一、Android 中IPC 跨进程通信方式
1.1、Linux下的进程通信:
- 进程间隔离
进程与进程间内存是不共享的。两个进程就像两个平行的世界 - 进程空间划分:用户空间(User Space)/内核空间(Kernel Space)。
现在的操作系统采用的都是虚拟存储技术。对32位系统而言,它的虚拟存储空间是2^32,就是4GB。 操作系统的核心是内核,可以访问一些受保护的内存空间,也可以访问底层硬件的权限。为了保护内核的安全,逻辑上将虚拟内存空间划分为了用户空间和内核空间。 - 系统调用(SystemCall):用户态->内核态
虽然划分了用户空间和内核空间,但是用户某些情况下需要访问内核资源,就需要系统调用。
当进程执行系统调用而陷入内核代码中执行时,称此进程处于内核态;当进程在执行自己的代码时,称为用户态。
copy内存相关的系统调用
copy_from_user()//将数据从用户空间拷贝到内核空间
copy_to_user()//将数据从内核空间拷贝到用户空间
传统ipc
IPC 进程通信的基础:
每个进程的用户空间是彼此独立的,但是内核空间是进程共享的。进程A 可以将数据从用户空间copy到内核空间,然后由内核做中转,再将数据copy到 进程B中,达到跨进程通信的目的
传统的IPC通信存在两个问题
- 一次数据传递需要经过两次拷贝:内存缓存区-->内核缓存区-->内存缓存区
- 浪费空间和时间:接收区的缓存区由数据接收进程提供,但接收进程并不知道要多大的空间来存放传递过来的数据。
1.2、Binder IPC 通信
Binder IPC 也是利用内核空间 做数据中转,不同的是Binder在内核空间 为Server端用户进程做了内存映射。
Binder驱动内存映射: binder_mmap
- 首先在内核虚拟地址空间申请一块和用户虚拟内存大小相同的内存;
- 再申请1个page(页)大小的物理内存;
- 再将同一块物理内存分别映射到内核虚拟地址空间和用户虚拟内存空间。
这样就实现了用户空间的Buffer和内核空间的Buffer同步操作的功能。这使得Binder通信只需要从用户空间复制一次信息到内核空间就可以了
image1.3、Binder 与linux进程通信的区别
Linux进程通信机制大概分为:管道、消息队列、socket和共享内存。
- 效率方面:消息队列、socket、管道 都需要两次拷贝,Binder只需要一次拷贝。
两个进程都使用mmap就是共享内存;一个进程是用copy_from_user,另一个进程使用mmap 就是用的Binder机制。
- 稳定性方面:共享内存的性能优于Binder,但是共享内存需要处理并发同步问题,容易出现死锁和资源竞争,稳定性差。Binder基于C/S架构 ,Server端与Client端相对独立,稳定性较好。
- 安全性方面:传统Linux IPC的接收方无法获得对方进程可靠的UID/PID,从而无法鉴别对方身份;而Binder机制为每个进程分配了UID/PID,且在Binder通信时会根据UID/PID进行有效性检测。
二、Binder的实现原理和构成
binder实现原理和构成Android Binder 通信有两套方法,一个是ServierManager 直接open、ioctl操作binder driver,另一个是通过Binder IPC 框架来进行通信。
Binder 是基于C/S 结构的,自上而下Java层(FrameWork层)、Native(包括JNI)层和Driver层。
FrameWork层 Binder 主要包括:Binder、BinderProxy、BinderInternal
Natvie层Binder主要包括:JavaBBinder、BBinder、BpBinder、ProcessState、IPCThreadState、ServiceManager
Driver层 也叫Kernel层,它直接操作“/dev/binder” 设备节点。主要操作open、ioctl、mmap()等。
Java Binder是native Binder的一个映射(Mirror),Java Binder 跨进程通信依托于由native Binder实现具体功能,最后都经过driver设备节点。
下面分别就各个类做一下简答的介绍。
2.1、FrameWork Binder
-
IBinder接口常量FLAG_ONEWAY:客户端利用Binder跟服务端通信是阻塞式的,但如果设置了FLAG_ONEWAY,这成为非阻塞式的调用方式,客户端能立即返回,服务端采用回调方式来通知客户端完成情况。另外IBinder接口有一个内部接口DeathDecipient(死亡通告)
-
Binder 是服务端类,为对端提供具体服务。mObject持有native层的JavaBBinderHolder,子类onTransact()方法实现具体的操作。
public class Binder implements IBinder {
/**
* Raw native pointer to JavaBBinderHolder object. Owned by this Java object. Not null.
*/
private final long mObject; //指向native层的JavaBBinderHolder对象
private IInterface mOwner;
private String mDescriptor;
public Binder() {
mObject = getNativeBBinderHolder();
}
public @Nullable IInterface queryLocalInterface(@NonNull String descriptor) {
if (mDescriptor != null && mDescriptor.equals(descriptor)) {
return mOwner;
}
return null;
}
protected boolean onTransact(int code, @NonNull Parcel data, @Nullable Parcel reply,
int flags){
}
}
- BinderProxy Binder代理类,用户Client端,向Server端发送IPC请求
final class BinderProxy implements IBinder {
/**
* C++ pointer to BinderProxyNativeData. That consists of strong pointers to the
* native IBinder object, and a DeathRecipientList.
