Merge pull request #11966 from swiftlang/wasm-imports

[lldb][Wasm] Handle imports when parsing Wasm name sections (llvm#170960)

Author: JDevlieghere

lldb/source/Plugins/ObjectFile/wasm/ObjectFileWasm.cpp

@@ -307,6 +307,41 @@ struct WasmFunction {
   uint32_t size = 0;
 };
 
+static llvm::Expected<uint32_t> ParseImports(DataExtractor &import_data) {
+  // Currently this function just returns the number of imported functions.
+  // If we want to do anything with global names in the future, we'll also
+  // need to know those.
+  llvm::DataExtractor data = import_data.GetAsLLVM();
+  llvm::DataExtractor::Cursor c(0);
+
+  llvm::Expected<uint32_t> count = GetULEB32(data, c);
+  if (!count)
+    return count.takeError();
+
+  uint32_t function_imports = 0;
+  for (uint32_t i = 0; c && i < *count; ++i) {
+    // We don't need module and field names, so we can just get them as raw
+    // strings and discard.
+    if (!GetWasmString(data, c))
+      return llvm::createStringError("failed to parse module name");
+    if (!GetWasmString(data, c))
+      return llvm::createStringError("failed to parse field name");
+
+    uint8_t kind = data.getU8(c);
+    if (kind == llvm::wasm::WASM_EXTERNAL_FUNCTION)
+      function_imports++;
+
+    // For function imports, this is a type index. For others it's different.
+    // We don't need it, just need to parse it to advance the cursor.
+    data.getULEB128(c);
+  }
+
+  if (!c)
+    return c.takeError();
+
+  return function_imports;
+}
+
 static llvm::Expected<std::vector<WasmFunction>>
 ParseFunctions(DataExtractor &data) {
   lldb::offset_t offset = 0;
@@ -410,7 +445,8 @@ static llvm::Expected<std::vector<WasmSegment>> ParseData(DataExtractor &data) {
 static llvm::Expected<std::vector<Symbol>>
 ParseNames(SectionSP code_section_sp, DataExtractor &name_data,
            const std::vector<WasmFunction> &functions,
-           std::vector<WasmSegment> &segments) {
+           std::vector<WasmSegment> &segments,
+           uint32_t num_imported_functions) {
 
   llvm::DataExtractor data = name_data.GetAsLLVM();
   llvm::DataExtractor::Cursor c(0);
@@ -434,15 +470,29 @@ ParseNames(SectionSP code_section_sp, DataExtractor &name_data,
         llvm::Expected<std::string> name = GetWasmString(data, c);
         if (!name)
           return name.takeError();
-        if (*idx >= functions.size())
+        if (*idx >= num_imported_functions + functions.size())
           continue;
-        symbols.emplace_back(
-            symbols.size(), *name, lldb::eSymbolTypeCode,
-            /*external=*/false, /*is_debug=*/false, /*is_trampoline=*/false,
-            /*is_artificial=*/false, code_section_sp,
-            functions[i].section_offset, functions[i].size,
-            /*size_is_valid=*/true, /*contains_linker_annotations=*/false,
-            /*flags=*/0);
+
+        if (*idx < num_imported_functions) {
+          symbols.emplace_back(symbols.size(), *name, lldb::eSymbolTypeCode,
+                               /*external=*/true, /*is_debug=*/false,
+                               /*is_trampoline=*/false,
+                               /*is_artificial=*/false,
+                               /*section_sp=*/lldb::SectionSP(),
+                               /*value=*/0, /*size=*/0,
+                               /*size_is_valid=*/false,
+                               /*contains_linker_annotations=*/false,
+                               /*flags=*/0);
+        } else {
+          const WasmFunction &func = functions[*idx - num_imported_functions];
+          symbols.emplace_back(symbols.size(), *name, lldb::eSymbolTypeCode,
+                               /*external=*/false, /*is_debug=*/false,
+                               /*is_trampoline=*/false, /*is_artificial=*/false,
+                               code_section_sp, func.section_offset, func.size,
+                               /*size_is_valid=*/true,
+                               /*contains_linker_annotations=*/false,
+                               /*flags=*/0);
+        }
       }
     } break;
     case llvm::wasm::WASM_NAMES_DATA_SEGMENT: {
@@ -590,6 +640,20 @@ void ObjectFileWasm::CreateSections(SectionList &unified_section_list) {
     }
   }
 
