Most backends don't like init-stmts with multiple VarDeclarations inside
them, so we no longer emit IR that does this. This lets us avoid doing
backend-level fixups.
Change-Id: Ide839de18953a73e0f9c7a690df59a7bc3523f89
Bug: skia:11860
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/398221
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Before this CL, only select bots used reordering. After this CL,
all bots except the NUC5PPYH use the flag.
Once all our clients are migrated, we'll remove the flag altogether.
Bug: skia:10877
Change-Id: Iee734f8b99ae9eba6c1947009d8164032594b051
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/398219
Commit-Queue: Adlai Holler <adlai@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Reviewed-by: Ravi Mistry <rmistry@google.com>
This is a step toward removing the various GrTextureProducer classes
and moving their functionality into SkImage subclasses. Currently,
the new method just creates the appropriate producer and invokes
its methods to create the FP.
Use in SkImageShader.
Bug: skia:11877
Change-Id: Idd9776d219ad69a5763c22603d1d06ddadb7ad47
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/398220
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Change-Id: If0ab7ee6b8518f679ed5d9d806c52eb7a6df938b
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/398456
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Bug: skia:11876
Change-Id: I4fed0accb0797163add0dc622bbe93b24dec2641
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/398036
Reviewed-by: Eric Boren <borenet@google.com>
Commit-Queue: Joe Gregorio <jcgregorio@google.com>
b027dfeb83..b574643ef2
2021-04-19 geofflang@chromium.org D3D11: Skip blits if there is no intersection of dest areas
2021-04-19 jmadill@chromium.org Revert "infra: Switch to "angle_2" presubmit config."
2021-04-19 m.maiya@samsung.com Vulkan: Keep the output variable of TCS active for barrier
2021-04-19 jmadill@chromium.org Revert "Revert "Vulkan: Suppress draw-time push constant VVL warnings."
2021-04-19 jmadill@chromium.org infra: Switch to "angle_2" presubmit config.
2021-04-19 angle-autoroll@skia-public.iam.gserviceaccount.com Roll SwiftShader from f32791f5d53b to dac99e8221a8 (1 revision)
2021-04-19 angle-autoroll@skia-public.iam.gserviceaccount.com Roll Chromium from 11d568cbfed1 to cb5ad5e674fa (484 revisions)
If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/angle-skia-autoroll
Please CC jvanverth@google.com on the revert to ensure that a human
is aware of the problem.
To report a problem with the AutoRoller itself, please file a bug:
https://bugs.chromium.org/p/skia/issues/entry?template=Autoroller+Bug
Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+doc/master/autoroll/README.md
Cq-Include-Trybots: skia/skia.primary:Build-Debian10-Clang-x86_64-Release-ANGLE;skia/skia.primary:Test-Win10-Clang-AlphaR2-GPU-RadeonR9M470X-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-Golo-GPU-QuadroP400-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUC5i7RYH-GPU-IntelIris6100-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUC6i5SYK-GPU-IntelIris540-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUC8i5BEK-GPU-IntelIris655-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUCD34010WYKH-GPU-IntelHD4400-x86_64-Debug-All-ANGLE
Tbr: jvanverth@google.com
Change-Id: I1f9cdc8379a452ea6b84936d4dfe5719939bb8e9
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/398318
Reviewed-by: skia-autoroll <skia-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: skia-autoroll <skia-autoroll@skia-public.iam.gserviceaccount.com>
This meant that the new Shader/ColorFilter factories were always
compiling against the more lenient (old) rules.
Bug: skia:11813
Change-Id: I65038595edf2c6087aad6efcd27508338b24cce8
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/398229
Reviewed-by: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
In the old code, fFirstPt actually pointed to the first added point that
came *after* a call to setMovePt. The moved-to point that started the
contour was never remembered in the Convexicator, so it required the
SkPath::Iter to force close. When the iterator force-closed, it made
sure to add a line-to back to the start of the contour, which it could
do because the Iter was also tracking the contours.
- It's important that the old code's close()'s addPt(fFirstPt) actually
re-added the first point after the move-to because it catches
concavities that occur at vertex that started the contour. The first
time that point is added, we have no initial vector to check for
direction changes against.
In the new code, I've updated the state tracking to be more straight
forward. fCurrPt and fLastPt were redundant (they always equaled each
other once addPt returned). fPriorPt was also redundant when combined
with the expected direction enum, since it's only use was to determine
when to cache the initial vector. I also found it confusing to know
the difference between fLastPt and fPriorPt based just on variable
names.
Now fFirstPt refers to the location of the moveTo/"start of contour",
and fFirstVec is the direction vector leaving that point to the next
point in the contour. fLastPt is the point that was last passed to
addPt and fLastVec is the direction that comes into 'fLastPt'. The
close() function is updated to add a point back to fFirstPt (which does
nothing if the close is explicit, or handled by SkPath::Iter) AND it
checks for a dir change on the first vector.
This shouldn't change any behavior, so I wanted it as a standalone
change, but it will allow
https://skia-review.googlesource.com/c/skia/+/396056/ to
update computeConvexity() to move away from
SkPath::Iter(forceClose=true).