*/
private final long mNativeData;
private BinderProxy(long nativeData) {
mNativeData = nativeData;
}
public boolean transact(int code, Parcel data, Parcel reply, int flags) throws RemoteException {
Binder.checkParcel(this, code, data, "Unreasonably large binder buffer");
try {
return transactNative(code, data, reply, flags);
} finally {
}
}
public native boolean transactNative(int code, Parcel data, Parcel reply,
int flags) throws RemoteException;
}
- BinderInternal 主要负责Binder的回收、获取ServiceManagerProxy实例等。
public class BinderInternal {
//GcWatcher 用于管理Binder的销毁和回收。
static WeakReference<BinderInternal.GcWatcher> sGcWatcher = new WeakReference(new BinderInternal.GcWatcher());
public BinderInternal() {
}
public static void addGcWatcher(Runnable watcher) {
synchronized(sGcWatchers) {
sGcWatchers.add(watcher);
}
}
//当前线程加入Binder线程池
public static final native void joinThreadPool();
//getContextObject() 用于获取handle=0 的BpBinder,也就是ServiceManager的BpBinder
public static final native IBinder getContextObject();
}
通过BinderInternal.getContextObject() 可以获得代表ServiceManager(C++)的BpBinder,进而生成ServiceManagerProxy()
- Java 层的ServiceManager 管理所有的Service(IBinder)类,提供addService(注册)和getService(查询)两种操作。所有的系统服务都要向ServiceManager注册,使用某个服务时,要通过ServiceManager.getService(name) 进行查询,获得BinderProxy对象。
具体的addService、getService() 最终是通过C++层的ServiceManager实现的。
frameworks/base/core/java/android/os/ServiceManager.java
ServiceManager{
@UnsupportedAppUsage
public static IBinder getService(String name) {
try {
IBinder service = sCache.get(name);
if (service != null) {
return service;
} else {
return Binder.allowBlocking(rawGetService(name));
}
} catch (RemoteException e) {
Log.e(TAG, "error in getService", e);
}
return null;
}
public static void addService(String name, IBinder service, boolean allowIsolated,
int dumpPriority) {
try {
getIServiceManager().addService(name, service, allowIsolated, dumpPriority);
} catch (RemoteException e) {
Log.e(TAG, "error in addService", e);
}
}
private static IServiceManager getIServiceManager() {
if (sServiceManager != null) {
return sServiceManager;
}
// Find the service manager
sServiceManager = ServiceManagerNative
.asInterface(Binder.allowBlocking(BinderInternal.getContextObject()));
return sServiceManager;
}
}
public final class ServiceManagerNative {
private ServiceManagerNative() {}
@UnsupportedAppUsage
public static IServiceManager asInterface(IBinder obj) {
if (obj == null) {
return null;
}
// ServiceManager is never local
return new ServiceManagerProxy(obj);
}
}
2.2 Native 层Binder
2.2.1、 BpBinder: Native层 Binder代理类,持有handle属性,记录Binder节点的id,是Client端代理类。
2.2.2、JavaBBinder:C++类,继承BBinder类,是Server端Binder的代表;
继承关系:
JavaBBinder -> BBinder->IBinder->RefBase
BpBinder->IBinder->RefBase
持有关系:
|Java | Binder.mObject -> |C++| JavaBBinderHolder->JavaBBinder
|Java | BinderProxy.mNativeData -> |C++| BinderProxyNativeData -> BinderProxyNativeData.mObject-> BpBinder
Java层的Binder对象mObject属性持有JavaBBinderHolder(C++)类,JavaBBinderHolder持有JavaBBinder对象。
Java层的BinderProxy.mNativeData属性持有BinderProxyNativeData属性,BinderProxyNativeData持有BpBinder对象。
2.2.3、ProcessState类
ProcessState 是一个单例类,一个进程仅存在一个ProcessState实例。
ProcessState 实例化时 做了两件事情:
- 调用 open("/dev/binder")打开设备节点,持有Binder驱动的mDriverFD
static int open_driver(const char *driver)
{
//"/dev/binder"
int fd = open(driver, O_RDWR | O_CLOEXEC);
if (fd >= 0) {
int vers = 0;
status_t result = ioctl(fd, BINDER_VERSION, &vers);
size_t maxThreads = DEFAULT_MAX_BINDER_THREADS;
result = ioctl(fd, BINDER_SET_MAX_THREADS, &maxThreads);
} else {
ALOGW("Opening '%s' failed: %s\n", driver, strerror(errno));
}
return fd;
}
//调用mmap 申请共享内存,共享内存大小为1M-8k.
if (mDriverFD >= 0) {
// mmap the binder, providing a chunk of virtual address space to receive transactions.
mVMStart = mmap(nullptr, BINDER_VM_SIZE, PROT_READ, MAP_PRIVATE | MAP_NORESERVE, mDriverFD, 0);
}
2.4、IPCThreadState Binder的线程管理类
主要工作:
- 通过ProcessState提供的mDriverFD,执行ioctl()向binder driver 发送和接收数据。
- joinThreadPool 开启无限循环,以轮询的方式 从Binder Driver 读取数据。
IPCThreadState 内部有有mIn和mOut 两个Parcel类型的属性。
mIn - Parcel对象 用于接收/dev/binder驱动发来的消息
mOut - Parcel对象, 用于向/dev/binder驱动 发送消息
IPCThreadState::talkWithDriver 会从mOut中读取指令,发送给Binder Driver;Binder Driver 收到的指令会写到mInt中
三、Java Binder和Native Binder是如何建立联系的。
我们平时使用Binder用的是 Java层的Binder,但最终会调用Natvie 层的Binder,实现具体的功能,那么java 层的Binder和Natvie层的Binder是如何建立联系的呢?
在Android系统开机过程中,Zygote启动时会有一个虚拟机注册过程,该过程调用AndroidRunntime::startReg方法来完成jni方法的注册。
startReg
int AndroidRuntime::startReg(JNIEnv* env)
{
androidSetCreateThreadFunc((android_create_thread_fn) javaCreateThreadEtc);
env->PushLocalFrame(200);
//注册jni方法
if (register_jni_procs(gRegJNI, NELEM(gRegJNI), env) < 0) {
env->PopLocalFrame(NULL);
return -1;
}
env->PopLocalFrame(NULL);
return 0;
}
注册jni方法,其中gRegJNI是一个数组,记录所有需要注册的jni方法,其中有一项就是REG_JNI(register_android_os_Binder)。
register_android_os_Binder
int register_android_os_Binder(JNIEnv* env) {
// 注册Binder类的jni方法
if (int_register_android_os_Binder(env) < 0)
return -1;
// 注册BinderInternal类的jni方法
if (int_register_android_os_BinderInternal(env) < 0)
return -1;
// 注册BinderProxy类的jni方法
if (int_register_android_os_BinderProxy(env) < 0)
return -1;
...
return 0;
}
注册Binder
static int int_register_android_os_Binder(JNIEnv* env) {
//其中kBinderPathName = "android/os/Binder";查找kBinderPathName路径所属类
jclass clazz = FindClassOrDie(env, kBinderPathName);
//将Java层Binder类保存到mClass变量;
gBinderOffsets.mClass = MakeGlobalRefOrDie(env, clazz);
//将Java层execTransact()方法保存到mExecTransact变量;
gBinderOffsets.mExecTransact = GetMethodIDOrDie(env, clazz, "execTransact", "(IJJI)Z");
//将Java层mObject属性保存到mObject变量
gBinderOffsets.mObject = GetFieldIDOrDie(env, clazz, "mObject", "J");
//注册JNI方法
return RegisterMethodsOrDie(env, kBinderPathName, gBinderMethods,
