Rewrites tessellation atlases as normal render tasks instead of
"onFlush" tasks. These tasks get inserted into the DAG upfront, lay
out their atlases as dependent tasks get built and reference them, and
finally add their ops to render themselves during onMakeClosed. Doing it
this way allows us to pause the flush and re-render the atlas whenever
it runs out of room.
Bug: b/188794626
Bug: chromium:928984
Change-Id: Id59a5527924c63d5ff7c5bce46a88368e79fc3ef
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420556
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Adlai Holler <adlai@google.com>
The tessellation atlas task will need a recording context in order to
record its ops during onMakeClosed.
Bug: b/188794626
Bug: chromium:928984
Change-Id: Ie8719a514899a5748c52ae78a6ecd997b1439ab7
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420879
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
The tessellation atlas needs this to call visitProxies.
Bug: b/188794626
Bug: chromium:928984
Change-Id: I5bfb2559abcaf5c7602393e96adb846bcfbce971
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420878
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Change-Id: I16301e6fb7193ba67b9fea292f931fda480c8ac3
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420956
Commit-Queue: Herb Derby <herb@google.com>
Reviewed-by: Mike Reed <reed@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Change-Id: I87defa005fab8b1be92a700bf38c91b92c0ab8a7
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420616
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Bug: chromium:1203652
Change-Id: Id83ac81c40eda2653e97a7c8ae9326c273f0f00b
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420537
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Bug: skia:11844
Change-Id: I8a5fdd6b074ba070022b6da3e31d8642b2f35181
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420896
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Change-Id: I015cc2401c9f4a87fa7a02a93890bcf0a6704a91
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420901
Commit-Queue: Greg Daniel <egdaniel@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Auto-Submit: Greg Daniel <egdaniel@google.com>
Reviewed-by: Chris Dalton <csmartdalton@google.com>
This reverts commit 3366efd581.
Reason for revert: Perf regression root cause fixed.
Original change's description:
> Revert "Replace GrColorMatrixFragmentProcessor with a runtime FP"
>
> Triggered a perf regression, somehow.
>
> This reverts commit 311010405e.
>
> Change-Id: Ic468d34c8ccd28aaaed2d3028ce9559115860844
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420816
> Reviewed-by: Brian Osman <brianosman@google.com>
> Commit-Queue: Brian Osman <brianosman@google.com>
# Not skipping CQ checks because this is a reland.
Change-Id: Iea53fba22ef08b0ed97220b9cd2a9a0d054ecf57
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420900
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
We were checking the flags for each uniform, but never advancing the
uniform iterator. As a result, for each uniform that was specialized,
we'd just fold in another copy of the *first* uniform's data.
This led to a nasty performance regression (extra shader compiles) in
certain scenarios. (And it was a correctness issue, as well).
Change-Id: Iaac878a7b9eb37e98815f56dc55a58d1f8e85c22
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420820
Reviewed-by: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
This unifies various chunks of error handling code and makes the DSL
more gracefully handle failed object construction.
Change-Id: I4edc581c9eba276ce2ee2a17125ea7bd6057134a
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420237
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Adding ovals reveals the clipping failure. The "actual" results with the
oval clips are currently incorrect.
Bug: skia:11837
Change-Id: I0789d5f758457425250ff6cecc0a5bb44f6e82ef
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420637
Commit-Queue: Robert Phillips <robertphillips@google.com>
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Triggered a perf regression, somehow.
This reverts commit 311010405e.
Change-Id: Ic468d34c8ccd28aaaed2d3028ce9559115860844
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420816
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Change-Id: Id1c0375d78c7b481e60aeeda53e3f89acf543c9e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420636
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Change-Id: I9dcfc0dd559fe1596019f0d8051f665f69a6960d
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/419163
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Jim Van Verth <jvanverth@google.com>
This array isn't used anywhere outside of a single function; we don't
need it to exist at global scope.
Change-Id: I66d65dcb92d5c7bbcfbd067ef6b6908e4dfd3dd6
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420617
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
This involves a significant number of overloads, as optionals can be
compared against other optionals, or nullopt, or a value; either side
can mix and match as desired.
Change-Id: I26b8420e417300d32569fc76a55bcfd8e01eb322
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420576
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Bug: skia:11837
Change-Id: I76e1cfdf7d2f7ca649612c8155571425684cf2d0
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420236
Commit-Queue: Robert Phillips <robertphillips@google.com>
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Bug: skia:11837
Change-Id: I07c7b3e3b55be8b0340a012e965bd23b92d27e1d
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420218
Commit-Queue: Robert Phillips <robertphillips@google.com>
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
This reverts commit 8cb7c3be75.
Just leave Win7 Golo tasks on python 2.
Change-Id: Idf1d6cc6d303107afd8e74cdec94e7a9ec91399e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/416316
Reviewed-by: Ravi Mistry <rmistry@google.com>
Commit-Queue: Eric Boren <borenet@google.com>
This allows SkM44 and optimize_intrinsic_call to share the same matrix
inversion logic.
Support for 2x2 and 3x3 matrix inversion was also added to
SkMatrixInvert, but is not currently used anywhere in Skia besides
optimize_intrinsic_call. (The perspective case of SkMatrix::ComputeInv
could probably be adapted to use the shared 3x3 code.)
