Mo, as pixel wrangler, can you figure out what to revert?

Issued a revert: https://chromium-review.googlesource.com/c/chromium/src/+/1388645

Pixel colors are a little off with decode swap chain or d3d11 video decoder:

AssertionError: Expected pixel at [5, 5] (actual pixel (5, 5))  to be [44, 255, 16] but got [61, 255, 25]

Locals:
  msg : 'Expected pixel at [5, 5] (actual pixel (5, 5))  to be [44, 255, 16] but got [61, 255, 25]'

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/311d14e9566d4f9f082d6b5c2fd309ed4c450e8e

commit 311d14e9566d4f9f082d6b5c2fd309ed4c450e8e
Author: Sunny Sachanandani <sunnyps@chromium.org>
Date: Fri Dec 21 17:26:26 2018

Revert "Enable decode swap chain feature on waterfall"

This reverts commit e1f01a1580387ffd655ad2504e9336d79b4cb5c1.

Reason for revert: Makes video pixel tests flaky: crbug.com/917399

Original change's description:
> Enable decode swap chain feature on waterfall
> 
> Enables DirectCompositionUseNV12DecodeSwapChain and D3D11VideoDecoder
> features on waterfall.
> 
> Bug: 848093
> Change-Id: I561f5910a32cbe2adb1c47bd9c35df355bdce1a4
> Reviewed-on: https://chromium-review.googlesource.com/c/1364341
> Commit-Queue: Sunny Sachanandani <sunnyps@chromium.org>
> Reviewed-by: Ilya Sherman <isherman@chromium.org>
> Reviewed-by: Frank Liberato <liberato@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#618447}

TBR=isherman@chromium.org,sunnyps@chromium.org,liberato@chromium.org

Change-Id: I13cb41b59140475eb46e789e8c94fc4e610ab614
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: 848093, 917399
Reviewed-on: https://chromium-review.googlesource.com/c/1388645
Reviewed-by: Sunny Sachanandani <sunnyps@chromium.org>
Commit-Queue: Sunny Sachanandani <sunnyps@chromium.org>
Cr-Commit-Position: refs/heads/master@{#618549}
[modify] https://crrev.com/311d14e9566d4f9f082d6b5c2fd309ed4c450e8e/testing/variations/fieldtrial_testing_config.json

Thanks Sunny.

This makes sense since we don't override the decode swap chain color space.  When I used Rec 709 last time I got muted colors (on crosvideo.appspot.com) which I believe is emblematic of limited range not being expanded. <- dalecurtis@

The color space is specified using DXGI_MULTIPLANE_OVERLAY_YCbCr_FLAGS:

typedef enum DXGI_MULTIPLANE_OVERLAY_YCbCr_FLAGS {
  DXGI_MULTIPLANE_OVERLAY_YCbCr_FLAG_NOMINAL_RANGE,
  DXGI_MULTIPLANE_OVERLAY_YCbCr_FLAG_BT709,
  DXGI_MULTIPLANE_OVERLAY_YCbCr_FLAG_xvYCC
};

These are bitwise-OR'd together.  The values start at 1 so I don't know what 0 means or what the default actually is.

https://docs.microsoft.com/en-us/windows/desktop/api/dxgi1_3/ne-dxgi1_3-dxgi_multiplane_overlay_ycbcr_flags

BTW on my Pixelbook, the colors on crosvideo don't change when transitioning between overlay and compositing, but the video shifts a miniscule amount, probably 1 pixel.  <- dcastagna@

What confuses me is why it's flaky.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/dbf4142cb21869051c3a7c7e3301abf3138485e0

commit dbf4142cb21869051c3a7c7e3301abf3138485e0
Author: Sunny Sachanandani <sunnyps@chromium.org>
Date: Sat Dec 22 03:08:57 2018

Set color space on decode swap chain

Use BT 709 primaries by default, use nominal (limited) range if color
space is Rec 709 or 601, and use xvYCC for full range encoding.

Bug: 848093, 917399
Change-Id: Iab54c9523826853265cf92cdd5b806fd82cd03e1
Reviewed-on: https://chromium-review.googlesource.com/c/1389504
Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Commit-Queue: Sunny Sachanandani <sunnyps@chromium.org>
Cr-Commit-Position: refs/heads/master@{#618719}
[modify] https://crrev.com/dbf4142cb21869051c3a7c7e3301abf3138485e0/gpu/ipc/service/direct_composition_surface_win.cc

Doesn't seem to be flaky anymore. Can we close this?