GrFence is implemented by a single MTLSharedEvent where we increase the
value with each new invocation. GrSemaphore uses a MTLEvent (the
assumption here is that we are signaling and waiting on the same device)
with an associated value that defaults to 1. For generating a large
number of GrSemaphores at once it should be possible to use the same
MTLEvent but with different assigned values.
Bug: skia:8243
Change-Id: Ic7de2d9d295fbe51e67bc7c3c4354257cb0774d4
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/233416
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Change-Id: Ifbe4bcd40bcf316868d23c2edf73b2371c6cb312
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/237119
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
We've failed to get these bots to report where issues happen
any more finely than the name of the executable, which makes
them a real pain to fix when they go red.
We don't expect we'll be able to run cleanly in this mode for
long without bots enforcing it, so remove support from GN too.
Change-Id: Ie86f0cbf2f5f859ac2ddb869da7e5b8d31b33fa0
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/237195
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
GrDawnGpu vends ring buffer slices out of a persistent (larger) buffer.
GrDawnProgramDataManager::setData() now returns a BindGroup containing
the (possibly new) UBO bindings, as well as the texture and sampler
bindings.
Change-Id: Id6694d6f44a815cfbffe4293779bf9bf558a2365
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/235866
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
In a follow up CL I will be merging GrRenderTargetOpList and GrOpList
since we no longer have need for that separation.
Change-Id: I267ead3beeceaece504968ca69165c425d38f761
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/236337
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>
This also means copies are no longer part of OpList or GpuCommandBuffer.
Change-Id: I234de2c93ca13e069d7e4423a3c8b292be2eb500
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/235831
Commit-Queue: Greg Daniel <egdaniel@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
AFAICT these three remaining uses of GrPendingIOResource were just real-ref placeholders (i.e., they just added 1 pending read and then held on to the GrPendingIOResource until object destruction).
Change-Id: Iba6bb199477d846d079999f09794422e240c675a
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/235458
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
These Ops don't need to go through the GrGpuCommandBuffer.
Change-Id: Ifa9c7d801c97c225989d2542709c89f36ddfdc5e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/235799
Commit-Queue: Greg Daniel <egdaniel@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
gn_to_bp: wrap defines in ifndef
create_apk.py assumes you are either run from the aosp tree, or you are
being run from make_universal_apk.py, which now defers to create_apk.py
for all functionality, even argument parsing.
tools/skqp/generate_gn_args moved some functionity into skqp_gn_args.py,
which is now used by create_apk.py
create_apk now accepts android sdk license for you.
create_apk and make_universal_apk.py now are better about cleaning up
after exceptions happen.
Old script make_apk.sh now just points at make_universal_apk.py
CQ_INCLUDE_TRYBOTS=skia.primary:Build-Debian9-Clang-x86-devrel-Android_SKQP,Test-Debian9-Clang-NUC7i5BNK-CPU-Emulator-x86-devrel-All-Android_SKQP
Change-Id: I2dba20ef7017987cabb2bd49f070e2b1594785d5
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/235678
Reviewed-by: Ben Wagner <bungeman@google.com>
Commit-Queue: Hal Canary <halcanary@google.com>
Move everything to SkRemoteGlyphCache.cpp. Add a call
to SkStrikeServer (AddGlyphForTest), to make tests work.
* Misc cleanups
Change-Id: I22bea686838f60cf96d097e1ea025cb5399903e7
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/235096
Reviewed-by: Ben Wagner <bungeman@google.com>
Commit-Queue: Herb Derby <herb@google.com>
The new namespace 'skif' will contain many more types down the road
to help with image filter implementations. Too many to make inner
classes like Context was, but once the namespace exists, it doesn't
make sense for Context to be the only inner class.
Bug: skia:9282
Change-Id: I4ed76bd5f6c8f0b8ca2f2e1c5d76006c7cccf91e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/234581
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Change-Id: I127c979670c3dc7dac2e35908a795afbdefca8f6
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/234902
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Change-Id: I0e6bc516eb43e7f7062f9d19e7ef8093324b551d
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/233997
Commit-Queue: Stephen White <senorblanco@chromium.org>
Reviewed-by: Greg Daniel <egdaniel@google.com>
Change-Id: I9c636c939dffbb0b395f3e9ed3d9a55784a6d390
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/233981
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
Use STDMETHODIMP because it's the right thing to do, use
COM_DECLSPEC_NOTHROW because STDMETHOD uses it but STDMETHODIMP does
not, leading to warnings from clang and maybe also from msvc, see
https://developercommunity.visualstudio.com/content/problem/325764/msvc-1582-generates-warning-c4986-in-atl-header-fi.html
This also removes SkBlockComRef, since even WRL has abandoned the
similar RemoveIUnknown. These classes were helpful in the transition to
using smart pointers, but are incompatible with final implementations.
