Please welcome casperd daemon. It (and its services) will be responsible for

giving access to functionality that is not available in capability mode
sandbox. The functionality can be precisely restricted.

Start with the following services:
- system.dns - provides API compatible to:
	- gethostbyname(3),
	- gethostbyname2(3),
	- gethostbyaddr(3),
	- getaddrinfo(3),
	- getnameinfo(3),
- system.grp - provides getgrent(3)-compatible API,
- system.pwd - provides getpwent(3)-compatible API,
- system.random - allows to obtain entropy from /dev/random,
- system.sysctl - provides sysctlbyname(3-compatible API.

Sponsored by:	The FreeBSD Foundation

5 files changed, 1112 insertions, 0 deletions

diff --git a/sbin/casperd/Makefile b/sbin/casperd/Makefile
new file mode 100644
index 000000000000..d436ce8f1150
--- /dev/null
+++ b/sbin/casperd/Makefile
@@ -0,0 +1,18 @@
+# $FreeBSD$
+
+PROG=	casperd
+
+SRCS=	casperd.c zygote.c
+
+DPADD=	${LIBCAPSICUM} ${LIBCASPER} ${LIBNV} ${LIBPJDLOG} ${LIBUTIL}
+LDADD=	-lcapsicum -lcasper -lnv -lpjdlog -lutil
+
+MAN=	casperd.8
+
+CFLAGS+=-I${.CURDIR}
+CFLAGS+=-I${.CURDIR}/../../lib/libcapsicum
+CFLAGS+=-I${.CURDIR}/../../lib/libcasper
+CFLAGS+=-I${.CURDIR}/../../lib/libnv
+CFLAGS+=-I${.CURDIR}/../../lib/libpjdlog
+
+.include <bsd.prog.mk>
diff --git a/sbin/casperd/casperd.8 b/sbin/casperd/casperd.8
new file mode 100644
index 000000000000..50f8ebed41d9
--- /dev/null
+++ b/sbin/casperd/casperd.8
@@ -0,0 +1,124 @@
+.\" Copyright (c) 2013 The FreeBSD Foundation
+.\" All rights reserved.
+.\"
+.\" This documentation was written by Pawel Jakub Dawidek under sponsorship
+.\" from the FreeBSD Foundation.
+.\"
+.\" 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 AUTHORS 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 AUTHORS 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.
+.\"
+.\" $FreeBSD$
+.\"
+.Dd October 26, 2013
+.Dt CASPERD 8
+.Os
+.Sh NAME
+.Nm casperd
+.Nd "Capability Services friendly daemon"
+.Sh SYNOPSIS
+.Nm
+[-Fhv] [-D servconfdir] [-P pidfile] [-S sockpath]
+.Op Fl Fhv
+.Op Fl D Ar servconfdir
+.Op Fl P Ar pidfile
+.Op Fl S Ar sockpath
+.Sh DESCRIPTION
+The
+.Nm
+daemon provides various services in capability-like fashion to programs
+running in capability mode sandbox.
+For example it is prohibited to send UDP packets to arbitrary destinations
+when operating in capability mode, which makes DNS resolution impossible.
+To make it possible the
+.Nm
+daemon provides
+.Nm system.dns
+service that allows to proxy DNS resolution requests through dedicated,
+non-sandboxed process provided by
+.Nm .
+.Pp
+The
+.Nm
+daemon can be started with the following command line arguments:
+.Bl -tag -width ".Fl D Ar servconfdir"
+.It Fl D Ar servconfdir
+Specify alternative location of the service configuration directory.
+The default location is
+.Pa /etc/casper/ .
+.It Fl F
+Start the
+.Nm
+daemon in the foreground.
+By default
+.Nm
+starts in the background.
+.It Fl h
+Print the
+.Nm
+usage message.
+.It Fl P Ar pidfile
+Specify alternative location of a file where main process PID will be
+stored.
+The default location is
+.Pa /var/run/casperd.pid .
+.It Fl S Ar sockpath
+Specify alternative location of a
+.Xr unix 4
+domain socket that can be used to connect to the
+.Nm
+daemon.
+The default location is
+.Pa /var/run/casper .
+.It Fl v
+Print or log verbose/debugging informations.
+This option can be specified multiple times to raise the verbosity
+level.
+.El
+.Sh FILES
+.Bl -tag -width ".Pa /var/run/casperd.pid" -compact
+.It Pa /etc/casper/
+The configuration directory for
+.Nm
+services.
