tcp: Don't assume power of 2 netisrs.

[dragonfly.git] / sys / net / rtsock.c

/*
 * Copyright (c) 2004, 2005 The DragonFly Project.  All rights reserved.
 *
 * This code is derived from software contributed to The DragonFly Project
 * by Jeffrey M. Hsu.
 *
 * 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 DragonFly 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 COPYRIGHT HOLDERS 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
 * COPYRIGHT HOLDERS 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) 1988, 1991, 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.
 *
 *      @(#)rtsock.c    8.7 (Berkeley) 10/12/95
 * $FreeBSD: src/sys/net/rtsock.c,v 1.44.2.11 2002/12/04 14:05:41 ru Exp $
 */

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/sysctl.h>
#include <sys/proc.h>
#include <sys/priv.h>
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/protosw.h>
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/domain.h>

#include <sys/thread2.h>
#include <sys/socketvar2.h>

#include <net/if.h>
#include <net/route.h>
#include <net/raw_cb.h>
#include <net/netmsg2.h>
#include <net/netisr2.h>

MALLOC_DEFINE(M_RTABLE, "routetbl", "routing tables");

static struct route_cb {
        int     ip_count;
        int     ip6_count;
        int     ns_count;
        int     any_count;
} route_cb;

static const struct sockaddr route_src = { 2, PF_ROUTE, };

struct walkarg {
        int     w_tmemsize;
        int     w_op, w_arg;
        void    *w_tmem;
        struct sysctl_req *w_req;
};

#ifndef RTTABLE_DUMP_MSGCNT_MAX
/* Should be large enough for dupkeys */
#define RTTABLE_DUMP_MSGCNT_MAX         64
#endif

struct rttable_walkarg {
        int     w_op;
        int     w_arg;
        int     w_bufsz;
        void    *w_buf;

        int     w_buflen;

        const char *w_key;
        const char *w_mask;

        struct sockaddr_storage w_key0;
        struct sockaddr_storage w_mask0;
};

struct netmsg_rttable_walk {
        struct netmsg_base      base;
        int                     af;
        struct rttable_walkarg  *w;
};

static struct mbuf *
                rt_msg_mbuf (int, struct rt_addrinfo *);
static void     rt_msg_buffer (int, struct rt_addrinfo *, void *buf, int len);
static int      rt_msgsize(int type, const struct rt_addrinfo *rtinfo);
static int      rt_xaddrs (char *, char *, struct rt_addrinfo *);
static int      sysctl_rttable(int af, struct sysctl_req *req, int op, int arg);
static int      sysctl_iflist (int af, struct walkarg *w);
static int      route_output(struct mbuf *, struct socket *, ...);
static void     rt_setmetrics (u_long, struct rt_metrics *,
                               struct rt_metrics *);

/*
 * It really doesn't make any sense at all for this code to share much
 * with raw_usrreq.c, since its functionality is so restricted.  XXX
 */
static void
rts_abort(netmsg_t msg)
{
        crit_enter();
        raw_usrreqs.pru_abort(msg);
        /* msg invalid now */
        crit_exit();
}

/* pru_accept is EOPNOTSUPP */

static void
rts_attach(netmsg_t msg)
{
        struct socket *so = msg->base.nm_so;
        struct pru_attach_info *ai = msg->attach.nm_ai;
        struct rawcb *rp;
        int proto = msg->attach.nm_proto;
        int error;

        crit_enter();
        if (sotorawcb(so) != NULL) {
                error = EISCONN;
                goto done;
        }

        rp = kmalloc(sizeof *rp, M_PCB, M_WAITOK | M_ZERO);

        /*
         * The critical section is necessary to block protocols from sending
         * error notifications (like RTM_REDIRECT or RTM_LOSING) while
         * this PCB is extant but incompletely initialized.
         * Probably we should try to do more of this work beforehand and
         * eliminate the critical section.
         */
        so->so_pcb = rp;
        soreference(so);        /* so_pcb assignment */
        error = raw_attach(so, proto, ai->sb_rlimit);
        rp = sotorawcb(so);
        if (error) {
                kfree(rp, M_PCB);
                goto done;
        }
        switch(rp->rcb_proto.sp_protocol) {
        case AF_INET:
                route_cb.ip_count++;
                break;
        case AF_INET6:
                route_cb.ip6_count++;
                break;
        }
        rp->rcb_faddr = &route_src;
        route_cb.any_count++;
        soisconnected(so);
        so->so_options |= SO_USELOOPBACK;
        error = 0;
done:
        crit_exit();
        lwkt_replymsg(&msg->lmsg, error);
}

static void
rts_bind(netmsg_t msg)
{
        crit_enter();
        raw_usrreqs.pru_bind(msg); /* xxx just EINVAL */
        /* msg invalid now */
        crit_exit();
}

static void
rts_connect(netmsg_t msg)
{
        crit_enter();
        raw_usrreqs.pru_connect(msg); /* XXX just EINVAL */
        /* msg invalid now */
        crit_exit();
}

/* pru_connect2 is EOPNOTSUPP */
/* pru_control is EOPNOTSUPP */

static void
rts_detach(netmsg_t msg)
{
        struct socket *so = msg->base.nm_so;
        struct rawcb *rp = sotorawcb(so);

        crit_enter();
        if (rp != NULL) {
                switch(rp->rcb_proto.sp_protocol) {
                case AF_INET:
                        route_cb.ip_count--;
                        break;
                case AF_INET6:
                        route_cb.ip6_count--;
                        break;
                }
                route_cb.any_count--;
        }
        raw_usrreqs.pru_detach(msg);
        /* msg invalid now */
        crit_exit();
}

static void
rts_disconnect(netmsg_t msg)
{
        crit_enter();
        raw_usrreqs.pru_disconnect(msg);
        /* msg invalid now */
        crit_exit();
}

/* pru_listen is EOPNOTSUPP */

static void
rts_peeraddr(netmsg_t msg)
{
        crit_enter();
        raw_usrreqs.pru_peeraddr(msg);
        /* msg invalid now */
        crit_exit();
}

/* pru_rcvd is EOPNOTSUPP */
/* pru_rcvoob is EOPNOTSUPP */

static void
rts_send(netmsg_t msg)
{
        crit_enter();
        raw_usrreqs.pru_send(msg);
        /* msg invalid now */
        crit_exit();
}

