1 //===- MachO.h - MachO object file implementation ---------------*- C++ -*-===//
3 // The LLVM Compiler Infrastructure
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
8 //===----------------------------------------------------------------------===//
10 // This file declares the MachOObjectFile class, which implement the ObjectFile
11 // interface for MachO files.
13 //===----------------------------------------------------------------------===//
15 #ifndef LLVM_OBJECT_MACHO_H
16 #define LLVM_OBJECT_MACHO_H
18 #include "llvm/ADT/ArrayRef.h"
19 #include "llvm/ADT/SmallVector.h"
20 #include "llvm/ADT/Triple.h"
21 #include "llvm/Object/ObjectFile.h"
22 #include "llvm/Support/MachO.h"
27 /// DiceRef - This is a value type class that represents a single
28 /// data in code entry in the table in a Mach-O object file.
30 DataRefImpl DicePimpl;
31 const ObjectFile *OwningObject;
34 DiceRef() : OwningObject(nullptr) { }
36 DiceRef(DataRefImpl DiceP, const ObjectFile *Owner);
38 bool operator==(const DiceRef &Other) const;
39 bool operator<(const DiceRef &Other) const;
43 std::error_code getOffset(uint32_t &Result) const;
44 std::error_code getLength(uint16_t &Result) const;
45 std::error_code getKind(uint16_t &Result) const;
47 DataRefImpl getRawDataRefImpl() const;
48 const ObjectFile *getObjectFile() const;
50 typedef content_iterator<DiceRef> dice_iterator;
52 class MachOObjectFile : public ObjectFile {
54 struct LoadCommandInfo {
55 const char *Ptr; // Where in memory the load command is.
56 MachO::load_command C; // The command itself.
59 MachOObjectFile(std::unique_ptr<MemoryBuffer> Object, bool IsLittleEndian,
60 bool Is64Bits, std::error_code &EC);
62 void moveSymbolNext(DataRefImpl &Symb) const override;
63 std::error_code getSymbolName(DataRefImpl Symb,
64 StringRef &Res) const override;
67 std::error_code getIndirectName(DataRefImpl Symb, StringRef &Res) const;
69 std::error_code getSymbolAddress(DataRefImpl Symb,
70 uint64_t &Res) const override;
71 std::error_code getSymbolAlignment(DataRefImpl Symb,
72 uint32_t &Res) const override;
73 std::error_code getSymbolSize(DataRefImpl Symb, uint64_t &Res) const override;
74 std::error_code getSymbolType(DataRefImpl Symb,
75 SymbolRef::Type &Res) const override;
76 uint32_t getSymbolFlags(DataRefImpl Symb) const override;
77 std::error_code getSymbolSection(DataRefImpl Symb,
78 section_iterator &Res) const override;
80 void moveSectionNext(DataRefImpl &Sec) const override;
81 std::error_code getSectionName(DataRefImpl Sec,
82 StringRef &Res) const override;
83 std::error_code getSectionAddress(DataRefImpl Sec,
84 uint64_t &Res) const override;
85 std::error_code getSectionSize(DataRefImpl Sec, uint64_t &Res) const override;
86 std::error_code getSectionContents(DataRefImpl Sec,
87 StringRef &Res) const override;
88 std::error_code getSectionAlignment(DataRefImpl Sec,
89 uint64_t &Res) const override;
90 std::error_code isSectionText(DataRefImpl Sec, bool &Res) const override;
91 std::error_code isSectionData(DataRefImpl Sec, bool &Res) const override;
92 std::error_code isSectionBSS(DataRefImpl Sec, bool &Res) const override;
93 std::error_code isSectionRequiredForExecution(DataRefImpl Sec,
94 bool &Res) const override;
95 std::error_code isSectionVirtual(DataRefImpl Sec, bool &Res) const override;
96 std::error_code isSectionZeroInit(DataRefImpl Sec, bool &Res) const override;
97 std::error_code isSectionReadOnlyData(DataRefImpl Sec,
98 bool &Res) const override;
99 std::error_code sectionContainsSymbol(DataRefImpl Sec, DataRefImpl Symb,
100 bool &Result) const override;
101 relocation_iterator section_rel_begin(DataRefImpl Sec) const override;
102 relocation_iterator section_rel_end(DataRefImpl Sec) const override;
104 void moveRelocationNext(DataRefImpl &Rel) const override;
105 std::error_code getRelocationAddress(DataRefImpl Rel,
106 uint64_t &Res) const override;
107 std::error_code getRelocationOffset(DataRefImpl Rel,
108 uint64_t &Res) const override;
109 symbol_iterator getRelocationSymbol(DataRefImpl Rel) const override;
110 std::error_code getRelocationType(DataRefImpl Rel,
111 uint64_t &Res) const override;
113 getRelocationTypeName(DataRefImpl Rel,
114 SmallVectorImpl<char> &Result) const override;
116 getRelocationValueString(DataRefImpl Rel,
117 SmallVectorImpl<char> &Result) const override;
118 std::error_code getRelocationHidden(DataRefImpl Rel,
119 bool &Result) const override;
122 std::error_code getLibraryShortNameByIndex(unsigned Index, StringRef &Res);
124 // TODO: Would be useful to have an iterator based version
125 // of the load command interface too.
