776 Commits

Author SHA1 Message Date
Bas Nieuwenhuizen
c9ffeec6e3 SPV: Add NonUniform decoration for constructors.
This is missing in particular for OpSampledImage, which can be
a direct argument for texture operations.
2020-09-10 23:02:54 +02:00
Bas Nieuwenhuizen
58064311be SPV: Add NonUniform decoration for OpImages created during lowering.
By directly creating the OpImage instructions we were not propagating
the appropriate decorations.

Since this had a lot of cases I centralized the OpImage creation
a bit too.
2020-09-10 23:02:54 +02:00
Bas Nieuwenhuizen
de949a2afc SPV: Add NonUniform decorations for stores.
The direct pointer argument to stores has to use
the NonUniform decoration but we were not using
qualifiers at all to decorate the NonUniform pointer.

(Test fixes by Greg Fischer <greg@lunarg.com>)
2020-09-10 23:02:29 +02:00
John Kessenich
8f0c6bd773
Merge pull request #2369 from ezdiy/c_api_export
GLSLANG_EXPORT for C APIs.
2020-08-17 03:02:44 -06:00
johnkslang
01384725c2 Fix #2366, fix #2358, correctly separate out numerical feature checking
We need separate concepts for
 - total set of extensions ever enabled, for the back end
 - current state of extensions, for parsing
 - the set of features currently enabled for building the AST
2020-08-14 08:40:06 -06:00
Ez Diy
1fde85ab96 GLSLANG_EXPORT for C APIs.
Fixes FFI breakage introduced in #2283
2020-08-10 22:26:41 +02:00
johnkslang
a1a497ffe7 SPV: Use more correct SPV-Tools environment, partially addressing #2290 2020-08-04 03:00:32 -06:00
johnkslang
f881f08358 SPV: Fix #2363: include trailing newline named text SPV output. 2020-08-04 02:13:50 -06:00
John Kessenich
5743eed4d1
Merge pull request #2362 from ShabbyX/optimize_for_angle
Use GLSLANG_ANGLE to strip features to what ANGLE requires
2020-08-03 01:28:52 -06:00
Shahbaz Youssefi
1ef2e250fc
Use GLSLANG_ANGLE to strip features to what ANGLE requires
This change strips a few features similar to GLSLANG_WEB but doesn't
remove every detail like the latter.  It also hardcodes profile/version
to core/450.

In particular, TBuiltIns::initialize is specialized to remove most of
what is not supported or won't be supported by ANGLE.  The result of
this function is parsed with TParseContext::parseShaderStrings which is
a performance bottleneck.

This change shaves about 300KB off of ANGLE's binary size and reduces
the cost of SetupBuiltinSymbolTable to nearly a sixth.

Signed-off-by: Shahbaz Youssefi <ShabbyX@gmail.com>
2020-07-31 12:44:49 -04:00
Ben Clayton
2a44064885 Revert changes that migrate to thread_local.
iOS 8 does not support `thread_local`, which is still in use.
Another approach will have to be found.

This change is a revert of the following changes:

a3845240 - "Simplify PoolAlloc with use of thread_local."
abf92c80 - "Deprecate InitializeDll functions"
33585c87 - "Limit visibility of symbols for internal libraries"

Issue: #2346
2020-07-31 07:13:24 +01:00
John Kessenich
7923b3e3d9
Merge pull request #2353 from vkushwaha-nv/SPV_EXT_shader_atomic_float
Add changes for SPV_EXT_shader_atomic_float_add
2020-07-22 22:06:01 +07:00
Vikram Kushwaha
79b93923d2 Add changes for SPV_EXT_shader_atomic_float_add 2020-07-22 05:52:22 -07:00
Ben Clayton
33585c871d Limit visibility of symbols for internal libraries
Also remove `SPIRV/doc.cpp` from the `SPVRemapper` target as this
is part of `SPIRV`, causing ODR violations. Instead have
`SPVRemapper` link against `SPIRV`.

