1. 概述
代码路径
framework/base/core/java/andorid/os/
- Handler.java
- Looper.java
- Message.java
- MessageQueue.java
1.1 架构介绍:
消息机制主要包含:
- Message:消息分为硬件产生的消息(如按钮、触摸)和软件生成的消息;
-
MessageQueue:消息队列的主要功能向消息池投递消息(
MessageQueue.enqueueMessage
)和取走消息池的消息(MessageQueue.next
); -
Handler:消息辅助类,主要功能向消息池发送各种消息事件(
Handler.sendMessage
)和处理相应消息事件(Handler.handleMessage
); -
Looper:不断循环执行(
Looper.loop
),按分发机制将消息分发给目标处理者
1.2 应用示例
class LooperThread extends Thread {
public Handler mHandler;
public void run() {
Looper.prepare();
mHandler = new Handler() {
public void handleMessage(Message msg) {
//TODO 定义消息处理逻辑.
}
};
Looper.loop();
}
}
1.3 ThreadLocal
ThreadLocal: 线程本地存储区(Thread Local Storage,简称为TLS)
每个线程都有自己的私有的本地存储区域,不同线程之间彼此不能访问对方的TLS区域。TLS常用的操作方法:
-
ThreadLocal.set(T value)
:将value存储到当前线程的TLS区域,源码如下:
public void set(T value) {
Thread currentThread = Thread.currentThread(); //获取当前线程
Values values = values(currentThread); //查找当前线程的本地储存区
if (values == null) {
//当线程本地存储区,尚未存储该线程相关信息时,则创建Values对象
values = initializeValues(currentThread);
}
//保存数据value到当前线程this
values.put(this, value);
}
-
ThreadLocal.get()
:获取当前线程TLS区域的数据,源码如下:
public T get() {
Thread currentThread = Thread.currentThread(); //获取当前线程
Values values = values(currentThread); //查找当前线程的本地储存区
if (values != null) {
Object[] table = values.table;
int index = hash & values.mask;
if (this.reference == table[index]) {
return (T) table[index + 1]; //返回当前线程储存区中的数据
}
} else {
//创建Values对象
values = initializeValues(currentThread);
}
return (T) values.getAfterMiss(this); //从目标线程存储区没有查询是则返回null
}
ThreadLocal的get()和set()方法操作的类型都是泛型,接着回到前面提到的sThreadLocal
变量,其定义如下:
static final ThreadLocal<Looper> sThreadLocal = new ThreadLocal<Looper>()
可见sThreadLocal
的get()和set()操作的类型都是Looper
类型。
1.4 Message
1.4.1 消息对象
每个消息用Message
表示,Message
主要包含以下内容:
数据类型 | 成员变量 | 解释 |
---|---|---|
int | what | 消息类别 |
long | when | 消息触发时间 |
int | arg1 | 参数1 |
int | arg2 | 参数2 |
Object | obj | 消息内容 |
Handler | target | 消息响应方 |
Runnable | callback | 回调方法 |
创建消息的过程,就是填充消息的上述内容的一项或多项。
1.4.2 消息池
在代码中,可能经常看到recycle()方法,咋一看,可能是在做虚拟机的gc()相关的工作,其实不然,这是用于把消息加入到消息池的作用。这样的好处是,当消息池不为空时,可以直接从消息池中获取Message对象,而不是直接创建,提高效率。
静态变量sPool
的数据类型为Message,通过next成员变量,维护一个消息池;静态变量MAX_POOL_SIZE
代表消息池的可用大小;消息池的默认大小为50。
消息池常用的操作方法是obtain()和recycle()。
public final Message obtainMessage() {
return Message.obtain(this);
}
obtain(),从消息池取Message,都是把消息池表头的Message取走,再把表头指向next;
public static Message obtain() {
synchronized (sPoolSync) {
if (sPool != null) {
Message m = sPool;
sPool = m.next;
m.next = null; //从sPool中取出一个Message对象,并消息链表断开
m.flags = 0; // 清除in-use flag
sPoolSize--; //消息池的可用大小进行减1操作
return m;
}
}
return new Message(); // 当消息池为空时,直接创建Message对象
}
recycle
recycle(),将Message加入到消息池的过程,都是把Message加到链表的表头;
public void recycle() {
if (isInUse()) { //判断消息是否正在使用
if (gCheckRecycle) { //Android 5.0以后的版本默认为true,之前的版本默认为false.
throw new IllegalStateException("This message cannot be recycled because it is still in use.");
}
return;
}
recycleUnchecked();
}
//对于不再使用的消息,加入到消息池
void recycleUnchecked() {
//将消息标示位置为IN_USE,并清空消息所有的参数。
flags = FLAG_IN_USE;
what = 0;
arg1 = 0;
arg2 = 0;
obj = null;
replyTo = null;
sendingUid = -1;
when = 0;
target = null;
callback = null;
data = null;
synchronized (sPoolSync) {
if (sPoolSize < MAX_POOL_SIZE) { //当消息池没有满时,将Message对象加入消息池
next = sPool;
sPool = this;
sPoolSize++; //消息池的可用大小进行加1操作
}
}
}
2. 消息队列的创建
-
可以在子线程创建handler么?
-
主线程 Looper 和子线程的 Looper 有什么区别?
-
Handler、 Looper 和 MessageQueue有什么关系?