/* pru_sense is null */

static void
rts_shutdown(netmsg_t msg)
{
        crit_enter();
        raw_usrreqs.pru_shutdown(msg);
        /* msg invalid now */
        crit_exit();
}

static void
rts_sockaddr(netmsg_t msg)
{
        crit_enter();
        raw_usrreqs.pru_sockaddr(msg);
        /* msg invalid now */
        crit_exit();
}

static struct pr_usrreqs route_usrreqs = {
        .pru_abort = rts_abort,
        .pru_accept = pr_generic_notsupp,
        .pru_attach = rts_attach,
        .pru_bind = rts_bind,
        .pru_connect = rts_connect,
        .pru_connect2 = pr_generic_notsupp,
        .pru_control = pr_generic_notsupp,
        .pru_detach = rts_detach,
        .pru_disconnect = rts_disconnect,
        .pru_listen = pr_generic_notsupp,
        .pru_peeraddr = rts_peeraddr,
        .pru_rcvd = pr_generic_notsupp,
        .pru_rcvoob = pr_generic_notsupp,
        .pru_send = rts_send,
        .pru_sense = pru_sense_null,
        .pru_shutdown = rts_shutdown,
        .pru_sockaddr = rts_sockaddr,
        .pru_sosend = sosend,
        .pru_soreceive = soreceive
};

static __inline sa_family_t
familyof(struct sockaddr *sa)
{
        return (sa != NULL ? sa->sa_family : 0);
}

/*
 * Routing socket input function.  The packet must be serialized onto cpu 0.
 * We use the cpu0_soport() netisr processing loop to handle it.
 *
 * This looks messy but it means that anyone, including interrupt code,
 * can send a message to the routing socket.
 */
static void
rts_input_handler(netmsg_t msg)
{
        static const struct sockaddr route_dst = { 2, PF_ROUTE, };
        struct sockproto route_proto;
        struct netmsg_packet *pmsg = &msg->packet;
        struct mbuf *m;
        sa_family_t family;
        struct rawcb *skip;

        family = pmsg->base.lmsg.u.ms_result;
        route_proto.sp_family = PF_ROUTE;
        route_proto.sp_protocol = family;

        m = pmsg->nm_packet;
        M_ASSERTPKTHDR(m);

        skip = m->m_pkthdr.header;
        m->m_pkthdr.header = NULL;

        raw_input(m, &route_proto, &route_src, &route_dst, skip);
}

static void
rts_input_skip(struct mbuf *m, sa_family_t family, struct rawcb *skip)
{
        struct netmsg_packet *pmsg;
        lwkt_port_t port;

        M_ASSERTPKTHDR(m);

        port = netisr_cpuport(0);       /* XXX same as for routing socket */
        pmsg = &m->m_hdr.mh_netmsg;
        netmsg_init(&pmsg->base, NULL, &netisr_apanic_rport,
                    0, rts_input_handler);
        pmsg->nm_packet = m;
        pmsg->base.lmsg.u.ms_result = family;
        m->m_pkthdr.header = skip; /* XXX steal field in pkthdr */
        lwkt_sendmsg(port, &pmsg->base.lmsg);
}

static __inline void
rts_input(struct mbuf *m, sa_family_t family)
{
        rts_input_skip(m, family, NULL);
}

static void *
reallocbuf_nofree(void *ptr, size_t len, size_t olen)
{
        void *newptr;

        newptr = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
        if (newptr == NULL)
                return NULL;
        bcopy(ptr, newptr, olen);
        return (newptr);
}

/*
 * Internal helper routine for route_output().
 */
static int
_fillrtmsg(struct rt_msghdr **prtm, struct rtentry *rt,
           struct rt_addrinfo *rtinfo)
{
        int msglen;
        struct rt_msghdr *rtm = *prtm;

        /* Fill in rt_addrinfo for call to rt_msg_buffer(). */
        rtinfo->rti_dst = rt_key(rt);
        rtinfo->rti_gateway = rt->rt_gateway;
        rtinfo->rti_netmask = rt_mask(rt);              /* might be NULL */
        rtinfo->rti_genmask = rt->rt_genmask;           /* might be NULL */
        if (rtm->rtm_addrs & (RTA_IFP | RTA_IFA)) {
                if (rt->rt_ifp != NULL) {
                        rtinfo->rti_ifpaddr =
                            TAILQ_FIRST(&rt->rt_ifp->if_addrheads[mycpuid])
                            ->ifa->ifa_addr;
                        rtinfo->rti_ifaaddr = rt->rt_ifa->ifa_addr;
                        if (rt->rt_ifp->if_flags & IFF_POINTOPOINT)
                                rtinfo->rti_bcastaddr = rt->rt_ifa->ifa_dstaddr;
                        rtm->rtm_index = rt->rt_ifp->if_index;
                } else {
                        rtinfo->rti_ifpaddr = NULL;
                        rtinfo->rti_ifaaddr = NULL;
                }
        } else if (rt->rt_ifp != NULL) {
                rtm->rtm_index = rt->rt_ifp->if_index;
        }

        msglen = rt_msgsize(rtm->rtm_type, rtinfo);
        if (rtm->rtm_msglen < msglen) {
                /* NOTE: Caller will free the old rtm accordingly */
                rtm = reallocbuf_nofree(rtm, msglen, rtm->rtm_msglen);
                if (rtm == NULL)
                        return (ENOBUFS);
                *prtm = rtm;
        }
        rt_msg_buffer(rtm->rtm_type, rtinfo, rtm, msglen);

        rtm->rtm_flags = rt->rt_flags;
        rtm->rtm_rmx = rt->rt_rmx;
        rtm->rtm_addrs = rtinfo->rti_addrs;

        return (0);
}

struct rtm_arg {
        struct rt_msghdr        *bak_rtm;
        struct rt_msghdr        *new_rtm;
};

static int
fillrtmsg(struct rtm_arg *arg, struct rtentry *rt,
          struct rt_addrinfo *rtinfo)
{
        struct rt_msghdr *rtm = arg->new_rtm;
        int error;

        error = _fillrtmsg(&rtm, rt, rtinfo);
        if (!error) {
                if (arg->new_rtm != rtm) {
                        /*
                         * _fillrtmsg() just allocated a new rtm;
                         * if the previously allocated rtm is not
                         * the backing rtm, it should be freed.
                         */
                        if (arg->new_rtm != arg->bak_rtm)
                                kfree(arg->new_rtm, M_RTABLE);
                        arg->new_rtm = rtm;
                }
        }
        return error;
}

static void route_output_add_callback(int, int, struct rt_addrinfo *,
                                        struct rtentry *, void *);
static void route_output_delete_callback(int, int, struct rt_addrinfo *,
                                        struct rtentry *, void *);
static int route_output_get_callback(int, struct rt_addrinfo *,
                                     struct rtentry *, void *, int);
static int route_output_change_callback(int, struct rt_addrinfo *,
                                        struct rtentry *, void *, int);
static int route_output_lock_callback(int, struct rt_addrinfo *,
                                      struct rtentry *, void *, int);

/*ARGSUSED*/
static int
route_output(struct mbuf *m, struct socket *so, ...)
{
        struct rtm_arg arg;
        struct rt_msghdr *rtm = NULL;
        struct rawcb *rp = NULL;
        struct pr_output_info *oi;
        struct rt_addrinfo rtinfo;
        sa_family_t family;
        int len, error = 0;
        __va_list ap;

        M_ASSERTPKTHDR(m);

        __va_start(ap, so);
        oi = __va_arg(ap, struct pr_output_info *);
        __va_end(ap);

        family = familyof(NULL);

