uipc_domain.c   [plain text]

/*
 * Copyright (c) 1998-2013 Apple Inc. All rights reserved.
 *
 * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
 *
 * This file contains Original Code and/or Modifications of Original Code
 * as defined in and that are subject to the Apple Public Source License
 * Version 2.0 (the 'License'). You may not use this file except in
 * compliance with the License. The rights granted to you under the License
 * may not be used to create, or enable the creation or redistribution of,
 * unlawful or unlicensed copies of an Apple operating system, or to
 * circumvent, violate, or enable the circumvention or violation of, any
 * terms of an Apple operating system software license agreement.
 *
 * Please obtain a copy of the License at
 * http://www.opensource.apple.com/apsl/ and read it before using this file.
 *
 * The Original Code and all software distributed under the License are
 * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
 * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
 * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
 * Please see the License for the specific language governing rights and
 * limitations under the License.
 *
 * @APPLE_OSREFERENCE_LICENSE_HEADER_END@
 */
/* Copyright (c) 1995 NeXT Computer, Inc. All Rights Reserved */
/*
 * Copyright (c) 1982, 1986, 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. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *	This product includes software developed by the University of
 *	California, Berkeley and its contributors.
 * 4. 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.
 *
 *	@(#)uipc_domain.c	8.3 (Berkeley) 2/14/95
 */

#include <sys/param.h>
#include <sys/socket.h>
#include <sys/protosw.h>
#include <sys/domain.h>
#include <sys/mcache.h>
#include <sys/mbuf.h>
#include <sys/time.h>
#include <sys/kernel.h>
#include <sys/systm.h>
#include <sys/proc_internal.h>
#include <sys/sysctl.h>
#include <sys/syslog.h>
#include <sys/queue.h>

#include <net/dlil.h>

#include <mach/boolean.h>
#include <pexpert/pexpert.h>

static void pr_init_old(struct protosw *, struct domain *);
static void init_proto(struct protosw *, struct domain *);
static void attach_proto(struct protosw *, struct domain *);
static void detach_proto(struct protosw *, struct domain *);
static void dom_init_old(struct domain *);
static void init_domain(struct domain *);
static void attach_domain(struct domain *);
static void detach_domain(struct domain *);
static struct protosw *pffindprotonotype_locked(int, int, int);
static struct domain *pffinddomain_locked(int);

static boolean_t domain_timeout_run;	/* domain timer is scheduled to run */
static boolean_t domain_draining;
static void domain_sched_timeout(void);
static void domain_timeout(void *);

lck_grp_t	*domain_proto_mtx_grp;
lck_attr_t	*domain_proto_mtx_attr;
static lck_grp_attr_t	*domain_proto_mtx_grp_attr;
decl_lck_mtx_data(static, domain_proto_mtx);
decl_lck_mtx_data(static, domain_timeout_mtx);

extern sysctlfn net_sysctl;

static u_int64_t _net_uptime;

static void
pr_init_old(struct protosw *pp, struct domain *dp)
{
#pragma unused(dp)
	VERIFY(pp->pr_flags & PR_OLD);
	VERIFY(pp->pr_old != NULL);

	if (pp->pr_old->pr_init != NULL)
		pp->pr_old->pr_init();
}

static void
init_proto(struct protosw *pp, struct domain *dp)
{
	VERIFY(pp->pr_flags & PR_ATTACHED);

	if (!(pp->pr_flags & PR_INITIALIZED)) {
		TAILQ_INIT(&pp->pr_filter_head);
		if (pp->pr_init != NULL)
			pp->pr_init(pp, dp);
		pp->pr_flags |= PR_INITIALIZED;
	}
}

static void
attach_proto(struct protosw *pp, struct domain *dp)
{
	domain_proto_mtx_lock_assert_held();
	VERIFY(!(pp->pr_flags & PR_ATTACHED));
	VERIFY(pp->pr_domain == NULL);
	VERIFY(pp->pr_protosw == NULL);

	TAILQ_INSERT_TAIL(&dp->dom_protosw, pp, pr_entry);
	pp->pr_flags |= PR_ATTACHED;
	pp->pr_domain = dp;
	pp->pr_protosw = pp;

