Cause DeferredTextureImageData functionality to support low bit depth
(4444, 565) image formats (with dithering).
Bug: 720105
Change-Id: Ie3b5768ebc393d9b0a5322461c722bf37c80b791
Reviewed-on: https://skia-review.googlesource.com/18945
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Eric Karl <ericrk@chromium.org>
SkCodec sets fRequiredFrame to be the earliest possible frame that a
given frame can depend on. e.g.
- Frame A fills the screen, Keep
- Frame B does not cover A, Keep
- Frame C covers B but not A, and is opaque
Frame C can depend on either A or B. SkCodec already reports that C
depends on A. This CL allows a client of SkCodec to use either A or
B to create C.
Also expose the DisposalMethod. Since any frame between A and C can
be used to create C except for DisposePrevious frames, the client
needs to be able to know the disposal method so they do not try to
use such a frame to create C.
Further, the disposal method can be used to give the client a better
idea whether they will continue to need a frame. (e.g. if frame i is
DisposePrevious and depends on i-1, the client may not want to steal
i-1 to create i, since i+1 may also depend on i-1.)
TODO: Share code for decoding prior frames between GIF and WEBP
Change-Id: I91a5ae22ba3d8dfbe0bde833fa67ae3da0d81ed6
Reviewed-on: https://skia-review.googlesource.com/13722
Reviewed-by: Mike Reed <reed@google.com>
Reviewed-by: Chris Blume <cblume@chromium.org>
Reviewed-by: Matt Sarett <msarett@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
Adds a DynamicState struct to GrPipeline that has a field for the
scissor rect. Eventually this should become the only way to specify a
scissor rectangle and may grow to contain more fields. Adds an array of
DynamicStates to GrGpuCommandBuffer::draw and implements support in GL
and Vulkan.
Bug: skia:
Change-Id: If5aebbf9da5d192acf7e68e7def4674ffc7ec310
Reviewed-on: https://skia-review.googlesource.com/18510
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
For streams that are memory backed (stream->getMemoryBase() returns
a non-null ptr and hasLength() returns true), handle the stream
with skjpeg_mem_source_mgr, which directly assigns memory base to
the source manager. For other non memory backed streams, handle the
stream with skjpeg_buffered_source_mgr, which is renamed from
the old skjpeg_source_mgr with no implementation change.
Original Author: cjbao <cathy.bao@intel.com>
https://skia-review.googlesource.com/c/17296/
Bug: skia:
Change-Id: I0885c5114083f976a9c08afe5954434196762e39
Reviewed-on: https://skia-review.googlesource.com/18933
Reviewed-by: Mike Klein <mtklein@chromium.org>
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>
The coverage blitter is an optimization of regular blitting to A8.
It can go faster because it never tries to blend with itself... because
it was intended to be called only by fill-path, which is written to
never draw twice in the same place. Hairlines don't make this promise,
so we can use the coverage optimization for hairlines.
Change-Id: I0a4816f886462017ad81378d5928357ef038ae8f
Reviewed-on: https://skia-review.googlesource.com/18666
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
For ANGLE D3D9 ES2 we're seeing GL_CHROMIUM_framebuffer_multisample but not GL_ANGLE_framebuffer_multisample.
The same occurs for ANGLE D3D11 ES2 but kAlpha8 is renderable in that case.
Bug: 730082
Change-Id: I7eda844b0ad48150d19406517d1dfbe453e9e478
Reviewed-on: https://skia-review.googlesource.com/18980
Commit-Queue: Robert Phillips <robertphillips@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
The stroke-to-rect test was accidentally doing both non-aa draws and
non-filtered texture sampling on exact half pixel boundaries. So
sometimes the CPU and GPU rounded in different directions. This CL
updates the draws and sampling to fall on integer boundaries.
Bug: skia:5522
Change-Id: If968bcbb327106eba8919e4abf3a7211c5326daf
Reviewed-on: https://skia-review.googlesource.com/18960
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
- Call conversion_possible with the proper alpha type for the frame.
- Always use kUnpremul for the transform. Previously we used the alpha
type for the first frame. If it was opaque and a later frame was not,
this would be incorrect.
Also fix Codec_frames test. Most of the tests were not running due to
a return statement in a loop. Change that to continue, and correct
errors in the test. Provide better debugging information.
