systemd关机流程

命令行调用流程

• 在系统中去使用poweoff，shutdown，halt等命令关机，其实都是是指向systemctl的软链接。

uos@uos-PC:~/Desktop$ file /usr/sbin/poweroff 
/usr/sbin/poweroff: symbolic link to /bin/systemctl

uos@uos-PC:~/Desktop$ file /usr/sbin/shutdown 
/usr/sbin/shutdown: symbolic link to /bin/systemctl

uos@uos-PC:~/Desktop$ file /usr/sbin/halt 
/usr/sbin/halt: symbolic link to /bin/systemctl

• 通过dbus调用到systemd-logind进程中的某个方法

tatic int logind_schedule_shutdown(void) {
.....
        r = sd_bus_call_method(
                        bus,
                        "org.freedesktop.login1",
                        "/org/freedesktop/login1",
                        "org.freedesktop.login1.Manager",
                        "ScheduleShutdown",
                        &error,
                        NULL,
                        "st",
                        action,
                        arg_when);
.....
 }

• 在systemd-logind中会去执行execute_shutdown_or_sleep

• 然后再通过dbus调用１号进程的method_start_unit方法，参数是poweroff.target

        r = sd_bus_call_method(
                        m->bus,
                        "org.freedesktop.systemd1",
                        "/org/freedesktop/systemd1",
                        "org.freedesktop.systemd1.Manager",
                        "StartUnit",
                        error,
                        &reply,
                        "ss", unit_name, "replace-irreversibly");

• 以上执行完后会退出１号进程，执行systemd-shutdown

                case MANAGER_KEXEC: {
                        static const char * const table[_MANAGER_OBJECTIVE_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->objective]);
                        *ret_fds = NULL;
                        *ret_switch_root_dir = *ret_switch_root_init = NULL;

                        return 0;
                }

static int become_shutdown(
                const char *shutdown_verb,
                int retval) {
                
                .....
                execve(SYSTEMD_SHUTDOWN_BINARY_PATH, (char **) command_line, env_block);
                .....
                
                }

调用该二进制systemd-shutdown，它的源码shutdow.c

        log_info("Sending SIGTERM to remaining processes...");
        broadcast_signal(SIGTERM, true, true, arg_timeout);

        log_info("Sending SIGKILL to remaining processes...");
        broadcast_signal(SIGKILL, true, false, arg_timeout);

 /* Unmount all mountpoints, swaps, and loopback devices */
        for (;;) {
                bool changed = false;

                if (use_watchdog)
                        watchdog_ping();

                /* Let's trim the cgroup tree on each iteration so
                   that we leave an empty cgroup tree around, so that
                   container managers get a nice notify event when we
                   are down */
                if (cgroup)
                        cg_trim(SYSTEMD_CGROUP_CONTROLLER, cgroup, false);

                if (need_umount) {
                        log_info("Unmounting file systems.");
                        r = umount_all(&changed, umount_log_level);
                        if (r == 0) {
                                need_umount = false;
                                log_info("All filesystems unmounted.");
                        } else if (r > 0)
                                log_info("Not all file systems unmounted, %d left.", r);
                        else
                                log_error_errno(r, "Failed to unmount file systems: %m");
                }

                if (need_swapoff) {
                        log_info("Deactivating swaps.");
                        r = swapoff_all(&changed);
                        if (r == 0) {
                                need_swapoff = false;
                                log_info("All swaps deactivated.");
                        } else if (r > 0)
                                log_info("Not all swaps deactivated, %d left.", r);
                        else
                                log_error_errno(r, "Failed to deactivate swaps: %m");
                }

                if (need_loop_detach) {
                        log_info("Detaching loop devices.");
                        r = loopback_detach_all(&changed, umount_log_level);
                        if (r == 0) {
                                need_loop_detach = false;
                                log_info("All loop devices detached.");
                        } else if (r > 0)
                                log_info("Not all loop devices detached, %d left.", r);
                        else
                                log_error_errno(r, "Failed to detach loop devices: %m");
                }

                if (need_dm_detach) {
                        log_info("Detaching DM devices.");
                        r = dm_detach_all(&changed, umount_log_level);
                        if (r == 0) {
                                need_dm_detach = false;
                                log_info("All DM devices detached.");
                        } else if (r > 0)
                                log_info("Not all DM devices detached, %d left.", r);
                        else
                                log_error_errno(r, "Failed to detach DM devices: %m");
                }

                if (!need_umount && !need_swapoff && !need_loop_detach && !need_dm_detach) {
                        log_info("All filesystems, swaps, loop devices and DM devices detached.");
                        /* Yay, done */
                        break;
                }

                if (!changed && umount_log_level == LOG_INFO && !can_initrd) {
                        /* There are things we cannot get rid of. Loop one more time
                         * with LOG_ERR to inform the user. Note that we don't need
                         * to do this if there is a initrd to switch to, because that
                         * one is likely to get rid of the remounting mounts. If not,
                         * it will log about them. */
                        umount_log_level = LOG_ERR;
                        continue;
                }

                /* If in this iteration we didn't manage to
                 * unmount/deactivate anything, we simply give up */
                if (!changed) {
                        log_info("Cannot finalize remaining%s%s%s%s continuing.",
                                 need_umount ? " file systems," : "",
                                 need_swapoff ? " swap devices," : "",
                                 need_loop_detach ? " loop devices," : "",
                                 need_dm_detach ? " DM devices," : "");
                        break;
                }

                log_debug("Couldn't finalize remaining %s%s%s%s trying again.",
                          need_umount ? " file systems," : "",
                          need_swapoff ? " swap devices," : "",
                          need_loop_detach ? " loop devices," : "",
                          need_dm_detach ? " DM devices," : "");
        }

