Android Framework中的线程Thread及它的threadLoop方法

当初跟踪Camera的代码中的时候一直追到了HAL层，而在Framework中的代码看见了许很多多的Thread。它们普遍的特点就是有一个threadLoop方法。依照字面的意思应该是这个线程能够循环处理数据。相应我想到到了java上层中的HandlerThread，这个预计也差点儿相同，但当时心里总有一个疙瘩。想弄清楚它为什么能够循环。还有它究竟是怎么循环起来的？

Android中java世界的Thread

我们先来看看java是怎么创建一个线程的。这个是最舒服的，也是我最熟悉的。

new Thread(new Runnable() {

        @Override
        public void run() {
            // TODO Auto-generated method stub
                ...
        }
}).start();

当然，你也能够在android中创建一个消息循环的HandlerThread

HandlerThread mThread = new HandlerThread("test");
mThread.start();
Handler mHandler = new Handler(mThread.getLooper()){

        @Override
        public void handleMessage(Message msg) {
            // TODO Auto-generated method stub
            super.handleMessage(msg);
        }

};

上面中通过mHandler发送消息就能够在mThread中处理了。而且这个mThread不是UIThread，不会堵塞主线程。

HandlerThread是一个好东西，在源代码中处处可见，希望对此不熟悉的新手及时去学习下它的使用方法。

Linux下c语言的Thread

java世界的Thread非常方便，那么c呢？

Android基于linux所以，多线程编程也应该基于linux下的多线程。linux下的c语言用pthread。大家能够看这篇文章。

linux下C/C++,多线程pthread

我把里面的样例改良了一下

test.c

#include <stdio.h>
#include <stdlib.h>
#include <pthread.h>
#include <unistd.h>

//线程函数
void *test(void *ptr)
{
    int i;
    for(i=0;i<8;i++)
    {
        printf("the pthread running ,count: %d\n",i);
        sleep(1);
    }

}

int main(void)
{
    pthread_t pId;
    int i,ret;
    //创建子线程，线程id为pId
    ret = pthread_create(&pId,NULL,test,NULL);

    if(ret != 0)
    {
        printf("create pthread error!\n");
        exit(1);
    }

    for(i=0;i < 5;i++)
    {
        printf("main thread running ,count : %d\n",i);
        sleep(1);
    }

    printf("main thread will exit when pthread is over\n");
    //等待线程pId的完毕
    pthread_join(pId,NULL);
    printf("main thread  exit\n");

    return 0;

}

然后编译

gcc -o test test.c -lpthread
./test

执行结果例如以下

main thread running ,count : 0
the pthread running ,count: 0
main thread running ,count : 1
the pthread running ,count: 1
main thread running ,count : 2
the pthread running ,count: 2
main thread running ,count : 3
the pthread running ,count: 3
main thread running ,count : 4
the pthread running ,count: 4
main thread will exit when pthread is over
the pthread running ,count: 5
the pthread running ,count: 6
the pthread running ,count: 7
main thread  exit

样例比較简单，主要是创建一个线程。然后主线程等待子线程执行完毕再退出。

Android Framework中的Thread

以下焦点回到文章的主题其中，我们来看看Framework中经常使用的Thread是个何种形态。

先看看活生生的样例。

在源代码中搜索threadLoop，当然也能够搜索thread,然后随便挑选一个Thread子类进行研究。

这里挑选

/frameworks/av/services/audioflinger/AudioWatchdog.h

#ifndef AUDIO_WATCHDOG_H
#define AUDIO_WATCHDOG_H

#include <time.h>
#include <utils/Thread.h>

namespace android {

......

class AudioWatchdog : public Thread {

public:
    AudioWatchdog(unsigned periodMs = 50) : Thread(false /*canCallJava*/), mPaused(false),
            mPeriodNs(periodMs * 1000000), mMaxCycleNs(mPeriodNs * 2),
            // mOldTs
            // mLogTs initialized below
            mOldTsValid(false), mUnderruns(0), mLogs(0), mDump(&mDummyDump)
        {
#define MIN_TIME_BETWEEN_LOGS_SEC 60
            // force an immediate log on first underrun
            mLogTs.tv_sec = MIN_TIME_BETWEEN_LOGS_SEC;
            mLogTs.tv_nsec = 0;
        }
    virtual         ~AudioWatchdog() { }

     // Do not call Thread::requestExitAndWait() without first calling requestExit().
    // Thread::requestExitAndWait() is not virtual, and the implementation doesn't do enough.
    virtual void        requestExit();