	/* do some cleaning up on user request callbacks */
	pru_sanitize(pp->pr_usrreqs);
}

static void
detach_proto(struct protosw *pp, struct domain *dp)
{
	domain_proto_mtx_lock_assert_held();
	VERIFY(pp->pr_flags & PR_ATTACHED);
	VERIFY(pp->pr_domain == dp);
	VERIFY(pp->pr_protosw == pp);

	TAILQ_REMOVE(&dp->dom_protosw, pp, pr_entry);
	pp->pr_flags &= ~PR_ATTACHED;
	pp->pr_domain = NULL;
	pp->pr_protosw = NULL;
}

static void
dom_init_old(struct domain *dp)
{
	VERIFY(dp->dom_flags & DOM_OLD);
	VERIFY(dp->dom_old != NULL);

	if (dp->dom_old->dom_init != NULL)
		dp->dom_old->dom_init();
}

static void
init_domain(struct domain *dp)
{
	VERIFY(dp->dom_flags & DOM_ATTACHED);

	if (!(dp->dom_flags & DOM_INITIALIZED)) {
		lck_mtx_init(&dp->dom_mtx_s, domain_proto_mtx_grp,
		    domain_proto_mtx_attr);
		dp->dom_mtx = &dp->dom_mtx_s;
		TAILQ_INIT(&dp->dom_protosw);
		if (dp->dom_init != NULL)
			dp->dom_init(dp);
		dp->dom_flags |= DOM_INITIALIZED;
	}

	/* Recompute for new protocol */
	if (_max_linkhdr < 16)		/* XXX - Sheesh; everything's ether? */
		_max_linkhdr = 16;
	_max_linkhdr = max_linkhdr;	/* round it up */

	if (dp->dom_protohdrlen > _max_protohdr)
		_max_protohdr = dp->dom_protohdrlen;
	_max_protohdr = max_protohdr;	/* round it up */

	max_hdr = max_linkhdr + max_protohdr;
	max_datalen = MHLEN - max_hdr;
}

static void
attach_domain(struct domain *dp)
{
	domain_proto_mtx_lock_assert_held();
	VERIFY(!(dp->dom_flags & DOM_ATTACHED));

	TAILQ_INSERT_TAIL(&domains, dp, dom_entry);
	dp->dom_flags |= DOM_ATTACHED;
}

static void
detach_domain(struct domain *dp)
{
	domain_proto_mtx_lock_assert_held();
	VERIFY(dp->dom_flags & DOM_ATTACHED);

	TAILQ_REMOVE(&domains, dp, dom_entry);
	dp->dom_flags &= ~DOM_ATTACHED;

	if (dp->dom_flags & DOM_OLD) {
		struct domain_old *odp = dp->dom_old;

		VERIFY(odp != NULL);
		odp->dom_next = NULL;
		odp->dom_mtx = NULL;
	}
}

/*
 * Exported (private) routine, indirection of net_add_domain.
 */
void
net_add_domain_old(struct domain_old *odp)
{
	struct domain *dp;
	domain_guard_t guard;

	VERIFY(odp != NULL);

	guard = domain_guard_deploy();
	if ((dp = pffinddomain_locked(odp->dom_family)) != NULL) {
		/*
		 * There is really nothing better than to panic here,
		 * as the caller would not have been able to handle
		 * any failures otherwise.
		 */
		panic("%s: domain (%d,%s) already exists for %s\n", __func__,
		    dp->dom_family, dp->dom_name, odp->dom_name);
		/* NOTREACHED */
	}

	/* Make sure nothing is currently pointing to the odp. */
	TAILQ_FOREACH(dp, &domains, dom_entry) {
		if (dp->dom_old == odp) {
			panic("%s: domain %p (%d,%s) is already "
			    "associated with %p (%d,%s)\n", __func__,
			    odp, odp->dom_family, odp->dom_name, dp,
			    dp->dom_family, dp->dom_name);
			/* NOTREACHED */
		}
	}

	if (odp->dom_protosw != NULL) {
		panic("%s: domain (%d,%s) protocols need to added "
		    "via net_add_proto\n", __func__, odp->dom_family,
		    odp->dom_name);
		/* NOTREACHED */
	}

	dp = _MALLOC(sizeof (*dp), M_TEMP, M_WAITOK | M_ZERO);
	if (dp == NULL) {
		/*
		 * There is really nothing better than to panic here,
		 * as the caller would not have been able to handle
		 * any failures otherwise.
		 */
		panic("%s: unable to allocate memory for domain family "
		    "%d (%s)\n", __func__, odp->dom_family, odp->dom_name);
		/* NOTREACHED */
	}