#define gotoerr(e) { error = e; goto flush;}

        if (m == NULL ||
            (m->m_len < sizeof(long) &&
             (m = m_pullup(m, sizeof(long))) == NULL))
                return (ENOBUFS);
        len = m->m_pkthdr.len;
        if (len < sizeof(struct rt_msghdr) ||
            len != mtod(m, struct rt_msghdr *)->rtm_msglen)
                gotoerr(EINVAL);

        rtm = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
        if (rtm == NULL)
                gotoerr(ENOBUFS);

        m_copydata(m, 0, len, (caddr_t)rtm);
        if (rtm->rtm_version != RTM_VERSION)
                gotoerr(EPROTONOSUPPORT);

        rtm->rtm_pid = oi->p_pid;
        bzero(&rtinfo, sizeof(struct rt_addrinfo));
        rtinfo.rti_addrs = rtm->rtm_addrs;
        if (rt_xaddrs((char *)(rtm + 1), (char *)rtm + len, &rtinfo) != 0)
                gotoerr(EINVAL);

        rtinfo.rti_flags = rtm->rtm_flags;
        if (rtinfo.rti_dst == NULL || rtinfo.rti_dst->sa_family >= AF_MAX ||
            (rtinfo.rti_gateway && rtinfo.rti_gateway->sa_family >= AF_MAX))
                gotoerr(EINVAL);

        family = familyof(rtinfo.rti_dst);

        /*
         * Verify that the caller has the appropriate privilege; RTM_GET
         * is the only operation the non-superuser is allowed.
         */
        if (rtm->rtm_type != RTM_GET &&
            priv_check_cred(so->so_cred, PRIV_ROOT, 0) != 0)
                gotoerr(EPERM);

        if (rtinfo.rti_genmask != NULL) {
                error = rtmask_add_global(rtinfo.rti_genmask,
                    rtm->rtm_type != RTM_GET ?
                    RTREQ_PRIO_HIGH : RTREQ_PRIO_NORM);
                if (error)
                        goto flush;
        }

        switch (rtm->rtm_type) {
        case RTM_ADD:
                if (rtinfo.rti_gateway == NULL) {
                        error = EINVAL;
                } else {
                        error = rtrequest1_global(RTM_ADD, &rtinfo,
                            route_output_add_callback, rtm, RTREQ_PRIO_HIGH);
                }
                break;
        case RTM_DELETE:
                /*
                 * Backing rtm (bak_rtm) could _not_ be freed during
                 * rtrequest1_global or rtsearch_global, even if the
                 * callback reallocates the rtm due to its size changes,
                 * since rtinfo points to the backing rtm's memory area.
                 * After rtrequest1_global or rtsearch_global returns,
                 * it is safe to free the backing rtm, since rtinfo will
                 * not be used anymore.
                 *
                 * new_rtm will be used to save the new rtm allocated
                 * by rtrequest1_global or rtsearch_global.
                 */
                arg.bak_rtm = rtm;
                arg.new_rtm = rtm;
                error = rtrequest1_global(RTM_DELETE, &rtinfo,
                    route_output_delete_callback, &arg, RTREQ_PRIO_HIGH);
                rtm = arg.new_rtm;
                if (rtm != arg.bak_rtm)
                        kfree(arg.bak_rtm, M_RTABLE);
                break;
        case RTM_GET:
                /* See the comment in RTM_DELETE */
                arg.bak_rtm = rtm;
                arg.new_rtm = rtm;
                error = rtsearch_global(RTM_GET, &rtinfo,
                    route_output_get_callback, &arg, RTS_NOEXACTMATCH,
                    RTREQ_PRIO_NORM);
                rtm = arg.new_rtm;
                if (rtm != arg.bak_rtm)
                        kfree(arg.bak_rtm, M_RTABLE);
                break;
        case RTM_CHANGE:
                error = rtsearch_global(RTM_CHANGE, &rtinfo,
                    route_output_change_callback, rtm, RTS_EXACTMATCH,
                    RTREQ_PRIO_HIGH);
                break;
        case RTM_LOCK:
                error = rtsearch_global(RTM_LOCK, &rtinfo,
                    route_output_lock_callback, rtm, RTS_EXACTMATCH,
                    RTREQ_PRIO_HIGH);
                break;
        default:
                error = EOPNOTSUPP;
                break;
        }
flush:
        if (rtm != NULL) {
                if (error != 0)
                        rtm->rtm_errno = error;
                else
                        rtm->rtm_flags |= RTF_DONE;
        }

        /*
         * Check to see if we don't want our own messages.
         */
        if (!(so->so_options & SO_USELOOPBACK)) {
                if (route_cb.any_count <= 1) {
                        if (rtm != NULL)
                                kfree(rtm, M_RTABLE);
                        m_freem(m);
                        return (error);
                }
                /* There is another listener, so construct message */
                rp = sotorawcb(so);
        }
        if (rtm != NULL) {
                m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm);
                if (m->m_pkthdr.len < rtm->rtm_msglen) {
                        m_freem(m);
                        m = NULL;
                } else if (m->m_pkthdr.len > rtm->rtm_msglen)
                        m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len);
                kfree(rtm, M_RTABLE);
        }
        if (m != NULL)
                rts_input_skip(m, family, rp);
        return (error);
}

static void
route_output_add_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
                          struct rtentry *rt, void *arg)
{
        struct rt_msghdr *rtm = arg;

        if (error == 0 && rt != NULL) {
                rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx,
                    &rt->rt_rmx);
                rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits);
                rt->rt_rmx.rmx_locks |=
                    (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks);
                if (rtinfo->rti_genmask != NULL) {
                        rt->rt_genmask = rtmask_purelookup(rtinfo->rti_genmask);
                        if (rt->rt_genmask == NULL) {
                                /*
                                 * This should not happen, since we
                                 * have already installed genmask
                                 * on each CPU before we reach here.
                                 */
                                panic("genmask is gone!?");
                        }
                } else {
                        rt->rt_genmask = NULL;
                }
                rtm->rtm_index = rt->rt_ifp->if_index;
        }
}

static void
route_output_delete_callback(int cmd, int error, struct rt_addrinfo *rtinfo,
                          struct rtentry *rt, void *arg)
{
        if (error == 0 && rt) {
                ++rt->rt_refcnt;
                if (fillrtmsg(arg, rt, rtinfo) != 0) {
                        error = ENOBUFS;
                        /* XXX no way to return the error */
                }
                --rt->rt_refcnt;
        }
        if (rt && rt->rt_refcnt == 0) {
                ++rt->rt_refcnt;
                rtfree(rt);
        }
}

static int
route_output_get_callback(int cmd, struct rt_addrinfo *rtinfo,
                          struct rtentry *rt, void *arg, int found_cnt)
{
        int error, found = 0;

        if (((rtinfo->rti_flags ^ rt->rt_flags) & RTF_HOST) == 0)
                found = 1;

        error = fillrtmsg(arg, rt, rtinfo);
        if (!error && found) {
                /* Got the exact match, we could return now! */
                error = EJUSTRETURN;
        }
        return error;
}

static int
route_output_change_callback(int cmd, struct rt_addrinfo *rtinfo,
                             struct rtentry *rt, void *arg, int found_cnt)
{
        struct rt_msghdr *rtm = arg;
        struct ifaddr *ifa;
        int error = 0;

