This reverts commit 4926b07217.
Reason for revert: fix wip
Original change's description:
> Revert "Improve scissor state tracking in GrRTC"
>
> This reverts commit 3b923a880b.
>
> Reason for revert: GrAppliedHardClip isn't tracking scissor state properly
>
> Original change's description:
> > Improve scissor state tracking in GrRTC
> >
> > At a low level, this changes GrScissorState from a rect+bool to a rect+size.
> > The scissor test is considered enablebd if the rect does not fill the
> > device bounds rect specified by the size. This has a number of benefits:
> >
> > 1. We can always access the scissor rect and know that it will be
> > restricted to the render target dimensions.
> > 2. It helps consolidate code that previously had to test the scissor rect
> > and render target bounds separately.
> > 3. The clear operations can now match the proper backing store dimensions
> > of the render target.
> > 4. It makes it easier to reason about scissors applying to the logical
> > dimensions of the render target vs. its backing store dimensions.
> >
> > Originally, I was going to have the extra scissor guards for the logical
> > dimensions be added in a separate CL (with the cleanup for
> > attemptQuadOptimization). However, it became difficult to ensure correct
> > behavior respecting the vulkan render pass bounds without applying this
> > new logic at the same time.
> >
> > So now, with this CL, GrAppliedClips are sized to the backing store
> > dimensions of the render target. GrOpsTasks also clip bounds to the
> > backing store dimensions instead of the logical dimensions (which seems
> > more correct since that's where the auto-clipping happens). Then when
> > we convert a GrClip to a GrAppliedClip, the GrRTC automatically enforces
> > the logical dimensions scissor if we have stencil settings (to ensure
> > the padded pixels don't get corrupted). It also may remove the scissor
> > if the draw was just a color buffer update.
> >
> > Change-Id: I75671c9cc921f4696b1dd5231e02486090aa4282
> > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/290654
> > Commit-Queue: Michael Ludwig <michaelludwig@google.com>
> > Reviewed-by: Brian Salomon <bsalomon@google.com>
>
> TBR=bsalomon@google.com,csmartdalton@google.com,michaelludwig@google.com
>
> Change-Id: Ie98d084158e3a537604ab0fecee69bde3e744d1b
> No-Presubmit: true
> No-Tree-Checks: true
> No-Try: true
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/294340
> Reviewed-by: Michael Ludwig <michaelludwig@google.com>
> Commit-Queue: Michael Ludwig <michaelludwig@google.com>
TBR=bsalomon@google.com,csmartdalton@google.com,michaelludwig@google.com
# Not skipping CQ checks because this is a reland.
Change-Id: I2116e52146890ee4b7ea007f3c3d5c3e532e4bdd
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/294257
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
This reverts commit 3b923a880b.
Reason for revert: GrAppliedHardClip isn't tracking scissor state properly
Original change's description:
> Improve scissor state tracking in GrRTC
>
> At a low level, this changes GrScissorState from a rect+bool to a rect+size.
> The scissor test is considered enablebd if the rect does not fill the
> device bounds rect specified by the size. This has a number of benefits:
>
> 1. We can always access the scissor rect and know that it will be
> restricted to the render target dimensions.
> 2. It helps consolidate code that previously had to test the scissor rect
> and render target bounds separately.
> 3. The clear operations can now match the proper backing store dimensions
> of the render target.
> 4. It makes it easier to reason about scissors applying to the logical
> dimensions of the render target vs. its backing store dimensions.
>
> Originally, I was going to have the extra scissor guards for the logical
> dimensions be added in a separate CL (with the cleanup for
> attemptQuadOptimization). However, it became difficult to ensure correct
> behavior respecting the vulkan render pass bounds without applying this
> new logic at the same time.
>
> So now, with this CL, GrAppliedClips are sized to the backing store
> dimensions of the render target. GrOpsTasks also clip bounds to the
> backing store dimensions instead of the logical dimensions (which seems
> more correct since that's where the auto-clipping happens). Then when
> we convert a GrClip to a GrAppliedClip, the GrRTC automatically enforces
> the logical dimensions scissor if we have stencil settings (to ensure
> the padded pixels don't get corrupted). It also may remove the scissor
> if the draw was just a color buffer update.
>
> Change-Id: I75671c9cc921f4696b1dd5231e02486090aa4282
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/290654
> Commit-Queue: Michael Ludwig <michaelludwig@google.com>
> Reviewed-by: Brian Salomon <bsalomon@google.com>
TBR=bsalomon@google.com,csmartdalton@google.com,michaelludwig@google.com
Change-Id: Ie98d084158e3a537604ab0fecee69bde3e744d1b
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/294340
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
At a low level, this changes GrScissorState from a rect+bool to a rect+size.