	/* Copy everything but dom_init, dom_mtx, dom_next and dom_refs */
	dp->dom_family		= odp->dom_family;
	dp->dom_flags		= (odp->dom_flags & DOMF_USERFLAGS) | DOM_OLD;
	dp->dom_name		= odp->dom_name;
	dp->dom_init		= dom_init_old;
	dp->dom_externalize	= odp->dom_externalize;
	dp->dom_dispose		= odp->dom_dispose;
	dp->dom_rtattach	= odp->dom_rtattach;
	dp->dom_rtoffset	= odp->dom_rtoffset;
	dp->dom_maxrtkey	= odp->dom_maxrtkey;
	dp->dom_protohdrlen	= odp->dom_protohdrlen;
	dp->dom_old		= odp;

	attach_domain(dp);
	init_domain(dp);

	/* Point the mutex back to the internal structure's */
	odp->dom_mtx		= dp->dom_mtx;
	domain_guard_release(guard);
}

/*
 * Exported (private) routine, indirection of net_del_domain.
 */
int
net_del_domain_old(struct domain_old *odp)
{
	struct domain *dp1, *dp2;
	int error = 0;
	domain_guard_t guard;

	VERIFY(odp != NULL);

	guard = domain_guard_deploy();
	if (odp->dom_refs != 0) {
		error = EBUSY;
		goto done;
	}

	TAILQ_FOREACH_SAFE(dp1, &domains, dom_entry, dp2) {
		if (!(dp1->dom_flags & DOM_OLD))
			continue;
		VERIFY(dp1->dom_old != NULL);
		if (odp == dp1->dom_old)
			break;
	}
	if (dp1 != NULL) {
		struct protosw *pp1, *pp2;

		VERIFY(dp1->dom_flags & DOM_OLD);
		VERIFY(dp1->dom_old == odp);

		/* Remove all protocols attached to this domain */
		TAILQ_FOREACH_SAFE(pp1, &dp1->dom_protosw, pr_entry, pp2) {
			detach_proto(pp1, dp1);
			if (pp1->pr_usrreqs->pru_flags & PRUF_OLD)
				FREE(pp1->pr_usrreqs, M_TEMP);
			if (pp1->pr_flags & PR_OLD)
				FREE(pp1, M_TEMP);
		}

		detach_domain(dp1);
		FREE(dp1, M_TEMP);
	} else {
		error = EPFNOSUPPORT;
	}
done:
	domain_guard_release(guard);
	return (error);
}

/*
 * Internal routine, not exported.
 *
 * net_add_proto - link a protosw into a domain's protosw chain
 *
 * NOTE: Caller must have acquired domain_proto_mtx
 */
int
net_add_proto(struct protosw *pp, struct domain *dp, int doinit)
{
	struct protosw *pp1;

	/*
	 * This could be called as part of initializing the domain,
	 * and thus DOM_INITIALIZED may not be set (yet).
	 */
	domain_proto_mtx_lock_assert_held();
	VERIFY(!(pp->pr_flags & PR_ATTACHED));

	/* pr_domain is set only after the protocol is attached */
	if (pp->pr_domain != NULL) {
		panic("%s: domain (%d,%s), proto %d has non-NULL pr_domain!\n",
		    __func__, dp->dom_family, dp->dom_name, pp->pr_protocol);
		/* NOTREACHED */
	}

	if (pp->pr_usrreqs == NULL) {
		panic("%s: domain (%d,%s), proto %d has no usrreqs!\n",
		    __func__, dp->dom_family, dp->dom_name, pp->pr_protocol);
		/* NOTREACHED */
	}

