Android系统进程间通信（IPC）机制Binder中的Server启动过程源代码分析

From:http://blog.csdn.net/luoshengyang/article/details/6629298

在前面一篇文章浅谈Android系统进程间通信（IPC）机制Binder中的Server和Client获得Service Manager接口之路中，介绍了在Android系统中Binder进程间通信机制中的Server角色是如何获得Service Manager远程接口的，即defaultServiceManager函数的实现。Server获得了Service Manager远程接口之后，就要把自己的Service添加到Service Manager中去，然后把自己启动起来，等待Client的请求。本文将通过分析源代码了解Server的启动过程是怎么样的。

        本文通过一个具体的例子来说明Binder机制中Server的启动过程。我们知道，在Android系统中，提供了多媒体播放的功能，这个功能是以服务的形式来提供的。这里，我们就通过分析MediaPlayerService的实现来了解Media Server的启动过程。

        首先，看一下MediaPlayerService的类图，以便我们理解下面要描述的内容。

        我们将要介绍的主角MediaPlayerService继承于BnMediaPlayerService类，熟悉Binder机制的同学应该知道BnMediaPlayerService是一个Binder Native类，用来处理Client请求的。BnMediaPlayerService继承于BnInterface<IMediaPlayerService>类，BnInterface是一个模板类，它定义在frameworks/base/include/binder/IInterface.h文件中：

[cpp] view plain copy print ?

1. template<typename INTERFACE>  
2. class BnInterface : public INTERFACE, public BBinder  
3. {  
4. public:  
5.     virtual sp<IInterface>      queryLocalInterface(const String16& _descriptor);  
6.     virtual const String16&     getInterfaceDescriptor() const;  
7.   
8. protected:  
9.     virtual IBinder*            onAsBinder();  
10. };  

template<typename INTERFACE>
class BnInterface : public INTERFACE, public BBinder
{
public:
    virtual sp<IInterface>      queryLocalInterface(const String16& _descriptor);
    virtual const String16&     getInterfaceDescriptor() const;

protected:
    virtual IBinder*            onAsBinder();
};

       这里可以看出，BnMediaPlayerService实际是继承了IMediaPlayerService和BBinder类。IMediaPlayerService和BBinder类又分别继承了IInterface和IBinder类，IInterface和IBinder类又同时继承了RefBase类。

       实际上，BnMediaPlayerService并不是直接接收到Client处发送过来的请求，而是使用了IPCThreadState接收Client处发送过来的请求，而IPCThreadState又借助了ProcessState类来与Binder驱动程序交互。有关IPCThreadState和ProcessState的关系，可以参考上一篇文章浅谈Android系统进程间通信（IPC）机制Binder中的Server和Client获得Service Manager接口之路，接下来也会有相应的描述。IPCThreadState接收到了Client处的请求后，就会调用BBinder类的transact函数，并传入相关参数，BBinder类的transact函数最终调用BnMediaPlayerService类的onTransact函数，于是，就开始真正地处理Client的请求了。

      了解了MediaPlayerService类结构之后，就要开始进入到本文的主题了。

      首先，看看MediaPlayerService是如何启动的。启动MediaPlayerService的代码位于frameworks/base/media/mediaserver/main_mediaserver.cpp文件中：

[cpp] view plain copy print ?

1. int main(int argc, char** argv)  
2. {  
3.     sp<ProcessState> proc(ProcessState::self());  
4.     sp<IServiceManager> sm = defaultServiceManager();  
5.     LOGI("ServiceManager: %p", sm.get());  
6.     AudioFlinger::instantiate();  
7.     MediaPlayerService::instantiate();  
8.     CameraService::instantiate();  
9.     AudioPolicyService::instantiate();  
10.     ProcessState::self()->startThreadPool();  
11.     IPCThreadState::self()->joinThreadPool();  
12. }  

int main(int argc, char** argv)
{
    sp<ProcessState> proc(ProcessState::self());
    sp<IServiceManager> sm = defaultServiceManager();
    LOGI("ServiceManager: %p", sm.get());
    AudioFlinger::instantiate();
    MediaPlayerService::instantiate();
    CameraService::instantiate();
    AudioPolicyService::instantiate();
    ProcessState::self()->startThreadPool();
    IPCThreadState::self()->joinThreadPool();
}

       这里我们不关注AudioFlinger和CameraService相关的代码。

       先看下面这句代码：

[cpp] view plain copy print ?

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

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

       这句代码的作用是通过ProcessState::self()调用创建一个ProcessState实例。ProcessState::self()是ProcessState类的一个静态成员变量，定义在frameworks/base/libs/binder/ProcessState.cpp文件中：

[cpp] view plain copy print ?

1. sp<ProcessState> ProcessState::self()  
2. {  
3.     if (gProcess != NULL) return gProcess;  
4.       
5.     AutoMutex _l(gProcessMutex);  
6.     if (gProcess == NULL) gProcess = new ProcessState;  
7.     return gProcess;  
8. }  

sp<ProcessState> ProcessState::self()
{
    if (gProcess != NULL) return gProcess;
    
    AutoMutex _l(gProcessMutex);
    if (gProcess == NULL) gProcess = new ProcessState;
    return gProcess;
}

       这里可以看出，这个函数作用是返回一个全局唯一的ProcessState实例gProcess。全局唯一实例变量gProcess定义在frameworks/base/libs/binder/Static.cpp文件中：

[cpp] view plain copy print ?

1. Mutex gProcessMutex;  
2. sp<ProcessState> gProcess;  

Mutex gProcessMutex;
sp<ProcessState> gProcess;

       再来看ProcessState的构造函数：

[cpp] view plain copy print ?

1. ProcessState::ProcessState()  
2.     : mDriverFD(open_driver())  
3.     , mVMStart(MAP_FAILED)  
4.     , mManagesContexts(false)  
5.     , mBinderContextCheckFunc(NULL)  
6.     , mBinderContextUserData(NULL)  
7.     , mThreadPoolStarted(false)  
8.     , mThreadPoolSeq(1)  
9. {  
10.     if (mDriverFD >= 0) {  
11.         // XXX Ideally, there should be a specific define for whether we   
12.         // have mmap (or whether we could possibly have the kernel module   
13.         // availabla).   
14. #if !defined(HAVE_WIN32_IPC)   
15.         // mmap the binder, providing a chunk of virtual address space to receive transactions.   
16.         mVMStart = mmap(0, BINDER_VM_SIZE, PROT_READ, MAP_PRIVATE | MAP_NORESERVE, mDriverFD, 0);  
17.         if (mVMStart == MAP_FAILED) {  
18.             // *sigh*   
19.             LOGE("Using /dev/binder failed: unable to mmap transaction memory.\n");  
20.             close(mDriverFD);  
21.             mDriverFD = -1;  
22.         }  
23. #else   
24.         mDriverFD = -1;  
25. #endif   
26.     }  
27.     if (mDriverFD < 0) {  
28.         // Need to run without the driver, starting our own thread pool.   
29.     }  
30. }  

ProcessState::ProcessState()
    : mDriverFD(open_driver())
    , mVMStart(MAP_FAILED)
    , mManagesContexts(false)
    , mBinderContextCheckFunc(NULL)
    , mBinderContextUserData(NULL)
    , mThreadPoolStarted(false)
    , mThreadPoolSeq(1)
{
    if (mDriverFD >= 0) {
        // XXX Ideally, there should be a specific define for whether we
        // have mmap (or whether we could possibly have the kernel module
        // availabla).
#if !defined(HAVE_WIN32_IPC)
        // mmap the binder, providing a chunk of virtual address space to receive transactions.
        mVMStart = mmap(0, BINDER_VM_SIZE, PROT_READ, MAP_PRIVATE | MAP_NORESERVE, mDriverFD, 0);
        if (mVMStart == MAP_FAILED) {
            // *sigh*
            LOGE("Using /dev/binder failed: unable to mmap transaction memory.\n");
            close(mDriverFD);
            mDriverFD = -1;
        }
#else
        mDriverFD = -1;
#endif
    }
    if (mDriverFD < 0) {
        // Need to run without the driver, starting our own thread pool.
    }
}

        这个函数有两个关键地方，一是通过open_driver函数打开Binder设备文件/dev/binder，并将打开设备文件描述符保存在成员变量mDriverFD中；二是通过mmap来把设备文件/dev/binder映射到内存中。

        先看open_driver函数的实现，这个函数同样位于frameworks/base/libs/binder/ProcessState.cpp文件中：

[cpp] view plain copy print ?

1. static int open_driver()  
2. {  
3.     if (gSingleProcess) {  
4.         return -1;  
5.     }  
6.   
7.     int fd = open("/dev/binder", O_RDWR);  
8.     if (fd >= 0) {  
9.         fcntl(fd, F_SETFD, FD_CLOEXEC);  
10.         int vers;  
11. #if defined(HAVE_ANDROID_OS)   
12.         status_t result = ioctl(fd, BINDER_VERSION, &vers);  
13. #else   
14.         status_t result = -1;  
15.         errno = EPERM;  
16. #endif   
17.         if (result == -1) {  
18.             LOGE("Binder ioctl to obtain version failed: %s", strerror(errno));  
19.             close(fd);  
20.             fd = -1;  
21.         }  
22.         if (result != 0 || vers != BINDER_CURRENT_PROTOCOL_VERSION) {  
23.             LOGE("Binder driver protocol does not match user space protocol!");  
24.             close(fd);  
25.             fd = -1;  
26.         }  
27. #if defined(HAVE_ANDROID_OS)   
28.         size_t maxThreads = 15;  
29.         result = ioctl(fd, BINDER_SET_MAX_THREADS, &maxThreads);  
30.         if (result == -1) {  
31.             LOGE("Binder ioctl to set max threads failed: %s", strerror(errno));  
32.         }  
33. #endif   
34.           
35.     } else {  
36.         LOGW("Opening '/dev/binder' failed: %s\n", strerror(errno));  
37.     }  
38.     return fd;  
39. }  

static int open_driver()
{
    if (gSingleProcess) {
        return -1;
    }

    int fd = open("/dev/binder", O_RDWR);
    if (fd >= 0) {
        fcntl(fd, F_SETFD, FD_CLOEXEC);
        int vers;
#if defined(HAVE_ANDROID_OS)
        status_t result = ioctl(fd, BINDER_VERSION, &vers);
#else
        status_t result = -1;
        errno = EPERM;
#endif
        if (result == -1) {
            LOGE("Binder ioctl to obtain version failed: %s", strerror(errno));
            close(fd);
            fd = -1;
        }
        if (result != 0 || vers != BINDER_CURRENT_PROTOCOL_VERSION) {
            LOGE("Binder driver protocol does not match user space protocol!");
            close(fd);
            fd = -1;
        }
#if defined(HAVE_ANDROID_OS)
        size_t maxThreads = 15;
        result = ioctl(fd, BINDER_SET_MAX_THREADS, &maxThreads);
        if (result == -1) {
            LOGE("Binder ioctl to set max threads failed: %s", strerror(errno));
        }
#endif
        
    } else {
        LOGW("Opening '/dev/binder' failed: %s\n", strerror(errno));
    }
    return fd;
}

        这个函数的作用主要是通过open文件操作函数来打开/dev/binder设备文件，然后再调用ioctl文件控制函数来分别执行BINDER_VERSION和BINDER_SET_MAX_THREADS两个命令来和Binder驱动程序进行交互，前者用于获得当前Binder驱动程序的版本号，后者用于通知Binder驱动程序，MediaPlayerService最多可同时启动15个线程来处理Client端的请求。

        open在Binder驱动程序中的具体实现，请参考前面一篇文章浅谈Service Manager成为Android进程间通信（IPC）机制Binder守护进程之路，这里不再重复描述。打开/dev/binder设备文件后，Binder驱动程序就为MediaPlayerService进程创建了一个struct binder_proc结构体实例来维护MediaPlayerService进程上下文相关信息。

        我们来看一下ioctl文件操作函数执行BINDER_VERSION命令的过程：

[cpp] view plain copy print ?

1. status_t result = ioctl(fd, BINDER_VERSION, &vers);  

status_t result = ioctl(fd, BINDER_VERSION, &vers);

        这个函数调用最终进入到Binder驱动程序的binder_ioctl函数中，我们只关注BINDER_VERSION相关的部分逻辑：

[cpp] view plain copy print ?

1. static long binder_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)  
2. {  
3.     int ret;  
4.     struct binder_proc *proc = filp->private_data;  
5.     struct binder_thread *thread;  
6.     unsigned int size = _IOC_SIZE(cmd);  
7.     void __user *ubuf = (void __user *)arg;  
8.   
9.     /*printk(KERN_INFO "binder_ioctl: %d:%d %x %lx\n", proc->pid, current->pid, cmd, arg);*/  
10.   
11.     ret = wait_event_interruptible(binder_user_error_wait, binder_stop_on_user_error < 2);  
12.     if (ret)  
13.         return ret;  
14.   
15.     mutex_lock(&binder_lock);  
16.     thread = binder_get_thread(proc);  
17.     if (thread == NULL) {  
18.         ret = -ENOMEM;  
19.         goto err;  
20.     }  
21.   
22.     switch (cmd) {  
23.     ......  
24.     case BINDER_VERSION:  
25.         if (size != sizeof(struct binder_version)) {  
26.             ret = -EINVAL;  
27.             goto err;  
28.         }  
29.         if (put_user(BINDER_CURRENT_PROTOCOL_VERSION, &((struct binder_version *)ubuf)->protocol_version)) {  
30.             ret = -EINVAL;  
31.             goto err;  
32.         }  
33.         break;  
34.     ......  
35.     }  
36.     ret = 0;  
37. err:  
38.         ......  
39.     return ret;  
40. }  

static long binder_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
{
	int ret;
	struct binder_proc *proc = filp->private_data;
	struct binder_thread *thread;
	unsigned int size = _IOC_SIZE(cmd);
	void __user *ubuf = (void __user *)arg;

	/*printk(KERN_INFO "binder_ioctl: %d:%d %x %lx\n", proc->pid, current->pid, cmd, arg);*/

	ret = wait_event_interruptible(binder_user_error_wait, binder_stop_on_user_error < 2);
	if (ret)
		return ret;

	mutex_lock(&binder_lock);
	thread = binder_get_thread(proc);
	if (thread == NULL) {
		ret = -ENOMEM;
		goto err;
	}

	switch (cmd) {
	......
	case BINDER_VERSION:
		if (size != sizeof(struct binder_version)) {
			ret = -EINVAL;
			goto err;
		}
		if (put_user(BINDER_CURRENT_PROTOCOL_VERSION, &((struct binder_version *)ubuf)->protocol_version)) {
			ret = -EINVAL;
			goto err;
		}
		break;
	......
	}
	ret = 0;
err:
        ......
	return ret;
}

        很简单，只是将BINDER_CURRENT_PROTOCOL_VERSION写入到传入的参数arg指向的用户缓冲区中去就返回了。BINDER_CURRENT_PROTOCOL_VERSION是一个宏，定义在kernel/common/drivers/staging/android/binder.h文件中：

[cpp] view plain copy print ?

1. /* This is the current protocol version. */  
2. #define BINDER_CURRENT_PROTOCOL_VERSION 7  

/* This is the current protocol version. */
#define BINDER_CURRENT_PROTOCOL_VERSION 7

       这里为什么要把ubuf转换成struct binder_version之后，再通过其protocol_version成员变量再来写入呢，转了一圈，最终内容还是写入到ubuf中。我们看一下struct binder_version的定义就会明白，同样是在kernel/common/drivers/staging/android/binder.h文件中：

[cpp] view plain copy print ?

1. /* Use with BINDER_VERSION, driver fills in fields. */  
2. struct binder_version {  
3.     /* driver protocol version -- increment with incompatible change */  
4.     signed long protocol_version;  
5. };  

/* Use with BINDER_VERSION, driver fills in fields. */
struct binder_version {
	/* driver protocol version -- increment with incompatible change */
	signed long	protocol_version;
};

        从注释中可以看出来，这里是考虑到兼容性，因为以后很有可能不是用signed long来表示版本号。

        这里有一个重要的地方要注意的是，由于这里是打开设备文件/dev/binder之后，第一次进入到binder_ioctl函数，因此，这里调用binder_get_thread的时候，就会为当前线程创建一个struct binder_thread结构体变量来维护线程上下文信息，具体可以参考浅谈Service Manager成为Android进程间通信（IPC）机制Binder守护进程之路一文。

        接着我们再来看一下ioctl文件操作函数执行BINDER_SET_MAX_THREADS命令的过程：

[cpp] view plain copy print ?

1. result = ioctl(fd, BINDER_SET_MAX_THREADS, &maxThreads);  

result = ioctl(fd, BINDER_SET_MAX_THREADS, &maxThreads);

        这个函数调用最终进入到Binder驱动程序的binder_ioctl函数中，我们只关注BINDER_SET_MAX_THREADS相关的部分逻辑：

[cpp] view plain copy print ?

1. static long binder_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)  
2. {  
3.     int ret;  
4.     struct binder_proc *proc = filp->private_data;  
5.     struct binder_thread *thread;  
6.     unsigned int size = _IOC_SIZE(cmd);  
7.     void __user *ubuf = (void __user *)arg;  
8.   
9.     /*printk(KERN_INFO "binder_ioctl: %d:%d %x %lx\n", proc->pid, current->pid, cmd, arg);*/  
10.   
11.     ret = wait_event_interruptible(binder_user_error_wait, binder_stop_on_user_error < 2);  
12.     if (ret)  
13.         return ret;  
14.   
15.     mutex_lock(&binder_lock);  
16.     thread = binder_get_thread(proc);  
17.     if (thread == NULL) {  
18.         ret = -ENOMEM;  
19.         goto err;  
20.     }  
21.   
22.     switch (cmd) {  
23.     ......  
24.     case BINDER_SET_MAX_THREADS:  
25.         if (copy_from_user(&proc->max_threads, ubuf, sizeof(proc->max_threads))) {  
26.             ret = -EINVAL;  
27.             goto err;  
28.         }  
29.         break;  
30.     ......  
31.     }  
32.     ret = 0;  
33. err:  
34.     ......  
35.     return ret;  
36. }  

static long binder_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
{
	int ret;
	struct binder_proc *proc = filp->private_data;
	struct binder_thread *thread;
	unsigned int size = _IOC_SIZE(cmd);
	void __user *ubuf = (void __user *)arg;

	/*printk(KERN_INFO "binder_ioctl: %d:%d %x %lx\n", proc->pid, current->pid, cmd, arg);*/

	ret = wait_event_interruptible(binder_user_error_wait, binder_stop_on_user_error < 2);
	if (ret)
		return ret;

	mutex_lock(&binder_lock);
	thread = binder_get_thread(proc);
	if (thread == NULL) {
		ret = -ENOMEM;
		goto err;
	}

	switch (cmd) {
	......
	case BINDER_SET_MAX_THREADS:
		if (copy_from_user(&proc->max_threads, ubuf, sizeof(proc->max_threads))) {
			ret = -EINVAL;
			goto err;
		}
		break;
	......
	}
	ret = 0;
err:
	......
	return ret;
}

        这里实现也是非常简单，只是简单地把用户传进来的参数保存在proc->max_threads中就完毕了。注意，这里再调用binder_get_thread函数的时候，就可以在proc->threads中找到当前线程对应的struct binder_thread结构了，因为前面已经创建好并保存在proc->threads红黑树中。

        回到ProcessState的构造函数中，这里还通过mmap函数来把设备文件/dev/binder映射到内存中，这个函数在浅谈Service Manager成为Android进程间通信（IPC）机制Binder守护进程之路一文也已经有详细介绍，这里不再重复描述。宏BINDER_VM_SIZE就定义在ProcessState.cpp文件中：

[cpp] view plain copy print ?

1. #define BINDER_VM_SIZE ((1*1024*1024) - (4096 *2))  

#define BINDER_VM_SIZE ((1*1024*1024) - (4096 *2))

        mmap函数调用完成之后，Binder驱动程序就为当前进程预留了BINDER_VM_SIZE大小的内存空间了。

        这样，ProcessState全局唯一变量gProcess就创建完毕了，回到frameworks/base/media/mediaserver/main_mediaserver.cpp文件中的main函数，下一步是调用defaultServiceManager函数来获得Service Manager的远程接口，这个已经在上一篇文章浅谈Android系统进程间通信（IPC）机制Binder中的Server和Client获得Service Manager接口之路有详细描述，读者可以回过头去参考一下。

        再接下来，就进入到MediaPlayerService::instantiate函数把MediaPlayerService添加到Service Manger中去了。这个函数定义在frameworks/base/media/libmediaplayerservice/MediaPlayerService.cpp文件中：

[cpp] view plain copy print ?

1. void MediaPlayerService::instantiate() {  
2.     defaultServiceManager()->addService(  
3.             String16("media.player"), new MediaPlayerService());  
4. }  

void MediaPlayerService::instantiate() {
    defaultServiceManager()->addService(
            String16("media.player"), new MediaPlayerService());
}

        我们重点看一下IServiceManger::addService的过程，这有助于我们加深对Binder机制的理解。

        在上一篇文章浅谈Android系统进程间通信（IPC）机制Binder中的Server和Client获得Service Manager接口之路中说到，defaultServiceManager返回的实际是一个BpServiceManger类实例，因此，我们看一下BpServiceManger::addService的实现，这个函数实现在frameworks/base/libs/binder/IServiceManager.cpp文件中：

[cpp] view plain copy print ?

1. class BpServiceManager : public BpInterface<IServiceManager>  
2. {  
3. public:  
4.     BpServiceManager(const sp<IBinder>& impl)  
5.         : BpInterface<IServiceManager>(impl)  
6.     {  
7.     }  
8.   
9.     ......  
10.   
11.     virtual status_t addService(const String16& name, const sp<IBinder>& service)  
12.     {  
13.         Parcel data, reply;  
14.         data.writeInterfaceToken(IServiceManager::getInterfaceDescriptor());  
15.         data.writeString16(name);  
16.         data.writeStrongBinder(service);  
17.         status_t err = remote()->transact(ADD_SERVICE_TRANSACTION, data, &reply);  
18.         return err == NO_ERROR ? reply.readExceptionCode()   
19.     }  
20.   
21.     ......  
22.   
23. };  

class BpServiceManager : public BpInterface<IServiceManager>
{
public:
	BpServiceManager(const sp<IBinder>& impl)
		: BpInterface<IServiceManager>(impl)
	{
	}

	......

	virtual status_t addService(const String16& name, const sp<IBinder>& service)
	{
		Parcel data, reply;
		data.writeInterfaceToken(IServiceManager::getInterfaceDescriptor());
		data.writeString16(name);
		data.writeStrongBinder(service);
		status_t err = remote()->transact(ADD_SERVICE_TRANSACTION, data, &reply);
		return err == NO_ERROR ? reply.readExceptionCode() 
	}

	......

};

 

         这里的Parcel类是用来于序列化进程间通信数据用的。

         先来看这一句的调用：

[cpp] view plain copy print ?

1. data.writeInterfaceToken(IServiceManager::getInterfaceDescriptor());  

data.writeInterfaceToken(IServiceManager::getInterfaceDescriptor());

         IServiceManager::getInterfaceDescriptor()返回来的是一个字符串，即"android.os.IServiceManager"，具体可以参考IServiceManger的实现。我们看一下Parcel::writeInterfaceToken的实现，位于frameworks/base/libs/binder/Parcel.cpp文件中：

[cpp] view plain copy print ?

1. // Write RPC headers.  (previously just the interface token)   
2. status_t Parcel::writeInterfaceToken(const String16& interface)  
3. {  
4.     writeInt32(IPCThreadState::self()->getStrictModePolicy() |  
5.                STRICT_MODE_PENALTY_GATHER);  
6.     // currently the interface identification token is just its name as a string   
7.     return writeString16(interface);  
8. }  

// Write RPC headers.  (previously just the interface token)
status_t Parcel::writeInterfaceToken(const String16& interface)
{
    writeInt32(IPCThreadState::self()->getStrictModePolicy() |
               STRICT_MODE_PENALTY_GATHER);
    // currently the interface identification token is just its name as a string
    return writeString16(interface);
}

         它的作用是写入一个整数和一个字符串到Parcel中去。

         再来看下面的调用：

[cpp] view plain copy print ?

1. data.writeString16(name);  

data.writeString16(name);

        这里又是写入一个字符串到Parcel中去，这里的name即是上面传进来的“media.player”字符串。

        往下看：

[cpp] view plain copy print ?

1. data.writeStrongBinder(service);  

data.writeStrongBinder(service);

        这里定入一个Binder对象到Parcel去。我们重点看一下这个函数的实现，因为它涉及到进程间传输Binder实体的问题，比较复杂，需要重点关注，同时，也是理解Binder机制的一个重点所在。注意，这里的service参数是一个MediaPlayerService对象。

[cpp] view plain copy print ?