        /*
         * new gateway could require new ifaddr, ifp;
         * flags may also be different; ifp may be specified
         * by ll sockaddr when protocol address is ambiguous
         */
        if (((rt->rt_flags & RTF_GATEWAY) && rtinfo->rti_gateway != NULL) ||
            rtinfo->rti_ifpaddr != NULL ||
            (rtinfo->rti_ifaaddr != NULL &&
             !sa_equal(rtinfo->rti_ifaaddr, rt->rt_ifa->ifa_addr))) {
                error = rt_getifa(rtinfo);
                if (error != 0)
                        goto done;
        }
        if (rtinfo->rti_gateway != NULL) {
                /*
                 * We only need to generate rtmsg upon the
                 * first route to be changed.
                 */
                error = rt_setgate(rt, rt_key(rt), rtinfo->rti_gateway,
                        found_cnt == 1 ? RTL_REPORTMSG : RTL_DONTREPORT);
                if (error != 0)
                        goto done;
        }
        if ((ifa = rtinfo->rti_ifa) != NULL) {
                struct ifaddr *oifa = rt->rt_ifa;

                if (oifa != ifa) {
                        if (oifa && oifa->ifa_rtrequest)
                                oifa->ifa_rtrequest(RTM_DELETE, rt);
                        IFAFREE(rt->rt_ifa);
                        IFAREF(ifa);
                        rt->rt_ifa = ifa;
                        rt->rt_ifp = rtinfo->rti_ifp;
                }
        }
        rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, &rt->rt_rmx);
        if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest)
                rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt);
        if (rtinfo->rti_genmask != NULL) {
                rt->rt_genmask = rtmask_purelookup(rtinfo->rti_genmask);
                if (rt->rt_genmask == NULL) {
                        /*
                         * This should not happen, since we
                         * have already installed genmask
                         * on each CPU before we reach here.
                         */
                        panic("genmask is gone!?");
                }
        }
        rtm->rtm_index = rt->rt_ifp->if_index;
done:
        return error;
}

static int
route_output_lock_callback(int cmd, struct rt_addrinfo *rtinfo,
                           struct rtentry *rt, void *arg,
                           int found_cnt __unused)
{
        struct rt_msghdr *rtm = arg;

        rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits);
        rt->rt_rmx.rmx_locks |=
                (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks);
        return 0;
}

static void
rt_setmetrics(u_long which, struct rt_metrics *in, struct rt_metrics *out)
{
#define setmetric(flag, elt) if (which & (flag)) out->elt = in->elt;
        setmetric(RTV_RPIPE, rmx_recvpipe);
        setmetric(RTV_SPIPE, rmx_sendpipe);
        setmetric(RTV_SSTHRESH, rmx_ssthresh);
        setmetric(RTV_RTT, rmx_rtt);
        setmetric(RTV_RTTVAR, rmx_rttvar);
        setmetric(RTV_HOPCOUNT, rmx_hopcount);
        setmetric(RTV_MTU, rmx_mtu);
        setmetric(RTV_EXPIRE, rmx_expire);
        setmetric(RTV_MSL, rmx_msl);
        setmetric(RTV_IWMAXSEGS, rmx_iwmaxsegs);
        setmetric(RTV_IWCAPSEGS, rmx_iwcapsegs);
#undef setmetric
}

/*
 * Extract the addresses of the passed sockaddrs.
 * Do a little sanity checking so as to avoid bad memory references.
 * This data is derived straight from userland.
 */
static int
rt_xaddrs(char *cp, char *cplim, struct rt_addrinfo *rtinfo)
{
        struct sockaddr *sa;
        int i;

        for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) {
                if ((rtinfo->rti_addrs & (1 << i)) == 0)
                        continue;
                sa = (struct sockaddr *)cp;
                /*
                 * It won't fit.
                 */
                if ((cp + sa->sa_len) > cplim) {
                        return (EINVAL);
                }

                /*
                 * There are no more...  Quit now.
                 * If there are more bits, they are in error.
                 * I've seen this.  route(1) can evidently generate these. 
                 * This causes kernel to core dump.
                 * For compatibility, if we see this, point to a safe address.
                 */
                if (sa->sa_len == 0) {
                        static struct sockaddr sa_zero = {
                                sizeof sa_zero, AF_INET,
                        };

                        rtinfo->rti_info[i] = &sa_zero;
                        kprintf("rtsock: received more addr bits than sockaddrs.\n");
                        return (0); /* should be EINVAL but for compat */
                }

                /* Accept the sockaddr. */
                rtinfo->rti_info[i] = sa;
                cp += RT_ROUNDUP(sa->sa_len);
        }
        return (0);
}

static int
rt_msghdrsize(int type)
{
        switch (type) {
        case RTM_DELADDR:
        case RTM_NEWADDR:
                return sizeof(struct ifa_msghdr);
        case RTM_DELMADDR:
        case RTM_NEWMADDR:
                return sizeof(struct ifma_msghdr);
        case RTM_IFINFO:
                return sizeof(struct if_msghdr);
        case RTM_IFANNOUNCE:
        case RTM_IEEE80211:
                return sizeof(struct if_announcemsghdr);
        default:
                return sizeof(struct rt_msghdr);
        }
}

static int
rt_msgsize(int type, const struct rt_addrinfo *rtinfo)
{
        int len, i;

        len = rt_msghdrsize(type);
        for (i = 0; i < RTAX_MAX; i++) {
                if (rtinfo->rti_info[i] != NULL)
                        len += RT_ROUNDUP(rtinfo->rti_info[i]->sa_len);
        }
        len = ALIGN(len);
        return len;
}

/*
 * Build a routing message in a buffer.
 * Copy the addresses in the rtinfo->rti_info[] sockaddr array
 * to the end of the buffer after the message header.
 *
 * Set the rtinfo->rti_addrs bitmask of addresses present in rtinfo->rti_info[].
 * This side-effect can be avoided if we reorder the addrs bitmask field in all
 * the route messages to line up so we can set it here instead of back in the
 * calling routine.
 */
static void
rt_msg_buffer(int type, struct rt_addrinfo *rtinfo, void *buf, int msglen)
{
        struct rt_msghdr *rtm;
        char *cp;
        int dlen, i;

        rtm = (struct rt_msghdr *) buf;
        rtm->rtm_version = RTM_VERSION;
        rtm->rtm_type = type;
        rtm->rtm_msglen = msglen;

        cp = (char *)buf + rt_msghdrsize(type);
        rtinfo->rti_addrs = 0;
        for (i = 0; i < RTAX_MAX; i++) {
                struct sockaddr *sa;

                if ((sa = rtinfo->rti_info[i]) == NULL)
                        continue;
                rtinfo->rti_addrs |= (1 << i);
                dlen = RT_ROUNDUP(sa->sa_len);
                bcopy(sa, cp, dlen);
                cp += dlen;
        }
}