Change-Id: Icd40c09526b1d599168bfe90d93d8ddcdd9ca20f
Reviewed-on: https://skia-review.googlesource.com/18935
Reviewed-by: Matt Sarett <msarett@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
This reverts commit 6383b29848.
Reason for revert: Test failures
Original change's description:
> Handle too many (or too large) paths in GrDefaultPathRenderer
>
> PathGeoBuilder constructs the geometry with the same basic
> technique as before, but allows interrupting the process
> to emit multiple draws.
>
> Original test case was 2000 non-AA stroked circles, which
> created ~66000 vertices. That now renders, as do various
> tests with a single large path (as well as filled paths).
>
> TODO: I think that this could be extracted and re-used for
> MSAA path renderer without too much work? I need to read
> that code more carefully to make sure it lines up.
>
> Bug: skia:6695
> Change-Id: I18983ba3d4f475ae0651946958b4911008aa623f
> Reviewed-on: https://skia-review.googlesource.com/18360
> Reviewed-by: Brian Salomon <bsalomon@google.com>
> Commit-Queue: Brian Osman <brianosman@google.com>
TBR=bsalomon@google.com,robertphillips@google.com,brianosman@google.com
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:6695
Change-Id: I78ce9879a2e45e19f53027ca506cc2e8fae664b3
Reviewed-on: https://skia-review.googlesource.com/18981
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
The versions of showmiplevel that take both a width and height, round the
coords of draws to integers. The basic version did not do this which caused
sampling bugs on adreno when using nearest neighbor at half pixel coords.
Since this GM isn't testing the GPU rendering of the mips, this workaround is
okay.
Bug: skia:5905
Change-Id: I4080532e8c1f37d74c60089970c5d0fc83cd5373
Reviewed-on: https://skia-review.googlesource.com/18939
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>
Had to add some logic to avoid touch and mouse cross-talk, because
(at least on my laptop), the touch screen generates both kinds of
events.
This seems really useful [1] for the many [2] Skia developers with
touch-enabled Windows devices.
----------
1: No, not really.
2: N = 1?
Bug: skia:
Change-Id: Ib888bf4198f2cc0a29a31581ec4b64d3d9008c33
Reviewed-on: https://skia-review.googlesource.com/18920
Reviewed-by: Yuqian Li <liyuqian@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Bug: chromium:727678
Change-Id: I4c59d9222d47b866b3c30408322ec456f304aa53
Reviewed-on: https://skia-review.googlesource.com/18938
Commit-Queue: Brian Osman <brianosman@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Mike Klein <mtklein@chromium.org>
PathGeoBuilder constructs the geometry with the same basic
technique as before, but allows interrupting the process
to emit multiple draws.
Original test case was 2000 non-AA stroked circles, which
created ~66000 vertices. That now renders, as do various
tests with a single large path (as well as filled paths).
TODO: I think that this could be extracted and re-used for
MSAA path renderer without too much work? I need to read
that code more carefully to make sure it lines up.
Bug: skia:6695
Change-Id: I18983ba3d4f475ae0651946958b4911008aa623f
Reviewed-on: https://skia-review.googlesource.com/18360
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Bug: skia:5587
Change-Id: I00f377c06228deb9053b1cf7d9f413e0bf2fb490
Reviewed-on: https://skia-review.googlesource.com/18936
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
clang-tidy pointed out this opportunity in the Google3 build.
Change-Id: Id3c7b291fec12cbf860e7fc2583c1a34ad23f697
Reviewed-on: https://skia-review.googlesource.com/18927
Reviewed-by: Yuqian Li <liyuqian@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
Very new versions of Clang (5.0 TOT) call out this problem.
Bug: skia:
Change-Id: Ie86d3936c26769c0d3387a81fd2c247ea1272d1c
Reviewed-on: https://skia-review.googlesource.com/18930
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Mike Klein <mtklein@chromium.org>
clang-tidy pointed out this opportunity in the Google3 build.
Change-Id: I2ab01602fae2d3132a229006c06625371342e18a
Reviewed-on: https://skia-review.googlesource.com/18926
Commit-Queue: Matt Sarett <msarett@google.com>
Reviewed-by: Matt Sarett <msarett@google.com>
Due to floating point inaccuracy, when intersecting edges, the
intersection point may fall above one of the edges' top vertices
or below one of the bottom vertices. In these cases, we were simply
splitting one edge on the relevant endpoint of the other edge. This
is incorrect if the intersection is far from the endpoint (e.g.,
the test case in the linked bug, where one of the intersected edges
is near-horizontal but the intersection falls below both of its
endpoints, in the middle of the edge.)
