Kotlin协程的线程调度示例详解

在我们分析了协程启动创建过程启动过程后，们将着重剖析协程中协程调度的逻辑流程。

主要是分析解答如下2个问题：

• 涉及到协程方法器是如何将协程代码调度到特定的线程执行？
• 子协程执行完又是如何切换0回父协程的线程环境？

一、协程的分发器作用

1.1 测试代码

GlobalScope.launch {
//协程体1
Log.d(TAG, "before suspend job.")
withContext(Dispatchers.Main) {
//协程体2
Log.d(TAG, "print in Main thread.")
}
Log.d(TAG, "after suspend job.")
}

• 此次的协程测试用例中，我们默认的launch一个协程，我们简单的将launch需要执行的这外层逻辑为协程体1。
• 在协程体1中，我们使用withContext将协程切换到主线程执行，打印日志。我们将这里面执行的协程逻辑为协程体2。
• 协程体2执行完成后，切回协程体1中执行并打印Log。
• 注意，根据我们之前《协程的创建与启动》文章中分析的，Kotlin编译器针对协程体1和协程体2分别生成一个继承与SuspenLamabda的类型，比如：class MainActivity#onCreate$1 : SuspenLambda{...}。我们在讲协程体时，也同时代指这个类实例。

继续跟踪launch()函数执行逻辑，这次跟踪过程不同与《协程的创建与启动》篇章，我们会将侧重点放在启动过程中协程调度器是如何起作用的？接下来见1.2

1.2 CoroutineScope.launch

public fun CoroutineScope.launch(
context: CoroutineContext = EmptyCoroutineContext,
start: CoroutineStart = CoroutineStart.DEFAULT,
block: suspend CoroutineScope.() -> Unit
): Job {
//1. 见1.2.1
val newContext = newCoroutineContext(context)
val coroutine = if (start.isLazy)
LazyStandaloneCoroutine(newContext, block) else
StandaloneCoroutine(newContext, active = true)
//2. 详见1.3
coroutine.start(start, coroutine, block)
return coroutine
}

• 这里会新建一个CoroutineContext，详见1.2.1
• 根据之前的分析，这个里最终会调用到startCoroutineCancellable()方法，详见1.3流程。

1.2.1 newCoroutineContext

public actual fun CoroutineScope.newCoroutineContext(context: CoroutineContext): CoroutineContext {
val combined = foldCopies(coroutineContext, context, true)
val debug = if (DEBUG) combined + CoroutineId(COROUTINE_ID.incrementAndGet()) else combined
return 
if (combined !== Dispatchers.Default && combined[ContinuationInterceptor] == null)
debug + Dispatchers.Default
else 
	debug
}

coroutineContext：coroutineContext是CoroutineScope的成员变量，当此时为GlobalScope.coroutineContext==EmptyCoroutineContext

context：由于调用launch时没有指定Context，所以传到此处也是EmptyCoroutineContext。foldCopies()函数将2个context相加并拷贝，最终combied==EmptyCoroutineContext。

而在return这最后判断返回的是debug+Dispatchers.Defatult，所以此时默认的分发器为Dispatchers.Defatult。

这里涉及到的协程Context运算不做深入剖析，简单可以认为协程重写了“+”运算，使得Context之间可以使用“+”来叠加，没有的Element类型会被添加到Element集合，集合中已有的Element类型会被覆盖。

1.3 startCoroutineCancellable

internal fun <R, T> (suspend (R) -> T).startCoroutineCancellable(
receiver: R, completion: Continuation<T>,
onCancellation: ((cause: Throwable) -> Unit)? = null
) =
runSafely(completion) {
	//1. 创建SuspendLambda协程体
createCoroutineUnintercepted(receiver, completion)
//2. 拦截：取出分发器，并构建方法器Continuation。详见1.3.1
.intercepted()
//3. 调用方法器Continuation的resume方法，详见1.4
.resumeCancellableWith(Result.success(Unit), onCancellation)
}

• 这里的构建协程体在《协程的创建与启动》一节中已经剖析，不再赘述。
• 进行拦截，注意：这里其实会根据方法器再构建出一个DispatchedContinuation对象，它也是一个续体类型，这是对协程体的一次包装。详见1.3.1小节。
• 调用拦截器续体的resumeCancellableWith()开始状态机流转，执行分发流程详见1.4小节。