Bug: skia:1187385
Change-Id: Iff383693956b28e3c21b23e5a5c6ce814227bb27
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/398018
Reviewed-by: Chris Dalton <csmartdalton@google.com>
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
This is another strange, experimental feature that clutters the
implementation and isn't used by anyone (to my knowledge).
Change-Id: I538b7eca0cd28aab32f4739b23459731ade9105e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/398226
Commit-Queue: John Stiles <johnstiles@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
The "raster" window on macOS and iOS is actually backed by GL. Fix the
build rules and code conditions to reflect this. This allows for some
sk_app applications to run on macOS and iOS with skia_use_gl=false.
> Revert:
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/397737
> Commit-Queue: Brian Osman <brianosman@google.com>
> Reviewed-by: Brian Osman <brianosman@google.com>
> Land:
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/397256
> Commit-Queue: Ben Wagner <bungeman@google.com>
> Reviewed-by: Brian Osman <brianosman@google.com>
Change-Id: Ia8a421f4818856dd90cb4847095eee0d1836d1e6
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/398056
Reviewed-by: Ben Wagner <bungeman@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Ben Wagner <bungeman@google.com>
In D3D we can alias a resource's allocation and re-use it for
another resource. For example, this is used in the mipmap generator
when we have an sRGB texture. The sRGB formats can't be used in a
UAV, but we can use an aliased texture with the corresponding linear
format, and do the sRGB conversion in the compute shader.
Change-Id: I6995cdecf5be36aafbed5c48e639c09f41ec4f8b
Bug: skia:10446
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/397996
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Jim Van Verth <jvanverth@google.com>
This was an experimental feature. It worked (but only the GPU backend).
It was never adopted or used by anyone, to my knowledge. It's a large
amount of code, and a strange corner of SkSL for users to stumble into.
Bug: skia:10680
Change-Id: I0dda0364bce7dbffa58c32de4c7801ec2a6bc42e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/398222
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Font was missing getMetrics()
However, seems like getShapedRuns() needs to return its
particular line spacing choices...
Change-Id: I574ebf789fb03b79c8e09198f1cb6e09dac3441b
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/397916
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Julia Lavrova <jlavrova@google.com>
This will let us see the reordered_dags_over_budget stat
in perf so that we know our fallback code path is getting tested.
Bug: skia:10877
Change-Id: I073c8622ddb8f3449511193bac045bfebc4e277c
Cq-Include-Trybots: luci.skia.skia.primary:Perf-Debian10-Clang-NUC7i5BNK-GPU-IntelIris640-x86_64-Release-All
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/398217
Reviewed-by: Ravi Mistry <rmistry@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
Auto-Submit: Adlai Holler <adlai@google.com>
Unused InterfaceBlocks were not added to the ProgramUsage map. The
ProgramUsageVisitor now makes sure to account for them during its
initial scan.
Change-Id: If3afac8e954c5b685ddc6b63b0f771d8c0b8f207
Bug: oss-fuzz:33405
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/398016
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
As explained, this is *very* conservative. It only works when the child
is sampled from within main, and using a direct reference to the coords
parameter. If that parameter is ever modified (even after being used),
the optimization doesn't happen. For most cases, this is fine.
Bug: skia:11869
Change-Id: Ia06181730a6d07e2a4fe2de4cc8e8c3402f0dc52
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/397320
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
It should now be rare to require a RGBA texture from a YUVA image (e.g.
in image filter)
Removes a thread-safety issue. A step towards allowing clients to pump
new data into the planes and be sure Skia wont draw the old contents.
Bug: skia:11873
Change-Id: I007cf28e477155f85ed9f1c6cc4547f9dbbb73fd
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/397319
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
f2aa9d5da6..b027dfeb83
2021-04-19 syoussefi@chromium.org Revert "Vulkan: Suppress draw-time push constant VVL warnings.
2021-04-17 syoussefi@chromium.org Suppress fifa mobile on intel/windows
2021-04-17 cnorthrop@google.com Tests: Add 3DMark Slingshot Test 2 trace
2021-04-17 cnorthrop@google.com Tests: Add 3DMark Slingshot Test 1 trace
2021-04-17 syoussefi@chromium.org Vulkan: Remove inappropriate use of VK_NULL_HANDLE
2021-04-17 cclao@google.com Vulkan: Add support for FBO with unequal sized attachments
2021-04-16 cclao@google.com Vulkan: Add test for FBO with unequal sized attachments
2021-04-16 syoussefi@chromium.org Skip failing test on AMD/Linux
2021-04-16 xiaoxuan.liu@arm.com Reland "Reland "Add support for Linux GPU info with Vulkan backend""
2021-04-16 cnorthrop@google.com Migrate more scripts to python3
2021-04-16 geofflang@google.com Remove 3pp files in third_party/jdk.
2021-04-16 m.maiya@samsung.com Vulkan: Handle non array type varying error case in TS
2021-04-16 angle-autoroll@skia-public.iam.gserviceaccount.com Roll Chromium from cac0b331e3fc to 11d568cbfed1 (483 revisions)
If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/angle-skia-autoroll
Please CC jvanverth@google.com on the revert to ensure that a human
is aware of the problem.
