Android6.0 图像合成过程详解(二) doComposition函数

1、Android6.0 图像合成过程详解（二） doComposition函数一、doComposition合成图层doComposition这个函数就是合成所有层的图像cpp view plain copy 在CODE上查看代码片派生到我的代码片void SurfaceFlinger:doComposition() ATRACE_CALL(); const bool repaintEverything = android_atomic_and(0, &mRepaintEverything); for (size_t dpy=0 ; dpymDisplays.size() ; dpy+) con

2、st sp& hw(mDisplaysdpy); if (hw-isDisplayOn() / transform the dirty region into this screens coordinate space const Region dirtyRegion(hw-getDirtyRegion(repaintEverything); / repaint the framebuffer (if needed) doDisplayComposition(hw, dirtyRegion); hw-dirtyRegion.clear(); hw-flip(hw-swapRegion); hw

3、-swapRegion.clear(); / inform the h/w that were done compositing hw-compositionComplete(); postFramebuffer(); 上面函数遍历所有的DisplayDevice然后调用doDisplayComposition函数。然后我们再看看doDisplayComposition函数cpp view plain copy 在CODE上查看代码片派生到我的代码片void SurfaceFlinger:doDisplayComposition(const sp& hw, const Region& inDi

4、rtyRegion) bool isHwcDisplay = hw-getHwcDisplayId() = 0; if (!isHwcDisplay & inDirtyRegion.isEmpty() return; Region dirtyRegion(inDirtyRegion); /swapRegion设置为需要更新的区域 hw-swapRegion.orSelf(dirtyRegion); uint32_t flags = hw-getFlags();/获得显示设备支持的更新方式标志 if (flags & DisplayDevice:SWAP_RECTANGLE) /支持矩阵更新 d

5、irtyRegion.set(hw-swapRegion.bounds(); else if (flags & DisplayDevice:PARTIAL_UPDATES) /支持部分更新 dirtyRegion.set(hw-swapRegion.bounds(); else /将更新区域调整为整个窗口大小 dirtyRegion.set(hw-bounds(); hw-swapRegion = dirtyRegion; if (CC_LIKELY(!mDaltonize & !mHasColorMatrix) if (!doComposeSurfaces(hw, dirtyRegion)

6、return;/合成 else RenderEngine& engine(getRenderEngine(); mat4 colorMatrix = mColorMatrix; if (mDaltonize) colorMatrix = colorMatrix * mDaltonizer(); mat4 oldMatrix = engine.setupColorTransform(colorMatrix); doComposeSurfaces(hw, dirtyRegion);/合成 engine.setupColorTransform(oldMatrix); / update the swa

7、p region and clear the dirty region hw-swapRegion.orSelf(dirtyRegion); / swap buffers (presentation) hw-swapBuffers(getHwComposer();/使用egl将egl中的合成好的图像，输出到DisplayDevice的mSurface中 这个函数设置下需要更新的区域，后面调用doComposeSurfaces函数来合成图层，调用完doComposeSurfaces函数后，如果需要egl合成图像话，在这个函数中合成好。而最后调用swapBuffers只是将egl合成好的图像输出到

8、DisplayDevice的mSurface中。我们再来看看doComposeSurfaces函数,我们先来看一开始的代码，先判断是否有egl合成，然后再看是否有hwc合成（硬件合成）cpp view plain copy 在CODE上查看代码片派生到我的代码片bool SurfaceFlinger:doComposeSurfaces(const sp& hw, const Region& dirty) RenderEngine& engine(getRenderEngine(); const int32_t id = hw-getHwcDisplayId(); HWComposer& hwc

9、(getHwComposer(); HWComposer:LayerListIterator cur = hwc.begin(id); const HWComposer:LayerListIterator end = hwc.end(id); bool hasGlesComposition = hwc.hasGlesComposition(id); if (hasGlesComposition) /是否有egl合成 if (!hw-makeCurrent(mEGLDisplay, mEGLContext) ALOGW(DisplayDevice:makeCurrent failed. Abor

10、ting surface composition for display %s, hw-getDisplayName().string(); eglMakeCurrent(mEGLDisplay, EGL_NO_SURFACE, EGL_NO_SURFACE, EGL_NO_CONTEXT); if(!getDefaultDisplayDevice()-makeCurrent(mEGLDisplay, mEGLContext) ALOGE(DisplayDevice:makeCurrent on default display failed. Aborting.); return false;

