#include "sd-bus.h"
#include "sd-daemon.h"
+#include "sd-messages.h"
#include "alloc-util.h"
#include "architecture.h"
static unsigned arg_default_start_limit_burst = DEFAULT_START_LIMIT_BURST;
static usec_t arg_runtime_watchdog = 0;
static usec_t arg_shutdown_watchdog = 10 * USEC_PER_MINUTE;
+static char *arg_watchdog_device = NULL;
static char **arg_default_environment = NULL;
static struct rlimit *arg_default_rlimit[_RLIMIT_MAX] = {};
static uint64_t arg_capability_bounding_set = CAP_ALL;
if (arg_default_timeout_start_usec <= 0)
arg_default_timeout_start_usec = USEC_INFINITY;
+ } else if (proc_cmdline_key_streq(key, "systemd.watchdog_device")) {
+
+ if (proc_cmdline_value_missing(key, value))
+ return 0;
+
+ parse_path_argument_and_warn(value, false, &arg_watchdog_device);
+
} else if (streq(key, "quiet") && !value) {
if (arg_show_status == _SHOW_STATUS_UNSET)
{ "Manager", "JoinControllers", config_parse_join_controllers, 0, &arg_join_controllers },
{ "Manager", "RuntimeWatchdogSec", config_parse_sec, 0, &arg_runtime_watchdog },
{ "Manager", "ShutdownWatchdogSec", config_parse_sec, 0, &arg_shutdown_watchdog },
+ { "Manager", "WatchdogDevice", config_parse_path, 0, &arg_watchdog_device },
{ "Manager", "CapabilityBoundingSet", config_parse_capability_set, 0, &arg_capability_bounding_set },
#if HAVE_SECCOMP
{ "Manager", "SystemCallArchitectures", config_parse_syscall_archs, 0, &arg_syscall_archs },
/* Get current RLIMIT_NOFILE maximum compiled into the kernel. */
r = read_one_line_file("/proc/sys/fs/nr_open", &nr_open);
- if (r == 0)
+ if (r >= 0)
r = safe_atoi(nr_open, &min_max);
/* If we fail, fallback to the hard-coded kernel limit of 1024 * 1024. */
if (r < 0)
static int fixup_environment(void) {
_cleanup_free_ char *term = NULL;
+ const char *t;
int r;
/* We expect the environment to be set correctly
r = proc_cmdline_get_key("TERM", 0, &term);
if (r < 0)
return r;
- if (r == 0) {
- term = strdup(default_term_for_tty("/dev/console"));
- if (!term)
- return -ENOMEM;
- }
- if (setenv("TERM", term, 1) < 0)
+ t = term ?: default_term_for_tty("/dev/console");
+
+ if (setenv("TERM", t, 1) < 0)
return -errno;
return 0;
static int become_shutdown(
const char *shutdown_verb,
- int retval,
- bool arm_reboot_watchdog) {
+ int retval) {
char log_level[DECIMAL_STR_MAX(int) + 1],
exit_code[DECIMAL_STR_MAX(uint8_t) + 1];
size_t pos = 5;
int r;
- assert(command_line[pos] == NULL);
+ assert(shutdown_verb);
+ assert(!command_line[pos]);
env_block = strv_copy(environ);
xsprintf(log_level, "%d", log_get_max_level());
assert(pos < ELEMENTSOF(command_line));
- if (arm_reboot_watchdog && arg_shutdown_watchdog > 0 && arg_shutdown_watchdog != USEC_INFINITY) {
+ if (streq(shutdown_verb, "reboot") &&
+ arg_shutdown_watchdog > 0 &&
+ arg_shutdown_watchdog != USEC_INFINITY) {
+
char *e;
/* If we reboot let's set the shutdown
/* Tell the binary how often to ping, ignore failure */
if (asprintf(&e, "WATCHDOG_USEC="USEC_FMT, arg_shutdown_watchdog) > 0)
- (void) strv_push(&env_block, e);
+ (void) strv_consume(&env_block, e);
+
+ if (arg_watchdog_device &&
+ asprintf(&e, "WATCHDOG_DEVICE=%s", arg_watchdog_device) > 0)
+ (void) strv_consume(&env_block, e);
} else
watchdog_close(true);
(void) write_string_file("/proc/sys/kernel/core_pattern", "|/bin/false", 0);
}
-int main(int argc, char *argv[]) {
- Manager *m = NULL;
- int r, retval = EXIT_FAILURE;
- usec_t before_startup, after_startup;
- char timespan[FORMAT_TIMESPAN_MAX];
- FDSet *fds = NULL;
- bool reexecute = false;
- const char *shutdown_verb = NULL;
- dual_timestamp initrd_timestamp = DUAL_TIMESTAMP_NULL;
- dual_timestamp userspace_timestamp = DUAL_TIMESTAMP_NULL;
- dual_timestamp kernel_timestamp = DUAL_TIMESTAMP_NULL;
- dual_timestamp security_start_timestamp = DUAL_TIMESTAMP_NULL;
- dual_timestamp security_finish_timestamp = DUAL_TIMESTAMP_NULL;
- static char systemd[] = "systemd";
- bool skip_setup = false;
- unsigned j;
- bool loaded_policy = false;
- bool arm_reboot_watchdog = false;
- bool queue_default_job = false;
- bool first_boot = false;
- char *switch_root_dir = NULL, *switch_root_init = NULL;
- struct rlimit saved_rlimit_nofile = RLIMIT_MAKE_CONST(0), saved_rlimit_memlock = RLIMIT_MAKE_CONST((rlim_t) -1);
- const char *error_message = NULL;
+static void do_reexecute(
+ int argc,
+ char *argv[],
+ const struct rlimit *saved_rlimit_nofile,
+ const struct rlimit *saved_rlimit_memlock,
+ FDSet *fds,
+ const char *switch_root_dir,
+ const char *switch_root_init,
+ const char **ret_error_message) {
+
+ unsigned i, j, args_size;
+ const char **args;
+ int r;
- redirect_telinit(argc, argv);
+ assert(saved_rlimit_nofile);
+ assert(saved_rlimit_memlock);
+ assert(ret_error_message);
- dual_timestamp_from_monotonic(&kernel_timestamp, 0);
- dual_timestamp_get(&userspace_timestamp);
+ /* Close and disarm the watchdog, so that the new instance can reinitialize it, but doesn't get rebooted while
+ * we do that */
+ watchdog_close(true);
- /* Determine if this is a reexecution or normal bootup. We do
- * the full command line parsing much later, so let's just
- * have a quick peek here. */
- if (strv_find(argv+1, "--deserialize"))
- skip_setup = true;
+ /* Reset the RLIMIT_NOFILE to the kernel default, so that the new systemd can pass the kernel default to its
+ * child processes */
- /* If we have switched root, do all the special setup
- * things */
- if (strv_find(argv+1, "--switched-root"))
- skip_setup = false;
+ if (saved_rlimit_nofile->rlim_cur > 0)
+ (void) setrlimit(RLIMIT_NOFILE, saved_rlimit_nofile);
+ if (saved_rlimit_memlock->rlim_cur != (rlim_t) -1)
+ (void) setrlimit(RLIMIT_MEMLOCK, saved_rlimit_memlock);
- /* If we get started via the /sbin/init symlink then we are
- called 'init'. After a subsequent reexecution we are then
- called 'systemd'. That is confusing, hence let's call us
- systemd right-away. */
- program_invocation_short_name = systemd;
- (void) prctl(PR_SET_NAME, systemd);
+ if (switch_root_dir) {
