freebsd/sys/net/if_ipsec.c

/*-
 * Copyright (c) 2016 Yandex LLC
 * Copyright (c) 2016 Andrey V. Elsukov <ae@FreeBSD.org>
 * 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.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``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 AUTHOR 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.
 */

#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");

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

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/fnv_hash.h>
#include <sys/jail.h>
#include <sys/lock.h>
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/module.h>
#include <sys/rmlock.h>
#include <sys/socket.h>
#include <sys/sockio.h>
#include <sys/sx.h>
#include <sys/errno.h>
#include <sys/sysctl.h>
#include <sys/priv.h>
#include <sys/proc.h>
#include <sys/conf.h>

#include <net/if.h>
#include <net/if_var.h>
#include <net/if_clone.h>
#include <net/if_types.h>
#include <net/bpf.h>
#include <net/route.h>
#include <net/vnet.h>

#include <netinet/in.h>
#include <netinet/in_var.h>
#include <netinet/ip.h>

#include <netinet/ip6.h>
#include <netinet6/in6_var.h>
#include <netinet6/scope6_var.h>

#include <netipsec/ipsec.h>
#ifdef INET6
#include <netipsec/ipsec6.h>
#endif

#include <net/if_ipsec.h>
#include <netipsec/key.h>

#include <security/mac/mac_framework.h>

static MALLOC_DEFINE(M_IPSEC, "ipsec", "IPsec Virtual Tunnel Interface");
static const char ipsecname[] = "ipsec";

#if defined(INET) && defined(INET6)
#define	IPSEC_SPCOUNT		4
#else
#define	IPSEC_SPCOUNT		2
#endif

struct ipsec_softc {
	struct ifnet		*ifp;

	struct rmlock		lock;
	struct secpolicy	*sp[IPSEC_SPCOUNT];

	uint32_t		reqid;
	u_int			family;
	u_int			fibnum;
	LIST_ENTRY(ipsec_softc)	chain;
	LIST_ENTRY(ipsec_softc) hash;
};

#define	IPSEC_LOCK_INIT(sc)	rm_init(&(sc)->lock, "if_ipsec softc")
#define	IPSEC_LOCK_DESTROY(sc)	rm_destroy(&(sc)->lock)
#define	IPSEC_RLOCK_TRACKER	struct rm_priotracker ipsec_tracker
#define	IPSEC_RLOCK(sc)		rm_rlock(&(sc)->lock, &ipsec_tracker)
#define	IPSEC_RUNLOCK(sc)	rm_runlock(&(sc)->lock, &ipsec_tracker)
#define	IPSEC_RLOCK_ASSERT(sc)	rm_assert(&(sc)->lock, RA_RLOCKED)
#define	IPSEC_WLOCK(sc)		rm_wlock(&(sc)->lock)
#define	IPSEC_WUNLOCK(sc)	rm_wunlock(&(sc)->lock)
#define	IPSEC_WLOCK_ASSERT(sc)	rm_assert(&(sc)->lock, RA_WLOCKED)

static struct rmlock ipsec_sc_lock;
RM_SYSINIT(ipsec_sc_lock, &ipsec_sc_lock, "if_ipsec softc list");

#define	IPSEC_SC_RLOCK_TRACKER	struct rm_priotracker ipsec_sc_tracker
#define	IPSEC_SC_RLOCK()	rm_rlock(&ipsec_sc_lock, &ipsec_sc_tracker)
#define	IPSEC_SC_RUNLOCK()	rm_runlock(&ipsec_sc_lock, &ipsec_sc_tracker)
#define	IPSEC_SC_RLOCK_ASSERT()	rm_assert(&ipsec_sc_lock, RA_RLOCKED)
#define	IPSEC_SC_WLOCK()	rm_wlock(&ipsec_sc_lock)
#define	IPSEC_SC_WUNLOCK()	rm_wunlock(&ipsec_sc_lock)
#define	IPSEC_SC_WLOCK_ASSERT()	rm_assert(&ipsec_sc_lock, RA_WLOCKED)

