1 /* 2 * Copyright (c) 2021-2024 Huawei Device Co., Ltd. 3 * Licensed under the Apache License, Version 2.0 (the "License"); 4 * you may not use this file except in compliance with the License. 5 * You may obtain a copy of the License at 6 * 7 * http://www.apache.org/licenses/LICENSE-2.0 8 * 9 * Unless required by applicable law or agreed to in writing, software 10 * distributed under the License is distributed on an "AS IS" BASIS, 11 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 12 * See the License for the specific language governing permissions and 13 * limitations under the License. 14 */ 15 16 #ifndef DISASM_LIB_H_INCLUDED 17 #define DISASM_LIB_H_INCLUDED 18 19 #include "macros.h" 20 #include "utils/type_helpers.h" 21 #include "utils/span.h" 22 23 #include "class_data_accessor-inl.h" 24 #include "code_data_accessor-inl.h" 25 #include "debug_data_accessor-inl.h" 26 #include "debug_info_extractor.h" 27 #include "field_data_accessor-inl.h" 28 #include "method_data_accessor-inl.h" 29 #include "literal_data_accessor-inl.h" 30 #include "param_annotations_data_accessor.h" 31 #include "annotation_data_accessor.h" 32 #include "proto_data_accessor-inl.h" 33 #include "file-inl.h" 34 #include "file.h" 35 #include "os/file.h" 36 37 #include "runtime/profiling/profiling-disasm-inl.h" 38 39 #include "assembly-program.h" 40 #include "assembly-ins.h" 41 42 #include "bytecode_instruction-inl.h" 43 #include "libpandabase/macros.h" 44 45 #include <map> 46 #include <memory> 47 #include <string> 48 49 #include "accumulators.h" 50 51 namespace ark::disasm { 52 class Disassembler { 53 public: 54 NO_COPY_SEMANTIC(Disassembler); 55 DEFAULT_MOVE_SEMANTIC(Disassembler); 56 57 Disassembler() = default; ~Disassembler()58 ~Disassembler() 59 { 60 profiling::DestroyProfile(profile_, fileLanguage_); 61 } 62 63 PANDA_PUBLIC_API void Serialize(std::ostream &os, bool addSeparators = false, bool printInformation = false) const; 64 PANDA_PUBLIC_API void Disassemble(std::string_view filenameIn, bool quiet = false, bool skipStrings = false); 65 PANDA_PUBLIC_API void Disassemble(const panda_file::File &file, bool quiet = false, bool skipStrings = false); 66 PANDA_PUBLIC_API void Disassemble(std::unique_ptr<const panda_file::File> &file, bool quiet = false, 67 bool skipStrings = false); 68 PANDA_PUBLIC_API void CollectInfo(); 69 70 PANDA_PUBLIC_API void SetProfile(std::string_view fname); 71 72 PANDA_PUBLIC_API void SetFile(std::unique_ptr<const panda_file::File> &file); 73 PANDA_PUBLIC_API void SetFile(const panda_file::File &file); 74 75 PANDA_PUBLIC_API void Serialize(const pandasm::Function &method, std::ostream &os, bool printInformation = false, 76 panda_file::LineNumberTable *lineTable = nullptr) const; 77 PANDA_PUBLIC_API void GetMethod(pandasm::Function *method, const panda_file::File::EntityId &methodId); 78 GetProgInfo()79 const ProgInfo &GetProgInfo() const 80 { 81 return progInfo_; 82 } 83 84 private: 85 void DisassembleImpl(const bool quiet = false, const bool skipStrings = false); 86 static inline bool IsSystemType(const std::string &typeName); 87 88 void GetRecord(pandasm::Record &record, const panda_file::File::EntityId &recordId); 89 void AddMethodToTables(const panda_file::File::EntityId &methodId); 90 void GetLiteralArray(pandasm::LiteralArray *litArray, size_t index); 91 template <typename T> 92 void FillLiteralArrayData(pandasm::LiteralArray *litArray, const panda_file::LiteralTag &tag, 93 const panda_file::LiteralDataAccessor::LiteralValue &value) const; 94 void FillLiteralData(pandasm::LiteralArray *litArray, const panda_file::LiteralDataAccessor::LiteralValue &value, 95 const panda_file::LiteralTag &tag) const; 96 97 static inline bool IsPandasmFriendly(char c); 98 void GetLiteralArrays(); 99 void GetRecords(); 100 void GetFields(pandasm::Record &record, const panda_file::File::EntityId &recordId); 101 102 void GetField(pandasm::Field &field, const panda_file::FieldDataAccessor &fieldAccessor); 103 