Go to file
2024-06-28 11:00:54 -04:00
.github Bump lukka/get-cmake from 3.29.5 to 3.29.6 2024-06-26 13:18:45 -04:00
build_overrides Add basic GN configurations 2020-10-16 13:15:47 +01:00
External Fix continous_deployment 2023-12-05 13:24:11 -05:00
glslang Adding the preprocessor value DISABLE_THREAD_SUPPORT to allow compilation of glslang without thread support for WASI. 2024-06-28 11:00:54 -04:00
gtests Move the spv.debugInfo.frag test back into runtests 2024-06-14 19:01:46 -04:00
kokoro kokoro: explicitly set ENABLE_OPT=0 in cmake builds 2023-09-06 13:16:16 -04:00
ndk_test Remove OGLCompiler and HLSL stub libraries from build 2023-11-29 08:20:27 -05:00
SPIRV refactor: Make type component/constituent counts unsigned 2024-06-24 13:52:56 -04:00
StandAlone Support files with UTF8BOM character 2024-03-13 13:03:10 -04:00
Test Generate vector constructions more efficiently when sizes match 2024-06-24 13:52:56 -04:00
_config.yml Make file formatting comply with POSIX and Unix standards 2020-03-21 03:20:25 -04:00
.clang-format clang-format correction and typo (clang format likely not complete) 2019-04-19 00:13:43 -06:00
.gitattributes Remove executable bits from code/data files (#2420) 2020-10-12 10:08:47 -06:00
.gitignore git: Ignore CMakeUserPresets.json 2023-07-24 18:43:25 -06:00
.gn Add basic GN configurations 2020-10-16 13:15:47 +01:00
.mailmap Add a .mailmap file 2023-03-25 12:26:55 -04:00
Android.mk Remove OGLCompiler and HLSL stub libraries from build 2023-11-29 08:20:27 -05:00
build_info.h.tmpl Fix version check macros 2022-07-12 16:59:22 +02:00
build_info.py Explicitly use Python 3 for scripts 2022-02-10 09:36:50 -08:00
BUILD.gn Remove OGLCompiler and HLSL stub libraries from build 2023-11-29 08:20:27 -05:00
CHANGES.md Update CHANGES for 14.3.0 2024-06-25 18:42:43 -04:00
CMakeLists.txt cmake: Enable implicit fallthrough warnings for gcc and clang 2024-02-20 19:12:06 -05:00
CODE_OF_CONDUCT.md Create CODE_OF_CONDUCT.md 2018-10-10 15:48:43 -04:00
DEPS Fix GN build and presubmits 2020-10-21 15:49:08 +01:00
gen_extension_headers.py Use nullptr where possible instead of NULL or 0 2022-11-30 09:33:28 -07:00
known_good_khr.json Use googletest version 1.14.0 2023-10-31 16:47:37 -04:00
known_good.json Update known_good.json 2024-06-25 13:59:16 -04:00
license-checker.cfg Implement NonSemantic.Shader.DebugInfo.100 2022-08-26 16:17:54 -06:00
LICENSE.txt Update LICENSE.txt 2022-01-13 12:14:34 +00:00
parse_version.cmake Code review. Thx ben-clayton for comments. 2021-02-09 12:14:52 +03:00
README-spirv-remap.txt Update the README-spirv-remap.txt 2024-04-15 20:26:17 -04:00
README.md readme: added a note about STACK_SIZE Emscripten compiler setting 2024-04-23 19:08:45 -04:00
SECURITY.md Create a Security Policy (#3169) 2023-07-13 13:58:05 -04:00
standalone.gclient Add basic GN configurations 2020-10-16 13:15:47 +01:00
update_glslang_sources.py Fix update_glslang_sources.py to not use distutils 2023-11-07 15:51:45 -07:00

Continuous Integration Continuous Deployment OpenSSF Scorecard

News

  1. OGLCompiler and HLSL stub libraries have been fully removed from the build.

  2. OVERRIDE_MSVCCRT has been removed in favor of CMAKE_MSVC_RUNTIME_LIBRARY

Users are encouraged to utilize the standard approach via CMAKE_MSVC_RUNTIME_LIBRARY.

Glslang Components and Status

There are several components:

Reference Validator and GLSL/ESSL -> AST Front End

An OpenGL GLSL and OpenGL|ES GLSL (ESSL) front-end for reference validation and translation of GLSL/ESSL into an internal abstract syntax tree (AST).