-
MessageQueue是怎么创建的?
2.1 Handler构造函数
一个例子:
new Thread() {
@Override
public void run() {
new Handler();
}
}.start();
//子线程创建Handler 会抛异常
//java.lang.RuntimeException: Can't create handler inside thread that has not called Looper.prepare
上面的例子中,子线程中创建Handler,抛异常
源码分析:
public Handler() {
this(null, false);
}
public Handler(Callback callback, boolean async) {
//匿名类、内部类或本地类都必须申明为static,否则会警告可能出现内存泄露
if (FIND_POTENTIAL_LEAKS) {
final Class<? extends Handler> klass = getClass();
if ((klass.isAnonymousClass() || klass.isMemberClass() || klass.isLocalClass()) &&
(klass.getModifiers() & Modifier.STATIC) == 0) {
Log.w(TAG, "The following Handler class should be static or leaks might occur: " +
klass.getCanonicalName());
}
}
//必须先执行Looper.prepare(),才能获取Looper对象,否则为null.
mLooper = Looper.myLooper(); //从当前线程的TLS中获取Looper对象
//获取不到mLooper对象,抛出异常(必须先调用Looper.prepare())
if (mLooper == null) {
throw new RuntimeException(
"Can't create handler inside thread " + Thread.currentThread()
+ " that has not called Looper.prepare()");
}
//消息队列,来自Looper对象
mQueue = mLooper.mQueue;
mCallback = callback; //回调方法
mAsynchronous = async; //设置消息是否为异步处理方式
}
2.2 Looper 对象的创建和获取
2.2.1 Looper 创建
Looper.prepare()每个线程只允许执行一次
//对于无参的情况,默认调用 prepare(true),表示这个Looper允许退出
//prepare(false),表示不允许退出
public static void prepare() {
prepare(true);
}
private static void prepare(boolean quitAllowed) {
//每个线程只允许执行一次该方法,第二次执行时线程的TLS已有数据,则会抛出异常
if (sThreadLocal.get() != null) {
throw new RuntimeException("Only one Looper may be created per thread");
}
//创建Looper对象,并保存到当前线程的TLS区域
sThreadLocal.set(new Looper(quitAllowed));
}
与prepare() 功能相近,该方法主要在ActivityThread类中使用。
public static void prepareMainLooper() {
prepare(false); //设置不允许退出的Looper
synchronized (Looper.class) {
//将当前的Looper保存为主Looper,每个线程只允许执行一次。
if (sMainLooper != null) {
throw new IllegalStateException("The main Looper has already been prepared.");
}
sMainLooper = myLooper();
}
}
/**
* Return the Looper object associated with the current thread. Returns
* null if the calling thread is not associated with a Looper.
*/
public static @Nullable Looper myLooper() {
return sThreadLocal.get();
}
2.2.2 Looper 构造函数
private Looper(boolean quitAllowed) {
//创建MessageQueue对象
mQueue = new MessageQueue(quitAllowed);
//记录当前线程.
mThread = Thread.currentThread();
}
2.3 MessageQueue 对象的创建
2.3.1 java层
MessageQueue(boolean quitAllowed) {
mQuitAllowed = quitAllowed;
//通过native方法初始化消息队列,其中mPtr是供native代码使用
mPtr = nativeInit();
}
2.3.2 native层调用过程
1. android_os_MessageQueue_nativeInit()
android_os_MessageQueue.cpp
static jlong android_os_MessageQueue_nativeInit(JNIEnv* env, jclass clazz) {
//初始化 NativeMessageQueue
NativeMessageQueue* nativeMessageQueue = new NativeMessageQueue();
if (!nativeMessageQueue) {
jniThrowRuntimeException(env, "Unable to allocate native queue");
return 0;
}
//增加引用计数
nativeMessageQueue->incStrong(env);
return reinterpret_cast<jlong>(nativeMessageQueue);
}
2. NativeMessageQueue()
class NativeMessageQueue : public MessageQueue, public LooperCallback {
public:
NativeMessageQueue();
virtual ~NativeMessageQueue();
virtual void raiseException(JNIEnv* env, const char* msg, jthrowable exceptionObj);
void pollOnce(JNIEnv* env, jobject obj, int timeoutMillis);
void wake();
void setFileDescriptorEvents(int fd, int events);
virtual int handleEvent(int fd, int events, void* data);
private:
JNIEnv* mPollEnv;
jobject mPollObj;
jthrowable mExceptionObj;
};
NativeMessageQueue::NativeMessageQueue() :
mPollEnv(NULL), mPollObj(NULL), mExceptionObj(NULL) {
//功能类比于Java层的Looper.myLooper(); 获取TLS中的Looper对象
mLooper = Looper::getForThread();
if (mLooper == NULL) {
//创建native层的Looper
mLooper = new Looper(false);
//功能类比于Java层的ThreadLocal.set(); 保存native层的Looper到TLS
Looper::setForThread(mLooper);
}
}
3. Looper()
Looper::Looper(bool allowNonCallbacks) :
mAllowNonCallbacks(allowNonCallbacks), mSendingMessage(false),
mPolling(false), mEpollFd(-1), mEpollRebuildRequired(false),
mNextRequestSeq(0), mResponseIndex(0), mNextMessageUptime(LLONG_MAX) {
//构造唤醒事件的fd
mWakeEventFd = eventfd(0, EFD_NONBLOCK | EFD_CLOEXEC);
LOG_ALWAYS_FATAL_IF(mWakeEventFd < 0, "Could not make wake event fd: %s",
strerror(errno));
AutoMutex _l(mLock);
//重建Epoll事件
rebuildEpollLocked();
}
Looper对象中的mWakeEventFd添加到epoll监控,以及mRequests也添加到epoll的监控范围内。
void Looper::rebuildEpollLocked() {
// Close old epoll instance if we have one.