To report a problem with the AutoRoller itself, please file a bug:
https://bugs.chromium.org/p/skia/issues/entry?template=Autoroller+Bug
Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+doc/master/autoroll/README.md
Cq-Include-Trybots: skia/skia.primary:Build-Debian10-Clang-x86_64-Release-ANGLE;skia/skia.primary:Test-Win10-Clang-AlphaR2-GPU-RadeonR9M470X-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-Golo-GPU-QuadroP400-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUC5i7RYH-GPU-IntelIris6100-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUC6i5SYK-GPU-IntelIris540-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUC8i5BEK-GPU-IntelIris655-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUCD34010WYKH-GPU-IntelHD4400-x86_64-Debug-All-ANGLE
Tbr: jvanverth@google.com
Test: Test: FramebufferTest_ES3.AttachmentWithUnequalDimensionsTest: Test: KHR-GLES32.core.tessellation_shader.compilation_and_linking_errors.te_non_arrayed_per_vertex_input_blocksTest: Test: angle_perftests --gtest_filter="*slingshot_test1*"Test: Test: angle_perftests --gtest_filter="*slingshot_test2*"Test: Test: python3 scripts/run_code_generation.py
Change-Id: Ie29de29aa0708d54d19b905df38177e97d0797e0
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/397841
Reviewed-by: skia-autoroll <skia-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: skia-autoroll <skia-autoroll@skia-public.iam.gserviceaccount.com>
This reverts commit 163ba10dde.
Reason for revert: Mac linker errors
Original change's description:
> Fix sk_app macOS raster window build conditions.
>
> The "raster" window on macOS is actually backed by GL. Fix the build
> rules and code conditions to reflect this. This allows for some sk_app
> applications to run on macOS with skia_use_gl=false.
>
> Change-Id: I5d7b37c4172079e163690faa4e55a622a6d4f844
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/397256
> Commit-Queue: Ben Wagner <bungeman@google.com>
> Reviewed-by: Brian Osman <brianosman@google.com>
TBR=bungeman@google.com,brianosman@google.com
Change-Id: Ie5fa24138e4387784c21559f28528a4c4d335626
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/397737
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
The "raster" window on macOS is actually backed by GL. Fix the build
rules and code conditions to reflect this. This allows for some sk_app
applications to run on macOS with skia_use_gl=false.
Change-Id: I5d7b37c4172079e163690faa4e55a622a6d4f844
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/397256
Commit-Queue: Ben Wagner <bungeman@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
SPIR-V would previously jump to the top of the do-while loop when
encountering a `continue`. This is not correct; it should jump to the
loop-exit test. We don't have any real shaders which test this case, and
it can't be checked from inside Runtime Effects since do-while is not
part of ES2.
Change-Id: Iac58e9e7ad0a94640e6418afea096b19ba50a2df
Bug: skia:11874
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/397576
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
These are rounded rectangles whose corners are superellipses instead
of ovals. This builds on a lot of work done by zakcohen@google.com to
define a consistent, visually appealing shape that can be used
reliably in the Android UI.
Change-Id: I97fde7be870152c9ee776fde53118019d0caa251
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/397416
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Change-Id: Ic5192b072c98617dee85f28f8f4214c7a805a1fd
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/397318
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Jim Van Verth <jvanverth@google.com>
This bot also runs reduceOpsTaskSplitting, so this allows us
to test the over-memory-budget fallback behavior.
At 16MB my desktop falls back on the following SKPs: desk_carsvg,
desk_twitter, desk_yahoosports, desk_linkedin, desk_facebook.
Bug: skia:10877
Change-Id: I60934e990581f762b72c077c6e8a40cc0c1d3d8c
Cq-Include-Trybots: luci.skia.skia.primary:Test-Debian10-Clang-NUC7i5BNK-GPU-IntelIris640-x86_64-Debug-All-ASAN_Vulkan,Test-Debian10-Clang-NUC7i5BNK-GPU-IntelIris640-x86_64-Debug-All-DDL3_ASAN
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/397461
Commit-Queue: Adlai Holler <adlai@google.com>
Reviewed-by: Eric Boren <borenet@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
A Block() at the top-level scope would previously disappear silently.
Change-Id: Ic3bac058361658d2e5bd8de1e3718dedf4cef1f7
Bug: skia:11854
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/397516
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
This also connects a lot of the wires needed to use dynamic MSAA in
vulkan. By using the framebuffer object in the render pass we can figure
out the specific framebuffer we want in one place, GrVkGpu::onGetOpsRenderPass,
and then the render pass itself doesn't need any explicit knowledge of
dmsaa stuff.
Bug: skia:11809
Change-Id: I3e4e71fa6f9536fdaf915d5369a2f8a24bf48c9b
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/397156
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>
We need to wait until there's a resourceCache to set the value on.
Bug: skia:10877
Change-Id: I1558af6c369e857fbafb450179d3b341250ad16a
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/397459
Auto-Submit: Adlai Holler <adlai@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Adlai Holler <adlai@google.com>
