Bug 36932:
Adds a lower limit when fuzzing dash path effects, since it can produce
paths with > 140k verbs. While this is not that much memory on its own,
the triangulating path renderer can require 3+GB to complete its work
(although it doesn't actually fail).
Bug 36945, 37042:
Also has PathToTriangles check for finite paths before starting any
triangulation work. These paths were created with infinities and NaNs.
Normally such a path would be rejected at a higher level in SkCanvas.
Since the triangulator is being fuzzed directly, this emulates this.
It's included in GrTriangulator and not the fuzzer's main function
because it's a cheap test and theoretically we could encounter a path
that was built lower down (e.g. dashing or transformed to device space)
that then overflowed.
Bug: oss-fuzz:36923, oss-fuzz:36945, oss-fuzz:37042
Change-Id: If97212bf410f771b42cebaedb5733af1abbfc4b2
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/449520
Reviewed-by: Greg Daniel <egdaniel@google.com>
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Also update RELEASE_NOTES to describe new syntax.
Change-Id: I2666551b98f80b61ae3a48c92a9e306cdc7242b0
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/444735
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
Bug: skia:12302
Change-Id: I8cf958acf9214d0de903a4097647afd74f2a659e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/441541
Reviewed-by: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
If the passed-in shader references RTFlip (i.e., sk_FragCoord is used),
the settings must contain RTFlip layout info; otherwise, an error
occurs. Originally, the fuzzer detected this as a problem because the
error was being delivered via SK_ABORT, but it's failing more cleanly
now that Ethan's new error handling code is in place (causing the fuzzer
to report that the bug was "fixed"). With this CL, the oss-fuzz shader
will actually compile successfully in SPIR-V instead of leading to an
error.
Change-Id: I3268e84bd8e01c95a25ed0845a37324e98033c4b
Bug: oss-fuzz:35916
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/439779
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
This is the other half of making everything in gpu/ops be v1-only.
Bug: skia:11837
Change-Id: I5d77a499ef02eba69208d5bd634650433d02f6fb
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/440216
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
These changes will allow us to enable shadowed-variable warnings.
Change-Id: I24ee7e198c1c77b58836237c37557c00452680e1
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/439476
Reviewed-by: Ben Wagner <bungeman@google.com>
Commit-Queue: Ben Wagner <bungeman@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Change-Id: I0e093fd35b11e9a765ef9c09f3b6346086ff66bf
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/435983
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
The SkClipOpPriv.h header will be going away soon, but a number of
places still use its kIntersect_SkClipOp definitions instead of the
equivalent SkClipOp::kIntersect. Besides updating these references,
a number of unnecessary includes to SkClipOpPriv.h are removed and
some test cases exercising expanding clip ops are deleted since they
will be unnecessary shortly.
Bug: skia:10208
Change-Id: I2bbdd6bb39869134c6a80ef2b4482e6cabdeb0b7
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/436157
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
This CL doesn't actually allow the callbacks to be invoked from SkSL
yet, but the callbacks now exist and are now threaded through the
various callsites which will need them.
Change-Id: I00f43ff94de9da8d93daf2e59885eea6f87c2e3e
Bug: skia:12257
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/431696
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Passing a color to a shader isn't supported in SkSL anymore.
Change-Id: Ic2743e519b69fa8add4c010166032b4dab2152a7
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/431938
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
We now have `sampleShader` and `sampleColorFilter`. This CL prepares us
for the addition of a third sampleable type, SkBlender, with a unique
signature (two colors, no coordinates).
Change-Id: I77ba9ba94a1c447d0ef10f5acb5ffe9b9395cfe1
Bug: skia:12257
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/431697
Commit-Queue: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
This is a reland of e58831cd95
Original change's description:
> Add format-specifier warnings to SkDebugf.
>
> This CL fixes up many existing format-specifier violations in Skia.
> Note that GCC has a warning for formatting nothing, so existing calls to
> `SkDebugf("")` have been removed, or replaced with `SkDebugf("%s", "")`.
> These were apparently meant to be used as a place to set a breakpoint.
>
> Some of our clients also use SkDebug with bad format specifiers, so this
> check is currently only enabled when SKIA_IMPLEMENTATION is true.