Fixes ODR violations.
2020-07-22 10:53:50 +01:00
johnkslang
c63502cf67 SPV: Update to the latest SPIR-V headers. 2020-07-21 21:05:25 -06:00
John Kessenich
f80ef74c80 Fix recently found non-determinism with gl_WorldToObject3x4EXT. 2020-07-14 01:44:35 -06:00
John Kessenich
4f32f93d7e Revert "Merge pull request #2330 from ShabbyX/optimize_for_angle"
This reverts commit 1ee5d1c0bb683daea245e8329fd619940a35a5b4, reversing
changes made to 906d48a7e828701e25487c9cf34418ded831c814.
2020-07-13 03:39:08 -06:00
John Kessenich
5799f5befe Fix a couple lines that were too long, to retrigger bots. 2020-07-13 02:52:11 -06:00
John Kessenich
4e13c90944 Fix #2329: don't use invalid initializers. 2020-07-13 00:35:58 -06:00
John Kessenich
1ee5d1c0bb
Merge pull request #2330 from ShabbyX/optimize_for_angle
Optimize for angle
2020-07-12 00:32:52 +07:00
Ben Clayton
fbe9a23baf Generate build information from CHANGES.md
This PR significantly reworks the way glslang is versioned.

Instead of committing changes to the `GLSLANG_MINOR_VERSION` define in
`glslang/Public/ShaderLang.h`, and using `make-revision` to generate
`GLSLANG_PATCH_LEVEL` in `glslang/Include/revision.h`, all version
information is now derived from the new `CHANGES.md` file.

`CHANGES.md` acts as the single source of truth for glslang version
information, along with a convenient place to put all release notes for
each notable change made.

`CHANGES.md` is parsed using the new `build_info.py` python script.
This script can read basic template files to produce new source files,
which it does to read the new `build_info.h.tmpl` to generate (at build
time) a glslang private header at
`<build-dir>/include/glslang/build_info.h`.
I've written generators for each of the CMake, Bazel, gn, and
`Android.mk` build scripts.

The new version code conforms to the Semantic Versioning 2.0 spec.

This new version is also used by the CMake rules to produce versioned
shared objects, including a major-versioned SONAME.

New APIs:
---------

* `glslang::GetVersion()` returns a `Version` struct with the version
  major, minor, patch and flavor.

Breaking API changes:
---------------------

* The public defines `GLSLANG_MINOR_VERSION` and `GLSLANG_PATCH_LEVEL`
  have been entirely removed.
* `glslang/Public/ShaderLang.h` and `glslang/Include/revision.h` have
  been deleted.
* Instead, `<build-dir>/include/glslang/build_info.h` is created in
  the build directory, and `<build-dir>/include` is a CMake `PUBLIC`
  (dependee-inherited) include directory for the glslang targets.
* `<build-dir>/include/glslang/build_info.h` contains the following
   new #defines:
   `GLSLANG_VERSION_MAJOR`, `GLSLANG_VERSION_MINOR`,
   `GLSLANG_VERSION_PATCH`, `GLSLANG_VERSION_FLAVOR`,
   `GLSLANG_VERSION_GREATER_THAN(major, minor, patch)`,
   `GLSLANG_VERSION_GREATER_OR_EQUAL_TO(major, minor, patch)`,
   `GLSLANG_VERSION_LESS_THAN(major, minor, patch)`,
   `GLSLANG_VERSION_LESS_OR_EQUAL_TO(major, minor, patch)`
*  The CMake install output directory contains a copy of
   `build_info.h` at: `include/glslang/build_info.h`
*  Python3 is now always required to build glslang (likely always
   required for transitive dependency builds).
2020-07-09 12:34:02 +01:00
Shahbaz Youssefi
8c49d15fbf
Use GLSLANG_ANGLE to strip features to what ANGLE requires
This change strips a few features similar to GLSLANG_WEB but doesn't
remove every detail like the latter.  It also hardcodes profile/version
to core/450.

In particular, TBuiltIns::initialize is specialized to remove most of
what is not supported or won't be supported by ANGLE.  The result of
this function is parsed with TParseContext::parseShaderStrings which is
a performance bottleneck.

This change shaves about 300KB off of ANGLE's binary size and reduces
the cost of SetupBuiltinSymbolTable to nearly a sixth.