Change-Id: I53a618ee629af638d9d8625ccd5acb0db6529950
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/233988
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Ben Wagner <bungeman@google.com>
This is a reland of fe19203eb7
Original change's description:
> Initiate regeneration of mipmaps from proxy DAG land
>
> This allows us to resolve all the textures before executing a command
> buffer, rather than interrupting execution as is currently done in the
> GL backend.
>
> Change-Id: I998546b312d24cf47fc2c837e7c94fbf95571af0
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/230636
> Commit-Queue: Chris Dalton <csmartdalton@google.com>
> Reviewed-by: Robert Phillips <robertphillips@google.com>
Change-Id: Ic904d0b1bcb451bcb74cfae2204fb7297eaff108
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/234016
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
The math isn't that complex and the existing code has intergralization of
the profile size that makes getting the texture coords correct tricky
(and currently wrong).
Bug: skia:9319
Change-Id: Ic928737ce4c40d28ee0696c3628e914f35ffe371
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/233985
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
I'm aware of two uses for cc_wrapper:
- Goma support for linking is experimental, and normally just falls
back to the local compiler.
- ccache site says it always falls back to the real compiler for
linking. https://ccache.dev/
The msvc toolchain already does not use cc_wrapper for linking; this
change removes it for gcc_like_toolchain. This matches Chromium, e.g.
2cde7f21ab/build/toolchain/gcc_toolchain.gni (221)
Change-Id: I7199c23343233a2435e4d28e7269f67efac36d72
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/234316
Commit-Queue: Mike Klein <mtklein@google.com>
Auto-Submit: Ben Wagner aka dogben <benjaminwagner@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Release and debug builds draw differently on my machine,
and if I tack #pragma clang optimize off onto cull_rect()
in SkDashPath.cpp, they both draw like a debug build.
I'm not very familiar with this code, so it may help my
diagnosis to see what other bots draw.
Bug: skia:9331
Change-Id: I1798503c0e5956ec767d8e0021ccca7a1d2a8ac3
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/233999
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
Previously YCbCr Vulkan samplers were supported only on Android for
external images, while Vulkan requires YCbCr sampler for I420 YUV image
formats such as VK_FORMAT_G8_B8_R8_3PLANE_420_UNORM,
VK_FORMAT_G8_B8R8_2PLANE_420_UNORM.
This CL:
- Adds VK_FORMAT_G8_B8_R8_3PLANE_420_UNORM and
VK_FORMAT_G8_B8R8_2PLANE_420_UNORM as supported Vulkan formats
- Updates GrVkYcbcrConversionInfo to add fFormat field and allow
fExternalFormat=0.
- Removes assertions format=VK_FORMAT_UNDEFINED for all images that
have ycbcr info.
Bug: chromium:981022
Change-Id: Id4d81b20d9fda4d9ad0831f77e6025eed3db2bfd
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/233776
Commit-Queue: Greg Daniel <egdaniel@google.com>
Reviewed-by: Greg Daniel <egdaniel@google.com>
Auto-Submit: Sergey Ulanov <sergeyu@chromium.org>
This reverts commit fe19203eb7.
Reason for revert: Turned a lot of bots red
Original change's description:
> Initiate regeneration of mipmaps from proxy DAG land
>
> This allows us to resolve all the textures before executing a command
> buffer, rather than interrupting execution as is currently done in the
> GL backend.
>
> Change-Id: I998546b312d24cf47fc2c837e7c94fbf95571af0
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/230636
> Commit-Queue: Chris Dalton <csmartdalton@google.com>
> Reviewed-by: Robert Phillips <robertphillips@google.com>
TBR=bsalomon@google.com,robertphillips@google.com,csmartdalton@google.com
Change-Id: Iaa8391b8fbd5a4f851ca41834c1dfeee4be6fedd
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/233836
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
This allows us to resolve all the textures before executing a command
buffer, rather than interrupting execution as is currently done in the
GL backend.
Change-Id: I998546b312d24cf47fc2c837e7c94fbf95571af0
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/230636
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Remove beginDraw()/endDraw() in favour of applyState().
This allows for varying primitive topology between meshes in a single
draw: call applyState() prior to drawing each mesh in case the
primitive type changed.
Initial support for onClear() and onClearStencil() in GrNXTGpuRTCommandBuffer.
Modify beginRenderPass() to take the LoadOps for color and stencil, so we
can encode render passes for stencil operations.
Add support for reverse-subtract blend equation.
Also apply options overrides in caps initialization.