Status: Virtually complete, with results carrying similar weight as the specifications.

HLSL -> AST Front End

An HLSL front-end for translation of an approximation of HLSL to glslang's AST form.

Status: Partially complete. Semantics are not reference quality and input is not validated. This is in contrast to the DXC project, which receives a much larger investment and attempts to have definitive/reference-level semantics.

See issue 362 and issue 701 for current status.

AST -> SPIR-V Back End

Translates glslang's AST to the Khronos-specified SPIR-V intermediate language.

Status: Virtually complete.

Reflector

An API for getting reflection information from the AST, reflection types/variables/etc. from the HLL source (not the SPIR-V).

Status: There is a large amount of functionality present, but no specification/goal to measure completeness against. It is accurate for the input HLL and AST, but only approximate for what would later be emitted for SPIR-V.

Standalone Wrapper

glslang is command-line tool for accessing the functionality above.

Status: Complete.

Tasks waiting to be done are documented as GitHub issues.

Other References

Also see the Khronos landing page for glslang as a reference front end:

https://www.khronos.org/opengles/sdk/tools/Reference-Compiler/

The above page, while not kept up to date, includes additional information regarding glslang as a reference validator.

How to Use Glslang

Execution of Standalone Wrapper

To use the standalone binary form, execute glslang, and it will print a usage statement. Basic operation is to give it a file containing a shader, and it will print out warnings/errors and optionally an AST.

The applied stage-specific rules are based on the file extension:

  • .vert for a vertex shader
  • .tesc for a tessellation control shader
  • .tese for a tessellation evaluation shader
  • .geom for a geometry shader
  • .frag for a fragment shader
  • .comp for a compute shader

For ray tracing pipeline shaders:

  • .rgen for a ray generation shader
  • .rint for a ray intersection shader
  • .rahit for a ray any-hit shader
  • .rchit for a ray closest-hit shader
  • .rmiss for a ray miss shader
  • .rcall for a callable shader

There is also a non-shader extension:

  • .conf for a configuration file of limits, see usage statement for example

Building (CMake)

Instead of building manually, you can also download the binaries for your platform directly from the main-tot release on GitHub. Those binaries are automatically uploaded by the buildbots after successful testing and they always reflect the current top of the tree of the main branch.

Dependencies

  • A C++17 compiler. (For MSVS: use 2019 or later.)
  • CMake: for generating compilation targets.
  • make: Linux, ninja is an alternative, if configured.
  • Python 3.x: for executing SPIRV-Tools scripts. (Optional if not using SPIRV-Tools and the 'External' subdirectory does not exist.)
  • bison: optional, but needed when changing the grammar (glslang.y).
  • googletest: optional, but should use if making any changes to glslang.

Build steps

The following steps assume a Bash shell. On Windows, that could be the Git Bash shell or some other shell of your choosing.

1) Check-Out this project

cd <parent of where you want glslang to be>
git clone https://github.com/KhronosGroup/glslang.git

2) Check-Out External Projects

./update_glslang_sources.py

3) Configure

Assume the source directory is $SOURCE_DIR and the build directory is $BUILD_DIR. First ensure the build directory exists, then navigate to it:

mkdir -p $BUILD_DIR
cd $BUILD_DIR

For building on Linux:

cmake -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX="$(pwd)/install" $SOURCE_DIR
# "Release" (for CMAKE_BUILD_TYPE) could also be "Debug" or "RelWithDebInfo"

For building on Android:

cmake $SOURCE_DIR -G "Unix Makefiles" -DCMAKE_INSTALL_PREFIX="$(pwd)/install" -DANDROID_ABI=arm64-v8a -DCMAKE_BUILD_TYPE=Release -DANDROID_STL=c++_static -DANDROID_PLATFORM=android-24 -DCMAKE_SYSTEM_NAME=Android -DANDROID_TOOLCHAIN=clang -DANDROID_ARM_MODE=arm -DCMAKE_MAKE_PROGRAM=$ANDROID_NDK_HOME/prebuilt/linux-x86_64/bin/make -DCMAKE_TOOLCHAIN_FILE=$ANDROID_NDK_HOME/build/cmake/android.toolchain.cmake
# If on Windows will be -DCMAKE_MAKE_PROGRAM=%ANDROID_NDK_HOME%\prebuilt\windows-x86_64\bin\make.exe
# -G is needed for building on Windows
# -DANDROID_ABI can also be armeabi-v7a for 32 bit

For building on Windows:

cmake $SOURCE_DIR -DCMAKE_INSTALL_PREFIX="$(pwd)/install"
# The CMAKE_INSTALL_PREFIX part is for testing (explained later).

