[mlir] Expose MLIR_CUDA_CONVERSIONS_ENABLED in mlir-config.h. (#83004)

That macro was not defined in some cases and thus yielded warnings if
compiled with `-Wundef`. In particular, they were not defined in the
BUILD files, so the GPU targets were broken when built with Bazel. This
commit exposes mentioned CMake variable through mlir-config.h and uses
the macro that is introduced with the same name. This replaces the macro
MLIR_CUDA_CONVERSIONS_ENABLED, which the CMake files previously defined
manually.

Author: ingomueller-net

mlir/CMakeLists.txt

@@ -111,8 +111,6 @@ if ("NVPTX" IN_LIST LLVM_TARGETS_TO_BUILD)
 else()
   set(MLIR_ENABLE_CUDA_CONVERSIONS 0)
 endif()
-# TODO: we should use a config.h file like LLVM does
-add_definitions(-DMLIR_CUDA_CONVERSIONS_ENABLED=${MLIR_ENABLE_CUDA_CONVERSIONS})
 
 # Build the ROCm conversions and run according tests if the AMDGPU backend
 # is available.

mlir/include/mlir/Config/mlir-config.h.cmake

@@ -29,4 +29,8 @@
 /* If set, enables PDL usage. */
 #cmakedefine01 MLIR_ENABLE_PDL_IN_PATTERNMATCH
 
+/* If set, enables CUDA-related features in CUDA-related transforms, pipelines,
+   and targets. */
+#cmakedefine01 MLIR_ENABLE_CUDA_CONVERSIONS
+
 #endif

mlir/include/mlir/InitAllPasses.h

@@ -14,6 +14,7 @@
 #ifndef MLIR_INITALLPASSES_H_
 #define MLIR_INITALLPASSES_H_
 
+#include "mlir/Config/mlir-config.h"
 #include "mlir/Conversion/Passes.h"
 #include "mlir/Dialect/AMDGPU/Transforms/Passes.h"
 #include "mlir/Dialect/Affine/Passes.h"
@@ -96,7 +97,7 @@ inline void registerAllPasses() {
   bufferization::registerBufferizationPipelines();
   sparse_tensor::registerSparseTensorPipelines();
   tosa::registerTosaToLinalgPipelines();
-#if MLIR_CUDA_CONVERSIONS_ENABLED
+#if MLIR_ENABLE_CUDA_CONVERSIONS
   gpu::registerGPUToNVVMPipeline();
 #endif
 }

mlir/lib/Dialect/GPU/Pipelines/GPUToNVVMPipeline.cpp

@@ -11,6 +11,7 @@
 //
 //===----------------------------------------------------------------------===//
 
+#include "mlir/Config/mlir-config.h"
 #include "mlir/Conversion/AffineToStandard/AffineToStandard.h"
 #include "mlir/Conversion/ArithToLLVM/ArithToLLVM.h"
 #include "mlir/Conversion/FuncToLLVM/ConvertFuncToLLVMPass.h"
@@ -38,7 +39,7 @@
 
 using namespace mlir;
 
-#if MLIR_CUDA_CONVERSIONS_ENABLED
+#if MLIR_ENABLE_CUDA_CONVERSIONS
 namespace {
 
 //===----------------------------------------------------------------------===//
@@ -127,4 +128,4 @@ void mlir::gpu::registerGPUToNVVMPipeline() {
       buildLowerToNVVMPassPipeline);
 }
 
-#endif // MLIR_CUDA_CONVERSIONS_ENABLED
+#endif // MLIR_ENABLE_CUDA_CONVERSIONS

mlir/lib/Dialect/GPU/Transforms/ModuleToBinary.cpp

@@ -13,6 +13,7 @@
 
 #include "mlir/Dialect/GPU/Transforms/Passes.h"
 
+#include "mlir/Config/mlir-config.h"
 #include "mlir/Dialect/Func/IR/FuncOps.h"
 #include "mlir/Dialect/GPU/IR/GPUDialect.h"
 #include "mlir/Dialect/LLVMIR/LLVMDialect.h"
@@ -48,7 +49,7 @@ void GpuModuleToBinaryPass::getDependentDialects(
   // Register all GPU related translations.
   registry.insert<gpu::GPUDialect>();
   registry.insert<LLVM::LLVMDialect>();
-#if MLIR_CUDA_CONVERSIONS_ENABLED == 1
+#if MLIR_ENABLE_CUDA_CONVERSIONS
   registry.insert<NVVM::NVVMDialect>();
 #endif
 #if MLIR_ROCM_CONVERSIONS_ENABLED == 1

mlir/lib/Target/LLVM/NVVM/Target.cpp

@@ -13,6 +13,7 @@
 
 #include "mlir/Target/LLVM/NVVM/Target.h"
 