Change-Id: I8547961c414187dbc763a67d6c7ec4383d6ce6f6
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/232136
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
This reverts commit cc4573dfed.
Reason for revert: Seems to be crashing in Win Intel vulkan bots
Original change's description:
> Enable YCbCr sampler support on platforms other than Android
>
> Previously YCbCr Vulkan samplers were supported only on Android for
> external images, while Vulkan requires YCbCr sampler for I420 YUV image
> formats such as VK_FORMAT_G8_B8_R8_3PLANE_420_UNORM,
> VK_FORMAT_G8_B8R8_2PLANE_420_UNORM.
> This CL:
> - Adds VK_FORMAT_G8_B8_R8_3PLANE_420_UNORM and
> VK_FORMAT_G8_B8R8_2PLANE_420_UNORM as supported Vulkan formats
> - Updates GrVkYcbcrConversionInfo to add fFormat field and allow
> fExternalFormat=0.
> - Removes assertions format=VK_FORMAT_UNDEFINED for all images that
> have ycbcr info.
>
> Bug: chromium:981022
> Change-Id: I3989f72e918a257a081b81ae20429a6771229ad6
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/230918
> Commit-Queue: Sergey Ulanov <sergeyu@chromium.org>
> Commit-Queue: Greg Daniel <egdaniel@google.com>
> Auto-Submit: Sergey Ulanov <sergeyu@chromium.org>
> Reviewed-by: Greg Daniel <egdaniel@google.com>
TBR=egdaniel@google.com,sergeyu@chromium.org
Change-Id: Ia9c229fbec4cd397fb04304753fa49f34336ed3a
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: chromium:981022
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/233563
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>
Previously YCbCr Vulkan samplers were supported only on Android for
external images, while Vulkan requires YCbCr sampler for I420 YUV image
formats such as VK_FORMAT_G8_B8_R8_3PLANE_420_UNORM,
VK_FORMAT_G8_B8R8_2PLANE_420_UNORM.
This CL:
- Adds VK_FORMAT_G8_B8_R8_3PLANE_420_UNORM and
VK_FORMAT_G8_B8R8_2PLANE_420_UNORM as supported Vulkan formats
- Updates GrVkYcbcrConversionInfo to add fFormat field and allow
fExternalFormat=0.
- Removes assertions format=VK_FORMAT_UNDEFINED for all images that
have ycbcr info.
Bug: chromium:981022
Change-Id: I3989f72e918a257a081b81ae20429a6771229ad6
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/230918
Commit-Queue: Sergey Ulanov <sergeyu@chromium.org>
Commit-Queue: Greg Daniel <egdaniel@google.com>
Auto-Submit: Sergey Ulanov <sergeyu@chromium.org>
Reviewed-by: Greg Daniel <egdaniel@google.com>
This reverts commit adfc22df86.
Reason for revert: breaks building DM in Android, probably Google3 too. Shame.
Original change's description:
> add a stub for running fiddles as GMs
>
> This is about the 15th time I've hacked this up.
>
> Change-Id: I0c75c0d04133d867e942c39eba693fd7227ba0f5
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/233240
> Reviewed-by: Brian Salomon <bsalomon@google.com>
> Commit-Queue: Mike Klein <mtklein@google.com>
TBR=mtklein@google.com,bsalomon@google.com
Change-Id: I56207630eeba4fb411111fd40e6bc532cef39407
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/233241
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
Nothing is cached in this initial version; all the magic happens in
GrDawnGpuRTCommandBuffer::beginDraw(). It builds the program,
sets the uniform data, converts the vertex attributes, initializes
rasterization state and blend state, creates a pipeline and bind group
and sets them.
send[Indexed]InstancedMeshToGpu() encodes the Dawn draw commands.
GrDawnGpuProgramBuilder::beginRenderPass() then finishes
the command encoder and submits the resulting command buffer.
Change-Id: Ie9c160cd18a2c4ef1bde66f86a52227e3bf1b570
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/230457
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
This is about the 15th time I've hacked this up.
Change-Id: I0c75c0d04133d867e942c39eba693fd7227ba0f5
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/233240
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
Maximum ring buffer size is based on the maxBufferLength of
the system. Also includes a sample for testing upload pressure.
Bug: skia:8243
Change-Id: I36863b725ba9cbcb02d575b2ab9695d186fcd529
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/232759
Reviewed-by: Greg Daniel <egdaniel@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Jim Van Verth <jvanverth@google.com>
The current use of MMD appears to ignore all includes which are resolved
from pointy bracket includes or system paths. This can lead to very
interesting incomplete rebuilds (or no rebuild at all) when updating
third_party headers or when the system headers are updated. The use of
MD makes the build depend on all the headers transitively included no
matter how they were included. This greatly increases the dependencies,
but only so much as to make them correct (barring the ever difficult
issue of a file being created with the same name earlier in the search
path).
