您的位置:首页 > 移动开发 > Android开发

Android休眠唤醒机制简介(二)

2014-02-22 16:27 459 查看
******************************************************************

作者:sean

日期:2012-11-29

修改历史:

******************************************************************
接上一节,结合code来分析一下:

具体流程

下面我将分别以两条路线(第一:获得wakelock唤醒锁。第二:系统进入睡眠。)来分别说明各自的流程,让读者对android睡眠唤醒机制有更深入的理解!

第一部分:获得wakelock唤醒锁

比如在应用程序中,当获得wakelock唤醒锁的时候,它首先是调用/android/frameworks/base/core/java/

android/os/PowerManager类中的public void acquire()方法,而该方法通过android特有的通讯机制,会接着调用到PowerManagerService类中的public void acquireWakeLock。

public void acquireWakeLock(int flags, IBinder lock, String tag, WorkSource ws) {
int uid = Binder.getCallingUid();
int pid = Binder.getCallingPid();
if (uid != Process.myUid()) {
mContext.enforceCallingOrSelfPermission(android.Manifest.permission.WAKE_LOCK, null);
}
if (ws != null) {
enforceWakeSourcePermission(uid, pid);
}
long ident = Binder.clearCallingIdentity();
try {
synchronized (mLocks) {
acquireWakeLockLocked(flags, lock, uid, pid, tag, ws);
}

} finally {
Binder.restoreCallingIdentity(ident);
}
}
而 public void acquireWakeLock方法又调用了acquireWakeLockLocked。

public void acquireWakeLockLocked(int flags, IBinder lock, int uid, int pid, String tag,
WorkSource ws)
{
if (mSpew) {
Slog.d(TAG, "acquireWakeLock flags=0x" + Integer.toHexString(flags) + " tag=" + tag); }
if (ws != null && ws.size() == 0) {ws = null;}
int index = mLocks.getIndex(lock);
WakeLock wl;
boolean newlock;
boolean diffsource;
WorkSource oldsource;
.
中间代码省略
.
Power.acquireWakeLock(Power.PARTIAL_WAKE_LOCK,PARTIAL_NAME);
}
if (diffsource) {
// If the lock sources have changed, need to first release the
// old ones.
noteStopWakeLocked(wl, oldsource);
}
if (newlock || diffsource) {
noteStartWakeLocked(wl, ws);
}
}

我们可以看到在acquireWakeLockLocked 方法调用Power类中的public static native void acquireWakeLock(int lock, String id)方法。而该方法是调用android_os_Power.cpp中的static void acquireWakeLock()函数。

static void acquireWakeLock(JNIEnv *env, jobject clazz, jint lock, jstring idObj)

{
if (idObj == NULL) {
throw_NullPointerException(env, "id is null");
return ;
}
const char *id = env->GetStringUTFChars(idObj, NULL);
acquire_wake_lock(lock, id);
env->ReleaseStringUTFChars(idObj, id);
}
函数 acquire_wake_lock()的实现在 power.c中,其定义如下:
int acquire_wake_lock(int lock, const char* id)
{
initialize_fds();
// LOGI("acquire_wake_lock lock=%d id='%s'\n", lock, id);
if (g_error) return g_error;
int fd;
if (lock == PARTIAL_WAKE_LOCK) {
fd = g_fds[ACQUIRE_PARTIAL_WAKE_LOCK];
}
else {
return EINVAL;
}
return write(fd, id, strlen(id));
}
到现在为止,我们的代码流程已经走了一大半了,我们一开始介绍的android的上面几层Framework层、JNI层、HAL层都已经介绍了就剩下Kernel层了。下面就应该是和kernel层进行交互了。

但是在android/hardware/libhardware_legacy/power/power.c中的acquire_wake_lock()函数似乎没法和kernel层进行通信啊??不急 要淡定!!在这个函数的最后不是还有一个返回语句return write(fd, id, strlen(id))嘛!!有人会说这句话看不出什么啊,我一开始用Source Insight代码阅读器跟踪的时候也没有找到它的原型,那个叫急啊!!呵呵 最后经过我的继续不断的努力查找(其实方法很简单,既然我从上往下的路断了,那我就换个方向,我最后又从下往上顺着代码走了一遍),终于被我发现了。

