The error was being reported at the position of the var declaration,
rather than the position of the reference. And since the declaration
was in a module, its position was both incorrect (with respect to the
program source) and could be past the end.
Change-Id: I443b9fbbe016c43b93d457abfefd17025e451d8a
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/521522
Reviewed-by: Brian Osman <brianosman@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
This CL switches almost all instances of line tracking over to track
Positions instead. This does not yet add full range support - only the
start offsets will be correct currently. Followup CLs will extend the
ranges to fully cover their nodes.
Change-Id: Ie49aee02f35dcb30a3adb8a35f3e4914ba6939d2
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/518137
Reviewed-by: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
MSL does not support the unary "-" operator on matrix types. Similarly
the SPIR-V OpFNegate/OpSNegate operations only work on scalar and vector
type.
* An expression such as "-<mat>" is now transformed to "-1.0 * <mat>" when
generating MSL.
* The same expression now generates a component-wise negation in SPIR-V,
matching what glslang outputs for GLSL.
* A unary "+" is now treated as NOP for MSL, matching the SPIR-V backend.
An expression such as "+<expr>" is now evaluated as "<expr>".
* The shared/Negation.sksl has been moved to folding/ as much of its
contents exercise constant-folding of comparison expressions.
* The shared/UnaryPositiveNegative.sksl test has been extended to
exercise scalar and matrix types.
NOTE: The SPIR-V backend changes have caused a minor re-ordering of SSA
IDs generated when writing out a prefix-expression. The affected gold
files have been updated.
Bug: skia:12627, skia:12992
Change-Id: Iec5cdafc591aed7e49b3b52bda42a02661380bab
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/513976
Auto-Submit: Arman Uguray <armansito@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
Commit-Queue: Arman Uguray <armansito@google.com>
These tests only generate an error in the SPIR-V or GLSL backends. We
will soon enforce that everything in errors/ must actually fail to
compile.
Change-Id: Ic54707eb3bfa19287b4ed52335066fc0fbf19ec1
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/505397
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
This adds new validation rules that we were breaking.
Binding and DescriptorSet can't be applied to push constants, nor to
struct members.
Bug: skia:12670
Bug: chromium:1270328
Change-Id: I332f77717b08d9945c8e5b79c5bf649a8f5f2043
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/474056
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Our SPIR-V code generator did not implement support for negating a uint.
However, this is something that GLSL allows (as does the rest of SkSL).
I checked glslang and it uses OpSNegate here. The SPIR-V docs indicate
that OpSNegate allows any type of integer, and the validator lets it
pass, so we now use OpSNegate here as well.
http://screen/33mkq92uxAT5Xu8http://screen/4YBTh3gCWz8eZx7http://screen/388HtXyytcN5vLZ
Change-Id: I8c142018fd5e162dcd051abe1bc5d69a6e034794
Bug: oss-fuzz:37627
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/441880
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@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>
Now that the various Metal and SPIR-V bugs have been shaken out, we can
enable these tests. Knock on wood.
Change-Id: If4b4e302cfdd91464aaf00bc9639989de5e49aac
Bug: skia:11985
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/408640
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
All the pieces of the puzzle were already here to support componentwise
addition and subtraction of matrices. Division was just a forgotten gap
in the implementation and is now patched up to match + and -.
NOTE: if you read the SPIR-V output very closely, you may be surprised
that there are fewer FDiv operations than you'd expect from reading the
input SkSL. As it turns out, a preexisting optimization is rewriting
`mat / 4` into `mat * 0.25` (see line 2689), and this rewritten form can
use the dedicated MatrixTimesScalar op. So we only get componentwise
FDivs for the `4 / mat` lines in the source.
Change-Id: I011c859f5b3a031fbb95a2956f1194a5f3b3794b
Bug: skia:11985
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/408639
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Multiplication is handled just as before. Ops other than multiplication
are handled by splatting the scalar into a matrix, then performing the
op as a componentwise matrix-op-matrix binary expression.
Change-Id: I654715c45bf5c91b8e9660fdf1c1c6d6818b621a
Bug: skia:11985
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/408637
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
This is a clone of the Metal test. (This can be moved into shared/ and
enabled as a real test once the codegen is fixed.)
At present, this test generates broken code; everything is writing an
SpvOpMatrixTimesScalar opcode regardless of the actual operation being
performed.