Change-Id: I852e617f23b4ff0c5b4eb3068455428c9568e078
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/232762
Commit-Queue: Ben Wagner <bungeman@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Adds a GrTextureResolveManager class and plumbs it through calls to
generate the drawing manager's dependency DAG. This new class is
currently unimplemented, but it wraps GrDrawingManager and will
eventually give limited access to functionality for making new tasks
that regenerate mipmaps and/or resolve MSAA. We will use this object
to move mipmap generation up to the DAG/proxy level.
Bug: skia:
Change-Id: Ie1deef21e7ae579a0262f2eeb93d451f0740d823
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/232633
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
2x faster than calling drawVertices as the impl.
Lots more to do in future CLs
- much of the time is spent in malloc, as we cons-up private shaders.
we *could* create a private shader and wack its data for each draw
(breaking the immutable contract, but that may be ok for a raster-only
internal-use shader...)
- also spend time building the pipeline for each draw, even though
all that has changed is a color payload (and the ctm). A custome
stage(s) that exposed its private data could be reused with new
data...
Bug: skia:
Change-Id: I0ff15155e37c0af7931abd34c0883701a47a048a
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/203168
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Mike Reed <reed@google.com>
This is a more abstract representation of something that can
participate in the DAG by modifying a GrSurfaceProxy's pixels. This
will serve as the direct base class for future DAG participants that
aren't opLists, and help to eventually remove GrTextureOpList.
Bug: skia:
Change-Id: Ia903c0188e0f6a6602781b90ee8da87ba3cc13f8
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/231277
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Conversion of SkSL to (Dawn-friendly) SPIR-V.
Conversion of GlBlendCoeff -> dawn::BlendFactor.
Conversion of GlBlendEquation -> dawn::BlendOperation.
Creation of dawn::ColorState from color format, write mask, blend state.
Building of basic BindGroupLayouts and BindGroups (the equivalent of
Vulkan's DescriptorSets).
Handling the RTAdjustment uniform, which handles the viewing transform.
Change-Id: If8aef19d71f81bc805defcf5fd7c4ebe3b9547db
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/231177
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
In preparation for moving the filter-specific headers from include/effects and into src/effects/imagefilters, this exposes a more convenient API where all filters provided by Skia are accessible from the same factory.
It also only exposes the factories that take the shared enums (SkColorChannel, SkTileMode), so the per-filter enums can be phased out (SkBlurImageFilter::TileMode, etc.). It also accepts an SkIRect* instead of an SkImageFilter::CropRect* since all uses seem to be for integer rectangles and flags are either always on or always off. This will allow us to later remove the CropRect type.
Bug: skia:9280
Change-Id: I71f77ff44b573f87b837d6523b9eb565521f0f36
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/230876
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
I'm having a hard time implementing SkRasterPipeline with SkVM,
so maybe it makes sense to try stepping over it instead?
Have not done much timing, except that it's noticeably slower
to generate GMs with this blitter turned on than turned off:
use of uniforms is forcing us into the interpreter.
I decided to start off with a perfect div255()... I'm curious to
see if we can live with perfect instead of an approximation.
Change-Id: I752f4a32ebe32ca550c9762aaad9ad86670b79d5
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/231117
Commit-Queue: Mike Klein <mtklein@google.com>
Reviewed-by: Mike Reed <reed@google.com>
This will decrease our reliance on GrPixelConfig for testing.
Bug: skia:6718
Change-Id: I951f57cce229792d994391dbba74dd206bbfb923
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/230893
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
Also add GrDawnCaps to friends of GrShaderCaps, as the other backends do.
Change-Id: Iae60cae7168ad9194f7808c2f51f41de10e53071
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/230747
Commit-Queue: Stephen White <senorblanco@chromium.org>
Reviewed-by: Greg Daniel <egdaniel@google.com>
Change-Id: I49a233a28d16622b286ffc4150446b33f899a891
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/230557
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
Change-Id: I76824eca6e72505004b4dfb03fc35c2849176ecb
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/230459
Commit-Queue: Stephen White <senorblanco@chromium.org>
Reviewed-by: Greg Daniel <egdaniel@google.com>
We can't remove the loopers themselves, as they are still used
by android and chrome (they just don't ever pass them to skia).
Eventually each of those clients will resolve this, but for now
we just keep the classes (and tests) in skia.
Bug: skia:4783
Change-Id: I5f507e6bb82280f2bc7c0b21eebe59c287aa9265
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/230579
Reviewed-by: Florin Malita <fmalita@chromium.org>
Commit-Queue: Mike Reed <reed@google.com>