+ /* Kill all remaining processes from the initrd, but don't wait for them, so that we can handle the
+ * SIGCHLD for them after deserializing. */
+ broadcast_signal(SIGTERM, false, true);
- saved_argv = argv;
- saved_argc = argc;
+ /* And switch root with MS_MOVE, because we remove the old directory afterwards and detach it. */
+ r = switch_root(switch_root_dir, "/mnt", true, MS_MOVE);
+ if (r < 0)
+ log_error_errno(r, "Failed to switch root, trying to continue: %m");
+ }
- log_set_upgrade_syslog_to_journal(true);
+ args_size = MAX(6, argc+1);
+ args = newa(const char*, args_size);
- if (getpid_cached() == 1) {
- /* Disable the umask logic */
- umask(0);
+ if (!switch_root_init) {
+ char sfd[DECIMAL_STR_MAX(int) + 1];
- /* Always reopen /dev/console when running as PID 1 or one of its pre-execve() children. This is
- * important so that we never end up logging to any foreign stderr, for example if we have to log in a
- * child process right before execve()'ing the actual binary, at a point in time where socket
- * activation stderr/stdout area already set up. */
- log_set_always_reopen_console(true);
- }
+ /* First try to spawn ourselves with the right path, and with full serialization. We do this only if
+ * the user didn't specify an explicit init to spawn. */
- if (getpid_cached() == 1 && detect_container() <= 0) {
+ assert(arg_serialization);
+ assert(fds);
- /* Running outside of a container as PID 1 */
- arg_system = true;
- log_set_target(LOG_TARGET_KMSG);
- log_open();
+ xsprintf(sfd, "%i", fileno(arg_serialization));
- if (in_initrd())
- initrd_timestamp = userspace_timestamp;
+ i = 0;
+ args[i++] = SYSTEMD_BINARY_PATH;
+ if (switch_root_dir)
+ args[i++] = "--switched-root";
+ args[i++] = arg_system ? "--system" : "--user";
+ args[i++] = "--deserialize";
+ args[i++] = sfd;
+ args[i++] = NULL;
- if (!skip_setup) {
- r = mount_setup_early();
- if (r < 0) {
- error_message = "Failed to mount early API filesystems";
- goto finish;
- }
+ assert(i <= args_size);
- dual_timestamp_get(&security_start_timestamp);
- if (mac_selinux_setup(&loaded_policy) < 0) {
- error_message = "Failed to load SELinux policy";
- goto finish;
- } else if (mac_smack_setup(&loaded_policy) < 0) {
- error_message = "Failed to load SMACK policy";
- goto finish;
- } else if (ima_setup() < 0) {
- error_message = "Failed to load IMA policy";
- goto finish;
- }
- dual_timestamp_get(&security_finish_timestamp);
- }
+ /*
+ * We want valgrind to print its memory usage summary before reexecution. Valgrind won't do this is on
+ * its own on exec(), but it will do it on exit(). Hence, to ensure we get a summary here, fork() off
+ * a child, let it exit() cleanly, so that it prints the summary, and wait() for it in the parent,
+ * before proceeding into the exec().
+ */
+ valgrind_summary_hack();
- if (mac_selinux_init() < 0) {
- error_message = "Failed to initialize SELinux policy";
- goto finish;
- }
+ (void) execv(args[0], (char* const*) args);
+ log_debug_errno(errno, "Failed to execute our own binary, trying fallback: %m");
+ }
- if (!skip_setup)
- initialize_clock();
+ /* Try the fallback, if there is any, without any serialization. We pass the original argv[] and envp[]. (Well,
+ * modulo the ordering changes due to getopt() in argv[], and some cleanups in envp[], but let's hope that
+ * doesn't matter.) */
- /* Set the default for later on, but don't actually
- * open the logs like this for now. Note that if we
- * are transitioning from the initrd there might still
- * be journal fd open, and we shouldn't attempt
- * opening that before we parsed /proc/cmdline which
- * might redirect output elsewhere. */
- log_set_target(LOG_TARGET_JOURNAL_OR_KMSG);
+ arg_serialization = safe_fclose(arg_serialization);
+ fds = fdset_free(fds);
- } else if (getpid_cached() == 1) {
- /* Running inside a container, as PID 1 */
- arg_system = true;
- log_set_target(LOG_TARGET_CONSOLE);
- log_close_console(); /* force reopen of /dev/console */
- log_open();
+ /* Reopen the console */
+ (void) make_console_stdio();
- /* For later on, see above... */
- log_set_target(LOG_TARGET_JOURNAL);
+ for (j = 1, i = 1; j < (unsigned) argc; j++)
+ args[i++] = argv[j];
+ args[i++] = NULL;
+ assert(i <= args_size);
- /* clear the kernel timestamp,
- * because we are in a container */
- kernel_timestamp = DUAL_TIMESTAMP_NULL;
- } else {
- /* Running as user instance */
- arg_system = false;
- log_set_target(LOG_TARGET_AUTO);
- log_open();
+ /* Reenable any blocked signals, especially important if we switch from initial ramdisk to init=... */
+ (void) reset_all_signal_handlers();
+ (void) reset_signal_mask();
- /* clear the kernel timestamp,
- * because we are not PID 1 */
- kernel_timestamp = DUAL_TIMESTAMP_NULL;
+ if (switch_root_init) {
+ args[0] = switch_root_init;
+ (void) execv(args[0], (char* const*) args);
+ log_warning_errno(errno, "Failed to execute configured init, trying fallback: %m");
}
- initialize_coredump(skip_setup);
+ args[0] = "/sbin/init";
+ (void) execv(args[0], (char* const*) args);
+ r = -errno;
- if (arg_system) {
- if (fixup_environment() < 0) {
- error_message = "Failed to fix up PID1 environment";
- goto finish;
- }
+ manager_status_printf(NULL, STATUS_TYPE_EMERGENCY,
+ ANSI_HIGHLIGHT_RED " !! " ANSI_NORMAL,
+ "Failed to execute /sbin/init");
- /* Try to figure out if we can use colors with the console. No
- * need to do that for user instances since they never log
- * into the console. */
- log_show_color(colors_enabled());
- r = make_null_stdio();
- if (r < 0)
- log_warning_errno(r, "Failed to redirect standard streams to /dev/null: %m");
- }
+ if (r == -ENOENT) {
+ log_warning("No /sbin/init, trying fallback");
- r = initialize_join_controllers();
- if (r < 0) {
- error_message = "Failed to initialize cgroup controllers";
- goto finish;
- }
+ args[0] = "/bin/sh";
+ args[1] = NULL;
+ (void) execv(args[0], (char* const*) args);
+ log_error_errno(errno, "Failed to execute /bin/sh, giving up: %m");