127 basic_symbol_iterator symbol_begin_impl() const override;
128 basic_symbol_iterator symbol_end_impl() const override;
131 basic_symbol_iterator getSymbolByIndex(unsigned Index) const;
133 section_iterator section_begin() const override;
134 section_iterator section_end() const override;
136 uint8_t getBytesInAddress() const override;
138 StringRef getFileFormatName() const override;
139 unsigned getArch() const override;
140 Triple getArch(const char **McpuDefault, Triple *ThumbTriple) const;
142 relocation_iterator section_rel_begin(unsigned Index) const;
143 relocation_iterator section_rel_end(unsigned Index) const;
145 dice_iterator begin_dices() const;
146 dice_iterator end_dices() const;
148 // In a MachO file, sections have a segment name. This is used in the .o
149 // files. They have a single segment, but this field specifies which segment
150 // a section should be put in in the final object.
151 StringRef getSectionFinalSegmentName(DataRefImpl Sec) const;
153 // Names are stored as 16 bytes. These returns the raw 16 bytes without
154 // interpreting them as a C string.
155 ArrayRef<char> getSectionRawName(DataRefImpl Sec) const;
156 ArrayRef<char> getSectionRawFinalSegmentName(DataRefImpl Sec) const;
158 // MachO specific Info about relocations.
159 bool isRelocationScattered(const MachO::any_relocation_info &RE) const;
160 unsigned getPlainRelocationSymbolNum(
161 const MachO::any_relocation_info &RE) const;
162 bool getPlainRelocationExternal(const MachO::any_relocation_info &RE) const;
163 bool getScatteredRelocationScattered(
164 const MachO::any_relocation_info &RE) const;
165 uint32_t getScatteredRelocationValue(
166 const MachO::any_relocation_info &RE) const;
167 unsigned getAnyRelocationAddress(const MachO::any_relocation_info &RE) const;
168 unsigned getAnyRelocationPCRel(const MachO::any_relocation_info &RE) const;
169 unsigned getAnyRelocationLength(const MachO::any_relocation_info &RE) const;
170 unsigned getAnyRelocationType(const MachO::any_relocation_info &RE) const;
171 SectionRef getRelocationSection(const MachO::any_relocation_info &RE) const;
173 // Walk load commands.
174 LoadCommandInfo getFirstLoadCommandInfo() const;
175 LoadCommandInfo getNextLoadCommandInfo(const LoadCommandInfo &L) const;
177 // MachO specific structures.