Change-Id: Ie55aab536b4bd0c0a04f2c24527747670ca68fbb
Bug: skia:12054
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420219
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Change-Id: I657ad573fe30e7793d20548c48b824229a3df3d7
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420538
Reviewed-by: Ravi Mistry <rmistry@google.com>
Commit-Queue: Eric Boren <borenet@google.com>
Bug: skia:11837
Change-Id: I8b209b362238e18a110fc518cb1c0b2f2244c22f
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/419840
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
Change-Id: Ife4dcd5627851b2dac1ed05b38c551d5d258e39c
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/419797
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
As is, all 3 formats have a median time of about 0.12 milliseconds.
Perhaps the performance hit is more strongly felt on larger font
files.
Bug: skia:12112
Change-Id: Ifb4b0aa5a8240d93e94f4b674d0a50e6009887ec
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420097
Reviewed-by: Kevin Lubick <kjlubick@google.com>
f0d8cd4094..9412ac9c7c
2021-06-22 angle-autoroll@skia-public.iam.gserviceaccount.com Roll Chromium from 414f62ab7363 to 637b4cdf0c64 (416 revisions)
2021-06-21 jplate@google.com CL: Make CL front end and back end thread-safe
2021-06-21 timvp@google.com Fix roll_aosp.sh 'gn gen' failure
2021-06-21 jplate@google.com CL: Refactor TRY macro and fix more conformance bugs
2021-06-21 angle-autoroll@skia-public.iam.gserviceaccount.com Roll vulkan-deps from 65149e19efd9 to 6cbd7212ad42 (2 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 tdenniston@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: tdenniston@google.com
Change-Id: Iebf2bd314b4b672ac124bc2594aa21fcf0254f69
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420287
Reviewed-by: skia-autoroll <skia-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: skia-autoroll <skia-autoroll@skia-public.iam.gserviceaccount.com>
Convexity is determined with a two pass algorithm, first by a sign
measure and then by a winding measure. Convexity also is meant to not be
affected by leading moveTos (other than the last leading moveTo before
real verbs) and not affected by trailing moveTos (since no additional
contour has actually started).
The old code would incorrectly reduce pointCount when the last moveTo
index was greater than 0, so the BySign pass was skipped or calculated
on an incomplete set of points. When a path (as the one added in this
CL's new test) is convex by winding but not by sign, it would be
incorrectly identified as convex. This led to further cascading issues
during rasterization.
However, the old code also had the effect of correctly ignoring any
last trailing moveTo from being included in the BySign test. Without
the new loop decrementing pointCount, trailing moveTo locations
would incorrectly create concave paths (and would in fact be concave
if the verb was anything other than a move).
I also realized that if the last moveTo index is not at the end of the
initial leading block, or at the end of the path entirely, then it
means the path must have multiple contours, at which point the path
cannot be convex, so we take the early out.
TBR: reed@google.com, bsalomon@google.com
Bug: skia:1220754
Change-Id: I9bd38f2eaaa3dbee135c190ade46fce0bd20257a
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420238
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Change-Id: I51013e7ab1b0361bd4bc9d2df1525164a142a892
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420216
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
When writing runtime FPs, this lets us confirm that we're going to get
the constant output for constant input optimization.
Change-Id: I358507585423e5b15c92a0b2e9278b16f8727fe8
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/419798
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Change-Id: Ic0876160c936132c53b55ac9d27e0e422e1ccd26
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420217
Reviewed-by: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Chrome does not rely on the default – this option is under a finch trial
named "ReduceOpsTaskSplitting." See crbug.com/1210578.
Android does not rely on the default – the flag is controlled by a
property named "renderthread.skia.reduceopstasksplitting", and enabled
by default. See b/183612348.
Flutter does not rely on the default – they explicitly disable the
option. See https://github.com/flutter/engine/pull/26568.
Our tools do not rely on the default – they enable the flag unless the
user passes --dontReduceOpsTaskSplitting on command line. Our bots
run with the flag enabled except the NUC5PPYH, for the sake of coverage.
The flag has no effect for Adreno 620 and 640, on drivers older than
v571, due to a driver bug.
Bug: skia:10877
Change-Id: Iafc84f3463194f0cf6cf135b2be05ec8eb08e446
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/419836
Auto-Submit: Adlai Holler <adlai@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
An upcoming CL will need to refer to the (now) first uniform IDs while
processing the second set, so arranging them like this makes more sense.
Landing this reordering first, to avoid making the new CL even more
confusing.
Change-Id: If19933edbabe840dd427525e5fceac8cd7fb1714
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420120
Commit-Queue: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: Brian Osman <brianosman@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
In the old layer creation, I believe nan matrices were creating
non-finite bounds that then failed intersection tests, which led to no
layer or device being made. With the new system focusing on device-to-
device relationships, the nan's weren't impacting bounds but led to
issues inverting the global-to-device matrix.
It's better to just be explicit about this anyways.
Bug: chromium:1217062
Change-Id: I44e792d0bfad91025c5a727b6bcc74b09bdfcc44
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/419856
Commit-Queue: Robert Phillips <robertphillips@google.com>
Auto-Submit: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
This change bumps the SkPicture version up by one.
Change-Id: Ib23214277cef1777edcc75b9c0cca9209adb9778
Bug: skia:12080
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/419804
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Mike Reed <reed@google.com>