/*
 * Build a routing message in a mbuf chain.
 * Copy the addresses in the rtinfo->rti_info[] sockaddr array
 * to the end of the mbuf after the message header.
 *
 * Set the rtinfo->rti_addrs bitmask of addresses present in rtinfo->rti_info[].
 * This side-effect can be avoided if we reorder the addrs bitmask field in all
 * the route messages to line up so we can set it here instead of back in the
 * calling routine.
 */
static struct mbuf *
rt_msg_mbuf(int type, struct rt_addrinfo *rtinfo)
{
        struct mbuf *m;
        struct rt_msghdr *rtm;
        int hlen, len;
        int i;

        hlen = rt_msghdrsize(type);
        KASSERT(hlen <= MCLBYTES, ("rt_msg_mbuf: hlen %d doesn't fit", hlen));

        m = m_getl(hlen, M_NOWAIT, MT_DATA, M_PKTHDR, NULL);
        if (m == NULL)
                return (NULL);
        mbuftrackid(m, 32);
        m->m_pkthdr.len = m->m_len = hlen;
        m->m_pkthdr.rcvif = NULL;
        rtinfo->rti_addrs = 0;
        len = hlen;
        for (i = 0; i < RTAX_MAX; i++) {
                struct sockaddr *sa;
                int dlen;

                if ((sa = rtinfo->rti_info[i]) == NULL)
                        continue;
                rtinfo->rti_addrs |= (1 << i);
                dlen = RT_ROUNDUP(sa->sa_len);
                m_copyback(m, len, dlen, (caddr_t)sa); /* can grow mbuf chain */
                len += dlen;
        }
        if (m->m_pkthdr.len != len) { /* one of the m_copyback() calls failed */
                m_freem(m);
                return (NULL);
        }
        rtm = mtod(m, struct rt_msghdr *);
        bzero(rtm, hlen);
        rtm->rtm_msglen = len;
        rtm->rtm_version = RTM_VERSION;
        rtm->rtm_type = type;
        return (m);
}

/*
 * This routine is called to generate a message from the routing
 * socket indicating that a redirect has occurred, a routing lookup
 * has failed, or that a protocol has detected timeouts to a particular
 * destination.
 */
void
rt_missmsg(int type, struct rt_addrinfo *rtinfo, int flags, int error)
{
        struct sockaddr *dst = rtinfo->rti_info[RTAX_DST];
        struct rt_msghdr *rtm;
        struct mbuf *m;

        if (route_cb.any_count == 0)
                return;
        m = rt_msg_mbuf(type, rtinfo);
        if (m == NULL)
                return;
        rtm = mtod(m, struct rt_msghdr *);
        rtm->rtm_flags = RTF_DONE | flags;
        rtm->rtm_errno = error;
        rtm->rtm_addrs = rtinfo->rti_addrs;
        rts_input(m, familyof(dst));
}

void
rt_dstmsg(int type, struct sockaddr *dst, int error)
{
        struct rt_msghdr *rtm;
        struct rt_addrinfo addrs;
        struct mbuf *m;

        if (route_cb.any_count == 0)
                return;
        bzero(&addrs, sizeof(struct rt_addrinfo));
        addrs.rti_info[RTAX_DST] = dst;
        m = rt_msg_mbuf(type, &addrs);
        if (m == NULL)
                return;
        rtm = mtod(m, struct rt_msghdr *);
        rtm->rtm_flags = RTF_DONE;
        rtm->rtm_errno = error;
        rtm->rtm_addrs = addrs.rti_addrs;
        rts_input(m, familyof(dst));
}

/*
 * This routine is called to generate a message from the routing
 * socket indicating that the status of a network interface has changed.
 */
void
rt_ifmsg(struct ifnet *ifp)
{
        struct if_msghdr *ifm;
        struct mbuf *m;
        struct rt_addrinfo rtinfo;

        if (route_cb.any_count == 0)
                return;
        bzero(&rtinfo, sizeof(struct rt_addrinfo));
        m = rt_msg_mbuf(RTM_IFINFO, &rtinfo);
        if (m == NULL)
                return;
        ifm = mtod(m, struct if_msghdr *);
        ifm->ifm_index = ifp->if_index;
        ifm->ifm_flags = ifp->if_flags;
        ifm->ifm_data = ifp->if_data;
        ifm->ifm_addrs = 0;
        rts_input(m, 0);
}

static void
rt_ifamsg(int cmd, struct ifaddr *ifa)
{
        struct ifa_msghdr *ifam;
        struct rt_addrinfo rtinfo;
        struct mbuf *m;
        struct ifnet *ifp = ifa->ifa_ifp;

        bzero(&rtinfo, sizeof(struct rt_addrinfo));
        rtinfo.rti_ifaaddr = ifa->ifa_addr;
        rtinfo.rti_ifpaddr =
                TAILQ_FIRST(&ifp->if_addrheads[mycpuid])->ifa->ifa_addr;
        rtinfo.rti_netmask = ifa->ifa_netmask;
        rtinfo.rti_bcastaddr = ifa->ifa_dstaddr;

        m = rt_msg_mbuf(cmd, &rtinfo);
        if (m == NULL)
                return;

        ifam = mtod(m, struct ifa_msghdr *);
        ifam->ifam_index = ifp->if_index;
        ifam->ifam_metric = ifa->ifa_metric;
        ifam->ifam_flags = ifa->ifa_flags;
        ifam->ifam_addrs = rtinfo.rti_addrs;

        rts_input(m, familyof(ifa->ifa_addr));
}

void
rt_rtmsg(int cmd, struct rtentry *rt, struct ifnet *ifp, int error)
{
        struct rt_msghdr *rtm;
        struct rt_addrinfo rtinfo;
        struct mbuf *m;
        struct sockaddr *dst;

        if (rt == NULL)
                return;

        bzero(&rtinfo, sizeof(struct rt_addrinfo));
        rtinfo.rti_dst = dst = rt_key(rt);
        rtinfo.rti_gateway = rt->rt_gateway;
        rtinfo.rti_netmask = rt_mask(rt);
        if (ifp != NULL) {
                rtinfo.rti_ifpaddr =
                TAILQ_FIRST(&ifp->if_addrheads[mycpuid])->ifa->ifa_addr;
        }
        rtinfo.rti_ifaaddr = rt->rt_ifa->ifa_addr;

        m = rt_msg_mbuf(cmd, &rtinfo);
        if (m == NULL)
                return;

        rtm = mtod(m, struct rt_msghdr *);
        if (ifp != NULL)
                rtm->rtm_index = ifp->if_index;
        rtm->rtm_flags |= rt->rt_flags;
        rtm->rtm_errno = error;
        rtm->rtm_addrs = rtinfo.rti_addrs;

        rts_input(m, familyof(dst));
}

/*
 * This is called to generate messages from the routing socket
 * indicating a network interface has had addresses associated with it.
 * if we ever reverse the logic and replace messages TO the routing
 * socket indicate a request to configure interfaces, then it will
 * be unnecessary as the routing socket will automatically generate
 * copies of it.
 */
void
rt_newaddrmsg(int cmd, struct ifaddr *ifa, int error, struct rtentry *rt)
{
        if (route_cb.any_count == 0)
                return;

        if (cmd == RTM_ADD) {
                rt_ifamsg(RTM_NEWADDR, ifa);
                rt_rtmsg(RTM_ADD, rt, ifa->ifa_ifp, error);
        } else {
                KASSERT((cmd == RTM_DELETE), ("unknown cmd %d", cmd));
                rt_rtmsg(RTM_DELETE, rt, ifa->ifa_ifp, error);
                rt_ifamsg(RTM_DELADDR, ifa);
        }
}

