Android4 0 3 显示系统深入理解

分享一下我老师大神的人工智能教程!零基础,通俗易懂!http://blog.csdn.net/jiangjunshow

也欢迎大家转载本篇文章。分享知识,造福人民,实现我们中华民族伟大复兴!

                 

1. 简介      

        网上已经有很多兄弟对Android的显示系统做了深入解剖,很是佩服。可最近小弟在研究Android4.0时发现出入比较大,也许是Android4.0的修改比较多吧!因为小弟没有看Android4.0以前的代码。

       面对这么复杂一个Android显示系统,如何入手呢? 根据以前的经验,不管它有多么复杂,其功能不就是以下三步曲吗?

  1)显示系统的创建及初始化

       2)画图

       3)销毁

       哪我的分析就从显示系统的创建及初始化开始吧!由于小弟对Java没有什么研究兴趣,所有重点就分析Native部分。当然Native的入口就在android_view_Surface.cpp中,此文件主要包含以下两部分给Java层调用:

       1)gSurfaceSessionMethods: 操作SurfaceSession的方法

       2)gSurfaceMethods:操作Surface的方法

2. android_view_Surface.cpp

 

2.1 SurfaceSession操作方法

 

static JNINativeMethod gSurfaceSessionMethods[] = {      {"init",     "()V",  (void*)SurfaceSession_init }, //创建SurfaceComposerClient     {"destroy""()V",  (void*)SurfaceSession_destroy }, //直接销毁SurfaceComposerClient     {"kill",     "()V",  (void*)SurfaceSession_kill },//先clear,再销毁SurfaceComposerClient};  



 2.1.1 SurfaceSession_init

        其功能如下:

        1)创建SurfaceComposerClient对象

        2)调用SurfaceComposerClient::onFirstRef方法

        现在已经进入到SurfaceComposerClient的地盘,根据其名字含义,它应该是一个进行Surface合成的客户端,通过它发命令给SurfaceFlinger来进行需要的操作。其初始化流程如下图所示:

2.1.2 SurfaceComposerClient.cpp中的宝贝

        为了方便后面的理解,先看看SurfaceComposerClient中有些什么宝贝来完成这个任务。在其中定义了如下几个类:

2.1.2.1 ComposerService(获取SurfaceFlinger服务)

        一看到名字为Service,应该是用于从SurfaceFlinger中获取Service以建立连接关系<它是一个单实例,一个进程有且只有一个实例对象>,然后供后面进行相关的操作。其构造函数代码如下:      

 

class ComposerService : public Singleton<ComposerService>{    //实质为BpSurfaceComposer,通过它与SurfaceFlinger进行通信,    //BnSurfaceComposer是SurfaceFlinger基类中的一个    sp<ISurfaceComposer> mComposerService;    //实质为BpMemoryHeap,它在SurfaceFlinger中对应为管理一个4096字节的    //一个MemoryHeapBase对象,在SurfaceFlinger::readyToRun中创建    sp<IMemoryHeap> mServerCblkMemory;        //为MemoryHeapBase管理的内存在用户空间的基地址,通过mmap而来,    //具体见MemoryHeapBase::mapfd    surface_flinger_cblk_t volatile* mServerCblk;    ComposerService();    friend class Singleton<ComposerService>;public:    static sp<ISurfaceComposer> getComposerService();    static surface_flinger_cblk_t const volatile * getControlBlock();};ComposerService::ComposerService(): Singleton<ComposerService>() {    const String16 name("SurfaceFlinger");    //获取SurfaceFlinger服务,即BpSurfaceComposer对象    while (getService(name, &mComposerService) != NO_ERROR) {        usleep(250000);    }    //获取共享内存块    mServerCblkMemory = mComposerService->getCblk();    //获取共享内存块基地址    mServerCblk = static_cast<surface_flinger_cblk_t volatile *>(            mServerCblkMemory->getBase());}

      由此可见,ComposerService主要是获取SurfaceFlinger服务、获取在SurfaceFlinger::readyToRun中创建的共享内存块及其基地址。在Client中,谁要想与SurfaceFlinger通信,需要通过接口getComposerService来获取此BpSurfaceComposer。

     此ComposerService是在调用ComposerService::getInstance时进行有且只有一个的实例化,因为前面讲过,它是一个单实例。

 

2.1.2.2 Composer

      它也是一个单实例,管理并发送每个layer的ComposerState。其定义如下:

