BufferQueue笔记

应用程序的一个窗口通过ViewRootImpl的relayout来向SurfaceFlinger请求创建Surface时，会在SurfaceFlinger这边创建一个Layer对象。Layer对象创建时会创建一个Buffer队列。因为是triple buffer，所以MaxDequeuedBufferCount被设置为2。

void Layer::onFirstRef() {
    // Creates a custom BufferQueue for SurfaceFlingerConsumer to use
    sp<IGraphicBufferProducer> producer;
    sp<IGraphicBufferConsumer> consumer;
    BufferQueue::createBufferQueue(&producer, &consumer);
    mProducer = new MonitoredProducer(producer, mFlinger);
    mSurfaceFlingerConsumer = new SurfaceFlingerConsumer(consumer, mTextureName);

#ifdef TARGET_DISABLE_TRIPLE_BUFFERING
#warning "disabling triple buffering"
#else
    mProducer->setMaxDequeuedBufferCount(2);
#endif

    const sp<const DisplayDevice> hw(mFlinger->getDefaultDisplayDevice());
    updateTransformHint(hw);
}

void BufferQueue::createBufferQueue(sp<IGraphicBufferProducer>* outProducer,
        sp<IGraphicBufferConsumer>* outConsumer,
        const sp<IGraphicBufferAlloc>& allocator) {
    ....
    sp<BufferQueueCore> core(new BufferQueueCore(allocator));
    LOG_ALWAYS_FATAL_IF(core == NULL,
            "BufferQueue: failed to create BufferQueueCore");

    sp<IGraphicBufferProducer> producer(new BufferQueueProducer(core));
    LOG_ALWAYS_FATAL_IF(producer == NULL,
            "BufferQueue: failed to create BufferQueueProducer");

    sp<IGraphicBufferConsumer> consumer(new BufferQueueConsumer(core));
    LOG_ALWAYS_FATAL_IF(consumer == NULL,
            "BufferQueue: failed to create BufferQueueConsumer");

    *outProducer = producer;
    *outConsumer = consumer;
}

};

BufferQueueCore的构造函数请求SurfaceFlinger创建一个GraphicBufferAlloc

BufferQueueCore::BufferQueueCore(const sp<IGraphicBufferAlloc>& allocator) :
    mAllocator(allocator),
......
    mMaxAcquiredBufferCount(1),
    mMaxDequeuedBufferCount(1),
.....
{
    if (allocator == NULL) {
        sp<ISurfaceComposer> composer(ComposerService::getComposerService());
        mAllocator = composer->createGraphicBufferAlloc();
        if (mAllocator == NULL) {
            BQ_LOGE("createGraphicBufferAlloc failed");
        }
    }

    int numStartingBuffers = getMaxBufferCountLocked(); //
    for (int s = 0; s < numStartingBuffers; s++) {
        mFreeSlots.insert(s);
    }
    for (int s = numStartingBuffers; s < BufferQueueDefs::NUM_BUFFER_SLOTS;
            s++) {
        mUnusedSlots.push_front(s);
    }
}

int BufferQueueCore::getMaxBufferCountLocked() const {
    int maxBufferCount = mMaxAcquiredBufferCount + mMaxDequeuedBufferCount +
            ((mAsyncMode || mDequeueBufferCannotBlock) ? 1 : 0);

    // limit maxBufferCount by mMaxBufferCount always
    maxBufferCount = std::min(mMaxBufferCount, maxBufferCount);

    return maxBufferCount;
}

sp<IGraphicBufferAlloc> SurfaceFlinger::createGraphicBufferAlloc()
{
    sp<GraphicBufferAlloc> gba(new GraphicBufferAlloc());
    return gba;
}

BufferQueueCore里面有个SlotsType的变量mSlots，其实是一个BufferSlot数组。一个BufferSlot包含了一块buffer所对应GraphicBuffer，还有buffer的状态。

SlotsType BufferQueueCore.mSlots: 

enum { NUM_BUFFER_SLOTS = 64 };

typedef BufferSlot SlotsType[NUM_BUFFER_SLOTS];

struct BufferSlot {

    // mGraphicBuffer points to the buffer allocated for this slot or is NULL
    // if no buffer has been allocated.
    sp<GraphicBuffer> mGraphicBuffer;

    // mEglDisplay is the EGLDisplay used to create EGLSyncKHR objects.
    EGLDisplay mEglDisplay;

    // mBufferState is the current state of this buffer slot.
    BufferState mBufferState;