The correct solution is to split both edges as normal, and take care
to produce edges with the correct ordering where the intersection is
above or below an edge. However, since the new vertex may be above
the current vertex, simply restarting intersection checks at the
current vertex won't work. We need to process the intersection
vertex before the current one.
This introduces another problem: unlike all other splitting modes
(which always shorten edges), splitting an edge above the top or
below the bottom can lengthen it, causing it to violate the AEL
with an adjacent edge which then shortens it back to the original
point (in cleanup_active_edges()). Since the splitting and merging
code can't agree, we loop forever.
Instead of simply fusing neighboring edges in cleanup_active_edges(),
the proper fix to this problem is to detect the AEL violation and
rewind all processing to the vertex above it. For performance, we
only rewind when we detect that a split edge is no longer ordered
within the mesh (merge_enclosing_edges()) or within the the AEL
(rewind_if_necessary()). We also store the enclosing edges of each
vertex, which allows us to rewind quickly, since we know exactly which
edges need to be added/removed from the AEL.
cleanup_active_edges(), fix_active_state() and Vertex::fProcessed have
been removed. In their place are rewind_active_edges() and
rewind_if_necessary(), which uses the same logic as
cleanup_active_edges() but uses it to know when to rewind.
Bug: skia:5026
Change-Id: I3638a429f5428498d6df6bb7b98c67374dc291aa
Reviewed-on: https://skia-review.googlesource.com/18900
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
This reverts commit 8b06ed7c9f.
Reason for revert: try again.
Original change's description:
> Revert "Workaround Adreno driver issue with stencil clears."
>
> This reverts commit 4e8c581f2a.
>
> Reason for revert: breaks Google3 roll
>
> Original change's description:
> > Workaround Adreno driver issue with stencil clears.
> >
> > This also removes the "debug wire rect" which was not used and not implemented for Vulkan.
> >
> > Also some declared but not implemented methods are removed from GrGLGpu.
> >
> > Bug: skia:5587
> >
> > Change-Id: I750051e90e6cfbfad6a6fe20792226182f698bcf
> > Reviewed-on: https://skia-review.googlesource.com/18639
> > Reviewed-by: Chris Dalton <csmartdalton@google.com>
> > Commit-Queue: Brian Salomon <bsalomon@google.com>
>
> TBR=bsalomon@google.com,csmartdalton@google.com
> No-Presubmit: true
> No-Tree-Checks: true
> No-Try: true
> Bug: skia:5587
>
> Change-Id: I65aa16b3f8c70cdef56ff16e16304ba09604c475
> Reviewed-on: https://skia-review.googlesource.com/18924
> Reviewed-by: Mike Klein <mtklein@google.com>
> Commit-Queue: Mike Klein <mtklein@google.com>
TBR=mtklein@google.com,bsalomon@google.com,reviews@skia.org,csmartdalton@google.com
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:5587
Change-Id: Ice49027bda1c1ff7e0362d0680341ac862159850
Reviewed-on: https://skia-review.googlesource.com/18928
Reviewed-by: Mike Klein <mtklein@google.com>
This reverts commit 4e8c581f2a.
Reason for revert: breaks Google3 roll
Original change's description:
> Workaround Adreno driver issue with stencil clears.
>
> This also removes the "debug wire rect" which was not used and not implemented for Vulkan.
>
> Also some declared but not implemented methods are removed from GrGLGpu.
>
> Bug: skia:5587
>
> Change-Id: I750051e90e6cfbfad6a6fe20792226182f698bcf
> Reviewed-on: https://skia-review.googlesource.com/18639
> Reviewed-by: Chris Dalton <csmartdalton@google.com>
> Commit-Queue: Brian Salomon <bsalomon@google.com>
TBR=bsalomon@google.com,csmartdalton@google.com
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:5587
Change-Id: I65aa16b3f8c70cdef56ff16e16304ba09604c475
Reviewed-on: https://skia-review.googlesource.com/18924
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
When drawing a single point it's valid for the first and last indices
to be equal.