The scissor test is considered enablebd if the rect does not fill the
device bounds rect specified by the size. This has a number of benefits:
1. We can always access the scissor rect and know that it will be
restricted to the render target dimensions.
2. It helps consolidate code that previously had to test the scissor rect
and render target bounds separately.
3. The clear operations can now match the proper backing store dimensions
of the render target.
4. It makes it easier to reason about scissors applying to the logical
dimensions of the render target vs. its backing store dimensions.
Originally, I was going to have the extra scissor guards for the logical
dimensions be added in a separate CL (with the cleanup for
attemptQuadOptimization). However, it became difficult to ensure correct
behavior respecting the vulkan render pass bounds without applying this
new logic at the same time.
So now, with this CL, GrAppliedClips are sized to the backing store
dimensions of the render target. GrOpsTasks also clip bounds to the
backing store dimensions instead of the logical dimensions (which seems
more correct since that's where the auto-clipping happens). Then when
we convert a GrClip to a GrAppliedClip, the GrRTC automatically enforces
the logical dimensions scissor if we have stencil settings (to ensure
the padded pixels don't get corrupted). It also may remove the scissor
if the draw was just a color buffer update.
Change-Id: I75671c9cc921f4696b1dd5231e02486090aa4282
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/290654
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
This is a prerequisite change for the upcoming sample(child, matrix)
function. By itself, this CL doesn't really change anything; it just
adds an ownership tracking feature which sample(child, matrix) depends
on.
Change-Id: I98b12e5fb062a2535af367931e7a932ea9c63a59
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/281337
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Fuses the relevant functionality into the set() calls.
Change-Id: I3a764219425c35967942366c60a6b9ee6ec48e11
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/276836
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
This CL also incidentally adds:
1) a GrMeshDrawOp::Target 'outputView' virtual and switches GrOpFlushState over to overriding it.
2) a createProgramInfo helper to GrSimpleMeshDrawOpHelper
Bug: skia:9455
Change-Id: I88ce51c585b1458ee79a9aaa7024190e9f19198c
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/274506
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
This reverts commit d3606518fa.
Reason for revert: DDL3 bots are red
Original change's description:
> Add onPrePrepareDraws & createProgramInfo methods to GrDrawVerticesOp
>
> This CL also incidentally adds:
>
> 1) a GrMeshDrawOp::Target 'outputView' virtual and switches GrOpFlushState over to overriding it.
>
> 2) a createProgramInfo helper to GrSimpleMeshDrawOpHelper
>
> Bug: skia:9455
> Change-Id: Iecd712d3ac76038651bd2e0512134e310930d527
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/274551
> Reviewed-by: Greg Daniel <egdaniel@google.com>
> Commit-Queue: Robert Phillips <robertphillips@google.com>
TBR=egdaniel@google.com,robertphillips@google.com
Change-Id: I6e44ba3bc47df27479af9344af946e27f0a2e937
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:9455
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/274505
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
This CL also incidentally adds:
1) a GrMeshDrawOp::Target 'outputView' virtual and switches GrOpFlushState over to overriding it.
2) a createProgramInfo helper to GrSimpleMeshDrawOpHelper
Bug: skia:9455
Change-Id: Iecd712d3ac76038651bd2e0512134e310930d527
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/274551
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
Each such op now has a base createProgramInfo method along with a helper version that just takes a flushState.
This CL also makes GrSimpleMeshDrawOpHelper::CreateProgramInfo a chokepoint for programInfo allocation.
Bug: skia:9455
Change-Id: Ibbf0e9cd6f24453e1b87a5e072a30248811a1237
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/274054
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
This is just a simple clean up from a CL request.
Change-Id: Ibc09e19aeb8a160e4e4e03573fcf56cb1d58525b
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/273983
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
Clients now must call bindPipeline() before drawing.
Also renames GrOpsRenderPass::draw() to drawMeshes(), in order to
ensure every call site gets updated. drawMeshes() will soon be
replaced by individual calls for each draw type (indexed, instanced,
indexed-patterned, indirect, etc.).
Change-Id: I93ef579ded7d0048c5aa1bf1d7c0eb7bc1cd27b2
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/270424
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
When iterating over the coord transforms or texture samplers of a
FP also have access to the owning FP.
Pass a coord transform range to GPs rather than a pointer to an
iterator.
Change-Id: If7c829a67dce6600d7f49e12d6f49f685dcace3a
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256216
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
This is complete overkill for what these GMs require but it will help centralize things.
Change-Id: If30cbd9a9cfc8fcc1fe96fc9ca1b4cb17cdeb4bd
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/255824
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
As the first real op to be updated this required the plumbing of the DstProxyView and the de-constification of the GrAppliedClip.
