Android (Linux) Suspend流程

1. Linux Suspend简介

    Linux Suspend主要有以下三步：

   1) 冻结用户态进程和内核态任务
   2) 调用注册的设备的suspend的回调函数，顺序是按照注册顺序
   3) 休眠核心设备和使CPU进入休眠态。
   冻结进程(suspend_freeze_processes)是内核把进程列表中所有的进程的状态都设置为停止，并且保存所有进程的上下文。 当这些进程被解冻(suspend_thaw_processes)的时候，他们是不知道自己被冻结过的，只是简单的继续执行。如何让Linux进入Suspend呢？用户可以通过读写sys文件/sys /power/state 是实现控制系统进入休眠，比如：

   # echo standby > /sys/power/state

2. Suspend流程

    Suspend主要流程如下图所示：

 

3. enter_state(PM_SUSPEND_MEM)

      其主要功能如下：

     1) suspend_prepare:   准备进入suspend，并冻结所有进程

     2) suspend_devices_and_enter: suspend所有外设，并进入sleep状态，只有当唤醒时，此函数才返回

     3) suspend_finish: suspend结束，并被唤醒

       

     enter_state代码如下：  

 

// kernel/kernel/power/suspend.c
int enter_state(suspend_state_t state)
{
    int error;
    if (!valid_state(state))
        return -ENODEV;

    if (!mutex_trylock(&pm_mutex))
        return -EBUSY;

#ifdef CONFIG_SUSPEND_SYNC_WORKQUEUE
    suspend_sys_sync_queue();
#else
    sys_sync();
    printk("done.\n");
#endif

    pr_debug("PM: Preparing system for %s sleep\n", pm_states[state]);

    error = suspend_prepare(); //准备进入suspend,并冻结所有进程
    if (error)
        goto Unlock;

    if (suspend_test(TEST_FREEZER))
        goto Finish;

    pr_debug("PM: Entering %s sleep\n", pm_states[state]);
    pm_restrict_gfp_mask();
    error = suspend_devices_and_enter(state); // suspend外部设备
    pm_restore_gfp_mask();

 Finish:
    pr_debug("PM: Finishing wakeup.\n");
    suspend_finish();  // 结束suspend,并被唤醒
 Unlock:
    mutex_unlock(&pm_mutex);
    return error;
}

// kernel/kernel/power/suspend.c
int enter_state(suspend_state_t state)
{
	int error;
	if (!valid_state(state))
		return -ENODEV;

	if (!mutex_trylock(&pm_mutex))
		return -EBUSY;

#ifdef CONFIG_SUSPEND_SYNC_WORKQUEUE
	suspend_sys_sync_queue();
#else
	sys_sync();
	printk("done.\n");
#endif

	pr_debug("PM: Preparing system for %s sleep\n", pm_states[state]);

	error = suspend_prepare(); //准备进入suspend,并冻结所有进程
	if (error)
		goto Unlock;

	if (suspend_test(TEST_FREEZER))
		goto Finish;

	pr_debug("PM: Entering %s sleep\n", pm_states[state]);
	pm_restrict_gfp_mask();
	error = suspend_devices_and_enter(state); // suspend外部设备
	pm_restore_gfp_mask();

 Finish:
	pr_debug("PM: Finishing wakeup.\n");
	suspend_finish();  // 结束suspend,并被唤醒
 Unlock:
	mutex_unlock(&pm_mutex);
	return error;
}

3.1 准备并冻结进程(suspend_prepare)

      在suspend_prepare()中它将完成以下任务：

      1) 给suspend分配一个虚拟终端来输出信息；

      2) 然后广播一个系统要进入suspend的Notify；

      3) 关闭掉用户态的helper进程；

      4) 最后调用suspend_freeze_processes()冻结所有的进程，这里将保存所有进程 当前的状态，也许有一些进程会拒绝进入冻结状态，当有这样的进程存在的时候，会导致冻结失败，此函数就会放弃冻结进程，并且解冻刚才冻结的所有进程。

      其详细代码如下：

static int suspend_prepare(void)
{
    int error;

    if (!suspend_ops || !suspend_ops->enter)
        return -EPERM;

    pm_prepare_console();  // 分配一个console

    error = pm_notifier_call_chain(PM_SUSPEND_PREPARE); // 发送suspend notify
    if (error)
        goto Finish;

    error = usermodehelper_disable(); // disable用户态的helper进程
    if (error)
        goto Finish;

    error = suspend_freeze_processes(); // 冻结所有进程
    if (!error)
        return 0;

    suspend_thaw_processes();
    usermodehelper_enable();
 Finish:
    pm_notifier_call_chain(PM_POST_SUSPEND);
    pm_restore_console();
    return error;
}