	TAILQ_FOREACH(pp1, &dp->dom_protosw, pr_entry) {
		if (pp1->pr_type == pp->pr_type &&
		    pp1->pr_protocol == pp->pr_protocol)
			return (EEXIST);
	}

	attach_proto(pp, dp);
	if (doinit)
		net_init_proto(pp, dp);

	return (0);
}

void
net_init_proto(struct protosw *pp, struct domain *dp)
{
	/*
	 * This could be called as part of initializing the domain,
	 * and thus DOM_INITIALIZED may not be set (yet).  The protocol
	 * must have been attached via net_addr_protosw() by now.
	 */
	domain_proto_mtx_lock_assert_held();
	VERIFY(pp->pr_flags & PR_ATTACHED);

	init_proto(pp, dp);
}

/*
 * Exported (private) routine, indirection of net_add_proto.
 */
int
net_add_proto_old(struct protosw_old *opp, struct domain_old *odp)
{
	struct pr_usrreqs_old *opru;
	struct pr_usrreqs *pru = NULL;
	struct protosw *pp = NULL, *pp1;
	int error = 0;
	struct domain *dp;
	domain_guard_t guard;

	/*
	 * This could be called as part of initializing the domain,
	 * and thus DOM_INITIALIZED may not be set (yet).
	 */
	guard = domain_guard_deploy();

	/* Make sure the domain has been added via net_add_domain */
	TAILQ_FOREACH(dp, &domains, dom_entry) {
		if (!(dp->dom_flags & DOM_OLD))
			continue;
		if (dp->dom_old == odp)
			break;
	}
	if (dp == NULL) {
		error = EINVAL;
		goto done;
	}

	TAILQ_FOREACH(pp1, &dp->dom_protosw, pr_entry) {
		if (pp1->pr_type == opp->pr_type &&
		    pp1->pr_protocol == opp->pr_protocol) {
			error = EEXIST;
			goto done;
		}
	}

	if ((opru = opp->pr_usrreqs) == NULL) {
		panic("%s: domain (%d,%s), proto %d has no usrreqs!\n",
		    __func__, odp->dom_family, odp->dom_name, opp->pr_protocol);
		/* NOTREACHED */
	}

	pru = _MALLOC(sizeof (*pru), M_TEMP, M_WAITOK | M_ZERO);
	if (pru == NULL) {
		error = ENOMEM;
		goto done;
	}

	pru->pru_flags		= PRUF_OLD;
	pru->pru_abort		= opru->pru_abort;
	pru->pru_accept		= opru->pru_accept;
	pru->pru_attach		= opru->pru_attach;
	pru->pru_bind		= opru->pru_bind;
	pru->pru_connect	= opru->pru_connect;
	pru->pru_connect2	= opru->pru_connect2;
	pru->pru_control	= opru->pru_control;
	pru->pru_detach		= opru->pru_detach;
	pru->pru_disconnect	= opru->pru_disconnect;
	pru->pru_listen		= opru->pru_listen;
	pru->pru_peeraddr	= opru->pru_peeraddr;
	pru->pru_rcvd		= opru->pru_rcvd;
	pru->pru_rcvoob		= opru->pru_rcvoob;
	pru->pru_send		= opru->pru_send;
	pru->pru_sense		= opru->pru_sense;
	pru->pru_shutdown	= opru->pru_shutdown;
	pru->pru_sockaddr	= opru->pru_sockaddr;
	pru->pru_sosend		= opru->pru_sosend;
	pru->pru_soreceive	= opru->pru_soreceive;
	pru->pru_sopoll		= opru->pru_sopoll;

	pp = _MALLOC(sizeof (*pp), M_TEMP, M_WAITOK | M_ZERO);
	if (pp == NULL) {
		error = ENOMEM;
		goto done;
	}

	/*
	 * Protocol fast and slow timers are now deprecated.
	 */
	if (opp->pr_unused != NULL) {
		printf("%s: domain (%d,%s), proto %d: pr_fasttimo is "
		    "deprecated and won't be called\n", __func__,
		    odp->dom_family, odp->dom_name, opp->pr_protocol);
	}
	if (opp->pr_unused2 != NULL) {
		printf("%s: domain (%d,%s), proto %d: pr_slowtimo is "
		    "deprecated and won't be called\n", __func__,
		    odp->dom_family, odp->dom_name, opp->pr_protocol);
	}