1.3.1 intercepted()

 public fun intercepted(): Continuation<Any?> =
intercepted?: (
//1. 取出拦截器
context[ContinuationInterceptor]?
//2.构建拦截器续体
.interceptContinuation(this)?: this)
.also { intercepted = it }

• 取出当前上下文中的拦截器类型，根据之前1.2.1小节的分析，这里取出来的是Dispatchers.Defatult。
• interceptContinuation(this)为构建拦截器续体，注意这里传入的this是协程体1。 详见1.3.2。

1.3.2 CoroutineDispatcher

//Base class to be extended by all coroutine dispatcher implementations.
public abstract class CoroutineDispatcher :
AbstractCoroutineContextElement(ContinuationInterceptor), ContinuationInterceptor {
public final override fun <T> interceptContinuation(continuation: Continuation<T>):
//详见1.4
Continuation<T> = DispatchedContinuation(this, continuation)
}

直接新建了一个DispatchedContinuation对象实例这里需要注意传入的构建参数：

• this：当前Dispatcher，也就是Dispatchers.Defatult。
• continuation：协程体1。

1.3.3 小结

自此Continuation.intercepted()方法就分析结束，最终的结果是：用上下文中的Dispatcher和当前Contination对象也就是协程体1，共同作为构建参数，新建了一个DispatchedContinuation对象。

接下来接着1.3中的第三点，调用DispatchedContinuation.resumeCancellableWith()方法开始分析。

1.4 DispatchedContinuation

internal class DispatchedContinuation<in T>(
//1. 分发器
@JvmField val dispatcher: CoroutineDispatcher,
	//2. 注意这里将Continuation的实现委托给了continuation成员变量。
@JvmField val continuation: Continuation<T>
) : DispatchedTask<T>(MODE_UNINITIALIZED)
, CoroutineStackFrame,
Continuation<T> by continuation {
	//3. 复写属性delegate为自己
	override val delegate: Continuation<T>
get() = this
...
// We inline it to save an entry on the stack in cases where it shows (unconfined dispatcher)
// It is used only in Continuation<T>.resumeCancellableWith
@Suppress("NOTHING_TO_INLINE")
inline fun resumeCancellableWith(
result: Result<T>,
noinline onCancellation: ((cause: Throwable) -> Unit)?
) {
val state = result.toState(onCancellation)
//默认为true
if (dispatcher.isDispatchNeeded(context)) {
_state = state
resumeMode = MODE_CANCELLABLE
//4. 详细见
dispatcher.dispatch(context, this)
} else {
executeUnconfined(state, MODE_CANCELLABLE) {
if (!resumeCancelled(state)) {
resumeUndispatchedWith(result)
}
}
}
}
}

这里的dispatcher==Dispatchers.Defatult，所以接下来需要解析Dispatchers.Defatult到底是什么东西。详见1.5

• 成员变量dispatcher==Dispatchers.Default。
• 成员变量continucation==协程体1（SuspenLambda类型实例）。同时DispatchedContinuation继承于Continuation接口，它将Continuation接口的实现委托给了成员变量continuation。
• deleagte为复写了DispatchedTask.delegate属性，将其返回自己。
• 调用分发器也就是Dispatchers.Defatult的dispatch()方法，注意这里传入的参数：

context：来自Continuation接口的属性，由于委托给了成员变量continuation，所以此context==continuation.context。

this：分发器本身Dispatchers.Defatult

自此这个方法的分析结束：调用分发器的进行分发，接下来分析就开始分析协程方法器CoroutineDispatcher

1.5 DefaultScheduler

//Dispathcer.kt
@JvmStatic
public actual val Default: CoroutineDispatcher = DefaultScheduler
//Dispathcer.kt
// Instance of Dispatchers.Default
internal object DefaultScheduler : SchedulerCoroutineDispatcher(
CORE_POOL_SIZE, MAX_POOL_SIZE,
IDLE_WORKER_KEEP_ALIVE_NS, DEFAULT_SCHEDULER_NAME
) {
...
}