/*
 * This is the analogue to the rt_newaddrmsg which performs the same
 * function but for multicast group memberhips.  This is easier since
 * there is no route state to worry about.
 */
void
rt_newmaddrmsg(int cmd, struct ifmultiaddr *ifma)
{
        struct rt_addrinfo rtinfo;
        struct mbuf *m = NULL;
        struct ifnet *ifp = ifma->ifma_ifp;
        struct ifma_msghdr *ifmam;

        if (route_cb.any_count == 0)
                return;

        bzero(&rtinfo, sizeof(struct rt_addrinfo));
        rtinfo.rti_ifaaddr = ifma->ifma_addr;
        if (ifp != NULL && !TAILQ_EMPTY(&ifp->if_addrheads[mycpuid])) {
                rtinfo.rti_ifpaddr =
                TAILQ_FIRST(&ifp->if_addrheads[mycpuid])->ifa->ifa_addr;
        }
        /*
         * If a link-layer address is present, present it as a ``gateway''
         * (similarly to how ARP entries, e.g., are presented).
         */
        rtinfo.rti_gateway = ifma->ifma_lladdr;

        m = rt_msg_mbuf(cmd, &rtinfo);
        if (m == NULL)
                return;

        ifmam = mtod(m, struct ifma_msghdr *);
        ifmam->ifmam_index = ifp->if_index;
        ifmam->ifmam_addrs = rtinfo.rti_addrs;

        rts_input(m, familyof(ifma->ifma_addr));
}

static struct mbuf *
rt_makeifannouncemsg(struct ifnet *ifp, int type, int what,
                     struct rt_addrinfo *info)
{
        struct if_announcemsghdr *ifan;
        struct mbuf *m;

        if (route_cb.any_count == 0)
                return NULL;

        bzero(info, sizeof(*info));
        m = rt_msg_mbuf(type, info);
        if (m == NULL)
                return NULL;

        ifan = mtod(m, struct if_announcemsghdr *);
        ifan->ifan_index = ifp->if_index;
        strlcpy(ifan->ifan_name, ifp->if_xname, sizeof ifan->ifan_name);
        ifan->ifan_what = what;
        return m;
}

/*
 * This is called to generate routing socket messages indicating
 * IEEE80211 wireless events.
 * XXX we piggyback on the RTM_IFANNOUNCE msg format in a clumsy way.
 */
void
rt_ieee80211msg(struct ifnet *ifp, int what, void *data, size_t data_len)
{
        struct rt_addrinfo info;
        struct mbuf *m;

        m = rt_makeifannouncemsg(ifp, RTM_IEEE80211, what, &info);
        if (m == NULL)
                return;

        /*
         * Append the ieee80211 data.  Try to stick it in the
         * mbuf containing the ifannounce msg; otherwise allocate
         * a new mbuf and append.
         *
         * NB: we assume m is a single mbuf.
         */
        if (data_len > M_TRAILINGSPACE(m)) {
                /* XXX use m_getb(data_len, M_NOWAIT, MT_DATA, 0); */
                struct mbuf *n = m_get(M_NOWAIT, MT_DATA);
                if (n == NULL) {
                        m_freem(m);
                        return;
                }
                KKASSERT(data_len <= M_TRAILINGSPACE(n));
                bcopy(data, mtod(n, void *), data_len);
                n->m_len = data_len;
                m->m_next = n;
        } else if (data_len > 0) {
                bcopy(data, mtod(m, u_int8_t *) + m->m_len, data_len);
                m->m_len += data_len;
        }
        mbuftrackid(m, 33);
        if (m->m_flags & M_PKTHDR)
                m->m_pkthdr.len += data_len;
        mtod(m, struct if_announcemsghdr *)->ifan_msglen += data_len;
        rts_input(m, 0);
}

/*
 * This is called to generate routing socket messages indicating
 * network interface arrival and departure.
 */
void
rt_ifannouncemsg(struct ifnet *ifp, int what)
{
        struct rt_addrinfo addrinfo;
        struct mbuf *m;

        m = rt_makeifannouncemsg(ifp, RTM_IFANNOUNCE, what, &addrinfo);
        if (m != NULL)
                rts_input(m, 0);
}

static int
resizewalkarg(struct walkarg *w, int len)
{
        void *newptr;

        newptr = kmalloc(len, M_RTABLE, M_INTWAIT | M_NULLOK);
        if (newptr == NULL)
                return (ENOMEM);
        if (w->w_tmem != NULL)
                kfree(w->w_tmem, M_RTABLE);
        w->w_tmem = newptr;
        w->w_tmemsize = len;
        return (0);
}

static void
ifnet_compute_stats(struct ifnet *ifp)
{
        IFNET_STAT_GET(ifp, ipackets, ifp->if_ipackets);
        IFNET_STAT_GET(ifp, ierrors, ifp->if_ierrors);
        IFNET_STAT_GET(ifp, opackets, ifp->if_opackets);
        IFNET_STAT_GET(ifp, collisions, ifp->if_collisions);
        IFNET_STAT_GET(ifp, ibytes, ifp->if_ibytes);
        IFNET_STAT_GET(ifp, obytes, ifp->if_obytes);
        IFNET_STAT_GET(ifp, imcasts, ifp->if_imcasts);
        IFNET_STAT_GET(ifp, omcasts, ifp->if_omcasts);
        IFNET_STAT_GET(ifp, iqdrops, ifp->if_iqdrops);
        IFNET_STAT_GET(ifp, noproto, ifp->if_noproto);
        IFNET_STAT_GET(ifp, oqdrops, ifp->if_oqdrops);
}

static int
sysctl_iflist(int af, struct walkarg *w)
{
        struct ifnet *ifp;
        struct rt_addrinfo rtinfo;
        int msglen, error;

        bzero(&rtinfo, sizeof(struct rt_addrinfo));

        ifnet_lock();
        TAILQ_FOREACH(ifp, &ifnetlist, if_link) {
                struct ifaddr_container *ifac, *ifac_mark;
                struct ifaddr_marker mark;
                struct ifaddrhead *head;
                struct ifaddr *ifa;

                if (w->w_arg && w->w_arg != ifp->if_index)
                        continue;
                head = &ifp->if_addrheads[mycpuid];
                /*
                 * There is no need to reference the first ifaddr
                 * even if the following resizewalkarg() blocks,
                 * since the first ifaddr will not be destroyed
                 * when the ifnet lock is held.
                 */
                ifac = TAILQ_FIRST(head);
                ifa = ifac->ifa;
                rtinfo.rti_ifpaddr = ifa->ifa_addr;
                msglen = rt_msgsize(RTM_IFINFO, &rtinfo);
                if (w->w_tmemsize < msglen && resizewalkarg(w, msglen) != 0) {
                        ifnet_unlock();
                        return (ENOMEM);
                }
                rt_msg_buffer(RTM_IFINFO, &rtinfo, w->w_tmem, msglen);
                rtinfo.rti_ifpaddr = NULL;
                if (w->w_req != NULL && w->w_tmem != NULL) {
                        struct if_msghdr *ifm = w->w_tmem;