1. status_t Parcel::writeStrongBinder(const sp<IBinder>& val)  
2. {  
3.     return flatten_binder(ProcessState::self(), val, this);  
4. }  

status_t Parcel::writeStrongBinder(const sp<IBinder>& val)
{
    return flatten_binder(ProcessState::self(), val, this);
}

        看到flatten_binder函数，是不是似曾相识的感觉？我们在前面一篇文章浅谈Service Manager成为Android进程间通信（IPC）机制Binder守护进程之路中，曾经提到在Binder驱动程序中，使用struct flat_binder_object来表示传输中的一个binder对象，它的定义如下所示：

[cpp] view plain copy print ?

1. /* 
2.  * This is the flattened representation of a Binder object for transfer 
3.  * between processes.  The 'offsets' supplied as part of a binder transaction 
4.  * contains offsets into the data where these structures occur.  The Binder 
5.  * driver takes care of re-writing the structure type and data as it moves 
6.  * between processes. 
7.  */  
8. struct flat_binder_object {  
9.     /* 8 bytes for large_flat_header. */  
10.     unsigned long       type;  
11.     unsigned long       flags;  
12.   
13.     /* 8 bytes of data. */  
14.     union {  
15.         void        *binder;    /* local object */  
16.         signed long handle;     /* remote object */  
17.     };  
18.   
19.     /* extra data associated with local object */  
20.     void            *cookie;  
21. };  

/*
 * This is the flattened representation of a Binder object for transfer
 * between processes.  The 'offsets' supplied as part of a binder transaction
 * contains offsets into the data where these structures occur.  The Binder
 * driver takes care of re-writing the structure type and data as it moves
 * between processes.
 */
struct flat_binder_object {
	/* 8 bytes for large_flat_header. */
	unsigned long		type;
	unsigned long		flags;

	/* 8 bytes of data. */
	union {
		void		*binder;	/* local object */
		signed long	handle;		/* remote object */
	};

	/* extra data associated with local object */
	void			*cookie;
};

        各个成员变量的含义请参考资料Android Binder设计与实现。

        我们进入到flatten_binder函数看看：

[cpp] view plain copy print ?

1. status_t flatten_binder(const sp<ProcessState>& proc,  
2.     const sp<IBinder>& binder, Parcel* out)  
3. {  
4.     flat_binder_object obj;  
5.       
6.     obj.flags = 0x7f | FLAT_BINDER_FLAG_ACCEPTS_FDS;  
7.     if (binder != NULL) {  
8.         IBinder *local = binder->localBinder();  
9.         if (!local) {  
10.             BpBinder *proxy = binder->remoteBinder();  
11.             if (proxy == NULL) {  
12.                 LOGE("null proxy");  
13.             }  
14.             const int32_t handle = proxy ? proxy->handle() : 0;  
15.             obj.type = BINDER_TYPE_HANDLE;  
16.             obj.handle = handle;  
17.             obj.cookie = NULL;  
18.         } else {  
19.             obj.type = BINDER_TYPE_BINDER;  
20.             obj.binder = local->getWeakRefs();  
21.             obj.cookie = local;  
22.         }  
23.     } else {  
24.         obj.type = BINDER_TYPE_BINDER;  
25.         obj.binder = NULL;  
26.         obj.cookie = NULL;  
27.     }  
28.       
29.     return finish_flatten_binder(binder, obj, out);  
30. }  

status_t flatten_binder(const sp<ProcessState>& proc,
    const sp<IBinder>& binder, Parcel* out)
{
    flat_binder_object obj;
    
    obj.flags = 0x7f | FLAT_BINDER_FLAG_ACCEPTS_FDS;
    if (binder != NULL) {
        IBinder *local = binder->localBinder();
        if (!local) {
            BpBinder *proxy = binder->remoteBinder();
            if (proxy == NULL) {
                LOGE("null proxy");
            }
            const int32_t handle = proxy ? proxy->handle() : 0;
            obj.type = BINDER_TYPE_HANDLE;
            obj.handle = handle;
            obj.cookie = NULL;
        } else {
            obj.type = BINDER_TYPE_BINDER;
            obj.binder = local->getWeakRefs();
            obj.cookie = local;
        }
    } else {
        obj.type = BINDER_TYPE_BINDER;
        obj.binder = NULL;
        obj.cookie = NULL;
    }
    
    return finish_flatten_binder(binder, obj, out);
}

        首先是初始化flat_binder_object的flags域：

[cpp] view plain copy print ?

1. obj.flags = 0x7f | FLAT_BINDER_FLAG_ACCEPTS_FDS;  

obj.flags = 0x7f | FLAT_BINDER_FLAG_ACCEPTS_FDS;

        0x7f表示处理本Binder实体请求数据包的线程的最低优先级，FLAT_BINDER_FLAG_ACCEPTS_FDS表示这个Binder实体可以接受文件描述符，Binder实体在收到文件描述符时，就会在本进程中打开这个文件。

       传进来的binder即为MediaPlayerService::instantiate函数中new出来的MediaPlayerService实例，因此，不为空。又由于MediaPlayerService继承自BBinder类，它是一个本地Binder实体，因此binder->localBinder返回一个BBinder指针，而且肯定不为空，于是执行下面语句：

[cpp] view plain copy print ?

1. obj.type = BINDER_TYPE_BINDER;  
2. obj.binder = local->getWeakRefs();  
3. obj.cookie = local;  

obj.type = BINDER_TYPE_BINDER;
obj.binder = local->getWeakRefs();
obj.cookie = local;

        设置了flat_binder_obj的其他成员变量，注意，指向这个Binder实体地址的指针local保存在flat_binder_obj的成员变量cookie中。

        函数调用finish_flatten_binder来将这个flat_binder_obj写入到Parcel中去：

[cpp] view plain copy print ?

1. inline static status_t finish_flatten_binder(  
2.     const sp<IBinder>& binder, const flat_binder_object& flat, Parcel* out)  
3. {  
4.     return out->writeObject(flat, false);  
5. }  

inline static status_t finish_flatten_binder(
    const sp<IBinder>& binder, const flat_binder_object& flat, Parcel* out)
{
    return out->writeObject(flat, false);
}

       Parcel::writeObject的实现如下：

[cpp] view plain copy print ?

1. status_t Parcel::writeObject(const flat_binder_object& val, bool nullMetaData)  
2. {  
3.     const bool enoughData = (mDataPos+sizeof(val)) <= mDataCapacity;  
4.     const bool enoughObjects = mObjectsSize < mObjectsCapacity;  
5.     if (enoughData && enoughObjects) {  
6. restart_write:  
7.         *reinterpret_cast<flat_binder_object*>(mData+mDataPos) = val;  
8.           
9.         // Need to write meta-data?   
10.         if (nullMetaData || val.binder != NULL) {  
11.             mObjects[mObjectsSize] = mDataPos;  
12.             acquire_object(ProcessState::self(), val, this);  
13.             mObjectsSize++;  
14.         }  
15.           
16.         // remember if it's a file descriptor   
17.         if (val.type == BINDER_TYPE_FD) {  
18.             mHasFds = mFdsKnown = true;  
19.         }  
20.   
21.         return finishWrite(sizeof(flat_binder_object));  
22.     }  
23.   
24.     if (!enoughData) {  
25.         const status_t err = growData(sizeof(val));  
26.         if (err != NO_ERROR) return err;  
27.     }  
28.     if (!enoughObjects) {  
29.         size_t newSize = ((mObjectsSize+2)*3)/2;  
30.         size_t* objects = (size_t*)realloc(mObjects, newSize*sizeof(size_t));  
31.         if (objects == NULL) return NO_MEMORY;  
32.         mObjects = objects;  
33.         mObjectsCapacity = newSize;  
34.     }  
35.       
36.     goto restart_write;  
37. }  

status_t Parcel::writeObject(const flat_binder_object& val, bool nullMetaData)
{
    const bool enoughData = (mDataPos+sizeof(val)) <= mDataCapacity;
    const bool enoughObjects = mObjectsSize < mObjectsCapacity;
    if (enoughData && enoughObjects) {
restart_write:
        *reinterpret_cast<flat_binder_object*>(mData+mDataPos) = val;
        
        // Need to write meta-data?
        if (nullMetaData || val.binder != NULL) {
            mObjects[mObjectsSize] = mDataPos;
            acquire_object(ProcessState::self(), val, this);
            mObjectsSize++;
        }
        
        // remember if it's a file descriptor
        if (val.type == BINDER_TYPE_FD) {
            mHasFds = mFdsKnown = true;
        }

        return finishWrite(sizeof(flat_binder_object));
    }

    if (!enoughData) {
        const status_t err = growData(sizeof(val));
        if (err != NO_ERROR) return err;
    }
    if (!enoughObjects) {
        size_t newSize = ((mObjectsSize+2)*3)/2;
        size_t* objects = (size_t*)realloc(mObjects, newSize*sizeof(size_t));
        if (objects == NULL) return NO_MEMORY;
        mObjects = objects;
        mObjectsCapacity = newSize;
    }
    
    goto restart_write;
}

        这里除了把flat_binder_obj写到Parcel里面之内，还要记录这个flat_binder_obj在Parcel里面的偏移位置：

[cpp] view plain copy print ?

1. mObjects[mObjectsSize] = mDataPos;  

mObjects[mObjectsSize] = mDataPos;

       这里因为，如果进程间传输的数据间带有Binder对象的时候，Binder驱动程序需要作进一步的处理，以维护各个Binder实体的一致性，下面我们将会看到Binder驱动程序是怎么处理这些Binder对象的。

       再回到BpServiceManager::addService函数中，调用下面语句：

[cpp] view plain copy print ?

1. status_t err = remote()->transact(ADD_SERVICE_TRANSACTION, data, &reply);  

status_t err = remote()->transact(ADD_SERVICE_TRANSACTION, data, &reply);

       回到浅谈Android系统进程间通信（IPC）机制Binder中的Server和Client获得Service Manager接口之路一文中的类图中去看一下，这里的remote成员函数来自于BpRefBase类，它返回一个BpBinder指针。因此，我们继续进入到BpBinder::transact函数中去看看：

[cpp] view plain copy print ?

1. status_t BpBinder::transact(  
2.     uint32_t code, const Parcel& data, Parcel* reply, uint32_t flags)  
3. {  
4.     // Once a binder has died, it will never come back to life.   
5.     if (mAlive) {  
6.         status_t status = IPCThreadState::self()->transact(  
7.             mHandle, code, data, reply, flags);  
8.         if (status == DEAD_OBJECT) mAlive = 0;  
9.         return status;  
10.     }  
11.   
12.     return DEAD_OBJECT;  
13. }  

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;
}

       这里又调用了IPCThreadState::transact进执行实际的操作。注意，这里的mHandle为0，code为ADD_SERVICE_TRANSACTION。ADD_SERVICE_TRANSACTION是上面以参数形式传进来的，那mHandle为什么是0呢？因为这里表示的是Service Manager远程接口，它的句柄值一定是0，具体请参考浅谈Android系统进程间通信（IPC）机制Binder中的Server和Client获得Service Manager接口之路一文。
       再进入到IPCThreadState::transact函数，看看做了些什么事情：

[cpp] view plain copy print ?

1. status_t IPCThreadState::transact(int32_t handle,  
2.                                   uint32_t code, const Parcel& data,  
3.                                   Parcel* reply, uint32_t flags)  
4. {  
5.     status_t err = data.errorCheck();  
6.   
7.     flags |= TF_ACCEPT_FDS;  
8.   
9.     IF_LOG_TRANSACTIONS() {  
10.         TextOutput::Bundle _b(alog);  
11.         alog << "BC_TRANSACTION thr " << (void*)pthread_self() << " / hand "  
12.             << handle << " / code " << TypeCode(code) << ": "  
13.             << indent << data << dedent << endl;  
14.     }  
15.       
16.     if (err == NO_ERROR) {  
17.         LOG_ONEWAY(">>>> SEND from pid %d uid %d %s", getpid(), getuid(),  
18.             (flags & TF_ONE_WAY) == 0 ? "READ REPLY" : "ONE WAY");  
19.         err = writeTransactionData(BC_TRANSACTION, flags, handle, code, data, NULL);  
20.     }  
21.       
22.     if (err != NO_ERROR) {  
23.         if (reply) reply->setError(err);  
24.         return (mLastError = err);  
25.     }  
26.       
27.     if ((flags & TF_ONE_WAY) == 0) {  
28.         #if 0   
29.         if (code == 4) { // relayout   
30.             LOGI(">>>>>> CALLING transaction 4");  
31.         } else {  
32.             LOGI(">>>>>> CALLING transaction %d", code);  
33.         }  
34.         #endif   
35.         if (reply) {  
36.             err = waitForResponse(reply);  
37.         } else {  
38.             Parcel fakeReply;  
39.             err = waitForResponse(&fakeReply);  
40.         }  
41.         #if 0   
42.         if (code == 4) { // relayout   
43.             LOGI("<<<<<< RETURNING transaction 4");  
44.         } else {  
45.             LOGI("<<<<<< RETURNING transaction %d", code);  
46.         }  
47.         #endif   
48.           
49.         IF_LOG_TRANSACTIONS() {  
50.             TextOutput::Bundle _b(alog);  
51.             alog << "BR_REPLY thr " << (void*)pthread_self() << " / hand "  
52.                 << handle << ": ";  
53.             if (reply) alog << indent << *reply << dedent << endl;  
54.             else alog << "(none requested)" << endl;  
55.         }  
56.     } else {  
57.         err = waitForResponse(NULL, NULL);  
58.     }  
59.       
60.     return err;  
61. }  

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
            LOGI(">>>>>> CALLING transaction 4");
        } else {
            LOGI(">>>>>> CALLING transaction %d", code);
        }
        #endif
        if (reply) {
            err = waitForResponse(reply);
        } else {
            Parcel fakeReply;
            err = waitForResponse(&fakeReply);
        }
        #if 0
        if (code == 4) { // relayout
            LOGI("<<<<<< RETURNING transaction 4");
        } else {
            LOGI("<<<<<< 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;
}

        IPCThreadState::transact函数的参数flags是一个默认值为0的参数，上面没有传相应的实参进来，因此，这里就为0。

        函数首先调用writeTransactionData函数准备好一个struct binder_transaction_data结构体变量，这个是等一下要传输给Binder驱动程序的。struct binder_transaction_data的定义我们在浅谈Service Manager成为Android进程间通信（IPC）机制Binder守护进程之路一文中有详细描述，读者不妨回过去读一下。这里为了方便描述，将struct binder_transaction_data的定义再次列出来：

[cpp] view plain copy print ?

1. struct binder_transaction_data {  
2.     /* The first two are only used for bcTRANSACTION and brTRANSACTION, 
3.      * identifying the target and contents of the transaction. 
4.      */  
5.     union {  
6.         size_t  handle; /* target descriptor of command transaction */  
7.         void    *ptr;   /* target descriptor of return transaction */  
8.     } target;  
9.     void        *cookie;    /* target object cookie */  
10.     unsigned int    code;       /* transaction command */  
11.   
12.     /* General information about the transaction. */  
13.     unsigned int    flags;  
14.     pid_t       sender_pid;  
15.     uid_t       sender_euid;  
16.     size_t      data_size;  /* number of bytes of data */  
17.     size_t      offsets_size;   /* number of bytes of offsets */  
18.   
19.     /* If this transaction is inline, the data immediately 
20.      * follows here; otherwise, it ends with a pointer to 
21.      * the data buffer. 
22.      */  
23.     union {  
24.         struct {  
25.             /* transaction data */  
26.             const void  *buffer;  
27.             /* offsets from buffer to flat_binder_object structs */  
28.             const void  *offsets;  
29.         } ptr;  
30.         uint8_t buf[8];  
31.     } data;  
32. };  

struct binder_transaction_data {
	/* The first two are only used for bcTRANSACTION and brTRANSACTION,
	 * identifying the target and contents of the transaction.
	 */
	union {
		size_t	handle;	/* target descriptor of command transaction */
		void	*ptr;	/* target descriptor of return transaction */
	} target;
	void		*cookie;	/* target object cookie */
	unsigned int	code;		/* transaction command */

	/* General information about the transaction. */
	unsigned int	flags;
	pid_t		sender_pid;
	uid_t		sender_euid;
	size_t		data_size;	/* number of bytes of data */
	size_t		offsets_size;	/* number of bytes of offsets */

	/* If this transaction is inline, the data immediately
	 * follows here; otherwise, it ends with a pointer to
	 * the data buffer.
	 */
	union {
		struct {
			/* transaction data */
			const void	*buffer;
			/* offsets from buffer to flat_binder_object structs */
			const void	*offsets;
		} ptr;
		uint8_t	buf[8];
	} data;
};

         writeTransactionData函数的实现如下：

[cpp] view plain copy print ?

1. status_t IPCThreadState::writeTransactionData(int32_t cmd, uint32_t binderFlags,  
2.     int32_t handle, uint32_t code, const Parcel& data, status_t* statusBuffer)  
3. {  
4.     binder_transaction_data tr;  
5.   
6.     tr.target.handle = handle;  
7.     tr.code = code;  
8.     tr.flags = binderFlags;  
9.       
10.     const status_t err = data.errorCheck();  
11.     if (err == NO_ERROR) {  
12.         tr.data_size = data.ipcDataSize();  
13.         tr.data.ptr.buffer = data.ipcData();  
14.         tr.offsets_size = data.ipcObjectsCount()*sizeof(size_t);  
15.         tr.data.ptr.offsets = data.ipcObjects();  
16.     } else if (statusBuffer) {  
17.         tr.flags |= TF_STATUS_CODE;  
18.         *statusBuffer = err;  
19.         tr.data_size = sizeof(status_t);  
20.         tr.data.ptr.buffer = statusBuffer;  
21.         tr.offsets_size = 0;  
22.         tr.data.ptr.offsets = NULL;  
23.     } else {  
24.         return (mLastError = err);  
25.     }  
26.       
27.     mOut.writeInt32(cmd);  
28.     mOut.write(&tr, sizeof(tr));  
29.       
30.     return NO_ERROR;  
31. }  

status_t IPCThreadState::writeTransactionData(int32_t cmd, uint32_t binderFlags,
    int32_t handle, uint32_t code, const Parcel& data, status_t* statusBuffer)
{
    binder_transaction_data tr;

    tr.target.handle = handle;
    tr.code = code;
    tr.flags = binderFlags;
    
    const status_t err = data.errorCheck();
    if (err == NO_ERROR) {
        tr.data_size = data.ipcDataSize();
        tr.data.ptr.buffer = data.ipcData();
        tr.offsets_size = data.ipcObjectsCount()*sizeof(size_t);
        tr.data.ptr.offsets = data.ipcObjects();
    } else if (statusBuffer) {
        tr.flags |= TF_STATUS_CODE;
        *statusBuffer = err;
        tr.data_size = sizeof(status_t);
        tr.data.ptr.buffer = statusBuffer;
        tr.offsets_size = 0;
        tr.data.ptr.offsets = NULL;
    } else {
        return (mLastError = err);
    }
    
    mOut.writeInt32(cmd);
    mOut.write(&tr, sizeof(tr));
    
    return NO_ERROR;
}

        注意，这里的cmd为BC_TRANSACTION。 这个函数很简单，在这个场景下，就是执行下面语句来初始化本地变量tr：

[cpp] view plain copy print ?

1. tr.data_size = data.ipcDataSize();  
2. tr.data.ptr.buffer = data.ipcData();  
3. tr.offsets_size = data.ipcObjectsCount()*sizeof(size_t);  
4. tr.data.ptr.offsets = data.ipcObjects();  

tr.data_size = data.ipcDataSize();
tr.data.ptr.buffer = data.ipcData();
tr.offsets_size = data.ipcObjectsCount()*sizeof(size_t);
tr.data.ptr.offsets = data.ipcObjects();

       回忆一下上面的内容，写入到tr.data.ptr.buffer的内容相当于下面的内容：

[cpp] view plain copy print ?

1. writeInt32(IPCThreadState::self()->getStrictModePolicy() |  
2.                STRICT_MODE_PENALTY_GATHER);  
3. writeString16("android.os.IServiceManager");  
4. writeString16("media.player");  
5. writeStrongBinder(new MediaPlayerService());  

writeInt32(IPCThreadState::self()->getStrictModePolicy() |
               STRICT_MODE_PENALTY_GATHER);
writeString16("android.os.IServiceManager");
writeString16("media.player");
writeStrongBinder(new MediaPlayerService());

       其中包含了一个Binder实体MediaPlayerService，因此需要设置tr.offsets_size就为1，tr.data.ptr.offsets就指向了这个MediaPlayerService的地址在tr.data.ptr.buffer中的偏移量。最后，将tr的内容保存在IPCThreadState的成员变量mOut中。
       回到IPCThreadState::transact函数中，接下去看，(flags & TF_ONE_WAY) == 0为true，并且reply不为空，所以最终进入到waitForResponse(reply)这条路径来。我们看一下waitForResponse函数的实现：

[cpp] view plain copy print ?

1. status_t IPCThreadState::waitForResponse(Parcel *reply, status_t *acquireResult)  
2. {  
3.     int32_t cmd;  
4.     int32_t err;  
5.   
6.     while (1) {  
7.         if ((err=talkWithDriver()) < NO_ERROR) break;  
8.         err = mIn.errorCheck();  
9.         if (err < NO_ERROR) break;  
10.         if (mIn.dataAvail() == 0) continue;  
11.           
12.         cmd = mIn.readInt32();  
13.           
14.         IF_LOG_COMMANDS() {  
15.             alog << "Processing waitForResponse Command: "  
16.                 << getReturnString(cmd) << endl;  
17.         }  
18.   
19.         switch (cmd) {  
20.         case BR_TRANSACTION_COMPLETE:  
21.             if (!reply && !acquireResult) goto finish;  
22.             break;  
23.           
24.         case BR_DEAD_REPLY:  
25.             err = DEAD_OBJECT;  
26.             goto finish;  
27.   
28.         case BR_FAILED_REPLY:  
29.             err = FAILED_TRANSACTION;  
30.             goto finish;  
31.           
32.         case BR_ACQUIRE_RESULT:  
33.             {  
34.                 LOG_ASSERT(acquireResult != NULL, "Unexpected brACQUIRE_RESULT");  
35.                 const int32_t result = mIn.readInt32();  
36.                 if (!acquireResult) continue;  
37.                 *acquireResult = result ? NO_ERROR : INVALID_OPERATION;  
38.             }  
39.             goto finish;  
40.           
41.         case BR_REPLY:  
42.             {  
43.                 binder_transaction_data tr;  
44.                 err = mIn.read(&tr, sizeof(tr));  
45.                 LOG_ASSERT(err == NO_ERROR, "Not enough command data for brREPLY");  
46.                 if (err != NO_ERROR) goto finish;  
47.   
48.                 if (reply) {  
49.                     if ((tr.flags & TF_STATUS_CODE) == 0) {  
50.                         reply->ipcSetDataReference(  
51.                             reinterpret_cast<const uint8_t*>(tr.data.ptr.buffer),  
52.                             tr.data_size,  
53.                             reinterpret_cast<const size_t*>(tr.data.ptr.offsets),  
54.                             tr.offsets_size/sizeof(size_t),  
55.                             freeBuffer, this);  
56.                     } else {  
57.                         err = *static_cast<const status_t*>(tr.data.ptr.buffer);  
58.                         freeBuffer(NULL,  
59.                             reinterpret_cast<const uint8_t*>(tr.data.ptr.buffer),  
60.                             tr.data_size,  
61.                             reinterpret_cast<const size_t*>(tr.data.ptr.offsets),  
62.                             tr.offsets_size/sizeof(size_t), this);  
63.                     }  
64.                 } else {  
65.                     freeBuffer(NULL,  
66.                         reinterpret_cast<const uint8_t*>(tr.data.ptr.buffer),  
67.                         tr.data_size,  
68.                         reinterpret_cast<const size_t*>(tr.data.ptr.offsets),  
69.                         tr.offsets_size/sizeof(size_t), this);  
70.                     continue;  
71.                 }  
72.             }  
73.             goto finish;  
74.   
75.         default:  
76.             err = executeCommand(cmd);  
77.             if (err != NO_ERROR) goto finish;  
78.             break;  
79.         }  
80.     }  
81.   
82. finish:  
83.     if (err != NO_ERROR) {  
84.         if (acquireResult) *acquireResult = err;  
85.         if (reply) reply->setError(err);  
86.         mLastError = err;  
87.     }  
88.       
89.     return err;  
90. }  

status_t IPCThreadState::waitForResponse(Parcel *reply, status_t *acquireResult)
{
    int32_t cmd;
    int32_t err;

    while (1) {
        if ((err=talkWithDriver()) < NO_ERROR) break;
        err = mIn.errorCheck();
        if (err < NO_ERROR) break;
        if (mIn.dataAvail() == 0) continue;
        
        cmd = mIn.readInt32();
        
        IF_LOG_COMMANDS() {
            alog << "Processing waitForResponse Command: "
                << getReturnString(cmd) << endl;
        }

        switch (cmd) {
        case BR_TRANSACTION_COMPLETE:
            if (!reply && !acquireResult) goto finish;
            break;
        
        case BR_DEAD_REPLY:
            err = DEAD_OBJECT;
            goto finish;

        case BR_FAILED_REPLY:
            err = FAILED_TRANSACTION;
            goto finish;
        
        case BR_ACQUIRE_RESULT:
            {
                LOG_ASSERT(acquireResult != NULL, "Unexpected brACQUIRE_RESULT");
                const int32_t result = mIn.readInt32();
                if (!acquireResult) continue;
                *acquireResult = result ? NO_ERROR : INVALID_OPERATION;
            }
            goto finish;
        
        case BR_REPLY:
            {
                binder_transaction_data tr;
                err = mIn.read(&tr, sizeof(tr));
                LOG_ASSERT(err == NO_ERROR, "Not enough command data for brREPLY");
                if (err != NO_ERROR) goto finish;

                if (reply) {
                    if ((tr.flags & TF_STATUS_CODE) == 0) {
                        reply->ipcSetDataReference(
                            reinterpret_cast<const uint8_t*>(tr.data.ptr.buffer),
                            tr.data_size,
                            reinterpret_cast<const size_t*>(tr.data.ptr.offsets),
                            tr.offsets_size/sizeof(size_t),
                            freeBuffer, this);
                    } else {
                        err = *static_cast<const status_t*>(tr.data.ptr.buffer);
                        freeBuffer(NULL,
                            reinterpret_cast<const uint8_t*>(tr.data.ptr.buffer),
                            tr.data_size,
                            reinterpret_cast<const size_t*>(tr.data.ptr.offsets),
                            tr.offsets_size/sizeof(size_t), this);
                    }
                } else {
                    freeBuffer(NULL,
                        reinterpret_cast<const uint8_t*>(tr.data.ptr.buffer),
                        tr.data_size,
                        reinterpret_cast<const size_t*>(tr.data.ptr.offsets),
                        tr.offsets_size/sizeof(size_t), this);
                    continue;
                }
            }
            goto finish;

        default:
            err = executeCommand(cmd);
            if (err != NO_ERROR) goto finish;
            break;
        }
    }

finish:
    if (err != NO_ERROR) {
        if (acquireResult) *acquireResult = err;
        if (reply) reply->setError(err);
        mLastError = err;
    }
    
    return err;
}

        这个函数虽然很长，但是主要调用了talkWithDriver函数来与Binder驱动程序进行交互：

[cpp] view plain copy print ?