实际上是继承 SchedulerCoroutineDispatcher类型。详见1.5.1

1.5.1 SchedulerCoroutineDispatcher

internal open class SchedulerCoroutineDispatcher(
private val corePoolSize: Int = CORE_POOL_SIZE,
private val maxPoolSize: Int = MAX_POOL_SIZE,
private val idleWorkerKeepAliveNs: Long = IDLE_WORKER_KEEP_ALIVE_NS,
private val schedulerName: String = "CoroutineScheduler",
) : ExecutorCoroutineDispatcher() {
override val executor: Executor
get() = coroutineScheduler
// This is variable for test purposes, so that we can reinitialize from clean state
private var coroutineScheduler = createScheduler()
private fun createScheduler() =
//1. 详见1.5.2
CoroutineScheduler(corePoolSize, maxPoolSize, idleWorkerKeepAliveNs, schedulerName)
//2. 详见1.5.2
override fun dispatch(context: CoroutineContext, block: Runnable): Unit 
= coroutineScheduler.dispatch(block)
...
}
//Executors.kt
//2. 实际上是继承ExecutorCoroutineDispatcher
public abstract class ExecutorCoroutineDispatcher: CoroutineDispatcher(), Closeable {
...
}

• 可以看到实际上调用了CoroutineScheduler.dispatch方法。此时发现，第二个参数是Runnable类型的，而在1.4小节中，我们知道传入的是this也就是DispatchedContinuation，所以DispatchedContinuation继承的父类中，必定有继承了Runnable接口，而他的run方法的实现也在父类中，这块我们暂时按下不表，接着看继续跟踪coroutineScheduler.dispatch(block)。

1.5.2 CoroutineScheduler

internal class CoroutineScheduler(
@JvmField val corePoolSize: Int,
@JvmField val maxPoolSize: Int,
@JvmField val idleWorkerKeepAliveNs: Long = IDLE_WORKER_KEEP_ALIVE_NS,
@JvmField val schedulerName: String = DEFAULT_SCHEDULER_NAME
) : Executor, Closeable {
	... 
override fun execute(command: Runnable) = dispatch(command)
fun dispatch(block: Runnable, taskContext: TaskContext = NonBlockingContext, tailDispatch: Boolean = false) {
trackTask() // this is needed for virtual time support
val task = createTask(block, taskContext)
// try to submit the task to the local queue and act depending on the result
val currentWorker = currentWorker()
val notAdded = currentWorker.submitToLocalQueue(task, tailDispatch)
if (notAdded != null) {
if (!addToGlobalQueue(notAdded)) {
// Global queue is closed in the last step of close/shutdown -- no more tasks should be accepted
throw RejectedExecutionException("$schedulerName was terminated")
}
}
val skipUnpark = tailDispatch && currentWorker != null
// Checking 'task' instead of 'notAdded' is completely okay
if (task.mode == TASK_NON_BLOCKING) {
if (skipUnpark) return
signalCpuWork()
} else {
// Increment blocking tasks anyway
signalBlockingWork(skipUnpark = skipUnpark)
}
}
}

• 该类继承了Executor类，而且它的构建参数可看到是线程池的参数，所以可以知道这个其实是Kotlin协程实现的一个线程池，具体就不跟进去了。
• execute()过程也是dispatch过程：将任务投递到任务队列，然后通知线程去取任务执行，自此完成了线程切换动作。
• 而在新线程里执行的Runnable为1.4中的调用代码：dispatcher.dispatch(context, this)中的this，也就是DispatchedContinuation。DispatchedContinuation.kt并没有实现run方法，那么一定是他继承的父类实现了Runnable接口并实现，所以需要接着看它继承的父类：DispatchedTask类。