+ } else
+ log_warning_errno(r, "Failed to execute /sbin/init, giving up: %m");
- /* Mount /proc, /sys and friends, so that /proc/cmdline and
- * /proc/$PID/fd is available. */
- if (getpid_cached() == 1) {
+ *ret_error_message = "Failed to execute fallback shell";
+}
- /* Load the kernel modules early. */
- if (!skip_setup)
- kmod_setup();
+static int invoke_main_loop(
+ Manager *m,
+ bool *ret_reexecute,
+ int *ret_retval, /* Return parameters relevant for shutting down */
+ const char **ret_shutdown_verb, /* … */
+ FDSet **ret_fds, /* Return parameters for reexecuting */
+ char **ret_switch_root_dir, /* … */
+ char **ret_switch_root_init, /* … */
+ const char **ret_error_message) {
- r = mount_setup(loaded_policy);
+ int r;
+
+ assert(m);
+ assert(ret_reexecute);
+ assert(ret_retval);
+ assert(ret_shutdown_verb);
+ assert(ret_fds);
+ assert(ret_switch_root_dir);
+ assert(ret_switch_root_init);
+ assert(ret_error_message);
+
+ for (;;) {
+ r = manager_loop(m);
if (r < 0) {
- error_message = "Failed to mount API filesystems";
- goto finish;
+ *ret_error_message = "Failed to run main loop";
+ return log_emergency_errno(r, "Failed to run main loop: %m");
}
- }
- /* Reset all signal handlers. */
- (void) reset_all_signal_handlers();
- (void) ignore_signals(SIGNALS_IGNORE, -1);
+ switch (m->exit_code) {
- arg_default_tasks_max = system_tasks_max_scale(DEFAULT_TASKS_MAX_PERCENTAGE, 100U);
+ case MANAGER_RELOAD:
+ log_info("Reloading.");
- if (parse_config_file() < 0) {
- error_message = "Failed to parse config file";
- goto finish;
- }
+ r = parse_config_file();
+ if (r < 0)
+ log_warning_errno(r, "Failed to parse config file, ignoring: %m");
- if (arg_system) {
- r = proc_cmdline_parse(parse_proc_cmdline_item, NULL, 0);
- if (r < 0)
- log_warning_errno(r, "Failed to parse kernel command line, ignoring: %m");
- }
+ set_manager_defaults(m);
- /* Note that this also parses bits from the kernel command
- * line, including "debug". */
- log_parse_environment();
+ r = manager_reload(m);
+ if (r < 0)
+ log_warning_errno(r, "Failed to reload, ignoring: %m");
- if (parse_argv(argc, argv) < 0) {
- error_message = "Failed to parse commandline arguments";
- goto finish;
- }
+ break;
- /* Initialize default unit */
- if (!arg_default_unit) {
- arg_default_unit = strdup(SPECIAL_DEFAULT_TARGET);
- if (!arg_default_unit) {
- r = log_oom();
- error_message = "Failed to set default unit";
- goto finish;
- }
- }
+ case MANAGER_REEXECUTE:
- if (arg_action == ACTION_TEST &&
- geteuid() == 0) {
- log_error("Don't run test mode as root.");
- goto finish;
- }
+ r = prepare_reexecute(m, &arg_serialization, ret_fds, false);
+ if (r < 0) {
+ *ret_error_message = "Failed to prepare for reexecution";
+ return r;
+ }
- if (!arg_system &&
- arg_action == ACTION_RUN &&
- sd_booted() <= 0) {
- log_error("Trying to run as user instance, but the system has not been booted with systemd.");
- goto finish;
- }
+ log_notice("Reexecuting.");
- if (arg_system &&
- arg_action == ACTION_RUN &&
- running_in_chroot() > 0) {
- log_error("Cannot be run in a chroot() environment.");
- goto finish;
- }
+ *ret_reexecute = true;
+ *ret_retval = EXIT_SUCCESS;
+ *ret_shutdown_verb = NULL;
+ *ret_switch_root_dir = *ret_switch_root_init = NULL;
- if (IN_SET(arg_action, ACTION_TEST, ACTION_HELP)) {
- pager_open(arg_no_pager, false);
- skip_setup = true;
- }
+ return 0;
- if (arg_action == ACTION_HELP) {
- retval = help();
- goto finish;
- } else if (arg_action == ACTION_VERSION) {
- retval = version();
- goto finish;
- } else if (arg_action == ACTION_DUMP_CONFIGURATION_ITEMS) {
- pager_open(arg_no_pager, false);
- unit_dump_config_items(stdout);
- retval = EXIT_SUCCESS;
- goto finish;
- }
+ case MANAGER_SWITCH_ROOT:
+ if (!m->switch_root_init) {
+ r = prepare_reexecute(m, &arg_serialization, ret_fds, true);
+ if (r < 0) {
+ *ret_error_message = "Failed to prepare for reexecution";
+ return r;
+ }
+ } else
+ *ret_fds = NULL;
- if (!arg_system &&
- !getenv("XDG_RUNTIME_DIR")) {
- log_error("Trying to run as user instance, but $XDG_RUNTIME_DIR is not set.");
- goto finish;
- }
+ log_notice("Switching root.");
- assert_se(IN_SET(arg_action, ACTION_RUN, ACTION_TEST));
+ *ret_reexecute = true;
+ *ret_retval = EXIT_SUCCESS;
+ *ret_shutdown_verb = NULL;
- /* Close logging fds, in order not to confuse fdset below */
- log_close();
+ /* Steal the switch root parameters */
+ *ret_switch_root_dir = m->switch_root;
+ *ret_switch_root_init = m->switch_root_init;
+ m->switch_root = m->switch_root_init = NULL;
- /* Remember open file descriptors for later deserialization */
- if (arg_action == ACTION_RUN) {
- r = fdset_new_fill(&fds);
- if (r < 0) {
- log_emergency_errno(r, "Failed to allocate fd set: %m");
- error_message = "Failed to allocate fd set";
- goto finish;
- } else
- fdset_cloexec(fds, true);
+ return 0;
+
+ case MANAGER_EXIT:
+
+ if (MANAGER_IS_USER(m)) {
+ log_debug("Exit.");
+
+ *ret_reexecute = false;
+ *ret_retval = m->return_value;
+ *ret_shutdown_verb = NULL;
+ *ret_fds = NULL;
+ *ret_switch_root_dir = *ret_switch_root_init = NULL;
+
+ return 0;
+ }
+
+ _fallthrough_;
+ case MANAGER_REBOOT:
+ case MANAGER_POWEROFF:
+ case MANAGER_HALT:
+ case MANAGER_KEXEC: {
+ static const char * const table[_MANAGER_EXIT_CODE_MAX] = {
+ [MANAGER_EXIT] = "exit",
+ [MANAGER_REBOOT] = "reboot",
+ [MANAGER_POWEROFF] = "poweroff",
+ [MANAGER_HALT] = "halt",
+ [MANAGER_KEXEC] = "kexec"
+ };
+
+ log_notice("Shutting down.");
+
+ *ret_reexecute = false;
+ *ret_retval = m->return_value;
+ assert_se(*ret_shutdown_verb = table[m->exit_code]);
+ *ret_fds = NULL;
+ *ret_switch_root_dir = *ret_switch_root_init = NULL;
+
+ return 0;
+ }
+
+ default:
+ assert_not_reached("Unknown exit code.");
+ }
+ }
+}
+
+static void log_execution_mode(bool *ret_first_boot) {
+ assert(ret_first_boot);
+
+ if (arg_system) {
+ int v;
+
+ log_info(PACKAGE_STRING " running in %ssystem mode. (" SYSTEMD_FEATURES ")",
+ arg_action == ACTION_TEST ? "test " : "" );