NELEM(gBinderMethods));
}
把有关Binder的重要变量都保存到了gBinderOffsets中。主要有三步:
- mClass:保存java层的Binder类
- mExecTransact:保存execTransact()方法
- mObject:保存mObject属性
gBinderOffsets
gBinderOffsets是全局静态结构体,定义如下:
static struct bindernative_offsets_t
{
jclass mClass; //记录Binder类
jmethodID mExecTransact; //记录execTransact()方法
jfieldID mObject; //记录mObject属性
} gBinderOffsets;
gBinderOffsets保存了Binder.java类本身以及成员方法execTransact()和成员属性mObject,这为JNI层访问java层提供通道。
gBidnerOffsets结构体保存binder类信息也是一种空间换时间的方法,不需要每次查找binder类信息,提高了查询效率。
gBinderMethods
static const JNINativeMethod gBinderMethods[] = {
/* 名称, 签名, 函数指针 */
{ "getCallingPid", "()I", (void*)android_os_Binder_getCallingPid },
{ "getCallingUid", "()I", (void*)android_os_Binder_getCallingUid },
{ "clearCallingIdentity", "()J", (void*)android_os_Binder_clearCallingIdentity },
{ "restoreCallingIdentity", "(J)V", (void*)android_os_Binder_restoreCallingIdentity },
{ "setThreadStrictModePolicy", "(I)V", (void*)android_os_Binder_setThreadStrictModePolicy },
{ "getThreadStrictModePolicy", "()I", (void*)android_os_Binder_getThreadStrictModePolicy },
{ "flushPendingCommands", "()V", (void*)android_os_Binder_flushPendingCommands },
{ "init", "()V", (void*)android_os_Binder_init },
{ "destroy", "()V", (void*)android_os_Binder_destroy },
{ "blockUntilThreadAvailable", "()V", (void*)android_os_Binder_blockUntilThreadAvailable }
};
RegisterMethodsOrDie()中为gBinderMethods数组中的方法建立了一一映射关系,从而为java层访问JNI层提供通道。
int_register_android_os_Binder方法的主要功能:
- 通过gBinderOffsets,保存java层Binder类的信息,为JNI层访问java层提供通道
- 通过RegisterMethodsOrDie,将gBinderMethods数组完成映射关系,从而为java层访问jni层提供通道。
也就是说该过程建立了Binder类在Native层与framework层之间相互调用的桥梁。
注册BinderInternal
static int int_register_android_os_BinderInternal(JNIEnv* env) {
//其中kBinderInternalPathName = "com/android/internal/os/BinderInternal"
jclass clazz = FindClassOrDie(env, kBinderInternalPathName);
gBinderInternalOffsets.mClass = MakeGlobalRefOrDie(env, clazz);
gBinderInternalOffsets.mForceGc = GetStaticMethodIDOrDie(env, clazz, "forceBinderGc", "()V");
return RegisterMethodsOrDie(
env, kBinderInternalPathName,
gBinderInternalMethods, NELEM(gBinderInternalMethods));
}
注册了BinderInternal类的jni方法,gBinderInternalOffsets保存了BinderInternal的forceBinderGC()方法。
下面是BinderInternal类的jni方法注册
static const JNINativeMethod gBinderInternalMethods[] = {
{ "getContextObject", "()Landroid/os/IBinder;", (void*)android_os_BinderInternal_getContextObject },
{ "joinThreadPool", "()V", (void*)android_os_BinderInternal_joinThreadPool },
{ "disableBackgroundScheduling", "(Z)V", (void*)android_os_BinderInternal_disableBackgroundScheduling },
{ "handleGc", "()V", (void*)android_os_BinderInternal_handleGc }
};
和注册Binder非常类似,该过程建立了BinderInternal类在Native层与framework层之间的相互调用的桥梁。
注册BinderProxy
static int int_register_android_os_BinderProxy(JNIEnv* env) {
//gErrorOffsets保存了Error类信息
jclass clazz = FindClassOrDie(env, "java/lang/Error");
gErrorOffsets.mClass = MakeGlobalRefOrDie(env, clazz);
//gBinderProxyOffsets保存了BinderProxy类的信息
//其中kBinderProxyPathName = "android/os/BinderProxy"
clazz = FindClassOrDie(env, kBinderProxyPathName);
gBinderProxyOffsets.mClass = MakeGlobalRefOrDie(env, clazz);
gBinderProxyOffsets.mConstructor = GetMethodIDOrDie(env, clazz, "<init>", "()V");
gBinderProxyOffsets.mSendDeathNotice = GetStaticMethodIDOrDie(env, clazz, "sendDeathNotice", "(Landroid/os/IBinder$DeathRecipient;)V");
gBinderProxyOffsets.mObject = GetFieldIDOrDie(env, clazz, "mObject", "J");
gBinderProxyOffsets.mSelf = GetFieldIDOrDie(env, clazz, "mSelf", "Ljava/lang/ref/WeakReference;");
gBinderProxyOffsets.mOrgue = GetFieldIDOrDie(env, clazz, "mOrgue", "J");
//gClassOffsets保存了Class.getName()方法
clazz = FindClassOrDie(env, "java/lang/Class");
gClassOffsets.mGetName = GetMethodIDOrDie(env, clazz, "getName", "()Ljava/lang/String;");
return RegisterMethodsOrDie(
env, kBinderProxyPathName,
gBinderProxyMethods, NELEM(gBinderProxyMethods));
}
注册BinderProxy类的jni方法,gBinderProxyOffsets保存了BinderProxy的构造方法,sendDeathNotice(),mObject,mSelf,mOrgue信息。
BinderProxy类在Binder类中,是本地IBinder对象的java代理,BinderProxy 是由C++层创建后,传递到java层的。
ibinderForJavaObject:Java类转换成C++ IBinder类
将java Binder 类对象转换成JavaBBinder;
将Java BinderProxy 转换成BpBinder
sp<IBinder> ibinderForJavaObject(JNIEnv* env, jobject obj)
{
if (obj == NULL) return NULL;
// Instance of Binder?
if (env->IsInstanceOf(obj, gBinderOffsets.mClass)) {
JavaBBinderHolder* jbh = (JavaBBinderHolder*)
env->GetLongField(obj, gBinderOffsets.mObject);
return jbh->get(env, obj);
}
// Instance of BinderProxy?
if (env->IsInstanceOf(obj, gBinderProxyOffsets.mClass)) {
return getBPNativeData(env, obj)->mObject;
}
ALOGW("ibinderForJavaObject: %p is not a Binder object", obj);
return NULL;
}
javaObjectForIBinder C++ IBinder类转换成对应的Java类
C++类 JavaBBBinder-> Java类 Binder
C++类 BpBinder -> Java类的 BinderProxy
object javaObjectForIBinder(JNIEnv* env, const sp<IBinder>& val) {
if (val == NULL) return NULL;
if (val->checkSubclass(&gBinderOffsets)) { //返回false
jobject object = static_cast<JavaBBinder*>(val.get())->object();
return object;
}
AutoMutex _l(mProxyLock);
jobject object = (jobject)val->findObject(&gBinderProxyOffsets);
if (object != NULL) { //第一次object为null
jobject res = jniGetReferent(env, object);
if (res != NULL) {
return res;
}
android_atomic_dec(&gNumProxyRefs);
val->detachObject(&gBinderProxyOffsets);
env->DeleteGlobalRef(object);
}
//创建BinderProxy对象
object = env->NewObject(gBinderProxyOffsets.mClass, gBinderProxyOffsets.mConstructor);
if (object != NULL) {
//BinderProxy.mObject成员变量记录BpBinder对象
env->SetLongField(object, gBinderProxyOffsets.mObject, (jlong)val.get());
val->incStrong((void*)javaObjectForIBinder);
jobject refObject = env->NewGlobalRef(
env->GetObjectField(object, gBinderProxyOffsets.mSelf));
