/*      $FreeBSD: src/sys/netinet6/ip6_output.c,v 1.13.2.18 2003/01/24 05:11:35 sam Exp $       */
/*      $KAME: ip6_output.c,v 1.279 2002/01/26 06:12:30 jinmei Exp $    */

/*
 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. Neither the name of the project nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 */

/*
 * Copyright (c) 1982, 1986, 1988, 1990, 1993
 *      The Regents of the University of California.  All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. Neither the name of the University nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 *
 *      @(#)ip_output.c 8.3 (Berkeley) 1/21/94
 */

#include "opt_ip6fw.h"
#include "opt_inet.h"
#include "opt_inet6.h"

#include <sys/param.h>
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/errno.h>
#include <sys/protosw.h>
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/proc.h>
#include <sys/caps.h>

#include <sys/msgport2.h>

#include <net/if.h>
#include <net/route.h>
#include <net/pfil.h>

#include <netinet/in.h>
#include <netinet/in_var.h>
#include <netinet6/in6_var.h>
#include <netinet/ip6.h>
#include <netinet/icmp6.h>
#include <netinet6/ip6_var.h>
#include <netinet/in_pcb.h>
#include <netinet6/nd6.h>
#include <netinet6/ip6protosw.h>

#include <net/ip6fw/ip6_fw.h>

#include <net/net_osdep.h>

static MALLOC_DEFINE(M_IPMOPTS, "ip6_moptions", "internet multicast options");

struct ip6_exthdrs {
        struct mbuf *ip6e_ip6;
        struct mbuf *ip6e_hbh;
        struct mbuf *ip6e_dest1;
        struct mbuf *ip6e_rthdr;
        struct mbuf *ip6e_dest2;
};

static int ip6_pcbopt (int, u_char *, int, struct ip6_pktopts **, int);
static int ip6_setpktoption (int, u_char *, int, struct ip6_pktopts *,
        int, int, int, int);
static int ip6_pcbopts(struct ip6_pktopts **, struct mbuf *, struct socket *,
        struct sockopt *);
static int ip6_getpcbopt(struct ip6_pktopts *, int, struct sockopt *);
static int ip6_setmoptions (int, struct ip6_moptions **, struct mbuf *);
static int ip6_getmoptions (int, struct ip6_moptions *, struct mbuf **);
static int ip6_getpmtu(struct route_in6 *, struct route_in6 *,
        struct ifnet *, struct in6_addr *, u_long *, int *);
static int copyexthdr (void *, struct mbuf **);
static int ip6_insertfraghdr(struct mbuf *, struct mbuf *, int,
        struct ip6_frag **);
static int ip6_insert_jumboopt (struct ip6_exthdrs *, u_int32_t);
static struct mbuf *ip6_splithdr (struct mbuf *);
static int copypktopts(struct ip6_pktopts *, struct ip6_pktopts *, int);

/*
 * IP6 output. The packet in mbuf chain m contains a skeletal IP6
 * header (with pri, len, nxt, hlim, src, dst).
 * This function may modify ver and hlim only.
 * The mbuf chain containing the packet will be freed.
 * The mbuf opt, if present, will not be freed.
 *
 * type of "mtu": rt_rmx.rmx_mtu is u_long, ifnet.ifr_mtu is int, and
 * nd_ifinfo.linkmtu is u_int32_t.  so we use u_long to hold largest one,
 * which is rt_rmx.rmx_mtu.
 */
int
ip6_output(struct mbuf *m0, struct ip6_pktopts *opt, struct route_in6 *ro,
           int flags, struct ip6_moptions *im6o,
           struct ifnet **ifpp,         /* XXX: just for statistics */
           struct inpcb *inp)
{
        struct ip6_hdr *ip6, *mhip6;
        struct ifnet *ifp, *origifp;
        struct mbuf *m = m0;
        struct mbuf *mprev;
        u_char *nexthdrp;
        int hlen, tlen, len, off;
        struct route_in6 ip6route;
        struct sockaddr_in6 *dst;
        int error = 0;
        struct in6_ifaddr *ia = NULL;
        u_long mtu;
        int alwaysfrag, dontfrag;
        u_int32_t optlen, plen = 0, unfragpartlen;
        struct ip6_exthdrs exthdrs;
        struct in6_addr finaldst;
        struct route_in6 *ro_pmtu = NULL;
        boolean_t hdrsplit = FALSE;

        bzero(&exthdrs, sizeof exthdrs);

        if (opt) {
                if ((error = copyexthdr(opt->ip6po_hbh, &exthdrs.ip6e_hbh)))
                        goto freehdrs;
                if ((error = copyexthdr(opt->ip6po_dest1, &exthdrs.ip6e_dest1)))
                        goto freehdrs;
                if ((error = copyexthdr(opt->ip6po_rthdr, &exthdrs.ip6e_rthdr)))
                        goto freehdrs;
                if ((error = copyexthdr(opt->ip6po_dest2, &exthdrs.ip6e_dest2)))
                        goto freehdrs;
        }

        /*
         * Calculate the total length of the extension header chain.
         * Keep the length of the unfragmentable part for fragmentation.
         */
        optlen = m_lengthm(exthdrs.ip6e_hbh, NULL) +
            m_lengthm(exthdrs.ip6e_dest1, NULL) +
            m_lengthm(exthdrs.ip6e_rthdr, NULL);

        unfragpartlen = optlen + sizeof(struct ip6_hdr);

        /* NOTE: we don't add AH/ESP length here. do that later. */
        optlen += m_lengthm(exthdrs.ip6e_dest2, NULL);

        /*
         * If there is at least one extension header,
         * separate IP6 header from the payload.
         */
        if (optlen && !hdrsplit) {
                exthdrs.ip6e_ip6 = ip6_splithdr(m);
                if (exthdrs.ip6e_ip6 == NULL) {
                        error = ENOBUFS;
                        goto freehdrs;
                }
                m = exthdrs.ip6e_ip6;
                hdrsplit = TRUE;
        }

        /* adjust pointer */
        ip6 = mtod(m, struct ip6_hdr *);

        /* adjust mbuf packet header length */
        m->m_pkthdr.len += optlen;
        plen = m->m_pkthdr.len - sizeof(*ip6);

        /* If this is a jumbo payload, insert a jumbo payload option. */
        if (plen > IPV6_MAXPACKET) {
                if (!hdrsplit) {
                        exthdrs.ip6e_ip6 = ip6_splithdr(m);
                        if (exthdrs.ip6e_ip6 == NULL) {
                                error = ENOBUFS;
                                goto freehdrs;
                        }
                        m = exthdrs.ip6e_ip6;
                        hdrsplit = TRUE;
                }
                /* adjust pointer */
                ip6 = mtod(m, struct ip6_hdr *);
                if ((error = ip6_insert_jumboopt(&exthdrs, plen)) != 0)
                        goto freehdrs;
                ip6->ip6_plen = 0;
        } else
                ip6->ip6_plen = htons(plen);

        /*
         * Concatenate headers and fill in next header fields.
         * Here we have, on "m"
         *      IPv6 payload
         * and we insert headers accordingly.  Finally, we should be getting:
         *      IPv6 hbh dest1 rthdr ah* [esp* dest2 payload]
         *
         * during the header composing process, "m" points to IPv6 header.
         * "mprev" points to an extension header prior to esp.
         */

        nexthdrp = &ip6->ip6_nxt;
        mprev = m;

        /*
         * we treat dest2 specially. the goal here is to make mprev point the
         * mbuf prior to dest2.
         *
         * result: IPv6 dest2 payload
         * m and mprev will point to IPv6 header.
         */
        if (exthdrs.ip6e_dest2) {
                if (!hdrsplit)
                        panic("assumption failed: hdr not split");
                exthdrs.ip6e_dest2->m_next = m->m_next;
                m->m_next = exthdrs.ip6e_dest2;
                *mtod(exthdrs.ip6e_dest2, u_char *) = ip6->ip6_nxt;
                ip6->ip6_nxt = IPPROTO_DSTOPTS;
        }

/*
 * Place m1 after mprev.
 */
#define MAKE_CHAIN(m1, mprev, nexthdrp, i)\
    do {\
        if (m1) {\
                if (!hdrsplit)\
                        panic("assumption failed: hdr not split");\
                *mtod(m1, u_char *) = *nexthdrp;\
                *nexthdrp = (i);\
                nexthdrp = mtod(m1, u_char *);\
                m1->m_next = mprev->m_next;\
                mprev->m_next = m1;\
                mprev = m1;\
        }\
    } while (0)

        /*
         * result: IPv6 hbh dest1 rthdr dest2 payload
         * m will point to IPv6 header.  mprev will point to the
         * extension header prior to dest2 (rthdr in the above case).
         */
        MAKE_CHAIN(exthdrs.ip6e_hbh, mprev, nexthdrp, IPPROTO_HOPOPTS);
        MAKE_CHAIN(exthdrs.ip6e_dest1, mprev, nexthdrp, IPPROTO_DSTOPTS);
        MAKE_CHAIN(exthdrs.ip6e_rthdr, mprev, nexthdrp, IPPROTO_ROUTING);

        /*
         * If there is a routing header, replace the destination address field
         * with the first hop of the routing header.
         */
        if (exthdrs.ip6e_rthdr) {
                struct ip6_rthdr *rh;

                finaldst = ip6->ip6_dst;
                rh = mtod(exthdrs.ip6e_rthdr, struct ip6_rthdr *);
                switch (rh->ip6r_type) {
                default:        /* is it possible? */
                         error = EINVAL;
                         goto bad;
                }
        }

        /* Source address validation */
        if (IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_src) &&
            !(flags & IPV6_DADOUTPUT)) {
                error = EOPNOTSUPP;
                ip6stat.ip6s_badscope++;
                goto bad;
        }
        if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_src)) {
                error = EOPNOTSUPP;
                ip6stat.ip6s_badscope++;
                goto bad;
        }

        ip6stat.ip6s_localout++;

        /*
         * Route packet.
         */
        if (ro == NULL) {
                ro = &ip6route;
                bzero(ro, sizeof(*ro));
        }
        ro_pmtu = ro;
        if (opt && opt->ip6po_rthdr)
                ro = &opt->ip6po_route;
        dst = (struct sockaddr_in6 *)&ro->ro_dst;

        /*
         * If there is a cached route,
         * check that it is to the same destination
         * and is still up. If not, free it and try again.
         */
        if (ro->ro_rt != NULL &&
            (!(ro->ro_rt->rt_flags & RTF_UP) || dst->sin6_family != AF_INET6 ||
             !IN6_ARE_ADDR_EQUAL(&dst->sin6_addr, &ip6->ip6_dst))) {
                RTFREE(ro->ro_rt);
                ro->ro_rt = NULL;
        }
        if (ro->ro_rt == NULL) {
                bzero(dst, sizeof(*dst));
                dst->sin6_family = AF_INET6;
                dst->sin6_len = sizeof(struct sockaddr_in6);
                dst->sin6_addr = ip6->ip6_dst;
        }
        if (!IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) {
                /* Unicast */

#define ifatoia6(ifa)   ((struct in6_ifaddr *)(ifa))
#define sin6tosa(sin6)  ((struct sockaddr *)(sin6))
                /* xxx
                 * interface selection comes here
                 * if an interface is specified from an upper layer,
                 * ifp must point it.
                 */
                if (ro->ro_rt == NULL) {
                        /*
                         * non-bsdi always clone routes, if parent is
                         * PRF_CLONING.
                         */
                        rtalloc((struct route *)ro);
                }
                if (ro->ro_rt == NULL) {
                        ip6stat.ip6s_noroute++;
                        error = EHOSTUNREACH;
                        /* XXX in6_ifstat_inc(ifp, ifs6_out_discard); */
                        goto bad;
                }
                ia = ifatoia6(ro->ro_rt->rt_ifa);
                ifp = ro->ro_rt->rt_ifp;
                ro->ro_rt->rt_use++;
                if (ro->ro_rt->rt_flags & RTF_GATEWAY)
                        dst = (struct sockaddr_in6 *)ro->ro_rt->rt_gateway;
                m->m_flags &= ~(M_BCAST | M_MCAST);     /* just in case */

                in6_ifstat_inc(ifp, ifs6_out_request);

                /*
                 * Check if the outgoing interface conflicts with
                 * the interface specified by ifi6_ifindex (if specified).
                 * Note that loopback interface is always okay.
                 * (this may happen when we are sending a packet to one of
                 *  our own addresses.)
                 */
                if (opt && opt->ip6po_pktinfo
                 && opt->ip6po_pktinfo->ipi6_ifindex) {
                        if (!(ifp->if_flags & IFF_LOOPBACK)
                         && ifp->if_index != opt->ip6po_pktinfo->ipi6_ifindex) {
                                ip6stat.ip6s_noroute++;
                                in6_ifstat_inc(ifp, ifs6_out_discard);
                                error = EHOSTUNREACH;
                                goto bad;
                        }
                }

                if (opt && opt->ip6po_hlim != -1)
                        ip6->ip6_hlim = opt->ip6po_hlim & 0xff;
        } else {
                /* Multicast */
                struct  in6_multi *in6m;

                m->m_flags = (m->m_flags & ~M_BCAST) | M_MCAST;