1. status_t IPCThreadState::talkWithDriver(bool doReceive)  
2. {  
3.     LOG_ASSERT(mProcess->mDriverFD >= 0, "Binder driver is not opened");  
4.       
5.     binder_write_read bwr;  
6.       
7.     // Is the read buffer empty?   
8.     const bool needRead = mIn.dataPosition() >= mIn.dataSize();  
9.       
10.     // We don't want to write anything if we are still reading   
11.     // from data left in the input buffer and the caller   
12.     // has requested to read the next data.   
13.     const size_t outAvail = (!doReceive || needRead) ? mOut.dataSize() : 0;  
14.       
15.     bwr.write_size = outAvail;  
16.     bwr.write_buffer = (long unsigned int)mOut.data();  
17.   
18.     // This is what we'll read.   
19.     if (doReceive && needRead) {  
20.         bwr.read_size = mIn.dataCapacity();  
21.         bwr.read_buffer = (long unsigned int)mIn.data();  
22.     } else {  
23.         bwr.read_size = 0;  
24.     }  
25.       
26.     IF_LOG_COMMANDS() {  
27.         TextOutput::Bundle _b(alog);  
28.         if (outAvail != 0) {  
29.             alog << "Sending commands to driver: " << indent;  
30.             const void* cmds = (const void*)bwr.write_buffer;  
31.             const void* end = ((const uint8_t*)cmds)+bwr.write_size;  
32.             alog << HexDump(cmds, bwr.write_size) << endl;  
33.             while (cmds < end) cmds = printCommand(alog, cmds);  
34.             alog << dedent;  
35.         }  
36.         alog << "Size of receive buffer: " << bwr.read_size  
37.             << ", needRead: " << needRead << ", doReceive: " << doReceive << endl;  
38.     }  
39.       
40.     // Return immediately if there is nothing to do.   
41.     if ((bwr.write_size == 0) && (bwr.read_size == 0)) return NO_ERROR;  
42.       
43.     bwr.write_consumed = 0;  
44.     bwr.read_consumed = 0;  
45.     status_t err;  
46.     do {  
47.         IF_LOG_COMMANDS() {  
48.             alog << "About to read/write, write size = " << mOut.dataSize() << endl;  
49.         }  
50. #if defined(HAVE_ANDROID_OS)   
51.         if (ioctl(mProcess->mDriverFD, BINDER_WRITE_READ, &bwr) >= 0)  
52.             err = NO_ERROR;  
53.         else  
54.             err = -errno;  
55. #else   
56.         err = INVALID_OPERATION;  
57. #endif   
58.         IF_LOG_COMMANDS() {  
59.             alog << "Finished read/write, write size = " << mOut.dataSize() << endl;  
60.         }  
61.     } while (err == -EINTR);  
62.       
63.     IF_LOG_COMMANDS() {  
64.         alog << "Our err: " << (void*)err << ", write consumed: "  
65.             << bwr.write_consumed << " (of " << mOut.dataSize()  
66.             << "), read consumed: " << bwr.read_consumed << endl;  
67.     }  
68.   
69.     if (err >= NO_ERROR) {  
70.         if (bwr.write_consumed > 0) {  
71.             if (bwr.write_consumed < (ssize_t)mOut.dataSize())  
72.                 mOut.remove(0, bwr.write_consumed);  
73.             else  
74.                 mOut.setDataSize(0);  
75.         }  
76.         if (bwr.read_consumed > 0) {  
77.             mIn.setDataSize(bwr.read_consumed);  
78.             mIn.setDataPosition(0);  
79.         }  
80.         IF_LOG_COMMANDS() {  
81.             TextOutput::Bundle _b(alog);  
82.             alog << "Remaining data size: " << mOut.dataSize() << endl;  
83.             alog << "Received commands from driver: " << indent;  
84.             const void* cmds = mIn.data();  
85.             const void* end = mIn.data() + mIn.dataSize();  
86.             alog << HexDump(cmds, mIn.dataSize()) << endl;  
87.             while (cmds < end) cmds = printReturnCommand(alog, cmds);  
88.             alog << dedent;  
89.         }  
90.         return NO_ERROR;  
91.     }  
92.       
93.     return err;  
94. }  

status_t IPCThreadState::talkWithDriver(bool doReceive)
{
    LOG_ASSERT(mProcess->mDriverFD >= 0, "Binder driver is not opened");
    
    binder_write_read bwr;
    
    // Is the read buffer empty?
    const bool needRead = mIn.dataPosition() >= mIn.dataSize();
    
    // We don't want to write anything if we are still reading
    // from data left in the input buffer and the caller
    // has requested to read the next data.
    const size_t outAvail = (!doReceive || needRead) ? mOut.dataSize() : 0;
    
    bwr.write_size = outAvail;
    bwr.write_buffer = (long unsigned int)mOut.data();

    // This is what we'll read.
    if (doReceive && needRead) {
        bwr.read_size = mIn.dataCapacity();
        bwr.read_buffer = (long unsigned int)mIn.data();
    } else {
        bwr.read_size = 0;
    }
    
    IF_LOG_COMMANDS() {
        TextOutput::Bundle _b(alog);
        if (outAvail != 0) {
            alog << "Sending commands to driver: " << indent;
            const void* cmds = (const void*)bwr.write_buffer;
            const void* end = ((const uint8_t*)cmds)+bwr.write_size;
            alog << HexDump(cmds, bwr.write_size) << endl;
            while (cmds < end) cmds = printCommand(alog, cmds);
            alog << dedent;
        }
        alog << "Size of receive buffer: " << bwr.read_size
            << ", needRead: " << needRead << ", doReceive: " << doReceive << endl;
    }
    
    // Return immediately if there is nothing to do.
    if ((bwr.write_size == 0) && (bwr.read_size == 0)) return NO_ERROR;
    
    bwr.write_consumed = 0;
    bwr.read_consumed = 0;
    status_t err;
    do {
        IF_LOG_COMMANDS() {
            alog << "About to read/write, write size = " << mOut.dataSize() << endl;
        }
#if defined(HAVE_ANDROID_OS)
        if (ioctl(mProcess->mDriverFD, BINDER_WRITE_READ, &bwr) >= 0)
            err = NO_ERROR;
        else
            err = -errno;
#else
        err = INVALID_OPERATION;
#endif
        IF_LOG_COMMANDS() {
            alog << "Finished read/write, write size = " << mOut.dataSize() << endl;
        }
    } while (err == -EINTR);
    
    IF_LOG_COMMANDS() {
        alog << "Our err: " << (void*)err << ", write consumed: "
            << bwr.write_consumed << " (of " << mOut.dataSize()
			<< "), read consumed: " << bwr.read_consumed << endl;
    }

    if (err >= NO_ERROR) {
        if (bwr.write_consumed > 0) {
            if (bwr.write_consumed < (ssize_t)mOut.dataSize())
                mOut.remove(0, bwr.write_consumed);
            else
                mOut.setDataSize(0);
        }
        if (bwr.read_consumed > 0) {
            mIn.setDataSize(bwr.read_consumed);
            mIn.setDataPosition(0);
        }
        IF_LOG_COMMANDS() {
            TextOutput::Bundle _b(alog);
            alog << "Remaining data size: " << mOut.dataSize() << endl;
            alog << "Received commands from driver: " << indent;
            const void* cmds = mIn.data();
            const void* end = mIn.data() + mIn.dataSize();
            alog << HexDump(cmds, mIn.dataSize()) << endl;
            while (cmds < end) cmds = printReturnCommand(alog, cmds);
            alog << dedent;
        }
        return NO_ERROR;
    }
    
    return err;
}

        这里doReceive和needRead均为1，有兴趣的读者可以自已分析一下。因此，这里告诉Binder驱动程序，先执行write操作，再执行read操作，下面我们将会看到。

        最后，通过ioctl(mProcess->mDriverFD, BINDER_WRITE_READ, &bwr)进行到Binder驱动程序的binder_ioctl函数，我们只关注cmd为BINDER_WRITE_READ的逻辑：

[cpp] view plain copy print ?

1. static long binder_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)  
2. {  
3.     int ret;  
4.     struct binder_proc *proc = filp->private_data;  
5.     struct binder_thread *thread;  
6.     unsigned int size = _IOC_SIZE(cmd);  
7.     void __user *ubuf = (void __user *)arg;  
8.   
9.     /*printk(KERN_INFO "binder_ioctl: %d:%d %x %lx\n", proc->pid, current->pid, cmd, arg);*/  
10.   
11.     ret = wait_event_interruptible(binder_user_error_wait, binder_stop_on_user_error < 2);  
12.     if (ret)  
13.         return ret;  
14.   
15.     mutex_lock(&binder_lock);  
16.     thread = binder_get_thread(proc);  
17.     if (thread == NULL) {  
18.         ret = -ENOMEM;  
19.         goto err;  
20.     }  
21.   
22.     switch (cmd) {  
23.     case BINDER_WRITE_READ: {  
24.         struct binder_write_read bwr;  
25.         if (size != sizeof(struct binder_write_read)) {  
26.             ret = -EINVAL;  
27.             goto err;  
28.         }  
29.         if (copy_from_user(&bwr, ubuf, sizeof(bwr))) {  
30.             ret = -EFAULT;  
31.             goto err;  
32.         }  
33.         if (binder_debug_mask & BINDER_DEBUG_READ_WRITE)  
34.             printk(KERN_INFO "binder: %d:%d write %ld at %08lx, read %ld at %08lx\n",  
35.             proc->pid, thread->pid, bwr.write_size, bwr.write_buffer, bwr.read_size, bwr.read_buffer);  
36.         if (bwr.write_size > 0) {  
37.             ret = binder_thread_write(proc, thread, (void __user *)bwr.write_buffer, bwr.write_size, &bwr.write_consumed);  
38.             if (ret < 0) {  
39.                 bwr.read_consumed = 0;  
40.                 if (copy_to_user(ubuf, &bwr, sizeof(bwr)))  
41.                     ret = -EFAULT;  
42.                 goto err;  
43.             }  
44.         }  
45.         if (bwr.read_size > 0) {  
46.             ret = binder_thread_read(proc, thread, (void __user *)bwr.read_buffer, bwr.read_size, &bwr.read_consumed, filp->f_flags & O_NONBLOCK);  
47.             if (!list_empty(&proc->todo))  
48.                 wake_up_interruptible(&proc->wait);  
49.             if (ret < 0) {  
50.                 if (copy_to_user(ubuf, &bwr, sizeof(bwr)))  
51.                     ret = -EFAULT;  
52.                 goto err;  
53.             }  
54.         }  
55.         if (binder_debug_mask & BINDER_DEBUG_READ_WRITE)  
56.             printk(KERN_INFO "binder: %d:%d wrote %ld of %ld, read return %ld of %ld\n",  
57.             proc->pid, thread->pid, bwr.write_consumed, bwr.write_size, bwr.read_consumed, bwr.read_size);  
58.         if (copy_to_user(ubuf, &bwr, sizeof(bwr))) {  
59.             ret = -EFAULT;  
60.             goto err;  
61.         }  
62.         break;  
63.     }  
64.     ......  
65.     }  
66.     ret = 0;  
67. err:  
68.     ......  
69.     return ret;  
70. }  

static long binder_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
{
	int ret;
	struct binder_proc *proc = filp->private_data;
	struct binder_thread *thread;
	unsigned int size = _IOC_SIZE(cmd);
	void __user *ubuf = (void __user *)arg;

	/*printk(KERN_INFO "binder_ioctl: %d:%d %x %lx\n", proc->pid, current->pid, cmd, arg);*/

	ret = wait_event_interruptible(binder_user_error_wait, binder_stop_on_user_error < 2);
	if (ret)
		return ret;

	mutex_lock(&binder_lock);
	thread = binder_get_thread(proc);
	if (thread == NULL) {
		ret = -ENOMEM;
		goto err;
	}

	switch (cmd) {
	case BINDER_WRITE_READ: {
		struct binder_write_read bwr;
		if (size != sizeof(struct binder_write_read)) {
			ret = -EINVAL;
			goto err;
		}
		if (copy_from_user(&bwr, ubuf, sizeof(bwr))) {
			ret = -EFAULT;
			goto err;
		}
		if (binder_debug_mask & BINDER_DEBUG_READ_WRITE)
			printk(KERN_INFO "binder: %d:%d write %ld at %08lx, read %ld at %08lx\n",
			proc->pid, thread->pid, bwr.write_size, bwr.write_buffer, bwr.read_size, bwr.read_buffer);
		if (bwr.write_size > 0) {
			ret = binder_thread_write(proc, thread, (void __user *)bwr.write_buffer, bwr.write_size, &bwr.write_consumed);
			if (ret < 0) {
				bwr.read_consumed = 0;
				if (copy_to_user(ubuf, &bwr, sizeof(bwr)))
					ret = -EFAULT;
				goto err;
			}
		}
		if (bwr.read_size > 0) {
			ret = binder_thread_read(proc, thread, (void __user *)bwr.read_buffer, bwr.read_size, &bwr.read_consumed, filp->f_flags & O_NONBLOCK);
			if (!list_empty(&proc->todo))
				wake_up_interruptible(&proc->wait);
			if (ret < 0) {
				if (copy_to_user(ubuf, &bwr, sizeof(bwr)))
					ret = -EFAULT;
				goto err;
			}
		}
		if (binder_debug_mask & BINDER_DEBUG_READ_WRITE)
			printk(KERN_INFO "binder: %d:%d wrote %ld of %ld, read return %ld of %ld\n",
			proc->pid, thread->pid, bwr.write_consumed, bwr.write_size, bwr.read_consumed, bwr.read_size);
		if (copy_to_user(ubuf, &bwr, sizeof(bwr))) {
			ret = -EFAULT;
			goto err;
		}
		break;
	}
	......
	}
	ret = 0;
err:
	......
	return ret;
}

         函数首先是将用户传进来的参数拷贝到本地变量struct binder_write_read bwr中去。这里bwr.write_size > 0为true，因此，进入到binder_thread_write函数中，我们只关注BC_TRANSACTION部分的逻辑：

[cpp] view plain copy print ?

1. binder_thread_write(struct binder_proc *proc, struct binder_thread *thread,  
2.                     void __user *buffer, int size, signed long *consumed)  
3. {  
4.     uint32_t cmd;  
5.     void __user *ptr = buffer + *consumed;  
6.     void __user *end = buffer + size;  
7.   
8.     while (ptr < end && thread->return_error == BR_OK) {  
9.         if (get_user(cmd, (uint32_t __user *)ptr))  
10.             return -EFAULT;  
11.         ptr += sizeof(uint32_t);  
12.         if (_IOC_NR(cmd) < ARRAY_SIZE(binder_stats.bc)) {  
13.             binder_stats.bc[_IOC_NR(cmd)]++;  
14.             proc->stats.bc[_IOC_NR(cmd)]++;  
15.             thread->stats.bc[_IOC_NR(cmd)]++;  
16.         }  
17.         switch (cmd) {  
18.             .....  
19.         case BC_TRANSACTION:  
20.         case BC_REPLY: {  
21.             struct binder_transaction_data tr;  
22.   
23.             if (copy_from_user(&tr, ptr, sizeof(tr)))  
24.                 return -EFAULT;  
25.             ptr += sizeof(tr);  
26.             binder_transaction(proc, thread, &tr, cmd == BC_REPLY);  
27.             break;  
28.         }  
29.         ......  
30.         }  
31.         *consumed = ptr - buffer;  
32.     }  
33.     return 0;  
34. }  

binder_thread_write(struct binder_proc *proc, struct binder_thread *thread,
					void __user *buffer, int size, signed long *consumed)
{
	uint32_t cmd;
	void __user *ptr = buffer + *consumed;
	void __user *end = buffer + size;

	while (ptr < end && thread->return_error == BR_OK) {
		if (get_user(cmd, (uint32_t __user *)ptr))
			return -EFAULT;
		ptr += sizeof(uint32_t);
		if (_IOC_NR(cmd) < ARRAY_SIZE(binder_stats.bc)) {
			binder_stats.bc[_IOC_NR(cmd)]++;
			proc->stats.bc[_IOC_NR(cmd)]++;
			thread->stats.bc[_IOC_NR(cmd)]++;
		}
		switch (cmd) {
	        .....
		case BC_TRANSACTION:
		case BC_REPLY: {
			struct binder_transaction_data tr;

			if (copy_from_user(&tr, ptr, sizeof(tr)))
				return -EFAULT;
			ptr += sizeof(tr);
			binder_transaction(proc, thread, &tr, cmd == BC_REPLY);
			break;
		}
		......
		}
		*consumed = ptr - buffer;
	}
	return 0;
}

         首先将用户传进来的transact参数拷贝在本地变量struct binder_transaction_data tr中去，接着调用binder_transaction函数进一步处理，这里我们忽略掉无关代码：

[cpp] view plain copy print ?

1. static void  
2. binder_transaction(struct binder_proc *proc, struct binder_thread *thread,  
3. struct binder_transaction_data *tr, int reply)  
4. {  
5.     struct binder_transaction *t;  
6.     struct binder_work *tcomplete;  
7.     size_t *offp, *off_end;  
8.     struct binder_proc *target_proc;  
9.     struct binder_thread *target_thread = NULL;  
10.     struct binder_node *target_node = NULL;  
11.     struct list_head *target_list;  
12.     wait_queue_head_t *target_wait;  
13.     struct binder_transaction *in_reply_to = NULL;  
14.     struct binder_transaction_log_entry *e;  
15.     uint32_t return_error;  
16.   
17.         ......  
18.   
19.     if (reply) {  
20.          ......  
21.     } else {  
22.         if (tr->target.handle) {  
23.             ......  
24.         } else {  
25.             target_node = binder_context_mgr_node;  
26.             if (target_node == NULL) {  
27.                 return_error = BR_DEAD_REPLY;  
28.                 goto err_no_context_mgr_node;  
29.             }  
30.         }  
31.         ......  
32.         target_proc = target_node->proc;  
33.         if (target_proc == NULL) {  
34.             return_error = BR_DEAD_REPLY;  
35.             goto err_dead_binder;  
36.         }  
37.         ......  
38.     }  
39.     if (target_thread) {  
40.         ......  
41.     } else {  
42.         target_list = &target_proc->todo;  
43.         target_wait = &target_proc->wait;  
44.     }  
45.       
46.     ......  
47.   
48.     /* TODO: reuse incoming transaction for reply */  
49.     t = kzalloc(sizeof(*t), GFP_KERNEL);  
50.     if (t == NULL) {  
51.         return_error = BR_FAILED_REPLY;  
52.         goto err_alloc_t_failed;  
53.     }  
54.     ......  
55.   
56.     tcomplete = kzalloc(sizeof(*tcomplete), GFP_KERNEL);  
57.     if (tcomplete == NULL) {  
58.         return_error = BR_FAILED_REPLY;  
59.         goto err_alloc_tcomplete_failed;  
60.     }  
61.       
62.     ......  
63.   
64.     if (!reply && !(tr->flags & TF_ONE_WAY))  
65.         t->from = thread;  
66.     else  
67.         t->from = NULL;  
68.     t->sender_euid = proc->tsk->cred->euid;  
69.     t->to_proc = target_proc;  
70.     t->to_thread = target_thread;  
71.     t->code = tr->code;  
72.     t->flags = tr->flags;  
73.     t->priority = task_nice(current);  
74.     t->buffer = binder_alloc_buf(target_proc, tr->data_size,  
75.         tr->offsets_size, !reply && (t->flags & TF_ONE_WAY));  
76.     if (t->buffer == NULL) {  
77.         return_error = BR_FAILED_REPLY;  
78.         goto err_binder_alloc_buf_failed;  
79.     }  
80.     t->buffer->allow_user_free = 0;  
81.     t->buffer->debug_id = t->debug_id;  
82.     t->buffer->transaction = t;  
83.     t->buffer->target_node = target_node;  
84.     if (target_node)  
85.         binder_inc_node(target_node, 1, 0, NULL);  
86.   
87.     offp = (size_t *)(t->buffer->data + ALIGN(tr->data_size, sizeof(void *)));  
88.   
89.     if (copy_from_user(t->buffer->data, tr->data.ptr.buffer, tr->data_size)) {  
90.         ......  
91.         return_error = BR_FAILED_REPLY;  
92.         goto err_copy_data_failed;  
93.     }  
94.     if (copy_from_user(offp, tr->data.ptr.offsets, tr->offsets_size)) {  
95.         ......  
96.         return_error = BR_FAILED_REPLY;  
97.         goto err_copy_data_failed;  
98.     }  
99.     ......  
100.   
101.     off_end = (void *)offp + tr->offsets_size;  
102.     for (; offp < off_end; offp++) {  
103.         struct flat_binder_object *fp;  
104.         ......  
105.         fp = (struct flat_binder_object *)(t->buffer->data + *offp);  
106.         switch (fp->type) {  
107.         case BINDER_TYPE_BINDER:  
108.         case BINDER_TYPE_WEAK_BINDER: {  
109.             struct binder_ref *ref;  
110.             struct binder_node *node = binder_get_node(proc, fp->binder);  
111.             if (node == NULL) {  
112.                 node = binder_new_node(proc, fp->binder, fp->cookie);  
113.                 if (node == NULL) {  
114.                     return_error = BR_FAILED_REPLY;  
115.                     goto err_binder_new_node_failed;  
116.                 }  
117.                 node->min_priority = fp->flags & FLAT_BINDER_FLAG_PRIORITY_MASK;  
118.                 node->accept_fds = !!(fp->flags & FLAT_BINDER_FLAG_ACCEPTS_FDS);  
119.             }  
120.             if (fp->cookie != node->cookie) {  
121.                 ......  
122.                 goto err_binder_get_ref_for_node_failed;  
123.             }  
124.             ref = binder_get_ref_for_node(target_proc, node);  
125.             if (ref == NULL) {  
126.                 return_error = BR_FAILED_REPLY;  
127.                 goto err_binder_get_ref_for_node_failed;  
128.             }  
129.             if (fp->type == BINDER_TYPE_BINDER)  
130.                 fp->type = BINDER_TYPE_HANDLE;  
131.             else  
132.                 fp->type = BINDER_TYPE_WEAK_HANDLE;  
133.             fp->handle = ref->desc;  
134.             binder_inc_ref(ref, fp->type == BINDER_TYPE_HANDLE, &thread->todo);  
135.             ......  
136.                                 
137.         } break;  
138.         ......  
139.         }  
140.     }  
141.   
142.     if (reply) {  
143.         ......  
144.     } else if (!(t->flags & TF_ONE_WAY)) {  
145.         BUG_ON(t->buffer->async_transaction != 0);  
146.         t->need_reply = 1;  
147.         t->from_parent = thread->transaction_stack;  
148.         thread->transaction_stack = t;  
149.     } else {  
150.         ......  
151.     }  
152.     t->work.type = BINDER_WORK_TRANSACTION;  
153.     list_add_tail(&t->work.entry, target_list);  
154.     tcomplete->type = BINDER_WORK_TRANSACTION_COMPLETE;  
155.     list_add_tail(&tcomplete->entry, &thread->todo);  
156.     if (target_wait)  
157.         wake_up_interruptible(target_wait);  
158.     return;  
159.     ......  
160. }  

static void
binder_transaction(struct binder_proc *proc, struct binder_thread *thread,
struct binder_transaction_data *tr, int reply)
{
	struct binder_transaction *t;
	struct binder_work *tcomplete;
	size_t *offp, *off_end;
	struct binder_proc *target_proc;
	struct binder_thread *target_thread = NULL;
	struct binder_node *target_node = NULL;
	struct list_head *target_list;
	wait_queue_head_t *target_wait;
	struct binder_transaction *in_reply_to = NULL;
	struct binder_transaction_log_entry *e;
	uint32_t return_error;

        ......

	if (reply) {
         ......
	} else {
		if (tr->target.handle) {
            ......
		} else {
			target_node = binder_context_mgr_node;
			if (target_node == NULL) {
				return_error = BR_DEAD_REPLY;
				goto err_no_context_mgr_node;
			}
		}
		......
		target_proc = target_node->proc;
		if (target_proc == NULL) {
			return_error = BR_DEAD_REPLY;
			goto err_dead_binder;
		}
		......
	}
	if (target_thread) {
		......
	} else {
		target_list = &target_proc->todo;
		target_wait = &target_proc->wait;
	}
	
	......