+.It Pa /var/run/casper
+.Xr unix 4
+domain socket used to connect to the
+.Nm
+daemon.
+.It Pa /var/run/casperd.pid
+The default location of the
+.Nm
+PID file.
+.El
+.Sh EXIT STATUS
+Exit status is 0 on success, and > 0 of an error occurs.
+.Sh SEE ALSO
+.Xr cap_enter 2 ,
+.Xr libcapsicum 3 ,
+.Xr pidfile 3 ,
+.Xr capsicum 4 ,
+.Xr unix 4
+.Sh AUTHORS
+The
+.Nm
+was implemented by
+.An Pawel Jakub Dawidek Aq pawel@dawidek.net
+under sponsorship from the FreeBSD Foundation.
diff --git a/sbin/casperd/casperd.c b/sbin/casperd/casperd.c
new file mode 100644
index 000000000000..e918534619f6
--- /dev/null
+++ b/sbin/casperd/casperd.c
@@ -0,0 +1,711 @@
+/*-
+ * Copyright (c) 2012 The FreeBSD Foundation
+ * All rights reserved.
+ *
+ * This software was developed by Pawel Jakub Dawidek under sponsorship from
+ * the FreeBSD Foundation.
+ *
+ * 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 AUTHORS 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 AUTHORS 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.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <sys/types.h>
+#include <sys/capability.h>
+#include <sys/queue.h>
+#include <sys/socket.h>
+#include <sys/stat.h>
+#include <sys/un.h>
+
+#include <assert.h>
+#include <dirent.h>
+#include <err.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <libutil.h>
+#include <paths.h>
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <strings.h>
+#include <unistd.h>
+
+#include <libcapsicum.h>
+#include <libcapsicum_impl.h>
+#include <libcasper.h>
+#include <libcasper_impl.h>
+#include <msgio.h>
+#include <nv.h>
+#include <pjdlog.h>
+
+#include "msgio.h"
+
+#include "zygote.h"
+
+#define	CASPERD_PIDFILE		"/var/run/casperd.pid"
+#define	CASPERD_SERVCONFDIR	"/etc/casper"
+#define	CASPERD_SOCKPATH	"/var/run/casper"
+
+typedef void service_function_t(struct service_connection *, const nvlist_t *,
+    nvlist_t *);
+
+struct casper_service {
+	const char	*cs_name;
+	const char	*cs_execpath;
+	struct service	*cs_service;
+	nvlist_t	*cs_attrs;
+	TAILQ_ENTRY(casper_service) cs_next;
+};
+
+static TAILQ_HEAD(, casper_service) casper_services =
+    TAILQ_HEAD_INITIALIZER(casper_services);
+
+#define	SERVICE_IS_CORE(service)	((service)->cs_execpath == NULL)
+
+static void service_external_execute(int chanfd);
+
+#define KEEP_ERRNO(work)	do {					\
+	int _serrno;							\
+									\
+	_serrno = errno;						\
+	work;								\
+	errno = _serrno;						\
+} while (0)
+
+static struct casper_service *
+service_find(const char *name)
+{
+	struct casper_service *casserv;
+
+	TAILQ_FOREACH(casserv, &casper_services, cs_next) {
+		if (strcmp(casserv->cs_name, name) == 0)
+			break;
+	}
+	return (casserv);
+}
+
+/*
+ * Function always consumes the given attrs.