                /*
                 * See if the caller provided any multicast options
                 */
                ifp = NULL;
                if (im6o != NULL) {
                        ip6->ip6_hlim = im6o->im6o_multicast_hlim;
                        if (im6o->im6o_multicast_ifp != NULL)
                                ifp = im6o->im6o_multicast_ifp;
                } else
                        ip6->ip6_hlim = ip6_defmcasthlim;

                /*
                 * See if the caller provided the outgoing interface
                 * as an ancillary data.
                 * Boundary check for ifindex is assumed to be already done.
                 */
                if (opt && opt->ip6po_pktinfo && opt->ip6po_pktinfo->ipi6_ifindex)
                        ifp = ifindex2ifnet[opt->ip6po_pktinfo->ipi6_ifindex];

                /*
                 * If the destination is a node-local scope multicast,
                 * the packet should be loop-backed only.
                 */
                if (IN6_IS_ADDR_MC_INTFACELOCAL(&ip6->ip6_dst)) {
                        /*
                         * If the outgoing interface is already specified,
                         * it should be a loopback interface.
                         */
                        if (ifp && !(ifp->if_flags & IFF_LOOPBACK)) {
                                ip6stat.ip6s_badscope++;
                                error = ENETUNREACH; /* XXX: better error? */
                                /* XXX correct ifp? */
                                in6_ifstat_inc(ifp, ifs6_out_discard);
                                goto bad;
                        } else {
                                ifp = loif;
                        }
                }

                if (opt && opt->ip6po_hlim != -1)
                        ip6->ip6_hlim = opt->ip6po_hlim & 0xff;

                /*
                 * If caller did not provide an interface lookup a
                 * default in the routing table.  This is either a
                 * default for the speicfied group (i.e. a host
                 * route), or a multicast default (a route for the
                 * ``net'' ff00::/8).
                 */
                if (ifp == NULL) {
                        if (ro->ro_rt == NULL) {
                                ro->ro_rt =
                                  rtpurelookup((struct sockaddr *)&ro->ro_dst);
                        }
                        if (ro->ro_rt == NULL) {
                                ip6stat.ip6s_noroute++;
                                error = EHOSTUNREACH;
                                /* XXX in6_ifstat_inc(ifp, ifs6_out_discard) */
                                goto bad;
                        }
                        ia = ifatoia6(ro->ro_rt->rt_ifa);
                        ifp = ro->ro_rt->rt_ifp;
                        ro->ro_rt->rt_use++;
                }

                if (!(flags & IPV6_FORWARDING))
                        in6_ifstat_inc(ifp, ifs6_out_request);
                in6_ifstat_inc(ifp, ifs6_out_mcast);

                /*
                 * Confirm that the outgoing interface supports multicast.
                 */
                if (!(ifp->if_flags & IFF_MULTICAST)) {
                        ip6stat.ip6s_noroute++;
                        in6_ifstat_inc(ifp, ifs6_out_discard);
                        error = ENETUNREACH;
                        goto bad;
                }
                in6m = IN6_LOOKUP_MULTI(&ip6->ip6_dst, ifp);
                if (in6m != NULL &&
                   (im6o == NULL || im6o->im6o_multicast_loop)) {
                        /*
                         * If we belong to the destination multicast group
                         * on the outgoing interface, and the caller did not
                         * forbid loopback, loop back a copy.
                         */
                        ip6_mloopback(ifp, m, dst);
                } else {
                        /*
                         * If we are acting as a multicast router, perform
                         * multicast forwarding as if the packet had just
                         * arrived on the interface to which we are about
                         * to send.  The multicast forwarding function
                         * recursively calls this function, using the
                         * IPV6_FORWARDING flag to prevent infinite recursion.
                         *
                         * Multicasts that are looped back by ip6_mloopback(),
                         * above, will be forwarded by the ip6_input() routine,
                         * if necessary.
                         */
                        if (ip6_mrouter && !(flags & IPV6_FORWARDING)) {
                                if (ip6_mforward(ip6, ifp, m) != 0) {
                                        m_freem(m);
                                        goto done;
                                }
                        }
                }
                /*
                 * Multicasts with a hoplimit of zero may be looped back,
                 * above, but must not be transmitted on a network.
                 * Also, multicasts addressed to the loopback interface
                 * are not sent -- the above call to ip6_mloopback() will
                 * loop back a copy if this host actually belongs to the
                 * destination group on the loopback interface.
                 */
                if (ip6->ip6_hlim == 0 || (ifp->if_flags & IFF_LOOPBACK)) {
                        m_freem(m);
                        goto done;
                }
        }

        /*
         * Fill the outgoing inteface to tell the upper layer
         * to increment per-interface statistics.
         */
        if (ifpp)
                *ifpp = ifp;

        /* Determine path MTU. */
        if ((error = ip6_getpmtu(ro_pmtu, ro, ifp, &finaldst, &mtu,
            &alwaysfrag)) != 0)
                goto bad;

        /*
         * The caller of this function may specify to use the minimum MTU
         * in some cases.
         * An advanced API option (IPV6_USE_MIN_MTU) can also override MTU
         * setting.  The logic is a bit complicated; by default, unicast
         * packets will follow path MTU while multicast packets will be sent at
         * the minimum MTU.  If IP6PO_MINMTU_ALL is specified, all packets
         * including unicast ones will be sent at the minimum MTU.  Multicast
         * packets will always be sent at the minimum MTU unless
         * IP6PO_MINMTU_DISABLE is explicitly specified.
         * See RFC 3542 for more details.
         */
        if (mtu > IPV6_MMTU) {
                if ((flags & IPV6_MINMTU))
                        mtu = IPV6_MMTU;
                else if (opt && opt->ip6po_minmtu == IP6PO_MINMTU_ALL)
                        mtu = IPV6_MMTU;
                else if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) &&
                         (opt == NULL ||
                          opt->ip6po_minmtu != IP6PO_MINMTU_DISABLE)) {
                        mtu = IPV6_MMTU;
                }
        }

        /* Fake scoped addresses */
        if ((ifp->if_flags & IFF_LOOPBACK) != 0) {
                /*
                 * If source or destination address is a scoped address, and
                 * the packet is going to be sent to a loopback interface,
                 * we should keep the original interface.
                 */

                /*
                 * XXX: this is a very experimental and temporary solution.
                 * We eventually have sockaddr_in6 and use the sin6_scope_id
                 * field of the structure here.
                 * We rely on the consistency between two scope zone ids
                 * of source and destination, which should already be assured.
                 * Larger scopes than link will be supported in the future.
                 */
                origifp = NULL;
                if (IN6_IS_SCOPE_LINKLOCAL(&ip6->ip6_src))
                        origifp = ifindex2ifnet[ntohs(ip6->ip6_src.s6_addr16[1])];
                else if (IN6_IS_SCOPE_LINKLOCAL(&ip6->ip6_dst))
                        origifp = ifindex2ifnet[ntohs(ip6->ip6_dst.s6_addr16[1])];
                /*
                 * XXX: origifp can be NULL even in those two cases above.
                 * For example, if we remove the (only) link-local address
                 * from the loopback interface, and try to send a link-local
                 * address without link-id information.  Then the source
                 * address is ::1, and the destination address is the
                 * link-local address with its s6_addr16[1] being zero.
                 * What is worse, if the packet goes to the loopback interface
                 * by a default rejected route, the null pointer would be
                 * passed to looutput, and the kernel would hang.
                 * The following last resort would prevent such disaster.
                 */
                if (origifp == NULL)
                        origifp = ifp;
        }
        else
                origifp = 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);

        /*
         * Check with the firewall...
         */
        if (ip6_fw_enable && ip6_fw_chk_ptr) {
                u_short port = 0;

                m->m_pkthdr.rcvif = NULL;       /* XXX */
                /* If ipfw says divert, we have to just drop packet */
                if ((*ip6_fw_chk_ptr)(&ip6, ifp, &port, &m)) {
                        m_freem(m);
                        goto done;
                }
                if (!m) {
                        error = EACCES;
                        goto done;
                }
        }

        /*
         * If the outgoing packet contains a hop-by-hop options header,
         * it must be examined and processed even by the source node.
         * (RFC 2460, section 4.)
         */
        if (exthdrs.ip6e_hbh) {
                struct ip6_hbh *hbh = mtod(exthdrs.ip6e_hbh, struct ip6_hbh *);
                u_int32_t dummy1; /* XXX unused */
                u_int32_t dummy2; /* XXX unused */

#ifdef DIAGNOSTIC
                if ((hbh->ip6h_len + 1) << 3 > exthdrs.ip6e_hbh->m_len)
                        panic("ip6e_hbh is not continuous");
#endif
                /*
                 *  XXX: if we have to send an ICMPv6 error to the sender,
                 *       we need the M_LOOP flag since icmp6_error() expects
                 *       the IPv6 and the hop-by-hop options header are
                 *       continuous unless the flag is set.
                 */
                m->m_flags |= M_LOOP;
                m->m_pkthdr.rcvif = ifp;
                if (ip6_process_hopopts(m, (u_int8_t *)(hbh + 1),
                    ((hbh->ip6h_len + 1) << 3) - sizeof(struct ip6_hbh),
                    &dummy1, &dummy2) < 0) {
                        /* m was already freed at this point */
                        error = EINVAL;/* better error? */
                        goto done;
                }
                m->m_flags &= ~M_LOOP; /* XXX */
                m->m_pkthdr.rcvif = NULL;
        }

        /*
         * Run through list of hooks for output packets.
         */
        if (pfil_has_hooks(&inet6_pfil_hook)) {
                error = pfil_run_hooks(&inet6_pfil_hook, &m, ifp, PFIL_OUT);
                if (error != 0 || m == NULL)
                        goto done;
                ip6 = mtod(m, struct ip6_hdr *);
        }

        /*
         * Send the packet to the outgoing interface.
         * If necessary, do IPv6 fragmentation before sending.
         *
         * the logic here is rather complex:
         * 1: normal case (dontfrag == 0, alwaysfrag == 0)
         * 1-a: send as is if tlen <= path mtu
         * 1-b: fragment if tlen > path mtu
         *
         * 2: if user asks us not to fragment (dontfrag == 1)
         * 2-a: send as is if tlen <= interface mtu
         * 2-b: error if tlen > interface mtu
         *
         * 3: if we always need to attach fragment header (alwaysfrag == 1)
         *      always fragment
         *
         * 4: if dontfrag == 1 && alwaysfrag == 1
         *      error, as we cannot handle this conflicting request
         */
        tlen = m->m_pkthdr.len;

        if (opt && (opt->ip6po_flags & IP6PO_DONTFRAG))
                dontfrag = 1;
        else
                dontfrag = 0;
        if (dontfrag && alwaysfrag) {   /* case 4 */
                /* conflicting request - can't transmit */
                error = EMSGSIZE;
                goto bad;
        }
        if (dontfrag && tlen > IN6_LINKMTU(ifp)) {      /* case 2-b */
                /*
                 * Even if the DONTFRAG option is specified, we cannot send the
                 * packet when the data length is larger than the MTU of the
                 * outgoing interface.
                 * Notify the error by sending IPV6_PATHMTU ancillary data as
                 * well as returning an error code (the latter is not described
                 * in the API spec.)
                 */
                u_int32_t mtu32;
                struct ip6ctlparam ip6cp;

                mtu32 = (u_int32_t)mtu;
                bzero(&ip6cp, sizeof(ip6cp));
                ip6cp.ip6c_cmdarg = (void *)&mtu32;
                kpfctlinput2(PRC_MSGSIZE, (struct sockaddr *)&ro_pmtu->ro_dst,
                    (void *)&ip6cp);

                error = EMSGSIZE;
                goto bad;
        }

        /*
         * transmit packet without fragmentation
         */
        if (dontfrag || (!alwaysfrag && tlen <= mtu)) { /* case 1-a and 2-a */
                struct in6_ifaddr *ia6;

                ip6 = mtod(m, struct ip6_hdr *);
                ia6 = in6_ifawithifp(ifp, &ip6->ip6_src);
                if (ia6) {
                        /* Record statistics for this interface address. */
                        IFA_STAT_INC(&ia6->ia_ifa, opackets, 1);
                        IFA_STAT_INC(&ia6->ia_ifa, obytes, m->m_pkthdr.len);
                }
                error = nd6_output(ifp, origifp, m, dst, ro->ro_rt);
                goto done;
        }

        /*
         * try to fragment the packet.  case 1-b and 3
         */
        if (mtu < IPV6_MMTU) {
                /*
                 * note that path MTU is never less than IPV6_MMTU
                 * (see icmp6_input).
                 */
                error = EMSGSIZE;
                in6_ifstat_inc(ifp, ifs6_out_fragfail);
                goto bad;
        } else if (ip6->ip6_plen == 0) {
                /* jumbo payload cannot be fragmented */
                error = EMSGSIZE;
                in6_ifstat_inc(ifp, ifs6_out_fragfail);
                goto bad;
        } else {
                struct mbuf **mnext, *m_frgpart;
                struct ip6_frag *ip6f;
                u_int32_t id = htonl(ip6_id++);
                u_char nextproto;