if (mEpollFd >= 0) {
//关闭旧的epoll实例
close(mEpollFd);
}
// Allocate the new epoll instance and register the wake pipe.
mEpollFd = epoll_create(EPOLL_SIZE_HINT);
struct epoll_event eventItem;
//把未使用的数据区域进行置0操作
memset(& eventItem, 0, sizeof(epoll_event)); // zero out unused members of data field union
eventItem.events = EPOLLIN;//可读事件
eventItem.data.fd = mWakeEventFd;
//将唤醒事件(mWakeEventFd)添加到epoll实例(mEpollFd)
int result = epoll_ctl(mEpollFd, EPOLL_CTL_ADD, mWakeEventFd, & eventItem);
for (size_t i = 0; i < mRequests.size(); i++) {
const Request& request = mRequests.valueAt(i);
struct epoll_event eventItem;
request.initEventItem(&eventItem);
//将request队列的事件,分别添加到epoll实例
int epollResult = epoll_ctl(mEpollFd, EPOLL_CTL_ADD, request.fd, & eventItem);
if (epollResult < 0) {
ALOGE("Error adding epoll events for fd %d while rebuilding epoll set: %s",
request.fd, strerror(errno));
}
}
}
3. 消息传递机制
消息是怎么发送的?
- Handler.sendMessage
消息循环过程是怎样的?
- Looper.loop
消息是怎么处理的?
- HandlerDispatchMessage
3.1 消息的发送
Handler.java
public final boolean sendMessageDelayed(Message msg, long delayMillis) {
if (delayMillis < 0) {
delayMillis = 0;
}
//SystemClock.uptimeMillis() + delayMillis 当前时间加上需要延迟的时间
return sendMessageAtTime(msg, SystemClock.uptimeMillis() + delayMillis);
}
public boolean sendMessageAtTime(Message msg, long uptimeMillis) {
MessageQueue queue = mQueue;
if (queue == null) {
return false;
}
return enqueueMessage(queue, msg, uptimeMillis);
}
该方法通过设置消息的触发时间为0,从而使Message加入到消息队列的队头。
public final boolean sendMessageAtFrontOfQueue(Message msg) {
//Handler初始化的时候 mQueue = mLooper.mQueue;
MessageQueue queue = mQueue;
if (queue == null) {
RuntimeException e = new RuntimeException(
this + " sendMessageAtTime() called with no mQueue");
Log.w("Looper", e.getMessage(), e);
return false;
}
return enqueueMessage(queue, msg, 0);
}
3.1.1 java层:enqueueMessage
Handler.equeueMessage
handler发送消息最终都会调用该方法
private boolean enqueueMessage(MessageQueue queue, Message msg, long uptimeMillis) {
//设置处理消息的target为当前发送消息的handler
msg.target = this;
//调用默认构造方法 mAsynchronous = false; 默认消息非异步
if (mAsynchronous) {
msg.setAsynchronous(true);
}
return queue.enqueueMessage(msg, uptimeMillis);
}
MessageQueue.enqueueMessage
boolean enqueueMessage(Message msg, long when) {
// 每一个普通Message必须有一个target
if (msg.target == null) {
throw new IllegalArgumentException("Message must have a target.");
}
if (msg.isInUse()) {
throw new IllegalStateException(msg + " This message is already in use.");
}
synchronized (this) {
//正在退出时,直接返回
if (mQuitting) {
msg.recycle();
return false;
}
//标记msg正在使用中
msg.markInUse();
msg.when = when;
//mMessages为链表的头结点
Message p = mMessages;
boolean needWake;
//p==null 说明当前消息队列为空
//调用sendMessageAtFrontOfQueue发消息时when==0
//when < p.when 表示新来的消息比消息队列头部的消息要早
if (p == null || when == 0 || when < p.when) {
//当前消息msg放到链表的头结点
msg.next = p;
mMessages = msg;
needWake = mBlocked; //当阻塞时需要唤醒
} else {
//如果这个消息不能插入到链表的头结点,接下来找一个合适的位置插进去
//将消息按时间顺序插入到MessageQueue。
//需要唤醒情况:阻塞了 && 此时消息对列头节点是barrier消息 && 新来的消息是异步消息
needWake = mBlocked && p.target == null && msg.isAsynchronous();
Message prev;
//遍历单链表找到新来消息的合适的位置
for (;;) {
prev = p;
p = p.next;
if (p == null || when < p.when) {
break;
}
//新消息未找到合适位置之前,遍历过程中发现异步消息&&需要唤醒时,needWake设置为false
if (needWake && p.isAsynchronous()) {
needWake = false;
}
}
//找到合适位置插入链表中
msg.next = p;
prev.next = msg;
}
//需要唤醒,进入nativeWake流程
if (needWake) {
nativeWake(mPtr);
}
}
return true;
}
3.1.2 native层:nativeWake(mPtr)
【1】android_os_MessageQueue_nativeWake()