    // FIXME merge API and implementation with AudioTrackThread
    void            pause();        // suspend thread from execution at next loop boundary
    void            resume();       // allow thread to execute, if not requested to exit

    // Where to store the dump, or NULL to not update
    void            setDump(AudioWatchdogDump* dump);

private:
    virtual bool    threadLoop();

    Mutex           mMyLock;        // Thread::mLock is private
    Condition       mMyCond;        // Thread::mThreadExitedCondition is private
    bool            mPaused;        // whether thread is currently paused

    ......
};

}   // namespace android

#endif  // AUDIO_WATCHDOG_H

我们能够看到AudioWatchDog确实是Thread的子类，那好。以下看实现。

/frameworks/av/services/audioflinger/AudioWatchdog.cpp

#define LOG_TAG "AudioWatchdog"
//#define LOG_NDEBUG 0

#include <utils/Log.h>
#include "AudioWatchdog.h"

namespace android {

bool AudioWatchdog::threadLoop()
{
    {
        AutoMutex _l(mMyLock);
        if (mPaused) {
            mMyCond.wait(mMyLock);
            // ignore previous timestamp after resume()
            mOldTsValid = false;
            // force an immediate log on first underrun after resume()
            mLogTs.tv_sec = MIN_TIME_BETWEEN_LOGS_SEC;
            mLogTs.tv_nsec = 0;
            // caller will check for exitPending()
            return true;
        }
    }
    struct timespec newTs;
    int rc = clock_gettime(CLOCK_MONOTONIC, &newTs);
    if (rc != 0) {
        pause();
        return false;
    }
    if (!mOldTsValid) {
        mOldTs = newTs;
        mOldTsValid = true;
        return true;
    }
    time_t sec = newTs.tv_sec - mOldTs.tv_sec;
    long nsec = newTs.tv_nsec - mOldTs.tv_nsec;
    if (nsec < 0) {
        --sec;
        nsec += 1000000000;
    }
    mOldTs = newTs;
    // cycleNs is same as sec*1e9 + nsec, but limited to about 4 seconds
    uint32_t cycleNs = nsec;
    if (sec > 0) {
        if (sec < 4) {
            cycleNs += sec * 1000000000;
        } else {
            cycleNs = 4000000000u;
        }
    }
    mLogTs.tv_sec += sec;
    if ((mLogTs.tv_nsec += nsec) >= 1000000000) {
        mLogTs.tv_sec++;
        mLogTs.tv_nsec -= 1000000000;
    }
    if (cycleNs > mMaxCycleNs) {
        mDump->mUnderruns = ++mUnderruns;
        if (mLogTs.tv_sec >= MIN_TIME_BETWEEN_LOGS_SEC) {
            mDump->mLogs = ++mLogs;
            mDump->mMostRecent = time(NULL);
            ALOGW("Insufficient CPU for load: expected=%.1f actual=%.1f ms; underruns=%u logs=%u",
                mPeriodNs * 1e-6, cycleNs * 1e-6, mUnderruns, mLogs);
            mLogTs.tv_sec = 0;
            mLogTs.tv_nsec = 0;
        }
    }
    struct timespec req;
    req.tv_sec = 0;
    req.tv_nsec = mPeriodNs;
    rc = nanosleep(&req, NULL);
    if (!((rc == 0) || (rc == -1 && errno == EINTR))) {
        pause();
        return false;
    }
    return true;
}

void AudioWatchdog::requestExit()
{
    // must be in this order to avoid a race condition
    Thread::requestExit();
    resume();
}

void AudioWatchdog::pause()
{
    AutoMutex _l(mMyLock);
    mPaused = true;
}

void AudioWatchdog::resume()
{
    AutoMutex _l(mMyLock);
    if (mPaused) {
        mPaused = false;
        mMyCond.signal();
    }
}

}   // namespace android

非常明显，它的核心方法就是threadLoop()，在本文中我们不关心它具体的功能。仅仅想确定它是怎么启动的呢？又是怎么循环执行的呢？带着疑问我又在源代码中搜索关键字AudioWatchdog

结果发现有两个地方引用了。

/frameworks/av/services/audioflinger/AudioFlinger.h
/frameworks/av/services/audioflinger/AudioFlinger.cpp