+
+ v = detect_virtualization();
+ if (v > 0)
+ log_info("Detected virtualization %s.", virtualization_to_string(v));
+
+ log_info("Detected architecture %s.", architecture_to_string(uname_architecture()));
+
+ if (in_initrd()) {
+ *ret_first_boot = false;
+ log_info("Running in initial RAM disk.");
+ } else {
+ /* Let's check whether we are in first boot, i.e. whether /etc is still unpopulated. We use
+ * /etc/machine-id as flag file, for this: if it exists we assume /etc is populated, if it
+ * doesn't it's unpopulated. This allows container managers and installers to provision a
+ * couple of files already. If the container manager wants to provision the machine ID itself
+ * it should pass $container_uuid to PID 1. */
+
+ *ret_first_boot = access("/etc/machine-id", F_OK) < 0;
+ if (*ret_first_boot)
+ log_info("Running with unpopulated /etc.");
+ }
+ } else {
+ _cleanup_free_ char *t;
+
+ t = uid_to_name(getuid());
+ log_debug(PACKAGE_STRING " running in %suser mode for user " UID_FMT "/%s. (" SYSTEMD_FEATURES ")",
+ arg_action == ACTION_TEST ? " test" : "", getuid(), strna(t));
+
+ *ret_first_boot = false;
+ }
+}
+
+static int initialize_runtime(
+ bool skip_setup,
+ struct rlimit *saved_rlimit_nofile,
+ struct rlimit *saved_rlimit_memlock,
+ const char **ret_error_message) {
+
+ int r;
+
+ assert(ret_error_message);
+
+ /* Sets up various runtime parameters. Many of these initializations are conditionalized:
+ *
+ * - Some only apply to --system instances
+ * - Some only apply to --user instances
+ * - Some only apply when we first start up, but not when we reexecute
+ */
+
+ if (arg_system && !skip_setup) {
+ if (arg_show_status > 0)
+ status_welcome();
+
+ hostname_setup();
+ machine_id_setup(NULL, arg_machine_id, NULL);
+ loopback_setup();
+ bump_unix_max_dgram_qlen();
+ test_usr();
+ write_container_id();
+ }
+
+ if (arg_system && arg_watchdog_device) {
+ r = watchdog_set_device(arg_watchdog_device);
+ if (r < 0)
+ log_warning_errno(r, "Failed to set watchdog device to %s, ignoring: %m",
+ arg_watchdog_device);
+ }
+
+ if (arg_system && arg_runtime_watchdog > 0 && arg_runtime_watchdog != USEC_INFINITY)
+ watchdog_set_timeout(&arg_runtime_watchdog);
+
+ if (arg_timer_slack_nsec != NSEC_INFINITY)
+ if (prctl(PR_SET_TIMERSLACK, arg_timer_slack_nsec) < 0)
+ log_error_errno(errno, "Failed to adjust timer slack: %m");
+
+ if (arg_system && !cap_test_all(arg_capability_bounding_set)) {
+ r = capability_bounding_set_drop_usermode(arg_capability_bounding_set);
+ if (r < 0) {
+ *ret_error_message = "Failed to drop capability bounding set of usermode helpers";
+ return log_emergency_errno(r, "Failed to drop capability bounding set of usermode helpers: %m");
+ }
+
+ r = capability_bounding_set_drop(arg_capability_bounding_set, true);
+ if (r < 0) {
+ *ret_error_message = "Failed to drop capability bounding set";
+ return log_emergency_errno(r, "Failed to drop capability bounding set: %m");
+ }
+ }
+
+ if (arg_syscall_archs) {
+ r = enforce_syscall_archs(arg_syscall_archs);
+ if (r < 0) {
+ *ret_error_message = "Failed to set syscall architectures";
+ return r;
+ }
+ }
+
+ if (!arg_system)
+ /* Become reaper of our children */
+ if (prctl(PR_SET_CHILD_SUBREAPER, 1) < 0)
+ log_warning_errno(errno, "Failed to make us a subreaper: %m");
+
+ if (arg_system) {
+ /* Bump up RLIMIT_NOFILE for systemd itself */
+ (void) bump_rlimit_nofile(saved_rlimit_nofile);
+ (void) bump_rlimit_memlock(saved_rlimit_memlock);
+ }
+
+ return 0;
+}
+
+static int do_queue_default_job(
+ Manager *m,
+ const char **ret_error_message) {
+
+ _cleanup_(sd_bus_error_free) sd_bus_error error = SD_BUS_ERROR_NULL;
+ Job *default_unit_job;
+ Unit *target = NULL;
+ int r;
+
+ log_debug("Activating default unit: %s", arg_default_unit);
+
+ r = manager_load_unit(m, arg_default_unit, NULL, &error, &target);
+ if (r < 0)
+ log_error("Failed to load default target: %s", bus_error_message(&error, r));
+ else if (IN_SET(target->load_state, UNIT_ERROR, UNIT_NOT_FOUND))
+ log_error_errno(target->load_error, "Failed to load default target: %m");
+ else if (target->load_state == UNIT_MASKED)
+ log_error("Default target masked.");
+
+ if (!target || target->load_state != UNIT_LOADED) {
+ log_info("Trying to load rescue target...");
+
+ r = manager_load_unit(m, SPECIAL_RESCUE_TARGET, NULL, &error, &target);
+ if (r < 0) {
+ *ret_error_message = "Failed to load rescue target";
+ return log_emergency_errno(r, "Failed to load rescue target: %s", bus_error_message(&error, r));
+ } else if (IN_SET(target->load_state, UNIT_ERROR, UNIT_NOT_FOUND)) {
+ *ret_error_message = "Failed to load rescue target";
+ return log_emergency_errno(target->load_error, "Failed to load rescue target: %m");
+ } else if (target->load_state == UNIT_MASKED) {
+ *ret_error_message = "Rescue target masked";
+ log_emergency("Rescue target masked.");
+ return -ERFKILL;
+ }
+ }
+
+ assert(target->load_state == UNIT_LOADED);
+
+ r = manager_add_job(m, JOB_START, target, JOB_ISOLATE, &error, &default_unit_job);
+ if (r == -EPERM) {
+ log_debug_errno(r, "Default target could not be isolated, starting instead: %s", bus_error_message(&error, r));
+
+ sd_bus_error_free(&error);
+
+ r = manager_add_job(m, JOB_START, target, JOB_REPLACE, &error, &default_unit_job);
+ if (r < 0) {
+ *ret_error_message = "Failed to start default target";
+ return log_emergency_errno(r, "Failed to start default target: %s", bus_error_message(&error, r));
+ }
+
+ } else if (r < 0) {
+ *ret_error_message = "Failed to isolate default target";
+ return log_emergency_errno(r, "Failed to isolate default target: %s", bus_error_message(&error, r));
+ }
+
+ m->default_unit_job_id = default_unit_job->id;
+
+ return 0;
+}
+
+static void free_arguments(void) {
+ size_t j;
+
+ /* Frees all arg_* variables, with the exception of arg_serialization */
+
+ for (j = 0; j < ELEMENTSOF(arg_default_rlimit); j++)
+ arg_default_rlimit[j] = mfree(arg_default_rlimit[j]);