	/* Copy everything but pr_init, pr_next, pr_domain, pr_protosw */
	pp->pr_type		= opp->pr_type;
	pp->pr_protocol		= opp->pr_protocol;
	pp->pr_flags		= (opp->pr_flags & PRF_USERFLAGS) | PR_OLD;
	pp->pr_input		= opp->pr_input;
	pp->pr_output		= opp->pr_output;
	pp->pr_ctlinput		= opp->pr_ctlinput;
	pp->pr_ctloutput	= opp->pr_ctloutput;
	pp->pr_usrreqs		= pru;
	pp->pr_init		= pr_init_old;
	pp->pr_drain		= opp->pr_drain;
	pp->pr_sysctl		= opp->pr_sysctl;
	pp->pr_lock		= opp->pr_lock;
	pp->pr_unlock		= opp->pr_unlock;
	pp->pr_getlock		= opp->pr_getlock;
	pp->pr_old		= opp;

	/* attach as well as initialize */
	attach_proto(pp, dp);
	net_init_proto(pp, dp);
done:
	if (error != 0) {
		printf("%s: domain (%d,%s), proto %d: failed to attach, "
		    "error %d\n", __func__, odp->dom_family,
		    odp->dom_name, opp->pr_protocol, error);

		if (pru != NULL)
			FREE(pru, M_TEMP);
		if (pp != NULL)
			FREE(pp, M_TEMP);
	}

	domain_guard_release(guard);
	return (error);
}

/*
 * Internal routine, not exported.
 *
 * net_del_proto - remove a protosw from a domain's protosw chain.
 * Search the protosw chain for the element with matching data.
 * Then unlink and return.
 *
 * NOTE: Caller must have acquired domain_proto_mtx
 */
int
net_del_proto(int type, int protocol, struct domain *dp)
{
	struct protosw *pp;

	/*
	 * This could be called as part of initializing the domain,
	 * and thus DOM_INITIALIZED may not be set (yet).
	 */
	domain_proto_mtx_lock_assert_held();

	TAILQ_FOREACH(pp, &dp->dom_protosw, pr_entry) {
		if (pp->pr_type == type && pp->pr_protocol == protocol)
			break;
	}
	if (pp == NULL)
		return (ENXIO);

	detach_proto(pp, dp);
	if (pp->pr_usrreqs->pru_flags & PRUF_OLD)
		FREE(pp->pr_usrreqs, M_TEMP);
	if (pp->pr_flags & PR_OLD)
		FREE(pp, M_TEMP);

	return (0);
}

/*
 * Exported (private) routine, indirection of net_del_proto.
 */
int
net_del_proto_old(int type, int protocol, struct domain_old *odp)
{
	int error = 0;
	struct protosw *pp;
	struct domain *dp;
	domain_guard_t guard;

	/*
	 * This could be called as part of initializing the domain,
	 * and thus DOM_INITIALIZED may not be set (yet).
	 */
	guard = domain_guard_deploy();

	/* Make sure the domain has been added via net_add_domain */
	TAILQ_FOREACH(dp, &domains, dom_entry) {
		if (!(dp->dom_flags & DOM_OLD))
			continue;
		if (dp->dom_old == odp)
			break;
	}
	if (dp == NULL) {
		error = ENXIO;
		goto done;
	}

	TAILQ_FOREACH(pp, &dp->dom_protosw, pr_entry) {
		if (pp->pr_type == type && pp->pr_protocol == protocol)
			break;
	}
	if (pp == NULL) {
		error = ENXIO;
		goto done;
	}
	detach_proto(pp, dp);
	if (pp->pr_usrreqs->pru_flags & PRUF_OLD)
		FREE(pp->pr_usrreqs, M_TEMP);
	if (pp->pr_flags & PR_OLD)
		FREE(pp, M_TEMP);

done:
	domain_guard_release(guard);
	return (error);
}

static void
domain_sched_timeout(void)
{
	lck_mtx_assert(&domain_timeout_mtx, LCK_MTX_ASSERT_OWNED);

	if (!domain_timeout_run && domain_draining) {
		domain_timeout_run = TRUE;
		timeout(domain_timeout, NULL, hz);
	}
}

void
net_drain_domains(void)
{
	lck_mtx_lock(&domain_timeout_mtx);
	domain_draining = TRUE;
	domain_sched_timeout();
	lck_mtx_unlock(&domain_timeout_mtx);
}