                /*
                 * Too large for the destination or interface;
                 * fragment if possible.
                 * Must be able to put at least 8 bytes per fragment.
                 */
                hlen = unfragpartlen;
                if (mtu > IPV6_MAXPACKET)
                        mtu = IPV6_MAXPACKET;

                len = (mtu - hlen - sizeof(struct ip6_frag)) & ~7;
                if (len < 8) {
                        error = EMSGSIZE;
                        in6_ifstat_inc(ifp, ifs6_out_fragfail);
                        goto bad;
                }

                mnext = &m->m_nextpkt;

                /*
                 * Change the next header field of the last header in the
                 * unfragmentable part.
                 */
                if (exthdrs.ip6e_rthdr) {
                        nextproto = *mtod(exthdrs.ip6e_rthdr, u_char *);
                        *mtod(exthdrs.ip6e_rthdr, u_char *) = IPPROTO_FRAGMENT;
                } else if (exthdrs.ip6e_dest1) {
                        nextproto = *mtod(exthdrs.ip6e_dest1, u_char *);
                        *mtod(exthdrs.ip6e_dest1, u_char *) = IPPROTO_FRAGMENT;
                } else if (exthdrs.ip6e_hbh) {
                        nextproto = *mtod(exthdrs.ip6e_hbh, u_char *);
                        *mtod(exthdrs.ip6e_hbh, u_char *) = IPPROTO_FRAGMENT;
                } else {
                        nextproto = ip6->ip6_nxt;
                        ip6->ip6_nxt = IPPROTO_FRAGMENT;
                }

                /*
                 * Loop through length of segment after first fragment,
                 * make new header and copy data of each part and link onto
                 * chain.
                 */
                m0 = m;
                for (off = hlen; off < tlen; off += len) {
                        MGETHDR(m, M_NOWAIT, MT_HEADER);
                        if (!m) {
                                error = ENOBUFS;
                                ip6stat.ip6s_odropped++;
                                goto sendorfree;
                        }
                        m->m_pkthdr.rcvif = NULL;
                        m->m_flags = m0->m_flags & M_COPYFLAGS;
                        *mnext = m;
                        mnext = &m->m_nextpkt;
                        m->m_data += max_linkhdr;
                        mhip6 = mtod(m, struct ip6_hdr *);
                        *mhip6 = *ip6;
                        m->m_len = sizeof(*mhip6);
                        error = ip6_insertfraghdr(m0, m, hlen, &ip6f);
                        if (error) {
                                ip6stat.ip6s_odropped++;
                                goto sendorfree;
                        }
                        ip6f->ip6f_offlg = htons((u_short)((off - hlen) & ~7));
                        if (off + len >= tlen)
                                len = tlen - off;
                        else
                                ip6f->ip6f_offlg |= IP6F_MORE_FRAG;
                        mhip6->ip6_plen = htons((u_short)(len + hlen +
                            sizeof(*ip6f) - sizeof(struct ip6_hdr)));
                        m_frgpart = m_copym(m0, off, len, M_NOWAIT);
                        if (m_frgpart == NULL) {
                                error = ENOBUFS;
                                ip6stat.ip6s_odropped++;
                                goto sendorfree;
                        }
                        m_cat(m, m_frgpart);
                        m->m_pkthdr.len = len + hlen + sizeof(*ip6f);
                        m->m_pkthdr.rcvif = NULL;
                        ip6f->ip6f_reserved = 0;
                        ip6f->ip6f_ident = id;
                        ip6f->ip6f_nxt = nextproto;
                        ip6stat.ip6s_ofragments++;
                        in6_ifstat_inc(ifp, ifs6_out_fragcreat);
                }

                in6_ifstat_inc(ifp, ifs6_out_fragok);
        }

        /*
         * Remove leading garbages.
         */
sendorfree:
        m = m0->m_nextpkt;
        m0->m_nextpkt = NULL;
        m_freem(m0);
        for (m0 = m; m; m = m0) {
                m0 = m->m_nextpkt;
                m->m_nextpkt = NULL;
                if (error == 0) {
                        /* Record statistics for this interface address. */
                        if (ia) {
                                IFA_STAT_INC(&ia->ia_ifa, opackets, 1);
                                IFA_STAT_INC(&ia->ia_ifa, obytes,
                                    m->m_pkthdr.len);
                        }
                        error = nd6_output(ifp, origifp, m, dst, ro->ro_rt);
                } else
                        m_freem(m);
        }

        if (error == 0)
                ip6stat.ip6s_fragmented++;

done:
        if (ro == &ip6route && ro->ro_rt) { /* brace necessary for RTFREE */
                RTFREE(ro->ro_rt);
        } else if (ro_pmtu == &ip6route && ro_pmtu->ro_rt) {
                RTFREE(ro_pmtu->ro_rt);
        }

        return (error);

freehdrs:
        m_freem(exthdrs.ip6e_hbh);      /* m_freem will check if mbuf is 0 */
        m_freem(exthdrs.ip6e_dest1);
        m_freem(exthdrs.ip6e_rthdr);
        m_freem(exthdrs.ip6e_dest2);
        /* FALLTHROUGH */
bad:
        m_freem(m);
        goto done;
}

static int
copyexthdr(void *h, struct mbuf **mp)
{
        struct ip6_ext *hdr = h;
        int hlen;
        struct mbuf *m;

        if (hdr == NULL)
                return 0;

        hlen = (hdr->ip6e_len + 1) * 8;
        if (hlen > MCLBYTES)
                return ENOBUFS; /* XXX */

        m = m_getb(hlen, M_NOWAIT, MT_DATA, 0);
        if (!m)
                return ENOBUFS;
        m->m_len = hlen;

        bcopy(hdr, mtod(m, caddr_t), hlen);

        *mp = m;
        return 0;
}

/*
 * Insert jumbo payload option.
 */
static int
ip6_insert_jumboopt(struct ip6_exthdrs *exthdrs, u_int32_t plen)
{
        struct mbuf *mopt;
        u_char *optbuf;
        u_int32_t v;

#define JUMBOOPTLEN     8       /* length of jumbo payload option and padding */

        /*
         * If there is no hop-by-hop options header, allocate new one.
         * If there is one but it doesn't have enough space to store the
         * jumbo payload option, allocate a cluster to store the whole options.
         * Otherwise, use it to store the options.
         */
        if (exthdrs->ip6e_hbh == NULL) {
                MGET(mopt, M_NOWAIT, MT_DATA);
                if (mopt == NULL)
                        return (ENOBUFS);
                mopt->m_len = JUMBOOPTLEN;
                optbuf = mtod(mopt, u_char *);
                optbuf[1] = 0;  /* = ((JUMBOOPTLEN) >> 3) - 1 */
                exthdrs->ip6e_hbh = mopt;
        } else {
                struct ip6_hbh *hbh;

                mopt = exthdrs->ip6e_hbh;
                if (M_TRAILINGSPACE(mopt) < JUMBOOPTLEN) {
                        /*
                         * XXX assumption:
                         * - exthdrs->ip6e_hbh is not referenced from places
                         *   other than exthdrs.
                         * - exthdrs->ip6e_hbh is not an mbuf chain.
                         */
                        int oldoptlen = mopt->m_len;
                        struct mbuf *n;

                        /*
                         * XXX: give up if the whole (new) hbh header does
                         * not fit even in an mbuf cluster.
                         */
                        if (oldoptlen + JUMBOOPTLEN > MCLBYTES)
                                return (ENOBUFS);

                        /*
                         * As a consequence, we must always prepare a cluster
                         * at this point.
                         */
                        n = m_getcl(M_NOWAIT, MT_DATA, 0);
                        if (!n)
                                return (ENOBUFS);
                        n->m_len = oldoptlen + JUMBOOPTLEN;
                        bcopy(mtod(mopt, caddr_t), mtod(n, caddr_t), oldoptlen);
                        optbuf = mtod(n, caddr_t) + oldoptlen;
                        m_freem(mopt);
                        mopt = exthdrs->ip6e_hbh = n;
                } else {
                        optbuf = mtod(mopt, u_char *) + mopt->m_len;
                        mopt->m_len += JUMBOOPTLEN;
                }
                optbuf[0] = IP6OPT_PADN;
                optbuf[1] = 1;

                /*
                 * Adjust the header length according to the pad and
                 * the jumbo payload option.
                 */
                hbh = mtod(mopt, struct ip6_hbh *);
                hbh->ip6h_len += (JUMBOOPTLEN >> 3);
        }

        /* fill in the option. */
        optbuf[2] = IP6OPT_JUMBO;
        optbuf[3] = 4;
        v = (u_int32_t)htonl(plen + JUMBOOPTLEN);
        bcopy(&v, &optbuf[4], sizeof(u_int32_t));

        /* finally, adjust the packet header length */
        exthdrs->ip6e_ip6->m_pkthdr.len += JUMBOOPTLEN;

        return (0);
#undef JUMBOOPTLEN
}

/*
 * Insert fragment header and copy unfragmentable header portions.
 */
static int
ip6_insertfraghdr(struct mbuf *m0, struct mbuf *m, int hlen,
                  struct ip6_frag **frghdrp)
{
        struct mbuf *n, *mlast;

        if (hlen > sizeof(struct ip6_hdr)) {
                n = m_copym(m0, sizeof(struct ip6_hdr),
                            hlen - sizeof(struct ip6_hdr), M_NOWAIT);
                if (n == NULL)
                        return (ENOBUFS);
                m->m_next = n;
        } else
                n = m;

        /* Search for the last mbuf of unfragmentable part. */
        for (mlast = n; mlast->m_next; mlast = mlast->m_next)
                ;

        if (!(mlast->m_flags & M_EXT) &&
            M_TRAILINGSPACE(mlast) >= sizeof(struct ip6_frag)) {
                /* use the trailing space of the last mbuf for the fragment hdr */
                *frghdrp = (struct ip6_frag *)
                    (mtod(mlast, caddr_t) + mlast->m_len);
                mlast->m_len += sizeof(struct ip6_frag);
                m->m_pkthdr.len += sizeof(struct ip6_frag);
        } else {
                /* allocate a new mbuf for the fragment header */
                struct mbuf *mfrg;

                MGET(mfrg, M_NOWAIT, MT_DATA);
                if (mfrg == NULL)
                        return (ENOBUFS);
                mfrg->m_len = sizeof(struct ip6_frag);
                *frghdrp = mtod(mfrg, struct ip6_frag *);
                mlast->m_next = mfrg;
        }

        return (0);
}

static int
ip6_getpmtu(struct route_in6 *ro_pmtu, struct route_in6 *ro,
    struct ifnet *ifp, struct in6_addr *dst, u_long *mtup,
    int *alwaysfragp)
{
        u_int32_t mtu = 0;
        int alwaysfrag = 0;
        int error = 0;

        if (ro_pmtu != ro) {
                /* The first hop and the final destination may differ. */
                struct sockaddr_in6 *sa6_dst =
                    (struct sockaddr_in6 *)&ro_pmtu->ro_dst;
                if (ro_pmtu->ro_rt &&
                    ((ro_pmtu->ro_rt->rt_flags & RTF_UP) == 0 ||
                     !IN6_ARE_ADDR_EQUAL(&sa6_dst->sin6_addr, dst))) {
                        RTFREE(ro_pmtu->ro_rt);
                        ro_pmtu->ro_rt = NULL;
                }
                if (ro_pmtu->ro_rt == NULL) {
                        bzero(sa6_dst, sizeof(*sa6_dst));
                        sa6_dst->sin6_family = AF_INET6;
                        sa6_dst->sin6_len = sizeof(struct sockaddr_in6);
                        sa6_dst->sin6_addr = *dst;

                        rtalloc((struct route *)ro_pmtu);
                }
        }
        if (ro_pmtu->ro_rt) {
                u_int32_t ifmtu;

                if (ifp == NULL)
                        ifp = ro_pmtu->ro_rt->rt_ifp;
                ifmtu = IN6_LINKMTU(ifp);
                mtu = ro_pmtu->ro_rt->rt_rmx.rmx_mtu;
                if (mtu == 0) {
                        mtu = ifmtu;
                } else if (mtu < IPV6_MMTU) {
                        /*
                         * RFC2460 section 5, last paragraph:
                         * if we record ICMPv6 too big message with
                         * mtu < IPV6_MMTU, transmit packets sized IPV6_MMTU
                         * or smaller, with framgent header attached.
                         * (fragment header is needed regardless from the
                         * packet size, for translators to identify packets)
                         */
                        alwaysfrag = 1;
                        mtu = IPV6_MMTU;
                } else if (mtu > ifmtu) {
                        /*
                         * The MTU on the route is larger than the MTU on
                         * the interface!  This shouldn't happen, unless the
                         * MTU of the interface has been changed after the
                         * interface was brought up.  Change the MTU in the
                         * route to match the interface MTU (as long as the
                         * field isn't locked).
                         */
                        mtu = ifmtu;
                        ro_pmtu->ro_rt->rt_rmx.rmx_mtu = mtu;
                }
        } else if (ifp) {
                mtu = IN6_LINKMTU(ifp);
        } else {
                error = EHOSTUNREACH; /* XXX */
        }

        *mtup = mtu;
        if (alwaysfragp)
                *alwaysfragp = alwaysfrag;
        return (error);
}