+
+ arg_default_unit = mfree(arg_default_unit);
+ arg_confirm_spawn = mfree(arg_confirm_spawn);
+ arg_join_controllers = strv_free_free(arg_join_controllers);
+ arg_default_environment = strv_free(arg_default_environment);
+ arg_syscall_archs = set_free(arg_syscall_archs);
+}
+
+static int load_configuration(int argc, char **argv, const char **ret_error_message) {
+ int r;
+
+ assert(ret_error_message);
+
+ arg_default_tasks_max = system_tasks_max_scale(DEFAULT_TASKS_MAX_PERCENTAGE, 100U);
+
+ r = parse_config_file();
+ if (r < 0) {
+ *ret_error_message = "Failed to parse config file";
+ return r;
+ }
+
+ if (arg_system) {
+ r = proc_cmdline_parse(parse_proc_cmdline_item, NULL, 0);
+ if (r < 0)
+ log_warning_errno(r, "Failed to parse kernel command line, ignoring: %m");
+ }
+
+ /* Note that this also parses bits from the kernel command line, including "debug". */
+ log_parse_environment();
+
+ r = parse_argv(argc, argv);
+ if (r < 0) {
+ *ret_error_message = "Failed to parse commandline arguments";
+ return r;
+ }
+
+ /* Initialize default unit */
+ if (!arg_default_unit) {
+ arg_default_unit = strdup(SPECIAL_DEFAULT_TARGET);
+ if (!arg_default_unit) {
+ *ret_error_message = "Failed to set default unit";
+ return log_oom();
+ }
+ }
+
+ return 0;
+}
+
+static int safety_checks(void) {
+
+ if (arg_action == ACTION_TEST &&
+ geteuid() == 0) {
+ log_error("Don't run test mode as root.");
+ return -EPERM;
+ }
+
+ if (!arg_system &&
+ arg_action == ACTION_RUN &&
+ sd_booted() <= 0) {
+ log_error("Trying to run as user instance, but the system has not been booted with systemd.");
+ return -EOPNOTSUPP;
+ }
+
+ if (!arg_system &&
+ arg_action == ACTION_RUN &&
+ !getenv("XDG_RUNTIME_DIR")) {
+ log_error("Trying to run as user instance, but $XDG_RUNTIME_DIR is not set.");
+ return -EUNATCH;
+ }
+
+ if (arg_system &&
+ arg_action == ACTION_RUN &&
+ running_in_chroot() > 0) {
+ log_error("Cannot be run in a chroot() environment.");
+ return -EOPNOTSUPP;
+ }
+
+ return 0;
+}
+
+static int initialize_security(
+ bool *loaded_policy,
+ dual_timestamp *security_start_timestamp,
+ dual_timestamp *security_finish_timestamp,
+ const char **ret_error_message) {
+
+ int r;
+
+ assert(loaded_policy);
+ assert(security_start_timestamp);
+ assert(security_finish_timestamp);
+ assert(ret_error_message);
+
+ dual_timestamp_get(security_start_timestamp);
+
+ r = mac_selinux_setup(loaded_policy) < 0;
+ if (r < 0) {
+ *ret_error_message = "Failed to load SELinux policy";
+ return r;
+ }
+
+ r = mac_smack_setup(loaded_policy);
+ if (r < 0) {
+ *ret_error_message = "Failed to load SMACK policy";
+ return r;
+ }
+
+ r = ima_setup();
+ if (r < 0) {
+ *ret_error_message = "Failed to load IMA policy";
+ return r;
+ }
+
+ dual_timestamp_get(security_finish_timestamp);
+ return 0;
+}
+
+int main(int argc, char *argv[]) {
+ Manager *m = NULL;
+ int r, retval = EXIT_FAILURE;
+ usec_t before_startup, after_startup;
+ char timespan[FORMAT_TIMESPAN_MAX];
+ FDSet *fds = NULL;
+ bool reexecute = false;
+ const char *shutdown_verb = NULL;
+ dual_timestamp initrd_timestamp = DUAL_TIMESTAMP_NULL;
+ dual_timestamp userspace_timestamp = DUAL_TIMESTAMP_NULL;
+ dual_timestamp kernel_timestamp = DUAL_TIMESTAMP_NULL;
+ dual_timestamp security_start_timestamp = DUAL_TIMESTAMP_NULL;
+ dual_timestamp security_finish_timestamp = DUAL_TIMESTAMP_NULL;
+ static char systemd[] = "systemd";
+ bool skip_setup = false;
+ bool loaded_policy = false;
+ bool queue_default_job = false;
+ bool first_boot = false;
+ char *switch_root_dir = NULL, *switch_root_init = NULL;
+ struct rlimit saved_rlimit_nofile = RLIMIT_MAKE_CONST(0), saved_rlimit_memlock = RLIMIT_MAKE_CONST((rlim_t) -1);
+ const char *error_message = NULL;
+
+ redirect_telinit(argc, argv);
+
+ dual_timestamp_from_monotonic(&kernel_timestamp, 0);
+ dual_timestamp_get(&userspace_timestamp);
+
+ /* Determine if this is a reexecution or normal bootup. We do
+ * the full command line parsing much later, so let's just
+ * have a quick peek here. */
+ if (strv_find(argv+1, "--deserialize"))
+ skip_setup = true;
+
+ /* If we have switched root, do all the special setup
+ * things */
+ if (strv_find(argv+1, "--switched-root"))
+ skip_setup = false;
+
+ /* If we get started via the /sbin/init symlink then we are
+ called 'init'. After a subsequent reexecution we are then
+ called 'systemd'. That is confusing, hence let's call us
+ systemd right-away. */
+ program_invocation_short_name = systemd;
+ (void) prctl(PR_SET_NAME, systemd);
+
+ saved_argv = argv;
+ saved_argc = argc;
+
+ log_set_upgrade_syslog_to_journal(true);
+
+ if (getpid_cached() == 1) {
+ /* Disable the umask logic */
+ umask(0);
+
+ /* Always reopen /dev/console when running as PID 1 or one of its pre-execve() children. This is
+ * important so that we never end up logging to any foreign stderr, for example if we have to log in a
+ * child process right before execve()'ing the actual binary, at a point in time where socket
+ * activation stderr/stdout area already set up. */
+ log_set_always_reopen_console(true);
+ }
+
+ if (getpid_cached() == 1 && detect_container() <= 0) {
+
+ /* Running outside of a container as PID 1 */
+ arg_system = true;
+ log_set_target(LOG_TARGET_KMSG);
+ log_open();
+
+ if (in_initrd())
+ initrd_timestamp = userspace_timestamp;
+
+ if (!skip_setup) {
+ r = mount_setup_early();
+ if (r < 0) {
+ error_message = "Failed to mount early API filesystems";
+ goto finish;
+ }
+
+ r = initialize_security(
+ &loaded_policy,
+ &security_start_timestamp,
+ &security_finish_timestamp,
+ &error_message);
+ if (r < 0)
+ goto finish;
+ }
+
+ if (mac_selinux_init() < 0) {
+ error_message = "Failed to initialize SELinux policy";
+ goto finish;
+ }
+
+ if (!skip_setup)
+ initialize_clock();
+
+ /* Set the default for later on, but don't actually
+ * open the logs like this for now. Note that if we
+ * are transitioning from the initrd there might still
+ * be journal fd open, and we shouldn't attempt