#if INET6
extern struct domain inet6domain_s;
#endif
#if IPSEC
extern struct domain keydomain_s;
#endif

extern struct domain routedomain_s, ndrvdomain_s, inetdomain_s;
extern struct domain systemdomain_s, localdomain_s;

#if MULTIPATH
extern struct domain mpdomain_s;
#endif /* MULTIPATH */

static void
domain_timeout(void *arg)
{
#pragma unused(arg)
	struct protosw *pp;
	struct domain *dp;
	domain_guard_t guard;

	lck_mtx_lock(&domain_timeout_mtx);
	if (domain_draining) {
		domain_draining = FALSE;
		lck_mtx_unlock(&domain_timeout_mtx);

		guard = domain_guard_deploy();
		TAILQ_FOREACH(dp, &domains, dom_entry) {
			TAILQ_FOREACH(pp, &dp->dom_protosw, pr_entry) {
				if (pp->pr_drain != NULL)
					(*pp->pr_drain)();
			}
		}
		domain_guard_release(guard);

		lck_mtx_lock(&domain_timeout_mtx);
	}

	/* re-arm the timer if there's work to do */
	domain_timeout_run = FALSE;
	domain_sched_timeout();
	lck_mtx_unlock(&domain_timeout_mtx);
}

void
domaininit(void)
{
	struct domain *dp;
	domain_guard_t guard;

	/*
	 * allocate lock group attribute and group for domain mutexes
	 */
	domain_proto_mtx_grp_attr = lck_grp_attr_alloc_init();

	domain_proto_mtx_grp = lck_grp_alloc_init("domain",
	    domain_proto_mtx_grp_attr);

	/*
	 * allocate the lock attribute for per domain mutexes
	 */
	domain_proto_mtx_attr = lck_attr_alloc_init();

	lck_mtx_init(&domain_proto_mtx, domain_proto_mtx_grp,
	    domain_proto_mtx_attr);
	lck_mtx_init(&domain_timeout_mtx, domain_proto_mtx_grp,
	    domain_proto_mtx_attr);

	guard = domain_guard_deploy();
	/*
	 * Add all the static domains to the domains list.  route domain
	 * gets added and initialized last, since we need it to attach
	 * rt_tables[] to everything that's already there.  This also
	 * means that domains added after this point won't get their
	 * dom_rtattach() called on rt_tables[].
	 */
	attach_domain(&inetdomain_s);
#if INET6
	attach_domain(&inet6domain_s);
#endif /* INET6 */
#if MULTIPATH
	attach_domain(&mpdomain_s);
#endif /* MULTIPATH */
	attach_domain(&systemdomain_s);
	attach_domain(&localdomain_s);
#if IPSEC
	attach_domain(&keydomain_s);
#endif /* IPSEC */
	attach_domain(&ndrvdomain_s);
	attach_domain(&routedomain_s);	/* must be last domain */

	/*
	 * Now ask them all to init (XXX including the routing domain,
	 * see above)
	 */
	TAILQ_FOREACH(dp, &domains, dom_entry)
		init_domain(dp);

	domain_guard_release(guard);
}

static __inline__ struct domain *
pffinddomain_locked(int pf)
{
	struct domain *dp;

	domain_proto_mtx_lock_assert_held();

	TAILQ_FOREACH(dp, &domains, dom_entry) {
		if (dp->dom_family == pf)
			break;
	}
	return (dp);
}

struct protosw *
pffindtype(int family, int type)
{
	struct protosw *pp = NULL;
	struct domain *dp;
	domain_guard_t guard;

	guard = domain_guard_deploy();
	if ((dp = pffinddomain_locked(family)) == NULL)
		goto done;

	TAILQ_FOREACH(pp, &dp->dom_protosw, pr_entry) {
		if (pp->pr_type != 0 && pp->pr_type == type)
			goto done;
	}
done:
	domain_guard_release(guard);
	return (pp);
}

/*
 * Internal routine, not exported.
 */
struct domain *
pffinddomain(int pf)
{
	struct domain *dp;
	domain_guard_t guard;

	guard = domain_guard_deploy();
	dp = pffinddomain_locked(pf);
	domain_guard_release(guard);
	return (dp);
}

