summaryrefslogtreecommitdiff
path: root/src/systemd-nspawn
diff options
context:
space:
mode:
authorLuke Shumaker <lukeshu@lukeshu.com>2017-05-10 15:54:51 -0400
committerLuke Shumaker <lukeshu@lukeshu.com>2017-05-10 15:54:51 -0400
commit51bf8f984974c8d60e6576b623b4ea7e0e20a470 (patch)
tree3ff14638977c50bffd735a6628842a675fc1710b /src/systemd-nspawn
parent88cdc57b277ec240d4c5ddfbbfaddfed2b2408c1 (diff)
./tools/notsd-move
Diffstat (limited to 'src/systemd-nspawn')
-rw-r--r--src/systemd-nspawn/.gitignore1
l---------src/systemd-nspawn/GNUmakefile1
-rw-r--r--src/systemd-nspawn/Makefile93
-rw-r--r--src/systemd-nspawn/nspawn-cgroup.c185
-rw-r--r--src/systemd-nspawn/nspawn-cgroup.h29
-rw-r--r--src/systemd-nspawn/nspawn-expose-ports.c245
-rw-r--r--src/systemd-nspawn/nspawn-expose-ports.h44
-rw-r--r--src/systemd-nspawn/nspawn-gperf.gperf47
-rw-r--r--src/systemd-nspawn/nspawn-mount.c1164
-rw-r--r--src/systemd-nspawn/nspawn-mount.h71
-rw-r--r--src/systemd-nspawn/nspawn-network.c696
-rw-r--r--src/systemd-nspawn/nspawn-network.h39
-rw-r--r--src/systemd-nspawn/nspawn-patch-uid.c483
-rw-r--r--src/systemd-nspawn/nspawn-patch-uid.h23
-rw-r--r--src/systemd-nspawn/nspawn-register.c227
-rw-r--r--src/systemd-nspawn/nspawn-register.h29
-rw-r--r--src/systemd-nspawn/nspawn-seccomp.c186
-rw-r--r--src/systemd-nspawn/nspawn-seccomp.h24
-rw-r--r--src/systemd-nspawn/nspawn-settings.c515
-rw-r--r--src/systemd-nspawn/nspawn-settings.h119
-rw-r--r--src/systemd-nspawn/nspawn-setuid.c271
-rw-r--r--src/systemd-nspawn/nspawn-setuid.h22
-rw-r--r--src/systemd-nspawn/nspawn-stub-pid1.c171
-rw-r--r--src/systemd-nspawn/nspawn-stub-pid1.h22
-rw-r--r--src/systemd-nspawn/nspawn.c4319
-rw-r--r--src/systemd-nspawn/systemd-nspawn.completion.bash155
-rw-r--r--src/systemd-nspawn/systemd-nspawn.completion.zsh50
-rw-r--r--src/systemd-nspawn/systemd-nspawn.tmpfiles23
-rw-r--r--src/systemd-nspawn/systemd-nspawn.xml1088
-rw-r--r--src/systemd-nspawn/systemd-nspawn@.service.in39
-rw-r--r--src/systemd-nspawn/test-patch-uid.c62
31 files changed, 10443 insertions, 0 deletions
diff --git a/src/systemd-nspawn/.gitignore b/src/systemd-nspawn/.gitignore
new file mode 100644
index 0000000000..85c81fff24
--- /dev/null
+++ b/src/systemd-nspawn/.gitignore
@@ -0,0 +1 @@
+/nspawn-gperf.c
diff --git a/src/systemd-nspawn/GNUmakefile b/src/systemd-nspawn/GNUmakefile
new file mode 120000
index 0000000000..54fdd42278
--- /dev/null
+++ b/src/systemd-nspawn/GNUmakefile
@@ -0,0 +1 @@
+../../GNUmakefile \ No newline at end of file
diff --git a/src/systemd-nspawn/Makefile b/src/systemd-nspawn/Makefile
new file mode 100644
index 0000000000..c0fc3bd98e
--- /dev/null
+++ b/src/systemd-nspawn/Makefile
@@ -0,0 +1,93 @@
+# -*- Mode: makefile; indent-tabs-mode: t -*-
+#
+# This file is part of systemd.
+#
+# Copyright 2010-2012 Lennart Poettering
+# Copyright 2010-2012 Kay Sievers
+# Copyright 2013 Zbigniew Jędrzejewski-Szmek
+# Copyright 2013 David Strauss
+# Copyright 2016 Luke Shumaker
+#
+# systemd is free software; you can redistribute it and/or modify it
+# under the terms of the GNU Lesser General Public License as published by
+# the Free Software Foundation; either version 2.1 of the License, or
+# (at your option) any later version.
+#
+# systemd is distributed in the hope that it will be useful, but
+# WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+# Lesser General Public License for more details.
+#
+# You should have received a copy of the GNU Lesser General Public License
+# along with systemd; If not, see <http://www.gnu.org/licenses/>.
+include $(dir $(lastword $(MAKEFILE_LIST)))/../../config.mk
+include $(topsrcdir)/build-aux/Makefile.head.mk
+
+bin_PROGRAMS += systemd-nspawn
+systemd_nspawn_SOURCES = \
+ src/nspawn/nspawn.c \
+ src/nspawn/nspawn-settings.c \
+ src/nspawn/nspawn-settings.h \
+ src/nspawn/nspawn-mount.c \
+ src/nspawn/nspawn-mount.h \
+ src/nspawn/nspawn-network.c \
+ src/nspawn/nspawn-network.h \
+ src/nspawn/nspawn-expose-ports.c \
+ src/nspawn/nspawn-expose-ports.h \
+ src/nspawn/nspawn-cgroup.c \
+ src/nspawn/nspawn-cgroup.h \
+ src/nspawn/nspawn-seccomp.c \
+ src/nspawn/nspawn-seccomp.h \
+ src/nspawn/nspawn-register.c \
+ src/nspawn/nspawn-register.h \
+ src/nspawn/nspawn-setuid.c \
+ src/nspawn/nspawn-setuid.h \
+ src/nspawn/nspawn-stub-pid1.c \
+ src/nspawn/nspawn-stub-pid1.h \
+ src/nspawn/nspawn-patch-uid.c \
+ src/nspawn/nspawn-patch-uid.h \
+ src/core/mount-setup.c \
+ src/core/mount-setup.h \
+ src/core/loopback-setup.c \
+ src/core/loopback-setup.h
+
+nodist_systemd_nspawn_SOURCES = \
+ src/nspawn/nspawn-gperf.c
+
+gperf_gperf_sources += \
+ src/nspawn/nspawn-gperf.gperf
+
+systemd_nspawn_CFLAGS = \
+ $(ACL_CFLAGS) \
+ $(BLKID_CFLAGS) \
+ $(SECCOMP_CFLAGS) \
+ $(SELINUX_CFLAGS)
+
+systemd_nspawn_LDADD = \
+ libsystemd-shared.la \
+ $(ACL_LIBS) \
+ $(BLKID_LIBS) \
+ $(SECCOMP_LIBS) \
+ $(SELINUX_LIBS)
+
+ifneq ($(HAVE_LIBIPTC),)
+systemd_nspawn_LDADD += \
+ libsystemd-firewall.la
+endif # HAVE_LIBIPTC
+
+test_patch_uid_SOURCES = \
+ src/nspawn/nspawn-patch-uid.c \
+ src/nspawn/nspawn-patch-uid.h \
+ src/nspawn/test-patch-uid.c
+
+test_patch_uid_CFLAGS = \
+ $(ACL_CFLAGS)
+
+test_patch_uid_LDADD = \
+ libsystemd-shared.la \
+ $(ACL_LIBS)
+
+manual_tests += \
+ test-patch-uid
+
+include $(topsrcdir)/build-aux/Makefile.tail.mk
diff --git a/src/systemd-nspawn/nspawn-cgroup.c b/src/systemd-nspawn/nspawn-cgroup.c
new file mode 100644
index 0000000000..9e793d85f1
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-cgroup.c
@@ -0,0 +1,185 @@
+/***
+ This file is part of systemd.
+
+ Copyright 2015 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include <sys/mount.h>
+
+#include "systemd-basic/alloc-util.h"
+#include "systemd-basic/fd-util.h"
+#include "systemd-basic/fileio.h"
+#include "systemd-basic/mkdir.h"
+#include "systemd-basic/mount-util.h"
+#include "systemd-basic/rm-rf.h"
+#include "systemd-basic/string-util.h"
+#include "systemd-basic/strv.h"
+#include "systemd-basic/util.h"
+
+#include "nspawn-cgroup.h"
+
+static int chown_cgroup_path(const char *path, uid_t uid_shift) {
+ _cleanup_close_ int fd = -1;
+ const char *fn;
+
+ fd = open(path, O_RDONLY|O_CLOEXEC|O_DIRECTORY);
+ if (fd < 0)
+ return -errno;
+
+ FOREACH_STRING(fn,
+ ".",
+ "tasks",
+ "notify_on_release",
+ "cgroup.procs",
+ "cgroup.events",
+ "cgroup.clone_children",
+ "cgroup.controllers",
+ "cgroup.subtree_control")
+ if (fchownat(fd, fn, uid_shift, uid_shift, 0) < 0)
+ log_full_errno(errno == ENOENT ? LOG_DEBUG : LOG_WARNING, errno,
+ "Failed to chown() cgroup file %s, ignoring: %m", fn);
+
+ return 0;
+}
+
+int chown_cgroup(pid_t pid, uid_t uid_shift) {
+ _cleanup_free_ char *path = NULL, *fs = NULL;
+ int r;
+
+ r = cg_pid_get_path(NULL, pid, &path);
+ if (r < 0)
+ return log_error_errno(r, "Failed to get container cgroup path: %m");
+
+ r = cg_get_path(SYSTEMD_CGROUP_CONTROLLER, path, NULL, &fs);
+ if (r < 0)
+ return log_error_errno(r, "Failed to get file system path for container cgroup: %m");
+
+ r = chown_cgroup_path(fs, uid_shift);
+ if (r < 0)
+ return log_error_errno(r, "Failed to chown() cgroup %s: %m", fs);
+
+ return 0;
+}
+
+int sync_cgroup(pid_t pid, CGroupUnified unified_requested, uid_t arg_uid_shift) {
+ _cleanup_free_ char *cgroup = NULL;
+ char tree[] = "/tmp/unifiedXXXXXX", pid_string[DECIMAL_STR_MAX(pid) + 1];
+ bool undo_mount = false;
+ const char *fn;
+ int unified, r;
+
+ unified = cg_unified(SYSTEMD_CGROUP_CONTROLLER);
+ if (unified < 0)
+ return log_error_errno(unified, "Failed to determine whether the unified hierarchy is used: %m");
+
+ if ((unified > 0) == (unified_requested >= CGROUP_UNIFIED_SYSTEMD))
+ return 0;
+
+ /* When the host uses the legacy cgroup setup, but the
+ * container shall use the unified hierarchy, let's make sure
+ * we copy the path from the name=systemd hierarchy into the
+ * unified hierarchy. Similar for the reverse situation. */
+
+ r = cg_pid_get_path(SYSTEMD_CGROUP_CONTROLLER, pid, &cgroup);
+ if (r < 0)
+ return log_error_errno(r, "Failed to get control group of " PID_FMT ": %m", pid);
+
+ /* In order to access the unified hierarchy we need to mount it */
+ if (!mkdtemp(tree))
+ return log_error_errno(errno, "Failed to generate temporary mount point for unified hierarchy: %m");
+
+ if (unified)
+ r = mount_verbose(LOG_ERR, "cgroup", tree, "cgroup",
+ MS_NOSUID|MS_NOEXEC|MS_NODEV, "none,name=systemd,xattr");
+ else
+ r = mount_verbose(LOG_ERR, "cgroup", tree, "cgroup2",
+ MS_NOSUID|MS_NOEXEC|MS_NODEV, NULL);
+ if (r < 0)
+ goto finish;
+
+ undo_mount = true;
+
+ /* If nspawn dies abruptly the cgroup hierarchy created below
+ * its unit isn't cleaned up. So, let's remove it
+ * https://github.com/systemd/systemd/pull/4223#issuecomment-252519810 */
+ fn = strjoina(tree, cgroup);
+ (void) rm_rf(fn, REMOVE_ROOT|REMOVE_ONLY_DIRECTORIES);
+
+ fn = strjoina(tree, cgroup, "/cgroup.procs");
+ (void) mkdir_parents(fn, 0755);
+
+ sprintf(pid_string, PID_FMT, pid);
+ r = write_string_file(fn, pid_string, 0);
+ if (r < 0) {
+ log_error_errno(r, "Failed to move process: %m");
+ goto finish;
+ }
+
+ fn = strjoina(tree, cgroup);
+ r = chown_cgroup_path(fn, arg_uid_shift);
+ if (r < 0)
+ log_error_errno(r, "Failed to chown() cgroup %s: %m", fn);
+finish:
+ if (undo_mount)
+ (void) umount_verbose(tree);
+
+ (void) rmdir(tree);
+ return r;
+}
+
+int create_subcgroup(pid_t pid, CGroupUnified unified_requested) {
+ _cleanup_free_ char *cgroup = NULL;
+ const char *child;
+ int unified, r;
+ CGroupMask supported;
+
+ /* In the unified hierarchy inner nodes may only contain
+ * subgroups, but not processes. Hence, if we running in the
+ * unified hierarchy and the container does the same, and we
+ * did not create a scope unit for the container move us and
+ * the container into two separate subcgroups. */
+
+ if (unified_requested == CGROUP_UNIFIED_NONE)
+ return 0;
+
+ unified = cg_unified(SYSTEMD_CGROUP_CONTROLLER);
+ if (unified < 0)
+ return log_error_errno(unified, "Failed to determine whether the unified hierarchy is used: %m");
+ if (unified == 0)
+ return 0;
+
+ r = cg_mask_supported(&supported);
+ if (r < 0)
+ return log_error_errno(r, "Failed to determine supported controllers: %m");
+
+ r = cg_pid_get_path(SYSTEMD_CGROUP_CONTROLLER, 0, &cgroup);
+ if (r < 0)
+ return log_error_errno(r, "Failed to get our control group: %m");
+
+ child = strjoina(cgroup, "/payload");
+ r = cg_create_and_attach(SYSTEMD_CGROUP_CONTROLLER, child, pid);
+ if (r < 0)
+ return log_error_errno(r, "Failed to create %s subcgroup: %m", child);
+
+ child = strjoina(cgroup, "/supervisor");
+ r = cg_create_and_attach(SYSTEMD_CGROUP_CONTROLLER, child, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to create %s subcgroup: %m", child);
+
+ /* Try to enable as many controllers as possible for the new payload. */
+ (void) cg_enable_everywhere(supported, supported, cgroup);
+ return 0;
+}
diff --git a/src/systemd-nspawn/nspawn-cgroup.h b/src/systemd-nspawn/nspawn-cgroup.h
new file mode 100644
index 0000000000..6c0ddfc7de
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-cgroup.h
@@ -0,0 +1,29 @@
+#pragma once
+
+/***
+ This file is part of systemd.
+
+ Copyright 2015 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include <stdbool.h>
+#include <sys/types.h>
+
+#include "systemd-basic/cgroup-util.h"
+
+int chown_cgroup(pid_t pid, uid_t uid_shift);
+int sync_cgroup(pid_t pid, CGroupUnified unified_requested, uid_t uid_shift);
+int create_subcgroup(pid_t pid, CGroupUnified unified_requested);
diff --git a/src/systemd-nspawn/nspawn-expose-ports.c b/src/systemd-nspawn/nspawn-expose-ports.c
new file mode 100644
index 0000000000..1ecebe8e30
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-expose-ports.c
@@ -0,0 +1,245 @@
+/***
+ This file is part of systemd.
+
+ Copyright 2015 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include "sd-netlink/local-addresses.h"
+#include "sd-netlink/netlink-util.h"
+#include "systemd-basic/alloc-util.h"
+#include "systemd-basic/fd-util.h"
+#include "systemd-basic/in-addr-util.h"
+#include "systemd-basic/parse-util.h"
+#include "systemd-basic/socket-util.h"
+#include "systemd-basic/string-util.h"
+#include "systemd-basic/util.h"
+#include "systemd-firewall/firewall-util.h"
+#include "systemd-staging/sd-netlink.h"
+
+#include "nspawn-expose-ports.h"
+
+int expose_port_parse(ExposePort **l, const char *s) {
+
+ const char *split, *e;
+ uint16_t container_port, host_port;
+ int protocol;
+ ExposePort *p;
+ int r;
+
+ assert(l);
+ assert(s);
+
+ if ((e = startswith(s, "tcp:")))
+ protocol = IPPROTO_TCP;
+ else if ((e = startswith(s, "udp:")))
+ protocol = IPPROTO_UDP;
+ else {
+ e = s;
+ protocol = IPPROTO_TCP;
+ }
+
+ split = strchr(e, ':');
+ if (split) {
+ char v[split - e + 1];
+
+ memcpy(v, e, split - e);
+ v[split - e] = 0;
+
+ r = safe_atou16(v, &host_port);
+ if (r < 0 || host_port <= 0)
+ return -EINVAL;
+
+ r = safe_atou16(split + 1, &container_port);
+ } else {
+ r = safe_atou16(e, &container_port);
+ host_port = container_port;
+ }
+
+ if (r < 0 || container_port <= 0)
+ return -EINVAL;
+
+ LIST_FOREACH(ports, p, *l)
+ if (p->protocol == protocol && p->host_port == host_port)
+ return -EEXIST;
+
+ p = new(ExposePort, 1);
+ if (!p)
+ return -ENOMEM;
+
+ p->protocol = protocol;
+ p->host_port = host_port;
+ p->container_port = container_port;
+
+ LIST_PREPEND(ports, *l, p);
+
+ return 0;
+}
+
+void expose_port_free_all(ExposePort *p) {
+
+ while (p) {
+ ExposePort *q = p;
+ LIST_REMOVE(ports, p, q);
+ free(q);
+ }
+}
+
+int expose_port_flush(ExposePort* l, union in_addr_union *exposed) {
+ ExposePort *p;
+ int r, af = AF_INET;
+
+ assert(exposed);
+
+ if (!l)
+ return 0;
+
+ if (in_addr_is_null(af, exposed))
+ return 0;
+
+ log_debug("Lost IP address.");
+
+ LIST_FOREACH(ports, p, l) {
+ r = fw_add_local_dnat(false,
+ af,
+ p->protocol,
+ NULL,
+ NULL, 0,
+ NULL, 0,
+ p->host_port,
+ exposed,
+ p->container_port,
+ NULL);
+ if (r < 0)
+ log_warning_errno(r, "Failed to modify firewall: %m");
+ }
+
+ *exposed = IN_ADDR_NULL;
+ return 0;
+}
+
+int expose_port_execute(sd_netlink *rtnl, ExposePort *l, union in_addr_union *exposed) {
+ _cleanup_free_ struct local_address *addresses = NULL;
+ _cleanup_free_ char *pretty = NULL;
+ union in_addr_union new_exposed;
+ ExposePort *p;
+ bool add;
+ int af = AF_INET, r;
+
+ assert(exposed);
+
+ /* Invoked each time an address is added or removed inside the
+ * container */
+
+ if (!l)
+ return 0;
+
+ r = local_addresses(rtnl, 0, af, &addresses);
+ if (r < 0)
+ return log_error_errno(r, "Failed to enumerate local addresses: %m");
+
+ add = r > 0 &&
+ addresses[0].family == af &&
+ addresses[0].scope < RT_SCOPE_LINK;
+
+ if (!add)
+ return expose_port_flush(l, exposed);
+
+ new_exposed = addresses[0].address;
+ if (in_addr_equal(af, exposed, &new_exposed))
+ return 0;
+
+ in_addr_to_string(af, &new_exposed, &pretty);
+ log_debug("New container IP is %s.", strna(pretty));
+
+ LIST_FOREACH(ports, p, l) {
+
+ r = fw_add_local_dnat(true,
+ af,
+ p->protocol,
+ NULL,
+ NULL, 0,
+ NULL, 0,
+ p->host_port,
+ &new_exposed,
+ p->container_port,
+ in_addr_is_null(af, exposed) ? NULL : exposed);
+ if (r < 0)
+ log_warning_errno(r, "Failed to modify firewall: %m");
+ }
+
+ *exposed = new_exposed;
+ return 0;
+}
+
+int expose_port_send_rtnl(int send_fd) {
+ _cleanup_close_ int fd = -1;
+ int r;
+
+ assert(send_fd >= 0);
+
+ fd = socket(PF_NETLINK, SOCK_RAW|SOCK_CLOEXEC|SOCK_NONBLOCK, NETLINK_ROUTE);
+ if (fd < 0)
+ return log_error_errno(errno, "Failed to allocate container netlink: %m");
+
+ /* Store away the fd in the socket, so that it stays open as
+ * long as we run the child */
+ r = send_one_fd(send_fd, fd, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to send netlink fd: %m");
+
+ return 0;
+}
+
+int expose_port_watch_rtnl(
+ sd_event *event,
+ int recv_fd,
+ sd_netlink_message_handler_t handler,
+ union in_addr_union *exposed,
+ sd_netlink **ret) {
+ _cleanup_(sd_netlink_unrefp) sd_netlink *rtnl = NULL;
+ int fd, r;
+
+ assert(event);
+ assert(recv_fd >= 0);
+ assert(ret);
+
+ fd = receive_one_fd(recv_fd, 0);
+ if (fd < 0)
+ return log_error_errno(fd, "Failed to recv netlink fd: %m");
+
+ r = sd_netlink_open_fd(&rtnl, fd);
+ if (r < 0) {
+ safe_close(fd);
+ return log_error_errno(r, "Failed to create rtnl object: %m");
+ }
+
+ r = sd_netlink_add_match(rtnl, RTM_NEWADDR, handler, exposed);
+ if (r < 0)
+ return log_error_errno(r, "Failed to subscribe to RTM_NEWADDR messages: %m");
+
+ r = sd_netlink_add_match(rtnl, RTM_DELADDR, handler, exposed);
+ if (r < 0)
+ return log_error_errno(r, "Failed to subscribe to RTM_DELADDR messages: %m");
+
+ r = sd_netlink_attach_event(rtnl, event, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to add to even loop: %m");
+
+ *ret = rtnl;
+ rtnl = NULL;
+
+ return 0;
+}
diff --git a/src/systemd-nspawn/nspawn-expose-ports.h b/src/systemd-nspawn/nspawn-expose-ports.h
new file mode 100644
index 0000000000..61134fcba6
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-expose-ports.h
@@ -0,0 +1,44 @@
+#pragma once
+
+/***
+ This file is part of systemd.
+
+ Copyright 2015 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include <inttypes.h>
+
+#include <systemd/sd-event.h>
+
+#include "systemd-basic/in-addr-util.h"
+#include "systemd-basic/list.h"
+#include "systemd-staging/sd-netlink.h"
+
+typedef struct ExposePort {
+ int protocol;
+ uint16_t host_port;
+ uint16_t container_port;
+ LIST_FIELDS(struct ExposePort, ports);
+} ExposePort;
+
+void expose_port_free_all(ExposePort *p);
+int expose_port_parse(ExposePort **l, const char *s);
+
+int expose_port_watch_rtnl(sd_event *event, int recv_fd, sd_netlink_message_handler_t handler, union in_addr_union *exposed, sd_netlink **ret);
+int expose_port_send_rtnl(int send_fd);
+
+int expose_port_execute(sd_netlink *rtnl, ExposePort *l, union in_addr_union *exposed);
+int expose_port_flush(ExposePort* l, union in_addr_union *exposed);
diff --git a/src/systemd-nspawn/nspawn-gperf.gperf b/src/systemd-nspawn/nspawn-gperf.gperf
new file mode 100644
index 0000000000..0c1dd98ff6
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-gperf.gperf
@@ -0,0 +1,47 @@
+%{
+#include <stddef.h>
+
+#include "systemd-shared/conf-parser.h"
+
+#include "nspawn-expose-ports.h"
+#include "nspawn-settings.h"
+%}
+struct ConfigPerfItem;
+%null_strings
+%language=ANSI-C
+%define slot-name section_and_lvalue
+%define hash-function-name nspawn_gperf_hash
+%define lookup-function-name nspawn_gperf_lookup
+%readonly-tables
+%omit-struct-type
+%struct-type
+%includes
+%%
+Exec.Boot, config_parse_boot, 0, 0
+Exec.ProcessTwo, config_parse_pid2, 0, 0
+Exec.Parameters, config_parse_strv, 0, offsetof(Settings, parameters)
+Exec.Environment, config_parse_strv, 0, offsetof(Settings, environment)
+Exec.User, config_parse_string, 0, offsetof(Settings, user)
+Exec.Capability, config_parse_capability, 0, offsetof(Settings, capability)
+Exec.DropCapability, config_parse_capability, 0, offsetof(Settings, drop_capability)
+Exec.KillSignal, config_parse_signal, 0, offsetof(Settings, kill_signal)
+Exec.Personality, config_parse_personality, 0, offsetof(Settings, personality)
+Exec.MachineID, config_parse_id128, 0, offsetof(Settings, machine_id)
+Exec.WorkingDirectory, config_parse_path, 0, offsetof(Settings, working_directory)
+Exec.PrivateUsers, config_parse_private_users, 0, 0
+Exec.NotifyReady, config_parse_bool, 0, offsetof(Settings, notify_ready)
+Files.ReadOnly, config_parse_tristate, 0, offsetof(Settings, read_only)
+Files.Volatile, config_parse_volatile_mode, 0, offsetof(Settings, volatile_mode)
+Files.Bind, config_parse_bind, 0, 0
+Files.BindReadOnly, config_parse_bind, 1, 0
+Files.TemporaryFileSystem, config_parse_tmpfs, 0, 0
+Files.PrivateUsersChown, config_parse_tristate, 0, offsetof(Settings, userns_chown)
+Network.Private, config_parse_tristate, 0, offsetof(Settings, private_network)
+Network.Interface, config_parse_strv, 0, offsetof(Settings, network_interfaces)
+Network.MACVLAN, config_parse_strv, 0, offsetof(Settings, network_macvlan)
+Network.IPVLAN, config_parse_strv, 0, offsetof(Settings, network_ipvlan)
+Network.VirtualEthernet, config_parse_tristate, 0, offsetof(Settings, network_veth)
+Network.VirtualEthernetExtra, config_parse_veth_extra, 0, 0
+Network.Bridge, config_parse_ifname, 0, offsetof(Settings, network_bridge)
+Network.Zone, config_parse_network_zone, 0, 0
+Network.Port, config_parse_expose_port, 0, 0
diff --git a/src/systemd-nspawn/nspawn-mount.c b/src/systemd-nspawn/nspawn-mount.c
new file mode 100644
index 0000000000..aac04efd4b
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-mount.c
@@ -0,0 +1,1164 @@
+/***
+ This file is part of systemd.
+
+ Copyright 2015 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include <sys/mount.h>
+
+#include <linux/magic.h>
+
+#include "systemd-basic/alloc-util.h"
+#include "systemd-basic/escape.h"
+#include "systemd-basic/fd-util.h"
+#include "systemd-basic/fileio.h"
+#include "systemd-basic/fs-util.h"
+#include "systemd-basic/label.h"
+#include "systemd-basic/mkdir.h"
+#include "systemd-basic/mount-util.h"
+#include "systemd-basic/parse-util.h"
+#include "systemd-basic/path-util.h"
+#include "systemd-basic/rm-rf.h"
+#include "systemd-basic/set.h"
+#include "systemd-basic/stat-util.h"
+#include "systemd-basic/string-util.h"
+#include "systemd-basic/strv.h"
+#include "systemd-basic/user-util.h"
+#include "systemd-basic/util.h"
+
+#include "nspawn-mount.h"
+
+CustomMount* custom_mount_add(CustomMount **l, unsigned *n, CustomMountType t) {
+ CustomMount *c, *ret;
+
+ assert(l);
+ assert(n);
+ assert(t >= 0);
+ assert(t < _CUSTOM_MOUNT_TYPE_MAX);
+
+ c = realloc(*l, (*n + 1) * sizeof(CustomMount));
+ if (!c)
+ return NULL;
+
+ *l = c;
+ ret = *l + *n;
+ (*n)++;
+
+ *ret = (CustomMount) { .type = t };
+
+ return ret;
+}
+
+void custom_mount_free_all(CustomMount *l, unsigned n) {
+ unsigned i;
+
+ for (i = 0; i < n; i++) {
+ CustomMount *m = l + i;
+
+ free(m->source);
+ free(m->destination);
+ free(m->options);
+
+ if (m->work_dir) {
+ (void) rm_rf(m->work_dir, REMOVE_ROOT|REMOVE_PHYSICAL);
+ free(m->work_dir);
+ }
+
+ strv_free(m->lower);
+ }
+
+ free(l);
+}
+
+int custom_mount_compare(const void *a, const void *b) {
+ const CustomMount *x = a, *y = b;
+ int r;
+
+ r = path_compare(x->destination, y->destination);
+ if (r != 0)
+ return r;
+
+ if (x->type < y->type)
+ return -1;
+ if (x->type > y->type)
+ return 1;
+
+ return 0;
+}
+
+int bind_mount_parse(CustomMount **l, unsigned *n, const char *s, bool read_only) {
+ _cleanup_free_ char *source = NULL, *destination = NULL, *opts = NULL;
+ const char *p = s;
+ CustomMount *m;
+ int r;
+
+ assert(l);
+ assert(n);
+
+ r = extract_many_words(&p, ":", EXTRACT_DONT_COALESCE_SEPARATORS, &source, &destination, NULL);
+ if (r < 0)
+ return r;
+ if (r == 0)
+ return -EINVAL;
+
+ if (r == 1) {
+ destination = strdup(source);
+ if (!destination)
+ return -ENOMEM;
+ }
+
+ if (r == 2 && !isempty(p)) {
+ opts = strdup(p);
+ if (!opts)
+ return -ENOMEM;
+ }
+
+ if (!path_is_absolute(source))
+ return -EINVAL;
+
+ if (!path_is_absolute(destination))
+ return -EINVAL;
+
+ m = custom_mount_add(l, n, CUSTOM_MOUNT_BIND);
+ if (!m)
+ return log_oom();
+
+ m->source = source;
+ m->destination = destination;
+ m->read_only = read_only;
+ m->options = opts;
+
+ source = destination = opts = NULL;
+ return 0;
+}
+
+int tmpfs_mount_parse(CustomMount **l, unsigned *n, const char *s) {
+ _cleanup_free_ char *path = NULL, *opts = NULL;
+ const char *p = s;
+ CustomMount *m;
+ int r;
+
+ assert(l);
+ assert(n);
+ assert(s);
+
+ r = extract_first_word(&p, &path, ":", EXTRACT_DONT_COALESCE_SEPARATORS);
+ if (r < 0)
+ return r;
+ if (r == 0)
+ return -EINVAL;
+
+ if (isempty(p))
+ opts = strdup("mode=0755");
+ else
+ opts = strdup(p);
+ if (!opts)
+ return -ENOMEM;
+
+ if (!path_is_absolute(path))
+ return -EINVAL;
+
+ m = custom_mount_add(l, n, CUSTOM_MOUNT_TMPFS);
+ if (!m)
+ return -ENOMEM;
+
+ m->destination = path;
+ m->options = opts;
+
+ path = opts = NULL;
+ return 0;
+}
+
+static int tmpfs_patch_options(
+ const char *options,
+ bool userns,
+ uid_t uid_shift, uid_t uid_range,
+ bool patch_ids,
+ const char *selinux_apifs_context,
+ char **ret) {
+
+ char *buf = NULL;
+
+ if ((userns && uid_shift != 0) || patch_ids) {
+ assert(uid_shift != UID_INVALID);
+
+ if (options)
+ (void) asprintf(&buf, "%s,uid=" UID_FMT ",gid=" UID_FMT, options, uid_shift, uid_shift);
+ else
+ (void) asprintf(&buf, "uid=" UID_FMT ",gid=" UID_FMT, uid_shift, uid_shift);
+ if (!buf)
+ return -ENOMEM;
+
+ options = buf;
+ }
+
+#ifdef HAVE_SELINUX
+ if (selinux_apifs_context) {
+ char *t;
+
+ if (options)
+ t = strjoin(options, ",context=\"", selinux_apifs_context, "\"", NULL);
+ else
+ t = strjoin("context=\"", selinux_apifs_context, "\"", NULL);
+ if (!t) {
+ free(buf);
+ return -ENOMEM;
+ }
+
+ free(buf);
+ buf = t;
+ }
+#endif
+
+ if (!buf && options) {
+ buf = strdup(options);
+ if (!buf)
+ return -ENOMEM;
+ }
+ *ret = buf;
+
+ return !!buf;
+}
+
+int mount_sysfs(const char *dest) {
+ const char *full, *top, *x;
+ int r;
+
+ top = prefix_roota(dest, "/sys");
+ r = path_check_fstype(top, SYSFS_MAGIC);
+ if (r < 0)
+ return log_error_errno(r, "Failed to determine filesystem type of %s: %m", top);
+ /* /sys might already be mounted as sysfs by the outer child in the
+ * !netns case. In this case, it's all good. Don't touch it because we
+ * don't have the right to do so, see https://github.com/systemd/systemd/issues/1555.
+ */
+ if (r > 0)
+ return 0;
+
+ full = prefix_roota(top, "/full");
+
+ (void) mkdir(full, 0755);
+
+ r = mount_verbose(LOG_ERR, "sysfs", full, "sysfs",
+ MS_RDONLY|MS_NOSUID|MS_NOEXEC|MS_NODEV, NULL);
+ if (r < 0)
+ return r;
+
+ FOREACH_STRING(x, "block", "bus", "class", "dev", "devices", "kernel") {
+ _cleanup_free_ char *from = NULL, *to = NULL;
+
+ from = prefix_root(full, x);
+ if (!from)
+ return log_oom();
+
+ to = prefix_root(top, x);
+ if (!to)
+ return log_oom();
+
+ (void) mkdir(to, 0755);
+
+ r = mount_verbose(LOG_ERR, from, to, NULL, MS_BIND, NULL);
+ if (r < 0)
+ return r;
+
+ r = mount_verbose(LOG_ERR, NULL, to, NULL,
+ MS_BIND|MS_RDONLY|MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_REMOUNT, NULL);
+ if (r < 0)
+ return r;
+ }
+
+ r = umount_verbose(full);
+ if (r < 0)
+ return r;
+
+ if (rmdir(full) < 0)
+ return log_error_errno(errno, "Failed to remove %s: %m", full);
+
+ x = prefix_roota(top, "/fs/kdbus");
+ (void) mkdir_p(x, 0755);
+
+ /* Create mountpoint for cgroups. Otherwise we are not allowed since we
+ * remount /sys read-only.
+ */
+ if (cg_ns_supported()) {
+ x = prefix_roota(top, "/fs/cgroup");
+ (void) mkdir_p(x, 0755);
+ }
+
+ return mount_verbose(LOG_ERR, NULL, top, NULL,
+ MS_BIND|MS_RDONLY|MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_REMOUNT, NULL);
+}
+
+static int mkdir_userns(const char *path, mode_t mode, bool in_userns, uid_t uid_shift) {
+ int r;
+
+ assert(path);
+
+ r = mkdir(path, mode);
+ if (r < 0 && errno != EEXIST)
+ return -errno;
+
+ if (!in_userns) {
+ r = lchown(path, uid_shift, uid_shift);
+ if (r < 0)
+ return -errno;
+ }
+
+ return 0;
+}
+
+static int mkdir_userns_p(const char *prefix, const char *path, mode_t mode, bool in_userns, uid_t uid_shift) {
+ const char *p, *e;
+ int r;
+
+ assert(path);
+
+ if (prefix && !path_startswith(path, prefix))
+ return -ENOTDIR;
+
+ /* create every parent directory in the path, except the last component */
+ p = path + strspn(path, "/");
+ for (;;) {
+ char t[strlen(path) + 1];
+
+ e = p + strcspn(p, "/");
+ p = e + strspn(e, "/");
+
+ /* Is this the last component? If so, then we're done */
+ if (*p == 0)
+ break;
+
+ memcpy(t, path, e - path);
+ t[e-path] = 0;
+
+ if (prefix && path_startswith(prefix, t))
+ continue;
+
+ r = mkdir_userns(t, mode, in_userns, uid_shift);
+ if (r < 0)
+ return r;
+ }
+
+ return mkdir_userns(path, mode, in_userns, uid_shift);
+}
+
+int mount_all(const char *dest,
+ bool use_userns, bool in_userns,
+ bool use_netns,
+ uid_t uid_shift, uid_t uid_range,
+ const char *selinux_apifs_context) {
+
+ typedef struct MountPoint {
+ const char *what;
+ const char *where;
+ const char *type;
+ const char *options;
+ unsigned long flags;
+ bool fatal;
+ bool in_userns;
+ bool use_netns;
+ } MountPoint;
+
+ static const MountPoint mount_table[] = {
+ { "proc", "/proc", "proc", NULL, MS_NOSUID|MS_NOEXEC|MS_NODEV, true, true, false },
+ { "/proc/sys", "/proc/sys", NULL, NULL, MS_BIND, true, true, false }, /* Bind mount first ...*/
+ { "/proc/sys/net", "/proc/sys/net", NULL, NULL, MS_BIND, true, true, true }, /* (except for this) */
+ { NULL, "/proc/sys", NULL, NULL, MS_BIND|MS_RDONLY|MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_REMOUNT, true, true, false }, /* ... then, make it r/o */
+ { "/proc/sysrq-trigger", "/proc/sysrq-trigger", NULL, NULL, MS_BIND, false, true, false }, /* Bind mount first ...*/
+ { NULL, "/proc/sysrq-trigger", NULL, NULL, MS_BIND|MS_RDONLY|MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_REMOUNT, false, true, false }, /* ... then, make it r/o */
+ { "tmpfs", "/sys", "tmpfs", "mode=755", MS_NOSUID|MS_NOEXEC|MS_NODEV, true, false, true },
+ { "sysfs", "/sys", "sysfs", NULL, MS_RDONLY|MS_NOSUID|MS_NOEXEC|MS_NODEV, true, false, false },
+ { "tmpfs", "/dev", "tmpfs", "mode=755", MS_NOSUID|MS_STRICTATIME, true, false, false },
+ { "tmpfs", "/dev/shm", "tmpfs", "mode=1777", MS_NOSUID|MS_NODEV|MS_STRICTATIME, true, false, false },
+ { "tmpfs", "/run", "tmpfs", "mode=755", MS_NOSUID|MS_NODEV|MS_STRICTATIME, true, false, false },
+ { "tmpfs", "/tmp", "tmpfs", "mode=1777", MS_STRICTATIME, true, false, false },
+#ifdef HAVE_SELINUX
+ { "/sys/fs/selinux", "/sys/fs/selinux", NULL, NULL, MS_BIND, false, false, false }, /* Bind mount first */
+ { NULL, "/sys/fs/selinux", NULL, NULL, MS_BIND|MS_RDONLY|MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_REMOUNT, false, false, false }, /* Then, make it r/o */
+#endif
+ };
+
+ unsigned k;
+ int r;
+
+ for (k = 0; k < ELEMENTSOF(mount_table); k++) {
+ _cleanup_free_ char *where = NULL, *options = NULL;
+ const char *o;
+
+ if (in_userns != mount_table[k].in_userns)
+ continue;
+
+ if (!use_netns && mount_table[k].use_netns)
+ continue;
+
+ where = prefix_root(dest, mount_table[k].where);
+ if (!where)
+ return log_oom();
+
+ r = path_is_mount_point(where, AT_SYMLINK_FOLLOW);
+ if (r < 0 && r != -ENOENT)
+ return log_error_errno(r, "Failed to detect whether %s is a mount point: %m", where);
+
+ /* Skip this entry if it is not a remount. */
+ if (mount_table[k].what && r > 0)
+ continue;
+
+ r = mkdir_userns_p(dest, where, 0755, in_userns, uid_shift);
+ if (r < 0 && r != -EEXIST) {
+ if (mount_table[k].fatal)
+ return log_error_errno(r, "Failed to create directory %s: %m", where);
+
+ log_debug_errno(r, "Failed to create directory %s: %m", where);
+ continue;
+ }
+
+ o = mount_table[k].options;
+ if (streq_ptr(mount_table[k].type, "tmpfs")) {
+ if (in_userns)
+ r = tmpfs_patch_options(o, use_userns, 0, uid_range, true, selinux_apifs_context, &options);
+ else
+ r = tmpfs_patch_options(o, use_userns, uid_shift, uid_range, false, selinux_apifs_context, &options);
+ if (r < 0)
+ return log_oom();
+ if (r > 0)
+ o = options;
+ }
+
+ r = mount_verbose(mount_table[k].fatal ? LOG_ERR : LOG_WARNING,
+ mount_table[k].what,
+ where,
+ mount_table[k].type,
+ mount_table[k].flags,
+ o);
+ if (r < 0 && mount_table[k].fatal)
+ return r;
+ }
+
+ return 0;
+}
+
+static int parse_mount_bind_options(const char *options, unsigned long *mount_flags, char **mount_opts) {
+ const char *p = options;
+ unsigned long flags = *mount_flags;
+ char *opts = NULL;
+
+ assert(options);
+
+ for (;;) {
+ _cleanup_free_ char *word = NULL;
+ int r = extract_first_word(&p, &word, ",", 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to extract mount option: %m");
+ if (r == 0)
+ break;
+
+ if (streq(word, "rbind"))
+ flags |= MS_REC;
+ else if (streq(word, "norbind"))
+ flags &= ~MS_REC;
+ else {
+ log_error("Invalid bind mount option: %s", word);
+ return -EINVAL;
+ }
+ }
+
+ *mount_flags = flags;
+ /* in the future mount_opts will hold string options for mount(2) */
+ *mount_opts = opts;
+
+ return 0;
+}
+
+static int mount_bind(const char *dest, CustomMount *m) {
+ struct stat source_st, dest_st;
+ const char *where;
+ unsigned long mount_flags = MS_BIND | MS_REC;
+ _cleanup_free_ char *mount_opts = NULL;
+ int r;
+
+ assert(m);
+
+ if (m->options) {
+ r = parse_mount_bind_options(m->options, &mount_flags, &mount_opts);
+ if (r < 0)
+ return r;
+ }
+
+ if (stat(m->source, &source_st) < 0)
+ return log_error_errno(errno, "Failed to stat %s: %m", m->source);
+
+ where = prefix_roota(dest, m->destination);
+
+ if (stat(where, &dest_st) >= 0) {
+ if (S_ISDIR(source_st.st_mode) && !S_ISDIR(dest_st.st_mode)) {
+ log_error("Cannot bind mount directory %s on file %s.", m->source, where);
+ return -EINVAL;
+ }
+
+ if (!S_ISDIR(source_st.st_mode) && S_ISDIR(dest_st.st_mode)) {
+ log_error("Cannot bind mount file %s on directory %s.", m->source, where);
+ return -EINVAL;
+ }
+
+ } else if (errno == ENOENT) {
+ r = mkdir_parents_label(where, 0755);
+ if (r < 0)
+ return log_error_errno(r, "Failed to make parents of %s: %m", where);
+
+ /* Create the mount point. Any non-directory file can be
+ * mounted on any non-directory file (regular, fifo, socket,
+ * char, block).
+ */
+ if (S_ISDIR(source_st.st_mode))
+ r = mkdir_label(where, 0755);
+ else
+ r = touch(where);
+ if (r < 0)
+ return log_error_errno(r, "Failed to create mount point %s: %m", where);
+
+ } else
+ return log_error_errno(errno, "Failed to stat %s: %m", where);
+
+ r = mount_verbose(LOG_ERR, m->source, where, NULL, mount_flags, mount_opts);
+ if (r < 0)
+ return r;
+
+ if (m->read_only) {
+ r = bind_remount_recursive(where, true, NULL);
+ if (r < 0)
+ return log_error_errno(r, "Read-only bind mount failed: %m");
+ }
+
+ return 0;
+}
+
+static int mount_tmpfs(
+ const char *dest,
+ CustomMount *m,
+ bool userns, uid_t uid_shift, uid_t uid_range,
+ const char *selinux_apifs_context) {
+
+ const char *where, *options;
+ _cleanup_free_ char *buf = NULL;
+ int r;
+
+ assert(dest);
+ assert(m);
+
+ where = prefix_roota(dest, m->destination);
+
+ r = mkdir_p_label(where, 0755);
+ if (r < 0 && r != -EEXIST)
+ return log_error_errno(r, "Creating mount point for tmpfs %s failed: %m", where);
+
+ r = tmpfs_patch_options(m->options, userns, uid_shift, uid_range, false, selinux_apifs_context, &buf);
+ if (r < 0)
+ return log_oom();
+ options = r > 0 ? buf : m->options;
+
+ return mount_verbose(LOG_ERR, "tmpfs", where, "tmpfs", MS_NODEV|MS_STRICTATIME, options);
+}
+
+static char *joined_and_escaped_lower_dirs(char * const *lower) {
+ _cleanup_strv_free_ char **sv = NULL;
+
+ sv = strv_copy(lower);
+ if (!sv)
+ return NULL;
+
+ strv_reverse(sv);
+
+ if (!strv_shell_escape(sv, ",:"))
+ return NULL;
+
+ return strv_join(sv, ":");
+}
+
+static int mount_overlay(const char *dest, CustomMount *m) {
+ _cleanup_free_ char *lower = NULL;
+ const char *where, *options;
+ int r;
+
+ assert(dest);
+ assert(m);
+
+ where = prefix_roota(dest, m->destination);
+
+ r = mkdir_label(where, 0755);
+ if (r < 0 && r != -EEXIST)
+ return log_error_errno(r, "Creating mount point for overlay %s failed: %m", where);
+
+ (void) mkdir_p_label(m->source, 0755);
+
+ lower = joined_and_escaped_lower_dirs(m->lower);
+ if (!lower)
+ return log_oom();
+
+ if (m->read_only) {
+ _cleanup_free_ char *escaped_source = NULL;
+
+ escaped_source = shell_escape(m->source, ",:");
+ if (!escaped_source)
+ return log_oom();
+
+ options = strjoina("lowerdir=", escaped_source, ":", lower);
+ } else {
+ _cleanup_free_ char *escaped_source = NULL, *escaped_work_dir = NULL;
+
+ assert(m->work_dir);
+ (void) mkdir_label(m->work_dir, 0700);
+
+ escaped_source = shell_escape(m->source, ",:");
+ if (!escaped_source)
+ return log_oom();
+ escaped_work_dir = shell_escape(m->work_dir, ",:");
+ if (!escaped_work_dir)
+ return log_oom();
+
+ options = strjoina("lowerdir=", lower, ",upperdir=", escaped_source, ",workdir=", escaped_work_dir);
+ }
+
+ return mount_verbose(LOG_ERR, "overlay", where, "overlay", m->read_only ? MS_RDONLY : 0, options);
+}
+
+int mount_custom(
+ const char *dest,
+ CustomMount *mounts, unsigned n,
+ bool userns, uid_t uid_shift, uid_t uid_range,
+ const char *selinux_apifs_context) {
+
+ unsigned i;
+ int r;
+
+ assert(dest);
+
+ for (i = 0; i < n; i++) {
+ CustomMount *m = mounts + i;
+
+ switch (m->type) {
+
+ case CUSTOM_MOUNT_BIND:
+ r = mount_bind(dest, m);
+ break;
+
+ case CUSTOM_MOUNT_TMPFS:
+ r = mount_tmpfs(dest, m, userns, uid_shift, uid_range, selinux_apifs_context);
+ break;
+
+ case CUSTOM_MOUNT_OVERLAY:
+ r = mount_overlay(dest, m);
+ break;
+
+ default:
+ assert_not_reached("Unknown custom mount type");
+ }
+
+ if (r < 0)
+ return r;
+ }
+
+ return 0;
+}
+
+/* Retrieve existing subsystems. This function is called in a new cgroup
+ * namespace.