+ * opening that before we parsed /proc/cmdline which
+ * might redirect output elsewhere. */
+ log_set_target(LOG_TARGET_JOURNAL_OR_KMSG);
+
+ } else if (getpid_cached() == 1) {
+ /* Running inside a container, as PID 1 */
+ arg_system = true;
+ log_set_target(LOG_TARGET_CONSOLE);
+ log_close_console(); /* force reopen of /dev/console */
+ log_open();
+
+ /* For later on, see above... */
+ log_set_target(LOG_TARGET_JOURNAL);
+
+ /* clear the kernel timestamp,
+ * because we are in a container */
+ kernel_timestamp = DUAL_TIMESTAMP_NULL;
+ } else {
+ /* Running as user instance */
+ arg_system = false;
+ log_set_target(LOG_TARGET_AUTO);
+ log_open();
+
+ /* clear the kernel timestamp,
+ * because we are not PID 1 */
+ kernel_timestamp = DUAL_TIMESTAMP_NULL;
+ }
+
+ initialize_coredump(skip_setup);
+
+ if (arg_system) {
+ if (fixup_environment() < 0) {
+ error_message = "Failed to fix up PID1 environment";
+ goto finish;
+ }
+
+ /* Try to figure out if we can use colors with the console. No
+ * need to do that for user instances since they never log
+ * into the console. */
+ log_show_color(colors_enabled());
+ r = make_null_stdio();
+ if (r < 0)
+ log_warning_errno(r, "Failed to redirect standard streams to /dev/null: %m");
+ }
+
+ r = initialize_join_controllers();
+ if (r < 0) {
+ error_message = "Failed to initialize cgroup controllers";
+ goto finish;
+ }
+
+ /* Mount /proc, /sys and friends, so that /proc/cmdline and
+ * /proc/$PID/fd is available. */
+ if (getpid_cached() == 1) {
+
+ /* Load the kernel modules early. */
+ if (!skip_setup)
+ kmod_setup();
+
+ r = mount_setup(loaded_policy);
+ if (r < 0) {
+ error_message = "Failed to mount API filesystems";
+ goto finish;
+ }
+ }
+
+ /* Reset all signal handlers. */
+ (void) reset_all_signal_handlers();
+ (void) ignore_signals(SIGNALS_IGNORE, -1);
+
+ r = load_configuration(argc, argv, &error_message);
+ if (r < 0)
+ goto finish;
+
+ r = safety_checks();
+ if (r < 0)
+ goto finish;
+
+ if (IN_SET(arg_action, ACTION_TEST, ACTION_HELP, ACTION_DUMP_CONFIGURATION_ITEMS))
+ pager_open(arg_no_pager, false);
+
+ if (arg_action != ACTION_RUN)
+ skip_setup = true;
+
+ if (arg_action == ACTION_HELP) {
+ retval = help();
+ goto finish;
+ } else if (arg_action == ACTION_VERSION) {
+ retval = version();
+ goto finish;
+ } else if (arg_action == ACTION_DUMP_CONFIGURATION_ITEMS) {
+ unit_dump_config_items(stdout);
+ retval = EXIT_SUCCESS;
+ goto finish;
+ }
+
+ assert_se(IN_SET(arg_action, ACTION_RUN, ACTION_TEST));
+
+ /* Close logging fds, in order not to confuse fdset below */
+ log_close();
+
+ /* Remember open file descriptors for later deserialization */
+ if (arg_action == ACTION_RUN) {
+ r = fdset_new_fill(&fds);
+ if (r < 0) {
+ log_emergency_errno(r, "Failed to allocate fd set: %m");
+ error_message = "Failed to allocate fd set";
+ goto finish;
+ } else
+ fdset_cloexec(fds, true);
if (arg_serialization)
assert_se(fdset_remove(fds, fileno(arg_serialization)) >= 0);
goto finish;
}
- if (arg_system) {
- int v;
-
- log_info(PACKAGE_STRING " running in %ssystem mode. (" SYSTEMD_FEATURES ")",
- arg_action == ACTION_TEST ? "test " : "" );
-
- v = detect_virtualization();
- if (v > 0)
- log_info("Detected virtualization %s.", virtualization_to_string(v));
-
- write_container_id();
-
- log_info("Detected architecture %s.", architecture_to_string(uname_architecture()));
-
- if (in_initrd())
- log_info("Running in initial RAM disk.");
- else {
- /* Let's check whether we are in first boot, i.e. whether /etc is still unpopulated. We use
- * /etc/machine-id as flag file, for this: if it exists we assume /etc is populated, if it
- * doesn't it's unpopulated. This allows container managers and installers to provision a
- * couple of files already. If the container manager wants to provision the machine ID itself
- * it should pass $container_uuid to PID 1. */
-
- first_boot = access("/etc/machine-id", F_OK) < 0;
- if (first_boot)
- log_info("Running with unpopulated /etc.");
- }
- } else {
- _cleanup_free_ char *t;
-
- t = uid_to_name(getuid());
- log_debug(PACKAGE_STRING " running in %suser mode for user " UID_FMT "/%s. (" SYSTEMD_FEATURES ")",
- arg_action == ACTION_TEST ? " test" : "", getuid(), strna(t));
- }
+ log_execution_mode(&first_boot);
if (arg_action == ACTION_RUN) {
- if (arg_system && !skip_setup) {
- if (arg_show_status > 0)
- status_welcome();
-
- hostname_setup();
- machine_id_setup(NULL, arg_machine_id, NULL);
- loopback_setup();
- bump_unix_max_dgram_qlen();
-
- test_usr();
- }
-
- if (arg_system && arg_runtime_watchdog > 0 && arg_runtime_watchdog != USEC_INFINITY)
- watchdog_set_timeout(&arg_runtime_watchdog);
-
- if (arg_timer_slack_nsec != NSEC_INFINITY)
- if (prctl(PR_SET_TIMERSLACK, arg_timer_slack_nsec) < 0)
- log_error_errno(errno, "Failed to adjust timer slack: %m");
-
- if (arg_system && !cap_test_all(arg_capability_bounding_set)) {
- r = capability_bounding_set_drop_usermode(arg_capability_bounding_set);
- if (r < 0) {
- log_emergency_errno(r, "Failed to drop capability bounding set of usermode helpers: %m");
- error_message = "Failed to drop capability bounding set of usermode helpers";
- goto finish;
- }
- r = capability_bounding_set_drop(arg_capability_bounding_set, true);
- if (r < 0) {
- log_emergency_errno(r, "Failed to drop capability bounding set: %m");
- error_message = "Failed to drop capability bounding set";
- goto finish;
- }
- }
-
- if (arg_syscall_archs) {
- r = enforce_syscall_archs(arg_syscall_archs);
- if (r < 0) {
- error_message = "Failed to set syscall architectures";
- goto finish;
- }
- }
-
- if (!arg_system)
- /* Become reaper of our children */
- if (prctl(PR_SET_CHILD_SUBREAPER, 1) < 0)
- log_warning_errno(errno, "Failed to make us a subreaper: %m");
-
- if (arg_system) {
- /* Bump up RLIMIT_NOFILE for systemd itself */
- (void) bump_rlimit_nofile(&saved_rlimit_nofile);
- (void) bump_rlimit_memlock(&saved_rlimit_memlock);
- }
+ r = initialize_runtime(skip_setup,
