V4L2SVDA/VAAPIVDA: use visible size from decoder and pass to client

media/gpu/vp9_decoder.h

Index: media/gpu/vp9_decoder.h
diff --git a/media/gpu/vp9_decoder.h b/media/gpu/vp9_decoder.h
index d2264b595f42b6544bc85377cdc2d1b777e8ba75..2aa2d48c2b5252e9ca2394faa5c8f1d9d4a96c2b 100644
--- a/media/gpu/vp9_decoder.h
+++ b/media/gpu/vp9_decoder.h
@@ -17,6 +17,7 @@
 #include "media/filters/vp9_parser.h"
 #include "media/gpu/accelerated_video_decoder.h"
 #include "media/gpu/vp9_picture.h"
+#include "ui/gfx/geometry/rect.h"
 
 namespace media {
 
@@ -76,7 +77,8 @@ class MEDIA_GPU_EXPORT VP9Decoder : public AcceleratedVideoDecoder {
     // immediately after calling this method.
     //
     // Return true when successful, false otherwise.
-    virtual bool OutputPicture(const scoped_refptr<VP9Picture>& pic) = 0;
+    virtual bool OutputPicture(const scoped_refptr<VP9Picture>& pic,
+                               const gfx::Rect& visible_rect) = 0;
 
     // Return true if the accelerator requires the client to provide frame
     // context in order to decode. If so, the Vp9FrameHeader provided by the
@@ -139,6 +141,8 @@ class MEDIA_GPU_EXPORT VP9Decoder : public AcceleratedVideoDecoder {
 
   // Current coded resolution.
   gfx::Size pic_size_;
+  // Current render resolution.
+  gfx::Rect render_rect_;
 
   // VP9Accelerator instance owned by the client.
   VP9Accelerator* accelerator_;