                        ifm->ifm_index = ifp->if_index;
                        ifm->ifm_flags = ifp->if_flags;
                        ifnet_compute_stats(ifp);
                        ifm->ifm_data = ifp->if_data;
                        ifm->ifm_addrs = rtinfo.rti_addrs;
                        error = SYSCTL_OUT(w->w_req, ifm, msglen);
                        if (error) {
                                ifnet_unlock();
                                return (error);
                        }
                }
                /*
                 * Add a marker, since SYSCTL_OUT() could block and during
                 * that period the list could be changed.
                 */
                ifa_marker_init(&mark, ifp);
                ifac_mark = &mark.ifac;
                TAILQ_INSERT_AFTER(head, ifac, ifac_mark, ifa_link);
                while ((ifac = TAILQ_NEXT(ifac_mark, ifa_link)) != NULL) {
                        TAILQ_REMOVE(head, ifac_mark, ifa_link);
                        TAILQ_INSERT_AFTER(head, ifac, ifac_mark, ifa_link);

                        ifa = ifac->ifa;

                        /* Ignore marker */
                        if (ifa->ifa_addr->sa_family == AF_UNSPEC)
                                continue;

                        if (af && af != ifa->ifa_addr->sa_family)
                                continue;
                        if (curproc->p_ucred->cr_prison &&
                            prison_if(curproc->p_ucred, ifa->ifa_addr))
                                continue;
                        rtinfo.rti_ifaaddr = ifa->ifa_addr;
                        rtinfo.rti_netmask = ifa->ifa_netmask;
                        rtinfo.rti_bcastaddr = ifa->ifa_dstaddr;
                        msglen = rt_msgsize(RTM_NEWADDR, &rtinfo);
                        /*
                         * Keep a reference on this ifaddr, so that it will
                         * not be destroyed if the following resizewalkarg()
                         * blocks.
                         */
                        IFAREF(ifa);
                        if (w->w_tmemsize < msglen &&
                            resizewalkarg(w, msglen) != 0) {
                                IFAFREE(ifa);
                                TAILQ_REMOVE(head, ifac_mark, ifa_link);
                                ifnet_unlock();
                                return (ENOMEM);
                        }
                        rt_msg_buffer(RTM_NEWADDR, &rtinfo, w->w_tmem, msglen);
                        if (w->w_req != NULL) {
                                struct ifa_msghdr *ifam = w->w_tmem;

                                ifam->ifam_index = ifa->ifa_ifp->if_index;
                                ifam->ifam_flags = ifa->ifa_flags;
                                ifam->ifam_metric = ifa->ifa_metric;
                                ifam->ifam_addrs = rtinfo.rti_addrs;
                                error = SYSCTL_OUT(w->w_req, w->w_tmem, msglen);
                                if (error) {
                                        IFAFREE(ifa);
                                        TAILQ_REMOVE(head, ifac_mark, ifa_link);
                                        ifnet_unlock();
                                        return (error);
                                }
                        }
                        IFAFREE(ifa);
                }
                TAILQ_REMOVE(head, ifac_mark, ifa_link);
                rtinfo.rti_netmask = NULL;
                rtinfo.rti_ifaaddr = NULL;
                rtinfo.rti_bcastaddr = NULL;
        }
        ifnet_unlock();
        return (0);
}

static int
rttable_walkarg_create(struct rttable_walkarg *w, int op, int arg)
{
        struct rt_addrinfo rtinfo;
        struct sockaddr_storage ss;
        int i, msglen;

        memset(w, 0, sizeof(*w));
        w->w_op = op;
        w->w_arg = arg;

        memset(&ss, 0, sizeof(ss));
        ss.ss_len = sizeof(ss);

        memset(&rtinfo, 0, sizeof(rtinfo));
        for (i = 0; i < RTAX_MAX; ++i)
                rtinfo.rti_info[i] = (struct sockaddr *)&ss;
        msglen = rt_msgsize(RTM_GET, &rtinfo);

        w->w_bufsz = msglen * RTTABLE_DUMP_MSGCNT_MAX;
        w->w_buf = kmalloc(w->w_bufsz, M_TEMP, M_WAITOK | M_NULLOK);
        if (w->w_buf == NULL)
                return ENOMEM;
        return 0;
}

static void
rttable_walkarg_destroy(struct rttable_walkarg *w)
{
        kfree(w->w_buf, M_TEMP);
}

static void
rttable_entry_rtinfo(struct rt_addrinfo *rtinfo, struct radix_node *rn)
{
        struct rtentry *rt = (struct rtentry *)rn;

        bzero(rtinfo, sizeof(*rtinfo));
        rtinfo->rti_dst = rt_key(rt);
        rtinfo->rti_gateway = rt->rt_gateway;
        rtinfo->rti_netmask = rt_mask(rt);
        rtinfo->rti_genmask = rt->rt_genmask;
        if (rt->rt_ifp != NULL) {
                rtinfo->rti_ifpaddr =
                TAILQ_FIRST(&rt->rt_ifp->if_addrheads[mycpuid])->ifa->ifa_addr;
                rtinfo->rti_ifaaddr = rt->rt_ifa->ifa_addr;
                if (rt->rt_ifp->if_flags & IFF_POINTOPOINT)
                        rtinfo->rti_bcastaddr = rt->rt_ifa->ifa_dstaddr;
        }
}

static int
rttable_walk_entry(struct radix_node *rn, void *xw)
{
        struct rttable_walkarg *w = xw;
        struct rtentry *rt = (struct rtentry *)rn;
        struct rt_addrinfo rtinfo;
        struct rt_msghdr *rtm;
        boolean_t save = FALSE;
        int msglen, w_bufleft;
        void *ptr;

        rttable_entry_rtinfo(&rtinfo, rn);
        msglen = rt_msgsize(RTM_GET, &rtinfo);

        w_bufleft = w->w_bufsz - w->w_buflen;

        if (rn->rn_dupedkey != NULL) {
                struct radix_node *rn1 = rn;
                int total_msglen = msglen;

                /*
                 * Make sure that we have enough space left for all
                 * dupedkeys, since rn_walktree_at always starts
                 * from the first dupedkey.
                 */
                while ((rn1 = rn1->rn_dupedkey) != NULL) {
                        struct rt_addrinfo rtinfo1;
                        int msglen1;