	/* TODO: reuse incoming transaction for reply */
	t = kzalloc(sizeof(*t), GFP_KERNEL);
	if (t == NULL) {
		return_error = BR_FAILED_REPLY;
		goto err_alloc_t_failed;
	}
	......

	tcomplete = kzalloc(sizeof(*tcomplete), GFP_KERNEL);
	if (tcomplete == NULL) {
		return_error = BR_FAILED_REPLY;
		goto err_alloc_tcomplete_failed;
	}
	
	......

	if (!reply && !(tr->flags & TF_ONE_WAY))
		t->from = thread;
	else
		t->from = NULL;
	t->sender_euid = proc->tsk->cred->euid;
	t->to_proc = target_proc;
	t->to_thread = target_thread;
	t->code = tr->code;
	t->flags = tr->flags;
	t->priority = task_nice(current);
	t->buffer = binder_alloc_buf(target_proc, tr->data_size,
		tr->offsets_size, !reply && (t->flags & TF_ONE_WAY));
	if (t->buffer == NULL) {
		return_error = BR_FAILED_REPLY;
		goto err_binder_alloc_buf_failed;
	}
	t->buffer->allow_user_free = 0;
	t->buffer->debug_id = t->debug_id;
	t->buffer->transaction = t;
	t->buffer->target_node = target_node;
	if (target_node)
		binder_inc_node(target_node, 1, 0, NULL);

	offp = (size_t *)(t->buffer->data + ALIGN(tr->data_size, sizeof(void *)));

	if (copy_from_user(t->buffer->data, tr->data.ptr.buffer, tr->data_size)) {
		......
		return_error = BR_FAILED_REPLY;
		goto err_copy_data_failed;
	}
	if (copy_from_user(offp, tr->data.ptr.offsets, tr->offsets_size)) {
		......
		return_error = BR_FAILED_REPLY;
		goto err_copy_data_failed;
	}
	......

	off_end = (void *)offp + tr->offsets_size;
	for (; offp < off_end; offp++) {
		struct flat_binder_object *fp;
		......
		fp = (struct flat_binder_object *)(t->buffer->data + *offp);
		switch (fp->type) {
		case BINDER_TYPE_BINDER:
		case BINDER_TYPE_WEAK_BINDER: {
			struct binder_ref *ref;
			struct binder_node *node = binder_get_node(proc, fp->binder);
			if (node == NULL) {
				node = binder_new_node(proc, fp->binder, fp->cookie);
				if (node == NULL) {
					return_error = BR_FAILED_REPLY;
					goto err_binder_new_node_failed;
				}
				node->min_priority = fp->flags & FLAT_BINDER_FLAG_PRIORITY_MASK;
				node->accept_fds = !!(fp->flags & FLAT_BINDER_FLAG_ACCEPTS_FDS);
			}
			if (fp->cookie != node->cookie) {
				......
				goto err_binder_get_ref_for_node_failed;
			}
			ref = binder_get_ref_for_node(target_proc, node);
			if (ref == NULL) {
				return_error = BR_FAILED_REPLY;
				goto err_binder_get_ref_for_node_failed;
			}
			if (fp->type == BINDER_TYPE_BINDER)
				fp->type = BINDER_TYPE_HANDLE;
			else
				fp->type = BINDER_TYPE_WEAK_HANDLE;
			fp->handle = ref->desc;
			binder_inc_ref(ref, fp->type == BINDER_TYPE_HANDLE, &thread->todo);
			......
							  
		} break;
		......
		}
	}

	if (reply) {
		......
	} else if (!(t->flags & TF_ONE_WAY)) {
		BUG_ON(t->buffer->async_transaction != 0);
		t->need_reply = 1;
		t->from_parent = thread->transaction_stack;
		thread->transaction_stack = t;
	} else {
		......
	}
	t->work.type = BINDER_WORK_TRANSACTION;
	list_add_tail(&t->work.entry, target_list);
	tcomplete->type = BINDER_WORK_TRANSACTION_COMPLETE;
	list_add_tail(&tcomplete->entry, &thread->todo);
	if (target_wait)
		wake_up_interruptible(target_wait);
	return;
    ......
}

       注意，这里传进来的参数reply为0，tr->target.handle也为0。因此，target_proc、target_thread、target_node、target_list和target_wait的值分别为：

[cpp] view plain copy print ?

1. target_node = binder_context_mgr_node;  
2. target_proc = target_node->proc;  
3. target_list = &target_proc->todo;  
4. target_wait = &target_proc->wait;   

target_node = binder_context_mgr_node;
target_proc = target_node->proc;
target_list = &target_proc->todo;
target_wait = &target_proc->wait; 

       接着，分配了一个待处理事务t和一个待完成工作项tcomplete，并执行初始化工作：

[cpp] view plain copy print ?

1. /* TODO: reuse incoming transaction for reply */  
2. t = kzalloc(sizeof(*t), GFP_KERNEL);  
3. if (t == NULL) {  
4.     return_error = BR_FAILED_REPLY;  
5.     goto err_alloc_t_failed;  
6. }  
7. ......  
8.   
9. tcomplete = kzalloc(sizeof(*tcomplete), GFP_KERNEL);  
10. if (tcomplete == NULL) {  
11.     return_error = BR_FAILED_REPLY;  
12.     goto err_alloc_tcomplete_failed;  
13. }  
14.   
15. ......  
16.   
17. if (!reply && !(tr->flags & TF_ONE_WAY))  
18.     t->from = thread;  
19. else  
20.     t->from = NULL;  
21. t->sender_euid = proc->tsk->cred->euid;  
22. t->to_proc = target_proc;  
23. t->to_thread = target_thread;  
24. t->code = tr->code;  
25. t->flags = tr->flags;  
26. t->priority = task_nice(current);  
27. t->buffer = binder_alloc_buf(target_proc, tr->data_size,  
28.     tr->offsets_size, !reply && (t->flags & TF_ONE_WAY));  
29. if (t->buffer == NULL) {  
30.     return_error = BR_FAILED_REPLY;  
31.     goto err_binder_alloc_buf_failed;  
32. }  
33. t->buffer->allow_user_free = 0;  
34. t->buffer->debug_id = t->debug_id;  
35. t->buffer->transaction = t;  
36. t->buffer->target_node = target_node;  
37. if (target_node)  
38.     binder_inc_node(target_node, 1, 0, NULL);  
39.   
40. offp = (size_t *)(t->buffer->data + ALIGN(tr->data_size, sizeof(void *)));  
41.   
42. if (copy_from_user(t->buffer->data, tr->data.ptr.buffer, tr->data_size)) {  
43.     ......  
44.     return_error = BR_FAILED_REPLY;  
45.     goto err_copy_data_failed;  
46. }  
47. if (copy_from_user(offp, tr->data.ptr.offsets, tr->offsets_size)) {  
48.     ......  
49.     return_error = BR_FAILED_REPLY;  
50.     goto err_copy_data_failed;  
51. }  

	/* TODO: reuse incoming transaction for reply */
	t = kzalloc(sizeof(*t), GFP_KERNEL);
	if (t == NULL) {
		return_error = BR_FAILED_REPLY;
		goto err_alloc_t_failed;
	}
	......

	tcomplete = kzalloc(sizeof(*tcomplete), GFP_KERNEL);
	if (tcomplete == NULL) {
		return_error = BR_FAILED_REPLY;
		goto err_alloc_tcomplete_failed;
	}
	
	......

	if (!reply && !(tr->flags & TF_ONE_WAY))
		t->from = thread;
	else
		t->from = NULL;
	t->sender_euid = proc->tsk->cred->euid;
	t->to_proc = target_proc;
	t->to_thread = target_thread;
	t->code = tr->code;
	t->flags = tr->flags;
	t->priority = task_nice(current);
	t->buffer = binder_alloc_buf(target_proc, tr->data_size,
		tr->offsets_size, !reply && (t->flags & TF_ONE_WAY));
	if (t->buffer == NULL) {
		return_error = BR_FAILED_REPLY;
		goto err_binder_alloc_buf_failed;
	}
	t->buffer->allow_user_free = 0;
	t->buffer->debug_id = t->debug_id;
	t->buffer->transaction = t;
	t->buffer->target_node = target_node;
	if (target_node)
		binder_inc_node(target_node, 1, 0, NULL);

	offp = (size_t *)(t->buffer->data + ALIGN(tr->data_size, sizeof(void *)));

	if (copy_from_user(t->buffer->data, tr->data.ptr.buffer, tr->data_size)) {
		......
		return_error = BR_FAILED_REPLY;
		goto err_copy_data_failed;
	}
	if (copy_from_user(offp, tr->data.ptr.offsets, tr->offsets_size)) {
		......
		return_error = BR_FAILED_REPLY;
		goto err_copy_data_failed;
	}

         注意，这里的事务t是要交给target_proc处理的，在这个场景之下，就是Service Manager了。因此，下面的语句：

[cpp] view plain copy print ?

1. t->buffer = binder_alloc_buf(target_proc, tr->data_size,  
2.         tr->offsets_size, !reply && (t->flags & TF_ONE_WAY));  

t->buffer = binder_alloc_buf(target_proc, tr->data_size,
		tr->offsets_size, !reply && (t->flags & TF_ONE_WAY));

         就是在Service Manager的进程空间中分配一块内存来保存用户传进入的参数了：

[cpp] view plain copy print ?

1. if (copy_from_user(t->buffer->data, tr->data.ptr.buffer, tr->data_size)) {  
2.     ......  
3.     return_error = BR_FAILED_REPLY;  
4.     goto err_copy_data_failed;  
5. }  
6. if (copy_from_user(offp, tr->data.ptr.offsets, tr->offsets_size)) {  
7.     ......  
8.     return_error = BR_FAILED_REPLY;  
9.     goto err_copy_data_failed;  
10. }  

	if (copy_from_user(t->buffer->data, tr->data.ptr.buffer, tr->data_size)) {
		......
		return_error = BR_FAILED_REPLY;
		goto err_copy_data_failed;
	}
	if (copy_from_user(offp, tr->data.ptr.offsets, tr->offsets_size)) {
		......
		return_error = BR_FAILED_REPLY;
		goto err_copy_data_failed;
	}

         由于现在target_node要被使用了，增加它的引用计数：

[cpp] view plain copy print ?

1. if (target_node)  
2.         binder_inc_node(target_node, 1, 0, NULL);  

if (target_node)
		binder_inc_node(target_node, 1, 0, NULL);

        接下去的for循环，就是用来处理传输数据中的Binder对象了。在我们的场景中，有一个类型为BINDER_TYPE_BINDER的Binder实体MediaPlayerService：

[cpp] view plain copy print ?

1.    switch (fp->type) {  
2.    case BINDER_TYPE_BINDER:  
3.    case BINDER_TYPE_WEAK_BINDER: {  
4. struct binder_ref *ref;  
5. struct binder_node *node = binder_get_node(proc, fp->binder);  
6. if (node == NULL) {  
7.     node = binder_new_node(proc, fp->binder, fp->cookie);  
8.     if (node == NULL) {  
9.         return_error = BR_FAILED_REPLY;  
10.         goto err_binder_new_node_failed;  
11.     }  
12.     node->min_priority = fp->flags & FLAT_BINDER_FLAG_PRIORITY_MASK;  
13.     node->accept_fds = !!(fp->flags & FLAT_BINDER_FLAG_ACCEPTS_FDS);  
14. }  
15. if (fp->cookie != node->cookie) {  
16.     ......  
17.     goto err_binder_get_ref_for_node_failed;  
18. }  
19. ref = binder_get_ref_for_node(target_proc, node);  
20. if (ref == NULL) {  
21.     return_error = BR_FAILED_REPLY;  
22.     goto err_binder_get_ref_for_node_failed;  
23. }  
24. if (fp->type == BINDER_TYPE_BINDER)  
25.     fp->type = BINDER_TYPE_HANDLE;  
26. else  
27.     fp->type = BINDER_TYPE_WEAK_HANDLE;  
28. fp->handle = ref->desc;  
29. binder_inc_ref(ref, fp->type == BINDER_TYPE_HANDLE, &thread->todo);  
30. ......  
31.                             
32. } break;  

    switch (fp->type) {
    case BINDER_TYPE_BINDER:
    case BINDER_TYPE_WEAK_BINDER: {
	struct binder_ref *ref;
	struct binder_node *node = binder_get_node(proc, fp->binder);
	if (node == NULL) {
		node = binder_new_node(proc, fp->binder, fp->cookie);
		if (node == NULL) {
			return_error = BR_FAILED_REPLY;
			goto err_binder_new_node_failed;
		}
		node->min_priority = fp->flags & FLAT_BINDER_FLAG_PRIORITY_MASK;
		node->accept_fds = !!(fp->flags & FLAT_BINDER_FLAG_ACCEPTS_FDS);
	}
	if (fp->cookie != node->cookie) {
		......
		goto err_binder_get_ref_for_node_failed;
	}
	ref = binder_get_ref_for_node(target_proc, node);
	if (ref == NULL) {
		return_error = BR_FAILED_REPLY;
		goto err_binder_get_ref_for_node_failed;
	}
	if (fp->type == BINDER_TYPE_BINDER)
		fp->type = BINDER_TYPE_HANDLE;
	else
		fp->type = BINDER_TYPE_WEAK_HANDLE;
	fp->handle = ref->desc;
	binder_inc_ref(ref, fp->type == BINDER_TYPE_HANDLE, &thread->todo);
	......
							  
	} break;

        由于是第一次在Binder驱动程序中传输这个MediaPlayerService，调用binder_get_node函数查询这个Binder实体时，会返回空，于是binder_new_node在proc中新建一个，下次就可以直接使用了。

        现在，由于要把这个Binder实体MediaPlayerService交给target_proc，也就是Service Manager来管理，也就是说Service Manager要引用这个MediaPlayerService了，于是通过binder_get_ref_for_node为MediaPlayerService创建一个引用，并且通过binder_inc_ref来增加这个引用计数，防止这个引用还在使用过程当中就被销毁。注意，到了这里的时候，t->buffer中的flat_binder_obj的type已经改为BINDER_TYPE_HANDLE，handle已经改为ref->desc，跟原来不一样了，因为这个flat_binder_obj是最终是要传给Service Manager的，而Service Manager只能够通过句柄值来引用这个Binder实体。

        最后，把待处理事务加入到target_list列表中去：

[cpp] view plain copy print ?

1. list_add_tail(&t->work.entry, target_list);  

list_add_tail(&t->work.entry, target_list);

        并且把待完成工作项加入到本线程的todo等待执行列表中去：

[cpp] view plain copy print ?

1. list_add_tail(&tcomplete->entry, &thread->todo);  

list_add_tail(&tcomplete->entry, &thread->todo);

        现在目标进程有事情可做了，于是唤醒它：

[cpp] view plain copy print ?

1. if (target_wait)  
2.     wake_up_interruptible(target_wait);  

if (target_wait)
	wake_up_interruptible(target_wait);

       这里就是要唤醒Service Manager进程了。回忆一下前面浅谈Service Manager成为Android进程间通信（IPC）机制Binder守护进程之路这篇文章，此时， Service Manager正在binder_thread_read函数中调用wait_event_interruptible进入休眠状态。

       这里我们先忽略一下Service Manager被唤醒之后的场景，继续MedaPlayerService的启动过程，然后再回来。

       回到binder_ioctl函数，bwr.read_size > 0为true，于是进入binder_thread_read函数：

[cpp] view plain copy print ?

1. static int  
2. binder_thread_read(struct binder_proc *proc, struct binder_thread *thread,  
3.                    void  __user *buffer, int size, signed long *consumed, int non_block)  
4. {  
5.     void __user *ptr = buffer + *consumed;  
6.     void __user *end = buffer + size;  
7.   
8.     int ret = 0;  
9.     int wait_for_proc_work;  
10.   
11.     if (*consumed == 0) {  
12.         if (put_user(BR_NOOP, (uint32_t __user *)ptr))  
13.             return -EFAULT;  
14.         ptr += sizeof(uint32_t);  
15.     }  
16.   
17. retry:  
18.     wait_for_proc_work = thread->transaction_stack == NULL && list_empty(&thread->todo);  
19.       
20.     .......  
21.   
22.     if (wait_for_proc_work) {  
23.         .......  
24.     } else {  
25.         if (non_block) {  
26.             if (!binder_has_thread_work(thread))  
27.                 ret = -EAGAIN;  
28.         } else  
29.             ret = wait_event_interruptible(thread->wait, binder_has_thread_work(thread));  
30.     }  
31.       
32.     ......  
33.   
34.     while (1) {  
35.         uint32_t cmd;  
36.         struct binder_transaction_data tr;  
37.         struct binder_work *w;  
38.         struct binder_transaction *t = NULL;  
39.   
40.         if (!list_empty(&thread->todo))  
41.             w = list_first_entry(&thread->todo, struct binder_work, entry);  
42.         else if (!list_empty(&proc->todo) && wait_for_proc_work)  
43.             w = list_first_entry(&proc->todo, struct binder_work, entry);  
44.         else {  
45.             if (ptr - buffer == 4 && !(thread->looper & BINDER_LOOPER_STATE_NEED_RETURN)) /* no data added */  
46.                 goto retry;  
47.             break;  
48.         }  
49.   
50.         if (end - ptr < sizeof(tr) + 4)  
51.             break;  
52.   
53.         switch (w->type) {  
54.         ......  
55.         case BINDER_WORK_TRANSACTION_COMPLETE: {  
56.             cmd = BR_TRANSACTION_COMPLETE;  
57.             if (put_user(cmd, (uint32_t __user *)ptr))  
58.                 return -EFAULT;  
59.             ptr += sizeof(uint32_t);  
60.   
61.             binder_stat_br(proc, thread, cmd);  
62.             if (binder_debug_mask & BINDER_DEBUG_TRANSACTION_COMPLETE)  
63.                 printk(KERN_INFO "binder: %d:%d BR_TRANSACTION_COMPLETE\n",  
64.                 proc->pid, thread->pid);  
65.   
66.             list_del(&w->entry);  
67.             kfree(w);  
68.             binder_stats.obj_deleted[BINDER_STAT_TRANSACTION_COMPLETE]++;  
69.                                                } break;  
70.         ......  
71.         }  
72.   
73.         if (!t)  
74.             continue;  
75.   
76.         ......  
77.     }  
78.   
79. done:  
80.     ......  
81.     return 0;  
82. }  

static int
binder_thread_read(struct binder_proc *proc, struct binder_thread *thread,
				   void  __user *buffer, int size, signed long *consumed, int non_block)
{
	void __user *ptr = buffer + *consumed;
	void __user *end = buffer + size;

	int ret = 0;
	int wait_for_proc_work;

	if (*consumed == 0) {
		if (put_user(BR_NOOP, (uint32_t __user *)ptr))
			return -EFAULT;
		ptr += sizeof(uint32_t);
	}

retry:
	wait_for_proc_work = thread->transaction_stack == NULL && list_empty(&thread->todo);
	
	.......

	if (wait_for_proc_work) {
		.......
	} else {
		if (non_block) {
			if (!binder_has_thread_work(thread))
				ret = -EAGAIN;
		} else
			ret = wait_event_interruptible(thread->wait, binder_has_thread_work(thread));
	}
	
	......

	while (1) {
		uint32_t cmd;
		struct binder_transaction_data tr;
		struct binder_work *w;
		struct binder_transaction *t = NULL;

		if (!list_empty(&thread->todo))
			w = list_first_entry(&thread->todo, struct binder_work, entry);
		else if (!list_empty(&proc->todo) && wait_for_proc_work)
			w = list_first_entry(&proc->todo, struct binder_work, entry);
		else {
			if (ptr - buffer == 4 && !(thread->looper & BINDER_LOOPER_STATE_NEED_RETURN)) /* no data added */
				goto retry;
			break;
		}

		if (end - ptr < sizeof(tr) + 4)
			break;

		switch (w->type) {
		......
		case BINDER_WORK_TRANSACTION_COMPLETE: {
			cmd = BR_TRANSACTION_COMPLETE;
			if (put_user(cmd, (uint32_t __user *)ptr))
				return -EFAULT;
			ptr += sizeof(uint32_t);

			binder_stat_br(proc, thread, cmd);
			if (binder_debug_mask & BINDER_DEBUG_TRANSACTION_COMPLETE)
				printk(KERN_INFO "binder: %d:%d BR_TRANSACTION_COMPLETE\n",
				proc->pid, thread->pid);

			list_del(&w->entry);
			kfree(w);
			binder_stats.obj_deleted[BINDER_STAT_TRANSACTION_COMPLETE]++;
											   } break;
		......
		}

		if (!t)
			continue;

		......
	}

done:
	......
	return 0;
}

        这里，thread->transaction_stack和thread->todo均不为空，于是wait_for_proc_work为false，由于binder_has_thread_work的时候，返回true，这里因为thread->todo不为空，因此，线程虽然调用了wait_event_interruptible，但是不会睡眠，于是继续往下执行。

        由于thread->todo不为空，执行下列语句：

[cpp] view plain copy print ?

1. if (!list_empty(&thread->todo))  
2.      w = list_first_entry(&thread->todo, struct binder_work, entry);  

if (!list_empty(&thread->todo))
     w = list_first_entry(&thread->todo, struct binder_work, entry);

        w->type为BINDER_WORK_TRANSACTION_COMPLETE，这是在上面的binder_transaction函数设置的，于是执行：

[cpp] view plain copy print ?

1.    switch (w->type) {  
2.    ......  
3.    case BINDER_WORK_TRANSACTION_COMPLETE: {  
4. cmd = BR_TRANSACTION_COMPLETE;  
5. if (put_user(cmd, (uint32_t __user *)ptr))  
6.     return -EFAULT;  
7. ptr += sizeof(uint32_t);  
8.   
9.        ......  
10. list_del(&w->entry);  
11. kfree(w);  
12.           
13. } break;  
14. ......  
15.    }  

    switch (w->type) {
    ......
    case BINDER_WORK_TRANSACTION_COMPLETE: {
	cmd = BR_TRANSACTION_COMPLETE;
	if (put_user(cmd, (uint32_t __user *)ptr))
		return -EFAULT;
	ptr += sizeof(uint32_t);

        ......
	list_del(&w->entry);
	kfree(w);
			
	} break;
	......
    }

        这里就将w从thread->todo删除了。由于这里t为空，重新执行while循环，这时由于已经没有事情可做了，最后就返回到binder_ioctl函数中。注间，这里一共往用户传进来的缓冲区buffer写入了两个整数，分别是BR_NOOP和BR_TRANSACTION_COMPLETE。

        binder_ioctl函数返回到用户空间之前，把数据消耗情况拷贝回用户空间中：

[cpp] view plain copy print ?

1. if (copy_to_user(ubuf, &bwr, sizeof(bwr))) {  
2.     ret = -EFAULT;  
3.     goto err;  
4. }  

if (copy_to_user(ubuf, &bwr, sizeof(bwr))) {
	ret = -EFAULT;
	goto err;
}

        最后返回到IPCThreadState::talkWithDriver函数中，执行下面语句：

[cpp] view plain copy print ?