>
> Change-Id: I8177a1298a624c6936adc24e0d8f481362a356d0
> Bug: skia:12143
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420902
> Auto-Submit: John Stiles <johnstiles@google.com>
> Commit-Queue: Brian Osman <brianosman@google.com>
> Reviewed-by: Brian Osman <brianosman@google.com>
Bug: skia:12143
Change-Id: Id3c0c21436ebd13899908d5ed5d44c42a0e23921
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/421918
Commit-Queue: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
This reverts commit e58831cd95.
Reason for revert: looks like breaking a few build bots
Original change's description:
> Add format-specifier warnings to SkDebugf.
>
> This CL fixes up many existing format-specifier violations in Skia.
> Note that GCC has a warning for formatting nothing, so existing calls to
> `SkDebugf("")` have been removed, or replaced with `SkDebugf("%s", "")`.
> These were apparently meant to be used as a place to set a breakpoint.
>
> Some of our clients also use SkDebug with bad format specifiers, so this
> check is currently only enabled when SKIA_IMPLEMENTATION is true.
>
> Change-Id: I8177a1298a624c6936adc24e0d8f481362a356d0
> Bug: skia:12143
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420902
> Auto-Submit: John Stiles <johnstiles@google.com>
> Commit-Queue: Brian Osman <brianosman@google.com>
> Reviewed-by: Brian Osman <brianosman@google.com>
TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com
Change-Id: I07848c1bf8992925c9498e916744d0840355a077
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:12143
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/421917
Reviewed-by: Tyler Denniston <tdenniston@google.com>
Commit-Queue: Tyler Denniston <tdenniston@google.com>
This CL fixes up many existing format-specifier violations in Skia.
Note that GCC has a warning for formatting nothing, so existing calls to
`SkDebugf("")` have been removed, or replaced with `SkDebugf("%s", "")`.
These were apparently meant to be used as a place to set a breakpoint.
Some of our clients also use SkDebug with bad format specifiers, so this
check is currently only enabled when SKIA_IMPLEMENTATION is true.
Change-Id: I8177a1298a624c6936adc24e0d8f481362a356d0
Bug: skia:12143
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/420902
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
This CL preserves the "StringFragment" name as an alias for
string_view to reduce the impact. The StringFragment alias
will be removed in a followup CL.
Change-Id: I89209bc626b0be0d0190823b6217f4c83cafe1bc
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/416736
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
Previously, when converting a program, the pipeline stage could assume
that an input to main() of type half4/float4 was always the input color.
This was a good assumption since the only possible inputs were
coordinates, or the input color.
This CL now recognizes that when a second float4 is passed to the main()
function, it should be a SK_DEST_COLOR_BUILTIN. This will let blend
functions pass in two colors.
ProgramToSkVM now takes a dest-color argument as well, but existing call
sites won't reference it (since they aren't for blend functions). I've
just passed the input-color a second time, since the value will never
actually be accessed.
Change-Id: I4214586bda605c6d287aa25b1b099e6ef5ba15a4
Bug: skia:12080
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/417261
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Change-Id: I674f038600afd6d49316c1ece515941ee5579068
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/406939
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
It took a few "independent" changes in order to get the linked fuzzer
bugs to pass without failing. Leaving any out triggered an issue :/
1. I changed nearly_flat to be <= epsilon, since if it's == epsilon and
we split the edge, then the new coordinates would have
difference < epsilon.
2. I updated double_to_clamped_scalar to also snap very small values to
0 (right now 16 * float epsilon).
3. double_to_clamped_scalar is now used to clean up the computed
intersection of two edges, and is used to process all initial
vertices (in case the curve evaluation generates lots of denormals
etc.)
4. I updated the use of nearly_flat in checkForIntersection to report
no intersection if both lines are nearly_flat. The comments suggest
nearly_flat means you can't split along that line since the new
coord is incalculable. So if both lines are flat, it's a really
tough numerical scenario and I just punt.
Then I made a few other changes for the fuzzer and debugging:
1. Added more logging messages and updated the code so that it compiles
correctly if TRIANGULATOR_LOGGING is defined.
2. I was also getting asserts in the fuzzer because the vertex
allocator expects the vertex buffer to be detached by the path
renderer before its destroyed, so I just have the fuzzer detach and
discard it. Running locally, the fuzzer test cases from the two
linked bugs pass successfully without oom'ing or timing out.