+ */
+static int get_controllers(Set *subsystems) {
+ _cleanup_fclose_ FILE *f = NULL;
+ char line[LINE_MAX];
+
+ assert(subsystems);
+
+ f = fopen("/proc/self/cgroup", "re");
+ if (!f)
+ return errno == ENOENT ? -ESRCH : -errno;
+
+ FOREACH_LINE(line, f, return -errno) {
+ int r;
+ char *e, *l, *p;
+
+ l = strchr(line, ':');
+ if (!l)
+ continue;
+
+ l++;
+ e = strchr(l, ':');
+ if (!e)
+ continue;
+
+ *e = 0;
+
+ if (STR_IN_SET(l, "", "name=systemd"))
+ continue;
+
+ p = strdup(l);
+ if (!p)
+ return -ENOMEM;
+
+ r = set_consume(subsystems, p);
+ if (r < 0)
+ return r;
+ }
+
+ return 0;
+}
+
+static int mount_legacy_cgroup_hierarchy(const char *dest, const char *controller, const char *hierarchy,
+ CGroupUnified unified_requested, bool read_only) {
+ const char *to, *fstype, *opts;
+ int r;
+
+ to = strjoina(strempty(dest), "/sys/fs/cgroup/", hierarchy);
+
+ r = path_is_mount_point(to, 0);
+ if (r < 0 && r != -ENOENT)
+ return log_error_errno(r, "Failed to determine if %s is mounted already: %m", to);
+ if (r > 0)
+ return 0;
+
+ mkdir_p(to, 0755);
+
+ /* The superblock mount options of the mount point need to be
+ * identical to the hosts', and hence writable... */
+ if (streq(controller, SYSTEMD_CGROUP_CONTROLLER)) {
+ if (unified_requested >= CGROUP_UNIFIED_SYSTEMD) {
+ fstype = "cgroup2";
+ opts = NULL;
+ } else {
+ fstype = "cgroup";
+ opts = "none,name=systemd,xattr";
+ }
+ } else {
+ fstype = "cgroup";
+ opts = controller;
+ }
+
+ r = mount_verbose(LOG_ERR, "cgroup", to, fstype, MS_NOSUID|MS_NOEXEC|MS_NODEV, opts);
+ if (r < 0)
+ return r;
+
+ /* ... hence let's only make the bind mount read-only, not the superblock. */
+ if (read_only) {
+ r = mount_verbose(LOG_ERR, NULL, to, NULL,
+ MS_BIND|MS_REMOUNT|MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_RDONLY, NULL);
+ if (r < 0)
+ return r;
+ }
+
+ return 1;
+}
+
+/* Mount a legacy cgroup hierarchy when cgroup namespaces are supported. */
+static int mount_legacy_cgns_supported(
+ CGroupUnified unified_requested, bool userns, uid_t uid_shift,
+ uid_t uid_range, const char *selinux_apifs_context) {
+ _cleanup_set_free_free_ Set *controllers = NULL;
+ const char *cgroup_root = "/sys/fs/cgroup", *c;
+ int r;
+
+ (void) mkdir_p(cgroup_root, 0755);
+
+ /* Mount a tmpfs to /sys/fs/cgroup if it's not mounted there yet. */
+ r = path_is_mount_point(cgroup_root, AT_SYMLINK_FOLLOW);
+ if (r < 0)
+ return log_error_errno(r, "Failed to determine if /sys/fs/cgroup is already mounted: %m");
+ if (r == 0) {
+ _cleanup_free_ char *options = NULL;
+
+ /* When cgroup namespaces are enabled and user namespaces are
+ * used then the mount of the cgroupfs is done *inside* the new
+ * user namespace. We're root in the new user namespace and the
+ * kernel will happily translate our uid/gid to the correct
+ * uid/gid as seen from e.g. /proc/1/mountinfo. So we simply
+ * pass uid 0 and not uid_shift to tmpfs_patch_options().
+ */
+ r = tmpfs_patch_options("mode=755", userns, 0, uid_range, true, selinux_apifs_context, &options);
+ if (r < 0)
+ return log_oom();
+
+ r = mount_verbose(LOG_ERR, "tmpfs", cgroup_root, "tmpfs",
+ MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_STRICTATIME, options);
+ if (r < 0)
+ return r;
+ }
+
+ if (cg_all_unified() > 0)
+ goto skip_controllers;
+
+ controllers = set_new(&string_hash_ops);
+ if (!controllers)
+ return log_oom();
+
+ r = get_controllers(controllers);
+ if (r < 0)
+ return log_error_errno(r, "Failed to determine cgroup controllers: %m");
+
+ for (;;) {
+ _cleanup_free_ const char *controller = NULL;
+
+ controller = set_steal_first(controllers);
+ if (!controller)
+ break;
+
+ r = mount_legacy_cgroup_hierarchy("", controller, controller, unified_requested, !userns);
+ if (r < 0)
+ return r;
+
+ /* When multiple hierarchies are co-mounted, make their
+ * constituting individual hierarchies a symlink to the
+ * co-mount.
+ */
+ c = controller;
+ for (;;) {
+ _cleanup_free_ char *target = NULL, *tok = NULL;
+
+ r = extract_first_word(&c, &tok, ",", 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to extract co-mounted cgroup controller: %m");
+ if (r == 0)
+ break;
+
+ target = prefix_root("/sys/fs/cgroup", tok);
+ if (!target)
+ return log_oom();
+
+ if (streq(controller, tok))
+ break;
+
+ r = symlink_idempotent(controller, target);
+ if (r == -EINVAL)
+ return log_error_errno(r, "Invalid existing symlink for combined hierarchy: %m");
+ if (r < 0)
+ return log_error_errno(r, "Failed to create symlink for combined hierarchy: %m");
+ }
+ }
+
+skip_controllers:
+ r = mount_legacy_cgroup_hierarchy("", SYSTEMD_CGROUP_CONTROLLER, "systemd", unified_requested, false);
+ if (r < 0)
+ return r;
+
+ if (!userns)
+ return mount_verbose(LOG_ERR, NULL, cgroup_root, NULL,
+ MS_REMOUNT|MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_STRICTATIME|MS_RDONLY, "mode=755");
+
+ return 0;
+}
+
+/* Mount legacy cgroup hierarchy when cgroup namespaces are unsupported. */
+static int mount_legacy_cgns_unsupported(
+ const char *dest,
+ CGroupUnified unified_requested, bool userns, uid_t uid_shift, uid_t uid_range,
+ const char *selinux_apifs_context) {
+ _cleanup_set_free_free_ Set *controllers = NULL;
+ const char *cgroup_root;
+ int r;
+
+ cgroup_root = prefix_roota(dest, "/sys/fs/cgroup");
+
+ (void) mkdir_p(cgroup_root, 0755);
+
+ /* Mount a tmpfs to /sys/fs/cgroup if it's not mounted there yet. */
+ r = path_is_mount_point(cgroup_root, AT_SYMLINK_FOLLOW);
+ if (r < 0)
+ return log_error_errno(r, "Failed to determine if /sys/fs/cgroup is already mounted: %m");
+ if (r == 0) {
+ _cleanup_free_ char *options = NULL;
+
+ r = tmpfs_patch_options("mode=755", userns, uid_shift, uid_range, false, selinux_apifs_context, &options);
+ if (r < 0)
+ return log_oom();
+
+ r = mount_verbose(LOG_ERR, "tmpfs", cgroup_root, "tmpfs",
+ MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_STRICTATIME, options);
+ if (r < 0)
+ return r;
+ }
+
+ if (cg_all_unified() > 0)
+ goto skip_controllers;
+
+ controllers = set_new(&string_hash_ops);
+ if (!controllers)
+ return log_oom();
+
+ r = cg_kernel_controllers(controllers);
+ if (r < 0)
+ return log_error_errno(r, "Failed to determine cgroup controllers: %m");
+
+ for (;;) {
+ _cleanup_free_ char *controller = NULL, *origin = NULL, *combined = NULL;
+
+ controller = set_steal_first(controllers);
+ if (!controller)
+ break;
+
+ origin = prefix_root("/sys/fs/cgroup/", controller);
+ if (!origin)
+ return log_oom();
+
+ r = readlink_malloc(origin, &combined);
+ if (r == -EINVAL) {
+ /* Not a symbolic link, but directly a single cgroup hierarchy */
+
+ r = mount_legacy_cgroup_hierarchy(dest, controller, controller, unified_requested, true);
+ if (r < 0)
+ return r;
+
+ } else if (r < 0)
+ return log_error_errno(r, "Failed to read link %s: %m", origin);
+ else {
+ _cleanup_free_ char *target = NULL;
+
+ target = prefix_root(dest, origin);
+ if (!target)
+ return log_oom();
+
+ /* A symbolic link, a combination of controllers in one hierarchy */
+
+ if (!filename_is_valid(combined)) {
+ log_warning("Ignoring invalid combined hierarchy %s.", combined);
+ continue;
+ }
+
+ r = mount_legacy_cgroup_hierarchy(dest, combined, combined, unified_requested, true);
+ if (r < 0)
+ return r;
+
+ r = symlink_idempotent(combined, target);
+ if (r == -EINVAL)
+ return log_error_errno(r, "Invalid existing symlink for combined hierarchy: %m");
+ if (r < 0)
+ return log_error_errno(r, "Failed to create symlink for combined hierarchy: %m");
+ }
+ }
+
+skip_controllers:
+ r = mount_legacy_cgroup_hierarchy(dest, SYSTEMD_CGROUP_CONTROLLER, "systemd", unified_requested, false);
+ if (r < 0)
+ return r;
+
+ return mount_verbose(LOG_ERR, NULL, cgroup_root, NULL,
+ MS_REMOUNT|MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_STRICTATIME|MS_RDONLY, "mode=755");
+}
+
+static int mount_unified_cgroups(const char *dest) {
+ const char *p;
+ int r;
+
+ assert(dest);
+
+ p = prefix_roota(dest, "/sys/fs/cgroup");
+
+ (void) mkdir_p(p, 0755);
+
+ r = path_is_mount_point(p, AT_SYMLINK_FOLLOW);
+ if (r < 0)
+ return log_error_errno(r, "Failed to determine if %s is mounted already: %m", p);
+ if (r > 0) {
+ p = prefix_roota(dest, "/sys/fs/cgroup/cgroup.procs");
+ if (access(p, F_OK) >= 0)
+ return 0;
+ if (errno != ENOENT)
+ return log_error_errno(errno, "Failed to determine if mount point %s contains the unified cgroup hierarchy: %m", p);
+
+ log_error("%s is already mounted but not a unified cgroup hierarchy. Refusing.", p);
+ return -EINVAL;
+ }
+
+ return mount_verbose(LOG_ERR, "cgroup", p, "cgroup2", MS_NOSUID|MS_NOEXEC|MS_NODEV, NULL);
+}
+
+int mount_cgroups(
+ const char *dest,
+ CGroupUnified unified_requested,
+ bool userns, uid_t uid_shift, uid_t uid_range,
+ const char *selinux_apifs_context,
+ bool use_cgns) {
+
+ if (unified_requested >= CGROUP_UNIFIED_ALL)
+ return mount_unified_cgroups(dest);
+ else if (use_cgns)
+ return mount_legacy_cgns_supported(unified_requested, userns, uid_shift, uid_range, selinux_apifs_context);
+
+ return mount_legacy_cgns_unsupported(dest, unified_requested, userns, uid_shift, uid_range, selinux_apifs_context);
+}
+
+int mount_systemd_cgroup_writable(
+ const char *dest,
+ CGroupUnified unified_requested) {
+
+ _cleanup_free_ char *own_cgroup_path = NULL;
+ const char *systemd_root, *systemd_own;
+ int r;
+
+ assert(dest);
+
+ r = cg_pid_get_path(NULL, 0, &own_cgroup_path);
+ if (r < 0)
+ return log_error_errno(r, "Failed to determine our own cgroup path: %m");
+
+ /* If we are living in the top-level, then there's nothing to do... */
+ if (path_equal(own_cgroup_path, "/"))
+ return 0;
+
+ if (unified_requested >= CGROUP_UNIFIED_ALL) {
+ systemd_own = strjoina(dest, "/sys/fs/cgroup", own_cgroup_path);
+ systemd_root = prefix_roota(dest, "/sys/fs/cgroup");
+ } else {
+ systemd_own = strjoina(dest, "/sys/fs/cgroup/systemd", own_cgroup_path);
+ systemd_root = prefix_roota(dest, "/sys/fs/cgroup/systemd");
+ }
+
+ /* Make our own cgroup a (writable) bind mount */
+ r = mount_verbose(LOG_ERR, systemd_own, systemd_own, NULL, MS_BIND, NULL);
+ if (r < 0)
+ return r;
+
+ /* And then remount the systemd cgroup root read-only */
+ return mount_verbose(LOG_ERR, NULL, systemd_root, NULL,
+ MS_BIND|MS_REMOUNT|MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_RDONLY, NULL);
+}
+
+int setup_volatile_state(
+ const char *directory,
+ VolatileMode mode,
+ bool userns, uid_t uid_shift, uid_t uid_range,
+ const char *selinux_apifs_context) {
+
+ _cleanup_free_ char *buf = NULL;
+ const char *p, *options;
+ int r;
+
+ assert(directory);
+
+ if (mode != VOLATILE_STATE)
+ return 0;
+
+ /* --volatile=state means we simply overmount /var
+ with a tmpfs, and the rest read-only. */
+
+ r = bind_remount_recursive(directory, true, NULL);
+ if (r < 0)
+ return log_error_errno(r, "Failed to remount %s read-only: %m", directory);
+
+ p = prefix_roota(directory, "/var");
+ r = mkdir(p, 0755);
+ if (r < 0 && errno != EEXIST)
+ return log_error_errno(errno, "Failed to create %s: %m", directory);
+
+ options = "mode=755";
+ r = tmpfs_patch_options(options, userns, uid_shift, uid_range, false, selinux_apifs_context, &buf);
+ if (r < 0)
+ return log_oom();
+ if (r > 0)
+ options = buf;
+
+ return mount_verbose(LOG_ERR, "tmpfs", p, "tmpfs", MS_STRICTATIME, options);
+}
+
+int setup_volatile(
+ const char *directory,
+ VolatileMode mode,
+ bool userns, uid_t uid_shift, uid_t uid_range,
+ const char *selinux_apifs_context) {
+
+ bool tmpfs_mounted = false, bind_mounted = false;
+ char template[] = "/tmp/nspawn-volatile-XXXXXX";
+ _cleanup_free_ char *buf = NULL;
+ const char *f, *t, *options;
+ int r;
+
+ assert(directory);
+
+ if (mode != VOLATILE_YES)
+ return 0;
+
+ /* --volatile=yes means we mount a tmpfs to the root dir, and
+ the original /usr to use inside it, and that read-only. */
+
+ if (!mkdtemp(template))
+ return log_error_errno(errno, "Failed to create temporary directory: %m");
+
+ options = "mode=755";
+ r = tmpfs_patch_options(options, userns, uid_shift, uid_range, false, selinux_apifs_context, &buf);
+ if (r < 0)
+ return log_oom();
+ if (r > 0)
+ options = buf;
+
+ r = mount_verbose(LOG_ERR, "tmpfs", template, "tmpfs", MS_STRICTATIME, options);
+ if (r < 0)
+ goto fail;
+
+ tmpfs_mounted = true;
+
+ f = prefix_roota(directory, "/usr");
+ t = prefix_roota(template, "/usr");
+
+ r = mkdir(t, 0755);
+ if (r < 0 && errno != EEXIST) {
+ r = log_error_errno(errno, "Failed to create %s: %m", t);
+ goto fail;
+ }
+
+ r = mount_verbose(LOG_ERR, f, t, NULL, MS_BIND|MS_REC, NULL);
+ if (r < 0)
+ goto fail;
+
+ bind_mounted = true;
+
+ r = bind_remount_recursive(t, true, NULL);
+ if (r < 0) {
+ log_error_errno(r, "Failed to remount %s read-only: %m", t);
+ goto fail;
+ }
+
+ r = mount_verbose(LOG_ERR, template, directory, NULL, MS_MOVE, NULL);
+ if (r < 0)
+ goto fail;
+
+ (void) rmdir(template);
+
+ return 0;
+
+fail:
+ if (bind_mounted)
+ (void) umount_verbose(t);
+
+ if (tmpfs_mounted)
+ (void) umount_verbose(template);
+ (void) rmdir(template);
+ return r;
+}
+
+VolatileMode volatile_mode_from_string(const char *s) {
+ int b;
+
+ if (isempty(s))
+ return _VOLATILE_MODE_INVALID;
+
+ b = parse_boolean(s);
+ if (b > 0)
+ return VOLATILE_YES;
+ if (b == 0)
+ return VOLATILE_NO;
+
+ if (streq(s, "state"))
+ return VOLATILE_STATE;
+
+ return _VOLATILE_MODE_INVALID;
+}
diff --git a/src/systemd-nspawn/nspawn-mount.h b/src/systemd-nspawn/nspawn-mount.h
new file mode 100644
index 0000000000..0d58439df3
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-mount.h
@@ -0,0 +1,71 @@
+#pragma once
+
+/***
+ This file is part of systemd.
+
+ Copyright 2015 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include <stdbool.h>
+
+#include "systemd-basic/cgroup-util.h"
+
+typedef enum VolatileMode {
+ VOLATILE_NO,
+ VOLATILE_YES,
+ VOLATILE_STATE,
+ _VOLATILE_MODE_MAX,
+ _VOLATILE_MODE_INVALID = -1
+} VolatileMode;
+
+typedef enum CustomMountType {
+ CUSTOM_MOUNT_BIND,
+ CUSTOM_MOUNT_TMPFS,
+ CUSTOM_MOUNT_OVERLAY,
+ _CUSTOM_MOUNT_TYPE_MAX,
+ _CUSTOM_MOUNT_TYPE_INVALID = -1
+} CustomMountType;
+
+typedef struct CustomMount {
+ CustomMountType type;
+ bool read_only;
+ char *source; /* for overlayfs this is the upper directory */
+ char *destination;
+ char *options;
+ char *work_dir;
+ char **lower;
+} CustomMount;
+
+CustomMount* custom_mount_add(CustomMount **l, unsigned *n, CustomMountType t);
+
+void custom_mount_free_all(CustomMount *l, unsigned n);
+int bind_mount_parse(CustomMount **l, unsigned *n, const char *s, bool read_only);
+int tmpfs_mount_parse(CustomMount **l, unsigned *n, const char *s);
+
+int custom_mount_compare(const void *a, const void *b);
+
+int mount_all(const char *dest, bool use_userns, bool in_userns, bool use_netns, uid_t uid_shift, uid_t uid_range, const char *selinux_apifs_context);
+int mount_sysfs(const char *dest);
+
+int mount_cgroups(const char *dest, CGroupUnified unified_requested, bool userns, uid_t uid_shift, uid_t uid_range, const char *selinux_apifs_context, bool use_cgns);
+int mount_systemd_cgroup_writable(const char *dest, CGroupUnified unified_requested);
+
+int mount_custom(const char *dest, CustomMount *mounts, unsigned n, bool userns, uid_t uid_shift, uid_t uid_range, const char *selinux_apifs_context);
+
+int setup_volatile(const char *directory, VolatileMode mode, bool userns, uid_t uid_shift, uid_t uid_range, const char *selinux_apifs_context);
+int setup_volatile_state(const char *directory, VolatileMode mode, bool userns, uid_t uid_shift, uid_t uid_range, const char *selinux_apifs_context);
+
+VolatileMode volatile_mode_from_string(const char *s);
diff --git a/src/systemd-nspawn/nspawn-network.c b/src/systemd-nspawn/nspawn-network.c
new file mode 100644
index 0000000000..2e0760ec86
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-network.c
@@ -0,0 +1,696 @@
+/***
+ This file is part of systemd.
+
+ Copyright 2015 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include <net/if.h>
+
+#include <linux/veth.h>
+
+#include <libudev.h>
+#include <systemd/sd-id128.h>
+
+#include "sd-netlink/netlink-util.h"
+#include "systemd-basic/alloc-util.h"
+#include "systemd-basic/ether-addr-util.h"
+#include "systemd-basic/lockfile-util.h"
+#include "systemd-basic/siphash24.h"
+#include "systemd-basic/socket-util.h"
+#include "systemd-basic/stat-util.h"
+#include "systemd-basic/string-util.h"
+#include "systemd-basic/util.h"
+#include "systemd-shared/udev-util.h"
+#include "systemd-staging/sd-netlink.h"
+
+#include "nspawn-network.h"
+
+#define HOST_HASH_KEY SD_ID128_MAKE(1a,37,6f,c7,46,ec,45,0b,ad,a3,d5,31,06,60,5d,b1)
+#define CONTAINER_HASH_KEY SD_ID128_MAKE(c3,c4,f9,19,b5,57,b2,1c,e6,cf,14,27,03,9c,ee,a2)
+#define VETH_EXTRA_HOST_HASH_KEY SD_ID128_MAKE(48,c7,f6,b7,ea,9d,4c,9e,b7,28,d4,de,91,d5,bf,66)
+#define VETH_EXTRA_CONTAINER_HASH_KEY SD_ID128_MAKE(af,50,17,61,ce,f9,4d,35,84,0d,2b,20,54,be,ce,59)
+#define MACVLAN_HASH_KEY SD_ID128_MAKE(00,13,6d,bc,66,83,44,81,bb,0c,f9,51,1f,24,a6,6f)
+
+static int remove_one_link(sd_netlink *rtnl, const char *name) {
+ _cleanup_(sd_netlink_message_unrefp) sd_netlink_message *m = NULL;
+ int r;
+
+ if (isempty(name))
+ return 0;
+
+ r = sd_rtnl_message_new_link(rtnl, &m, RTM_DELLINK, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to allocate netlink message: %m");
+
+ r = sd_netlink_message_append_string(m, IFLA_IFNAME, name);
+ if (r < 0)
+ return log_error_errno(r, "Failed to add netlink interface name: %m");
+
+ r = sd_netlink_call(rtnl, m, 0, NULL);
+ if (r == -ENODEV) /* Already gone */
+ return 0;
+ if (r < 0)
+ return log_error_errno(r, "Failed to remove interface %s: %m", name);
+
+ return 1;
+}
+
+static int generate_mac(
+ const char *machine_name,
+ struct ether_addr *mac,
+ sd_id128_t hash_key,
+ uint64_t idx) {
+
+ uint64_t result;
+ size_t l, sz;
+ uint8_t *v, *i;
+ int r;
+
+ l = strlen(machine_name);
+ sz = sizeof(sd_id128_t) + l;
+ if (idx > 0)
+ sz += sizeof(idx);
+
+ v = alloca(sz);
+
+ /* fetch some persistent data unique to the host */
+ r = sd_id128_get_machine((sd_id128_t*) v);
+ if (r < 0)
+ return r;
+
+ /* combine with some data unique (on this host) to this
+ * container instance */
+ i = mempcpy(v + sizeof(sd_id128_t), machine_name, l);
+ if (idx > 0) {
+ idx = htole64(idx);
+ memcpy(i, &idx, sizeof(idx));
+ }
+
+ /* Let's hash the host machine ID plus the container name. We
+ * use a fixed, but originally randomly created hash key here. */
+ result = htole64(siphash24(v, sz, hash_key.bytes));
+
+ assert_cc(ETH_ALEN <= sizeof(result));
+ memcpy(mac->ether_addr_octet, &result, ETH_ALEN);
+
+ /* see eth_random_addr in the kernel */
+ mac->ether_addr_octet[0] &= 0xfe; /* clear multicast bit */
+ mac->ether_addr_octet[0] |= 0x02; /* set local assignment bit (IEEE802) */
+
+ return 0;
+}
+
+static int add_veth(
+ sd_netlink *rtnl,
+ pid_t pid,
+ const char *ifname_host,
+ const struct ether_addr *mac_host,
+ const char *ifname_container,
+ const struct ether_addr *mac_container) {
+
+ _cleanup_(sd_netlink_message_unrefp) sd_netlink_message *m = NULL;
+ int r;
+
+ assert(rtnl);
+ assert(ifname_host);
+ assert(mac_host);
+ assert(ifname_container);
+ assert(mac_container);
+
+ r = sd_rtnl_message_new_link(rtnl, &m, RTM_NEWLINK, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to allocate netlink message: %m");
+
+ r = sd_netlink_message_append_string(m, IFLA_IFNAME, ifname_host);
+ if (r < 0)
+ return log_error_errno(r, "Failed to add netlink interface name: %m");
+
+ r = sd_netlink_message_append_ether_addr(m, IFLA_ADDRESS, mac_host);
+ if (r < 0)
+ return log_error_errno(r, "Failed to add netlink MAC address: %m");
+
+ r = sd_netlink_message_open_container(m, IFLA_LINKINFO);
+ if (r < 0)
+ return log_error_errno(r, "Failed to open netlink container: %m");
+
+ r = sd_netlink_message_open_container_union(m, IFLA_INFO_DATA, "veth");
+ if (r < 0)
+ return log_error_errno(r, "Failed to open netlink container: %m");
+
+ r = sd_netlink_message_open_container(m, VETH_INFO_PEER);
+ if (r < 0)
+ return log_error_errno(r, "Failed to open netlink container: %m");
+
+ r = sd_netlink_message_append_string(m, IFLA_IFNAME, ifname_container);
+ if (r < 0)
+ return log_error_errno(r, "Failed to add netlink interface name: %m");
+
+ r = sd_netlink_message_append_ether_addr(m, IFLA_ADDRESS, mac_container);
+ if (r < 0)
+ return log_error_errno(r, "Failed to add netlink MAC address: %m");
+
+ r = sd_netlink_message_append_u32(m, IFLA_NET_NS_PID, pid);
+ if (r < 0)
+ return log_error_errno(r, "Failed to add netlink namespace field: %m");
+
+ r = sd_netlink_message_close_container(m);
+ if (r < 0)
+ return log_error_errno(r, "Failed to close netlink container: %m");
+
+ r = sd_netlink_message_close_container(m);
+ if (r < 0)
+ return log_error_errno(r, "Failed to close netlink container: %m");
+
+ r = sd_netlink_message_close_container(m);
+ if (r < 0)
+ return log_error_errno(r, "Failed to close netlink container: %m");
+
+ r = sd_netlink_call(rtnl, m, 0, NULL);
+ if (r < 0)
+ return log_error_errno(r, "Failed to add new veth interfaces (%s:%s): %m", ifname_host, ifname_container);
+
+ return 0;
+}
+
+int setup_veth(const char *machine_name,
+ pid_t pid,
+ char iface_name[IFNAMSIZ],
+ bool bridge) {
+
+ _cleanup_(sd_netlink_unrefp) sd_netlink *rtnl = NULL;
+ struct ether_addr mac_host, mac_container;
+ int r, i;
+
+ assert(machine_name);
+ assert(pid > 0);
+ assert(iface_name);
+
+ /* Use two different interface name prefixes depending whether
+ * we are in bridge mode or not. */
+ snprintf(iface_name, IFNAMSIZ - 1, "%s-%s",
+ bridge ? "vb" : "ve", machine_name);
+
+ r = generate_mac(machine_name, &mac_container, CONTAINER_HASH_KEY, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to generate predictable MAC address for container side: %m");
+
+ r = generate_mac(machine_name, &mac_host, HOST_HASH_KEY, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to generate predictable MAC address for host side: %m");
+
+ r = sd_netlink_open(&rtnl);
+ if (r < 0)
+ return log_error_errno(r, "Failed to connect to netlink: %m");
+
+ r = add_veth(rtnl, pid, iface_name, &mac_host, "host0", &mac_container);
+ if (r < 0)
+ return r;
+
+ i = (int) if_nametoindex(iface_name);
+ if (i <= 0)
+ return log_error_errno(errno, "Failed to resolve interface %s: %m", iface_name);
+
+ return i;
+}
+
+int setup_veth_extra(
+ const char *machine_name,
+ pid_t pid,
+ char **pairs) {
+
+ _cleanup_(sd_netlink_unrefp) sd_netlink *rtnl = NULL;
+ uint64_t idx = 0;
+ char **a, **b;
+ int r;
+
+ assert(machine_name);
+ assert(pid > 0);
+
+ if (strv_isempty(pairs))
+ return 0;
+
+ r = sd_netlink_open(&rtnl);
+ if (r < 0)
+ return log_error_errno(r, "Failed to connect to netlink: %m");
+
+ STRV_FOREACH_PAIR(a, b, pairs) {
+ struct ether_addr mac_host, mac_container;
+
+ r = generate_mac(machine_name, &mac_container, VETH_EXTRA_CONTAINER_HASH_KEY, idx);
+ if (r < 0)
+ return log_error_errno(r, "Failed to generate predictable MAC address for container side of extra veth link: %m");
+
+ r = generate_mac(machine_name, &mac_host, VETH_EXTRA_HOST_HASH_KEY, idx);
+ if (r < 0)
+ return log_error_errno(r, "Failed to generate predictable MAC address for container side of extra veth link: %m");
+
+ r = add_veth(rtnl, pid, *a, &mac_host, *b, &mac_container);
+ if (r < 0)
+ return r;
+
+ idx++;
+ }
+
+ return 0;
+}
+
+static int join_bridge(sd_netlink *rtnl, const char *veth_name, const char *bridge_name) {
+ _cleanup_(sd_netlink_message_unrefp) sd_netlink_message *m = NULL;
+ int r, bridge_ifi;
+
+ assert(rtnl);
+ assert(veth_name);
+ assert(bridge_name);
+
+ bridge_ifi = (int) if_nametoindex(bridge_name);
+ if (bridge_ifi <= 0)
+ return -errno;
+
+ r = sd_rtnl_message_new_link(rtnl, &m, RTM_SETLINK, 0);
+ if (r < 0)
+ return r;
+
+ r = sd_rtnl_message_link_set_flags(m, IFF_UP, IFF_UP);
+ if (r < 0)
+ return r;
+
+ r = sd_netlink_message_append_string(m, IFLA_IFNAME, veth_name);
+ if (r < 0)
+ return r;
+
+ r = sd_netlink_message_append_u32(m, IFLA_MASTER, bridge_ifi);
+ if (r < 0)
+ return r;
+
+ r = sd_netlink_call(rtnl, m, 0, NULL);
+ if (r < 0)
+ return r;
+
+ return bridge_ifi;
+}
+
+static int create_bridge(sd_netlink *rtnl, const char *bridge_name) {
+ _cleanup_(sd_netlink_message_unrefp) sd_netlink_message *m = NULL;
+ int r;
+
+ r = sd_rtnl_message_new_link(rtnl, &m, RTM_NEWLINK, 0);
+ if (r < 0)
+ return r;
+
+ r = sd_netlink_message_append_string(m, IFLA_IFNAME, bridge_name);
+ if (r < 0)
+ return r;
+
+ r = sd_netlink_message_open_container(m, IFLA_LINKINFO);
+ if (r < 0)
+ return r;
+
+ r = sd_netlink_message_open_container_union(m, IFLA_INFO_DATA, "bridge");
+ if (r < 0)
+ return r;
+
+ r = sd_netlink_message_close_container(m);
+ if (r < 0)
+ return r;
+
+ r = sd_netlink_message_close_container(m);
+ if (r < 0)
+ return r;
+
+ r = sd_netlink_call(rtnl, m, 0, NULL);
+ if (r < 0)
+ return r;
+
+ return 0;
+}
+
+int setup_bridge(const char *veth_name, const char *bridge_name, bool create) {
+ _cleanup_release_lock_file_ LockFile bridge_lock = LOCK_FILE_INIT;
+ _cleanup_(sd_netlink_unrefp) sd_netlink *rtnl = NULL;
+ int r, bridge_ifi;
+ unsigned n = 0;
+
+ assert(veth_name);
+ assert(bridge_name);
+
+ r = sd_netlink_open(&rtnl);
+ if (r < 0)
+ return log_error_errno(r, "Failed to connect to netlink: %m");
+
+ if (create) {
+ /* We take a system-wide lock here, so that we can safely check whether there's still a member in the
+ * bridge before removing it, without risking interference from other nspawn instances. */
+
+ r = make_lock_file("/run/systemd/nspawn-network-zone", LOCK_EX, &bridge_lock);
+ if (r < 0)
+ return log_error_errno(r, "Failed to take network zone lock: %m");
+ }
+
+ for (;;) {
+ bridge_ifi = join_bridge(rtnl, veth_name, bridge_name);
+ if (bridge_ifi >= 0)
+ return bridge_ifi;
+ if (bridge_ifi != -ENODEV || !create || n > 10)
+ return log_error_errno(bridge_ifi, "Failed to add interface %s to bridge %s: %m", veth_name, bridge_name);
+
+ /* Count attempts, so that we don't enter an endless loop here. */
+ n++;
+
+ /* The bridge doesn't exist yet. Let's create it */
+ r = create_bridge(rtnl, bridge_name);
+ if (r < 0)
+ return log_error_errno(r, "Failed to create bridge interface %s: %m", bridge_name);
+
+ /* Try again, now that the bridge exists */
+ }
+}
+
+int remove_bridge(const char *bridge_name) {
+ _cleanup_release_lock_file_ LockFile bridge_lock = LOCK_FILE_INIT;
+ _cleanup_(sd_netlink_unrefp) sd_netlink *rtnl = NULL;
+ const char *path;
+ int r;
+
+ /* Removes the specified bridge, but only if it is currently empty */
+
+ if (isempty(bridge_name))
+ return 0;
+
+ r = make_lock_file("/run/systemd/nspawn-network-zone", LOCK_EX, &bridge_lock);
+ if (r < 0)
+ return log_error_errno(r, "Failed to take network zone lock: %m");
+
+ path = strjoina("/sys/class/net/", bridge_name, "/brif");
+
+ r = dir_is_empty(path);
+ if (r == -ENOENT) /* Already gone? */
+ return 0;
+ if (r < 0)
+ return log_error_errno(r, "Can't detect if bridge %s is empty: %m", bridge_name);
+ if (r == 0) /* Still populated, leave it around */
+ return 0;
+
+ r = sd_netlink_open(&rtnl);
+ if (r < 0)
+ return log_error_errno(r, "Failed to connect to netlink: %m");
+
+ return remove_one_link(rtnl, bridge_name);
+}
+
+static int parse_interface(struct udev *udev, const char *name) {
+ _cleanup_udev_device_unref_ struct udev_device *d = NULL;
+ char ifi_str[2 + DECIMAL_STR_MAX(int)];
+ int ifi;
+
+ ifi = (int) if_nametoindex(name);
+ if (ifi <= 0)
+ return log_error_errno(errno, "Failed to resolve interface %s: %m", name);
+
+ sprintf(ifi_str, "n%i", ifi);
+ d = udev_device_new_from_device_id(udev, ifi_str);
+ if (!d)
+ return log_error_errno(errno, "Failed to get udev device for interface %s: %m", name);
+
+ if (udev_device_get_is_initialized(d) <= 0) {
+ log_error("Network interface %s is not initialized yet.", name);
+ return -EBUSY;
+ }
+
+ return ifi;
+}
+
+int move_network_interfaces(pid_t pid, char **ifaces) {
+ _cleanup_udev_unref_ struct udev *udev = NULL;
+ _cleanup_(sd_netlink_unrefp) sd_netlink *rtnl = NULL;
+ char **i;
+ int r;
+
+ if (strv_isempty(ifaces))
+ return 0;
+
+ r = sd_netlink_open(&rtnl);
+ if (r < 0)
+ return log_error_errno(r, "Failed to connect to netlink: %m");
+
+ udev = udev_new();
+ if (!udev) {
+ log_error("Failed to connect to udev.");
+ return -ENOMEM;
+ }
+
+ STRV_FOREACH(i, ifaces) {
+ _cleanup_(sd_netlink_message_unrefp) sd_netlink_message *m = NULL;
+ int ifi;
+
+ ifi = parse_interface(udev, *i);
+ if (ifi < 0)
+ return ifi;
+
+ r = sd_rtnl_message_new_link(rtnl, &m, RTM_SETLINK, ifi);
+ if (r < 0)
+ return log_error_errno(r, "Failed to allocate netlink message: %m");
+
+ r = sd_netlink_message_append_u32(m, IFLA_NET_NS_PID, pid);
+ if (r < 0)
+ return log_error_errno(r, "Failed to append namespace PID to netlink message: %m");
+
+ r = sd_netlink_call(rtnl, m, 0, NULL);
+ if (r < 0)
+ return log_error_errno(r, "Failed to move interface %s to namespace: %m", *i);
+ }
+
+ return 0;
+}
+
+int setup_macvlan(const char *machine_name, pid_t pid, char **ifaces) {
+ _cleanup_udev_unref_ struct udev *udev = NULL;
+ _cleanup_(sd_netlink_unrefp) sd_netlink *rtnl = NULL;
+ unsigned idx = 0;
+ char **i;
+ int r;
+
+ if (strv_isempty(ifaces))
+ return 0;
+
+ r = sd_netlink_open(&rtnl);
+ if (r < 0)
+ return log_error_errno(r, "Failed to connect to netlink: %m");
+
+ udev = udev_new();
+ if (!udev) {
+ log_error("Failed to connect to udev.");
+ return -ENOMEM;
+ }
+
+ STRV_FOREACH(i, ifaces) {
+ _cleanup_(sd_netlink_message_unrefp) sd_netlink_message *m = NULL;
+ _cleanup_free_ char *n = NULL;
+ struct ether_addr mac;
+ int ifi;
+
+ ifi = parse_interface(udev, *i);
+ if (ifi < 0)
+ return ifi;
+
+ r = generate_mac(machine_name, &mac, MACVLAN_HASH_KEY, idx++);
+ if (r < 0)
+ return log_error_errno(r, "Failed to create MACVLAN MAC address: %m");
+
+ r = sd_rtnl_message_new_link(rtnl, &m, RTM_NEWLINK, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to allocate netlink message: %m");
+
+ r = sd_netlink_message_append_u32(m, IFLA_LINK, ifi);
+ if (r < 0)
+ return log_error_errno(r, "Failed to add netlink interface index: %m");
+
+ n = strappend("mv-", *i);
+ if (!n)
+ return log_oom();
+
+ strshorten(n, IFNAMSIZ-1);
+
+ r = sd_netlink_message_append_string(m, IFLA_IFNAME, n);
+ if (r < 0)
+ return log_error_errno(r, "Failed to add netlink interface name: %m");
+
+ r = sd_netlink_message_append_ether_addr(m, IFLA_ADDRESS, &mac);
+ if (r < 0)
+ return log_error_errno(r, "Failed to add netlink MAC address: %m");
+
+ r = sd_netlink_message_append_u32(m, IFLA_NET_NS_PID, pid);
+ if (r < 0)
+ return log_error_errno(r, "Failed to add netlink namespace field: %m");
+
+ r = sd_netlink_message_open_container(m, IFLA_LINKINFO);
+ if (r < 0)
+ return log_error_errno(r, "Failed to open netlink container: %m");
+
+ r = sd_netlink_message_open_container_union(m, IFLA_INFO_DATA, "macvlan");
+ if (r < 0)
+ return log_error_errno(r, "Failed to open netlink container: %m");
+
+ r = sd_netlink_message_append_u32(m, IFLA_MACVLAN_MODE, MACVLAN_MODE_BRIDGE);
+ if (r < 0)
+ return log_error_errno(r, "Failed to append macvlan mode: %m");
+
+ r = sd_netlink_message_close_container(m);
+ if (r < 0)
+ return log_error_errno(r, "Failed to close netlink container: %m");
+
+ r = sd_netlink_message_close_container(m);
+ if (r < 0)
+ return log_error_errno(r, "Failed to close netlink container: %m");
+
+ r = sd_netlink_call(rtnl, m, 0, NULL);
+ if (r < 0)
+ return log_error_errno(r, "Failed to add new macvlan interfaces: %m");
+ }
+
+ return 0;
+}
+
+int setup_ipvlan(const char *machine_name, pid_t pid, char **ifaces) {
+ _cleanup_udev_unref_ struct udev *udev = NULL;
+ _cleanup_(sd_netlink_unrefp) sd_netlink *rtnl = NULL;
+ char **i;
+ int r;
+
+ if (strv_isempty(ifaces))
+ return 0;
+
+ r = sd_netlink_open(&rtnl);
+ if (r < 0)
+ return log_error_errno(r, "Failed to connect to netlink: %m");
+
+ udev = udev_new();
+ if (!udev) {
+ log_error("Failed to connect to udev.");
+ return -ENOMEM;
+ }
+
+ STRV_FOREACH(i, ifaces) {
+ _cleanup_(sd_netlink_message_unrefp) sd_netlink_message *m = NULL;
+ _cleanup_free_ char *n = NULL;
+ int ifi;
+
+ ifi = parse_interface(udev, *i);
+ if (ifi < 0)
+ return ifi;
+
+ r = sd_rtnl_message_new_link(rtnl, &m, RTM_NEWLINK, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to allocate netlink message: %m");
+
+ r = sd_netlink_message_append_u32(m, IFLA_LINK, ifi);
+ if (r < 0)
+ return log_error_errno(r, "Failed to add netlink interface index: %m");
+
+ n = strappend("iv-", *i);
+ if (!n)
+ return log_oom();
+
+ strshorten(n, IFNAMSIZ-1);
+
+ r = sd_netlink_message_append_string(m, IFLA_IFNAME, n);
+ if (r < 0)
+ return log_error_errno(r, "Failed to add netlink interface name: %m");
+
+ r = sd_netlink_message_append_u32(m, IFLA_NET_NS_PID, pid);
+ if (r < 0)
+ return log_error_errno(r, "Failed to add netlink namespace field: %m");
+
+ r = sd_netlink_message_open_container(m, IFLA_LINKINFO);
+ if (r < 0)
+ return log_error_errno(r, "Failed to open netlink container: %m");
+
+ r = sd_netlink_message_open_container_union(m, IFLA_INFO_DATA, "ipvlan");
+ if (r < 0)
+ return log_error_errno(r, "Failed to open netlink container: %m");
+
+ r = sd_netlink_message_append_u16(m, IFLA_IPVLAN_MODE, IPVLAN_MODE_L2);
+ if (r < 0)
+ return log_error_errno(r, "Failed to add ipvlan mode: %m");
+
+ r = sd_netlink_message_close_container(m);
+ if (r < 0)
+ return log_error_errno(r, "Failed to close netlink container: %m");
+
+ r = sd_netlink_message_close_container(m);
+ if (r < 0)
+ return log_error_errno(r, "Failed to close netlink container: %m");
+
+ r = sd_netlink_call(rtnl, m, 0, NULL);
+ if (r < 0)
+ return log_error_errno(r, "Failed to add new ipvlan interfaces: %m");
+ }
+
+ return 0;
+}
+
+int veth_extra_parse(char ***l, const char *p) {
+ _cleanup_free_ char *a = NULL, *b = NULL;
+ int r;
+
+ r = extract_first_word(&p, &a, ":", EXTRACT_DONT_COALESCE_SEPARATORS);
+ if (r < 0)
+ return r;
+ if (r == 0 || !ifname_valid(a))
+ return -EINVAL;
+
+ r = extract_first_word(&p, &b, ":", EXTRACT_DONT_COALESCE_SEPARATORS);
+ if (r < 0)
+ return r;
+ if (r == 0 || !ifname_valid(b)) {
+ free(b);
+ b = strdup(a);
+ if (!b)
+ return -ENOMEM;
+ }
+
+ if (p)
+ return -EINVAL;
+
+ r = strv_push_pair(l, a, b);
+ if (r < 0)
+ return -ENOMEM;
+
+ a = b = NULL;
+ return 0;
+}
+
+int remove_veth_links(const char *primary, char **pairs) {
+ _cleanup_(sd_netlink_unrefp) sd_netlink *rtnl = NULL;
+ char **a, **b;
+ int r;
+
+ /* In some cases the kernel might pin the veth links between host and container even after the namespace
+ * died. Hence, let's better remove them explicitly too. */
+
+ if (isempty(primary) && strv_isempty(pairs))
+ return 0;
+
+ r = sd_netlink_open(&rtnl);
+ if (r < 0)
+ return log_error_errno(r, "Failed to connect to netlink: %m");
+
+ remove_one_link(rtnl, primary);
+
+ STRV_FOREACH_PAIR(a, b, pairs)
+ remove_one_link(rtnl, *a);
+
+ return 0;
+}
diff --git a/src/systemd-nspawn/nspawn-network.h b/src/systemd-nspawn/nspawn-network.h
new file mode 100644
index 0000000000..3d8861e1e5
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-network.h
@@ -0,0 +1,39 @@
+#pragma once
+
+/***
+ This file is part of systemd.
