[tools][wasm] Add "wami", the Wasm Module Inspector

Initial feature: list functions in a module, as follows:

$ gm x64.release wami
$ out/x64.release/wami --list-functions my_module.wasm

More to come.

Change-Id: I9580437d51153e1b5ccc291fdb6a6a67315be07d
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3742700
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Manos Koukoutos <manoskouk@chromium.org>
Cr-Commit-Position: refs/heads/main@{#81515}

src/wasm/module-decoder-impl.h

@@ -35,7 +35,7 @@ constexpr char kBranchHintsString[] = "metadata.code.branch_hint";
 constexpr char kDebugInfoString[] = ".debug_info";
 constexpr char kExternalDebugInfoString[] = "external_debug_info";
 
-const char* ExternalKindName(ImportExportKindCode kind) {
+inline const char* ExternalKindName(ImportExportKindCode kind) {
   switch (kind) {
     case kExternalFunction:
       return "function";
@@ -51,7 +51,7 @@ const char* ExternalKindName(ImportExportKindCode kind) {
   return "unknown";
 }
 
-const char* SectionName(SectionCode code) {
+inline const char* SectionName(SectionCode code) {
   switch (code) {
     case kUnknownSectionCode:
       return "Unknown";
@@ -100,7 +100,7 @@ const char* SectionName(SectionCode code) {
   }
 }
 
-bool validate_utf8(Decoder* decoder, WireBytesRef string) {
+inline bool validate_utf8(Decoder* decoder, WireBytesRef string) {
   return unibrow::Utf8::ValidateEncoding(
       decoder->start() + decoder->GetBufferRelativeOffset(string.offset()),
       string.length());
@@ -110,7 +110,8 @@ enum class StringValidation { kNone, kUtf8, kWtf8 };
 
 // Reads a length-prefixed string, checking that it is within bounds. Returns
 // the offset of the string, and the length as an out parameter.
-WireBytesRef consume_string(Decoder* decoder, StringValidation validation,
+inline WireBytesRef consume_string(Decoder* decoder,
+                                   StringValidation validation,
                                    const char* name) {
   uint32_t length = decoder->consume_u32v("string length");
   uint32_t offset = decoder->pc_offset();
@@ -138,11 +139,11 @@ WireBytesRef consume_string(Decoder* decoder, StringValidation validation,
   return {offset, decoder->failed() ? 0 : length};
 }
 
-WireBytesRef consume_utf8_string(Decoder* decoder, const char* name) {
+inline WireBytesRef consume_utf8_string(Decoder* decoder, const char* name) {
   return consume_string(decoder, StringValidation::kUtf8, name);
 }
 
-SectionCode IdentifyUnknownSectionInternal(Decoder* decoder) {
+inline SectionCode IdentifyUnknownSectionInternal(Decoder* decoder) {
   WireBytesRef string = consume_utf8_string(decoder, "section name");
   if (decoder->failed()) {
     return kUnknownSectionCode;

src/wasm/module-decoder.cc

@@ -76,6 +76,15 @@ ModuleResult DecodeWasmModule(
   return result;
 }
 
+ModuleResult DecodeWasmModuleForDisassembler(const byte* module_start,
+                                             const byte* module_end,
+                                             AccountingAllocator* allocator) {
+  constexpr bool verify_functions = false;
+  ModuleDecoderImpl decoder(WasmFeatures::All(), module_start, module_end,
+                            kWasmOrigin);
+  return decoder.DecodeModule(nullptr, allocator, verify_functions);
+}
+
 ModuleDecoder::ModuleDecoder(const WasmFeatures& enabled)
     : enabled_features_(enabled) {}
 

src/wasm/module-decoder.h

@@ -91,6 +91,10 @@ V8_EXPORT_PRIVATE ModuleResult DecodeWasmModule(
     Counters* counters, std::shared_ptr<metrics::Recorder> metrics_recorder,
     v8::metrics::Recorder::ContextId context_id, DecodingMethod decoding_method,
     AccountingAllocator* allocator);
+// Stripped down version for disassembler needs.
+V8_EXPORT_PRIVATE ModuleResult DecodeWasmModuleForDisassembler(
+    const byte* module_start, const byte* module_end,
+    AccountingAllocator* allocator);
 