static int suspend_prepare(void)
{
	int error;

	if (!suspend_ops || !suspend_ops->enter)
		return -EPERM;

	pm_prepare_console();  // 分配一个console

	error = pm_notifier_call_chain(PM_SUSPEND_PREPARE); // 发送suspend notify
	if (error)
		goto Finish;

	error = usermodehelper_disable(); // disable用户态的helper进程
	if (error)
		goto Finish;

	error = suspend_freeze_processes(); // 冻结所有进程
	if (!error)
		return 0;

	suspend_thaw_processes();
	usermodehelper_enable();
 Finish:
	pm_notifier_call_chain(PM_POST_SUSPEND);
	pm_restore_console();
	return error;
}

3.2 Suspend外部设备(suspend_devices_and_enter)

       现在， 所有的进程(也包括workqueue/kthread) 都已经停止了，内核态进程有可能在停止的时候握有一些信号量， 所以如果这时候在外设里面去解锁这个信号量有可能会发生死锁， 所以在外设的suspend()函数里面进行lock/unlock锁要非常小心，建议设计时不要在suspend()里面等待锁。而且因为suspend的时候，有一些Log是无法输出的，所以一旦出现问题，非常难调试。
       suspend_devices_and_enter的主要功能为：

       1) suspend_console: Suspend console子系统，即printk将不能打印了

       2) dpm_suspend_start: Suspend所有非系统设备，即调用所有注册设备的suspend回调函数

       3) suspend_enter: 使系统进入要求的sleep状态，然后停在这儿，只有当系统被中断或者其他事件唤醒时，此函数才返回

           以下函数只有当wakeup时才被执行：

       4) dpm_resume_end:  resume所有非系统设备，即执行所有注册设备的resume回调函数

       5) resume_console: resume console子系统，即printk可用了

       详细代码如下所示：

            kernel/kernel/power/suspend.c

int suspend_devices_and_enter(suspend_state_t state)
{
    int error;
        /* suspend_pos通过suspend_set_ops来进行注册，
           它在kernel/arch/arm/mach-xx/pm.c中定义，其函数名
           可能为xx_pm_ops,例子如下：
       static struct platform_suspend_ops rk30_pm_ops = {
        .enter      = xx_pm_enter,
        .valid      = suspend_valid_only_mem,
        .prepare    = xx_pm_prepare,
        .finish     = xx_pm_finish,
       };
    */

    if (!suspend_ops)
        return -ENOSYS;

    trace_machine_suspend(state);
    if (suspend_ops->begin) {
        error = suspend_ops->begin(state);
        if (error)
            goto Close;
    }
    suspend_console(); // suspend console子系统，printk将不能打印了
    suspend_test_start();
    error = dpm_suspend_start(PMSG_SUSPEND);  // suspend所有非系统设备
                                                  // 即执行所有设备的suspend回调函数
    if (error) {
        printk(KERN_ERR "PM: Some devices failed to suspend\n");
        goto Recover_platform;
    }
    suspend_test_finish("suspend devices");
    if (suspend_test(TEST_DEVICES))
        goto Recover_platform;

    error = suspend_enter(state); // 系统进入要求的sleep状态，
                                      // 只有当wakeup时，此函数才返回

 Resume_devices:
    suspend_test_start();
    dpm_resume_end(PMSG_RESUME);  // resume所有非系统设备
                                      // 即执行所有设备的resume回调函数
    suspend_test_finish("resume devices");
    resume_console();             // resume console子系统，即printk可用了
 Close:
    if (suspend_ops->end)
        suspend_ops->end();
    trace_machine_suspend(PWR_EVENT_EXIT);
    return error;

 Recover_platform:
    if (suspend_ops->recover)
        suspend_ops->recover();
    goto Resume_devices;
}

int suspend_devices_and_enter(suspend_state_t state)
{
	int error;
        /* suspend_pos通过suspend_set_ops来进行注册，
           它在kernel/arch/arm/mach-xx/pm.c中定义，其函数名
           可能为xx_pm_ops,例子如下：
	   static struct platform_suspend_ops rk30_pm_ops = {
		.enter		= xx_pm_enter,
		.valid		= suspend_valid_only_mem,
		.prepare 	= xx_pm_prepare,
		.finish		= xx_pm_finish,
	   };
	*/

	if (!suspend_ops)
		return -ENOSYS;

	trace_machine_suspend(state);
	if (suspend_ops->begin) {
		error = suspend_ops->begin(state);
		if (error)
			goto Close;
	}
	suspend_console(); // suspend console子系统，printk将不能打印了
	suspend_test_start();
	error = dpm_suspend_start(PMSG_SUSPEND);  // suspend所有非系统设备
                                                  // 即执行所有设备的suspend回调函数
	if (error) {
		printk(KERN_ERR "PM: Some devices failed to suspend\n");
		goto Recover_platform;
	}
	suspend_test_finish("suspend devices");
	if (suspend_test(TEST_DEVICES))
		goto Recover_platform;

	error = suspend_enter(state); // 系统进入要求的sleep状态，
                                      // 只有当wakeup时，此函数才返回