/*
 * IP6 socket option processing.
 */
void
ip6_ctloutput_dispatch(netmsg_t msg)
{
        int error;

        error = ip6_ctloutput(msg->ctloutput.base.nm_so,
                              msg->ctloutput.nm_sopt);
        lwkt_replymsg(&msg->ctloutput.base.lmsg, error);
}

int
ip6_ctloutput(struct socket *so, struct sockopt *sopt)
{
        int optdatalen,uproto;
        int privileged;
        struct inpcb *in6p = so->so_pcb;
        void *optdata;
        int error, optval;
        int level, op, optname;
        int optlen;
        struct thread *td;

        if (sopt) {
                level = sopt->sopt_level;
                op = sopt->sopt_dir;
                optname = sopt->sopt_name;
                optlen = sopt->sopt_valsize;
                td = sopt->sopt_td;
        } else {
                panic("ip6_ctloutput: arg soopt is NULL");
                /* NOT REACHED */
                td = NULL;
        }
        error = optval = 0;

        uproto = (int)so->so_proto->pr_protocol;
        privileged = (td == NULL ||
                      caps_priv_check_td(td, SYSCAP_RESTRICTEDROOT)) ?
                     0 : 1;

        if (level == IPPROTO_IPV6) {
                switch (op) {

                case SOPT_SET:
                        switch (optname) {
                        case IPV6_2292PKTOPTIONS:
#ifdef IPV6_PKTOPTIONS
                        case IPV6_PKTOPTIONS:
#endif
                        {
                                struct mbuf *m;

                                error = soopt_getm(sopt, &m); /* XXX */
                                if (error != 0)
                                        break;
                                soopt_to_mbuf(sopt, m); /* XXX */
                                error = ip6_pcbopts(&in6p->in6p_outputopts,
                                                    m, so, sopt);
                                m_freem(m); /* XXX */
                                break;
                        }

                        /*
                         * Use of some Hop-by-Hop options or some
                         * Destination options, might require special
                         * privilege.  That is, normal applications
                         * (without special privilege) might be forbidden
                         * from setting certain options in outgoing packets,
                         * and might never see certain options in received
                         * packets. [RFC 2292 Section 6]
                         * KAME specific note:
                         *  KAME prevents non-privileged users from sending or
                         *  receiving ANY hbh/dst options in order to avoid
                         *  overhead of parsing options in the kernel.
                         */
                        case IPV6_RECVHOPOPTS:
                        case IPV6_RECVDSTOPTS:
                        case IPV6_RECVRTHDRDSTOPTS:
                                if (!privileged)
                                        return (EPERM);
                        case IPV6_RECVPKTINFO:
                        case IPV6_RECVHOPLIMIT:
                        case IPV6_RECVRTHDR:
                        case IPV6_RECVPATHMTU:
                        case IPV6_RECVTCLASS:
                        case IPV6_AUTOFLOWLABEL:
                        case IPV6_HOPLIMIT:
                        /* FALLTHROUGH */
                        case IPV6_UNICAST_HOPS:

                        case IPV6_V6ONLY:
                                if (optlen != sizeof(int)) {
                                        error = EINVAL;
                                        break;
                                }
                                error = soopt_to_kbuf(sopt, &optval,
                                        sizeof optval, sizeof optval);
                                if (error)
                                        break;
                                switch (optname) {

                                case IPV6_UNICAST_HOPS:
                                        if (optval < -1 || optval >= 256)
                                                error = EINVAL;
                                        else {
                                                /* -1 = kernel default */
                                                in6p->in6p_hops = optval;
                                        }
                                        break;
#define OPTSET(bit) \
do { \
        if (optval) \
                in6p->in6p_flags |= (bit); \
        else \
                in6p->in6p_flags &= ~(bit); \
} while (0)
#define OPTBIT(bit) (in6p->in6p_flags & (bit) ? 1 : 0)
/*
 * Although changed to RFC3542, It's better to also support RFC2292 API
 */
#define OPTSET2292(bit) \
do { \
        in6p->in6p_flags |= IN6P_RFC2292; \
        if (optval) \
                in6p->in6p_flags |= (bit); \
        else \
                in6p->in6p_flags &= ~(bit); \
} while (/*CONSTCOND*/ 0)

                                case IPV6_RECVPKTINFO:
                                        /* cannot mix with RFC2292 */
                                        if (OPTBIT(IN6P_RFC2292)) {
                                                error = EINVAL;
                                                break;
                                        }
                                        OPTSET(IN6P_PKTINFO);
                                        break;

                                case IPV6_HOPLIMIT:
                                {
                                        struct ip6_pktopts **optp;

                                        /* cannot mix with RFC2292 */
                                        if (OPTBIT(IN6P_RFC2292)) {
                                                error = EINVAL;
                                                break;
                                        }
                                        optp = &in6p->in6p_outputopts;
                                        error = ip6_pcbopt(IPV6_HOPLIMIT,
                                            (u_char *)&optval, sizeof(optval),
                                            optp, uproto);
                                        break;
                                }

                                case IPV6_RECVHOPLIMIT:
                                        /* cannot mix with RFC2292 */
                                        if (OPTBIT(IN6P_RFC2292)) {
                                                error = EINVAL;
                                                break;
                                        }
                                        OPTSET(IN6P_HOPLIMIT);
                                        break;

                                case IPV6_RECVHOPOPTS:
                                        /* cannot mix with RFC2292 */
                                        if (OPTBIT(IN6P_RFC2292)) {
                                                error = EINVAL;
                                                break;
                                        }
                                        OPTSET(IN6P_HOPOPTS);
                                        break;

                                case IPV6_RECVDSTOPTS:
                                        /* cannot mix with RFC2292 */
                                        if (OPTBIT(IN6P_RFC2292)) {
                                                error = EINVAL;
                                                break;
                                        }
                                        OPTSET(IN6P_DSTOPTS);
                                        break;

                                case IPV6_RECVRTHDRDSTOPTS:
                                        /* cannot mix with RFC2292 */
                                        if (OPTBIT(IN6P_RFC2292)) {
                                                error = EINVAL;
                                                break;
                                        }
                                        OPTSET(IN6P_RTHDRDSTOPTS);
                                        break;

                                case IPV6_RECVRTHDR:
                                        /* cannot mix with RFC2292 */
                                        if (OPTBIT(IN6P_RFC2292)) {
                                                error = EINVAL;
                                                break;
                                        }
                                        OPTSET(IN6P_RTHDR);
                                        break;

                                case IPV6_RECVPATHMTU:
                                        /*
                                         * We ignore this option for TCP
                                         * sockets.
                                         * (RFC3542 leaves this case
                                         * unspecified.)
                                         */
                                        if (uproto != IPPROTO_TCP)
                                                OPTSET(IN6P_MTU);
                                        break;

                                case IPV6_RECVTCLASS:
                                        /* cannot mix with RFC2292 XXX */
                                        if (OPTBIT(IN6P_RFC2292)) {
                                                error = EINVAL;
                                                break;
                                        }
                                        OPTSET(IN6P_TCLASS);
                                        break;

                                case IPV6_AUTOFLOWLABEL:
                                        OPTSET(IN6P_AUTOFLOWLABEL);
                                        break;

                                case IPV6_V6ONLY:
                                        /*
                                         * make setsockopt(IPV6_V6ONLY)
                                         * available only prior to bind(2).
                                         */
                                        if (in6p->in6p_lport ||
                                            !IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_laddr))
                                        {
                                                error = EINVAL;
                                                break;
                                        }

                                        /*
                                         * Since we don't support v6->v4
                                         * mapping any more this option does
                                         * nothing.  But apparently some
                                         * ports and libraries (e.g. libuv)
                                         * actually try to set the value to
                                         * 0 so just silently ignore the value
                                         * entirely.
                                         *
                                         * (also fixes named which uses libuv,
                                         * and a few other apps)
                                         */
#if 0
                                        if (!optval) {
                                                /* Don't allow v4-mapped */
                                                error = EOPNOTSUPP;
                                        }
#endif
                                        break;
                                }
                                break;

                        case IPV6_TCLASS:
                        case IPV6_DONTFRAG:
                        case IPV6_USE_MIN_MTU:
                        case IPV6_PREFER_TEMPADDR:
                                if (optlen != sizeof(optval)) {
                                        error = EINVAL;
                                        break;
                                }
                                error = soopt_to_kbuf(sopt, &optval,
                                        sizeof optval, sizeof optval);
                                if (error)
                                        break;
                                {
                                        struct ip6_pktopts **optp;
                                        optp = &in6p->in6p_outputopts;
                                        error = ip6_pcbopt(optname,
                                            (u_char *)&optval, sizeof(optval),
                                            optp, uproto);
                                        break;
                                }

                        case IPV6_2292PKTINFO:
                        case IPV6_2292HOPLIMIT:
                        case IPV6_2292HOPOPTS:
                        case IPV6_2292DSTOPTS:
                        case IPV6_2292RTHDR:
                                /* RFC 2292 */
                                if (optlen != sizeof(int)) {
                                        error = EINVAL;
                                        break;
                                }
                                error = soopt_to_kbuf(sopt, &optval,
                                        sizeof optval, sizeof optval);
                                if (error)
                                        break;
                                switch (optname) {
                                case IPV6_2292PKTINFO:
                                        OPTSET2292(IN6P_PKTINFO);
                                        break;
                                case IPV6_2292HOPLIMIT:
                                        OPTSET2292(IN6P_HOPLIMIT);
                                        break;
                                case IPV6_2292HOPOPTS:
                                        /*
                                         * Check super-user privilege.
                                         * See comments for IPV6_RECVHOPOPTS.
                                         */
                                        if (!privileged)
                                                return (EPERM);
                                        OPTSET2292(IN6P_HOPOPTS);
                                        break;
                                case IPV6_2292DSTOPTS:
                                        if (!privileged)
                                                return (EPERM);
                                        OPTSET2292(IN6P_DSTOPTS|IN6P_RTHDRDSTOPTS); /* XXX */
                                        break;
                                case IPV6_2292RTHDR:
                                        OPTSET2292(IN6P_RTHDR);
                                        break;
                                }
                                break;

                        case IPV6_PKTINFO:
                        case IPV6_HOPOPTS:
                        case IPV6_RTHDR:
                        case IPV6_DSTOPTS:
                        case IPV6_RTHDRDSTOPTS:
                        case IPV6_NEXTHOP:
                        {
                                /*
                                 * New advanced API (RFC3542)
                                 */
                                u_char *optbuf;
                                u_char optbuf_storage[MCLBYTES];
                                int optlen;
                                struct ip6_pktopts **optp;

                                /* cannot mix with RFC2292 */
                                if (OPTBIT(IN6P_RFC2292)) {
                                        error = EINVAL;
                                        break;
                                }

                                /*
                                 * We only ensure valsize is not too large
                                 * here.  Further validation will be done
                                 * later.
                                 */
                                error = soopt_to_kbuf(sopt, optbuf_storage,
                                    sizeof(optbuf_storage), 0);
                                if (error)
                                        break;
                                optlen = sopt->sopt_valsize;
                                optbuf = optbuf_storage;
                                optp = &in6p->in6p_outputopts;
                                error = ip6_pcbopt(optname, optbuf, optlen,
                                    optp, uproto);
                                break;
                        }
#undef OPTSET

                        case IPV6_MULTICAST_IF:
                        case IPV6_MULTICAST_HOPS:
                        case IPV6_MULTICAST_LOOP:
                        case IPV6_JOIN_GROUP:
                        case IPV6_LEAVE_GROUP:
                            {
                                struct mbuf *m;

                                if (sopt->sopt_valsize > MLEN) {
                                        error = EMSGSIZE;
                                        break;
                                }
                                /* XXX */
                                MGET(m, sopt->sopt_td ? M_WAITOK : M_NOWAIT, MT_HEADER);
                                if (m == NULL) {
                                        error = ENOBUFS;
                                        break;
                                }
                                m->m_len = sopt->sopt_valsize;
                                error = soopt_to_kbuf(sopt, mtod(m, char *),
                                                    m->m_len, m->m_len);
                                error = ip6_setmoptions(sopt->sopt_name,
                                                        &in6p->in6p_moptions,
                                                        m);
                                m_free(m);
                            }
                                break;

                        case IPV6_PORTRANGE:
                                error = soopt_to_kbuf(sopt, &optval,
                                    sizeof optval, sizeof optval);
                                if (error)
                                        break;

                                switch (optval) {
                                case IPV6_PORTRANGE_DEFAULT:
                                        in6p->in6p_flags &= ~(IN6P_LOWPORT);
                                        in6p->in6p_flags &= ~(IN6P_HIGHPORT);
                                        break;

                                case IPV6_PORTRANGE_HIGH:
                                        in6p->in6p_flags &= ~(IN6P_LOWPORT);
                                        in6p->in6p_flags |= IN6P_HIGHPORT;
                                        break;

                                case IPV6_PORTRANGE_LOW:
                                        in6p->in6p_flags &= ~(IN6P_HIGHPORT);
                                        in6p->in6p_flags |= IN6P_LOWPORT;
                                        break;

                                default:
                                        error = EINVAL;
                                        break;
                                }
                                break;

                        case IPV6_FW_ADD:
                        case IPV6_FW_DEL:
                        case IPV6_FW_FLUSH:
                        case IPV6_FW_ZERO:
                            {
                                struct mbuf *m;
                                struct mbuf **mp = &m;

                                if (ip6_fw_ctl_ptr == NULL)
                                        return EINVAL;
                                /* XXX */
                                if ((error = soopt_getm(sopt, &m)) != 0)
                                        break;
                                /* XXX */
                                soopt_to_mbuf(sopt, m);
                                error = (*ip6_fw_ctl_ptr)(optname, mp);
                                m = *mp;
                            }
                                break;

