Android4.4深入浅出之SurfaceFlinger与Client通信框架（一）

               SurfaceFlinger框架是基于Binder进程间通信机制搭建的，SF作为一个服务进程，用户程序想要跟它通信必然要经过Binder机制。首先说一下，用户要跟SF通信，那么SF必须出现在ServiceManager中，因为SF也是一个服务，所有的服务都由ServiceManager来进行统一管理。在系统启动的过程中，SF就在ServiceManager中注册好了，注册好之后，SF在后台中监视一些surface的变化从而做出处理。

             而启动之后，用户程序想操作一些跟surface有关的动作，就必须和SF进行交互。而这种交互是基于Binder进程间通信机制的。下面是一张图简单说明了SF的基本框架：

大体结构是这样，有些地方有些小错误用图形不能很好的表达。用文字来说明一下：

当应用程序需要请求SurfaceFlinger服务时，首先需要构造SurfaceComposerClient对象，通过SurfaceComposerClient对象就可以访问SurfaceFlinger服务了。基本的服务流程都是这么走的，我们看一下SurfaceComposerClient的构造函数：

 

SurfaceComposerClient::SurfaceComposerClient()
    : mStatus(NO_INIT), mComposer(Composer::getInstance())
{
}

可想而知没做啥事啊，SurfaceComposerClient 继承于RefBase类，在创建SurfaceComposerClient对象，系统第一次引用SurfaceComposerClient对象时，onFirstRef函数自动调用，我们看一下它的onFirstRef

 

void SurfaceComposerClient::onFirstRef() {
    sp<ISurfaceComposer> sm(ComposerService::getComposerService());
    if (sm != 0) {
        sp<ISurfaceComposerClient> conn = sm->createConnection();
        if (conn != 0) {
            mClient = conn;
            mStatus = NO_ERROR;
        }
    }
}

上面代码，第二行的作用是获得SurfaceFlinger这个服务，我们跟踪一下：

 

sp<ISurfaceComposer> ComposerService::getComposerService() {
    ComposerService& instance = ComposerService::getInstance();
    Mutex::Autolock _l(instance.mLock);
    if (instance.mComposerService == NULL) {
        ComposerService::getInstance().connectLocked();
        assert(instance.mComposerService != NULL);
        ALOGD("ComposerService reconnected");
    }
    return instance.mComposerService;
}

第二行得到了ComposerService的对象，是为了下一步访问它的成员变量mComposerService（ISurfaceComposer类的），跟踪到第五行的函数 connectLocked（）：

 

void ComposerService::connectLocked() {
    const String16 name("SurfaceFlinger");
    while (getService(name, &mComposerService) != NO_ERROR) {
        usleep(250000);
    }
    assert(mComposerService != NULL);                                                                                                         //省略了一些东西。。。
 
           }

这里看到第三行getService（），第一个参数是“SurfaceFlinger”，第二个参数是mComposerService，记住它的类型的强引用的ISurfaceComposer，对就是这句得到了一个SurfaceFlinger服务的代理对象，看一下这个的函数的实体：

template<typename INTERFACE>
status_t getService(const String16& name, sp<INTERFACE>* outService)
{
    const sp<IServiceManager> sm = defaultServiceManager();
    if (sm != NULL) {
        *outService = interface_cast<INTERFACE>(sm->getService(name));
        if ((*outService) != NULL) return NO_ERROR;
    }
    return NAME_NOT_FOUND;
}

到第四行这里跟之前启动过程中SurfaceFlinger的创建一模一样了，最后outService会被强转化成BpSurfaceComposer类，BpSurfaceComposer就是SurfaceFlinger在客户端这边的Binder代理，与之对应的是BnSurfaceComposer相当于SurfaceFlinger这边与BpSurfaceComposer进行交互的一方。获得了代理之后，我们再回到之前的onFirstRef函数：

void SurfaceComposerClient::onFirstRef() {
    sp<ISurfaceComposer> sm(ComposerService::getComposerService());
    if (sm != 0) {
        sp<ISurfaceComposerClient> conn = sm->createConnection();
        if (conn != 0) {
            mClient = conn;
            mStatus = NO_ERROR;
        }
    }
}