LIST_HEAD(ipsec_iflist, ipsec_softc);
static VNET_DEFINE(struct ipsec_iflist, ipsec_sc_list);
static VNET_DEFINE(struct ipsec_iflist *, ipsec_sc_htbl);
static VNET_DEFINE(u_long, ipsec_sc_hmask);
#define	V_ipsec_sc_list		VNET(ipsec_sc_list)
#define	V_ipsec_sc_htbl		VNET(ipsec_sc_htbl)
#define	V_ipsec_sc_hmask	VNET(ipsec_sc_hmask)

static uint32_t
ipsec_hash(uint32_t id)
{

	return (fnv_32_buf(&id, sizeof(id), FNV1_32_INIT));
}

#define	SCHASH_NHASH_LOG2	5
#define	SCHASH_NHASH		(1 << SCHASH_NHASH_LOG2)
#define	SCHASH_HASHVAL(id)	(ipsec_hash((id)) & V_ipsec_sc_hmask)
#define	SCHASH_HASH(id)		&V_ipsec_sc_htbl[SCHASH_HASHVAL(id)]

/*
 * ipsec_ioctl_sx protects from concurrent ioctls.
 */
static struct sx ipsec_ioctl_sx;
SX_SYSINIT(ipsec_ioctl_sx, &ipsec_ioctl_sx, "ipsec_ioctl");

static int	ipsec_init_reqid(struct ipsec_softc *);
static int	ipsec_set_tunnel(struct ipsec_softc *, struct sockaddr *,
    struct sockaddr *, uint32_t);
static void	ipsec_delete_tunnel(struct ifnet *, int);

static int	ipsec_set_addresses(struct ifnet *, struct sockaddr *,
    struct sockaddr *);
static int	ipsec_set_reqid(struct ifnet *, uint32_t);

static int	ipsec_ioctl(struct ifnet *, u_long, caddr_t);
static int	ipsec_transmit(struct ifnet *, struct mbuf *);
static int	ipsec_output(struct ifnet *, struct mbuf *,
    const struct sockaddr *, struct route *);
static void	ipsec_qflush(struct ifnet *);
static int	ipsec_clone_create(struct if_clone *, int, caddr_t);
static void	ipsec_clone_destroy(struct ifnet *);

static VNET_DEFINE(struct if_clone *, ipsec_cloner);
#define	V_ipsec_cloner		VNET(ipsec_cloner)

static int
ipsec_clone_create(struct if_clone *ifc, int unit, caddr_t params)
{
	struct ipsec_softc *sc;
	struct ifnet *ifp;

	sc = malloc(sizeof(*sc), M_IPSEC, M_WAITOK | M_ZERO);
	sc->fibnum = curthread->td_proc->p_fibnum;
	sc->ifp = ifp = if_alloc(IFT_TUNNEL);
	IPSEC_LOCK_INIT(sc);
	ifp->if_softc = sc;
	if_initname(ifp, ipsecname, unit);

	ifp->if_addrlen = 0;
	ifp->if_mtu = IPSEC_MTU;
	ifp->if_flags  = IFF_POINTOPOINT | IFF_MULTICAST;
	ifp->if_ioctl  = ipsec_ioctl;
	ifp->if_transmit  = ipsec_transmit;
	ifp->if_qflush  = ipsec_qflush;
	ifp->if_output = ipsec_output;
	if_attach(ifp);
	bpfattach(ifp, DLT_NULL, sizeof(uint32_t));

	IPSEC_SC_WLOCK();
	LIST_INSERT_HEAD(&V_ipsec_sc_list, sc, chain);
	IPSEC_SC_WUNLOCK();
	return (0);
}

static void
ipsec_clone_destroy(struct ifnet *ifp)
{
	struct ipsec_softc *sc;

	sx_xlock(&ipsec_ioctl_sx);
	sc = ifp->if_softc;

	IPSEC_SC_WLOCK();
	ipsec_delete_tunnel(ifp, 1);
	LIST_REMOVE(sc, chain);
	IPSEC_SC_WUNLOCK();

	bpfdetach(ifp);
	if_detach(ifp);
	ifp->if_softc = NULL;
	sx_xunlock(&ipsec_ioctl_sx);

	if_free(ifp);
	IPSEC_LOCK_DESTROY(sc);
	free(sc, M_IPSEC);
}

static void
vnet_ipsec_init(const void *unused __unused)
{

	LIST_INIT(&V_ipsec_sc_list);
	V_ipsec_sc_htbl = hashinit(SCHASH_NHASH, M_IPSEC, &V_ipsec_sc_hmask);
	V_ipsec_cloner = if_clone_simple(ipsecname, ipsec_clone_create,
	    ipsec_clone_destroy, 0);
}
VNET_SYSINIT(vnet_ipsec_init, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY,
    vnet_ipsec_init, NULL);