11、 / Never touch the framebuffer if we dont have any framebuffer layers const bool hasHwcComposition = hwc.hasHwcComposition(id); if (hasHwcComposition) /是否有hwc合成 / when using overlays, we assume a fully transparent framebuffer / NOTE: we could reduce how much we need to clear, for instance / remove w

12、here there are opaque FB layers. however, on some / GPUs doing a clean slate clear might be more efficient. / Well revisit later if needed. engine.clearWithColor(0, 0, 0, 0); else / we start with the whole screen area const Region bounds(hw-getBounds(); / we remove the scissor part / were left with

13、the letterbox region / (common case is that letterbox ends-up being empty) const Region letterbox(bounds.subtract(hw-getScissor(); / compute the area to clear Region region(hw-undefinedRegion.merge(letterbox); / but limit it to the dirty region region.andSelf(dirty); / screen is already cleared here

14、 if (!region.isEmpty() / can happen with SurfaceView drawWormhole(hw, region); if (hw-getDisplayType() != DisplayDevice:DISPLAY_PRIMARY) / just to be on the safe side, we dont set the / scissor on the main display. It should never be needed / anyways (though in theory it could since the API allows i

15、t). const Rect& bounds(hw-getBounds(); const Rect& scissor(hw-getScissor(); if (scissor != bounds) / scissor doesnt match the screens dimensions, so we / need to clear everything outside of it and enable / the GL scissor so we dont draw anything where we shouldnt / enable scissor for this frame cons

16、t uint32_t height = hw-getHeight(); engine.setScissor(scissor.left, height - scissor.bottom, scissor.getWidth(), scissor.getHeight(); . 我们来看hasGlesComposition函数和hasHwcComposition函数，就是看其对应的DisplayData中是否有hasFbComp和hasOvComp。cpp view plain copy 在CODE上查看代码片派生到我的代码片bool HWComposer:hasGlesComposition(int

17、32_t id) const if (!mHwc | uint32_t(id)31 | !mAllocatedDisplayIDs.hasBit(id) return true; return mDisplayDataid.hasFbComp; cpp view plain copy 在CODE上查看代码片派生到我的代码片bool HWComposer:hasHwcComposition(int32_t id) const if (!mHwc | uint32_t(id)31 | !mAllocatedDisplayIDs.hasBit(id) return false; return mDi

18、splayDataid.hasOvComp; 而这两个值是在prepare中调用Hwc的prepare函数之后赋值的cpp view plain copy 在CODE上查看代码片派生到我的代码片status_t HWComposer:prepare() . int err = mHwc-prepare(mHwc, mNumDisplays, mLists); ALOGE_IF(err, HWComposer: prepare failed (%s), strerror(-err); if (err = NO_ERROR) / here were just making sure that sk

19、ip layers are set / to HWC_FRAMEBUFFER and were also counting how many layers / we have of each type. / / If there are no window layers, we treat the display has having FB / composition, because SurfaceFlinger will use GLES to draw the / wormhole region. for (size_t i=0 ; imNumDisplays ; i+) Display

20、Data& disp(mDisplayDatai); disp.hasFbComp = false; disp.hasOvComp = false; if (disp.list) for (size_t i=0 ; inumHwLayers ; i+) hwc_layer_1_t& l = disp.list-hwLayersi; /ALOGD(prepare: %d, type=%d, handle=%p, / i, positionType, l.handle); if (l.flags & HWC_SKIP_LAYER) positionType = HWC_FRAMEBUFFER; i

21、f (positionType = HWC_FRAMEBUFFER) disp.hasFbComp = true;/只要有一个layer是HWC_FRAMEBUFFER if (positionType = HWC_OVERLAY) disp.hasOvComp = true;/有一个layer是HWC_OVERLAY if (positionType = HWC_CURSOR_OVERLAY) disp.hasOvComp = true;/有一个layer是HWC_CURSOR_OVERLAY if (disp.list-numHwLayers = (disp.framebufferTarg

22、et ? 1 : 0) /layer的数量 有framebufferTarget为1 没有为0 disp.hasFbComp = true; else disp.hasFbComp = true;/没有list return (status_t)err; 我们继续看doComposeSurfaces函数，下面这个函数当cur!=end代表起码有两个以上图层，然后遍历图层，当layer是HWC_FRAMEBUFFER代表是需要egl合成的，而HWC_FRAMEBUFFER_TARGET是egl合成后使用的直接就跳了，HWC_CURSOR_OVERLAY和HWC_OVERLAY是用HWC模块（硬件

