This replaces GrXferProcessor::getOptimizations with a new function on GrXPFactory. The results are made available via FragmentProcessorAnalysis.
Bug: skia:
Change-Id: I535985458c9d13ad858cac94e957e2fdbe332036
Reviewed-on: https://skia-review.googlesource.com/10218
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Assign names that indicate that they aren't just for the input phase since I plan to use them at the boundary between FPs and XPs as well.
Renamed GrProcOptInfo to GrColorFragmentProcessorAnalysis. This is now only used on the color side and the new name seems clearer to me.
Change GrMeshDrawOp::getFragmentProcessorAnalysisInputs to use the new color/coverage types directly rather than a class that has been reduced to simply bundling them together.
Change-Id: If93bae74c9d590486eecdf63f302418c96deab65
Reviewed-on: https://skia-review.googlesource.com/10161
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Greg Daniel <egdaniel@google.com>
This reverts commit 2bf4b3a97b.
Reason for revert: nanobench assertion
Original change's description:
> Remove GrPipeline from GrDrawOp.
>
> GrDrawOp subclasses are now free to construct their pipelines at flush time and now in theory could use multiple GrPipelines for multipass rendering.
>
> GrProcessorSet may be used to retain the processors from a GrPaint with "pending execution" style refs.
>
> NVPR and Instanced rendering are updated to create their pipelines at flush time without a GrPipelineBuilder.
>
> The monolithic pipeline creation/management that was on GrDrawOp is moved to GrMeshDrawOp. However, this is temporary and will be removed in coming changes.
>
> Change-Id: I124282e3cea5d070970b5460c8a679fcaf7a8eff
> Reviewed-on: https://skia-review.googlesource.com/7279
> Commit-Queue: Brian Salomon <bsalomon@google.com>
> Reviewed-by: Robert Phillips <robertphillips@google.com>
>
TBR=bsalomon@google.com,robertphillips@google.com,csmartdalton@google.com,reviews@skia.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Change-Id: I1bc64f6cbbd5f482417637a034342c2b5371dc5c
Reviewed-on: https://skia-review.googlesource.com/9817
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
GrDrawOp subclasses are now free to construct their pipelines at flush time and now in theory could use multiple GrPipelines for multipass rendering.
GrProcessorSet may be used to retain the processors from a GrPaint with "pending execution" style refs.
NVPR and Instanced rendering are updated to create their pipelines at flush time without a GrPipelineBuilder.
The monolithic pipeline creation/management that was on GrDrawOp is moved to GrMeshDrawOp. However, this is temporary and will be removed in coming changes.
Change-Id: I124282e3cea5d070970b5460c8a679fcaf7a8eff
Reviewed-on: https://skia-review.googlesource.com/7279
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
We will be storing GrAppliedClips alongside ops. The op already stores the
clipped bounds. If GrAppliedClip has draw bounds then as ops combine the
GrAppliedClip's bounds should be merged to be consistent. However, we won't
actually ever use those bounds again so it would be wasteful to merge them.
Change-Id: I4ef3010dc04761e256120a2e0e074bc3c6ff6ca1
Reviewed-on: https://skia-review.googlesource.com/9642
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Chris Dalton <csmartdalton@google.com>
This compacts the object so that it is more efficient for ops to store it.
It also adds a new constructor and query that will allow ops to use the analysis to also store the GrPaint's color.
This has the side effect of limiting the number of color processors on a GrProcessorSet to 64K which is just under 64K more than should ever be needed.
Change-Id: I4e6bc8e3f81bb2ff6a73af685beb6fb928a3de67
Reviewed-on: https://skia-review.googlesource.com/8972
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
The only use case for this was using the blend constant for LCD text. Now instead of overriding the op's color with an alpha we upload the alpha as a uniform.
This also removes two unused parameters from GrXferProcessor::getOptimizations.
Change-Id: I8268da9904a5d26649c6ae81a5705b0930893904
Reviewed-on: https://skia-review.googlesource.com/9221
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Greg Daniel <egdaniel@google.com>
Combine texture provider and resource provider
Largely mechanical. Only three places that were calling createApprox
via texture provider (ie without flags), so that was simple.
BUG=skia:
Change-Id: I876367bcdc6a8db736deedab1028de1972015509
Reviewed-on: https://skia-review.googlesource.com/9176
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
This seems like extra complexity for a impractical case. Also, if this is important a lot more work could be saved by catching this upstack (e.g. SkCanvas or SkPaintToGrPaint).