Bug: skia:6728
Change-Id: Ice8cd89b306ed9564cdd9a77c557628662e18abb
Reviewed-on: https://skia-review.googlesource.com/18904
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Applies the quadratic formula from "Numerical Recipes in C",
Section 5.6, to the homogeneous quadratic equations that find cubic
inflection points and loop intersections. Also addresses KLM
orientation ahead of time, rather than negating K and L after the
fact.
Bug: skia:
Change-Id: Ic7e0818e2fe49b7724f9b583bae52281cfb1aea1
Reviewed-on: https://skia-review.googlesource.com/13481
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Cary Clark <caryclark@google.com>
The current roll is failing due to double linking:
multiple definition of 'SkEncodeImage(SkWStream*, SkPixmap const&, SkEncodedImageFormat, int)'
src/images/SkImageEncoder.pic.o: previous definition here
Change-Id: I64ada3591b4c4c0cb8deada0dc796d9a0376b6a7
Reviewed-on: https://skia-review.googlesource.com/18855
Reviewed-by: Matt Sarett <msarett@google.com>
This also removes the "debug wire rect" which was not used and not implemented for Vulkan.
Also some declared but not implemented methods are removed from GrGLGpu.
Bug: skia:5587
Change-Id: I750051e90e6cfbfad6a6fe20792226182f698bcf
Reviewed-on: https://skia-review.googlesource.com/18639
Reviewed-by: Chris Dalton <csmartdalton@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Add a dedicated virtual for shaders to select burst mode.
Enabled for linear gradients with more than 8 stops, as a start.
BUG=skia:6710
Change-Id: I1b19124f42d1d805de27a5db6a26601c386bb9ff
Reviewed-on: https://skia-review.googlesource.com/18628
Commit-Queue: Florin Malita <fmalita@chromium.org>
Reviewed-by: Mike Reed <reed@google.com>
Reviewed-by: Mike Klein <mtklein@chromium.org>
This reverts commit 96cc36124d.
Reason for revert: I think this is break one of our test bots. I will reland once that is figured out.
Original change's description:
> Handle different types of streams in different jpeg source managers
>
> For streams that are memory backed (stream->getMemoryBase() returns
> a non-null ptr and hasLength() returns true), handle the stream
> with skjpeg_mem_source_mgr, which directly assigns memory base to
> the source manager. For other non memory backed streams, handle the
> stream with skjpeg_buffered_source_mgr, which is renamed from
> the old skjpeg_source_mgr with no implementation change.
>
> Signed-off-by: cjbao <cathy.bao@intel.com>
> Bug: skia:
> Change-Id: I748de0bdba726bbb318922c08497135e73e37329
> Reviewed-on: https://skia-review.googlesource.com/17296
> Reviewed-by: Leon Scroggins <scroggo@google.com>
> Reviewed-by: Matt Sarett <msarett@google.com>
> Commit-Queue: Matt Sarett <msarett@google.com>
>
TBR=msarett@google.com,scroggo@google.com,cathy.bao@intel.com
# Not skipping CQ checks because original CL landed > 1 day ago.
Bug: skia:
Change-Id: Idf6c426468cc959b3f4661c0b3e86cb4d5e93688
Reviewed-on: https://skia-review.googlesource.com/18850
Reviewed-by: Matt Sarett <msarett@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>
which is used in the Nvidia Jetson TX2 dev kit. It is Linux on ARM64.
Bug: skia:
Change-Id: I97b99226a8eb2b28190d3cf281d4a1a8278ae5b0
Reviewed-on: https://skia-review.googlesource.com/18884
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
This is the same logic from constant_color, covering all the other
places where we convert from float to fixed, e.g. scale_1_float.
This isn't quite ideal yet. We replace mulss+cvttss2si for addss+movd,
which is great, but this leads to a silly sequence of code:
addss %xmm2, %xmm0
movd %xmm0, %r9d
movd %r9d, %xmm0
pshuflw $0x0, %xmm0, %xmm0
Those two movd are pointless...
Again, all diffs due to switching from truncation to rounding.
Change-Id: Icf6f3b6eb370fe41cea0cebcfda0b8907e055f41
Reviewed-on: https://skia-review.googlesource.com/18846
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
This is as good as we can get without switching away from float inputs.