Change-Id: If06b4196e7d9be36e41c5c60c006b2a713cc25d8
Bug: skia:11985
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/408297
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
In particular, this optimizes abs() and sign() when all inputs are known
at compile time. This resolves a TODO on a test case in
`IllegalIndexing.rts`.
Change-Id: Ica310522a85b42dc7ae255bd25004a6629d04176
Bug: skia:10835
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/405676
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Making a VectorCast from a compile-time constant will perform the cast
at compile-time instead; previously, we did not apply this optimization.
This simplified a few test outputs in subtle ways. (In particular, the
SPIR-V codegen used to occasionally decorate OpConstantComposite of
constant numbers with RelaxedPrecision, and no longer appears to do
this. This should have no effect on results either way AFAICS.)
Because we don't return VectorCast constructors containing compile-time
constant values, we do not need to implement compareConstant for this
constructor; they only wrap non-compile-time-constant expressions.
Change-Id: I28c1f337f64d6f20fb86bc0f58e225af4bd7b26c
Bug: skia:11032
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/392197
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
This reverts commit 7bba1f55e8.
Change-Id: I707a3c215f37376086e22eaa43916afeed6da4c7
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/388456
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
This reverts commit a9c187e5cc.
Change-Id: Icbfb8abdfc67fc2e6428d97a6cdede2726fb56e4
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/385596
Auto-Submit: Ethan Nicholas <ethannicholas@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
This reverts commit 9e476b798f.
Reason for revert: Angry Vulkan bots
Original change's description:
> Refactored SPIR-V RelaxedPrecision handling
>
> The RelaxedPrecision decoration is now handled by nextId(), to make it
> easier to see all spots where a RelaxedPrecision decoration might be
> necessary. The goal of this initial refactor is not to actually fix the
> issues with RelaxedPrecision decorations, but rather to lay the
> groundwork for doing so in followup CLs.
>
> The initial intent of this change was to not affect the SPIR-V at all,
> saving modifications for followups, but there ended up being three kinds
> of changes to the output:
>
> 1. Doing things at nextId() time rather than later means some
> decorations move to an earlier spot in the output. This results in
> diffs, but should not cause any behavioral changes.
> 2. We were incorrectly tagging bools as RelaxedPrecision in some
> situations. By funneling things through fewer code paths, the refactor
> would have caused this to happen in even more situations, and the code
> responsible for the bug was being rewritten in this CL anyway, so it
> seemed worth just fixing the issue as part of this change.
> 3. Funneling things through fewer code paths ended up adding
> (correct) RelaxedPrecision modifiers to binary operations that had
> previously been missing them. It seemed better to just let it happen
> than to try to maintain bug-for-bug compatibility with the previous
> approach.
>
> Change-Id: Ia9654d6b5754e2c797e02226660cb618c9189b36
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/384318
> Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
> Reviewed-by: Brian Osman <brianosman@google.com>
TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com
Change-Id: I9ada728e5fd5798bc1179640560c2e6045b7efd1
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/385158
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
The RelaxedPrecision decoration is now handled by nextId(), to make it
easier to see all spots where a RelaxedPrecision decoration might be
necessary. The goal of this initial refactor is not to actually fix the
issues with RelaxedPrecision decorations, but rather to lay the
groundwork for doing so in followup CLs.
The initial intent of this change was to not affect the SPIR-V at all,
saving modifications for followups, but there ended up being three kinds
of changes to the output:
1. Doing things at nextId() time rather than later means some
decorations move to an earlier spot in the output. This results in
diffs, but should not cause any behavioral changes.
2. We were incorrectly tagging bools as RelaxedPrecision in some
situations. By funneling things through fewer code paths, the refactor
would have caused this to happen in even more situations, and the code
responsible for the bug was being rewritten in this CL anyway, so it
seemed worth just fixing the issue as part of this change.
3. Funneling things through fewer code paths ended up adding
(correct) RelaxedPrecision modifiers to binary operations that had
previously been missing them. It seemed better to just let it happen
than to try to maintain bug-for-bug compatibility with the previous
approach.
Change-Id: Ia9654d6b5754e2c797e02226660cb618c9189b36
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/384318
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
This reverts commit a04692f69e.
Reason for revert: Angry Vulkan bots.
Original change's description:
> Fixed a number of spots where we should have been using RelaxedPrecision
>
> Our SPIR-V output was missing many RelaxedPrecision decorations, which
> was presumably impacting performance.