==> android_os_MessageQueue.cpp
static void android_os_MessageQueue_nativeWake(JNIEnv* env, jclass clazz, jlong ptr) {
NativeMessageQueue* nativeMessageQueue = reinterpret_cast<NativeMessageQueue*>(ptr);
nativeMessageQueue->wake(); 【3】
}
【2】NativeMessageQueue::wake()
==> android_os_MessageQueue.cpp
void NativeMessageQueue::wake() {
mLooper->wake(); 【4】
}
【3】Looper::wake()
==> Looper.cpp
void Looper::wake() {
uint64_t inc = 1;
// 向管道mWakeEventFd写入字符1
ssize_t nWrite = TEMP_FAILURE_RETRY(write(mWakeEventFd, &inc, sizeof(uint64_t)));
if (nWrite != sizeof(uint64_t)) {
if (errno != EAGAIN) {
ALOGW("Could not write wake signal, errno=%d", errno);
}
}
}
其中TEMP_FAILURE_RETRY
是一个宏定义, 当执行write
失败后,会不断重复执行,直到执行成功为止。
3.2 消息的循环
消息循环的重点:
-
取消息 queue.next()
-
分发消息 msg.target.dispatchMessage(msg);
-
分发后的Message回收到消息池,以便重复利用
3.2.1 Looper.loop()
public static void loop() {
final Looper me = myLooper(); //获取TLS存储的Looper对象
if (me == null) {
throw new RuntimeException("No Looper; Looper.prepare() wasn't called on this thread.");
}
//获取Looper对象中的消息队列
final MessageQueue queue = me.mQueue;
Binder.clearCallingIdentity();
//确保在权限检查时基于本地进程,而不是调用进程。
final long ident = Binder.clearCallingIdentity();
//进入loop的主循环方法
for (;;) {
//取下一条消息, 可能会阻塞
Message msg = queue.next();
//msg == null 表示Looper结束了,直接退出循环
if (msg == null) {
return;
}
...
try {
//用于分发Message
msg.target.dispatchMessage(msg);
} finally {
...
}
...
//恢复调用者信息
final long newIdent = Binder.clearCallingIdentity();
...
//将Message放入消息池 重置一些状态,放入链表中
msg.recycleUnchecked();
}
}
3.2.2 java层 : queue.next()
取下一条消息
Message next() {
final long ptr = mPtr;
if (ptr == 0) { //当消息循环已经退出,则直接返回
return null;
}
int pendingIdleHandlerCount = -1;
//初始值设为0 ,表示立即返回
int nextPollTimeoutMillis = 0;
for (;;) {
if (nextPollTimeoutMillis != 0) {
Binder.flushPendingCommands();
}
//阻塞操作,当等待nextPollTimeoutMillis时长,或者消息队列被唤醒,都会返回
//nextPollTimeoutMillis = -1 表示一直阻塞,直到有消息
//第一次调用时nextPollTimeoutMillis = 0 不会阻塞,直接返回
nativePollOnce(ptr, nextPollTimeoutMillis);
synchronized (this) {
final long now = SystemClock.uptimeMillis();
Message prevMsg = null;
//从链表的头部取一条消息
Message msg = mMessages;
//如果第一条消息就是屏障,就往后遍历,看看有没有异步消息
//如果没有,就休眠,等待别人唤醒
//如果有,就看离这个消息出发时间还有多久,设置一个超时,继续休眠
if (msg != null && msg.target == null) {
//当查询到异步消息,则立刻退出循环
do {
prevMsg = msg;
msg = msg.next;
} while (msg != null && !msg.isAsynchronous());
}
//拿到一条有用的消息
if (msg != null) {
//当消息触发时间大于当前时间,则该消息触发时间还没到,重新设置超时时间
if (now < msg.when) {
nextPollTimeoutMillis = (int) Math.min(msg.when - now, Integer.MAX_VALUE);
} else {
// 获取一条消息,并返回
mBlocked = false;
if (prevMsg != null) {
prevMsg.next = msg.next;
} else {
mMessages = msg.next;
}
msg.next = null;
if (DEBUG) Log.v(TAG, "Returning message: " + msg);
msg.markInUse(); //设置消息的使用状态,即flags |= FLAG_IN_USE
//成功地获取MessageQueue中的下一条即将要执行的消息
return msg;
}
} else {
// msg==null 没有消息的时候,设置该值为-1,下次循环时,会一直阻塞直到有消息
nextPollTimeoutMillis = -1;
}
//消息正在退出,返回null
if (mQuitting) {
dispose();
return null;
}
//当消息队列为空 msg == null,或者是消息队列的第一个消息还没到时间
if (pendingIdleHandlerCount < 0
&& (mMessages == null || now < mMessages.when)) {
//获取IdleHandlers的个数
pendingIdleHandlerCount = mIdleHandlers.size();
}
if (pendingIdleHandlerCount <= 0) {
//没有idle handlers 需要运行,则跳出本次循环
mBlocked = true;
continue;
}
//有需要运行的 IdleHandler
if (mPendingIdleHandlers == null) {
mPendingIdleHandlers = new IdleHandler[Math.max(pendingIdleHandlerCount, 4)];
}
mPendingIdleHandlers = mIdleHandlers.toArray(mPendingIdleHandlers);
}
//只有第一次循环时,会运行idleHandlers,执行完成后,重置pendingIdleHandlerCount为0.