+
+ Copyright 2015 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include <net/if.h>
+#include <stdbool.h>
+#include <sys/types.h>
+
+int setup_veth(const char *machine_name, pid_t pid, char iface_name[IFNAMSIZ], bool bridge);
+int setup_veth_extra(const char *machine_name, pid_t pid, char **pairs);
+
+int setup_bridge(const char *veth_name, const char *bridge_name, bool create);
+int remove_bridge(const char *bridge_name);
+
+int setup_macvlan(const char *machine_name, pid_t pid, char **ifaces);
+int setup_ipvlan(const char *machine_name, pid_t pid, char **ifaces);
+
+int move_network_interfaces(pid_t pid, char **ifaces);
+
+int veth_extra_parse(char ***l, const char *p);
+
+int remove_veth_links(const char *primary, char **pairs);
diff --git a/src/systemd-nspawn/nspawn-patch-uid.c b/src/systemd-nspawn/nspawn-patch-uid.c
new file mode 100644
index 0000000000..ef297ed915
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-patch-uid.c
@@ -0,0 +1,483 @@
+/***
+ This file is part of systemd.
+
+ Copyright 2016 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include <fcntl.h>
+
+#include <linux/magic.h>
+#ifdef HAVE_ACL
+#include <sys/acl.h>
+#endif
+#include <sys/stat.h>
+#include <sys/vfs.h>
+#include <unistd.h>
+
+#include "systemd-basic/dirent-util.h"
+#include "systemd-basic/fd-util.h"
+#include "systemd-basic/missing.h"
+#include "systemd-basic/stat-util.h"
+#include "systemd-basic/stdio-util.h"
+#include "systemd-basic/string-util.h"
+#include "systemd-basic/strv.h"
+#include "systemd-basic/user-util.h"
+#include "systemd-shared/acl-util.h"
+
+#include "nspawn-patch-uid.h"
+
+#ifdef HAVE_ACL
+
+static int get_acl(int fd, const char *name, acl_type_t type, acl_t *ret) {
+ char procfs_path[strlen("/proc/self/fd/") + DECIMAL_STR_MAX(int) + 1];
+ acl_t acl;
+
+ assert(fd >= 0);
+ assert(ret);
+
+ if (name) {
+ _cleanup_close_ int child_fd = -1;
+
+ child_fd = openat(fd, name, O_PATH|O_CLOEXEC|O_NOFOLLOW);
+ if (child_fd < 0)
+ return -errno;
+
+ xsprintf(procfs_path, "/proc/self/fd/%i", child_fd);
+ acl = acl_get_file(procfs_path, type);
+ } else if (type == ACL_TYPE_ACCESS)
+ acl = acl_get_fd(fd);
+ else {
+ xsprintf(procfs_path, "/proc/self/fd/%i", fd);
+ acl = acl_get_file(procfs_path, type);
+ }
+ if (!acl)
+ return -errno;
+
+ *ret = acl;
+ return 0;
+}
+
+static int set_acl(int fd, const char *name, acl_type_t type, acl_t acl) {
+ char procfs_path[strlen("/proc/self/fd/") + DECIMAL_STR_MAX(int) + 1];
+ int r;
+
+ assert(fd >= 0);
+ assert(acl);
+
+ if (name) {
+ _cleanup_close_ int child_fd = -1;
+
+ child_fd = openat(fd, name, O_PATH|O_CLOEXEC|O_NOFOLLOW);
+ if (child_fd < 0)
+ return -errno;
+
+ xsprintf(procfs_path, "/proc/self/fd/%i", child_fd);
+ r = acl_set_file(procfs_path, type, acl);
+ } else if (type == ACL_TYPE_ACCESS)
+ r = acl_set_fd(fd, acl);
+ else {
+ xsprintf(procfs_path, "/proc/self/fd/%i", fd);
+ r = acl_set_file(procfs_path, type, acl);
+ }
+ if (r < 0)
+ return -errno;
+
+ return 0;
+}
+
+static int shift_acl(acl_t acl, uid_t shift, acl_t *ret) {
+ _cleanup_(acl_freep) acl_t copy = NULL;
+ acl_entry_t i;
+ int r;
+
+ assert(acl);
+ assert(ret);
+
+ r = acl_get_entry(acl, ACL_FIRST_ENTRY, &i);
+ if (r < 0)
+ return -errno;
+ while (r > 0) {
+ uid_t *old_uid, new_uid;
+ bool modify = false;
+ acl_tag_t tag;
+
+ if (acl_get_tag_type(i, &tag) < 0)
+ return -errno;
+
+ if (IN_SET(tag, ACL_USER, ACL_GROUP)) {
+
+ /* We don't distuingish here between uid_t and gid_t, let's make sure the compiler checks that
+ * this is actually OK */
+ assert_cc(sizeof(uid_t) == sizeof(gid_t));
+
+ old_uid = acl_get_qualifier(i);
+ if (!old_uid)
+ return -errno;
+
+ new_uid = shift | (*old_uid & UINT32_C(0xFFFF));
+ if (!uid_is_valid(new_uid))
+ return -EINVAL;
+
+ modify = new_uid != *old_uid;
+ if (modify && !copy) {
+ int n;
+
+ /* There's no copy of the ACL yet? if so, let's create one, and start the loop from the
+ * beginning, so that we copy all entries, starting from the first, this time. */
+
+ n = acl_entries(acl);
+ if (n < 0)
+ return -errno;
+
+ copy = acl_init(n);
+ if (!copy)
+ return -errno;
+
+ /* Seek back to the beginning */
+ r = acl_get_entry(acl, ACL_FIRST_ENTRY, &i);
+ if (r < 0)
+ return -errno;
+ continue;
+ }
+ }
+
+ if (copy) {
+ acl_entry_t new_entry;
+
+ if (acl_create_entry(&copy, &new_entry) < 0)
+ return -errno;
+
+ if (acl_copy_entry(new_entry, i) < 0)
+ return -errno;
+
+ if (modify)
+ if (acl_set_qualifier(new_entry, &new_uid) < 0)
+ return -errno;
+ }
+
+ r = acl_get_entry(acl, ACL_NEXT_ENTRY, &i);
+ if (r < 0)
+ return -errno;
+ }
+
+ *ret = copy;
+ copy = NULL;
+
+ return !!*ret;
+}
+
+static int patch_acls(int fd, const char *name, const struct stat *st, uid_t shift) {
+ _cleanup_(acl_freep) acl_t acl = NULL, shifted = NULL;
+ bool changed = false;
+ int r;
+
+ assert(fd >= 0);
+ assert(st);
+
+ /* ACLs are not supported on symlinks, there's no point in trying */
+ if (S_ISLNK(st->st_mode))
+ return 0;
+
+ r = get_acl(fd, name, ACL_TYPE_ACCESS, &acl);
+ if (r == -EOPNOTSUPP)
+ return 0;
+ if (r < 0)
+ return r;
+
+ r = shift_acl(acl, shift, &shifted);
+ if (r < 0)
+ return r;
+ if (r > 0) {
+ r = set_acl(fd, name, ACL_TYPE_ACCESS, shifted);
+ if (r < 0)
+ return r;
+
+ changed = true;
+ }
+
+ if (S_ISDIR(st->st_mode)) {
+ acl_free(acl);
+ acl_free(shifted);
+
+ acl = shifted = NULL;
+
+ r = get_acl(fd, name, ACL_TYPE_DEFAULT, &acl);
+ if (r < 0)
+ return r;
+
+ r = shift_acl(acl, shift, &shifted);
+ if (r < 0)
+ return r;
+ if (r > 0) {
+ r = set_acl(fd, name, ACL_TYPE_DEFAULT, shifted);
+ if (r < 0)
+ return r;
+
+ changed = true;
+ }
+ }
+
+ return changed;
+}
+
+#else
+
+static int patch_acls(int fd, const char *name, const struct stat *st, uid_t shift) {
+ return 0;
+}
+
+#endif
+
+static int patch_fd(int fd, const char *name, const struct stat *st, uid_t shift) {
+ uid_t new_uid;
+ gid_t new_gid;
+ bool changed = false;
+ int r;
+
+ assert(fd >= 0);
+ assert(st);
+
+ new_uid = shift | (st->st_uid & UINT32_C(0xFFFF));
+ new_gid = (gid_t) shift | (st->st_gid & UINT32_C(0xFFFF));
+
+ if (!uid_is_valid(new_uid) || !gid_is_valid(new_gid))
+ return -EINVAL;
+
+ if (st->st_uid != new_uid || st->st_gid != new_gid) {
+ if (name)
+ r = fchownat(fd, name, new_uid, new_gid, AT_SYMLINK_NOFOLLOW);
+ else
+ r = fchown(fd, new_uid, new_gid);
+ if (r < 0)
+ return -errno;
+
+ /* The Linux kernel alters the mode in some cases of chown(). Let's undo this. */
+ if (name) {
+ if (!S_ISLNK(st->st_mode))
+ r = fchmodat(fd, name, st->st_mode, 0);
+ else /* AT_SYMLINK_NOFOLLOW is not available for fchmodat() */
+ r = 0;
+ } else
+ r = fchmod(fd, st->st_mode);
+ if (r < 0)
+ return -errno;
+
+ changed = true;
+ }
+
+ r = patch_acls(fd, name, st, shift);
+ if (r < 0)
+ return r;
+
+ return r > 0 || changed;
+}
+
+/*
+ * Check if the filesystem is fully compatible with user namespaces or
+ * UID/GID patching. Some filesystems in this list can be fully mounted inside
+ * user namespaces, however their inodes may relate to host resources or only
+ * valid in the global user namespace, therefore no patching should be applied.
+ */
+static int is_fs_fully_userns_compatible(int fd) {
+ struct statfs sfs;
+
+ assert(fd >= 0);
+
+ if (fstatfs(fd, &sfs) < 0)
+ return -errno;
+
+ return F_TYPE_EQUAL(sfs.f_type, BINFMTFS_MAGIC) ||
+ F_TYPE_EQUAL(sfs.f_type, CGROUP_SUPER_MAGIC) ||
+ F_TYPE_EQUAL(sfs.f_type, CGROUP2_SUPER_MAGIC) ||
+ F_TYPE_EQUAL(sfs.f_type, DEBUGFS_MAGIC) ||
+ F_TYPE_EQUAL(sfs.f_type, DEVPTS_SUPER_MAGIC) ||
+ F_TYPE_EQUAL(sfs.f_type, EFIVARFS_MAGIC) ||
+ F_TYPE_EQUAL(sfs.f_type, HUGETLBFS_MAGIC) ||
+ F_TYPE_EQUAL(sfs.f_type, MQUEUE_MAGIC) ||
+ F_TYPE_EQUAL(sfs.f_type, PROC_SUPER_MAGIC) ||
+ F_TYPE_EQUAL(sfs.f_type, PSTOREFS_MAGIC) ||
+ F_TYPE_EQUAL(sfs.f_type, SELINUX_MAGIC) ||
+ F_TYPE_EQUAL(sfs.f_type, SMACK_MAGIC) ||
+ F_TYPE_EQUAL(sfs.f_type, SECURITYFS_MAGIC) ||
+ F_TYPE_EQUAL(sfs.f_type, BPF_FS_MAGIC) ||
+ F_TYPE_EQUAL(sfs.f_type, TRACEFS_MAGIC) ||
+ F_TYPE_EQUAL(sfs.f_type, SYSFS_MAGIC);
+}
+
+static int recurse_fd(int fd, bool donate_fd, const struct stat *st, uid_t shift, bool is_toplevel) {
+ bool changed = false;
+ int r;
+
+ assert(fd >= 0);
+
+ /* We generally want to permit crossing of mount boundaries when patching the UIDs/GIDs. However, we
+ * probably shouldn't do this for /proc and /sys if that is already mounted into place. Hence, let's
+ * stop the recursion when we hit procfs, sysfs or some other special file systems. */
+ r = is_fs_fully_userns_compatible(fd);
+ if (r < 0)
+ goto finish;
+ if (r > 0) {
+ r = 0; /* don't recurse */
+ goto finish;
+ }
+
+ r = patch_fd(fd, NULL, st, shift);
+ if (r == -EROFS) {
+ _cleanup_free_ char *name = NULL;
+
+ if (!is_toplevel) {
+ /* When we hit a ready-only subtree we simply skip it, but log about it. */
+ (void) fd_get_path(fd, &name);
+ log_debug("Skippping read-only file or directory %s.", strna(name));
+ r = 0;
+ }
+
+ goto finish;
+ }
+ if (r < 0)
+ goto finish;
+
+ if (S_ISDIR(st->st_mode)) {
+ _cleanup_closedir_ DIR *d = NULL;
+ struct dirent *de;
+
+ if (!donate_fd) {
+ int copy;
+
+ copy = fcntl(fd, F_DUPFD_CLOEXEC, 3);
+ if (copy < 0) {
+ r = -errno;
+ goto finish;
+ }
+
+ fd = copy;
+ donate_fd = true;
+ }
+
+ d = fdopendir(fd);
+ if (!d) {
+ r = -errno;
+ goto finish;
+ }
+ fd = -1;
+
+ FOREACH_DIRENT_ALL(de, d, r = -errno; goto finish) {
+ struct stat fst;
+
+ if (STR_IN_SET(de->d_name, ".", ".."))
+ continue;
+
+ if (fstatat(dirfd(d), de->d_name, &fst, AT_SYMLINK_NOFOLLOW) < 0) {
+ r = -errno;
+ goto finish;
+ }
+
+ if (S_ISDIR(fst.st_mode)) {
+ int subdir_fd;
+
+ subdir_fd = openat(dirfd(d), de->d_name, O_RDONLY|O_NONBLOCK|O_DIRECTORY|O_CLOEXEC|O_NOFOLLOW|O_NOATIME);
+ if (subdir_fd < 0) {
+ r = -errno;
+ goto finish;
+
+ }
+
+ r = recurse_fd(subdir_fd, true, &fst, shift, false);
+ if (r < 0)
+ goto finish;
+ if (r > 0)
+ changed = true;
+
+ } else {
+ r = patch_fd(dirfd(d), de->d_name, &fst, shift);
+ if (r < 0)
+ goto finish;
+ if (r > 0)
+ changed = true;
+ }
+ }
+ }
+
+ r = changed;
+
+finish:
+ if (donate_fd)
+ safe_close(fd);
+
+ return r;
+}
+
+static int fd_patch_uid_internal(int fd, bool donate_fd, uid_t shift, uid_t range) {
+ struct stat st;
+ int r;
+
+ assert(fd >= 0);
+
+ /* Recursively adjusts the UID/GIDs of all files of a directory tree. This is used to automatically fix up an
+ * OS tree to the used user namespace UID range. Note that this automatic adjustment only works for UID ranges
+ * following the concept that the upper 16bit of a UID identify the container, and the lower 16bit are the actual
+ * UID within the container. */
+
+ if ((shift & 0xFFFF) != 0) {
+ /* We only support containers where the shift starts at a 2^16 boundary */
+ r = -EOPNOTSUPP;
+ goto finish;
+ }
+
+ if (range != 0x10000) {
+ /* We only support containers with 16bit UID ranges for the patching logic */
+ r = -EOPNOTSUPP;
+ goto finish;
+ }
+
+ if (fstat(fd, &st) < 0) {
+ r = -errno;
+ goto finish;
+ }
+
+ if ((uint32_t) st.st_uid >> 16 != (uint32_t) st.st_gid >> 16) {
+ /* We only support containers where the uid/gid container ID match */
+ r = -EBADE;
+ goto finish;
+ }
+
+ /* Try to detect if the range is already right. Of course, this a pretty drastic optimization, as we assume
+ * that if the top-level dir has the right upper 16bit assigned, then everything below will have too... */
+ if (((uint32_t) (st.st_uid ^ shift) >> 16) == 0)
+ return 0;
+
+ return recurse_fd(fd, donate_fd, &st, shift, true);
+
+finish:
+ if (donate_fd)
+ safe_close(fd);
+
+ return r;
+}
+
+int fd_patch_uid(int fd, uid_t shift, uid_t range) {
+ return fd_patch_uid_internal(fd, false, shift, range);
+}
+
+int path_patch_uid(const char *path, uid_t shift, uid_t range) {
+ int fd;
+
+ fd = open(path, O_RDONLY|O_NONBLOCK|O_DIRECTORY|O_CLOEXEC|O_NOFOLLOW|O_NOATIME);
+ if (fd < 0)
+ return -errno;
+
+ return fd_patch_uid_internal(fd, true, shift, range);
+}
diff --git a/src/systemd-nspawn/nspawn-patch-uid.h b/src/systemd-nspawn/nspawn-patch-uid.h
new file mode 100644
index 0000000000..55d0990016
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-patch-uid.h
@@ -0,0 +1,23 @@
+/***
+ This file is part of systemd.
+
+ Copyright 2016 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include <sys/types.h>
+
+int fd_patch_uid(int fd, uid_t shift, uid_t range);
+int path_patch_uid(const char *path, uid_t shift, uid_t range);
diff --git a/src/systemd-nspawn/nspawn-register.c b/src/systemd-nspawn/nspawn-register.c
new file mode 100644
index 0000000000..de3ee23f09
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-register.c
@@ -0,0 +1,227 @@
+/***
+ This file is part of systemd.
+
+ Copyright 2015 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include <systemd/sd-bus.h>
+
+#include "sd-bus/bus-error.h"
+#include "sd-bus/bus-util.h"
+#include "systemd-basic/stat-util.h"
+#include "systemd-basic/strv.h"
+#include "systemd-basic/util.h"
+#include "systemd-shared/bus-unit-util.h"
+
+#include "nspawn-register.h"
+
+int register_machine(
+ const char *machine_name,
+ pid_t pid,
+ const char *directory,
+ sd_id128_t uuid,
+ int local_ifindex,
+ const char *slice,
+ CustomMount *mounts,
+ unsigned n_mounts,
+ int kill_signal,
+ char **properties,
+ bool keep_unit,
+ const char *service) {
+
+ _cleanup_(sd_bus_error_free) sd_bus_error error = SD_BUS_ERROR_NULL;
+ _cleanup_(sd_bus_flush_close_unrefp) sd_bus *bus = NULL;
+ int r;
+
+ r = sd_bus_default_system(&bus);
+ if (r < 0)
+ return log_error_errno(r, "Failed to open system bus: %m");
+
+ if (keep_unit) {
+ r = sd_bus_call_method(
+ bus,
+ "org.freedesktop.machine1",
+ "/org/freedesktop/machine1",
+ "org.freedesktop.machine1.Manager",
+ "RegisterMachineWithNetwork",
+ &error,
+ NULL,
+ "sayssusai",
+ machine_name,
+ SD_BUS_MESSAGE_APPEND_ID128(uuid),
+ service,
+ "container",
+ (uint32_t) pid,
+ strempty(directory),
+ local_ifindex > 0 ? 1 : 0, local_ifindex);
+ } else {
+ _cleanup_(sd_bus_message_unrefp) sd_bus_message *m = NULL;
+ unsigned j;
+
+ r = sd_bus_message_new_method_call(
+ bus,
+ &m,
+ "org.freedesktop.machine1",
+ "/org/freedesktop/machine1",
+ "org.freedesktop.machine1.Manager",
+ "CreateMachineWithNetwork");
+ if (r < 0)
+ return bus_log_create_error(r);
+
+ r = sd_bus_message_append(
+ m,
+ "sayssusai",
+ machine_name,
+ SD_BUS_MESSAGE_APPEND_ID128(uuid),
+ service,
+ "container",
+ (uint32_t) pid,
+ strempty(directory),
+ local_ifindex > 0 ? 1 : 0, local_ifindex);
+ if (r < 0)
+ return bus_log_create_error(r);
+
+ r = sd_bus_message_open_container(m, 'a', "(sv)");
+ if (r < 0)
+ return bus_log_create_error(r);
+
+ if (!isempty(slice)) {
+ r = sd_bus_message_append(m, "(sv)", "Slice", "s", slice);
+ if (r < 0)
+ return bus_log_create_error(r);
+ }
+
+ r = sd_bus_message_append(m, "(sv)", "DevicePolicy", "s", "closed");
+ if (r < 0)
+ return bus_log_create_error(r);
+
+ /* If you make changes here, also make sure to update
+ * systemd-nspawn@.service, to keep the device
+ * policies in sync regardless if we are run with or
+ * without the --keep-unit switch. */
+ r = sd_bus_message_append(m, "(sv)", "DeviceAllow", "a(ss)", 2,
+ /* Allow the container to
+ * access and create the API
+ * device nodes, so that
+ * PrivateDevices= in the
+ * container can work
+ * fine */
+ "/dev/net/tun", "rwm",
+ /* Allow the container
+ * access to ptys. However,
+ * do not permit the
+ * container to ever create
+ * these device nodes. */
+ "char-pts", "rw");
+ if (r < 0)
+ return bus_log_create_error(r);
+
+ for (j = 0; j < n_mounts; j++) {
+ CustomMount *cm = mounts + j;
+
+ if (cm->type != CUSTOM_MOUNT_BIND)
+ continue;
+
+ r = is_device_node(cm->source);
+ if (r < 0)
+ return log_error_errno(r, "Failed to stat %s: %m", cm->source);
+
+ if (r) {
+ r = sd_bus_message_append(m, "(sv)", "DeviceAllow", "a(ss)", 1,
+ cm->source, cm->read_only ? "r" : "rw");
+ if (r < 0)
+ return log_error_errno(r, "Failed to append message arguments: %m");
+ }
+ }
+
+ if (kill_signal != 0) {
+ r = sd_bus_message_append(m, "(sv)", "KillSignal", "i", kill_signal);
+ if (r < 0)
+ return bus_log_create_error(r);
+
+ r = sd_bus_message_append(m, "(sv)", "KillMode", "s", "mixed");
+ if (r < 0)
+ return bus_log_create_error(r);
+ }
+
+ r = bus_append_unit_property_assignment_many(m, properties);
+ if (r < 0)
+ return r;
+
+ r = sd_bus_message_close_container(m);
+ if (r < 0)
+ return bus_log_create_error(r);
+
+ r = sd_bus_call(bus, m, 0, &error, NULL);
+ }
+
+ if (r < 0) {
+ log_error("Failed to register machine: %s", bus_error_message(&error, r));
+ return r;
+ }
+
+ return 0;
+}
+
+int terminate_machine(pid_t pid) {
+ _cleanup_(sd_bus_error_free) sd_bus_error error = SD_BUS_ERROR_NULL;
+ _cleanup_(sd_bus_message_unrefp) sd_bus_message *reply = NULL;
+ _cleanup_(sd_bus_flush_close_unrefp) sd_bus *bus = NULL;
+ const char *path;
+ int r;
+
+ r = sd_bus_default_system(&bus);
+ if (r < 0)
+ return log_error_errno(r, "Failed to open system bus: %m");
+
+ r = sd_bus_call_method(
+ bus,
+ "org.freedesktop.machine1",
+ "/org/freedesktop/machine1",
+ "org.freedesktop.machine1.Manager",
+ "GetMachineByPID",
+ &error,
+ &reply,
+ "u",
+ (uint32_t) pid);
+ if (r < 0) {
+ /* Note that the machine might already have been
+ * cleaned up automatically, hence don't consider it a
+ * failure if we cannot get the machine object. */
+ log_debug("Failed to get machine: %s", bus_error_message(&error, r));
+ return 0;
+ }
+
+ r = sd_bus_message_read(reply, "o", &path);
+ if (r < 0)
+ return bus_log_parse_error(r);
+
+ r = sd_bus_call_method(
+ bus,
+ "org.freedesktop.machine1",
+ path,
+ "org.freedesktop.machine1.Machine",
+ "Terminate",
+ &error,
+ NULL,
+ NULL);
+ if (r < 0) {
+ log_debug("Failed to terminate machine: %s", bus_error_message(&error, r));
+ return 0;
+ }
+
+ return 0;
+}
diff --git a/src/systemd-nspawn/nspawn-register.h b/src/systemd-nspawn/nspawn-register.h
new file mode 100644
index 0000000000..c7a50f7477
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-register.h
@@ -0,0 +1,29 @@
+#pragma once
+
+/***
+ This file is part of systemd.
+
+ Copyright 2015 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include <sys/types.h>
+
+#include <systemd/sd-id128.h>
+
+#include "nspawn-mount.h"
+
+int register_machine(const char *machine_name, pid_t pid, const char *directory, sd_id128_t uuid, int local_ifindex, const char *slice, CustomMount *mounts, unsigned n_mounts, int kill_signal, char **properties, bool keep_unit, const char *service);
+int terminate_machine(pid_t pid);
diff --git a/src/systemd-nspawn/nspawn-seccomp.c b/src/systemd-nspawn/nspawn-seccomp.c
new file mode 100644
index 0000000000..e5cfe789a1
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-seccomp.c
@@ -0,0 +1,186 @@
+/***
+ This file is part of systemd.
+
+ Copyright 2016 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include <errno.h>
+#include <sys/capability.h>
+#include <sys/types.h>
+
+#include <linux/netlink.h>
+
+#ifdef HAVE_SECCOMP
+#include <seccomp.h>
+#endif
+
+#include "systemd-basic/log.h"
+
+#ifdef HAVE_SECCOMP
+#include "systemd-shared/seccomp-util.h"
+#endif
+
+#include "nspawn-seccomp.h"
+
+#ifdef HAVE_SECCOMP
+
+static int seccomp_add_default_syscall_filter(scmp_filter_ctx ctx,
+ uint64_t cap_list_retain) {
+ unsigned i;
+ int r;
+ static const struct {
+ uint64_t capability;
+ int syscall_num;
+ } blacklist[] = {
+ { 0, SCMP_SYS(_sysctl) }, /* obsolete syscall */
+ { 0, SCMP_SYS(add_key) }, /* keyring is not namespaced */
+ { 0, SCMP_SYS(afs_syscall) }, /* obsolete syscall */
+ { 0, SCMP_SYS(bdflush) },
+#ifdef __NR_bpf
+ { 0, SCMP_SYS(bpf) },
+#endif
+ { 0, SCMP_SYS(break) }, /* obsolete syscall */
+ { 0, SCMP_SYS(create_module) }, /* obsolete syscall */
+ { 0, SCMP_SYS(ftime) }, /* obsolete syscall */
+ { 0, SCMP_SYS(get_kernel_syms) }, /* obsolete syscall */
+ { 0, SCMP_SYS(getpmsg) }, /* obsolete syscall */
+ { 0, SCMP_SYS(gtty) }, /* obsolete syscall */
+#ifdef __NR_kexec_file_load
+ { 0, SCMP_SYS(kexec_file_load) },
+#endif
+ { 0, SCMP_SYS(kexec_load) },
+ { 0, SCMP_SYS(keyctl) }, /* keyring is not namespaced */
+ { 0, SCMP_SYS(lock) }, /* obsolete syscall */
+ { 0, SCMP_SYS(lookup_dcookie) },
+ { 0, SCMP_SYS(mpx) }, /* obsolete syscall */
+ { 0, SCMP_SYS(nfsservctl) }, /* obsolete syscall */
+ { 0, SCMP_SYS(open_by_handle_at) },
+ { 0, SCMP_SYS(perf_event_open) },
+ { 0, SCMP_SYS(prof) }, /* obsolete syscall */
+ { 0, SCMP_SYS(profil) }, /* obsolete syscall */
+ { 0, SCMP_SYS(putpmsg) }, /* obsolete syscall */
+ { 0, SCMP_SYS(query_module) }, /* obsolete syscall */
+ { 0, SCMP_SYS(quotactl) },
+ { 0, SCMP_SYS(request_key) }, /* keyring is not namespaced */
+ { 0, SCMP_SYS(security) }, /* obsolete syscall */
+ { 0, SCMP_SYS(sgetmask) }, /* obsolete syscall */
+ { 0, SCMP_SYS(ssetmask) }, /* obsolete syscall */
+ { 0, SCMP_SYS(stty) }, /* obsolete syscall */
+ { 0, SCMP_SYS(swapoff) },
+ { 0, SCMP_SYS(swapon) },
+ { 0, SCMP_SYS(sysfs) }, /* obsolete syscall */
+ { 0, SCMP_SYS(tuxcall) }, /* obsolete syscall */
+ { 0, SCMP_SYS(ulimit) }, /* obsolete syscall */
+ { 0, SCMP_SYS(uselib) }, /* obsolete syscall */
+ { 0, SCMP_SYS(ustat) }, /* obsolete syscall */
+ { 0, SCMP_SYS(vserver) }, /* obsolete syscall */
+ { CAP_SYSLOG, SCMP_SYS(syslog) },
+ { CAP_SYS_MODULE, SCMP_SYS(delete_module) },
+ { CAP_SYS_MODULE, SCMP_SYS(finit_module) },
+ { CAP_SYS_MODULE, SCMP_SYS(init_module) },
+ { CAP_SYS_PACCT, SCMP_SYS(acct) },
+ { CAP_SYS_PTRACE, SCMP_SYS(process_vm_readv) },
+ { CAP_SYS_PTRACE, SCMP_SYS(process_vm_writev) },
+ { CAP_SYS_PTRACE, SCMP_SYS(ptrace) },
+ { CAP_SYS_RAWIO, SCMP_SYS(ioperm) },
+ { CAP_SYS_RAWIO, SCMP_SYS(iopl) },
+ { CAP_SYS_RAWIO, SCMP_SYS(pciconfig_iobase) },
+ { CAP_SYS_RAWIO, SCMP_SYS(pciconfig_read) },
+ { CAP_SYS_RAWIO, SCMP_SYS(pciconfig_write) },
+#ifdef __NR_s390_pci_mmio_read
+ { CAP_SYS_RAWIO, SCMP_SYS(s390_pci_mmio_read) },
+#endif
+#ifdef __NR_s390_pci_mmio_write
+ { CAP_SYS_RAWIO, SCMP_SYS(s390_pci_mmio_write) },
+#endif
+ { CAP_SYS_TIME, SCMP_SYS(adjtimex) },
+ { CAP_SYS_TIME, SCMP_SYS(clock_adjtime) },
+ { CAP_SYS_TIME, SCMP_SYS(clock_settime) },
+ { CAP_SYS_TIME, SCMP_SYS(settimeofday) },
+ { CAP_SYS_TIME, SCMP_SYS(stime) },
+ };
+
+ for (i = 0; i < ELEMENTSOF(blacklist); i++) {
+ if (blacklist[i].capability != 0 && (cap_list_retain & (1ULL << blacklist[i].capability)))
+ continue;
+
+ r = seccomp_rule_add(ctx, SCMP_ACT_ERRNO(EPERM), blacklist[i].syscall_num, 0);
+ if (r == -EFAULT)
+ continue; /* unknown syscall */
+ if (r < 0)
+ return log_error_errno(r, "Failed to block syscall: %m");
+ }
+
+ return 0;
+}
+
+int setup_seccomp(uint64_t cap_list_retain) {
+ scmp_filter_ctx seccomp;
+ int r;
+
+ if (!is_seccomp_available()) {
+ log_debug("SECCOMP features not detected in the kernel, disabling SECCOMP audit filter");
+ return 0;
+ }
+
+ r = seccomp_init_conservative(&seccomp, SCMP_ACT_ALLOW);
+ if (r < 0)
+ return log_error_errno(r, "Failed to allocate seccomp object: %m");
+
+ r = seccomp_add_default_syscall_filter(seccomp, cap_list_retain);
+ if (r < 0)
+ goto finish;
+
+ /*
+ Audit is broken in containers, much of the userspace audit
+ hookup will fail if running inside a container. We don't
+ care and just turn off creation of audit sockets.
+
+ This will make socket(AF_NETLINK, *, NETLINK_AUDIT) fail
+ with EAFNOSUPPORT which audit userspace uses as indication
+ that audit is disabled in the kernel.
+ */
+
+ r = seccomp_rule_add(
+ seccomp,
+ SCMP_ACT_ERRNO(EAFNOSUPPORT),
+ SCMP_SYS(socket),
+ 2,
+ SCMP_A0(SCMP_CMP_EQ, AF_NETLINK),
+ SCMP_A2(SCMP_CMP_EQ, NETLINK_AUDIT));
+ if (r < 0) {
+ log_error_errno(r, "Failed to add audit seccomp rule: %m");
+ goto finish;
+ }
+
+ r = seccomp_load(seccomp);
+ if (r < 0) {
+ log_error_errno(r, "Failed to install seccomp audit filter: %m");
+ goto finish;
+ }
+
+finish:
+ seccomp_release(seccomp);
+ return r;
+}
+
+#else
+
+int setup_seccomp(uint64_t cap_list_retain) {
+ return 0;
+}
+
+#endif
diff --git a/src/systemd-nspawn/nspawn-seccomp.h b/src/systemd-nspawn/nspawn-seccomp.h
new file mode 100644
index 0000000000..5bde16faf9
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-seccomp.h
@@ -0,0 +1,24 @@
+#pragma once
+
+/***
+ This file is part of systemd.
+
+ Copyright 2016 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include <sys/types.h>
+
+int setup_seccomp(uint64_t cap_list_retain);
diff --git a/src/systemd-nspawn/nspawn-settings.c b/src/systemd-nspawn/nspawn-settings.c
new file mode 100644
index 0000000000..6c1614b276
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-settings.c
@@ -0,0 +1,515 @@
+/***
+ This file is part of systemd.
+
+ Copyright 2015 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include "systemd-basic/alloc-util.h"
+#include "systemd-basic/cap-list.h"
+#include "systemd-basic/parse-util.h"
+#include "systemd-basic/process-util.h"
+#include "systemd-basic/socket-util.h"
+#include "systemd-basic/string-util.h"
+#include "systemd-basic/strv.h"
+#include "systemd-basic/user-util.h"
+#include "systemd-basic/util.h"
+#include "systemd-shared/conf-parser.h"
+
+#include "nspawn-network.h"
+#include "nspawn-settings.h"
+
+int settings_load(FILE *f, const char *path, Settings **ret) {
+ _cleanup_(settings_freep) Settings *s = NULL;
+ int r;
+
+ assert(path);
+ assert(ret);
+
+ s = new0(Settings, 1);
+ if (!s)
+ return -ENOMEM;
+
+ s->start_mode = _START_MODE_INVALID;
+ s->personality = PERSONALITY_INVALID;
+ s->userns_mode = _USER_NAMESPACE_MODE_INVALID;
+ s->uid_shift = UID_INVALID;
+ s->uid_range = UID_INVALID;
+
+ s->read_only = -1;
+ s->volatile_mode = _VOLATILE_MODE_INVALID;
+ s->userns_chown = -1;
+
+ s->private_network = -1;
+ s->network_veth = -1;
+
+ r = config_parse(NULL, path, f,
+ "Exec\0"
+ "Network\0"
+ "Files\0",
+ config_item_perf_lookup, nspawn_gperf_lookup,
+ false,
+ false,
+ true,
+ s);
+ if (r < 0)
+ return r;
+
+ /* Make sure that if userns_mode is set, userns_chown is set to something appropriate, and vice versa. Either
+ * both fields shall be initialized or neither. */
+ if (s->userns_mode == USER_NAMESPACE_PICK)
+ s->userns_chown = true;
+ else if (s->userns_mode != _USER_NAMESPACE_MODE_INVALID && s->userns_chown < 0)
+ s->userns_chown = false;
+
+ if (s->userns_chown >= 0 && s->userns_mode == _USER_NAMESPACE_MODE_INVALID)
+ s->userns_mode = USER_NAMESPACE_NO;
+
+ *ret = s;
+ s = NULL;
+
+ return 0;
+}
+
+Settings* settings_free(Settings *s) {
+
+ if (!s)
+ return NULL;
+
+ strv_free(s->parameters);
+ strv_free(s->environment);
+ free(s->user);
+ free(s->working_directory);
+
+ strv_free(s->network_interfaces);
+ strv_free(s->network_macvlan);
+ strv_free(s->network_ipvlan);
+ strv_free(s->network_veth_extra);
+ free(s->network_bridge);
+ free(s->network_zone);
+ expose_port_free_all(s->expose_ports);
+
+ custom_mount_free_all(s->custom_mounts, s->n_custom_mounts);
+ return mfree(s);
+}
+
+bool settings_private_network(Settings *s) {
+ assert(s);
+
+ return
+ s->private_network > 0 ||
+ s->network_veth > 0 ||
+ s->network_bridge ||
+ s->network_zone ||
+ s->network_interfaces ||
+ s->network_macvlan ||
+ s->network_ipvlan ||
+ s->network_veth_extra;
+}
+
+bool settings_network_veth(Settings *s) {
+ assert(s);
+
+ return
+ s->network_veth > 0 ||
+ s->network_bridge ||
+ s->network_zone;
+}
+
+DEFINE_CONFIG_PARSE_ENUM(config_parse_volatile_mode, volatile_mode, VolatileMode, "Failed to parse volatile mode");
+
+int config_parse_expose_port(
+ const char *unit,
+ const char *filename,
+ unsigned line,
+ const char *section,
+ unsigned section_line,
+ const char *lvalue,
+ int ltype,
+ const char *rvalue,
+ void *data,
+ void *userdata) {
+
+ Settings *s = data;
+ int r;
+
+ assert(filename);
+ assert(lvalue);
+ assert(rvalue);
+
+ r = expose_port_parse(&s->expose_ports, rvalue);
+ if (r == -EEXIST) {
+ log_syntax(unit, LOG_ERR, filename, line, r, "Duplicate port specification, ignoring: %s", rvalue);
+ return 0;
+ }
+ if (r < 0) {
+ log_syntax(unit, LOG_ERR, filename, line, r, "Failed to parse host port %s: %m", rvalue);
+ return 0;
+ }
+
+ return 0;
+}
+
+int config_parse_capability(
+ const char *unit,
+ const char *filename,
+ unsigned line,
+ const char *section,
+ unsigned section_line,
+ const char *lvalue,
+ int ltype,
+ const char *rvalue,
+ void *data,
+ void *userdata) {
+
+ uint64_t u = 0, *result = data;
+ int r;
+
+ assert(filename);
+ assert(lvalue);
+ assert(rvalue);
+
+ for (;;) {
+ _cleanup_free_ char *word = NULL;
+ int cap;
+
+ r = extract_first_word(&rvalue, &word, NULL, 0);
+ if (r < 0) {
+ log_syntax(unit, LOG_ERR, filename, line, r, "Failed to extract capability string, ignoring: %s", rvalue);
+ return 0;
+ }
+ if (r == 0)
+ break;
+
+ cap = capability_from_name(word);
+ if (cap < 0) {
+ log_syntax(unit, LOG_ERR, filename, line, 0, "Failed to parse capability, ignoring: %s", word);
+ continue;
+ }
+
+ u |= 1 << ((uint64_t) cap);
+ }
+
+ if (u == 0)
+ return 0;
+
+ *result |= u;
+ return 0;
+}
+
+int config_parse_id128(
+ const char *unit,
+ const char *filename,
+ unsigned line,
+ const char *section,
+ unsigned section_line,
+ const char *lvalue,
+ int ltype,
+ const char *rvalue,
+ void *data,
+ void *userdata) {
+
+ sd_id128_t t, *result = data;
+ int r;
+
+ assert(filename);
+ assert(lvalue);
+ assert(rvalue);
+
+ r = sd_id128_from_string(rvalue, &t);
+ if (r < 0) {
+ log_syntax(unit, LOG_ERR, filename, line, r, "Failed to parse 128bit ID/UUID, ignoring: %s", rvalue);
+ return 0;
+ }
+
+ *result = t;
+ return 0;
+}
+
+int config_parse_bind(
+ const char *unit,
+ const char *filename,
+ unsigned line,
+ const char *section,
+ unsigned section_line,
+ const char *lvalue,
+ int ltype,
+ const char *rvalue,
+ void *data,
+ void *userdata) {
+
+ Settings *settings = data;
+ int r;
+
+ assert(filename);
+ assert(lvalue);
+ assert(rvalue);
+
+ r = bind_mount_parse(&settings->custom_mounts, &settings->n_custom_mounts, rvalue, ltype);
+ if (r < 0) {
+ log_syntax(unit, LOG_ERR, filename, line, r, "Invalid bind mount specification %s: %m", rvalue);
+ return 0;
+ }
+
+ return 0;
+}
+
+int config_parse_tmpfs(
+ const char *unit,
+ const char *filename,
+ unsigned line,
+ const char *section,
+ unsigned section_line,
+ const char *lvalue,
+ int ltype,
+ const char *rvalue,
+ void *data,
+ void *userdata) {
+
+ Settings *settings = data;
+ int r;
+
+ assert(filename);
+ assert(lvalue);
+ assert(rvalue);
+
+ r = tmpfs_mount_parse(&settings->custom_mounts, &settings->n_custom_mounts, rvalue);
+ if (r < 0) {
+ log_syntax(unit, LOG_ERR, filename, line, r, "Invalid temporary file system specification %s: %m", rvalue);
+ return 0;
+ }
+
+ return 0;
+}
+
+int config_parse_veth_extra(
+ const char *unit,
+ const char *filename,
+ unsigned line,
+ const char *section,
+ unsigned section_line,
+ const char *lvalue,
+ int ltype,
+ const char *rvalue,
+ void *data,
+ void *userdata) {
+
+ Settings *settings = data;
+ int r;
+
+ assert(filename);
+ assert(lvalue);
+ assert(rvalue);
+
+ r = veth_extra_parse(&settings->network_veth_extra, rvalue);
+ if (r < 0) {
+ log_syntax(unit, LOG_ERR, filename, line, r, "Invalid extra virtual Ethernet link specification %s: %m", rvalue);
+ return 0;
+ }
+
+ return 0;
+}
+
+int config_parse_network_zone(
+ const char *unit,
+ const char *filename,
+ unsigned line,
+ const char *section,
+ unsigned section_line,
+ const char *lvalue,
+ int ltype,
+ const char *rvalue,
+ void *data,
+ void *userdata) {
+
+ Settings *settings = data;
+ _cleanup_free_ char *j = NULL;
+
+ assert(filename);
+ assert(lvalue);
+ assert(rvalue);
+
+ j = strappend("vz-", rvalue);
+ if (!ifname_valid(j)) {
+ log_syntax(unit, LOG_ERR, filename, line, 0, "Invalid network zone name %s, ignoring: %m", rvalue);
+ return 0;
+ }
+
+ free(settings->network_zone);
+ settings->network_zone = j;
+ j = NULL;
+
+ return 0;
+}
+
+int config_parse_boot(
+ const char *unit,
+ const char *filename,
+ unsigned line,
+ const char *section,
+ unsigned section_line,
+ const char *lvalue,
+ int ltype,
+ const char *rvalue,
+ void *data,
+ void *userdata) {
+
+ Settings *settings = data;
+ int r;
+
+ assert(filename);
+ assert(lvalue);
+ assert(rvalue);
+
+ r = parse_boolean(rvalue);
+ if (r < 0) {
+ log_syntax(unit, LOG_ERR, filename, line, r, "Failed to parse Boot= parameter %s, ignoring: %m", rvalue);
+ return 0;
+ }
+
+ if (r > 0) {
+ if (settings->start_mode == START_PID2)
+ goto conflict;
+
+ settings->start_mode = START_BOOT;
+ } else {
+ if (settings->start_mode == START_BOOT)
+ goto conflict;
+
+ if (settings->start_mode < 0)
+ settings->start_mode = START_PID1;
+ }
+
+ return 0;
+
+conflict:
+ log_syntax(unit, LOG_ERR, filename, line, r, "Conflicting Boot= or ProcessTwo= setting found. Ignoring.");
+ return 0;
+}
+
+int config_parse_pid2(
+ const char *unit,
+ const char *filename,
+ unsigned line,
+ const char *section,
+ unsigned section_line,
+ const char *lvalue,
+ int ltype,
+ const char *rvalue,
+ void *data,
+ void *userdata) {
+
+ Settings *settings = data;
+ int r;
+
+ assert(filename);
+ assert(lvalue);
+ assert(rvalue);
+
+ r = parse_boolean(rvalue);
+ if (r < 0) {
+ log_syntax(unit, LOG_ERR, filename, line, r, "Failed to parse ProcessTwo= parameter %s, ignoring: %m", rvalue);
+ return 0;
+ }
+
+ if (r > 0) {
+ if (settings->start_mode == START_BOOT)
+ goto conflict;
+
+ settings->start_mode = START_PID2;
+ } else {
+ if (settings->start_mode == START_PID2)
+ goto conflict;
+
+ if (settings->start_mode < 0)
+ settings->start_mode = START_PID1;
+ }
+
+ return 0;
+
+conflict:
+ log_syntax(unit, LOG_ERR, filename, line, r, "Conflicting Boot= or ProcessTwo= setting found. Ignoring.");
+ return 0;
+}
+
+int config_parse_private_users(
+ const char *unit,
+ const char *filename,
+ unsigned line,
+ const char *section,
+ unsigned section_line,
+ const char *lvalue,
+ int ltype,
+ const char *rvalue,
+ void *data,
+ void *userdata) {
+
+ Settings *settings = data;
+ int r;
+
+ assert(filename);
+ assert(lvalue);
+ assert(rvalue);
+
+ r = parse_boolean(rvalue);
+ if (r == 0) {
+ /* no: User namespacing off */
+ settings->userns_mode = USER_NAMESPACE_NO;
+ settings->uid_shift = UID_INVALID;
+ settings->uid_range = UINT32_C(0x10000);
+ } else if (r > 0) {
+ /* yes: User namespacing on, UID range is read from root dir */
+ settings->userns_mode = USER_NAMESPACE_FIXED;
+ settings->uid_shift = UID_INVALID;
+ settings->uid_range = UINT32_C(0x10000);
+ } else if (streq(rvalue, "pick")) {
+ /* pick: User namespacing on, UID range is picked randomly */
+ settings->userns_mode = USER_NAMESPACE_PICK;
+ settings->uid_shift = UID_INVALID;
+ settings->uid_range = UINT32_C(0x10000);
+ } else {
+ const char *range, *shift;
+ uid_t sh, rn;
+
+ /* anything else: User namespacing on, UID range is explicitly configured */
+
+ range = strchr(rvalue, ':');
+ if (range) {
+ shift = strndupa(rvalue, range - rvalue);
+ range++;
+
+ r = safe_atou32(range, &rn);
+ if (r < 0 || rn <= 0) {
+ log_syntax(unit, LOG_ERR, filename, line, r, "UID/GID range invalid, ignoring: %s", range);
+ return 0;
+ }
+ } else {
+ shift = rvalue;
+ rn = UINT32_C(0x10000);
+ }
+
+ r = parse_uid(shift, &sh);
+ if (r < 0) {
+ log_syntax(unit, LOG_ERR, filename, line, r, "UID/GID shift invalid, ignoring: %s", range);
+ return 0;
+ }
+
+ settings->userns_mode = USER_NAMESPACE_FIXED;
+ settings->uid_shift = sh;
+ settings->uid_range = rn;
+ }
+
+ return 0;
+}
diff --git a/src/systemd-nspawn/nspawn-settings.h b/src/systemd-nspawn/nspawn-settings.h
new file mode 100644
index 0000000000..2e57072eec
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-settings.h
@@ -0,0 +1,119 @@
+#pragma once
+
+/***
+ This file is part of systemd.