static void
vnet_ipsec_uninit(const void *unused __unused)
{

	if_clone_detach(V_ipsec_cloner);
	hashdestroy(V_ipsec_sc_htbl, M_IPSEC, V_ipsec_sc_hmask);
}
VNET_SYSUNINIT(vnet_ipsec_uninit, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY,
    vnet_ipsec_uninit, NULL);

static struct secpolicy *
ipsec_getpolicy(struct ipsec_softc *sc, int dir, sa_family_t af)
{

	switch (af) {
#ifdef INET
	case AF_INET:
		return (sc->sp[(dir == IPSEC_DIR_INBOUND ? 0: 1)]);
#endif
#ifdef INET6
	case AF_INET6:
		return (sc->sp[(dir == IPSEC_DIR_INBOUND ? 0: 1)
#ifdef INET
			+ 2
#endif
		]);
#endif
	}
	return (NULL);
}

static struct secasindex *
ipsec_getsaidx(struct ipsec_softc *sc, int dir, sa_family_t af)
{
	struct secpolicy *sp;

	sp = ipsec_getpolicy(sc, dir, af);
	if (sp == NULL)
		return (NULL);
	return (&sp->req[0]->saidx);
}

static int
ipsec_transmit(struct ifnet *ifp, struct mbuf *m)
{
	IPSEC_RLOCK_TRACKER;
	struct ipsec_softc *sc;
	struct secpolicy *sp;
	struct ip *ip;
	uint32_t af;
	int error;

#ifdef MAC
	error = mac_ifnet_check_transmit(ifp, m);
	if (error) {
		m_freem(m);
		goto err;
	}
#endif
	error = ENETDOWN;
	sc = ifp->if_softc;
	if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0 ||
	    (ifp->if_flags & IFF_MONITOR) != 0 ||
	    (ifp->if_flags & IFF_UP) == 0) {
		m_freem(m);
		goto err;
	}

	/* Determine address family to correctly handle packet in BPF */
	ip = mtod(m, struct ip *);
	switch (ip->ip_v) {
#ifdef INET
	case IPVERSION:
		af = AF_INET;
		break;
#endif
#ifdef INET6
	case (IPV6_VERSION >> 4):
		af = AF_INET6;
		break;
#endif
	default:
		error = EAFNOSUPPORT;
		m_freem(m);
		goto err;
	}

	/*
	 * Loop prevention.
	 * XXX: for now just check presence of IPSEC_OUT_DONE mbuf tag.
	 *      We can read full chain and compare destination address,
	 *      proto and mode from xform_history with values from softc.
	 */
	if (m_tag_find(m, PACKET_TAG_IPSEC_OUT_DONE, NULL) != NULL) {
		m_freem(m);
		goto err;
	}

	IPSEC_RLOCK(sc);
	if (sc->family == 0) {
		IPSEC_RUNLOCK(sc);
		m_freem(m);
		goto err;
	}
	sp = ipsec_getpolicy(sc, IPSEC_DIR_OUTBOUND, af);
	key_addref(sp);
	M_SETFIB(m, sc->fibnum);
	IPSEC_RUNLOCK(sc);

	BPF_MTAP2(ifp, &af, sizeof(af), m);
	if_inc_counter(ifp, IFCOUNTER_OPACKETS, 1);
	if_inc_counter(ifp, IFCOUNTER_OBYTES, m->m_pkthdr.len);

	switch (af) {
#ifdef INET
	case AF_INET:
		error = ipsec4_process_packet(m, sp, NULL);
		break;
#endif
#ifdef INET6
	case AF_INET6:
		error = ipsec6_process_packet(m, sp, NULL);
		break;
#endif
	default:
		panic("%s: unknown address family\n", __func__);
	}
err:
	if (error != 0)
		if_inc_counter(ifp, IFCOUNTER_OERRORS, 1);
	return (error);
}

static void
ipsec_qflush(struct ifnet *ifp __unused)
{

}

static int
ipsec_output(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *dst,
	struct route *ro)
{

	return (ifp->if_transmit(ifp, m));
}

int
ipsec_if_input(struct mbuf *m, struct secasvar *sav, uint32_t af)
{
	IPSEC_SC_RLOCK_TRACKER;
	struct secasindex *saidx;
	struct ipsec_softc *sc;
	struct ifnet *ifp;

	if (sav->state != SADB_SASTATE_MATURE &&
	    sav->state != SADB_SASTATE_DYING) {
		m_freem(m);
		return (ENETDOWN);
	}

	if (sav->sah->saidx.mode != IPSEC_MODE_TUNNEL ||
	    sav->sah->saidx.proto != IPPROTO_ESP)
		return (0);