在AudioFlinger.h中MixerThread中有个AudioWatchdog的sp对象

 class MixerThread : public PlaybackThread {
    public:
        MixerThread (const sp<AudioFlinger>& audioFlinger,
                     AudioStreamOut* output,
                     audio_io_handle_t id,
                     audio_devices_t device,
                     type_t type = MIXER);
        virtual             ~MixerThread();

    protected:

                    AudioMixer* mAudioMixer;    // normal mixer
    private:

                    sp<AudioWatchdog> mAudioWatchdog; // non-0 if there is an audio watchdog thread

    };

我们再看代码

/frameworks/av/services/audioflinger/AudioFlinger.cpp

AudioFlinger::MixerThread::MixerThread(const sp<AudioFlinger>& audioFlinger, AudioStreamOut* output,
        audio_io_handle_t id, audio_devices_t device, type_t type)
    :   PlaybackThread(audioFlinger, output, id, device, type),
        // mAudioMixer below
        // mFastMixer below
        mFastMixerFutex(0)
        // mOutputSink below
        // mPipeSink below
        // mNormalSink below
{

......
#ifdef AUDIO_WATCHDOG
        // create and start the watchdog
        mAudioWatchdog = new AudioWatchdog();
        mAudioWatchdog->setDump(&mAudioWatchdogDump);
        //AudioWatchdog的run方法在此调用，所以线程启动
        mAudioWatchdog->run("AudioWatchdog", PRIORITY_URGENT_AUDIO);
        tid = mAudioWatchdog->getTid();
        err = requestPriority(getpid_cached, tid, kPriorityFastMixer);
        if (err != 0) {
            ALOGW("Policy SCHED_FIFO priority %d is unavailable for pid %d tid %d; error %d",
                    kPriorityFastMixer, getpid_cached, tid, err);
        }
#endif

......
}

删掉不相关代码，我们看到AudioWatchdog对象确实创建了，而且调用了它的run方法。在java中Thread的run方法就是启动，这个也应该如此。可是如之前的源代码所看到的AudioWatchdog.cpp中并没有实现run方法，怎么办呢？别紧张，它还有父类Thread.

/frameworks/native/include/utils/Thread.h

#ifndef _LIBS_UTILS_THREAD_H
#define _LIBS_UTILS_THREAD_H

#include <stdint.h>
#include <sys/types.h>
#include <time.h>

#if defined(HAVE_PTHREADS)
# include <pthread.h>
#endif

#include <utils/Condition.h>
#include <utils/Errors.h>
#include <utils/Mutex.h>
#include <utils/RefBase.h>
#include <utils/Timers.h>
#include <utils/ThreadDefs.h>

// ---------------------------------------------------------------------------
namespace android {
// ---------------------------------------------------------------------------

class Thread : virtual public RefBase
{
public:
    // Create a Thread object, but doesn't create or start the associated
    // thread. See the run() method.
                        Thread(bool canCallJava = true);
    virtual             ~Thread();

    // Start the thread in threadLoop() which needs to be implemented.
    virtual status_t    run(    const char* name = 0,
                                int32_t priority = PRIORITY_DEFAULT,
                                size_t stack = 0);

    // Ask this object's thread to exit. This function is asynchronous, when the
    // function returns the thread might still be running. Of course, this
    // function can be called from a different thread.
    virtual void        requestExit();

    // Good place to do one-time initializations
    virtual status_t    readyToRun();

    // Call requestExit() and wait until this object's thread exits.
    // BE VERY CAREFUL of deadlocks. In particular, it would be silly to call
    // this function from this object's thread. Will return WOULD_BLOCK in
    // that case.
            status_t    requestExitAndWait();

    // Wait until this object's thread exits. Returns immediately if not yet running.
    // Do not call from this object's thread; will return WOULD_BLOCK in that case.
            status_t    join();

#ifdef HAVE_ANDROID_OS
    // Return the thread's kernel ID, same as the thread itself calling gettid() or
    // androidGetTid(), or -1 if the thread is not running.
            pid_t       getTid() const;
#endif

protected:
    // exitPending() returns true if requestExit() has been called.
            bool        exitPending() const;

private:
    // Derived class must implement threadLoop(). The thread starts its life
    // here. There are two ways of using the Thread object:
    // 1) loop: if threadLoop() returns true, it will be called again if
    //          requestExit() wasn't called.
    // 2) once: if threadLoop() returns false, the thread will exit upon return.
    virtual bool        threadLoop() = 0;

private:
    Thread& operator=(const Thread&);
    static  int             _threadLoop(void* user);
    const   bool            mCanCallJava;
    // always hold mLock when reading or writing
            thread_id_t     mThread;
    mutable Mutex           mLock;
            Condition       mThreadExitedCondition;
            status_t        mStatus;
    // note that all accesses of mExitPending and mRunning need to hold mLock
    volatile bool           mExitPending;
    volatile bool           mRunning;
            sp<Thread>      mHoldSelf;
#ifdef HAVE_ANDROID_OS
    // legacy for debugging, not used by getTid() as it is set by the child thread
    // and so is not initialized until the child reaches that point
            pid_t           mTid;
#endif
};