+#include "mlir/Config/mlir-config.h"
 #include "mlir/Dialect/GPU/IR/GPUDialect.h"
 #include "mlir/Dialect/LLVMIR/NVVMDialect.h"
 #include "mlir/Target/LLVM/NVVM/Utils.h"
@@ -156,7 +157,7 @@ SerializeGPUModuleBase::loadBitcodeFiles(llvm::Module &module) {
   return std::move(bcFiles);
 }
 
-#if MLIR_CUDA_CONVERSIONS_ENABLED == 1
+#if MLIR_ENABLE_CUDA_CONVERSIONS
 namespace {
 class NVPTXSerializer : public SerializeGPUModuleBase {
 public:
@@ -562,7 +563,7 @@ NVPTXSerializer::moduleToObject(llvm::Module &llvmModule) {
   return compileToBinary(*serializedISA);
 #endif // MLIR_NVPTXCOMPILER_ENABLED == 1
 }
-#endif // MLIR_CUDA_CONVERSIONS_ENABLED == 1
+#endif // MLIR_ENABLE_CUDA_CONVERSIONS
 
 std::optional<SmallVector<char, 0>>
 NVVMTargetAttrImpl::serializeToObject(Attribute attribute, Operation *module,
@@ -574,15 +575,15 @@ NVVMTargetAttrImpl::serializeToObject(Attribute attribute, Operation *module,
     module->emitError("Module must be a GPU module.");
     return std::nullopt;
   }
-#if MLIR_CUDA_CONVERSIONS_ENABLED == 1
+#if MLIR_ENABLE_CUDA_CONVERSIONS
   NVPTXSerializer serializer(*module, cast<NVVMTargetAttr>(attribute), options);
   serializer.init();
   return serializer.run();
 #else
   module->emitError(
       "The `NVPTX` target was not built. Please enable it when building LLVM.");
   return std::nullopt;
-#endif // MLIR_CUDA_CONVERSIONS_ENABLED == 1
+#endif // MLIR_ENABLE_CUDA_CONVERSIONS
 }
 
 Attribute

mlir/unittests/Target/LLVM/SerializeNVVMTarget.cpp

@@ -6,6 +6,7 @@
 //
 //===----------------------------------------------------------------------===//
 
+#include "mlir/Config/mlir-config.h"
 #include "mlir/Dialect/GPU/IR/GPUDialect.h"
 #include "mlir/Dialect/LLVMIR/NVVMDialect.h"
 #include "mlir/IR/MLIRContext.h"
@@ -29,10 +30,10 @@
 using namespace mlir;
 
 // Skip the test if the NVPTX target was not built.
-#if MLIR_CUDA_CONVERSIONS_ENABLED == 0
-#define SKIP_WITHOUT_NVPTX(x) DISABLED_##x
-#else
+#if MLIR_ENABLE_CUDA_CONVERSIONS
 #define SKIP_WITHOUT_NVPTX(x) x
+#else
+#define SKIP_WITHOUT_NVPTX(x) DISABLED_##x
 #endif
 
 class MLIRTargetLLVMNVVM : public ::testing::Test {

utils/bazel/llvm-project-overlay/mlir/BUILD.bazel

@@ -6,7 +6,6 @@
 #   The MLIR "Multi-Level Intermediate Representation" Compiler Infrastructure
 
 load("@bazel_skylib//rules:expand_template.bzl", "expand_template")
-load("@bazel_skylib//rules:write_file.bzl", "write_file")
 load(
     ":build_defs.bzl",
     "cc_headers_only",
@@ -36,7 +35,10 @@ expand_template(
         "#cmakedefine01 MLIR_ENABLE_EXPENSIVE_PATTERN_API_CHECKS": "#define MLIR_ENABLE_EXPENSIVE_PATTERN_API_CHECKS 0",
         "#cmakedefine MLIR_GREEDY_REWRITE_RANDOMIZER_SEED ${MLIR_GREEDY_REWRITE_RANDOMIZER_SEED}": "/* #undef MLIR_GREEDY_REWRITE_RANDOMIZER_SEED */",
         "#cmakedefine01 MLIR_ENABLE_PDL_IN_PATTERNMATCH": "#define MLIR_ENABLE_PDL_IN_PATTERNMATCH 1",
-    },
+    } | if_cuda_available(
+        {"#cmakedefine01 MLIR_ENABLE_CUDA_CONVERSIONS": "#define MLIR_ENABLE_CUDA_CONVERSIONS 1"},
+        {"#cmakedefine01 MLIR_ENABLE_CUDA_CONVERSIONS": "#define MLIR_ENABLE_CUDA_CONVERSIONS 0"},
+    ),
     template = "include/mlir/Config/mlir-config.h.cmake",
 )
 