	IPSEC_SC_RLOCK();
	/*
	 * We only acquire SC_RLOCK() while we are doing search in
	 * ipsec_sc_htbl. It is safe, because removing softc or changing
	 * of reqid/addresses requires removing from hash table.
	 */
	LIST_FOREACH(sc, SCHASH_HASH(sav->sah->saidx.reqid), hash) {
		saidx = ipsec_getsaidx(sc, IPSEC_DIR_INBOUND,
		    sav->sah->saidx.src.sa.sa_family);
		/* SA's reqid should match reqid in SP */
		if (saidx == NULL ||
		    sav->sah->saidx.reqid != saidx->reqid)
			continue;
		/* SAH's addresses should match tunnel endpoints. */
		if (key_sockaddrcmp(&sav->sah->saidx.dst.sa,
		    &saidx->dst.sa, 0) != 0)
			continue;
		if (key_sockaddrcmp(&sav->sah->saidx.src.sa,
		    &saidx->src.sa, 0) == 0)
			break;
	}
	if (sc == NULL) {
		IPSEC_SC_RUNLOCK();
		/* Tunnel was not found. Nothing to do. */
		return (0);
	}
	ifp = sc->ifp;
	if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0 ||
	    (ifp->if_flags & IFF_UP) == 0) {
		IPSEC_SC_RUNLOCK();
		m_freem(m);
		return (ENETDOWN);
	}
	/*
	 * We found matching and working tunnel.
	 * Set its ifnet as receiving interface.
	 */
	m->m_pkthdr.rcvif = ifp;
	IPSEC_SC_RUNLOCK();

	/* m_clrprotoflags(m); */
	M_SETFIB(m, ifp->if_fib);
	BPF_MTAP2(ifp, &af, sizeof(af), m);
	if_inc_counter(ifp, IFCOUNTER_IPACKETS, 1);
	if_inc_counter(ifp, IFCOUNTER_IBYTES, m->m_pkthdr.len);
	if ((ifp->if_flags & IFF_MONITOR) != 0) {
		m_freem(m);
		return (ENETDOWN);
	}
	return (0);
}

/* XXX how should we handle IPv6 scope on SIOC[GS]IFPHYADDR? */
int
ipsec_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
{
	IPSEC_RLOCK_TRACKER;
	struct ifreq *ifr = (struct ifreq*)data;
	struct sockaddr *dst, *src;
	struct ipsec_softc *sc;
	struct secasindex *saidx;
#ifdef INET
	struct sockaddr_in *sin = NULL;
#endif
#ifdef INET6
	struct sockaddr_in6 *sin6 = NULL;
#endif
	uint32_t reqid;
	int error;

	switch (cmd) {
	case SIOCSIFADDR:
		ifp->if_flags |= IFF_UP;
	case SIOCADDMULTI:
	case SIOCDELMULTI:
	case SIOCGIFMTU:
	case SIOCSIFFLAGS:
		return (0);
	case SIOCSIFMTU:
		if (ifr->ifr_mtu < IPSEC_MTU_MIN ||
		    ifr->ifr_mtu > IPSEC_MTU_MAX)
			return (EINVAL);
		else
			ifp->if_mtu = ifr->ifr_mtu;
		return (0);
	}
	sx_xlock(&ipsec_ioctl_sx);
	sc = ifp->if_softc;
	/* Check that softc is still here */
	if (sc == NULL) {
		error = ENXIO;
		goto bad;
	}
	error = 0;
	switch (cmd) {
	case SIOCSIFPHYADDR:
#ifdef INET6
	case SIOCSIFPHYADDR_IN6:
#endif
		error = EINVAL;
		switch (cmd) {
#ifdef INET
		case SIOCSIFPHYADDR:
			src = (struct sockaddr *)
				&(((struct in_aliasreq *)data)->ifra_addr);
			dst = (struct sockaddr *)
				&(((struct in_aliasreq *)data)->ifra_dstaddr);
			break;
#endif
#ifdef INET6
		case SIOCSIFPHYADDR_IN6:
			src = (struct sockaddr *)
				&(((struct in6_aliasreq *)data)->ifra_addr);
			dst = (struct sockaddr *)
				&(((struct in6_aliasreq *)data)->ifra_dstaddr);
			break;
#endif
		default:
			goto bad;
		}
		/* sa_family must be equal */
		if (src->sa_family != dst->sa_family ||
		    src->sa_len != dst->sa_len)
			goto bad;