//将BinderProxy对象信息附加到BpBinder的成员变量mObjects中
val->attachObject(&gBinderProxyOffsets, refObject,
jnienv_to_javavm(env), proxy_cleanup);
sp<DeathRecipientList> drl = new DeathRecipientList;
drl->incStrong((void*)javaObjectForIBinder);
//BinderProxy.mOrgue成员变量记录死亡通知对象
env->SetLongField(object, gBinderProxyOffsets.mOrgue, reinterpret_cast<jlong>(drl.get()));
android_atomic_inc(&gNumProxyRefs);
incRefsCreated(env);
}
return object;
}
四、Binder 数据传递过程分析
Android系统中的众多服务(Binder)都需要向ServerManager进行注册,当需要使用某个Service时 再从ServiceManager进行查询。
ServiceManager.addService() 和getService()来简单看下,ServiceManager是如何注册和维护Service(IBinder)的
4.1、SM.addService()
public static void addService(String name, IBinder service, boolean allowIsolated) {
try {
//先获取SMP对象,则执行注册服务操作
getIServiceManager().addService(name, service, allowIsolated);
} catch (RemoteException e) {
Log.e(TAG, "error in addService", e);
}
}
getIServiceManager
private static IServiceManager getIServiceManager() {
if (sServiceManager != null) {
return sServiceManager;
}
//这里调用了ServiceManagerNative获取sServiceManager
sServiceManager = ServiceManagerNative.asInterface(BinderInternal.getContextObject());
return sServiceManager;
}
采用单例模式获取ServiceManager getIServiceManager()返回的是ServiceManagerProxy对象
getContextObject
static jobject android_os_BinderInternal_getContextObject(JNIEnv* env, jobject clazz)
{
//获取一个BpBinder(0) C++对象
sp<IBinder> b = ProcessState::self()->getContextObject(NULL);
//由BpBinder(0) 创建一个BinderProxy()对象
return javaObjectForIBinder(env, b);
}
BinderInternal.java中有一个native方法getContextObject(),JNI调用执行上述方法。
这里有个熟悉的ProcessState::self()->getContextObject()在分析获取ServiceManager的时候调用过,就相当于是new BpBinder(0)
javaObjectForIBinder
object javaObjectForIBinder(JNIEnv* env, const sp<IBinder>& val) {
if (val == NULL) return NULL;
if (val->checkSubclass(&gBinderOffsets)) { //返回false
jobject object = static_cast<JavaBBinder*>(val.get())->object();
return object;
}
AutoMutex _l(mProxyLock);
jobject object = (jobject)val->findObject(&gBinderProxyOffsets);
if (object != NULL) { //第一次object为null
jobject res = jniGetReferent(env, object);
if (res != NULL) {
return res;
}
android_atomic_dec(&gNumProxyRefs);
val->detachObject(&gBinderProxyOffsets);
env->DeleteGlobalRef(object);
}
//创建BinderProxy对象
object = env->NewObject(gBinderProxyOffsets.mClass, gBinderProxyOffsets.mConstructor);
if (object != NULL) {
//BinderProxy.mObject成员变量记录BpBinder对象
env->SetLongField(object, gBinderProxyOffsets.mObject, (jlong)val.get());
val->incStrong((void*)javaObjectForIBinder);
jobject refObject = env->NewGlobalRef(
env->GetObjectField(object, gBinderProxyOffsets.mSelf));
//将BinderProxy对象信息附加到BpBinder的成员变量mObjects中
val->attachObject(&gBinderProxyOffsets, refObject,
jnienv_to_javavm(env), proxy_cleanup);
sp<DeathRecipientList> drl = new DeathRecipientList;
drl->incStrong((void*)javaObjectForIBinder);
//BinderProxy.mOrgue成员变量记录死亡通知对象
env->SetLongField(object, gBinderProxyOffsets.mOrgue, reinterpret_cast<jlong>(drl.get()));
android_atomic_inc(&gNumProxyRefs);
incRefsCreated(env);
}
return object;
}
根据BpBinder(c++)生成BinderProxy(java)对象,并把BpBinder对象地址保存到BinderProxy.mObject成员变量。
ServiceManagerNative.asInterface(BinderInternal.getContextObject())
等价于
ServiceManagerNative.asInterface(new BinderProxy());
SMN.asInterface
static public IServiceManager asInterface(IBinder obj)
{
if (obj == null) { //obj为BpBinder
return null;
}
//由于obj为BpBinder,该方法默认返回null
IServiceManager in = (IServiceManager)obj.queryLocalInterface(descriptor);
if (in != null) {
return in;
}
return new ServiceManagerProxy(obj);
}
这里可以知道ServiceManagerNative.asInterface(new BinderProxy())等价于ServiceManagerProxy(new BinderProxy())
ServiceManagerProxy初始化
class ServiceManagerProxy implements IServiceManager {
public ServiceManagerProxy(IBinder remote) {
mRemote = remote;
}
}
mRemote为BinderProxy对象,该BinderProxy对象对应于BpBinder(0),其作为Binder代理端,指向Native层大管家SerivceManager。
framework层的ServiceManager的调用实际上是交给了ServiceManagerProxy的成员变量BinderProxy;
BinderProxy通过jni方式,最终会调用BpBinder对象。
4.2、SMP.addService()
public void addService(String name, IBinder service, boolean allowIsolated) throws RemoteException {
Parcel data = Parcel.obtain();
Parcel reply = Parcel.obtain();
data.writeInterfaceToken(IServiceManager.descriptor);
data.writeString(name);
data.writeStrongBinder(service);
data.writeInt(allowIsolated ? 1 : 0);
//mRemote为BinderProxy
mRemote.transact(ADD_SERVICE_TRANSACTION, data, reply, 0);
reply.recycle();
data.recycle();
}
大概做了两件事情:
1、 data.writeStrongBinder(service); 利用Parcel对Binder做处理后
2、 mRemote.transact(ADD_SERVICE_TRANSACTION, data, reply, 0);调用BinderProxy.trasanct() 向Binder Driver发消息.
writeStrongBinder
public writeStrongBinder(IBinder val){
//此处为Native调用
nativewriteStrongBinder(mNativePtr, val);
}
android_os_Parcel_writeStrongBinder
static void android_os_Parcel_writeStrongBinder(JNIEnv* env, jclass clazz, jlong nativePtr, jobject object) {
//将java层Parcel转换为native层Parcel
Parcel* parcel = reinterpret_cast<Parcel*>(nativePtr);
if (parcel != NULL) {
const status_t err = parcel->writeStrongBinder(ibinderForJavaObject(env, object));
if (err != NO_ERROR) {
signalExceptionForError(env, clazz, err);
}
}
}
ibinderForJavaObject
sp<IBinder> ibinderForJavaObject(JNIEnv* env, jobject obj)
{
if (obj == NULL) return NULL;
//Java层的Binder对象
if (env->IsInstanceOf(obj, gBinderOffsets.mClass)) {
JavaBBinderHolder* jbh = (JavaBBinderHolder*)
env->GetLongField(obj, gBinderOffsets.mObject);
return jbh != NULL ? jbh->get(env, obj) : NULL;
}
//Java层的BinderProxy对象
if (env->IsInstanceOf(obj, gBinderProxyOffsets.mClass)) {
return (IBinder*)env->GetLongField(obj, gBinderProxyOffsets.mObject);
}
return NULL;
}
frameworks/base/core/jni/android_util_Binder.cpp
JavaBBinderHolder
class JavaBBinderHolder