+
+ Copyright 2015 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include <stdio.h>
+
+#include "systemd-basic/macro.h"
+
+#include "nspawn-expose-ports.h"
+#include "nspawn-mount.h"
+
+typedef enum StartMode {
+ START_PID1, /* Run parameters as command line as process 1 */
+ START_PID2, /* Use stub init process as PID 1, run parameters as command line as process 2 */
+ START_BOOT, /* Search for init system, pass arguments as parameters */
+ _START_MODE_MAX,
+ _START_MODE_INVALID = -1
+} StartMode;
+
+typedef enum UserNamespaceMode {
+ USER_NAMESPACE_NO,
+ USER_NAMESPACE_FIXED,
+ USER_NAMESPACE_PICK,
+ _USER_NAMESPACE_MODE_MAX,
+ _USER_NAMESPACE_MODE_INVALID = -1,
+} UserNamespaceMode;
+
+typedef enum SettingsMask {
+ SETTING_START_MODE = 1 << 0,
+ SETTING_ENVIRONMENT = 1 << 1,
+ SETTING_USER = 1 << 2,
+ SETTING_CAPABILITY = 1 << 3,
+ SETTING_KILL_SIGNAL = 1 << 4,
+ SETTING_PERSONALITY = 1 << 5,
+ SETTING_MACHINE_ID = 1 << 6,
+ SETTING_NETWORK = 1 << 7,
+ SETTING_EXPOSE_PORTS = 1 << 8,
+ SETTING_READ_ONLY = 1 << 9,
+ SETTING_VOLATILE_MODE = 1 << 10,
+ SETTING_CUSTOM_MOUNTS = 1 << 11,
+ SETTING_WORKING_DIRECTORY = 1 << 12,
+ SETTING_USERNS = 1 << 13,
+ SETTING_NOTIFY_READY = 1 << 14,
+ _SETTINGS_MASK_ALL = (1 << 15) -1
+} SettingsMask;
+
+typedef struct Settings {
+ /* [Run] */
+ StartMode start_mode;
+ char **parameters;
+ char **environment;
+ char *user;
+ uint64_t capability;
+ uint64_t drop_capability;
+ int kill_signal;
+ unsigned long personality;
+ sd_id128_t machine_id;
+ char *working_directory;
+ UserNamespaceMode userns_mode;
+ uid_t uid_shift, uid_range;
+ bool notify_ready;
+
+ /* [Image] */
+ int read_only;
+ VolatileMode volatile_mode;
+ CustomMount *custom_mounts;
+ unsigned n_custom_mounts;
+ int userns_chown;
+
+ /* [Network] */
+ int private_network;
+ int network_veth;
+ char *network_bridge;
+ char *network_zone;
+ char **network_interfaces;
+ char **network_macvlan;
+ char **network_ipvlan;
+ char **network_veth_extra;
+ ExposePort *expose_ports;
+} Settings;
+
+int settings_load(FILE *f, const char *path, Settings **ret);
+Settings* settings_free(Settings *s);
+
+bool settings_network_veth(Settings *s);
+bool settings_private_network(Settings *s);
+
+DEFINE_TRIVIAL_CLEANUP_FUNC(Settings*, settings_free);
+
+const struct ConfigPerfItem* nspawn_gperf_lookup(const char *key, GPERF_LEN_TYPE length);
+
+int config_parse_capability(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
+int config_parse_id128(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
+int config_parse_expose_port(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
+int config_parse_volatile_mode(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
+int config_parse_bind(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
+int config_parse_tmpfs(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
+int config_parse_veth_extra(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
+int config_parse_network_zone(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
+int config_parse_boot(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
+int config_parse_pid2(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
+int config_parse_private_users(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
diff --git a/src/systemd-nspawn/nspawn-setuid.c b/src/systemd-nspawn/nspawn-setuid.c
new file mode 100644
index 0000000000..129d3acc5f
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-setuid.c
@@ -0,0 +1,271 @@
+/***
+ This file is part of systemd.
+
+ Copyright 2015 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include <grp.h>
+#include <sys/types.h>
+#include <unistd.h>
+
+#include "systemd-basic/alloc-util.h"
+#include "systemd-basic/fd-util.h"
+#include "systemd-basic/mkdir.h"
+#include "systemd-basic/process-util.h"
+#include "systemd-basic/signal-util.h"
+#include "systemd-basic/string-util.h"
+#include "systemd-basic/user-util.h"
+#include "systemd-basic/util.h"
+
+#include "nspawn-setuid.h"
+
+static int spawn_getent(const char *database, const char *key, pid_t *rpid) {
+ int pipe_fds[2];
+ pid_t pid;
+
+ assert(database);
+ assert(key);
+ assert(rpid);
+
+ if (pipe2(pipe_fds, O_CLOEXEC) < 0)
+ return log_error_errno(errno, "Failed to allocate pipe: %m");
+
+ pid = fork();
+ if (pid < 0)
+ return log_error_errno(errno, "Failed to fork getent child: %m");
+ else if (pid == 0) {
+ int nullfd;
+ char *empty_env = NULL;
+
+ if (dup3(pipe_fds[1], STDOUT_FILENO, 0) < 0)
+ _exit(EXIT_FAILURE);
+
+ if (pipe_fds[0] > 2)
+ safe_close(pipe_fds[0]);
+ if (pipe_fds[1] > 2)
+ safe_close(pipe_fds[1]);
+
+ nullfd = open("/dev/null", O_RDWR);
+ if (nullfd < 0)
+ _exit(EXIT_FAILURE);
+
+ if (dup3(nullfd, STDIN_FILENO, 0) < 0)
+ _exit(EXIT_FAILURE);
+
+ if (dup3(nullfd, STDERR_FILENO, 0) < 0)
+ _exit(EXIT_FAILURE);
+
+ if (nullfd > 2)
+ safe_close(nullfd);
+
+ (void) reset_all_signal_handlers();
+ (void) reset_signal_mask();
+ close_all_fds(NULL, 0);
+
+ execle("/usr/bin/getent", "getent", database, key, NULL, &empty_env);
+ execle("/bin/getent", "getent", database, key, NULL, &empty_env);
+ _exit(EXIT_FAILURE);
+ }
+
+ pipe_fds[1] = safe_close(pipe_fds[1]);
+
+ *rpid = pid;
+
+ return pipe_fds[0];
+}
+
+int change_uid_gid(const char *user, char **_home) {
+ char line[LINE_MAX], *x, *u, *g, *h;
+ const char *word, *state;
+ _cleanup_free_ uid_t *uids = NULL;
+ _cleanup_free_ char *home = NULL;
+ _cleanup_fclose_ FILE *f = NULL;
+ _cleanup_close_ int fd = -1;
+ unsigned n_uids = 0;
+ size_t sz = 0, l;
+ uid_t uid;
+ gid_t gid;
+ pid_t pid;
+ int r;
+
+ assert(_home);
+
+ if (!user || streq(user, "root") || streq(user, "0")) {
+ /* Reset everything fully to 0, just in case */
+
+ r = reset_uid_gid();
+ if (r < 0)
+ return log_error_errno(r, "Failed to become root: %m");
+
+ *_home = NULL;
+ return 0;
+ }
+
+ /* First, get user credentials */
+ fd = spawn_getent("passwd", user, &pid);
+ if (fd < 0)
+ return fd;
+
+ f = fdopen(fd, "r");
+ if (!f)
+ return log_oom();
+ fd = -1;
+
+ if (!fgets(line, sizeof(line), f)) {
+ if (!ferror(f)) {
+ log_error("Failed to resolve user %s.", user);
+ return -ESRCH;
+ }
+
+ return log_error_errno(errno, "Failed to read from getent: %m");
+ }
+
+ truncate_nl(line);
+
+ wait_for_terminate_and_warn("getent passwd", pid, true);
+
+ x = strchr(line, ':');
+ if (!x) {
+ log_error("/etc/passwd entry has invalid user field.");
+ return -EIO;
+ }
+
+ u = strchr(x+1, ':');
+ if (!u) {
+ log_error("/etc/passwd entry has invalid password field.");
+ return -EIO;
+ }
+
+ u++;
+ g = strchr(u, ':');
+ if (!g) {
+ log_error("/etc/passwd entry has invalid UID field.");
+ return -EIO;
+ }
+
+ *g = 0;
+ g++;
+ x = strchr(g, ':');
+ if (!x) {
+ log_error("/etc/passwd entry has invalid GID field.");
+ return -EIO;
+ }
+
+ *x = 0;
+ h = strchr(x+1, ':');
+ if (!h) {
+ log_error("/etc/passwd entry has invalid GECOS field.");
+ return -EIO;
+ }
+
+ h++;
+ x = strchr(h, ':');
+ if (!x) {
+ log_error("/etc/passwd entry has invalid home directory field.");
+ return -EIO;
+ }
+
+ *x = 0;
+
+ r = parse_uid(u, &uid);
+ if (r < 0) {
+ log_error("Failed to parse UID of user.");
+ return -EIO;
+ }
+
+ r = parse_gid(g, &gid);
+ if (r < 0) {
+ log_error("Failed to parse GID of user.");
+ return -EIO;
+ }
+
+ home = strdup(h);
+ if (!home)
+ return log_oom();
+
+ /* Second, get group memberships */
+ fd = spawn_getent("initgroups", user, &pid);
+ if (fd < 0)
+ return fd;
+
+ fclose(f);
+ f = fdopen(fd, "r");
+ if (!f)
+ return log_oom();
+ fd = -1;
+
+ if (!fgets(line, sizeof(line), f)) {
+ if (!ferror(f)) {
+ log_error("Failed to resolve user %s.", user);
+ return -ESRCH;
+ }
+
+ return log_error_errno(errno, "Failed to read from getent: %m");
+ }
+
+ truncate_nl(line);
+
+ wait_for_terminate_and_warn("getent initgroups", pid, true);
+
+ /* Skip over the username and subsequent separator whitespace */
+ x = line;
+ x += strcspn(x, WHITESPACE);
+ x += strspn(x, WHITESPACE);
+
+ FOREACH_WORD(word, l, x, state) {
+ char c[l+1];
+
+ memcpy(c, word, l);
+ c[l] = 0;
+
+ if (!GREEDY_REALLOC(uids, sz, n_uids+1))
+ return log_oom();
+
+ r = parse_uid(c, &uids[n_uids++]);
+ if (r < 0) {
+ log_error("Failed to parse group data from getent.");
+ return -EIO;
+ }
+ }
+
+ r = mkdir_parents(home, 0775);
+ if (r < 0)
+ return log_error_errno(r, "Failed to make home root directory: %m");
+
+ r = mkdir_safe(home, 0755, uid, gid);
+ if (r < 0 && r != -EEXIST)
+ return log_error_errno(r, "Failed to make home directory: %m");
+
+ (void) fchown(STDIN_FILENO, uid, gid);
+ (void) fchown(STDOUT_FILENO, uid, gid);
+ (void) fchown(STDERR_FILENO, uid, gid);
+
+ if (setgroups(n_uids, uids) < 0)
+ return log_error_errno(errno, "Failed to set auxiliary groups: %m");
+
+ if (setresgid(gid, gid, gid) < 0)
+ return log_error_errno(errno, "setresgid() failed: %m");
+
+ if (setresuid(uid, uid, uid) < 0)
+ return log_error_errno(errno, "setresuid() failed: %m");
+
+ if (_home) {
+ *_home = home;
+ home = NULL;
+ }
+
+ return 0;
+}
diff --git a/src/systemd-nspawn/nspawn-setuid.h b/src/systemd-nspawn/nspawn-setuid.h
new file mode 100644
index 0000000000..b4968ba1fc
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-setuid.h
@@ -0,0 +1,22 @@
+#pragma once
+
+/***
+ This file is part of systemd.
+
+ Copyright 2015 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+int change_uid_gid(const char *user, char **ret);
diff --git a/src/systemd-nspawn/nspawn-stub-pid1.c b/src/systemd-nspawn/nspawn-stub-pid1.c
new file mode 100644
index 0000000000..36d7bfc7c4
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-stub-pid1.c
@@ -0,0 +1,171 @@
+/***
+ This file is part of systemd.
+
+ Copyright 2016 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include <sys/reboot.h>
+#include <sys/unistd.h>
+#include <sys/wait.h>
+
+#include "systemd-basic/def.h"
+#include "systemd-basic/fd-util.h"
+#include "systemd-basic/log.h"
+#include "systemd-basic/process-util.h"
+#include "systemd-basic/signal-util.h"
+#include "systemd-basic/time-util.h"
+
+#include "nspawn-stub-pid1.h"
+
+int stub_pid1(void) {
+ enum {
+ STATE_RUNNING,
+ STATE_REBOOT,
+ STATE_POWEROFF,
+ } state = STATE_RUNNING;
+
+ sigset_t fullmask, oldmask, waitmask;
+ usec_t quit_usec = USEC_INFINITY;
+ pid_t pid;
+ int r;
+
+ /* Implements a stub PID 1, that reaps all processes and processes a couple of standard signals. This is useful
+ * for allowing arbitrary processes run in a container, and still have all zombies reaped. */
+
+ assert_se(sigfillset(&fullmask) >= 0);
+ assert_se(sigprocmask(SIG_BLOCK, &fullmask, &oldmask) >= 0);
+
+ pid = fork();
+ if (pid < 0)
+ return log_error_errno(errno, "Failed to fork child pid: %m");
+
+ if (pid == 0) {
+ /* Return in the child */
+ assert_se(sigprocmask(SIG_SETMASK, &oldmask, NULL) >= 0);
+ setsid();
+ return 0;
+ }
+
+ reset_all_signal_handlers();
+
+ log_close();
+ close_all_fds(NULL, 0);
+ log_open();
+
+ rename_process("STUBINIT");
+
+ assert_se(sigemptyset(&waitmask) >= 0);
+ assert_se(sigset_add_many(&waitmask,
+ SIGCHLD, /* posix: process died */
+ SIGINT, /* sysv: ctrl-alt-del */
+ SIGRTMIN+3, /* systemd: halt */
+ SIGRTMIN+4, /* systemd: poweroff */
+ SIGRTMIN+5, /* systemd: reboot */
+ SIGRTMIN+6, /* systemd: kexec */
+ SIGRTMIN+13, /* systemd: halt */
+ SIGRTMIN+14, /* systemd: poweroff */
+ SIGRTMIN+15, /* systemd: reboot */
+ SIGRTMIN+16, /* systemd: kexec */
+ -1) >= 0);
+
+ /* Note that we ignore SIGTERM (sysv's reexec), SIGHUP (reload), and all other signals here, since we don't
+ * support reexec/reloading in this stub process. */
+
+ for (;;) {
+ siginfo_t si;
+ usec_t current_usec;
+
+ si.si_pid = 0;
+ r = waitid(P_ALL, 0, &si, WEXITED|WNOHANG);
+ if (r < 0) {
+ r = log_error_errno(errno, "Failed to reap children: %m");
+ goto finish;
+ }
+
+ current_usec = now(CLOCK_MONOTONIC);
+
+ if (si.si_pid == pid || current_usec >= quit_usec) {
+
+ /* The child we started ourselves died or we reached a timeout. */
+
+ if (state == STATE_REBOOT) { /* dispatch a queued reboot */
+ (void) reboot(RB_AUTOBOOT);
+ r = log_error_errno(errno, "Failed to reboot: %m");
+ goto finish;
+
+ } else if (state == STATE_POWEROFF)
+ (void) reboot(RB_POWER_OFF); /* if this fails, fall back to normal exit. */
+
+ if (si.si_pid == pid && si.si_code == CLD_EXITED)
+ r = si.si_status; /* pass on exit code */
+ else
+ r = 255; /* signal, coredump, timeout, … */
+
+ goto finish;
+ }
+ if (si.si_pid != 0)
+ /* We reaped something. Retry until there's nothing more to reap. */
+ continue;
+
+ if (quit_usec == USEC_INFINITY)
+ r = sigwaitinfo(&waitmask, &si);
+ else {
+ struct timespec ts;
+ r = sigtimedwait(&waitmask, &si, timespec_store(&ts, quit_usec - current_usec));
+ }
+ if (r < 0) {
+ if (errno == EINTR) /* strace -p attach can result in EINTR, let's handle this nicely. */
+ continue;
+ if (errno == EAGAIN) /* timeout reached */
+ continue;
+
+ r = log_error_errno(errno, "Failed to wait for signal: %m");
+ goto finish;
+ }
+
+ if (si.si_signo == SIGCHLD)
+ continue; /* Let's reap this */
+
+ if (state != STATE_RUNNING)
+ continue;
+
+ /* Would love to use a switch() statement here, but SIGRTMIN is actually a function call, not a
+ * constant… */
+
+ if (si.si_signo == SIGRTMIN+3 ||
+ si.si_signo == SIGRTMIN+4 ||
+ si.si_signo == SIGRTMIN+13 ||
+ si.si_signo == SIGRTMIN+14)
+
+ state = STATE_POWEROFF;
+
+ else if (si.si_signo == SIGINT ||
+ si.si_signo == SIGRTMIN+5 ||
+ si.si_signo == SIGRTMIN+6 ||
+ si.si_signo == SIGRTMIN+15 ||
+ si.si_signo == SIGRTMIN+16)
+
+ state = STATE_REBOOT;
+ else
+ assert_not_reached("Got unexpected signal");
+
+ /* (void) kill_and_sigcont(pid, SIGTERM); */
+ quit_usec = now(CLOCK_MONOTONIC) + DEFAULT_TIMEOUT_USEC;
+ }
+
+finish:
+ _exit(r < 0 ? EXIT_FAILURE : r);
+}
diff --git a/src/systemd-nspawn/nspawn-stub-pid1.h b/src/systemd-nspawn/nspawn-stub-pid1.h
new file mode 100644
index 0000000000..36c1aaf5dd
--- /dev/null
+++ b/src/systemd-nspawn/nspawn-stub-pid1.h
@@ -0,0 +1,22 @@
+#pragma once
+
+/***
+ This file is part of systemd.
+
+ Copyright 2016 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+int stub_pid1(void);
diff --git a/src/systemd-nspawn/nspawn.c b/src/systemd-nspawn/nspawn.c
new file mode 100644
index 0000000000..5a3624fad1
--- /dev/null
+++ b/src/systemd-nspawn/nspawn.c
@@ -0,0 +1,4319 @@
+/***
+ This file is part of systemd.
+
+ Copyright 2010 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#ifdef HAVE_BLKID
+#include <blkid/blkid.h>
+#endif
+#include <errno.h>
+#include <getopt.h>
+#include <grp.h>
+#include <pwd.h>
+#include <sched.h>
+
+#include <linux/loop.h>
+#ifdef HAVE_SELINUX
+#include <selinux/selinux.h>
+#endif
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/file.h>
+#include <sys/mount.h>
+#include <sys/personality.h>
+#include <sys/prctl.h>
+#include <sys/types.h>
+#include <unistd.h>
+
+#include <systemd/sd-daemon.h>
+#include <systemd/sd-id128.h>
+
+#include "core/loopback-setup.h"
+#include "sd-id128/id128-util.h"
+#include "sd-netlink/netlink-util.h"
+#include "systemd-basic/alloc-util.h"
+#include "systemd-basic/barrier.h"
+#include "systemd-basic/btrfs-util.h"
+#include "systemd-basic/cap-list.h"
+#include "systemd-basic/capability-util.h"
+#include "systemd-basic/cgroup-util.h"
+#include "systemd-basic/copy.h"
+#include "systemd-basic/env-util.h"
+#include "systemd-basic/fd-util.h"
+#include "systemd-basic/fileio.h"
+#include "systemd-basic/formats-util.h"
+#include "systemd-basic/fs-util.h"
+#include "systemd-basic/hostname-util.h"
+#include "systemd-basic/log.h"
+#include "systemd-basic/macro.h"
+#include "systemd-basic/missing.h"
+#include "systemd-basic/mkdir.h"
+#include "systemd-basic/mount-util.h"
+#include "systemd-basic/parse-util.h"
+#include "systemd-basic/path-util.h"
+#include "systemd-basic/process-util.h"
+#include "systemd-basic/random-util.h"
+#include "systemd-basic/raw-clone.h"
+#include "systemd-basic/rm-rf.h"
+#include "systemd-basic/selinux-util.h"
+#include "systemd-basic/signal-util.h"
+#include "systemd-basic/socket-util.h"
+#include "systemd-basic/stat-util.h"
+#include "systemd-basic/stdio-util.h"
+#include "systemd-basic/string-util.h"
+#include "systemd-basic/strv.h"
+#include "systemd-basic/terminal-util.h"
+#include "systemd-basic/umask-util.h"
+#include "systemd-basic/user-util.h"
+#include "systemd-basic/util.h"
+#include "systemd-blkid/blkid-util.h"
+#include "systemd-shared/base-filesystem.h"
+#include "systemd-shared/dev-setup.h"
+#include "systemd-shared/fdset.h"
+#include "systemd-shared/gpt.h"
+#include "systemd-shared/machine-image.h"
+#include "systemd-shared/ptyfwd.h"
+#include "systemd-shared/udev-util.h"
+
+#include "nspawn-cgroup.h"
+#include "nspawn-expose-ports.h"
+#include "nspawn-mount.h"
+#include "nspawn-network.h"
+#include "nspawn-patch-uid.h"
+#include "nspawn-register.h"
+#include "nspawn-seccomp.h"
+#include "nspawn-settings.h"
+#include "nspawn-setuid.h"
+#include "nspawn-stub-pid1.h"
+
+/* Note that devpts's gid= parameter parses GIDs as signed values, hence we stay away from the upper half of the 32bit
+ * UID range here. We leave a bit of room at the lower end and a lot of room at the upper end, so that other subsystems
+ * may have their own allocation ranges too. */
+#define UID_SHIFT_PICK_MIN ((uid_t) UINT32_C(0x00080000))
+#define UID_SHIFT_PICK_MAX ((uid_t) UINT32_C(0x6FFF0000))
+
+/* nspawn is listening on the socket at the path in the constant nspawn_notify_socket_path
+ * nspawn_notify_socket_path is relative to the container
+ * the init process in the container pid can send messages to nspawn following the sd_notify(3) protocol */
+#define NSPAWN_NOTIFY_SOCKET_PATH "/run/systemd/nspawn/notify"
+
+typedef enum ContainerStatus {
+ CONTAINER_TERMINATED,
+ CONTAINER_REBOOTED
+} ContainerStatus;
+
+typedef enum LinkJournal {
+ LINK_NO,
+ LINK_AUTO,
+ LINK_HOST,
+ LINK_GUEST
+} LinkJournal;
+
+static char *arg_directory = NULL;
+static char *arg_template = NULL;
+static char *arg_chdir = NULL;
+static char *arg_user = NULL;
+static sd_id128_t arg_uuid = {};
+static char *arg_machine = NULL;
+static const char *arg_selinux_context = NULL;
+static const char *arg_selinux_apifs_context = NULL;
+static const char *arg_slice = NULL;
+static bool arg_private_network = false;
+static bool arg_read_only = false;
+static StartMode arg_start_mode = START_PID1;
+static bool arg_ephemeral = false;
+static LinkJournal arg_link_journal = LINK_AUTO;
+static bool arg_link_journal_try = false;
+static uint64_t arg_caps_retain =
+ (1ULL << CAP_AUDIT_CONTROL) |
+ (1ULL << CAP_AUDIT_WRITE) |
+ (1ULL << CAP_CHOWN) |
+ (1ULL << CAP_DAC_OVERRIDE) |
+ (1ULL << CAP_DAC_READ_SEARCH) |
+ (1ULL << CAP_FOWNER) |
+ (1ULL << CAP_FSETID) |
+ (1ULL << CAP_IPC_OWNER) |
+ (1ULL << CAP_KILL) |
+ (1ULL << CAP_LEASE) |
+ (1ULL << CAP_LINUX_IMMUTABLE) |
+ (1ULL << CAP_MKNOD) |
+ (1ULL << CAP_NET_BIND_SERVICE) |
+ (1ULL << CAP_NET_BROADCAST) |
+ (1ULL << CAP_NET_RAW) |
+ (1ULL << CAP_SETFCAP) |
+ (1ULL << CAP_SETGID) |
+ (1ULL << CAP_SETPCAP) |
+ (1ULL << CAP_SETUID) |
+ (1ULL << CAP_SYS_ADMIN) |
+ (1ULL << CAP_SYS_BOOT) |
+ (1ULL << CAP_SYS_CHROOT) |
+ (1ULL << CAP_SYS_NICE) |
+ (1ULL << CAP_SYS_PTRACE) |
+ (1ULL << CAP_SYS_RESOURCE) |
+ (1ULL << CAP_SYS_TTY_CONFIG);
+static CustomMount *arg_custom_mounts = NULL;
+static unsigned arg_n_custom_mounts = 0;
+static char **arg_setenv = NULL;
+static bool arg_quiet = false;
+static bool arg_register = true;
+static bool arg_keep_unit = false;
+static char **arg_network_interfaces = NULL;
+static char **arg_network_macvlan = NULL;
+static char **arg_network_ipvlan = NULL;
+static bool arg_network_veth = false;
+static char **arg_network_veth_extra = NULL;
+static char *arg_network_bridge = NULL;
+static char *arg_network_zone = NULL;
+static unsigned long arg_personality = PERSONALITY_INVALID;
+static char *arg_image = NULL;
+static VolatileMode arg_volatile_mode = VOLATILE_NO;
+static ExposePort *arg_expose_ports = NULL;
+static char **arg_property = NULL;
+static UserNamespaceMode arg_userns_mode = USER_NAMESPACE_NO;
+static uid_t arg_uid_shift = UID_INVALID, arg_uid_range = 0x10000U;
+static bool arg_userns_chown = false;
+static int arg_kill_signal = 0;
+static CGroupUnified arg_unified_cgroup_hierarchy = CGROUP_UNIFIED_UNKNOWN;
+static SettingsMask arg_settings_mask = 0;
+static int arg_settings_trusted = -1;
+static char **arg_parameters = NULL;
+static const char *arg_container_service_name = "systemd-nspawn";
+static bool arg_notify_ready = false;
+static bool arg_use_cgns = true;
+static unsigned long arg_clone_ns_flags = CLONE_NEWIPC|CLONE_NEWPID|CLONE_NEWUTS;
+
+static void help(void) {
+ printf("%s [OPTIONS...] [PATH] [ARGUMENTS...]\n\n"
+ "Spawn a minimal namespace container for debugging, testing and building.\n\n"
+ " -h --help Show this help\n"
+ " --version Print version string\n"
+ " -q --quiet Do not show status information\n"
+ " -D --directory=PATH Root directory for the container\n"
+ " --template=PATH Initialize root directory from template directory,\n"
+ " if missing\n"
+ " -x --ephemeral Run container with snapshot of root directory, and\n"
+ " remove it after exit\n"
+ " -i --image=PATH File system device or disk image for the container\n"
+ " -a --as-pid2 Maintain a stub init as PID1, invoke binary as PID2\n"
+ " -b --boot Boot up full system (i.e. invoke init)\n"
+ " --chdir=PATH Set working directory in the container\n"
+ " -u --user=USER Run the command under specified user or uid\n"
+ " -M --machine=NAME Set the machine name for the container\n"
+ " --uuid=UUID Set a specific machine UUID for the container\n"
+ " -S --slice=SLICE Place the container in the specified slice\n"
+ " --property=NAME=VALUE Set scope unit property\n"
+ " -U --private-users=pick Run within user namespace, autoselect UID/GID range\n"
+ " --private-users[=UIDBASE[:NUIDS]]\n"
+ " Similar, but with user configured UID/GID range\n"
+ " --private-users-chown Adjust OS tree ownership to private UID/GID range\n"
+ " --private-network Disable network in container\n"
+ " --network-interface=INTERFACE\n"
+ " Assign an existing network interface to the\n"
+ " container\n"
+ " --network-macvlan=INTERFACE\n"
+ " Create a macvlan network interface based on an\n"
+ " existing network interface to the container\n"
+ " --network-ipvlan=INTERFACE\n"
+ " Create a ipvlan network interface based on an\n"
+ " existing network interface to the container\n"
+ " -n --network-veth Add a virtual Ethernet connection between host\n"
+ " and container\n"
+ " --network-veth-extra=HOSTIF[:CONTAINERIF]\n"
+ " Add an additional virtual Ethernet link between\n"
+ " host and container\n"
+ " --network-bridge=INTERFACE\n"
+ " Add a virtual Ethernet connection to the container\n"
+ " and attach it to an existing bridge on the host\n"
+ " --network-zone=NAME Similar, but attach the new interface to an\n"
+ " an automatically managed bridge interface\n"
+ " -p --port=[PROTOCOL:]HOSTPORT[:CONTAINERPORT]\n"
+ " Expose a container IP port on the host\n"
+ " -Z --selinux-context=SECLABEL\n"
+ " Set the SELinux security context to be used by\n"
+ " processes in the container\n"
+ " -L --selinux-apifs-context=SECLABEL\n"
+ " Set the SELinux security context to be used by\n"
+ " API/tmpfs file systems in the container\n"
+ " --capability=CAP In addition to the default, retain specified\n"
+ " capability\n"
+ " --drop-capability=CAP Drop the specified capability from the default set\n"
+ " --kill-signal=SIGNAL Select signal to use for shutting down PID 1\n"
+ " --link-journal=MODE Link up guest journal, one of no, auto, guest, \n"
+ " host, try-guest, try-host\n"
+ " -j Equivalent to --link-journal=try-guest\n"
+ " --read-only Mount the root directory read-only\n"
+ " --bind=PATH[:PATH[:OPTIONS]]\n"
+ " Bind mount a file or directory from the host into\n"
+ " the container\n"
+ " --bind-ro=PATH[:PATH[:OPTIONS]\n"
+ " Similar, but creates a read-only bind mount\n"
+ " --tmpfs=PATH:[OPTIONS] Mount an empty tmpfs to the specified directory\n"
+ " --overlay=PATH[:PATH...]:PATH\n"
+ " Create an overlay mount from the host to \n"
+ " the container\n"
+ " --overlay-ro=PATH[:PATH...]:PATH\n"
+ " Similar, but creates a read-only overlay mount\n"
+ " -E --setenv=NAME=VALUE Pass an environment variable to PID 1\n"
+ " --register=BOOLEAN Register container as machine\n"
+ " --keep-unit Do not register a scope for the machine, reuse\n"
+ " the service unit nspawn is running in\n"
+ " --volatile[=MODE] Run the system in volatile mode\n"
+ " --settings=BOOLEAN Load additional settings from .nspawn file\n"
+ " --notify-ready=BOOLEAN Receive notifications from the child init process\n"
+ , program_invocation_short_name);
+}
+
+static int custom_mounts_prepare(void) {
+ unsigned i;
+ int r;
+
+ /* Ensure the mounts are applied prefix first. */
+ qsort_safe(arg_custom_mounts, arg_n_custom_mounts, sizeof(CustomMount), custom_mount_compare);
+
+ /* Allocate working directories for the overlay file systems that need it */
+ for (i = 0; i < arg_n_custom_mounts; i++) {
+ CustomMount *m = &arg_custom_mounts[i];
+
+ if (path_equal(m->destination, "/") && arg_userns_mode != USER_NAMESPACE_NO) {
+
+ if (arg_userns_chown) {
+ log_error("--private-users-chown may not be combined with custom root mounts.");
+ return -EINVAL;
+ } else if (arg_uid_shift == UID_INVALID) {
+ log_error("--private-users with automatic UID shift may not be combined with custom root mounts.");
+ return -EINVAL;
+ }
+ }
+
+ if (m->type != CUSTOM_MOUNT_OVERLAY)
+ continue;
+
+ if (m->work_dir)
+ continue;
+
+ if (m->read_only)
+ continue;
+
+ r = tempfn_random(m->source, NULL, &m->work_dir);
+ if (r < 0)
+ return log_error_errno(r, "Failed to generate work directory from %s: %m", m->source);
+ }
+
+ return 0;
+}
+
+static int detect_unified_cgroup_hierarchy(const char *directory) {
+ const char *e;
+ int r, all_unified, systemd_unified;
+
+ /* Allow the user to control whether the unified hierarchy is used */
+ e = getenv("UNIFIED_CGROUP_HIERARCHY");
+ if (e) {
+ r = parse_boolean(e);
+ if (r < 0)
+ return log_error_errno(r, "Failed to parse $UNIFIED_CGROUP_HIERARCHY.");
+ if (r > 0)
+ arg_unified_cgroup_hierarchy = CGROUP_UNIFIED_ALL;
+ else
+ arg_unified_cgroup_hierarchy = CGROUP_UNIFIED_NONE;
+
+ return 0;
+ }
+
+ all_unified = cg_all_unified();
+ systemd_unified = cg_unified(SYSTEMD_CGROUP_CONTROLLER);
+
+ if (all_unified < 0 || systemd_unified < 0)
+ return log_error_errno(all_unified < 0 ? all_unified : systemd_unified,
+ "Failed to determine whether the unified cgroups hierarchy is used: %m");
+
+ /* Otherwise inherit the default from the host system */
+ if (all_unified > 0) {
+ /* Unified cgroup hierarchy support was added in 230. Unfortunately the detection
+ * routine only detects 231, so we'll have a false negative here for 230. */
+ r = systemd_installation_has_version(directory, 230);
+ if (r < 0)
+ return log_error_errno(r, "Failed to determine systemd version in container: %m");
+ if (r > 0)
+ arg_unified_cgroup_hierarchy = CGROUP_UNIFIED_ALL;
+ else
+ arg_unified_cgroup_hierarchy = CGROUP_UNIFIED_NONE;
+ } else if (systemd_unified > 0) {
+ /* Mixed cgroup hierarchy support was added in 232 */
+ r = systemd_installation_has_version(directory, 232);
+ if (r < 0)
+ return log_error_errno(r, "Failed to determine systemd version in container: %m");
+ if (r > 0)
+ arg_unified_cgroup_hierarchy = CGROUP_UNIFIED_SYSTEMD;
+ else
+ arg_unified_cgroup_hierarchy = CGROUP_UNIFIED_NONE;
+ } else
+ arg_unified_cgroup_hierarchy = CGROUP_UNIFIED_NONE;
+
+ return 0;
+}
+
+static void parse_share_ns_env(const char *name, unsigned long ns_flag) {
+ int r;
+
+ r = getenv_bool(name);
+ if (r == -ENXIO)
+ return;
+ if (r < 0)
+ log_warning_errno(r, "Failed to parse %s from environment, defaulting to false.", name);
+ arg_clone_ns_flags = (arg_clone_ns_flags & ~ns_flag) | (r > 0 ? 0 : ns_flag);
+}
+
+static int parse_argv(int argc, char *argv[]) {
+
+ enum {
+ ARG_VERSION = 0x100,
+ ARG_PRIVATE_NETWORK,
+ ARG_UUID,
+ ARG_READ_ONLY,
+ ARG_CAPABILITY,
+ ARG_DROP_CAPABILITY,
+ ARG_LINK_JOURNAL,
+ ARG_BIND,
+ ARG_BIND_RO,
+ ARG_TMPFS,
+ ARG_OVERLAY,
+ ARG_OVERLAY_RO,
+ ARG_SHARE_SYSTEM,
+ ARG_REGISTER,
+ ARG_KEEP_UNIT,
+ ARG_NETWORK_INTERFACE,
+ ARG_NETWORK_MACVLAN,
+ ARG_NETWORK_IPVLAN,
+ ARG_NETWORK_BRIDGE,
+ ARG_NETWORK_ZONE,
+ ARG_NETWORK_VETH_EXTRA,
+ ARG_PERSONALITY,
+ ARG_VOLATILE,
+ ARG_TEMPLATE,
+ ARG_PROPERTY,
+ ARG_PRIVATE_USERS,
+ ARG_KILL_SIGNAL,
+ ARG_SETTINGS,
+ ARG_CHDIR,
+ ARG_PRIVATE_USERS_CHOWN,
+ ARG_NOTIFY_READY,
+ };
+
+ static const struct option options[] = {
+ { "help", no_argument, NULL, 'h' },
+ { "version", no_argument, NULL, ARG_VERSION },
+ { "directory", required_argument, NULL, 'D' },
+ { "template", required_argument, NULL, ARG_TEMPLATE },
+ { "ephemeral", no_argument, NULL, 'x' },
+ { "user", required_argument, NULL, 'u' },
+ { "private-network", no_argument, NULL, ARG_PRIVATE_NETWORK },
+ { "as-pid2", no_argument, NULL, 'a' },
+ { "boot", no_argument, NULL, 'b' },
+ { "uuid", required_argument, NULL, ARG_UUID },
+ { "read-only", no_argument, NULL, ARG_READ_ONLY },
+ { "capability", required_argument, NULL, ARG_CAPABILITY },
+ { "drop-capability", required_argument, NULL, ARG_DROP_CAPABILITY },
+ { "link-journal", required_argument, NULL, ARG_LINK_JOURNAL },
+ { "bind", required_argument, NULL, ARG_BIND },
+ { "bind-ro", required_argument, NULL, ARG_BIND_RO },
+ { "tmpfs", required_argument, NULL, ARG_TMPFS },
+ { "overlay", required_argument, NULL, ARG_OVERLAY },
+ { "overlay-ro", required_argument, NULL, ARG_OVERLAY_RO },
+ { "machine", required_argument, NULL, 'M' },
+ { "slice", required_argument, NULL, 'S' },
+ { "setenv", required_argument, NULL, 'E' },
+ { "selinux-context", required_argument, NULL, 'Z' },
+ { "selinux-apifs-context", required_argument, NULL, 'L' },
+ { "quiet", no_argument, NULL, 'q' },
+ { "share-system", no_argument, NULL, ARG_SHARE_SYSTEM }, /* not documented */
+ { "register", required_argument, NULL, ARG_REGISTER },
+ { "keep-unit", no_argument, NULL, ARG_KEEP_UNIT },
+ { "network-interface", required_argument, NULL, ARG_NETWORK_INTERFACE },
+ { "network-macvlan", required_argument, NULL, ARG_NETWORK_MACVLAN },
+ { "network-ipvlan", required_argument, NULL, ARG_NETWORK_IPVLAN },
+ { "network-veth", no_argument, NULL, 'n' },
+ { "network-veth-extra", required_argument, NULL, ARG_NETWORK_VETH_EXTRA },
+ { "network-bridge", required_argument, NULL, ARG_NETWORK_BRIDGE },
+ { "network-zone", required_argument, NULL, ARG_NETWORK_ZONE },
+ { "personality", required_argument, NULL, ARG_PERSONALITY },
+ { "image", required_argument, NULL, 'i' },
+ { "volatile", optional_argument, NULL, ARG_VOLATILE },
+ { "port", required_argument, NULL, 'p' },
+ { "property", required_argument, NULL, ARG_PROPERTY },
+ { "private-users", optional_argument, NULL, ARG_PRIVATE_USERS },
+ { "private-users-chown", optional_argument, NULL, ARG_PRIVATE_USERS_CHOWN },
+ { "kill-signal", required_argument, NULL, ARG_KILL_SIGNAL },
+ { "settings", required_argument, NULL, ARG_SETTINGS },
+ { "chdir", required_argument, NULL, ARG_CHDIR },
+ { "notify-ready", required_argument, NULL, ARG_NOTIFY_READY },
+ {}
+ };
+
+ int c, r;
+ const char *p, *e;
+ uint64_t plus = 0, minus = 0;
+ bool mask_all_settings = false, mask_no_settings = false;
+
+ assert(argc >= 0);
+ assert(argv);
+
+ while ((c = getopt_long(argc, argv, "+hD:u:abL:M:jS:Z:qi:xp:nU", options, NULL)) >= 0)
+
+ switch (c) {
+
+ case 'h':
+ help();
+ return 0;
+
+ case ARG_VERSION:
+ return version();
+
+ case 'D':
+ r = parse_path_argument_and_warn(optarg, false, &arg_directory);
+ if (r < 0)
+ return r;
+ break;
+
+ case ARG_TEMPLATE:
+ r = parse_path_argument_and_warn(optarg, false, &arg_template);
+ if (r < 0)
+ return r;
+ break;
+
+ case 'i':
+ r = parse_path_argument_and_warn(optarg, false, &arg_image);
+ if (r < 0)
+ return r;
+ break;
+
+ case 'x':
+ arg_ephemeral = true;
+ break;
+
+ case 'u':
+ r = free_and_strdup(&arg_user, optarg);
+ if (r < 0)
+ return log_oom();
+
+ arg_settings_mask |= SETTING_USER;
+ break;
+
+ case ARG_NETWORK_ZONE: {
+ char *j;
+
+ j = strappend("vz-", optarg);
+ if (!j)
+ return log_oom();
+
+ if (!ifname_valid(j)) {
+ log_error("Network zone name not valid: %s", j);
+ free(j);
+ return -EINVAL;
+ }
+
+ free(arg_network_zone);
+ arg_network_zone = j;
+
+ arg_network_veth = true;
+ arg_private_network = true;
+ arg_settings_mask |= SETTING_NETWORK;
+ break;
+ }
+
+ case ARG_NETWORK_BRIDGE:
+
+ if (!ifname_valid(optarg)) {
+ log_error("Bridge interface name not valid: %s", optarg);
+ return -EINVAL;
+ }
+
+ r = free_and_strdup(&arg_network_bridge, optarg);
+ if (r < 0)
+ return log_oom();
+
+ /* fall through */
+
+ case 'n':
+ arg_network_veth = true;
+ arg_private_network = true;
+ arg_settings_mask |= SETTING_NETWORK;
+ break;
+
+ case ARG_NETWORK_VETH_EXTRA:
+ r = veth_extra_parse(&arg_network_veth_extra, optarg);
+ if (r < 0)
+ return log_error_errno(r, "Failed to parse --network-veth-extra= parameter: %s", optarg);
+
+ arg_private_network = true;
+ arg_settings_mask |= SETTING_NETWORK;
+ break;
+
+ case ARG_NETWORK_INTERFACE:
+
+ if (!ifname_valid(optarg)) {
+ log_error("Network interface name not valid: %s", optarg);
+ return -EINVAL;
+ }
+
+ if (strv_extend(&arg_network_interfaces, optarg) < 0)
+ return log_oom();
+
+ arg_private_network = true;
+ arg_settings_mask |= SETTING_NETWORK;
+ break;
+
+ case ARG_NETWORK_MACVLAN:
+
+ if (!ifname_valid(optarg)) {
+ log_error("MACVLAN network interface name not valid: %s", optarg);
+ return -EINVAL;
+ }
+
+ if (strv_extend(&arg_network_macvlan, optarg) < 0)
+ return log_oom();
+
+ arg_private_network = true;
+ arg_settings_mask |= SETTING_NETWORK;
+ break;
+
+ case ARG_NETWORK_IPVLAN:
+
+ if (!ifname_valid(optarg)) {
+ log_error("IPVLAN network interface name not valid: %s", optarg);
+ return -EINVAL;
+ }
+
+ if (strv_extend(&arg_network_ipvlan, optarg) < 0)
+ return log_oom();
+
+ /* fall through */
+
+ case ARG_PRIVATE_NETWORK:
+ arg_private_network = true;
+ arg_settings_mask |= SETTING_NETWORK;
+ break;
+
+ case 'b':
+ if (arg_start_mode == START_PID2) {
+ log_error("--boot and --as-pid2 may not be combined.");
+ return -EINVAL;
+ }
+
+ arg_start_mode = START_BOOT;
+ arg_settings_mask |= SETTING_START_MODE;
+ break;
+
+ case 'a':
+ if (arg_start_mode == START_BOOT) {
+ log_error("--boot and --as-pid2 may not be combined.");
+ return -EINVAL;
+ }
+
+ arg_start_mode = START_PID2;
+ arg_settings_mask |= SETTING_START_MODE;
+ break;
+
+ case ARG_UUID:
+ r = sd_id128_from_string(optarg, &arg_uuid);
+ if (r < 0)
+ return log_error_errno(r, "Invalid UUID: %s", optarg);
+
+ if (sd_id128_is_null(arg_uuid)) {
+ log_error("Machine UUID may not be all zeroes.");
+ return -EINVAL;
+ }
+
+ arg_settings_mask |= SETTING_MACHINE_ID;
+ break;
+
+ case 'S':
+ arg_slice = optarg;
+ break;
+
+ case 'M':
+ if (isempty(optarg))