 

struct ComposerState {    sp<ISurfaceComposerClient> client;    layer_state_t state;    status_t    write(Parcel& output) const;    status_t    read(const Parcel& input);};class Composer : public Singleton<Composer>{    friend class Singleton<Composer>;    mutable Mutex               mLock;    //SurfaceComposerClient+SurfaceID与一个ComposerState一一对应    SortedVector<ComposerState> mStates;         int                         mOrientation;//整个屏幕的方向    Composer() : Singleton<Composer>(),        mOrientation(ISurfaceComposer::eOrientationUnchanged) { }    //通过BpSurfaceComposer把mStates发送给SurfaceFlinger处理    void closeGlobalTransactionImpl();    //根据client和id从mStates中获取对应原ComposerState,从而获取对应的layer_state_t    layer_state_t* getLayerStateLocked(            const sp<SurfaceComposerClient>& client, SurfaceID id);public:    //设置与client和id对应的layer_state_t中的位置信息,并保存在mStates中    status_t setPosition(const sp<SurfaceComposerClient>& client, SurfaceID id,            float x, float y);    //设置与client和id对应的layer_state_t中的Size信息,并保存在mStates中    status_t setSize(const sp<SurfaceComposerClient>& client, SurfaceID id,            uint32_t w, uint32_t h);    //设置与client和id对应的layer_state_t中的z-order信息,并保存在mStates中    status_t setLayer(const sp<SurfaceComposerClient>& client, SurfaceID id,            int32_t z);    //设置与client和id对应的layer_state_t中的flags信息,并保存在mStates中    status_t setFlags(const sp<SurfaceComposerClient>& client, SurfaceID id,            uint32_t flags, uint32_t mask);    //设置与client和id对应的layer_state_t中的透明区域信息,并保存在mStates中    status_t setTransparentRegionHint(            const sp<SurfaceComposerClient>& client, SurfaceID id,            const Region& transparentRegion);    //设置与client和id对应的layer_state_t中的alpha信息,并保存在mStates中    status_t setAlpha(const sp<SurfaceComposerClient>& client, SurfaceID id,            float alpha);    //设置与client和id对应的layer_state_t中的矩阵信息,并保存在mStates中    status_t setMatrix(const sp<SurfaceComposerClient>& client, SurfaceID id,            float dsdx, float dtdx, float dsdy, float dtdy);    //设置与client和id对应的layer_state_t中的位置信息,并保存在mStates中    status_t setFreezeTint(            const sp<SurfaceComposerClient>& client, SurfaceID id,            uint32_t tint);    //设置整个屏幕的方向    status_t setOrientation(int orientation);    //通过BpSurfaceComposer把mStates发送给SurfaceFlinger处理    static void closeGlobalTransaction() {        Composer::getInstance().closeGlobalTransactionImpl();    }}

      把上面的comments看完就明白了,Composer管理每个SurfaceComposerClient中的每一个Surface的状态,并记录在ComposerState的layer_state_t中,然后调用者可以调用其closeGlobalTransaction方法把这些mStates发送给SurfaceFlinger处理(处理函数为:SurfaceFlinger::setTransactionState)。

      谁来调用它的方法设置层的属性及发送mStates呢? -----答案是由SurfaceComposerClient来调用。

  2.1.2.3  SurfaceComposerClient

       前面介绍的两个类一个用于获取SurfaceFlinger服务;一个用于记录每个Layer的状态,且可按要求把这些CoposerState发送给SurfaceFlinger。这个类是不是来使用前面两个类提供的服务呢? --答案是肯定的。其定义及详细注释如下:

 