for (int i = 0; i < pendingIdleHandlerCount; i++) {
final IdleHandler idler = mPendingIdleHandlers[i];
mPendingIdleHandlers[i] = null;
boolean keep = false;
try {
keep = idler.queueIdle(); //idle时执行的方法
} catch (Throwable t) {
Log.wtf(TAG, "IdleHandler threw exception", t);
}
//idler.queueIdle() 返回 false, 移除 IdleHandler
if (!keep) {
synchronized (this) {
mIdleHandlers.remove(idler);
}
}
}
//重置idle handler个数为0,以保证不会再次重复运行
pendingIdleHandlerCount = 0;
//当调用一个空闲handler时,给nextPollTimeoutMillis设为0,下次循环无需等待直接返回
nextPollTimeoutMillis = 0;
}
}
3.2.3 native层:nativePollOnce
代码路径
frameworks\base\core\jni\android_os_MessageQueue.cpp
system\core\libutils\Looper.cpp
frameworks\base\core\jni\android_os_MessageQueue.h
system\core\include\utils\Looper.h
【1】android_os_MessageQueue_nativePollOnce()
==> android_os_MessageQueue.cpp
//初次循环的 timeoutMillis == 0
static void android_os_MessageQueue_nativePollOnce(JNIEnv* env, jobject obj, jlong ptr, jint timeoutMillis) {
//先将java层传递下来的mPtr转换为nativeMessageQueue
NativeMessageQueue* nativeMessageQueue = reinterpret_cast<NativeMessageQueue*>(ptr);
nativeMessageQueue->pollOnce(env, obj, timeoutMillis);
}
【2】NativeMessageQueue::pollOnce()
==> android_os_MessageQueue.cpp
void NativeMessageQueue::pollOnce(JNIEnv* env, jobject pollObj, int timeoutMillis) {
mPollEnv = env;
mPollObj = pollObj;
//调用到Looper->pollOnce(timeoutMillis)
mLooper->pollOnce(timeoutMillis);
mPollObj = NULL;
mPollEnv = NULL;
if (mExceptionObj) {
env->Throw(mExceptionObj);
env->DeleteLocalRef(mExceptionObj);
mExceptionObj = NULL;
}
}
【3】Looper::pollOnce()
==> Looper.h
inline int pollOnce(int timeoutMillis) {
return pollOnce(timeoutMillis, NULL, NULL, NULL);
}
【4】 Looper::pollOnce()
==> Looper.cpp
/**
*- timeoutMillis:超时时长
*- outFd:发生事件的文件描述符
*- outEvents:当前outFd上发生的事件,包含以下4类事件
* - EVENT_INPUT 可读
* - EVENT_OUTPUT 可写
* - EVENT_ERROR 错误
* - EVENT_HANGUP 中断
*- outData:上下文数据
*/
int Looper::pollOnce(int timeoutMillis, int* outFd, int* outEvents, void** outData) {
int result = 0;
for (;;) {
// 先处理没有Callback方法的 Response事件
while (mResponseIndex < mResponses.size()) {
const Response& response = mResponses.itemAt(mResponseIndex++);
int ident = response.request.ident;
if (ident >= 0) { //ident大于0,则表示没有callback, 因为POLL_CALLBACK = -2,
int fd = response.request.fd;
int events = response.events;
void* data = response.request.data;
if (outFd != NULL) *outFd = fd;
if (outEvents != NULL) *outEvents = events;
if (outData != NULL) *outData = data;
return ident;
}
}
if (result != 0) {
if (outFd != NULL) *outFd = 0;
if (outEvents != NULL) *outEvents = 0;
if (outData != NULL) *outData = NULL;
return result;
}
// 再处理内部轮询
result = pollInner(timeoutMillis); 【5】
}
}
【5】Looper::pollInner()
==> Looper.cpp
pollOnce返回值说明:
- POLL_WAKE: 表示由wake()触发,即pipe写端的write事件触发;
- POLL_CALLBACK: 表示某个被监听fd被触发。
- POLL_TIMEOUT: 表示等待超时;
- POLL_ERROR:表示等待期间发生错误;
int Looper::pollInner(int timeoutMillis) {
...