    // mRequestBufferCalled is used for validating that the producer did
    // call requestBuffer() when told to do so. Technically this is not
    // needed but useful for debugging and catching producer bugs.
    bool mRequestBufferCalled;

    // mFrameNumber is the number of the queued frame for this slot.  This
    // is used to dequeue buffers in LRU order (useful because buffers
    // may be released before their release fence is signaled).
    uint64_t mFrameNumber;

    // mEglFence is the EGL sync object that must signal before the buffer
    // associated with this buffer slot may be dequeued. It is initialized
    // to EGL_NO_SYNC_KHR when the buffer is created and may be set to a
    // new sync object in releaseBuffer.  (This is deprecated in favor of
    // mFence, below.)
    EGLSyncKHR mEglFence;

    // mFence is a fence which will signal when work initiated by the
    // previous owner of the buffer is finished. When the buffer is FREE,
    // the fence indicates when the consumer has finished reading
    // from the buffer, or when the producer has finished writing if it
    // called cancelBuffer after queueing some writes. When the buffer is
    // QUEUED, it indicates when the producer has finished filling the
    // buffer. When the buffer is DEQUEUED or ACQUIRED, the fence has been
    // passed to the consumer or producer along with ownership of the
    // buffer, and mFence is set to NO_FENCE.
    sp<Fence> mFence;

    // Indicates whether this buffer has been seen by a consumer yet
    bool mAcquireCalled;

    // Indicates whether the buffer was re-allocated without notifying the
    // producer. If so, it needs to set the BUFFER_NEEDS_REALLOCATION flag when
    // dequeued to prevent the producer from using a stale cached buffer.
    bool mNeedsReallocation;
};

struct BufferState {

    uint32_t mDequeueCount;
    uint32_t mQueueCount;
    uint32_t mAcquireCount;
    bool mShared;

    // A buffer can be in one of five states, represented as below:
    //
    //         | mShared | mDequeueCount | mQueueCount | mAcquireCount |
    // --------|---------|---------------|-------------|---------------|
    // FREE    |  false  |       0       |      0      |       0       |
    // DEQUEUED|  false  |       1       |      0      |       0       |
    // QUEUED  |  false  |       0       |      1      |       0       |
    // ACQUIRED|  false  |       0       |      0      |       1       |
    // SHARED  |  true   |      any      |     any     |      any      |
}

一块Buffer状态转移：绘制图像时，Surface会从BufferQueue中dequeue中一块buffer出来进行绘制，最多可以dequeue两块，绘制完成后queue到bufferQueue中，然后BufferConsumer（SurfaceFlinger）可以从BufferQueue中acquire一块buffer进行多个窗口layer的合成渲染，然后显示到屏幕上。

buffer_state.png

以硬件渲染UI为例，分析一下BufferQueue的使用过程

ViewRootImpl.performTraversal会用Surface初始化ThreadedRenderer

hwInitialized = mAttachInfo.mHardwareRenderer.initialize(mSurface);

initialize方法内部调用native方法。

static void android_view_ThreadedRenderer_initialize(JNIEnv* env, jobject clazz,
        jlong proxyPtr, jobject jsurface) {
    RenderProxy* proxy = reinterpret_cast<RenderProxy*>(proxyPtr);
    sp<Surface> surface = android_view_Surface_getSurface(env, jsurface);
    proxy->initialize(surface);
}

最后调用CanvasContext.initialize

void CanvasContext::initialize(Surface* surface) {
    setSurface(surface);
}

void CanvasContext::setSurface(Surface* surface) {
    mNativeSurface = surface;

    if (surface) {
        mEglSurface = mEglManager.createSurface(surface);
    }

EGLSurface EglManager::createSurface(EGLNativeWindowType window) {
    EGLSurface surface = eglCreateWindowSurface(mEglDisplay, mEglConfig, window, nullptr);

    return surface;
}

/frameworks/native/opengl/libs/EGL/eglApi.cpp

EGLSurface eglCreateWindowSurface(  EGLDisplay dpy, EGLConfig config,
                                    NativeWindowType window,
                                    const EGLint *attrib_list)
{
    int result = native_window_api_connect(window, NATIVE_WINDOW_API_EGL);
    
    EGLSurface surface = cnx->egl.eglCreateWindowSurface(
            iDpy, config, window, attrib_list);
    if (surface != EGL_NO_SURFACE) {
        egl_surface_t* s = new egl_surface_t(dp.get(), config, window,
                surface, cnx);
        return s;
    }
    return EGL_NO_SURFACE;
}