Bug: skia:9455
Change-Id: Ieaa16adfa1d114021b2ad60d0ef8ecbe31d9cc82
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/255136
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
This CL also introduces creating the geometryProcessor in the arena (which bypasses the GrProcessor memory pool).
Bug: skia:9455
Change-Id: Ia9d5c0d886ea918d9e4209078a1f7e8595234d0e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/254963
Commit-Queue: Robert Phillips <robertphillips@google.com>
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Hopefully this makes it clearer that the GrProgramInfo isn't copying these
objects.
Bug: skia:9455
Change-Id: Id43b7978d63b6d53d531601f3446745f33afcf2d
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/254970
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
Replaces numSamples with numRasterSamples, and adds isMixedSampled.
The sample count that vulkan and metal actually want to know is how
many samples the rasterizer will compute, which may not match the
number of samples in the render target when we have mixed samples.
They will also need to know whether a program is mixed sampled in
order to set up coverage modulation.
Change-Id: I133c11f74b7dc6a7580818ef73d6deec1d201b64
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/253550
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
This further consolidates the information required to compute the program key (esp. for Vulkan). This CL mainly comprises the plumbing portion - a follow up CL will actually use it.
Bug: skia:9455
Change-Id: Iaac716c289916981a1757a333bfa57b3051fd35b
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/252161
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
This reverts commit 63a3873957.
Reason for revert: fail
Original change's description:
> Reduce inputs to GrProgramDesc::Build method(s)
>
> This moves the computation of the GrStencilSettings into GrProgramDesc::Build for the Vulkan backend
>
> This makes it clear that the stencilSettings are determined by the GrProgramInfo.
>
> Note that this does change some behavior in that the number of stencil bits is now determined (and set up for) the number requested not the number actually available on the attached stencil buffer at flush time.
>
> Bug: skia:9455
> Change-Id: Ia5fa4e30e34952c1a492b8d571094abf2291b2a4
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/248696
> Reviewed-by: Greg Daniel <egdaniel@google.com>
> Commit-Queue: Robert Phillips <robertphillips@google.com>
TBR=egdaniel@google.com,robertphillips@google.com
Change-Id: Ib0b337c19acea8000b5592f3dfeda181fd6ef157
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:9455
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/249222
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
This moves the computation of the GrStencilSettings into GrProgramDesc::Build for the Vulkan backend
This makes it clear that the stencilSettings are determined by the GrProgramInfo.
Note that this does change some behavior in that the number of stencil bits is now determined (and set up for) the number requested not the number actually available on the attached stencil buffer at flush time.
Bug: skia:9455
Change-Id: Ia5fa4e30e34952c1a492b8d571094abf2291b2a4
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/248696
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
This improves the encapsulation of the GrProgramInfo class.
Bug: skia:9455
Change-Id: Ic241ba4a2c2edea3e774522065371e5ed8cc2da3
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/247438
Commit-Queue: Robert Phillips <robertphillips@google.com>
Reviewed-by: Greg Daniel <egdaniel@google.com>
This is the first step in moving the marshaling of program information earlier in renderTask processing (i.e., to onPrePrepare).
Change-Id: I91e3baed9a128e845bd32f9dbbacd9b21d852a3d
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/244118
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
Since we want to collect shader information early, this attempts to make clear that numSamples, origin and outputSwizzle can all be obtained from the target renderTargetProxy.
Change-Id: I42e0fd79e2163f17673ccdd344a31fbaadac5f53
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/246298
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
Need for a follow on change, we need all ops to do all their prep work
which includes creating needed buffers before we start calling
execute on ops. This change updates the few ops that were creating their
buffers in onExecute.
Change-Id: I0f43b0cc35359e1e7ccfb3b159e79bdfe288dcb1
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/239436
Reviewed-by: Chris Dalton <csmartdalton@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>
From now on, sample counts always refer to the number of actual color
samples, and render targets don't have separate color and stencil
sample counts.
If mixed samples support is available when making a
"GrAAType::kCoverage" draw, then an op may attach and use a mixed
sampled stencil buffer internally. But this will all be invisible to
the client.
After this CL, we temporarily won't have a mode to use nvpr with mixed
samples. That will soon be fixed by a follow-on CL that enables nvpr
with mixed samples in the normal "gl" and "gles" configs.
Bug: skia:
Change-Id: I1cb8277f0d2d0d371f24bb9f39cd473ed5c5c83b
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/221878
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
This almost gets gms to be iwyu clean. The last bit is around gm.cpp
and the tracing framework and its use of atomic. Will also need a way
of keeping things from regressing, which is difficult due to needing to
do this outside-in.