 Resume_devices:
	suspend_test_start();
	dpm_resume_end(PMSG_RESUME);  // resume所有非系统设备
                                      // 即执行所有设备的resume回调函数
	suspend_test_finish("resume devices");
	resume_console();             // resume console子系统，即printk可用了
 Close:
	if (suspend_ops->end)
		suspend_ops->end();
	trace_machine_suspend(PWR_EVENT_EXIT);
	return error;

 Recover_platform:
	if (suspend_ops->recover)
		suspend_ops->recover();
	goto Resume_devices;
}

 

3.2.1 suspend_console

        Suspend console子系统，即printk将不能打印了

 

void suspend_console(void)
{
    if (!console_suspend_enabled)
        return;
    printk("Suspending console(s) (use no_console_suspend to debug)\n");
    console_lock();
    console_suspended = 1;
    up(&console_sem);
}

void suspend_console(void)
{
	if (!console_suspend_enabled)
		return;
	printk("Suspending console(s) (use no_console_suspend to debug)\n");
	console_lock();
	console_suspended = 1;
	up(&console_sem);
}

 

3.2.2 dpm_suspend_start  (PMSG_SUSPEND)

        Suspend所有非系统设备，即调用所有注册设备的suspend回调函数

/**
 * dpm_suspend_start - Prepare devices for PM transition and suspend them.
 * @state: PM transition of the system being carried out.
 *
 * Prepare all non-sysdev devices for system PM transition and execute "suspend"
 * callbacks for them.
 */
int dpm_suspend_start(pm_message_t state)
{
    int error;

    error = dpm_prepare(state);  // 根据dpm_list生成dpm_prepared_list
    if (!error)
        error = dpm_suspend(state); //根据dpm_prepared_list生成dpm_suspended_list
    return error;
}

/**
 * dpm_suspend_start - Prepare devices for PM transition and suspend them.
 * @state: PM transition of the system being carried out.
 *
 * Prepare all non-sysdev devices for system PM transition and execute "suspend"
 * callbacks for them.
 */
int dpm_suspend_start(pm_message_t state)
{
	int error;

	error = dpm_prepare(state);  // 根据dpm_list生成dpm_prepared_list
	if (!error)
		error = dpm_suspend(state); //根据dpm_prepared_list生成dpm_suspended_list
	return error;
}

  

3.2.3 suspend_enter 

  

    使系统进入要求的sleep状态，然后停在这儿，只有当系统被中断或者其他事件唤醒时，此函数才返回，其详细代码如下：

/**
 *  suspend_enter - enter the desired system sleep state.
 *  @state:     state to enter
 *
 *  This function should be called after devices have been suspended.
 */
static int suspend_enter(suspend_state_t state)
{
    int error;

    if (suspend_ops->prepare) {
        error = suspend_ops->prepare(); //即执行xx_pm_prepare
        if (error)
            goto Platform_finish;
    }

    error = dpm_suspend_noirq(PMSG_SUSPEND); //使所有外设驱动不再接收中断
    if (error) {
        printk(KERN_ERR "PM: Some devices failed to power down\n");
        goto Platform_finish;
    }

    if (suspend_ops->prepare_late) {
        error = suspend_ops->prepare_late();
        if (error)
            goto Platform_wake;
    }

    if (suspend_test(TEST_PLATFORM))
        goto Platform_wake;

    error = disable_nonboot_cpus(); // 停止非启动CPU
    if (error || suspend_test(TEST_CPUS))
        goto Enable_cpus;

    arch_suspend_disable_irqs(); // 关闭中断
    BUG_ON(!irqs_disabled());

    error = syscore_suspend();  // 执行注册在syscore_ops_list的syscore_ops的suspend函数
    if (!error) {
        if (!(suspend_test(TEST_CORE) || pm_wakeup_pending())) {
            error = suspend_ops->enter(state); // KEY: 即执行xx_pm_enter，唤醒时才返回
            events_check_enabled = false;
        }
        syscore_resume(); // 执行注册在syscore_ops_list的syscore_ops的resume函数
    }

    arch_suspend_enable_irqs(); // 打开中断
    BUG_ON(irqs_disabled());