1.6 DispatchedTask.run()

internal abstract class DispatchedTask<in T>(
@JvmField public var resumeMode: Int
) : SchedulerTask() {
	...
internal abstract val delegate: Continuation<T>
@Suppress("UNCHECKED_CAST")
internal open fun <T> getSuccessfulResult(state: Any?): T =
state as T
internal open fun getExceptionalResult(state: Any?): Throwable? =
(state as? CompletedExceptionally)?.cause
public final override fun run() {
assert { resumeMode != MODE_UNINITIALIZED } // should have been set before dispatching
val taskContext = this.taskContext
var fatalException: Throwable? = null
try {
val delegate = delegate as DispatchedContinuation<T>
//1. 取出代理商的续体
val continuation = delegate.continuation
withContinuationContext(continuation, delegate.countOrElement) {
val context = continuation.context
val state = takeState() // NOTE: Must take state in any case, even if cancelled
val exception = getExceptionalResult(state)
val job = if (exception == null && resumeMode.isCancellableMode) context[Job] else null
if (job != null && !job.isActive) {
val cause = job.getCancellationException()
cancelCompletedResult(state, cause)
continuation.resumeWithStackTrace(cause)
} else {
if (exception != null) {
continuation.resumeWithException(exception)
} else {
//1. 被包装的续体的resume方法，真正的开始出发其协程状态机代码。
continuation.resume(getSuccessfulResult(state))
}
}
}
} catch (e: Throwable) {
// This instead of runCatching to have nicer stacktrace and debug experience
fatalException = e
} finally {
val result = runCatching { taskContext.afterTask() }
handleFatalException(fatalException, result.exceptionOrNull())
}
}
}

• 将delegate转为DispatchedContinuation，应该注意1.4 小节中DispatchedContinuation继承DispatchTask时，便对此delegate进行了复写：

override val delegate: Continuation

get() = this

而此delegate.continucation便是当初newDispatchedContinuation(this)时传入的this，此this就是Kotlin编译器一开始为协程体生成的SuspendLambda类型对象。具体可以回看1.3小节。

• 调用了continuation.resume()方法触发了协程的状态机进而开始执行协程业务逻辑代码，结合之前1.5.2的分析可以知道，这个方法的调用已经是被dispatch到特定线程，完成线程切换后执行的。所以协程状态机的代码也是跑在新线程上的。

1.7 总结

至此，协程的线程调度分析结束，关键有如下几个要点：

• 创建SuspendLambda时，他的协程上下文对象来自于comletion.context，默认就是Dispatcher.Default。
• SuspendLambda启动时调用了intercept()进行一层包装，得到DispatchedContinuation，后续协程启动是启动的DispatchedContinuation协程。
• DispatchedContinuation继承于Runnable接口，协程启动时将自己投递到分发器dispatcher执行run方法，从而达到了线程切换效果。
• 在DispatchedContinuation的run方法中，调用SuspendLambda.resume()启动状态机。在新线程执行协程状态机代码。

这一小节中，介绍了如何将协程调度到目的线程执行，接下来分析如何做到随意切换线程后，然后再恢复到原来线程的。

二、协程中的线程切换

在第一小节中，我们搞清楚了协程启动时，协程调度器是如何在其中起作用的。这一小节旨在剖析在协程用分发器切换线程执行新的挂起函数后，是如何切换会原来线程继续执行剩下的逻辑的。

为此，我们需要将1.1的测试代码反编译出来实际代码进而分析。

2.1 反编译代码

2.1.1 MainActivityonCreateonCreateonCreate1

final class MainActivity$onCreate$1 extends SuspendLambda implements Function2<CoroutineScope, Continuation<? super Unit>, Object> {
...
@Override // kotlin.coroutines.jvm.internal.BaseContinuationImpl
public final Object invokeSuspend(Object $result) {
Object coroutine_suspended = IntrinsicsKt.getCOROUTINE_SUSPENDED();
switch (this.label) {
case 0:
ResultKt.throwOnFailure($result);
Log.d(MainActivity.TAG, LiveLiterals$MainActivityKt.INSTANCE.m4147xf96cab04());
this.label = 1;
//1. 新建编译器自动生成的继承于SuspendLambda的类型。
AnonymousClass1 anonymousClass1 = new AnonymousClass1(null);
//2. 调用withContext
	Object res = BuildersKt.withContext(Dispatchers.getIO(), anonymousClass1, this);
if (res != coroutine_suspended) {
break;
} else {
//挂起
return coroutine_suspended;
}
case 1:
ResultKt.throwOnFailure($result);
break;
default:
throw new IllegalStateException("call to 'resume' before 'invoke' with coroutine");
}
Log.d(MainActivity.TAG, LiveLiterals$MainActivityKt.INSTANCE.m4148xe0c1b328());
return Unit.INSTANCE;
}
}