23、合成）的，也就不用调用Layer的draw方法。而如果图层只要1个或者没有，那么直接使用egl合成。cpp view plain copy 在CODE上查看代码片派生到我的代码片 HWComposer:LayerListIterator cur = hwc.begin(id); const HWComposer:LayerListIterator end = hwc.end(id); . const Vector sp & layers(hw-getVisibleLayersSortedByZ(); const size_t count = layers.size(); const Trans

24、form& tr = hw-getTransform(); if (cur != end) /代表起码有两个以上图层 / were using h/w composer for (size_t i=0 ; icount & cur!=end ; +i, +cur) /遍历图层 const sp& layer(layersi); const Region clip(ersect(tr.transform(layer-visibleRegion); if (!clip.isEmpty() switch (cur-getCompositionType() case HWC_CURS

25、OR_OVERLAY: case HWC_OVERLAY: const Layer:State& state(layer-getDrawingState(); if (cur-getHints() & HWC_HINT_CLEAR_FB) & i & layer-isOpaque(state) & (state.alpha = 0xFF) & hasGlesComposition) / never clear the very first layer since were / guaranteed the FB is already cleared layer-clearWithOpenGL(

26、hw, clip); break; case HWC_FRAMEBUFFER: layer-draw(hw, clip);/只有是HWC_FRAMEBUFFER才会调用Layer的draw合成 break; case HWC_FRAMEBUFFER_TARGET: / this should not happen as the iterator shouldnt / let us get there. ALOGW(HWC_FRAMEBUFFER_TARGET found in hwc list (index=%zu), i); break; layer-setAcquireFence(hw,

27、*cur); else / were not using h/w composer for (size_t i=0 ; icount ; +i) /只有一个或者没有图层 就直接使用Layer的draw合成 const sp& layer(layersi); const Region clip(ersect( tr.transform(layer-visibleRegion); if (!clip.isEmpty() layer-draw(hw, clip); / disable scissor at the end of the frame engine.disableSci

28、ssor(); return true; Layer的draw我们就不看了主要是使用egl合成纹理，但是有一点疑问，我们从来没有把layer中的mActiveBuffer放到egl中去，那么egl又是怎么合成各个layer的呢，我想肯定客户进程在绘制各个layer的时候，也是用egl绘制的，所有后面合成的时候egl有各个layer的buffer。后面我们再来看下DisplayDevice:swapBuffers函数，是使用eglSwapBuffers来把egl合成的数据放到mSurface中去。cpp view plain copy 在CODE上查看代码片派生到我的代码片void Displa

29、yDevice:swapBuffers(HWComposer& hwc) const / We need to call eglSwapBuffers() if: / (1) we dont have a hardware composer, or / (2) we did GLES composition this frame, and either / (a) we have framebuffer target support (not present on legacy / devices, where HWComposer:commit() handles things); or /

30、 (b) this is a virtual display if (hwc.initCheck() != NO_ERROR | (hwc.hasGlesComposition(mHwcDisplayId) & (hwc.supportsFramebufferTarget() | mType = DISPLAY_VIRTUAL) EGLBoolean success = eglSwapBuffers(mDisplay, mSurface); if (!success) EGLint error = eglGetError(); if (error = EGL_CONTEXT_LOST | mT

31、ype = DisplayDevice:DISPLAY_PRIMARY) LOG_ALWAYS_FATAL(eglSwapBuffers(%p, %p) failed with 0x%08x, mDisplay, mSurface, error); else ALOGE(eglSwapBuffers(%p, %p) failed with 0x%08x, mDisplay, mSurface, error); else if(hwc.supportsFramebufferTarget() | mType = DISPLAY_VIRTUAL) EGLBoolean success = eglSw

32、apBuffersVIV(mDisplay, mSurface); if (!success) EGLint error = eglGetError(); ALOGE(eglSwapBuffersVIV(%p, %p) failed with 0x%08x, mDisplay, mSurface, error); status_t result = mDisplaySurface-advanceFrame(); if (result != NO_ERROR) ALOGE(%s failed pushing new frame to HWC: %d, mDisplayName.string(),

33、 result); 二、FramebufferSurface收到egl合成数据之前分析DisplayDevice时候，还分析了FramebufferSurface，我们这里再来看下。在SurfaceFlinger.cpp中的init函数，在创建DisplayDevice之前，我们先调用createBufferQueue来创建了一个buffer的生产者和消费者，然后把消费者放入了FramebufferSurface，生产者放入了DisplayDevice中。cpp view plain copy 在CODE上查看代码片派生到我的代码片sp producer; sp consumer; Buffe