/*
 * Exported (private) routine, indirection of pffinddomain.
 */
struct domain_old *
pffinddomain_old(int pf)
{
	struct domain_old *odp = NULL;
	struct domain *dp;
	domain_guard_t guard;

	guard = domain_guard_deploy();
	if ((dp = pffinddomain_locked(pf)) != NULL && (dp->dom_flags & DOM_OLD))
		odp = dp->dom_old;
	domain_guard_release(guard);
	return (odp);
}

/*
 * Internal routine, not exported.
 */
struct protosw *
pffindproto(int family, int protocol, int type)
{
	struct protosw *pp;
	domain_guard_t guard;

	guard = domain_guard_deploy();
	pp = pffindproto_locked(family, protocol, type);
	domain_guard_release(guard);
	return (pp);
}

struct protosw *
pffindproto_locked(int family, int protocol, int type)
{
	struct protosw *maybe = NULL;
	struct protosw *pp;
	struct domain *dp;

	domain_proto_mtx_lock_assert_held();

	if (family == 0)
		return (0);

	dp = pffinddomain_locked(family);
	if (dp == NULL)
		return (NULL);

	TAILQ_FOREACH(pp, &dp->dom_protosw, pr_entry) {
		if ((pp->pr_protocol == protocol) && (pp->pr_type == type))
			return (pp);

		if (type == SOCK_RAW && pp->pr_type == SOCK_RAW &&
		    pp->pr_protocol == 0 && maybe == NULL)
			maybe = pp;
	}
	return (maybe);
}

/*
 * Exported (private) routine, indirection of pffindproto.
 */
struct protosw_old *
pffindproto_old(int family, int protocol, int type)
{
	struct protosw_old *opr = NULL;
	struct protosw *pp;
	domain_guard_t guard;

	guard = domain_guard_deploy();
	if ((pp = pffindproto_locked(family, protocol, type)) != NULL &&
	    (pp->pr_flags & PR_OLD))
		opr = pp->pr_old;
	domain_guard_release(guard);
	return (opr);
}

static struct protosw *
pffindprotonotype_locked(int family, int protocol, int type)
{
#pragma unused(type)
	struct domain *dp;
	struct protosw *pp;

	domain_proto_mtx_lock_assert_held();

	if (family == 0)
		return (0);

	dp = pffinddomain_locked(family);
	if (dp == NULL)
		return (NULL);

	TAILQ_FOREACH(pp, &dp->dom_protosw, pr_entry) {
		if (pp->pr_protocol == protocol)
			return (pp);
	}
	return (NULL);
}

struct protosw *
pffindprotonotype(int family, int protocol)
{
	struct protosw *pp;
	domain_guard_t guard;

	if (protocol == 0)
		return (NULL);

	guard = domain_guard_deploy();
	pp = pffindprotonotype_locked(family, protocol, 0);
	domain_guard_release(guard);
	return (pp);
}

int
net_sysctl(int *name, u_int namelen, user_addr_t oldp, size_t *oldlenp,
    user_addr_t newp, size_t newlen, struct proc *p)
{
#pragma unused(p)
	int family, protocol, error = 0;
	struct domain *dp;
	struct protosw *pp;
	domain_guard_t guard;

	/*
	 * All sysctl names at this level are nonterminal;
	 * next two components are protocol family and protocol number,
	 * then at least one addition component.
	 */
	if (namelen < 3)
		return (EISDIR);		/* overloaded */
	family = name[0];
	protocol = name[1];

	if (family == 0)
		return (0);

	guard = domain_guard_deploy();
	TAILQ_FOREACH(dp, &domains, dom_entry) {
		if (dp->dom_family == family)
			break;
	}
	if (dp == NULL) {
		error = ENOPROTOOPT;
		goto done;
	}

