vasilito
f26371ae30
libclc was vendored at recipes/dev/llvm21/source/libclc/ as a
build artifact of the llvm21 fork. This promoted it from an
implicit Mesa dependency (the CLC bitcode Mesa consumes) to a
first-class local project, mirroring how tlc, redbear-*, cub, etc.
are organized per local/AGENTS.md.
Local layout follows the existing in-house pattern:
recipes/dev/libclc -> ../../local/recipes/dev/libclc
local/recipes/dev/libclc/recipe.toml
local/recipes/dev/libclc/source/ (vendored from llvm-project)
The recipe builds libclc as a host-side CMake project using the
cookbook-sysroot clang21 binary (not the cross-compiler) to compile
OpenCL kernels to .bc bitcode. Output targets are amdgcn, amdgcn-amdhsa,
r600, generic, and spirv — the union that Mesa's clc tool consumes
for the iris/radeonsi/llvmpipe drivers. ptx-nvidiacl and clspv are
excluded because they don't map onto any Mesa hardware driver.
The recipe also wires the cookbook's native-cmake generation with
an override that points CMAKE_C_COMPILER etc. at the cookbook
sysroot's clang-21 binary. This is needed because libclc must
run on the build host (output is bitcode, not a Redox binary),
and the cookbook's default cross-toolchain would link the build
against relibc instead of host glibc.
libclc.pc is installed at usr/lib/pkgconfig/, satisfying Mesa's
dependency('libclc') lookup. The .bc files are installed under
usr/share/clc/, which is what the libclc.pc libexecdir pointer
(libexecdir=<prefix>/share/clc) expects.
This is the prerequisite for re-enabling iris, radeonsi, and the
Intel Vulkan driver in the Mesa recipe (Phase 4-6 of
local/docs/3D-DRIVER-PLAN.md). With libclc available in the
cross-sysroot, mesa's meson '-Dwith_clc -> dependency(libclc)'
resolves and the gallium drivers compile.
libclc
libclc is an open source implementation of the library requirements of the OpenCL C programming language, as specified by the OpenCL 1.1 Specification. The following sections of the specification impose library requirements:
- 6.1: Supported Data Types
- 6.2.3: Explicit Conversions
- 6.2.4.2: Reinterpreting Types Using as_type() and as_typen()
- 6.9: Preprocessor Directives and Macros
- 6.11: Built-in Functions
- 9.3: Double Precision Floating-Point
- 9.4: 64-bit Atomics
- 9.5: Writing to 3D image memory objects
- 9.6: Half Precision Floating-Point
libclc is intended to be used with the Clang compiler's OpenCL frontend.
libclc is designed to be portable and extensible. To this end, it provides generic implementations of most library requirements, allowing the target to override the generic implementation at the granularity of individual functions.
libclc currently supports PTX, AMDGPU, SPIRV and CLSPV targets, but support for more targets is welcome.
Compiling and installing
(in the following instructions you can use make or ninja)
For an in-tree build, Clang must also be built at the same time:
$ cmake <path-to>/llvm-project/llvm/CMakeLists.txt -DLLVM_ENABLE_PROJECTS="libclc;clang" \
-DCMAKE_BUILD_TYPE=Release -G Ninja
$ ninja
Then install:
$ ninja install
Note you can use the DESTDIR Makefile variable to do staged installs.
$ DESTDIR=/path/for/staged/install ninja install
To build out of tree, or in other words, against an existing LLVM build or install:
$ cmake <path-to>/llvm-project/libclc/CMakeLists.txt -DCMAKE_BUILD_TYPE=Release \
-G Ninja -DLLVM_DIR=$(<path-to>/llvm-config --cmakedir)
$ ninja
Then install as before.
In both cases this will include all supported targets. You can choose which
targets are enabled by passing -DLIBCLC_TARGETS_TO_BUILD to CMake. The default
is all.
In both cases, the LLVM used must include the targets you want libclc support for
(AMDGPU and NVPTX are enabled in LLVM by default). Apart from SPIRV where you do
not need an LLVM target but you do need the
llvm-spirv tool available.
Either build this in-tree, or place it in the directory pointed to by
LLVM_TOOLS_BINARY_DIR.