其中native_window_api_connect会调用到Surface::connect，api=NATIVE_WINDOW_API_EGL

int Surface::connect(int api, const sp<IProducerListener>& listener) {

    int err = mGraphicBufferProducer->connect(listener, api, mProducerControlledByApp, &output);

    return err;
}

BufferQueueProducer::connect里设置了BufferQueueCore里的一些参数。

status_t BufferQueueProducer::connect(const sp<IProducerListener>& listener,
        int api, bool producerControlledByApp, QueueBufferOutput *output) {

    int delta = mCore->getMaxBufferCountLocked(mCore->mAsyncMode,
            mDequeueTimeout < 0 ?
            mCore->mConsumerControlledByApp && producerControlledByApp : false,
            mCore->mMaxBufferCount) -
            mCore->getMaxBufferCountLocked();
    if (!mCore->adjustAvailableSlotsLocked(delta)) {
        BQ_LOGE("connect: BufferQueue failed to adjust the number of available "
                "slots. Delta = %d", delta);
        return BAD_VALUE;
    }

    mCore->mBufferHasBeenQueued = false;
    mCore->mDequeueBufferCannotBlock = false;
    if (mDequeueTimeout < 0) {
        mCore->mDequeueBufferCannotBlock =
                mCore->mConsumerControlledByApp && producerControlledByApp;
    }

    mCore->mAllowAllocation = true;

}

/frameworks/native/opengl/libagl/egl.cpp

EGLSurface eglCreateWindowSurface(  EGLDisplay dpy, EGLConfig config,
                                    NativeWindowType window,
                                    const EGLint *attrib_list)
{
    return createWindowSurface(dpy, config, window, attrib_list);
}

EGLSurface EglManager::createSurface(EGLNativeWindowType window) {
    initialize();
    EGLSurface surface = eglCreateWindowSurface(mEglDisplay, mEglConfig, window, nullptr);
    .....
    return surface;
}

static EGLSurface createWindowSurface(EGLDisplay dpy, EGLConfig config,
        NativeWindowType window, const EGLint* /*attrib_list*/)
{

    egl_surface_t* surface;
    surface = new egl_window_surface_v2_t(dpy, config, depthFormat,
            static_cast<ANativeWindow*>(window));

    return surface;
}

到此为止创建了一个EglSurface保存在CanvasContext中

当使用ThreadedRenderer去draw一帧displaylist时，会调用nSyncAndDrawFrame：

void draw(View view, AttachInfo attachInfo, HardwareDrawCallbacks callbacks) {
    updateRootDisplayList(view, callbacks);
   
    int syncResult = nSyncAndDrawFrame(mNativeProxy, frameInfo, frameInfo.length);
}

最终会调用到CanvasContext的draw方法：

void DrawFrameTask::run() {
    if (CC_LIKELY(canDrawThisFrame)) {
        context->draw();
    }
}

void CanvasContext::draw() {

    Frame frame = mEglManager.beginFrame(mEglSurface);
    ....
    if (drew || mEglManager.damageRequiresSwap()) {
        if (CC_UNLIKELY(!mEglManager.swapBuffers(frame, screenDirty))) {
            setSurface(nullptr);
        }
    }
}

CanvasContext的draw方法会先调用mEglManager.beginFrame，然后调用mEglManager.swapBuffers

Frame EglManager::beginFrame(EGLSurface surface) {
    .....
    makeCurrent(surface);
    ....
}
bool EglManager::makeCurrent(EGLSurface surface, EGLint* errOut) {

    if (!eglMakeCurrent(mEglDisplay, surface, surface, mEglContext)) {
        ....
    }
    .....
}

EGLBoolean eglMakeCurrent(  EGLDisplay dpy, EGLSurface draw,
                            EGLSurface read, EGLContext ctx)
{

    if (ctx == EGL_NO_CONTEXT) {
        // if we're detaching, we need the current context
        current_ctx = (EGLContext)getGlThreadSpecific();
    } else {
       
        egl_surface_t* d = (egl_surface_t*)draw;
    }

    if (d) {
        if (d->connect() == EGL_FALSE) {
            return EGL_FALSE;
        }
    }
    return setError(EGL_BAD_ACCESS, EGL_FALSE);
}

EGLBoolean egl_window_surface_v2_t::connect()
{

    // dequeue a buffer
    int fenceFd = -1;
    if (nativeWindow->dequeueBuffer(nativeWindow, &buffer,
            &fenceFd) != NO_ERROR) {
        return setError(EGL_BAD_ALLOC, EGL_FALSE);
    }