 Enable_cpus:
    enable_nonboot_cpus(); // 启动非启动CPU

 Platform_wake:
    if (suspend_ops->wake)
        suspend_ops->wake();

    dpm_resume_noirq(PMSG_RESUME); //使所有外设驱动接收中断

 Platform_finish:
    if (suspend_ops->finish)
        suspend_ops->finish(); //即执行xx_pm_finish

    return error;
}

/**
 *	suspend_enter - enter the desired system sleep state.
 *	@state:		state to enter
 *
 *	This function should be called after devices have been suspended.
 */
static int suspend_enter(suspend_state_t state)
{
	int error;

	if (suspend_ops->prepare) {
		error = suspend_ops->prepare(); //即执行xx_pm_prepare
		if (error)
			goto Platform_finish;
	}

	error = dpm_suspend_noirq(PMSG_SUSPEND); //使所有外设驱动不再接收中断
	if (error) {
		printk(KERN_ERR "PM: Some devices failed to power down\n");
		goto Platform_finish;
	}

	if (suspend_ops->prepare_late) {
		error = suspend_ops->prepare_late();
		if (error)
			goto Platform_wake;
	}

	if (suspend_test(TEST_PLATFORM))
		goto Platform_wake;

	error = disable_nonboot_cpus(); // 停止非启动CPU
	if (error || suspend_test(TEST_CPUS))
		goto Enable_cpus;

	arch_suspend_disable_irqs(); // 关闭中断
	BUG_ON(!irqs_disabled());

	error = syscore_suspend();  // 执行注册在syscore_ops_list的syscore_ops的suspend函数
	if (!error) {
		if (!(suspend_test(TEST_CORE) || pm_wakeup_pending())) {
			error = suspend_ops->enter(state); // KEY: 即执行xx_pm_enter，唤醒时才返回
			events_check_enabled = false;
		}
		syscore_resume(); // 执行注册在syscore_ops_list的syscore_ops的resume函数
	}

	arch_suspend_enable_irqs(); // 打开中断
	BUG_ON(irqs_disabled());

 Enable_cpus:
	enable_nonboot_cpus(); // 启动非启动CPU

 Platform_wake:
	if (suspend_ops->wake)
		suspend_ops->wake();

	dpm_resume_noirq(PMSG_RESUME); //使所有外设驱动接收中断

 Platform_finish:
	if (suspend_ops->finish)
		suspend_ops->finish(); //即执行xx_pm_finish

	return error;
}

  

3.2.4 dpm_resume_end  (PMSG_RESUME)

        resume所有非系统设备，即执行所有注册设备的resume回调函数

/**
 * dpm_resume_end - Execute "resume" callbacks and complete system transition.
 * @state: PM transition of the system being carried out.
 *
 * Execute "resume" callbacks for all devices and complete the PM transition of
 * the system.
 */
void dpm_resume_end(pm_message_t state)
{
    dpm_resume(state); //根据dpm_suspended_list生成dpm_prepared_list
    dpm_complete(state); //根据dpm_prepared_list生成dpm_list
}

/**
 * dpm_resume_end - Execute "resume" callbacks and complete system transition.
 * @state: PM transition of the system being carried out.
 *
 * Execute "resume" callbacks for all devices and complete the PM transition of
 * the system.
 */
void dpm_resume_end(pm_message_t state)
{
	dpm_resume(state); //根据dpm_suspended_list生成dpm_prepared_list
	dpm_complete(state); //根据dpm_prepared_list生成dpm_list
}

 

3.2.5 resume_console

         resume console子系统，即printk可用了

void resume_console(void)
{
    if (!console_suspend_enabled)
        return;
    down(&console_sem);
    console_suspended = 0;
    console_unlock();
}

void resume_console(void)
{
	if (!console_suspend_enabled)
		return;
	down(&console_sem);
	console_suspended = 0;
	console_unlock();
}

 

3.3 Suspend结束(suspend_finish)

    其主要功能如下(它是suspend_prepare的逆过程)：

      1) 解冻所有进程；

      2) 打开用户态helper进程；

      3) 广播系系统suspend结束的Notify；

      4) 释放分配的虚拟终端。

     其详细代码如下：
 

static void suspend_finish(void)
{
    suspend_thaw_processes();  //解冻所有进程
    usermodehelper_enable();   // 打开用户态helper进程
    pm_notifier_call_chain(PM_POST_SUSPEND); // 广播系系统suspend结束的Notify
    pm_restore_console();  // 释放分配的虚拟终端
}

static void suspend_finish(void)
{
	suspend_thaw_processes();  //解冻所有进程
	usermodehelper_enable();   // 打开用户态helper进程
	pm_notifier_call_chain(PM_POST_SUSPEND); // 广播系系统suspend结束的Notify
	pm_restore_console();  // 释放分配的虚拟终端
}