我们先看一下android/kernel/kernel/power/main.c中的一段代码,我将会做简单的分析,之后你就会明白刚才上面所产生的疑问了。

#ifdef CONFIG_USER_WAKELOCK
power_attr(wake_lock);
power_attr(wake_unlock);
#endif

static struct attribute * g[] = {
&state_attr.attr,
#ifdef CONFIG_PM_TRACE
&pm_trace_attr.attr,
#endif

#ifdef CONFIG_PM_SLEEP
&pm_async_attr.attr,
#ifdef CONFIG_PM_DEBUG
&pm_test_attr.attr,
#endif

#ifdef CONFIG_USER_WAKELOCK
&wake_lock_attr.attr,
&wake_unlock_attr.attr,
#endif

#endif
NULL,
};

static struct attribute_group attr_group = {
.attrs = g,
};

#ifdef CONFIG_PM_RUNTIME
struct workqueue_struct *pm_wq;
EXPORT_SYMBOL_GPL(pm_wq);

static int __init pm_start_workqueue(void)
{
pm_wq = create_freezeable_workqueue("pm");
return pm_wq ? 0 : -ENOMEM;
}
#else
static inline int pm_start_workqueue(void) { return 0; }
#endif
static int __init pm_init(void)
{
int error = pm_start_workqueue();
if (error)
return error;
power_kobj = kobject_create_and_add("power", NULL);
if (!power_kobj)
return -ENOMEM;
return sysfs_create_group(power_kobj, &attr_group);
}
core_initcall(pm_init);

这段代码虽然简短,但看起来是不是还是比较费劲,没关系,我们倒过来看就比较清楚了。上面代码中的sysfs_create_group(power_kobj, &attr_group);的意思就是当我们在对sysfs/下相对的节点进行操作的时候会调用与attr_group里的相关函数,再往上面看其实就是指&wake_lock_attr.attr(对不同情况的操作会调用不同的attr_group,在第二条路的里面我们还会再次接触到这里)。power_attr(wake_lock)就是使具体的操作函数与其挂钩。我们现在来看一看这个挂钩过程是怎么实现的。

#define power_attr(_name) \
static struct kobj_attribute _name##_attr = {\
.attr = { \
.name = __stringify(_name),\
.mode = 0644, \
}, \
.show = _name##_show,\
.store = _name##_store,\
}

在该函数中##的作用通俗点讲就是“连接”的意思,比如power_attr(wake_lock),
static struct kobj_attribute wake_lock_attr = {\
.attr = { \
.name = __stringify(wake_lock),\
.mode = 0644, \
}, \
.show = wake_lock_show, \
.store = wake_lock_store, \
}

函数wake_lock_store和wake_lock_show就定义在android/kernel/kernel/power/userwakelock.c
中。因此当我们对/sys/power/wake_lock进行操作的时候就会调用到userwakelock.c中定义的
wake_lock_store()函数。
好了,我们该回到原来我们产生疑问的地方了,在 power.c中我们将重新研究一下这这段代码,这时我们还得关注其中的另一个函数initialize_fds()。

int acquire_wake_lock(int lock, const char* id)
{
initialize_fds();
// LOGI("acquire_wake_lock lock=%d id='%s'\n", lock, id);
if (g_error) return g_error;
int fd;
if (lock == PARTIAL_WAKE_LOCK) {
fd = g_fds[ACQUIRE_PARTIAL_WAKE_LOCK];
}
else {
return EINVAL;
}
return write(fd, id, strlen(id));
}
initialize_fds(void)
{
// XXX: should be this:
//pthread_once(&g_initialized, open_file_descriptors);
// XXX: not this:
if (g_initialized == 0) {
if(open_file_descriptors(NEW_PATHS) < 0) {
open_file_descriptors(OLD_PATHS);
on_state = "wake";
off_state = "standby";
}
g_initialized = 1;
}
}
其实这个函数中最和新的步骤就是open_file_descriptors(NEW_PATHS) ;而
const char * const NEW_PATHS[] = {
"/sys/power/wake_lock",
"/sys/power/wake_unlock",
"/sys/power/state"
};
总之经过着一些列的步骤后,最终我们将在 return write(fd, id, strlen(id));时调用android/kernel/kernel/power/userwakelock.c 中的 wake_lock_store()函数。