                        if (rn1->rn_flags & RNF_ROOT)
                                continue;

                        rttable_entry_rtinfo(&rtinfo1, rn1);
                        msglen1 = rt_msgsize(RTM_GET, &rtinfo1);
                        total_msglen += msglen1;
                }

                if (total_msglen > w_bufleft) {
                        if (total_msglen > w->w_bufsz) {
                                static int logged = 0;

                                if (!logged) {
                                        kprintf("buffer is too small for "
                                            "all dupedkeys, increase "
                                            "RTTABLE_DUMP_MSGCNT_MAX\n");
                                        logged = 1;
                                }
                                return ENOMEM;
                        }
                        save = TRUE;
                }
        } else if (msglen > w_bufleft) {
                save = TRUE;
        }

        if (save) {
                /*
                 * Not enough buffer left; remember the position
                 * to start from upon next round.
                 */
                KASSERT(msglen <= w->w_bufsz, ("msg too long %d", msglen));

                KASSERT(rtinfo.rti_dst->sa_len <= sizeof(w->w_key0),
                    ("key too long %d", rtinfo.rti_dst->sa_len));
                memset(&w->w_key0, 0, sizeof(w->w_key0));
                memcpy(&w->w_key0, rtinfo.rti_dst, rtinfo.rti_dst->sa_len);
                w->w_key = (const char *)&w->w_key0;

                if (rtinfo.rti_netmask != NULL) {
                        KASSERT(
                            rtinfo.rti_netmask->sa_len <= sizeof(w->w_mask0),
                            ("mask too long %d", rtinfo.rti_netmask->sa_len));
                        memset(&w->w_mask0, 0, sizeof(w->w_mask0));
                        memcpy(&w->w_mask0, rtinfo.rti_netmask,
                            rtinfo.rti_netmask->sa_len);
                        w->w_mask = (const char *)&w->w_mask0;
                } else {
                        w->w_mask = NULL;
                }
                return EJUSTRETURN;
        }

        if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg))
                return 0;

        ptr = ((uint8_t *)w->w_buf) + w->w_buflen;
        rt_msg_buffer(RTM_GET, &rtinfo, ptr, msglen);

        rtm = (struct rt_msghdr *)ptr;
        rtm->rtm_flags = rt->rt_flags;
        rtm->rtm_use = rt->rt_use;
        rtm->rtm_rmx = rt->rt_rmx;
        rtm->rtm_index = rt->rt_ifp->if_index;
        rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0;
        rtm->rtm_addrs = rtinfo.rti_addrs;

        w->w_buflen += msglen;

        return 0;
}

static void
rttable_walk_dispatch(netmsg_t msg)
{
        struct netmsg_rttable_walk *nmsg = (struct netmsg_rttable_walk *)msg;
        struct radix_node_head *rnh = rt_tables[mycpuid][nmsg->af];
        struct rttable_walkarg *w = nmsg->w;
        int error;

        error = rnh->rnh_walktree_at(rnh, w->w_key, w->w_mask,
            rttable_walk_entry, w);
        lwkt_replymsg(&nmsg->base.lmsg, error);
}

static int
sysctl_rttable(int af, struct sysctl_req *req, int op, int arg)
{
        struct rttable_walkarg w;
        int error, i;

        error = rttable_walkarg_create(&w, op, arg);
        if (error)
                return error;

        error = EINVAL;
        for (i = 1; i <= AF_MAX; i++) {
                if (rt_tables[mycpuid][i] != NULL && (af == 0 || af == i)) {
                        w.w_key = NULL;
                        w.w_mask = NULL;
                        for (;;) {
                                struct netmsg_rttable_walk nmsg;

                                netmsg_init(&nmsg.base, NULL,
                                    &curthread->td_msgport, 0,
                                    rttable_walk_dispatch);
                                nmsg.af = i;
                                nmsg.w = &w;

                                w.w_buflen = 0;

                                error = lwkt_domsg(netisr_cpuport(mycpuid),
                                    &nmsg.base.lmsg, 0);
                                if (error && error != EJUSTRETURN)
                                        goto done;

                                if (req != NULL && w.w_buflen > 0) {
                                        int error1;

                                        error1 = SYSCTL_OUT(req, w.w_buf,
                                            w.w_buflen);
                                        if (error1) {
                                                error = error1;
                                                goto done;
                                        }
                                }
                                if (error == 0) /* done */
                                        break;
                        }
                }
        }
done:
        rttable_walkarg_destroy(&w);
        return error;
}

static int
sysctl_rtsock(SYSCTL_HANDLER_ARGS)
{
        int     *name = (int *)arg1;
        u_int   namelen = arg2;
        int     error = EINVAL;
        int     origcpu, cpu;
        u_char  af;
        struct  walkarg w;

        name ++;
        namelen--;
        if (req->newptr)
                return (EPERM);
        if (namelen != 3 && namelen != 4)
                return (EINVAL);
        af = name[0];
        bzero(&w, sizeof w);
        w.w_op = name[1];
        w.w_arg = name[2];
        w.w_req = req;

        /*
         * Optional third argument specifies cpu, used primarily for
         * debugging the route table.
         */
        if (namelen == 4) {
                if (name[3] < 0 || name[3] >= ncpus)
                        return (EINVAL);
                cpu = name[3];
        } else {
                /*
                 * Target cpu is not specified, use cpu0 then, so that
                 * the result set will be relatively stable.
                 */
                cpu = 0;
        }
        origcpu = mycpuid;
        lwkt_migratecpu(cpu);

        switch (w.w_op) {
        case NET_RT_DUMP:
        case NET_RT_FLAGS:
                error = sysctl_rttable(af, w.w_req, w.w_op, w.w_arg);
                break;

        case NET_RT_IFLIST:
                error = sysctl_iflist(af, &w);
                break;
        }
        if (w.w_tmem != NULL)
                kfree(w.w_tmem, M_RTABLE);

        lwkt_migratecpu(origcpu);
        return (error);
}

SYSCTL_NODE(_net, PF_ROUTE, routetable, CTLFLAG_RD, sysctl_rtsock, "");

/*
 * Definitions of protocols supported in the ROUTE domain.
 */

static struct domain routedomain;               /* or at least forward */

static struct protosw routesw[] = {
    {
        .pr_type = SOCK_RAW,
        .pr_domain = &routedomain,
        .pr_protocol = 0,
        .pr_flags = PR_ATOMIC|PR_ADDR,
        .pr_input = NULL,
        .pr_output = route_output,
        .pr_ctlinput = raw_ctlinput,
        .pr_ctloutput = NULL,
        .pr_ctlport = cpu0_ctlport,

        .pr_init = raw_init,
        .pr_usrreqs = &route_usrreqs
    }
};

static struct domain routedomain = {
        PF_ROUTE, "route", NULL, NULL, NULL,
        routesw, &routesw[(sizeof routesw)/(sizeof routesw[0])],
};

DOMAIN_SET(route);