sp<JavaBBinder> get(JNIEnv* env, jobject obj)
{
AutoMutex _l(mLock);
sp<JavaBBinder> b = mBinder.promote();
if (b == NULL) {
b = new JavaBBinder(env, obj);
if (mVintf) {
::android::internal::Stability::markVintf(b.get());
}
if (mExtension != nullptr) {
b.get()->setExtension(mExtension);
}
mBinder = b;
ALOGV("Creating JavaBinder %p (refs %p) for Object %p, weakCount=%" PRId32 "\n",
b.get(), b->getWeakRefs(), obj, b->getWeakRefs()->getWeakCount());
}
return b;
}
Java Binder创建时 会自动创建一个JavaBBinderHolder对象。
JavaBBinderHolder有一个成员变量mBinder,保存当前创建的JavaBBinder对象,这是一个弱引用类型的,可能会被垃圾回收器给回收,所以每次使用前都需要判断是否存在。
当量mBinder不存在时,会创建一个新的JavaBBinder()实例,并且JavaBBinder持有Java 层的Binder对象。
此时的引用关系如下:
(Java) Binder->JavaBBinderHolder->JavaBBinder
JavaBBinder->(Java) Binder
data.writeStrongBinder(service)等价于parcel->writeStrongBinder(new JavaBinder(env,obj));
writeStrongBinder(C++)
status_t Parcel::writeStrongBinder(const sp<IBinder>& val)
{
return flatten_binder(ProcessState::self(), val, this);
}
flatten_binder
flatten_binder 将Binder信息扁平化,保存在flat_binder_object结构体中。
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();
const int32_t handle = proxy ? proxy->handle() : 0;
obj.type = BINDER_TYPE_HANDLE; //远程Binder
obj.binder = 0;
obj.handle = handle;
obj.cookie = 0;
} else {
obj.type = BINDER_TYPE_BINDER; //本地Binder,进入该分支
obj.binder = reinterpret_cast<uintptr_t>(local->getWeakRefs());
obj.cookie = reinterpret_cast<uintptr_t>(local);
}
} else {
obj.type = BINDER_TYPE_BINDER; //本地Binder
obj.binder = 0;
obj.cookie = 0;
}
return finish_flatten_binder(binder, obj, out);
}
- 对于Binder实体,则cookie记录Binder实体指针,type= BINDER_TYPE_BINDER
- 对于Binder代理,则handle记录Binder代理的句柄;type = BINDER_TYPE_HANDLE
由此可知在IPC前,Binder对象的处理过程:
- Parce.writeStrongBinder() - 输入Binder
- ibinderForJavaObject - 返回BpBinder
- flatten_binder - 返回flat_binder_object
- (C++) parcel.writeObject(flat); flat_binder_object保存到parcel
BinderProxy.transact
Binder对象被保存在Parcel对象后,调用BinderProxy.transact 执行进入到IPC的调用流程中。
public boolean transact(int code, Parcel data, Parcel reply, int flags) throws RemoteException {
//用于检测Parcel大小是否大于800k
Binder.checkParcel(this, code, data, "Unreasonably large binder buffer");
return transactNative(code, data, reply, flags);
}
mRemote是BinderProxy。transactNative经过jni调用,进入下面的方法
android_os_BinderProxy_transact
static jboolean android_os_BinderProxy_transact(JNIEnv* env, jobject obj,
jint code, jobject dataObj, jobject replyObj, jint flags)
{
...
//java Parcel转为native Parcel
Parcel* data = parcelForJavaObject(env, dataObj);
Parcel* reply = parcelForJavaObject(env, replyObj);
...
//gBinderProxyOffsets.mObject中保存的是new BpBinder(0)对象
IBinder* target = (IBinder*)
env->GetLongField(obj, gBinderProxyOffsets.mObject);
...
//此处便是BpBinder::transact(), 经过native层,进入Binder驱动程序
status_t err = target->transact(code, *data, reply, flags);
...
return JNI_FALSE;
}
java层的BinderProxy.transact()最终交由Native层的BpBinder::transact()完成。
addService的核心过程:
public void addService(String name, IBinder service, boolean allowIsolated) throws RemoteException {
...
Parcel data = Parcel.obtain(); //此处还需要将java层的Parcel转为Native层的Parcel
data->writeStrongBinder(new JavaBBinder(env, obj));
BpBinder::transact(ADD_SERVICE_TRANSACTION, *data, reply, 0); //与Binder驱动交互
...
}
4.2、 SM.getService
ServiceManager.java
public static IBinder getService(String name) {
try {
IBinder service = sCache.get(name); //先从缓存中查看
if (service != null) {
return service;
} else {
return getIServiceManager().getService(name); 【见4.2】
}
} catch (RemoteException e) {
Log.e(TAG, "error in getService", e);
}
return null;
}
优先从sCache缓存中 获取已查询的IBinder,若缓存未命中,则从ServiceManagerProxy()中来查询
SMP.getService
class ServiceManagerProxy implements IServiceManager {
public IBinder getService(String name) throws RemoteException {
Parcel data = Parcel.obtain();
Parcel reply = Parcel.obtain();
data.writeInterfaceToken(IServiceManager.descriptor);
data.writeString(name);
//mRemote为BinderProxy 【见4.3】
mRemote.transact(GET_SERVICE_TRANSACTION, data, reply, 0);
//从reply里面解析出获取的IBinder对象【见4.8】
IBinder binder = reply.readStrongBinder();
reply.recycle();
data.recycle();
return binder;
}
}
BinderProxy.transact
android_util_Binder.cpp
final class BinderProxy implements IBinder {
public boolean transact(int code, Parcel data, Parcel reply, int flags) throws RemoteException {
Binder.checkParcel(this, code, data, "Unreasonably large binder buffer");
return transactNative(code, data, reply, flags);
}
}
android_os_BinderProxy_transact
static jboolean android_os_BinderProxy_transact(JNIEnv* env, jobject obj,
jint code, jobject dataObj, jobject replyObj, jint flags)
{
...
//java Parcel转为native Parcel
Parcel* data = parcelForJavaObject(env, dataObj);
Parcel* reply = parcelForJavaObject(env, replyObj);
...
//gBinderProxyOffsets.mObject中保存的是new BpBinder(0)对象
IBinder* target = (IBinder*)
env->GetLongField(obj, gBinderProxyOffsets.mObject);
...
//此处便是BpBinder::transact(), 经过native层[见小节4.5]
status_t err = target->transact(code, *data, reply, flags);
...
return JNI_FALSE;
}
BpBinder.transact
BpBinder.cpp
status_t BpBinder::transact(
uint32_t code, const Parcel& data, Parcel* reply, uint32_t flags)
{
if (mAlive) {
// [见小节4.6]
status_t status = IPCThreadState::self()->transact(
mHandle, code, data, reply, flags);
if (status == DEAD_OBJECT) mAlive = 0;
return status;
}
return DEAD_OBJECT;
}
IPC.transact
IPCThreadState.cpp
status_t IPCThreadState::transact(int32_t handle,
uint32_t code, const Parcel& data,
Parcel* reply, uint32_t flags)
{
status_t err = data.errorCheck(); //数据错误检查
flags |= TF_ACCEPT_FDS;
....
if (err == NO_ERROR) {
//(1)将handle、code、data暂存在binder_transaction_data结构体,然后写如到mOut(Parcel)中
err = writeTransactionData(BC_TRANSACTION, flags, handle, code, data, NULL);
}
...