Change-Id: I1393531e99da8b0f1a29f55c53c86d53f459af7d
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/211593
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Ben Wagner <bungeman@google.com>
Current strategy: everything from the top
Things to look at first are the manual changes:
- added tools/rewrite_includes.py
- removed -Idirectives from BUILD.gn
- various compile.sh simplifications
- tweak tools/embed_resources.py
- update gn/find_headers.py to write paths from the top
- update gn/gn_to_bp.py SkUserConfig.h layout
so that #include "include/config/SkUserConfig.h" always
gets the header we want.
No-Presubmit: true
Change-Id: I73a4b181654e0e38d229bc456c0d0854bae3363e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/209706
Commit-Queue: Mike Klein <mtklein@google.com>
Reviewed-by: Hal Canary <halcanary@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Reviewed-by: Florin Malita <fmalita@chromium.org>
sk_tool_utils doesn't really fit the naming convention
the rest of code under tools/ tends to use.
Change-Id: I45326a174101c6eb4b6149e9c742f658f2fd23b1
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/202313
Auto-Submit: Mike Klein <mtklein@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Previously, we relied on ops to deduce whether a draw would have
hardware coverage modulation as a result mixed samples. This is
problematic because *any* draw can have mixed samples coverage if
there is a multisampled stencil clip. No ops were checking for stencil
clip, and most just said they never used mixed samples.
Now that the only usecase for mixed samples is the stencil buffer,
this CL makes the processorSet automatically deduce mixed samples
coverage from the stencil settings and fsaaType.
Bug: skia:
Change-Id: Ib69b84bc03b12f6efb8e7d6ed721ae1612785315
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/197281
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
This is useful, for example, when trying to test a hardware feature
that isn't supported in the current context.
Bug: skia:8731
Change-Id: I9a363159300c92e4039bfd05400238c27002efb1
Reviewed-on: https://skia-review.googlesource.com/c/189133
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
This is a reland of f5efa78902
Original change's description:
> Add a GpuGM class for GPU-only tests
>
> Bug: skia:8731
> Change-Id: Ic81f21621c0183bb356601eb663fa1ee334bca55
> Reviewed-on: https://skia-review.googlesource.com/c/189134
> Reviewed-by: Brian Osman <brianosman@google.com>
> Reviewed-by: Mike Klein <mtklein@google.com>
> Commit-Queue: Chris Dalton <csmartdalton@google.com>
TBR=mtklein@google.com
Bug: skia:8731
Change-Id: Ia322d7e9280b67ebea1edbefc820cae72cd9c3a6
Reviewed-on: https://skia-review.googlesource.com/c/190148
Reviewed-by: Chris Dalton <csmartdalton@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
GrBuffer is a base class for GrGpuBuffer and GrCpuBuffer. GrGpuBuffer is a
GrGpuResource and the others are not. This allows GrCpuBuffers to exist
outside of the GrGpuResourceCache.
Also removes flags from GrResourceProvider buffer factory function. The
only flag still in use was kRequireGpuMemory. Now CPU buffers are made
without using GrResourceProvider.
Change-Id: I82670d1316e28fd6331ca36b26c8c4ead33846f9
Reviewed-on: https://skia-review.googlesource.com/c/188823
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Bug: skia:8731
Change-Id: Ic81f21621c0183bb356601eb663fa1ee334bca55
Reviewed-on: https://skia-review.googlesource.com/c/189134
Reviewed-by: Brian Osman <brianosman@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
This is a reland of 91c1d08bc3
Original change's description:
> SkSL is now pickier about type conversions
>
> Bug: skia:
> Change-Id: I4e8b8f229f4e4344f160b0dbb41832764d0b75bd
> Reviewed-on: https://skia-review.googlesource.com/c/188311
> Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
> Reviewed-by: Brian Salomon <bsalomon@google.com>
Bug: skia:
Change-Id: I727cad061afc0a5ee6f4d2df789330d809dd110a
Reviewed-on: https://skia-review.googlesource.com/c/189643
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Renamed to GrGpuBufferType in anticipation of splitting GrBuffer
into GrGpuBuffer and GrCpuBuffer types.
There were two unused values in the enum that are removed, DrawIndirect
and Texel.
Change-Id: Icb6b3da689adbd8e10495c10fd0470a6ee0120b5
Reviewed-on: https://skia-review.googlesource.com/c/189280
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Mechanical. This makes the priv() accessor the same for all the context types.
Change-Id: I40850eb05a33b8d7cc3eabdd42226d24b2ba58aa
Reviewed-on: https://skia-review.googlesource.com/c/189164
Commit-Queue: Robert Phillips <robertphillips@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Bug: skia:8731
Change-Id: If73216bd427a1ce773fa41044a45c1bbd7ea08e9
Reviewed-on: https://skia-review.googlesource.com/c/189124
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