34、rQueue:createBufferQueue(&producer, &consumer, new GraphicBufferAlloc(); sp fbs = new FramebufferSurface(*mHwc, i, consumer); int32_t hwcId = allocateHwcDisplayId(type); sp hw = new DisplayDevice(this, type, hwcId, mHwc-getFormat(hwcId), isSecure, token, fbs, producer, mRenderEngine-getEGLConfig();

35、我们先来看生产者，下面是DisplayDevice的构造函数，生产者作为参数直接新建了一个Surface，然后把这个Surface作为参数调用eglCreateWindowSurface返回的就是mSurface，之前我们分析最后egl合成的数据时调用eglSwapBuffers并且把数据放到mSurface，这样最后肯定就到消费者（FramebufferSurface）去了。cpp view plain copy 在CODE上查看代码片派生到我的代码片mNativeWindow = new Surface(producer, false); ANativeWindow* const wind

36、ow = mNativeWindow.get(); /* * Create our displays surface */ EGLSurface surface; EGLDisplay display = eglGetDisplay(EGL_DEFAULT_DISPLAY); if (config = EGL_NO_CONFIG) config = RenderEngine:chooseEglConfig(display, format); surface = eglCreateWindowSurface(display, config, window, NULL); 最后到消费者那端的onF

37、rameAvailable，也就是FramebufferSurface的onFrameAvailable中，我们现在来分析下这个过程，也就解答了一个onFrameAvailable的疑惑。FramebufferSurface的父类是ConsumerBase类，我们来看其构造函数。先是构造了mConsumer，这里其实就是BufferQueueConsumer类，后面调用了其consumerConnect方法。cpp view plain copy 在CODE上查看代码片派生到我的代码片ConsumerBase:ConsumerBase(const sp& bufferQueue, bool c

38、ontrolledByApp) : mAbandoned(false), mConsumer(bufferQueue) mName = String8:format(unnamed-%d-%d, getpid(), createProcessUniqueId(); wp listener = static_cast(this); sp proxy = new BufferQueue:ProxyConsumerListener(listener); status_t err = mConsumer-consumerConnect(proxy, controlledByApp); if (err

39、!= NO_ERROR) CB_LOGE(ConsumerBase: error connecting to BufferQueue: %s (%d), strerror(-err), err); else mConsumer-setConsumerName(mName); 我们来看下BufferQueueConsumer类的consumerConnect方法，就是调用了connect方法。cpp view plain copy 在CODE上查看代码片派生到我的代码片virtual status_t consumerConnect(const sp& consumer, bool contro

40、lledByApp) return connect(consumer, controlledByApp); 这个方法中将mCore-mConsumerListener = consumerListener，这个mCore就是BufferQueueCore类。我们再从ConsumerBase的构造函数看这个consumerListener参数其实就是FrameBufferSurface对象本身。cpp view plain copy 在CODE上查看代码片派生到我的代码片status_t BufferQueueConsumer:connect( const sp& consumerListene

41、r, bool controlledByApp) ATRACE_CALL(); if (consumerListener = NULL) BQ_LOGE(connect(C): consumerListener may not be NULL); return BAD_VALUE; BQ_LOGV(connect(C): controlledByApp=%s, controlledByApp ? true : false); Mutex:Autolock lock(mCore-mMutex); if (mCore-mIsAbandoned) BQ_LOGE(connect(C): Buffer

42、Queue has been abandoned); return NO_INIT; mCore-mConsumerListener = consumerListener;/设置回调 mCore-mConsumerControlledByApp = controlledByApp; return NO_ERROR; 我们再看BufferQueueProducer:queueBuffer函数，这个函数应该是生产者已经使用好buffer了，这个使用会调用如下代码这个listener就是BufferQueueCore的mConsumerListener，传输的数据时BufferItem。再传之前把BufferItem的mGraphicBuffer清了，因为消费者可以自己获取buffer，不用通过BufferItem传。cpp view plain copy 在CODE上查看代码片派生到我的代码片item.mGraphicBuffer.clear(); item.mSlot = BufferItem:INVALID_BUFFER_SLOT; / Call back without the main BufferQueue lock held, but with the callback / lock held so we can ensure that callbacks occur