Change-Id: Ib47be9f3cdc8ce9e5b12d9e9eac5266f04c337a9
Reviewed-on: https://skia-review.googlesource.com/8949
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
It is renamed to FragmentProcessorAnalysis since it represents the outputs of the final FPs.
It now stores the analysis results that are subsequently needed rather than exposing GrProcOptInfo.
GrProcOptInfo is now only used on color FPs (not coverage).
Miscellaneous related renamings.
Change-Id: I95c518a7a76df6dc294a9fa67c611f8f653247bc
Reviewed-on: https://skia-review.googlesource.com/8534
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
This was the only non-test user of GrXPFactory::isConstantPreCoverageBlendedColor which is now removed.
Change-Id: Ic1c130d30a44e1a955b89f8912433a3c9df2e61e
Reviewed-on: https://skia-review.googlesource.com/8776
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
We don't use these anywhere downstream except to check for opaqueness.
Change-Id: I897137135d69004ed45c0f4c1e7297183f49fc6d
Reviewed-on: https://skia-review.googlesource.com/8402
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Greg Daniel <egdaniel@google.com>
This doesn't yet delete the old virtuals.
It still uses the color and component flags model for the pipeline input and blended output but I'm planning to change those as well.
Change-Id: I64e2ec0fe9ed9fae3aabf1ca8c9bc0582fc7565a
Reviewed-on: https://skia-review.googlesource.com/7760
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Greg Daniel <egdaniel@google.com>
Removes unused single channel tracking.
Makes it so that only the op/gp can initiate lcd coverage.
Makes GrProcOptInfo fragment processor analysis continuable.
Change-Id: I003a8aa3836bb64d04b230ddee581dc500e613a9
Reviewed-on: https://skia-review.googlesource.com/7039
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
This reverts commit a8f80de2bc.
Reason for revert: nanobench failing on windows bots, possibly others
Change-Id: Iacb8c650064a28654c165665be057377ffb02ba5
Reviewed-on: https://skia-review.googlesource.com/6802
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
All GrXPFactory instances are static constexpr.
Change-Id: If1086b08534166201e53b3fd9379104e361eb5e6
Reviewed-on: https://skia-review.googlesource.com/6701
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Removes the feedback to GrDrawOp via GrPipelineOptimizations.
Change-Id: I3cb17cad41779af292a92385fcd5ac23ae5a1ffd
Reviewed-on: https://skia-review.googlesource.com/6561
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Change-Id: I80f951976558a284e55386e0a368f08bd835d8ca
Reviewed-on: https://skia-review.googlesource.com/6359
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Change-Id: I6410eae41f051ce38bef6f38d670924c3483c325
Reviewed-on: https://skia-review.googlesource.com/6163
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Change-Id: I3ebe5a471477ce1b71c150b0bde4982d113fd8a7
Reviewed-on: https://skia-review.googlesource.com/5468
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Change-Id: I27b6324f8040899fafeda23ca524bc54a4dbf090
Reviewed-on: https://skia-review.googlesource.com/5392
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
This also makes the required changed to src, tests, and tools. The few
public APIs modified by this change appear to be unused outside of Skia.
Removing these from the public API makes it easier to ensure users are
no longer using them.
This also updates GrGpu::wrapBackendXXX and the
::onWrapBackendXXX methods to clarify ownership.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2448593002
Review-Url: https://codereview.chromium.org/2448593002
Use DEF_GPUTEST_FOR_*_CONTEXT macros to obtain the
test GPU context.
Makes changing the context -related classes easier,
since not all tests need to be changed.
BUG=skia:2992
Review URL: https://codereview.chromium.org/1448873002
Renames getInvariantOutput to getInvariantBlendedColor on GrXPFactory
and redefines it to not account for coverage conflation. This is the
information that all the callsites actually wanted to know.
BUG=skia:
Review URL: https://codereview.chromium.org/1161273005
Removes the runtime logic used by PorterDuffXferProcessor to decide
blend coeffs and shader outputs, and instead uses a compile-time
constant table of pre-selected blend formulas. Separates out the dst
read fallback into its own XP.
Introduces a new blend strategy for srcCoeff=0 that can apply coverage
with a reverse subtract blend equation instead of dual source
blending.
Adds new macros in GrBlend.h to analyze blend formulas both runtime.
Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no
longer used.
Adds a GM that verifies all xfermodes, including arithmetic, with the
color/coverage invariants used by Porter Duff.
Adds a unit test that verifies each Porter Duff formula with every
color/coverage invariant.
Major changes:
* Uses a reverse subtract blend equation for coverage when srcCoeff=0
(clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't
support dual source blending no longer require a dst copy for
dst-in and modulate.
