1 /* 2 * Copyright (c) 2021-2022 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 DISASSEMBLER_DISASSEMBLER_H 17 #define DISASSEMBLER_DISASSEMBLER_H 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 "libpandafile/module_data_accessor-inl.h" 31 #include "param_annotations_data_accessor.h" 32 #include "annotation_data_accessor.h" 33 #include "proto_data_accessor-inl.h" 34 #include "file-inl.h" 35 #include "file.h" 36 #include "os/file.h" 37 38 #include "assembly-program.h" 39 #include "assembly-ins.h" 40 41 #include "bytecode_instruction-inl.h" 42 #include "libpandabase/macros.h" 43 44 #include <map> 45 #include <memory> 46 #include <string> 47 48 #include "accumulators.h" 49 50 namespace panda::disasm { 51 class Disassembler { 52 public: 53 NO_COPY_SEMANTIC(Disassembler); 54 DEFAULT_MOVE_SEMANTIC(Disassembler); 55 56 Disassembler() = default; 57 ~Disassembler() = default; 58 59 void Disassemble(const std::string &filename_in, const bool quiet = false, const bool skip_strings = false); 60 void CollectInfo(); 61 void Serialize(std::ostream &os, bool add_separators = false, bool print_information = false) const; 62 63 static inline bool IsPandasmFriendly(const char c); 64 static inline bool IsSystemType(const std::string &type_name); 65 66 void GetRecord(pandasm::Record *record, const panda_file::File::EntityId &record_id); 67 void AddMethodToTables(const panda_file::File::EntityId &method_id); 68 void GetMethod(pandasm::Function *method, const panda_file::File::EntityId &method_id); 69 void GetLiteralArray(pandasm::LiteralArray *lit_array, size_t index) const; 70 std::optional<std::vector<std::string>> GetAnnotationByMethodName(const std::string &method_name) const; 71 std::vector<std::string> GetStrings() const; 72 std::vector<std::string> GetModuleLiterals() const; 73 template <typename T> 74 void FillLiteralArrayData(pandasm::LiteralArray *lit_array, const panda_file::LiteralTag &tag, 75 const panda_file::LiteralDataAccessor::LiteralValue &value) const; 76 pandasm::Ins BytecodeInstructionToPandasmInstruction(BytecodeInstruction bc_ins, 77 panda_file::File::EntityId method_id) const; 78 GetProgInfo()79 const ProgInfo &GetProgInfo() const 80 { 81 return prog_info_; 82 } 83 84 std::vector<size_t> GetColumnNumber(); 85 std::vector<size_t> GetLineNumber(); 86 87 private: 88 void GetLiteralArrays(); 89 void FillLiteralData(pandasm::LiteralArray *lit_array, const panda_file::LiteralDataAccessor::LiteralValue &value, 90 const panda_file::LiteralTag &tag) const; 91 void GetRecords(); 92 void GetFields(pandasm::Record *record, const panda_file::File::EntityId &record_id); 93 94 void GetMethods(const panda_file::File::EntityId &record_id); 95 void GetMethodAnnotations(pandasm::Function &method, const panda_file::File::EntityId &method_id); 96 void CreateAnnotationElement(panda_file::AnnotationDataAccessor &ada, pandasm::Function &method, 97 const std::string &ann_name, const std::string &ann_elem_name, 98 const std::string &ann_elem_index); 99 void AddAnnotationElement(pandasm::Function &method, const std::string &annotation_name, 100 const std::string &key, const uint32_t &value); 101 void GetParams(pandasm::Function *method, const panda_file::File::EntityId &code_id) const; 102 IdList GetInstructions(pandasm::Function *method, panda_file::File::EntityId method_id, 103 panda_file::File::EntityId code_id) const; 104 LabelTable GetExceptions(pandasm::Function *method, panda_file::File::EntityId method_id, 105 panda_file::File::EntityId code_id) const; 106 bool LocateTryBlock(const BytecodeInstruction &bc_ins, const BytecodeInstruction &bc_ins_last, 107 const panda_file::CodeDataAccessor::TryBlock &try_block, 108 pandasm::Function::CatchBlock *catch_block_pa, LabelTable *label_table, size_t try_idx) const; 109 bool LocateCatchBlock(const BytecodeInstruction &bc_ins, const BytecodeInstruction &bc_ins_last, 110 const panda_file::CodeDataAccessor::CatchBlock &catch_block, 111 pandasm::Function::CatchBlock *catch_block_pa, LabelTable *label_table, size_t try_idx, 112 size_t catch_idx) const; 113 114 void GetMetaData(pandasm::Record *record, const panda_file::File::EntityId &record_id) const; 115 void GetMetaData(pandasm::Function *method, const panda_file::File::EntityId &method_id) const; 116 void GetMetaData(pandasm::Field *field, const panda_file::File::EntityId &field_id, bool is_scope_names_record); 117 118 void GetLanguageSpecificMetadata(); 119 120 std::string AnnotationTagToString(const char tag) const; 121 122 std::string ScalarValueToString(const panda_file::ScalarValue &value, const std::string &type); 123 std::string ArrayValueToString(const panda_file::ArrayValue &value, const std::string &type, const size_t idx); 124 125 std::string GetFullMethodName(const panda_file::File::EntityId &method_id) const; 126 std::string GetMethodSignature(const panda_file::File::EntityId &method_id) const; 127 std::string GetFullRecordName(const panda_file::File::EntityId &class_id) const; 128 129 void GetRecordInfo(const panda_file::File::EntityId &record_id, RecordInfo *record_info) const; 130 void GetMethodInfo(const panda_file::File::EntityId &method_id, MethodInfo *method_info) const; 131 