这时候走到if，sm有值了，所以往下走，看名字可以得出这是客户端在请求连接到SurfaceFlinger服务。我们跟踪一下createConnection：

 

virtual sp<ISurfaceComposerClient> createConnection()
    {
        uint32_t n;
        Parcel data, reply;
        data.writeInterfaceToken(ISurfaceComposer::getInterfaceDescriptor());
        remote()->transact(BnSurfaceComposer::CREATE_CONNECTION, data, &reply);
        return interface_cast<ISurfaceComposerClient>(reply.readStrongBinder());
    }

函数的大概意思是把一些连接信息写到数据包data里，然后通过transact这个函数传出去，remote（）之前有研究过他是IBinder类的，这里重点看transact这个函数，他是IBinder类的成员函数故它的实现必定在BpBinder类里，我们跳到这个函数的实体：

 

status_t BpBinder::transact(
    uint32_t code, const Parcel& data, Parcel* reply, uint32_t flags)
{
    // Once a binder has died, it will never come back to life.
    if (mAlive) {
        status_t status = IPCThreadState::self()->transact(
            mHandle, code, data, reply, flags);
        if (status == DEAD_OBJECT) mAlive = 0;
        return status;
    }

    return DEAD_OBJECT;
}

发现他又调用了transact这个函数，不同的是 是在IPCThreadState类中的 我们继续跟踪：

status_t IPCThreadState::transact(int32_t handle,
                                  uint32_t code, const Parcel& data,
                                  Parcel* reply, uint32_t flags)
{
    status_t err = data.errorCheck();

    flags |= TF_ACCEPT_FDS;

    IF_LOG_TRANSACTIONS() {
        TextOutput::Bundle _b(alog);
        alog << "BC_TRANSACTION thr " << (void*)pthread_self() << " / hand "
            << handle << " / code " << TypeCode(code) << ": "
            << indent << data << dedent << endl;
    }
    
    if (err == NO_ERROR) {
        LOG_ONEWAY(">>>> SEND from pid %d uid %d %s", getpid(), getuid(),
            (flags & TF_ONE_WAY) == 0 ? "READ REPLY" : "ONE WAY");
        err = writeTransactionData(BC_TRANSACTION, flags, handle, code, data, NULL);
    }
    
    if (err != NO_ERROR) {
        if (reply) reply->setError(err);
        return (mLastError = err);
    }
    
    if ((flags & TF_ONE_WAY) == 0) {
        #if 0
        if (code == 4) { // relayout
            ALOGI(">>>>>> CALLING transaction 4");
        } else {
            ALOGI(">>>>>> CALLING transaction %d", code);
        }
        #endif
        if (reply) {
            err = waitForResponse(reply);
        } else {
            Parcel fakeReply;
            err = waitForResponse(&fakeReply);
        }
        #if 0
        if (code == 4) { // relayout
            ALOGI("<<<<<< RETURNING transaction 4");
        } else {
            ALOGI("<<<<<< RETURNING transaction %d", code);
        }
        #endif
        
        IF_LOG_TRANSACTIONS() {
            TextOutput::Bundle _b(alog);
            alog << "BR_REPLY thr " << (void*)pthread_self() << " / hand "
                << handle << ": ";
            if (reply) alog << indent << *reply << dedent << endl;
            else alog << "(none requested)" << endl;
        }
    } else {
        err = waitForResponse(NULL, NULL);
    }
    
    return err;
}

这里已经扯到Binder机制的基本原理了，简单说就是客户端的SF代理通过BpBinder跟/dev/binder设备进行交互，向binder里写东西，函数err = writeTransactionData(BC_TRANSACTION, flags, handle, code, data, NULL);就是实现了这个功能，写完后他又等待服务端（SurfaceFlinger）那端的回应，err = waitForResponse(reply);当然服务端那边同样有人跟他进行交互。到这里，如果服务端收到消息并且返回了一个消息给客户端，这说明客户端请求连接成功。

     请求成功后conn就不为0了，回到onFirstRef，成功后把coon赋给mClient。

既然客户端有请求了，那么服务端肯定有东西会去留意这个消息，回到更以前，做出向服务端请求连接这件事其实是SurfaceFlinger在客户端这边的代理BpSurfaceComposer完成的，那么之前说过，与之对应的就是BnSurfaceComposer。我们看一下他类的定义：