@@ -5468,7 +5470,6 @@ cc_library(
     srcs = ["lib/Dialect/GPU/Pipelines/GPUToNVVMPipeline.cpp"],
     hdrs = ["include/mlir/Dialect/GPU/Pipelines/Passes.h"],
     includes = ["include"],
-    local_defines = ["MLIR_CUDA_CONVERSIONS_ENABLED"],
     deps = [
         ":AffineToStandard",
         ":ArithToLLVM",
@@ -5492,6 +5493,7 @@ cc_library(
         ":Transforms",
         ":VectorToLLVM",
         ":VectorToSCF",
+        ":config",
     ],
 )
 
@@ -5541,6 +5543,7 @@ cc_library(
         ":Transforms",
         ":VCIXToLLVMIRTranslation",
         ":VectorDialect",
+        ":config",
         "//llvm:Core",
         "//llvm:MC",
         "//llvm:Support",
@@ -6176,6 +6179,7 @@ cc_library(
         ":NVVMToLLVMIRTranslation",
         ":TargetLLVM",
         ":ToLLVMIRTranslation",
+        ":config",
         "//llvm:NVPTXCodeGen",
         "//llvm:Support",
     ],
@@ -9131,6 +9135,7 @@ cc_library(
         ":VectorTransforms",
         ":X86VectorDialect",
         ":X86VectorTransforms",
+        ":config",
     ],
 )
 