Bug: oss-fuzz:33672, oss-fuzz:33620
Change-Id: I7687b920db0a9e200b3fa79b323974b7812e52ff
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/404120
Reviewed-by: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
This is a reland of adadb95a9f
... adds a temporary workaround for some Android framework code.
Original change's description:
> Better first-class shader & color filter support in runtime effects
>
> This does a few things, because they're all intertwined:
>
> 1) SkRuntimeEffect's API now includes details about children (which Skia
> stage was declared, not just the name). The factories verify that the
> declared types in the SkSL match up with the C++ types being passed.
> Today, we still only support adding children of the same type, so the
> checks are simple. Once we allow mixing types, we'll be testing the
> declared type against the actual C++ type supplied for each slot.
> 2) Adds sample variants that supply the input color to the child. This
> is now the only way to invoke a colorFilter child. Internally, we
> support passing a color when invoking a child shader, but I'm not
> exposing that. It's not clearly part of the semantics of the Skia
> pipeline, and is almost never useful. It also exposes users to
> several inconsistencies (skbug.com/11942).
> 3) Because of #2, it's possible that we can't compute a reusable program
> to filter individual colors. In that case, we don't set the constant
> output for constant input optimization, and filterColor4f falls back
> to the slower base-class implementation.
>
> Bug: skia:11813 skia:11942
> Change-Id: I06c41e1b35056e486f3163a72acf6b9535d7fed4
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/401917
> Commit-Queue: Brian Osman <brianosman@google.com>
> Reviewed-by: Mike Klein <mtklein@google.com>
Bug: skia:11813 skia:11942
Change-Id: I2c31b147ed86fa8c4dddefb7066bc1d07fe0d285
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/404637
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
This reverts commit adadb95a9f.
Reason for revert: breaking android
Original change's description:
> Better first-class shader & color filter support in runtime effects
>
> This does a few things, because they're all intertwined:
>
> 1) SkRuntimeEffect's API now includes details about children (which Skia
> stage was declared, not just the name). The factories verify that the
> declared types in the SkSL match up with the C++ types being passed.
> Today, we still only support adding children of the same type, so the
> checks are simple. Once we allow mixing types, we'll be testing the
> declared type against the actual C++ type supplied for each slot.
> 2) Adds sample variants that supply the input color to the child. This
> is now the only way to invoke a colorFilter child. Internally, we
> support passing a color when invoking a child shader, but I'm not
> exposing that. It's not clearly part of the semantics of the Skia
> pipeline, and is almost never useful. It also exposes users to
> several inconsistencies (skbug.com/11942).
> 3) Because of #2, it's possible that we can't compute a reusable program
> to filter individual colors. In that case, we don't set the constant
> output for constant input optimization, and filterColor4f falls back
> to the slower base-class implementation.
>
> Bug: skia:11813 skia:11942
> Change-Id: I06c41e1b35056e486f3163a72acf6b9535d7fed4
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/401917
> Commit-Queue: Brian Osman <brianosman@google.com>
> Reviewed-by: Mike Klein <mtklein@google.com>
TBR=mtklein@google.com,bsalomon@google.com,brianosman@google.com
Change-Id: I94ba57e73305b2302f86fd0c1d76f667d4e45b92
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:11813 skia:11942
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/404117
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>
This does a few things, because they're all intertwined:
1) SkRuntimeEffect's API now includes details about children (which Skia
stage was declared, not just the name). The factories verify that the
declared types in the SkSL match up with the C++ types being passed.
Today, we still only support adding children of the same type, so the
checks are simple. Once we allow mixing types, we'll be testing the
declared type against the actual C++ type supplied for each slot.
2) Adds sample variants that supply the input color to the child. This
is now the only way to invoke a colorFilter child. Internally, we
support passing a color when invoking a child shader, but I'm not
exposing that. It's not clearly part of the semantics of the Skia
pipeline, and is almost never useful. It also exposes users to
several inconsistencies (skbug.com/11942).
3) Because of #2, it's possible that we can't compute a reusable program
to filter individual colors. In that case, we don't set the constant
output for constant input optimization, and filterColor4f falls back
to the slower base-class implementation.