		/* validate sa_len */
		switch (src->sa_family) {
#ifdef INET
		case AF_INET:
			if (src->sa_len != sizeof(struct sockaddr_in))
				goto bad;
			break;
#endif
#ifdef INET6
		case AF_INET6:
			if (src->sa_len != sizeof(struct sockaddr_in6))
				goto bad;
			break;
#endif
		default:
			error = EAFNOSUPPORT;
			goto bad;
		}
		/* check sa_family looks sane for the cmd */
		error = EAFNOSUPPORT;
		switch (cmd) {
#ifdef INET
		case SIOCSIFPHYADDR:
			if (src->sa_family == AF_INET)
				break;
			goto bad;
#endif
#ifdef INET6
		case SIOCSIFPHYADDR_IN6:
			if (src->sa_family == AF_INET6)
				break;
			goto bad;
#endif
		}
		error = EADDRNOTAVAIL;
		switch (src->sa_family) {
#ifdef INET
		case AF_INET:
			if (satosin(src)->sin_addr.s_addr == INADDR_ANY ||
			    satosin(dst)->sin_addr.s_addr == INADDR_ANY)
				goto bad;
			break;
#endif
#ifdef INET6
		case AF_INET6:
			if (IN6_IS_ADDR_UNSPECIFIED(&satosin6(src)->sin6_addr)
			    ||
			    IN6_IS_ADDR_UNSPECIFIED(&satosin6(dst)->sin6_addr))
				goto bad;
			/*
			 * Check validity of the scope zone ID of the
			 * addresses, and convert it into the kernel
			 * internal form if necessary.
			 */
			error = sa6_embedscope(satosin6(src), 0);
			if (error != 0)
				goto bad;
			error = sa6_embedscope(satosin6(dst), 0);
			if (error != 0)
				goto bad;
#endif
		};
		error = ipsec_set_addresses(ifp, src, dst);
		break;
	case SIOCDIFPHYADDR:
		ipsec_delete_tunnel(ifp, 0);
		break;
	case SIOCGIFPSRCADDR:
	case SIOCGIFPDSTADDR:
#ifdef INET6
	case SIOCGIFPSRCADDR_IN6:
	case SIOCGIFPDSTADDR_IN6:
#endif
		IPSEC_RLOCK(sc);
		if (sc->family == 0) {
			IPSEC_RUNLOCK(sc);
			error = EADDRNOTAVAIL;
			break;
		}
		saidx = ipsec_getsaidx(sc, IPSEC_DIR_OUTBOUND, sc->family);
		switch (cmd) {
#ifdef INET
		case SIOCGIFPSRCADDR:
		case SIOCGIFPDSTADDR:
			if (saidx->src.sa.sa_family != AF_INET) {
				error = EADDRNOTAVAIL;
				break;
			}
			sin = (struct sockaddr_in *)&ifr->ifr_addr;
			memset(sin, 0, sizeof(*sin));
			sin->sin_family = AF_INET;
			sin->sin_len = sizeof(*sin);
			break;
#endif
#ifdef INET6
		case SIOCGIFPSRCADDR_IN6:
		case SIOCGIFPDSTADDR_IN6:
			if (saidx->src.sa.sa_family != AF_INET6) {
				error = EADDRNOTAVAIL;
				break;
			}
			sin6 = (struct sockaddr_in6 *)
				&(((struct in6_ifreq *)data)->ifr_addr);
			memset(sin6, 0, sizeof(*sin6));
			sin6->sin6_family = AF_INET6;
			sin6->sin6_len = sizeof(*sin6);
			break;
#endif
		default:
			error = EAFNOSUPPORT;
		}
		if (error == 0) {
			switch (cmd) {
#ifdef INET
			case SIOCGIFPSRCADDR:
				sin->sin_addr = saidx->src.sin.sin_addr;
				break;
			case SIOCGIFPDSTADDR:
				sin->sin_addr = saidx->dst.sin.sin_addr;
				break;
#endif
#ifdef INET6
			case SIOCGIFPSRCADDR_IN6:
				sin6->sin6_addr = saidx->src.sin6.sin6_addr;
				break;
			case SIOCGIFPDSTADDR_IN6:
				sin6->sin6_addr = saidx->dst.sin6.sin6_addr;
				break;
#endif
			}
		}
		IPSEC_RUNLOCK(sc);
		if (error != 0)
			break;
		switch (cmd) {
#ifdef INET
		case SIOCGIFPSRCADDR:
		case SIOCGIFPDSTADDR:
			error = prison_if(curthread->td_ucred,
			    (struct sockaddr *)sin);
			if (error != 0)
				memset(sin, 0, sizeof(*sin));
			break;
#endif
#ifdef INET6
		case SIOCGIFPSRCADDR_IN6:
		case SIOCGIFPDSTADDR_IN6:
			error = prison_if(curthread->td_ucred,
			    (struct sockaddr *)sin6);
			if (error == 0)
				error = sa6_recoverscope(sin6);
			if (error != 0)
				memset(sin6, 0, sizeof(*sin6));
#endif
		}
		break;
	case SIOCGTUNFIB:
		ifr->ifr_fib = sc->fibnum;
		break;
	case SIOCSTUNFIB:
		if ((error = priv_check(curthread, PRIV_NET_SETIFFIB)) != 0)
			break;
		if (ifr->ifr_fib >= rt_numfibs)
			error = EINVAL;
		else
			sc->fibnum = ifr->ifr_fib;
		break;
	case IPSECGREQID:
		reqid = sc->reqid;
		error = copyout(&reqid, ifr->ifr_data, sizeof(reqid));
		break;
	case IPSECSREQID:
		if ((error = priv_check(curthread, PRIV_NET_SETIFCAP)) != 0)
			break;
		error = copyin(ifr->ifr_data, &reqid, sizeof(reqid));
		if (error != 0)
			break;
		error = ipsec_set_reqid(ifp, reqid);
		break;
	default:
		error = EINVAL;
		break;
	}
bad:
	sx_xunlock(&ipsec_ioctl_sx);
	return (error);
}