// 默认情况下,都是采用非oneway的方式, 也就是需要等待服务端的返回结果
if ((flags & TF_ONE_WAY) == 0) {
if (reply) {
//(2)执行waitForResponse,等待回应事件
err = waitForResponse(reply);
}else {
Parcel fakeReply;
err = waitForResponse(&fakeReply);
}
} else {
err = waitForResponse(NULL, NULL);
}
return err;
}
writeTransactionData
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.ptr = 0; /* Don't pass uninitialized stack data to a remote process */
tr.target.handle = handle;
tr.code = code;
tr.flags = binderFlags;
tr.cookie = 0;
tr.sender_pid = 0;
tr.sender_euid = 0;
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(binder_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 = reinterpret_cast<uintptr_t>(statusBuffer);
tr.offsets_size = 0;
tr.data.ptr.offsets = 0;
} else {
return (mLastError = err);
}
mOut.writeInt32(cmd);
mOut.write(&tr, sizeof(tr));
return NO_ERROR;
}
IPC.waitForResponse
status_t IPCThreadState::waitForResponse(Parcel *reply, status_t *acquireResult)
{
int32_t cmd;
int32_t err;
while (1) {
if ((err=talkWithDriver()) < NO_ERROR) break; // 【见小节2.11】
err = mIn.errorCheck();
if (err < NO_ERROR) break; //当存在error则退出循环
//每当跟Driver交互一次,若mIn收到数据则往下执行一次BR命令
if (mIn.dataAvail() == 0) continue;
cmd = mIn.readInt32();
switch (cmd) {
case BR_TRANSACTION_COMPLETE:
//只有当不需要reply, 也就是oneway时 才会跳出循环,否则还需要等待.
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_REPLY: ... goto finish;
default:
err = executeCommand(cmd); //【见小节2.12】
if (err != NO_ERROR) goto finish;
break;
}
}
finish:
if (err != NO_ERROR) {
if (reply) reply->setError(err); //将发送的错误代码返回给最初的调用者
}
return err;
}
在这个过程中, 收到以下任一BR_命令,处理后便会退出waitForResponse()的状态:
- BR_TRANSACTION_COMPLETE binder驱动收到BC_TRANSACTION事件后的应答消息; 对于oneway transaction,当收到该消息,则完成了本次Binder通信;
- BR_DEAD_REPLY:回复失败,往往是线程或节点为空. 则结束本次通信Binder;
- BR_FAILED_REPLY::回复失败,往往是transaction出错导致. 则结束本次通信Binder;
- BR_REPLY: Binder驱动向Client端发送回应消息; 对于非oneway transaction时,当收到该消息,则完整地完成本次Binder通信;
除了以上命令,其他命令 执行executeCommand()
这里SM.addService() 采用的是非ONE_WAY方法,所以waitForResponse中,收到BC_REPLAY时,返回调用者进程。
class ServiceManagerProxy implements IServiceManager {
public IBinder getService(String name) throws RemoteException {
Parcel data = Parcel.obtain();
Parcel reply = Parcel.obtain();
//(1)设置parcel参数
data.writeInterfaceToken(IServiceManager.descriptor);
data.writeString(name);
//(2)mRemote为BinderProxy 【见4.3】
mRemote.transact(GET_SERVICE_TRANSACTION, data, reply, 0);
//(3)从reply里面解析出获取的IBinder对象【见4.8】
IBinder binder = reply.readStrongBinder();
reply.recycle();
data.recycle();
return binder;
}
}
也就是会回到了(3)处
IBinder binder = reply.readStrongBinder();
Parcel.java
static jobject android_os_Parcel_readStrongBinder(JNIEnv* env, jclass clazz, jlong nativePtr) {
Parcel* parcel = reinterpret_cast<Parcel*>(nativePtr);
if (parcel != NULL) {
//【见小节4.8.1】
return javaObjectForIBinder(env, parcel->readStrongBinder());
}
return NULL;
}
javaObjectForIBinder 将native层BpBinder对象转换为Java层BinderProxy对象。
readStrongBinder(C++)
sp<IBinder> Parcel::readStrongBinder() const
{
sp<IBinder> val;
//【见小节4.8.2】
unflatten_binder(ProcessState::self(), *this, &val);
return val;
}
unflatten_binder
status_t unflatten_binder(const sp<ProcessState>& proc,
const Parcel& in, sp<IBinder>* out)
{
const flat_binder_object* flat = in.readObject(false);
if (flat) {
switch (flat->type) {
case BINDER_TYPE_BINDER:
*out = reinterpret_cast<IBinder*>(flat->cookie);
return finish_unflatten_binder(NULL, *flat, in);
case BINDER_TYPE_HANDLE:
//进入该分支【见4.8.3】
*out = proc->getStrongProxyForHandle(flat->handle);
//创建BpBinder对象
return finish_unflatten_binder(
static_cast<BpBinder*>(out->get()), *flat, in);
}
}
return BAD_TYPE;
}
getStrongProxyForHandle
ProcessState.cpp
sp<IBinder> ProcessState::getStrongProxyForHandle(int32_t handle)
{
sp<IBinder> result;
AutoMutex _l(mLock);
//查找handle对应的资源项
handle_entry* e = lookupHandleLocked(handle);
if (e != NULL) {
IBinder* b = e->binder;
if (b == NULL || !e->refs->attemptIncWeak(this)) {
...
//当handle值所对应的IBinder不存在或弱引用无效时,则创建BpBinder对象
b = new BpBinder(handle);
e->binder = b;
if (b) e->refs = b->getWeakRefs();
result = b;
} else {
result.force_set(b);
e->refs->decWeak(this);
}
}
return result;
}
经过该方法,最终创建了指向Binder服务端的BpBinder代理对象。
javaObjectForIBinder将native层BpBinder对象转换为Java层BinderProxy对象。
getService()最终获取了指向目标Binder服务端的代理对象BinderProxy。
Binder.readStrongBinder 具体过程:
- Parcel.unflatten_binder 返回flat_binder_object
- ProcessState.getStrongProxyForHandle(handle):返回BpBinder
- javaObjectForIBinder 返回BinderProxy对象
getService的核心过程:
public static IBinder getService(String name) {
...
Parcel reply = Parcel.obtain(); //此处还需要将java层的Parcel转为Native层的Parcel
BpBinder::transact(GET_SERVICE_TRANSACTION, *data, reply, 0); //与Binder驱动交互
IBinder binder = javaObjectForIBinder(env, new BpBinder(handle));
...