                        default:
                                error = ENOPROTOOPT;
                                break;
                        }
                        break;

                case SOPT_GET:
                        switch (optname) {
                        case IPV6_2292PKTOPTIONS:
#ifdef IPV6_PKTOPTIONS
                        case IPV6_PKTOPTIONS:
#endif
                                /*
                                 * RFC3542 (effectively) deprecated the
                                 * semantics of the 2292-style pktoptions.
                                 * Since it was not reliable in nature (i.e.,
                                 * applications had to expect the lack of some
                                 * information after all), it would make sense
                                 * to simplify this part by always returning
                                 * empty data.
                                 */
                                if (in6p->in6p_options) {
                                        struct mbuf *m;
                                        m = m_copym(in6p->in6p_options,
                                            0, M_COPYALL, M_WAITOK);
                                        error = soopt_from_mbuf(sopt, m);
                                        if (error == 0)
                                                m_freem(m);
                                } else
                                        sopt->sopt_valsize = 0;
                                break;

                        case IPV6_RECVHOPOPTS:
                        case IPV6_RECVDSTOPTS:
                        case IPV6_RECVRTHDRDSTOPTS:
                        case IPV6_UNICAST_HOPS:
                        case IPV6_RECVPKTINFO:
                        case IPV6_RECVHOPLIMIT:
                        case IPV6_RECVRTHDR:
                        case IPV6_RECVPATHMTU:
                        case IPV6_RECVTCLASS:
                        case IPV6_AUTOFLOWLABEL:
                        case IPV6_V6ONLY:
                        case IPV6_PORTRANGE:
                                switch (optname) {

                                case IPV6_RECVHOPOPTS:
                                        optval = OPTBIT(IN6P_HOPOPTS);
                                        break;

                                case IPV6_RECVDSTOPTS:
                                        optval = OPTBIT(IN6P_DSTOPTS);
                                        break;

                                case IPV6_RECVRTHDRDSTOPTS:
                                        optval = OPTBIT(IN6P_RTHDRDSTOPTS);
                                        break;

                                case IPV6_RECVPKTINFO:
                                        optval = OPTBIT(IN6P_PKTINFO);
                                        break;

                                case IPV6_RECVHOPLIMIT:
                                        optval = OPTBIT(IN6P_HOPLIMIT);
                                        break;

                                case IPV6_RECVRTHDR:
                                        optval = OPTBIT(IN6P_RTHDR);
                                        break;

                                case IPV6_RECVPATHMTU:
                                        optval = OPTBIT(IN6P_MTU);
                                        break;

                                case IPV6_RECVTCLASS:
                                        optval = OPTBIT(IN6P_TCLASS);
                                        break;

                                case IPV6_AUTOFLOWLABEL:
                                        optval = OPTBIT(IN6P_AUTOFLOWLABEL);
                                        break;


                                case IPV6_UNICAST_HOPS:
                                        optval = in6p->in6p_hops;
                                        break;

                                case IPV6_V6ONLY:
                                        optval = 1;
                                        break;

                                case IPV6_PORTRANGE:
                                    {
                                        int flags;
                                        flags = in6p->in6p_flags;
                                        if (flags & IN6P_HIGHPORT)
                                                optval = IPV6_PORTRANGE_HIGH;
                                        else if (flags & IN6P_LOWPORT)
                                                optval = IPV6_PORTRANGE_LOW;
                                        else
                                                optval = 0;
                                        break;
                                    }
                                }
                                soopt_from_kbuf(sopt, &optval,
                                        sizeof optval);
                                break;

                        case IPV6_PATHMTU:
                        {
                                u_long pmtu = 0;
                                struct ip6_mtuinfo mtuinfo;
                                struct route_in6 sro;

                                bzero(&sro, sizeof(sro));

                                if (!(so->so_state & SS_ISCONNECTED))
                                        return (ENOTCONN);
                                /*
                                 * XXX: we dot not consider the case of source
                                 * routing, or optional information to specify
                                 * the outgoing interface.
                                 */
                                error = ip6_getpmtu(&sro, NULL, NULL,
                                    &in6p->in6p_faddr, &pmtu, NULL);
                                if (sro.ro_rt)
                                        RTFREE(sro.ro_rt);
                                if (error)
                                        break;
                                if (pmtu > IPV6_MAXPACKET)
                                        pmtu = IPV6_MAXPACKET;

                                bzero(&mtuinfo, sizeof(mtuinfo));
                                mtuinfo.ip6m_mtu = (u_int32_t)pmtu;
                                optdata = (void *)&mtuinfo;
                                optdatalen = sizeof(mtuinfo);
                                soopt_from_kbuf(sopt, optdata,
                                    optdatalen);
                                break;
                        }

                        case IPV6_2292PKTINFO:
                        case IPV6_2292HOPLIMIT:
                        case IPV6_2292HOPOPTS:
                        case IPV6_2292RTHDR:
                        case IPV6_2292DSTOPTS:
                                if (optname == IPV6_2292HOPOPTS ||
                                    optname == IPV6_2292DSTOPTS ||
                                    !privileged)
                                        return (EPERM);
                                switch (optname) {
                                case IPV6_2292PKTINFO:
                                        optval = OPTBIT(IN6P_PKTINFO);
                                        break;
                                case IPV6_2292HOPLIMIT:
                                        optval = OPTBIT(IN6P_HOPLIMIT);
                                        break;
                                case IPV6_2292HOPOPTS:
                                        if (!privileged)
                                                return (EPERM);
                                        optval = OPTBIT(IN6P_HOPOPTS);
                                        break;
                                case IPV6_2292RTHDR:
                                        optval = OPTBIT(IN6P_RTHDR);
                                        break;
                                case IPV6_2292DSTOPTS:
                                        if (!privileged)
                                                return (EPERM);
                                        optval = OPTBIT(IN6P_DSTOPTS|IN6P_RTHDRDSTOPTS);
                                        break;
                                }
                                soopt_from_kbuf(sopt, &optval,
                                        sizeof optval);
                                break;

                        case IPV6_PKTINFO:
                        case IPV6_HOPOPTS:
                        case IPV6_RTHDR:
                        case IPV6_DSTOPTS:
                        case IPV6_RTHDRDSTOPTS:
                        case IPV6_NEXTHOP:
                        case IPV6_TCLASS:
                        case IPV6_DONTFRAG:
                        case IPV6_USE_MIN_MTU:
                        case IPV6_PREFER_TEMPADDR:
                                error = ip6_getpcbopt(in6p->in6p_outputopts,
                                    optname, sopt);
                                break;

                        case IPV6_MULTICAST_IF:
                        case IPV6_MULTICAST_HOPS:
                        case IPV6_MULTICAST_LOOP:
                        case IPV6_JOIN_GROUP:
                        case IPV6_LEAVE_GROUP:
                            {
                                struct mbuf *m;
                                error = ip6_getmoptions(sopt->sopt_name,
                                    in6p->in6p_moptions, &m);
                                if (error == 0) {
                                        soopt_from_kbuf(sopt,
                                            mtod(m, char *), m->m_len);
                                }
                                m_freem(m);
                            }
                                break;

                        case IPV6_FW_GET:
                          {
                                struct mbuf *m;
                                struct mbuf **mp = &m;

                                if (ip6_fw_ctl_ptr == NULL)
                                {
                                        return EINVAL;
                                }
                                error = (*ip6_fw_ctl_ptr)(optname, mp);
                                if (error == 0)
                                        error = soopt_from_mbuf(sopt, m); /* XXX */
                                if (error == 0 && m != NULL)
                                        m_freem(m);
                          }
                                break;

                        default:
                                error = ENOPROTOOPT;
                                break;
                        }
                        break;
                }
        } else {
                error = EINVAL;
        }
        return (error);
}

int
ip6_raw_ctloutput(struct socket *so, struct sockopt *sopt)
{
        int error = 0, optval, optlen;
        const int icmp6off = offsetof(struct icmp6_hdr, icmp6_cksum);
        struct in6pcb *in6p = sotoin6pcb(so);
        int level, op, optname;

        if (sopt) {
                level = sopt->sopt_level;
                op = sopt->sopt_dir;
                optname = sopt->sopt_name;
                optlen = sopt->sopt_valsize;
        } else
                panic("ip6_raw_ctloutput: arg soopt is NULL");

        if (level != IPPROTO_IPV6) {
                return (EINVAL);
        }

        switch (optname) {
        case IPV6_CHECKSUM:
                /*
                 * For ICMPv6 sockets, no modification allowed for checksum
                 * offset, permit "no change" values to help existing apps.
                 *
                 * RFC3542 says: "An attempt to set IPV6_CHECKSUM
                 * for an ICMPv6 socket will fail."
                 * The current behavior does not meet RFC3542.
                 */
                switch (op) {
                case SOPT_SET:
                        if (optlen != sizeof(int)) {
                                error = EINVAL;
                                break;
                        }
                        error = soopt_to_kbuf(sopt, &optval,
                                    sizeof optval, sizeof optval);
                        if (error)
                                break;
                        if ((optval % 2) != 0) {
                                /* the API assumes even offset values */
                                error = EINVAL;
                        } else if (so->so_proto->pr_protocol ==
                            IPPROTO_ICMPV6) {
                                if (optval != icmp6off)
                                        error = EINVAL;
                        } else
                                in6p->in6p_cksum = optval;
                        break;

                case SOPT_GET:
                        if (so->so_proto->pr_protocol == IPPROTO_ICMPV6)
                                optval = icmp6off;
                        else
                                optval = in6p->in6p_cksum;

                        soopt_from_kbuf(sopt, &optval, sizeof(optval));
                        break;

                default:
                        error = EINVAL;
                        break;
                }
                break;

        default:
                error = ENOPROTOOPT;
                break;
        }

        return (error);
}

/*
 * Set up IP6 options in pcb for insertion in output packets or
 * specifying behavior of outgoing packets.
 */
static int
ip6_pcbopts(struct ip6_pktopts **pktopt, struct mbuf *m,
    struct socket *so, struct sockopt *sopt)
{
        int priv = 0;
        struct ip6_pktopts *opt = *pktopt;
        int error = 0;

        /* turn off any old options. */
        if (opt) {
#ifdef DIAGNOSTIC
                if (opt->ip6po_pktinfo || opt->ip6po_nexthop ||
                    opt->ip6po_hbh || opt->ip6po_dest1 || opt->ip6po_dest2 ||
                    opt->ip6po_rhinfo.ip6po_rhi_rthdr)
                        kprintf("ip6_pcbopts: all specified options are cleared.\n");
#endif
                ip6_clearpktopts(opt, -1);
        } else
                opt = kmalloc(sizeof(*opt), M_IP6OPT, M_WAITOK);
        *pktopt = NULL;

        if (!m || m->m_len == 0) {
                /*
                 * Only turning off any previous options, regardless of
                 * whether the opt is just created or given.
                 */
                kfree(opt, M_IP6OPT);
                return (0);
        }

        /*  set options specified by user. */
        if ((error = ip6_setpktoptions(m, opt, NULL, so->so_proto->pr_protocol, priv)) != 0) {
                ip6_clearpktopts(opt, -1); /* XXX: discard all options */
                kfree(opt, M_IP6OPT);
                return (error);
        }
        *pktopt = opt;
        return (0);
}