All diffs due to rounding (from the +256.0f).
Change-Id: I0d314f111d313577ce9078660178be17e865f11e
Reviewed-on: https://skia-review.googlesource.com/18845
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Mike Reed <reed@google.com>
For compatibility with older system headers, instead of looking for HWCAP_
values in asm/hwcap.h, just define the bits we want to test ourselves.
This lets us compile this code on systems before those bits were defined.
At runtime the bits will harmlessly test as zero.
Change-Id: I44b6aba7d6f0fc2c5df08ad262c2b0537d900209
Reviewed-on: https://skia-review.googlesource.com/18844
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
Similar to the raster pipeline stage, transform the stops into the dest
color space before interpolation.
Change-Id: I626b6ef18606fd2308d7da166ce70d05f3951e21
Reviewed-on: https://skia-review.googlesource.com/18767
Reviewed-by: Florin Malita <fmalita@chromium.org>
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Florin Malita <fmalita@chromium.org>
This should be immediately useful in the Skia-Android
rendering pipeline.
Possible future uses include creating a "renderable"
SkImage from a bitmap with a funny color space.
Inspired by:
https://skia-review.googlesource.com/c/13981/
Bug: b/62347704
Change-Id: I388c7af1fc43834b8ad22022d0caf3ac90b734c8
Reviewed-on: https://skia-review.googlesource.com/18598
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>
Add Git recipe module to easily use this version of Git anywhere.
This fixes recipe bundling and unblocks the recipe roll.
Bug: skia:
Change-Id: Ib4d1361b7a52676e1992025b29e630ea3ada173b
Reviewed-on: https://skia-review.googlesource.com/18833
Reviewed-by: Ravi Mistry <rmistry@google.com>
Commit-Queue: Eric Boren <borenet@google.com>
Change-Id: I8a292bc98135b41ceedb4242451436c3657616fc
Reviewed-on: https://skia-review.googlesource.com/18722
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
The chromium compositor wants to lean on Skia internal serialization (at
least temporarily, but possibly long term as well) for certain types.
As flattening types requires an SkWriteBuffer, make SkWriteBuffer
have public visibility in component builds.
Change-Id: I635f89bcf816aa376682bd7f7ef46de7d5669e12
Reviewed-on: https://skia-review.googlesource.com/18700
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Mike Klein <mtklein@chromium.org>
This seems to make Instruments work a lot better.
Change-Id: I078f005d32e427b4eb31bc92c731e3444f2faffb
Reviewed-on: https://skia-review.googlesource.com/18635
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
Most SkCodec subclasses did the following to apply their
SkColorSpaceXform:
dstFormat = select_xform_format(dstInfo.colorType());
srcFormat = select_xform_format(<something that doesn't change>);
xformAlphaType = select_xform_alpha(dstInfo.alphaType(),
this->getInfo().alphaType());
this->colorXform()->apply(dstFormat, dst, srcFormat, src, width,
xformAlphaType);
Consolidate the computation of these parameters into SkCodec and add a
new method to SkCodec that calls apply() with those parameters.
Add a SkColorSpaceXform::ColorFormat to SkCodec. This allows the new
method SkCodec::applyColorXform to supply the ColorFormat.
TBR=reed@google.com
(No change to public API.)
Change-Id: I8ea7ba4c0024be827a9f9359796c778744330f6e
Reviewed-on: https://skia-review.googlesource.com/18523
Reviewed-by: Leon Scroggins <scroggo@google.com>
Reviewed-by: Matt Sarett <msarett@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
Some assertions cannot be relied upon due to FP error.
Bug: skia:5990
Change-Id: I32445b320b9100ae2f80d2f762707d823da77805
Reviewed-on: https://skia-review.googlesource.com/18602
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
(instead of finishing enque before draw). The highlight is that we can now
achieve 9x speedup compared to 5x in all our previous approaches
(including multi-picture draw).
The schedulers here are experimental. I'd like to move on to try initializing
once for each draw before further polishing and optimizing the schedule
mechanism.
Bug: skia:
Change-Id: Idc3d030d475af9645c24c5372ff62b9a402206cc
Reviewed-on: https://skia-review.googlesource.com/17826
Reviewed-by: Mike Reed <reed@google.com>
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Yuqian Li <liyuqian@google.com>