根据之前的文章分析，这里suspend lambda 的类型都自动生成继承于SuspendLambda的类型。详见2.1.2。

将anonymousClass1传入withContext，而且注意这里传入了this==MainActivity$onCreate$1，详见2.2。

2.1.2 AnonymousClass1

/* compiled from: MainActivity.kt */
public static final class AnonymousClass1 extends SuspendLambda implements Function2<CoroutineScope, Continuation<? super Integer>, Object> {
int label
...
@Override // kotlin.coroutines.jvm.internal.BaseContinuationImpl
public final Object invokeSuspend(Object obj) {
IntrinsicsKt.getCOROUTINE_SUSPENDED();
switch (this.label) {
case 0:
ResultKt.throwOnFailure(obj);
return Boxing.boxInt(Log.d(MainActivity.TAG, LiveLiterals$MainActivityKt.INSTANCE.m4146x7c0f011f()));
default:
throw new IllegalStateException("call to 'resume' before 'invoke' with coroutine");
}
}
}

2.2 withContext

public suspend fun <T> withContext(
context: CoroutineContext,
block: suspend CoroutineScope.() -> T
): T {
contract {
callsInPlace(block, InvocationKind.EXACTLY_ONCE)
}
//1. 获取当前协程, 注意这里的uCont就是当前续体，也就是MainActivity$onCreate$1
return suspendCoroutineUninterceptedOrReturn sc@ { uCont ->
//2. 计算获的新的协程上下文
val oldContext = uCont.context
val newContext = oldContext + context
//3. 快速判断：新上下文和旧上下文一致的情况快速处理。
// always check for cancellation of new context
newContext.ensureActive()
// FAST PATH #1 -- new context is the same as the old one
if (newContext === oldContext) {
val coroutine = ScopeCoroutine(newContext, uCont)
return@sc coroutine.startUndispatchedOrReturn(coroutine, block)
}
// FAST PATH #2 -- the new dispatcher is the same as the old one (something else changed)
// `equals` is used by design (see equals implementation is wrapper context like ExecutorCoroutineDispatcher)
if (newContext[ContinuationInterceptor] == oldContext[ContinuationInterceptor]) {
val coroutine = UndispatchedCoroutine(newContext, uCont)
// There are changes in the context, so this thread needs to be updated
withCoroutineContext(newContext, null) {
return@sc coroutine.startUndispatchedOrReturn(coroutine, block)
}
}
// SLOW PATH -- use new dispatcher
//4. 新建一个DispatchedCoroutine
val coroutine = DispatchedCoroutine(newContext, uCont)
//5. 启动协程
block.startCoroutineCancellable(coroutine, coroutine)
coroutine.getResult()
}
}

• suspendCoroutineUninterceptedOrReturn这个函数直接步进是看不到实现的，它的实现是由Kotlin编译器生成的，它的作用是用来获取当前续体的，并且通过uCont返回，这里就是MainActivity$onCreate$1。
• 将旧协程上下文和新的上下文一起。计算得到最终的上下文。这里的context==Dispatchers.getIO()。
• 快速判断，不用看。
• 新建一个DispatchedCoroutine，注意这里传入了新的协程上下文和当前续体对象。
• 调用startCoroutineCancellable()启动协程。这里的同1.3.2小节分析一样，详见 2.2.1

2.2.1 startCoroutineCancellable

internal fun <R, T> (suspend (R) -> T).startCoroutineCancellable(
receiver: R, completion: Continuation<T>,
onCancellation: ((cause: Throwable) -> Unit)? = null
) =
runSafely(completion) {
	//1. 创建SuspendLambda协程体
createCoroutineUnintercepted(receiver, completion)
//2. 拦截：取出分发器，并构建方法器Continuation。详见1.3.1
.intercepted()
//3. 调用方法器Continuation的resume方法，详见1.4
.resumeCancellableWith(Result.success(Unit), onCancellation)
}