+ */
+static void
+service_register(nvlist_t *attrs)
+{
+	struct casper_service *casserv;
+	const char *name;
+
+	PJDLOG_ASSERT(nvlist_exists_string(attrs, "name"));
+	PJDLOG_ASSERT(nvlist_exists_string(attrs, "execpath") ||
+	    (nvlist_exists_number(attrs, "commandfunc") &&
+	     nvlist_exists_number(attrs, "limitfunc")));
+
+	name = nvlist_get_string(attrs, "name");
+	PJDLOG_ASSERT(name != NULL);
+	if (name[0] == '\0') {
+		pjdlog_error("Unable to register service with an empty name.");
+		nvlist_destroy(attrs);
+		return;
+	}
+	if (service_find(name) != NULL) {
+		pjdlog_error("Service \"%s\" is already registered.", name);
+		nvlist_destroy(attrs);
+		return;
+	}
+
+	casserv = malloc(sizeof(*casserv));
+	if (casserv == NULL) {
+		pjdlog_errno(LOG_ERR, "Unable to register service \"%s\"",
+		    name);
+		nvlist_destroy(attrs);
+		return;
+	}
+	casserv->cs_name = name;
+	if (nvlist_exists_string(attrs, "execpath")) {
+		struct stat sb;
+
+		PJDLOG_ASSERT(!nvlist_exists_number(attrs, "commandfunc"));
+		PJDLOG_ASSERT(!nvlist_exists_number(attrs, "limitfunc"));
+
+		casserv->cs_service = NULL;
+
+		casserv->cs_execpath = nvlist_get_string(attrs, "execpath");
+		if (casserv->cs_execpath == NULL ||
+		    casserv->cs_execpath[0] == '\0') {
+			pjdlog_error("Unable to register service with an empty execpath.");
+			free(casserv);
+			nvlist_destroy(attrs);
+			return;
+		}
+		if (stat(casserv->cs_execpath, &sb) == -1) {
+			pjdlog_errno(LOG_ERR,
+			    "Unable to register service \"%s\", problem with executable \"%s\"",
+			    name, casserv->cs_execpath);
+			free(casserv);
+			nvlist_destroy(attrs);
+			return;
+		}
+	} else /* if (nvlist_exists_number(attrs, "commandfunc")) */ {
+		PJDLOG_ASSERT(!nvlist_exists_string(attrs, "execpath"));
+
+		casserv->cs_execpath = NULL;
+
+		casserv->cs_service = service_alloc(name,
+		    (void *)(uintptr_t)nvlist_get_number(attrs, "limitfunc"),
+		    (void *)(uintptr_t)nvlist_get_number(attrs, "commandfunc"));
+		if (casserv->cs_service == NULL) {
+			pjdlog_errno(LOG_ERR,
+			    "Unable to register service \"%s\"", name);
+			free(casserv);
+			nvlist_destroy(attrs);
+			return;
+		}
+	}
+	casserv->cs_attrs = attrs;
+	TAILQ_INSERT_TAIL(&casper_services, casserv, cs_next);
+	pjdlog_debug(1, "Service %s successfully registered.",
+	    casserv->cs_name);
+}
+
+static bool
+casper_allowed_service(const nvlist_t *limits, const char *service)
+{
+
+	if (limits == NULL)
+		return (true);
+
+	if (nvlist_exists_null(limits, service))
+		return (true);
+
+	return (false);
+}
+
+static int
+casper_limit(const nvlist_t *oldlimits, const nvlist_t *newlimits)
+{
+	const char *name;
+	int type;
+	void *cookie;
+
+	while ((name = nvlist_next(newlimits, &type, &cookie)) != NULL) {
+		if (type != NV_TYPE_NULL)
+			return (EINVAL);
+		if (!casper_allowed_service(oldlimits, name))
+			return (ENOTCAPABLE);
+	}
+
+	return (0);
+}
+
+static int
+casper_command(const char *cmd, const nvlist_t *limits, nvlist_t *nvlin,
+    nvlist_t *nvlout)
+{
+	struct casper_service *casserv;
+	const char *servname;
+	nvlist_t *nvl;
+	int chanfd, execfd, procfd, error;
+
+	if (strcmp(cmd, "open") != 0)
+		return (EINVAL);
+	if (!nvlist_exists_string(nvlin, "service"))
+		return (EINVAL);
+