* Sets BlendInfo::fWriteColor to false when the blend does not modify
the dst. GrGLGpu will now use glColorMask instead of blending for
these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]).
* Converts all SA blend coeffs to One for opaque inputs, and ISA to
Zero if there is also no coverage. (We keep ISA around when there
is coverage because we use it to tweak alpha for coverage.)
* Abandons solid white optimizations for the sake of simplicity
(screen was the only mode that previous had solid white opts).
Minor differences:
* Inconsequential differences in opt flags (e.g. we now return
kCanTweakAlphaForCoverage_OptFlag even when there is no coverage).
* Src coeffs when the shader outputs 0.
* IS2C vs IS2A when the secondary output is scalar.
BUG=skia:
Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414
Review URL: https://codereview.chromium.org/1124373002
Reason for revert:
Blocking DEPS roll into Chromium. Crashing virtual/gpu/fast/canvas/canvas-composite-*.html tests with the assert
../../third_party/skia/src/gpu/gl/builders/GrGLFragmentShaderBuilder.cpp:281: failed assertion "k110_GrGLSLGeneration != gpu->glslGeneration() || fOutputs.empty()"
Original issue's description:
> Implement Porter Duff XP with a blend table
>
> Removes the runtime logic used by PorterDuffXferProcessor to decide
> blend coeffs and shader outputs, and instead uses a compile-time
> constant table of pre-selected blend formulas.
>
> Introduces a new blend strategy for srcCoeff=0 that can apply coverage
> with a reverse subtract blend equation instead of dual source
> blending.
>
> Adds new macros in GrBlend.h to analyze blend formulas both runtime.
>
> Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no
> longer used.
>
> Adds a GM that verifies all xfermodes, including arithmetic, with the
> color/coverage invariants used by Porter Duff.
>
> Adds a unit test that verifies each Porter Duff formula with every
> color/coverage invariant.
>
> Major changes:
>
> * Uses a reverse subtract blend equation for coverage when srcCoeff=0
> (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't
> support dual source blending no longer require a dst copy for
> dst-in and modulate.
>
> * Sets BlendInfo::fWriteColor to false when the blend does not modify
> the dst. GrGLGpu will now use glColorMask instead of blending for
> these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]).
>
> * Converts all SA blend coeffs to One for opaque inputs, and ISA to
> Zero if there is also no coverage. (We keep ISA around when there
> is coverage because we use it to tweak alpha for coverage.)
>
> * Abandons solid white optimizations for the sake of simplicity
> (screen was the only mode that previous had solid white opts).
>
> Minor differences:
>
> * Inconsequential differences in opt flags (e.g. we now return
> kCanTweakAlphaForCoverage_OptFlag even when there is no coverage).
>
> * Src coeffs when the shader outputs 0.
>
> * IS2C vs IS2A when the secondary output is scalar.
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414TBR=egdaniel@google.com,bsalomon@google.com,cdalton@nvidia.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Review URL: https://codereview.chromium.org/1153993002
Removes the runtime logic used by PorterDuffXferProcessor to decide
blend coeffs and shader outputs, and instead uses a compile-time
constant table of pre-selected blend formulas.
Introduces a new blend strategy for srcCoeff=0 that can apply coverage
with a reverse subtract blend equation instead of dual source
blending.
Adds new macros in GrBlend.h to analyze blend formulas both runtime.
Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no
longer used.
Adds a GM that verifies all xfermodes, including arithmetic, with the
color/coverage invariants used by Porter Duff.
Adds a unit test that verifies each Porter Duff formula with every
color/coverage invariant.
Major changes:
* Uses a reverse subtract blend equation for coverage when srcCoeff=0
(clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't
support dual source blending no longer require a dst copy for
dst-in and modulate.
* Sets BlendInfo::fWriteColor to false when the blend does not modify
the dst. GrGLGpu will now use glColorMask instead of blending for
these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]).
* Converts all SA blend coeffs to One for opaque inputs, and ISA to
Zero if there is also no coverage. (We keep ISA around when there
is coverage because we use it to tweak alpha for coverage.)
* Abandons solid white optimizations for the sake of simplicity
(screen was the only mode that previous had solid white opts).
Minor differences:
* Inconsequential differences in opt flags (e.g. we now return
kCanTweakAlphaForCoverage_OptFlag even when there is no coverage).
* Src coeffs when the shader outputs 0.
* IS2C vs IS2A when the secondary output is scalar.
BUG=skia:
Review URL: https://codereview.chromium.org/1124373002