    // wait for the buffer
    sp<Fence> fence(new Fence(fenceFd));
    if (fence->wait(Fence::TIMEOUT_NEVER) != NO_ERROR) {
        nativeWindow->cancelBuffer(nativeWindow, buffer, fenceFd);
        return setError(EGL_BAD_ALLOC, EGL_FALSE);
    }

    return EGL_TRUE;
}

beginFrame其实就是调用ANativeWindow的dequeueBuffer来获取一个buffer。最终调用到Surface的dequeueBuffer

int Surface::dequeueBuffer(android_native_buffer_t** buffer, int* fenceFd) {
    ....
    int buf = -1;
    sp<Fence> fence;
    status_t result = mGraphicBufferProducer->dequeueBuffer(&buf, &fence,
            reqWidth, reqHeight, reqFormat, reqUsage);

    if (result < 0) {
        ALOGV("dequeueBuffer: IGraphicBufferProducer::dequeueBuffer"
                "(%d, %d, %d, %d) failed: %d", reqWidth, reqHeight, reqFormat,
                reqUsage, result);
        return result;
    }

    Mutex::Autolock lock(mMutex);

    sp<GraphicBuffer>& gbuf(mSlots[buf].buffer);

    // this should never happen
    ALOGE_IF(fence == NULL, "Surface::dequeueBuffer: received null Fence! buf=%d", buf);

    if (result & IGraphicBufferProducer::RELEASE_ALL_BUFFERS) {
        freeAllBuffers();
    }

    if ((result & IGraphicBufferProducer::BUFFER_NEEDS_REALLOCATION) || gbuf == 0) {
        result = mGraphicBufferProducer->requestBuffer(buf, &gbuf);
        if (result != NO_ERROR) {
            ALOGE("dequeueBuffer: IGraphicBufferProducer::requestBuffer failed: %d", result);
            mGraphicBufferProducer->cancelBuffer(buf, fence);
            return result;
        }
    }

    if (fence->isValid()) {
        *fenceFd = fence->dup();
        if (*fenceFd == -1) {
            ALOGE("dequeueBuffer: error duping fence: %d", errno);
            // dup() should never fail; something is badly wrong. Soldier on
            // and hope for the best; the worst that should happen is some
            // visible corruption that lasts until the next frame.
        }
    } else {
        *fenceFd = -1;
    }

    *buffer = gbuf.get();

    return OK;
}

GraphicBufferProducer::dequeueBuffer

status_t BufferQueueProducer::dequeueBuffer(int *outSlot,
        sp<android::Fence> *outFence, uint32_t width, uint32_t height,
        PixelFormat format, uint32_t usage) {

    { // Autolock scope
        Mutex::Autolock lock(mCore->mMutex);
        mCore->waitWhileAllocatingLocked();

        if (format == 0) {
            format = mCore->mDefaultBufferFormat;
        }

        // Enable the usage bits the consumer requested
        usage |= mCore->mConsumerUsageBits;

        const bool useDefaultSize = !width && !height;
        if (useDefaultSize) {
            width = mCore->mDefaultWidth;
            height = mCore->mDefaultHeight;
        }

        int found = BufferItem::INVALID_BUFFER_SLOT;
        while (found == BufferItem::INVALID_BUFFER_SLOT) {
            status_t status = waitForFreeSlotThenRelock(FreeSlotCaller::Dequeue,
                    &found);
            if (status != NO_ERROR) {
                return status;
            }
        }

        if (mCore->mSharedBufferSlot != found) {
            mCore->mActiveBuffers.insert(found);
        }
        *outSlot = found;
        ATRACE_BUFFER_INDEX(found);

        attachedByConsumer = mSlots[found].mNeedsReallocation;
        mSlots[found].mNeedsReallocation = false;

        mSlots[found].mBufferState.dequeue();

        if ((buffer == NULL) ||
                buffer->needsReallocation(width, height, format, usage))
        {
            mSlots[found].mAcquireCalled = false;
            mSlots[found].mGraphicBuffer = NULL;
            mSlots[found].mRequestBufferCalled = false;
            mSlots[found].mEglDisplay = EGL_NO_DISPLAY;
            mSlots[found].mEglFence = EGL_NO_SYNC_KHR;
            mSlots[found].mFence = Fence::NO_FENCE;
            mCore->mBufferAge = 0;
            mCore->mIsAllocating = true;

