mewin/glslang
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
4,207 Commits 1 Branch 0 Tags
Commit Graph

1756 Commits

Author SHA1 Message Date
Jeff Bolz
e135a4897a More precisely check for the qualifiers that enable the vulkan memory model for buffer references. 
Not all memory qualifiers necessarily need the vulkan memory model, e.g. volatile/restrict that correspond to core SPIR-V features do not.
 2019-01-20 00:59:38 -06:00
Rex Xu
75c5603ada Fix xfb_stride incorrectness(#1654) 
Add int64 support in XFB. Change containsDouble to contains64BitType. Make
it more general.
 2019-01-14 12:02:58 +08:00
David Neto
c4d5ac6bfa Update SPIRV-Tools known good: relax Uconvert check 
Permit UConvert as an OpSpecConstantOp operation when
SPV_AMD_gpu_shader_int_16 is used.
 2019-01-09 10:32:32 -05:00
John Kessenich
69932ab47f
Merge pull request #1645 from KhronosGroup/dupe-accelstruct 
Don't output duplicate OpTypeAccelerationStructureNV declarations
 2019-01-08 09:39:32 +07:00
Piers Daniell
3cf3f96292 Don't output duplicate TypeAccelerationStructureNV declarations 2019-01-07 14:17:44 -07:00
Jeff Bolz
9f2aec49e9 GL_EXT_buffer_reference 2019-01-07 12:36:13 -06:00
David Neto
082184bcc9 Mark spv.int16.amd.frag.out as failing validation 
See #1642

Also updates SPIRV-Tools and SPIRV-Headers known_good.
 2019-01-03 14:36:20 -05:00
Greg Fischer
d445bb2867 Create separate OpSource for each included file 2018-12-28 12:06:43 -07:00
John Kessenich
9ed38739b9
Merge pull request #1627 from greg-lunarg/lines8 
Add passes to spirv-opt passes to preserve source line info when compiling -g.
 2018-12-18 10:38:32 -07:00
Alan Baker
3b656c79f4 Move memory model scope tests to Vulkan 1.1 batch 
Fixes #1610

* Updated expected result due to version change
 2018-12-17 17:07:55 -05:00
Greg Fischer
5620a2f95e Add passes to legalization to preserve source line info when generated. 2018-12-16 13:13:46 -07:00
Jeff Bolz
61a0cd19cc Fix bug where shaderRecordNV buffers were using StorageBuffer storage class 2018-12-14 20:59:53 -06:00
John Kessenich
236eb0d325 GLSL/SPV: XFB: No streams on types, but support them on built-in blocks. 
From internal Khronos discussions, work, and testing.
 2018-12-13 12:06:12 -07:00
Jeff Bolz
0a93cfbb79 Update to newest SPIRV-Tools. It checks the Vulkan rule that all resources must have a set and binding decoration, which many tests fail. So add code to set a default value of zero. Also disable PCH for Ninja builds. 2018-12-11 20:58:29 -06:00
Rémi Verschelde
ebfd91a719 Remove Unix executable permission from text files 2018-12-10 10:49:16 +01:00
John Kessenich
bd1c1831d5 Manually merge ClemensRognerSD-dx9-sampler and resolve conflicts. 2018-12-07 18:38:26 -07:00
greg-lunarg
5d43c4aac7 SPV: Fix #1575, fix #1593: Support HLSL #line 
SPIR-V OpLines now contain filenames from HLSL-style #lines.
 2018-12-07 17:36:33 -07:00
Alan Baker
ed777b0fbd Modify testing to only record validation pass/fail 
* Changed unit tests to only record known the validation pass/fail
status
 * errors are output as part of the failure message if the result is
 unexpected
 * can turn off validation for each test individually
* Moved some SPV_KHR_vulkan_memory_model tests to be compiled for Vulkan
1.1
 2018-12-07 13:58:46 -05:00
Alexander Galazin
32def9f467 Rename GL_KHX_shader_explicit_arithmetic_types to GL_EXT_shader_explicit_arithmetic_types 2018-12-03 13:16:59 +01:00
John Kessenich
667506a5ea
Merge pull request #1602 from Igalia/apinheiro/prevent-xfbstride-twice 
GlslangToSpv: prevent add XfbStride twice
 2018-11-29 10:43:25 -07:00
Alejandro Piñeiro
16ef9da372 GlslangToSpv: prevent add XfbStride twice 
Without this commit, if the XfbStride was explicitly set, the
decoration was added twice on the shader.

v2 (changes after Jonh Kessenich first review)
   * Simplified by just removing the firs assignment
   * Removed assert
 2018-11-29 11:49:04 +01:00
John Kessenich
8c869679f2 SPV: For GLSL only, bitcast OpArrayLength result back to int. 
This makes the rest of the AST consuming the result match GLSL
semantics and hence get complete type matching.
 2018-11-28 07:01:37 -07:00
John Kessenich
e045c23a34 SPV: Generate OpArrayLength with no signedness. 2018-11-27 09:19:10 -07:00
John Kessenich
c7a84a9247
Merge pull request #1567 from Igalia/apinheiro/xfb-1535 
ParseHelper: assign global XfbBuffer to a block missing it
 2018-11-27 07:17:10 -07:00
John Kessenich
d7baecfab7
Merge pull request #1595 from dgkoch/EXT_fragment_invocation_density 
Implement GL_EXT_fragment_invocation_density
 2018-11-27 06:41:36 -07:00
John Kessenich
2173c650ac
Merge pull request #1597 from sheredom/add_conversion_folding 
Add conversion folding when the source is a constant.
 2018-11-27 03:53:28 -07:00
Alan Baker
440f95b1e7 Update known good SPIRV tools and headers 
* Fixes use of optimizer
* Updated test outputs
 * GeometryStreams and TransformFeedback no longer fail validation
 2018-11-26 11:24:22 -05:00
Daniel Koch
5154db5183 Implement GL_EXT_fragment_invocation_density 
including SPV generation using SPV_EXT_fragment_invocation_density.
This is an alias of the functionality in SPV_NV_shading_rate, and thus in some
cases we can only have one set of the tokens present (switch statements), so
we have picked the EXT version. This required updating the expected test
results for SPV_NV_shading_rate.

Also updated the known-good for spirv-headers so that the validator in
spirv-tools knows about the new extension.
 2018-11-26 10:01:58 -05:00
Neil Henning
81a63f1de0 Add conversion folding when the source is a constant. 
This change adds unary conversion folding when the source is a constant.
This fixes an ISV issue whereby:

```
const float16_t f = float16_t(42.0);
```

Wouldn't compile because the conversion operator would always produce an
EvqTemporary when it could have produced an EvqConst.

I've also added a test case that proves out that all basic-type to
basic-type conversions work.
 2018-11-26 12:45:33 +00:00
John Kessenich
f2675d4d6c GLSL: Fix #1591: Require the id in subgroupBroadcast to be constant. 2018-11-21 15:16:30 -07:00
David Neto
2d1739d57a Update to newer SPIRV-Tools, including VK_EXT_scalar_block_layout 
Validator has more refined messages about what kind of block layout
rules have been applied.
 2018-11-16 16:16:31 -05:00
John Kessenich
e3b76ed608 Merge branch 'scalar_block_layout' of https://github.com/jeffbolznv/glslang into jeffbolznv-scalar_block_layout 2018-11-16 09:57:33 -07:00
John Kessenich
30cb924538 Move to next SPIRV-Tools known good. 2018-11-15 23:43:35 -07:00
John Kessenich
f4ba9522ec GLSL: Increase supported sizes of XBF strides and offsets. 2018-11-15 15:12:07 -07:00
John Kessenich
579ccece45
Merge pull request #1583 from TiemoJung/fix_dot_int 
[HLSL/Spir-V] fix for incorrect spir-v on int dot(int, int)
 2018-11-15 11:29:12 -07:00
John Kessenich
75cbfcc278 Update to latest SPIR-V Tools. 2018-11-15 09:57:43 -07:00
t.jung
b16bea801c [HLSL/Spir-V] fix for incorrect spir-v on int dot(int, int) 
Decomposes OpDot into corresponding multiplies and additions.
 2018-11-15 11:02:30 +01:00
John Kessenich
e3cc0d35f3
Merge pull request #1579 from TiemoJung/handle_c_register 
HLSL: implemented c register handling
 2018-11-14 11:40:45 -07:00
John Kessenich
844e0773a3
Merge pull request #1580 from TiemoJung/float1_times_scalar_fix 
HLSL: fix invalid spir-v vector * scalar for float1
 2018-11-14 11:39:22 -07:00
Jeff Bolz
7da39ed968 Implement GL_EXT_scalar_block_layout 2018-11-14 09:30:53 -06:00
t.jung
697fdf015f HLSL: fix invalid spir-v vector * scalar for float1 
Missing check for float results in always emitting OpIMul even for float types
 2018-11-14 14:27:22 +01:00
t.jung
31cd98c7b2 HLSL: implemented c register handling 
Adjusted a test not to use c register for
a structured buffer, they are supposed to
use t registers

Added comments with hints for what are the
register types are used for.
 2018-11-14 13:12:12 +01:00
t.jung
8279f0728a Updated type conversion rules for HLSL frontend 2018-11-14 08:13:32 +01:00
Rex Xu
7bf5fcc541 Fix parser issue: redeclare gl_FragStencilRefARB is not supported 2018-11-12 15:15:57 +08:00
Alejandro Piñeiro
adba7dfc68 ParseHelper: assign global XfbBuffer to a block missing it 
If a block has assigned a XfbOffset it is assumed that it would
inherit the current global XfbBuffer. This commit fixes two use cases:

1) Getting the members of a Block with a XfbOffset to be assigned an
   offset, as explained on GLSL 4.60 spec, section "4.4.2 Output
   Layout Qualifiers", subsection "Transform Feedback Layout
   Qualifiers".

2) Compute properly an error on overlapping ranges if a block is
   assigned a XfbOffset and one of it members is assigned a explicit
   one. This gets working because when the members of a block get
   assigned a Offset/Buffer at fixBlockXfbOffsets, then the block is
   deassigned the Offsets, so ranges are computed only with the block
   members.

BTW, this is already done when redeclaring block builtins.

Fixes #1535
 2018-11-10 11:35:50 +01:00
John Kessenich
918caa7ee4
Merge pull request #1547 from Igalia/apinheiro/xfb-struct-assignment 
Apinheiro/xfb struct assignment
 2018-11-09 12:27:39 -07:00
John Kessenich
4207c97b93 Fix #1563: Corrupt token when an object macro miscalls a function macro 2018-11-07 18:25:56 -07:00
John Kessenich
4ee2f75294 PP: More clearly distinguish funtion-like and object-like macros 2018-11-07 17:22:17 -07:00
John Kessenich
8b782e77df Tests: Add test for ways of calling no-argument macro with (). 2018-11-06 14:51:27 -07:00
John Kessenich
cdf7a85b14
Merge pull request #1557 from nicebyte/spec-const-decorations 
Add names for composite spec constants in SPIR-V
 2018-11-02 10:23:22 -06:00