}; // namespace android

// ---------------------------------------------------------------------------
#endif // _LIBS_UTILS_THREAD_H
// 

能够看到确实有run方法。那以下看看它的实现

status_t Thread::run(const char* name, int32_t priority, size_t stack)
{
    Mutex::Autolock _l(mLock);

    if (mRunning) {
        // thread already started
        return INVALID_OPERATION;
    }

    // reset status and exitPending to their default value, so we can
    // try again after an error happened (either below, or in readyToRun())
    mStatus = NO_ERROR;
    mExitPending = false;
    mThread = thread_id_t(-1);

    // hold a strong reference on ourself
    mHoldSelf = this;

    mRunning = true;

    bool res;
    if (mCanCallJava) {
        res = createThreadEtc(_threadLoop,
                this, name, priority, stack, &mThread);
    } else {
        res = androidCreateRawThreadEtc(_threadLoop,
                this, name, priority, stack, &mThread);
    }

    if (res == false) {
        mStatus = UNKNOWN_ERROR;   // something happened!
        mRunning = false;
        mThread = thread_id_t(-1);
        mHoldSelf.clear();  // "this" may have gone away after this.

        return UNKNOWN_ERROR;
    }

    // Do not refer to mStatus here: The thread is already running (may, in fact
    // already have exited with a valid mStatus result). The NO_ERROR indication
    // here merely indicates successfully starting the thread and does not
    // imply successful termination/execution.
    return NO_ERROR;

    // Exiting scope of mLock is a memory barrier and allows new thread to run
}

run（）方法中有这么一段

if (mCanCallJava) {
        res = createThreadEtc(_threadLoop,
                this, name, priority, stack, &mThread);
    } else {
        res = androidCreateRawThreadEtc(_threadLoop,
                this, name, priority, stack, &mThread);
    }

mCanCallJava的意思是能不能被JNI层调用，然后依据值去创建Thread，这里有两个分支，我们就选择createThreadEtc()

终于代码会走到这里

int androidCreateRawThreadEtc(android_thread_func_t entryFunction,
                               void *userData,
                               const char* threadName,
                               int32_t threadPriority,
                               size_t threadStackSize,
                               android_thread_id_t *threadId)
{
    ......
        entryFunction = (android_thread_func_t)&thread_data_t::trampoline;
        userData = t;
    }
#endif

    if (threadStackSize) {
        pthread_attr_setstacksize(&attr, threadStackSize);
    }

    errno = 0;
    pthread_t thread;
    //在此创建，文章开头我有写样例怎么用c语言创建一个线程
    int result = pthread_create(&thread, &attr,
                    (android_pthread_entry)entryFunction, userData);
    pthread_attr_destroy(&attr);
    if (result != 0) {
        ALOGE("androidCreateRawThreadEtc failed (entry=%p, res=%d, errno=%d)\n"
             "(android threadPriority=%d)",
            entryFunction, result, errno, threadPriority);
        return 0;
    }

    ......
    return 1;
}

删除了不相关代码，大家看看是不是非常熟悉啊。我在文章開始的部分就写出了linux下c语言pthread创建线程的样例，大家能够回头看看。也就是pthread_create()。

这里面传进来的entryFunction是Thread中的_threadLoop()

int Thread::_threadLoop(void* user)
{
    Thread* const self = static_cast<Thread*>(user);

    sp<Thread> strong(self->mHoldSelf);
    wp<Thread> weak(strong);
    self->mHoldSelf.clear();