	TAILQ_FOREACH(pp, &dp->dom_protosw, pr_entry) {
		if (pp->pr_protocol == protocol && pp->pr_sysctl != NULL) {
			error = (*pp->pr_sysctl)(name + 2, namelen - 2,
			    (void *)(uintptr_t)oldp, oldlenp,
			    (void *)(uintptr_t)newp, newlen);
			goto done;
		}
	}
	error = ENOPROTOOPT;
done:
	domain_guard_release(guard);
	return (error);
}

void
pfctlinput(int cmd, struct sockaddr *sa)
{
	pfctlinput2(cmd, sa, NULL);
}

void
pfctlinput2(int cmd, struct sockaddr *sa, void *ctlparam)
{
	struct domain *dp;
	struct protosw *pp;
	domain_guard_t guard;

	if (sa == NULL)
		return;

	guard = domain_guard_deploy();
	TAILQ_FOREACH(dp, &domains, dom_entry) {
		TAILQ_FOREACH(pp, &dp->dom_protosw, pr_entry) {
			if (pp->pr_ctlinput != NULL)
				(*pp->pr_ctlinput)(cmd, sa, ctlparam);
		}
	}
	domain_guard_release(guard);
}

void
net_update_uptime(void)
{
	struct timeval tv;

	microuptime(&tv);
	_net_uptime = tv.tv_sec;
	/*
	 * Round up the timer to the nearest integer value because otherwise
	 * we might setup networking timers that are off by almost 1 second.
	 */
	if (tv.tv_usec > 500000)
		_net_uptime++;
}

void
net_update_uptime_secs(uint64_t secs)
{
	_net_uptime = secs;
}

/*
 * Convert our uin64_t net_uptime to a struct timeval.
 */
void
net_uptime2timeval(struct timeval *tv)
{
	if (tv == NULL)
		return;

	tv->tv_usec = 0;
	tv->tv_sec = net_uptime();
}

/*
 * An alternative way to obtain the coarse-grained uptime (in seconds)
 * for networking code which do not require high-precision timestamp,
 * as this is significantly cheaper than microuptime().
 */
u_int64_t
net_uptime(void)
{
	if (_net_uptime == 0)
		net_update_uptime();

	return (_net_uptime);
}

void
domain_proto_mtx_lock_assert_held(void)
{
	lck_mtx_assert(&domain_proto_mtx, LCK_MTX_ASSERT_OWNED);
}

void
domain_proto_mtx_lock_assert_notheld(void)
{
	lck_mtx_assert(&domain_proto_mtx, LCK_MTX_ASSERT_NOTOWNED);
}

domain_guard_t
domain_guard_deploy(void)
{
	net_thread_marks_t marks;

	marks = net_thread_marks_push(NET_THREAD_HELD_DOMAIN);
	if (marks != net_thread_marks_none) {
		lck_mtx_assert(&domain_proto_mtx, LCK_MTX_ASSERT_NOTOWNED);
		lck_mtx_lock(&domain_proto_mtx);
	}
	else
		lck_mtx_assert(&domain_proto_mtx, LCK_MTX_ASSERT_OWNED);

	return ((domain_guard_t)(const void*)marks);
}

void
domain_guard_release(domain_guard_t guard)
{
	net_thread_marks_t marks = (net_thread_marks_t)(const void*)guard;

	if (marks != net_thread_marks_none) {
		lck_mtx_assert(&domain_proto_mtx, LCK_MTX_ASSERT_OWNED);
		lck_mtx_unlock(&domain_proto_mtx);
		net_thread_marks_pop(marks);
	}
	else
		lck_mtx_assert(&domain_proto_mtx, LCK_MTX_ASSERT_NOTOWNED);
}

domain_unguard_t
domain_unguard_deploy(void)
{
	net_thread_marks_t marks;

	marks = net_thread_unmarks_push(NET_THREAD_HELD_DOMAIN);
	if (marks != net_thread_marks_none) {
		lck_mtx_assert(&domain_proto_mtx, LCK_MTX_ASSERT_OWNED);
		lck_mtx_unlock(&domain_proto_mtx);
	}
	else
		lck_mtx_assert(&domain_proto_mtx, LCK_MTX_ASSERT_NOTOWNED);

	return ((domain_unguard_t)(const void*)marks);
}

void
domain_unguard_release(domain_unguard_t unguard)
{
	net_thread_marks_t marks = (net_thread_marks_t)(const void*)unguard;

	if (marks != net_thread_marks_none) {
		lck_mtx_assert(&domain_proto_mtx, LCK_MTX_ASSERT_NOTOWNED);
		lck_mtx_lock(&domain_proto_mtx);
		net_thread_unmarks_pop(marks);
	}
	else
		lck_mtx_assert(&domain_proto_mtx, LCK_MTX_ASSERT_OWNED);
}