utils/bazel/llvm-project-overlay/mlir/test/BUILD.bazel

@@ -552,7 +552,6 @@ cc_library(
 cc_library(
     name = "TestTransforms",
     srcs = glob(["lib/Transforms/*.cpp"]),
-    defines = ["MLIR_CUDA_CONVERSIONS_ENABLED"],
     includes = ["lib/Dialect/Test"],
     deps = [
         ":TestDialect",
@@ -579,7 +578,6 @@ cc_library(
 cc_library(
     name = "TestFuncToLLVM",
     srcs = glob(["lib/Conversion/FuncToLLVM/*.cpp"]),
-    defines = ["MLIR_CUDA_CONVERSIONS_ENABLED"],
     includes = ["lib/Dialect/Test"],
     deps = [
         ":TestDialect",
@@ -594,7 +592,6 @@ cc_library(
 cc_library(
     name = "TestOneToNTypeConversion",
     srcs = glob(["lib/Conversion/OneToNTypeConversion/*.cpp"]),
-    defines = ["MLIR_CUDA_CONVERSIONS_ENABLED"],
     includes = ["lib/Dialect/Test"],
     deps = [
         ":TestDialect",
@@ -653,7 +650,6 @@ cc_library(
 cc_library(
     name = "TestDLTI",
     srcs = glob(["lib/Dialect/DLTI/*.cpp"]),
-    defines = ["MLIR_CUDA_CONVERSIONS_ENABLED"],
     includes = ["lib/Dialect/Test"],
     deps = [
         ":TestDialect",
@@ -667,7 +663,7 @@ cc_library(
 cc_library(
     name = "TestGPU",
     srcs = glob(["lib/Dialect/GPU/*.cpp"]),
-    defines = ["MLIR_CUDA_CONVERSIONS_ENABLED"] + if_cuda_available([
+    defines = if_cuda_available([
         "MLIR_GPU_TO_CUBIN_PASS_ENABLE",
     ]),
     includes = ["lib/Dialect/Test"],
@@ -714,7 +710,6 @@ cc_library(
 cc_library(
     name = "TestLinalg",
     srcs = glob(["lib/Dialect/Linalg/*.cpp"]),
-    defines = ["MLIR_CUDA_CONVERSIONS_ENABLED"],
     includes = ["lib/Dialect/Test"],
     deps = [
         "//llvm:Support",
@@ -748,7 +743,6 @@ cc_library(
 cc_library(
     name = "TestLLVM",
     srcs = glob(["lib/Dialect/LLVM/*.cpp"]),
-    defines = ["MLIR_CUDA_CONVERSIONS_ENABLED"],
     includes = ["lib/Dialect/Test"],
     deps = [
         "//mlir:AffineToStandard",
@@ -773,7 +767,6 @@ cc_library(
 cc_library(
     name = "TestMath",
     srcs = glob(["lib/Dialect/Math/*.cpp"]),
-    defines = ["MLIR_CUDA_CONVERSIONS_ENABLED"],
     includes = ["lib/Dialect/Test"],
     deps = [
         "//mlir:ArithDialect",
@@ -790,7 +783,6 @@ cc_library(
 cc_library(
     name = "TestMathToVCIX",
     srcs = glob(["lib/Conversion/MathToVCIX/*.cpp"]),
-    defines = ["MLIR_CUDA_CONVERSIONS_ENABLED"],
     includes = ["lib/Dialect/Test"],
     deps = [
         "//mlir:ArithDialect",
@@ -807,7 +799,6 @@ cc_library(
 cc_library(
     name = "TestMemRef",
     srcs = glob(["lib/Dialect/MemRef/*.cpp"]),
-    defines = ["MLIR_CUDA_CONVERSIONS_ENABLED"],
     includes = ["lib/Dialect/Test"],
     deps = [
         ":TestDialect",
@@ -847,7 +838,6 @@ cc_library(
 cc_library(
     name = "TestNVGPU",
     srcs = glob(["lib/Dialect/NVGPU/*.cpp"]),
-    defines = ["MLIR_CUDA_CONVERSIONS_ENABLED"],
     includes = ["lib/Dialect/Test"],
     deps = [
         "//mlir:AffineDialect",
@@ -871,7 +861,6 @@ cc_library(
 cc_library(
     name = "TestSCF",
     srcs = glob(["lib/Dialect/SCF/*.cpp"]),
-    defines = ["MLIR_CUDA_CONVERSIONS_ENABLED"],
     includes = ["lib/Dialect/Test"],
     deps = [
         "//llvm:Support",
@@ -891,7 +880,6 @@ cc_library(
 cc_library(
     name = "TestArith",
     srcs = glob(["lib/Dialect/Arith/*.cpp"]),
-    defines = ["MLIR_CUDA_CONVERSIONS_ENABLED"],
     includes = ["lib/Dialect/Test"],
     deps = [
         "//mlir:ArithDialect",
@@ -908,7 +896,6 @@ cc_library(
 cc_library(
     name = "TestArmSME",
     srcs = glob(["lib/Dialect/ArmSME/*.cpp"]),
-    defines = ["MLIR_CUDA_CONVERSIONS_ENABLED"],
     includes = ["lib/Dialect/Test"],
     deps = [
         "//mlir:ArithToArmSME",
@@ -927,7 +914,6 @@ cc_library(
 cc_library(
     name = "TestBufferization",
     srcs = glob(["lib/Dialect/Bufferization/*.cpp"]),
-    defines = ["MLIR_CUDA_CONVERSIONS_ENABLED"],
     includes = ["lib/Dialect/Test"],
     deps = [
         "//mlir:BufferizationDialect",
@@ -989,7 +975,6 @@ cc_library(
 cc_library(
     name = "TestFunc",
     srcs = glob(["lib/Dialect/Func/*.cpp"]),
-    defines = ["MLIR_CUDA_CONVERSIONS_ENABLED"],
     includes = ["lib/Dialect/Test"],
     deps = [
         ":TestDialect",
@@ -1005,7 +990,6 @@ cc_library(
 cc_library(
     name = "TestTensor",
     srcs = glob(["lib/Dialect/Tensor/*.cpp"]),
-    defines = ["MLIR_CUDA_CONVERSIONS_ENABLED"],
     includes = ["lib/Dialect/Test"],
     deps = [
         "//mlir:ArithDialect",
@@ -1022,7 +1006,6 @@ cc_library(
 cc_library(
     name = "TestVector",
     srcs = glob(["lib/Dialect/Vector/*.cpp"]),
-    defines = ["MLIR_CUDA_CONVERSIONS_ENABLED"],
     includes = ["lib/Dialect/Test"],
     deps = [
         "//mlir:AffineDialect",