media/gpu/v4l2svda: use visible size as natural size of wrapping frame

Using the coded size as the natural size of a frame does not make much
sense. Since this parameter is not essential here, use the visible size
for consistency.

BUG=None
TEST=E2E Youtube decoding working on Kukui.

Change-Id: I6958cd6d1568941e28bd5ca881a9afb6683c9392
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2011782
Auto-Submit: Alexandre Courbot <acourbot@chromium.org>
Reviewed-by: Hirokazu Honda <hiroh@chromium.org>
Commit-Queue: Alexandre Courbot <acourbot@chromium.org>
Cr-Commit-Position: refs/heads/master@{#733861}

media/gpu/v4l2/v4l2_slice_video_decode_accelerator.cc

@@ -2160,7 +2160,7 @@ bool V4L2SliceVideoDecodeAccelerator::ProcessFrame(
     // imported frame into another one that we can destruct.
     scoped_refptr<VideoFrame> wrapped_frame = VideoFrame::WrapVideoFrame(
         output_frame, output_frame->format(), output_frame->visible_rect(),
-        output_frame->coded_size());
+        output_frame->visible_rect().size());
     DCHECK(wrapped_frame);
 
     image_processor_->Process(