Signed-off-by: Shahbaz Youssefi <ShabbyX@gmail.com>
2020-07-08 15:33:20 -04:00
Ben Clayton
cb261e3c21 License headers: s/Google/The Khronos Group
This was a copy-paste screwup, where the first line of the copyright had the company name was updated, but the company name mid way though was not.
2020-07-03 12:21:01 +01:00
Ben Clayton
1d21270ebf Add missing copyright headers
Add copyright headers to build files and scripts.
Simplifies automated scanning for bad license headers.
2020-07-02 16:03:48 +01:00
John Kessenich
8d3f3b7dac
Merge pull request #2302 from KhronosGroup/fix-texture-precision
SPV: RelaxedPrecision: use the result precision for texture sampling.
2020-06-30 12:41:15 -06:00
John Kessenich
d5b5215a88
Merge pull request #2300 from KhronosGroup/generalize-precision
SPV: RelaxedPrecision: Generalize fix #2293 to cover more operations.
2020-06-30 10:57:16 -06:00
Ben Clayton
b8c3386ec0 CMake: break up glslang into smaller static libs
Add `GenericCodeGen` and `MachineIndependent` static library targets.
Privately import both of these into the `glslang` target.
Privately import `MachineIndependent` into the  `SPIRV` target.

This is done to break the dependency of `libglslang.so` non-public APIs from `libspirv.so`,
which will become problematic once `glslang` hides its non-public symbols.

| File                      | Before     |     After  |
|---------------------------|-----------:|-----------:|
| `libGenericCodeGen.a`     |  -         |   `527716` |
| `libglslang.a`            | `68175944` |   `512938` |
| `libHLSL.a`               |     `1428` |     `1428` |
| `libMachineIndependent.a` |  -         | `67132202` |
| `libOGLCompiler.a`        |    `75908` |    `75908` |
| `libOSDependent.a`        |    `23768` |    `23768` |
| `libSPIRV.a`              | `15710210` | `15710210` |
| `libSPVRemapper.a`        |  `3250894` |  `3250894` |

| File                                    | Before     |     After  |
|-----------------------------------------|-----------:|-----------:|
| `libglslang-default-resource-limits.so` |   `117032` |   `117032` |
| `libglslang.so`                         | `22380688` | `22368216` |
| `libHLSL.so`                            |     `7520` |     `7520` |
| `libOGLCompiler.a`                      |    `75908` |    `75908` |
| `libOSDependent.a`                      |    `23768` |    `23768` |
| `libSPIRV.so`                           |  `7288336` | `28151016` |
| `libSPVRemapper.so`                     |  `1940208` |  `1940208` |

Issues: #2283, #1484
2020-06-30 14:57:40 +01:00
John Kessenich
12c155f303 SPV: RelaxedPrecision: use the result precision for texture sampling.
Fix #2298.

The AST has two precisions, an operation precision and a result precision.
Actual use of GLSL with mediump samplers wants the result precision, so
pick that up instead of the operation precision.
2020-06-30 07:52:05 -06:00
John Kessenich
435dd8028b SPV: RelaxedPrecision: Generalize fix #2293 to cover more operations.
This simplifies and enforces use of precision in many more places,
to help avoid accidental loss of RelaxedPrecision through intermediate
operations. Known fixes are:
- ?:
- function return values with mis-matched precision
- precision of function return values when a copy was needed to fix types
2020-06-30 02:44:52 -06:00
John Kessenich
bf6efd0316 SPV: Fix #2293: keep relaxed precision on arg passed to relaxed param
When arguments are copied to make space for a writable formal parameter,
and the formal parameter is relaxed precision, make the copy also
relaxed precision.
2020-06-26 09:05:31 -06:00
John Kessenich
4df10335e6 SPV: Partially address #2293: correct "const in" precision matching.
Track whether formal parameters declare reduced precision and match
that with arguments, and if they differ, make a copy to promote the
precision.
2020-06-26 08:37:06 -06:00
Graeme Leese
060882fd37 Update SPIR-V generator version
Change to 10 to reflect the change to generating unordered !=
operations.
2020-06-22 11:03:46 +01:00
Graeme Leese
65ce566cbc Use OpFUnordNotEqual for floating-point !=
The normal IEEE not equal operation tests whether operands are unordered
or not equal (so comparison with a NaN returns true). This corresponds
to the SPIR-V OpFUnordNotEqual, so change to using that.
2020-06-22 10:59:33 +01:00
Shahbaz Youssefi
d52dce5067
Add -g0 command line argument
Analogous to gcc, -g0 would strip all debug info.  This is done
regardless of whether optimizations are enabled.

