[mlir][gpu] Add the OffloadEmbeddingAttr offloading translation attr

This patch adds the offloading translation attribute. This attribute uses LLVM
offloading infrastructure to embed GPU binaries in the IR. At the program start,
the LLVM offloading mechanism registers kernels and variables with the runtime
library: CUDA RT, HIP RT, or LibOMPTarget.

The offloading mechanism relies on the runtime library to dispatch the correct
kernel based on the registered symbols.

This patch is 3/4 on introducing the OffloadEmbeddingAttr GPU translation
attribute.

Note: Ignore the base commits; those are being reviewed in PRs llvm#78057, llvm#78098,
and llvm#78073.

Author: fabianmcg

mlir/include/mlir/Dialect/GPU/IR/CompilationAttrs.td

@@ -107,4 +107,39 @@ def GPU_SelectObjectAttr : GPU_Attr<"SelectObject", "select_object", [
   let genVerifyDecl = 1;
 }
 
+//===----------------------------------------------------------------------===//
+// GPU LLVM offload attribute.
+//===----------------------------------------------------------------------===//
+def GPU_OffloadOpenMP : I32EnumAttrCase<"OpenMP", 1, "omp">;
+def GPU_OffloadCUDA : I32EnumAttrCase<"CUDA", 2, "cuda">;
+def GPU_OffloadHIP : I32EnumAttrCase<"HIP", 3, "hip">;
+def GPU_OffloadKindEnum : GPU_I32Enum<
+  "OffloadKind", "GPU offload kind", [
+    GPU_OffloadOpenMP,
+    GPU_OffloadCUDA,
+    GPU_OffloadHIP
+  ]>;
+
+def GPU_OffloadEmbeddingAttr : GPU_Attr<"OffloadEmbedding", "offload_embedding", [
+    OffloadingTranslationAttrTrait
+  ]> {
+  let description = [{
+    This GPU offloading handler uses LLVM offloading infrastructure to embed GPU
+    binaries in the IR. At program start, the LLVM offloading mechanism registers
+    kernels and variables with the runtime library: CUDA RT, HIP RT or
+    LibOMPTarget.
+    The offloading mechanism relies on the runtime library to dispatch the
+    correct kernel based on the registered symbols.
+    This offload mechanism requires to specify which runtime is being called,
+    this is done by the `kind` parameter.
+    Example:
+    ```mlir
+    gpu.binary @binary <#gpu.offload_embedding<omp>> [...]
+    gpu.binary @binary <#gpu.offload_embedding<cuda>> [...]
+    ```
+  }];
+  let parameters = (ins "gpu::OffloadKind":$kind);
+  let assemblyFormat = [{ `<` $kind `>` }];
+}
+
 #endif // GPU_COMPILATION_ATTRS

mlir/lib/Target/LLVMIR/Dialect/GPU/CMakeLists.txt

@@ -1,14 +1,17 @@
 add_mlir_translation_library(MLIRGPUToLLVMIRTranslation
   GPUToLLVMIRTranslation.cpp
-  SelectObjectAttr.cpp
+  OffloadingTranslationAttrs.cpp
 
   LINK_COMPONENTS
   Core
+  FrontendOffloading
+  Object
 
   LINK_LIBS PUBLIC
   MLIRIR
   MLIRGPUDialect
   MLIRLLVMDialect
   MLIRSupport
+  MLIRTargetLLVM
   MLIRTargetLLVMIRExport
   )