gpu video: optimize HW video to SW canvas and implement it for WebRTC.

content/renderer/media/renderer_gpu_video_accelerator_factories.h

Index: content/renderer/media/renderer_gpu_video_accelerator_factories.h
diff --git a/content/renderer/media/renderer_gpu_video_accelerator_factories.h b/content/renderer/media/renderer_gpu_video_accelerator_factories.h
index 2d25de37f8da8eb5eddfd3fb163a777265da0c69..26ebf6dde16b3736ef5d27bd812a3c613225efef 100644
--- a/content/renderer/media/renderer_gpu_video_accelerator_factories.h
+++ b/content/renderer/media/renderer_gpu_video_accelerator_factories.h
@@ -58,9 +58,6 @@ class CONTENT_EXPORT RendererGpuVideoAcceleratorFactories
                       uint32 texture_target) override;
   void DeleteTexture(uint32 texture_id) override;
   void WaitSyncPoint(uint32 sync_point) override;
-  void ReadPixels(uint32 texture_id,
-                  const gfx::Rect& visible_rect,
-                  const SkBitmap& pixels) override;
   scoped_ptr<base::SharedMemory> CreateSharedMemory(size_t size) override;
   scoped_refptr<base::SingleThreadTaskRunner> GetTaskRunner() override;
   std::vector<media::VideoEncodeAccelerator::SupportedProfile>