此方法在之前1.3小节已经分析过，针对此此次调用，其中的改变是协程上下文中的分发器已经被设置为Dispatchers.Main。

• 创建了SuspendLambda对象，此对象的CoroutineContext为completion.context。而其中的ContinuationInterceptor类型Element就是我们之前传入的Dispatchers.Main。
• 创建一个DispatchedContinuation。
• 将协程SuspendLambda的状态机逻辑通过Dispatcher.Main调度到主线程执行，调度过程参考第一下节。分发逻辑详见2.7小节。
• 当SuspendLambda的状态机invokeSuspend()逻辑执行完成后，会返回到BaseContinuationImpl.resumeWith()，我们需要接此方法分析，来得到协程在切换到主线程执行后，又是怎么切回协程体1的执行线程的，详见2.3。

2.3 resumeWith

public final override fun resumeWith(result: Result<Any?>) {
// This loop unrolls recursion in current.resumeWith(param) to make saner and shorter stack traces on resume
var current = this
var param = result
while (true) {
// Invoke "resume" debug probe on every resumed continuation, so that a debugging library infrastructure
// can precisely track what part of suspended callstack was already resumed
probeCoroutineResumed(current)
with(current) {
val completion = completion!! // fail fast when trying to resume continuation without completion
val outcome: Result<Any?> =
try {
val outcome = invokeSuspend(param)
if (outcome === COROUTINE_SUSPENDED) return
Result.success(outcome)
} catch (exception: Throwable) {
Result.failure(exception)
}
releaseIntercepted() // this state machine instance is terminating
if (completion is BaseContinuationImpl) {
// unrolling recursion via loop
current = completion
param = outcome
} else {
//1. 进入此判断
// top-level completion reached -- invoke and return
completion.resumeWith(outcome)
return
}
}
}
}

当状态机执行完后， 后进入到completion的类型判断，由2.2和2.2.1可以知道，当初传入的completion是DispatchedCoroutine类型，所以加入到else分支，调用了DispatchedCoroutine.resumeWith()，接下来分析此方法。

在此之前，我们需要看下DispatchedCoroutine的继承关系，详见2.4.1。如果想直接跟踪流程，可以直接看2.4.2。

2.4 DispatchedCoroutine

2.4.1 DispatchedCoroutine 的继承关系

internal class DispatchedCoroutine<in T>(
context: CoroutineContext,
uCont: Continuation<T>
) : ScopeCoroutine<T>(context, uCont) {
}

继承于ScopeCoroutine

internal open class ScopeCoroutine<in T>(
context: CoroutineContext,
@JvmField val uCont: Continuation<T> // unintercepted continuation
) : AbstractCoroutine<T>(context, true, true), CoroutineStackFrame {
}

继承于AbstractCoroutine

public abstract class AbstractCoroutine<in T>(
parentContext: CoroutineContext,
initParentJob: Boolean,
active: Boolean
) : JobSupport(active), Job, Continuation<T>, CoroutineScope {
}

2.5 协程线程的恢复

2.5.1 AbstractCoroutine.resumeWith()

public final override fun resumeWith(result: Result<T>) {
val state = makeCompletingOnce(result.toState())
if (state === COMPLETING_WAITING_CHILDREN) return
afterResume(state)
}

调用了afterResume方法，此方法在DispatchedCoroutine类型有具体实现。见2.5.2

2.5.2 afterResume

//DispatchedCoroutine
override fun afterResume(state: Any?) {
if (tryResume()) return // completed before getResult invocation -- bail out
// Resume in a cancellable way because we have to switch back to the original dispatcher
uCont.intercepted().resumeCancellableWith(recoverResult(state, uCont))
}

• 取出当前续体uCont，这个续体根据之前的分析：2.2小节，可以知道它等于MainActivity$onCreate$1。
• intercepted()：取出其分发拦截器
• resumeCancellableWith：使用方法拦截器协程体，将uCont续体的状态机逻辑调度到相对应的线程环境执行，这里就是之前的Dispatcher.Default。注意其注释：“将其切换到原先的分发器”。2⃣而这一过程其实和1.3小节的过程一致。
• 恢复到Dispatcher.Default继续执行状态机时，由于label已经被更新，所以会往下继续执行，打印最后一句log。

2.6 总结

withContext(Dispatcher.Main)启动的协程时，取得当前协程续体uCount也就是MainActivity$onCreate$1，会计算出新的协程context，然后用它们创建一个DispatchedCoroutine。