+	servname = nvlist_get_string(nvlin, "service");
+
+	casserv = service_find(servname);
+	if (casserv == NULL)
+		return (ENOENT);
+
+	if (!casper_allowed_service(limits, servname))
+		return (ENOTCAPABLE);
+
+#ifdef O_EXEC_WORKING
+	execfd = open(casserv->cs_execpath, O_EXEC);
+#else
+	execfd = open(casserv->cs_execpath, O_RDONLY);
+#endif
+	if (execfd < -1) {
+		error = errno;
+		pjdlog_errno(LOG_ERR,
+		    "Unable to open executable '%s' of service '%s'",
+		    casserv->cs_execpath, casserv->cs_name);
+		return (error);
+	}
+
+	if (zygote_clone(service_external_execute, 0, &chanfd, &procfd) == -1) {
+		error = errno;
+		close(execfd);
+		return (error);
+	}
+
+	nvl = nvlist_create(0);
+	nvlist_add_string(nvl, "service", casserv->cs_name);
+	if (nvlist_exists_descriptor(nvlin, "stderrfd")) {
+		nvlist_move_descriptor(nvl, "stderrfd",
+		    nvlist_take_descriptor(nvlin, "stderrfd"));
+	}
+	nvlist_move_descriptor(nvl, "execfd", execfd);
+	nvlist_move_descriptor(nvl, "procfd", procfd);
+	if (nvlist_send(chanfd, nvl) == -1) {
+		error = errno;
+		pjdlog_errno(LOG_ERR, "Unable to send nvlist");
+		nvlist_destroy(nvl);
+		close(chanfd);
+		return (error);
+	}
+	nvlist_destroy(nvl);
+
+	nvlist_move_descriptor(nvlout, "chanfd", chanfd);
+
+	return (0);
+}
+
+static void
+fdswap(int *fd0, int *fd1)
+{
+	int tmpfd;
+
+	PJDLOG_VERIFY((tmpfd = dup(*fd0)) != -1);
+	PJDLOG_VERIFY(dup2(*fd1, *fd0) != -1);
+	PJDLOG_VERIFY(dup2(tmpfd, *fd1) != -1);
+	close(tmpfd);
+	tmpfd = *fd0;
+	*fd0 = *fd1;
+	*fd1 = tmpfd;
+}
+
+static void
+fdmove(int *oldfdp, int newfd)
+{
+
+	if (*oldfdp != newfd) {
+		PJDLOG_VERIFY(dup2(*oldfdp, newfd) != -1);
+		close(*oldfdp);
+		*oldfdp = newfd;
+	}
+}
+
+static void
+fdcloexec(int fd)
+{
+	int flags;
+
+	flags = fcntl(fd, F_GETFD);
+	PJDLOG_ASSERT(flags != -1);
+	if ((flags & FD_CLOEXEC) != 0)
+		PJDLOG_VERIFY(fcntl(fd, F_SETFD, flags & ~FD_CLOEXEC) != -1);
+}
+
+static void
+service_register_core(void)
+{
+	nvlist_t *nvl;
+
+	nvl = nvlist_create(0);
+	nvlist_add_string(nvl, "name", "core.casper");
+	nvlist_add_number(nvl, "limitfunc", (uint64_t)(uintptr_t)casper_limit);
+	nvlist_add_number(nvl, "commandfunc",
+	    (uint64_t)(uintptr_t)casper_command);
+	service_register(nvl);
+}
+
+static int
+setup_creds(int sock)
+{
+	struct cmsgcred cred;
+
+	if (cred_recv(sock, &cred) == -1)
+		return (-1);
+
+	if (setgroups((int)cred.cmcred_ngroups, cred.cmcred_groups) == -1)
+		return (-1);
+
+	if (setgid(cred.cmcred_groups[0]) == -1)
+		return (-1);
+
+	if (setuid(cred.cmcred_euid) == -1)
+		return (-1);
+
+	return (0);
+}
+
+static void
+service_external_execute(int chanfd)
+{
+	char *service, *argv[3];
+	int stderrfd, execfd, procfd;
+	nvlist_t *nvl;
+
+	nvl = nvlist_recv(chanfd);
+	if (nvl == NULL)
+		pjdlog_exit(1, "Unable to receive nvlist");
+	service = nvlist_take_string(nvl, "service");
+	PJDLOG_ASSERT(service != NULL);
+	if (nvlist_exists_descriptor(nvl, "stderrfd")) {
+		stderrfd = nvlist_take_descriptor(nvl, "stderrfd");
+	} else {
+		stderrfd = open(_PATH_DEVNULL, O_RDWR);
+		if (stderrfd < 0)
+			pjdlog_exit(1, "Unable to open %s", _PATH_DEVNULL);
+	}
+	execfd = nvlist_take_descriptor(nvl, "execfd");
+	procfd = nvlist_take_descriptor(nvl, "procfd");
+	nvlist_destroy(nvl);
+
+	/*
+	 * Move all descriptors into right slots.