The CMake GUI also works for Windows (version 3.4.1 tested).

Also, consider using git config --global core.fileMode false (or with --local) on Windows to prevent the addition of execution permission on files.

4) Build and Install

# for Linux:
make -j4 install

# for Windows:
cmake --build . --config Release --target install
# "Release" (for --config) could also be "Debug", "MinSizeRel", or "RelWithDebInfo"

If using MSVC, after running CMake to configure, use the Configuration Manager to check the INSTALL project.

If you want to enable testing via CMake set GLSLANG_TESTS=ON when configuring the build.

GLSLANG_TESTS is off by default to streamline the packaging / Vulkan SDK process.

Building (GN)

glslang can also be built with the GN build system.

1) Install depot_tools

Download depot_tools.zip, extract to a directory, and add this directory to your PATH.

2) Synchronize dependencies and generate build files

This only needs to be done once after updating glslang.

With the current directory set to your glslang checkout, type:

./update_glslang_sources.py
gclient sync --gclientfile=standalone.gclient
gn gen out/Default

3) Build

With the current directory set to your glslang checkout, type:

cd out/Default
ninja

If you need to change the GLSL grammar

The grammar in glslang/MachineIndependent/glslang.y has to be recompiled with bison if it changes, the output files are committed to the repo to avoid every developer needing to have bison configured to compile the project when grammar changes are quite infrequent. For windows you can get binaries from GnuWin32.

The command to rebuild is:

bison --defines=MachineIndependent/glslang_tab.cpp.h
      -t MachineIndependent/glslang.y
      -o MachineIndependent/glslang_tab.cpp

The above command is also available in the bash script in updateGrammar, when executed from the glslang subdirectory of the glslang repository.

Building to WASM for the Web and Node

Building a standalone JS/WASM library for the Web and Node

Use the steps in Build Steps, with the following notes/exceptions:

  • emsdk needs to be present in your executable search path, PATH for Bash-like environments:
  • Wrap cmake call: emcmake cmake
  • Set -DENABLE_OPT=OFF.
  • Set -DENABLE_HLSL=OFF if HLSL is not needed.
  • For a standalone JS/WASM library, turn on -DENABLE_GLSLANG_JS=ON.
  • To get a fully minimized build, make sure to use brotli to compress the .js and .wasm files
  • Note that by default, Emscripten allocates a very small stack size, which may cause stack overflows when compiling large shaders. Use the STACK_SIZE compiler setting to increase the stack size.

Example:

emcmake cmake -DCMAKE_BUILD_TYPE=Release -DENABLE_GLSLANG_JS=ON \
    -DENABLE_HLSL=OFF -DENABLE_OPT=OFF ..

Building glslang - Using vcpkg

You can download and install glslang using the vcpkg dependency manager:

git clone https://github.com/Microsoft/vcpkg.git
cd vcpkg
./bootstrap-vcpkg.sh
./vcpkg integrate install
./vcpkg install glslang

The glslang port in vcpkg is kept up to date by Microsoft team members and community contributors. If the version is out of date, please create an issue or pull request on the vcpkg repository.

Testing

Right now, there are two test harnesses existing in glslang: one is Google Test, one is the runtests script. The former runs unit tests and single-shader single-threaded integration tests, while the latter runs multiple-shader linking tests and multi-threaded tests.

Tests may erroneously fail or pass if using ALLOW_EXTERNAL_SPIRV_TOOLS with any commit other than the one specified in known_good.json.

Running tests

The runtests script requires compiled binaries to be installed into $BUILD_DIR/install. Please make sure you have supplied the correct configuration to CMake (using -DCMAKE_INSTALL_PREFIX) when building; otherwise, you may want to modify the path in the runtests script.