int result = POLL_WAKE;
mResponses.clear();
mResponseIndex = 0;
mPolling = true; //即将处于idle状态
struct epoll_event eventItems[EPOLL_MAX_EVENTS]; //fd最大个数为16
//等待事件发生或者超时,在nativeWake()方法,向管道写端写入字符,则该方法会返回;
//出错eventCount = -1 超时 eventCount = 0
int eventCount = epoll_wait(mEpollFd, eventItems, EPOLL_MAX_EVENTS, timeoutMillis);
mPolling = false; //不再处于idle状态
mLock.lock(); //请求锁
if (mEpollRebuildRequired) {
mEpollRebuildRequired = false;
rebuildEpollLocked(); // epoll重建,直接跳转Done;
goto Done;
}
// epoll事件个数小于0,发生错误,直接跳转Done;
if (eventCount < 0) {
if (errno == EINTR) {
goto Done;
}
result = POLL_ERROR;
goto Done;
}
//epoll事件个数等于0,发生超时,直接跳转Done;
if (eventCount == 0) {
result = POLL_TIMEOUT;
goto Done;
}
//循环遍历,处理所有的事件
for (int i = 0; i < eventCount; i++) {
int fd = eventItems[i].data.fd;
uint32_t epollEvents = eventItems[i].events;
if (fd == mWakeEventFd) {
if (epollEvents & EPOLLIN) {
//处理这个事件,即读事件
awoken(); //已经唤醒了,则读取并清空管道数据
}
} else {
ssize_t requestIndex = mRequests.indexOfKey(fd);
if (requestIndex >= 0) {
int events = 0;
if (epollEvents & EPOLLIN) events |= EVENT_INPUT;
if (epollEvents & EPOLLOUT) events |= EVENT_OUTPUT;
if (epollEvents & EPOLLERR) events |= EVENT_ERROR;
if (epollEvents & EPOLLHUP) events |= EVENT_HANGUP;
//处理request,生成对应的reponse对象,push到响应数组
pushResponse(events, mRequests.valueAt(requestIndex));
}
}
}
Done: ;
//再处理Native的Message,调用相应回调方法
mNextMessageUptime = LLONG_MAX;
while (mMessageEnvelopes.size() != 0) {
nsecs_t now = systemTime(SYSTEM_TIME_MONOTONIC);
const MessageEnvelope& messageEnvelope = mMessageEnvelopes.itemAt(0);
if (messageEnvelope.uptime <= now) {
{
sp<MessageHandler> handler = messageEnvelope.handler;
Message message = messageEnvelope.message;
mMessageEnvelopes.removeAt(0);
mSendingMessage = true;
mLock.unlock(); //释放锁
handler->handleMessage(message); // 处理消息事件
}
mLock.lock(); //请求锁
mSendingMessage = false;
result = POLL_CALLBACK; // 发生回调
} else {
mNextMessageUptime = messageEnvelope.uptime;
break;
}
}
mLock.unlock(); //释放锁
//处理带有Callback()方法的Response事件,执行Reponse相应的回调方法
for (size_t i = 0; i < mResponses.size(); i++) {
Response& response = mResponses.editItemAt(i);
if (response.request.ident == POLL_CALLBACK) {
int fd = response.request.fd;
int events = response.events;
void* data = response.request.data;
// 处理请求的回调方法
int callbackResult = response.request.callback->handleEvent(fd, events, data);
if (callbackResult == 0) {
removeFd(fd, response.request.seq); //移除fd
}
response.request.callback.clear(); //清除reponse引用的回调方法
result = POLL_CALLBACK; // 发生回调
}
}
return result;
}
【6】Looper::awoken()
void Looper::awoken() {
uint64_t counter;
//不断读取管道数据,目的就是为了清空管道内容
TEMP_FAILURE_RETRY(read(mWakeEventFd, &counter, sizeof(uint64_t)));
}
poll小结
pollInner()方法的处理流程:
- 先调用epoll_wait(),这是阻塞方法,用于等待事件发生或者超时;
- 对于epoll_wait()返回,当且仅当以下3种情况出现:
- POLL_ERROR,发生错误,直接跳转到Done;
- POLL_TIMEOUT,发生超时,直接跳转到Done;
- 检测到管道有事件发生,则再根据情况做相应处理:
- 如果是管道读端产生事件,则直接读取管道的数据;
- 如果是其他事件,则处理request,生成对应的reponse对象,push到reponse数组;
- 进入Done标记位的代码段:
- 先处理Native的Message,调用Native 的Handler来处理该Message;
- 再处理Response数组,POLL_CALLBACK类型的事件;
从上面的流程,可以发现对于Request先收集,一并放入reponse数组,而不是马上执行。真正在Done开始执行的时候,是先处理native Message,再处理Request,说明native Message的优先级高于Request请求的优先级。
另外pollOnce()方法中,先处理Response数组中不带Callback的事件,再调用了pollInner()方法。
3.3 消息的处理
public void dispatchMessage(Message msg) {
if (msg.callback != null) {
//当Message存在回调方法,回调msg.callback.run()方法;
handleCallback(msg);
} else {
if (mCallback != null) {
//当Handler存在Callback成员变量时,回调方法handleMessage();
if (mCallback.handleMessage(msg)) {
return;
}
}
//Handler自身的回调方法handleMessage()
handleMessage(msg);
}
}
public interface Callback {
public boolean handleMessage(Message msg);
}
Handler中使用 post提交: callback = r
public final boolean post(Runnable r) {
return sendMessageDelayed(getPostMessage(r), 0);
}
private static Message getPostMessage(Runnable r) {
Message m = Message.obtain();
m.callback = r;
return m;
}
消息分发的优先级:
- Message的回调方法:
message.callback.run()
,优先级最高; - Handler的回调方法:
Handler.mCallback.handleMessage(msg)
,优先级仅次于1; - Handler的默认方法:
Handler.handleMessage(msg)
,优先级最低。
消息缓存:
为了提供效率,提供了一个大小为50的Message缓存队列,减少对象不断创建与销毁的过程
3.4 消息的移除
public final void removeMessages(int what) {
mQueue.removeMessages(this, what, null);
}
void removeMessages(Handler h, int what, Object object) {
if (h == null) {
return;
}
synchronized (this) {
Message p = mMessages;
//从消息队列的头部开始,移除所有符合条件的消息
while (p != null && p.target == h && p.what == what
&& (object == null || p.obj == object)) {
Message n = p.next;
mMessages = n;
p.recycleUnchecked();
p = n;
}
//移除剩余的符合要求的消息
while (p != null) {
Message n = p.next;
if (n != null) {
if (n.target == h && n.what == what
&& (object == null || n.obj == object)) {
Message nn = n.next;
n.recycleUnchecked();
p.next = nn;
continue;
}
}
p = n;
}
}
}
4. 消息的延时机制
-
handler 消息延时机制是如何实现的?