+  // Parse the import section. The number of functions is needed because the
+  // function index space used in the name section includes imports.
+  if (std::optional<section_info> info =
+          GetSectionInfo(llvm::wasm::WASM_SEC_IMPORT)) {
+    DataExtractor import_data = ReadImageData(info->offset, info->size);
+    llvm::Expected<uint32_t> num_imports = ParseImports(import_data);
+    if (!num_imports) {
+      LLDB_LOG_ERROR(log, num_imports.takeError(),
+                     "Failed to parse Wasm import section: {0}");
+    } else {
+      m_num_imported_functions = *num_imports;
+    }
+  }
+
   // Parse the data section.
   std::optional<section_info> data_info =
       GetSectionInfo(llvm::wasm::WASM_SEC_DATA);
@@ -609,7 +673,7 @@ void ObjectFileWasm::CreateSections(SectionList &unified_section_list) {
     DataExtractor names_data = ReadImageData(info->offset, info->size);
     llvm::Expected<std::vector<Symbol>> symbols = ParseNames(
         m_sections_up->FindSectionByType(lldb::eSectionTypeCode, false),
-        names_data, functions, segments);
+        names_data, functions, segments, m_num_imported_functions);
     if (!symbols) {
       LLDB_LOG_ERROR(log, symbols.takeError(),
                      "Failed to parse Wasm names: {0}");

lldb/source/Plugins/ObjectFile/wasm/ObjectFileWasm.h

@@ -146,6 +146,7 @@ class ObjectFileWasm : public ObjectFile {
   /// \}
 
   std::vector<section_info> m_sect_infos;
+  uint32_t m_num_imported_functions = 0;
   std::vector<Symbol> m_symbols;
   ArchSpec m_arch;
   UUID m_uuid;

lldb/test/Shell/Symtab/Inputs/simple.wasm.yaml

@@ -10,6 +10,7 @@
 #   int j = 2;
 #   return add(i, j);
 # }
+# Additional imports have been manually added and indexes adjusted after compiling
 --- !WASM
 FileHeader:
   Version:         0x1
@@ -29,6 +30,21 @@ Sections:
         ParamTypes:      []
         ReturnTypes:
           - I32
+  - Type:            IMPORT
+    Imports:
+      - Module:          env
+        Field:           foo
+        Kind:            FUNCTION
+        SigIndex:        0
+      - Module:          env
+        Field:           another_import
+        Kind:            FUNCTION
+        SigIndex:        0
+      - Module:          env
+        Field:           bar
+        Kind:            GLOBAL
+        GlobalType:      I32
+        GlobalMutable:   true
   - Type:            FUNCTION
     FunctionTypes:   [ 0, 1, 2, 1 ]
   - Type:            TABLE
@@ -44,73 +60,73 @@ Sections:
       - Minimum:         0x2
   - Type:            GLOBAL
     Globals:
-      - Index:           0
+      - Index:           1
         Type:            I32
         Mutable:         true
         InitExpr:
           Opcode:          I32_CONST
           Value:           66576
-      - Index:           1
+      - Index:           2
         Type:            I32
         Mutable:         false
         InitExpr:
           Opcode:          I32_CONST
           Value:           1036
-      - Index:           2
+      - Index:           3
         Type:            I32
         Mutable:         false
         InitExpr:
           Opcode:          I32_CONST
           Value:           1024
-      - Index:           3
+      - Index:           4
         Type:            I32
         Mutable:         false
         InitExpr:
           Opcode:          I32_CONST
           Value:           1040
-      - Index:           4
+      - Index:           5
         Type:            I32
         Mutable:         false
         InitExpr:
           Opcode:          I32_CONST
           Value:           1040
-      - Index:           5
+      - Index:           6
         Type:            I32
         Mutable:         false
         InitExpr:
           Opcode:          I32_CONST
           Value:           66576
-      - Index:           6
+      - Index:           7
         Type:            I32
         Mutable:         false
         InitExpr:
           Opcode:          I32_CONST
           Value:           1024
-      - Index:           7
+      - Index:           8
         Type:            I32
         Mutable:         false
         InitExpr:
           Opcode:          I32_CONST
           Value:           66576
-      - Index:           8
+      - Index:           9
         Type:            I32
         Mutable:         false
         InitExpr:
           Opcode:          I32_CONST
           Value:           131072
-      - Index:           9
+      - Index:           10
         Type:            I32
         Mutable:         false
         InitExpr:
           Opcode:          I32_CONST
           Value:           0
-      - Index:           10
+      - Index:           11
         Type:            I32
         Mutable:         false
         InitExpr:
           Opcode:          I32_CONST
           Value:           1
-      - Index:           11
+      - Index:           12
         Type:            I32
         Mutable:         false
         InitExpr:
@@ -123,16 +139,16 @@ Sections:
         Index:           0
       - Name:            __wasm_call_ctors
         Kind:            FUNCTION
-        Index:           0
+        Index:           2
       - Name:            add
         Kind:            FUNCTION
-        Index:           1
+        Index:           3
       - Name:            __original_main
         Kind:            FUNCTION
-        Index:           2
+        Index:           4
       - Name:            main
         Kind:            FUNCTION
-        Index:           3
+        Index:           5
       - Name:            str
         Kind:            GLOBAL
         Index:           1
@@ -174,20 +190,20 @@ Sections:
         Index:           11
   - Type:            CODE
     Functions:
-      - Index:           0
+      - Index:           2
         Locals:          []
         Body:            0B
-      - Index:           1
+      - Index:           3
         Locals:
           - Type:            I32
             Count:           1
         Body:            23808080800041106B21022002200036020C20022001360208200228020C20022802086A0F0B
-      - Index:           2
+      - Index:           4
         Locals:
           - Type:            I32
             Count:           2
         Body:            23808080800041106B210020002480808080002000410036020C2000410136020820004102360204200028020820002802041081808080002101200041106A24808080800020010F0B
-      - Index:           3
+      - Index:           5
         Locals:          []
         Body:            1082808080000F0B
   - Type:            DATA
@@ -208,15 +224,19 @@ Sections:
     Name:            name
     FunctionNames:
       - Index:           0
-        Name:            __wasm_call_ctors
+        Name:            foo
       - Index:           1
-        Name:            add
+        Name:            another_import
       - Index:           2
-        Name:            __original_main
+        Name:            __wasm_call_ctors
       - Index:           3
+        Name:            add
+      - Index:           4
+        Name:            __original_main
+      - Index:           5
         Name:            main
     GlobalNames:
-      - Index:           0
+      - Index:           1
         Name:            __stack_pointer
     DataSegmentNames:
       - Index:           0