/*
 * Allocate new private security policies for tunneling interface.
 * Each tunneling interface has following security policies for
 * both AF:
 *   0.0.0.0/0[any] 0.0.0.0/0[any] -P in \
 *	ipsec esp/tunnel/RemoteIP-LocalIP/unique:reqid
 *   0.0.0.0/0[any] 0.0.0.0/0[any] -P out \
 *	ipsec esp/tunnel/LocalIP-RemoteIP/unique:reqid
 */
static int
ipsec_newpolicies(struct secpolicy *sp[IPSEC_SPCOUNT],
    const struct sockaddr *src, const struct sockaddr *dst, uint32_t reqid)
{
	struct ipsecrequest *isr;
	int i;

	memset(sp, 0, sizeof(struct secpolicy *) * IPSEC_SPCOUNT);
	for (i = 0; i < IPSEC_SPCOUNT; i++) {
		if ((sp[i] = key_newsp()) == NULL)
			goto fail;
		if ((isr = ipsec_newisr()) == NULL)
			goto fail;

		sp[i]->policy = IPSEC_POLICY_IPSEC;
		sp[i]->state = IPSEC_SPSTATE_DEAD;
		sp[i]->req[sp[i]->tcount++] = isr;
		sp[i]->created = time_second;
		isr->level = IPSEC_LEVEL_UNIQUE;
		isr->saidx.proto = IPPROTO_ESP;
		isr->saidx.mode = IPSEC_MODE_TUNNEL;
		isr->saidx.reqid = reqid;
		if (i % 2 == 0) {
			sp[i]->spidx.dir = IPSEC_DIR_INBOUND;
			bcopy(src, &isr->saidx.dst, src->sa_len);
			bcopy(dst, &isr->saidx.src, dst->sa_len);
		} else {
			sp[i]->spidx.dir = IPSEC_DIR_OUTBOUND;
			bcopy(src, &isr->saidx.src, src->sa_len);
			bcopy(dst, &isr->saidx.dst, dst->sa_len);
		}
		sp[i]->spidx.ul_proto = IPSEC_ULPROTO_ANY;
#ifdef INET
		if (i < 2) {
			sp[i]->spidx.src.sa.sa_family =
			    sp[i]->spidx.dst.sa.sa_family = AF_INET;
			sp[i]->spidx.src.sa.sa_len =
			    sp[i]->spidx.dst.sa.sa_len =
			    sizeof(struct sockaddr_in);
			continue;
		}
#endif
#ifdef INET6
		sp[i]->spidx.src.sa.sa_family =
		    sp[i]->spidx.dst.sa.sa_family = AF_INET6;
		sp[i]->spidx.src.sa.sa_len =
		    sp[i]->spidx.dst.sa.sa_len = sizeof(struct sockaddr_in6);
#endif
	}
	return (0);
fail:
	for (i = 0; i < IPSEC_SPCOUNT; i++)
		key_freesp(&sp[i]);
	return (ENOMEM);
}

