[WebAssembly] Add support for named data sections in wasm binaries

Followup to https://reviews.llvm.org/D91769 which added support
for names globals.

Differential Revision: https://reviews.llvm.org/D92909

include/llvm/BinaryFormat/Wasm.h

@@ -202,7 +202,8 @@ struct WasmSymbolInfo {
 
 enum class NameType {
   FUNCTION,
-  GLOBAL
+  GLOBAL,
+  DATA_SEGMENT,
 };
 
 struct WasmDebugName {
@@ -313,9 +314,10 @@ enum : uint8_t {
 
 // Kind codes used in the custom "name" section
 enum : unsigned {
-  WASM_NAMES_FUNCTION = 0x1,
+  WASM_NAMES_FUNCTION = 1,
-  WASM_NAMES_LOCAL = 0x2,
+  WASM_NAMES_LOCAL = 2,
-  WASM_NAMES_GLOBAL = 0x7,
+  WASM_NAMES_GLOBAL = 7,
+  WASM_NAMES_DATA_SEGMENT = 9,
 };
 
 // Kind codes used in the custom "linking" section

include/llvm/ObjectYAML/WasmYAML.h

@@ -222,6 +222,7 @@ struct NameSection : CustomSection {
 
   std::vector<NameEntry> FunctionNames;
   std::vector<NameEntry> GlobalNames;
+  std::vector<NameEntry> DataSegmentNames;
 };
 
 struct LinkingSection : CustomSection {

lib/Object/WasmObjectFile.cpp

@@ -357,6 +357,7 @@ Error WasmObjectFile::parseDylinkSection(ReadContext &Ctx) {
 Error WasmObjectFile::parseNameSection(ReadContext &Ctx) {
   llvm::DenseSet<uint64_t> SeenFunctions;
   llvm::DenseSet<uint64_t> SeenGlobals;
+  llvm::DenseSet<uint64_t> SeenSegments;
   if (FunctionTypes.size() && !SeenCodeSection) {
     return make_error<GenericBinaryError>("Names must come after code section",
                                           object_error::parse_failed);
@@ -368,11 +369,13 @@ Error WasmObjectFile::parseNameSection(ReadContext &Ctx) {
     const uint8_t *SubSectionEnd = Ctx.Ptr + Size;
     switch (Type) {
     case wasm::WASM_NAMES_FUNCTION:
-    case wasm::WASM_NAMES_GLOBAL: {
+    case wasm::WASM_NAMES_GLOBAL:
+    case wasm::WASM_NAMES_DATA_SEGMENT: {
       uint32_t Count = readVaruint32(Ctx);
       while (Count--) {
         uint32_t Index = readVaruint32(Ctx);
         StringRef Name = readString(Ctx);
+        wasm::NameType nameType = wasm::NameType::FUNCTION;
         if (Type == wasm::WASM_NAMES_FUNCTION) {
           if (!SeenFunctions.insert(Index).second)
             return make_error<GenericBinaryError>(
@@ -383,18 +386,24 @@ Error WasmObjectFile::parseNameSection(ReadContext &Ctx) {
 
           if (isDefinedFunctionIndex(Index))
             getDefinedFunction(Index).DebugName = Name;
-        } else {
+        } else if (Type == wasm::WASM_NAMES_GLOBAL) {
+          nameType = wasm::NameType::GLOBAL;
           if (!SeenGlobals.insert(Index).second)
             return make_error<GenericBinaryError>("Global named more than once",
                                                   object_error::parse_failed);
           if (!isValidGlobalIndex(Index) || Name.empty())
             return make_error<GenericBinaryError>("Invalid name entry",
                                                   object_error::parse_failed);
+        } else {
+          nameType = wasm::NameType::DATA_SEGMENT;
+          if (!SeenSegments.insert(Index).second)
+            return make_error<GenericBinaryError>(
+                "Segment named more than once", object_error::parse_failed);
+          if (Index > DataSegments.size())
+            return make_error<GenericBinaryError>("Invalid named data segment",
+                                                  object_error::parse_failed);
         }
-        wasm::NameType T = Type == wasm::WASM_NAMES_FUNCTION
+        DebugNames.push_back(wasm::WasmDebugName{nameType, Index, Name});
-                               ? wasm::NameType::FUNCTION
-                               : wasm::NameType::GLOBAL;
-        DebugNames.push_back(wasm::WasmDebugName{T, Index, Name});
       }
       break;
     }

lib/ObjectYAML/WasmEmitter.cpp

@@ -281,6 +281,19 @@ void WasmWriter::writeSectionContent(raw_ostream &OS,
       writeStringRef(NameEntry.Name, SubSection.getStream());
     }
 
+    SubSection.done();
+  }
+  if (Section.DataSegmentNames.size()) {
+    writeUint8(OS, wasm::WASM_NAMES_DATA_SEGMENT);
+
+    SubSectionWriter SubSection(OS);
+
+    encodeULEB128(Section.DataSegmentNames.size(), SubSection.getStream());
+    for (const WasmYAML::NameEntry &NameEntry : Section.DataSegmentNames) {
+      encodeULEB128(NameEntry.Index, SubSection.getStream());
+      writeStringRef(NameEntry.Name, SubSection.getStream());
+    }
+
     SubSection.done();
   }
 }

lib/ObjectYAML/WasmYAML.cpp

@@ -62,6 +62,7 @@ static void sectionMapping(IO &IO, WasmYAML::NameSection &Section) {
   IO.mapRequired("Name", Section.Name);
   IO.mapOptional("FunctionNames", Section.FunctionNames);
   IO.mapOptional("GlobalNames", Section.GlobalNames);
+  IO.mapOptional("DataSegmentNames", Section.DataSegmentNames);
 }
 
 static void sectionMapping(IO &IO, WasmYAML::LinkingSection &Section) {

tools/obj2yaml/wasm2yaml.cpp

@@ -70,9 +70,11 @@ WasmDumper::dumpCustomSection(const WasmSection &WasmSec) {
       NameEntry.Index = Name.Index;
       if (Name.Type == llvm::wasm::NameType::FUNCTION) {
         NameSec->FunctionNames.push_back(NameEntry);
-      } else {
+      } else if (Name.Type == llvm::wasm::NameType::GLOBAL) {
-        assert(Name.Type == llvm::wasm::NameType::GLOBAL);
         NameSec->GlobalNames.push_back(NameEntry);
+      } else {
+        assert(Name.Type == llvm::wasm::NameType::DATA_SEGMENT);
+        NameSec->DataSegmentNames.push_back(NameEntry);
       }
     }
     CustomSec = std::move(NameSec);