[MLIR][WASM] Introduce an importer for Wasm binaries (llvm#152131)

First step in introducing the wasm-import target to mlir-translate. 
This is the first PR to introduce the pass, with this PR, there is very
little support for the actual WebAssembly language, it's mostly there to
introduce the skeleton of the importer. A follow-up will come with
support for a wider range of operators. It was split to make it easier
to review, since it's a good chunk of work.

---------

Co-authored-by: Luc Forget <[email protected]>
Co-authored-by: Ferdinand Lemaire <[email protected]>
Co-authored-by: Jessica Paquette <[email protected]>
Co-authored-by: Luc Forget <[email protected]>

Author: 5 people

mlir/include/mlir/InitAllTranslations.h

@@ -20,6 +20,7 @@ namespace mlir {
 
 void registerFromLLVMIRTranslation();
 void registerFromSPIRVTranslation();
+void registerFromWasmTranslation();
 void registerToCppTranslation();
 void registerToLLVMIRTranslation();
 void registerToSPIRVTranslation();
@@ -36,6 +37,7 @@ inline void registerAllTranslations() {
     registerFromLLVMIRTranslation();
     registerFromSPIRVTranslation();
     registerIRDLToCppTranslation();
+    registerFromWasmTranslation();
     registerToCppTranslation();
     registerToLLVMIRTranslation();
     registerToSPIRVTranslation();

mlir/include/mlir/Target/Wasm/WasmBinaryEncoding.h

@@ -0,0 +1,74 @@
+//===- WasmBinaryEncoding.h - Byte encodings for Wasm binary format ===----===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+// Define various flags used to encode instructions, types, etc. in
+// WebAssembly binary format.
+//
+// These encodings are defined in the WebAssembly binary format specification.
+//
+//===----------------------------------------------------------------------===//
+#ifndef MLIR_TARGET_WASMBINARYENCODING
+#define MLIR_TARGET_WASMBINARYENCODING
+
+#include <cstddef>
+
+namespace mlir {
+struct WasmBinaryEncoding {
+  /// Byte encodings for Wasm instructions.
+  struct OpCode {
+    // Locals, globals, constants.
+    static constexpr std::byte constI32{0x41};
+    static constexpr std::byte constI64{0x42};
+    static constexpr std::byte constFP32{0x43};
+    static constexpr std::byte constFP64{0x44};
+  };
+
+  /// Byte encodings of types in Wasm binaries
+  struct Type {
+    static constexpr std::byte emptyBlockType{0x40};
+    static constexpr std::byte funcType{0x60};
+    static constexpr std::byte externRef{0x6F};
+    static constexpr std::byte funcRef{0x70};
+    static constexpr std::byte v128{0x7B};
+    static constexpr std::byte f64{0x7C};
+    static constexpr std::byte f32{0x7D};
+    static constexpr std::byte i64{0x7E};
+    static constexpr std::byte i32{0x7F};
+  };
+
+  /// Byte encodings of Wasm imports.
+  struct Import {
+    static constexpr std::byte typeID{0x00};
+    static constexpr std::byte tableType{0x01};
+    static constexpr std::byte memType{0x02};
+    static constexpr std::byte globalType{0x03};
+  };
+
+  /// Byte encodings for Wasm limits.
+  struct LimitHeader {
+    static constexpr std::byte lowLimitOnly{0x00};
+    static constexpr std::byte bothLimits{0x01};
+  };
+
+  /// Byte encodings describing the mutability of globals.
+  struct GlobalMutability {
+    static constexpr std::byte isConst{0x00};
+    static constexpr std::byte isMutable{0x01};
+  };
+
+  /// Byte encodings describing Wasm exports.
+  struct Export {
+    static constexpr std::byte function{0x00};
+    static constexpr std::byte table{0x01};
+    static constexpr std::byte memory{0x02};
+    static constexpr std::byte global{0x03};
+  };
+
+  static constexpr std::byte endByte{0x0B};
+};
+} // namespace mlir
+
+#endif

mlir/include/mlir/Target/Wasm/WasmImporter.h

@@ -0,0 +1,31 @@
+//===- WasmImporter.h - Helpers to create WebAssembly emitter ---*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file defines helpers to import WebAssembly code using the WebAssembly
+// dialect.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef MLIR_TARGET_WASM_WASMIMPORTER_H
+#define MLIR_TARGET_WASM_WASMIMPORTER_H
+
+#include "mlir/IR/BuiltinOps.h"
+#include "mlir/IR/MLIRContext.h"
+#include "mlir/IR/OwningOpRef.h"
+#include "llvm/Support/SourceMgr.h"
+
+namespace mlir::wasm {
+
+/// If `source` contains a valid Wasm binary file, this function returns a
+/// a ModuleOp containing the representation of the Wasm module encoded in
+/// the source file in the `wasmssa` dialect.
+OwningOpRef<ModuleOp> importWebAssemblyToModule(llvm::SourceMgr &source,
+                                                MLIRContext *context);
+} // namespace mlir::wasm
+
+#endif // MLIR_TARGET_WASM_WASMIMPORTER_H

mlir/lib/Target/CMakeLists.txt

@@ -4,3 +4,4 @@ add_subdirectory(SPIRV)
 add_subdirectory(LLVMIR)
 add_subdirectory(LLVM)
 add_subdirectory(SMTLIB)
+add_subdirectory(Wasm)

mlir/lib/Target/Wasm/CMakeLists.txt

@@ -0,0 +1,13 @@
+add_mlir_translation_library(MLIRTargetWasmImport
+  TranslateRegistration.cpp
+  TranslateFromWasm.cpp
+
+  ADDITIONAL_HEADER_DIRS
+  ${MLIR_MAIN_INCLUDE_DIR}/Target/Wasm
+
+  LINK_LIBS PUBLIC
+  MLIRWasmSSADialect
+  MLIRIR
+  MLIRSupport
+  MLIRTranslateLib
+)