Running Google Test-backed tests:

cd $BUILD_DIR

# for Linux:
ctest

# for Windows:
ctest -C {Debug|Release|RelWithDebInfo|MinSizeRel}

# or, run the test binary directly
# (which gives more fine-grained control like filtering):
<dir-to-glslangtests-in-build-dir>/glslangtests

Running runtests script-backed tests:

cd $SOURCE_DIR/Test && ./runtests

If some tests fail with validation errors, there may be a mismatch between the version of spirv-val on the system and the version of glslang. In this case, it is necessary to run update_glslang_sources.py. See "Check-Out External Projects" above for more details.

Contributing tests

Test results should always be included with a pull request that modifies functionality.

If you are writing unit tests, please use the Google Test framework and place the tests under the gtests/ directory.

Integration tests are placed in the Test/ directory. It contains test input and a subdirectory baseResults/ that contains the expected results of the tests. Both the tests and baseResults/ are under source-code control.

Google Test runs those integration tests by reading the test input, compiling them, and then compare against the expected results in baseResults/. The integration tests to run via Google Test is registered in various gtests/*.FromFile.cpp source files. glslangtests provides a command-line option --update-mode, which, if supplied, will overwrite the golden files under the baseResults/ directory with real output from that invocation. For more information, please check gtests/ directory's README.

For the runtests script, it will generate current results in the localResults/ directory and diff them against the baseResults/. When you want to update the tracked test results, they need to be copied from localResults/ to baseResults/. This can be done by the bump shell script.

You can add your own private list of tests, not tracked publicly, by using localtestlist to list non-tracked tests. This is automatically read by runtests and included in the diff and bump process.

Programmatic Interfaces

Another piece of software can programmatically translate shaders to an AST using one of two different interfaces:

  • A new C++ class-oriented interface, or
  • The original C functional interface

The main() in StandAlone/StandAlone.cpp shows examples using both styles.

C++ Class Interface (new, preferred)

This interface is in roughly the last 1/3 of ShaderLang.h. It is in the glslang namespace and contains the following, here with suggested calls for generating SPIR-V:

const char* GetEsslVersionString();
const char* GetGlslVersionString();
bool InitializeProcess();
void FinalizeProcess();

class TShader
    setStrings(...);
    setEnvInput(EShSourceHlsl or EShSourceGlsl, stage,  EShClientVulkan or EShClientOpenGL, 100);
    setEnvClient(EShClientVulkan or EShClientOpenGL, EShTargetVulkan_1_0 or EShTargetVulkan_1_1 or EShTargetOpenGL_450);
    setEnvTarget(EShTargetSpv, EShTargetSpv_1_0 or EShTargetSpv_1_3);
    bool parse(...);
    const char* getInfoLog();

class TProgram
    void addShader(...);
    bool link(...);
    const char* getInfoLog();
    Reflection queries

For just validating (not generating code), substitute these calls:

    setEnvInput(EShSourceHlsl or EShSourceGlsl, stage,  EShClientNone, 0);
    setEnvClient(EShClientNone, 0);
    setEnvTarget(EShTargetNone, 0);

See ShaderLang.h and the usage of it in StandAlone/StandAlone.cpp for more details. There is a block comment giving more detail above the calls for setEnvInput, setEnvClient, and setEnvTarget.

C Functional Interface (original)

This interface is in roughly the first 2/3 of ShaderLang.h, and referred to as the Sh*() interface, as all the entry points start Sh.

The Sh*() interface takes a "compiler" call-back object, which it calls after building call back that is passed the AST and can then execute a back end on it.

The following is a simplified resulting run-time call stack:

ShCompile(shader, compiler) -> compiler(AST) -> <back end>

In practice, ShCompile() takes shader strings, default version, and warning/error and other options for controlling compilation.

C Functional Interface (new)

This interface is located glslang_c_interface.h and exposes functionality similar to the C++ interface. The following snippet is a complete example showing how to compile GLSL into SPIR-V 1.5 for Vulkan 1.2.

#include <glslang/Include/glslang_c_interface.h>

// Required for use of glslang_default_resource
#include <glslang/Public/resource_limits_c.h>

typedef struct SpirVBinary {
    uint32_t *words; // SPIR-V words
    int size; // number of words in SPIR-V binary
} SpirVBinary;

SpirVBinary compileShaderToSPIRV_Vulkan(glslang_stage_t stage, const char* shaderSource, const char* fileName) {
    const glslang_input_t input = {
        .language = GLSLANG_SOURCE_GLSL,
        .stage = stage,
        .client = GLSLANG_CLIENT_VULKAN,
        .client_version = GLSLANG_TARGET_VULKAN_1_2,
        .target_language = GLSLANG_TARGET_SPV,
        .target_language_version = GLSLANG_TARGET_SPV_1_5,
        .code = shaderSource,
        .default_version = 100,
        .default_profile = GLSLANG_NO_PROFILE,
        .force_default_version_and_profile = false,
        .forward_compatible = false,
        .messages = GLSLANG_MSG_DEFAULT_BIT,
        .resource = glslang_default_resource(),
    };

    glslang_shader_t* shader = glslang_shader_create(&input);

    SpirVBinary bin = {
        .words = NULL,
        .size = 0,
    };
    if (!glslang_shader_preprocess(shader, &input))	{
        printf("GLSL preprocessing failed %s\n", fileName);
        printf("%s\n", glslang_shader_get_info_log(shader));
        printf("%s\n", glslang_shader_get_info_debug_log(shader));
        printf("%s\n", input.code);
        glslang_shader_delete(shader);
        return bin;
    }

    if (!glslang_shader_parse(shader, &input)) {
        printf("GLSL parsing failed %s\n", fileName);
        printf("%s\n", glslang_shader_get_info_log(shader));
        printf("%s\n", glslang_shader_get_info_debug_log(shader));
        printf("%s\n", glslang_shader_get_preprocessed_code(shader));
        glslang_shader_delete(shader);
        return bin;
    }

    glslang_program_t* program = glslang_program_create();
    glslang_program_add_shader(program, shader);

    if (!glslang_program_link(program, GLSLANG_MSG_SPV_RULES_BIT | GLSLANG_MSG_VULKAN_RULES_BIT)) {
        printf("GLSL linking failed %s\n", fileName);
        printf("%s\n", glslang_program_get_info_log(program));
        printf("%s\n", glslang_program_get_info_debug_log(program));
        glslang_program_delete(program);
        glslang_shader_delete(shader);
        return bin;
    }

    glslang_program_SPIRV_generate(program, stage);

    bin.size = glslang_program_SPIRV_get_size(program);
    bin.words = malloc(bin.size * sizeof(uint32_t));
    glslang_program_SPIRV_get(program, bin.words);

    const char* spirv_messages = glslang_program_SPIRV_get_messages(program);
    if (spirv_messages)
        printf("(%s) %s\b", fileName, spirv_messages);

    glslang_program_delete(program);
    glslang_shader_delete(shader);

    return bin;
}

Basic Internal Operation

  • Initial lexical analysis is done by the preprocessor in MachineIndependent/Preprocessor, and then refined by a GLSL scanner in MachineIndependent/Scan.cpp. There is currently no use of flex.

  • Code is parsed using bison on MachineIndependent/glslang.y with the aid of a symbol table and an AST. The symbol table is not passed on to the back-end; the intermediate representation stands on its own. The tree is built by the grammar productions, many of which are offloaded into ParseHelper.cpp, and by Intermediate.cpp.

  • The intermediate representation is very high-level, and represented as an in-memory tree. This serves to lose no information from the original program, and to have efficient transfer of the result from parsing to the back-end. In the AST, constants are propagated and folded, and a very small amount of dead code is eliminated.

    To aid linking and reflection, the last top-level branch in the AST lists all global symbols.

  • The primary algorithm of the back-end compiler is to traverse the tree (high-level intermediate representation), and create an internal object code representation. There is an example of how to do this in MachineIndependent/intermOut.cpp.

  • Reduction of the tree to a linear byte-code style low-level intermediate representation is likely a good way to generate fully optimized code.

  • There is currently some dead old-style linker-type code still lying around.

  • Memory pool: parsing uses types derived from C++ std types, using a custom allocator that puts them in a memory pool. This makes allocation of individual container/contents just few cycles and deallocation free. This pool is popped after the AST is made and processed.

    The use is simple: if you are going to call new, there are three cases:

    • the object comes from the pool (its base class has the macro POOL_ALLOCATOR_NEW_DELETE in it) and you do not have to call delete

    • it is a TString, in which case call NewPoolTString(), which gets it from the pool, and there is no corresponding delete

    • the object does not come from the pool, and you have to do normal C++ memory management of what you new

  • Features can be protected by version/extension/stage/profile: See the comment in glslang/MachineIndependent/Versions.cpp.