+ arg_machine = mfree(arg_machine);
+ else {
+ if (!machine_name_is_valid(optarg)) {
+ log_error("Invalid machine name: %s", optarg);
+ return -EINVAL;
+ }
+
+ r = free_and_strdup(&arg_machine, optarg);
+ if (r < 0)
+ return log_oom();
+
+ break;
+ }
+
+ case 'Z':
+ arg_selinux_context = optarg;
+ break;
+
+ case 'L':
+ arg_selinux_apifs_context = optarg;
+ break;
+
+ case ARG_READ_ONLY:
+ arg_read_only = true;
+ arg_settings_mask |= SETTING_READ_ONLY;
+ break;
+
+ case ARG_CAPABILITY:
+ case ARG_DROP_CAPABILITY: {
+ p = optarg;
+ for (;;) {
+ _cleanup_free_ char *t = NULL;
+
+ r = extract_first_word(&p, &t, ",", 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to parse capability %s.", t);
+
+ if (r == 0)
+ break;
+
+ if (streq(t, "all")) {
+ if (c == ARG_CAPABILITY)
+ plus = (uint64_t) -1;
+ else
+ minus = (uint64_t) -1;
+ } else {
+ int cap;
+
+ cap = capability_from_name(t);
+ if (cap < 0) {
+ log_error("Failed to parse capability %s.", t);
+ return -EINVAL;
+ }
+
+ if (c == ARG_CAPABILITY)
+ plus |= 1ULL << (uint64_t) cap;
+ else
+ minus |= 1ULL << (uint64_t) cap;
+ }
+ }
+
+ arg_settings_mask |= SETTING_CAPABILITY;
+ break;
+ }
+
+ case 'j':
+ arg_link_journal = LINK_GUEST;
+ arg_link_journal_try = true;
+ break;
+
+ case ARG_LINK_JOURNAL:
+ if (streq(optarg, "auto")) {
+ arg_link_journal = LINK_AUTO;
+ arg_link_journal_try = false;
+ } else if (streq(optarg, "no")) {
+ arg_link_journal = LINK_NO;
+ arg_link_journal_try = false;
+ } else if (streq(optarg, "guest")) {
+ arg_link_journal = LINK_GUEST;
+ arg_link_journal_try = false;
+ } else if (streq(optarg, "host")) {
+ arg_link_journal = LINK_HOST;
+ arg_link_journal_try = false;
+ } else if (streq(optarg, "try-guest")) {
+ arg_link_journal = LINK_GUEST;
+ arg_link_journal_try = true;
+ } else if (streq(optarg, "try-host")) {
+ arg_link_journal = LINK_HOST;
+ arg_link_journal_try = true;
+ } else {
+ log_error("Failed to parse link journal mode %s", optarg);
+ return -EINVAL;
+ }
+
+ break;
+
+ case ARG_BIND:
+ case ARG_BIND_RO:
+ r = bind_mount_parse(&arg_custom_mounts, &arg_n_custom_mounts, optarg, c == ARG_BIND_RO);
+ if (r < 0)
+ return log_error_errno(r, "Failed to parse --bind(-ro)= argument %s: %m", optarg);
+
+ arg_settings_mask |= SETTING_CUSTOM_MOUNTS;
+ break;
+
+ case ARG_TMPFS:
+ r = tmpfs_mount_parse(&arg_custom_mounts, &arg_n_custom_mounts, optarg);
+ if (r < 0)
+ return log_error_errno(r, "Failed to parse --tmpfs= argument %s: %m", optarg);
+
+ arg_settings_mask |= SETTING_CUSTOM_MOUNTS;
+ break;
+
+ case ARG_OVERLAY:
+ case ARG_OVERLAY_RO: {
+ _cleanup_free_ char *upper = NULL, *destination = NULL;
+ _cleanup_strv_free_ char **lower = NULL;
+ CustomMount *m;
+ unsigned n = 0;
+ char **i;
+
+ r = strv_split_extract(&lower, optarg, ":", EXTRACT_DONT_COALESCE_SEPARATORS);
+ if (r == -ENOMEM)
+ return log_oom();
+ else if (r < 0) {
+ log_error("Invalid overlay specification: %s", optarg);
+ return r;
+ }
+
+ STRV_FOREACH(i, lower) {
+ if (!path_is_absolute(*i)) {
+ log_error("Overlay path %s is not absolute.", *i);
+ return -EINVAL;
+ }
+
+ n++;
+ }
+
+ if (n < 2) {
+ log_error("--overlay= needs at least two colon-separated directories specified.");
+ return -EINVAL;
+ }
+
+ if (n == 2) {
+ /* If two parameters are specified,
+ * the first one is the lower, the
+ * second one the upper directory. And
+ * we'll also define the destination
+ * mount point the same as the upper. */
+ upper = lower[1];
+ lower[1] = NULL;
+
+ destination = strdup(upper);
+ if (!destination)
+ return log_oom();
+
+ } else {
+ upper = lower[n - 2];
+ destination = lower[n - 1];
+ lower[n - 2] = NULL;
+ }
+
+ m = custom_mount_add(&arg_custom_mounts, &arg_n_custom_mounts, CUSTOM_MOUNT_OVERLAY);
+ if (!m)
+ return log_oom();
+
+ m->destination = destination;
+ m->source = upper;
+ m->lower = lower;
+ m->read_only = c == ARG_OVERLAY_RO;
+
+ upper = destination = NULL;
+ lower = NULL;
+
+ arg_settings_mask |= SETTING_CUSTOM_MOUNTS;
+ break;
+ }
+
+ case 'E': {
+ char **n;
+
+ if (!env_assignment_is_valid(optarg)) {
+ log_error("Environment variable assignment '%s' is not valid.", optarg);
+ return -EINVAL;
+ }
+
+ n = strv_env_set(arg_setenv, optarg);
+ if (!n)
+ return log_oom();
+
+ strv_free(arg_setenv);
+ arg_setenv = n;
+
+ arg_settings_mask |= SETTING_ENVIRONMENT;
+ break;
+ }
+
+ case 'q':
+ arg_quiet = true;
+ break;
+
+ case ARG_SHARE_SYSTEM:
+ /* We don't officially support this anymore, except for compat reasons. People should use the
+ * $SYSTEMD_NSPAWN_SHARE_* environment variables instead. */
+ arg_clone_ns_flags = 0;
+ break;
+
+ case ARG_REGISTER:
+ r = parse_boolean(optarg);
+ if (r < 0) {
+ log_error("Failed to parse --register= argument: %s", optarg);
+ return r;
+ }
+
+ arg_register = r;
+ break;
+
+ case ARG_KEEP_UNIT:
+ arg_keep_unit = true;
+ break;
+
+ case ARG_PERSONALITY:
+
+ arg_personality = personality_from_string(optarg);
+ if (arg_personality == PERSONALITY_INVALID) {
+ log_error("Unknown or unsupported personality '%s'.", optarg);
+ return -EINVAL;
+ }
+
+ arg_settings_mask |= SETTING_PERSONALITY;
+ break;
+
+ case ARG_VOLATILE:
+
+ if (!optarg)
+ arg_volatile_mode = VOLATILE_YES;
+ else {
+ VolatileMode m;
+
+ m = volatile_mode_from_string(optarg);
+ if (m < 0) {
+ log_error("Failed to parse --volatile= argument: %s", optarg);
+ return -EINVAL;
+ } else
+ arg_volatile_mode = m;
+ }
+
+ arg_settings_mask |= SETTING_VOLATILE_MODE;
+ break;
+
+ case 'p':
+ r = expose_port_parse(&arg_expose_ports, optarg);
+ if (r == -EEXIST)
+ return log_error_errno(r, "Duplicate port specification: %s", optarg);
+ if (r < 0)
+ return log_error_errno(r, "Failed to parse host port %s: %m", optarg);
+
+ arg_settings_mask |= SETTING_EXPOSE_PORTS;
+ break;
+
+ case ARG_PROPERTY:
+ if (strv_extend(&arg_property, optarg) < 0)
+ return log_oom();
+
+ break;
+
+ case ARG_PRIVATE_USERS: {
+ int boolean = -1;
+
+ if (!optarg)
+ boolean = true;
+ else if (!in_charset(optarg, DIGITS))
+ /* do *not* parse numbers as booleans */
+ boolean = parse_boolean(optarg);
+
+ if (boolean == false) {
+ /* no: User namespacing off */
+ arg_userns_mode = USER_NAMESPACE_NO;
+ arg_uid_shift = UID_INVALID;
+ arg_uid_range = UINT32_C(0x10000);
+ } else if (boolean == true) {
+ /* yes: User namespacing on, UID range is read from root dir */
+ arg_userns_mode = USER_NAMESPACE_FIXED;
+ arg_uid_shift = UID_INVALID;
+ arg_uid_range = UINT32_C(0x10000);
+ } else if (streq(optarg, "pick")) {
+ /* pick: User namespacing on, UID range is picked randomly */
+ arg_userns_mode = USER_NAMESPACE_PICK;
+ arg_uid_shift = UID_INVALID;
+ arg_uid_range = UINT32_C(0x10000);
+ } else {
+ _cleanup_free_ char *buffer = NULL;
+ const char *range, *shift;
+
+ /* anything else: User namespacing on, UID range is explicitly configured */
+
+ range = strchr(optarg, ':');
+ if (range) {
+ buffer = strndup(optarg, range - optarg);
+ if (!buffer)
+ return log_oom();
+ shift = buffer;
+
+ range++;
+ r = safe_atou32(range, &arg_uid_range);
+ if (r < 0)
+ return log_error_errno(r, "Failed to parse UID range \"%s\": %m", range);
+ } else
+ shift = optarg;
+
+ r = parse_uid(shift, &arg_uid_shift);
+ if (r < 0)
+ return log_error_errno(r, "Failed to parse UID \"%s\": %m", optarg);
+
+ arg_userns_mode = USER_NAMESPACE_FIXED;
+ }
+
+ if (arg_uid_range <= 0) {
+ log_error("UID range cannot be 0.");
+ return -EINVAL;
+ }
+
+ arg_settings_mask |= SETTING_USERNS;
+ break;
+ }
+
+ case 'U':
+ if (userns_supported()) {
+ arg_userns_mode = USER_NAMESPACE_PICK;
+ arg_uid_shift = UID_INVALID;
+ arg_uid_range = UINT32_C(0x10000);
+
+ arg_settings_mask |= SETTING_USERNS;
+ }
+
+ break;
+
+ case ARG_PRIVATE_USERS_CHOWN:
+ arg_userns_chown = true;
+
+ arg_settings_mask |= SETTING_USERNS;
+ break;
+
+ case ARG_KILL_SIGNAL:
+ arg_kill_signal = signal_from_string_try_harder(optarg);
+ if (arg_kill_signal < 0) {
+ log_error("Cannot parse signal: %s", optarg);
+ return -EINVAL;
+ }
+
+ arg_settings_mask |= SETTING_KILL_SIGNAL;
+ break;
+
+ case ARG_SETTINGS:
+
+ /* no → do not read files
+ * yes → read files, do not override cmdline, trust only subset
+ * override → read files, override cmdline, trust only subset
+ * trusted → read files, do not override cmdline, trust all
+ */
+
+ r = parse_boolean(optarg);
+ if (r < 0) {
+ if (streq(optarg, "trusted")) {
+ mask_all_settings = false;
+ mask_no_settings = false;
+ arg_settings_trusted = true;
+
+ } else if (streq(optarg, "override")) {
+ mask_all_settings = false;
+ mask_no_settings = true;
+ arg_settings_trusted = -1;
+ } else
+ return log_error_errno(r, "Failed to parse --settings= argument: %s", optarg);
+ } else if (r > 0) {
+ /* yes */
+ mask_all_settings = false;
+ mask_no_settings = false;
+ arg_settings_trusted = -1;
+ } else {
+ /* no */
+ mask_all_settings = true;
+ mask_no_settings = false;
+ arg_settings_trusted = false;
+ }
+
+ break;
+
+ case ARG_CHDIR:
+ if (!path_is_absolute(optarg)) {
+ log_error("Working directory %s is not an absolute path.", optarg);
+ return -EINVAL;
+ }
+
+ r = free_and_strdup(&arg_chdir, optarg);
+ if (r < 0)
+ return log_oom();
+
+ arg_settings_mask |= SETTING_WORKING_DIRECTORY;
+ break;
+
+ case ARG_NOTIFY_READY:
+ r = parse_boolean(optarg);
+ if (r < 0) {
+ log_error("%s is not a valid notify mode. Valid modes are: yes, no, and ready.", optarg);
+ return -EINVAL;
+ }
+ arg_notify_ready = r;
+ arg_settings_mask |= SETTING_NOTIFY_READY;
+ break;
+
+ case '?':
+ return -EINVAL;
+
+ default:
+ assert_not_reached("Unhandled option");
+ }
+
+ parse_share_ns_env("SYSTEMD_NSPAWN_SHARE_NS_IPC", CLONE_NEWIPC);
+ parse_share_ns_env("SYSTEMD_NSPAWN_SHARE_NS_PID", CLONE_NEWPID);
+ parse_share_ns_env("SYSTEMD_NSPAWN_SHARE_NS_UTS", CLONE_NEWUTS);
+ parse_share_ns_env("SYSTEMD_NSPAWN_SHARE_SYSTEM", CLONE_NEWIPC|CLONE_NEWPID|CLONE_NEWUTS);
+
+ if (!(arg_clone_ns_flags & CLONE_NEWPID) ||
+ !(arg_clone_ns_flags & CLONE_NEWUTS)) {
+ arg_register = false;
+ if (arg_start_mode != START_PID1) {
+ log_error("--boot cannot be used without namespacing.");
+ return -EINVAL;
+ }
+ }
+
+ if (arg_userns_mode == USER_NAMESPACE_PICK)
+ arg_userns_chown = true;
+
+ if (arg_keep_unit && cg_pid_get_owner_uid(0, NULL) >= 0) {
+ log_error("--keep-unit may not be used when invoked from a user session.");
+ return -EINVAL;
+ }
+
+ if (arg_directory && arg_image) {
+ log_error("--directory= and --image= may not be combined.");
+ return -EINVAL;
+ }
+
+ if (arg_template && arg_image) {
+ log_error("--template= and --image= may not be combined.");
+ return -EINVAL;
+ }
+
+ if (arg_template && !(arg_directory || arg_machine)) {
+ log_error("--template= needs --directory= or --machine=.");
+ return -EINVAL;
+ }
+
+ if (arg_ephemeral && arg_template) {
+ log_error("--ephemeral and --template= may not be combined.");
+ return -EINVAL;
+ }
+
+ if (arg_ephemeral && arg_image) {
+ log_error("--ephemeral and --image= may not be combined.");
+ return -EINVAL;
+ }
+
+ if (arg_ephemeral && !IN_SET(arg_link_journal, LINK_NO, LINK_AUTO)) {
+ log_error("--ephemeral and --link-journal= may not be combined.");
+ return -EINVAL;
+ }
+
+ if (arg_userns_mode != USER_NAMESPACE_NO && !userns_supported()) {
+ log_error("--private-users= is not supported, kernel compiled without user namespace support.");
+ return -EOPNOTSUPP;
+ }
+
+ if (arg_userns_chown && arg_read_only) {
+ log_error("--read-only and --private-users-chown may not be combined.");
+ return -EINVAL;
+ }
+
+ if (arg_network_bridge && arg_network_zone) {
+ log_error("--network-bridge= and --network-zone= may not be combined.");
+ return -EINVAL;
+ }
+
+ if (argc > optind) {
+ arg_parameters = strv_copy(argv + optind);
+ if (!arg_parameters)
+ return log_oom();
+
+ arg_settings_mask |= SETTING_START_MODE;
+ }
+
+ /* Load all settings from .nspawn files */
+ if (mask_no_settings)
+ arg_settings_mask = 0;
+
+ /* Don't load any settings from .nspawn files */
+ if (mask_all_settings)
+ arg_settings_mask = _SETTINGS_MASK_ALL;
+
+ arg_caps_retain = (arg_caps_retain | plus | (arg_private_network ? 1ULL << CAP_NET_ADMIN : 0)) & ~minus;
+
+ e = getenv("SYSTEMD_NSPAWN_CONTAINER_SERVICE");
+ if (e)
+ arg_container_service_name = e;
+
+ r = getenv_bool("SYSTEMD_NSPAWN_USE_CGNS");
+ if (r < 0)
+ arg_use_cgns = cg_ns_supported();
+ else
+ arg_use_cgns = r;
+
+ return 1;
+}
+
+static int verify_arguments(void) {
+
+ if (arg_volatile_mode != VOLATILE_NO && arg_read_only) {
+ log_error("Cannot combine --read-only with --volatile. Note that --volatile already implies a read-only base hierarchy.");
+ return -EINVAL;
+ }
+
+ if (arg_expose_ports && !arg_private_network) {
+ log_error("Cannot use --port= without private networking.");
+ return -EINVAL;
+ }
+
+#ifndef HAVE_LIBIPTC
+ if (arg_expose_ports) {
+ log_error("--port= is not supported, compiled without libiptc support.");
+ return -EOPNOTSUPP;
+ }
+#endif
+
+ if (arg_start_mode == START_BOOT && arg_kill_signal <= 0)
+ arg_kill_signal = SIGRTMIN+3;
+
+ return 0;
+}
+
+static int userns_lchown(const char *p, uid_t uid, gid_t gid) {
+ assert(p);
+
+ if (arg_userns_mode == USER_NAMESPACE_NO)
+ return 0;
+
+ if (uid == UID_INVALID && gid == GID_INVALID)
+ return 0;
+
+ if (uid != UID_INVALID) {
+ uid += arg_uid_shift;
+
+ if (uid < arg_uid_shift || uid >= arg_uid_shift + arg_uid_range)
+ return -EOVERFLOW;
+ }
+
+ if (gid != GID_INVALID) {
+ gid += (gid_t) arg_uid_shift;
+
+ if (gid < (gid_t) arg_uid_shift || gid >= (gid_t) (arg_uid_shift + arg_uid_range))
+ return -EOVERFLOW;
+ }
+
+ if (lchown(p, uid, gid) < 0)
+ return -errno;
+
+ return 0;
+}
+
+static int userns_mkdir(const char *root, const char *path, mode_t mode, uid_t uid, gid_t gid) {
+ const char *q;
+
+ q = prefix_roota(root, path);
+ if (mkdir(q, mode) < 0) {
+ if (errno == EEXIST)
+ return 0;
+ return -errno;
+ }
+
+ return userns_lchown(q, uid, gid);
+}
+
+static int setup_timezone(const char *dest) {
+ _cleanup_free_ char *p = NULL, *q = NULL;
+ const char *where, *check, *what;
+ char *z, *y;
+ int r;
+
+ assert(dest);
+
+ /* Fix the timezone, if possible */
+ r = readlink_malloc("/etc/localtime", &p);
+ if (r < 0) {
+ log_warning("host's /etc/localtime is not a symlink, not updating container timezone.");
+ /* to handle warning, delete /etc/localtime and replace it
+ * with a symbolic link to a time zone data file.
+ *
+ * Example:
+ * ln -s /usr/share/zoneinfo/UTC /etc/localtime
+ */
+ return 0;
+ }
+
+ z = path_startswith(p, "../usr/share/zoneinfo/");
+ if (!z)
+ z = path_startswith(p, "/usr/share/zoneinfo/");
+ if (!z) {
+ log_warning("/etc/localtime does not point into /usr/share/zoneinfo/, not updating container timezone.");
+ return 0;
+ }
+
+ where = prefix_roota(dest, "/etc/localtime");
+ r = readlink_malloc(where, &q);
+ if (r >= 0) {
+ y = path_startswith(q, "../usr/share/zoneinfo/");
+ if (!y)
+ y = path_startswith(q, "/usr/share/zoneinfo/");
+
+ /* Already pointing to the right place? Then do nothing .. */
+ if (y && streq(y, z))
+ return 0;
+ }
+
+ check = strjoina("/usr/share/zoneinfo/", z);
+ check = prefix_roota(dest, check);
+ if (laccess(check, F_OK) < 0) {
+ log_warning("Timezone %s does not exist in container, not updating container timezone.", z);
+ return 0;
+ }
+
+ r = unlink(where);
+ if (r < 0 && errno != ENOENT) {
+ log_error_errno(errno, "Failed to remove existing timezone info %s in container: %m", where);
+ return 0;
+ }
+
+ what = strjoina("../usr/share/zoneinfo/", z);
+ if (symlink(what, where) < 0) {
+ log_error_errno(errno, "Failed to correct timezone of container: %m");
+ return 0;
+ }
+
+ r = userns_lchown(where, 0, 0);
+ if (r < 0)
+ return log_warning_errno(r, "Failed to chown /etc/localtime: %m");
+
+ return 0;
+}
+
+static int setup_resolv_conf(const char *dest) {
+ const char *where = NULL;
+ int r;
+
+ assert(dest);
+
+ if (arg_private_network)
+ return 0;
+
+ /* Fix resolv.conf, if possible */
+ where = prefix_roota(dest, "/etc/resolv.conf");
+
+ r = copy_file("/etc/resolv.conf", where, O_TRUNC|O_NOFOLLOW, 0644, 0);
+ if (r < 0) {
+ /* If the file already exists as symlink, let's
+ * suppress the warning, under the assumption that
+ * resolved or something similar runs inside and the
+ * symlink points there.
+ *
+ * If the disk image is read-only, there's also no
+ * point in complaining.
+ */
+ log_full_errno(IN_SET(r, -ELOOP, -EROFS) ? LOG_DEBUG : LOG_WARNING, r,
+ "Failed to copy /etc/resolv.conf to %s: %m", where);
+ return 0;
+ }
+
+ r = userns_lchown(where, 0, 0);
+ if (r < 0)
+ log_warning_errno(r, "Failed to chown /etc/resolv.conf: %m");
+
+ return 0;
+}
+
+static int setup_boot_id(const char *dest) {
+ sd_id128_t rnd = SD_ID128_NULL;
+ const char *from, *to;
+ int r;
+
+ /* Generate a new randomized boot ID, so that each boot-up of
+ * the container gets a new one */
+
+ from = prefix_roota(dest, "/run/proc-sys-kernel-random-boot-id");
+ to = prefix_roota(dest, "/proc/sys/kernel/random/boot_id");
+
+ r = sd_id128_randomize(&rnd);
+ if (r < 0)
+ return log_error_errno(r, "Failed to generate random boot id: %m");
+
+ r = id128_write(from, ID128_UUID, rnd, false);
+ if (r < 0)
+ return log_error_errno(r, "Failed to write boot id: %m");
+
+ r = mount_verbose(LOG_ERR, from, to, NULL, MS_BIND, NULL);
+ if (r >= 0)
+ r = mount_verbose(LOG_ERR, NULL, to, NULL,
+ MS_BIND|MS_REMOUNT|MS_RDONLY|MS_NOSUID|MS_NODEV, NULL);
+
+ (void) unlink(from);
+ return r;
+}
+
+static int copy_devnodes(const char *dest) {
+
+ static const char devnodes[] =
+ "null\0"
+ "zero\0"
+ "full\0"
+ "random\0"
+ "urandom\0"
+ "tty\0"
+ "net/tun\0";
+
+ const char *d;
+ int r = 0;
+ _cleanup_umask_ mode_t u;
+
+ assert(dest);
+
+ u = umask(0000);
+
+ /* Create /dev/net, so that we can create /dev/net/tun in it */
+ if (userns_mkdir(dest, "/dev/net", 0755, 0, 0) < 0)
+ return log_error_errno(r, "Failed to create /dev/net directory: %m");
+
+ NULSTR_FOREACH(d, devnodes) {
+ _cleanup_free_ char *from = NULL, *to = NULL;
+ struct stat st;
+
+ from = strappend("/dev/", d);
+ to = prefix_root(dest, from);
+
+ if (stat(from, &st) < 0) {
+
+ if (errno != ENOENT)
+ return log_error_errno(errno, "Failed to stat %s: %m", from);
+
+ } else if (!S_ISCHR(st.st_mode) && !S_ISBLK(st.st_mode)) {
+
+ log_error("%s is not a char or block device, cannot copy.", from);
+ return -EIO;
+
+ } else {
+ if (mknod(to, st.st_mode, st.st_rdev) < 0) {
+ /*
+ * This is some sort of protection too against
+ * recursive userns chown on shared /dev/
+ */
+ if (errno == EEXIST)
+ log_notice("%s/dev/ should be an empty directory", dest);
+ if (errno != EPERM)
+ return log_error_errno(errno, "mknod(%s) failed: %m", to);
+
+ /* Some systems abusively restrict mknod but
+ * allow bind mounts. */
+ r = touch(to);
+ if (r < 0)
+ return log_error_errno(r, "touch (%s) failed: %m", to);
+ r = mount_verbose(LOG_DEBUG, from, to, NULL, MS_BIND, NULL);
+ if (r < 0)
+ return log_error_errno(r, "Both mknod and bind mount (%s) failed: %m", to);
+ }
+
+ r = userns_lchown(to, 0, 0);
+ if (r < 0)
+ return log_error_errno(r, "chown() of device node %s failed: %m", to);
+ }
+ }
+
+ return r;
+}
+
+static int setup_pts(const char *dest) {
+ _cleanup_free_ char *options = NULL;
+ const char *p;
+ int r;
+
+#ifdef HAVE_SELINUX
+ if (arg_selinux_apifs_context)
+ (void) asprintf(&options,
+ "newinstance,ptmxmode=0666,mode=620,gid=" GID_FMT ",context=\"%s\"",
+ arg_uid_shift + TTY_GID,
+ arg_selinux_apifs_context);
+ else
+#endif
+ (void) asprintf(&options,
+ "newinstance,ptmxmode=0666,mode=620,gid=" GID_FMT,
+ arg_uid_shift + TTY_GID);
+
+ if (!options)
+ return log_oom();
+
+ /* Mount /dev/pts itself */
+ p = prefix_roota(dest, "/dev/pts");
+ if (mkdir(p, 0755) < 0)
+ return log_error_errno(errno, "Failed to create /dev/pts: %m");
+ r = mount_verbose(LOG_ERR, "devpts", p, "devpts", MS_NOSUID|MS_NOEXEC, options);
+ if (r < 0)
+ return r;
+ r = userns_lchown(p, 0, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to chown /dev/pts: %m");
+
+ /* Create /dev/ptmx symlink */
+ p = prefix_roota(dest, "/dev/ptmx");
+ if (symlink("pts/ptmx", p) < 0)
+ return log_error_errno(errno, "Failed to create /dev/ptmx symlink: %m");
+ r = userns_lchown(p, 0, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to chown /dev/ptmx: %m");
+
+ /* And fix /dev/pts/ptmx ownership */
+ p = prefix_roota(dest, "/dev/pts/ptmx");
+ r = userns_lchown(p, 0, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to chown /dev/pts/ptmx: %m");
+
+ return 0;
+}
+
+static int setup_dev_console(const char *dest, const char *console) {
+ _cleanup_umask_ mode_t u;
+ const char *to;
+ int r;
+
+ assert(dest);
+ assert(console);
+
+ u = umask(0000);
+
+ r = chmod_and_chown(console, 0600, arg_uid_shift, arg_uid_shift);
+ if (r < 0)
+ return log_error_errno(r, "Failed to correct access mode for TTY: %m");
+
+ /* We need to bind mount the right tty to /dev/console since
+ * ptys can only exist on pts file systems. To have something
+ * to bind mount things on we create a empty regular file. */
+
+ to = prefix_roota(dest, "/dev/console");
+ r = touch(to);
+ if (r < 0)
+ return log_error_errno(r, "touch() for /dev/console failed: %m");
+
+ return mount_verbose(LOG_ERR, console, to, NULL, MS_BIND, NULL);
+}
+
+static int setup_kmsg(const char *dest, int kmsg_socket) {
+ const char *from, *to;
+ _cleanup_umask_ mode_t u;
+ int fd, r;
+
+ assert(kmsg_socket >= 0);
+
+ u = umask(0000);
+
+ /* We create the kmsg FIFO as /run/kmsg, but immediately
+ * delete it after bind mounting it to /proc/kmsg. While FIFOs
+ * on the reading side behave very similar to /proc/kmsg,
+ * their writing side behaves differently from /dev/kmsg in
+ * that writing blocks when nothing is reading. In order to
+ * avoid any problems with containers deadlocking due to this
+ * we simply make /dev/kmsg unavailable to the container. */
+ from = prefix_roota(dest, "/run/kmsg");
+ to = prefix_roota(dest, "/proc/kmsg");
+
+ if (mkfifo(from, 0600) < 0)
+ return log_error_errno(errno, "mkfifo() for /run/kmsg failed: %m");
+ r = mount_verbose(LOG_ERR, from, to, NULL, MS_BIND, NULL);
+ if (r < 0)
+ return r;
+
+ fd = open(from, O_RDWR|O_NDELAY|O_CLOEXEC);
+ if (fd < 0)
+ return log_error_errno(errno, "Failed to open fifo: %m");
+
+ /* Store away the fd in the socket, so that it stays open as
+ * long as we run the child */
+ r = send_one_fd(kmsg_socket, fd, 0);
+ safe_close(fd);
+
+ if (r < 0)
+ return log_error_errno(r, "Failed to send FIFO fd: %m");
+
+ /* And now make the FIFO unavailable as /run/kmsg... */
+ (void) unlink(from);
+
+ return 0;
+}
+
+static int on_address_change(sd_netlink *rtnl, sd_netlink_message *m, void *userdata) {
+ union in_addr_union *exposed = userdata;
+
+ assert(rtnl);
+ assert(m);
+ assert(exposed);
+
+ expose_port_execute(rtnl, arg_expose_ports, exposed);
+ return 0;
+}
+
+static int setup_hostname(void) {
+
+ if ((arg_clone_ns_flags & CLONE_NEWUTS) == 0)
+ return 0;
+
+ if (sethostname_idempotent(arg_machine) < 0)
+ return -errno;
+
+ return 0;
+}
+
+static int setup_journal(const char *directory) {
+ sd_id128_t this_id;
+ _cleanup_free_ char *d = NULL;
+ const char *p, *q;
+ bool try;
+ char id[33];
+ int r;
+
+ /* Don't link journals in ephemeral mode */
+ if (arg_ephemeral)
+ return 0;
+
+ if (arg_link_journal == LINK_NO)
+ return 0;
+
+ try = arg_link_journal_try || arg_link_journal == LINK_AUTO;
+
+ r = sd_id128_get_machine(&this_id);
+ if (r < 0)
+ return log_error_errno(r, "Failed to retrieve machine ID: %m");
+
+ if (sd_id128_equal(arg_uuid, this_id)) {
+ log_full(try ? LOG_WARNING : LOG_ERR,
+ "Host and machine ids are equal (%s): refusing to link journals", sd_id128_to_string(arg_uuid, id));
+ if (try)
+ return 0;
+ return -EEXIST;
+ }
+
+ r = userns_mkdir(directory, "/var", 0755, 0, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to create /var: %m");
+
+ r = userns_mkdir(directory, "/var/log", 0755, 0, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to create /var/log: %m");
+
+ r = userns_mkdir(directory, "/var/log/journal", 0755, 0, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to create /var/log/journal: %m");
+
+ (void) sd_id128_to_string(arg_uuid, id);
+
+ p = strjoina("/var/log/journal/", id);
+ q = prefix_roota(directory, p);
+
+ if (path_is_mount_point(p, 0) > 0) {
+ if (try)
+ return 0;
+
+ log_error("%s: already a mount point, refusing to use for journal", p);
+ return -EEXIST;
+ }
+
+ if (path_is_mount_point(q, 0) > 0) {
+ if (try)
+ return 0;
+
+ log_error("%s: already a mount point, refusing to use for journal", q);
+ return -EEXIST;
+ }
+
+ r = readlink_and_make_absolute(p, &d);
+ if (r >= 0) {
+ if ((arg_link_journal == LINK_GUEST ||
+ arg_link_journal == LINK_AUTO) &&
+ path_equal(d, q)) {
+
+ r = userns_mkdir(directory, p, 0755, 0, 0);
+ if (r < 0)
+ log_warning_errno(r, "Failed to create directory %s: %m", q);
+ return 0;
+ }
+
+ if (unlink(p) < 0)
+ return log_error_errno(errno, "Failed to remove symlink %s: %m", p);
+ } else if (r == -EINVAL) {
+
+ if (arg_link_journal == LINK_GUEST &&
+ rmdir(p) < 0) {
+
+ if (errno == ENOTDIR) {
+ log_error("%s already exists and is neither a symlink nor a directory", p);
+ return r;
+ } else
+ return log_error_errno(errno, "Failed to remove %s: %m", p);
+ }
+ } else if (r != -ENOENT)
+ return log_error_errno(r, "readlink(%s) failed: %m", p);
+
+ if (arg_link_journal == LINK_GUEST) {
+
+ if (symlink(q, p) < 0) {
+ if (try) {
+ log_debug_errno(errno, "Failed to symlink %s to %s, skipping journal setup: %m", q, p);
+ return 0;
+ } else
+ return log_error_errno(errno, "Failed to symlink %s to %s: %m", q, p);
+ }
+
+ r = userns_mkdir(directory, p, 0755, 0, 0);
+ if (r < 0)
+ log_warning_errno(r, "Failed to create directory %s: %m", q);
+ return 0;
+ }
+
+ if (arg_link_journal == LINK_HOST) {
+ /* don't create parents here — if the host doesn't have
+ * permanent journal set up, don't force it here */
+
+ if (mkdir(p, 0755) < 0 && errno != EEXIST) {
+ if (try) {
+ log_debug_errno(errno, "Failed to create %s, skipping journal setup: %m", p);
+ return 0;
+ } else
+ return log_error_errno(errno, "Failed to create %s: %m", p);
+ }
+
+ } else if (access(p, F_OK) < 0)
+ return 0;
+
+ if (dir_is_empty(q) == 0)
+ log_warning("%s is not empty, proceeding anyway.", q);
+
+ r = userns_mkdir(directory, p, 0755, 0, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to create %s: %m", q);
+
+ r = mount_verbose(LOG_DEBUG, p, q, NULL, MS_BIND, NULL);
+ if (r < 0)
+ return log_error_errno(errno, "Failed to bind mount journal from host into guest: %m");
+
+ return 0;
+}
+
+static int drop_capabilities(void) {
+ return capability_bounding_set_drop(arg_caps_retain, false);
+}
+
+static int reset_audit_loginuid(void) {
+ _cleanup_free_ char *p = NULL;
+ int r;
+
+ if ((arg_clone_ns_flags & CLONE_NEWPID) == 0)
+ return 0;
+
+ r = read_one_line_file("/proc/self/loginuid", &p);
+ if (r == -ENOENT)
+ return 0;
+ if (r < 0)
+ return log_error_errno(r, "Failed to read /proc/self/loginuid: %m");
+
+ /* Already reset? */
+ if (streq(p, "4294967295"))
+ return 0;
+
+ r = write_string_file("/proc/self/loginuid", "4294967295", 0);
+ if (r < 0) {
+ log_error_errno(r,
+ "Failed to reset audit login UID. This probably means that your kernel is too\n"
+ "old and you have audit enabled. Note that the auditing subsystem is known to\n"
+ "be incompatible with containers on old kernels. Please make sure to upgrade\n"
+ "your kernel or to off auditing with 'audit=0' on the kernel command line before\n"
+ "using systemd-nspawn. Sleeping for 5s... (%m)");
+
+ sleep(5);
+ }
+
+ return 0;
+}
+
+
+static int setup_propagate(const char *root) {
+ const char *p, *q;
+ int r;
+
+ (void) mkdir_p("/run/systemd/nspawn/", 0755);
+ (void) mkdir_p("/run/systemd/nspawn/propagate", 0600);
+ p = strjoina("/run/systemd/nspawn/propagate/", arg_machine);
+ (void) mkdir_p(p, 0600);
+
+ r = userns_mkdir(root, "/run/systemd", 0755, 0, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to create /run/systemd: %m");
+
+ r = userns_mkdir(root, "/run/systemd/nspawn", 0755, 0, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to create /run/systemd/nspawn: %m");
+
+ r = userns_mkdir(root, "/run/systemd/nspawn/incoming", 0600, 0, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to create /run/systemd/nspawn/incoming: %m");
+
+ q = prefix_roota(root, "/run/systemd/nspawn/incoming");
+ r = mount_verbose(LOG_ERR, p, q, NULL, MS_BIND, NULL);
+ if (r < 0)
+ return r;
+
+ r = mount_verbose(LOG_ERR, NULL, q, NULL, MS_BIND|MS_REMOUNT|MS_RDONLY, NULL);
+ if (r < 0)
+ return r;
+
+ /* machined will MS_MOVE into that directory, and that's only
+ * supported for non-shared mounts. */
+ return mount_verbose(LOG_ERR, NULL, q, NULL, MS_SLAVE, NULL);
+}
+
+static int setup_image(char **device_path, int *loop_nr) {
+ struct loop_info64 info = {
+ .lo_flags = LO_FLAGS_AUTOCLEAR|LO_FLAGS_PARTSCAN
+ };
+ _cleanup_close_ int fd = -1, control = -1, loop = -1;
+ _cleanup_free_ char* loopdev = NULL;
+ struct stat st;
+ int r, nr;
+
+ assert(device_path);
+ assert(loop_nr);
+ assert(arg_image);
+
+ fd = open(arg_image, O_CLOEXEC|(arg_read_only ? O_RDONLY : O_RDWR)|O_NONBLOCK|O_NOCTTY);
+ if (fd < 0)
+ return log_error_errno(errno, "Failed to open %s: %m", arg_image);
+
+ if (fstat(fd, &st) < 0)
+ return log_error_errno(errno, "Failed to stat %s: %m", arg_image);
+
+ if (S_ISBLK(st.st_mode)) {
+ char *p;
+
+ p = strdup(arg_image);
+ if (!p)
+ return log_oom();
+
+ *device_path = p;
+
+ *loop_nr = -1;
+
+ r = fd;
+ fd = -1;
+
+ return r;
+ }
+
+ if (!S_ISREG(st.st_mode)) {
+ log_error("%s is not a regular file or block device.", arg_image);
+ return -EINVAL;
+ }
+
+ control = open("/dev/loop-control", O_RDWR|O_CLOEXEC|O_NOCTTY|O_NONBLOCK);
+ if (control < 0)
+ return log_error_errno(errno, "Failed to open /dev/loop-control: %m");
+
+ nr = ioctl(control, LOOP_CTL_GET_FREE);
+ if (nr < 0)
+ return log_error_errno(errno, "Failed to allocate loop device: %m");
+
+ if (asprintf(&loopdev, "/dev/loop%i", nr) < 0)
+ return log_oom();
+
+ loop = open(loopdev, O_CLOEXEC|(arg_read_only ? O_RDONLY : O_RDWR)|O_NONBLOCK|O_NOCTTY);
+ if (loop < 0)
+ return log_error_errno(errno, "Failed to open loop device %s: %m", loopdev);
+
+ if (ioctl(loop, LOOP_SET_FD, fd) < 0)
+ return log_error_errno(errno, "Failed to set loopback file descriptor on %s: %m", loopdev);
+
+ if (arg_read_only)
+ info.lo_flags |= LO_FLAGS_READ_ONLY;
+
+ if (ioctl(loop, LOOP_SET_STATUS64, &info) < 0)
+ return log_error_errno(errno, "Failed to set loopback settings on %s: %m", loopdev);
+
+ *device_path = loopdev;
+ loopdev = NULL;
+
+ *loop_nr = nr;
+
+ r = loop;
+ loop = -1;
+
+ return r;
+}
+
+#define PARTITION_TABLE_BLURB \
+ "Note that the disk image needs to either contain only a single MBR partition of\n" \
+ "type 0x83 that is marked bootable, or a single GPT partition of type " \
+ "0FC63DAF-8483-4772-8E79-3D69D8477DE4 or follow\n" \
+ " http://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec/\n" \
+ "to be bootable with systemd-nspawn."
+
+static int dissect_image(
+ int fd,
+ char **root_device, bool *root_device_rw,
+ char **home_device, bool *home_device_rw,
+ char **srv_device, bool *srv_device_rw,
+ char **esp_device,
+ bool *secondary) {
+
+#ifdef HAVE_BLKID
+ int home_nr = -1, srv_nr = -1, esp_nr = -1;
+#ifdef GPT_ROOT_NATIVE
+ int root_nr = -1;
+#endif
+#ifdef GPT_ROOT_SECONDARY
+ int secondary_root_nr = -1;
+#endif
+ _cleanup_free_ char *home = NULL, *root = NULL, *secondary_root = NULL, *srv = NULL, *esp = NULL, *generic = NULL;
+ _cleanup_udev_enumerate_unref_ struct udev_enumerate *e = NULL;
+ _cleanup_udev_device_unref_ struct udev_device *d = NULL;
+ _cleanup_blkid_free_probe_ blkid_probe b = NULL;
+ _cleanup_udev_unref_ struct udev *udev = NULL;
+ struct udev_list_entry *first, *item;
+ bool home_rw = true, root_rw = true, secondary_root_rw = true, srv_rw = true, generic_rw = true;
+ bool is_gpt, is_mbr, multiple_generic = false;
+ const char *pttype = NULL;
+ blkid_partlist pl;
+ struct stat st;
+ unsigned i;
+ int r;
+
+ assert(fd >= 0);
+ assert(root_device);
+ assert(home_device);
+ assert(srv_device);
+ assert(esp_device);
+ assert(secondary);
+ assert(arg_image);
+
+ b = blkid_new_probe();
+ if (!b)
+ return log_oom();
+
+ errno = 0;
+ r = blkid_probe_set_device(b, fd, 0, 0);
+ if (r != 0) {
+ if (errno == 0)
+ return log_oom();
+
+ return log_error_errno(errno, "Failed to set device on blkid probe: %m");
+ }
+
+ blkid_probe_enable_partitions(b, 1);
+ blkid_probe_set_partitions_flags(b, BLKID_PARTS_ENTRY_DETAILS);
+
+ errno = 0;
+ r = blkid_do_safeprobe(b);
+ if (r == -2 || r == 1) {
+ log_error("Failed to identify any partition table on\n"
+ " %s\n"
+ PARTITION_TABLE_BLURB, arg_image);
+ return -EINVAL;
+ } else if (r != 0) {
+ if (errno == 0)
+ errno = EIO;
+ return log_error_errno(errno, "Failed to probe: %m");
+ }
+
+ (void) blkid_probe_lookup_value(b, "PTTYPE", &pttype, NULL);
+
+ is_gpt = streq_ptr(pttype, "gpt");
+ is_mbr = streq_ptr(pttype, "dos");
+
+ if (!is_gpt && !is_mbr) {
+ log_error("No GPT or MBR partition table discovered on\n"
+ " %s\n"
+ PARTITION_TABLE_BLURB, arg_image);
+ return -EINVAL;
+ }
+
+ errno = 0;
+ pl = blkid_probe_get_partitions(b);
+ if (!pl) {
+ if (errno == 0)
+ return log_oom();
+
+ log_error("Failed to list partitions of %s", arg_image);
+ return -errno;
+ }
+
+ udev = udev_new();
+ if (!udev)
+ return log_oom();
+
+ if (fstat(fd, &st) < 0)
+ return log_error_errno(errno, "Failed to stat block device: %m");
+
+ d = udev_device_new_from_devnum(udev, 'b', st.st_rdev);
+ if (!d)
+ return log_oom();
+
+ for (i = 0;; i++) {
+ int n, m;
+
+ if (i >= 10) {
+ log_error("Kernel partitions never appeared.");
+ return -ENXIO;
+ }
+
+ e = udev_enumerate_new(udev);
+ if (!e)
+ return log_oom();
+
+ r = udev_enumerate_add_match_parent(e, d);
+ if (r < 0)
+ return log_oom();
+
+ r = udev_enumerate_scan_devices(e);
+ if (r < 0)
+ return log_error_errno(r, "Failed to scan for partition devices of %s: %m", arg_image);
+
+ /* Count the partitions enumerated by the kernel */
+ n = 0;
+ first = udev_enumerate_get_list_entry(e);
+ udev_list_entry_foreach(item, first)
+ n++;
+
+ /* Count the partitions enumerated by blkid */
+ m = blkid_partlist_numof_partitions(pl);
+ if (n == m + 1)
+ break;
+ if (n > m + 1) {
+ log_error("blkid and kernel partition list do not match.");
+ return -EIO;
+ }
+ if (n < m + 1) {
+ unsigned j;
+
+ /* The kernel has probed fewer partitions than
+ * blkid? Maybe the kernel prober is still
+ * running or it got EBUSY because udev
+ * already opened the device. Let's reprobe
+ * the device, which is a synchronous call
+ * that waits until probing is complete. */
+
+ for (j = 0; j < 20; j++) {
+
+ r = ioctl(fd, BLKRRPART, 0);
+ if (r < 0)
+ r = -errno;
+ if (r >= 0 || r != -EBUSY)
+ break;
+
+ /* If something else has the device
+ * open, such as an udev rule, the
+ * ioctl will return EBUSY. Since
+ * there's no way to wait until it
+ * isn't busy anymore, let's just wait
+ * a bit, and try again.