ssize_t wake_lock_store(
struct kobject *kobj, struct kobj_attribute *attr,
const char *buf, size_t n)
{
long timeout;
struct user_wake_lock *l;
mutex_lock(&tree_lock);
l = lookup_wake_lock_name(buf, 1, &timeout);
if (IS_ERR(l)) {
n = PTR_ERR(l);
goto bad_name;
}
if (debug_mask & DEBUG_ACCESS)
pr_info("wake_lock_store: %s, timeout %ld\n", l->name, timeout);
if (timeout)
wake_lock_timeout(&l->wake_lock, timeout);
else
wake_lock(&l->wake_lock);
bad_name:
mutex_unlock(&tree_lock);
return n;
}
该函数执行的基本流程为:首先调用lookup_wake_lock_name()来获得指定的唤醒锁,若延迟参数timeout为零的话,就调用 wake_lock()否则就调用wake_lock_timeout(),但不管调用哪个最后都会调用到android/kernel/kernel/power/wakelock.c中的函数static void wake_lock_internal()。

static void wake_lock_internal(struct wake_lock *lock, long timeout, int has_timeout)
{
int type;
unsigned long irqflags;
long expire_in;
spin_lock_irqsave(&list_lock, irqflags);
type = lock->flags & WAKE_LOCK_TYPE_MASK;
BUG_ON(type >= WAKE_LOCK_TYPE_COUNT);
BUG_ON(!(lock->flags & WAKE_LOCK_INITIALIZED));
#ifdef CONFIG_WAKELOCK_STAT
if (type == WAKE_LOCK_SUSPEND && wait_for_wakeup) {
if (debug_mask & DEBUG_WAKEUP)
pr_info("wakeup wake lock: %s\n", lock->name);
wait_for_wakeup = 0;
lock->stat.wakeup_count++;
}

if ((lock->flags & WAKE_LOCK_AUTO_EXPIRE) &&
(long)(lock->expires - jiffies) <= 0) {
wake_unlock_stat_locked(lock, 0);
lock->stat.last_time = ktime_get();
}
#endif
if (!(lock->flags & WAKE_LOCK_ACTIVE)) {
lock->flags |= WAKE_LOCK_ACTIVE;
#ifdef CONFIG_WAKELOCK_STAT
lock->stat.last_time = ktime_get();
#endif
}
list_del(&lock->link);
if (has_timeout) {
if (debug_mask & DEBUG_WAKE_LOCK)
pr_info("wake_lock: %s, type %d, timeout %ld.lu\n",
lock->name, type, timeout / HZ,
(timeout % HZ) * MSEC_PER_SEC / HZ);
lock->expires = jiffies + timeout;
lock->flags |= WAKE_LOCK_AUTO_EXPIRE;
list_add_tail(&lock->link, &active_wake_locks[type]);
} else {
if (debug_mask & DEBUG_WAKE_LOCK)
pr_info("wake_lock: %s, type %d\n", lock->name, type);
lock->expires = LONG_MAX;
lock->flags &= ~WAKE_LOCK_AUTO_EXPIRE;
list_add(&lock->link, &active_wake_locks[type]);
}
if (type == WAKE_LOCK_SUSPEND) {
current_event_num++;
#ifdef CONFIG_WAKELOCK_STAT
if (lock == &main_wake_lock)
update_sleep_wait_stats_locked(1);
else if (!wake_lock_active(&main_wake_lock))
update_sleep_wait_stats_locked(0);
#endif
if (has_timeout)
expire_in = has_wake_lock_locked(type);
else
expire_in = -1;
if (expire_in > 0) {
if (debug_mask & DEBUG_EXPIRE)
pr_info("wake_lock: %s, start expire timer, "
"%ld\n", lock->name, expire_in);
mod_timer(&expire_timer, jiffies + expire_in);
} else {
if (del_timer(&expire_timer))
if (debug_mask & DEBUG_EXPIRE)
pr_info("wake_lock: %s, stop expire timer\n",
lock->name);
if (expire_in == 0)
queue_work(suspend_work_queue, &suspend_work);
}
}
spin_unlock_irqrestore(&list_lock, irqflags);
}