1.     if (err >= NO_ERROR) {  
2.         if (bwr.write_consumed > 0) {  
3.             if (bwr.write_consumed < (ssize_t)mOut.dataSize())  
4.                 mOut.remove(0, bwr.write_consumed);  
5.             else  
6.                 mOut.setDataSize(0);  
7.         }  
8.         if (bwr.read_consumed > 0) {  
9. <PRE class=cpp name="code" snippet_file_name="blog_20131230_54_6706870" code_snippet_id="134056">            mIn.setDataSize(bwr.read_consumed);  
10.             mIn.setDataPosition(0);</PRE>        }        ......        return NO_ERROR;    }  

    if (err >= NO_ERROR) {
        if (bwr.write_consumed > 0) {
            if (bwr.write_consumed < (ssize_t)mOut.dataSize())
                mOut.remove(0, bwr.write_consumed);
            else
                mOut.setDataSize(0);
        }
        if (bwr.read_consumed > 0) {
<div class="dp-highlighter bg_cpp" sizcache="28" sizset="400"><div class="bar" sizcache="28" sizset="400"><div class="tools" sizcache="28" sizset="400"><strong>[cpp]</strong> <a target=_blank title="view plain" class="ViewSource" href="http://blog.csdn.net/luoshengyang/article/details/6629298#">view plain</a><a target=_blank title="copy" class="CopyToClipboard" href="http://blog.csdn.net/luoshengyang/article/details/6629298#">copy</a><a target=_blank title="print" class="PrintSource" href="http://blog.csdn.net/luoshengyang/article/details/6629298#">print</a><a target=_blank title="?" class="About" href="http://blog.csdn.net/luoshengyang/article/details/6629298#">?</a><a target=_blank title="在CODE上查看代码片" style="TEXT-INDENT: 0px" href="https://code.csdn.net/snippets/134056" target="_blank"><img style="POSITION: relative; LEFT: 2px; TOP: 1px" alt="在CODE上查看代码片" src="https://code.csdn.net/assets/CODE_ico.png" width="12" height="12" /></a><a target=_blank title="派生到我的代码片" style="TEXT-INDENT: 0px" href="https://code.csdn.net/snippets/134056/fork" target="_blank"><img style="POSITION: relative; LEFT: 2px; TOP: 2px" alt="派生到我的代码片" src="https://code.csdn.net/assets/ico_fork.svg" width="12" height="12" /></a></div></div><ol class="dp-cpp"><li class="alt"><span><span>mIn.setDataSize(bwr.read_consumed);  </span></span></li><li><span>mIn.setDataPosition(0);  </span></li></ol></div><pre class="cpp" style="DISPLAY: none" code_snippet_id="134056" snippet_file_name="blog_20131230_54_6706870" name="code">            mIn.setDataSize(bwr.read_consumed);
            mIn.setDataPosition(0);

} ...... return NO_ERROR; }

        首先是把mOut的数据清空：

[cpp] view plain copy print ?

1. mOut.setDataSize(0);  

    mOut.setDataSize(0);

        然后设置已经读取的内容的大小：

[cpp] view plain copy print ?

1. mIn.setDataSize(bwr.read_consumed);  
2. mIn.setDataPosition(0);  

    mIn.setDataSize(bwr.read_consumed);
    mIn.setDataPosition(0);

        然后返回到IPCThreadState::waitForResponse函数中。在IPCThreadState::waitForResponse函数，先是从mIn读出一个整数，这个便是BR_NOOP了，这是一个空操作，什么也不做。然后继续进入IPCThreadState::talkWithDriver函数中。
        这时候，下面语句执行后：

[cpp] view plain copy print ?

1. const bool needRead = mIn.dataPosition() >= mIn.dataSize();  

const bool needRead = mIn.dataPosition() >= mIn.dataSize();

        needRead为false，因为在mIn中，尚有一个整数BR_TRANSACTION_COMPLETE未读出。

       这时候，下面语句执行后：

[cpp] view plain copy print ?

1. const size_t outAvail = (!doReceive || needRead) ? mOut.dataSize() : 0;  

const size_t outAvail = (!doReceive || needRead) ? mOut.dataSize() : 0;

        outAvail等于0。因此，最后bwr.write_size和bwr.read_size均为0，IPCThreadState::talkWithDriver函数什么也不做，直接返回到IPCThreadState::waitForResponse函数中。在IPCThreadState::waitForResponse函数，又继续从mIn读出一个整数，这个便是BR_TRANSACTION_COMPLETE：

[cpp] view plain copy print ?

1. switch (cmd) {  
2. case BR_TRANSACTION_COMPLETE:  
3.        if (!reply && !acquireResult) goto finish;  
4.        break;  
5. ......  
6. }  

switch (cmd) {
case BR_TRANSACTION_COMPLETE:
       if (!reply && !acquireResult) goto finish;
       break;
......
}

        reply不为NULL，因此，IPCThreadState::waitForResponse的循环没有结束，继续执行，又进入到IPCThreadState::talkWithDrive中。

        这次，needRead就为true了，而outAvail仍为0，所以bwr.read_size不为0，bwr.write_size为0。于是通过：

[cpp] view plain copy print ?

1. ioctl(mProcess->mDriverFD, BINDER_WRITE_READ, &bwr)  

ioctl(mProcess->mDriverFD, BINDER_WRITE_READ, &bwr)

        进入到Binder驱动程序中的binder_ioctl函数中。由于bwr.write_size为0，bwr.read_size不为0，这次直接就进入到binder_thread_read函数中。这时候，thread->transaction_stack不等于0，但是thread->todo为空，于是线程就通过：

[cpp] view plain copy print ?

1. wait_event_interruptible(thread->wait, binder_has_thread_work(thread));  

wait_event_interruptible(thread->wait, binder_has_thread_work(thread));

        进入睡眠状态，等待Service Manager来唤醒了。

        现在，我们可以回到Service Manager被唤醒的过程了。我们接着前面浅谈Service Manager成为Android进程间通信（IPC）机制Binder守护进程之路这篇文章的最后，继续描述。此时， Service Manager正在binder_thread_read函数中调用wait_event_interruptible_exclusive进入休眠状态。上面被MediaPlayerService启动后进程唤醒后，继续执行binder_thread_read函数：

[cpp] view plain copy print ?

1. static int  
2. binder_thread_read(struct binder_proc *proc, struct binder_thread *thread,  
3.                    void  __user *buffer, int size, signed long *consumed, int non_block)  
4. {  
5.     void __user *ptr = buffer + *consumed;  
6.     void __user *end = buffer + size;  
7.   
8.     int ret = 0;  
9.     int wait_for_proc_work;  
10.   
11.     if (*consumed == 0) {  
12.         if (put_user(BR_NOOP, (uint32_t __user *)ptr))  
13.             return -EFAULT;  
14.         ptr += sizeof(uint32_t);  
15.     }  
16.   
17. retry:  
18.     wait_for_proc_work = thread->transaction_stack == NULL && list_empty(&thread->todo);  
19.   
20.     ......  
21.   
22.     if (wait_for_proc_work) {  
23.         ......  
24.         if (non_block) {  
25.             if (!binder_has_proc_work(proc, thread))  
26.                 ret = -EAGAIN;  
27.         } else  
28.             ret = wait_event_interruptible_exclusive(proc->wait, binder_has_proc_work(proc, thread));  
29.     } else {  
30.         ......  
31.     }  
32.       
33.     ......  
34.   
35.     while (1) {  
36.         uint32_t cmd;  
37.         struct binder_transaction_data tr;  
38.         struct binder_work *w;  
39.         struct binder_transaction *t = NULL;  
40.   
41.         if (!list_empty(&thread->todo))  
42.             w = list_first_entry(&thread->todo, struct binder_work, entry);  
43.         else if (!list_empty(&proc->todo) && wait_for_proc_work)  
44.             w = list_first_entry(&proc->todo, struct binder_work, entry);  
45.         else {  
46.             if (ptr - buffer == 4 && !(thread->looper & BINDER_LOOPER_STATE_NEED_RETURN)) /* no data added */  
47.                 goto retry;  
48.             break;  
49.         }  
50.   
51.         if (end - ptr < sizeof(tr) + 4)  
52.             break;  
53.   
54.         switch (w->type) {  
55.         case BINDER_WORK_TRANSACTION: {  
56.             t = container_of(w, struct binder_transaction, work);  
57.                                       } break;  
58.         ......  
59.         }  
60.   
61.         if (!t)  
62.             continue;  
63.   
64.         BUG_ON(t->buffer == NULL);  
65.         if (t->buffer->target_node) {  
66.             struct binder_node *target_node = t->buffer->target_node;  
67.             tr.target.ptr = target_node->ptr;  
68.             tr.cookie =  target_node->cookie;  
69.             ......  
70.             cmd = BR_TRANSACTION;  
71.         } else {  
72.             ......  
73.         }  
74.         tr.code = t->code;  
75.         tr.flags = t->flags;  
76.         tr.sender_euid = t->sender_euid;  
77.   
78.         if (t->from) {  
79.             struct task_struct *sender = t->from->proc->tsk;  
80.             tr.sender_pid = task_tgid_nr_ns(sender, current->nsproxy->pid_ns);  
81.         } else {  
82.             tr.sender_pid = 0;  
83.         }  
84.   
85.         tr.data_size = t->buffer->data_size;  
86.         tr.offsets_size = t->buffer->offsets_size;  
87.         tr.data.ptr.buffer = (void *)t->buffer->data + proc->user_buffer_offset;  
88.         tr.data.ptr.offsets = tr.data.ptr.buffer + ALIGN(t->buffer->data_size, sizeof(void *));  
89.   
90.         if (put_user(cmd, (uint32_t __user *)ptr))  
91.             return -EFAULT;  
92.         ptr += sizeof(uint32_t);  
93.         if (copy_to_user(ptr, &tr, sizeof(tr)))  
94.             return -EFAULT;  
95.         ptr += sizeof(tr);  
96.   
97.         ......  
98.   
99.         list_del(&t->work.entry);  
100.         t->buffer->allow_user_free = 1;  
101.         if (cmd == BR_TRANSACTION && !(t->flags & TF_ONE_WAY)) {  
102.             t->to_parent = thread->transaction_stack;  
103.             t->to_thread = thread;  
104.             thread->transaction_stack = t;  
105.         } else {  
106.             t->buffer->transaction = NULL;  
107.             kfree(t);  
108.             binder_stats.obj_deleted[BINDER_STAT_TRANSACTION]++;  
109.         }  
110.         break;  
111.     }  
112.   
113. done:  
114.   
115.     ......  
116.     return 0;  
117. }  

static int
binder_thread_read(struct binder_proc *proc, struct binder_thread *thread,
				   void  __user *buffer, int size, signed long *consumed, int non_block)
{
	void __user *ptr = buffer + *consumed;
	void __user *end = buffer + size;

	int ret = 0;
	int wait_for_proc_work;

	if (*consumed == 0) {
		if (put_user(BR_NOOP, (uint32_t __user *)ptr))
			return -EFAULT;
		ptr += sizeof(uint32_t);
	}

retry:
	wait_for_proc_work = thread->transaction_stack == NULL && list_empty(&thread->todo);

	......

	if (wait_for_proc_work) {
		......
		if (non_block) {
			if (!binder_has_proc_work(proc, thread))
				ret = -EAGAIN;
		} else
			ret = wait_event_interruptible_exclusive(proc->wait, binder_has_proc_work(proc, thread));
	} else {
		......
	}
	
	......

	while (1) {
		uint32_t cmd;
		struct binder_transaction_data tr;
		struct binder_work *w;
		struct binder_transaction *t = NULL;

		if (!list_empty(&thread->todo))
			w = list_first_entry(&thread->todo, struct binder_work, entry);
		else if (!list_empty(&proc->todo) && wait_for_proc_work)
			w = list_first_entry(&proc->todo, struct binder_work, entry);
		else {
			if (ptr - buffer == 4 && !(thread->looper & BINDER_LOOPER_STATE_NEED_RETURN)) /* no data added */
				goto retry;
			break;
		}

		if (end - ptr < sizeof(tr) + 4)
			break;

		switch (w->type) {
		case BINDER_WORK_TRANSACTION: {
			t = container_of(w, struct binder_transaction, work);
									  } break;
		......
		}

		if (!t)
			continue;

		BUG_ON(t->buffer == NULL);
		if (t->buffer->target_node) {
			struct binder_node *target_node = t->buffer->target_node;
			tr.target.ptr = target_node->ptr;
			tr.cookie =  target_node->cookie;
			......
			cmd = BR_TRANSACTION;
		} else {
			......
		}
		tr.code = t->code;
		tr.flags = t->flags;
		tr.sender_euid = t->sender_euid;

		if (t->from) {
			struct task_struct *sender = t->from->proc->tsk;
			tr.sender_pid = task_tgid_nr_ns(sender, current->nsproxy->pid_ns);
		} else {
			tr.sender_pid = 0;
		}

		tr.data_size = t->buffer->data_size;
		tr.offsets_size = t->buffer->offsets_size;
		tr.data.ptr.buffer = (void *)t->buffer->data + proc->user_buffer_offset;
		tr.data.ptr.offsets = tr.data.ptr.buffer + ALIGN(t->buffer->data_size, sizeof(void *));

		if (put_user(cmd, (uint32_t __user *)ptr))
			return -EFAULT;
		ptr += sizeof(uint32_t);
		if (copy_to_user(ptr, &tr, sizeof(tr)))
			return -EFAULT;
		ptr += sizeof(tr);

		......

		list_del(&t->work.entry);
		t->buffer->allow_user_free = 1;
		if (cmd == BR_TRANSACTION && !(t->flags & TF_ONE_WAY)) {
			t->to_parent = thread->transaction_stack;
			t->to_thread = thread;
			thread->transaction_stack = t;
		} else {
			t->buffer->transaction = NULL;
			kfree(t);
			binder_stats.obj_deleted[BINDER_STAT_TRANSACTION]++;
		}
		break;
	}

done:

    ......
	return 0;
}

        Service Manager被唤醒之后，就进入while循环开始处理事务了。这里wait_for_proc_work等于1，并且proc->todo不为空，所以从proc->todo列表中得到第一个工作项：

[cpp] view plain copy print ?

1. w = list_first_entry(&proc->todo, struct binder_work, entry);  

w = list_first_entry(&proc->todo, struct binder_work, entry);

        从上面的描述中，我们知道，这个工作项的类型为BINDER_WORK_TRANSACTION，于是通过下面语句得到事务项：

[cpp] view plain copy print ?

1. t = container_of(w, struct binder_transaction, work);  

t = container_of(w, struct binder_transaction, work);

       接着就是把事务项t中的数据拷贝到本地局部变量struct binder_transaction_data tr中去了：

[cpp] view plain copy print ?

1. if (t->buffer->target_node) {  
2.     struct binder_node *target_node = t->buffer->target_node;  
3.     tr.target.ptr = target_node->ptr;  
4.     tr.cookie =  target_node->cookie;  
5.     ......  
6.     cmd = BR_TRANSACTION;  
7. } else {  
8.     ......  
9. }  
10. tr.code = t->code;  
11. tr.flags = t->flags;  
12. tr.sender_euid = t->sender_euid;  
13.   
14. if (t->from) {  
15.     struct task_struct *sender = t->from->proc->tsk;  
16.     tr.sender_pid = task_tgid_nr_ns(sender, current->nsproxy->pid_ns);  
17. } else {  
18.     tr.sender_pid = 0;  
19. }  
20.   
21. tr.data_size = t->buffer->data_size;  
22. tr.offsets_size = t->buffer->offsets_size;  
23. tr.data.ptr.buffer = (void *)t->buffer->data + proc->user_buffer_offset;  
24. tr.data.ptr.offsets = tr.data.ptr.buffer + ALIGN(t->buffer->data_size, sizeof(void *));  

if (t->buffer->target_node) {
	struct binder_node *target_node = t->buffer->target_node;
	tr.target.ptr = target_node->ptr;
	tr.cookie =  target_node->cookie;
	......
	cmd = BR_TRANSACTION;
} else {
	......
}
tr.code = t->code;
tr.flags = t->flags;
tr.sender_euid = t->sender_euid;

if (t->from) {
	struct task_struct *sender = t->from->proc->tsk;
	tr.sender_pid = task_tgid_nr_ns(sender, current->nsproxy->pid_ns);
} else {
	tr.sender_pid = 0;
}

tr.data_size = t->buffer->data_size;
tr.offsets_size = t->buffer->offsets_size;
tr.data.ptr.buffer = (void *)t->buffer->data + proc->user_buffer_offset;
tr.data.ptr.offsets = tr.data.ptr.buffer + ALIGN(t->buffer->data_size, sizeof(void *));

        这里有一个非常重要的地方，是Binder进程间通信机制的精髓所在：

[cpp] view plain copy print ?

1. tr.data.ptr.buffer = (void *)t->buffer->data + proc->user_buffer_offset;  
2. tr.data.ptr.offsets = tr.data.ptr.buffer + ALIGN(t->buffer->data_size, sizeof(void *));  

tr.data.ptr.buffer = (void *)t->buffer->data + proc->user_buffer_offset;
tr.data.ptr.offsets = tr.data.ptr.buffer + ALIGN(t->buffer->data_size, sizeof(void *));

        t->buffer->data所指向的地址是内核空间的，现在要把数据返回给Service Manager进程的用户空间，而Service Manager进程的用户空间是不能访问内核空间的数据的，所以这里要作一下处理。怎么处理呢？我们在学面向对象语言的时候，对象的拷贝有深拷贝和浅拷贝之分，深拷贝是把另外分配一块新内存，然后把原始对象的内容搬过去，浅拷贝是并没有为新对象分配一块新空间，而只是分配一个引用，而个引用指向原始对象。Binder机制用的是类似浅拷贝的方法，通过在用户空间分配一个虚拟地址，然后让这个用户空间虚拟地址与 t->buffer->data这个内核空间虚拟地址指向同一个物理地址，这样就可以实现浅拷贝了。怎么样用户空间和内核空间的虚拟地址同时指向同一个物理地址呢？请参考前面一篇文章浅谈Service Manager成为Android进程间通信（IPC）机制Binder守护进程之路，那里有详细描述。这里只要将t->buffer->data加上一个偏移值proc->user_buffer_offset就可以得到t->buffer->data对应的用户空间虚拟地址了。调整了tr.data.ptr.buffer的值之后，不要忘记也要一起调整tr.data.ptr.offsets的值。

        接着就是把tr的内容拷贝到用户传进来的缓冲区去了，指针ptr指向这个用户缓冲区的地址：

[cpp] view plain copy print ?

1. if (put_user(cmd, (uint32_t __user *)ptr))  
2.     return -EFAULT;  
3. ptr += sizeof(uint32_t);  
4. if (copy_to_user(ptr, &tr, sizeof(tr)))  
5.     return -EFAULT;  
6. ptr += sizeof(tr);  

if (put_user(cmd, (uint32_t __user *)ptr))
	return -EFAULT;
ptr += sizeof(uint32_t);
if (copy_to_user(ptr, &tr, sizeof(tr)))
	return -EFAULT;
ptr += sizeof(tr);

         这里可以看出，这里只是对作tr.data.ptr.bufferr和tr.data.ptr.offsets的内容作了浅拷贝。

         最后，由于已经处理了这个事务，要把它从todo列表中删除：

[cpp] view plain copy print ?

1. list_del(&t->work.entry);  
2. t->buffer->allow_user_free = 1;  
3. if (cmd == BR_TRANSACTION && !(t->flags & TF_ONE_WAY)) {  
4.     t->to_parent = thread->transaction_stack;  
5.     t->to_thread = thread;  
6.     thread->transaction_stack = t;  
7. } else {  
8.     t->buffer->transaction = NULL;  
9.     kfree(t);  
10.     binder_stats.obj_deleted[BINDER_STAT_TRANSACTION]++;  
11. }  

list_del(&t->work.entry);
t->buffer->allow_user_free = 1;
if (cmd == BR_TRANSACTION && !(t->flags & TF_ONE_WAY)) {
	t->to_parent = thread->transaction_stack;
	t->to_thread = thread;
	thread->transaction_stack = t;
} else {
	t->buffer->transaction = NULL;
	kfree(t);
	binder_stats.obj_deleted[BINDER_STAT_TRANSACTION]++;
}

         注意，这里的cmd == BR_TRANSACTION && !(t->flags & TF_ONE_WAY)为true，表明这个事务虽然在驱动程序中已经处理完了，但是它仍然要等待Service Manager完成之后，给驱动程序一个确认，也就是需要等待回复，于是把当前事务t放在thread->transaction_stack队列的头部：

[cpp] view plain copy print ?

1. t->to_parent = thread->transaction_stack;  
2. t->to_thread = thread;  
3. thread->transaction_stack = t;  

t->to_parent = thread->transaction_stack;
t->to_thread = thread;
thread->transaction_stack = t;

         如果cmd == BR_TRANSACTION && !(t->flags & TF_ONE_WAY)为false，那就不需要等待回复了，直接把事务t删掉。

         这个while最后通过一个break跳了出来，最后返回到binder_ioctl函数中：

[cpp] view plain copy print ?

1. static long binder_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)  
2. {  
3.     int ret;  
4.     struct binder_proc *proc = filp->private_data;  
5.     struct binder_thread *thread;  
6.     unsigned int size = _IOC_SIZE(cmd);  
7.     void __user *ubuf = (void __user *)arg;  
8.   
9.     ......  
10.   
11.     switch (cmd) {  
12.     case BINDER_WRITE_READ: {  
13.         struct binder_write_read bwr;  
14.         if (size != sizeof(struct binder_write_read)) {  
15.             ret = -EINVAL;  
16.             goto err;  
17.         }  
18.         if (copy_from_user(&bwr, ubuf, sizeof(bwr))) {  
19.             ret = -EFAULT;  
20.             goto err;  
21.         }  
22.         ......  
23.         if (bwr.read_size > 0) {  
24.             ret = binder_thread_read(proc, thread, (void __user *)bwr.read_buffer, bwr.read_size, &bwr.read_consumed, filp->f_flags & O_NONBLOCK);  
25.             if (!list_empty(&proc->todo))  
26.                 wake_up_interruptible(&proc->wait);  
27.             if (ret < 0) {  
28.                 if (copy_to_user(ubuf, &bwr, sizeof(bwr)))  
29.                     ret = -EFAULT;  
30.                 goto err;  
31.             }  
32.         }  
33.         ......  
34.         if (copy_to_user(ubuf, &bwr, sizeof(bwr))) {  
35.             ret = -EFAULT;  
36.             goto err;  
37.         }  
38.         break;  
39.         }  
40.     ......  
41.     default:  
42.         ret = -EINVAL;  
43.         goto err;  
44.     }  
45.     ret = 0;  
46. err:  
47.     ......  
48.     return ret;  
49. }  

static long binder_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
{
	int ret;
	struct binder_proc *proc = filp->private_data;
	struct binder_thread *thread;
	unsigned int size = _IOC_SIZE(cmd);
	void __user *ubuf = (void __user *)arg;

	......

	switch (cmd) {
	case BINDER_WRITE_READ: {
		struct binder_write_read bwr;
		if (size != sizeof(struct binder_write_read)) {
			ret = -EINVAL;
			goto err;
		}
		if (copy_from_user(&bwr, ubuf, sizeof(bwr))) {
			ret = -EFAULT;
			goto err;
		}
		......
		if (bwr.read_size > 0) {
			ret = binder_thread_read(proc, thread, (void __user *)bwr.read_buffer, bwr.read_size, &bwr.read_consumed, filp->f_flags & O_NONBLOCK);
			if (!list_empty(&proc->todo))
				wake_up_interruptible(&proc->wait);
			if (ret < 0) {
				if (copy_to_user(ubuf, &bwr, sizeof(bwr)))
					ret = -EFAULT;
				goto err;
			}
		}
		......
		if (copy_to_user(ubuf, &bwr, sizeof(bwr))) {
			ret = -EFAULT;
			goto err;
		}
		break;
	    }
	......
	default:
		ret = -EINVAL;
		goto err;
	}
	ret = 0;
err:
	......
	return ret;
}

         从binder_thread_read返回来后，再看看proc->todo是否还有事务等待处理，如果是，就把睡眠在proc->wait队列的线程唤醒来处理。最后，把本地变量struct binder_write_read bwr的内容拷贝回到用户传进来的缓冲区中，就返回了。

        这里就是返回到frameworks/base/cmds/servicemanager/binder.c文件中的binder_loop函数了：

[cpp] view plain copy print ?