}
五、Binder 线程创建
Binder线程的创建都是伴随着Java进程的创建而创建的。
ava层进程的创建都是通过Process.start()方法,向Zygote进程发出创建进程的socket消息。Zygote收到消息后会调用Zygote.forkAndSpecialize()来fork出新进程,在新进程中会调用到RuntimeInit.nativeZygoteInit方法,该方法经过jni映射,最终会调用到app_main.cpp中的onZygoteInit。
- Process.start()
- Zygote.forkAndSpecialize()
- RuntimeInit.nativeZygoteInit
- onZygoteInit
5.1、onZygoteInit
[-> app_main.cpp]
virtual void onZygoteInit() {
//获取ProcessState对象
sp<ProcessState> proc = ProcessState::self();
//启动新binder线程 【见小节2.2】
proc->startThreadPool();
}
ProcessState::self()是单例模式,主要工作是调用open()打开/dev/binder驱动设备,再利用mmap()映射内核的地址空间,将Binder驱动的fd赋值ProcessState对象中的变量mDriverFD,用于交互操作。startThreadPool()是创建一个新的binder线程,不断进行talkWithDriver()。
5.2、 PS.startThreadPool
[-> ProcessState.cpp]
void ProcessState::startThreadPool()
{
AutoMutex _l(mLock); //多线程同步
if (!mThreadPoolStarted) {
mThreadPoolStarted = true;
spawnPooledThread(true); 【见小节2.3】
}
启动Binder线程池后, 则设置mThreadPoolStarted=true. 通过变量mThreadPoolStarted来保证每个应用进程只允许启动一个binder主线程池。本次创建的是binder主线程(isMain=true). 其余binder线程池中的线程都是由Binder驱动来控制创建的。
5.3、PS.spawnPooledThread
void ProcessState::spawnPooledThread(bool isMain)
{
if (mThreadPoolStarted) {
//获取Binder线程名【见小节2.3.1】
String8 name = makeBinderThreadName();
//此处isMain=true【见小节2.3.2】
sp<Thread> t = new PoolThread(isMain);
t->run(name.string());
}
}
5.4 PoolThread.run
[-> ProcessState.cpp]
class PoolThread : public Thread
{
public:
PoolThread(bool isMain)
: mIsMain(isMain)
{
}
protected:
virtual bool threadLoop() {
IPCThreadState::self()->joinThreadPool(mIsMain); //【见小节2.4】
return false;
}
const bool mIsMain;
};
从函数名看起来是创建线程池,其实就只是创建一个线程,该PoolThread继承Thread类。t->run()方法最终调用 PoolThread的threadLoop()方法。
5.5、IPC.joinThreadPool
[-> IPCThreadState.cpp]
void IPCThreadState::joinThreadPool(bool isMain)
{
//创建Binder线程
mOut.writeInt32(isMain ? BC_ENTER_LOOPER : BC_REGISTER_LOOPER);
set_sched_policy(mMyThreadId, SP_FOREGROUND); //设置前台调度策略
status_t result;
do {
processPendingDerefs(); //清除队列的引用
result = getAndExecuteCommand(); //处理下一条指令
if (result < NO_ERROR && result != TIMED_OUT
&& result != -ECONNREFUSED && result != -EBADF) {
abort();
}
if(result == TIMED_OUT && !isMain) {
break; //非主线程出现timeout则线程退出
}
} while (result != -ECONNREFUSED && result != -EBADF);
mOut.writeInt32(BC_EXIT_LOOPER); // 线程退出循环
talkWithDriver(false); //false代表bwr数据的read_buffer为空
}
- 对于isMain=true的情况下, command为BC_ENTER_LOOPER,代表的是Binder主线程,不会退出的线程;
- 对于isMain=false的情况下,command为BC_REGISTER_LOOPER,表示是由binder驱动创建的线程。非主线程出现timeout则线程退出。
5.6、IPC.getAndExecuteCommand
status_t IPCThreadState::getAndExecuteCommand()
{
status_t result;
int32_t cmd;
result = talkWithDriver(); //与binder进行交互
if (result >= NO_ERROR) {
size_t IN = mIn.dataAvail();
if (IN < sizeof(int32_t)) return result;
cmd = mIn.readInt32();
pthread_mutex_lock(&mProcess->mThreadCountLock);
mProcess->mExecutingThreadsCount++;
pthread_mutex_unlock(&mProcess->mThreadCountLock);
result = executeCommand(cmd); //执行Binder响应码
pthread_mutex_lock(&mProcess->mThreadCountLock);
mProcess->mExecutingThreadsCount--;
pthread_cond_broadcast(&mProcess->mThreadCountDecrement);
pthread_mutex_unlock(&mProcess->mThreadCountLock);
set_sched_policy(mMyThreadId, SP_FOREGROUND);
}
return result;
}
5.7、talkWithDriver
//mOut有数据,mIn还没有数据。doReceive默认值为true
status_t IPCThreadState::talkWithDriver(bool doReceive)
{
binder_write_read bwr;
...
// 当同时没有输入和输出数据则直接返回
if ((bwr.write_size == 0) && (bwr.read_size == 0)) return NO_ERROR;
...
do {
//ioctl执行binder读写操作,经过syscall,进入Binder驱动。调用Binder_ioctl
if (ioctl(mProcess->mDriverFD, BINDER_WRITE_READ, &bwr) >= 0)
err = NO_ERROR;
...
} while (err == -EINTR);
...
return err;
}
5.8、binder_thread_write
[-> binder.c]
static int binder_thread_write(struct binder_proc *proc,
struct binder_thread *thread,
binder_uintptr_t binder_buffer, size_t size,
binder_size_t *consumed)
{
uint32_t cmd;
void __user *buffer = (void __user *)(uintptr_t)binder_buffer;
void __user *ptr = buffer + *consumed;
void __user *end = buffer + size;
while (ptr < end && thread->return_error == BR_OK) {
//拷贝用户空间的cmd命令,此时为BC_ENTER_LOOPER
if (get_user(cmd, (uint32_t __user *)ptr)) -EFAULT;
ptr += sizeof(uint32_t);
switch (cmd) {
case BC_REGISTER_LOOPER:
if (thread->looper & BINDER_LOOPER_STATE_ENTERED) {
//出错原因:线程调用完BC_ENTER_LOOPER,不能执行该分支
thread->looper |= BINDER_LOOPER_STATE_INVALID;
} else if (proc->requested_threads == 0) {
//出错原因:没有请求就创建线程
thread->looper |= BINDER_LOOPER_STATE_INVALID;
} else {
proc->requested_threads--;
proc->requested_threads_started++;
}
thread->looper |= BINDER_LOOPER_STATE_REGISTERED;
break;
case BC_ENTER_LOOPER:
if (thread->looper & BINDER_LOOPER_STATE_REGISTERED) {
//出错原因:线程调用完BC_REGISTER_LOOPER,不能立刻执行该分支
thread->looper |= BINDER_LOOPER_STATE_INVALID;
}
//创建Binder主线程
thread->looper |= BINDER_LOOPER_STATE_ENTERED;
break;
case BC_EXIT_LOOPER:
thread->looper |= BINDER_LOOPER_STATE_EXITED;
break;
}
...
}
*consumed = ptr - buffer;
}
return 0;
}
处理完BC_ENTER_LOOPER命令后,一般情况下会设置成功,为thread->looper增加BINDER_LOOPER_STATE_ENTERED标志。
thread->looper |= BINDER_LOOPER_STATE_ENTERED。
至此主Binder线程的已经创建完成。之后就可以正常接收Binder消息了。
5.9、Binder驱动主动创建Binder线程的时机。
binder_thread_read
binder_thread_read(){
...
retry:
//当前线程todo队列为空且transaction栈为空,则代表该线程是空闲的
wait_for_proc_work = thread->transaction_stack == NULL &&
list_empty(&thread->todo);
if (thread->return_error != BR_OK && ptr < end) {
...
put_user(thread->return_error, (uint32_t __user *)ptr);
ptr += sizeof(uint32_t);
goto done; //发生error,则直接进入done
}
thread->looper |= BINDER_LOOPER_STATE_WAITING;
if (wait_for_proc_work)
proc->ready_threads++; //可用线程个数+1
binder_unlock(__func__);
if (wait_for_proc_work) {
if (non_block) {
...
} else
//当进程todo队列没有数据,则进入休眠等待状态
ret = wait_event_freezable_exclusive(proc->wait, binder_has_proc_work(proc, thread));
} else {
if (non_block) {
...