>
> Change-Id: Iee32d4a42f37af167fe0e45f3db94c2142129695
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/384178
> Reviewed-by: Brian Osman <brianosman@google.com>
> Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
TBR=egdaniel@google.com,brianosman@google.com,ethannicholas@google.com,johnstiles@google.com
Change-Id: If4fe945cb363c9b61b5a4abfde649a437689d2eb
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/384217
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Our SPIR-V output was missing many RelaxedPrecision decorations, which
was presumably impacting performance.
Change-Id: Iee32d4a42f37af167fe0e45f3db94c2142129695
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/384178
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Added asserts that verify we don't try to emit the same struct or array
with two different memory layout rules. Some code paths were failing to
inspect the associated variable, leading to incorrect errors about the
attached offsets of members.
Added a test case that triggered that error, and also triggers the new
asserts.
Then, fixed the underlying cause: writing out the struct definition as a
side effect of accessing a member in getLValue().
Bug: skia:11205
Change-Id: I6e5fb76ea918ec9ff10425f2d519ddbc54404b27
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/357436
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
This will allow us to load these inputs for unit testing in `dm`.
Change-Id: Id256ba7c30d3ec94b98048e47af44cf9efe580d5
Bug: skia:11009
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/357282
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
The test diffs look scary, but the only actual change is a minor
renumbering of IDs. The actual logic is the same.
Change-Id: I5ecc26c8581a4c01834932ff0291deba7d9e4618
Bug: skia:11171
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/353622
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Many of our shaders generate the same vector constant dozens of times,
e.g. Gaussian blur uses float4(1) repeatedly. This change avoids
re-emitting redundant vector constants.
Change-Id: I22a71cd8b2783fb997f52d485b49031f64ca6d96
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/350701
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
This fix is overly conservative in some situations (identity conversions
among vectors with the same component type), but fixes errors in two
existing unit test cases.
Bug: skia:11116
Change-Id: If852f8591fb26817528fdc37191c49129e17d6b3
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/347053
Reviewed-by: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Change-Id: I1a96060b2e52cddb50948a48520aab30bd097bbd
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/343577
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Change-Id: Ib9117dbd1bcd2c3581fba02416d9eabda1dfc6dd
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/341458
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: Ethan Nicholas <ethannicholas@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
GLSL only supports arrays of samplers in very limited ways; they aren't
supported at all by SkSL. We now detect arrays of opaque objects and
reject the code.
We have several paths through the IR generator that create and process
array types; the unit test covers global and local variables, and array
on the type versus array on the variable.
Change-Id: I5b45e88e31cf4005723c3bf35561622d65321f7b
Bug: skia:11008
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/339317
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
This CL addresses the root cause of the fuzzer issue, by checking for
LayoutIsSupported before getting the MemoryLayout of a type. However,
this array ought to be detected as an error everywhere, as samplers are
opaque types; at present, this code compiles without error in GLSL and
Metal. This is an issue for followup CLs.
GLSL's actual support for arrays of samplers is interesting and probably
too nuanced for us to try to emulate:
https://www.khronos.org/opengl/wiki/Data_Type_(GLSL)#Opaque_arrays
"Under GLSL version 3.30, Sampler arrays (the only opaque type 3.30
provides) can be declared, but they can only be accessed by compile-time
integral Constant Expressions. So you cannot loop over an array of
samplers, no matter what the array initializer, offset and comparison
expressions are.
Under GLSL 4.00 and above, array indices leading to an opaque value can
be accessed by non-compile-time constants, but these index values must
be dynamically uniform. The value of those indices must be the same
value, in the same execution order, regardless of any non-uniform
parameter values, for all shader invocations in the invocation group."
Change-Id: Ib382f5c3b563f996b3c8f1eb6b021b6d31fa9ce7
Bug: oss-fuzz:28107
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/339159
Commit-Queue: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
There were a surprisingly small number of dedicated SPIR-V tests.
SkSLSPIRVBadOffset was the only test that didn't already exist in the
golden outputs, although it actually contained two tests.
The SPIRVTest.cpp file has been converted to SPIRVTestbed.cpp, which can
be used for local debugging of SPIR-V issues via dm (like GLSLTestbed
and MetalTestbed).
Change-Id: I978d8a7cf5735af7f537113d2b9411ce42cfcf88
Bug: skia:10694
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/338756
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>