#define NUM_DISPLAY_MAX 4  //最多支持四个显示屏struct display_cblk_t //每个显示屏的配置参数{    uint16_t    w;    uint16_t    h;    uint8_t     format;    uint8_t     orientation;    uint8_t     reserved[2];    float       fps;    float       density;    float       xdpi;    float       ydpi;    uint32_t    pad[2];};//在SurfaceFlinger::readyToRun中创建的共享控制块struct surface_flinger_cblk_t   // 4KB max,管理系统中所有的显示屏{    uint8_t         connected; //每一个bit表示一个显示屏    uint8_t         reserved[3];    uint32_t        pad[7];    display_cblk_t  displays[NUM_DISPLAY_MAX];};class SurfaceComposerClient : public RefBase{    friend class Composer;public:             //获取Composer实例,并保存在mComposer中                SurfaceComposerClient();    virtual     ~SurfaceComposerClient();    //通过BpSurfaceComposerClient<mClient>创建Surface,    //同时通过ISurfaceComposerClient::surface_data_t返回SurfaceID.然后创建一个SurfaceControl    //并把返回的BpSurface和当前的SurfaceComposerClient保存在SurfaceControl中,    //然后返回此SurfaceControl    sp<SurfaceControl> createSurface(            const String8& name,// name of the surface            DisplayID display,  // Display to create this surface on            uint32_t w,         // width in pixel            uint32_t h,         // height in pixel            PixelFormat format, // pixel-format desired            uint32_t flags = 0  // usage flags    );    // Composer parameters <合成参数>    //所有的合成参数必须在一个transaction中被修改,多个surface可在一个transaction中被更新,    //所有的变化在关闭transaction时被一次性提交(通过调用closeGlobalTransaction来提交所有变化)。    //什么都没有做    static void openGlobalTransaction();            //通过调用Composer::closeGlobalTransaction(),    // 把Composer中记录的ComposerState(即mStates)发送给SurfaceFlinger    static void closeGlobalTransaction();        //什么都没做    static status_t freezeDisplay(DisplayID dpy, uint32_t flags = 0);            //什么都没做    static status_t unfreezeDisplay(DisplayID dpy, uint32_t flags = 0);    //把新的显示方向保存在Composer实例中    static int setOrientation(DisplayID dpy, int orientation, uint32_t flags);    //从surface_flinger_cblk_t.connected中获取显示屏个数    static ssize_t getNumberOfDisplays();    //获取显示屏的信息    static status_t getDisplayInfo(DisplayID dpy, DisplayInfo* info);    static ssize_t getDisplayWidth(DisplayID dpy);    static ssize_t getDisplayHeight(DisplayID dpy);    static ssize_t getDisplayOrientation(DisplayID dpy);        //通过注册,当Binder异常退出时,可以获得通知    status_t linkToComposerDeath(const sp<IBinder::DeathRecipient>& recipient,            void* cookie = NULL, uint32_t flags = 0);    //Start####: 以下函数都是把相应的修改状态记录在Composer的mStates中    //调用Composer::setFlags来设置对应(client+id)的layer状态〈即ComposerState中的layer_state_t〉    status_t    hide(SurfaceID id);    status_t    show(SurfaceID id, int32_t layer = -1);    status_t    freeze(SurfaceID id);    status_t    unfreeze(SurfaceID id);    status_t    setFlags(SurfaceID id, uint32_t flags, uint32_t mask);    //调用Composer::setTransparentRegionHint    status_t    setTransparentRegionHint(SurfaceID id, const Region& transparent);    //调用Composer::setLayer    status_t    setLayer(SurfaceID id, int32_t layer);    //调用Composer::setAlpha    status_t    setAlpha(SurfaceID id, float alpha=1.0f);    //调用Composer::setFreezeTint    status_t    setFreezeTint(SurfaceID id, uint32_t tint);    //调用Composer::setMatrix    status_t    setMatrix(SurfaceID id, float dsdx, float dtdx, float dsdy, float dtdy);    //调用Composer::setPosition    status_t    setPosition(SurfaceID id, float x, float y);    //调用Composer::setSize    status_t    setSize(SurfaceID id, uint32_t w, uint32_t h);    //End####:    status_t    destroySurface(SurfaceID sid);//通过BpSurfaceComposerClient销毁Surfaceprivate:    //通过BpSurfaceComposer从SurfaceFlinger获取BpSurfaceComposerClient,    //并把它保存在mClient中    virtual void onFirstRef();    Composer& getComposer();    mutable     Mutex                       mLock;                status_t                    mStatus;                //实质为BpSurfaceComposerClient,与SurfaceFlinger.cpp中的Client相对应                   sp<ISurfaceComposerClient>  mClient;                 //Composer实例                  Composer&                   mComposer;}

    其功能列表如下:

     1)获取BpSurfaceComposerClient(即mClient),在onFirstRef中实现

     2)通过BpSurfaceComposerClient(即mClient)创建和销毁Surface

     3)通过Composer来记录Surface和显示屏状态变化,及在Composer中通过BpSurfaceComposer把状态变化发给SurfaceFlinger处理

     至此,SurfaceComposerClient功能已经分析清楚。可是从这三个类中,我们已经看到三个Bp(BpSurfaceComposer,BpSurfaceComposerClient和BpSurface)及三个对应的接口。下面总结一下,每个接口的功能,在客户端由谁使用,在服务器端谁来实现。

 2.1.2.4 Surface相关接口总结

     

 

 2.2 Surface操作

其相关接口如下:

static JNINativeMethod gSurfaceMethods[] = {    {"nativeClassInit",     "()V",  (void*)nativeClassInit },    {"init",                "(Landroid/view/SurfaceSession;ILjava/lang/String;IIIII)V",  (void*)Surface_init },    {"init",                "(Landroid/os/Parcel;)V",  (void*)Surface_initParcel },    {"initFromSurfaceTexture", "(Landroid/graphics/SurfaceTexture;)V", (void*)Surface_initFromSurfaceTexture },    {"getIdentity",         "()I",  (void*)Surface_getIdentity },    {"destroy",             "()V",  (void*)Surface_destroy },    {"release",             "()V",  (void*)Surface_release },    {"copyFrom",            "(Landroid/view/Surface;)V",  (void*)Surface_copyFrom },    {"isValid",             "()Z",  (void*)Surface_isValid },    {"lockCanvasNative",    "(Landroid/graphics/Rect;)Landroid/graphics/Canvas;",  (void*)Surface_lockCanvas },    {"unlockCanvasAndPost", "(Landroid/graphics/Canvas;)V", (void*)Surface_unlockCanvasAndPost },    {"unlockCanvas",        "(Landroid/graphics/Canvas;)V", (void*)Surface_unlockCanvas },    {"openTransaction",     "()V",  (void*)Surface_openTransaction },    {"closeTransaction",    "()V",  (void*)Surface_closeTransaction },    {"setOrientation",      "(III)V", (void*)Surface_setOrientation },    {"freezeDisplay",       "(I)V", (void*)Surface_freezeDisplay },    {"unfreezeDisplay",     "(I)V", (void*)Surface_unfreezeDisplay },    {"screenshot",          "(II)Landroid/graphics/Bitmap;", (void*)Surface_screenshotAll },    {"screenshot",          "(IIII)Landroid/graphics/Bitmap;", (void*)Surface_screenshot },    {"setLayer",            "(I)V", (void*)Surface_setLayer },    {"setPosition",         "(FF)V",(void*)Surface_setPosition },    {"setSize",             "(II)V",(void*)Surface_setSize },    {"hide",                "()V",  (void*)Surface_hide },    {"show",                "()V",  (void*)Surface_show },    {"freeze",              "()V",  (void*)Surface_freeze },    {"unfreeze",            "()V",  (void*)Surface_unfreeze },    {"setFlags",            "(II)V",(void*)Surface_setFlags },    {"setTransparentRegionHint","(Landroid/graphics/Region;)V", (void*)Surface_setTransparentRegion },    {"setAlpha",            "(F)V", (void*)Surface_setAlpha },    {"setMatrix",           "(FFFF)V",  (void*)Surface_setMatrix },    {"setFreezeTint",       "(I)V",  (void*)Surface_setFreezeTint },    {"readFromParcel",      "(Landroid/os/Parcel;)V", (void*)Surface_readFromParcel },    {"writeToParcel",       "(Landroid/os/Parcel;I)V", (void*)Surface_writeToParcel },};

2.2.1 Surface_init调用流程

     在SurfaceFlinger端创建BSurface,在客户端返回SurfaceControl,同时在SurfaceControl中拥有了BpSurface用于与BSurface交互。

 

2.2.1.1 调用流程分析

     BpSurfaceComposerClient->createSurface返回BpSurface。且通过参数返回ISurfaceComposerClient::surface_data_t,其定义如下:

     其中token在SurfaceComposerClient的函数参数中,对应于SurfaceID。即在客户端,它就是SurfaceID。

     token: 加入到Client::mLayers中的序号,在Client中单调递增,初始值为:1,一个Layer创建一个BSurface

     identity: LayerBaseClient中的mIdentity,在所有的Layer中单调递增,初始值为:1

    struct surface_data_t {        int32_t             token;  //加入到Client::mLayers中的序号,在Client中单调递增,初始值为:1        int32_t             identity; //LayerBaseClient中的mIdentity,在所有的Layer中单调递增,初始值为:1        status_t readFromParcel(const Parcel& parcel);        status_t writeToParcel(Parcel* parcel) const;    };
2.2.1.2 创建真正的Surface

      在Layer::createSurface中创建真正的BSurface,在SurfaceFlinger::createSurface中调用layer->getSurface时创建的。此BSurface定义如下:

sp<ISurface> Layer::createSurface(){    class BSurface : public BnSurface, public LayerCleaner {        wp<const Layer> mOwner;        virtual sp<ISurfaceTexture> getSurfaceTexture() const { //实现了ISurface的接口            sp<ISurfaceTexture> res;            sp<const Layer> that( mOwner.promote() );            if (that != NULL) {                res = that->mSurfaceTexture;            }            return res;        }    public:        BSurface(const sp<SurfaceFlinger>& flinger,                const sp<Layer>& layer)            : LayerCleaner(flinger, layer), mOwner(layer) { }    };    sp<ISurface> sur(new BSurface(mFlinger, this));    return sur;}

         在此BSurface中实现了ISurface的接口getSurfaceTexture,在此接口中返回Layer::mSurfaceTexture(类型为:SurfaceTextureLayer,它才是真正操作内存的东东),此成员在Layer::onFirstRef中创建,SurfaceTextureLayer是SurfaceTexture的派生类,代码如下:

void Layer::onFirstRef(){    LayerBaseClient::onFirstRef();    struct FrameQueuedListener : public SurfaceTexture::FrameAvailableListener {        FrameQueuedListener(Layer* layer) : mLayer(layer) { }    private:        wp<Layer> mLayer;        virtual void onFrameAvailable() {            sp<Layer> that(mLayer.promote());            if (that != 0) {                that->onFrameQueued();            }        }    };    mSurfaceTexture = new SurfaceTextureLayer(mTextureName, this); //创建Layer中的mSurfaceTexture    mSurfaceTexture->setFrameAvailableListener(new FrameQueuedListener(this));    mSurfaceTexture->setSynchronousMode(true);    mSurfaceTexture->setBufferCountServer(2);}

2.2.1.3 不得不说的SurfaceControl

       本来Surface_init调用SurfaceComposerClient::createSurface创建一个Surface,可却返回了一个SurfaceControl,下面看看SurfaceCotrol到底做了些什么,以及如何做的?