到这里为止,我们走的第一条路就到目的地了,这个函数具体做了什么,在这里就不仔细分析了,大家可以自己再跟下或者上网查相关资料,理解这个函数不难。

第二部分:系统进入睡眠

有了上面第一部分的学习,再看第二部分的话,会容易很多。假如现在我们按了PAD上的power睡眠键,经过一些列的事件处理后,它会调用到PowerManager类中的

public void goToSleep(long time)
{
try {
mService.goToSleep(time);
} catch (RemoteException e) {
}
}
而该函数会调用到PowerManagerService类中的public void goToSleep()方法;
public void goToSleep(long time)
{
goToSleepWithReason(time, WindowManagerPolicy.OFF_BECAUSE_OF_USER);
}
goToSleepWithReason()会调用goToSleepLocked()方法,接着会调用setPowerState();而setPowerState()方法里会调用setScreenStateLocked(),setScreenStateLocked()又会调用到Power类中的JNI接口setScreenState(),其具体实现是在android_os_Power.cpp文件中;
static int setScreenState(JNIEnv *env, jobject clazz, jboolean on)
{
return set_screen_state(on);
}

函数中return set_screen_state()的实现是android/hardware/libhardware_legacy/power/power.c
set_screen_state(int on)
{
QEMU_FALLBACK(set_screen_state(on));
LOGI("*** set_screen_state %d", on);
initialize_fds();
//LOGI("go_to_sleep eventTime=%lld now=%lld g_error=%s\n", eventTime,
// systemTime(), strerror(g_error));
if (g_error) return g_error;
char buf[32];
int len;
if(on)
len = snprintf(buf, sizeof(buf), "%s", on_state);
else
len = snprintf(buf, sizeof(buf), "%s", off_state);
buf[sizeof(buf) - 1] = '\0';
len = write(g_fds[REQUEST_STATE], buf, len);
if(len < 0) {
LOGE("Failed setting last user activity: g_error=%d\n", g_error);
}
return 0;

看!!代码到这里是不是跟第一部分很相似?不错,如果接着往下分析的话,可以套用上面第一部分的分析思路,最终len = write(g_fds[REQUEST_STATE], buf, len);语句调用的是android//kernel/kernel/power/main.c中的set_screen_state( );

当我们在sys/power/state(android/hardware/libhardware_legacy/power/power.c)进行读写操作的时候,(linux/kernel/power/main.c)中的state_store()函数会被调用,在该函数中会分成两个分支:

static ssize_t state_store(struct kobject *kobj, struct kobj_attribute *attr, const char *buf, size_t n)
{
#ifdef CONFIG_SUSPEND
#ifdef CONFIG_EARLYSUSPEND
suspend_state_t state = PM_SUSPEND_ON;
#else
suspend_state_t state = PM_SUSPEND_STANDBY;
#endif
const char * const *s;
#endif
char *p;
int len;
int error = -EINVAL;
p = memchr(buf, '\n', n);
len = p ? p - buf : n;
if (len == 4 && !strncmp(buf, "disk", len)) {
error = hibernate();
goto Exit;
}
#ifdef CONFIG_SUSPEND
for (s = &pm_states[state]; state < PM_SUSPEND_MAX; s++, state++) {
if (*s && len == strlen(*s) && !strncmp(buf, *s, len))
break;
}
if (state < PM_SUSPEND_MAX && *s)
#ifdef CONFIG_EARLYSUSPEND
if (state == PM_SUSPEND_ON || valid_state(state)) {
error = 0;
request_suspend_state(state);
}
#else
error = enter_state(state);
#endif
#endif
Exit:
return error ? error : n;
}
Android特有的earlysuspend: request_suspend_state(state)
Linux标准的suspend: enter_state(state)

注意:如果CONFIG_EARLYSUSPEND宏开的话,kernel会先走earlysuspend,反之则直接走suspend;从这里开始就要分两个分支了,如果支持earlysuspend的话就进入 request_suspend_state(state)函数,如果不支持的话就进入标准Linux的enter_state(state)函数。、

这两个函数分别在两个文件中kernel/kernel/power/earlysuspend.c和suspend.c。现在再回过头来看的话,感觉整个android中睡眠唤醒机制还是很清晰的。这两个函数体里又做了什么,在这里就不再做具体分析,大家可以自己对照代码或者上网查资料,因为本文的主旨是带读者从最上层应用层一直到最底层kernel层,把整个android的睡眠唤醒机制给走通。

PowerManager.java goToSleep( )
PowerManagerService.java goToSleep()
PowerManagerService.java goToSleepWithReason()
PowerManagerService.java setPowerState()
PowerManagerService.java SetScreenStateLocked ()
Power.java setScreenState()
android_os_Power.cpp setScreenState()
power.c set_screen_state( )
main.c state_store( )
内容来自用户分享和网络整理,不保证内容的准确性,如有侵权内容,可联系管理员处理 点击这里给我发消息
标签: