1717 Commits

Author SHA1 Message Date
Pankaj Mistry
228c67228a GL_EXT_mesh_shader/SPV_EXT_mesh_shader implementation
Added following updates to GL_EXT_mesh_shader implementation:

1. Added SPIRV and GLSL test cases
2. Added checks to ensure NV and EXT mesh shader builtins cannot be used interchangeably.
3. Updated the language name by removing the postfix "NV" to MeshShader and TaskShader.
4. Added checks for grammar checking to comply with the spec.

5. Added gl_NumWorkGroups builtin to Mesh shader
6. Fixed data type of gl_PrimitiveLineIndicesEXT and gl_PrimitiveTriangleIndicesEXT
7. Added new constants to the resources table
8. Updates to handle new storage qualifier "taskPayloadSharedEXT"
9. Updated test cases by replacing "taskEXT" with storage qualifier "taskPayloadSharedEXT"

Addressed  Review comments
1. Fixed instruction description used by glslang disassembly.
2. Updated OpEmitMeshTasksEXT as per spec update
3. Fixed implementation that errors out if there are more then one taskPayloadSharedEXT varjables.
4. Fixed miscellaneous error logs and removed unwanted code.

SPIRV 1.6 related build failure fixes
- Update SPIRV header to 1.6
- Fix conflict wiht SPIRV 1.6 change, where localSizeId is used for execution mode for mesh/task shaders

Enable SPIRV generated for EXT_mesh_shader to be version 1.4

GL_EXT_mesh_shader: Add checks for atomic support and corresponding test cases
2022-09-01 18:02:21 -07:00
David Neto
f28022c9f9
Avoid double-free in functions cloned for vulkan relaxed mode (#2987)
* Avoid double-free in functions cloned for vulkan relaxed mode

When rewriting function calls atomicCounterIncrement and
atoicCounterDecrement, clone the parameters so that the TParameter
'type' field is cloned.  This avoids double-free when both the original
and transformed functions are deleted by the parser.

Fixes a ubsan failure.
2022-08-02 18:07:01 -06:00
Greg Fischer
22d39cd684
Merge pull request #2977 from AMD-dwang/AMD_shader_early_and_late_fragment_tests
Add SPV_AMD_shader_early_and_late_fragment_tests
2022-07-27 11:23:16 -06:00
Thomas Aven
374c124025 Make GL_KHR_ray_query provide EOpConstructAccStruct
Previously, GL_KHR_ray_tracing was a required extension to generate
OpConvertUToAccelerationStructureKHR conversion instructions from uint64
and uvec2. However, both GL_KHR_ray_tracing and GL_KHR_ray_query should
provide this construction.

Change-Id: I6564c127fd28d9b527d334958a5adc168f5cdd9a
2022-07-21 11:00:34 +02:00
dwang102
070863af69 Add SPV_AMD_shader_early_and_late_fragment_tests 2022-07-18 14:20:18 +08:00
Greg Fischer
8e5f1ac954 Fix getEnhancedMsgs to work when HLSL not enabled
Fixes #2969
2022-07-07 11:40:02 -06:00
stusmith
ebf45697be Add support for VK_EXT_fragment_shader_barycentric 2022-05-25 13:02:02 +01:00
Malcolm Bechard
67384dd18b fix structure indexing reassignment during block merging
For EOpIndexDirectStruct binaries, we want to visit the left symbol (the
structure) before we visit the binary, so it gets updated first.
That way we are comparing the updated structure against the target
'unitType', not the original structure.
2022-05-17 00:53:52 -04:00
Greg Fischer
14f6e27304
Merge pull request #2940 from alelenv/spirv_intrinsics_rt_ops
Disable layout error check for RT ops in presence of EXT_spirv_intrinsics
2022-05-06 11:09:55 -06:00
alelenv
521216aaaa
Disable layout error check for RT ops in presence of EXT_spirv_intrinsics
Fixes #2935
2022-05-05 21:46:58 -07:00
Marius Bjorge
3015d00ee0 Adding support for GL_EXT_ray_cull_mask 2022-05-05 12:56:04 +02:00
Greg Fischer
f1d286fcc2
Merge pull request #2851 from chaoticbob/version-override
Added GLSL version override functionality and CLI
2022-03-24 17:51:54 -06:00
sfricke-samsung
610fd6edf3
Prevent Push Constant blocks being an array (#2904)
* Prevent Push Constant blocks being an array

* Add push constant array error test

Co-authored-by: Greg Fischer <greg@lunarg.com>
2022-03-23 11:42:21 -06:00
ZhiqianXia
07f677028c The first redeclarations of gl_FragCoord must appear before any
use of gl_FragCoord.
2022-03-11 14:42:11 +08:00
ZhiqianXia
438999d24f 1. refine the check for "origin_upper_left" and "pixel_center_integer"
2. gl_FragCoord can be used at ogl140 with extension "GL_ARB_fragment_coord_conventions".




Signed-off-by: ZhiqianXia <xzq0528@outlook.com>
2022-03-01 15:06:04 +08:00
Greg Fischer
79a35ace7c Don't do updatePrecision on float16_t operations
float16_t does not take GLSL precisions and SPIR-V does not
support RelaxedPrecision on float16_t.

Fixes #2894
2022-02-24 15:03:13 -07:00
Greg Fischer
bbe692e731 Improve error message for image/sampler functions for enhanced-msgs
For recent GLSL versions, if texture2D function call appears, the error
message reports an unsupported type constructor. Change message to
unsupported function. Likewise for other removed texture* function calls.
2022-02-23 15:55:49 -07:00
Malcolm Bechard
be202ef9ba avoid leaving decorations on auto-push-constants
uniform blocks that are upgraded to push_constants shouldn't have
set/binding etc. decorations applied to them
2022-02-09 18:44:22 -05:00
Greg Fischer
16e3b403aa
Merge pull request #2883 from greg-lunarg/link2
Enhance readability of error messages for GLSL
2022-02-01 15:47:19 -07:00
Greg Fischer
3005da6e6e
Merge pull request #2874 from mbechard/master
fix cases where symbols in the tree didn't get updated during block merging
2022-02-01 15:46:25 -07:00
Greg Fischer
ca0d54d51b Enhance readability of error messages for GLSL
Specifically, make GLSL link error messages more specific and output
only information relevant to the error.

Also change type printing to more closely reflect GLSL syntax. This
is the default for link error messages, but must me enabled with the
new option --enhanced-msgs for compilation error messages.

Also with --enhanced-msgs, only emit one error message per source
line.
2022-02-01 12:10:08 -07:00
Malcolm Bechard
b8338311ea fix cases where symbols in the tree didn't get updated during block merging
For GL_EXT_vulkan_glsl_relaxed. When merging the default uniform block,
there were cases where symbols in the tree wern't updated to match the
new block structure after merging blocks together.

This change traverses the symbol tree and updates any references to the
merged block.
2022-01-28 23:38:56 -05:00
Greg Fischer
9ebd8ff6c1 Add Vulkan 1.3 support
Also update known goods to Vulkan 1.3 support
Also re-enable SPIR-V 1.6 tests with vulkan1.3 target
Also re-cache SPIRV 1.6 header which somehow regressed back to 1.5
2022-01-26 17:57:31 -07:00
Greg Fischer
364b169807
Merge branch 'master' into EndStreamPrimitive 2022-01-10 11:49:19 -07:00
ZhiqianXia
0a59dd5107 Support the floatBitsToint function at GL_ARB_gpu_shader5 Extension. 2022-01-07 13:50:30 +08:00
Greg Fischer
17232d21d7
Merge pull request #2847 from jeremy-lunarg/hayes-fix-issue-2846
Unblock unsized error from assert
2022-01-05 11:34:32 -07:00
Jeremy Hayes
95e15366e7 Unblock unsized error from assert
Fix issue #2846.
2022-01-04 21:24:05 -07:00
ZhiqianXia
1b01aaaf29 EndStreamPrimitive not supported when there is #extension GL_ARB_gpu_shader5
Signed-off-by: ZhiqianXia <xzq0528@outlook.com>
2022-01-04 10:40:04 +08:00
Greg Fischer
950c6ddc11
Merge pull request #2849 from jeremy-lunarg/hayes-fix-issue-2848
Fix seg fault
2022-01-03 16:39:45 -07:00
Hai Nguyen
1f10dddac4 Revisions to GLSL version override
- Reverted public function interface changes for C++.
- Added override member variable to TShader.
- Added accessor TShader::setOverrideVersion.
- Reverted changes to tests.
2022-01-01 19:02:14 -05:00
Hai Nguyen
9a98d32366 Added GLSL version override interface and CLI
This change list allows a user to override the GLSL version from the
command line or through the C and C++ interfaces. This will override the
override happens in ProcessDeferred() before DeduceVersionProfile() so
the process should still error out if the version is insufficient for
the shader code.

- Added --glsl-version <version> to CLI.
- Added parameter to route glslVersion as override version to
  preprocessor and parse functions to C++ interface.
- Updated C interface with function to override GLSL version.
2022-01-01 04:31:44 -05:00
Jeremy Hayes
46466be045 Fix seg fault
Check types before accessing typeName.

Fix #2848.
2021-12-28 17:09:57 -07:00
Greg Fischer
7a49192d23 Add support for spirv1.6
Add command line support which enables the following behavior:

- Remap discard

  Map discard to DemoteToHelperInvocation for HLSL shaders. Map to
  OpTerminateInvocation for GLSL shaders.

- Decorate HelperInvocation with Volatile

- Use localSizeId for execution mode

  WorkGroupSize is deprecated in spirv1.6

Also update known goods to SPIRV 1.6
2021-12-16 15:01:42 -07:00
Rex Xu
6c1db7fd6c Fix validation failures of test cases regarding GL_EXT_spirv_intrinsics
1. The test cases of OpReadClockKHR is invalid. The return type is
   unsigned integer rather than signed integer.
2. When SPIR-V decorate or SPIR-V type is specified, we should avoid
   auto location mapping because the semantics are totally decided by
   SPIR-V tokens.
2021-11-25 16:24:37 +08:00
Greg Fischer
69f9dce708
Merge branch 'master' into GL_ARB_fragment_coord_conventions 2021-11-24 11:52:08 -07:00
Greg Fischer
a40e4ac3f7
Merge branch 'master' into textureQueryLOD 2021-11-22 10:39:03 -07:00
Greg Fischer
2c54c20813
Merge pull request #2782 from kevin-mccullough/FixIomapperSeperableShaderValidation
Fix issue with separable shader validation in iomapper
2021-11-19 15:19:27 -07:00
Greg Fischer
a0c7b6fe87
Merge pull request #2811 from mbechard/master
remove unneeded extra constructions now that pool allocation is fixed
2021-11-18 10:18:22 -07:00
Malcolm Bechard
eebb48f5fc remove unneeded extra constructions now that pool allocation is fixed 2021-11-17 18:11:10 -05:00
Kevin McCullough
cbab732905 Fix issue with separable shader validation in iomapper 2021-11-17 14:28:16 -08:00
Greg Fischer
74b0abb9cd
Merge pull request #2825 from greg-lunarg/i2244
Add --hlsl-dx-position-w option
2021-11-17 09:55:17 -07:00
Greg Fischer
e9564feb55 Add --hlsl-dx-position-w option
This reciprocates the w component of SV_Position in HLSL fragment shaders
to provide DirectX compatibility for HLSL shaders in Vulkan.

Fixes #2244
2021-11-16 18:42:12 -07:00
Kevin Athey
cd187e201a Initialize member TSymbol::uniqueId. 2021-11-15 16:33:08 -08:00
Greg Fischer
7141f89a89
Merge pull request #2806 from ZhiqianXia/bestFunctionMatching
#extension GL_ARB_gpu_shader5 support the implicit conversion ,
2021-11-15 13:28:53 -07:00
Greg Fischer
7d27f667a3
Merge pull request #2818 from mhillenbrand/promoteConstantUnion_signed_char
TIntermediate::promoteConstantUnion(): fix conversion to int8
2021-11-15 10:52:12 -07:00
Greg Fischer
3074ec94bf
Merge pull request #2814 from mhillenbrand/isinf_isnan2
Use intermOut.cpp's IsNan and IsInfinity for parse-time constant folding (updated)
2021-11-12 17:38:11 -07:00
Marius Hillenbrand
f1fa8afa25 TIntermediate::promoteConstantUnion(): fix conversion to int8
The signedness of type char is implementation-defined in C++. The
conversion to (signed) int8 currently uses a cast to char, which is
undefined for negative values when the type char is implemented as
unsigned. Thus, fix to cast to "signed char", which has the intended
semantic on all implementations.

Fixes #2807
2021-11-12 10:35:31 +01:00
Greg Fischer
1f8c8b88c7 Revert port of GL_EXT_shader_realtime_clock to GL_EXT_spirv_intrinsics
Could not be found by glslang library interface. Blocking SDK build.
2021-11-11 23:39:47 -07:00
Marius Hillenbrand
0eda343970 Use intermOut.cpp's IsNan and IsInfinity for parse-time constant folding (updated)
There were two implementations of isInf() and isNan(), in Constant.cpp
and in intermOut.cpp. The former only works on little-endian systems,
the latter is a wrapper for library functions and works regardless of
endianness. Move the second version into Common.h and adopt it in both
places. Thereby avoid the duplication and fix for big-endian systems.

A previous commit with the same intent and purpose had missed a required
header for builds on Windows.

On s390x, this fixes the test case
Glsl/CompileToAstTest.FromFile/constFold_frag.

Fixes #2802
2021-11-11 16:49:40 +01:00
Greg Fischer
600c5037ba
Merge pull request #2810 from greg-lunarg/push1
Generate error for binding on push_constant
2021-11-11 08:18:59 -07:00