static int
ipsec_check_reqid(uint32_t reqid)
{
	struct ipsec_softc *sc;

	IPSEC_SC_RLOCK_ASSERT();
	LIST_FOREACH(sc, &V_ipsec_sc_list, chain) {
		if (sc->reqid == reqid)
			return (EEXIST);
	}
	return (0);
}

/*
 * We use key_newreqid() to automatically obtain unique reqid.
 * Then we check that given id is unique, i.e. it is not used by
 * another if_ipsec(4) interface. This macro limits the number of
 * tries to get unique id.
 */
#define	IPSEC_REQID_TRYCNT	64
static int
ipsec_init_reqid(struct ipsec_softc *sc)
{
	uint32_t reqid;
	int trycount;

	IPSEC_SC_RLOCK_ASSERT();

	if (sc->reqid != 0) /* already initialized */
		return (0);

	trycount = IPSEC_REQID_TRYCNT;
	while (--trycount > 0) {
		reqid = key_newreqid();
		if (ipsec_check_reqid(reqid) == 0)
			break;
	}
	if (trycount == 0)
		return (EEXIST);
	sc->reqid = reqid;
	return (0);
}

/*
 * Set or update reqid for given tunneling interface.
 * When specified reqid is zero, generate new one.
 * We are protected by ioctl_sx lock from concurrent id generation.
 * Also softc would not disappear while we hold ioctl_sx lock.
 */
static int
ipsec_set_reqid(struct ifnet *ifp, uint32_t reqid)
{
	IPSEC_SC_RLOCK_TRACKER;
	struct ipsec_softc *sc;
	struct secasindex *saidx;

	sx_assert(&ipsec_ioctl_sx, SA_XLOCKED);

	sc = ifp->if_softc;
	if (sc->reqid == reqid && reqid != 0)
		return (0);

	IPSEC_SC_RLOCK();
	if (reqid != 0) {
		/* Check that specified reqid doesn't exist */
		if (ipsec_check_reqid(reqid) != 0) {
			IPSEC_SC_RUNLOCK();
			return (EEXIST);
		}
		sc->reqid = reqid;
	} else {
		/* Generate new reqid */
		if (ipsec_init_reqid(sc) != 0) {
			IPSEC_SC_RUNLOCK();
			return (EEXIST);
		}
	}
	IPSEC_SC_RUNLOCK();

	/* Tunnel isn't fully configured, just return. */
	if (sc->family == 0)
		return (0);

	saidx = ipsec_getsaidx(sc, IPSEC_DIR_OUTBOUND, sc->family);
	KASSERT(saidx != NULL,
	    ("saidx is NULL, but family is %d", sc->family));
	return (ipsec_set_tunnel(sc, &saidx->src.sa, &saidx->dst.sa,
	    sc->reqid));
}

/*
 * Set tunnel endpoints addresses.
 */
static int
ipsec_set_addresses(struct ifnet *ifp, struct sockaddr *src,
    struct sockaddr *dst)
{
	IPSEC_SC_RLOCK_TRACKER;
	struct ipsec_softc *sc, *tsc;
	struct secasindex *saidx;

	sx_assert(&ipsec_ioctl_sx, SA_XLOCKED);

	sc = ifp->if_softc;
	if (sc->family != 0) {
		saidx = ipsec_getsaidx(sc, IPSEC_DIR_OUTBOUND,
		    src->sa_family);
		if (saidx != NULL && saidx->reqid == sc->reqid &&
		    key_sockaddrcmp(&saidx->src.sa, src, 0) == 0 &&
		    key_sockaddrcmp(&saidx->dst.sa, dst, 0) == 0)
			return (0); /* Nothing has been changed. */