        相关数据结构如下图所示:

 

 SurfaceControl定义如下:

 

class SurfaceControl : public RefBase{public:    // release surface data from java    void        clear();        //调用SurfaceComposerClient中对应方法,把对应信息保存在    //Composer的ComposerState中        status_t    setLayer(int32_t layer);    status_t    setPosition(int32_t x, int32_t y);    status_t    setSize(uint32_t w, uint32_t h);    status_t    hide();    status_t    show(int32_t layer = -1);    status_t    freeze();    status_t    unfreeze();    status_t    setFlags(uint32_t flags, uint32_t mask);    status_t    setTransparentRegionHint(const Region& transparent);    status_t    setAlpha(float alpha=1.0f);    status_t    setMatrix(float dsdx, float dtdx, float dsdy, float dtdy);    status_t    setFreezeTint(uint32_t tint);        //把SurfaceControl中的mSurface和mIdentity写入parcel    static status_t writeSurfaceToParcel(            const sp<SurfaceControl>& control, Parcel* parcel);        //以SurfaceControl为参数创建一个Surface返回,此Surface派生关系如下:    //class Surface : public SurfaceTextureClient    //class SurfaceTextureClient: public ANativeWindow, RefBase    //struct ANativeWindow    sp<Surface> getSurface() const;private:    SurfaceControl(            const sp<SurfaceComposerClient>& client,            const sp<ISurface>& surface,            const ISurfaceComposerClient::surface_data_t& data);    ~SurfaceControl();    void destroy();        sp<SurfaceComposerClient>   mClient;    sp<ISurface>                mSurface;    SurfaceID                   mToken;  //对应SurfaceID,在Client中单调递增    uint32_t                    mIdentity;  //Layer在系统中唯一的序列号,在系统中单调递增    mutable Mutex               mLock;        mutable sp<Surface>         mSurfaceData;}


       从其定义中可以看出,在getSurface中将有新花样,其它操作函数都是直接以mToken作为SurfaceID,直接调用SurfaceComposerClient中对应方法。 经过这样一分析,SurfaceControl也没什么神秘的了。但它的getSurface到有点神秘。

 

 2.2.2 getSurface流程

      getSurface在客户端返回Surface(派生于SurfaceTextureClient),并在Surface的mSurfaceTexture域中保存了BpSurfaceTexture。

      前面Surface初始化之后,就可以getSurface了。getSurface流程如下图所示:

           有了Surface,且在Surface中又有了BpSurfaceTexture,下一步就操作GraphicBuffer了。

 

 3. 画图流程

对于画图流程,可以从ViewRootImpl(ViewRootImpl.java)的draw函数看起,在画图之间,它要调用java层的surface.lockCanvas,画完图之后调用surface.unlockCanvasAndPost来提交显示。

surface.lockCanvas->

lockCanvasNative(Java)->

(C++)Surface_lockCanvas<android_view_Surface.cpp>

surface.unlockCanvasAndPost(Java)->

(C++)Surface_unlockCanvasAndPost<android_view_Surface.cpp>

本章主要分析这两个函数到底做了些什么>

 3.1 Surface_lockCanvas

       Android图形系统中一个重要的概念是surface。View及其子类(如TextView, Button)要画在surface上。每个surface创建一个Canvas对象(但属性时常改变),用来管理view在surface上的绘图操作,如画点画线。每个canvas对象对应一个bitmap,存储画在surface上的内容。

3.1.1 相关数据结构定义
 3.1.1.1 ANativeWindow_Buffer
typedef struct ANativeWindow_Buffer {    // The number of pixels that are show horizontally.    int32_t width;    // The number of pixels that are shown vertically.    int32_t height;    // The number of *pixels* that a line in the buffer takes in    // memory.  This may be >= width.    int32_t stride;    // The format of the buffer.  One of WINDOW_FORMAT_*    int32_t format;    // The actual bits.    void* bits;  //显示内存基地址,通过服务器端fd通过flat_binder_object传给客户端, 然后客户端通过mmap获取。        // Do not touch.    uint32_t reserved[6];} ANativeWindow_Buffer;
3.1.1.2 SurfaceInfo
    struct SurfaceInfo {        uint32_t    w;        uint32_t    h;        uint32_t    s;        uint32_t    usage;        PixelFormat format;        void*       bits;//显示内存基地址,通过服务器端fd通过flat_binder_object传给客户端, 然后客户端通过mmap获取。        uint32_t    reserved[2];    };