+ *
+ * This is really something they
+ * should fix in the kernel! */
+
+ usleep(50 * USEC_PER_MSEC);
+ }
+
+ if (r < 0)
+ return log_error_errno(r, "Failed to reread partition table: %m");
+ }
+
+ e = udev_enumerate_unref(e);
+ }
+
+ first = udev_enumerate_get_list_entry(e);
+ udev_list_entry_foreach(item, first) {
+ _cleanup_udev_device_unref_ struct udev_device *q;
+ const char *node;
+ unsigned long long flags;
+ blkid_partition pp;
+ dev_t qn;
+ int nr;
+
+ errno = 0;
+ q = udev_device_new_from_syspath(udev, udev_list_entry_get_name(item));
+ if (!q) {
+ if (!errno)
+ errno = ENOMEM;
+
+ return log_error_errno(errno, "Failed to get partition device of %s: %m", arg_image);
+ }
+
+ qn = udev_device_get_devnum(q);
+ if (major(qn) == 0)
+ continue;
+
+ if (st.st_rdev == qn)
+ continue;
+
+ node = udev_device_get_devnode(q);
+ if (!node)
+ continue;
+
+ pp = blkid_partlist_devno_to_partition(pl, qn);
+ if (!pp)
+ continue;
+
+ flags = blkid_partition_get_flags(pp);
+
+ nr = blkid_partition_get_partno(pp);
+ if (nr < 0)
+ continue;
+
+ if (is_gpt) {
+ sd_id128_t type_id;
+ const char *stype;
+
+ if (flags & GPT_FLAG_NO_AUTO)
+ continue;
+
+ stype = blkid_partition_get_type_string(pp);
+ if (!stype)
+ continue;
+
+ if (sd_id128_from_string(stype, &type_id) < 0)
+ continue;
+
+ if (sd_id128_equal(type_id, GPT_HOME)) {
+
+ if (home && nr >= home_nr)
+ continue;
+
+ home_nr = nr;
+ home_rw = !(flags & GPT_FLAG_READ_ONLY);
+
+ r = free_and_strdup(&home, node);
+ if (r < 0)
+ return log_oom();
+
+ } else if (sd_id128_equal(type_id, GPT_SRV)) {
+
+ if (srv && nr >= srv_nr)
+ continue;
+
+ srv_nr = nr;
+ srv_rw = !(flags & GPT_FLAG_READ_ONLY);
+
+ r = free_and_strdup(&srv, node);
+ if (r < 0)
+ return log_oom();
+ } else if (sd_id128_equal(type_id, GPT_ESP)) {
+
+ if (esp && nr >= esp_nr)
+ continue;
+
+ esp_nr = nr;
+
+ r = free_and_strdup(&esp, node);
+ if (r < 0)
+ return log_oom();
+ }
+#ifdef GPT_ROOT_NATIVE
+ else if (sd_id128_equal(type_id, GPT_ROOT_NATIVE)) {
+
+ if (root && nr >= root_nr)
+ continue;
+
+ root_nr = nr;
+ root_rw = !(flags & GPT_FLAG_READ_ONLY);
+
+ r = free_and_strdup(&root, node);
+ if (r < 0)
+ return log_oom();
+ }
+#endif
+#ifdef GPT_ROOT_SECONDARY
+ else if (sd_id128_equal(type_id, GPT_ROOT_SECONDARY)) {
+
+ if (secondary_root && nr >= secondary_root_nr)
+ continue;
+
+ secondary_root_nr = nr;
+ secondary_root_rw = !(flags & GPT_FLAG_READ_ONLY);
+
+ r = free_and_strdup(&secondary_root, node);
+ if (r < 0)
+ return log_oom();
+ }
+#endif
+ else if (sd_id128_equal(type_id, GPT_LINUX_GENERIC)) {
+
+ if (generic)
+ multiple_generic = true;
+ else {
+ generic_rw = !(flags & GPT_FLAG_READ_ONLY);
+
+ r = free_and_strdup(&generic, node);
+ if (r < 0)
+ return log_oom();
+ }
+ }
+
+ } else if (is_mbr) {
+ int type;
+
+ if (flags != 0x80) /* Bootable flag */
+ continue;
+
+ type = blkid_partition_get_type(pp);
+ if (type != 0x83) /* Linux partition */
+ continue;
+
+ if (generic)
+ multiple_generic = true;
+ else {
+ generic_rw = true;
+
+ r = free_and_strdup(&root, node);
+ if (r < 0)
+ return log_oom();
+ }
+ }
+ }
+
+ if (root) {
+ *root_device = root;
+ root = NULL;
+
+ *root_device_rw = root_rw;
+ *secondary = false;
+ } else if (secondary_root) {
+ *root_device = secondary_root;
+ secondary_root = NULL;
+
+ *root_device_rw = secondary_root_rw;
+ *secondary = true;
+ } else if (generic) {
+
+ /* There were no partitions with precise meanings
+ * around, but we found generic partitions. In this
+ * case, if there's only one, we can go ahead and boot
+ * it, otherwise we bail out, because we really cannot
+ * make any sense of it. */
+
+ if (multiple_generic) {
+ log_error("Identified multiple bootable Linux partitions on\n"
+ " %s\n"
+ PARTITION_TABLE_BLURB, arg_image);
+ return -EINVAL;
+ }
+
+ *root_device = generic;
+ generic = NULL;
+
+ *root_device_rw = generic_rw;
+ *secondary = false;
+ } else {
+ log_error("Failed to identify root partition in disk image\n"
+ " %s\n"
+ PARTITION_TABLE_BLURB, arg_image);
+ return -EINVAL;
+ }
+
+ if (home) {
+ *home_device = home;
+ home = NULL;
+
+ *home_device_rw = home_rw;
+ }
+
+ if (srv) {
+ *srv_device = srv;
+ srv = NULL;
+
+ *srv_device_rw = srv_rw;
+ }
+
+ if (esp) {
+ *esp_device = esp;
+ esp = NULL;
+ }
+
+ return 0;
+#else
+ log_error("--image= is not supported, compiled without blkid support.");
+ return -EOPNOTSUPP;
+#endif
+}
+
+static int mount_device(const char *what, const char *where, const char *directory, bool rw) {
+#ifdef HAVE_BLKID
+ _cleanup_blkid_free_probe_ blkid_probe b = NULL;
+ const char *fstype, *p, *options;
+ int r;
+
+ assert(what);
+ assert(where);
+
+ if (arg_read_only)
+ rw = false;
+
+ if (directory)
+ p = strjoina(where, directory);
+ else
+ p = where;
+
+ errno = 0;
+ b = blkid_new_probe_from_filename(what);
+ if (!b) {
+ if (errno == 0)
+ return log_oom();
+ return log_error_errno(errno, "Failed to allocate prober for %s: %m", what);
+ }
+
+ blkid_probe_enable_superblocks(b, 1);
+ blkid_probe_set_superblocks_flags(b, BLKID_SUBLKS_TYPE);
+
+ errno = 0;
+ r = blkid_do_safeprobe(b);
+ if (r == -1 || r == 1) {
+ log_error("Cannot determine file system type of %s", what);
+ return -EINVAL;
+ } else if (r != 0) {
+ if (errno == 0)
+ errno = EIO;
+ return log_error_errno(errno, "Failed to probe %s: %m", what);
+ }
+
+ errno = 0;
+ if (blkid_probe_lookup_value(b, "TYPE", &fstype, NULL) < 0) {
+ if (errno == 0)
+ errno = EINVAL;
+ log_error("Failed to determine file system type of %s", what);
+ return -errno;
+ }
+
+ if (streq(fstype, "crypto_LUKS")) {
+ log_error("nspawn currently does not support LUKS disk images.");
+ return -EOPNOTSUPP;
+ }
+
+ /* If this is a loopback device then let's mount the image with discard, so that the underlying file remains
+ * sparse when possible. */
+ if (STR_IN_SET(fstype, "btrfs", "ext4", "vfat", "xfs")) {
+ const char *l;
+
+ l = path_startswith(what, "/dev");
+ if (l && startswith(l, "loop"))
+ options = "discard";
+ }
+
+ return mount_verbose(LOG_ERR, what, p, fstype, MS_NODEV|(rw ? 0 : MS_RDONLY), options);
+#else
+ log_error("--image= is not supported, compiled without blkid support.");
+ return -EOPNOTSUPP;
+#endif
+}
+
+static int setup_machine_id(const char *directory) {
+ const char *etc_machine_id;
+ sd_id128_t id;
+ int r;
+
+ /* If the UUID in the container is already set, then that's what counts, and we use. If it isn't set, and the
+ * caller passed --uuid=, then we'll pass it in the $container_uuid env var to PID 1 of the container. The
+ * assumption is that PID 1 will then write it to /etc/machine-id to make it persistent. If --uuid= is not
+ * passed we generate a random UUID, and pass it via $container_uuid. In effect this means that /etc/machine-id
+ * in the container and our idea of the container UUID will always be in sync (at least if PID 1 in the
+ * container behaves nicely). */
+
+ etc_machine_id = prefix_roota(directory, "/etc/machine-id");
+
+ r = id128_read(etc_machine_id, ID128_PLAIN, &id);
+ if (r < 0) {
+ if (!IN_SET(r, -ENOENT, -ENOMEDIUM)) /* If the file is missing or empty, we don't mind */
+ return log_error_errno(r, "Failed to read machine ID from container image: %m");
+
+ if (sd_id128_is_null(arg_uuid)) {
+ r = sd_id128_randomize(&arg_uuid);
+ if (r < 0)
+ return log_error_errno(r, "Failed to acquire randomized machine UUID: %m");
+ }
+ } else {
+ if (sd_id128_is_null(id)) {
+ log_error("Machine ID in container image is zero, refusing.");
+ return -EINVAL;
+ }
+
+ arg_uuid = id;
+ }
+
+ return 0;
+}
+
+static int recursive_chown(const char *directory, uid_t shift, uid_t range) {
+ int r;
+
+ assert(directory);
+
+ if (arg_userns_mode == USER_NAMESPACE_NO || !arg_userns_chown)
+ return 0;
+
+ r = path_patch_uid(directory, arg_uid_shift, arg_uid_range);
+ if (r == -EOPNOTSUPP)
+ return log_error_errno(r, "Automatic UID/GID adjusting is only supported for UID/GID ranges starting at multiples of 2^16 with a range of 2^16.");
+ if (r == -EBADE)
+ return log_error_errno(r, "Upper 16 bits of root directory UID and GID do not match.");
+ if (r < 0)
+ return log_error_errno(r, "Failed to adjust UID/GID shift of OS tree: %m");
+ if (r == 0)
+ log_debug("Root directory of image is already owned by the right UID/GID range, skipping recursive chown operation.");
+ else
+ log_debug("Patched directory tree to match UID/GID range.");
+
+ return r;
+}
+
+static int mount_devices(
+ const char *where,
+ const char *root_device, bool root_device_rw,
+ const char *home_device, bool home_device_rw,
+ const char *srv_device, bool srv_device_rw,
+ const char *esp_device) {
+ int r;
+
+ assert(where);
+
+ if (root_device) {
+ r = mount_device(root_device, arg_directory, NULL, root_device_rw);
+ if (r < 0)
+ return log_error_errno(r, "Failed to mount root directory: %m");
+ }
+
+ if (home_device) {
+ r = mount_device(home_device, arg_directory, "/home", home_device_rw);
+ if (r < 0)
+ return log_error_errno(r, "Failed to mount home directory: %m");
+ }
+
+ if (srv_device) {
+ r = mount_device(srv_device, arg_directory, "/srv", srv_device_rw);
+ if (r < 0)
+ return log_error_errno(r, "Failed to mount server data directory: %m");
+ }
+
+ if (esp_device) {
+ const char *mp, *x;
+
+ /* Mount the ESP to /efi if it exists and is empty. If it doesn't exist, use /boot instead. */
+
+ mp = "/efi";
+ x = strjoina(arg_directory, mp);
+ r = dir_is_empty(x);
+ if (r == -ENOENT) {
+ mp = "/boot";
+ x = strjoina(arg_directory, mp);
+ r = dir_is_empty(x);
+ }
+
+ if (r > 0) {
+ r = mount_device(esp_device, arg_directory, mp, true);
+ if (r < 0)
+ return log_error_errno(r, "Failed to mount ESP: %m");
+ }
+ }
+
+ return 0;
+}
+
+static void loop_remove(int nr, int *image_fd) {
+ _cleanup_close_ int control = -1;
+ int r;
+
+ if (nr < 0)
+ return;
+
+ if (image_fd && *image_fd >= 0) {
+ r = ioctl(*image_fd, LOOP_CLR_FD);
+ if (r < 0)
+ log_debug_errno(errno, "Failed to close loop image: %m");
+ *image_fd = safe_close(*image_fd);
+ }
+
+ control = open("/dev/loop-control", O_RDWR|O_CLOEXEC|O_NOCTTY|O_NONBLOCK);
+ if (control < 0) {
+ log_warning_errno(errno, "Failed to open /dev/loop-control: %m");
+ return;
+ }
+
+ r = ioctl(control, LOOP_CTL_REMOVE, nr);
+ if (r < 0)
+ log_debug_errno(errno, "Failed to remove loop %d: %m", nr);
+}
+
+/*
+ * Return values:
+ * < 0 : wait_for_terminate() failed to get the state of the
+ * container, the container was terminated by a signal, or
+ * failed for an unknown reason. No change is made to the
+ * container argument.
+ * > 0 : The program executed in the container terminated with an
+ * error. The exit code of the program executed in the
+ * container is returned. The container argument has been set
+ * to CONTAINER_TERMINATED.
+ * 0 : The container is being rebooted, has been shut down or exited
+ * successfully. The container argument has been set to either
+ * CONTAINER_TERMINATED or CONTAINER_REBOOTED.
+ *
+ * That is, success is indicated by a return value of zero, and an
+ * error is indicated by a non-zero value.
+ */
+static int wait_for_container(pid_t pid, ContainerStatus *container) {
+ siginfo_t status;
+ int r;
+
+ r = wait_for_terminate(pid, &status);
+ if (r < 0)
+ return log_warning_errno(r, "Failed to wait for container: %m");
+
+ switch (status.si_code) {
+
+ case CLD_EXITED:
+ if (status.si_status == 0)
+ log_full(arg_quiet ? LOG_DEBUG : LOG_INFO, "Container %s exited successfully.", arg_machine);
+ else
+ log_full(arg_quiet ? LOG_DEBUG : LOG_INFO, "Container %s failed with error code %i.", arg_machine, status.si_status);
+
+ *container = CONTAINER_TERMINATED;
+ return status.si_status;
+
+ case CLD_KILLED:
+ if (status.si_status == SIGINT) {
+ log_full(arg_quiet ? LOG_DEBUG : LOG_INFO, "Container %s has been shut down.", arg_machine);
+ *container = CONTAINER_TERMINATED;
+ return 0;
+
+ } else if (status.si_status == SIGHUP) {
+ log_full(arg_quiet ? LOG_DEBUG : LOG_INFO, "Container %s is being rebooted.", arg_machine);
+ *container = CONTAINER_REBOOTED;
+ return 0;
+ }
+
+ /* CLD_KILLED fallthrough */
+
+ case CLD_DUMPED:
+ log_error("Container %s terminated by signal %s.", arg_machine, signal_to_string(status.si_status));
+ return -EIO;
+
+ default:
+ log_error("Container %s failed due to unknown reason.", arg_machine);
+ return -EIO;
+ }
+}
+
+static int on_orderly_shutdown(sd_event_source *s, const struct signalfd_siginfo *si, void *userdata) {
+ pid_t pid;
+
+ pid = PTR_TO_PID(userdata);
+ if (pid > 0) {
+ if (kill(pid, arg_kill_signal) >= 0) {
+ log_info("Trying to halt container. Send SIGTERM again to trigger immediate termination.");
+ sd_event_source_set_userdata(s, NULL);
+ return 0;
+ }
+ }
+
+ sd_event_exit(sd_event_source_get_event(s), 0);
+ return 0;
+}
+
+static int determine_names(void) {
+ int r;
+
+ if (arg_template && !arg_directory && arg_machine) {
+
+ /* If --template= was specified then we should not
+ * search for a machine, but instead create a new one
+ * in /var/lib/machine. */
+
+ arg_directory = strjoin("/var/lib/machines/", arg_machine, NULL);
+ if (!arg_directory)
+ return log_oom();
+ }
+
+ if (!arg_image && !arg_directory) {
+ if (arg_machine) {
+ _cleanup_(image_unrefp) Image *i = NULL;
+
+ r = image_find(arg_machine, &i);
+ if (r < 0)
+ return log_error_errno(r, "Failed to find image for machine '%s': %m", arg_machine);
+ else if (r == 0) {
+ log_error("No image for machine '%s': %m", arg_machine);
+ return -ENOENT;
+ }
+
+ if (i->type == IMAGE_RAW)
+ r = free_and_strdup(&arg_image, i->path);
+ else
+ r = free_and_strdup(&arg_directory, i->path);
+ if (r < 0)
+ return log_error_errno(r, "Invalid image directory: %m");
+
+ if (!arg_ephemeral)
+ arg_read_only = arg_read_only || i->read_only;
+ } else
+ arg_directory = get_current_dir_name();
+
+ if (!arg_directory && !arg_machine) {
+ log_error("Failed to determine path, please use -D or -i.");
+ return -EINVAL;
+ }
+ }
+
+ if (!arg_machine) {
+ if (arg_directory && path_equal(arg_directory, "/"))
+ arg_machine = gethostname_malloc();
+ else
+ arg_machine = strdup(basename(arg_image ?: arg_directory));
+
+ if (!arg_machine)
+ return log_oom();
+
+ hostname_cleanup(arg_machine);
+ if (!machine_name_is_valid(arg_machine)) {
+ log_error("Failed to determine machine name automatically, please use -M.");
+ return -EINVAL;
+ }
+
+ if (arg_ephemeral) {
+ char *b;
+
+ /* Add a random suffix when this is an
+ * ephemeral machine, so that we can run many
+ * instances at once without manually having
+ * to specify -M each time. */
+
+ if (asprintf(&b, "%s-%016" PRIx64, arg_machine, random_u64()) < 0)
+ return log_oom();
+
+ free(arg_machine);
+ arg_machine = b;
+ }
+ }
+
+ return 0;
+}
+
+static int determine_uid_shift(const char *directory) {
+ int r;
+
+ if (arg_userns_mode == USER_NAMESPACE_NO) {
+ arg_uid_shift = 0;
+ return 0;
+ }
+
+ if (arg_uid_shift == UID_INVALID) {
+ struct stat st;
+
+ r = stat(directory, &st);
+ if (r < 0)
+ return log_error_errno(errno, "Failed to determine UID base of %s: %m", directory);
+
+ arg_uid_shift = st.st_uid & UINT32_C(0xffff0000);
+
+ if (arg_uid_shift != (st.st_gid & UINT32_C(0xffff0000))) {
+ log_error("UID and GID base of %s don't match.", directory);
+ return -EINVAL;
+ }
+
+ arg_uid_range = UINT32_C(0x10000);
+ }
+
+ if (arg_uid_shift > (uid_t) -1 - arg_uid_range) {
+ log_error("UID base too high for UID range.");
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+static int inner_child(
+ Barrier *barrier,
+ const char *directory,
+ bool secondary,
+ int kmsg_socket,
+ int rtnl_socket,
+ FDSet *fds) {
+
+ _cleanup_free_ char *home = NULL;
+ char as_uuid[37];
+ unsigned n_env = 1;
+ const char *envp[] = {
+ "PATH=" DEFAULT_PATH_SPLIT_USR,
+ NULL, /* container */
+ NULL, /* TERM */
+ NULL, /* HOME */
+ NULL, /* USER */
+ NULL, /* LOGNAME */
+ NULL, /* container_uuid */
+ NULL, /* LISTEN_FDS */
+ NULL, /* LISTEN_PID */
+ NULL, /* NOTIFY_SOCKET */
+ NULL
+ };
+
+ _cleanup_strv_free_ char **env_use = NULL;
+ int r;
+
+ assert(barrier);
+ assert(directory);
+ assert(kmsg_socket >= 0);
+
+ cg_unified_flush();
+
+ if (arg_userns_mode != USER_NAMESPACE_NO) {
+ /* Tell the parent, that it now can write the UID map. */
+ (void) barrier_place(barrier); /* #1 */
+
+ /* Wait until the parent wrote the UID map */
+ if (!barrier_place_and_sync(barrier)) { /* #2 */
+ log_error("Parent died too early");
+ return -ESRCH;
+ }
+ }
+
+ r = reset_uid_gid();
+ if (r < 0)
+ return log_error_errno(r, "Couldn't become new root: %m");
+
+ r = mount_all(NULL,
+ arg_userns_mode != USER_NAMESPACE_NO,
+ true,
+ arg_private_network,
+ arg_uid_shift,
+ arg_uid_range,
+ arg_selinux_apifs_context);
+
+ if (r < 0)
+ return r;
+
+ r = mount_sysfs(NULL);
+ if (r < 0)
+ return r;
+
+ /* Wait until we are cgroup-ified, so that we
+ * can mount the right cgroup path writable */
+ if (!barrier_place_and_sync(barrier)) { /* #3 */
+ log_error("Parent died too early");
+ return -ESRCH;
+ }
+
+ if (arg_use_cgns && cg_ns_supported()) {
+ r = unshare(CLONE_NEWCGROUP);
+ if (r < 0)
+ return log_error_errno(errno, "Failed to unshare cgroup namespace");
+ r = mount_cgroups(
+ "",
+ arg_unified_cgroup_hierarchy,
+ arg_userns_mode != USER_NAMESPACE_NO,
+ arg_uid_shift,
+ arg_uid_range,
+ arg_selinux_apifs_context,
+ true);
+ if (r < 0)
+ return r;
+ } else {
+ r = mount_systemd_cgroup_writable("", arg_unified_cgroup_hierarchy);
+ if (r < 0)
+ return r;
+ }
+
+ r = setup_boot_id(NULL);
+ if (r < 0)
+ return r;
+
+ r = setup_kmsg(NULL, kmsg_socket);
+ if (r < 0)
+ return r;
+ kmsg_socket = safe_close(kmsg_socket);
+
+ umask(0022);
+
+ if (setsid() < 0)
+ return log_error_errno(errno, "setsid() failed: %m");
+
+ if (arg_private_network)
+ loopback_setup();
+
+ if (arg_expose_ports) {
+ r = expose_port_send_rtnl(rtnl_socket);
+ if (r < 0)
+ return r;
+ rtnl_socket = safe_close(rtnl_socket);
+ }
+
+ r = drop_capabilities();
+ if (r < 0)
+ return log_error_errno(r, "drop_capabilities() failed: %m");
+
+ setup_hostname();
+
+ if (arg_personality != PERSONALITY_INVALID) {
+ if (personality(arg_personality) < 0)
+ return log_error_errno(errno, "personality() failed: %m");
+ } else if (secondary) {
+ if (personality(PER_LINUX32) < 0)
+ return log_error_errno(errno, "personality() failed: %m");
+ }
+
+#ifdef HAVE_SELINUX
+ if (arg_selinux_context)
+ if (setexeccon(arg_selinux_context) < 0)
+ return log_error_errno(errno, "setexeccon(\"%s\") failed: %m", arg_selinux_context);
+#endif
+
+ r = change_uid_gid(arg_user, &home);
+ if (r < 0)
+ return r;
+
+ /* LXC sets container=lxc, so follow the scheme here */
+ envp[n_env++] = strjoina("container=", arg_container_service_name);
+
+ envp[n_env] = strv_find_prefix(environ, "TERM=");
+ if (envp[n_env])
+ n_env++;
+
+ if ((asprintf((char**)(envp + n_env++), "HOME=%s", home ? home: "/root") < 0) ||
+ (asprintf((char**)(envp + n_env++), "USER=%s", arg_user ? arg_user : "root") < 0) ||
+ (asprintf((char**)(envp + n_env++), "LOGNAME=%s", arg_user ? arg_user : "root") < 0))
+ return log_oom();
+
+ assert(!sd_id128_is_null(arg_uuid));
+
+ if (asprintf((char**)(envp + n_env++), "container_uuid=%s", id128_to_uuid_string(arg_uuid, as_uuid)) < 0)
+ return log_oom();
+
+ if (fdset_size(fds) > 0) {
+ r = fdset_cloexec(fds, false);
+ if (r < 0)
+ return log_error_errno(r, "Failed to unset O_CLOEXEC for file descriptors.");
+
+ if ((asprintf((char **)(envp + n_env++), "LISTEN_FDS=%u", fdset_size(fds)) < 0) ||
+ (asprintf((char **)(envp + n_env++), "LISTEN_PID=1") < 0))
+ return log_oom();
+ }
+ if (asprintf((char **)(envp + n_env++), "NOTIFY_SOCKET=%s", NSPAWN_NOTIFY_SOCKET_PATH) < 0)
+ return log_oom();
+
+ env_use = strv_env_merge(2, envp, arg_setenv);
+ if (!env_use)
+ return log_oom();
+
+ /* Let the parent know that we are ready and
+ * wait until the parent is ready with the
+ * setup, too... */
+ if (!barrier_place_and_sync(barrier)) { /* #4 */
+ log_error("Parent died too early");
+ return -ESRCH;
+ }
+
+ if (arg_chdir)
+ if (chdir(arg_chdir) < 0)
+ return log_error_errno(errno, "Failed to change to specified working directory %s: %m", arg_chdir);
+
+ if (arg_start_mode == START_PID2) {
+ r = stub_pid1();
+ if (r < 0)
+ return r;
+ }
+
+ /* Now, explicitly close the log, so that we
+ * then can close all remaining fds. Closing
+ * the log explicitly first has the benefit
+ * that the logging subsystem knows about it,
+ * and is thus ready to be reopened should we
+ * need it again. Note that the other fds
+ * closed here are at least the locking and
+ * barrier fds. */
+ log_close();
+ (void) fdset_close_others(fds);
+
+ if (arg_start_mode == START_BOOT) {
+ char **a;
+ size_t m;
+
+ /* Automatically search for the init system */
+
+ m = strv_length(arg_parameters);
+ a = newa(char*, m + 2);
+ memcpy_safe(a + 1, arg_parameters, m * sizeof(char*));
+ a[1 + m] = NULL;
+
+ a[0] = (char*) "/usr/lib/systemd/systemd";
+ execve(a[0], a, env_use);
+
+ a[0] = (char*) "/lib/systemd/systemd";
+ execve(a[0], a, env_use);
+
+ a[0] = (char*) "/sbin/init";
+ execve(a[0], a, env_use);
+ } else if (!strv_isempty(arg_parameters))
+ execvpe(arg_parameters[0], arg_parameters, env_use);
+ else {
+ if (!arg_chdir)
+ /* If we cannot change the directory, we'll end up in /, that is expected. */
+ (void) chdir(home ?: "/root");
+
+ execle("/bin/bash", "-bash", NULL, env_use);
+ execle("/bin/sh", "-sh", NULL, env_use);
+ }
+
+ r = -errno;
+ (void) log_open();
+ return log_error_errno(r, "execv() failed: %m");
+}
+
+static int setup_sd_notify_child(void) {
+ static const int one = 1;
+ int fd = -1;
+ union sockaddr_union sa = {
+ .sa.sa_family = AF_UNIX,
+ };
+ int r;
+
+ fd = socket(AF_UNIX, SOCK_DGRAM|SOCK_CLOEXEC|SOCK_NONBLOCK, 0);
+ if (fd < 0)
+ return log_error_errno(errno, "Failed to allocate notification socket: %m");
+
+ (void) mkdir_parents(NSPAWN_NOTIFY_SOCKET_PATH, 0755);
+ (void) unlink(NSPAWN_NOTIFY_SOCKET_PATH);
+
+ strncpy(sa.un.sun_path, NSPAWN_NOTIFY_SOCKET_PATH, sizeof(sa.un.sun_path)-1);
+ r = bind(fd, &sa.sa, SOCKADDR_UN_LEN(sa.un));
+ if (r < 0) {
+ safe_close(fd);
+ return log_error_errno(errno, "bind(%s) failed: %m", sa.un.sun_path);
+ }
+
+ r = setsockopt(fd, SOL_SOCKET, SO_PASSCRED, &one, sizeof(one));
+ if (r < 0) {
+ safe_close(fd);
+ return log_error_errno(errno, "SO_PASSCRED failed: %m");
+ }
+
+ return fd;
+}
+
+static int outer_child(
+ Barrier *barrier,
+ const char *directory,
+ const char *console,
+ const char *root_device, bool root_device_rw,
+ const char *home_device, bool home_device_rw,
+ const char *srv_device, bool srv_device_rw,
+ const char *esp_device,
+ bool interactive,
+ bool secondary,
+ int pid_socket,
+ int uuid_socket,
+ int notify_socket,
+ int kmsg_socket,
+ int rtnl_socket,
+ int uid_shift_socket,
+ FDSet *fds) {
+
+ pid_t pid;
+ ssize_t l;
+ int r;
+ _cleanup_close_ int fd = -1;
+
+ assert(barrier);
+ assert(directory);
+ assert(console);
+ assert(pid_socket >= 0);
+ assert(uuid_socket >= 0);
+ assert(notify_socket >= 0);
+ assert(kmsg_socket >= 0);
+
+ cg_unified_flush();
+
+ if (prctl(PR_SET_PDEATHSIG, SIGKILL) < 0)
+ return log_error_errno(errno, "PR_SET_PDEATHSIG failed: %m");
+
+ if (interactive) {
+ close_nointr(STDIN_FILENO);
+ close_nointr(STDOUT_FILENO);
+ close_nointr(STDERR_FILENO);
+
+ r = open_terminal(console, O_RDWR);
+ if (r != STDIN_FILENO) {
+ if (r >= 0) {
+ safe_close(r);
+ r = -EINVAL;
+ }
+
+ return log_error_errno(r, "Failed to open console: %m");
+ }
+
+ if (dup2(STDIN_FILENO, STDOUT_FILENO) != STDOUT_FILENO ||
+ dup2(STDIN_FILENO, STDERR_FILENO) != STDERR_FILENO)
+ return log_error_errno(errno, "Failed to duplicate console: %m");
+ }
+
+ r = reset_audit_loginuid();
+ if (r < 0)
+ return r;
+
+ /* Mark everything as slave, so that we still
+ * receive mounts from the real root, but don't
+ * propagate mounts to the real root. */
+ r = mount_verbose(LOG_ERR, NULL, "/", NULL, MS_SLAVE|MS_REC, NULL);
+ if (r < 0)
+ return r;
+
+ r = mount_devices(directory,
+ root_device, root_device_rw,
+ home_device, home_device_rw,
+ srv_device, srv_device_rw,
+ esp_device);
+ if (r < 0)
+ return r;
+
+ r = determine_uid_shift(directory);
+ if (r < 0)
+ return r;
+
+ r = detect_unified_cgroup_hierarchy(directory);
+ if (r < 0)
+ return r;
+
+ if (arg_userns_mode != USER_NAMESPACE_NO) {
+ /* Let the parent know which UID shift we read from the image */
+ l = send(uid_shift_socket, &arg_uid_shift, sizeof(arg_uid_shift), MSG_NOSIGNAL);
+ if (l < 0)
+ return log_error_errno(errno, "Failed to send UID shift: %m");
+ if (l != sizeof(arg_uid_shift)) {
+ log_error("Short write while sending UID shift.");
+ return -EIO;
+ }
+
+ if (arg_userns_mode == USER_NAMESPACE_PICK) {
+ /* When we are supposed to pick the UID shift, the parent will check now whether the UID shift
+ * we just read from the image is available. If yes, it will send the UID shift back to us, if
+ * not it will pick a different one, and send it back to us. */
+
+ l = recv(uid_shift_socket, &arg_uid_shift, sizeof(arg_uid_shift), 0);
+ if (l < 0)
+ return log_error_errno(errno, "Failed to recv UID shift: %m");
+ if (l != sizeof(arg_uid_shift)) {
+ log_error("Short read while receiving UID shift.");
+ return -EIO;
+ }
+ }
+
+ log_info("Selected user namespace base " UID_FMT " and range " UID_FMT ".", arg_uid_shift, arg_uid_range);
+ }
+
+ /* Turn directory into bind mount */
+ r = mount_verbose(LOG_ERR, directory, directory, NULL, MS_BIND|MS_REC, NULL);
+ if (r < 0)
+ return r;
+
+ /* Mark everything as shared so our mounts get propagated down. This is
+ * required to make new bind mounts available in systemd services
+ * inside the containter that create a new mount namespace.
+ * See https://github.com/systemd/systemd/issues/3860
+ * Further submounts (such as /dev) done after this will inherit the
+ * shared propagation mode.*/
+ r = mount_verbose(LOG_ERR, NULL, directory, NULL, MS_SHARED|MS_REC, NULL);
+ if (r < 0)
+ return r;
+
+ r = recursive_chown(directory, arg_uid_shift, arg_uid_range);
+ if (r < 0)
+ return r;
+
+ r = setup_volatile(
+ directory,
+ arg_volatile_mode,
+ arg_userns_mode != USER_NAMESPACE_NO,
+ arg_uid_shift,
+ arg_uid_range,
+ arg_selinux_context);
+ if (r < 0)
+ return r;
+
+ r = setup_volatile_state(
+ directory,
+ arg_volatile_mode,
+ arg_userns_mode != USER_NAMESPACE_NO,
+ arg_uid_shift,
+ arg_uid_range,
+ arg_selinux_context);
+ if (r < 0)
+ return r;
+
+ r = base_filesystem_create(directory, arg_uid_shift, (gid_t) arg_uid_shift);
+ if (r < 0)
+ return r;
+
+ if (arg_read_only) {
+ r = bind_remount_recursive(directory, true, NULL);
+ if (r < 0)
+ return log_error_errno(r, "Failed to make tree read-only: %m");
+ }
+
+ r = mount_all(directory,
+ arg_userns_mode != USER_NAMESPACE_NO,
+ false,
+ arg_private_network,
+ arg_uid_shift,
+ arg_uid_range,
+ arg_selinux_apifs_context);
+ if (r < 0)
+ return r;
+
+ r = copy_devnodes(directory);
+ if (r < 0)
+ return r;
+
+ dev_setup(directory, arg_uid_shift, arg_uid_shift);
+
+ r = setup_pts(directory);
+ if (r < 0)
+ return r;
+
+ r = setup_propagate(directory);
+ if (r < 0)
+ return r;
+
+ r = setup_dev_console(directory, console);
+ if (r < 0)
+ return r;
+
+ r = setup_seccomp(arg_caps_retain);
+ if (r < 0)
+ return r;
+
+ r = setup_timezone(directory);
+ if (r < 0)
+ return r;
+
+ r = setup_resolv_conf(directory);
+ if (r < 0)
+ return r;
+
+ r = setup_machine_id(directory);
+ if (r < 0)
+ return r;
+
+ r = setup_journal(directory);
+ if (r < 0)
+ return r;
+
+ r = mount_custom(
+ directory,
+ arg_custom_mounts,
+ arg_n_custom_mounts,
+ arg_userns_mode != USER_NAMESPACE_NO,
+ arg_uid_shift,
+ arg_uid_range,
+ arg_selinux_apifs_context);
+ if (r < 0)
+ return r;
+
+ if (!arg_use_cgns || !cg_ns_supported()) {
+ r = mount_cgroups(
+ directory,
+ arg_unified_cgroup_hierarchy,
+ arg_userns_mode != USER_NAMESPACE_NO,
+ arg_uid_shift,
+ arg_uid_range,
+ arg_selinux_apifs_context,
+ false);
+ if (r < 0)
+ return r;
+ }
+
+ r = mount_move_root(directory);
+ if (r < 0)
+ return log_error_errno(r, "Failed to move root directory: %m");
+
+ fd = setup_sd_notify_child();
+ if (fd < 0)
+ return fd;
+
+ pid = raw_clone(SIGCHLD|CLONE_NEWNS|
+ arg_clone_ns_flags |
+ (arg_private_network ? CLONE_NEWNET : 0) |
+ (arg_userns_mode != USER_NAMESPACE_NO ? CLONE_NEWUSER : 0));
+ if (pid < 0)
+ return log_error_errno(errno, "Failed to fork inner child: %m");
+ if (pid == 0) {
+ pid_socket = safe_close(pid_socket);
+ uuid_socket = safe_close(uuid_socket);
+ notify_socket = safe_close(notify_socket);
+ uid_shift_socket = safe_close(uid_shift_socket);
+
+ /* The inner child has all namespaces that are
+ * requested, so that we all are owned by the user if
+ * user namespaces are turned on. */
+
+ r = inner_child(barrier, directory, secondary, kmsg_socket, rtnl_socket, fds);
+ if (r < 0)
+ _exit(EXIT_FAILURE);
+
+ _exit(EXIT_SUCCESS);
+ }
+
+ l = send(pid_socket, &pid, sizeof(pid), MSG_NOSIGNAL);
+ if (l < 0)
+ return log_error_errno(errno, "Failed to send PID: %m");
+ if (l != sizeof(pid)) {
+ log_error("Short write while sending PID.");
+ return -EIO;
+ }
+
+ l = send(uuid_socket, &arg_uuid, sizeof(arg_uuid), MSG_NOSIGNAL);
+ if (l < 0)
+ return log_error_errno(errno, "Failed to send machine ID: %m");
+ if (l != sizeof(arg_uuid)) {
+ log_error("Short write while sending machine ID.");
+ return -EIO;
+ }
+
+ l = send_one_fd(notify_socket, fd, 0);
+ if (l < 0)
+ return log_error_errno(errno, "Failed to send notify fd: %m");
+
+ pid_socket = safe_close(pid_socket);
+ uuid_socket = safe_close(uuid_socket);
+ notify_socket = safe_close(notify_socket);
+ kmsg_socket = safe_close(kmsg_socket);
+ rtnl_socket = safe_close(rtnl_socket);
+
+ return 0;
+}
+
+static int uid_shift_pick(uid_t *shift, LockFile *ret_lock_file) {
+ unsigned n_tries = 100;
+ uid_t candidate;
+ int r;
+
+ assert(shift);
+ assert(ret_lock_file);
+ assert(arg_userns_mode == USER_NAMESPACE_PICK);
+ assert(arg_uid_range == 0x10000U);
+
+ candidate = *shift;
+
+ (void) mkdir("/run/systemd/nspawn-uid", 0755);
+
+ for (;;) {
+ char lock_path[strlen("/run/systemd/nspawn-uid/") + DECIMAL_STR_MAX(uid_t) + 1];
+ _cleanup_release_lock_file_ LockFile lf = LOCK_FILE_INIT;
+
+ if (--n_tries <= 0)
+ return -EBUSY;
+
+ if (candidate < UID_SHIFT_PICK_MIN || candidate > UID_SHIFT_PICK_MAX)
+ goto next;
+ if ((candidate & UINT32_C(0xFFFF)) != 0)
+ goto next;
+
+ xsprintf(lock_path, "/run/systemd/nspawn-uid/" UID_FMT, candidate);
+ r = make_lock_file(lock_path, LOCK_EX|LOCK_NB, &lf);
+ if (r == -EBUSY) /* Range already taken by another nspawn instance */
+ goto next;
+ if (r < 0)
+ return r;
+
+ /* Make some superficial checks whether the range is currently known in the user database */
+ if (getpwuid(candidate))
+ goto next;
+ if (getpwuid(candidate + UINT32_C(0xFFFE)))
+ goto next;
+ if (getgrgid(candidate))
+ goto next;
+ if (getgrgid(candidate + UINT32_C(0xFFFE)))
+ goto next;
+
+ *ret_lock_file = lf;
+ lf = (struct LockFile) LOCK_FILE_INIT;
+ *shift = candidate;
+ return 0;
+
+ next:
+ random_bytes(&candidate, sizeof(candidate));
+ candidate = (candidate % (UID_SHIFT_PICK_MAX - UID_SHIFT_PICK_MIN)) + UID_SHIFT_PICK_MIN;
+ candidate &= (uid_t) UINT32_C(0xFFFF0000);
+ }
+}
+
+static int setup_uid_map(pid_t pid) {
+ char uid_map[strlen("/proc//uid_map") + DECIMAL_STR_MAX(uid_t) + 1], line[DECIMAL_STR_MAX(uid_t)*3+3+1];
+ int r;
+
+ assert(pid > 1);
+
+ xsprintf(uid_map, "/proc/" PID_FMT "/uid_map", pid);
+ xsprintf(line, UID_FMT " " UID_FMT " " UID_FMT "\n", 0, arg_uid_shift, arg_uid_range);
+ r = write_string_file(uid_map, line, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to write UID map: %m");
+
+ /* We always assign the same UID and GID ranges */
+ xsprintf(uid_map, "/proc/" PID_FMT "/gid_map", pid);
+ r = write_string_file(uid_map, line, 0);
+ if (r < 0)
+ return log_error_errno(r, "Failed to write GID map: %m");
+
+ return 0;
+}
+
+static int nspawn_dispatch_notify_fd(sd_event_source *source, int fd, uint32_t revents, void *userdata) {
+ char buf[NOTIFY_BUFFER_MAX+1];
+ char *p = NULL;
+ struct iovec iovec = {
+ .iov_base = buf,
+ .iov_len = sizeof(buf)-1,
+ };
+ union {
+ struct cmsghdr cmsghdr;
+ uint8_t buf[CMSG_SPACE(sizeof(struct ucred)) +
+ CMSG_SPACE(sizeof(int) * NOTIFY_FD_MAX)];
+ } control = {};
+ struct msghdr msghdr = {
+ .msg_iov = &iovec,
+ .msg_iovlen = 1,
+ .msg_control = &control,
+ .msg_controllen = sizeof(control),
+ };
+ struct cmsghdr *cmsg;
+ struct ucred *ucred = NULL;
+ ssize_t n;
+ pid_t inner_child_pid;
+ _cleanup_strv_free_ char **tags = NULL;
+
+ assert(userdata);
+
+ inner_child_pid = PTR_TO_PID(userdata);
+
+ if (revents != EPOLLIN) {
+ log_warning("Got unexpected poll event for notify fd.");
+ return 0;
+ }
+
+ n = recvmsg(fd, &msghdr, MSG_DONTWAIT|MSG_CMSG_CLOEXEC);
+ if (n < 0) {
+ if (errno == EAGAIN || errno == EINTR)
+ return 0;
+
+ return log_warning_errno(errno, "Couldn't read notification socket: %m");
+ }
+ cmsg_close_all(&msghdr);
+
+ CMSG_FOREACH(cmsg, &msghdr) {
+ if (cmsg->cmsg_level == SOL_SOCKET &&
+ cmsg->cmsg_type == SCM_CREDENTIALS &&
+ cmsg->cmsg_len == CMSG_LEN(sizeof(struct ucred))) {
+
+ ucred = (struct ucred*) CMSG_DATA(cmsg);
+ }
+ }
+
+ if (!ucred || ucred->pid != inner_child_pid) {
+ log_warning("Received notify message without valid credentials. Ignoring.");
+ return 0;
+ }
+
+ if ((size_t) n >= sizeof(buf)) {
+ log_warning("Received notify message exceeded maximum size. Ignoring.");
+ return 0;
+ }
+
+ buf[n] = 0;
+ tags = strv_split(buf, "\n\r");
+ if (!tags)
+ return log_oom();
+
+ if (strv_find(tags, "READY=1"))
+ sd_notifyf(false, "READY=1\n");
+
+ p = strv_find_startswith(tags, "STATUS=");
+ if (p)
+ sd_notifyf(false, "STATUS=Container running: %s", p);
+
+ return 0;
+}
+
+static int setup_sd_notify_parent(sd_event *event, int fd, pid_t *inner_child_pid) {
+ int r;
+ sd_event_source *notify_event_source;
+
+ r = sd_event_add_io(event, &notify_event_source, fd, EPOLLIN, nspawn_dispatch_notify_fd, inner_child_pid);
+ if (r < 0)
+ return log_error_errno(r, "Failed to allocate notify event source: %m");
+
+ (void) sd_event_source_set_description(notify_event_source, "nspawn-notify");
+
+ return 0;
+}
+
+static int load_settings(void) {
+ _cleanup_(settings_freep) Settings *settings = NULL;
+ _cleanup_fclose_ FILE *f = NULL;
+ _cleanup_free_ char *p = NULL;
+ const char *fn, *i;
+ int r;
+
+ /* If all settings are masked, there's no point in looking for
+ * the settings file */
+ if ((arg_settings_mask & _SETTINGS_MASK_ALL) == _SETTINGS_MASK_ALL)
+ return 0;
+
+ fn = strjoina(arg_machine, ".nspawn");
+
+ /* We first look in the admin's directories in /etc and /run */
+ FOREACH_STRING(i, "/etc/systemd/nspawn", "/run/systemd/nspawn") {
+ _cleanup_free_ char *j = NULL;
+
+ j = strjoin(i, "/", fn, NULL);
+ if (!j)
+ return log_oom();
+
+ f = fopen(j, "re");
+ if (f) {
+ p = j;
+ j = NULL;
+
+ /* By default, we trust configuration from /etc and /run */
+ if (arg_settings_trusted < 0)
+ arg_settings_trusted = true;
+
+ break;
+ }
+
+ if (errno != ENOENT)
+ return log_error_errno(errno, "Failed to open %s: %m", j);
+ }
+
+ if (!f) {
+ /* After that, let's look for a file next to the
+ * actual image we shall boot. */
+
+ if (arg_image) {
+ p = file_in_same_dir(arg_image, fn);
+ if (!p)
+ return log_oom();
+ } else if (arg_directory) {
+ p = file_in_same_dir(arg_directory, fn);
+ if (!p)
+ return log_oom();
+ }
+
+ if (p) {
+ f = fopen(p, "re");
+ if (!f && errno != ENOENT)
+ return log_error_errno(errno, "Failed to open %s: %m", p);
+
+ /* By default, we do not trust configuration from /var/lib/machines */
+ if (arg_settings_trusted < 0)
+ arg_settings_trusted = false;
+ }
+ }
+
+ if (!f)
+ return 0;
+
+ log_debug("Settings are trusted: %s", yes_no(arg_settings_trusted));
+
+ r = settings_load(f, p, &settings);
+ if (r < 0)
+ return r;
+
+ /* Copy over bits from the settings, unless they have been
+ * explicitly masked by command line switches. */
+
+ if ((arg_settings_mask & SETTING_START_MODE) == 0 &&
+ settings->start_mode >= 0) {
+ arg_start_mode = settings->start_mode;
+
+ strv_free(arg_parameters);
+ arg_parameters = settings->parameters;
+ settings->parameters = NULL;
+ }
+
+ if ((arg_settings_mask & SETTING_WORKING_DIRECTORY) == 0 &&
+ settings->working_directory) {
+ free(arg_chdir);
+ arg_chdir = settings->working_directory;
+ settings->working_directory = NULL;
+ }
+
+ if ((arg_settings_mask & SETTING_ENVIRONMENT) == 0 &&
+ settings->environment) {
+ strv_free(arg_setenv);
+ arg_setenv = settings->environment;
+ settings->environment = NULL;
+ }
+
+ if ((arg_settings_mask & SETTING_USER) == 0 &&
+ settings->user) {
+ free(arg_user);
+ arg_user = settings->user;
+ settings->user = NULL;
+ }
+
+ if ((arg_settings_mask & SETTING_CAPABILITY) == 0) {
+ uint64_t plus;
+
+ plus = settings->capability;
+ if (settings_private_network(settings))
+ plus |= (1ULL << CAP_NET_ADMIN);
+
+ if (!arg_settings_trusted && plus != 0) {
+ if (settings->capability != 0)
+ log_warning("Ignoring Capability= setting, file %s is not trusted.", p);
+ } else
+ arg_caps_retain |= plus;
+
+ arg_caps_retain &= ~settings->drop_capability;
+ }
+
+ if ((arg_settings_mask & SETTING_KILL_SIGNAL) == 0 &&
+ settings->kill_signal > 0)
+ arg_kill_signal = settings->kill_signal;
+
+ if ((arg_settings_mask & SETTING_PERSONALITY) == 0 &&
+ settings->personality != PERSONALITY_INVALID)
+ arg_personality = settings->personality;
+
+ if ((arg_settings_mask & SETTING_MACHINE_ID) == 0 &&
+ !sd_id128_is_null(settings->machine_id)) {
+
+ if (!arg_settings_trusted)
+ log_warning("Ignoring MachineID= setting, file %s is not trusted.", p);
+ else
+ arg_uuid = settings->machine_id;
+ }
+
+ if ((arg_settings_mask & SETTING_READ_ONLY) == 0 &&
+ settings->read_only >= 0)
+ arg_read_only = settings->read_only;
+
+ if ((arg_settings_mask & SETTING_VOLATILE_MODE) == 0 &&
+ settings->volatile_mode != _VOLATILE_MODE_INVALID)
+ arg_volatile_mode = settings->volatile_mode;
+
+ if ((arg_settings_mask & SETTING_CUSTOM_MOUNTS) == 0 &&
+ settings->n_custom_mounts > 0) {
+
+ if (!arg_settings_trusted)
+ log_warning("Ignoring TemporaryFileSystem=, Bind= and BindReadOnly= settings, file %s is not trusted.", p);
+ else {
+ custom_mount_free_all(arg_custom_mounts, arg_n_custom_mounts);
+ arg_custom_mounts = settings->custom_mounts;
+ arg_n_custom_mounts = settings->n_custom_mounts;
+
+ settings->custom_mounts = NULL;
+ settings->n_custom_mounts = 0;
+ }
+ }
+
+ if ((arg_settings_mask & SETTING_NETWORK) == 0 &&
+ (settings->private_network >= 0 ||
+ settings->network_veth >= 0 ||
+ settings->network_bridge ||
+ settings->network_zone ||
+ settings->network_interfaces ||
+ settings->network_macvlan ||
+ settings->network_ipvlan ||
+ settings->network_veth_extra)) {
+
+ if (!arg_settings_trusted)
+ log_warning("Ignoring network settings, file %s is not trusted.", p);
+ else {
+ arg_network_veth = settings_network_veth(settings);
+ arg_private_network = settings_private_network(settings);
+
+ strv_free(arg_network_interfaces);
+ arg_network_interfaces = settings->network_interfaces;
+ settings->network_interfaces = NULL;
+
+ strv_free(arg_network_macvlan);
+ arg_network_macvlan = settings->network_macvlan;
+ settings->network_macvlan = NULL;
+
+ strv_free(arg_network_ipvlan);
+ arg_network_ipvlan = settings->network_ipvlan;
+ settings->network_ipvlan = NULL;
+
+ strv_free(arg_network_veth_extra);
+ arg_network_veth_extra = settings->network_veth_extra;
+ settings->network_veth_extra = NULL;
+
+ free(arg_network_bridge);
+ arg_network_bridge = settings->network_bridge;
+ settings->network_bridge = NULL;
+
+ free(arg_network_zone);
+ arg_network_zone = settings->network_zone;
+ settings->network_zone = NULL;
+ }
+ }
+
+ if ((arg_settings_mask & SETTING_EXPOSE_PORTS) == 0 &&
+ settings->expose_ports) {
+
+ if (!arg_settings_trusted)
+ log_warning("Ignoring Port= setting, file %s is not trusted.", p);
+ else {
+ expose_port_free_all(arg_expose_ports);
+ arg_expose_ports = settings->expose_ports;
+ settings->expose_ports = NULL;
+ }
+ }
+
+ if ((arg_settings_mask & SETTING_USERNS) == 0 &&
+ settings->userns_mode != _USER_NAMESPACE_MODE_INVALID) {
+
+ if (!arg_settings_trusted)
+ log_warning("Ignoring PrivateUsers= and PrivateUsersChown= settings, file %s is not trusted.", p);
+ else {
+ arg_userns_mode = settings->userns_mode;
+ arg_uid_shift = settings->uid_shift;
+ arg_uid_range = settings->uid_range;
+ arg_userns_chown = settings->userns_chown;
+ }
+ }
+
+ if ((arg_settings_mask & SETTING_NOTIFY_READY) == 0)
+ arg_notify_ready = settings->notify_ready;
+
+ return 0;
+}
+
+static int run(int master,
+ const char* console,
+ const char *root_device, bool root_device_rw,
+ const char *home_device, bool home_device_rw,
+ const char *srv_device, bool srv_device_rw,
+ const char *esp_device,
+ bool interactive,
+ bool secondary,
+ FDSet *fds,
+ char veth_name[IFNAMSIZ], bool *veth_created,
+ union in_addr_union *exposed,
+ pid_t *pid, int *ret) {
+
+ static const struct sigaction sa = {
+ .sa_handler = nop_signal_handler,
+ .sa_flags = SA_NOCLDSTOP,
+ };
+
+ _cleanup_release_lock_file_ LockFile uid_shift_lock = LOCK_FILE_INIT;
+ _cleanup_close_ int etc_passwd_lock = -1;
+ _cleanup_close_pair_ int
+ kmsg_socket_pair[2] = { -1, -1 },
+ rtnl_socket_pair[2] = { -1, -1 },
+ pid_socket_pair[2] = { -1, -1 },
+ uuid_socket_pair[2] = { -1, -1 },
+ notify_socket_pair[2] = { -1, -1 },
+ uid_shift_socket_pair[2] = { -1, -1 };
+ _cleanup_close_ int notify_socket= -1;
+ _cleanup_(barrier_destroy) Barrier barrier = BARRIER_NULL;
+ _cleanup_(sd_event_unrefp) sd_event *event = NULL;
+ _cleanup_(pty_forward_freep) PTYForward *forward = NULL;
+ _cleanup_(sd_netlink_unrefp) sd_netlink *rtnl = NULL;
+ ContainerStatus container_status = 0;
+ char last_char = 0;
+ int ifi = 0, r;
+ ssize_t l;
+ sigset_t mask_chld;
+
+ assert_se(sigemptyset(&mask_chld) == 0);
+ assert_se(sigaddset(&mask_chld, SIGCHLD) == 0);
+
+ if (arg_userns_mode == USER_NAMESPACE_PICK) {
+ /* When we shall pick the UID/GID range, let's first lock /etc/passwd, so that we can safely
+ * check with getpwuid() if the specific user already exists. Note that /etc might be
+ * read-only, in which case this will fail with EROFS. But that's really OK, as in that case we
+ * can be reasonably sure that no users are going to be added. Note that getpwuid() checks are
+ * really just an extra safety net. We kinda assume that the UID range we allocate from is
+ * really ours. */
+
+ etc_passwd_lock = take_etc_passwd_lock(NULL);
+ if (etc_passwd_lock < 0 && etc_passwd_lock != -EROFS)
+ return log_error_errno(etc_passwd_lock, "Failed to take /etc/passwd lock: %m");
+ }
+
+ r = barrier_create(&barrier);
+ if (r < 0)
+ return log_error_errno(r, "Cannot initialize IPC barrier: %m");
+
+ if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, kmsg_socket_pair) < 0)
+ return log_error_errno(errno, "Failed to create kmsg socket pair: %m");
+
+ if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, rtnl_socket_pair) < 0)
+ return log_error_errno(errno, "Failed to create rtnl socket pair: %m");
+
+ if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, pid_socket_pair) < 0)
+ return log_error_errno(errno, "Failed to create pid socket pair: %m");
+
+ if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, uuid_socket_pair) < 0)
+ return log_error_errno(errno, "Failed to create id socket pair: %m");
+
+ if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, notify_socket_pair) < 0)
+ return log_error_errno(errno, "Failed to create notify socket pair: %m");
+
+ if (arg_userns_mode != USER_NAMESPACE_NO)
+ if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, uid_shift_socket_pair) < 0)
+ return log_error_errno(errno, "Failed to create uid shift socket pair: %m");
+
+ /* Child can be killed before execv(), so handle SIGCHLD in order to interrupt
+ * parent's blocking calls and give it a chance to call wait() and terminate. */
+ r = sigprocmask(SIG_UNBLOCK, &mask_chld, NULL);
+ if (r < 0)
+ return log_error_errno(errno, "Failed to change the signal mask: %m");
+
+ r = sigaction(SIGCHLD, &sa, NULL);
+ if (r < 0)
+ return log_error_errno(errno, "Failed to install SIGCHLD handler: %m");
+
+ *pid = raw_clone(SIGCHLD|CLONE_NEWNS);
+ if (*pid < 0)
+ return log_error_errno(errno, "clone() failed%s: %m",
+ errno == EINVAL ?