 

class BnSurfaceComposer: public BnInterface<ISurfaceComposer> {
public:
    enum {
        // Note: BOOT_FINISHED must remain this value, it is called from
        // Java by ActivityManagerService.
        BOOT_FINISHED = IBinder::FIRST_CALL_TRANSACTION,
        CREATE_CONNECTION,
        CREATE_GRAPHIC_BUFFER_ALLOC,
        CREATE_DISPLAY_EVENT_CONNECTION,
        CREATE_DISPLAY,
        DESTROY_DISPLAY,
        GET_BUILT_IN_DISPLAY,
        SET_TRANSACTION_STATE,
        AUTHENTICATE_SURFACE,
        BLANK,
        UNBLANK,
        GET_DISPLAY_INFO,
        CONNECT_DISPLAY,
        CAPTURE_SCREEN,
    };

    virtual status_t onTransact(uint32_t code, const Parcel& data,
            Parcel* reply, uint32_t flags = 0);
 ｝

发现他没有构造函数只有一个成员函数 onTransact。跟踪一下：

status_t BnSurfaceComposer::onTransact(
    uint32_t code, const Parcel& data, Parcel* reply, uint32_t flags)
{
    switch(code) {
        case CREATE_CONNECTION: {
            CHECK_INTERFACE(ISurfaceComposer, data, reply);
            sp<IBinder> b = createConnection()->asBinder();
            reply->writeStrongBinder(b);
            return NO_ERROR;
        }
        case CREATE_GRAPHIC_BUFFER_ALLOC: {
            CHECK_INTERFACE(ISurfaceComposer, data, reply);
            sp<IBinder> b = createGraphicBufferAlloc()->asBinder();
            reply->writeStrongBinder(b);
            return NO_ERROR;
        }
        case SET_TRANSACTION_STATE: {
            CHECK_INTERFACE(ISurfaceComposer, data, reply);
            size_t count = data.readInt32();
            ComposerState s;
            Vector<ComposerState> state;
            state.setCapacity(count);
            for (size_t i=0 ; i<count ; i++) {
                s.read(data);
                state.add(s);
            }
            count = data.readInt32();
            DisplayState d;
            Vector<DisplayState> displays;
            displays.setCapacity(count);
            for (size_t i=0 ; i<count ; i++) {
                d.read(data);
                displays.add(d);
            }
            uint32_t flags = data.readInt32();
            setTransactionState(state, displays, flags);
            return NO_ERROR;
        }
        case BOOT_FINISHED: {
            CHECK_INTERFACE(ISurfaceComposer, data, reply);
            bootFinished();
            return NO_ERROR;
        }
        case CAPTURE_SCREEN: {
            CHECK_INTERFACE(ISurfaceComposer, data, reply);
            sp<IBinder> display = data.readStrongBinder();
            sp<IGraphicBufferProducer> producer =
                    interface_cast<IGraphicBufferProducer>(data.readStrongBinder());
            uint32_t reqWidth = data.readInt32();
            uint32_t reqHeight = data.readInt32();
            uint32_t minLayerZ = data.readInt32();
            uint32_t maxLayerZ = data.readInt32();
            status_t res = captureScreen(display, producer,
                    reqWidth, reqHeight, minLayerZ, maxLayerZ);
            reply->writeInt32(res);
            return NO_ERROR;
        }
        case AUTHENTICATE_SURFACE: {
            CHECK_INTERFACE(ISurfaceComposer, data, reply);
            sp<IGraphicBufferProducer> bufferProducer =
                    interface_cast<IGraphicBufferProducer>(data.readStrongBinder());
            int32_t result = authenticateSurfaceTexture(bufferProducer) ? 1 : 0;
            reply->writeInt32(result);
            return NO_ERROR;
        }
        case CREATE_DISPLAY_EVENT_CONNECTION: {
            CHECK_INTERFACE(ISurfaceComposer, data, reply);
            sp<IDisplayEventConnection> connection(createDisplayEventConnection());
            reply->writeStrongBinder(connection->asBinder());
            return NO_ERROR;
        }
        case CREATE_DISPLAY: {
            CHECK_INTERFACE(ISurfaceComposer, data, reply);
            String8 displayName = data.readString8();
            bool secure = bool(data.readInt32());
            sp<IBinder> display(createDisplay(displayName, secure));
            reply->writeStrongBinder(display);
            return NO_ERROR;
        }
        case DESTROY_DISPLAY: {
            CHECK_INTERFACE(ISurfaceComposer, data, reply);
            sp<IBinder> display = data.readStrongBinder();
            destroyDisplay(display);
            return NO_ERROR;
        }
        case GET_BUILT_IN_DISPLAY: {
            CHECK_INTERFACE(ISurfaceComposer, data, reply);
            int32_t id = data.readInt32();
            sp<IBinder> display(getBuiltInDisplay(id));
            reply->writeStrongBinder(display);
            return NO_ERROR;
        }
        case BLANK: {
            CHECK_INTERFACE(ISurfaceComposer, data, reply);
            sp<IBinder> display = data.readStrongBinder();
            blank(display);
            return NO_ERROR;
        }
        case UNBLANK: {
            CHECK_INTERFACE(ISurfaceComposer, data, reply);
            sp<IBinder> display = data.readStrongBinder();
            unblank(display);
            return NO_ERROR;
        }
        case GET_DISPLAY_INFO: {
            CHECK_INTERFACE(ISurfaceComposer, data, reply);
            DisplayInfo info;
            sp<IBinder> display = data.readStrongBinder();
            status_t result = getDisplayInfo(display, &info);
            memcpy(reply->writeInplace(sizeof(DisplayInfo)), &info, sizeof(DisplayInfo));
            reply->writeInt32(result);
            return NO_ERROR;
        }
        default: {
            return BBinder::onTransact(code, data, reply, flags);
        }
    }
    // should be unreachable
    return NO_ERROR;
}