/*
 * Below three functions are introduced by merge to RFC3542
 */

static int
ip6_getpcbopt(struct ip6_pktopts *pktopt, int optname, struct sockopt *sopt)
{
        void *optdata = NULL;
        int optdatalen = 0;
        struct ip6_ext *ip6e;
        int error = 0;
        struct in6_pktinfo null_pktinfo;
        int deftclass = 0, on;
        int defminmtu = IP6PO_MINMTU_MCASTONLY;
        int defpreftemp = IP6PO_TEMPADDR_SYSTEM;

        switch (optname) {
        case IPV6_PKTINFO:
                if (pktopt && pktopt->ip6po_pktinfo)
                        optdata = (void *)pktopt->ip6po_pktinfo;
                else {
                        /* XXX: we don't have to do this every time... */
                        bzero(&null_pktinfo, sizeof(null_pktinfo));
                        optdata = (void *)&null_pktinfo;
                }
                optdatalen = sizeof(struct in6_pktinfo);
                break;
        case IPV6_TCLASS:
                if (pktopt && pktopt->ip6po_tclass >= 0)
                        optdata = (void *)&pktopt->ip6po_tclass;
                else
                        optdata = (void *)&deftclass;
                optdatalen = sizeof(int);
                break;
        case IPV6_HOPOPTS:
                if (pktopt && pktopt->ip6po_hbh) {
                        optdata = (void *)pktopt->ip6po_hbh;
                        ip6e = (struct ip6_ext *)pktopt->ip6po_hbh;
                        optdatalen = (ip6e->ip6e_len + 1) << 3;
                }
                break;
        case IPV6_RTHDR:
                if (pktopt && pktopt->ip6po_rthdr) {
                        optdata = (void *)pktopt->ip6po_rthdr;
                        ip6e = (struct ip6_ext *)pktopt->ip6po_rthdr;
                        optdatalen = (ip6e->ip6e_len + 1) << 3;
                }
                break;
        case IPV6_RTHDRDSTOPTS:
                if (pktopt && pktopt->ip6po_dest1) {
                        optdata = (void *)pktopt->ip6po_dest1;
                        ip6e = (struct ip6_ext *)pktopt->ip6po_dest1;
                        optdatalen = (ip6e->ip6e_len + 1) << 3;
                }
                break;
        case IPV6_DSTOPTS:
                if (pktopt && pktopt->ip6po_dest2) {
                        optdata = (void *)pktopt->ip6po_dest2;
                        ip6e = (struct ip6_ext *)pktopt->ip6po_dest2;
                        optdatalen = (ip6e->ip6e_len + 1) << 3;
                }
                break;
        case IPV6_NEXTHOP:
                if (pktopt && pktopt->ip6po_nexthop) {
                        optdata = (void *)pktopt->ip6po_nexthop;
                        optdatalen = pktopt->ip6po_nexthop->sa_len;
                }
                break;
        case IPV6_USE_MIN_MTU:
                if (pktopt)
                        optdata = (void *)&pktopt->ip6po_minmtu;
                else
                        optdata = (void *)&defminmtu;
                optdatalen = sizeof(int);
                break;
        case IPV6_DONTFRAG:
                if (pktopt && ((pktopt->ip6po_flags) & IP6PO_DONTFRAG))
                        on = 1;
                else
                        on = 0;
                optdata = (void *)&on;
                optdatalen = sizeof(on);
                break;
        case IPV6_PREFER_TEMPADDR:
                if (pktopt)
                        optdata = (void *)&pktopt->ip6po_prefer_tempaddr;
                else
                        optdata = (void *)&defpreftemp;
                optdatalen = sizeof(int);
                break;
        default:                /* should not happen */
#ifdef DIAGNOSTIC
                panic("ip6_getpcbopt: unexpected option");
#endif
                return (ENOPROTOOPT);
        }

        soopt_from_kbuf(sopt, optdata, optdatalen);

        return (error);
}

/*
 * initialize ip6_pktopts.  beware that there are non-zero default values in
 * the struct.
 */

static int
ip6_pcbopt(int optname, u_char *buf, int len, struct ip6_pktopts **pktopt, int uproto)
{
        struct ip6_pktopts *opt;
        int priv =0;
        if (*pktopt == NULL) {
                *pktopt = kmalloc(sizeof(*opt), M_IP6OPT, M_WAITOK);
                init_ip6pktopts(*pktopt);
        }
        opt = *pktopt;

        return (ip6_setpktoption(optname, buf, len, opt, 1, 0, uproto, priv));
}

/*
 * initialize ip6_pktopts.  beware that there are non-zero default values in
 * the struct.
 */
void
init_ip6pktopts(struct ip6_pktopts *opt)
{

        bzero(opt, sizeof(*opt));
        opt->ip6po_hlim = -1;   /* -1 means default hop limit */
        opt->ip6po_tclass = -1; /* -1 means default traffic class */
        opt->ip6po_minmtu = IP6PO_MINMTU_MCASTONLY;
        opt->ip6po_prefer_tempaddr = IP6PO_TEMPADDR_SYSTEM;
}

void
ip6_clearpktopts(struct ip6_pktopts *pktopt, int optname)
{
        if (pktopt == NULL)
                return;

        if (optname == -1 || optname == IPV6_PKTINFO) {
                if (pktopt->ip6po_pktinfo)
                        kfree(pktopt->ip6po_pktinfo, M_IP6OPT);
                pktopt->ip6po_pktinfo = NULL;
        }
        if (optname == -1 || optname == IPV6_HOPLIMIT)
                pktopt->ip6po_hlim = -1;
        if (optname == -1 || optname == IPV6_TCLASS)
                pktopt->ip6po_tclass = -1;
        if (optname == -1 || optname == IPV6_NEXTHOP) {
                if (pktopt->ip6po_nextroute.ro_rt) {
                        RTFREE(pktopt->ip6po_nextroute.ro_rt);
                        pktopt->ip6po_nextroute.ro_rt = NULL;
                }
                if (pktopt->ip6po_nexthop)
                        kfree(pktopt->ip6po_nexthop, M_IP6OPT);
                pktopt->ip6po_nexthop = NULL;
        }
        if (optname == -1 || optname == IPV6_HOPOPTS) {
                if (pktopt->ip6po_hbh)
                        kfree(pktopt->ip6po_hbh, M_IP6OPT);
                pktopt->ip6po_hbh = NULL;
        }
        if (optname == -1 || optname == IPV6_RTHDRDSTOPTS) {
                if (pktopt->ip6po_dest1)
                        kfree(pktopt->ip6po_dest1, M_IP6OPT);
                pktopt->ip6po_dest1 = NULL;
        }
        if (optname == -1 || optname == IPV6_RTHDR) {
                if (pktopt->ip6po_rhinfo.ip6po_rhi_rthdr)
                        kfree(pktopt->ip6po_rhinfo.ip6po_rhi_rthdr, M_IP6OPT);
                pktopt->ip6po_rhinfo.ip6po_rhi_rthdr = NULL;
                if (pktopt->ip6po_route.ro_rt) {
                        RTFREE(pktopt->ip6po_route.ro_rt);
                        pktopt->ip6po_route.ro_rt = NULL;
                }
        }
        if (optname == -1 || optname == IPV6_DSTOPTS) {
                if (pktopt->ip6po_dest2)
                        kfree(pktopt->ip6po_dest2, M_IP6OPT);
                pktopt->ip6po_dest2 = NULL;
        }
}

#define PKTOPT_EXTHDRCPY(type) \
do {\
        if (src->type) {\
                int hlen = (((struct ip6_ext *)src->type)->ip6e_len + 1) << 3;\
                dst->type = kmalloc(hlen, M_IP6OPT, canwait);\
                if (dst->type == NULL)\
                        goto bad;\
                bcopy(src->type, dst->type, hlen);\
        }\
} while (0)

struct ip6_pktopts *
ip6_copypktopts(struct ip6_pktopts *src, int canwait)
{
        struct ip6_pktopts *dst;

        if (src == NULL) {
                kprintf("ip6_clearpktopts: invalid argument\n");
                return (NULL);
        }

        dst = kmalloc(sizeof(*dst), M_IP6OPT, canwait | M_ZERO);
        if (dst == NULL)
                return (NULL);

        dst->ip6po_hlim = src->ip6po_hlim;
        if (src->ip6po_pktinfo) {
                dst->ip6po_pktinfo = kmalloc(sizeof(*dst->ip6po_pktinfo),
                    M_IP6OPT, canwait);
                if (dst->ip6po_pktinfo == NULL)
                        goto bad;
                *dst->ip6po_pktinfo = *src->ip6po_pktinfo;
        }
        if (src->ip6po_nexthop) {
                dst->ip6po_nexthop = kmalloc(src->ip6po_nexthop->sa_len,
                    M_IP6OPT, canwait);
                if (dst->ip6po_nexthop == NULL)
                        goto bad;
                bcopy(src->ip6po_nexthop, dst->ip6po_nexthop,
                    src->ip6po_nexthop->sa_len);
        }
        PKTOPT_EXTHDRCPY(ip6po_hbh);
        PKTOPT_EXTHDRCPY(ip6po_dest1);
        PKTOPT_EXTHDRCPY(ip6po_dest2);
        PKTOPT_EXTHDRCPY(ip6po_rthdr); /* not copy the cached route */
        return (dst);

bad:
        if (dst->ip6po_pktinfo) kfree(dst->ip6po_pktinfo, M_IP6OPT);
        if (dst->ip6po_nexthop) kfree(dst->ip6po_nexthop, M_IP6OPT);
        if (dst->ip6po_hbh) kfree(dst->ip6po_hbh, M_IP6OPT);
        if (dst->ip6po_dest1) kfree(dst->ip6po_dest1, M_IP6OPT);
        if (dst->ip6po_dest2) kfree(dst->ip6po_dest2, M_IP6OPT);
        if (dst->ip6po_rthdr) kfree(dst->ip6po_rthdr, M_IP6OPT);
        kfree(dst, M_IP6OPT);
        return (NULL);
}

static int
copypktopts(struct ip6_pktopts *dst, struct ip6_pktopts *src, int canwait)
{
        if (dst == NULL || src == NULL)  {
#ifdef DIAGNOSTIC
                kprintf("ip6_clearpktopts: invalid argument\n");
#endif
                return (EINVAL);
        }

        dst->ip6po_hlim = src->ip6po_hlim;
        dst->ip6po_tclass = src->ip6po_tclass;
        dst->ip6po_flags = src->ip6po_flags;
        if (src->ip6po_pktinfo) {
                dst->ip6po_pktinfo = kmalloc(sizeof(*dst->ip6po_pktinfo),
                    M_IP6OPT, canwait);
                if (dst->ip6po_pktinfo == NULL)
                        goto bad;
                *dst->ip6po_pktinfo = *src->ip6po_pktinfo;
        }
        if (src->ip6po_nexthop) {
                dst->ip6po_nexthop = kmalloc(src->ip6po_nexthop->sa_len,
                    M_IP6OPT, canwait);
                if (dst->ip6po_nexthop == NULL)
                        goto bad;
                bcopy(src->ip6po_nexthop, dst->ip6po_nexthop,
                    src->ip6po_nexthop->sa_len);
        }
        PKTOPT_EXTHDRCPY(ip6po_hbh);
        PKTOPT_EXTHDRCPY(ip6po_dest1);
        PKTOPT_EXTHDRCPY(ip6po_dest2);
        PKTOPT_EXTHDRCPY(ip6po_rthdr); /* not copy the cached route */
        return (0);

  bad:
        ip6_clearpktopts(dst, -1);
        return (ENOBUFS);
}
#undef PKTOPT_EXTHDRCPY

void
ip6_freepcbopts(struct ip6_pktopts *pktopt)
{
        if (pktopt == NULL)
                return;

        ip6_clearpktopts(pktopt, -1);

        kfree(pktopt, M_IP6OPT);
}

/*
 * Set the IP6 multicast options in response to user setsockopt().
 */
static int
ip6_setmoptions(int optname, struct ip6_moptions **im6op, struct mbuf *m)
{
        int error = 0;
        u_int loop, ifindex;
        struct ipv6_mreq *mreq;
        struct ifnet *ifp;
        struct ip6_moptions *im6o = *im6op;
        struct route_in6 ro;
        struct sockaddr_in6 *dst;
        struct in6_multi_mship *imm;

        if (im6o == NULL) {
                /*
                 * No multicast option buffer attached to the pcb;
                 * allocate one and initialize to default values.
                 */
                im6o = (struct ip6_moptions *)
                        kmalloc(sizeof(*im6o), M_IPMOPTS, M_WAITOK);

                *im6op = im6o;
                im6o->im6o_multicast_ifp = NULL;
                im6o->im6o_multicast_hlim = ip6_defmcasthlim;
                im6o->im6o_multicast_loop = IPV6_DEFAULT_MULTICAST_LOOP;
                LIST_INIT(&im6o->im6o_memberships);
        }

        switch (optname) {

        case IPV6_MULTICAST_IF:
                /*
                 * Select the interface for outgoing multicast packets.
                 */
                if (m == NULL || m->m_len != sizeof(u_int)) {
                        error = EINVAL;
                        break;
                }
                bcopy(mtod(m, u_int *), &ifindex, sizeof(ifindex));
                if (ifindex < 0 || if_index < ifindex) {
                        error = ENXIO;  /* XXX EINVAL? */
                        break;
                }
                ifp = ifindex2ifnet[ifindex];
                if (ifp == NULL || !(ifp->if_flags & IFF_MULTICAST)) {
                        error = EADDRNOTAVAIL;
                        break;
                }
                im6o->im6o_multicast_ifp = ifp;
                break;

        case IPV6_MULTICAST_HOPS:
            {
                /*
                 * Set the IP6 hoplimit for outgoing multicast packets.
                 */
                int optval;
                if (m == NULL || m->m_len != sizeof(int)) {
                        error = EINVAL;
                        break;
                }
                bcopy(mtod(m, u_int *), &optval, sizeof(optval));
                if (optval < -1 || optval >= 256)
                        error = EINVAL;
                else if (optval == -1)
                        im6o->im6o_multicast_hlim = ip6_defmcasthlim;
                else
                        im6o->im6o_multicast_hlim = optval;
                break;
            }

        case IPV6_MULTICAST_LOOP:
                /*
                 * Set the loopback flag for outgoing multicast packets.
                 * Must be zero or one.
                 */
                if (m == NULL || m->m_len != sizeof(u_int)) {
                        error = EINVAL;
                        break;
                }
                bcopy(mtod(m, u_int *), &loop, sizeof(loop));
                if (loop > 1) {
                        error = EINVAL;
                        break;
                }
                im6o->im6o_multicast_loop = loop;
                break;

        case IPV6_JOIN_GROUP:
                /*
                 * Add a multicast group membership.
                 * Group must be a valid IP6 multicast address.
                 */
                if (m == NULL || m->m_len != sizeof(struct ipv6_mreq)) {
                        error = EINVAL;
                        break;
                }
                mreq = mtod(m, struct ipv6_mreq *);
                if (IN6_IS_ADDR_UNSPECIFIED(&mreq->ipv6mr_multiaddr)) {
                        /*
                         * We use the unspecified address to specify to accept
                         * all multicast addresses. Only super user is allowed
                         * to do this.
                         */
                        if (caps_priv_check_self(SYSCAP_RESTRICTEDROOT)) {
                                error = EACCES;
                                break;
                        }
                } else if (!IN6_IS_ADDR_MULTICAST(&mreq->ipv6mr_multiaddr)) {
                        error = EINVAL;
                        break;
                }