+ ", do you have namespace support enabled in your kernel? (You need UTS, IPC, PID and NET namespacing built in)" : "");
+
+ if (*pid == 0) {
+ /* The outer child only has a file system namespace. */
+ barrier_set_role(&barrier, BARRIER_CHILD);
+
+ master = safe_close(master);
+
+ kmsg_socket_pair[0] = safe_close(kmsg_socket_pair[0]);
+ rtnl_socket_pair[0] = safe_close(rtnl_socket_pair[0]);
+ pid_socket_pair[0] = safe_close(pid_socket_pair[0]);
+ uuid_socket_pair[0] = safe_close(uuid_socket_pair[0]);
+ notify_socket_pair[0] = safe_close(notify_socket_pair[0]);
+ uid_shift_socket_pair[0] = safe_close(uid_shift_socket_pair[0]);
+
+ (void) reset_all_signal_handlers();
+ (void) reset_signal_mask();
+
+ r = outer_child(&barrier,
+ arg_directory,
+ console,
+ root_device, root_device_rw,
+ home_device, home_device_rw,
+ srv_device, srv_device_rw,
+ esp_device,
+ interactive,
+ secondary,
+ pid_socket_pair[1],
+ uuid_socket_pair[1],
+ notify_socket_pair[1],
+ kmsg_socket_pair[1],
+ rtnl_socket_pair[1],
+ uid_shift_socket_pair[1],
+ fds);
+ if (r < 0)
+ _exit(EXIT_FAILURE);
+
+ _exit(EXIT_SUCCESS);
+ }
+
+ barrier_set_role(&barrier, BARRIER_PARENT);
+
+ fds = fdset_free(fds);
+
+ kmsg_socket_pair[1] = safe_close(kmsg_socket_pair[1]);
+ rtnl_socket_pair[1] = safe_close(rtnl_socket_pair[1]);
+ pid_socket_pair[1] = safe_close(pid_socket_pair[1]);
+ uuid_socket_pair[1] = safe_close(uuid_socket_pair[1]);
+ notify_socket_pair[1] = safe_close(notify_socket_pair[1]);
+ uid_shift_socket_pair[1] = safe_close(uid_shift_socket_pair[1]);
+
+ if (arg_userns_mode != USER_NAMESPACE_NO) {
+ /* The child just let us know the UID shift it might have read from the image. */
+ l = recv(uid_shift_socket_pair[0], &arg_uid_shift, sizeof arg_uid_shift, 0);
+ if (l < 0)
+ return log_error_errno(errno, "Failed to read UID shift: %m");
+
+ if (l != sizeof arg_uid_shift) {
+ log_error("Short read while reading UID shift.");
+ return -EIO;
+ }
+
+ if (arg_userns_mode == USER_NAMESPACE_PICK) {
+ /* If we are supposed to pick the UID shift, let's try to use the shift read from the
+ * image, but if that's already in use, pick a new one, and report back to the child,
+ * which one we now picked. */
+
+ r = uid_shift_pick(&arg_uid_shift, &uid_shift_lock);
+ if (r < 0)
+ return log_error_errno(r, "Failed to pick suitable UID/GID range: %m");
+
+ l = send(uid_shift_socket_pair[0], &arg_uid_shift, sizeof arg_uid_shift, MSG_NOSIGNAL);
+ if (l < 0)
+ return log_error_errno(errno, "Failed to send UID shift: %m");
+ if (l != sizeof arg_uid_shift) {
+ log_error("Short write while writing UID shift.");
+ return -EIO;
+ }
+ }
+ }
+
+ /* Wait for the outer child. */
+ r = wait_for_terminate_and_warn("namespace helper", *pid, NULL);
+ if (r != 0)
+ return r < 0 ? r : -EIO;
+
+ /* And now retrieve the PID of the inner child. */
+ l = recv(pid_socket_pair[0], pid, sizeof *pid, 0);
+ if (l < 0)
+ return log_error_errno(errno, "Failed to read inner child PID: %m");
+ if (l != sizeof *pid) {
+ log_error("Short read while reading inner child PID.");
+ return -EIO;
+ }
+
+ /* We also retrieve container UUID in case it was generated by outer child */
+ l = recv(uuid_socket_pair[0], &arg_uuid, sizeof arg_uuid, 0);
+ if (l < 0)
+ return log_error_errno(errno, "Failed to read container machine ID: %m");
+ if (l != sizeof(arg_uuid)) {
+ log_error("Short read while reading container machined ID.");
+ return -EIO;
+ }
+
+ /* We also retrieve the socket used for notifications generated by outer child */
+ notify_socket = receive_one_fd(notify_socket_pair[0], 0);
+ if (notify_socket < 0)
+ return log_error_errno(notify_socket,
+ "Failed to receive notification socket from the outer child: %m");
+
+ log_debug("Init process invoked as PID "PID_FMT, *pid);
+
+ if (arg_userns_mode != USER_NAMESPACE_NO) {
+ if (!barrier_place_and_sync(&barrier)) { /* #1 */
+ log_error("Child died too early.");
+ return -ESRCH;
+ }
+
+ r = setup_uid_map(*pid);
+ if (r < 0)
+ return r;
+
+ (void) barrier_place(&barrier); /* #2 */
+ }
+
+ if (arg_private_network) {
+
+ r = move_network_interfaces(*pid, arg_network_interfaces);
+ if (r < 0)
+ return r;
+
+ if (arg_network_veth) {
+ r = setup_veth(arg_machine, *pid, veth_name,
+ arg_network_bridge || arg_network_zone);
+ if (r < 0)
+ return r;
+ else if (r > 0)
+ ifi = r;
+
+ if (arg_network_bridge) {
+ /* Add the interface to a bridge */
+ r = setup_bridge(veth_name, arg_network_bridge, false);
+ if (r < 0)
+ return r;
+ if (r > 0)
+ ifi = r;
+ } else if (arg_network_zone) {
+ /* Add the interface to a bridge, possibly creating it */
+ r = setup_bridge(veth_name, arg_network_zone, true);
+ if (r < 0)
+ return r;
+ if (r > 0)
+ ifi = r;
+ }
+ }
+
+ r = setup_veth_extra(arg_machine, *pid, arg_network_veth_extra);
+ if (r < 0)
+ return r;
+
+ /* We created the primary and extra veth links now; let's remember this, so that we know to
+ remove them later on. Note that we don't bother with removing veth links that were created
+ here when their setup failed half-way, because in that case the kernel should be able to
+ remove them on its own, since they cannot be referenced by anything yet. */
+ *veth_created = true;
+
+ r = setup_macvlan(arg_machine, *pid, arg_network_macvlan);
+ if (r < 0)
+ return r;
+
+ r = setup_ipvlan(arg_machine, *pid, arg_network_ipvlan);
+ if (r < 0)
+ return r;
+ }
+
+ if (arg_register) {
+ r = register_machine(
+ arg_machine,
+ *pid,
+ arg_directory,
+ arg_uuid,
+ ifi,
+ arg_slice,
+ arg_custom_mounts, arg_n_custom_mounts,
+ arg_kill_signal,
+ arg_property,
+ arg_keep_unit,
+ arg_container_service_name);
+ if (r < 0)
+ return r;
+ }
+
+ r = sync_cgroup(*pid, arg_unified_cgroup_hierarchy, arg_uid_shift);
+ if (r < 0)
+ return r;
+
+ if (arg_keep_unit) {
+ r = create_subcgroup(*pid, arg_unified_cgroup_hierarchy);
+ if (r < 0)
+ return r;
+ }
+
+ r = chown_cgroup(*pid, arg_uid_shift);
+ if (r < 0)
+ return r;
+
+ /* Notify the child that the parent is ready with all
+ * its setup (including cgroup-ification), and that
+ * the child can now hand over control to the code to
+ * run inside the container. */
+ (void) barrier_place(&barrier); /* #3 */
+
+ /* Block SIGCHLD here, before notifying child.
+ * process_pty() will handle it with the other signals. */
+ assert_se(sigprocmask(SIG_BLOCK, &mask_chld, NULL) >= 0);
+
+ /* Reset signal to default */
+ r = default_signals(SIGCHLD, -1);
+ if (r < 0)
+ return log_error_errno(r, "Failed to reset SIGCHLD: %m");
+
+ r = sd_event_new(&event);
+ if (r < 0)
+ return log_error_errno(r, "Failed to get default event source: %m");
+
+ r = setup_sd_notify_parent(event, notify_socket, PID_TO_PTR(*pid));
+ if (r < 0)
+ return r;
+
+ /* Let the child know that we are ready and wait that the child is completely ready now. */
+ if (!barrier_place_and_sync(&barrier)) { /* #4 */
+ log_error("Child died too early.");
+ return -ESRCH;
+ }
+
+ /* At this point we have made use of the UID we picked, and thus nss-mymachines
+ * will make them appear in getpwuid(), thus we can release the /etc/passwd lock. */
+ etc_passwd_lock = safe_close(etc_passwd_lock);
+
+ sd_notifyf(false,
+ "STATUS=Container running.\n"
+ "X_NSPAWN_LEADER_PID=" PID_FMT, *pid);
+ if (!arg_notify_ready)
+ sd_notify(false, "READY=1\n");
+
+ if (arg_kill_signal > 0) {
+ /* Try to kill the init system on SIGINT or SIGTERM */
+ sd_event_add_signal(event, NULL, SIGINT, on_orderly_shutdown, PID_TO_PTR(*pid));
+ sd_event_add_signal(event, NULL, SIGTERM, on_orderly_shutdown, PID_TO_PTR(*pid));
+ } else {
+ /* Immediately exit */
+ sd_event_add_signal(event, NULL, SIGINT, NULL, NULL);
+ sd_event_add_signal(event, NULL, SIGTERM, NULL, NULL);
+ }
+
+ /* simply exit on sigchld */
+ sd_event_add_signal(event, NULL, SIGCHLD, NULL, NULL);
+
+ if (arg_expose_ports) {
+ r = expose_port_watch_rtnl(event, rtnl_socket_pair[0], on_address_change, exposed, &rtnl);
+ if (r < 0)
+ return r;
+
+ (void) expose_port_execute(rtnl, arg_expose_ports, exposed);
+ }
+
+ rtnl_socket_pair[0] = safe_close(rtnl_socket_pair[0]);
+
+ r = pty_forward_new(event, master,
+ PTY_FORWARD_IGNORE_VHANGUP | (interactive ? 0 : PTY_FORWARD_READ_ONLY),
+ &forward);
+ if (r < 0)
+ return log_error_errno(r, "Failed to create PTY forwarder: %m");
+
+ r = sd_event_loop(event);
+ if (r < 0)
+ return log_error_errno(r, "Failed to run event loop: %m");
+
+ pty_forward_get_last_char(forward, &last_char);
+
+ forward = pty_forward_free(forward);
+
+ if (!arg_quiet && last_char != '\n')
+ putc('\n', stdout);
+
+ /* Kill if it is not dead yet anyway */
+ if (arg_register && !arg_keep_unit)
+ terminate_machine(*pid);
+
+ /* Normally redundant, but better safe than sorry */
+ kill(*pid, SIGKILL);
+
+ r = wait_for_container(*pid, &container_status);
+ *pid = 0;
+
+ if (r < 0)
+ /* We failed to wait for the container, or the container exited abnormally. */
+ return r;
+ if (r > 0 || container_status == CONTAINER_TERMINATED) {
+ /* r > 0 → The container exited with a non-zero status.
+ * As a special case, we need to replace 133 with a different value,
+ * because 133 is special-cased in the service file to reboot the container.
+ * otherwise → The container exited with zero status and a reboot was not requested.
+ */
+ if (r == 133)
+ r = EXIT_FAILURE; /* replace 133 with the general failure code */
+ *ret = r;
+ return 0; /* finito */
+ }
+
+ /* CONTAINER_REBOOTED, loop again */
+
+ if (arg_keep_unit) {
+ /* Special handling if we are running as a service: instead of simply
+ * restarting the machine we want to restart the entire service, so let's
+ * inform systemd about this with the special exit code 133. The service
+ * file uses RestartForceExitStatus=133 so that this results in a full
+ * nspawn restart. This is necessary since we might have cgroup parameters
+ * set we want to have flushed out. */
+ *ret = 0;
+ return 133;
+ }
+
+ expose_port_flush(arg_expose_ports, exposed);
+
+ (void) remove_veth_links(veth_name, arg_network_veth_extra);
+ *veth_created = false;
+ return 1; /* loop again */
+}
+
+int main(int argc, char *argv[]) {
+
+ _cleanup_free_ char *device_path = NULL, *root_device = NULL, *home_device = NULL, *srv_device = NULL, *esp_device = NULL, *console = NULL;
+ bool root_device_rw = true, home_device_rw = true, srv_device_rw = true;
+ _cleanup_close_ int master = -1, image_fd = -1;
+ _cleanup_fdset_free_ FDSet *fds = NULL;
+ int r, n_fd_passed, loop_nr = -1, ret = EXIT_SUCCESS;
+ char veth_name[IFNAMSIZ] = "";
+ bool secondary = false, remove_subvol = false;
+ pid_t pid = 0;
+ union in_addr_union exposed = {};
+ _cleanup_release_lock_file_ LockFile tree_global_lock = LOCK_FILE_INIT, tree_local_lock = LOCK_FILE_INIT;
+ bool interactive, veth_created = false;
+
+ log_parse_environment();
+ log_open();
+
+ /* Make sure rename_process() in the stub init process can work */
+ saved_argv = argv;
+ saved_argc = argc;
+
+ r = parse_argv(argc, argv);
+ if (r <= 0)
+ goto finish;
+
+ if (geteuid() != 0) {
+ log_error("Need to be root.");
+ r = -EPERM;
+ goto finish;
+ }
+ r = determine_names();
+ if (r < 0)
+ goto finish;
+
+ r = load_settings();
+ if (r < 0)
+ goto finish;
+
+ r = verify_arguments();
+ if (r < 0)
+ goto finish;
+
+ n_fd_passed = sd_listen_fds(false);
+ if (n_fd_passed > 0) {
+ r = fdset_new_listen_fds(&fds, false);
+ if (r < 0) {
+ log_error_errno(r, "Failed to collect file descriptors: %m");
+ goto finish;
+ }
+ }
+
+ if (arg_directory) {
+ assert(!arg_image);
+
+ if (path_equal(arg_directory, "/") && !arg_ephemeral) {
+ log_error("Spawning container on root directory is not supported. Consider using --ephemeral.");
+ r = -EINVAL;
+ goto finish;
+ }
+
+ if (arg_ephemeral) {
+ _cleanup_free_ char *np = NULL;
+
+ /* If the specified path is a mount point we
+ * generate the new snapshot immediately
+ * inside it under a random name. However if
+ * the specified is not a mount point we
+ * create the new snapshot in the parent
+ * directory, just next to it. */
+ r = path_is_mount_point(arg_directory, 0);
+ if (r < 0) {
+ log_error_errno(r, "Failed to determine whether directory %s is mount point: %m", arg_directory);
+ goto finish;
+ }
+ if (r > 0)
+ r = tempfn_random_child(arg_directory, "machine.", &np);
+ else
+ r = tempfn_random(arg_directory, "machine.", &np);
+ if (r < 0) {
+ log_error_errno(r, "Failed to generate name for snapshot: %m");
+ goto finish;
+ }
+
+ r = image_path_lock(np, (arg_read_only ? LOCK_SH : LOCK_EX) | LOCK_NB, &tree_global_lock, &tree_local_lock);
+ if (r < 0) {
+ log_error_errno(r, "Failed to lock %s: %m", np);
+ goto finish;
+ }
+
+ r = btrfs_subvol_snapshot(arg_directory, np, (arg_read_only ? BTRFS_SNAPSHOT_READ_ONLY : 0) | BTRFS_SNAPSHOT_FALLBACK_COPY | BTRFS_SNAPSHOT_RECURSIVE | BTRFS_SNAPSHOT_QUOTA);
+ if (r < 0) {
+ log_error_errno(r, "Failed to create snapshot %s from %s: %m", np, arg_directory);
+ goto finish;
+ }
+
+ free(arg_directory);
+ arg_directory = np;
+ np = NULL;
+
+ remove_subvol = true;
+
+ } else {
+ r = image_path_lock(arg_directory, (arg_read_only ? LOCK_SH : LOCK_EX) | LOCK_NB, &tree_global_lock, &tree_local_lock);
+ if (r == -EBUSY) {
+ log_error_errno(r, "Directory tree %s is currently busy.", arg_directory);
+ goto finish;
+ }
+ if (r < 0) {
+ log_error_errno(r, "Failed to lock %s: %m", arg_directory);
+ goto finish;
+ }
+
+ if (arg_template) {
+ r = btrfs_subvol_snapshot(arg_template, arg_directory, (arg_read_only ? BTRFS_SNAPSHOT_READ_ONLY : 0) | BTRFS_SNAPSHOT_FALLBACK_COPY | BTRFS_SNAPSHOT_RECURSIVE | BTRFS_SNAPSHOT_QUOTA);
+ if (r == -EEXIST) {
+ if (!arg_quiet)
+ log_info("Directory %s already exists, not populating from template %s.", arg_directory, arg_template);
+ } else if (r < 0) {
+ log_error_errno(r, "Couldn't create snapshot %s from %s: %m", arg_directory, arg_template);
+ goto finish;
+ } else {
+ if (!arg_quiet)
+ log_info("Populated %s from template %s.", arg_directory, arg_template);
+ }
+ }
+ }
+
+ if (arg_start_mode == START_BOOT) {
+ if (path_is_os_tree(arg_directory) <= 0) {
+ log_error("Directory %s doesn't look like an OS root directory (os-release file is missing). Refusing.", arg_directory);
+ r = -EINVAL;
+ goto finish;
+ }
+ } else {
+ const char *p;
+
+ p = strjoina(arg_directory, "/usr/");
+ if (laccess(p, F_OK) < 0) {
+ log_error("Directory %s doesn't look like it has an OS tree. Refusing.", arg_directory);
+ r = -EINVAL;
+ goto finish;
+ }
+ }
+
+ } else {
+ char template[] = "/tmp/nspawn-root-XXXXXX";
+
+ assert(arg_image);
+ assert(!arg_template);
+
+ r = image_path_lock(arg_image, (arg_read_only ? LOCK_SH : LOCK_EX) | LOCK_NB, &tree_global_lock, &tree_local_lock);
+ if (r == -EBUSY) {
+ r = log_error_errno(r, "Disk image %s is currently busy.", arg_image);
+ goto finish;
+ }
+ if (r < 0) {
+ r = log_error_errno(r, "Failed to create image lock: %m");
+ goto finish;
+ }
+
+ if (!mkdtemp(template)) {
+ log_error_errno(errno, "Failed to create temporary directory: %m");
+ r = -errno;
+ goto finish;
+ }
+
+ arg_directory = strdup(template);
+ if (!arg_directory) {
+ r = log_oom();
+ goto finish;
+ }
+
+ image_fd = setup_image(&device_path, &loop_nr);
+ if (image_fd < 0) {
+ r = image_fd;
+ goto finish;
+ }
+
+ r = dissect_image(image_fd,
+ &root_device, &root_device_rw,
+ &home_device, &home_device_rw,
+ &srv_device, &srv_device_rw,
+ &esp_device,
+ &secondary);
+ if (r < 0)
+ goto finish;
+ }
+
+ r = custom_mounts_prepare();
+ if (r < 0)
+ goto finish;
+
+ interactive =
+ isatty(STDIN_FILENO) > 0 &&
+ isatty(STDOUT_FILENO) > 0;
+
+ master = posix_openpt(O_RDWR|O_NOCTTY|O_CLOEXEC|O_NDELAY);
+ if (master < 0) {
+ r = log_error_errno(errno, "Failed to acquire pseudo tty: %m");
+ goto finish;
+ }
+
+ r = ptsname_malloc(master, &console);
+ if (r < 0) {
+ r = log_error_errno(r, "Failed to determine tty name: %m");
+ goto finish;
+ }
+
+ if (arg_selinux_apifs_context) {
+ r = mac_selinux_apply(console, arg_selinux_apifs_context);
+ if (r < 0)
+ goto finish;
+ }
+
+ if (unlockpt(master) < 0) {
+ r = log_error_errno(errno, "Failed to unlock tty: %m");
+ goto finish;
+ }
+
+ if (!arg_quiet)
+ log_info("Spawning container %s on %s.\nPress ^] three times within 1s to kill container.",
+ arg_machine, arg_image ?: arg_directory);
+
+ assert_se(sigprocmask_many(SIG_BLOCK, NULL, SIGCHLD, SIGWINCH, SIGTERM, SIGINT, -1) >= 0);
+
+ if (prctl(PR_SET_CHILD_SUBREAPER, 1) < 0) {
+ r = log_error_errno(errno, "Failed to become subreaper: %m");
+ goto finish;
+ }
+
+ for (;;) {
+ r = run(master,
+ console,
+ root_device, root_device_rw,
+ home_device, home_device_rw,
+ srv_device, srv_device_rw,
+ esp_device,
+ interactive, secondary,
+ fds,
+ veth_name, &veth_created,
+ &exposed,
+ &pid, &ret);
+ if (r <= 0)
+ break;
+ }
+
+finish:
+ sd_notify(false,
+ "STOPPING=1\n"
+ "STATUS=Terminating...");
+
+ if (pid > 0)
+ kill(pid, SIGKILL);
+
+ /* Try to flush whatever is still queued in the pty */
+ if (master >= 0)
+ (void) copy_bytes(master, STDOUT_FILENO, (uint64_t) -1, false);
+
+ loop_remove(loop_nr, &image_fd);
+
+ if (remove_subvol && arg_directory) {
+ int k;
+
+ k = btrfs_subvol_remove(arg_directory, BTRFS_REMOVE_RECURSIVE|BTRFS_REMOVE_QUOTA);
+ if (k < 0)
+ log_warning_errno(k, "Cannot remove subvolume '%s', ignoring: %m", arg_directory);
+ }
+
+ if (arg_machine) {
+ const char *p;
+
+ p = strjoina("/run/systemd/nspawn/propagate/", arg_machine);
+ (void) rm_rf(p, REMOVE_ROOT);
+ }
+
+ expose_port_flush(arg_expose_ports, &exposed);
+
+ if (veth_created)
+ (void) remove_veth_links(veth_name, arg_network_veth_extra);
+ (void) remove_bridge(arg_network_zone);
+
+ free(arg_directory);
+ free(arg_template);
+ free(arg_image);
+ free(arg_machine);
+ free(arg_user);
+ free(arg_chdir);
+ strv_free(arg_setenv);
+ free(arg_network_bridge);
+ strv_free(arg_network_interfaces);
+ strv_free(arg_network_macvlan);
+ strv_free(arg_network_ipvlan);
+ strv_free(arg_network_veth_extra);
+ strv_free(arg_parameters);
+ custom_mount_free_all(arg_custom_mounts, arg_n_custom_mounts);
+ expose_port_free_all(arg_expose_ports);
+
+ return r < 0 ? EXIT_FAILURE : ret;
+}
diff --git a/src/systemd-nspawn/systemd-nspawn.completion.bash b/src/systemd-nspawn/systemd-nspawn.completion.bash
new file mode 100644
index 0000000000..ea4a5e1f43
--- /dev/null
+++ b/src/systemd-nspawn/systemd-nspawn.completion.bash
@@ -0,0 +1,155 @@
+# systemd-nspawn(1) completion -*- shell-script -*-
+#
+# This file is part of systemd.
+#
+# Copyright 2014 Thomas H.P. Andersen
+#
+# systemd is free software; you can redistribute it and/or modify it
+# under the terms of the GNU Lesser General Public License as published by
+# the Free Software Foundation; either version 2.1 of the License, or
+# (at your option) any later version.
+#
+# systemd is distributed in the hope that it will be useful, but
+# WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+# General Public License for more details.
+#
+# You should have received a copy of the GNU Lesser General Public License
+# along with systemd; If not, see <http://www.gnu.org/licenses/>.
+
+__contains_word() {
+ local w word=$1; shift
+ for w in "$@"; do
+ [[ $w = "$word" ]] && return
+ done
+}
+
+__get_users() {
+ local a b
+ loginctl list-users --no-legend --no-pager | { while read a b; do echo " $b"; done; };
+}
+
+__get_slices() {
+ local a b
+ systemctl list-units -t slice --no-legend --no-pager | { while read a b; do echo " $a"; done; };
+}
+
+__get_machines() {
+ local a b
+ machinectl list --no-legend --no-pager | { while read a b; do echo " $a"; done; };
+}
+
+__get_env() {
+ local a
+ env | { while read a; do echo " ${a%%=*}"; done; };
+}
+
+__get_interfaces(){
+ { cd /sys/class/net && echo *; } | \
+ while read -d' ' -r name; do
+ [[ "$name" != "lo" ]] && echo "$name"
+ done
+}
+
+_systemd_nspawn() {
+ local cur=${COMP_WORDS[COMP_CWORD]} prev=${COMP_WORDS[COMP_CWORD-1]}
+ local i verb comps
+
+ local -A OPTS=(
+ [STANDALONE]='-h --help --version --private-network -b --boot --read-only -q --quiet --share-system --keep-unit --network-veth -j'
+ [ARG]='-D --directory -u --user --uuid --capability --drop-capability --link-journal --bind --bind-ro -M --machine
+ -S --slice --setenv -Z --selinux-context -L --selinux-apifs-context --register --network-interface --network-bridge
+ --personality -i --image --tmpfs --volatile
+ --network-macvlan --kill-signal --template
+ --notify-ready'
+ )
+
+ _init_completion || return
+
+ if __contains_word "$prev" ${OPTS[ARG]}; then
+ case $prev in
+ --directory|-D|--template)
+ compopt -o nospace
+ comps=$(compgen -S/ -A directory -- "$cur" )
+ ;;
+ --user|-u)
+ comps=$( __get_users )
+ ;;
+ --uuid)
+ comps=''
+ ;;
+ --capability)
+ comps='CAP_BLOCK_SUSPEND CAP_IPC_LOCK CAP_MAC_ADMIN CAP_MAC_OVERRIDE CAP_SYS_MODULE CAP_SYS_PACCT CAP_SYS_RAWIO
+ CAP_SYS_TIME CAP_SYSLOG CAP_WAKE_ALARM CAP_NET_ADMIN'
+ ;;
+ --drop-capability)
+ comps='CAP_AUDIT_CONTROL CAP_AUDIT_WRITE CAP_CHOWN CAP_DAC_OVERRIDE CAP_DAC_READ_SEARCH CAP_FOWNER CAP_FSETID
+ CAP_IPC_OWNER CAP_KILL CAP_LEASE CAP_LINUX_IMMUTABLE CAP_MKNOD CAP_NET_ADMIN CAP_NET_BIND_SERVICE
+ CAP_NET_BROADCAST CAP_NET_RAW CAP_SETFCAP CAP_SETGID CAP_SETPCAP CAP_SETUID CAP_SYS_ADMIN CAP_SYS_BOOT
+ CAP_SYS_CHROOT CAP_SYS_NICE CAP_SYS_PTRACE CAP_SYS_RESOURCE CAP_SYS_TTY_CONFIG'
+ ;;
+ --link-journal)
+ comps='no auto guest try-guest host try-host'
+ ;;
+ --bind|--bind-ro)
+ compopt -o nospace
+ comps=$(compgen -S/ -A directory -- "$cur" )
+ ;;
+ --tmpfs)
+ compopt -o nospace
+ comps=$(compgen -S/ -A directory -- "$cur" )
+ ;;
+ --machine|-M)
+ comps=$( __get_machines )
+ ;;
+ --slice|-S)
+ comps=$( __get_slices )
+ ;;
+ --setenv)
+ comps=$( __get_env )
+ ;;
+ --selinux-context|-Z)
+ comps=''
+ ;;
+ --selinux-apifs-context|-L)
+ comps=''
+ ;;
+ --register)
+ comps='yes no'
+ ;;
+ --network-interface)
+ comps=$(__get_interfaces)
+ ;;
+ --network-bridge)
+ comps=''
+ ;;
+ --network-macvlan)
+ comps=''
+ ;;
+ --personality)
+ comps='x86 x86-64'
+ ;;
+ --volatile)
+ comps='yes state no'
+ ;;
+ --image|-i)
+ compopt -o nospace
+ comps=$( compgen -A file -- "$cur" )
+ ;;
+ --kill-signal)
+ _signals
+ return
+ ;;
+ --notify-ready)
+ comps='yes no'
+ return
+ ;;
+ esac
+ COMPREPLY=( $(compgen -W '$comps' -- "$cur") )
+ return 0
+ fi
+
+ COMPREPLY=( $(compgen -W '${OPTS[*]}' -- "$cur") )
+}
+
+complete -F _systemd_nspawn systemd-nspawn
diff --git a/src/systemd-nspawn/systemd-nspawn.completion.zsh b/src/systemd-nspawn/systemd-nspawn.completion.zsh
new file mode 100644
index 0000000000..77b2e7cd7c
--- /dev/null
+++ b/src/systemd-nspawn/systemd-nspawn.completion.zsh
@@ -0,0 +1,50 @@
+#compdef systemd-nspawn
+
+_nspawn-caps(){
+ local -a _caps
+ _caps=( CAP_CHOWN CAP_DAC_OVERRIDE CAP_DAC_READ_SEARCH
+ CAP_FOWNER CAP_FSETID CAP_IPC_OWNER CAP_KILL CAP_LEASE CAP_LINUX_IMMUTABLE
+ CAP_NET_BIND_SERVICE CAP_NET_BROADCAST CAP_NET_RAW CAP_SETGID CAP_SETFCAP CAP_SETPCAP
+ CAP_SETUID CAP_SYS_ADMIN CAP_SYS_CHROOT CAP_SYS_NICE CAP_SYS_PTRACE CAP_SYS_TTY_CONFIG
+ CAP_SYS_RESOURCE CAP_SYS_BOOT )
+ _values -s , 'capabilities' "$_caps[@]"
+}
+
+_arguments \
+ {-h,--help}'[Show this help.]' \
+ '--version[Print a short version string and exit.]' \
+ {--quiet,-q}'[Turns off any status output by the tool itself.]' \
+ {--directory=,-D+}'[Directory to use as file system root for the namespace container. If omitted the current directory will be used.]:directories:_directories' \
+ '--template=[Initialize root directory from template directory, if missing.]:template:_directories' \
+ {--ephemeral,-x}'[Run container with snapshot of root directory, and remove it after exit.]' \
+ {--image=,-i+}'[Disk image to mount the root directory for the container from.]:disk image: _files' \
+ {--boot,-b}'[Automatically search for an init binary and invoke it instead of a shell or a user supplied program.]' \
+ {--user=,-u+}'[Run the command under specified user, create home directory and cd into it.]:user:_users' \
+ {--machine=,-M+}'[Sets the machine name for this container.]: : _message "container name"' \
+ '--uuid=[Set the specified uuid for the container.]: : _message "container UUID"' \
+ {--slice=,-S+}'[Make the container part of the specified slice, instead of the default machine.slice.]: : _message slice' \
+ '--private-network[Disconnect networking of the container from the host.]' \
+ '--network-interface=[Assign the specified network interface to the container.]: : _net_interfaces' \
+ '--network-macvlan=[Create a "macvlan" interface of the specified Ethernet network interface and add it to the container.]: : _net_interfaces' \
+ '--network-ipvlan=[Create a ipvlan network interface based on an existing network interface to the container.]: : _net_interfaces' \
+ {--network-veth,-n}'[Create a virtual Ethernet link (veth) between host and container.]' \
+ '--network-bridge=[Adds the host side of the Ethernet link created with --network-veth to the specified bridge.]: : _net_interfaces' \
+ {--port=,-p+}'[Expose a container IP port on the host.]: : _message port' \
+ {--selinux-context=,-Z+}'[Sets the SELinux security context to be used to label processes in the container.]: : _message "SELinux context"' \
+ {--selinux-apifs-context=,-L+}'[Sets the SELinux security context to be used to label files in the virtual API file systems in the container.]: : _message "SELinux context"' \
+ '--capability=[List one or more additional capabilities to grant the container.]:capabilities:_nspawn-caps' \
+ '--drop-capability=[Specify one or more additional capabilities to drop for the containerm]:capabilities:_nspawn-caps' \
+ "--link-journal=[Control whether the container's journal shall be made visible to the host system.]:options:(no host guest auto)" \
+ '-j[Equivalent to --link-journal=guest.]' \
+ '--read-only[Mount the root file system read only for the container.]' \
+ '--bind=[Bind mount a file or directory from the host into the container.]: : _files' \
+ '--bind-ro=[Bind mount a file or directory from the host into the container (read-only).]: : _files' \
+ '--tmpfs=[Mount an empty tmpfs to the specified directory.]: : _files' \
+ '--setenv=[Specifies an environment variable assignment to pass to the init process in the container, in the format "NAME=VALUE".]: : _message "environment variables"' \
+ '--share-system[Allows the container to share certain system facilities with the host.]' \
+ '--register=[Controls whether the container is registered with systemd-machined(8).]:systemd-machined registration:( yes no )' \
+ '--keep-unit[Instead of creating a transient scope unit to run the container in, simply register the service or scope unit systemd-nspawn has been invoked in with systemd-machined(8).]' \
+ '--personality=[Control the architecture ("personality") reported by uname(2) in the container.]:architecture:(x86 x86-64)' \
+ '--volatile=[Run the system in volatile mode.]:volatile:(no yes state)' \
+ "--notify-ready=[Control when the ready notification is sent]:options:(yes no)" \
+ '*:: : _normal'
diff --git a/src/systemd-nspawn/systemd-nspawn.tmpfiles b/src/systemd-nspawn/systemd-nspawn.tmpfiles
new file mode 100644
index 0000000000..78bd1c670e
--- /dev/null
+++ b/src/systemd-nspawn/systemd-nspawn.tmpfiles
@@ -0,0 +1,23 @@
+# This file is part of systemd.
+#
+# systemd is free software; you can redistribute it and/or modify it
+# under the terms of the GNU Lesser General Public License as published by
+# the Free Software Foundation; either version 2.1 of the License, or
+# (at your option) any later version.
+
+# See tmpfiles.d(5) for details
+
+Q /var/lib/machines 0700 - - -
+
+# Remove old temporary snapshots, but only at boot. Ideally we'd have
+# "self-destroying" btrfs snapshots that go away if the last
+# reference to it does. To mimic a scheme like this at least remove
+# the old snapshots on fresh boots, where we know they cannot be
+# referenced anymore. Note that we actually remove all temporary files
+# in /var/lib/machines/ at boot, which should be safe since the
+# directory has defined semantics. In the root directory (where
+# systemd-nspawn --ephemeral places snapshots) we are more strict, to
+# avoid removing unrelated temporary files.
+
+R! /var/lib/machines/.#*
+R! /.#machine.*
diff --git a/src/systemd-nspawn/systemd-nspawn.xml b/src/systemd-nspawn/systemd-nspawn.xml
new file mode 100644
index 0000000000..c449edee89
--- /dev/null
+++ b/src/systemd-nspawn/systemd-nspawn.xml
@@ -0,0 +1,1088 @@
+<?xml version='1.0'?> <!--*- Mode: nxml; nxml-child-indent: 2; indent-tabs-mode: nil -*-->
+<!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN"
+ "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd">
+
+<!--
+ This file is part of systemd.