Signed-off-by: Shahbaz Youssefi <ShabbyX@gmail.com>
2020-06-18 10:16:40 -04:00
Ben Clayton
1a6c8ecb20 C Interface: Split SPIR-V C interface to own file
This breaks a cyclic dependency between the SPIRV and glslang build targets.
2020-06-15 13:22:19 +01:00
John Kessenich
3641ff7378 Bump code gen version, due to removal of OpEntryPoint operands. 2020-06-10 07:40:56 -06:00
John Kessenich
c30d335377 Fix #2264: OpEntryPoint incorrectly including function parameters. 2020-06-10 07:17:11 -06:00
Alejandro Piñeiro
ff6dcca575
spirv: Support initializers on uniforms (#1588)
If a uniform has an initializer it will now be given as the optional
initializer operand to the OpVariable instruction.

Fixes: https://github.com/KhronosGroup/glslang/issues/1259

Signed-off-by: Neil Roberts <nroberts@igalia.com> (the code)
Signed-off-by: Alejandro Piñeiro <apinheiro@igalia.com> (the tests)
Signed-off-by: Arcady Goldmints-Orlov <agoldmints@igalia.com>

Co-authored-by: Neil Roberts <nroberts@igalia.com>
2020-06-04 01:39:31 -06:00
Ricardo Garcia
232ba0d848
Add SPIR-V capabilities needed for spec constants (#2199)
Fixes #2198.
2020-06-03 07:52:55 -06:00
rdb
d8edfd8e66
HLSL: fix handling of uniform qualifier in entry point parameters (#2254)
* HLSL: Fix handling of uniforms in entry point parameters

* HLSL: fix handling of "uniform in"

* Tests: Update baseResults of hlsl.function.frag.out for #2254

* HLSL: fix uniforms in function parameters for opaque types
2020-06-02 00:30:07 -06:00
alelenv
999d4fdcdd
Add default descriptorset decoration for acceleration structure (#2257)
variables.
2020-06-02 00:24:41 -06:00
dan sinclair
f15982ebeb
Remove install to the SPIRV/ folder. (#2256)
This CL updates the build scripts to only install to glslang/SPIRV
instead of also installing to the SPIRV/ folder. The deprecation notice
is also removed.

Note, this may cause downstream build issues if include directories have
not been updated

Fixes #1964 #2216
2020-06-01 23:21:03 -06:00
alelenv
59216d5cd8
Add support for primitive culling layout qualifier. (#2220)
* Add support for primitive culling layout qualifier.

* Add error checks for primitive flags and negative test.
2020-05-21 05:38:41 -06:00
Pankaj Mistry
2a8ead2109 Add support for extension GL_EXT_shader_implicit_conversions
Updated extension management in TIntermediate class.
2020-04-30 08:41:23 -07:00
John Kessenich
bcf6a2430e Move to SPIR-V 1.5 Rev. 3, bump revision, remove a status from README. 2020-04-27 02:12:05 -06:00
John Kessenich
52d34b4c65 Move to latest SPIR-V header, and bump glslang revision. 2020-04-24 08:46:32 -06:00
Cody Northrop
4d2298bfd7
Support multiple swizzled out operands (#2175)
Swizzled out operands were added in bbbd9a2a. This was sufficient
for most tests, but we ran into problems with umulExtended and
imulExtended, which have two.

This CL converts the tracking values to vectors so multiple operands
can be supported.

Test: KHR-GLES31.core.shader_bitfield_operation.*
Test: ctest
2020-04-13 21:59:49 -06:00
alelenv
75de196cec
Add support for EXT_ray_flags_primitive_culling. (#2173)
Fixes issue #2169.
2020-04-08 22:09:20 -06:00
John Kessenich
1fff362355 Build warning: Fix #2167: Remove nested reuse of 'unreachable'. 2020-04-01 00:46:57 -06:00
ntfs.hard
7fc409eecc bitwise on boolean 2020-03-30 01:32:23 +03:00