1. void binder_loop(struct binder_state *bs, binder_handler func)  
2. {  
3.     int res;  
4.     struct binder_write_read bwr;  
5.     unsigned readbuf[32];  
6.   
7.     bwr.write_size = 0;  
8.     bwr.write_consumed = 0;  
9.     bwr.write_buffer = 0;  
10.       
11.     readbuf[0] = BC_ENTER_LOOPER;  
12.     binder_write(bs, readbuf, sizeof(unsigned));  
13.   
14.     for (;;) {  
15.         bwr.read_size = sizeof(readbuf);  
16.         bwr.read_consumed = 0;  
17.         bwr.read_buffer = (unsigned) readbuf;  
18.   
19.         res = ioctl(bs->fd, BINDER_WRITE_READ, &bwr);  
20.   
21.         if (res < 0) {  
22.             LOGE("binder_loop: ioctl failed (%s)\n", strerror(errno));  
23.             break;  
24.         }  
25.   
26.         res = binder_parse(bs, 0, readbuf, bwr.read_consumed, func);  
27.         if (res == 0) {  
28.             LOGE("binder_loop: unexpected reply?!\n");  
29.             break;  
30.         }  
31.         if (res < 0) {  
32.             LOGE("binder_loop: io error %d %s\n", res, strerror(errno));  
33.             break;  
34.         }  
35.     }  
36. }  

void binder_loop(struct binder_state *bs, binder_handler func)
{
    int res;
    struct binder_write_read bwr;
    unsigned readbuf[32];

    bwr.write_size = 0;
    bwr.write_consumed = 0;
    bwr.write_buffer = 0;
    
    readbuf[0] = BC_ENTER_LOOPER;
    binder_write(bs, readbuf, sizeof(unsigned));

    for (;;) {
        bwr.read_size = sizeof(readbuf);
        bwr.read_consumed = 0;
        bwr.read_buffer = (unsigned) readbuf;

        res = ioctl(bs->fd, BINDER_WRITE_READ, &bwr);

        if (res < 0) {
            LOGE("binder_loop: ioctl failed (%s)\n", strerror(errno));
            break;
        }

        res = binder_parse(bs, 0, readbuf, bwr.read_consumed, func);
        if (res == 0) {
            LOGE("binder_loop: unexpected reply?!\n");
            break;
        }
        if (res < 0) {
            LOGE("binder_loop: io error %d %s\n", res, strerror(errno));
            break;
        }
    }
}

       返回来的数据都放在readbuf中，接着调用binder_parse进行解析：

[cpp] view plain copy print ?

1. int binder_parse(struct binder_state *bs, struct binder_io *bio,  
2.                  uint32_t *ptr, uint32_t size, binder_handler func)  
3. {  
4.     int r = 1;  
5.     uint32_t *end = ptr + (size / 4);  
6.   
7.     while (ptr < end) {  
8.         uint32_t cmd = *ptr++;  
9.         ......  
10.         case BR_TRANSACTION: {  
11.             struct binder_txn *txn = (void *) ptr;  
12.             if ((end - ptr) * sizeof(uint32_t) < sizeof(struct binder_txn)) {  
13.                 LOGE("parse: txn too small!\n");  
14.                 return -1;  
15.             }  
16.             binder_dump_txn(txn);  
17.             if (func) {  
18.                 unsigned rdata[256/4];  
19.                 struct binder_io msg;  
20.                 struct binder_io reply;  
21.                 int res;  
22.   
23.                 bio_init(&reply, rdata, sizeof(rdata), 4);  
24.                 bio_init_from_txn(&msg, txn);  
25.                 res = func(bs, txn, &msg, &reply);  
26.                 binder_send_reply(bs, &reply, txn->data, res);  
27.             }  
28.             ptr += sizeof(*txn) / sizeof(uint32_t);  
29.             break;  
30.                              }  
31.         ......  
32.         default:  
33.             LOGE("parse: OOPS %d\n", cmd);  
34.             return -1;  
35.         }  
36.     }  
37.   
38.     return r;  
39. }  

int binder_parse(struct binder_state *bs, struct binder_io *bio,
				 uint32_t *ptr, uint32_t size, binder_handler func)
{
	int r = 1;
	uint32_t *end = ptr + (size / 4);

	while (ptr < end) {
		uint32_t cmd = *ptr++;
        ......
		case BR_TRANSACTION: {
			struct binder_txn *txn = (void *) ptr;
			if ((end - ptr) * sizeof(uint32_t) < sizeof(struct binder_txn)) {
				LOGE("parse: txn too small!\n");
				return -1;
			}
			binder_dump_txn(txn);
			if (func) {
				unsigned rdata[256/4];
				struct binder_io msg;
				struct binder_io reply;
				int res;

				bio_init(&reply, rdata, sizeof(rdata), 4);
				bio_init_from_txn(&msg, txn);
				res = func(bs, txn, &msg, &reply);
				binder_send_reply(bs, &reply, txn->data, res);
			}
			ptr += sizeof(*txn) / sizeof(uint32_t);
			break;
							 }
		......
		default:
			LOGE("parse: OOPS %d\n", cmd);
			return -1;
		}
	}

	return r;
}

        首先把从Binder驱动程序读出来的数据转换为一个struct binder_txn结构体，保存在txn本地变量中，struct binder_txn定义在frameworks/base/cmds/servicemanager/binder.h文件中：

[cpp] view plain copy print ?

1. struct binder_txn  
2. {  
3.     void *target;  
4.     void *cookie;  
5.     uint32_t code;  
6.     uint32_t flags;  
7.   
8.     uint32_t sender_pid;  
9.     uint32_t sender_euid;  
10.   
11.     uint32_t data_size;  
12.     uint32_t offs_size;  
13.     void *data;  
14.     void *offs;  
15. };  

struct binder_txn
{
    void *target;
    void *cookie;
    uint32_t code;
    uint32_t flags;

    uint32_t sender_pid;
    uint32_t sender_euid;

    uint32_t data_size;
    uint32_t offs_size;
    void *data;
    void *offs;
};

       函数中还用到了另外一个数据结构struct binder_io，也是定义在frameworks/base/cmds/servicemanager/binder.h文件中：

[cpp] view plain copy print ?

1. struct binder_io  
2. {  
3.     char *data;            /* pointer to read/write from */  
4.     uint32_t *offs;        /* array of offsets */  
5.     uint32_t data_avail;   /* bytes available in data buffer */  
6.     uint32_t offs_avail;   /* entries available in offsets array */  
7.   
8.     char *data0;           /* start of data buffer */  
9.     uint32_t *offs0;       /* start of offsets buffer */  
10.     uint32_t flags;  
11.     uint32_t unused;  
12. };  

struct binder_io
{
    char *data;            /* pointer to read/write from */
    uint32_t *offs;        /* array of offsets */
    uint32_t data_avail;   /* bytes available in data buffer */
    uint32_t offs_avail;   /* entries available in offsets array */

    char *data0;           /* start of data buffer */
    uint32_t *offs0;       /* start of offsets buffer */
    uint32_t flags;
    uint32_t unused;
};

       接着往下看，函数调bio_init来初始化reply变量：

[cpp] view plain copy print ?

1. void bio_init(struct binder_io *bio, void *data,  
2.               uint32_t maxdata, uint32_t maxoffs)  
3. {  
4.     uint32_t n = maxoffs * sizeof(uint32_t);  
5.   
6.     if (n > maxdata) {  
7.         bio->flags = BIO_F_OVERFLOW;  
8.         bio->data_avail = 0;  
9.         bio->offs_avail = 0;  
10.         return;  
11.     }  
12.   
13.     bio->data = bio->data0 = data + n;  
14.     bio->offs = bio->offs0 = data;  
15.     bio->data_avail = maxdata - n;  
16.     bio->offs_avail = maxoffs;  
17.     bio->flags = 0;  
18. }  

void bio_init(struct binder_io *bio, void *data,
              uint32_t maxdata, uint32_t maxoffs)
{
    uint32_t n = maxoffs * sizeof(uint32_t);

    if (n > maxdata) {
        bio->flags = BIO_F_OVERFLOW;
        bio->data_avail = 0;
        bio->offs_avail = 0;
        return;
    }

    bio->data = bio->data0 = data + n;
    bio->offs = bio->offs0 = data;
    bio->data_avail = maxdata - n;
    bio->offs_avail = maxoffs;
    bio->flags = 0;
}

       接着又调用bio_init_from_txn来初始化msg变量：

[cpp] view plain copy print ?

1. void bio_init_from_txn(struct binder_io *bio, struct binder_txn *txn)  
2. {  
3.     bio->data = bio->data0 = txn->data;  
4.     bio->offs = bio->offs0 = txn->offs;  
5.     bio->data_avail = txn->data_size;  
6.     bio->offs_avail = txn->offs_size / 4;  
7.     bio->flags = BIO_F_SHARED;  
8. }  

void bio_init_from_txn(struct binder_io *bio, struct binder_txn *txn)
{
    bio->data = bio->data0 = txn->data;
    bio->offs = bio->offs0 = txn->offs;
    bio->data_avail = txn->data_size;
    bio->offs_avail = txn->offs_size / 4;
    bio->flags = BIO_F_SHARED;
}

      最后，真正进行处理的函数是从参数中传进来的函数指针func，这里就是定义在frameworks/base/cmds/servicemanager/service_manager.c文件中的svcmgr_handler函数：

[cpp] view plain copy print ?

1. int svcmgr_handler(struct binder_state *bs,  
2.                    struct binder_txn *txn,  
3.                    struct binder_io *msg,  
4.                    struct binder_io *reply)  
5. {  
6.     struct svcinfo *si;  
7.     uint16_t *s;  
8.     unsigned len;  
9.     void *ptr;  
10.     uint32_t strict_policy;  
11.   
12.     if (txn->target != svcmgr_handle)  
13.         return -1;  
14.   
15.     // Equivalent to Parcel::enforceInterface(), reading the RPC   
16.     // header with the strict mode policy mask and the interface name.   
17.     // Note that we ignore the strict_policy and don't propagate it   
18.     // further (since we do no outbound RPCs anyway).   
19.     strict_policy = bio_get_uint32(msg);  
20.     s = bio_get_string16(msg, &len);  
21.     if ((len != (sizeof(svcmgr_id) / 2)) ||  
22.         memcmp(svcmgr_id, s, sizeof(svcmgr_id))) {  
23.             fprintf(stderr,"invalid id %s\n", str8(s));  
24.             return -1;  
25.     }  
26.   
27.     switch(txn->code) {  
28.     ......  
29.     case SVC_MGR_ADD_SERVICE:  
30.         s = bio_get_string16(msg, &len);  
31.         ptr = bio_get_ref(msg);  
32.         if (do_add_service(bs, s, len, ptr, txn->sender_euid))  
33.             return -1;  
34.         break;  
35.     ......  
36.     }  
37.   
38.     bio_put_uint32(reply, 0);  
39.     return 0;  
40. }  

int svcmgr_handler(struct binder_state *bs,
				   struct binder_txn *txn,
				   struct binder_io *msg,
				   struct binder_io *reply)
{
	struct svcinfo *si;
	uint16_t *s;
	unsigned len;
	void *ptr;
	uint32_t strict_policy;

	if (txn->target != svcmgr_handle)
		return -1;

	// Equivalent to Parcel::enforceInterface(), reading the RPC
	// header with the strict mode policy mask and the interface name.
	// Note that we ignore the strict_policy and don't propagate it
	// further (since we do no outbound RPCs anyway).
	strict_policy = bio_get_uint32(msg);
	s = bio_get_string16(msg, &len);
	if ((len != (sizeof(svcmgr_id) / 2)) ||
		memcmp(svcmgr_id, s, sizeof(svcmgr_id))) {
			fprintf(stderr,"invalid id %s\n", str8(s));
			return -1;
	}

	switch(txn->code) {
	......
	case SVC_MGR_ADD_SERVICE:
		s = bio_get_string16(msg, &len);
		ptr = bio_get_ref(msg);
		if (do_add_service(bs, s, len, ptr, txn->sender_euid))
			return -1;
		break;
	......
	}

	bio_put_uint32(reply, 0);
	return 0;
}

         回忆一下，在BpServiceManager::addService时，传给Binder驱动程序的参数为：

[cpp] view plain copy print ?

1. writeInt32(IPCThreadState::self()->getStrictModePolicy() | STRICT_MODE_PENALTY_GATHER);  
2. writeString16("android.os.IServiceManager");  
3. writeString16("media.player");  
4. writeStrongBinder(new MediaPlayerService());  

writeInt32(IPCThreadState::self()->getStrictModePolicy() | STRICT_MODE_PENALTY_GATHER);
writeString16("android.os.IServiceManager");
writeString16("media.player");
writeStrongBinder(new MediaPlayerService());

         这里的语句：

[cpp] view plain copy print ?

1. strict_policy = bio_get_uint32(msg);  
2. s = bio_get_string16(msg, &len);  
3. s = bio_get_string16(msg, &len);  
4. ptr = bio_get_ref(msg);  

strict_policy = bio_get_uint32(msg);
s = bio_get_string16(msg, &len);
s = bio_get_string16(msg, &len);
ptr = bio_get_ref(msg);

         就是依次把它们读取出来了，这里，我们只要看一下bio_get_ref的实现。先看一个数据结构struct binder_obj的定义：

[cpp] view plain copy print ?

1. struct binder_object  
2. {  
3.     uint32_t type;  
4.     uint32_t flags;  
5.     void *pointer;  
6.     void *cookie;  
7. };  

struct binder_object
{
    uint32_t type;
    uint32_t flags;
    void *pointer;
    void *cookie;
};

        这个结构体其实就是对应struct flat_binder_obj的。

        接着看bio_get_ref实现：

[cpp] view plain copy print ?

1. void *bio_get_ref(struct binder_io *bio)  
2. {  
3.     struct binder_object *obj;  
4.   
5.     obj = _bio_get_obj(bio);  
6.     if (!obj)  
7.         return 0;  
8.   
9.     if (obj->type == BINDER_TYPE_HANDLE)  
10.         return obj->pointer;  
11.   
12.     return 0;  
13. }  

void *bio_get_ref(struct binder_io *bio)
{
    struct binder_object *obj;

    obj = _bio_get_obj(bio);
    if (!obj)
        return 0;

    if (obj->type == BINDER_TYPE_HANDLE)
        return obj->pointer;

    return 0;
}

       _bio_get_obj这个函数就不跟进去看了，它的作用就是从binder_io中取得第一个还没取获取过的binder_object。在这个场景下，就是我们最开始传过来代表MediaPlayerService的flat_binder_obj了，这个原始的flat_binder_obj的type为BINDER_TYPE_BINDER，binder为指向MediaPlayerService的弱引用的地址。在前面我们说过，在Binder驱动驱动程序里面，会把这个flat_binder_obj的type改为BINDER_TYPE_HANDLE，handle改为一个句柄值。这里的handle值就等于obj->pointer的值。

        回到svcmgr_handler函数，调用do_add_service进一步处理：

[cpp] view plain copy print ?

1. int do_add_service(struct binder_state *bs,  
2.                    uint16_t *s, unsigned len,  
3.                    void *ptr, unsigned uid)  
4. {  
5.     struct svcinfo *si;  
6. //    LOGI("add_service('%s',%p) uid=%d\n", str8(s), ptr, uid);   
7.   
8.     if (!ptr || (len == 0) || (len > 127))  
9.         return -1;  
10.   
11.     if (!svc_can_register(uid, s)) {  
12.         LOGE("add_service('%s',%p) uid=%d - PERMISSION DENIED\n",  
13.              str8(s), ptr, uid);  
14.         return -1;  
15.     }  
16.   
17.     si = find_svc(s, len);  
18.     if (si) {  
19.         if (si->ptr) {  
20.             LOGE("add_service('%s',%p) uid=%d - ALREADY REGISTERED\n",  
21.                  str8(s), ptr, uid);  
22.             return -1;  
23.         }  
24.         si->ptr = ptr;  
25.     } else {  
26.         si = malloc(sizeof(*si) + (len + 1) * sizeof(uint16_t));  
27.         if (!si) {  
28.             LOGE("add_service('%s',%p) uid=%d - OUT OF MEMORY\n",  
29.                  str8(s), ptr, uid);  
30.             return -1;  
31.         }  
32.         si->ptr = ptr;  
33.         si->len = len;  
34.         memcpy(si->name, s, (len + 1) * sizeof(uint16_t));  
35.         si->name[len] = '\0';  
36.         si->death.func = svcinfo_death;  
37.         si->death.ptr = si;  
38.         si->next = svclist;  
39.         svclist = si;  
40.     }  
41.   
42.     binder_acquire(bs, ptr);  
43.     binder_link_to_death(bs, ptr, &si->death);  
44.     return 0;  
45. }  

int do_add_service(struct binder_state *bs,
                   uint16_t *s, unsigned len,
                   void *ptr, unsigned uid)
{
    struct svcinfo *si;
//    LOGI("add_service('%s',%p) uid=%d\n", str8(s), ptr, uid);

    if (!ptr || (len == 0) || (len > 127))
        return -1;

    if (!svc_can_register(uid, s)) {
        LOGE("add_service('%s',%p) uid=%d - PERMISSION DENIED\n",
             str8(s), ptr, uid);
        return -1;
    }

    si = find_svc(s, len);
    if (si) {
        if (si->ptr) {
            LOGE("add_service('%s',%p) uid=%d - ALREADY REGISTERED\n",
                 str8(s), ptr, uid);
            return -1;
        }
        si->ptr = ptr;
    } else {
        si = malloc(sizeof(*si) + (len + 1) * sizeof(uint16_t));
        if (!si) {
            LOGE("add_service('%s',%p) uid=%d - OUT OF MEMORY\n",
                 str8(s), ptr, uid);
            return -1;
        }
        si->ptr = ptr;
        si->len = len;
        memcpy(si->name, s, (len + 1) * sizeof(uint16_t));
        si->name[len] = '\0';
        si->death.func = svcinfo_death;
        si->death.ptr = si;
        si->next = svclist;
        svclist = si;
    }

    binder_acquire(bs, ptr);
    binder_link_to_death(bs, ptr, &si->death);
    return 0;
}

        这个函数的实现很简单，就是把MediaPlayerService这个Binder实体的引用写到一个struct svcinfo结构体中，主要是它的名称和句柄值，然后插入到链接svclist的头部去。这样，Client来向Service Manager查询服务接口时，只要给定服务名称，Service Manger就可以返回相应的句柄值了。

        这个函数执行完成后，返回到svcmgr_handler函数，函数的最后，将一个错误码0写到reply变量中去，表示一切正常：

[cpp] view plain copy print ?

1. bio_put_uint32(reply, 0);  

bio_put_uint32(reply, 0);

       svcmgr_handler函数执行完成后，返回到binder_parse函数，执行下面语句：

[cpp] view plain copy print ?

1. binder_send_reply(bs, &reply, txn->data, res);  

binder_send_reply(bs, &reply, txn->data, res);

       我们看一下binder_send_reply的实现，从函数名就可以猜到它要做什么了，告诉Binder驱动程序，它完成了Binder驱动程序交给它的任务了。

[cpp] view plain copy print ?

1. void binder_send_reply(struct binder_state *bs,  
2.                        struct binder_io *reply,  
3.                        void *buffer_to_free,  
4.                        int status)  
5. {  
6.     struct {  
7.         uint32_t cmd_free;  
8.         void *buffer;  
9.         uint32_t cmd_reply;  
10.         struct binder_txn txn;  
11.     } __attribute__((packed)) data;  
12.   
13.     data.cmd_free = BC_FREE_BUFFER;  
14.     data.buffer = buffer_to_free;  
15.     data.cmd_reply = BC_REPLY;  
16.     data.txn.target = 0;  
17.     data.txn.cookie = 0;  
18.     data.txn.code = 0;  
19.     if (status) {  
20.         data.txn.flags = TF_STATUS_CODE;  
21.         data.txn.data_size = sizeof(int);  
22.         data.txn.offs_size = 0;  
23.         data.txn.data = &status;  
24.         data.txn.offs = 0;  
25.     } else {  
26.         data.txn.flags = 0;  
27.         data.txn.data_size = reply->data - reply->data0;  
28.         data.txn.offs_size = ((char*) reply->offs) - ((char*) reply->offs0);  
29.         data.txn.data = reply->data0;  
30.         data.txn.offs = reply->offs0;  
31.     }  
32.     binder_write(bs, &data, sizeof(data));  
33. }  

void binder_send_reply(struct binder_state *bs,
                       struct binder_io *reply,
                       void *buffer_to_free,
                       int status)
{
    struct {
        uint32_t cmd_free;
        void *buffer;
        uint32_t cmd_reply;
        struct binder_txn txn;
    } __attribute__((packed)) data;

    data.cmd_free = BC_FREE_BUFFER;
    data.buffer = buffer_to_free;
    data.cmd_reply = BC_REPLY;
    data.txn.target = 0;
    data.txn.cookie = 0;
    data.txn.code = 0;
    if (status) {
        data.txn.flags = TF_STATUS_CODE;
        data.txn.data_size = sizeof(int);
        data.txn.offs_size = 0;
        data.txn.data = &status;
        data.txn.offs = 0;
    } else {
        data.txn.flags = 0;
        data.txn.data_size = reply->data - reply->data0;
        data.txn.offs_size = ((char*) reply->offs) - ((char*) reply->offs0);
        data.txn.data = reply->data0;
        data.txn.offs = reply->offs0;
    }
    binder_write(bs, &data, sizeof(data));
}

       从这里可以看出，binder_send_reply告诉Binder驱动程序执行BC_FREE_BUFFER和BC_REPLY命令，前者释放之前在binder_transaction分配的空间，地址为buffer_to_free，buffer_to_free这个地址是Binder驱动程序把自己在内核空间用的地址转换成用户空间地址再传给Service Manager的，所以Binder驱动程序拿到这个地址后，知道怎么样释放这个空间；后者告诉MediaPlayerService，它的addService操作已经完成了，错误码是0，保存在data.txn.data中。

       再来看binder_write函数：

[cpp] view plain copy print ?

1. int binder_write(struct binder_state *bs, void *data, unsigned len)  
2. {  
3.     struct binder_write_read bwr;  
4.     int res;  
5.     bwr.write_size = len;  
6.     bwr.write_consumed = 0;  
7.     bwr.write_buffer = (unsigned) data;  
8.     bwr.read_size = 0;  
9.     bwr.read_consumed = 0;  
10.     bwr.read_buffer = 0;  
11.     res = ioctl(bs->fd, BINDER_WRITE_READ, &bwr);  
12.     if (res < 0) {  
13.         fprintf(stderr,"binder_write: ioctl failed (%s)\n",  
14.                 strerror(errno));  
15.     }  
16.     return res;  
17. }  

int binder_write(struct binder_state *bs, void *data, unsigned len)
{
    struct binder_write_read bwr;
    int res;
    bwr.write_size = len;
    bwr.write_consumed = 0;
    bwr.write_buffer = (unsigned) data;
    bwr.read_size = 0;
    bwr.read_consumed = 0;
    bwr.read_buffer = 0;
    res = ioctl(bs->fd, BINDER_WRITE_READ, &bwr);
    if (res < 0) {
        fprintf(stderr,"binder_write: ioctl failed (%s)\n",
                strerror(errno));
    }
    return res;
}

       这里可以看出，只有写操作，没有读操作，即read_size为0。

       这里又是一个ioctl的BINDER_WRITE_READ操作。直入到驱动程序的binder_ioctl函数后，执行BINDER_WRITE_READ命令，这里就不累述了。

       最后，从binder_ioctl执行到binder_thread_write函数，我们首先看第一个命令BC_FREE_BUFFER：

[cpp] view plain copy print ?