void AddExternalFieldsToRecords(); 104 void AddExternalFieldsInfoToRecords(); 105 106 void GetMethods(const panda_file::File::EntityId &recordId); 107 void GetParams(pandasm::Function *method, const panda_file::File::EntityId &protoId) const; 108 IdList GetInstructions(pandasm::Function *method, panda_file::File::EntityId methodId, 109 panda_file::File::EntityId codeId); 110 LabelTable GetExceptions(pandasm::Function *method, panda_file::File::EntityId methodId, 111 panda_file::File::EntityId codeId) const; 112 bool LocateTryBlock(const BytecodeInstruction &bcIns, const BytecodeInstruction &bcInsLast, 113 const panda_file::CodeDataAccessor::TryBlock &tryBlock, 114 pandasm::Function::CatchBlock *catchBlockPa, LabelTable *labelTable, size_t tryIdx) const; 115 bool LocateCatchBlock(const BytecodeInstruction &bcIns, const BytecodeInstruction &bcInsLast, 116 const panda_file::CodeDataAccessor::CatchBlock &catchBlock, 117 pandasm::Function::CatchBlock *catchBlockPa, LabelTable *labelTable, size_t tryIdx, 118 size_t catchIdx) const; 119 120 void GetMetaData(pandasm::Record *record, const panda_file::File::EntityId &recordId) const; 121 void GetMetaData(pandasm::Function *method, const panda_file::File::EntityId &methodId) const; 122 void GetMetaData(pandasm::Field *field, const panda_file::File::EntityId &fieldId) const; 123 124 void GetLanguageSpecificMetadata(); 125 126 std::string AnnotationTagToString(const char tag) const; 127 128 std::string ScalarValueToString(const panda_file::ScalarValue &value, const std::string &type); 129 std::string ArrayValueToString(const panda_file::ArrayValue &value, const std::string &type, size_t idx); 130 131 std::string GetFullMethodName(const panda_file::File::EntityId &methodId) const; 132 std::string GetMethodSignature(const panda_file::File::EntityId &methodId) const; 133 std::string GetFullRecordName(const panda_file::File::EntityId &classId) const; 134 135 void GetRecordInfo(const panda_file::File::EntityId &recordId, RecordInfo *recordInfo) const; 136 void GetMethodInfo(const panda_file::File::EntityId &methodId, MethodInfo *methodInfo) const; 137 void GetInsInfo(panda_file::MethodDataAccessor &mda, const panda_file::File::EntityId &codeId, 138 MethodInfo *methodInfo) const; 139 template <typename T> 140 void SerializeValue(bool isSigned, const pandasm::LiteralArray::Literal &lit, std::ostream &os) const; 141 void Serialize(const std::string &name, const pandasm::LiteralArray &litArray, std::ostream &os) const; 142 void SerializeValues(const pandasm::LiteralArray &litArray, bool isConst, std::ostream &os) const; 143 std::string LiteralTagToString(const panda_file::LiteralTag &tag) const; 144 std::string LiteralValueToString(const pandasm::LiteralArray::Literal &lit) const; 145 void Serialize(const pandasm::Record &record, std::ostream &os, bool printInformation = false) const; 146 void SerializeFields(const pandasm::Record &record, std::ostream &os, bool printInformation, 147 bool isUnion = false) const; 148 void SerializeUnionFields(const pandasm::Record &record, std::ostream &os, bool printInformation) const; 149 void Serialize(const pandasm::Function::CatchBlock &catchBlock, std::ostream &os) const; 150 void Serialize(const pandasm::ItemMetadata &meta, const AnnotationList &annList, std::ostream &os) const; 151 void SerializeLineNumberTable(const panda_file::LineNumberTable &lineNumberTable, std::ostream &os) const; 152 void SerializeLocalVariableTable(const panda_file::LocalVariableTable &localVariableTable, 153 const pandasm::Function &method, std::ostream &os) const; 154 void SerializeLanguage(std::ostream &os) const; 155 void SerializeFilename(std::ostream &os) const; 156 void SerializeLitArrays(std::ostream &os, bool addSeparators = false) const; 157 void SerializeRecords(std::ostream &os, bool addSeparators = false, bool printInformation = false) const; 158 void SerializeMethods(std::ostream &os, bool addSeparators = false, bool printInformation = false) const; 159 void SerializePrintStartInfo(const pandasm::Function &method, std::ostringstream &headerSs) const; 160 void SerializeCheckEnd(const pandasm::Function &method, std::ostream &os, bool printMethodInfo, 161 const MethodInfo *&methodInfo) const; 162 size_t SerializeIfPrintMethodInfo( 163 const pandasm::Function &method, bool printMethodInfo, std::ostringstream &headerSs, 164 const MethodInfo *&methodInfo, 165 std::map<std::string, ark::disasm::MethodInfo>::const_iterator &methodInfoIt) const; 166 167 void EnumerateAnnotations(panda_file::AnnotationDataAccessor &annotationAccessor, AnnotationList &annList, 168 const std::string &type, const panda_file::File::EntityId &annotationId); 169 170 void HandleArrayAnnotation(panda_file::AnnotationDataAccessor &annotationAccessor, AnnotationList &annList, 171 size_t i); 172 173 pandasm::Type PFTypeToPandasmType(const panda_file::Type &type, panda_file::ProtoDataAccessor &pda, 174 size_t &refIdx) const; 175 176 pandasm::Type FieldTypeToPandasmType(const uint32_t &type) const; 177 StringDataToString(panda_file::File::StringData sd)178 static std::string StringDataToString(panda_file::File::StringData sd) 179 { 180 std::string str = std::string(utf::Mutf8AsCString(sd.data)); 181 size_t symPos = 0; 182 while (symPos = str.find_first_of("\a\b\f\n\r\t\v\'\?\\", symPos), symPos != std::string::npos) { 183 std::string sym; 184 switch (str[symPos]) { 185 case '\a': 186 sym = R"(\a)"; 187 break; 188 case '\b': 189 sym = R"(\b)"; 190 break; 191 case '\f': 192 sym = R"(\f)"; 193 break; 194 case '\n': 195 sym = R"(\n)"; 196 break; 197 case '\r': 198 sym = R"(\r)"; 199 break; 200 case '\t': 201 sym = R"(\t)"; 202 break; 203 case '\v': 204 sym = R"(\v)"; 205 break; 206 case '\'': 207 sym = R"(\')"; 208 break; 209 case '\?': 210 sym = R"(\?)"; 211 break; 212 case '\\': 213 sym = R"(\\)"; 214 break; 215 default: 216 UNREACHABLE(); 217 } 218 str = str.replace(symPos, 1, sym); 219 ASSERT(sym.size() == 2U); 220 symPos += 2U; 221 } 222 return str; 223 } 224 225 template <typename T> LiteralIntegralValueToString(const pandasm::LiteralArray::Literal & lit)226 std::string LiteralIntegralValueToString(const pandasm::LiteralArray::Literal &lit) const 227 { 228 std::stringstream res {}; 229 T value = std::get<T>(lit.value); 230 if (lit.IsSigned()) { 231 res << +static_cast<typename std::make_signed<T>::type>(value); 232 } else { 233 res << +value; 234 } 235 return res.str(); 236 } 237 238 pandasm::Opcode BytecodeOpcodeToPandasmOpcode(BytecodeInstruction::Opcode o) const; 239 pandasm::Opcode BytecodeOpcodeToPandasmOpcode(uint8_t o) const; 240 void CollectExternalFields(const panda_file::FieldDataAccessor &fieldAccessor); 241 242 pandasm::Ins BytecodeInstructionToPandasmInstruction(BytecodeInstruction bcIns, 243 panda_file::File::EntityId methodId) const; 244 245 std::string IDToString(BytecodeInstruction bcIns, panda_file::File::EntityId methodId) const; 246 247 ark::panda_file::SourceLang GetRecordLanguage(panda_file::File::EntityId classId) const; 248 249 std::unique_ptr<const panda_file::File> fileHolder_ {nullptr}; 250 const panda_file::File *file_ {}; 251 pandasm::Program prog_ {}; 252 253 ark::panda_file::SourceLang fileLanguage_ = ark::panda_file::SourceLang::PANDA_ASSEMBLY; 254 255 std::map<std::string, panda_file::File::EntityId> recordNameToId_ {}; 256 std::map<std::string, panda_file::File::EntityId> methodNameToId_ {}; 257 258 std::map<std::string, std::vector<pandasm::Field>> externalFieldTable_ {}; 259 std::map<std::string, std::vector<std::string>> externalFieldsInfoTable_ {}; 260 261 ProgAnnotations progAnn_ {}; 262 263 ProgInfo progInfo_ {}; 264 265 profiling::ProfileContainer profile_ {profiling::INVALID_PROFILE}; 266 267 std::unique_ptr<panda_file::DebugInfoExtractor> debugInfoExtractor_ {nullptr}; 268 269 bool quiet_ {false}; 270 bool skipStrings_ {false}; 271 272 #include "disasm_plugins.inc" 273 }; 274 } // namespace ark::disasm 275 276 #endif 277