+ &saved_rlimit_nofile,
+ &saved_rlimit_memlock,
+ &error_message);
+ if (r < 0)
+ goto finish;
}
r = manager_new(arg_system ? UNIT_FILE_SYSTEM : UNIT_FILE_USER,
goto finish;
}
- m->userspace_timestamp = userspace_timestamp;
- m->kernel_timestamp = kernel_timestamp;
- m->initrd_timestamp = initrd_timestamp;
- m->security_start_timestamp = security_start_timestamp;
- m->security_finish_timestamp = security_finish_timestamp;
+ m->timestamps[MANAGER_TIMESTAMP_KERNEL] = kernel_timestamp;
+ m->timestamps[MANAGER_TIMESTAMP_INITRD] = initrd_timestamp;
+ m->timestamps[MANAGER_TIMESTAMP_USERSPACE] = userspace_timestamp;
+ m->timestamps[MANAGER_TIMESTAMP_SECURITY_START] = security_start_timestamp;
+ m->timestamps[MANAGER_TIMESTAMP_SECURITY_FINISH] = security_finish_timestamp;
set_manager_defaults(m);
set_manager_settings(m);
r = manager_startup(m, arg_serialization, fds);
if (r < 0) {
log_error_errno(r, "Failed to fully start up daemon: %m");
+ error_message = "Failed to start up manager";
goto finish;
}
- /* This will close all file descriptors that were opened, but
- * not claimed by any unit. */
+ /* This will close all file descriptors that were opened, but not claimed by any unit. */
fds = fdset_free(fds);
-
arg_serialization = safe_fclose(arg_serialization);
if (queue_default_job) {
- _cleanup_(sd_bus_error_free) sd_bus_error error = SD_BUS_ERROR_NULL;
- Unit *target = NULL;
- Job *default_unit_job;
-
- log_debug("Activating default unit: %s", arg_default_unit);
-
- r = manager_load_unit(m, arg_default_unit, NULL, &error, &target);
+ r = do_queue_default_job(m, &error_message);
if (r < 0)
- log_error("Failed to load default target: %s", bus_error_message(&error, r));
- else if (IN_SET(target->load_state, UNIT_ERROR, UNIT_NOT_FOUND))
- log_error_errno(target->load_error, "Failed to load default target: %m");
- else if (target->load_state == UNIT_MASKED)
- log_error("Default target masked.");
-
- if (!target || target->load_state != UNIT_LOADED) {
- log_info("Trying to load rescue target...");
-
- r = manager_load_unit(m, SPECIAL_RESCUE_TARGET, NULL, &error, &target);
- if (r < 0) {
- log_emergency("Failed to load rescue target: %s", bus_error_message(&error, r));
- error_message = "Failed to load rescue target";
- goto finish;
- } else if (IN_SET(target->load_state, UNIT_ERROR, UNIT_NOT_FOUND)) {
- log_emergency_errno(target->load_error, "Failed to load rescue target: %m");
- error_message = "Failed to load rescue target";
- goto finish;
- } else if (target->load_state == UNIT_MASKED) {
- log_emergency("Rescue target masked.");
- error_message = "Rescue target masked";
- goto finish;
- }
- }
-
- assert(target->load_state == UNIT_LOADED);
-
- if (arg_action == ACTION_TEST) {
- printf("-> By units:\n");
- manager_dump_units(m, stdout, "\t");
- }
-
- r = manager_add_job(m, JOB_START, target, JOB_ISOLATE, &error, &default_unit_job);
- if (r == -EPERM) {
- log_debug("Default target could not be isolated, starting instead: %s", bus_error_message(&error, r));
-
- sd_bus_error_free(&error);
-
- r = manager_add_job(m, JOB_START, target, JOB_REPLACE, &error, &default_unit_job);
- if (r < 0) {
- log_emergency("Failed to start default target: %s", bus_error_message(&error, r));
- error_message = "Failed to start default target";
- goto finish;
- }
- } else if (r < 0) {
- log_emergency("Failed to isolate default target: %s", bus_error_message(&error, r));
- error_message = "Failed to isolate default target";
- goto finish;
- }
-
- m->default_unit_job_id = default_unit_job->id;
-
- after_startup = now(CLOCK_MONOTONIC);
- log_full(arg_action == ACTION_TEST ? LOG_INFO : LOG_DEBUG,
- "Loaded units and determined initial transaction in %s.",
- format_timespan(timespan, sizeof(timespan), after_startup - before_startup, 100 * USEC_PER_MSEC));
-
- if (arg_action == ACTION_TEST) {
- printf("-> By jobs:\n");
- manager_dump_jobs(m, stdout, "\t");
- retval = EXIT_SUCCESS;
goto finish;
- }
}
- for (;;) {
- r = manager_loop(m);
- if (r < 0) {
- log_emergency_errno(r, "Failed to run main loop: %m");
- error_message = "Failed to run main loop";
- goto finish;
- }
-
- switch (m->exit_code) {
-
- case MANAGER_RELOAD:
- log_info("Reloading.");
-
- r = parse_config_file();
- if (r < 0)
- log_error("Failed to parse config file.");
-
- set_manager_defaults(m);
-
- r = manager_reload(m);
- if (r < 0)
- log_error_errno(r, "Failed to reload: %m");
- break;
-
- case MANAGER_REEXECUTE:
-
- if (prepare_reexecute(m, &arg_serialization, &fds, false) < 0) {
- error_message = "Failed to prepare for reexecution";
- goto finish;
- }
-
- reexecute = true;
- log_notice("Reexecuting.");
- goto finish;
-
- case MANAGER_SWITCH_ROOT:
- /* Steal the switch root parameters */
- switch_root_dir = m->switch_root;
- switch_root_init = m->switch_root_init;
- m->switch_root = m->switch_root_init = NULL;
-
- if (!switch_root_init)
- if (prepare_reexecute(m, &arg_serialization, &fds, true) < 0) {
- error_message = "Failed to prepare for reexecution";
- goto finish;
- }
-
- reexecute = true;
- log_notice("Switching root.");
- goto finish;
-
- case MANAGER_EXIT:
- retval = m->return_value;
-
- if (MANAGER_IS_USER(m)) {
- log_debug("Exit.");
- goto finish;
- }
+ after_startup = now(CLOCK_MONOTONIC);
- /* fallthrough */
- case MANAGER_REBOOT:
- case MANAGER_POWEROFF:
- case MANAGER_HALT:
- case MANAGER_KEXEC: {
- static const char * const table[_MANAGER_EXIT_CODE_MAX] = {
- [MANAGER_EXIT] = "exit",
- [MANAGER_REBOOT] = "reboot",
- [MANAGER_POWEROFF] = "poweroff",
- [MANAGER_HALT] = "halt",
- [MANAGER_KEXEC] = "kexec"
- };
+ log_full(arg_action == ACTION_TEST ? LOG_INFO : LOG_DEBUG,
+ "Loaded units and determined initial transaction in %s.",
+ format_timespan(timespan, sizeof(timespan), after_startup - before_startup, 100 * USEC_PER_MSEC));
- assert_se(shutdown_verb = table[m->exit_code]);
- arm_reboot_watchdog = m->exit_code == MANAGER_REBOOT;
+ if (arg_system) {
+ _cleanup_free_ char *taint;
+