178 MachO::section getSection(DataRefImpl DRI) const;
179 MachO::section_64 getSection64(DataRefImpl DRI) const;
180 MachO::section getSection(const LoadCommandInfo &L, unsigned Index) const;
181 MachO::section_64 getSection64(const LoadCommandInfo &L,unsigned Index) const;
182 MachO::nlist getSymbolTableEntry(DataRefImpl DRI) const;
183 MachO::nlist_64 getSymbol64TableEntry(DataRefImpl DRI) const;
185 MachO::linkedit_data_command
186 getLinkeditDataLoadCommand(const LoadCommandInfo &L) const;
187 MachO::segment_command
188 getSegmentLoadCommand(const LoadCommandInfo &L) const;
189 MachO::segment_command_64
190 getSegment64LoadCommand(const LoadCommandInfo &L) const;
191 MachO::linker_options_command
192 getLinkerOptionsLoadCommand(const LoadCommandInfo &L) const;
193 MachO::version_min_command
194 getVersionMinLoadCommand(const LoadCommandInfo &L) const;
196 getDylibIDLoadCommand(const LoadCommandInfo &L) const;
198 MachO::any_relocation_info getRelocation(DataRefImpl Rel) const;
199 MachO::data_in_code_entry getDice(DataRefImpl Rel) const;
200 MachO::mach_header getHeader() const;
201 MachO::mach_header_64 getHeader64() const;
203 getIndirectSymbolTableEntry(const MachO::dysymtab_command &DLC,
204 unsigned Index) const;
205 MachO::data_in_code_entry getDataInCodeTableEntry(uint32_t DataOffset,
206 unsigned Index) const;
207 MachO::symtab_command getSymtabLoadCommand() const;
208 MachO::dysymtab_command getDysymtabLoadCommand() const;
209 MachO::linkedit_data_command getDataInCodeLoadCommand() const;
211 StringRef getStringTableData() const;
212 bool is64Bit() const;
213 void ReadULEB128s(uint64_t Index, SmallVectorImpl<uint64_t> &Out) const;
215 static StringRef guessLibraryShortName(StringRef Name, bool &isFramework,
218 static Triple::ArchType getArch(uint32_t CPUType);
219 static Triple getArch(uint32_t CPUType, uint32_t CPUSubType,
220 const char **McpuDefault = nullptr);
221 static Triple getThumbArch(uint32_t CPUType, uint32_t CPUSubType,
222 const char **McpuDefault = nullptr);
223 static Triple getArch(uint32_t CPUType, uint32_t CPUSubType,
224 const char **McpuDefault, Triple *ThumbTriple);
225 static bool isValidArch(StringRef ArchFlag);
226 static Triple getHostArch();
228 bool isRelocatableObject() const override;
230 static bool classof(const Binary *v) {
235 typedef SmallVector<const char*, 1> SectionList;
236 SectionList Sections;
237 typedef SmallVector<const char*, 1> LibraryList;
238 LibraryList Libraries;
239 typedef SmallVector<StringRef, 1> LibraryShortName;
240 LibraryShortName LibrariesShortNames;
241 const char *SymtabLoadCmd;
242 const char *DysymtabLoadCmd;
243 const char *DataInCodeLoadCmd;
247 inline DiceRef::DiceRef(DataRefImpl DiceP, const ObjectFile *Owner)
248 : DicePimpl(DiceP) , OwningObject(Owner) {}
250 inline bool DiceRef::operator==(const DiceRef &Other) const {
251 return DicePimpl == Other.DicePimpl;
254 inline bool DiceRef::operator<(const DiceRef &Other) const {
255 return DicePimpl < Other.DicePimpl;
258 inline void DiceRef::moveNext() {
259 const MachO::data_in_code_entry *P =
260 reinterpret_cast<const MachO::data_in_code_entry *>(DicePimpl.p);
261 DicePimpl.p = reinterpret_cast<uintptr_t>(P + 1);
264 // Since a Mach-O data in code reference, a DiceRef, can only be created when
265 // the OwningObject ObjectFile is a MachOObjectFile a static_cast<> is used for
266 // the methods that get the values of the fields of the reference.
268 inline std::error_code DiceRef::getOffset(uint32_t &Result) const {
269 const MachOObjectFile *MachOOF =
270 static_cast<const MachOObjectFile *>(OwningObject);
271 MachO::data_in_code_entry Dice = MachOOF->getDice(DicePimpl);
272 Result = Dice.offset;
273 return object_error::success;
276 inline std::error_code DiceRef::getLength(uint16_t &Result) const {
277 const MachOObjectFile *MachOOF =
278 static_cast<const MachOObjectFile *>(OwningObject);
279 MachO::data_in_code_entry Dice = MachOOF->getDice(DicePimpl);
280 Result = Dice.length;
281 return object_error::success;
284 inline std::error_code DiceRef::getKind(uint16_t &Result) const {
285 const MachOObjectFile *MachOOF =
286 static_cast<const MachOObjectFile *>(OwningObject);
287 MachO::data_in_code_entry Dice = MachOOF->getDice(DicePimpl);
289 return object_error::success;
292 inline DataRefImpl DiceRef::getRawDataRefImpl() const {
296 inline const ObjectFile *DiceRef::getObjectFile() const {