void GetInsInfo(const panda_file::File::EntityId &code_id, MethodInfo *method_info) const; 132 133 template <typename T> 134 void SerializeValues(const pandasm::LiteralArray &lit_array, T &os) const; 135 std::string SerializeLiteralArray(const pandasm::LiteralArray &lit_array) const; 136 void Serialize(const std::string &key, const pandasm::LiteralArray &lit_array, std::ostream &os) const; 137 void Serialize(const std::string &module_offset, const std::vector<std::string> &module_array, 138 std::ostream &os) const; 139 template <typename T> 140 void SerializeLiterals(const pandasm::LiteralArray &lit_array, T &os) const; 141 std::string LiteralTagToString(const panda_file::LiteralTag &tag) const; 142 void Serialize(const pandasm::Record &record, std::ostream &os, bool print_information = false) const; 143 void SerializeFields(const pandasm::Record &record, std::ostream &os, bool print_information) const; 144 void Serialize(const pandasm::Function &method, std::ostream &os, bool print_information = false) const; 145 void SerializeInstructions(const pandasm::Function &method, std::ostream &os, 146 const std::map<std::string, MethodInfo>::const_iterator &method_info_it, 147 bool print_method_info = false) const; 148 void SerializeMethodAnnotations(const pandasm::Function &method, std::ostream &os) const; 149 void SerializeStrings(const panda_file::File::EntityId &offset, const std::string &name_value, 150 std::ostream &os) const; 151 void Serialize(const pandasm::Function::CatchBlock &catch_block, std::ostream &os) const; 152 void Serialize(const pandasm::ItemMetadata &meta, const AnnotationList &ann_list, std::ostream &os) const; 153 void SerializeLineNumberTable(const panda_file::LineNumberTable &line_number_table, std::ostream &os) const; 154 void SerializeColumnNumberTable(const panda_file::ColumnNumberTable &column_number_table, 155 std::ostream &os) const; 156 void SerializeLocalVariableTable(const panda_file::LocalVariableTable &local_variable_table, 157 const pandasm::Function &method, std::ostream &os) const; 158 bool IsModuleLiteralOffset(const panda_file::File::EntityId &id) const; SerializeLanguage(std::ostream & os)159 inline void SerializeLanguage(std::ostream &os) const 160 { 161 os << ".language " << panda::panda_file::LanguageToString(file_language_) << "\n\n"; 162 } 163 164 pandasm::Type PFTypeToPandasmType(const panda_file::Type &type, panda_file::ProtoDataAccessor &pda, 165 size_t &ref_idx) const; 166 167 pandasm::Type FieldTypeToPandasmType(const uint32_t &type) const; 168 StringDataToString(panda_file::File::StringData sd)169 static inline std::string StringDataToString(panda_file::File::StringData sd) 170 { 171 return std::string(utf::Mutf8AsCString(sd.data)); 172 } 173 174 pandasm::Opcode BytecodeOpcodeToPandasmOpcode(BytecodeInstruction::Opcode o) const; 175 pandasm::Opcode BytecodeOpcodeToPandasmOpcode(uint8_t o) const; 176 177 std::string IDToString(BytecodeInstruction bc_ins, panda_file::File::EntityId method_id, size_t idx) const; 178 179 panda::panda_file::SourceLang GetRecordLanguage(panda_file::File::EntityId class_id) const; 180 void GetLiteralArrayByOffset(pandasm::LiteralArray *lit_array, panda_file::File::EntityId offset) const; 181 182 std::vector<std::string> GetModuleLiteralArray(panda_file::File::EntityId &module_id) const; 183 std::string SerializeModuleLiteralArray(const std::vector<std::string> &module_array) const; 184 std::string ModuleTagToString(panda_file::ModuleTag &tag) const; 185 186 std::unique_ptr<const panda_file::File> file_; 187 pandasm::Program prog_; 188 GetStringByOffset(uint32_t offset)189 inline std::string GetStringByOffset(uint32_t offset) const 190 { 191 const auto sd = file_->GetStringData(panda_file::File::EntityId(offset)); 192 return std::string(utf::Mutf8AsCString(sd.data)); 193 } 194 IsValidOffset(uint32_t offset)195 inline bool IsValidOffset(uint32_t offset) const 196 { 197 return panda_file::File::EntityId(offset).IsValid() && offset < file_->GetHeader()->file_size; 198 } 199 GetFileNameByPath(const std::string & file_path)200 inline std::string GetFileNameByPath(const std::string &file_path) const 201 { 202 size_t pos = file_path.find_last_of(panda::os::file::File::GetPathDelim()); 203 if (pos == std::string::npos) { 204 return file_path; 205 } 206 207 return file_path.substr(pos + 1); 208 } 209 210 panda::panda_file::SourceLang file_language_ = panda::panda_file::SourceLang::PANDA_ASSEMBLY; 211 212 std::map<std::string, panda_file::File::EntityId> record_name_to_id_; 213 std::map<std::string, panda_file::File::EntityId> method_name_to_id_; 214 std::map<std::string, std::vector<std::string>> modulearray_table_; 215 mutable std::map<panda_file::File::EntityId, std::string> string_offset_to_name_; 216 217 ProgAnnotations prog_ann_; 218 219 ProgInfo prog_info_; 220 221 std::unique_ptr<panda_file::DebugInfoExtractor> debug_info_extractor_; 222 223 bool quiet_; 224 bool skip_strings_; 225 std::unordered_set<uint32_t> module_literals_; 226 std::unordered_set<uint32_t> module_request_phase_literals_; 227 #include "disasm_plugins.inc" 228 }; 229 } // namespace panda::disasm 230 231 #endif // DISASSEMBLER_DISASSEMBLER_H 232