                /*
                 * If the interface is specified, validate it.
                 */
                if (mreq->ipv6mr_interface < 0
                 || if_index < mreq->ipv6mr_interface) {
                        error = ENXIO;  /* XXX EINVAL? */
                        break;
                }
                /*
                 * If no interface was explicitly specified, choose an
                 * appropriate one according to the given multicast address.
                 */
                if (mreq->ipv6mr_interface == 0) {
                        /*
                         * If the multicast address is in node-local scope,
                         * the interface should be a loopback interface.
                         * Otherwise, look up the routing table for the
                         * address, and choose the outgoing interface.
                         *   XXX: is it a good approach?
                         */
                        if (IN6_IS_ADDR_MC_INTFACELOCAL(&mreq->ipv6mr_multiaddr)) {
                                ifp = loif;
                        } else {
                                ro.ro_rt = NULL;
                                dst = (struct sockaddr_in6 *)&ro.ro_dst;
                                bzero(dst, sizeof(*dst));
                                dst->sin6_len = sizeof(struct sockaddr_in6);
                                dst->sin6_family = AF_INET6;
                                dst->sin6_addr = mreq->ipv6mr_multiaddr;
                                rtalloc((struct route *)&ro);
                                if (ro.ro_rt == NULL) {
                                        error = EADDRNOTAVAIL;
                                        break;
                                }
                                ifp = ro.ro_rt->rt_ifp;
                                rtfree(ro.ro_rt);
                        }
                } else
                        ifp = ifindex2ifnet[mreq->ipv6mr_interface];

                /*
                 * See if we found an interface, and confirm that it
                 * supports multicast
                 */
                if (ifp == NULL || !(ifp->if_flags & IFF_MULTICAST)) {
                        error = EADDRNOTAVAIL;
                        break;
                }
                /*
                 * Put interface index into the multicast address,
                 * if the address has link-local scope.
                 */
                if (IN6_IS_ADDR_MC_LINKLOCAL(&mreq->ipv6mr_multiaddr)) {
                        mreq->ipv6mr_multiaddr.s6_addr16[1]
                                = htons(mreq->ipv6mr_interface);
                }
                /*
                 * See if the membership already exists.
                 */
                for (imm = im6o->im6o_memberships.lh_first;
                     imm != NULL; imm = imm->i6mm_chain.le_next)
                        if (imm->i6mm_maddr->in6m_ifp == ifp &&
                            IN6_ARE_ADDR_EQUAL(&imm->i6mm_maddr->in6m_addr,
                                               &mreq->ipv6mr_multiaddr))
                                break;
                if (imm != NULL) {
                        error = EADDRINUSE;
                        break;
                }
                /*
                 * Everything looks good; add a new record to the multicast
                 * address list for the given interface.
                 */
                imm = kmalloc(sizeof(*imm), M_IPMADDR, M_WAITOK);
                if ((imm->i6mm_maddr =
                     in6_addmulti(&mreq->ipv6mr_multiaddr, ifp, &error)) == NULL) {
                        kfree(imm, M_IPMADDR);
                        break;
                }
                LIST_INSERT_HEAD(&im6o->im6o_memberships, imm, i6mm_chain);
                break;

        case IPV6_LEAVE_GROUP:
                /*
                 * Drop a multicast group membership.
                 * Group must be a valid IP6 multicast address.
                 */
                if (m == NULL || m->m_len != sizeof(struct ipv6_mreq)) {
                        error = EINVAL;
                        break;
                }
                mreq = mtod(m, struct ipv6_mreq *);
                if (IN6_IS_ADDR_UNSPECIFIED(&mreq->ipv6mr_multiaddr)) {
                        if (caps_priv_check_self(SYSCAP_RESTRICTEDROOT)) {
                                error = EACCES;
                                break;
                        }
                } else if (!IN6_IS_ADDR_MULTICAST(&mreq->ipv6mr_multiaddr)) {
                        error = EINVAL;
                        break;
                }
                /*
                 * If an interface address was specified, get a pointer
                 * to its ifnet structure.
                 */
                if (mreq->ipv6mr_interface < 0
                 || if_index < mreq->ipv6mr_interface) {
                        error = ENXIO;  /* XXX EINVAL? */
                        break;
                }
                ifp = ifindex2ifnet[mreq->ipv6mr_interface];
                /*
                 * Put interface index into the multicast address,
                 * if the address has link-local scope.
                 */
                if (IN6_IS_ADDR_MC_LINKLOCAL(&mreq->ipv6mr_multiaddr)) {
                        mreq->ipv6mr_multiaddr.s6_addr16[1]
                                = htons(mreq->ipv6mr_interface);
                }

                /*
                 * Find the membership in the membership list.
                 */
                for (imm = im6o->im6o_memberships.lh_first;
                     imm != NULL; imm = imm->i6mm_chain.le_next) {
                        if ((ifp == NULL || imm->i6mm_maddr->in6m_ifp == ifp) &&
                            IN6_ARE_ADDR_EQUAL(&imm->i6mm_maddr->in6m_addr,
                            &mreq->ipv6mr_multiaddr))
                                break;
                }
                if (imm == NULL) {
                        /* Unable to resolve interface */
                        error = EADDRNOTAVAIL;
                        break;
                }
                /*
                 * Give up the multicast address record to which the
                 * membership points.
                 */
                LIST_REMOVE(imm, i6mm_chain);
                in6_delmulti(imm->i6mm_maddr);
                kfree(imm, M_IPMADDR);
                break;

        default:
                error = EOPNOTSUPP;
                break;
        }

        /*
         * If all options have default values, no need to keep the mbuf.
         */
        if (im6o->im6o_multicast_ifp == NULL &&
            im6o->im6o_multicast_hlim == ip6_defmcasthlim &&
            im6o->im6o_multicast_loop == IPV6_DEFAULT_MULTICAST_LOOP &&
            im6o->im6o_memberships.lh_first == NULL) {
                kfree(*im6op, M_IPMOPTS);
                *im6op = NULL;
        }

        return (error);
}

/*
 * Return the IP6 multicast options in response to user getsockopt().
 */
static int
ip6_getmoptions(int optname, struct ip6_moptions *im6o, struct mbuf **mp)
{
        u_int *hlim, *loop, *ifindex;

        *mp = m_get(M_WAITOK, MT_HEADER);               /* XXX */

        switch (optname) {

        case IPV6_MULTICAST_IF:
                ifindex = mtod(*mp, u_int *);
                (*mp)->m_len = sizeof(u_int);
                if (im6o == NULL || im6o->im6o_multicast_ifp == NULL)
                        *ifindex = 0;
                else
                        *ifindex = im6o->im6o_multicast_ifp->if_index;
                return (0);

        case IPV6_MULTICAST_HOPS:
                hlim = mtod(*mp, u_int *);
                (*mp)->m_len = sizeof(u_int);
                if (im6o == NULL)
                        *hlim = ip6_defmcasthlim;
                else
                        *hlim = im6o->im6o_multicast_hlim;
                return (0);

        case IPV6_MULTICAST_LOOP:
                loop = mtod(*mp, u_int *);
                (*mp)->m_len = sizeof(u_int);
                if (im6o == NULL)
                        *loop = ip6_defmcasthlim;
                else
                        *loop = im6o->im6o_multicast_loop;
                return (0);

        default:
                return (EOPNOTSUPP);
        }
}

/*
 * Discard the IP6 multicast options.
 */
void
ip6_freemoptions(struct ip6_moptions *im6o)
{
        struct in6_multi_mship *imm;

        if (im6o == NULL)
                return;

        while ((imm = im6o->im6o_memberships.lh_first) != NULL) {
                LIST_REMOVE(imm, i6mm_chain);
                if (imm->i6mm_maddr)
                        in6_delmulti(imm->i6mm_maddr);
                kfree(imm, M_IPMADDR);
        }
        kfree(im6o, M_IPMOPTS);
}

/*
 * Set a particular packet option, as a sticky option or an ancillary data
 * item.  "len" can be 0 only when it's a sticky option.
 * We have 4 cases of combination of "sticky" and "cmsg":
 * "sticky=0, cmsg=0": impossible
 * "sticky=0, cmsg=1": RFC2292 or RFC3542 ancillary data
 * "sticky=1, cmsg=0": RFC3542 socket option
 * "sticky=1, cmsg=1": RFC2292 socket option
 */
static int
ip6_setpktoption(int optname, u_char *buf, int len, struct ip6_pktopts *opt,
     int sticky, int cmsg, int uproto, int priv)
{
        int minmtupolicy, preftemp;
        //int error;

        if (!sticky && !cmsg) {
                kprintf("ip6_setpktoption: impossible case\n");
                return (EINVAL);
        }

        /*
         * IPV6_2292xxx is for backward compatibility to RFC2292, and should
         * not be specified in the context of RFC3542.  Conversely,
         * RFC3542 types should not be specified in the context of RFC2292.
         */
        if (!cmsg) {
                switch (optname) {
                case IPV6_2292PKTINFO:
                case IPV6_2292HOPLIMIT:
                case IPV6_2292NEXTHOP:
                case IPV6_2292HOPOPTS:
                case IPV6_2292DSTOPTS:
                case IPV6_2292RTHDR:
                case IPV6_2292PKTOPTIONS:
                        return (ENOPROTOOPT);
                }
        }
        if (sticky && cmsg) {
                switch (optname) {
                case IPV6_PKTINFO:
                case IPV6_HOPLIMIT:
                case IPV6_NEXTHOP:
                case IPV6_HOPOPTS:
                case IPV6_DSTOPTS:
                case IPV6_RTHDRDSTOPTS:
                case IPV6_RTHDR:
                case IPV6_USE_MIN_MTU:
                case IPV6_DONTFRAG:
                case IPV6_TCLASS:
                case IPV6_PREFER_TEMPADDR: /* XXX: not an RFC3542 option */
                        return (ENOPROTOOPT);
                }
        }

        switch (optname) {
        case IPV6_2292PKTINFO:
        case IPV6_PKTINFO:
        {
                struct in6_pktinfo *pktinfo;
                if (len != sizeof(struct in6_pktinfo))
                        return (EINVAL);
                pktinfo = (struct in6_pktinfo *)buf;

                /*
                 * An application can clear any sticky IPV6_PKTINFO option by
                 * doing a "regular" setsockopt with ipi6_addr being
                 * in6addr_any and ipi6_ifindex being zero.
                 * [RFC 3542, Section 6]
                 */
                if (optname == IPV6_PKTINFO && opt->ip6po_pktinfo &&
                    pktinfo->ipi6_ifindex == 0 &&
                    IN6_IS_ADDR_UNSPECIFIED(&pktinfo->ipi6_addr)) {
                        ip6_clearpktopts(opt, optname);
                        break;
                }

                if (uproto == IPPROTO_TCP && optname == IPV6_PKTINFO &&
                    sticky && !IN6_IS_ADDR_UNSPECIFIED(&pktinfo->ipi6_addr)) {
                        return (EINVAL);
                }

                /* validate the interface index if specified. */
                if (pktinfo->ipi6_ifindex > if_index ||
                    pktinfo->ipi6_ifindex < 0) {
                         return (ENXIO);
                }
                /*
                 * Check if the requested source address is indeed a
                 * unicast address assigned to the node, and can be
                 * used as the packet's source address.
                 */
                if (opt->ip6po_pktinfo != NULL &&
                    !IN6_IS_ADDR_UNSPECIFIED(&opt->ip6po_pktinfo->ipi6_addr)) {
                        struct in6_ifaddr *ia6;
                        struct sockaddr_in6 sin6;

                        bzero(&sin6, sizeof(sin6));
                        sin6.sin6_len = sizeof(sin6);
                        sin6.sin6_family = AF_INET6;
                        sin6.sin6_addr =
                        opt->ip6po_pktinfo->ipi6_addr;
                        ia6 = (struct in6_ifaddr *)ifa_ifwithaddr(sin6tosa(&sin6));
                        if (ia6 == NULL ||
                                (ia6->ia6_flags & (IN6_IFF_ANYCAST |
                                        IN6_IFF_NOTREADY)) != 0)
                        return (EADDRNOTAVAIL);
                }

                /*
                 * We store the address anyway, and let in6_selectsrc()
                 * validate the specified address.  This is because ipi6_addr
                 * may not have enough information about its scope zone, and
                 * we may need additional information (such as outgoing
                 * interface or the scope zone of a destination address) to
                 * disambiguate the scope.
                 * XXX: the delay of the validation may confuse the
                 * application when it is used as a sticky option.
                 */
                if (opt->ip6po_pktinfo == NULL) {
                        opt->ip6po_pktinfo = kmalloc(sizeof(*pktinfo),
                            M_IP6OPT, M_NOWAIT);
                        if (opt->ip6po_pktinfo == NULL)
                                return (ENOBUFS);
                }
                bcopy(pktinfo, opt->ip6po_pktinfo, sizeof(*pktinfo));
                break;
        }

        case IPV6_2292HOPLIMIT:
        case IPV6_HOPLIMIT:
        {
                int *hlimp;