1. int  
2. binder_thread_write(struct binder_proc *proc, struct binder_thread *thread,  
3.                     void __user *buffer, int size, signed long *consumed)  
4. {  
5.     uint32_t cmd;  
6.     void __user *ptr = buffer + *consumed;  
7.     void __user *end = buffer + size;  
8.   
9.     while (ptr < end && thread->return_error == BR_OK) {  
10.         if (get_user(cmd, (uint32_t __user *)ptr))  
11.             return -EFAULT;  
12.         ptr += sizeof(uint32_t);  
13.         if (_IOC_NR(cmd) < ARRAY_SIZE(binder_stats.bc)) {  
14.             binder_stats.bc[_IOC_NR(cmd)]++;  
15.             proc->stats.bc[_IOC_NR(cmd)]++;  
16.             thread->stats.bc[_IOC_NR(cmd)]++;  
17.         }  
18.         switch (cmd) {  
19.         ......  
20.         case BC_FREE_BUFFER: {  
21.             void __user *data_ptr;  
22.             struct binder_buffer *buffer;  
23.   
24.             if (get_user(data_ptr, (void * __user *)ptr))  
25.                 return -EFAULT;  
26.             ptr += sizeof(void *);  
27.   
28.             buffer = binder_buffer_lookup(proc, data_ptr);  
29.             if (buffer == NULL) {  
30.                 binder_user_error("binder: %d:%d "  
31.                     "BC_FREE_BUFFER u%p no match\n",  
32.                     proc->pid, thread->pid, data_ptr);  
33.                 break;  
34.             }  
35.             if (!buffer->allow_user_free) {  
36.                 binder_user_error("binder: %d:%d "  
37.                     "BC_FREE_BUFFER u%p matched "  
38.                     "unreturned buffer\n",  
39.                     proc->pid, thread->pid, data_ptr);  
40.                 break;  
41.             }  
42.             if (binder_debug_mask & BINDER_DEBUG_FREE_BUFFER)  
43.                 printk(KERN_INFO "binder: %d:%d BC_FREE_BUFFER u%p found buffer %d for %s transaction\n",  
44.                 proc->pid, thread->pid, data_ptr, buffer->debug_id,  
45.                 buffer->transaction ? "active" : "finished");  
46.   
47.             if (buffer->transaction) {  
48.                 buffer->transaction->buffer = NULL;  
49.                 buffer->transaction = NULL;  
50.             }  
51.             if (buffer->async_transaction && buffer->target_node) {  
52.                 BUG_ON(!buffer->target_node->has_async_transaction);  
53.                 if (list_empty(&buffer->target_node->async_todo))  
54.                     buffer->target_node->has_async_transaction = 0;  
55.                 else  
56.                     list_move_tail(buffer->target_node->async_todo.next, &thread->todo);  
57.             }  
58.             binder_transaction_buffer_release(proc, buffer, NULL);  
59.             binder_free_buf(proc, buffer);  
60.             break;  
61.                              }  
62.   
63.         ......  
64.         *consumed = ptr - buffer;  
65.     }  
66.     return 0;  
67. }  

int
binder_thread_write(struct binder_proc *proc, struct binder_thread *thread,
					void __user *buffer, int size, signed long *consumed)
{
	uint32_t cmd;
	void __user *ptr = buffer + *consumed;
	void __user *end = buffer + size;

	while (ptr < end && thread->return_error == BR_OK) {
		if (get_user(cmd, (uint32_t __user *)ptr))
			return -EFAULT;
		ptr += sizeof(uint32_t);
		if (_IOC_NR(cmd) < ARRAY_SIZE(binder_stats.bc)) {
			binder_stats.bc[_IOC_NR(cmd)]++;
			proc->stats.bc[_IOC_NR(cmd)]++;
			thread->stats.bc[_IOC_NR(cmd)]++;
		}
		switch (cmd) {
		......
		case BC_FREE_BUFFER: {
			void __user *data_ptr;
			struct binder_buffer *buffer;

			if (get_user(data_ptr, (void * __user *)ptr))
				return -EFAULT;
			ptr += sizeof(void *);

			buffer = binder_buffer_lookup(proc, data_ptr);
			if (buffer == NULL) {
				binder_user_error("binder: %d:%d "
					"BC_FREE_BUFFER u%p no match\n",
					proc->pid, thread->pid, data_ptr);
				break;
			}
			if (!buffer->allow_user_free) {
				binder_user_error("binder: %d:%d "
					"BC_FREE_BUFFER u%p matched "
					"unreturned buffer\n",
					proc->pid, thread->pid, data_ptr);
				break;
			}
			if (binder_debug_mask & BINDER_DEBUG_FREE_BUFFER)
				printk(KERN_INFO "binder: %d:%d BC_FREE_BUFFER u%p found buffer %d for %s transaction\n",
				proc->pid, thread->pid, data_ptr, buffer->debug_id,
				buffer->transaction ? "active" : "finished");

			if (buffer->transaction) {
				buffer->transaction->buffer = NULL;
				buffer->transaction = NULL;
			}
			if (buffer->async_transaction && buffer->target_node) {
				BUG_ON(!buffer->target_node->has_async_transaction);
				if (list_empty(&buffer->target_node->async_todo))
					buffer->target_node->has_async_transaction = 0;
				else
					list_move_tail(buffer->target_node->async_todo.next, &thread->todo);
			}
			binder_transaction_buffer_release(proc, buffer, NULL);
			binder_free_buf(proc, buffer);
			break;
							 }

		......
		*consumed = ptr - buffer;
	}
	return 0;
}

       首先通过看这个语句：

[cpp] view plain copy print ?

1. get_user(data_ptr, (void * __user *)ptr)  

get_user(data_ptr, (void * __user *)ptr)

       这个是获得要删除的Buffer的用户空间地址，接着通过下面这个语句来找到这个地址对应的struct binder_buffer信息：

[cpp] view plain copy print ?

1. buffer = binder_buffer_lookup(proc, data_ptr);  

buffer = binder_buffer_lookup(proc, data_ptr);

       因为这个空间是前面在binder_transaction里面分配的，所以这里一定能找到。

       最后，就可以释放这块内存了：

[cpp] view plain copy print ?

1. binder_transaction_buffer_release(proc, buffer, NULL);  
2. binder_free_buf(proc, buffer);  

binder_transaction_buffer_release(proc, buffer, NULL);
binder_free_buf(proc, buffer);

       再来看另外一个命令BC_REPLY：

[cpp] view plain copy print ?

1. int  
2. binder_thread_write(struct binder_proc *proc, struct binder_thread *thread,  
3.                     void __user *buffer, int size, signed long *consumed)  
4. {  
5.     uint32_t cmd;  
6.     void __user *ptr = buffer + *consumed;  
7.     void __user *end = buffer + size;  
8.   
9.     while (ptr < end && thread->return_error == BR_OK) {  
10.         if (get_user(cmd, (uint32_t __user *)ptr))  
11.             return -EFAULT;  
12.         ptr += sizeof(uint32_t);  
13.         if (_IOC_NR(cmd) < ARRAY_SIZE(binder_stats.bc)) {  
14.             binder_stats.bc[_IOC_NR(cmd)]++;  
15.             proc->stats.bc[_IOC_NR(cmd)]++;  
16.             thread->stats.bc[_IOC_NR(cmd)]++;  
17.         }  
18.         switch (cmd) {  
19.         ......  
20.         case BC_TRANSACTION:  
21.         case BC_REPLY: {  
22.             struct binder_transaction_data tr;  
23.   
24.             if (copy_from_user(&tr, ptr, sizeof(tr)))  
25.                 return -EFAULT;  
26.             ptr += sizeof(tr);  
27.             binder_transaction(proc, thread, &tr, cmd == BC_REPLY);  
28.             break;  
29.                        }  
30.   
31.         ......  
32.         *consumed = ptr - buffer;  
33.     }  
34.     return 0;  
35. }  

int
binder_thread_write(struct binder_proc *proc, struct binder_thread *thread,
					void __user *buffer, int size, signed long *consumed)
{
	uint32_t cmd;
	void __user *ptr = buffer + *consumed;
	void __user *end = buffer + size;

	while (ptr < end && thread->return_error == BR_OK) {
		if (get_user(cmd, (uint32_t __user *)ptr))
			return -EFAULT;
		ptr += sizeof(uint32_t);
		if (_IOC_NR(cmd) < ARRAY_SIZE(binder_stats.bc)) {
			binder_stats.bc[_IOC_NR(cmd)]++;
			proc->stats.bc[_IOC_NR(cmd)]++;
			thread->stats.bc[_IOC_NR(cmd)]++;
		}
		switch (cmd) {
		......
		case BC_TRANSACTION:
		case BC_REPLY: {
			struct binder_transaction_data tr;

			if (copy_from_user(&tr, ptr, sizeof(tr)))
				return -EFAULT;
			ptr += sizeof(tr);
			binder_transaction(proc, thread, &tr, cmd == BC_REPLY);
			break;
					   }

		......
		*consumed = ptr - buffer;
	}
	return 0;
}

       又再次进入到binder_transaction函数：

[cpp] view plain copy print ?

1. static void  
2. binder_transaction(struct binder_proc *proc, struct binder_thread *thread,  
3. struct binder_transaction_data *tr, int reply)  
4. {  
5.     struct binder_transaction *t;  
6.     struct binder_work *tcomplete;  
7.     size_t *offp, *off_end;  
8.     struct binder_proc *target_proc;  
9.     struct binder_thread *target_thread = NULL;  
10.     struct binder_node *target_node = NULL;  
11.     struct list_head *target_list;  
12.     wait_queue_head_t *target_wait;  
13.     struct binder_transaction *in_reply_to = NULL;  
14.     struct binder_transaction_log_entry *e;  
15.     uint32_t return_error;  
16.   
17.     ......  
18.   
19.     if (reply) {  
20.         in_reply_to = thread->transaction_stack;  
21.         if (in_reply_to == NULL) {  
22.             ......  
23.             return_error = BR_FAILED_REPLY;  
24.             goto err_empty_call_stack;  
25.         }  
26.         binder_set_nice(in_reply_to->saved_priority);  
27.         if (in_reply_to->to_thread != thread) {  
28.             .......  
29.             goto err_bad_call_stack;  
30.         }  
31.         thread->transaction_stack = in_reply_to->to_parent;  
32.         target_thread = in_reply_to->from;  
33.         if (target_thread == NULL) {  
34.             return_error = BR_DEAD_REPLY;  
35.             goto err_dead_binder;  
36.         }  
37.         if (target_thread->transaction_stack != in_reply_to) {  
38.             ......  
39.             return_error = BR_FAILED_REPLY;  
40.             in_reply_to = NULL;  
41.             target_thread = NULL;  
42.             goto err_dead_binder;  
43.         }  
44.         target_proc = target_thread->proc;  
45.     } else {  
46.         ......  
47.     }  
48.     if (target_thread) {  
49.         e->to_thread = target_thread->pid;  
50.         target_list = &target_thread->todo;  
51.         target_wait = &target_thread->wait;  
52.     } else {  
53.         ......  
54.     }  
55.   
56.   
57.     /* TODO: reuse incoming transaction for reply */  
58.     t = kzalloc(sizeof(*t), GFP_KERNEL);  
59.     if (t == NULL) {  
60.         return_error = BR_FAILED_REPLY;  
61.         goto err_alloc_t_failed;  
62.     }  
63.       
64.   
65.     tcomplete = kzalloc(sizeof(*tcomplete), GFP_KERNEL);  
66.     if (tcomplete == NULL) {  
67.         return_error = BR_FAILED_REPLY;  
68.         goto err_alloc_tcomplete_failed;  
69.     }  
70.   
71.     if (!reply && !(tr->flags & TF_ONE_WAY))  
72.         t->from = thread;  
73.     else  
74.         t->from = NULL;  
75.     t->sender_euid = proc->tsk->cred->euid;  
76.     t->to_proc = target_proc;  
77.     t->to_thread = target_thread;  
78.     t->code = tr->code;  
79.     t->flags = tr->flags;  
80.     t->priority = task_nice(current);  
81.     t->buffer = binder_alloc_buf(target_proc, tr->data_size,  
82.         tr->offsets_size, !reply && (t->flags & TF_ONE_WAY));  
83.     if (t->buffer == NULL) {  
84.         return_error = BR_FAILED_REPLY;  
85.         goto err_binder_alloc_buf_failed;  
86.     }  
87.     t->buffer->allow_user_free = 0;  
88.     t->buffer->debug_id = t->debug_id;  
89.     t->buffer->transaction = t;  
90.     t->buffer->target_node = target_node;  
91.     if (target_node)  
92.         binder_inc_node(target_node, 1, 0, NULL);  
93.   
94.     offp = (size_t *)(t->buffer->data + ALIGN(tr->data_size, sizeof(void *)));  
95.   
96.     if (copy_from_user(t->buffer->data, tr->data.ptr.buffer, tr->data_size)) {  
97.         binder_user_error("binder: %d:%d got transaction with invalid "  
98.             "data ptr\n", proc->pid, thread->pid);  
99.         return_error = BR_FAILED_REPLY;  
100.         goto err_copy_data_failed;  
101.     }  
102.     if (copy_from_user(offp, tr->data.ptr.offsets, tr->offsets_size)) {  
103.         binder_user_error("binder: %d:%d got transaction with invalid "  
104.             "offsets ptr\n", proc->pid, thread->pid);  
105.         return_error = BR_FAILED_REPLY;  
106.         goto err_copy_data_failed;  
107.     }  
108.       
109.     ......  
110.   
111.     if (reply) {  
112.         BUG_ON(t->buffer->async_transaction != 0);  
113.         binder_pop_transaction(target_thread, in_reply_to);  
114.     } else if (!(t->flags & TF_ONE_WAY)) {  
115.         ......  
116.     } else {  
117.         ......  
118.     }  
119.     t->work.type = BINDER_WORK_TRANSACTION;  
120.     list_add_tail(&t->work.entry, target_list);  
121.     tcomplete->type = BINDER_WORK_TRANSACTION_COMPLETE;  
122.     list_add_tail(&tcomplete->entry, &thread->todo);  
123.     if (target_wait)  
124.         wake_up_interruptible(target_wait);  
125.     return;  
126.     ......  
127. }  

static void
binder_transaction(struct binder_proc *proc, struct binder_thread *thread,
struct binder_transaction_data *tr, int reply)
{
	struct binder_transaction *t;
	struct binder_work *tcomplete;
	size_t *offp, *off_end;
	struct binder_proc *target_proc;
	struct binder_thread *target_thread = NULL;
	struct binder_node *target_node = NULL;
	struct list_head *target_list;
	wait_queue_head_t *target_wait;
	struct binder_transaction *in_reply_to = NULL;
	struct binder_transaction_log_entry *e;
	uint32_t return_error;

	......

	if (reply) {
		in_reply_to = thread->transaction_stack;
		if (in_reply_to == NULL) {
			......
			return_error = BR_FAILED_REPLY;
			goto err_empty_call_stack;
		}
		binder_set_nice(in_reply_to->saved_priority);
		if (in_reply_to->to_thread != thread) {
			.......
			goto err_bad_call_stack;
		}
		thread->transaction_stack = in_reply_to->to_parent;
		target_thread = in_reply_to->from;
		if (target_thread == NULL) {
			return_error = BR_DEAD_REPLY;
			goto err_dead_binder;
		}
		if (target_thread->transaction_stack != in_reply_to) {
			......
			return_error = BR_FAILED_REPLY;
			in_reply_to = NULL;
			target_thread = NULL;
			goto err_dead_binder;
		}
		target_proc = target_thread->proc;
	} else {
		......
	}
	if (target_thread) {
		e->to_thread = target_thread->pid;
		target_list = &target_thread->todo;
		target_wait = &target_thread->wait;
	} else {
		......
	}


	/* TODO: reuse incoming transaction for reply */
	t = kzalloc(sizeof(*t), GFP_KERNEL);
	if (t == NULL) {
		return_error = BR_FAILED_REPLY;
		goto err_alloc_t_failed;
	}
	

	tcomplete = kzalloc(sizeof(*tcomplete), GFP_KERNEL);
	if (tcomplete == NULL) {
		return_error = BR_FAILED_REPLY;
		goto err_alloc_tcomplete_failed;
	}

	if (!reply && !(tr->flags & TF_ONE_WAY))
		t->from = thread;
	else
		t->from = NULL;
	t->sender_euid = proc->tsk->cred->euid;
	t->to_proc = target_proc;
	t->to_thread = target_thread;
	t->code = tr->code;
	t->flags = tr->flags;
	t->priority = task_nice(current);
	t->buffer = binder_alloc_buf(target_proc, tr->data_size,
		tr->offsets_size, !reply && (t->flags & TF_ONE_WAY));
	if (t->buffer == NULL) {
		return_error = BR_FAILED_REPLY;
		goto err_binder_alloc_buf_failed;
	}
	t->buffer->allow_user_free = 0;
	t->buffer->debug_id = t->debug_id;
	t->buffer->transaction = t;
	t->buffer->target_node = target_node;
	if (target_node)
		binder_inc_node(target_node, 1, 0, NULL);

	offp = (size_t *)(t->buffer->data + ALIGN(tr->data_size, sizeof(void *)));

	if (copy_from_user(t->buffer->data, tr->data.ptr.buffer, tr->data_size)) {
		binder_user_error("binder: %d:%d got transaction with invalid "
			"data ptr\n", proc->pid, thread->pid);
		return_error = BR_FAILED_REPLY;
		goto err_copy_data_failed;
	}
	if (copy_from_user(offp, tr->data.ptr.offsets, tr->offsets_size)) {
		binder_user_error("binder: %d:%d got transaction with invalid "
			"offsets ptr\n", proc->pid, thread->pid);
		return_error = BR_FAILED_REPLY;
		goto err_copy_data_failed;
	}
	
    ......

	if (reply) {
		BUG_ON(t->buffer->async_transaction != 0);
		binder_pop_transaction(target_thread, in_reply_to);
	} else if (!(t->flags & TF_ONE_WAY)) {
		......
	} else {
		......
	}
	t->work.type = BINDER_WORK_TRANSACTION;
	list_add_tail(&t->work.entry, target_list);
	tcomplete->type = BINDER_WORK_TRANSACTION_COMPLETE;
	list_add_tail(&tcomplete->entry, &thread->todo);
	if (target_wait)
		wake_up_interruptible(target_wait);
	return;
    ......
}

       注意，这里的reply为1，我们忽略掉其它无关代码。

       前面Service Manager正在binder_thread_read函数中被MediaPlayerService启动后进程唤醒后，在最后会把当前处理完的事务放在thread->transaction_stack中：

[cpp] view plain copy print ?

1. if (cmd == BR_TRANSACTION && !(t->flags & TF_ONE_WAY)) {  
2.     t->to_parent = thread->transaction_stack;  
3.     t->to_thread = thread;  
4.     thread->transaction_stack = t;  
5. }   

if (cmd == BR_TRANSACTION && !(t->flags & TF_ONE_WAY)) {
	t->to_parent = thread->transaction_stack;
	t->to_thread = thread;
	thread->transaction_stack = t;
} 

       所以，这里，首先是把它这个binder_transaction取回来，并且放在本地变量in_reply_to中：

[cpp] view plain copy print ?

1. in_reply_to = thread->transaction_stack;  

in_reply_to = thread->transaction_stack;

       接着就可以通过in_reply_to得到最终发出这个事务请求的线程和进程：

[cpp] view plain copy print ?

1. target_thread = in_reply_to->from;  
2. target_proc = target_thread->proc;  

target_thread = in_reply_to->from;
target_proc = target_thread->proc;

        然后得到target_list和target_wait：

[cpp] view plain copy print ?

1. target_list = &target_thread->todo;  
2. target_wait = &target_thread->wait;  

target_list = &target_thread->todo;
target_wait = &target_thread->wait;

       下面这一段代码：

[cpp] view plain copy print ?

1. /* TODO: reuse incoming transaction for reply */  
2. t = kzalloc(sizeof(*t), GFP_KERNEL);  
3. if (t == NULL) {  
4.     return_error = BR_FAILED_REPLY;  
5.     goto err_alloc_t_failed;  
6. }  
7.   
8.   
9. tcomplete = kzalloc(sizeof(*tcomplete), GFP_KERNEL);  
10. if (tcomplete == NULL) {  
11.     return_error = BR_FAILED_REPLY;  
12.     goto err_alloc_tcomplete_failed;  
13. }  
14.   
15. if (!reply && !(tr->flags & TF_ONE_WAY))  
16.     t->from = thread;  
17. else  
18.     t->from = NULL;  
19. t->sender_euid = proc->tsk->cred->euid;  
20. t->to_proc = target_proc;  
21. t->to_thread = target_thread;  
22. t->code = tr->code;  
23. t->flags = tr->flags;  
24. t->priority = task_nice(current);  
25. t->buffer = binder_alloc_buf(target_proc, tr->data_size,  
26.     tr->offsets_size, !reply && (t->flags & TF_ONE_WAY));  
27. if (t->buffer == NULL) {  
28.     return_error = BR_FAILED_REPLY;  
29.     goto err_binder_alloc_buf_failed;  
30. }  
31. t->buffer->allow_user_free = 0;  
32. t->buffer->debug_id = t->debug_id;  
33. t->buffer->transaction = t;  
34. t->buffer->target_node = target_node;  
35. if (target_node)  
36.     binder_inc_node(target_node, 1, 0, NULL);  
37.   
38. offp = (size_t *)(t->buffer->data + ALIGN(tr->data_size, sizeof(void *)));  
39.   
40. if (copy_from_user(t->buffer->data, tr->data.ptr.buffer, tr->data_size)) {  
41.     binder_user_error("binder: %d:%d got transaction with invalid "  
42.         "data ptr\n", proc->pid, thread->pid);  
43.     return_error = BR_FAILED_REPLY;  
44.     goto err_copy_data_failed;  
45. }  
46. if (copy_from_user(offp, tr->data.ptr.offsets, tr->offsets_size)) {  
47.     binder_user_error("binder: %d:%d got transaction with invalid "  
48.         "offsets ptr\n", proc->pid, thread->pid);  
49.     return_error = BR_FAILED_REPLY;  
50.     goto err_copy_data_failed;  
51. }  

	/* TODO: reuse incoming transaction for reply */
	t = kzalloc(sizeof(*t), GFP_KERNEL);
	if (t == NULL) {
		return_error = BR_FAILED_REPLY;
		goto err_alloc_t_failed;
	}
	

	tcomplete = kzalloc(sizeof(*tcomplete), GFP_KERNEL);
	if (tcomplete == NULL) {
		return_error = BR_FAILED_REPLY;
		goto err_alloc_tcomplete_failed;
	}

	if (!reply && !(tr->flags & TF_ONE_WAY))
		t->from = thread;
	else
		t->from = NULL;
	t->sender_euid = proc->tsk->cred->euid;
	t->to_proc = target_proc;
	t->to_thread = target_thread;
	t->code = tr->code;
	t->flags = tr->flags;
	t->priority = task_nice(current);
	t->buffer = binder_alloc_buf(target_proc, tr->data_size,
		tr->offsets_size, !reply && (t->flags & TF_ONE_WAY));
	if (t->buffer == NULL) {
		return_error = BR_FAILED_REPLY;
		goto err_binder_alloc_buf_failed;
	}
	t->buffer->allow_user_free = 0;
	t->buffer->debug_id = t->debug_id;
	t->buffer->transaction = t;
	t->buffer->target_node = target_node;
	if (target_node)
		binder_inc_node(target_node, 1, 0, NULL);

	offp = (size_t *)(t->buffer->data + ALIGN(tr->data_size, sizeof(void *)));

	if (copy_from_user(t->buffer->data, tr->data.ptr.buffer, tr->data_size)) {
		binder_user_error("binder: %d:%d got transaction with invalid "
			"data ptr\n", proc->pid, thread->pid);
		return_error = BR_FAILED_REPLY;
		goto err_copy_data_failed;
	}
	if (copy_from_user(offp, tr->data.ptr.offsets, tr->offsets_size)) {
		binder_user_error("binder: %d:%d got transaction with invalid "
			"offsets ptr\n", proc->pid, thread->pid);
		return_error = BR_FAILED_REPLY;
		goto err_copy_data_failed;
	}

          我们在前面已经分析过了，这里不再重复。但是有一点要注意的是，这里target_node为NULL，因此，t->buffer->target_node也为NULL。

          函数本来有一个for循环，用来处理数据中的Binder对象，这里由于没有Binder对象，所以就略过了。到了下面这句代码：

[cpp] view plain copy print ?

1. binder_pop_transaction(target_thread, in_reply_to);  

binder_pop_transaction(target_thread, in_reply_to);

          我们看看做了什么事情：

[cpp] view plain copy print ?

1. static void  
2. binder_pop_transaction(  
3.     struct binder_thread *target_thread, struct binder_transaction *t)  
4. {  
5.     if (target_thread) {  
6.         BUG_ON(target_thread->transaction_stack != t);  
7.         BUG_ON(target_thread->transaction_stack->from != target_thread);  
8.         target_thread->transaction_stack =  
9.             target_thread->transaction_stack->from_parent;  
10.         t->from = NULL;  
11.     }  
12.     t->need_reply = 0;  
13.     if (t->buffer)  
14.         t->buffer->transaction = NULL;  
15.     kfree(t);  
16.     binder_stats.obj_deleted[BINDER_STAT_TRANSACTION]++;  
17. }  

static void
binder_pop_transaction(
	struct binder_thread *target_thread, struct binder_transaction *t)
{
	if (target_thread) {
		BUG_ON(target_thread->transaction_stack != t);
		BUG_ON(target_thread->transaction_stack->from != target_thread);
		target_thread->transaction_stack =
			target_thread->transaction_stack->from_parent;
		t->from = NULL;
	}
	t->need_reply = 0;
	if (t->buffer)
		t->buffer->transaction = NULL;
	kfree(t);
	binder_stats.obj_deleted[BINDER_STAT_TRANSACTION]++;
}

        由于到了这里，已经不需要in_reply_to这个transaction了，就把它删掉。

        回到binder_transaction函数：

[cpp] view plain copy print ?