#ifdef HAVE_ANDROID_OS
    // this is very useful for debugging with gdb
    self->mTid = gettid();
#endif

    bool first = true;

    do {
        bool result;
        if (first) {
            first = false;
            self->mStatus = self->readyToRun();
            result = (self->mStatus == NO_ERROR);

            if (result && !self->exitPending()) {
                // Binder threads (and maybe others) rely on threadLoop
                // running at least once after a successful ::readyToRun()
                // (unless, of course, the thread has already been asked to exit
                // at that point).
                // This is because threads are essentially used like this:
                //   (new ThreadSubclass())->run();
                // The caller therefore does not retain a strong reference to
                // the thread and the thread would simply disappear after the
                // successful ::readyToRun() call instead of entering the
                // threadLoop at least once.
                //调用threadLoop()
                result = self->threadLoop();
            }
        } else {
            result = self->threadLoop();
        }

        // establish a scope for mLock
        {
        Mutex::Autolock _l(self->mLock);
        if (result == false || self->mExitPending) {
            self->mExitPending = true;
            self->mRunning = false;
            // clear thread ID so that requestExitAndWait() does not exit if
            // called by a new thread using the same thread ID as this one.
            self->mThread = thread_id_t(-1);
            // note that interested observers blocked in requestExitAndWait are
            // awoken by broadcast, but blocked on mLock until break exits scope
            self->mThreadExitedCondition.broadcast();
            break;
        }
        }

        // Release our strong reference, to let a chance to the thread
        // to die a peaceful death.
        strong.clear();
        // And immediately, re-acquire a strong reference for the next loop
        strong = weak.promote();
    } while(strong != 0);

    return 0;
}

_threadLoop()这种方法就是Thread的最大秘密，它是一个while循环。

1、创建线程时。会sp和wp一次线程本身。

2、假设是第一次执行会执行线程的readyToRun()方法，再执行threadLoop()，否则，直接执行threadLoop()。

3、threadLoop()方法有返回值，假设threadLoop()返回false的时候，线程会做清理工作。然后退出while循环，结束执行。

所以在这里。我開始时的疑问—为什么线程Thread中的threadLoop()能够循环处理数据就到此做了说明。

Thread被创建，

Thread中的run被调用。

__threadLoop()被调用，

readyToRun()被调用。

然后循环调用threadLoop()。

而且在threadLoop()返回false时。能够退出循环。

！

！！

!!!

还有。最关键的一点是threadLoop能够循环事实上是由于调用它的_threadLoop()方法里面有一个while循环。

特殊情况

有的时候Android Framework中Thread的run()方法非常难发如今哪里被调用。如SurfaceFlinger它也是一个Thread子类。

在源代码中搜索能够发现它的创建位置

class SurfaceFlinger : public BinderService<SurfaceFlinger>,
                       public BnSurfaceComposer,
                       private IBinder::DeathRecipient,
                       private Thread,
                       private HWComposer::EventHandler
{
public:
    static char const* getServiceName() {
        return "SurfaceFlinger";
    }

    SurfaceFlinger();

    /* ------------------------------------------------------------------------
     * Thread interface
     */
    virtual bool threadLoop();
    virtual status_t readyToRun();
    virtual void onFirstRef();

};

// ---------------------------------------------------------------------------
}; // namespace android

#endif // ANDROID_SURFACE_FLINGER_H

去找它创建的地方

/frameworks/base/cmds/system_server/library/system_init.cpp

extern "C" status_t system_init()
{
    ALOGI("Entered system_init()");

    sp<ProcessState> proc(ProcessState::self());

    sp<IServiceManager> sm = defaultServiceManager();
    ALOGI("ServiceManager: %p\n", sm.get());

    char propBuf[PROPERTY_VALUE_MAX];
    property_get("system_init.startsurfaceflinger", propBuf, "1");
    if (strcmp(propBuf, "1") == 0) {
        // Start the SurfaceFlinger
        SurfaceFlinger::instantiate();
    }

    // And now start the Android runtime.  We have to do this bit
    // of nastiness because the Android runtime initialization requires
    // some of the core system services to already be started.
    // All other servers should just start the Android runtime at
    // the beginning of their processes's main(), before calling
    // the init function.
    ALOGI("System server: starting Android runtime.\n");
    AndroidRuntime* runtime = AndroidRuntime::getRuntime();

    ALOGI("System server: starting Android services.\n");
    JNIEnv* env = runtime->getJNIEnv();
    if (env == NULL) {
        return UNKNOWN_ERROR;
    }
    jclass clazz = env->FindClass("com/android/server/SystemServer");
    if (clazz == NULL) {
        return UNKNOWN_ERROR;
    }
    jmethodID methodId = env->GetStaticMethodID(clazz, "init2", "()V");
    if (methodId == NULL) {
        return UNKNOWN_ERROR;
    }
    env->CallStaticVoidMethod(clazz, methodId);