Bug: skia:11813 skia:11942
Change-Id: I06c41e1b35056e486f3163a72acf6b9535d7fed4
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/401917
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
All internal usage has migrated to MakeFor..., this removes the old
program kind, and updates some tests.
Bug: skia:11813
Change-Id: I56733b071270e1ae3fab5d851e23acf6c02e3361
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/402536
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
Simplifies SampleUsage quite a bit (no need to track multiple kinds of
sampling, variable matrices don't exist any more, etc...).
Change-Id: I58b8de7218d00c4d882d2650672e5fe01892a062
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/402177
Reviewed-by: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
New entry point that does stricter checking on SkSL validity.
Bug: skia:11813
Change-Id: Icc8501c108af278e2fe1029859a552ea6ab6eb08
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/401056
Reviewed-by: Kevin Lubick <kjlubick@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
For now, just bolt this onto the existing runtime effects. The next step
is to add dedicated modes to the compiler for shader vs. color filter.
Once we get there, we will be much more strict about main signature in
each mode (and start adding other per-mode error checking).
Bug: skia:11813
Change-Id: I27e27600209e9844ae107364baea2fb949b47c3f
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/395838
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
We are up to having seven distinct types of codegen, and will soon have
an 8th (DSL C++).
Change-Id: I6758328390c234ba1d5c30c118199dbc820af52a
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/395817
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Previously, the act of painting a Runtime Effect was causing its helper
functions to get inlined, even if inlining was disabled during the
initial SkSL generation. This meant that the "NoInline" path was not
actually very effective.
Change-Id: If8e3933be61df4a49d2e11d916d7fff22876315e
Bug: skia:11362
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/388099
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
This was landed but not turned on. D'oh!
Bug: skia:10286
Change-Id: I65682370046c87c854d806253db32795ef3a9d14
Cq-Include-Trybots: luci.skia.skia.primary:Fuzz-Debian10-Clang
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/375736
Reviewed-by: Kevin Lubick <kjlubick@google.com>
Commit-Queue: Adlai Holler <adlai@google.com>
This helps us find any issues with the promise image sharing.
Bug: skia:10286
Change-Id: I393655c2de76f896d9f376765894f84c015b2760
Cq-Include-Trybots: luci.skia.skia.primary:Fuzz-Debian10-Clang
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/374317
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Adlai Holler <adlai@google.com>
Bug: skia:11295
Change-Id: Iec11f3f4d26eb5b1c07707b3cedd09096bad80d0
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/371478
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Change-Id: I7be9a95bc190760245966c36ed088afd68108a5f
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/371316
Auto-Submit: Brian Osman <brianosman@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
This change will allow these types to be forward-declared; C++ doesn't
allow forward declaration of types declared inside a struct. Moving
these types out of Programs resulted in a large diff.
The Settings::Value helper class has been moved inside of the
IRGenerator. In practice, it was actually just an implementation detail
of how IRGenerator looks up caps-values by name. It seems very unlikely
that this will be necessary elsewhere going forward.
Change-Id: I6119417fae608f1c492a27de746d2b550ef8ca20
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/370836
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Small diffs in a GM -- due to the dual nature of kMedium (cpu and gpu)
Another reason to stop using it, and switch to the more-explicit
sampling.
Bug: skia:7650
Change-Id: Ie7575071b19778626da7f94804abaaa11861a050
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/370259
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Reed <reed@google.com>
Uses the pipeline-stage callback mechanism. It mangles the type name
(with a test to verify that this works), and then calls defineStruct
with the entire SkSL struct definition string.
Bug: skia:10939
Change-Id: If14cf1b11faaa80ad8d4086cdacf68532bac43fc
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/368809
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
Structs, enums, and global variables are all going to require name
mangling. Move that to a separate virtual on the callback. More
importantly, the generator is going to need to do type-name substitution
inside the function declaration string, so the contract has shifted:
The generator constructs the entire function declaration line (using the
mangled name it gets from the new callback), then it calls
defineFunction with two strings, and a flag indicating if this is main
or not.
Bug: skia:10939, skia:11295, skia:11296
Change-Id: I535eee9bfbb2337013b539908fe3d658ec3b2dbd
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/368397
Reviewed-by: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>