            returnFlags |= BUFFER_NEEDS_REALLOCATION;
        } 
    } // Autolock scope

    if (returnFlags & BUFFER_NEEDS_REALLOCATION) {
        status_t error;
        BQ_LOGV("dequeueBuffer: allocating a new buffer for slot %d", *outSlot);
        sp<GraphicBuffer> graphicBuffer(mCore->mAllocator->createGraphicBuffer(
                width, height, format, usage, &error));
        { // Autolock scope
            Mutex::Autolock lock(mCore->mMutex);

            if (graphicBuffer != NULL && !mCore->mIsAbandoned) {
                graphicBuffer->setGenerationNumber(mCore->mGenerationNumber);
                mSlots[*outSlot].mGraphicBuffer = graphicBuffer;
            }

            mCore->mIsAllocating = false;
            mCore->mIsAllocatingCondition.broadcast();
        }
    }


    return returnFlags;
}

GraphicBufferProducer::dequeueBuffer中会从BufferQueueCore中分配一个free slot，然后这个slot对应的mSlots[found]的mGraphicBuffer是空的，这时会走到

 sp<GraphicBuffer> graphicBuffer(mCore->mAllocator->createGraphicBuffer(width, height, format, usage, &error));

会调用GraphicBufferAlloc.createGraphicBuffer去从SurfaceFlinger分配一块GraphicBuffer。GraphicBufferAlloc继承自BnGraphicBufferAlloc

class GraphicBufferAlloc : public BnGraphicBufferAlloc

sp<GraphicBuffer> GraphicBufferAlloc::createGraphicBuffer(uint32_t width,
        uint32_t height, PixelFormat format, uint32_t usage, status_t* error) {
    sp<GraphicBuffer> graphicBuffer(
            new GraphicBuffer(width, height, format, usage));
    status_t err = graphicBuffer->initCheck();
    *error = err;
    if (err != 0 || graphicBuffer->handle == 0) {
        if (err == NO_MEMORY) {
            GraphicBuffer::dumpAllocationsToSystemLog();
        }
        ALOGE("GraphicBufferAlloc::createGraphicBuffer(w=%d, h=%d) "
             "failed (%s), handle=%p",
                width, height, strerror(-err), graphicBuffer->handle);
        return 0;
    }
    return graphicBuffer;
}

GraphicBuffer实现了Flattenable接口来进行序列化，同时实现了ANativeWindowBuffer接口来供ANativeWindow来进行绘制图像：

class GraphicBuffer
    : public ANativeObjectBase< ANativeWindowBuffer, GraphicBuffer, RefBase >,
      public Flattenable<GraphicBuffer>

GraphicBuffer的构造函数里请求gralloc来分配缓存。这块缓存应该是SurfaceFlinger和App共享的。

GraphicBuffer::GraphicBuffer(uint32_t inWidth, uint32_t inHeight,
        PixelFormat inFormat, uint32_t inUsage)
    : BASE(), mOwner(ownData), mBufferMapper(GraphicBufferMapper::get()),
      mInitCheck(NO_ERROR), mId(getUniqueId()), mGenerationNumber(0)
{
    width  =
    height =
    stride =
    format =
    usage  = 0;
    handle = NULL;
    mInitCheck = initSize(inWidth, inHeight, inFormat, inUsage);
}

status_t GraphicBuffer::initSize(uint32_t inWidth, uint32_t inHeight,
        PixelFormat inFormat, uint32_t inUsage)
{
    GraphicBufferAllocator& allocator = GraphicBufferAllocator::get();
    status_t err = allocator.alloc(inWidth, inHeight, inFormat, inUsage,
            &handle, &outStride); //通过GraphicBufferAllocator来分配缓存
    return err;
}

status_t GraphicBufferAllocator::alloc(uint32_t width, uint32_t height,
        PixelFormat format, uint32_t usage, buffer_handle_t* handle,
        uint32_t* stride)
{

    err = mAllocDev->alloc(mAllocDev, static_cast<int>(width),
            static_cast<int>(height), format, static_cast<int>(usage), handle,
            &outStride);
    *stride = static_cast<uint32_t>(outStride);

    return err;
}

GraphicBufferAllocator::GraphicBufferAllocator()
    : mAllocDev(0)
{
    hw_module_t const* module;
    int err = hw_get_module(GRALLOC_HARDWARE_MODULE_ID, &module);
    ALOGE_IF(err, "FATAL: can't find the %s module", GRALLOC_HARDWARE_MODULE_ID);
    if (err == 0) {
        gralloc_open(module, &mAllocDev);
    }
}

mAllocDev应该是通过gralloc_open赋值的。