消息队列按消息触发时间顺序排序
-
消息延时做了什么特殊处理?
设置epoll_wait 的超时时间,使其在特定时间唤醒
-
是发送延时,还是消息处理延时?
-
延时精度怎么样?
精度不高,有可能有些消息的处理比较耗时
public final boolean sendMessageDelayed(Message msg, long delayMillis) {
if (delayMillis < 0) {
delayMillis = 0;
}
return sendMessageAtTime(msg, SystemClock.uptimeMillis() + delayMillis);
}
public boolean sendMessageAtTime(Message msg, long uptimeMillis) {
MessageQueue queue = mQueue;
if (queue == null) {
return false;
}
return enqueueMessage(queue, msg, uptimeMillis);
}
private boolean enqueueMessage(MessageQueue queue, Message msg, long uptimeMillis) {
//设置处理消息的target为当前发送消息的handler
msg.target = this;
if (mAsynchronous) {
msg.setAsynchronous(true);
}
return queue.enqueueMessage(msg, uptimeMillis);
}
接下来的分析可以看 3. 消息传递机制-消息的发送
5. IdleHandler
App启动优化涉及到
了解IdleHandler 的作用以及调用方式
了解IdleHandler 有哪些使用场景
熟悉IdleHandler 的实现原理
5.1 IdleHandler原理
MessageQueue.java
/**
* Callback interface for discovering when a thread is going to block
* waiting for more messages.
*/
public static interface IdleHandler {
/**
* Called when the message queue has run out of messages and will now
* wait for more. Return true to keep your idle handler active, false
* to have it removed. This may be called if there are still messages
* pending in the queue, but they are all scheduled to be dispatched
* after the current time.
*/
boolean queueIdle();
}
Looper.myQueue().addIdleHandler(new MessageQueue.IdleHandler() {
@Override
public boolean queueIdle() {
//true false 区别
return false;
}
});
/**
* Add a new {@link IdleHandler} to this message queue. This may be
* removed automatically for you by returning false from
* {@link IdleHandler#queueIdle IdleHandler.queueIdle()} when it is
* invoked, or explicitly removing it with {@link #removeIdleHandler}.
*
* <p>This method is safe to call from any thread.
*
* @param handler The IdleHandler to be added.
*/
public void addIdleHandler(@NonNull IdleHandler handler) {
if (handler == null) {
throw new NullPointerException("Can't add a null IdleHandler");
}
synchronized (this) {
mIdleHandlers.add(handler);
}
}
Message next() {
...
for (;;) {
nativePollOnce(ptr, nextPollTimeoutMillis);
synchronized (this) {
...
// If first time idle, then get the number of idlers to run.
// Idle handles only run if the queue is empty or if the first message
// in the queue (possibly a barrier) is due to be handled in the future.
//看消息列表是否有消息可以分发,如果有,就返回该消息
//走到这里说明,没有消息可以分发,下一个for循环就要进入休眠了
if (pendingIdleHandlerCount < 0
&& (mMessages == null || now < mMessages.when)) {
pendingIdleHandlerCount = mIdleHandlers.size();
}
if (pendingIdleHandlerCount <= 0) {
// No idle handlers to run. Loop and wait some more.
mBlocked = true;
continue;
}
if (mPendingIdleHandlers == null) {
mPendingIdleHandlers = new IdleHandler[Math.max(pendingIdleHandlerCount, 4)];
}
mPendingIdleHandlers = mIdleHandlers.toArray(mPendingIdleHandlers);
}
// Run the idle handlers.
// We only ever reach this code block during the first iteration.
for (int i = 0; i < pendingIdleHandlerCount; i++) {
final IdleHandler idler = mPendingIdleHandlers[i];
mPendingIdleHandlers[i] = null; // release the reference to the handler
boolean keep = false;
try {
keep = idler.queueIdle();
} catch (Throwable t) {
Log.wtf(TAG, "IdleHandler threw exception", t);
}
//如果keep == flase 需要remove掉 idler
if (!keep) {
synchronized (this) {
mIdleHandlers.remove(idler);
}
}
}
// Reset the idle handler count to 0 so we do not run them again.
pendingIdleHandlerCount = 0;
// While calling an idle handler, a new message could have been delivered
// so go back and look again for a pending message without waiting.