+ taint = manager_taint_string(m);
+ if (!isempty(taint))
+ log_struct(LOG_NOTICE,
+ LOG_MESSAGE("System is tainted: %s", taint),
+ "TAINT=%s", taint,
+ "MESSAGE_ID=" SD_MESSAGE_TAINTED_STR,
+ NULL);
+ }
- log_notice("Shutting down.");
- goto finish;
- }
+ if (arg_action == ACTION_TEST) {
+ printf("-> By units:\n");
+ manager_dump_units(m, stdout, "\t");
- default:
- assert_not_reached("Unknown exit code.");
- }
+ printf("-> By jobs:\n");
+ manager_dump_jobs(m, stdout, "\t");
+ retval = EXIT_SUCCESS;
+ goto finish;
}
+ r = invoke_main_loop(m,
+ &reexecute,
+ &retval,
+ &shutdown_verb,
+ &fds,
+ &switch_root_dir,
+ &switch_root_init,
+ &error_message);
+
finish:
pager_close();
m = manager_free(m);
- for (j = 0; j < ELEMENTSOF(arg_default_rlimit); j++)
- arg_default_rlimit[j] = mfree(arg_default_rlimit[j]);
-
- arg_default_unit = mfree(arg_default_unit);
- arg_confirm_spawn = mfree(arg_confirm_spawn);
- arg_join_controllers = strv_free_free(arg_join_controllers);
- arg_default_environment = strv_free(arg_default_environment);
- arg_syscall_archs = set_free(arg_syscall_archs);
-
+ free_arguments();
mac_selinux_finish();
- if (reexecute) {
- const char **args;
- unsigned i, args_size;
-
- /* Close and disarm the watchdog, so that the new
- * instance can reinitialize it, but doesn't get
- * rebooted while we do that */
- watchdog_close(true);
-
- /* Reset the RLIMIT_NOFILE to the kernel default, so
- * that the new systemd can pass the kernel default to
- * its child processes */
- if (saved_rlimit_nofile.rlim_cur > 0)
- (void) setrlimit(RLIMIT_NOFILE, &saved_rlimit_nofile);
- if (saved_rlimit_memlock.rlim_cur != (rlim_t) -1)
- (void) setrlimit(RLIMIT_MEMLOCK, &saved_rlimit_memlock);
-
- if (switch_root_dir) {
- /* Kill all remaining processes from the
- * initrd, but don't wait for them, so that we
- * can handle the SIGCHLD for them after
- * deserializing. */
- broadcast_signal(SIGTERM, false, true);
-
- /* And switch root with MS_MOVE, because we remove the old directory afterwards and detach it. */
- r = switch_root(switch_root_dir, "/mnt", true, MS_MOVE);
- if (r < 0)
- log_error_errno(r, "Failed to switch root, trying to continue: %m");
- }
-
- args_size = MAX(6, argc+1);
- args = newa(const char*, args_size);
-
- if (!switch_root_init) {
- char sfd[DECIMAL_STR_MAX(int) + 1];
-
- /* First try to spawn ourselves with the right
- * path, and with full serialization. We do
- * this only if the user didn't specify an
- * explicit init to spawn. */
-
- assert(arg_serialization);
- assert(fds);
-
- xsprintf(sfd, "%i", fileno(arg_serialization));
-
- i = 0;
- args[i++] = SYSTEMD_BINARY_PATH;
- if (switch_root_dir)
- args[i++] = "--switched-root";
- args[i++] = arg_system ? "--system" : "--user";
- args[i++] = "--deserialize";
- args[i++] = sfd;
- args[i++] = NULL;
-
- assert(i <= args_size);
-
- /*
- * We want valgrind to print its memory usage summary before reexecution.
- * Valgrind won't do this is on its own on exec(), but it will do it on exit().
- * Hence, to ensure we get a summary here, fork() off a child, let it exit() cleanly,
- * so that it prints the summary, and wait() for it in the parent, before proceeding into the exec().
- */
- valgrind_summary_hack();
-
- (void) execv(args[0], (char* const*) args);
- }
-
- /* Try the fallback, if there is any, without any
- * serialization. We pass the original argv[] and
- * envp[]. (Well, modulo the ordering changes due to
- * getopt() in argv[], and some cleanups in envp[],
- * but let's hope that doesn't matter.) */
-
- arg_serialization = safe_fclose(arg_serialization);
- fds = fdset_free(fds);
-
- /* Reopen the console */
- (void) make_console_stdio();
-
- for (j = 1, i = 1; j < (unsigned) argc; j++)
- args[i++] = argv[j];
- args[i++] = NULL;
- assert(i <= args_size);
-
- /* Reenable any blocked signals, especially important
- * if we switch from initial ramdisk to init=... */
- (void) reset_all_signal_handlers();
- (void) reset_signal_mask();
-
- if (switch_root_init) {
- args[0] = switch_root_init;
- (void) execv(args[0], (char* const*) args);
- log_warning_errno(errno, "Failed to execute configured init, trying fallback: %m");
- }
-
- args[0] = "/sbin/init";
- (void) execv(args[0], (char* const*) args);
-
- manager_status_printf(NULL, STATUS_TYPE_EMERGENCY,
- ANSI_HIGHLIGHT_RED " !! " ANSI_NORMAL,
- "Failed to execute /sbin/init");
-
- if (errno == ENOENT) {
- log_warning("No /sbin/init, trying fallback");
-
- args[0] = "/bin/sh";
- args[1] = NULL;
- (void) execv(args[0], (char* const*) args);
- log_error_errno(errno, "Failed to execute /bin/sh, giving up: %m");
- } else
- log_warning_errno(errno, "Failed to execute /sbin/init, giving up: %m");
-
- error_message = "Failed to execute fallback shell";
- }
+ if (reexecute)
+ do_reexecute(argc, argv,
+ &saved_rlimit_nofile,
+ &saved_rlimit_memlock,
+ fds,
+ switch_root_dir,
+ switch_root_init,
+ &error_message); /* This only returns if reexecution failed */
arg_serialization = safe_fclose(arg_serialization);
fds = fdset_free(fds);
* here explicitly. valgrind will only generate nice output on
* exit(), not on exec(), hence let's do the former not the
* latter here. */
- if (getpid_cached() == 1 && RUNNING_ON_VALGRIND)
+ if (getpid_cached() == 1 && RUNNING_ON_VALGRIND) {
+ /* Cleanup watchdog_device strings for valgrind. We need them
+ * in become_shutdown() so normally we cannot free them yet. */
+ watchdog_free_device();
+ arg_watchdog_device = mfree(arg_watchdog_device);
return 0;
+ }
#endif
if (shutdown_verb) {
- r = become_shutdown(shutdown_verb, retval, arm_reboot_watchdog);
+ r = become_shutdown(shutdown_verb, retval);
log_error_errno(r, "Failed to execute shutdown binary, %s: %m", getpid_cached() == 1 ? "freezing" : "quitting");
error_message = "Failed to execute shutdown binary";
}
+ watchdog_free_device();
+ arg_watchdog_device = mfree(arg_watchdog_device);
+
if (getpid_cached() == 1) {
if (error_message)
manager_status_printf(NULL, STATUS_TYPE_EMERGENCY,