+	 */
+
+	if (stderrfd != STDERR_FILENO) {
+		if (chanfd == STDERR_FILENO)
+			fdswap(&stderrfd, &chanfd);
+		else if (execfd == STDERR_FILENO)
+			fdswap(&stderrfd, &execfd);
+		else if (procfd == STDERR_FILENO)
+			fdswap(&stderrfd, &procfd);
+		fdmove(&stderrfd, STDERR_FILENO);
+	}
+	fdcloexec(stderrfd);
+
+	if (chanfd != PARENT_FILENO) {
+		if (execfd == PARENT_FILENO)
+			fdswap(&chanfd, &execfd);
+		else if (procfd == PARENT_FILENO)
+			fdswap(&chanfd, &procfd);
+		fdmove(&chanfd, PARENT_FILENO);
+	}
+	fdcloexec(chanfd);
+
+	if (execfd != EXECUTABLE_FILENO) {
+		if (procfd == EXECUTABLE_FILENO)
+			fdswap(&execfd, &procfd);
+		fdmove(&execfd, EXECUTABLE_FILENO);
+	}
+	fdcloexec(execfd);
+
+	if (procfd != PROC_FILENO)
+		fdmove(&procfd, PROC_FILENO);
+	fdcloexec(procfd);
+
+	/*
+	 * Use credentials of the caller process.
+	 */
+	setup_creds(chanfd);
+
+	argv[0] = service;
+	asprintf(&argv[1], "%d", pjdlog_debug_get());
+	argv[2] = NULL;
+
+	fexecve(execfd, argv, NULL);
+	pjdlog_exit(1, "Unable to execute service %s", service);
+}
+
+static void
+service_register_external_one(const char *dirpath, int dfd,
+    const char *filename)
+{
+	char execpath[FILENAME_MAX];
+	nvlist_t *nvl;
+	ssize_t done;
+	int fd;
+
+	fd = openat(dfd, filename, O_RDONLY);
+	if (fd == -1) {
+		pjdlog_errno(LOG_ERR, "Unable to open \"%s/%s\"", dirpath,
+		    filename);
+		return;
+	}
+
+	done = read(fd, execpath, sizeof(execpath));
+	if (done == -1) {
+		pjdlog_errno(LOG_ERR, "Unable to read content of \"%s/%s\"",
+		    dirpath, filename);
+		close(fd);
+		return;
+	}
+	close(fd);
+	if (done == sizeof(execpath)) {
+		pjdlog_error("Executable path too long in \"%s/%s\".", dirpath,
+		    filename);
+		return;
+	}
+	execpath[done] = '\0';
+	while (done > 0) {
+		if (execpath[--done] == '\n')
+			execpath[done] = '\0';
+	}
+
+	nvl = nvlist_create(0);
+	nvlist_add_string(nvl, "name", filename);
+	nvlist_add_string(nvl, "execpath", execpath);
+	if (nvlist_error(nvl) != 0) {
+		pjdlog_common(LOG_ERR, 0, nvlist_error(nvl),
+		    "Unable to allocate attributes for service \"%s/%s\"",
+		    dirpath, filename);
+		nvlist_destroy(nvl);
+		return;
+	}
+
+	service_register(nvl);
+	/* service_register() consumed nvl. */
+}
+
+static uint8_t
+file_type(int dfd, const char *filename)
+{
+	struct stat sb;
+
+	if (fstatat(dfd, filename, &sb, AT_SYMLINK_NOFOLLOW) == -1) {
+		pjdlog_errno(LOG_ERR, "Unable to stat \"%s\"", filename);
+		return (DT_UNKNOWN);
+	}
+	return (IFTODT(sb.st_mode));
+}
+
+static void
+service_register_external(const char *dirpath)
+{
+	DIR *dirp;
+	struct dirent *dp;
+	int dfd;
+
+	dirp = opendir(dirpath);
+	if (dirp == NULL) {
+		pjdlog_errno(LOG_WARNING, "Unable to open \"%s\"", dirpath);
+		return;
+	}
+	dfd = dirfd(dirp);
+	PJDLOG_ASSERT(dfd >= 0);
+	while ((dp = readdir(dirp)) != NULL) {
+		dp->d_type = file_type(dfd, dp->d_name);
+		/* We are only interested in regular files, skip the rest. */
+		if (dp->d_type != DT_REG) {
+			pjdlog_debug(1,
+			    "File \"%s/%s\" is not a regular file, skipping.",
+			    dirpath, dp->d_name);
+			continue;
+		}
+		service_register_external_one(dirpath, dfd, dp->d_name);
+	}
+	closedir(dirp);
+}
+
+static void
+casper_accept(int lsock)
+{
+	struct casper_service *casserv;
+	struct service_connection *sconn;
+	int sock;
+
+	sock = accept(lsock, NULL, NULL);
+	if (sock == -1) {