3.1.1.3 二者对应关系
SurfaceInfo* other;ANativeWindow_Buffer outBuffer;other->w = uint32_t(outBuffer.width);other->h = uint32_t(outBuffer.height);other->s = uint32_t(outBuffer.stride);other->usage = GRALLOC_USAGE_SW_READ_OFTEN | GRALLOC_USAGE_SW_WRITE_OFTEN;other->format = uint32_t(outBuffer.format);other->bits = outBuffer.bits;

 3.1.1.4 GraphicBuffer

     在分析下面的流程时, 不得不对GraphicBuffer进行深入了解,特别是其Flattenable interface,这是实现画图buffer的关键。其相关定义如下:

typedef struct native_handle{    int version;        /* sizeof(native_handle_t) */    int numFds;         /* number of file-descriptors at &data[0] */    int numInts;        /* number of ints at &data[numFds] */    int data[0];        /* numFds + numInts ints */} native_handle_t;typedef const native_handle_t* buffer_handle_t;class GraphicBuffer    : public EGLNativeBase<        ANativeWindowBuffer,        GraphicBuffer,         LightRefBase<GraphicBuffer> >, public Flattenable{    ...    // Flattenable interface    size_t getFlattenedSize() const;    size_t getFdCount() const;    status_t flatten(void* buffer, size_t size,            int fds[], size_t count) const;    status_t unflatten(void const* buffer, size_t size,            int fds[], size_t count);    ...    buffer_handle_t handle; //定义于基类ANativeWindowBuffer中};
3.1.1.5 Flattenable interface

下面看看每个Flattenable interface是如何实现的:

3.1.1.5.1 getFlattenedSize
size_t GraphicBuffer::getFlattenedSize() const {    return (8 + (handle ? handle->numInts : 0))*sizeof(int);}
3.1.1.5.2 getFdCount
size_t GraphicBuffer::getFdCount() const {    return handle ? handle->numFds : 0;}
3.1.1.5.3 flatten
status_t GraphicBuffer::flatten(void* buffer, size_t size,        int fds[], size_t count) const{    size_t sizeNeeded = GraphicBuffer::getFlattenedSize();    if (size < sizeNeeded) return NO_MEMORY;    size_t fdCountNeeded = GraphicBuffer::getFdCount();    if (count < fdCountNeeded) return NO_MEMORY;    int* buf = static_cast<int*>(buffer);    buf[0] = 'GBFR';    buf[1] = width;    buf[2] = height;    buf[3] = stride;    buf[4] = format;    buf[5] = usage;    buf[6] = 0;    buf[7] = 0;    if (handle) {        buf[6] = handle->numFds;        buf[7] = handle->numInts;        native_handle_t const* const h = handle;        memcpy(fds,     h->data,             h->numFds*sizeof(int));        memcpy(&buf[8], h->data + h->numFds, h->numInts*sizeof(int));    }    return NO_ERROR;}


     把handle中的numFds拷贝到fds中,把handle中的numInts拷贝到buffer中。

 3.1.1.5.4 unflatten 
status_t GraphicBuffer::unflatten(void const* buffer, size_t size,        int fds[], size_t count){    if (size < 8*sizeof(int)) return NO_MEMORY;    int const* buf = static_cast<int const*>(buffer);    if (buf[0] != 'GBFR') return BAD_TYPE;    const size_t numFds  = buf[6];    const size_t numInts = buf[7];    const size_t sizeNeeded = (8 + numInts) * sizeof(int);    if (size < sizeNeeded) return NO_MEMORY;    size_t fdCountNeeded = 0;    if (count < fdCountNeeded) return NO_MEMORY;    if (handle) {        // free previous handle if any        free_handle();    }    if (numFds || numInts) {        width  = buf[1];        height = buf[2];        stride = buf[3];        format = buf[4];        usage  = buf[5];        native_handle* h = native_handle_create(numFds, numInts);        memcpy(h->data,          fds,     numFds*sizeof(int));        memcpy(h->data + numFds, &buf[8], numInts*sizeof(int));        handle = h;    } else {        width = height = stride = format = usage = 0;        handle = NULL;    }    mOwner = ownHandle;    if (handle != 0) {        mBufferMapper.registerBuffer(handle);    }    return NO_ERROR;}