AnonymousClass1协程启动时，用DispatchedCoroutine作为completion参数，然后启动，此时会调度主线程执行协程。

当协程执行完成后，AnonymousClass1.resumeWith()方法会调用completion.resumeWith()。

DispatchedCoroutine.resumeWith()方法会调用uCount.intercepted().resumeCancellableWith()，使得父协程进行调度并接着执行状态机逻辑。

2.7 Dispatchers.Main

@JvmStatic
public actual val Main: MainCoroutineDispatcher get() 
= MainDispatcherLoader.dispatcher

直接详见2.7.1

2.7.1 MainDispatcherLoader

internal object MainDispatcherLoader {
private val FAST_SERVICE_LOADER_ENABLED = systemProp(FAST_SERVICE_LOADER_PROPERTY_NAME, true)
@JvmField
val dispatcher: MainCoroutineDispatcher = loadMainDispatcher()
private fun loadMainDispatcher(): MainCoroutineDispatcher {
return try {
val factories = if (FAST_SERVICE_LOADER_ENABLED) {
FastServiceLoader.loadMainDispatcherFactory()
} else {
// We are explicitly using the
// `ServiceLoader.load(MyClass::class.java, MyClass::class.java.classLoader).iterator()`
// form of the ServiceLoader call to enable R8 optimization when compiled on Android.
// 1.获得MainDispatcherFactory的实现类
ServiceLoader.load(
MainDispatcherFactory::class.java,
MainDispatcherFactory::class.java.classLoader
).iterator().asSequence().toList()
}
@Suppress("ConstantConditionIf")
factories.maxByOrNull { it.loadPriority }?.tryCreateDispatcher(factories)
?: createMissingDispatcher()
} catch (e: Throwable) {
// Service loader can throw an exception as well
createMissingDispatcher(e)
}
}
}

• 通过ServiceLoad机制获取MainDispatcherFactory的实现类，而在源码里面，其实现类为AndroidDispatcherFactory
• 调用tryCreateDispatcher()创建分发器，详见2.7.2。

2.7.2 AndroidDispatcherFactory

internal class AndroidDispatcherFactory : MainDispatcherFactory {
override fun createDispatcher(allFactories: List<MainDispatcherFactory>) =
HandlerContext(Looper.getMainLooper().asHandler(async = true))
override fun hintOnError(): String = "For tests Dispatchers.setMain from kotlinx-coroutines-test module can be used"
override val loadPriority: Int
get() = Int.MAX_VALUE / 2
}

根据createDispatcher分发，主线程分发器的实现类为HandlerContext类型，传入用MainLooper构建的Handler。详见2.7.3。

2.7.3 HandlerContext

internal class HandlerContext private constructor(
private val handler: Handler,
private val name: String?,
private val invokeImmediately: Boolean
) : HandlerDispatcher(), Delay {
/**
 * Creates [CoroutineDispatcher] for the given Android [handler].
 *
 * @param handler a handler.
 * @param name an optional name for debugging.
 */
constructor(
handler: Handler,
name: String? = null
) : this(handler, name, false)
@Volatile
private var _immediate: HandlerContext? = if (invokeImmediately) this else null
override val immediate: HandlerContext = _immediate ?:
HandlerContext(handler, name, true).also { _immediate = it }
override fun isDispatchNeeded(context: CoroutineContext): Boolean {
return !invokeImmediately || Looper.myLooper() != handler.looper
}
override fun dispatch(context: CoroutineContext, block: Runnable) {
if (!handler.post(block)) {
cancelOnRejection(context, block)
}
}
override fun scheduleResumeAfterDelay(timeMillis: Long, continuation: CancellableContinuation<Unit>) {
val block = Runnable {
with(continuation) { resumeUndispatched(Unit) }
}
if (handler.postDelayed(block, timeMillis.coerceAtMost(MAX_DELAY))) {
continuation.invokeOnCancellation { handler.removeCallbacks(block) }
} else {
cancelOnRejection(continuation.context, block)
}
}
   ...
}

HandlerContext继承于HandlerDispatcher，而他的dispatch方法，可以看到，就是将block丢到设置MainLooper的handler执行。所以续体将会在主线程执行状态机，达到切换到主线程执行协程的目的。