+		pjdlog_errno(LOG_ERR, "Unable to accept casper connection");
+		return;
+	}
+	casserv = service_find("core.casper");
+	PJDLOG_ASSERT(casserv != NULL);
+
+	sconn = service_connection_add(casserv->cs_service, sock, NULL);
+	if (sconn == NULL) {
+		close(sock);
+		return;
+	}
+}
+
+static void
+main_loop(const char *sockpath, struct pidfh *pfh)
+{
+	fd_set fds;
+	struct sockaddr_un sun;
+	struct casper_service *casserv;
+	struct service_connection *sconn, *sconntmp;
+	int lsock, sock, maxfd, ret;
+
+	lsock = socket(AF_UNIX, SOCK_STREAM, 0);
+	if (lsock == -1)
+		pjdlog_exit(1, "Unable to create socket");
+
+	(void)unlink(sockpath);
+
+	bzero(&sun, sizeof(sun));
+	sun.sun_family = AF_UNIX;
+	PJDLOG_VERIFY(strlcpy(sun.sun_path, sockpath, sizeof(sun.sun_path)) <
+	    sizeof(sun.sun_path));
+	sun.sun_len = SUN_LEN(&sun);
+
+	if (bind(lsock, (struct sockaddr *)&sun, sizeof(sun)) == -1)
+		pjdlog_exit(1, "Unable to bind to %s", sockpath);
+	if (listen(lsock, 8) == -1)
+		pjdlog_exit(1, "Unable to listen on %s", sockpath);
+
+	for (;;) {
+		FD_ZERO(&fds);
+		FD_SET(lsock, &fds);
+		maxfd = lsock;
+		TAILQ_FOREACH(casserv, &casper_services, cs_next) {
+			/* We handle only core services. */
+			if (!SERVICE_IS_CORE(casserv))
+				continue;
+			for (sconn = service_connection_first(casserv->cs_service);
+			    sconn != NULL;
+			    sconn = service_connection_next(sconn)) {
+				sock = service_connection_get_sock(sconn);
+				FD_SET(sock, &fds);
+				maxfd = sock > maxfd ? sock : maxfd;
+			}
+		}
+		maxfd++;
+
+		PJDLOG_ASSERT(maxfd <= (int)FD_SETSIZE);
+		ret = select(maxfd, &fds, NULL, NULL, NULL);
+		PJDLOG_ASSERT(ret == -1 || ret > 0);	/* select() cannot timeout */
+		if (ret == -1) {
+			if (errno == EINTR)
+				continue;
+			KEEP_ERRNO((void)pidfile_remove(pfh));
+			pjdlog_exit(1, "select() failed");
+		}
+
+		if (FD_ISSET(lsock, &fds))
+			casper_accept(lsock);
+		TAILQ_FOREACH(casserv, &casper_services, cs_next) {
+			/* We handle only core services. */
+			if (!SERVICE_IS_CORE(casserv))
+				continue;
+			for (sconn = service_connection_first(casserv->cs_service);
+			    sconn != NULL; sconn = sconntmp) {
+				/*
+				 * Prepare for connection to be removed from
+				 * the list on failure.
+				 */
+				sconntmp = service_connection_next(sconn);
+				sock = service_connection_get_sock(sconn);
+				if (FD_ISSET(sock, &fds)) {
+					service_message(casserv->cs_service,
+					    sconn);
+				}
+			}
+		}
+	}
+}
+
+static void
+usage(void)
+{
+
+	pjdlog_exitx(1,
+	    "usage: casperd [-Fhv] [-D servconfdir] [-P pidfile] [-S sockpath]");
+}
+
+int
+main(int argc, char *argv[])
+{
+	struct pidfh *pfh;
+	const char *pidfile, *servconfdir, *sockpath;
+	pid_t otherpid;
+	int ch, debug;
+	bool foreground;
+
+	pjdlog_init(PJDLOG_MODE_STD);
+
+	debug = 0;
+	foreground = false;
+	pidfile = CASPERD_PIDFILE;
+	servconfdir = CASPERD_SERVCONFDIR;
+	sockpath = CASPERD_SOCKPATH;
+
+	while ((ch = getopt(argc, argv, "D:FhP:S:v")) != -1) {
+		switch (ch) {
+		case 'D':
+			servconfdir = optarg;
+			break;
+		case 'F':
+			foreground = true;
+			break;
+		case 'P':
+			pidfile = optarg;
+			break;
+		case 'S':
+			sockpath = optarg;
+			break;
+		case 'v':
+			debug++;
+			break;
+		case 'h':
+		default:
+			usage();
+		}
+	}
+	argc -= optind;
+	argv += optind;
+
+	if (argc != 0)
+		usage();
+
+	if (!foreground)
+		pjdlog_mode_set(PJDLOG_MODE_SYSLOG);
+	pjdlog_prefix_set("(casperd) ");