+
+ Copyright 2010 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+-->
+
+<refentry id="systemd-nspawn"
+ xmlns:xi="http://www.w3.org/2001/XInclude">
+
+ <refentryinfo>
+ <title>systemd-nspawn</title>
+ <productname>systemd</productname>
+
+ <authorgroup>
+ <author>
+ <contrib>Developer</contrib>
+ <firstname>Lennart</firstname>
+ <surname>Poettering</surname>
+ <email>lennart@poettering.net</email>
+ </author>
+ </authorgroup>
+ </refentryinfo>
+
+ <refmeta>
+ <refentrytitle>systemd-nspawn</refentrytitle>
+ <manvolnum>1</manvolnum>
+ </refmeta>
+
+ <refnamediv>
+ <refname>systemd-nspawn</refname>
+ <refpurpose>Spawn a namespace container for debugging, testing and building</refpurpose>
+ </refnamediv>
+
+ <refsynopsisdiv>
+ <cmdsynopsis>
+ <command>systemd-nspawn</command>
+ <arg choice="opt" rep="repeat">OPTIONS</arg>
+ <arg choice="opt"><replaceable>COMMAND</replaceable>
+ <arg choice="opt" rep="repeat">ARGS</arg>
+ </arg>
+ </cmdsynopsis>
+ <cmdsynopsis>
+ <command>systemd-nspawn</command>
+ <arg choice="plain">--boot</arg>
+ <arg choice="opt" rep="repeat">OPTIONS</arg>
+ <arg choice="opt" rep="repeat">ARGS</arg>
+ </cmdsynopsis>
+ </refsynopsisdiv>
+
+ <refsect1>
+ <title>Description</title>
+
+ <para><command>systemd-nspawn</command> may be used to run a command or OS in a light-weight namespace
+ container. In many ways it is similar to <citerefentry
+ project='man-pages'><refentrytitle>chroot</refentrytitle><manvolnum>1</manvolnum></citerefentry>, but more powerful
+ since it fully virtualizes the file system hierarchy, as well as the process tree, the various IPC subsystems and
+ the host and domain name.</para>
+
+ <para><command>systemd-nspawn</command> may be invoked on any directory tree containing an operating system tree,
+ using the <option>--directory=</option> command line option. By using the <option>--machine=</option> option an OS
+ tree is automatically searched for in a couple of locations, most importantly in
+ <filename>/var/lib/machines</filename>, the suggested directory to place container images installed on the
+ system.</para>
+
+ <para>In contrast to <citerefentry
+ project='man-pages'><refentrytitle>chroot</refentrytitle><manvolnum>1</manvolnum></citerefentry> <command>systemd-nspawn</command>
+ may be used to boot full Linux-based operating systems in a container.</para>
+
+ <para><command>systemd-nspawn</command> limits access to various kernel interfaces in the container to read-only,
+ such as <filename>/sys</filename>, <filename>/proc/sys</filename> or <filename>/sys/fs/selinux</filename>. The
+ host's network interfaces and the system clock may not be changed from within the container. Device nodes may not
+ be created. The host system cannot be rebooted and kernel modules may not be loaded from within the
+ container.</para>
+
+ <para>Use a tool like <citerefentry
+ project='mankier'><refentrytitle>dnf</refentrytitle><manvolnum>8</manvolnum></citerefentry>, <citerefentry
+ project='die-net'><refentrytitle>debootstrap</refentrytitle><manvolnum>8</manvolnum></citerefentry>, or
+ <citerefentry project='archlinux'><refentrytitle>pacman</refentrytitle><manvolnum>8</manvolnum></citerefentry> to
+ set up an OS directory tree suitable as file system hierarchy for <command>systemd-nspawn</command> containers. See
+ the Examples section below for details on suitable invocation of these commands.</para>
+
+ <para>As a safety check <command>systemd-nspawn</command> will verify the existence of
+ <filename>/usr/lib/os-release</filename> or <filename>/etc/os-release</filename> in the container tree before
+ starting the container (see
+ <citerefentry><refentrytitle>os-release</refentrytitle><manvolnum>5</manvolnum></citerefentry>). It might be
+ necessary to add this file to the container tree manually if the OS of the container is too old to contain this
+ file out-of-the-box.</para>
+
+ <para><command>systemd-nspawn</command> may be invoked directly from the interactive command line or run as system
+ service in the background. In this mode each container instance runs as its own service instance; a default
+ template unit file <filename>systemd-nspawn@.service</filename> is provided to make this easy, taking the container
+ name as instance identifier. Note that different default options apply when <command>systemd-nspawn</command> is
+ invoked by the template unit file than interactively on the command line. Most importantly the template unit file
+ makes use of the <option>--boot</option> which is not the default in case <command>systemd-nspawn</command> is
+ invoked from the interactive command line. Further differences with the defaults are documented along with the
+ various supported options below.</para>
+
+ <para>The <citerefentry><refentrytitle>machinectl</refentrytitle><manvolnum>1</manvolnum></citerefentry> tool may
+ be used to execute a number of operations on containers. In particular it provides easy-to-use commands to run
+ containers as system services using the <filename>systemd-nspawn@.service</filename> template unit
+ file.</para>
+
+ <para>Along with each container a settings file with the <filename>.nspawn</filename> suffix may exist, containing
+ additional settings to apply when running the container. See
+ <citerefentry><refentrytitle>systemd.nspawn</refentrytitle><manvolnum>5</manvolnum></citerefentry> for
+ details. Settings files override the default options used by the <filename>systemd-nspawn@.service</filename>
+ template unit file, making it usually unnecessary to alter this template file directly.</para>
+
+ <para>Note that <command>systemd-nspawn</command> will mount file systems private to the container to
+ <filename>/dev</filename>, <filename>/run</filename> and similar. These will not be visible outside of the
+ container, and their contents will be lost when the container exits.</para>
+
+ <para>Note that running two <command>systemd-nspawn</command> containers from the same directory tree will not make
+ processes in them see each other. The PID namespace separation of the two containers is complete and the containers
+ will share very few runtime objects except for the underlying file system. Use
+ <citerefentry><refentrytitle>machinectl</refentrytitle><manvolnum>1</manvolnum></citerefentry>'s
+ <command>login</command> or <command>shell</command> commands to request an additional login session in a running
+ container.</para>
+
+ <para><command>systemd-nspawn</command> implements the <ulink
+ url="http://www.freedesktop.org/wiki/Software/systemd/ContainerInterface">Container Interface</ulink>
+ specification.</para>
+
+ <para>While running, containers invoked with <command>systemd-nspawn</command> are registered with the
+ <citerefentry><refentrytitle>systemd-machined</refentrytitle><manvolnum>8</manvolnum></citerefentry> service that
+ keeps track of running containers, and provides programming interfaces to interact with them.</para>
+ </refsect1>
+
+ <refsect1>
+ <title>Options</title>
+
+ <para>If option <option>-b</option> is specified, the arguments
+ are used as arguments for the init binary. Otherwise,
+ <replaceable>COMMAND</replaceable> specifies the program to launch
+ in the container, and the remaining arguments are used as
+ arguments for this program. If <option>--boot</option> is not used and
+ no arguments are specified, a shell is launched in the
+ container.</para>
+
+ <para>The following options are understood:</para>
+
+ <variablelist>
+ <varlistentry>
+ <term><option>-D</option></term>
+ <term><option>--directory=</option></term>
+
+ <listitem><para>Directory to use as file system root for the
+ container.</para>
+
+ <para>If neither <option>--directory=</option>, nor
+ <option>--image=</option> is specified the directory is
+ determined by searching for a directory named the same as the
+ machine name specified with <option>--machine=</option>. See
+ <citerefentry><refentrytitle>machinectl</refentrytitle><manvolnum>1</manvolnum></citerefentry>
+ section "Files and Directories" for the precise search path.</para>
+
+ <para>If neither <option>--directory=</option>,
+ <option>--image=</option>, nor <option>--machine=</option>
+ are specified, the current directory will
+ be used. May not be specified together with
+ <option>--image=</option>.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--template=</option></term>
+
+ <listitem><para>Directory or <literal>btrfs</literal>
+ subvolume to use as template for the container's root
+ directory. If this is specified and the container's root
+ directory (as configured by <option>--directory=</option>)
+ does not yet exist it is created as <literal>btrfs</literal>
+ subvolume and populated from this template tree. Ideally, the
+ specified template path refers to the root of a
+ <literal>btrfs</literal> subvolume, in which case a simple
+ copy-on-write snapshot is taken, and populating the root
+ directory is instant. If the specified template path does not
+ refer to the root of a <literal>btrfs</literal> subvolume (or
+ not even to a <literal>btrfs</literal> file system at all),
+ the tree is copied, which can be substantially more
+ time-consuming. Note that if this option is used the
+ container's root directory (in contrast to the template
+ directory!) must be located on a <literal>btrfs</literal> file
+ system, so that the <literal>btrfs</literal> subvolume may be
+ created. May not be specified together with
+ <option>--image=</option> or
+ <option>--ephemeral</option>.</para>
+
+ <para>Note that this switch leaves host name, machine ID and
+ all other settings that could identify the instance
+ unmodified.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>-x</option></term>
+ <term><option>--ephemeral</option></term>
+
+ <listitem><para>If specified, the container is run with a
+ temporary <literal>btrfs</literal> snapshot of its root
+ directory (as configured with <option>--directory=</option>),
+ that is removed immediately when the container terminates.
+ This option is only supported if the root file system is
+ <literal>btrfs</literal>. May not be specified together with
+ <option>--image=</option> or
+ <option>--template=</option>.</para>
+ <para>Note that this switch leaves host name, machine ID and
+ all other settings that could identify the instance
+ unmodified.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>-i</option></term>
+ <term><option>--image=</option></term>
+
+ <listitem><para>Disk image to mount the root directory for the
+ container from. Takes a path to a regular file or to a block
+ device node. The file or block device must contain
+ either:</para>
+
+ <itemizedlist>
+ <listitem><para>An MBR partition table with a single
+ partition of type 0x83 that is marked
+ bootable.</para></listitem>
+
+ <listitem><para>A GUID partition table (GPT) with a single
+ partition of type
+ 0fc63daf-8483-4772-8e79-3d69d8477de4.</para></listitem>
+
+ <listitem><para>A GUID partition table (GPT) with a marked
+ root partition which is mounted as the root directory of the
+ container. Optionally, GPT images may contain a home and/or
+ a server data partition which are mounted to the appropriate
+ places in the container. All these partitions must be
+ identified by the partition types defined by the <ulink
+ url="http://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec/">Discoverable
+ Partitions Specification</ulink>.</para></listitem>
+ </itemizedlist>
+
+ <para>Any other partitions, such as foreign partitions, swap
+ partitions or EFI system partitions are not mounted. May not
+ be specified together with <option>--directory=</option>,
+ <option>--template=</option> or
+ <option>--ephemeral</option>.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>-a</option></term>
+ <term><option>--as-pid2</option></term>
+
+ <listitem><para>Invoke the shell or specified program as process ID (PID) 2 instead of PID 1 (init). By
+ default, if neither this option nor <option>--boot</option> is used, the selected binary is run as process with
+ PID 1, a mode only suitable for programs that are aware of the special semantics that the process with PID 1
+ has on UNIX. For example, it needs to reap all processes reparented to it, and should implement
+ <command>sysvinit</command> compatible signal handling (specifically: it needs to reboot on SIGINT, reexecute
+ on SIGTERM, reload configuration on SIGHUP, and so on). With <option>--as-pid2</option> a minimal stub init
+ process is run as PID 1 and the selected binary is executed as PID 2 (and hence does not need to implement any
+ special semantics). The stub init process will reap processes as necessary and react appropriately to
+ signals. It is recommended to use this mode to invoke arbitrary commands in containers, unless they have been
+ modified to run correctly as PID 1. Or in other words: this switch should be used for pretty much all commands,
+ except when the command refers to an init or shell implementation, as these are generally capable of running
+ correctly as PID 1. This option may not be combined with <option>--boot</option>.</para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>-b</option></term>
+ <term><option>--boot</option></term>
+
+ <listitem><para>Automatically search for an init binary and invoke it as PID 1, instead of a shell or a user
+ supplied program. If this option is used, arguments specified on the command line are used as arguments for the
+ init binary. This option may not be combined with <option>--as-pid2</option>.</para>
+
+ <para>The following table explains the different modes of invocation and relationship to
+ <option>--as-pid2</option> (see above):</para>
+
+ <table>
+ <title>Invocation Mode</title>
+ <tgroup cols='2' align='left' colsep='1' rowsep='1'>
+ <colspec colname="switch" />
+ <colspec colname="explanation" />
+ <thead>
+ <row>
+ <entry>Switch</entry>
+ <entry>Explanation</entry>
+ </row>
+ </thead>
+ <tbody>
+ <row>
+ <entry>Neither <option>--as-pid2</option> nor <option>--boot</option> specified</entry>
+ <entry>The passed parameters are interpreted as the command line, which is executed as PID 1 in the container.</entry>
+ </row>
+
+ <row>
+ <entry><option>--as-pid2</option> specified</entry>
+ <entry>The passed parameters are interpreted as the command line, which is executed as PID 2 in the container. A stub init process is run as PID 1.</entry>
+ </row>
+
+ <row>
+ <entry><option>--boot</option> specified</entry>
+ <entry>An init binary as automatically searched and run as PID 1 in the container. The passed parameters are used as invocation parameters for this process.</entry>
+ </row>
+
+ </tbody>
+ </tgroup>
+ </table>
+
+ <para>Note that <option>--boot</option> is the default mode of operation if the
+ <filename>systemd-nspawn@.service</filename> template unit file is used.</para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--chdir=</option></term>
+
+ <listitem><para>Change to the specified working directory before invoking the process in the container. Expects
+ an absolute path in the container's file system namespace.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>-u</option></term>
+ <term><option>--user=</option></term>
+
+ <listitem><para>After transitioning into the container, change
+ to the specified user-defined in the container's user
+ database. Like all other systemd-nspawn features, this is not
+ a security feature and provides protection against accidental
+ destructive operations only.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>-M</option></term>
+ <term><option>--machine=</option></term>
+
+ <listitem><para>Sets the machine name for this container. This
+ name may be used to identify this container during its runtime
+ (for example in tools like
+ <citerefentry><refentrytitle>machinectl</refentrytitle><manvolnum>1</manvolnum></citerefentry>
+ and similar), and is used to initialize the container's
+ hostname (which the container can choose to override,
+ however). If not specified, the last component of the root
+ directory path of the container is used, possibly suffixed
+ with a random identifier in case <option>--ephemeral</option>
+ mode is selected. If the root directory selected is the host's
+ root directory the host's hostname is used as default
+ instead.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--uuid=</option></term>
+
+ <listitem><para>Set the specified UUID for the container. The
+ init system will initialize
+ <filename>/etc/machine-id</filename> from this if this file is
+ not set yet. Note that this option takes effect only if
+ <filename>/etc/machine-id</filename> in the container is
+ unpopulated.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--slice=</option></term>
+
+ <listitem><para>Make the container part of the specified
+ slice, instead of the default
+ <filename>machine.slice</filename>. This is only applies if
+ the machine is run in its own scope unit, i.e. if
+ <option>--keep-unit</option> is not used.</para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--property=</option></term>
+
+ <listitem><para>Set a unit property on the scope unit to
+ register for the machine. This only applies if the machine is
+ run in its own scope unit, i.e. if
+ <option>--keep-unit</option> is not used. Takes unit property
+ assignments in the same format as <command>systemctl
+ set-property</command>. This is useful to set memory limits
+ and similar for machines.</para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--private-users=</option></term>
+
+ <listitem><para>Controls user namespacing. If enabled, the container will run with its own private set of UNIX
+ user and group ids (UIDs and GIDs). This involves mapping the private UIDs/GIDs used in the container (starting
+ with the container's root user 0 and up) to a range of UIDs/GIDs on the host that are not used for other
+ purposes (usually in the range beyond the host's UID/GID 65536). The parameter may be specified as follows:</para>
+
+ <orderedlist>
+ <listitem><para>If one or two colon-separated numbers are specified, user namespacing is turned on. The first
+ parameter specifies the first host UID/GID to assign to the container, the second parameter specifies the
+ number of host UIDs/GIDs to assign to the container. If the second parameter is omitted, 65536 UIDs/GIDs are
+ assigned.</para></listitem>
+
+ <listitem><para>If the parameter is omitted, or true, user namespacing is turned on. The UID/GID range to
+ use is determined automatically from the file ownership of the root directory of the container's directory
+ tree. To use this option, make sure to prepare the directory tree in advance, and ensure that all files and
+ directories in it are owned by UIDs/GIDs in the range you'd like to use. Also, make sure that used file ACLs
+ exclusively reference UIDs/GIDs in the appropriate range. If this mode is used the number of UIDs/GIDs
+ assigned to the container for use is 65536, and the UID/GID of the root directory must be a multiple of
+ 65536.</para></listitem>
+
+ <listitem><para>If the parameter is false, user namespacing is turned off. This is the default.</para>
+ </listitem>
+
+ <listitem><para>The special value <literal>pick</literal> turns on user namespacing. In this case the UID/GID
+ range is automatically chosen. As first step, the file owner of the root directory of the container's
+ directory tree is read, and it is checked that it is currently not used by the system otherwise (in
+ particular, that no other container is using it). If this check is successful, the UID/GID range determined
+ this way is used, similar to the behavior if "yes" is specified. If the check is not successful (and thus
+ the UID/GID range indicated in the root directory's file owner is already used elsewhere) a new – currently
+ unused – UID/GID range of 65536 UIDs/GIDs is randomly chosen between the host UID/GIDs of 524288 and
+ 1878982656, always starting at a multiple of 65536. This setting implies
+ <option>--private-users-chown</option> (see below), which has the effect that the files and directories in
+ the container's directory tree will be owned by the appropriate users of the range picked. Using this option
+ makes user namespace behavior fully automatic. Note that the first invocation of a previously unused
+ container image might result in picking a new UID/GID range for it, and thus in the (possibly expensive) file
+ ownership adjustment operation. However, subsequent invocations of the container will be cheap (unless of
+ course the picked UID/GID range is assigned to a different use by then).</para></listitem>
+ </orderedlist>
+
+ <para>It is recommended to assign at least 65536 UIDs/GIDs to each container, so that the usable UID/GID range in the
+ container covers 16 bit. For best security, do not assign overlapping UID/GID ranges to multiple containers. It is
+ hence a good idea to use the upper 16 bit of the host 32-bit UIDs/GIDs as container identifier, while the lower 16
+ bit encode the container UID/GID used. This is in fact the behavior enforced by the
+ <option>--private-users=pick</option> option.</para>
+
+ <para>When user namespaces are used, the GID range assigned to each container is always chosen identical to the
+ UID range.</para>
+
+ <para>In most cases, using <option>--private-users=pick</option> is the recommended option as it enhances
+ container security massively and operates fully automatically in most cases.</para>
+
+ <para>Note that the picked UID/GID range is not written to <filename>/etc/passwd</filename> or
+ <filename>/etc/group</filename>. In fact, the allocation of the range is not stored persistently anywhere,
+ except in the file ownership of the files and directories of the container.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--private-users-chown</option></term>
+
+ <listitem><para>If specified, all files and directories in the container's directory tree will adjusted so that
+ they are owned to the appropriate UIDs/GIDs selected for the container (see above). This operation is
+ potentially expensive, as it involves descending and iterating through the full directory tree of the
+ container. Besides actual file ownership, file ACLs are adjusted as well.</para>
+
+ <para>This option is implied if <option>--private-users=pick</option> is used. This option has no effect if
+ user namespacing is not used.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>-U</option></term>
+
+ <listitem><para>If the kernel supports the user namespaces feature, equivalent to
+ <option>--private-users=pick --private-users-chown</option>, otherwise equivalent to
+ <option>--private-users=no</option>.</para>
+
+ <para>Note that <option>-U</option> is the default if the
+ <filename>systemd-nspawn@.service</filename> template unit file is used.</para>
+
+ <para>Note: it is possible to undo the effect of <option>--private-users-chown</option> (or
+ <option>-U</option>) on the file system by redoing the operation with the first UID of 0:</para>
+
+ <programlisting>systemd-nspawn … --private-users=0 --private-users-chown</programlisting>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--private-network</option></term>
+
+ <listitem><para>Disconnect networking of the container from
+ the host. This makes all network interfaces unavailable in the
+ container, with the exception of the loopback device and those
+ specified with <option>--network-interface=</option> and
+ configured with <option>--network-veth</option>. If this
+ option is specified, the CAP_NET_ADMIN capability will be
+ added to the set of capabilities the container retains. The
+ latter may be disabled by using
+ <option>--drop-capability=</option>.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--network-interface=</option></term>
+
+ <listitem><para>Assign the specified network interface to the
+ container. This will remove the specified interface from the
+ calling namespace and place it in the container. When the
+ container terminates, it is moved back to the host namespace.
+ Note that <option>--network-interface=</option> implies
+ <option>--private-network</option>. This option may be used
+ more than once to add multiple network interfaces to the
+ container.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--network-macvlan=</option></term>
+
+ <listitem><para>Create a <literal>macvlan</literal> interface
+ of the specified Ethernet network interface and add it to the
+ container. A <literal>macvlan</literal> interface is a virtual
+ interface that adds a second MAC address to an existing
+ physical Ethernet link. The interface in the container will be
+ named after the interface on the host, prefixed with
+ <literal>mv-</literal>. Note that
+ <option>--network-macvlan=</option> implies
+ <option>--private-network</option>. This option may be used
+ more than once to add multiple network interfaces to the
+ container.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--network-ipvlan=</option></term>
+
+ <listitem><para>Create an <literal>ipvlan</literal> interface
+ of the specified Ethernet network interface and add it to the
+ container. An <literal>ipvlan</literal> interface is a virtual
+ interface, similar to a <literal>macvlan</literal> interface,
+ which uses the same MAC address as the underlying interface.
+ The interface in the container will be named after the
+ interface on the host, prefixed with <literal>iv-</literal>.
+ Note that <option>--network-ipvlan=</option> implies
+ <option>--private-network</option>. This option may be used
+ more than once to add multiple network interfaces to the
+ container.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>-n</option></term>
+ <term><option>--network-veth</option></term>
+
+ <listitem><para>Create a virtual Ethernet link (<literal>veth</literal>) between host and container. The host
+ side of the Ethernet link will be available as a network interface named after the container's name (as
+ specified with <option>--machine=</option>), prefixed with <literal>ve-</literal>. The container side of the
+ Ethernet link will be named <literal>host0</literal>. The <option>--network-veth</option> option implies
+ <option>--private-network</option>.</para>
+
+ <para>Note that
+ <citerefentry><refentrytitle>systemd-networkd.service</refentrytitle><manvolnum>8</manvolnum></citerefentry>
+ includes by default a network file <filename>/usr/lib/systemd/network/80-container-ve.network</filename>
+ matching the host-side interfaces created this way, which contains settings to enable automatic address
+ provisioning on the created virtual link via DHCP, as well as automatic IP routing onto the host's external
+ network interfaces. It also contains <filename>/usr/lib/systemd/network/80-container-host0.network</filename>
+ matching the container-side interface created this way, containing settings to enable client side address
+ assignment via DHCP. In case <filename>systemd-networkd</filename> is running on both the host and inside the
+ container, automatic IP communication from the container to the host is thus available, with further
+ connectivity to the external network.</para>
+
+ <para>Note that <option>--network-veth</option> is the default if the
+ <filename>systemd-nspawn@.service</filename> template unit file is used.</para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--network-veth-extra=</option></term>
+
+ <listitem><para>Adds an additional virtual Ethernet link
+ between host and container. Takes a colon-separated pair of
+ host interface name and container interface name. The latter
+ may be omitted in which case the container and host sides will
+ be assigned the same name. This switch is independent of
+ <option>--network-veth</option>, and — in contrast — may be
+ used multiple times, and allows configuration of the network
+ interface names. Note that <option>--network-bridge=</option>
+ has no effect on interfaces created with
+ <option>--network-veth-extra=</option>.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--network-bridge=</option></term>
+
+ <listitem><para>Adds the host side of the Ethernet link created with <option>--network-veth</option> to the
+ specified Ethernet bridge interface. Expects a valid network interface name of a bridge device as
+ argument. Note that <option>--network-bridge=</option> implies <option>--network-veth</option>. If this option
+ is used, the host side of the Ethernet link will use the <literal>vb-</literal> prefix instead of
+ <literal>ve-</literal>.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--network-zone=</option></term>
+
+ <listitem><para>Creates a virtual Ethernet link (<literal>veth</literal>) to the container and adds it to an
+ automatically managed Ethernet bridge interface. The bridge interface is named after the passed argument,
+ prefixed with <literal>vz-</literal>. The bridge interface is automatically created when the first container
+ configured for its name is started, and is automatically removed when the last container configured for its
+ name exits. Hence, each bridge interface configured this way exists only as long as there's at least one
+ container referencing it running. This option is very similar to <option>--network-bridge=</option>, besides
+ this automatic creation/removal of the bridge device.</para>
+
+ <para>This setting makes it easy to place multiple related containers on a common, virtual Ethernet-based
+ broadcast domain, here called a "zone". Each container may only be part of one zone, but each zone may contain
+ any number of containers. Each zone is referenced by its name. Names may be chosen freely (as long as they form
+ valid network interface names when prefixed with <literal>vz-</literal>), and it is sufficient to pass the same
+ name to the <option>--network-zones=</option> switch of the various concurrently running containers to join
+ them in one zone.</para>
+
+ <para>Note that
+ <citerefentry><refentrytitle>systemd-networkd.service</refentrytitle><manvolnum>8</manvolnum></citerefentry>
+ includes by default a network file <filename>/usr/lib/systemd/network/80-container-vz.network</filename>
+ matching the bridge interfaces created this way, which contains settings to enable automatic address
+ provisioning on the created virtual network via DHCP, as well as automatic IP routing onto the host's external
+ network interfaces. Using <option>--network-zone=</option> is hence in most cases fully automatic and
+ sufficient to connect multiple local containers in a joined broadcast domain to the host, with further
+ connectivity to the external network.</para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>-p</option></term>
+ <term><option>--port=</option></term>
+
+ <listitem><para>If private networking is enabled, maps an IP
+ port on the host onto an IP port on the container. Takes a
+ protocol specifier (either <literal>tcp</literal> or
+ <literal>udp</literal>), separated by a colon from a host port
+ number in the range 1 to 65535, separated by a colon from a
+ container port number in the range from 1 to 65535. The
+ protocol specifier and its separating colon may be omitted, in
+ which case <literal>tcp</literal> is assumed. The container
+ port number and its colon may be omitted, in which case the
+ same port as the host port is implied. This option is only
+ supported if private networking is used, such as with
+ <option>--network-veth</option>, <option>--network-zone=</option>
+ <option>--network-bridge=</option>.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>-Z</option></term>
+ <term><option>--selinux-context=</option></term>
+
+ <listitem><para>Sets the SELinux security context to be used
+ to label processes in the container.</para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>-L</option></term>
+ <term><option>--selinux-apifs-context=</option></term>
+
+ <listitem><para>Sets the SELinux security context to be used
+ to label files in the virtual API file systems in the
+ container.</para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--capability=</option></term>
+
+ <listitem><para>List one or more additional capabilities to
+ grant the container. Takes a comma-separated list of
+ capability names, see
+ <citerefentry project='man-pages'><refentrytitle>capabilities</refentrytitle><manvolnum>7</manvolnum></citerefentry>
+ for more information. Note that the following capabilities
+ will be granted in any way: CAP_CHOWN, CAP_DAC_OVERRIDE,
+ CAP_DAC_READ_SEARCH, CAP_FOWNER, CAP_FSETID, CAP_IPC_OWNER,
+ CAP_KILL, CAP_LEASE, CAP_LINUX_IMMUTABLE,
+ CAP_NET_BIND_SERVICE, CAP_NET_BROADCAST, CAP_NET_RAW,
+ CAP_SETGID, CAP_SETFCAP, CAP_SETPCAP, CAP_SETUID,
+ CAP_SYS_ADMIN, CAP_SYS_CHROOT, CAP_SYS_NICE, CAP_SYS_PTRACE,
+ CAP_SYS_TTY_CONFIG, CAP_SYS_RESOURCE, CAP_SYS_BOOT,
+ CAP_AUDIT_WRITE, CAP_AUDIT_CONTROL. Also CAP_NET_ADMIN is
+ retained if <option>--private-network</option> is specified.
+ If the special value <literal>all</literal> is passed, all
+ capabilities are retained.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--drop-capability=</option></term>
+
+ <listitem><para>Specify one or more additional capabilities to
+ drop for the container. This allows running the container with
+ fewer capabilities than the default (see
+ above).</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--kill-signal=</option></term>
+
+ <listitem><para>Specify the process signal to send to the
+ container's PID 1 when nspawn itself receives SIGTERM, in
+ order to trigger an orderly shutdown of the
+ container. Defaults to SIGRTMIN+3 if <option>--boot</option>
+ is used (on systemd-compatible init systems SIGRTMIN+3
+ triggers an orderly shutdown). For a list of valid signals, see
+ <citerefentry project='man-pages'><refentrytitle>signal</refentrytitle><manvolnum>7</manvolnum></citerefentry>.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--link-journal=</option></term>
+
+ <listitem><para>Control whether the container's journal shall
+ be made visible to the host system. If enabled, allows viewing
+ the container's journal files from the host (but not vice
+ versa). Takes one of <literal>no</literal>,
+ <literal>host</literal>, <literal>try-host</literal>,
+ <literal>guest</literal>, <literal>try-guest</literal>,
+ <literal>auto</literal>. If <literal>no</literal>, the journal
+ is not linked. If <literal>host</literal>, the journal files
+ are stored on the host file system (beneath
+ <filename>/var/log/journal/<replaceable>machine-id</replaceable></filename>)
+ and the subdirectory is bind-mounted into the container at the
+ same location. If <literal>guest</literal>, the journal files
+ are stored on the guest file system (beneath
+ <filename>/var/log/journal/<replaceable>machine-id</replaceable></filename>)
+ and the subdirectory is symlinked into the host at the same
+ location. <literal>try-host</literal> and
+ <literal>try-guest</literal> do the same but do not fail if
+ the host does not have persistent journaling enabled. If
+ <literal>auto</literal> (the default), and the right
+ subdirectory of <filename>/var/log/journal</filename> exists,
+ it will be bind mounted into the container. If the
+ subdirectory does not exist, no linking is performed.
+ Effectively, booting a container once with
+ <literal>guest</literal> or <literal>host</literal> will link
+ the journal persistently if further on the default of
+ <literal>auto</literal> is used.</para>
+
+ <para>Note that <option>--link-journal=try-guest</option> is the default if the
+ <filename>systemd-nspawn@.service</filename> template unit file is used.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>-j</option></term>
+
+ <listitem><para>Equivalent to
+ <option>--link-journal=try-guest</option>.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--read-only</option></term>
+
+ <listitem><para>Mount the root file system read-only for the
+ container.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--bind=</option></term>
+ <term><option>--bind-ro=</option></term>
+
+ <listitem><para>Bind mount a file or directory from the host
+ into the container. Takes one of: a path argument — in which
+ case the specified path will be mounted from the host to the
+ same path in the container —, or a colon-separated pair of
+ paths — in which case the first specified path is the source
+ in the host, and the second path is the destination in the
+ container —, or a colon-separated triple of source path,
+ destination path and mount options. Mount options are
+ comma-separated and currently, only "rbind" and "norbind"
+ are allowed. Defaults to "rbind". Backslash escapes are interpreted, so
+ <literal>\:</literal> may be used to embed colons in either path.
+ This option may be specified multiple times for
+ creating multiple independent bind mount points. The
+ <option>--bind-ro=</option> option creates read-only bind
+ mounts.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--tmpfs=</option></term>
+
+ <listitem><para>Mount a tmpfs file system into the container.
+ Takes a single absolute path argument that specifies where to
+ mount the tmpfs instance to (in which case the directory
+ access mode will be chosen as 0755, owned by root/root), or
+ optionally a colon-separated pair of path and mount option
+ string that is used for mounting (in which case the kernel
+ default for access mode and owner will be chosen, unless
+ otherwise specified). This option is particularly useful for
+ mounting directories such as <filename>/var</filename> as
+ tmpfs, to allow state-less systems, in particular when
+ combined with <option>--read-only</option>.
+ Backslash escapes are interpreted in the path, so
+ <literal>\:</literal> may be used to embed colons in the path.
+ </para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--overlay=</option></term>
+ <term><option>--overlay-ro=</option></term>
+
+ <listitem><para>Combine multiple directory trees into one
+ overlay file system and mount it into the container. Takes a
+ list of colon-separated paths to the directory trees to
+ combine and the destination mount point.</para>
+
+ <para>Backslash escapes are interpreted in the paths, so
+ <literal>\:</literal> may be used to embed colons in the paths.
+ </para>
+
+ <para>If three or more paths are specified, then the last
+ specified path is the destination mount point in the
+ container, all paths specified before refer to directory trees
+ on the host and are combined in the specified order into one
+ overlay file system. The left-most path is hence the lowest
+ directory tree, the second-to-last path the highest directory
+ tree in the stacking order. If <option>--overlay-ro=</option>
+ is used instead of <option>--overlay=</option>, a read-only
+ overlay file system is created. If a writable overlay file
+ system is created, all changes made to it are written to the
+ highest directory tree in the stacking order, i.e. the
+ second-to-last specified.</para>
+
+ <para>If only two paths are specified, then the second
+ specified path is used both as the top-level directory tree in
+ the stacking order as seen from the host, as well as the mount
+ point for the overlay file system in the container. At least
+ two paths have to be specified.</para>
+
+ <para>For details about overlay file systems, see <ulink
+ url="https://www.kernel.org/doc/Documentation/filesystems/overlayfs.txt">overlayfs.txt</ulink>. Note
+ that the semantics of overlay file systems are substantially
+ different from normal file systems, in particular regarding
+ reported device and inode information. Device and inode
+ information may change for a file while it is being written
+ to, and processes might see out-of-date versions of files at
+ times. Note that this switch automatically derives the
+ <literal>workdir=</literal> mount option for the overlay file
+ system from the top-level directory tree, making it a sibling
+ of it. It is hence essential that the top-level directory tree
+ is not a mount point itself (since the working directory must
+ be on the same file system as the top-most directory
+ tree). Also note that the <literal>lowerdir=</literal> mount
+ option receives the paths to stack in the opposite order of
+ this switch.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>-E <replaceable>NAME</replaceable>=<replaceable>VALUE</replaceable></option></term>
+ <term><option>--setenv=<replaceable>NAME</replaceable>=<replaceable>VALUE</replaceable></option></term>
+
+ <listitem><para>Specifies an environment variable assignment
+ to pass to the init process in the container, in the format
+ <literal>NAME=VALUE</literal>. This may be used to override
+ the default variables or to set additional variables. This
+ parameter may be used more than once.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--register=</option></term>
+
+ <listitem><para>Controls whether the container is registered
+ with
+ <citerefentry><refentrytitle>systemd-machined</refentrytitle><manvolnum>8</manvolnum></citerefentry>.
+ Takes a boolean argument, which defaults to <literal>yes</literal>.
+ This option should be enabled when the container runs a full
+ Operating System (more specifically: an init system), and is
+ useful to ensure that the container is accessible via
+ <citerefentry><refentrytitle>machinectl</refentrytitle><manvolnum>1</manvolnum></citerefentry>
+ and shown by tools such as
+ <citerefentry project='man-pages'><refentrytitle>ps</refentrytitle><manvolnum>1</manvolnum></citerefentry>.
+ If the container does not run an init system, it is
+ recommended to set this option to <literal>no</literal>.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--keep-unit</option></term>
+
+ <listitem><para>Instead of creating a transient scope unit to
+ run the container in, simply register the service or scope
+ unit <command>systemd-nspawn</command> has been invoked in
+ with
+ <citerefentry><refentrytitle>systemd-machined</refentrytitle><manvolnum>8</manvolnum></citerefentry>.
+ This has no effect if <option>--register=no</option> is used.
+ This switch should be used if
+ <command>systemd-nspawn</command> is invoked from within a
+ service unit, and the service unit's sole purpose is to run a
+ single <command>systemd-nspawn</command> container. This
+ option is not available if run from a user
+ session.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--personality=</option></term>
+
+ <listitem><para>Control the architecture ("personality")
+ reported by
+ <citerefentry project='man-pages'><refentrytitle>uname</refentrytitle><manvolnum>2</manvolnum></citerefentry>
+ in the container. Currently, only <literal>x86</literal> and
+ <literal>x86-64</literal> are supported. This is useful when
+ running a 32-bit container on a 64-bit host. If this setting
+ is not used, the personality reported in the container is the
+ same as the one reported on the host.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>-q</option></term>
+ <term><option>--quiet</option></term>
+
+ <listitem><para>Turns off any status output by the tool
+ itself. When this switch is used, the only output from nspawn
+ will be the console output of the container OS
+ itself.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--volatile</option></term>
+ <term><option>--volatile=</option><replaceable>MODE</replaceable></term>
+
+ <listitem><para>Boots the container in volatile mode. When no
+ mode parameter is passed or when mode is specified as
+ <option>yes</option>, full volatile mode is enabled. This
+ means the root directory is mounted as a mostly unpopulated
+ <literal>tmpfs</literal> instance, and
+ <filename>/usr</filename> from the OS tree is mounted into it
+ in read-only mode (the system thus starts up with read-only OS
+ image, but pristine state and configuration, any changes
+ are lost on shutdown). When the mode parameter
+ is specified as <option>state</option>, the OS tree is
+ mounted read-only, but <filename>/var</filename> is mounted as
+ a <literal>tmpfs</literal> instance into it (the system thus
+ starts up with read-only OS resources and configuration, but
+ pristine state, and any changes to the latter are lost on
+ shutdown). When the mode parameter is specified as
+ <option>no</option> (the default), the whole OS tree is made
+ available writable.</para>
+
+ <para>Note that setting this to <option>yes</option> or
+ <option>state</option> will only work correctly with
+ operating systems in the container that can boot up with only
+ <filename>/usr</filename> mounted, and are able to populate
+ <filename>/var</filename> automatically, as
+ needed.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--settings=</option><replaceable>MODE</replaceable></term>
+
+ <listitem><para>Controls whether
+ <command>systemd-nspawn</command> shall search for and use
+ additional per-container settings from
+ <filename>.nspawn</filename> files. Takes a boolean or the
+ special values <option>override</option> or
+ <option>trusted</option>.</para>
+
+ <para>If enabled (the default), a settings file named after the
+ machine (as specified with the <option>--machine=</option>
+ setting, or derived from the directory or image file name)
+ with the suffix <filename>.nspawn</filename> is searched in
+ <filename>/etc/systemd/nspawn/</filename> and
+ <filename>/run/systemd/nspawn/</filename>. If it is found
+ there, its settings are read and used. If it is not found
+ there, it is subsequently searched in the same directory as the
+ image file or in the immediate parent of the root directory of
+ the container. In this case, if the file is found, its settings
+ will be also read and used, but potentially unsafe settings
+ are ignored. Note that in both these cases, settings on the
+ command line take precedence over the corresponding settings
+ from loaded <filename>.nspawn</filename> files, if both are
+ specified. Unsafe settings are considered all settings that
+ elevate the container's privileges or grant access to
+ additional resources such as files or directories of the
+ host. For details about the format and contents of
+ <filename>.nspawn</filename> files, consult
+ <citerefentry><refentrytitle>systemd.nspawn</refentrytitle><manvolnum>5</manvolnum></citerefentry>.</para>
+
+ <para>If this option is set to <option>override</option>, the
+ file is searched, read and used the same way, however, the order of
+ precedence is reversed: settings read from the
+ <filename>.nspawn</filename> file will take precedence over
+ the corresponding command line options, if both are
+ specified.</para>
+
+ <para>If this option is set to <option>trusted</option>, the
+ file is searched, read and used the same way, but regardless
+ of being found in <filename>/etc/systemd/nspawn/</filename>,
+ <filename>/run/systemd/nspawn/</filename> or next to the image
+ file or container root directory, all settings will take
+ effect, however, command line arguments still take precedence
+ over corresponding settings.</para>
+
+ <para>If disabled, no <filename>.nspawn</filename> file is read
+ and no settings except the ones on the command line are in
+ effect.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><option>--notify-ready=</option></term>
+
+ <listitem><para>Configures support for notifications from the container's init process.
+ <option>--notify-ready=</option> takes a boolean (<option>no</option> and <option>yes</option>).
+ With option <option>no</option> systemd-nspawn notifies systemd
+ with a <literal>READY=1</literal> message when the init process is created.
+ With option <option>yes</option> systemd-nspawn waits for the
+ <literal>READY=1</literal> message from the init process in the container
+ before sending its own to systemd. For more details about notifications
+ see <citerefentry><refentrytitle>sd_notify</refentrytitle><manvolnum>3</manvolnum></citerefentry>).</para></listitem>
+ </varlistentry>
+
+ <xi:include href="standard-options.xml" xpointer="help" />
+ <xi:include href="standard-options.xml" xpointer="version" />
+ </variablelist>
+
+ </refsect1>
+
+ <refsect1>
+ <title>Examples</title>
+
+ <example>
+ <title>Build and boot a minimal BLAG distribution in a container</title>
+
+ <programlisting># dnf -y --releasever=210k --installroot=/srv/mycontainer --disablerepo='*' --enablerepo=blag --enablerepo=updates install systemd passwd dnf blag-release vim-minimal
+# systemd-nspawn -bD /srv/mycontainer</programlisting>
+
+ <para>This installs a minimal BLAG distribution into the
+ directory <filename noindex='true'>/srv/mycontainer/</filename>
+ and then boots an OS in a namespace container in it.</para>
+ </example>
+
+ <example>
+ <title>Spawn a shell in a container of a minimal gNewSense Ucclia distribution</title>
+
+ <programlisting># debootstrap --arch=amd64 ucclia ~/gnewsense-tree/
+# systemd-nspawn -D ~/gnewsense-tree/</programlisting>
+
+ <para>This installs a minimal gNewSense unstable distribution into
+ the directory <filename>~/gnewsense-tree/</filename> and then
+ spawns a shell in a namespace container in it.</para>
+ </example>
+
+ <example>
+ <title>Boot a minimal Parabola distribution in a container</title>
+
+ <programlisting># pacstrap -c -d ~/parabola-tree/ base
+# systemd-nspawn -bD ~/parabola-tree/</programlisting>
+
+ <para>This installs a minimal Parabola distribution into the
+ directory <filename>~/parabola-tree/</filename> and then boots an OS
+ in a namespace container in it.</para>
+ </example>
+
+ <example>
+ <title>Boot into an ephemeral <literal>btrfs</literal> snapshot of the host system</title>
+
+ <programlisting># systemd-nspawn -D / -xb</programlisting>
+
+ <para>This runs a copy of the host system in a
+ <literal>btrfs</literal> snapshot which is removed immediately
+ when the container exits. All file system changes made during
+ runtime will be lost on shutdown, hence.</para>
+ </example>
+
+ <example>
+ <title>Run a container with SELinux sandbox security contexts</title>
+
+ <programlisting># chcon system_u:object_r:svirt_sandbox_file_t:s0:c0,c1 -R /srv/container
+# systemd-nspawn -L system_u:object_r:svirt_sandbox_file_t:s0:c0,c1 -Z system_u:system_r:svirt_lxc_net_t:s0:c0,c1 -D /srv/container /bin/sh</programlisting>
+ </example>
+ </refsect1>
+
+ <refsect1>
+ <title>Exit status</title>
+
+ <para>The exit code of the program executed in the container is
+ returned.</para>
+ </refsect1>
+
+ <refsect1>
+ <title>See Also</title>
+ <para>
+ <citerefentry><refentrytitle>systemd</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
+ <citerefentry><refentrytitle>systemd.nspawn</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
+ <citerefentry project='man-pages'><refentrytitle>chroot</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
+ <citerefentry project='mankier'><refentrytitle>dnf</refentrytitle><manvolnum>8</manvolnum></citerefentry>,
+ <citerefentry project='die-net'><refentrytitle>debootstrap</refentrytitle><manvolnum>8</manvolnum></citerefentry>,
+ <citerefentry project='archlinux'><refentrytitle>pacman</refentrytitle><manvolnum>8</manvolnum></citerefentry>,
+ <citerefentry><refentrytitle>systemd.slice</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
+ <citerefentry><refentrytitle>machinectl</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
+ <citerefentry project='man-pages'><refentrytitle>btrfs</refentrytitle><manvolnum>8</manvolnum></citerefentry>
+ </para>
+ </refsect1>
+
+</refentry>
diff --git a/src/systemd-nspawn/systemd-nspawn@.service.in b/src/systemd-nspawn/systemd-nspawn@.service.in
new file mode 100644
index 0000000000..c8141639b6
--- /dev/null
+++ b/src/systemd-nspawn/systemd-nspawn@.service.in
@@ -0,0 +1,39 @@
+# This file is part of systemd.
+#
+# systemd is free software; you can redistribute it and/or modify it
+# under the terms of the GNU Lesser General Public License as published by
+# the Free Software Foundation; either version 2.1 of the License, or
+# (at your option) any later version.
+
+[Unit]
+Description=Container %i
+Documentation=man:systemd-nspawn(1)
+PartOf=machines.target
+Before=machines.target
+After=network.target
+
+[Service]
+ExecStart=@bindir@/systemd-nspawn --quiet --keep-unit --boot --link-journal=try-guest --network-veth -U --settings=override --machine=%i
+KillMode=mixed
+Type=notify
+RestartForceExitStatus=133
+SuccessExitStatus=133
+Slice=machine.slice
+Delegate=yes
+TasksMax=16384
+
+# Enforce a strict device policy, similar to the one nspawn configures
+# when it allocates its own scope unit. Make sure to keep these
+# policies in sync if you change them!
+DevicePolicy=closed
+DeviceAllow=/dev/net/tun rwm
+DeviceAllow=char-pts rw
+
+# nspawn itself needs access to /dev/loop-control and /dev/loop, to
+# implement the --image= option. Add these here, too.
+DeviceAllow=/dev/loop-control rw
+DeviceAllow=block-loop rw
+DeviceAllow=block-blkext rw
+
+[Install]
+WantedBy=machines.target
diff --git a/src/systemd-nspawn/test-patch-uid.c b/src/systemd-nspawn/test-patch-uid.c
new file mode 100644
index 0000000000..4297fea6fe
--- /dev/null
+++ b/src/systemd-nspawn/test-patch-uid.c
@@ -0,0 +1,62 @@
+/***
+ This file is part of systemd.
+
+ Copyright 2016 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
+ (at your option) any later version.
+
+ systemd is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include <stdlib.h>
+
+#include "systemd-basic/log.h"
+#include "systemd-basic/user-util.h"
+#include "systemd-basic/util.h"
+
+#include "nspawn-patch-uid.h"
+
+int main(int argc, char *argv[]) {
+ uid_t shift, range;
+ int r;
+
+ log_set_max_level(LOG_DEBUG);
+ log_parse_environment();
+ log_open();
+
+ if (argc != 4) {
+ log_error("Expected PATH SHIFT RANGE parameters.");
+ return EXIT_FAILURE;
+ }
+
+ r = parse_uid(argv[2], &shift);
+ if (r < 0) {
+ log_error_errno(r, "Failed to parse UID shift %s.", argv[2]);
+ return EXIT_FAILURE;
+ }
+
+ r = parse_gid(argv[3], &range);
+ if (r < 0) {
+ log_error_errno(r, "Failed to parse UID range %s.", argv[3]);
+ return EXIT_FAILURE;
+ }
+
+ r = path_patch_uid(argv[1], shift, range);
+ if (r < 0) {
+ log_error_errno(r, "Failed to patch directory tree: %m");
+ return EXIT_FAILURE;
+ }
+
+ log_info("Changed: %s", yes_no(r));
+
+ return EXIT_SUCCESS;
+}