} else
//当线程todo队列没有数据,则进入休眠等待状态
ret = wait_event_freezable(thread->wait, binder_has_thread_work(thread));
}
binder_lock(__func__);
if (wait_for_proc_work)
proc->ready_threads--; //可用线程个数-1
thread->looper &= ~BINDER_LOOPER_STATE_WAITING;
if (ret)
return ret; //对于非阻塞的调用,直接返回
while (1) {
uint32_t cmd;
struct binder_transaction_data tr;
struct binder_work *w;
struct binder_transaction *t = NULL;
//先考虑从线程todo队列获取事务数据
if (!list_empty(&thread->todo)) {
w = list_first_entry(&thread->todo, struct binder_work, entry);
//线程todo队列没有数据, 则从进程todo对获取事务数据
} else if (!list_empty(&proc->todo) && wait_for_proc_work) {
w = list_first_entry(&proc->todo, struct binder_work, entry);
} else {
... //没有数据,则返回retry
}
switch (w->type) {
case BINDER_WORK_TRANSACTION: ... break;
case BINDER_WORK_TRANSACTION_COMPLETE:... break;
case BINDER_WORK_NODE: ... break;
case BINDER_WORK_DEAD_BINDER:
case BINDER_WORK_DEAD_BINDER_AND_CLEAR:
case BINDER_WORK_CLEAR_DEATH_NOTIFICATION:
struct binder_ref_death *death;
uint32_t cmd;
death = container_of(w, struct binder_ref_death, work);
if (w->type == BINDER_WORK_CLEAR_DEATH_NOTIFICATION)
cmd = BR_CLEAR_DEATH_NOTIFICATION_DONE;
else
cmd = BR_DEAD_BINDER;
put_user(cmd, (uint32_t __user *)ptr;
ptr += sizeof(uint32_t);
put_user(death->cookie, (void * __user *)ptr);
ptr += sizeof(void *);
...
if (cmd == BR_DEAD_BINDER)
goto done; //Binder驱动向client端发送死亡通知,则进入done
break;
}
if (!t)
continue; //只有BINDER_WORK_TRANSACTION命令才能继续往下执行
...
break;
}
done:
*consumed = ptr - buffer;
//创建线程的条件
if (proc->requested_threads + proc->ready_threads == 0 &&
proc->requested_threads_started < proc->max_threads &&
(thread->looper & (BINDER_LOOPER_STATE_REGISTERED |
BINDER_LOOPER_STATE_ENTERED))) {
proc->requested_threads++;
// 生成BR_SPAWN_LOOPER命令,用于创建新的线程
put_user(BR_SPAWN_LOOPER, (uint32_t __user *)buffer);
}
return 0;
}
当发生以下3种情况之一,便会进入done:
- 当前线程的return_error发生error的情况;
- 当Binder驱动向client端发送死亡通知的情况;
- 当类型为BINDER_WORK_TRANSACTION(即收到命令是BC_TRANSACTION或BC_REPLY)的情况;
任何一个Binder线程当同时满足以下条件,则会生成用于创建新线程的BR_SPAWN_LOOPER命令:
- 当前进程没有空闲可用的binder线程,即ready_threads = 0;(线程进入休眠状态的个数就是空闲线程数)
- 当前进程已启动线程个数小于最大上限(默认15);
- 当前线程已接收到BC_ENTER_LOOPER或者BC_REGISTER_LOOPER命令,即当前处于BINDER_LOOPER_STATE_REGISTERED或者BINDER_LOOPER_STATE_ENTERED状态。
从system_server的binder线程一直的执行流: IPC.joinThreadPool –> IPC.getAndExecuteCommand() -> IPC.talkWithDriver() ,但talkWithDriver收到事务之后, 便进入IPC.executeCommand(), 接下来,从executeCommand说起.
status_t IPCThreadState::executeCommand(int32_t cmd)
{
status_t result = NO_ERROR;
switch ((uint32_t)cmd) {
...
case BR_SPAWN_LOOPER:
//创建新的binder线程 【见小节2.3】
mProcess->spawnPooledThread(false);
break;
...
}
return result;
}
Binder主线程的创建是在其所在进程创建的过程一起创建的,后面再创建的普通binder线程是由spawnPooledThread(false)方法所创建的。
思考:
每个进程的binder线程池的线程个数上限为15,该上限不统计通过BC_ENTER_LOOPER命令创建的binder主线程, 只计算BC_REGISTER_LOOPER命令创建的线程。
某个进程的主线程执行如下方法,那么该进程可创建的binder线程个数上限是多少呢?
ProcessState::self()->setThreadPoolMaxThreadCount(6); // 6个线程
ProcessState::self()->startThreadPool(); // 1个线程
IPCThread::self()->joinThreadPool(); // 1个线程
首先由Binder驱动创建的的binder线程个数上限为6个,通过startThreadPool()创建的主线程不算在最大线程上限,最后一句是将当前线程成为binder线程,所以说可创建的binder线程个数上限为8。
Binder线程创建 小结
image每次由Zygote fork出新进程的过程中,伴随着创建binder线程池,调用spawnPooledThread来创建binder主线程。当线程执行binder_thread_read的过程中,发现当前没有空闲线程,没有请求创建线程,且没有达到上限,则创建新的binder线程。
Binder系统中可分为3类binder线程:
- Binder主线程:进程创建过程会调用startThreadPool()过程中再进入spawnPooledThread(true),来创建Binder主线程。编号从1开始,也就是意味着binder主线程名为binder_1,并且主线程是不会退出的。
- Binder普通线程:是由Binder Driver来根据是否有空闲的binder线程来决定是否创建binder线程,回调spawnPooledThread(false) ,isMain=false,该线程名格式为binder_x。
- Binder其他线程:其他线程是指并没有调用spawnPooledThread方法,而是直接调用IPC.joinThreadPool(),将当前线程直接加入binder线程队列。例如: mediaserver和servicemanager的主线程都是binder线程,但system_server的主线程并非binder线程。
六、Binder IPC 的一般调用流程。
Binder通信架构上图为以IActivityManager.startService为例,Binder IPC 跨进程通信的流程图。
Client端 BinderProxy.transact -> Server端 Binder.onTransact Binder,为一个通用的调用过程。
6.1、发起端线程向Binder Driver发起binder ioctl请求后, 便采用环不断talkWithDriver,此时该线程处于阻塞状态, 直到收到如下BR_XXX命令才会结束该过程.
- BR_TRANSACTION_COMPLETE: oneway模式下,收到该命令则退出
- BR_REPLY: 非oneway模式下,收到该命令才退出;
- BR_DEAD_REPLY: 目标进程/线程/binder实体为空, 以及释放正在等待reply的binder thread或者binder buffer;
- BR_FAILED_REPLY: 情况较多,比如非法handle, 错误事务栈, security, 内存不足, buffer不足, 数据拷贝失败, 节点创建失败, 各种不匹配等问题
- BR_ACQUIRE_RESULT: 目前未使用的协议;
6.2、目标Binder线程创建后, 便进入joinThreadPool()方法, 采用循环不断地循环执行getAndExecuteCommand()方法, 当bwr的读写buffer都没有数据时,则阻塞在binder_thread_read的wait_event过程. 另外,正常情况下binder线程一旦创建则不会退出
七、参考文章
https://www.jianshu.com/p/1bef7e865498
http://gityuan.com/2015/11/21/binder-framework/#getiservicemanager
网友评论