	}
	/*
	 * We cannot service IPsec tunnel when source address is
	 * not our own.
	 */
#ifdef INET
	if (src->sa_family == AF_INET &&
	    in_localip(satosin(src)->sin_addr) == 0)
		return (EADDRNOTAVAIL);
#endif
#ifdef INET6
	/*
	 * NOTE: IPv6 addresses are in kernel internal form with
	 * embedded scope zone id.
	 */
	if (src->sa_family == AF_INET6 &&
	    in6_localip(&satosin6(src)->sin6_addr) == 0)
		return (EADDRNOTAVAIL);
#endif
	/* Check that given addresses aren't already configured */
	IPSEC_SC_RLOCK();
	LIST_FOREACH(tsc, &V_ipsec_sc_list, chain) {
		if (tsc == sc || tsc->family != src->sa_family)
			continue;
		saidx = ipsec_getsaidx(tsc, IPSEC_DIR_OUTBOUND, tsc->family);
		if (key_sockaddrcmp(&saidx->src.sa, src, 0) == 0 &&
		    key_sockaddrcmp(&saidx->dst.sa, dst, 0) == 0) {
			/* We already have tunnel with such addresses */
			IPSEC_SC_RUNLOCK();
			return (EADDRNOTAVAIL);
		}
	}
	/* If reqid is not set, generate new one. */
	if (ipsec_init_reqid(sc) != 0) {
		IPSEC_SC_RUNLOCK();
		return (EEXIST);
	}
	IPSEC_SC_RUNLOCK();
	return (ipsec_set_tunnel(sc, src, dst, sc->reqid));
}

static int
ipsec_set_tunnel(struct ipsec_softc *sc, struct sockaddr *src,
    struct sockaddr *dst, uint32_t reqid)
{
	struct secpolicy *sp[IPSEC_SPCOUNT];
	struct secpolicy *oldsp[IPSEC_SPCOUNT];
	int i, f;

	sx_assert(&ipsec_ioctl_sx, SA_XLOCKED);

	/* Allocate SP with new addresses. */
	if (ipsec_newpolicies(sp, src, dst, reqid) == 0) {
		/* Add new policies to SPDB */
		if (key_register_ifnet(sp, IPSEC_SPCOUNT) != 0) {
			for (i = 0; i < IPSEC_SPCOUNT; i++)
				key_freesp(&sp[i]);
			return (EAGAIN);
		}
		IPSEC_SC_WLOCK();
		if ((f = sc->family) != 0)
			LIST_REMOVE(sc, hash);
		IPSEC_WLOCK(sc);
		for (i = 0; i < IPSEC_SPCOUNT; i++) {
			oldsp[i] = sc->sp[i];
			sc->sp[i] = sp[i];
		}
		sc->family = src->sa_family;
		IPSEC_WUNLOCK(sc);
		LIST_INSERT_HEAD(SCHASH_HASH(sc->reqid), sc, hash);
		IPSEC_SC_WUNLOCK();
	} else {
		sc->ifp->if_drv_flags &= ~IFF_DRV_RUNNING;
		return (ENOMEM);
	}

	sc->ifp->if_drv_flags |= IFF_DRV_RUNNING;
	if (f != 0) {
		key_unregister_ifnet(oldsp, IPSEC_SPCOUNT);
		for (i = 0; i < IPSEC_SPCOUNT; i++)
			key_freesp(&oldsp[i]);
	}
	return (0);
}

static void
ipsec_delete_tunnel(struct ifnet *ifp, int locked)
{
	struct ipsec_softc *sc = ifp->if_softc;
	struct secpolicy *oldsp[IPSEC_SPCOUNT];
	int i;

	sx_assert(&ipsec_ioctl_sx, SA_XLOCKED);

	ifp->if_drv_flags &= ~IFF_DRV_RUNNING;
	if (sc->family != 0) {
		if (!locked)
			IPSEC_SC_WLOCK();
		/* Remove from hash table */
		LIST_REMOVE(sc, hash);
		IPSEC_WLOCK(sc);
		for (i = 0; i < IPSEC_SPCOUNT; i++) {
			oldsp[i] = sc->sp[i];
			sc->sp[i] = NULL;
		}
		sc->family = 0;
		IPSEC_WUNLOCK(sc);
		if (!locked)
			IPSEC_SC_WUNLOCK();
		key_unregister_ifnet(oldsp, IPSEC_SPCOUNT);
		for (i = 0; i < IPSEC_SPCOUNT; i++)
			key_freesp(&oldsp[i]);
	}
}