media: Introduce Y8 and Y16 video pixel format

cc/resources/video_resource_updater.cc

 // Copyright 2013 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #include "cc/resources/video_resource_updater.h"
 
 #include <stddef.h>
 #include <stdint.h>
 
 #include <algorithm>
@@ skipping 65 matching lines @@
     case media::PIXEL_FORMAT_RGB24:
     case media::PIXEL_FORMAT_RGB32:
     case media::PIXEL_FORMAT_MJPEG:
     case media::PIXEL_FORMAT_MT21:
     case media::PIXEL_FORMAT_YUV420P9:
     case media::PIXEL_FORMAT_YUV422P9:
     case media::PIXEL_FORMAT_YUV444P9:
     case media::PIXEL_FORMAT_YUV420P10:
     case media::PIXEL_FORMAT_YUV422P10:
     case media::PIXEL_FORMAT_YUV444P10:
+    case media::PIXEL_FORMAT_Y8:
+    case media::PIXEL_FORMAT_Y16:
     case media::PIXEL_FORMAT_UNKNOWN:
       break;
   }
   return VideoFrameExternalResources::NONE;
 }
 
 class SyncTokenClientImpl : public media::VideoFrame::SyncTokenClient {
  public:
   SyncTokenClientImpl(gpu::gles2::GLES2Interface* gl,
                       const gpu::SyncToken& sync_token)
@@ skipping 167 matching lines @@
     case media::PIXEL_FORMAT_NV12:
     case media::PIXEL_FORMAT_NV21:
     case media::PIXEL_FORMAT_UYVY:
     case media::PIXEL_FORMAT_YUY2:
     case media::PIXEL_FORMAT_ARGB:
     case media::PIXEL_FORMAT_XRGB:
     case media::PIXEL_FORMAT_RGB24:
     case media::PIXEL_FORMAT_RGB32:
     case media::PIXEL_FORMAT_MJPEG:
     case media::PIXEL_FORMAT_MT21:
+    case media::PIXEL_FORMAT_Y8:
       bits_per_channel = 8;
       break;
     case media::PIXEL_FORMAT_YUV420P9:
     case media::PIXEL_FORMAT_YUV422P9:
     case media::PIXEL_FORMAT_YUV444P9:
       bits_per_channel = 9;
       break;
     case media::PIXEL_FORMAT_YUV420P10:
     case media::PIXEL_FORMAT_YUV422P10:
     case media::PIXEL_FORMAT_YUV444P10:
       bits_per_channel = 10;
       break;
+    case media::PIXEL_FORMAT_Y16:

[hubbe, 2016/07/06 17:39:00]

The code below doesn't support bits_per_channel higher than 10.
Thus, this will not work unless you fix the convert-to-half-float code below.

[dshwang, 2016/07/06 17:45:18]

yes, I know. It will be handled in the next CL;
https://codereview.chromium.org/2122573003/

[Ken Russell (switch to Gerrit), 2016/07/06 18:24:42]

Is it wise to land this patch without the next CL? While the other CL is huge,
breaking the code temporarily doesn't seem to me a good path to take.

[dshwang, 2016/07/07 12:41:08]

It doesn't break any code, because any video decoders don't use Y16 yet. This
line is added to avoid build failure.

Next CL has concern about R16 vs. RB8, but this CL will be intact, no matter
what decision happens.

IMO, it's good to land.

[hubbe, 2016/07/07 17:51:39]

You need to do the proper error handling.
Always assume that your next CL is going to land a long time from now and
someone else might try to use or change your code in between.

[dshwang, 2016/07/07 18:11:38]

Ok, I understand. Done.

+      bits_per_channel = 16;
+      break;
   }
 
   // Only YUV software video frames are supported.
   if (!media::IsYuvPlanar(input_frame_format)) {
     NOTREACHED() << media::VideoPixelFormatToString(input_frame_format);
     return VideoFrameExternalResources();
   }
 
   const bool software_compositor = context_provider_ == NULL;
 
@@ skipping 394 matching lines @@
   if (lost_resource) {
     resource_it->clear_refs();
     updater->DeleteResource(resource_it);
     return;
   }
 
   resource_it->remove_ref();
 }
 
 }  // namespace cc