+	pjdlog_debug_set(debug);
+
+	if (zygote_init() < 0)
+		pjdlog_exit(1, "Unable to create zygote process");
+
+	pfh = pidfile_open(pidfile, 0600, &otherpid);
+	if (pfh == NULL) {
+		if (errno == EEXIST) {
+			pjdlog_exitx(1, "casperd already running, pid: %jd.",
+			    (intmax_t)otherpid);
+		}
+		pjdlog_errno(LOG_WARNING, "Cannot open or create pidfile %s",
+		    pidfile);
+	}
+
+	if (!foreground) {
+		if (daemon(0, 0) == -1) {
+			KEEP_ERRNO((void)pidfile_remove(pfh));
+			pjdlog_exit(1, "Unable to go into background");
+		}
+	}
+
+	/* Write PID to a file. */
+	if (pidfile_write(pfh) == -1) {
+		pjdlog_errno(LOG_WARNING, "Unable to write to pidfile %s",
+		    pidfile);
+	} else {
+		pjdlog_debug(1, "PID stored in %s.", pidfile);
+	}
+
+	/*
+	 * Register core services.
+	 */
+	service_register_core();
+	/*
+	 * Register external services.
+	 */
+	service_register_external(servconfdir);
+
+	/*
+	 * Wait for connections.
+	 */
+	main_loop(sockpath, pfh);
+}
diff --git a/sbin/casperd/zygote.c b/sbin/casperd/zygote.c
new file mode 100644
index 000000000000..70594ced1c68
--- /dev/null
+++ b/sbin/casperd/zygote.c
@@ -0,0 +1,219 @@
+/*-
+ * Copyright (c) 2012 The FreeBSD Foundation
+ * All rights reserved.
+ *
+ * This software was developed by Pawel Jakub Dawidek under sponsorship from
+ * the FreeBSD Foundation.
+ *
+ * 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 AUTHORS 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 AUTHORS 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.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <sys/types.h>
+#include <sys/capability.h>
+#include <sys/procdesc.h>
+#include <sys/socket.h>
+
+#include <assert.h>
+#include <err.h>
+#include <errno.h>
+#include <paths.h>
+#include <stdbool.h>
+#include <stdlib.h>
+#include <strings.h>
+#include <unistd.h>
+
+#include <libcapsicum.h>
+#include <libcapsicum_impl.h>
+#include <nv.h>
+#include <pjdlog.h>
+
+#include "zygote.h"
+
+/* Zygote info. */
+static int	zygote_sock = -1;
+
+static void
+stdnull(void)
+{
+	int fd;
+
+	fd = open(_PATH_DEVNULL, O_RDWR);
+	if (fd == -1)
+		errx(1, "Unable to open %s", _PATH_DEVNULL);
+
+	if (dup2(fd, STDIN_FILENO) == -1)
+		errx(1, "Unable to cover stdin");
+	if (dup2(fd, STDOUT_FILENO) == -1)
+		errx(1, "Unable to cover stdout");
+	if (dup2(fd, STDERR_FILENO) == -1)
+		errx(1, "Unable to cover stderr");
+
+	close(fd);
+}
+
+int
+zygote_clone(zygote_func_t *func, int flags, int *chanfdp, int *procfdp)
+{
+	nvlist_t *nvl;
+	int error;
+
+	if (zygote_sock == -1) {
+		/* Zygote didn't start. */
+		errno = ENXIO;
+		return (-1);
+	}
+
+	nvl = nvlist_create(0);
+	nvlist_add_number(nvl, "func", (uint64_t)(uintptr_t)func);
+	nvlist_add_number(nvl, "flags", (uint64_t)flags);
+	nvl = nvlist_xfer(zygote_sock, nvl);
+	if (nvl == NULL)
+		return (-1);
+	if (nvlist_exists_number(nvl, "error")) {
+		error = (int)nvlist_get_number(nvl, "error");
+		nvlist_destroy(nvl);
+		errno = error;
+		return (-1);
+	}
+
+	*chanfdp = nvlist_take_descriptor(nvl, "chanfd");
+	*procfdp = nvlist_take_descriptor(nvl, "procfd");
+
+	nvlist_destroy(nvl);
+	return (0);
+}
+
+/*
+ * This function creates sandboxes on-demand whoever has access to it via
+ * 'sock' socket. Function sends two descriptors to the caller: process
+ * descriptor of the sandbox and socket pair descriptor for communication
+ * between sandbox and its owner.