代码虽然有点长，但是整体框架就是一个swith语句，看到第一个分支，马上就能看出来这是一个对请求连接信号的处理，这里大概能知道这个onTransact的功能就是判断收到的消息即参数code是哪种类型从而做出相应的处理。在server测，一开始创建SF服务的时候，在线城池创建的时候就已经开启线程去监听binder设备了我们来跟踪下：

 从最开始SF服务启动的源文件开始：

int main(int argc, char** argv) {
    // When SF is launched in its own process, limit the number of
    // binder threads to 4.
    ProcessState::self()->setThreadPoolMaxThreadCount(4);

    // start the thread pool
    sp<ProcessState> ps(ProcessState::self());
    ps->startThreadPool();

是一个片段，最后一句就是开始线程池。跟踪其代码：

void ProcessState::startThreadPool()
{
    AutoMutex _l(mLock);
    if (!mThreadPoolStarted) {
        mThreadPoolStarted = true;
        spawnPooledThread(true);
    }
}

核心也在最后一句，

void ProcessState::spawnPooledThread(bool isMain)
{
    if (mThreadPoolStarted) {
        String8 name = makeBinderThreadName();
        ALOGV("Spawning new pooled thread, name=%s\n", name.string());
        sp<Thread> t = new PoolThread(isMain);
        t->run(name.string());
    }
}

到这里，可以看到实质的线程创建和开始运行。接下来我用一张图来表示他的函数调用的走向：

这里一旦检测到有连接请求消息就会跳到executeCommand的switch分支BR_TRANSACTION，去执行BBinder的transact，transact函数会调用onTransact函数，这个onTransact由其子类实现，在这里由BnSurfaceComposer完成，即调用的是BnSurfaceComposer::onTransact（），函数的具体内容在上面已贴出，就是处理不同的请求。

     到此为止，客户端已经与SF进行连接，也就是交手过了，那么自然连接了，一定要做些事情比如客户端请求渲染一个surface，或者等等其他的。接下去的工作就是SurfaceFlinger的事，对客户端不同的请求而进行不同的处理，这才是SF核心工作所在。当然所有的通信机制都是基于binder机制，而求也有负责这个方面的客户端这边的binder代理BpSurfaceComposerClient和服务端这边的本地对象 BnSurfaceComposerClient，而这里有点不一样的是BnSurfaceComposerClient派生出Client所以事情都交与Client做了。