lldb/test/Shell/Symtab/symtab-wasm.test

@@ -1,16 +1,18 @@
 # RUN: yaml2obj %S/Inputs/simple.wasm.yaml -o %t.wasm
 
 # RUN: %lldb %t.wasm -o 'image dump symtab' | FileCheck %s --check-prefix SYMTAB
+SYMTAB: Code 0x0000000000000000 0x0000000000000000 0x00000000 foo
+SYMTAB: Code 0x0000000000000000 0x0000000000000000 0x00000000 another_import
 SYMTAB: Code 0x0000000000000002 0x0000000000000002 0x00000000 __wasm_call_ctors
 SYMTAB: Code 0x0000000000000005 0x0000000000000029 0x00000000 add
 SYMTAB: Code 0x000000000000002f 0x000000000000004c 0x00000000 __original_main
 SYMTAB: Code 0x000000000000007c 0x0000000000000009 0x00000000 main
 
 # RUN: %lldb %t.wasm -o 'image dump sections' | FileCheck %s --check-prefix SECTIONS
-SECTIONS: 0x0000000000000001 code                   [0x0000000000000000-0x0000000000000085)  ---  0x000001a1 0x00000085 0x00000000 symtab-wasm.test.tmp.wasm.code
-SECTIONS: 0x0000000000000040 wasm-name                                                       ---  0x00000251 0x00000059 0x00000000 symtab-wasm.test.tmp.wasm.name
-SECTIONS: 0x0000000000000100 data                   [0x0000000000000400-0x0000000000000409)  ---  0x00000233 0x00000009 0x00000000 symtab-wasm.test.tmp.wasm..rodata
-SECTIONS: 0x0000000000000200 data                   [0x000000000000040c-0x0000000000000410)  ---  0x00000242 0x00000004 0x00000000 symtab-wasm.test.tmp.wasm..data
+SECTIONS: 0x0000000000000001 code                   [0x0000000000000000-0x0000000000000085)  ---  0x000001d2 0x00000085 0x00000000 symtab-wasm.test.tmp.wasm.code
+SECTIONS: 0x0000000000000040 wasm-name                                                       ---  0x00000282 0x0000006e 0x00000000 symtab-wasm.test.tmp.wasm.name
+SECTIONS: 0x0000000000000100 data                   [0x0000000000000400-0x0000000000000409)  ---  0x00000264 0x00000009 0x00000000 symtab-wasm.test.tmp.wasm..rodata
+SECTIONS: 0x0000000000000200 data                   [0x000000000000040c-0x0000000000000410)  ---  0x00000273 0x00000004 0x00000000 symtab-wasm.test.tmp.wasm..data
 
 # RUN: %lldb %t.wasm -o 'x/s 0x0000000000000400' | FileCheck %s --check-prefix STR
 STR: "data str"