so.
*/
rt2 = rtalloc1((struct sockaddr *)sin, 0,
RTF_CLONING | RTF_PRCLONING);
if (rt2) {
if (rt2->rt_flags & RTF_LLINFO &&
rt2->rt_flags & RTF_HOST &&
rt2->rt_gateway &&
rt2->rt_gateway->sa_family == AF_LINK) {
rtrequest(RTM_DELETE,
(struct sockaddr *)rt_key(rt2),
rt2->rt_gateway,
rt_mask(rt2), rt2->rt_flags, 0);
ret = rn_addroute(v_arg, n_arg, head,
treenodes);
}
RTFREE(rt2);
}
}
return ret;
}
/*
* NOTE: in6_ifdetach() does not support loopback if at this moment.
* We don't need this function in bsdi, because interfaces are never removed
* from the ifnet list in bsdi.
*/
void
in6_ifdetach(struct ifnet *ifp)
{
struct in6_ifaddr *ia, *oia;
struct ifaddr *ifa, *next;
struct rtentry *rt;
short rtflags;
struct in6_multi_mship *imm;
/* remove ip6_mrouter stuff */
ip6_mrouter_detach(ifp);
/* remove neighbor management table */
nd6_purge(ifp);
/* XXX this code is duplicated in in6_purgeif() --dyoung */
/* nuke any of IPv6 addresses we have */
if_purgeaddrs(ifp, AF_INET6, in6_purgeaddr);
/* XXX isn't this code is redundant, given the above? --dyoung */
/* XXX doesn't this code replicate code in in6_purgeaddr() ? --dyoung */
/* undo everything done by in6_ifattach(), just in case */
for (ifa = IFADDR_FIRST(ifp); ifa != NULL; ifa = next) {
next = IFADDR_NEXT(ifa);
if (ifa->ifa_addr->sa_family != AF_INET6
|| !IN6_IS_ADDR_LINKLOCAL(&satosin6(&ifa->ifa_addr)->sin6_addr)) {
continue;
}
ia = (struct in6_ifaddr *)ifa;
/*
* leave from multicast groups we have joined for the interface
*/
while ((imm = LIST_FIRST(&ia->ia6_memberships)) != NULL) {
LIST_REMOVE(imm, i6mm_chain);
in6_leavegroup(imm);
}
/* remove from the routing table */
if ((ia->ia_flags & IFA_ROUTE) &&
(rt = rtalloc1((struct sockaddr *)&ia->ia_addr, 0))) {
rtflags = rt->rt_flags;
rtfree(rt);
rtrequest(RTM_DELETE, (struct sockaddr *)&ia->ia_addr,
(struct sockaddr *)&ia->ia_addr,
(struct sockaddr *)&ia->ia_prefixmask,
rtflags, NULL);
}
/* remove from the linked list */
ifa_remove(ifp, &ia->ia_ifa);
/* also remove from the IPv6 address chain(itojun&jinmei) */
oia = ia;
if (oia == (ia = in6_ifaddr))
in6_ifaddr = ia->ia_next;
else {
while (ia->ia_next && (ia->ia_next != oia))
ia = ia->ia_next;
if (ia->ia_next)
ia->ia_next = oia->ia_next;
else {
nd6log((LOG_ERR,
"%s: didn't unlink in6ifaddr from list\n",
if_name(ifp)));
}
}
ifafree(&oia->ia_ifa);
}
/* cleanup multicast address kludge table, if there is any */
in6_purgemkludge(ifp);
/*
* remove neighbor management table. we call it twice just to make
* sure we nuke everything. maybe we need just one call.
* XXX: since the first call did not release addresses, some prefixes
* might remain. We should call nd6_purge() again to release the
* prefixes after removing all addresses above.
* (Or can we just delay calling nd6_purge until at this point?)
*/
nd6_purge(ifp);
}