       把width,height,stride,format和usage保存到成员变量中,并创建一个native_handle,然后把numFds和numInts拷贝到handle的data中。同时把此handle注册到mBufferMapper中,mBufferMapper的注册函数实现代码如下:

status_t GraphicBufferMapper::registerBuffer(buffer_handle_t handle){    status_t err;    //gralloc_module_t const *mAllocMod;是一个硬件抽象层实现。通过hw_get_module(GRALLOC_HARDWARE_MODULE_ID, &module)方式获取    err = mAllocMod->registerBuffer(mAllocMod, handle);    LOGW_IF(err, "registerBuffer(%p) failed %d (%s)",            handle, err, strerror(-err));    return err;}

 

GraphicBufferMapper::GraphicBufferMapper()    : mAllocMod(0){    hw_module_t const* module;    int err = hw_get_module(GRALLOC_HARDWARE_MODULE_ID, &module);    LOGE_IF(err, "FATAL: can't find the %s module", GRALLOC_HARDWARE_MODULE_ID);    if (err == 0) {        mAllocMod = (gralloc_module_t const *)module;    }}
 3.1.1.5.4  GRALLOC_HARDWARE_MODULE_ID实例

      对于GRALLOC_HARDWARE_MODULE_ID,以hardware/msm7k/libgralloc/gralloc.cpp为例进行分析。其registerBuffer实现函数:gralloc_register_buffer(hardware/msm7k/libgralloc/mapper.cpp),其相关代码如下:

int gralloc_register_buffer(gralloc_module_t const* module,        buffer_handle_t handle){    if (private_handle_t::validate(handle) < 0)        return -EINVAL;    // if this handle was created in this process, then we keep it as is.    int err = 0;    private_handle_t* hnd = (private_handle_t*)handle;    if (hnd->pid != getpid()) {        hnd->base = NULL;        if (!(hnd->flags & private_handle_t::PRIV_FLAGS_USES_GPU)) {            void *vaddr;            err = gralloc_map(module, handle, &vaddr);        }    }    return err;}static int gralloc_map(gralloc_module_t const* module,        buffer_handle_t handle,        void** vaddr){    private_handle_t* hnd = (private_handle_t*)handle;    if (!(hnd->flags & private_handle_t::PRIV_FLAGS_FRAMEBUFFER)) {        size_t size = hnd->size;#if PMEM_HACK        size += hnd->offset;#endif        void* mappedAddress = mmap(0, size,                PROT_READ|PROT_WRITE, MAP_SHARED, hnd->fd, 0);        if (mappedAddress == MAP_FAILED) {            LOGE("Could not mmap handle %p, fd=%d (%s)",                    handle, hnd->fd, strerror(errno));            hnd->base = 0;            return -errno;        }        hnd->base = intptr_t(mappedAddress) + hnd->offset;        //LOGD("gralloc_map() succeeded fd=%d, off=%d, size=%d, vaddr=%p",         //        hnd->fd, hnd->offset, hnd->size, mappedAddress);    }    *vaddr = (void*)hnd->base;    return 0;}

    从gralloc_map可以看出,这个registerBuffer主要做了一件事:

    1)根据handle中传过来的fd和size进行mmap映射(把kernel中的内存映射到用户空间),映射之后的地址再加上hnd->offset便获得hnd->base供后面使用。

      从这里可以初步看出,这个图形buffer数据并不是真正的从client传递到server,而是在lock是从server把fd传递给client,由客户端进行mmap,然后进行使用。关于这个是怎么实现的,后面将详细分析其实现过程。

    对于如何从native_handle转换为private_handle_t,且在private_handle_t中可以获取fd和offset? 看一下其数据结构和flatten的实现方式就可以得知:
native_handle:

typedef struct native_handle{    int version;        /* sizeof(native_handle_t) */    int numFds;         /* number of file-descriptors at &data[0] */    int numInts;        /* number of ints at &data[numFds] */    int data[0];        /* numFds + numInts ints */} native_handle_t;

     这个data[0]是关键,虽然分配了哪么多buffer,但实质上native_handle只占了3个int.其它的数据由包含它的数据结构来解析。

private_handle_t:

struct private_handle_t {    native_handle_t nativeHandle;#endif        enum {        PRIV_FLAGS_FRAMEBUFFER = 0x00000001,        PRIV_FLAGS_USES_PMEM   = 0x00000002,        PRIV_FLAGS_USES_GPU    = 0x00000004,    };    // file-descriptors    int     fd;    // ints    int     magic;    int     flags;    int     size;    int     offset;    int     gpu_fd; // stored as an int, b/c we don't want it marshalled    // FIXME: the attributes below should be out-of-line    int     base;    int     map_offset;    int     pid;#ifdef __cplusplus    static const int sNumInts = 8//numInts在这儿明确指定    static const int sNumFds = 1;   //numFds在这儿明确指定    static const int sMagic = 'gmsm';    private_handle_t(int fd, int size, int flags) :        fd(fd), magic(sMagic), flags(flags), size(size), offset(0),        base(0), pid(getpid())    {        version = sizeof(native_handle);        numInts = sNumInts;        numFds = sNumFds;    }    ~private_handle_t() {        magic = 0;    }    static int validate(const native_handle* h) {        const private_handle_t* hnd = (const private_handle_t*)h;        if (!h || h->version != sizeof(native_handle) ||                h->numInts != sNumInts || h->numFds != sNumFds ||                hnd->magic != sMagic)         {            LOGE("invalid gralloc handle (at %p)", h);            return -EINVAL;        }        return 0;    }#endif}
3.1.2 Surface_lockCanvas执行流程

查看高清大图


3.1.3 Surface_lockCanvas总结 

功能:Surface_lockCanvas获取显示buffer在本进程用户空间的地址,并据此创建一个SkBitmap给Java使用。
关键技术:BINDER_TYPE_FD类型的Binder、mmap、gralloc硬件抽象层

 

3.1.4 SurfaceTexture::dequeueBuffer如何创建GraphicBuffer

       相关代码如下:

    const sp<GraphicBuffer>& buffer(mSlots[buf].mGraphicBuffer);    if ((buffer == NULL) ||        (uint32_t(buffer->width)  != w) ||        (uint32_t(buffer->height) != h) ||        (uint32_t(buffer->format) != format) ||        ((uint32_t(buffer->usage) & usage) != usage))    {        usage |= GraphicBuffer::USAGE_HW_TEXTURE;        status_t error;        sp<GraphicBuffer> graphicBuffer( //创建GraphicBuffer                mGraphicBufferAlloc->createGraphicBuffer(                        w, h, format, usage, &error));        if (graphicBuffer == 0) {            ST_LOGE("dequeueBuffer: SurfaceComposer::createGraphicBuffer "                    "failed");            return error;        }        if (updateFormat) {            mPixelFormat = format;        }        mSlots[buf].mGraphicBuffer = graphicBuffer;        mSlots[buf].mRequestBufferCalled = false;        if (mSlots[buf].mEglImage != EGL_NO_IMAGE_KHR) {            eglDestroyImageKHR(mSlots[buf].mEglDisplay, mSlots[buf].mEglImage);            mSlots[buf].mEglImage = EGL_NO_IMAGE_KHR;            mSlots[buf].mEglDisplay = EGL_NO_DISPLAY;        }        returnFlags |= ISurfaceTexture::BUFFER_NEEDS_REALLOCATION;    }


     mGraphicBufferAlloc也是通过调用BpSurfaceComposer->createGraphicBufferAlloc而获取,它对应的服务器为SufaceFlinger中的GraphicBufferAlloc。

     mGraphicBufferAlloc实质为一个BpGraphicBufferAlloc,它真正创建GraphicBuffer的代码位于GraphicBufferAlloc::createGraphicBuffer中。代码关键调用流程如下:

     new GraphicBuffer(w, h, format, usage)->

       initSize(w, h, reqFormat, reqUsage)->

         GraphicBufferAllocator::get()->

            allocator.alloc(w, h, format, reqUsage, &handle, &stride)->

            返回handle,此handle为ANativeWindowBuffer成员,类型为native_handle。

              GraphicBufferAllocator::alloc->

                mAllocDev->alloc->

                 mAllocDev类型为alloc_device_t,它通过gralloc_open向

                 GRALLOC_HARDWARE_MODULE_ID获取,根据上面的实例msm7k,

                 它最终执行gralloc_device_open而获取gralloc_context_t.device.common,

                  alloc的实现函数为gralloc_alloc.

                   gralloc_alloc->

                    gralloc_alloc_buffer->

                     1)获取GPU内存(调用SimpleBestFitAllocator::allocate进行分配)

                     2)fd = open("/dev/null", O_RDONLY)获取fd

                     3)根据fd、size和flags创建private_handle_t,其相关代码如下:

 

        private_handle_t* hnd = new private_handle_t(fd, size, flags);        if (base == NULL) {...            }        } else {            private_module_t* m = reinterpret_cast<private_module_t*>(                    dev->common.module);            hnd->offset = offset;            hnd->base = int(base)+offset;            hnd->gpu_fd = gpu_fd;            hnd->map_offset = m->fb_map_offset;            *pHandle = hnd;        }


3.2 Surface_unlockCanvasAndPost

 

 

 

 

 

 

 

 

 

 

           

给我老师的人工智能教程打call!http://blog.csdn.net/jiangjunshow
这里写图片描述
  • 0
    点赞
  • 1
    收藏
    觉得还不错? 一键收藏
  • 0
    评论

“相关推荐”对你有帮助么?

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值