 // Exposed for testing. Decodes a single function signature, allocating it
 // in the given zone. Returns {nullptr} upon failure.

src/wasm/names-provider.h

@@ -22,7 +22,7 @@ namespace wasm {
 class DecodedNameSection;
 class StringBuilder;
 
-class NamesProvider {
+class V8_EXPORT_PRIVATE NamesProvider {
  public:
   // {kWasmInternal}: only return raw name from name section.
   // {kDevTools}: prepend '$', use import/export names as fallback,

tools/BUILD.gn

@@ -18,6 +18,10 @@ group("gn_all") {
   if (is_win) {
     data_deps += [ "v8windbg" ]
   }
+
+  if (v8_enable_webassembly) {
+    data_deps += [ "wasm:wami" ]
+  }
 }
 
 group("v8_check_static_initializers") {

tools/dev/gm.py

@@ -56,9 +56,11 @@ MODES = {
 # Modes that get built/run when you don't specify any.
 DEFAULT_MODES = ["release", "debug"]
 # Build targets that can be manually specified.
-TARGETS = ["d8", "cctest", "unittests", "v8_fuzzers", "wasm_api_tests", "wee8",
+TARGETS = [
+    "d8", "cctest", "unittests", "v8_fuzzers", "wasm_api_tests", "wee8",
     "mkgrokdump", "generate-bytecode-expectations", "inspector-test",
-           "bigint_shell"]
+    "bigint_shell", "wami"
+]
 # Build targets that get built when you don't specify any (and specified tests
 # don't imply any other targets).
 DEFAULT_TARGETS = ["d8"]

tools/wasm/BUILD.gn

+# Copyright 2022 the V8 project authors. All rights reserved.
+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+import("../../gni/v8.gni")
+
+config("internal_config") {
+  visibility = [ ":*" ]  # Only targets in this file can depend on this.
+
+  include_dirs = [
+    ".",
+    "../..",
+    "../../include/",
+    "$target_gen_dir/../..",
+  ]
+}
+
+# Wasm Module Inspector
+v8_executable("wami") {
+  testonly = true
+
+  deps = [
+    "../..:v8",
+    "../..:v8_libbase",
+    "../..:v8_libplatform",
+  ]
+
+  sources = [ "module-inspector.cc" ]
+
+  defines = []
+
+  configs = [ ":internal_config" ]
+  if (v8_enable_i18n_support) {
+    configs += [ "//third_party/icu:icu_config" ]
+  }
+}