     (void) reboot(cmd);
        if (errno == EPERM && in_container) {
                /* If we are in a container, and we lacked
                 * CAP_SYS_BOOT just exit, this will kill our
                 * container for good. */
                log_info("Exiting container.");
                return EXIT_SUCCESS;
        }

系统调用/usr/include/x86_64-linux-gnu/sys/reboot.h

#ifndef _SYS_REBOOT_H
#define _SYS_REBOOT_H   1

#include <features.h>

/* Perform a hard reset now.  */
#define RB_AUTOBOOT     0x01234567

/* Halt the system.  */
#define RB_HALT_SYSTEM  0xcdef0123

/* Enable reboot using Ctrl-Alt-Delete keystroke.  */
#define RB_ENABLE_CAD   0x89abcdef

/* Disable reboot using Ctrl-Alt-Delete keystroke.  */
#define RB_DISABLE_CAD  0

/* Stop system and switch power off if possible.  */
#define RB_POWER_OFF    0x4321fedc

/* Suspend system using software suspend.  */
#define RB_SW_SUSPEND   0xd000fce2

/* Reboot system into new kernel.  */
#define RB_KEXEC        0x45584543

__BEGIN_DECLS

/* Reboot or halt the system.  */
extern int reboot (int __howto) __THROW;

__END_DECLS

#endif  /* _SYS_REBOOT_H */

Reboot系统调用的实现位于“kernel/reboot.c”，函数如下：

/*
 * Reboot system call: for obvious reasons only root may call it,
 * and even root needs to set up some magic numbers in the registers
 * so that some mistake won't make this reboot the whole machine.
 * You can also set the meaning of the ctrl-alt-del-key here.
 *
 * reboot doesn't sync: do that yourself before calling this.
 */
SYSCALL_DEFINE4(reboot, int, magic1, int, magic2, unsigned int, cmd,
		void __user *, arg)
{
	struct pid_namespace *pid_ns = task_active_pid_ns(current);
	char buffer[256];
	int ret = 0;

	/* We only trust the superuser with rebooting the system. */
	if (!ns_capable(pid_ns->user_ns, CAP_SYS_BOOT))
		return -EPERM;

	/* For safety, we require "magic" arguments. */
	if (magic1 != LINUX_REBOOT_MAGIC1 ||
			(magic2 != LINUX_REBOOT_MAGIC2 &&
			magic2 != LINUX_REBOOT_MAGIC2A &&
			magic2 != LINUX_REBOOT_MAGIC2B &&
			magic2 != LINUX_REBOOT_MAGIC2C))
		return -EINVAL;

	/*
	 * If pid namespaces are enabled and the current task is in a child
	 * pid_namespace, the command is handled by reboot_pid_ns() which will
	 * call do_exit().
	 */
	ret = reboot_pid_ns(pid_ns, cmd);
	if (ret)
		return ret;

	/* Instead of trying to make the power_off code look like
	 * halt when pm_power_off is not set do it the easy way.
	 */
	if ((cmd == LINUX_REBOOT_CMD_POWER_OFF) && !pm_power_off)
		cmd = LINUX_REBOOT_CMD_HALT;

	mutex_lock(&reboot_mutex);
	switch (cmd) {
	case LINUX_REBOOT_CMD_RESTART:
		kernel_restart(NULL);
		break;

	case LINUX_REBOOT_CMD_CAD_ON:
		C_A_D = 1;
		break;

	case LINUX_REBOOT_CMD_CAD_OFF:
		C_A_D = 0;
		break;

	case LINUX_REBOOT_CMD_HALT:
		kernel_halt();
		do_exit(0);
		panic("cannot halt");

	case LINUX_REBOOT_CMD_POWER_OFF:
		kernel_power_off();
		do_exit(0);
		break;

	case LINUX_REBOOT_CMD_RESTART2:
		ret = strncpy_from_user(&buffer[0], arg, sizeof(buffer) - 1);
		if (ret < 0) {
			ret = -EFAULT;
			break;
		}
		buffer[sizeof(buffer) - 1] = '\0';

		kernel_restart(buffer);
		break;

#ifdef CONFIG_KEXEC_CORE
	case LINUX_REBOOT_CMD_KEXEC:
		ret = kernel_kexec();
		break;
#endif

#ifdef CONFIG_HIBERNATION
	case LINUX_REBOOT_CMD_SW_SUSPEND:
		ret = hibernate();
		break;
#endif

	default:
		ret = -EINVAL;
		break;
	}
	mutex_unlock(&reboot_mutex);
	return ret;
}

reboot该系统调用的名称。
magic1、magic2两个int类型的“魔力数”，用于防止误操作。
arg其它的额外参数。
函数内部流程：

1. 判断调用者的用户权限，如果不是超级用户（superuser），则直接返回错误（这也是我们再用户空间执行reboot、halt、poweroff等命令时，必须是root用户的原因。
2. 判断传入的magicnumber是否匹配，如果不匹配，直接返回错误。这样就可以尽可能的防止误动作发生
3. 调用reboot_pid_ns接口，检查是否需要由该接口处理reboot请求。
4. 如果是POWER_OFF命令，且没有注册poweroff的machine处理函数（pm_power_off），把该命令转换为HALT命令。
5. 根据具体的cmd命令，执行具体的处理，包括，如果是RESTART或者RESTART2命令，调用kernel_restart。如果是CAD_ON或CAD_OFF命令，更新C_A_D的值，表示是否允许通过Ctrl+Alt+Del组合键重启系统。如果是HALT命令，调用kernel_halt。如果是POWER_OFF命令，调用kernel_power_off。如果是KEXEC命令，调用kernel_kexec接口。如果是SW_SUSPEND，调用hibernate接口。
6. 返回上述的处理结果，系统调用结束。