                /*
                 * RFC 3542 deprecated the usage of sticky IPV6_HOPLIMIT
                 * to simplify the ordering among hoplimit options.
                 */
                if (optname == IPV6_HOPLIMIT && sticky)
                        return (ENOPROTOOPT);

                if (len != sizeof(int))
                        return (EINVAL);
                hlimp = (int *)buf;
                if (*hlimp < -1 || *hlimp > 255)
                        return (EINVAL);

                opt->ip6po_hlim = *hlimp;
                break;
        }

        case IPV6_TCLASS:
        {
                int tclass;

                if (len != sizeof(int))
                        return (EINVAL);
                tclass = *(int *)buf;
                if (tclass < -1 || tclass > 255)
                        return (EINVAL);

                opt->ip6po_tclass = tclass;
                break;
        }

        case IPV6_2292NEXTHOP:
        case IPV6_NEXTHOP:
                if (!priv)
                        return (EPERM);

                if (len == 0) { /* just remove the option */
                        ip6_clearpktopts(opt, IPV6_NEXTHOP);
                        break;
                }

                /* check if cmsg_len is large enough for sa_len */
                if (len < sizeof(struct sockaddr) || len < *buf)
                        return (EINVAL);

                switch (((struct sockaddr *)buf)->sa_family) {
                case AF_INET6:
                {
                        struct sockaddr_in6 *sa6 = (struct sockaddr_in6 *)buf;
                        //int error;

                        if (sa6->sin6_len != sizeof(struct sockaddr_in6))
                                return (EINVAL);

                        if (IN6_IS_ADDR_UNSPECIFIED(&sa6->sin6_addr) ||
                            IN6_IS_ADDR_MULTICAST(&sa6->sin6_addr)) {
                                return (EINVAL);
                        }
                        break;
                }
                case AF_LINK:   /* should eventually be supported */
                default:
                        return (EAFNOSUPPORT);
                }

                /* turn off the previous option, then set the new option. */
                ip6_clearpktopts(opt, IPV6_NEXTHOP);
                opt->ip6po_nexthop = kmalloc(*buf, M_IP6OPT, M_NOWAIT);
                if (opt->ip6po_nexthop == NULL)
                        return (ENOBUFS);
                bcopy(buf, opt->ip6po_nexthop, *buf);
                break;

        case IPV6_2292HOPOPTS:
        case IPV6_HOPOPTS:
        {
                struct ip6_hbh *hbh;
                int hbhlen;

                /*
                 * XXX: We don't allow a non-privileged user to set ANY HbH
                 * options, since per-option restriction has too much
                 * overhead.
                 */
                if (!priv)
                        return (EPERM);
                if (len == 0) {
                        ip6_clearpktopts(opt, IPV6_HOPOPTS);
                        break;  /* just remove the option */
                }

                /* message length validation */
                if (len < sizeof(struct ip6_hbh))
                        return (EINVAL);
                hbh = (struct ip6_hbh *)buf;
                hbhlen = (hbh->ip6h_len + 1) << 3;
                if (len != hbhlen)
                        return (EINVAL);

                /* turn off the previous option, then set the new option. */
                ip6_clearpktopts(opt, IPV6_HOPOPTS);
                opt->ip6po_hbh = kmalloc(hbhlen, M_IP6OPT, M_NOWAIT);
                if (opt->ip6po_hbh == NULL)
                        return (ENOBUFS);
                bcopy(hbh, opt->ip6po_hbh, hbhlen);

                break;
        }

        case IPV6_2292DSTOPTS:
        case IPV6_DSTOPTS:
        case IPV6_RTHDRDSTOPTS:
        {
                struct ip6_dest *dest, **newdest = NULL;
                int destlen;
                if (!priv)
                        return (EPERM);

                if (len == 0) {
                        ip6_clearpktopts(opt, optname);
                        break;  /* just remove the option */
                }

                /* message length validation */
                if (len < sizeof(struct ip6_dest))
                        return (EINVAL);
                dest = (struct ip6_dest *)buf;
                destlen = (dest->ip6d_len + 1) << 3;
                if (len != destlen)
                        return (EINVAL);

                /*
                 * Determine the position that the destination options header
                 * should be inserted; before or after the routing header.
                 */
                switch (optname) {
                case IPV6_2292DSTOPTS:
                        /*
                         * The old advacned API is ambiguous on this point.
                         * Our approach is to determine the position based
                         * according to the existence of a routing header.
                         * Note, however, that this depends on the order of the
                         * extension headers in the ancillary data; the 1st
                         * part of the destination options header must appear
                         * before the routing header in the ancillary data,
                         * too.
                         * RFC3542 solved the ambiguity by introducing
                         * separate ancillary data or option types.
                         */
                        if (opt->ip6po_rthdr == NULL)
                                newdest = &opt->ip6po_dest1;
                        else
                                newdest = &opt->ip6po_dest2;
                        break;
                case IPV6_RTHDRDSTOPTS:
                        newdest = &opt->ip6po_dest1;
                        break;
                case IPV6_DSTOPTS:
                        newdest = &opt->ip6po_dest2;
                        break;
                }

                /* turn off the previous option, then set the new option. */
                ip6_clearpktopts(opt, optname);
                *newdest = kmalloc(destlen, M_IP6OPT, M_NOWAIT);
                if (*newdest == NULL)
                        return (ENOBUFS);
                bcopy(dest, *newdest, destlen);

                break;
        }

        case IPV6_2292RTHDR:
        case IPV6_RTHDR:
        {
                struct ip6_rthdr *rth;
                int rthlen;

                if (len == 0) {
                        ip6_clearpktopts(opt, IPV6_RTHDR);
                        break;  /* just remove the option */
                }

                /* message length validation */
                if (len < sizeof(struct ip6_rthdr))
                        return (EINVAL);
                rth = (struct ip6_rthdr *)buf;
                rthlen = (rth->ip6r_len + 1) << 3;
                if (len != rthlen)
                        return (EINVAL);

                switch (rth->ip6r_type) {
                default:
                        return (EINVAL);        /* not supported */
                }

                /* turn off the previous option */
                ip6_clearpktopts(opt, IPV6_RTHDR);
                opt->ip6po_rthdr = kmalloc(rthlen, M_IP6OPT, M_NOWAIT);
                if (opt->ip6po_rthdr == NULL)
                        return (ENOBUFS);
                bcopy(rth, opt->ip6po_rthdr, rthlen);

                break;
        }

        case IPV6_USE_MIN_MTU:
                if (len != sizeof(int))
                        return (EINVAL);
                minmtupolicy = *(int *)buf;
                if (minmtupolicy != IP6PO_MINMTU_MCASTONLY &&
                    minmtupolicy != IP6PO_MINMTU_DISABLE &&
                    minmtupolicy != IP6PO_MINMTU_ALL) {
                        return (EINVAL);
                }
                opt->ip6po_minmtu = minmtupolicy;
                break;

        case IPV6_DONTFRAG:
                if (len != sizeof(int))
                        return (EINVAL);

                if (uproto == IPPROTO_TCP || *(int *)buf == 0) {
                        /*
                         * we ignore this option for TCP sockets.
                         * (RFC3542 leaves this case unspecified.)
                         */
                        opt->ip6po_flags &= ~IP6PO_DONTFRAG;
                } else
                        opt->ip6po_flags |= IP6PO_DONTFRAG;
                break;

        case IPV6_PREFER_TEMPADDR:
                if (len != sizeof(int))
                        return (EINVAL);
                preftemp = *(int *)buf;
                if (preftemp != IP6PO_TEMPADDR_SYSTEM &&
                    preftemp != IP6PO_TEMPADDR_NOTPREFER &&
                    preftemp != IP6PO_TEMPADDR_PREFER) {
                        return (EINVAL);
                }
                opt->ip6po_prefer_tempaddr = preftemp;
                break;

        default:
                return (ENOPROTOOPT);
        } /* end of switch */

        return (0);
}


/*
 * Set IPv6 outgoing packet options based on advanced API.
 */
int
ip6_setpktoptions(struct mbuf *control, struct ip6_pktopts *opt,
    struct ip6_pktopts *stickyopt, int uproto, int priv)
{
        struct cmsghdr *cm = NULL;

        if (control == NULL || opt == NULL)
                return (EINVAL);

        init_ip6pktopts(opt);

        /*
         * XXX: Currently, we assume all the optional information is stored
         * in a single mbuf.
         */
        if (stickyopt) {
                int error;

                /*
                 * If stickyopt is provided, make a local copy of the options
                 * for this particular packet, then override them by ancillary
                 * objects.
                 * XXX: copypktopts() does not copy the cached route to a next
                 * hop (if any).  This is not very good in terms of efficiency,
                 * but we can allow this since this option should be rarely
                 * used.
                 */
                if ((error = copypktopts(opt, stickyopt, M_NOWAIT)) != 0)
                        return (error);
        }

        /*
         * XXX: Currently, we assume all the optional information is stored
         * in a single mbuf.
         */
        if (control->m_next)
                return (EINVAL);

        for (;;) {
                int error;

                if (control->m_len == 0)
                        break;
                if (control->m_len < sizeof(*cm))
                        return EINVAL;

                cm = mtod(control, struct cmsghdr *);
                if (cm->cmsg_len == 0 || cm->cmsg_len > control->m_len)
                        return (EINVAL);
                if (cm->cmsg_level == IPPROTO_IPV6) {
                        error = ip6_setpktoption(cm->cmsg_type, CMSG_DATA(cm),
                                                 cm->cmsg_len - CMSG_LEN(0),
                                                 opt, 0, 1, uproto, priv);
                        if (error)
                                return (error);
                }

                /*
                 * The cmsg fit, but the aligned step for the next one might
                 * not.  Check the case and terminate normally (allows the
                 * cmsg_len to not be aligned).
                 */
                if (CMSG_ALIGN(cm->cmsg_len) >= control->m_len) {
                        control->m_data += control->m_len;
                        control->m_len = 0;
                        break;
                }
                control->m_data += CMSG_ALIGN(cm->cmsg_len);
                control->m_len -= CMSG_ALIGN(cm->cmsg_len);
        }

        return (0);
}

/*
 * Routine called from ip6_output() to loop back a copy of an IP6 multicast
 * packet to the input queue of a specified interface.  Note that this
 * calls the output routine of the loopback "driver", but with an interface
 * pointer that might NOT be loif -- easier than replicating that code here.
 */
void
ip6_mloopback(struct ifnet *ifp, struct mbuf *m, struct sockaddr_in6 *dst)
{
        struct mbuf *copym;
        struct ip6_hdr *ip6;

        copym = m_copym(m, 0, M_COPYALL, M_NOWAIT);
        if (copym == NULL)
                return;

        /*
         * Make sure to deep-copy IPv6 header portion in case the data
         * is in an mbuf cluster, so that we can safely override the IPv6
         * header portion later.
         */
        if ((copym->m_flags & M_EXT) != 0 ||
            copym->m_len < sizeof(struct ip6_hdr)) {
                copym = m_pullup(copym, sizeof(struct ip6_hdr));
                if (copym == NULL)
                        return;
        }

#ifdef DIAGNOSTIC
        if (copym->m_len < sizeof(*ip6)) {
                m_freem(copym);
                return;
        }
#endif

        ip6 = mtod(copym, struct ip6_hdr *);
        /*
         * 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);

        if_simloop(ifp, copym, dst->sin6_family, 0);
}

/*
 * Separate the IPv6 header from the payload into its own mbuf.
 *
 * Returns the new mbuf chain or the original mbuf if no payload.
 * Returns NULL if can't allocate new mbuf for header.
 */
static struct mbuf *
ip6_splithdr(struct mbuf *m)
{
        struct mbuf *mh;

        if (m->m_len <= sizeof(struct ip6_hdr))         /* no payload */
                return (m);

        MGETHDR(mh, M_NOWAIT, MT_HEADER);
        if (mh == NULL)
                return (NULL);
        mh->m_len = sizeof(struct ip6_hdr);
        M_MOVE_PKTHDR(mh, m);
        MH_ALIGN(mh, sizeof(struct ip6_hdr));
        bcopy(mtod(m, caddr_t), mtod(mh, caddr_t), sizeof(struct ip6_hdr));
        m->m_data += sizeof(struct ip6_hdr);
        m->m_len -= sizeof(struct ip6_hdr);
        mh->m_next = m;
        return (mh);
}

/*
 * Compute IPv6 extension header length.
 */
int
ip6_optlen(struct in6pcb *in6p)
{
        int len;

        if (!in6p->in6p_outputopts)
                return 0;

        len = 0;
#define elen(x) \
    (((struct ip6_ext *)(x)) ? (((struct ip6_ext *)(x))->ip6e_len + 1) << 3 : 0)

        len += elen(in6p->in6p_outputopts->ip6po_hbh);
        if (in6p->in6p_outputopts->ip6po_rthdr)
                /* dest1 is valid with rthdr only */
                len += elen(in6p->in6p_outputopts->ip6po_dest1);
        len += elen(in6p->in6p_outputopts->ip6po_rthdr);
        len += elen(in6p->in6p_outputopts->ip6po_dest2);
        return len;
#undef elen
}