    ALOGI("System server: entering thread pool.\n");
    ProcessState::self()->startThreadPool();
    IPCThreadState::self()->joinThreadPool();
    ALOGI("System server: exiting thread pool.\n");

    return NO_ERROR;
}

我们能够看到

 SurfaceFlinger::instantiate();

但它本身并没有实现instantiate()方法，那之类找它的父类了。

/frameworks/native/include/binder/BinderService.h

namespace android {

template<typename SERVICE>
class BinderService
{
public:
    static status_t publish(bool allowIsolated = false) {
        sp<IServiceManager> sm(defaultServiceManager());
        return sm->addService(String16(SERVICE::getServiceName()), new SERVICE(), allowIsolated);
    }

    static void publishAndJoinThreadPool(bool allowIsolated = false) {
        sp<IServiceManager> sm(defaultServiceManager());
        sm->addService(String16(SERVICE::getServiceName()), new SERVICE(), allowIsolated);
        ProcessState::self()->startThreadPool();
        IPCThreadState::self()->joinThreadPool();
    }

    static void instantiate() { publish(); }

    static status_t shutdown() {
        return NO_ERROR;
    }
};

}; // namespace android
// ---------------------------------------------------------------------------
#endif // ANDROID_BINDER_SERVICE_H

会调用publish()方法。

而SERVICE在这里是一个模板类。在这里SERVICE自然相应SurfaceFlinger

所以publish()会向ServiceManager加入一个Service这个Service就是Surfaceflinger。

然后我们看SurfaceFlinger的构造函数

/frameworks/native/services/surfaceflinger/SurfaceFlinger.cpp

SurfaceFlinger::SurfaceFlinger()
    :   BnSurfaceComposer(), Thread(false),
        mTransactionFlags(0),
        mTransactionPending(false),
        mAnimTransactionPending(false),
        mLayersRemoved(false),
        mRepaintEverything(0),
        mBootTime(systemTime()),
        mVisibleRegionsDirty(false),
        mHwWorkListDirty(false),
        mDebugRegion(0),
        mDebugDDMS(0),
        mDebugDisableHWC(0),
        mDebugDisableTransformHint(0),
        mDebugInSwapBuffers(0),
        mLastSwapBufferTime(0),
        mDebugInTransaction(0),
        mLastTransactionTime(0),
        mBootFinished(false)
{
    ALOGI("SurfaceFlinger is starting");

    // debugging stuff...
    char value[PROPERTY_VALUE_MAX];

    property_get("debug.sf.showupdates", value, "0");
    mDebugRegion = atoi(value);

    property_get("debug.sf.ddms", value, "0");
    mDebugDDMS = atoi(value);
    if (mDebugDDMS) {
        if (!startDdmConnection()) {
            // start failed, and DDMS debugging not enabled
            mDebugDDMS = 0;
        }
    }
    ALOGI_IF(mDebugRegion, "showupdates enabled");
    ALOGI_IF(mDebugDDMS, "DDMS debugging enabled");
}

可是遗憾的是没有发现run()方法的影踪，没有办法仅仅得去父类构造方法看

结果发现也没有！

。。

没有办法，继续在源代码中搜索SurfaceFlinger。结果发现与之相关的信息大多是sp<SurfaceFlinger>

就看看sp吧。

sp是Android在c++中搞得相似java中弱引用、强引用的一套指针概念，那应该是方便回收吧。

而Android Framework中的c++世界，RefBase这个类有点像java中的Object.

而sp是一个模板类。

这部分内容。请看点击以下链接

Android中的sp和wp指针

上面的链接讲得还算具体。这里纠结sp的过多细节。长话短说。总之第一次对SurfaceFlinger引用调用sp<SurfaceFlinger>时会调用SurfaceFlinger的onFirstRef()方法。

那好。看代码吧

void SurfaceFlinger::onFirstRef()
{
    mEventQueue.init(this);
    //run方法在此被调用
    run("SurfaceFlinger", PRIORITY_URGENT_DISPLAY);

    // Wait for the main thread to be done with its initialization
    mReadyToRunBarrier.wait();
}

看见没有？run()方法在这里调用了。

所以，在Framework中假设你找不到一个Thread在何处被启动，那么去它的onFirstRef()方法中去看看吧