nextPollTimeoutMillis = 0;
}
}
5.2 IdleHandler 在Framework中的使用
void scheduleGcIdler() {
if (!mGcIdleScheduled) {
mGcIdleScheduled = true;
Looper.myQueue().addIdleHandler(mGcIdler);
}
mH.removeMessage(H.GC_WHEN_IDLE);
}
final class GcIdler implements MessageQueue.IdleHandler{
@Override
public boolean queueIdle() {
//实际上调用BinderInternal.forceGc("bg");
doGcIfNeeded();
return false;
}
}
一个例子:等待线程 idle 触发回调
public void waitForIdle(Runnable recipient) {
mMessageQueue.addIdleHandler(new Idler(recipient));
mThread.getHandler().post(new EmptyRunnable());
}
private static final class Idler implements MessageQueueIdleHandler{
public final boolean queueIdle() {
if (mCallback != null) {
mCallback.run();
}
//返回false 表示回调是一次性的
return false;
}
}
5.3 使用场景
- 延时执行
@Override
protected void onCreate(Bundle savedInstanceState) {
super.onCreate(savedInstanceState);
setContentView(R.layout.activity_main);
mHandler.postDelayed(new Runnable() {
@Override
public void run() {
//执行一些耗时任务,但是要延时多久呢?
//可以写一个IdleHandler 在主线程空闲的时候执行耗时任务,return false
doSomething();
}
}, 1000);
}
- 批量任务
任务密集
只关注最终结果
开一个工作线程,每一个任务都封装成一个消息,丢到工作线程中。等工作线程空闲下来之后,汇总消息,刷新界面
6. 主线程进入Looper为什么没有ANR
- 了解ANR触发的原理
- 了解应用的启动流程
- 了解线程的消息循环机制
- 了解应用和系统服务通信的过程
ANR是什么
AMS中 在SystemServer进程
final void appNotResponding() {
Message msg = Message.obtain();
msg.what = SHOW_NOT_RESPONDING_MSG;
...
//mUiHandler 是 SystemServer的一个子线程
mUiHandler.sendMessage(msg);
}
Dialog d = new AppNotResbondingDialog(...);
d.show();
ANR场景:
- Service Timeout
- BroadcastQueue Timeout
- ContentProvider Timeout
- InputDispatching Timeout
以service为例,看看ANR如何触发
void realStartServiceLocked(ServiceRecord r, ProcessRecord app, boolean execInFg) {
...
bumpServiceExecutingLocked(r, execInFg, "create");
app.thread.scheduleCreateService(r, ...)
}
void bumpServiceExecutingLocked(ServiceRecord r, boolean fg, String why) {
...
//启动一个超时机制,如果应用端没有在规定时间启动一个 Service ,就会ANR
scheduleServiceTimeoutLocked(r.app);
}
void scheduleServiceTimeoutLocked(ProcessRecord proc) {
long now = SystemClock.uptimeMillis();
Message msg = mAm.mHandler.obtainMessage(SERVICE_TIMEOUT_MSG);
mAm.mHandler.sendMessageAtTime(msg, now + SERVICE_TIMEOUT);
}
void serviceTimeout(ProcessRecord proc) {
...
mAm.appNotResponding(...);
}
应用端收到系统服务发过来的启动 service 服务之后
private void handleCreateService(CreateServiceData data) {
service = (Service)cl.loadClass(data.info.name).newInstance();
ContextImpl context = ContextImpl.createAppContext(this, packageInfo);
Application app = packageInfo.makeApplication(false, mInstrumentation);
service.attach(...);
service.onCreate();
ActivityManagerNative.getDefault().serviceDoneExecuting(...);
}
//跨进程调用
private void serviceDoneExecutingLocked(ServiceRecord r, ...) {
mAm.mHandler.removeMessages(SERVICE_TIMEOUT_MSG, r.app);
}
- anr 是应用没有在规定的时间内完成AMS指定的任务导致的
- AMS请求调到应用端Binder线程,再丢消息去唤醒主线程来处理
- ANR不是因为主线程loop循环,而是因为主线程中有耗时任务
7. 消息屏障
消息有三种 :
normal
barrier
async
怎么往消息队列发送消息屏障?
postSyncBarrier只对同步消息产生影响,对于异步消息没有任何差别
private int postSyncBarrier(long when) {
synchronized(this) {
final int token = mNextBarrierToken++;
final Message msg = Message.obtain();
//没有target 不需要分发,可以根据target是否为null 判断是否是消息屏障
msg.makeInUse();
msg.when = when;
msg.arg1 = token;
//给这个msg按时间顺序插到消息队列
return token;
}
}
-
消息屏障只会影响它后面的消息,它前面的消息不受影响
-
消息队列可以插入多个消息屏障
-
消息屏障插到消息队列没有唤醒线程
-
插入消息屏障会返回一个 token(屏障的一个序列号) 凭借这个token在消息队列中查找消息屏障,然后移除它
-
只能通过反射调用 postSyncBarrier
删除屏障
public void removeSyncBarrier(int token) {
synchronized (this) {
Message prev = null;
Message p = mMessages;
//从消息队列找到 target为空,并且token相等的Message
while (p != null && (p.target != null || p.arg1 != token)) {
prev = p;
p = p.next;
}
final boolean needWake;
if (prev != null) {
prev.next = p.next;
needWake = false;
} else {
mMessages = p.next;
needWake = mMessages == null || mMessages.target != null;
}
p.recycleUnchecked();
//如果这个线程就是因为这个屏障block住,那么撤除屏障之后,需要唤醒线程
if (needWake && !mQuitting) {
nativeWake(mPtr);
}
}
}
网友评论