1. t->work.type = BINDER_WORK_TRANSACTION;  
2. list_add_tail(&t->work.entry, target_list);  
3. tcomplete->type = BINDER_WORK_TRANSACTION_COMPLETE;  
4. list_add_tail(&tcomplete->entry, &thread->todo);  

t->work.type = BINDER_WORK_TRANSACTION;
list_add_tail(&t->work.entry, target_list);
tcomplete->type = BINDER_WORK_TRANSACTION_COMPLETE;
list_add_tail(&tcomplete->entry, &thread->todo);

         和前面一样，分别把t和tcomplete分别放在target_list和thread->todo队列中，这里的target_list指的就是最初调用IServiceManager::addService的MediaPlayerService的Server主线程的的thread->todo队列了，而thread->todo指的是Service Manager中用来回复IServiceManager::addService请求的线程。

        最后，唤醒等待在target_wait队列上的线程了，就是最初调用IServiceManager::addService的MediaPlayerService的Server主线程了，它最后在binder_thread_read函数中睡眠在thread->wait上，就是这里的target_wait了：

[cpp] view plain copy print ?

1. if (target_wait)  
2.     wake_up_interruptible(target_wait);  

if (target_wait)
	wake_up_interruptible(target_wait);

        这样，Service Manger回复调用IServiceManager::addService请求就算完成了，重新回到frameworks/base/cmds/servicemanager/binder.c文件中的binder_loop函数等待下一个Client请求的到来。事实上，Service Manger回到binder_loop函数再次执行ioctl函数时候，又会再次进入到binder_thread_read函数。这时个会发现thread->todo不为空，这是因为刚才我们调用了：

[cpp] view plain copy print ?

1. list_add_tail(&tcomplete->entry, &thread->todo);  

list_add_tail(&tcomplete->entry, &thread->todo);

          把一个工作项tcompelete放在了在thread->todo中，这个tcompelete的type为BINDER_WORK_TRANSACTION_COMPLETE，因此，Binder驱动程序会执行下面操作：

[cpp] view plain copy print ?

1. switch (w->type) {  
2. case BINDER_WORK_TRANSACTION_COMPLETE: {  
3.     cmd = BR_TRANSACTION_COMPLETE;  
4.     if (put_user(cmd, (uint32_t __user *)ptr))  
5.         return -EFAULT;  
6.     ptr += sizeof(uint32_t);  
7.   
8.     list_del(&w->entry);  
9.     kfree(w);  
10.       
11.     } break;  
12.     ......  
13. }  

switch (w->type) {
case BINDER_WORK_TRANSACTION_COMPLETE: {
	cmd = BR_TRANSACTION_COMPLETE;
	if (put_user(cmd, (uint32_t __user *)ptr))
		return -EFAULT;
	ptr += sizeof(uint32_t);

	list_del(&w->entry);
	kfree(w);
	
	} break;
	......
}

        binder_loop函数执行完这个ioctl调用后，才会在下一次调用ioctl进入到Binder驱动程序进入休眠状态，等待下一次Client的请求。

        上面讲到调用IServiceManager::addService的MediaPlayerService的Server主线程被唤醒了，于是，重新执行binder_thread_read函数：

[cpp] view plain copy print ?

1. static int  
2. binder_thread_read(struct binder_proc *proc, struct binder_thread *thread,  
3.                    void  __user *buffer, int size, signed long *consumed, int non_block)  
4. {  
5.     void __user *ptr = buffer + *consumed;  
6.     void __user *end = buffer + size;  
7.   
8.     int ret = 0;  
9.     int wait_for_proc_work;  
10.   
11.     if (*consumed == 0) {  
12.         if (put_user(BR_NOOP, (uint32_t __user *)ptr))  
13.             return -EFAULT;  
14.         ptr += sizeof(uint32_t);  
15.     }  
16.   
17. retry:  
18.     wait_for_proc_work = thread->transaction_stack == NULL && list_empty(&thread->todo);  
19.   
20.     ......  
21.   
22.     if (wait_for_proc_work) {  
23.         ......  
24.     } else {  
25.         if (non_block) {  
26.             if (!binder_has_thread_work(thread))  
27.                 ret = -EAGAIN;  
28.         } else  
29.             ret = wait_event_interruptible(thread->wait, binder_has_thread_work(thread));  
30.     }  
31.       
32.     ......  
33.   
34.     while (1) {  
35.         uint32_t cmd;  
36.         struct binder_transaction_data tr;  
37.         struct binder_work *w;  
38.         struct binder_transaction *t = NULL;  
39.   
40.         if (!list_empty(&thread->todo))  
41.             w = list_first_entry(&thread->todo, struct binder_work, entry);  
42.         else if (!list_empty(&proc->todo) && wait_for_proc_work)  
43.             w = list_first_entry(&proc->todo, struct binder_work, entry);  
44.         else {  
45.             if (ptr - buffer == 4 && !(thread->looper & BINDER_LOOPER_STATE_NEED_RETURN)) /* no data added */  
46.                 goto retry;  
47.             break;  
48.         }  
49.   
50.         ......  
51.   
52.         switch (w->type) {  
53.         case BINDER_WORK_TRANSACTION: {  
54.             t = container_of(w, struct binder_transaction, work);  
55.                                       } break;  
56.         ......  
57.         }  
58.   
59.         if (!t)  
60.             continue;  
61.   
62.         BUG_ON(t->buffer == NULL);  
63.         if (t->buffer->target_node) {  
64.             ......  
65.         } else {  
66.             tr.target.ptr = NULL;  
67.             tr.cookie = NULL;  
68.             cmd = BR_REPLY;  
69.         }  
70.         tr.code = t->code;  
71.         tr.flags = t->flags;  
72.         tr.sender_euid = t->sender_euid;  
73.   
74.         if (t->from) {  
75.             ......  
76.         } else {  
77.             tr.sender_pid = 0;  
78.         }  
79.   
80.         tr.data_size = t->buffer->data_size;  
81.         tr.offsets_size = t->buffer->offsets_size;  
82.         tr.data.ptr.buffer = (void *)t->buffer->data + proc->user_buffer_offset;  
83.         tr.data.ptr.offsets = tr.data.ptr.buffer + ALIGN(t->buffer->data_size, sizeof(void *));  
84.   
85.         if (put_user(cmd, (uint32_t __user *)ptr))  
86.             return -EFAULT;  
87.         ptr += sizeof(uint32_t);  
88.         if (copy_to_user(ptr, &tr, sizeof(tr)))  
89.             return -EFAULT;  
90.         ptr += sizeof(tr);  
91.   
92.         ......  
93.   
94.         list_del(&t->work.entry);  
95.         t->buffer->allow_user_free = 1;  
96.         if (cmd == BR_TRANSACTION && !(t->flags & TF_ONE_WAY)) {  
97.             ......  
98.         } else {  
99.             t->buffer->transaction = NULL;  
100.             kfree(t);  
101.             binder_stats.obj_deleted[BINDER_STAT_TRANSACTION]++;  
102.         }  
103.         break;  
104.     }  
105.   
106. done:  
107.     ......  
108.     return 0;  
109. }  

static int
binder_thread_read(struct binder_proc *proc, struct binder_thread *thread,
				   void  __user *buffer, int size, signed long *consumed, int non_block)
{
	void __user *ptr = buffer + *consumed;
	void __user *end = buffer + size;

	int ret = 0;
	int wait_for_proc_work;

	if (*consumed == 0) {
		if (put_user(BR_NOOP, (uint32_t __user *)ptr))
			return -EFAULT;
		ptr += sizeof(uint32_t);
	}

retry:
	wait_for_proc_work = thread->transaction_stack == NULL && list_empty(&thread->todo);

	......

	if (wait_for_proc_work) {
		......
	} else {
		if (non_block) {
			if (!binder_has_thread_work(thread))
				ret = -EAGAIN;
		} else
			ret = wait_event_interruptible(thread->wait, binder_has_thread_work(thread));
	}
	
	......

	while (1) {
		uint32_t cmd;
		struct binder_transaction_data tr;
		struct binder_work *w;
		struct binder_transaction *t = NULL;

		if (!list_empty(&thread->todo))
			w = list_first_entry(&thread->todo, struct binder_work, entry);
		else if (!list_empty(&proc->todo) && wait_for_proc_work)
			w = list_first_entry(&proc->todo, struct binder_work, entry);
		else {
			if (ptr - buffer == 4 && !(thread->looper & BINDER_LOOPER_STATE_NEED_RETURN)) /* no data added */
				goto retry;
			break;
		}

		......

		switch (w->type) {
		case BINDER_WORK_TRANSACTION: {
			t = container_of(w, struct binder_transaction, work);
									  } break;
		......
		}

		if (!t)
			continue;

		BUG_ON(t->buffer == NULL);
		if (t->buffer->target_node) {
			......
		} else {
			tr.target.ptr = NULL;
			tr.cookie = NULL;
			cmd = BR_REPLY;
		}
		tr.code = t->code;
		tr.flags = t->flags;
		tr.sender_euid = t->sender_euid;

		if (t->from) {
			......
		} else {
			tr.sender_pid = 0;
		}

		tr.data_size = t->buffer->data_size;
		tr.offsets_size = t->buffer->offsets_size;
		tr.data.ptr.buffer = (void *)t->buffer->data + proc->user_buffer_offset;
		tr.data.ptr.offsets = tr.data.ptr.buffer + ALIGN(t->buffer->data_size, sizeof(void *));

		if (put_user(cmd, (uint32_t __user *)ptr))
			return -EFAULT;
		ptr += sizeof(uint32_t);
		if (copy_to_user(ptr, &tr, sizeof(tr)))
			return -EFAULT;
		ptr += sizeof(tr);

		......

		list_del(&t->work.entry);
		t->buffer->allow_user_free = 1;
		if (cmd == BR_TRANSACTION && !(t->flags & TF_ONE_WAY)) {
			......
		} else {
			t->buffer->transaction = NULL;
			kfree(t);
			binder_stats.obj_deleted[BINDER_STAT_TRANSACTION]++;
		}
		break;
	}

done:
	......
	return 0;
}

         在while循环中，从thread->todo得到w，w->type为BINDER_WORK_TRANSACTION，于是，得到t。从上面可以知道，Service Manager反回了一个0回来，写在t->buffer->data里面，现在把t->buffer->data加上proc->user_buffer_offset，得到用户空间地址，保存在tr.data.ptr.buffer里面，这样用户空间就可以访问这个返回码了。由于cmd不等于BR_TRANSACTION，这时就可以把t删除掉了，因为以后都不需要用了。

         执行完这个函数后，就返回到binder_ioctl函数，执行下面语句，把数据返回给用户空间：

[cpp] view plain copy print ?

1. if (copy_to_user(ubuf, &bwr, sizeof(bwr))) {  
2.     ret = -EFAULT;  
3.     goto err;  
4. }  

if (copy_to_user(ubuf, &bwr, sizeof(bwr))) {
    ret = -EFAULT;
    goto err;
}

         接着返回到用户空间IPCThreadState::talkWithDriver函数，最后返回到IPCThreadState::waitForResponse函数，最终执行到下面语句：

[cpp] view plain copy print ?

1. status_t IPCThreadState::waitForResponse(Parcel *reply, status_t *acquireResult)  
2. {  
3.     int32_t cmd;  
4.     int32_t err;  
5.   
6.     while (1) {  
7.         if ((err=talkWithDriver()) < NO_ERROR) break;  
8.           
9.         ......  
10.   
11.         cmd = mIn.readInt32();  
12.   
13.         ......  
14.   
15.         switch (cmd) {  
16.         ......  
17.         case BR_REPLY:  
18.             {  
19.                 binder_transaction_data tr;  
20.                 err = mIn.read(&tr, sizeof(tr));  
21.                 LOG_ASSERT(err == NO_ERROR, "Not enough command data for brREPLY");  
22.                 if (err != NO_ERROR) goto finish;  
23.   
24.                 if (reply) {  
25.                     if ((tr.flags & TF_STATUS_CODE) == 0) {  
26.                         reply->ipcSetDataReference(  
27.                             reinterpret_cast<const uint8_t*>(tr.data.ptr.buffer),  
28.                             tr.data_size,  
29.                             reinterpret_cast<const size_t*>(tr.data.ptr.offsets),  
30.                             tr.offsets_size/sizeof(size_t),  
31.                             freeBuffer, this);  
32.                     } else {  
33.                         ......  
34.                     }  
35.                 } else {  
36.                     ......  
37.                 }  
38.             }  
39.             goto finish;  
40.   
41.         ......  
42.         }  
43.     }  
44.   
45. finish:  
46.     ......  
47.     return err;  
48. }  

status_t IPCThreadState::waitForResponse(Parcel *reply, status_t *acquireResult)
{
	int32_t cmd;
	int32_t err;

	while (1) {
		if ((err=talkWithDriver()) < NO_ERROR) break;
		
		......

		cmd = mIn.readInt32();

		......

		switch (cmd) {
		......
		case BR_REPLY:
			{
				binder_transaction_data tr;
				err = mIn.read(&tr, sizeof(tr));
				LOG_ASSERT(err == NO_ERROR, "Not enough command data for brREPLY");
				if (err != NO_ERROR) goto finish;

				if (reply) {
					if ((tr.flags & TF_STATUS_CODE) == 0) {
						reply->ipcSetDataReference(
							reinterpret_cast<const uint8_t*>(tr.data.ptr.buffer),
							tr.data_size,
							reinterpret_cast<const size_t*>(tr.data.ptr.offsets),
							tr.offsets_size/sizeof(size_t),
							freeBuffer, this);
					} else {
						......
					}
				} else {
					......
				}
			}
			goto finish;

		......
		}
	}

finish:
	......
	return err;
}

        注意，这里的tr.flags等于0，这个是在上面的binder_send_reply函数里设置的。最终把结果保存在reply了：

[cpp] view plain copy print ?

1. reply->ipcSetDataReference(  
2.        reinterpret_cast<const uint8_t*>(tr.data.ptr.buffer),  
3.        tr.data_size,  
4.        reinterpret_cast<const size_t*>(tr.data.ptr.offsets),  
5.        tr.offsets_size/sizeof(size_t),  
6.        freeBuffer, this);  

reply->ipcSetDataReference(
       reinterpret_cast<const uint8_t*>(tr.data.ptr.buffer),
       tr.data_size,
       reinterpret_cast<const size_t*>(tr.data.ptr.offsets),
       tr.offsets_size/sizeof(size_t),
       freeBuffer, this);

       这个函数我们就不看了，有兴趣的读者可以研究一下。

       从这里层层返回，最后回到MediaPlayerService::instantiate函数中。

       至此，IServiceManager::addService终于执行完毕了。这个过程非常复杂，但是如果我们能够深刻地理解这一过程，将能很好地理解Binder机制的设计思想和实现过程。这里，对IServiceManager::addService过程中MediaPlayerService、ServiceManager和BinderDriver之间的交互作一个小结：

        回到frameworks/base/media/mediaserver/main_mediaserver.cpp文件中的main函数，接下去还要执行下面两个函数：

[cpp] view plain copy print ?

1. ProcessState::self()->startThreadPool();  
2. IPCThreadState::self()->joinThreadPool();  

    ProcessState::self()->startThreadPool();
    IPCThreadState::self()->joinThreadPool();

        首先看ProcessState::startThreadPool函数的实现：

[cpp] view plain copy print ?

1. void ProcessState::startThreadPool()  
2. {  
3.     AutoMutex _l(mLock);  
4.     if (!mThreadPoolStarted) {  
5.         mThreadPoolStarted = true;  
6.         spawnPooledThread(true);  
7.     }  
8. }  

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

       这里调用spwanPooledThread：

[cpp] view plain copy print ?

1. void ProcessState::spawnPooledThread(bool isMain)  
2. {  
3.     if (mThreadPoolStarted) {  
4.         int32_t s = android_atomic_add(1, &mThreadPoolSeq);  
5.         char buf[32];  
6.         sprintf(buf, "Binder Thread #%d", s);  
7.         LOGV("Spawning new pooled thread, name=%s\n", buf);  
8.         sp<Thread> t = new PoolThread(isMain);  
9.         t->run(buf);  
10.     }  
11. }  

void ProcessState::spawnPooledThread(bool isMain)
{
    if (mThreadPoolStarted) {
        int32_t s = android_atomic_add(1, &mThreadPoolSeq);
        char buf[32];
        sprintf(buf, "Binder Thread #%d", s);
        LOGV("Spawning new pooled thread, name=%s\n", buf);
        sp<Thread> t = new PoolThread(isMain);
        t->run(buf);
    }
}

       这里主要是创建一个线程，PoolThread继续Thread类，Thread类定义在frameworks/base/libs/utils/Threads.cpp文件中，其run函数最终调用子类的threadLoop函数，这里即为PoolThread::threadLoop函数：

[cpp] view plain copy print ?

1. virtual bool threadLoop()  
2. {  
3.     IPCThreadState::self()->joinThreadPool(mIsMain);  
4.     return false;  
5. }  

    virtual bool threadLoop()
    {
        IPCThreadState::self()->joinThreadPool(mIsMain);
        return false;
    }

       这里和frameworks/base/media/mediaserver/main_mediaserver.cpp文件中的main函数一样，最终都是调用了IPCThreadState::joinThreadPool函数，它们的区别是，一个参数是true，一个是默认值false。我们来看一下这个函数的实现：

[cpp] view plain copy print ?

1. void IPCThreadState::joinThreadPool(bool isMain)  
2. {  
3.     LOG_THREADPOOL("**** THREAD %p (PID %d) IS JOINING THE THREAD POOL\n", (void*)pthread_self(), getpid());  
4.   
5.     mOut.writeInt32(isMain ? BC_ENTER_LOOPER : BC_REGISTER_LOOPER);  
6.   
7.     ......  
8.   
9.     status_t result;  
10.     do {  
11.         int32_t cmd;  
12.   
13.         .......  
14.   
15.         // now get the next command to be processed, waiting if necessary   
16.         result = talkWithDriver();  
17.         if (result >= NO_ERROR) {  
18.             size_t IN = mIn.dataAvail();  
19.             if (IN < sizeof(int32_t)) continue;  
20.             cmd = mIn.readInt32();  
21.             ......  
22.             }  
23.   
24.             result = executeCommand(cmd);  
25.         }  
26.   
27.         ......  
28.     } while (result != -ECONNREFUSED && result != -EBADF);  
29.   
30.     .......  
31.   
32.     mOut.writeInt32(BC_EXIT_LOOPER);  
33.     talkWithDriver(false);  
34. }  

void IPCThreadState::joinThreadPool(bool isMain)
{
	LOG_THREADPOOL("**** THREAD %p (PID %d) IS JOINING THE THREAD POOL\n", (void*)pthread_self(), getpid());

	mOut.writeInt32(isMain ? BC_ENTER_LOOPER : BC_REGISTER_LOOPER);

	......

	status_t result;
	do {
		int32_t cmd;

		.......

		// now get the next command to be processed, waiting if necessary
		result = talkWithDriver();
		if (result >= NO_ERROR) {
			size_t IN = mIn.dataAvail();
			if (IN < sizeof(int32_t)) continue;
			cmd = mIn.readInt32();
			......
			}

			result = executeCommand(cmd);
		}

		......
	} while (result != -ECONNREFUSED && result != -EBADF);

	.......

	mOut.writeInt32(BC_EXIT_LOOPER);
	talkWithDriver(false);
}

        这个函数最终是在一个无穷循环中，通过调用talkWithDriver函数来和Binder驱动程序进行交互，实际上就是调用talkWithDriver来等待Client的请求，然后再调用executeCommand来处理请求，而在executeCommand函数中，最终会调用BBinder::transact来真正处理Client的请求：

[cpp] view plain copy print ?

1. status_t IPCThreadState::executeCommand(int32_t cmd)  
2. {  
3.     BBinder* obj;  
4.     RefBase::weakref_type* refs;  
5.     status_t result = NO_ERROR;  
6.   
7.     switch (cmd) {  
8.     ......  
9.   
10.     case BR_TRANSACTION:  
11.         {  
12.             binder_transaction_data tr;  
13.             result = mIn.read(&tr, sizeof(tr));  
14.               
15.             ......  
16.   
17.             Parcel reply;  
18.               
19.             ......  
20.   
21.             if (tr.target.ptr) {  
22.                 sp<BBinder> b((BBinder*)tr.cookie);  
23.                 const status_t error = b->transact(tr.code, buffer, &reply, tr.flags);  
24.                 if (error < NO_ERROR) reply.setError(error);  
25.   
26.             } else {  
27.                 const status_t error = the_context_object->transact(tr.code, buffer, &reply, tr.flags);  
28.                 if (error < NO_ERROR) reply.setError(error);  
29.             }  
30.   
31.             ......  
32.         }  
33.         break;  
34.   
35.     .......  
36.     }  
37.   
38.     if (result != NO_ERROR) {  
39.         mLastError = result;  
40.     }  
41.   
42.     return result;  
43. }  

status_t IPCThreadState::executeCommand(int32_t cmd)
{
	BBinder* obj;
	RefBase::weakref_type* refs;
	status_t result = NO_ERROR;

	switch (cmd) {
	......

	case BR_TRANSACTION:
		{
			binder_transaction_data tr;
			result = mIn.read(&tr, sizeof(tr));
			
			......

			Parcel reply;
			
			......

			if (tr.target.ptr) {
				sp<BBinder> b((BBinder*)tr.cookie);
				const status_t error = b->transact(tr.code, buffer, &reply, tr.flags);
				if (error < NO_ERROR) reply.setError(error);

			} else {
				const status_t error = the_context_object->transact(tr.code, buffer, &reply, tr.flags);
				if (error < NO_ERROR) reply.setError(error);
			}

			......
		}
		break;

	.......
	}

	if (result != NO_ERROR) {
		mLastError = result;
	}

	return result;
}

        接下来再看一下BBinder::transact的实现：

[cpp] view plain copy print ?

1. status_t BBinder::transact(  
2.     uint32_t code, const Parcel& data, Parcel* reply, uint32_t flags)  
3. {  
4.     data.setDataPosition(0);  
5.   
6.     status_t err = NO_ERROR;  
7.     switch (code) {  
8.         case PING_TRANSACTION:  
9.             reply->writeInt32(pingBinder());  
10.             break;  
11.         default:  
12.             err = onTransact(code, data, reply, flags);  
13.             break;  
14.     }  
15.   
16.     if (reply != NULL) {  
17.         reply->setDataPosition(0);  
18.     }  
19.   
20.     return err;  
21. }  

status_t BBinder::transact(
    uint32_t code, const Parcel& data, Parcel* reply, uint32_t flags)
{
    data.setDataPosition(0);

    status_t err = NO_ERROR;
    switch (code) {
        case PING_TRANSACTION:
            reply->writeInt32(pingBinder());
            break;
        default:
            err = onTransact(code, data, reply, flags);
            break;
    }

    if (reply != NULL) {
        reply->setDataPosition(0);
    }

    return err;
}

       最终会调用onTransact函数来处理。在这个场景中，BnMediaPlayerService继承了BBinder类，并且重载了onTransact函数，因此，这里实际上是调用了BnMediaPlayerService::onTransact函数，这个函数定义在frameworks/base/libs/media/libmedia/IMediaPlayerService.cpp文件中：

[cpp] view plain copy print ?

1. status_t BnMediaPlayerService::onTransact(  
2.     uint32_t code, const Parcel& data, Parcel* reply, uint32_t flags)  
3. {  
4.     switch(code) {  
5.         case CREATE_URL: {  
6.             ......  
7.                          } break;  
8.         case CREATE_FD: {  
9.             ......  
10.                         } break;  
11.         case DECODE_URL: {  
12.             ......  
13.                          } break;  
14.         case DECODE_FD: {  
15.             ......  
16.                         } break;  
17.         case CREATE_MEDIA_RECORDER: {  
18.             ......  
19.                                     } break;  
20.         case CREATE_METADATA_RETRIEVER: {  
21.             ......  
22.                                         } break;  
23.         case GET_OMX: {  
24.             ......  
25.                       } break;  
26.         default:  
27.             return BBinder::onTransact(code, data, reply, flags);  
28.     }  
29. }  

status_t BnMediaPlayerService::onTransact(
	uint32_t code, const Parcel& data, Parcel* reply, uint32_t flags)
{
	switch(code) {
		case CREATE_URL: {
			......
						 } break;
		case CREATE_FD: {
			......
						} break;
		case DECODE_URL: {
			......
						 } break;
		case DECODE_FD: {
			......
						} break;
		case CREATE_MEDIA_RECORDER: {
			......
									} break;
		case CREATE_METADATA_RETRIEVER: {
			......
										} break;
		case GET_OMX: {
			......
					  } break;
		default:
			return BBinder::onTransact(code, data, reply, flags);
	}
}

       至此，我们就以MediaPlayerService为例，完整地介绍了Android系统进程间通信Binder机制中的Server启动过程。Server启动起来之后，就会在一个无穷循环中等待Client的请求了。在下一篇文章中，我们将介绍Client如何通过Service Manager远程接口来获得Server远程接口，进而调用Server远程接口来使用Server提供的服务，敬请关注。