+ */
+static void
+zygote_main(int sock)
+{
+	int error, fd, flags, procfd;
+	int chanfd[2];
+	nvlist_t *nvlin, *nvlout;
+	zygote_func_t *func;
+	pid_t pid;
+
+	assert(sock > STDERR_FILENO);
+
+	if (pjdlog_mode_get() != PJDLOG_MODE_STD)
+		stdnull();
+	for (fd = STDERR_FILENO + 1; fd < sock; fd++)
+		close(fd);
+	closefrom(sock + 1);
+
+	for (;;) {
+		nvlin = nvlist_recv(sock);
+		if (nvlin == NULL)
+			continue;
+		func = (zygote_func_t *)(uintptr_t)nvlist_get_number(nvlin,
+		    "func");
+		flags = (int)nvlist_get_number(nvlin, "flags");
+		nvlist_destroy(nvlin);
+
+		/*
+		 * Someone is requesting a new process, create one.
+		 */
+		procfd = -1;
+		chanfd[0] = -1;
+		chanfd[1] = -1;
+		error = 0;
+		if (socketpair(PF_UNIX, SOCK_STREAM | SOCK_NONBLOCK, 0,
+		    chanfd) == -1) {
+			error = errno;
+			goto send;
+		}
+		pid = pdfork(&procfd, 0);
+		switch (pid) {
+		case -1:
+			/* Failure. */
+			error = errno;
+			break;
+		case 0:
+			/* Child. */
+			close(sock);
+			close(chanfd[0]);
+			func(chanfd[1]);
+			/* NOTREACHED */
+			exit(1);
+		default:
+			/* Parent. */
+			close(chanfd[1]);
+			break;
+		}
+send:
+		nvlout = nvlist_create(0);
+		if (error != 0) {
+			nvlist_add_number(nvlout, "error", (uint64_t)error);
+			if (chanfd[0] >= 0)
+				close(chanfd[0]);
+			if (procfd >= 0)
+				close(procfd);
+		} else {
+			nvlist_move_descriptor(nvlout, "chanfd", chanfd[0]);
+			nvlist_move_descriptor(nvlout, "procfd", procfd);
+		}
+		(void)nvlist_send(sock, nvlout);
+		nvlist_destroy(nvlout);
+	}
+	/* NOTREACHED */
+}
+
+int
+zygote_init(void)
+{
+	int serrno, sp[2];
+	pid_t pid;
+
+	if (socketpair(PF_UNIX, SOCK_STREAM | SOCK_NONBLOCK, 0, sp) == -1)
+		return (-1);
+
+	pid = fork();
+	switch (pid) {
+	case -1:
+		/* Failure. */
+		serrno = errno;
+		close(sp[0]);
+		close(sp[1]);
+		errno = serrno;
+		return (-1);
+	case 0:
+		/* Child. */
+		close(sp[0]);
+		zygote_main(sp[1]);
+		/* NOTREACHED */
+		abort();
+	default:
+		/* Parent. */
+		zygote_sock = sp[0];
+		close(sp[1]);
+		return (0);
+	}
+	/* NOTREACHED */
+}
diff --git a/sbin/casperd/zygote.h b/sbin/casperd/zygote.h
new file mode 100644
index 000000000000..75ef2efda486
--- /dev/null
+++ b/sbin/casperd/zygote.h
@@ -0,0 +1,40 @@
+/*-
+ * Copyright (c) 2012 The FreeBSD Foundation
+ * All rights reserved.
+ *
+ * This software was developed by Pawel Jakub Dawidek under sponsorship from
+ * the FreeBSD Foundation.
+ *
+ * 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 AUTHORS 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 AUTHORS 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.
+ *
+ * $FreeBSD$
+ */
+
+#ifndef _ZYGOTE_H_
+#define	_ZYGOTE_H_
+
+typedef void zygote_func_t(int);
+
+int zygote_init(void);
+int zygote_clone(zygote_func_t *func, int flags, int *chanfdp, int *procfdp);
+
+#endif	/* !_ZYGOTE_H_ */