tools/wasm/module-inspector.cc

+// Copyright 2022 the V8 project authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include <fstream>
+#include <iomanip>
+#include <iostream>
+#include <vector>
+
+#include "include/libplatform/libplatform.h"
+#include "include/v8-initialization.h"
+#include "src/wasm/module-decoder-impl.h"
+#include "src/wasm/names-provider.h"
+#include "src/wasm/string-builder.h"
+#include "src/wasm/wasm-opcodes-inl.h"
+
+int PrintHelp(char** argv) {
+  std::cerr << "Usage: Specify an action and a module name in any order.\n"
+            << "The action can be any of:\n"
+
+            << " --help\n"
+            << "     Print this help and exit.\n"
+
+            << " --list-functions\n"
+            << "     List functions in the given module\n"
+
+            << "The module name must be a file name.\n";
+  return 1;
+}
+
+namespace v8 {
+namespace internal {
+namespace wasm {
+
+class FormatConverter {
+ public:
+  explicit FormatConverter(std::string path) {
+    std::ifstream input(path, std::ios::binary);
+    if (!input.is_open()) {
+      std::cerr << "Failed to open " << path << "!\n";
+      return;
+    }
+    raw_bytes_ = std::vector<byte>(std::istreambuf_iterator<char>(input), {});
+    if (raw_bytes_.size() < 8 || raw_bytes_[0] != 0 || raw_bytes_[1] != 'a' ||
+        raw_bytes_[2] != 's' || raw_bytes_[3] != 'm') {
+      std::cerr << "That's not a Wasm module!\n";
+      return;
+    }
+    base::Vector<const byte> wire_bytes(raw_bytes_.data(), raw_bytes_.size());
+    ModuleResult result =
+        DecodeWasmModuleForDisassembler(start(), end(), &allocator_);
+    if (result.failed()) {
+      WasmError error = result.error();
+      std::cerr << "Decoding error: " << error.message() << " at offset "
+                << error.offset() << "\n";
+      // TODO(jkummerow): Show some disassembly.
+      return;
+    }
+    ok_ = true;
+    module_ = result.value();
+    names_provider_ =
+        std::make_unique<NamesProvider>(module_.get(), wire_bytes);
+  }
+
+  bool ok() const { return ok_; }
+
+  void ListFunctions() {
+    DCHECK(ok_);
+    const WasmModule* m = module();
+    uint32_t num_functions = static_cast<uint32_t>(m->functions.size());
+    std::cout << "There are " << num_functions << " functions ("
+              << m->num_imported_functions << " imported, "
+              << m->num_declared_functions
+              << " locally defined); the following have names:\n";
+    for (uint32_t i = 0; i < num_functions; i++) {
+      StringBuilder sb;
+      names()->PrintFunctionName(sb, i);
+      if (sb.length() == 0) continue;
+      std::string name(sb.start(), sb.length());
+      std::cout << i << " " << name << "\n";
+    }
+  }
+
+ private:
+  byte* start() { return raw_bytes_.data(); }
+  byte* end() { return start() + raw_bytes_.size(); }
+  const WasmModule* module() { return module_.get(); }
+  NamesProvider* names() { return names_provider_.get(); }
+
+  AccountingAllocator allocator_;
+  bool ok_{false};
+  std::vector<byte> raw_bytes_;
+  std::shared_ptr<WasmModule> module_;
+  std::unique_ptr<NamesProvider> names_provider_;
+};
+
+}  // namespace wasm
+}  // namespace internal
+}  // namespace v8
+
+using FormatConverter = v8::internal::wasm::FormatConverter;
+
+enum class Action {
+  kUnset,
+  kHelp,
+  kListFunctions,
+};
+
+struct Options {
+  const char* filename = nullptr;
+  Action action = Action::kUnset;
+};
+
+void ListFunctions(const Options& options) {
+  FormatConverter fc(options.filename);
+  if (fc.ok()) fc.ListFunctions();
+}
+
+int ParseOptions(int argc, char** argv, Options* options) {
+  for (int i = 1; i < argc; i++) {
+    if (strcmp(argv[i], "--help") == 0 || strcmp(argv[i], "-h") == 0 ||
+        strcmp(argv[i], "help") == 0) {
+      options->action = Action::kHelp;
+    } else if (strcmp(argv[i], "--list-functions") == 0) {
+      options->action = Action::kListFunctions;
+    } else if (options->filename != nullptr) {
+      return PrintHelp(argv);
+    } else {
+      options->filename = argv[i];
+    }
+  }
+  if (options->action == Action::kUnset || options->filename == nullptr) {
+    return PrintHelp(argv);
+  }
+  return 0;
+}
+
+int main(int argc, char** argv) {
+  Options options;
+  if (ParseOptions(argc, argv, &options) != 0) return 1;
+  // Bootstrap the basics.
+  v8::V8::InitializeICUDefaultLocation(argv[0]);
+  v8::V8::InitializeExternalStartupData(argv[0]);
+  std::unique_ptr<v8::Platform> platform = v8::platform::NewDefaultPlatform();
+  v8::V8::InitializePlatform(platform.get());
+#ifdef V8_ENABLE_SANDBOX
+  if (!v8::V8::InitializeSandbox()) {
+    fprintf(stderr, "Error initializing the V8 sandbox\n");
+    return 1;
+  }
+#endif
+  v8::V8::Initialize();
+
+  switch (options.action) {
+    // clang-format off
+    case Action::kHelp:          PrintHelp(argv);             break;
+    case Action::kListFunctions: ListFunctions(options);      break;
+    case Action::kUnset:         UNREACHABLE();
+      // clang-format on
+  }
+
+  v8::V8::Dispose();
+  v8::V8::DisposePlatform();
+  return 0;
+}